I’m currently working on a Mask-RCNN project. I’ve been following the TorchVision Object Detection Finetuning Tutorial and it helped me a lot.
But I’d like to be able to get some evaluation metrics in order to compare various models. Things like training and validation losses, Precision, Recall, F1 curve, mAP, losses, and confusion matrix. I haven’t been able to find a neat implementation. I’d like to be able to store information somewhere (like a csv file) and do some plotting.
So far I haven’t been able to find many tutorials explaining this part. Could anyone point me to a project or an example that does this?