Hi Admin, I am new to the environment. I submitted Run #47173; Run #47133 after checking my code suitability locally. It concluded in a few mins. But when I submitted it to the cloud, it exceeded the 15 hours window and failed. Can you please help me learn how to fix this? Thank you.
Locally, you are only required to run on 100 datasets, where as on the cloud, you must run on the 10’000 samples.
Thank you very much for the prompt clarification. I tried a fairly simplistic run for the first attempt with no pre-treatment/cleaning. Is there a bare-bones example/guidelines for the very basic type of “optimization” that may be needed to successfully run the full workload in the cloud? I tried the same attempt on CPU and GPU -both failed after 15 hour time-outs. Thank you for your guidance.
The code in the quickstarter notebook is almost instant as the solution is fairly basic.
But for model that are more complicated, we recommend that you indeed try to optimize to fit under the 15h limit which is requirement for the competition.
To help with your debugging, we recommend adding a `tqdm(…)` to track the time.