llama

The open world for autonomous AI agents on Solana Trade. Build. Fight. Earn. Explore. Connect your AI agent to a persistent shared world. Trade real SOL, build structures, form guilds, fight for territory, complete bounties, gather resources across 7 biomes. Not a simulation — everything is real.

Training LLaMA language model with MMEngine! It supports LoRA fine-tuning!

Preprint: Asymmetry in Low-Rank Adapters of Foundation Models

Ollama Zig library

SORSA: Singular Values and Orthonormal Regularized Singular Vectors Adaptation of Large Language Models

An interactive AI chatbot for querying and discussing the contents of PDF documents using Streamlit and OpenAI/other AI language models.

Piazza-Updater automates updates to a Weaviate database with real-time vectorial data. By continuously searching the internet and integrating with Verba repositories, it enhances retrieval-augmented generation (RAG) capabilities, keeping your applications informed and responsive.

Run 70B+ LLMs on Apple Silicon by using SSD as extended memory — intelligent layer streaming and caching for Mac

🚀 Welcome to AskSage Open Source Repo, a community aimed at providing code 🧑‍💻, documentation📝, and proof-of-concept projects for interacting with the AskSage API. This repository serves as an unofficial resource for paid subscribers seeking practical guidance on utilizing the API. If you have any questions email us support@asksage.ai

Universal local LLM runtime. Drop-in Ollama replacement with HuggingFace integration, pipelines, structured output. Free & open source.

Caption Creator is a fast and portable tool for generating high-quality image captions and tags - ideal for custom dataset creation. Works seamlessly for both training and image generation.

CSVs of the Huggingface and LMArena LLM leaderboards, along with the code to generate them in Python and R.

A fluent, scalable, and easy-to-use LLM data processing framework.

a simple lightweight large language model pipeline framework.

Fine tune LLM with HuggingFace

Rental app with in-depth data analysis, interactive data visualization, predictive analytics, and chatbot using Dash. Selected as Plotly-Dash’s Top GenAI Project

Training-free KV cache compression via E8 lattice VQ. 2-bit KV that preserves retrieval (30/30 NIAH vs TurboQuant 0/30). Calibration-free, 9 architectures validated.

A lightweight and intuitive Node.js client for the Paxsenix AI API.

A distributed training framework for large language models powered by Lightning.

📘 Taiwan-LLM Tutor: Large Language Models for Taiwanese Secondary Education