Github Top Repositories

🚀 Meet rasbt/LLMs-from-scratch: a gem from today's GitHub trending list. 🔗 https://github.com/rasbt/LLMs-from-scratch 📝 Implement a ChatGPT-like LLM in PyTorch from scratch, step by step ────────────────────────────── The rasbt/LLMs-from-scratch GitHub repository provides a comprehensive guide to building a large language model (LLM) from scratch. The repository is the official code companion to the book Build a Large Language Model (From Scratch). It covers the development, pretraining, and finetuning of a GPT-like LLM, with a focus on hands-on implementation using PyTorch. The repository includes code notebooks for each chapter, covering topics such as text data processing, attention mechanisms, and model finetuning. It also provides a troubleshooting guide and a setup guide for installing Python and required packages. The target audience is developers and researchers interested in LLMs, with a strong foundation in Python programming. The repository is designed to be accessible, with code that can run on conventional laptops. In summary, the rasbt/LLMs-from-scratch repository offers a unique opportunity to learn about LLMs by building one from scratch - and that's a powerful way to learn. ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

💡 millionco/react-doctor just hit the trending charts — here's why it matters. 🔗 https://github.com/millionco/react-doctor 📝 Your agent writes bad React. This catches it ────────────────────────────── The react-doctor GitHub repository is a diagnostic tool designed to help developers identify and fix issues in their React applications. Its primary purpose is to simplify the debugging process by providing a simple and straightforward way to diagnose common problems. The key features of react-doctor include its ability to detect a wide range of issues, from incorrect prop types to potential memory leaks. Usage is relatively straightforward, with users able to integrate the tool into their existing workflows with minimal setup. From a technical standpoint, react-doctor is built using a combination of React and other popular JavaScript libraries. The codebase is well-organized and easy to navigate, making it a great resource for developers looking to learn more about debugging and diagnostics. react-doctor is ideal for anyone working with React, from seasoned veterans to newcomers to the framework. Whether you're trying to track down a tricky bug or simply looking to improve your overall development workflow, react-doctor is definitely worth checking out. You can use react-doctor to diagnose your React application by following the setup instructions in the repository. The takeaway: Debug your React apps like a pro with react-doctor! ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

🔍 Deep-diving into anonfaded/FadCam — fresh off the trending list. 🔗 https://github.com/anonfaded/FadCam 📝 Open-source, ad-free Android multimedia recorder with background video recording, screen recording, live streaming, and remote camera control ────────────────────────────── FadCam is a privacy-focused Android multimedia suite that offers various features like background video recording, dashcam, screen recording with annotations, live streaming over local networks, and remote camera control. The app is designed for ethical use only, with examples including personal security, content creation, remote monitoring, and research and documentation. The app has gained popularity and has been featured on several platforms, including YouTube videos and articles. You can download FadCam from various sources, including GitHub, F-Droid, and the Amazon appstore. From a technical perspective, FadCam is an open-source application, and its development is part of the FadSec Lab suite. The app is designed to be ad-free, ensuring a seamless user experience. The target audience for FadCam includes individuals looking for a privacy-focused multimedia suite for their Android devices. Whether you're a content creator, a researcher, or someone who wants to use the app for personal security, FadCam has something to offer. Checkout the FadCam GitHub repository for more information on how to use and contribute to the project. In short, FadCam is the ultimate privacy-focused Android app for all your multimedia needs - and it's completely free! ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

🔥 mattpocock/skills is trending — and it deserves your attention. 🔗 https://github.com/mattpocock/skills 📝 Skills for Real Engineers. Straight from my .claude directory. ────────────────────────────── The mattpocock/skills GitHub repository is a collection of skills designed to help engineers work more efficiently with coding agents. The skills are small, easy to adapt, and composable, making them suitable for use with any model. They are based on decades of engineering experience and are designed to be used daily. The skills can be quickly set up using the npx skills@latest add mattpocock/skills command, followed by selecting the desired skills and running the /setup-matt-pocock-skills command in your agent. This will guide you through setting up the necessary configuration, such as choosing an issue tracker and defining labels for triage. The skills are categorized into engineering, productivity, and misc tools. Engineering skills include diagnose, grill-with-docs, and tdd, which help with debugging, grilling sessions, and test-driven development. Productivity skills like grill-me and handoff help with communication and workflow. The skills are designed to address common failure modes in software development, such as misalignment, verbosity, and code that doesn't work. They promote a shared language, feedback loops, and disciplined diagnosis. With these skills, engineers can ship better apps and enjoy a more efficient development process. In short, mattpocock/skills is a game-changer for engineers working with coding agents - it's like having a superpower in your coding toolkit! ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

