ggml-org/llama.cpp

Pull Requests Commits

server: disable debug release type sanitizer, simplify trigger (#6047)

phymbert committed 1 year ago

Verified 43241adf

llama : fix typo

ggerganov committed 1 year ago

Verified a44bc969

llama : optimize defrag moves + fix fragmentation calculation (#6037)

Xarbirus committed 1 year ago

Verified 2c4fb692

gguf-py : add support for I8, I16 and I32 (#6045)

certik committed 1 year ago

Verified 3ca23481

ggml : designate enum vals for integer types (#6050)

ggerganov committed 1 year ago

Verified 3fe8d7a1

embedding : print all resulting embeddings (#899)

ggerganov committed 1 year ago

Verified 68265ebf

metal : build metallib + fix embed path (#6015)

ggerganov committed 1 year ago

Verified 381da2d9

embedding : print cosine similarity (#899)

ggerganov committed 1 year ago

Verified 0fd6c1f0

readme : update details about running llama in Termux on Android (#6039)

wanix1988 committed 1 year ago

Verified 19885d20

readme : update API changes and hot topics

ggerganov committed 1 year ago

Verified 76a936c8

grammar : handle missing "root" node (#6004)

HanClinto committed 1 year ago

Verified 46362837

llama : add pipeline parallelism support (#6017)

slaren committed 1 year ago

Verified f30ea47a

test-backend-ops : skip CPU backend by default (#6028)

slaren committed 1 year ago

Verified d8fd0ccf

Update get version (#6025)

AidanBeltonS committed 1 year ago

Verified b3d97860

Server: Use multi-task for embeddings endpoint (#6001)

ngxson committed 1 year ago

Verified 99b71c06

ci : remove tidy-review (#6021)

slaren committed 1 year ago

Verified 306d34be

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

Newer Older