metadata

inference: false
license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
model-index:
  - name: AI-generated_images_detector
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: validation
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.9735697557711609

AI-generated_images_detector

This model achieves the following results on the evaluation set:

Loss: 0.0987
Accuracy: 0.9736

To utilize this model


from PIL import Image
from transformers import pipeline
classifier = pipeline("image-classification", model="NYUAD-ComNets/NYUAD_AI-generated_images_detector")
image=Image.open("path_to_image")
pred=classifier(image)
print(pred)

Training and evaluation data

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.0431	0.55	100	0.1672	0.9568
0.0139	1.1	200	0.2338	0.9398
0.0201	1.66	300	0.1291	0.9655
0.0023	2.21	400	0.1147	0.9709
0.0033	2.76	500	0.0987	0.9736

BibTeX entry and citation info

@article{aldahoul2024detecting,
  title={Detecting AI-Generated Images Using Vision Transformers: A Robust Approach for Safeguarding Visual Media Integrity},
  author={AlDahoul, Nouar and Zaki, Yasir},
  journal={Available at SSRN},
  year={2024}
}

@misc{ComNets,
      url={https://huggingface.co/NYUAD-ComNets/NYUAD_AI-generated_images_detector](https://huggingface.co/NYUAD-ComNets/NYUAD_AI-generated_images_detector)},
      title={NYUAD_AI-generated_images_detector},
      author={Nouar AlDahoul, Yasir Zaki}
}