Meta 用蒸馏实现 2 步扩散实时视频流
Live video streaming needs to respond instantly while remaining visually consistent over long conversations.
AI 导读
实时视频流需要即时响应，同时在长时间对话中保持视觉一致性。 我们通过将带 3-way CFG 的大型 40 步扩散教师模型（每个视频块 120 次评估）蒸馏为带固定长度 KV cache 的无引导 2 步因果学生模型，实现了这一点。 Self-forcing 帮助学生模型抵抗漂移，以少 60 倍的评估次数保持接近教师模型的质量。
正文
当前显示原文，中文全文译文尚未就绪。
Live video streaming needs to respond instantly while remaining visually consistent over long conversations.
We achieved this by distilling a large 40-step diffusion teacher with 3-way CFG (120 evaluations per video chunk) into an unguided 2-step causal student with a fixed-length KV cache.
Self-forcing helps the student resist drift and maintain near teacher quality with 60x fewer evaluations.