Turn Videos Into Stylized Animation with A.I | Disco Diffusion & After Effects Tutorial
Vložit
- čas přidán 28. 07. 2024
- In this video, we look at how you can use AI technology to turn real-life footage into a stylized animation.
Disco Diffusion v5.2: colab.research.google.com/git...
*If you keep getting errors, use the newest version: discodiffusion.com/
📁Project Files:
/ 74331108
Assets:
Original clip: www.pexels.com/video/man-wear...
enigmatic_e:
/ @enigmatic_e
-------------------------------------------------------------------------------------
🔗 Software & Plugins:
VC Saber: bit.ly/2LIVms4
After Effects: prf.hn/l/ZYg0GWV
Topaz Video AI: bit.ly/3t04Otl
-------------------------------------------------------------------------------------
💻My Laptop:
ConceptD 7: acer.co/MDMZ-ConceptD7
Web: conceptd.acer.com/
Social: acer.co/ConceptD-Social
Disclaimer: Some links in the description are affiliate links. If you make a purchase through them, I may earn a small commission at no extra cost to you.
⏲ Chapters:
Intro 0:00
Tools needed 0:23
Introducing Disco Diffusion 0:40
How to Stylize 0:52
AI Prompts 4:58
Enigmatic_e 5:24
How to run the AI 5:46
Convert Frames to Video 7:06
Electric Transition Effect 8:19
Export Video 10:38
Enhance Animation Quality 10:53
Best Prompts:
['cypunk, futuristic, mechanic, dystopia, trending on artstaion, by greg rutkowski, 8k']
['digital painting, nature, colorful, illustration, sketch, spring, greenery, fantasy, featured on artstation, 8k']
['digital painting, nature, kyoto, japan, fantasy, featured on artstation, 8k']
["spooky, halloween, scary, fantasy, mystical, mist, scary, fantasy, featured on artstation, 8k"]
🎵 Where I get my Music:
bit.ly/3boTeyv
🎤 My Microphone:
amzn.to/3kuHeki
🔈 Join my Discord server:
bit.ly/3qixniz
Join me!
Instagram: / justmdmz
Tiktok: / justmdmz
Twitter: / justmdmz
Facebook: / justmdmz
Website: medmehrez.com/
-----------------------------------------
Who am I?
My name is Mohamed Mehrez and I create videos about visual effects and filmmaking
techniques. Over the past few years, I've gained a community of viewers who are interested in
digital art, video effects, and using AI in their creative projects.
#discodiffusion #ai #unitedstates - Krátké a kreslené filmy
Disco diffusion produces some really good results, but it's becoming more and more unstable, if you encounter errors. use Waprfusion instead: czcams.com/video/tUHCtQaBWCw/video.html
GREAT tutorial. I been designing 14 years and stuff like this keeps me learning and perfecting my craft. Took a few hours , but I got it, and already sold my first design using this. Thanks so much for the "sauce" you dropped! You got a new subscriber!
CUDA says its out of memory. What can i do to overcome this?
reduce the dimensions to 512*512
This one took a little too long, I know! I Can't wait to see what you guys create with this, make sure to share your outputs with me 😉
Instagram:instagram.com/medmehrez
Discord: bit.ly/3qixniz
One request for your next video🥲❤️🥺
what is it
Hey bro I made 1 edit, where can I share you?
But it needs more smoothness
CUDA says its out of memory. What can i do to overcome this?
This is insane... I respect people with knowledge, and you're one of them. Thank you for this tutorial.
AND ... Obviously, I subscribe ! 👍😊👍
I appreciate that!
CUDA says its out of memory. What can i do to overcome this?
@@MDMZ How to fix this error?
NameError: name 'animation_mode' is not defined
Amazing video bro.. The way you explain without rushing anything is appreciable. Thank you so much. subbed
Thanks for the sub!
CUDA says its out of memory. What can i do to overcome this?
@@MDMZ I want to try this effect for my wedding project but I can't create video
please help me on how to fix this error HUHU
NameError: name 'animation_mode' is not defined
This is a huge world of possibilities, Wow... That intro
Indeed
Oh yes! An huge world full of unemployed artists
@@rb3980 Also all the excitement goes away.
