Add support for image batch processing

[zhliuworks]

May I inquire if the author has implemented batch processing for multiple images? It appears that the current detector's __call__ function only allows for inference on a single image.

class CannyDetector:
    def __call__(self, input_image=None, low_threshold=100, high_threshold=200, detect_resolution=512, image_resolution=512, output_type=None, **kwargs):

[patrickvonplaten]

I have not looked into batch processing inputs here

[mehmetsat]

for video processing it would be great to add batch processing. Because it takes too long to process one by one