onnx_diagnostic.torch_export_patches.patches.patch_torch¶
- onnx_diagnostic.torch_export_patches.patches.patch_torch.patch__check_input_constraints_for_graph(previous_function: Callable, input_placeholders: list[Node], flat_args_with_path, range_constraints, verbose: int = 0) None [source][source]¶
- onnx_diagnostic.torch_export_patches.patches.patch_torch.patched__broadcast_shapes(*_shapes)[source][source]¶
Patches
torch._refs._broadcast_shapes
.
- onnx_diagnostic.torch_export_patches.patches.patch_torch.patched_infer_size(a, b)[source][source]¶
Patches
torch._subclasses.fake_impls.infer_size
.
- onnx_diagnostic.torch_export_patches.patches.patch_torch.patched_vmap(func, in_dims=0, out_dims=0)[source][source]¶
Python implementation of
torch.vmap()
. The implementation raises an issue when it is being exported withtorch.export.export()
when the function is called with non tensors arguments and the batch size is dynamic.