Projects

Overview: Cognio is a Model Context Protocol (MCP) server designed for AI assistants, providing persistent semantic memory that survives across conversations. It supports multi-project context management and seamless multilingual support, making it ideal for personal knowledge bases, research notes, and conversation history. Key features: - **Semantic Search**: Finds memories by meaning. - **LEANN Vector Search (Optional)**: Lazy-built index for efficient on-demand recomputation. - **Multilingual Support**: Searches in 100+ languages. - **Persistent Storage**: SQLite-based storage that persists across sessions. - **Project Organization**: Organizes memories by project and tags. - **Auto-Tagging**: Automatically generates tags via LLMs. - **Text Summarization**: Extractive and abstractive summarization for long texts. - **MCP Integration**: One-click setup for VS Code, Claude, Cursor, and more. - **RESTful API**: Standard HTTP API with OpenAPI documentation. - **Export Capabilities**: Export memories to JSON or Markdown format. - **Docker Support**: Simple deployment with docker-compose. Tech stack: Python. Why it matters: Cognio offers a reliable, persistent memory solution for AI assistants, ensuring that their knowledge and context are preserved and easily accessible across different sessions and projects. Repository: 0xReLogic/Cognio GitHub stars: 12 GitHub description: Persistent semantic memory server for AI assistants via Model Context Protocol (MCP).

Overview: This repository contains a collection of alternative ChatGPT models and fine-tuning methods for language modeling. It includes models such as THUDM/ChatGLM-6B, Stanford Alpaca, PaLM-rlhf-pytorch, ChatLLaMA, and OpenChatKit. Key features: - THUDM/ChatGLM-6B: A 62-GB General Language Model with support for Chinese and English. - Stanford Alpaca: A 52K Alpaca model finetuned from OpenAI's text-davinci-003, closely resembling the Alpaca 7B model. - PaLM-rlhf-pytorch: An RLHF implementation on the PaLM architecture. - ChatLLaMA: A smaller LLaMA model with RLHF training, similar to ChatGPT but with lower training costs. - OpenChatKit: A 200-GB model created by former Hugging Face employees, trained on OIG-43M data. Tech stack: Python, PyTorch, Hugging Face, and other relevant libraries. Why it matters: This repository provides a comprehensive collection of alternative ChatGPT models and fine-tuning methods, offering developers a variety of options to experiment with and potentially improve their language modeling capabilities.

### Overview: The leCore is a self-contained engine for the leOS operating system. It represents everything—memory, meaning, 3-D geometry, physics, images—as points in a single, very large space. It operates using a handful of simple, reversible operations and is entirely based on NumPy, without the use of neural networks or pretrained models. This project is designed to be a key component of the leOS ecosystem, providing a foundational engine for various applications. ### Key features: - **Memory Representation**: All data is stored and manipulated as points in a single, large space. - **Pure NumPy**: The engine operates entirely on NumPy, ensuring deterministic behavior. - **Long-Horizon Memory (LongMemEval)**: Achieves 1.000 across all six categories and six seeds, demonstrating robust memory handling. - **Grounded Answering**: Achieves a hit rate of 0.75 with 1.00 answer quality, showcasing effective grounded answering capabilities. - **Memory-First Speed**: Performs significantly faster than generating, with a speed improvement of 97×. - **In-Model Fact Install and Preservation**: Preserves key perturbations with a preservation rate of 0.1014 → 2.69e-09. - **Exact Uninstall**: Ensures that the weight delta after reverting is 3.7e-09, maintaining the integrity of the system. - **Wiring Health**: The system has 12 organ groups, no import failures, and a

