• Stars
    star
    2,782
  • Rank 15,786 (Top 0.4 %)
  • Language
    Python
  • License
    Other
  • Created over 5 years ago
  • Updated about 1 year ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

An Automatic Model Compression (AutoMC) framework for developing smaller and faster AI applications.

PocketFlow

PocketFlow is an open-source framework for compressing and accelerating deep learning models with minimal human effort. Deep learning is widely used in various areas, such as computer vision, speech recognition, and natural language translation. However, deep learning models are often computational expensive, which limits further applications on mobile devices with limited computational resources.

PocketFlow aims at providing an easy-to-use toolkit for developers to improve the inference efficiency with little or no performance degradation. Developers only needs to specify the desired compression and/or acceleration ratios and then PocketFlow will automatically choose proper hyper-parameters to generate a highly efficient compressed model for deployment.

PocketFlow was originally developed by researchers and engineers working on machine learning team within Tencent AI Lab for the purposes of compacting deep neural networks with industrial applications.

For full documentation, please refer to PocketFlow's GitHub Pages. To start with, you may be interested in the installation guide and the tutorial on how to train a compressed model and deploy it on mobile devices.

For general discussions about PocketFlow development and directions please refer to PocketFlow Google Group. If you need a general help, please direct to Stack Overflow. You can report issues, bug reports, and feature requests on GitHub Issue Page.

News: we have created a QQ group (ID: 827277965) for technical discussions. Welcome to join us! qr_code

Framework

The proposed framework mainly consists of two categories of algorithm components, i.e. learners and hyper-parameter optimizers, as depicted in the figure below. Given an uncompressed original model, the learner module generates a candidate compressed model using some randomly chosen hyper-parameter combination. The candidate model's accuracy and computation efficiency is then evaluated and used by hyper-parameter optimizer module as the feedback signal to determine the next hyper-parameter combination to be explored by the learner module. After a few iterations, the best one of all the candidate models is output as the final compressed model.

Framework Design

Learners

A learner refers to some model compression algorithm augmented with several training techniques as shown in the figure above. Below is a list of model compression algorithms supported in PocketFlow:

Name Description
ChannelPrunedLearner channel pruning with LASSO-based channel selection (He et al., 2017)
DisChnPrunedLearner discrimination-aware channel pruning (Zhuang et al., 2018)
WeightSparseLearner weight sparsification with dynamic pruning schedule (Zhu & Gupta, 2017)
UniformQuantLearner weight quantization with uniform reconstruction levels (Jacob et al., 2018)
UniformQuantTFLearner weight quantization with uniform reconstruction levels and TensorFlow APIs
NonUniformQuantLearner weight quantization with non-uniform reconstruction levels (Han et al., 2016)

All the above model compression algorithms can trained with fast fine-tuning, which is to directly derive a compressed model from the original one by applying either pruning masks or quantization functions. The resulting model can be fine-tuned with a few iterations to recover the accuracy to some extent. Alternatively, the compressed model can be re-trained with the full training data, which leads to higher accuracy but usually takes longer to complete.

To further reduce the compressed model's performance degradation, we adopt network distillation to augment its training process with an extra loss term, using the original uncompressed model's outputs as soft labels. Additionally, multi-GPU distributed training is enabled for all learners to speed-up the time-consuming training process.

Hyper-parameter Optimizers

For model compression algorithms, there are several hyper-parameters that may have a large impact on the final compressed model's performance. It can be quite difficult to manually determine proper values for these hyper-parameters, especially for developers that are not very familiar with algorithm details. Recently, several AutoML systems, e.g. Cloud AutoML from Google, have been developed to train high-quality machine learning models with minimal human effort. Particularly, the AMC algorithm (He et al., 2018) presents promising results for adopting reinforcement learning for automated model compression with channel pruning and fine-grained pruning.

