# Hugging Face — Unlocking the conversion of Web Screenshots into HTML Code with the WebSight Dataset

- Company: Hugging Face (huggingface.co)
- Announced: 2024-03-15T00:00:00+00:00
- Category: research-paper
- Subject: Platform
- Models affected: Sightseer, WebSight-v0.1, WebSight-v0.2
- Source: https://huggingface.co/blog/websight
- Record: https://forck.live/items/1928-unlocking-the-conversion-of-web-screenshots-into-html-code-with-the-websight

Hugging Face introduces the WebSight dataset, a large synthetic dataset of screenshot/HTML code pairs for training AI models to convert web screenshots into HTML code. The dataset has been updated to version 0.2, featuring 2 million examples and improvements like real images and Tailwind CSS. A model fine-tuned on this dataset, Sightseer, is also presented.

## Evidence

Verbatim from https://huggingface.co/blog/websight:

> WebSight is a new dataset that aims at building AI systems capable of transforming screenshots to HTML code.

---

Record: https://forck.live/items/1928-unlocking-the-conversion-of-web-screenshots-into-html-code-with-the-websight
Catalogue: https://forck.live/llms.txt
Feed: https://forck.live/feed.md