Stateful CharRNN Text Generation Example

Instead of feeding every single character every frame to predict the next character, we feed only the last character, and instruct the CharRNN to remember its internal state.

This example uses a pre-trained model on a corpus of Virginia Woolf

seed text:

temperature: 0.5

Loading Model


This work was supported through a residency at the Frank-Ratchye STUDIO for Creative Inquiry at Carnegie Mellon University, Pittsburgh, October 2018, with additional support from the Sylvia and David Steiner Speaker Series.