Numpy indexing warnings

[twiecki]

/Users/twiecki/anaconda3/lib/python3.6/site-packages/theano/tensor/basic.py:6592: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  result[diagonal_slice] = x
/Users/twiecki/anaconda3/lib/python3.6/site-packages/theano/tensor/subtensor.py:2339: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  out[0][inputs[2:]] = inputs[1]
/Users/twiecki/anaconda3/lib/python3.6/site-packages/theano/tensor/subtensor.py:2197: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  rval = inputs[0].__getitem__(inputs[1:])
/Users/twiecki/anaconda3/lib/python3.6/site-packages/theano/tensor/subtensor.py:2339: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  out[0][inputs[2:]] = inputs[1]
/Users/twiecki/anaconda3/lib/python3.6/site-packages/theano/tensor/basic.py:6592: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  result[diagonal_slice] = x
/Users/twiecki/anaconda3/lib/python3.6/site-packages/theano/tensor/subtensor.py:2197: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  rval = inputs[0].__getitem__(inputs[1:])

[atreyasha]

Hello @twiecki! I am also encountering the same issue, might be an internal theano issue?

[nouiz]

This is a futurwarning. You can ignore it for now. It is NumPy that raise this warning. It tell us that its interface will change. So to support future NumPy version, some changes will be needed in Theano. If someone have time to make a PR, it would be great. But you can continue to use it without any changes.

…

On Thu, Apr 11, 2019 at 7:07 AM Atreya Shankar ***@***.***> wrote: Hello @twiecki <https://github.com/twiecki>! I am also encountering the same issue, might be an internal theano issue? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#6695 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AALC-_Nf0mHvSsmkh6LsntFcUeO01EX5ks5vfxd5gaJpZM4b766H> .

[twiecki]

@nouiz Can you point me to where in the code this likely originates?

[nouiz]

The error tell it: /Users/twiecki/anaconda3/lib/python3.6/site-packages/theano/tensor/basic.py:6592: ... Check the start of each line. It give the filename and line number that triggered the warning.

…

On Thu, Apr 11, 2019 at 11:43 AM Thomas Wiecki ***@***.***> wrote: @nouiz <https://github.com/nouiz> Can you point me to where in the code this likely originates? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#6695 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AALC-324yK8AhHMncR_yz_zbK42-bg9zks5vf1gPgaJpZM4b766H> .

[atreyasha]

It is also discussed as a numpy issue here: numpy/numpy#11411

[twiecki]

@AtreyaSh Want to do a PR?

[atreyasha]

@twiecki Yes we could do one. For me, I found the line causing the warning:

Which can be found here:

Theano/theano/tensor/subtensor.py

Line 2197 in 9feed78

rval = inputs[0].__getitem__(inputs[1:])

The entire function is:

def perform(self, node, inputs, out_):
    out, = out_
    check_advanced_indexing_dimensions(inputs[0], inputs[1:])
    rval = inputs[0].__getitem__(inputs[1:]) # this is the line triggering the warning
    # When there are no arrays, we are not actually doing advanced
    # indexing, so __getitem__ will not return a copy.
    # Since no view_map is set, we need to copy the returned value
    if not any(isinstance(v.type, TensorType) and v.ndim > 0
               for v in node.inputs[1:]):
        rval = rval.copy()
    out[0] = rval

Any suggestions on what to change?

[twiecki]

Hm, not sure what is going on there. Is inputs multidimensional?

[lucianopaz]

Isn't it the same as #6667? I made the warnings be handled like exceptions to get the full trace. I don't know how to make a fix though.