8. Access to datasets and models

Chapter 8 overview:

This section is for researchers, developers, humanitarian technology practitioners, and program managers who need to access and use TWB Voice datasets and models.

We cover:

• how to use Hugging Face as our main platform for sharing data

• how to understand different types of licensing

• how to access open and gated datasets

• how to work with pre-trained speech AI models.

For most of this section, you don’t need much technical expertise at all. It focuses on practical guidance to help you find and download resources. But if you are planning to use these datasets and models in your own projects, it would be good to have some basic knowledge of programming. This is particularly true with programmatic data access through Python libraries.

TWB Voice creates useful speech datasets and AI models. We aim to help the broader world of research and the humanitarian technology community. In this section, we explain how to access and use these resources through our main data sharing platform, Hugging Face.