ggml-org/llama.cpp

Pull Requests Commits

kv-cache : do not quantize SWA KV cache

ggerganov committed 4 hours ago

Verified 85d5d08f

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

martin-klacer-arm committed 5 hours ago

Verified 6de97b9d

Update Dawn version in WebGPU CI (#20784)

nikhilJain17 committed 5 hours ago

Verified 5a0ed515

hexagon: improve RMS_NORM and DIV accuracy (#21251)

aparmp-quic committed 6 hours ago

Verified 8710e5f9

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

jbuchananr committed 7 hours ago

Verified 1d6d4cf7

llama : rotate activations for better quantization (#21038)

ggerganov committed 8 hours ago

Verified 744c0c73

scripts: add function call test script (#21234)

ngxson committed 8 hours ago

Verified 0356e33a

ggerganov committed 9 hours ago

6422036f

ggml : bump version to 0.9.10 (ggml/1454)

ggerganov committed 9 hours ago

296bc053

sycl : support nvfp4 type in mul_mat (#21227)

arthw committed 11 hours ago

Verified 6b949d10

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

michaelw9999 committed 12 hours ago

Verified 84f82e84

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

mudler committed 12 hours ago

Verified e1cb8174

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

IMbackK committed 13 hours ago

Verified 88d5f8ff

ggml : fix RWKV ops thread assignment (#21226)

ggerganov committed 14 hours ago

Verified d43375ff

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

taimur-10x committed 14 hours ago

Verified 2b86e5ca

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

anavp-nvidia committed 15 hours ago

Verified 88458164

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

EAddario committed 16 hours ago

Verified 49512502

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

reeselevine committed 16 hours ago

Verified 82764c34

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

abhijitramesh committed 23 hours ago

Verified 825eb91a

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

allozaur committed 1 day ago

Verified 0fcb3760

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

angt committed 1 day ago

Verified 6307ec07

CANN: fix multi-thread set_tensor race conditions (#20151)

hipudding committed 1 day ago

Verified 632219af

server: (webui) no more gzip compression (#21073)

ngxson committed 1 day ago

Verified 4a00bbfe

common : gpt-oss handle builtin and unsolicited tool calls (#21213)

aldehir committed 1 day ago

Verified 624733d6

fix: correct misspellings in code comments (#21217)

lainon1 committed 1 day ago

Verified 0b6ff479

CI: Enable CPU and Vulkan ARM64 Release (#21207)

ehfd committed 1 day ago

Verified eec6f85d

ggerganov committed 1 day ago

9281dd13

ggml : bump version to 0.9.9 (ggml/1449)

ggerganov committed 1 day ago

0be6c7c9

common : move up common_init() and fix Windows UTF-8 logs (#21176)

angt committed 1 day ago

Verified 41361c85

sycl : enhance fattn perf (#21185)

arthw committed 1 day ago

Verified 62278ced

Older