Open-source deep-learning software
2785 open-source libraries, tools and applications tagged deep-learning, ranked by how closely deep-learning is their primary purpose, then by adoption and maintenance health. Each entry links to a page with sources, use cases and when to choose or avoid it.
2785 products, primary matches first, then adoption-weighted; health v2 shown.
| Product | Health v2 | Stars | Maturity |
|---|---|---|---|
| Omni-Avatar/OmniAvatar OmniAvatar is an audio-driven full-body avatar video generation model built on Wan2.1 text-to-video diffusion models with LoRA-based audio … | 31 | 1862 | active |
| laekov/fastmoe FastMoE is a PyTorch library providing efficient Mixture of Experts (MoE) layers with custom C/CUDA operators. It supports distributed expe… | 23 | 1860 | active |
| OpenNMT OpenNMT is an open-source ecosystem for neural machine translation and sequence learning, with PyTorch (OpenNMT-py) and TensorFlow (OpenNMT… | 40 | 7018 | maintenance |
| LuChengTHU/dpm-solver Official PyTorch implementation of DPM-Solver and DPM-Solver++, fast high-order ODE solvers for diffusion probabilistic model sampling that… | 32 | 1859 | stable |
| TorchSharp TorchSharp is a .NET library providing bindings to LibTorch, the library that powers PyTorch, with a focus on tensors and a PyTorch-like AP… | 69 | 1858 | active |
| facebookresearch/MetaCLIP Meta's research code and models for Meta CLIP, a reimplementation and scaling recipe for CLIP-style contrastive vision-language models, inc… | 76 | 1854 | active |
| trevin-creator/autoresearch-mlx An Apple Silicon (MLX) port of Karpathy's autoresearch that runs autonomous AI research loops on a Mac without PyTorch or CUDA. A coding ag… | 52 | 1852 | active |
| probcomp/Gen.jl Gen.jl is a general-purpose probabilistic programming system embedded in Julia that lets users write generative models as probabilistic pro… | 58 | 1851 | active |
| Tencent-Hunyuan/HunyuanVideo-I2V HunyuanVideo-I2V is Tencent's open-source image-to-video generation framework built on the HunyuanVideo diffusion model, providing PyTorch … | 51 | 1845 | active |
| NVlabs/stylegan3 Official PyTorch implementation of StyleGAN3 (Alias-Free GANs), a state-of-the-art generative adversarial network for high-fidelity image s… | 32 | 6945 | maintenance |
| NVIDIA/pix2pixHD PyTorch implementation of pix2pixHD, a conditional GAN method for synthesizing and manipulating high-resolution (2048x1024) photorealistic … | 32 | 6925 | maintenance |
| ytongbai/LVM LVM is a large vision model trained with sequential next-token prediction over 'visual sentences', using no linguistic data. It builds on O… | 31 | 1836 | active |
| cazala/synaptic Synaptic is an architecture-free neural network library for JavaScript that runs in both Node.js and the browser. It supports building and … | 62 | 6912 | maintenance |
| tile-ai/TileRT TileRT is a tile-based runtime for ultra-low-latency LLM inference, achieving hundreds to over 1000 tokens/s decode speeds on frontier mode… | 75 | 1831 | active |
| ZFTurbo/Weighted-Boxes-Fusion A Python library implementing several methods for ensembling bounding boxes from multiple object detection models, including Non-maximum Su… | 66 | 1831 | stable |
| openai/point-e Point-E is OpenAI's official release of models and code for generating 3D point clouds from text prompts or images using diffusion models. … | 32 | 6892 | maintenance |
| VAST-AI-Research/TripoSG TripoSG is an open-source image-to-3D generation foundation model that produces high-fidelity 3D meshes from single images using large-scal… | 24 | 1815 | active |
| zju3dv/4K4D 4K4D is a research implementation of a 4D point cloud representation for real-time dynamic view synthesis at up to 4K resolution, built on … | 70 | 1814 | active |
