Top Related Projects
Image Super-Resolution for Anime-Style Art
Video, Image and GIF upscale/enlarge(Super-Resolution) and Video frame interpolation. Achieved with Waifu2x, Real-ESRGAN, Real-CUGAN, RTX Video Super Resolution VSR, SRMD, RealSR, Anime4K, RIFE, IFRNet, CAIN, DAIN, and ACNet.
waifu2x converter ncnn version, runs fast on intel / amd / nvidia / apple-silicon GPU with vulkan
Quick Overview
Waifu2x-caffe is an image upscaling and noise reduction tool specifically designed for anime-style art and images. It uses convolutional neural networks to enhance image quality and is implemented using the Caffe deep learning framework. This project is a Windows port of the original waifu2x.
Pros
- High-quality upscaling and noise reduction for anime-style images
- Supports both CPU and GPU processing for faster performance
- Offers a user-friendly graphical interface for easy use
- Provides command-line interface for advanced users and batch processing
Cons
- Limited to Windows operating system
- Requires a powerful GPU for optimal performance, especially with large images
- May produce artifacts or undesired results with non-anime style images
- Requires additional setup and dependencies compared to web-based alternatives
Code Examples
As waifu2x-caffe is primarily a standalone application rather than a code library, there are no direct code examples for integration. However, here are some command-line usage examples:
# Upscale an image by 2x with noise reduction
waifu2x-caffe-cui.exe -i input.png -o output.png -m noise_scale -n 1 -s 2
# Denoise an image without upscaling
waifu2x-caffe-cui.exe -i input.png -o output.png -m noise -n 2
# Upscale an image by 4x without noise reduction
waifu2x-caffe-cui.exe -i input.png -o output.png -m scale -s 4
Getting Started
- Download the latest release from the GitHub repository.
- Extract the files to a desired location on your Windows machine.
- Install CUDA and cuDNN if you want to use GPU acceleration.
- Run
waifu2x-caffe.exe
for the graphical interface or usewaifu2x-caffe-cui.exe
for command-line operations. - For the GUI: Select an input image, choose processing options, and click "Process" to start.
- For CLI: Use the examples provided above, adjusting parameters as needed.
Note: Refer to the project's README for detailed information on installation, dependencies, and usage options.
Competitor Comparisons
Image Super-Resolution for Anime-Style Art
Pros of waifu2x
- Original implementation with a longer development history
- Supports both CPU and CUDA (GPU) processing
- Includes a web-based demo for easy testing
Cons of waifu2x
- Written in Lua, which may be less familiar to some developers
- Requires Torch7 framework, which has a steeper learning curve
- Limited platform support (mainly Linux and macOS)
Code Comparison
waifu2x:
require 'image'
local iproc = require 'iproc'
local reconstruct = require 'reconstruct'
local image = image.load(input_file)
local output = reconstruct(image, model, offset, crop_size)
waifu2x-caffe:
#include "caffe/caffe.hpp"
#include "waifu2x.hpp"
Waifu2x w2x(modelDir, gpuNo);
cv::Mat output;
w2x.waifu2x(input, output, noise_level, scale_ratio);
The waifu2x repository uses Lua with the Torch7 framework, while waifu2x-caffe is implemented in C++ using the Caffe deep learning framework. This difference in implementation languages and frameworks affects performance, ease of use, and platform compatibility.
Video, Image and GIF upscale/enlarge(Super-Resolution) and Video frame interpolation. Achieved with Waifu2x, Real-ESRGAN, Real-CUGAN, RTX Video Super Resolution VSR, SRMD, RealSR, Anime4K, RIFE, IFRNet, CAIN, DAIN, and ACNet.
Pros of Waifu2x-Extension-GUI
- User-friendly graphical interface for easier operation
- Supports multiple AI models, including waifu2x, Real-ESRGAN, and Anime4K
- Offers video processing capabilities in addition to image upscaling
Cons of Waifu2x-Extension-GUI
- May have higher system requirements due to additional features
- Potentially slower processing speed compared to the lightweight waifu2x-caffe
- More complex setup and configuration options
Code Comparison
While both projects are based on the waifu2x algorithm, their implementations differ significantly. Waifu2x-caffe is primarily focused on the core upscaling functionality, while Waifu2x-Extension-GUI provides a wrapper and additional features.
Waifu2x-caffe (C++):
cv::Mat process(cv::Mat &image) {
// Image processing logic
}
Waifu2x-Extension-GUI (Python):
def process_image(image_path, model, scale):
# Image processing with multiple AI models
pass
The code structures reflect their different approaches, with waifu2x-caffe focusing on low-level image processing and Waifu2x-Extension-GUI emphasizing high-level integration of multiple AI models and user interface interactions.
waifu2x converter ncnn version, runs fast on intel / amd / nvidia / apple-silicon GPU with vulkan
Pros of waifu2x-ncnn-vulkan
- Utilizes Vulkan API for improved GPU acceleration and cross-platform support
- Generally faster processing times, especially on modern GPUs
- Smaller file size and easier deployment due to fewer dependencies
Cons of waifu2x-ncnn-vulkan
- May have slightly lower image quality in some cases compared to waifu2x-caffe
- Less extensive model support and customization options
- Requires a Vulkan-compatible GPU for optimal performance
Code Comparison
waifu2x-ncnn-vulkan:
ncnn::VulkanDevice* vkdev = ncnn::get_gpu_device();
ncnn::VkAllocator* blob_vkallocator = vkdev->acquire_blob_allocator();
ncnn::VkAllocator* staging_vkallocator = vkdev->acquire_staging_allocator();
waifu2x-caffe:
caffe::Caffe::set_mode(caffe::Caffe::GPU);
caffe::Net<float> net(model_file, caffe::TEST);
net.CopyTrainedLayersFrom(trained_file);
The code snippets show the initialization process for each project. waifu2x-ncnn-vulkan uses Vulkan-specific allocators, while waifu2x-caffe initializes a Caffe network for GPU processing.
Pros of ailab
- More comprehensive AI research focus beyond just image upscaling
- Likely has more active development and resources from Bilibili
- Potentially offers a wider range of AI-related tools and models
Cons of ailab
- May be less specialized for image upscaling compared to waifu2x-caffe
- Possibly more complex to use for those only interested in image enhancement
- Could have a steeper learning curve due to broader scope
Code Comparison
waifu2x-caffe:
void filter_CV(cv::Mat &dst, cv::Mat &src, std::vector<cv::Mat> &kernels) {
int nkernels = kernels.size();
for (int i = 0; i < nkernels; i++) {
cv::filter2D(src, dst, CV_32F, kernels[i]);
}
}
ailab:
def process_image(image, model):
preprocessed = preprocess(image)
output = model(preprocessed)
result = postprocess(output)
return result
Note: The code snippets are hypothetical examples to illustrate potential differences in approach and language used. waifu2x-caffe likely uses C++ with OpenCV, while ailab might use Python with deep learning frameworks.
Convert designs to code with AI
Introducing Visual Copilot: A new AI model to turn Figma designs to high quality code using your components.
