Unverified50% confidenceFactTime unknown
Stable Diffusion series models focus on text-to-image generation tasks.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026
Sources
Related Claims
Verified扩散模型以 DALL-E 2、Stable Diffusion、Midjourney 为代表，重新定义了图像生成任务76% similarUnverified扩散模型（如Stable Diffusion）通过逐步去噪还原图像，擅长细节纹理；自回归多模态模型（如GPT Image 2）在理解复杂文字指令和保持角色一致性方面更具优势71% similarVerifiedStable Diffusion基于潜在扩散模型（Latent Diffusion Model），通过在压缩的潜在空间中进行去噪过程来生成高质量图像70% similarUnverified扩散模型的数学本质是基于随机微分方程和得分匹配理论，支撑着Stable Diffusion、DALL-E等图像生成系统69% similarUnverified当前主流的文生图模型包括Stable Diffusion、DALL-E、Midjourney，文生视频模型包括Sora、Runway Gen系列、即梦AI底层模型69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/54532API
curl https://kongchang.com/api/v1/knowledge/claims/54532MCP
get_claim(id=54532)