MNIST Handwriting Recognition

This repository contains two Python scripts for creating and training a Convolutional Neural Network (CNN) to recognize handwritten digits using the MNIST dataset. It is designed as a beginner-friendly project, and the following files are included:

本存儲庫包含兩個Python腳本，用於創建和訓練卷積神經網絡（CNN）以使用MNIST數據集識別手寫數字。這是一個適合初學者的項目，包括以下文件：

1. mnist.py: This script provides a simple graphical user interface (GUI) for drawing handwritten digits and using the trained model to predict them. It uses the TensorFlow library for deep learning and tkinter for the GUI.

   mnist.py腳本提供了一個簡單的圖形用戶界面（GUI），用於繪製手寫數字並使用已訓練的模型進行預測。它使用TensorFlow庫進行深度學習，使用tkinter進行GUI。

2. train.py: This script trains a CNN model on the MNIST dataset and saves the trained model as mnist_model.h5. You can use this model with the mnist.py script for digit recognition.

   train.py腳本在MNIST數據集上訓練了一個CNN模型，並將訓練好的模型保存為mnist_model.h5。您可以將此模型與mnist.py腳本一起用於數字識別。

Requirements (要求)

To run the mnist.py and train.py scripts, you need to have the following Python libraries installed:

要運行mnist.py和train.py腳本，您需要安裝以下Python庫：

• TensorFlow
• tkinter
• numpy
• PIL (Pillow)
• matplotlib

You can install these libraries using pip with the following command:

您可以使用以下命令使用pip安裝這些庫：

pip install tensorflow tkinter numpy pillow matplotlib