| apple/ml-4m 4M is a framework from Apple and EPFL for training any-to-any multimodal foundation models using masked modeling over discrete tokens acros… | 68 | 1814 | active |
| triple-mu/YOLOv8-TensorRT A library for running YOLOv8 inference accelerated with NVIDIA TensorRT, supporting detection, segmentation, pose estimation, oriented boun… | 72 | 1812 | active |
| RAVE RAVE is the official PyTorch implementation of a realtime audio variational autoencoder for fast, high-quality neural audio synthesis. It s… | 51 | 1811 | active |
| yerfor/GeneFacePlusPlus GeneFace++ is the official PyTorch implementation of a NeRF-based system for generalized and stable real-time 3D talking face generation. I… | 26 | 1808 | active |
| thu-ml/RoboticsDiffusionTransformer RDT-1B is a 1B-parameter diffusion foundation model for robot bimanual manipulation, pre-trained on 1M+ multi-robot episodes to predict rob… | 48 | 1806 | active |
| Vitis AI AMD Vitis AI is an integrated development environment and stack for accelerating AI inference on AMD/Xilinx adaptable platforms, including … | 64 | 1805 | active |
| HuangJunJie2017/BEVDet BEVDet is a Python research codebase implementing the BEVDet series of bird's-eye-view (BEV) 3D object detection models for autonomous driv… | 23 | 1805 | active |
| lehaifeng/T-GCN A collection of research source code implementing Temporal Graph Convolutional Networks (T-GCN) and related variants for urban traffic flow… | 46 | 1787 | active |
| octo-models/octo Octo is an open-source generalist robot policy: a transformer-based diffusion policy pretrained on 800k robot trajectories from the Open X-… | 18 | 1787 | active |
| CalculatedContent/WeightWatcher WeightWatcher is an open-source Python diagnostic tool for analyzing pre-trained deep neural networks without access to training or test da… | 56 | 1777 | active |
| Totoro97/NeuS Official PyTorch implementation of NeuS, a neural implicit surface reconstruction method that learns SDF-based surfaces via volume renderin… | 32 | 1775 | stable |
| huchenlei/ComfyUI-layerdiffuse A ComfyUI custom node plugin implementing LayerDiffuse, enabling generation of transparent images with RGBA output and foreground/backgroun… | 26 | 1771 | active |
| character-ai/Ovi Ovi is a video-plus-audio generation model from Character AI that simultaneously generates synchronized video and audio from text or text+i… | 38 | 1763 | active |
| AnswerDotAI/ModernBERT ModernBERT is the research repository for a modernized BERT-family bidirectional encoder trained on 2 trillion tokens with an 8192-token co… | 52 | 1758 | active |
| xinntao/ESRGAN ESRGAN (Enhanced SRGAN) is a PyTorch-based image super-resolution model that won the PIRM 2018 Challenge on Perceptual Super-Resolution. Th… | 32 | 6571 | maintenance |
| Maximal Update Parametrization (μP) The `mup` Python package implements Maximal Update Parametrization (μP) for PyTorch models, enabling optimal hyperparameters to remain stab… | 10 | 1756 | stable |
| T5 The official T5 library from Google Research, implementing the text-to-text transfer transformer for NLP tasks like summarization, question… | 67 | 6562 | maintenance |
| SamsungSAILMontreal/TinyRecursiveModels Official codebase for the Tiny Recursive Model (TRM) paper, a recursive reasoning approach where a tiny 7M-parameter neural network iterati… | 10 | 6561 | maintenance |
| mahmoodlab/CLAM CLAM is an open-source Python toolkit for data-efficient, weakly supervised classification of whole-slide images (WSIs) in computational pa… | 36 | 1755 | active |
| webonnx/wonnx Wonnx is a GPU-accelerated ONNX inference runtime written entirely in Rust, built on wgpu and usable natively or in the browser via WebGPU … | 10 | 1754 | active |
| facebookresearch/metaseq Metaseq is a PyTorch codebase from Meta AI for training and working with large-scale Open Pre-trained Transformers (OPT), forked from fairs… | 10 | 6544 | maintenance |
