What is the Base model for this checkpoint?

#3
by zolicsaki - opened

Thank you for the great release, what base model was used for this?

They usually don't release base models for whatever reason unfortunately, would like to see that change though

Cohere For AI org

Hey @zolicsaki

Apologies for the delay in response. But Aya 23 models are based on Cohere's Command series models. You can check out our technical report for more details. Aya23-35B is specifically a fine-tune of Command-R whose weights are released.

I think they where referring to base model rather than instruct tune? I don't think this is a non fined tuned version, which is what they where asking about I think? Unless this is a fine tune on top of the already existing instruct tune, but the base model for that isn't released I think for whatever reason. I am not here to judge but personally would like that if it's something Cohere is comfortable doing

Sign up or log in to comment