deluca.agents.Zero

class deluca.agents.Zero(*args, **kwargs)

Zero: agent that plays the zero action

Public Data Attributes:

Inherited from JaxObject

name
attrs

Public Methods:

__init__(shape)	Description: iniitalizes the Zero agent
__call__(state)	Description: provide an action given a state

Inherited from Agent

__init_subclass__(*args, **kwargs)	For avoiding a decorator for each subclass
__call__(state)	Description: provide an action given a state
reset()
feed(reward)

Inherited from JaxObject

__new__(cls, *args, **kwargs)	For avoiding super().__init__()
__init_subclass__(*args, **kwargs)	For avoiding a decorator for each subclass
__str__()	Return str(self).
__setattr__(key, val)	Implement setattr(self, name, value).
save(path)
load(path)
throw(err, msg)

---

__call__(state: jax._src.numpy.lax_numpy.ndarray) → jax._src.numpy.lax_numpy.ndarray

Description: provide an action given a state

Parameters

state (jnp.ndarray) – current state

Returns

action to take

Return type

action (jnp.ndarray)

__init__(shape: Sequence[int]) → None

Description: iniitalizes the Zero agent

Parameters

shape – shape of the action the agent should take

Returns

None

Notes

• The Zero agent cannot, in this implementation, return a scalar that passes jnp.isscalar; we can emulate via Zero(shape=()), whic would return jnp.array(0.0)