Deprecate sparse attention - #8493
Conversation
Signed-off-by: Hongwei Chen <hongweichen@microsoft.com>
Signed-off-by: Hongwei Chen <hongweichen@microsoft.com>
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: cf61b7154a
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| "quantize_training": | ||
| "Mixture-of-Quantization (MoQ) / 'quantize_training' has been removed. See " | ||
| f"{_REMOVED_FEATURES_ISSUE}.", | ||
| "sparse_attention": |
There was a problem hiding this comment.
This is a non-merge commit, but its commit message has no Signed-off-by trailer, violating the repository's mandatory commit requirement. Add a sign-off using the configured Git name and email before merging.
AGENTS.md reference: AGENTS.md:L8-L8
Useful? React with 👍 / 👎.
1 participant
Removes DeepSpeed’s legacy Sparse Attention feature as part of #8489, including its Python/Triton/C++ implementation, op builder,
sparse_attninstallation extra, and associated documentation and build/packaging entries.The sparse-attention APIs are no longer available, and configurations containing
sparse_attentionnow fail explicitly with DeepSpeedConfigError .