INSTALL UNCENSORED InstantStyle Ai WebUI! COPY ANY STYLE!
HTML-код
- Опубликовано: 15 апр 2024
- Unleash the artist within yourself with the InstantStyle AI WebUI, the AMAZING New tool that lets you clone any image style for FREE! Ever wondered how to replicate the style of a famous artwork or a trendy design with just 1 click? Well, now you can do that with InstantStyle! So in this video, I'll show you how to install InstantStyle on your local computer easily and without any coding skills! I will also show you how to transform any photo by applying new styles derived from your favorite images. It's fun, it's free, and it's fully customizable!
What do you think of InstantStyle? Let me know in the comments!
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
SOCIAL MEDIA LINKS!
✨ Support my work on Patreon: / aitrepreneur
⚔️ Join the Discord server: / discord
🧠 My Second Channel THE MAKER LAIR: bit.ly/themakerlair
📧 Business Contact: theaitrepreneur@gmail.com
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
✨ PATREON LINK: / aitrepreneur
My InstantStyle Repo: huggingface.co/Aitrepreneur/I...
InstantStyle Project: instantstyle.github.io/
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
►► My PC & Favorite Gear:
i9-12900K: amzn.to/3L03tLG
RTX 3090 Gigabyte Vision OC : amzn.to/40ANaue
SAMSUNG 980 PRO SSD 2TB PCIe NVMe: amzn.to/3oBR0WO
Kingston FURY Beast 64GB 3200MHz DDR4 : amzn.to/3osdZ6z
iCUE 4000X - White: amzn.to/40y9BAk
ASRock Z690 DDR4 : amzn.to/3Amcxph
Corsair RM850 - White : amzn.to/3NbXlm2
Corsair iCUE SP120 : amzn.to/43WR9nW
Noctua NH-D15 chromax.Black : amzn.to/3H7qQSa
EDUP PCIe WiFi 6E Card Bluetooth : amzn.to/40t5Lsk
Recording Gear:
Rode PodMic : amzn.to/43ZvYlm
Rode AI-1 USB Audio Interface : amzn.to/3N6ybFk
Rode WS2 Microphone Pop Filter : amzn.to/3oIo9Qw
Elgato Wave Mic Arm : amzn.to/3LosH7D
Stagg XLR Cable - Black - 6M : amzn.to/3L5Fuue
FetHead Microphone Preamp : amzn.to/41TWQ4o
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
Special thanks to Royal Emperor:
- Totoro
- TNSEE
- RG
- Gluthoric
- Peter Bernaiche
Thank you so much for your support on Patreon! You are truly a glory to behold! Your generosity is immense, and it means the world to me. Thank you for helping me keep the lights on and the content flowing. Thank you very much!
#instantstyle #ai #imagegeneration
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
WATCH MY MOST POPULAR VIDEOS:
RECOMMENDED WATCHING - All LLM & ChatGPT Video:
►► • CHATGPT
RECOMMENDED WATCHING - My "Tutorial" Playlist:
►► bit.ly/TuTPlaylist
Disclosure: Bear in mind that some of the links in this post are affiliate links and if you go through them to make a purchase I will earn a commission. Keep in mind that I link these companies and their products because of their quality and not because of the commission I receive from your purchases. The decision is yours, and whether or not you decide to buy something is completely up to you.
![DRAGON BALL: Sparking! ZERO - Fused Warriors Trailer [BUDOKAI TENKAICHI Series]](http://i.ytimg.com/vi/84p4Lvyfyes/mqdefault.jpg)
HELLO HUMANS! Thank you for watching & do NOT forget to LIKE and SUBSCRIBE For More Ai Updates. Thx
IP adapter is not the same?
you should do a video about chat RTX and if it's worth it
@@Mente_Fugaz chat RTX supports only handful of models. "LM Studio" is much more compatible and can split performance between CPU and GPU allowing you to take in larger models than with chat RTX. Only upperhand that Chat TRX has against LM Studio is reading documents, but you can run CrewAI to achieve similar thing.
