mengf1/cher

Curriculum-guided Hindsight Experience Replay (NeurIPS-2019)

cher's Issues

cannot install because there is no setup document

there is no setup.py at root directory.

Success rate and hyperparameters of CHER on FetchSlide

Hi!

Have you tried launching the CHER algorithm on the FetchSlide-v1 environment? Did you get a good result? And if so, could you share hyperparameters for CHER on FetchSlide?
I have only achieved about 50% of success on FetchSlide.
I have seen several papers, where CHER performs well on FetchSlide. In particular, in the Soft HER paper(https://arxiv.org/abs/2002.02089) CHER shows better results on FetchSlide (success rate is around 0.65). And I am trying to reach it by tuning lambda and batch size, but it only slightly improves outcomes.
Could you help, please?

Thanks

The reproduction results are too far apart！Ask for advice!

Dear author, Your article has taught me a lot, and I understand your algorithm and experimental process. However, when I used your code to reproduce the experimental results, I found that the results were not satisfactory and doubled from the results in the paper. For example, DDPG + CHER experiment in the HandReach environment, I got a lower success rate than the worst success rate on this, so I would like to ask sincerely how to improve to get close to the experimental results on the paper.
Thanks,
Best wishes

results_plotter.py

Hi,

After the training is finished for the environment "FetchReach-v1", how can I use results_plotter.py to plot the results ?

Thank you