| CompVis/taming-transformers The official implementation of 'Taming Transformers for High-Resolution Image Synthesis' (CVPR 2021), combining a convolutional VQGAN codeb… | 32 | 6530 | maintenance |
| codertimo/BERT-pytorch A PyTorch implementation of Google AI's 2018 BERT model with simple, readable code. It provides CLI tools for building vocabulary and pre-t… | 23 | 6529 | maintenance |
| Informer The official PyTorch implementation of Informer, an efficient Transformer architecture for long sequence time-series forecasting that won t… | 41 | 6528 | maintenance |
| Stability-AI/StableCascade Official codebase for Stable Cascade, a text-to-image generation model built on the Würstchen architecture with a highly compressed latent … | 26 | 6528 | maintenance |
| rusty1s/pytorch_scatter A PyTorch extension library providing highly optimized scatter and segment (sparse update) operations with sum, mean, min, and max reductio… | 57 | 1747 | stable |
| TencentARC/BrushNet BrushNet is the official PyTorch implementation of an ECCV 2024 plug-and-play image inpainting model that embeds pixel-level masked image f… | 26 | 1744 | active |
| facebookresearch/multimodal TorchMultimodal is a PyTorch library from Meta for training state-of-the-art multimodal multi-task models at scale, covering both content u… | 77 | 1740 | active |
| SHI-Labs/OneFormer OneFormer is a universal image segmentation framework (CVPR 2023) that unifies semantic, instance, and panoptic segmentation in a single tr… | 32 | 1740 | stable |
| MultimediaTechLab/YOLO Official MIT-licensed implementation of the YOLOv9, YOLOv7, and YOLO-RD real-time object detection models, including pre-trained weights, t… | 51 | 1738 | active |
| Xiaojiu-z/EasyControl EasyControl is the official implementation of an ICCV 2025 paper adding efficient and flexible conditional control to Diffusion Transformer… | 32 | 1738 | active |
| openai/consistency_models Official PyTorch implementation of Consistency Models, a generative image model family from OpenAI supporting consistency distillation, con… | 10 | 6483 | maintenance |
| google/automl Google Brain's AutoML repository containing implementations of AutoML models and libraries such as EfficientNet, EfficientNetV2, and Effici… | 10 | 6474 | maintenance |
| muscriptor/muscriptor MuScriptor is a multi-instrument music transcription model by Kyutai and Mirelo that converts audio recordings into MIDI and sheet music. I… | 78 | 1732 | active |
| cleverhans-lab/cleverhans CleverHans is a Python library for benchmarking machine learning systems' vulnerability to adversarial examples, providing reference implem… | 23 | 6455 | maintenance |
| NVIDIA/FasterTransformer NVIDIA's highly optimized C++/CUDA library for fast inference of Transformer-based models such as BERT, GPT, and encoder-decoder models, wi… | 23 | 6450 | maintenance |
| AvaLovelace1/BrickGPT BrickGPT is the official implementation of an ICCV 2025 Best Paper approach that generates physically stable, buildable toy brick (LEGO) mo… | 54 | 1726 | active |
| meta-pytorch/tnt TNT (torchtnt) is a lightweight library from Meta providing tools and utilities for PyTorch training workflows. It offers building blocks f… | 77 | 1723 | active |
| facebookresearch/coconut Official PyTorch implementation of Coconut, a method for training large language models to reason in a continuous latent space instead of e… | 58 | 1723 | active |
| facebookresearch/ConvNeXt Official PyTorch implementation of ConvNeXt, a pure convolutional neural network architecture from the CVPR 2022 paper 'A ConvNet for the 2… | 10 | 6408 | maintenance |
| tenstorrent/tt-metal TT-Metal is Tenstorrent's open-source software stack containing TT-NN, a Python and C++ neural network operator library, and TT-Metalium, a… | 99 | 1718 | active |
