Table of contents
1. PyTorch learning tutorials and manuals
3. NLP & PyTorch actual combat
5. PyTorch paper recommendation
Six, PyTorch book recommendation
PyTorch official Chinese tutorial site:
The first PyTorch learning website:
Introduction - Dive-into-DL-PyTorch (tangshusen.me)
1. PyTorch learning tutorials and manuals
(1) PyTorch English version official manual: https://pytorch.org/tutorials/. For students who are good at English, I highly recommend this official PyTorch document, which will take you from entry to proficiency step by step. This document introduces in detail from the basics to how to use PyTorch to build a deep neural network, as well as PyTorch syntax and some high-quality cases.
(2) PyTorch Chinese official documentation: https://pytorch-cn.readthedocs.io/zh/latest/. Students who have difficulty reading the above English documents are welcome. We have prepared a relatively official PyTorch Chinese document for you. The document introduces each function in great detail and can be used as a quick reference book for PyTorch.
(3) This is a PyTorch code tutorial that is more partial to the actual combat of algorithms. It has a high star on github, https://github.com/yunjey/pytorch-tutorial. It is recommended that you study the above two basic PyTorch tutorials before reading this document.
(4) Next, I will introduce an open source book: https://github.com/zergtant/pytorch-handbook.
This is an open source book, the goal is to help those who want to use PyTorch for deep learning development and research to get started quickly. However, the content of this document is not complete, and it is still being updated.
(5) Finally, I recommend a simple and easy-to-use PyTorch Chinese document, which is very suitable for novices to learn:
https://github.com/fendouai/pytorch1.0-cn. This document starts from the introduction of what is PyTorch, to the installation of neural network and PyTorch, to the image classifier and data parallel processing, and introduces the knowledge system of PyTorch in great detail, which is suitable for beginners to learn. The official website of the document: http://pytorchchina.com.
2. PyTorch video tutorial
(1) The first one is a PyTorch video tutorial with a very high click-through rate in station B: https://www.bilibili.com/video/av31914351/. Although the video content is only eight episodes, it is very exciting and easy to understand. . It's just that there are no Chinese subtitles, it's time for you guys to practice English...
(2) Another video tutorial by a foreign tycoon has a high hit rate on YouTube, and it is also a pure English video: https://www.youtube.com/watch?v=SKq-pmkekTk, do you feel foreign No matter how complicated the problem is, the teaching video can be very vivid and simple?
(3) Next, I have to give everyone Amway the PyTorch video tutorial of Teacher Mofan, https://morvanzhou.github.io/tutorials/machine-learning/torch/. I believe everyone should be familiar with Mr. Mofan. His series of videos on Python and deep learning have a high click-through rate on Bilibili and YouTube.
(4) The PyTorch series of video courses of the Artificial Intelligence 101 Academy are more detailed and cover a wider range of knowledge points, https://www.bilibili.com/video/av49008640/, interested friends can listen to it.
(5) Finally, I would like to recommend the leading artificial intelligence education platform in China - the PyTorch introductory and practical series online in July: https://www.julyedu.com/course/getDetail/140/. Although the course is a paid course, the course includes PyTorch grammar, deep learning foundation, word vector foundation, NLP and CV project application, actual combat, etc. The combination of theory and actual combat is indeed more detailed than other courses. I recommend it to everyone.
3. NLP & PyTorch actual combat
（1）Pytorch text
(https://github.com/pytorch/text): Torchtext is a very useful library that can help us solve the problem of text preprocessing. This github repository consists of two parts:
torchText.data: Generic data loaders, abstractions and iterators for text (including vocabularies and word vectors)
torchText.datasets: pre-trained loader for common NLP datasets
We only need to install torchtext through pip install torchtext, and then we can start to experience the various conveniences of Torchtext.
(2)Pytorch-Seq2seq
(https://github.com/IBM/pytorch-seq2seq): Seq2seq is a rapidly developing field where new techniques and frameworks are frequently released. This library is a framework for the Seq2seq model implemented in PyTorch. The framework provides modular and extensible components for the training and prediction of the Seq2seq model. This github project is a basic version with the goal of promoting these techniques and applications. development.