Try Visual CopilotREADME
waifu2x-caffe (for Windows)
å¶ä½è : lltcggie
æ¬ã½ããã¯ãç»åå¤æã½ããã¦ã§ã¢ãwaifu2xãã®å¤ææ©è½ã®ã¿ãã Caffeãç¨ãã¦æ¸ãç´ããWindowsåãã«ãã«ãããã½ããã§ãã CPUã§å¤æãããã¨ãåºæ¥ã¾ãããCUDA(ãããã¯cuDNN)ã使ãã¨CPUããé«éã«å¤æãããã¨ãåºæ¥ã¾ãã
GUI supports English, Japanese, Simplified Chinese, Traditional Chinese, Korean, Turkish, Spanish, Russian, and French.
ã½ããã®ãã¦ã³ãã¼ãã¯ãã¡ãã®releasesãã¼ã¸ã®ãDownloadsãã®é ã§åºæ¥ã¾ãã
è¦æ±ç°å¢
ãã®ã½ãããåä½ãããã«ã¯æä½ã§ã以ä¸ã®ç°å¢ãå¿ è¦ã§ãã
- OS : Windows Vista以é 64bit (32bitç¨exeã¯ããã¾ãã)
- ã¡ã¢ãª : 空ãã¡ã¢ãª1GBä»¥ä¸ (ãã ããå¤æããç»åãµã¤ãºã«ãã)
- GPU : Compute Capability 3.5 以ä¸ã®NVIDIA製GPU(CPUã§å¤æããå ´åã¯ä¸è¦)
- Microsoft Visual C++ 2015 åé å¸å¯è½ããã±ã¼ã¸ Update 3(x64ç)ãã¤ã³ã¹ãã¼ã«ããã¦ãããã¨(å¿
é )
- ä¸è¨ããã±ã¼ã¸ã¯ãã¡ã
ãã¦ã³ãã¼ã
ãã¿ã³ãæ¼ããå¾ãvcredist_x64.exe
ãé¸æãããã¦ã³ãã¼ãã»ã¤ã³ã¹ãã¼ã«ãè¡ã£ã¦ä¸ããã- è¦ã¤ãããªãå ´åã¯ããVisual C++ 2015 åé å¸å¯è½ããã±ã¼ã¸ Update 3ãã§æ¤ç´¢ãã¦ã¿ã¦ä¸ããã
èªåã®GPUã®Compute Capabilityãç¥ãããå ´åã¯ãã¡ãã®ãã¼ã¸ã§èª¿ã¹ã¦ä¸ããã
使ãæ¹(GUIç)
ãwaifu2x-caffe.exeãã¯GUIã½ããã§ããããã«ã¯ãªãã¯ã§èµ·åãã¾ãã ã¾ãã¯ã¨ã¯ã¹ããã¼ã©ã§ãwaifu2x-caffe.exeãã«ãã¡ã¤ã«ããã©ã«ãããã©ãã°&ããããã§æ¾ãè¾¼ãã¨ãååèµ·åæã®è¨å®ã§å¤æãè¡ãã¾ãã ãã®å ´åãè¨å®ã«ãã£ã¦ã¯å¤æã«æåãããèªåã§ãã¤ã¢ãã°ãéãããã¾ãã ã¾ããGUIã§ãã³ãã³ãã©ã¤ã³ã«ãããªãã·ã§ã³è¨å®ãè¡ããã¨ãåºæ¥ã¾ãã 詳ããã¯ã³ãã³ãã©ã¤ã³ãªãã·ã§ã³(å ±é)ã¨ã³ãã³ãã©ã¤ã³ãªãã·ã§ã³(GUIç)ã®é ããèªã¿ãã ããã
èµ·åå¾ããå ¥åãã¹ãæ¬ã«ç»åããã©ã«ãããã©ãã°&ããããã§æ¾ãè¾¼ãã¨ãåºåãã¹ãæ¬ãèªåã§è¨å®ããã¾ãã åºåå ãå¤ãããå ´åã¯ãåºåãã¹ãæ¬ãå¤æ´ãã¦ä¸ããã
好ã¿ã«åããã¦å¤æè¨å®ãå¤æ´ãããã¨ãåºæ¥ã¾ãã
å ¥åºåè¨å®
ãã¡ã¤ã«ã®å
¥åºåã«é¢ããè¨å®é
ç®ç¾¤ã§ãã
ãå ¥åãã¹ã
å¤æããããã¡ã¤ã«ã®ãã¹ãæå®ãã¾ãã
ãã©ã«ããæå®ããå ´åã¯ããµããã©ã«ãå
ãå«ããããã©ã«ãå
ã®å¤æããæ¡å¼µåããä»ããã¡ã¤ã«ãå¤æ対象ã¨ãã¾ãã
è¤æ°ã®ãã¡ã¤ã«ããã©ã«ãããã©ãã°ã§æå®ãããã¨ãåºæ¥ã¾ãã
ãã®å ´åã¯æ°ãããã©ã«ãã®ä¸ã«ãã¡ã¤ã«ããã©ã«ãæ§é ãç¶æããã¾ã¾åºåããã¾ãã
ï¼å
¥åãã¹æ¬ã«ã¯ã(Multi Files)ãã¨è¡¨ç¤ºããã¾ããåºåãã©ã«ãåã¯ãã¦ã¹ã§æ´ãã§ãããã¡ã¤ã«ããã©ã«ãåããçæããã¾ãï¼
åç
§ãã¿ã³ãæ¼ãã¦ãã¡ã¤ã«ãé¸æããå ´åãåä½ã®ãã¡ã¤ã«ããã©ã«ãããè¤æ°ã®ãã¡ã¤ã«ãé¸æã§ãã¾ãã
ãåºåãã¹ã
å¤æå¾ã®ç»åãä¿åãããã¹ãæå®ãã¾ãã
ãå
¥åãã¹ã㧠ãã©ã«ããæå®ããå ´åã¯ãæå®ãããã©ã«ãã®ä¸ã«å¤æãããã¡ã¤ã«ã(ãã©ã«ãæ§é ããã®ã¾ã¾ã§)ä¿åãã¾ããæå®ãããã©ã«ãããªãå ´åã¯èªåã§ä½æãã¾ãã
ããã©ã«ãå ã®å¤æããæ¡å¼µåã
ãå
¥åãã¹ãããã©ã«ãã®å ´åã®ããã©ã«ãå
ã®å¤æããç»åã®æ¡å¼µåãæå®ãã¾ãã
ããã©ã«ãå¤ã¯`png:jpg:jpeg:tif:tiff:bmp:tga`ã§ãã
ã¾ããåºåãæåã¯`:`ã§ãã
大æåå°æåã¯åºå¥ãã¾ããã
ä¾. png:jpg:jpeg:tif:tiff:bmp:tga
ãåºåæ¡å¼µåã
å¤æå¾ã®ç»åã®å½¢å¼ãæå®ãã¾ãã
ãåºåç»è³ªè¨å®ãã¨ãåºå深度ãããæ°ãã«è¨å®ã§ããå¤ã¯ããã§æå®ããå½¢å¼ã«ããç°ãªãã¾ãã
ãåºåç»è³ªè¨å®ã
å¤æå¾ã®ç»åã®ç»è³ªãæå®ãã¾ãã
