Wow, you are the first content creator that I found who doesn't shy away from the hard topics and challenges in stable diffusion and actually make a comprehensive coverage on it. Kudos!
hey my old friend lost your channel when my GPU got cooked and I stopped doing AI art because of it, finally found you again by accident today I love your tutorials and I have to say Wolverine is not going to appreciate his cousin shaving his beard like that
first of all, I'm sorry that your gpu got fkd and I hope you are able to solve that issue. And thank you for the kind words Second... hahahahhahhahahahha dont worry it'll be back in 3 days xD
@@Not4Talent_AI Nah it's dead and it doesn't look like I'm gonna be able to replace it I live in Iran my wage is 150$ a month there is no way I can afford a new GPU after all the bills but I'm still extremely fascinated by AI art which is the reason I still watch and love your videos, P.S: as a middle eastern with a thick beard I completely understand lol
As an artist, holy shit, I'm glad I know how to draw. Those expressions should take 10-20 seconds to rough depending then give it to AI, that'll be my approach. With the AI its like a multiple stage processes with like 12 parameters and a bunch of specific loras for a generic expression. Good example of doing simple things the hard way but it will come in clutch if things are complicated maybe.
Yup, 100% When it comes to control in the image, being an artist is and probably always will be the best and fastest method jahahaha. But getting to the point where you are able to draw good facial expressions on any style is extremely hard. Even still I think its super worth to learn how to draw if you care about having control over the final result
Hi, thanks for sharing so much. I'm learning a lot from you. As someone said, don't shy away from complex topics, unlike most content creators. I do have one question. Where did you get that OpenPose Maya rig? Did you make it? It would be awesome to see a video of it. It's too bad you don't have a Patreon because I'm sure many of us would support you. The wealth of information that you share is invaluable. Thanks again.
thank you so much!!! I made it yeah. But there is a better version of that rigg made for free for blender. And it has a lot more stuff. toyxyz.gumroad.com/l/ciojz
hey, sorry for the random question, but how can I have this interface for the 'lora' tab, shown at 0:34? Is this because of the version you use or do you have an extension to make it like this? Btw, thanks for the great videos and knowledge you share with us!! 🤙
@@Not4Talent_AIMine is 'version: v1.8.0', but the layout is different, the 'tree view' of the folders appears on the side, I tried clicking on all the buttons and looking for settings to change it, but I didn't find anything. Anyway, thank you very much!!
Really great video as usual sir. Have you considered using the Krita open source paint programme with Stable Diffusion plugin? Might be a huge time saver for some of these issues 👍
Hi there. Have tried downloading the lora pack 5 times but it gets to around 1gb and fails, and wont let me resume. It always starts at fast speed then drops to 300kb/s too, any chance you have mirrored it alsewhere?
hi!! probably something to do with the kofee page atm. I currently dont have it mirrored, but if it keeps failing let me know and I'll try to upload it somewhere else too. (I got a pretty sad internet connection so it will take a few hours) . Srry for the inconvenience
Just click the lora and it will appear on the prompt. Idk if that was what you are asking though. If you want to know where to put the downloaded loras, its in: Stable-diffusion-webui > models > Lora
Muchas gracias! Ojalá tener tiempo y lo hacia en español tmb. A ver si puedo encontrar a alguien pa que lo doble. O pillar una IA que lo doble al español por mi 😂😂
You could use online sites like google colabs or even places like: dreamlook.ai/?via=N4T (I am sponsored by them, but it's a great service). I'm sure there are a lot of other places to go to tho. leaonardo.ai, www.thinkdiffusion.com/ ...
This might be a dumb question but: How do you view the meta data on the loras like it shows up for you? When ever I click meta data on my end it just shows a text screen with what looks like some code. Am I missing a plug in or something?
I used Sdxl inpainting at the begining of the example for the man. It really depends on the model imo. I was generating with colossusproject by afromanforpeace. But yeah, over all I use mainly sd1.5 because it has 90% of stuff. Like cn models, good extension support, way more loras and models.... Even though over time sdxl is getting those and getting better
Yeah, like 95% 😂 didnt find much for realistic. Some of them do yes. Idk if most of them, havent tried them all on realistic. Usually slider loras work well on realistic to a certain extent. As well as expressions that are simple and not too anime-centered. But Id say that maybe 10% of anime loras work with realistic unfortunatelly
@@Not4Talent_AI I've never tried LORAs with only anime pictures as examples because I expected them to not work with photorealistic images. I'll give it a try anyway.
