Sign in to view source links and access this dataset
Description
A dataset for GUI element grounding tasks, curated from five major GUI interaction sources: Aria-UI, OmniAct, Widget Caption, UI-Vision, and OS-Atlas. It was created by Salesforce and last updated on October 3, 2025. The dataset combines and standardizes annotations to provide diverse, well-annotated interface interactions.
Use Cases
Train models for GUI element detection based on standardized annotations from multiple sources.
Benchmark grounding models for interface interaction tasks using the combined dataset.
Develop cross-dataset evaluation methods for GUI understanding leveraging the five integrated sources.
Strengths
Combines annotations from five established GUI interaction datasets, suggesting breadth of coverage.
Described as a high-quality dataset curated from multiple authoritative sources.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
Salesforce
Collection Method
Curated and standardized from five major GUI interaction datasets: Aria-UI, OmniAct, Widget Caption, UI-Vision, OS-Atlas.
Freshness
Last updated 2025-10-03 15:55:47; freshness should be verified.
License is unknown and must be verified before use.