[XPU] [CI] Xpu ci lock PaddlePaddle Version#5218
Merged
plusNew001 merged 10 commits intoPaddlePaddle:developfrom Nov 25, 2025
Merged
[XPU] [CI] Xpu ci lock PaddlePaddle Version#5218plusNew001 merged 10 commits intoPaddlePaddle:developfrom
plusNew001 merged 10 commits intoPaddlePaddle:developfrom
Conversation
Comment out the previous paddlepaddle-xpu installation command and replace it with a specific version installation due to EP parallel error.
|
Thanks for your contribution! |
Contributor
There was a problem hiding this comment.
Pull request overview
This PR addresses a CI bug in the XPU environment by temporarily pinning the paddlepaddle-xpu package to a specific development version (3.3.0.dev20251123) instead of using the latest nightly build, due to EP (Executor Parallelism) related errors.
Key Changes:
- Commented out the original nightly build installation command
- Added a specific paddlepaddle-xpu version installation from a direct URL
- Included a Chinese comment explaining the temporary fix
Comment on lines
+47
to
+49
| # python -m pip install paddlepaddle-xpu -i https://www.paddlepaddle.org.cn/packages/nightly/xpu-p800/ | ||
| # 由于ep并行报错暂时锁死paddle版本 | ||
| python -m pip install https://paddle-whl.bj.bcebos.com/nightly/xpu-p800/paddlepaddle-xpu/paddlepaddle_xpu-3.3.0.dev20251123-cp310-cp310-linux_x86_64.whl |
There was a problem hiding this comment.
The PR description is incomplete. Please provide:
- Motivation: Explain why this change is needed (e.g., "Due to EP parallelism errors in the nightly build")
- Modifications: Detail what was changed (e.g., "Temporarily pinned paddlepaddle-xpu to version 3.3.0.dev20251123 instead of using the latest nightly build")
- Usage or Command: Include the command to run the CI test
- Checklist: Complete the checklist items, especially regarding unit tests and accuracy tests
Comment on lines
+48
to
+49
| # 由于ep并行报错暂时锁死paddle版本 | ||
| python -m pip install https://paddle-whl.bj.bcebos.com/nightly/xpu-p800/paddlepaddle-xpu/paddlepaddle_xpu-3.3.0.dev20251123-cp310-cp310-linux_x86_64.whl |
There was a problem hiding this comment.
Hardcoding a specific development version URL creates maintainability issues. Consider:
- Adding a version variable at the top of the script for easy updates (e.g.,
PADDLE_XPU_VERSION="3.3.0.dev20251123") - Adding a TODO comment with a tracking issue number to revert to the nightly build once the EP parallelism bug is fixed
- Documenting when this temporary fix should be removed
Example:
# TODO(issue-#XXXX): Revert to nightly build once EP parallelism bug is fixed
PADDLE_XPU_VERSION="3.3.0.dev20251123"
python -m pip install https://paddle-whl.bj.bcebos.com/nightly/xpu-p800/paddlepaddle-xpu/paddlepaddle_xpu-${PADDLE_XPU_VERSION}-cp310-cp310-linux_x86_64.whl
Suggested change
| # 由于ep并行报错暂时锁死paddle版本 | |
| python -m pip install https://paddle-whl.bj.bcebos.com/nightly/xpu-p800/paddlepaddle-xpu/paddlepaddle_xpu-3.3.0.dev20251123-cp310-cp310-linux_x86_64.whl | |
| # TODO(issue-#XXXX): Revert to nightly build once EP parallelism bug is fixed. Temporary fix: lock paddlepaddle-xpu version due to EP parallelism error. | |
| PADDLE_XPU_VERSION="3.3.0.dev20251123" | |
| python -m pip install https://paddle-whl.bj.bcebos.com/nightly/xpu-p800/paddlepaddle-xpu/paddlepaddle_xpu-${PADDLE_XPU_VERSION}-cp310-cp310-linux_x86_64.whl |
| python -m pip install paddlepaddle-xpu -i https://www.paddlepaddle.org.cn/packages/nightly/xpu-p800/ | ||
|
|
||
| # python -m pip install paddlepaddle-xpu -i https://www.paddlepaddle.org.cn/packages/nightly/xpu-p800/ | ||
| # 由于ep并行报错暂时锁死paddle版本 |
There was a problem hiding this comment.
