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.

Sign up for GitHub

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

Jump to bottom

[Unified Checkpoint] Cherry-pick for newest paddle #9361

Closed

DesmonDay wants to merge 28 commits into PaddlePaddle:develop from DesmonDay:origin_release_3.0-beta2

Contributor

DesmonDay commented Nov 4, 2024

PR types

Bug fixes

PR changes

Others

Description

Cherry-pick #9360

gongel and others added 28 commits

September 30, 2024 11:49


          fix speed ppt eval hang (PaddlePaddle#9218)

fbe5bc1


          Update branch version to 3.0.0b2 (PaddlePaddle#9220)

b4017a5

* fix speed ppt eval hang

* update version


          Update branch version to 3.0.0b2 (PaddlePaddle#9221)

57ce8a7

* fix speed ppt eval hang

* update version

* update version


          Revert "fix speed ppt eval hang (PaddlePaddle#9218)" (PaddlePaddle#9229)

81de41a

This reverts commit fbe5bc1.


          Revert "Revert "fix speed ppt eval hang (PaddlePaddle#9218)" (PaddleP…

73f1a20

…addle#9229)" (PaddlePaddle#9230)

This reverts commit 81de41a.


          [Unified Checkpoint] Support expert parallel (PaddlePaddle#9055)

2c24e99

* update code


          [Unified Checkpoint] Fix generation config save (PaddlePaddle#9223)

0ea031c


          [Unified Checkpoint] update async_save_info in develop (PaddlePaddle#…

6af261d

…9173)


          [Unified Checkpoint] update async save logic (PaddlePaddle#9274)

4f66a19

* update async save signal

* fix async save hang


          bug fix

8f6e88b


          [Unified checkpoint] PR cherry-picks (PaddlePaddle#9277)

37b03fc

* [Unified Checkpoint] Support expert parallel (PaddlePaddle#9055)

* update code

* [Unified Checkpoint] Fix generation config save (PaddlePaddle#9223)

* [Unified Checkpoint] update async_save_info in develop (PaddlePaddle#9173)

* [Unified Checkpoint] update async save logic (PaddlePaddle#9274)

* update async save signal

* fix async save hang

* bug fix

---------

Co-authored-by: Weiguo Zhu <[email protected]>


          bug fix

bf4003c


          Merge branch 'release/3.0-beta2' of https://github.com/PaddlePaddle/P…

a557aac

…addleNLP into origin_release_3.0-beta2


          [BugFix] fix get_last_checkpoint usage (PaddlePaddle#9283)

036e1cd


          [Trainer] fix save_model (PaddlePaddle#9286)

a43371d

* bug fix

* bug fix


          Merge branch 'release/3.0-beta2' of https://github.com/PaddlePaddle/P…

1cf3094

…addleNLP into origin_release_3.0-beta2


          [Trainer] fix save_model (PaddlePaddle#9287)

8da4248

* [Unified Checkpoint] Support expert parallel (PaddlePaddle#9055)

* update code

* [Unified Checkpoint] Fix generation config save (PaddlePaddle#9223)

* [Unified Checkpoint] update async_save_info in develop (PaddlePaddle#9173)

* [Unified Checkpoint] update async save logic (PaddlePaddle#9274)

* update async save signal

* fix async save hang

* bug fix

* bug fix

* [Trainer] fix save_model (PaddlePaddle#9286)

* bug fix

* bug fix

---------

Co-authored-by: Weiguo Zhu <[email protected]>


          [Cherry-pick] fix non-pipelinelayer to distributed PaddlePaddle#9 (Pa…

513a966

…ddlePaddle#9313)

* fix speed ppt eval hang

* update version

* update version

* fix non-pipelayer to distributed


          [Unified Checkpoint] Accelerate loading checkpoint by multi-thread (P…

bf5551d

…addlePaddle#9034)

