Unable to obtain torch tensor after reduction

[magrenimish]

Hi,
I am new to using the tool and am trying out some of its functionalities! I tried the following code and was expecting 'm' to be a torch tensor after performing the argmax reduction on 'out' but 'm' turns out to be a Lazy Tensor.

`from pykeops.torch import LazyTensor
a = torch.tensor([[0.1, 0.2, 0.7],
[0, 0.2, 0.8]], requires_grad=True)

b = torch.tensor([[1, 1, 1],
[0, 1, 0],
[0, 0, 1]], dtype=torch.float)

l = LazyTensor(logits_nc[None,:,:,None,None,10]) #shape = (1,2,3)

d = LazyTensor(desc_matrix_cc[:,None,:,None,10])#shape = (3,1,3)

out = (l1*d1)#shape = (3,2,3)

m = out.argmax(dim=-1)`
Am I missing correct symbolic axes assignment when creating the lazy tensors?