Showing 2026Show all
3 papers · 1 filter
cs.CV2026
CT-OPD: Counterfactual Trace On-Policy Distillation for Diffusion Vision-Language Models
Long Qian, Bingke Zhu, Jiaqi Wei +3
Diffusion vision-language models generate answers by gradually resolving masked tokens, making accurate conditional prediction in partially resolved states central to post-training…
cs.CV2026
OmniMoE-VL: A Sparse Vision-Language Model with Coupled Visual-Depth Routing
Long Qian, Bingke Zhu, Jiaqi Wei +2
Vision-language models (VLMs) increasingly use sparse mixture-of-experts (MoE) to scale language-side computation, yet visual information is typically routed only after passing thr…
cs.CV2026
Messages, Not Tokens: Grounded Coresets for Faithful VLM Compression
Long Qian, Jiaqi Wei, Bingke Zhu +2
Modern vision language models (VLMs) turn high-resolution images into long sequences of visual tokens. Every token traverses the language decoder and persists in its prompt KV cach…