Software Engineering Courses (SECourses)•3y ago

Yeah I think thats what Fast Stable Diffusion by TheLastBen does, and thats what RunPod use in their

Yeah I think thats what Fast Stable Diffusion by TheLastBen does, and thats what RunPod use in their ridiculously expensive dreambooth API endpoint.

Aashley Yeah I think thats what Fast Stable Diffusion by TheLastBen does, and thats what...

Furkan Gözükara SECourses•6/6/23, 4:19 PM

do you have any idea regarding this : https://github.com/Mikubill/sd-webui-controlnet/issues/1573

GitHub

FileNotFoundError: [Errno 2] - works but result not displayed in gr...

Is there an existing issue for this? I have searched the existing issues and checked the recent builds/commits of both this extension and the webui What happened? Just made a fresh installation The...

Furkan Gözükara SECourses•6/6/23, 4:19 PM

this happened on a patreon supporter

Furkan Gözükara SECourses•6/6/23, 4:19 PM

i just did a fresh install

Furkan Gözükara SECourses•6/6/23, 4:19 PM

definitely some caching issue

Furkan Gözükara SECourses•6/6/23, 4:27 PM

solution found

Furkan Gözükara SECourses•6/6/23, 4:27 PM

playereckksdee•6/6/23, 4:29 PM

Anyone have kohya ss settings that give similar output to dreambooth?

FFurkan Gözükara SECourses solution found

ashleyOP•6/6/23, 4:29 PM

What was the solution?

Furkan Gözükara SECourses•6/6/23, 4:35 PM

https://github.com/Mikubill/sd-webui-controlnet/issues/1563

GitHub

[Bug]: FileNotFoundError: [Errno 2]... tmp2514rfci.png' · Issue #15...

Is there an existing issue for this? I have searched the existing issues and checked the recent builds/commits of both this extension and the webui What happened? Hello! since last update(today) go...

Furkan Gözükara SECourses•6/6/23, 4:35 PM

Just running any generated image with any model using controlnet and no embeddings showed this. I actually just fixed it by adding a gradio folder to my temp location. Not sure why it wasnt automatically generating this.

C:\Users"Username"\AppData\Local\Temp

Furkan Gözükara SECourses•6/6/23, 4:48 PM

this realism amazing

Furkan Gözükara SECourses•6/6/23, 4:48 PM

even my wife said

Furkan Gözükara SECourses•6/6/23, 4:48 PM

FFurkan Gözükara SECourses Click to see attachment

playereckksdee•6/6/23, 4:50 PM

kohya or dreambooth?

Pplayereckksdee kohya or dreambooth?

Furkan Gözükara SECourses•6/6/23, 4:50 PM

dreambooth of web ui

Furkan Gözükara SECourses•6/6/23, 4:50 PM

hopefully will make a video

Furkan Gözükara SECourses•6/6/23, 4:50 PM

i am processing school lecture

FFurkan Gözükara SECourses this realism amazing

felixsanz•6/6/23, 4:52 PM

what about flexibility?

felixsanz•6/6/23, 4:53 PM

can you share here the first 5 results of "ohwx man wearing a police costume, looking at camera" ?

felixsanz•6/6/23, 4:53 PM

not cherry-picking

Ffelixsanz what about flexibility?

Furkan Gözükara SECourses•6/6/23, 4:54 PM

flexbility bad

felixsanz•6/6/23, 4:54 PM

haha

Furkan Gözükara SECourses•6/6/23, 4:55 PM

this is for realism

Furkan Gözükara SECourses•6/6/23, 4:55 PM

i think it is bad

Furkan Gözükara SECourses•6/6/23, 4:55 PM

because of the resolution

Furkan Gözükara SECourses•6/6/23, 4:55 PM

because underlying model dont have info

Furkan Gözükara SECourses•6/6/23, 4:55 PM

in that resoluion

felixsanz•6/6/23, 4:55 PM

flexibility is important

felixsanz•6/6/23, 4:55 PM

you can't prompt at all if no flexibility

Furkan Gözükara SECourses•6/6/23, 4:55 PM

diferent purposes

felixsanz•6/6/23, 4:55 PM

yup

Furkan Gözükara SECourses•6/6/23, 4:55 PM

if you need shot for linked in

Furkan Gözükara SECourses•6/6/23, 4:55 PM

then this is it

felixsanz•6/6/23, 4:56 PM

depends

felixsanz•6/6/23, 4:56 PM

if you want a front facing picture looking at camera

felixsanz•6/6/23, 4:56 PM

how do you generate it?

felixsanz•6/6/23, 4:56 PM

haha

Furkan Gözükara SECourses•6/6/23, 4:57 PM

Furkan Gözükara SECourses•6/6/23, 4:57 PM

making some shots right now

Furkan Gözükara SECourses•6/6/23, 4:58 PM

will sort with AI

felixsanz•6/6/23, 4:58 PM

some pictures look like your face is photoshopped inside the suit

felixsanz•6/6/23, 4:58 PM

but yeah, good quality

Le•6/6/23, 5:02 PM

I noticed you got full body results from just using your face to train the LoRA. Do I need full body pictures to have a LoRA that respects prompt?

Original message was deleted

ashleyOP•6/6/23, 5:02 PM

Don't you have to adjust the width and height from 512x512 for this?

Ffelixsanz some pictures look like your face is photoshopped inside the suit

Furkan Gözükara SECourses•6/6/23, 5:02 PM

i will use AI to sort

Aashley Don't you have to adjust the width and height from 512x512 for this?

Le•6/6/23, 5:03 PM

it is 512x513 for a missclick. But resolution doesn't impact anything. As you can see, I have full body pictures using the same resolution. If I use any textual inversion, I also get good results.

Le•6/6/23, 5:03 PM

No results on my LoRA

Original message was deleted

Furkan Gözükara SECourses•6/6/23, 5:03 PM

for distant shots you need either higher resolution

Le•6/6/23, 5:03 PM

I got TONS of face pictures using (((full body:1.3))) on the prompt

Furkan Gözükara SECourses•6/6/23, 5:03 PM

or fixing with inpaint

Yeah I think thats what Fast Stable Diffusion by TheLastBen does, and thats what RunPod use in their

Similar Threads

Similar Threads

Similar Threads