𝗙𝗿𝘂𝗴𝗮𝗹𝗶𝘁𝘆 𝗶𝘀 𝗻𝗼𝘁 𝗮𝗯𝗼𝘂𝘁 𝗯𝗲𝗶𝗻𝗴 "𝗰𝗵𝗲𝗮𝗽", 𝗯𝘂𝘁 𝗮𝗯𝗼𝘂𝘁 𝗱𝗼𝗶𝗻𝗴 𝗺𝗼𝗿𝗲 𝘄𝗶𝘁𝗵 𝗹𝗲𝘀𝘀. 💚 There's an edge of ingenuity and creativity to it that is often ignored. In the land of #GenAI, nothing represents the values of frugality better than the 𝘥𝘰-𝘮𝘰𝘳𝘦-𝘸𝘪𝘵𝘩-𝘸𝘦𝘢𝘬𝘦𝘳-𝘮𝘰𝘥𝘦𝘭𝘴 approach characteristic of #LLMCascades. In my new blog post, I'm putting LLM cascades on the spot and breaking them apart just for fun. Inside you'll learn how they work, when to use them and how to implement a #FrugalGPT-style LLM cascade on top of Amazon Bedrock ⛰️ using LangChain 🦜🔗 and LangGraph 🦜🕸️ to tackle one of the biggest challenges in the LLM world: 𝔥𝙖𝙡𝖑𝓊𝐜𝙞𝗻𝖆𝙩𝒾𝘰𝙣𝘴 🍄 https://lnkd.in/dw9WZ-rS // Don't forget to give it a thumbs if you're a big fan of frugality 👍 Kudos to Vectara for their HHEM model and making it available on Hugging Face 🙏 #QotD "I think frugality drives innovation, just like other constraints do.One of the only ways to get out of a tight box is to invent your way out." ―Jeff Bezos
João Galego’s Post
More Relevant Posts
-
From pioneering AI projects to leading ML teams in groundbreaking data science initiatives, my journey has been one of innovation and tangible impact. Projects like ethscore.net, where I leveraged ML to predict fraudulent Ethereum wallets, stand as testament to my commitment to invest in learning and practicing AI. However, amidst this dynamic landscape, I find myself challenged by the rapid evolution of LLM technology. Despite my hands-on experience in the domain, I am unable to keep up with the newest and the latest. And I wonder: What am I missing here? #AI #machinelearning #llm #openai
To view or add a comment, sign in
-
According to Anthropic's announcement, Claude 3.5 surpassed all the competitors' top-tier models. But does it actually perform better in an actual use case in our hands? Here is my quick experiment using LangChain's Agent framework. #llm #claude #gpt #genAI #langchain
Claude 3.5 vs. GPT-4: Performance Comparison Using LangChain’s SQL Agent
medium.com
To view or add a comment, sign in
-
Hi #connections, I tried to explain why we cannot see the overlapping data even if we have passed overlapping chunk size while using RecursiveCharacterTextSplitter. tagging my connections Vetrivel PS Sunny Savita Boktiar Ahmed Bappy sudhanshu kumar Nitin Jadhav Tania Laha Abhijith Reddy .Rajdeep Mallick Shajee Pulukkool Ashwini Ugale Rohit Madas Arijit Shingha Roy Eeshani Agrawal Subha Karmakar. Sinchini Dhar Maria Vechtomova Raphaël Hoogvliets Himanshu Ramchandani #ai #generativeai #langchain #textsplitter #chunking #datascience #datanalysis
To view or add a comment, sign in
-
30 Seconds me PDF Q&A System Banayein | LangChain & Gemini AI Tutorial #AICoding #LangChain #GeminiAPI #PDFProcessing #Shorts #AITutorial #CodingTips
To view or add a comment, sign in
-
Part 2 of the Plotly AI Agent for multiple Stock Sentiment comparison. Grab the latest news articles on Multiple stock, and visualize the provided sentiment and relevance of each stock with a Plotly AI agent. We’ll use #Plotly, Dash, #Groq, #Llama3, #Lanchain, and #AlphaVantage. I was using Llama3 with Groq, but I was wondering if any of you had better experiences with other LLMs? Thank you Troy Scribner for the the code in the Alphavantage Queries file.
To view or add a comment, sign in
-
Excited to share that I've completed the 2024 Google Machine Learning Bootcamp, starting with Andrew Ng’s Deep Learning Specialization and a challenging Kaggle Featured Competition, where I am currently ranked within the top 12% globally. 🚀 For the Gemma Sprint, I fine-tuned the Gemma-2 9 bilion model into a crypto agent, using LoRA and 4-bit quantization to optimize it for interpreting and translating crypto news. This specialized model is key for providing real-time, accurate analysis in the fast-moving crypto space. #GemmaSprint Hugging Face Website: https://lnkd.in/gCjzVbst Looking forward to more ML adventures ahead! 💡 #MachineLearning #Kaggle #GoogleBootcamp #DeepLearning #GemmaSprint
To view or add a comment, sign in
-
Hardest problem of AI LLM is solved!! Another sign that LLM progress has not hit brick wall. In short term, we overestimate but in long term we underestimate technology advancement. Noam Brown from OpenAI recently said in an interview how much he underestimated that test-time compute could be achieved quickly. In a conversation with Ilya, he estimated that the TTT solution would take around 10 years, but in the end it was 2-3 years. And he believes that TTT was the hardest problem to solve. There are other challenges, but none come close to solving TTT. There are few experts who know what they are talking about better than OpenAi's Noam Brown. So we can safely say that we have the hardest part behind us. And progress will continue for a long time to come.
To view or add a comment, sign in
-
🚀 Excited to share my latest project! 🎉 I've built a cutting-edge LLM model that can read multiple engineering PDFs and provide precise answers to your questions. Utilizing Google Gemini, Streamlit, and the LangChain framework, this tool revolutionizes how we interact with technical documents. #AI #MachineLearning #Engineering #GoogleGemini #Streamlit #LangChain #Innovation #Tech
To view or add a comment, sign in
-
🌟 Day 21 - #Geekstreak2024 🌟 Today's challenge was Count Inversions, a problem that evaluates how far an array is from being sorted. This classic question emphasizes understanding inversion counts, showcasing the importance of divide-and-conquer strategies like merge sort for efficient O(nlogn) solutions. I successfully tackled the problem, finding the number of inversions that indicate the "disorder" in an array. Each challenge like this sharpens my algorithmic thinking and prepares me for real-world problem-solving! 🚀 #CountInversions #ProblemSolving GeeksforGeeks #Algorithms #DataStructures #100DaysOfCode
To view or add a comment, sign in
-
My LangChain course just crossed 2,000 learners and has a 4.9/5 rating! Thanks to everyone who has taken it and the 2,700+ who bookmarked it...what are you waiting for? Finish the damn thing! It's a fairly comprehensive course, covers a lot of ground, and goes deep into prompting techniques and strategies. Plus, you get to spend ~6 hours with me and my sexy voice. Who doesn't want that? My LlamaIndex course will be publishing soon, too. Keep an eye out for that. Check out the course here 👉🏼 https://lnkd.in/gyTvnV-q #langchain #generativeai #artificialintelligence
To view or add a comment, sign in