add backward inplace for dygraph #35412
Merged
Conversation
This file contains 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
|
Thanks for your contribution! |
sneaxiy
reviewed
Sep 6, 2021
| paddle::framework::OpInfoMap::Instance().Get(op_type).infer_inplace_; | ||
|
|
||
| if (infer_inplace) { | ||
| auto in_to_outs = infer_inplace(true); |
There was a problem hiding this comment.
Discussed offline, the parameter being true and false has no difference now.
AnnaTrainingG
pushed a commit
to AnnaTrainingG/Paddle
that referenced
this pull request
Sep 29, 2021
* add backward inplace for dygraph * fix bug * support gradient accumulation
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
PR types
New features
PR changes
Others
Describe
add backward inplace for dygraph
how
In dygraph mode, the forward part is not determined, while the backward computation graph is determined after forward. So the graph optimization can be performed. And, the reference count used to guide inplace optimization can be easily get
shared_ptrof input tensor.Two benefits: (1) reduce d2d copy and speed up training, (2) reduce memory usage
performance
GPT model, speed up +2.8%.
