pyrofolium/MNIST_tutorial

Sample tutorial for MNIST dataset.

Before you start make sure to check requirements.txt to make sure you have all of those packages installed. (Anaconda should take care of this for you)

Usage

First generate the model:

$ ./classify.py

This creates a 'model.pkl' file, which predict.py will open in order to predict user-drawn inputs. Run ./predict.py and draw a digit in the input window, which will then be scaled and normalized to conform (more or less) to the format of the training data. The predicted digit will be printed to the screen.

##Tutorial You should start your jupyter-notebook service and then open classify.ipynb in your web browser. You can then run the code and see some swet graphs to visualize the process.