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

Introducing the AI Evaluation Studio – Yes, we're shipping NON-STOP, folks! This is the NEW IBM 🚀 This new capability, part of watsonx.governance, allows you to: 𝟭. 𝗘𝗺𝗽𝗼𝘄𝗲𝗿 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿𝘀 𝗮𝗻𝗱 𝗠𝗼𝗱𝗲𝗹 𝗘𝘃𝗮𝗹𝘂𝗮𝘁𝗶𝗼𝗻𝘀 This tool helps reduce the time and effort spent on manually reviewing individual AI assets, paving the way for rapid experimentation in AI development. 𝟮. 𝗙𝗮𝘀𝘁𝗲𝗿 𝗧𝗶𝗺𝗲 𝘁𝗼 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 With actionable dashboards that perform assessments and provide results based on quantitative criteria, helping users select the best asset for each specific AI use case and evaluate multiple methods efficiently at scale. 𝟯. 𝗦𝗮𝗳𝗲𝗴𝘂𝗮𝗿𝗱 𝗬𝗼𝘂𝗿 𝗕𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗳𝗿𝗼𝗺 𝗘𝗺𝗲𝗿𝗴𝗶𝗻𝗴 𝗔𝗜 𝗥𝗶𝘀𝗸𝘀 Customers are exposed to new risks such as toxicity, hallucinations, copyright infringement, and prompt-based attacks, making AI governance essential for businesses aiming to scale AI adoption. Happy safe AI development and enterprise production!

The Granite 3.0 paper is a 46-page technical masterpiece. It is a must-read for everyone who cares deeply about the details. It goes over the following components to deliver Granite to the open-source community: - Overview of Granite 3.0 LLMs, targeting enterprise needs. - Model Architecture: Describes dense and Mixture-of-Expert (MoE) models for efficiency. - Training Data: Uses curated web and synthetic data to train models. - Pre-training: Two-stage process: broad data in Stage 1, refined in Stage 2. - Training Hyperparameters: Explains hyperparameter tuning methods for efficient training. - Post-training: Focuses on instruction fine-tuning and structured interaction templates. - Model Alignment: Uses techniques like PPO and BRAIn to align models with human preferences. - Model Merging: Combines specialized models for overall best performance. - Infrastructure, Energy Consumption, and Carbon Emissions: Describes training infrastructure and efforts to minimize carbon footprint. - Evaluation: Assesses model performance on various tasks and benchmarks. - Data: Details the datasets used for training and evaluation. - Socio-Technical Harms and Risks: Discusses potential risks like bias and misinformation. -Final Conclusion: Summarizes key results. 🤯 Link to full paper available here: https://lnkd.in/gdUikiHG Happy reading! ❓Share in the comments what your most surprising fact from the paper is.

So, as some of you might’ve heard, our EY team ( EYers) snagged second place in the global Hackathon organized by IBM at the #TechXchange conference! Pretty cool, right? Huge shoutout to to my teammates: Daniel, Łukasz. We even ran an AMA (Ask Me Anything) session about our idea. It was a great way to chat about #BussyBee, get some feedback, and answer a few burning questions. We also had the chance to get feedback from a speaker coach and a Toastmaster pro( thanks Paul). The tips were super helpful, and I mean, who knew one of them involved glasses? Congrats to the first place team (Tushar and Suhas) for their idea on boosting productivity in restaurants and the third place crew (Itelvina and others) for helping students make data-driven decisions about their futures! Once I’m back in Poland, I’ll share a full breakdown with numbers, outcomes, and what's next. Big thanks to everyone who had their fingers crossed for us! #Hackathon #SecondPlace #EYers

Rodrigo Andrade, Senior Product Manager, IBM, gives his take on the new Granite models and why even non-developers should take notice. Check it out here👇

Get an overview of our recently released Granite models here👇

Introducing the watsonx․ai Developer Toolkit — providing access to cutting-edge tools, APIs, and models to build #AI in a single intuitive, collaborative development experience. Designed to meet the needs of both novice and experienced developers, it facilitates faster development cycles and a smoother path to deployment for a variety of AI applications, including agents and RAG-based solutions. Discover how the Developer Toolkit can accelerate your AI initiatives: https://ibm.biz/BdaGL7

Did you know that IBM Granite LLM is now available through the Generative AI Hub in SAP AI Core? Learn more here👇

A product is only a good as the team behind it 💪

CEO Arvind Krishna on $IBM's 3Q 2024 earnings results. Read more: https://ibm.co/4fcgm2u