Software Engineering Courses (SECourses)•3y ago

Feel free to use my custom template that has Dreambooth preinstalled and working. Also has Kohya_ss

Feel free to use my custom template that has Dreambooth preinstalled and working. Also has Kohya_ss preinstalled.
https://runpod.io/gsc?template=ya6013lj5a&ref=w18gds2n

Aashley Feel free to use my custom template that has Dreambooth preinstalled and working...

Knechtruprecht•7/14/23, 12:13 PM

thank you for the quick answer, i will try your custom template for my next training

monsieurBoutte•7/14/23, 1:03 PM

Loved the new video you posted on getting SDXL running on run pod

KKnechtruprecht Hi Guys, i am following the current manual installation guide for the dreambooth...

Furkan Gözükara SECourses•7/14/23, 1:03 PM

remove skip install

Furkan Gözükara SECourses•7/14/23, 1:03 PM

restart

Furkan Gözükara SECourses•7/14/23, 1:03 PM

skip install need to be added last

MmonsieurBoutte Loved the new video you posted on getting SDXL running on run pod

Furkan Gözükara SECourses•7/14/23, 1:03 PM

ty. i am also hopefully going to release

Furkan Gözükara SECourses•7/14/23, 1:03 PM

sdxl on runpod colab pc with comfy ui

Furkan Gözükara SECourses•7/14/23, 1:03 PM

prepared files

Furkan Gözükara SECourses•7/14/23, 1:04 PM

even auto runpod comfyui installer posted

Furkan Gözükara SECourses•7/14/23, 1:04 PM

on patreon

monsieurBoutte•7/14/23, 1:05 PM

I’m looking to experiment with stability Ai uncropuncrop at the API level but not seeing it in the docs.

I’m also wondering if image-to-image/masking is a better way of doing what I’m trying to do

I’m wanting to augment my webcam background.
- So I start off with a 16:9
- then I crop in on the photo, keeping the subject in frame
- then uncrop back to 16:9, letting generative fill do its magic

This works well in the uncrop UI

MmonsieurBoutte I’m looking to experiment with stability Ai `uncrop` at the API level but not se...

Furkan Gözükara SECourses•7/14/23, 1:06 PM

interesting idea

Furkan Gözükara SECourses•7/14/23, 1:06 PM

you want this to be real time?=

monsieurBoutte•7/14/23, 1:06 PM

Id love to hear any recommendations on how I could go about doing this

monsieurBoutte•7/14/23, 1:07 PM

Real time as in a POST request to an API and hopefully getting something back within a couple seconds

monsieurBoutte•7/14/23, 1:07 PM

Ideally I’m generating a few at a time to let the user cycle through some options

MmonsieurBoutte Id love to hear any recommendations on how I could go about doing this

Furkan Gözükara SECourses•7/14/23, 1:14 PM

no idea atm

monsieurBoutte•7/14/23, 1:35 PM

word. I'll report back when I figure it out ( :

felixsanz•7/14/23, 1:52 PM

finally I finished the article !!!

felixsanz•7/14/23, 1:53 PM

holy shit. what i'm doing with my life

felixsanz•7/14/23, 2:00 PM

https://www.felixsanz.dev/articles/how-to-create-qr-codes-with-stable-diffusion

How to create QR codes with Stable Diffusion

Create decorative, professional and 100% functional QR codes using Stable Diffusion and ControlNet.

felixsanz•7/14/23, 2:00 PM

tell me that you like it!!

felixsanz•7/14/23, 2:01 PM

i think this method is unique. all qr codes work and i've included a script that automatically check generated qr codes to keep the working ones

felixsanz•7/14/23, 2:01 PM

@ashleyk this one is for you

Ffelixsanz @ashleyk this one is for you 😆

ashleyOP•7/14/23, 2:13 PM

Not really, QR codes are a gimmick that I lost interest in like a month ago already.

ashleyOP•7/14/23, 2:14 PM

I am more interested in using AI for things that are actually useful, like this.

