Skip to content

Weird behavior when Iterating over tensors #124

@StenSipma

Description

@StenSipma

Whilst working with the tensor.Iterator functionality I found some strange behavior, when using iterator in a for loop like follows:

it = t.Iterator()  // t is a tensor.Dense
for i, err := it.Start(); err == nil; i, err = it.Next() {
	fmt.Printf("i = %v, coord = %v\n", i, it.Coord())
}
  1. When using the iterator as stated above, the 'first' element of the tensor is always last in the iteration. This is not really a big issue if you want to iterate over all elements and order does not matter, but it is weird nonetheless. As an example consider the tensor [1, 2, 3, 4] with shape (2, 2). It will visit the elements in order:
    i = 0, coord = [0 1]
    i = 1, coord = [1 0]
    i = 2, coord = [1 1]
    i = 3, coord = [0 0]  <------------ should be first
    
  2. When iterating over a vector or a scalar, the indices are off by one. Again same tensor, but as a vector (shape (4,)):
    i = 0, coord = [1]
    i = 1, coord = [2]
    i = 2, coord = [3]
    i = 3, coord = [4]  <----------- gives index error when used
    
    The same thing happens with 'scalar-like' tensors, like: [1] with shape (1, 1, 1):
    i = 0, coord = [1 0 0]  <--------- also index error, should be [0 0 0] always for scalars
    
    Interestingly, when the tensor is a vector/scalar View of a tensor that is not a vector/scalar (i.e. obtained by slicing), the issue does not happen.

What I found to work properly is to use the following for loop instead, but I don't think this is the intended way of iterating.

it = t.Iterator()
for it.Reset(); !it.Done(); it.Next() {
	...
}

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions