Create a network of dense layers classifying the dataset fashion MNIST. Get inspired by the Colab notebook classifying MNIST digits in Chapter 2 of Deep Learning with Python. Make some improvements:

• Set aside 5000 samples from the training set as a validation set (look at Chapter 4, the IMDB example, to see how)
• Use them to plot the accurracy as in Chapter 4
• Experiment with more layers and their sizes
• Try to use regularization by using additional parameters of the layers.Dense method, visit Keras documentation to see how
• You should obtain validation accurracy over 0.9 

Submit a sharing link to your Colab notebook (allow access to all with the link), and an image of your accurracy plot.