I have really loved the environment and the progression of tasks, some looking impossible at first, but doable at last.
Great game design.
PS: yes it would be nice to have the test seeds, because it was really slow to test.
Is there any way to know which tag corresponds to a given submission from the issue tab?
Now there are several queued but it is hard to know to which commit they correspond.
It does not seem to work for us
Thanks. We now think that overfitting is the problem as well.
I did not expect to overfit such a complex environment, but we probably did.
On the training seeds we average out above 20 over 5 runs, but on the evaluation seeds seem to be a totally different story.
Now the bot is not even picking up new submissions. Anyone else experiencing the same issues?
we are experiencing very different results between what we get when evaluating compared to the local and debug mode.
Locally and in debug mode we get the same expected score, while in evaluation we get unexpected low or a lot lower performance.
For example episode 1, we always stop at level 5 in evaluation. According to our stats we have 99% success across the training seeds at level 5 and indeed we never fail at 5 locally and in debug mode.
Now I understand that the evaluation seeds are different, but we cannot understand how there can be such a difference. We tried to change model at level 5 but the behavior is the same, fine locally and in debug
For the admins this is one debug test for instance:
This is one evaluation:
It would be interesting to have some info on episode 1 to understand (a video?), to know how it dies?
Maybe it’s reporting steps in action-steps. Although there are 3000 time units at start each env step is 5 time units.
Nevertheless we also find the evaluation results a bit weird.