To obtain the BAM dataset, follow the instructions here. Once finished, you should have be able to download a .sqlite file. Save this file in the /data folder. The download script requires sqlite3 and parallel, which can be obtained by running sudo apt install sqlite3 parallel. Then run the download script from the /data folder:

./get_data.sh

To process the data into the appropriate format, run:

python main.py --mode=gen_data --update

If the data is already saved, run:

python main.py --mode=gen_data --no-update

To train your model, run:

python main.py --mode=train --log_folder=1

This will create a subfolder named 1 under the folder results.