Unverified85% confidenceFactTime unknown
早期的离散扩散语言模型如 MDLM、SEDD 通过在 Token 嵌入空间或掩码空间中定义扩散过程来解决文本离散性挑战
1
Sources
85%
Confidence
Long-term
Relevance
5/31/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe underlying architecture of mainstream text-to-image models such as Stable Diffusion, Midjourney, and DALL-E typically combines a diffusion model with a CLIP text encoder.67% similarUnverified掩码扩散框架将加噪重新定义为token掩码替换，使扩散过程在离散文本域得以自洽66% similarUnverified研究者提出了掩码扩散（Masked Diffusion）、吸收态扩散（Absorbing Diffusion）等变体以解决文本离散空间的扩散挑战66% similarUnverified通过将模糊的风格描述拆解为具体设计参数（如无衬线字体、大量留白等），可以大幅收窄大语言模型的输出空间，提升结果一致性和可控性65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/14055API
curl https://kongchang.com/api/v1/knowledge/claims/14055MCP
get_claim(id=14055)