Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Flux trainer autocaption inconsistency #211

Open
achuinard opened this issue Nov 5, 2024 · 1 comment
Open

Flux trainer autocaption inconsistency #211

achuinard opened this issue Nov 5, 2024 · 1 comment

Comments

@achuinard
Copy link

achuinard commented Nov 5, 2024 •
edited
Loading

I'm using the Flux trainer on Replicate, but I notice that when I autocaption and require the prefix to be 'a photo of TOK man', only like 75% of the auto-generated captions actually show that.

image

image

Why is this the case? Am I doing something wrong? It is making me want to figure out an alternative solution for captionining.
@xsvfat
Copy link

xsvfat commented Nov 26, 2024 •
edited
Loading

I have the same problem - did you ever figure this out? (I had one training where it only added in the caption prefix to 1 out of the 10 images). (my trigger_word was just "TOK" and autocaption_prefix was photo of a TOK man)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@achuinard @xsvfat