Deep-Learning-in-Production

In this repository, I will share some useful notes and references about deploying deep learning-based models in production.

Convert PyTorch Models in Production:

• PyTorch Production Level Tutorials [Fantastic]
• The road to 1.0: production ready PyTorch
• PyTorch 1.0 tracing JIT and LibTorch C++ API to integrate PyTorch into NodeJS [Good Article]
• Model Serving in PyTorch
• PyTorch Summer Hackathon [Very Important]
• Deploying PyTorch and Building a REST API using Flask [Important]
• PyTorch model recognizing hotdogs and not-hotdogs deployed on flask
• Serving PyTorch 1.0 Models as a Web Server in C++ [Useful Example]
• PyTorch Internals [Interesting & Useful Article]
• Flask application to support pytorch model prediction
• Serving PyTorch Model on Flask Thread-Safety
• Serving PyTorch Models on AWS Lambda with Caffe2 & ONNX
• Serving PyTorch Models on AWS Lambda with Caffe2 & ONNX (Another Version)
• EuclidesDB - multi-model machine learning feature database with PyTorch
• EuclidesDB - GitHub
• WebDNN: Fastest DNN Execution Framework on Web Browser
• FastAI PyTorch Serverless API (with AWS Lambda)
• FastAI PyTorch in Production (discussion)

Convert PyTorch Models to C++:

• Loading a PyTorch Model in C++ [Fantastic]
• PyTorch C++ API [Bravo]
• An Introduction To Torch (Pytorch) C++ Front-End [Very Good]
• Blogs on using PyTorch C++ API [Good]
• ATen: A TENsor library
• Important Issue about PyTorch-like C++ interface
• PyTorch C++ API Test
• PyTorch via C++ [Useful Notes]
• AUTOGRADPP
• PyTorch C++ Library
• Direct C++ Interface to PyTorch
• A Python module for compiling PyTorch graphs to C

Deploy TensorFlow Models in Production:

• How to deploy Machine Learning models with TensorFlow - Part1
• How to deploy Machine Learning models with TensorFlow - Part2
• How to deploy Machine Learning models with TensorFlow - Part3
• Neural Structured Learning (NSL) in TensorFlow [Great]
• Building Robust Production-Ready Deep Learning Vision Models
• Creating REST API for TensorFlow models
• "How to Deploy a Tensorflow Model in Production" by Siraj Raval on YouTube
• Code for the "How to Deploy a Tensorflow Model in Production" by Siraj Raval on YouTube
• How to deploy an Object Detection Model with TensorFlow serving [Very Good Tutorial]
• Freeze Tensorflow models and serve on web [Very Good Tutorial]
• How to deploy TensorFlow models to production using TF Serving [Good]
• How Zendesk Serves TensorFlow Models in Production
• TensorFlow Serving Example Projects
• Serving Models in Production with TensorFlow Serving [TensorFlow Dev Summit 2017 Video]
• Building TensorFlow as a Standalone Project
• TensorFlow C++ API Example
• TensorFlow.js
• Introducing TensorFlow.js: Machine Learning in Javascript

Convert Keras Models in Production:

• Deep learning in production with Keras, Redis, Flask, and Apache [Rank: 1st & General Usefult Tutorial]
• Deploying a Keras Deep Learning Model as a Web Application in Python [Very Good]
• Deploying a Python Web App on AWS [Very Good]
• Deploying Deep Learning Models Part 1: Preparing the Model
• Deploying your Keras model
• Deploying your Keras model using Keras.JS
• "How to Deploy a Keras Model to Production" by Siraj Raval on Youtube
• Deploy Keras Model with Flask as Web App in 10 Minutes [Good Repository]
• Deploying Keras Deep Learning Models with Flask
• keras2cpp

Deploy MXNet Models in Production:

• Model Server for Apache MXNet
• Running the Model Server
• Exporting Models for Use with MMS
• Single Shot Multi Object Detection Inference Service
• Amazon SageMaker
• How can we serve MXNet models built with gluon api
• MXNet C++ Package
• MXNet C++ Package Examples
• MXNet Image Classification Example of C++
• MXNet C++ Tutorial
• An introduction to the MXNet API [Very Good Tutorial for Learning MXNet]
• GluonCV
• GluonNLP
• Model Quantization for Production-Level Neural Network Inference [Excellent]

Deploy Machine Learning Models with Go:

• Cortex: Deploy machine learning models in production
• Cortex - Main Page
• Why we deploy machine learning models with Go — not Python

Huawei Deep Learning Framework:

• MindSpore - Huawei Deep Learning Framework
• MindSpore - Tutorial

General Deep Learning Compiler Stack:

• TVM Stack

Model Conversion between Deep Learning Frameworks:

• ONNX (Open Neural Network Exchange)
• Tutorials for using ONNX
• MMdnn [Fantastic]
• Convert Full ImageNet Pre-trained Model from MXNet to PyTorch [Fantastic, & Full ImageNet model means the model trained on ~ 14M images]

Some Caffe2 Tutorials:

• Mnist using caffe2
• Caffe2 C++ Tutorials and Examples
• Make Transfer Learning of SqueezeNet on Caffe2
• Build Basic program by using Caffe2 framework in C++

