Skip to content

LLM inference in C/C++

C++ ggml

Commits on master

opencl: initial q1_0 support (#25160)
lhez committed 4 months ago
fd1a057
model : register t_layer_inp for qwen3next (#25141)
Jürgen Schmied committed 4 months ago
4f31eed
CUDA: fix Gemma E4B MTP FlashAttention (#25148)
Johannes Gäßler committed 4 months ago
e495d1e
f708a5b
ggml-webgpu: add support for NVFP4 (#25143)
Masashi Yoshimura committed 4 months ago
6c5de1c
6f4f53f
DeepSeek V4 (#24162)
Aman Gupta committed 4 months ago
8c146a8
6cb18b2
277a105
jinja, chat: add --reasoning-preserve flag (#25105)
Xuan-Son Nguyen committed 4 months ago
b3fed31
dflash: refactor draft model conversion (#25110)
Ruixiang Wang committed 4 months ago
fa72bc6
chat : implement minicpm5 parser (#24889)
Aldehir Rojas committed 4 months ago
c818263
jinja: add --dump-prog for debugging (#25086)
Xuan-Son Nguyen committed 4 months ago
f68a788
spec : add DFlash support (#22105)
Ruixiang Wang committed 4 months ago
d1b3425
common : allow --offline in llama download (#25091)
Adrien Gallouët committed 4 months ago
c1a1c8e
logs : reduce v2 (#25078)
Georgi Gerganov committed 4 months ago
27c8bb4
opencl: flash attention improvement (#25069)
Hongqiang Wang committed 4 months ago
ebd048f
sycl : fix failed ut cases of norm (#25044)
Neo Zhang committed 4 months ago
9bebfcb