ggml-org/llama.cpp

Pull Requests Commits

CLBlast: Fix handling of on-device tensor data

shibe2 committed 1 year ago

e2583cbc

server : fix incorrect num_tokens_predicted (#3480)

jhen0409 committed 1 year ago

Verified e8b8d32e

swift : disable ACCELERATE_NEW_LAPACK (#3481)

jhen0409 committed 1 year ago

Verified 8f3a642e

ci : add swift build via xcodebuild (#3482)

jhen0409 committed 1 year ago

Verified 07453844

convert : fix Baichuan2 models by using vocab size in config.json (#3299)

KerfuffleV2 committed 1 year ago

Verified 019ba1dc

readme : add project status link

ggerganov committed 1 year ago

Verified beabc8cf

ggml : fix build after #3329

ggerganov committed 1 year ago

Verified 0d152b37

llm : add Refact model (#3329)

ds5t5 committed 1 year ago

Verified f8c90cdb

sync : ggml (conv 1d + 2d updates, UB fixes) (#3468)

ggerganov committed 1 year ago

Verified f93af024

finetune : readme fix typo (#3465)

iammerrick committed 1 year ago

Verified f72f8f22

ggml : add RISC-V Vector Support for K-Quants and improved the existing intrinsics (#3453)

Tameem-10xE committed 1 year ago

Verified 79f34abd

main : consistent prefix/suffix coloring (#3425)

h-h-h-h committed 1 year ago

Verified 8186242b

llama : fix session saving/loading (#3400)

ggerganov committed 1 year ago

Verified ac2219fe

llama : expose model's rope_freq_scale in the API (#3418)

grencez committed 1 year ago

Verified 48be797f

metal : alibi for arbitrary number of heads (#3426)

li-plus committed 1 year ago

Verified f56e1bae

cmake : make LLAMA_NATIVE flag actually use the instructions supported by the processor (#3273)

netrunnereve committed 1 year ago

Verified 017efe89

Work on the BPE tokenizer (#3252)

goerch committed 1 year ago

Verified ff5a3f0c

convert : fix vocab size when not defined in hparams (#3421)

cebtenzzre committed 1 year ago

Verified 1c84003c

cmake : increase minimum version for add_link_options (#3444)

cebtenzzre committed 1 year ago

Verified e78f0b0d

CLBlast: Add broadcast support for matrix multiplication (#3402)

shibe2 committed 1 year ago

Verified 665018c7

gguf : add BERT, MPT, and GPT-J arch info (#3408)

cebtenzzre committed 1 year ago

Verified 29a404a9

gguf : general usability improvements (#3409)

cebtenzzre committed 1 year ago

Verified 0fe32103

cmake : make CUDA flags more similar to the Makefile (#3420)

cebtenzzre committed 1 year ago

Verified 9476b012

finetune : fix #3404 (#3437)

xaedes committed 1 year ago

Verified a03ce384

metal : set log callback before initializing (#3427)

phronmophobic committed 1 year ago

Verified a8476769

cmake : fix transient definitions in find pkg (#3411)

bandoti committed 1 year ago

Verified 095231df

docker : ignore Git files (#3314)

kevinji committed 1 year ago

Verified ea55295a

infill : add new example + extend server API (#3296)

vvhg1 committed 1 year ago

Verified c97f01c3

ggml-cuda : perform cublas mat mul of quantized types as f16 (#3412)

slaren committed 1 year ago

Verified f5ef5cfb

llama.cpp : add documentation about rope_freq_base and scale values (#3401)

slaren committed 1 year ago

Verified 40e07a60

Newer Older