Releases: fkodom/grouped-query-attention-pytorch
Releases · fkodom/grouped-query-attention-pytorch
0.3.0
0.2.0
0.1.0
First full release candidate for PyPI. 🎉
- Working version of GQA
- Reproduced runtime benchmarks
- Prototype
GQATransformer
modules - (Experimental) Convert T5 models to GQA
Full Changelog: 0.1.0rc1...0.1.0