English | 简体中文
This project is an assignment for my undergraduate machine learning course. The project content includes:
- Use PyTorch to build LeNet5 model;
- Training LeNet5 model on MNIST dataset;
- Make a handwritten digit recognition applet.
After running the main program, a whiteboard will appear. Press and hold the left mouse button to write on the whiteboard. After releasing the left mouse button, the program will automatically recognize the numbers on the whiteboard and display the recognition results below the whiteboard. Press the right button of the mouse to clear the whiteboard.
Digits
├── docs # Project Documents
├── demo.gif # Demo
└── README_cn.md # Chinese Description Document
├── model # LeNet Package
└── __init__.py # LeNet Model
├── install.sh # Install program
├── LICENSE # LICENSE
├── main.py # Handwriting Digit Recognition Application
├── README.md # English Description Document
├── requirements.txt # List of Requirements
└── train.py # Training Program
git clone https://github.com/Yue-0/Digits.git
cd ./Digits
sh install.sh
python train.py --lr 0.1 --epoch 10 --batch 64
Due to the simple network structure, this project uses CPU for training and inference.
python main.py