Resize Mask RCNN transform

Hi there,

I resized the MasK RCNN transform as follow:

 res_ = GeneralizedRCNNTransform(224, 224, [0.485, 0.456, 0.406],
                                        [0.229, 0.224, 0.225])

I am using weights from a model I trained for the backbone. I was wondering if this is correct or if it will impact the COCO metrics. Any help is highly regarded.