.torch_dynamo¶
modules
onnx_custom_backend¶
- experimental_experiment.torch_dynamo.onnx_custom_backend(graph_module: torch.fx.GraphModule, args: List[torch.Tensor], target_opset: int | None = None, backend: str = 'ort', verbose: int | Tuple[int, int] = 0, dump_prefix: None = None, dump_patterns: str | None = None, providers: Tuple[str] | None = None, raise_exc: bool = True, storage: Dict[str, Any] | None = None, enable_pattern: str | List[str | type] | None = 'default', disable_pattern: str | List[str | type] | None = None, pre_ort_model_transforms: Callable[[ModelProto], ModelProto] | List[Callable[[ModelProto], ModelProto]] | None = None, ort_optimization_level: str | None = None, dispatcher: Dispatcher | None = None, rename_inputs: bool = True, optimize: bool = True, exporter: str | None = None, processor: str = 'CPU', order_algorithm: str | None = None, options: OptimizationOptions | None = None, export_options: str | ExportOptions | None = None) Callable[source]¶
Custom backend to export torch models into onnx (see torch.compiler). This backend relies on onnxruntime and tries to be as efficient as possible.
- Parameters:
graph_module – graph to export
args – arguments
target_opset – opset to use for the conversion
backend – only ‘ort’ is allowed
verbose – adjust verbosity, if tuple, if gives different verbosity level to the exporter and the runtime
dump_prefix – to dump the models and the inputs
dump_patterns – dump the patterns as well
providers – where to run the model, by default
raise_exc – raise an exception whenever something goes wrong
storage – to store any interesting objects during the process
enable_pattern – optimization patterns to enable
disable_pattern – optimization patterns to disable
pre_ort_model_transforms – list of transformations applied on the final ModelProto
ort_optimization_level – graph optimization level for onnxruntime, the default value is the same as what onnxruntime defines
dispatcher – see
experimental_experiment.torch_interpreter.Dispatcherrename_inputs – rename the inputs
optimize – enable or disable the optimization
exporter – use a different exporter
processor – optimization should be made for this processor or this list of processors (comma separated value)
order_algorithm – algorithm optimizing the order the onnx node, none by default
options – to define custom Optimization options, in that case, any other optimization parameter is ignored
export_options – see
ExportOptions
- Returns:
Callable
See 301: Compares LLAMA exporters for onnxrt backend or 101: A custom backend for torch for examples. If not empty, storage keeps the memory of the data generated, onnx models, graph module as well the inputs and outputs when the model is run.
The following example shows how to use the custom backend (based on onnxruntime).
<<<
import torch from experimental_experiment.torch_dynamo import onnx_custom_backend class MLP(torch.nn.Module): def __init__(self): super().__init__() self.layers = torch.nn.Sequential( torch.nn.Linear(10, 32), torch.nn.Sigmoid(), torch.nn.Linear(32, 1), ) def forward(self, x): return self.layers(x) x = torch.randn(3, 10, dtype=torch.float32) mlp = MLP() expected = mlp(x) compiled_model = torch.compile( mlp, backend=lambda *args, **kwargs: onnx_custom_backend(*args, verbose=1, **kwargs), dynamic=False, fullgraph=True, ) try: got = compiled_model(x) diff = (expected - got).max() print(f"discrepancies: {diff}") except (ImportError, AttributeError) as e: print("onnxruntime-training is not installed", e)
>>>
[onnx_custom_backend] starts conversion to onnx. [to_onnx] build the graph module from <class 'torch.fx.graph_module.GraphModule.__new__.<locals>.GraphModuleImpl'>, type(args)=<class 'tuple'> [to_onnx] build the graph module with input_names=['input0', 'input1', 'input2', 'input3', 'input4'] [_make_builder_interpreter] use existing <class 'torch.fx.graph_module.GraphModule.__new__.<locals>.GraphModuleImpl'> [to_onnx] graph module done in 0.0004353780022938736 s [to_onnx] start creating the onnx nodes [to_onnx] interpreter.function_options=FunctionOptions(export_as_function=True, name='*', domain='*', external_threshold=256, move_initializer_to_constant=True, return_initializer=True, merge_allowed=True, rename_allowed=True) [to_onnx] 13 onnx nodes done in 0.0013873110001441091 s [to_onnx] start conversion to onnx (before optimization) mask_outputs=None [GraphBuilder-DBC._add_shape_information] dynamic shapes replacements={} [GraphBuilder-DBC.optimize] start with 13 nodes [GraphBuilder-DBC.optimize] #patterns=72 [GraphBuilder-DBC.optimize] start with subgraphs [GraphBuilder-DBC.optimize] done with subgraphs [GraphBuilderPatternOptimization-DBC.optimize] start with 7 nodes, 0 initializers, 72 patterns, priorities=[0, 1, 3], max_iter=30 [GraphBuilderPatternOptimization-DBC.optimize] iteration 0: 7 nodes, priority=0 [GraphBuilderPatternOptimization-DBC.optimize] increase priority to 1 [GraphBuilderPatternOptimization-DBC.optimize] iteration 1: 7 nodes, priority=1 [GraphBuilderPatternOptimization-DBC.optimize] applies 2 matches, 1*TransposeEqualReshapePattern, 1*TransposeMatMulPattern - time=0.001 | max_time=TransposeMatMulPattern:0.000 [GraphBuilderPatternOptimization-DBC.optimize] iteration 2: 6 nodes, priority=1 [GraphBuilderPatternOptimization-DBC.optimize] increase priority to 3 [GraphBuilderPatternOptimization-DBC.optimize] iteration 3: 6 nodes, priority=3 [GraphBuilderPatternOptimization-DBC.optimize] applies 2 matches, 2*MatMulAddPattern - time=0.001 | max_time=MatMulAddPattern:0.000 [GraphBuilderPatternOptimization-DBC.optimize] iteration 4: 4 nodes, priority=3 [GraphBuilderPatternOptimization-DBC.optimize] stops current_priority_index=3, priorities=[0, 1, 3] [GraphBuilderPatternOptimization-DBC.optimize] done after 5 iterations with 4 nodes in 0.007 [GraphBuilder-DBC.optimize] done with 4 nodes in 0.007 [GraphBuilder-DBC.to_onnx] make_model 1 inits 0 params [GraphBuilder-DBC.time_evaluation_constants_] 0 [GraphBuilder-DBC._build_initializers] start with 1 initializers, large_model=False, external_threshold=1024 [GraphBuilder-DBC._build_initializers] switch low/high order [GraphBuilder-DBC._build_initializers] done in 7.499984349124134e-07s with 1 initializers, 0 large initializers [GraphBuilder-DBC._add_shape_information] dynamic shapes replacements={} [to_onnx] to_onnx done in 0.008354208002856467s and 4 nodes, 1 initializers, 5 inputs, 1 outputs [onnx_custom_backend] to_onnx done in 0.010672976000932977 with 4 nodes and 0 local functions. [onnx_custom_backend] starts creating InferenceSession [onnx_custom_backend] InferenceSession done in 0.0038633379954262637 discrepancies: 0.0
onnx_debug_backend¶
- experimental_experiment.torch_dynamo.onnx_debug_backend(graph_module: torch.fx.GraphModule, args: List[torch.Tensor | torch.SymInt | torch.SymFloat], target_opset: int | None = None, backend: str | Callable[[ModelProto, bool | None], Any] = 'ort', verbose: int | Tuple[int, int] = 0, dump_prefix: None = None, dump_patterns: str | None = None, providers: Tuple[str] | None = None, raise_exc: bool = True, storage: Dict[str, Any] | None = None, raise_list: Set[str] | None = None, enable_pattern: str | List[str | type] | None = 'default', disable_pattern: str | List[str | type] | None = None, pre_ort_model_transforms: Callable[[ModelProto], ModelProto] | List[Callable[[ModelProto], ModelProto]] | None = None, ort_optimization_level: str | None = None, dispatcher: Dispatcher | None = None, rename_inputs: bool = True, optimize: bool = True, processor: str = 'CPU', order_algorithm: str | None = None) Callable[source]¶
Custom backend to export torch models into onnx (see torch.compiler). This backend is not meant to be efficient, it is more to check the conversion is ok. It relies either on onnxruntime or the python reference implementation.
- Parameters:
graph_module – graph to export
args – arguments
target_opset – opset to use for the conversion
backend – after the conversion, the model is executed with a runtime, onnxruntime or the reference implementation, it must be a value among ‘ort’, ‘ref’ or a class, it can be a function as well which returns an object behaving the same way
verbose – adjust verbosity, if tuple, if gives different verbosity level to the exporter and the runtime
dump_prefix – prefix used to dump the model generated by the backend
dump_patterns – dump the patterns as well
providers – where to run the model, by default
raise_exc – raise an exception whenever something goes wrong
storage – to store any interesting objects during the process
raise_list – the builder stops any time a name falls into that list, this is a debbuging tool
enable_pattern – optimization patterns to enable
disable_pattern – optimization patterns to disable
pre_ort_model_transforms – list of transformations applied on the final ModelProto
ort_optimization_level – graph optimization level for onnxruntime, the default value is the same as what onnxruntime defines
dispatcher – see
experimental_experiment.torch_interpreter.Dispatcherrename_inputs – rename inputs into
input_{i}optimize – enable or disable the optimization
processor – specifies the processor it is optimized for
order_algorithm – algorithm optimizing the order the onnx node, none by default
- Returns:
Callable
See 301: Compares LLAMA exporters for onnxrt backend for an example. If not empty, storage keeps the memory of the data generated, onnx models, graph module as well the inputs and outputs when the model is run.
The following example shows how to use the reference implementation (
experimental_experiment.reference.ExtendedReferenceEvaluator) to run the onnx model and display the intermediate results.<<<
import torch from experimental_experiment.torch_dynamo import onnx_debug_backend class MLP(torch.nn.Module): def __init__(self): super().__init__() self.layers = torch.nn.Sequential( torch.nn.Linear(10, 32), torch.nn.Sigmoid(), torch.nn.Linear(32, 1), ) def forward(self, x): return self.layers(x) x = torch.randn(3, 10, dtype=torch.float32) mlp = MLP() expected = mlp(x) compiled_model = torch.compile( mlp, backend=lambda *args, **kwargs: onnx_debug_backend( *args, verbose=(1, 10), backend="ref", **kwargs ), dynamic=False, fullgraph=True, ) got = compiled_model(x) diff = (expected - got).max() print(f"discrepancies: {diff}")
>>>
[to_onnx] build the graph module from <class 'torch.fx.graph_module.GraphModule.__new__.<locals>.GraphModuleImpl'>, type(args)=<class 'tuple'> [to_onnx] build the graph module with input_names=['input0', 'input1', 'input2', 'input3', 'input4'] [_make_builder_interpreter] use existing <class 'torch.fx.graph_module.GraphModule.__new__.<locals>.GraphModuleImpl'> [to_onnx] graph module done in 0.0006473399989772588 s [to_onnx] start creating the onnx nodes [to_onnx] interpreter.function_options=FunctionOptions(export_as_function=True, name='*', domain='*', external_threshold=256, move_initializer_to_constant=True, return_initializer=True, merge_allowed=True, rename_allowed=True) [to_onnx] 13 onnx nodes done in 0.0020159810010227375 s [to_onnx] start conversion to onnx (before optimization) mask_outputs=None [GraphBuilder-QNG._add_shape_information] dynamic shapes replacements={} [GraphBuilder-QNG.optimize] start with 13 nodes [GraphBuilder-QNG.optimize] #patterns=72 [GraphBuilder-QNG.optimize] start with subgraphs [GraphBuilder-QNG.optimize] done with subgraphs [GraphBuilderPatternOptimization-QNG.optimize] start with 7 nodes, 0 initializers, 72 patterns, priorities=[0, 1, 3], max_iter=30 [GraphBuilderPatternOptimization-QNG.optimize] iteration 0: 7 nodes, priority=0 [GraphBuilderPatternOptimization-QNG.optimize] increase priority to 1 [GraphBuilderPatternOptimization-QNG.optimize] iteration 1: 7 nodes, priority=1 [GraphBuilderPatternOptimization-QNG.optimize] applies 2 matches, 1*TransposeEqualReshapePattern, 1*TransposeMatMulPattern - time=0.002 | max_time=ReshapeMatMulReshapePattern:0.001 [GraphBuilderPatternOptimization-QNG.optimize] iteration 2: 6 nodes, priority=1 [GraphBuilderPatternOptimization-QNG.optimize] increase priority to 3 [GraphBuilderPatternOptimization-QNG.optimize] iteration 3: 6 nodes, priority=3 [GraphBuilderPatternOptimization-QNG.optimize] applies 2 matches, 2*MatMulAddPattern - time=0.001 | max_time=ShapeBasedReshapeIsSqueezePattern:0.000 [GraphBuilderPatternOptimization-QNG.optimize] iteration 4: 4 nodes, priority=3 [GraphBuilderPatternOptimization-QNG.optimize] stops current_priority_index=3, priorities=[0, 1, 3] [GraphBuilderPatternOptimization-QNG.optimize] done after 5 iterations with 4 nodes in 0.017 [GraphBuilder-QNG.optimize] done with 4 nodes in 0.017 [GraphBuilder-QNG.to_onnx] make_model 1 inits 0 params [GraphBuilder-QNG.time_evaluation_constants_] 0 [GraphBuilder-QNG._build_initializers] start with 1 initializers, large_model=False, external_threshold=1024 [GraphBuilder-QNG._build_initializers] switch low/high order [GraphBuilder-QNG._build_initializers] done in 1.0470030247233808e-06s with 1 initializers, 0 large initializers [GraphBuilder-QNG._add_shape_information] dynamic shapes replacements={} [to_onnx] to_onnx done in 0.01890074699622346s and 4 nodes, 1 initializers, 5 inputs, 1 outputs +C init7_s2_-1_1: int64:(2,):[-1, 1] +I input0: float32:(32, 10):0.1540030986070633,-0.2804732918739319,0.08394774794578552,-0.05558472499251366,-0.09353578835725784... +I input1: float32:(32,):-0.06670594960451126,0.06225314736366272,0.13869209587574005,-0.08152256906032562,-0.27877023816108704... +I input2: float32:(3, 10):0.4001654386520386,-0.3588574230670929,-0.18088537454605103,0.44526371359825134,1.5276145935058594... +I input3: float32:(1, 32):-0.07608781009912491,-0.10965229570865631,-0.03746930509805679,-0.007769898045808077,-0.1556672900915146... +I input4: float32:(1,):[-0.018923070281744003] Gemm(input2, input0, input1) -> input_1 + input_1: float32:(3, 32):0.5776692628860474,-0.7482751607894897,-0.08798091113567352,0.2877032160758972,0.4855598509311676... Sigmoid(input_1) -> input_2 + input_2: float32:(3, 32):0.6405309438705444,0.3211972415447235,0.47801893949508667,0.5714337229728699,0.6190598607063293... Reshape(input3, init7_s2_-1_1) -> l_self_modules_layers_modules_2_parameters_weight_::T10 + l_self_modules_layers_modules_2_parameters_weight_::T10: float32:(32, 1):-0.07608781009912491,-0.10965229570865631,-0.03746930509805679,-0.007769898045808077,-0.1556672900915146... Gemm(input_2, l_self_modules_layers_modules_2_parameters_weight_::T10, input4) -> output_0 + output_0: float32:(3, 1):[-0.017499592155218124, -0.02169310674071312, -0.05856211110949516] discrepancies: 1.4901161193847656e-08
dynger_backend¶
- experimental_experiment.torch_dynamo.dynger_backend(graph_module: GraphModule, args: List[Tensor | SymInt | SymFloat], dynamic_shapes: Dict[str, Any] | Tuple[Any] | None = None, optimize: bool = True, verbose: int | Tuple[int, int] = 0) Callable[source]¶
Eager backend for dynamo.
- Parameters:
graph_module – graph to export
args – arguments
optimize – optimize or not, those optimization would be done on the graph module itself
verbose – adjust verbosity, if tuple, if gives different verbosity level to the exporter and the runtime
- Returns:
Callable
Next examples shows how to display intermediate results while executing the graph produced by torch dynamo.
<<<
import torch from experimental_experiment.torch_dynamo import dynger_backend class MLP(torch.nn.Module): def __init__(self): super().__init__() self.layers = torch.nn.Sequential( torch.nn.Linear(10, 32), torch.nn.Sigmoid(), torch.nn.Linear(32, 1), ) def forward(self, x): return self.layers(x) x = torch.randn(3, 10, dtype=torch.float32) mlp = MLP() expected = mlp(x) compiled_model = torch.compile( mlp, backend=lambda *args, **kwargs: dynger_backend(*args, verbose=10, **kwargs), dynamic=False, fullgraph=True, ) got = compiled_model(x) diff = (expected - got).max() print(f"discrepancies: {diff}")
>>>
[dynger_backend] use existing <class 'torch.fx.graph_module.GraphModule.__new__.<locals>.GraphModuleImpl'> [dynger_backend] begin execution with 9 nodes <built-in function linear>((l_x_, l_self_modules_layers_modules_0_parameters_weight_, l_self_modules_layers_modules_0_parameters_bias_)) -> input_1 + input_1: torch.float32:torch.Size([3, 32]):-0.905686616897583,-0.8605302572250366,-0.09786485135555267,-0.111294686794281,-0.32955700159072876... <built-in method sigmoid of type object at 0x7777f8322ba0>((input_1,)) -> input_2 + input_2: torch.float32:torch.Size([3, 32]):0.28788331151008606,0.2972285747528076,0.4755533039569855,0.4722050130367279,0.41834843158721924... <built-in function linear>((input_2, l_self_modules_layers_modules_2_parameters_weight_, l_self_modules_layers_modules_2_parameters_bias_)) -> input_3 + input_3: torch.float32:torch.Size([3, 1]):0.11127372086048126,0.15405121445655823,0.20040298998355865 [dynger_backend] done discrepancies: 0.0
Other functions¶
- experimental_experiment.torch_dynamo.filter_decomposition_table(existing_table: Dict | None = None, filter_fct: Callable[[Any], bool] | None = None) Dict[source]¶
Returns the decomposition table when some conversions because their translation in ONNX is less efficient.
- Parameters:
existing_table – dictionary of decompositions, by default, it is
torch._decomp.decomposition_table.filter_fct – if specified, a decomposition function is remove if the function returns false
- Returns:
new table
import torch from torch._dynamo.backends.common import aot_autograd from experimental_experiment.torch_dynamo import filter_decomposition_table aot_compiler = aot_autograd( fw_compiler=backend_debug, decompositions=filter_decomposition_table() ) compiled_model = torch.compile( model, backend=aot_compiler, dynamic=dynamic, fullgraph=fullgraph, )
The value is:
<<<
import pprint from experimental_experiment.torch_dynamo import filter_decomposition_table pprint.pprint(filter_decomposition_table())
>>>
{<torch._higher_order_ops.out_dtype.OutDtypeOperator object at 0x7776e22f2ed0>: <function out_dtype_decomp at 0x7776e1d97740>, <OpOverload(op='aten.fft_hfftn', overload='default')>: <function hfftn at 0x7776e19d99e0>, <OpOverload(op='aten.fft_hfftn', overload='out')>: <function hfftn at 0x7776e19d99e0>, <OpOverload(op='aten.fft_ifft2', overload='default')>: <function ifft2 at 0x7776e19d96c0>, <OpOverload(op='aten.fft_ifft2', overload='out')>: <function ifft2 at 0x7776e19d96c0>, <OpOverload(op='aten.fft_rfft2', overload='default')>: <function rfft2 at 0x7776e19d8860>, <OpOverload(op='aten.fft_rfft2', overload='out')>: <function rfft2 at 0x7776e19d8860>, <OpOverload(op='aten.fft_irfft2', overload='default')>: <function irfft2 at 0x7776e19d9da0>, <OpOverload(op='aten.fft_irfft2', overload='out')>: <function irfft2 at 0x7776e19d9da0>, <OpOverload(op='aten.fft_hfft2', overload='default')>: <function hfft2 at 0x7776e19da020>, <OpOverload(op='aten.fft_hfft2', overload='out')>: <function hfft2 at 0x7776e19da020>, <OpOverload(op='aten.fft_ihfft2', overload='default')>: <function ihfft2 at 0x7776e19da2a0>, <OpOverload(op='aten.fft_ihfft2', overload='out')>: <function ihfft2 at 0x7776e19da2a0>, <OpOverload(op='aten.fft_fftshift', overload='default')>: <function fftshift at 0x7776e19da340>, <OpOverload(op='aten.fft_ifftshift', overload='default')>: <function ifftshift at 0x7776e19da3e0>, <OpOverload(op='aten.linalg_cross', overload='default')>: <function cross at 0x7776e19dab60>, <OpOverload(op='aten.linalg_cross', overload='out')>: <function cross at 0x7776e19dab60>, <OpOverload(op='aten.linalg_vector_norm', overload='default')>: <function vector_norm at 0x7776e19daf20>, <OpOverload(op='aten.linalg_vector_norm', overload='out')>: <function vector_norm at 0x7776e19daf20>, <OpOverload(op='aten.alpha_dropout', overload='default')>: <function alpha_dropout at 0x7776e19dbec0>, <OpOverload(op='aten.celu', overload='default')>: <function celu at 0x7776e1a04180>, <OpOverload(op='aten.celu', overload='out')>: <function celu at 0x7776e1a04180>, <OpOverload(op='aten.elu', overload='default')>: <function elu at 0x7776e1a04860>, <OpOverload(op='aten.elu', overload='out')>: <function elu at 0x7776e1a04860>, <OpOverload(op='aten.relu', overload='default')>: <function relu at 0x7776e1a04cc0>, <OpOverload(op='aten.relu', overload='out')>: <function relu at 0x7776e1a04cc0>, <OpOverload(op='aten.channel_shuffle', overload='default')>: <function channel_shuffle at 0x7776e1a05300>, <OpOverload(op='aten.channel_shuffle', overload='out')>: <function channel_shuffle at 0x7776e1a05300>, <OpOverload(op='aten.leaky_relu', overload='default')>: <function leaky_relu at 0x7776e1a054e0>, <OpOverload(op='aten.leaky_relu', overload='out')>: <function leaky_relu at 0x7776e1a054e0>, <OpOverload(op='aten.mish', overload='default')>: <function mish at 0x7776e1a05260>, <OpOverload(op='aten.mish', overload='out')>: <function mish at 0x7776e1a05260>, <OpOverload(op='aten.hardshrink', overload='default')>: <function hardshrink at 0x7776e1a06660>, <OpOverload(op='aten.selu', overload='default')>: <function selu at 0x7776e1a059e0>, <OpOverload(op='aten.softshrink', overload='default')>: <function softshrink at 0x7776e1a06980>, <OpOverload(op='aten.softplus', overload='default')>: <function softplus at 0x7776e1a05f80>, <OpOverload(op='aten.softplus', overload='out')>: <function softplus at 0x7776e1a05f80>, <OpOverload(op='aten.hardshrink', overload='out')>: <function hardshrink at 0x7776e1a06660>, <OpOverload(op='aten.softshrink', overload='out')>: <function softshrink at 0x7776e1a06980>, <OpOverload(op='aten.margin_ranking_loss', overload='default')>: <function margin_ranking_loss at 0x7776e1a07100>, <OpOverload(op='aten.hinge_embedding_loss', overload='default')>: <function hinge_embedding_loss at 0x7776e1a07420>, <OpOverload(op='aten.nll_loss', overload='default')>: <function nll_loss at 0x7776e1a079c0>, <OpOverload(op='aten.nll_loss', overload='out')>: <function nll_loss at 0x7776e1a079c0>, <OpOverload(op='aten.huber_loss', overload='default')>: <function huber_loss at 0x7776e1a07d80>, <OpOverload(op='aten.huber_loss', overload='out')>: <function huber_loss at 0x7776e1a07d80>, <OpOverload(op='aten.threshold', overload='default')>: <function threshold at 0x7776e1a200e0>, <OpOverload(op='aten.threshold', overload='out')>: <function threshold at 0x7776e1a200e0>, <OpOverload(op='aten.special_bessel_j0', overload='default')>: <function bessel_j0 at 0x7776e1a220c0>, <OpOverload(op='aten.hardtanh', overload='default')>: <function hardtanh at 0x7776e1a07240>, <OpOverload(op='aten.hardtanh', overload='out')>: <function hardtanh at 0x7776e1a07240>, <OpOverload(op='aten.special_xlog1py', overload='out')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.gelu', overload='default')>: <function gelu at 0x7776e1a20900>, <OpOverload(op='aten.gelu', overload='out')>: <function gelu at 0x7776e1a20900>, <OpOverload(op='aten.special_xlog1py', overload='self_scalar')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.prelu', overload='default')>: <function prelu at 0x7776e1a20b80>, <OpOverload(op='aten.glu', overload='default')>: <function glu at 0x7776e1a21440>, <OpOverload(op='aten.glu', overload='out')>: <function glu at 0x7776e1a21440>, <OpOverload(op='aten.pairwise_distance', overload='default')>: <function pairwise_distance at 0x7776e1a216c0>, <OpOverload(op='aten.pdist', overload='default')>: <function pdist at 0x7776e1a21a80>, <OpOverload(op='aten.pixel_shuffle', overload='default')>: <function pixel_shuffle at 0x7776e1a21d00>, <OpOverload(op='aten.pixel_shuffle', overload='out')>: <function pixel_shuffle at 0x7776e1a21d00>, <OpOverload(op='aten.pixel_unshuffle', overload='default')>: <function pixel_unshuffle at 0x7776e1a21f80>, <OpOverload(op='aten.pixel_unshuffle', overload='out')>: <function pixel_unshuffle at 0x7776e1a21f80>, <OpOverload(op='aten.celu_', overload='default')>: <function celu at 0x7776e1a20a40>, <OpOverload(op='aten.elu_', overload='default')>: <function elu at 0x7776e1a21da0>, <OpOverload(op='aten.mish_', overload='default')>: <function mish at 0x7776e1a22020>, <OpOverload(op='aten.selu_', overload='default')>: <function selu at 0x7776e1a22160>, <OpOverload(op='aten.threshold_', overload='default')>: <function threshold at 0x7776e1a222a0>, <OpOverload(op='aten.special_bessel_j0', overload='out')>: <function bessel_j0 at 0x7776e1a220c0>, <OpOverload(op='aten.special_bessel_j1', overload='default')>: <function bessel_j1 at 0x7776e1a207c0>, <OpOverload(op='aten.special_bessel_j1', overload='out')>: <function bessel_j1 at 0x7776e1a207c0>, <OpOverload(op='aten.special_entr', overload='default')>: <function entr at 0x7776e1a22d40>, <OpOverload(op='aten.special_entr', overload='out')>: <function entr at 0x7776e1a22d40>, <OpOverload(op='aten.special_erfcx', overload='default')>: <function erfcx at 0x7776e1a23100>, <OpOverload(op='aten.special_erfcx', overload='out')>: <function erfcx at 0x7776e1a23100>, <OpOverload(op='aten.special_i0e', overload='default')>: <function i0e at 0x7776e1a236a0>, <OpOverload(op='aten.special_i0e', overload='out')>: <function i0e at 0x7776e1a236a0>, <OpOverload(op='aten.special_i1', overload='default')>: <function i1 at 0x7776e1a23ba0>, <OpOverload(op='aten.special_i1', overload='out')>: <function i1 at 0x7776e1a23ba0>, <OpOverload(op='aten.special_i1e', overload='default')>: <function i1e at 0x7776e1a400e0>, <OpOverload(op='aten.special_i1e', overload='out')>: <function i1e at 0x7776e1a400e0>, <OpOverload(op='aten.special_log_ndtr', overload='default')>: <function log_ndtr at 0x7776e1a404a0>, <OpOverload(op='aten.special_log_ndtr', overload='out')>: <function log_ndtr at 0x7776e1a404a0>, <OpOverload(op='aten.special_xlog1py', overload='other_scalar_out')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.logit', overload='default')>: <function logit at 0x7776e1a40860>, <OpOverload(op='aten.logit', overload='out')>: <function logit at 0x7776e1a40860>, <OpOverload(op='aten.special_xlog1py', overload='default')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.special_xlog1py', overload='other_scalar')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.special_xlog1py', overload='self_scalar_out')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.mvlgamma', overload='default')>: <function multigammaln at 0x7776e1a21ee0>, <OpOverload(op='aten.mvlgamma', overload='out')>: <function multigammaln at 0x7776e1a21ee0>, <OpOverload(op='aten.special_ndtr', overload='default')>: <function ndtr at 0x7776e1a40d60>, <OpOverload(op='aten.special_ndtr', overload='out')>: <function ndtr at 0x7776e1a40d60>, <OpOverload(op='aten.special_ndtri', overload='default')>: <function ndtri at 0x7776e1a41120>, <OpOverload(op='aten.special_ndtri', overload='out')>: <function ndtri at 0x7776e1a41120>, <OpOverload(op='aten.special_spherical_bessel_j0', overload='default')>: <function spherical_bessel_j0 at 0x7776e1a41760>, <OpOverload(op='aten.special_spherical_bessel_j0', overload='out')>: <function spherical_bessel_j0 at 0x7776e1a41760>, <OpOverload(op='aten.special_zeta', overload='default')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='other_scalar')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='self_scalar')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='out')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='self_scalar_out')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='other_scalar_out')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.clamp_max', overload='default')>: <function clamp_max at 0x7776e1ad1120>, <OpOverload(op='aten.clamp_max', overload='Tensor')>: <function clamp_max at 0x7776e1ad1120>, <OpOverload(op='aten.clamp_max', overload='out')>: <function clamp_max at 0x7776e1ad1120>, <OpOverload(op='aten.clamp_max', overload='Tensor_out')>: <function clamp_max at 0x7776e1ad1120>, <OpOverload(op='aten.all', overload='default')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dim')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dims')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dims_out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='all_out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dimname')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dimname_out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.any', overload='default')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dim')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dims')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dims_out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='all_out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dimname')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dimname_out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.std_mean', overload='correction_out')>: <function std_mean at 0x7776e1c656c0>, <OpOverload(op='aten.std_mean', overload='correction_names')>: <function std_mean at 0x7776e1c656c0>, <OpOverload(op='aten.std_mean', overload='names_dim')>: <function std_mean at 0x7776e1c656c0>, <OpOverload(op='aten.std_mean', overload='default')>: <function std_mean at 0x7776e1c656c0>, <OpOverload(op='aten.std', overload='default')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.std', overload='dim')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.std', overload='correction')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.std', overload='correction_names')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.std', overload='names_out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.std', overload='out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.mean', overload='default')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.std', overload='correction_names_out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.mean', overload='dim')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.mean', overload='names_dim')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.mean', overload='dtype_out')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.mean', overload='out')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.mean', overload='names_out')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.native_layer_norm', overload='default')>: <function native_layer_norm at 0x7776e1af5e40>, <OpOverload(op='aten.native_layer_norm', overload='out')>: <function native_layer_norm at 0x7776e1af5e40>, <OpOverload(op='aten.var_mean', overload='default')>: <function var_mean at 0x7776e1ad2160>, <OpOverload(op='aten.var_mean', overload='dim')>: <function var_mean at 0x7776e1ad2160>, <OpOverload(op='aten.var_mean', overload='correction')>: <function var_mean at 0x7776e1ad2160>, <OpOverload(op='aten.var_mean', overload='correction_names')>: <function var_mean at 0x7776e1ad2160>, <OpOverload(op='aten.var_mean', overload='names_dim')>: <function var_mean at 0x7776e1ad2160>, <OpOverload(op='aten.var_mean', overload='correction_out')>: <function var_mean at 0x7776e1ad2160>, <OpOverload(op='aten.addr', overload='default')>: <function addr at 0x7776e1ad3b00>, <OpOverload(op='aten.addr', overload='out')>: <function addr at 0x7776e1ad3b00>, <OpOverload(op='aten.index_select', overload='dimname_out')>: <function index_select at 0x7776e1b10220>, <OpOverload(op='aten.index_select', overload='dimname')>: <function index_select at 0x7776e1b10220>, <OpOverload(op='aten.constant_pad_nd', overload='default')>: <function constant_pad_nd at 0x7776e1af4c20>, <OpOverload(op='aten.constant_pad_nd', overload='out')>: <function constant_pad_nd at 0x7776e1af4c20>, <OpOverload(op='aten.native_group_norm', overload='default')>: <function native_group_norm at 0x7776e1af5760>, <OpOverload(op='aten.expand', overload='default')>: <function expand at 0x7776e1af4e00>, <OpOverload(op='aten.flip', overload='default')>: <function flip at 0x7776e1af5440>, <OpOverload(op='aten.flip', overload='out')>: <function flip at 0x7776e1af5440>, <OpOverload(op='aten.permute', overload='default')>: <function permute at 0x7776e1af56c0>, <OpOverload(op='aten.renorm', overload='default')>: <function renorm at 0x7776e1af4680>, <OpOverload(op='aten.renorm', overload='out')>: <function renorm at 0x7776e1af4680>, <OpOverload(op='aten.repeat', overload='default')>: <function repeat at 0x7776e1af6340>, <OpOverload(op='aten.repeat', overload='out')>: <function repeat at 0x7776e1af6340>, <OpOverload(op='aten.roll', overload='default')>: <function roll at 0x7776e1af6980>, <OpOverload(op='aten.roll', overload='out')>: <function roll at 0x7776e1af6980>, <OpOverload(op='aten.rot90', overload='default')>: <function rot90 at 0x7776e1af6c00>, <OpOverload(op='aten.rot90', overload='out')>: <function rot90 at 0x7776e1af6c00>, <OpOverload(op='aten.stack', overload='default')>: <function stack at 0x7776e1af6f20>, <OpOverload(op='aten.stack', overload='out')>: <function stack at 0x7776e1af6f20>, <OpOverload(op='aten.unbind', overload='int')>: <function unbind at 0x7776e1af7560>, <OpOverload(op='aten.unbind', overload='Dimname')>: <function unbind at 0x7776e1af7560>, <OpOverload(op='aten.split_with_sizes', overload='default')>: <function split_with_sizes at 0x7776e1af7b00>, <OpOverload(op='aten.index_fill', overload='int_Tensor')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill', overload='int_Scalar')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill', overload='Dimname_Tensor')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill', overload='int_Scalar_out')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill', overload='int_Tensor_out')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill_', overload='Dimname_Tensor')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.diag', overload='out')>: <function diag at 0x7776e1b10180>, <OpOverload(op='aten.index_select', overload='default')>: <function index_select at 0x7776e1b10220>, <OpOverload(op='aten.index_select', overload='out')>: <function index_select at 0x7776e1b10220>, <OpOverload(op='aten.t', overload='default')>: <function t at 0x7776e1b10cc0>, <OpOverload(op='aten.diagonal_scatter', overload='out')>: <function diagonal_scatter at 0x7776e1b104a0>, <OpOverload(op='aten.diagonal', overload='Dimname')>: <function diagonal at 0x7776e1b102c0>, <OpOverload(op='aten.diag_embed', overload='default')>: <function diag_embed at 0x7776e1b107c0>, <OpOverload(op='aten.diag_embed', overload='out')>: <function diag_embed at 0x7776e1b107c0>, <OpOverload(op='aten.block_diag', overload='default')>: <function _block_diag_iterable at 0x7776e1b10b80>, <OpOverload(op='aten.alias', overload='default')>: <function alias at 0x7776e1b10ea0>, <OpOverload(op='aten.unfold', overload='default')>: <function unfold at 0x7776e1b11080>, <OpOverload(op='aten.unfold_copy', overload='default')>: <function unfold_copy at 0x7776e1b114e0>, <OpOverload(op='aten.unfold_copy', overload='out')>: <function unfold_copy at 0x7776e1b114e0>, <OpOverload(op='aten.cumsum', overload='default')>: <function cumsum at 0x7776e1b11580>, <OpOverload(op='aten.cumsum', overload='dimname')>: <function cumsum at 0x7776e1b11580>, <OpOverload(op='aten.cumsum', overload='dimname_out')>: <function cumsum at 0x7776e1b11580>, <OpOverload(op='aten.cumsum', overload='out')>: <function cumsum at 0x7776e1b11580>, <OpOverload(op='aten.cumprod', overload='default')>: <function cumprod at 0x7776e1b11620>, <OpOverload(op='aten.cumprod', overload='dimname')>: <function cumprod at 0x7776e1b11620>, <OpOverload(op='aten.cumprod', overload='dimname_out')>: <function cumprod at 0x7776e1b11620>, <OpOverload(op='aten.cumprod', overload='out')>: <function cumprod at 0x7776e1b11620>, <OpOverload(op='aten.unsqueeze', overload='default')>: <function unsqueeze at 0x7776e1b11760>, <OpOverload(op='aten.arange', overload='start_out')>: <function arange at 0x7776e1b12e80>, <OpOverload(op='aten.arange', overload='start_step')>: <function arange at 0x7776e1b12e80>, <OpOverload(op='aten.eye', overload='m')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.eye', overload='m_out')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.eye', overload='out')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.eye', overload='default')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.randn', overload='default')>: <function randn at 0x7776e1b40400>, <OpOverload(op='aten.tril_indices', overload='out')>: <function tril_indices at 0x7776e1b41a80>, <OpOverload(op='aten.lerp', overload='Scalar')>: <function lerp at 0x7776e1b13240>, <OpOverload(op='aten.lerp', overload='Tensor')>: <function lerp at 0x7776e1b13240>, <OpOverload(op='aten.lerp', overload='Scalar_out')>: <function lerp at 0x7776e1b13240>, <OpOverload(op='aten.lerp', overload='Tensor_out')>: <function lerp at 0x7776e1b13240>, <OpOverload(op='aten.linspace', overload='Tensor_Tensor')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Tensor_Scalar')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Scalar_Tensor')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='default')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Tensor_Tensor_out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Scalar_Tensor_out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Tensor_Scalar_out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.logspace', overload='Tensor_Tensor')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Tensor_Scalar')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Scalar_Tensor')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='default')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Tensor_Tensor_out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Tensor_Scalar_out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Scalar_Tensor_out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.meshgrid', overload='default')>: <function meshgrid at 0x7776e1b13920>, <OpOverload(op='aten.meshgrid', overload='indexing')>: <function meshgrid at 0x7776e1b13920>, <OpOverload(op='aten.triu_indices', overload='out')>: <function triu_indices at 0x7776e1b419e0>, <OpOverload(op='aten.triu_indices', overload='default')>: <function triu_indices at 0x7776e1b419e0>, <OpOverload(op='aten.masked_fill', overload='Scalar')>: <function masked_fill at 0x7776e1b407c0>, <OpOverload(op='aten.masked_fill', overload='Tensor')>: <function masked_fill at 0x7776e1b407c0>, <OpOverload(op='aten.masked_fill', overload='Scalar_out')>: <function masked_fill at 0x7776e1b407c0>, <OpOverload(op='aten.masked_fill', overload='Tensor_out')>: <function masked_fill at 0x7776e1b407c0>, <OpOverload(op='aten.masked_fill_', overload='Scalar')>: <function masked_fill_ at 0x7776e1b405e0>, <OpOverload(op='aten.masked_fill_', overload='Tensor')>: <function masked_fill_ at 0x7776e1b405e0>, <OpOverload(op='aten.norm', overload='Scalar')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dim')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='names_ScalarOpt_dim')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='dtype_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dtype')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dim_dtype')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='names_dtype_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dtype_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='Scalar_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='names_ScalarOpt_dim_dtype')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='names_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.trace', overload='default')>: <function trace at 0x7776e1b40f40>, <OpOverload(op='aten.triu', overload='default')>: <function triu at 0x7776e1b41440>, <OpOverload(op='aten.triu', overload='out')>: <function triu at 0x7776e1b41440>, <OpOverload(op='aten.tril', overload='default')>: <function tril at 0x7776e1b416c0>, <OpOverload(op='aten.tril', overload='out')>: <function tril at 0x7776e1b416c0>, <OpOverload(op='aten.tril_indices', overload='default')>: <function tril_indices at 0x7776e1b41a80>, <OpOverload(op='aten.bucketize', overload='Tensor')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.bucketize', overload='Scalar')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.bucketize', overload='Tensor_out')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.bucketize', overload='Scalar_out')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.cauchy', overload='default')>: <function cauchy at 0x7776e1b41c60>, <OpOverload(op='aten.cauchy', overload='out')>: <function cauchy at 0x7776e1b41c60>, <OpOverload(op='aten.exponential', overload='default')>: <function exponential at 0x7776e1b42020>, <OpOverload(op='aten.exponential', overload='out')>: <function exponential at 0x7776e1b42020>, <OpOverload(op='aten.dot', overload='out')>: <function dot at 0x7776e1b41f80>, <OpOverload(op='aten.geometric', overload='default')>: <function geometric at 0x7776e1b423e0>, <OpOverload(op='aten.geometric', overload='out')>: <function geometric at 0x7776e1b423e0>, <OpOverload(op='aten.dot', overload='default')>: <function dot at 0x7776e1b41f80>, <OpOverload(op='aten.log_normal', overload='default')>: <function log_normal at 0x7776e1b427a0>, <OpOverload(op='aten.log_normal', overload='out')>: <function log_normal at 0x7776e1b427a0>, <OpOverload(op='aten.normal_', overload='default')>: <function normal_ at 0x7776e1b42840>, <OpOverload(op='aten.asinh_', overload='default')>: <function asinh at 0x7776e1b645e0>, <OpOverload(op='aten.addcmul_', overload='default')>: <function addcmul at 0x7776e1b64220>, <OpOverload(op='aten.rad2deg', overload='out')>: <function rad2deg at 0x7776e1b431a0>, <OpOverload(op='aten.deg2rad', overload='default')>: <function deg2rad at 0x7776e1b436a0>, <OpOverload(op='aten.deg2rad', overload='out')>: <function deg2rad at 0x7776e1b436a0>, <OpOverload(op='aten.count_nonzero', overload='dim_IntList')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.count_nonzero', overload='dim_IntList_out')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.count_nonzero', overload='default')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.count_nonzero', overload='out')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.cumprod_', overload='default')>: <function cumprod at 0x7776e1b65940>, <OpOverload(op='aten.cumprod_', overload='dimname')>: <function cumprod at 0x7776e1b65940>, <OpOverload(op='aten.vdot', overload='default')>: <function vdot at 0x7776e1b43c40>, <OpOverload(op='aten.vdot', overload='out')>: <function vdot at 0x7776e1b43c40>, <OpOverload(op='aten.le_', overload='Tensor')>: <function le at 0x7776e1b66d40>, <OpOverload(op='aten.select_scatter', overload='out')>: <function select_scatter at 0x7776e1b43ec0>, <OpOverload(op='aten.abs_', overload='default')>: <function abs at 0x7776e1b439c0>, <OpOverload(op='aten.acos_', overload='default')>: <function acos at 0x7776e1b43ce0>, <OpOverload(op='aten.cumsum_', overload='default')>: <function cumsum at 0x7776e1b659e0>, <OpOverload(op='aten.cumsum_', overload='dimname')>: <function cumsum at 0x7776e1b659e0>, <OpOverload(op='aten.acosh_', overload='default')>: <function acosh at 0x7776e1b43f60>, <OpOverload(op='aten.add_', overload='Tensor')>: <function add at 0x7776e1b640e0>, <OpOverload(op='aten.add_', overload='Scalar')>: <function add at 0x7776e1b640e0>, <OpOverload(op='aten.cosh_', overload='default')>: <function cosh at 0x7776e1b42340>, <OpOverload(op='aten.addcdiv_', overload='default')>: <function addcdiv at 0x7776e1b64360>, <OpOverload(op='aten.asin_', overload='default')>: <function asin at 0x7776e1b644a0>, <OpOverload(op='aten.cos_', overload='default')>: <function cos at 0x7776e1b43ba0>, <OpOverload(op='aten.atan_', overload='default')>: <function atan at 0x7776e1b64720>, <OpOverload(op='aten.atanh_', overload='default')>: <function atanh at 0x7776e1b64860>, <OpOverload(op='aten.copysign_', overload='Scalar')>: <function copysign at 0x7776e1b43e20>, <OpOverload(op='aten.atan2_', overload='default')>: <function atan2 at 0x7776e1b649a0>, <OpOverload(op='aten.bitwise_and_', overload='Tensor')>: <function bitwise_and at 0x7776e1b64ae0>, <OpOverload(op='aten.bitwise_and_', overload='Scalar')>: <function bitwise_and at 0x7776e1b64ae0>, <OpOverload(op='aten.copysign_', overload='Tensor')>: <function copysign at 0x7776e1b43e20>, <OpOverload(op='aten.bitwise_left_shift_', overload='Tensor_Scalar')>: <function bitwise_left_shift at 0x7776e1b64c20>, <OpOverload(op='aten.bitwise_left_shift_', overload='Tensor')>: <function bitwise_left_shift at 0x7776e1b64c20>, <OpOverload(op='aten.bitwise_not_', overload='default')>: <function bitwise_not at 0x7776e1b64d60>, <OpOverload(op='aten.bitwise_or_', overload='Tensor')>: <function bitwise_or at 0x7776e1b64ea0>, <OpOverload(op='aten.bitwise_or_', overload='Scalar')>: <function bitwise_or at 0x7776e1b64ea0>, <OpOverload(op='aten.bitwise_right_shift_', overload='Tensor_Scalar')>: <function bitwise_right_shift at 0x7776e1b64fe0>, <OpOverload(op='aten.bitwise_right_shift_', overload='Tensor')>: <function bitwise_right_shift at 0x7776e1b64fe0>, <OpOverload(op='aten.bitwise_xor_', overload='Tensor')>: <function bitwise_xor at 0x7776e1b65120>, <OpOverload(op='aten.bitwise_xor_', overload='Scalar')>: <function bitwise_xor at 0x7776e1b65120>, <OpOverload(op='aten.ceil_', overload='default')>: <function ceil at 0x7776e1b65260>, <OpOverload(op='aten.conj_physical_', overload='default')>: <function conj_physical at 0x7776e1b65760>, <OpOverload(op='aten.clamp_', overload='default')>: <function clamp at 0x7776e1b653a0>, <OpOverload(op='aten.clamp_', overload='Tensor')>: <function clamp at 0x7776e1b653a0>, <OpOverload(op='aten.deg2rad_', overload='default')>: <function deg2rad at 0x7776e1b656c0>, <OpOverload(op='aten.clamp_min_', overload='default')>: <function clamp_min at 0x7776e1b654e0>, <OpOverload(op='aten.clamp_min_', overload='Tensor')>: <function clamp_min at 0x7776e1b654e0>, <OpOverload(op='aten.lt_', overload='Scalar')>: <function lt at 0x7776e1b65c60>, <OpOverload(op='aten.digamma_', overload='default')>: <function digamma at 0x7776e1b65440>, <OpOverload(op='aten.clamp_max_', overload='default')>: <function clamp_max at 0x7776e1b65620>, <OpOverload(op='aten.clamp_max_', overload='Tensor')>: <function clamp_max at 0x7776e1b65620>, <OpOverload(op='aten.div_', overload='Tensor')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.div_', overload='Tensor_mode')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.div_', overload='Scalar')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.div_', overload='Scalar_mode')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.logical_xor_', overload='default')>: <function logical_xor at 0x7776e1b65ee0>, <OpOverload(op='aten.cauchy_', overload='default')>: <function cauchy at 0x7776e1b677e0>, <OpOverload(op='aten.eq_', overload='Scalar')>: <function eq at 0x7776e1b64f40>, <OpOverload(op='aten.eq_', overload='Tensor')>: <function eq at 0x7776e1b64f40>, <OpOverload(op='aten.erf_', overload='default')>: <function erf at 0x7776e1b64cc0>, <OpOverload(op='aten.logical_or_', overload='default')>: <function logical_or at 0x7776e1b66160>, <OpOverload(op='aten.erfc_', overload='default')>: <function erfc at 0x7776e1b64a40>, <OpOverload(op='aten.erfinv_', overload='default')>: <function erfinv at 0x7776e1b647c0>, <OpOverload(op='aten.exponential_', overload='default')>: <function exponential at 0x7776e1b67560>, <OpOverload(op='aten.exp_', overload='default')>: <function exp at 0x7776e1b64540>, <OpOverload(op='aten.logical_not_', overload='default')>: <function logical_not at 0x7776e1b663e0>, <OpOverload(op='aten.exp2_', overload='default')>: <function exp2 at 0x7776e1b642c0>, <OpOverload(op='aten.expm1_', overload='default')>: <function expm1 at 0x7776e1b64040>, <OpOverload(op='aten.logical_and_', overload='default')>: <function logical_and at 0x7776e1b66660>, <OpOverload(op='aten.float_power_', overload='Tensor')>: <function float_power at 0x7776e1b65bc0>, <OpOverload(op='aten.float_power_', overload='Scalar')>: <function float_power at 0x7776e1b65bc0>, <OpOverload(op='aten.floor_', overload='default')>: <function floor at 0x7776e1b65d00>, <OpOverload(op='aten.floor_divide_', overload='Scalar')>: <function floor_divide at 0x7776e1b65e40>, <OpOverload(op='aten.floor_divide_', overload='Tensor')>: <function floor_divide at 0x7776e1b65e40>, <OpOverload(op='aten.log_', overload='default')>: <function log at 0x7776e1b668e0>, <OpOverload(op='aten.fmod_', overload='Tensor')>: <function fmod at 0x7776e1b65f80>, <OpOverload(op='aten.fmod_', overload='Scalar')>: <function fmod at 0x7776e1b65f80>, <OpOverload(op='aten.frac_', overload='default')>: <function frac at 0x7776e1b660c0>, <OpOverload(op='aten.geometric_', overload='default')>: <function geometric at 0x7776e1b672e0>, <OpOverload(op='aten.log2_', overload='default')>: <function log2 at 0x7776e1b66b60>, <OpOverload(op='aten.gcd_', overload='default')>: <function gcd at 0x7776e1b66200>, <OpOverload(op='aten.ge_', overload='Scalar')>: <function ge at 0x7776e1b66340>, <OpOverload(op='aten.ge_', overload='Tensor')>: <function ge at 0x7776e1b66340>, <OpOverload(op='aten.gt_', overload='Scalar')>: <function gt at 0x7776e1b66480>, <OpOverload(op='aten.gt_', overload='Tensor')>: <function gt at 0x7776e1b66480>, <OpOverload(op='aten.log1p_', overload='default')>: <function log1p at 0x7776e1b66de0>, <OpOverload(op='aten.heaviside_', overload='default')>: <function heaviside at 0x7776e1b665c0>, <OpOverload(op='aten.log_normal_', overload='default')>: <function log_normal at 0x7776e1b658a0>, <OpOverload(op='aten.hypot_', overload='default')>: <function hypot at 0x7776e1b66700>, <OpOverload(op='aten.igamma_', overload='default')>: <function igamma at 0x7776e1b66840>, <OpOverload(op='aten.igammac_', overload='default')>: <function igammac at 0x7776e1b66980>, <OpOverload(op='aten.lgamma_', overload='default')>: <function lgamma at 0x7776e1b66fc0>, <OpOverload(op='aten.zero_', overload='default')>: <function zero at 0x7776e1b65300>, <OpOverload(op='aten.i0_', overload='default')>: <function i0 at 0x7776e1b66ac0>, <OpOverload(op='aten.lcm_', overload='default')>: <function lcm at 0x7776e1b66c00>, <OpOverload(op='aten.lerp_', overload='Scalar')>: <function lerp at 0x7776e1b66e80>, <OpOverload(op='aten.le_', overload='Scalar')>: <function le at 0x7776e1b66d40>, <OpOverload(op='aten.lt_', overload='Tensor')>: <function lt at 0x7776e1b65c60>, <OpOverload(op='aten.mul_', overload='Tensor')>: <function mul at 0x7776e1b64180>, <OpOverload(op='aten.mul_', overload='Scalar')>: <function mul at 0x7776e1b64180>, <OpOverload(op='aten.mvlgamma_', overload='default')>: <function _make_alias.<locals>._fn at 0x7776e1b64680>, <OpOverload(op='aten.nan_to_num_', overload='default')>: <function nan_to_num at 0x7776e1b64b80>, <OpOverload(op='aten.xlogy_', overload='Tensor')>: <function xlogy at 0x7776e1b67a60>, <OpOverload(op='aten.xlogy_', overload='Scalar_Other')>: <function xlogy at 0x7776e1b67a60>, <OpOverload(op='aten.ne_', overload='Scalar')>: <function ne at 0x7776e1b65080>, <OpOverload(op='aten.ne_', overload='Tensor')>: <function ne at 0x7776e1b65080>, <OpOverload(op='aten.neg_', overload='default')>: <function neg at 0x7776e1b65580>, <OpOverload(op='aten.nextafter_', overload='default')>: <function nextafter at 0x7776e1b65a80>, <OpOverload(op='aten.pow_', overload='Scalar')>: <function pow at 0x7776e1b67100>, <OpOverload(op='aten.pow_', overload='Tensor')>: <function pow at 0x7776e1b67100>, <OpOverload(op='aten.trunc_', overload='default')>: <function trunc at 0x7776e1b67ce0>, <OpOverload(op='aten.rad2deg_', overload='default')>: <function rad2deg at 0x7776e1b67240>, <OpOverload(op='aten.reciprocal_', overload='default')>: <function reciprocal at 0x7776e1b67380>, <OpOverload(op='aten.true_divide_', overload='Scalar')>: <function true_divide at 0x7776e1b67e20>, <OpOverload(op='aten.remainder_', overload='Tensor')>: <function remainder at 0x7776e1b674c0>, <OpOverload(op='aten.remainder_', overload='Scalar')>: <function remainder at 0x7776e1b674c0>, <OpOverload(op='aten.rsqrt_', overload='default')>: <function rsqrt at 0x7776e1b67600>, <OpOverload(op='aten.true_divide_', overload='Tensor')>: <function true_divide at 0x7776e1b67e20>, <OpOverload(op='aten.transpose_copy', overload='int')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.transpose at 0x7776e1b64400>, <OpOverload(op='aten.sgn_', overload='default')>: <function sgn at 0x7776e1b67740>, <OpOverload(op='aten.sigmoid_', overload='default')>: <function sigmoid at 0x7776e1b67880>, <OpOverload(op='aten.triu_', overload='default')>: <function triu at 0x7776e1b67f60>, <OpOverload(op='aten.sign_', overload='default')>: <function sign at 0x7776e1b679c0>, <OpOverload(op='aten.sin_', overload='default')>: <function sin at 0x7776e1b67b00>, <OpOverload(op='aten.transpose_copy', overload='int_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.transpose at 0x7776e1b64400>, <OpOverload(op='aten.sinc_', overload='default')>: <function sinc at 0x7776e1b67c40>, <OpOverload(op='aten.sinh_', overload='default')>: <function sinh at 0x7776e1b67d80>, <OpOverload(op='aten.sqrt_', overload='default')>: <function sqrt at 0x7776e1b67ec0>, <OpOverload(op='aten.square_', overload='default')>: <function square at 0x7776e1b98040>, <OpOverload(op='aten.tril_', overload='default')>: <function tril at 0x7776e1b98540>, <OpOverload(op='aten.sub_', overload='Tensor')>: <function sub at 0x7776e1b98180>, <OpOverload(op='aten.sub_', overload='Scalar')>: <function sub at 0x7776e1b98180>, <OpOverload(op='aten.tan_', overload='default')>: <function tan at 0x7776e1b982c0>, <OpOverload(op='aten.tanh_', overload='default')>: <function tanh at 0x7776e1b98400>, <OpOverload(op='aten.unbind_copy', overload='int')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unbind at 0x7776e1b65800>, <OpOverload(op='aten.alias_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.alias at 0x7776e1b65b20>, <OpOverload(op='aten.alias_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.alias at 0x7776e1b65b20>, <OpOverload(op='aten.as_strided_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.as_strided at 0x7776e1b667a0>, <OpOverload(op='aten.as_strided_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.as_strided at 0x7776e1b667a0>, <OpOverload(op='aten.diagonal_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.diagonal at 0x7776e1b987c0>, <OpOverload(op='aten.diagonal_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.diagonal at 0x7776e1b987c0>, <OpOverload(op='aten.expand_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.expand at 0x7776e1b98360>, <OpOverload(op='aten.expand_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.expand at 0x7776e1b98360>, <OpOverload(op='aten.narrow_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.narrow at 0x7776e1b98a40>, <OpOverload(op='aten.squeeze_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='dims')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='dim_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='dims_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.permute_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.permute at 0x7776e1b99080>, <OpOverload(op='aten.permute_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.permute at 0x7776e1b99080>, <OpOverload(op='aten.t_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.t at 0x7776e1b993a0>, <OpOverload(op='aten.t_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.t at 0x7776e1b993a0>, <OpOverload(op='aten.unbind_copy', overload='int_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unbind at 0x7776e1b65800>, <OpOverload(op='aten.unsqueeze_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unsqueeze at 0x7776e1b67ba0>, <OpOverload(op='aten.unsqueeze_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unsqueeze at 0x7776e1b67ba0>, <OpOverload(op='aten.view_copy', overload='dtype')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.view_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.view_copy', overload='dtype_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.complex', overload='default')>: <function complex at 0x7776e1b9ab60>, <OpOverload(op='aten.complex', overload='out')>: <function complex at 0x7776e1b9ab60>, <OpOverload(op='aten.fft_ifft', overload='out')>: <function ifft at 0x7776e1b9bb00>, <OpOverload(op='aten.polar', overload='default')>: <function polar at 0x7776e1b9ade0>, <OpOverload(op='aten.polar', overload='out')>: <function polar at 0x7776e1b9ade0>, <OpOverload(op='aten.fft_fft', overload='default')>: <function fft at 0x7776e1b9b880>, <OpOverload(op='aten.fft_fft', overload='out')>: <function fft at 0x7776e1b9b880>, <OpOverload(op='aten.fft_ifft', overload='default')>: <function ifft at 0x7776e1b9bb00>, <OpOverload(op='aten.fft_rfft', overload='default')>: <function rfft at 0x7776e1b9b7e0>, <OpOverload(op='aten.fft_rfft', overload='out')>: <function rfft at 0x7776e1b9b7e0>, <OpOverload(op='aten.fft_irfft', overload='default')>: <function irfft at 0x7776e1b98e00>, <OpOverload(op='aten.fft_irfft', overload='out')>: <function irfft at 0x7776e1b98e00>, <OpOverload(op='aten.fft_hfft', overload='default')>: <function hfft at 0x7776e1b9bd80>, <OpOverload(op='aten.fft_hfft', overload='out')>: <function hfft at 0x7776e1b9bd80>, <OpOverload(op='aten.fft_fft2', overload='out')>: <function fft2 at 0x7776e1b9bce0>, <OpOverload(op='aten.fft_ihfft', overload='default')>: <function ihfft at 0x7776e1b9bf60>, <OpOverload(op='aten.fft_ihfft', overload='out')>: <function ihfft at 0x7776e1b9bf60>, <OpOverload(op='aten.fft_fft2', overload='default')>: <function fft2 at 0x7776e1b9bce0>, <OpOverload(op='aten.fft_fftn', overload='default')>: <function fftn at 0x7776e19d8900>, <OpOverload(op='aten.fft_fftn', overload='out')>: <function fftn at 0x7776e19d8900>, <OpOverload(op='aten.fft_ifftn', overload='default')>: <function ifftn at 0x7776e19d8b80>, <OpOverload(op='aten.fft_rfftn', overload='default')>: <function rfftn at 0x7776e19d8e00>, <OpOverload(op='aten.fft_rfftn', overload='out')>: <function rfftn at 0x7776e19d8e00>, <OpOverload(op='aten.fft_ihfftn', overload='default')>: <function ihfftn at 0x7776e19d9080>, <OpOverload(op='aten.fft_ihfftn', overload='out')>: <function ihfftn at 0x7776e19d9080>, <OpOverload(op='aten.fft_irfftn', overload='default')>: <function irfftn at 0x7776e19d9760>, <OpOverload(op='aten.fft_irfftn', overload='out')>: <function irfftn at 0x7776e19d9760>, <OpOverload(op='aten.scatter_', overload='src')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_', overload='value')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_', overload='reduce')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_', overload='value_reduce')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_add_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc680>, <OpOverload(op='aten.scatter_reduce_', overload='two')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbe020>, <OpOverload(op='aten.silu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbe160>, <OpOverload(op='aten.is_complex', overload='default')>: <function is_complex at 0x7776e1c3de40>, <OpOverload(op='aten.erfinv', overload='default')>: <function erfinv at 0x7776e1c3f7e0>, <OpOverload(op='aten.erfinv', overload='out')>: <function erfinv at 0x7776e1c3f7e0>, <OpOverload(op='aten.zero', overload='default')>: <function zero at 0x7776e1c4d260>, <OpOverload(op='aten.zero', overload='out')>: <function zero at 0x7776e1c4d260>, <OpOverload(op='aten.block_diag', overload='out')>: <function _block_diag_iterable at 0x7776e1b10b80>, <OpOverload(op='aten.frac', overload='default')>: <function frac at 0x7776e1c4dc60>, <OpOverload(op='aten.frac', overload='out')>: <function frac at 0x7776e1c4dc60>, <OpOverload(op='aten.isinf', overload='default')>: <function isinf at 0x7776e1c4e660>, <OpOverload(op='aten.isinf', overload='out')>: <function isinf at 0x7776e1c4e660>, <OpOverload(op='aten.isposinf', overload='default')>: <function isposinf at 0x7776e1c4d800>, <OpOverload(op='aten.isposinf', overload='out')>: <function isposinf at 0x7776e1c4d800>, <OpOverload(op='aten.isneginf', overload='default')>: <function isneginf at 0x7776e1c4e980>, <OpOverload(op='aten.isneginf', overload='out')>: <function isneginf at 0x7776e1c4e980>, <OpOverload(op='aten.isnan', overload='default')>: <function isnan at 0x7776e1c4ee80>, <OpOverload(op='aten.isnan', overload='out')>: <function isnan at 0x7776e1c4ee80>, <OpOverload(op='aten.i0', overload='default')>: <function i0 at 0x7776e1c4f880>, <OpOverload(op='aten.i0', overload='out')>: <function i0 at 0x7776e1c4f880>, <OpOverload(op='aten.index_fill_', overload='Dimname_Scalar')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.logsumexp', overload='names_out')>: <function logsumexp at 