File size: 482 Bytes
87bed78 66c848c 87bed78 66c848c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 |
---
license: other
license_name: llama3
license_link: LICENSE
language:
- en
tags:
- llama
- ORPO
---
Barcenas Llama3 8b ORPO
Model trained with the novel new ORPO method, based on the recent Llama 3 8b, specifically: VAGOsolutions/Llama-3-SauerkrautLM-8b-Instruct
The model was trained with the dataset: reciperesearch/dolphin-sft-v0.1-preference which uses Dolphin data with GPT 4 to improve its conversation sections.
Made with ❤️ in Guadalupe, Nuevo Leon, Mexico 🇲🇽 |