Image_netayume_lumina_t2i ignores the prompt

Desktop Feedback Windows
,
1

image_netayume_lumina_t2i always generate using the same default prompt.

How to replicate:

  1. Go to template browser
  2. Choose image_netayume_lumina_t2i template
  3. Run it. It will generate an anime of a girl near trees.
  4. Now change prompt with something entirely different.
  5. The model still generates an anime of a girl near a tree.

Tried to reboot, it doesn’t help.

How can I fix it? Is it a bug? I’m just using the default nodes in the template. I didn’t change anything.

2

It’s been 26 days, so you probably already got this fixed. If not or someone else is looking for the fix, here it is. Both the positive and negative CLIP Text Encode prompts have a hidden hard code that over writes whatever you enter. Click edit on each one and disconnect the Concatenate node output from the Prompt text input. You can also delete the System Prompt, Prompt and Concatenate nodes.

I think it was hard coded for demo purposes.

3

This topic was automatically closed 5 hours after the last reply. New replies are no longer allowed.