Method predict_all_actions

rl/approx_control.py:74–75 · view source on GitHub ↗

(self, s)

Source from the content-addressed store, hash-verified

72	return x @ self.w
73
74	def predict_all_actions(self, s):
75	return [self.predict(s, a) for a in ALL_POSSIBLE_ACTIONS]
76
77	def grad(self, s, a):
78	sa = merge_state_action(s, a)

epsilon_greedyFunction · 0.45

approx_control.pyFile · 0.45

predictMethod · 0.95

no test coverage detected