ngxson/llama.cpp

Pull Requests Commits

ngxson committed 3 days ago

ea8230eb

fix gguf conversion for audio/vision mmproj

ngxson committed 3 days ago

deb65438

common : add commentary rules for gpt-oss-20b (#21286)

aldehir committed 3 days ago

Verified 22337374

Relax prefill parser to allow space. (#21240)

pwilkin committed 4 days ago

Verified e15efe00

chat : add Granite 4.0 chat template with correct tool_call role mapping (#20804)

jesus-talavera-ibm committed 4 days ago

Verified 6137c325

kv-cache : do not quantize SWA KV cache (#21277)

ggerganov committed 4 days ago

Verified 17193cce

Ignore Transfer-Encoding header. (#20269)

crmky committed 4 days ago

Verified d6dac92b

ggerganov committed 4 days ago

dae2bf41

ggml : bump version to 0.9.11 (ggml/1456)

ggerganov committed 4 days ago

bc07d559

sycl : fix llama_kv_cache hang when kv_cache is huge: 5GB (#21283)

arthw committed 4 days ago

Verified 4888137b

hexagon : add cumsum op support (#21246)

tboinovski1 committed 4 days ago

Verified fbd441c3

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

ngxson committed 4 days ago

Verified c30e0122

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

lhez committed 4 days ago

Verified 95a6ebab

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

allozaur committed 4 days ago

Verified 12dbf1da

CUDA: fix FA kernel selection logic (#21271)

JohannesGaessler committed 4 days ago

Verified 86221cf6

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

martin-klacer-arm committed 4 days ago

Verified 6de97b9d

Update Dawn version in WebGPU CI (#20784)

nikhilJain17 committed 4 days ago

Verified 5a0ed515

hexagon: improve RMS_NORM and DIV accuracy (#21251)

aparmp-quic committed 4 days ago

Verified 8710e5f9

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

jbuchananr committed 4 days ago

Verified 1d6d4cf7

llama : rotate activations for better quantization (#21038)

ggerganov committed 4 days ago

Verified 744c0c73

scripts: add function call test script (#21234)

ngxson committed 5 days ago

Verified 0356e33a

ggerganov committed 5 days ago

6422036f

ggml : bump version to 0.9.10 (ggml/1454)

ggerganov committed 5 days ago

296bc053

sycl : support nvfp4 type in mul_mat (#21227)

arthw committed 5 days ago

Verified 6b949d10

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

michaelw9999 committed 5 days ago

Verified 84f82e84

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

mudler committed 5 days ago

Verified e1cb8174

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

IMbackK committed 5 days ago

Verified 88d5f8ff

ggml : fix RWKV ops thread assignment (#21226)

ggerganov committed 5 days ago

Verified d43375ff

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

taimur-10x committed 5 days ago

Verified 2b86e5ca

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

anavp-nvidia committed 5 days ago

Verified 88458164

Older