ggerganov/whisper.cpp

Pull Requests Commits

ruby : add `VAD::Context#segments_from_samples`, allow Pathname, etc. (#3633)

KitaitiMakoto committed 1 day ago

Verified aa1bc0d1

scripts : Fix dSYMs path case for macOS xcframework build (#3630)

friederbluemle committed 1 day ago

Verified bf422cb7

cuda : fix compile warnings (#0)

ggerganov committed 1 day ago

acbace05

talk-llama : sync llama.cpp

ggerganov committed 1 day ago

953e503f

ggerganov committed 1 day ago

b529c061

add tensor type checking as part of cuda graph properties (llama/19186)

bssrdf committed 1 day ago

5dca0db9

sycl: implement GGML_UNARY_OP_SOFTPLUS (llama/19114)

s8322 committed 1 day ago

2a16e7a6

sycl: implement GGML_OP_TRI (llama/19089)

RachelMantel committed 1 day ago

1b3c27ef

ggml-webgpu: improve flastAttention performance by software pipelining (llama/19151)

ArberSephirotheca committed 1 day ago

829e7004

hexagon: enable offloading to Hexagon on Windows on Snapdragon (llama/19150)

tboinovski1 committed 1 day ago

2a89a3f3

cuda : fix nkvo, offload and cuda graph node properties matching (llama/19165)

ggerganov committed 1 day ago

b997e690

HIP: add mmf for CDNA (llama/18896)

zhang-hui-yulo committed 1 day ago

34a3e28a

ggml-zendnn : resolve ZenDNN backend cross-module symbol dependency (llama/19159)

z-vishal committed 1 day ago

e0a21829

CUDA: refactor topk-moe to enable more models (GLM 4.7, Nemotron etc.) (llama/19126)

am17an committed 1 day ago

62ba8b53

sycl: fix norm kernels: l2_norm, group_norm, rms_norm by remove assert to support more cases (llama/19154)

arthw committed 1 day ago

f0e85bb1

Vulkan Flash Attention Coopmat1 Refactor (llama/19075)

0cc4m committed 1 day ago

33148bb5

ggml-sycl: remove unused syclcompat header (llama/19140)

PatKamin committed 1 day ago

cc0c103b

vulkan: handle device dedup on MacOS + Vega II Duo cards (llama/19058)

okuvshynov committed 1 day ago

dda7d9cd

ggml: new backend for Virglrenderer API Remoting acceleration (v2) (llama/18718)

kpouget committed 1 day ago

531d7b67

ggml-cpu: arm64: Q4_K scale unroll and vectorization (llama/19108)

Alcpz committed 1 day ago

3701413a

cuda : fix "V is K view" check for non-unified KV cache (llama/19145)

ggerganov committed 1 day ago

7fb0f823

CUDA: tune GLM 4.7 Flash FA kernel selection logic (DGX Spark) (llama/19142)

ggerganov committed 1 day ago

f28a7330

ggml webgpu: Split shared state (webgpu_context) into global state and per-thread state (llama/18976)

nikhilJain17 committed 1 day ago

dfdd2fee

ggml-zendnn : update ZenDNN git tag to main branch (llama/19133)

z-vishal committed 1 day ago

9c75c793

CUDA: tune GLM 4.7 Flash FA kernel selection logic (llama/19097)

JohannesGaessler committed 1 day ago

9d94d0f7

ggml-cpu: aarm64: q6_K repack gemm and gemv (and generic) implementations (i8mm) #18860 (llama/18888)

Alcpz committed 1 day ago

00885e08

Reduce CPU-side stalls due to the CUDA command buffer being full (llama/19042)

gaugarg-nv committed 1 day ago

5fcbbdc0

ggml-cpu: Enable FP16 MMA kernels on PPC (llama/19060)

shalinib-ibm committed 1 day ago

b2e20328

opencl: add flattened q6_K mv (llama/19054)

lhez committed 1 day ago

56f82a9f

CUDA: fix padding of GQA to power of 2 in FA (llama/19115)

JohannesGaessler committed 1 day ago

41d5d7bb

Older