Skip to content
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

Add support for GPTQ Marlin #2052

Merged
merged 1 commit into from
Jun 14, 2024
Merged

Add support for GPTQ Marlin #2052

merged 1 commit into from
Jun 14, 2024

Conversation

danieldk
Copy link
Member

@danieldk danieldk commented Jun 11, 2024

What does this PR do?

Add support for GPTQ Marlin kernels. GPTQ Marlin extends the Marlin kernels to support common GPTQ configurations:

  • bits: 4 or 8
  • groupsize: -1, 32, 64, or 128
  • desc_act: true/false

Using the GPTQ Marlin kernels requires repacking the parameters in the Marlin quantizer format.

The kernels were contributed by Neural Magic to VLLM. We vendor them here for convenience.

Before submitting

  • This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
  • Did you read the contributor guideline,
    Pull Request section?
  • Was this discussed/approved via a Github issue or the forum? Please add a link
    to it if that's the case.
  • Did you make sure to update the documentation with your changes? Here are the
    documentation guidelines, and
    here are tips on formatting docstrings.
  • Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

@danieldk danieldk force-pushed the feature/marlin-gptq branch 2 times, most recently from f0187ef to fd32b58 Compare June 12, 2024 11:59
@danieldk danieldk marked this pull request as ready for review June 12, 2024 12:02
GPTQ Marlin extends the Marlin kernels to support common GPTQ
configurations:

- bits: 4 or 8
- groupsize: -1, 32, 64, or 128
- desc_act: true/false

Using the GPTQ Marlin kernels requires repacking the parameters in the
Marlin quantizer format.

The kernels were contributed by Neural Magic to VLLM. We vendor them
here for convenience.
@danieldk danieldk force-pushed the feature/marlin-gptq branch from fd32b58 to 60ee0b5 Compare June 13, 2024 07:07
@@ -9,6 +10,15 @@
from text_generation_server.utils.log import log_once


@dataclass
class _GPTQParams:
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

not important: but is there a reason this class is prefixed with _?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

To make it extra clear that the class is not meant for external consumption (not a public API).

Copy link
Collaborator

@drbh drbh left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

lgtm!

@danieldk danieldk merged commit 093a27c into main Jun 14, 2024
6 checks passed
@danieldk danieldk deleted the feature/marlin-gptq branch June 14, 2024 07:45
@Narsil Narsil mentioned this pull request Jun 24, 2024
5 tasks
yuanwu2017 pushed a commit to yuanwu2017/tgi-gaudi that referenced this pull request Sep 26, 2024
Add support for GPTQ Marlin kernels

GPTQ Marlin extends the Marlin kernels to support common GPTQ
configurations:

- bits: 4 or 8
- groupsize: -1, 32, 64, or 128
- desc_act: true/false

Using the GPTQ Marlin kernels requires repacking the parameters in the
Marlin quantizer format.

The kernels were contributed by Neural Magic to VLLM. We vendor them
here for convenience.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants