Georgi Gerganov 7094ea5e75

whisper : use flash attention (#2152) ...

* whisper : use flash attention in the encoder

* whisper : add kv_pad

* whisper : remove extra backend instance (huh?)

* whisper : use FA for cross-attention

* whisper : use FA for self-attention

* whisper : simplify encoder FA

* whisper : add flash_attn runtime parameter

* scripts : add bench log

* scripts : add M1 Pro bench log

2024-05-15 09:38:19 +03:00

..

bench-all-gg.txt

whisper : use flash attention (#2152)

2024-05-15 09:38:19 +03:00

bench-all.sh

whisper : use flash attention (#2152)

2024-05-15 09:38:19 +03:00

bench-wts.sh

files : rename ./extra to ./scripts

2024-04-09 20:13:41 +03:00

bench.py

files : rename ./extra to ./scripts

2024-04-09 20:13:41 +03:00

convert-all.sh

files : rename ./extra to ./scripts

2024-04-09 20:13:41 +03:00

deploy-wasm.sh

files : rename ./extra to ./scripts

2024-04-09 20:13:41 +03:00

gen-authors.sh

license : update copyright notice + add AUTHORS

2024-04-09 20:27:44 +03:00

quantize-all.sh

files : rename ./extra to ./scripts

2024-04-09 20:13:41 +03:00

sha-all.sh

files : rename ./extra to ./scripts

2024-04-09 20:13:41 +03:00

sync-ggml-am.sh

scripts : sync ggml-rpc

2024-05-14 19:15:35 +03:00

sync-ggml.last

sync : ggml

2024-05-14 19:16:32 +03:00

sync-ggml.sh

scripts : sync ggml-rpc

2024-05-14 19:15:35 +03:00

sync-llama.sh

files : rename ./extra to ./scripts

2024-04-09 20:13:41 +03:00