Presented to the Atlanta Deep Learning Meetup group on January 13th 2020

The repository contains the powerpoint presentation along with a Google Colab notebook showing how to fine-tune GPT2. You just need to download the file Transformer_Demo and go to colab and upload the notebook. Note, that your own dataset will be required to do the fine-tuning. In Google Colab make sure you set the runtime to GPU by changing the runtime type. Due to the memory constraints of colab it is only possible to fine-tune on small gpt2 model.