Pamela Fox’s Post

GitHub Models (LLM /SLM catalog) launched! We have just lunched new super-cool feature on GitHub, that provide access to popular LLMs /SLMs like GPT-4o, GPT-40 mini, LLaMa 3.1, Phi-3, and others. They can be tested for free in the built-in Playground as well. Why not give it a try? It's a seamless process for developers to efficiently test different models." #GitHub #LLM #SLM #GenAI

Check out our RAG demo

Graph RAG is now available on GitHub! https://lnkd.in/g9gVP2nX I'm working on various Project of AOAI RAG, so understand Graph RAG from video and blog of Microsoft Research. Graph RAG has the ability to perform any question that requires a holistic understanding "all the content". Typical vector rag is excellent at performing for highly specific facts. => ex. searching about "one" thing. Graph RAG is excellent at performing for holistic understanding of all the content. => ex. searching and finding "abstract" from all content. https://lnkd.in/gRiYQ6fb #Azure #AOAI #RAG #GPT

🤓 If you enjoyed my DataCamp courses, you're going to love this next step in your learning journey: a live, online, 4-week course on Maven: "Building LLM Applications for Data Scientists and SWEs." I'm stoked to be teaching it with Stefan Krawczyk (Dagworks, ex-StitchFix). 🤖 Most of you are stuck in GenAI "proof of concept purgatory," and it doesn’t have to be that way. Stefan and I will teach you the entire software development lifecycle (illustrated wonderfully below by Hamel H.): how to prototype, test, evaluate, iterate, and ship LLM-powered software. This course will include:  🛠️ Hands-on sessions building and using LLM applications  🔄 Iteration and experimentation to improve GenAI applications  🗣️ Group discussions and reflection on the building process and the end-user experience  ❓ Q&As with expert practitioners, including guest lectures from our network (watch this space for some exciting people)  💬 A private Discord group to chat with myself, Stefan, and more about DS, ML, and AI  ☁️ Over $1,000 in cloud and compute credits (thanks Charles Frye and Modal!) We have an early bird discount for our January cohort: it's $800 until the end of November. If you'd like a further 25% off, DM me for a coupon (and with over $1K in cloud/compute credits, now you're making money 😉). You can sign up here: https://lnkd.in/gBKePdph We’re also hosting a free lightning lesson next week to demo some of the course content. You can sign up for that here: https://lnkd.in/g9nkZ5ur Let’s move beyond the prototype stage and start building robust and reliable GenAI applications! 

🚀 Excited to share my latest project on GitHub: Deployment with MLOps and MLFlow! In this repository, I demonstrate proficiency in deploying machine learning models using MLFlow, Dagshub, AWS S3, and MLOps principles. Whether you're new to MLOps or looking to streamline your deployment process, this project offers practical insights and examples to help you succeed. Key Features: - MLFlow Integration: Track experiments, log parameters, and manage model versions effortlessly. - Dagshub Collaboration: Enable version control and collaboration across machine learning projects. - AWS S3 Integration: Store artifacts and datasets securely, ensuring scalability and accessibility. - MLOps Best Practices: Follow best practices to streamline the machine learning lifecycle from development to deployment. Now, with the addition of running models like Linear Regression and Support Vector Regressor models, you can compare their performance using MLFlow UI, providing valuable insights for model selection and optimization. Check out the repository and dive into the world of MLOps: https://lnkd.in/d66WdZk4. Feel free to contribute, share, and let me know your thoughts! Your feedback is highly appreciated. Let's advance the field of machine learning together! 💡 #MLFlow #MLOps #MachineLearning #GitHub #OpenSource #DataScience

Glad to share my final project in the MLOps Zoomcamp #mlopszoomcamp course, by DataTalksClub (and the amazing team lead by Alexey Grigorev). My understanding of machine learning operations has been greatly increased by the new information I've gained. You can find the “Cyberbullying Detection” project on my GitHub here: https://lnkd.in/d46c_ZfZ. This project intends to promote proactive efforts towards combatting this by creating models to automatically flag potentially harmful tweets and break down the patterns of hatred. Project highlights: - Procured infrastructure on #AWS with #Terraform, setup #MLflow for experiments tracking and model registry - Set up #Prefect for model training pipeline deployment, unit tests and integration tests for quality testing - The model deployment code is containerized by #Docker, deployed in streaming services by using #Lambda, #Kinesis and #ECR, to flag potentially harmful tweets automatically and send warnings to users - Utilized #Evidently to calculate metrics and #Grafana to build a dashboard for model performance monitoring

