ggml-org/whisper.cpp

Pull Requests Commits
cuda : avoid async allocs in CUDA mel code
ggerganov committed 2 years ago
Verified 267e15a4
cuda : fix HIPBLAS build (#2234)
ggerganov committed 2 years ago
Verified 420b6abc
cuda : fix bounds check for src0 rows in MMVQ kernel (#2231)
ggerganov committed 2 years ago
Verified 99804b0f
ci : fix CUDA builds (#2232)
ggerganov committed 2 years ago
Verified c55964c9
whisper : auto-grow working areas for mel_calc_cuda (#2227)
iboB committed 2 years ago
Verified 20c542c7
whisper : free whisper_mel instances (#2220)
ggerganov committed 2 years ago
Verified c2bdb960
whisper : whisper_state/backend fixes (#2217)
ggerganov committed 2 years ago
Verified 87acd6d6
whisper : calculate mel spectrogram directly into a ggml_tensor (#2208)
iboB committed 2 years ago
Verified f842d311
whisper : add CUDA-specific computation mel spectrograms (#2206)
iboB committed 2 years ago
Verified ffef323c
whisper : remove `speed_up` and `phase_vocoder*` functions (#2198)
iboB committed 2 years ago
Verified af5833e2
readme : add conan badge (#2196)
MartinDelille committed 2 years ago
Verified b87494bb
readme : add install instructions for Conan (#2189)
czoido committed 2 years ago
Verified ad130431
whisper: use global cache for sin/cos vals and Hann window (#2194)
iboB committed 2 years ago
Verified e130b666
release : v1.6.2
ggerganov committed 2 years ago
Verified c7b69886
Revert "whisper : remove extra backend instance (huh?)" (#2182)
ggerganov committed 2 years ago
Verified 05042a78
server : fix typo (#2181)
dvaldivia committed 2 years ago
Verified a7dc2aab
ruby : update bindings (#2154)
taf2 committed 2 years ago
Verified 22d46b7b
release : v1.6.1
ggerganov committed 2 years ago
Verified c10db6ea
examples : add support for decoding input with ffmpeg (Linux) (#2133)
WilliamTambellini committed 2 years ago
Verified 1b51fdf1
node : add flash_attn param (#2170)
pprobst committed 2 years ago
Verified adee3f9c
ci: Update build.yml to suppress warnings about node.js versions (#2166)
tamo committed 2 years ago
Verified 4798be1f
release : v1.6.0
ggerganov committed 2 years ago
Verified 08981d1b
whisper : use flash attention (#2152)
ggerganov committed 2 years ago
Verified 7094ea5e
talk-llama : reject runs without required arguments (#2153)
petterreinholdtsen committed 2 years ago
Verified 9d5771ae
sync : ggml
ggerganov committed 2 years ago
Verified f56b8305
metal : support FA without mask + add asserts (llama/7278)
ggerganov committed 2 years ago
Verified 1056ad76
ggml : add RPC backend (llama/6829)
rgerganov committed 2 years ago
Verified c451080c
rm wait() (llama/7233)
arthw committed 2 years ago
Verified 8e7c22fb
CUDA: add FP32 FlashAttention vector kernel (llama/7188)
JohannesGaessler committed 2 years ago
Verified e57e95eb
scripts : sync ggml-rpc
ggerganov committed 2 years ago
Verified 130f43e4
Terms Of Use Privacy Policy Refund Policy Impressum