Convergence issue during model training

[NadiaLiz]

Hello, I am facing issues with the training of my model. Despite several attempts, the loss does not converge. I am using a batch size of 4, and I have noticed that the IoU values and related metrics do not seem to improve. I would appreciate any suggestions to address this issue. Could it be related to the model configuration, hyperparameters, or the quality of the dataset?