MNIST-handwritten digit classification

This is a lab of python in deep learning. we want to generate a multilayer perceptron for handwritten digit classification from scratch: Using only NumPy and matplotlib libraries. This repo based on the Intelligent Architectures(5LIL0) of TU/e.

这是一个深度学习中的python实验。我想从头开始写一个用于手写数字分类的多层感知器。Part 1部分只使用NumPy和matplotlib库，Part 2使用Pytorch库。本仓库基于TU/e的智能架构(5LIL0)课程

setup of jupyter notebok 开始使用jupyter notebook

English Verision

• Install Anacoda
• open Anacoda Powershell Prompt and input conda init --all
• Open Windows PowerShell as an administrator (win + x) and enter set-executionpolicy remotesigned.
• Open folder of this repo notebook, and Right-click to open in terminal.
• Input jupyter notebook, then you can do the lab.

Or

• Install Anacoda
• open Anacoda Powershell Prompt
• Use D: to go to the disk where your folder is located(there I go to D disk)
• Input jupyter notebook, then you can choose folder where lab is located.

中文版Chinese Verision

• 安装 Anacoda
• 打开 Anacoda Prompt 然后输入 conda init --all
• 以管理员身份打开 PowerShell (win + x) 然后输入 set-executionpolicy remotesigned.
• 在 notebook文件夹下右键从终端中打开。
• 输入 jupyter notebook即可开始实验。

或者

• 安装 Anacoda
• 打开 Anacoda Powershell Prompt
• 使用命令 D: 切换当前磁盘，这里是进入D盘
• 直接输入 jupyter notebook, 然后在弹出的浏览器中选择lab文件所在位置开始实验。