In a significant step towards advancing its enterprise-ready AI and data platform, IBM has announced plans to integrate Meta’s Llama 2-Chat 70 billion parameter model into the watsonx.ai studio. This move will grant select clients and partners early access to the powerful AI model, marking a continuation of IBM’s collaboration with Meta on open innovation for AI.
The integration of Llama 2-Chat into the watsonx.ai studio is expected to build on the success of their joint efforts, which have included contributions to open source projects like the PyTorch machine learning framework and the Presto query engine utilized in watsonx.data.
The addition of Llama 2-Chat model aligns with IBM’s vision to provide a diverse range of AI models, both proprietary and third-party. Presently, within the watsonx.ai environment, AI developers can harness models from IBM and the Hugging Face community. These models come pre-trained to address a wide spectrum of Natural Language Processing (NLP) tasks, encompassing question answering, content generation, summarization, text classification, and extraction.
Anticipated to be a notable milestone on IBM’s path in generative AI, the inclusion of Llama 2-Chat in watsonx.ai will pave the way for forthcoming releases such as the AI Tuning Studio, additional AI models within the platform, and the introduction of FactSheets in watsonx.governance.
IBM’s commitment to upholding principles of trust and security remains a focal point as it continues to expand its generative AI capabilities. Notably, users engaging with the Llama 2-Chat model through the prompt lab in watsonx.ai will have access to the AI guardrails function, a feature designed to automatically eliminate harmful language from both input prompt text and the generated output. Meta has also provided insights into the fine-tuning methodology employed in the development of their expansive language models.
IBM Consulting has over 21,000 data, AI, and automation consultants. This is complemented by the Center of Excellence for Generative AI, consisting of more than 1,000 specialized consultants who stand ready to collaborate with clients. Their goal is to fine-tune and operationalize models for specific use cases, tailored to individual business requirements.
Both IBM and Meta share a steadfast commitment to open innovation. Recognizing the value of a diverse community of AI builders and researchers, they actively encourage testing, feedback sharing, and collaborative efforts to drive innovation further. The unveiling of Llama 2 and other upcoming models on the watsonx platform is eagerly anticipated, with industry insiders excited to witness the innovative solutions that will emerge from this partnership.