Unverified50% confidenceFactExact time
In stage two, SUPRA-TEX projects the predicted PBR maps onto the visible surface and converts them, through a pretrained material VAE encoder, into 'native observation latents' that share the same latent space as the generation target, in order to achieve hyper-alignment.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedNative RGBA output means that the model simultaneously predicts the alpha transparency channel for each pixel during the generation phase, enabling it to directly output PNG images with transparent backgrounds, eliminating the need for post-processing.69% similarUnverified在Midjourney所使用的扩散模型架构中，图像先在压缩的低维潜空间中完成逐步去噪，再通过解码器还原为完整图像67% similarUnverified多模态实现路径分为早融合（预训练阶段混合多模态数据）和晚融合（先用独立视觉编码器ViT提取图像特征，再通过投影层对齐到语言模型embedding空间）66% similarUnverifiedCheckpoint是预训练的扩散模型权重文件通常2-7GB，VAE负责像素空间和潜空间之间转换图像，CFG Scale控制生成图像对文本提示词的遵循程度66% similarUnverifiedDINOv2的训练框架以iBOT为主干，结合DINO的自蒸馏目标和SwAV的聚类一致性损失，并加入masked image modeling任务64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1042491API
curl https://kongchang.com/api/v1/knowledge/claims/1042491MCP
get_claim(id=1042491)