ggml-org/llama.cpp

Pull Requests Commits

ggml : try another fix

ggerganov committed 2 years ago

Verified 974e3cad

ggml : no cpu_set_t on Android

ggerganov committed 2 years ago

Verified e9caab61

llava : removed excess free(NULL) operation (#5531)

GermanAizek committed 2 years ago

Verified 4cb07276

llama : minor fixed return int value (#5529)

GermanAizek committed 2 years ago

Verified 65085c71

server : add "samplers" param to control the samplers order (#5494)

z80maniac committed 2 years ago

Verified 6dcc02d2

server : fix system prompt cli (#5516)

An0nie committed 2 years ago

Verified 5f5808ca

ggml : add numa options (#5377)

bmtwl committed 2 years ago

Verified f486f6e1

llava : fix clip-model-is-vision flag in README.md (#5509)

danbev committed 2 years ago

Verified 60ed04cf

ci : fix BERT model download and convert

ggerganov committed 2 years ago

Verified 594845aa

Use correct type of pooling for embedding models (#5500)

iamlemec committed 2 years ago

Verified 4524290e

clip : fix wrong loop condition

ggerganov committed 2 years ago

Verified c06e45d7

cuda : print message when initialization fails (#5512)

slaren committed 2 years ago

Verified 9060a1e9

scripts : add hf.sh helper script (#5501)

ggerganov committed 2 years ago

Verified 9350a1cf

fix(gguf-py): special tokens are no longer skipped when add_<token>_token is set to false (#5487)

vriesdemichael committed 2 years ago

Verified 73122473

llava : fix memory management bug (#5491)

Elbios committed 2 years ago

Verified 0d417712

llaba : hotfix for llava-1.6 image number (#5495)

cmp-nct committed 2 years ago

Verified 7930a8a6

vulkan: Find optimal memory type but with fallback (#5381)

luciferous committed 2 years ago

Verified 704359e2

readme : fix typo (#5490)

Rune-AI committed 2 years ago

Verified 594fca3f

llava : update README.md (#5489)

cmp-nct committed 2 years ago

Verified ccbb277f

cmake : ARM intrinsics detection for MSVC (#5401)

Xarbirus committed 2 years ago

Verified 8084d554

llava : support v1.6 (#5267)

cmp-nct committed 2 years ago

Verified aa234129

Early return for zero size calls to get_tensor. (#5482)

manyoso committed 2 years ago

Verified f5ca0548

gguf : add python reader example (#5216)

cmp-nct committed 2 years ago

Verified 6c00a066

llama : add support for Nomic Embed (#5468)

cebtenzzre committed 2 years ago

Verified ea9c8e11

llama : allow raw byte in SPM vocabs; don't crash on nl 404 (#5478)

akx committed 2 years ago

Verified c4e6dd59

llama : make load error reporting more granular (#5477)

akx committed 2 years ago

Verified 037259be

finetune : rename feed-forward tensors (w1/w2/w3) (#4839)

danbev committed 2 years ago

Verified 26397890

tests : multi-thread the tokenizer tests (#5474)

ggerganov committed 2 years ago

Verified cf45252a

llama : support batched embeddings (#5466)

iamlemec committed 2 years ago

Verified 03bf161e

make: add error message for bad CUDA version (#5444)

JohannesGaessler committed 2 years ago

Verified ad014bba

Older