è¨å®ã§ããå¤ã¯æ´æ°ã§ãã
æå®ã§ããå¤ã®ç¯å²ã¨æå³ã¯ãåºåæ¡å¼µåãã§è¨å®ããå½¢å¼ã«ããç°ãªãã¾ãã
* .jpg : å¤ã®ç¯å²(0ï½100) æ°åãé«ãã»ã©é«ç»è³ª
* .webp : å¤ã®ç¯å²(1ï½100) æ°åãé«ãã»ã©é«ç»è³ª 100ã ã¨å¯éå§ç¸®
* .tga : å¤ã®ç¯å²(0ï½1) 0ãªãå§ç¸®ãªãã1ãªãRLEå§ç¸®
ãåºå深度ãããæ°ã
å¤æå¾ã®ç»åã®1ãã£ã³ãã«ãããã®ãããæ°ãæå®ãã¾ãã
æå®ã§ããå¤ã¯ãåºåæ¡å¼µåãã§è¨å®ããå½¢å¼ã«ããç°ãªãã¾ãã
å¤æç»è³ªã»å¦çè¨å®
ãã¡ã¤ã«å¤æã®å¦çæ¹æ³ãç»è³ªã«é¢ããè¨å®é
ç®ç¾¤ã§ãã
ãå¤æã¢ã¼ãã
å¤æã¢ã¼ããæå®ãã¾ãã
* ãã¤ãºé¤å»ã¨æ¡å¤§ : ãã¤ãºé¤å»ã¨æ¡å¤§ãè¡ãã¾ã
* æ¡å¤§ : æ¡å¤§ãè¡ãã¾ã
* ãã¤ãºé¤å» : ãã¤ãºé¤å»ãè¡ãã¾ã
* ãã¤ãºé¤å»(èªåå¤å¥)ã¨æ¡å¤§ : æ¡å¤§ãè¡ãã¾ããå
¥åãJPEGç»åã®å ´åã®ã¿ãã¤ãºé¤å»ãè¡ãã¾ã
ãJPEGãã¤ãºé¤å»ã¬ãã«ã
ãã¤ãºé¤å»ã¬ãã«ãæå®ãã¾ããã¬ãã«ã®é«ãã»ããããå¼·åã«ãã¤ãºãé¤å»ãã¾ãããã®ã£ãºãã¨ããçµµã«ãªãå¯è½æ§ãããã¾ãã
ãæ¡å¤§ãµã¤ãºã
æ¡å¤§å¾ã®ãµã¤ãºã®è¨å®ãè¡ãã¾ãã
* æ¡å¤§çã§æå® : ç»åãæå®ã®æ¡å¤§çã§æ¡å¤§ãã¾ã
* å¤æå¾ã®æ¨ªå¹
ã§æå® : ç»åã®ç¸¦æ¨ªæ¯ãç¶æããã¾ã¾ãæå®ããã横å¹
ã«ãªãããã«æ¡å¤§ãã¾ã(åä½ã¯ãã¯ã»ã«)
* å¤æå¾ã®ç¸¦å¹
ã§æå® : ç»åã®ç¸¦æ¨ªæ¯ãç¶æããã¾ã¾ãæå®ããã縦å¹
ã«ãªãããã«æ¡å¤§ãã¾ã(åä½ã¯ãã¯ã»ã«)
* å¤æå¾ã®ç¸¦æ¨ªå¹
ã§æå® : æå®ããã縦横å¹
ã«ãªãããã«æ¡å¤§ãã¾ããã1920x1080ãã®ããã«æå®ãã¾ã(åä½ã¯ãã¯ã»ã«)
2åãè¶
ããæ¡å¤§çã®å ´åã(ãã¤ãºãé¤å»ããå ´åã¯æåã®1åã ãè¡ã)æå®ãããæ¡å¤§çãè¶
ããã¾ã§2åãã¤æ¡å¤§ãã2ã®ç´¯ä¹åã§ãªãæ¡å¤§çã®å ´åã¯æå¾ã«ç¸®å°ããã¨ããå¦çãè¡ãã¾ãããã®ããå¤æçµæãã®ã£ãºãã¨ããçµµã«ãªãå¯è½æ§ãããã¾ãã
ãã¢ãã«ã
使ç¨ããã¢ãã«ãæå®ãã¾ãã
å¤æ対象ã®ç»åã«ãã£ã¦æé©ãªã¢ãã«ã¯ç°ãªãã®ã§ãæ§ã
ãªã¢ãã«ã試ãã¦ã¿ããã¨ããããããã¾ãã
* 2次å
ã¤ã©ã¹ã(RGBã¢ãã«) : ç»åã®RGBãã¹ã¦ãå¤æãã2次å
ã¤ã©ã¹ãç¨ã¢ãã«
* åçã»ã¢ãã¡(Photoã¢ãã«) : åçã»ã¢ãã¡ç¨ã®ã¢ãã«
* 2次å
ã¤ã©ã¹ã(UpRGBã¢ãã«) : 2次å
ã¤ã©ã¹ã(RGBã¢ãã«)ããé«éãã¤åç以ä¸ã®ç»è³ªã§å¤æããã¢ãã«ããã ãRGBã¢ãã«ããæ¶è²»ããã¡ã¢ãª(VRAM)ã®éãå¤ãã®ã§ãå¤æä¸ã«å¼·å¶çµäºããå ´åã¯åå²ãµã¤ãºã調ç¯ãããã¨
* åçã»ã¢ãã¡(UpPhotoã¢ãã«) : åçã»ã¢ãã¡(Photoã¢ãã«)ããé«éãã¤åç以ä¸ã®ç»è³ªã§å¤æããã¢ãã«ããã ãPhotoã¢ãã«ããæ¶è²»ããã¡ã¢ãª(VRAM)ã®éãå¤ãã®ã§ãå¤æä¸ã«å¼·å¶çµäºããå ´åã¯åå²ãµã¤ãºã調ç¯ãããã¨
* 2次å
ã¤ã©ã¹ã(Yã¢ãã«) : ç»åã®è¼åº¦ã®ã¿ãå¤æãã2次å
ã¤ã©ã¹ãç¨ã¢ãã«
* 2次å
ã¤ã©ã¹ã(UpResNet10ã¢ãã«) : 2次å
ã¤ã©ã¹ã(UpRGBã¢ãã«)ããé«ç»è³ªã§å¤æããã¢ãã«ããã®ã¢ãã«ã¯åå²ãµã¤ãºãéãã¨åºåçµæãå¤ããã®ã§æ³¨æãããã¨
* 2次å
ã¤ã©ã¹ã(CUnetã¢ãã«) : 2次å
ã¤ã©ã¹ããå梱ã®ã¢ãã«ã§ä¸çªé«ç»è³ªã§å¤æã§ããã¢ãã«ããã®ã¢ãã«ã¯åå²ãµã¤ãºãéãã¨åºåçµæãå¤ããã®ã§æ³¨æãããã¨
ãTTAã¢ã¼ãã使ãã
TTA(Test-Time Augmentation)ã¢ã¼ãã使ããã©ãããæå®ãã¾ãã
TTAã¢ã¼ãã使ãã¨å¤æã8åé
ããªã代ããã«ãPSNR(ç»åã®è©ä¾¡ææ°ã®ä¸ã¤)ã0.15ãããä¸ããããã§ãã
å¦çé度è¨å®
ç»åå¤æã®å¦çé度ã«å½±é¿ããè¨å®é
ç®ç¾¤ã§ãã
ãåå²ãµã¤ãºã
å
é¨ã§åå²ãã¦å¦çãè¡ãéã®å¹
ï¼ãã¯ã»ã«åä½ï¼ãæå®ãã¾ãã
æé©ãª(å¦çãæéã§çµãã)æ°å¤ã®æ±ºãæ¹ã¯ãåå²ãµã¤ãºãã®é
ã§èª¬æãã¾ãã
ã-------ãã§åºåããã¦ããä¸ã®æ¹ã¯å
¥åãããç»åã®ç¸¦æ¨ªãµã¤ãºã®ç´æ°ã
ä¸ã®æ¹ã¯ãcrop_size_list.txtãããèªã¿åºããæ±ç¨çãªåå²ãµã¤ãºã§ãã
åå²ãµã¤ãºã大ããããå ´åãè¦æ±ãããã¡ã¢ãªã®é(GPUã使ãå ´åã¯VRAMã®é)ãPCã§ä½¿ç¨ã§ããã¡ã¢ãªãè¶
ãã¦ã½ãããå¼·å¶çµäºããã®ã§æ°ãã¤ãã¦ãã ããã
å¦çé度ã«ãããªãã«å½±é¿ããã®ã§ãåãç»åãµã¤ãºã®ç»åããã©ã«ãæå®ã§å¤§éã«å¤æããã¨ãã¯ãæé©ãªåå²ãµã¤ãºã調ã¹ã¦ããå¤æãããã¨ããããããã¾ãã
ãã ããã¢ãã«ã«ãã£ã¦ã¯åå²ãµã¤ãºãå¤æ´ããã¨åºåçµæãå¤ãããã®ãããã®ã§æ°ãä»ãã¦ãã ããã
(ãã®å ´åã¯åå²ãµã¤ãºã¯ããã©ã«ãã使ç¨ããããããµã¤ãºã調æ´ãããã¨ã§å¦çé度ãä¸ãããã¨ãã§ãã¾ã)
ãããããµã¤ãºã
å
é¨ã§ã¾ã¨ãã¦å¦çãè¡ãéã®ãµã¤ãºãæå®ãã¾ãã
ããããµã¤ãºã大ããããã¨å¦çé度ãéããªããã¨ãããã¾ãã
åå²ãµã¤ãºã¨åæ§ã«è¦æ±ãããã¡ã¢ãªã®éãPCã§ä½¿ç¨ã§ããã¡ã¢ãªãè¶
ããªãããã«æ°ãä»ãã¦ãã ããã
åä½è¨å®
ãã¾ãå¤æ´ããæ©ä¼ããªãã¨æãããåä½è¨å®ãã¾ã¨ããè¨å®ç¾¤ã§ãã
ããã¡ã¤ã«å ¥åæèªåå¤æéå§è¨å®ã
åç
§ãã¿ã³ããã©ãã°ã¢ã³ãããããã§å
¥åãã¡ã¤ã«ãæå®ããéã«èªåã§å¤æãéå§ããã®ãè¨å®ãè¡ãã¾ãã
exeã«å
¥åãã¡ã¤ã«ãå¼æ°ã§ä¸ããå ´åã§ã¯ãã®é
ç®ã®è¨å®å
容ã¯å½±é¿ãã¾ããã
* èªåã§éå§ããªã : ãã¡ã¤ã«ãå
¥åãã¦ãèªåã§å¤æãéå§ãã¾ãã
* ãã¡ã¤ã«ã1ã¤ã§ãå
¥åãããéå§ : ãã¡ã¤ã«ã1ã¤ã§ãå
¥åãããèªåã§å¤æãéå§ãã¾ã
* ãã©ã«ããããã¯è¤æ°ãã¡ã¤ã«ãå
¥åãããéå§ : ãã©ã«ããè¤æ°ãã¡ã¤ã«ãå
¥åãããèªåã§å¤æãéå§ãã¾ããåä½ã®ç»åãã¡ã¤ã«ãå¤æããã®ã¯å¤æè¨å®ã®èª¿ç¯ãè¡ãã¨ãã ãã ãã¨ããæã«ã©ãã
ã使ç¨ããã»ããµã¼ã
å¤æãè¡ãããã»ããµã¼ãæå®ãã¾ãã
* CUDA(使ãããcuDNN) : CUDA(GPU)ã使ã£ã¦å¤æãè¡ãã¾ã(cuDNNã使ããå ´åã¯cuDNNã使ããã¾ã)
* CPU : CPUã®ã¿ã使ã£ã¦å¤æãè¡ãã¾ã
ãåºåãã¡ã¤ã«ãä¸æ¸ãããªãã
ãã®è¨å®ãONã®å ´åãç»åã®æ¸ãè¾¼ã¿å
ã«ååã®ãã¡ã¤ã«ãåå¨ããå ´åã¯å¤æãè¡ãã¾ããã
ãå¼æ°ä»ãèµ·åæè¨å®ã
exeã«å
¥åãã¡ã¤ã«ãå¼æ°ã§ä¸ããå ´åã§ã®åä½ãè¨å®ãã¾ãã
* èµ·åæã«å¤æãã : èµ·åæã«èªåã§å¤æãéå§ãã¾ã
* æåæã«çµäºãã : å¤æçµäºæã«å¤±æãã¦ããªããã°èªåã§çµäºãã¾ã
ã使ç¨GPU Noã
GPUãè¤æ°æããå ´åã«ä½¿ç¨ããããã¤ã¹çªå·ãæå®ã§ãã¾ããCPUã¢ã¼ãæãç¡å¹ãªããã¤ã¹çªå·ãæå®ããå ´åã¯ç¡è¦ããã¾ãã
ãå ¥ååç §æåºå®ãã©ã«ãã
å
¥åã®åç
§ãã¿ã³ãæ¼ããéã«æåã«è¡¨ç¤ºããããã©ã«ããããã§è¨å®ãããã©ã«ãã«åºå®ãã¾ãã
ãåºååç §æåºå®ãã©ã«ãã
å¤æããç»åã®åºåå
ãã©ã«ããããã§è¨å®ãããã©ã«ãã«åºå®ãã¾ãã
ã¾ããåºåã®åç
§ãã¿ã³ãæ¼ããéã«æåã«è¡¨ç¤ºããããã©ã«ããããã§è¨å®ãããã©ã«ãã«åºå®ãã¾ãã
ãã®ä»
ãã®ä»ã®è¨å®é
ç®ç¾¤ã§ãã
ãUIã®è¨èªã
UIã®è¨èªãè¨å®ãã¾ãã
ååèµ·åæã¯PCã®è¨èªè¨å®ã¨åãè¨èªãé¸ã°ãã¾ãã(åå¨ããªãå ´åã¯è±èªã«ãªãã¾ã)
ãcuDNNãã§ãã¯ã
ãcuDNNãã§ãã¯ããã¿ã³ãæ¼ãã¨cuDNNã使ããã調ã¹ããã¨ãåºæ¥ã¾ãã
cuDNNã使ããªãå ´åã¯çç±ã表示ããã¾ãã
ãå®è¡ããã¿ã³ãæ¼ãã¨å¤æãå§ã¾ãã¾ãã éä¸ã§ãã£ã³ã»ã«ãããå ´åã¯ããã£ã³ã»ã«ããã¿ã³ãæ¼ãã¾ãã ãã ããå®éã«åæ¢ããã¾ã§ã¿ã¤ã ã©ã°ãããã¾ãã ããã°ã¬ã¹ãã¼ã¯è¤æ°æã®ç»åãå¤æ´ããéã®é²è¡åº¦åãã示ãã¦ãã¾ãã ãã°ã«æ®ãäºæ³æéã表示ããã¾ãããããã¯åãç¸¦å¹ ãæ¨ªå¹ ã®è¤æ°ãã¡ã¤ã«ãå¦çããã¨ãã®äºæ³ã§ãã ãªã®ã§ããã¡ã¤ã«ã®å¤§ããããã©ãã©ãªå ´åã¯å½¹ã«ç«ã¡ã¾ããããå¦çããç»åã2æ以ä¸ã®æã¯ãä¸æãã¨ãã表示ããã¾ããã
