KInference

KInference is a library that makes possible execution of complex ML models (written via ONNX) in vanilla Kotlin. ONNX is a popular ecosystem for building, training, evaluating, and exchanging ML and DL models. It makes the process much simpler and divides the model into building blocks that can be switched or tuned to one's liking. However, ONNX carries with itself a lot of dependencies and requirements that complicate its use in some cases. Our library does not require anything but vanilla Kotlin. KInference is lightweight but fast, and supports numerous ONNX operators. This makes it easier to use and is especially useful for various applications that require the models to be run on the users' machines.