Software Engineering Courses (SECourses)•3y ago

Hey guys, since I have to make an impression on the influencers based on a real one, I should slight

Hey guys, since I have to make an impression on the influencers based on a real one, I should slightly change their face with stable diffusion. In your opinion, how could it be done?

Ssexyboi hmmm interesting, i haven't received an out of memory error on that with a batch...

Furkan Gözükara SECourses•12/24/23, 11:29 PM

then something must be wrong

FFurkan Gözükara SECourses then something must be wrong

sexyboi•12/24/23, 11:30 PM

okay gotcha. Yeah, I guess that should use double the amount of vram, exceeding what the a5000 can do

Ssexyboi okay gotcha. Yeah, I guess that should use double the amount of vram, exceeding...

Furkan Gözükara SECourses•12/24/23, 11:37 PM

yep

Furkan Gözükara SECourses•12/24/23, 11:37 PM

40 gb card can do 3 batch size

Furkan Gözükara SECourses•12/24/23, 11:37 PM

48 gb can do 4

Furkan Gözükara SECourses•12/24/23, 11:37 PM

i dont remember testing 2 on 24

sexyboi•12/24/23, 11:37 PM

okay gotcha, thanks for that info

MMax Hey guys, since I have to make an impression on the influencers based on a real ...

MaxOP•12/24/23, 11:38 PM

I was thinking about controlNet, for you am I on the right way?

sexyboi•12/24/23, 11:42 PM

@Dr. Furkan Gözükara when you use a batch size of 4 on a 48gb card, did you find you needed to adjust your learning rate?

MMax I was thinking about controlNet, for you am I on the right way?

Furkan Gözükara SECourses•12/24/23, 11:54 PM

ye may work

Ssexyboi @Dr. Furkan Gözükara when you use a batch size of 4 on a 48gb card, did you find...

Furkan Gözükara SECourses•12/24/23, 11:54 PM

i use same

Furkan Gözükara SECourses•12/24/23, 11:54 PM

but i didnt test actually

Furkan Gözükara SECourses•12/24/23, 11:54 PM

however it works

Furkan Gözükara SECourses•12/24/23, 11:54 PM

i trained several times

Stefanoid9343•12/25/23, 2:34 AM

Guys any of you knows if I upload my data to chat gpt is possible that they take it to re train their model? or someone else get the data trough a question that someone can ask?

୧୧ʕ•̀ᴥ•́ʔ୨ btw we're in the wrong channel, this is about stable diffusion

Stefanoid9343•12/25/23, 2:40 AM

yes I know. @Dr. Furkan Gözükara please let us know if there is another channel to ask this type of questions without interfere with stable difussion questions

SStefanoid9343 Guys any of you knows if I upload my data to chat gpt is possible that they take...

sexyboi•12/25/23, 4:37 AM

Their terms claim that any data you send into their APIs directly will not be used for training. Using ChatGPT (the web UI), however, the data will be used for training in the future https://community.openai.com/t/where-can-i-read-the-terms-and-conditions-for-the-api-service/365738

FFurkan Gözükara SECourses Click to see attachment

Enlil Wolf•12/25/23, 5:09 AM

These are amazing :Fire:

FFurkan Gözükara SECourses yes many times faster with batch size 4

Vexmachina•12/25/23, 5:32 AM

So i just tried an a6000 with 48GB memory, using your 48GB_TextEncoder.json. Tried to set batch size higher than 1, but I get an out of memory error on 4, 3, or 2, so I had to keep it at 1. Doesn't seem any faster than 4090 with batch size 1.

Ren•12/25/23, 8:59 AM

Merry Christmas to everyone

Ren•12/25/23, 9:00 AM

I am new to join here

Ren•12/25/23, 9:00 AM

I am trying to create my own SDXL lora using diffusers

Ren•12/25/23, 9:00 AM

But not working well

Ren•12/25/23, 9:00 AM

Does anyone can help me?

thisisrahul.•12/25/23, 9:03 AM

You can try creating dreambooth

Ren•12/25/23, 9:15 AM

Yeah, I am trying to

Ren•12/25/23, 9:15 AM

What I am trying to do is using diffusers to create a dreambooth lora model

VVexmachina So i just tried an a6000 with 48GB memory, using your 48GB_TextEncoder.json. Tri...

