Propulsion: Steering LLM with Tiny Fine-Tuning

[Paper] [Code]

---

🙋 Please let us know if you find out a mistake or have any suggestions!

🌟 If you find this resource helpful, please consider to star this repository and cite our research:

@article{kowsher2024propulsion,
  title={Propulsion: Steering LLM with Tiny Fine-Tuning},
  author={Kowsher, Md and Prottasha, Nusrat Jahan and Bhat, Prakash},
  journal={arXiv preprint arXiv:2409.10927},
  year={2024}
}

Introduction

Propulsion, a parameter-efficient fine-tuning (PEFT) method designed to optimize task-specific performance while drastically reducing computational overhead

Requirements

Use python 3.11 from MiniConda

• torch==2.3.0
• accelerate==0.33.0
• einops==0.7.0
• matplotlib==3.7.0
• numpy==1.23.5
• pandas==1.5.3
• scikit_learn==1.2.2
• scipy==1.12.0
• tqdm==4.65.0
• peft==0.12.0
• transformers==4.44.0
• deepspeed==0.15.1
• sentencepiece==0.2.0

To install all dependencies:

pip install -r requirements.txt

Datasets

You can access the datasets from hugginface

Quick Demos

To get started with propulsion, follow these simple steps:

1. Import the necessary modules:

   import propulsion
   from transformers import RobertaForSequenceClassification

2. Load a pre-trained model and apply PEFT:

   model = RobertaForSequenceClassification.from_pretrained('model_name')
   propulsion.PEFT(model)

3. Now you're ready to fine-tune your model using propulsion.