3ce4fd552a89c6fd5e479ad3713418b036337ff6,mushroom_rl/algorithms/value/batch_td/fqi.py,FQI,fit,#FQI#Any#,52

Before Change

        for _ in trange(self._n_iterations, dynamic_ncols=True,
                        disable=self._quiet, leave=False):
            fit(dataset)

    def _fit(self, x):
        
        Single fit iteration.

After Change

            if self._target is None:
                self._target = reward
            else:
                q = self.approximator.predict(next_state)
                if np.any(absorbing):
                    q *= 1 - absorbing.reshape(-1, 1)

                max_q = np.max(q, axis=1)
                self._target = reward + self.mdp_info.gamma * max_q

            self.approximator.fit(state, action, self._target, **self._fit_params)

Instances