Fine-tuned EfficientNet and MobileNetV2 Models for Intel Images Classification

This study looks at using two types of deep learning models, MobileNetV2 and EfficientNet, for sorting pictures in Intel's image collections. The study looks at how well it works by testing with different numbers for learning rate; these are 0.01 and 0.09 faster than the last ones, seen as goals to reach success high in speed needed fast. The study is about checking how good the models' guesses are since it shows how well they work. The MobileNetV2 model did very well, reaching the best accuracy of 99.67% at point 50 in training sessions. Instead, the EfficientNet showed a very good accuracy of 92.11% at the same time period. The side-by-side look shows that MobileNetV2 does better than EfficientNet in image sorting jobs, winning in this special situation. Changing the learning rate helps us understand how sensitive our models are to fine-tuning. This can help make training smoother and better. The results show how important it is to choose the right model and adjust its settings to get better accuracy when classifying images. This could have many uses in different areas, like seeing what objects are shown on images or finding out patterns from them.