[PT FE] Improve torch.export tests#33980
Open
mvafin wants to merge 1 commit intoopenvinotoolkit:masterfrom
Open
[PT FE] Improve torch.export tests#33980mvafin wants to merge 1 commit intoopenvinotoolkit:masterfrom
mvafin wants to merge 1 commit intoopenvinotoolkit:masterfrom
Conversation
Contributor
There was a problem hiding this comment.
Pull request overview
This PR enhances torch.export tests by adding validation of FX graph operations. The changes ensure that tests properly verify the presence of expected operations in the FX graph representation, improving test accuracy and catching potential regressions.
Changes:
- Added
fx_kindparameter validation to verify expected FX operations exist in exported models - Implemented helper methods to check FX operation existence and auto-derive FX operation names from TorchScript kinds
- Updated test methods across multiple files to specify expected FX operations
Reviewed changes
Copilot reviewed 31 out of 31 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
| tests/layer_tests/pytorch_tests/pytorch_layer_test_class.py | Core infrastructure: added FX operation validation methods and assertion logic |
| tests/layer_tests/pytorch_tests/test_unary_ops.py | Added fx_kind parameter for abs operation validation |
| tests/layer_tests/pytorch_tests/test_split.py | Added fx_kind parameters for split operations and modernized super() calls |
| tests/layer_tests/pytorch_tests/test_slice.py | Added fx_kind for unsqueeze validation and cleaned up model |
| tests/layer_tests/pytorch_tests/test_shift_operations.py | Added fx_kind for shift operations |
| tests/layer_tests/pytorch_tests/test_scatter.py | Added fx_kind for scatter_add validation |
| tests/layer_tests/pytorch_tests/test_resolve_conj_neg.py | Added fx_kind for resolve operations |
| tests/layer_tests/pytorch_tests/test_repeat.py | Added fx_kind and removed precommit_torch_export marker |
| tests/layer_tests/pytorch_tests/test_permute.py | Added fx_kind for permute validation |
| tests/layer_tests/pytorch_tests/test_neg.py | Added fx_kind for built-in neg function |
| tests/layer_tests/pytorch_tests/test_min_max.py | Added fx_kind for min/max operations |
| tests/layer_tests/pytorch_tests/test_matmul.py | Formatting cleanup (trailing comma) |
| tests/layer_tests/pytorch_tests/test_masked_fill.py | Used skip_if_export for inplace parameter |
| tests/layer_tests/pytorch_tests/test_loop.py | Added fx_kind for while_loop validation |
| tests/layer_tests/pytorch_tests/test_logsumexp.py | Added fx_kind for logsumexp validation |
| tests/layer_tests/pytorch_tests/test_lerp.py | Added fx_kind with conditional logic for inplace variant |
| tests/layer_tests/pytorch_tests/test_index_copy_.py | Used skip_if_export for inplace parameter |
| tests/layer_tests/pytorch_tests/test_fake_quantize.py | Added fx_kind for fake quantization operations |
| tests/layer_tests/pytorch_tests/test_eye.py | Added fx_kind for eye operations |
| tests/layer_tests/pytorch_tests/test_expand.py | Added fx_kind for expand/broadcast validation |
| tests/layer_tests/pytorch_tests/test_device.py | Removed precommit_torch_export markers |
| tests/layer_tests/pytorch_tests/test_convolution.py | Added fx_kind for convolution operations |
| tests/layer_tests/pytorch_tests/test_cond.py | Added fx_kind for conditional operations |
| tests/layer_tests/pytorch_tests/test_clamp.py | Added fx_kind for clamp operations |
| tests/layer_tests/pytorch_tests/test_broadcast_tensors.py | Added fx_kind and refactored parametrization to reduce test combinations |
| tests/layer_tests/pytorch_tests/test_bitwise_ops.py | Added fx_kind for bitwise operations |
| tests/layer_tests/pytorch_tests/test_atan2.py | Added fx_kind for atan2 validation |
| tests/layer_tests/pytorch_tests/test_as_strided.py | Added fx_kind for as_strided validation |
| tests/layer_tests/pytorch_tests/test_aliases.py | Added fx_kind for alias operations |
| tests/layer_tests/pytorch_tests/test_add.py | Added fx_kind for add and removed precommit_torch_export marker |
| tests/layer_tests/pytorch_tests/test_strided_const.py | Modernized super() call and updated return statement |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| if fx_kind is None: | ||
| fx_kind = self._derive_fx_kind_from_kind(kind) | ||
|
|
||
| assert fx_kind is not None, f"fx_kind must be provided explicitly or derivable from kind. Graph:\n{em.graph}" |
There was a problem hiding this comment.
The error message could be more helpful by specifying what 'kind' value was provided. Consider: f\"fx_kind must be provided explicitly or derivable from kind='{kind}'. Graph:\\n{em.graph}\"
Suggested change
| assert fx_kind is not None, f"fx_kind must be provided explicitly or derivable from kind. Graph:\n{em.graph}" | |
| assert fx_kind is not None, f"fx_kind must be provided explicitly or derivable from kind='{kind}'. Graph:\n{em.graph}" |
| if not isinstance(fx_kind, (tuple, list)): | ||
| fx_kind = [fx_kind] | ||
| for op in fx_kind: | ||
| assert self._check_fx_op_exist(em, op), f"Operation {op} doesn't exist in FX graph. Graph:\n{em.graph}" |
There was a problem hiding this comment.
The error message uses the contraction "doesn't" which should be spelled out as "does not" for consistency with professional error messages.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Details:
Tickets: