Skip to content
/ waifu2x-ncnn-vulkan-macos Public
forked from moeoverflow/waifu2x-ncnn-vulkan-macos

Tool for double image magnification using the waifu2x model and Vulkan API.

License

MIT license
0 stars 8 forks Branches Tags Activity
Star
Notifications You must be signed in to change notification settings

EETagent/waifu2x-ncnn-vulkan-macos

BranchesTags
 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

80 Commits
.github/img
.github/img
 
 
waifu2x.xcodeproj
waifu2x.xcodeproj
 
 
waifu2x
waifu2x
 
 
.gitignore
.gitignore
 
 
.gitmodules
.gitmodules
 
 
LICENSE
LICENSE
 
 
README.md
README.md
 
 

Repository files navigation

waifu2x-ncnn-vulkan-macos

Tool for double image magnification using the waifu2x model and Vulkan API.

Acknowledgement

Thanks

Thanks to @moeoverflow for the original repo.

Thanks to @shincurry for contributing to the UI of this project.

Usage

Single Mode

  1. Click Single in the top tab.
  2. Drag the image you want to enlarge to the left cell of the app.
  3. Adjust the settings at the bottom.
  4. Click 2x! and the generated image will be displayed on the right.
  5. Drag the image from the right cell to the location where you want to save it (the file name will be waifu2x-output.png).

screenshot

Multiple Mode

  1. Click Multiple in the top tab
  2. Drag and drop images or directories into the table. (Only decodable images will be processed.)
  3. Adjust the settings at the bottom
  4. Click on 2x! and the generated images will be saved to the location from where they originated with the .png extension. For example:

Input

.
├── 1
│   ├── IMG_2185.JPG
│   ├── IMG_2211.JPG
│   └── IMG_2212.JPG
├── 2
│   └── IMG_2208.PNG
└── IMG_2213.JPG

Output

.
.
├── 1
│   ├── IMG_2185.JPG
│   ├── IMG_2185.JPG.png
│   ├── IMG_2211.JPG
│   ├── IMG_2211.JPG.png
│   ├── IMG_2212.JPG
│   └── IMG_2212.JPG.png
├── 2
│   ├── IMG_2208.PNG
│   └── IMG_2208.PNG.png
├── IMG_2213.JPG
└── IMG_2213.JPG.png

screenshot

Benchmark

To run the benchmark, click on waifu2x-gui -> Benchmark.

screenshot

Build Instructions (macOS arm64)

Download the lastest Vulkan SDK at https://vulkan.lunarg.com/sdk/home#mac.

At the time of writing this README.md file, the latest version for macOS was 1.3.250.1.

Copy VulkanSDK/1.3.250.1 (or other version) as waifu2x-ncnn-vulkan-macos/waifu2x/VulkanSDK

