Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/KaimingHe/deep-residual-networks

Deep Residual Learning for Image Recognition
https://github.com/KaimingHe/deep-residual-networks

Last synced: about 2 months ago
JSON representation

Deep Residual Learning for Image Recognition

Awesome Lists containing this project

README 

        
# Deep Residual Networks



By [Kaiming He](http://kaiminghe.com), [Xiangyu Zhang](https://scholar.google.com/citations?user=yuB-cfoAAAAJ&hl=en), [Shaoqing Ren](http://home.ustc.edu.cn/~sqren/), [Jian Sun](http://research.microsoft.com/en-us/people/jiansun/).



Microsoft Research Asia (MSRA).



### Table of Contents

0. [Introduction](#introduction)

0. [Citation](#citation)

0. [Disclaimer and known issues](#disclaimer-and-known-issues)

0. [Models](#models)

0. [Results](#results)

0. [Third-party re-implementations](#third-party-re-implementations)



### Introduction



This repository contains the original models (ResNet-50, ResNet-101, and ResNet-152) described in the paper "Deep Residual Learning for Image Recognition" (http://arxiv.org/abs/1512.03385). These models are those used in [ILSVRC] (http://image-net.org/challenges/LSVRC/2015/) and [COCO](http://mscoco.org/dataset/#detections-challenge2015) 2015 competitions, which won the 1st places in: ImageNet classification, ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.



**Note**



0. Re-implementations with **training code** and models from Facebook AI Research (FAIR): [blog](http://torch.ch/blog/2016/02/04/resnets.html), [code](https://github.com/facebook/fb.resnet.torch)

0. Code of improved **1K-layer ResNets** with 4.62% test error on CIFAR-10 in our new arXiv paper: https://github.com/KaimingHe/resnet-1k-layers



### Citation



If you use these models in your research, please cite:



	@article{He2015,

		author = {Kaiming He and Xiangyu Zhang and Shaoqing Ren and Jian Sun},

		title = {Deep Residual Learning for Image Recognition},

		journal = {arXiv preprint arXiv:1512.03385},

		year = {2015}

	}



### Disclaimer and known issues



0. These models are converted from our own implementation to a recent version of Caffe (2016/2/3, b590f1d). The numerical results using this code are as in the tables below.

0. These models are for the usage of testing or fine-tuning.

0. These models were **not** trained using this version of Caffe.

0. If you want to train these models using this version of Caffe without modifications, please notice that:

	- GPU memory might be insufficient for extremely deep models.

	- Changes of mini-batch size should impact accuracy (we use a mini-batch of 256 images on 8 GPUs, that is, 32 images per GPU).

	- Implementation of data augmentation might be different (see our paper about the data augmentation we used).

	- We randomly shuffle data at the beginning of every epoch.

	- There might be some other untested issues.

0. In our BN layers, the provided mean and variance are strictly computed using average (**not** moving average) on a sufficiently large training batch after the training procedure. The numerical results are very stable (variation of val error < 0.1%). Using moving average might lead to different results. 

0. In the BN paper, the BN layer learns gamma/beta. To implement BN in this version of Caffe, we use its provided "batch_norm_layer" (which has no gamma/beta learned) followed by "scale_layer" (which learns gamma/beta).

0. We use Caffe's implementation of SGD with momentum: v := momentum\*v + lr\*g. **If you want to port these models to other libraries (e.g., Torch, CNTK), please pay careful attention to the possibly different implementation of SGD with momentum**: v := momentum\*v + (1-momentum)\*lr\*g, which changes the effective learning rates.



