Cannot create large ndarray from numpy

[eric-haibin-lin]

>>> import mxnet as mx
>>> import numpy as np
>>> a = np.empty((5000000000,))
>>> a
array([0., 0., 0., ..., 0., 0., 0.])
>>> b = mx.nd.array(a)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/ubuntu/.local/lib/python2.7/site-packages/mxnet/ndarray/utils.py", line 146, in array
    return _array(source_array, ctx=ctx, dtype=dtype)
  File "/home/ubuntu/.local/lib/python2.7/site-packages/mxnet/ndarray/ndarray.py", line 2488, in array
    arr = empty(source_array.shape, ctx, dtype)
  File "/home/ubuntu/.local/lib/python2.7/site-packages/mxnet/ndarray/ndarray.py", line 3877, in empty
    return NDArray(handle=_new_alloc_handle(shape, ctx, False, dtype))
  File "/home/ubuntu/.local/lib/python2.7/site-packages/mxnet/ndarray/ndarray.py", line 134, in _new_alloc_handle
    c_array_buf(mx_uint, native_array('I', shape)),
OverflowError: unsigned int is greater than maximum
>>>

[wkcn]

Similar Issue: #13036

[apeforest]

This issue has been fixed. In 1.5.0 release, user need to build MXNet from source with the compilation flag USE_INT64_TENSOR_SIZE=1. We are working to make this flag on by default and available in pip package in next minor release. Closing this issue for now.