Sygil-Dev / sygil-webui

Stable Diffusion web UI
GNU Affero General Public License v3.0
7.86k stars 883 forks source link

Feature: Streamlit ImageLab? #1537

Open Anyeos opened 1 year ago

Anyeos commented 1 year ago

What happened?

Are there plans to include a ImageLab tab on StreamLit?

I need to have it because I just generate an image with low specs and if that is what I want then I apply more processing like face fixing and upscale.

So I miss that tab on StreamLit :(

In my opinion having postprocessing is generally a waste of time. I prefer to choose manually the results because I discard a lot, and only a few are useful. So I dont think using more time to postprocess is good. I prefer to save that time and generate more results instead. Then having the ImageLab is good for it, I only will postprocess what really matters.

Version

0.0.1 (Default)

What browsers are you seeing the problem on?

Firefox, Chrome

Where are you running the webui?

Linux

Custom settings

No response

Relevant log output

No response

Code of Conduct

ZeroCool940711 commented 1 year ago

The post-processing tab on the streamlit version is being worked on, might take a bit but it is planned, regarding the need for automatic post-processing I've said it before but some of us, me included, like to have it done automatically and for us a second or two added to the total duration of the generation is a lot better than having to spend 1-2 minutes doing manual post-processing after each generation.

Anyeos commented 1 year ago

The post-processing tab on the streamlit version is being worked on, might take a bit but it is planned, regarding the need for automatic post-processing I've said it before but some of us, me included, like to have it done automatically and for us a second or two added to the total duration of the generation is a lot better than having to spend 1-2 minutes doing manual post-processing after each generation.

Ok, great, but I don't mean to remove the automatic option because that can be useful for other cases.

I explain how I work so you can understand me better. I generate a lot of images that fit some art style, for each 20 images maybe 1 is choosen. Next I will do a second selection but not using the IA generation.

So my work consist of two stages:

1) In first stage I do a fast selection, I take note of the promt and the parameters too. 2) In a second stage I am more selective and I do it not using the IA but viewing the before generated images.

So, I need to save time on the first stage because more generation is better. I generate a lot of images in that stage. At the same time I am selecting or saving the information if I choose one image.

The second stage is not mixed with the first. I first do the first stage and later when I have a lot of images generated I go to the second stage.

And that is basically why I need the ImageLab feature.

In the last stage when I have a few selected images I regenerate each with the same promt and parameteres but with more details. And I do some fixes or scaled up too if needed.

Finally I manually edit some of the images if I need too.

And If I have an own image (painted by myself, rendered from 3D or edited manually from a previous one) I use ImageLab to implement some processings like upscale after image-to-image step. Because the same: not all resulting images from that are acceptable and I repeat generation a lot of times so each second is a save for me in that case.

That is all about my use but I think the automatic option can be useful in other cases. I don't want it to be removed. I only wanted the ImageLab feature to be added.

I apologies If I was not clear at beginning.