hidden_states_list error

[mrsahabu]

Hi, while training I am getting this error.

(lstm) D:\LSTM Object Detection>python train_ssd.py
2019-12-23 16:37:41,477 - root - INFO - cuda
2019-12-23 16:37:41,478 - root - INFO - Use GPU
2019-12-23 16:37:41,478 - root - INFO - Namespace(balance_data=False, base_net=None, base_net_lr=None, batch_size=2, checkpoint_folder='models/', dataset_type='voc', datasets='./data', debug_steps=1, extra_layers_lr=None, freeze_base_net=False, freeze_net=False, gamma=0.1, lr=0.001, mb2_width_mult=1.0, milestones='80,100', momentum=0.9, net='vgg16-ssd', num_epochs=200, num_workers=0, pretrained_ssd=None, resume=None, scheduler='multi-step', t_max=120, use_cuda=True, validation_dataset=None, validation_epochs=5, weight_decay=0.0005)
2019-12-23 16:37:41,480 - root - INFO - Prepare training datasets.
2019-12-23 16:37:41,481 - root - INFO - No labels file, using default VOC classes.
2019-12-23 16:37:41,481 - root - INFO - No labels file, using default VOC classes.
2019-12-23 16:37:41,482 - root - INFO - No labels file, using default VOC classes.
2019-12-23 16:37:41,483 - root - INFO - No labels file, using default VOC classes.
2019-12-23 16:37:41,483 - root - INFO - No labels file, using default VOC classes.
2019-12-23 16:37:41,484 - root - INFO - No labels file, using default VOC classes.
2019-12-23 16:37:41,484 - root - INFO - Stored labels into file D:\LSTM Object Detection\labels.txt.
2019-12-23 16:37:41,485 - root - INFO - Train dataset size: 18
2019-12-23 16:37:41,485 - root - INFO - Prepare Validation datasets.
2019-12-23 16:37:41,485 - root - INFO - No labels file, using default VOC classes.
2019-12-23 16:37:41,485 - root - INFO - validation dataset size: 3
2019-12-23 16:37:41,485 - root - INFO - Build network.
2019-12-23 16:37:42,800 - root - INFO - Took 0.00 seconds to load the model.
2019-12-23 16:37:45,074 - root - INFO - Learning rate: 0.001, Base net learning rate: 0.001, Extra Layers learning rate: 0.001.
2019-12-23 16:37:53,236 - root - INFO - Uses MultiStepLR scheduler.
2019-12-23 16:37:53,238 - root - INFO - Start training from epoch 0.
Traceback (most recent call last):
  File "train_ssd.py", line 381, in <module>
    device=DEVICE, debug_steps=args.debug_steps, epoch=epoch)
  File "train_ssd.py", line 154, in train
    hidden_states_list, confidence, locations = net(video, hidden_states_list)
  File "C:\Users\abdul.ahad\AppData\Local\Continuum\anaconda3\envs\lstm\lib\site-packages\torch\nn\modules\module.py", line 489, in __call__
    result = self.forward(*input, **kwargs)
  File "D:\LSTM Object Detection\vision\ssd\ssd.py", line 148, in forward
    state = self.LSTM_list[i](y, hidden_states[i])
  File "C:\Users\abdul.ahad\AppData\Local\Continuum\anaconda3\envs\lstm\lib\site-packages\torch\nn\modules\module.py", line 489, in __call__
    result = self.forward(*input, **kwargs)
  File "D:\LSTM Object Detection\vision\nn\conv_lstm.py", line 41, in forward
    stacked_inputs = torch.cat((input_, prev_hidden), 1)
RuntimeError: Expected object of backend CUDA but got backend CPU for sequence element 1 in sequence argument at position #1 'tensors'

While this was working on CPU. I have changed hidden_states_list to tensor.zeros but then got TypeError: iteration over a 0-d tensor.
Versions pytorch = 1.0.0 Cuda = 9.2 python 3