ggml-org/llama.cpp

Pull Requests Commits

fix part of mul_mat_id

airMeng committed 1 year ago

ff0aa3ab

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

chore: clean useless beam search param (#7985)

thxCode committed 1 year ago

Verified b96f9afb

readme : update UI list (#7943)

abgulati committed 1 year ago

Verified 11937781

ggerganov committed 1 year ago

Verified 5326bcce

whisper : use ggml_backend_sched (whisper/2239)

ggerganov committed 1 year ago

Verified e6ecc2be

update: support Qwen2-57B-A14B (#7835)

legraphista committed 1 year ago

Verified a94e6ff8

Make updates to type cast based on compiler instead of OS (#7851)

Srihari-mcw committed 1 year ago

Verified 5b6da187

llama : disable FA if KV head size do not match (#7982)

ggerganov committed 1 year ago

Verified 7c26775a

Add Nix and Flox install instructions (#7899)

bryanhonof committed 1 year ago

Verified b473e950

sched : offload_op also requires supports_op (#7977)

slaren committed 1 year ago

Verified 99052cd2

fix: divide 0 exception in mamba (#7932)

thxCode committed 1 year ago

Verified c637fcd3

Implement non-mapped async IO for CUDA on Windows. (#7896)

mtavenrath committed 1 year ago

Verified 6a2f0b34

rpc : fix load/store misaligned addresses (#7948)

ggerganov committed 1 year ago

Verified 21be9cab

gguf-dump.py: add --markdown dump output (#7853)

mofosyne committed 1 year ago

Verified 006167aa

[SYCL] Update README-sycl.md for Chapter "Recommended release" and "News" (#7946)

arthw committed 1 year ago

Verified df68d4fa

Older