| Enzyme Enzyme is a high-performance automatic differentiation plugin for LLVM and MLIR that computes derivatives and gradients of arbitrary existi… | 95 | 1718 | active |
| lightly-ai/lightly-train LightlyTrain is a Python framework for training computer vision models, covering pretraining of vision foundation models (DINOv2/v3) on unl… | 83 | 1716 | active |
| sihyun-yu/REPA REPA is the official PyTorch implementation of the ICLR 2025 paper 'Representation Alignment for Generation', a regularization technique th… | 24 | 1716 | active |
| cocodataset/cocoapi Official API for the COCO (Common Objects in Context) dataset, providing Matlab, Python, and Lua interfaces to load, parse, and visualize C… | 32 | 6387 | maintenance |
| kingoflolz/mesh-transformer-jax A JAX/Haiku library implementing model-parallel training and inference of transformer models using xmap/pjit operators, similar to Megatron… | 32 | 6384 | maintenance |
| lsdefine/simple_GRPO A minimal (~200 lines) Python implementation of GRPO reinforcement learning for training LLMs to develop r1-like reasoning, built on PyTorc… | 41 | 1714 | active |
| feizc/FluxMusic FluxMusic is the official PyTorch implementation of a rectified flow Transformer model for text-to-music generation, from the paper 'Flux t… | 24 | 1709 | active |
| iMoonLab/yolov13 Official PyTorch implementation of YOLOv13, a real-time object detection model family (Nano to X-Large) featuring Hypergraph-based Adaptive… | 29 | 1708 | active |
| jiaweizzhao/GaLore GaLore is a PyTorch library implementing Gradient Low-Rank Projection for memory-efficient full-parameter training of large language models… | 26 | 1708 | active |
| omerbt/TokenFlow TokenFlow is the official PyTorch implementation of an ICLR 2024 paper for text-driven, temporally consistent video editing using a pre-tra… | 29 | 1706 | stable |
| BindsNET/bindsnet BindsNET is a Python package for simulating spiking neural networks (SNNs) built on PyTorch tensor functionality, running on CPUs or GPUs. … | 81 | 1704 | active |
| PaddlePaddle/PaddleVideo PaddleVideo is a video understanding toolkit built on PaddlePaddle, offering state-of-the-art models for action recognition, temporal actio… | 23 | 1702 | active |
| lucidrains/alphafold3-pytorch A PyTorch implementation of Google DeepMind's Alphafold 3 model for protein and molecular structure prediction. It is a research-oriented r… | 67 | 1699 | active |
| shubham-goel/4D-Humans 4DHumans is a Python research codebase implementing HMR 2.0, a transformer-based model for 3D human mesh recovery from single images, plus … | 56 | 1695 | active |
| TensorPack Tensorpack is a high-level neural network training interface built on graph-mode TensorFlow, focused on training speed and flexibility for … | 23 | 6284 | maintenance |
| elixir-nx/axon Axon is a neural network library for Elixir built on top of Nx numerical definitions. It provides a functional API, a high-level model crea… | 89 | 1692 | active |
| intelligent-machine-learning/dlrover DLRover is an automatic distributed deep learning system that manages training of large AI models on Kubernetes and Ray clusters. It provid… | 80 | 1689 | active |
| takuseno/d3rlpy d3rlpy is a Python library for offline and online deep reinforcement learning built on PyTorch, offering state-of-the-art algorithms throug… | 45 | 1684 | active |
| franciszzj/Leffa Leffa is a diffusion-based framework for controllable person image generation, supporting virtual try-on and pose transfer via a regulariza… | 37 | 1683 | active |
| google/vizier Open Source Vizier is a Python-based service and library for black-box and hyperparameter optimization, based on Google's internal Vizier t… | 79 | 1680 | active |