In PocketFlow, we introduce the hyper-parameter optimizer module to iteratively search for the optimal hyper-parameter setting. We provide several implementations of hyper-parameter optimizer, based on models including Gaussian Processes (GP, Mockus, 1975), Tree-structured Parzen Estimator (TPE, Bergstra et al., 2013), and Deterministic Deep Policy Gradients (DDPG, Lillicrap et al., 2016). The hyper-parameter setting is optimized through an iterative process. In each iteration, the hyper-parameter optimizer chooses a combination of hyper-parameter values, and the learner generates a candidate model with fast fast-tuning. The candidate model is evaluated to calculate the reward of the current hyper-parameter setting. After that, the hyper-parameter optimizer updates its model to improve its estimation on the hyper-parameter space. Finally, when the best candidate model (and corresponding hyper-parameter setting) is selected after some iterations, this model can be re-trained with full data to further reduce the performance loss.

Performance

In this section, we present some of our results for applying various model compression methods for ResNet and MobileNet models on the ImageNet classification task, including channel pruning, weight sparsification, and uniform quantization. For complete evaluation results, please refer to here.

Channel Pruning

We adopt the DDPG algorithm as the RL agent to find the optimal layer-wise pruning ratios, and use group fine-tuning to further improve the compressed model's accuracy:

Model FLOPs Uniform RL-based RL-based + Group Fine-tuning
MobileNet-v1 50% 66.5% 67.8% (+1.3%) 67.9% (+1.4%)
MobileNet-v1 40% 66.2% 66.9% (+0.7%) 67.0% (+0.8%)
MobileNet-v1 30% 64.4% 64.5% (+0.1%) 64.8% (+0.4%)
Mobilenet-v1 20% 61.4% 61.4% (+0.0%) 62.2% (+0.8%)

Weight Sparsification

Comparing with the original algorithm (Zhu & Gupta, 2017) which uses the same sparsity for all layers, we incorporate the DDPG algorithm to iteratively search for the optimal sparsity of each layer, which leads to the increased accuracy:

Model Sparsity (Zhu & Gupta, 2017) RL-based
MobileNet-v1 50% 69.5% 70.5% (+1.0%)
MobileNet-v1 75% 67.7% 68.5% (+0.8%)
MobileNet-v1 90% 61.8% 63.4% (+1.6%)
MobileNet-v1 95% 53.6% 56.8% (+3.2%)

Uniform Quantization

We show that models with 32-bit floating-point number weights can be safely quantized into their 8-bit counterpart without accuracy loss (sometimes even better!). The resulting model can be deployed on mobile devices for faster inference (Device: XiaoMi 8 with a Snapdragon 845 CPU):

Model Acc. (32-bit) Acc. (8-bit) Time (32-bit) Time (8-bit)
MobileNet-v1 70.89% 71.29% (+0.40%) 124.53 56.12 (2.22x)
MobileNet-v2 71.84% 72.26% (+0.42%) 120.59 49.04 (2.46x)
  • All the reported time are in milliseconds.

Citation

Please cite PocketFlow in your publications if it helps your research:

@incollection{wu2018pocketflow,
  author = {Jiaxiang Wu and Yao Zhang and Haoli Bai and Huasong Zhong and Jinlong Hou and Wei Liu and Junzhou Huang},
  title = {PocketFlow: An Automated Framework for Compressing and Accelerating Deep Neural Networks},
  booktitle = {Advances in Neural Information Processing Systems (NIPS), Workshop on Compact Deep Neural Networks with Industrial Applications},
  year = {2018},
}

