I went through the document and didn't find a way to roll the tdqm progress bar back. Trying to update the progress bar with a negative value just end up with an error.

In [31]: bar.update(-1)
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-31-707a754749a6> in <module>()
----> 1 bar.update(-1)
/usr/lib/python3/dist-packages/tqdm/_tqdm.py in update(self, n)
    925
    926         if n < 0:
--> 927             raise ValueError("n ({0}) cannot be negative".format(n))
    928         self.n += n
    929
ValueError: n (-1) cannot be negative

Actually I intended to print the machine learning model accuracy (\in [0,1]) with a progress bar library,
and the accuracy value while training is NOT monotonously increasing. Is there any workaround?

Thanks.