IBM watsonx

Introducing the next generation of watsonx Code Assistant – helping developers code smarter, not harder. 💻 Powered by IBM Granite code models, this new generative #AI tool helps simplify software development and application modernization at the enterprise level. See how it can help you and your business: https://ibm.co/3Yhsb0x

We're excited to announce the general availability of the API Assistant feature in IBM API Connect, powered by IBM watsonx. 🙌 See how this new virtual assistant can help developers create better APIs faster: https://ibm.co/3AyQ6AO

Let us know below 👀

At IBM, we’re always excited to see how our partners are driving innovation 💡 and transforming industries with the power of AI #watsonx. One fantastic example is Tutore, a dynamic educational platform serving over 335,000 students across Poland 🇵🇱 . Working with our partner TUATARA, Tutore implemented a virtual assistant—Tutorek—powered by IBM watsonx™ Assistant #watsonx to enhance both student engagement and operational efficiency. In just 9 months, Tutorek has handled over 90,000 messages and resolved 95% of queries, allowing Tutore’s support staff to focus on high-value tasks and improving the overall student experience. This partnership is a testament to how AI can help organizations scale, innovate, and deliver real value. Congrats to the Tutore and TUATARA teams for this incredible success—this is exactly the type of impact we strive for! #ibmpartners #innovation #cocreation Piotr Laszczewski Marcin Gajdziński Maija Tanner Dejan Vukovic Evita Stoop Stefano Rebattoni Martin Svik Katie Axam Kevin Redmond Vincent R. Visser MBA Ewa Królik

Disclosing the full list of datasets used to train IBM LLMs Granite 3.0. This is true transparency - no other LLM provider shares such detailed information about their training datasets. WEB Data - FineWeb: More than 15T tokens of cleaned and deduplicated English data from CommonCrawl. - Webhose: Unstructured web content in English converted into machine-readable data. - DCLM-Baseline: A 4T token / 3B document pretraining dataset that achieves strong performance on language model benchmarks. CODE - Code Pile: Sourced from publicly available datasets like GitHub Code Clean and StarCoderdata. - FineWeb-Code: Contains programming/coding-related documents filtered from the FineWeb dataset using annotation. - CodeContests: Competitive programming dataset with problems, test cases, and human solutions in multiple languages. DOMAIN - USPTO: Collection of US patents granted from 1975 to 2023. - Free Law: Public-domain legal opinions from US federal and state courts. - PubMed Central: Biomedical and life sciences papers. - EDGAR Filings: Annual reports from US publicly traded companies over 25 years. MULTILINGUAL - Multilingual Wikipedia: Data from 11 languages to support multilingual capabilities. - Multilingual Webhose: Multilingual web content converted into machine-readable data feeds. - MADLAD-12: Document-level multilingual dataset covering 12 languages. INSTRUCTIONS - Code Instructions Alpaca: Instruction-response pairs about code generation problems. - Glaive Function Calling: Dataset focused on function calling in real scenarios. ACADEMIC - peS2o: A collection of 40M open-access academic papers for pre-training. - arXiv: Scientific paper pre-prints posted to arXiv. Full author acknowledgement can be found here. - IEEE: Technical content from IEEE acquired by IBM. TECHNICAL - Wikipedia: Technical articles sourced from Wikipedia. - Library of Congress Public Domain Books: More than 140,000 public domain English books. - Directory of Open Access Books: Publicly available technical books from the Directory of Open Access Books. - Cosmopedia: Synthetic textbooks, blog posts, stories, and WikiHow articles. MATH - OpenWebMath: Mathematical text from the internet, filtered from 200B HTML files. - Algebraic-Stack: Mathematical code dataset including numerical computing and formal mathematics. - Stack Exchange: User-contributed content from the Stack Exchange network. - MetaMathQA: Dataset of rewritten mathematical questions. - StackMathQA: A curated collection of 2 million mathematical questions from Stack Exchange. - MathInstruct: Focused on chain-of-thought (CoT) and program-of-thought (PoT) rationales for mathematical reasoning. - TemplateGSM: Collection of over 7 million grade-school math problems with code and natural language solutions. BOOM!

How IBM is simplifying #GenAI across the stack for CIOs 🤖Tune in now for our analyst highlights from #IBMARFORUM theCUBE.net #theCUBE is joined by Bruno Aziza, VP of product management, data and AI at IBM, to discuss how the company enables CIOs to progress with gen AI and achieve their goals. “If you look at the stats in the market, you'll find that 60% of CIOs are dissatisfied with the progress they've made with gen AI. The reality is that the world of the CIO today… is what I call the multi-world. You have to deal with multi-cloud and models of multiple sizes. So, you must have a small and a large model strategy,” Aziza shares. “You have multi-steps because you're going from assistants to agents, multi interfaces. For CIOs, it's very complex. The opportunity for us is to simplify across the stack. First, you're going to manage and work with data, then orchestrate and activate it. What matters the most is working with a partner that doesn't have a horse in the race of which cloud or model you're going to choose, but is aligned with your goals as a CIO,” he adds. 📺 Tune in now for more insights! https://lnkd.in/gDrYabYM 🧠 Learn more at: https://lnkd.in/gk6nByuN #EnterpriseAI #SmallLanguageModels #Multicloud #DataManagement #CIOinsights #AI #IBM IBM Partner Plus

#ad Controlling a robot with my hand was a NEW experience! I think I did pretty good? I recently attended #IBMTechXchange in Las Vegas, where IBM made a significant announcement about Granite 3.0, which was released with an Apache 2.0 open-source license, offering faster performance, greater cost-efficiencies, and a focus on responsible AI. During the conference, I got to check out various innovative demonstrations, including a Python-controlled robot that I could manipulate through hand movements, which was actually easier to control than I expected! I also get to attend some hands-on labs and sessions, with a big focus on the transparency of Granite 3.0. Which I was really excited about since I really love coding and building with Granite already. If you haven’t used Granite yet, make sure to go check it out and let me know what you build! #IBM Learn more here: https://ibm.biz/Bdapbb

Why should devs trust AI models that are open source vs. closed systems? Listen to host Albert Lawrence and guest JJ Asghar, Developer Advocate at IBM, as they discuss how open-source models could be the future of security and governance: https://ibm.biz/BdapmU

How can you use Granite Guardian models to manage AI safety? Find out in our latest episode of Mixture of Experts👇

Missed out on the IBM TechXchange 2024 conference last week? IBMer Ayal Steinberg has you covered. From the launch of IBM Granite 3.0 foundation models to unveiling the new watsonx Code Assistant, our Las Vegas event showcased how #AI is actively transforming the ways we work. Curious to learn more? Explore more announcements from TechXchange here: https://ibm.biz/BdabC9