This repository contains the code of [SSLAPP] Self-Supervised Learning with Attention-based Latent Signal Augmentation for Sleep Staging with Limited Labeled Data If the code or the paper has been useful in your research, please add a citation to our work:

• CPU or NVIDIA GPU
• Python 3.9.7 x64
• pytorch 1.11.0
• numpy 1.21.5
• scikit-learn 1.0.2
• scipy 1.8.0

• Dataset used in our paper, SleepEDFX and ISRUC can be downloaded from here.
• Each data is sliced as segments composed of one epoch (30 seconds).
• 2 channels, EEG and EOG are used in our research.
• 'EEG_train' and 'EOG_train' in SleepEDFX is used for representation learning, and the others as finetuning and evaluation.

In order to train a model for SSLLAP, use main.py script. Following are the main parameters for training:

--klwt : importance of pair wise representation loss 
--segment : number of segment you want to use
--lambda_g : importance of global loss
--lambda_l : importance of local loss

For example, to train our model, use the following:

python main.py --klwt 10 --segment 6 --lambda_g 0.5 --lambda_l 0.5

After the representation learning stage, use finetune.py to evaluate our model.

python finetune.py