Implementasi Deep Learning EfficientNet dengan Fine Tuning Pada Klasifikasi Ras Kucing

Cats have various types with different characteristics. This study aims to classify various types of cat breeds using the EfficientNet architecture model, which is known to have high computational efficiency and superior performance in image classification. Fine-tuning, data augmentation, and dropout techniques are applied to improve model performance. In this study, the EfficientNetB0 model is used for cat breed classification, by comparing the performance between the base model without additional methods and the optimized model. The EfficientNetB0 model that applies data augmentation, fine-tuning, and dropout successfully achieves 97% accuracy on training data, 97% on validation data, and 91% on test
data.