I’m using torchvision.models.detection.maskrcnn_resnet50_fpn to detect object, and get detection output (type=List[Dict])，but output[image_id]["labels"] returns int, so how can I get the label in str ?
I found that #class of default pretrained MaskRCNN in COCO_train_2017 is 91, but actually the #class of the dataset is 80, and plus the background is 81, it’s weird.