Realtime AI Generation Local Install Guide (AI Drawing) for FREE on your PC!

Today we dive into an Easy guide for installing ENTIRELY Local Realtime AI Drawing & SDXL for FREE on your machine at home! HUGE Thanks to Gigabyte @AORUS for sponsoring today's video!
GIGABYTE LAPTOP GIVEAWAY: www.gigabyte.com/us/unleash-a...
Shop AORUS laptops: aorus.shop/MattVidPro-AI
▼ Link(s) From Today’s Video:
Step 1: Can we run this?
Step 2: 7 Zip: www.7-zip.org/download.html
Step 3: Start SDXL download: huggingface.co/stabilityai/sd...
Step 4: Download Upscalers: drive.google.com/file/d/1HNmZ...
Step 5: install comfyui github.com/comfyanonymous/Com...
Step 6: install Git: git-scm.com/download/win
Step 7: install Comfyui manager (makes things easier) github.com/ltdrdata/ComfyUI-M...
.Bat File code:
.\python_embeded\python.exe -s -m pip install gitpython
.\python_embeded\python.exe -c "import git; git.Repo.clone_from('github.com/ltdrdata/ComfyUI-M..., './ComfyUI/custom_nodes/ComfyUI-Manager')"
Step 8: Install The Canvas Tab and Efficiency Nodes
Step 9: Load custom workflow. ctrl+m to enable/disable upscaling nodes. drive.google.com/file/d/1yzvd...
► MattVidPro Discord: / discord
► Follow Me on Twitter: / mattvidpro
-------------------------------------------------
▼ Extra Links of Interest:
✩ AI LINKS MASTER LIST: www.futurepedia.io/
✩ General AI Playlist: • General MattVidPro AI ...
✩ AI I use to edit videos: www.descript.com/?lmref=nA4fDg
✩ Instagram: mattvidpro
✩ Tiktok: tiktok.com/@mattvidpro
✩ Second Channel: / @matt_pie
-------------------------------------------------
Thanks for watching Matt Video Productions! I make all sorts of videos here on KZbin! Technology, Tutorials, and Reviews! Enjoy Your stay here, and subscribe!
All Suggestions, Thoughts And Comments Are Greatly Appreciated… Because I Actually Read Them.
-------------------------------------------------
► Business Contact: [e-mail]

[@MattVidPro]

*Please Read

[@Sulayman.786]

I gave you 100 like

[@LouisGedo]

👋

[@AmazingArends]

You put a lot of work into this! Thanks!

[@AORUS]

We're happy to learn more about AI and its application with you and everyone together! Can't wait to create more AI drawing with our AORUS laptops after watching this tutorial!

[@AnthonyGoodley]

Matt you do a great job of covering the fast moving field of AI. I can't imagine what 2024 holds for AI.

[@MattVidPro]

Thank you!!

[@MyAmazingUsername]

Ever since I saw realtime editors, I kept thinking "all we need is SDXL Turbo, an empty canvas to draw on, convert it to a latent with a variable amount of noise, and have it redraw the latent based on our prompt". It's really surprising that it took so long for someone to come up with this. I guess because turbo speed models are a pretty new thing. But it would have been doable even in SD 1.0 days with a custom latent input.

[@MattVidPro]

I essentially learned comfyui in order to make this tutorial.. it is indeed suprising

[@Alex52YT]

@@MattVidProyou should make a tutorial for Stable Video Diffusion, it’s basically just a comfyui install with a SVD specialized workflow and you obviously use the models

[@SumairaSiddiquiMasood]

Don't forget to do the 2023 Year Recap! 🎆

[@MiaoYing]

Good job providing this installation guide for everyone here, really appreciate you making this as easy and simple as possible!

[@jtjames79]

Would it be possible to point a webcam as the input? I want to make a "magic mirror" for a haunted house.

[@MattVidPro]

I bet it would be. Not sure if comfy UI has an add on for this

[@user-cz9bl6jp8b]

With my 8 Gig card (lower than a 3060), it still takes about 4 secs to auto update the image but it's still good to see the weekly advancements in A.I. We need a strong competitor to NVidia because the costs of the high end video cards are unbelievable.

[@Jwoodill2112]

Installing this right now. really excited. was actually waiting for a good tutorial for a while.

[@Jjflokn]

Any updates?? How is it?

[@induhcat]

You are amazing at providing great information on the current state or the AI field! Keep up the great work 🧡

[@apre8566]

Fascinating technology, and great to see it running locally. SDXL is a massive step forward.

