Transposed Convolution Quiz Solution

One possible answer is using tf.layers.conv2d_transpose(x, 3, (2, 2), (2, 2)) to upsample.

• The second argument 3 is the number of kernels/output channels.
• The third argument is the kernel size, (2, 2). Note that the kernel size could also be (1, 1) and the output shape would be the same. However, if it were changed to (3, 3) note the shape would be (9, 9), at least with 'VALID' padding.
• The fourth argument, the number of strides, is how we get from a height and width from (4, 4) to (8, 8). If this were a regular convolution the output height and width would be (2, 2).

Now that you've learned how to use transposed convolution, let's learn about the third technique in FCNs.