IBM has unearthed plans for new generative AI foundation models and enhancements coming to watsonx – its AI and data platform that helps enterprises scale and utilize AI.
The enhancements include a technical preview for watsonx.governance, new generative AI data services coming to watsonx.data and the planned integration of watsonx.ai foundation models across select software and infrastructure products.
The new IBM and third-party generative AI models coming to watsonx.ai feature Granite series models. The models use “Decoder” architecture, which enables large language models (LLM) to predict the next word in a sequence. In addition, it can support enterprise NLP tasks like summarization, content generation and insight extraction.
IBM is also now offering Meta’s Llama 2-chat 70 billion parameter model and the StarCoder LLM for code generation in watsonx.ai on IBM Cloud.
The watsonx platform will see new capabilities being introduced with IBM’s release of the first iteration of its Tuning Studio in Q3. This capability will include prompt tuning – a low-cost way for users to adapt foundation models to their downstream tasks with their own enterprise data.
As part of the launch, IBM has introduced a synthetic data generator to assist users in creating artificial tabular data sets from custom data schemas or internal data sets.
IBM will look to implement watsonx.ai generative AI capabilities into watsonx.data to enable users to discover, augment, visualize and refine data for AI through a self-service experience. It also plans to integrate a vector database capability into watsonx.data to support watsonx.ai retrieval augmented generation use cases.
Dinesh Nirmal, senior vice president, products, IBM Software, said: “As demonstrated by the ongoing rollout of the watsonx platform within just a few months since launch, we are here to support clients through the entire AI lifecycle. As a transformation partner, IBM is collaborating with clients to help them scale AI in a trustworthy way – from helping to institute foundational elements of their data strategies to tuning models for their specific business uses cases to helping them govern models beyond that.”