📌 Spotted on GitHub Trending: apernet/hysteria — let's break it down. 🔗 https://github.com/apernet/hysteria 📝 Hysteria is a powerful, lightning fast and censorship resistant proxy. ────────────────────────────── Meet Hysteria, a powerful and lightning-fast proxy that's resistant to censorship. Its key features include a wide range of modes like SOCKS5, HTTP Proxy, and TCP/UDP Forwarding, making it a jack of all trades. It's also incredibly fast, thanks to a customized QUIC protocol, and can masquerade as standard HTTP/3 traffic to evade detection. Hysteria is cross-platform, with builds for every major platform and architecture, and is easy to integrate into your infrastructure. The project has a chill and supportive community, with well-documented specifications and code for developers to contribute and build their own apps. Whether you're a developer looking for a reliable proxy solution or just someone who wants to bypass censorship, Hysteria is definitely worth checking out. So why not give it a try and see how it can help you - Hysteria: proxying just got a whole lot faster and freer. ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

📌 Spotted on GitHub Trending: CloakHQ/CloakBrowser — let's break it down. 🔗 https://github.com/CloakHQ/CloakBrowser 📝 Stealth Chromium that passes every bot detection test. Drop-in Playwright replacement with source-level fingerprint patches. 30/30 tests passed. ────────────────────────────── Meet CloakBrowser, a stealthy Chromium browser that passes every bot detection test. It's not a patched config or JS injection, but a real Chromium binary with fingerprints modified at the C++ source level. Antibot systems score it as a normal browser — because it is a normal browser. Key features include: - 49 source-level C++ patches for canvas, WebGL, audio, and more - humanize=True for human-like mouse curves, keyboard timing, and scroll patterns - 0.9 reCAPTCHA v3 score and passing Cloudflare Turnstile, FingerprintJS, and BrowserScan - Auto-updating binary with background update checks - Free and open source with no subscriptions or usage limits You can use CloakBrowser as a drop-in replacement for Playwright or Puppeteer in Python or JavaScript. It's easy to install with pip install cloakbrowser or npm install cloakbrowser, and the binary auto-downloads on first run. The target audience is developers who need to automate browser interactions without being detected as bots. CloakBrowser is perfect for web scraping, automation, and AI agents. In a nutshell, CloakBrowser is the ultimate solution for stealthy browser automation — it's a game-changer for anyone tired of bot detection roadblocks. ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

🎯 rohitg00/agentmemory landed on trending. Worth a proper look. 🔗 https://github.com/rohitg00/agentmemory 📝 #1 Persistent memory for AI coding agents based on real-world benchmarks ────────────────────────────── Agentmemory is a game-changer for AI coding agents, providing persistent memory that captures what your agent does, compresses it into searchable memory, and injects the right context when the next session starts. With npx @agentmemory/agentmemory, you can unlock the full potential of your agent, eliminating the need for re-explaining and copy-pasting. Key features include auto-capture with 12 hooks, search using BM25, vector, and graph fusion, and multi-agent support via MCP, REST, leases, and signals. The project boasts impressive benchmarks, with 95.2% retrieval accuracy and significant token savings. Technical highlights include a memory engine, MCP server, and zero external database dependencies. Usage is straightforward, with a simple command to get started. The target audience includes developers working with various AI coding agents, such as Claude Code, Cursor, Gemini CLI, and more. In short, agentmemory revolutionizes the way you work with AI coding agents, making it an indispensable tool for any developer looking to boost productivity and efficiency. With its robust features and impressive performance, agentmemory is a must-have for anyone working with AI coding agents - say goodbye to re-explaining, and hello to seamless coding sessions. ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

🔍 Deep-diving into tinyhumansai/openhuman — fresh off the trending list. 🔗 https://github.com/tinyhumansai/openhuman 📝 Your Personal AI super intelligence. Private, Simple and extremely powerful. ────────────────────────────── OpenHuman is your personal AI superintelligence, designed to integrate with your daily life in a private, simple, and powerful way. With a clean desktop experience and short onboarding paths, you can get started in just a few clicks - no config-first setup or terminal required. Key features include 118+ third-party integrations with one-click OAuth, a local-first knowledge base built from your data and activity, and a memory tree that compresses everything into Markdown files stored intelligently in an Obsidian-compatible vault. To get started, you can either download from the website or run the installation script using curl -fsSL https://raw.githubusercontent.com/tinyhumansai/openhuman/main/scripts/install.sh | bash for MacOS/Linux or irm https://raw.githubusercontent.com/tinyhumansai/openhuman/main/scripts/install.ps1 | iex for Windows. Technical highlights include smart token compression, model routing, and native voice capabilities. This project is ideal for individuals looking for a powerful and private AI assistant. In short, OpenHuman is a game-changer - it's like having your own personal AI assistant that learns and adapts to your habits in minutes, not weeks. ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

🚀 Meet rohitg00/agentmemory: a gem from today's GitHub trending list. 🔗 https://github.com/rohitg00/agentmemory 📝 #1 Persistent memory for AI coding agents based on real-world benchmarks ────────────────────────────── The agentmemory GitHub repository provides a persistent memory solution for AI coding agents, allowing them to recall previous conversations and maintain context. This means you won't have to re-explain things to your agent every time you interact with it. Key features include automatic capture of agent interactions, searchable memory, and integration with multiple agents via MCP or REST API. To get started, simply run npx @agentmemory/agentmemory and follow the prompts. The repository includes benchmarks demonstrating the effectiveness of agentmemory, including a 95.2% retrieval accuracy and 92% fewer tokens required for context. Technical highlights of the project include its use of iii-engine and SQLite for data storage, as well as its zero external dependencies. The project is suitable for developers and users of AI coding agents, particularly those who work with multiple agents or need to maintain complex conversations. One command, npx @agentmemory/agentmemory, and your agent will never forget again. ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

