benchmarks?

#1
by vince62s - opened

Do you guys have some benchmarks? my first tests show results just under the 9B model.

UTTER - Unified Transcription and Translation for Extended Reality org

Hi @vince62s . I believe in our benchmarks it was a bit better than the 9B model. But note that the model is only a half trained checkpoint... it was only trained for 2T tokens and the goal is to train it with at least 4T tokens

(the 9B model is fully trained with 4T tokens already)

Sign up or log in to comment