[Python] Visualize Filters of CNN using Tensor Flow

Im working on Visualizing Filters of a CNN using TensorFlow - Guided Project in coursera. I encountered the error when trying to call the plot_image function. I defined the size of the image error screenshot as (96,96,3) but it takes as (1,96,96,3)

def create_image():

im = tf.random.uniform((96,96,3),minval=0.5,maxval=0.5)

return im

def plot_image(image,title ='random'):

image =image - tf.math.reduce_min(image),

image =image / tf.math.reduce_max(image)

plt.imshow(image)

plt.xticks([])

plt.yticks([])

plt.title(title)

plt.imshow(image)

plt.show()

image = create_image()
plot_image(image)


I tried using

image_data_channels_first = np.random.rand(3, 96, 96)
image_data_channels_last = np.moveaxis(image_data_channels_first, 0, -1)


Squeeze method also i tried, it didnt work. Im encountering same error when calling the function.

Continue reading...