Unverified50% confidenceFactExact time
Stable Diffusion通过预训练的VAE将512×512×3的图像编码为64×64×4的潜在张量，计算量降低约48倍
1
Sources
50%
Confidence
Long-term
Relevance
8/27/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedStable Diffusion将扩散过程从像素空间压缩至潜在空间，由变分自编码器（VAE）实现69% similarUnverifiedStable Diffusion XL约有34亿参数，推理一张图像需要执行20至50步去噪迭代66% similarUnverified扩散模型（Diffusion Model）的兴起使Stable Diffusion、Midjourney等工具将图像生成门槛降至消费级硬件可运行水平64% similarUnverifiedIn image generation, distillation techniques can reduce the number of sampling steps from 50-100 steps down to 4-8 steps while achieving comparable results64% similarVerifiedLatent Diffusion Model由Rombach等人于2022年提出，将扩散过程与VAE的潜在空间结合，将计算维度从512×512×3降低到64×64×464% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/810103API
curl https://kongchang.com/api/v1/knowledge/claims/810103MCP
get_claim(id=810103)