We developed the language model and so on. This research is based on recurrent neural networks, LSTM networks and the new Transformer architecture and attention mechanism. - View it on GitHub