Open Instruct Stanford Alpaca 7b

🚀 Open-Instruct Stanford Alpaca 7B

This is a 7B LLaMa model finetuned on the Stanford Alpaca dataset. Note that this is a model diff - see below for usage instructions.

This model was trained as part of the paper How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources. The codebase for training and evaluating this model can be found at https://github.com/allenai/open-instruct.

This model is licensed under the AI model license in LICENSE.txt along with the original Llama license (llama_license.txt).

🚀 Quick Start

Prerequisites

We assume you already have access to a LLaMa model in HF format. You can find details on getting access and converting the model here: https://huggingface.co/docs/transformers/main/model_doc/llama.

Steps

1. Clone https://github.com/allenai/open-instruct and install the required dependencies, or just copy scripts/weight_diff.py and install the minimal requirements listed in weight-diff-requirements.txt.
2. Download or clone this model diff to the same machine.
3. Run the following command:

python scripts/weight_diff.py recover --path_raw ${hf_llama_path} --path_tuned ${output_path} --path_diff ${diff_location}

After running the command, you will get a recovered model. Note that this process consumes a significant amount of RAM, especially for larger models.

💻 Usage Examples

Input Format

The model is trained to use the following format (note the newlines):

<|user|>
Your message here!
<|assistant|>

For best results, format all inputs in this way.

📚 Documentation

Performance

Here is the performance of this model across benchmarks explored in our paper How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources:

Citation

If you use this model, please cite our work, the llama paper, and the original dataset:

@misc{wang2023far,
      title={How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources}, 
      author={Yizhong Wang and Hamish Ivison and Pradeep Dasigi and Jack Hessel and Tushar Khot and Khyathi Raghavi Chandu and David Wadden and Kelsey MacMillan and Noah A. Smith and Iz Beltagy and Hannaneh Hajishirzi},
      year={2023},
      eprint={2306.04751},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

@misc{touvron2023llama,
      title={LLaMA: Open and Efficient Foundation Language Models}, 
      author={Hugo Touvron and Thibaut Lavril and Gautier Izacard and Xavier Martinet and Marie-Anne Lachaux and Timothée Lacroix and Baptiste Rozière and Naman Goyal and Eric Hambro and Faisal Azhar and Aurelien Rodriguez and Armand Joulin and Edouard Grave and Guillaume Lample},
      year={2023},
      eprint={2302.13971},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

@misc{alpaca,
  author = {Rohan Taori and Ishaan Gulrajani and Tianyi Zhang and Yann Dubois and Xuechen Li and Carlos Guestrin and Percy Liang and Tatsunori B. Hashimoto },
  title = {Stanford Alpaca: An Instruction-following LLaMA model},
  year = {2023},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/tatsu-lab/stanford_alpaca}},
}

📄 License

This model is licensed under the AI model license given in LICENSE.txt along with the original Llama license (llama_license.txt).