Unverified50% confidenceFactExact time
AnimateDiff was proposed by Shanghai AI Laboratory in 2023, achieving motion modeling by inserting trainable temporal attention layers into the UNet architecture of Stable Diffusion while freezing the original spatial layer weights.
1
Sources
50%
Confidence
Long-term
Relevance
9/10/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedChina released the Interim Measures for the Management of Generative Artificial Intelligence Services in 2023, requiring that both training data and generated content be compliant.66% similarUnverifiedControlNet was proposed by Lvmin Zhang in 2023, injecting structural conditions into pre-trained diffusion models without modifying the original model weights.64% similarUnverifiedV-JEPA was released by Meta AI in 2024, training models to predict masked regions in representation space by masking up to 90% of the spatiotemporal extent of a video.64% similarUnverifiedMeta于2023年发布Segment Anything图像分割基础模型，训练了超过11亿个掩码，实现zero-shot通用分割能力62% similarUnverifiedZero-1-to-3等工作于2022年前后首次系统性证明预训练图像扩散模型具备隐式3D空间理解能力62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/889213API
curl https://kongchang.com/api/v1/knowledge/claims/889213MCP
get_claim(id=889213)