COMMITS
July 30, 2026
Y
Remove max-num-batched-tokens from DeepSeek config (#700)
yiminghub2024 committed
F
Update Nemotron-3-Super DGX Spark config (#637)
Faradawn Yang committed
July 29, 2026
S
Let Kimi recipes auto-select MLA prefill backend (#701)
Simon Mo committed
P
DGX Station recipes (#657)
prodney-dev committed
F
[ROCm] Add tuned AITER MXFP4 MoE config for Kimi-K2.5 on MI355X (#661)
Fangzhou Ai committed
R
Update kimi k3 notes (#697)
Roy Wang committed
R
update Kimi config (#696)
Roy Wang committed
July 28, 2026
T
Sort latest recipes by catalog addition date (#686)
Tiezhen WANG committed
H
[ROCm] Add Kimi-K2.5 AMD AITER MoE env vars (#680)
haic0 committed
H
[ROCm] Add MiniMax-M3 AMD AITER MoE env vars (#681)
haic0 committed
P
J
mindlab-research/Macaron-V1-Venti (#668)
Jared Wen committed
H
[AMD] [K3] update spec-decoding with amd override (#691)
Hongxia Yang committed
F
Minimax M3 agg (#682)
Faradawn Yang committed
R
Fix version & PD setup (#690)
Roy Wang committed
Y
add kimi-k3 b300 best practice (#689)
yiminghub2024 committed
July 27, 2026
T
K3 recipes (#684)
Tiezhen WANG committed
July 24, 2026
J
Enable AMD MI355X-tuned serving defaults for Hy3 (#677)
jotsaiamd committed
July 23, 2026
July 22, 2026
M
Remove redundant CUDA attention backend overrides (#667)
Matthew Bonanni committed
M
Let Mistral Small 4 auto-select attention backends (#666)
Matthew Bonanni committed
P
Add RTX Pro 6000 recipes (#648)
prodney-dev committed
July 21, 2026
M
Validate speculative decoding modes (#662)
Michael Goin committed
F
[ROCm] Document AITER MXFP4 MoE backend in MiniMax-M3 MI355X reproduction (#660)
Fangzhou Ai committed
J
Add Laguna-S-2.1 recipe (#659)
Joe Rowell committed
R
Remove multi node tp pp support for TML. (#658)
Roy Wang committed