Aashley Not really, QR codes are a gimmick that I lost interest in like a month ago alre...

felixsanz•7/14/23, 2:20 PM

check my article! all codes work

felixsanz•7/14/23, 2:22 PM

felixsanz•7/14/23, 2:22 PM

this is not a gimmick! xD

ashleyOP•7/14/23, 2:26 PM

Sure it's not

felixsanz•7/14/23, 2:28 PM

it's totally neccesary

felixsanz•7/14/23, 2:28 PM

who wants an upscaler to restore moon landing videos when you can have a QR code with a cute puppy dog

Aashley I tried it and it sucks, got very bad results. Kohya_ss and Dreambooth results a...

Khang•7/14/23, 2:38 PM

That's unfortunate

I enjoy exploring new tools, hope they improve it in the near future

KKhang That's unfortunate 🥹 I enjoy exploring new tools, hope they improve it in the ...

ashleyOP•7/14/23, 2:41 PM

Well as @rafraf said, it should work better with massive datasets. (He is actually the one who recommended it to me in the first place, but I guess he had not used it properly yet when he made that recommendation).

Khang•7/14/23, 3:04 PM

@ashleyk is Auto1111 in the template working Ashleyk? I couldn't generate new images this afternoon and I thought that maybe just a bug but this evening I made a new pod and SD still can't generate any thing. It can't even load extension list. I clicked "Generate" or "Load from" button but nothing happens.

KKhang @ashleyk is Auto1111 in the template working Ashleyk? I couldn't generate new im...

ashleyOP•7/14/23, 3:06 PM

Should be, I'll check

Khang•7/14/23, 3:25 PM

@Dr. Furkan Gözükara I found this very interesting extension Dr.Furkan https://github.com/ModelSurge/sd-webui-comfyui
With this we may use Queue prompt and other features in Auto1111

GitHub

GitHub - ModelSurge/sd-webui-comfyui: A1111 extension to embed comf...

A1111 extension to embed comfyui in its own tab. Contribute to ModelSurge/sd-webui-comfyui development by creating an account on GitHub.

KKhang @ashleyk is Auto1111 in the template working Ashleyk? I couldn't generate new im...

ashleyOP•7/14/23, 3:46 PM

Thanks for letting me know, my apologies, should have tested it better, I've reverted the template to use the previous Docker image again.

xlads•7/14/23, 3:52 PM

Hello hello , Sdxl is only for txt to img model correct?

Xxlads Hello hello , Sdxl is only for txt to img model correct?

ashleyOP•7/14/23, 3:57 PM

I would wait until its officially released into A1111 etc personally.

DeadMan1999•7/14/23, 6:17 PM

https://huggingface.co/papers/2307.06949 looks at this guys!

Paper page - HyperDreamBooth: HyperNetworks for Fast Personalizatio...

DDeadMan1999 https://huggingface.co/papers/2307.06949 looks at this guys!

xlads•7/14/23, 6:48 PM

have you done this?

xlads•7/14/23, 6:49 PM

im geting this error OSError: [Errno 122] Disk quota exceeded: 'outputs/txt2img-images/2023-07-14' when trying to use my dreambooth model in txt to img

Xxlads have you done this?

ashleyOP•7/14/23, 6:52 PM

Where is this? RunPod?

xlads•7/14/23, 6:52 PM

i deleted the few imgs i had its working

xlads•7/14/23, 6:52 PM

yes runpod

Xxlads yes runpod

ashleyOP•7/14/23, 6:52 PM

Is it a Network Volume?

ashleyOP•7/14/23, 6:53 PM

Looks like a Network Volume error, you need to increase your Network Volume size if that happens.

xlads•7/14/23, 7:06 PM

im also getting this error when i try to click on an image File Load Error for sample_980-1.png
Unhandled error

xlads•7/14/23, 7:06 PM

inside of juypter on runpod

Feel free to use my custom template that has Dreambooth preinstalled and working. Also has Kohya_ss

Similar Threads

Similar Threads

Similar Threads