@@MDMZ Its extraordinary how you are digging your own grave LOL
@@rb3980 Don't tell them, I see all those youtubers on Ai, I hope AI will click on their video soon LOL.
Just Found Your Channel it's totally awesome!
I needed channels like this so I can learn VFX and those small small Animation,etc we do in After Effects
Happy to help!
CUDA says its out of memory. What can i do to overcome this?
Thanks for sharing, buddy. Great work with excellent timing.
U r welcome :)
CUDA says its out of memory. What can i do to overcome this?
What a great tutorial!! Love it! You just got yourself a new subscriber!
Welcome aboard
CUDA says its out of memory. What can i do to overcome this?
incredible. really appreciate you finding this and providing a way in. all the best from Shanghai.
My pleasure!
How did you get access to google in shanghai ? 😁
Which VPN? Genuine questions! Not trolling.
Thanks Bro, I've been seeing this stuff everywhere and found it to be very interesting so I was trying to figure out how to do it to play around with it on a video of mine that I made in Blender so I really appreciate your effort here taking the time to walk us through doing it.
thank you so much! have fun
CUDA says its out of memory. What can i do to overcome this?
Excellent!!!! Going to try this today!
Bruh. EPIC! I can't wait to try this!
A big thanks for the video. I was having errors generated each time but looking at some of your replies below seems to have sorted things. Thanks again
Awesome! Great to hear!
Yep definitely saving this guide when I try this out later
Nicee
This is ridiculous tutorial! I love your work so much!
Thank you! Cheers!
CUDA says its out of memory. What can i do to overcome this?
Thank you sir! 😊
Much appreciated for this fantastic tutorial!
Glad it was helpful!
CUDA says its out of memory. What can i do to overcome this?
Hi everyone, I noticed many of you are encountring this error
1.4 Define Midas functions
FileNotFoundError: [Errno 2] No such file or directory: './externals/Next_ViT/classification/nextvit.py'
it seems to be an issue with all DD versions that needs to be fixed by the developers, I'm looking into the issue and will post updates when I have them.
*Update* : Try deleting the 1.4 cell, this should be a temporary fix as long as you're working in video mode
For more technical support, join the Disco Diffusion discord: discord.gg/msEZBy4HxA
thanks _ workin now
I now get an error under Define optical flow functions for Video input animation mode only
/content/drive/MyDrive/AI/Disco_Diffusion/RAFT/core
---------------------------------------------------------------------------
ModuleNotFoundError Traceback (most recent call last)
in
12 get_ipython().run_line_magic('cd', '{path}')
13
---> 14 from utils.utils import InputPadder
15 from raft import RAFT
16 # from utils import flow_viz
ModuleNotFoundError: No module named 'utils.utils'
any idea ?
@@hanley-darlington same
@@hanley-darlington maybe you can try re-copying the original Disco diffusion file into your drive and start all over, or jump over to the newest version of DD v5.61, it's only a little bit different
where can i find the 1.4 cell?
What an intro
Awesome video
Thank you for sharing
You r welcome 😊
Great video bro! Thanks for the shout out!
Any time! I couldn't have done this one without your help. Cheers!
great work man ♥
Thank youuuu
I just don't know what to say, because of that knowledge you have shared with us.
Thank you so much and much respect 🙇🏽♂️🙇🏽♂️🙇🏽♂️
You are so welcome
CUDA says its out of memory. What can i do to overcome this?
BEAUTIFUL WHAT A TOOL YOU HAVE TOLD US> AMAZING CAN'T WAIT TO TRY IT. LOVE IT
glad to hear that :)
Am in a wrong class 😂😂😂 let me first go back to the basics of adobe after effects i will come back here after some years, i love the effects and editing am a beginner ❤
AmazIng! Great work
Thanks a lot!
That so coool bro! I really like it and thank you for your tutorial video ! 🤩🤩🤩
Thank you for the kind words
Great job bro ❤️🤯
Thank youuuu bro
Thanks so much for this great tutorial!!!
Glad it was helpful!
Dude Thank you so much going to try to implement this soon
Have fun
Great explanation thank you
Really helpful. Thanks man
Glad it helped!
This video is so amazing.
Wow, This is Awesome. Thanks bro.
Glad you like it!
BRUTAL, THANKS!!! 👍🏼👍🏼
Thank you for sharing your creation
My pleasure 😊
Epic! We use AI a bit for "Yosi Bunny" Video clip as well, but what you did here it's amazing!!!
I'll check it out!
@@MDMZ Your support is appreciated Thank you!!
Unreal dude!!
🙏
Super cool
Thank you so much for the information brother
🙏🙏
cools 😍😍😍 thanks for tutorial
Thank you so much for sharing 🥰
u r welcome :)
Thank you, great tutorial!
Glad you enjoyed it!
looks absolutely amazing!! time, and dedication to making it but the end result is worth it for sure!!
Thanks a ton!
not so much work
@@merlinthelemurian3197 Salty af
CUDA says its out of memory. What can i do to overcome this?
@@MDMZ I want to try this effect for my wedding project but I can't create video
please help me on how to fix this error HUHU
NameError: name 'animation_mode' is not defined
This is truly insane
VERY COOL TUTORIAL!!!
Thank you
Awesome tutorial
Thank you! Cheers!
Thanks for the tutorial 🔥🫶🏻
You’re welcome 😊
thanx bro
this is sick man 👏👍
Glad you like it!
Thank you!
@MDMZ, what is your batch's size? I mean, how long is your batch? For example, my video is like 13 seconds and the system gives me memory error. All the sources I checked recommend reducing the batch size. Are you familiar with the memory error? My GPU is 16 GB (8+8) btw and it is not enough.
Grazie!
Very cool
so cool man
Hi, thanks for the tutorial.
I have some confusion I hope you can help me.
In the prompt section there is this line "100: ["This set of prompts start at frame 100","This prompt has weight five:5"],
}" should I delete this?
And for the random seed, I saw you type in numbers what is that and how it is better than Random seeds
Thank you so much
very cool!
loveeee this tutorial
Glad u do 😊
Wow thanks for the tutorial
You’re welcome 😊
Great work
Thank you
@@MDMZ welcome
What model of stable diffusion are the videos based on? Can I create videos based on deliberate or dreamshaper models, for example?
CUDA says its out of memory. What can i do to overcome this?
Awesome bro 😎
Thanks ✌️
Thank you so much
🙏 u are welcome
Thank you so much for your informative videos! I have a question: do you know anything about Stable Warpfusion? Is it another AI or version of SD or it is a model, embedding or promt?
This is it, it's just a variation of DIsco difusion
amazing...subbed!
Awesome, thank you!
big thanks man !
Happy to help
Thanks!
I appreciate your support buddy 🙌
I really like ur videos
wow this is so wow 😳
Thx for your work ! How to be sure to have the good size for a youtube video clip please?
Wow very nice
Thank you so much
Hi, Thanks for the tutorial, nice and clear. Just checking regarding the steps - 333 seems kind of arbitrary and it makes the process much heavier on GPU. I usually use between 30-70 steps in the Deforum colab for stable diffusion. Wondering if you have been testing and experimenting with the step settings?
Hi Leandro, I have tested lower values yes, it did work for me on several occasions, but like I said the combination I showcased worked on this particular video, you can definitely try lower steps value, if it works it should save you so much time
CUDA says its out of memory. What can i do to overcome this?
Thx, gonna try this
try and share with us
This is what AI art should be. Not those who use AI with stolen image and made huge amount profits with it
U got it right, most people are confusing both aspects
at no 3. Settings it is showing me red error , running stops please tell what i can do?
what does the error say
Hey! How would I go about reducing the 'single frame' process time down for quicker renders? Yours in the video is */33 and mine is rendering at */100?
getting colab pro will make it run much faster
This video is amazing, but there's 1 thing I'm wondering, my step 4 is very slow, it looks like gonna take a whole day for a 120 frames video, so if I subscribe for Colab Pro and got the cloud GPU supported, will it boost the process even faster?
yes from my experience subscribing to the pro plan makes things faster, but not dramatically.
