[Relax] Add torch exported program ingestion capability for Tensor.detach(), Tensor.copy_, and aten.lift_fresh_copy #17723
Conversation
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
MasterJH5574
approved these changes
Mar 11, 2025
ShiboXing
pushed a commit
to ShiboXing/tvm
that referenced
this pull request
Aug 10, 2025
…tach(), Tensor.copy_, and aten.lift_fresh_copy (apache#17723) * detach and copy * copy_ implemenation. Unit test passes * restore test_frontend * don't specify syspath * todo for _detach() * Black formatter * black formatting with version 22.12.0 * cleanup unit tests and ran Black Formatter with version 22 * restore unmodified frontend test * fix vm in assert_torch_output_vs_tvm_from_exported_to_cuda * lint with Python Black formatter * update todo * update explanation for _detach
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
Exported program translator now supports Tensor.detach(), Tensor.copy_, and aten.lift_fresh_copy.
I also added unit tests. The last test I created does not pass (it's commented out) because I found no way to correctly implement detach(). The output of detach() should share the same memory as the input. It is not the case right now.
In other words, ideally, this test would pass, but it doesn't : https://github.com/hugolatendresse/tvm/blob/456845811ba01c0bea07737a4f7a333a0b45ea92/tests/python/relax/test_from_exported_to_cuda.py#L98
cc: @MasterJH5574