GitHub - Leozhou0724/Manually_Implemented_LeNet5: manually implement a LeNet-5 convolutional NN using Numpy (no automatic gradient)

Leozhou0724 / Manually_Implemented_LeNet5 Public

Notifications You must be signed in to change notification settings
Fork 0
Star 1

manually implement a LeNet-5 convolutional NN using Numpy (no automatic gradient)

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 1 Commit
LeNet.py		LeNet.py
MLProject_report.pdf		MLProject_report.pdf
README.txt		README.txt
Result.txt		Result.txt
layers.py		layers.py

Repository files navigation

Machine Learning Final Project
Yuhang Zhou
yz853
5/8/2019

Type-A Project
It's a manually implemented LeNet-5 Neural Network for the MNIST dataset based on numpy
The MNIST dataset is already included in the floder which is the "mnist.pkl" file

All the layer class are written in the "layer.py" file and 
the LeNet-5 model is in the "LeNet.py" file

Run the "LeNet.py" to run the training program, 
it's very very slow in my laptopm, so I just run 
the training process for a while and watch the loss, details in the "Result.txt" file