Thank you so much I need suggest, I use for TikTok ratio. How resolution must I choose as creating an animation? I mean ratio for cell phones. widthxheight
How do you pick the style of animation or theme? Mostly the theme, or is it random
Fantastic
Thank youuu 🙏😊
cool bro i like your video
hello
can you help me please ?
i follow the steps all but during the render it stops and show that's it completed , even it got a red play icon from "Generate optical flow and consistency maps" to the rest of setting list down it , and in google drive i just found the folder of " images_out" only not the png one i should get
Great video, on tab 4 Diffuse I get the error message n_batches ignored with animation modes.¡ do you know how to solve this pass?
I have been having a reoccurring issue with some of my animations, the frame tends to slide off screen into a black screen part way through each animation. its hard to describe but I cant seem to figure out which setting might affect this
This may be happening cause u have lots of camera movement in your footage, if i had to guess.
CUDA says its out of memory. What can i do to overcome this?
tks man!
Happy to help
How to do make an input for the Disco Diffusion to make specific style? Or is it random/one style only?
that's what the prompt is for
First, I thank you for this great information. I have a problem. The maximum level it reaches is 6% framerate, and it stops. And the call sign remains on Connecting. The signal does not return until I do disconnect and delete runtime?
Does this work on AMD graphics card for theo ther versions of this?
Good work
Thank you! Cheers!
1 question, can't we do the same effect in stable diffusion rather than Disco Diffusion? because for i am always getting error in Disco Diffusion, any way we can do it in automatic 1111?
OMG: your content, voice and speed is gold. thank you very much!!
Glad you enjoy it!
According to your settings. Do you know exactly how much storage (on google) do you need during the proccess? i supossed to have 6GB free, but now it appears that is full. it only worked 5 frames from 357 in total. BTW, Great work on this video.
Hi Ozner, I don't think I used more than 3gb for this one , but it's possible that the source files use up much space, what you can do is back up the files locally, delete the AI folder and start over, this will save you so much space
For some reason when I render, even with default settings all the frames are just Black with neon light streaks across them, even with bright input footage
don't render with default settings
@@MDMZ I copied your settings exactly at first and that gave me the problem. I then tried with default and I still got the exact same result
@WigWoo1 try gradually reducing the clip_guidance_scale or the steps, make sure cut_n_batches is set to more than 0, lastly check my pinned comment for more things you can try
how to turn on these little guides (when you move the object purple) like in photoshop?
which part of the video ?
Cool! I've already reviewed it several times. But can I achieve the same effect on the local Stable diffusion (automatic1111) in my PC? Do I need to download a specific model for this?
you can push SD for a similar effect, I believe there are certain models that gets you there, not exactly the same look, but at least it works a lot faster
thx alot
Welcome!
Hello. How would you sort out "CUDA out of memory"?
I keep getting CUDA out of memory error at "Generate optical flow and consistency maps" and i don't know how to fix it.
Do you have to keep the tab open to continue proccessing ?
Why at the prompt you only did 0: and removed the others ?
Thanks,
yes, I wanted to use the same prompt throughout the video
Question: I did everything, and I can see in the diffuse section pictures are being generated, but when it stores in my google drive, it does not show me any of the newly generated files. All I can see are the regular frames of the video in the "frames" folder. How/What should I try to change to fix this?
Hi there, I put some steps together in the pinned comment for such issues, hope that helps :)
Do i need to do this same procedure everytime I wanna use this effect? (all the steps from the beginning like copy the code and open it on a new page
etc)
no, the settings will be saved in your notebook
Thanks so much it is so awesome! Question: How did you make only part of your video in AI? Did you create it and then combined the rest in After Effect?
Yes, exactly, it's all in the tutorial 😉
😂😂😅
CUDA says its out of memory. What can i do to overcome this?