IBM recently unveiled its plans to host Meta’s Llama 2, a 70 billion parameter model, on its AI and data platform, watsonx.ai. This initiative is part of IBM’s ongoing collaboration with Meta and its commitment to offering both proprietary and third-party AI models on watsonx.
Expanding the Reach of AI with Open-Source Models
This development follows IBM’s recent launch of its geospatial foundation model on the open-source AI platform, Hugging Face. Created using NASA’s satellite data, this model represents IBM’s efforts to democratize access to cutting-edge AI technologies.
Llama 2: Fostering Responsible AI Use
Llama 2, Meta’s latest open-source AI model, seeks to encourage responsible and safe AI usage within the tech industry. This model, trained on publicly available online data sources, comes with a comprehensive user guide outlining best practices for responsibly creating products powered by Large Language Models (LLMs).
Building Trustworthy AI in an Open Environment
IBM’s watsonx platform enables businesses to construct and adapt AI models using their own data in a reliable, open setting. As part of this initiative, IBM will host Meta’s Llama 2 model in the watsonx.ai studio, working on open-source projects developed by Meta.
Keeping Trust and Security at the Forefront
Both IBM and Meta prioritize trust and security in AI development. By utilizing watsonx.ai, users can effortlessly filter out offensive language from their input prompts and model-generated outputs when running the Llama 2 model. This automated process ensures a safer and more refined experience, maintaining the original meaning while enhancing the overall quality of the content. Meta also discloses its fine-tuning methodology for large language models (LLMs), underscoring its commitment to transparency.
The Imperative of Responsible AI Development
As more businesses incorporate generative AI capabilities into their operations, responsible and ethical AI development becomes increasingly critical. Collaborations between industry giants like IBM and Meta play a significant role in driving positive advancements in this area.
IBM’s Expertise in Generative AI
With a team of 21,000 data, AI, and automation consultants and a dedicated Centre of Excellence for generative AI, IBM Consulting is well-positioned to help clients tailor AI models to specific business needs.
Championing Open Innovation in AI
Both IBM and Meta champion open innovation in AI, recognizing the value of engaging with a community of builders and researchers to test, share feedback, and collaborate on digital transformation technologies.