Since the documentation for setting up custom active learning models is pretty sparse. I'm creating this issue to find help in setting up a Keras Retina-Net-like(https://github.com/fizyr/keras-retinanet) model in VoTT.
This model requires some pre and post-processing of the inputs and outputs. My Tensorflow Js model does not do that. I pre-process the image outside the model, pass it in, get some predictions, post-process them, and produce the final output.
Is there even a way of embedding such a model in VoTT?
Any help will be greatly appreciated.
I just want to use predictions from Retina-Net to ease the burden of creating a bunch of labels from scratch.
Since the documentation for setting up custom active learning models is pretty sparse. I'm creating this issue to find help in setting up a Keras Retina-Net-like(https://github.com/fizyr/keras-retinanet) model in VoTT.
This model requires some pre and post-processing of the inputs and outputs. My Tensorflow Js model does not do that. I pre-process the image outside the model, pass it in, get some predictions, post-process them, and produce the final output.
Is there even a way of embedding such a model in VoTT?
Any help will be greatly appreciated.
I just want to use predictions from Retina-Net to ease the burden of creating a bunch of labels from scratch.