Sequence parallel strategy support. #819
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
1. Add sequence parallel strategy for GPTModelHybrid 2. Output has been checked layer by layer both in forward and backward progress, and its loss curve of the beginning 5000 steps fits the peer 3. Performance is improved for about 10% with sequence_parallel strategy compared with pretrain_gpt_1.3B_mp8
… sequence_parallel
1. Add sequence_parallel option for GPTModel 2. When mp=1, sequence_parallel option should always be set False
… sequence_parallel
sneaxiy
approved these changes
Oct 17, 2022
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Outdated suggestions cannot be applied.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.
Add sequence support strategy support for GPT pipeline parallel model.
_is_valid_send_recv_partial()in paddle/distributed/fleet/meta_parallel/pp_utils/p2p_communication.py should been modified. Corresponding PR of paddle repo will be submitted.