I'm working on an object detection task using Keras RetinaNet. I would appreciate guidance on how to compute F1 score, precision, and recall for my RetinaNet model. Specifically, I'm interested in understanding how to leverage the model's predicted bounding boxes, class labels, and confidence scores to calculate these key evaluation metrics.
I have already made progress in training and predicting with the RetinaNet model, but I'm now looking to assess its performance more comprehensively. Any insights or code snippets that demonstrate the calculation of these metrics within the context of RetinaNet would be extremely helpful.
also i'm using google colab and tensorboared Thank you in advance for your assistance!