ggerganov/llama.cpp

Pull Requests Commits

rpc : update README for cache usage

rgerganov committed 88 days ago

c875e03f

rpc : send hash when tensor data is above some fixed threshold (#12496)

rgerganov committed 88 days ago

Verified ab6ab8f8

server : Support listening on a unix socket (#12613)

p1-0tr committed 89 days ago

Verified 2099a9d5

media : add SVG logo [no ci] (#12616)

ggerganov committed 89 days ago

Verified 29690198

opencl: add multi and vision rope, `gelu_quick` and `im2col` (#12600)

lhez committed 89 days ago

Verified 5dec47dc

llama : add PLM GGUF Conversion & Inference Support (#12457)

Si1w committed 89 days ago

Verified f125b8dc

model : restore support for T5Encoder (#12590)

HighDoping committed 89 days ago

Verified 953c2a62

convert : Support Qwen2_5_VLForConditionalGeneration (#12595)

csabakecskemeti committed 89 days ago

Verified d5c6309d

ggerganov committed 89 days ago

029c693f

scripts : update sync + fix cmake merge

ggerganov committed 89 days ago

771d8437

ggerganov committed 89 days ago

df0665a4

cmake : sync/merge PowerPC build commands (#0)

ggerganov committed 89 days ago

0306aad1

llamafile : ppc64le MMA implementation for Q4_0. (#12489)

amritahs-ibm committed 89 days ago

Verified c7b43ab6

ggml : riscv: add 128-bit RVV support (#12530)

xctan committed 89 days ago

Verified 24feaec0

llama : make loras compatible with repacking (#12593)

ggerganov committed 89 days ago

Verified f28bc4c2

SYCL: implement memset ggml backend buffer interface (#12580)

qnixsynapse committed 90 days ago

Verified f17a3bb4

HIP: Add support for RDNA4 targets (#12372)

slojosic-amd committed 90 days ago

Verified bd40678d

metal : refactor mat-vec code (#12569)

ggerganov committed 90 days ago

Verified b3298fa4

upgrade to llguidance 0.7.10 (#12576)

mmoskal committed 90 days ago

Verified 2447ad8a

clip: Fix llama-llava-clip-quantize-cli quantization error under CUDA backend (#12566)

Ivy233 committed 90 days ago

Verified 02082f15

convert : fix squeeze for ssm_conv tensors (#12573)

ggerganov committed 90 days ago

Verified df4d20cd

ggml : fix MUL_MAT_ID repack with Q8_K (#12544)

ggerganov committed 90 days ago

Verified 5ed38b68

doc: [MUSA] minor changes (#12583)

yeahdongcn committed 90 days ago

Verified fd7855f8

convert: fix Mistral3/Gemma3 model hparams init (#12571)

CISC committed 91 days ago

Verified 53af4dba

run: de-duplicate fmt and format functions and optimize (#11596)

ericcurtin committed 91 days ago

Verified ef19c717

ggml-cpu : update KleidiAI to v1.5.0 (#12568)

eddnjjn committed 91 days ago

Verified 053b3f9a

SYCL: disable Q4_0 reorder optimization (#12560)

qnixsynapse committed 91 days ago

Verified e2f56017

docs : add build instructions for KleidiAI (#12563)

eddnjjn committed 91 days ago

Verified 36ee06dd

ci: [MUSA] add CI and update doc (#12562)

yeahdongcn committed 91 days ago

Verified 3cd3a395

context : fix worst-case reserve outputs (#12545)

ggerganov committed 91 days ago

Verified 2d77d88e

Older