Generating New Recipes using GPT-2

In this 2-hour project, you will learn how to preprocess a text dataset comprising recipes and split it into training and validation sets. The course covers using the HuggingFace library to fine-tune a deep, generative model and training the model on Google Colab. You will also learn how to effectively use GPT-2 to create realistic and unique recipes from lists of ingredients based on the dataset.

Key takeaways include:

• Cleaning and preprocessing text data for modeling
• Creating datasets for large-scale language generation
• Fine-tuning a large-scale language model for the specific task of generating recipes

This project aims to provide insights into the resources required for large-scale model training, as well as the efficacy of knowledge distillation in such use cases. Please note that this course is most suitable for learners in the North America region.