dpctl.tensor.asarray raises TypeError with [[], []] as input

[densmirn]

Reproducer:

In [1]: import dpctl.tensor as dpt

In [2]: dpt.asarray([[],[]])
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-7-daad18362542> in <module>
----> 1 dpt.asarray([[],[]])

/localdisk/work/denissmi/dpctl/dpctl/tensor/_ctors.py in asarray(obj, dtype, device, copy, usm_type, sycl_queue, order)
    385                 "Converting Python sequence to usm_ndarray requires a copy"
    386             )
--> 387         _, dt, devs = _array_info_sequence(obj)
    388         if devs == _host_set:
    389             return _asarray_from_numpy_ndarray(

/localdisk/work/denissmi/dpctl/dpctl/tensor/_ctors.py in _array_info_sequence(li)
     57         if dim is None:
     58             dim = el_dim
---> 59             dt = np.promote_types(el_dt, el_dt)
     60             device = device.union(el_dev)
     61         elif el_dim == dim:

TypeError: did not understand one of the types

dpctl version:

In [3]: import dpctl

In [4]: dpctl.__version__
Out[9]: '0.11.0+31.gd9cfb43a'