How to use it efficiently?
#3
by
Varkoyote
- opened
I've been trying all these 14b Qwen tunes but every time it goes into incoherence and repeats itself quickly and horribly like old 7b models, any idea why? With these issues NeMo is still far superior, and it's getting old :(
Edit: I meant to post this on the most recent version of the model, my bad...