Implemented a Neural Machine Translator in Keras and TensorFlow, with Encoder-Decoder architecture, LSTM layers and Attention Mechanism to improve translations. The model takes French sentences as input and translates it to English. This project utilizes the concepts of NLP, Word Embeddings (Word2Vec and Glove), Neural Networks & Attention Mechanism.
The accuracy of the model is measured using BLEU score. For French to English language pair, the BLEU score is 37.88.