Python Deep Learning Projects

Going deeper

The convolution classifier with max pooling and dropout seems to be the best classifier so far. However, we also noticed that there was a slight amount of overfitting on the train data.

Let's build a deeper model to see if we can create a classifier that is more accurate than the other models we have trained so far, and see if we can get it to reach an even better minima.

We will build a deeper model by adding two more convolution layers to our best model so far:

• The first layer is a convolution 2-D layer with 32 filters of size 3*3 with activation as relu, followed by downsampling with max pooling of size 2*2, followed by Dropout as the regularizer
• The second layer is a convolution 2-D layer with 64 filters of size 3*3 with ...