ggerganov/llama.cpp

Pull Requests Commits

vulkan: Implement "fast divide" (mul+shift) for unary ops like copy (#10642)

jeffbolznv committed 337 days ago

Verified 2759916d

SYCL : Move to compile time oneMKL interface backend selection for NVIDIA backend (#10584)

s-Nick committed 337 days ago

Verified 40c6d79f

fix typo of README.md (#10605)

WrRan committed 337 days ago

Verified 98036d56

Avoid using __fp16 on ARM with old nvcc (#10616)

frankier committed 337 days ago

Verified cd2f37b3

Add docs for creating a static build (#10268) (#10630)

mostlygeek committed 337 days ago

Verified da6aac91

clip : add sycl support (#10574)

piDack committed 337 days ago

Verified 01e6d9bb

vulkan: optimize and reenable split_k (#10637)

jeffbolznv committed 337 days ago

Verified cc98896d

server : (web ui) Various improvements, now use vite as bundler (#10599)

ngxson committed 337 days ago

Verified 91c36c26

scripts : remove amx sync

ggerganov committed 337 days ago

1cd3df46

ggerganov committed 337 days ago

c5054718

CUDA: remove unnecessary warp reduce in FA (ggml/1032)

mahorozte committed 337 days ago

e9e661bd

feat: add `GGML_UNARY_OP_ARGMAX` Metal kernel (ggml/1019)

PABannier committed 337 days ago

efb6ae96

metal : add `GGML_OP_CONV_TRANSPOSE_1D` kernels (ggml/1026)

PABannier committed 337 days ago

667d70d1

llama : add missing LLAMA_API for llama_chat_builtin_templates (#10636)

ngxson committed 338 days ago

Verified 3b4f2e33

readme : add option, update default value, fix formatting (#10271)

pothitos committed 338 days ago

Verified 82bca225

metal : small-batch mat-mul kernels (#10581)

ggerganov committed 338 days ago

Verified 0115df2f

github : minify link [no ci] (revert)

ggerganov committed 338 days ago

Verified 515d4e53

github : minify link [no ci]

ggerganov committed 338 days ago

Verified 844e2e1f

server : fix default draft model parameters (#10586)

ggerganov committed 338 days ago

Verified 70b98fad

llama : add enum for built-in chat templates (#10623)

ngxson committed 338 days ago

Verified 642330ac

make : deprecate (#10514)

ggerganov committed 338 days ago

Verified 8648c521

server: Add "tokens per second" information in the backend (#10548)

lhpqaq committed 339 days ago

Verified 64ed2091

SYCL: Fix and switch to GGML_LOG system instead of fprintf (#10579)

qnixsynapse committed 339 days ago

Verified 991f8aab

contrib : refresh (#10593)

ggerganov committed 339 days ago

Verified 4cb003dd

Add `mistral-v1`, `mistral-v3`, `mistral-v3-tekken` and `mistral-v7` chat template types (#10572)

jukofyork committed 339 days ago

Verified 917786f4

grammars : add English-only grammar (#10612)

ggerganov committed 339 days ago

Verified 5e1ed955

ci: add error handling for Python venv creation in run.sh (#10608)

wangqin0 committed 339 days ago

Verified 5c7a5aa0

ggml : automatic selection of best CPU backend (#10606)

slaren committed 340 days ago

Verified 3420909d

server : bind to any port when specified (#10590)

alek3y committed 340 days ago

Verified 86dc11c5

readme : update the usage section with examples (#10596)

ggerganov committed 340 days ago

Verified 6acce397

Newer Older