AutoML与轻量模型大列表

导读

一份高质量(最新的)AutoML工作和轻量级模型的列表，包括神经结构搜索，轻量级结构，模型压缩和加速，超参数优化，自动特征工程。

作者 | guan-yuan 

编译 | Xiaowen 

Github： 

https://github.com/guan-yuan/awesome-AutoML-and-Lightweight-Models

1.) Neural Architecture Search

[论文]

Gradient:

• ASAP: Architecture Search, Anneal and Prune | [2019/04]

• Single-Path NAS: Designing Hardware-Efficient ConvNets in less than 4 Hours | [2019/04]

• dstamoulis/single-path-nas | [Tensorflow]

• Automatic Convolutional Neural Architecture Search for Image Classification Under Different Scenes | [IEEE Access 2019]

• sharpDARTS: Faster and More Accurate Differentiable Architecture Search | [2019/03]

• Learning Implicitly Recurrent CNNs Through Parameter Sharing | [ICLR 2019]

• lolemacs/soft-sharing | [Pytorch]

• Probabilistic Neural Architecture Search | [2019/02]

• Auto-DeepLab: Hierarchical Neural Architecture Search for Semantic Image Segmentation | [2019/01]

• SNAS: Stochastic Neural Architecture Search | [ICLR 2019]

• FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable Neural Architecture Search | [2018/12]

• Neural Architecture Optimization | [NIPS 2018]

• renqianluo/NAO | [Tensorflow]

• DARTS: Differentiable Architecture Search | [2018/06]

• quark0/darts | [Pytorch]

• khanrc/pt.darts | [Pytorch]

• dragen1860/DARTS-PyTorch | [Pytorch]

Reinforcement Learning:

• Template-Based Automatic Search of Compact Semantic Segmentation Architectures | [2019/04]

• Understanding Neural Architecture Search Techniques | [2019/03]

• Fast, Accurate and Lightweight Super-Resolution with Neural Architecture Search | [2019/01]

• falsr/FALSR | [Tensorflow]

• Multi-Objective Reinforced Evolution in Mobile Neural Architecture Search | [2019/01]

• moremnas/MoreMNAS | [Tensorflow]

• ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware | [ICLR 2019]

• MIT-HAN-LAB/ProxylessNAS | [Pytorch, Tensorflow]

• Transfer Learning with Neural AutoML | [NIPS 2018]

• Learning Transferable Architectures for Scalable Image Recognition | [2018/07]

• wandering007/nasnet-pytorch | [Pytorch]

• tensorflow/models/research/slim/nets/nasnet | [Tensorflow]

• MnasNet: Platform-Aware Neural Architecture Search for Mobile | [2018/07]

• AnjieZheng/MnasNet-PyTorch | [Pytorch]

• Practical Block-wise Neural Network Architecture Generation | [CVPR 2018]

• Efficient Neural Architecture Search via Parameter Sharing | [ICML 2018]

• melodyguan/enas | [Tensorflow]

• carpedm20/ENAS-pytorch | [Pytorch]

• Efficient Architecture Search by Network Transformation | [AAAI 2018]

Evolutionary Algorithm:

• Single Path One-Shot Neural Architecture Search with Uniform Sampling | [2019/04]

• DetNAS: Neural Architecture Search on Object Detection | [2019/03]

• The Evolved Transformer | [2019/01]

• Designing neural networks through neuroevolution | [Nature Machine Intelligence 2019]

• EAT-NAS: Elastic Architecture Transfer for Accelerating Large-scale Neural Architecture Search | [2019/01]

• Efficient Multi-objective Neural Architecture Search via Lamarckian Evolution | [ICLR 2019]

SMBO:

• MFAS: Multimodal Fusion Architecture Search | [CVPR 2019]

• DPP-Net: Device-aware Progressive Search for Pareto-optimal Neural Architectures | [ECCV 2018]

• Progressive Neural Architecture Search | [ECCV 2018]

• titu1994/progressive-neural-architecture-search | [Keras, Tensorflow]

• chenxi116/PNASNet.pytorch | [Pytorch]

Random Search:

• Exploring Randomly Wired Neural Networks for Image Recognition | [2019/04]

• Searching for Efficient Multi-Scale Architectures for Dense Image Prediction | [NIPS 2018]

Hypernetwork:

• Graph HyperNetworks for Neural Architecture Search | [ICLR 2019]

Bayesian Optimization:

• Inductive Transfer for Neural Architecture Optimization | [2019/03]

Partial Order Pruning

• Partial Order Pruning: for Best Speed/Accuracy Trade-off in Neural Architecture Search | [CVPR 2019]

• lixincn2015/Partial-Order-Pruning | [Caffe]

Knowledge Distillation

• Improving Neural Architecture Search Image Classifiers via Ensemble Learning | [2019/03]

[项目]

• Microsoft/nni | [Python]

2.) Lightweight Structures

[论文]

