I have a MaskRCNN model for instance segmentation with Resnet 50 - FPN backbone trained in detectron2. And I want to extract the embedding/feature vectors for visualizing input and hopefully detecting outliers. Which place would be the best to extract the embedding vector? I tried p5 of FPN whose shape is (1,256,51,68), but it's a tensor should I flatten this and feed it into the visualizer?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.