Welcome to 🤗 fast.ai Sprint! ❤️ In this document, you will find everything you need to host your model and build demos successfully 🤗

The goal of this sprint is to train fast.ai learners and build demos for them and release them openly.

(Spoiler: There’s also swag involved at the end of the sprint 😏✨ See bottom of this page.)

For each submission, the participants will share dataset and model repositories and a Space made with those models. 🌌 (If this sounds like alien language to you, don’t worry, just keep reading! 👽)

Visit this page for our previous work and to get a sense of what we’re trying to achieve 🤩

Getting Started 🚀

To get started with the sprint, join us in fast.ai discord in #hf-fastai channel and simply leave a comment on the owner section of this spreadsheet to take ownership of example you wish to submit (we will fill the spreadsheet for you). If you have another dataset or a project idea, you can comment the link to it. Then you can get started with building! 🛠

Dataset Conversion

After you take ownership, you need to convert Kaggle dataset to Hugging Face dataset and push to 🤗 Hub. Please follow the instructions in this notebook to authenticate yourself on Kaggle & Hugging Face Hub, and then convert the dataset.

Model Hosting on the Hub

Essentially all models and Spaces are hosted inside git repositories on the Hub. You have to provide a token to push your code to those repositories. You can do it either by logging in or providing your token to pushing functions.

Simply install Hugging Face Hub library like below: pip install huggingface_hub["fastai"]

Logging In

Notebook Login

If you’re pushing a model from a notebook, you can just type