Skip to content

Wowe-Peanut/Number-Identification-AI

Repository files navigation

Number-Identification-AI

Inspired and guided by 3b1b's video on how backpropagation machine learning works. All code, however, is original and my own. Able to acheive 93% testing accuracy with layer sizes of 784,16,16,10; batchsize of 32, learning rate of 0.1, and 10 epochs. Now, I wrote this several months ago and don't remember exactly how long it took to train but definetly something I ran overnight. My code is very inefficient as the main goal of the project was developing my understanding of how basic machine learning works, not making the best code possible :P

Training and Testing Data Used: http://yann.lecun.com/exdb/mnist/

3b1b Video (1/4): https://www.youtube.com/watch?v=aircAruvnKk&ab_channel=3Blue1Brown

Libraries Used: Numpy (Matrix Math), mnist (Importing Training and Testing Data), and Pickle (Storing Trained Weights and Baises for later)

About

Inspired and guided by 3b1b's video on how backpropagation machine learning works. All code, however, is original and my own.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages