Skip to content

[Quantization] From config Quantization for FP8 - #42944

Closed
MekkCyber wants to merge 5 commits into
mainfrom
from-config-quantization
Closed

MekkCyber wants to merge 5 commits into
mainfrom
from-config-quantization

Conversation

@MekkCyber

Copy link
Copy Markdown
Contributor

What does this PR do?

Make the FP8 logic accessible from from_config instead of having to call from_pretrained

@HuggingFaceDocBuilderDev

Copy link
Copy Markdown

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

@ArthurZucker

Copy link
Copy Markdown
Collaborator

@MekkCyber this one has been waiting since 2025-12-19 (overlap with #46324 + quantizer refactor #49548). Still planned, or should we close it?

@ArthurZucker

Copy link
Copy Markdown
Collaborator

Hi! This overlaps with #46324, which makes cls(config) / from_config honor quantization_config through a generic per-quantizer opt-in, with FP8 as the first user. All three files here also conflict with main now. Should we consolidate on #46324 and close this one, or is there something here (e.g. the FP8 weight init) that should be carried over?

@MekkCyber

Copy link
Copy Markdown
Contributor Author

@ArthurZucker feel free to close it, the logic in #46324 should be enough, Thanks for the notice

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