Repository History
206 repositories tagged with AI
Colibri: Run 744B GLM-5.2 MoE on Consumer Machines with Pure C
Colibri is an innovative project that enables running the massive 744B-parameter GLM-5.2 Mixture-of-Experts (MoE) model on consumer-grade machines with as little as 25GB of RAM. It achieves this remarkable feat through a pure C engine with zero dependencies, streaming model experts from disk on demand. This allows users to interact with a frontier-class LLM without requiring expensive GPU hardware.
Lamini: The Official Python Client for Generative AI API
Lamini is the official Python client and SDK designed to interact with the Lamini API, enabling developers to create their own Generative AI applications. It provides a straightforward interface for integrating powerful AI capabilities into Python projects. This package simplifies the process of building and deploying generative AI solutions.
RL4LMs: A Modular RL Library for Fine-tuning Language Models
RL4LMs is a powerful and modular reinforcement learning library designed to fine-tune language models to human preferences. It offers easily customizable building blocks for training, including on-policy algorithms, reward functions, and metrics. Thoroughly tested and benchmarked, RL4LMs supports a wide range of NLP tasks and models.
torchtune: PyTorch Native Library for LLM Post-Training and Experimentation
torchtune is a PyTorch native library designed for authoring, post-training, and experimenting with Large Language Models (LLMs). It offers hackable training recipes, simple PyTorch implementations of popular LLMs, and best-in-class memory efficiency. Please note: torchtune is no longer actively maintained as of 2025.
RouteLLM: Optimize LLM Costs and Maintain Quality with Intelligent Routing
RouteLLM is a powerful framework designed to serve and evaluate LLM routers, enabling significant cost savings without compromising response quality. It intelligently routes simpler queries to cheaper models while maintaining high performance, offering a drop-in replacement for existing OpenAI clients or a compatible server. This solution helps balance the dilemma of LLM deployment costs versus model capabilities.
RAGChecker: A Fine-grained Framework for Diagnosing RAG Systems
RAGChecker is an advanced automatic evaluation framework developed by Amazon Science, specifically designed to assess and diagnose Retrieval-Augmented Generation (RAG) systems. It offers a comprehensive suite of metrics and tools for in-depth analysis of RAG performance. This framework empowers developers and researchers to thoroughly evaluate and enhance their RAG systems with precision.
LLM Compressor: Optimize LLMs for Deployment with vLLM
LLM Compressor is a Transformers-compatible Python library designed to apply various compression algorithms to Large Language Models (LLMs). It enables optimized deployment, especially with vLLM, by offering a comprehensive set of quantization techniques for weights, activations, and KV Cache. This tool seamlessly integrates with Hugging Face models, making LLM optimization accessible and efficient.
Llama Cloud Services: Knowledge Agents and Management in the Cloud
Llama Cloud Services offers tools for building knowledge agents and managing data in the cloud. It provides robust capabilities for parsing various document types, including PDF, DOCX, and PPTX, into structured formats. Users should note that this repository is deprecated, with migration recommended to the new `llama-cloud` packages for continued support and improved performance.
OpenWebAgent: An Open Toolkit for LLM- and LMM-based Web Agents
OpenWebAgent is an open toolkit designed to empower model-based web agents, streamlining human-computer interactions by automating tasks on webpages. It offers a convenient framework for developing LLM- and LMM-based web agents, providing both plugin and server source code for easy integration and customization. This project was featured as an ACL'24 Demo, showcasing its innovative approach to web automation.
ChatArena: Multi-Agent Language Game Environments for LLMs
ChatArena is a Python library designed to provide multi-agent language game environments for Large Language Models (LLMs), aiming to foster the development of communication and collaboration capabilities in AI. It offers a flexible framework for defining players, environments, and interactions based on Markov Decision Processes. Please note that as of August 11, 2025, this project has been deprecated due to a lack of widespread community use and is no longer receiving updates or support.
Agentarium: A Python Framework for AI Agent Simulations
Agentarium is an open-source Python framework designed for creating and managing simulations with AI-powered agents. It offers an intuitive platform for designing complex, interactive environments where agents can act, learn, and evolve. This powerful tool simplifies the orchestration of multiple AI agents and their interactions.
Lighteval: Your All-in-One Toolkit for LLM Evaluation
Lighteval is a comprehensive toolkit from Hugging Face for evaluating Large Language Models (LLMs) across various backends. It enables users to dive deep into model performance by saving detailed, sample-by-sample results and supports over 1000 evaluation tasks. The framework offers extensive customization options, allowing users to create custom tasks and metrics tailored to their specific needs.