	

### Models



0. Visualizations of network structures (tools from [ethereon](http://ethereon.github.io/netscope/quickstart.html)):

	- [ResNet-50] (http://ethereon.github.io/netscope/#/gist/db945b393d40bfa26006)

	- [ResNet-101] (http://ethereon.github.io/netscope/#/gist/b21e2aae116dc1ac7b50)

	- [ResNet-152] (http://ethereon.github.io/netscope/#/gist/d38f3e6091952b45198b)



0. Model files:

	- ~~MSR download: [link] (http://research.microsoft.com/en-us/um/people/kahe/resnet/models.zip)~~

	- OneDrive download: [link](https://onedrive.live.com/?authkey=%21AAFW2-FVoxeVRck&id=4006CBB8476FF777%2117887&cid=4006CBB8476FF777)



### Results

0. Curves on ImageNet (solid lines: 1-crop val error; dashed lines: training error):

	![Training curves](https://cloud.githubusercontent.com/assets/11435359/13046277/e904c04c-d412-11e5-9260-efc5b8301e2f.jpg)



0. 1-crop validation error on ImageNet (center 224x224 crop from resized image with shorter side=256):



	model|top-1|top-5

	:---:|:---:|:---:

	[VGG-16](http://www.vlfeat.org/matconvnet/pretrained/)|[28.5%](http://www.vlfeat.org/matconvnet/pretrained/)|[9.9%](http://www.vlfeat.org/matconvnet/pretrained/)

	ResNet-50|24.7%|7.8%

	ResNet-101|23.6%|7.1%

	ResNet-152|23.0%|6.7%

	

0. 10-crop validation error on ImageNet (averaging softmax scores of 10 224x224 crops from resized image with shorter side=256), the same as those in the paper:



	model|top-1|top-5

	:---:|:---:|:---:

	ResNet-50|22.9%|6.7%

	ResNet-101|21.8%|6.1%

	ResNet-152|21.4%|5.7%

	

### Third-party re-implementations



Deep residual networks are very easy to implement and train. We recommend to see also the following third-party re-implementations and extensions:



0. By Facebook AI Research (FAIR), with **training code in Torch and pre-trained ResNet-18/34/50/101 models for ImageNet**: [blog](http://torch.ch/blog/2016/02/04/resnets.html), [code](https://github.com/facebook/fb.resnet.torch)

0. Torch, CIFAR-10, with ResNet-20 to ResNet-110, training code, and curves: [code](https://github.com/gcr/torch-residual-networks)

0. Lasagne, CIFAR-10, with ResNet-32 and ResNet-56 and training code: [code](https://github.com/Lasagne/Recipes/tree/master/papers/deep_residual_learning)

0. Neon, CIFAR-10, with pre-trained ResNet-32 to ResNet-110 models, training code, and curves: [code](https://github.com/apark263/cfmz)

0. Torch, MNIST, 100 layers: [blog](https://deepmlblog.wordpress.com/2016/01/05/residual-networks-in-torch-mnist/), [code](https://github.com/arunpatala/residual.mnist)

0. A winning entry in Kaggle's right whale recognition challenge: [blog](http://blog.kaggle.com/2016/02/04/noaa-right-whale-recognition-winners-interview-2nd-place-felix-lau/), [code](https://github.com/felixlaumon/kaggle-right-whale)

0. Neon, Place2 (mini), 40 layers: [blog](http://www.nervanasys.com/using-neon-for-scene-recognition-mini-places2/), [code](https://github.com/hunterlang/mpmz/)

0. MatConvNet, CIFAR-10, with ResNet-20 to ResNet-110, training code, and curves: [code](https://github.com/suhangpro/matresnet)

0. TensorFlow, CIFAR-10, with ResNet-32,110,182 training code and curves:

[code](https://github.com/ppwwyyxx/tensorpack/tree/master/examples/ResNet)

0. MatConvNet, reproducing CIFAR-10 and ImageNet experiments (supporting official MatConvNet), training code and curves: [blog](https://zhanghang1989.github.io/ResNet/), [code](https://github.com/zhanghang1989/ResNet-Matconvnet)

0. Keras, ResNet-50: [code](https://github.com/raghakot/keras-resnet)



Converters:



0. MatConvNet: [url](http://www.vlfeat.org/matconvnet/pretrained/#imagenet-ilsvrc-classification)

0. TensorFlow: [url](https://github.com/ry/tensorflow-resnet)