Hey, I am working on gaze estimation in Browser environment. While I was able to convert the provided models to TFJS and able to run them in the browser; the inference time is 4s.
Has anyone attempted at optimizing the model.
Or if I were to go about it, how should I do it?
Hey, I am working on gaze estimation in Browser environment. While I was able to convert the provided models to TFJS and able to run them in the browser; the inference time is 4s.
Has anyone attempted at optimizing the model.
Or if I were to go about it, how should I do it?