When training the model, it stops at a level far below a given epoch.

[Choikyungho9]

When training data, the trained model is created as the training ends without filling the set epoch.

If you look at the attached picture, the total epoch is 100, and it stops at the 10th training.

How do I fill in 100 trainings of my model?

[sctse999]

This is called early stopping. You can configure EARLY_STOP_PATIENCE in myconfig.py to a larger value. However, if the training is not getting better for 5 epoches, you may want to observe if it improves in the latter epoches. If it doesn't, it is probably that the data is not good.

[Ezward]

Early stopping is a feature. I'm closing this as 'as intended'.