wax.modules package

Gym modules

In WAX-ML, an agent and environments are simple functions:

A Gym feedback loops can be represented with the diagram:

Equivalently, it can be described with the pair of init and apply functions:

gym_feedback

Gym feedback between an agent and a Gym environment.

Online Learning

WAX-ML contains a module to perform online learning for supervised problems.

online_supervised_learner

Online supervised learner.

Other Haiku modules

buffer	Implement buffering mechanism.
diff	Implement difference of values on sequential data.
ewma	Compute exponentioal moving average.
ewmcov	Compute exponentially weighted covariance.
ewmvar	Compute exponentially weighted variance.
fill_nan_inf	Fill nan, posinf and neginf values.
has_changed	Detect if something has changed.
lag	Delay operator.
ohlc	Open-High-Low-Close binning.
pct_change	Relative change between the current and a prior element.
rolling_mean	Rolling mean.
update_on_event	Apply a module when an event occur otherwise return last computed output.