Dev-docs’ Post

You can now use Dev-docs to create contextual documentation that can automatically be used as context with Continue! 📚 Andrew van Beek shows you how here: https://lnkd.in/gVtS4d5U

Excited about our recent collab with Continue. Continue is the leading open-source AI code assistant that allows you to connect any model and any context to compose custom autocomplete and chat experiences inside the IDE. It's been awesome to see how Dev-docs' auto-generated markdown content allows for more context aware answers from Continue's code assistant. This creates a true 1+1=3 scenario. Here's to helping devs Continue to onboard and ship faster! (pun intended 🙃 )

🦑 Informative Demystifying Tech Terms: A Beginner’s Guide https://lnkd.in/d3iy3FV9

I'm delighted to unveil my latest venture into prompt engineering for dialogue summarization using the pre-trained FLAN-T5 (Text-to-Text Transfer Transformer) model from Huggingface (google/flan-t5-base). ✨ 📌 𝗣𝗿𝗼𝗷𝗲𝗰𝘁 𝗢𝘃𝗲𝗿𝘃𝗶𝗲𝘄: I've been exploring the power of prompt engineering in generating dialogue summaries using the pre-trained LLM FLAN T5 model from Huggingface. You can check out the project and dive into the code on GitHub: 𝗚𝗶𝘁𝗵𝘂𝗯 𝗟𝗶𝗻𝗸: https://lnkd.in/guzJ7iNu Built and Ran : 𝗔𝗺𝗮𝘇𝗼𝗻 𝗦𝗮𝗴𝗲𝗠𝗮𝗸𝗲𝗿 🚀 𝗪𝗵𝗮𝘁'𝘀 𝗜𝗻𝘀𝗶𝗱𝗲? Explore how prompt engineering takes dialogue summarization to the next level! I've used simple dialogues from the Dialogsum Huggingface dataset to showcase the effectiveness of prompt engineering. The notebooks provide the following scenarios. 1. 𝗗𝗶𝗮𝗹𝗼𝗴𝘂𝗲 𝗦𝘂𝗺𝗺𝗮𝗿𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝗪𝗶𝘁𝗵𝗼𝘂𝘁 𝗣𝗿𝗼𝗺𝗽𝘁 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 2. 𝗗𝗶𝗮𝗹𝗼𝗴𝘂𝗲 𝗦𝘂𝗺𝗺𝗮𝗿𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝘄𝗶𝘁𝗵 𝗭𝗲𝗿𝗼-𝘀𝗵𝗼𝘁 𝗣𝗿𝗼𝗺𝗽𝘁 𝗜𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 3. 𝗗𝗶𝗮𝗹𝗼𝗴𝘂𝗲 𝗦𝘂𝗺𝗺𝗮𝗿𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝘄𝗶𝘁𝗵 𝗢𝗻𝗲-𝘀𝗵𝗼𝘁 𝗣𝗿𝗼𝗺𝗽𝘁 𝗜𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 4. 𝗗𝗶𝗮𝗹𝗼𝗴𝘂𝗲 𝗦𝘂𝗺𝗺𝗮𝗿𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝘄𝗶𝘁𝗵 𝗙𝗲𝘄-𝘀𝗵𝗼𝘁 𝗣𝗿𝗼𝗺𝗽𝘁 𝗜𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 👉 𝗖𝗵𝗲𝗰𝗸 𝗶𝘁 𝗢𝘂𝘁! Dive into the code, play around, and witness the magic of prompt engineering in action. Don't miss out on the opportunity to understand its impact on dialogue summarization. 𝐹𝑒𝑒𝑙 𝑓𝑟𝑒𝑒 𝑡𝑜 𝑙𝑒𝑎𝑣𝑒 𝑦𝑜𝑢𝑟 𝑡ℎ𝑜𝑢𝑔ℎ𝑡𝑠, 𝑠𝑢𝑔𝑔𝑒𝑠𝑡𝑖𝑜𝑛𝑠, 𝑜𝑟 𝑞𝑢𝑒𝑠𝑡𝑖𝑜𝑛𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑐𝑜𝑚𝑚𝑒𝑛𝑡𝑠.  #GenerativeAI #PromptEngineering #AIInnovation #GitHub #DialogueSummarization #amazonsagemaker #TechInnovation #MachineLearning

