Learner1D.load throws exception when file is empty

Should be a pretty simple fix.

l = adaptive.Learner1D(lambda x: x, [0, 1])
l.save('foo')
l.load('foo')
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-30-af0e4613be23> in <module>
      1 l = adaptive.Learner1D(lambda x: x, [0, 1])
      2 l.save('foo')
----> 3 l.load('foo')

~/.conda/envs/dev/lib/python3.6/site-packages/adaptive/learner/base_learner.py in load(self, fname, compress)
    136         with suppress(FileNotFoundError, EOFError):
    137             data = load(fname, compress)
--> 138             self._set_data(data)
    139 
    140     def __getstate__(self):

~/.conda/envs/dev/lib/python3.6/site-packages/adaptive/learner/learner1D.py in _set_data(self, data)
    655 
    656     def _set_data(self, data):
--> 657         self.tell_many(*zip(*data.items()))
    658 
    659 

TypeError: tell_many() missing 2 required positional arguments: 'xs' and 'ys'