（3）BERT NER
(https://github.com/kamalkraj/BERT-NER): BERT is a pre-trained language model proposed by Google in 2018. Since its birth, it has broken a series of NLP tasks, so it has always been very important in the field of nlp. important influence. The github library is the PyTorch version of BERT, which has built-in many powerful pre-training models, which is very convenient and easy to use.
（4）Fairseq
(https://github.com/pytorch/fairseq): Fairseq is a sequence modeling toolkit that allows researchers and developers to train custom models for translation, summarization, language modeling, and other text generation tasks. It also provides Reference implementations of various Seq2seq models. The github repository contains instructions on getting started, training new models, extending Fairseq with new models and tasks, and those who are interested in the model can click the link above to learn.
（5）Quick-nlp
(https://github.com/outcastofmusic/quick-nlp): Quick-nlp is a deep learning Nlp library inspired by the fast.ai library. It follows the same API as Fastai and extends it to allow running NLP models quickly and easily.
（6）OpenNMT-py
(https://github.com/OpenNMT/OpenNMT-py): This is a PyTorch implementation of OpenNMT, an open source neural network machine translation system. It is designed to facilitate research, try out new ideas, and try out new ideas in translation, summarization, image-to-text, morphology, and many other fields. Some companies have demonstrated that the code can be used in real industrial projects, more details about this github can be found in the link above.
4. CV&PyTorch actual combat
（1）pytorch vision
(https://github.com/pytorch/vision): Torchvision is a convenient tool library for image manipulation independent of pytorch. It mainly includes: vision.datasets, vision.models, vision.transforms, vision.utils several packages, installation and use are very simple, interested friends can refer to the above link.
（2）OpenFacePytorch
(https://github.com/thnkim/OpenFacePytorch): This github library is the implementation of OpenFace in Pytorch. The code requires the input image to be aligned and cropped in the same way as the original OpenFace.
（3）TorchCV
(https://github.com/donnyyou/torchcv): TorchCV is a PyTorch-based computer vision deep learning framework that supports most vision task training and deployment. This github library provides source code for most deep learning-based CV problems , what are you waiting for, friends who are interested in the CV direction?
（4）Pytorch-cnn-finetune
(https://github.com/creafz/pytorch-cnn-finetune): This github library uses pytorch to fine-tune the pre-trained convolutional neural network. The supported architectures and models include: ResNet, DenseNet, Inception v3, VGG, SqueezeNet, AlexNet, etc.
（5）Pt-styletransfer
(https://github.com/tymokvo/pt-styletransfer#pt-styletransfer): This github project is neural style transfer in Pytorch. There are several points to note:
StyleTransferNet as a class that can be imported by other scripts;
VGG support (this was before pretrained VGG models were available in PyTorch)
Ability to save intermediate styles and content targets for display
Can be used as a function of image inspection map matrix
Automatic style, content and product image preservation
Matplotlib plot of loss over time and hyperparameter logging to track favorable outcomes
（6）Face-alignment
(https://github.com/1adrianb/face-alignment#face-recognition): Face-alignment is a 2D and 3D face alignment library implemented in pytorch, using the world's most accurate face alignment network to detect from Python Facial landmarks, capable of detecting points in 2D and 3D coordinates. The github library introduces in detail the basic process of using Face-alignment for face alignment. Interested students are welcome to learn.
5. PyTorch paper recommendation
（1）Google_evolution
(https://github.com/neuralix/google_evolution): This paper implements the result network that realizes the large-scale evolution of the image classifier proposed by Esteban Real et al. Before the experiment, we need to install PyTorch, Scikit-learn and download the CIFAR10 dataset
（https://www.cs.toronto.edu/~kriz/cifar.html）。
（2）PyTorch-value-iteration-networks
(https://github.com/onlytailei/Value-Iteration-Networks-PyTorch): This paper is based on the author's original Theano implementation and Abhishek Kumar's Tensoflow implementation, including the implementation of the value iteration network (VIN) in PyTorch. Vin won the Best Paper Award at NIPS 2016.
（3）Pytorch Highway
(https://github.com/kefirski/pytorch_Highway): Highway Netowrks allows information to pass through each layer at high speed without hindrance. It is inspired by the gate mechanism in Long Short Term Memory (LSTM) recurrent networks, allowing information to It passes through many layers to achieve the effect of training a deep neural network, so that the deep neural network no longer has the effect of a shallow neural network. This paper is a Pytorch-based implementation of the Highway network.
（4）Pyscatwave
(https://github.com/edouardoyallon/pyscatwave): Scattering implementation for Cupy/Pythorn. A scatternet is a convolutional network whose filters are pre-defined as wavelets, does not require learning, and can be used for visual tasks such as image classification. Scattering transforms can significantly reduce the spatial resolution of the input (e.g. 224x224 -> 14x14) with a significantly negative pun power loss.
（5）Pytorch_NEG_loss
(https://github.com/kefirski/pytorch_NEG_loss): This paper is a Pytorch implementation of Negative Sampling Loss. Negative Sampling is a method to solve the word2vec model. It abandons the Huffman tree and uses the Negative Sampling (negative sampling) method to solve it. This paper is a study on the loss function of Negative Sampling. Interested partners can Click on the paper link above to study.
（6）Pytorch_TDNN
(https://github.com/kefirski/pytorch_TDNN): This paper is a Pytorch implementation of Time Delayed NN. The paper describes the principle and implementation process of TDNN in detail.
Six, PyTorch book recommendation
Compared with the current situation where Tensorflow-type books have been rotten, there are not so many PyTorch-type books that have been published. The author recommends four PyTorch books that I think are not bad.
"Introduction to Deep Learning: PyTorch", Electronic Industry Press, author: Liao Xingyu. This "Introduction to Deep Learning: PyTorch" is a relatively early one published among all PyTorch books. The author explained the grammar, principles and actual combat of PyTorch in a simple and simple way with his own way of getting started with deep learning. It is suitable for Introductory learning for novices. But the downside is that there are many imprecise and rote places in the book, which require readers to carefully identify.
Recommended index: ★★★
"PyTorch Deep Learning", People's Posts and Telecommunications Press, author: Wang Hailing, Liu Jiangfeng. This book is an English translation book, the original authors are two Indian bosses, in addition to the basic grammar and functions of PyTorch, the book also covers advanced neural network architectures such as ResNET, Inception, DenseNet and their application cases . This book is suitable for readers who have some theoretical basis and practical experience, such as data analysts and data scientists, and it is not recommended as an entry choice for novices.
Recommended index: ★★★
"Introduction and Practice of Deep Learning Framework PyTorch", Electronic Industry Press, author: Chen Yun. This is a PyTorch book released in 2018. It includes two parts: introduction to theory and practical projects. Compared with other books of the same type, the case of this book is very detailed, including: classic projects in Kaggle competitions, GAN to generate animation avatars, AI Filters, RNN poetry writing, image description tasks, etc. The content setting of theory + actual combat is also more suitable for beginners and practitioners of deep learning.
Recommended index: ★★★★
"PyTorch machine learning from entry to actual combat", Machinery Industry Press, author: Xiaobao Online, Sun Lin, etc. This book is also a Pytorch tutorial that combines theory with practice. Compared with the previous introductory + practical tutorial, the feature of this book is that the theoretical part of deep learning is very detailed, and the actual combat projects behind are more comprehensive. Overall, this book is also a good PyTorch introductory book suitable for novices to learn.
Recommended index: ★★★
[Written at the end] This article summarizes various learning resources of PyTorch for everyone, which is very suitable for beginners.
Github address: https://github.com/INTERMT/Awesome-PyTorch-Chinese
PyTorch official Chinese tutorial site:
Welcome to the official Chinese tutorial site of PyTorch:
http://pytorch.panchuang.net/