[@MattVidPro]

Absolutely!

[@AmazingArends]

I would like to see a proper developer come out with an SDXL art generator that runs locally with a proper warranty, tech-support etc. I would definitely pay for that! Unfortunately, there's not a lot of profit or incentive to create a real AI app that runs locally 😢

[@user-kp3ef5cv5u]

Thanks Matt! Your Video's are always simplified and very informative, I've learned so much by following you! Love the software and images I'm generating, it's all so smooth!

[@fabiano8888]

Hey Matt, big thanks for putting together that awesome tutorial! Your workflow is spot-on for gaining control over the results in image generation, way better than just relying on prompts. ComfyUI is a game-changer and feels like the future of generative AI. The real-time image generation is mind-blowing, and what's even more impressive is that anyone with the right computer can dive into it for free using open-source software. We're living in an exciting time!

[@cgrydu1249]

Great info in a clear delivery format

[@outcast6187]

Nice work, this tutorial video is much better quality than a lot of others out there, even if such content doesn't interest the widest common denominator the people who do watch it really do appreciate it.

[@MattVidPro]

I appreciate that!

[@axelsauer9589]

thanks for building awesome stuff on top of our work, love the video ❤

[@MattVidPro]

Appreciate the kind words!

[@opeyemiademosu1496]

Thanks man. I wish my system could run it. Hoping to get the AORUA giveaway. It'll be a blast. I could finally take my design to a whole new level. Thanks for the opportunity bro.

[@soullodolo]

It's fun to follow along, thank you for the guide

[@dbizseven4]

Very useful!! Thanks a lot!

[@newadventure5609]

Thanks for all you do!

[@NewYorkGiant518]

Excellent video. Thank you.

[@letsgetfamous5899]

Worked like a charm ❤

[@sportscommentaries4396]

I’d love to learn more about this, it’s amazing to see how far technology is growing now.

[@MadRango]

Great tutorial, thank you!

[@oxygenkiosk]

Love it, nice playing with it, on my new Gigabyte PC too! thank you!!

[@xyzmoral262]

It's going to really help me with my projects... Thanks, executed successfully ❤

[@normantran90]

AI has come such a long way! It's such a technical process to get set up but this video explains it very clearly!

[@MarianoMattei]

Nice - well done! Thank you!

[@techendofficial]

Such a nice content, keep it up ❤❤❤

[@david-sb5mp]

this is simply amazing

[@joelface]

Great tutorial. Wish my PC wasn't so old and feeble. haha. One day I'll run local AI, but not in the cards on my current budget. Still thought it was amazing how clear you made this tutorial... felt VERY doable.

[@konstantinlozev2272]

Just fantastic!

[@ganante]

awesome video, keep on the great job :)

[@GES1985]

This video earned that like/sub for sure. Thank you sir, and please keep them coming. Do you have a bunch of different workflows available for download? ComfyUI is awesome, but learning to write custom workflows seems like such a daunting task!

[@lainastanford9762]

Good content. Keep up the awesome videos.

[@ReidKimball]

Excellent tutorial and I got everything working. Never used Comfy UI or downloaded models before. Thanks Matt! One weird issue I had with my MX vertical mouse when scrolling, it went too far zoomed in or too far zoomed out. Had to turn off smooth scrolling in the Logitech options app.

[@centurionstrengthandfitnes3694]

Useful. Thanks, Matt.

[@MattVidPro]

Glad it was helpful!

[@Cledwyn-E]

I like how simple you make everything seem. Even if you are a 70 year old you can probably follow this easily.

[@drendelous]

agism

[@pumpkiono6330]

And what convincing way to sell one of these new fangled laptops to a 70 year old who's been banging away on a HP laptop with 8 gigs of ram for the last 5 years.😂😂😂$$$

[@RancidBaptism]

I'm in my mid 40s and this is accurate. I went from zero computer background to messing around with A.I. because he does make it so simple.

[@brainwithani5693]

Great tutorial Matt, if i ever get the hardware I'll try this.

[@MattVidPro]

Thanks! It’s amazing, and I hope you do aquire the hardware!

[@manacloud618]

great video learned a lot

[@henriquebasler]

sweeeet! can't wait!

[@wildscorpion73]

Nice tutorial!

[@Dutchruler]

In comfy you can create a group. Then put all your upscale logic in the group. That way you can mute/unmute the group rather than each control individually

[@MattVidPro]

Thanks for the tip!!!

[@stephenrodwell]

Thanks Matt! 🙏🏼

[@MattVidPro]

My pleasure!

