Ecosyste.ms: Awesome

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

https://github.com/ZhouYanzhao/ORN

Oriented Response Networks, in CVPR 2017
https://github.com/ZhouYanzhao/ORN

caffe cvpr pytorch rotation-invariant-features torch

Last synced: 3 months ago
JSON representation

Oriented Response Networks, in CVPR 2017

Lists

README 

        
# Oriented Response Networks

[![](https://img.shields.io/badge/torch-V1.0-green.svg)](https://github.com/ZhouYanzhao/ORN/tree/torch) [![](https://img.shields.io/badge/pytorch-alpha-blue.svg)](https://github.com/ZhouYanzhao/ORN/tree/pytorch-v2) [![](https://img.shields.io/badge/caffe-alpha-red.svg)](https://github.com/ZhouYanzhao/ORN/tree/caffe)



[[Home]](http://yzhou.work) [[Project]](http://yzhou.work/ORN) [[Paper]](https://arxiv.org/pdf/1701.01833) [[Supp]](http://yzhou.work/ORN/Supplementary.pdf) [[Poster]](http://yzhou.work/ORN/0160_POSTER.pdf)



![illustration](illustration.png)



## 🎉Update: Reimplemented ORN that supports modern PyTorch.

* Tested with PyTorch 1.12.0 (Ubuntu / GTX 2080 Ti)

* A New helper function `upgrade_to_orn` for easy model conversion.

* Predefined ORN-upgraded models (OR-VGG, OR-ResNet, OR-Inception, OR-WRN, etc.).



Please check the [pytorch-v2 branch](https://github.com/ZhouYanzhao/ORN/tree/pytorch-v2) for more details.



## Torch Implementation

The [torch branch](https://github.com/ZhouYanzhao/ORN/tree/torch) contains:



* the official **torch** implementation of ORN.

* the **MNIST-Variants** demo.



Please follow the instruction below to install it and run the experiment demo.



### Prerequisites

* Linux (tested on ubuntu 14.04LTS)

* NVIDIA GPU + CUDA CuDNN (CPU mode and CUDA without CuDNN mode are also available but significantly slower)

* [Torch7](http://torch.ch/docs/getting-started.html)



### Getting started

You can setup everything via a single command `wget -O - https://git.io/vHCMI | bash` **or** do it manually in case something goes wrong:



1. install the dependencies (required by the demo code):

    * [torchnet](https://github.com/torchnet/torchnet): `luarocks install torchnet`

    * [optnet](https://github.com/fmassa/optimize-net): `luarocks install optnet`



2. clone the torch branch: 



    ```bash

    # git version must be greater than 1.9.10

    git clone https://github.com/ZhouYanzhao/ORN.git -b torch --single-branch ORN.torch

    cd ORN.torch

    export DIR=$(pwd)

    ```



3. install ORN: 



    ```bash

    cd $DIR/install

    # install the CPU/GPU/CuDNN version ORN.

    bash install.sh

    ```



4. unzip the MNIST dataset:



    ```bash

    cd $DIR/demo/datasets

    unzip MNIST

    ```



5. run the MNIST-Variants demo:



    ```bash

    cd $DIR/demo

    # you can modify the script to test different hyper-parameters

    bash ./scripts/Train_MNIST.sh

    ```



### Trouble shooting

If you run into `'cudnn.find' not found`, update Torch7 to the latest version via `cd  && bash ./update.sh` then re-install everything.



### More experiments



**CIFAR 10/100**



You can train the [OR-WideResNet](https://gist.github.com/ZhouYanzhao/c7f75cd8ea3c92e2044d71ac7bc30fab/raw/or-wrn.lua) model (converted from WideResNet by simply replacing Conv layers with ORConv layers) on CIFAR dataset with [WRN](https://github.com/szagoruyko/wide-residual-networks).

```bash

dataset=cifar10_original.t7 model=or-wrn widen_factor=4 depth=40 ./scripts/train_cifar.sh

```

With exactly the same settings, ORN-augmented WideResNet achieves state-of-the-art result while using significantly fewer parameters.



![CIFAR](CIFAR.png)



Network          | Params   | CIFAR-10 (ZCA) | CIFAR-10 (mean/std) | CIFAR-100 (ZCA) | CIFAR-100 (mean/std) 

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

DenseNet-100-12-dropout | 7.0M     | -    | 4.10 | -  | 20.20 |

DenseNet-190-40-dropout | 25.6M    | -    | 3.46 | -  | 17.18 |

WRN-40-4         | 8.9M     | 4.97 | 4.53 | 22.89 | 21.18 |

WRN-28-10-dropout| 36.5M    | 4.17 | 3.89 | 20.50 | 18.85 |

WRN-40-10-dropout| 55.8M    | -    | 3.80 | -  | 18.3  |

ORN-40-4(1/2)    | 4.5M     | 4.13 | 3.43 | 21.24 | 18.82 |

ORN-28-10(1/2)-dropout | 18.2M | 3.52 | **2.98** | 19.22 | **16.15** |



Table.1 Test error (%) on CIFAR10/100 dataset with flip/translation augmentation) 



**ImageNet**



![ILSVRC2012](ILSVRC2012.png)



The effectiveness of ORN is further verified on large scale data. The OR-ResNet-18 model upgraded from [ResNet-18](https://github.com/facebook/fb.resnet.torch) yields significant better performance when using similar parameters.



| Network      | Params | Top1-Error | Top5-Error |

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

| ResNet-18    |  11.7M |   30.614   |    10.98   |

| OR-ResNet-18 |  11.4M |   **28.916**   |    **9.88**    |



Table.2 Validation error (%) on ILSVRC-2012 dataset.



