d641cd7cf8d44d76e34076ed6df3c04154976bf4,mushroom/algorithms/td.py,TD,fit,#TD#Any#Any#,18

Before Change

        assert n_iterations == 1 and len(dataset) == 1

        sample = dataset[0]
        sa = [np.array([sample[0]]), np.array([sample[1]])]

        q_current = self.approximator.predict(sa)
        q_next = self._next_q(np.array([sample[3]])) if not sample[4] else 0.

        q = q_current + self.learning_rate(sa) * (
            sample[2] + self._gamma * q_next - q_current)

        self.approximator.fit(sa, q, **self.params["fit_params"])

    def __str__(self):

After Change

        assert n_iterations == 1 and len(dataset) == 1

        s, a, r, ss, ab = self._parse(dataset)
        self._update(s, a, r, ss, ab)

    def _parse(self, dataset):
        sample = dataset[0]
        s = np.array([sample[0]])

Instances