SUGARRA-3B

SUGARRA: Modelo 3B experimental ChatML euskaldun.

SUGARRA es el resultado de finetunear el modelo open_llama_3b_v2 con las instrucciones OpenAssistant v2 traducidas automáticamente al euskera usando recursos de Helsinki-NLP y tratadas en formato ChatML.

Prompt Template

SUGARRA usa el prompt template ChatML:

<|im_start|>user
Nor zen Isaac Newton?<|im_end|>
<|im_start|>assistant\n

<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>

Referències

@software{xaviviro2024sugarra,
  author = {xaviviro},
  title = {SUGARRA: Modelo 3B experimental ChatML euskaldun.},
  month = January,
  year = 2024,
  url = {https://huggingface.co/xaviviro/SUGARRA-3B}
}

@software{openlm2023openllama,
  author = {Geng, Xinyang and Liu, Hao},
  title = {OpenLLaMA: An Open Reproduction of LLaMA},
  month = May,
  year = 2023,
  url = {https://github.com/openlm-research/open_llama}
}

@software{together2023redpajama,
  author = {Together Computer},
  title = {RedPajama-Data: An Open Source Recipe to Reproduce LLaMA training dataset},
  month = April,
  year = 2023,
  url = {https://github.com/togethercomputer/RedPajama-Data}
}

@article{touvron2023llama,
  title={Llama: Open and efficient foundation language models},
  author={Touvron, Hugo and Lavril, Thibaut and Izacard, Gautier and Martinet, Xavier and Lachaux, Marie-Anne and Lacroix, Timoth{\'e}e and Rozi{\`e}re, Baptiste and Goyal, Naman and Hambro, Eric and Azhar, Faisal and others},
  journal={arXiv preprint arXiv:2302.13971},
  year={2023}
}