README.md

---
title: README
emoji: 🏆
colorFrom: purple
colorTo: indigo
sdk: static
pinned: false
---

<p align="center">
    <img src="https://huggingface.co/spaces/HuggingFaceM4/README/resolve/main/IDEFICS.png" alt="Idefics-Obelics logo" width="200" height="100">
</p>

HuggingFaceM4 is the multimodal team at Hugging Face, working on vision-language models.

Within this organization on the Hugging Face hub, you can access the Idefics models (version 1 [IDEFICS](https://huggingface.co/HuggingFaceM4/idefics-80b-instruct), version 2 [Idefics2](https://huggingface.co/HuggingFaceM4/idefics2-8b), version 3 [Idefics3](https://huggingface.co/HuggingFaceM4/Idefics3-8B-Llama3)), datasets used for the training like [OBELICS](https://huggingface.co/datasets/HuggingFaceM4/OBELICS), [WebSight](https://huggingface.co/datasets/HuggingFaceM4/WebSight), [The Cauldron](https://huggingface.co/datasets/HuggingFaceM4/the_cauldron) or [Docmatix](https://huggingface.co/datasets/HuggingFaceM4/Docmatix), and interactive tools to visualize the results.