0 Index1 Preparation2 Install3 Input4 Run5 What6 Discussion7 Play8 Share9 Presentation

Let’s play

Once the model has analysed and trained on a certain set of input data, we can now access any chackpoint that has been created to now reproduce text.

Run rnn.py again

This can even be done in a new Terminal window while still training in another. After you typed in your project-directory press p for play. You can now either play the newest created checkpoint or type in the filename of a previous one e.g. 20180730-033023-12000000.meta.

From the learned state, the model will now generate new text in real time that is saved into the òutput folder of your project.