- seq2seq (-> decoder with attention)
- AutoEncoder(->LSTM encoder & decoder layers)
- Transformer (-> Encoder & Decoder with Attention)
- SWAT dataset (not uploaded in this repo due to large size)
parser.py
Initial data preparation/preprocessing and normalization, converts .csv data to .dat format
swat_dataset.py
Class extends Torch Dataset to create train and test dataset with sliding window size
encoder.py
Implementation of LSTMAutoEncoder model
seq2seq.py
Implementation of Seq2Seq model
transformer.py
Implementaion of TransformerEncoder model
train.py
training all models, saved models are saved in checkpoints/ folder
evaluate.py
evaluating model with ground truth values (labels)
utils.py
helper functions for plotting graphs and calculating ROC-AUC results