# SigLIP Classification Eval

- Model: /hub_data3/seohyun/outputs/siglip_ecva

- Test repo: happy8825/siglip_test (split: train)

- Samples visualized: 50 of 924

## Metrics

- accuracy: 0.5736

- f1: 0.5321

- total: 924, correct: 530


## Confusion Matrix

![](confusion_matrix.png)