0x7776e1c65760>, <OpOverload(op='aten.logsumexp', overload='names')>: <function logsumexp at 0x7776e1c65760>, <OpOverload(op='aten.logsumexp', overload='out')>: <function logsumexp at 0x7776e1c65760>, <OpOverload(op='aten.logsumexp', overload='default')>: <function logsumexp at 0x7776e1c65760>, <OpOverload(op='aten.nan_to_num', overload='default')>: <function nan_to_num at 0x7776e1c4d1c0>, <OpOverload(op='aten.nan_to_num', overload='out')>: <function nan_to_num at 0x7776e1c4d1c0>, <OpOverload(op='aten.sigmoid', overload='default')>: <function sigmoid at 0x7776e1c66de0>, <OpOverload(op='aten.sigmoid', overload='out')>: <function sigmoid at 0x7776e1c66de0>, <OpOverload(op='aten.std_mean', overload='dim')>: <function std_mean at 0x7776e1c656c0>, <OpOverload(op='aten.sgn', overload='default')>: <function sgn at 0x7776e1c672e0>, <OpOverload(op='aten.sgn', overload='out')>: <function sgn at 0x7776e1c672e0>, <OpOverload(op='aten.sinc', overload='default')>: <function sinc at 0x7776e1c74720>, <OpOverload(op='aten.sinc', overload='out')>: <function sinc at 0x7776e1c74720>, <OpOverload(op='aten.std', overload='names_dim')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.bitwise_right_shift', overload='Tensor_Scalar_out')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_right_shift', overload='Scalar_Tensor_out')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_left_shift', overload='Tensor_out')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.xlogy', overload='OutTensor')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.bitwise_left_shift', overload='Tensor')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_left_shift', overload='Tensor_Scalar')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_left_shift', overload='Scalar_Tensor')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_left_shift', overload='Tensor_Scalar_out')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_left_shift', overload='Scalar_Tensor_out')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_right_shift', overload='Tensor')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_right_shift', overload='Tensor_Scalar')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_right_shift', overload='Tensor_out')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_right_shift', overload='Scalar_Tensor')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.copysign', overload='Tensor')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.copysign', overload='Scalar')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.copysign', overload='out')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.copysign', overload='Scalar_out')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.lerp_', overload='Tensor')>: <function lerp at 0x7776e1b66e80>, <OpOverload(op='aten.trace', overload='out')>: <function trace at 0x7776e1b40f40>, <OpOverload(op='aten.heaviside', overload='default')>: <function heaviside at 0x7776e1c96ca0>, <OpOverload(op='aten.heaviside', overload='out')>: <function heaviside at 0x7776e1c96ca0>, <OpOverload(op='aten.logical_and', overload='out')>: <function logical_and at 0x7776e1cad120>, <OpOverload(op='aten.logical_and', overload='default')>: <function logical_and at 0x7776e1cad120>, <OpOverload(op='aten.std', overload='correction_out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.lcm', overload='default')>: <function lcm at 0x7776e1c97f60>, <OpOverload(op='aten.lcm', overload='out')>: <function lcm at 0x7776e1c97f60>, <OpOverload(op='aten.squeeze_copy', overload='dim')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.logaddexp', overload='default')>: <function logaddexp at 0x7776e1cac860>, <OpOverload(op='aten.logaddexp', overload='out')>: <function logaddexp at 0x7776e1cac860>, <OpOverload(op='aten.logaddexp2', overload='default')>: <function logaddexp2 at 0x7776e1caccc0>, <OpOverload(op='aten.logaddexp2', overload='out')>: <function logaddexp2 at 0x7776e1caccc0>, <OpOverload(op='aten.logical_not', overload='default')>: <function logical_not at 0x7776e1cad080>, <OpOverload(op='aten.logical_not', overload='out')>: <function logical_not at 0x7776e1cad080>, <OpOverload(op='aten.logical_or', overload='default')>: <function logical_or at 0x7776e1cad300>, <OpOverload(op='aten.logical_or', overload='out')>: <function logical_or at 0x7776e1cad300>, <OpOverload(op='aten.logical_xor', overload='default')>: <function logical_xor at 0x7776e1cad760>, <OpOverload(op='aten.logical_xor', overload='out')>: <function logical_xor at 0x7776e1cad760>, <OpOverload(op='aten.rsub', overload='Tensor')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.rsub', overload='Scalar')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.rsub', overload='Tensor_out')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.rsub', overload='Scalar_out')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.xlogy', overload='Scalar_Other')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.xlogy', overload='Scalar_Self')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.xlogy', overload='OutScalar_Other')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.xlogy', overload='OutScalar_Self')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.addcdiv', overload='default')>: <function addcdiv at 0x7776e1ad04a0>, <OpOverload(op='aten.addcdiv', overload='out')>: <function addcdiv at 0x7776e1ad04a0>, <OpOverload(op='aten.addcmul', overload='default')>: <function addcmul at 0x7776e1ad0860>, <OpOverload(op='aten.addcmul', overload='out')>: <function addcmul at 0x7776e1ad0860>, <OpOverload(op='aten.clamp', overload='default')>: <function clamp at 0x7776e1ad0c20>, <OpOverload(op='aten.clamp', overload='Tensor')>: <function clamp at 0x7776e1ad0c20>, <OpOverload(op='aten.clamp', overload='out')>: <function clamp at 0x7776e1ad0c20>, <OpOverload(op='aten.clamp', overload='Tensor_out')>: <function clamp at 0x7776e1ad0c20>, <OpOverload(op='aten.clamp_min', overload='default')>: <function clamp_min at 0x7776e1ad0ea0>, <OpOverload(op='aten.clamp_min', overload='Tensor')>: <function clamp_min at 0x7776e1ad0ea0>, <OpOverload(op='aten.clamp_min', overload='out')>: <function clamp_min at 0x7776e1ad0ea0>, <OpOverload(op='aten.clamp_min', overload='Tensor_out')>: <function clamp_min at 0x7776e1ad0ea0>, <OpOverload(op='aten.frexp', overload='Tensor_out')>: <function frexp at 0x7776e1c95f80>, <OpOverload(op='aten._adaptive_avg_pool2d', overload='default')>: <function adaptive_avg_pool2d at 0x7776e1d5a700>, <OpOverload(op='aten.relu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd440>, <OpOverload(op='aten.addmm', overload='dtype_out')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.sigmoid_backward', overload='grad_input')>: <function sigmoid_backward at 0x7776e1cde020>, <OpOverload(op='aten.sigmoid_backward', overload='default')>: <function sigmoid_backward at 0x7776e1cde020>, <OpOverload(op='aten.hardswish', overload='out')>: <function hardswish at 0x7776e1cdef20>, <OpOverload(op='aten.softplus_backward', overload='default')>: <function softplus_backward at 0x7776e1cde3e0>, <OpOverload(op='aten.softplus_backward', overload='grad_input')>: <function softplus_backward at 0x7776e1cde3e0>, <OpOverload(op='aten.hardswish', overload='default')>: <function hardswish at 0x7776e1cdef20>, <OpOverload(op='aten.elu_backward', overload='default')>: <function elu_backward at 0x7776e1cddf80>, <OpOverload(op='aten.hardsigmoid', overload='default')>: <function hardsigmoid at 0x7776e1cdefc0>, <OpOverload(op='aten.hardsigmoid', overload='out')>: <function hardsigmoid at 0x7776e1cdefc0>, <OpOverload(op='aten.hardsigmoid_backward', overload='default')>: <function hardsigmoid_backward at 0x7776e1cdf060>, <OpOverload(op='aten.hardsigmoid_backward', overload='grad_input')>: <function hardsigmoid_backward at 0x7776e1cdf060>, <OpOverload(op='aten.hardtanh_backward', overload='default')>: <function hardtanh_backward at 0x7776e1cdf420>, <OpOverload(op='aten.hardtanh_backward', overload='grad_input')>: <function hardtanh_backward at 0x7776e1cdf420>, <OpOverload(op='aten.hardswish_backward', overload='default')>: <function hardswish_backward at 0x7776e1cde200>, <OpOverload(op='aten.threshold_backward', overload='default')>: <function threshold_backward at 0x7776e1cddda0>, <OpOverload(op='aten.threshold_backward', overload='grad_input')>: <function threshold_backward at 0x7776e1cddda0>, <OpOverload(op='aten.leaky_relu_backward', overload='default')>: <function leaky_relu_backward at 0x7776e1cdfb00>, <OpOverload(op='aten.leaky_relu_backward', overload='grad_input')>: <function leaky_relu_backward at 0x7776e1cdfb00>, <OpOverload(op='aten.gelu_backward', overload='default')>: <function gelu_backward at 0x7776e1cdfec0>, <OpOverload(op='aten.gelu_backward', overload='grad_input')>: <function gelu_backward at 0x7776e1cdfec0>, <OpOverload(op='aten.mish_backward', overload='default')>: <function mish_backward at 0x7776e1cfc2c0>, <OpOverload(op='aten.silu', overload='default')>: <function silu at 0x7776e1cfc720>, <OpOverload(op='aten.silu', overload='out')>: <function silu at 0x7776e1cfc720>, <OpOverload(op='aten.silu_backward', overload='default')>: <function silu_backward at 0x7776e1cfc7c0>, <OpOverload(op='aten.silu_backward', overload='grad_input')>: <function silu_backward at 0x7776e1cfc7c0>, <OpOverload(op='aten._prelu_kernel', overload='default')>: <function _prelu_kernel at 0x7776e1cfcae0>, <OpOverload(op='aten._prelu_kernel_backward', overload='default')>: <function _prelu_kernel_backward at 0x7776e1cfcb80>, <OpOverload(op='aten.rrelu_with_noise_backward', overload='default')>: <function rrelu_with_noise_backward at 0x7776e1cfd080>, <OpOverload(op='aten.rrelu_with_noise_backward', overload='out')>: <function rrelu_with_noise_backward at 0x7776e1cfd080>, <OpOverload(op='aten.mse_loss', overload='out')>: <function mse_loss at 0x7776e1cfd8a0>, <OpOverload(op='aten.mse_loss', overload='default')>: <function mse_loss at 0x7776e1cfd8a0>, <OpOverload(op='aten.log_sigmoid_backward', overload='grad_input')>: <function log_sigmoid_backward at 0x7776e1cfd120>, <OpOverload(op='aten.smooth_l1_loss_backward', overload='default')>: <function smooth_l1_loss_backward at 0x7776e1cfd9e0>, <OpOverload(op='aten.mse_loss_backward', overload='default')>: <function mse_loss_backward at 0x7776e1cdf240>, <OpOverload(op='aten.mse_loss_backward', overload='grad_input')>: <function mse_loss_backward at 0x7776e1cdf240>, <OpOverload(op='aten._safe_softmax', overload='default')>: <function safe_softmax at 0x7776e1cfcd60>, <OpOverload(op='aten.smooth_l1_loss_backward', overload='grad_input')>: <function smooth_l1_loss_backward_out at 0x7776e1cfdbc0>, <OpOverload(op='aten.smooth_l1_loss', overload='default')>: <function smooth_l1_loss at 0x7776e1cfd940>, <OpOverload(op='aten.smooth_l1_loss', overload='out')>: <function smooth_l1_loss at 0x7776e1cfd940>, <OpOverload(op='aten.huber_loss_backward', overload='default')>: <function huber_loss_backward at 0x7776e1cfdda0>, <OpOverload(op='aten.huber_loss_backward', overload='out')>: <function huber_loss_backward_out at 0x7776e1cfdf80>, <OpOverload(op='aten.glu_backward', overload='default')>: <function glu_backward at 0x7776e1cfe200>, <OpOverload(op='aten.glu_backward', overload='grad_input')>: <function glu_backward at 0x7776e1cfe200>, <OpOverload(op='aten.nll_loss_backward', overload='grad_input')>: <function nll_loss_backward at 0x7776e1cfe5c0>, <OpOverload(op='aten.nll_loss2d_backward', overload='default')>: <function nll_loss2d_backward at 0x7776e1cfe8e0>, <OpOverload(op='aten.nll_loss2d_backward', overload='grad_input')>: <function nll_loss2d_backward at 0x7776e1cfe8e0>, <OpOverload(op='aten.binary_cross_entropy', overload='default')>: <function binary_cross_entropy at 0x7776e1cfee80>, <OpOverload(op='aten.binary_cross_entropy', overload='out')>: <function binary_cross_entropy at 0x7776e1cfee80>, <OpOverload(op='aten.binary_cross_entropy_backward', overload='grad_input')>: <function binary_cross_entropy_backward at 0x7776e1cfef20>, <OpOverload(op='aten.soft_margin_loss_backward', overload='default')>: <function soft_margin_loss_backward at 0x7776e1cff4c0>, <OpOverload(op='aten.soft_margin_loss', overload='default')>: <function soft_margin_loss at 0x7776e1cff560>, <OpOverload(op='aten.unfold_backward', overload='out')>: <function unfold_backward at 0x7776e1cfdd00>, <OpOverload(op='aten.dist', overload='default')>: <function dist at 0x7776e1cfdb20>, <OpOverload(op='aten.dist', overload='out')>: <function dist at 0x7776e1cfdb20>, <OpOverload(op='aten._euclidean_dist', overload='default')>: <function _euclidean_dist at 0x7776e1cff2e0>, <OpOverload(op='aten._euclidean_dist', overload='out')>: <function _euclidean_dist at 0x7776e1cff2e0>, <OpOverload(op='aten.slice_backward', overload='out')>: <function slice_backward at 0x7776e1cff7e0>, <OpOverload(op='aten.select_backward', overload='default')>: <function select_backward at 0x7776e1cffec0>, <OpOverload(op='aten.diagonal_backward', overload='default')>: <function diagonal_backward at 0x7776e1d28180>, <OpOverload(op='aten.logit_backward', overload='default')>: <function logit_backward at 0x7776e1cfea20>, <OpOverload(op='aten._softmax_backward_data', overload='default')>: <function _softmax_backward_data at 0x7776e1d282c0>, <OpOverload(op='aten._softmax_backward_data', overload='out')>: <function _softmax_backward_data at 0x7776e1d282c0>, <OpOverload(op='aten._log_softmax_backward_data', overload='default')>: <function _log_softmax_backward_data at 0x7776e1d28900>, <OpOverload(op='aten._log_softmax_backward_data', overload='out')>: <function _log_softmax_backward_data at 0x7776e1d28900>, <OpOverload(op='aten.im2col', overload='out')>: <function im2col at 0x7776e1d28c20>, <OpOverload(op='aten.native_dropout_backward', overload='out')>: <function native_dropout_backward at 0x7776e1d29260>, <OpOverload(op='aten.col2im', overload='out')>: <function col2im at 0x7776e1d28fe0>, <OpOverload(op='aten.col2im', overload='default')>: <function col2im at 0x7776e1d28fe0>, <OpOverload(op='aten.native_layer_norm_backward', overload='default')>: <function native_layer_norm_backward at 0x7776e1d29080>, <OpOverload(op='aten.lift', overload='out')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.native_dropout_backward', overload='default')>: <function native_dropout_backward at 0x7776e1d29260>, <OpOverload(op='aten.dropout', overload='default')>: <function dropout at 0x7776e1d291c0>, <OpOverload(op='aten.native_dropout', overload='default')>: <function native_dropout at 0x7776e1d29620>, <OpOverload(op='aten.native_dropout', overload='out')>: <function native_dropout at 0x7776e1d29620>, <OpOverload(op='aten._softmax', overload='default')>: <function _softmax at 0x7776e1d299e0>, <OpOverload(op='aten._softmax', overload='out')>: <function _softmax at 0x7776e1d299e0>, <OpOverload(op='aten._log_softmax', overload='default')>: <function _log_softmax at 0x7776e1d29c60>, <OpOverload(op='aten._log_softmax', overload='out')>: <function _log_softmax at 0x7776e1d29c60>, <OpOverload(op='aten.embedding', overload='default')>: <function embedding at 0x7776e1d29ee0>, <OpOverload(op='aten.embedding', overload='out')>: <function embedding at 0x7776e1d29ee0>, <OpOverload(op='aten._chunk_cat', overload='default')>: <function _chunk_cat at 0x7776e1d2a480>, <OpOverload(op='aten._chunk_cat', overload='out')>: <function _chunk_cat at 0x7776e1d2a480>, <OpOverload(op='aten.embedding_dense_backward', overload='default')>: <function embedding_dense_backward at 0x7776e1d2a160>, <OpOverload(op='aten.embedding_dense_backward', overload='out')>: <function embedding_dense_backward at 0x7776e1d2a160>, <OpOverload(op='aten.split_with_sizes_copy', overload='default')>: <function split_with_sizes_copy at 0x7776e1d2a520>, <OpOverload(op='aten.split_with_sizes_copy', overload='out')>: <function split_with_sizes_copy at 0x7776e1d2a520>, <OpOverload(op='aten._addmm_activation', overload='default')>: <function _addmm_activation at 0x7776e1d2aac0>, <OpOverload(op='aten.unsafe_split', overload='Tensor')>: <function unsafe_split at 0x7776e1d2a660>, <OpOverload(op='aten.unsafe_split_with_sizes', overload='default')>: <function unsafe_split_with_sizes at 0x7776e1d2a7a0>, <OpOverload(op='aten.split', overload='Tensor')>: <function split at 0x7776e1d2a8e0>, <OpOverload(op='aten.native_group_norm_backward', overload='out')>: <function native_group_norm_backward_out at 0x7776e1d29bc0>, <OpOverload(op='aten.addmm', overload='default')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.addmm', overload='out')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.addmm', overload='dtype')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.native_group_norm_backward', overload='default')>: <function native_group_norm_backward at 0x7776e1d29e40>, <OpOverload(op='aten.addmv', overload='default')>: <function addmv at 0x7776e1d2a0c0>, <OpOverload(op='aten.native_batch_norm_backward', overload='default')>: <function native_batch_norm_backward at 0x7776e1d59300>, <OpOverload(op='aten._fused_rms_norm_backward', overload='default')>: <function _fused_rms_norm_backward at 0x7776e1d2b240>, <OpOverload(op='aten.native_batch_norm_backward', overload='out')>: <function native_batch_norm_backward_out at 0x7776e1d59440>, <OpOverload(op='aten._native_batch_norm_legit', overload='default')>: <function _native_batch_norm_legit at 0x7776e1d2bec0>, <OpOverload(op='aten.native_batch_norm', overload='default')>: <function native_batch_norm at 0x7776e1d2ba60>, <OpOverload(op='aten.native_batch_norm', overload='out')>: <function native_batch_norm at 0x7776e1d2ba60>, <OpOverload(op='aten._native_batch_norm_legit', overload='no_stats')>: <function _native_batch_norm_legit_no_stats at 0x7776e1d58040>, <OpOverload(op='aten._native_batch_norm_legit_functional', overload='default')>: <function _native_batch_norm_legit_functional at 0x7776e1d58180>, <OpOverload(op='aten._batch_norm_with_update', overload='default')>: <function _batch_norm_with_update at 0x7776e1d58400>, <OpOverload(op='aten._batch_norm_with_update_functional', overload='default')>: <function _batch_norm_with_update_functional at 0x7776e1d584a0>, <OpOverload(op='aten._batch_norm_no_update', overload='default')>: <function _batch_norm_no_update at 0x7776e1d585e0>, <OpOverload(op='aten.batch_norm_backward', overload='default')>: <function batch_norm_backward at 0x7776e1d580e0>, <OpOverload(op='aten._to_copy', overload='default')>: <function _to_copy at 0x7776e1d59260>, <OpOverload(op='aten._to_copy', overload='out')>: <function _to_copy at 0x7776e1d59260>, <OpOverload(op='aten._fused_dropout', overload='default')>: <function _fused_dropout_decomposition at 0x7776e1d58ea0>, <OpOverload(op='aten._fused_dropout', overload='out')>: <function _fused_dropout_decomposition at 0x7776e1d58ea0>, <OpOverload(op='aten.cudnn_batch_norm', overload='out')>: <function cudnn_batch_norm at 0x7776e1d2a5c0>, <OpOverload(op='aten.index_copy_', overload='default')>: <function index_copy_ at 0x7776e1d5ade0>, <OpOverload(op='aten.index_copy_', overload='dimname')>: <function index_copy_ at 0x7776e1d5ade0>, <OpOverload(op='aten.miopen_batch_norm_backward', overload='default')>: <function miopen_batch_norm_backward at 0x7776e1d59bc0>, <OpOverload(op='aten.miopen_batch_norm_backward', overload='out')>: <function miopen_batch_norm_backward at 0x7776e1d59bc0>, <OpOverload(op='aten.cudnn_batch_norm_backward', overload='default')>: <function cudnn_batch_norm_backward at 0x7776e1d5a2a0>, <OpOverload(op='aten.cudnn_batch_norm_backward', overload='out')>: <function cudnn_batch_norm_backward at 0x7776e1d5a2a0>, <OpOverload(op='aten._upsample_nearest_exact3d', overload='out')>: <function _upsample_nearest_exact3d at 0x7776e1d5ae80>, <OpOverload(op='aten.max_unpool2d', overload='default')>: <function max_unpool2d at 0x7776e1d5aa20>, <OpOverload(op='aten.max_unpool2d', overload='out')>: <function max_unpool2d at 0x7776e1d5aa20>, <OpOverload(op='aten.index_add', overload='out')>: <function index_add at 0x7776e1d5b060>, <OpOverload(op='aten.index_add', overload='dimname')>: <function index_add at 0x7776e1d5b060>, <OpOverload(op='aten.pad_sequence', overload='default')>: <function pad_sequence at 0x7776e1d5ac00>, <OpOverload(op='aten.index_add', overload='default')>: <function index_add at 0x7776e1d5b060>, <OpOverload(op='aten.max_unpool3d', overload='default')>: <function max_unpool3d at 0x7776e1d5aca0>, <OpOverload(op='aten.max_unpool3d', overload='out')>: <function max_unpool3d at 0x7776e1d5aca0>, <OpOverload(op='aten.index_add_', overload='default')>: <function index_add_ at 0x7776e1d5aac0>, <OpOverload(op='aten._upsample_nearest_exact3d', overload='default')>: <function _upsample_nearest_exact3d at 0x7776e1d5ae80>, <OpOverload(op='aten.index_copy', overload='default')>: <function index_copy at 0x7776e1d58220>, <OpOverload(op='aten.index_copy', overload='dimname')>: <function index_copy at 0x7776e1d58220>, <OpOverload(op='aten.index_copy', overload='out')>: <function index_copy at 0x7776e1d58220>, <OpOverload(op='aten.log_sigmoid_forward', overload='default')>: <function log_sigmoid_forward at 0x7776e1d5b6a0>, <OpOverload(op='aten.log_sigmoid_forward', overload='output')>: <function log_sigmoid_forward at 0x7776e1d5b6a0>, <OpOverload(op='aten._upsample_nearest_exact1d', overload='vec')>: <function _upsample_nearest_exact_vec at 0x7776e1d78680>, <OpOverload(op='aten.uniform_', overload='default')>: <function uniform_ at 0x7776e1d5b880>, <OpOverload(op='aten.upsample_nearest1d', overload='default')>: <function upsample_nearest1d at 0x7776e1d78a40>, <OpOverload(op='aten._upsample_nearest_exact1d', overload='default')>: <function upsample_nearest_exact1d at 0x7776e1d78d60>, <OpOverload(op='aten.upsample_nearest3d', overload='vec')>: <function _upsample_nearest_vec at 0x7776e1d78220>, <OpOverload(op='aten.rnn_tanh', overload='data')>: <function rnn_tanh_data at 0x7776e1d79d00>, <OpOverload(op='aten.upsample_nearest1d', overload='out')>: <function upsample_nearest1d at 0x7776e1d78a40>, <OpOverload(op='aten._upsample_nearest_exact2d', overload='vec')>: <function _upsample_nearest_exact_vec at 0x7776e1d78680>, <OpOverload(op='aten.upsample_nearest2d', overload='default')>: <function upsample_nearest2d at 0x7776e1d79080>, <OpOverload(op='aten._upsample_nearest_exact3d', overload='vec')>: <function _upsample_nearest_exact_vec at 0x7776e1d78680>, <OpOverload(op='aten._upsample_nearest_exact1d', overload='out')>: <function upsample_nearest_exact1d at 0x7776e1d78d60>, <OpOverload(op='aten.rnn_relu', overload='input')>: <function rnn_relu_input at 0x7776e1d79940>, <OpOverload(op='aten._upsample_nearest_exact2d', overload='default')>: <function _upsample_nearest_exact2d at 0x7776e1d793a0>, <OpOverload(op='aten.rnn_relu', overload='data')>: <function rnn_relu_data at 0x7776e1d79b20>, <OpOverload(op='aten.upsample_nearest2d', overload='out')>: <function upsample_nearest2d at 0x7776e1d79080>, <OpOverload(op='aten._upsample_nearest_exact2d', overload='out')>: <function _upsample_nearest_exact2d at 0x7776e1d793a0>, <OpOverload(op='aten.upsample_nearest3d', overload='default')>: <function upsample_nearest3d at 0x7776e1d5bd80>, <OpOverload(op='aten.upsample_nearest3d', overload='out')>: <function upsample_nearest3d at 0x7776e1d5bd80>, <OpOverload(op='aten.lstm', overload='input')>: <function lstm_impl at 0x7776e1d7a160>, <OpOverload(op='aten.lstm', overload='data')>: <function lstm_data_impl at 0x7776e1d7a340>, <OpOverload(op='aten._upsample_bilinear2d_aa', overload='vec')>: <function upsample_bilinear2d_aa_vec at 0x7776e1d7aa20>, <OpOverload(op='aten.gru', overload='data')>: <function gru_impl_data at 0x7776e1d7a660>, <OpOverload(op='aten.gru', overload='input')>: <function gru_impl at 0x7776e1d7a840>, <OpOverload(op='aten._upsample_bicubic2d_aa', overload='vec')>: <function upsample_bicubic2d_aa_vec at 0x7776e1d7ac00>, <OpOverload(op='aten.upsample_trilinear3d', overload='out')>: <function upsample_trilinear3d at 0x7776e1d7b880>, <OpOverload(op='aten.upsample_linear1d', overload='out')>: <function upsample_linear1d at 0x7776e1d7b2e0>, <OpOverload(op='aten.upsample_bilinear2d', overload='default')>: <function upsample_bilinear2d at 0x7776e1d7b600>, <OpOverload(op='aten.upsample_bilinear2d', overload='out')>: <function upsample_bilinear2d at 0x7776e1d7b600>, <OpOverload(op='aten.upsample_trilinear3d', overload='vec')>: <function _upsample_linear_vec at 0x7776e1d7b100>, <OpOverload(op='aten.nll_loss_forward', overload='default')>: <function nll_loss_forward at 0x7776e1d7a7a0>, <OpOverload(op='aten.upsample_trilinear3d', overload='default')>: <function upsample_trilinear3d at 0x7776e1d7b880>, <OpOverload(op='aten.nll_loss_forward', overload='output')>: <function nll_loss_forward at 0x7776e1d7a7a0>, <OpOverload(op='aten.is_same_size', overload='default')>: <function is_same_size at 0x7776e1d7bc40>, <OpOverload(op='aten._reshape_alias', overload='default')>: <function _reshape_alias at 0x7776e1d7bf60>, <OpOverload(op='aten._unsafe_view', overload='out')>: <function _reshape_alias at 0x7776e1d7bf60>, <OpOverload(op='aten._unsafe_index', overload='Tensor')>: <function _unsafe_index at 0x7776e1d7be20>, <OpOverload(op='aten._unsafe_masked_index', overload='default')>: <function _unsafe_masked_index at 0x7776e1d94220>, <OpOverload(op='aten._unsafe_masked_index_put_accumulate', overload='default')>: <function _unsafe_masked_index_put_accumulate at 0x7776e1d94360>, <OpOverload(op='aten.nll_loss2d_forward', overload='default')>: <function nll_loss2d_forward at 0x7776e1d7a0c0>, <OpOverload(op='aten.affine_grid_generator', overload='default')>: <function affine_grid_generator at 0x7776e1d94ea0>, <OpOverload(op='aten.affine_grid_generator', overload='out')>: <function affine_grid_generator at 0x7776e1d94ea0>, <OpOverload(op='aten.grid_sampler_2d', overload='default')>: <function grid_sampler_2d at 0x7776e1d95260>, <OpOverload(op='aten.grid_sampler_2d', overload='out')>: <function grid_sampler_2d at 0x7776e1d95260>, <OpOverload(op='aten.mv', overload='default')>: <function mv at 0x7776e1d95580>, <OpOverload(op='aten.upsample_bicubic2d', overload='vec')>: <function upsample_bicubic2d_vec at 0x7776e1d963e0>, <OpOverload(op='aten.binary_cross_entropy_with_logits', overload='default')>: <function binary_cross_entropy_with_logits at 0x7776e1d95800>, <OpOverload(op='aten.binary_cross_entropy_with_logits', overload='out')>: <function binary_cross_entropy_with_logits at 0x7776e1d95800>, <OpOverload(op='aten.upsample_bicubic2d', overload='default')>: <function upsample_bicubic2d_default at 0x7776e1d95f80>, <OpOverload(op='aten.upsample_bicubic2d', overload='out')>: <function upsample_bicubic2d_default at 0x7776e1d95f80>, <OpOverload(op='aten.reflection_pad1d', overload='default')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.reflection_pad1d', overload='out')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.replication_pad1d', overload='default')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.__ilshift__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdda0>, <OpOverload(op='aten.reflection_pad2d', overload='default')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.reflection_pad3d', overload='default')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.reflection_pad3d', overload='out')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.reflection_pad2d', overload='out')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.replication_pad2d', overload='default')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.replication_pad1d', overload='out')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.replication_pad2d', overload='out')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.replication_pad3d', overload='default')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.replication_pad3d', overload='out')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.reflection_pad1d_backward', overload='default')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.reflection_pad2d_backward', overload='default')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.reflection_pad1d_backward', overload='grad_input')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.reflection_pad3d_backward', overload='default')>: <function _reflection_pad_backward at 0x7776e1d96840>, <OpOverload(op='aten._scaled_dot_product_flash_attention_for_cpu', overload='default')>: <function scaled_dot_product_flash_attention_for_cpu at 0x7776e1d97c40>, <OpOverload(op='aten.arange', overload='out')>: <function arange_default at 0x7776e1d976a0>, <OpOverload(op='aten.aminmax', overload='default')>: <function aminmax at 0x7776e1d97060>, <OpOverload(op='aten.aminmax', overload='out')>: <function aminmax at 0x7776e1d97060>, <OpOverload(op='aten.multilabel_margin_loss_forward', overload='output')>: <function multilabel_margin_loss_forward at 0x7776e1d97880>, <OpOverload(op='aten.nansum', overload='default')>: <function nansum at 0x7776e1d97420>, <OpOverload(op='aten.nansum', overload='out')>: <function nansum at 0x7776e1d97420>, <OpOverload(op='aten.multilabel_margin_loss_forward', overload='default')>: <function multilabel_margin_loss_forward at 0x7776e1d97880>, <OpOverload(op='aten.multi_margin_loss', overload='default')>: <function multi_margin_loss at 0x7776e1d97ce0>, <OpOverload(op='aten.multi_margin_loss', overload='out')>: <function multi_margin_loss at 0x7776e1d97ce0>, <OpOverload(op='aten.baddbmm', overload='dtype_out')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.baddbmm', overload='default')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.baddbmm', overload='out')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.baddbmm', overload='dtype')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.__irshift__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbde40>, <OpOverload(op='aten.floor_divide', overload='default')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.floor_divide', overload='Scalar')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.floor_divide', overload='out')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.floor_divide', overload='Scalar_out')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.__irshift__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbde40>, <OpOverload(op='aten.bernoulli', overload='default')>: <function bernoulli at 0x7776e1dbc9a0>, <OpOverload(op='aten._weight_norm_interface', overload='default')>: <function _weight_norm_interface at 0x7776e1dbc720>, <OpOverload(op='aten._weight_norm_interface', overload='out')>: <function _weight_norm_interface at 0x7776e1dbc720>, <OpOverload(op='aten.isin', overload='Tensor_Tensor')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Tensor_Scalar')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Scalar_Tensor')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Tensor_Tensor_out')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Tensor_Scalar_out')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Scalar_Tensor_out')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.take', overload='default')>: <function take at 0x7776e1dbd080>, <OpOverload(op='aten.take', overload='out')>: <function take at 0x7776e1dbd080>, <OpOverload(op='aten.resize_as', overload='default')>: <function resize_as at 0x7776e1dbcea0>, <OpOverload(op='aten.resize_as', overload='out')>: <function resize_as at 0x7776e1dbcea0>, <OpOverload(op='aten.__ior__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdee0>, <OpOverload(op='aten.__ior__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdee0>, <OpOverload(op='aten.addbmm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd120>, <OpOverload(op='aten.addmm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd260>, <OpOverload(op='aten.addmv_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd3a0>, <OpOverload(op='aten.baddbmm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd4e0>, <OpOverload(op='aten.fill_', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd620>, <OpOverload(op='aten.fill_', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd620>, <OpOverload(op='aten.gelu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd760>, <OpOverload(op='aten.index_reduce_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1d977e0>, <OpOverload(op='aten.hardswish_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd8a0>, <OpOverload(op='aten.hardtanh_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd9e0>, <OpOverload(op='aten.hardsigmoid_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdb20>, <OpOverload(op='aten.__iand__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdc60>, <OpOverload(op='aten.__iand__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdc60>, <OpOverload(op='aten.__ilshift__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdda0>, <OpOverload(op='aten.__ixor__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdbc0>, <OpOverload(op='aten.__ixor__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdbc0>, <OpOverload(op='aten.leaky_relu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd940>, <OpOverload(op='aten.logit_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd6c0>, <OpOverload(op='aten.renorm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd1c0>, <OpOverload(op='aten.round_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbccc0>, <OpOverload(op='aten.round_', overload='decimals')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbccc0>, <OpOverload(op='aten.ne', overload='Scalar_out')>: <function ne at 0x7776e1caed40>, <OpOverload(op='aten.ne', overload='Scalar')>: <function ne at 0x7776e1caed40>, <OpOverload(op='aten.ne', overload='Tensor_out')>: <function ne at 0x7776e1caed40>, <OpOverload(op='aten.nextafter', overload='out')>: <function nextafter at 0x7776e1caf1a0>, <OpOverload(op='aten.nextafter', overload='default')>: <function nextafter at 0x7776e1caf1a0>, <OpOverload(op='aten.pow', overload='Tensor_Tensor')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Tensor_Scalar')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Scalar')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Tensor_Scalar_out')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Scalar_out')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Tensor_Tensor_out')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.remainder', overload='Tensor')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Scalar_Tensor')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Scalar')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.full', overload='default')>: <function full at 0x7776e1b12de0>, <OpOverload(op='aten.remainder', overload='Scalar_Tensor_out')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Scalar_out')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Tensor_out')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.sub', overload='Scalar')>: <function sub at 0x7776e1caede0>, <OpOverload(op='aten.sub', overload='Tensor')>: <function sub at 0x7776e1caede0>, <OpOverload(op='aten.upsample_linear1d', overload='default')>: <function upsample_linear1d at 0x7776e1d7b2e0>, <OpOverload(op='aten.sub', overload='out')>: <function sub at 0x7776e1caede0>, <OpOverload(op='aten.sub', overload='Scalar_out')>: <function sub at 0x7776e1caede0>, <OpOverload(op='aten.reflection_pad3d_backward', overload='grad_input')>: <function _reflection_pad_backward at 0x7776e1d96840>, <OpOverload(op='aten.unfold_backward', overload='default')>: <function unfold_backward at 0x7776e1cfdd00>, <OpOverload(op='aten._addmm_activation', overload='out')>: <function _addmm_activation at 0x7776e1d2aac0>, <OpOverload(op='aten.squeeze', overload='default')>: <function squeeze_default at 0x7776e1dbc5e0>, <OpOverload(op='aten.squeeze', overload='dim')>: <function squeeze_default at 0x7776e1dbc5e0>, <OpOverload(op='aten.squeeze', overload='dims')>: <function squeeze at 0x7776e1af7ce0>, <OpOverload(op='aten.transpose', overload='int')>: <function transpose at 0x7776e1b10f40>, <OpOverload(op='aten.elu_backward', overload='grad_input')>: <function elu_backward at 0x7776e1cddf80>, <OpOverload(op='aten.as_strided_scatter', overload='out')>: <function as_strided_scatter at 0x7776e1af40e0>, <OpOverload(op='aten.fft_ifftn', overload='out')>: <function ifftn at 0x7776e19d8b80>, <OpOverload(op='aten.cat', overload='names')>: <function cat at 0x7776e1af4720>, <OpOverload(op='aten.cat', overload='default')>: <function cat at 0x7776e1af4720>, <OpOverload(op='aten.cat', overload='names_out')>: <function cat at 0x7776e1af4720>, <OpOverload(op='aten.cat', overload='out')>: <function cat at 0x7776e1af4720>, <OpOverload(op='aten.ones', overload='default')>: <function ones at 0x7776e1b10e00>, <OpOverload(op='aten.std_mean', overload='correction')>: <function std_mean at 0x7776e1c656c0>, <OpOverload(op='aten.where', overload='self')>: <function where at 0x7776e1ad1760>, <OpOverload(op='aten.where', overload='ScalarOther')>: <function where at 0x7776e1ad1760>, <OpOverload(op='aten.where', overload='ScalarSelf')>: <function where at 0x7776e1ad1760>, <OpOverload(op='aten.where', overload='Scalar')>: <function where at 0x7776e1ad1760>, <OpOverload(op='aten.where', overload='self_out')>: <function where at 0x7776e1ad1760>, <OpOverload(op='aten.arange', overload='start')>: <function arange_start at 0x7776e1d974c0>, <OpOverload(op='aten.soft_margin_loss', overload='out')>: <function soft_margin_loss at 0x7776e1cff560>, <OpOverload(op='aten.nll_loss_backward', overload='default')>: <function nll_loss_backward at 0x7776e1cfe5c0>, <OpOverload(op='aten.sum', overload='dim_IntList')>: <function sum at 0x7776e1ad22a0>, <OpOverload(op='aten.sum', overload='default')>: <function sum_default at 0x7776e1dbc400>, <OpOverload(op='aten.sum', overload='IntList_out')>: <function sum at 0x7776e1ad22a0>, <OpOverload(op='aten.relu6', overload='default')>: <function relu6 at 0x7776e1a20e00>, <OpOverload(op='aten.soft_margin_loss_backward', overload='grad_input')>: <function soft_margin_loss_backward at 0x7776e1cff4c0>, <OpOverload(op='aten.prod', overload='dim_int')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='default')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='dim_Dimname')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='int_out')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='Dimname_out')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='out')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.var', overload='default')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='correction')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='dim')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.normal', overload='float_float_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.var', overload='correction_out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='names_dim')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='names_out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='correction_names')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='correction_names_out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.amax', overload='out')>: <function amax at 0x7776e1ad2840>, <OpOverload(op='aten.nll_loss2d_forward', overload='output')>: <function nll_loss2d_forward at 0x7776e1d7a0c0>, <OpOverload(op='aten.amax', overload='default')>: <function amax at 0x7776e1ad2840>, <OpOverload(op='aten.amin', overload='default')>: <function amin at 0x7776e1ad2700>, <OpOverload(op='aten.amin', overload='out')>: <function amin at 0x7776e1ad2700>, <OpOverload(op='aten.im2col', overload='default')>: <function im2col at 0x7776e1d28c20>, <OpOverload(op='aten.lift', overload='default')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.empty_strided', overload='out')>: <function empty_strided at 0x7776e1b13d80>, <OpOverload(op='aten.reflection_pad2d_backward', overload='grad_input')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.full', overload='out')>: <function full at 0x7776e1b12de0>, <OpOverload(op='aten.arange', overload='default')>: <function arange_default at 0x7776e1d976a0>, <OpOverload(op='aten._adaptive_avg_pool2d', overload='out')>: <function adaptive_avg_pool2d at 0x7776e1d5a700>, <OpOverload(op='aten.diagonal_backward', overload='out')>: <function diagonal_backward at 0x7776e1d28180>, <OpOverload(op='aten.normal', overload='Tensor_float')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.native_layer_norm_backward', overload='out')>: <function native_layer_norm_backward_out at 0x7776e1d2b100>, <OpOverload(op='aten.normal', overload='Tensor_float_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='Tensor_Tensor')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='float_Tensor_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='float_Tensor')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='Tensor_Tensor_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.uniform', overload='default')>: <function uniform at 0x7776e1d5ba60>, <OpOverload(op='aten.uniform', overload='out')>: <function uniform at 0x7776e1d5ba60>, <OpOverload(op='aten.normal', overload='float_float')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.mv', overload='out')>: <function mv at 0x7776e1d95580>, <OpOverload(op='aten.index_fill_', overload='int_Tensor')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.upsample_nearest1d', overload='vec')>: <function _upsample_nearest_vec at 0x7776e1d78220>, <OpOverload(op='aten.frexp', overload='Tensor')>: <function frexp at 0x7776e1c95f80>, <OpOverload(op='aten.div', overload='Tensor_mode')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.lgamma', overload='out')>: <function lgamma at 0x7776e1c4fd80>, <OpOverload(op='aten.log', overload='default')>: <function log at 0x7776e1c642c0>, <OpOverload(op='aten.log', overload='out')>: <function log at 0x7776e1c642c0>, <OpOverload(op='aten.bitwise_xor', overload='Tensor_out')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.log1p', overload='out')>: <function log1p at 0x7776e1c647c0>, <OpOverload(op='aten.log1p', overload='default')>: <function log1p at 0x7776e1c647c0>, <OpOverload(op='aten._native_batch_norm_legit_no_training', overload='default')>: <function _native_batch_norm_legit_no_training at 0x7776e1d2be20>, <OpOverload(op='aten.log2', overload='default')>: <function log2 at 0x7776e1c64cc0>, <OpOverload(op='aten.log10', overload='default')>: <function log10 at 0x7776e1c651c0>, <OpOverload(op='aten.log10', overload='out')>: <function log10 at 0x7776e1c651c0>, <OpOverload(op='aten.rsqrt', overload='default')>: <function rsqrt at 0x7776e1c668e0>, <OpOverload(op='aten.reciprocal', overload='out')>: <function reciprocal at 0x7776e1c66020>, <OpOverload(op='aten.reciprocal', overload='default')>: <function reciprocal at 0x7776e1c66020>, <OpOverload(op='aten.neg', overload='default')>: <function neg at 0x7776e1c659e0>, <OpOverload(op='aten.div', overload='Scalar_mode')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.neg', overload='out')>: <function neg at 0x7776e1c659e0>, <OpOverload(op='aten.log_sigmoid_backward', overload='default')>: <function log_sigmoid_backward at 0x7776e1cfd120>, <OpOverload(op='aten.tan', overload='default')>: <function tan at 0x7776e1c753a0>, <OpOverload(op='aten.tan', overload='out')>: <function tan at 0x7776e1c753a0>, <OpOverload(op='aten.round', overload='decimals')>: <function round at 0x7776e1c663e0>, <OpOverload(op='aten.round', overload='default')>: <function round at 0x7776e1c663e0>, <OpOverload(op='aten.round', overload='decimals_out')>: <function round at 0x7776e1c663e0>, <OpOverload(op='aten.round', overload='out')>: <function round at 0x7776e1c663e0>, <OpOverload(op='aten.tanh_backward', overload='grad_input')>: <function tanh_backward at 0x7776e2122840>, <OpOverload(op='aten.rsqrt', overload='out')>: <function rsqrt at 0x7776e1c668e0>, <OpOverload(op='aten.sign', overload='out')>: <function sign at 0x7776e1c677e0>, <OpOverload(op='aten.signbit', overload='default')>: <function signbit at 0x7776e1c67ce0>, <OpOverload(op='aten.signbit', overload='out')>: <function signbit at 0x7776e1c67ce0>, <OpOverload(op='aten.sin', overload='out')>: <function sin at 0x7776e1c74220>, <OpOverload(op='aten.sin', overload='default')>: <function sin at 0x7776e1c74220>, <OpOverload(op='aten.bitwise_and', overload='Tensor_out')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.sinh', overload='default')>: <function sinh at 0x7776e1c66980>, <OpOverload(op='aten.cudnn_batch_norm', overload='default')>: <function cudnn_batch_norm at 0x7776e1d2a5c0>, <OpOverload(op='aten.sqrt', overload='out')>: <function sqrt at 0x7776e1c749a0>, <OpOverload(op='aten.sqrt', overload='default')>: <function sqrt at 0x7776e1c749a0>, <OpOverload(op='aten.lt', overload='Tensor')>: <function lt at 0x7776e1cadbc0>, <OpOverload(op='aten.tanh', overload='out')>: <function tanh at 0x7776e1c758a0>, <OpOverload(op='aten.tanh', overload='default')>: <function tanh at 0x7776e1c758a0>, <OpOverload(op='aten.trunc', overload='out')>: <function trunc at 0x7776e1c75da0>, <OpOverload(op='aten.trunc', overload='default')>: <function trunc at 0x7776e1c75da0>, <OpOverload(op='aten.add', overload='Scalar')>: <function add at 0x7776e1c762a0>, <OpOverload(op='aten.add', overload='Tensor')>: <function add at 0x7776e1c762a0>, <OpOverload(op='aten.add', overload='out')>: <function add at 0x7776e1c762a0>, <OpOverload(op='aten.add', overload='Scalar_out')>: <function add at 0x7776e1c762a0>, <OpOverload(op='aten.narrow_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.narrow at 0x7776e1b98a40>, <OpOverload(op='aten.atan2', overload='out')>: <function atan2 at 0x7776e1c76700>, <OpOverload(op='aten.atan2', overload='default')>: <function atan2 at 0x7776e1c76700>, <OpOverload(op='aten.bitwise_and', overload='Tensor')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_and', overload='Scalar')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_and', overload='Scalar_out')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_and', overload='Scalar_Tensor_out')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_or', overload='Scalar')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_or', overload='Tensor_out')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_or', overload='Tensor')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_or', overload='Scalar_Tensor_out')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_or', overload='Scalar_out')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_xor', overload='Scalar_Tensor')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.bitwise_xor', overload='Scalar_Tensor_out')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.bitwise_xor', overload='Scalar')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.bitwise_xor', overload='Tensor')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.bitwise_xor', overload='Scalar_out')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.div', overload='out')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='out_mode')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='Scalar_out')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='Scalar_mode_out')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.eq', overload='Tensor')>: <function eq at 0x7776e1c94180>, <OpOverload(op='aten.eq', overload='Scalar')>: <function eq at 0x7776e1c94180>, <OpOverload(op='aten.eq', overload='Scalar_out')>: <function eq at 0x7776e1c94180>, <OpOverload(op='aten.eq', overload='Tensor_out')>: <function eq at 0x7776e1c94180>, <OpOverload(op='aten.fmax', overload='out')>: <function fmax at 0x7776e1c95120>, <OpOverload(op='aten.binary_cross_entropy_backward', overload='default')>: <function binary_cross_entropy_backward at 0x7776e1cfef20>, <OpOverload(op='aten.fmax', overload='default')>: <function fmax at 0x7776e1c95120>, <OpOverload(op='aten.fmin', overload='default')>: <function fmin at 0x7776e1c95580>, <OpOverload(op='aten.fmin', overload='out')>: <function fmin at 0x7776e1c95580>, <OpOverload(op='aten.fmod', overload='Scalar')>: <function fmod at 0x7776e1c959e0>, <OpOverload(op='aten.log10_', overload='default')>: <function log10 at 0x7776e1b67060>, <OpOverload(op='aten.fmod', overload='Tensor')>: <function fmod at 0x7776e1c959e0>, <OpOverload(op='aten.fmod', overload='Tensor_out')>: <function fmod at 0x7776e1c959e0>, <OpOverload(op='aten.igammac', overload='out')>: <function igammac at 0x7776e1c979c0>, <OpOverload(op='aten.fmod', overload='Scalar_out')>: <function fmod at 0x7776e1c959e0>, <OpOverload(op='aten.gcd', overload='default')>: <function gcd at 0x7776e1c76200>, <OpOverload(op='aten.gcd', overload='out')>: <function gcd at 0x7776e1c76200>, <OpOverload(op='aten.addmv', overload='out')>: <function addmv at 0x7776e1d2a0c0>, <OpOverload(op='aten.ge', overload='Tensor')>: <function ge at 0x7776e1c963e0>, <OpOverload(op='aten.ge', overload='Scalar')>: <function ge at 0x7776e1c963e0>, <OpOverload(op='aten.ge', overload='Tensor_out')>: <function ge at 0x7776e1c963e0>, <OpOverload(op='aten.xlogy', overload='Tensor')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.ge', overload='Scalar_out')>: <function ge at 0x7776e1c963e0>, <OpOverload(op='aten.gt', overload='Scalar')>: <function gt at 0x7776e1c96840>, <OpOverload(op='aten.gt', overload='Tensor')>: <function gt at 0x7776e1c96840>, <OpOverload(op='aten.hardswish_backward', overload='out')>: <function hardswish_backward at 0x7776e1cde200>, <OpOverload(op='aten.gt', overload='Scalar_out')>: <function gt at 0x7776e1c96840>, <OpOverload(op='aten.gt', overload='Tensor_out')>: <function gt at 0x7776e1c96840>, <OpOverload(op='aten.hypot', overload='default')>: <function hypot at 0x7776e1c97100>, <OpOverload(op='aten.hypot', overload='out')>: <function hypot at 0x7776e1c97100>, <OpOverload(op='aten.index_fill', overload='Dimname_Scalar')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.igamma', overload='out')>: <function igamma at 0x7776e1c97560>, <OpOverload(op='aten.igamma', overload='default')>: <function igamma at 0x7776e1c97560>, <OpOverload(op='aten.sum', overload='out')>: <function sum_default at 0x7776e1dbc400>, <OpOverload(op='aten.igammac', overload='default')>: <function igammac at 0x7776e1c979c0>, <OpOverload(op='aten.le', overload='Tensor')>: <function le at 0x7776e1cac400>, <OpOverload(op='aten.le', overload='Scalar')>: <function le at 0x7776e1cac400>, <OpOverload(op='aten.le', overload='Tensor_out')>: <function le at 0x7776e1cac400>, <OpOverload(op='aten.le', overload='Scalar_out')>: <function le at 0x7776e1cac400>, <OpOverload(op='aten.lt', overload='Scalar')>: <function lt at 0x7776e1cadbc0>, <OpOverload(op='aten.lt', overload='Scalar_out')>: <function lt at 0x7776e1cadbc0>, <OpOverload(op='aten.lt', overload='Tensor_out')>: <function lt at 0x7776e1cadbc0>, <OpOverload(op='aten.maximum', overload='default')>: <function maximum at 0x7776e1cae020>, <OpOverload(op='aten.maximum', overload='out')>: <function maximum at 0x7776e1cae020>, <OpOverload(op='aten.minimum', overload='default')>: <function minimum at 0x7776e1cae480>, <OpOverload(op='aten.minimum', overload='out')>: <function minimum at 0x7776e1cae480>, <OpOverload(op='aten.mul', overload='Scalar_out')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.rnn_tanh', overload='input')>: <function rnn_tanh_input at 0x7776e1d79760>, <OpOverload(op='aten.ne', overload='Tensor')>: <function ne at 0x7776e1caed40>, <OpOverload(op='aten.acosh', overload='out')>: <function acosh at 0x7776e1c3c400>, <OpOverload(op='aten.acosh', overload='default')>: <function acosh at 0x7776e1c3c400>, <OpOverload(op='aten.asin', overload='out')>: <function asin at 0x7776e1c3c900>, <OpOverload(op='aten.asin', overload='default')>: <function asin at 0x7776e1c3c900>, <OpOverload(op='aten.asinh', overload='default')>: <function asinh at 0x7776e1c3ce00>, <OpOverload(op='aten.asinh', overload='out')>: <function asinh at 0x7776e1c3ce00>, <OpOverload(op='aten.lgamma', overload='default')>: <function lgamma at 0x7776e1c4fd80>, <OpOverload(op='aten.atan', overload='default')>: <function atan at 0x7776e1c3d300>, <OpOverload(op='aten.atanh', overload='out')>: <function atanh at 0x7776e1c3d800>, <OpOverload(op='aten.atan', overload='out')>: <function atan at 0x7776e1c3d300>, <OpOverload(op='aten.atanh', overload='default')>: <function atanh at 0x7776e1c3d800>, <OpOverload(op='aten.cos', overload='default')>: <function cos at 0x7776e1c3eac0>, <OpOverload(op='aten.cos', overload='out')>: <function cos at 0x7776e1c3eac0>, <OpOverload(op='aten.cosh', overload='out')>: <function cosh at 0x7776e1c3efc0>, <OpOverload(op='aten.cosh', overload='default')>: <function cosh at 0x7776e1c3efc0>, <OpOverload(op='aten.bitwise_and', overload='Scalar_Tensor')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_not', overload='out')>: <function bitwise_not at 0x7776e1c3dd00>, <OpOverload(op='aten.bitwise_not', overload='default')>: <function bitwise_not at 0x7776e1c3dd00>, <OpOverload(op='aten.ceil', overload='default')>: <function ceil at 0x7776e1c3e200>, <OpOverload(op='aten.index_fill_', overload='int_Scalar')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.clone', overload='default')>: <function clone at 0x7776e1ad19e0>, <OpOverload(op='aten.ceil', overload='out')>: <function ceil at 0x7776e1c3e200>, <OpOverload(op='aten.conj_physical', overload='out')>: <function conj_physical at 0x7776e1c3e5c0>, <OpOverload(op='aten.conj_physical', overload='default')>: <function conj_physical at 0x7776e1c3e5c0>, <OpOverload(op='aten.digamma', overload='default')>: <function digamma at 0x7776e1c3f4c0>, <OpOverload(op='aten.bitwise_or', overload='Scalar_Tensor')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.digamma', overload='out')>: <function digamma at 0x7776e1c3f4c0>, <OpOverload(op='aten.sinh', overload='out')>: <function sinh at 0x7776e1c66980>, <OpOverload(op='aten.erf', overload='default')>: <function erf at 0x7776e1c3d3a0>, <OpOverload(op='aten.erf', overload='out')>: <function erf at 0x7776e1c3d3a0>, <OpOverload(op='aten.erfc', overload='default')>: <function erfc at 0x7776e1c3fce0>, <OpOverload(op='aten.sign', overload='default')>: <function sign at 0x7776e1c677e0>, <OpOverload(op='aten.exp', overload='default')>: <function exp at 0x7776e1c4c220>, <OpOverload(op='aten.exp', overload='out')>: <function exp at 0x7776e1c4c220>, <OpOverload(op='aten.expm1', overload='out')>: <function expm1 at 0x7776e1c4c720>, <OpOverload(op='aten.expm1', overload='default')>: <function expm1 at 0x7776e1c4c720>, <OpOverload(op='aten.exp2', overload='default')>: <function exp2 at 0x7776e1c4cc20>, <OpOverload(op='aten.exp2', overload='out')>: <function exp2 at 0x7776e1c4cc20>, <OpOverload(op='aten.fill', overload='Scalar')>: <function fill_scalar at 0x7776e1cdea20>, <OpOverload(op='aten.transpose', overload='Dimname')>: <function transpose at 0x7776e1b10f40>, <OpOverload(op='aten.mul', overload='out')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.fill', overload='Tensor')>: <function fill_tensor at 0x7776e1cdeac0>, <OpOverload(op='aten.floor', overload='out')>: <function floor at 0x7776e1c4d760>, <OpOverload(op='aten.floor', overload='default')>: <function floor at 0x7776e1c4d760>, <OpOverload(op='aten.clone', overload='out')>: <function clone at 0x7776e1ad19e0>, <OpOverload(op='aten.abs', overload='default')>: <function abs at 0x7776e1dbede0>, <OpOverload(op='aten.acos', overload='out')>: <function acos at 0x7776e1dbfd80>, <OpOverload(op='aten.acos', overload='default')>: <function acos at 0x7776e1dbfd80>, <OpOverload(op='aten.log2', overload='out')>: <function log2 at 0x7776e1c64cc0>, <OpOverload(op='aten.rad2deg', overload='default')>: <function rad2deg at 0x7776e1b431a0>, <OpOverload(op='aten.mul', overload='Tensor')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.mul', overload='Scalar')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.div', overload='Tensor')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='Scalar')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.upsample_nearest2d', overload='vec')>: <function _upsample_nearest_vec at 0x7776e1d78220>, <OpOverload(op='aten.tanh_backward', overload='default')>: <function tanh_backward at 0x7776e2122840>, <OpOverload(op='aten.abs', overload='out')>: <function abs at 0x7776e1dbede0>, <OpOverload(op='aten.sym_numel', overload='default')>: <function sym_numel at 0x7776e1d96340>, <OpOverload(op='aten.diagonal_scatter', overload='default')>: <function diagonal_scatter at 0x7776e1b104a0>, <OpOverload(op='aten.diagonal', overload='default')>: <function diagonal at 0x7776e1b102c0>, <OpOverload(op='aten.select_scatter', overload='default')>: <function select_scatter at 0x7776e1b43ec0>, <OpOverload(op='aten.upsample_bilinear2d', overload='vec')>: <function _upsample_linear_vec at 0x7776e1d7b100>, <OpOverload(op='aten.zeros', overload='default')>: <function zeros at 0x7776e1b11a80>, <OpOverload(op='aten.detach', overload='default')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.as_strided_scatter', overload='default')>: <function as_strided_scatter at 0x7776e1af40e0>, <OpOverload(op='aten.empty_strided', overload='default')>: <function empty_strided at 0x7776e1b13d80>, <OpOverload(op='aten.view', overload='default')>: <function view at 0x7776e1b118a0>, <OpOverload(op='aten._unsafe_view', overload='default')>: <function _reshape_alias at 0x7776e1d7bf60>, <OpOverload(op='aten.lift_fresh', overload='default')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.empty_like', overload='default')>: <function empty_like at 0x7776e1b12c00>, <OpOverload(op='aten.ones_like', overload='out')>: <function ones_like at 0x7776e1b40180>, <OpOverload(op='aten.zeros_like', overload='default')>: <function zeros_like at 0x7776e1b116c0>, <OpOverload(op='aten.zeros_like', overload='out')>: <function zeros_like at 0x7776e1b116c0>, <OpOverload(op='aten.new_empty', overload='default')>: <function new_empty at 0x7776e1b12480>, <OpOverload(op='aten.new_empty', overload='out')>: <function new_empty at 0x7776e1b12480>, <OpOverload(op='aten.new_empty_strided', overload='out')>: <function new_empty_strided at 0x7776e1b12700>, <OpOverload(op='aten.new_full', overload='default')>: <function new_full at 0x7776e1b12840>, <OpOverload(op='aten.new_full', overload='out')>: <function new_full at 0x7776e1b12840>, <OpOverload(op='aten.new_zeros', overload='default')>: <function new_zeros at 0x7776e1b11440>, <OpOverload(op='aten.new_zeros', overload='out')>: <function new_zeros at 0x7776e1b11440>, <OpOverload(op='aten.new_ones', overload='default')>: <function new_ones at 0x7776e1b10400>, <OpOverload(op='aten.new_ones', overload='out')>: <function new_ones at 0x7776e1b10400>, <OpOverload(op='aten.view_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.item', overload='default')>: <function item at 0x7776e1ad1800>, <OpOverload(op='aten._unsafe_index_put', overload='default')>: <function _unsafe_index_put at 0x7776e1d940e0>, <OpOverload(op='aten.slice_scatter', overload='default')>: <function slice_scatter at 0x7776e1cffc40>, <OpOverload(op='aten.slice_scatter', overload='out')>: <function slice_scatter at 0x7776e1cffc40>, <OpOverload(op='aten.index_put_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1d96ac0>, <OpOverload(op='aten.erfc', overload='out')>: <function erfc at 0x7776e1c3fce0>, <OpOverload(op='aten.new_empty_strided', overload='default')>: <function new_empty_strided at 0x7776e1b12700>, <OpOverload(op='aten.ones_like', overload='default')>: <function ones_like at 0x7776e1b40180>, <OpOverload(op='aten.empty_like', overload='out')>: <function empty_like at 0x7776e1b12c00>}
- experimental_experiment.torch_dynamo.get_decomposition_table()[source]¶
Returns the decomposition table needed to translate backward graph into onnx. It should used as follows:
import torch from torch._dynamo.backends.common import aot_autograd from experimental_experiment.torch_dynamo import get_decomposition_table aot_compiler = aot_autograd( fw_compiler=backend_debug, decompositions=get_decomposition_table() ) compiled_model = torch.compile( model, backend=aot_compiler, dynamic=dynamic, fullgraph=fullgraph, )
The value is:
<<<
import pprint from experimental_experiment.torch_dynamo import get_decomposition_table pprint.pprint(get_decomposition_table())
>>>
{<OpOverload(op='aten.native_layer_norm_backward', overload='default')>: <function native_layer_norm_backward at 0x7776e1d29080>, <OpOverload(op='aten.embedding_dense_backward', overload='default')>: <function embedding_dense_backward at 0x7776e1d2a160>}
- experimental_experiment.torch_dynamo.get_decomposition_table_by_name(name: str)[source]¶
Returns a predefined decomposition table.
