Index_select, take all but specified indices?

Hello,

index_select allows me to retrieve rows of a tensor by specifying their respective indices. Now suppose my tensor is very large and I want almost all of these indices. Instead of specifying all row indices I want, is their a function where I specify all rows I do not want instead?

May be you can do like this

x = torch.randn(3, 4)
x=torch.FloatTensor(np.delete(x.data.numpy(), 1, 0))
print(x)