Software Engineering Courses (SECourses)•3y ago

i can't generate a checkpoint cause I get this error Exception compiling checkpoint: The size of ten

i can't generate a checkpoint cause I get this error Exception compiling checkpoint: The size of tensor a (4) must match the size of tensor b (320) at non-singleton dimension 1

ZZono50 i can't generate a checkpoint cause I get this error Exception compiling checkpo...

Major Epidemic•3/19/23, 10:36 PM

That beyond my level of expertise I am afraid.

Zono50OP•3/19/23, 10:36 PM

so i generate a lora that doesn't look like my model, and then can't convert it to a checkpoint either

Major Epidemic•3/19/23, 10:36 PM

Welcome to the bleeding edge of ai image generation.

Zono50OP•3/19/23, 10:37 PM

lol another frustrating thing is when i click save settings, it doesn't actually save my settings

Major Epidemic•3/19/23, 10:38 PM

I am sure you have done this but have you tried closing the cmd window for automatic, not the web browser, and then starting it again. As long as you don’t touch the web browser until it’s fully loaded you can usually save settings.

Zono50OP•3/19/23, 10:40 PM

i wonder if the issue is using lora extended

Zono50OP•3/19/23, 10:40 PM

if i provide all my own pictures, can i set class images per instance image to 0?

Major Epidemic•3/19/23, 10:41 PM

Ah yes. I don’t mess with that. Also I recommend trying Kohya for Lora. I found it much easier.

Major Epidemic•3/19/23, 10:41 PM

Yes

Zono50OP•3/19/23, 10:41 PM

yeah i tried kohya and that doesn't work on my system either

Zono50OP•3/19/23, 10:41 PM

it won't train a model, gives me an error, it's bugged atm

Major Epidemic•3/19/23, 10:41 PM

Even for Lora?

Zono50OP•3/19/23, 10:41 PM

yup

Zono50OP•3/19/23, 10:41 PM

i'm doing it through linux

Major Epidemic•3/19/23, 10:41 PM

Oooh.

Major Epidemic•3/19/23, 10:41 PM

Ha bleeding edge again.

Zono50OP•3/19/23, 10:45 PM

one of your videos says to apply horizontal flip for variation, then the other says not to for faces

Furkan Gözükara SECourses•3/19/23, 10:53 PM

FFurkan Gözükara SECourses Click to see attachment

Haashim•3/20/23, 12:00 AM

Awesome !

Furkan Gözükara SECourses•3/20/23, 1:02 AM

exporting video. hopefully will be on channel tomorrow

Furkan Gözükara SECourses•3/20/23, 1:02 AM

man spent like 4 days on this :d

Furkan Gözükara SECourses•3/20/23, 1:02 AM

i am going to sleep hopefully see you tomorrow. already 4 am here :/

Zono50OP•3/20/23, 1:07 AM

alrighty, see you in the morning doc

Zono50OP•3/20/23, 1:07 AM

just trained another model and don't even see my model in any of the images, fml

Zono50OP•3/20/23, 1:08 AM

trained it on a bunch of pictures of my female friend, and this is what it puts out

sample_6600-019e6204dcb99ce71783560fe79f0082009dc78de.png

The Laughing Rottie•3/20/23, 1:50 AM

Hi guys, recently a patrean (no status update yet, no hurry Dr. Furkan

I'm trying to Create a Model (Dreambooth), however my folder that was created (model>dreambooth) is empty and only contains a folder names "working" (also empty). So its hard to continue the tutorial videos.

I was able to continue to change settings with this Empty Model Folder and save the settings, change concept and stuff, got all the way of the tutorial where you generate samples, but it was all glitchy images (tried restarting). But I think the overlining issue the empty model folder (strange)

Sorry for the bad english not my first language.

I'm following this guide:
https://www.youtube.com/watch?v=Bdl-jWR3Ukc

I have a laptop with a RTX 3080 with 8GB of vram btw.

Did any of you had this issue?

Many Thanks!

Dincman•3/20/23, 2:47 AM

Hi there, I am in need of assistance. I have been trying to use stable diffusion for a month now, and only accomplished to somehow run a1111 but cannot seem to restart the same setup another day. There's too much info about too many things and I can't seem to connect that information together.
I saw https://www.youtube.com/watch?v=kIyqAdd_i10
and https://www.youtube.com/watch?v=mnCY8uM7E50&list=PL_pbwdIyffsmclLl0O144nQRnezKlNdx3&index=1
these two videos of yours seem very informative but before training dreambooth models I can't yet understand:
-which web ui is the best up to date and which models to use in colab.
-how to install and reboot with those models.
-how to use inpainting, outpainting, control net, upscaling ...etc
-what is lora and how its relevant?
-what other things should I know in what order?
I don't know any coding and even tried to look at some online videos so I could understand the basics colab and git hub, but I am an artist and the technology seem to move faster than I can understand now.
This thing is giving me pain for weeks now. It took me a week to understand that I couldn't run A1111 locally because of low GPU.
I will be happy even if you give me a youtube watch order and won't refuse a little bit more explanation.

YouTubeSECourses - Software Engineering Courses

Stable Diffusion Google Colab, Continue, Directory, Transfer, Clone...

Our Discord : https://discord.gg/HbqgGaZVmr. This is the video where you will learn how to use Google Colab for Stable Diffusion. If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on

https://www.patreon.com/SECourses

Playlist of Stable Diffusion Tutorials, #Automatic1111...

YouTubeSECourses - Software Engineering Courses

Transform Your Selfie into a Stunning AI Avatar with Stable Diffusi...

Our Discord : https://discord.gg/HbqgGaZVmr. How to do free Stable Diffusion DreamBooth training on Google Colab for free. If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on

https://www.patreon.com/SECourses

Playlist of Stable Diffusion Tutorials, Automatic1111 and Goo...

TThe Laughing Rottie Hi guys, recently a patrean (no status update yet, no hurry Dr. Furkan 😉 I'm ...

Furkan Gözükara SECourses•3/20/23, 10:11 AM

hi welcome

Furkan Gözükara SECourses•3/20/23, 10:11 AM

you need to add this to the command line arguments

Furkan Gözükara SECourses•3/20/23, 10:11 AM

--ckptfix

--ckptfix

Furkan Gözükara SECourses•3/20/23, 10:11 AM

because when creating model it uses GPU and it causes out of memory error. this will move model generation part into cpu

DDincman Hi there, I am in need of assistance. I have been trying to use stable diffusion...

Furkan Gözükara SECourses•3/20/23, 10:12 AM

you cant use automatic1111 in colab as you are expecting. there are some repos but i didnt make any video of them yet since they are not as fluent as pc

Furkan Gözükara SECourses•3/20/23, 10:13 AM

however you can use automatic1111 in runpod

Furkan Gözükara SECourses•3/20/23, 10:13 AM

as in your pc

Zono50OP•3/20/23, 10:19 AM

I'm trying to figure out why when i train my model, it's doing everything except what I want. i created captions for each picture like this