search

thundaga / SD-webui-txt2img-script

Stable diffusion webui script to batch process images within the txt2img section
38 stars 5 forks source link

Bug? Images losing prompt input #27

Open mattstobber opened 23 hours ago

mattstobber commented 23 hours ago

I have a batch of images in a folder and randomly some of the images will be a blank gray output because it isn't taking the positive/negative prompts from the input. If I drag the original image to PNG Info, it has the prompt so I know the data exists there. If I remove the Prompt and Negative Prompt from the "Assign from input image" section then it works fine(since its using the seed from the image). This means that in order to use this script properly I basically have to rely on not changing the prompt in the ui. This is still a MASSIVE time saver and I love this script, but this bug makes it slightly less usable. Thank you for any help.

00021

OPCOXE commented 23 hours ago

We need the metadata of the image that you use with the script not the one that is generated using it