| maziarraissi/PINNs The original reference implementation of Physics-Informed Neural Networks (PINNs), which train neural networks to solve and discover nonlin… | 58 | 6217 | maintenance |
| microsoft/Mage Mage is a family of lightweight 4B-parameter multimodal models from Microsoft, including Mage-VL for image and video understanding and Mage… | 55 | 1677 | active |
| elixir-nx/bumblebee Bumblebee is an Elixir library providing pre-trained neural network models built on Axon, with integration for downloading models from Hugg… | 94 | 1674 | active |
| sunlabuiuc/PyHealth PyHealth is an open-source Python toolkit for clinical deep learning, unifying healthcare datasets (EHRs, physiological signals, medical im… | 95 | 1673 | active |
| Gen-Verse/MMaDA MMaDA is an open-source family of multimodal large diffusion language models that unify textual reasoning, multimodal understanding, and te… | 46 | 1673 | active |
| ZrrSkywalker/Personalize-SAM PerSAM is the official implementation of 'Personalize Segment Anything Model with One Shot', which customizes the Segment Anything Model (S… | 30 | 1673 | active |
| intel/auto-round AutoRound is Intel's advanced quantization toolkit for LLMs and vision-language models, using sign-gradient descent to achieve high accurac… | 94 | 1672 | active |
| zihangdai/xlnet XLNet is a generalized autoregressive pretraining method for language representation learning, built on Transformer-XL. This repo provides … | 32 | 6189 | maintenance |
| Tencent/AngelSlim AngelSlim is a Python toolkit from Tencent for compressing large language models and related architectures (VLMs, diffusion, audio models) … | 70 | 1669 | active |
| mit-han-lab/torchquantum TorchQuantum is a PyTorch-based framework for quantum computing simulation, supporting statevector and pulse-level simulation on GPUs with … | 60 | 1669 | active |
| tkarras/progressive_growing_of_gans Official TensorFlow implementation of the ICLR 2018 NVIDIA paper 'Progressive Growing of GANs', which trains generators and discriminators … | 32 | 6176 | maintenance |
| JiuhaiChen/BLIP3o Official implementation of the BLIP3o-Series, a unified autoregressive-plus-diffusion model for text-to-image generation and editing. It co… | 41 | 1667 | active |
| thtrieu/darkflow Darkflow is a Python library that translates Darknet's YOLO neural network definitions to TensorFlow, enabling real-time object detection a… | 32 | 6139 | maintenance |
| thunil/TecoGAN TecoGAN is the official source code for a temporally coherent GAN for video super-resolution, published at SIGGRAPH/ACM TOG. It includes in… | 32 | 6133 | maintenance |
| XueZeyue/DanceGRPO Official implementation of DanceGRPO, a framework applying Group Relative Policy Optimization (GRPO) to fine-tune visual generation models … | 37 | 1654 | active |
| chineseocr A Python OCR toolkit that combines YOLO3-based text detection with CRNN/Dense recognition for Chinese and English text in natural scene ima… | 32 | 6113 | maintenance |
| taki0112/UGATIT Official TensorFlow implementation of U-GAT-IT, an unsupervised image-to-image translation model using attention modules and adaptive layer… | 32 | 6113 | maintenance |
| mhamilton723/FeatUp FeatUp is a model-agnostic framework that upsamples the spatial resolution of deep neural network features by 16-32x without changing their… | 17 | 1653 | active |
| GoMLX GoMLX is an accelerated machine learning and math framework for Go, comparable to PyTorch/JAX/TensorFlow. It offers differentiable operator… | 96 | 1651 | active |
| InterDigitalInc/CompressAI CompressAI is a PyTorch library and evaluation platform for end-to-end learned data compression research. It provides custom layers, entrop… | 76 | 1648 | active |
| JIA-Lab-research/ControlNeXt ControlNeXt is the official implementation of a controllable generation method for images and videos, built on Stable Diffusion XL, Stable … | 24 | 1646 | active |