Arcee AI has introduced SuperNova, a 70-billion parameter language model designed for enterprise deployment. The model offers advanced instruction-following capabilities and full customization options, aiming to provide an alternative to API-based services.
SuperNova is built on Meta's Llama-3.1-70 billion-Instruct architecture and employs a novel post-training process. It can be deployed within an enterprise's cloud environment, starting with AWS Marketplace availability. The model features three main components: a distilled version from Llama 405 billion, a model trained with Arcee's EvolKit repository, and a model with exhaustive DPO on top of Llama 3 instructions.
Arcee AI claims that SuperNova offers superior instruction adherence and adaptability to specific business needs and addresses key concerns around data privacy and model stability. The company also highlights the model's ability to be fine-tuned and retrained within the enterprise environment, allowing continuous improvement and customization without data leaving the organization's system.
By using this site, you agree to allow SPEEDA Edge and our partners to use cookies for analytics and personalization. Visit our privacy policy for more information about our data collection practices.