ggml-org/llama.cpp

Pull Requests Commits

ggml : testing GPU FP precision via quantized CPY

ggerganov committed 1 year ago

Verified f64e4f04

ggml : add ggml_cpu_has_avx_vnni() (#4589)

tikikun committed 1 year ago

Verified 24a447e2

CUDA: fix tensor core logic for Pascal and HIP (#4682)

JohannesGaessler committed 1 year ago

Verified a20f3c74

clip : use ggml_backend_buffer_is_host (#4205)

ggerganov committed 1 year ago

Verified 0235b9b5

clip : enable gpu backend (#4205)

FSSRepo committed 1 year ago

Verified ce18d727

cuda: fix vmm oom issue on NVIDIA AGX Orin (#4687)

hydai committed 1 year ago

Verified 91bb39ce

python : add check-requirements.sh and GitHub workflow (#4585)

crasm committed 1 year ago

Verified 04ac0607

flake.nix : rewrite (#4605)

philiptaron committed 1 year ago

Verified 68eccbdc

cmake : fix ld warning duplicate libraries libllama.a (#4671)

nguoithichkhampha committed 1 year ago

Verified 97bbca6e

llava-cli : refactor to use sampling library (#4669)

jart committed 1 year ago

Verified 4af48015

server : replace sleep with condition variables (#4673)

jart committed 1 year ago

Verified db49ff8e

server : fix OpenAI server sampling w.r.t. penalty. (#4675)

sakura-umi committed 1 year ago

Verified 60f55e88

server : allow to generate multimodal embeddings (#4681)

kseth committed 1 year ago

Verified b93edd22

main-cmake-pkg : fix build issue (#4665)

andrijdavid committed 1 year ago

Verified 82d6eab2

llama.swiftui : fix infinite loop, ouput timings, buff UI (#4674)

psugihara committed 1 year ago

Verified afd997ab

scripts : print list of sync commits

ggerganov committed 1 year ago

Verified c8255f8a

ci : build with CLBlast + ggml-opencl use GGML_API (whisper/1576)

tamo committed 1 year ago

Verified 441f51dc

ggerganov committed 1 year ago

Verified 38b3de46

ggml : fix some mul mat cases + add tests for src1 F16 (ggml/669)

bssrdf committed 1 year ago

Verified afc8c192

scripts : do not sync commits from this repo

ggerganov committed 1 year ago

Verified ca38b8d3

Fix OpenAI server sampling w.r.t. temp and seed (#4668)

jart committed 1 year ago

Verified 65e5f6da

gpt2 : Add gpt2 architecture integration (#4555)

manikbhandari committed 1 year ago

Verified ea5497df

llama : add AWQ for llama, llama2, mpt, and mistral models (#4593)

namtranase committed 1 year ago

Verified f6793491

finetune : fix output formatting in print_params (#4653)

danbev committed 1 year ago

Verified 879b690a

scripts : add sync-ggml-am.sh

ggerganov committed 1 year ago

Verified b47879b0

ggml : fix dot product for ARM (#4630)

ggerganov committed 1 year ago

Verified 951010fa

Add byte token type when tokenizer.model is not exists (#4641)

strutive07 committed 1 year ago

Verified f56d6077

cuda : fix vmm pool with multi GPU (#4620)

slaren committed 1 year ago

Verified dc68f005

Update comment for AdamW implementation reference. (#4604)

WillCorticesAI committed 1 year ago

Verified de8e4964

Fix new CUDA10 compilation errors (#4635)

FantasyGmm committed 1 year ago

Verified 77465dad

Older