I want to run pytorch program with my dataset.Please help me in solving the problem.The error is as follows:
Traceback (most recent call last):
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 311, in
fire.Fire(demo)
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\fire\core.py”, line 138, in Fire
component_trace = _Fire(component, args, parsed_flag_args, context, name)
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\fire\core.py”, line 468, in _Fire
target=component.name)
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\fire\core.py”, line 672, in _CallAndUpdateTrace
component = fn(*varargs, **kwargs)
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 289, in demo
n_epochs=n_epochs, batch_size=batch_size, seed=seed)
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 168, in train
n_epochs=n_epochs,
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 42, in train_epoch
for batch_idx, (input, target) in enumerate(loader):
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\torch\utils\data\dataloader.py”, line 346, in next
data = self._next_data()
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\torch\utils\data\dataloader.py”, line 386, in _next_data
data = self._dataset_fetcher.fetch(index) # may raise StopIteration
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\torch\utils\data_utils\fetch.py”, line 47, in fetch
return self.collate_fn(data)
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\torch\utils\data_utils\collate.py”, line 87, in default_collate
return [default_collate(samples) for samples in transposed]
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\torch\utils\data_utils\collate.py”, line 87, in
return [default_collate(samples) for samples in transposed]
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\torch\utils\data_utils\collate.py”, line 72, in default_collate
return default_collate([torch.as_tensor(b) for b in batch])
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\torch\utils\data_utils\collate.py”, line 63, in default_collate
return torch.stack(batch, 0, out=out)
RuntimeError: stack expects each tensor to be equal size, but got [650] at entry 0 and [108] at entry 1
Thanks in advance!!
Based on the error message if seems that the tensors stored in batch, which should be created in the Dataset.__getitem__ method, do not have the same shape.
If you are dealing with variable input shapes, you could use a custom collate function as described here.
Thank you very much for answering.But the error appears:
Traceback (most recent call last):
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 320, in
fire.Fire(demo)
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\fire\core.py”, line 138, in Fire
component_trace = _Fire(component, args, parsed_flag_args, context, name)
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\fire\core.py”, line 468, in _Fire
target=component.name)
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\fire\core.py”, line 672, in _CallAndUpdateTrace
component = fn(*varargs, **kwargs)
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 298, in demo
n_epochs=n_epochs, batch_size=batch_size, seed=seed)
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 177, in train
n_epochs=n_epochs,
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 54, in train_epoch
input = input.cuda()
AttributeError: ‘numpy.ndarray’ object has no attribute 'cuda
Although ‘input’ is converted as torch.tensor,ValueError: expected sequence of length 528 at dim 1 (got 76) appears.Please answer how can I solve.Thank a lot!
The first error is created, since input seems to be a numpy array.
You can check the type via print(type(input)).
Could you post the complete error message for the second issue and which line of code is raising this error?
Thank you very much!When’print(type(input))’,the line shows :type(input) <type(input) <class ‘list’>
The code rising this error:
for batch_idx,(input, target) in enumerate(loader):
# Create vaiables
if torch.cuda.is_available():
print(‘type(input)’,type(input))
input = input.cuda()
The error message is :Traceback (most recent call last):
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 323, in
fire.Fire(demo)
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\fire\core.py”, line 138, in Fire
component_trace = _Fire(component, args, parsed_flag_args, context, name)
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\fire\core.py”, line 468, in _Fire
target=component.name)
File “C:\Users\hp\Anaconda3\envs\tf-gpu\lib\site-packages\fire\core.py”, line 672, in _CallAndUpdateTrace
component = fn(*varargs, **kwargs)
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 301, in demo
n_epochs=n_epochs, batch_size=batch_size, seed=seed)
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 180, in train
n_epochs=n_epochs,
File “C:\Users\hp\Downloads\efficient_densenet_pytorch-master\demoEmotion.py”, line 56, in train_epoch
input = input.cuda()
AttributeError: ‘list’ object has no attribute ‘cuda’
Now when I run again,list object has no ‘cuda’. Please answer!Thanks a lot in advance!
As the error explains, list objects do not have the .cuda() method, which is a tensor method.
Could you unpack the list and call cuda() on each tensor inside the list?
Thanks!How to unpack and call?Please explain.When I convert to torch.tensor,ValueError: expected sequence of length 528 at dim 1 (got 76) appears.
Since it’s a list, you could get each value via:
tensor0 = input[0]
tensor1 = input[1]
tensor2 = input[2]
...
Thank you very much!The problem above is solved.But ‘RuntimeError: Expected 4-dimensional input for 4-dimensional weight [24, 3, 3, 3], but got 3-dimensional input of size [3, 3, 158] instead’ error appears.Thanks a lot!
The input tensor might be missing the batch dimension, if the current shapes correspond to [channels, height, width].
If so, you could add the batch dimension via x = x.unsqueeze(0).
Thanks a lot!However the error ‘RuntimeError: Given input size: (150x1x79). Calculated output size: (150x0x39). Output size is too small’ appears.Why does the error appear?Thanks!