Datasets documentation
Overview
Get started
Tutorials
OverviewInstallationHugging Face HubThe Dataset objectTrain with 🤗 DatasetsEvaluate predictionsUpload a dataset to the Hub
How-to guides
OverviewLoadProcessProcess audio dataProcess image dataStreamShareCreate a dataset loading scriptCreate a dataset cardStructure your repositoryCache managementCloud storageSearch indexMetricsBeam Datasets
Conceptual guides
Reference
You are viewing v2.0.0 version. A newer version v4.8.4 is available.
Overview
Welcome to the 🤗 Datasets tutorial!
The goal of the tutorials are to help new users build up a basic understanding of 🤗 Datasets. You will learn to:
- Setup a virtual environment and install 🤗 Datasets.
- Load a dataset.
- Explore what’s inside a Dataset object.
- Use a dataset with PyTorch and TensorFlow.
- Evaluate your model predictions with a metric.
- Easily upload a dataset to the Hugging Face Hub.
After completing the tutorials, we hope you will have the necessary skills to start using our library in your own projects!
We understand that people who want to use 🤗 Datasets come from a wide and diverse range of disciplines. The tutorials are designed to be as accessible as possible to people without a developer background. If you already have some experience, take a look at our Quick Start to see an end-to-end code example in context.