ggml-org/whisper.cpp

Pull Requests Commits

parakeet : update parakeet test models generation

danbev committed 74 days ago

Verified 0aec3b24

ggerganov committed 75 days ago

Verified 2f1f1a86

ggml : bump version to 0.18.0 (ggml/1576)

ggerganov committed 75 days ago

Verified 6f6960d6

CUDA: add Q2_0 support (llama/25707)

khosravipasha committed 75 days ago

Verified 6366ca02

metal: fix memory unwire if model is freed without any GPU operations (llama/26082)

nikwen committed 75 days ago

Verified 2b8f0f18

ggml : Fix issue with kleidiai ci and stringop overflow warning (llama/26277)

JonathanC-ARM committed 75 days ago

Verified c2b2ecb6

ggml-cuda : disable MMQ on devices with less than 48 KiB shared memory (llama/26141)

KakaruHayate committed 75 days ago

Verified a816f406

sycl: contiguous fast path + 32-bit index math for unary elementwise ops (llama/25946)

Titaniumtown committed 75 days ago

Verified e103c721

RPC: add tensor_memset (llama/25912)

am17an committed 75 days ago

Verified 58e109a9

add rdna3.5, and 3 to mmq configs so they can be tuned independently. (llama/26199)

Geramy committed 75 days ago

Verified 3ab356fe

ggml-webgpu: Fix some binding alias issues to support all archs, fix recurrent-state-rollback test (llama/25931)

reeselevine committed 75 days ago

Verified 2709c63e

opencl: skip the Adreno KQ/KQV image kernels for multi-stream batches (llama/26189)

wanghqc committed 75 days ago

Verified 3583bc38

ggml : set output of view src (llama/25729)

am17an committed 75 days ago

Verified 98dc1760

vulkan: add iq4_nl support back to FA (llama/24585)

jeffbolznv committed 75 days ago

Verified cb8a0ad8

ggml-cuda: add chunked SSD matmul for Mamba-2 prefill acceleration (llama/22675)

BLSharda committed 75 days ago

Verified 1e4d28c3

sycl: fix use-after-return of the SDPA scale in the oneDNN flash-attention path (llama/25880)

meatposes committed 75 days ago

Verified e43559f5

ggml-metal: FWHT kernel for metal backend (llama/25924)

nicklafleur committed 75 days ago

Verified 00009c89

Disable -ffast-math on HIP (llama/25495)

Beinsezii committed 75 days ago

Verified 0c1bd5dc

sycl(build): parallelize ocloc invocations (llama/25903)

Titaniumtown committed 75 days ago

Verified 7ebe1c2b

ggml : adjust logic for offloading ops to weight's backend (llama/25832)

ggerganov committed 75 days ago

Verified 2fa1559f

ggml-cpu: Enable BF16 tiled gemm optimization on PowerPC (llama/26068)

shalinib-ibm committed 75 days ago

Verified 98a532e1

opencl: fix fused RMS norm mul view offset (llama/26085)

happyyzy committed 75 days ago

Verified 2e5b78c1

hexagon: partial im2col support (llama/26007)

tboinovski1 committed 75 days ago

Verified a798a7e3

Update ggml/src/gguf.cpp : Defined virtual keyword for destructor of gguf_writer_base (llama/25867)

yymin1022 committed 75 days ago

Verified 6d4dc97c

HIP: remove rocWMMA FlashAttention (llama/26046)

JohannesGaessler committed 75 days ago

Verified 9dbbfa47

opencl: cache compiled cl_program binaries on disk (llama/26050)

wanghqc committed 75 days ago

Verified 526d0ed9

opencl: do not treat NULL-mask flash attention as causal (llama/25771)

kumaal committed 75 days ago

Verified ce735813

hexagon: further improved pipeline of the core bits (L2, DMA, MM, FA) (llama/26049)

max-krasnyansky committed 75 days ago

Verified ddc63beb

hexagon: fix Windows crash when op_poll is enabled (llama/26029)

adgup committed 75 days ago

Verified 47a3239d

CUDA: fix external compilation of q1_0 MMQ (llama/25778)

JohannesGaessler committed 75 days ago

Verified 49d48deb

Older