The GPT2 model is fine-tuned on the Shakespeare dataset.

In this repository you can find the code for training the model and generating text from it.