The function get_batch_generator just work correctly if we use batch_size=1. I think the module detection.py L:727 must be X = np.array([compute_input(image) for image in images], dtype="float32") instead of X = compute_input(images) if we want to use batch_size > 1.
The function
get_batch_generator
just work correctly if we use batch_size=1. I think the moduledetection.py L:727
must beX = np.array([compute_input(image) for image in images], dtype="float32")
instead ofX = compute_input(images)
if we want to use batch_size > 1.