How to Turn Anime into Realistic Photos for FREE
ฝัง
- เผยแพร่เมื่อ 2 ก.ค. 2024
- I show you how to turn any anime/cartoon into a realistic photo, using FREE AI image generators. No GPU or fancy hardware required. #ai #aitools #stablediffusion
0:00 Intro
0:35 SeaArt
www.seaart.ai/
7:03 MyVocal AI
MyVocal is a quick & easy way to clone voices and use them for text-to-speech. It has emotion recognition and multilingual support. Try it out for free:
www.myvocal.ai/
8:04 Automatic1111 & Colab
A1111 Github: github.com/AUTOMATIC1111/stab...
How to install locally: stable-diffusion-art.com/inst...
Civitai: civitai.com/?...
Nolanatama 1-click notebooks: github.com/nolanaatama/sd-1cl...
Colab: colab.research.google.com/
Prompts used:
Spirited away: girl, brown hair with ponytail, white t-shirt with green stripe, white dragon
Goku: asian man, black hair, orange shirt with blue edges, muscular
Emilia: girl with long white hair, white dress with purple edges
Ai Hoshino: girl with long purple hair, pink dress, pink gloves, purple eyes, tongue sticking out
Negative: easynegative,(red eyes, green eyes, orange eyes, pink eyes, purple eyes ) simple background,(worst quality:2), (low quality:2), (normal quality:2), lowres, normal quality, skin spots, acnes, skin blemishes, age spot, glans, (watermark:2),(dusty sunbeams:1.0),(abs, muscular, rib:1.0), (depth of field, bokeh, blurry:1.4), (greyscale, monochrome:1.0),(malformed hands:1.4),(poorly drawn hands:1.4),(mutated fingers:1.4),(extra limbs:1.35),(poorly drawn face:1.4), pointy ears,((logo,text)), (extra arms:1.15), (extra feet:1.6), (extra legs:1.65), (fused fingers:1.05), (too many fingers)
Discover thousands of AI Tools. Also available in 中文, español, 日本語:
ai-search.io/
Here's our equipment, in case you're wondering:
GPU: RTX 4080 amzn.to/3OCOJ8e
Secondary GPU: GTX 1080 (too old, would not recommend)
Mic: Shure SM7B amzn.to/3DErjt1
Secondary mic: Maono PD400x amzn.to/3Klhwvu
Audio interface: Scarlett Solo amzn.to/3qELMeu
CPU: i9 11900K amzn.to/3KmYs0b
Mouse: Logi G502 amzn.to/44e7KCF
If you found this helpful, consider supporting me here. Hopefully I can turn this from a side-hustle into a full-time thing!
ko-fi.com/aisearch - วิทยาศาสตร์และเทคโนโลยี
Your videos don't disappoint. Thank you!
Glad you like them!
honestly i wasn't looking for this but im loving SeaArt!
glad you like it!
Awesome video, and great tips!
Thank you!
The idea that any cartoon show can be transformed into real live people show, cool.
Great tutorial
Thank you! Cheers!
HorseGron! look cool sir, 👍🤯👍
For different anime and drawing styles you will have to use different models and parameters. With that said its much better than MidJourney
woow awesome
great video
Thanks!
Thank you for the great content, it’s so helpful for newbies. I’m wondering if it’s possible to convert a real picture to a certain style cartoon image? Say to convert a girl’s photo to a Simpson figure? Could you do a tutorial video on such conversion? 😊
Yes you can! Short answer is, it's the same workflow. Import a realistic picture to img2img, and use a cartoon or anime model (checkpoint) to generate the image
I wouldn't recommend Google collabs it's been cracking down on people using it for AI picture and voice synthesis. The collabs keep disconnecting after a short amount of time.
That happened to me when I wanted to train an AI voice. It suddenly disconnected and I thought it was a bug... But thanks to your comment I know why :,)
Same here so frustrating 😭
@@theAIsearchI've been hearing it from multiple Discord users. Thank goodness I have a decent GPU.
@@alexiscoutinho8078 Don't count on the internet being there much longer. I do everything offline when possible. Websites can disappear, start charging, or the government can shut off the internet in the name of "public safety".
Yeah, it seems Google is cracking down on people using colab for stable diffusion. Best if you can run A1111 locally, otherwise use seaart
nice tutorial. I forgot you dont really need controlnet to do this kind of stuff.
Thank you!
Could you please write down the negative prompt you used in this video?
Oh wow, the quality got better
Good info, but something is wrong with Automatic 1111: after the first generation an error pops up saying "Something went wrong Connection errored out". Any fix? I heard Google allows Stable Diffusion usage through Collab for payment only, maybe that's the reason?
Hmmm that shouldn't happen. I can generate lots of images from one session in colab. What error message do you see in the logs (like in the colab page)?
@@theAIsearch "Something went wrong Connection errored out" - that's the error in the Stable Diffusion window in the upper-right corner. In the collab the cell icon becomes red and its animation stops. Don't see anything else. It all happens after one or two image generations.
It happened to me when I poked some other options that weren't covered in the video like the Interrogate buttons and Styles refresh. But otherwise working good.
Ok, I got an error out of nowhere. And nothing is said in the collab output.
