upbeat orbit
I tried to fine tune ada model for multi classification, following this cookbook
After fine tuning completed, I checked the result and both columns classification/accuracy and classification/weighted_f1_score in result.csv are NaN.
Why is this happening?
As for data, it was like this:
{"prompt":"What a rollercoaster ->","completion":" 14"}
{"prompt":"Well good luck! Hoping to see you in my diamond games soon. Psn is LokleLad if you ever wanna stack ->","completion":" 9"}
{"prompt":"Just learn from this and move on bro. Give yourself time to mourn. ->","completion":" 17"}
{"prompt":"So what should I do?? I’ve been dating her for eight years and am confused on what to do please help me. ->","completion":" 25"}
{"prompt":"Glad to hear it :) ->","completion":" 22"}
(I am fine tuning for emotion detection from text)
Is my data structure wrong or something?