Our Terms & Conditions | Our Privacy Policy
IBM introduces Granite 3.0; Aims to upgrade enterprise AI solutions for organizations – Digital Transformation News
Amidst organizations widely incorporating artificial intelligence (AI), IBM also brings an enhanced enterprise AI model. Named Granite 3.0, the latest AI model by IBM aims to provide autonomy and flexibility to its users.
According to an official blog Granite 3.0 brings a range of sophisticated AI models specially curated for enterprise applications. The company further added that the new Granite model is specially designed for enterprises. The new model is IBM’s third generation Granite flagship language models.
Decoding Granite 3.0
IBM said that the Granite models are released under the permissive Apache 2.0 license. This not only makes it unique but is also a better version of the older version of Granite. The new model is made to perform tasks such as classification, Retrieval Augmented Generation (RAG), summarisation, entity extraction and tool use. These compact and versatile models are designed to match with enterprise data and get integrated across diverse business environments and workflows.
LTIM deploys AI use cases for 40 top clients
Companies make a beeline to set up RE storage capacities
Nissan will launch an affordable EV SUV in India by 2026, aiming to price it competitively with ICE cars
Mental health is a top priority for Gen Z, report revealed
This upgraded version of Granite works on ‘real open source enterprise AI’ the company explained.Granite 3.5 will include three models. These includes the following:
- General Purpose/Language: This includes the Granite 3.0 8B Instruct, Granite 3.0 8B Base, Granite 3.0 2B Instruct and the Granite 3.0 2B Base
- Guardrails & Safety: It includes the Granite Guardian 3.0 8B and the Granite Guardian 3.0 2B
- Mixture-of-Experts: This will consists of the Granite 3.0 3B-A800M Instruct, Granite 3.0 3B-A800M Base, Granite 3.0 1B-A400M Instruct and the Granite 3.0 1B-A400M Base
The way ahead
The Granite 3.0 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, leveraging results from several thousand experiments designed to optimize data quality, data selection, and training parameters The company aims to improve some specific sectors. These include application development, IT automation, customer service, Business Process Outsourcing (BPO)and cybersecurity.
Furthermore, “Our new Granite models address this by offering truly open and cost-efficient AI solutions, optimized for tasks like RAG, classification, summarization, and entity extraction. Combined with InstructLab, Granite 8B models enable businesses to integrate enterprise data into LLMs at up to 23x cost savings in areas such as finance, HR, and customer support.”, Viswanath Ramaswamy, VP, Technology, IBM India & South Asia, concluded.
Follow FE Tech Bytes on Twitter, Instagram, LinkedIn, Facebook.
Images are for reference only.Images and contents gathered automatic from google or 3rd party sources.All rights on the images and contents are with their legal original owners.
Comments are closed.