Skip to content

feat(shortgpt): add new script and polish existing code (typos, paths)#10882

Closed
tianyumyum wants to merge 2 commits into
PaddlePaddle:developfrom
tianyumyum:add_shortgpt
Closed

feat(shortgpt): add new script and polish existing code (typos, paths)#10882
tianyumyum wants to merge 2 commits into
PaddlePaddle:developfrom
tianyumyum:add_shortgpt

Conversation

@tianyumyum

@tianyumyum tianyumyum commented Jul 23, 2025

Copy link
Copy Markdown
Contributor

Before submitting

  • Lint code. If there are lint issues, please format the code first.
# Install and register `pre-commit` in the project folder
pip install pre-commit && pre-commit install

# Process previous code files separately
pre-commit run --file XXXX.py
  • Add test cases into tests folder. If there are codecov issues, please add tests cases first.

PR types

  1. 新增shortgpt删层操作，用于模型推理加速，减少显存使用
  2. 修复原项目中的typo，模型路径不对齐等问题

@paddle-bot

paddle-bot Bot commented Jul 23, 2025

Copy link
Copy Markdown

Thanks for your contribution!

@CLAassistant

CLAassistant commented Jul 23, 2025 •
edited
Loading

Copy link
Copy Markdown

CLA assistant check
All committers have signed the CLA.

DrownFish19
DrownFish19 reviewed Jul 23, 2025
View reviewed changes
Comment thread slm/pipelines/examples/contrastive_training/README.md
# 向量检索模型训练

推荐安装 gpu 版本的[PaddlePaddle](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/install/conda/linux-conda.html)，以 cuda12.3的 paddle 为例，安装命令如下：
推荐安装 gpu 版本的[PaddlePaddle](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/install/conda/linux-conda.html)，以 cuda11.8的 paddle 为例，安装命令如下：

@DrownFish19 DrownFish19 Jul 23, 2025

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

推荐使用更高版本的paddle，低版本保留时间不会太长

@tianyumyum tianyumyum Jul 23, 2025 •
edited
Loading

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

适配更高cuda版本的话，请问具体哪个版本呢，nightly or stable。目前我使用服务器中的cuda version是12.1

@DrownFish19 DrownFish19 Jul 31, 2025

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

可以考虑当前的cuda 12.6和12.9

@tianyumyum tianyumyum marked this pull request as ready for review July 23, 2025 02:37
@DrownFish19

DrownFish19 commented Jul 31, 2025

Copy link
Copy Markdown
Collaborator

我看Test测试还有些问题，可以试试拉最新代码，然后重新过Test试试

@DrownFish19 DrownFish19 reopened this Jul 31, 2025
@PaddlePaddle PaddlePaddle locked and limited conversation to collaborators Aug 18, 2025
@PaddlePaddle PaddlePaddle unlocked this conversation Aug 18, 2025
@PaddlePaddle PaddlePaddle locked and limited conversation to collaborators Aug 18, 2025
@PaddlePaddle PaddlePaddle unlocked this conversation Aug 18, 2025
@Liujie0926 Liujie0926 closed this Aug 18, 2025
@Liujie0926 Liujie0926 reopened this Aug 18, 2025
@Liujie0926

Liujie0926 commented Aug 20, 2025

Copy link
Copy Markdown
Collaborator

PaddleNLP-CI、PaddleNLP-CI-Unittest-GPU、PaddleNLP-gpt-A100任务调起失败，怀疑是因为第一次提交pr未能成功调起，建议重新提交一个新的pr尝试。
另外Test任务网络问题已修复，辛苦merge下develop代码

@tianyumyum tianyumyum closed this Aug 21, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@DrownFish19 DrownFish19 DrownFish19 left review comments

Assignees

@KB-Ding KB-Ding

Labels

Beijing Innovation Consortium contributor

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@tianyumyum @CLAassistant @DrownFish19 @Liujie0926 @KB-Ding