Earlier this month @huggingface released a number of notebooks that walk users through some NLP basics. The three-part series, written by @MorganFunto, covers tokenizers, transformers, and pipelines utilizing Hugging Face’s transformer library. The notebooks cover the basics on a high level and get you working in the code quickly. The notebooks written in Colab allows anyone to run the code in the browser. Here’s the intro from the tokenization notebook:

Before going deep into any Machine Learning or Deep Learning Natural Language Processing models, every practitioner should find a way to map raw input strings to a representation understandable by a trainable model. One very simple approach would be to split inputs over every space and assign an identifier to each word.

The repo contains official notebooks provided by hugging face but also has a call for transformer notebooks from the community:

…we would like to list here interesting content created by the community. If you wrote some notebook(s) leveraging transformers and would like be listed here, please open a Pull Request and we’ll review it so it can be included here.

In addition to the three-part series described above, there are notebooks on “How to train a language model” and “How to generate text“.  You can find more details about the transformer library in their repo or paper. You can also use a transformer to generate text in the browser with their “write with transformer” tool.

Written by Rebecca Minich, Product Analyst, Data Science at Google. Opinions expressed are solely my own and do not express the views or opinions of my employer.