Faeghe kashani’s Post

Random forest for open source LLMs! And it beats GPT-4 omni, significantly! Image allowing multiple LLMs to work together as agents to produce the best possible output. The results? It can actually match the performance of generalist models like GPT-4o , but with much lower costs and greater efficiency. This novel approach, called 'Mixture of Agents' (MoA), is introduced in a research paper by Together AI. MoA combines multiple agents (like AI systems, robots) with different strengths and weaknesses, similar to random forest in ML, but with LLMs (in this paper) instead of decision trees. As an agentic framework, MoA is more complex, but incredibly powerful when allowing specialized LLMs (each specializing in a specific domain or task) to work together, each doing what they're best at. Reference / Paper https://lnkd.in/g5E5MBXS GitHub https://lnkd.in/gtREPZ6X How Random Forest works: https://lnkd.in/gY9pkGZa _____________ I share my learning journey here. Join me and let's grow together. #harshitahluwalia #llms #generativeai

🧩Fill in the Copilot prompt: “How many _____ would it take to cover an area the size of _____?” Get creative and drop your completed version in the comments. ⬇️ Together we’ll picture what Copilot can do, using AI turn the results into a sharable image and featuring our favorites at #MSIgnite.

Thrilled to announce that I've developed an ML model named **ChatSphere** during the Guvi challenge! 🚀 This chatbot leverages advanced algorithms to provide engaging and insightful conversations. Excited to see where this journey takes me in the world of AI and machine learning! #MachineLearning #Chatbot #GuviChallenge

After many requests over the past couple of years, we’ve worked on expanding the license terms for the Arkitscene dataset. It’s now available for commercial use under certain conditions. This change is all about making the dataset more accessible so it can drive more innovation and impact across the industry. Checkout the detailed license terms on our GitHub page here: https://lnkd.in/gvd5yEKg #ARkit #3DSceneUnderstanding #AI #AR #MachineLearning #3DDataset

The tidychatmodels R package “provides a simple interface to chat with your favorite AI chatbot from R. It is inspired by the modular nature of tidymodels where you can easily swap out any ML model for another one but keep the other parts of the workflow the same.” Current support for OpenAI, Mistral.ai, and Ollama. By Dr. Albert Rapp on GitHub https://lnkd.in/e4xuPus4 #rstats #GenAI #LLM #LLMs #AI

Finally Completed My Hand Gesture Recognition Model. The Dataset I used was HAGRID Dataset and Model is YOLOv8. It detects 18 gesture types for now. A short demo video of it's performance. Still need a little bit of tweaking #MachineLearning #ML #ArtificialIntelligence #AI #YOLO #ObjectDetection #ComputerVision #DeepLearning #Model #NeuralNetworks

Random Forest for Open Source LLMs! 🤖 Better than GPT-4 Omni ?! Imagine allowing multiple LLMs to work together as agents to produce the best possible output. What would that result ? It can actually match the performance of generalist models like GPT-4, but with much lower costs and greater efficiency. This novel approach, called ‘Mixture of Agents’ (MoA), is introduced in a research paper by Together AI. How does MoA work? MoA combines multiple agents (like AI systems, robots) with different strengths and weaknesses, similar to a random forest in machine learning, but with LLMs instead of decision trees. As an agentic framework, MoA is more complex, but incredibly powerful when allowing specialized LLMs (each specializing in a specific domain or task) to work together, each doing what they’re best at. Check out the full paper here: https://lnkd.in/g5E5MBXS #AI #MachineLearning #NaturalLanguageProcessing #MoA #LLMs #Innovation #Research #TechAdvancements

Ready to train on your custom dataset? Here's a quick guide using Roboflow Prepare Dataset: Collect and label images on Roboflow. Download: Export your dataset in YOLO format. Set Up YOLOv10: Clone the YOLOv10 repo and configure paths. Train: Start training. Deploy: Use your model for real-time detection! Start now and share your results! 💡📷 Check out our tutorial for more details: https://lnkd.in/d8J6QyEg #YOLOv10 #ObjectDetection #MachineLearning #AI #Roboflow #TechInnovation

Only 1 Week Left—Don’t Miss Our Free Microsoft Copilot Webinar! On November 5th, we’ll be diving into how to go from “What do I do with this?” to “Wow, look at those results!” using Microsoft’s innovative new AI tool, Copilot. In this webinar, you’ll learn: - Easy ways to start using AI to boost your team’s productivity - Our 5-step strategy for adopting Copilot successfully - A live look at our Copilot Proof of Concept in action If you’ve been wondering how to make the most of this powerful tool, this is your chance! One week to go—register now and get ready to level up your AI game. 📅 Tuesday, November 5 at 12:00 PM CST 🔗 https://buff.ly/3NGlxfJ #MicrosoftCopilot #AI #ProductivityBoost #Microsoft365

I just completed The Missing Link Copilot course created by Matt Dunn, and I was impressed by the level of detail covered in the course. I learned several tips and tricks on how Copilot and AI can assist me with my day-to-day tasks. I highly recommend this course to anyone looking to explore how they can leverage AI to their advantage. https://lnkd.in/gqC-6Xvb