ggerganov/llama.cpp

Pull Requests Commits

Implement llama-pull tool

ericcurtin committed 2 days ago

17ca6ed5

ggerganov committed 2 days ago

7f766929

ggml : introduce semantic versioning (ggml/1336)

danbev committed 2 days ago

405921dc

CUDA : conditionally add cuda architectures (ggml/1341)

gjasny committed 2 days ago

fa6383ca

vulkan: use vec dot for matrix matrix multiplications (#16056)

0cc4m committed 2 days ago

Verified 803dac2e

server: fix SSE and OpenAI compatibility for error messages when streaming (#16109)

BenjaminBruenau committed 2 days ago

Verified 459c0c2c

llama-bench: add --devices and --list-devices support (#16039)

ssweens committed 2 days ago

Verified be79d9fd

chat: Fix streaming parser for granite models (#15682)

shun095 committed 3 days ago

Verified f432d8d8

feat: Improve mobile UI for Settings Dialog (#16084)

allozaur committed 3 days ago

Verified 4067f07f

chat : fix build on arm64 (#16101)

ngxson committed 3 days ago

Verified 4b8560ab

ggml : refactor forward_dup for cpu backend (#16062)

ngxson committed 3 days ago

Verified 0dd58b68

ggml-amx : fix ggml_amx_init() on generic Linux (#16049)

angt committed 3 days ago

Verified 69ffd891

cmake : fix static linking for OpenMP on Unix-like systems (#16031)

angt committed 3 days ago

Verified 246c0d9c

opencl: optimize mxfp4 kernels (#16037)

shawngu-quic committed 3 days ago

Verified 3edd87cd

rename optimize_graph to graph_optimize (#16082)

jeffbolznv committed 3 days ago

Verified c0b45097

CUDA: Optimize PAD_REFLECT_1D (#15957)

bugparty committed 3 days ago

Verified 38dbdf4c

CUDA: fix compilation on CC 6.0 (#16091)

JohannesGaessler committed 3 days ago

Verified 368560a1

Add resumable downloads for llama-server model loading (#15963)

ericcurtin committed 4 days ago

Verified 4ca088b0

metal : use function constants for mul_mv_ext kernels (#16074)

ggerganov committed 4 days ago

Verified 703f9e32

cuda : add missing F32<->I32 entries in ggml_cuda_cpy_fn (#16060)

CISC committed 4 days ago

Verified ad6bd908

server : include usage statistics only when user request them (#16052)

rgerganov committed 4 days ago

Verified 2b6b55a5

llama : bump max seq limit from 64 to 256 (#15916)

ggerganov committed 4 days ago

Verified e58174ce

metal : improve F32, F16 and BF16 mat-vec multiplication (#16057)

ggerganov committed 4 days ago

Verified b213fce8

metal : avoid call free for non-owned buffer (#16067)

jhen0409 committed 4 days ago

Verified e00f3fd8

metal : handle nil cv during pipeline creation (#16065)

ggerganov committed 4 days ago

Verified f2f28380

CANN: Remove print (#16044)

noemotiovon committed 4 days ago

Verified 62c3b645

GGML WebGPU: Support for ADD, MUL, RMS_NORM, GET_ROWS operators (#16018)

reeselevine committed 4 days ago

Verified d304f459

metal : refactor + optimize v2 (#15995)

ggerganov committed 4 days ago

Verified 0320ac52

SvelteKit-based WebUI (#14839)

allozaur committed 4 days ago

Verified a7a98e0f

convert : add Llama4ForCausalLM (#16042)

ngxson committed 4 days ago

Verified 8f8f2274

Older