Super RVRT

This article is a VSR article just published in June this year.It introduced a model that combines the recurrent network structure and the Transformer structure - Recurrent Video Restoration Transformer (RVRT).RVRT completes the balance between the Recurrent-based model and the Vision-Transformer model, thereby achieving a trade-off between model size, computational efficiency, and expressiveness.

Note:

1. This article only explores the VSR part.

Reference documents:
①Source code