brew install protobuf libomp

# if you have installed libomp before, you may have to forcefully reinstall it
brew link --force libomp

# clone this repo first
git clone --recursive --depth=1 https://github.com/EETagent/waifu2x-ncnn-vulkan-macos

# check your cmake installation
which cmake

# copy your VulkanSDK

cp -r /Users/mac/VulkanSDK/1.3.250.1 waifu2x-ncnn-vulkan-macos/waifu2x/VulkanSDK

# replace signing data in Xcode project and build app
xcodebuild

Notice

Double check if all included libraries in the Xcode exists

Speed Comparison between Macs

Environment 0

  • MacBook Pro 14-inch 2023
  • macOS 14.0 (23A5286i)
  • Apple M2 Max
  • Apple M2 Max 30-core 32 GB
Model Image Size Target Size Block Size Total Time(sec) GPU Memory(MB)
models-cunet 200x200 400x400 400/200/100 0.25/0.23/0.22 240/240/49
models-cunet 400x400 800x800 400/200/100 0.38/0.34/0.36 944/240/49
models-cunet 1000x1000 2000x2000 400/200/100 1.17/1.17/1.28 962/241/49
models-cunet 2000x2000 4000x4000 400/200/100 3.86/4.03/4.48 964/258/50
models-cunet 4000x4000 8000x8000 400/200/100 14.53/15.21/17.23 987/261/67
models-upconv_7_anime_style_art_rgb 200x200 400x400 400/200/100 0.18/0.15/0.15 141/141/33
models-upconv_7_anime_style_art_rgb 400x400 800x800 400/200/100 0.22/0.20/0.21 539/141/33
models-upconv_7_anime_style_art_rgb 1000x1000 2000x2000 400/200/100 0.58/0.59/0.62 541/142/49
models-upconv_7_anime_style_art_rgb 2000x2000 4000x4000 400/200/100 1.89/1.93/2.06 543/159/50
models-upconv_7_anime_style_art_rgb 4000x4000 8000x8000 400/200/100 6.68/6.88/7.43 566/162/51

noise: 2, scale: 2, gpuid: 0, tta mode: NO

Environment 1

  • MacBook Pro 15-inch 2018
  • macOS 10.14.6 (18G103)
  • Intel Core i9 8950HK
  • AMD Radeon Pro Vega 20
Model Image Size Target Size Block Size Total Time(sec) GPU Memory(MB)
models-cunet 200x200 400x400 400/200/100 0.47/0.43/0.49 613/613/172
models-cunet 400x400 800x800 400/200/100 0.97/0.88/0.95 2407/614/173
models-cunet 1000x1000 2000x2000 400/200/100 3.56/3.61/4.18 2415/617/175
models-cunet 2000x2000 4000x4000 400/200/100 12.72/13.22/15.49 2420/669/193
models-cunet 4000x4000 8000x8000 400/200/100 49.79/51.51/60.60 2452/645/197
models-upconv_7_anime_style_art_rgb 200x200 400x400 400/200/100 0.26/0.24/0.20 460/460/119
models-upconv_7_anime_style_art_rgb 400x400 800x800 400/200/100 0.47/0.40/0.40 1741/460/120
models-upconv_7_anime_style_art_rgb 1000x1000 2000x2000 400/200/100 1.67/1.64/1.73 1765/463/121
models-upconv_7_anime_style_art_rgb 2000x2000 4000x4000 400/200/100 6.12/6.11/6.49 1769/466/122
models-upconv_7_anime_style_art_rgb 4000x4000 8000x8000 400/200/100 23.75/23.71/25.27 1801/489/142

noise: 2, scale: 2, gpuid: 0, tta mode: NO

Environment 2

  • MacBook Pro 15-inch 2018
  • macOS 10.14.6 (18G84)
  • Intel Core i9 8850HK
  • AMD Radeon 560X
Model Image Size Target Size Block Size Total Time(sec) GPU Memory(MB)
models-cunet 200x200 400x400 400/200/100 2.00/0.53/0.56 613/613/172
models-cunet 400x400 800x800 400/200/100 1.29/1.18/1.26 2407/614/173
models-cunet 1000x1000 2000x2000 400/200/100 5.20/5.17/5.85 2415/617/175
models-cunet 2000x2000 4000x4000 400/200/100 19.07/19.35/22.25 2420/669/193
models-cunet 4000x4000 8000x8000 400/200/100 74.49/76.73/88.12 2452/644/197
models-upconv_7_anime_style_art_rgb 200x200 400x400 400/200/100 0.31/0.27/0.27 460/460/119
models-upconv_7_anime_style_art_rgb 400x400 800x800 400/200/100 0.65/0.54/0.55 1741/460/119
models-upconv_7_anime_style_art_rgb 1000x1000 2000x2000 400/200/100 2.41/2.34/2.48 1765/463/121
models-upconv_7_anime_style_art_rgb 2000x2000 4000x4000 400/200/100 8.74/8.84/9.43 1769/466/122
models-upconv_7_anime_style_art_rgb 4000x4000 8000x8000 400/200/100 32.66/33.00/35.29 1801/489/142

