Skip to content
Discussion options

You must be logged in to vote

Can try again after pulling in the latest changes?

What weighting scheme are you using? Have you tried training with "logit_normal"?

There is no such parameter 'text_input_ids_list' in encode_prompt - script fails if --train_text_encoder is set.

This is being fixed in #8630

Replies: 1 comment 1 reply

Comment options

You must be logged in to vote
1 reply
@r-aristov
Comment options

Answer selected by r-aristov
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Category
Q&A
Labels
None yet
2 participants