Team 7: 3D UNet for atmospheric aerosol prediction

• Place training.nc and evaluation.nc in ./data/
• Run Preprocessing.ipynb to generate normalized data used for training and evaluation
• sh ./run.sh to run train.py training script across multiple GPUs
• Evaluation.ipynb for evaluating relative error after models have been saved to ./checkpoints/
• Visualization.ipynb for visualization.

Link to presentation:

https://docs.google.com/presentation/d/1hwNx4CFVkG0f0ljpc_z0XGmhjDXFAs5nDaur5fw35Zs/edit?usp=sharing