Reference

  • [Bergstra et al., 2013] J. Bergstra, D. Yamins, and D. D. Cox. Making a Science of Model Search: Hyperparameter Optimization in Hundreds of Dimensions for Vision Architectures. In International Conference on Machine Learning (ICML), pages 115-123, Jun 2013.
  • [Han et al., 2016] Song Han, Huizi Mao, and William J. Dally. Deep Compression: Compressing Deep Neural Network with Pruning, Trained Quantization and Huffman Coding. In International Conference on Learning Representations (ICLR), 2016.
  • [He et al., 2017] Yihui He, Xiangyu Zhang, and Jian Sun. Channel Pruning for Accelerating Very Deep Neural Networks. In IEEE International Conference on Computer Vision (ICCV), pages 1389-1397, 2017.
  • [He et al., 2018] Yihui He, Ji Lin, Zhijian Liu, Hanrui Wang, Li-Jia Li, and Song Han. AMC: AutoML for Model Compression and Acceleration on Mobile Devices. In European Conference on Computer Vision (ECCV), pages 784-800, 2018.
  • [Jacob et al., 2018] Benoit Jacob, Skirmantas Kligys, Bo Chen, Menglong Zhu, Matthew Tang, Andrew Howard, Hartwig Adam, and Dmitry Kalenichenko. Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 2704-2713, 2018.
  • [Lillicrap et al., 2016] Timothy P. Lillicrap, Jonathan J. Hunt, Alexander Pritzel, Nicolas Heess, Tom Erez, Yuval Tassa, David Silver, and Daan Wierstra. Continuous Control with Deep Reinforcement Learning. In International Conference on Learning Representations (ICLR), 2016.
  • [Mockus, 1975] J. Mockus. On Bayesian Methods for Seeking the Extremum. In Optimization Techniques IFIP Technical Conference, pages 400-404, 1975.
  • [Zhu & Gupta, 2017] Michael Zhu and Suyog Gupta. To Prune, or Not to Prune: Exploring the Efficacy of Pruning for Model Compression. CoRR, abs/1710.01878, 2017.
  • [Zhuang et al., 2018] Zhuangwei Zhuang, Mingkui Tan, Bohan Zhuang, Jing Liu, Jiezhang Cao, Qingyao Wu, Junzhou Huang, and Jinhui Zhu. Discrimination-aware Channel Pruning for Deep Neural Networks. In Annual Conference on Neural Information Processing Systems (NIPS), 2018.

Contributing

If you are interested in contributing, check out the CONTRIBUTING.md, also join our Tencent OpenSource Plan.

More Repositories

1

weui

A UI library by WeChat official design team, includes the most useful widgets/modules in mobile web applications.
Less
27,053
star
2

wepy

小程序组件化开发框架
JavaScript
22,396
star
3

ncnn

ncnn is a high-performance neural network inference framework optimized for the mobile platform
C++
19,310
star
4

mars

Mars is a cross-platform network component developed by WeChat.
C++
17,137
star
5

tinker

Tinker is a hot-fix solution library for Android, it supports dex, library and resources update without reinstall apk.
Java
17,056
star
6

MMKV

An efficient, small mobile key-value storage framework developed by WeChat. Works on Android, iOS, macOS, Windows, and POSIX.
C++
16,913
star
7

APIJSON

🏆 零代码、全功能、强安全 ORM 库 🚀 后端接口和文档零代码,前端(客户端) 定制返回 JSON 的数据和结构。 🏆 A JSON Transmission Protocol and an ORM Library 🚀 provides APIs and Docs without writing any code.
Java
16,681
star
8

vConsole

A lightweight, extendable front-end developer tool for mobile web page.
TypeScript
16,379
star
9

weui-wxss

A UI library by WeChat official design team, includes the most useful widgets/modules.
Less
14,966
star
10

QMUI_Android

提高 Android UI 开发效率的 UI 库
Java
14,336
star
11

rapidjson

A fast JSON parser/generator for C++ with both SAX/DOM style API
C++
13,803
star
12

secguide

面向开发人员梳理的代码安全指南
13,093
star
13

omi

Web Components Framework - Web组件框架
TypeScript
12,926
star
14

VasSonic

VasSonic is a lightweight and high-performance Hybrid framework developed by tencent VAS team, which is intended to speed up the first screen of websites working on Android and iOS platform.
Java
11,779
star
15

matrix

Matrix is a plugin style, non-invasive APM system developed by WeChat.
Java
11,417
star
16

wcdb

WCDB is a cross-platform database framework developed by WeChat.
C
10,509
star
17

xLua

xLua is a lua programming solution for C# ( Unity, .Net, Mono) , it supports android, ios, windows, linux, osx, etc.
C
9,133
star
18

libco

libco is a coroutine library which is widely used in wechat back-end service. It has been running on tens of thousands of machines since 2013.
C++
7,998
star
19

