Home>Programming related>Other source code
Download

Awesome-Pytorch-list｜Awesome Pytorch project

English Version

Contents｜Content

  • Awesome-Pytorch-list｜Awesome Pytorch project
    • English Version
    • Contents｜Content
    • Pytorch & related libraries｜Pytorch & related libraries
      • NLP & Speech Processing｜Natural Language Processing & Speech Processing
      • CV｜Computer Vision
      • Probabilistic/Generated Libraries｜Probability Library and Generation Library
      • Other libraries｜Other libraries
    • Tutorials & books & examples｜Tutorials & Books & Examples
    • Paper implementations｜Paper implementation
    • Talks & conferences｜Reports & conferences
    • Pytorch elsewhere | Pytorch related

Pytorch & related libraries｜Pytorch & related libraries

  1. pytorch: Tensors and Dynamic neural networks in Python with strong GPU acceleration | Python tensor computing and dynamic neural networks using strong GPU acceleration.

NLP & Speech Processing｜Natural Language Processing & Speech Processing:

  1. 2800+ text: Data loading and abstraction for text data and NLP datasets.
  2. 1300+ pytorch-seq2seq: An open source framework for handling seq2seq in Pytorch.
  3. 1000- anuvada: NLP interpretable model.
  4. 1300+ audio: Simple audio I/O.
  5. 1000- loop: A method of speech generation across multiple speakers.
  6. null faireq: The Sequence-to-Sequence python toolkit developed by Facebook.
  7. 1000- speech: End-to-end model implementation of speech to text.
  8. 5100+ OpenNMT-py: Open Source Neural Machine Translation http://opennmt.net.
  9. 2300+ neuralcoref: Use neural networks in spaCy to achieve rapid co-referential digestion.
  10. 1000+ sentiment-discovery: The application of scale-based unsupervised language model in robust sentiment classification.
  11. 2800+ MUSE: A multilingual unsupervised or supervised word embedding library.
  12. 1000- nmtpytorch: Sequence-to-Sequence framework in PyTorch.
  13. 1000- pytorch-wavenet: Quickly generate WaveNet implementation.
  14. 1000- Tacotron-pytorch: Tacotron: End-to-end speech synthesis.
  15. 10300+ AllenNLP: Open source NLP research library, based on PyTorch. http://www.allennlp.org/
  16. 1900+ PyTorch-NLP: A library established to accelerate NLP research, including neural network layers, text processing modules and numerous data sets. pytorchnlp.readthedocs.io
  17. 1000- quick-nlp: Pytorch NLP library based on FastAI.
  18. 4900+ TTS: A deep learning framework for text to speech.
  19. 2800+ LASER: LASER is a library used to calculate and use multilingual statement embeddings.
  20. 1100+ pyannote-audio: Neural building blocks for speaker classification: speech activity detection, speaker change detection, speaker embedding.
  21. 1000- gensen: A universal sentence representation based on large-scale multitasking learning.
  22. 1000- translate: Translation - a PyTorch language library.
  23. 3900+ espnet: End-to-end voice processing toolset. espnet.github.io/espnet
  24. 4500+ pythia: A modular framework for multimodal research on visual and languages ​​originating from FAIR (Facebook AI Research).
  25. 1400+ UnsupervisedMT: Neural unsupervised machine translation based on phrases.
  26. 1300+ jiant: A jiant toolkit for general text understanding models. https://jiant.info
  27. 4300+ BERT-PyTorch: Google AI 2018 BERT's Pytorch implementation, accompanied by simple comments.
  28. 2100+ InferSent: NLI sentence embedding (InferSent) and training code.
  29. 1300+ uis-rnn: Infinite interleaved state recursive neural network (UIS-RNN) algorithm, which can distinguish sound from noisy environments, corresponding to paper Fully Supervised Speaker Diarization. arxiv.org/abs/1810.04719
  30. 10600+ flair: A simple framework for state-of-the-art NLP.
  31. 6200+ pytext: Natural language modeling framework based on PyTorch. fb.me/pytextdocs
  32. 1000- voicefilter: Unofficial implementation of Google AI's VoiceFilter. http://swpark.me/voicefilter
  33. 1000- BERT-NER: Named-Entity-Recognition based on BERT.
  34. 1000- transfer-nlp: NLP library designed for replicable experimental management.
  35. 1000- texar-pytorch: Machine Learning and Text Generation Toolkit. texar.io
  36. 2000+ pytorch-kaldi: pytorch-kaldi is a state-of-the-art dnn/rnn hybrid speech recognition system in development. Its DNN part is implemented by PyTorch, while feature extraction, label calculation and decoding are done by the kaldi toolkit.
  37. 2900+ NeMo: Neural Module: Conversational AI Toolset nvidia.github.io/NeMo
  38. 1000- pytorch-struct: The tested GPU implementation library implements some core structured algorithms in deep learning, such as HMM, Dep Trees, CKY, ...
  39. 1000- espresso: Espresso: A fast set of end-to-end neural speech recognition tools.
  40. 48900+ transformers: huggingface Transformers: The most advanced NLP tools on TensorFlow 2.0 and PyTorch. huggingface.co/transformers
  41. 1500+ reformer-pytorch: PyTorch version of Reformer.
  42. 1000- torch-metrics: Model evaluation metrics in PyTorch.
  43. 2600+ speechbrain: SpeechBrain is an open-source and all-in-one speech toolkit based on PyTorch.
  44. 1000- Backprop: Backprop makes it simple to use, finetune, and deploy state-of-the-art ML models.

CV｜Computer Vision:

  1. 9400+ pytorch vision: TorchVision contains popular data sets, model architectures, and commonly used image transformations in computer vision.
  2. 1000- pt-styletransfer: Neural style transfer as a class in PyTorch.
  3. 1000- OpenFacePytorch: PyTorch module using OpenFace's nn4.small2.v1.t7 model.
  4. 1000- img_classification_pk_pytorch: Quickly compare your image classification model with state-of-the-art models (such as DenseNet, ResNet, ...)
  5. 1400+ SparseConvNet: Submanifold sparse convolutional neural network.
  6. 1000- Convolution_LSTM_pytorch: Multi-layer convolution LSTM (Long Short-term Memory Network) module.
  7. 5000+ face-alignment: 2D and 3D face alignment libraries based on PyTorch. adrianbulat.com
  8. 1500+ pytorch-semantic-segmentation: Semantic segmentation.
  9. 1000- RoIAlign.pytorch: PyTorch version of RoIAlign. Its implementation is based on crop_and_resize and supports forward and backward on CPU and GPU.
  10. 1000- pytorch-cnn-finetune: Pre-training convolutional neural networks with PyTorch fine-tuning.
  11. 1000- detectorch: Detectorch - PyTorch version of detectron framework, currently only detectron's inference and evaluation functions, without training functions.
  12. 4400+ Augmentor: Image enhancement library for machine learning. http://augmentor.readthedocs.io
  13. 1000- s2cnn: Spherical CNNs: PyTorch implementation of spherical convolutional networks. (eg all-round image, global signal)
  14. 2100+ TorchCV: A computer vision deep learning framework based on PyTorch.
  15. 8400+ maskrcnn-benchmark: A fast modular reference implementation of instance segmentation and object detection.
  16. 2200+ image-classification-mobile: Computer vision convolution network training sandbox, containing a collection of training classification models on ImageNet-1K.
  17. 1000- medicaltorch: A medical imaging framework. http://medicaltorch.readthedocs.io
  18. 8400+ accountings: Easy-to-use wrapper for fast image enhancement libraries and other libraries.
  19. 4200+ kornia: Open source microcomputer vision library. https://kornia.org
  20. 1000- text-detector: Detect and translate text.
  21. 2200+ facenet-pytorch: Pre-trained Pytorch face detection and recognition model, ported from davidsandberg/facenet.
  22. 17300+ detectron2: Detectron2 is FAIR's next generation target detection and segmentation research platform.
  23. 1000- vedaseg: A semantic segmentation toolbox based on PyTorch.
  24. 1300+ ClassyVision: A End-to-end PyTorch framework for image and video classification. https://classyvision.ai
  25. 1000- detecto: Use 5 lines of code to build a fully functional computer vision model. https://detecto.readthedocs.io/
  26. 5000+ pytorch3d: PyTorch3d is an efficient, reusable 3D computer vision library for deep learning. https://pytorch3d.org/
  27. 15700+ MMDetection: MMDetection is an open source object detection toolbox, belonging to the OpenMMLab project.
  28. 1000- neural-dream: The PyTorch implementation of the DeepDream algorithm can create dream-like hallucination visual effects.
  29. 1000- FlashTorch: Visualization toolkit for neural networks in PyTorch!
  30. 1000- Lucent: Tensorflow and OpenAI Clarity's Lucid adapted for PyTorch.
  31. 1300+ MMDetection3D: MMDetection3D is OpenMMLab's next-generation platform for general 3D object detection, a part of the OpenMMLab project.
  32. 2100+ MMSegmentation: MMSegmentation is a semantic segmentation toolbox and benchmark, a part of the OpenMMLab project.
  33. 2200+ MMEditing: MMEditing is a image and video editing toolbox, a part of the OpenMMLab project.
  34. 1000+ MMAction2: MMAction2 is OpenMMLab's next generation action understanding toolbox and benchmark, a part of the OpenMMLab project.
  35. 1000+ MMPose: MMPose is a pose estimation toolbox and benchmark, a part of the OpenMMLab project.
  36. 1000+ lightly - Lightly is a computer vision framework for self-supervised learning.

Probabilistic/Generated Libraries｜Probability Library and Generation Library:

  1. 1000- ptstat: Probability programming and statistical inference.
  2. 7000+ pyro: A deep universal probability programming library based on Python and PyTorch. http://pyro.ai
  3. 1000- probtorch: Probabilistic Torch is a library that extends PyTorch's deep generation model.
  4. 1000- paysage: Unsupervised learning and generative model library based on Python/PyTorch.
  5. 1000- pyvarinf: Python package, which promotes the application of Bayesian deep learning methods with variational inference in pytorch.
  6. 1000- pyprob: A library for probabilistic programming and inference compilation based on PyTorch.
  7. 1000- mia: A library that runs member reasoning attacks against machine learning models.
  8. 1000- pro_gan_pytorch: ProGAN package as an extension of PyTorch nn.Module.
  9. 2000+ botorch: Bayesian optimization in PyTorch.

