Stability AI has introduced StableLM, an open-source language Model that is poised to transform the field of AI. The model, currently available in an Alpha version with 3 billion and 7 billion parameters, is just the beginning, as Stability AI plans to release models with up to 65 billion parameters in the near future.
StableLM is not just any language model; it represents Stability AI’s commitment to making foundational AI technology accessible to all. Under the CC BY-SA-4.0 license, the model can be freely used and adapted for commercial or research purposes, making it a valuable tool for various applications.
This launch builds on Stability AI’s previous open-source initiatives, such as Stable Diffusion, an open and scalable image model released in 2022. StableLM is designed to generate text and code, enabling its use in downstream applications and demonstrating the ability of small, efficiently trained models to deliver high performance.
Stability AI’s experience with EleutherAI, a nonprofit research hub, has been influential in the development of StableLM. Based on a new experimental dataset that is three times larger than The Pile open-source dataset, StableLM delivers excellent performance in conversational and coding tasks.
In addition to StableLM, Stability AI is releasing research models that are fine-tuned using a combination of five recent open-source datasets for conversational agents. These research models, however, are only available for noncommercial use under the CC BY-NC-SA 4.0 license.
Stability AI champions transparency and believes in open-sourcing its models to promote understanding and trust. By allowing researchers and AI enthusiasts to explore the models, Stability AI encourages verification of performance claims and the development of interpretation techniques. This process also helps identify potential risks and promotes the development of proactive safeguards.
The open-source models from Stability AI are beneficial not only to the research community but also to organizations in the public and private sectors. These models can be customized to meet specific requirements without compromising data security or relinquishing control over AI capabilities.
Stability AI will continue to contribute to the collaborative and inclusive AI ecosystem by launching a crowd-sourced RLHF program and collaborating with community efforts like Open Assistant to create an open-source dataset for AI assistants.
For more information and to download the StableLM language model, visit Stability AI’s official website or their GitHub repository, which will be regularly updated with new checkpoints and patches.
The post Stability AI Unveils StableLM: A Groundbreaking Open-Source Language Model appeared first on TS2 SPACE.