noise: 2, scale: 2, gpuid: 0, tta mode: NO

Environment 3

  • MacBook Pro 15-inch 2018
  • macOS 10.14.6 (18G103)
  • Intel Core i9 8950HK
  • Intel UHD Graphics 630
Model Image Size Target Size Block Size Total Time(sec) GPU Memory(MB)
models-cunet 200x200 400x400 400/200/100 0.95/0.99/0.87 616/616/176
models-cunet 400x400 800x800 400/200/100 2.53/2.27/2.48 2408/616/176
models-cunet 1000x1000 2000x2000 400/200/100 12.40/12.00/13.31 2408/616/176
models-cunet 2000x2000 4000x4000 400/200/100 44.50/46.81/52.13 2408/669/196
models-cunet 4000x4000 8000x8000 400/200/100 175.64/185.56/222.58 2431/637/196
models-upconv_7_anime_style_art_rgb 200x200 400x400 400/200/100 1.07/1.27/0.83 466/466/125
models-upconv_7_anime_style_art_rgb 400x400 800x800 400/200/100 2.30/1.09/1.11 1746/466/125
models-upconv_7_anime_style_art_rgb 1000x1000 2000x2000 400/200/100 6.30/5.70/5.92 1762/466/125
models-upconv_7_anime_style_art_rgb 2000x2000 4000x4000 400/200/100 22.12/22.48/23.71 1762/466/125
models-upconv_7_anime_style_art_rgb 4000x4000 8000x8000 400/200/100 87.28/89.11/93.98 1780/482/141

noise: 2, scale: 2, gpuid: 1, tta mode: NO

Environment 4

  • MacBook 12-inch Early 2016
  • macOS 10.14.6 (18G84)
  • Intel Core m7 6Y75
  • Intel HD Graphics 515
Model Image Size Target Size Block Size Total Time(sec) GPU Memory(MB)
models-cunet 200x200 400x400 400/200/100 1.12/1.41/1.23 616/616/176
models-cunet 400x400 800x800 400/200/100 3.06/2.90/3.34 2408/616/176
models-cunet 1000x1000 2000x2000 400/200/100 18.29/17.81/19.89 2408/616/176
models-cunet 2000x2000 4000x4000 400/200/100 66.55/71.79/85.83 2408/665/196
models-cunet 4000x4000 8000x8000 400/200/100 288.38/337.44/385.85 2431/637/196
models-upconv_7_anime_style_art_rgb 200x200 400x400 400/200/100 0.63/0.69/0.53 466/466/125
models-upconv_7_anime_style_art_rgb 400x400 800x800 400/200/100 1.62/1.43/1.47 1746/466/125
models-upconv_7_anime_style_art_rgb 1000x1000 2000x2000 400/200/100 9.19/9.06/9.46 1762/466/125
models-upconv_7_anime_style_art_rgb 2000x2000 4000x4000 400/200/100 35.52/37.66/41.57 1762/466/125
models-upconv_7_anime_style_art_rgb 4000x4000 8000x8000 400/200/100 199.20/182.04/159.11 1780/482/141

noise: 2, scale: 2, gpuid: 0, tta mode: NO

Speed Comparison (not really) with waifu2x-caffe-cui & waifu2x-ncnn-vulkan

Environment (waifu2x-caffe-cui & waifu2x-ncnn-vulkan)

  • Windows 10 1809
  • AMD R7-1700
  • Nvidia GTX-1070
  • Nvidia driver 419.67
  • CUDA 10.1.105
  • cuDNN 10.1

Environment (waifu2x-ncnn-vulkan-macos)

