how to decrease identity loss, the final result looks too much like 70newface-30%oldface mix no not like swap to fully new face #254

Open
opened 2022-05-03 14:11:58 +02:00 by 2blackbar · 2 comments
2blackbar commented 2022-05-03 14:11:58 +02:00 (Migrated from github.com)
No description provided.
momstouch commented 2022-05-09 06:50:44 +02:00 (Migrated from github.com)

see line 218 in train.py
you can control id-loss using opt.lambda_id

see line 218 in train.py you can control id-loss using opt.lambda_id
Fibonacci134 commented 2022-05-10 06:39:37 +02:00 (Migrated from github.com)

I believe he means with the current model, in which case not much can be done. you can play with the masking and the swapped img a bit if youre familiar with opencv.

I believe he means with the current model, in which case not much can be done. you can play with the masking and the swapped img a bit if youre familiar with opencv.
Sign in to join this conversation.