Some Useful Resources for Designing UI (Front-End Development):

• ReactJS vs Angular5 vs Vue.js
• A comparison between Angular and React and their core languages
• A Guide to Becoming a Full-Stack Developer [Very Good Tutorial]
• Roadmap to becoming a web developer in 2018 [Very Good Repository]
• Modern Frontend Developer in 2018
• Roadmap to becoming a React developer in 2018
• 2019 UI and UX Design Trends [Good]
• Streamlit [The fastest way to build custom ML tools]
• Web Developer Monthly
• 23 Best React UI Component Frameworks
• 9 React Styled-Components UI Libraries for 2018
• 35 New Tools for UI Design
• 5 Tools To Speed Up Your App Development [Very Good]
• How to use ReactJS with Webpack 4, Babel 7, and Material Design
• Adobe Typekit [Great fonts, where you need them]
• Build A Real World Beautiful Web APP with Angular 6
• You Don't Know JS
• JavaScript Top 10 Articles
• Web Design with Adobe XD
• INSPINIA Bootstrap Web Theme
• A Learning Tracker for Front-End Developers
• The best front-end hacking cheatsheets — all in one place [Useful & Interesting]
• GUI-fying the Machine Learning Workflow (Machine Flow)
• Electron - Build cross platform desktop apps with JavaScript [Very Good]

Mobile & Embedded Devices Development:

• PyTorch Mobile [Excellent]
• Mobile UI Design Trends In 2018
• ncnn - high-performance neural network inference framework optimized for the mobile platform [Useful]
• Alibaba - MNN
• Awesome Mobile Machine Learning
• EMDL - Embedded and Mobile Deep Learning
• Fritz - machine learning platform for iOS and Android
• TensorFlow Lite
• NVIDIA Jetson Inference [Great]

Back-End Development Part:

• Modern Backend Developer in 2018
• Deploying frontend applications — the fun way [Very Good]
• RabbitMQ [Message Broker Software]
• Celery [Distributed Task Queue]
• Kafka [Distributed Streaming Platform]
• Kubernetes - GitHub
• Deploy Machine Learning Pipeline on Google Kubernetes Engine
• An introduction to Kubernetes for Data Scientists
• Jenkins and Kubernetes with Docker Desktop
• Create Cluster using docker swarm
• deepo - Docker Image for all DL Framewors
• Kubeflow [deployments of ML workflows on Kubernetes]
• kubespray - Deploy a Production Ready Kubernetes Cluster
• KFServing - Kubernetes for Serving ML Models
• Deploying a HuggingFace NLP Model with KFServing [Interesting]
• Machine Learning: serving models with Kubeflow on Ubuntu, Part 1
• CoreWeave Kubernetes Cloud
• MLOps References [DevOps for ML]
• Data Version Control - DVC [Great]
• PySyft - A library for encrypted, privacy preserving deep learning
• LocalStack - A fully functional local AWS cloud stack

GPU Management Libraries:

• GPUtil
• py3nvml [Python 3 binding to the NVIDIA Management Library]
• PyCUDA - GitHub
• PyCUDA
• PyCUDA Tutorial
• setGPU
• Monitor your GPUs [Excellent]
• Grafana - Monitoring and Observability [Excellent]
• Prometheus [Excellent for monitoring solution & extract required metrics]

Speed-up & Scalabale Python Codes:

• Numba - makes Python code fast
• Dask - natively scales Python
• What is Dask
• Ray - running distributed applications
• Neural Network Distiller [Distillation & Quantization of Deep Learning Models in PyTorch]
• PyTorch Pruning Tutorial
• Can you remove 99% of a neural network without losing accuracy? - An introduction to weight pruning
• PocketFlow - An Automatic Model Compression (AutoMC) framework [Great]
• Introducing the Model Optimization Toolkit for TensorFlow
• TensorFlow Model Optimization Toolkit — Post-Training Integer Quantization
• TensorFlow Post-training Quantization
• Dynamic Quantization in PyTorch
• Static Quantization in PyTorch
• NVIDIA DALI - highly optimized data pre-processing in deep learning
• Horovod - Distributed training framework
• ONNX Float32 to Float16
• Speeding Up Deep Learning Inference Using TensorRT
• Speed up Training
• JAX - Composable transformations of Python+NumPy programs
• TensorRTx - popular DL networks with tensorrt
• Speeding up Deep Learning Inference Using TensorFlow, ONNX, and TensorRT
• TensorRT Developer Guide

Other:

• A Guide to Production Level Deep Learning
• Facebook Says Developers Will Love PyTorch 1.0
• Some PyTorch Workflow Changes
• wandb - A tool for visualizing and tracking your machine learning experiments
• PyTorch and Caffe2 repos getting closer together
• PyTorch or TensorFlow?
• Choosing a Deep Learning Framework in 2018: Tensorflow or Pytorch?
• Deep Learning War between PyTorch & TensorFlow
• Embedding Machine Learning Models to Web Apps (Part-1)
• Deploying deep learning models: Part 1 an overview
• Machine Learning in Production
• how you can get a 2–6x speed-up on your data pre-processing with Python
• Making your C library callable from Python
• MIL WebDNN
• Multi-GPU Framework Comparisons [Great]