COMMITS
August 12, 2026
S
[Recipe] Add IndexTTS-2.5 (#772)
Sy03 committed
P
Change DSv4 backend on DGX Spark to b12x (#786)
prodney-dev committed
S
[ROCm] Fix AITER MoE SITU v2 A8W4 env var name for Kimi-K3 (#788)
seungrokj committed
August 11, 2026
P
DGX Spark DeepSeek V4 flash recipe (#784)
prodney-dev committed
Y
Remove source oneapi for xpu (#783)
Yejing Lai committed
T
[ROCm] Fix MiniMax-M3 EAGLE draft backend recipe (#740)
Tuukka Sarvi committed
August 10, 2026
T
D
Add FP8_BLOCK variant for Muse-Glimmer-30B (#777)
Dipika Sikka committed
T
Muse glimmer (#774)
Tiezhen WANG committed
Z
Add inclusionAI/Ling-3.0-flash-FP8 recipe (#756)
zexplorerhj committed
V
Update GLM-5.2 mi355x mxfp4 recipes (#770)
vllmellm committed
August 9, 2026
Y
add MiniMax-H3 cookbook for dgx spark gb10 (#769)
yiminghub2024 committed
August 8, 2026
S
docs(minimax-h3): add six request mode presets (#765)
SYLAR committed
S
docs(minimax-h3): add RTX PRO 5000 and modular service (#742)
SYLAR committed
August 7, 2026
Y
fix deepseek-v4-flash-0731 rtx pro 6000 deploy bug (#760)
yiminghub2024 committed
August 6, 2026
I
Enable the Offloading options on the Nemotron-3 / Nemotron-H family (#755)
Itay Etelis committed
I
Enable the Offloading options on GLM-5.2 (#754)
Itay Etelis committed
R
Y
add deepseek-v4-flash-0731 rtx pro 6000 cookbook (#758)
yiminghub2024 committed
Y
[DOCS]add minimax-H3 rtx pro 6000 cookbook & fix 4090 userguide bug (#757)
yiminghub2024 committed
T
Add AMD recipes for Inkling models and fix selector deadlocks (#747)
Tan Pin Siang committed
T
[Kimi-K3] Remove MegaMoE for non-GB. use flashinfer for MLA prefill (#752)
Thien Tran committed
Z
Add inclusionAI/Ling-3.0-flash recipe (#743)
zexplorerhj committed
B
[DSV4] Use probabilistic draft sampling (#751)
Benjamin Chislett committed
August 5, 2026
F
Update MiniMax M3 for B300 NVFP4 (#730)
Faradawn Yang committed
K
Add RedHatAI/Kimi-K3-NVFP4 recipe (#744)
Kyle Sayers committed
Y
Fix Simple CPU offload boolean (#746)
Yifan Qiao committed
Y
Add Intel XPU(B60, B70) deployment recipe for 4 models (#693)
Yejing Lai committed
August 4, 2026
A
Add MI325X recipe for Laguna S 2.1 (#665)
akii96 committed
S
Add MiniMax-H3 two-GPU DLO deployment recipe (#739)
SYLAR committed