• ax.flatten(): Transform n*m to 1*nm 1-D Array

    fig, ax = plt.subplots(nrows=2,ncols=2,sharex='all',sharey='all')
    ax = ax.flatten()  
 
    for i in range(4):
        img = image[i].reshape(28, 28)
        ax[i].imshow(img, cmap='Greys', interpolation='nearest')
        
    # ax[i] is available

• Without flatten()

    fig, ax = plt.subplots(nrows=2,ncols=2,sharex='all',sharey='all')  
    for i in range(4):
        img = image[i].reshape(28, 28)
        ax[0, 0].imshow(img, cmap='Greys', interpolation='nearest') 
        ax[0, 1].imshow(img, cmap='Greys', interpolation='nearest')
        ax[1, 0].imshow(img, cmap='Greys', interpolation='nearest')
        ax[1, 1].imshow(img, cmap='Greys', interpolation='nearest')
        
    # ax[i] is unavailable

• Convolutional Neural Networks

1. We often use Flatten, converting matrice to vectors.
2. After flattening, then feed the vectors to Fully Connected Layers.
3. CNN -> Pooling -> CNN -> Pooling...-> Flatten -> Fully Connected Layers -> Softmax -> Probabilities