-
-
Notifications
You must be signed in to change notification settings - Fork 4.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Model] Support quantization of PixtralHFTransformer for PixtralHF #9921
[Model] Support quantization of PixtralHFTransformer for PixtralHF #9921
Conversation
Signed-off-by: mgoin <[email protected]>
👋 Hi! Thank you for contributing to the vLLM project. Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging. To run CI, PR reviewers can do one of these:
🚀 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM! Not related to this PR in particular but does it make sense to use get_act_and_mul_fn
and act_and_mul
everywhere?
Yeah this was my thinking @ywang96, with |
…llm-project#9921) Signed-off-by: mgoin <[email protected]> Signed-off-by: Loc Huynh <[email protected]>
…llm-project#9921) Signed-off-by: mgoin <[email protected]> Signed-off-by: Sumit Dubey <[email protected]>
Allows us to support quantization of the vision encoder for PixtralHF, where before we could only quantize the language model. Also converts the raw torch nn.Linear modules to proper merged vLLM Linear modules
Tested with
neuralmagic/pixtral-12b-FP8-dynamic
(onlyMistralModel
quantized)nm-testing/pixtral-12b-FP8-dynamic-all
(MistralModel
andPixtralHFTransformer
quantized, just notlm_head
)