使ãæ¹(CUIç)
ãwaifu2x-caffe-cui.exeãã¯ã³ãã³ãã©ã¤ã³ãã¼ã«ã§ãã
ã³ãã³ãããã³ãã
ãç«ã¡ä¸ãã次ã®ããã«ã³ãã³ããæã¡è¾¼ã¿ã使ç¨ãã¦ä¸ããã
以ä¸ã®ã³ãã³ãã¯ã使ãæ¹ãç»é¢ã«åºåãã¾ãã
waifu2x-caffe-cui.exe --help
以ä¸ã®ã³ãã³ãã¯ãç»åå¤æãå®è¡ããã³ãã³ãã®ä¾ã§ãã
waifu2x-caffe-cui.exe -i mywaifu.png -m noise_scale --scale_ratio 1.6 --noise_level 2
以ä¸ãå®è¡ããã¨ãmywaifu(noise_scale)(Level2)(x1.600000).png
ã«å¤æçµæãä¿åããã¾ãã
ã³ãã³ããªã¹ããåã³ãã³ãã®è©³ç´°ã¯ã³ãã³ãã©ã¤ã³ãªãã·ã§ã³(å ±é)ã¨ã³ãã³ãã©ã¤ã³ãªãã·ã§ã³(CUIç)ã®é ããèªã¿ãã ããã
ã³ãã³ãã©ã¤ã³ãªãã·ã§ã³(å ±é)
æ¬ã½ããã§ã¯ã以ä¸ã®ãªãã·ã§ã³ãæå®ãããã¨ãåºæ¥ã¾ãã GUIçã§ã¯å ¥åãã¡ã¤ã«ä»¥å¤ã®ã³ãã³ãã©ã¤ã³ãªãã·ã§ã³ãæå®ãã¦èµ·åããå ´åãç¾å¨ãªãã·ã§ã³ã®ãã¡ã¤ã«ä¿åãè¡ãã¾ããã ã¾ããGUIçã§ã¯æå®ãããªãã£ããªãã·ã§ã³ã¯ååçµäºæã®ãªãã·ã§ã³ã使ç¨ããã¾ãã
-l <æåå>, --input_extention_list <æåå>
input_fileããã©ã«ãã®å ´åã®ããã©ã«ãå
ã®å¤æããç»åã®æ¡å¼µåãæå®ãã¾ãã
ããã©ã«ãå¤ã¯`png:jpg:jpeg:tif:tiff:bmp:tga`ã§ãã
ã¾ããåºåãæåã¯`:`ã§ãã
ä¾. png:jpg:jpeg:tif:tiff:bmp:tga
-e <æåå>, --output_extention <æåå>
input_fileããã©ã«ãã®å ´åã®ãåºåç»åã®æ¡å¼µåãæå®ãã¾ãã
ããã©ã«ãå¤ã¯`png`ã§ãã
-m <noise|scale|noise_scale>, --mode <noise|scale|noise_scale>
å¤æã¢ã¼ããæå®ãã¾ããæå®ããªãã£ãå ´åã¯`noise_scale`ãé¸æããã¾ãã
* noise : ãã¤ãºé¤å»ãè¡ãã¾ã (æ£ç¢ºã«ã¯ããã¤ãºé¤å»ç¨ã®ã¢ãã«ãç¨ãã¦ç»åå¤æãè¡ãã¾ã)
* scale : æ¡å¤§ãè¡ãã¾ã (æ£ç¢ºã«ã¯ãæ¢åã¢ã«ã´ãªãºã ã§æ¡å¤§ããå¾ã«ãæ¡å¤§ç»åè£å®ç¨ã®ã¢ãã«ãç¨ãã¦ç»åå¤æãè¡ãã¾ã)
* noise_scale : ãã¤ãºé¤å»ã¨æ¡å¤§ãè¡ãã¾ã (ãã¤ãºé¤å»ãè¡ã£ãå¾ã«ãå¼ãç¶ãæ¡å¤§å¦çãè¡ãã¾ã)
* auto_scale : æ¡å¤§ãè¡ãã¾ããå
¥åãJPEGç»åã®å ´åã®ã¿ãã¤ãºé¤å»ãè¡ãã¾ã
-s <å°æ°ç¹ä»ãæ°å¤>, --scale_ratio <å°æ°ç¹ä»ãæ°å¤>
ç»åãä½åã«æ¡å¤§ããããæå®ãã¾ããããã©ã«ãå¤ã¯`2.0`ã§ããã2.0å以å¤ãæå®ã§ãã¾ãã
scale_widthãscale_heightãæå®ãããå ´åããã¡ããåªå
ããã¾ãã
2.0以å¤ã®æ°å¤ãæå®ããã¨ã次ã®ãããªå¦çãè¡ãã¾ãã
* ã¾ããæå®ãããåçãå¿
è¦ååã«ã«ãã¼ããããã«ã2åæ¡å¤§ãç¹°ãè¿ãè¡ãã¾ãã
* 2ã®ç´¯ä¹ä»¥å¤ã®æ°å¤ãæå®ããã¦ããå ´åã¯ãæå®åçã«ãªãããã«æ¡å¤§ããç»åã縮å°ãã¾ãã
-w <æ´æ°>, --scale_width <æ´æ°>
ç»åã®ç¸¦æ¨ªæ¯ãç¶æããã¾ã¾ãæå®ããã横å¹
ã«ãªãããã«æ¡å¤§ãã¾ã(åä½ã¯ãã¯ã»ã«)ã
scale_heightã¨åæã«æå®ããã¨ãæå®ããã縦横å¹
ã«ãªãããã«ç»åãæ¡å¤§ãã¾ãã
-h <æ´æ°>, --scale_height <æ´æ°>
ç»åã®ç¸¦æ¨ªæ¯ãç¶æããã¾ã¾ãæå®ããã縦å¹
ã«ãªãããã«æ¡å¤§ãã¾ã(åä½ã¯ãã¯ã»ã«)ã
scale_widthã¨åæã«æå®ããã¨ãæå®ããã縦横å¹
ã«ãªãããã«ç»åãæ¡å¤§ãã¾ãã
-n <0|1|2|3>, --noise_level <0|1|2|3>
ãã¤ãºé¤å»ã¬ãã«ãæå®ãã¾ãããã¤ãºé¤å»ç¨ã®ã¢ãã«ã¯ã¬ãã«0ï½3ã®ã¿ç¨æããã¦ããã®ã§ã
0 ã 1 ã 2 ã 3 ãæå®ãã¦ä¸ããã
ããã©ã«ãå¤ã¯`0`ã§ãã
-p <cpu|gpu|cudnn>, --process <cpu|gpu|cudnn>
å¦çã«ä½¿ãããã»ããµã¼ãæå®ãã¾ããããã©ã«ãå¤ã¯`gpu`ã§ãã
* cpu : CPUã使ã£ã¦å¤æãè¡ãã¾ãã
* gpu : CUDA(GPU)ã使ã£ã¦å¤æãè¡ãã¾ããWindowsçã§ã®ã¿ãcuDNNã使ãããªãcuDNNã使ãã¾ãã
* cudnn : cuDNNã使ã£ã¦å¤æãè¡ãã¾ãã
-c <æ´æ°>, --crop_size <æ´æ°>
åå²ãµã¤ãºãæå®ãã¾ããããã©ã«ãå¤ã¯`128`ã§ãã
-q <æ´æ°>, --output_quality <æ´æ°>
å¤æå¾ã®ç»åã®ç»è³ªãè¨å®ãã¾ããããã©ã«ãå¤ã¯`-1`ã§ã
æå®ã§ããå¤ã¨æå³ã¯ãåºåæ¡å¼µåãã§è¨å®ããå½¢å¼ã«ããç°ãªãã¾ãã
-1ã®å ´åã¯ãåç»åå½¢å¼ã®ããã©ã«ãå¤ã使ããã¾ãã
-d <æ´æ°>, --output_depth <æ´æ°>
å¤æå¾ã®ç»åã®1ãã£ã³ãã«ãããã®ãããæ°ãæå®ãã¾ããããã©ã«ãå¤ã¯`8`ã§ãã
æå®ã§ããå¤ã¯ãåºåæ¡å¼µåãã§è¨å®ããå½¢å¼ã«ããç°ãªãã¾ãã
-b <æ´æ°>, --batch_size <æ´æ°>
mini-batchãµã¤ãºãæå®ãã¾ããããã©ã«ãå¤ã¯`1`ã§ãã
mini-batchãµã¤ãºã¯ç»åããåå²ãµã¤ãºãã§åå²ãããããã¯ãåæã«å¦çããæ°ã®ãã¨ã§ããä¾ãã°`2`ãæå®ããå ´åã2ãããã¯ãã¤å¤æãã¦ããã¾ãã
mini-batchãµã¤ãºã大ããããã¨åå²ãµã¤ãºã大ããããã¨ã®åæ§ã«GPUã®ä½¿ç¨çãé«ããªãã¾ãããè¨æ¸¬ããæãã ã¨åå²ãµã¤ãºã大ããããæ¹ãå¹æãé«ãã§ãã
(ä¾ãã°åå²ãµã¤ãºã`64`ãmini-batchãµã¤ãºã`4`ã«ãããããåå²ãµã¤ãºã`128`ãmini-batchãµã¤ãºã`1`ã«ããæ¹ãå¦çãéãçµãã)
--gpu <æ´æ°>
å¦çã«ä½¿ãGPUããã¤ã¹çªå·ãæå®ãã¾ããããã©ã«ãå¤ã¯`0`ã§ãã
GPUããã¤ã¹çªå·ã¯0ããå§ã¾ããã¨ã«æ³¨æãã¦ãã ããã
å¦çã«GPUã使ããªãå ´åã¯ç¡è¦ããã¾ãã
ã¾ããåå¨ããªãGPUããã¤ã¹çªå·ãæå®ããå ´åã¯ããã©ã«ãã®GPUã§å®è¡ããã¾ãã
-t <0|1>, --tta <0|1>
`1`ãæå®ããã¨TTAã¢ã¼ãã使ç¨ãã¾ããããã©ã«ãå¤ã¯`0`ã§ãã
--, --ignore_rest
ãã®ãªãã·ã§ã³ãæå®ãããå¾ã®å
¨ã¦ã®ãªãã·ã§ã³ãç¡è¦ãã¾ãã
ã¹ã¯ãªããã»ããããã¡ã¤ã«ç¨ã§ãã
ã³ãã³ãã©ã¤ã³ãªãã·ã§ã³(GUIç)
GUIçã§ã¯ãªãã·ã§ã³æå®ã«å½ã¦ã¯ã¾ããªãã£ãå¼æ°ã¯å ¥åãã¡ã¤ã«ã¨ãã¦èªèããã¾ãã å ¥åãã¡ã¤ã«ã¯ãã¡ã¤ã«ããã©ã«ããè¤æ°ããã¡ã¤ã«ã¨ãã©ã«ãåæã«æå®ã§ãã¾ãã
-o , --output_folder
å¤æãããç»åãä¿åãããã©ã«ãã¸ã®ãã¹ãè¨å®ãã¾ãã
æå®ããããã©ã«ãã®ä¸ã«å¤æå¾ã®ãã¡ã¤ã«ãä¿åãã¾ãã
å¤æå¾ã®ãã¡ã¤ã«ã®å½åè¦åã¯GUIã§å
¥åãã¡ã¤ã«ãè¨å®ããæã«èªåã§æ±ºå®ãããåºåãã¡ã¤ã«åã¨åãã§ãã
æå®ãããªãã£ãå ´åãã²ã¨ã¤ç®ã®å
¥åãã¡ã¤ã«ã¨åããã©ã«ãã«ä¿åããã¾ãã
--auto_start <0|1>
`1`ãæå®ããã¨èµ·åæã«èªåã§å¤æãéå§ãã¾ãã
--auto_exit <0|1>
`1`ãæå®ããã¨ãèµ·åæã«èªåã§å¤æããå ´åã«å¤æãæåããã¨èªåã§çµäºãã¾ãã
--no_overwrite <0|1>
`1`ãæå®ããã¨ãç»åã®æ¸ãè¾¼ã¿å
ã«ååã®ãã¡ã¤ã«ãåå¨ããå ´åã¯å¤æãè¡ãã¾ããã
-y <upconv_7_anime_style_art_rgb|upconv_7_photo|anime_style_art_rgb|photo|anime_style_art_y|upresnet10|cunet>, --model_type <upconv_7_anime_style_art_rgb|upconv_7_photo|anime_style_art_rgb|photo|anime_style_art_y|upresnet10|cunet>
使ç¨ããã¢ãã«ãæå®ãã¾ãã
GUIã§ã®è¨å®é
ç®ãã¢ãã«ãã¨ä»¥ä¸ã®ããã«å¯¾å¿ãã¦ãã¾ãã
* upconv_7_anime_style_art_rgb : 2次å
ã¤ã©ã¹ã(UpRGBã¢ãã«)
* upconv_7_photo : åçã»ã¢ãã¡(UpPhotoã¢ãã«)
* anime_style_art_rgb : 2次å
ã¤ã©ã¹ã(RGBã¢ãã«)
* photo : åçã»ã¢ãã¡(Photoã¢ãã«)
* anime_style_art_y : 2次å
ã¤ã©ã¹ã(Yã¢ãã«)
* upresnet10 : 2次å
ã¤ã©ã¹ã(UpResNet10ã¢ãã«)
* cunet : 2次å
ã¤ã©ã¹ã(CUnetã¢ãã«)
ã³ãã³ãã©ã¤ã³ãªãã·ã§ã³(CUIç)
