e2d3382bb4132ddb8aa586bf3c4c570be414f6af,tensorforce/models/vpg_model.py,VPGModel,update,#VPGModel#Any#,43

Before Change

        // Merge episode inputs into single arrays
        _, _, actions, batch_advantage, states = self.merge_episodes(batch)

        log_probs, loss, _ = self.session.run([self.log_probabilities, self.loss, self.optimize_op],
                                              {self.state: states,
                                               self.actions: actions,
                                               self.advantage: batch_advantage})
       // print("log probs:" + str(log_probs))
       // print("loss:" + str(loss))

After Change

        fetches = [self.optimize_op, self.log_probabilities, self.loss]
        fetches.extend(self.network.internal_state_outputs)

        feed_dict = {self.state: states, self.path_length: path_lengths, self.actions: actions, self.advantage: batch_advantage}
        feed_dict.update({internal_state: self.internal_states[n] for n, internal_state in enumerate(self.network.internal_state_inputs)})

        fetched = self.session.run(fetches, feed_dict)
        log_probs = fetched[1]
        loss = fetched[2]

Instances