ggerganov/llama.cpp

Pull Requests Commits

ggml : optimize non-SIMD Q4_0 vector dot product (#703)

sw committed 2 years ago

Verified 6232f2d7

ggml : introduce GGML_ALIGNED_MALLOC/GGML_ALIGNED_FREE macros (#884)

prusnak committed 2 years ago

Verified 6c248707

fix whitespace (#944)

CRD716 committed 2 years ago

Verified 8cda5c98

readme : remove python 3.10 warning (#929)

CRD716 committed 2 years ago

Verified ec292721

readme : llama node binding (#911)

hlhr202 committed 2 years ago

Verified 7e941b95

flake.nix: add all binaries from bin (#848)

prusnak committed 2 years ago

Verified c729ff73

zig : update build.zig (#872)

foldl committed 2 years ago

Verified 4579af95

ggml : update cblas_sgemm columns var to be more reasonable (#838)

bogdad committed 2 years ago

Verified 8c3ffc2f

examples : add -n to alpaca and gpt4all scripts (#706)

niansa committed 2 years ago

Verified 107980d9

cmake : add explicit F16C option (x86) (#576)

anzz1 committed 2 years ago

Verified 585d91a1

benchmark : add tool for timing q4_0 matrix multiplication (#653)

SebastianApel committed 2 years ago

Verified 95ea26f6

do not force the prompt file to end with a new line (#908)

prusnak committed 2 years ago

Verified 82d146df

Don't crash on ftype (formerly f16) == 4 (#917)

sw committed 2 years ago

Verified e7f6997f

readme : change "GPU support" link to discussion

ggerganov committed 2 years ago

Verified f76cb3a3

readme : update hot topics with link to "GPU support" issue

ggerganov committed 2 years ago

Verified 78243807

readme: link to sha256sums file (#902)

NicoWeio committed 2 years ago

Verified 4dbbd407

Fix whitespace, add .editorconfig, add GitHub workflow (#883)

prusnak committed 2 years ago

Verified 8b679987

Add enum llama_ftype, sync ggml_type to model files (#709)

sw committed 2 years ago

Verified 3e6e70d8

Windows fixes (#890)

comex committed 2 years ago

Verified 2663d2c6

Add BAIR's Koala to supported models (#877)

qouoq committed 2 years ago

Verified a0caa34b

ggml : fix WASM build

ggerganov committed 2 years ago

Verified 461ba9e6

ggml : add ggml_cont() + optimize ggml_cpy() for contiguous dst

ggerganov committed 2 years ago

Verified c3ac702e

ggml : remove trailing whitespaces

ggerganov committed 2 years ago

Verified 9d634ef4

Simplify to include lower-case windows.h always, fix compile on mingw32 (#747)

marcom committed 2 years ago

Verified d9a239c4

ggml : fix quantize_row_q4_1() ARM_NEON (close #876)

ggerganov committed 2 years ago

Verified 684da259

Print model version.

comex committed 2 years ago

180b693a

Rewrite loading code to try to satisfy everyone:

comex committed 2 years ago

f963b63a

fix for windows utf-8 input (#840)

aroidzap committed 2 years ago

Verified aaf3b23d

cmake should link openblas properly with -lopenblas like how it's done in the makefile (#839)

eiery committed 2 years ago

Verified f2d1c472

Add new binaries to flake.nix (#847)

longregen committed 2 years ago

Verified 317fb12f

Newer Older