The research paper for this project can be found by navigating to the sentiment manifolds pdf in this repo.

After downloading IMBD data from kaggle site (https://goo.gl/1of8KR), copy the directory train into the training_data directory.

So the directory structure should be: training_data : train : {pos,neg} where pos and neg are directories containing reviews.

To train network do the following: modelData is a tuple containing compiled model, Xtrain,ytrain,Xtest,ytest where xtest,ytest are dev sets:

from network_architecture_7 import build_CNN_model,train_CNN_model

modelData = build_CNN_model()

train_CNN_model(modelData)