Minghuan Liu
Minghuan Liu
I find that with a larger batch size, HER still not work, do you know why?
Hi, I've tried 4 processes and 2 processes, they both work but a single process with 2048 batch size cannot work.
If this is true, we should be able to succeed by scaling the number of episodes by K times? However, it seems not work either.
Great and many thanks. I did this because I find my own implementation of HER can only reach a success rate of 70-80% and I am figuring out what really...
This seems an important reason, but when I run with a single process, it just can not get any evidence of learning... (at least the avg gradient of 2 processes...
I think the learning rates for both the policy network and the value network are important hyper-parameters for these goal-conditioned tasks, after fine-tune some values I found that with only...
No... But maybe it is not the path problem. I think it is the error that the authors had mentioned [here](https://github.com/sisl/ngsim_env/blob/master/scripts/imitation/README.md), which says "this is somehow related to julia processes...
Did you work well with the NGSIM dataset that the authors packaged up? I encountered this error when I use the packaged NGSIM dataset.
> Hi all > I am not sure if this works for all, but I think the problem can be solved by removing the cache file under folder "~/julia/.julia/compiled/v1.1/PyCall/****.jl". It...
No. Actually I already gave it up ...