- Parameters:
name – name see below
- Returns:
decomposition table
‘none’: do not decompose
‘default’:
get_decomposition_table()‘onnxscript’:
get_decomposition_table_onnxscript()‘dynamo’:
get_decomposition_table_dynamo()
- experimental_experiment.torch_dynamo.get_decomposition_table_dynamo(onnx_registry=None)[source]¶
Returns the decomposition table needed for the dynamo exporter.
- Parameters:
onnx_registry – sent to
create_onnx_friendly_decomposition_table
The value is:
<<<
import pprint from experimental_experiment.torch_dynamo import get_decomposition_table_dynamo pprint.pprint(get_decomposition_table_dynamo())
>>>
{<OpOverload(op='aten._backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354c220>, kernel=<OpOverload(op='aten._backward', overload='default')>), <OpOverload(op='aten._test_check_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968ea0>, kernel=<OpOverload(op='aten._test_check_tensor', overload='default')>), <OpOverload(op='image._is_compiled_against_turbo', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c31b240>, kernel=<OpOverload(op='image._is_compiled_against_turbo', overload='default')>), <OpOverload(op='aten.linalg_svd', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329919e0>, kernel=<OpOverload(op='aten.linalg_svd', overload='default')>), <OpOverload(op='aten.gradient', overload='tensorarrayint')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329913a0>, kernel=<OpOverload(op='aten.gradient', overload='tensorarrayint')>), <OpOverload(op='aten.masked_select_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329916c0>, kernel=<OpOverload(op='aten.masked_select_backward', overload='default')>), <OpOverload(op='aten._grid_sampler_2d_cpu_fallback_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632d191c0>, kernel=<OpOverload(op='aten._grid_sampler_2d_cpu_fallback_backward', overload='default')>), <OpOverload(op='aten._saturate_weight_to_fp16', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354dee0>, kernel=<OpOverload(op='aten._saturate_weight_to_fp16', overload='default')>), <OpOverload(op='aten.quantile', overload='scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539e980>, kernel=<OpOverload(op='aten.quantile', overload='scalar')>), <OpOverload(op='aten._pad_enum', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b740>, kernel=<OpOverload(op='aten._pad_enum', overload='default')>), <OpOverload(op='aten.lstm_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394af20>, kernel=<OpOverload(op='aten.lstm_cell', overload='default')>), <OpOverload(op='aten._weight_norm_differentiable_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7740>, kernel=<OpOverload(op='aten._weight_norm_differentiable_backward', overload='default')>), <OpOverload(op='aten.less_equal', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ebba0>, kernel=<OpOverload(op='aten.less_equal', overload='Tensor')>), <OpOverload(op='aten.logdet', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776345bb2e0>, kernel=<OpOverload(op='aten.logdet', overload='default')>), <OpOverload(op='aten.linalg_solve_ex', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539cfe0>, kernel=<OpOverload(op='aten.linalg_solve_ex', overload='default')>), <OpOverload(op='aten.sparse_csr_tensor', overload='crow_col_value_size')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496c900>, kernel=<OpOverload(op='aten.sparse_csr_tensor', overload='crow_col_value_size')>), <OpOverload(op='aten.gradient', overload='scalararray')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296aa20>, kernel=<OpOverload(op='aten.gradient', overload='scalararray')>), <OpOverload(op='aten.retains_grad', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b240>, kernel=<OpOverload(op='aten.retains_grad', overload='default')>), <OpOverload(op='aten.to_sparse_bsc', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354c2c0>, kernel=<OpOverload(op='aten.to_sparse_bsc', overload='default')>), <OpOverload(op='aten.linalg_svdvals', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d696c0>, kernel=<OpOverload(op='aten.linalg_svdvals', overload='default')>), <OpOverload(op='aten.special_gammainc', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7ec0>, kernel=<OpOverload(op='aten.special_gammainc', overload='default')>), <OpOverload(op='image.decode_image', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d7337e0>, kernel=<OpOverload(op='image.decode_image', overload='default')>), <OpOverload(op='aten._test_serialization_subcmul', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d1c0>, kernel=<OpOverload(op='aten._test_serialization_subcmul', overload='default')>), <OpOverload(op='aten.fake_quantize_per_tensor_affine', overload='tensor_qparams')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634179c60>, kernel=<OpOverload(op='aten.fake_quantize_per_tensor_affine', overload='tensor_qparams')>), <OpOverload(op='image.encode_png', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de40180>, kernel=<OpOverload(op='image.encode_png', overload='default')>), <OpOverload(op='aten.sparse_csc_tensor', overload='ccol_row_value')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b9c0>, kernel=<OpOverload(op='aten.sparse_csc_tensor', overload='ccol_row_value')>), <OpOverload(op='aten.sparse_bsc_tensor', overload='ccol_row_value')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a0c0>, kernel=<OpOverload(op='aten.sparse_bsc_tensor', overload='ccol_row_value')>), <OpOverload(op='aten.clip', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eaf20>, kernel=<OpOverload(op='aten.clip', overload='Tensor')>), <OpOverload(op='aten.scatter_add', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992160>, kernel=<OpOverload(op='aten.scatter_add', overload='dimname')>), <OpOverload(op='aten.unflatten_dense_tensors', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296afc0>, kernel=<OpOverload(op='aten.unflatten_dense_tensors', overload='default')>), <OpOverload(op='image.decode_webp', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c5eb9c0>, kernel=<OpOverload(op='image.decode_webp', overload='default')>), <OpOverload(op='aten._choose_qparams_per_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539d120>, kernel=<OpOverload(op='aten._choose_qparams_per_tensor', overload='default')>), <OpOverload(op='aten.thnn_conv2d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296ae80>, kernel=<OpOverload(op='aten.thnn_conv2d', overload='default')>), <OpOverload(op='aten.linalg_det', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969e40>, kernel=<OpOverload(op='aten.linalg_det', overload='default')>), <OpOverload(op='aten.concat', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a020>, kernel=<OpOverload(op='aten.concat', overload='default')>), <OpOverload(op='aten.is_vulkan_available', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969440>, kernel=<OpOverload(op='aten.is_vulkan_available', overload='default')>), <OpOverload(op='aten.concat', overload='names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776deec4f40>, kernel=<OpOverload(op='aten.concat', overload='names')>), <OpOverload(op='aten.sym_is_contiguous', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632957c40>, kernel=<OpOverload(op='aten.sym_is_contiguous', overload='default')>), <OpOverload(op='aten.linalg_matrix_rank', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296bc40>, kernel=<OpOverload(op='aten.linalg_matrix_rank', overload='default')>), <OpOverload(op='aten.linalg_matrix_power', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354e5c0>, kernel=<OpOverload(op='aten.linalg_matrix_power', overload='default')>), <OpOverload(op='aten.is_distributed', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d694e0>, kernel=<OpOverload(op='aten.is_distributed', overload='default')>), <OpOverload(op='aten.arctanh', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e65c0>, kernel=<OpOverload(op='aten.arctanh', overload='default')>), <OpOverload(op='aten.linalg_tensorinv', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b1a0>, kernel=<OpOverload(op='aten.linalg_tensorinv', overload='default')>), <OpOverload(op='aten.softmax', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296be20>, kernel=<OpOverload(op='aten.softmax', overload='Dimname')>), <OpOverload(op='aten.scatter', overload='dimname_value')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968c20>, kernel=<OpOverload(op='aten.scatter', overload='dimname_value')>), <OpOverload(op='aten.vander', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632957ce0>, kernel=<OpOverload(op='aten.vander', overload='default')>), <OpOverload(op='aten.cumulative_trapezoid', overload='x')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296aac0>, kernel=<OpOverload(op='aten.cumulative_trapezoid', overload='x')>), <OpOverload(op='aten.arcsinh', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539f880>, kernel=<OpOverload(op='aten.arcsinh', overload='default')>), <OpOverload(op='aten.hstack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296ab60>, kernel=<OpOverload(op='aten.hstack', overload='default')>), <OpOverload(op='aten.subtract', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d68d60>, kernel=<OpOverload(op='aten.subtract', overload='Scalar')>), <OpOverload(op='aten.linalg_matrix_rank', overload='atol_rtol_float')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992520>, kernel=<OpOverload(op='aten.linalg_matrix_rank', overload='atol_rtol_float')>), <OpOverload(op='aten.greater_equal', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a3e0>, kernel=<OpOverload(op='aten.greater_equal', overload='Tensor')>), <OpOverload(op='aten.to_sparse', overload='sparse_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a480>, kernel=<OpOverload(op='aten.to_sparse', overload='sparse_dim')>), <OpOverload(op='aten.ger', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573fb00>, kernel=<OpOverload(op='aten.ger', overload='default')>), <OpOverload(op='aten.is_signed', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776341791c0>, kernel=<OpOverload(op='aten.is_signed', overload='default')>), <OpOverload(op='aten.type_as', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763417a020>, kernel=<OpOverload(op='aten.type_as', overload='default')>), <OpOverload(op='aten.conv2d', overload='padding')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394bb00>, kernel=<OpOverload(op='aten.conv2d', overload='padding')>), <OpOverload(op='aten.atleast_2d', overload='Sequence')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539d1c0>, kernel=<OpOverload(op='aten.atleast_2d', overload='Sequence')>), <OpOverload(op='aten._convolution', overload='deprecated')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763417aac0>, kernel=<OpOverload(op='aten._convolution', overload='deprecated')>), <OpOverload(op='aten.linalg_multi_dot', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776defd89a0>, kernel=<OpOverload(op='aten.linalg_multi_dot', overload='default')>), <OpOverload(op='aten.linalg_slogdet', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6bec0>, kernel=<OpOverload(op='aten.linalg_slogdet', overload='default')>), <OpOverload(op='aten.linalg_matrix_rank', overload='atol_rtol_tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763417aa20>, kernel=<OpOverload(op='aten.linalg_matrix_rank', overload='atol_rtol_tensor')>), <OpOverload(op='aten._version', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394aca0>, kernel=<OpOverload(op='aten._version', overload='default')>), <OpOverload(op='aten.matrix_exp_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394aa20>, kernel=<OpOverload(op='aten.matrix_exp_backward', overload='default')>), <OpOverload(op='aten.divide', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990a40>, kernel=<OpOverload(op='aten.divide', overload='Tensor')>), <OpOverload(op='aten.get_gradients', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990fe0>, kernel=<OpOverload(op='aten.get_gradients', overload='default')>), <OpOverload(op='aten.rms_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7ce0>, kernel=<OpOverload(op='aten.rms_norm', overload='default')>), <OpOverload(op='aten.special_psi', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990d60>, kernel=<OpOverload(op='aten.special_psi', overload='default')>), <OpOverload(op='c10d_functional.all_reduce', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c3bc5e0>, kernel=<OpOverload(op='c10d_functional.all_reduce', overload='default')>), <OpOverload(op='aten.embedding_bag', overload='padding_idx')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990f40>, kernel=<OpOverload(op='aten.embedding_bag', overload='padding_idx')>), <OpOverload(op='aten._propagate_xla_data', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329684a0>, kernel=<OpOverload(op='aten._propagate_xla_data', overload='default')>), <OpOverload(op='aten.outer', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329696c0>, kernel=<OpOverload(op='aten.outer', overload='default')>), <OpOverload(op='aten.choose_qparams_optimized', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968360>, kernel=<OpOverload(op='aten.choose_qparams_optimized', overload='default')>), <OpOverload(op='aten.conv_transpose3d', overload='input')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394bce0>, kernel=<OpOverload(op='aten.conv_transpose3d', overload='input')>), <OpOverload(op='aten.fbgemm_linear_int8_weight_fp32_activation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e79c0>, kernel=<OpOverload(op='aten.fbgemm_linear_int8_weight_fp32_activation', overload='default')>), <OpOverload(op='aten.min', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d69580>, kernel=<OpOverload(op='aten.min', overload='names_dim')>), <OpOverload(op='aten._wrapped_linear_prepack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6ae80>, kernel=<OpOverload(op='aten._wrapped_linear_prepack', overload='default')>), <OpOverload(op='aten.output_nr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991e40>, kernel=<OpOverload(op='aten.output_nr', overload='default')>), <OpOverload(op='aten._sparse_mm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394b600>, kernel=<OpOverload(op='aten._sparse_mm', overload='default')>), <OpOverload(op='aten.conv_transpose2d', overload='input')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573e3e0>, kernel=<OpOverload(op='aten.conv_transpose2d', overload='input')>), <OpOverload(op='aten.cumprod_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a160>, kernel=<OpOverload(op='aten.cumprod_backward', overload='default')>), <OpOverload(op='aten.stride', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b605e0>, kernel=<OpOverload(op='aten.stride', overload='Dimname')>), <OpOverload(op='aten.cummin', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329923e0>, kernel=<OpOverload(op='aten.cummin', overload='dimname')>), <OpOverload(op='aten.max', overload='other')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eb740>, kernel=<OpOverload(op='aten.max', overload='other')>), <OpOverload(op='image.decode_jpegs_cuda', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9c60>, kernel=<OpOverload(op='image.decode_jpegs_cuda', overload='default')>), <OpOverload(op='aten.cummax', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eba60>, kernel=<OpOverload(op='aten.cummax', overload='dimname')>), <OpOverload(op='aten._use_cudnn_rnn_flatten_weight', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b880>, kernel=<OpOverload(op='aten._use_cudnn_rnn_flatten_weight', overload='default')>), <OpOverload(op='aten.nll_loss2d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776345bb1a0>, kernel=<OpOverload(op='aten.nll_loss2d', overload='default')>), <OpOverload(op='aten.isreal', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991bc0>, kernel=<OpOverload(op='aten.isreal', overload='default')>), <OpOverload(op='aten.rrelu', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991800>, kernel=<OpOverload(op='aten.rrelu', overload='default')>), <OpOverload(op='prepacked.unpack_prepacked_sizes_linear', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de42980>, kernel=<OpOverload(op='prepacked.unpack_prepacked_sizes_linear', overload='default')>), <OpOverload(op='aten.not_equal', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573e520>, kernel=<OpOverload(op='aten.not_equal', overload='Tensor')>), <OpOverload(op='aten._dim_arange', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573e020>, kernel=<OpOverload(op='aten._dim_arange', overload='default')>), <OpOverload(op='aten.linalg_tensorsolve', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e3e0>, kernel=<OpOverload(op='aten.linalg_tensorsolve', overload='default')>), <OpOverload(op='_test.get_first', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c2ab7e0>, kernel=<OpOverload(op='_test.get_first', overload='default')>), <OpOverload(op='c10d_functional.reduce_scatter_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c3384a0>, kernel=<OpOverload(op='c10d_functional.reduce_scatter_tensor', overload='default')>), <OpOverload(op='aten._convolution_double_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991ee0>, kernel=<OpOverload(op='aten._convolution_double_backward', overload='default')>), <OpOverload(op='aten.fbgemm_linear_fp16_weight_fp32_activation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992020>, kernel=<OpOverload(op='aten.fbgemm_linear_fp16_weight_fp32_activation', overload='default')>), <OpOverload(op='aten.less', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e480>, kernel=<OpOverload(op='aten.less', overload='Scalar')>), <OpOverload(op='aten._remove_batch_dim', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7f60>, kernel=<OpOverload(op='aten._remove_batch_dim', overload='default')>), <OpOverload(op='aten.frobenius_norm', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d68ae0>, kernel=<OpOverload(op='aten.frobenius_norm', overload='dim')>), <OpOverload(op='aten.special_xlogy', overload='other_scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632d19260>, kernel=<OpOverload(op='aten.special_xlogy', overload='other_scalar')>), <OpOverload(op='aten.linalg_vecdot', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a980>, kernel=<OpOverload(op='aten.linalg_vecdot', overload='default')>), <OpOverload(op='aten.linalg_cholesky', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991300>, kernel=<OpOverload(op='aten.linalg_cholesky', overload='default')>), <OpOverload(op='aten.special_i0', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990ae0>, kernel=<OpOverload(op='aten.special_i0', overload='default')>), <OpOverload(op='aten.inner', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991b20>, kernel=<OpOverload(op='aten.inner', overload='default')>), <OpOverload(op='aten.multiply', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329920c0>, kernel=<OpOverload(op='aten.multiply', overload='Scalar')>), <OpOverload(op='aten.absolute', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969120>, kernel=<OpOverload(op='aten.absolute', overload='default')>), <OpOverload(op='aten._pad_packed_sequence', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354e160>, kernel=<OpOverload(op='aten._pad_packed_sequence', overload='default')>), <OpOverload(op='aten._cufft_get_plan_cache_size', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b61440>, kernel=<OpOverload(op='aten._cufft_get_plan_cache_size', overload='default')>), <OpOverload(op='aten.linalg_cond', overload='p_str')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d940>, kernel=<OpOverload(op='aten.linalg_cond', overload='p_str')>), <OpOverload(op='aten.special_xlogy', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632955e40>, kernel=<OpOverload(op='aten.special_xlogy', overload='default')>), <OpOverload(op='aten.greater', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968220>, kernel=<OpOverload(op='aten.greater', overload='Scalar')>), <OpOverload(op='aten.linalg_norm', overload='ord_str')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573e5c0>, kernel=<OpOverload(op='aten.linalg_norm', overload='ord_str')>), <OpOverload(op='aten.l1_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632993100>, kernel=<OpOverload(op='aten.l1_loss', overload='default')>), <OpOverload(op='aten.bilinear', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eae80>, kernel=<OpOverload(op='aten.bilinear', overload='default')>), <OpOverload(op='aten.fbgemm_pack_gemm_matrix_fp16', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992480>, kernel=<OpOverload(op='aten.fbgemm_pack_gemm_matrix_fp16', overload='default')>), <OpOverload(op='aten.conv_tbc_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a5c0>, kernel=<OpOverload(op='aten.conv_tbc_backward', overload='default')>), <OpOverload(op='aten._cast_Long', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634178c20>, kernel=<OpOverload(op='aten._cast_Long', overload='default')>), <OpOverload(op='aten._validate_sparse_compressed_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eb1a0>, kernel=<OpOverload(op='aten._validate_sparse_compressed_tensor_args', overload='default')>), <OpOverload(op='aten.sparse_bsr_tensor', overload='crow_col_value_size')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763417a8e0>, kernel=<OpOverload(op='aten.sparse_bsr_tensor', overload='crow_col_value_size')>), <OpOverload(op='aten.repeat_interleave', overload='self_int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573ede0>, kernel=<OpOverload(op='aten.repeat_interleave', overload='self_int')>), <OpOverload(op='aten.gather', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763417ab60>, kernel=<OpOverload(op='aten.gather', overload='dimname')>), <OpOverload(op='aten.cross', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d9e0>, kernel=<OpOverload(op='aten.cross', overload='default')>), <OpOverload(op='aten._lu_with_info', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b60900>, kernel=<OpOverload(op='aten._lu_with_info', overload='default')>), <OpOverload(op='aten.data', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992a20>, kernel=<OpOverload(op='aten.data', overload='default')>), <OpOverload(op='aten._is_zerotensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b637e0>, kernel=<OpOverload(op='aten._is_zerotensor', overload='default')>), <OpOverload(op='aten.min', overload='other')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b380>, kernel=<OpOverload(op='aten.min', overload='other')>), <OpOverload(op='aten.cudnn_is_acceptable', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573eac0>, kernel=<OpOverload(op='aten.cudnn_is_acceptable', overload='default')>), <OpOverload(op='aten.linalg_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573e340>, kernel=<OpOverload(op='aten.linalg_norm', overload='default')>), <OpOverload(op='aten.gradient', overload='scalarrayint')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573f4c0>, kernel=<OpOverload(op='aten.gradient', overload='scalarrayint')>), <OpOverload(op='aten._sparse_softmax', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573fba0>, kernel=<OpOverload(op='aten._sparse_softmax', overload='int')>), <OpOverload(op='aten.nanmedian', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969800>, kernel=<OpOverload(op='aten.nanmedian', overload='names_dim')>), <OpOverload(op='aten.result_type', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573f6a0>, kernel=<OpOverload(op='aten.result_type', overload='Scalar')>), <OpOverload(op='image.decode_gif', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c399080>, kernel=<OpOverload(op='image.decode_gif', overload='default')>), <OpOverload(op='profiler._record_function_exit', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de42840>, kernel=<OpOverload(op='profiler._record_function_exit', overload='default')>), <OpOverload(op='aten.linalg_eigvalsh', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990860>, kernel=<OpOverload(op='aten.linalg_eigvalsh', overload='default')>), <OpOverload(op='aten._cast_Byte', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e340>, kernel=<OpOverload(op='aten._cast_Byte', overload='default')>), <OpOverload(op='aten.to_sparse', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496c7c0>, kernel=<OpOverload(op='aten.to_sparse', overload='default')>), <OpOverload(op='aten._test_ambiguous_defaults', overload='a')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634178180>, kernel=<OpOverload(op='aten._test_ambiguous_defaults', overload='a')>), <OpOverload(op='aten.conv1d', overload='padding')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b1a0>, kernel=<OpOverload(op='aten.conv1d', overload='padding')>), <OpOverload(op='aten._validate_sparse_csc_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968900>, kernel=<OpOverload(op='aten._validate_sparse_csc_tensor_args', overload='default')>), <OpOverload(op='aten.avg_pool1d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969c60>, kernel=<OpOverload(op='aten.avg_pool1d', overload='default')>), <OpOverload(op='aten._cast_Int', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eb560>, kernel=<OpOverload(op='aten._cast_Int', overload='default')>), <OpOverload(op='aten._sparse_coo_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969d00>, kernel=<OpOverload(op='aten._sparse_coo_tensor_unsafe', overload='default')>), <OpOverload(op='aten.fake_quantize_per_tensor_affine_cachemask_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ea840>, kernel=<OpOverload(op='aten.fake_quantize_per_tensor_affine_cachemask_backward', overload='default')>), <OpOverload(op='aten.special_log1p', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573de40>, kernel=<OpOverload(op='aten.special_log1p', overload='default')>), <OpOverload(op='aten.slogdet', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e520>, kernel=<OpOverload(op='aten.slogdet', overload='default')>), <OpOverload(op='aten.special_logit', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573f9c0>, kernel=<OpOverload(op='aten.special_logit', overload='default')>), <OpOverload(op='aten.adaptive_avg_pool2d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eb4c0>, kernel=<OpOverload(op='aten.adaptive_avg_pool2d', overload='default')>), <OpOverload(op='aten.fake_quantize_per_channel_affine', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990040>, kernel=<OpOverload(op='aten.fake_quantize_per_channel_affine', overload='default')>), <OpOverload(op='aten.rnn_tanh_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a7a0>, kernel=<OpOverload(op='aten.rnn_tanh_cell', overload='default')>), <OpOverload(op='aten.atleast_1d', overload='Sequence')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394b6a0>, kernel=<OpOverload(op='aten.atleast_1d', overload='Sequence')>), <OpOverload(op='aten.max_pool1d_with_indices', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ebc40>, kernel=<OpOverload(op='aten.max_pool1d_with_indices', overload='default')>), <OpOverload(op='aten._sparse_sum', overload='dtype')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990cc0>, kernel=<OpOverload(op='aten._sparse_sum', overload='dtype')>), <OpOverload(op='aten.concatenate', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b63ec0>, kernel=<OpOverload(op='aten.concatenate', overload='default')>), <OpOverload(op='aten.true_divide', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496cc20>, kernel=<OpOverload(op='aten.true_divide', overload='Scalar')>), <OpOverload(op='aten.special_polygamma', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329902c0>, kernel=<OpOverload(op='aten.special_polygamma', overload='default')>), <OpOverload(op='aten.special_exp2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e5c0>, kernel=<OpOverload(op='aten.special_exp2', overload='default')>), <OpOverload(op='aten.gather_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329907c0>, kernel=<OpOverload(op='aten.gather_backward', overload='default')>), <OpOverload(op='aten.native_channel_shuffle', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329544a0>, kernel=<OpOverload(op='aten.native_channel_shuffle', overload='default')>), <OpOverload(op='aten.arccos', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394bc40>, kernel=<OpOverload(op='aten.arccos', overload='default')>), <OpOverload(op='aten.sym_size', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d300>, kernel=<OpOverload(op='aten.sym_size', overload='int')>), <OpOverload(op='aten.sym_stride', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573f600>, kernel=<OpOverload(op='aten.sym_stride', overload='int')>), <OpOverload(op='aten.ctc_loss', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776345bb560>, kernel=<OpOverload(op='aten.ctc_loss', overload='Tensor')>), <OpOverload(op='aten._cast_Half', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634178b80>, kernel=<OpOverload(op='aten._cast_Half', overload='default')>), <OpOverload(op='aten._cast_Float', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969da0>, kernel=<OpOverload(op='aten._cast_Float', overload='default')>), <OpOverload(op='aten._sparse_softmax', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394ba60>, kernel=<OpOverload(op='aten._sparse_softmax', overload='Dimname')>), <OpOverload(op='aten.sparse_bsr_tensor', overload='crow_col_value')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632d18b80>, kernel=<OpOverload(op='aten.sparse_bsr_tensor', overload='crow_col_value')>), <OpOverload(op='aten.trapz', overload='x')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573dc60>, kernel=<OpOverload(op='aten.trapz', overload='x')>), <OpOverload(op='aten.tensordot', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b634c0>, kernel=<OpOverload(op='aten.tensordot', overload='default')>), <OpOverload(op='aten.sparse_csc_tensor', overload='ccol_row_value_size')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6bf60>, kernel=<OpOverload(op='aten.sparse_csc_tensor', overload='ccol_row_value_size')>), <OpOverload(op='aten.nanmean', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776defd8a40>, kernel=<OpOverload(op='aten.nanmean', overload='default')>), <OpOverload(op='aten.align_as', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e67a0>, kernel=<OpOverload(op='aten.align_as', overload='default')>), <OpOverload(op='aten.__or__', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6bce0>, kernel=<OpOverload(op='aten.__or__', overload='Scalar')>), <OpOverload(op='aten.special_erf', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632d18ae0>, kernel=<OpOverload(op='aten.special_erf', overload='default')>), <OpOverload(op='aten.grid_sampler', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539f920>, kernel=<OpOverload(op='aten.grid_sampler', overload='default')>), <OpOverload(op='aten.gradient', overload='array')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632d19d00>, kernel=<OpOverload(op='aten.gradient', overload='array')>), <OpOverload(op='aten.arctan', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6ba60>, kernel=<OpOverload(op='aten.arctan', overload='default')>), <OpOverload(op='aten.chalf', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7880>, kernel=<OpOverload(op='aten.chalf', overload='default')>), <OpOverload(op='aten.det', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394a660>, kernel=<OpOverload(op='aten.det', overload='default')>), <OpOverload(op='image._jpeg_version', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776e18653a0>, kernel=<OpOverload(op='image._jpeg_version', overload='default')>), <OpOverload(op='aten.ldexp', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6ade0>, kernel=<OpOverload(op='aten.ldexp', overload='Tensor')>), <OpOverload(op='aten.log_softmax', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d684a0>, kernel=<OpOverload(op='aten.log_softmax', overload='Dimname')>), <OpOverload(op='aten._sobol_engine_draw', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6840>, kernel=<OpOverload(op='aten._sobol_engine_draw', overload='default')>), <OpOverload(op='aten.trapezoid', overload='x')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394bf60>, kernel=<OpOverload(op='aten.trapezoid', overload='x')>), <OpOverload(op='aten.linalg_solve', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d69440>, kernel=<OpOverload(op='aten.linalg_solve', overload='default')>), <OpOverload(op='aten.kthvalue', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354c360>, kernel=<OpOverload(op='aten.kthvalue', overload='dimname')>), <OpOverload(op='aten.histogramdd', overload='int_bins')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394bba0>, kernel=<OpOverload(op='aten.histogramdd', overload='int_bins')>), <OpOverload(op='c10d_functional.broadcast', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c31bc40>, kernel=<OpOverload(op='c10d_functional.broadcast', overload='default')>), <OpOverload(op='aten._pack_padded_sequence_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6660>, kernel=<OpOverload(op='aten._pack_padded_sequence_backward', overload='default')>), <OpOverload(op='aten._reshape_from_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329698a0>, kernel=<OpOverload(op='aten._reshape_from_tensor', overload='default')>), <OpOverload(op='aten.sparse_bsc_tensor', overload='ccol_row_value_size')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632d187c0>, kernel=<OpOverload(op='aten.sparse_bsc_tensor', overload='ccol_row_value_size')>), <OpOverload(op='aten.triplet_margin_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6e80>, kernel=<OpOverload(op='aten.triplet_margin_loss', overload='default')>), <OpOverload(op='quantized.conv2d_unpack_sizes', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de40680>, kernel=<OpOverload(op='quantized.conv2d_unpack_sizes', overload='default')>), <OpOverload(op='aten.multiply', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354e340>, kernel=<OpOverload(op='aten.multiply', overload='Tensor')>), <OpOverload(op='aten.linalg_eigh', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296af20>, kernel=<OpOverload(op='aten.linalg_eigh', overload='default')>), <OpOverload(op='aten.is_leaf', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b60b80>, kernel=<OpOverload(op='aten.is_leaf', overload='default')>), <OpOverload(op='aten.repeat_interleave', overload='self_Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7060>, kernel=<OpOverload(op='aten.repeat_interleave', overload='self_Tensor')>), <OpOverload(op='aten._has_compatible_shallow_copy_type', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496c9a0>, kernel=<OpOverload(op='aten._has_compatible_shallow_copy_type', overload='default')>), <OpOverload(op='torchvision._cuda_version', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c39b560>, kernel=<OpOverload(op='torchvision._cuda_version', overload='default')>), <OpOverload(op='aten.cumulative_trapezoid', overload='dx')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968d60>, kernel=<OpOverload(op='aten.cumulative_trapezoid', overload='dx')>), <OpOverload(op='_test.leaky_relu', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c3985e0>, kernel=<OpOverload(op='_test.leaky_relu', overload='default')>), <OpOverload(op='aten.cummaxmin_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573dd00>, kernel=<OpOverload(op='aten.cummaxmin_backward', overload='default')>), <OpOverload(op='aten._sparse_sum', overload='dim_dtype')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d760>, kernel=<OpOverload(op='aten._sparse_sum', overload='dim_dtype')>), <OpOverload(op='aten.sparse_csr_tensor', overload='crow_col_value')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573c0e0>, kernel=<OpOverload(op='aten.sparse_csr_tensor', overload='crow_col_value')>), <OpOverload(op='aten._batch_norm_impl_index_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632957d80>, kernel=<OpOverload(op='aten._batch_norm_impl_index_backward', overload='default')>), <OpOverload(op='aten.fliplr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539cea0>, kernel=<OpOverload(op='aten.fliplr', overload='default')>), <OpOverload(op='aten.argwhere', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b240>, kernel=<OpOverload(op='aten.argwhere', overload='default')>), <OpOverload(op='aten.rnn_relu_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6be20>, kernel=<OpOverload(op='aten.rnn_relu_cell', overload='default')>), <OpOverload(op='aten.softmax', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296bec0>, kernel=<OpOverload(op='aten.softmax', overload='int')>), <OpOverload(op='aten.row_stack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b920>, kernel=<OpOverload(op='aten.row_stack', overload='default')>), <OpOverload(op='aten.__and__', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394bd80>, kernel=<OpOverload(op='aten.__and__', overload='Tensor')>), <OpOverload(op='aten.adaptive_avg_pool1d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6b60>, kernel=<OpOverload(op='aten.adaptive_avg_pool1d', overload='default')>), <OpOverload(op='aten.is_conj', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d68b80>, kernel=<OpOverload(op='aten.is_conj', overload='default')>), <OpOverload(op='aten.argsort', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329682c0>, kernel=<OpOverload(op='aten.argsort', overload='default')>), <OpOverload(op='aten.cov', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e76a0>, kernel=<OpOverload(op='aten.cov', overload='default')>), <OpOverload(op='aten.value_selecting_reduction_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394ae80>, kernel=<OpOverload(op='aten.value_selecting_reduction_backward', overload='default')>), <OpOverload(op='aten.to_dense', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6af20>, kernel=<OpOverload(op='aten.to_dense', overload='default')>), <OpOverload(op='aten.diagflat', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539d3a0>, kernel=<OpOverload(op='aten.diagflat', overload='default')>), <OpOverload(op='aten.is_inference', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632954d60>, kernel=<OpOverload(op='aten.is_inference', overload='default')>), <OpOverload(op='aten.to_sparse_bsr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b60a40>, kernel=<OpOverload(op='aten.to_sparse_bsr', overload='default')>), <OpOverload(op='aten.linalg_pinv', overload='atol_rtol_float')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539f7e0>, kernel=<OpOverload(op='aten.linalg_pinv', overload='atol_rtol_float')>), <OpOverload(op='aten.__and__', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969940>, kernel=<OpOverload(op='aten.__and__', overload='Scalar')>), <OpOverload(op='aten.fbgemm_linear_fp16_weight', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6520>, kernel=<OpOverload(op='aten.fbgemm_linear_fp16_weight', overload='default')>), <OpOverload(op='aten.__xor__', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632993380>, kernel=<OpOverload(op='aten.__xor__', overload='Scalar')>), <OpOverload(op='aten.mode', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539c4a0>, kernel=<OpOverload(op='aten.mode', overload='dimname')>), <OpOverload(op='aten.to_dense_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539d080>, kernel=<OpOverload(op='aten.to_dense_backward', overload='default')>), <OpOverload(op='image.encode_jpegs_cuda', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d7339c0>, kernel=<OpOverload(op='image.encode_jpegs_cuda', overload='default')>), <OpOverload(op='aten.qr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969a80>, kernel=<OpOverload(op='aten.qr', overload='default')>), <OpOverload(op='aten.special_round', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354c0e0>, kernel=<OpOverload(op='aten.special_round', overload='default')>), <OpOverload(op='aten.flipud', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6fc0>, kernel=<OpOverload(op='aten.flipud', overload='default')>), <OpOverload(op='aten._pad_circular', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969760>, kernel=<OpOverload(op='aten._pad_circular', overload='default')>), <OpOverload(op='aten._cast_Double', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b4c0>, kernel=<OpOverload(op='aten._cast_Double', overload='default')>), <OpOverload(op='aten.special_log_softmax', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991a80>, kernel=<OpOverload(op='aten.special_log_softmax', overload='default')>), <OpOverload(op='aten.index_select_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b600>, kernel=<OpOverload(op='aten.index_select_backward', overload='default')>), <OpOverload(op='aten.nuclear_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a8e0>, kernel=<OpOverload(op='aten.nuclear_norm', overload='default')>), <OpOverload(op='aten.linalg_matmul', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a660>, kernel=<OpOverload(op='aten.linalg_matmul', overload='default')>), <OpOverload(op='_test.cat', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c3bda80>, kernel=<OpOverload(op='_test.cat', overload='default')>), <OpOverload(op='aten.argsort', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776339494e0>, kernel=<OpOverload(op='aten.argsort', overload='dimname')>), <OpOverload(op='aten.msort', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354c180>, kernel=<OpOverload(op='aten.msort', overload='default')>), <OpOverload(op='mkldnn._is_mkldnn_bf16_supported', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d733ce0>, kernel=<OpOverload(op='mkldnn._is_mkldnn_bf16_supported', overload='default')>), <OpOverload(op='aten.linalg_cond', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b9c0>, kernel=<OpOverload(op='aten.linalg_cond', overload='default')>), <OpOverload(op='aten.vstack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776345bb420>, kernel=<OpOverload(op='aten.vstack', overload='default')>), <OpOverload(op='aten.matrix_power', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969f80>, kernel=<OpOverload(op='aten.matrix_power', overload='default')>), <OpOverload(op='aten.align_tensors', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634178040>, kernel=<OpOverload(op='aten.align_tensors', overload='default')>), <OpOverload(op='aten.one_hot', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394b060>, kernel=<OpOverload(op='aten.one_hot', overload='default')>), <OpOverload(op='aten.greater_equal', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573e160>, kernel=<OpOverload(op='aten.greater_equal', overload='Scalar')>), <OpOverload(op='aten.special_expm1', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b100>, kernel=<OpOverload(op='aten.special_expm1', overload='default')>), <OpOverload(op='aten.conv1d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b740>, kernel=<OpOverload(op='aten.conv1d', overload='default')>), <OpOverload(op='aten.cdist', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763417b380>, kernel=<OpOverload(op='aten.cdist', overload='default')>), <OpOverload(op='aten.special_gammaln', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354e520>, kernel=<OpOverload(op='aten.special_gammaln', overload='default')>), <OpOverload(op='aten._to_cpu', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634178e00>, kernel=<OpOverload(op='aten._to_cpu', overload='default')>), <OpOverload(op='aten.to_sparse_csr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b4c0>, kernel=<OpOverload(op='aten.to_sparse_csr', overload='default')>), <OpOverload(op='aten.linalg_pinv', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539ce00>, kernel=<OpOverload(op='aten.linalg_pinv', overload='default')>), <OpOverload(op='aten.kron', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634178220>, kernel=<OpOverload(op='aten.kron', overload='default')>), <OpOverload(op='aten.group_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992f20>, kernel=<OpOverload(op='aten.group_norm', overload='default')>), <OpOverload(op='aten.fbgemm_pack_quantized_matrix', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573e0c0>, kernel=<OpOverload(op='aten.fbgemm_pack_quantized_matrix', overload='default')>), <OpOverload(op='aten.ctc_loss', overload='IntList')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329685e0>, kernel=<OpOverload(op='aten.ctc_loss', overload='IntList')>), <OpOverload(op='aten.linalg_vander', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632955a80>, kernel=<OpOverload(op='aten.linalg_vander', overload='default')>), <OpOverload(op='aten.log_softmax', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6480>, kernel=<OpOverload(op='aten.log_softmax', overload='int')>), <OpOverload(op='aten._scaled_dot_product_attention_math', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991580>, kernel=<OpOverload(op='aten._scaled_dot_product_attention_math', overload='default')>), <OpOverload(op='aten.embedding_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6d40>, kernel=<OpOverload(op='aten.embedding_backward', overload='default')>), <OpOverload(op='aten.special_gammaincc', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776345bb6a0>, kernel=<OpOverload(op='aten.special_gammaincc', overload='default')>), <OpOverload(op='image.decode_jpeg', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c338900>, kernel=<OpOverload(op='image.decode_jpeg', overload='default')>), <OpOverload(op='aten.pad', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539ccc0>, kernel=<OpOverload(op='aten.pad', overload='default')>), <OpOverload(op='aten.inverse', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d800>, kernel=<OpOverload(op='aten.inverse', overload='default')>), <OpOverload(op='aten._sparse_bsc_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992d40>, kernel=<OpOverload(op='aten._sparse_bsc_tensor_unsafe', overload='default')>), <OpOverload(op='aten.stride', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b632e0>, kernel=<OpOverload(op='aten.stride', overload='int')>), <OpOverload(op='aten._sparse_compressed_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969b20>, kernel=<OpOverload(op='aten._sparse_compressed_tensor_unsafe', overload='default')>), <OpOverload(op='aten.nested_to_padded_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329694e0>, kernel=<OpOverload(op='aten.nested_to_padded_tensor', overload='default')>), <OpOverload(op='aten.linalg_matrix_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992de0>, kernel=<OpOverload(op='aten.linalg_matrix_norm', overload='default')>), <OpOverload(op='aten._cufft_clear_plan_cache', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296aca0>, kernel=<OpOverload(op='aten._cufft_clear_plan_cache', overload='default')>), <OpOverload(op='aten._wrapped_quantized_linear_prepacked', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329922a0>, kernel=<OpOverload(op='aten._wrapped_quantized_linear_prepacked', overload='default')>), <OpOverload(op='aten.special_erfinv', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b63740>, kernel=<OpOverload(op='aten.special_erfinv', overload='default')>), <OpOverload(op='aten.fix', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632954e00>, kernel=<OpOverload(op='aten.fix', overload='default')>), <OpOverload(op='aten.size', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b420>, kernel=<OpOverload(op='aten.size', overload='int')>), <OpOverload(op='aten.quantized_rnn_tanh_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991260>, kernel=<OpOverload(op='aten.quantized_rnn_tanh_cell', overload='default')>), <OpOverload(op='aten.atleast_3d', overload='Sequence')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632d18c20>, kernel=<OpOverload(op='aten.atleast_3d', overload='Sequence')>), <OpOverload(op='aten.pinverse', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329925c0>, kernel=<OpOverload(op='aten.pinverse', overload='default')>), <OpOverload(op='aten._sparse_bsr_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329563e0>, kernel=<OpOverload(op='aten._sparse_bsr_tensor_unsafe', overload='default')>), <OpOverload(op='aten.fbgemm_linear_quantize_weight', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968f40>, kernel=<OpOverload(op='aten.fbgemm_linear_quantize_weight', overload='default')>), <OpOverload(op='aten._add_batch_dim', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296bb00>, kernel=<OpOverload(op='aten._add_batch_dim', overload='default')>), <OpOverload(op='aten._nnpack_available', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7a60>, kernel=<OpOverload(op='aten._nnpack_available', overload='default')>), <OpOverload(op='aten.fbgemm_pack_quantized_matrix', overload='KN')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632956020>, kernel=<OpOverload(op='aten.fbgemm_pack_quantized_matrix', overload='KN')>), <OpOverload(op='aten.special_logsumexp', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394bec0>, kernel=<OpOverload(op='aten.special_logsumexp', overload='default')>), <OpOverload(op='aten._cast_Char', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d580>, kernel=<OpOverload(op='aten._cast_Char', overload='default')>), <OpOverload(op='aten.is_nonzero', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d69620>, kernel=<OpOverload(op='aten.is_nonzero', overload='default')>), <OpOverload(op='mkldnn._is_mkldnn_fp16_supported', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d733c40>, kernel=<OpOverload(op='mkldnn._is_mkldnn_fp16_supported', overload='default')>), <OpOverload(op='aten.nanquantile', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992e80>, kernel=<OpOverload(op='aten.nanquantile', overload='default')>), <OpOverload(op='aten.is_neg', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6afc0>, kernel=<OpOverload(op='aten.is_neg', overload='default')>), <OpOverload(op='aten.layer_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6bba0>, kernel=<OpOverload(op='aten.layer_norm', overload='default')>), <OpOverload(op='aten._rowwise_prune', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539fb00>, kernel=<OpOverload(op='aten._rowwise_prune', overload='default')>), <OpOverload(op='aten.column_stack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7d80>, kernel=<OpOverload(op='aten.column_stack', overload='default')>), <OpOverload(op='aten.linalg_pinv', overload='rcond_tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6700>, kernel=<OpOverload(op='aten.linalg_pinv', overload='rcond_tensor')>), <OpOverload(op='aten.size', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296ac00>, kernel=<OpOverload(op='aten.size', overload='Dimname')>), <OpOverload(op='aten.sort', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969620>, kernel=<OpOverload(op='aten.sort', overload='dimname')>), <OpOverload(op='aten.result_type', overload='Scalar_Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394a5c0>, kernel=<OpOverload(op='aten.result_type', overload='Scalar_Tensor')>), <OpOverload(op='aten.fake_quantize_per_tensor_affine', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9760>, kernel=<OpOverload(op='aten.fake_quantize_per_tensor_affine', overload='default')>), <OpOverload(op='aten.less', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329904a0>, kernel=<OpOverload(op='aten.less', overload='Tensor')>), <OpOverload(op='aten.diff', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990ea0>, kernel=<OpOverload(op='aten.diff', overload='default')>), <OpOverload(op='aten.special_expit', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634179b20>, kernel=<OpOverload(op='aten.special_expit', overload='default')>), <OpOverload(op='aten.sparse_coo_tensor', overload='indices')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b380>, kernel=<OpOverload(op='aten.sparse_coo_tensor', overload='indices')>), <OpOverload(op='aten._thnn_differentiable_lstm_cell_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329693a0>, kernel=<OpOverload(op='aten._thnn_differentiable_lstm_cell_backward', overload='default')>), <OpOverload(op='aten.cosine_embedding_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991940>, kernel=<OpOverload(op='aten.cosine_embedding_loss', overload='default')>), <OpOverload(op='aten.arccosh', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968e00>, kernel=<OpOverload(op='aten.arccosh', overload='default')>), <OpOverload(op='aten._embedding_bag_sparse_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a340>, kernel=<OpOverload(op='aten._embedding_bag_sparse_backward', overload='default')>), <OpOverload(op='image.write_file', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d733ec0>, kernel=<OpOverload(op='image.write_file', overload='default')>), <OpOverload(op='aten.special_softmax', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a2a0>, kernel=<OpOverload(op='aten.special_softmax', overload='default')>), <OpOverload(op='mkldnn._is_mkldnn_acl_supported', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c39bec0>, kernel=<OpOverload(op='mkldnn._is_mkldnn_acl_supported', overload='default')>), <OpOverload(op='aten.result_type', overload='Scalar_Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968b80>, kernel=<OpOverload(op='aten.result_type', overload='Scalar_Scalar')>), <OpOverload(op='aten.max_pool1d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296ba60>, kernel=<OpOverload(op='aten.max_pool1d', overload='default')>), <OpOverload(op='aten.dstack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e020>, kernel=<OpOverload(op='aten.dstack', overload='default')>), <OpOverload(op='aten.is_floating_point', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992700>, kernel=<OpOverload(op='aten.is_floating_point', overload='default')>), <OpOverload(op='aten.subtract', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968400>, kernel=<OpOverload(op='aten.subtract', overload='Tensor')>), <OpOverload(op='aten.greater', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b920>, kernel=<OpOverload(op='aten.greater', overload='Tensor')>), <OpOverload(op='aten.isclose', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b2e0>, kernel=<OpOverload(op='aten.isclose', overload='default')>), <OpOverload(op='aten._sparse_log_softmax', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573f880>, kernel=<OpOverload(op='aten._sparse_log_softmax', overload='Dimname')>), <OpOverload(op='aten._test_autograd_multiple_dispatch', overload='ntonly')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6ac0>, kernel=<OpOverload(op='aten._test_autograd_multiple_dispatch', overload='ntonly')>), <OpOverload(op='c10d_functional.all_reduce_coalesced', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c2ab6a0>, kernel=<OpOverload(op='c10d_functional.all_reduce_coalesced', overload='default')>), <OpOverload(op='aten.arctan2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992ca0>, kernel=<OpOverload(op='aten.arctan2', overload='default')>), <OpOverload(op='aten._test_ambiguous_defaults', overload='b')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573db20>, kernel=<OpOverload(op='aten._test_ambiguous_defaults', overload='b')>), <OpOverload(op='aten.special_digamma', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573c180>, kernel=<OpOverload(op='aten.special_digamma', overload='default')>), <OpOverload(op='aten.stft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968040>, kernel=<OpOverload(op='aten.stft', overload='default')>), <OpOverload(op='aten.result_type', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a700>, kernel=<OpOverload(op='aten.result_type', overload='Tensor')>), <OpOverload(op='aten.tile', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632993240>, kernel=<OpOverload(op='aten.tile', overload='default')>), <OpOverload(op='aten.gradient', overload='tensorarray')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329905e0>, kernel=<OpOverload(op='aten.gradient', overload='tensorarray')>), <OpOverload(op='aten._validate_sparse_bsr_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d4e0>, kernel=<OpOverload(op='aten._validate_sparse_bsr_tensor_args', overload='default')>), <OpOverload(op='aten.divide', overload='Tensor_mode')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b636a0>, kernel=<OpOverload(op='aten.divide', overload='Tensor_mode')>), <OpOverload(op='aten.linalg_eigvals', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d620>, kernel=<OpOverload(op='aten.linalg_eigvals', overload='default')>), <OpOverload(op='aten._debug_has_internal_overlap', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633949440>, kernel=<OpOverload(op='aten._debug_has_internal_overlap', overload='default')>), <OpOverload(op='c10d_functional.all_gather_into_tensor_coalesced', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c3382c0>, kernel=<OpOverload(op='c10d_functional.all_gather_into_tensor_coalesced', overload='default')>), <OpOverload(op='aten.max', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b63420>, kernel=<OpOverload(op='aten.max', overload='names_dim')>), <OpOverload(op='aten.__xor__', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632954f40>, kernel=<OpOverload(op='aten.