Computing big IBM just lately introduced {that a} distilled model of DeepSeek-R1 could be deployed on watsonx.ai, an enterprise-grade AI growth studio that facilitates the creation, coaching, and deployment of AI fashions inside enterprise purposes. Builders can add and deploy customized basis fashions to utilise watsonx.ai’s inference capabilities.
The corporate additionally outlined an in depth stepwise process to assist customers deploy distilled variants of DeepSeek-R1 for inference.
“The discharge of DeepSeek-R1, an open-sourced reasoning mannequin on par with OpenAI’s o1 collection of fashions, is a big step on this route,” IBM mentioned within the announcement. “We hope that it’s going to encourage different mannequin suppliers to observe go well with and contribute to the expansion of the open-source AI ecosystem.”
On Wednesday, IBM additionally introduced its earnings outcomes for the fourth quarter of 2024. The corporate reported a income of $17.6 billion (up 1%), whereas its software program income grew by 10%. IBM’s gross revenue margin elevated by 40 foundation factors to 59.5%. Nonetheless, consulting income was down 2%, and infrastructure income was down 8%.
“Our generative AI ebook of enterprise now stands at greater than $5 billion inception-to-date, up practically $2 billion quarter over quarter,” mentioned Arvind Krishna, CEO of IBM. He additionally revealed that about eight factors of development had been organically led by the corporate’s AI merchandise, together with IBM Live performance and different AI assistants.
The corporate additionally expects a income development of no less than 5% for 2025, together with a free money stream of round $13.5 billion.
“We’re assured that our funding in partnerships and abilities, in addition to our early management in GenAI, place us to speed up consulting development as we transfer ahead,” Krishna added.
Final November, the corporate launched a brand new variant of their open-source AI mannequin, the Granite 3.0. Armand Ruiz, VP of product at IBM’s AI platform, publicly disclosed the datasets on which the mannequin is skilled. This can be a observe IBM has adhered to even previously with new mannequin releases.
“That is true transparency. No different LLM supplier shares such detailed details about its coaching datasets,” Ruiz additional mentioned.
The submit IBM Allows DeepSeek R1 on watsonx.ai appeared first on Analytics India Journal.
