# Swinv2 backbone network

The DPT (Dense Prediction Transformer) model is trained on 1.4 million images for monocular depth estimation. This model uses Swinv2 as the backbone network and is suitable for high-precision depth prediction tasks.