【PyTorch Bug】RuntimeError: Boolean value of Tensor with more than one value is ambiguous

Case 1

Will contain Two or more The Boolean tensor of is used in if Judging conditions ：

a = torch.tensor([True, False])
if a:
    pass

One of the possible reasons for this error is to judge a Not for None, At this time, it should be changed to the following statement

if a is not None:

It should be noted that , If a Contains only one Boolean value , Then the judgment will not be wrong ：

a = torch.tensor([True])
if a:
    print(1)
# 1

Case 2

When using cross entropy loss Not instantiated first ：

inputs = torch.randn(6, 4)
target = torch.randint(4, (6, ))
loss = nn.CrossEntropyLoss(inputs, target)

The loss should be calculated after instantiation ：

criterion = nn.CrossEntropyLoss()
loss = criterion(inputs, target)

Case 3

Yes, it contains Two or more The Boolean tensor of performs or、and、not This kind of operation ：

a = torch.tensor([True, False])
b = torch.tensor([False, True])
"""  The following three operations will report errors  """
print(a or b)
print(a and b)
print(not a)

It should be noted that , If a and b All contain only one Boolean value , There will be no mistakes ：

a = torch.tensor([True])
b = torch.tensor([False])
print(a or b)
# tensor([True])
print(a and b)
# tensor([False])
print(not a)
# False