sunnyguang/digit-recognizer

Digit Recognizer

Introduction:

Digit Recognizer is the basic project for Deep Learning. It exsits 42,000 training data. Because it is the Kaggle entry level project. The dataset is also not complex, and I do not need to improve the dataset again. I need to convert 28*28 image to the matrix for training, and I need to build the model to train. I will try to use python and Deep learning (CNN) to build the algorithm that could detect the objects of the images.

Platform

Windows 10 system Kernals in Kaggle.com

Why do I choose this project?

I want to learn python and machine learning through the project, to have better understanding for machine learning and expert to use python after I finish the project.

Instructions:

• This program is a Convolutional Neural Network design by python to identify digit recognizer. It loads the input dataset and perform by running the code, does not require the test code.
• Runing the code need to using jupyter notebook. It also need some python package such as matplotlib and keras.
• Step one: git clone the code to the local PC, using "git clone + repo_link"
• Step two: using juypter notebook to open the Python Project.ipynb file.
• Step three: use pip install + package_name to install the required packages
• Step four: run the code
• Result picture shows in "_result_files" folder