Unlocking the Power of Optical Character Recognition with Ollama OCR In today’s digital world, extracting text from images efficiently is crucial. Enter Ollama OCR! 🚀 An open-source Optical Character Recognition (OCR) application that uses advanced vision language models to transform how we interact with text in images. Inspired by LlamaOCR, Ollama OCR combines cutting-edge technology with a user-friendly interface. ------ Why Choose Ollama OCR? 🤔 ------ High-Precision Text Recognition 🎯 LLaVA 7B: Fast and efficient for real-time processing. Llama 3.2 Vision: High accuracy for complex documents. These models ensure reliable text extraction from various image types!Versatile ------ Output Formats 📄 Ollama OCR supports multiple output formats:Markdown: Preserves formatting with headers and lists. Plain Text: Clean and simple extraction. JSON: Structured data for easy integration. Structured Data: Extracts tables and organizes information. Key-Value Pairs: Efficiently extracts labeled information. ------- User-Friendly Interface 🖥️ Key features include:Drag-and-Drop Upload: Simplifies image addition. Real-Time Processing: Instant results enhance workflow. Download Options: Save extracted text in various formats. Image Preview: Review images before extraction. Technical Excellence 💪 -------- Ollama OCR is built on a robust architecture using Streamlit for the frontend and Python for the backend, ensuring smooth operation and reliable performance. Comprehensive error handling enhances user experience.Join the Community 🌍 -------- As an open-source and free tool, Ollama OCR welcomes contributions from developers and enthusiasts. Your involvement can help shape the future of OCR technology! Transform your text extraction experience today! Check out Ollama OCR on GitHub:  https://lnkd.in/grnN_yMk

Recently, I came across Hydra, a tool for elegantly configuring complex applications. Intrigued by what it offers, I decided to incorporate it into a simple classification model. Here's a breakdown of the key technologies I used and their roles: 1. Hydra: It completely changed my configuration management, allowing me to effortlessly handle complex settings and experiment with different hyperparameters. It enhanced the model's flexibility and reproducibility in ways I hadn't imagined possible. 2. PyTorch Lightning: Building on Hydra's configuration prowess, I used PyTorch Lightning to streamline the training process. It made organizing code easier and allowed me to focus on the core machine learning logic. The modular approach significantly improved code readability and maintainability. 3. Pytest & Coverage: To ensure the reliability of my code, I implemented comprehensive unit tests using pytest. The coverage tool was instrumental in identifying areas that needed more test coverage, promoting a test-driven development approach. 4. GitHub Actions: To tie it all together, I set up GitHub Actions to automate the CI/CD pipeline. It ran test cases and created artifacts on each push, streamlining the development process and catching potential issues early. Key Learnings: • The transformative impact of proper configuration management in ML projects • The power of modular design in simplifying complex ML workflows Have you used Hydra or similar tools in your ML projects? I'd love to hear about your experiences and any tips you might have! You can check out the entire project here - https://lnkd.in/dBcDVF6E #MachineLearning #Hydra #PyTorch #DevOps #DataScience

Excited to share my latest project: Coder Assistants! This project is developing AI-powered code assistants to streamline the development workflow. It utilizes large language models (LLMs) like GPT-3.5 and Gemma-2b-it to aid Python and SQL development teams. Key functionalities include: - Code & Query Generation: Assisting developers with writing code and SQL queries. - Technical Management: Assigning tasks to appropriate AI agents based on complexity. - Code & Query Review: Facilitating code and query testing to ensure functionality. - Code Migration: Supporting migration between various programming languages (e.g., SAS to PySpark). This project is still under development, and I'm exploring further advancements in: - Multi-API endpoint support through LiteLLM. - Integration of LangFuse for improved monitoring. - Control Vector implementation for fine-tuning LLM outputs. - Local LLM execution using Ollama or Llama.cpp. - Looking to connect with developers and AI enthusiasts! Feel free to share your thoughts and suggestions in the comments below! #AI #DeveloperTools #CodeAssistants P.S. Check out the project on GitHub for more details: https://lnkd.in/gX-iNx63

