Public repository for a minimum working example of the topographs project.
- Associated paper: arxiv
- Setup the environment.
- You can setup your own environment using the requirements.txt file. The project was tested with python3.8
- Alternatively you can build a docker image with the provided Dockerfile
- Download the data from zenodo data
- Make sure to adjust the paths in the configuration files to point to the data.
To train a model, simply run python train.py configs/config_full.yaml log_dir. This will train a model on complete events only saving all outputs in the directory log_dir which will be created.
Alternatively you can use configs/config_partial.yaml to train including partial events.