Beginner using Tensorflow. Started reading the documentation and checking the examples provided in the website:
https://www.tensorflow.org/api_docs/python/tf/keras/layers/Conv2D
https://www.tensorflow.org/tutorials/images/cnn

From a mathematical point of view I understand the convolution, but it is not clear to me how it is implemented using Conv2D.
From the second link, the convolution is implemented as:

layers.Conv2D(32, (3, 3), activation='relu', input_shape=(32, 32, 3))

From the convolution point of view, how is the filter defined by only one number? How does the convolution works in Tensorflow?

Thanks,
Gus