0.6.3.post1 is released on 21th April 2025

What's Changed

• fix sage_fp16_triton repeat by @ITerydh in #146
• bump to 0.6.3.post1 by @feifeibear in #147

New Contributors

• @ITerydh made their first contribution in #146

Full Changelog: 0.6.3...0.6.3.post1

0.6.3 is released on 18th April

What's Changed

• Add an option for using a triton kernel for sageattention by @intervitens in #141
• add more sage attn impl. by @feifeibear in #143
• bump to version 0.6.3 by @feifeibear in #144

New Contributors

• @intervitens made their first contribution in #141

Full Changelog: 0.6.2...0.6.3

0.6.2 is released 8th April 2025

What's Changed

• (WIP) feat: support sparse_sage_attn by @Eigensystem in #132
• remove useless tests and benchmarks by @feifeibear in #134
• feat: support sparse sage attn, passed unit test by @Eigensystem in #136
• rename test args for sparse sage by @feifeibear in #138
• feat: support flashinfer for ring attention by @ZDJeffrey in #139
• bump version to 0.6.2 by @feifeibear in #140

New Contributors

• @ZDJeffrey made their first contribution in #139

Full Changelog: 0.6.1...0.6.2

0.6.1 is released on 20th March 2025!

What's Changed

• fix qkvpacked_attn backward bugs by @feifeibear in #121
• [Fix] remove unnecessary print by @xibosun in #123
• updated pytorch attention to flash attention by @FrankLeeeee in #124
• feat: support sage attention by @Eigensystem in #126
• fix sageattention bugs by @feifeibear in #127
• fix sageattention lse return shape error by @feifeibear in #128
• fix(sage attention): layout bug by @Eigensystem in #129
• bump version to 0.6.1 by @feifeibear in #130

New Contributors

• @FrankLeeeee made their first contribution in #124

Full Changelog: 0.6.0...0.6.1

0.6.0 is released on 26th December 2024

What's Changed

• FlashAttentionImpl -> AttnType by @feifeibear in #117
• bump to 0.6.0 by @feifeibear in #118
• fix license classifier error by @feifeibear in #119

Full Changelog: 0.5.1...0.6.0

0.5.1 is released on December 26th 2024

What's Changed

• update readme pictures by @feifeibear in #111
• test with flash_attn 2.6.3 and more flexible test by @feifeibear in #113
• pytorch ring attention by @feifeibear in #115
• version bump to 0.5.1 by @feifeibear in #116

Full Changelog: 0.5.0...0.5.1

0.5.0 is released on 11th December

What's Changed

• support fa 2.7.0 and support torch only ulysses by @feifeibear in #108
• update readme with fav3 and no fa usage by @feifeibear in #109
• bump to 0.5.0 by @feifeibear in #110

Full Changelog: 0.4.3...0.5.0

Version 0.4.3 is released on 5th Dec 2024

What's Changed

• support flash_attn 2.7.0 by @xibosun in #107

New Contributors

• @xibosun made their first contribution in #107

Full Changelog: 0.4.2...0.4.3

Version 0.4.2 is released on 19th Nov 2024

What's Changed

• ulysses in benchmark by @feifeibear in #104
• flash_attn3 not directly import by @feifeibear in #105
• version to 0.4.2 by @feifeibear in #106

Full Changelog: 0.4.1...0.4.2

0.4.1 is released on Nov 15th 2024

What's Changed

• feat: add use_sync switch to ulysses by @Eigensystem in #103

Full Changelog: 0.4.0...0.4.1