__xor__', overload='Tensor')>), <OpOverload(op='aten.adaptive_avg_pool3d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d080>, kernel=<OpOverload(op='aten.adaptive_avg_pool3d', overload='default')>), <OpOverload(op='aten.max_pool2d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7ba0>, kernel=<OpOverload(op='aten.max_pool2d', overload='default')>), <OpOverload(op='aten.nanquantile', overload='scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990680>, kernel=<OpOverload(op='aten.nanquantile', overload='scalar')>), <OpOverload(op='c10d_functional.all_to_all_single', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c3385e0>, kernel=<OpOverload(op='c10d_functional.all_to_all_single', overload='default')>), <OpOverload(op='aten._validate_sparse_bsc_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329928e0>, kernel=<OpOverload(op='aten._validate_sparse_bsc_tensor_args', overload='default')>), <OpOverload(op='aten.conv3d', overload='padding')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ea340>, kernel=<OpOverload(op='aten.conv3d', overload='padding')>), <OpOverload(op='aten.trapezoid', overload='dx')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992b60>, kernel=<OpOverload(op='aten.trapezoid', overload='dx')>), <OpOverload(op='aten.divide', overload='Scalar_mode')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296ade0>, kernel=<OpOverload(op='aten.divide', overload='Scalar_mode')>), <OpOverload(op='image.encode_jpeg', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de404a0>, kernel=<OpOverload(op='image.encode_jpeg', overload='default')>), <OpOverload(op='aten.fbgemm_linear_int8_weight', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eb060>, kernel=<OpOverload(op='aten.fbgemm_linear_int8_weight', overload='default')>), <OpOverload(op='aten.scatter', overload='dimname_src')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b600>, kernel=<OpOverload(op='aten.scatter', overload='dimname_src')>), <OpOverload(op='aten.linalg_lu_factor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632954860>, kernel=<OpOverload(op='aten.linalg_lu_factor', overload='default')>), <OpOverload(op='aten.to_sparse_csc', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632956520>, kernel=<OpOverload(op='aten.to_sparse_csc', overload='default')>), <OpOverload(op='aten.gradient', overload='scalarint')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633948180>, kernel=<OpOverload(op='aten.gradient', overload='scalarint')>), <OpOverload(op='aten.concatenate', overload='names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539cf40>, kernel=<OpOverload(op='aten.concatenate', overload='names')>), <OpOverload(op='aten._sparse_mm', overload='reduce')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539d260>, kernel=<OpOverload(op='aten._sparse_mm', overload='reduce')>), <OpOverload(op='aten.flatten_dense_tensors', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329689a0>, kernel=<OpOverload(op='aten.flatten_dense_tensors', overload='default')>), <OpOverload(op='profiler._record_function_enter_new', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c31af20>, kernel=<OpOverload(op='profiler._record_function_enter_new', overload='default')>), <OpOverload(op='aten.linalg_inv', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632957f60>, kernel=<OpOverload(op='aten.linalg_inv', overload='default')>), <OpOverload(op='aten.sparse_coo_tensor', overload='indices_size')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329559e0>, kernel=<OpOverload(op='aten.sparse_coo_tensor', overload='indices_size')>), <OpOverload(op='aten.diag', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991120>, kernel=<OpOverload(op='aten.diag', overload='default')>), <OpOverload(op='aten._fused_rms_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991da0>, kernel=<OpOverload(op='aten._fused_rms_norm', overload='default')>), <OpOverload(op='aten.float_power', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329932e0>, kernel=<OpOverload(op='aten.float_power', overload='Scalar')>), <OpOverload(op='aten.float_power', overload='Tensor_Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634178d60>, kernel=<OpOverload(op='aten.float_power', overload='Tensor_Scalar')>), <OpOverload(op='aten.float_power', overload='Tensor_Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991f80>, kernel=<OpOverload(op='aten.float_power', overload='Tensor_Tensor')>), <OpOverload(op='aten.square', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496ca40>, kernel=<OpOverload(op='aten.square', overload='default')>), <OpOverload(op='aten.true_divide', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573f420>, kernel=<OpOverload(op='aten.true_divide', overload='Tensor')>), <OpOverload(op='aten.clip', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7e20>, kernel=<OpOverload(op='aten.clip', overload='default')>), <OpOverload(op='aten.adaptive_max_pool1d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329579c0>, kernel=<OpOverload(op='aten.adaptive_max_pool1d', overload='default')>), <OpOverload(op='aten._test_string_default', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539d300>, kernel=<OpOverload(op='aten._test_string_default', overload='default')>), <OpOverload(op='aten.einsum', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776345bb4c0>, kernel=<OpOverload(op='aten.einsum', overload='default')>), <OpOverload(op='aten._validate_sparse_coo_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763417ac00>, kernel=<OpOverload(op='aten._validate_sparse_coo_tensor_args', overload='default')>), <OpOverload(op='c10d_functional.wait_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d733740>, kernel=<OpOverload(op='c10d_functional.wait_tensor', overload='default')>), <OpOverload(op='aten.combinations', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9e40>, kernel=<OpOverload(op='aten.combinations', overload='default')>), <OpOverload(op='aten.quantized_gru_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354e3e0>, kernel=<OpOverload(op='aten.quantized_gru_cell', overload='default')>), <OpOverload(op='aten.conv3d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9f80>, kernel=<OpOverload(op='aten.conv3d', overload='default')>), <OpOverload(op='aten.logcumsumexp', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d68cc0>, kernel=<OpOverload(op='aten.logcumsumexp', overload='dimname')>), <OpOverload(op='aten.cosine_similarity', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7100>, kernel=<OpOverload(op='aten.cosine_similarity', overload='default')>), <OpOverload(op='aten._validate_sparse_csr_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394b560>, kernel=<OpOverload(op='aten._validate_sparse_csr_tensor_args', overload='default')>), <OpOverload(op='aten.less_equal', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776341780e0>, kernel=<OpOverload(op='aten.less_equal', overload='Scalar')>), <OpOverload(op='image.decode_png', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d733d80>, kernel=<OpOverload(op='image.decode_png', overload='default')>), <OpOverload(op='aten.log_sigmoid', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776345bb380>, kernel=<OpOverload(op='aten.log_sigmoid', overload='default')>), <OpOverload(op='aten.chain_matmul', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632957b00>, kernel=<OpOverload(op='aten.chain_matmul', overload='default')>), <OpOverload(op='aten._cufft_set_plan_cache_max_size', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991c60>, kernel=<OpOverload(op='aten._cufft_set_plan_cache_max_size', overload='default')>), <OpOverload(op='aten.argsort', overload='stable')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990e00>, kernel=<OpOverload(op='aten.argsort', overload='stable')>), <OpOverload(op='aten._thnn_differentiable_gru_cell_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968a40>, kernel=<OpOverload(op='aten._thnn_differentiable_gru_cell_backward', overload='default')>), <OpOverload(op='inductor._alloc_from_pool', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c31be20>, kernel=<OpOverload(op='inductor._alloc_from_pool', overload='default')>), <OpOverload(op='aten.embedding_bag', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990360>, kernel=<OpOverload(op='aten.embedding_bag', overload='default')>), <OpOverload(op='aten.max_pool3d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991080>, kernel=<OpOverload(op='aten.max_pool3d', overload='default')>), <OpOverload(op='aten.nuclear_norm', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991440>, kernel=<OpOverload(op='aten.nuclear_norm', overload='dim')>), <OpOverload(op='aten.special_sinc', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b060>, kernel=<OpOverload(op='aten.special_sinc', overload='default')>), <OpOverload(op='aten.can_cast', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539c400>, kernel=<OpOverload(op='aten.can_cast', overload='default')>), <OpOverload(op='image.read_file', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de43e20>, kernel=<OpOverload(op='image.read_file', overload='default')>), <OpOverload(op='aten._sparse_log_softmax', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776345bb240>, kernel=<OpOverload(op='aten._sparse_log_softmax', overload='int')>), <OpOverload(op='aten.nll_loss_nd', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b2e0>, kernel=<OpOverload(op='aten.nll_loss_nd', overload='default')>), <OpOverload(op='aten.slow_conv3d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9d00>, kernel=<OpOverload(op='aten.slow_conv3d', overload='default')>), <OpOverload(op='aten.divide', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573fc40>, kernel=<OpOverload(op='aten.divide', overload='Scalar')>), <OpOverload(op='aten.smm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b7e0>, kernel=<OpOverload(op='aten.smm', overload='default')>), <OpOverload(op='aten.trace_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eb7e0>, kernel=<OpOverload(op='aten.trace_backward', overload='default')>), <OpOverload(op='aten.gradient', overload='scalarrayarray')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d6c0>, kernel=<OpOverload(op='aten.gradient', overload='scalarrayarray')>), <OpOverload(op='c10d_functional.all_gather_into_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c31bb00>, kernel=<OpOverload(op='c10d_functional.all_gather_into_tensor', overload='default')>), <OpOverload(op='aten.linalg_ldl_factor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329565c0>, kernel=<OpOverload(op='aten.linalg_ldl_factor', overload='default')>), <OpOverload(op='aten.arcsin', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573e980>, kernel=<OpOverload(op='aten.arcsin', overload='default')>), <OpOverload(op='aten._cast_Short', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632d18d60>, kernel=<OpOverload(op='aten._cast_Short', overload='default')>), <OpOverload(op='aten._sparse_sum', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eb600>, kernel=<OpOverload(op='aten._sparse_sum', overload='default')>), <OpOverload(op='aten.fake_quantize_per_channel_affine_cachemask_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ea480>, kernel=<OpOverload(op='aten.fake_quantize_per_channel_affine_cachemask_backward', overload='default')>), <OpOverload(op='aten.sort', overload='dimname_stable')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e660>, kernel=<OpOverload(op='aten.sort', overload='dimname_stable')>), <OpOverload(op='aten.embedding_sparse_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573e480>, kernel=<OpOverload(op='aten.embedding_sparse_backward', overload='default')>), <OpOverload(op='aten._sparse_csr_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573f060>, kernel=<OpOverload(op='aten._sparse_csr_tensor_unsafe', overload='default')>), <OpOverload(op='aten.special_multigammaln', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573ed40>, kernel=<OpOverload(op='aten.special_multigammaln', overload='default')>), <OpOverload(op='aten.special_erfc', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eb6a0>, kernel=<OpOverload(op='aten.special_erfc', overload='default')>), <OpOverload(op='aten.multilabel_margin_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e840>, kernel=<OpOverload(op='aten.multilabel_margin_loss', overload='default')>), <OpOverload(op='aten.median', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632993060>, kernel=<OpOverload(op='aten.median', overload='names_dim')>), <OpOverload(op='aten.instance_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ea0c0>, kernel=<OpOverload(op='aten.instance_norm', overload='default')>), <OpOverload(op='aten.linalg_matrix_rank', overload='tol_tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b604a0>, kernel=<OpOverload(op='aten.linalg_matrix_rank', overload='tol_tensor')>), <OpOverload(op='aten.sspaddmm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ebe20>, kernel=<OpOverload(op='aten.sspaddmm', overload='default')>), <OpOverload(op='aten.orgqr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e160>, kernel=<OpOverload(op='aten.orgqr', overload='default')>), <OpOverload(op='aten.take_along_dim', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573dda0>, kernel=<OpOverload(op='aten.take_along_dim', overload='default')>), <OpOverload(op='aten._cufft_get_plan_cache_max_size', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496dda0>, kernel=<OpOverload(op='aten._cufft_get_plan_cache_max_size', overload='default')>), <OpOverload(op='aten.not_equal', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e94e0>, kernel=<OpOverload(op='aten.not_equal', overload='Scalar')>), <OpOverload(op='aten.kl_div', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992c00>, kernel=<OpOverload(op='aten.kl_div', overload='default')>), <OpOverload(op='aten._weight_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b63240>, kernel=<OpOverload(op='aten._weight_norm', overload='default')>), <OpOverload(op='aten.special_xlogy', overload='self_scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9b20>, kernel=<OpOverload(op='aten.special_xlogy', overload='self_scalar')>), <OpOverload(op='aten.to_mkldnn_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496d6c0>, kernel=<OpOverload(op='aten.to_mkldnn_backward', overload='default')>), <OpOverload(op='aten.scaled_dot_product_attention', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e2a0>, kernel=<OpOverload(op='aten.scaled_dot_product_attention', overload='default')>), <OpOverload(op='aten._shape_as_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992840>, kernel=<OpOverload(op='aten._shape_as_tensor', overload='default')>), <OpOverload(op='prepacked.unpack_prepacked_sizes_conv2d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c3380e0>, kernel=<OpOverload(op='prepacked.unpack_prepacked_sizes_conv2d', overload='default')>), <OpOverload(op='aten.cross_entropy_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d3a0>, kernel=<OpOverload(op='aten.cross_entropy_loss', overload='default')>), <OpOverload(op='aten.__or__', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496d300>, kernel=<OpOverload(op='aten.__or__', overload='Tensor')>), <OpOverload(op='aten.histogramdd', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992980>, kernel=<OpOverload(op='aten.histogramdd', overload='default')>), <OpOverload(op='aten.linalg_matrix_norm', overload='str_ord')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b60680>, kernel=<OpOverload(op='aten.linalg_matrix_norm', overload='str_ord')>), <OpOverload(op='profiler._record_function_enter', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c2ab560>, kernel=<OpOverload(op='profiler._record_function_enter', overload='default')>), <OpOverload(op='aten.quantized_rnn_relu_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573f920>, kernel=<OpOverload(op='aten.quantized_rnn_relu_cell', overload='default')>), <OpOverload(op='aten.trapz', overload='dx')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ebec0>, kernel=<OpOverload(op='aten.trapz', overload='dx')>), <OpOverload(op='aten.sum_to_size', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e200>, kernel=<OpOverload(op='aten.sum_to_size', overload='default')>), <OpOverload(op='aten.matrix_exp', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573c220>, kernel=<OpOverload(op='aten.matrix_exp', overload='default')>), <OpOverload(op='aten.infinitely_differentiable_gelu_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329909a0>, kernel=<OpOverload(op='aten.infinitely_differentiable_gelu_backward', overload='default')>), <OpOverload(op='aten.histogramdd', overload='TensorList_bins')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ebf60>, kernel=<OpOverload(op='aten.histogramdd', overload='TensorList_bins')>), <OpOverload(op='aten.quantized_lstm_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992fc0>, kernel=<OpOverload(op='aten.quantized_lstm_cell', overload='default')>), <OpOverload(op='aten.affine_grid_generator_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573f740>, kernel=<OpOverload(op='aten.affine_grid_generator_backward', overload='default')>), <OpOverload(op='aten._sparse_csc_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b63560>, kernel=<OpOverload(op='aten._sparse_csc_tensor_unsafe', overload='default')>), <OpOverload(op='aten._thnn_fused_lstm_cell_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b60860>, kernel=<OpOverload(op='aten._thnn_fused_lstm_cell_backward', overload='default')>), <OpOverload(op='aten.norm_except_dim', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990900>, kernel=<OpOverload(op='aten.norm_except_dim', overload='default')>), <OpOverload(op='aten.gru_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573e8e0>, kernel=<OpOverload(op='aten.gru_cell', overload='default')>), <OpOverload(op='c10d_functional.reduce_scatter_tensor_coalesced', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c3984a0>, kernel=<OpOverload(op='c10d_functional.reduce_scatter_tensor_coalesced', overload='default')>), <OpOverload(op='aten.poisson_nll_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573ec00>, kernel=<OpOverload(op='aten.poisson_nll_loss', overload='default')>), <OpOverload(op='aten.lu_solve', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496c5e0>, kernel=<OpOverload(op='aten.lu_solve', overload='default')>), <OpOverload(op='aten._convolution_mode', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ebb00>, kernel=<OpOverload(op='aten._convolution_mode', overload='default')>), <OpOverload(op='aten._gather_sparse_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990220>, kernel=<OpOverload(op='aten._gather_sparse_backward', overload='default')>), <OpOverload(op='aten.negative', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b63f60>, kernel=<OpOverload(op='aten.negative', overload='default')>), <OpOverload(op='aten.conv_transpose1d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573ef20>, kernel=<OpOverload(op='aten.conv_transpose1d', overload='default')>), <OpOverload(op='aten.promote_types', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e8e0>, kernel=<OpOverload(op='aten.promote_types', overload='default')>), <OpOverload(op='aten.quantile', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777635b609a0>, kernel=<OpOverload(op='aten.quantile', overload='default')>), <OpOverload(op='aten.corrcoef', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573eb60>, kernel=<OpOverload(op='aten.corrcoef', overload='default')>), <OpOverload(op='aten.fft_hfftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968540>, kernel=<OpOverload(op='aten.fft_hfftn', overload='default')>), <OpOverload(op='aten.fft_hfftn', overload='out')>: <function hfftn at 0x7776e19d99e0>, <OpOverload(op='aten.fft_ifft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329691c0>, kernel=<OpOverload(op='aten.fft_ifft2', overload='default')>), <OpOverload(op='aten.fft_ifft2', overload='out')>: <function ifft2 at 0x7776e19d96c0>, <OpOverload(op='aten.fft_rfft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968680>, kernel=<OpOverload(op='aten.fft_rfft2', overload='default')>), <OpOverload(op='aten.fft_rfft2', overload='out')>: <function rfft2 at 0x7776e19d8860>, <OpOverload(op='aten.fft_irfft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329918a0>, kernel=<OpOverload(op='aten.fft_irfft2', overload='default')>), <OpOverload(op='aten.fft_irfft2', overload='out')>: <function irfft2 at 0x7776e19d9da0>, <OpOverload(op='aten.fft_hfft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992200>, kernel=<OpOverload(op='aten.fft_hfft2', overload='default')>), <OpOverload(op='aten.fft_hfft2', overload='out')>: <function hfft2 at 0x7776e19da020>, <OpOverload(op='aten.fft_ihfft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329914e0>, kernel=<OpOverload(op='aten.fft_ihfft2', overload='default')>), <OpOverload(op='aten.fft_ihfft2', overload='out')>: <function ihfft2 at 0x7776e19da2a0>, <OpOverload(op='aten.fft_fftshift', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990720>, kernel=<OpOverload(op='aten.fft_fftshift', overload='default')>), <OpOverload(op='aten.fft_ifftshift', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329911c0>, kernel=<OpOverload(op='aten.fft_ifftshift', overload='default')>), <OpOverload(op='aten.linalg_cross', overload='default')>: <function cross at 0x7776e19dab60>, <OpOverload(op='aten.linalg_cross', overload='out')>: <function cross at 0x7776e19dab60>, <OpOverload(op='aten.linalg_vector_norm', overload='default')>: <function vector_norm at 0x7776e19daf20>, <OpOverload(op='aten.linalg_vector_norm', overload='out')>: <function vector_norm at 0x7776e19daf20>, <OpOverload(op='aten.alpha_dropout', overload='default')>: <function alpha_dropout at 0x7776e19dbec0>, <OpOverload(op='aten.celu', overload='default')>: <function celu at 0x7776e1a04180>, <OpOverload(op='aten.celu', overload='out')>: <function celu at 0x7776e1a04180>, <OpOverload(op='aten.elu', overload='default')>: <function elu at 0x7776e1a04860>, <OpOverload(op='aten.elu', overload='out')>: <function elu at 0x7776e1a04860>, <OpOverload(op='aten.relu', overload='default')>: <function relu at 0x7776e1a04cc0>, <OpOverload(op='aten.relu', overload='out')>: <function relu at 0x7776e1a04cc0>, <OpOverload(op='aten.channel_shuffle', overload='default')>: <function channel_shuffle at 0x7776e1a05300>, <OpOverload(op='aten.channel_shuffle', overload='out')>: <function channel_shuffle at 0x7776e1a05300>, <OpOverload(op='aten.leaky_relu', overload='default')>: <function leaky_relu at 0x7776e1a054e0>, <OpOverload(op='aten.leaky_relu', overload='out')>: <function leaky_relu at 0x7776e1a054e0>, <OpOverload(op='aten.mish', overload='default')>: <function mish at 0x7776e1a05260>, <OpOverload(op='aten.mish', overload='out')>: <function mish at 0x7776e1a05260>, <OpOverload(op='aten.hardshrink', overload='default')>: <function hardshrink at 0x7776e1a06660>, <OpOverload(op='aten.selu', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354c040>, kernel=<OpOverload(op='aten.selu', overload='default')>), <OpOverload(op='aten.softshrink', overload='default')>: <function softshrink at 0x7776e1a06980>, <OpOverload(op='aten.softplus', overload='default')>: <function softplus at 0x7776e1a05f80>, <OpOverload(op='aten.softplus', overload='out')>: <function softplus at 0x7776e1a05f80>, <OpOverload(op='aten.hardshrink', overload='out')>: <function hardshrink at 0x7776e1a06660>, <OpOverload(op='aten.softshrink', overload='out')>: <function softshrink at 0x7776e1a06980>, <OpOverload(op='aten.margin_ranking_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776345bb600>, kernel=<OpOverload(op='aten.margin_ranking_loss', overload='default')>), <OpOverload(op='aten.hinge_embedding_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d693a0>, kernel=<OpOverload(op='aten.hinge_embedding_loss', overload='default')>), <OpOverload(op='aten.nll_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634178cc0>, kernel=<OpOverload(op='aten.nll_loss', overload='default')>), <OpOverload(op='aten.nll_loss', overload='out')>: <function nll_loss at 0x7776e1a079c0>, <OpOverload(op='aten.huber_loss', overload='default')>: <function huber_loss at 0x7776e1a07d80>, <OpOverload(op='aten.huber_loss', overload='out')>: <function huber_loss at 0x7776e1a07d80>, <OpOverload(op='aten.threshold', overload='default')>: <function threshold at 0x7776e1a200e0>, <OpOverload(op='aten.threshold', overload='out')>: <function threshold at 0x7776e1a200e0>, <OpOverload(op='aten.special_bessel_j0', overload='default')>: <function bessel_j0 at 0x7776e1a220c0>, <OpOverload(op='aten.hardtanh', overload='default')>: <function hardtanh at 0x7776e1a07240>, <OpOverload(op='aten.hardtanh', overload='out')>: <function hardtanh at 0x7776e1a07240>, <OpOverload(op='aten.special_xlog1py', overload='out')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.gelu', overload='default')>: <function gelu at 0x7776e1a20900>, <OpOverload(op='aten.gelu', overload='out')>: <function gelu at 0x7776e1a20900>, <OpOverload(op='aten.special_xlog1py', overload='self_scalar')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.prelu', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b560>, kernel=<OpOverload(op='aten.prelu', overload='default')>), <OpOverload(op='aten.glu', overload='default')>: <function glu at 0x7776e1a21440>, <OpOverload(op='aten.glu', overload='out')>: <function glu at 0x7776e1a21440>, <OpOverload(op='aten.pairwise_distance', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632956660>, kernel=<OpOverload(op='aten.pairwise_distance', overload='default')>), <OpOverload(op='aten.pdist', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969080>, kernel=<OpOverload(op='aten.pdist', overload='default')>), <OpOverload(op='aten.pixel_shuffle', overload='default')>: <function pixel_shuffle at 0x7776e1a21d00>, <OpOverload(op='aten.pixel_shuffle', overload='out')>: <function pixel_shuffle at 0x7776e1a21d00>, <OpOverload(op='aten.pixel_unshuffle', overload='default')>: <function pixel_unshuffle at 0x7776e1a21f80>, <OpOverload(op='aten.pixel_unshuffle', overload='out')>: <function pixel_unshuffle at 0x7776e1a21f80>, <OpOverload(op='aten.celu_', overload='default')>: <function celu at 0x7776e1a20a40>, <OpOverload(op='aten.elu_', overload='default')>: <function elu at 0x7776e1a21da0>, <OpOverload(op='aten.mish_', overload='default')>: <function mish at 0x7776e1a22020>, <OpOverload(op='aten.selu_', overload='default')>: <function selu at 0x7776e1a22160>, <OpOverload(op='aten.threshold_', overload='default')>: <function threshold at 0x7776e1a222a0>, <OpOverload(op='aten.special_bessel_j0', overload='out')>: <function bessel_j0 at 0x7776e1a220c0>, <OpOverload(op='aten.special_bessel_j1', overload='default')>: <function bessel_j1 at 0x7776e1a207c0>, <OpOverload(op='aten.special_bessel_j1', overload='out')>: <function bessel_j1 at 0x7776e1a207c0>, <OpOverload(op='aten.special_entr', overload='default')>: <function entr at 0x7776e1a22d40>, <OpOverload(op='aten.special_entr', overload='out')>: <function entr at 0x7776e1a22d40>, <OpOverload(op='aten.special_erfcx', overload='default')>: <function erfcx at 0x7776e1a23100>, <OpOverload(op='aten.special_erfcx', overload='out')>: <function erfcx at 0x7776e1a23100>, <OpOverload(op='aten.special_i0e', overload='default')>: <function i0e at 0x7776e1a236a0>, <OpOverload(op='aten.special_i0e', overload='out')>: <function i0e at 0x7776e1a236a0>, <OpOverload(op='aten.special_i1', overload='default')>: <function i1 at 0x7776e1a23ba0>, <OpOverload(op='aten.special_i1', overload='out')>: <function i1 at 0x7776e1a23ba0>, <OpOverload(op='aten.special_i1e', overload='default')>: <function i1e at 0x7776e1a400e0>, <OpOverload(op='aten.special_i1e', overload='out')>: <function i1e at 0x7776e1a400e0>, <OpOverload(op='aten.special_log_ndtr', overload='default')>: <function log_ndtr at 0x7776e1a404a0>, <OpOverload(op='aten.special_log_ndtr', overload='out')>: <function log_ndtr at 0x7776e1a404a0>, <OpOverload(op='aten.special_xlog1py', overload='other_scalar_out')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.logit', overload='default')>: <function logit at 0x7776e1a40860>, <OpOverload(op='aten.logit', overload='out')>: <function logit at 0x7776e1a40860>, <OpOverload(op='aten.special_xlog1py', overload='default')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.special_xlog1py', overload='other_scalar')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.special_xlog1py', overload='self_scalar_out')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.mvlgamma', overload='default')>: <function multigammaln at 0x7776e1a21ee0>, <OpOverload(op='aten.mvlgamma', overload='out')>: <function multigammaln at 0x7776e1a21ee0>, <OpOverload(op='aten.special_ndtr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990b80>, kernel=<OpOverload(op='aten.special_ndtr', overload='default')>), <OpOverload(op='aten.special_ndtr', overload='out')>: <function ndtr at 0x7776e1a40d60>, <OpOverload(op='aten.special_ndtri', overload='default')>: <function ndtri at 0x7776e1a41120>, <OpOverload(op='aten.special_ndtri', overload='out')>: <function ndtri at 0x7776e1a41120>, <OpOverload(op='aten.special_spherical_bessel_j0', overload='default')>: <function spherical_bessel_j0 at 0x7776e1a41760>, <OpOverload(op='aten.special_spherical_bessel_j0', overload='out')>: <function spherical_bessel_j0 at 0x7776e1a41760>, <OpOverload(op='aten.special_zeta', overload='default')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='other_scalar')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='self_scalar')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='out')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='self_scalar_out')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='other_scalar_out')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.clamp_max', overload='default')>: <function clamp_max at 0x7776e1ad1120>, <OpOverload(op='aten.clamp_max', overload='Tensor')>: <function clamp_max at 0x7776e1ad1120>, <OpOverload(op='aten.clamp_max', overload='out')>: <function clamp_max at 0x7776e1ad1120>, <OpOverload(op='aten.clamp_max', overload='Tensor_out')>: <function clamp_max at 0x7776e1ad1120>, <OpOverload(op='aten.all', overload='default')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dim')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dims')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dims_out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='all_out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992660>, kernel=<OpOverload(op='aten.all', overload='dimname')>), <OpOverload(op='aten.all', overload='dimname_out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.any', overload='default')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dim')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dims')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dims_out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='all_out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632992ac0>, kernel=<OpOverload(op='aten.any', overload='dimname')>), <OpOverload(op='aten.any', overload='dimname_out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.std_mean', overload='correction_out')>: <function std_mean at 0x7776e1c656c0>, <OpOverload(op='aten.std_mean', overload='correction_names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632d19300>, kernel=<OpOverload(op='aten.std_mean', overload='correction_names')>), <OpOverload(op='aten.std_mean', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632d18cc0>, kernel=<OpOverload(op='aten.std_mean', overload='names_dim')>), <OpOverload(op='aten.std_mean', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e7c40>, kernel=<OpOverload(op='aten.std_mean', overload='default')>), <OpOverload(op='aten.std', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394ad40>, kernel=<OpOverload(op='aten.std', overload='default')>), <OpOverload(op='aten.std', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394b4c0>, kernel=<OpOverload(op='aten.std', overload='dim')>), <OpOverload(op='aten.std', overload='correction')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.std', overload='correction_names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6b880>, kernel=<OpOverload(op='aten.std', overload='correction_names')>), <OpOverload(op='aten.std', overload='names_out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.std', overload='out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.mean', overload='default')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.std', overload='correction_names_out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.mean', overload='dim')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.mean', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763394be20>, kernel=<OpOverload(op='aten.mean', overload='names_dim')>), <OpOverload(op='aten.mean', overload='dtype_out')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.mean', overload='out')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.mean', overload='names_out')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.native_layer_norm', overload='default')>: <function native_layer_norm at 0x7776e1af5e40>, <OpOverload(op='aten.native_layer_norm', overload='out')>: <function native_layer_norm at 0x7776e1af5e40>, <OpOverload(op='aten.stft', overload='center')>: <function stft at 0x7776e1af53a0>, <OpOverload(op='aten.broadcast_tensors', overload='default')>: <function broadcast_tensors at 0x7776e1af4220>, <OpOverload(op='aten.var_mean', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6c00>, kernel=<OpOverload(op='aten.var_mean', overload='default')>), <OpOverload(op='aten.var_mean', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354e480>, kernel=<OpOverload(op='aten.var_mean', overload='dim')>), <OpOverload(op='aten.var_mean', overload='correction')>: <function var_mean at 0x7776e1ad2160>, <OpOverload(op='aten.var_mean', overload='correction_names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354e2a0>, kernel=<OpOverload(op='aten.var_mean', overload='correction_names')>), <OpOverload(op='aten.var_mean', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763354df80>, kernel=<OpOverload(op='aten.var_mean', overload='names_dim')>), <OpOverload(op='aten.var_mean', overload='correction_out')>: <function var_mean at 0x7776e1ad2160>, <OpOverload(op='aten.addr', overload='default')>: <function addr at 0x7776e1ad3b00>, <OpOverload(op='aten.addr', overload='out')>: <function addr at 0x7776e1ad3b00>, <OpOverload(op='aten.index_select', overload='dimname_out')>: <function index_select at 0x7776e1b10220>, <OpOverload(op='aten.index_select', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968cc0>, kernel=<OpOverload(op='aten.index_select', overload='dimname')>), <OpOverload(op='aten.constant_pad_nd', overload='default')>: <function constant_pad_nd at 0x7776e1af4c20>, <OpOverload(op='aten.constant_pad_nd', overload='out')>: <function constant_pad_nd at 0x7776e1af4c20>, <OpOverload(op='aten.native_group_norm', overload='default')>: <function native_group_norm at 0x7776e1af5760>, <OpOverload(op='aten.expand', overload='default')>: <function expand at 0x7776e1af4e00>, <OpOverload(op='aten.flip', overload='default')>: <function flip at 0x7776e1af5440>, <OpOverload(op='aten.flip', overload='out')>: <function flip at 0x7776e1af5440>, <OpOverload(op='aten.permute', overload='default')>: <function permute at 0x7776e1af56c0>, <OpOverload(op='aten.istft', overload='default')>: <function istft at 0x7776e1af5f80>, <OpOverload(op='aten.renorm', overload='default')>: <function renorm at 0x7776e1af4680>, <OpOverload(op='aten.renorm', overload='out')>: <function renorm at 0x7776e1af4680>, <OpOverload(op='aten.repeat', overload='default')>: <function repeat at 0x7776e1af6340>, <OpOverload(op='aten.repeat', overload='out')>: <function repeat at 0x7776e1af6340>, <OpOverload(op='aten.roll', overload='default')>: <function roll at 0x7776e1af6980>, <OpOverload(op='aten.roll', overload='out')>: <function roll at 0x7776e1af6980>, <OpOverload(op='aten.rot90', overload='default')>: <function rot90 at 0x7776e1af6c00>, <OpOverload(op='aten.rot90', overload='out')>: <function rot90 at 0x7776e1af6c00>, <OpOverload(op='aten.stack', overload='default')>: <function stack at 0x7776e1af6f20>, <OpOverload(op='aten.stack', overload='out')>: <function stack at 0x7776e1af6f20>, <OpOverload(op='aten.unbind', overload='int')>: <function unbind at 0x7776e1af7560>, <OpOverload(op='aten.unbind', overload='Dimname')>: <function unbind at 0x7776e1af7560>, <OpOverload(op='aten.split_with_sizes', overload='default')>: <function split_with_sizes at 0x7776e1af7b00>, <OpOverload(op='aten.index_fill', overload='int_Tensor')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill', overload='int_Scalar')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill', overload='Dimname_Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968ae0>, kernel=<OpOverload(op='aten.index_fill', overload='Dimname_Tensor')>), <OpOverload(op='aten.index_fill', overload='int_Scalar_out')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill', overload='int_Tensor_out')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill_', overload='Dimname_Tensor')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.diag', overload='out')>: <function diag at 0x7776e1b10180>, <OpOverload(op='aten.index_select', overload='default')>: <function index_select at 0x7776e1b10220>, <OpOverload(op='aten.index_select', overload='out')>: <function index_select at 0x7776e1b10220>, <OpOverload(op='aten.t', overload='default')>: <function t at 0x7776e1b10cc0>, <OpOverload(op='aten.diagonal_scatter', overload='out')>: <function diagonal_scatter at 0x7776e1b104a0>, <OpOverload(op='aten.diagonal', overload='Dimname')>: <function diagonal at 0x7776e1b102c0>, <OpOverload(op='aten.diag_embed', overload='default')>: <function diag_embed at 0x7776e1b107c0>, <OpOverload(op='aten.diag_embed', overload='out')>: <function diag_embed at 0x7776e1b107c0>, <OpOverload(op='aten.block_diag', overload='default')>: <function _block_diag_iterable at 0x7776e1b10b80>, <OpOverload(op='aten.alias', overload='default')>: <function alias at 0x7776e1b10ea0>, <OpOverload(op='aten.unfold', overload='default')>: <function unfold at 0x7776e1b11080>, <OpOverload(op='aten.unfold_copy', overload='default')>: <function unfold_copy at 0x7776e1b114e0>, <OpOverload(op='aten.unfold_copy', overload='out')>: <function unfold_copy at 0x7776e1b114e0>, <OpOverload(op='aten.cumsum', overload='default')>: <function cumsum at 0x7776e1b11580>, <OpOverload(op='aten.cumsum', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329931a0>, kernel=<OpOverload(op='aten.cumsum', overload='dimname')>), <OpOverload(op='aten.cumsum', overload='dimname_out')>: <function cumsum at 0x7776e1b11580>, <OpOverload(op='aten.cumsum', overload='out')>: <function cumsum at 0x7776e1b11580>, <OpOverload(op='aten.cumprod', overload='default')>: <function cumprod at 0x7776e1b11620>, <OpOverload(op='aten.cumprod', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632968fe0>, kernel=<OpOverload(op='aten.cumprod', overload='dimname')>), <OpOverload(op='aten.cumprod', overload='dimname_out')>: <function cumprod at 0x7776e1b11620>, <OpOverload(op='aten.cumprod', overload='out')>: <function cumprod at 0x7776e1b11620>, <OpOverload(op='aten.unsqueeze', overload='default')>: <function unsqueeze at 0x7776e1b11760>, <OpOverload(op='aten.arange', overload='start_out')>: <function arange at 0x7776e1b12e80>, <OpOverload(op='aten.arange', overload='start_step')>: <function arange at 0x7776e1b12e80>, <OpOverload(op='aten.eye', overload='m')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.eye', overload='m_out')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.eye', overload='out')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.eye', overload='default')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.randn', overload='default')>: <function randn at 0x7776e1b40400>, <OpOverload(op='aten.tril_indices', overload='out')>: <function tril_indices at 0x7776e1b41a80>, <OpOverload(op='aten.lerp', overload='Scalar')>: <function lerp at 0x7776e1b13240>, <OpOverload(op='aten.lerp', overload='Tensor')>: <function lerp at 0x7776e1b13240>, <OpOverload(op='aten.lerp', overload='Scalar_out')>: <function lerp at 0x7776e1b13240>, <OpOverload(op='aten.lerp', overload='Tensor_out')>: <function lerp at 0x7776e1b13240>, <OpOverload(op='aten.linspace', overload='Tensor_Tensor')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Tensor_Scalar')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Scalar_Tensor')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='default')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Tensor_Tensor_out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Scalar_Tensor_out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Tensor_Scalar_out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.logspace', overload='Tensor_Tensor')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Tensor_Scalar')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Scalar_Tensor')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='default')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Tensor_Tensor_out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Tensor_Scalar_out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Scalar_Tensor_out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.meshgrid', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573dee0>, kernel=<OpOverload(op='aten.meshgrid', overload='default')>), <OpOverload(op='aten.meshgrid', overload='indexing')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573fce0>, kernel=<OpOverload(op='aten.meshgrid', overload='indexing')>), <OpOverload(op='aten.triu_indices', overload='out')>: <function triu_indices at 0x7776e1b419e0>, <OpOverload(op='aten.triu_indices', overload='default')>: <function triu_indices at 0x7776e1b419e0>, <OpOverload(op='aten.masked_fill', overload='Scalar')>: <function masked_fill at 0x7776e1b407c0>, <OpOverload(op='aten.masked_fill', overload='Tensor')>: <function masked_fill at 0x7776e1b407c0>, <OpOverload(op='aten.masked_fill', overload='Scalar_out')>: <function masked_fill at 0x7776e1b407c0>, <OpOverload(op='aten.masked_fill', overload='Tensor_out')>: <function masked_fill at 0x7776e1b407c0>, <OpOverload(op='aten.masked_fill_', overload='Scalar')>: <function masked_fill_ at 0x7776e1b405e0>, <OpOverload(op='aten.masked_fill_', overload='Tensor')>: <function masked_fill_ at 0x7776e1b405e0>, <OpOverload(op='aten.norm', overload='Scalar')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dim')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='names_ScalarOpt_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573ee80>, kernel=<OpOverload(op='aten.norm', overload='names_ScalarOpt_dim')>), <OpOverload(op='aten.norm', overload='dtype_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dtype')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dim_dtype')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='names_dtype_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dtype_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='Scalar_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='names_ScalarOpt_dim_dtype')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776defd8ae0>, kernel=<OpOverload(op='aten.norm', overload='names_ScalarOpt_dim_dtype')>), <OpOverload(op='aten.norm', overload='names_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.trace', overload='default')>: <function trace at 0x7776e1b40f40>, <OpOverload(op='aten.triu', overload='default')>: <function triu at 0x7776e1b41440>, <OpOverload(op='aten.triu', overload='out')>: <function triu at 0x7776e1b41440>, <OpOverload(op='aten.tril', overload='default')>: <function tril at 0x7776e1b416c0>, <OpOverload(op='aten.tril', overload='out')>: <function tril at 0x7776e1b416c0>, <OpOverload(op='aten.tril_indices', overload='default')>: <function tril_indices at 0x7776e1b41a80>, <OpOverload(op='aten.bucketize', overload='Tensor')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.bucketize', overload='Scalar')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.bucketize', overload='Tensor_out')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.bucketize', overload='Scalar_out')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.cauchy', overload='default')>: <function cauchy at 0x7776e1b41c60>, <OpOverload(op='aten.cauchy', overload='out')>: <function cauchy at 0x7776e1b41c60>, <OpOverload(op='aten.exponential', overload='default')>: <function exponential at 0x7776e1b42020>, <OpOverload(op='aten.exponential', overload='out')>: <function exponential at 0x7776e1b42020>, <OpOverload(op='aten.dot', overload='out')>: <function dot at 0x7776e1b41f80>, <OpOverload(op='aten.geometric', overload='default')>: <function geometric at 0x7776e1b423e0>, <OpOverload(op='aten.geometric', overload='out')>: <function geometric at 0x7776e1b423e0>, <OpOverload(op='aten.dot', overload='default')>: <function dot at 0x7776e1b41f80>, <OpOverload(op='aten.log_normal', overload='default')>: <function log_normal at 0x7776e1b427a0>, <OpOverload(op='aten.log_normal', overload='out')>: <function log_normal at 0x7776e1b427a0>, <OpOverload(op='aten.normal_', overload='default')>: <function normal_ at 0x7776e1b42840>, <OpOverload(op='aten.asinh_', overload='default')>: <function asinh at 0x7776e1b645e0>, <OpOverload(op='aten.addcmul_', overload='default')>: <function addcmul at 0x7776e1b64220>, <OpOverload(op='aten.rad2deg', overload='out')>: <function rad2deg at 0x7776e1b431a0>, <OpOverload(op='aten.deg2rad', overload='default')>: <function deg2rad at 0x7776e1b436a0>, <OpOverload(op='aten.deg2rad', overload='out')>: <function deg2rad at 0x7776e1b436a0>, <OpOverload(op='aten.count_nonzero', overload='dim_IntList')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.count_nonzero', overload='dim_IntList_out')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.count_nonzero', overload='default')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.count_nonzero', overload='out')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.cumprod_', overload='default')>: <function cumprod at 0x7776e1b65940>, <OpOverload(op='aten.cumprod_', overload='dimname')>: <function cumprod at 0x7776e1b65940>, <OpOverload(op='aten.vdot', overload='default')>: <function vdot at 0x7776e1b43c40>, <OpOverload(op='aten.vdot', overload='out')>: <function vdot at 0x7776e1b43c40>, <OpOverload(op='aten.le_', overload='Tensor')>: <function le at 0x7776e1b66d40>, <OpOverload(op='aten.select_scatter', overload='out')>: <function select_scatter at 0x7776e1b43ec0>, <OpOverload(op='aten.abs_', overload='default')>: <function abs at 0x7776e1b439c0>, <OpOverload(op='aten.acos_', overload='default')>: <function acos at 0x7776e1b43ce0>, <OpOverload(op='aten.cumsum_', overload='default')>: <function cumsum at 0x7776e1b659e0>, <OpOverload(op='aten.cumsum_', overload='dimname')>: <function cumsum at 0x7776e1b659e0>, <OpOverload(op='aten.acosh_', overload='default')>: <function acosh at 0x7776e1b43f60>, <OpOverload(op='aten.add_', overload='Tensor')>: <function add at 0x7776e1b640e0>, <OpOverload(op='aten.add_', overload='Scalar')>: <function add at 0x7776e1b640e0>, <OpOverload(op='aten.cosh_', overload='default')>: <function cosh at 0x7776e1b42340>, <OpOverload(op='aten.addcdiv_', overload='default')>: <function addcdiv at 0x7776e1b64360>, <OpOverload(op='aten.asin_', overload='default')>: <function asin at 0x7776e1b644a0>, <OpOverload(op='aten.cos_', overload='default')>: <function cos at 0x7776e1b43ba0>, <OpOverload(op='aten.atan_', overload='default')>: <function atan at 0x7776e1b64720>, <OpOverload(op='aten.atanh_', overload='default')>: <function atanh at 0x7776e1b64860>, <OpOverload(op='aten.copysign_', overload='Scalar')>: <function copysign at 0x7776e1b43e20>, <OpOverload(op='aten.atan2_', overload='default')>: <function atan2 at 0x7776e1b649a0>, <OpOverload(op='aten.bitwise_and_', overload='Tensor')>: <function bitwise_and at 0x7776e1b64ae0>, <OpOverload(op='aten.bitwise_and_', overload='Scalar')>: <function bitwise_and at 0x7776e1b64ae0>, <OpOverload(op='aten.copysign_', overload='Tensor')>: <function copysign at 0x7776e1b43e20>, <OpOverload(op='aten.bitwise_left_shift_', overload='Tensor_Scalar')>: <function bitwise_left_shift at 0x7776e1b64c20>, <OpOverload(op='aten.bitwise_left_shift_', overload='Tensor')>: <function bitwise_left_shift at 0x7776e1b64c20>, <OpOverload(op='aten.bitwise_not_', overload='default')>: <function bitwise_not at 0x7776e1b64d60>, <OpOverload(op='aten.bitwise_or_', overload='Tensor')>: <function bitwise_or at 0x7776e1b64ea0>, <OpOverload(op='aten.bitwise_or_', overload='Scalar')>: <function bitwise_or at 0x7776e1b64ea0>, <OpOverload(op='aten.bitwise_right_shift_', overload='Tensor_Scalar')>: <function bitwise_right_shift at 0x7776e1b64fe0>, <OpOverload(op='aten.bitwise_right_shift_', overload='Tensor')>: <function bitwise_right_shift at 0x7776e1b64fe0>, <OpOverload(op='aten.bitwise_xor_', overload='Tensor')>: <function bitwise_xor at 0x7776e1b65120>, <OpOverload(op='aten.bitwise_xor_', overload='Scalar')>: <function bitwise_xor at 0x7776e1b65120>, <OpOverload(op='aten.ceil_', overload='default')>: <function ceil at 0x7776e1b65260>, <OpOverload(op='aten.conj_physical_', overload='default')>: <function conj_physical at 0x7776e1b65760>, <OpOverload(op='aten.clamp_', overload='default')>: <function clamp at 0x7776e1b653a0>, <OpOverload(op='aten.clamp_', overload='Tensor')>: <function clamp at 0x7776e1b653a0>, <OpOverload(op='aten.deg2rad_', overload='default')>: <function deg2rad at 0x7776e1b656c0>, <OpOverload(op='aten.clamp_min_', overload='default')>: <function clamp_min at 0x7776e1b654e0>, <OpOverload(op='aten.clamp_min_', overload='Tensor')>: <function clamp_min at 0x7776e1b654e0>, <OpOverload(op='aten.lt_', overload='Scalar')>: <function lt at 0x7776e1b65c60>, <OpOverload(op='aten.digamma_', overload='default')>: <function digamma at 0x7776e1b65440>, <OpOverload(op='aten.clamp_max_', overload='default')>: <function clamp_max at 0x7776e1b65620>, <OpOverload(op='aten.clamp_max_', overload='Tensor')>: <function clamp_max at 0x7776e1b65620>, <OpOverload(op='aten.div_', overload='Tensor')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.div_', overload='Tensor_mode')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.div_', overload='Scalar')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.div_', overload='Scalar_mode')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.logical_xor_', overload='default')>: <function logical_xor at 0x7776e1b65ee0>, <OpOverload(op='aten.cauchy_', overload='default')>: <function cauchy at 0x7776e1b677e0>, <OpOverload(op='aten.eq_', overload='Scalar')>: <function eq at 0x7776e1b64f40>, <OpOverload(op='aten.eq_', overload='Tensor')>: <function eq at 0x7776e1b64f40>, <OpOverload(op='aten.erf_', overload='default')>: <function erf at 0x7776e1b64cc0>, <OpOverload(op='aten.logical_or_', overload='default')>: <function logical_or at 0x7776e1b66160>, <OpOverload(op='aten.erfc_', overload='default')>: <function erfc at 0x7776e1b64a40>, <OpOverload(op='aten.erfinv_', overload='default')>: <function erfinv at 0x7776e1b647c0>, <OpOverload(op='aten.exponential_', overload='default')>: <function exponential at 0x7776e1b67560>, <OpOverload(op='aten.exp_', overload='default')>: <function exp at 0x7776e1b64540>, <OpOverload(op='aten.logical_not_', overload='default')>: <function logical_not at 0x7776e1b663e0>, <OpOverload(op='aten.exp2_', overload='default')>: <function exp2 at 0x7776e1b642c0>, <OpOverload(op='aten.expm1_', overload='default')>: <function expm1 at 0x7776e1b64040>, <OpOverload(op='aten.logical_and_', overload='default')>: <function logical_and at 0x7776e1b66660>, <OpOverload(op='aten.float_power_', overload='Tensor')>: <function float_power at 0x7776e1b65bc0>, <OpOverload(op='aten.float_power_', overload='Scalar')>: <function float_power at 0x7776e1b65bc0>, <OpOverload(op='aten.floor_', overload='default')>: <function floor at 0x7776e1b65d00>, <OpOverload(op='aten.floor_divide_', overload='Scalar')>: <function floor_divide at 0x7776e1b65e40>, <OpOverload(op='aten.floor_divide_', overload='Tensor')>: <function floor_divide at 0x7776e1b65e40>, <OpOverload(op='aten.log_', overload='default')>: <function log at 0x7776e1b668e0>, <OpOverload(op='aten.fmod_', overload='Tensor')>: <function fmod at 0x7776e1b65f80>, <OpOverload(op='aten.fmod_', overload='Scalar')>: <function fmod at 0x7776e1b65f80>, <OpOverload(op='aten.frac_', overload='default')>: <function frac at 0x7776e1b660c0>, <OpOverload(op='aten.geometric_', overload='default')>: <function geometric at 0x7776e1b672e0>, <OpOverload(op='aten.log2_', overload='default')>: <function log2 at 0x7776e1b66b60>, <OpOverload(op='aten.gcd_', overload='default')>: <function gcd at 0x7776e1b66200>, <OpOverload(op='aten.ge_', overload='Scalar')>: <function ge at 0x7776e1b66340>, <OpOverload(op='aten.ge_', overload='Tensor')>: <function ge at 0x7776e1b66340>, <OpOverload(op='aten.gt_', overload='Scalar')>: <function gt at 0x7776e1b66480>, <OpOverload(op='aten.gt_', overload='Tensor')>: <function gt at 0x7776e1b66480>, <OpOverload(op='aten.log1p_', overload='default')>: <function log1p at 0x7776e1b66de0>, <OpOverload(op='aten.heaviside_', overload='default')>: <function heaviside at 0x7776e1b665c0>, <OpOverload(op='aten.log_normal_', overload='default')>: <function log_normal at 0x7776e1b658a0>, <OpOverload(op='aten.hypot_', overload='default')>: <function hypot at 0x7776e1b66700>, <OpOverload(op='aten.igamma_', overload='default')>: <function igamma at 0x7776e1b66840>, <OpOverload(op='aten.igammac_', overload='default')>: <function igammac at 0x7776e1b66980>, <OpOverload(op='aten.lgamma_', overload='default')>: <function lgamma at 0x7776e1b66fc0>, <OpOverload(op='aten.zero_', overload='default')>: <function zero at 0x7776e1b65300>, <OpOverload(op='aten.i0_', overload='default')>: <function i0 at 0x7776e1b66ac0>, <OpOverload(op='aten.lcm_', overload='default')>: <function lcm at 0x7776e1b66c00>, <OpOverload(op='aten.lerp_', overload='Scalar')>: <function lerp at 0x7776e1b66e80>, <OpOverload(op='aten.le_', overload='Scalar')>: <function le at 0x7776e1b66d40>, <OpOverload(op='aten.lt_', overload='Tensor')>: <function lt at 0x7776e1b65c60>, <OpOverload(op='aten.mul_', overload='Tensor')>: <function mul at 0x7776e1b64180>, <OpOverload(op='aten.mul_', overload='Scalar')>: <function mul at 0x7776e1b64180>, <OpOverload(op='aten.mvlgamma_', overload='default')>: <function _make_alias.<locals>._fn at 0x7776e1b64680>, <OpOverload(op='aten.nan_to_num_', overload='default')>: <function nan_to_num at 0x7776e1b64b80>, <OpOverload(op='aten.xlogy_', overload='Tensor')>: <function xlogy at 0x7776e1b67a60>, <OpOverload(op='aten.xlogy_', overload='Scalar_Other')>: <function xlogy at 0x7776e1b67a60>, <OpOverload(op='aten.ne_', overload='Scalar')>: <function ne at 0x7776e1b65080>, <OpOverload(op='aten.ne_', overload='Tensor')>: <function ne at 0x7776e1b65080>, <OpOverload(op='aten.neg_', overload='default')>: <function neg at 0x7776e1b65580>, <OpOverload(op='aten.nextafter_', overload='default')>: <function nextafter at 0x7776e1b65a80>, <OpOverload(op='aten.pow_', overload='Scalar')>: <function pow at 0x7776e1b67100>, <OpOverload(op='aten.pow_', overload='Tensor')>: <function pow at 0x7776e1b67100>, <OpOverload(op='aten.trunc_', overload='default')>: <function trunc at 0x7776e1b67ce0>, <OpOverload(op='aten.rad2deg_', overload='default')>: <function rad2deg at 0x7776e1b67240>, <OpOverload(op='aten.reciprocal_', overload='default')>: <function reciprocal at 0x7776e1b67380>, <OpOverload(op='aten.true_divide_', overload='Scalar')>: <function true_divide at 0x7776e1b67e20>, <OpOverload(op='aten.remainder_', overload='Tensor')>: <function remainder at 0x7776e1b674c0>, <OpOverload(op='aten.remainder_', overload='Scalar')>: <function remainder at 0x7776e1b674c0>, <OpOverload(op='aten.rsqrt_', overload='default')>: <function rsqrt at 0x7776e1b67600>, <OpOverload(op='aten.true_divide_', overload='Tensor')>: <function true_divide at 0x7776e1b67e20>, <OpOverload(op='aten.transpose_copy', overload='int')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.transpose at 0x7776e1b64400>, <OpOverload(op='aten.sgn_', overload='default')>: <function sgn at 0x7776e1b67740>, <OpOverload(op='aten.sigmoid_', overload='default')>: <function sigmoid at 0x7776e1b67880>, <OpOverload(op='aten.triu_', overload='default')>: <function triu at 0x7776e1b67f60>, <OpOverload(op='aten.sign_', overload='default')>: <function sign at 0x7776e1b679c0>, <OpOverload(op='aten.sin_', overload='default')>: <function sin at 0x7776e1b67b00>, <OpOverload(op='aten.transpose_copy', overload='int_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.transpose at 0x7776e1b64400>, <OpOverload(op='aten.sinc_', overload='default')>: <function sinc at 0x7776e1b67c40>, <OpOverload(op='aten.sinh_', overload='default')>: <function sinh at 0x7776e1b67d80>, <OpOverload(op='aten.sqrt_', overload='default')>: <function sqrt at 0x7776e1b67ec0>, <OpOverload(op='aten.square_', overload='default')>: <function square at 0x7776e1b98040>, <OpOverload(op='aten.tril_', overload='default')>: <function tril at 0x7776e1b98540>, <OpOverload(op='aten.sub_', overload='Tensor')>: <function sub at 0x7776e1b98180>, <OpOverload(op='aten.sub_', overload='Scalar')>: <function sub at 0x7776e1b98180>, <OpOverload(op='aten.tan_', overload='default')>: <function tan at 0x7776e1b982c0>, <OpOverload(op='aten.tanh_', overload='default')>: <function tanh at 0x7776e1b98400>, <OpOverload(op='aten.unbind_copy', overload='int')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unbind at 0x7776e1b65800>, <OpOverload(op='aten.alias_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.alias at 0x7776e1b65b20>, <OpOverload(op='aten.alias_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.alias at 0x7776e1b65b20>, <OpOverload(op='aten.as_strided_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.as_strided at 0x7776e1b667a0>, <OpOverload(op='aten.as_strided_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.as_strided at 0x7776e1b667a0>, <OpOverload(op='aten.diagonal_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.diagonal at 0x7776e1b987c0>, <OpOverload(op='aten.diagonal_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.diagonal at 0x7776e1b987c0>, <OpOverload(op='aten.expand_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.expand at 0x7776e1b98360>, <OpOverload(op='aten.expand_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.expand at 0x7776e1b98360>, <OpOverload(op='aten.narrow_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.narrow at 0x7776e1b98a40>, <OpOverload(op='aten.squeeze_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='dims')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='dim_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='dims_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.permute_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.permute at 0x7776e1b99080>, <OpOverload(op='aten.permute_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.permute at 0x7776e1b99080>, <OpOverload(op='aten.t_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.t at 0x7776e1b993a0>, <OpOverload(op='aten.t_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.t at 0x7776e1b993a0>, <OpOverload(op='aten.unbind_copy', overload='int_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unbind at 0x7776e1b65800>, <OpOverload(op='aten.unsqueeze_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unsqueeze at 0x7776e1b67ba0>, <OpOverload(op='aten.unsqueeze_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unsqueeze at 0x7776e1b67ba0>, <OpOverload(op='aten.view_copy', overload='dtype')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.view_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.view_copy', overload='dtype_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.complex', overload='default')>: <function complex at 0x7776e1b9ab60>, <OpOverload(op='aten.complex', overload='out')>: <function complex at 0x7776e1b9ab60>, <OpOverload(op='aten.fft_ifft', overload='out')>: <function ifft at 0x7776e1b9bb00>, <OpOverload(op='aten.polar', overload='default')>: <function polar at 0x7776e1b9ade0>, <OpOverload(op='aten.polar', overload='out')>: <function polar at 0x7776e1b9ade0>, <OpOverload(op='aten.fft_fft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776345bb100>, kernel=<OpOverload(op='aten.fft_fft', overload='default')>), <OpOverload(op='aten.fft_fft', overload='out')>: <function fft at 0x7776e1b9b880>, <OpOverload(op='aten.fft_ifft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763539cd60>, kernel=<OpOverload(op='aten.fft_ifft', overload='default')>), <OpOverload(op='aten.fft_rfft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763417b880>, kernel=<OpOverload(op='aten.fft_rfft', overload='default')>), <OpOverload(op='aten.fft_rfft', overload='out')>: <function rfft at 0x7776e1b9b7e0>, <OpOverload(op='aten.fft_irfft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d68c20>, kernel=<OpOverload(op='aten.fft_irfft', overload='default')>), <OpOverload(op='aten.fft_irfft', overload='out')>: <function irfft at 0x7776e1b98e00>, <OpOverload(op='aten.fft_hfft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777634178ea0>, kernel=<OpOverload(op='aten.fft_hfft', overload='default')>), <OpOverload(op='aten.fft_hfft', overload='out')>: <function hfft at 0x7776e1b9bd80>, <OpOverload(op='aten.fft_fft2', overload='out')>: <function fft2 at 0x7776e1b9bce0>, <OpOverload(op='aten.fft_ihfft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763417a980>, kernel=<OpOverload(op='aten.fft_ihfft', overload='default')>), <OpOverload(op='aten.fft_ihfft', overload='out')>: <function ihfft at 0x7776e1b9bf60>, <OpOverload(op='aten.fft_fft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632957ba0>, kernel=<OpOverload(op='aten.fft_fft2', overload='default')>), <OpOverload(op='aten.fft_fftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296a520>, kernel=<OpOverload(op='aten.fft_fftn', overload='default')>), <OpOverload(op='aten.fft_fftn', overload='out')>: <function fftn at 0x7776e19d8900>, <OpOverload(op='aten.fft_ifftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969ee0>, kernel=<OpOverload(op='aten.fft_ifftn', overload='default')>), <OpOverload(op='aten.fft_rfftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b6a0>, kernel=<OpOverload(op='aten.fft_rfftn', overload='default')>), <OpOverload(op='aten.fft_rfftn', overload='out')>: <function rfftn at 0x7776e19d8e00>, <OpOverload(op='aten.fft_ihfftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b060>, kernel=<OpOverload(op='aten.fft_ihfftn', overload='default')>), <OpOverload(op='aten.fft_ihfftn', overload='out')>: <function ihfftn at 0x7776e19d9080>, <OpOverload(op='aten.fft_irfftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b100>, kernel=<OpOverload(op='aten.fft_irfftn', overload='default')>), <OpOverload(op='aten.fft_irfftn', overload='out')>: <function irfftn at 0x7776e19d9760>, <OpOverload(op='aten.scatter_', overload='src')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_', overload='value')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_', overload='reduce')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_', overload='value_reduce')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_add_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc680>, <OpOverload(op='aten.scatter_reduce_', overload='two')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbe020>, <OpOverload(op='aten.silu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbe160>, <OpOverload(op='aten.is_complex', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d69300>, kernel=<OpOverload(op='aten.is_complex', overload='default')>), <OpOverload(op='aten.erfinv', overload='default')>: <function erfinv at 0x7776e1c3f7e0>, <OpOverload(op='aten.erfinv', overload='out')>: <function erfinv at 0x7776e1c3f7e0>, <OpOverload(op='aten.zero', overload='default')>: <function zero at 0x7776e1c4d260>, <OpOverload(op='aten.zero', overload='out')>: <function zero at 0x7776e1c4d260>, <OpOverload(op='aten.block_diag', overload='out')>: <function _block_diag_iterable at 0x7776e1b10b80>, <OpOverload(op='aten.frac', overload='default')>: <function frac at 0x7776e1c4dc60>, <OpOverload(op='aten.frac', overload='out')>: <function frac at 0x7776e1c4dc60>, <OpOverload(op='aten.isinf', overload='default')>: <function isinf at 0x7776e1c4e660>, <OpOverload(op='aten.isinf', overload='out')>: <function isinf at 0x7776e1c4e660>, <OpOverload(op='aten.isposinf', overload='default')>: <function isposinf at 0x7776e1c4d800>, <OpOverload(op='aten.isposinf', overload='out')>: <function isposinf at 0x7776e1c4d800>, <OpOverload(op='aten.isneginf', overload='default')>: <function isneginf at 0x7776e1c4e980>, <OpOverload(op='aten.isneginf', overload='out')>: <function isneginf at 0x7776e1c4e980>, <OpOverload(op='aten.isnan', overload='default')>: <function isnan at 0x7776e1c4ee80>, <OpOverload(op='aten.isnan', overload='out')>: <function isnan at 0x7776e1c4ee80>, <OpOverload(op='aten.i0', overload='default')>: <function i0 at 0x7776e1c4f880>, <OpOverload(op='aten.i0', overload='out')>: <function i0 at 0x7776e1c4f880>, <OpOverload(op='aten.index_fill_', overload='Dimname_Scalar')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.logsumexp', overload='names_out')>: <function logsumexp at 0x7776e1c65760>, <OpOverload(op='aten.logsumexp', overload='names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969580>, kernel=<OpOverload(op='aten.logsumexp', overload='names')>), <OpOverload(op='aten.logsumexp', overload='out')>: <function logsumexp at 0x7776e1c65760>, <OpOverload(op='aten.logsumexp', overload='default')>: <function logsumexp at 0x7776e1c65760>, <OpOverload(op='aten.nan_to_num', overload='default')>: <function nan_to_num at 0x7776e1c4d1c0>, <OpOverload(op='aten.nan_to_num', overload='out')>: <function nan_to_num at 0x7776e1c4d1c0>, <OpOverload(op='aten.sigmoid', overload='default')>: <function sigmoid at 0x7776e1c66de0>, <OpOverload(op='aten.sigmoid', overload='out')>: <function sigmoid at 0x7776e1c66de0>, <OpOverload(op='aten.std_mean', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776331e6f20>, kernel=<OpOverload(op='aten.std_mean', overload='dim')>), <OpOverload(op='aten.sgn', overload='default')>: <function sgn at 0x7776e1c672e0>, <OpOverload(op='aten.sgn', overload='out')>: <function sgn at 0x7776e1c672e0>, <OpOverload(op='aten.sinc', overload='default')>: <function sinc at 0x7776e1c74720>, <OpOverload(op='aten.sinc', overload='out')>: <function sinc at 0x7776e1c74720>, <OpOverload(op='aten.std', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777633d6bd80>, kernel=<OpOverload(op='aten.std', overload='names_dim')>), <OpOverload(op='aten.bitwise_right_shift', overload='Tensor_Scalar_out')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_right_shift', overload='Scalar_Tensor_out')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_left_shift', overload='Tensor_out')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.xlogy', overload='OutTensor')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.bitwise_left_shift', overload='Tensor')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_left_shift', overload='Tensor_Scalar')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_left_shift', overload='Scalar_Tensor')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_left_shift', overload='Tensor_Scalar_out')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_left_shift', overload='Scalar_Tensor_out')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_right_shift', overload='Tensor')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_right_shift', overload='Tensor_Scalar')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_right_shift', overload='Tensor_out')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_right_shift', overload='Scalar_Tensor')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.copysign', overload='Tensor')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.copysign', overload='Scalar')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.copysign', overload='out')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.copysign', overload='Scalar_out')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.lerp_', overload='Tensor')>: <function lerp at 0x7776e1b66e80>, <OpOverload(op='aten.trace', overload='out')>: <function trace at 0x7776e1b40f40>, <OpOverload(op='aten.heaviside', overload='default')>: <function heaviside at 0x7776e1c96ca0>, <OpOverload(op='aten.heaviside', overload='out')>: <function heaviside at 0x7776e1c96ca0>, <OpOverload(op='aten.logical_and', overload='out')>: <function logical_and at 0x7776e1cad120>, <OpOverload(op='aten.logical_and', overload='default')>: <function logical_and at 0x7776e1cad120>, <OpOverload(op='aten.std', overload='correction_out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.lcm', overload='default')>: <function lcm at 0x7776e1c97f60>, <OpOverload(op='aten.lcm', overload='out')>: <function lcm at 0x7776e1c97f60>, <OpOverload(op='aten.squeeze_copy', overload='dim')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.logaddexp', overload='default')>: <function logaddexp at 0x7776e1cac860>, <OpOverload(op='aten.logaddexp', overload='out')>: <function logaddexp at 0x7776e1cac860>, <OpOverload(op='aten.logaddexp2', overload='default')>: <function logaddexp2 at 0x7776e1caccc0>, <OpOverload(op='aten.logaddexp2', overload='out')>: <function logaddexp2 at 0x7776e1caccc0>, <OpOverload(op='aten.logical_not', overload='default')>: <function logical_not at 0x7776e1cad080>, <OpOverload(op='aten.logical_not', overload='out')>: <function logical_not at 0x7776e1cad080>, <OpOverload(op='aten.logical_or', overload='default')>: <function logical_or at 0x7776e1cad300>, <OpOverload(op='aten.logical_or', overload='out')>: <function logical_or at 0x7776e1cad300>, <OpOverload(op='aten.logical_xor', overload='default')>: <function logical_xor at 0x7776e1cad760>, <OpOverload(op='aten.logical_xor', overload='out')>: <function logical_xor at 0x7776e1cad760>, <OpOverload(op='aten.rsub', overload='Tensor')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.rsub', overload='Scalar')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.rsub', overload='Tensor_out')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.rsub', overload='Scalar_out')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.xlogy', overload='Scalar_Other')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.xlogy', overload='Scalar_Self')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.xlogy', overload='OutScalar_Other')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.xlogy', overload='OutScalar_Self')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.addcdiv', overload='default')>: <function addcdiv at 0x7776e1ad04a0>, <OpOverload(op='aten.addcdiv', overload='out')>: <function addcdiv at 0x7776e1ad04a0>, <OpOverload(op='aten.addcmul', overload='default')>: <function addcmul at 0x7776e1ad0860>, <OpOverload(op='aten.addcmul', overload='out')>: <function addcmul at 0x7776e1ad0860>, <OpOverload(op='aten.clamp', overload='default')>: <function clamp at 0x7776e1ad0c20>, <OpOverload(op='aten.clamp', overload='Tensor')>: <function clamp at 0x7776e1ad0c20>, <OpOverload(op='aten.clamp', overload='out')>: <function clamp at 0x7776e1ad0c20>, <OpOverload(op='aten.clamp', overload='Tensor_out')>: <function clamp at 0x7776e1ad0c20>, <OpOverload(op='aten.clamp_min', overload='default')>: <function clamp_min at 0x7776e1ad0ea0>, <OpOverload(op='aten.clamp_min', overload='Tensor')>: <function clamp_min at 0x7776e1ad0ea0>, <OpOverload(op='aten.clamp_min', overload='out')>: <function clamp_min at 0x7776e1ad0ea0>, <OpOverload(op='aten.clamp_min', overload='Tensor_out')>: <function clamp_min at 0x7776e1ad0ea0>, <OpOverload(op='aten.frexp', overload='Tensor_out')>: <function frexp at 0x7776e1c95f80>, <OpOverload(op='aten._adaptive_avg_pool2d', overload='default')>: <function adaptive_avg_pool2d at 0x7776e1d5a700>, <OpOverload(op='aten.relu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd440>, <OpOverload(op='aten.addmm', overload='dtype_out')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.sigmoid_backward', overload='grad_input')>: <function sigmoid_backward at 0x7776e1cde020>, <OpOverload(op='aten.sigmoid_backward', overload='default')>: <function sigmoid_backward at 0x7776e1cde020>, <OpOverload(op='aten.hardswish', overload='out')>: <function hardswish at 0x7776e1cdef20>, <OpOverload(op='aten.softplus_backward', overload='default')>: <function softplus_backward at 0x7776e1cde3e0>, <OpOverload(op='aten.softplus_backward', overload='grad_input')>: <function softplus_backward at 0x7776e1cde3e0>, <OpOverload(op='aten.hardswish', overload='default')>: <function hardswish at 0x7776e1cdef20>, <OpOverload(op='aten.elu_backward', overload='default')>: <function elu_backward at 0x7776e1cddf80>, <OpOverload(op='aten.hardsigmoid', overload='default')>: <function hardsigmoid at 0x7776e1cdefc0>, <OpOverload(op='aten.hardsigmoid', overload='out')>: <function hardsigmoid at 0x7776e1cdefc0>, <OpOverload(op='aten.hardsigmoid_backward', overload='default')>: <function hardsigmoid_backward at 0x7776e1cdf060>, <OpOverload(op='aten.hardsigmoid_backward', overload='grad_input')>: <function hardsigmoid_backward at 0x7776e1cdf060>, <OpOverload(op='aten.hardtanh_backward', overload='default')>: <function hardtanh_backward at 0x7776e1cdf420>, <OpOverload(op='aten.hardtanh_backward', overload='grad_input')>: <function hardtanh_backward at 0x7776e1cdf420>, <OpOverload(op='aten.hardswish_backward', overload='default')>: <function hardswish_backward at 0x7776e1cde200>, <OpOverload(op='aten.threshold_backward', overload='default')>: <function threshold_backward at 0x7776e1cddda0>, <OpOverload(op='aten.threshold_backward', overload='grad_input')>: <function threshold_backward at 0x7776e1cddda0>, <OpOverload(op='aten.leaky_relu_backward', overload='default')>: <function leaky_relu_backward at 0x7776e1cdfb00>, <OpOverload(op='aten.leaky_relu_backward', overload='grad_input')>: <function leaky_relu_backward at 0x7776e1cdfb00>, <OpOverload(op='aten.gelu_backward', overload='default')>: <function gelu_backward at 0x7776e1cdfec0>, <OpOverload(op='aten.gelu_backward', overload='grad_input')>: <function gelu_backward at 0x7776e1cdfec0>, <OpOverload(op='aten.mish_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eb880>, kernel=<OpOverload(op='aten.mish_backward', overload='default')>), <OpOverload(op='aten.silu', overload='default')>: <function silu at 0x7776e1cfc720>, <OpOverload(op='aten.silu', overload='out')>: <function silu at 0x7776e1cfc720>, <OpOverload(op='aten.silu_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9bc0>, kernel=<OpOverload(op='aten.silu_backward', overload='default')>), <OpOverload(op='aten.silu_backward', overload='grad_input')>: <function silu_backward at 0x7776e1cfc7c0>, <OpOverload(op='aten._prelu_kernel', overload='default')>: <function _prelu_kernel at 0x7776e1cfcae0>, <OpOverload(op='aten._prelu_kernel_backward', overload='default')>: <function _prelu_kernel_backward at 0x7776e1cfcb80>, <OpOverload(op='aten.rrelu_with_noise_backward', overload='default')>: <function rrelu_with_noise_backward at 0x7776e1cfd080>, <OpOverload(op='aten.rrelu_with_noise_backward', overload='out')>: <function rrelu_with_noise_backward at 0x7776e1cfd080>, <OpOverload(op='aten.mse_loss', overload='out')>: <function mse_loss at 0x7776e1cfd8a0>, <OpOverload(op='aten.mse_loss', overload='default')>: <function mse_loss at 0x7776e1cfd8a0>, <OpOverload(op='aten.log_sigmoid_backward', overload='grad_input')>: <function log_sigmoid_backward at 0x7776e1cfd120>, <OpOverload(op='aten.smooth_l1_loss_backward', overload='default')>: <function smooth_l1_loss_backward at 0x7776e1cfd9e0>, <OpOverload(op='aten.mse_loss_backward', overload='default')>: <function mse_loss_backward at 0x7776e1cdf240>, <OpOverload(op='aten.mse_loss_backward', overload='grad_input')>: <function mse_loss_backward at 0x7776e1cdf240>, <OpOverload(op='aten._safe_softmax', overload='default')>: <function safe_softmax at 0x7776e1cfcd60>, <OpOverload(op='aten.smooth_l1_loss_backward', overload='grad_input')>: <function smooth_l1_loss_backward_out at 0x7776e1cfdbc0>, <OpOverload(op='aten.smooth_l1_loss', overload='default')>: <function smooth_l1_loss at 0x7776e1cfd940>, <OpOverload(op='aten.smooth_l1_loss', overload='out')>: <function smooth_l1_loss at 0x7776e1cfd940>, <OpOverload(op='aten.huber_loss_backward', overload='default')>: <function huber_loss_backward at 0x7776e1cfdda0>, <OpOverload(op='aten.huber_loss_backward', overload='out')>: <function huber_loss_backward_out at 0x7776e1cfdf80>, <OpOverload(op='aten.glu_backward', overload='default')>: <function glu_backward at 0x7776e1cfe200>, <OpOverload(op='aten.glu_backward', overload='grad_input')>: <function glu_backward at 0x7776e1cfe200>, <OpOverload(op='aten.nll_loss_backward', overload='grad_input')>: <function nll_loss_backward at 0x7776e1cfe5c0>, <OpOverload(op='aten.nll_loss2d_backward', overload='default')>: <function nll_loss2d_backward at 0x7776e1cfe8e0>, <OpOverload(op='aten.nll_loss2d_backward', overload='grad_input')>: <function nll_loss2d_backward at 0x7776e1cfe8e0>, <OpOverload(op='aten.binary_cross_entropy', overload='default')>: <function binary_cross_entropy at 0x7776e1cfee80>, <OpOverload(op='aten.binary_cross_entropy', overload='out')>: <function binary_cross_entropy at 0x7776e1cfee80>, <OpOverload(op='aten.binary_cross_entropy_backward', overload='grad_input')>: <function binary_cross_entropy_backward at 0x7776e1cfef20>, <OpOverload(op='aten.soft_margin_loss_backward', overload='default')>: <function soft_margin_loss_backward at 0x7776e1cff4c0>, <OpOverload(op='aten.soft_margin_loss', overload='default')>: <function soft_margin_loss at 0x7776e1cff560>, <OpOverload(op='aten.unfold_backward', overload='out')>: <function unfold_backward at 0x7776e1cfdd00>, <OpOverload(op='aten.dist', overload='default')>: <function dist at 0x7776e1cfdb20>, <OpOverload(op='aten.dist', overload='out')>: <function dist at 0x7776e1cfdb20>, <OpOverload(op='aten._euclidean_dist', overload='default')>: <function _euclidean_dist at 0x7776e1cff2e0>, <OpOverload(op='aten._euclidean_dist', overload='out')>: <function _euclidean_dist at 0x7776e1cff2e0>, <OpOverload(op='aten.slice_backward', overload='out')>: <function slice_backward at 0x7776e1cff7e0>, <OpOverload(op='aten.slice_backward', overload='default')>: <function slice_backward at 0x7776e1cff7e0>, <OpOverload(op='aten.select_backward', overload='default')>: <function select_backward at 0x7776e1cffec0>, <OpOverload(op='aten.select_backward', overload='out')>: <function select_backward at 0x7776e1cffec0>, <OpOverload(op='aten.diagonal_backward', overload='default')>: <function diagonal_backward at 0x7776e1d28180>, <OpOverload(op='aten.logit_backward', overload='default')>: <function logit_backward at 0x7776e1cfea20>, <OpOverload(op='aten._softmax_backward_data', overload='default')>: <function _softmax_backward_data at 0x7776e1d282c0>, <OpOverload(op='aten._softmax_backward_data', overload='out')>: <function _softmax_backward_data at 0x7776e1d282c0>, <OpOverload(op='aten._log_softmax_backward_data', overload='default')>: <function _log_softmax_backward_data at 0x7776e1d28900>, <OpOverload(op='aten._log_softmax_backward_data', overload='out')>: <function _log_softmax_backward_data at 0x7776e1d28900>, <OpOverload(op='aten.im2col', overload='out')>: <function im2col at 0x7776e1d28c20>, <OpOverload(op='aten.native_dropout_backward', overload='out')>: <function native_dropout_backward at 0x7776e1d29260>, <OpOverload(op='aten.col2im', overload='out')>: <function col2im at 0x7776e1d28fe0>, <OpOverload(op='aten.col2im', overload='default')>: <function col2im at 0x7776e1d28fe0>, <OpOverload(op='aten.native_layer_norm_backward', overload='default')>: <function native_layer_norm_backward at 0x7776e1d29080>, <OpOverload(op='aten.lift', overload='out')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.native_dropout_backward', overload='default')>: <function native_dropout_backward at 0x7776e1d29260>, <OpOverload(op='aten.dropout', overload='default')>: <function dropout at 0x7776e1d291c0>, <OpOverload(op='aten.native_dropout', overload='default')>: <function native_dropout at 0x7776e1d29620>, <OpOverload(op='aten.native_dropout', overload='out')>: <function native_dropout at 0x7776e1d29620>, <OpOverload(op='aten._softmax', overload='default')>: <function _softmax at 0x7776e1d299e0>, <OpOverload(op='aten._softmax', overload='out')>: <function _softmax at 0x7776e1d299e0>, <OpOverload(op='aten._log_softmax', overload='default')>: <function _log_softmax at 0x7776e1d29c60>, <OpOverload(op='aten._log_softmax', overload='out')>: <function _log_softmax at 0x7776e1d29c60>, <OpOverload(op='aten.embedding', overload='default')>: <function embedding at 0x7776e1d29ee0>, <OpOverload(op='aten.embedding', overload='out')>: <function embedding at 0x7776e1d29ee0>, <OpOverload(op='aten._chunk_cat', overload='default')>: <function _chunk_cat at 0x7776e1d2a480>, <OpOverload(op='aten._chunk_cat', overload='out')>: <function _chunk_cat at 0x7776e1d2a480>, <OpOverload(op='aten.embedding_dense_backward', overload='default')>: <function embedding_dense_backward at 0x7776e1d2a160>, <OpOverload(op='aten.embedding_dense_backward', overload='out')>: <function embedding_dense_backward at 0x7776e1d2a160>, <OpOverload(op='aten.split_with_sizes_copy', overload='default')>: <function split_with_sizes_copy at 0x7776e1d2a520>, <OpOverload(op='aten.split_with_sizes_copy', overload='out')>: <function split_with_sizes_copy at 0x7776e1d2a520>, <OpOverload(op='aten._addmm_activation', overload='default')>: <function _addmm_activation at 0x7776e1d2aac0>, <OpOverload(op='aten.unsafe_split', overload='Tensor')>: <function unsafe_split at 0x7776e1d2a660>, <OpOverload(op='aten.unsafe_split_with_sizes', overload='default')>: <function unsafe_split_with_sizes at 0x7776e1d2a7a0>, <OpOverload(op='aten.split', overload='Tensor')>: <function split at 0x7776e1d2a8e0>, <OpOverload(op='aten.native_group_norm_backward', overload='out')>: <function native_group_norm_backward_out at 0x7776e1d29bc0>, <OpOverload(op='aten.addmm', overload='default')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.addmm', overload='out')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.addmm', overload='dtype')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.native_group_norm_backward', overload='default')>: <function native_group_norm_backward at 0x7776e1d29e40>, <OpOverload(op='aten.addmv', overload='default')>: <function addmv at 0x7776e1d2a0c0>, <OpOverload(op='aten.native_batch_norm_backward', overload='default')>: <function native_batch_norm_backward at 0x7776e1d59300>, <OpOverload(op='aten._fused_rms_norm_backward', overload='default')>: <function _fused_rms_norm_backward at 0x7776e1d2b240>, <OpOverload(op='aten.native_batch_norm_backward', overload='out')>: <function native_batch_norm_backward_out at 0x7776e1d59440>, <OpOverload(op='aten._native_batch_norm_legit', overload='default')>: <function _native_batch_norm_legit at 0x7776e1d2bec0>, <OpOverload(op='aten.native_batch_norm', overload='default')>: <function native_batch_norm at 0x7776e1d2ba60>, <OpOverload(op='aten.native_batch_norm', overload='out')>: <function native_batch_norm at 0x7776e1d2ba60>, <OpOverload(op='aten._native_batch_norm_legit', overload='no_stats')>: <function _native_batch_norm_legit_no_stats at 0x7776e1d58040>, <OpOverload(op='aten._native_batch_norm_legit_functional', overload='default')>: <function _native_batch_norm_legit_functional at 0x7776e1d58180>, <OpOverload(op='aten._batch_norm_with_update', overload='default')>: <function _batch_norm_with_update at 0x7776e1d58400>, <OpOverload(op='aten._batch_norm_with_update_functional', overload='default')>: <function _batch_norm_with_update_functional at 0x7776e1d584a0>, <OpOverload(op='aten._batch_norm_no_update', overload='default')>: <function _batch_norm_no_update at 0x7776e1d585e0>, <OpOverload(op='aten.batch_norm_backward', overload='default')>: <function batch_norm_backward at 0x7776e1d580e0>, <OpOverload(op='aten._to_copy', overload='default')>: <function _to_copy at 0x7776e1d59260>, <OpOverload(op='aten._to_copy', overload='out')>: <function _to_copy at 0x7776e1d59260>, <OpOverload(op='aten._fused_dropout', overload='default')>: <function _fused_dropout_decomposition at 0x7776e1d58ea0>, <OpOverload(op='aten._fused_dropout', overload='out')>: <function _fused_dropout_decomposition at 0x7776e1d58ea0>, <OpOverload(op='aten.cudnn_batch_norm', overload='out')>: <function cudnn_batch_norm at 0x7776e1d2a5c0>, <OpOverload(op='aten.index_copy_', overload='default')>: <function index_copy_ at 0x7776e1d5ade0>, <OpOverload(op='aten.index_copy_', overload='dimname')>: <function index_copy_ at 0x7776e1d5ade0>, <OpOverload(op='aten.miopen_batch_norm_backward', overload='default')>: <function miopen_batch_norm_backward at 0x7776e1d59bc0>, <OpOverload(op='aten.miopen_batch_norm_backward', overload='out')>: <function miopen_batch_norm_backward at 0x7776e1d59bc0>, <OpOverload(op='aten.cudnn_batch_norm_backward', overload='default')>: <function cudnn_batch_norm_backward at 0x7776e1d5a2a0>, <OpOverload(op='aten.cudnn_batch_norm_backward', overload='out')>: <function cudnn_batch_norm_backward at 0x7776e1d5a2a0>, <OpOverload(op='aten._upsample_nearest_exact3d', overload='out')>: <function _upsample_nearest_exact3d at 0x7776e1d5ae80>, <OpOverload(op='aten.max_unpool2d', overload='default')>: <function max_unpool2d at 0x7776e1d5aa20>, <OpOverload(op='aten.max_unpool2d', overload='out')>: <function max_unpool2d at 0x7776e1d5aa20>, <OpOverload(op='aten.index_add', overload='out')>: <function index_add at 0x7776e1d5b060>, <OpOverload(op='aten.index_add', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296ad40>, kernel=<OpOverload(op='aten.index_add', overload='dimname')>), <OpOverload(op='aten.pad_sequence', overload='default')>: <function pad_sequence at 0x7776e1d5ac00>, <OpOverload(op='aten.index_add', overload='default')>: <function index_add at 0x7776e1d5b060>, <OpOverload(op='aten.max_unpool3d', overload='default')>: <function max_unpool3d at 0x7776e1d5aca0>, <OpOverload(op='aten.max_unpool3d', overload='out')>: <function max_unpool3d at 0x7776e1d5aca0>, <OpOverload(op='aten.index_add_', overload='default')>: <function index_add_ at 0x7776e1d5aac0>, <OpOverload(op='aten._upsample_nearest_exact3d', overload='default')>: <function _upsample_nearest_exact3d at 0x7776e1d5ae80>, <OpOverload(op='aten.index_copy', overload='default')>: <function index_copy at 0x7776e1d58220>, <OpOverload(op='aten.index_copy', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632969300>, kernel=<OpOverload(op='aten.index_copy', overload='dimname')>), <OpOverload(op='aten.index_copy', overload='out')>: <function index_copy at 0x7776e1d58220>, <OpOverload(op='aten.log_sigmoid_forward', overload='default')>: <function log_sigmoid_forward at 0x7776e1d5b6a0>, <OpOverload(op='aten.log_sigmoid_forward', overload='output')>: <function log_sigmoid_forward at 0x7776e1d5b6a0>, <OpOverload(op='aten._upsample_nearest_exact1d', overload='vec')>: <function _upsample_nearest_exact_vec at 0x7776e1d78680>, <OpOverload(op='aten.uniform_', overload='default')>: <function uniform_ at 0x7776e1d5b880>, <OpOverload(op='aten.upsample_nearest1d', overload='default')>: <function upsample_nearest1d at 0x7776e1d78a40>, <OpOverload(op='aten._upsample_nearest_exact1d', overload='default')>: <function upsample_nearest_exact1d at 0x7776e1d78d60>, <OpOverload(op='aten.upsample_nearest3d', overload='vec')>: <function _upsample_nearest_vec at 0x7776e1d78220>, <OpOverload(op='aten.rnn_tanh', overload='data')>: <function rnn_tanh_data at 0x7776e1d79d00>, <OpOverload(op='aten.upsample_nearest1d', overload='out')>: <function upsample_nearest1d at 0x7776e1d78a40>, <OpOverload(op='aten._upsample_nearest_exact2d', overload='vec')>: <function _upsample_nearest_exact_vec at 0x7776e1d78680>, <OpOverload(op='aten.upsample_nearest2d', overload='default')>: <function upsample_nearest2d at 0x7776e1d79080>, <OpOverload(op='aten._upsample_nearest_exact3d', overload='vec')>: <function _upsample_nearest_exact_vec at 0x7776e1d78680>, <OpOverload(op='aten._upsample_nearest_exact1d', overload='out')>: <function upsample_nearest_exact1d at 0x7776e1d78d60>, <OpOverload(op='aten.rnn_relu', overload='input')>: <function rnn_relu_input at 0x7776e1d79940>, <OpOverload(op='aten._upsample_nearest_exact2d', overload='default')>: <function _upsample_nearest_exact2d at 0x7776e1d793a0>, <OpOverload(op='aten.rnn_relu', overload='data')>: <function rnn_relu_data at 0x7776e1d79b20>, <OpOverload(op='aten.upsample_nearest2d', overload='out')>: <function upsample_nearest2d at 0x7776e1d79080>, <OpOverload(op='aten._upsample_nearest_exact2d', overload='out')>: <function _upsample_nearest_exact2d at 0x7776e1d793a0>, <OpOverload(op='aten.upsample_nearest3d', overload='default')>: <function upsample_nearest3d at 0x7776e1d5bd80>, <OpOverload(op='aten.upsample_nearest3d', overload='out')>: <function upsample_nearest3d at 0x7776e1d5bd80>, <OpOverload(op='aten.lstm', overload='input')>: <function lstm_impl at 0x7776e1d7a160>, <OpOverload(op='aten.lstm', overload='data')>: <function lstm_data_impl at 0x7776e1d7a340>, <OpOverload(op='aten._upsample_bilinear2d_aa', overload='vec')>: <function upsample_bilinear2d_aa_vec at 0x7776e1d7aa20>, <OpOverload(op='aten.gru', overload='data')>: <function gru_impl_data at 0x7776e1d7a660>, <OpOverload(op='aten.gru', overload='input')>: <function gru_impl at 0x7776e1d7a840>, <OpOverload(op='aten._upsample_bicubic2d_aa', overload='vec')>: <function upsample_bicubic2d_aa_vec at 0x7776e1d7ac00>, <OpOverload(op='aten.upsample_trilinear3d', overload='out')>: <function upsample_trilinear3d at 0x7776e1d7b880>, <OpOverload(op='aten.upsample_linear1d', overload='out')>: <function upsample_linear1d at 0x7776e1d7b2e0>, <OpOverload(op='aten.upsample_bilinear2d', overload='default')>: <function upsample_bilinear2d at 0x7776e1d7b600>, <OpOverload(op='aten.upsample_bilinear2d', overload='out')>: <function upsample_bilinear2d at 0x7776e1d7b600>, <OpOverload(op='aten.upsample_trilinear3d', overload='vec')>: <function _upsample_linear_vec at 0x7776e1d7b100>, <OpOverload(op='aten.nll_loss_forward', overload='default')>: <function nll_loss_forward at 0x7776e1d7a7a0>, <OpOverload(op='aten.upsample_trilinear3d', overload='default')>: <function upsample_trilinear3d at 0x7776e1d7b880>, <OpOverload(op='aten.nll_loss_forward', overload='output')>: <function nll_loss_forward at 0x7776e1d7a7a0>, <OpOverload(op='aten.is_same_size', overload='default')>: <function is_same_size at 0x7776e1d7bc40>, <OpOverload(op='aten._reshape_alias', overload='default')>: <function _reshape_alias at 0x7776e1d7bf60>, <OpOverload(op='aten._unsafe_view', overload='out')>: <function _reshape_alias at 0x7776e1d7bf60>, <OpOverload(op='aten._unsafe_index', overload='Tensor')>: <function _unsafe_index at 0x7776e1d7be20>, <OpOverload(op='aten._unsafe_masked_index', overload='default')>: <function _unsafe_masked_index at 0x7776e1d94220>, <OpOverload(op='aten._unsafe_masked_index_put_accumulate', overload='default')>: <function _unsafe_masked_index_put_accumulate at 0x7776e1d94360>, <OpOverload(op='aten.nll_loss2d_forward', overload='default')>: <function nll_loss2d_forward at 0x7776e1d7a0c0>, <OpOverload(op='aten.affine_grid_generator', overload='default')>: <function affine_grid_generator at 0x7776e1d94ea0>, <OpOverload(op='aten.affine_grid_generator', overload='out')>: <function affine_grid_generator at 0x7776e1d94ea0>, <OpOverload(op='aten.grid_sampler_2d', overload='default')>: <function grid_sampler_2d at 0x7776e1d95260>, <OpOverload(op='aten.grid_sampler_2d', overload='out')>: <function grid_sampler_2d at 0x7776e1d95260>, <OpOverload(op='aten.matmul', overload='default')>: <function matmul at 0x7776e1d95bc0>, <OpOverload(op='aten.mv', overload='default')>: <function mv at 0x7776e1d95580>, <OpOverload(op='aten.upsample_bicubic2d', overload='vec')>: <function upsample_bicubic2d_vec at 0x7776e1d963e0>, <OpOverload(op='aten.binary_cross_entropy_with_logits', overload='default')>: <function binary_cross_entropy_with_logits at 0x7776e1d95800>, <OpOverload(op='aten.binary_cross_entropy_with_logits', overload='out')>: <function binary_cross_entropy_with_logits at 0x7776e1d95800>, <OpOverload(op='aten.upsample_bicubic2d', overload='default')>: <function upsample_bicubic2d_default at 0x7776e1d95f80>, <OpOverload(op='aten.upsample_bicubic2d', overload='out')>: <function upsample_bicubic2d_default at 0x7776e1d95f80>, <OpOverload(op='aten.reflection_pad1d', overload='default')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.reflection_pad1d', overload='out')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.replication_pad1d', overload='default')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.__ilshift__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdda0>, <OpOverload(op='aten.reflection_pad2d', overload='default')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.reflection_pad3d', overload='default')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.reflection_pad3d', overload='out')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.reflection_pad2d', overload='out')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.replication_pad2d', overload='default')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.replication_pad1d', overload='out')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.replication_pad2d', overload='out')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.replication_pad3d', overload='default')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.replication_pad3d', overload='out')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.reflection_pad1d_backward', overload='default')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.reflection_pad2d_backward', overload='default')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.reflection_pad1d_backward', overload='grad_input')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.reflection_pad3d_backward', overload='default')>: <function _reflection_pad_backward at 0x7776e1d96840>, <OpOverload(op='aten._scaled_dot_product_flash_attention_for_cpu', overload='default')>: <function scaled_dot_product_flash_attention_for_cpu at 0x7776e1d97c40>, <OpOverload(op='aten.arange', overload='out')>: <function arange_default at 0x7776e1d976a0>, <OpOverload(op='aten.aminmax', overload='default')>: <function aminmax at 0x7776e1d97060>, <OpOverload(op='aten.aminmax', overload='out')>: <function aminmax at 0x7776e1d97060>, <OpOverload(op='aten.multilabel_margin_loss_forward', overload='output')>: <function multilabel_margin_loss_forward at 0x7776e1d97880>, <OpOverload(op='aten.nansum', overload='default')>: <function nansum at 0x7776e1d97420>, <OpOverload(op='aten.nansum', overload='out')>: <function nansum at 0x7776e1d97420>, <OpOverload(op='aten.multilabel_margin_loss_forward', overload='default')>: <function multilabel_margin_loss_forward at 0x7776e1d97880>, <OpOverload(op='aten.multi_margin_loss', overload='default')>: <function multi_margin_loss at 0x7776e1d97ce0>, <OpOverload(op='aten.multi_margin_loss', overload='out')>: <function multi_margin_loss at 0x7776e1d97ce0>, <OpOverload(op='aten.baddbmm', overload='dtype_out')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.baddbmm', overload='default')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.baddbmm', overload='out')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.baddbmm', overload='dtype')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.__irshift__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbde40>, <OpOverload(op='aten.floor_divide', overload='default')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.floor_divide', overload='Scalar')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.floor_divide', overload='out')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.floor_divide', overload='Scalar_out')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.__irshift__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbde40>, <OpOverload(op='aten.bernoulli', overload='default')>: <function bernoulli at 0x7776e1dbc9a0>, <OpOverload(op='aten._weight_norm_interface', overload='default')>: <function _weight_norm_interface at 0x7776e1dbc720>, <OpOverload(op='aten._weight_norm_interface', overload='out')>: <function _weight_norm_interface at 0x7776e1dbc720>, <OpOverload(op='aten.isin', overload='Tensor_Tensor')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Tensor_Scalar')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Scalar_Tensor')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Tensor_Tensor_out')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Tensor_Scalar_out')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Scalar_Tensor_out')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.take', overload='default')>: <function take at 0x7776e1dbd080>, <OpOverload(op='aten.take', overload='out')>: <function take at 0x7776e1dbd080>, <OpOverload(op='aten.resize_as', overload='default')>: <function resize_as at 0x7776e1dbcea0>, <OpOverload(op='aten.resize_as', overload='out')>: <function resize_as at 0x7776e1dbcea0>, <OpOverload(op='aten.__ior__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdee0>, <OpOverload(op='aten.__ior__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdee0>, <OpOverload(op='aten.addbmm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd120>, <OpOverload(op='aten.addmm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd260>, <OpOverload(op='aten.addmv_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd3a0>, <OpOverload(op='aten.baddbmm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd4e0>, <OpOverload(op='aten.fill_', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd620>, <OpOverload(op='aten.fill_', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd620>, <OpOverload(op='aten.gelu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd760>, <OpOverload(op='aten.index_reduce_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1d977e0>, <OpOverload(op='aten.hardswish_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd8a0>, <OpOverload(op='aten.hardtanh_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd9e0>, <OpOverload(op='aten.hardsigmoid_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdb20>, <OpOverload(op='aten.__iand__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdc60>, <OpOverload(op='aten.__iand__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdc60>, <OpOverload(op='aten.__ilshift__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdda0>, <OpOverload(op='aten.__ixor__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdbc0>, <OpOverload(op='aten.__ixor__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdbc0>, <OpOverload(op='aten.leaky_relu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd940>, <OpOverload(op='aten.logit_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd6c0>, <OpOverload(op='aten.renorm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd1c0>, <OpOverload(op='aten.round_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbccc0>, <OpOverload(op='aten.round_', overload='decimals')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbccc0>, <OpOverload(op='aten.ne', overload='Scalar_out')>: <function ne at 0x7776e1caed40>, <OpOverload(op='aten.ne', overload='Scalar')>: <function ne at 0x7776e1caed40>, <OpOverload(op='aten.ne', overload='Tensor_out')>: <function ne at 0x7776e1caed40>, <OpOverload(op='aten.nextafter', overload='out')>: <function nextafter at 0x7776e1caf1a0>, <OpOverload(op='aten.nextafter', overload='default')>: <function nextafter at 0x7776e1caf1a0>, <OpOverload(op='aten.pow', overload='Tensor_Tensor')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Tensor_Scalar')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Scalar')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Tensor_Scalar_out')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Scalar_out')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Tensor_Tensor_out')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.remainder', overload='Tensor')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Scalar_Tensor')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Scalar')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.full', overload='default')>: <function full at 0x7776e1b12de0>, <OpOverload(op='aten.remainder', overload='Scalar_Tensor_out')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Scalar_out')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Tensor_out')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.sub', overload='Scalar')>: <function sub at 0x7776e1caede0>, <OpOverload(op='aten.sub', overload='Tensor')>: <function sub at 0x7776e1caede0>, <OpOverload(op='aten.upsample_linear1d', overload='default')>: <function upsample_linear1d at 0x7776e1d7b2e0>, <OpOverload(op='aten.sub', overload='out')>: <function sub at 0x7776e1caede0>, <OpOverload(op='aten.sub', overload='Scalar_out')>: <function sub at 0x7776e1caede0>, <OpOverload(op='aten.reflection_pad3d_backward', overload='grad_input')>: <function _reflection_pad_backward at 0x7776e1d96840>, <OpOverload(op='aten.unfold_backward', overload='default')>: <function unfold_backward at 0x7776e1cfdd00>, <OpOverload(op='aten.svd', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e980>, kernel=<OpOverload(op='aten.svd', overload='default')>), <OpOverload(op='aten._addmm_activation', overload='out')>: <function _addmm_activation at 0x7776e1d2aac0>, <OpOverload(op='aten.squeeze', overload='default')>: <function squeeze_default at 0x7776e1dbc5e0>, <OpOverload(op='aten.squeeze', overload='dim')>: <function squeeze_default at 0x7776e1dbc5e0>, <OpOverload(op='aten.squeeze', overload='dims')>: <function squeeze at 0x7776e1af7ce0>, <OpOverload(op='aten.transpose', overload='int')>: <function transpose at 0x7776e1b10f40>, <OpOverload(op='aten.elu_backward', overload='grad_input')>: <function elu_backward at 0x7776e1cddf80>, <OpOverload(op='aten.as_strided_scatter', overload='out')>: <function as_strided_scatter at 0x7776e1af40e0>, <OpOverload(op='aten.fft_ifftn', overload='out')>: <function ifftn at 0x7776e19d8b80>, <OpOverload(op='aten.cat', overload='names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367ebd80>, kernel=<OpOverload(op='aten.cat', overload='names')>), <OpOverload(op='aten.cat', overload='default')>: <function cat at 0x7776e1af4720>, <OpOverload(op='aten.cat', overload='names_out')>: <function cat at 0x7776e1af4720>, <OpOverload(op='aten.cat', overload='out')>: <function cat at 0x7776e1af4720>, <OpOverload(op='aten.ones', overload='default')>: <function ones at 0x7776e1b10e00>, <OpOverload(op='aten.std_mean', overload='correction')>: <function std_mean at 0x7776e1c656c0>, <OpOverload(op='aten.where', overload='self')>: <function where at 0x7776e1ad1760>, <OpOverload(op='aten.where', overload='ScalarOther')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496ccc0>, kernel=<OpOverload(op='aten.where', overload='ScalarOther')>), <OpOverload(op='aten.where', overload='ScalarSelf')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496ce00>, kernel=<OpOverload(op='aten.where', overload='ScalarSelf')>), <OpOverload(op='aten.where', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632990400>, kernel=<OpOverload(op='aten.where', overload='Scalar')>), <OpOverload(op='aten.where', overload='self_out')>: <function where at 0x7776e1ad1760>, <OpOverload(op='aten.where', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e700>, kernel=<OpOverload(op='aten.where', overload='default')>), <OpOverload(op='aten.arange', overload='start')>: <function arange_start at 0x7776e1d974c0>, <OpOverload(op='aten.soft_margin_loss', overload='out')>: <function soft_margin_loss at 0x7776e1cff560>, <OpOverload(op='aten.upsample_linear1d', overload='vec')>: <function _upsample_linear_vec at 0x7776e1d7b100>, <OpOverload(op='aten.nll_loss_backward', overload='default')>: <function nll_loss_backward at 0x7776e1cfe5c0>, <OpOverload(op='aten.sum', overload='dim_IntList')>: <function sum at 0x7776e1ad22a0>, <OpOverload(op='aten.sum', overload='default')>: <function sum_default at 0x7776e1dbc400>, <OpOverload(op='aten.sum', overload='IntList_out')>: <function sum at 0x7776e1ad22a0>, <OpOverload(op='aten.sum', overload='dim_DimnameList')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763496e0c0>, kernel=<OpOverload(op='aten.sum', overload='dim_DimnameList')>), <OpOverload(op='aten.relu6', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329577e0>, kernel=<OpOverload(op='aten.relu6', overload='default')>), <OpOverload(op='aten.soft_margin_loss_backward', overload='grad_input')>: <function soft_margin_loss_backward at 0x7776e1cff4c0>, <OpOverload(op='aten.prod', overload='dim_int')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='default')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='dim_Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d440>, kernel=<OpOverload(op='aten.prod', overload='dim_Dimname')>), <OpOverload(op='aten.prod', overload='int_out')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='Dimname_out')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='out')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.var', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573efc0>, kernel=<OpOverload(op='aten.var', overload='default')>), <OpOverload(op='aten.var', overload='out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='correction')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d260>, kernel=<OpOverload(op='aten.var', overload='dim')>), <OpOverload(op='aten.normal', overload='float_float_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.var', overload='correction_out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d88c360>, kernel=<OpOverload(op='aten.var', overload='names_dim')>), <OpOverload(op='aten.var', overload='names_out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='correction_names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776defd8900>, kernel=<OpOverload(op='aten.var', overload='correction_names')>), <OpOverload(op='aten.var', overload='correction_names_out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.amax', overload='out')>: <function amax at 0x7776e1ad2840>, <OpOverload(op='aten.nll_loss2d_forward', overload='output')>: <function nll_loss2d_forward at 0x7776e1d7a0c0>, <OpOverload(op='aten.amax', overload='default')>: <function amax at 0x7776e1ad2840>, <OpOverload(op='aten.amin', overload='default')>: <function amin at 0x7776e1ad2700>, <OpOverload(op='aten.amin', overload='out')>: <function amin at 0x7776e1ad2700>, <OpOverload(op='aten.im2col', overload='default')>: <function im2col at 0x7776e1d28c20>, <OpOverload(op='aten.lift', overload='default')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.empty_strided', overload='out')>: <function empty_strided at 0x7776e1b13d80>, <OpOverload(op='aten.reflection_pad2d_backward', overload='grad_input')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.full', overload='out')>: <function full at 0x7776e1b12de0>, <OpOverload(op='aten.arange', overload='default')>: <function arange_default at 0x7776e1d976a0>, <OpOverload(op='aten._adaptive_avg_pool2d', overload='out')>: <function adaptive_avg_pool2d at 0x7776e1d5a700>, <OpOverload(op='aten.diagonal_backward', overload='out')>: <function diagonal_backward at 0x7776e1d28180>, <OpOverload(op='aten.normal', overload='Tensor_float')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.native_layer_norm_backward', overload='out')>: <function native_layer_norm_backward_out at 0x7776e1d2b100>, <OpOverload(op='aten.normal', overload='Tensor_float_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='Tensor_Tensor')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='float_Tensor_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='float_Tensor')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='Tensor_Tensor_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.uniform', overload='default')>: <function uniform at 0x7776e1d5ba60>, <OpOverload(op='aten.uniform', overload='out')>: <function uniform at 0x7776e1d5ba60>, <OpOverload(op='aten.normal', overload='float_float')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.mv', overload='out')>: <function mv at 0x7776e1d95580>, <OpOverload(op='aten.index_fill_', overload='int_Tensor')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.upsample_nearest1d', overload='vec')>: <function _upsample_nearest_vec at 0x7776e1d78220>, <OpOverload(op='aten.frexp', overload='Tensor')>: <function frexp at 0x7776e1c95f80>, <OpOverload(op='aten.isfinite', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367eb9c0>, kernel=<OpOverload(op='aten.isfinite', overload='default')>), <OpOverload(op='aten.div', overload='Tensor_mode')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.lgamma', overload='out')>: <function lgamma at 0x7776e1c4fd80>, <OpOverload(op='aten.log', overload='default')>: <function log at 0x7776e1c642c0>, <OpOverload(op='aten.log', overload='out')>: <function log at 0x7776e1c642c0>, <OpOverload(op='aten.bitwise_xor', overload='Tensor_out')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.log1p', overload='out')>: <function log1p at 0x7776e1c647c0>, <OpOverload(op='aten.log1p', overload='default')>: <function log1p at 0x7776e1c647c0>, <OpOverload(op='aten._native_batch_norm_legit_no_training', overload='default')>: <function _native_batch_norm_legit_no_training at 0x7776e1d2be20>, <OpOverload(op='aten.log2', overload='default')>: <function log2 at 0x7776e1c64cc0>, <OpOverload(op='aten.log10', overload='default')>: <function log10 at 0x7776e1c651c0>, <OpOverload(op='aten.log10', overload='out')>: <function log10 at 0x7776e1c651c0>, <OpOverload(op='aten.rsqrt', overload='default')>: <function rsqrt at 0x7776e1c668e0>, <OpOverload(op='aten.reciprocal', overload='out')>: <function reciprocal at 0x7776e1c66020>, <OpOverload(op='aten.reciprocal', overload='default')>: <function reciprocal at 0x7776e1c66020>, <OpOverload(op='aten.neg', overload='default')>: <function neg at 0x7776e1c659e0>, <OpOverload(op='aten.div', overload='Scalar_mode')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.neg', overload='out')>: <function neg at 0x7776e1c659e0>, <OpOverload(op='aten.log_sigmoid_backward', overload='default')>: <function log_sigmoid_backward at 0x7776e1cfd120>, <OpOverload(op='aten.tan', overload='default')>: <function tan at 0x7776e1c753a0>, <OpOverload(op='aten.tan', overload='out')>: <function tan at 0x7776e1c753a0>, <OpOverload(op='aten.round', overload='decimals')>: <function round at 0x7776e1c663e0>, <OpOverload(op='aten.round', overload='default')>: <function round at 0x7776e1c663e0>, <OpOverload(op='aten.round', overload='decimals_out')>: <function round at 0x7776e1c663e0>, <OpOverload(op='aten.round', overload='out')>: <function round at 0x7776e1c663e0>, <OpOverload(op='aten.tanh_backward', overload='grad_input')>: <function tanh_backward at 0x7776e2122840>, <OpOverload(op='aten.rsqrt', overload='out')>: <function rsqrt at 0x7776e1c668e0>, <OpOverload(op='aten.sign', overload='out')>: <function sign at 0x7776e1c677e0>, <OpOverload(op='aten.signbit', overload='default')>: <function signbit at 0x7776e1c67ce0>, <OpOverload(op='aten.signbit', overload='out')>: <function signbit at 0x7776e1c67ce0>, <OpOverload(op='aten.sin', overload='out')>: <function sin at 0x7776e1c74220>, <OpOverload(op='aten.sin', overload='default')>: <function sin at 0x7776e1c74220>, <OpOverload(op='aten.bitwise_and', overload='Tensor_out')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.sinh', overload='default')>: <function sinh at 0x7776e1c66980>, <OpOverload(op='aten.cudnn_batch_norm', overload='default')>: <function cudnn_batch_norm at 0x7776e1d2a5c0>, <OpOverload(op='aten.sqrt', overload='out')>: <function sqrt at 0x7776e1c749a0>, <OpOverload(op='aten.sqrt', overload='default')>: <function sqrt at 0x7776e1c749a0>, <OpOverload(op='aten.lt', overload='Tensor')>: <function lt at 0x7776e1cadbc0>, <OpOverload(op='aten.tanh', overload='out')>: <function tanh at 0x7776e1c758a0>, <OpOverload(op='aten.tanh', overload='default')>: <function tanh at 0x7776e1c758a0>, <OpOverload(op='aten.trunc', overload='out')>: <function trunc at 0x7776e1c75da0>, <OpOverload(op='aten.trunc', overload='default')>: <function trunc at 0x7776e1c75da0>, <OpOverload(op='aten.add', overload='Scalar')>: <function add at 0x7776e1c762a0>, <OpOverload(op='aten.add', overload='Tensor')>: <function add at 0x7776e1c762a0>, <OpOverload(op='aten.add', overload='out')>: <function add at 0x7776e1c762a0>, <OpOverload(op='aten.add', overload='Scalar_out')>: <function add at 0x7776e1c762a0>, <OpOverload(op='aten.narrow_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.narrow at 0x7776e1b98a40>, <OpOverload(op='aten.atan2', overload='out')>: <function atan2 at 0x7776e1c76700>, <OpOverload(op='aten.atan2', overload='default')>: <function atan2 at 0x7776e1c76700>, <OpOverload(op='aten.bitwise_and', overload='Tensor')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_and', overload='Scalar')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_and', overload='Scalar_out')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_and', overload='Scalar_Tensor_out')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_or', overload='Scalar')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_or', overload='Tensor_out')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_or', overload='Tensor')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_or', overload='Scalar_Tensor_out')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_or', overload='Scalar_out')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_xor', overload='Scalar_Tensor')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.bitwise_xor', overload='Scalar_Tensor_out')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.bitwise_xor', overload='Scalar')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.bitwise_xor', overload='Tensor')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.bitwise_xor', overload='Scalar_out')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.div', overload='out')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='out_mode')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='Scalar_out')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='Scalar_mode_out')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.eq', overload='Tensor')>: <function eq at 0x7776e1c94180>, <OpOverload(op='aten.eq', overload='Scalar')>: <function eq at 0x7776e1c94180>, <OpOverload(op='aten.eq', overload='Scalar_out')>: <function eq at 0x7776e1c94180>, <OpOverload(op='aten.eq', overload='Tensor_out')>: <function eq at 0x7776e1c94180>, <OpOverload(op='aten.fmax', overload='out')>: <function fmax at 0x7776e1c95120>, <OpOverload(op='aten.binary_cross_entropy_backward', overload='default')>: <function binary_cross_entropy_backward at 0x7776e1cfef20>, <OpOverload(op='aten.fmax', overload='default')>: <function fmax at 0x7776e1c95120>, <OpOverload(op='aten.fmin', overload='default')>: <function fmin at 0x7776e1c95580>, <OpOverload(op='aten.fmin', overload='out')>: <function fmin at 0x7776e1c95580>, <OpOverload(op='aten.fmod', overload='Scalar')>: <function fmod at 0x7776e1c959e0>, <OpOverload(op='aten.log10_', overload='default')>: <function log10 at 0x7776e1b67060>, <OpOverload(op='aten.fmod', overload='Tensor')>: <function fmod at 0x7776e1c959e0>, <OpOverload(op='aten.fmod', overload='Tensor_out')>: <function fmod at 0x7776e1c959e0>, <OpOverload(op='aten.igammac', overload='out')>: <function igammac at 0x7776e1c979c0>, <OpOverload(op='aten.fmod', overload='Scalar_out')>: <function fmod at 0x7776e1c959e0>, <OpOverload(op='aten.gcd', overload='default')>: <function gcd at 0x7776e1c76200>, <OpOverload(op='aten.gcd', overload='out')>: <function gcd at 0x7776e1c76200>, <OpOverload(op='aten.addmv', overload='out')>: <function addmv at 0x7776e1d2a0c0>, <OpOverload(op='aten.ge', overload='Tensor')>: <function ge at 0x7776e1c963e0>, <OpOverload(op='aten.ge', overload='Scalar')>: <function ge at 0x7776e1c963e0>, <OpOverload(op='aten.ge', overload='Tensor_out')>: <function ge at 0x7776e1c963e0>, <OpOverload(op='aten.xlogy', overload='Tensor')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.ge', overload='Scalar_out')>: <function ge at 0x7776e1c963e0>, <OpOverload(op='aten.gt', overload='Scalar')>: <function gt at 0x7776e1c96840>, <OpOverload(op='aten.gt', overload='Tensor')>: <function gt at 0x7776e1c96840>, <OpOverload(op='aten.hardswish_backward', overload='out')>: <function hardswish_backward at 0x7776e1cde200>, <OpOverload(op='aten.gt', overload='Scalar_out')>: <function gt at 0x7776e1c96840>, <OpOverload(op='aten.gt', overload='Tensor_out')>: <function gt at 0x7776e1c96840>, <OpOverload(op='aten.hypot', overload='default')>: <function hypot at 0x7776e1c97100>, <OpOverload(op='aten.hypot', overload='out')>: <function hypot at 0x7776e1c97100>, <OpOverload(op='aten.index_fill', overload='Dimname_Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329680e0>, kernel=<OpOverload(op='aten.index_fill', overload='Dimname_Scalar')>), <OpOverload(op='aten.igamma', overload='out')>: <function igamma at 0x7776e1c97560>, <OpOverload(op='aten.igamma', overload='default')>: <function igamma at 0x7776e1c97560>, <OpOverload(op='aten.sum', overload='out')>: <function sum_default at 0x7776e1dbc400>, <OpOverload(op='aten.igammac', overload='default')>: <function igammac at 0x7776e1c979c0>, <OpOverload(op='aten.le', overload='Tensor')>: <function le at 0x7776e1cac400>, <OpOverload(op='aten.le', overload='Scalar')>: <function le at 0x7776e1cac400>, <OpOverload(op='aten.le', overload='Tensor_out')>: <function le at 0x7776e1cac400>, <OpOverload(op='aten.le', overload='Scalar_out')>: <function le at 0x7776e1cac400>, <OpOverload(op='aten.lt', overload='Scalar')>: <function lt at 0x7776e1cadbc0>, <OpOverload(op='aten.lt', overload='Scalar_out')>: <function lt at 0x7776e1cadbc0>, <OpOverload(op='aten.lt', overload='Tensor_out')>: <function lt at 0x7776e1cadbc0>, <OpOverload(op='aten.maximum', overload='default')>: <function maximum at 0x7776e1cae020>, <OpOverload(op='aten.maximum', overload='out')>: <function maximum at 0x7776e1cae020>, <OpOverload(op='aten.minimum', overload='default')>: <function minimum at 0x7776e1cae480>, <OpOverload(op='aten.minimum', overload='out')>: <function minimum at 0x7776e1cae480>, <OpOverload(op='aten.mul', overload='Scalar_out')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.rnn_tanh', overload='input')>: <function rnn_tanh_input at 0x7776e1d79760>, <OpOverload(op='aten.ne', overload='Tensor')>: <function ne at 0x7776e1caed40>, <OpOverload(op='aten.acosh', overload='out')>: <function acosh at 0x7776e1c3c400>, <OpOverload(op='aten.acosh', overload='default')>: <function acosh at 0x7776e1c3c400>, <OpOverload(op='aten.asin', overload='out')>: <function asin at 0x7776e1c3c900>, <OpOverload(op='aten.asin', overload='default')>: <function asin at 0x7776e1c3c900>, <OpOverload(op='aten.asinh', overload='default')>: <function asinh at 0x7776e1c3ce00>, <OpOverload(op='aten.asinh', overload='out')>: <function asinh at 0x7776e1c3ce00>, <OpOverload(op='aten.lgamma', overload='default')>: <function lgamma at 0x7776e1c4fd80>, <OpOverload(op='aten.atan', overload='default')>: <function atan at 0x7776e1c3d300>, <OpOverload(op='aten.atanh', overload='out')>: <function atanh at 0x7776e1c3d800>, <OpOverload(op='aten.atan', overload='out')>: <function atan at 0x7776e1c3d300>, <OpOverload(op='aten.atanh', overload='default')>: <function atanh at 0x7776e1c3d800>, <OpOverload(op='aten.cos', overload='default')>: <function cos at 0x7776e1c3eac0>, <OpOverload(op='aten.cos', overload='out')>: <function cos at 0x7776e1c3eac0>, <OpOverload(op='aten.cosh', overload='out')>: <function cosh at 0x7776e1c3efc0>, <OpOverload(op='aten.cosh', overload='default')>: <function cosh at 0x7776e1c3efc0>, <OpOverload(op='aten.bitwise_and', overload='Scalar_Tensor')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_not', overload='out')>: <function bitwise_not at 0x7776e1c3dd00>, <OpOverload(op='aten.bitwise_not', overload='default')>: <function bitwise_not at 0x7776e1c3dd00>, <OpOverload(op='aten.ceil', overload='default')>: <function ceil at 0x7776e1c3e200>, <OpOverload(op='aten.index_fill_', overload='int_Scalar')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.clone', overload='default')>: <function clone at 0x7776e1ad19e0>, <OpOverload(op='aten.ceil', overload='out')>: <function ceil at 0x7776e1c3e200>, <OpOverload(op='aten.conj_physical', overload='out')>: <function conj_physical at 0x7776e1c3e5c0>, <OpOverload(op='aten.conj_physical', overload='default')>: <function conj_physical at 0x7776e1c3e5c0>, <OpOverload(op='aten.digamma', overload='default')>: <function digamma at 0x7776e1c3f4c0>, <OpOverload(op='aten.bitwise_or', overload='Scalar_Tensor')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.digamma', overload='out')>: <function digamma at 0x7776e1c3f4c0>, <OpOverload(op='aten.sinh', overload='out')>: <function sinh at 0x7776e1c66980>, <OpOverload(op='aten.erf', overload='default')>: <function erf at 0x7776e1c3d3a0>, <OpOverload(op='aten.erf', overload='out')>: <function erf at 0x7776e1c3d3a0>, <OpOverload(op='aten.erfc', overload='default')>: <function erfc at 0x7776e1c3fce0>, <OpOverload(op='aten.sign', overload='default')>: <function sign at 0x7776e1c677e0>, <OpOverload(op='aten.exp', overload='default')>: <function exp at 0x7776e1c4c220>, <OpOverload(op='aten.exp', overload='out')>: <function exp at 0x7776e1c4c220>, <OpOverload(op='aten.expm1', overload='out')>: <function expm1 at 0x7776e1c4c720>, <OpOverload(op='aten.expm1', overload='default')>: <function expm1 at 0x7776e1c4c720>, <OpOverload(op='aten.exp2', overload='default')>: <function exp2 at 0x7776e1c4cc20>, <OpOverload(op='aten.exp2', overload='out')>: <function exp2 at 0x7776e1c4cc20>, <OpOverload(op='aten.fill', overload='Scalar')>: <function fill_scalar at 0x7776e1cdea20>, <OpOverload(op='aten.transpose', overload='Dimname')>: <function transpose at 0x7776e1b10f40>, <OpOverload(op='aten.mul', overload='out')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.fill', overload='Tensor')>: <function fill_tensor at 0x7776e1cdeac0>, <OpOverload(op='aten.floor', overload='out')>: <function floor at 0x7776e1c4d760>, <OpOverload(op='aten.floor', overload='default')>: <function floor at 0x7776e1c4d760>, <OpOverload(op='aten.clone', overload='out')>: <function clone at 0x7776e1ad19e0>, <OpOverload(op='aten.abs', overload='default')>: <function abs at 0x7776e1dbede0>, <OpOverload(op='aten.acos', overload='out')>: <function acos at 0x7776e1dbfd80>, <OpOverload(op='aten.acos', overload='default')>: <function acos at 0x7776e1dbfd80>, <OpOverload(op='aten.log2', overload='out')>: <function log2 at 0x7776e1c64cc0>, <OpOverload(op='aten.rad2deg', overload='default')>: <function rad2deg at 0x7776e1b431a0>, <OpOverload(op='aten.linear', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573fa60>, kernel=<OpOverload(op='aten.linear', overload='default')>), <OpOverload(op='aten.conv2d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763573d120>, kernel=<OpOverload(op='aten.conv2d', overload='default')>), <OpOverload(op='aten.mul', overload='Tensor')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.mul', overload='Scalar')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.div', overload='Tensor')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='Scalar')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.upsample_nearest2d', overload='vec')>: <function _upsample_nearest_vec at 0x7776e1d78220>, <OpOverload(op='aten.tanh_backward', overload='default')>: <function tanh_backward at 0x7776e2122840>, <OpOverload(op='aten.abs', overload='out')>: <function abs at 0x7776e1dbede0>, <OpOverload(op='aten.sym_numel', overload='default')>: <function sym_numel at 0x7776e1d96340>, <OpOverload(op='aten.slice', overload='Tensor')>: <function slice_forward at 0x7776e1cff600>, <OpOverload(op='aten.diagonal_scatter', overload='default')>: <function diagonal_scatter at 0x7776e1b104a0>, <OpOverload(op='aten.diagonal', overload='default')>: <function diagonal at 0x7776e1b102c0>, <OpOverload(op='aten.select_scatter', overload='default')>: <function select_scatter at 0x7776e1b43ec0>, <OpOverload(op='aten.upsample_bilinear2d', overload='vec')>: <function _upsample_linear_vec at 0x7776e1d7b100>, <OpOverload(op='aten.zeros', overload='default')>: <function zeros at 0x7776e1b11a80>, <OpOverload(op='aten.detach', overload='default')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.as_strided_scatter', overload='default')>: <function as_strided_scatter at 0x7776e1af40e0>, <OpOverload(op='aten.empty_strided', overload='default')>: <function empty_strided at 0x7776e1b13d80>, <OpOverload(op='aten.view', overload='default')>: <function view at 0x7776e1b118a0>, <OpOverload(op='aten._unsafe_view', overload='default')>: <function _reshape_alias at 0x7776e1d7bf60>, <OpOverload(op='aten.lift_fresh', overload='default')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.empty_like', overload='default')>: <function empty_like at 0x7776e1b12c00>, <OpOverload(op='aten.ones_like', overload='out')>: <function ones_like at 0x7776e1b40180>, <OpOverload(op='aten.zeros_like', overload='default')>: <function zeros_like at 0x7776e1b116c0>, <OpOverload(op='aten.zeros_like', overload='out')>: <function zeros_like at 0x7776e1b116c0>, <OpOverload(op='aten.new_empty', overload='default')>: <function new_empty at 0x7776e1b12480>, <OpOverload(op='aten.new_empty', overload='out')>: <function new_empty at 0x7776e1b12480>, <OpOverload(op='aten.new_empty_strided', overload='out')>: <function new_empty_strided at 0x7776e1b12700>, <OpOverload(op='aten.new_full', overload='default')>: <function new_full at 0x7776e1b12840>, <OpOverload(op='aten.new_full', overload='out')>: <function new_full at 0x7776e1b12840>, <OpOverload(op='aten.new_zeros', overload='default')>: <function new_zeros at 0x7776e1b11440>, <OpOverload(op='aten.new_zeros', overload='out')>: <function new_zeros at 0x7776e1b11440>, <OpOverload(op='aten.new_ones', overload='default')>: <function new_ones at 0x7776e1b10400>, <OpOverload(op='aten.new_ones', overload='out')>: <function new_ones at 0x7776e1b10400>, <OpOverload(op='aten.view_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.item', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776329927a0>, kernel=<OpOverload(op='aten.item', overload='default')>), <OpOverload(op='aten.nonzero_numpy', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x777632991620>, kernel=<OpOverload(op='aten.nonzero_numpy', overload='default')>), <OpOverload(op='aten._unsafe_index_put', overload='default')>: <function _unsafe_index_put at 0x7776e1d940e0>, <OpOverload(op='aten.slice_scatter', overload='default')>: <function slice_scatter at 0x7776e1cffc40>, <OpOverload(op='aten.slice_scatter', overload='out')>: <function slice_scatter at 0x7776e1cffc40>, <OpOverload(op='aten.index_put_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1d96ac0>, <OpOverload(op='aten.erfc', overload='out')>: <function erfc at 0x7776e1c3fce0>, <OpOverload(op='aten.new_empty_strided', overload='default')>: <function new_empty_strided at 0x7776e1b12700>, <OpOverload(op='aten.ones_like', overload='default')>: <function ones_like at 0x7776e1b40180>, <OpOverload(op='aten.empty_like', overload='out')>: <function empty_like at 0x7776e1b12c00>, <OpOverload(op='quantized.make_quantized_cell_params_fp16', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e99e0>, kernel=<OpOverload(op='quantized.make_quantized_cell_params_fp16', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_stride', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9800>, kernel=<OpOverload(op='quantized.conv_transpose3d_stride', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_groups', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d733560>, kernel=<OpOverload(op='quantized.conv_transpose3d_groups', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_stride', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c2ab4c0>, kernel=<OpOverload(op='quantized.conv_transpose2d_stride', overload='default')>), <OpOverload(op='quantized.conv2d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d733920>, kernel=<OpOverload(op='quantized.conv2d_unpack', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c39be20>, kernel=<OpOverload(op='quantized.conv_transpose3d_padding', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_dilation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de41120>, kernel=<OpOverload(op='quantized.conv_transpose3d_dilation', overload='default')>), <OpOverload(op='quantized.conv2d_dilation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c39b4c0>, kernel=<OpOverload(op='quantized.conv2d_dilation', overload='default')>), <OpOverload(op='quantized.conv2d_output_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c39a480>, kernel=<OpOverload(op='quantized.conv2d_output_padding', overload='default')>), <OpOverload(op='quantized.conv2d_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c3bc4a0>, kernel=<OpOverload(op='quantized.conv2d_padding', overload='default')>), <OpOverload(op='quantized.conv3d_dilation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de41940>, kernel=<OpOverload(op='quantized.conv3d_dilation', overload='default')>), <OpOverload(op='quantized.conv1d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9300>, kernel=<OpOverload(op='quantized.conv1d_unpack', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776e1865260>, kernel=<OpOverload(op='quantized.conv_transpose3d_unpack', overload='default')>), <OpOverload(op='quantized.conv3d_groups', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d733ba0>, kernel=<OpOverload(op='quantized.conv3d_groups', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_output_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9440>, kernel=<OpOverload(op='quantized.conv_transpose3d_output_padding', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_transpose', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c31bba0>, kernel=<OpOverload(op='quantized.conv_transpose3d_transpose', overload='default')>), <OpOverload(op='quantized.linear_unpack_fp16', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de42d40>, kernel=<OpOverload(op='quantized.linear_unpack_fp16', overload='default')>), <OpOverload(op='quantized.conv3d_stride', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c31bd80>, kernel=<OpOverload(op='quantized.conv3d_stride', overload='default')>), <OpOverload(op='quantized.conv2d_groups', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de416c0>, kernel=<OpOverload(op='quantized.conv2d_groups', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_output_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c398fe0>, kernel=<OpOverload(op='quantized.conv_transpose2d_output_padding', overload='default')>), <OpOverload(op='quantized.conv2d_transpose', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x7776367e9a80>, kernel=<OpOverload(op='quantized.conv2d_transpose', overload='default')>), <OpOverload(op='quantized.conv_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c3bdb20>, kernel=<OpOverload(op='quantized.conv_unpack', overload='default')>), <OpOverload(op='quantized.embedding_bag_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c31ad40>, kernel=<OpOverload(op='quantized.embedding_bag_unpack', overload='default')>), <OpOverload(op='quantized.linear_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c31bce0>, kernel=<OpOverload(op='quantized.linear_unpack', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_transpose', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d733a60>, kernel=<OpOverload(op='quantized.conv_transpose2d_transpose', overload='default')>), <OpOverload(op='quantized.conv2d_stride', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de413a0>, kernel=<OpOverload(op='quantized.conv2d_stride', overload='default')>), <OpOverload(op='quantized.conv3d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de422a0>, kernel=<OpOverload(op='quantized.conv3d_unpack', overload='default')>), <OpOverload(op='sparse.qlinear_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de439c0>, kernel=<OpOverload(op='sparse.qlinear_unpack', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_dilation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d733880>, kernel=<OpOverload(op='quantized.conv_transpose2d_dilation', overload='default')>), <OpOverload(op='quantized.conv_transpose1d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c2ab880>, kernel=<OpOverload(op='quantized.conv_transpose1d_unpack', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_groups', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767de40360>, kernel=<OpOverload(op='quantized.conv_transpose2d_groups', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77767d7336a0>, kernel=<OpOverload(op='quantized.conv_transpose2d_padding', overload='default')>), <OpOverload(op='quantized.conv3d_output_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c39af20>, kernel=<OpOverload(op='quantized.conv3d_output_padding', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c398540>, kernel=<OpOverload(op='quantized.conv_transpose2d_unpack', overload='default')>), <OpOverload(op='quantized.conv3d_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c398680>, kernel=<OpOverload(op='quantized.conv3d_padding', overload='default')>), <OpOverload(op='quantized.conv3d_transpose', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c399580>, kernel=<OpOverload(op='quantized.conv3d_transpose', overload='default')>), <OpOverload(op='profiler._record_function_exit', overload='_RecordFunction')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77762c31bf60>, kernel=<OpOverload(op='profiler._record_function_exit', overload='_RecordFunction')>)}
- experimental_experiment.torch_dynamo.get_decomposition_table_onnxscript()[source]¶
Returns the decomposition table used by
torch.onnx.export().The value is:
<<<
import pprint from experimental_experiment.torch_dynamo import get_decomposition_table_onnxscript pprint.pprint(get_decomposition_table_onnxscript())
>>>
{<OpOverload(op='aten._backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5300>, kernel=<OpOverload(op='aten._backward', overload='default')>), <OpOverload(op='aten._test_check_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3560>, kernel=<OpOverload(op='aten._test_check_tensor', overload='default')>), <OpOverload(op='image._is_compiled_against_turbo', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f44a0>, kernel=<OpOverload(op='image._is_compiled_against_turbo', overload='default')>), <OpOverload(op='aten.linalg_svd', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb6a0>, kernel=<OpOverload(op='aten.linalg_svd', overload='default')>), <OpOverload(op='aten.gradient', overload='tensorarrayint')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cbec0>, kernel=<OpOverload(op='aten.gradient', overload='tensorarrayint')>), <OpOverload(op='aten.masked_select_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb880>, kernel=<OpOverload(op='aten.masked_select_backward', overload='default')>), <OpOverload(op='aten._grid_sampler_2d_cpu_fallback_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a55cc20>, kernel=<OpOverload(op='aten._grid_sampler_2d_cpu_fallback_backward', overload='default')>), <OpOverload(op='aten._saturate_weight_to_fp16', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5080>, kernel=<OpOverload(op='aten._saturate_weight_to_fp16', overload='default')>), <OpOverload(op='aten.quantile', overload='scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0d60>, kernel=<OpOverload(op='aten.quantile', overload='scalar')>), <OpOverload(op='aten._pad_enum', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9440>, kernel=<OpOverload(op='aten._pad_enum', overload='default')>), <OpOverload(op='aten.lstm_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b59e0>, kernel=<OpOverload(op='aten.lstm_cell', overload='default')>), <OpOverload(op='aten._weight_norm_differentiable_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4b80>, kernel=<OpOverload(op='aten._weight_norm_differentiable_backward', overload='default')>), <OpOverload(op='aten.linalg_solve_ex', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c16c0>, kernel=<OpOverload(op='aten.linalg_solve_ex', overload='default')>), <OpOverload(op='aten.sparse_csr_tensor', overload='crow_col_value_size')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d19e0>, kernel=<OpOverload(op='aten.sparse_csr_tensor', overload='crow_col_value_size')>), <OpOverload(op='aten.gradient', overload='scalararray')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8680>, kernel=<OpOverload(op='aten.gradient', overload='scalararray')>), <OpOverload(op='aten.retains_grad', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7600>, kernel=<OpOverload(op='aten.retains_grad', overload='default')>), <OpOverload(op='aten.to_sparse_bsc', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5440>, kernel=<OpOverload(op='aten.to_sparse_bsc', overload='default')>), <OpOverload(op='aten.linalg_svdvals', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b77e0>, kernel=<OpOverload(op='aten.linalg_svdvals', overload='default')>), <OpOverload(op='aten.special_gammainc', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4f40>, kernel=<OpOverload(op='aten.special_gammainc', overload='default')>), <OpOverload(op='image.decode_image', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db1a0>, kernel=<OpOverload(op='image.decode_image', overload='default')>), <OpOverload(op='aten._test_serialization_subcmul', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9260>, kernel=<OpOverload(op='aten._test_serialization_subcmul', overload='default')>), <OpOverload(op='aten.fake_quantize_per_tensor_affine', overload='tensor_qparams')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0040>, kernel=<OpOverload(op='aten.fake_quantize_per_tensor_affine', overload='tensor_qparams')>), <OpOverload(op='image.encode_png', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db9c0>, kernel=<OpOverload(op='image.encode_png', overload='default')>), <OpOverload(op='aten.sparse_csc_tensor', overload='ccol_row_value')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7420>, kernel=<OpOverload(op='aten.sparse_csc_tensor', overload='ccol_row_value')>), <OpOverload(op='aten.sparse_bsc_tensor', overload='ccol_row_value')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c98a0>, kernel=<OpOverload(op='aten.sparse_bsc_tensor', overload='ccol_row_value')>), <OpOverload(op='aten.clip', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9e40>, kernel=<OpOverload(op='aten.clip', overload='Tensor')>), <OpOverload(op='aten.scatter_add', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cafc0>, kernel=<OpOverload(op='aten.scatter_add', overload='dimname')>), <OpOverload(op='aten.unflatten_dense_tensors', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8e00>, kernel=<OpOverload(op='aten.unflatten_dense_tensors', overload='default')>), <OpOverload(op='image.decode_webp', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f40e0>, kernel=<OpOverload(op='image.decode_webp', overload='default')>), <OpOverload(op='aten._choose_qparams_per_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c14e0>, kernel=<OpOverload(op='aten._choose_qparams_per_tensor', overload='default')>), <OpOverload(op='aten.thnn_conv2d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8ae0>, kernel=<OpOverload(op='aten.thnn_conv2d', overload='default')>), <OpOverload(op='aten.is_vulkan_available', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3060>, kernel=<OpOverload(op='aten.is_vulkan_available', overload='default')>), <OpOverload(op='aten.concat', overload='names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1f80>, kernel=<OpOverload(op='aten.concat', overload='names')>), <OpOverload(op='aten.sym_is_contiguous', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2ca0>, kernel=<OpOverload(op='aten.sym_is_contiguous', overload='default')>), <OpOverload(op='aten.linalg_matrix_rank', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca2a0>, kernel=<OpOverload(op='aten.linalg_matrix_rank', overload='default')>), <OpOverload(op='aten.linalg_matrix_power', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5800>, kernel=<OpOverload(op='aten.linalg_matrix_power', overload='default')>), <OpOverload(op='aten.is_distributed', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7e20>, kernel=<OpOverload(op='aten.is_distributed', overload='default')>), <OpOverload(op='aten.arctanh', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b42c0>, kernel=<OpOverload(op='aten.arctanh', overload='default')>), <OpOverload(op='aten.linalg_tensorinv', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8fe0>, kernel=<OpOverload(op='aten.linalg_tensorinv', overload='default')>), <OpOverload(op='aten.softmax', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c39c0>, kernel=<OpOverload(op='aten.softmax', overload='Dimname')>), <OpOverload(op='aten.scatter', overload='dimname_value')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3880>, kernel=<OpOverload(op='aten.scatter', overload='dimname_value')>), <OpOverload(op='aten.vander', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2840>, kernel=<OpOverload(op='aten.vander', overload='default')>), <OpOverload(op='aten.cumulative_trapezoid', overload='x')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8720>, kernel=<OpOverload(op='aten.cumulative_trapezoid', overload='x')>), <OpOverload(op='aten.arcsinh', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1120>, kernel=<OpOverload(op='aten.arcsinh', overload='default')>), <OpOverload(op='aten.hstack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c89a0>, kernel=<OpOverload(op='aten.hstack', overload='default')>), <OpOverload(op='aten.linalg_matrix_rank', overload='atol_rtol_float')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cade0>, kernel=<OpOverload(op='aten.linalg_matrix_rank', overload='atol_rtol_float')>), <OpOverload(op='aten.to_sparse', overload='sparse_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8360>, kernel=<OpOverload(op='aten.to_sparse', overload='sparse_dim')>), <OpOverload(op='aten.ger', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8040>, kernel=<OpOverload(op='aten.ger', overload='default')>), <OpOverload(op='aten.is_signed', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0720>, kernel=<OpOverload(op='aten.is_signed', overload='default')>), <OpOverload(op='aten.conv2d', overload='padding')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5da0>, kernel=<OpOverload(op='aten.conv2d', overload='padding')>), <OpOverload(op='aten._convolution', overload='deprecated')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0360>, kernel=<OpOverload(op='aten._convolution', overload='deprecated')>), <OpOverload(op='aten.linalg_multi_dot', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9580>, kernel=<OpOverload(op='aten.linalg_multi_dot', overload='default')>), <OpOverload(op='aten.linalg_slogdet', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6f20>, kernel=<OpOverload(op='aten.linalg_slogdet', overload='default')>), <OpOverload(op='aten.linalg_matrix_rank', overload='atol_rtol_tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c04a0>, kernel=<OpOverload(op='aten.linalg_matrix_rank', overload='atol_rtol_tensor')>), <OpOverload(op='aten._version', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5bc0>, kernel=<OpOverload(op='aten._version', overload='default')>), <OpOverload(op='aten.matrix_exp_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b63e0>, kernel=<OpOverload(op='aten.matrix_exp_backward', overload='default')>), <OpOverload(op='aten.get_gradients', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0180>, kernel=<OpOverload(op='aten.get_gradients', overload='default')>), <OpOverload(op='aten.rms_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4400>, kernel=<OpOverload(op='aten.rms_norm', overload='default')>), <OpOverload(op='aten.special_psi', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cbba0>, kernel=<OpOverload(op='aten.special_psi', overload='default')>), <OpOverload(op='c10d_functional.all_reduce', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f42c0>, kernel=<OpOverload(op='c10d_functional.all_reduce', overload='default')>), <OpOverload(op='aten._propagate_xla_data', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca480>, kernel=<OpOverload(op='aten._propagate_xla_data', overload='default')>), <OpOverload(op='aten.outer', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9a80>, kernel=<OpOverload(op='aten.outer', overload='default')>), <OpOverload(op='aten.choose_qparams_optimized', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca020>, kernel=<OpOverload(op='aten.choose_qparams_optimized', overload='default')>), <OpOverload(op='aten.conv_transpose3d', overload='input')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b60c0>, kernel=<OpOverload(op='aten.conv_transpose3d', overload='input')>), <OpOverload(op='aten.fbgemm_linear_int8_weight_fp32_activation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b45e0>, kernel=<OpOverload(op='aten.fbgemm_linear_int8_weight_fp32_activation', overload='default')>), <OpOverload(op='aten.min', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7d80>, kernel=<OpOverload(op='aten.min', overload='names_dim')>), <OpOverload(op='aten._wrapped_linear_prepack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6a20>, kernel=<OpOverload(op='aten._wrapped_linear_prepack', overload='default')>), <OpOverload(op='aten.output_nr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb420>, kernel=<OpOverload(op='aten.output_nr', overload='default')>), <OpOverload(op='aten._sparse_mm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b65c0>, kernel=<OpOverload(op='aten._sparse_mm', overload='default')>), <OpOverload(op='aten.conv_transpose2d', overload='input')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d31a0>, kernel=<OpOverload(op='aten.conv_transpose2d', overload='input')>), <OpOverload(op='aten.cumprod_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9800>, kernel=<OpOverload(op='aten.cumprod_backward', overload='default')>), <OpOverload(op='aten.stride', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d27a0>, kernel=<OpOverload(op='aten.stride', overload='Dimname')>), <OpOverload(op='aten.cummin', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cae80>, kernel=<OpOverload(op='aten.cummin', overload='dimname')>), <OpOverload(op='aten.max', overload='other')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da160>, kernel=<OpOverload(op='aten.max', overload='other')>), <OpOverload(op='image.decode_jpegs_cuda', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dad40>, kernel=<OpOverload(op='image.decode_jpegs_cuda', overload='default')>), <OpOverload(op='aten.cummax', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da660>, kernel=<OpOverload(op='aten.cummax', overload='dimname')>), <OpOverload(op='aten._use_cudnn_rnn_flatten_weight', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9580>, kernel=<OpOverload(op='aten._use_cudnn_rnn_flatten_weight', overload='default')>), <OpOverload(op='aten.nll_loss2d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1ee0>, kernel=<OpOverload(op='aten.nll_loss2d', overload='default')>), <OpOverload(op='aten.isreal', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0e00>, kernel=<OpOverload(op='aten.isreal', overload='default')>), <OpOverload(op='aten.rrelu', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0d60>, kernel=<OpOverload(op='aten.rrelu', overload='default')>), <OpOverload(op='prepacked.unpack_prepacked_sizes_linear', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dbba0>, kernel=<OpOverload(op='prepacked.unpack_prepacked_sizes_linear', overload='default')>), <OpOverload(op='aten.not_equal', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3560>, kernel=<OpOverload(op='aten.not_equal', overload='Tensor')>), <OpOverload(op='aten._dim_arange', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3420>, kernel=<OpOverload(op='aten._dim_arange', overload='default')>), <OpOverload(op='aten.linalg_tensorsolve', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2160>, kernel=<OpOverload(op='aten.linalg_tensorsolve', overload='default')>), <OpOverload(op='_test.get_first', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4d60>, kernel=<OpOverload(op='_test.get_first', overload='default')>), <OpOverload(op='c10d_functional.reduce_scatter_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4c20>, kernel=<OpOverload(op='c10d_functional.reduce_scatter_tensor', overload='default')>), <OpOverload(op='aten._convolution_double_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb380>, kernel=<OpOverload(op='aten._convolution_double_backward', overload='default')>), <OpOverload(op='aten.fbgemm_linear_fp16_weight_fp32_activation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb240>, kernel=<OpOverload(op='aten.fbgemm_linear_fp16_weight_fp32_activation', overload='default')>), <OpOverload(op='aten.less', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d20c0>, kernel=<OpOverload(op='aten.less', overload='Scalar')>), <OpOverload(op='aten._remove_batch_dim', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4360>, kernel=<OpOverload(op='aten._remove_batch_dim', overload='default')>), <OpOverload(op='aten.frobenius_norm', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7880>, kernel=<OpOverload(op='aten.frobenius_norm', overload='dim')>), <OpOverload(op='aten.special_xlogy', overload='other_scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a273ce0>, kernel=<OpOverload(op='aten.special_xlogy', overload='other_scalar')>), <OpOverload(op='aten.linalg_vecdot', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c87c0>, kernel=<OpOverload(op='aten.linalg_vecdot', overload='default')>), <OpOverload(op='aten.linalg_cholesky', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb920>, kernel=<OpOverload(op='aten.linalg_cholesky', overload='default')>), <OpOverload(op='aten.special_i0', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0220>, kernel=<OpOverload(op='aten.special_i0', overload='default')>), <OpOverload(op='aten.inner', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb560>, kernel=<OpOverload(op='aten.inner', overload='default')>), <OpOverload(op='aten.multiply', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb100>, kernel=<OpOverload(op='aten.multiply', overload='Scalar')>), <OpOverload(op='aten.absolute', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c37e0>, kernel=<OpOverload(op='aten.absolute', overload='default')>), <OpOverload(op='aten._pad_packed_sequence', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b56c0>, kernel=<OpOverload(op='aten._pad_packed_sequence', overload='default')>), <OpOverload(op='aten._cufft_get_plan_cache_size', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2660>, kernel=<OpOverload(op='aten._cufft_get_plan_cache_size', overload='default')>), <OpOverload(op='aten.linalg_cond', overload='p_str')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3740>, kernel=<OpOverload(op='aten.linalg_cond', overload='p_str')>), <OpOverload(op='aten.special_xlogy', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c27a0>, kernel=<OpOverload(op='aten.special_xlogy', overload='default')>), <OpOverload(op='aten.greater', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3380>, kernel=<OpOverload(op='aten.greater', overload='Scalar')>), <OpOverload(op='aten.linalg_norm', overload='ord_str')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d37e0>, kernel=<OpOverload(op='aten.linalg_norm', overload='ord_str')>), <OpOverload(op='aten.l1_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca700>, kernel=<OpOverload(op='aten.l1_loss', overload='default')>), <OpOverload(op='aten.bilinear', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9760>, kernel=<OpOverload(op='aten.bilinear', overload='default')>), <OpOverload(op='aten.fbgemm_pack_gemm_matrix_fp16', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0caca0>, kernel=<OpOverload(op='aten.fbgemm_pack_gemm_matrix_fp16', overload='default')>), <OpOverload(op='aten.conv_tbc_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c85e0>, kernel=<OpOverload(op='aten.conv_tbc_backward', overload='default')>), <OpOverload(op='aten._cast_Long', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0860>, kernel=<OpOverload(op='aten._cast_Long', overload='default')>), <OpOverload(op='aten._validate_sparse_compressed_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d99e0>, kernel=<OpOverload(op='aten._validate_sparse_compressed_tensor_args', overload='default')>), <OpOverload(op='aten.sparse_bsr_tensor', overload='crow_col_value_size')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c05e0>, kernel=<OpOverload(op='aten.sparse_bsr_tensor', overload='crow_col_value_size')>), <OpOverload(op='aten.gather', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0400>, kernel=<OpOverload(op='aten.gather', overload='dimname')>), <OpOverload(op='aten._lu_with_info', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2ac0>, kernel=<OpOverload(op='aten._lu_with_info', overload='default')>), <OpOverload(op='aten.data', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0f40>, kernel=<OpOverload(op='aten.data', overload='default')>), <OpOverload(op='aten._is_zerotensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d25c0>, kernel=<OpOverload(op='aten._is_zerotensor', overload='default')>), <OpOverload(op='aten.min', overload='other')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7060>, kernel=<OpOverload(op='aten.min', overload='other')>), <OpOverload(op='aten.cudnn_is_acceptable', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3880>, kernel=<OpOverload(op='aten.cudnn_is_acceptable', overload='default')>), <OpOverload(op='aten.linalg_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3600>, kernel=<OpOverload(op='aten.linalg_norm', overload='default')>), <OpOverload(op='aten.gradient', overload='scalarrayint')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3ec0>, kernel=<OpOverload(op='aten.gradient', overload='scalarrayint')>), <OpOverload(op='aten._sparse_softmax', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3e20>, kernel=<OpOverload(op='aten._sparse_softmax', overload='int')>), <OpOverload(op='aten.nanmedian', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9b20>, kernel=<OpOverload(op='aten.nanmedian', overload='names_dim')>), <OpOverload(op='aten.result_type', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3c40>, kernel=<OpOverload(op='aten.result_type', overload='Scalar')>), <OpOverload(op='image.decode_gif', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f5120>, kernel=<OpOverload(op='image.decode_gif', overload='default')>), <OpOverload(op='profiler._record_function_exit', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dbc40>, kernel=<OpOverload(op='profiler._record_function_exit', overload='default')>), <OpOverload(op='aten.linalg_eigvalsh', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d04a0>, kernel=<OpOverload(op='aten.linalg_eigvalsh', overload='default')>), <OpOverload(op='aten._cast_Byte', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d22a0>, kernel=<OpOverload(op='aten._cast_Byte', overload='default')>), <OpOverload(op='aten.to_sparse', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1d00>, kernel=<OpOverload(op='aten.to_sparse', overload='default')>), <OpOverload(op='aten._test_ambiguous_defaults', overload='a')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c09a0>, kernel=<OpOverload(op='aten._test_ambiguous_defaults', overload='a')>), <OpOverload(op='aten.conv1d', overload='padding')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b74c0>, kernel=<OpOverload(op='aten.conv1d', overload='padding')>), <OpOverload(op='aten._validate_sparse_csc_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3a60>, kernel=<OpOverload(op='aten._validate_sparse_csc_tensor_args', overload='default')>), <OpOverload(op='aten._cast_Int', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da200>, kernel=<OpOverload(op='aten._cast_Int', overload='default')>), <OpOverload(op='aten._sparse_coo_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9d00>, kernel=<OpOverload(op='aten._sparse_coo_tensor_unsafe', overload='default')>), <OpOverload(op='aten.fake_quantize_per_tensor_affine_cachemask_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9800>, kernel=<OpOverload(op='aten.fake_quantize_per_tensor_affine_cachemask_backward', overload='default')>), <OpOverload(op='aten.special_log1p', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3920>, kernel=<OpOverload(op='aten.special_log1p', overload='default')>), <OpOverload(op='aten.slogdet', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1f80>, kernel=<OpOverload(op='aten.slogdet', overload='default')>), <OpOverload(op='aten.special_logit', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d80e0>, kernel=<OpOverload(op='aten.special_logit', overload='default')>), <OpOverload(op='aten.adaptive_avg_pool2d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da0c0>, kernel=<OpOverload(op='aten.adaptive_avg_pool2d', overload='default')>), <OpOverload(op='aten.fake_quantize_per_channel_affine', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d09a0>, kernel=<OpOverload(op='aten.fake_quantize_per_channel_affine', overload='default')>), <OpOverload(op='aten.rnn_tanh_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca0c0>, kernel=<OpOverload(op='aten.rnn_tanh_cell', overload='default')>), <OpOverload(op='aten._sparse_sum', overload='dtype')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0c20>, kernel=<OpOverload(op='aten._sparse_sum', overload='dtype')>), <OpOverload(op='aten.special_polygamma', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0860>, kernel=<OpOverload(op='aten.special_polygamma', overload='default')>), <OpOverload(op='aten.special_exp2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2020>, kernel=<OpOverload(op='aten.special_exp2', overload='default')>), <OpOverload(op='aten.gather_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0540>, kernel=<OpOverload(op='aten.gather_backward', overload='default')>), <OpOverload(op='aten.native_channel_shuffle', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2160>, kernel=<OpOverload(op='aten.native_channel_shuffle', overload='default')>), <OpOverload(op='aten.arccos', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5f80>, kernel=<OpOverload(op='aten.arccos', overload='default')>), <OpOverload(op='aten.sym_stride', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8360>, kernel=<OpOverload(op='aten.sym_stride', overload='int')>), <OpOverload(op='aten.ctc_loss', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1b20>, kernel=<OpOverload(op='aten.ctc_loss', overload='Tensor')>), <OpOverload(op='aten._cast_Half', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0680>, kernel=<OpOverload(op='aten._cast_Half', overload='default')>), <OpOverload(op='aten._cast_Float', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3ec0>, kernel=<OpOverload(op='aten._cast_Float', overload='default')>), <OpOverload(op='aten._sparse_softmax', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5e40>, kernel=<OpOverload(op='aten._sparse_softmax', overload='Dimname')>), <OpOverload(op='aten.sparse_bsr_tensor', overload='crow_col_value')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77763296b880>, kernel=<OpOverload(op='aten.sparse_bsr_tensor', overload='crow_col_value')>), <OpOverload(op='aten.trapz', overload='x')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8860>, kernel=<OpOverload(op='aten.trapz', overload='x')>), <OpOverload(op='aten.tensordot', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2d40>, kernel=<OpOverload(op='aten.tensordot', overload='default')>), <OpOverload(op='aten.sparse_csc_tensor', overload='ccol_row_value_size')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6e80>, kernel=<OpOverload(op='aten.sparse_csc_tensor', overload='ccol_row_value_size')>), <OpOverload(op='aten.nanmean', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d94e0>, kernel=<OpOverload(op='aten.nanmean', overload='default')>), <OpOverload(op='aten.align_as', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4900>, kernel=<OpOverload(op='aten.align_as', overload='default')>), <OpOverload(op='aten.__or__', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6ca0>, kernel=<OpOverload(op='aten.__or__', overload='Scalar')>), <OpOverload(op='aten.gradient', overload='array')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a273a60>, kernel=<OpOverload(op='aten.gradient', overload='array')>), <OpOverload(op='aten.arctan', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7380>, kernel=<OpOverload(op='aten.arctan', overload='default')>), <OpOverload(op='aten.chalf', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4fe0>, kernel=<OpOverload(op='aten.chalf', overload='default')>), <OpOverload(op='image._jpeg_version', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dbf60>, kernel=<OpOverload(op='image._jpeg_version', overload='default')>), <OpOverload(op='aten.ldexp', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6980>, kernel=<OpOverload(op='aten.ldexp', overload='Tensor')>), <OpOverload(op='aten.log_softmax', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6c00>, kernel=<OpOverload(op='aten.log_softmax', overload='Dimname')>), <OpOverload(op='aten._sobol_engine_draw', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b47c0>, kernel=<OpOverload(op='aten._sobol_engine_draw', overload='default')>), <OpOverload(op='aten.trapezoid', overload='x')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5c60>, kernel=<OpOverload(op='aten.trapezoid', overload='x')>), <OpOverload(op='aten.linalg_solve', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7c40>, kernel=<OpOverload(op='aten.linalg_solve', overload='default')>), <OpOverload(op='aten.kthvalue', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b53a0>, kernel=<OpOverload(op='aten.kthvalue', overload='dimname')>), <OpOverload(op='aten.histogramdd', overload='int_bins')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5ee0>, kernel=<OpOverload(op='aten.histogramdd', overload='int_bins')>), <OpOverload(op='c10d_functional.broadcast', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4540>, kernel=<OpOverload(op='c10d_functional.broadcast', overload='default')>), <OpOverload(op='aten._pack_padded_sequence_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b49a0>, kernel=<OpOverload(op='aten._pack_padded_sequence_backward', overload='default')>), <OpOverload(op='aten._reshape_from_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c99e0>, kernel=<OpOverload(op='aten._reshape_from_tensor', overload='default')>), <OpOverload(op='aten.sparse_bsc_tensor', overload='ccol_row_value_size')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a273b00>, kernel=<OpOverload(op='aten.sparse_bsc_tensor', overload='ccol_row_value_size')>), <OpOverload(op='aten.triplet_margin_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4cc0>, kernel=<OpOverload(op='aten.triplet_margin_loss', overload='default')>), <OpOverload(op='quantized.conv2d_unpack_sizes', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dbd80>, kernel=<OpOverload(op='quantized.conv2d_unpack_sizes', overload='default')>), <OpOverload(op='aten.linalg_eigh', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8ea0>, kernel=<OpOverload(op='aten.linalg_eigh', overload='default')>), <OpOverload(op='aten.is_leaf', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2fc0>, kernel=<OpOverload(op='aten.is_leaf', overload='default')>), <OpOverload(op='aten.repeat_interleave', overload='self_Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4ea0>, kernel=<OpOverload(op='aten.repeat_interleave', overload='self_Tensor')>), <OpOverload(op='aten._has_compatible_shallow_copy_type', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1620>, kernel=<OpOverload(op='aten._has_compatible_shallow_copy_type', overload='default')>), <OpOverload(op='torchvision._cuda_version', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f5260>, kernel=<OpOverload(op='torchvision._cuda_version', overload='default')>), <OpOverload(op='aten.cumulative_trapezoid', overload='dx')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9f80>, kernel=<OpOverload(op='aten.cumulative_trapezoid', overload='dx')>), <OpOverload(op='_test.leaky_relu', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f5620>, kernel=<OpOverload(op='_test.leaky_relu', overload='default')>), <OpOverload(op='aten.cummaxmin_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8900>, kernel=<OpOverload(op='aten.cummaxmin_backward', overload='default')>), <OpOverload(op='aten._sparse_sum', overload='dim_dtype')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d85e0>, kernel=<OpOverload(op='aten._sparse_sum', overload='dim_dtype')>), <OpOverload(op='aten.sparse_csr_tensor', overload='crow_col_value')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d93a0>, kernel=<OpOverload(op='aten.sparse_csr_tensor', overload='crow_col_value')>), <OpOverload(op='aten._batch_norm_impl_index_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2a20>, kernel=<OpOverload(op='aten._batch_norm_impl_index_backward', overload='default')>), <OpOverload(op='aten.fliplr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1760>, kernel=<OpOverload(op='aten.fliplr', overload='default')>), <OpOverload(op='aten.argwhere', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8f40>, kernel=<OpOverload(op='aten.argwhere', overload='default')>), <OpOverload(op='aten.rnn_relu_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6d40>, kernel=<OpOverload(op='aten.rnn_relu_cell', overload='default')>), <OpOverload(op='aten.row_stack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7560>, kernel=<OpOverload(op='aten.row_stack', overload='default')>), <OpOverload(op='aten.__and__', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6020>, kernel=<OpOverload(op='aten.__and__', overload='Tensor')>), <OpOverload(op='aten.adaptive_avg_pool1d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a273ec0>, kernel=<OpOverload(op='aten.adaptive_avg_pool1d', overload='default')>), <OpOverload(op='aten.is_conj', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b79c0>, kernel=<OpOverload(op='aten.is_conj', overload='default')>), <OpOverload(op='aten.argsort', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca340>, kernel=<OpOverload(op='aten.argsort', overload='default')>), <OpOverload(op='aten.cov', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4a40>, kernel=<OpOverload(op='aten.cov', overload='default')>), <OpOverload(op='aten.value_selecting_reduction_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6200>, kernel=<OpOverload(op='aten.value_selecting_reduction_backward', overload='default')>), <OpOverload(op='aten.to_dense', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6ac0>, kernel=<OpOverload(op='aten.to_dense', overload='default')>), <OpOverload(op='aten.diagflat', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1300>, kernel=<OpOverload(op='aten.diagflat', overload='default')>), <OpOverload(op='aten.is_inference', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c20c0>, kernel=<OpOverload(op='aten.is_inference', overload='default')>), <OpOverload(op='aten.to_sparse_bsr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2a20>, kernel=<OpOverload(op='aten.to_sparse_bsr', overload='default')>), <OpOverload(op='aten.linalg_pinv', overload='atol_rtol_float')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0fe0>, kernel=<OpOverload(op='aten.linalg_pinv', overload='atol_rtol_float')>), <OpOverload(op='aten.__and__', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9940>, kernel=<OpOverload(op='aten.__and__', overload='Scalar')>), <OpOverload(op='aten.fbgemm_linear_fp16_weight', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4860>, kernel=<OpOverload(op='aten.fbgemm_linear_fp16_weight', overload='default')>), <OpOverload(op='aten.__xor__', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca8e0>, kernel=<OpOverload(op='aten.__xor__', overload='Scalar')>), <OpOverload(op='aten.mode', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0ea0>, kernel=<OpOverload(op='aten.mode', overload='dimname')>), <OpOverload(op='aten.to_dense_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1580>, kernel=<OpOverload(op='aten.to_dense_backward', overload='default')>), <OpOverload(op='image.encode_jpegs_cuda', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db240>, kernel=<OpOverload(op='image.encode_jpegs_cuda', overload='default')>), <OpOverload(op='aten.qr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9c60>, kernel=<OpOverload(op='aten.qr', overload='default')>), <OpOverload(op='aten.special_round', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b51c0>, kernel=<OpOverload(op='aten.special_round', overload='default')>), <OpOverload(op='aten.flipud', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4680>, kernel=<OpOverload(op='aten.flipud', overload='default')>), <OpOverload(op='aten._pad_circular', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3f60>, kernel=<OpOverload(op='aten._pad_circular', overload='default')>), <OpOverload(op='aten._cast_Double', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9300>, kernel=<OpOverload(op='aten._cast_Double', overload='default')>), <OpOverload(op='aten.index_select_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b68e0>, kernel=<OpOverload(op='aten.index_select_backward', overload='default')>), <OpOverload(op='aten.nuclear_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8860>, kernel=<OpOverload(op='aten.nuclear_norm', overload='default')>), <OpOverload(op='aten.linalg_matmul', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8540>, kernel=<OpOverload(op='aten.linalg_matmul', overload='default')>), <OpOverload(op='_test.cat', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4180>, kernel=<OpOverload(op='_test.cat', overload='default')>), <OpOverload(op='aten.argsort', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6480>, kernel=<OpOverload(op='aten.argsort', overload='dimname')>), <OpOverload(op='aten.msort', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5260>, kernel=<OpOverload(op='aten.msort', overload='default')>), <OpOverload(op='mkldnn._is_mkldnn_bf16_supported', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db100>, kernel=<OpOverload(op='mkldnn._is_mkldnn_bf16_supported', overload='default')>), <OpOverload(op='aten.linalg_cond', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9620>, kernel=<OpOverload(op='aten.linalg_cond', overload='default')>), <OpOverload(op='aten.vstack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1bc0>, kernel=<OpOverload(op='aten.vstack', overload='default')>), <OpOverload(op='aten.matrix_power', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8040>, kernel=<OpOverload(op='aten.matrix_power', overload='default')>), <OpOverload(op='aten.align_tensors', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0a40>, kernel=<OpOverload(op='aten.align_tensors', overload='default')>), <OpOverload(op='aten.one_hot', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5d00>, kernel=<OpOverload(op='aten.one_hot', overload='default')>), <OpOverload(op='aten.greater_equal', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3b00>, kernel=<OpOverload(op='aten.greater_equal', overload='Scalar')>), <OpOverload(op='aten.cdist', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7f60>, kernel=<OpOverload(op='aten.cdist', overload='default')>), <OpOverload(op='aten.special_gammaln', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b54e0>, kernel=<OpOverload(op='aten.special_gammaln', overload='default')>), <OpOverload(op='aten._to_cpu', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0cc0>, kernel=<OpOverload(op='aten._to_cpu', overload='default')>), <OpOverload(op='aten.to_sparse_csr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7740>, kernel=<OpOverload(op='aten.to_sparse_csr', overload='default')>), <OpOverload(op='aten.linalg_pinv', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1800>, kernel=<OpOverload(op='aten.linalg_pinv', overload='default')>), <OpOverload(op='aten.kron', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0ae0>, kernel=<OpOverload(op='aten.kron', overload='default')>), <OpOverload(op='aten.fbgemm_pack_quantized_matrix', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3a60>, kernel=<OpOverload(op='aten.fbgemm_pack_quantized_matrix', overload='default')>), <OpOverload(op='aten.ctc_loss', overload='IntList')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2e80>, kernel=<OpOverload(op='aten.ctc_loss', overload='IntList')>), <OpOverload(op='aten.linalg_vander', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2980>, kernel=<OpOverload(op='aten.linalg_vander', overload='default')>), <OpOverload(op='aten._scaled_dot_product_attention_math', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cbf60>, kernel=<OpOverload(op='aten._scaled_dot_product_attention_math', overload='default')>), <OpOverload(op='aten.embedding_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4e00>, kernel=<OpOverload(op='aten.embedding_backward', overload='default')>), <OpOverload(op='aten.special_gammaincc', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c19e0>, kernel=<OpOverload(op='aten.special_gammaincc', overload='default')>), <OpOverload(op='image.decode_jpeg', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4a40>, kernel=<OpOverload(op='image.decode_jpeg', overload='default')>), <OpOverload(op='aten.inverse', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d87c0>, kernel=<OpOverload(op='aten.inverse', overload='default')>), <OpOverload(op='aten._sparse_bsc_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca7a0>, kernel=<OpOverload(op='aten._sparse_bsc_tensor_unsafe', overload='default')>), <OpOverload(op='aten.stride', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2520>, kernel=<OpOverload(op='aten.stride', overload='int')>), <OpOverload(op='aten._sparse_compressed_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9bc0>, kernel=<OpOverload(op='aten._sparse_compressed_tensor_unsafe', overload='default')>), <OpOverload(op='aten.nested_to_padded_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9ee0>, kernel=<OpOverload(op='aten.nested_to_padded_tensor', overload='default')>), <OpOverload(op='aten.linalg_matrix_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca520>, kernel=<OpOverload(op='aten.linalg_matrix_norm', overload='default')>), <OpOverload(op='aten._cufft_clear_plan_cache', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8c20>, kernel=<OpOverload(op='aten._cufft_clear_plan_cache', overload='default')>), <OpOverload(op='aten._wrapped_quantized_linear_prepacked', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb1a0>, kernel=<OpOverload(op='aten._wrapped_quantized_linear_prepacked', overload='default')>), <OpOverload(op='aten.special_erfinv', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2e80>, kernel=<OpOverload(op='aten.special_erfinv', overload='default')>), <OpOverload(op='aten.fix', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2200>, kernel=<OpOverload(op='aten.fix', overload='default')>), <OpOverload(op='aten.size', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c93a0>, kernel=<OpOverload(op='aten.size', overload='int')>), <OpOverload(op='aten.quantized_rnn_tanh_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cbc40>, kernel=<OpOverload(op='aten.quantized_rnn_tanh_cell', overload='default')>), <OpOverload(op='aten.pinverse', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cad40>, kernel=<OpOverload(op='aten.pinverse', overload='default')>), <OpOverload(op='aten._sparse_bsr_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c23e0>, kernel=<OpOverload(op='aten._sparse_bsr_tensor_unsafe', overload='default')>), <OpOverload(op='aten.fbgemm_linear_quantize_weight', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3420>, kernel=<OpOverload(op='aten.fbgemm_linear_quantize_weight', overload='default')>), <OpOverload(op='aten._add_batch_dim', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3ce0>, kernel=<OpOverload(op='aten._add_batch_dim', overload='default')>), <OpOverload(op='aten._nnpack_available', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b40e0>, kernel=<OpOverload(op='aten._nnpack_available', overload='default')>), <OpOverload(op='aten.fbgemm_pack_quantized_matrix', overload='KN')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2ac0>, kernel=<OpOverload(op='aten.fbgemm_pack_quantized_matrix', overload='KN')>), <OpOverload(op='aten.special_logsumexp', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6660>, kernel=<OpOverload(op='aten.special_logsumexp', overload='default')>), <OpOverload(op='aten._cast_Char', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8f40>, kernel=<OpOverload(op='aten._cast_Char', overload='default')>), <OpOverload(op='mkldnn._is_mkldnn_fp16_supported', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db060>, kernel=<OpOverload(op='mkldnn._is_mkldnn_fp16_supported', overload='default')>), <OpOverload(op='aten.nanquantile', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1440>, kernel=<OpOverload(op='aten.nanquantile', overload='default')>), <OpOverload(op='aten.is_neg', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6b60>, kernel=<OpOverload(op='aten.is_neg', overload='default')>), <OpOverload(op='aten._rowwise_prune', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0f40>, kernel=<OpOverload(op='aten._rowwise_prune', overload='default')>), <OpOverload(op='aten.column_stack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a273f60>, kernel=<OpOverload(op='aten.column_stack', overload='default')>), <OpOverload(op='aten.linalg_pinv', overload='rcond_tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4d60>, kernel=<OpOverload(op='aten.linalg_pinv', overload='rcond_tensor')>), <OpOverload(op='aten.size', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8900>, kernel=<OpOverload(op='aten.size', overload='Dimname')>), <OpOverload(op='aten.sort', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c36a0>, kernel=<OpOverload(op='aten.sort', overload='dimname')>), <OpOverload(op='aten.result_type', overload='Scalar_Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b67a0>, kernel=<OpOverload(op='aten.result_type', overload='Scalar_Tensor')>), <OpOverload(op='aten.fake_quantize_per_tensor_affine', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d98a0>, kernel=<OpOverload(op='aten.fake_quantize_per_tensor_affine', overload='default')>), <OpOverload(op='aten.diff', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cba60>, kernel=<OpOverload(op='aten.diff', overload='default')>), <OpOverload(op='aten.special_expit', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0180>, kernel=<OpOverload(op='aten.special_expit', overload='default')>), <OpOverload(op='aten.sparse_coo_tensor', overload='indices')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9080>, kernel=<OpOverload(op='aten.sparse_coo_tensor', overload='indices')>), <OpOverload(op='aten._thnn_differentiable_lstm_cell_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c34c0>, kernel=<OpOverload(op='aten._thnn_differentiable_lstm_cell_backward', overload='default')>), <OpOverload(op='aten.cosine_embedding_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb740>, kernel=<OpOverload(op='aten.cosine_embedding_loss', overload='default')>), <OpOverload(op='aten.arccosh', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3920>, kernel=<OpOverload(op='aten.arccosh', overload='default')>), <OpOverload(op='aten._embedding_bag_sparse_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c84a0>, kernel=<OpOverload(op='aten._embedding_bag_sparse_backward', overload='default')>), <OpOverload(op='image.write_file', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db600>, kernel=<OpOverload(op='image.write_file', overload='default')>), <OpOverload(op='mkldnn._is_mkldnn_acl_supported', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f5300>, kernel=<OpOverload(op='mkldnn._is_mkldnn_acl_supported', overload='default')>), <OpOverload(op='aten.result_type', overload='Scalar_Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3600>, kernel=<OpOverload(op='aten.result_type', overload='Scalar_Scalar')>), <OpOverload(op='aten.dstack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1da0>, kernel=<OpOverload(op='aten.dstack', overload='default')>), <OpOverload(op='aten.is_floating_point', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cab60>, kernel=<OpOverload(op='aten.is_floating_point', overload='default')>), <OpOverload(op='aten._sparse_log_softmax', overload='Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3100>, kernel=<OpOverload(op='aten._sparse_log_softmax', overload='Dimname')>), <OpOverload(op='aten._test_autograd_multiple_dispatch', overload='ntonly')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b44a0>, kernel=<OpOverload(op='aten._test_autograd_multiple_dispatch', overload='ntonly')>), <OpOverload(op='c10d_functional.all_reduce_coalesced', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4e00>, kernel=<OpOverload(op='c10d_functional.all_reduce_coalesced', overload='default')>), <OpOverload(op='aten.arctan2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca840>, kernel=<OpOverload(op='aten.arctan2', overload='default')>), <OpOverload(op='aten._test_ambiguous_defaults', overload='b')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8400>, kernel=<OpOverload(op='aten._test_ambiguous_defaults', overload='b')>), <OpOverload(op='aten.special_digamma', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d82c0>, kernel=<OpOverload(op='aten.special_digamma', overload='default')>), <OpOverload(op='aten.stft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca3e0>, kernel=<OpOverload(op='aten.stft', overload='default')>), <OpOverload(op='aten.result_type', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca160>, kernel=<OpOverload(op='aten.result_type', overload='Tensor')>), <OpOverload(op='aten.gradient', overload='tensorarray')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0b80>, kernel=<OpOverload(op='aten.gradient', overload='tensorarray')>), <OpOverload(op='aten._validate_sparse_bsr_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8ea0>, kernel=<OpOverload(op='aten._validate_sparse_bsr_tensor_args', overload='default')>), <OpOverload(op='aten.divide', overload='Tensor_mode')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2c00>, kernel=<OpOverload(op='aten.divide', overload='Tensor_mode')>), <OpOverload(op='aten.linalg_eigvals', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8d60>, kernel=<OpOverload(op='aten.linalg_eigvals', overload='default')>), <OpOverload(op='aten._debug_has_internal_overlap', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b62a0>, kernel=<OpOverload(op='aten._debug_has_internal_overlap', overload='default')>), <OpOverload(op='c10d_functional.all_gather_into_tensor_coalesced', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4cc0>, kernel=<OpOverload(op='c10d_functional.all_gather_into_tensor_coalesced', overload='default')>), <OpOverload(op='aten.max', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2de0>, kernel=<OpOverload(op='aten.max', overload='names_dim')>), <OpOverload(op='aten.