An open API service indexing awesome lists of open source software.

https://github.com/samo135/interactive-mnist

Web app using an MNIST trained neural network to classify images of numbers drawn by the user. An interactive way to play around with a simple neural network.
https://github.com/samo135/interactive-mnist

css docker fastapi google-cloud-platform html mnist neural-network python pytorch rest-api

Last synced: about 1 year ago
JSON representation

Web app using an MNIST trained neural network to classify images of numbers drawn by the user. An interactive way to play around with a simple neural network.

Awesome Lists containing this project

README

          

# interactive-mnist

[Link to website](https://interactive-mnist-128295820748.europe-west9.run.app/)

Draw a number (0 - 9) in the box. When you click submit, the image will be passed through a simple convolutional neural network trained on the MNIST dataset which will try to accurately classify the number you have drawn. The model scored an accuracy of ~98% on the validation dataset. To the right of the box you will see what probabilities the model has assigned each number 0 - 9.