ggerganov/llama.cpp

Pull Requests Commits

examples: add model-backend-compare tool to compare intermediate device tensors with CPU reference

0cc4m committed 13 days ago

05429433

CANN: supports out_prod operator for F32 and F16 (#17406)

TianHao324 committed 13 days ago

Verified 064c90d8

webui: add rehype plugin to restore HTML in Markdown table cells (#17477)

ServeurpersoCom committed 14 days ago

Verified b1846f1c

vulkan: Use fewer rows for scalar FA when HS is not a multiple of 16 (#17455)

jeffbolznv committed 14 days ago

Verified d414db02

llama: introduce support for model-embedded sampling parameters (#17120)

taronaeo committed 14 days ago

Verified 877566d5

vulkan: more FA details in vk_perf_logger (#17443)

jeffbolznv committed 14 days ago

Verified 3d07caa9

llama : skip output reordering for single token batches (#17466)

danbev committed 14 days ago

Verified 134e6940

HIP: WMMA-MMQ kernels for RDNA 4 (#17156)

jiachengjason committed 14 days ago

Verified 0543f928

convert : allow quantizing lora again (#17453)

CISC committed 14 days ago

Verified b61de2b2

server: split server.cpp code into server/common/task/queue (#17362)

ngxson committed 14 days ago

Verified b8372eec

examples : add -kvu to batched usage example [no ci] (#17469)

danbev committed 14 days ago

Verified 6ab8eacd

ggerganov committed 14 days ago

2d50b9d8

ggml : remove dirty flag from version string (ggml/1391)

danbev committed 14 days ago

697edfee

ggml-cpu: arm64: q4_K repack gemm and gemv implementations (i8mm) (#16739)

Alcpz committed 14 days ago

Verified dbb852b5

ggml: add RISC-V cpu-feats (#17461)

ixgbe committed 14 days ago

Verified 5f55c385

models : Added support for RND1 Diffusion Language Model (#17433)

wp4032 committed 15 days ago

Verified 4902eebe

hexagon: add support for ROPE_NEOX (#17458)

max-krasnyansky committed 15 days ago

Verified 923ae3c6

CANN: Define `cann_graph_update_required` before macro (#17434)

rauletorresc committed 15 days ago

Verified 01ad35e6

ggml-hexagon: Initial Hexagon v68/v69 support (#17394)

mediouni-m committed 15 days ago

Verified fcb01384

ggml-hexagon: add `hex_supported_buffer` for better buffer supported check (#17212)

chraac committed 15 days ago

Verified d5bc1ad1

webui: minor settings reorganization and add disable autoscroll option (#17452)

ServeurpersoCom committed 15 days ago

Verified 0c7220db

cuda : support non-contiguous i32 to i32 copy (#17326)

CISC committed 15 days ago

Verified 96ac5a23

vulkan: Update docker image to Ubuntu 26.04 to enable glslc features (#17439)

ericcurtin committed 15 days ago

Verified bc809e9c

vulkan: remove a couple unnecessary switches (#17419)

jeffbolznv committed 16 days ago

Verified 54d83bbe

ci : switch to BoringSSL on Server workflow (#17441)

angt committed 16 days ago

Verified 4949ac0f

Revive MUL_MAT_ID to perf testing (#17397)

rillomas committed 16 days ago

Verified 3f3a4fb9

HIP: RDNA4 tensor core support for MMF (#17077)

zhang-hui-yulo committed 17 days ago

Verified 028f93ef

opencl: refine condition for kqv mm (#17392)

lhez committed 17 days ago

Verified 8e9ddba6

model : detect GigaChat3-10-A1.8B as deepseek lite (#17420)

ubergarm committed 17 days ago

Verified 23bc779a

cmake : add option to build and link BoringSSL (#17205)

angt committed 17 days ago

Verified 28175f85

Older