All Posts
Filter by keyword, tag, and category.
InsForge: The Backend Development Platform Built for AI Agents
InsForge is a backend development platform specifically built for AI coding agents and AI code editors. Acting as a semantic layer between agents and backend infrastructure like databases and authentication, it enables AI to autonomously understand, configure, and inspect backend states through backend context engineering. This facilitates the automated delivery of full-stack applications.
NousResearch/hermes-agent: An Open-Source AI Agent Framework with Self-Evolving Capabilities
The open-source hermes-agent by NousResearch is an AI agent framework with self-evolving capabilities. It features a built-in learning loop to extract skills from experience and supports various mainstream large models. With its minimalist installation experience and innovative self-improvement mechanism, this project is becoming a popular choice for developers to automate workflows.
Official Google Cloud Generative AI Guide: Analysis of the Practical Codebase Based on Vertex AI and Gemini
This project is an official open-source collection of generative AI sample code and Jupyter Notebooks by Google Cloud, focusing on invoking models like Gemini, Imagen, and Chirp via Vertex AI. Providing comprehensive use cases from basic introductions to advanced function calling, this codebase serves as a standard reference implementation for developers to build, test, and deploy large language models and multimodal AI applications within the Google Cloud ecosystem.
Claude Skills: An Open-Source Plugin Library Injecting 170+ Production-Grade Domain Knowledge into AI Coding Assistants
alirezarezvani/claude-skills is an open-source project providing over 170 production-grade skills and plugins for AI coding agents like Claude Code and OpenAI Codex. Through modular instruction packages, it injects expert knowledge from engineering, product, and compliance domains into AI assistants, supporting native cross-platform execution. This project is ideal for developers and teams needing to expand the professional capabilities of their AI assistants.
Microsoft Open-Sources HVE Core: Hypervelocity Engineering Prompts and Component Library for GitHub Copilot
Microsoft's open-source hve-core (Hypervelocity Engineering Core) is a prompt and component library specifically designed for GitHub Copilot. By providing validated instructions, agents, and skills, it helps developers build constraint-based AI workflows. This maximizes the effectiveness of AI programming assistants across various projects, ultimately achieving standardization and increased efficiency in the research and development process.
POET-X: Memory-efficient Large Language Model Training by Scaling Orthogonal Transformation
Large language model training often faces memory bottlenecks. This paper proposes the POET-X framework, which significantly reduces computational overhead and memory footprint by scaling orthogonal equivalence transformations while maintaining training stability and generalization. Experiments show that POET-X can pre-train a billion-parameter LLM on a single Nvidia H100 GPU, whereas AdamW runs out of memory under the same conditions. This provides a highly valuable training solution for resource-constrained teams.
Exploring the Application of AI Multi-Agents in Quantitative Trading: An Analysis of the ai-hedge-fund Project
virattt/ai-hedge-fund is a Python-based proof-of-concept AI hedge fund project. Through multi-agent collaboration, the system simulates the trading strategies of renowned investment masters, including Charlie Munger and Cathie Wood. Primarily intended for educational purposes, it aims to explore the potential of large language models in financial trading decisions. The project has currently garnered over 46,000 stars on GitHub.
RoboPocket: Instant Robot-Free Policy Iteration with Smartphones
Data collection efficiency in imitation learning has always been a bottleneck in robotics. This paper proposes the RoboPocket system, which utilizes ordinary smartphones and AR visual foresight technology to achieve instant robot-free policy iteration. By visualizing predicted trajectories through remote inference and combining asynchronous online fine-tuning, the system doubles data efficiency. It provides a novel, low-cost, and high-efficiency paradigm for large-scale robot data collection.