ggerganov/llama.cpp

Pull Requests Commits

sampling : one sequence per sampling context

ggerganov committed 1 year ago

Verified 5261aee8

examples: support LLaVA v1.5 (multimodal model) (#3436)

monatis committed 1 year ago

Verified 370359e5

docs : fix typo GOMP_CPU_AFFINITY (#3597)

maekawatoshiki committed 1 year ago

Verified 9e24cc6e

cmake : fix add_compile_options on macOS

ggerganov committed 1 year ago

Verified d28e572c

typo : it is `--n-gpu-layers` not `--gpu-layers` (#3592)

ianscrivener committed 1 year ago

Verified f3040bea

ci : check if there is enough VRAM (#3596)

ggerganov committed 1 year ago

Verified 1a8c8795

server : add completion mode (no chat) (#3582)

akx committed 1 year ago

Verified b016596d

prompts : add mnemonics.txt

ggerganov committed 1 year ago

Verified 6b3ae4da

server : fix kv cache management (#3588)

ggerganov committed 1 year ago

Verified 57dd55e2

main : fix session loading bug (#3400)

ggerganov committed 1 year ago

Verified b8fe4b5c

server : add parameter -tb N, --threads-batch N (#3584)

m18coppola committed 1 year ago

Verified a8bdd655

common : fix mirostat state when using multiple sequences (#3543)

KerfuffleV2 committed 1 year ago

Verified 70c29da1

batched : add bench tool (#3545)

ggerganov committed 1 year ago

Verified 8c70a5ff

examples : add batched.swift + improve CI for swift (#3562)

zshannon committed 1 year ago

Verified 24ba3d82

Add MPT model to supported models in README.md (#3574)

Galunid committed 1 year ago

Verified 9f6ede19

Minor improvements in GPT2 tokenizer (#3567)

goerch committed 1 year ago

Verified 233fc1c6

readme : add bloom (#3570)

xingchensong committed 1 year ago

Verified c5b49360

llm : add bloom models (#3553)

xingchensong committed 1 year ago

Verified 02d2875d

swift : improvements and fixes (#3564)

jhen0409 committed 1 year ago

Verified 0aa6595a

llm : add MPT support (#3417)

jploski committed 1 year ago

Verified f5f9121d

infill. : fix tokenization (#3508)

vvhg1 committed 1 year ago

Verified 11ea5c7d

ggml-alloc : fix assert in debug builds (#3555)

slaren committed 1 year ago

Verified 95bd60a0

refact : fix convert script + zero out KV cache to avoid nans (#3523)

ggerganov committed 1 year ago

Verified fcca0a70

metal : do not use mul_mm kernels when ne00 < 64 (#3542)

ggerganov committed 1 year ago

Verified dcc09d25

sync : ggml (ggml-backend) (#3548)

ggerganov committed 1 year ago

Verified db3abcc1

ci : add Zig CI/CD and fix build (#2996)

kassane committed 1 year ago

Verified eee42c67

api_like_OAI.py : compat with Microsoft Guidance (#2746)

ryderwishart committed 1 year ago

Verified 8e6716a1

api_like_OAI.py : simplify function (#2796)

arcrank committed 1 year ago

Verified 9c38d181

k-quants : fix comments about block sizing (#3499)

jrudolph committed 1 year ago

Verified a1202a31

ci : enable on obj-c changes + fix metal build (#3540)

ggerganov committed 1 year ago

Verified 94e502df

Older