Hippy

Hippy is designed to easily build cross-platform dynamic apps. 👏
C++
7,840
star
20

Shadow

零反射全动态Android插件框架
Java
7,316
star
21

QMUI_iOS

QMUI iOS——致力于提高项目 UI 开发效率的解决方案
Objective-C
7,030
star
22

MLeaksFinder

Find memory leaks in your iOS app at develop time.
Objective-C
5,399
star
23

lemon-cleaner

腾讯柠檬清理是针对macOS系统专属制定的清理工具。主要功能包括重复文件和相似照片的识别、软件的定制化垃圾扫描、可视化的全盘空间分析、内存释放、浏览器隐私清理以及设备实时状态的监控等。重点聚焦清理功能,对上百款软件提供定制化的清理方案,提供专业的清理建议,帮助用户轻松完成一键式清理。
Objective-C
5,188
star
24

kbone

一个致力于微信小程序和 Web 端同构的解决方案
JavaScript
4,742
star
25

libpag

The official rendering library for PAG (Portable Animated Graphics) files that renders After Effects animations natively across multiple platforms.
C++
4,729
star
26

puerts

PUER(普洱) Typescript. Let's write your game in UE or Unity with TypeScript.
C++
4,661
star
27

GT

GT (Great Tit) is a portable debugging tool for bug hunting and performance tuning on smartphones anytime and anywhere just as listening music with Walkman. GT can act as the Integrated Debug Environment by directly running on smartphones.
Java
4,385
star
28

TNN

TNN: developed by Tencent Youtu Lab and Guangying Lab, a uniform deep learning inference framework for mobile、desktop and server. TNN is distinguished by several outstanding features, including its cross-platform capability, high performance, model compression and code pruning. Based on ncnn and Rapidnet, TNN further strengthens the support and performance optimization for mobile devices, and also draws on the advantages of good extensibility and high performance from existed open source efforts. TNN has been deployed in multiple Apps from Tencent, such as Mobile QQ, Weishi, Pitu, etc. Contributions are welcome to work in collaborative with us and make TNN a better framework.
C++
4,297
star
29

westore

小程序项目分层架构
JavaScript
4,216
star
30

tmagic-editor

TypeScript
4,037
star
31

wujie

极致的微前端框架
TypeScript
3,801
star
32

vap

VAP是企鹅电竞开发,用于播放特效动画的实现方案。具有高压缩率、硬件解码等优点。同时支持 iOS,Android,Web 平台。
Objective-C
3,794
star
33

phxpaxos

The Paxos library implemented in C++ that has been used in the WeChat production environment.
C++
3,301
star
34

WeFlow

A web developer workflow tool by WeChat team based on tmt-workflow, with cross-platform supported and environment ready.
JavaScript
3,224
star
35

cherry-markdown

✨ A Markdown Editor
JavaScript
3,195
star
36

weui.js

A lightweight javascript library for WeUI.
JavaScript
3,157
star
37

spring-cloud-tencent

Spring Cloud Tencent is a Spring Cloud based Service Governance Framework provided by Tencent.
Java
3,116
star
38

tencent-ml-images

Largest multi-label image database; ResNet-101 model; 80.73% top-1 acc on ImageNet
Python
3,046
star
39

tdesign

Enterprise Design System
Vue
3,010
star
40

VasDolly

Android V1 and V2 Signature Channel Package Plugin
Java
2,999
star
41

FaceDetection-DSFD

腾讯优图高精度双分支人脸检测器
Python
2,863
star
42

PhoenixGo

Go AI program which implements the AlphaGo Zero paper
C++
2,863
star
43

Tendis

Tendis is a high-performance distributed storage system fully compatible with the Redis protocol.
C++
2,837
star
44

behaviac

behaviac is a framework of the game AI development, and it also can be used as a rapid game prototype design tool. behaviac supports the behavior tree, finite state machine and hierarchical task network(BT, FSM, HTN)
C#
2,784
star
45