Other libraries｜Other libraries:

  1. 1000- pytorch extras: Additional features of PyTorch.
  2. 1000- functional zoo: Model definition and pre-training weights for PyTorch and Tensorflow.
  3. 1600+ torch-sampling: Pytorch sampling, advanced training, data augmentation and utilities.
  4. 1000- torchcraft-py: TorchCraft Python wrapper, TorchCraft is the bridge connecting Torch and StarCraft.
  5. 1000- aorun: Aorun attempts to implement a Keras-like API using PyTorch as the backend.
  6. 1000- logger: machine learning logger.
  7. 1000- PyTorch-docset: PyTorch offline documentation, combined with Dash, Zeal, Velocity or LovelyDocs.
  8. 1000- convert_torch_to_pytorch: Converts the Torch t7 model to the PyTorch model.
  9. 8000+ pretrained-models.pytorch: PyTorch Pretrained convolutional neural networks: NASNet, ResNeXt, ResNet, InceptionV4, InceptionResnetV2, Xception, DPN, etc. The goal of the project is to help replicate the results of the research paper.
  10. 1000- pytorch_fft: PyTorch wrapper for CUDA FFTs.
  11. 1000- caffe_to_torch_to_pytorch: Caffe model is converted to PyTorch/Torch model, and Torch model is converted to PyTorch model.
  12. 1000- pytorch-extension: PyTorch's CUDA extension example, computes the Hadamard product of two tensors.
  13. 7000+ tensorboard-pytorch: This module saves PyTorch tensors in tensorboard format for inspection. Currently, it supports scalar, image, audio, histogram and other features in tensorboard.
  14. 2400+ gpytorch: GPyTorch is a Gaussian process library implemented with PyTorch. It makes it easy to create scalable, flexible and modular Gaussian process models.
  15. 2500+ spotlight: In-depth recommended model.
  16. 1000- pytorch-cns: Generalized Compressed Network Search based on PyTorch.
  17. 1000- pyinn: CuPy implements fused PyTorch operation.
  18. 1000- inferno: About PyTorch's utility library.
  19. 1000- pytorch-fitmodule: A super simple fitting method for PyTorch module.
  20. 4000+ inferno-sklearn: A scikit-learn-compatible neural network library based on PyTorch encapsulation.
  21. null pytorch-caffe-darknet-convert: Convert between pytorch, caffe prototxt/weights and darknet cfg/weights.
  22. 1000- pytorch2caffe: Convert PyTorch model to Caffe model.
  23. 1000- pytorch-tools: PyTorch tool.
  24. 1900+ sru: Training RNNs is as fast as training CNNs. (arxiv.org/abs/1709.02755)
  25. 1000- torch2coreml: Torch7 -> CoreML, a tool that converts the Torch7 model to Apple CoreML format for running on Apple devices.
  26. 1700+ PyTorch-Encoding: PyTorch Deep Texture Encoding Network (Deep Texture Encoding Network) http://hangzh.com/PyTorch-Encoding
  27. 1000- pytorch-ctc: PyTorch-CTC implements Beam Search decoding by CTC (Connectionist Temporal Classification). The C++ code borrows from TensorFlow and adds flexibility with some improvements.
  28. 1000- candlegp: Gaussian process in Pytorch.
  29. 1000- dpwa: Distributed learning based on pairwise averaging.
  30. 1000- dni-pytorch: PyTorch decoupled neural interface based on synthetic gradients.
  31. 4000+ skorch: A scikit-learn-compatible neural network library based on PyTorch encapsulation.
  32. 3600+ ignite: Ignite is a high-level library that helps you train neural networks in PyTorch.
  33. 1000- Arnold: Arnold - DOOM Game Agent.
  34. 1000- pytorch-mcn: Convert MatConvNet model to PyTorch model.
  35. 3200+ simple-faster-rcnn-pytorch: A simplified implementation of Faster R-CNN, with performance comparable to that of the original paper.
  36. 1000- generative_zoo: generative_zoo provides working implementations of some generative models in PyTorch.
  37. 1800+ pytorchviz: Visualize the running diagram of PyTorch.
  38. 1000- cogitare: Cogitare - A modern, fast and modular deep learning and machine learning framework.
  39. 1000- pydlt: A deep learning toolbox based on PyTorch.
  40. 1000- semi-supervised-pytorch: Implementation of various VAE-based semi-supervised models and generative models.
  41. 1000- pytorch_cluster: PyTorch extension library for optimizing graph cluster algorithm.
  42. 1000- neural-assembly-compiler: PyTorch neural assembly compiler based on adaptive neural compilation.
  43. 1000- caffemodel2pytorch: Convert Caffe model to PyTorch model.
  44. 1000- extension-cpp: C++ extension in PyTorch.
  45. 1000- pytoune: Class Keras framework and utility.
  46. 1000- jetson-reinforcement: NVIDIA Jetson deep reinforcement learning GPU library simulated using PyTorch, OpenAI Gym and Gazebo robots.
  47. 1000- matchbox: Write PyTorch code for a single example and run efficiently in small batches.
  48. 1000- torch-two-sample: PyTorch two-sample test library.
  49. 3100+ pytorch-summary: PyTorch model summary, similar to model.summary() in Keras.
  50. 1000- mpl.pytorch: PyTorch implementation of MaxPoolingLoss.
  51. null scVI-dev: The link is invalid.
  52. 5500+ apex: A PyTorch extension: oriented towards thin-simplified hybrid precision and distributed training.
  53. 3100+ ELF: ELF: A game research platform, recreating AlphaGoZero/AlphaZero.
  54. 1000- Torchlite: Pytorch builds a high-level library on popular machine learning frameworks such as sklearn, Pytorch and Tensorflow.
  55. 1000- joint-vae: PyTorch implementation of JointVAE, a framework for separating continuous and discrete variability factors?
  56. 1000- SLM-Lab: PyTorch modular deep reinforcement learning framework.
  57. 1000- bindsnet: A Python package that simulates pulsed neural networks on CPUs or GPUs with the help of the PyTorch Tensor function.
  58. 1000- pro_gan_pytorch: ProGAN package as an extension of PyTorch nn.Module.
  59. 11500+ pytorch_geometric: PyTorch geometry deep learning extension library.
  60. 1000- torchplus: Implement the + operator on PyTorch modules to return the sequence.
  61. 1000- lagom: lagom: A lightweight PyTorch architecture for rapid prototyping of reinforcement learning algorithms.
  62. 1000- torchbearer: torchbearer: PyTorch model fitting library.
  63. 1000- pytorch-maml-rl: Model-Agnostic Meta-Learning in reinforcement learning.
  64. 1000- NALU: Basic implementation of PyTorch of Neural Arithmetic Logic Units, paper: arxiv.org/pdf/1808.00508.pdf.
  65. 1000- QuCumber: Reconstruction of multi-body wave function in neural networks.
  66. 1000- magnet: Self-established deep learning project. http://magnet-dl.readthedocs.io/
  67. 1000- opencv_transforms: OpenCV implements image segmentation of Torchvision.
  68. 21100+ fastai: fast.ai deep learning library, courses and tutorials.
  69. 1000- pytorch-dense-correspondence: Code for the article "Dense Object Nets: Learning Dense Visual Object Descriptors By and For Robotic Manipulation".
  70. 1000- colorization-pytorch: PyTorch implements interactive deep colorization. richzhang.github.io/ideepcolor
  71. 1000- beauty-net: PyTorch A simple, flexible, and extensible PyTorch template.
  72. 1000- OpenChem: OpenChem: Deep Learning Toolkit for Computational Chemistry and Drug Design Research mariewelt.github.io/OpenChem.
  73. 1000- torchani: PyTorch accurate neural network potential. aiqm.github.io/torchani
  74. 1000- PyTorch-LBFGS: PyTorch implements L-BFGS.
  75. 2400+ gpytorch: An efficient and modular implementation of Gaussian processes in PyTorch.
  76. 1000- hessian: PyTorch version of hessian.
  77. 1000- vel: Speed ​​in deep learning research.
  78. 1000- nonechucks: Dynamically process bad samples in the dataset, using transformations as filters.
  79. 1000+ torchstat: Model analyzer in PyTorch.
  80. 1400+ QNNPACK: Quantitative neural network package—Mobile optimization implementation of quantized neural network operator.
  81. 3600+ torchdiffeq: PyTorch solves the ordinary differential equation (ODE), using a backpropagation algorithm supported by full GPU and O(1) memory complexity.
  82. 1000+ redner: Differentiable Monte Carlo path tracker.
  83. 1000- pixyz: A library for developing deep generative models in a more concise, intuitive and scalable way.
  84. 1000- euclidesdb: A multi-model machine learning feature embedded database. http://euclidesdb.readthedocs.io
  85. 1000- pytorch2keras: Convert PyTorch model to Keras model.
  86. 1000- salad: Domain adaptation and semi-supervised learning toolbox.
  87. 1000- netharn: Parameterized fitting and prediction harnesses for PyTorch.
  88. 7700+ dgl: Python package, based on the existing DL framework, is used to simplify deep learning of graphics. http://dgl.ai.
  89. 1600+ gandissect: A PyTorch-based tool for visualizing and understanding GAN's neurons. gandissect.csail.mit.edu
  90. 1000- delira: A lightweight framework for rapid prototyping and training deep neural networks based on PyTorch and Tensorlow for medical imaging. delira.rtfd.io
  91. 1000- mushroom: Python library for reinforcement learning experiments.
  92. 1000- Xlearn: Transfer learning library.
  93. 1000- geoopt: Riemann adaptive optimization method based on PyTorch optimization.
  94. 1000- vegans: Contains multiple existing GANs.
  95. 4200+ kornia: PyTorch open source microcomputer vision library. https://kornia.org
  96. 1000- AdverTorch: A toolbox for researching against robustness.
  97. 2800+ AdaBound: An optimizer that trains as fast as Adam and as good as SGD.
  98. 1000- fenchel-young-losses: Use Fenchel-Young loss as a loss function for probability classification in PyTorch/TensorFlow/scikit-learn.
  99. 2700+ pytorch-OpCounter: Statistics MACs/FLOPs of PyTorch model.
  100. 1000- Tor10: A general tensor network library designed for quantum simulation based on PyTorch.
  101. 2600+ Catalyst: Advanced utility for PyTorch DL&RL research. Its development focus is repeatability, rapid experimentation and code/idea reuse. Being able to research/develop something new instead of writing another regular training loop.
  102. 1500+ Ax: Adaptive experimental platform.
  103. 1000- pywick: High-level PyTorch neural network training library.
  104. 1000- torchgpipe: PyTorch implements GPipe. torchgpipe.readthedocs.io
  105. 1000+ hub: Pytorch Hub is a pre-trained model library used to improve the repeatability of research.
  106. 14600+ pytorch-lightning: A lightweight PyTorch wrapper for ML researchers. Scaling the model, writing less templates.
  107. 1000- Tor10: A general tensor network library designed for quantum simulation based on pytorch.
  108. 3100+ tensorwatch: Debugging, monitoring and visualization of Python machine learning and data science.
  109. 1000- wavetorch: Numerical solution and counterpropagation of wave equations. arxiv.org/abs/1904.12831
  110. 1000- diffdist: diffdist is a Python library for PyTorch. It extends the default functionality of torch.autograd and adds support for inter-process microcommunication.
  111. 1000- torchprof: The minimum dependency library for layer-by-layer analysis of Pytorch models.
  112. 1000- osqpth: PyTorch microscopic OSQP solver.
  113. 1000- mctorch: A manifold optimization library for deep learning.
  114. 1000- pytorch-hessian-eigenthings: Efficient PyTorch Hessian feature decomposition using Hessian vector product and random power iteration.
  115. 1200+ MinkowskiEngine: Minkowski Engine is an automatic differential method library for generalized sparse convolution and high-dimensional sparse tensors.
  116. 1000- pytorch-cpp-rl: CppRl is a reinforcement learning framework written in the PyTorch C++ front-end.
  117. 1000+ pytorch-toolbelt: PyTorch extension for rapid R&D prototype development and Kaggle code collection.
  118. 1000- argus-tensor-stream: A library for decoding live video streams into CUDA memory. tensorstream.argus-ai.com
  119. 1000- macarico: Learn to search in PyTorch.
  120. 1900+ rlpyt: Reinforcement learning in PyTorch.
  121. 1000- pywarm: A cleaner way to build neural networks for PyTorch. https://blue-season.github.io/pywarm/
  122. 1300+ learn2learn: PyTorch meta-learning framework. http://learn2learn.net
  123. 1000- torchbeast: A PyTorch platform for distributed reinforcement learning.
  124. 1100+ higher: higher is a PyTorch library that allows users to obtain higher-order gradients that span training loops rather than losses in a single training step.
  125. null Torchelie: Torchélie is a collection of a series of tool functions, layers, losses, models, trainers, etc. for PyTorch. https://torchelie.readthedocs.org/
  126. 1000- CrypTen: CrypTen is a privacy-protected machine learning framework written in PyTorch, allowing researchers and developers to train models using encrypted data. CrypTen currently supports Secure Multiparty Computation as its encryption mechanism.
  127. 1000+ cvxpylayers: cvxpylayers is a Python library for constructing microconvex optimization layers in PyTorch.
  128. 1000+ RepDistiller: A benchmark for comparing representation distillation (CRD) and the latest knowledge of distillation methods.
  129. 2500+ kaolin: A PyTorch library designed to accelerate 3D deep learning research.
  130. 1000- PySNN: An efficient spike neural network framework built on PyTorch for GPU acceleration.
  131. 1000- sparktorch: Train and run PyTorch models on Apache Spark.
  132. 3400+ pytorch-metric-learning: The easiest way to use metric learning in your application. Modular, flexible and scalable. Built with PyTorch.
  133. 1000- autonomous-learning-library: A PyTorch library for building deep reinforcement learning agents.
  134. 1000- flambe: An ML framework for accelerating research and its production paths. https://flambe.ai
  135. 1900+ pytorch-optimizer: Collections of modern optimization algorithms for PyTorch, include: AccSGD, AdaBound, AdaMod, DiffGrad, Lamb, RAdam, RAdam, Yogi.
  136. 2200+ PyTorch-VAE: A Collection of Variational Autoencoders (VAE) in PyTorch.
  137. 16700+ ray: A fast and simple framework for building and running distributed applications. Ray is packaged with RLlib, a scalable reformer learning library, and Tune, a scalable hyperparameter tuning library. ray.io
  138. 1000- Pytorch Geometric Temporal: A temporal extension library for PyTorch Geometric.
  139. 1000- Poutyne: A Keras-like framework for PyTorch that handles much of the boilerpulating code needed to train neural networks.
  140. 1000- Pytorch-Toolbox: This is toolbox project for Pytorch. Aiming to make you write Pytorch code more easier, readable and concise.
  141. 1000- Pytorch-contrib: It contains reviewed implementations of ideas from recent machine learning papers.
  142. 6200+ EfficientNet PyTorch: It contains an op-for-op PyTorch reimplementation of EfficientNet, along with pre-trained models and examples.
  143. 1300+ PyTorch/XLA: PyTorch/XLA is a Python package that uses the XLA deep learning compiler to connect the PyTorch deep learning framework and Cloud TPUs.
  144. 1000- webdataset: WebDataset is a PyTorch Dataset (IterableDataset) implementation providing efficient access to datasets stored in POSIX tar archives.
  145. 1000- volksdep: volksdep is an open-source toolbox for deploying and accelerating PyTorch, Onnx and Tensorflow models with TensorRT.
  146. 1700+ PyTorch-StudioGAN: StudioGAN is a Pytorch library providing implementations of representative General Adversarial Networks (GANs) for conditional/unconditional image generation. StudioGAN aims to offer an identical playground for modern GANs so that machine learning researchers can readily compare and analyze a new idea.
  147. null torchdrift: drift detection library
  148. 1600+ accelerate : A simple way to train and use PyTorch models with multi-GPU, TPU, mixed-precision.
  149. 1000- lightning-transformers: Flexible interface for high-performance research using SOTA Transformers leveraging Pytorch Lightning, Transformers, and Hydra.

