"use the same prompt for each image in a batch" doesn't work as intended
diregoblin opened this issue · 6 comments
Thanks for the prompt collections! I've played with them a bit, and it seems when I uncheck the option in question in the WebUI, it only works for the 1st batch.
I ran batches of four, and in the 1st batch images 1,2,3 had different prompts; image 1.4 had a same one as 1.3. In the 2nd batch image 2.1 was similar to 1.4 and 1.3; 2.2, 2.3, 2.4, 3.1 had a shared prompt; 3.2, 3.3, 3.4 shared a prompt as well. From batch 4 onwards the prompt was shared by the entire batch.
Yup. Batches don't work in Umi AI properly. That's why we don't use 'em. Unless you know python/how to fix it, it's unlikely to get fixed. Both of my python guys said it's an issue with WebUI's limitations.
Correction: Batches don't work with the negative prompts in Wildcards, or certain wildcard parameters. This isn't limited to Umi. It's in the other wildcard systems too.
I'm not a coder, so I can't explain it well. If you'd like, you can jump on the discord and we can talk about it there! I have my other two coders who might be able to help explain.
I do believe one of my coder guys solved the issue! Try again?
Umi enables defining negative prompts in positive prompts by wrapping text between **neg prompt**. This does not work with batches because the txt2img script accepts an array of prompts for a batch but only a single string for the negative prompt. Earlier versions of txt2img did support arrays of negative prompts but it was dropped when token weights were introduced.
Reimplementing it is not really needed because Umi really uses neg prompts only for a couple presets which still work more or less fine without them. The fix in #9 ignores negative prompts defined with ** when a batch is rendered. The regular negative prompt window still works for batches.
Closing as IMO it's fixed.