Skip to content

[Eager grad] Refactor partial grad logic #40693

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
Merged
merged 53 commits into from
Mar 21, 2022
Merged

[Eager grad] Refactor partial grad logic #40693

merged 53 commits into from
Mar 21, 2022

Conversation

@veyron95
Copy link
Contributor

@veyron95 veyron95 commented Mar 18, 2022 •
edited
Loading

PR types

Function optimization

PR changes

Others

Describe

  1. Refactor partial grad logic
  2. Add DuplicateCheck for partial grad interface
  3. PartialGrad rename to GeneralGrad
veyron95 added 30 commits March 7, 2022 12:42
@veyron95
[Eager] Support eager grad interface, draft version
9fc70fe
@veyron95
Support eager grad interface with allow_unused and multi startup_op
ba8d79e
@veyron95
Fix code format
137db9d
@veyron95
Fix allow_unused case, return PyNone if tensor not initialize
1a18aa2
@veyron95
Support output's stop_gradient related to create_graph
d09ec3b
@veyron95
Support grad exception case in eager mode, fix coverage CI
f84f2be
@veyron95
Update ToPyObject, return PyNone if not initialize
733672e
@veyron95
AccumulationNode add FLAGS_retain_grad_for_all_tensor
68b1991
@veyron95
Fix ci issue
7665d63
@veyron95
Fix CI issue
86393f5
@veyron95
fix, use core.eager.Tensor
c653ec0
@veyron95
Add func SetBufferSlotRankZeros for GradTensorHolder
9156cea
@veyron95
Support retain_graph by using ClearTensorWrappers
6fd613d
@veyron95
Support retain_graph by using ClearTensorWrappers
58731e9
@veyron95
Update retain_graph and no_grad_vars related test case
a88f9b1
@veyron95
Update code gen logic for ClearTensorWrappers
778719b
@veyron95
Fix by override statement
65cf9e3
@veyron95
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
af7b919 
… support_partial_grad
@veyron95
fix override func args
4d3b57d
@veyron95
Support retain_graph, update unit tests
415ff65
@veyron95
Updated ClearTensorWrappers logic
bb283ce
@veyron95
fix grad python interface
e548c22
@veyron95
Use deep copy and update unit tests
519c9a6
@veyron95
Polish code
1fbc61b
@veyron95
Polish code
c0a2b8b
@veyron95
Fix CI issue, Deep copy only use when user set grad_tensors
536a28b
@veyron95
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
2417858 
… support_partial_grad
@veyron95
Fix CI, use Backward instead RunBackward
34fa7c0
@veyron95
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
1b89072 
… support_partial_grad
@veyron95
Fix CI, Declare kernel explicitly in test file
af7f058
veyron95 added 20 commits March 14, 2022 13:14
@veyron95
Polish, remove vector of TensorWrapper
f397b8f
@veyron95
Refactor the logic of grad/backward, polish codes
e3f9826
@veyron95
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
27830a9 
… support_partial_grad
@veyron95
Update code after merge upstream develop
7ede919
@veyron95
Polish after merge upstream develop
f9adf49
@veyron95
Update to adapt new GradNodeBase superclass
2fe3b9f
@veyron95
Fix error introduced during conflict resolution
90e97d6
@veyron95
Update purify potential_startup_nodes logic
d18697a
@veyron95
Fix errors
1b5eac2
@veyron95
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
f4e42e2 
… support_partial_grad
@veyron95
Polish code
58a03b5
@veyron95
Remove useless args for ToPyObject
b04e9a9
@veyron95
Remove useless TensorWrappersSet
c7bd6fc
@veyron95
Fix code-format, re-install pre-commit
ac85d81
@veyron95
Fix pre-process logic for potential_startup_ops
8312d2d
@veyron95
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
441bc81 
… support_partial_grad
@veyron95
Update unit tests, use eager mode
326eee5
@veyron95
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
a0ec433 
… support_partial_grad
@veyron95
Fix conflicts
df99eea
@veyron95
Refactor partial_grad/backward logic
7b1841a
@paddle-bot-old
Copy link

paddle-bot-old bot commented Mar 18, 2022 •
edited
Loading

✅ This PR's description meets the template requirements!
Please wait for other CI results.
@paddle-bot-old
Copy link

paddle-bot-old bot commented Mar 18, 2022

你的PR提交成功，感谢你对开源项目的贡献!
请关注后续CI自动化测试结果，详情请参考Paddle-CI手册。
Your PR has been submitted. Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.
JiabinYang
JiabinYang approved these changes Mar 21, 2022
View reviewed changes
Copy link
Contributor

@JiabinYang JiabinYang left a comment

Choose a reason for hiding this comment

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

LGTM
chenwhql
chenwhql approved these changes Mar 21, 2022
View reviewed changes
Copy link
Contributor

@chenwhql chenwhql left a comment

Choose a reason for hiding this comment

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

LGTM for PADDLE_ENFORCE
@veyron95 veyron95 merged commit facda82 into PaddlePaddle:develop Mar 21, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

4 participants

@veyron95 @jim19930609 @JiabinYang @chenwhql