- unsloth by unslothai, 80k stars. Local UI for running and fine-tuning LLMs and diffusion models
- DeepSpeed by deepspeedai, 43k stars. Library for distributed training and inference of large PyTorch models
- airllm by lyogavin, 42k stars. Runs large language models layer by layer so they fit in a small GPU's memory
- heretic by p-e-w, 41k stars. Training-free tool that edits language models to reduce refusals and sycophancy
- open-r1 by huggingface, 26k stars. Open training pipeline that reproduces the DeepSeek-R1 reasoning model
- ktransformers by kvcache-ai, 20k stars. Framework for LLM inference and fine-tuning on mixed hardware
- agent-lightning by microsoft, 19k stars. Reinforcement learning framework for training existing AI agents
- train-llm-from-scratch by FareedKhan-dev, 13k stars. Walkthrough for training a language model from scratch, from data to generation
- OpenLLM by bentoml, 13k stars. Serve open-source LLMs like Llama and DeepSeek behind an OpenAI-compatible API
- amazon-sagemaker-examples by aws, 11k stars. Jupyter notebook examples for training and deploying ML models on SageMaker
- ART by OpenPipe, 11k stars. Reinforcement learning library for training multi-step LLM agents with GRPO
- Soup by MakazhanAlpamys, 11k stars. Fine-tunes LLMs from a YAML config, streaming layers to fit small GPUs
- Sana by NVlabs, 9.8k stars. Efficient diffusion models and inference engines for image and video
- slime by THUDM, 8.8k stars. Reinforcement learning framework for scaling LLM training
- notebooks by unslothai, 5.8k stars. Colab and Jupyter notebooks for fine-tuning text, vision, audio and TTS models
- transformerlab-app by transformerlab, 5.2k stars. Desktop app to train, evaluate and scale models from laptop to GPU cluster
- Awesome-LLM-Post-training by mbzuai-oryx, 2.6k stars. Curated list of papers and guides on reasoning and post-training for LLMs
- jeff by firelex, 1.6k stars. Small open model that picks between options with calibrated probabilities
- microgpt-c by vixhal-baraiya, 853 stars. Dependency-free GPT training and inference in plain C
- LoongForge by baidu-baige, 846 stars. Framework for pretraining and tuning LLMs, VLMs and embodied models on GPUs
- overmind by overmind-core, 759 stars. Platform for evaluating and fine-tuning LLM agents over time
- ai-engineer-notebooks by calmrocks, 749 stars. Framework-free Colab notebooks on model APIs, RAG, evals and agents
- TetreesAgent_EX by TetreesEX, 289 stars. MCP server for training external agent intelligence and trading it on Tetrees EX
- llm-rl-environments-lil-course by anakin87, 245 stars. Short course on building reinforcement learning environments for LLMs
- ai-trains-ai by Danau5tin, 243 stars. RL training setup where an agent is trained to train other AI agents
- High-Angle_Robust_Fast_FaceAlignment by PINTO0309, 206 stars. Dataset and training code for aligning faces at high head angles
- Mixture-of-LoRA-Harness by MindLab-Research, 199 stars. Serving system that routes and composes many LoRA adapters on one base model
- lora-speedrun by Saivineeth147, 151 stars. Public wall-clock leaderboard for LoRA fine-tuning on a fixed task
- autoresearch-rl by vivekvkashyap, 113 stars. Reinforcement learning loop for automated research with an agent
- deleting-weights by agentic-learning-ai-lab, 101 stars. Paper code on deleting weights in LLM post-training
- LLM by The-School-of-AI, 71 stars. Reference pretraining pipeline for small to mid-size mixture-of-experts language models
- nano-llm-posttraining by pochenai, 32 stars. Small LLM post-training experiments on one 8GB GPU, measuring forgetting and variance
- nanochat-jax by tucan9389, 12 stars. Low-cost ChatGPT-style model training on TPUs using JAX
- GADC by 2018cx, 11 stars. PyTorch code for dataset condensation to train diffusion models, ICML 2026
New repos every two days, stars refreshed daily. Visits are counted without cookies, and the numbers are public. How it works Data RSS Newsletter