Software Engineering Courses (SECourses)•3y ago

@Dr. Furkan Gözükara definitely check out LLaVa for image captioning. So much better than BLIP

Sstillmotion @Dr. Furkan Gözükara definitely check out LLaVa for image captioning. So much be...

Furkan Gözükara SECourses•10/23/23, 8:49 PM

i tested their online version

Furkan Gözükara SECourses•10/23/23, 8:49 PM

waiting them to release windows

Furkan Gözükara SECourses•10/23/23, 8:49 PM

not working on windows yet sadly

stillmotionOP•10/23/23, 8:49 PM

ah no good

Furkan Gözükara SECourses•10/23/23, 8:49 PM

it has checkbox

KKallamamran My img2img ControlNet is borked 😦

Furkan Gözükara SECourses•10/23/23, 8:49 PM

you see there is checkbox

stillmotionOP•10/23/23, 8:50 PM

These guys used LLaVa to caption an entire dataset https://pixart-alpha.github.io/

PIXART-α:Fast Training of Diffusion Transformer for Photorealistic ...

Sstillmotion These guys used LLaVa to caption an entire dataset https://pixart-alpha.github.i...

Furkan Gözükara SECourses•10/23/23, 8:51 PM

ye it is a good idea

Furkan Gözükara SECourses•10/23/23, 8:51 PM

leave a reply here

Furkan Gözükara SECourses•10/23/23, 8:51 PM

https://github.com/haotian-liu/LLaVA/issues/521

GitHub

Can you add venv installer? · Issue #521 · haotian-liu/LLaVA

Describe the issue Conda failing on Windows I have miniconda and anaconda as well (base) PS C:\Users\King> cd "G:\LLaVA" (base) PS G:\LLaVA> conda create -n llava python=3.10 -y Ret...

Furkan Gözükara SECourses•10/23/23, 9:38 PM

rtx 4090 speed : https://twitter.com/GozukaraFurkan/status/1716569763940372822

Kallamamran•10/23/23, 9:42 PM

Seen FreeU? https://github.com/ChenyangSi/FreeU

GitHub

GitHub - ChenyangSi/FreeU: FreeU: Free Lunch in Diffusion U-Net

FreeU: Free Lunch in Diffusion U-Net. Contribute to ChenyangSi/FreeU development by creating an account on GitHub.

KKallamamran Seen FreeU? https://github.com/ChenyangSi/FreeU

Furkan Gözükara SECourses•10/23/23, 9:44 PM

yes only difference is their prompting

Furkan Gözükara SECourses•10/23/23, 9:44 PM

ah this is not that

Furkan Gözükara SECourses•10/23/23, 9:44 PM

ye i tested it

Furkan Gözükara SECourses•10/23/23, 9:45 PM

never gotten better results

Kallamamran•10/23/23, 9:45 PM

Ok. Looked good

Furkan Gözükara SECourses•10/23/23, 9:45 PM

i tested extension

Kallamamran•10/23/23, 9:45 PM

Didin't find one for A1111

Kallamamran•10/23/23, 9:45 PM

Only Comfy

KKallamamran Didin't find one for A1111

Furkan Gözükara SECourses•10/23/23, 9:54 PM

there is

Furkan Gözükara SECourses•10/23/23, 9:55 PM

https://github.com/ljleb/sd-webui-freeu

GitHub

GitHub - ljleb/sd-webui-freeu: a1111 implementation of https://gith...

a1111 implementation of https://github.com/ChenyangSi/FreeU - GitHub - ljleb/sd-webui-freeu: a1111 implementation of https://github.com/ChenyangSi/FreeU

Kallamamran•10/23/23, 9:55 PM

Well I take your word for it that it is not worth the time

Furkan Gözükara SECourses•10/23/23, 9:57 PM

well at least it was for me

Furkan Gözükara SECourses•10/23/23, 9:57 PM

give it a try

iamannakogan•10/23/23, 9:58 PM

How long should a training take on a machine with A10 GPU? 24Gb parameters config with 14 photos -> 4480 iterations says it'll take 2.5 hours - do you think it's OK or too slow?

Iiamannakogan How long should a training take on a machine with A10 GPU? 24Gb parameters confi...