--version
ãã¼ã¸ã§ã³æ
å ±ãåºåããçµäºãã¾ãã
-?, --help
使ãæ¹ã表示ããçµäºãã¾ãã
æ軽ã«ä½¿ãæ¹ã確èªãããæãªã©ã«ã©ããã
-i <æåå>, --input_file <æåå>
(å¿
é ) å¤æããç»åã¸ã®ãã¹
ãã©ã«ããæå®ããå ´åããã®ãã©ã«ã以ä¸ã®ç»åãã¡ã¤ã«ãå
¨ã¦å¤æãã¦output_fileã§æå®ãããã©ã«ãã¸åºåãã¾ãã
-o , --output_file
å¤æãããç»åãä¿åãããã¡ã¤ã«ã¸ã®ãã¹
(input_fileããã©ã«ãã®å ´å)å¤æãããç»åãä¿åãããã©ã«ãã¸ã®ãã¹
(input_fileãç»åãã¡ã¤ã«ã®å ´å)æ¡å¼µå(æå¾ã®.pngãªã©)ã¯å¿
ãå
¥åããããã«ãã¦ä¸ããã
æå®ããªãã£ãå ´åã¯èªåã§ãã¡ã¤ã«åã決å®ãããã®ãã¡ã¤ã«ã«ä¿åãã¾ãã
ãã¡ã¤ã«åã®æ±ºå®ã«ã¼ã«ã¯ã
`[å
ã®ç»åãã¡ã¤ã«å]``(ã¢ãã«å)``(ã¢ã¼ãå)``(ãã¤ãºé¤å»ã¬ãã«(ãã¤ãºé¤å»ã¢ã¼ãã®å ´å))``(æ¡å¤§ç(æ¡å¤§ã¢ã¼ãã®å ´å))``(åºåãããæ°(8ããã以å¤ã®å ´å))``.åºåæ¡å¼µå`
ã®ããã«ãªã£ã¦ãã¾ãã
ä¿åãããå ´æã¯ãåºæ¬çã«ã¯å
¥åç»åã¨åããã£ã¬ã¯ããªã«ãªãã¾ãã
--model_dir <æåå>
ã¢ãã«ãæ ¼ç´ããã¦ãããã£ã¬ã¯ããªã¸ã®ãã¹ãæå®ãã¾ããããã©ã«ãå¤ã¯`models/cunet`ã§ãã
æ¨æºã§ã¯ä»¥ä¸ã®ã¢ãã«ãä»å±ãã¦ãã¾ãã
* `models/anime_style_art_rgb` : 2次å
ã¤ã©ã¹ã(RGBã¢ãã«)
* `models/anime_style_art` : 2次å
ã¤ã©ã¹ã(Yã¢ãã«)
* `models/photo` : åçã»ã¢ãã¡(Photoã¢ãã«)
* `models/upconv_7_anime_style_art_rgb` : 2次å
ã¤ã©ã¹ã(UpRGBã¢ãã«)
* `models/upconv_7_photo` : åçã»ã¢ãã¡(UpPhotoã¢ãã«)
* `models/upresnet10` : 2次å
ã¤ã©ã¹ã(UpResNet10ã¢ãã«)
* `models/cunet` : 2次å
ã¤ã©ã¹ã(CUnetã¢ãã«)
* `models/ukbench` : æ§å¼ã®åçç¨ã¢ãã«(æ¡å¤§ããã¢ãã«ã®ã¿ä»å±ãã¦ãã¾ãããã¤ãºé¤å»ã¯åºæ¥ã¾ãã)
åºæ¬çã«ã¯æå®ããªãã¦ã大ä¸å¤«ã§ããããã©ã«ã以å¤ã®ã¢ãã«ãèªä½ã®ã¢ãã«ã使ç¨ããæãªã©ã«æå®ãã¦ä¸ããã
--crop_w <æ´æ°>
åå²ãµã¤ãº(横å¹
)ãæå®ãã¾ããè¨å®ããªãã£ãå ´åã¯crop_sizeã®å¤ã使ç¨ããã¾ãã
å
¥åããç»åã®æ¨ªå¹
ã®ç´æ°ãæå®ããã¨ããé«éã«å¤æã§ããå¯è½æ§ãããã¾ãã
--crop_h <æ´æ°>
åå²ãµã¤ãº(縦å¹
)ãæå®ãã¾ããè¨å®ããªãã£ãå ´åã¯crop_sizeã®å¤ã使ç¨ããã¾ãã
å
¥åããç»åã®ç¸¦å¹
ã®ç´æ°ãæå®ããã¨ããé«éã«å¤æã§ãã¾å¯è½æ§ãããã¾ãã
åå²ãµã¤ãº
waifu2x-caffe(waifu2xãã§ãã)ã¯ç»åãå¤æããæã ç»åãä¸å®ã®ãµã¤ãºæ¯ã«åå²ãã¦ä¸ã¤ãã¤å¤æãè¡ããæå¾ã«çµåãã¦ä¸æã®ç»åã«æ»ããã¨ããå¦çããã¦ãã¾ãã åå²ãµã¤ãº(crop_size)ã¨ã¯ããã®ç»åãåå²ããéã®å¹ ï¼ãã¯ã»ã«åä½ï¼ã®äºã§ãã
CUDAã§å¤æä¸ã§ãGPUã使ãåãã¦ããªãï¼GPUã®ä½¿ç¨çã100%è¿ãã¾ã§ãã£ã¦ããªãï¼å ´åã ãã®æ°å¤ã大ãããããã¨ã§å¦çãæ©ãçµããå¯è½æ§ãããã¾ããï¼GPUã使ãåããã¨ãåºæ¥ãæ§ã«ãªãããï¼ GPU-Zãªã©ã§GPU Load(GPU使ç¨ç)ã¨Memory Used(VRAM使ç¨ç)ãè¦ãªãã調ç¯ãã¦ã¿ã¦ä¸ããã ã¾ãã以ä¸ã®æ§ãªç¹æ§ãããã®ã§åèã«ãã¦ä¸ããã
- å¿ ãããæ°å¤ã大ãããã°å¤§ããã»ã©éããªãããã§ã¯ç¡ã
- åå²ãµã¤ãºãç»åã®ç¸¦æ¨ªãµã¤ãºã®ç´æ°ï¼ãããã¯å²ã£ãã¨ãã«ä½ããå°ãªãæ°ï¼ã ã¨ãç¡é§ã«æ¼ç®ããéãæ¸ã£ã¦éããªãã (ãã®æ¡ä»¶ã«ãã¾ãå½ã¦ã¯ã¾ããªãæ°å¤ãæéã«ãªãã±ã¼ã¹ããããããï¼)
- æ°å¤ã2åã«ããå ´åãçè«ä¸ã¯ä½¿ç¨ããã¡ã¢ãªéã¯4åã«ãªã(å®éã¯3ï½4åã¨ãã£ãã¨ãã)ã®ã§ã½ãããè½ã¡ãªãããã«æ³¨æãç¹ã«CUDAã¯cuDNNã«æ¯ã¹ã¦ã¡ã¢ãªã®æ¶è²»éãã¨ã¦ãå¤ãã®ã§æ°ãã¤ãããã¨
