ggerganov/llama.cpp

Pull Requests Commits

py : try to fix flake stuff

ggerganov committed 1 year ago

Verified 1fb563eb

convert : update phi-2 to latest HF repo

ggerganov committed 1 year ago

Verified fe252237

ggerganov committed 1 year ago

Verified de473f5f

ggml : fix 32-bit ARM compat for IQ2_XS (whisper/1758)

ggerganov committed 1 year ago

Verified f2384612

backend_sched : fix assignments

slaren committed 1 year ago

Verified fa5c1fb4

examples : add pydantic models to GBNF grammar generator (#4883)

Maximilian-Winter committed 1 year ago

Verified 52ee4540

CUDA: faster q8_0 -> f16 dequantization (#4895)

JohannesGaessler committed 1 year ago

Verified 3fe81781

llama : ggml-backend integration (#4766)

slaren committed 1 year ago

Verified e7e4df03

llama : remove redundant assert for StableLM (#4901)

ggerganov committed 1 year ago

Verified 584d674b

export-lora : use LLAMA_FILE_MAGIC_GGLA (#4894)

danbev committed 1 year ago

Verified 930f907d

llama.swiftui : update models layout (#4826)

isaiahbjork committed 1 year ago

Verified e790eef2

gitignore : imatrix

ggerganov committed 1 year ago

Verified 5537d9d3

CUDA: fix softmax compile for old CUDA versions (#4862)

JohannesGaessler committed 1 year ago

Verified 1b280c9f

llama : fix typo "imp_embd" -> "inp_embd"

ggerganov committed 1 year ago

Verified 3cabe806

common : streamline the formatting of help (#4890)

howlger committed 1 year ago

Verified 4315a943

py : fix lint (#4889)

ggerganov committed 1 year ago

Verified 2d00741e

llama : fix llm_build_k_shift to use correct n_rot (#4889)

ggerganov committed 1 year ago

Verified f445c0e6

Importance Matrix calculation (#4861)

ikawrakow committed 1 year ago

Verified 326b418b

server : fix infill when prompt is empty (#4833)

ggerganov committed 1 year ago

Verified 1d118386

main : better name for variable n_print (#4874)

ggerganov committed 1 year ago

Verified 7edefbd7

main : disable token count by default (#4874)

ggerganov committed 1 year ago

Verified 3ca63b45

swift : track ggml release branch (#4867)

ggerganov committed 1 year ago

Verified b0377875

llama : restore intended k-quants mixes for MoE models (#4872)

ikawrakow committed 1 year ago

Verified 469e75d0

ggml : SOTA 2-bit quants (add IQ2_XS) (#4856)

ikawrakow committed 1 year ago

Verified 49662cbe

swift : pin ggml commit + remove ggml.h from spm-headers (#4878)

ggerganov committed 1 year ago

Verified 3ba5b8ca

server : implement credentialed CORS (#4514)

Azeirah committed 1 year ago

Verified 4330bd83

server : support for multiple api keys (#4864)

m18coppola committed 1 year ago

Verified 27379455

server : add `LOG_INFO` when model is successfully loaded (#4881)

ibehnam committed 1 year ago

Verified eab67950

ci: nix-flake-update: new token with pr permissions (#4879)

SomeoneSerge committed 1 year ago

Verified d8d90aa3

main : print total token count and tokens consumed so far (#4874)

pudepiedj committed 1 year ago

Verified 43f76bf1

Older