metal-flash-attention-rs

public

9 stars

0 forks

0 issues

Commits

List of commits on branch main.

Unverified

56171a690723f623fb5ffe6a86c48d4e114b0ba0

Refactoring

iivarflakstad committed a year ago

Unverified

6caf34fe7e82bff31933fddb8dfd9ddec4c18833

Almost works...

iivarflakstad committed a year ago

Unverified

e8ce37531bfbad31b2614ac32cffe6511281beea

Add README

iivarflakstad committed a year ago

Unverified

09aa7a93b4094d45ff748f25150d0eb02c56bf4f

Setting up metal flash attention integration

iivarflakstad committed a year ago

Unverified

b98891e6a08136c92489237360d7972baf8cf7b9

Metal flash attention submodule

iivarflakstad committed a year ago

Unverified

10ce16dfce455abbb63c2b4c4513c14ee2b65e77

Init

iivarflakstad committed a year ago

README

The README file for this repository.

Metal Flash Attention Rust API

Work in progress!

Use MFA from Rust. Metal Flash Attention is a faster alternative to Metal Performance Shaders, a reference implementation of modern GPU algorithms, and a step toward defragmenting the AI ecosystem. See metal flash attention