Issue with training on user data about rt-detr HOT 3 CLOSED

Please use this torchvision version. It has ConvertDtype. This pr remove ConvertDtype pytorch/vision#7759

torch                              2.0.1
torchvision                        0.15.2

from rt-detr.

@aiexperts-andreas could you please share with us how you managed to solve the KeyError:'_pymodule'?

from rt-detr.

Changed versions and got this now?

python tools/train.py -c configs/rtdetr/rtdetr_r50vd_6x_coco.yml

Not init distributed mode.
Start training
Load PResNet50 state_dict
Initial lr: [1e-05, 0.0001, 0.0001, 0.0001]
Traceback (most recent call last):
File "C:\Users\copyc\AiExperts\Eide\RT-DETR\rtdetr_pytorch\tools\train.py", line 38, in
main(args)
File "C:\Users\copyc\AiExperts\Eide\RT-DETR\rtdetr_pytorch\tools\train.py", line 24, in main
solver.fit()
File "C:\Users\copyc\AiExperts\Eide\RT-DETR\rtdetr_pytorch\tools..\src\solver\det_solver.py", line 21, in fit
self.train()
File "C:\Users\copyc\AiExperts\Eide\RT-DETR\rtdetr_pytorch\tools..\src\solver\solver.py", line 48, in train
self.train_dataloader = dist.warp_loader(self.cfg.train_dataloader,
^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\copyc\AiExperts\Eide\RT-DETR\rtdetr_pytorch\tools..\src\core\yaml_config.py", line 83, in train_dataloader
self._train_dataloader = create('train_dataloader')
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\copyc\AiExperts\Eide\RT-DETR\rtdetr_pytorch\tools..\src\core\yaml_utils.py", line 95, in create
cls = getattr(cfg['_pymodule'], name)
~~~^^^^^^^^^^^^^
KeyError: '_pymodule'

from rt-detr.