MSEC

Mass Service Engine in Cluster(MSEC) is opened source by QQ team from Tencent. It is a backend DEV &OPS engine, including RPC,name finding,load balance,monitoring,release and capacity management.
Java
2,745
star
46

phxsql

A high availability MySQL cluster that guarantees data consistency between a master and slaves.
C++
2,463
star
47

OOMDetector

OOMDetector is a memory monitoring component for iOS which provides you with OOM monitoring, memory allocation monitoring, memory leak detection and other functions.
Objective-C++
2,298
star
48

tsf

coroutine and Swoole based php server framework in tencent
PHP
2,179
star
49

tmt-workflow

A web developer workflow used by WeChat team based on Gulp, with cross-platform supported and solutions prepared.
CSS
2,175
star
50

Hardcoder

Hardcoder is a solution which allows Android APP and Android System to communicate with each other directly, solving the problem that Android APP could only use system standard API rather than the hardware resource of system.
C++
2,145
star
51

LKImageKit

A high-performance image framework, including a series of capabilities such as image views, image downloader, memory caches, disk caches, image decoders and image processors.
Objective-C
2,079
star
52

UnLua

A feature-rich, easy-learning and highly optimized Lua scripting plugin for UE.
C++
2,053
star
53

TubeMQ

TubeMQ has been donated to the Apache Software Foundation and renamed to InLong, please visit the new Apache repository: https://github.com/apache/incubator-inlong
2,027
star
54

ObjectDetection-OneStageDet

单阶段通用目标检测器
Python
1,962
star
55

cloudbase-framework

腾讯云开发云原生一体化部署工具 🚀 CloudBase Framework:一键部署,不限框架语言,云端一体化开发,基于Serverless 架构。A front-end and back-end integrated deployment tool. One-click deploy to serverless architecture. https://docs.cloudbase.net/framework/index
JavaScript
1,936
star
56

InjectFix

InjectFix is a hot-fix solution library for Unity
C#
1,933
star
57

TscanCode

A static code analyzer for C++, C#, Lua
C++
1,916
star
58

phxrpc

A simple C++ based RPC framework.
C++
1,905
star
59

soter

A secure and quick biometric authentication standard and platform in Android held by Tencent.
Java
1,902
star
60

phxqueue

A high-availability, high-throughput and highly reliable distributed queue based on the Paxos algorithm.
C++
1,891
star
61

plato

腾讯高性能分布式图计算框架Plato
C++
1,889
star
62

GameAISDK

基于图像的游戏AI自动化框架
C++
1,861
star
63

MedicalNet

Many studies have shown that the performance on deep learning is significantly affected by volume of training data. The MedicalNet project provides a series of 3D-ResNet pre-trained models and relative code.
Python
1,837
star
64

TSW

Tencent Server Web
TypeScript
1,802
star
65

NeuralNLP-NeuralClassifier

An Open-source Neural Hierarchical Multi-label Text Classification Toolkit
Python
1,781
star
66

QMUI_Web

An efficient front-end framework for developers building UI on the web.
JavaScript
1,719
star
67

Biny

Biny is a tiny, high-performance PHP framework for web applications
PHP
1,690
star
68

sluaunreal

lua dev plugin for unreal engine 4 or 5
C++
1,687
star
69

paxosstore

PaxosStore has been deployed in WeChat production for more than two years, providing storage services for the core businesses of WeChat backend. Now PaxosStore is running on thousands of machines, and is able to afford billions of peak TPS.
C++
1,658
star
70

Metis

Metis is a learnware platform in the field of AIOps.
Python
1,644
star
71

CodeAnalysis

Static Code Analysis - 静态代码分析
Python
1,585
star
72

TurboTransformers

a fast and user-friendly runtime for transformer inference (Bert, Albert, GPT2, Decoders, etc) on CPU and GPU.
C++
1,442
star
73

