Call Params

#1
by Lya - opened

Hi @joehdownardkainos , thank you for sharing this model;
I am getting different results when I try this model via API on this model page vs from local download using pipeline or AutoModelForSeq2SeqLM. API results are A LOT better - could you please share any additional parameters (like beam, min max len etc) please?

Sign up or log in to comment