ggml-org/llama.cpp

Pull Requests Commits

llama : uniform variable names + struct init

ggerganov committed 2 years ago

Verified 26cc1bd7

update readme, update baked includes

tobi committed 2 years ago

Verified 4fe95c69

export llama_timings as struct and expose them in server

tobi committed 2 years ago

Verified 30d973dc

use javascript generators as much cleaner API

tobi committed 2 years ago

Verified a76ce02a

Update Server Instructions (#2113)

jessejohnson committed 2 years ago

Verified 983b555e

ggml : fix bug introduced in #1237

ggerganov committed 2 years ago

Verified ec326d35

tests : fix test-grad0

ggerganov committed 2 years ago

Verified 1b6efeab

ggml : generalize `quantize_fns` for simpler FP16 handling (#1237)

sw committed 2 years ago

Verified 1b107b85

Update server instructions for web front end (#2103)

jessejohnson committed 2 years ago

Verified 8567c76b

Quantized dot products for CUDA mul mat vec (#2067)

JohannesGaessler committed 2 years ago

Verified 924dd22f

llama: Don't double count the sampling time (#2107)

howard0su committed 2 years ago

Verified 051c70dc

Fixed OpenCL offloading prints (#2082)

JohannesGaessler committed 2 years ago

Verified 9e4475f5

embd-input: Fix input embedding example unsigned int seed (#2105)

pnb committed 2 years ago

Verified 7f0e9a77

readme : add link web chat PR

ggerganov committed 2 years ago

Verified b472f3fc

ggml : sync latest (new ops, macros, refactoring) (#2106)

ggerganov committed 2 years ago

Verified ed9a54e5

Add an API example using server.cpp similar to OAI. (#2009)

jwj7140 committed 2 years ago

Verified f257fd25

Simple webchat for server (#1998)

tobi committed 2 years ago

Verified 7ee76e45

Allow old Make to build server. (#2098)

SlyEcho committed 2 years ago

Verified acc111ca

Update Makefile: clean simple (#2097)

ZhouYuChen committed 2 years ago

Verified 23c7c6fc

CI: make the brew update temporarily optional. (#2092)

Green-Sky committed 2 years ago

Verified 698efad5

[ggml] fix index for ne03 value in ggml_cl_mul_f32 (#2088)

gotope committed 2 years ago

Verified 14a2cc71

fix server crashes (#2076)

SlyEcho committed 2 years ago

Verified 1cf14cce

Fix crash of test-tokenizer-0 under Debug build (#2064)

howard0su committed 2 years ago

Verified cc45a7fe

[llama] No need to check file version when loading vocab score (#2079)

howard0su committed 2 years ago

Verified 55dbb915

server: add option to output probabilities for completion (#1962)

WangHaoranRobin committed 2 years ago

Verified d7d2e6a0

ggml : fix build with OpenBLAS (close #2066)

ggerganov committed 2 years ago

46088f72

Better CUDA synchronization logic (#2057)

JohannesGaessler committed 2 years ago

Verified 0bc2cdfc

Test-based VRAM scratch size + context adjustment (#2056)

JohannesGaessler committed 2 years ago

Verified befb3a35

cmake : don't force -mcpu=native on aarch64 (#2063)

dsd committed 2 years ago

Verified b2132270

metal : release buffers when freeing metal context (#2062)

apage43 committed 2 years ago

Verified 2f8cd979

Older