Question about using transforms with RCNN


I am currently using a pretrained RCNN ( torchvision.models.detection.fasterrcnn_resnet50_fpn ) and I noticed that i could not use the transforms functions defined in torchvision.transforms — Torchvision master documentation but only those defined in vision/ at master · pytorch/vision · GitHub otherwise I get an error. Do you know why the official functions does not work and do you know any way to make them functionnal because currently I can only use the 3 or 4 functions from the github.