Iwan Kawrakow 71f5b941bf Experimenting with flash attention on Zen4 ...

This version is finally faster up to 32k tokens.
At 32k tokens it bets no-FA by 23%, at 16k by 20%,
at 8k by 10%.

2024-08-30 15:37:52 +03:00

..

cmake

Merge mainline llama.cpp (#3)

2024-07-27 07:55:01 +02:00

include

Fused soft cap and SIMD-ified GeLU (#9)

2024-08-20 17:15:47 +03:00

src

Experimenting with flash attention on Zen4

2024-08-30 15:37:52 +03:00

.gitignore

Merge mainline llama.cpp (#3)

2024-07-27 07:55:01 +02:00

CMakeLists.txt

Merge mainline - Aug 12 2024 (#17)

2024-08-12 15:14:32 +02:00