[HybridParallel]Fix c_split op for TensorParallel#33207
Merged
ForFishes merged 11 commits intoPaddlePaddle:developfrom Jun 7, 2021
Merged
[HybridParallel]Fix c_split op for TensorParallel#33207ForFishes merged 11 commits intoPaddlePaddle:developfrom
ForFishes merged 11 commits intoPaddlePaddle:developfrom
Conversation
|
Thanks for your contribution! |
chenwhql
approved these changes
Jun 7, 2021
Contributor
chenwhql
left a comment
chenwhql
left a comment
There was a problem hiding this comment.
LGTM for PADDLE_ENFORCE in kernel, please fix comment in next PR
| PADDLE_ENFORCE_EQ( | ||
| table_dims.size(), 2, | ||
| platform::errors::InvalidArgument( | ||
| "ShapeError: The dimensions of the 'c_embedding' must be 2. " |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Function optimization
PR changes
Others
Describe
[HybridParallel]Fix c_split op for TensorParallel
优化MP性能，优化点如下：
1、fuse embedding op，对shard_index + embedding进行fuse，同时恢复embeding 的weight的正常维度。
2、解决_c_concat的api没有传入rank信息的bug。