LeNet-5 Architecture

Implementation of the LeNet-5 deep neural network model for the MNIST dataset using tensorflow.

The code can be found in the LeNet jupyter notebook.

Input

The LeNet architecture accepts a 32x32xC image as input, where C is the number of color channels. Since MNIST images are grayscale, C is 1 in this case.

Model Architecture

Layer 1: Input is 32x32x1.

• Convolutional with 5x5x6 Filter, output shape is 28x28x6.

• A Relu activation function is applied.

• A max pooling 2x2 with a stride of 2 is the output of the first layer (shape 14x14x6).

Layer 2: Input is 14x14x6.

• Convolutional with 5x5x16 Filter, output shape is 10x10x16.

• A Relu activation function is applied.

• A max pooling 2x2 with a stride of 2 is the output of the second layer (shape 5x5x16).

• Flatten. the output (400).

Layer 3:

• Fully Connected. (400, 120)
• A Relu activation function is applied.

Layer 4:

• Fully Connected. (120, 84)
• A Relu activation function is applied.

Layer 5:

• Fully Connected (Logits). (84, 10)