ngxson/llama.cpp

Pull Requests Commits

qwen3 asr working

ngxson committed 10 days ago

eefcfeed

convert ASR model ok

ngxson committed 10 days ago

172865e9

no more deepstack for audio

ngxson committed 10 days ago

792edb7d

ngxson committed 10 days ago

e0adcf72

Merge branch 'master' into xsn/qwen3a

ngxson committed 10 days ago

2bd09667

contrib : rewrite AGENTS.md, make it more clear about project values (#21270)

ngxson committed 10 days ago

Verified c30e0122

opencl: fix leak in Adreno q8_0 path (#21212)

lhez committed 10 days ago

Verified 95a6ebab

server: Bypass API Key validation for WebUI static bundle assets (#21269)

allozaur committed 10 days ago

Verified 12dbf1da

CUDA: fix FA kernel selection logic (#21271)

JohannesGaessler committed 10 days ago

Verified 86221cf6

kleidiai: add CPU feature detection to CI run script (#20394)

martin-klacer-arm committed 10 days ago

Verified 6de97b9d

Update Dawn version in WebGPU CI (#20784)

nikhilJain17 committed 10 days ago

Verified 5a0ed515

hexagon: improve RMS_NORM and DIV accuracy (#21251)

aparmp-quic committed 10 days ago

Verified 8710e5f9

fix: tool call parsing for LFM2 and LFM2.5 models (#21242)

jbuchananr committed 10 days ago

Verified 1d6d4cf7

llama : rotate activations for better quantization (#21038)

ggerganov committed 10 days ago

Verified 744c0c73

scripts: add function call test script (#21234)

ngxson committed 10 days ago

Verified 0356e33a

Merge branch 'master' into xsn/qwen3a

ngxson committed 10 days ago

948f614f

ggerganov committed 10 days ago

6422036f

ggml : bump version to 0.9.10 (ggml/1454)

ggerganov committed 10 days ago

296bc053

sycl : support nvfp4 type in mul_mat (#21227)

arthw committed 10 days ago

Verified 6b949d10

ggml-cuda: Add generic NVFP4 MMQ kernel (#21074)

michaelw9999 committed 10 days ago

Verified 84f82e84

memory: respect unified KV cache in hybrid memory for eval tasks (#21224)

mudler committed 10 days ago

Verified e1cb8174

CUDA/HIP: Fix kernel slection for mmvq mmid kernel to align host selection with device launch bounds (#21238)

IMbackK committed 10 days ago

Verified 88d5f8ff

ggml : fix RWKV ops thread assignment (#21226)

ggerganov committed 10 days ago

Verified d43375ff

ggml-cpu: fix fallback for RVV kernels without zvfh (#21157)

taimur-10x committed 10 days ago

Verified 2b86e5ca

CUDA: Add Flash Attention Support for Head Dimension 512 (#20998)

anavp-nvidia committed 10 days ago

Verified 88458164

llama : refactor llama_model_quantize_params to expose a pure C interface (#20346)

EAddario committed 10 days ago

Verified 49512502

ggml webgpu: quantized buffers to u32 + wider browser/device support (#21046)

reeselevine committed 10 days ago

Verified 82764c34

ggml-webgpu: port all AOT operators to JIT (#20728)

abhijitramesh committed 11 days ago

Verified 825eb91a

fix: Use lower-case proxy headers naming (#21235)

allozaur committed 11 days ago

Verified 0fcb3760

common : cleanup logs and modernize the progress bar (#21215)

angt committed 11 days ago

Verified 6307ec07

Older