Skip to content

Fix CUDA UT and HPU UT#1514

Merged
XuehaoSun merged 4 commits intomainfrom
xuehao/cuda-ut
Mar 10, 2026
Merged

Fix CUDA UT and HPU UT#1514
XuehaoSun merged 4 commits intomainfrom
xuehao/cuda-ut

Conversation

@XuehaoSun
Copy link
Copy Markdown
Contributor

@XuehaoSun XuehaoSun commented Mar 9, 2026

Description

Please briefly describe your main changes, the motivation.

Type of Change

  • Bug fix
  • New feature
  • Documentation update
  • Performance improvement
  • Code refactoring
  • Other (please specify):

Related Issues

Fixes or relates to #

Checklist Before Submitting

  • My code has been tested locally.
  • Documentation has been updated as needed.
  • New or updated tests are included where applicable.

Signed-off-by: Sun, Xuehao <xuehao.sun@intel.com>
…ments.txt

Signed-off-by: Sun, Xuehao <xuehao.sun@intel.com>
Signed-off-by: Sun, Xuehao <xuehao.sun@intel.com>
Copilot AI review requested due to automatic review settings March 9, 2026 02:54
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR aims to fix failing CUDA and HPU unit test runs by adjusting test dependency requirements and updating the CUDA UT pipeline script’s installation steps.

Changes:

  • Unpinned compressed_tensors in HPU test requirements.
  • Removed an invalid extra-index-url line from CUDA test requirements.
  • Reordered/relocated dependency and package installation in the CUDA UT pipeline script and made report artifact moves conditional on file existence.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File Description
test/test_hpu/requirements.txt Removes the temporary <0.14 restriction for compressed_tensors.
test/test_cuda/requirements.txt Removes an invalid requirements entry that can break pip -r parsing.
.azure-pipelines/scripts/ut/run_ut_cuda.sh Moves torch + repo installation into per-test functions and guards artifact moves.

cd ${REPO_PATH}/test
rm -rf .coverage* *.xml *.html
uv pip install torch==2.9.1 torchvision
uv pip install -r test_cuda/requirements_sglang.txt
Copy link

Copilot AI Mar 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

run_unit_test_sglang no longer installs/pins PyTorch (the previous torch==... line was removed), but the SGLang tests import torch and the repo install also depends on torch. Add an explicit uv pip install torch==... torchvision (matching the version used in the other CUDA UT functions) before uv pip install . so this job doesn’t end up with an unintended torch version or fail to import torch.

Suggested change
uv pip install -r test_cuda/requirements_sglang.txt
uv pip install -r test_cuda/requirements_sglang.txt
uv pip install torch==2.1.2 torchvision

Copilot uses AI. Check for mistakes.
@XuehaoSun XuehaoSun merged commit 89f99ae into main Mar 10, 2026
15 checks passed
@XuehaoSun XuehaoSun deleted the xuehao/cuda-ut branch March 10, 2026 01:47
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants