Multi-layer recurrent neural networks for training and sampling from texts, inspired by karpathy/char-rnn.
the Keras deep learning library.
All input data should be placed in the data/
directory. The example input.txt
is taken from the Nottingham Dataset (Cleaned).
To train the model with default settings:
$ python train.py
To sample the model:
$ python sample.py 100
Training loss/accuracy is stored in logs/training_log.csv
.