- tensorflow by tensorflow, 202k stars. Open source machine learning framework for building and training models
- whisper by openai, 113k stars. Speech recognition model trained on large-scale weakly supervised audio
- LLMs-from-scratch by rasbt, 108k stars. Step-by-step notebooks for building a GPT-style LLM in PyTorch
- pytorch by pytorch, 106k stars. Python deep learning framework with tensors, autograd and GPU acceleration
- llm-course by mlabonne, 85k stars. Course with roadmaps and Colab notebooks for learning large language models
- unsloth by unslothai, 80k stars. Local UI for running and fine-tuning LLMs and diffusion models
- models by tensorflow, 78k stars. Official collection of example and reference models built with TensorFlow
- AI-For-Beginners by microsoft, 71k stars. 12-week course with 24 lessons on AI fundamentals and practice
- annotated_deep_learning_paper_implementations by labmlai, 68k stars. 60+ deep learning paper implementations in PyTorch with side-by-side notes
- supervision by roboflow, 52k stars. Reusable Python helpers for computer vision tasks like detection and annotation
- Kronos by shiyu-coder, 48k stars. Foundation model trained on financial market data, in Python
- DeepSpeed by deepspeedai, 43k stars. Library for distributed training and inference of large PyTorch models
- heretic by p-e-w, 41k stars. Training-free tool that edits language models to reduce refusals and sycophancy
- laya by NandhaKishorM, 36k stars. One-pass model that returns typed choices, scores and yes/no answers for text
- detectron2 by facebookresearch, 35k stars. Library for object detection and segmentation built on PyTorch
- netron by lutzroeder, 34k stars. Viewer that draws the graph of neural network and ML model files
- Unlimited-OCR by baidu, 31k stars. OCR project for parsing long documents into text in a single pass
- label-studio by HumanSignal, 29k stars. Annotation tool for labeling text, images, audio and video data
- MiniCPM-V by OpenBMB, 27k stars. Vision-language model that runs on local devices for seeing and hearing
- open-r1 by huggingface, 26k stars. Open training pipeline that reproduces the DeepSeek-R1 reasoning model
- recommenders by recommenders-team, 22k stars. Examples and best practices for building recommendation systems
- awesome-production-machine-learning by EthicalML, 21k stars. Curated list of open source tools for deploying and monitoring ML
- Qwen3-VL by QwenLM, 20k stars. Series of multimodal vision-language models with notebook examples
- olmocr by allenai, 20k stars. Toolkit that converts PDFs into linear text for LLM training datasets
- ml-engineering by stas00, 20k stars. Open book on training and scaling large language models on clusters
- ktransformers by kvcache-ai, 20k stars. Framework for LLM inference and fine-tuning on mixed hardware
- LightGBM by lightgbm-org, 19k stars. Fast gradient boosting library for ranking, classification and regression
- agent-lightning by microsoft, 19k stars. Reinforcement learning framework for training existing AI agents
- Speech by NVIDIA-NeMo, 19k stars. Framework for speech recognition, text-to-speech, and translation models
- lingbot-map by Robbyant, 19k stars. Transformer model for streaming 3D reconstruction from video input
- WeClone by xming521, 18k stars. Trains a chat model on your message history to talk like you
- face-api.js by justadudewhohacks, 18k stars. Face detection and recognition in the browser and Node.js using TensorFlow.js
- labelme by wkentaro, 16k stars. Image annotation tool for polygons, boxes and AI-assisted labels
- OpenMythos by kyegomez, 15k stars. Theoretical reconstruction of a Claude-style model architecture from papers
- train-llm-from-scratch by FareedKhan-dev, 13k stars. Walkthrough for training a language model from scratch, from data to generation
- 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
- kev by jaredpalmer, 10.0k stars. Decision models built on Qwen3 that you can train and run yourself
- OBLITERATUS by elder-plinius, 9.5k stars. Tool for removing refusal behavior from open-weight language models
New repos every two days, stars refreshed daily. Visits are counted without cookies, and the numbers are public. How it works Data RSS Newsletter