EmbedSleepNet

This repository reproduces TinySleepNet for sleep stage prediction based on the signal channel EEG using PyTorch and implements a new smaller and faster network called EmbedSleepNet.

Environment Setup

1. Create a virtual environment with Python 3.8: virtualenv venv --python=python3.8
2. Activate the environment: source venv/bin/activate or venv\Scripts\activate.bat (for Windows)
3. Install dependencies: pip install -r requirements.txt
4. Download and extract Sleep-EDF dataset from https://www.physionet.org/content/sleep-edfx/1.0.0/
5. Preprocess the data by running python preprocess.py --data_path PATH_TO_sleep-cassete_FOLDER

Running

To start the training process simply run: python train.py --flavor=[embed, tiny], where tiny represents original TinySleepNet, and embed represents newly introduced EmbedSleepNet. You may additionally specify the number of epochs and model output name, for example:

python train.py --flavor embed --epochs 450 --model-name model