ggerganov/llama.cpp

Pull Requests Commits

server : add tests

ggerganov committed 278 days ago

Verified 81611bef

server : take into account speculative limits

ggerganov committed 278 days ago

Verified b436edaa

server : fix speculative decoding with context shift

ggerganov committed 279 days ago

Verified a5a915b5

vulkan: optimize and reenable split_k (#10637)

jeffbolznv committed 279 days ago

Verified cc98896d

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

ngxson committed 279 days ago

Verified 91c36c26

scripts : remove amx sync

ggerganov committed 279 days ago

1cd3df46

ggerganov committed 279 days ago

c5054718

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

mahorozte committed 279 days ago

e9e661bd

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

PABannier committed 279 days ago

efb6ae96

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

PABannier committed 279 days ago

667d70d1

llama : add missing LLAMA_API for llama_chat_builtin_templates (#10636)

ngxson committed 279 days ago

Verified 3b4f2e33

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

pothitos committed 279 days ago

Verified 82bca225

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

ggerganov committed 279 days ago

Verified 0115df2f

github : minify link [no ci] (revert)

ggerganov committed 279 days ago

Verified 515d4e53

github : minify link [no ci]

ggerganov committed 279 days ago

Verified 844e2e1f

server : fix default draft model parameters (#10586)

ggerganov committed 279 days ago

Verified 70b98fad

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

ngxson committed 280 days ago

Verified 642330ac

make : deprecate (#10514)

ggerganov committed 280 days ago

Verified 8648c521

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

lhpqaq committed 280 days ago

Verified 64ed2091

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

qnixsynapse committed 280 days ago

Verified 991f8aab

contrib : refresh (#10593)

ggerganov committed 280 days ago

Verified 4cb003dd

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

jukofyork committed 281 days ago

Verified 917786f4

grammars : add English-only grammar (#10612)

ggerganov committed 281 days ago

Verified 5e1ed955

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

wangqin0 committed 281 days ago

Verified 5c7a5aa0

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

slaren committed 281 days ago

Verified 3420909d

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

alek3y committed 281 days ago

Verified 86dc11c5

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

ggerganov committed 281 days ago

Verified 6acce397

build: update Makefile comments for C++ version change (#10598)

wangqin0 committed 281 days ago

Verified 43957ef2

ggml-cpu: replace AArch64 NEON assembly with intrinsics in ggml_gemv_q4_0_4x4_q8_0() (#10567)

angt committed 282 days ago

Verified 0c39f44d

readme : remove old badge

ggerganov committed 282 days ago

Verified 3e0ba0e6

Older