Furkan Gözükara SECourses•12/25/23, 11:49 AM

ye with that you have to enable xformers

RRen Merry Christmas to everyone

Furkan Gözükara SECourses•12/25/23, 11:49 AM

welcome

Furkan Gözükara SECourses•12/25/23, 11:49 AM

what help you need

Metaphysix•12/25/23, 12:19 PM

This is only 8 bit and yet it's highly accurate, thank you!

MMetaphysix This is only 8 bit and yet it's highly accurate, thank you!

Furkan Gözükara SECourses•12/25/23, 12:20 PM

yes llava is very powerful

Metaphysix•12/25/23, 12:21 PM

What GPU do you suggest on runpod for 16bit? 3090 is not enough

FFurkan Gözükara SECourses what help you need

Ren•12/25/23, 12:31 PM

I wanna create SDXL lora model on my own images, but not working well.

maya007•12/25/23, 12:41 PM

I want Sd WebUi API text2Image for multiple user so we can track every user process percentage runtime, Is there any extension or other repo availble for ready to used

MMetaphysix What GPU do you suggest on runpod for 16bit? 3090 is not enough

Furkan Gözükara SECourses•12/25/23, 12:57 PM

i think go with a6000

Furkan Gözükara SECourses•12/25/23, 12:57 PM

48 gb

RRen I wanna create SDXL lora model on my own images, but not working well.

Furkan Gözükara SECourses•12/25/23, 12:58 PM

which tutorial you following?

Mmaya007 I want Sd WebUi API text2Image for multiple user so we can track every user proc...

Furkan Gözükara SECourses•12/25/23, 12:58 PM

sadly i dont know

Furkan Gözükara SECourses•12/25/23, 2:23 PM

just published a great article

Furkan Gözükara SECourses•12/25/23, 2:23 PM

https://www.reddit.com/r/StableDiffusion/comments/18qjk3u/dreambooth_training_on_sdxl_anime_model_comparing/

From the StableDiffusion community on Reddit: DreamBooth Training o...

Explore this post and more from the StableDiffusion community

FFurkan Gözükara SECourses which tutorial you following?

Ren•12/25/23, 2:31 PM

I am using this code.
https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/SDXL_DreamBooth_LoRA_.ipynb#scrollTo=eTz6Zmfc0i-0

Google Colaboratory

Ren•12/25/23, 2:31 PM

But it doesn't work when I change the image folder

Furkan Gözükara SECourses•12/25/23, 2:40 PM

@Julián MANGOincode.com they are not ai

Furkan Gözükara SECourses•12/25/23, 2:40 PM

only face slightly changed

Furkan Gözükara SECourses•12/25/23, 2:40 PM

also not approritate for our channel

RRen I am using this code. https://colab.research.google.com/github/huggingface/noteb...

Furkan Gözükara SECourses•12/25/23, 2:40 PM

i dont have tutorial for that

Furkan Gözükara SECourses•12/25/23, 2:40 PM

if you dont have a gpu and cant pay runpod

Furkan Gözükara SECourses•12/25/23, 2:40 PM

we have free kaggle sdxl

Furkan Gözükara SECourses•12/25/23, 2:40 PM

https://youtu.be/16-b1AjvyBE

YouTubeSECourses

How To Do Stable Diffusion XL (SDXL) DreamBooth Training For Free -...

Master Stable Diffusion XL Training on Kaggle for Free!

Welcome to this comprehensive tutorial where I'll be guiding you through the exciting world of setting up and training Stable Diffusion XL (SDXL) with Kohya on a free Kaggle account. This video is your one-stop resource for learning everything from initiating a Kaggle session with dual ...

How To Do Stable Diffusion XL (SDXL) DreamBooth Training For Free -...

Hey guys, since I have to make an impression on the influencers based on a real one, I should slight

Similar Threads