You can use [facebook.resnet.torch](https://github.com/facebook/fb.resnet.torch) to train the [OR-ResNet-18](https://gist.github.com/ZhouYanzhao/c7f75cd8ea3c92e2044d71ac7bc30fab/raw/or-resnet.lua) model from scratch or finetune it on your data by using the [pre-trained weights](https://1drv.ms/u/s!Avhhrlo9ASwciWcEjJ_KWBgTWWyg).



```lua

-- To fill the model with the pre-trained weights:

model = require('or-resnet.lua')({tensorType='torch.CudaTensor', pretrained='or-resnet18_weights.t7'})

```



A more specific demo notebook of using the pre-trained OR-ResNet to classify images can be found [here](classify.ipynb).



## ~~PyTorch Implementation (Deprecated)~~

**Please check the [pytorch-v2 branch](https://github.com/ZhouYanzhao/ORN/tree/pytorch-v2) for more details.**



The [pytorch branch](https://github.com/ZhouYanzhao/ORN/tree/pytorch) contains:



* the official **pytorch** implementation of ORN *(alpha version supports 1x1/3x3 ARFs with 4/8 orientation channels only)*.

* the **MNIST-Variants** demo.



Please follow the instruction below to install it and run the experiment demo.



### Prerequisites

* Linux (tested on ubuntu 14.04LTS)

* NVIDIA GPU + CUDA CuDNN (CPU mode and CUDA without CuDNN mode are also available but significantly slower)

* [PyTorch](http://pytorch.org)



### Getting started



1. install the dependencies (required by the demo code):

    * [tqdm](https://github.com/noamraph/tqdm): `pip install tqdm`

    * [pillow](https://python-pillow.org): `pip install Pillow`



2. clone the pytorch branch: 



    ```bash

    # git version must be greater than 1.9.10

    git clone https://github.com/ZhouYanzhao/ORN.git -b pytorch --single-branch ORN.pytorch

    cd ORN.pytorch

    export DIR=$(pwd)

    ```



3. install ORN: 



    ```bash

    cd $DIR/install

    bash install.sh

    ```



4. run the MNIST-Variants demo:



    ```bash

    cd $DIR/demo

    # train ORN on MNIST-rot

    python main.py --use-arf

    # train baseline CNN

    python main.py

    ```



## Caffe Implementation 

The [caffe branch](https://github.com/ZhouYanzhao/ORN/tree/caffe) contains:



* the official **caffe** implementation of ORN *(alpha version supports 1x1/3x3 ARFs with 4/8 orientation channels only)*.

* the **MNIST-Variants** demo.



Please follow the instruction below to install it and run the experiment demo.



### Prerequisites

* Linux (tested on ubuntu 14.04LTS)

* NVIDIA GPU + CUDA CuDNN (CPU mode and CUDA without CuDNN mode are also available but significantly slower)

* [Caffe](http://caffe.berkeleyvision.org/)



### Getting started



1. install the dependency (required by the demo code):

    * [idx2numpy](https://github.com/ivanyu/idx2numpy): `pip install idx2numpy`



2. clone the caffe branch: 



    ```bash

    # git version must be greater than 1.9.10

    git clone https://github.com/ZhouYanzhao/ORN.git -b caffe --single-branch ORN.caffe

    cd ORN.caffe

    export DIR=$(pwd)

    ```



3. install ORN: 



    ```bash

    # modify Makefile.config first

    # compile ORN.caffe

    make clean && make -j"$(nproc)" all

    ```



4. run the MNIST-Variants demo:



    ```bash

    cd $DIR/examples/mnist

    bash get_mnist.sh

    # train ORN & CNN on MNIST-rot

    bash train.sh

    ```



### Note

Due to implementation differences,

* upgrading Conv layers to ORConv layers can be done by adding an `orn_param`

* num_output of ORConv layers should be multipied by nOrientation of ARFs



Example:

```YAML

layer {

  type: "Convolution"

  name: "ORConv" bottom: "Data" top: "ORConv"

  # add this line to replace regular filters with ARFs

  orn_param {orientations: 8}

  param { lr_mult: 1 decay_mult: 2}

  convolution_param {

    # this means 10 ARF feature maps

    num_output: 80

    kernel_size: 3

    stride: 1

    pad: 0

    weight_filler { type: "msra"}

    bias_filler { type: "constant" value: 0}

  }

}

```

Check the MNIST demo [prototxt](https://github.com/ZhouYanzhao/ORN/blob/caffe/examples/mnist/orn.prototxt) (and its [visualization](http://ethereon.github.io/netscope/#/gist/c7f75cd8ea3c92e2044d71ac7bc30fab)) for more details. 



## Citation 

If you use the code in your research, please cite:

```bibtex

@INPROCEEDINGS{Zhou2017ORN,

    author = {Zhou, Yanzhao and Ye, Qixiang and Qiu, Qiang and Jiao, Jianbin},

    title = {Oriented Response Networks},

    booktitle = {CVPR},

    year = {2017}

}

```