Introducing new audio and vision documentation in 🤗 Datasets

Published Time: 2022-07-28T00:00:00.108Z

# Introducing new audio and vision documentation in 🤗 Datasets

[Hugging Face](https://huggingface.co/)

* [Models](https://huggingface.co/models)
* [Datasets](https://huggingface.co/datasets)
* [Spaces](https://huggingface.co/spaces)
* [Buckets new](https://huggingface.co/storage)
* [Docs](https://huggingface.co/docs)
* [Enterprise](https://huggingface.co/enterprise)
* [Pricing](https://huggingface.co/pricing)
*
*
* * *

* [Log In](https://huggingface.co/login)
* [Sign Up](https://huggingface.co/join)

[Back to Articles](https://huggingface.co/blog)

# [](https://huggingface.co/blog/datasets-docs-update#introducing-new-audio-and-vision-documentation-in-%F0%9F%A4%97-datasets) Introducing new audio and vision documentation in 🤗 Datasets

Published July 28, 2022

[Update on GitHub](https://github.com/huggingface/blog/blob/main/datasets-docs-update.md)

[- [x] Upvote 1](https://huggingface.co/login?next=%2Fblog%2Fdatasets-docs-update)
* [](https://huggingface.co/Aanuoluwapo65 "Aanuoluwapo65")

[](https://huggingface.co/stevhliu)

[Steven Liu stevhliu Follow](https://huggingface.co/stevhliu)

* [Quickstart](https://huggingface.co/blog/datasets-docs-update#quickstart "Quickstart")

* [Dedicated guides](https://huggingface.co/blog/datasets-docs-update#dedicated-guides "Dedicated guides")

* [ImageFolder](https://huggingface.co/blog/datasets-docs-update#imagefolder "ImageFolder")

* [What’s next?](https://huggingface.co/blog/datasets-docs-update#whats-next "What’s next?")

Open and reproducible datasets are essential for advancing good machine learning. At the same time, datasets have grown tremendously in size as rocket fuel for large language models. In 2020, Hugging Face launched 🤗 Datasets, a library dedicated to:
1. Providing access to standardized datasets with a single line of code.
2. Tools for rapidly and efficiently processing large-scale datasets.

Thanks to the community, we added hundreds of NLP datasets in many languages and dialects during the [Datasets Sprint](https://discuss.huggingface.co/t/open-to-the-community-one-week-team-effort-to-reach-v2-0-of-hf-datasets-library/2176)! 🤗❤️

But text datasets are just the beginning. Data is represented in richer formats like 🎵 audio, 📸 images, and even a combination of audio and text or image and text. Models trained on these datasets enable awesome applications like describing what is in an image or answering questions about an image.

The 🤗Datasets team has been building tools and features to make working with these dataset types as simple as possible for the best developer experience. We added new documentation along the way to help you learn more about loading and processing audio and image datasets.

## [](https://huggingface.co/blog/datasets-docs-update#quickstart) Quickstart

The [Quickstart](https://huggingface.co/docs/datasets/quickstart) is one of the first places new users visit for a TLDR about a library’s features. That’s why we updated the Quickstart to include how you can use 🤗Datasets to work with audio and image datasets. Choose a dataset modality you want to work with and see an end-to-end example of how to load and process the dataset to get it ready for training with either PyTorch or TensorFlow.

Also new in the Quickstart is the `to_tf_dataset` function which takes care of converting a dataset into a `tf.data.Dataset` like a mama bear taking care of her cubs. This means you don’t have to write any code to shuffle and load batches from your dataset to get it to play nicely with TensorFlow. Once you’ve converted your dataset into a `tf.data.Dataset`, you can