🔥 What if you could turn $1,000 into $10,100 in just 2-3 days? Analysis of 5000+ investors shows a staggering 85-95% accuracy in our signals! Imagine following daily signals that not only guide your trades but transform your financial future. 👉 Here’s what you get: - 2-5 high-accuracy trading signals daily - Expert mentorship to elevate your skills - Investment options that can multiply your earnings Don’t let fear of missing out hold you back. DM us to unlock the potential of your investments! 💰✨ @MambaFxprofit #ad 📢 InsideAd

🚀 Meet NousResearch/hermes-agent: a gem from today's GitHub trending list. 🔗 https://github.com/NousResearch/hermes-agent 📝 The agent that grows with you ────────────────────────────── Hermes Agent is a self-improving AI agent that creates skills from experience, improves them during use, and builds a deepening model of the user across sessions. It can run on various platforms, including a $5 VPS, GPU cluster, or serverless infrastructure. Key features include a real terminal interface, cross-platform conversation continuity, a closed learning loop, scheduled automations, and delegates and parallelizes tasks. hermes model allows switching between different models and providers, such as Nous Portal, OpenRouter, and Hugging Face, without code changes. The agent can be used via Telegram, Discord, Slack, WhatsApp, Signal, or the command-line interface. To get started, users can run the curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash command for Linux, macOS, WSL2, or Termux, or use the PowerShell script for Windows. After installation, users can start chatting with the agent using hermes, configure tools with hermes tools, or set individual config values with hermes config set. The target audience includes researchers, developers, and individuals interested in AI and self-improving systems. With its extensive documentation and contributing guide, Hermes Agent is an open and collaborative project. In a nutshell, Hermes Agent is a cutting-edge AI tool that learns and adapts to its user, making it an exciting project to explore and contribute to - the future of AI is here, and it's conversational. ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

📌 Spotted on GitHub Trending: rasbt/LLMs-from-scratch — let's break it down. 🔗 https://github.com/rasbt/LLMs-from-scratch 📝 Implement a ChatGPT-like LLM in PyTorch from scratch, step by step ────────────────────────────── Check out the rasbt/LLMs-from-scratch GitHub repo, the official code repository for the book "Build a Large Language Model (From Scratch)". This repo provides a step-by-step guide to developing, pretraining, and finetuning a GPT-like Large Language Model (LLM) from scratch. The repo includes code examples and exercises for each chapter, allowing you to practice and learn by coding. You can download the repository using git clone --depth 1 https://github.com/rasbt/LLMs-from-scratch.git or click the "Download ZIP" button. Key features include: - Implementing a GPT model from scratch - Pretraining on unlabeled data - Finetuning for text classification and instruction following - Using PyTorch to implement the code This repo is perfect for developers and researchers looking to build and understand LLMs. The prerequisites include a strong foundation in Python programming and some experience with deep neural networks. You can run the code on your laptop, and it will automatically utilize GPUs if available. Build a Large Language Model from scratch and become a master of LLMs - the possibilities are endless! ────────────────────────────── 🧠 Channel: https://t.me/GithubRe

💡 AUTOMATIC1111/stable-diffusion-webui just hit the trending charts — here's why it matters. 🔗 https://github.com/AUTOMATIC1111/stable-diffusion-webui 📝 Stable Diffusion web UI ────────────────────────────── The Stable Diffusion web UI is a powerful tool that allows users to interact with the Stable Diffusion model through a user-friendly web interface. This interface offers a wide range of features, including txt2img and img2img modes, outpainting, inpainting, and more. Users can also adjust parameters such as attention, sampling methods, and negative prompts to refine their results. The UI is built using the Gradio library and supports various models, including Stable Diffusion 2.0 and Alt-Diffusion. It also includes features like batch processing, prompt editing, and a CLIP interrogator. The project is highly customizable, with options to change defaults, mix, and max values for UI elements via a text config. To get started, users can follow the installation instructions for their specific operating system, including Windows, Linux, and Apple Silicon. The project also has a wiki with detailed documentation and a contributing guide for those who want to add code to the repository. The Stable Diffusion web UI is perfect for artists, researchers, and anyone interested in exploring the capabilities of text-to-image models. With its intuitive interface and extensive features, it's an essential tool for anyone looking to push the boundaries of AI-generated art. One-liner takeaway: The Stable Diffusion web UI is a game-changer for text-to-image generation, offering a user-friendly interface and a vast array of features to unlock your creativity. ────────────────────────────── 🧠 Channel: https://t.me/GithubRe