Skip to content

Releases: lucidrains/memory-efficient-attention-pytorch

0.0.23

23 Jul 21:32
Compare
Choose a tag to compare
add tests, fix bugs

0.0.22

23 Jul 20:00
Compare
Choose a tag to compare
oops, actually add cosine sim flash attention

0.0.21

23 Jul 19:59
Compare
Choose a tag to compare
oops, actually add cosine sim flash attention

0.0.20

23 Jul 19:59
Compare
Choose a tag to compare
add cosine sim attention variant of flash attention

0.0.19

23 Jul 19:49
Compare
Choose a tag to compare
fix bugs

0.0.18

23 Jul 19:39
Compare
Choose a tag to compare
complete flash attention algorithm in plain pytorch (for educational …

0.0.17

22 Mar 17:38
Compare
Choose a tag to compare
directly calculate triu shift value for causal mask for memory savings

0.0.16

21 Mar 03:24
Compare
Choose a tag to compare
materialize causal mask only when needed, to reduce peak memory usage…

0.0.15

13 Mar 20:37
Compare
Choose a tag to compare
chunk feedforward for entirely memory efficient transformer

0.0.14

04 Mar 22:44
Compare
Choose a tag to compare
max for numerical stability should be taken after masking