ggerganov/llama.cpp

Pull Requests Commits

metal : make the NSG a function constant in mul_mv kernels

ggerganov committed 2 hours ago

Verified 64c6dcbe

metal : improve F32, F16 and BF16 mat-vec multiplication

ggerganov committed 3 hours ago

Verified 320f0296

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

jhen0409 committed 4 hours ago

Verified e00f3fd8

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

ggerganov committed 4 hours ago

Verified f2f28380

CANN: Remove print (#16044)

noemotiovon committed 9 hours ago

Verified 62c3b645

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

reeselevine committed 15 hours ago

Verified d304f459

metal : refactor + optimize v2 (#15995)

ggerganov committed 17 hours ago

Verified 0320ac52

SvelteKit-based WebUI (#14839)

allozaur committed 17 hours ago

Verified a7a98e0f

convert : add Llama4ForCausalLM (#16042)

ngxson committed 17 hours ago

Verified 8f8f2274

CUDA: fix FA occupancy, optimize tile kernel (#15982)

JohannesGaessler committed 21 hours ago

Verified c959b676

common : Fix corrupted memory error on json grammar initialization (#16038)

dralves committed 1 day ago

Verified cd08fc3e

vulkan: automatically remove unsupported devices (#15976)

netrunnereve committed 1 day ago

Verified cb5bb6cc

ci : revert back to macos-13 for macOS-latest-cmake-x64 (#16040)

danbev committed 1 day ago

Verified a91d035b

llama-quant : fix the verification of attention layers for encoder-decoder models (#16023)

DamonFool committed 1 day ago

Verified 745cbcf2

examples : support encoder-decoder models in the simple example (#16002)

DamonFool committed 1 day ago

Verified 1cbd80f8

model : add OLMo3 support (#16015)

2015aroras committed 1 day ago

Verified 85286f35

CANN: Optimize ggml_cann_set_device (#15935)

noemotiovon committed 1 day ago

Verified d5fabe36

llama-bench: add --n-cpu-moe support (#15952)

jacekpoplawski committed 1 day ago

Verified 8ff20609

ci : use macos-latest for arm64 webgpu build (#16029)

danbev committed 1 day ago

Verified 77475530

ggml : fix padding in timestep embedding kernels (#15932)

danbev committed 1 day ago

Verified 3913f873

ci : upload xcframework artifact from ios-xcode-build job (#16010)

danbev committed 1 day ago

Verified 76888d20

fix: apply clang-format to CUDA macros (#16017)

bugparty committed 2 days ago

Verified f1fbffb5

ci : update macos-latest* jobs to use macos-latest (#15938)

danbev committed 2 days ago

Verified 51abc96b

cmake : Do not install tools on iOS targets (#15903)

ykhrustalev committed 2 days ago

Verified 07808ebb

Add LLaDA-7b-MoE diffusion model (#16003)

am17an committed 2 days ago

Verified 6d758839

CUDA: fix im2col_3d to respect non-contiguous inputs (views) (#15956)

jakekarnes42 committed 2 days ago

Verified 3d4053f7

docker : enable rocWMMA in ROCm images, add gfx1151 (#15997)

slaren committed 2 days ago

Verified dc381aa9

releases : switch to rocWMMA develop branch, add gfx1151 (#15992)

slaren committed 2 days ago

Verified 10d19740

SYCL: Add COUNT_EQUAL operator support (#15991)

yael-works committed 2 days ago

Verified b907255f

llama-run: Fix model download on Windows (#15988)

npopov-vst committed 3 days ago

Verified 28c39da7

Older