Repository History
239 repositories tagged with python
Axolotl: Streamlining LLM Fine-tuning with a Powerful Open-Source Framework
Axolotl is a comprehensive, free, and open-source framework designed to simplify the post-training and fine-tuning processes for large language models (LLMs). It offers extensive model support, diverse training methods, and robust performance optimizations, making it an invaluable tool for researchers and developers. With easy configuration and cloud-ready deployment, Axolotl empowers users to efficiently customize and enhance LLMs.
Griptape: Modular Python Framework for AI Agents and Workflows
Griptape is a modular Python framework designed to simplify the development of generative AI applications. It provides a flexible set of abstractions for working with Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and various other AI components. With its structured approach, Griptape enables developers to build sophisticated AI agents and workflows efficiently.
torchchat: Run PyTorch LLMs Locally on Servers, Desktop, and Mobile
torchchat is a PyTorch-native codebase designed to showcase the ability to run large language models (LLMs) seamlessly across various platforms. It enables local execution of LLMs using Python, within C/C++ applications on desktop or servers, and directly on iOS and Android devices. Although no longer under active development, it remains a valuable resource for understanding and implementing local LLM deployment strategies.
Docling: Streamline Document Processing for Generative AI Applications
Docling is a powerful Python library designed to simplify document processing, preparing diverse formats for generative AI applications. It offers advanced parsing capabilities, including sophisticated PDF understanding, and provides a unified document representation. With seamless integrations into the AI ecosystem, Docling empowers developers to build robust AI solutions.
DeepFabric: High-Quality Synthetic Data for Agentic AI Systems
DeepFabric is an open-source Python library designed to generate high-quality synthetic training data for language models and agent evaluations. It excels at creating domain-specific datasets that teach models to think, plan, and act effectively, including correct tool usage and adherence to schema structures. This comprehensive pipeline also integrates training and evaluation capabilities, ensuring robust model development.
OpenMontage: The First Open-Source, Agentic Video Production System
OpenMontage is the world's first open-source, agentic video production system, designed to transform your AI coding assistant into a full video production studio. It features 12 pipelines, 52 tools, and over 500 agent skills, enabling end-to-end video creation from a simple prompt. This powerful tool handles research, scripting, asset generation, editing, and final composition, including the unique ability to produce real video from stock footage.
MarkLLM: An Open-Source Toolkit for LLM Watermarking
MarkLLM is an open-source toolkit designed to simplify the research and application of watermarking technologies for large language models (LLMs). It offers a unified framework for implementing various watermarking algorithms, alongside robust visualization and comprehensive evaluation tools. This toolkit helps researchers and the broader community understand and assess the authenticity and origin of machine-generated text.
Agent-Reach: Empower Your AI Agents with Internet Access, Zero API Fees
Agent-Reach is a powerful GitHub repository that equips AI agents with the ability to access and search the entire internet, including platforms like Twitter, Reddit, YouTube, and Bilibili. It provides a streamlined CLI experience, eliminating the need for complex API configurations and associated fees. This project ensures your AI agent can "see" and interact with web content effortlessly.
REAL Video Enhancer: AI-Powered Video Interpolation, Upscaling, and Denoising
REAL Video Enhancer is a powerful open-source application designed to enhance video quality across Linux, Windows, and macOS. It leverages AI models for advanced video processing tasks such as frame interpolation, upscaling, decompression, and denoising. This tool provides a modern alternative to older software, making high-quality video enhancement accessible to a wider audience.
NVIDIA PhysicsNeMo: Deep Learning Framework for Physics-ML Models
NVIDIA PhysicsNeMo is an open-source deep learning framework designed for building, training, and fine-tuning Physics AI models. It leverages state-of-the-art scientific machine learning methods, enabling real-time predictions by combining physics knowledge with data. This framework provides scalable, GPU-optimized tools for AI4Science and engineering applications.
Open-LLM-VTuber: Your Offline, Cross-Platform AI Companion with Live2D
Open-LLM-VTuber is an innovative open-source project that brings a customizable, voice-interactive AI companion to your desktop. It supports real-time voice interaction, visual perception, and dynamic Live2D avatars, all running completely offline across Windows, macOS, and Linux. This project allows users to create a personalized AI companion with extensive model support and advanced interaction features, ensuring privacy and flexibility.
Claude Ads: AI-Powered Paid Advertising Audit for Claude Code
Claude Ads is an AI-powered skill for Claude Code designed to automate comprehensive paid advertising audits. It performs over 250 checks across major ad platforms, providing a weighted health score, prioritized action plans, and AI creative generation. This tool drastically reduces the time and cost associated with manual ad account analysis.