# GroundGUI RL Dataset

Desktop grounding prompts selected from GroundCUA and formatted for click-based reinforcement learning with target-box rewards.

- Catalog URL: https://awesomecua.github.io/resources/url-4e71dc456c4112eed727/
- Original source: https://huggingface.co/datasets/PrentisAI/GroundGUI_RL
- Kind: dataset
- Review: curated
- Scope: core
- Environments: desktop
- Topics: grounding, reinforcement-learning, data-generation
- Evidence checked: 2026-10-17T01:48:10Z
- ID: url:4e71dc456c4112eed727

## Limitations

The card declares MIT and attributes the source examples to GroundCUA. Its separate claim about UI-Vision overlap is not independently validated by this catalog.

- Record JSON: https://awesomecua.github.io/resources/url-4e71dc456c4112eed727/record.json

## Structured profile

- formats: Parquet
- modalities: screenshots, text, coordinates
- splits: train
- supervision: instruction-to-click grounding, target bounding boxes

### Count observations

- 12,600 prompts — RL training corpus (reported; source: https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md)
- 11,205 screenshots — RL training corpus (reported; source: https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md)
- 87 desktop applications — RL training corpus (reported; source: https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md)

Publisher-reported counts cover a selection and reformatting of GroundCUA, not a new disjoint image collection. No independent image-overlap audit was performed.

## License declarations

- code: not recorded
- data: mit — https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md
- document: not recorded
- weights: not recorded

## Primary evidence

- https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md (retrieved 2026-10-17T01:48:10Z; SHA-256 34f0025f686d0e48e2baa5244056a4362018aaad77ac54abea7f1f6eb25c4276)
- https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md (retrieved 2026-10-17T01:48:10Z; SHA-256 34f0025f686d0e48e2baa5244056a4362018aaad77ac54abea7f1f6eb25c4276)