ggerganov/llama.cpp

Pull Requests Commits

llama : revert n_threads_batch logic

ggerganov committed 1 year ago

Verified 87f4102a

llama : use n_threads_batch only when n_tokens >= 32

ggerganov committed 1 year ago

Verified e9b7a5cb

ggml : use blas even if src0 is not F32

ggerganov committed 1 year ago

Verified f815fe43

readme : add Amica to UI list (#4230)

kasumi-1 committed 1 year ago

Verified 0dab8cd7

examples : iOS example with swift ui (#4159)

bachittle committed 1 year ago

Verified bb03290c

ggml : fix -Warray-bounds warning with gcc (#4231)

cebtenzzre committed 1 year ago

Verified f3b26981

lookahead : support `-n -1` infinite generation

ggerganov committed 1 year ago

Verified 3e73d31d

readme : update hot topics

ggerganov committed 1 year ago

Verified 9656026b

lookahead : add example for lookahead decoding (#4207)

ggerganov committed 1 year ago

Verified 922754a8

metal : fix yarn (#4220)

jxy committed 1 year ago

Verified 22da0553

scripts : Use mmap in torch load (#4202)

Galunid committed 1 year ago

Verified 1ddb52ec

llama : grammar `reserve` space in `decode_utf8` (#4210)

MarcusDunn committed 1 year ago

Verified f837c3a9

Update docs for yarn_ext_factor <0.0 as unspecified instead of NaN (#4189)

crasm committed 1 year ago

Verified 3014b541

readme : update hot topics

ggerganov committed 1 year ago

Verified 04814e71

server : OAI API compatibility (#4198)

ggerganov committed 1 year ago

Verified af19d357

llama : set metal log callback correctly (#4204)

slaren committed 1 year ago

Verified e9c13ff7

ggml-cuda : support stablelm rope (#4156)

slaren committed 1 year ago

Verified 8a052c13

convert : fix tensors using grad in some models (#4173)

Galunid committed 1 year ago

Verified 189d6844

main.swift : fix eos checking (#4197)

eastriverlee committed 1 year ago

Verified 2568a4bf

readme : use PATH for Windows ROCm (#4195)

jammm committed 1 year ago

Verified b35f3d0d

Fix incorrect format strings and uninitialized variables. (#4133)

haohui committed 1 year ago

Verified 55978ce0

llama : KV cache view API + better KV cache management (#4170)

ggerganov committed 1 year ago

Verified 6b0a7420

readme : update hot topics

ggerganov committed 1 year ago

Verified d103d935

examples : fix typo in parallel example doc comment (#4181)

danbev committed 1 year ago

Verified 9d5949f0

docs : add llama-star arch idea

ggerganov committed 1 year ago

Verified ff8238f7

stablelm : simplify + speedup generation (#4153)

Galunid committed 1 year ago

Verified 8e672efe

finetune - update readme to mention llama support only (#4148)

Galunid committed 1 year ago

Verified 0b871f1a

readme : update ROCm Windows instructions (#4122)

jammm committed 1 year ago

Verified dfc7cd48

main : Add ChatML functionality to main example (#4046)

Sebby37 committed 1 year ago

Verified 881800d1

ci : add flake8 to github actions (python linting) (#4129)

Galunid committed 1 year ago

Verified f23c0359

Older