Revert "Embedding parallel by attn_tp (#7623)" #7880
Conversation
This reverts commit ef8a29c.
There was a problem hiding this comment.
Summary of Changes
Hello @zhyncs, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!
This pull request serves to undo a recent modification to the DeepSeekV2 model's embedding layer configuration. The primary purpose is to revert the parallelism setting for the vocabulary embedding, restoring its behavior to be conditionally dependent on a global server argument related to data parallelism for attention, rather than a fixed attention tensor parallelism group.
Highlights
- Revert: This pull request reverts a previous change (commit
ef8a29c429f62e0b0d013cc5091264ac75f70f56) that modified the initialization ofVocabParallelEmbeddingin the DeepSeekV2 model. Specifically, it undoes the introduction ofuse_attn_tp_group=True. - Model Configuration: The
VocabParallelEmbeddinginitialization for the DeepSeekV2 model is reverted to its prior state, where theenable_tpparameter is set based on theglobal_server_args_dict["enable_dp_attention"]flag, rather than a fixeduse_attn_tp_group=True.
Using Gemini Code Assist
The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.
Invoking Gemini
You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.
| Feature | Command | Description |
|---|---|---|
| Code Review | /gemini review |
Performs a code review for the current pull request in its current state. |
| Pull Request Summary | /gemini summary |
Provides a summary of the current pull request in its current state. |
| Comment | @gemini-code-assist | Responds in comments when explicitly tagged, both in issue comments and review comments. |
| Help | /gemini help |
Displays a list of available commands. |
Customization
To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.
Limitations & Feedback
Gemini Code Assist is currently in preview and may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments to provide feedback.
You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.
Footnotes
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
There was a problem hiding this comment.
Code Review
This pull request reverts a previous change related to tensor parallelism for embeddings in DeepSeek-V2 models. The new code disables tensor parallelism for embeddings when data-parallel attention is on, which is a safer approach. A comment has been suggested to explain this logic.
| @@ -1938,7 +1938,7 @@ def __init__( | |||
| self.embed_tokens = VocabParallelEmbedding( | |||
| config.vocab_size, | |||
| config.hidden_size, | |||
| use_attn_tp_group=True, | |||
| enable_tp=not global_server_args_dict["enable_dp_attention"], | |||
There was a problem hiding this comment.
For better maintainability, consider adding a comment explaining why tensor parallelism for embeddings is disabled when data-parallel attention is enabled. This clarifies the rationale behind the logic.
# When data-parallel attention is enabled, the input batch is split across DP ranks.
# To avoid complex communication for the embedding lookup, we replicate the embedding layer by disabling its tensor parallelism.
self.embed_tokens = VocabParallelEmbedding(
config.vocab_size,
config.hidden_size,
enable_tp=not global_server_args_dict["enable_dp_attention"],
)
This reverts commit ef8a29c.
Motivation
Modifications
Checklist