__xor__', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c28e0>, kernel=<OpOverload(op='aten.__xor__', overload='Tensor')>), <OpOverload(op='aten.adaptive_avg_pool3d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8680>, kernel=<OpOverload(op='aten.adaptive_avg_pool3d', overload='default')>), <OpOverload(op='aten.nanquantile', overload='scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0680>, kernel=<OpOverload(op='aten.nanquantile', overload='scalar')>), <OpOverload(op='c10d_functional.all_to_all_single', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4b80>, kernel=<OpOverload(op='c10d_functional.all_to_all_single', overload='default')>), <OpOverload(op='aten._validate_sparse_bsc_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0fe0>, kernel=<OpOverload(op='aten._validate_sparse_bsc_tensor_args', overload='default')>), <OpOverload(op='aten.conv3d', overload='padding')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9f80>, kernel=<OpOverload(op='aten.conv3d', overload='padding')>), <OpOverload(op='aten.trapezoid', overload='dx')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1260>, kernel=<OpOverload(op='aten.trapezoid', overload='dx')>), <OpOverload(op='aten.divide', overload='Scalar_mode')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8b80>, kernel=<OpOverload(op='aten.divide', overload='Scalar_mode')>), <OpOverload(op='image.encode_jpeg', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db740>, kernel=<OpOverload(op='image.encode_jpeg', overload='default')>), <OpOverload(op='aten.fbgemm_linear_int8_weight', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9ee0>, kernel=<OpOverload(op='aten.fbgemm_linear_int8_weight', overload='default')>), <OpOverload(op='aten.scatter', overload='dimname_src')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c91c0>, kernel=<OpOverload(op='aten.scatter', overload='dimname_src')>), <OpOverload(op='aten.linalg_lu_factor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2b60>, kernel=<OpOverload(op='aten.linalg_lu_factor', overload='default')>), <OpOverload(op='aten.to_sparse_csc', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c25c0>, kernel=<OpOverload(op='aten.to_sparse_csc', overload='default')>), <OpOverload(op='aten.gradient', overload='scalarint')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5a80>, kernel=<OpOverload(op='aten.gradient', overload='scalarint')>), <OpOverload(op='aten.concatenate', overload='names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1620>, kernel=<OpOverload(op='aten.concatenate', overload='names')>), <OpOverload(op='aten._sparse_mm', overload='reduce')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1440>, kernel=<OpOverload(op='aten._sparse_mm', overload='reduce')>), <OpOverload(op='aten.flatten_dense_tensors', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3740>, kernel=<OpOverload(op='aten.flatten_dense_tensors', overload='default')>), <OpOverload(op='profiler._record_function_enter_new', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4680>, kernel=<OpOverload(op='profiler._record_function_enter_new', overload='default')>), <OpOverload(op='aten.linalg_inv', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2c00>, kernel=<OpOverload(op='aten.linalg_inv', overload='default')>), <OpOverload(op='aten.sparse_coo_tensor', overload='indices_size')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2700>, kernel=<OpOverload(op='aten.sparse_coo_tensor', overload='indices_size')>), <OpOverload(op='aten.diag', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cbd80>, kernel=<OpOverload(op='aten.diag', overload='default')>), <OpOverload(op='aten._fused_rms_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0caf20>, kernel=<OpOverload(op='aten._fused_rms_norm', overload='default')>), <OpOverload(op='aten.float_power', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca980>, kernel=<OpOverload(op='aten.float_power', overload='Scalar')>), <OpOverload(op='aten.float_power', overload='Tensor_Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c07c0>, kernel=<OpOverload(op='aten.float_power', overload='Tensor_Scalar')>), <OpOverload(op='aten.float_power', overload='Tensor_Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb2e0>, kernel=<OpOverload(op='aten.float_power', overload='Tensor_Tensor')>), <OpOverload(op='aten.square', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1580>, kernel=<OpOverload(op='aten.square', overload='default')>), <OpOverload(op='aten.clip', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4040>, kernel=<OpOverload(op='aten.clip', overload='default')>), <OpOverload(op='aten.adaptive_max_pool1d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2480>, kernel=<OpOverload(op='aten.adaptive_max_pool1d', overload='default')>), <OpOverload(op='aten._test_string_default', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1260>, kernel=<OpOverload(op='aten._test_string_default', overload='default')>), <OpOverload(op='aten._validate_sparse_coo_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c02c0>, kernel=<OpOverload(op='aten._validate_sparse_coo_tensor_args', overload='default')>), <OpOverload(op='c10d_functional.wait_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db4c0>, kernel=<OpOverload(op='c10d_functional.wait_tensor', overload='default')>), <OpOverload(op='aten.combinations', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da980>, kernel=<OpOverload(op='aten.combinations', overload='default')>), <OpOverload(op='aten.quantized_gru_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b58a0>, kernel=<OpOverload(op='aten.quantized_gru_cell', overload='default')>), <OpOverload(op='aten.logcumsumexp', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7ba0>, kernel=<OpOverload(op='aten.logcumsumexp', overload='dimname')>), <OpOverload(op='aten.cosine_similarity', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4ae0>, kernel=<OpOverload(op='aten.cosine_similarity', overload='default')>), <OpOverload(op='aten._validate_sparse_csr_tensor_args', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6520>, kernel=<OpOverload(op='aten._validate_sparse_csr_tensor_args', overload='default')>), <OpOverload(op='aten.less_equal', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0b80>, kernel=<OpOverload(op='aten.less_equal', overload='Scalar')>), <OpOverload(op='image.decode_png', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db2e0>, kernel=<OpOverload(op='image.decode_png', overload='default')>), <OpOverload(op='aten.chain_matmul', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2020>, kernel=<OpOverload(op='aten.chain_matmul', overload='default')>), <OpOverload(op='aten._cufft_set_plan_cache_max_size', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb4c0>, kernel=<OpOverload(op='aten._cufft_set_plan_cache_max_size', overload='default')>), <OpOverload(op='aten.argsort', overload='stable')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cbb00>, kernel=<OpOverload(op='aten.argsort', overload='stable')>), <OpOverload(op='aten._thnn_differentiable_gru_cell_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3b00>, kernel=<OpOverload(op='aten._thnn_differentiable_gru_cell_backward', overload='default')>), <OpOverload(op='inductor._alloc_from_pool', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4900>, kernel=<OpOverload(op='inductor._alloc_from_pool', overload='default')>), <OpOverload(op='aten.nuclear_norm', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cbe20>, kernel=<OpOverload(op='aten.nuclear_norm', overload='dim')>), <OpOverload(op='aten.can_cast', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0e00>, kernel=<OpOverload(op='aten.can_cast', overload='default')>), <OpOverload(op='image.read_file', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db7e0>, kernel=<OpOverload(op='image.read_file', overload='default')>), <OpOverload(op='aten._sparse_log_softmax', overload='int')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1d00>, kernel=<OpOverload(op='aten._sparse_log_softmax', overload='int')>), <OpOverload(op='aten.nll_loss_nd', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b76a0>, kernel=<OpOverload(op='aten.nll_loss_nd', overload='default')>), <OpOverload(op='aten.slow_conv3d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da8e0>, kernel=<OpOverload(op='aten.slow_conv3d', overload='default')>), <OpOverload(op='aten.smm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9760>, kernel=<OpOverload(op='aten.smm', overload='default')>), <OpOverload(op='aten.trace_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da520>, kernel=<OpOverload(op='aten.trace_backward', overload='default')>), <OpOverload(op='aten.gradient', overload='scalarrayarray')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8e00>, kernel=<OpOverload(op='aten.gradient', overload='scalarrayarray')>), <OpOverload(op='c10d_functional.all_gather_into_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f47c0>, kernel=<OpOverload(op='c10d_functional.all_gather_into_tensor', overload='default')>), <OpOverload(op='aten.linalg_ldl_factor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2340>, kernel=<OpOverload(op='aten.linalg_ldl_factor', overload='default')>), <OpOverload(op='aten.arcsin', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d36a0>, kernel=<OpOverload(op='aten.arcsin', overload='default')>), <OpOverload(op='aten._cast_Short', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a273ba0>, kernel=<OpOverload(op='aten._cast_Short', overload='default')>), <OpOverload(op='aten._sparse_sum', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da020>, kernel=<OpOverload(op='aten._sparse_sum', overload='default')>), <OpOverload(op='aten.fake_quantize_per_channel_affine_cachemask_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da3e0>, kernel=<OpOverload(op='aten.fake_quantize_per_channel_affine_cachemask_backward', overload='default')>), <OpOverload(op='aten.sort', overload='dimname_stable')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1e40>, kernel=<OpOverload(op='aten.sort', overload='dimname_stable')>), <OpOverload(op='aten.embedding_sparse_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d34c0>, kernel=<OpOverload(op='aten.embedding_sparse_backward', overload='default')>), <OpOverload(op='aten._sparse_csr_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8540>, kernel=<OpOverload(op='aten._sparse_csr_tensor_unsafe', overload='default')>), <OpOverload(op='aten.special_multigammaln', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8720>, kernel=<OpOverload(op='aten.special_multigammaln', overload='default')>), <OpOverload(op='aten.multilabel_margin_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d16c0>, kernel=<OpOverload(op='aten.multilabel_margin_loss', overload='default')>), <OpOverload(op='aten.median', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1300>, kernel=<OpOverload(op='aten.median', overload='names_dim')>), <OpOverload(op='aten.linalg_matrix_rank', overload='tol_tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2f20>, kernel=<OpOverload(op='aten.linalg_matrix_rank', overload='tol_tensor')>), <OpOverload(op='aten.sspaddmm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9b20>, kernel=<OpOverload(op='aten.sspaddmm', overload='default')>), <OpOverload(op='aten.orgqr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d23e0>, kernel=<OpOverload(op='aten.orgqr', overload='default')>), <OpOverload(op='aten.take_along_dim', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3240>, kernel=<OpOverload(op='aten.take_along_dim', overload='default')>), <OpOverload(op='aten._cufft_get_plan_cache_max_size', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1a80>, kernel=<OpOverload(op='aten._cufft_get_plan_cache_max_size', overload='default')>), <OpOverload(op='aten.not_equal', overload='Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9d00>, kernel=<OpOverload(op='aten.not_equal', overload='Scalar')>), <OpOverload(op='aten.kl_div', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d11c0>, kernel=<OpOverload(op='aten.kl_div', overload='default')>), <OpOverload(op='aten._weight_norm', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2ca0>, kernel=<OpOverload(op='aten._weight_norm', overload='default')>), <OpOverload(op='aten.special_xlogy', overload='self_scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9a80>, kernel=<OpOverload(op='aten.special_xlogy', overload='self_scalar')>), <OpOverload(op='aten.to_mkldnn_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1940>, kernel=<OpOverload(op='aten.to_mkldnn_backward', overload='default')>), <OpOverload(op='aten._shape_as_tensor', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1120>, kernel=<OpOverload(op='aten._shape_as_tensor', overload='default')>), <OpOverload(op='prepacked.unpack_prepacked_sizes_conv2d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4ae0>, kernel=<OpOverload(op='prepacked.unpack_prepacked_sizes_conv2d', overload='default')>), <OpOverload(op='aten.__or__', overload='Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1b20>, kernel=<OpOverload(op='aten.__or__', overload='Tensor')>), <OpOverload(op='aten.histogramdd', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0ea0>, kernel=<OpOverload(op='aten.histogramdd', overload='default')>), <OpOverload(op='aten.linalg_matrix_norm', overload='str_ord')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2840>, kernel=<OpOverload(op='aten.linalg_matrix_norm', overload='str_ord')>), <OpOverload(op='profiler._record_function_enter', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4ea0>, kernel=<OpOverload(op='profiler._record_function_enter', overload='default')>), <OpOverload(op='aten.quantized_rnn_relu_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3ce0>, kernel=<OpOverload(op='aten.quantized_rnn_relu_cell', overload='default')>), <OpOverload(op='aten.trapz', overload='dx')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9bc0>, kernel=<OpOverload(op='aten.trapz', overload='dx')>), <OpOverload(op='aten.sum_to_size', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2340>, kernel=<OpOverload(op='aten.sum_to_size', overload='default')>), <OpOverload(op='aten.matrix_exp', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8cc0>, kernel=<OpOverload(op='aten.matrix_exp', overload='default')>), <OpOverload(op='aten.infinitely_differentiable_gelu_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0720>, kernel=<OpOverload(op='aten.infinitely_differentiable_gelu_backward', overload='default')>), <OpOverload(op='aten.histogramdd', overload='TensorList_bins')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9c60>, kernel=<OpOverload(op='aten.histogramdd', overload='TensorList_bins')>), <OpOverload(op='aten.quantized_lstm_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d13a0>, kernel=<OpOverload(op='aten.quantized_lstm_cell', overload='default')>), <OpOverload(op='aten.affine_grid_generator_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8220>, kernel=<OpOverload(op='aten.affine_grid_generator_backward', overload='default')>), <OpOverload(op='aten._sparse_csc_tensor_unsafe', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3060>, kernel=<OpOverload(op='aten._sparse_csc_tensor_unsafe', overload='default')>), <OpOverload(op='aten._thnn_fused_lstm_cell_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2b60>, kernel=<OpOverload(op='aten._thnn_fused_lstm_cell_backward', overload='default')>), <OpOverload(op='aten.norm_except_dim', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d07c0>, kernel=<OpOverload(op='aten.norm_except_dim', overload='default')>), <OpOverload(op='aten.gru_cell', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8b80>, kernel=<OpOverload(op='aten.gru_cell', overload='default')>), <OpOverload(op='c10d_functional.reduce_scatter_tensor_coalesced', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f5580>, kernel=<OpOverload(op='c10d_functional.reduce_scatter_tensor_coalesced', overload='default')>), <OpOverload(op='aten.poisson_nll_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d39c0>, kernel=<OpOverload(op='aten.poisson_nll_loss', overload='default')>), <OpOverload(op='aten.lu_solve', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d18a0>, kernel=<OpOverload(op='aten.lu_solve', overload='default')>), <OpOverload(op='aten._convolution_mode', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da5c0>, kernel=<OpOverload(op='aten._convolution_mode', overload='default')>), <OpOverload(op='aten._gather_sparse_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0a40>, kernel=<OpOverload(op='aten._gather_sparse_backward', overload='default')>), <OpOverload(op='aten.negative', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2700>, kernel=<OpOverload(op='aten.negative', overload='default')>), <OpOverload(op='aten.conv_transpose1d', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8ae0>, kernel=<OpOverload(op='aten.conv_transpose1d', overload='default')>), <OpOverload(op='aten.promote_types', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1760>, kernel=<OpOverload(op='aten.promote_types', overload='default')>), <OpOverload(op='aten.quantile', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2980>, kernel=<OpOverload(op='aten.quantile', overload='default')>), <OpOverload(op='aten.corrcoef', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3380>, kernel=<OpOverload(op='aten.corrcoef', overload='default')>), <OpOverload(op='aten.fft_hfftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2fc0>, kernel=<OpOverload(op='aten.fft_hfftn', overload='default')>), <OpOverload(op='aten.fft_hfftn', overload='out')>: <function hfftn at 0x7776e19d99e0>, <OpOverload(op='aten.fft_ifft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3c40>, kernel=<OpOverload(op='aten.fft_ifft2', overload='default')>), <OpOverload(op='aten.fft_ifft2', overload='out')>: <function ifft2 at 0x7776e19d96c0>, <OpOverload(op='aten.fft_rfft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c31a0>, kernel=<OpOverload(op='aten.fft_rfft2', overload='default')>), <OpOverload(op='aten.fft_rfft2', overload='out')>: <function rfft2 at 0x7776e19d8860>, <OpOverload(op='aten.fft_irfft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb7e0>, kernel=<OpOverload(op='aten.fft_irfft2', overload='default')>), <OpOverload(op='aten.fft_irfft2', overload='out')>: <function irfft2 at 0x7776e19d9da0>, <OpOverload(op='aten.fft_hfft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cb060>, kernel=<OpOverload(op='aten.fft_hfft2', overload='default')>), <OpOverload(op='aten.fft_hfft2', overload='out')>: <function hfft2 at 0x7776e19da020>, <OpOverload(op='aten.fft_ihfft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0040>, kernel=<OpOverload(op='aten.fft_ihfft2', overload='default')>), <OpOverload(op='aten.fft_ihfft2', overload='out')>: <function ihfft2 at 0x7776e19da2a0>, <OpOverload(op='aten.fft_fftshift', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d05e0>, kernel=<OpOverload(op='aten.fft_fftshift', overload='default')>), <OpOverload(op='aten.fft_ifftshift', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cbce0>, kernel=<OpOverload(op='aten.fft_ifftshift', overload='default')>), <OpOverload(op='aten.linalg_cross', overload='out')>: <function cross at 0x7776e19dab60>, <OpOverload(op='aten.linalg_vector_norm', overload='out')>: <function vector_norm at 0x7776e19daf20>, <OpOverload(op='aten.alpha_dropout', overload='default')>: <function alpha_dropout at 0x7776e19dbec0>, <OpOverload(op='aten.celu', overload='out')>: <function celu at 0x7776e1a04180>, <OpOverload(op='aten.elu', overload='out')>: <function elu at 0x7776e1a04860>, <OpOverload(op='aten.relu', overload='out')>: <function relu at 0x7776e1a04cc0>, <OpOverload(op='aten.channel_shuffle', overload='default')>: <function channel_shuffle at 0x7776e1a05300>, <OpOverload(op='aten.channel_shuffle', overload='out')>: <function channel_shuffle at 0x7776e1a05300>, <OpOverload(op='aten.leaky_relu', overload='out')>: <function leaky_relu at 0x7776e1a054e0>, <OpOverload(op='aten.mish', overload='out')>: <function mish at 0x7776e1a05260>, <OpOverload(op='aten.hardshrink', overload='default')>: <function hardshrink at 0x7776e1a06660>, <OpOverload(op='aten.softshrink', overload='default')>: <function softshrink at 0x7776e1a06980>, <OpOverload(op='aten.softplus', overload='out')>: <function softplus at 0x7776e1a05f80>, <OpOverload(op='aten.hardshrink', overload='out')>: <function hardshrink at 0x7776e1a06660>, <OpOverload(op='aten.softshrink', overload='out')>: <function softshrink at 0x7776e1a06980>, <OpOverload(op='aten.margin_ranking_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1940>, kernel=<OpOverload(op='aten.margin_ranking_loss', overload='default')>), <OpOverload(op='aten.hinge_embedding_loss', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7ce0>, kernel=<OpOverload(op='aten.hinge_embedding_loss', overload='default')>), <OpOverload(op='aten.nll_loss', overload='out')>: <function nll_loss at 0x7776e1a079c0>, <OpOverload(op='aten.huber_loss', overload='default')>: <function huber_loss at 0x7776e1a07d80>, <OpOverload(op='aten.huber_loss', overload='out')>: <function huber_loss at 0x7776e1a07d80>, <OpOverload(op='aten.threshold', overload='default')>: <function threshold at 0x7776e1a200e0>, <OpOverload(op='aten.threshold', overload='out')>: <function threshold at 0x7776e1a200e0>, <OpOverload(op='aten.special_bessel_j0', overload='default')>: <function bessel_j0 at 0x7776e1a220c0>, <OpOverload(op='aten.hardtanh', overload='out')>: <function hardtanh at 0x7776e1a07240>, <OpOverload(op='aten.special_xlog1py', overload='out')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.gelu', overload='out')>: <function gelu at 0x7776e1a20900>, <OpOverload(op='aten.special_xlog1py', overload='self_scalar')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.glu', overload='out')>: <function glu at 0x7776e1a21440>, <OpOverload(op='aten.pairwise_distance', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2520>, kernel=<OpOverload(op='aten.pairwise_distance', overload='default')>), <OpOverload(op='aten.pdist', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3100>, kernel=<OpOverload(op='aten.pdist', overload='default')>), <OpOverload(op='aten.pixel_shuffle', overload='out')>: <function pixel_shuffle at 0x7776e1a21d00>, <OpOverload(op='aten.pixel_unshuffle', overload='out')>: <function pixel_unshuffle at 0x7776e1a21f80>, <OpOverload(op='aten.celu_', overload='default')>: <function celu at 0x7776e1a20a40>, <OpOverload(op='aten.elu_', overload='default')>: <function elu at 0x7776e1a21da0>, <OpOverload(op='aten.mish_', overload='default')>: <function mish at 0x7776e1a22020>, <OpOverload(op='aten.selu_', overload='default')>: <function selu at 0x7776e1a22160>, <OpOverload(op='aten.threshold_', overload='default')>: <function threshold at 0x7776e1a222a0>, <OpOverload(op='aten.special_bessel_j0', overload='out')>: <function bessel_j0 at 0x7776e1a220c0>, <OpOverload(op='aten.special_bessel_j1', overload='default')>: <function bessel_j1 at 0x7776e1a207c0>, <OpOverload(op='aten.special_bessel_j1', overload='out')>: <function bessel_j1 at 0x7776e1a207c0>, <OpOverload(op='aten.special_entr', overload='default')>: <function entr at 0x7776e1a22d40>, <OpOverload(op='aten.special_entr', overload='out')>: <function entr at 0x7776e1a22d40>, <OpOverload(op='aten.special_erfcx', overload='out')>: <function erfcx at 0x7776e1a23100>, <OpOverload(op='aten.special_i0e', overload='default')>: <function i0e at 0x7776e1a236a0>, <OpOverload(op='aten.special_i0e', overload='out')>: <function i0e at 0x7776e1a236a0>, <OpOverload(op='aten.special_i1', overload='default')>: <function i1 at 0x7776e1a23ba0>, <OpOverload(op='aten.special_i1', overload='out')>: <function i1 at 0x7776e1a23ba0>, <OpOverload(op='aten.special_i1e', overload='default')>: <function i1e at 0x7776e1a400e0>, <OpOverload(op='aten.special_i1e', overload='out')>: <function i1e at 0x7776e1a400e0>, <OpOverload(op='aten.special_log_ndtr', overload='default')>: <function log_ndtr at 0x7776e1a404a0>, <OpOverload(op='aten.special_log_ndtr', overload='out')>: <function log_ndtr at 0x7776e1a404a0>, <OpOverload(op='aten.special_xlog1py', overload='other_scalar_out')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.logit', overload='out')>: <function logit at 0x7776e1a40860>, <OpOverload(op='aten.special_xlog1py', overload='default')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.special_xlog1py', overload='other_scalar')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.special_xlog1py', overload='self_scalar_out')>: <function xlog1py at 0x7776e1a40c20>, <OpOverload(op='aten.mvlgamma', overload='default')>: <function multigammaln at 0x7776e1a21ee0>, <OpOverload(op='aten.mvlgamma', overload='out')>: <function multigammaln at 0x7776e1a21ee0>, <OpOverload(op='aten.special_ndtr', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0ae0>, kernel=<OpOverload(op='aten.special_ndtr', overload='default')>), <OpOverload(op='aten.special_ndtr', overload='out')>: <function ndtr at 0x7776e1a40d60>, <OpOverload(op='aten.special_ndtri', overload='default')>: <function ndtri at 0x7776e1a41120>, <OpOverload(op='aten.special_ndtri', overload='out')>: <function ndtri at 0x7776e1a41120>, <OpOverload(op='aten.special_spherical_bessel_j0', overload='default')>: <function spherical_bessel_j0 at 0x7776e1a41760>, <OpOverload(op='aten.special_spherical_bessel_j0', overload='out')>: <function spherical_bessel_j0 at 0x7776e1a41760>, <OpOverload(op='aten.special_zeta', overload='default')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='other_scalar')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='self_scalar')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='out')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='self_scalar_out')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.special_zeta', overload='other_scalar_out')>: <function zeta at 0x7776e1a41bc0>, <OpOverload(op='aten.clamp_max', overload='out')>: <function clamp_max at 0x7776e1ad1120>, <OpOverload(op='aten.clamp_max', overload='Tensor_out')>: <function clamp_max at 0x7776e1ad1120>, <OpOverload(op='aten.all', overload='out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dims_out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='all_out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.all', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0cac00>, kernel=<OpOverload(op='aten.all', overload='dimname')>), <OpOverload(op='aten.all', overload='dimname_out')>: <function all at 0x7776e1ad2200>, <OpOverload(op='aten.any', overload='out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dims_out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='all_out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.any', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0caac0>, kernel=<OpOverload(op='aten.any', overload='dimname')>), <OpOverload(op='aten.any', overload='dimname_out')>: <function any at 0x7776e1ad2480>, <OpOverload(op='aten.std_mean', overload='correction_out')>: <function std_mean at 0x7776e1c656c0>, <OpOverload(op='aten.std_mean', overload='correction_names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a273740>, kernel=<OpOverload(op='aten.std_mean', overload='correction_names')>), <OpOverload(op='aten.std_mean', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a273c40>, kernel=<OpOverload(op='aten.std_mean', overload='names_dim')>), <OpOverload(op='aten.std_mean', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4540>, kernel=<OpOverload(op='aten.std_mean', overload='default')>), <OpOverload(op='aten.std', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6340>, kernel=<OpOverload(op='aten.std', overload='default')>), <OpOverload(op='aten.std', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5b20>, kernel=<OpOverload(op='aten.std', overload='dim')>), <OpOverload(op='aten.std', overload='correction')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.std', overload='correction_names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6fc0>, kernel=<OpOverload(op='aten.std', overload='correction_names')>), <OpOverload(op='aten.std', overload='names_out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.std', overload='out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.std', overload='correction_names_out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.mean', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6700>, kernel=<OpOverload(op='aten.mean', overload='names_dim')>), <OpOverload(op='aten.mean', overload='dtype_out')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.mean', overload='out')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.mean', overload='names_out')>: <function mean at 0x7776e1ad2de0>, <OpOverload(op='aten.native_layer_norm', overload='out')>: <function native_layer_norm at 0x7776e1af5e40>, <OpOverload(op='aten.stft', overload='center')>: <function stft at 0x7776e1af53a0>, <OpOverload(op='aten.broadcast_tensors', overload='default')>: <function broadcast_tensors at 0x7776e1af4220>, <OpOverload(op='aten.var_mean', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4180>, kernel=<OpOverload(op='aten.var_mean', overload='default')>), <OpOverload(op='aten.var_mean', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5620>, kernel=<OpOverload(op='aten.var_mean', overload='dim')>), <OpOverload(op='aten.var_mean', overload='correction')>: <function var_mean at 0x7776e1ad2160>, <OpOverload(op='aten.var_mean', overload='correction_names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5760>, kernel=<OpOverload(op='aten.var_mean', overload='correction_names')>), <OpOverload(op='aten.var_mean', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b5940>, kernel=<OpOverload(op='aten.var_mean', overload='names_dim')>), <OpOverload(op='aten.var_mean', overload='correction_out')>: <function var_mean at 0x7776e1ad2160>, <OpOverload(op='aten.addr', overload='out')>: <function addr at 0x7776e1ad3b00>, <OpOverload(op='aten.index_select', overload='dimname_out')>: <function index_select at 0x7776e1b10220>, <OpOverload(op='aten.index_select', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c32e0>, kernel=<OpOverload(op='aten.index_select', overload='dimname')>), <OpOverload(op='aten.constant_pad_nd', overload='out')>: <function constant_pad_nd at 0x7776e1af4c20>, <OpOverload(op='aten.flip', overload='out')>: <function flip at 0x7776e1af5440>, <OpOverload(op='aten.istft', overload='default')>: <function istft at 0x7776e1af5f80>, <OpOverload(op='aten.renorm', overload='default')>: <function renorm at 0x7776e1af4680>, <OpOverload(op='aten.renorm', overload='out')>: <function renorm at 0x7776e1af4680>, <OpOverload(op='aten.repeat', overload='out')>: <function repeat at 0x7776e1af6340>, <OpOverload(op='aten.roll', overload='out')>: <function roll at 0x7776e1af6980>, <OpOverload(op='aten.rot90', overload='default')>: <function rot90 at 0x7776e1af6c00>, <OpOverload(op='aten.rot90', overload='out')>: <function rot90 at 0x7776e1af6c00>, <OpOverload(op='aten.stack', overload='out')>: <function stack at 0x7776e1af6f20>, <OpOverload(op='aten.unbind', overload='Dimname')>: <function unbind at 0x7776e1af7560>, <OpOverload(op='aten.index_fill', overload='int_Tensor')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill', overload='int_Scalar')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill', overload='Dimname_Tensor')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2de0>, kernel=<OpOverload(op='aten.index_fill', overload='Dimname_Tensor')>), <OpOverload(op='aten.index_fill', overload='int_Scalar_out')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill', overload='int_Tensor_out')>: <function index_fill at 0x7776e1af7ba0>, <OpOverload(op='aten.index_fill_', overload='Dimname_Tensor')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.diag', overload='out')>: <function diag at 0x7776e1b10180>, <OpOverload(op='aten.index_select', overload='out')>: <function index_select at 0x7776e1b10220>, <OpOverload(op='aten.diagonal_scatter', overload='out')>: <function diagonal_scatter at 0x7776e1b104a0>, <OpOverload(op='aten.diagonal', overload='Dimname')>: <function diagonal at 0x7776e1b102c0>, <OpOverload(op='aten.diag_embed', overload='default')>: <function diag_embed at 0x7776e1b107c0>, <OpOverload(op='aten.diag_embed', overload='out')>: <function diag_embed at 0x7776e1b107c0>, <OpOverload(op='aten.block_diag', overload='default')>: <function _block_diag_iterable at 0x7776e1b10b80>, <OpOverload(op='aten.unfold_copy', overload='default')>: <function unfold_copy at 0x7776e1b114e0>, <OpOverload(op='aten.unfold_copy', overload='out')>: <function unfold_copy at 0x7776e1b114e0>, <OpOverload(op='aten.cumsum', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0ca660>, kernel=<OpOverload(op='aten.cumsum', overload='dimname')>), <OpOverload(op='aten.cumsum', overload='dimname_out')>: <function cumsum at 0x7776e1b11580>, <OpOverload(op='aten.cumsum', overload='out')>: <function cumsum at 0x7776e1b11580>, <OpOverload(op='aten.cumprod', overload='default')>: <function cumprod at 0x7776e1b11620>, <OpOverload(op='aten.cumprod', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c9e40>, kernel=<OpOverload(op='aten.cumprod', overload='dimname')>), <OpOverload(op='aten.cumprod', overload='dimname_out')>: <function cumprod at 0x7776e1b11620>, <OpOverload(op='aten.cumprod', overload='out')>: <function cumprod at 0x7776e1b11620>, <OpOverload(op='aten.arange', overload='start_out')>: <function arange at 0x7776e1b12e80>, <OpOverload(op='aten.eye', overload='m')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.eye', overload='m_out')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.eye', overload='out')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.eye', overload='default')>: <function eye at 0x7776e1b13ce0>, <OpOverload(op='aten.tril_indices', overload='out')>: <function tril_indices at 0x7776e1b41a80>, <OpOverload(op='aten.lerp', overload='Scalar_out')>: <function lerp at 0x7776e1b13240>, <OpOverload(op='aten.lerp', overload='Tensor_out')>: <function lerp at 0x7776e1b13240>, <OpOverload(op='aten.linspace', overload='Tensor_Tensor')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Tensor_Scalar')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Scalar_Tensor')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Tensor_Tensor_out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Scalar_Tensor_out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.linspace', overload='Tensor_Scalar_out')>: <function linspace at 0x7776e1b134c0>, <OpOverload(op='aten.logspace', overload='Tensor_Tensor')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Tensor_Scalar')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Scalar_Tensor')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='default')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Tensor_Tensor_out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Tensor_Scalar_out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.logspace', overload='Scalar_Tensor_out')>: <function logspace at 0x7776e1b13740>, <OpOverload(op='aten.meshgrid', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3ba0>, kernel=<OpOverload(op='aten.meshgrid', overload='default')>), <OpOverload(op='aten.meshgrid', overload='indexing')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d3d80>, kernel=<OpOverload(op='aten.meshgrid', overload='indexing')>), <OpOverload(op='aten.triu_indices', overload='out')>: <function triu_indices at 0x7776e1b419e0>, <OpOverload(op='aten.triu_indices', overload='default')>: <function triu_indices at 0x7776e1b419e0>, <OpOverload(op='aten.masked_fill', overload='Scalar_out')>: <function masked_fill at 0x7776e1b407c0>, <OpOverload(op='aten.masked_fill', overload='Tensor_out')>: <function masked_fill at 0x7776e1b407c0>, <OpOverload(op='aten.masked_fill_', overload='Scalar')>: <function masked_fill_ at 0x7776e1b405e0>, <OpOverload(op='aten.masked_fill_', overload='Tensor')>: <function masked_fill_ at 0x7776e1b405e0>, <OpOverload(op='aten.norm', overload='Scalar')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dim')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='names_ScalarOpt_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d89a0>, kernel=<OpOverload(op='aten.norm', overload='names_ScalarOpt_dim')>), <OpOverload(op='aten.norm', overload='dtype_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dtype')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dim_dtype')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='names_dtype_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='ScalarOpt_dtype_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='Scalar_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.norm', overload='names_ScalarOpt_dim_dtype')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9440>, kernel=<OpOverload(op='aten.norm', overload='names_ScalarOpt_dim_dtype')>), <OpOverload(op='aten.norm', overload='names_out')>: <function norm at 0x7776e1b40cc0>, <OpOverload(op='aten.trace', overload='default')>: <function trace at 0x7776e1b40f40>, <OpOverload(op='aten.triu', overload='out')>: <function triu at 0x7776e1b41440>, <OpOverload(op='aten.tril', overload='out')>: <function tril at 0x7776e1b416c0>, <OpOverload(op='aten.tril_indices', overload='default')>: <function tril_indices at 0x7776e1b41a80>, <OpOverload(op='aten.bucketize', overload='Tensor')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.bucketize', overload='Scalar')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.bucketize', overload='Tensor_out')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.bucketize', overload='Scalar_out')>: <function bucketize at 0x7776e1b40c20>, <OpOverload(op='aten.cauchy', overload='default')>: <function cauchy at 0x7776e1b41c60>, <OpOverload(op='aten.cauchy', overload='out')>: <function cauchy at 0x7776e1b41c60>, <OpOverload(op='aten.exponential', overload='default')>: <function exponential at 0x7776e1b42020>, <OpOverload(op='aten.exponential', overload='out')>: <function exponential at 0x7776e1b42020>, <OpOverload(op='aten.dot', overload='out')>: <function dot at 0x7776e1b41f80>, <OpOverload(op='aten.geometric', overload='default')>: <function geometric at 0x7776e1b423e0>, <OpOverload(op='aten.geometric', overload='out')>: <function geometric at 0x7776e1b423e0>, <OpOverload(op='aten.log_normal', overload='default')>: <function log_normal at 0x7776e1b427a0>, <OpOverload(op='aten.log_normal', overload='out')>: <function log_normal at 0x7776e1b427a0>, <OpOverload(op='aten.normal_', overload='default')>: <function normal_ at 0x7776e1b42840>, <OpOverload(op='aten.asinh_', overload='default')>: <function asinh at 0x7776e1b645e0>, <OpOverload(op='aten.addcmul_', overload='default')>: <function addcmul at 0x7776e1b64220>, <OpOverload(op='aten.rad2deg', overload='out')>: <function rad2deg at 0x7776e1b431a0>, <OpOverload(op='aten.deg2rad', overload='out')>: <function deg2rad at 0x7776e1b436a0>, <OpOverload(op='aten.count_nonzero', overload='dim_IntList')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.count_nonzero', overload='dim_IntList_out')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.count_nonzero', overload='default')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.count_nonzero', overload='out')>: <function count_nonzero at 0x7776e1b43920>, <OpOverload(op='aten.cumprod_', overload='default')>: <function cumprod at 0x7776e1b65940>, <OpOverload(op='aten.cumprod_', overload='dimname')>: <function cumprod at 0x7776e1b65940>, <OpOverload(op='aten.vdot', overload='default')>: <function vdot at 0x7776e1b43c40>, <OpOverload(op='aten.vdot', overload='out')>: <function vdot at 0x7776e1b43c40>, <OpOverload(op='aten.le_', overload='Tensor')>: <function le at 0x7776e1b66d40>, <OpOverload(op='aten.select_scatter', overload='out')>: <function select_scatter at 0x7776e1b43ec0>, <OpOverload(op='aten.abs_', overload='default')>: <function abs at 0x7776e1b439c0>, <OpOverload(op='aten.acos_', overload='default')>: <function acos at 0x7776e1b43ce0>, <OpOverload(op='aten.cumsum_', overload='default')>: <function cumsum at 0x7776e1b659e0>, <OpOverload(op='aten.cumsum_', overload='dimname')>: <function cumsum at 0x7776e1b659e0>, <OpOverload(op='aten.acosh_', overload='default')>: <function acosh at 0x7776e1b43f60>, <OpOverload(op='aten.add_', overload='Tensor')>: <function add at 0x7776e1b640e0>, <OpOverload(op='aten.add_', overload='Scalar')>: <function add at 0x7776e1b640e0>, <OpOverload(op='aten.cosh_', overload='default')>: <function cosh at 0x7776e1b42340>, <OpOverload(op='aten.addcdiv_', overload='default')>: <function addcdiv at 0x7776e1b64360>, <OpOverload(op='aten.asin_', overload='default')>: <function asin at 0x7776e1b644a0>, <OpOverload(op='aten.cos_', overload='default')>: <function cos at 0x7776e1b43ba0>, <OpOverload(op='aten.atan_', overload='default')>: <function atan at 0x7776e1b64720>, <OpOverload(op='aten.atanh_', overload='default')>: <function atanh at 0x7776e1b64860>, <OpOverload(op='aten.copysign_', overload='Scalar')>: <function copysign at 0x7776e1b43e20>, <OpOverload(op='aten.atan2_', overload='default')>: <function atan2 at 0x7776e1b649a0>, <OpOverload(op='aten.bitwise_and_', overload='Tensor')>: <function bitwise_and at 0x7776e1b64ae0>, <OpOverload(op='aten.bitwise_and_', overload='Scalar')>: <function bitwise_and at 0x7776e1b64ae0>, <OpOverload(op='aten.copysign_', overload='Tensor')>: <function copysign at 0x7776e1b43e20>, <OpOverload(op='aten.bitwise_left_shift_', overload='Tensor_Scalar')>: <function bitwise_left_shift at 0x7776e1b64c20>, <OpOverload(op='aten.bitwise_left_shift_', overload='Tensor')>: <function bitwise_left_shift at 0x7776e1b64c20>, <OpOverload(op='aten.bitwise_not_', overload='default')>: <function bitwise_not at 0x7776e1b64d60>, <OpOverload(op='aten.bitwise_or_', overload='Tensor')>: <function bitwise_or at 0x7776e1b64ea0>, <OpOverload(op='aten.bitwise_or_', overload='Scalar')>: <function bitwise_or at 0x7776e1b64ea0>, <OpOverload(op='aten.bitwise_right_shift_', overload='Tensor_Scalar')>: <function bitwise_right_shift at 0x7776e1b64fe0>, <OpOverload(op='aten.bitwise_right_shift_', overload='Tensor')>: <function bitwise_right_shift at 0x7776e1b64fe0>, <OpOverload(op='aten.bitwise_xor_', overload='Tensor')>: <function bitwise_xor at 0x7776e1b65120>, <OpOverload(op='aten.bitwise_xor_', overload='Scalar')>: <function bitwise_xor at 0x7776e1b65120>, <OpOverload(op='aten.ceil_', overload='default')>: <function ceil at 0x7776e1b65260>, <OpOverload(op='aten.conj_physical_', overload='default')>: <function conj_physical at 0x7776e1b65760>, <OpOverload(op='aten.clamp_', overload='default')>: <function clamp at 0x7776e1b653a0>, <OpOverload(op='aten.clamp_', overload='Tensor')>: <function clamp at 0x7776e1b653a0>, <OpOverload(op='aten.deg2rad_', overload='default')>: <function deg2rad at 0x7776e1b656c0>, <OpOverload(op='aten.clamp_min_', overload='default')>: <function clamp_min at 0x7776e1b654e0>, <OpOverload(op='aten.clamp_min_', overload='Tensor')>: <function clamp_min at 0x7776e1b654e0>, <OpOverload(op='aten.lt_', overload='Scalar')>: <function lt at 0x7776e1b65c60>, <OpOverload(op='aten.digamma_', overload='default')>: <function digamma at 0x7776e1b65440>, <OpOverload(op='aten.clamp_max_', overload='default')>: <function clamp_max at 0x7776e1b65620>, <OpOverload(op='aten.clamp_max_', overload='Tensor')>: <function clamp_max at 0x7776e1b65620>, <OpOverload(op='aten.div_', overload='Tensor')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.div_', overload='Tensor_mode')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.div_', overload='Scalar')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.div_', overload='Scalar_mode')>: <function div at 0x7776e1b651c0>, <OpOverload(op='aten.logical_xor_', overload='default')>: <function logical_xor at 0x7776e1b65ee0>, <OpOverload(op='aten.cauchy_', overload='default')>: <function cauchy at 0x7776e1b677e0>, <OpOverload(op='aten.eq_', overload='Scalar')>: <function eq at 0x7776e1b64f40>, <OpOverload(op='aten.eq_', overload='Tensor')>: <function eq at 0x7776e1b64f40>, <OpOverload(op='aten.erf_', overload='default')>: <function erf at 0x7776e1b64cc0>, <OpOverload(op='aten.logical_or_', overload='default')>: <function logical_or at 0x7776e1b66160>, <OpOverload(op='aten.erfc_', overload='default')>: <function erfc at 0x7776e1b64a40>, <OpOverload(op='aten.erfinv_', overload='default')>: <function erfinv at 0x7776e1b647c0>, <OpOverload(op='aten.exponential_', overload='default')>: <function exponential at 0x7776e1b67560>, <OpOverload(op='aten.exp_', overload='default')>: <function exp at 0x7776e1b64540>, <OpOverload(op='aten.logical_not_', overload='default')>: <function logical_not at 0x7776e1b663e0>, <OpOverload(op='aten.exp2_', overload='default')>: <function exp2 at 0x7776e1b642c0>, <OpOverload(op='aten.expm1_', overload='default')>: <function expm1 at 0x7776e1b64040>, <OpOverload(op='aten.logical_and_', overload='default')>: <function logical_and at 0x7776e1b66660>, <OpOverload(op='aten.float_power_', overload='Tensor')>: <function float_power at 0x7776e1b65bc0>, <OpOverload(op='aten.float_power_', overload='Scalar')>: <function float_power at 0x7776e1b65bc0>, <OpOverload(op='aten.floor_', overload='default')>: <function floor at 0x7776e1b65d00>, <OpOverload(op='aten.floor_divide_', overload='Scalar')>: <function floor_divide at 0x7776e1b65e40>, <OpOverload(op='aten.floor_divide_', overload='Tensor')>: <function floor_divide at 0x7776e1b65e40>, <OpOverload(op='aten.log_', overload='default')>: <function log at 0x7776e1b668e0>, <OpOverload(op='aten.fmod_', overload='Tensor')>: <function fmod at 0x7776e1b65f80>, <OpOverload(op='aten.fmod_', overload='Scalar')>: <function fmod at 0x7776e1b65f80>, <OpOverload(op='aten.frac_', overload='default')>: <function frac at 0x7776e1b660c0>, <OpOverload(op='aten.geometric_', overload='default')>: <function geometric at 0x7776e1b672e0>, <OpOverload(op='aten.log2_', overload='default')>: <function log2 at 0x7776e1b66b60>, <OpOverload(op='aten.gcd_', overload='default')>: <function gcd at 0x7776e1b66200>, <OpOverload(op='aten.ge_', overload='Scalar')>: <function ge at 0x7776e1b66340>, <OpOverload(op='aten.ge_', overload='Tensor')>: <function ge at 0x7776e1b66340>, <OpOverload(op='aten.gt_', overload='Scalar')>: <function gt at 0x7776e1b66480>, <OpOverload(op='aten.gt_', overload='Tensor')>: <function gt at 0x7776e1b66480>, <OpOverload(op='aten.log1p_', overload='default')>: <function log1p at 0x7776e1b66de0>, <OpOverload(op='aten.heaviside_', overload='default')>: <function heaviside at 0x7776e1b665c0>, <OpOverload(op='aten.log_normal_', overload='default')>: <function log_normal at 0x7776e1b658a0>, <OpOverload(op='aten.hypot_', overload='default')>: <function hypot at 0x7776e1b66700>, <OpOverload(op='aten.igamma_', overload='default')>: <function igamma at 0x7776e1b66840>, <OpOverload(op='aten.igammac_', overload='default')>: <function igammac at 0x7776e1b66980>, <OpOverload(op='aten.lgamma_', overload='default')>: <function lgamma at 0x7776e1b66fc0>, <OpOverload(op='aten.zero_', overload='default')>: <function zero at 0x7776e1b65300>, <OpOverload(op='aten.i0_', overload='default')>: <function i0 at 0x7776e1b66ac0>, <OpOverload(op='aten.lcm_', overload='default')>: <function lcm at 0x7776e1b66c00>, <OpOverload(op='aten.lerp_', overload='Scalar')>: <function lerp at 0x7776e1b66e80>, <OpOverload(op='aten.le_', overload='Scalar')>: <function le at 0x7776e1b66d40>, <OpOverload(op='aten.lt_', overload='Tensor')>: <function lt at 0x7776e1b65c60>, <OpOverload(op='aten.mul_', overload='Tensor')>: <function mul at 0x7776e1b64180>, <OpOverload(op='aten.mul_', overload='Scalar')>: <function mul at 0x7776e1b64180>, <OpOverload(op='aten.mvlgamma_', overload='default')>: <function _make_alias.