Tutorials & books & examples｜Tutorials & Books & Examples

  1. 4200+ Practical Pytorch**: This tutorial explains different RNN models.
  2. DeepLearningForNLPInPytorch: IPython Notebook Deep Learning Tutorial, containing emphasis on natural language processing.
  3. 21300+ pytorch-tutorial: A deep learning tutorial for researchers, where most of the models have less than 30 lines of implementation code.
  4. 1000- pytorch-exercises: PyTorch exercise collection.
  5. 5200+ pytorch tutorials: Various PyTorch tutorials.
  6. 16500+ pytorch examples: PyTorch usage examples, application scenarios include vision, text, reinforcement learning, etc.
  7. 1000- pytorch practice: PyTorch example.
  8. 1000- pytorch mini tutorials: PyTorch minimalist tutorial, adapted from Alec Radford's Theano tutorial.
  9. 1000- pytorch text classification: PyTorch implements text classification based on CNN.
  10. 1000- cats vs dogs: Kaggle Contest Dogs vs. Cats Redux: Kernels Edition network fine-tuning example.
  11. 1000- convnet: A complete training example of a deep convolutional network on different data sets (ImageNet, Cifar10, Cifar100, MNIST).
  12. 1000- pytorch-generated-adversarial-networks: A simple adversarial generation network (GAN).
  13. 1000- pytorch containers: Simplified Torch containers in PyTorch.
  14. 1000- T-SNE in pytorch: t-SNE experiment.
  15. 1000- AAE_pytorch: PyTorch version against autoencoder.
  16. 1000- Kind_PyTorch_Tutorial: PyTorch beginner tutorial.
  17. 1000- pytorch-poetry-gen: char-RNN (character-level recurrent neural network) based on PyTorch.
  18. 1000- pytorch-REINFORCE: PyTorch implements discrete and continuous control under OpenAI gym.
  19. 6100+ PyTorch-Tutorial**: Build your own neural network simply and quickly. https://morvanzhou.github.io/tutorials/
  20. 1000- pytorch-intro: Demonstrate how to implement CNNs and RNNs in PyTorch.
  21. 1300+ pytorch-classification: A classification framework on CIFAR-10/100 and ImageNet datasets.
  22. 1000- pytorch_notebooks - hardmaru: Random tutorials written in NumPy and PyTorch.
  23. 1000- pytorch_tutoria-quick: PyTorch introduction and tutorial. Researchers in the fields of computer vision, graphics and machine learning require basic knowledge of neural network theory and commonly used neural network frameworks.
  24. 1000- Pytorch_fine_tuning_Tutorial: A short tutorial on fine-tuning or transfer learning in PyTorch.
  25. 1000- pytorch_exercises: PyTorch exercises.
  26. 1000- traffic-sign-detection: New York University 2018 Computer Vision Fall Course Example.
  27. 1000- mss_pytorch: No need for post-filtering, use loop inference algorithm to achieve singing and speech separation - PyTorch implementation. Demo: js-mim.github.io/mss_pytorch
  28. 2600+ DeepNLP-models-Pytorch PyTorch implementation of various deep NLP models in the cs-224n course. (Stanford Univ: NLP with Deep Learning)
  29. 1000- Mila introduction tutorials: Various tutorials for MILA freshmen. (MILA: Canadian Center for Artificial Intelligence Research)
  30. 1000- pytorch.rl.learning: Use PyTorch to learn reinforcement learning.
  31. 1000- minimum-seq2seq: Minimum Seq2Seq model focusing on neural machine translation.
  32. 1000- tensorly-notebooks: Use Python and TensorLy to implement tensor methods. tensorly.github.io/dev
  33. 1000- pytorch_bits: Related examples of timing prediction.
  34. 1000- skip-thoughts: PyTorch implements the Skip-Thought word vector model.
  35. 1000- video-caption-pytorch: Use PyTorch to add subtitles to videos.
  36. 1000- Capsule-Network-Tutorial: A simple and easy-to-learn Capsule Network tutorial.
  37. 2100+ code-of-learn-deep-learning-with-pytorch: Code in the book "Introduction to Deep Learning: PyTorch". item.jd.com/17915495606.html
  38. 2300+ RL-Adventure: Deep Q Learning tutorial for Pytorch version, simple, easy to learn, and strong code readability. It includes the PyTorch implementation of DQN / DDQN / Prioritized replay/ noisy networks/ distributional values/ Rainbow/ hierarchical RL.
  39. 1000- accelerated_dl_pytorch: Jupyter Day Atlanta II Accelerated deep learning algorithms for meetings, including PyTorch tutorials and conference speeches.
  40. 2500+ RL-Adventure-2: A tutorial on PyTorch0.4 version of the following: actor critic / proximal policy optimization / acer / ddpg / twin duel ddpg / soft actor critic / generalized adversarial imitation learning / hindsight experience replay.
  41. Generative Adversarial Networks (GANs) in 50 lines of code (PyTorch): 50 lines generate adversarial networks.
  42. adversarial-autoencoders-with-pytorch: PyTorch against autoencoders.
  43. transfer learning using pytorch: PyTorch transfer learning.
  44. how-to-implement-a-yolo-object-detector-in-pytorch: How to implement a YOLO (v3) object detector using PyTorch.
  45. pytorch-for-recommenders-101: Build a recommendation system using PyTorch.
  46. 1000- pytorch-for-numpy-users: PyTorch for Numpy users.
  47. PyTorch Tutorial: PyTorch Chinese Tutorial (PyTorch Chinese Network).
  48. 1000- grokking-pytorch: Teach you step by step how to learn PyTorch.
  49. 5200+ PyTorch-Deep-Learning-Minicourse: PyTorch Deep Learning Micro Courses.
  50. 1000- pytorch-custom-dataset-examples: Some examples of custom datasets for PyTorch.
  51. Multiplicative LSTM for sequence-based Recommenders: Multiplicative LSTM for sequence-based recommenders. / Recommended sequence implementation based on LSTM.
  52. 1000- deeplearning.ai-pytorch: PyTorch implementation of Coursera deep learning course (deeplearning.ai) tasks.
  53. 1000- MNIST_Pytorch_python_and_capi: Example: How to train an MNIST network in Python and run it in C++ with PyTorch1.0.
  54. 1000- torch_light: Tutorials and examples, including reinforcement learning, NLP, CV. Neural network models such as Logistic, CNN, RNN, LSTM are implemented by several lines of code, and some advanced examples are implemented by complex models.
  55. 1000- porttrain-gan: Encoding (decoding has not been implemented yet) Art-DCGAN generated portrait oil painting.
  56. 1000- mri-analysis-pytorch: Nuclear magnetic resonance (MRI) analysis was performed using PyTorch and MedicalTorch.
  57. 1000- cifar10-fast: Complete training of the ResNet model on the CIFAR10 dataset within 79 seconds and achieves a test accuracy of 94%. For related content, see blog series.
  58. Intro to Deep Learning with PyTorch: A free course jointly launched by Udacity and Facebook, including an introduction to PyTorch and an interview with Soumith Chintala, one of the authors of PyTorch.
  59. 2900+ pytorch-sentiment-analysis: PyTorch and TorchText semantic analysis tutorial.
  60. 11800+ pytorch-image-models: PyTorch image model, script, and training weights— (SE)ResNet/ResNeXT, DPN, EfficientNet, MobileNet-V3/V2/V1, MNASNet, Single-Path NAS, FBNet, etc.
  61. 1000- CIFAR-ZOO: PyTorch implementation of multiple CNN architectures based on CIFAR.
  62. 3700+ d2l-pytorch: This project attempts to copy the book "Dive into Deep Learning" (www.d2l.ai) and adapt the MXnet code into the PyTorch version.
  63. 1000- thinking-in-tensors-writing-in-pytorch: Tensor thinking, PyTorch practice (Introduction to deep learning).
  64. 1000- NER-BERT-pytorch: A PyTorch solution for naming test questions, using Google AI's pre-trained BERT model.
  65. 1000- pytorch-sync-batchnorm-example: How to use Cross Replica/Synchronized Batchnorm in PyTorch.
  66. 1000- SentimentAnalysis: Sentiment analysis neural network, obtained by fine-tuning BERT training on the Stanford Emotion Tree Library.
  67. 1100+ pytorch-cpp: Created for deep learning researchers, implementing PyTorch tutorial content in C++ (Python tutorial based on pytorch-tutorial).
  68. Deep Learning with PyTorch: Zero to GANs: Deep Learning with PyTorch (video) series of online courses.
  69. Deep Learning with PyTorch: Deep Learning with PyTorch teaches you how to implement deep learning algorithms with Python and PyTorch, the book includes a case study: building an algorithm capable of detecting malignant lung tumors using CT scans.
  70. Serverless Machine Learning in Action with PyTorch and AWS: Serverless Machine Learning in Action is a guide to bringing your experimental PyTorch machine learning code to production using serverless capabilities from major cloud providers like AWS, Azure, or GCP.
  71. 3200+ LabML NN: A collection of PyTorch implementations of neural networks architectures and algorithms with side-by-side notes.

Paper implementations｜Paper implementation

  1. 1000- google_evolution: One of the results networks that implements the article Large-scale evolution of image classes.
  2. 1000- pyscatwave: Fast scattering transformation based on CuPy/PyTorch, Scaling the Scattering Transform: Deep Hybrid Networks
  3. 1000- scalingscattering: This repository contains experiments from the article Scaling The Scattering Transform: Deep Hybrid Networks.
  4. 1000- deep-auto-punctuation: Automatically add punctuation through character-by-character learning.
  5. 1100+ Realtime_Multi-Person_Pose_Estimation: Multi-person human pose estimation based on PyTorch, original code.
  6. 1000- PyTorch-value-iteration-networks: PyTorch implements Value Iteration Networks (NIPS2016 Best Paper Award).
  7. 1000- pytorch_Highway: PyTorch implements highway networks (Highway Networks).
  8. 1000- pytorch_NEG_loss: PyTorch implements negative sampling loss (Negative Sampling Loss).
  9. 1000- pytorch_RVAE: Recursive variational autoencoder for generating sequence data implemented with PyTorch, related papers: Generating Sentences from a Continuous Space, Character-Aware Neural Language Models.
  10. 1000- pytorch_TDNN: Use PyTorch to implement time-delayed neural network (Time Delayed NN).
  11. 1000- eve.pytorch: An implementation of an Eve optimizer, related paper: Imploving Stochastic Gradient Descent with Feedback.
  12. 1000- e2e-model-learning: Task-based end-to-end model in stochastic optimization, https://arxiv.org/abs/1703.04529.
  13. 1000- pix2pix-pytorch: PyTorch实现“基于条件对抗网络的图像到图像翻译”。 论文：Image-to-Image Translation Using Conditional Adversarial Networks。
  14. 4300+ Single Shot MultiBox Detector: 单发多盒探测器，论文：Single Shot MultiBox Detector。
  15. 1000- DiscoGAN: 学习利用生成性对抗网络发现跨域关系。论文：Learning to Discover Cross-Domain Relations with Generative Adversarial Networks。
  16. 1000- official DiscoGAN implementation: 官方实现“学习利用生成性对抗网络发现跨域关系”。 论文：Learning to Discover Cross-Domain Relations with Generative Adversarial Networks。
  17. 1000- pytorch-es: 进化策略。论文：Evolution Strategies as a Scalable Alternative to Reinforcement Learning .
  18. 1000- piwise: 使用PyTorch对VOC2012数据集进行像素切割。
  19. 1000- pytorch-dqn: 深度Q学习网络。
  20. 1000+ neuraltalk2-pytorch: PyTorch图像字幕代码库(在分支“with_finetune”中有可微调CNN)。
  21. 1000- vnet.pytorch: PyTorch实现V-Net：全卷积神经网络在体医学图像分割中的应用。 http://mattmacy.io/vnet.pytorch/
  22. 1400+ pytorch-fcn: PyTorch 实现完全卷积网络。
  23. 1000- WideResNets: PyTorch实现WideResNets。该实现比官方Torch实现花费更少的GPU内存。实现: https://github.com/szagoruyko/wide-residual-networks .
  24. 1000- pytorch_highway_networks: PyTorch实现高速公路网络。
  25. 1000- pytorch-NeuCom: Pytorch实现DeepMind的可微神经计算机论文。
  26. 1000- captionGen: 使用PyTorch为图像生成标注。
  27. 1100+ AnimeGAN: 生成对抗网络的PyTorch简单实现，关注于动漫脸谱绘画。
  28. 1000- Cnn-text classification: PyTorch 实现Kim的基于卷积神经网络的句子分类论文。
  29. 1700+ deepspeech2: 使用Baidu Warp-CTC 实现DeepSpeech2。创造一个基于DeepSpeech2 架构的网络，用CTC 激活函数训练。
  30. 1000- seq2seq: 包含PyTorch中的Seq2Seq模型。
  31. 1000- Asynchronous Advantage Actor-Critic in PyTorch: PyTorch实现A3C(Asynchronous Advantage Actor-Critic)，论文：Asynchronous Methods for Deep Reinforcement Learning。由于PyTorch 可以轻松地在多进程内控制共享内存，我们可以轻易实现A3C这样的异步算法。
  32. 1000- densenet: This is a PyTorch 实现DenseNet-BC 架构，相关论文Densely Connected Convolutional Networks。该实现的CIFAR-10+ 100层错误率为4.77 增长率为12。官方实现和许多第三方库的链接参见liuzhuang13/DenseNet。
  33. 1000- nninit: PyTorch神经网络模块的权值初始化方案，这是nninit 的流行端口。
  34. 1500+ faster rcnn: PyTorch 实现Faster RCNN。该项目主要基于py-faster-rcnn 和TFFRCNN。更多关于R-CNN 的细节请参考论文Faster R-CNN：Towards Real-Time Object Detection with Region Proposal Network。
  35. 1000- doomnet: PyTorch版Doom-net，实现了ViZDoom环境下的RL模型。
  36. 1000- flownet: 通过Dosovitskiy等完成FlowNet的Pytorch实现。
  37. 1000- sqeezenet: 在CIFAR10数据集上用PyTorch实现Squeezenet模型，论文。
  38. 2800+ WassersteinGAN: PyTorch实现WassersteinGAN。
  39. 1000- optnet: 该仓库包含PyTorch源码，重现了论文OptNet: Differentiable Optimization as a Layer in Neural Networks中的实验。
  40. 1000- qp solver: PyTorch的一个快速和可微分的QP求解器。https://locuslab.github.io/qpth/
  41. 1000- Continuous Deep Q-Learning with Model-based Acceleration : 基于模型加速的连续深度Q学习的再实现。
  42. 1000- Learning to learn by gradient descent by gradient descent: PyTorch实现Learning to learn by gradient descent by gradient descent。
  43. 1000- fast-neural-style: PyTorch实现fast-neural-style，论文：Perceptual Losses for Real-Time Style Transfer and Super-Resolution。
  44. 1000- PytorchNeuralStyleTransfer: Pytorch中的神经风格转换。
  45. 1000- Fast Neural Style for Image Style Transform by Pytorch: 使用快速神经风格进行图像风格转换。
  46. 1000- neural style transfer: 通过神经风格算法介绍PyTorch，Neural-Style algorithm。
  47. 1000- VIN_PyTorch_Visdom: PyTorch实现价值迭代网络(VIN):干净、简单、模块化。利用Visdom进行可视化。
  48. 1400+ YOLO2: PyTorch中的YOLOv2。
  49. 1200+ attention-transfer: 通过注意转移改善卷积网络，ICLR2017会议论文。
  50. 1000- SVHNClassifier: PyTorch实现基于深度卷积神经网络的街景图像多位数识别。
  51. 1000- pytorch-deform-conv: PyTorch实现可变形卷积(Deformable Convolution)。
  52. 1000- BEGAN-pytorch: PyTorch实现边界均衡生成对抗网络（BEGAN）: Boundary Equilibrium Generative Adversarial Networks.
  53. 1000- treelstm.pytorch: PyTorch实现树形结构LSTM。
  54. 1000- AGE: 论文代码，原文：对抗生成编码器网络（Adversarial Generator-Encoder Networks）。
  55. 1000- ResNeXt.pytorch: 再现ResNet-V3 (深度神经网络的聚集残差变换)。
  56. 1000- pytorch-rl: 基于PyTorch和Visdom的深度强化学习。
  57. 1000- Deep-Leafsnap: 对比传统的计算机视觉方法，使用深度神经网络的LeafSnap能有效提高测试准确率。
  58. 15400+ pytorch-CycleGAN-and-pix2pix: PyTorch 实现图像风格迁移。
  59. 1000- A3C-PyTorch:PyTorch 实现A3C(Advantage async actor-critic)算法。
  60. 1000- pytorch-value-iteration-networks: PyTorch实现价值迭代网络Value Iteration Networks (NIPS 2016 最佳论文)。
  61. 1000- PyTorch-Style-Transfer: PyTorch实现实时转换多风格生成网络。
  62. 1000- pytorch-deeplab-resnet: PyTorch实现DeepLab resnet v2。
  63. 1100+ pointnet.pytorch: PyTorch实现"PointNet: 基于深度学习的3D点分类和分割模型" https://arxiv.org/abs/1612.00593
  64. 2100+ pytorch-playground: 包含常见的预训练模型和数据集(MNIST, SVHN, CIFAR10, CIFAR100, STL10, AlexNet, VGG16, VGG19, ResNet, Inception, SqueezeNet)**.
  65. 1000- pytorch-dnc: PyTorch/Visdom实现的神经机器翻译(NTM)&可微神经计算机(DNC)。
  66. 1000- pytorch_image_classifier: 使用PyTorch的最小但实用的图像分类器管道，在ResNet18上进行细化，在自己的小型数据集上获得99%的准确率。
  67. 1000- mnist-svhn-transfer: PyTorch实现CycleGAN和SGAN。
  68. null pytorch-yolo2: pytorch-yolo2
  69. 1000- dni: PyTorch实现使用合成梯度的解耦神经接口，论文：Decoupled Neural Interfaces using Synthetic Gradients。
  70. 1200+ wgan-gp: PyTorch实现论文"Improved Training of Wasserstein GANs".
  71. 1000- pytorch-seq2seq-intent-parsing: PyTorch使用seq2seq和注意力模型进行意图分析和空位填充。
  72. 1000- pyTorch_NCE: 复现噪音对比估计算法，论文：Noise-contrastive estimation: A new estimation principle for unnormalized statistical models。
  73. 1000- molencoder: 分子自动编码器。
  74. 1000- GAN-weight-norm: 论文代码，"生成对抗网络中批量和权重归一化的影响"
  75. 1000- lgamma: 实现polygamma、lgamma和beta函数。
  76. 1000- bigBatch: 论文代码，论文：“训练越久，泛化越好：关闭神经网络大批量训练的泛化间隙”。
  77. 1000- rl_a3c_pytorch: 针对Atari 2600 的强化学习，实现了A3C LSTM 。
  78. 1000- pytorch-retraining: PyTorch动物园模型转移学习(torchvision)。
  79. 1000- nmp_qc: 用于计算机视觉的神经消息传递。
  80. 2900+ grad-cam: PyTorch 实现Grad-CAM。
  81. null pytorch-trpo: PyTorch s实现置信域策略优化（Trust Region Policy Optimization (TRPO)）。
  82. 1000- pytorch-explain-black-box: PyTorch通过有意义扰动实现黑箱的可解释性解释，论文。
  83. 1000- vae_vpflows: 凸组合线性IAF与Householder流https://jmtomczak.github.io/deebmed.html 。
  84. 1000- relational-networks: Pytorch实现"用一个简单的神经网络模块来做关系推理"(关系网络)。
  85. 1000- vqa.pytorch: 视觉问答。
  86. 1300+ end-to-end-negotiator: 成交还是不成交？谈判对话的端到端学习。
  87. 1000- odin-pytorch: 神经网络失配实例的原则性检测。
  88. 1000- FreezeOut: 一种通过逐步冻结层加速神经网络训练的简单技术。
  89. 1000- ARAE: 论文代码，"对抗性正则化的自动编码器, ARAE"。
  90. 1000- forward-thinking-pytorch: PyTorch实现"前向思考：一次一层地建立和训练神经网络"。
  91. 1000- context_encoder_pytorch: PyTorch实现上下文编码器(Context Encoders)，可用于图像修复。
  92. 5500+ attention-is-all-you-need-pytorch: PyTorch在"Attention is All You Need"中实现转换模型，https://github.com/thnkim/OpenFacePytorch。
  93. 1000- OpenFacePytorch: 使用OpenFace's nn4.small2.v1.t7 模型的PyTorch模块。
  94. 1000- neural-combinatorial-rl-pytorch: PyTorch 实现"通过强化学习实现神经组合优化"。
  95. null pytorch-nec: PyTorch实现神经情景控制(NEC，Neural Episodic Control)。
  96. 1000- seq2seq.pytorch: 使用PyTorch进行Sequence-to-Sequence学习。
  97. 1000- Pytorch-Sketch-RNN: PyTorch实现“A Neural Representation of Sketch Drawings”。
  98. 1000- pytorch-pruning: PyTorch实现[1611.06440] 用于资源有效推理的剪枝卷积神经网络
  99. 1000- DrQA: PyTorch实现自动阅读维基百科并回答开放领域问题。
  100. 1000- YellowFin_Pytorch: 基于动量梯度下降（momentum SGD）的自动调优优化器，无需手动指定学习速率和动量。
  101. 1000- samplernn-pytorch: PyTorch实现SampleRNN: 一种无条件端到端神经音频生成模型。
