This repository is an unofficial Pytorch implementation of "High Quality Monocular Depth Estimation via Transfer Learning". The official implementation can be found here.
NYU Depth Dataset V2 dataset is required. You can download the preprocessed dataset here. You do not need to unzip this file.
-
Open the configs.py file and edit the dataset path and training options.
-
For training, use the following simple command.
$ python train_nyuv2.py
- Results for NYU Depth Dataset V2
- complete the KITTI data loader
- Visualization code