llama lora

LORA finetune on LLAMA 13B + VICUNA finetune, trained on ~500 limericks dataset. Since most limericks are dirty - that's what you get with this LORA too. You have been warned.

The priming is easy, just start:

There was a man in