Make Your First GAN Using PyTorch/

...

/

The Fashion MNIST CNN Classifier Performance

The Fashion MNIST CNN Classifier Performance

Learn about training the Fashion MNIST classifier and testing its performance.

We'll cover the following...

Training the CNN classifier

Let’s train this CNN and test its performance in the same way as our fully connected Fashion MNIST classifier.

Press + to interact
%%time 
# create neural network
C = Classifier()
# train network on MNIST data set
epochs = 3
for i in range(epochs):
    print('training epoch', i+1, "of", epochs)
    for label, image_data_tensor, target_tensor in fmnist_dataset:
        C.train(image_data_tensor.view(1, 1, 28, 28), target_tensor)
        pass
    pass

Let’s observe the loss charts.

The loss chart looks very similar to the chart for the fully connected network. The loss falls rapidly towards zero and the bulk of the loss values remain close to zero.

Classifier

...