Tensorflow Based Cnn Model

Abstract

Google's open source deep learning and machine learning framework is TensorFlow, which is flexible and convenient  to make the modern mainstream deep learning model. The advantage of a convolutional neural network over other deep learning models is the powerful feature extraction capabilities of convolutional blocks. A convolutional neural network model with two convolution layers was developed using the TensorFlow platform. The MNIST data collection was used to train and test the model. the test accuracy rate could reach 99.25 percent, compared to 98.69 percent with only one-convolution. Layer model, demonstrating that the two-convolution-layers convolutional natural network model has a stronger ability of feature extraction and classification decision-making. The advantage of a convolutional neural  network over other deep learning models is the powerful feature extraction capabilities of convolutional blocks

The,advantage of a convolutional neural network over other deep learning models is thepowerful feature extraction capabilities of convolutional blocks.

A convolutional neural network model with two convolution layers was developed using the TensorFlow platform. The MNIST data collection was used to train and test the model. The test accuracy rate was 99.15 percent, compared to 98.69 percent with a single-convolution-layer model, demonstrating that the two-convolution-layers convolutional neural network model has a stronger ability of feature extraction and classification decision-making.