* [Unified Checkpoint] speed up loading checkpoint by multi thread

* [Unified CHeckpoint] speed up load by multi-thread

* [Unified CHeckpoint] speed up load by multi-thread

* [Unified CHeckpoint] speed up load by multi-thread

* Unified CHeckpoint] speed up loading checkpoint by  multi-thread

* Unified CHeckpoint] speed up loading checkpoint by  multi-thread

* Unified CHeckpoint] speed up loading checkpoint by  multi-thread

* Unified CHeckpoint] speed up loading checkpoint by  multi-thread

* Unified CHeckpoint] speed up loading checkpoint by  multi-thread

* Unified CHeckpoint] speed up loading checkpoint by  multi-thread


          Merge branch 'release/3.0-beta2' of https://github.com/PaddlePaddle/P…

c7c8251

…addleNLP into origin_release_3.0-beta2


          [Unified Checkpoint] Support multi-threads loading (PaddlePaddle#9316)

---------

Co-authored-by: CrystalX <[email protected]>


          Merge branch 'release/3.0-beta2' of https://github.com/PaddlePaddle/P…

9a15f35

…addleNLP into origin_release_3.0-beta2


          [Unified Checkpoint] Add split param and refactor code (PaddlePaddle#…

…9240)

* [Unified checkpoint] update optimizer async save signal

* update paddlepaddle

* split param

* add save for split param

* fix save split_param

* add load uc split_param

* update uc files

* update uc files

* update split_param loading

* mkdir unified_checkpoint directory

* rename file

* update async handler

* update files

---------

Co-authored-by: gongenlei <[email protected]>


          [Unified Checkpoint] update non-merge checkpoint loading, move async_…

9c941b3

…save_info.json location (PaddlePaddle#9321)

* [Unified checkpoint] update optimizer async save signal

* update async_save_info.json file place


          [Cherry-pick] Support split_param and refactor uc code. (PaddlePaddle…

9cf618e

…#9329)


          Update load_save_single_card.py

e5cc07b


          Merge branch 'release/3.0-beta2' of https://github.com/PaddlePaddle/P…

62ecb24

…addleNLP into origin_release_3.0-beta2


          [Unified Checkpoint] Fix fp32 dtype for using newest paddle(PaddlePad…

e281a32

…dle#9360)

paddle-bot bot commented Nov 4, 2024

Thanks for your contribution!

DesmonDay closed this

CLAassistant commented Nov 4, 2024 •

edited

Loading

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
4 out of 5 committers have signed the CLA.

✅ gongel
✅ DrownFish19
✅ Crystal-X-111
✅ DesmonDay
❌ ZHUI
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

DesmonDay reopened this

DesmonDay closed this

codecov bot commented Nov 4, 2024 •

edited

Loading

Codecov Report

Attention: Patch coverage is 12.11640% with 1661 lines in your changes missing coverage. Please review.

Project coverage is 52.06%. Comparing base (f80a051) to head (e281a32).
Report is 310 commits behind head on develop.

Files with missing lines	Patch %	Lines
paddlenlp/trainer/unified_checkpoint/utils.py	11.46%	363 Missing ⚠️
...p/trainer/unified_checkpoint/unified_checkpoint.py	10.00%	288 Missing ⚠️
...ddlenlp/trainer/unified_checkpoint/load_dynamic.py	8.86%	257 Missing ⚠️
...r/unified_checkpoint/sharding_split_param_utils.py	7.97%	173 Missing ⚠️
...nlp/trainer/unified_checkpoint/check_completion.py	8.33%	143 Missing ⚠️
...dlenlp/trainer/unified_checkpoint/async_handler.py	11.32%	141 Missing ⚠️
paddlenlp/trainer/unified_checkpoint/load_local.py	11.62%	114 Missing ⚠️
...rainer/unified_checkpoint/load_save_single_card.py	14.28%	114 Missing ⚠️
paddlenlp/trainer/trainer.py	38.00%	31 Missing ⚠️
paddlenlp/transformers/model_utils.py	42.50%	23 Missing ⚠️
... and 3 more

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #9361      +/-   ##
===========================================
- Coverage    53.02%   52.06%   -0.96%     
===========================================
  Files          656      676      +20     
  Lines       106153   116200   +10047     
===========================================
+ Hits         56288    60501    +4213     
- Misses       49865    55699    +5834

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet