I am working on Swook GitHub code for GazeML and cannot do inference for a single image to find gaze vector, please help @parai