time taken to go through the paper and will try our best here to respond to as many of the questions raised as possible

Firstly we would like to thank all of the reviewers for their insightful and thorough feedback. To start with, three of the reviewers noted that we had not included any comparisons with other model-based approaches. For HER, we used the OpenAI Baselines implementation. If we compare with Plappert et al. To address some of reviewer 2's other questions: we agree that All GANs in the ensemble are trained from batches taken from the same replay buffer.