DSB_2018

Data Science Bowl 2018

1. SSH into the server, using your own account
2. Clone this repository into your folder
3. source activate dsb (shared conda environment)
4. cd lucas-u-net
5. python train.py: this should create a .csv file that you can then submit on Kaggle.

(You might have to create an empty folder called numpy_dumps in "DSB_2018" if it gives you an error about the directory being non-existent. This is where the pre-processed data gets stored for later use.)

The raw data itself is stored at /hdd/datasets/datasciencebowl2018/.