3 repositorios
Configures label sets and LLM settings for assigning categories to document pages.
Distinct from Classification Labelers: Distinct from Classification Labelers: uses LLMs for dynamic label assignment rather than fixed ML training labels.
Explore 3 awesome GitHub repositories matching data & databases · LLM-Based. Refine with filters or upvote what's useful.
Crucix is an open-source intelligence system comprising an OSINT aggregator, a geospatial intelligence dashboard, and an LLM intelligence agent. It functions as a real-time signal monitor and automated alerting system designed to collect, analyze, and visualize geopolitical, economic, and satellite data from diverse open-source intelligence sources. The system utilizes large language models to synthesize intelligence feeds, generate actionable trade ideas, and classify signal priority with confidence scores. It features a geospatial visualization interface that plots intelligence events, such
Employs large language models to assign semantic labels and priority scores to raw intelligence data.
Kreuzberg is a document extraction engine that converts PDFs, Office files, images, and over 90 other formats into clean, structured text and metadata. It is built around a compiled Rust core that can be used as a native library, a command-line tool, a REST API server, or a WebAssembly module for browser-based processing. The system is designed to run entirely on self-hosted infrastructure, with no data leaving the user's environment. What distinguishes Kreuzberg is its breadth of integration surfaces and its pipeline architecture. It exposes extraction capabilities through native bindings fo
Ships a configurable LLM-based page classification enrichment stage for document processing.
llama-fs es un gestor de directorios y organizador de sistemas de archivos automatizado que utiliza modelos de lenguaje grandes (LLM) para categorizar, renombrar y ordenar archivos en directorios estructurados. Funciona como un pipeline de datos local que analiza el contenido de los archivos para determinar jerarquías de carpetas y patrones de nombres apropiados. El proyecto enfatiza la privacidad al enrutar el procesamiento de datos a instancias de modelos locales, asegurando que la información sensible de los archivos permanezca en el dispositivo en lugar de enviarse a proveedores en la nube. Utiliza análisis multimodal para procesar imágenes, audio y texto, permitiendo al sistema identificar y clasificar diversos tipos de medios para una organización automática. El sistema incluye monitoreo del sistema de archivos en tiempo real para detectar cambios en directorios y activar tareas de limpieza automatizadas. Admite procesamiento por lotes para reestructurar árboles de directorios completos en una sola operación y síntesis recursiva de directorios para crear automáticamente jerarquías de carpetas anidadas.
Uses large language models to analyze file text and determine appropriate folder structures and filenames.