Need your help about Focs model train

[bovane]

@delmalih I'm training the focs model with google colab ,but it can't work when I follow the instructions. Colab configuration are as follows: torch.version.cuda----- Cuda compilation tools, release 10.1, V10.1.243 print(torch.version) ----- 1.4.0 print(torch.cuda.is_available()) print(torch.backends.cudnn.enabled)-------- True True

the details for error--------------- Traceback (most recent call last): File "train.py", line 180, in main() File "train.py", line 173, in main model = train(cfg, args.local_rank, args.distributed) File "train.py", line 79, in train arguments File "/content/drive/Shared drives/BoVane/dependency/FCOS/fcos_core/engine/trainer.py", line 69, in do_train loss_dict = model(images, targets) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 532, in call result = self.forward(*input, kwargs) File "/content/drive/Shared drives/BoVane/dependency/FCOS/fcos_core/modeling/detector/generalized_rcnn.py", line 50, in forward proposals, proposal_losses = self.rpn(images, features, targets) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 532, in call result = self.forward(*input, *kwargs) File "/content/drive/Shared drives/BoVane/dependency/FCOS/fcos_core/modeling/rpn/fcos/fcos.py", line 159, in forward centerness, targets File "/content/drive/Shared drives/BoVane/dependency/FCOS/fcos_core/modeling/rpn/fcos/fcos.py", line 169, in _forward_train locations, box_cls, box_regression, centerness, targets File "/content/drive/Shared drives/BoVane/dependency/FCOS/fcos_core/modeling/rpn/fcos/loss.py", line 257, in call labels_flatten.int() File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 532, in call result = self.forward(input, kwargs) File "/content/drive/Shared drives/BoVane/dependency/FCOS/fcos_core/layers/sigmoid_focal_loss.py", line 68, in forward loss = loss_func(logits, targets, self.gamma, self.alpha) File "/content/drive/Shared drives/BoVane/dependency/FCOS/fcos_core/layers/sigmoid_focal_loss.py", line 19, in forward logits, targets, num_classes, gamma, alpha RuntimeError: Not compiled with GPU support (SigmoidFocalLoss_forward at fcos_core/csrc/SigmoidFocalLoss.h:20)

Please help me !!! Thank you 🙏 😊

[delmalih]

For the FCOS obj. detector training, I'd recommend you to use the official repo

[bovane]

Thank you David，I have send a email to you，could you reply me ? please ! @delmalih

[bovane]

How can I get such output ？！！

[bovane]

@delmalih I am so sorry for my interruption，Please forgive me for interrupting. I have trained the faster-rcnn,fcos,and retinanet,but I can't get the output as your picture,please help me! I have sent the email to you,I introduced more information in the email,could you reply me? I need whole codes and documents about this project.I promise I will only use it for personal learning ！If you can help me, I will pray for you all day long！Thank you ! 🙏 😊 ☺️ David