CUDA: faster FlashAttention, kernel for bs == 1

Browse source

...

This commit is contained in:

Johannes Gäßler 2024-03-29 23:02:39 +01:00 • committed by Georgi Gerganov

parent 08e69c5008

commit 75aa7b4b18

1 changed files with 937 additions and 482 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download patch file Download diff file Expand all files Collapse all files

1419

ggml-cuda/fattn.cu

View file

File diff suppressed because it is too large Load diff