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 |
|---|---|---|---|
| leehomyc/Faster-High-Res-Neural-Inpainting Research code implementing the CVPR 2017 paper 'High-Resolution Image Inpainting using Multi-Scale Neural Patch Synthesis', written in Lua … | 32 | 1286 | abandoned |
| pluskid/Mocha.jl Mocha.jl is a deep learning framework for Julia, inspired by Caffe, providing modular layers, solvers, and support for training convolution… | 32 | 1284 | abandoned |
| Prinsphield/Wechat_AutoJump A Python bot that automatically plays the WeChat 'Jump Jump' mini-game using computer vision and a CNN coarse-to-fine model to locate the p… | 32 | 1280 | abandoned |
| rosejn/cortex Cortex is a Clojure machine learning library providing neural networks, regression, and feature learning on the JVM, with optional GPU acce… | 32 | 1272 | abandoned |
| hollance/Forge Forge is a Swift library that wraps Apple's MPSCNN (Metal Performance Shaders CNN) framework to make it easier to build and run deep neural… | 32 | 1265 | abandoned |
| gnes-ai/gnes GNES (Generic Neural Elastic Search) is a cloud-native framework for building semantic search systems over text, images, and video using de… | 10 | 1263 | abandoned |
| yahoo/CaffeOnSpark CaffeOnSpark is a Spark package that brings the Caffe deep learning framework to Hadoop and Spark clusters, enabling distributed neural net… | 10 | 1262 | abandoned |
| shekkizh/FCN.tensorflow A TensorFlow implementation of Fully Convolutional Networks (FCN) for semantic segmentation, based on the reference code from the original … | 32 | 1248 | abandoned |
| ivan-vasilev/neuralnetworks A Java library implementing deep learning algorithms and deep neural network architectures (MLPs, CNNs, RBMs, autoencoders, DBNs) with GPU … | 23 | 1236 | abandoned |
| atpaino/deep-text-corrector A TensorFlow research project that trains sequence-to-sequence deep learning models to correct small grammatical errors in short conversati… | 32 | 1235 | abandoned |
| google/prettytensor Pretty Tensor is a Python library providing a fluent, chainable builder API on top of TensorFlow for defining multi-layer neural networks. … | 10 | 1226 | abandoned |
| DmitryUlyanov/texture_nets Research code implementing the 'Texture Networks' paper for fast feed-forward texture synthesis and neural style transfer using Torch7. Onc… | 32 | 1224 | abandoned |
| facebookarchive/C3D C3D is a modified fork of BVLC Caffe that adds support for 3D convolution and pooling layers, enabling training and fine-tuning of 3D ConvN… | 10 | 1182 | abandoned |
| gpleiss/temperature_scaling A small PyTorch module implementing temperature scaling, a post-processing technique that calibrates the overconfident probability outputs … | 44 | 1180 | abandoned |
| hannes-brt/hebel Hebel is a GPU-accelerated deep learning library in Python built on PyCUDA, implementing feed-forward neural networks for classification an… | 23 | 1169 | abandoned |
| kootenpv/neural_complete Neural Complete is a generative LSTM-based autocomplete system trained on Python Keras source code to help write neural network code. It co… | 32 | 1160 | abandoned |
| Blocks Blocks is a Python framework for building and training neural network models on top of Theano. It provides parametrized operations called '… | 23 | 1151 | abandoned |
| NVIDIA/tensorflow NVIDIA's fork of TensorFlow 1.15 providing GPU support for newer NVIDIA hardware and CUDA libraries after Google stopped maintaining the TF… | 10 | 1147 | abandoned |
| eldar/pose-tensorflow A TensorFlow implementation of the DeeperCut and ArtTrack algorithms for human body pose estimation, supporting both single-person and mult… | 32 | 1139 | abandoned |
