Spaces:
Running
on
CPU Upgrade
Need help doing eval- they keep failing
I have been attempting to submit my latest models for evaluation, specifically the following:
Negative_LLAMA_70B
Boomer_Qwen_72B
Despite multiple attempts and adjustments to various settings, the submissions have consistently failed. The models initially appear as being evaluated but subsequently disappear from the queue.
Could someone please assist me in resolving this issue? Any guidance or support would be greatly appreciated.
Thank you for your time and consideration,
Sicarius
I am sure you meant this for the hugging face team, so I apologize for intruding, but in my experience, when an eval fails, it is usually some issue with one or more of three files: "config.json", "generation_config.json" and "tokenizer_config.json".
this isn't all there is to it, the reason can be elsewhere, but whenever I had an issue such as yours I always fixed it by replacing these files with ones I knew to be working, like from base/reference models.
I am almost sure you already tried this, but if not, I hope this helps...