Many thanks to the author for the study material and the compendium Lora's🙏I'm very ashamed that I took Lora's set for free, but monetary transactions from my country are forbidden as we unsuccessfully😮💨 try to destroy the entire civilized world.
this is merely experimentation and not prodution ready at all, a complete waste of precious time. these loras will be swiftly forgotten when the next better ai thing comes along
this is awesome, though wish there was a laughing lora lol that was the big reason i downloaded this pack, and most of the reactions are for cartoons/anime.. oh well.
Yeah sorry. Really tried to download all the loras I could, but after a week I had a ratio of like 1 realistic per 70 anime 😂 Some of the anime ones work for realistic too though, Id see if the laughing ones do. I remember a laugh slider that kinda worked with realistic as well. Maybe not for SDXL tho. (Maybe there are new loras on civit ai for it too btw. Havent updated the pack and SDXL has gained a lot of traction lately(
Wow, you are the first content creator that I found who doesn't shy away from the hard topics and challenges in stable diffusion and actually make a comprehensive coverage on it. Kudos!
hahahha Thank you !! Hope it was helpful!
Love your videos man. It shows the process, struggles and hardship to make complex AI Arts.
If anyone of my friends says that AI art is easy. Will be sure to send one of your videos. 😂
Hahahhahahaha
Ty!!! Yeah, i feel like creating cool stuff with AI is easy. But creating the cool stuf you have in mind is pretty tough
As soon as possible, i will take somes weeks to learn things on your videos, there are very helpful
thank you!! Good luck and have fun !
hey my old friend lost your channel when my GPU got cooked and I stopped doing AI art because of it, finally found you again by accident today I love your tutorials and I have to say Wolverine is not going to appreciate his cousin shaving his beard like that
first of all, I'm sorry that your gpu got fkd and I hope you are able to solve that issue. And thank you for the kind words
Second...
hahahahhahhahahahha dont worry it'll be back in 3 days xD
@@Not4Talent_AI Nah it's dead and it doesn't look like I'm gonna be able to replace it I live in Iran my wage is 150$ a month there is no way I can afford a new GPU after all the bills but I'm still extremely fascinated by AI art which is the reason I still watch and love your videos, P.S: as a middle eastern with a thick beard I completely understand lol
damn man that sucks. wish you the best of luck then! Hopefully economical situation gets better
@@sahanddoeseverything8373
oh hell so much useful information for proper poses
thank you! Glad it helps
@@Not4Talent_AI i am going to work a bit as a character illustrator in one project so your videos are really a bless for me
wow thats super nice to hear!! Best of lucks! (and thank you so much for the kind words) @@Navi-Val
As an artist, holy shit, I'm glad I know how to draw. Those expressions should take 10-20 seconds to rough depending then give it to AI, that'll be my approach. With the AI its like a multiple stage processes with like 12 parameters and a bunch of specific loras for a generic expression. Good example of doing simple things the hard way but it will come in clutch if things are complicated maybe.
Yup, 100%
When it comes to control in the image, being an artist is and probably always will be the best and fastest method jahahaha.
But getting to the point where you are able to draw good facial expressions on any style is extremely hard.
Even still I think its super worth to learn how to draw if you care about having control over the final result
The Goat 🐐🐐
tyty!
Hi, thanks for sharing so much. I'm learning a lot from you. As someone said, don't shy away from complex topics, unlike most content creators. I do have one question. Where did you get that OpenPose Maya rig? Did you make it? It would be awesome to see a video of it. It's too bad you don't have a Patreon because I'm sure many of us would support you. The wealth of information that you share is invaluable. Thanks again.
thank you so much!!! I made it yeah. But there is a better version of that rigg made for free for blender. And it has a lot more stuff.
toyxyz.gumroad.com/l/ciojz
@@Not4Talent_AI Yeah, I saw this one. Thanks!
i am impressed
hopefully for the best hahahaa
@@Not4Talent_AI of course!
hahaha thats nice to hear, Thank you then! @@pn4960
Banger video as always!
