Repository History
138 repositories tagged with LLM
Picotron: Minimalistic 4D-Parallelism Framework for LLM Training Education
Picotron is a minimalistic and hackable distributed training framework designed for educational purposes. Inspired by NanoGPT, it focuses on pre-training Llama-like models using 4D Parallelism, making complex concepts accessible. Its simple and readable codebase, with core files under 300 lines, provides an excellent tool for learning and experimentation in distributed machine learning.
Model Context Protocol TypeScript SDK: Build MCP Servers and Clients
The `modelcontextprotocol/typescript-sdk` is the official TypeScript SDK for interacting with Model Context Protocol (MCP) servers and clients. It provides a standardized way for applications to offer context to Large Language Models (LLMs), separating context provision from LLM interaction. Developers can use it to easily create MCP servers that expose resources, prompts, and tools, as well as build MCP clients to connect to any MCP server.
Aider: AI Pair Programming in Your Terminal
Aider is an open-source project that brings AI pair programming directly to your terminal, enabling developers to collaborate with large language models (LLMs). It helps in building new projects or enhancing existing codebases efficiently. With robust features like codebase mapping, Git integration, and multi-language support, Aider is a versatile tool for modern development workflows.
Ant Design X: Crafting AI-Driven Interfaces with React Components
Ant Design X is an innovative open-source project by Ant Design, designed to simplify the creation of AI-driven user interfaces. It offers a comprehensive suite of enterprise-level LLM components, tools for efficient data stream management, and a high-performance Markdown renderer. Built with TypeScript, it leverages React and Ant Design to provide a robust solution for modern AI applications.
Judgy: Correcting LLM Judge Bias for Reliable AI Model Evaluation
Judgy is a Python package designed to improve the reliability of evaluations performed by LLM-as-Judges. It provides tools to estimate the true success rate of a system by correcting for LLM judge bias and generating confidence intervals through bootstrapping. This helps ensure more accurate and trustworthy assessments of AI model performance.
AingDesk: A User-Friendly AI Assistant with Local Model and Knowledge Base Support
AingDesk is a rapidly growing, easy-to-use AI assistant designed for both beginners and experienced users. It offers robust features including support for local AI models, mainstream model APIs, knowledge bases, intelligent agents, and web search. This open-source project, built with TypeScript, simplifies the deployment and management of AI tools.
Open NotebookLM: Convert PDFs into Personalized Podcast Episodes
Open NotebookLM is an innovative open-source project that transforms any PDF document into an engaging podcast episode. Inspired by NotebookLM, it leverages powerful LLMs and text-to-speech models to generate natural dialogue from your documents. This tool provides a unique way to consume information, making learning and content absorption more accessible and enjoyable.
PyPriompt: Python Library for Priority-Based Prompt Design
PyPriompt is a Python library for designing prompts, inspired by web design libraries like React and FastHTML. It intelligently manages context windows by using priorities to decide what information to include in the prompt, ensuring efficient use of token limits. This tool helps developers create dynamic and adaptable prompts for large language models.
Attachments: The Python Funnel for LLM Context and Multimodal Data
Attachments simplifies providing context to Large Language Models by transforming various file types into model-ready text and images. This Python library acts as a universal funnel, enabling developers to integrate diverse data sources like PDFs, images, web content, and even entire code repositories with just a few lines of code. It supports popular LLM APIs and frameworks, making multimodal AI development more accessible.
CORE: A Unified Memory System for Your AI Applications
CORE by RedPlanetHQ is an open-source project designed to provide a persistent, unified memory layer for AI applications. It leverages a temporal knowledge graph to prevent context loss across various AI tools, ensuring LLMs retain past conversations, preferences, and project history. This system significantly enhances AI interactions by making context available across different sessions and platforms.
txtinstruct: Building Instruction-Tuned Models with Custom Data
txtinstruct is a Python framework designed for training instruction-tuned models. It focuses on supporting open data and models, enabling users to build their own instruction-following datasets and train models without licensing ambiguity. This project simplifies the process of creating custom instruction-tuned solutions.
Open R1: An Open-Source Reproduction of DeepSeek-R1 for Advanced LLM Training
Open R1 is a Hugging Face project dedicated to creating a fully open reproduction of DeepSeek-R1, a powerful reasoning language model. This initiative provides comprehensive tools and recipes for training, evaluating, and generating data for large language models. It fosters community collaboration in AI research, enabling developers to build upon and understand the complex R1 pipeline.