RandomCrop in PyTorch (1)

The code demonstrates the `RandomCrop` function from `torchvision.transforms.v2` for image manipulation. `RandomCrop` allows random cropping of images with customizable parameters. These parameters include size, padding, `pad_if_needed` flag, fill color, and padding mode. The size parameter determines the output image dimensions, accepting integers or lists/tuples. Padding adds extra pixels around the image before cropping; negative padding is allowed. `pad_if_needed` handles cases where the crop size exceeds the image dimensions. The fill parameter specifies the color used for padding, and padding mode defines how padding is applied. The code uses the OxfordIIITPet dataset to showcase `RandomCrop` with various parameter settings, illustrating the effects of size, padding, fill color, and padding mode on the resulting cropped images. Multiple examples are provided, demonstrating different combinations of these parameters. The code also includes visualization functions to display the original and transformed images. Two functions, `show_images1` and `show_images2`, are used for image display. The second function offers additional control over cropping parameters.