AlUlkesh/stable-diffusion-webui-images-browser

Feature request, formatted metadata

Wprussell opened this issue · 8 comments

Wprussell commented

It's really difficult to discern details through the current implementation. All text is jumbled together and in one big string. Can we break into separate parts that are clearly marked? For instance, checkpoint, prompt, etc

Thank you

AlUlkesh commented

Here's a first version. I'm sure it can be done prettier, but I hope this helps already.

You can now switch between the previous display and a formatted version:
image

There's also a new setting to set the initial display.

Wprussell commented
Wprussell commented

That's the ticket. But yes, It could be formatted and ordered a little better. Might even enable in .css or Format.txt etc, if it hasn't already been so the user can override the format.

Might I suggest,

Model:
VAE:
Prompt:
Neg Prompt:
Sampler:
Schedule:
Steps:
CFG:
Seed:
Etc....

I pretty sure we are all organizing our work now by Checkpoint, Its like an ocean out there. And Additionally, If we have something like the above (with the inclusion of the prompts) We don't even need to see the jumble)

On a personal note, I really appreciate the your dedication and the turn around.

AlUlkesh commented

Thanks.

I added a small sort function, these are displayed first now:

img_file_info_order = ["Model", "VAE", "Sampler", "Schedule", "Steps", "CFG scale", "Seed", "Size"]

When I have some more time, I could make this also customizable.

Wprussell commented
Wprussell commented
AlUlkesh commented

Ok, I have now switched to a Gradio component that can display tables and made the sort order customizable:

image

image

Wprussell commented