TheBloke

/

Luna-AI-Llama2-Uncensored-GGML

like 137

Hi, thank you as always. This model is exceptional, but I've noticed through llama.cpp that the Assistant will eventually get to a point that it just stops generating before max context.

It's randomly stopped around 700, 1100, 1200 a few times even with --context 2048

It's not only your repo, I've downloaded the identical model from 2 other repos and experienced the same abrupt stop.

I dunno if there's a problem with either the dataset, or llama.cpp

Any experience with this? Hear any similar issues?

Thank you.