### Overview: FloatChat AI is a conversational AI platform designed for seamless ARGO Ocean Data Discovery and Visualization. Built with Python, LangChain, Pandas, and machine learning models, it makes complex ocean data accessible and actionable for researchers and enthusiasts. ### Key Features: - **Advanced Conversational AI**: Empowers users to interact with ARGO data effortlessly and efficiently. - **Data Analysis and Visualization**: Utilizes machine learning models to generate intelligent insights from oceanographic datasets. - **User-Friendly Interface**: Ensures that both researchers and enthusiasts can explore, analyze, and visualize data seamlessly. - **Data Handling**: Supports efficient data manipulation, analysis, and transformation using Pandas and CSV/JSON data handling. - **Real-Time Data Discovery**: Enables real-time data discovery and updates through APIs and integration tools. - **Modular Architecture**: Ensures adaptability and ease of integration with diverse data sources. ### Tech Stack: Python, LangChain, Pandas, Machine Learning Models, Data Visualization Libraries, APIs, Integration Tools, User Interface Technologies ### Why it Matters: FloatChat AI provides a powerful and user-friendly solution for researchers and enthusiasts to explore, analyze, and visualize complex oceanographic datasets, making data-driven insights more accessible and actionable.

Overview: Kling AI Master is an AI environment offering model chaining, workflow automation, and output optimization. Key features: - Model Comparison - Privacy Controls - Plugin System Tech stack: Unknown Why it matters: This AI environment provides a comprehensive solution for model comparison, privacy controls, and plugin integration, making it useful for various AI applications and research.

Overview: This project aims to develop an innovative anomaly detection system using advanced data mining and deep learning techniques to accurately identify and localize defects in manufacturing components, thereby enhancing quality control processes and reducing production losses. Key features: - Advanced data mining techniques - Deep learning algorithms - Explainable AI for interpretable results Tech stack: Jupyter Notebook Why it matters: The project addresses the critical challenges faced by manufacturing industries, including rare defects, large scale, sensitivity, and diverse domains, by leveraging advanced machine learning and artificial intelligence techniques. This system can help ensure high-quality production, reduce losses, and improve overall efficiency, making it essential for industries like semiconductor manufacturing in the Philippines.

Overview: Polymer-Agent is a Python-based LLM-aided toolkit designed for initial polymer discovery, offering single-property and multi-property constrained optimization, latent-space decoding, and chemistry-aware constraints. Key features: - Single-property optimization - Multi-property constrained optimization - Latent-space decoding - Chemistry-aware constraints Tech stack: Python Why it matters: Polymer-Agent simplifies the process of polymer discovery by leveraging advanced machine learning techniques, making it an invaluable tool for researchers and developers in the field of polymer science and engineering. Repository: BaratiLab/Polymer-Agent Primary language: Python GitHub stars: 11

Overview: This project is a MCP server designed for AI image generation, transformation, and style management using the Recraft API. It includes 16 tools for various image processing tasks such as text-to-image, inpainting, background replacement, and more. Key features: - Generate: Text-to-image, image-to-image, inpainting, background replacement & generation - Process: Background removal, region erasing, vectorization (SVG), crisp & creative upscaling - Styles: Create custom styles from reference images, list/get/delete styles - Account: Check remaining credits and user info (included in the styles toolset) - Configurable: Filter toolsets, adjust cache TTL, control retries Tech stack: TypeScript Why it matters: This project provides a comprehensive solution for AI image processing, making it a valuable tool for developers and enthusiasts alike. GitHub stars: 12 GitHub description: MCP server for AI image generation, transformation, and style management via the Recraft API. 16 tools for text-to-image, inpainting, vectorization, upscaling, and more.

### Overview: This repository contains the GLM-5.3-Flash NVFP4 checkpoint for distributed execution on DGX Spark-class systems. It serves two versions: 1.x with TensorFold (TP=2 or TP=3, FP8 KV, text, tool calls, and images) and 2.x with TensorFold (TP=2 and TP=3, FP8 KV, text, tool calls, and images) on two or three nodes. The repository includes changelogs and directories for each serving line, ensuring compatibility with DGX Spark-compatible systems. ### Key features: - **GLM-5.3-Flash NVFP4 checkpoint** for distributed execution on DGX Spark-class systems. - **TensorFold (TP=2 or TP=3, FP8 KV, text, tool calls, and images)** for 2.x version. - **TensorFold (TP=2 and TP=3, FP8 KV, text, tool calls, and structured output)** for 3.x version. - **Pinned assets** for 1.x version, including a vLLM checkpoint. ### Tech stack: Python-based repository with Apache-2.0 code and MIT weights. ### Why it matters: This repository provides a business-use setup with commercially usable licensing, ensuring the intended use is within the scope of the repository's documentation. It serves GLM-5.3-Flash on DGX Spark-compatible systems, offering flexibility and compatibility for various deployment scenarios.

