ClassiFly is a deep learning-based image classification system utilizing EfficientNetB0 for accurate and efficient categorization of images. It supports data augmentation, transfer learning, and fine-tuning for optimal performance.
- Supports multiple classification algorithms with EfficientNetB0
- Image preprocessing and real-time data augmentation
- Model training with early stopping and learning rate reduction
- Fine-tuning support for improved accuracy
- Generates predictions with class label mapping
- Saves best model for reuse and future inference
- Launch the script to train the model:
python script.py
- Enable fine-tuning if needed:
python script.py --fine_tune
- View predictions after training:
cat submission.csv
- Python
- TensorFlow & Keras
- EfficientNetB0 (Transfer Learning)
- Scikit-learn
- Pandas & NumPy (for data processing)
- Matplotlib & Seaborn (for data visualization)