We learn sycl and machine learning algorithms from scratch together!
You need git lfs if you want to work with the datasets provided.
This software is clearly dependent on a working sycl compiler.
Please install one, such as AdaptiveCpp or intel oneAPI DPC++.
Then copy CMakeUserPresets.json.example to CMakeUserPresets.json and set the compiler to the correct path.
Now you can build using:
cmake --preset sycl-BUILD_TYPE
cmake --build --preset sycl-BUILD_TYPE
where BUILD_TYPE is one of debug and release.
As the GUI applications build raylib, you can opt out of building them by setting -DJMYML_BUILD_GUI=OFF.
GUI interface visualizing any idx file of rank at least 2, i.e. a collection of images in uncompressed grayscale format. This is very useful for example to visualize the MNIST dataset located in this repository under data/mnist.
LEFT/RIGHT arrow keys page through more images.
GUI interface to draw numbers with your mouse, which then are detected by a machine learning model implemented with this library.