This project uses Layers API of TensorFlow.js to create a text classifier for three books from Project Gutenberg
- The three books I picked are: Paradise Lost, Macbeth and Bible, each with the first 1000 sentences.
- All the details are shown in the jupyter notebook.
- Gutenberg-LSTM/index.html and Gutenberg-LSTM/index.js include the JavaScript to launch the demo
- Gutenberg-LSTM/python/ contains my preprocessing, Keras model with LSTM layer and the code for history callback plot
- Gutenberg-LSTM/model_js/ contains my TensorFlow.js model (model.json + shards) and metadata.json