[@addy2819]

This is super cool omg

[@Marc42]

Awesome, the capability I have always been craving ever since 2017's academic demos Two-Minute Papers.

[@MattVidPro]

Totally agree with that!!

[@elGenericos]

I love this tutorial, it looks like setting up sdxl easier, the only thing is get a better performance with ai I hope that changes with time

[@Cokeman439-Minecraft]

Great tutorial

[@calebweintraub1]

Nice work on this... my machine isn't up to par, but I'll save this video in case I can upgrade.

[@Valkyri]

Fast, helpful, free thanks for sharing this tips! Time to start drawing with AI!

[@johnnysakz]

Interesting video. I would like to try this on my PC.

[@Scn64]

Great tutorial! My cat is right here next to me and even he could almost follow it. I just had to help him with any typing.

[@MattVidPro]

Ahah thank you for watching!

[@ricardosequeira5422]

Awesome! Generative AI is better every time we blink

[@AcapellaFella]

Thanks Matt

[@MattVidPro]

No Problem!

[@jeffman112]

we love matt

[@nixthefolf7037]

It is gonna be amazing when AI is able to look at a basic drawing and be able to reproduce it almost perfectly with whatever you want

[@dawid2977]

Craaazy!

[@tyed94]

You are the best!

[@MattVidPro]

Thank you for the kind words!!

[@theillmeister]

AI is insane. I can understand the worries for traditional artists/creators, but AI is really something that is undeniably fascinating.

[@alexm7335]

awesome!!!

[@williammixson2541]

Hell yes man! Thank you. #AORUS

[@MattVidPro]

Anytime!

[@nomadame333]

Thanks, man!

[@MattVidPro]

Always here for yall!

[@adammolcan2448]

This is the future ❤

[@agnesslovehealz]

This is exciting as always and what's the swag box from gigabyte include? Do u know

[@TimothyGraupmann]

In the paint tool is there options to import an image or capture part of the screen or use OBS as a virtual camera as your input?

[@SumairaSiddiquiMasood]

👏👍💕💐 love the teaching videos!

[@amerindianrising]

Hi Matt! Love your channel, loves this tutorial, one thing, I can't seem to get the images to upscale! I've tried everything, should I reload the model? Thanks

[@SaraAB87]

Thank you

[@the_hatterdad]

Right on

[@OgyJoe]

Thank you.

[@abdulalfitouri646]

Thanks a lot

[@huni_19]

😭😭I need this cute laptop!!

[@TheOldKnight]

Open source will always be the future of AI

[@kuromiLayfe]

For some reason it only works with the base SDXL Turbo model .. any other mergers or trained models based on SDXL Turbo only output noise

[@Phoebe3315]

Can you copy/paste pictures in the canvas editor? How much can high quality art in the canvas impact the generation? When I tried with other tools they had trouble making fictional creatures look very good.

[@BlueCollarDev]

I would like to use Krita with the AI on my Wacom tablet. How feasible is that? I remember that way back this was an option

[@GES1985]

Can you put an image, or an overlay of an image, in the draw box, to trace over for specific changes

[@guzu672]

14:03 You can actually install more than one at a time. Click on the ticks on the lift, then click install on any of them.

[@LucidSurreal]

about to get a 3090 for AI purposes. can't wait to try this out with that extra power

[@ofercraft]

wow!

[@fkgmale]

Thank you Matt, this is an excellent tutorial! I noticed that ComfyUI actually saved all the preview images in the ComfyUI > temp folder. Is there any way to stop it saving the preview images?

[@MattVidPro]

Not sure about that, tbh I didn’t know that!

[@Jake-bh1hm]

Can you import an image or logo into the doodle window and have it somewhat preserve the logo and text ?

[@MattVidPro]

This is possible but not with the canvas app. You could modify my workflow to work with any 512 x 512 image quite easily

[@SpectacularSpidey670]

Matt please bring AI that helps in videogame development you promissed!

[@MattVidPro]

Still looking into making a video on this I haven't forgot about it!

[@darkknight9185]

how to use a reference picture from my pc or from screenshot ?

[@itaxiomclone145]

Why is all images so bad though? Using a pretty good prompt and know how to make good prompts. but they got extra limbs, wierd eyes, faces. How do I fix this?

[@dariusdt1214]

I've followed the tutorial, but I had to stop on one of the last steps. How do I run on AMD GPU instead of NVidia?

