Description

Function _numpy_arrays_to_lists resulted in traceback when 0-dim numpy array was passed as a parameter. How to reproduce:

>>> from wandb.sdk.data_types import _numpy_arrays_to_lists
>>> import numpy as np
>>> x = _numpy_arrays_to_lists(np.array(4.0))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/rpitonak/workspace/github/client/wandb/sdk/data_types.py", line 2713, in _numpy_arrays_to_lists
    return [_numpy_arrays_to_lists(v) for v in payload.tolist()]
TypeError: 'float' object is not iterable

The problem is the .tolist() call. Explanation is in the numpy docs:

If a.ndim is 0, then since the depth of the nested list is 0, it will not be a list at all, but a simple Python scalar.

Testing

PR includes one test scenario for this use-case. I was not sure what is the expected behaviour if it should return [value] or scalar value. From the function definition it make sense to me that it is [value] but feedback is welcomed.