Furkan Gözükara SECourses•10/23/23, 10:16 PM

it makes 4480 / 150

Furkan Gözükara SECourses•10/23/23, 10:16 PM

so around 2 second / it

Furkan Gözükara SECourses•10/23/23, 10:16 PM

a10 is probably slower than rtx 3090 ti

Furkan Gözükara SECourses•10/23/23, 10:16 PM

so looks accurate

vKorby•10/23/23, 10:50 PM

Hey guys jus joined in here, I've been trying to figure out if this is stable diffusion that they're using? if so I’m really trying to get this type of model because i have no idea what it is, would really appreciate if anyone could help. Thanks

Maelstromboni•10/23/23, 10:55 PM

Looks like semi-realism. Legible text looks like SDXL.

You could try:
* SD XL standard model
* DreamShaper XL

MMaelstromboni Looks like semi-realism. Legible text looks like SDXL. You could try: * SD XL s...

vKorby•10/23/23, 10:57 PM

thank you going to try that out, I've been literally trying to figure out what he's using for the past month lol

Maelstromboni•10/23/23, 10:58 PM

For specific prompts, I would browse a site like civitai.com and look at the comic section for realism or semi-realism. It's obviously not photo-realistic, but something like digital illustration. Cheers!

MMaelstromboni For specific prompts, I would browse a site like civitai.com and look at the com...

vKorby•10/23/23, 11:00 PM

yeah for sure not photo real. but got it, going to try that out too thanks.

VvKorby Hey guys jus joined in here, I've been trying to figure out if this is stable di...

Furkan Gözükara SECourses•10/23/23, 11:28 PM

it is a so lame model

Furkan Gözükara SECourses•10/23/23, 11:28 PM

use sdxl you will get much better images

FFurkan Gözükara SECourses use sdxl you will get much better images

vKorby•10/23/23, 11:30 PM

Oh ok thank you, is this it?

FFurkan Gözükara SECourses it is a so lame model

vKorby•10/23/23, 11:31 PM

yeah its jus he used it and it goes viral back to back

VvKorby Oh ok thank you, is this it?

Furkan Gözükara SECourses•10/23/23, 11:40 PM

ye it is a good one

VvKorby Hey guys jus joined in here, I've been trying to figure out if this is stable di...

EyeSpyBekoAI•10/23/23, 11:53 PM

Chiming in. Yes. I’m aware of that channel I was watching a few of their vids too. Definitely a SDXL based model. I forget which video but there were one or two that had pretty distinct style that I also see a lot in my generations. If you’ll notice also the guy in his videos prompts a pic based on his narration and puts them together as slides as videos. It’s a nice way to get a short without doing a full vid.

EEyeSpyBekoAI Chiming in. Yes. I’m aware of that channel I was watching a few of their vids to...

vKorby•10/23/23, 11:58 PM

Yeah, heres another vid

vKorby•10/23/23, 11:59 PM

he's the only one doing these type of images, running multiple accounts. Ik him but he gatekeeps lol

EyeSpyBekoAI•10/24/23, 12:00 AM

Yup. Sdxl based model. They have a bit of an anime style if it’s not sdxl base directly I’m sure you’ll find it on civitai. There aren’t too many sdxl based models yet.

VvKorby he's the only one doing these type of images, running multiple accounts. Ik him ...

EyeSpyBekoAI•10/24/23, 12:01 AM

I didn’t realize he was going viral. I thought I was seeing them because I watch AI based content constantly so figured that’s why I was seeing them

EEyeSpyBekoAI I didn’t realize he was going viral. I thought I was seeing them because I watch...

vKorby•10/24/23, 12:03 AM

Yeah lmfao, we have a group for brands that we promote. He's the only one doing these type of images, took me a whole month too figure out he's using SD. Bouta try it rn gonna see if i can get sum similar

Furkan Gözükara SECourses•10/24/23, 12:08 AM

i dont think sdxl

Furkan Gözükara SECourses•10/24/23, 12:08 AM

images are so low quality

FFurkan Gözükara SECourses images are so low quality

vKorby•10/24/23, 12:09 AM

what do u think it is?

VvKorby what do u think it is?

Furkan Gözükara SECourses•10/24/23, 12:15 AM

i dont know which sd model

@Dr. Furkan Gözükara definitely check out LLaVa for image captioning. So much better than BLIP

Similar Threads

Similar Threads

Similar Threads