For toxic Comment classification project, we use LSTM layers in the machine learning model. The architecture of our model consists of an embedding layer, an LSTM layer, and a Dense layer at the end.
The model obtained 99% accuracy on the training set and 99% accuracy on the test set.