[BUG] - Target in Object Detection Finetuning Tutorial is not beeing recognised as dict! #3137
Description
Add Link
https://pytorch.org/tutorials/intermediate/torchvision_tutorial.html
Describe the bug
I've taken the Code from the tutorial. I work with the first Alternative and was at the point 'testing forward() method (Optional)'. Runiing train.py, got:
Traceback (most recent call last):
File "/home/me/PycharmProjects2/00CNNundRNN/me/14PedestrianDatasetFirst/train.py", line 19, in
output = model(images, targets) # Returns losses and detections
^^^^^^^^^^^^^^^^^^^^^^
File "/home/me/miniconda3/envs/venv00/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1553, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/me/miniconda3/envs/venv00/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1562, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/me/miniconda3/envs/venv00/lib/python3.11/site-packages/torchvision/models/detection/generalized_rcnn.py", line 65, in forward
boxes = torch.stack([item['boxes'] for item in target]) # boxes = target["boxes"]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/me/miniconda3/envs/venv00/lib/python3.11/site-packages/torchvision/models/detection/generalized_rcnn.py", line 65, in
boxes = torch.stack([item['boxes'] for item in target]) # boxes = target["boxes"]
~~~~^^^^^^^^^
TypeError: string indices must be integers, not 'str'
Describe your environment
Ubuntu 22.4 LTS,
everthing is updated
Pytorch 2.4.0+cu121