A model is implemented for the classification of CIFAR-10 dataset images using a deep-learning programming framework. This data set consists of 60,000 color images with dimensions of 32x32 (including 50,000 training images and 10,000 test images). The images of this data set are divided into ten categories.
Convolutional network was created via following architecture:
- Convolutional algorithm: kernel size: 3x3, (number of output channels: 7 and activation function: ReLU)
- Convolutional algorithm: kernel size: 3x3, (number of output channels: 9 and activation function: ReLU)
- Maximum Pooling Layer: Kernel size: 2x2
- Drop_Out Layer with a 30% Probability
- Linear Layer with ten (number of groups) output neurons
This network was trained with the Adam optimizer, cross-categorical entropy error function, and a learning rate of 100% (with a batch size of 32).