| omoindrot/tensorflow-triplet-loss A TensorFlow implementation of triplet loss with online triplet mining strategies (batch all and batch hard), for training embedding models… | 32 | 1127 | abandoned |
| szad670401/end-to-end-for-chinese-plate-recognition An end-to-end Chinese license plate recognition model based on MXnet, using multi-label classification. It was trained on ~500k synthetic r… | 32 | 1116 | abandoned |
| pbloem/former A minimal, from-scratch implementation of the transformer architecture in PyTorch, accompanying a tutorial blog post on how modern transfor… | 34 | 1098 | abandoned |
| carpedm20/DiscoGAN-pytorch A PyTorch implementation of DiscoGAN, a generative adversarial network that discovers cross-domain relations in an unsupervised manner. It … | 32 | 1096 | abandoned |
| spragunr/deep_q_rl A Theano/Lasagne-based implementation of the Deep Q-Network (DQN) algorithm from DeepMind's Atari reinforcement learning papers. It trains … | 32 | 1094 | abandoned |
| dmlc/minpy MinPy is a Python library providing a pure NumPy interface backed by MXNet, automatically dispatching operators to GPU when available and f… | 10 | 1094 | abandoned |
| MarvinTeichmann/tensorflow-fcn A single-file TensorFlow implementation of Fully Convolutional Networks (FCN) for semantic segmentation, using VGG16 weights loaded from a … | 32 | 1093 | abandoned |
| SegNet A modified version of the Caffe deep learning framework implementing SegNet, a deep convolutional encoder-decoder architecture for semantic… | 32 | 1084 | abandoned |
| neuronika/neuronika Neuronika is a machine learning framework written in pure Rust providing tensors and dynamic neural networks with reverse-mode automatic di… | 32 | 1083 | abandoned |
| openai/InfoGAN Official research code from OpenAI for reproducing results of the InfoGAN paper, an information-theoretic extension to GANs that learns dis… | 10 | 1068 | abandoned |
| google/jaxopt JAXopt is a Python library providing hardware-accelerated, batchable, and differentiable optimizers built on JAX, including implicit differ… | 79 | 1066 | abandoned |
| isl-org/MultiObjectiveOptimization Source code for the NeurIPS 2018 paper 'Multi-Task Learning as Multi-Objective Optimization', implementing the MGDA_UB algorithm for multi-… | 10 | 1066 | abandoned |
| NVIDIA/sentiment-discovery A deprecated PyTorch codebase from NVIDIA for large-scale unsupervised language model pretraining and transfer to sentiment and emotion cla… | 23 | 1065 | abandoned |
| DeepLearningKit/DeepLearningKit DeepLearningKit is an open-source deep learning framework for Apple's iOS, OS X and tvOS, written in Swift and using Metal for GPU-accelera… | 32 | 1059 | abandoned |
| facebookarchive/fbcunn Facebook AI Research's CUDA-accelerated deep learning extension modules for the Torch-7 framework, including fast FFT-based convolutions, m… | 10 | 1058 | abandoned |
| TensorFlow.js tfjs-node was the Node.js native binding for TensorFlow.js, providing accelerated training and inference of ML models in JavaScript server … | 10 | 1055 | abandoned |
| shiyybua/NER A Chinese named entity recognition (NER) system built with TensorFlow, using a bidirectional RNN combined with a CRF layer for sequence lab… | 32 | 1054 | abandoned |
| basveeling/wavenet A Keras implementation of DeepMind's WaveNet, a generative neural network model for raw audio synthesis. It supports training on datasets l… | 32 | 1053 | abandoned |
| facebookresearch/VMZ VMZ is a model zoo from Facebook AI's Computer Vision team providing Caffe2 and PyTorch implementations of video classification models such… | 10 | 1053 | abandoned |
| carpedm20/NTM-tensorflow A TensorFlow implementation of the Neural Turing Machine (NTM) architecture with an LSTM controller and support for multiple read/write hea… | 32 | 1047 | abandoned |