🚀 Introducing Source-Code-Analysis-using-GenAI 🤖🔍 What is the power of artificial intelligence in source code analysis? 🌟 My project is called Source-Code-Analysis-using-GenAI. This innovative tool allows you to have interactive conversations with your code, providing invaluable insights and enhancing your coding experience. 💻🤝 GitHub Repository: [Source-Code-Analysis-using-GenAI](https://lnkd.in/g96QAbY2) With this AI-powered tool, you can provide your GitHub URL, and it will analyze your code, enabling you to have a chat-like conversation by mentioning specific files or process names. It's like having a knowledgeable companion who understands your code and offers valuable suggestions and insights. 🗣️💡 Tech Stack Used: - Python - LangChain - Flask - OpenAI - GPT 3 - ChoromaDB By leveraging the power of Python, LangChain, Flask, OpenAI's GPT-3, and ChoromaDB, this project brings cutting-edge AI capabilities to the world of source code analysis. Combining these technologies enables intelligent code understanding, making it easier for developers to identify issues, optimize performance, and improve code quality. 🚀🔬 Imagine having an AI assistant who can comprehend your codebase, help you navigate complex code structures, and provide valuable suggestions. Source-Code-Analysis-using-GenAI is here to revolutionize your coding journey and take it to new heights. 🌟💻 To get started, head over to the GitHub repository [here](https://lnkd.in/g96QAbY2) and explore the possibilities. Be sure to star the repository to show your support and stay updated on the latest developments. ⭐ Let's embrace the future of code analysis with AI and empower ourselves to write better, more efficient code. Together, we can unlock new horizons and push the boundaries of software development. 🌐📈 https://lnkd.in/g96QAbY2 #AI #ArtificialIntelligence #CodeAnalysis #GitHub #OpenSource #Python #Flask #GPT3 #LangChain #ChoromaDB #DeveloperTools #SoftwareDevelopment #TechInnovation

🚀 Exciting News! 🚀 Are you looking to supercharge your source code analysis process? Look no further! Introducing Source-Code-Analysis-using-GenAI, a cutting-edge solution that leverages the power of artificial intelligence to revolutionize how we analyze and understand code. With Source-Code-Analysis-using-GenAI, you can: 🔍 Gain deeper insights into your codebase. 📈 Identify patterns and trends to optimize code quality. 🔄 Automate repetitive tasks and streamline your development workflow. 🔒 Ensure code security and compliance with industry standards. 💡 Unlock new possibilities for innovation and collaboration. Here's how Source-Code-Analysis-using-GenAI can transform your development process: Deeper Insights: GenAI provides unparalleled insights into your codebase, uncovering hidden patterns, dependencies, and performance bottlenecks that may be difficult to detect manually. Code Quality Optimization: By analyzing your code for best practices, design patterns, and potential bugs, GenAI helps you optimize code quality and maintainability, leading to more robust and scalable software. Automated Workflows: Say goodbye to time-consuming manual code reviews. GenAI automates repetitive tasks, such as code formatting, style consistency checks, and even refactoring suggestions, allowing your team to focus on what truly matters—building great software. Enhanced Security and Compliance: With built-in security and compliance checks, GenAI helps ensure that your code meets industry standards and regulatory requirements, reducing the risk of security vulnerabilities and compliance issues. Innovation and Collaboration: By streamlining the code analysis process, GenAI frees up time and resources for innovation and collaboration. Whether you're working on new features, exploring cutting-edge technologies, or collaborating with team members, GenAI empowers you to do more, faster. Github link: https://lnkd.in/gYRn-mQ9 Say goodbye to manual code reviews and hello to intelligent, data-driven analysis with Source-Code-Analysis-using-GenAI. Discover how GenAI can transform your development process today! #SourceCodeAnalysis #GenAI #ArtificialIntelligence #CodeQuality #SoftwareDevelopment #Innovation #Automation sudhanshu kumar Krish Naik Boktiar Ahmed Bappy Sunny Savita iNeuron.ai

I am pleased to announce the completion of my first MLOps project, focusing on Large Language Models (LLMs). This project involved establishing a secure CI/CD pipeline for a Medical Chatbot using GitHub Actions. 𝐒𝐭𝐞𝐩𝐬 𝐔𝐧𝐝𝐞𝐫𝐭𝐚𝐤𝐞𝐧: 1-𝐂𝐨𝐝𝐞𝐛𝐚𝐬𝐞 𝐑𝐞𝐯𝐢𝐞𝐰: Conducted an initial review of the codebase to determine the appropriate pipeline required for deployment and delivery. 2-𝐃𝐞𝐩𝐞𝐧𝐝𝐞𝐧𝐜𝐲 𝐈𝐧𝐬𝐭𝐚𝐥𝐥𝐚𝐭𝐢𝐨𝐧: Installed necessary dependencies, implemented linters, and conducted unit tests. 3-𝐒𝐭𝐚𝐭𝐢𝐜 𝐂𝐨𝐝𝐞 𝐀𝐧𝐚𝐥𝐲𝐬𝐢𝐬: Integrated a Static Code Analyzer into the pipeline. While I utilized SonarQube, other tools are also compatible. 4-𝐂𝐨𝐝𝐞𝐐𝐋 𝐂𝐨𝐧𝐟𝐢𝐠𝐮𝐫𝐚𝐭𝐢𝐨𝐧: Configured CodeQL from the settings to scan for security vulnerabilities. 5-𝐂𝐨𝐧𝐭𝐚𝐢𝐧𝐞𝐫 𝐒𝐜𝐚𝐧𝐧𝐢𝐧𝐠: Performed an initial scan of the container to identify any critical vulnerabilities. 6-𝐁𝐮𝐢𝐥𝐝 𝐚𝐧𝐝 𝐃𝐞𝐩𝐥𝐨𝐲𝐦𝐞𝐧𝐭: Built and pushed an ECR image, followed by deploying the model using Docker and AWS. 𝐊𝐞𝐲 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠𝐬: 1-𝐂𝐈/𝐂𝐃 𝐟𝐨𝐫 𝐌𝐚𝐜𝐡𝐢𝐧𝐞 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠 𝐏𝐫𝐨𝐣𝐞𝐜𝐭𝐬: Developed skills in setting up CI/CD pipelines specifically for machine learning projects involving LLMs and Generative AI. 2-𝐒𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐀𝐰𝐚𝐫𝐞𝐧𝐞𝐬𝐬: Gained an understanding of the top machine learning vulnerabilities and learned methods to secure them through the pipeline. 𝐅𝐮𝐭𝐮𝐫𝐞 𝐒𝐭𝐞𝐩𝐬: 1-𝐌𝐨𝐝𝐞𝐥 𝐕𝐞𝐫𝐬𝐢𝐨𝐧𝐢𝐧𝐠 𝐚𝐧𝐝 𝐦𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭: Implement version control and management for models. 2-𝐋𝐨𝐠𝐠𝐢𝐧𝐠 𝐚𝐧𝐝 𝐌𝐨𝐧𝐢𝐭𝐨𝐫𝐢𝐧𝐠: Introduce comprehensive logging and monitoring mechanisms. 3-𝐄𝐧𝐡𝐚𝐧𝐜𝐞𝐝 𝐒𝐞𝐜𝐮𝐫𝐢𝐭𝐲: Design the pipeline with an increased focus on application security. 4-𝐊𝐮𝐛𝐞𝐫𝐧𝐞𝐭𝐞𝐬 𝐈𝐧𝐭𝐞𝐠𝐫𝐚𝐭𝐢𝐨𝐧: Plan to integrate Kubernetes for deployment. 5-𝐂𝐨𝐧𝐭𝐢𝐧𝐮𝐨𝐮𝐬 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠: Establish pipelines for continuous training and updating of ML models. This project has been a valuable learning experience, and I look forward to further refining and expanding these capabilities. Your feedback and suggestions would be greatly appreciated. Thank you Boktiar Ahmed Bappy for the repo which helped me implement the project using DevSecOps principles #MLOps #DevOps #MachineLearning #LLMs #GitHubActions #CI #CD #AWS #Docker #Kubernetes #ContinuousLearning 𝐏𝐫𝐨𝐣𝐞𝐜𝐭 𝐑𝐞𝐩𝐨: https://lnkd.in/dZbuzwzf