Metadata-Version: 1.1
Name: rounding
Version: 0.03
Summary: Stochastic Rounding
Home-page: http://dmaust.github.io/rounding/
Author: David Maust
Author-email: david@dmaust.net
License: Apache License, Version 2.0
Description: rounding
        ========
        
        Rounding methods
        
        `StandardRound` rounds to the nearest, and in the event of a tie, rounds up.
        
        `RoundToEven` rounds to the nearest, but in the event of a tie, rounds toward 
        the nearest even number.
        
        Stochastic rounding provides a mechanism to eliminate accumulated 
        roundoff error in the presence of a distribution where for 
        individual samples, the roundoff error is skewed. This is typically
        caused by small values.
        
        One drawback to stochastic rounding is the output is non-deterministic,
        but this can be avoided by providing a custom deterministic generator, 
        or invoking providing the random number generator with a fixed seed.
        
        .. code:: python
          
          r = random.Random()
          r.seed(123)
          sr = StochasticRound(precision=0, random_generator=r)
        
Platform: UNKNOWN
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: Apache Software License