  102. 1000- AEGeAN: 基于AE稳定的更深的深度卷积生成对抗网络(DCGAN, Deep Convolution Generative Adversarial Networks)。
  103. 1000- /pytorch-SRResNet: PyTorch实现“基于生成对抗网络的实感单幅图像超分辨率”。
  104. 1000- vsepp: 论文代码，"VSE++:使用难分样本(Hard Negative)改善视觉语义联合嵌入"。
  105. 1000- Pytorch-DPPO: Pytorch实现分布式近端策略优化(Distributed Proximal Policy Optimization)。
  106. 1700+ UNIT: 无监督的图像到图像转换网络，论文。
  107. 1300+ efficient_densenet_pytorch: DenseNets的内存高效实现。
  108. 1000- tsn-pytorch: PyTorch实现时间分割网络(TSN, Temporal Segment Networks)。
  109. 1000- SMASH: SMASH，一种高效地探索神经体系结构的实验技术。
  110. 1000- pytorch-retinanet: RetinaNet。
  111. 1000- biogans: 实现ICCV 2017 论文"利用GANs进行生物图像合成"。
  112. null Semantic Image Synthesis via Adversarial Learning: PyTorch 实现ICCV 2017 论文"基于对抗学习的语义图像合成"。
  113. 1000- fmpytorch: PyTorch在Cython中实现分析机（Factorization Machine）模块。
  114. 1000- ORN: PyTorch 实现CVPR 2017 论文"Oriented Response Networks"。
  115. 1000- pytorch-maml: PyTorch实现MAML（Model-Agnostic Meta-Learning，与模型无关的元学习）。
  116. 2200+ pytorch-generative-model-collections: PyTorch中的各种生成模型集合。
  117. 1000- vqa-winner-cvprw-2017: Pytorch 实现CVPR'17 VQA( Visual Question Answer，视觉问答) 挑战冠军。
  118. 1000- tacotron_pytorch: PyTorch 实现Tacotron 语音合成模型。
  119. 1000- pspnet-pytorch: PyTorch 实现PSPNet 语义分割网络。
  120. 1000- LM-LSTM-CRF: 《Empower Sequence Labeling with Task-Aware Language Model》 http://arxiv.org/abs/1709.04109
  121. 5000+ face-alignment: 使用PyTorch构建2D和3D人脸对齐库。
  122. 1000- DepthNet: PyTorch 在Still Box数据集上训练DepthNet。
  123. 1600+ EDSR-PyTorch: 论文《Enhanced Deep Residual Networks for Single Image Super-Resolution》的PyTorch实现版本。 (CVPRW 2017)
  124. 1000- e2c-pytorch: E2C，Embed to Control 实现。
  125. 2900+ 3D-ResNets-PyTorch: 基于3D残差网络的动作识别。
  126. 1000- bandit-nmt: EMNLP 2017 论文《Reinforcement Learning for Bandit Neural Machine Translation with Simulated Human Feedback》的代码,，改论文在神经编解码模型的基础上实现了A2C算法，并在模拟噪声激励下对组合进行了基准测试。
  127. 2400+ pytorch-a2c-ppo-acktr: PyTorch 实现Advantage Actor Critic (A2C), Proximal Policy Optimization (PPO，近端策略优化) 和可扩展信赖域（Trust Region）方法，这些算法使用Kronecker因子近似（ACKTR）和生成对抗模仿学习（GAIL）实现，可用于深度强化学习。
  128. 1000- zalando-pytorch: Fashion-MNIST数据集上的各种实验。
  129. 1000- sphereface_pytorch: PyTorch实现SphereFace，人脸识别相关，https://arxiv.org/abs/1704.08063 。
  130. 1000- Categorical DQN: PyTorch 版Categorical DQN，该模型来自论文《A Distributional Perspective on Reinforcement Learning》。
  131. 1000- pytorch-ntm: 神经网络图灵机。
  132. null mask_rcnn_pytorch: Mask RCNN in PyTorch.
  133. 1000- graph_convnets_pytorch: PyTorch 实现图卷积神经网络，NIPS'16。
  134. 1700+ pytorch-faster-rcnn: PyTorch实现faster RCNN 检测框架，基于Xinlei Chen 的tf-faster-rcnn，已不再维护。
  135. 1000- torchMoji: A pyTorch implementation of the DeepMoji model: state-of-the-art deep learning model for analyzing sentiment, emotion, sarcasm etc.
  136. 3900+ semantic-segmentation-pytorch: 在MIT ADE20K dataset数据集上实现语义分割/场景解析。
  137. 1200+ pytorch-qrnn: PyTorch implementation of the Quasi-Recurrent Neural Network - up to 16 times faster than NVIDIA's cuDNN LSTM
  138. 1000- pytorch-sgns: Skipgram Negative Sampling in PyTorch.
  139. 1000- SfmLearner-Pytorch : Pytorch version of SfmLearner from Tinghui Zhou et al.
  140. 1000- deformable-convolution-pytorch: PyTorch实现可变形卷积。
  141. 1000- skip-gram-pytorch: A complete pytorch implementation of skipgram model (with subsampling and negative sampling). The embedding result is tested with Spearman's rank correlation.
  142. 1000- stackGAN-v2: Pytorch implementation for reproducing StackGAN_v2 results in the paper StackGAN++: Realistic Image Synthesis with Stacked Generative Adversarial Networks by Han Zhang*, Tao Xu*, Hongsheng Li, Shaoting Zhang, Xiaogang Wang, Xiaolei Huang, Dimitris Metaxas.
  143. 1000- self-critical.pytorch: 非官方，PyTorch实现基于self-critical 序列训练的图像标注。
  144. 3600+ pygcn: 图卷积网络。
  145. 1000- dnc: 可微神经计算机、稀疏存取存储器与稀疏可微神经计算机。
  146. 1000- prog_gans_pytorch_inference: PyTorch inference for "Progressive Growing of GANs" with CelebA snapshot.
  147. 1000- pytorch-capsule: Pytorch implementation of Hinton's Dynamic Routing Between Capsules.
  148. 1000- PyramidNet-PyTorch: A PyTorch implementation for PyramidNets (Deep Pyramidal Residual Networks, arxiv.org/abs/1610.02915)
  149. 1000- radio-transformer-networks: A PyTorch implementation of Radio Transformer Networks from the paper "An Introduction to Deep Learning for the Physical Layer". arxiv.org/abs/1702.00832
  150. 1000- honk: PyTorch reimplementation of Google's TensorFlow CNNs for keyword spotting.
  151. 1000- DeepCORAL: A PyTorch implementation of 'Deep CORAL: Correlation Alignment for Deep Domain Adaptation.', ECCV 2016
  152. 1000- pytorch-pose: PyTorch工具包，用于2D人体姿态估计。
  153. 1000- lang-emerge-parlai: Implementation of EMNLP 2017 Paper "Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog" using PyTorch and ParlAI
  154. 1200+ Rainbow: Rainbow: Combining Improvements in Deep Reinforcement Learning
  155. 1000- pytorch_compact_bilinear_pooling v1: This repository has a pure Python implementation of Compact Bilinear Pooling and Count Sketch for PyTorch.
  156. 1000- CompactBilinearPooling-Pytorch v2: (Yang Gao, et al.) A Pytorch Implementation for Compact Bilinear Pooling.
  157. 1000- FewShotLearning: Pytorch implementation of the paper "Optimization as a Model for Few-Shot Learning"
  158. 1000- meProp: Codes for "meProp: Sparsified Back Propagation for Accelerated Deep Learning with Reduced Overfitting".
  159. 1000- SFD_pytorch: 单镜头尺度不变人脸检测器。
  160. 1000- GradientEpisodicMemory: Continuum Learning with GEM: Gradient Episodic Memory. https://arxiv.org/abs/1706.08840
  161. 1900+ DeblurGAN: Pytorch implementation of the paper DeblurGAN: Blind Motion Deblurring Using Conditional Adversarial Networks.
  162. 4800+ StarGAN: StarGAN: 多领域图像转换GAN 网络，https://arxiv.org/abs/1711.09020 。
  163. 1000- CapsNet-pytorch: PyTorch 实现NIPS 2017 论文“胶囊间的动态路由”。
  164. 1000- CondenseNet: CondenseNet: 面向移动设备的轻量级CNN。
  165. 6700+ deep-image-prior: 基于神经网络的图像修复，无学习过程。
  166. 1100+ deep-head-pose: 使用PyTorch进行深度学习头部姿势估计。
  167. 1000- Random-Erasing: 论文代码，论文："随机擦除数据增强"。
  168. 1000- FaderNetworks: Fader Networks: 通过滑动属性重构图像- NIPS 2017，https://arxiv.org/pdf/1706.00409.pdf 。
  169. 2300+ FlowNet 2.0: FlowNet 2.0: 深度网络中光流估计的演化。
  170. 5300+ pix2pixHD: 利用条件GANs 合成和处理HD 高清图像的PyTorch 实现，https://arxiv.org/pdf/1711.11585.pdf。
  171. 1000- pytorch-smoothgrad: SmoothGrad通过增加噪声来去除噪声。
  172. 1000- RetinaNet: RetinaNe实现。
  173. 6300+ faster-rcnn.pytorch: This project is a faster faster R-CNN implementation, aimed to accelerating the training of faster R-CNN object detection models.
  174. 1000- mixup_pytorch: A PyTorch implementation of the paper Mixup: Beyond Empirical Risk Minimization in PyTorch.
  175. 1100+ inplace_abn: In-Place Activated BatchNorm for Memory-Optimized Training of DNNs
  176. 1000- pytorch-pose-hg-3d: PyTorch implementation for 3D human pose estimation
  177. 1000- nmn-pytorch: Neural Module Network for VQA in Pytorch.
  178. 1000- bytenet: Pytorch implementation of bytenet from "Neural Machine Translation in Linear Time" paper
  179. 1000- bottom-up-attention-vqa: vqa, bottom-up-attention, pytorch
  180. 1000- yolo2-pytorch: The YOLOv2 is one of the most popular one-stage object detector. This project adopts PyTorch as the developing framework to increase productivity, and utilize ONNX to convert models into Caffe 2 to benifit engineering deployment.
  181. 1000- reseg-pytorch: PyTorch 实现ReSeg。 (https://arxiv.org/pdf/1511.07053.pdf)
  182. 1000- binary-stochastic-neurons: Binary Stochastic Neurons in PyTorch.
  183. 1000- pytorch-pose-estimation: PyTorch Implementation of Realtime Multi-Person Pose Estimation project.
  184. 1000- interaction_network_pytorch: Pytorch Implementation of Interaction Networks for Learning about Objects, Relations and Physics.
  185. 1000- NoisyNaturalGradient: Pytorch Implementation of paper "Noisy Natural Gradient as Variational Inference".
  186. 1000- ewc.pytorch: An implementation of Elastic Weight Consolidation (EWC), proposed in James Kirkpatrick et al. Overcoming catastrophic forgetting in neural networks 2016(10.1073/pnas.1611835114).
  187. 1000- pytorch-zssr: PyTorch implementation of 1712.06087 "Zero-Shot" Super-Resolution using Deep Internal Learning
  188. 1000- deep_image_prior: 基于未训练神经网络的图像重建算法实现。算法：Deep Image Prior。
  189. 1000- pytorch-transformer: PyTorch实现论文Attention Is All You Need。
  190. 1000- DeepRL-Grounding: PyTorch实现AAAI-18论文Gated-Attention Architectures for Task-Oriented Language Grounding。
