I'm creating a dataloader with a dataset.
train_dl = torch.utils.data.DataLoader(train_ds, batch_size=8, shuffle=True)
The dataset contains 2 elements per index: a matrix and an array.
a, b = train_ds.__getitem__(0)
print(type(a))
print(a)
print(type(b))
print(b)
returns:
<class 'numpy.ndarray'>
[[ 44.9329 46.08967 44.9329 ... 99.2188 99.735664 99.17029 ]
[ 44.9329 44.9329 44.9329 ... 114.164474 114.292244 114.11395 ]
[ 44.9329 45.03071 44.9329 ... 114.57378 114.56599 114.49552 ]
...
[ 44.9329 44.9329 44.9329 ... 52.242996 50.12293 44.9329 ]
[ 44.9329 44.9329 44.9329 ... 44.9329 44.9329 44.9329 ]
[ 44.9329 44.9329 44.9329 ... 44.9329 44.9329 44.9329 ]]
<class 'numpy.ndarray'>
[-0.002963486 -0.003033393 0.00371422 2.02e-06 0.004402838 -0.002704915
0.003289625 -0.002551801 -0.003632823 -0.003408553 -0.002707387
0.00278949 0.000828761 0.000849513 0.003992096 -0.002692624 0.001183484
9.43e-05 0.003836168 2.24e-05 0.003944455 -0.001950883 -0.000877485
0.001734729 -0.003225849 -0.000537016 6.53e-05 -0.003643878 -0.002444321
0.002499692 0.001538219 0.002263657 0.003073046 0.004134932 -0.002500862
-0.001662471 0.002273667 0.00375025 0.001866289 -0.002027481 0.002197658
-0.002243473 0.000943156 -0.000643054 -0.003169563 -0.003424202
0.00118924 -0.003570424 0.002273526]
But when trying to iterate my dataloader with:
for i, data in enumerate(train_dl):
I get the error:
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
in
1 num_epochs=1 # Just for demo, adjust this higher.
----> 2 training(myModel, train_dl, num_epochs)
in training(model, train_dl, num_epochs)
18
19 # Repeat for each batch in the training set
---> 20 for i, data in enumerate(train_dl):
21 # Get the input features and target labels, and put them on the GPU
22 inputs, labels = data[0].to(device), data[1].to(device)
~\AppData\Roaming\Python\Python38\site-packages\torch\utils\data\dataloader.py in __next__(self)
631 # TODO(https://github.com/pytorch/pytorch/issues/76750)
632 self._reset() # type: ignore[call-arg]
--> 633 data = self._next_data()
634 self._num_yielded += 1
635 if self._dataset_kind == _DatasetKind.Iterable and \
~\AppData\Roaming\Python\Python38\site-packages\torch\utils\data\dataloader.py in _next_data(self)
675 def _next_data(self):
676 index = self._next_index() # may raise StopIteration
--> 677 data = self._dataset_fetcher.fetch(index) # may raise StopIteration
678 if self._pin_memory:
679 data = _utils.pin_memory.pin_memory(data, self._pin_memory_device)
~\AppData\Roaming\Python\Python38\site-packages\torch\utils\data\_utils\fetch.py in fetch(self, possibly_batched_index)
52 else:
53 data = self.dataset[possibly_batched_index]
---> 54 return self.collate_fn(data)
~\AppData\Roaming\Python\Python38\site-packages\torch\utils\data\_utils\collate.py in default_collate(batch)
263 >>> default_collate(batch) # Handle `CustomType` automatically
264 """
--> 265 return collate(batch, collate_fn_map=default_collate_fn_map)
~\AppData\Roaming\Python\Python38\site-packages\torch\utils\data\_utils\collate.py in collate(batch, collate_fn_map)
140
141 if isinstance(elem, tuple):
--> 142 return [collate(samples, collate_fn_map=collate_fn_map) for samples in transposed] # Backwards compatibility.
143 else:
144 try:
~\AppData\Roaming\Python\Python38\site-packages\torch\utils\data\_utils\collate.py in (.0)
140
141 if isinstance(elem, tuple):
--> 142 return [collate(samples, collate_fn_map=collate_fn_map) for samples in transposed] # Backwards compatibility.
143 else:
144 try:
~\AppData\Roaming\Python\Python38\site-packages\torch\utils\data\_utils\collate.py in collate(batch, collate_fn_map)
117 if collate_fn_map is not None:
118 if elem_type in collate_fn_map:
--> 119 return collate_fn_map[elem_type](batch, collate_fn_map=collate_fn_map)
120
121 for collate_type in collate_fn_map:
~\AppData\Roaming\Python\Python38\site-packages\torch\utils\data\_utils\collate.py in collate_numpy_array_fn(batch, collate_fn_map)
167 # array of string classes and object
168 if np_str_obj_array_pattern.search(elem.dtype.str) is not None:
--> 169 raise TypeError(default_collate_err_msg_format.format(elem.dtype))
170
171 return collate([torch.as_tensor(b) for b in batch], collate_fn_map=collate_fn_map)
TypeError: default_collate: batch must contain tensors, numpy arrays, numbers, dicts or lists; found object
Why does it throw the error and returns that it only found an object?
The error you're encountering is because PyTorch's default collate function (default_collate) doesn't know how to handle the custom objects in your dataset.
The error message "TypeError: default_collate: batch must contain tensors, numpy arrays, numbers, dicts or lists; found object" indicates that default_collate was expecting one of the listed types, but found a different type ("object") instead.
In your case, you have numpy arrays in your dataset, which the DataLoader tries to collate (combine) into a batch. PyTorch's DataLoader expects the data to be in the form of PyTorch tensors. Therefore, before sending the data to the DataLoader, you should convert your numpy arrays to PyTorch tensors.
Here's how you can resolve the issue:
__getitem__ method.__getitem__ method.Here's a modified version of what your dataset's __getitem__ might look like:
import torch
class YourDataset(torch.utils.data.Dataset):
def __init__(self, ...): # your other initialization arguments
...
def __getitem__(self, index):
a, b = ... # however you're getting your numpy arrays currently
a_tensor = torch.from_numpy(a)
b_tensor = torch.from_numpy(b)
return a_tensor, b_tensor
By doing this, you're ensuring that whenever an item from the dataset is accessed, it's already in tensor form, which DataLoader can handle without any issues.