ggerganov/llama.cpp

Pull Requests Commits

server : fix JSON-Scheme typo (#7975)

akx committed 1 year ago

Verified 6a2f298b

Fix typo in llama_set_embeddings comment (#8077)

danbev committed 1 year ago

Verified 11318d9a

fix CI failures (#8066)

slaren committed 1 year ago

Verified b6b9a8e6

Refactor Vulkan backend to allow multiple contexts (#7961)

0cc4m committed 1 year ago

Verified 45c0e2e4

Removing extra blank lines that were breaking Lint. (#8067)

HanClinto committed 1 year ago

Verified b5a5f34e

cvector: fix CI + correct help message (#8064)

ngxson committed 1 year ago

Verified 3e58b0ee

cvector-generator: Moe Moe Fixie-Fixie for Lots of Formats~! ♡(ᐢ ᴥ ᐢ)♡ (#8052)

HatsuneMikuUwU33 committed 1 year ago

Verified adf480c3

convert-hf : change assert to exception (#8015)

0xspringtime committed 1 year ago

Verified 3aa184a8

Update llama-quantize ppl/file size output from LLaMA-v1 to Llama-3 values (#8058)

ddh0 committed 1 year ago

Verified 5b48cd53

JSON Schema to GBNF integration tests (#7790)

HanClinto committed 1 year ago

Verified c5a8d4b7

vulkan: detect multiple devices by deviceUUID instead of deviceID (#8022)

Adriankhl committed 1 year ago

Verified 557b653d

ggml : AVX IQ quants (#7845)

netrunnereve committed 1 year ago

Verified 7d5e8777

llama : optimize long word tokenization with WPM (#8034)

ggerganov committed 1 year ago

Verified a927b0f3

llama : allow pooled embeddings on any model (#7477)

iamlemec committed 1 year ago

Verified 80ea089d

swiftui : enable stream updating (#7754)

shu223 committed 1 year ago

Verified 0e64591e

requirements : Bump torch and numpy for python3.12 (#8041)

hamdoudhakem committed 1 year ago

Verified b1ef562b

convert-hf : Fix the encoding in the convert-hf-to-gguf-update.py (#8040)

hamdoudhakem committed 1 year ago

Verified 17b291a6

common: fix warning (#8036)

JohannesGaessler committed 1 year ago

Verified abd894ad

[SYCL] Fix windows build and inference (#8003)

luoyu-intel committed 1 year ago

Verified de391e4c

CUDA: stream-k decomposition for MMQ (#8018)

JohannesGaessler committed 1 year ago

Verified d50f8897

metal : fix `ggml_metal_supports_op` for BF16 (#8021)

mdegans committed 1 year ago

Verified 2075a66a

server : fix smart slot selection (#8020)

sasha0552 committed 1 year ago

Verified ba589931

un-ignore `build-info.cmake` and `build-info.sh` (#7996)

mdegans committed 1 year ago

Verified a7854743

ggml : synchronize threads using barriers (#7993)

slaren committed 1 year ago

Verified 9c77ec1d

codecov : remove (#8004)

ggerganov committed 1 year ago

Verified a04a953c

[SYCL] refactor (#6408)

airMeng committed 1 year ago

Verified 623494a4

tokenizer : BPE fixes (#7530)

jaime-m-p committed 1 year ago

Verified 37bef894

Only use FIM middle token if it exists (#7648)

CISC committed 1 year ago

Verified 91c188d6

Fix no gcc pragma on Windows (#7751)

jojorne committed 1 year ago

Verified 84f6de17

Allow compiling with CUDA without CUDA runtime installed (#7989)

drepper committed 1 year ago

Verified 61665277

Newer Older