We do not have unit tests but currently recommend the following to make sure nothing has broken:
- Run the notebook dense_correspondence/dataset/simple_datasets_test.ipynb
- Run the notebook dense_correspondence/correspondence_tools/pytorch-finding-correspondences.ipynb
- Run a training notebook
- Run qualitative eval notebook
- Run quantitative eval notebook