  191. 1000- deep-forecast-pytorch: 使用LSTMs进行风速预测，论文：Deep Forecast: Deep Learning-based Spatio-Temporal Forecasting。
  192. 1000- cat-net: 正则外观变换（Canonical Appearance Transformations）
  193. 1000- minimal_glo: Minimal PyTorch implementation of Generative Latent Optimization from the paper "Optimizing the Latent Space of Generative Networks"
  194. 1000- LearningToCompare-Pytorch: Pytorch Implementation for Paper: Learning to Compare: Relation Network for Few-Shot Learning.
  195. 1400+ poincare-embeddings: PyTorch implementation of the NIPS-17 paper "Poincaré Embeddings for Learning Hierarchical Representations".
  196. null pytorch-trpo(Hessian-vector product version): This is a PyTorch implementation of "Trust Region Policy Optimization (TRPO)" with exact Hessian-vector product instead of finite differences approximation.
  197. 1000- ggnn.pytorch: A PyTorch Implementation of Gated Graph Sequence Neural Networks (GGNN).
  198. 1000- visual-interaction-networks-pytorch: This's an implementation of deepmind Visual Interaction Networks paper using pytorch
  199. 1000- adversarial-patch: PyTorch实现对抗补丁。
  200. 1000- Prototypical-Networks-for-Few-shot-Learning-PyTorch: Implementation of Prototypical Networks for Few Shot Learning (arxiv.org/abs/1703.05175) in Pytorch
  201. 1000- Visual-Feature-Attribution-Using-Wasserstein-GANs-Pytorch: Implementation of Visual Feature Attribution using Wasserstein GANs (arxiv.org/abs/1711.08998) in PyTorch.
  202. 1000- PhotographicImageSynthesiswithCascadedRefinementNetworks-Pytorch: 用级联优化网络生成照片级图像，https://arxiv.org/abs/1707.09405 。
  203. 2400+ ENAS-pytorch: PyTorch实现"基于参数共享的高效神经网络结构搜索"。
  204. 1000- Neural-IMage-Assessment: 神经图片评估，https://arxiv.org/abs/1709.05424 。
  205. 1000- proxprop: 近端回传(Proximal Backpropagation) - 隐式梯度代替显式梯度的神经网络训练算法。
  206. 10500+ FastPhotoStyle: 照片级逼真的图像风格化的一个封闭解。
  207. 1000- Deep-Image-Analogy-PyTorch: 基于PyTorch的深度图像模拟的Python实现。
  208. 2700+ Person-reID_pytorch: 行人再识别Person-reID的PyTorch实现。
  209. 1000- pt-dilate-rnn: 空洞递归神经网络（Dilated RNNs）。
  210. 1000- pytorch-i-revnet: Pytorch实现i-RevNets。
  211. 1000- OrthNet: TensorFlow、PyTorch和Numpy层生成正交多项式。
  212. 1000- DRRN-pytorch: "超分辨率的深递归残差网络(DRRN)", CVPR 2017
  213. 1000- shampoo.pytorch: Shampoo算法实现。
  214. 1000- Neural-IMage-Assessment 2: 神经图片评估，https://arxiv.org/abs/1709.05424 。
  215. 2900+ TCN: Sequence modeling benchmarks and temporal convolutional networks locuslab/TCN
  216. 1000- DCC: This repository contains the source code and data for reproducing results of Deep Continuous Clustering paper.
  217. 1000- packnet: Code for PackNet: Adding Multiple Tasks to a Single Network by Iterative Pruning arxiv.org/abs/1711.05769
  218. 1000- PyTorch-progressive_growing_of_gans: PyTorch implementation of Progressive Growing of GANs for Improved Quality, Stability, and Variation.
  219. 1000- nonauto-nmt: PyTorch Implementation of "Non-Autoregressive Neural Machine Translation"
  220. 9800+ PyTorch-GAN: PyTorch implementations of Generative Adversarial Networks.
  221. 1000- PyTorchWavelets: PyTorch implementation of the wavelet analysis found in Torrence and Compo (1998)
  222. 1000- pytorch-made: MADE (Masked Autoencoder Density Estimation) implementation in PyTorch
  223. 1000- VRNN: Pytorch implementation of the Variational RNN (VRNN), from A Recurrent Latent Variable Model for Sequential Data.
  224. 1000- flow: Pytorch implementation of ICLR 2018 paper Deep Learning for Physical Processes: Integrating Prior Scientific Knowledge.
  225. 1600+ deepvoice3_pytorch: PyTorch实现基于卷积神经网络的语音合成模型。
  226. 1000- psmm: imlementation of the the Pointer Sentinel Mixture Model, as described in the paper by Stephen Merity et al.
  227. 3000+ tacotron2: Tacotron 2 - PyTorch implementation with faster-than-realtime inference.
  228. 1000- AccSGD: Implements pytorch code for the Accelerated SGD algorithm.
  229. 1000- QANet-pytorch: an implementation of QANet with PyTorch (EM/F1 = 70.5/77.2 after 20 epoches for about 20 hours on one 1080Ti card.)
  230. 1000- ConvE: Convolutional 2D Knowledge Graph Embeddings
  231. 1000- Structured-Self-Attention: Implementation for the paper A Structured Self-Attentive Sentence Embedding, which is published in ICLR 2017: arxiv.org/abs/1703.03130 .
  232. 1000- graphsage-simple: Simple reference implementation of GraphSAGE.
  233. 2800+ Detectron.pytorch: A pytorch implementation of Detectron. Both training from scratch and inferring directly from pretrained Detectron weights are available.
  234. 1000- R2Plus1D-PyTorch: PyTorch implementation of the R2Plus1D convolution based ResNet architecture described in the paper "A Closer Look at Spatiotemporal Convolutions for Action Recognition"
  235. 1000- StackNN: A PyTorch implementation of differentiable stacks for use in neural networks.
  236. 1000- translagent: Code for Emergent Translation in Multi-Agent Communication.
  237. 1000- ban-vqa: Bilinear attention networks for visual question answering.
  238. 1200+ pytorch-openai-transformer-lm: This is a PyTorch implementation of the TensorFlow code provided with OpenAI's paper "Improving Language Understanding by Generative Pre-Training" by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
  239. 1000- T2F: 使用深度学习进行Text-to-Face生成。该项目结合了StackGAN和ProGAN，这两个模型可以基于文字描述合成人脸。
  240. 1300+ pytorch - fid: A Port of Fréchet Inception Distance (FID score) to PyTorch
  241. 1000- vae_vpflows:Code in PyTorch for the convex combination linear IAF and the Householder Flow, JM Tomczak & M. Welling jmtomczak.github.io/deebmed.html
  242. 1000- CoordConv-pytorch: Pytorch implementation of CoordConv introduced in 'An intriguing failing of convolutional neural networks and the CoordConv solution' paper. (arxiv.org/pdf/1807.03247.pdf)
  243. 1000- SDPoint: Implementation of "Stochastic Downsampling for Cost-Adjustable Inference and Improved Regularization in Convolutional Networks", published in CVPR 2018.
  244. 1000- SRDenseNet-pytorch: 极深网络，SRDenseNet-pytorch，论文：基于密集跳跃连接的图像超分辨率（ICCV_2017）。
  245. 1000- GAN_stability: Code for paper "Which Training Methods for GANs do actually Converge? (ICML 2018)"
  246. 1000- Mask-RCNN: A PyTorch implementation of the architecture of Mask RCNN, serves as an introduction to working with PyTorch
  247. 1000- pytorch-coviar: Compressed Video Action Recognition
  248. 1000- PNASNet.pytorch: PyTorch implementation of PNASNet-5 on ImageNet.
  249. 1000- NALU-pytorch: Basic pytorch implementation of NAC/NALU from Neural Arithmetic Logic Units arxiv.org/pdf/1808.00508.pdf
  250. 1000- LOLA_DiCE: Pytorch 使用DiCE实现LOLA。
  251. 1000- generative-query-network-pytorch: Generative Query Network (GQN) in PyTorch as described in "Neural Scene Representation and Rendering"
  252. 1000- pytorch_hmax: 在PyTorch中实现HMAX(Hierarchical Model and X)视觉模型。
  253. 1000- FCN-pytorch-easiest: trying to be the most easiest and just get-to-use pytorch implementation of FCN (Fully Convolotional Networks)
  254. 1000- transducer: A Fast Sequence Transducer Implementation with PyTorch Bindings.
  255. 1000- AVO-pytorch: Implementation of Adversarial Variational Optimization in PyTorch.
  256. 1000- HCN-pytorch: A pytorch reimplementation of { Co-occurrence Feature Learning from Skeleton Data for Action Recognition and Detection with Hierarchical Aggregation }.
  257. 1000- binary-wide-resnet: PyTorch implementation of Wide Residual Networks with 1-bit weights by McDonnel (ICLR 2018)
  258. 1000- piggyback: Code for Piggyback: Adapting a Single Network to Multiple Tasks by Learning to Mask Weights arxiv.org/abs/1801.06519
  259. 7700+ vid2vid: Pytorch implementation of our method for high-resolution (eg 2048x1024) photorealistic video-to-video translation.
  260. 1000- poisson-convolution-sum: Implements an infinite sum of poisson-weighted convolutions
  261. 1000- tbd-nets: PyTorch implementation of "Transparency by Design: Closing the Gap Between Performance and Interpretability in Visual Reasoning" arxiv.org/abs/1803.05268
  262. 1000- attn2d: Pervasive Attention: 2D Convolutional Networks for Sequence-to-Sequence Prediction
  263. 7500+ yolov3: YOLOv3: 训练和推断，https://www.ultralytics.com 。
  264. 1000- deep-dream-in-pytorch: Pytorch implementation of the DeepDream computer vision algorithm.
  265. 1000- pytorch-flows: PyTorch implementations of algorithms for density estimation
  266. 1000- quantile-regression-dqn-pytorch: Quantile Regression DQN a Minimal Working Example
