Skip to content

LLM inference in C/C++

C++ ggml

Commits on master

minor: fix reasoning preserve var for DS4 [no ci] (#25999)
Piotr Wilkin (ilintar) committed 4 months ago
e8e6c7a
Fix DeepSeek4 crafted template (#25414)
Piotr Wilkin (ilintar) committed 4 months ago
f534da2
ggml: enable PowerPC backend variants on AIX (#25983)
shalinib-ibm committed 4 months ago
3ce7da2
b4d6c7d
cuda: GET_ROWS quants (#25962)
Pascal committed 4 months ago
c5a4a0b
llama-arch: fix DeepSeek4 APE tensor op (#25945)
helanfxz committed 4 months ago
67b9b0e
Add support for Laguna XS.2 & M.1 (#25165)
Joe Rowell committed 4 months ago
1f66c3c
66e4bf7
846e991
server : properly handle null llama_context (#25868)
fairydreaming committed 4 months ago
40b740a
CUDA: vectorize same-type get_rows with int4 copy (#25929)
Piotr Wilkin (ilintar) committed 4 months ago
305ba51
hexagon: add CLAMP op (#25934)
Todor Boinovski committed 4 months ago
76f46ad
llama_dsv4: write only used rows in state (#25325)
Aman Gupta committed 4 months ago
91d2fc3
model: rotate injected K/V cache for DFlash (#25823)
Ruixiang Wang committed 4 months ago
571d0d5