Thank you so much!!
se extrañaban tus videos, esta excelente el video como siempre
Ajajajjaja muchas muchas gracias!
As always, amazing video with powerful techniquies. Thanks a lot!
thank you for watching and leaving a positive comment!
GOAT
thank youuu!!! ya'll GOATed for watching!
my laptop couldnt run that sht for its life bruh T-T, it already takes 20 min to generate the pics
rip. Maybe try using a colab or an online site? I think civit AI has generation services
@@Not4Talent_AI I'll check it out thanks
this is the most useful AI channel
tysm!!!
Me salio en recomendados, tendras buen exito
Gracias!
HE'S BACK!
That's right! Back at it seriously again 😂
bruh watching you test out all these new things makes me wanna upgrade lmao, I really wanna hold out for the new cards tho haha@@Not4Talent_AI
hold it!! hahaahaa
but yeah the new pc goes fast af hahahaha@@Rasukix
meanwhile It's taking me 10 minutes to generate 8 images lmao@@Not4Talent_AI
Damn. The perfect video is exist
hahahahahahha tyty!!
hey, sorry for the random question, but how can I have this interface for the 'lora' tab, shown at 0:34?
Is this because of the version you use or do you have an extension to make it like this?
Btw, thanks for the great videos and knowledge you share with us!! 🤙
it comes naturally with versions of Automatic1111 over 1.6 I think
And tyty! hope it helps !
@@Not4Talent_AIMine is 'version: v1.8.0', but the layout is different, the 'tree view' of the folders appears on the side, I tried clicking on all the buttons and looking for settings to change it, but I didn't find anything.
Anyway, thank you very much!!
oh, weird. maybe they changed it again XD I'm on 1.6 still hahahahaa
np!! (idk how to change it either xd)@@rafaelsato3573
Mega helpful video!
Ty! Happy to hear it
thx, sir. respect you sir
tyty!!
super useful ! thks 👌
glad it helped, ty for watching!
I agree fixing in photoshop will save you typing more "outrages" final prompt
100%
Really great video as usual sir. Have you considered using the Krita open source paint programme with Stable Diffusion plugin? Might be a huge time saver for some of these issues 👍
Thank you!! I actually havent, I have to look into all of that stuff... I'll note it down to do soon or I'll forget xD
thanks for the suggestion!
Instant sub
hahhahaha thank you so much!
That league of legends emotion is spot on lol.
😂😂😂
When are you doing the Slider video for? Definitely something I’ve been wanting to do
I'll try to make a tutorial on how to train sliders with as much detail as I can figure out, since there is very little info on the topic
@@Not4Talent_AI that would grand. Do you have a discord?
yup! description of the video, second link (I think)
@@___x__x_r___xa__x_____f______
Eres un crack
muchísimas gracias! espero que el vídeo sirva
Hi there.
Have tried downloading the lora pack 5 times but it gets to around 1gb and fails, and wont let me resume. It always starts at fast speed then drops to 300kb/s too, any chance you have mirrored it alsewhere?
hi!! probably something to do with the kofee page atm.
I currently dont have it mirrored, but if it keeps failing let me know and I'll try to upload it somewhere else too. (I got a pretty sad internet connection so it will take a few hours) .
Srry for the inconvenience
will try again tomorrow and see how it goes and let you know. Cheers! @@Not4Talent_AI
@@Not4Talent_AIunfortunately it still doesn't work
really?? doest it stop downloading or it just doesnt let you? @@keylanoslokj1806 (its downloading fine for me)
@@Not4Talent_AI stops midway yes
Perfect!
ty so much!
in the end you still need a talent.... (manually edit using photoshop)
hahhahhahahhaa Idk if I'd consider that a talent (since it's very simple retouching), but I like the play on words xD
How about img2img? And batch sequencing……
Idk what you mean ngl 😂 img id rather inpainting to not change everything else
How do you implement the Loras on the model?
