Seed-42 LoRA adapter for the ECCV 2026 CORA release. It must be used with google/gemma-3-4b-it at revision 093f9f388b31de276ce2de164bdc2081324b9767 and the matching CORA config included as cora_config.yaml.

Evaluation

QuIC-360 test set (5349 query-caption pairs):

BLEU-4	METEOR	ROUGE-L	CIDEr	SPICE
0.0447	0.1149	0.2469	0.3502	0.1691

Three-seed aggregate results and exact split hashes are maintained in the CORA-360 repository.

Usage

bash

git clone --branch v2.0.0-eccv2026 https://github.com/wooseungw/CORA-360.git
cd CORA-360
./reproduce.sh evaluate gemma3-4b-vicreg /path/to/test.csv

Limitations and license

CORA was evaluated on English query-focused captions from QuIC-360. Performance outside that domain, on non-ERP imagery, or in safety-critical settings is not established. This adapter is subject to the Gemma Terms of Use in addition to the CORA source-code license.

Citation

bibtex

@inproceedings{woo2026cora,
  title={Overlap-Consistent View Decomposition for Adapting Vision--Language Models to 360-Degree Panoramas},
  author={Woo, Seungwoo and Jung, Daewon and Youm, Sekyoung},
  booktitle={European Conference on Computer Vision},
  year={2026}
}

wfwefw/cora-gemma3-4b-vicreg

CORA Gemma3-4B VICReg 50%

Evaluation

Usage

Limitations and license

Citation

wfwefw/
cora-gemma3-4b-vicreg