ZCS (Zeaware Consulting Services)’s Post

We are happy to announce MLflow 2.10.0 release!! 🚀 MLflow 2.10 includes several significant new features that are preparing the way for current and future enhanced support for Deep Learning use cases, new features to support a broadened support for #GenAI applications, and some quality-of-life improvements for the MLflow Deployments Server (formerly the AI Gateway). ✅ New MLflow Website - MLflow.org has been completely refreshed! The new site landing page is fresh, modern, and contains more content than ever. We're adding new content and blogs all of the time. ✅ Model Signature Supports Objects and Arrays - Objects and Arrays are now available as configurable input and output schema elements. These new types are particularly useful for GenAI-focused flavors that can have complex input and output types. ✅ Langchain Autologging🦜 - MLflow LangChain flavor has autologging support now! When you invoke a chain, with autologging enabled, we will automatically log most chain implementations, recording and storing your configured LLM application for you. . ✅ Prompt Templating for Transformers Models -The MLflow Transformers🤗 flavor now supports prompt templates. You can now specify an application-specific set of instructions to submit to your GenAI pipeline to simplify, streamline, and integrate sets of system prompts to be supplied with each input request. ✅ MLflow Deployments Server Enhancement - The https://lnkd.in/etuhUMie  now supports two new requested features: (1) OpenAI endpoints that support streaming responses. You can now configure an endpoint to return real-time responses for Chat and Completions instead of waiting for the entire text contents to be completed. (2) Rate limits can now be set per endpoint to help control cost overrun when using SaaS models.  ✅ Further Document Improvements - Enhanced documentation, guides, tutorials, and examples by expanding on core MLflow functionality as well as entirely new pages for GenAI flavors. For a comprehensive list of changes, see 👉 https://lnkd.in/eZHdA8-q #opensource #mlflow #mlops #machinelearning #oss #linuxfoundation

The latest MLflow brings noteworthy enhancements, such as automatic logging for LangChain pipelines, facilitating traceability and debugging at each stage.

The #prompt template and the #openai endpoint long waited features. I am looking forward to use these two and the others. #mlflow #mlops

🚀 Exciting Journey with ZenML, MLflow, and Streamlit! In the dynamic world of Machine Learning, managing and scaling ML projects can be quite daunting. Fortunately, tools like ZenML have made this a lot easier. I recently integrated ZenML with MLflow for seamless experiment tracking and model management, which has been a game-changer in maintaining model lineage and improving reproducibility. ZenML's modular design not only helped in orchestrating the workflows but also in scaling them effortlessly from a local setup to cloud environments. The Pythonic nature of ZenML made it incredibly easy to turn existing code into reusable and versioned pipeline components. To bring our models closer to the stakeholders, I utilized Streamlit to create interactive dashboard that provides live predictions. This allowed us to demonstrate the real-time capabilities of our models, making them more tangible and understandable for everyone involved. Here’s a quick breakdown of how I did it: Step 1 - #ZenML: Begin with ZenML to set up and orchestrate your ML pipelines. Step 2 - #ZenML + #MLflow: Integrate MLflow with ZenML for robust experiment tracking and model versioning. Step 3 - #Streamlit: Use Streamlit to deploy a live interactive application that provides real-time predictions from your models. Integrating these tools has not only streamlined the workflow but also provided me with invaluable insights into managing complex processes with ease. This experience has truly underscored the power of efficient tool integration in simplifying operational complexities. Would love to hear your thoughts or experiences with these tools! #MachineLearning #MLOps #ZenML #MLflow #Streamlit #DataScience

