# TAG attention-driven GUI grounding

Grounding and evaluation code that aggregates a pretrained vision-language model’s attention maps to localize requested interface elements without fine-tuning.

- Catalog URL: https://awesomecua.github.io/resources/github-heimingx-tag/
- Original source: https://github.com/HeimingX/TAG
- Kind: tool
- Review: curated
- Scope: core
- Environments: web
- Topics: grounding, evaluation, training-free
- Evidence checked: 2026-10-17T02:02:06Z
- ID: github:heimingx/tag

## Limitations

The release contains grounding and benchmark evaluation code. The inspected root repository metadata does not declare a code license. It uses external pretrained models rather than publishing a new trained checkpoint, and the README restricts redistribution of unzipped benchmark data.

- Record JSON: https://awesomecua.github.io/resources/github-heimingx-tag/record.json

## License declarations

- code: not recorded
- data: not recorded
- document: not recorded
- weights: not recorded

## Primary evidence

- https://raw.githubusercontent.com/HeimingX/TAG/HEAD/README.md (retrieved 2026-10-17T02:02:06Z; SHA-256 68cb0deb8a2eb49fb9236ae3d885b043d240dded0fce31cf7425130a56ff5819)
- https://api.github.com/repos/HeimingX/TAG (retrieved 2026-10-17T02:02:05Z; SHA-256 a33bee17685d64d23fb4d65b30b9c52f3a6dd5b21789d7b3509eb2e15145de6f)
- https://api.github.com/repos/HeimingX/TAG/git/trees/main?recursive=1 (retrieved 2026-10-17T02:02:05Z; SHA-256 fac4f5e38011bbcfef916063d495c21d49830759af4e8713a60d7bae992d2710)