Note
Go to the end to download the full example code.
Find and fix an export issue due to dynamic shapes¶
LLMs must be exported with dynamic shapes and it is common that
a static dimension turns into a static ones. The error message from
pytorch tells the user to define TORCH_LOGS="+dynamic"
but it shows a very long list of messages where we need
to find the string range_refined_to_singleton and that
does not really indicates where it comes from. The example
shows how to tweak pytorch to get that information until
it gets better.
A model with an export issue¶
The following model implies the first dimension of x is equal to 1
or equal to the number of element in the list ys.
It is not really dynamic. It looks obvious here but
it is difficult to find deep inside a big model.
import traceback
import torch
from onnx_diagnostic import doc
from onnx_diagnostic.torch_export_patches import torch_export_patches
class ModelWithIssue(torch.nn.Module):
def forward(self, x: torch.Tensor, ys: list[torch.Tensor]):
caty = torch.cat([y.unsqueeze(0) for y in ys], axis=0)
z = x * caty
return z
inputs = (torch.rand(2, 3, 1), [torch.rand(3, 4), torch.rand(3, 4)])
model = ModelWithIssue()
model(*inputs)
tensor([[[0.0519, 0.0844, 0.1746, 0.0404],
[0.5118, 0.2140, 0.5120, 0.5620],
[0.2474, 0.1650, 0.1219, 0.1728]],
[[0.6298, 0.4665, 0.4820, 0.2788],
[0.1256, 0.7841, 0.7342, 0.7996],
[0.2932, 0.7785, 0.6687, 0.1029]]])
Let’s export.
DYN = torch.export.Dim.DYNAMIC
dyn_shapes = ({0: DYN, 1: DYN}, [{0: DYN, 1: DYN}, {0: DYN, 1: DYN}])
try:
ep = torch.export.export(model, inputs, dynamic_shapes=dyn_shapes)
print(ep)
except Exception as e:
print("-- ERROR:")
print(e)
-- ERROR:
Found the following conflicts between user-specified ranges and inferred ranges from model tracing:
- Received user-specified dim hint Dim.DYNAMIC(min=None, max=None), but tracing inferred a static shape of 2 for dimension inputs['x'].shape[0].
The error shows:
Constraints violated (L['args'][0][0].size()[0])!
For more information, run with TORCH_LOGS="+dynamic".
- Not all values of RelaxedUnspecConstraint(L['args'][0][0].size()[0])
are valid because L['args'][0][0].size()[0] was inferred to be a constant (2).
Where does it happens? That’s a tricky question we need to answer. The message is raised from torch.fx.experimental.symbolic_shapes.ShapeEnv._set_replacement. One way to find the exact location is to retrieve a stack trace by inserting an assert such as the following:
assert msg != "range_refined_to_singleton", (
f"A dynamic dimension becomes static! "
f"a={a!r}, tgt={tgt!r}, msg={msg!r}, tgt_bound={tgt_bound}"
)
Stop when a dynamic dimension turns static¶
We use torch_export_patches
to replace torch implementation by a new one raising the exception
mentioned in previous section.
with torch_export_patches(stop_if_static=1, verbose=1):
try:
torch.export.export(model, inputs, dynamic_shapes=dyn_shapes)
except (AssertionError, torch._dynamo.exc.TorchRuntimeError) as e:
print("-- It failed as excepted.")
print(f"-- final error is {e}")
print("-- Stack Trace")
print(traceback.format_exc())
# The stack trace is quite long but the first line referring to this example
# is the following one. It points out the line turing a dynamic dimension into
# static.
#
# .. code-block::
#
# File "onnx-diagnostic/_doc/examples/plot_export_locate_issue.py", line 25, in forward
# z = x * caty
[torch_export_patches] replace torch.jit.isinstance, torch._dynamo.mark_static_address
[torch_export_patches] sympy.__version__='1.13.3'
[torch_export_patches] patch sympy
[torch_export_patches] torch.__version__='2.9.0.dev20250820+cu126'
[torch_export_patches] stop_if_static=1
[torch_export_patches] patch pytorch
[torch_export_patches] modifies shape constraints
[torch_export_patches] assert when a dynamic dimension turns static
[torch_export_patches] replaces ShapeEnv._set_replacement
[torch_export_patches] replaces ShapeEnv._log_guard
[torch_export_patches] done patching
-- It failed as excepted.
-- final error is patched_ShapeEnv: A dynamic dimension becomes static! a=s77, tgt=2, msg='range_refined_to_singleton', tgt_bound=VR[2, 2]
-- Stack Trace
Traceback (most recent call last):
File "~/github/onnx-diagnostic/_doc/examples/plot_export_locate_issue.py", line 90, in <module>
torch.export.export(model, inputs, dynamic_shapes=dyn_shapes)
File "~/vv/this312/lib/python3.12/site-packages/torch/export/__init__.py", line 307, in export
raise e
File "~/vv/this312/lib/python3.12/site-packages/torch/export/__init__.py", line 274, in export
return _export(
^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 1158, in wrapper
raise e
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 1124, in wrapper
ep = fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/exported_program.py", line 124, in wrapper
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 2192, in _export
ep = _export_for_training(
^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 1158, in wrapper
raise e
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 1124, in wrapper
ep = fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/exported_program.py", line 124, in wrapper
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 2055, in _export_for_training
export_artifact = export_func(
^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 1997, in _non_strict_export
aten_export_artifact = _to_aten_func( # type: ignore[operator]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 1788, in _export_to_aten_ir_make_fx
gm, graph_signature = transform(_make_fx_helper)(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 1917, in _aot_export_non_strict
gm, sig = aot_export(wrapped_mod, args, kwargs=kwargs, **flags)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 1701, in _make_fx_helper
gm = make_fx(
^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 2354, in wrapped
return make_fx_tracer.trace(f, *args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 2285, in trace
return self._trace_inner(f, *args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 2256, in _trace_inner
t = dispatch_trace(
^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_compile.py", line 53, in inner
return disable_fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_dynamo/eval_frame.py", line 1005, in _fn
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 1283, in dispatch_trace
graph = tracer.trace(root, concrete_args) # type: ignore[arg-type]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 1867, in trace
res = super().trace(root, concrete_args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/_symbolic_trace.py", line 850, in trace
(self.create_arg(fn(*args)),),
^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 1341, in wrapped
out = f(*tensors) # type:ignore[call-arg]
^^^^^^^^^^^
File "<string>", line 1, in <lambda>
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 1588, in wrapped_fn
return tuple(flat_fn(*args))
^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_functorch/_aot_autograd/utils.py", line 187, in flat_fn
tree_out = fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_functorch/_aot_autograd/graph_capture_wrappers.py", line 1354, in functional_call
out = mod(*args[params_len:], **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/_symbolic_trace.py", line 825, in module_call_wrapper
return self.call_module(mod, forward, args, kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 1937, in call_module
return Tracer.call_module(self, m, forward, args, kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/_symbolic_trace.py", line 542, in call_module
ret_val = forward(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/_symbolic_trace.py", line 818, in forward
return _orig_module_call(mod, *args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/export/_trace.py", line 1901, in forward
tree_out = mod(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/_symbolic_trace.py", line 825, in module_call_wrapper
return self.call_module(mod, forward, args, kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 1937, in call_module
return Tracer.call_module(self, m, forward, args, kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/_symbolic_trace.py", line 542, in call_module
ret_val = forward(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/_symbolic_trace.py", line 818, in forward
return _orig_module_call(mod, *args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/github/onnx-diagnostic/_doc/examples/plot_export_locate_issue.py", line 35, in forward
z = x * caty
~~^~~~~~
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 1389, in __torch_function__
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 1438, in __torch_function__
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_export/non_strict_utils.py", line 1067, in __torch_function__
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_ops.py", line 961, in handler
return torch._library.utils.handle_dispatch_mode(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_library/utils.py", line 286, in handle_dispatch_mode
return curr_mode.__torch_dispatch__(op_overload, overload_types, args, kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/utils/_stats.py", line 28, in wrapper
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 1493, in __torch_dispatch__
return proxy_call(self, func, self.pre_dispatch, args, kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/proxy_tensor.py", line 974, in proxy_call
out = func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_ops.py", line 840, in __call__
return self._op(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/utils/_stats.py", line 28, in wrapper
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_subclasses/fake_tensor.py", line 1376, in __torch_dispatch__
return self.dispatch(func, types, args, kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_subclasses/fake_tensor.py", line 2092, in dispatch
return self._cached_dispatch_impl(func, types, args, kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_subclasses/fake_tensor.py", line 1511, in _cached_dispatch_impl
output = self._dispatch_impl(func, types, args, kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_subclasses/fake_tensor.py", line 2611, in _dispatch_impl
return maybe_propagate_real_tensors(fast_impl(self, *args, **kwargs))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/_subclasses/fake_impls.py", line 1271, in fast_binary_impl
final_shape = infer_size(final_shape, shape)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/github/onnx-diagnostic/onnx_diagnostic/torch_export_patches/patches/patch_torch.py", line 119, in patched_infer_size
b3 = guard_size_oblivious(sizeA == sizeB)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/symbolic_shapes.py", line 476, in guard_size_oblivious
return expr.node.guard_size_oblivious("", 0)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/sym_node.py", line 596, in guard_size_oblivious
r = self.evaluate(size_oblivious=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/sym_node.py", line 512, in evaluate
return self.shape_env.evaluate_sym_node(self, size_oblivious)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/symbolic_shapes.py", line 7238, in evaluate_sym_node
return self.evaluate_expr(
^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/symbolic_shapes.py", line 7338, in evaluate_expr
return self._inner_evaluate_expr(
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/recording.py", line 272, in wrapper
return retlog(fn(*args, **kwargs))
^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/symbolic_shapes.py", line 7361, in _inner_evaluate_expr
return self._evaluate_expr(
^^^^^^^^^^^^^^^^^^^^
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/symbolic_shapes.py", line 7630, in _evaluate_expr
self._maybe_guard_rel(g)
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/symbolic_shapes.py", line 6883, in _maybe_guard_rel
self._refine_ranges(expr)
File "~/vv/this312/lib/python3.12/site-packages/torch/fx/experimental/symbolic_shapes.py", line 7834, in _refine_ranges
self._set_replacement(
File "~/github/onnx-diagnostic/onnx_diagnostic/torch_export_patches/patches/patch_torch.py", line 343, in _set_replacement
assert msg != "range_refined_to_singleton", (
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: patched_ShapeEnv: A dynamic dimension becomes static! a=s77, tgt=2, msg='range_refined_to_singleton', tgt_bound=VR[2, 2]
[torch_export_patches] remove patches
[torch_export_patches] restored sympy functions
[torch_export_patches] restored pytorch functions
[torch_export_patches] restored ShapeEnv._set_replacement
[torch_export_patches] restored ShapeEnv._log_guard
[torch_export_patches] restored shape constraints
doc.plot_legend(
"dynamic dimension\nwas inferred\nto be a constant", "torch.export.export", "tomato"
)
Total running time of the script: (0 minutes 0.277 seconds)
Related examples
Intermediate results with (ONNX) ReferenceEvaluator
Intermediate results with (ONNX) ReferenceEvaluator