TencentOS-kernel

腾讯针对云的场景研发的服务器操作系统
1,401
star
74

nohost

基于 Whistle 实现的多账号多环境远程配置及抓包调试平台
JavaScript
1,392
star
75

TBase

TBase is an enterprise-level distributed HTAP database. Through a single database cluster to provide users with highly consistent distributed database services and high-performance data warehouse services, a set of integrated enterprise-level solutions is formed.
C
1,372
star
76

WeDemo

WeDemo为微信团队开源项目,用于帮助微信开发者完成微信登录、微信分享等功能的接入和开发。开发者可参考源代码完成开发,也可以直接将代码应用到自己的App开发中,安全、便捷地在App中实现微信分享、微信登录功能。
Objective-C
1,365
star
77

feflow

🚀 A command line tool aims to improve front-end engineer workflow and standard, powered by TypeScript.
TypeScript
1,354
star
78

GAutomator

Automation for mobile games
Objective-C
1,318
star
79

tdesign-vue-next

A Vue3.x UI components lib for TDesign.
TypeScript
1,316
star
80

flare

Flare是广泛投产于腾讯广告后台的现代化C++开发框架,包含了基础库、RPC、各种客户端等。主要特点为易用性强、长尾延迟低。
C++
1,264
star
81

TFace

A trusty face analysis research platform developed by Tencent Youtu Lab
Python
1,236
star
82

LuaPanda

lua debug and code tools for VS Code
Lua
1,219
star
83

FeatherCNN

FeatherCNN is a high performance inference engine for convolutional neural networks.
C++
1,209
star
84

tdesign-miniprogram

A Wechat MiniProgram UI components lib for TDesign.
HTML
1,084
star
85

tgfx

A lightweight 2D graphics library for rendering texts, geometries, and images with high-performance APIs that work across various platforms.
C++
1,011
star
86

TencentPretrain

Tencent Pre-training framework in PyTorch & Pre-trained Model Zoo
Python
985
star
87

RapidView

RapidView is an android ui and lightapp development framework
Java
977
star
88

FAutoTest

A UI automated testing framework for H5 and applets
Python
930
star
89

TencentKona-8

Tencent Kona is a no-cost, production-ready distribution of the Open Java Development Kit (OpenJDK), Long-term support(LTS) with quarterly updates. Tencent Kona serves as the default JDK internally at Tencent Cloud for cloud computing and other Java applications.
Java
909
star
90

tquic

A high-performance, lightweight, and cross-platform QUIC library
Rust
900
star
91

hel

A module federation SDK which is unrelated to tool chain for module consumer. 工具链无关的运行时模块联邦sdk.
JavaScript
888
star
92

tdesign-vue

A Vue.js UI components lib for TDesign.
TypeScript
872
star
93

Pebble

Pebble分布式开发框架
C++
861
star
94

mxflutter

使用 TypeScript/JavaScript 来开发 Flutter 应用的框架。
Dart
834
star
95

Face2FaceTranslator

面对面翻译小程序是微信团队针对面对面沟通的场景开发的流式语音翻译小程序,通过微信同声传译插件提供了语音识别,文本翻译等功能。
JavaScript
822
star
96

tdesign-react

A React UI components lib for TDesign.
TypeScript
787
star
97

LightDiffusionFlow

This extension is developed for AUTOMATIC1111's Stable Diffusion web UI that provides import/export options for parameters.
JavaScript
764
star
98

Real-SR

Real-World Super-Resolution via Kernel Estimation and Noise Injection
Python
753
star
99

DCache

A distributed in-memory NOSQL system based on TARS framework, support LRU algorithm and data persists on back-end database. Users can easily deploy, publish, and scale services on the web interface.
C++
746
star
100

PatrickStar

PatrickStar enables Larger, Faster, Greener Pretrained Models for NLP and democratizes AI for everyone.
Python
741
star