d3-ultra: open 27B multimodal decision model (#2/230 on Jev Decision Index text), Apache-2.0 weights on Hugging Face.
Xunzhuo @XunzhuoLiu
EN
- 11.3k views
- 123 likes
- 85 bookmarks
- 13 reposts
About this case
Product/open-model note with media. XunzhuoLiu introduces d3-ultra, a 27B multimodal decision model ranked #2 of 230 on the Jev Decision Index text board (+3.03 over d3). Handles text, images, video with boxing/tracking over time. Open weights Apache-2.0 at huggingface.co/vllm-sr/d3-ultra.
Original posthttps://x.com/XunzhuoLiu/status/2114915380926017591