Skip to content
This repository has been archived by the owner on Aug 7, 2024. It is now read-only.

[wip] add axiswise granularity to Float8Tensor #352

Open
wants to merge 1 commit into
base: gh/vkuzo/61/base
Choose a base branch
from

Conversation

vkuzo
Copy link
Contributor

@vkuzo vkuzo commented Jul 29, 2024

Stack from ghstack (oldest at bottom):

Summary:

This PR adds the axiswise scaling granularity to Float8Tensor and
ensures that basic ops like transpose and torch._scaled_mm work as
expected.

A future PR will add integration with Float8Linear.

Test Plan:

TODO

Reviewers:

Subscribers:

Tasks:

Tags:

Summary:

This PR adds the axiswise scaling granularity to `Float8Tensor` and
ensures that basic ops like transpose and `torch._scaled_mm` work as
expected.

A future PR will add integration with `Float8Linear`.

Test Plan:

TODO

Reviewers:

Subscribers:

Tasks:

Tags:

[ghstack-poisoned]
vkuzo added a commit that referenced this pull request Jul 29, 2024
Summary:

This PR adds the axiswise scaling granularity to `Float8Tensor` and
ensures that basic ops like transpose and `torch._scaled_mm` work as
expected.

A future PR will add integration with `Float8Linear`.

Test Plan:

TODO

Reviewers:

Subscribers:

Tasks:

Tags:

ghstack-source-id: 163452e97ed6a26fa5dcba01c36f49eb744484a6
Pull Request resolved: #352
@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Jul 29, 2024
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants