This fork adds a batch-processing feature to the command-line interface (CLI) usage. See the added batch_convert_waifu2x.py file. Please scroll down to the "LIA'S FORK SPECIFIC INSTRUCTIONS" for more information. Feel free to add to Issues with any requests, I'll try my best to fulfill them.

Waifu2x port to macOS, also in Core ML and Metal. For other details please refer to waifu2x-ios.

I can't believe this really happened:

My update to the "waifu magnifier" on macOS just got rejected. I admit that it's my fault that I never appealed the so-called "Waifu2x" app which pirated my code in this repo to Apple.

I was still updating this project since that person pirated my code back in 2018. All of this suddenly becomes a joke now.

I am extremely upset about all of this. So I decided to completely abandon this project.

You can still download past releases and send PRs for improvements. But don't expect any update from me in the future.

Update (2021-10-28): The developer of "WaiFu2x" responded to my complaint through Apple. They have removed their App from the App Store and released the name. After calming down I have re-considered the future of this project. It was a huge mistake that I chose MIT as license. I'll still work on this in the future, but all newer changes will be released under a more strict license.

The app waifu2x on the MAS is based on waifu2x-ios instead of this project. It runs through Mac Catalyst.

I haven't published an app based on this project to the Mac App Store. If any app appears on the store that looks like this project, it has nothing to do with me.

• macOS 10.13+ (Running)
• Xcode 11.1+ (Building)

Note that all pre-built binaries are not signed. You can either sign them yourself or go to system settings to bypass the signature check.

Go to Releases to get pre-built binaries. Not all release versions contain both GUI and CLI.

Go to Actions to fetch the latest build for both GUI and CLI. They will be removed automatically 90 days after being built.

If you want to install with homebrew:

brew install imxieyi/waifu2x/waifu2x

It downloads pre-built binary from releases. Xcode is not needed.

1. Open the Terminal (⌘+Space, "terminal")
2. Drag build.sh from Finder to the Terminal window, and press Return to start building the app.
   All missing dependencies (including Xcode) will be installed automatically by the script if needed.
3. Once the build has completed, the waifu2x-mac-app application can be found in the build folder.

1. Build using waifu2x-mac-app scheme
2. To locate the built macOS app, expand the Products folder on Project Navigator (left pane) and right click on waifu2x-mac-app.app to select Show in Finder

The app can be dragged to any location you choose, such as /Applications.

If you would like to use the CLI version, right click on the app and select Show Package Contents. Navigate to Contents/MacOS. The CLI version is waifu2x.

If you would like to run the program anywhere, you must create a symbolic link by typing ln -s /path/to/waifu2x /usr/local/bin/waifu2x in a terminal. You can also drag the waifu2x executable after ln -s to get the file path in terminal automatically.

For example, if waifu2x-mac-app is in /Applications, you would run the following command to create a symlink:
ln -s /Applications/waifu2x-mac-app.app/Contents/MacOS/waifu2x /usr/local/bin/waifu2x

N.B.: You can not drag the CLI executable out and use it directly as it will not work. You must create a symbolic link as shown above if you want to use it without going into the waifu2x-mac-app.app directory.
Additionally the symbolic link will break if you move the macOS app. You can delete the old symlink with unlink /usr/local/bin/waifu2x and run ln -s again to create a new one.

Usage: waifu2x [options]
    -t, --type:
        Image type - a for anime (default), p for photo
    -s, --scale:
        Scale factor (1 or 2)
    -n, --noise:
        Denoise level (0-4)
    -i, --input:
        Input image file (any format as long as NSImage loads)
    -o, --output:
        Output image file (png)
    -h, --help:
        Print usage

WARNING: The CLI version is not a self-contained executable. waifu2x must remain in the same directory as CommandLineKit.framework and waifu2x_mac.framework.

1. Download the Script: Clone this repository or download the script batch_convert_waifu2x.py to your local machine.

2. Make the Script Executable: On Unix-based systems (Linux/macOS), open a terminal and run: chmod +x /path/to/batch_convert_waifu2x.py (This step is not necessary on Windows).

3. Create a Symlink (Optional, Unix-based systems): To run the script from anywhere, create a symlink in /usr/local/bin: ln -s /path/to/batch_convert_waifu2x.py /usr/local/bin/w2x (Replace /path/to/batch_convert_waifu2x.py with the actual path to the script).

Run the script from a terminal or command prompt. Here's the basic syntax: python3 /path/to/batch_convert_waifu2x.py [source_directory] [options] Or, if you created a symlink: w2x [source_directory] [options]

[source_directory]: The path to the directory containing the images you want to process.

[destination_directory] (Optional): Specify where you want to save the processed images. If not specified, the script will create a new directory adjacent to the source directory, appending -waifu-ed to its name.

--type [a|p]: Specify the image type (a for anime, p for photo). Default is a.

--scale [1|2]: Scale factor (1 for no scaling, 2 for doubling the size). Default is 1.

--noise [0-4]: Denoise level (0 for none, up to 4 for maximum denoising). Default is 1.

Upscale and denoise images in /path/to/images and save them to /path/to/processed_images:

python3 batch_convert_waifu2x.py /path/to/images /path/to/processed_images --scale 2 --noise 2

^The above assumes you are in the same directory as batch_convert_waifu2x.py. If you're not, either reference the symlink (w2x) or put the entire file path for batch_convert_waifu2x.py.

Remember to replace /path/to/batch_convert_waifu2x.py and /path/to/images (etc) with the actual paths on your system.

Upscale and denoise all .jpg and .png images in the source_images directory, saving the processed images to processed_images with default settings (anime style, scale factor 1, noise level 1).

`w2x /path/to/source_images /path/to/processed_images`

Or just:

`w2x /path/to/processed_images`

^For a quick run with the default settings, just specify the source directory. This is the simplest way to use the script.

Process images for a photo (not anime) with a high level of noise reduction but without upscaling. This example is useful for cleaning up noisy photographs.

`w2x /path/to/photos --type p --scale 1 --noise 4`

To maximize processing speed, specify the maximum number of worker threads. This can significantly speed up processing time on multi-core systems. Beware of resource usage! Default max workers is set to 4. The more workers the faster the processing speed (more images processed simulatenously), but the more resource-intensive it is.

`w2x /path/to/large_collection /path/to/output --max_workers 8`

Make sure to remove the extra ` notations at the beginning and end of each command.