LLMs are a complicated space! I have started an open-source repo to track frameworks, tutorials, and resources to help everyone navigate it. Contributions are much appreciated! https://lnkd.in/gihC5jvP #artificialintelligence #machinelearning #llms

Over the past several months, Patrick Hall and I have continued building out the Awesome Machine Learning Interpretability GitHub repository -- https://lnkd.in/eb5UD-YH. We're always on the lookout for new items to add, so please feel free to contribute! If you're searching for a wide array of resources in the responsible AI space, here's an overview of what we have so far: 🌐 Community and Official Guidance Resources - Community Frameworks & Guidance: 161 - Infographics & Cheat Sheets: 42 - AI Red-Teaming & Generative AI Explainability: 44 - Conferences & Workshops: 91 - Official Policies & Frameworks: 140 - Law Texts & Drafts: 22 📚 Educational Resources - Software Examples & Tutorials: 19 - Free-ish Books: 30 - Glossaries & Dictionaries: 46 - Open-ish Classes & Podcasts: 36 📝 AI Incidents, Critiques, and Research Resources - AI Incident Info Sharing: 10 - Bibliography of AI Incidents & Failures: 30 - AI Law, Policy, & Guidance Trackers: 14 - Challenges & Competitions: 4 - Critiques of AI: 94 - Curated Bibliographies and Lists: 37 - In particular, as its editor, I always recommend the AI Incident Database: https://lnkd.in/eH-9xriX 🛠 Technical Resources - Benchmarks: 23 - Common Datasets: 18 - ML Environment Management Tools: 7 - Personal Data Protection Tools: 1 - Responsible AI Python Packages: 153 - Responsible AI R Packages: 43 - Responsible AI C/C++, JavaScript, and Browser-based Packages: 12 Visit us at: https://lnkd.in/eb5UD-YH #MachineLearning #AI #ResponsibleAI #AIExplainability #AIPolicy #OpenSource #AIResearch #TechEthics #MLInterpretability #TrustworthyAI #GitHub

🚀 𝐁𝐮𝐢𝐥𝐝𝐢𝐧𝐠 𝐚 𝐐𝐮𝐞𝐬𝐭𝐢𝐨𝐧-𝐀𝐧𝐬𝐰𝐞𝐫𝐢𝐧𝐠 𝐒𝐲𝐬𝐭𝐞𝐦 𝐰𝐢𝐭𝐡 𝐀𝐳𝐮𝐫𝐞 𝐀𝐈! 🚀 I'm excited to share my latest project where I’ve integrated Azure Cognitive Services to build a Question Answering Solution! This Python-based application allows users to ask questions, and the system provides intelligent, confident answers using Azure's powerful AI capabilities. 🔍 𝐖𝐡𝐚𝐭 𝐢𝐭 𝐝𝐨𝐞𝐬: Uses Azure Cognitive Services' Question Answering API. Retrieves and displays answers along with confidence scores and sources. Simple to set up with configuration through a secure .env file for API credentials. I’m using this project as a hands-on way to learn and implement AI solutions for real-world applications! 💻🤖 🔗 𝐅𝐞𝐞𝐥 𝐟𝐫𝐞𝐞 𝐭𝐨 𝐜𝐡𝐞𝐜𝐤 𝐨𝐮𝐭 𝐭𝐡𝐞 𝐩𝐫𝐨𝐣𝐞𝐜𝐭 𝐨𝐧 𝐆𝐢𝐭𝐇𝐮𝐛: https://lnkd.in/gyZAT65J #Azure #AI #QuestionAnswering #Python #MachineLearning #AzureAI #TechInnovation #GitHub #ProjectShowcase Microsoft Azure ICT Group ICT Academy AI Engineer

Gen AI: RAG with AWS Bedrock Retrieval Augmented Generation (RAG) pipeline using AWS Bedrock. This includes setting up FAISS for vector-based document retrieval and leveraging the Bedrock LLM and embeddings from Amazon's Titan model. Solution integrates Langchain to enable effective question-answering systems by retrieving context from a vector store and utilizing prompt templates. Additionally, created a document ingestion process that loads and processes PDFs, transforming them into vector embeddings for efficient querying. This setup supports scalable and efficient information retrieval and summarization using LLMs. #OpenAI #RAG #Langchain #AWS Github repo :