Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train Qwen3, Llama 4, DeepSeek-R1, Gemma 3, TTS 2x faster with 70% less VRAM.

SGLang is a fast serving framework for large language models and vision language models.

Llama中文社区，实时汇总最新Llama学习资料，构建最好的中文Llama大模型开源生态，完全开源可商用

Scira (Formerly MiniPerplx) is a minimalistic AI-powered search engine that helps you find information on the internet and cites it too. Powered by Vercel AI SDK! Search with models like xAI's Grok 3.

Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 500+ LLMs (Qwen3, Qwen3-MoE, Llama4, InternLM3, DeepSeek-R1, ...) and 200+ MLLMs (Qwen2.5-VL, Qwen2.5-Omni, Qwen2-Audio, Ovis2, InternVL3, Llava, GLM4v, Phi4, ...) (AAAI 2025).

Open Source Deep Research Alternative to Reason and Search on Private Data. Written in Python.

A native macOS app that allows users to chat with a local LLM that can respond with information from files, folders and websites on your Mac without installing any other software. Powered by llama.cpp.

A simple, performant and scalable Jax LLM!

Make Discord your LLM frontend ● Supports any OpenAI compatible API (Ollama, LM Studio, vLLM, OpenRouter, xAI, Mistral, Groq and more)

Building LLaMA 4 MoE from Scratch

Efficient computer use agent powered by Meta Llama 4 Maverick

Rushikesh Nimkar's portfolio, accessible at rushikeshnimkar.xyz, is a modern website built with Next.js 15 and TypeScript. It features AI-powered email generation, an interactive chat with an AI version of Rushikesh, dynamic animations, responsive design, dark mode, and showcases his projects and contributions.

Comprehensive benchmark of 44 open source language models across creative writing, logic puzzles, counterfactual reasoning, and programming tasks. Tested on Apple M4 Max with detailed performance analysis.

Combining OCR for text extraction with LLMs for accurate, efficient document structuring.

The User Interface for NYTimes IntelliNews

Repository for BabyLM competition on 3 models in Strict and Strict-small tracks

MultiPage Invoice Parser

This project demonstrates three progressively advanced approaches to converting unstructured data into structured formats using AI