Performance on FetchReach-v0

Question

Performance on FetchReach-v0

ArshT opened this issue 2 years ago · 2 comments

The policy for the FetchReach task seems to converge much faster than in the original report, considering the fact that the given command in the ReadMe should result in 10 * 2 * 50 = 1000 timesteps per epoch, while the HER OpenAI report has about 95k timesteps in a single epoch. Is there a reason why this implementation does much better?? Am I missing something?

Answer 1 · 2024-02-13T08:20:05.000Z

did u figure out the reason behind this yet?

Answer 2 · 2024-07-06T02:42:24.000Z

Hi
I apologize for the late reply. I think the improved performance relates to the normalization, As far as I remember, the original paper did not have this.