Unverified50% confidenceFactExact time
The LoRA is of the Video-to-Video (V2V) type, with the core capability of re-solving a different camera path while preserving the subject content of the original footage.
1
Sources
50%
Confidence
Long-term
Relevance
9/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe LoRA is fine-tuned from the MiniMax H3 video model and completes cross-view transformation using a reference-frame-guided (Ref2VA) approach.75% similarUnverifiedVideo VQ-VAE, 3D convolution, and encoding uniformly sampled keyframes with ViT are the mainstream video tokenization methods.72% similarUnverifiedVideoViT 是将 Vision Transformer 架构扩展至视频域的技术路线，通过在时间维度引入注意力机制建模帧间关系，并使用 Tubelet Embedding 将连续帧时空体素化为三维 patch69% similarUnverifiedA Reddit user conducted a controlled variable comparative review of 11 physics-motion LoRAs on the WAN 2.1 image-to-video model.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/937577API
curl https://kongchang.com/api/v1/knowledge/claims/937577MCP
get_claim(id=937577)