Overview: This project introduces a method called **Vulnerability-Aware Alignment (VAA)**, which addresses the issue of uneven forgetting during harmful fine-tuning in machine learning models. VAA upweights and reinforces vulnerable data to enhance safety retention during customized fine-tuning. Key features: - Upweights and reinforces vulnerable data - Mitigates uneven forgetting in harmful fine-tuning - Supports both ERM and vaccine models Tech stack: Python Why it matters: VAA provides a safety alignment method that effectively mitigates the issue of uneven forgetting in harmful fine-tuning, ensuring more stable and safer fine-tuned models.

Overview: SynthForge is a toolkit designed for generating high-quality synthetic datasets specifically tailored for large language models (LLMs) for fine-tuning. It supports multiple backends and allows for filtering, making it efficient and user-friendly. Key features: - GPU-accelerated processing - Multi-backend support - Filterable datasets Tech stack: Python Why it matters: SynthForge offers a valuable solution for developers looking to optimize the training of LLMs by providing high-quality synthetic datasets, which can significantly improve model performance and reduce the need for real-world data. Repository: Dosi-f/synthforge Primary language: Python GitHub stars: 12 GitHub description: 🧪 SynthForge — toolkit for generating high-quality synthetic datasets for LLM fine-tuning. Multi-backend, filterable, GPU-accelerated. Early-stage experimental project.

Overview: This project focuses on developing a smoke and fire detection system using custom CNN and fine-tuned Google Vision Transformer for wild environment monitoring. Key features: - Custom CNN for smoke and fire detection - Fine-tuned Google Vision Transformer for smoke and fire detection - Self-built dataset for training and evaluation Tech stack: PyTorch with PyTorch Lightning Module Why it matters: The project showcases the effectiveness of combining custom CNNs with pre-trained Vision Transformers for smoke and fire detection in forest environments, providing a robust solution for early warning systems.

### Overview: This ComfyUI-stylebook repository contains 700+ visual styles, each with a rendered preview you can browse before committing. It includes 1000+ artists with written descriptors and modifier previews for lighting, color, era, period dress, finish, and mood. The styles ship with a written description, keyword list, negative prompt, and preview image. The package is offline, zero dependencies, and no install needed. ### Key features: - **700+ Visual Styles**: Each style has a rendered preview, keyword list, negative prompt, and preview image. - **1000+ Artists**: Descriptors and written descriptions for 1000+ artists. - **Modifier Gallery**: Tilted lighting, color, era, period dress, finish, and mood modifiers with rendered tiles and verbatim text. - **No Dependencies**: No install, no downloads, no network calls, no API keys required. ### Tech stack: Python ### Why it matters: Stop guessing whether "wet plate collodion" or "risograph" will do what you want. Open the gallery, look at it, click it. Every style ships four things: a written description, a keyword list, a matching negative prompt, and a preview image. No dependencies, no downloads, no network calls, no API keys. Every style ships four things: a written description, a keyword list, a matching negative prompt, and a preview image. 1050+ artists ship a written descriptor each, so the look still

Overview: This project is an application that uses Weaviate, a vector-based image similarity search engine, to find similar products across the market. The app leverages 62k images of products combined with their text descriptions to build vectors for search purposes. Key features: - Vector-based image similarity search for product recommendations - Integration with Weaviate for efficient image and text data storage and retrieval - Use of 62k product images and descriptions for accurate search results Tech stack: Python Why it matters: This project is useful for stores to identify if their products were also in their competitor's offerings, providing valuable insights into market trends and competition.

Overview: TechRAG is a hybrid ontology-driven retrieval engine designed to enhance the precision and traceability of AI-generated technical documentation. Key features: - Vector embeddings via LanceDB (OpenAI-compatible endpoint) - Domain ontology + knowledge graph following the SQLGraph approach - Atomic facts with source back-reference - Response audit metadata (evidence coverage, reasoning score, recovery cycles) Tech stack: Python Why it matters: TechRAG leverages a structured knowledge representation and adaptive vector search to provide precise and traceable context to LLMs, ensuring better quality and reliability of technical documentation. Repository: FBR65/TechRAG GitHub stars: 12 GitHub description: TechRAG is a hybrid ontology-driven retrieval engine that combines a SQLAlchemy-based knowledge graph and LanceDB vector store with OpenAI-compatible chat and embedding clients for retrieval-augmented generation with audit metadata and recovery cycles.