Just click the lora and it will appear on the prompt.
Idk if that was what you are asking though.
If you want to know where to put the downloaded loras, its in:
Stable-diffusion-webui > models > Lora
Awesome!
Thankss!! Hope it is helpful 😄
👋
hi! xD
excelente video, aunque lo quisiera en español
Muchas gracias! Ojalá tener tiempo y lo hacia en español tmb.
A ver si puedo encontrar a alguien pa que lo doble. O pillar una IA que lo doble al español por mi 😂😂
1:41 LMAO
I also want to make all this but I don't have a GPU.
You could use online sites like google colabs or even places like: dreamlook.ai/?via=N4T (I am sponsored by them, but it's a great service).
I'm sure there are a lot of other places to go to tho. leaonardo.ai, www.thinkdiffusion.com/ ...
This might be a dumb question but:
How do you view the meta data on the loras like it shows up for you? When ever I click meta data on my end it just shows a text screen with what looks like some code. Am I missing a plug in or something?
Im using automatic1111 version 1.6. something. It came with it installed. It'd the button to edit
@@Not4Talent_AI Thank you for the quick reply. Yeah, this was the problem. I had to update.
nice!! @@birdmanoo0
voy corriendo a bajarme 200+ loras
Al final son 164 porque no me dejaba subir las 200 (demasiado peso)😂
What do you think about sdxl? Seems like inpainting is a no go with it. Are you staying in 1.5 because of the lack of adoption?
I used Sdxl inpainting at the begining of the example for the man. It really depends on the model imo.
I was generating with colossusproject by afromanforpeace.
But yeah, over all I use mainly sd1.5 because it has 90% of stuff. Like cn models, good extension support, way more loras and models....
Even though over time sdxl is getting those and getting better
Most of these LORAs seem to be made for anime. Do they usually work for realistic images, too?
Yeah, like 95% 😂 didnt find much for realistic.
Some of them do yes. Idk if most of them, havent tried them all on realistic.
Usually slider loras work well on realistic to a certain extent. As well as expressions that are simple and not too anime-centered.
But Id say that maybe 10% of anime loras work with realistic unfortunatelly
@@Not4Talent_AI I've never tried LORAs with only anime pictures as examples because I expected them to not work with photorealistic images. I'll give it a try anyway.
your PC specs?
I upgraded right before this video (I posted a vid where I show a comparison with my old pc too). I have a 4090 now
Many thanks to the author for the study material and the compendium Lora's🙏I'm very ashamed that I took Lora's set for free, but monetary transactions from my country are forbidden as we unsuccessfully😮💨 try to destroy the entire civilized world.
hahhahhahaa The fact that you watch the video is more than enough! also, the pack is 100% free, 0 need to pay anything for it!
has anyone ever told you that you look like christopher moltisanti
ngl never, but thanks I guess XD
@@Not4Talent_AI dw hes a good looking guy
hahahahah ty then XD@@individuality4422
I just sketch the parts manually , way faster personally.
For sure, anyone that knows how to draw has such a massive advantage 😂
brother is every thing free?
yep
the lost brother of @Hydemagician
😂😂😂😂
peluchitos supa monos (very cute plushies in case you don't speak Spanish)
Hahahahhah tii
I doubt he doesn't speak Spanish 😂
😅@@sherpya
this is merely experimentation and not prodution ready at all, a complete waste of precious time. these loras will be swiftly forgotten when the next better ai thing comes along
yes, ofc. Thats how technology works
this is awesome, though wish there was a laughing lora lol that was the big reason i downloaded this pack, and most of the reactions are for cartoons/anime.. oh well.
Yeah sorry. Really tried to download all the loras I could, but after a week I had a ratio of like 1 realistic per 70 anime 😂
Some of the anime ones work for realistic too though, Id see if the laughing ones do. I remember a laugh slider that kinda worked with realistic as well.
Maybe not for SDXL tho. (Maybe there are new loras on civit ai for it too btw. Havent updated the pack and SDXL has gained a lot of traction lately(
The number of times I see that hairstyle… surely there has to be a LoRA with a different style.
hahahahahahha probably yeah XD