Llama 3.1 Super Nova Lite

Introduction

Llama-3.1-SuperNova-Lite is an 8 billion parameter language model developed by Arcee.ai. It is a distilled version of the larger Llama-3.1-405B-Instruct model, designed to maintain high performance while offering improved instruction-following capabilities and domain-specific adaptability.

Architecture

The model is based on the Meta-Llama-3.1-8B-Instruct architecture. It utilizes offline logits from the larger 405B parameter model to create a more efficient version, suited for organizations requiring robust performance with reduced computational resources.

Training

Llama-3.1-SuperNova-Lite was trained using a state-of-the-art distillation pipeline and an instruction dataset generated with EvolKit. This approach ensures the model's accuracy and efficiency across various tasks. Further information on its training process can be found at blog.arcee.ai.

Guide: Running Locally

1. Install Dependencies: Ensure you have Python and the transformers library installed.
2. Clone the Repository: Download the model files from the Hugging Face model card page.
3. Load the Model: Use the transformers library to load and initialize the model.
4. Run Inference: Input your text data to generate predictions.

For optimal performance, it is recommended to use a cloud GPU service such as AWS, Google Cloud, or Azure.

License

The Llama-3.1-SuperNova-Lite model is released under the llama3 license, which governs its use and distribution. Please refer to the license documentation for more details.