IP adapter is not the same?
@ thank you, now: chat rtx vs chatgpt4 vs LM studio,
can you enlight me?
Quite medicore to be honest. Also that could be better if this would be a addon for A1111 instead of another standalone app with own enviroment. Will pass this time.
Yeah, I see no reason this couldn't have been an extension rather than a different WebUI entirely. Seems rather wasteful on resources. Can't even get this one working right now anyway (and I'm not alone).
@@mattelder1971 not everybody uses auto1111 and you have more control over your software, options to set up premium features etc etc. It's like setting up a mall record store vs a standalone store that you can close after 9pm, or throw artist meet and greets. Opensource is great but a lot of folks don't even know how to add a star to the repo to help the developers. But yea, I probably would have said the same thing a few months ago before I got to know some cool devs that would much prefer comfy UI, their own gradio build or whatever else!
Sucks to have a middle man, but a lot of the great ones that I use (stand alone) open super fast, and get better updates (facefusion, wav2lip, lips2lora) But yea, if they started on A1111 and managed to get a fanbase (without advertisements or any mention of their site in the UI), they could launch a cooler version that transcends a1111's limits. Hope that makes sense and hope this hits a1111 (it likely has been submitted, but a1111 takes its time and is a tad rigorous with approvals)
Oh oh oh, or you know... since we're in the opensource world, you could actually make your own port for a111 called 'autostyle' or whatever and submit it! That would take work or at least LLM code vram..
I have a feeling that these purely negative comments may be more a result of the frustration of the installation that comes with a standalone. But it's faster and better and you don't want Mall Steak, you want it at a steak house. I have no idea what I'm talking about and why I'm still here.
This is a neat time to be alive. With hardware, this is free because cool people don't want AI tainted by paywalls. A1111 is awesome but I would personally rather build my UI from the ground, make it good, update it often, then do a Mall Version for A1111 Jr. version that.
Again I'm not really arguing, I'm just sharing something I've learned since I working with more devs. They usually do their own tech support too, which sounds like a nightmare.
This already is in a1111 in the form of IP_Adapter, and there are even better methods for style transfer through ComfyUI workflows. Lately Aitrepeneur seems to be out of the loop about SD. :/ With recent extensions like Ella and Incantation there's not much SD 1.5 can't do, it's still doing more and more things none of the closed source generators can do. Also AnimateDiff. Not new but it just keeps getting better.
@CrystalBreakfast Which Ip_Adapter is that supposed to be? It's a new one as far as I know
Yeah I see little reason for this to be standalone too, i can see they tried to simplify it that way, but in doing so alot of user control is removed, using it with image 2 image, (for style transfer) you get a complete and utter bastardization of your input image, its not just 'changing the style' its changing every aspect but the composition, and it doesnt seem to care much about the value set in the Controlnet conditioning scale slider.
that is very cool ! Thanks for your work, as always
Nice video thank you ^^
YOU ARE THE BEST!!
기대됩니다.😆
Nice to know about, but was so slow generating images on my desktop with a 12gb RTX 3060, that I probably won't get too much use out of it. Still good to see what new things are popping up, though. Love the channel, looking forward to the next video.
Do you use midjourney to generate your thumbnail images?
Can you use it with videos or batch sequencing?
It runs locally right? How much VRAM do you need?
Hey man, have you ever encountered with an error "TypeError: 'NoneType' object is not iterable? when using a new model? Also in the prompt word counter i get this ?/?
Does anyone know where is the output folder located? Thx in advance!
I use models with aauto awq model loader but it is very slow . any tips?
Fooocus is my jam. Easy, offline, uncensored.
it would be cool to have a gallery app that works with locally generated embeddings and has AI batch image editing capabilities. E.g. select a cluster of all images that are badly drawn and run some "professional" art style transfer over it.
I often find concept art that is great in spirit, but the artist just doesn't have the skills to depict it well. I don't want to discard those, but I also don't want to manually sift through my entire collection to find and apply some style transfer to them.
