where is that old attribute JointLoss, what's the name of it now?

[lucas2606-rs]

🐛 Bug

To Reproduce

Steps to reproduce the behavior:

1. 1. 1.

Expected behavior

Environment

• Pytorch-toolbelt version (e.g., 0.4.4):
• Pytorch version (e.g., 1.8.1):
• Python version (e.g., 3.7):
• OS (e.g., Linux):
• Any other relevant information:

Additional context

[BloodAxe]

It was removed in recent release. The reasoning behind this action is the following: Combining multiple losses in a hard coded way as JointLoss suggested is a wrong design and what you should be doing instead is to leverage some configuration based system to define what losses you want to have, their weights and what input/output keys should go in. At the moment of writing, JointLoss looked like a good idea, however after years of building pytorch proficiency it became clear to me that this class is rather a shortcut than a long term solution. So it was dropped. If you really want to have it I suggest you to copy-paste it's implementation from the past releases.