Cookies help us deliver our services. By using our services, you agree to our use of cookies. Learn more

crowdAI is shutting down - please read our blog post for more information

NIPS 2017: Learning to Run

Reinforcement learning environments with musculoskeletal models


Important Announcement : Round 2 Updates : Submission Screencast

Posted by spMohanty about 4 years ago

Hi everyone,

Here is a link to a screencast explaining the submission procedure for round2 : https://www.youtube.com/watch?v=2e495B8kmUk&feature=youtu.be.

Some more updates about rules for round2 are :

  • 1) Round 2 Deadline extended to Monday, November 13th, 12:00 UTC
  • 2) All eligible participants are allowed 5 successful submissions (instead of 3) and upto 2 failed submissions
  • 3) The state of the submitted models can now be reviewed at : https://www.crowdai.org/challenges/nips-2017-learning-to-run/submissions. In case that the grader fails to grade a particular container, a “failed” status will be reflected on the referenced page.
  • 4) The submission container will not have access to external network when being graded
  • 5) Each submission container is allowed to use a maximum memory of 5GB
  • 6) Each submission will be run for atleast N=10 simulations.
  • 6.5) In case of a tie between the top-2 participants, we will re-run their simulations with N=20 and the new scores will be used as a tie breaker.
  • 7) Timeout for a submission is 8hours. In case of N > 10, the timeout will be proportionally increased.
  • 8) Each team can use only one account in the second round
  • 9) A team with two or more accounts accepted in the second round is obliged to report this issue to the organizers immediately, before submitting any solution
  • 9.5) To be eligible for the prize as a team, the combined submissions from the accounts of all team members in round 2 has to be less than or equal to upto 5 successful submissions (+2 failed submissions).
  • 10) The winners will be asked to release the code and the trained models of the solution
  • 11) Violation of the rules or other unfair activity may result in disqualification for the prizes

Best of Luck,


Posted by SDU_VSISLab  about 4 years ago |  Quote

hello, the image I pulled is 4.835 GB ,I commit it and now it is 5.606GB who is large than 5GB, I tried to reduce it but failed…so can I submit my submission with a little bigger….

Posted by Yongjin  about 4 years ago |  Quote

Does the constraint of ‘maximum memory of 5GB’ imply that the size of our DNN model should be smaller than 5GB?

Posted by spMohanty  about 4 years ago |  Quote


The requirement mentioned is for the memory(RAM) used at runtime by your submitted docker container. There are no constraints on the size of the actual docker container at the moment. (But thats with the hope that participants do not abuse it. If we notice participants abusing the absence of a size limit on the actual tar dump, we will have to enforce an upper bound on the container size. But that will also be of the order of ~10gb, so feel free to submit your model.


Posted by Yongjin  about 4 years ago |  Quote

I am new to Docker and I have a question. I am using Tensorflow 1.3. Should I install TF into the docker image that would be submitted to the grader?