Find 10 suitable colour object images either online or from your own photos, crop and/or resize them to the right size (32 x 32 pixels), preprocess them using the transform function, feed the images to the network that we trained in the programming lab, and get the predicted labels for each image. Feel free to play around with images that come from the 10 trained categories versus images from other categories.
Output: Upload the following file to codePost: ps11.pdf. The file should display the 10 images. Also include the predicted category label for each image.