ggerganov/llama.cpp

Pull Requests Commits

Iwan Kawrakow committed 1 year ago

9f805264

iq1_s[SYCL]: remove unnecessary (unused) data

Iwan Kawrakow committed 1 year ago

9188523f

iq1_s: attempt to fix SYCL

Iwan Kawrakow committed 1 year ago

da5a6f05

ggml : reuse quantum structs across backends (#5943)

ggerganov committed 1 year ago

Verified 8030da7a

ggml : fix UB in IQ2_S and IQ3_S (#6012)

ggerganov committed 1 year ago

Verified 184215e7

sycl : update IQ1_S kernels (WIP - not working!) (#5995)

ggerganov committed 1 year ago

Verified 48358b2e

grammar : fix unnecessarily retained pointer to rules (#6003)

gliptic committed 1 year ago

Verified 5cdb3717

1.5 bit: we can do even better (#5999)

ikawrakow committed 1 year ago

Verified 44ca159f

llama : more consistent names of count variables (#5994)

ggerganov committed 1 year ago

Verified 05b06210

llama : refactor unicode stuff (#5992)

ggerganov committed 1 year ago

Verified 83796e62

Update server docker image URLs (#5997)

GetTuh committed 1 year ago

Verified 828defef

Server: format error to json (#5961)

ngxson committed 1 year ago

Verified caa106d4

ggml, ci : Windows ARM runner and build fixes (#5979)

Xarbirus committed 1 year ago

Verified 3202361c

server : maintain chat completion id for streaming responses (#5988)

mscheong01 committed 1 year ago

Verified 332bdfd7

cmake : fix subdir for `LLAMA_METAL_EMBED_LIBRARY` (#5985)

giladgd committed 1 year ago

Verified ecab1c75

llama : fix F16/F32 downcast + improve names (#5980)

ggerganov committed 1 year ago

Verified ee35600b

Better 1.5 bit quantization (#5971)

ikawrakow committed 1 year ago

Verified be858f62

[SYCL] Add q3_s and q1_s (#5886)

abhilash1910 committed 1 year ago

Verified ef3ced26

[SYCL] Add support for SYCL Nvidia target (#5738)

AidanBeltonS committed 1 year ago

Verified 3814a073

metal : move mm_id indices to shared mem (#5982)

ggerganov committed 1 year ago

Verified bb6d00bb

android : fix utf8 decoding error (#5935)

zhangfuwen committed 1 year ago

Verified 7ab7b733

readme : update hot topics

ggerganov committed 1 year ago

Verified d9f65c97

ggerganov committed 1 year ago

Verified b838b53a

ggml : try fix 32-bit arm compat (whisper/1938)

ggerganov committed 1 year ago

Verified df4dc3e7

ggml : remove __constant__ specifier for CUDA tables (#5940)

ggerganov committed 1 year ago

Verified bf47a5ee

server: ci: windows build and tests (#5968)

phymbert committed 1 year ago

Verified fa8a809a

llama : add support for GritLM (#5959)

dranger003 committed 1 year ago

Verified bcebd7db

grammar : verify parsed state (#5950)

HanClinto committed 1 year ago

Verified 2960eae8

nix: update flake.lock (#5969)

ggerganov committed 1 year ago

Verified c7854147

server: benchmark: chat/completions scenario and other llm servers comparison (#5941)

phymbert committed 1 year ago

Verified 621e86b3

Older