This repository was archived by the owner on Nov 17, 2023. It is now read-only.
This repository was archived by the owner on Nov 17, 2023. It is now read-only.
Description
Currently, MXNet's fancy indexing will fail when indexing with a list rather than a tuple (while it works with NumPy or PyTorch).
Minimal example to reproduce the issue:
import mxnet.ndarray as nd
# Creating a simple third order tensor
a = nd.arange(24).reshape((3, 4, 2))
# This works fine
a[slice(None), [1, 2, 1], [0, 1, 2]]
# This should work but does not
a[[slice(None), [1, 2, 1], [0, 1, 2]]]
# This works as expected
a[tuple([slice(None), [1, 2, 1], [0, 1, 2]])]