This Repo is the official CUDA implementation of ICCV 2019 Oral paper for CARAFE: Content-Aware ReAssembly of FEatures

Introduction

This Repo is the official CUDA implementation of ICCV 2019 Oral paper for CARAFE: Content-Aware ReAssembly of FEatures.

@inproceedings{Wang_2019_ICCV,
    title = {CARAFE: Content-Aware ReAssembly of FEatures},
    author = {Wang, Jiaqi and Chen, Kai and Xu, Rui and Liu, Ziwei and Loy, Chen Change and Lin, Dahua},
    booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
    month = {October},
    year = {2019}
}

Setup CARAFE Operator

There are two ways to setup CARAFE operator.

A. Install mmcv which contains CARAFE.

CARAFE is supported in mmcv. You may install mmcv following the official guideline.

https://github.com/open-mmlab/mmcv

B. Install CARAFE directly from GitHub.

Requirements:

CUDA >= 9.0, Pytorch >= 1.3, Python >= 3.6

Install with pip

pip install git+https://github.com/myownskyW7/[email protected]

Run gradient check to make sure the operator is successfully compiled

$ python

>>> from carafe import grad_check

C. Compile CARAFE from source.

Requirements:

CUDA >= 9.0, Pytorch >= 1.3, Python >= 3.6

Git clone this repo.

git clone https://github.com/myownskyW7/CARAFE

Setup CARAFE op.

cd CARAFE
python setup.py develop
# or "pip install -v -e ."

Run gradient check to make sure the operator is successfully compiled

$ python

>>> from carafe import grad_check

Usage

import torch
from mmcv.ops.carafe import CARAFEPack
# or "from carafe import CARAFEPack"


x = torch.rand(2, 40, 50, 70)
model = CARAFEPack(channels=40, scale_factor=2)

model = model.cuda()
x = x.cuda()

out = model(x)

print('original shape: ', x.shape)
print('upscaled shape: ', out.shape)

Applications

Projects with CARAFE operators

mmcv

mmdetection

mmediting