ggerganov/llama.cpp

Pull Requests Commits

sycl : fix grid type

ggerganov committed 1 year ago

Verified 76be02ae

sycl : iq1s_grid -> iq1s_grid_gpu

ggerganov committed 1 year ago

Verified cb5a702e

sycl : try to fix after IQ1_S changes

ggerganov committed 1 year ago

Verified 77d586f5

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

server : print chat template info

ggerganov committed 1 year ago

Verified 77d1ac7e

perplexity : support using multiple sequences to allow larger batch sizes (#5946)

slaren committed 1 year ago

Verified d894f352

readme : update hot topics

ggerganov committed 1 year ago

Verified 098dbaab

ggml : fix unnecessary f32 -> f16 -> f32 casts (mmla) (#5951)

ggerganov committed 1 year ago

Verified 8380ecfb

server : fix metrics init (#5964)

ggerganov committed 1 year ago

Verified 58308a0e

ggml : remove old quantization functions (#5942)

ggerganov committed 1 year ago

Verified 5b097973

server : clarify some items in the readme (#5957)

ggerganov committed 1 year ago

Verified 97c09585

server : normalize embeddings (#5956)

redlion0929 committed 1 year ago

Verified fb215c38

Older