Spaces:

HuggingFaceH4

/

open_llm_leaderboard

like 9.63k

Running on CPU Upgrade

Even if you never uploaded it, someone else could have. That message appears because there was an attempted run, but it failed:

https://huggingface.co/datasets/open-llm-leaderboard/requests/blob/main/PracticeLLM/Twice-KoSOLAR-16.1B-test_eval_request_False_float16_Original.json

You should make sure the model follows the requirements mentioned here: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard in the "Submit here!" tab.

If all looks ok (and it does indeed looks ok for that model), then you should open a thread (like this one) in the Community section of the leaderboard (here, where we are now) requesting a new rerun (see here an example: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard/discussions/492 ). When requesting a rerun, make sure you provide the leaderboard datasets link to the model (the first link I have in this message).

I think, everything is okay..

-> I tested using lm-harness
-> Model precision & type: FP16 & safetensor
-> Model card: fullfill

But, I cannot submit...
https://huggingface.co/datasets/open-llm-leaderboard/requests/blob/main/PracticeLLM/Twice-KoSOLAR-16.1B-test_eval_request_False_float16_Original.json

What's wrong?
Can you upload my model?

Thanks..!

I mentioned at the end of my message: you'll have to request a rerun - see the example I provided.

You can't resubmit it yourself with same parameters. You could resubmit it with another precision, though.

Thank you!

I think my discussion is asking for a rerun enough, should I close and reopen this discussion?

You're welcome! 🤗
I'm not sure if you should close this thread and reopen another one. This thread should be enough, I guess. The next step would be for someone from the HF team to see this message and rerun it so you'll have to wait for that.

Thank you so much for your kindness. 🤗
Also, thank you for the link to the json file.

Hi ! Thanks @Mihaiii for your help :) The model has been re-submited