It seems Google is cracking down on people using colab for stable diffusion. Best if you can run A1111 locally, otherwise use seaart
BTW for me easy diffusion is work the best. Have everything what automatic11 but easier and working on my 4gb vram old gamer laptop.
why mine still look like an anime, seems like they just change the version of it but still anime
Same thing here. Nothing changes, just the aspect ratio. Maybe the model is not the right one :(
i couldn't get either 1111 to work. not in the colab. the github pack pack would display the path and the site was ip address site wasn't accessible during the manual install
the ip address link wouldn't work on colab. try the gradio live link
Pretty cool, though I think the results for the two programs are extremely similar apart from if there are errors or not... Anyway thanks! Fun tools to play with.
Yep, they're both based off of stable diffusion
great tutorial, someone please can tell me which model I can use to change the clothes of one anime character? something like the video just uppdate an image and change the clothes
For that, it may be better to use inpaint in A1111.
@@theAIsearch thanks, I Will try
Man In future If there is an election for AI president I would recommend you and we will make you as AI president ❤ because you have given us many things as free thank you man.
Haha, thanks man!
@@theAIsearch ❤️❤️
I have a question.
Have a photo where the model is in a dress (any outfit), but they made the model without a face (the photo is cropped at the level of the neck) how to use AI to add the face to her of a beautiful Korean girl ?
u can use outpaint for that
I dont know how you do it.. this thing keep changing my prompts after pressing regenarate and never come out as real. Piss me off some much. Tech hate me I guess. Having a hard time with all this ai tools. Cant they just poop an image from anime to realistic in just 1 click? How hard it is....
Also Is there any free alternative for runway and Stable diffusion also if you could you please add to it and for us thanks
free alternative to runway would be pika: th-cam.com/video/tDidofPTwCc/w-d-xo.html
@@theAIsearch thanks man
Does any of this work on a Macbook Pro M2 Max, 96 GB? I see you do everything on Windows. My Windows PC is very slow :(
Yes! Everything is done in the browser
I thought everything was paid and could only try free once or twice. thanks for sharing
You're welcome!
yea um.... i type and click on all the settings.... finally I click enter and the photo comes out the same anime drawn as it was before?? what is up with that....only her eyes get distorted like an alien thats it
make denoising strength higher
bro thats no more goku its now goku musk
🔥🔥🔥
I cannot use this it shows error code is Something went wrong Connection errored out. How can I fix it. (It don't show finish picture. Thank you very much.
A1111 won't work in colab anymore. I'll make a new video next on how to STILL use stable diffusion in colab w/o disconnecting. will publish in a few days!
@@theAIsearch Thank you very much.
i started the run command on google colab and it told me that it cost money to run that program
colab is cracking down on people using this for image generation. It's best if you can install it locally
The collab disconnected a lot. Having a hard time to even train 1 voice
it's best to install A1111 locally. colab is cracking down on people using it for image generation
Uhh my pc broke down 3 months ago but i don't know if you can use a girl voice on mobile
There's no good solution for mobile at the moment
it looks like this ai can only understand mangas. I don't know but with my draws the end result looks like a transporter-beam-accident in Star Trek 😅
Cool, i guess
😃
Is there another way around?
what do you mean?
@@theAIsearch photo to anime. I been checking a few but please recommend as well.
it keeps saying No interface is running right now
r u using colab?
@@theAIsearch oh i just copeid the link, now its saying error
@@theAIsearch yes, i am
oh i have to re-run the cell
no that did not work
Emilia hot af
🔥🔥🔥
It's a scam. I've used the same Chihiro picture and set up the same settings and the result was terrible
Egg salad!
Just learn to install and run stable diffusion for free local no limit 😂
Just have a powerful PC.
Fisrt option only works with asian people tho
Does it make HENTAI?
u can make whatever on A1111
I believe nsfw is restricted for seaart
Can you give Google colab for kaggle...?
ive heard kaggle is even more stringent than colab for art generation
@@theAIsearch Google colab banned automatic 1111 to free user, so kaggle is the last way...
These "photos" look quite impressive, but they are only pseudo-realistic. The proportions of the humans are totally off, no real-life human has such a huge head as compared to the body, not even a baby. Especially in the image from Spirited away this is quite obvious. This is because the bot does not understand, how human proportions ought to be correct. Maybe future generative models can overcome this limitation.
how to download sea art ai
no need to download. just use in your internet browser
@@theAIsearch okkk thankss nice vid buddy 😉
bisa di android?
I used seaart on android
This is cool and all, but I want to know what is the purpose of this is.
AI art don't have a professional purpose right now except for YT thumbnails, making some pictures stories and the main and the only reason its developing so fast, corn.
Google did just block Colab?
seems like it. its best to run a1111 locally
That not goku .. thats chines elon musk cosplay as goku .. hahaha
Great video...(I hate anime)
i say nope. stay in 2d
true, some things r better left as 2d
My goodness... SeaArt is giving some terrible results following every step in video 👀 faces all deformed, strange artifacts everywhere... Hm 🤔
Play around w the denoising strength and sampling method
Can you please make an update? The interface of the platform already changed as for July 2024 :😢