Deepfake Image Detector (Celeb-DF-v1 Version 3)

Model Performance

  • Test Accuracy: 96.43%
  • Best Validation Accuracy: 96.41%
  • Best Epoch: 3
  • Planned Epochs: 6
  • Actual Epoches Trained: 4 (early stopping applied)

Dataset

  • Training: 2,452 images
  • Validation: 306 images
  • Test: 308 images

Training Details

Training was stopped early at epoch 4 due to early stopping criteria being met. The best model was achieved at epoch 3 with validation accuracy of 96.41%.

Usage

from transformers import ViTForImageClassification, ViTFeatureExtractor

model = ViTForImageClassification.from_pretrained('shivani1511/deepfake-image-detector-celebdf-v1-v3')
feature_extractor = ViTFeatureExtractor.from_pretrained('shivani1511/deepfake-image-detector-celebdf-v1-v3')
Downloads last month
6
Safetensors
Model size
85.8M params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support