34bae27ac7d3c39335a2b8c7d4cd4efba56aa398,rllib/execution/replay_buffer.py,PrioritizedReplayBuffer,sample,#PrioritizedReplayBuffer#Any#Any#,201

Before Change

            weights.append(weight / max_weight)
        weights = np.array(weights)
        encoded_sample = self._encode_sample(idxes)
        return tuple(list(encoded_sample) + [weights, idxes])

    @DeveloperAPI
    def update_priorities(self, idxes, priorities):

After Change

            weight = (p_sample * len(self._storage))**(-beta)
            count = self._storage[idx].count
            weights.extend([weight / max_weight] * count)
            batch_indexes.extend([idx] * count)
        batch = self._encode_sample(idxes)

        // Note: prioritization is not supported in lockstep replay mode.
        if isinstance(batch, SampleBatch):

Instances