Fix TinyGemmQBitsTensor move #246
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
This will allow to detach TinyGemmQBitsTensor that have different inner tensors (scale and shift are combined).
Since QBitsTensor ops are now all compatible with TinyGemmQBitsTensor we can remove the specific dispatch.
The QBitsTensor.create factory method checks for CUDA version, but the unit tests that bypass that method must check themselves.
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Outdated suggestions cannot be applied.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.
What does this PR do?
This fixes a few issues on CUDA with arch <= sm80: a TinyGemmQBitsTensor could be moved to the CUDA device without calling QBitsTensor::create.
A small change is also added to remove a dual dispatch for AWQ gemm. This slightly improves the decode latency for LLM (but does not impact much the end-to-end latency that is still dominated by the prefill latency).