[KongchangAI]
Verified65% confidenceFactExact time

现代主流图像生成模型普遍采用潜在扩散模型(LDM)框架,通过VAE将图像压缩至低维潜在空间,压缩步骤会导致高频细节信息损失

3
Sources
65%
Confidence
Long-term
Relevance
7/24/2026
First Seen

Sources

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/605012
API
curl https://kongchang.com/api/v1/knowledge/claims/605012
MCP
get_claim(id=605012)