COMMITS
May 14, 2026
Y
Add inclusionAI/Ring-2.6-1T recipe
yasong.wang committed
T
Verify Ling-2.6-flash and Ling-2.6-1T on MI300X (#464)
Tan Pin Siang committed
M
Update Kimi K2.5 Eagle3 MLA and add Kimi K2.6 NVFP4 variant (#462)
Mohammad Miadh Angkad committed
May 13, 2026
R
Add mistralai/Mistral-Small-4-119B-2603 recipe (#461)
Roy Wang committed
R
Add mistralai/Voxtral-Mini-4B-Realtime-2602 recipe (#460)
Roy Wang committed
May 12, 2026
R
M
Add quantized variants and speculative decoding support (#456)
Michael Goin committed
May 11, 2026
S
Fix the title of the DeepSeek/DeepSeek-OCR-2 Usage Guide (#436)
Serenus committed
M
mi300x verified tag for Qwen3Guard (#431)
matti-palomaki-amd committed
R
Add PaddlePaddle/PaddleOCR-VL-1.5 recipe (#454)
Roy Wang committed
R
Fix Gemma-4-31B VRAM and DeepSeek-R1 NVFP4 model_id (#455)
Roy Wang committed
May 10, 2026
M
Update GPT-OSS Eagle3 speculative decoding to v3 (#449)
Mohammad Miadh Angkad committed
May 9, 2026
R
Add jinaai/jina-embeddings-v5-text-small recipe (#448)
Roy Wang committed
May 8, 2026
A
Fix Gemma 4 limit-mm-per-prompt examples (#446)
Aratako committed
R
Add openai/gpt-oss-20b recipe and Laguna-XS.2 DFlash spec decoding (#447)
Roy Wang committed
R
Site polish: cu130 baseline, per-recipe state, provider naming (#445)
Roy Wang committed
May 7, 2026
L
fix: correct HuggingFace repo prefix in Gemma 4 MTP speculative decoding docs (#444)
Luciano Martins committed
M
MI300x verified tag for Ministral-3-8B-Reasoning-2512 (#443)
matti-palomaki-amd committed
May 6, 2026
S
Add homepage redirect to new recipes (#442)
Santino Ramos committed
May 5, 2026
L
fix: correct gemma-4-26B MTP name (#440)
Liana Koleva committed
R
Gemma4 MTP additional notes (#439)
Roger Wang committed
L
[Gemma4] Add MTP speculative decoding section to usage guide (#438)
Luciano Martins committed
R
Update model default tp size (#437)
Roy Wang committed
May 4, 2026
F
DeepSeek-V4-Pro: enable num_speculative_tokens=2 on Hopper (#435)
functionstackx committed
May 2, 2026
R
Update JSON API (#434)
Roy Wang committed
April 30, 2026
R
Update Eagle3 draft model for Kimi 2.6 (#428)
Roger Wang committed
R
add mistral medium (#427)
Roy Wang committed
April 29, 2026
D
fix(deepseek): set TP=8 for V4-Flash single_node_tp on 8-GPU nodes (#423)
Dhruv Singal committed
J
add Ling 2.6 1T (#426)
Jiangyun Zhu committed