# AutoGUI v1 training data

Screenshot-based grounding and captioning samples with UI functionality descriptions, target coordinates, and bounding boxes.

- Catalog URL: https://awesomecua.github.io/resources/url-79287058342289aa5d45/
- Original source: https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k
- Kind: dataset
- Review: curated
- Scope: core
- Environments: web, cross-platform
- Topics: grounding, screen-understanding
- Evidence checked: 2026-10-16T10:42:13Z
- ID: url:79287058342289aa5d45

## Limitations

The row count is not a unique-screen count. The data card documents image scaling conventions that affect coordinate interpretation.

- Record JSON: https://awesomecua.github.io/resources/url-79287058342289aa5d45/record.json

## Structured profile

- formats: not recorded
- modalities: screenshots, instructions, functionality labels, bounding boxes
- splits: train
- supervision: element grounding, functionality captioning

### Count observations

- 701,864 training rows — default/train in the dataset card metadata (reported; source: https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md)

Rows are not unique screenshots. The card documents image-downscaling conventions that affect coordinates; preserve the device and original image-size fields.

## License declarations

- code: not recorded
- data: cc-by-4.0 — https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md
- document: not recorded
- weights: not recorded

## Primary evidence

- https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md (retrieved 2026-10-16T10:42:13Z; SHA-256 c73b46acc2ff034a443fb8dbde327de1888535267199c9fb1819adf66898f1d7)
- https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md (retrieved 2026-10-16T10:42:13Z; SHA-256 c73b46acc2ff034a443fb8dbde327de1888535267199c9fb1819adf66898f1d7)