<locals>._fn at 0x7776e1b64680>, <OpOverload(op='aten.nan_to_num_', overload='default')>: <function nan_to_num at 0x7776e1b64b80>, <OpOverload(op='aten.xlogy_', overload='Tensor')>: <function xlogy at 0x7776e1b67a60>, <OpOverload(op='aten.xlogy_', overload='Scalar_Other')>: <function xlogy at 0x7776e1b67a60>, <OpOverload(op='aten.ne_', overload='Scalar')>: <function ne at 0x7776e1b65080>, <OpOverload(op='aten.ne_', overload='Tensor')>: <function ne at 0x7776e1b65080>, <OpOverload(op='aten.neg_', overload='default')>: <function neg at 0x7776e1b65580>, <OpOverload(op='aten.nextafter_', overload='default')>: <function nextafter at 0x7776e1b65a80>, <OpOverload(op='aten.pow_', overload='Scalar')>: <function pow at 0x7776e1b67100>, <OpOverload(op='aten.pow_', overload='Tensor')>: <function pow at 0x7776e1b67100>, <OpOverload(op='aten.trunc_', overload='default')>: <function trunc at 0x7776e1b67ce0>, <OpOverload(op='aten.rad2deg_', overload='default')>: <function rad2deg at 0x7776e1b67240>, <OpOverload(op='aten.reciprocal_', overload='default')>: <function reciprocal at 0x7776e1b67380>, <OpOverload(op='aten.true_divide_', overload='Scalar')>: <function true_divide at 0x7776e1b67e20>, <OpOverload(op='aten.remainder_', overload='Tensor')>: <function remainder at 0x7776e1b674c0>, <OpOverload(op='aten.remainder_', overload='Scalar')>: <function remainder at 0x7776e1b674c0>, <OpOverload(op='aten.rsqrt_', overload='default')>: <function rsqrt at 0x7776e1b67600>, <OpOverload(op='aten.true_divide_', overload='Tensor')>: <function true_divide at 0x7776e1b67e20>, <OpOverload(op='aten.transpose_copy', overload='int')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.transpose at 0x7776e1b64400>, <OpOverload(op='aten.sgn_', overload='default')>: <function sgn at 0x7776e1b67740>, <OpOverload(op='aten.sigmoid_', overload='default')>: <function sigmoid at 0x7776e1b67880>, <OpOverload(op='aten.triu_', overload='default')>: <function triu at 0x7776e1b67f60>, <OpOverload(op='aten.sign_', overload='default')>: <function sign at 0x7776e1b679c0>, <OpOverload(op='aten.sin_', overload='default')>: <function sin at 0x7776e1b67b00>, <OpOverload(op='aten.transpose_copy', overload='int_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.transpose at 0x7776e1b64400>, <OpOverload(op='aten.sinc_', overload='default')>: <function sinc at 0x7776e1b67c40>, <OpOverload(op='aten.sinh_', overload='default')>: <function sinh at 0x7776e1b67d80>, <OpOverload(op='aten.sqrt_', overload='default')>: <function sqrt at 0x7776e1b67ec0>, <OpOverload(op='aten.square_', overload='default')>: <function square at 0x7776e1b98040>, <OpOverload(op='aten.tril_', overload='default')>: <function tril at 0x7776e1b98540>, <OpOverload(op='aten.sub_', overload='Tensor')>: <function sub at 0x7776e1b98180>, <OpOverload(op='aten.sub_', overload='Scalar')>: <function sub at 0x7776e1b98180>, <OpOverload(op='aten.tan_', overload='default')>: <function tan at 0x7776e1b982c0>, <OpOverload(op='aten.tanh_', overload='default')>: <function tanh at 0x7776e1b98400>, <OpOverload(op='aten.unbind_copy', overload='int')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unbind at 0x7776e1b65800>, <OpOverload(op='aten.alias_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.alias at 0x7776e1b65b20>, <OpOverload(op='aten.alias_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.alias at 0x7776e1b65b20>, <OpOverload(op='aten.as_strided_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.as_strided at 0x7776e1b667a0>, <OpOverload(op='aten.as_strided_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.as_strided at 0x7776e1b667a0>, <OpOverload(op='aten.diagonal_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.diagonal at 0x7776e1b987c0>, <OpOverload(op='aten.expand_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.expand at 0x7776e1b98360>, <OpOverload(op='aten.expand_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.expand at 0x7776e1b98360>, <OpOverload(op='aten.narrow_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.narrow at 0x7776e1b98a40>, <OpOverload(op='aten.squeeze_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='dims')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='dim_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.squeeze_copy', overload='dims_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.permute_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.permute at 0x7776e1b99080>, <OpOverload(op='aten.permute_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.permute at 0x7776e1b99080>, <OpOverload(op='aten.t_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.t at 0x7776e1b993a0>, <OpOverload(op='aten.t_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.t at 0x7776e1b993a0>, <OpOverload(op='aten.unbind_copy', overload='int_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unbind at 0x7776e1b65800>, <OpOverload(op='aten.unsqueeze_copy', overload='default')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unsqueeze at 0x7776e1b67ba0>, <OpOverload(op='aten.unsqueeze_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.unsqueeze at 0x7776e1b67ba0>, <OpOverload(op='aten.view_copy', overload='dtype')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.view_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.view_copy', overload='dtype_out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.view at 0x7776e1b99120>, <OpOverload(op='aten.complex', overload='out')>: <function complex at 0x7776e1b9ab60>, <OpOverload(op='aten.fft_ifft', overload='out')>: <function ifft at 0x7776e1b9bb00>, <OpOverload(op='aten.polar', overload='out')>: <function polar at 0x7776e1b9ade0>, <OpOverload(op='aten.fft_fft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c1e40>, kernel=<OpOverload(op='aten.fft_fft', overload='default')>), <OpOverload(op='aten.fft_fft', overload='out')>: <function fft at 0x7776e1b9b880>, <OpOverload(op='aten.fft_ifft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c18a0>, kernel=<OpOverload(op='aten.fft_ifft', overload='default')>), <OpOverload(op='aten.fft_rfft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c00e0>, kernel=<OpOverload(op='aten.fft_rfft', overload='default')>), <OpOverload(op='aten.fft_rfft', overload='out')>: <function rfft at 0x7776e1b9b7e0>, <OpOverload(op='aten.fft_irfft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7920>, kernel=<OpOverload(op='aten.fft_irfft', overload='default')>), <OpOverload(op='aten.fft_irfft', overload='out')>: <function irfft at 0x7776e1b98e00>, <OpOverload(op='aten.fft_hfft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0c20>, kernel=<OpOverload(op='aten.fft_hfft', overload='default')>), <OpOverload(op='aten.fft_hfft', overload='out')>: <function hfft at 0x7776e1b9bd80>, <OpOverload(op='aten.fft_fft2', overload='out')>: <function fft2 at 0x7776e1b9bce0>, <OpOverload(op='aten.fft_ihfft', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c0540>, kernel=<OpOverload(op='aten.fft_ihfft', overload='default')>), <OpOverload(op='aten.fft_ihfft', overload='out')>: <function ihfft at 0x7776e1b9bf60>, <OpOverload(op='aten.fft_fft2', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c22a0>, kernel=<OpOverload(op='aten.fft_fft2', overload='default')>), <OpOverload(op='aten.fft_fftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c82c0>, kernel=<OpOverload(op='aten.fft_fftn', overload='default')>), <OpOverload(op='aten.fft_fftn', overload='out')>: <function fftn at 0x7776e19d8900>, <OpOverload(op='aten.fft_ifftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c80e0>, kernel=<OpOverload(op='aten.fft_ifftn', overload='default')>), <OpOverload(op='aten.fft_rfftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c94e0>, kernel=<OpOverload(op='aten.fft_rfftn', overload='default')>), <OpOverload(op='aten.fft_rfftn', overload='out')>: <function rfftn at 0x7776e19d8e00>, <OpOverload(op='aten.fft_ihfftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8d60>, kernel=<OpOverload(op='aten.fft_ihfftn', overload='default')>), <OpOverload(op='aten.fft_ihfftn', overload='out')>: <function ihfftn at 0x7776e19d9080>, <OpOverload(op='aten.fft_irfftn', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8cc0>, kernel=<OpOverload(op='aten.fft_irfftn', overload='default')>), <OpOverload(op='aten.fft_irfftn', overload='out')>: <function irfftn at 0x7776e19d9760>, <OpOverload(op='aten.scatter_', overload='src')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_', overload='value')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_', overload='reduce')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_', overload='value_reduce')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc900>, <OpOverload(op='aten.scatter_add_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbc680>, <OpOverload(op='aten.scatter_reduce_', overload='two')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbe020>, <OpOverload(op='aten.silu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbe160>, <OpOverload(op='aten.is_complex', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b7b00>, kernel=<OpOverload(op='aten.is_complex', overload='default')>), <OpOverload(op='aten.erfinv', overload='default')>: <function erfinv at 0x7776e1c3f7e0>, <OpOverload(op='aten.erfinv', overload='out')>: <function erfinv at 0x7776e1c3f7e0>, <OpOverload(op='aten.zero', overload='default')>: <function zero at 0x7776e1c4d260>, <OpOverload(op='aten.zero', overload='out')>: <function zero at 0x7776e1c4d260>, <OpOverload(op='aten.block_diag', overload='out')>: <function _block_diag_iterable at 0x7776e1b10b80>, <OpOverload(op='aten.frac', overload='out')>: <function frac at 0x7776e1c4dc60>, <OpOverload(op='aten.isinf', overload='out')>: <function isinf at 0x7776e1c4e660>, <OpOverload(op='aten.isposinf', overload='out')>: <function isposinf at 0x7776e1c4d800>, <OpOverload(op='aten.isneginf', overload='out')>: <function isneginf at 0x7776e1c4e980>, <OpOverload(op='aten.isnan', overload='out')>: <function isnan at 0x7776e1c4ee80>, <OpOverload(op='aten.i0', overload='default')>: <function i0 at 0x7776e1c4f880>, <OpOverload(op='aten.i0', overload='out')>: <function i0 at 0x7776e1c4f880>, <OpOverload(op='aten.index_fill_', overload='Dimname_Scalar')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.logsumexp', overload='names_out')>: <function logsumexp at 0x7776e1c65760>, <OpOverload(op='aten.logsumexp', overload='names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c3240>, kernel=<OpOverload(op='aten.logsumexp', overload='names')>), <OpOverload(op='aten.logsumexp', overload='out')>: <function logsumexp at 0x7776e1c65760>, <OpOverload(op='aten.nan_to_num', overload='default')>: <function nan_to_num at 0x7776e1c4d1c0>, <OpOverload(op='aten.nan_to_num', overload='out')>: <function nan_to_num at 0x7776e1c4d1c0>, <OpOverload(op='aten.sigmoid', overload='out')>: <function sigmoid at 0x7776e1c66de0>, <OpOverload(op='aten.std_mean', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b4c20>, kernel=<OpOverload(op='aten.std_mean', overload='dim')>), <OpOverload(op='aten.sgn', overload='default')>: <function sgn at 0x7776e1c672e0>, <OpOverload(op='aten.sgn', overload='out')>: <function sgn at 0x7776e1c672e0>, <OpOverload(op='aten.sinc', overload='out')>: <function sinc at 0x7776e1c74720>, <OpOverload(op='aten.std', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0b6de0>, kernel=<OpOverload(op='aten.std', overload='names_dim')>), <OpOverload(op='aten.bitwise_right_shift', overload='Tensor_Scalar_out')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_right_shift', overload='Scalar_Tensor_out')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.bitwise_left_shift', overload='Tensor_out')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.xlogy', overload='OutTensor')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.bitwise_left_shift', overload='Tensor_Scalar_out')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_left_shift', overload='Scalar_Tensor_out')>: <function bitwise_left_shift at 0x7776e1c76fc0>, <OpOverload(op='aten.bitwise_right_shift', overload='Tensor_out')>: <function bitwise_right_shift at 0x7776e1c77880>, <OpOverload(op='aten.copysign', overload='Tensor')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.copysign', overload='Scalar')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.copysign', overload='out')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.copysign', overload='Scalar_out')>: <function copysign at 0x7776e1c77a60>, <OpOverload(op='aten.lerp_', overload='Tensor')>: <function lerp at 0x7776e1b66e80>, <OpOverload(op='aten.trace', overload='out')>: <function trace at 0x7776e1b40f40>, <OpOverload(op='aten.heaviside', overload='out')>: <function heaviside at 0x7776e1c96ca0>, <OpOverload(op='aten.logical_and', overload='out')>: <function logical_and at 0x7776e1cad120>, <OpOverload(op='aten.std', overload='correction_out')>: <function std at 0x7776e1ad2fc0>, <OpOverload(op='aten.lcm', overload='default')>: <function lcm at 0x7776e1c97f60>, <OpOverload(op='aten.lcm', overload='out')>: <function lcm at 0x7776e1c97f60>, <OpOverload(op='aten.squeeze_copy', overload='dim')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.squeeze at 0x7776e1b98d60>, <OpOverload(op='aten.logaddexp', overload='out')>: <function logaddexp at 0x7776e1cac860>, <OpOverload(op='aten.logaddexp2', overload='out')>: <function logaddexp2 at 0x7776e1caccc0>, <OpOverload(op='aten.logical_not', overload='out')>: <function logical_not at 0x7776e1cad080>, <OpOverload(op='aten.logical_or', overload='out')>: <function logical_or at 0x7776e1cad300>, <OpOverload(op='aten.logical_xor', overload='out')>: <function logical_xor at 0x7776e1cad760>, <OpOverload(op='aten.rsub', overload='Tensor')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.rsub', overload='Scalar')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.rsub', overload='Tensor_out')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.rsub', overload='Scalar_out')>: <function rsub at 0x7776e1caf880>, <OpOverload(op='aten.xlogy', overload='OutScalar_Other')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.xlogy', overload='OutScalar_Self')>: <function xlogy at 0x7776e1cafce0>, <OpOverload(op='aten.addcdiv', overload='out')>: <function addcdiv at 0x7776e1ad04a0>, <OpOverload(op='aten.addcmul', overload='out')>: <function addcmul at 0x7776e1ad0860>, <OpOverload(op='aten.clamp', overload='out')>: <function clamp at 0x7776e1ad0c20>, <OpOverload(op='aten.clamp', overload='Tensor_out')>: <function clamp at 0x7776e1ad0c20>, <OpOverload(op='aten.clamp_min', overload='out')>: <function clamp_min at 0x7776e1ad0ea0>, <OpOverload(op='aten.clamp_min', overload='Tensor_out')>: <function clamp_min at 0x7776e1ad0ea0>, <OpOverload(op='aten.frexp', overload='Tensor_out')>: <function frexp at 0x7776e1c95f80>, <OpOverload(op='aten._adaptive_avg_pool2d', overload='default')>: <function adaptive_avg_pool2d at 0x7776e1d5a700>, <OpOverload(op='aten.relu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd440>, <OpOverload(op='aten.addmm', overload='dtype_out')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.sigmoid_backward', overload='grad_input')>: <function sigmoid_backward at 0x7776e1cde020>, <OpOverload(op='aten.sigmoid_backward', overload='default')>: <function sigmoid_backward at 0x7776e1cde020>, <OpOverload(op='aten.hardswish', overload='out')>: <function hardswish at 0x7776e1cdef20>, <OpOverload(op='aten.softplus_backward', overload='default')>: <function softplus_backward at 0x7776e1cde3e0>, <OpOverload(op='aten.softplus_backward', overload='grad_input')>: <function softplus_backward at 0x7776e1cde3e0>, <OpOverload(op='aten.elu_backward', overload='default')>: <function elu_backward at 0x7776e1cddf80>, <OpOverload(op='aten.hardsigmoid', overload='out')>: <function hardsigmoid at 0x7776e1cdefc0>, <OpOverload(op='aten.hardsigmoid_backward', overload='default')>: <function hardsigmoid_backward at 0x7776e1cdf060>, <OpOverload(op='aten.hardsigmoid_backward', overload='grad_input')>: <function hardsigmoid_backward at 0x7776e1cdf060>, <OpOverload(op='aten.hardtanh_backward', overload='grad_input')>: <function hardtanh_backward at 0x7776e1cdf420>, <OpOverload(op='aten.hardswish_backward', overload='default')>: <function hardswish_backward at 0x7776e1cde200>, <OpOverload(op='aten.threshold_backward', overload='default')>: <function threshold_backward at 0x7776e1cddda0>, <OpOverload(op='aten.threshold_backward', overload='grad_input')>: <function threshold_backward at 0x7776e1cddda0>, <OpOverload(op='aten.leaky_relu_backward', overload='default')>: <function leaky_relu_backward at 0x7776e1cdfb00>, <OpOverload(op='aten.leaky_relu_backward', overload='grad_input')>: <function leaky_relu_backward at 0x7776e1cdfb00>, <OpOverload(op='aten.gelu_backward', overload='default')>: <function gelu_backward at 0x7776e1cdfec0>, <OpOverload(op='aten.gelu_backward', overload='grad_input')>: <function gelu_backward at 0x7776e1cdfec0>, <OpOverload(op='aten.mish_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0da340>, kernel=<OpOverload(op='aten.mish_backward', overload='default')>), <OpOverload(op='aten.silu', overload='out')>: <function silu at 0x7776e1cfc720>, <OpOverload(op='aten.silu_backward', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0daa20>, kernel=<OpOverload(op='aten.silu_backward', overload='default')>), <OpOverload(op='aten.silu_backward', overload='grad_input')>: <function silu_backward at 0x7776e1cfc7c0>, <OpOverload(op='aten._prelu_kernel_backward', overload='default')>: <function _prelu_kernel_backward at 0x7776e1cfcb80>, <OpOverload(op='aten.rrelu_with_noise_backward', overload='default')>: <function rrelu_with_noise_backward at 0x7776e1cfd080>, <OpOverload(op='aten.rrelu_with_noise_backward', overload='out')>: <function rrelu_with_noise_backward at 0x7776e1cfd080>, <OpOverload(op='aten.mse_loss', overload='out')>: <function mse_loss at 0x7776e1cfd8a0>, <OpOverload(op='aten.log_sigmoid_backward', overload='grad_input')>: <function log_sigmoid_backward at 0x7776e1cfd120>, <OpOverload(op='aten.smooth_l1_loss_backward', overload='default')>: <function smooth_l1_loss_backward at 0x7776e1cfd9e0>, <OpOverload(op='aten.mse_loss_backward', overload='default')>: <function mse_loss_backward at 0x7776e1cdf240>, <OpOverload(op='aten.mse_loss_backward', overload='grad_input')>: <function mse_loss_backward at 0x7776e1cdf240>, <OpOverload(op='aten._safe_softmax', overload='default')>: <function safe_softmax at 0x7776e1cfcd60>, <OpOverload(op='aten.smooth_l1_loss_backward', overload='grad_input')>: <function smooth_l1_loss_backward_out at 0x7776e1cfdbc0>, <OpOverload(op='aten.smooth_l1_loss', overload='default')>: <function smooth_l1_loss at 0x7776e1cfd940>, <OpOverload(op='aten.smooth_l1_loss', overload='out')>: <function smooth_l1_loss at 0x7776e1cfd940>, <OpOverload(op='aten.huber_loss_backward', overload='default')>: <function huber_loss_backward at 0x7776e1cfdda0>, <OpOverload(op='aten.huber_loss_backward', overload='out')>: <function huber_loss_backward_out at 0x7776e1cfdf80>, <OpOverload(op='aten.glu_backward', overload='default')>: <function glu_backward at 0x7776e1cfe200>, <OpOverload(op='aten.glu_backward', overload='grad_input')>: <function glu_backward at 0x7776e1cfe200>, <OpOverload(op='aten.nll_loss_backward', overload='grad_input')>: <function nll_loss_backward at 0x7776e1cfe5c0>, <OpOverload(op='aten.nll_loss2d_backward', overload='default')>: <function nll_loss2d_backward at 0x7776e1cfe8e0>, <OpOverload(op='aten.nll_loss2d_backward', overload='grad_input')>: <function nll_loss2d_backward at 0x7776e1cfe8e0>, <OpOverload(op='aten.binary_cross_entropy', overload='default')>: <function binary_cross_entropy at 0x7776e1cfee80>, <OpOverload(op='aten.binary_cross_entropy', overload='out')>: <function binary_cross_entropy at 0x7776e1cfee80>, <OpOverload(op='aten.binary_cross_entropy_backward', overload='grad_input')>: <function binary_cross_entropy_backward at 0x7776e1cfef20>, <OpOverload(op='aten.soft_margin_loss_backward', overload='default')>: <function soft_margin_loss_backward at 0x7776e1cff4c0>, <OpOverload(op='aten.soft_margin_loss', overload='default')>: <function soft_margin_loss at 0x7776e1cff560>, <OpOverload(op='aten.unfold_backward', overload='out')>: <function unfold_backward at 0x7776e1cfdd00>, <OpOverload(op='aten.dist', overload='default')>: <function dist at 0x7776e1cfdb20>, <OpOverload(op='aten.dist', overload='out')>: <function dist at 0x7776e1cfdb20>, <OpOverload(op='aten._euclidean_dist', overload='default')>: <function _euclidean_dist at 0x7776e1cff2e0>, <OpOverload(op='aten._euclidean_dist', overload='out')>: <function _euclidean_dist at 0x7776e1cff2e0>, <OpOverload(op='aten.slice_backward', overload='out')>: <function slice_backward at 0x7776e1cff7e0>, <OpOverload(op='aten.slice_backward', overload='default')>: <function slice_backward at 0x7776e1cff7e0>, <OpOverload(op='aten.select_backward', overload='default')>: <function select_backward at 0x7776e1cffec0>, <OpOverload(op='aten.select_backward', overload='out')>: <function select_backward at 0x7776e1cffec0>, <OpOverload(op='aten.diagonal_backward', overload='default')>: <function diagonal_backward at 0x7776e1d28180>, <OpOverload(op='aten.logit_backward', overload='default')>: <function logit_backward at 0x7776e1cfea20>, <OpOverload(op='aten._softmax_backward_data', overload='default')>: <function _softmax_backward_data at 0x7776e1d282c0>, <OpOverload(op='aten._softmax_backward_data', overload='out')>: <function _softmax_backward_data at 0x7776e1d282c0>, <OpOverload(op='aten._log_softmax_backward_data', overload='default')>: <function _log_softmax_backward_data at 0x7776e1d28900>, <OpOverload(op='aten._log_softmax_backward_data', overload='out')>: <function _log_softmax_backward_data at 0x7776e1d28900>, <OpOverload(op='aten.im2col', overload='out')>: <function im2col at 0x7776e1d28c20>, <OpOverload(op='aten.native_dropout_backward', overload='out')>: <function native_dropout_backward at 0x7776e1d29260>, <OpOverload(op='aten.col2im', overload='out')>: <function col2im at 0x7776e1d28fe0>, <OpOverload(op='aten.native_layer_norm_backward', overload='default')>: <function native_layer_norm_backward at 0x7776e1d29080>, <OpOverload(op='aten.lift', overload='out')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.native_dropout_backward', overload='default')>: <function native_dropout_backward at 0x7776e1d29260>, <OpOverload(op='aten.native_dropout', overload='out')>: <function native_dropout at 0x7776e1d29620>, <OpOverload(op='aten._softmax', overload='out')>: <function _softmax at 0x7776e1d299e0>, <OpOverload(op='aten._log_softmax', overload='out')>: <function _log_softmax at 0x7776e1d29c60>, <OpOverload(op='aten.embedding', overload='out')>: <function embedding at 0x7776e1d29ee0>, <OpOverload(op='aten._chunk_cat', overload='default')>: <function _chunk_cat at 0x7776e1d2a480>, <OpOverload(op='aten._chunk_cat', overload='out')>: <function _chunk_cat at 0x7776e1d2a480>, <OpOverload(op='aten.embedding_dense_backward', overload='default')>: <function embedding_dense_backward at 0x7776e1d2a160>, <OpOverload(op='aten.embedding_dense_backward', overload='out')>: <function embedding_dense_backward at 0x7776e1d2a160>, <OpOverload(op='aten.split_with_sizes_copy', overload='default')>: <function split_with_sizes_copy at 0x7776e1d2a520>, <OpOverload(op='aten.split_with_sizes_copy', overload='out')>: <function split_with_sizes_copy at 0x7776e1d2a520>, <OpOverload(op='aten._addmm_activation', overload='default')>: <function _addmm_activation at 0x7776e1d2aac0>, <OpOverload(op='aten.unsafe_split_with_sizes', overload='default')>: <function unsafe_split_with_sizes at 0x7776e1d2a7a0>, <OpOverload(op='aten.native_group_norm_backward', overload='out')>: <function native_group_norm_backward_out at 0x7776e1d29bc0>, <OpOverload(op='aten.addmm', overload='out')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.addmm', overload='dtype')>: <function addmm at 0x7776e1d2af20>, <OpOverload(op='aten.native_group_norm_backward', overload='default')>: <function native_group_norm_backward at 0x7776e1d29e40>, <OpOverload(op='aten.native_batch_norm_backward', overload='default')>: <function native_batch_norm_backward at 0x7776e1d59300>, <OpOverload(op='aten._fused_rms_norm_backward', overload='default')>: <function _fused_rms_norm_backward at 0x7776e1d2b240>, <OpOverload(op='aten.native_batch_norm_backward', overload='out')>: <function native_batch_norm_backward_out at 0x7776e1d59440>, <OpOverload(op='aten.native_batch_norm', overload='out')>: <function native_batch_norm at 0x7776e1d2ba60>, <OpOverload(op='aten._batch_norm_with_update', overload='default')>: <function _batch_norm_with_update at 0x7776e1d58400>, <OpOverload(op='aten._batch_norm_with_update_functional', overload='default')>: <function _batch_norm_with_update_functional at 0x7776e1d584a0>, <OpOverload(op='aten._batch_norm_no_update', overload='default')>: <function _batch_norm_no_update at 0x7776e1d585e0>, <OpOverload(op='aten.batch_norm_backward', overload='default')>: <function batch_norm_backward at 0x7776e1d580e0>, <OpOverload(op='aten._to_copy', overload='out')>: <function _to_copy at 0x7776e1d59260>, <OpOverload(op='aten._fused_dropout', overload='default')>: <function _fused_dropout_decomposition at 0x7776e1d58ea0>, <OpOverload(op='aten._fused_dropout', overload='out')>: <function _fused_dropout_decomposition at 0x7776e1d58ea0>, <OpOverload(op='aten.cudnn_batch_norm', overload='out')>: <function cudnn_batch_norm at 0x7776e1d2a5c0>, <OpOverload(op='aten.index_copy_', overload='default')>: <function index_copy_ at 0x7776e1d5ade0>, <OpOverload(op='aten.index_copy_', overload='dimname')>: <function index_copy_ at 0x7776e1d5ade0>, <OpOverload(op='aten.miopen_batch_norm_backward', overload='default')>: <function miopen_batch_norm_backward at 0x7776e1d59bc0>, <OpOverload(op='aten.miopen_batch_norm_backward', overload='out')>: <function miopen_batch_norm_backward at 0x7776e1d59bc0>, <OpOverload(op='aten.cudnn_batch_norm_backward', overload='default')>: <function cudnn_batch_norm_backward at 0x7776e1d5a2a0>, <OpOverload(op='aten.cudnn_batch_norm_backward', overload='out')>: <function cudnn_batch_norm_backward at 0x7776e1d5a2a0>, <OpOverload(op='aten._upsample_nearest_exact3d', overload='out')>: <function _upsample_nearest_exact3d at 0x7776e1d5ae80>, <OpOverload(op='aten.max_unpool2d', overload='default')>: <function max_unpool2d at 0x7776e1d5aa20>, <OpOverload(op='aten.max_unpool2d', overload='out')>: <function max_unpool2d at 0x7776e1d5aa20>, <OpOverload(op='aten.index_add', overload='out')>: <function index_add at 0x7776e1d5b060>, <OpOverload(op='aten.index_add', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c8a40>, kernel=<OpOverload(op='aten.index_add', overload='dimname')>), <OpOverload(op='aten.pad_sequence', overload='default')>: <function pad_sequence at 0x7776e1d5ac00>, <OpOverload(op='aten.index_add', overload='default')>: <function index_add at 0x7776e1d5b060>, <OpOverload(op='aten.max_unpool3d', overload='default')>: <function max_unpool3d at 0x7776e1d5aca0>, <OpOverload(op='aten.max_unpool3d', overload='out')>: <function max_unpool3d at 0x7776e1d5aca0>, <OpOverload(op='aten.index_add_', overload='default')>: <function index_add_ at 0x7776e1d5aac0>, <OpOverload(op='aten._upsample_nearest_exact3d', overload='default')>: <function _upsample_nearest_exact3d at 0x7776e1d5ae80>, <OpOverload(op='aten.index_copy', overload='default')>: <function index_copy at 0x7776e1d58220>, <OpOverload(op='aten.index_copy', overload='dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2d40>, kernel=<OpOverload(op='aten.index_copy', overload='dimname')>), <OpOverload(op='aten.index_copy', overload='out')>: <function index_copy at 0x7776e1d58220>, <OpOverload(op='aten.log_sigmoid_forward', overload='default')>: <function log_sigmoid_forward at 0x7776e1d5b6a0>, <OpOverload(op='aten.log_sigmoid_forward', overload='output')>: <function log_sigmoid_forward at 0x7776e1d5b6a0>, <OpOverload(op='aten._upsample_nearest_exact1d', overload='vec')>: <function _upsample_nearest_exact_vec at 0x7776e1d78680>, <OpOverload(op='aten.uniform_', overload='default')>: <function uniform_ at 0x7776e1d5b880>, <OpOverload(op='aten._upsample_nearest_exact1d', overload='default')>: <function upsample_nearest_exact1d at 0x7776e1d78d60>, <OpOverload(op='aten.rnn_tanh', overload='data')>: <function rnn_tanh_data at 0x7776e1d79d00>, <OpOverload(op='aten.upsample_nearest1d', overload='out')>: <function upsample_nearest1d at 0x7776e1d78a40>, <OpOverload(op='aten._upsample_nearest_exact2d', overload='vec')>: <function _upsample_nearest_exact_vec at 0x7776e1d78680>, <OpOverload(op='aten._upsample_nearest_exact3d', overload='vec')>: <function _upsample_nearest_exact_vec at 0x7776e1d78680>, <OpOverload(op='aten._upsample_nearest_exact1d', overload='out')>: <function upsample_nearest_exact1d at 0x7776e1d78d60>, <OpOverload(op='aten.rnn_relu', overload='input')>: <function rnn_relu_input at 0x7776e1d79940>, <OpOverload(op='aten._upsample_nearest_exact2d', overload='default')>: <function _upsample_nearest_exact2d at 0x7776e1d793a0>, <OpOverload(op='aten.rnn_relu', overload='data')>: <function rnn_relu_data at 0x7776e1d79b20>, <OpOverload(op='aten.upsample_nearest2d', overload='out')>: <function upsample_nearest2d at 0x7776e1d79080>, <OpOverload(op='aten._upsample_nearest_exact2d', overload='out')>: <function _upsample_nearest_exact2d at 0x7776e1d793a0>, <OpOverload(op='aten.upsample_nearest3d', overload='out')>: <function upsample_nearest3d at 0x7776e1d5bd80>, <OpOverload(op='aten.lstm', overload='input')>: <function lstm_impl at 0x7776e1d7a160>, <OpOverload(op='aten.lstm', overload='data')>: <function lstm_data_impl at 0x7776e1d7a340>, <OpOverload(op='aten._upsample_bilinear2d_aa', overload='vec')>: <function upsample_bilinear2d_aa_vec at 0x7776e1d7aa20>, <OpOverload(op='aten.gru', overload='data')>: <function gru_impl_data at 0x7776e1d7a660>, <OpOverload(op='aten.gru', overload='input')>: <function gru_impl at 0x7776e1d7a840>, <OpOverload(op='aten._upsample_bicubic2d_aa', overload='vec')>: <function upsample_bicubic2d_aa_vec at 0x7776e1d7ac00>, <OpOverload(op='aten.upsample_trilinear3d', overload='out')>: <function upsample_trilinear3d at 0x7776e1d7b880>, <OpOverload(op='aten.upsample_linear1d', overload='out')>: <function upsample_linear1d at 0x7776e1d7b2e0>, <OpOverload(op='aten.upsample_bilinear2d', overload='out')>: <function upsample_bilinear2d at 0x7776e1d7b600>, <OpOverload(op='aten.nll_loss_forward', overload='output')>: <function nll_loss_forward at 0x7776e1d7a7a0>, <OpOverload(op='aten.is_same_size', overload='default')>: <function is_same_size at 0x7776e1d7bc40>, <OpOverload(op='aten._reshape_alias', overload='default')>: <function _reshape_alias at 0x7776e1d7bf60>, <OpOverload(op='aten._unsafe_view', overload='out')>: <function _reshape_alias at 0x7776e1d7bf60>, <OpOverload(op='aten._unsafe_masked_index', overload='default')>: <function _unsafe_masked_index at 0x7776e1d94220>, <OpOverload(op='aten._unsafe_masked_index_put_accumulate', overload='default')>: <function _unsafe_masked_index_put_accumulate at 0x7776e1d94360>, <OpOverload(op='aten.nll_loss2d_forward', overload='default')>: <function nll_loss2d_forward at 0x7776e1d7a0c0>, <OpOverload(op='aten.affine_grid_generator', overload='default')>: <function affine_grid_generator at 0x7776e1d94ea0>, <OpOverload(op='aten.affine_grid_generator', overload='out')>: <function affine_grid_generator at 0x7776e1d94ea0>, <OpOverload(op='aten.grid_sampler_2d', overload='out')>: <function grid_sampler_2d at 0x7776e1d95260>, <OpOverload(op='aten.binary_cross_entropy_with_logits', overload='default')>: <function binary_cross_entropy_with_logits at 0x7776e1d95800>, <OpOverload(op='aten.binary_cross_entropy_with_logits', overload='out')>: <function binary_cross_entropy_with_logits at 0x7776e1d95800>, <OpOverload(op='aten.upsample_bicubic2d', overload='out')>: <function upsample_bicubic2d_default at 0x7776e1d95f80>, <OpOverload(op='aten.reflection_pad1d', overload='out')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.__ilshift__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdda0>, <OpOverload(op='aten.reflection_pad3d', overload='out')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.reflection_pad2d', overload='out')>: <function _reflection_pad at 0x7776e1d96520>, <OpOverload(op='aten.replication_pad1d', overload='out')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.replication_pad2d', overload='out')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.replication_pad3d', overload='out')>: <function _replication_pad at 0x7776e1d967a0>, <OpOverload(op='aten.reflection_pad1d_backward', overload='default')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.reflection_pad2d_backward', overload='default')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.reflection_pad1d_backward', overload='grad_input')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.reflection_pad3d_backward', overload='default')>: <function _reflection_pad_backward at 0x7776e1d96840>, <OpOverload(op='aten.arange', overload='out')>: <function arange_default at 0x7776e1d976a0>, <OpOverload(op='aten.aminmax', overload='default')>: <function aminmax at 0x7776e1d97060>, <OpOverload(op='aten.aminmax', overload='out')>: <function aminmax at 0x7776e1d97060>, <OpOverload(op='aten.multilabel_margin_loss_forward', overload='output')>: <function multilabel_margin_loss_forward at 0x7776e1d97880>, <OpOverload(op='aten.nansum', overload='default')>: <function nansum at 0x7776e1d97420>, <OpOverload(op='aten.nansum', overload='out')>: <function nansum at 0x7776e1d97420>, <OpOverload(op='aten.multilabel_margin_loss_forward', overload='default')>: <function multilabel_margin_loss_forward at 0x7776e1d97880>, <OpOverload(op='aten.multi_margin_loss', overload='default')>: <function multi_margin_loss at 0x7776e1d97ce0>, <OpOverload(op='aten.multi_margin_loss', overload='out')>: <function multi_margin_loss at 0x7776e1d97ce0>, <OpOverload(op='aten.baddbmm', overload='dtype_out')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.baddbmm', overload='out')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.baddbmm', overload='dtype')>: <function baddbmm at 0x7776e1d96980>, <OpOverload(op='aten.__irshift__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbde40>, <OpOverload(op='aten.floor_divide', overload='Scalar')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.floor_divide', overload='out')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.floor_divide', overload='Scalar_out')>: <function floor_divide at 0x7776e1d95b20>, <OpOverload(op='aten.__irshift__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbde40>, <OpOverload(op='aten._weight_norm_interface', overload='default')>: <function _weight_norm_interface at 0x7776e1dbc720>, <OpOverload(op='aten._weight_norm_interface', overload='out')>: <function _weight_norm_interface at 0x7776e1dbc720>, <OpOverload(op='aten.isin', overload='Tensor_Tensor')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Tensor_Scalar')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Scalar_Tensor')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Tensor_Tensor_out')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Tensor_Scalar_out')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.isin', overload='Scalar_Tensor_out')>: <function isin at 0x7776e1dbcb80>, <OpOverload(op='aten.take', overload='default')>: <function take at 0x7776e1dbd080>, <OpOverload(op='aten.take', overload='out')>: <function take at 0x7776e1dbd080>, <OpOverload(op='aten.resize_as', overload='default')>: <function resize_as at 0x7776e1dbcea0>, <OpOverload(op='aten.resize_as', overload='out')>: <function resize_as at 0x7776e1dbcea0>, <OpOverload(op='aten.__ior__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdee0>, <OpOverload(op='aten.__ior__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdee0>, <OpOverload(op='aten.addbmm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd120>, <OpOverload(op='aten.addmm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd260>, <OpOverload(op='aten.addmv_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd3a0>, <OpOverload(op='aten.baddbmm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd4e0>, <OpOverload(op='aten.fill_', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd620>, <OpOverload(op='aten.fill_', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd620>, <OpOverload(op='aten.gelu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd760>, <OpOverload(op='aten.index_reduce_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1d977e0>, <OpOverload(op='aten.hardswish_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd8a0>, <OpOverload(op='aten.hardtanh_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd9e0>, <OpOverload(op='aten.hardsigmoid_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdb20>, <OpOverload(op='aten.__iand__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdc60>, <OpOverload(op='aten.__iand__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdc60>, <OpOverload(op='aten.__ilshift__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdda0>, <OpOverload(op='aten.__ixor__', overload='Tensor')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdbc0>, <OpOverload(op='aten.__ixor__', overload='Scalar')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbdbc0>, <OpOverload(op='aten.leaky_relu_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd940>, <OpOverload(op='aten.logit_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd6c0>, <OpOverload(op='aten.renorm_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbd1c0>, <OpOverload(op='aten.round_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbccc0>, <OpOverload(op='aten.round_', overload='decimals')>: <function register_inplace.<locals>.inplace_op at 0x7776e1dbccc0>, <OpOverload(op='aten.ne', overload='Scalar_out')>: <function ne at 0x7776e1caed40>, <OpOverload(op='aten.ne', overload='Tensor_out')>: <function ne at 0x7776e1caed40>, <OpOverload(op='aten.nextafter', overload='out')>: <function nextafter at 0x7776e1caf1a0>, <OpOverload(op='aten.nextafter', overload='default')>: <function nextafter at 0x7776e1caf1a0>, <OpOverload(op='aten.pow', overload='Tensor_Scalar_out')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Scalar_out')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.pow', overload='Tensor_Tensor_out')>: <function pow at 0x7776e1c945e0>, <OpOverload(op='aten.remainder', overload='Scalar_Tensor')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Scalar_Tensor_out')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Scalar_out')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.remainder', overload='Tensor_out')>: <function remainder at 0x7776e1caf600>, <OpOverload(op='aten.sub', overload='out')>: <function sub at 0x7776e1caede0>, <OpOverload(op='aten.sub', overload='Scalar_out')>: <function sub at 0x7776e1caede0>, <OpOverload(op='aten.reflection_pad3d_backward', overload='grad_input')>: <function _reflection_pad_backward at 0x7776e1d96840>, <OpOverload(op='aten.unfold_backward', overload='default')>: <function unfold_backward at 0x7776e1cfdd00>, <OpOverload(op='aten.svd', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1800>, kernel=<OpOverload(op='aten.svd', overload='default')>), <OpOverload(op='aten._addmm_activation', overload='out')>: <function _addmm_activation at 0x7776e1d2aac0>, <OpOverload(op='aten.squeeze', overload='dims')>: <function squeeze at 0x7776e1af7ce0>, <OpOverload(op='aten.elu_backward', overload='grad_input')>: <function elu_backward at 0x7776e1cddf80>, <OpOverload(op='aten.as_strided_scatter', overload='out')>: <function as_strided_scatter at 0x7776e1af40e0>, <OpOverload(op='aten.fft_ifftn', overload='out')>: <function ifftn at 0x7776e19d8b80>, <OpOverload(op='aten.cat', overload='names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9da0>, kernel=<OpOverload(op='aten.cat', overload='names')>), <OpOverload(op='aten.cat', overload='names_out')>: <function cat at 0x7776e1af4720>, <OpOverload(op='aten.cat', overload='out')>: <function cat at 0x7776e1af4720>, <OpOverload(op='aten.std_mean', overload='correction')>: <function std_mean at 0x7776e1c656c0>, <OpOverload(op='aten.where', overload='self_out')>: <function where at 0x7776e1ad1760>, <OpOverload(op='aten.where', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1ee0>, kernel=<OpOverload(op='aten.where', overload='default')>), <OpOverload(op='aten.soft_margin_loss', overload='out')>: <function soft_margin_loss at 0x7776e1cff560>, <OpOverload(op='aten.upsample_linear1d', overload='vec')>: <function _upsample_linear_vec at 0x7776e1d7b100>, <OpOverload(op='aten.nll_loss_backward', overload='default')>: <function nll_loss_backward at 0x7776e1cfe5c0>, <OpOverload(op='aten.sum', overload='IntList_out')>: <function sum at 0x7776e1ad22a0>, <OpOverload(op='aten.sum', overload='dim_DimnameList')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d2480>, kernel=<OpOverload(op='aten.sum', overload='dim_DimnameList')>), <OpOverload(op='aten.soft_margin_loss_backward', overload='grad_input')>: <function soft_margin_loss_backward at 0x7776e1cff4c0>, <OpOverload(op='aten.prod', overload='dim_Dimname')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8fe0>, kernel=<OpOverload(op='aten.prod', overload='dim_Dimname')>), <OpOverload(op='aten.prod', overload='int_out')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='Dimname_out')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.prod', overload='out')>: <function prod at 0x7776e1ad2660>, <OpOverload(op='aten.var', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d8c20>, kernel=<OpOverload(op='aten.var', overload='default')>), <OpOverload(op='aten.var', overload='out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='correction')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9120>, kernel=<OpOverload(op='aten.var', overload='dim')>), <OpOverload(op='aten.normal', overload='float_float_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.var', overload='correction_out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='names_dim')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d96c0>, kernel=<OpOverload(op='aten.var', overload='names_dim')>), <OpOverload(op='aten.var', overload='names_out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.var', overload='correction_names')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d9620>, kernel=<OpOverload(op='aten.var', overload='correction_names')>), <OpOverload(op='aten.var', overload='correction_names_out')>: <function var at 0x7776e1ad2d40>, <OpOverload(op='aten.amax', overload='out')>: <function amax at 0x7776e1ad2840>, <OpOverload(op='aten.nll_loss2d_forward', overload='output')>: <function nll_loss2d_forward at 0x7776e1d7a0c0>, <OpOverload(op='aten.amin', overload='out')>: <function amin at 0x7776e1ad2700>, <OpOverload(op='aten.lift', overload='default')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.empty_strided', overload='out')>: <function empty_strided at 0x7776e1b13d80>, <OpOverload(op='aten.reflection_pad2d_backward', overload='grad_input')>: <function _reflection_pad_backward at 0x7776e1d96a20>, <OpOverload(op='aten.full', overload='out')>: <function full at 0x7776e1b12de0>, <OpOverload(op='aten._adaptive_avg_pool2d', overload='out')>: <function adaptive_avg_pool2d at 0x7776e1d5a700>, <OpOverload(op='aten.diagonal_backward', overload='out')>: <function diagonal_backward at 0x7776e1d28180>, <OpOverload(op='aten.native_layer_norm_backward', overload='out')>: <function native_layer_norm_backward_out at 0x7776e1d2b100>, <OpOverload(op='aten.normal', overload='Tensor_float_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='float_Tensor_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.normal', overload='Tensor_Tensor_out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.uniform', overload='default')>: <function uniform at 0x7776e1d5ba60>, <OpOverload(op='aten.uniform', overload='out')>: <function uniform at 0x7776e1d5ba60>, <OpOverload(op='aten.normal', overload='out')>: <function normal at 0x7776e1b42b60>, <OpOverload(op='aten.mv', overload='out')>: <function mv at 0x7776e1d95580>, <OpOverload(op='aten.index_fill_', overload='int_Tensor')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.frexp', overload='Tensor')>: <function frexp at 0x7776e1c95f80>, <OpOverload(op='aten.lgamma', overload='out')>: <function lgamma at 0x7776e1c4fd80>, <OpOverload(op='aten.log', overload='out')>: <function log at 0x7776e1c642c0>, <OpOverload(op='aten.bitwise_xor', overload='Tensor_out')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.log1p', overload='out')>: <function log1p at 0x7776e1c647c0>, <OpOverload(op='aten.log10', overload='out')>: <function log10 at 0x7776e1c651c0>, <OpOverload(op='aten.reciprocal', overload='out')>: <function reciprocal at 0x7776e1c66020>, <OpOverload(op='aten.neg', overload='out')>: <function neg at 0x7776e1c659e0>, <OpOverload(op='aten.log_sigmoid_backward', overload='default')>: <function log_sigmoid_backward at 0x7776e1cfd120>, <OpOverload(op='aten.tan', overload='out')>: <function tan at 0x7776e1c753a0>, <OpOverload(op='aten.round', overload='decimals_out')>: <function round at 0x7776e1c663e0>, <OpOverload(op='aten.round', overload='out')>: <function round at 0x7776e1c663e0>, <OpOverload(op='aten.tanh_backward', overload='grad_input')>: <function tanh_backward at 0x7776e2122840>, <OpOverload(op='aten.rsqrt', overload='out')>: <function rsqrt at 0x7776e1c668e0>, <OpOverload(op='aten.sign', overload='out')>: <function sign at 0x7776e1c677e0>, <OpOverload(op='aten.signbit', overload='default')>: <function signbit at 0x7776e1c67ce0>, <OpOverload(op='aten.signbit', overload='out')>: <function signbit at 0x7776e1c67ce0>, <OpOverload(op='aten.sin', overload='out')>: <function sin at 0x7776e1c74220>, <OpOverload(op='aten.bitwise_and', overload='Tensor_out')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.cudnn_batch_norm', overload='default')>: <function cudnn_batch_norm at 0x7776e1d2a5c0>, <OpOverload(op='aten.sqrt', overload='out')>: <function sqrt at 0x7776e1c749a0>, <OpOverload(op='aten.tanh', overload='out')>: <function tanh at 0x7776e1c758a0>, <OpOverload(op='aten.trunc', overload='out')>: <function trunc at 0x7776e1c75da0>, <OpOverload(op='aten.add', overload='out')>: <function add at 0x7776e1c762a0>, <OpOverload(op='aten.add', overload='Scalar_out')>: <function add at 0x7776e1c762a0>, <OpOverload(op='aten.narrow_copy', overload='out')>: <function pybind11_detail_function_record_v1_system_libstdcpp_gxx_abi_1xxx_use_cxx11_abi_1.narrow at 0x7776e1b98a40>, <OpOverload(op='aten.atan2', overload='out')>: <function atan2 at 0x7776e1c76700>, <OpOverload(op='aten.bitwise_and', overload='Scalar_out')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_and', overload='Scalar_Tensor_out')>: <function bitwise_and at 0x7776e1c76b60>, <OpOverload(op='aten.bitwise_or', overload='Tensor_out')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_or', overload='Scalar_Tensor_out')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_or', overload='Scalar_out')>: <function bitwise_or at 0x7776e1c77420>, <OpOverload(op='aten.bitwise_xor', overload='Scalar_Tensor_out')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.bitwise_xor', overload='Scalar_out')>: <function bitwise_xor at 0x7776e1c767a0>, <OpOverload(op='aten.div', overload='out')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='out_mode')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='Scalar_out')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.div', overload='Scalar_mode_out')>: <function div at 0x7776e1c77ce0>, <OpOverload(op='aten.eq', overload='Scalar_out')>: <function eq at 0x7776e1c94180>, <OpOverload(op='aten.eq', overload='Tensor_out')>: <function eq at 0x7776e1c94180>, <OpOverload(op='aten.fmax', overload='out')>: <function fmax at 0x7776e1c95120>, <OpOverload(op='aten.binary_cross_entropy_backward', overload='default')>: <function binary_cross_entropy_backward at 0x7776e1cfef20>, <OpOverload(op='aten.fmax', overload='default')>: <function fmax at 0x7776e1c95120>, <OpOverload(op='aten.fmin', overload='default')>: <function fmin at 0x7776e1c95580>, <OpOverload(op='aten.fmin', overload='out')>: <function fmin at 0x7776e1c95580>, <OpOverload(op='aten.log10_', overload='default')>: <function log10 at 0x7776e1b67060>, <OpOverload(op='aten.fmod', overload='Tensor_out')>: <function fmod at 0x7776e1c959e0>, <OpOverload(op='aten.igammac', overload='out')>: <function igammac at 0x7776e1c979c0>, <OpOverload(op='aten.fmod', overload='Scalar_out')>: <function fmod at 0x7776e1c959e0>, <OpOverload(op='aten.gcd', overload='default')>: <function gcd at 0x7776e1c76200>, <OpOverload(op='aten.gcd', overload='out')>: <function gcd at 0x7776e1c76200>, <OpOverload(op='aten.addmv', overload='out')>: <function addmv at 0x7776e1d2a0c0>, <OpOverload(op='aten.ge', overload='Tensor_out')>: <function ge at 0x7776e1c963e0>, <OpOverload(op='aten.ge', overload='Scalar_out')>: <function ge at 0x7776e1c963e0>, <OpOverload(op='aten.hardswish_backward', overload='out')>: <function hardswish_backward at 0x7776e1cde200>, <OpOverload(op='aten.gt', overload='Scalar_out')>: <function gt at 0x7776e1c96840>, <OpOverload(op='aten.gt', overload='Tensor_out')>: <function gt at 0x7776e1c96840>, <OpOverload(op='aten.hypot', overload='default')>: <function hypot at 0x7776e1c97100>, <OpOverload(op='aten.hypot', overload='out')>: <function hypot at 0x7776e1c97100>, <OpOverload(op='aten.index_fill', overload='Dimname_Scalar')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0c2f20>, kernel=<OpOverload(op='aten.index_fill', overload='Dimname_Scalar')>), <OpOverload(op='aten.igamma', overload='out')>: <function igamma at 0x7776e1c97560>, <OpOverload(op='aten.igamma', overload='default')>: <function igamma at 0x7776e1c97560>, <OpOverload(op='aten.sum', overload='out')>: <function sum_default at 0x7776e1dbc400>, <OpOverload(op='aten.igammac', overload='default')>: <function igammac at 0x7776e1c979c0>, <OpOverload(op='aten.le', overload='Tensor_out')>: <function le at 0x7776e1cac400>, <OpOverload(op='aten.le', overload='Scalar_out')>: <function le at 0x7776e1cac400>, <OpOverload(op='aten.lt', overload='Scalar_out')>: <function lt at 0x7776e1cadbc0>, <OpOverload(op='aten.lt', overload='Tensor_out')>: <function lt at 0x7776e1cadbc0>, <OpOverload(op='aten.maximum', overload='out')>: <function maximum at 0x7776e1cae020>, <OpOverload(op='aten.minimum', overload='out')>: <function minimum at 0x7776e1cae480>, <OpOverload(op='aten.mul', overload='Scalar_out')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.rnn_tanh', overload='input')>: <function rnn_tanh_input at 0x7776e1d79760>, <OpOverload(op='aten.acosh', overload='out')>: <function acosh at 0x7776e1c3c400>, <OpOverload(op='aten.asin', overload='out')>: <function asin at 0x7776e1c3c900>, <OpOverload(op='aten.asinh', overload='out')>: <function asinh at 0x7776e1c3ce00>, <OpOverload(op='aten.lgamma', overload='default')>: <function lgamma at 0x7776e1c4fd80>, <OpOverload(op='aten.atanh', overload='out')>: <function atanh at 0x7776e1c3d800>, <OpOverload(op='aten.atan', overload='out')>: <function atan at 0x7776e1c3d300>, <OpOverload(op='aten.cos', overload='out')>: <function cos at 0x7776e1c3eac0>, <OpOverload(op='aten.cosh', overload='out')>: <function cosh at 0x7776e1c3efc0>, <OpOverload(op='aten.bitwise_not', overload='out')>: <function bitwise_not at 0x7776e1c3dd00>, <OpOverload(op='aten.index_fill_', overload='int_Scalar')>: <function index_fill_ at 0x7776e1af79c0>, <OpOverload(op='aten.ceil', overload='out')>: <function ceil at 0x7776e1c3e200>, <OpOverload(op='aten.conj_physical', overload='out')>: <function conj_physical at 0x7776e1c3e5c0>, <OpOverload(op='aten.conj_physical', overload='default')>: <function conj_physical at 0x7776e1c3e5c0>, <OpOverload(op='aten.digamma', overload='default')>: <function digamma at 0x7776e1c3f4c0>, <OpOverload(op='aten.digamma', overload='out')>: <function digamma at 0x7776e1c3f4c0>, <OpOverload(op='aten.sinh', overload='out')>: <function sinh at 0x7776e1c66980>, <OpOverload(op='aten.erf', overload='out')>: <function erf at 0x7776e1c3d3a0>, <OpOverload(op='aten.exp', overload='out')>: <function exp at 0x7776e1c4c220>, <OpOverload(op='aten.expm1', overload='out')>: <function expm1 at 0x7776e1c4c720>, <OpOverload(op='aten.exp2', overload='out')>: <function exp2 at 0x7776e1c4cc20>, <OpOverload(op='aten.transpose', overload='Dimname')>: <function transpose at 0x7776e1b10f40>, <OpOverload(op='aten.mul', overload='out')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.floor', overload='out')>: <function floor at 0x7776e1c4d760>, <OpOverload(op='aten.clone', overload='out')>: <function clone at 0x7776e1ad19e0>, <OpOverload(op='aten.acos', overload='out')>: <function acos at 0x7776e1dbfd80>, <OpOverload(op='aten.log2', overload='out')>: <function log2 at 0x7776e1c64cc0>, <OpOverload(op='aten.mul', overload='Scalar')>: <function mul at 0x7776e1cae8e0>, <OpOverload(op='aten.tanh_backward', overload='default')>: <function tanh_backward at 0x7776e2122840>, <OpOverload(op='aten.abs', overload='out')>: <function abs at 0x7776e1dbede0>, <OpOverload(op='aten.sym_numel', overload='default')>: <function sym_numel at 0x7776e1d96340>, <OpOverload(op='aten.diagonal_scatter', overload='default')>: <function diagonal_scatter at 0x7776e1b104a0>, <OpOverload(op='aten.as_strided_scatter', overload='default')>: <function as_strided_scatter at 0x7776e1af40e0>, <OpOverload(op='aten.lift_fresh', overload='default')>: <function nop_decomposition at 0x7776e1d29300>, <OpOverload(op='aten.ones_like', overload='out')>: <function ones_like at 0x7776e1b40180>, <OpOverload(op='aten.zeros_like', overload='out')>: <function zeros_like at 0x7776e1b116c0>, <OpOverload(op='aten.new_empty', overload='out')>: <function new_empty at 0x7776e1b12480>, <OpOverload(op='aten.new_empty_strided', overload='out')>: <function new_empty_strided at 0x7776e1b12700>, <OpOverload(op='aten.new_full', overload='out')>: <function new_full at 0x7776e1b12840>, <OpOverload(op='aten.new_zeros', overload='out')>: <function new_zeros at 0x7776e1b11440>, <OpOverload(op='aten.new_ones', overload='out')>: <function new_ones at 0x7776e1b10400>, <OpOverload(op='aten.item', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d1080>, kernel=<OpOverload(op='aten.item', overload='default')>), <OpOverload(op='aten.nonzero_numpy', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0d0cc0>, kernel=<OpOverload(op='aten.nonzero_numpy', overload='default')>), <OpOverload(op='aten.slice_scatter', overload='out')>: <function slice_scatter at 0x7776e1cffc40>, <OpOverload(op='aten.index_put_', overload='default')>: <function register_inplace.<locals>.inplace_op at 0x7776e1d96ac0>, <OpOverload(op='aten.erfc', overload='out')>: <function erfc at 0x7776e1c3fce0>, <OpOverload(op='aten.empty_like', overload='out')>: <function empty_like at 0x7776e1b12c00>, <OpOverload(op='quantized.make_quantized_cell_params_fp16', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dac00>, kernel=<OpOverload(op='quantized.make_quantized_cell_params_fp16', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_stride', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dade0>, kernel=<OpOverload(op='quantized.conv_transpose3d_stride', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_groups', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dae80>, kernel=<OpOverload(op='quantized.conv_transpose3d_groups', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_stride', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4f40>, kernel=<OpOverload(op='quantized.conv_transpose2d_stride', overload='default')>), <OpOverload(op='quantized.conv2d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db560>, kernel=<OpOverload(op='quantized.conv2d_unpack', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f5760>, kernel=<OpOverload(op='quantized.conv_transpose3d_padding', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_dilation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db920>, kernel=<OpOverload(op='quantized.conv_transpose3d_dilation', overload='default')>), <OpOverload(op='quantized.conv2d_dilation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f51c0>, kernel=<OpOverload(op='quantized.conv2d_dilation', overload='default')>), <OpOverload(op='quantized.conv2d_output_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f54e0>, kernel=<OpOverload(op='quantized.conv2d_output_padding', overload='default')>), <OpOverload(op='quantized.conv2d_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4360>, kernel=<OpOverload(op='quantized.conv2d_padding', overload='default')>), <OpOverload(op='quantized.conv3d_dilation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db880>, kernel=<OpOverload(op='quantized.conv3d_dilation', overload='default')>), <OpOverload(op='quantized.conv1d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0daca0>, kernel=<OpOverload(op='quantized.conv1d_unpack', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4040>, kernel=<OpOverload(op='quantized.conv_transpose3d_unpack', overload='default')>), <OpOverload(op='quantized.conv3d_groups', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0daf20>, kernel=<OpOverload(op='quantized.conv3d_groups', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_output_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dab60>, kernel=<OpOverload(op='quantized.conv_transpose3d_output_padding', overload='default')>), <OpOverload(op='quantized.conv_transpose3d_transpose', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f45e0>, kernel=<OpOverload(op='quantized.conv_transpose3d_transpose', overload='default')>), <OpOverload(op='quantized.linear_unpack_fp16', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dbb00>, kernel=<OpOverload(op='quantized.linear_unpack_fp16', overload='default')>), <OpOverload(op='quantized.conv3d_stride', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f49a0>, kernel=<OpOverload(op='quantized.conv3d_stride', overload='default')>), <OpOverload(op='quantized.conv2d_groups', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dbec0>, kernel=<OpOverload(op='quantized.conv2d_groups', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_output_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f5080>, kernel=<OpOverload(op='quantized.conv_transpose2d_output_padding', overload='default')>), <OpOverload(op='quantized.conv2d_transpose', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0daac0>, kernel=<OpOverload(op='quantized.conv2d_transpose', overload='default')>), <OpOverload(op='quantized.conv_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4220>, kernel=<OpOverload(op='quantized.conv_unpack', overload='default')>), <OpOverload(op='quantized.embedding_bag_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4720>, kernel=<OpOverload(op='quantized.embedding_bag_unpack', overload='default')>), <OpOverload(op='quantized.linear_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4400>, kernel=<OpOverload(op='quantized.linear_unpack', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_transpose', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db420>, kernel=<OpOverload(op='quantized.conv_transpose2d_transpose', overload='default')>), <OpOverload(op='quantized.conv2d_stride', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db6a0>, kernel=<OpOverload(op='quantized.conv2d_stride', overload='default')>), <OpOverload(op='quantized.conv3d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dbce0>, kernel=<OpOverload(op='quantized.conv3d_unpack', overload='default')>), <OpOverload(op='sparse.qlinear_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dba60>, kernel=<OpOverload(op='sparse.qlinear_unpack', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_dilation', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0db380>, kernel=<OpOverload(op='quantized.conv_transpose2d_dilation', overload='default')>), <OpOverload(op='quantized.conv_transpose1d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4fe0>, kernel=<OpOverload(op='quantized.conv_transpose1d_unpack', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_groups', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dbe20>, kernel=<OpOverload(op='quantized.conv_transpose2d_groups', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0dafc0>, kernel=<OpOverload(op='quantized.conv_transpose2d_padding', overload='default')>), <OpOverload(op='quantized.conv3d_output_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f5800>, kernel=<OpOverload(op='quantized.conv3d_output_padding', overload='default')>), <OpOverload(op='quantized.conv_transpose2d_unpack', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f56c0>, kernel=<OpOverload(op='quantized.conv_transpose2d_unpack', overload='default')>), <OpOverload(op='quantized.conv3d_padding', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f5440>, kernel=<OpOverload(op='quantized.conv3d_padding', overload='default')>), <OpOverload(op='quantized.conv3d_transpose', overload='default')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f53a0>, kernel=<OpOverload(op='quantized.conv3d_transpose', overload='default')>), <OpOverload(op='profiler._record_function_exit', overload='_RecordFunction')>: functools.partial(<function _get_decomp_for_cia.<locals>._special_op_to_decompose_cia at 0x77766a0f4860>, kernel=<OpOverload(op='profiler._record_function_exit', overload='_RecordFunction')>)}