ggerganov/llama.cpp

Pull Requests Commits

llava : change API to pure C style for Rust FFI bindgen (#6079)

tinglou committed 1 year ago

Verified 4e9a7f7f

cuda : disable unused cudaLaunchHostFunc code (#6078)

slaren committed 1 year ago

Verified 3020327f

fix set main gpu error (#6073)

NeoZhangJianyu committed 1 year ago

Verified 46acb367

make : ggml-metal.o depends on ggml.h

ggerganov committed 1 year ago

Verified 131b0584

[SYCL] Fix non-intel device selection (#6042)

AidanBeltonS committed 1 year ago

Verified 753e36f6

gguf : add support for I64 and F64 arrays (#6062)

certik committed 1 year ago

Verified 7ce2c77f

llama : add Orion chat template (#6066)

ngxson committed 1 year ago

Verified aab606a1

llama-bench : use random tokens to improve accuracy with mixtral (#6069)

slaren committed 1 year ago

Verified b0bc9f4a

llama : fix integer overflow during quantization (#6063)

ggerganov committed 1 year ago

Verified 4755afd1

gguf : fix resource leaks (#6061)

stevegrubb committed 1 year ago

Verified 6e0438da

gguf-py : bump version to 0.8.0 (#6060)

certik committed 1 year ago

Verified 72710770

llama : support models without vocabulary (#5798)

Xarbirus committed 1 year ago

Verified 69ff6139

embedding : add EOS token if not present (#899)

ggerganov committed 1 year ago

Verified 044ec4b2

gguf-py : fix dtype check (#6045)

ggerganov committed 1 year ago

Verified 77178eed

readme : improve readme for Llava-1.6 example (#6044)

jianliao committed 1 year ago

Verified 15a33326

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

Newer Older