AI Difficulty estimation for In The Groove (ITG) technical charts as investigated by AITG: A Machine Learning Approach to ITG Difficulty Analysis
After cloning/downloading the project repository and installing the required libraries as listed in requirements.txt;
Run
python main.py
for a list of available commands and arguments.
To execute a full training and prediction run of a new model named MyModel, while saving a copy of the model, run:
python main.py full --name=MyModel -s
This will train the model for a default 10 epochs (configurable via the --epochs= argument) on any charts located in ./data/dataset/, and start a prediction run using charts in ./data/testset/.
The per-chart prediction results will be printed to the terminal and a csv in the model directory, along with a graph comparing the user and machine generated values for the run.