ggerganov/whisper.cpp

Pull Requests Commits

whisper : add whisper_token_count helper

ggerganov committed 1 year ago

Verified ba695788

whisper : improve handling of prompts

ggerganov committed 1 year ago

Verified 5c2c07d4

ruby : fix build (#1980)

ggerganov committed 1 year ago

Verified 48a14520

docker : libcuda.so.1 in PATH (#1966)

tiagofassoni committed 1 year ago

Verified 79d5765e

readme : add Fedora dependencies (#1970)

Man2Dev committed 1 year ago

Verified 04e48094

whisper : token-level timestamps with DTW (#1485)

denersc committed 1 year ago

Verified 741abb16

examples : rename --audio-context to --audio-ctx per help text (#1953)

joliss committed 1 year ago

Verified e7794a86

whisper : set outputs from conv graph (#1959)

ggerganov committed 1 year ago

Verified 725350d4

alloc : fix allocation data of pre-allocated leafs

slaren committed 1 year ago

Verified 906c73b2

cmake : copy ggml-common.h to bin

ggerganov committed 1 year ago

Verified 00d80ff9

gitignore : .vimspector.json

ggerganov committed 1 year ago

Verified 1b553b98

talk-llama : sync llama.cpp

ggerganov committed 1 year ago

Verified de4d067f

ggerganov committed 1 year ago

Verified e715f6a6

update examples and tests

slaren committed 1 year ago

Verified f60ccfd8

ggml : add ggml-common.h

ggerganov committed 1 year ago

Verified 3753a2b2

ggml : designate enum vals for integer types (llama/6050)

ggerganov committed 1 year ago

Verified 592dd256

metal : build metallib + fix embed path (llama/6015)

ggerganov committed 1 year ago

Verified c8709d46

llama : add pipeline parallelism support (llama/6017)

slaren committed 1 year ago

Verified 8932c2d6

Update get version (llama/6025)

AidanBeltonS committed 1 year ago

Verified 2bddfdd7

ggml : reuse quantum structs across backends (llama/5943)

ggerganov committed 1 year ago

Verified 46e3c3f1

ggml : fix UB in IQ2_S and IQ3_S (llama/6012)

ggerganov committed 1 year ago

Verified ef24ae0c

sycl : update IQ1_S kernels (WIP - not working!) (llama/5995)

ggerganov committed 1 year ago

Verified a753926f

1.5 bit: we can do even better (llama/5999)

ikawrakow committed 1 year ago

Verified 9dc60fc0

ggml, ci : Windows ARM runner and build fixes (llama/5979)

Xarbirus committed 1 year ago

Verified d73a6362

Better 1.5 bit quantization (llama/5971)

ikawrakow committed 1 year ago

Verified f79d0d4f

Add q3_s and q1_s (llama/5886)

abhilash1910 committed 1 year ago

Verified 4f88940f

metal : move mm_id indices to shared mem (llama/5982)

ggerganov committed 1 year ago

Verified 7bdb1de9

ggml : fix unnecessary f32 -> f16 -> f32 casts (mmla) (llama/5951)

ggerganov committed 1 year ago

Verified 653d2e8f

ggml : remove old quantization functions (llama/5942)

ggerganov committed 1 year ago

Verified 2fef660d

ggml : add ggml-common.h to deduplicate shared code (llama/5940)

ggerganov committed 1 year ago

Verified 24eba5a2

Older