1 paper
Tianlin Pan, Jiayi Dai, Chenpu Yuan +7
Recent video editing models have achieved impressive results, but most still require large-scale paired datasets. Collecting such naturally aligned pairs at scale remains highly ch…