Computing giant IBM has broadened its catalogue of artificial intelligence (AI) large language models (LLMs), releasing the Granite 3.0 family of AI models.
Announced today at the company’s annual TechXchange event, the release comes amid the explosion of interest and investment in generative AI.
Studies also show organisations are embracing generative AI in their enterprise strategies, to drive significant results and achieve sustainable impacts on their bottom line.
According to IBM, Granite 3.0 – its third-generation Granite flagship language models – are built for business and aligned to the company’s commitment to open-source AI.
The Granite 3.0 release comprises dense general purpose LLMs, namely Granite 3.0 8B Instruct, Granite 3.0 2B Instruct, Granite 3.0 8B Base and Granite 3.0 2B Base.
Further, the LLM-based input-output guardrail models are made up of Granite-Guardian-3.0-8B and Granite-Guardian-3.0-2B.
The Granite 3.0 language models are trained on Blue Vela, powered by 100% renewable energy, says the company.
The models were trained on over 12 trillion tokens on data taken from 12 different natural languages and 116 different programming languages, using a novel two-stage training method.
By the end of the year, the 3.0 8B and 2B language models are expected to include support for an extended 128K context window and multi-modal document understanding capabilities, reveals the company in a statement.
“The new Granite 3.0 8B and 2B language models are designed as ‘workhorse’ models for enterprise AI, delivering strong performance for tasks such as retrieval augmented generation, classification, summarisation, entity extraction and tool use.”
IBM notes the models are designed to be fine-tuned with enterprise data and integrated across diverse business environments or workflows.
Additionally, the company provides an IP indemnity for all Granite models on watsonx.ai.
“While many LLMs are trained on publicly available data, a vast majority of enterprise data remains untapped. By combining a small Granite model with enterprise data, especially using alignment technique InstructLab – introduced by IBM and RedHat in May – businesses can achieve task-specific performance that rivals larger models.”
Share