ggml-org/llama.cpp

Pull Requests Commits

Fix gguf-convert-endian script (#4037)

monatis committed 1 year ago

Verified e86fc56f

server : fix crash when prompt exceeds context size (#3996)

z80maniac committed 1 year ago

Verified d96ca7de

gguf-py: Refactor and allow reading/modifying existing GGUF files (#3981)

KerfuffleV2 committed 1 year ago

Verified 34b0a082

server : allow continue edit on completion mode (#3950)

jhen0409 committed 1 year ago

Verified 4a4fd3ee

Unbreak persimmon after #3837 (#4010)

Galunid committed 1 year ago

Verified df9d1293

scripts: Generalize convert scripts (#3838)

Galunid committed 1 year ago

Verified a75fa576

server : add min_p param (#3877)

Mihaiii committed 1 year ago

Verified 57ad015d

ggml-alloc : fix backend assignments of views (#3982)

slaren committed 1 year ago

Verified 875fb428

gguf : track writer state, free unneeded tensors, cleanup (#3871)

cebtenzzre committed 1 year ago

Verified 0a7c980b

make : do not add linker flags when compiling static llava lib (#3977)

ggerganov committed 1 year ago

Verified 413503d4

ggml : fix backward rope after YaRN (#3974)

xaedes committed 1 year ago

Verified e9c1cecb

Use params when loading models in llava-cli (#3976)

tejom committed 1 year ago

Verified 54b4df88

cuda : supports running on CPU for GGML_USE_CUBLAS=ON build (#3946)

wsxiaoys committed 1 year ago

Verified 46876d2a

llava : expose as a shared library for downstream projects (#3613)

damian0815 committed 1 year ago

Verified 381efbf4

ggml-cuda : fix f16 mul mat (#3961)

slaren committed 1 year ago

Verified 2833a6f6

Allow common process_escapes to handle \x sequences (#3928)

KerfuffleV2 committed 1 year ago

Verified d9ccce2e

server : fix typo for --alias shortcut from -m to -a (#3958)

RoyalHeart committed 1 year ago

Verified bb60fd0b

cuda : fix disabling device with --tensor-split 1,0 (#3951)

cebtenzzre committed 1 year ago

Verified 132d25b8

llama : mark LLM_ARCH_STARCODER as full offload supported (#3945)

wsxiaoys committed 1 year ago

Verified 3d48f42e

cmake : MSVC instruction detection (fixed up #809) (#3923)

netrunnereve committed 1 year ago

Verified c41ea36e

ci : use intel sde when ci cpu doesn't support avx512 (#3949)

netrunnereve committed 1 year ago

Verified a7fac013

cuda : revert CUDA pool stuff (#3944)

slaren committed 1 year ago

Verified 48ade945

gguf-py: Support 01.AI Yi models (#3943)

KerfuffleV2 committed 1 year ago

Verified f28af0d8

metal : round up to 16 to fix MTLDebugComputeCommandEncoder assertion (#3938)

psugihara committed 1 year ago

Verified d9b33fe9

ggml-metal: fix yarn rope (#3937)

jxy committed 1 year ago

Verified 5ba37461

ggml-cuda : move row numbers to x grid dim in mmv kernels (#3921)

slaren committed 1 year ago

Verified abb77e73

speculative : change default p_accept to 0.5 + CLI args (#3919)

ggerganov committed 1 year ago

Verified 8f961abd

common : YAYF (yet another YARN fix) (#3925)

ggerganov committed 1 year ago

Verified 05816027

llama : change yarn_ext_factor placeholder to -1 (#3922)

cebtenzzre committed 1 year ago

Verified 3fdbe6b6

cuda : add ROCM aliases for CUDA pool stuff (#3918)

KerfuffleV2 committed 1 year ago

Verified 629f917c

Newer Older