I would like to compile a Python project with an ML model into a single executable binary. This means I want the model and the weights to be in the binary with the rest of the code. The size of the binary is not of my concern at first. Is this possible? How would I go about it? I am aware of ONNX and Nuitka. From my current understanding, with Nuitka, the resulting binary would load the ONNX models via a path. Continue reading...
