# UI-AGILE checkpoint

The released UI-AGILE model for GUI grounding with reward shaping, crop-based resampling and decomposed selection during inference.

- Catalog URL: https://awesomecua.github.io/resources/url-b94e285c3f757e2d0f3a/
- Original source: https://huggingface.co/KDEGroup/UI-AGILE
- Kind: model
- Review: curated
- Scope: core
- Environments: android
- Topics: grounding, reinforcement-learning, efficiency
- Evidence checked: 2026-10-17T02:44:30Z
- ID: url:b94e285c3f757e2d0f3a

## Limitations

Four safetensors shards and model/tokenizer files are public. The card declares MIT for these weights; the companion dataset has no declared license and is separate. The short card does not state a parameter count, so this entry does not infer one from shard sizes.

- Record JSON: https://awesomecua.github.io/resources/url-b94e285c3f757e2d0f3a/record.json

## License declarations

- code: not recorded
- data: not recorded
- document: not recorded
- weights: mit — https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md

## Primary evidence

- https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md (retrieved 2026-10-17T02:44:30Z; SHA-256 75aed1ebff7dd54ab1dc0112fe524a118d9e8845c2e079080d2bb7954f7ab36e)
- https://huggingface.co/api/models/KDEGroup/UI-AGILE (retrieved 2026-10-17T02:44:29Z; SHA-256 a52032bb185885d47466f0bea26c15ccefeb8d6ff89f3bacc5eb4192c35d45f7)
- https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/raw/main/README.md (retrieved 2026-10-17T02:43:19Z; SHA-256 ce35413bc1ef28338bc6fd9f132fa271a4f5e0bf23cecd7346f54adf0a964fd7)
- https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md (retrieved 2026-10-17T02:44:30Z; SHA-256 75aed1ebff7dd54ab1dc0112fe524a118d9e8845c2e079080d2bb7954f7ab36e)