ggml-org/whisper.cpp

Pull Requests Commits

ggml-ci: update input env variables to GG_BUILD_

redraskal committed 342 days ago

Verified 05ce7476

ggml-ci: add run.sh (#2877)

redraskal committed 342 days ago

Verified f11de0e7

examples : add dl to the list of libraries linked (#2875)

danbev committed 342 days ago

Verified d5cc27ee

whisper: add xcframework build script (#2873)

mdestagnol committed 343 days ago

Verified 5bb1d58c

objc : fix build, tmp remove GPU support, use C++17

ggerganov committed 348 days ago

7d140057

cmake : fix ggml-config (ggml/0)

ggerganov committed 348 days ago

4ffb8e3e

ggerganov committed 348 days ago

1d8d8ae5

ggml-cpu: faster AVX2 variant for IQ1_M (llama/12216)

remyoudompheng committed 348 days ago

eebf6bc0

metal : simplify kernel arguments using a struct (ggml/3229) (llama/12194)

BB-fat committed 348 days ago

dc8f423b

metal : fix default.metallib build (llama/12224)

danbev committed 348 days ago

548e7052

opencl: Noncontiguous `norm`, `rms_norm`, disable `fp16` for some ops (llama/12217)

lhez committed 348 days ago

a34cb73d

cmake : fix undefined reference errors for std::filesystem in ggml (#12092) (llama/12094)

hbuxiaofei committed 348 days ago

82f94966

CUDA: fix FA logic for PTX 7.0 and CC >= 7.5 (llama/12222)

JohannesGaessler committed 348 days ago

e3c85e75

HIP/CUDA: set the paramerter value in maintain_cuda_graph instead of replaceing it. (llama/12209)

IMbackK committed 348 days ago

b9eab73f

opencl : fix buffer alignment (llama/12197)

linehill committed 348 days ago

76385c83

opencl : fix `ulong` kernel args were set from `int` variables (llama/12174)

linehill committed 348 days ago

442cd1d2

opencl : fix profile-related errors (llama/12095)

simon886212 committed 348 days ago

bc8cb97e

ggml-cpu: Faster IQ1 mul_mat_vec on AVX2 using BMI2 instructions (llama/12154)

remyoudompheng committed 348 days ago

8dcadf73

SYCL: Disable f16 Unary OPs as not supported by the kernels (llama/12201)

qnixsynapse committed 348 days ago

93986b61

ggml : fix GGMLMetalClass ODR (llama/12200)

pminev committed 348 days ago

bd1a9e34

ggml : ggml_compute_forward_concat() for arbitrary tensor type (ggml/1118)

vmobilis committed 348 days ago

cc03608e

vulkan : sync (llama/0)

ggerganov committed 348 days ago

54a54fae

ggml : portability fixes for VS 2017 (llama/12150)

mgroeber9110 committed 348 days ago

96a92ecc

HIP: implement FlashAttention via rocWMMA for CDNA and RDNA3+ (llama/12032)

hjc4869 committed 348 days ago

edd1d868

ggml : fix kleidiai build (llama/12159)

ag2s20150909 committed 348 days ago

dc6f4e7c

SYCL: Move CPY kernels to a separate file and add few missing kernels (llama/12133)

qnixsynapse committed 348 days ago

74c85d15

ggml-backend : keep paths in native string type when possible (llama/12144)

slaren committed 348 days ago

eb2d8b6f

CUDA: compress mode option and default to size (llama/12029)

Green-Sky committed 348 days ago

b442dcd5

ggml : upgrade init_tensor API to return a ggml_status (llama/11854)

WilliamTambellini committed 348 days ago

c98681e6

vulkan: add specific MMV kernels for IQ2 and IQ3 quants + optimizations (llama/11595)

remyoudompheng committed 348 days ago

3bab8049

Older