-
Notifications
You must be signed in to change notification settings - Fork 2.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[FIX DDP] fix ddp #8549
[FIX DDP] fix ddp #8549
Conversation
Thanks for your contribution! |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## develop #8549 +/- ##
===========================================
+ Coverage 53.86% 54.96% +1.10%
===========================================
Files 620 620
Lines 97090 97109 +19
===========================================
+ Hits 52298 53377 +1079
+ Misses 44792 43732 -1060 ☔ View full report in Codecov by Sentry. |
@@ -1795,17 +1795,8 @@ def _wrap_model(self, model, training=True): | |||
in_cp_parallel_mode = self.args.context_parallel_degree > 1 | |||
|
|||
# Multi-gpu training |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
这个地方需要合入到2.8吗?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
* enable trainer tests.
* [Safetensors] Fix fast safe open slice. (PaddlePaddle#8512) * [FIX DDP] fix ddp (PaddlePaddle#8549)
* quick fix from pretrained. (#8487) * quick fix os.path.split (#8508) * Cp/fix (#8569) * [Safetensors] Fix fast safe open slice. (#8512) * [FIX DDP] fix ddp (#8549) * [BUG] Fix build train valid test datasets (#8823) * Update causal_dataset.py * Add twenty redundant data in post pretrain (#8777) * 给dataset再添加20条数据,防止blend dataset出现错误 * num_samples向下去整,防止数据集的溢出 (#8691) * update release_grads (#8834) * update release_grads (#8834) * [Trainer] Fix release_grads (#9085) * fix pp release_grads * add dataloader_drop_last to evaldataloader (#8773) * bugfix * Fix eval hang (#9052) * fix pipeline eval * fix eval dataloader_num_workers --------- Co-authored-by: Zhong Hui <zhonghui.net@gmail.com> Co-authored-by: yujun <50394665+JunnYu@users.noreply.github.com> Co-authored-by: gongel <ainlp88@qq.com>
PR types
Others
PR changes
Others
Description
Others