Can I create video from a list of images using Torchvision CUDA?

Is it possible to create videos from image list using Cuda version of torchvision. I can do it using opencv but I want to utilize my rtx gpu. I didn’t find any code example resources. Can anyone help ?
Thanks.