To build a multi-headed model that’s capable of detecting different types of of toxicity like threats, obscenity, insults and identity hate using Keras.
Dataset : Wikipedia Comments taken from Kaggle
Model: Bi-direcctional LSTM with two dense layers.