A Frozen Pixel-Space Diffusion Model Can Guide Itself with Its Own Samples

• 类型：arxiv
• 标识：2607.29122
• 链接：https://arxiv.org/abs/2607.29122
• 主分类：multimodal
• 形态：method
• 被引：1
• 被引来源：Semantic Scholar
• S2被引：1
• 影响力被引：0
• TLDR：This work attaches a lightweight prediction head to an intermediate layer, keep the backbone frozen, and use the discrepancy between the intermediate and final predictions as a self-guidance direction during sampling to train a frozen, pretrained pixel diffusion model that can guide itself.
• 副分类：engineering
• 待LLM分类：否
• 标题中文：冻结的像素空间扩散模型可借助自身采样进行自我引导
• TLDR中文：本文在中间层附加轻量预测头，保持 backbone 冻结，并利用中间预测与最终预测的差异作为采样时的自引导方向，训练一个能够自引导的冻结预训练像素扩散模型
• 来源文件：
• /inbox/tom/_candidates/2026-08-04-agent-rag-longcontext-candidates.json
• [S2 enrich]