---
datasets:
- PocketDoc/DansTestYard
---
Repository for my in progress LoRa files.
At the moment they are trained against https://huggingface.co/wcde/llama-13b-4bit-gr128 but they should work on any 4bit 128 group size 13b llama model.