ã¢ã«ãã¡ãã£ã³ãã«ä»ãç»åã«ã¤ãã¦
æ¬ã½ããã§ã¯ã¢ã«ãã¡ãã£ã³ãã«ä»ãç»åã®æ¡å¤§ã対å¿ãã¦ãã¾ãã ããããã¢ã«ãã¡ãã£ã³ãã«ãåä½ã§æ¡å¤§ããå¦çã«ãªã£ã¦ãããããã¢ã«ãã¡ãã£ã³ãã«ä»ãç»åã®æ¡å¤§ã¯ç¡ãå ´åã¨æ¯ã¹ã¦ããã2åã®æéããããã®ã§æ³¨æãã¦ãã ããã ãã ããã¢ã«ãã¡ãã£ã³ãã«ãåè²ã§æ§æããã¦ããå ´åã¯ãªãã®å ´åã¨ã»ã¼åãæéã§æ¡å¤§ã§ãã¾ãã
The format of language files
Language files format is JSON. If you create new language file, add language setting to 'lang/LangList.txt'. 'lang/LangList.txt' format is TSV(Tab-Separated Values).
- LangName : Language name
- LangID : Primary language See MSDN
- SubLangID : Sublanguage See MSDN
- FileName : Language file name
ex.
- Japanese LangID : 0x11(LANG_JAPANESE), SubLangID : 0x01(SUBLANG_JAPANESE_JAPAN)
- English(US) LangID : 0x09(LANG_ENGLISH), SubLangID : 0x01(SUBLANG_ENGLISH_US)
- English(UK) LangID : 0x09(LANG_ENGLISH), SubLangID : 0x02(SUBLANG_ENGLISH_UK)
ããã¨ãã
æ¬ã½ããã¯ç¡ä¿è¨¼ã§ãã å©ç¨è ã®å¤æã®ä¸ã«ä½¿ç¨ãã¦ä¸ããã å¶ä½è ã¯ãããªã義åãè² ããªããã®ã¨ãã¾ãã
è¬è¾
ãªãªã¸ãã«ã®waifu2xãåã³ã¢ãã«ã®å¶ä½ãè¡ããMITã©ã¤ã»ã³ã¹ã®ä¸ã§å
¬éãã¦ä¸ãã£ã ultraistããã
ãªãªã¸ãã«ã®waifu2xãå
ã«waifu2x-converterãä½æãã¦ä¸ãã£ã ã¢ãã¼ã´ãã(READMEãLICENSE.txtã®æ¸ãæ¹ãOpenCVã®ä½¿ãæ¹çããªãåèã«ããã¦ããã ãã¾ãã)
waifu2x-chainerãä½æãã¦ãªãªã¸ãã«ã®ã¢ãã«ã®å¶ä½ãè¡ããMITã©ã¤ã»ã³ã¹ã®ä¸ã§å
¬éãã¦ä¸ãã£ãtsurumesoãã
ã«ãæè¬ãã¾ãã
ã¾ããã¡ãã»ã¼ã¸ãè±è¨³ãã¦ãã ãã£ã @paul70078 ãããã¡ãã»ã¼ã¸ãä¸å½èª(ç°¡ä½å)ã«ç¿»è¨³ãã¦ãã ãã£ã @yoonhakcher ãããä¸å½èª(ç°¡ä½å)訳ã®ãã«ãªã¯ã¨ã¹ããä¸ãã£ã @mzhboy ããã
ã¡ãã»ã¼ã¸ãéå½èªã«ç¿»è¨³ãã¦ãã ãã£ã @kenin0726 ãããéå½èªè¨³ã®æ¹åãææ¡ãã¦ãã ãã£ã @aruhirin ããã
ã¡ãã»ã¼ã¸ãä¸å½èª(ç¹ä½å)ã«ç¿»è¨³ãã¦ãã ãã£ã @lizardon1995 ããã@yoonhakcher ããããã«ã³èªè¨³ã®ãã«ãªã¯ã¨ã¹ããä¸ãã£ã @Scharynche ããããã©ã³ã¹èªè¨³ã®ãã«ãªã¯ã¨ã¹ããä¸ãã£ã @Serized ããããã©ã¸ã«ãã«ãã¬ã«èªè¨³ã®ãã«ãªã¯ã¨ã¹ããä¸ãã£ã @simrafael ããããã¤ãèªè¨³ã®ãã«ãªã¯ã¨ã¹ããä¸ãã£ã @AndreasWebdev ããããã§ã³èªè¨³ã®ãã«ãªã¯ã¨ã¹ããä¸ãã£ã @07pepa ãããã¤ã¿ãªã¢èªè¨³ã®ãã«ãªã¯ã¨ã¹ããä¸ãã£ã @AndryGabry01 ããã
GUIçã®ã¢ã¤ã³ã³ãæä¾ãã¦ãã ãã£ã JYUNYAãã ã«æè¬ãã¾ãã
Top Related Projects
Image Super-Resolution for Anime-Style Art
Video, Image and GIF upscale/enlarge(Super-Resolution) and Video frame interpolation. Achieved with Waifu2x, Real-ESRGAN, Real-CUGAN, RTX Video Super Resolution VSR, SRMD, RealSR, Anime4K, RIFE, IFRNet, CAIN, DAIN, and ACNet.
waifu2x converter ncnn version, runs fast on intel / amd / nvidia / apple-silicon GPU with vulkan
Convert designs to code with AI
Introducing Visual Copilot: A new AI model to turn Figma designs to high quality code using your components.
Try Visual Copilot