Quanto precisa pagar pra fazer parte dessas vantagens que você oferece? Pois estou precisando daquela ferramenta clone de voz ilimitada, não consegui instalar seguindo aquele seu tutorial.
Am getting:
ERROR: Could not find a version that satisfies the requirement xformers (from versions: none)
ERROR: No matching distribution found for xformers
Any ideas?
same here
same problem here
same
Same :/
F to us
can i use this on low pc?
anyone can tell me where these thing have default output folder? or do i need save every time ?
It does not output images natively. Here's what I did to add it:
from datetime import datetime
now = datetime.now().strftime("%Y-%m-%d")
time = datetime.now().strftime("%H-%M-%S")
Now in def create_image() add this above return images:
if not os.path.exists(f"output/{now}"):
os.makedirs(now)
images[0].save(f"output/{now}/{seed}{time}.jpg")
x2
how do i run this on 3070 with 8gb? i keep getting CUDA out of memory
7:15 LOL 🤣🤣🤣
this is just IPadapter with less features and extra steps. I do love your content though!
Could you show us how to run SUNO AI on our pCs!!!!! its so fun
Your discord link is expired
Mate, your content keeps getting better by the hour. Amazing! Thank you for sharing!!
i cant get it to work. i try but it uses too much vram. i have 24gb
Nice video !
Great tutorial! It takes ~18mins on a RTX4070 for each picture 😬
Thanks for save me wasting time.👍
This guy never says the GPU requirements in his videos.
OK, so it's not just my system. Something is seriously screwed up with this thing. I get an odd error message when first trying to generate an image, but it does start, but was taking at least 10x longer than Forge does to generate an image (was going to be nearly an hour on my system which has a Quadro M6000 that will even generate a large, complex image in a couple of minutes).
@@ZeroCool22 I think he just assumes (incorrectly) that everyone is running the latest and greatest hardware. The system I'm on is about 6 years old, but would have been about a $20,000+ workstation level machine when it was released (Dual 10c/20t Xeons with 32GB RAM and a 12GB Quadro M6000). This thing was going so slow I thought my system had frozen. I might try it on my wife's M2 Mac mini just for laughs, but it will probably crash that thing.
~3 sec on 4090 ( pytorch + cu121)
my command "pip" doesn't work, i downloaded again python but same issue, any idea ?
Disable all antivirus. Add python too environment variables paths.
check it's not pip3 and python3 for you
You should add pip to PATH (environment variables)
Also pc reloading can help.
Don't turn off antivirus though.
Do you have a tutorial for linux users ?
Is it automatic 1111 based or something else???
Something else :)
automatic1111 uses gradio for its ui if thats what you are asking
there will be an Extension for this in some time I guess. happened to a lot of standalone repos like this
I feel like, copying styles to another image was a thing like 7 or so years ago. I remember using a really small model many years ago to apply starry nights to a picture of mine and it already looked great (it ran on my laptop too). So though this is cool, I wonder why I see this everywhere as a big new thing?
I suspect that model you used was a lora or embedding, so it would be limited to whatever it was trained on. As in, it could do Van Gogh but not Picasso, for example. This can use any style as long as you have a image of it.
@@Elwaves2925 it wasn't a lora or embedding. Tbh I don't remember what exactly it was, but the way you'd do it back then was to train a new model on the reference image that you wanted the style off (which took only a couple of seconds even on my low-end laptop) and then you could run that model on the picture that you wanted to change.
It's long ago and I couldn't find the project anymore, but it worked surprisingly well.
It was also rather simple. This is in fact more advanced, because you can give it a prompt. The thing back then was without prompts.
@@goodguidegreg Ah okay, it was a guess as that's what it sounded like and embeddings were the earliest function of that.
@@Elwaves2925 it might've been an embedding, when I think about it. It would make sense.
ip adapter in comfyui do exactly the same thing since months
please also start adding chapter to your videos
Please, stay relevant, and start a tut series about ComfyUI!