Segmentation:

• ESPNetv2: A Light-weight, Power Efficient, and General Purpose Convolutional Neural Network | [2018/11]

• sacmehta/ESPNetv2 | [Pytorch]

• ESPNet: Efficient Spatial Pyramid of Dilated Convolutions for Semantic Segmentation | [ECCV 2018]

• sacmehta/ESPNet | [Pytorch]

• BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation | [ECCV 2018]

• ooooverflow/BiSeNet | [Pytorch]

• ycszen/TorchSeg | [Pytorch]

• ERFNet: Efficient Residual Factorized ConvNet for Real-time Semantic Segmentation | [T-ITS 2017]

• Eromera/erfnet_pytorch | [Pytorch]

Object Detection:

• Pooling Pyramid Network for Object Detection | [2018/09]

• tensorflow/models | [Tensorflow]

• Tiny-DSOD: Lightweight Object Detection for Resource-Restricted Usages | [BMVC 2018]

• lyxok1/Tiny-DSOD | [Caffe]

• Pelee: A Real-Time Object Detection System on Mobile Devices | [NeurIPS 2018]

• Robert-JunWang/Pelee | [Caffe]

• Robert-JunWang/PeleeNet | [Pytorch]

• Receptive Field Block Net for Accurate and Fast Object Detection | [ECCV 2018]

• ruinmessi/RFBNet | [Pytorch]

• ShuangXieIrene/ssds.pytorch | [Pytorch]

• lzx1413/PytorchSSD | [Pytorch]

• FSSD: Feature Fusion Single Shot Multibox Detector | [2017/12]

• ShuangXieIrene/ssds.pytorch | [Pytorch]

• lzx1413/PytorchSSD | [Pytorch]

• dlyldxwl/fssd.pytorch | [Pytorch]

• Feature Pyramid Networks for Object Detection | [CVPR 2017]

• tensorflow/models | [Tensorflow]

3.) Model Compression & Acceleration

[论文]

Compression:

• Slimmable Neural Networks | [ICLR 2019]

• JiahuiYu/slimmable_networks | [Pytorch]

• AMC: AutoML for Model Compression and Acceleration on Mobile Devices | [ECCV 2018]

• AutoML for Model Compression (AMC): Trials and Tribulations | [Pytorch]

• Learning Efficient Convolutional Networks through Network Slimming | [ICCV 2017]

• foolwood/pytorch-slimming | [Pytorch]

• Channel Pruning for Accelerating Very Deep Neural Networks | [ICCV 2017]

• yihui-he/channel-pruning | [Caffe]

• Pruning Convolutional Neural Networks for Resource Efficient Inference | [ICLR 2017]

• jacobgil/pytorch-pruning | [Pytorch]

• Pruning Filters for Efficient ConvNets | [ICLR 2017]

Acceleration:

• Fast Algorithms for Convolutional Neural Networks | [CVPR 2016]

• andravin/wincnn | [Python]

[项目]

• NervanaSystems/distiller | [Pytorch]

• Tencent/PocketFlow | [Tensorflow]

[教程/博客]

• Introducing the CVPR 2018 On-Device Visual Intelligence Challenge

4.) Hyperparameter Optimization

[论文]

• Tuning Hyperparameters without Grad Students: Scalable and Robust Bayesian Optimisation with Dragonfly | [2019/03]

• dragonfly/dragonfly

• Google vizier: A service for black-box optimization | [SIGKDD 2017]

[项目]

• Microsoft/nni | [Python]

• dragonfly/dragonfly | [Python]

[教程/博客]

• Hyperparameter tuning in Cloud Machine Learning Engine using Bayesian Optimization

• Overview of Bayesian Optimization

• Bayesian optimization

• krasserm/bayesian-machine-learning | [Python]

5.) Automated Feature Engineering

Model Analyzer

• Netscope CNN Analyzer | [Caffe]

• sksq96/pytorch-summary | [Pytorch]

• Lyken17/pytorch-OpCounter | [Pytorch]

References

• LITERATURE ON NEURAL ARCHITECTURE SEARCH

• handong1587/handong1587.github.io

• hibayesian/awesome-automl-papers

• mrgloom/awesome-semantic-segmentation

• amusi/awesome-object-detection

-END-

专 · 知

专知，专业可信的人工智能知识分发，让认知协作更快更好！欢迎登录www.zhuanzhi.ai，注册登录专知，获取更多AI知识资料！

欢迎微信扫一扫加入专知人工智能知识星球群，获取最新AI专业干货知识教程视频资料和与专家交流咨询！

请加专知小助手微信（扫一扫如下二维码添加），加入专知人工智能主题群，咨询技术商务合作~

专知《深度学习:算法到实战》课程全部完成！530+位同学在学习，现在报名，限时优惠！网易云课堂人工智能畅销榜首位！

点击“阅读原文”，了解报名专知《深度学习:算法到实战》课程