A peek into how Neural Networks classify images from the MNIST dataset.
MNIST is the de facto "Hello World" of image processing.
- Run the Learner's Notebook(ipynb file) which will create the files necessary for visualization of a neural ntwork.
- Train the model by running
python train_model.py. 20 epochs, takes about 5 minutes on a CPU with average processing power. - model.h5 is created.
- Run the ml_server by executing
python ml_server.py. - With the flask server running, execute the virtual environment(Visualize) and start streamlit server using
run streamlit run app.py.
