Deep Learning Project at Denmark Technical University
After downloading the files from Google Drive Folder arrange them in the following structure:
02456-MonoDepth/
├── DepthAnything_MySet/
├── DepthAnything_vkitti2/
├── Zoe_MySet/
├── Zoe_MySet_vkitti2/
├── KITTI/
├── VirtualKITTI_2/
├── Zoe_MySet_vkitti2/
├── EvalSetMetrics.ipynb
├── EvalSetMetricsVirtual.ipynb
├── util_parse_3d_dataset.py
├── README.md
ZoeDepth predictions are obtained using the repo.
python evaluate.py -m zoedepth_nk -d <kitti | vkitti2>DepthAnything predictions are obtained using the repo.
python evaluate.py -m zoedepth --pretrained_resource="local::./checkpoints/depth_anything_metric_depth_outdoor.pt" -d <kitti | vkitti2>ZoeDepth is trained
python train_mono.py -m zoedepth -d <nyu | kitti> --pretrained_resource=""Training logs can be found here, further more wandb platform is used for detailed logging during training and can be found on my github page