I was curious how many changes are required to train SWAG model on a regression problem using the swag.py, would this require many changes or just a change of args to the arg parser? Thank you!
Hi @wrkhard, sorry for the late reply! SWAG should be generally applicable to regression models without significant modifications. In particular, we applied it to a synthetic regression problem here. We also apply it to UCI regression problems here.
izmailovpavel
commented
2 years ago
Hello,
I was curious how many changes are required to train SWAG model on a regression problem using the swag.py, would this require many changes or just a change of args to the arg parser? Thank you!