marrz/lenet-keras

Implementation of LeNet-5 with keras

This repo is implemented in Tensorflow

CNN Architecture

Architecture	Framework	Link
LeNet-5	keras(using tensorflow)	HERE
AlexNet	Tensorflow	-
GoogLeNet	-	-
ResNet	-	-

LeNet-5 with keras

LeNet-5 Architecture is perhaps the most widely known CNN architecture. This architecture was made by Yann LeCun. This architecture consist of the following layers.

LeNet-5 with keras

LeNet-5 Architecture is perhaps the most widely known CNN architecture. This architecture was made by Yann LeCun. This architecture consist of the following layers.

Layer	Feature Map	Size	Kernel Size	Stride	Activation
FC	-	10	-	-	softmax
FC	-	84	-	-	tanh
Conv2d	120	1x1	5x5	1	tanh
Average Pooling	16	5x5	2x2	2	tanh
Conv2d	16	10x10	5x5	1	tanh
Average Pooling	6	14x14	2x2	2	tanh
Conv2d	6	28x28	5x5	1	tanh
Input	1	32x32	-	-	-

• MNIST Image is 28x28 pixel, but it was 32x32 pixel because of applied zero padding.

code

data.py is class for MNIST Data preprocessing
lenet.py is LeNet model implemented with keras
plot.py is function for loss visualization
If you want to test, run main.py