I am having trouble understanding the curve val_loss
and loss
in keras after training my model.
Can anyone help me understand it? Also, is my model overfitting or underfitting?
Data Science Stack Exchange is a question and answer site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field. It only takes a minute to sign up.
Sign up to join this communityI am having trouble understanding the curve val_loss
and loss
in keras after training my model.
Can anyone help me understand it? Also, is my model overfitting or underfitting?
The loss curve shows what the model is trying to reduce. The training procedure tried to achieve the lowest loss possible. The loss is calculated using the number of training examples that the models gets right, versus the ones it gets wrong. Or how close it gets to the right answer for regression problems.
The loss curves are going smoothly down, meaning your model improves as it is training, which is good. Your test loss is slightly higher than your training loss, meaning your model is slightly overfitting the training data, but that’s inevitable, it doesn’t seem problematic. All seems ok from this plot.
Now your model is getting an accuracy of 30% or so. Unless you tell us what the model is doing and how you define accuracy, there’s no way of knowing if that’s ok or not.