3 ms·

Yeah, definitely not pretraining. I've been referring to fine-tuning as training, since that's what's happening when you "fine-tune" a llama model.