---
datasets:
- shuaishuaicdp/GUI-World
language:
- en
license: cc-by-4.0
metrics:
- bertscore
- LLM-as-a-Judge
tags:
- gui
- agent
pipeline_tag: video-text-to-text
---

This is the first VideoLLM with powerful GUI-oriented capabilities, retrained on [GUI-World](https://gui-world.github.io). 

It was presented in [GUI-WORLD: A Dataset for GUI-oriented Multimodal LLM-based Agents](https://huggingface.co/papers/2406.10819).

See [Github](https://github.com/Dongping-Chen/GUI-World) for how to use GUI-Vid for GUI understanding tasks.