Skip to content

Releases: lucidrains/native-sparse-attention-pytorch

0.0.77

04 Mar 18:14
Compare
Choose a tag to compare
fixes for triton pathway

0.0.76

04 Mar 16:55
Compare
Choose a tag to compare
causal flag for the transformer and setting correct flag for flex att…

0.0.75

04 Mar 16:33
Compare
Choose a tag to compare
some progress towards non-causal variant

0.0.73

03 Mar 16:53
Compare
Choose a tag to compare
fix

0.0.72

03 Mar 16:39
Compare
Choose a tag to compare
initial forward needs to return cache with rotated keys

0.0.71

03 Mar 16:31
Compare
Choose a tag to compare
update NSA inference so rotated queries and keys are cached

0.0.70

01 Mar 17:53
Compare
Choose a tag to compare
fix some padding issues for gating with importance score

0.0.69

01 Mar 17:42
Compare
Choose a tag to compare
move the gating back onto the selected keys for improved differentiab…

0.0.68

01 Mar 16:07
Compare
Choose a tag to compare
some intermittent issue with flex attention on sample, just disable a…

0.0.66

01 Mar 15:22
Compare
Choose a tag to compare
fix intermittent issue with triton nsa dk