buxtcodes's picture
Strip model card to minimal -- full docs live in the GitHub repo
9477e06 verified
|
Raw
History Blame Contribute Delete
620 Bytes
---
license: other
tags:
- image-classification
- ai-generated-image-detection
- dinov3
pipeline_tag: image-classification
---
# mLoRC
AI-generated-image detector (DINOv3 ViT-H+/16 + [LoRC](https://arxiv.org/abs/2608.20882)).
Code, results, and docs: [Buxt-Codes/AIGI-mLoRC](https://github.com/Buxt-Codes/AIGI-mLoRC).
```python
from mlorc import ModulatedLoRC
model = ModulatedLoRC.from_pretrained(hf_token="hf_...")
model.predict_image("photo.jpg")
```
## Files
- `mlorc-full.pt` — full checkpoint, LoRA merged into the backbone weights.
- `dinov3_config.json` — DINOv3 architecture config (no weights).