ggml-org/llama.cpp

Pull Requests Commits

one less magic number

ngxson committed 99 days ago

e83ef747

refactor softplus fn

ngxson committed 99 days ago

f643b957

split q_proj/gate

ngxson committed 99 days ago

46110e06

Getting to decode stage...

pwilkin committed 101 days ago

178230ee

Merge branch 'ggml-org:master' into qwen3_next

pwilkin committed 101 days ago

Verified c78f9fce

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

rgerganov committed 101 days ago

Verified 2b6b55a5

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

ggerganov committed 101 days ago

Verified e58174ce

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

ggerganov committed 101 days ago

Verified b213fce8

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

jhen0409 committed 101 days ago

Verified e00f3fd8

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

ggerganov committed 101 days ago

Verified f2f28380

CANN: Remove print (#16044)

noemotiovon committed 101 days ago

Verified 62c3b645

pwilkin committed 101 days ago

344331c2

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

reeselevine committed 102 days ago

Verified d304f459

metal : refactor + optimize v2 (#15995)

ggerganov committed 102 days ago

Verified 0320ac52

SvelteKit-based WebUI (#14839)

allozaur committed 102 days ago

Verified a7a98e0f

convert : add Llama4ForCausalLM (#16042)

ngxson committed 102 days ago

Verified 8f8f2274

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

JohannesGaessler committed 102 days ago

Verified c959b676

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

dralves committed 102 days ago

Verified cd08fc3e

vulkan: automatically remove unsupported devices (#15976)

netrunnereve committed 102 days ago

Verified cb5bb6cc

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

danbev committed 102 days ago

Verified a91d035b

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

DamonFool committed 102 days ago

Verified 745cbcf2

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

DamonFool committed 102 days ago

Verified 1cbd80f8

model : add OLMo3 support (#16015)

2015aroras committed 102 days ago

Verified 85286f35

CANN: Optimize ggml_cann_set_device (#15935)

noemotiovon committed 102 days ago

Verified d5fabe36

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

jacekpoplawski committed 103 days ago

Verified 8ff20609

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

danbev committed 103 days ago

Verified 77475530

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

danbev committed 103 days ago

Verified 3913f873

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

danbev committed 103 days ago

Verified 76888d20

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

bugparty committed 103 days ago

Verified f1fbffb5

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

danbev committed 103 days ago

Verified 51abc96b

Older