About This Project
Trained transfer learning models (EfficientNetV2B0) with TensorFlow on Food101 dataset subsets. Compared feature extraction vs fine-tuning with and without data augmentation. Implemented checkpointing for best model saving and achieved robust food image classification.
Key Features
- Transfer learning experiments
- Data augmentation pipeline
- Model checkpointing
- Performance comparison
