Spaces:
Running
on
CPU Upgrade
Multiple model failures
Hi! I've submitted a couple of models to the leaderboard in the last couple of days. One or two of them succeeded but most started, hung out in RUNNING for 2-5 hours, then show up as FAILED in open-llm-leaderboard/requests. I've ensured they can all be loaded using AutoModel and AutoTokenizer.
What can I do to diagnose what's going on here?
Thanks!
Hi
@chargoddard
,
Could you be more specific? Which models are you referring to?
Sure! As a specific example, earlier today I submitted Chronorctypus-Limarobormes-13b.
Thank you!
You model passed, but we had a small connectivity/auth problem over the weekend, so all models launched between Friday and today need to have their results pushed manually. I pushed a first batch this morning and will wait for model results tonight to push the rest.
If you still have problems with your models tomorrow ping me again :)
Great, thanks for looking at this! There's no hurry, I just wanted to know if I bungled something on my end. :)