Have you ever wanted to add features to applications - but you don't have access to the source code? You won't want to miss our re:Invent Builders' Session! You will learn how to: ✅ react to changes in the application's database in an event-driven fashion ✅ automatically update a knowledge base with new content for Retrieval Augmented Generation (RAG) ✅ build an AI powered chatbot that answers questions on the content of your application Absolutely stoked to deliver "𝗠𝗔𝗠 𝟯𝟮𝟲 - 𝗕𝘂𝗶𝗹𝗱 𝗻𝗲𝗮𝗿-𝗿𝗲𝗮𝗹-𝘁𝗶𝗺𝗲 𝗟𝗟𝗠 𝗽𝗼𝘄𝗲𝗿𝗲𝗱 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀 𝗳𝗼𝗿 𝗺𝗼𝗻𝗼𝗹𝗶𝘁𝗵𝗶𝗰 𝗮𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀" together with my dear colleagues Adrian Begg, Gaurav Arora, Marco Sommella and Franz Stefan in Las Vegas. Register for our session, bring your laptop and learn how to enhance existing applications that are black boxes to you! 🗓 Date and time: December 4th, 01:30 PM 📍 Location: MGM Grand Hotel ℹ Session details: https://lnkd.in/e_mA9iTk #AWS #AWSreInvent #modernization

Extend Power of AutoGen with Promptflow. Multi-agent framework is the next evolution of single agent chat bots and RAG (Retrieval Augmented Generation) applications. The Autogen framework is currently the most popular for multi-agent use cases. Promptflow is a tool that simplifies the process of building, evaluating, and deploying Language Learning Model (LLM) applications. This blog will guide you on how to build Autogen workflows with Promptflow.   What is Autogen?   AutoGen is a framework developed by Microsoft Research that simplifies the orchestration, optimization, and automation of workflows for large language models (LLMs).... #techcommunity #azure #microsoft https://lnkd.in/gUCEq4tb

TensorFlow has a comprehensive, flexible ecosystem of tools, libraries, and community resources that lets researchers push the state-of-the-art in #ML and developers easily build and deploy ML-powered applications. #MLflow has built-in support (we call it MLflow Tensorflow flavor) for Tensorflow workflow, at a high level in MLflow we provide a set of APIs for: ✅ 𝗦𝗶𝗺𝗽𝗹𝗶𝗳𝗶𝗲𝗱 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁 𝗧𝗿𝗮𝗰𝗸𝗶𝗻𝗴: Log parameters, metrics, and models during model training. ✅ 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁𝘀 𝗠𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁: Store your Tensorflow experiments in MLflow server, and you can view and share them from MLflow UI. ✅ 𝗘𝗳𝗳𝗼𝗿𝘁𝗹𝗲𝘀𝘀 𝗗𝗲𝗽𝗹𝗼𝘆𝗺𝗲𝗻𝘁: Deploy Tensorflow models with simple API calls, catering to a variety of production environments. #machinelearning #mlflow #mlops #deployment

Generative AI on AWS & Innovate faster with new capabilities. Generative AI–powered assistant to accelerate software development and business operations. Pick your choice of industry-leading foundation models (FMs), and get the most cost-effective infrastructure. #arravo #it #asw #cloudcomputing #cloudservices #generativeai

Deploying ML Models in Kubernetes: Best Practices Get the most out of the machine learning in Kubernetes! The latest post shares expert tips and best practices for the deploying ML models in Kubernetes, covering containerization and model serving. Check it out and take your ML skills to the next level! #MachineLearning #Kubernetes #Deployment Looking to streamline your ML workflow? Learn how to leverage the Kubernetes for efficient model deployment, management, and scaling. Read our latest post for insights on the containerization, model serving, and more! Useful Links to share: - Check out Kubeflow, an open-source toolkit for machine learning workflows in Kubernetes: https://lnkd.in/dY74V373 - Learn more about MLflow, a platform for managing the ML lifecycle: https://lnkd.in/dN9kyhuS

Moving from AIOps hype to AIOps hero - – that's our mission with stAIble.💪 In our "From Hype to Hero" series, we're diving into the nitty-gritty of AIOps. First stop: topology mapping. This crucial task involves creating a comprehensive visualization of your organization's IT infrastructure. By automatically discovering and mapping the relationships among your apps, servers, networks, and storage, AIOps paints a holistic picture of your entire system. This empowers IT teams to: ✅ Gain insights into how different elements interact ✅ Identify potential bottlenecks ✅ Optimize resource allocation ✅ Facilitate effective incident management We see that most companies have a topology, but it’s often outdated and that’s exactly where AIOps can help! Reach out and we’ll be happy to tell you all about it. ✉️ #AIops #topology #automation