ngxson/llama.cpp

Pull Requests Commits

ngxson committed 2 years ago

3bea2e6a

ngxson committed 2 years ago

468551e7

update server docs

ngxson committed 2 years ago

7fee2037

ngxson committed 2 years ago

a533aac1

protect /props endpoint

ngxson committed 2 years ago

7560ecb3

server : more explicit endpoint access settings

ngxson committed 2 years ago

460ec6f4

metal : single allocation of encode_async block (#9747)

ptsochantaris committed 2 years ago

Verified 96b69121

contrib : simplify + minor edits [no ci]

ggerganov committed 2 years ago

Verified d5cb8684

readme : fix typo [no ci]

ggerganov committed 2 years ago

Verified f4b2dcdf

sync : llama.cpp

ggerganov committed 2 years ago

Verified b6d6c528

vulkan : retry allocation with fallback flags (whisper/2451)

SRHMorris committed 2 years ago

Verified b0915d5b

rerank : use [SEP] token instead of [BOS] (#9737)

ggerganov committed 2 years ago

Verified 8c475b97

ggerganov committed 2 years ago

Verified 58b16695

metal : zero-init buffer contexts (whisper/0)

ggerganov committed 2 years ago

Verified 905f5485

Add Llama Assistant (#9744)

vietanhdev committed 2 years ago

Verified 71967c2a

ggerganov committed 2 years ago

Verified 17880771

ggml : fix typo in example usage ggml_gallocr_new (ggml/984)

danbev committed 2 years ago

Verified 55951c01

ggml : fixes after sync (ggml/983)

slaren committed 2 years ago

Verified ff565769

ci : fine-grant permission (#9710)

ngxson committed 2 years ago

Verified f3fdcfaa

Fixed RNG seed docs (#9723)

d-kleine committed 2 years ago

Verified 133c7b46

metal : remove abort (skip) (ggml/0)

ggerganov committed 2 years ago

Verified d5ed2b92

ggerganov committed 2 years ago

Verified 1bb8a64e

ggml/ex: calculate accuracy in graph, adapt MNIST (ggml/980)

JohannesGaessler committed 2 years ago

Verified fabdc3bd

ggml: refactor cross entropy loss CPU impl. (ggml/976)

JohannesGaessler committed 2 years ago

Verified eee39bdc

metal : fix compute pass descriptor autorelease crash (#9718)

jmousseau committed 2 years ago

Verified 5d5ab1e5

ggml-backend : add device description to CPU backend (#9720)

slaren committed 2 years ago

Verified a7ad5535

ggml: unify backend logging mechanism (#9709)

mtmcp committed 2 years ago

Verified d6fe7abf

convert : handle tokenizer merges format from transformers 4.45 (#9696)

compilade committed 2 years ago

Verified e3c355ba

rpc : enable vulkan (#9714)

rgerganov committed 2 years ago

Verified 841713e1

Fixed dequant precision issues in Q4_1 and Q5_1 (#9711)

OuadiElfarouki committed 2 years ago

Verified 56399714

Older