  267. 1000- relational-rnn-pytorch: An implementation of DeepMind's Relational Recurrent Neural Networks in PyTorch.
  268. 1000- DEXTR-PyTorch: 深度极端切割，http://www.vision.ee.ethz.ch/~cvlsegmentation/dextr 。
  269. 1000- PyTorch_GBW_LM: PyTorch Language Model for Google Billion Word Dataset.
  270. 1000- Pytorch-NCE: The Noise Contrastive Estimation for softmax output written in Pytorch
  271. 1000- generative-models: Annotated, understandable, and visually interpretable PyTorch implementations of: VAE, BIRVAE, NSGAN, MMGAN, WGAN, WGANGP, LSGAN, DRAGAN, BEGAN, RaGAN, InfoGAN, fGAN, FisherGAN.
  272. 1000- convnet-aig: PyTorch implementation for Convolutional Networks with Adaptive Inference Graphs.
  273. 1000- integrated-gradient-pytorch: This is the pytorch implementation of the paper - Axiomatic Attribution for Deep Networks.
  274. 1000- MalConv-Pytorch: Pytorch implementation of MalConv.
  275. 1000- trellisnet: Trellis Networks for Sequence Modeling
  276. 1000- Learning to Communicate with Deep Multi-Agent Reinforcement Learning: pytorch implementation of Learning to Communicate with Deep Multi-Agent Reinforcement Learning paper.
  277. 1000- pnn.pytorch: PyTorch implementation of CVPR'18 - Perturbative Neural Networks http://xujuefei.com/pnn.html.
  278. 1000- Face_Attention_Network: Pytorch implementation of face attention network as described in Face Attention Network: An Effective Face Detector for the Occluded Faces.
  279. 1800+ waveglow: 基于流的语音合成生成网络。
  280. 1000- deepfloat: This repository contains the SystemVerilog RTL, C++, HLS (Intel FPGA OpenCL to wrap RTL code) and Python needed to reproduce the numerical results in "Rethinking floating point for deep learning"
  281. 1000- EPSR: Pytorch implementation of Analyzing Perception-Distortion Tradeoff using Enhanced Perceptual Super-resolution Network. This work has won the first place in PIRM2018-SR competition (region 1) held as part of the ECCV 2018.
  282. 1000- ClariNet: Pytorch实现ClariNet。
  283. 48900+ pytorch-pretrained-BERT: PyTorch version of Google AI's BERT model with script to load Google's pre-trained models
  284. 1000- torch_waveglow: PyTorch实现WaveGlow: 基于流的语音合成生成网络。
  285. 3000+ 3DDFA: The pytorch improved re-implementation of TPAMI 2017 paper: Face Alignment in Full Pose Range: A 3D Total Solution.
  286. 1600+ loss-landscape: loss-landscape Code for visualizing the loss landscape of neural nets.
  287. 1000- famos:（非）参数图像风格化马赛克的对抗性框架。论文：http://arxiv.org/abs/1811.09236 。
  288. 1000- back2future.pytorch: This is a Pytorch implementation of Janai, J., Güney, F., Ranjan, A., Black, M. and Geiger, A., Unsupervised Learning of Multi-Frame Optical Flow with Occlusions. ECCV 2018.
  289. 1000- FFTNet: Unofficial Implementation of FFTNet vocode paper.
  290. 1000- FaceBoxes.PyTorch: PyTorch实现FaceBoxes。
  291. 2900+ Transformer-XL: Transformer-XL: Attentive Language Models Beyond a Fixed-Length Contexthttps://github.com/kimiyoung/transformer-xl
  292. 1000- associative_compression_networks: Associative Compression Networks for Representation Learning.
  293. 1000- fluidnet_cxx: FluidNet re-written with ATen tensor lib.
  294. 3700+ Deep-Reinforcement-Learning-Algorithms-with-PyTorch: This repository contains PyTorch implementations of deep reinforcement learning algorithms.
  295. 1000- Shufflenet-v2-Pytorch: This is a Pytorch implementation of faceplusplus's ShuffleNet-v2.
  296. 1000- GraphWaveletNeuralNetwork: This is a Pytorch implementation of Graph Wavelet Neural Network. ICLR 2019.
  297. 1000- AttentionWalk: This is a Pytorch implementation of Watch Your Step: Learning Node Embeddings via Graph Attention. NIPS 2018.
  298. 1000- SGCN: This is a Pytorch implementation of Signed Graph Convolutional Network. ICDM 2018.
  299. 1000- SINE: This is a Pytorch implementation of SINE: Scalable Incomplete Network Embedding. ICDM 2018.
  300. 1000- GAM: This is a Pytorch implementation of Graph Classification using Structural Attention. KDD 2018.
  301. 1000- neural-style-pt: PyTorch 实现Justin Johnson 的神经风格算法。论文：A Neural Algorithm of Artistic Style。
  302. 1000- TuckER: TuckER: Tensor Factorization for Knowledge Graph Completion.
  303. 1000- pytorch-prunes: Pruning neural networks: is it time to nip it in the bud?
  304. 1000- SimGNN: SimGNN: 一个快速图形相似度计算的神经网络方法。论文：A Neural Network Approach to Fast Graph Similarity Computation.
  305. 1000- Character CNN: PyTorch implementation of the Character-level Convolutional Networks for Text Classification paper.
  306. 2400+ XLM: PyTorch original implementation of Cross-lingual Language Model Pretraining.
  307. 1000- DiffAI: A provable defense against adversarial examples and library for building compatible PyTorch models.
  308. 1000- APPNP: Combining Neural Networks with Personalized PageRank for Classification on Graphs. ICLR 2019.
  309. 1000- NGCN: A Higher-Order Graph Convolutional Layer. NeurIPS 2018.
  310. 1000- gpt-2-Pytorch: Simple Text-Generator with OpenAI gpt-2 Pytorch Implementation
  311. 1000- Splitter: Splitter: Learning Node Representations that Capture Multiple Social Contexts. (WWW 2019).
  312. 1000+ CapsGNN: 胶囊图神经网络，Capsule Graph Neural Network。
  313. 2300+ BigGAN-PyTorch: PyTorch实现BigGAN（非官方）。
  314. 1000- ppo_pytorch_cpp: 近端策略优化算法的C++ API。
  315. 1000- RandWireNN: 基于随机连接神经网络性能的图像识别。
  316. 1000- Zero-shot Intent CapsNet: GPU-accelerated PyTorch implementation of "Zero-shot User Intent Detection via Capsule Neural Networks".
  317. 1000- SEAL-CI 半监督图分类：层次图视角，Semi-Supervised Graph Classification: A Hierarchical Graph Perspective. (WWW 2019)。
  318. 1000- MixHop: MixHop: Higher-Order Graph Convolutional Architectures via Sparsified Neighborhood Mixing. ICML 2019.
  319. 1000- densebody_pytorch: PyTorch implementation of CloudWalk's recent paper DenseBody.
  320. 1000- voicefilter: Unofficial PyTorch implementation of Google AI's VoiceFilter system http://swpark.me/voicefilter.
  321. 1300+ NVIDIA/semantic-segmentation: PyTorch实现“利用视频传播和标签松弛改进语义分割”。论文：Improving Semantic Segmentation via Video Propagation and Label Relaxation, In CVPR2019.
  322. 1000- ClusterGCN: A PyTorch implementation of "Cluster-GCN: An Efficient Algorithm for Training Deep and Large Graph Convolutional Networks" (KDD 2019).
  323. 1000+ NVlabs/DG-Net: A PyTorch implementation of "Joint Discriminative and Generative Learning for Person Re-identification" (CVPR19 Oral).
  324. 1000- NCRF: 基于神经网络条件随机场(NCRF)的肿瘤转移检测，相关论文：https://openreview.net/forum?id=S1aY66iiM。
  325. 1000- pytorch-sift: PyTorch实现SIFT（尺度不变特征变换匹配算法，Scale Invariant Feature Transform）描述子。
  326. 1000- brain-segmentation-pytorch: 深度学习分割网络U-Net的PyTorch模型实现，用于脑核磁共振中FLAIR异常的分割。
  327. 1000- glow-pytorch: PyTorch 实现"Glow, Generative Flow with Invertible 1x1 Convolutions"。
  328. 1000- EfficientNets-PyTorch: PyTorch实现EfficientNet: 卷积神经网络模型尺度的再思考。
  329. 1000- STEAL: STEAL - 从噪声标注中学习语义边界，https://nv-tlabs.github.io/STEAL/ 。
  330. 1000- EigenDamage-Pytorch: 官方实现ICML'19 论文"特征损伤：克罗内克分解特征基中的结构剪枝"。
  331. 1000- Aspect-level-sentiment: 论文代码和数据集，ACL2018论文："利用文档知识进行体层情感分类"。
  332. 1000- breast_cancer_classifier: 深层神经网络提高放射科医生乳腺癌筛查的效果，https://arxiv.org/abs/1903.08297 。
  333. 1000- DGC-Net: PyTorch实现"DGC-Net: 密集几何对应网络".
  334. 1000- universal-triggers: Universal Adversarial Triggers for Attacking and Analyzing NLP (EMNLP 2019)
  335. 3700+ Deep-Reinforcement-Learning-Algorithms-with-PyTorch: PyTorch implementations of deep reinforcement learning algorithms and environments.
  336. 1000- simple-effective-text-matching-pytorch: A pytorch implementation of the ACL2019 paper "Simple and Effective Text Matching with Richer Alignment Features".
  337. null Adaptive-segmentation-mask-attack (ASMA): A pytorch implementation of the MICCAI2019 paper "Impact of Adversarial Examples on Deep Learning Models for Biomedical Image Segmentation".
  338. 1000- NVIDIA/unsupervised-video-interpolation: A PyTorch Implementation of Unsupervised Video Interpolation Using Cycle Consistency, In ICCV 2019.
  339. 1000- Seg-Uncertainty: Unsupervised Scene Adaptation with Memory Regularization in vivo, In IJCAI 2020.
  340. 5700+ pulse: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models
  341. 1000- distance-encoding: Distance-Encoding - Design Provably More PowerfulGNNs for Structural Representation Learning.
  342. 1000- Pathfinder Discovery Networks: Pathfinder Discovery Networks for Neural Message Passing.
  343. 1000- PyKEEN: A Python library for learning and evaluating knowledge graph embeddings

