An implementation of vanilla recurrent neural network in python (without libraries) for a "many-to-one" classification task.
- models folder: model parameters are saved in this folder when
trainModel.py
is run - bestModel folder: has the parameters for the best trained model for this implementation
- src folder: has the implementation of the RNN
- testData folder: has the test data
- trainingData folder: has the training data and labels
- trainModel.py: code to train the network
- testModel.py: code to test the trained network
- testPrediction.csv: output of the trained network. Genetated when testModel.py is run
- python v3
- python libraries : numpy
For training the network
python3 trainModel.py
For testing the trained network
python3 testModel.py
This project is in progress and may contain issues.