How to train custom model?

Hi, I am shifting from opennmt-tf to opennmt-py and I got some problems training my models.

In opennmt-tf, you can always use a new .py file with a model() function to pass the new defined model instance into the trainer, but I could not found detailed tutorial to train custom model using opennmt-py. Is there any tutorial for this?


You may have a look at this post: How to build my own model using OpenNMT-py?