Barcelona, ES
Activity
Ratings Progression
Challenge Categories
Challenges Entered
Completed
Unity Technologies
A new benchmark for Artificial Intelligence (AI) research in Reinforcement Learning
Latest submissions
See All| graded | 8963 | ||
| graded | 8954 | ||
| graded | 8868 |
| Participant | Rating |
|---|
| Participant | Rating |
|---|
Unity Obstacle Tower Challenge
Submissions are stuck
Over 5 years agoIs 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.
g
Different results between training, debug and evaluation
Over 5 years agoThanks. 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.
g
Submissions are stuck
Over 5 years agoNow the bot is not even picking up new submissions. Anyone else experiencing the same issues?
Different results between training, debug and evaluation
Over 5 years agoHi,
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
Any idea?
For the admins this is one debug test for instance:
https://gitlab.aicrowd.com/giadefa/obstacle-tower-challenge/issues/16
This is one evaluation:
https://gitlab.aicrowd.com/giadefa/obstacle-tower-challenge/issues/31
It would be interesting to have some info on episode 1 to understand (a video?), to know how it dies?
Episode in evaluation ends prematurely
Over 5 years agoMaybe 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.
Release of the evaluation seeds?
Over 5 years agoI have really loved the environment and the progression of tasks, some looking impossible at first, but doable at last.
Great game design.
g
PS: yes it would be nice to have the test seeds, because it was really slow to test.