Research-on-Transfer-Learning-based-on-AlexNet-applied-in-Image-Classification

Transfer learning is a machine learning method which is widely used in the field of deep learning and image classification. The idea of transfer learning is to use the knowledge learned from one task to better solve another task. This paper discusses the application of transfer learning in deep learning image classification, and studies the impact of different learning methods and learning rates on the performance of the model, and then verifies the improvement of transfer learning on the performance of the model through experiments. This research implements image classification of transfer learning based on AlexNet, and explores the impact of different learning rates on model performance. In the experiment, pre training was conducted on the ImageNet dataset, and new learning and fine-tuning were performed on the CIFAR-10 dataset. The experimental results show that transfer learning can significantly improve the performance of the model, accelerate the convergence speed, and improve the accuracy of image classification. When the learning rate is 0.001, the test accuracy rate of transfer learning is 15 percentage points higher than that of new learning. With a learning rate of 0.01, transfer learning achieved the highest test accuracy of 92.97%.