الباحثون

Fangyu Lin

المنشورات 2

نسخة أولية وصول مفتوح

Enhancing Autoregressive Video Generation via Representation Adversarial Distillation

Fangyu Lin, Xingtong Ge, Lunjie Zhu وآخرون · 2026

Few-step autoregressive video generation enables efficient streaming synthesis, but errors introduced in early temporal blocks are reused as context and can propagate through subsequent rollouts, leading to detail degradation, structural drift, and unstable motion. Existing distribution matching distillation (DMD) prim …

نسخة أولية وصول مفتوح

Salt++: Context-Aligned Post-Training for Few-Step Streaming Multimodal Generation

Xingtong Ge, Yutong Wang, Lunjie Zhu وآخرون · 2026

Few-step streaming audio--video generation requires both causal modeling and step distillation, yet standard training recipes face two context-related challenges. Teacher forcing pairs clean history with a noisy target, but supervises predictive contextual representations only indirectly through velocity prediction. Me …

المؤلفون المشاركون