I’m using PyCharm as default IDE which is very powerful for several reasons.
Autocompletion of torch functions doesn’t work in several scenery such as using functions directly imported in torch module. Even though, when executed, it works perfectly.
Scenery 1
# It doesn't recognize functions like `cat`
torch.cat([a, b])
Scenery 2
class MyModel(nn.Module):
def __init__(self):
super(MyModel).__init__()
def forward(self, *input):
# Do something awesome
model = MyModel()
# Functional calling is not recognized and its really helpful
model(x)