Support local mask #5015
Open
Support local mask #5015
Conversation
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
Summary: This diff updates the code to enable BF16 enablement with latest Cutlass version. The changes include updating the code in the `blackwell_gen_impl.cu` and `collective/sm100_fmha_gen_mainloop_warpspecialized.hpp` files to support BF16 data type. The `fmha.hpp` file also includes a check to ensure that the SMEM usage does not exceed the capacity. Differential Revision: D84624233
Summary: Add stand-alone blackwell decode op. Supported mask: BlockDiagonalCausalWithOffsetPaddedKeysMask Differential Revision: D84630701
Summary: This diff introduces changes to support local masks in the decode attn implementation. The changes include adding window_left and window_right parameters to the decode function, modifying the GenRunner class to include a Mask template parameter, and modifying the collective_builder to include a Mask parameter. The changes also include modifying the load_cpasync_warpspecialized class to include window_size_left and window_size_right parameters. Currrently, softmax is applied in a 3-loop setting. Next: Optimize these iteration and benchmark perf. Differential Revision: D84778050
Contributor
|
@Aya-ZIbra has exported this pull request. If you are a Meta employee, you can view the originating Diff in D84778050. |
✅ Deploy Preview for pytorch-fbgemm-docs ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
Summary:
This diff introduces changes to support local masks in the decode attn implementation. The changes include adding window_left and window_right parameters to the decode function, modifying the GenRunner class to include a Mask template parameter, and modifying the collective_builder to include a Mask parameter. The changes also include modifying the load_cpasync_warpspecialized class to include window_size_left and window_size_right parameters.
Currrently, softmax is applied in a 3-loop setting.
Next: Optimize these iteration and benchmark perf.
Differential Revision: D84778050