Replies: 4 comments 1 reply
-
For 2 subjects, double the number of unet steps, keep the text_encoder below 500 |
Beta Was this translation helpful? Give feedback.
0 replies
-
For me 2 persons always mess.. even if the prompt is very unique. So I still prefer to use just one person trained per model. |
Beta Was this translation helpful? Give feedback.
1 reply
-
您的来信已收到,谢谢。。。。
|
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
The current dreambooth suggests 650 steps for 10 images. If I'm adding more than one subject, should that be increased? IE, if I have two subjects would the number of training steps be increased to 1300?
Also, I can't find this, but the original suggestion was that you should have full body shots among the 30 images. Is this still the case?
Thank you!
Beta Was this translation helpful? Give feedback.
All reactions