Talks & conferences｜报告& 会议

  1. PyTorch Conference 2018: 2018年首届PyTorch开发者大会。

Pytorch elsewhere ｜ Pytorch相关

  1. 8300+ the-incredible-pytorch**: 不可思议的Pythorch：一份PyTorch相关的教程、论文、项目、社区等的清单。
  2. 6500+ generative models: 各种生成模型，例如基于Pytorch和Tensorflow的GAN、VAE。 http://wiseodd.github.io
  3. pytorch vs tensorflow: Reddit上的PyTorch和TensorFlow的比较文章。
  4. Pytorch discussion forum: PyTorch论坛。
  5. null pytorch notebook: docker-stack: 类似于Jupyter Notebook Scientific Python Stack
  6. 1000- drawlikebobross: 使用神经网络作画！
  7. 1000- pytorch-tvmisc: 该仓库收集了作者用PyTorch实现的各种玩意儿。
  8. 1000- pytorch-a3c-mujoco: 该项目旨在解决Mujoco中的控制问题，高度基于pytorch-a3c。
  9. PyTorch in 5 Minutes.
  10. 1000- pytorch_chatbot: 用PyTorch实现的聊天机器人。
  11. 1000- malmo-challenge: Malmo协作人工智能挑战-Pig Catcher团队。
  12. 1000- sketchnet: 指导计算机作画。http://www.jtoy.net/projects/sketchnet/
  13. 1200+ Deep-Learning-Boot-Camp: 非盈利社区运营的5天深度学习训练营。 http://deep-ml.com.
  14. 1000- Amazon_Forest_Computer_Vision: 亚马逊森林计算机视觉：使用PyTorch标记卫星图像标记/Keras中的PyTorch技巧。
  15. 2400+ AlphaZero_Gomoku: 用AlphaZero算法玩五子棋。
  16. null pytorch-cv: null.
  17. 2800+ deep-person-reid: Pytorch实现深度学习行人重新识别方法。
  18. 2700+ pytorch-template: PyTorch深度学习模版。
  19. 1000- Deep Learning With Pytorch: 随书代码《Deep Learning With Pytorch TextBook》 PyTorch实用指南：使用PyTorch建立文本和视觉神经网络模型。亚马逊中国电子版
  20. 1000- compare-tensorflow-pytorch: 比较用Tensorflow编写的层和用Pytorch编写的层之间的输出。
  21. 1000- hasktorch: Haskell中的张量与神经网络。
  22. Deep Learning With Pytorch Deep Learning with PyTorch 教你如何用Python和PyTorch实现深度学习算法。
  23. 1000- nimtorch: PyTorch - Python + Nim，PyTorch的Nim前端。
  24. 1000- derplearning: 自动驾驶遥控车代码。
  25. 1000- pytorch-saltnet: Kaggle | TGS Salt Identification Challenge 第9名解决方案。
  26. 1000- pytorch-scripts: 一些脚本，使在Windows上使用PyTorch更加容易。
  27. 1000- pytorch_misc: 为PyTorch讨论板创建的代码片段。
  28. 1000- awesome-pytorch-scholarship: 收集了一系列优秀的PyTorch学术文章、指南、博客、课程和其他资源。
  29. 1000- MentisOculi: PyTorch版raytracer。(raynet?)
  30. 2400+ DoodleMaster: “画出UI！”("Don't code your UI, Draw it !")
  31. 1000- ocaml-torch: ocaml-torch为PyTorch张量库提供一些ocaml绑定。
  32. 1000- extension-script: TorchScript自定义C++/CUDA运算符的示例。
  33. 1000- pytorch-inference: Windows10 平台上Pytorch 1.0在C++ 中的推断。
  34. 1000- pytorch-cpp-inference: 包含使用PyTorch C++ API执行推断的各种示例。
  35. 1100+ tch-rs: PyTorch的Rust绑定。
  36. 1000- TorchSharp: Pytorch引擎的.NET绑定。
  37. 2000+ ML Workspace: 面向机器学习和数据科学的一体化Web IDE。包含Jupyter, VS Code, PyTorch 和许多其他工具或库，这些都集合在一个Docker映像中。
  38. 1100+ PyTorch Style Guide Style guide for PyTorch code. Consistent and good code style helps collaboration and prevents errors!

Feedback: If you have any ideas or you want any other content to be added to this list, feel free to contribute.

Expand
Additional Information
Related Applications
Recommended for You
Related Information All