| PRBonn/lidar-bonnetal A deep learning framework for training and deploying semantic segmentation of LiDAR point clouds using range-image representations, develop… | 10 | 1037 | abandoned |
| andrewgordonwilson/bayesgan TensorFlow implementation of the Bayesian GAN from the NIPS 2017 paper by Saatchi and Wilson, which places conditional posteriors over gene… | 32 | 1013 | abandoned |
| graykode/gpt-2-Pytorch A compact PyTorch reimplementation of OpenAI's GPT-2 for text generation, loading Hugging Face's converted pre-trained weights. It provides… | 32 | 1012 | abandoned |
| coreylynch/async-rl A TensorFlow + Keras implementation of asynchronous 1-step Q-learning from the DeepMind 'Asynchronous Methods for Deep Reinforcement Learni… | 32 | 1003 | abandoned |
| karpathy/autoresearch A framework by Andrej Karpathy where AI agents autonomously run LLM research experiments on a single-GPU nanochat training setup. The agent… | 45 | 98112 | active |
| hiyouga/LlamaFactory LlamaFactory is an open‑source Python platform that streamlines fine‑tuning, quantization, and inference of over 100 LLM and VLM models. It… | 81 | 75500 | active |
| Gradio Gradio is an open-source Python library for quickly building web-based demos and UIs for machine learning models, APIs, or arbitrary Python… | 95 | 43762 | stable |
| haotian-liu/LLaVA LLaVA (Large Language and Vision Assistant) is an open-source multimodal large language model framework implementing visual instruction tun… | 21 | 25067 | active |
| DeepFace DeepFace is a lightweight Python library for face recognition and facial attribute analysis, wrapping state-of-the-art models like VGG-Face… | 92 | 23482 | stable |
| KlingAIResearch/LivePortrait LivePortrait is a Python-based portrait animation tool from Kuaishou Technology that synthesizes lifelike videos from a single source image… | 59 | 19260 | active |
| apple/ml-stable-diffusion Apple's official tooling for running Stable Diffusion on Apple Silicon using Core ML. It includes a Python package for converting PyTorch m… | 65 | 17965 | active |
| NVlabs/instant-ngp NVIDIA's implementation of instant neural graphics primitives, training NeRFs, signed distance functions, neural images, and neural volumes… | 48 | 17557 | stable |
| cvat-ai/cvat CVAT (Computer Vision Annotation Tool) is an open-source, self-hosted platform for annotating images, videos, and 3D point clouds to build … | 94 | 16896 | active |
| HumanAIGC/AnimateAnyone Animate Anyone is the official research implementation of a diffusion-based image-to-video synthesis method that animates a static characte… | 43 | 14777 | active |
| CARLA Simulator CARLA is an open-source simulator for autonomous driving research built on Unreal Engine, providing open digital assets like urban layouts,… | 67 | 14465 | active |
| ACE-Step ACE-Step 1.5 is an open-source music generation foundation model combining a language model planner with a Diffusion Transformer to create … | 77 | 13070 | active |
| Orchestra-Research/AI-Research-SKILLs An open-source library of ~98 packaged 'skills' (knowledge modules) that turn AI coding agents like Claude Code, Codex, or Gemini into auto… | 69 | 12991 | active |
| cupy/cupy CuPy is a NumPy/SciPy-compatible array library for GPU-accelerated computing with Python, running on NVIDIA CUDA or AMD ROCm. It acts as a … | 92 | 12362 | stable |
| nerfstudio Nerfstudio is a Python library and CLI toolkit providing a simple, modular API for creating, training, and testing Neural Radiance Fields (… | 34 | 12019 | active |
| InstantID InstantID is a tuning-free, zero-shot identity-preserving image generation method built on diffusion models, generating customized images i… | 27 | 11994 | active |
| FiftyOne FiftyOne is an open-source Python library and GUI app for building high-quality computer vision datasets and models. It enables visualizing… | 98 | 11184 | active |
