search

TheLastBen / fast-stable-diffusion

fast-stable-diffusion + DreamBooth
MIT License
7.43k stars 1.28k forks source link

a simple gradio UI version with just a text box for the prompt #8

Closed GeorvityLabs closed 1 year ago

GeorvityLabs commented 1 year ago

both the notebooks used here , seems to be super complex. to keep things simple , i was wondering that : it would be great if there is a way to link 25% faster technique to a simple gradio ui , implemented within the notebook file itself. with just a text box and submit button. would be dope if someone could implement it with gradio blocks, such that the model is accessed from a local model path

TheLastBen commented 1 year ago

I will soon add a relaxed mode notebook, which will concentrate on the output image rather than the settings, like this one (I tuned for my personal use) : https://i.postimg.cc/6q9v3HZY/Capturerea.jpg

GeorvityLabs commented 1 year ago

that looks interesting @TheLastBen , looking forward to it.

GeorvityLabs commented 1 year ago

any way to strip your current relaxed mode , to make a super basic version , with just a prompt box and submit button , like the official hugginface demo for stable diffusion . or like DALL-E 2 demo , keeping it super clean and interesting to build upon.

TheLastBen commented 1 year ago

But the resolution and the number of steps and the cfg scale are very important in stable diffusion

GeorvityLabs commented 1 year ago

true @TheLastBen , those can be available as text box inputs

GuruVirus commented 1 year ago

Regarding the main topic of a simpler version of the GUI. On mobile devices the slider bars are imprecise and very laggy. I prefer the text box inputs as seen in the Image Creation section of this colab: https://colab.research.google.com/github/pharmapsychotic/ai-notebooks/blob/main/pharmapsychotic_Stable_Diffusion.ipynb

Honestly a simple set of text boxes on PC is way faster too so you can tab thru them. Maybe a checkbox at the top for "input fields instead of sliders" or have both sliders and a text field next to them?

I basically only run 704x704 or 512x896. Though the webUI has been so efficient I have been able to try larger resolutions. 1024x1024 I think went thru and saved, but then it crashed the Generate button (appears to be the same Gradio connector issue as the batch failure problem). Edit I was able to 1600x1600 with the server running on colab thru mobile and gradio on desktop.

AK391 commented 1 year ago

@GuruVirus for mobile you might want to try this gradio gui on huggingface, mininal version with less features: https://huggingface.co/spaces/stabilityai/stable-diffusion