Overview: COVID-VTS is a fact-checking dataset designed for short video platforms, focusing on automatically generating verifiable, trustworthy, and misleading claims on the platform. Key features: - Automatic generation of large-scale verifiable, trustworthy, and misleading claims - Proposed TwtrDetective framework for fact-checking on short video platforms Tech stack: Python, Golang, TensorFlow Why it matters: The dataset and framework provide a practical tool for researchers and practitioners to evaluate and improve fact-checking algorithms on short video platforms, contributing to the field of digital misinformation detection.

Overview: This repository contains open-source skills designed to maintain brand consistency in AI-generated visuals, specifically for Claude. It addresses the 30% problem of style drift by providing a disciplined, training-free workflow that ensures visual identity remains consistent across generations. Key features: - Style-consistent image & video generation - Disciplined, training-free workflow for brand consistency - Accurate, corrective guidelines for specific failures - Representative atlas for reference-based editing Tech stack: Python, PyTorch, and other dependencies. Why it matters: This approach offers a practical solution for maintaining brand consistency in AI-generated visuals, avoiding the downsides of LoRA training and ensuring high-quality, brand-consistent outputs without the need for expensive retraining.

Overview: This project aims to provide a demo of QDrant, a vector database, for NLP tasks, focusing on interactive query creation and display of results. Key features: - Supports creating and storing queries for QDrant Vector Database. - Displays positives, negatives, and results for each query. - Enables downloading datasets directly from Hugging Face. - Provides a loading sign while generating embeddings. - SQLiteDB for storing query results and names. - SQLiteDB for storing query results and names. Tech stack: Python Why it matters: This project serves as an introduction to vector databases, offering a simple way to interact with QDrant for NLP tasks. It demonstrates the capabilities of QDrant and provides a foundation for more advanced use cases. Repository: GeorgePearse/QDrant-NLP Primary language: Python GitHub stars: 12 GitHub description: QDrant-NLP

Overview: This repository showcases the implementation of a basic RAG (Reinforcement Augmented Generation) pipeline using LangChain in Supabase Edge Functions, providing a practical example for developers. Key features: - Basic RAG pipeline using LangChain - Integration of LangChain packages in Supabase Edge Functions - Installation of npm packages via `import_map.json` - Deployment of Supabase Edge Functions to production Tech stack: TypeScript Why it matters: This project offers a clear demonstration of how to integrate LangChain with Supabase Edge Functions, making it a valuable resource for developers looking to enhance their applications with advanced natural language processing capabilities.

### Overview: VelociRAG is a lightning-fast RAG solution designed for AI agents, leveraging ONNX Runtime for efficient vector search and four-layer retrieval fusion. It provides sub-200ms warm search times and integrates seamlessly with MCP servers, making it ideal for real-time applications without the need for GPUs or PyTorch. ### Key Features: - **4-layer Retrieval Fusion:** Combines vector similarity, BM25 keyword matching, knowledge graph traversal, and metadata filtering. - **ONNX Runtime:** Runs all components on local models, ensuring no GPU or PyTorch dependencies. - **Incremental Indexing:** Adds documents without rebuilding the entire index, making it easier to manage large datasets. - **MCP Server Integration:** Plugs into Claude, Cursor, and Windsurf for agent integration, providing a Unix socket daemon for warm queries. - **CLI and Python API:** Available for both command-line and Python integration, allowing seamless integration with various AI agents. ### Tech Stack: VelociRAG is built using Python, utilizing ONNX Runtime for inference and FAISS for vector similarity search. The solution is designed to be lightweight and efficient, making it suitable for real-time applications. ### Why It Matters: VelociRAG offers a powerful and efficient way to integrate AI agents into existing systems, providing fast, accurate search capabilities without the need for complex frameworks like PyTorch. Its lightweight nature and easy integration with MCP servers make it a valuable tool for developers looking to enhance their AI applications.