| zyddnys/manga-image-translator A Python tool that automatically detects, OCRs, translates, inpaints, and re-typesets text in images, primarily for manga and comics. It ru… | 66 | 10477 | active |
| InternVL InternVL is a family of open-source multimodal large language models (vision-language models) that combine vision transformers with LLMs to… | 34 | 10163 | active |
| unit8co/darts Darts is a Python library for user-friendly time series forecasting and anomaly detection. It provides a unified scikit-learn-like fit/pred… | 92 | 9533 | active |
| tile-ai/tilelang TileLang is a Pythonic domain-specific language and compiler (built on TVM) for writing high-performance GPU/CPU/accelerator kernels such a… | 90 | 8724 | active |
| CASIA-LMC-Lab/FastSAM FastSAM is a CNN-based Segment Anything Model trained on only 2% of the SA-1B dataset, achieving comparable segmentation performance to SAM… | 20 | 8421 | active |
| Ucas-HaoranWei/GOT-OCR2.0 Official implementation of GOT-OCR2.0, a unified end-to-end vision-language model for general OCR that converts images of text, documents, … | 23 | 8228 | active |
| NVIDIA/Isaac-GR00T NVIDIA Isaac GR00T N1.7 is an open vision-language-action (VLA) foundation model for generalized humanoid robot skills, taking language and… | 70 | 8116 | stable |
| leejet/stable-diffusion.cpp A pure C/C++ inference engine for diffusion models (Stable Diffusion, FLUX, Wan, Qwen Image, Z-Image, and more) built on ggml in the style … | 92 | 7734 | active |
| zai-org/GLM-OCR GLM-OCR is an open-source 0.9B-parameter multimodal OCR model built on the GLM-V encoder-decoder architecture for complex document understa… | 73 | 7504 | active |
| Background Matting V2 Official PyTorch implementation of the CVPR 2021 paper 'Real-Time High-Resolution Background Matting'. It produces state-of-the-art alpha m… | 23 | 7193 | stable |
| zai-org/GLM-4 Official repository for the GLM-4 series of open-weight, multilingual (primarily Chinese/English) multimodal chat language models from Z.ai… | 64 | 7069 | active |
| VAST-AI-Research/TripoSR TripoSR is an open-source model for fast feedforward 3D object reconstruction from a single image, developed by Tripo AI and Stability AI. … | 61 | 7005 | active |
| linkedin/Liger-Kernel Liger Kernel is a collection of Triton kernels designed to make LLM training faster and more memory-efficient, offering drop-in replacement… | 91 | 6664 | active |
| Yuliang-Liu/MonkeyOCR MonkeyOCR is a lightweight large multimodal model (LMM) for document parsing that uses a Structure-Recognition-Relation triplet paradigm to… | 58 | 6650 | active |
| Trusted-AI/adversarial-robustness-toolbox Adversarial Robustness Toolbox (ART) is a Python library for machine learning security covering evasion, poisoning, extraction, and inferen… | 56 | 6250 | stable |
| ByteDance-Seed/Bagel BAGEL is an open-source unified multimodal foundation model with 7B active parameters (14B total) trained on interleaved multimodal data. I… | 52 | 6191 | active |
| Doubiiu/ToonCrafter ToonCrafter is a generative model that interpolates two cartoon images into a short animation by leveraging pre-trained image-to-video diff… | 26 | 6033 | stable |
| PaddlePaddle/PaddleClas PaddleClas is a Python library and toolkit for image classification, recognition, and retrieval built on the PaddlePaddle deep learning fra… | 65 | 5847 | active |
| areal-project/AReaL AReaL is a large-scale asynchronous reinforcement learning system that bridges foundation model training with agent-based applications, sup… | 86 | 5842 | active |
| AidLux (AidLearning) AidLux (originally AidLearning) is an AIoT development platform that runs a native Ubuntu Linux environment with GUI, deep learning tooling… | 63 | 5812 | active |
