Location
AU
Badges
Activity
Ratings Progression
Challenge Categories
Challenges Entered
Completed
Carnegie Mellon University
Arrival
The first, open autonomous racing challenge.
Latest submissions
See All| failed | 181166 | ||
| graded | 178752 | ||
| failed | 178750 |
| Participant | Rating |
|---|---|
stefan_podgorski
|
0 |
james_bockman
|
0 |
SamBahrami
|
0 |
| Participant | Rating |
|---|
Learn-to-Race: Autonomous Racing Virtual Challenge
Ground truth segmentation image in training phase seems to be invalid
Over 2 years agoCheck your active sensors in config.py, afaik the order matters but I would check that assumption because list. Since we aren’t getting a dictionary it’s difficult to be certain but when I was previously checking input sizes they seemed to be fine.
Success rate is always 0.0 during evaluation
Over 2 years agoI’m not sure what your problem may be but I’ll add that the unmodified SAC agent doesn’t go around any of the tracks as far as I know. I don’t think it works at all from our experiments. I would say it is your actual agent your submitting. To debug my issues I print out a lot in the debugger to check what’s going on inside the sim like if the car is even moving, if images are being processed, etc.
Multi-camera evaluation has shape 800x800
Over 2 years agoI am trying to submit a multi-cam evaluation and get an error when I stack my images to run through my model. The images I get at test time for left and right are 800x800, while front is 512x384. As far as I know we do not set these, they are set on the simulation end.
[Round 2] Launch - Expected Date
Almost 3 years agoI think that maybe we were not meant to be able to submit before, but could. So they have blocked submissions now, if your submission says Submission failed : Failed to communicate with the grader. then it isn’t on your side.
Information available during training in round 2
Almost 3 years agoSorry, during the training step (1hr practice).
Information available during training in round 2
Almost 3 years agoWhat ground truth information is available in round two?
- Full ground truth segmentation?
- Additional cameras?
- All multimodal information?
Observation delay limits processing to maximum of 10FPS
Almost 3 years agoAre we meant to be able to change this? I am able to run my code locally at around 30FPS while on the server it runs at around 5-6FPS because of the obs_delay parameter in config.py being overwritten to 0.1s. This 0.1s delay essentially makes a hard limit that no ones code should run faster than 10FPS.
Is this and intentional limitation to make the processing power similar to what the real world car can actually achieve?
KeyError: ‘success_rate’
Almost 3 years ago@jyotish I don’t think ignoring the absence of the key is going to do any good.
Let’s say success rate is meant to be 100%, which is the case for my agent and it shows in my own logs, the key is missing and it is replaced with 0 giving a false positive result for that run.
This issue is likely the cause of not one vehicle completing the track 100% on the leader board.
I’m currently trying to figure out what the issue could be but I’m afraid I don’t have access to the files responsible for this.
Ground truth segmentation image in training phase seems to be invalid
Over 2 years agoUnfortunately having the agent running on local doesn’t mean it will run on the challenge server (re: my 20+ failed debugging submissions over the last month). Keep in mind that there is a significant amount of files overwritten for the test phase so changes to some files don’t result in changes during testing. I’m not sure what issue you are encountering, sorry I can’t be of any more help.