[@onzpo]

I want a laptop like that 😳

[@Alex52YT]

Cool

[@arturodelgadillo5244]

Do you think that open source 7b parameter models will surpass current gpt 4 in benchmarks in 25 days?

[@joelface]

no

[@agent_evandrake8027]

I am having a problem where the manager button is not showing. I feel like I followed everything you said to a T, and even did it two more times to make sure (rewinding and rewatching to ensure its proper) and i still don't have the manager button... What am I doing wrong?

[@MattVidPro]

did you create the batch file and then run it? Did you make sure you installed git? Perhaps this could help. github.com/ltdrdata/ComfyUI-Manager It seems to me like your comfyui manager is not installing

[@agent_evandrake8027]

@@MattVidPro Yes. I have followed every step with every link you have provided. I was thinking so too, but when I run the .bat file, it says "requirements already satisfied" for everything and closes in less than a second.

[@agent_evandrake8027]

@@MattVidPro And you will have to forgive me, I don't know what to do with that link...

[@MattVidPro]

@@agent_evandrake8027 strange. was comfyui closed when u first installed the manager? try restarting you PC to make sure the old comfyui isn't still running

[@agent_evandrake8027]

@@MattVidPro I guess that did it. Restarting did the trick. Thanks! This is the first time I have been able to install and run one of these. Lol

[@hostjhall]

Did they finally release this under a commercially permissive license?

[@moonshot85]

Just heard your voice narrating a random KZbin video about a chameleon a guy saved 😂

[@FusionDeveloper]

Nice

[@MattVidPro]

Thx!

[@levonkenney]

thats really cool! i tried but i gave up because nothing was working lol

[@AmazingArends]

I wish a real developer would come out with an SDXL generator that runs locally with a proper installer, warranty and tech-support. I would pay for it, but unfortunately there's not a lot of profit than that! 😢

[@sharkeys]

It sounds like an awesome thing but it barely follows what you draw, as in I draw a basic silhouette and its not even following the rough pose (even with a basic prompt that isnt pushing the model to pose differently).

[@MattVidPro]

With these realtime drawing models, (even on krea ai) you typically have to prompt the model accordingly. I also recommend trying different seeds!

[@shadowfax859]

I followed step by step, and when launching run Nvidia gpu, it pauses and says press any key to continue , which doing closes the window and nothing happens. No browser window opens, I wish you didn't skip that..EDIT: Disregard, was getting runtime error due to outdated graphics driver, updated and now it's working fine thank you

[@MichaelKafke]

Thanks for the tutorial. I am trying to run it on an AMD GPU. Worked so far, ComfyUI starts up, but when I "Queue prompt", nothing happens. Console output says "got prompt" - then nothing. Same when I started ComfyUI with CPU only (just to see if anything works at all). UPDATE: After 5 to 10 minutes I THEN get the error message that I don't have enough VRAM. Which is strange, because I have 8 GB of dedicated memory on my AMD GPU... Hm...

[@dcpuzzles2990]

I find with Comfy that, if I have a problem, I can usually find more information about what the issue is by going to the console and scrolling back up to the loading stage - it will often give you information about areas that have not loaded correctly and a pointer to the script, or issue that has caused the issue. I would recommend loading Comfy and then seeing whether there is anything in the console at that point that indicates that not everything is loaded correctly

[@RainbowSixIntel]

After downloading the missing nodes, I still get the error for the efficient loader. I have tried re downloading and restarting many times now, can someone help?

[@RainbowSixIntel]

Oh nevermind, updating comfyui worked!

[@MrEthanhines]

Neuby question but why is a dedicated graphics card required for large language models? It doesn't seem like the right tool for the job.

[@MattVidPro]

There are currently no better devices for processing what AI does. Nvidia is likely to come out with dedicated AI acceleration cards

[@MrEthanhines]

@@MattVidPro that's what i thought thanks for the quick reply love watching your vids for the latest in AI evolution.

[@AmazingArends]

It does seem strange, and it seems like even the experts were surprised that a GPU is best for AI, but Nvidia sure cashed in on it! 😅

[@dcpuzzles2990]

This is not based on specific knowledge, but I would suggest that the reason is because a GPU is just a processor that is designed to do many mathematical operations across many cores at the same time - this is something that is a massive part of the graphical creation process, particularly with 3D graphics that involve a lot of maths. My suspicion is that the use of LLMs include a lot of the same sort of calculations that the GPU is designed to handle and so the GPU capability has been co-opted to help make the use of LLMs more effective.

[@hamedmoosaei9288]

nice

[@VeganRashad]

Can I do this with a rtx 2070 super