| ZenML ZenML is an open-source MLOps framework for building, deploying, and managing production ML pipelines, LLM workflows, and AI agents on any … | 97 | 5601 | active |
| MiniGPT-4 Official code for MiniGPT-4 and MiniGPT-v2, vision-language models that align a frozen visual encoder with a frozen LLM (Vicuna) via a sing… | 30 | 25605 | maintenance |
| xxlong0/Wonder3D Wonder3D is a cross-domain diffusion model that reconstructs high-fidelity textured 3D meshes from a single image in 2-3 minutes. It genera… | 29 | 5437 | active |
| deepseek-ai/DeepSeek-VL2 DeepSeek-VL2 is a series of Mixture-of-Experts vision-language models (Tiny, Small, and 4.5B activated parameters) with inference code and … | 22 | 5376 | active |
| microsoft/SynapseML SynapseML (formerly MMLSpark) is an open-source machine learning library built on Apache Spark that provides simple, composable, distribute… | 86 | 5248 | active |
| yisol/IDM-VTON Official implementation of IDM-VTON, an ECCV 2024 paper that improves diffusion models for high-fidelity virtual try-on, swapping garments … | 27 | 5206 | active |
| hiyouga/EasyR1 EasyR1 is an efficient, scalable reinforcement learning training framework for large language models and vision-language models, built as a… | 62 | 5192 | active |
| UX-Decoder/Segment-Everything-Everywhere-All-At-Once SEEM is the official PyTorch implementation of the NeurIPS 2023 paper 'Segment Everything Everywhere All at Once', a model for universal im… | 21 | 4792 | stable |
| openxla/xla XLA (Accelerated Linear Algebra) is an open-source machine learning compiler that optimizes models from frameworks like PyTorch, TensorFlow… | 77 | 4605 | active |
| TencentARC/InstantMesh InstantMesh is a feed-forward framework for generating 3D meshes from a single image using sparse-view large reconstruction models (LRM/Ins… | 25 | 4545 | active |
| dreamgaussian/dreamgaussian DreamGaussian is the official PyTorch implementation of an ICLR 2024 Oral paper for efficient 3D content creation using generative Gaussian… | 19 | 4360 | active |
| SysCV/sam-hq HQ-SAM (Segment Anything in High Quality) upgrades Meta's Segment Anything Model with a learnable High-Quality Output Token for accurate ze… | 45 | 4262 | active |
| ZhengPeng7/BiRefNet BiRefNet is a PyTorch implementation of the CAAI AIR 2024 paper 'Bilateral Reference for High-Resolution Dichotomous Image Segmentation'. I… | 64 | 4247 | active |
| VectorSpaceLab/OmniGen2 OmniGen2 is an open-source unified multimodal generation model supporting text-to-image generation, instruction-guided image editing, and i… | 48 | 4113 | active |
| StarsfieldAI/R1-V R1-V is an open-source research codebase for training vision-language models with reinforcement learning (RLVR/GRPO), demonstrating strong … | 18 | 4063 | active |
| NVlabs/VILA VILA is a family of open-source vision language models (VLMs) optimized for efficient video and multi-image understanding, spanning edge, d… | 54 | 3865 | active |
| IDEA-Research/Grounded-SAM-2 Grounded SAM 2 is a foundation-model pipeline that combines open-set detectors (Grounding DINO, Grounding DINO 1.5/1.6, Florence-2, DINO-X)… | 34 | 3780 | active |
| AiuniAI/Unique3D Unique3D is the official implementation of a NeurIPS 2024 paper that generates high-quality textured 3D meshes from a single image in about… | 35 | 3589 | active |
| Anttwo/SuGaR SuGaR is the official PyTorch implementation of a CVPR 2024 method that extracts accurate, editable meshes from 3D Gaussian Splatting recon… | 27 | 3525 | active |
| MooreThreads/Moore-AnimateAnyone An open-source reproduction of AnimateAnyone that animates a character from a single reference image using pose sequences from a driving vi… | 27 | 3517 | active |