CLRNet
CLRNet copied to clipboard
What pre-training model is adopted?
Do you mean the backbone? We use resnet18, resnet34, resnet101 and DLA34 to serve as our backbone and do following experiments.