improve list_data_collate error messages #1833
Description
Is your feature request related to a problem? Please describe.
list_data_collate or the pytorch's default_collate may raise RuntimeError or TypeError, but the error messages are difficult to understand. This is a feature request to improve the error message especially for collating dictionary data.
Current example error messages:
type error (the list of dict data has mixed torch tensor and numpy ndarray, but we couldn't locate the key):
Traceback (most recent call last):
File "MONAI/monai/data/utils.py", line 257, in list_data_collate
return default_collate(data)
File "/usr/local/anaconda3/envs/py37/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 73, in default_collate
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/usr/local/anaconda3/envs/py37/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 73, in <dictcomp>
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/usr/local/anaconda3/envs/py37/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 83, in default_collate
return [default_collate(samples) for samples in transposed]
File "/usr/local/anaconda3/envs/py37/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 83, in <listcomp>
return [default_collate(samples) for samples in transposed]
File "/usr/local/anaconda3/envs/py37/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 73, in default_collate
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/usr/local/anaconda3/envs/py37/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 73, in <dictcomp>
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/usr/local/anaconda3/envs/py37/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 73, in default_collate
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/usr/local/anaconda3/envs/py37/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 73, in <dictcomp>
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/usr/local/anaconda3/envs/py37/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 55, in default_collate
return torch.stack(batch, 0, out=out)
TypeError: expected Tensor as element 4 in argument 0, but got numpy.ndarray
RuntimeError error (one key-val in list of data dict has inconsistent shapes but we couldn't locate the key):
RuntimeError: Caught RuntimeError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "/__w/MONAI/MONAI/monai/data/utils.py", line 257, in list_data_collate
return default_collate(data)
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 73, in default_collate
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 73, in <dictcomp>
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 83, in default_collate
return [default_collate(samples) for samples in transposed]
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 83, in <listcomp>
return [default_collate(samples) for samples in transposed]
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 73, in default_collate
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 73, in <dictcomp>
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 73, in default_collate
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 73, in <dictcomp>
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 63, in default_collate
return default_collate([torch.as_tensor(b) for b in batch])
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 55, in default_collate
return torch.stack(batch, 0, out=out)
RuntimeError: stack expects each tensor to be equal size, but got [3, 3] at entry 0 and [4, 4] at entry 4