Stability AI has just unleashed their latest gift to the world of AI – say hello to StableLM, Stability ai’s breakthrough language models. But what sets StableLM apart from other language models? Read on to find out.
Stability AI, a leading artificial intelligence (AI) company, has released a new open-source language model called StableLM. The model is available in two versions – 3 billion and 7 billion parameters and will be followed by larger 15 billion to 65 billion parameter models. StableLM is freely available for commercial or research purposes, subject to the terms of the CC BY-SA-4.0 license.
StableLM models can generate text and code, and they will power a range of downstream applications. The models are built on a new experimental dataset that is three times larger than The Pile, with 1.5 trillion tokens of content. Despite its small size, StableLM demonstrates surprisingly high performance in conversational and coding tasks, outperforming GPT-3 with its 175 billion parameters.
In addition to the base models, the company is also releasing a set of research models that are instruction fine-tuned. These models are intended for research use only and are released under a noncommercial CC BY-NC-SA 4.0 license.
Stability AI has a strong commitment to transparency, accessibility, and supportiveness. They open-source their models to promote transparency and foster trust, and design for the edge so that everyday users can run their models on local devices. They also build models to support users, not replace them.
The release of StableLM builds on the company’s experience in open-sourcing earlier language models with EleutherAI, a nonprofit research hub. StableLM is trained on an experimental dataset built on The Pile but is three times larger with 1.5 trillion tokens of content. But this isn’t the only new AI in town, after Elon Musk announced plans to launch TruthGPT
The models are now available on their GitHub repository, and they will be publishing a full technical report in the near future. The company is also growing their team and looking for passionate and experienced LLMs to join them in democratizing access to this technology.
With every day a new AI popping up in the market, Google CEO Sundar Pichai’s concerns that AI can be harmful if deployed wrongly seem less of a far-fetched idea. As the industry moves forward it should pay close attention to unregulated growth so that things don’t go out of hand.
Source: Stability AI Blog
