Unverified50% confidenceBenchmarkExact time
ChangeFormer采用层次化Transformer编码器，在LEVIR-CD和DSIFN-CD等基准上取得了SOTA性能
1
Sources
50%
Confidence
Long-term
Relevance
8/10/2026
First Seen
Sources
1500组物体形变配对图像数据集：损伤检测AI训练数据的价值解析
Related Claims
Partially VerifiedSora adopts the DiT (Diffusion Transformer) architecture, compressing video into spatiotemporal patch sequences before feeding them into a Transformer, replacing the traditional U-Net with a Transformer as the denoising backbone network.64% similarUnverified相比RNN/LSTM架构，Transformer支持大规模并行计算62% similarUnverifiedSwitch Transformer and the Mixtral series have widely adopted the MoE architecture.62% similarUnverifiedSD3引入了MMDiT（Multi-Modal Diffusion Transformer）架构，用Transformer替代了传统的U-Net骨干网络，并采用三重文本编码器（CLIP ViT-L、OpenCLIP ViT-bigG、T5-XXL）62% similarVerifiedFLUX adopts a DiT (Diffusion Transformer) architecture, replacing the traditional U-Net structure with a pure Transformer.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/723118API
curl https://kongchang.com/api/v1/knowledge/claims/723118MCP
get_claim(id=723118)