Train and use models for detection of events in eye tracking data
Clone the project
git clone https://github.com/martindue/eed.gitGo to the project directory
cd eedInstall dependencies (using conda)
conda env create -f environment.yamlActivate the environment
conda activate eedThe code currently works with the 6 .npy files in this folder: https://github.com/r-zemblys/irf/tree/master/etdata/lookAtPoint_EL
They have to be placed into the folders .data/raw/train_data and .data/raw/test_data/ which both need to be created manually by the user.
I plan to add support for more publicly available data.
The program uses the lightningCLI and jsonargparse command line interfaces.
Running training:
python src/ml/scripts/main.py fit --config src/ml/config/main.yamlMany options and settings can be set in the config file.
Hyperparamter search for sklearn with Optuna:
python src/ml/scripts/tune_sklearn.py --config src/ml/config/sklearn_tune.yaml