First: it fails to install xformers.
Second: it fails to generate an image with "RuntimeError: "LayerNormKernelImpl" not implemented for 'Half'"
I am getting the same error
Follow the instructions for the Patron version. LAUNCHER-InstantStyle wouldn't load the interface
Interesting so this is how Midjourney added their new style feature.
MJ and SD are different.
@@FenrirRobu I would want to know that, nobody has been able to know how exactly works Midjourney.
Not working. It needs much of vRams, I have only 8 gb.
fly like an eagle into the future and spend some dead prezzies on a 5090
@@wakegary 😁 Yeah but I have some kids to feed
Im just shocked that you used a picture of a usual midaged woman instead of sexbomb margot robbie.
He's used that image in some other videos. His wife or girlfriend maybe??
@@mattelder1971 "Woman at Lover's Bridge Tanjung Sepat" on wikipedia
The unpacking seems to be taking too long on my PC. I'm counting over twenty minutes.
Can't start the ui, can't find a .bat to run like with Automatic 111111
I love your channel man, but 2 requests, 1)You're talking way too fast now, please slow down a bit, 2) less and less do you cover requirements for new tools, what GPU size needed, etc, remember most of your audiance probabaly dont own an RTX 4090, i'm running Instant Style Web UI on a 3070, it works. but boy is it slow.
❤A🖖
Does this work for non SDXL models?
By the way: 20+ GB for this seems a little too bloated in my opinion
Edit: my bad, in fact this shit took more than 40 GB of space. Only the folder is 21 GB. Who knows where the rest is
Probably in the hugging face cache. Use tree size or file explorer to find it if you need to in your user dir. This is why one project for multiple things is usually better.
I don't really get what's new - we could do this in sd for over a year. And yeah, ok, adding the style to an existing image - sd couldn't do that directly, but that sort of tech existed as a separate demo even in 2016
It does not work on 10GB VRAM cards.
bruh can u make, a1111 1.8.1 google collab?
i love using google collab pro 💪
i still love you,.. i vould like one click LLM webuUi AI agents videos haks pls
Your biggest problem is having 24gb vram, while most viewers have 8gb
instant style is just 2 control nets
its ip adapter...
"ModuleNotFoundError: No module named 'triton'.
I'm pretty sure that's a linux thing you can ignore, unless you're using linux.
my 4090 seems to lag out trying to process a pictures with this :/@@AmeshinART
Here for the NSFW. Keep em coming K
This AI still conflates style and content. It makes day sky into a night sky for instance.
This is very Generic, LoRA will always be better..
GenAI has actually reached the limit?
I can't see any new progress since last year ...
And Suno is what it is, i don't see any developments more than "sound like" creating generic garbage for memes
Totally agree
I think its because we are now at a point where we can't just put bunch of craps to a model and it outperforms the majority.
@@melihcloud maybe the actual best content is already on the datasets...
And AI training on it's own outputs has even stronger limitations.
don't think that is true, although progress has definitely slowed.
Have you seen udio? Doubtful, that requires some effort to look at the new releases instead of bashing the RUclipsr who's just trying to share what they did find with their own looking for new releases.
Сразу слышу что ты русский
'env' is not recognized as an internal or external command,
thats also my problem
@@mrgk1735 you have to use this dash \ when typing this command and not this one /
Why are you promoting this
Not everyone that wants to draw something the way they want to draw it has the time and energy to learn how to do it the right way or money to pay someone else do to it for them.
Why not??
Because this is an AI centric channel?
because it is an AI centric channel.
What the f is problem promoting this ?
@@dantegotaro3368 unethical?
Destroys the motivation to artists to create new styles if they will be used just to feed a machine that aims to compete against themselves with their own artwork.
If there's no more new artists AI will stagnate..
As the video said in the intro , we are going to get bored about the same styles very fast...
Is the eternal debate "AI vs artists"
It's like the piracy dilema, but in this case the original content of the artist is not leaked, but unvisibilized and devaluated which is even more harmful to them Lmao