waifu2x

waifu2x

Image Super-Resolution for Anime-style art using Deep Convolutional Neural Networks. And it supports photo.

The demo application can be found at http://waifu2x.udp.jp/ .

Note that I only provide this website and this repository. Other software or website claiming "waifu2x" has nothing to do with me.

Summary

Click to see the slide show.

References

waifu2x is inspired by SRCNN [1]. 2D character picture (HatsuneMiku) is licensed under CC BY-NC by piapro [2].

• [1] Chao Dong, Chen Change Loy, Kaiming He, Xiaoou Tang, "Image Super-Resolution Using Deep Convolutional Networks", http://arxiv.org/abs/1501.00092
• [2] "For Creators", http://piapro.net/en_for_creators.html

Public AMI

TODO

Third Party Software

Third-Party

If you are a windows user, I recommend you to use waifu2x-caffe(Just download from releases tab), waifu2x-ncnn-vulkan or waifu2x-conver-cpp.

Dependencies

Hardware

• NVIDIA GPU

Platform

• Torch7
• NVIDIA CUDA

LuaRocks packages (excludes torch7's default packages)

• lua-csnappy
• md5
• uuid
• csvigo
• turbo

Installation

Setting Up the Command Line Tool Environment

(on Ubuntu 16.04)

Install CUDA

See: NVIDIA CUDA Getting Started Guide for Linux

Download CUDA

sudo dpkg -i cuda-repo-ubuntu1404_7.5-18_amd64.deb
sudo apt-get update
sudo apt-get install cuda

Install Package

sudo apt-get install libsnappy-dev
sudo apt-get install libgraphicsmagick1-dev
sudo apt-get install libssl1.0-dev # for web server

Note: waifu2x requires little-cms2 linked graphicsmagick. if you use macOS/homebrew, See #174.

Install Torch7

See: Getting started with Torch.

• For CUDA9.x/CUDA8.x, see #222
• For CUDA10.x, see #253

Getting waifu2x

git clone --depth 1 https://github.com/nagadomi/waifu2x.git

and install lua modules.

cd waifu2x
./install_lua_modules.sh

Validation

Testing the waifu2x command line tool.

th waifu2x.lua

Web Application

th web.lua

View at: http://localhost:8812/

Command line tools

Notes: If you have cuDNN library, than you can use cuDNN with -force_cudnn 1 option. cuDNN is too much faster than default kernel. If you got GPU out of memory error, you can avoid it with -crop_size option (e.g. -crop_size 128).

Noise Reduction

th waifu2x.lua -m noise -noise_level 1 -i input_image.png -o output_image.png

th waifu2x.lua -m noise -noise_level 0 -i input_image.png -o output_image.png
th waifu2x.lua -m noise -noise_level 2 -i input_image.png -o output_image.png
th waifu2x.lua -m noise -noise_level 3 -i input_image.png -o output_image.png

2x Upscaling

th waifu2x.lua -m scale -i input_image.png -o output_image.png

Noise Reduction + 2x Upscaling

th waifu2x.lua -m noise_scale -noise_level 1 -i input_image.png -o output_image.png

th waifu2x.lua -m noise_scale -noise_level 0 -i input_image.png -o output_image.png
th waifu2x.lua -m noise_scale -noise_level 2 -i input_image.png -o output_image.png
th waifu2x.lua -m noise_scale -noise_level 3 -i input_image.png -o output_image.png

Batch conversion

find /path/to/imagedir -name "*.png" -o -name "*.jpg" > image_list.txt
th waifu2x.lua -m scale -l ./image_list.txt -o /path/to/outputdir/prefix_%d.png

The output format supports %s and %d(e.g. %06d). %s will be replaced the basename of the source filename. %d will be replaced a sequence number. For example, when input filename is piyo.png, %s_%03d.png will be replaced piyo_001.png.

See also th waifu2x.lua -h.

Using photo model

Please add -model_dir models/photo to command line option, if you want to use photo model. For example,

th waifu2x.lua -model_dir models/photo -m scale -i input_image.png -o output_image.png

Video Encoding

* avconv is alias of ffmpeg on Ubuntu 14.04.

Extracting images and audio from a video. (range: 00:09:00 ~ 00:12:00)

mkdir frames
avconv -i data/raw.avi -ss 00:09:00 -t 00:03:00 -r 24 -f image2 frames/%06d.png
avconv -i data/raw.avi -ss 00:09:00 -t 00:03:00 audio.mp3

Generating a image list.

find ./frames -name "*.png" |sort > data/frame.txt

waifu2x (for example, noise reduction)

mkdir new_frames
th waifu2x.lua -m noise -noise_level 1 -resume 1 -l data/frame.txt -o new_frames/%d.png

Generating a video from waifu2xed images and audio.

avconv -f image2 -framerate 24 -i new_frames/%d.png -i audio.mp3 -r 24 -vcodec libx264 -crf 16 video.mp4

Train Your Own Model

Note1: If you have cuDNN library, you can use cudnn kernel with -backend cudnn option. And, you can convert trained cudnn model to cunn model with tools/rebuild.lua.

Note2: The command that was used to train for waifu2x's pretrained models is available at appendix/train_upconv_7_art.sh, appendix/train_upconv_7_photo.sh. Maybe it is helpful.

Data Preparation

Genrating a file list.

find /path/to/image/dir -name "*.png" > data/image_list.txt

You should use noise free images. In my case, waifu2x is trained with 6000 high-resolution-noise-free-PNG images.

Converting training data.

th convert_data.lua

Train a Noise Reduction(level1) model

mkdir models/my_model
th train.lua -model_dir models/my_model -method noise -noise_level 1 -test images/miku_noisy.png
# usage
th waifu2x.lua -model_dir models/my_model -m noise -noise_level 1 -i images/miku_noisy.png -o output.png

You can check the performance of model with models/my_model/noise1_best.png.

Train a Noise Reduction(level2) model

th train.lua -model_dir models/my_model -method noise -noise_level 2 -test images/miku_noisy.png
# usage
th waifu2x.lua -model_dir models/my_model -m noise -noise_level 2 -i images/miku_noisy.png -o output.png

You can check the performance of model with models/my_model/noise2_best.png.

Train a 2x UpScaling model

th train.lua -model upconv_7 -model_dir models/my_model -method scale -scale 2 -test images/miku_small.png
# usage
th waifu2x.lua -model_dir models/my_model -m scale -scale 2 -i images/miku_small.png -o output.png

You can check the performance of model with models/my_model/scale2.0x_best.png.

Train a 2x and noise reduction fusion model

th train.lua -model upconv_7 -model_dir models/my_model -method noise_scale -scale 2 -noise_level 1 -test images/miku_small.png
# usage
th waifu2x.lua -model_dir models/my_model -m noise_scale -scale 2 -noise_level 1 -i images/miku_small.png -o output.png

You can check the performance of model with models/my_model/noise1_scale2.0x_best.png.

Docker

( Docker image is available at https://hub.docker.com/r/nagadomi/waifu2x )

Requires nvidia-docker.

docker build -t waifu2x .
docker run --gpus all -p 8812:8812 waifu2x th web.lua
docker run --gpus all -v `pwd`/images:/images waifu2x th waifu2x.lua -force_cudnn 1 -m scale -scale 2 -i /images/miku_small.png -o /images/output.png

Note that running waifu2x in without JIT caching is very slow, which is what would happen if you use docker. For a workaround, you can mount a host volume to the CUDA_CACHE_PATH, for instance,

docker run --gpus all -v $PWD/ComputeCache:/root/.nv/ComputeCache waifu2x th waifu2x.lua --help