I wanted to use this line of code to wrap my image data and labels into one tensor, but the variable training_img_set is in the form of a list of images which gives an error. How do I convert it to a tensor format so I can use it as a parameter for the function?