python vizualize.py

python Runner.py 0 1

The first arg is the value of 'lambda' the regularization param
The second is the dropout keep probability. Both should be between 0 and 1, though no validation has been implemented.
0 and 1 meeans, lambda is 0 => no regularization,
the keep probability is 1 => no dropout since all the nodes are kept during training.

Train/Test accuracy: 0.9479 / 0.915

python Runner.py 0.7 1

Train/Test accuracy: 0.938 / 0.93

python Runner.py 0 0.86

Train/Test accuracy: 0.929 / 0.95