Hello everyone,
I am doing semi-sueprvised binary sementatic segmentation, although the accuracy is pretty good, 95% and a dice score of 0.94, after plotting the running training loss and running validation loss, I was curious if the model is overfitting.
The training data is 5100 images with augmentations (rotation, random resized crop, random horizontal flip, color jitter and gaussian blur) and 1400 validation images which are only center cropped.
Both data are normalised the same way.
Thank you!
The loss plots seem kind of weird to me and I don’t know how to interpret them.