  • macOS 10.14.6 (18G103)
  • Intel Core i9 8950HK
  • AMD Radeon Pro Vega 20

cunet

Image Size Target Size Block Size Total Time(s) GPU Memory(MB)
waifu2x-ncnn-vulkan-macOS 200x200 400x400 400/200/100 0.46/0.44/0.43 621/621/180
waifu2x-ncnn-vulkan 200x200 400x400 400/200/100 0.86/0.86/0.82 638/638/197
waifu2x-caffe-cui 200x200 400x400 400/200/100 2.54/2.39/2.36 3017/936/843
waifu2x-ncnn-vulkan-macOS 400x400 800x800 400/200/100 0.91/0.84/0.92 2415/621/180
waifu2x-ncnn-vulkan 400x400 800x800 400/200/100 1.17/1.04/1.02 2430/638/197
waifu2x-caffe-cui 400x400 800x800 400/200/100 2.91/2.43/2.7 3202/1389/1178
waifu2x-ncnn-vulkan-macOS 1000x1000 2000x2000 400/200/100 3.54/3.58/4.18 2422/624/182
waifu2x-ncnn-vulkan 1000x1000 2000x2000 400/200/100 2.35/2.26/2.46 2430/638/197
waifu2x-caffe-cui 1000x1000 2000x2000 400/200/100 4.04/3.79/4.35 3258/1582/1175
waifu2x-ncnn-vulkan-macOS 2000x2000 4000x4000 400/200/100 12.83/13.25/15.44 2426/676/200
waifu2x-ncnn-vulkan 2000x2000 4000x4000 400/200/100 6.46/6.59/7.49 2430/686/213
waifu2x-caffe-cui 2000x2000 4000x4000 400/200/100 7.01/7.54/10.11 3258/1499/1200
waifu2x-ncnn-vulkan-macOS 4000x4000 8000x8000 400/200/100 49.56/51.44/60.56 2459/651/203
waifu2x-ncnn-vulkan 4000x4000 8000x8000 400/200/100 22.78/23.78/27.61 2448/654/213
waifu2x-caffe-cui 4000x4000 8000x8000 400/200/100 18.45/21.85/31.82 3325/1652/1236

upconv_7_anime_style_art_rgb

Image Size Target Size Block Size Total Time(s) GPU Memory(MB)
waifu2x-ncnn-vulkan-macOS 200x200 400x400 400/200/100 0.23/0.20/0.22 465/465/125
waifu2x-ncnn-vulkan 200x200 400x400 400/200/100 0.74/0.75/0.72 482/482/142
waifu2x-caffe-cui 200x200 400x400 400/200/100 2.04/1.99/1.99 995/546/459
waifu2x-ncnn-vulkan-macOS 400x400 800x800 400/200/100 0.49/0.42/0.41 1747/466/125
waifu2x-ncnn-vulkan 400x400 800x800 400/200/100 0.95/0.83/0.81 1762/482/142
waifu2x-caffe-cui 400x400 800x800 400/200/100 2.08/2.12/2.11 995/546/459
waifu2x-ncnn-vulkan-macOS 1000x1000 2000x2000 400/200/100 1.67/1.60/1.68 1770/468/127
waifu2x-ncnn-vulkan 1000x1000 2000x2000 400/200/100 1.52/1.41/1.44 1778/482/142
waifu2x-caffe-cui 1000x1000 2000x2000 400/200/100 2.72/2.60/2.68 1015/570/459
waifu2x-ncnn-vulkan-macOS 2000x2000 4000x4000 400/200/100 6.11/5.89/6.18 1774/472/128
waifu2x-ncnn-vulkan 2000x2000 4000x4000 400/200/100 3.45/3.42/3.63 1778/482/142
waifu2x-caffe-cui 2000x2000 4000x4000 400/200/100 3.90/4.01/4.35 1015/521/462
waifu2x-ncnn-vulkan-macOS 4000x4000 8000x8000 400/200/100 22.92/22.70/24.16 1806/495/147
waifu2x-ncnn-vulkan 4000x4000 8000x8000 400/200/100 11.16/11.29/12.07 1796/498/158
waifu2x-caffe-cui 4000x4000 8000x8000 400/200/100 9.24/9.81/11.16 995/546/436

About

Tool for double image magnification using the waifu2x model and Vulkan API.

Topics

macos objective-c vulkan objective-c-plus-plus waifu2x arm64 aarch64 m2 upscaling m1 moltenvk waifu2x-gui waifu2x-ncnn-vulkan waifu2x-mac m1-mac

Resources

Readme

License

MIT license
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases 1

waifu2x-ncnn-vulkan-macos arm64 Latest
Jul 17, 2023

Packages

No packages published

Languages

  • Objective-C++ 62.5%
  • Objective-C 37.5%