建议在中文注释旁边添加英文翻译以便国际贡献者理解:
# 由于ep并行报错暂时锁死paddle版本
# Temporarily locking paddle version due to EP parallelism errorConsider adding an English translation alongside the Chinese comment for better accessibility to international contributors.
Suggested change
| # 由于ep并行报错暂时锁死paddle版本 | |
| # 由于ep并行报错暂时锁死paddle版本 | |
| # Temporarily locking paddle version due to EP parallelism error |
|
|
||
| # python -m pip install paddlepaddle-xpu -i https://www.paddlepaddle.org.cn/packages/nightly/xpu-p800/ | ||
| # 由于ep并行报错暂时锁死paddle版本 | ||
| python -m pip install https://paddle-whl.bj.bcebos.com/nightly/xpu-p800/paddlepaddle-xpu/paddlepaddle_xpu-3.3.0.dev20251123-cp310-cp310-linux_x86_64.whl |
There was a problem hiding this comment.
使用直接URL安装特定开发版本可能会导致问题：
- 该URL指向特定的Python版本(cp310),如果环境使用其他Python版本会失败
- 没有错误处理 - 如果下载失败,安装会静默失败
建议改进:
PADDLE_XPU_VERSION="3.3.0.dev20251123"
PADDLE_XPU_URL="https://paddle-whl.bj.bcebos.com/nightly/xpu-p800/paddlepaddle-xpu/paddlepaddle_xpu-${PADDLE_XPU_VERSION}-cp310-cp310-linux_x86_64.whl"
echo "Installing paddlepaddle-xpu ${PADDLE_XPU_VERSION}..."
python -m pip install "${PADDLE_XPU_URL}" || {
echo "Error: Failed to install paddlepaddle-xpu from ${PADDLE_XPU_URL}"
exit 1
}Installing from a direct URL to a specific dev version can cause issues:
- The URL targets a specific Python version (cp310), which will fail if the environment uses a different Python version
- No error handling - if the download fails, the installation fails silently
Consider improving with the code above.
Suggested change
| python -m pip install https://paddle-whl.bj.bcebos.com/nightly/xpu-p800/paddlepaddle-xpu/paddlepaddle_xpu-3.3.0.dev20251123-cp310-cp310-linux_x86_64.whl | |
| # 动态获取当前Python版本，构造对应的whl文件URL | |
| PYTHON_VERSION=$(python -c "import sys; print(f'cp{sys.version_info.major}{sys.version_info.minor}')") | |
| PADDLE_XPU_VERSION="3.3.0.dev20251123" | |
| PADDLE_XPU_URL="https://paddle-whl.bj.bcebos.com/nightly/xpu-p800/paddlepaddle-xpu/paddlepaddle_xpu-${PADDLE_XPU_VERSION}-${PYTHON_VERSION}-${PYTHON_VERSION}-linux_x86_64.whl" | |
| echo "Installing paddlepaddle-xpu ${PADDLE_XPU_VERSION} for Python ${PYTHON_VERSION}..." | |
| python -m pip install "${PADDLE_XPU_URL}" || { | |
| echo "Error: Failed to install paddlepaddle-xpu from ${PADDLE_XPU_URL}" | |
| exit 1 | |
| } |
plusNew001
changed the title
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## develop #5218 +/- ##
==========================================
Coverage ? 59.84%
==========================================
Files ? 317
Lines ? 38682
Branches ? 5813
==========================================
Hits ? 23150
Misses ? 13707
Partials ? 1825
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
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
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.
Motivation
Modifications
Usage or Command
Accuracy Tests
Checklist
[FDConfig],[APIServer],[Engine],[Scheduler],[PD Disaggregation],[Executor],[Graph Optimization],[Speculative Decoding],[RL],[Models],[Quantization],[Loader],[OP],[KVCache],[DataProcessor],[BugFix],[Docs],[CI],[Optimization],[Feature],[Benchmark],[Others],[XPU],[HPU],[GCU],[DCU],[Iluvatar],[Metax]]pre-commitbefore commit.releasebranch, make sure the PR has been submitted to thedevelopbranch, then cherry-pick it to thereleasebranch with the[Cherry-Pick]PR tag.