Training All Mask R-CNN layers - Val Loss not improving

I am trying to train a Mask R-CNN to detect power lines but the validation loss keeps fluctuating. I am using a dataset size of ~1300. Any help?

Hi @Paul_Addai

Welcome to the TensorFlow Forum!

Some are the factors which may cause loss fluctuations are like imbalanced dataset, data preprocessing and the used optimizer. Could you please share minimal reproducible code to replicate and understand the issue? Thank you.