Reshaping tensors

In many machine learning problems, you will need to reshape your input data. If, for instance, you loaded a 9-pixel, grayscale image of the letter H, it might have the following 2-dimensional representation:

\begin{equation} \begin{bmatrix} 255 & 0 & 255 \\ 255 & 255 & 255 \\ 255 & 0 & 255 \end{bmatrix} \end{equation}

Some models are designed to use image data as an input. Many, however, require you to transform the data into a vector. In this exercise, we will use the reshape() operation to practice transforming tensors. Note that ones() and reshape() operations have been imported.

1
Reshape the grayscale image into a 256x1 image_vector and a 4x4x4x4 image_tensor.
2
Adapt image by adding three color channels and then adjust image_vector and image_tensor accordingly.

Exercise

Reshaping tensors

Instructions 1/2