Deforum for AUTOMATIC1111 - Stable Diffusion Tutorial - AI Animation Part 1

This is the first part of a deep dive series for Deforum for AUTOMATIC1111.
Part 2: • Deforum for AUTOMATIC1...
SD Deforum Deepdive Playlist: • Deforum for AUTOMATIC1...
In this Stable Diffusion video, we'll be learning about how to install and setup Deforum for AUTO1111 to create your own AI animations.
Furthermore we will take a look at the first parameters of the Run settings and Keyframe Settings of this script. Every setting will be visualized with a comprehensive example, so you can see exactly how the changes affect the outcome. After watching this tutorial, you'll have a first overview of how we will be using Deforum!
--- Chapters ---
00:00 - Intro
01:12 - Installation & Setup
02:47 - Run Settings
05:20 - Keyframe Syntax
07:00 - Keyframe Settings
--- Relevant Links ---
Deforum github: github.com/deforum-art/deforu...
Automatic1111 github: github.com/AUTOMATIC1111/stab...
Deforum discord server: / discord
Official user guide:
Deforum guide by dreamingcomputers: dreamingcomputers.com/deforum...
ffmpeg setup: phoenixnap.com/kb/ffmpeg-windows
#stablediffusion
#aiart
#aianimation
--- Music Credit ---
Music from Uppbeat (free for Creators!):
uppbeat.io/t/adi-goldstein/i-...
License code: H1OLSAGGJQINAP66
uppbeat.io/t/prigida/adrift
License code: FPEQL0NRL3HKHYWV
uppbeat.io/t/prigida/cherry-b...
License code: 6PTJXRTSSQCHV49I

[@fictitiousness]

Check out Part 2 right here: kzbin.info/www/bejne/pIbZgZWhdruWkMk

[@alessandromolinari4256]

But the Images where did u get the images

[@ArisenProdigy]

This is seriously amazing. Thank you for the in-depth explanations and visual representations. Definitely looking forward to part 2!

[@duxington_]

This is the most comprehensive and easy to understand tutorial I've seen for deforum yet, I'm keen for part 2

[@Bukulmus]

I cannot thank you enough for this video! By and far the most detailed and easy to follow tutorial I have discovered on running Deforum locally. So much potential with this teaching style. Very grateful

[@charapoKKo]

Great video. The tutorial was clear, informative, well-paced, and easy to reference. The side by side output + command examples were incredible in reinforcing the concepts. I have never subbed+liked on a video so fast. Honestly can't wait to learn more from you. Thank you.

[@MrPlasmo]

Best video I have seen so far to go over the settings in one place. Nice job.

[@yolamontalvan9502]

So you finished college and now you decided to give us these amazing videos. Just what I needed. Thank you.

[@omegaelite3]

This is the most in depth, step by step video I've seen so far about deforum. No one else goes this in-depth. Good job, I def see you growing this channel to the moon.

[@st.michaelthearchangel7774]

In the truest sense of the word, all these AI programs coming out are awesome.

[@chacecampbell2697]

This is EXACTLY what I was looking for, thanks for taking the time to make this!

[@77110ful]

Thanks for the in-depth explanation and examples. Appreciated your efforts in making these tutorials. Much love 🤗

[@raneanubis]

This has got to be one of the best and clear tutorials for a noob like me. Thank you.

[@devdasshenoy3588]

Excellent Video. This was really useful especially for people just starting out. Looking forward to part 2!!!

[@moe_joe_man]

Awesome video and very well done!!! Thank you for this information it is greatly appreciated in the Deforum community 👍

[@BrianSheppard]

You are doing an amazing job with these tutorials!

[@sungod9481]

Subscribed. Fantastic Tutorial. Thank you, and thank you so much for the demo videos starting with the same original image. Incredibly helpful for visual learning.

[@junebarry7456]

Great stuff, the actual video exemples make all the difference!! Thank for you work.

[@ronami3176]

Thank you for taking your time and making this tutorial!!!

[@marianoequizzi]

one of the MOST clear video in the planet about DEFORUM, comme on dit en Francais : CHAPEAU!

[@elad3958]

This channel is love at first sight

[@marcdevinci893]

Thank you this is very helpful, looking forward to part 2!

[@adriansimancas]

this is the best ia tutorial ive seen so far on youtube, simnply perfect

[@ikizinteki]

these are the best video content about the a.i world. thanks mate this one really helps me. I can't wait for part 3.

[@hapoTT]

I came up with everything myself, but this video is great. Clear and cool, thank you.

[@FCCEO]

dude,,,,master......you are the best! I have been looking for deforum tutorials like this. Thank you so so so much.

[@iampjm]

Love where this is headed already!! I appreciate your calm tone and in depth explanation. Hoping in future parts of the series you will be able to cover the 'Video Input' mode more in depth. Thanks!

[@fictitiousness]

Everything will be covered :)

[@Datboynutty_nutelli2go]

It's an AI voice I believe lmao

[@tonyambito8028]

best tutorials ive seen on deforum..thx man

[@HMaxTube11]

Tremendously valuable exploration of the iteration. 👍👏🌟Subscribed

[@PrincessSleepyTV]

You have me following so effortlessly along to your tutorials thanks to your How it's made type voice XD

[@neilfpv]

This is amazing! This is the video I need! I think I'm 5 days old now with using SD and Deforum. It's absolutely amazing! Thank you for this tutorial! You got my sub!

[@majora2765]

epic tutorial i finally understand alot of it now. PLZ PART 2

[@ILOVEMUAYTHAI]

Thank you!

[@BryGuy_TV]

Amazing tutorial. There is so much information packed in this video

[@TheJarvins]

Really great video, looking forward Part2

[@mosmanya]

beautifully done. Thank you.

[@phlerpdesigns9273]

Wow this video is so well produced and so clear and concise. I will be coming back and referring this video for sure

[@phlerpdesigns9273]

Also the so helpful to understand the strength and that formula! Was always wondering why the first frame was so detailed and everything else got so blurry!

[@aaavisuals2950]

ur a life savior, insane quality of your video thanks !

[@Evijan]

amazing you just saved half my life trying to learn this

[@user-hv3ey2zg4i]

This video is a gem!

[@viralml]

Excellent explanation!

[@aispookystorytime]

Excellent video, thank you so much!

[@CaramelBass]

Love this tutorial!

[@CapApollo]

invaluable how to.. i will wait for part 2

[@showplusbiz]

Exemplary explanation. Thank you

[@p_p]

thanks. Very professional guide and great video edit

[@ALLinDRAMA]

Great video, got my sub ✌️

[@andreykochergin]

this is soo dope tutorial, actually the best one, my respect sir

[@nzrify]

Thank you for this :)

[@cinevizual]

this is amazing

[@ramsesmendoza8951]

men finally a good tutorial.

[@aishajamila]

Thanks 😀 I finally found something I could use

[@prazeem8359]

Thank You!

[@Michael-Humphrey]

Ok we need a part 2 asap!

[@spacedDiffusion]

Good Job, great video.

[@coloryvr]

Wow! ....that's helpful! Big FANX!

[@MrBademy]

Great video, epic info, big appreciation

[@AI.Life.encorp]

thx mate, you're my hero :D

[@MarkoKacanski]

Thank you. Very usefull. Great video!

[@Ahtraihue]

Thank you! very helpful, will not forget it when I am able~*

[@koa8299]

this is underrated. ken carson's music vids had me itching to learn ai

[@ficky5609]

Hi. Thank you for making these, I learn a lot from your videos. Maybe you can add a very basic tutorial in the future. I think it will be great if you can explain the technical terms like seed, or step..

[@mattjpellerito]

terrific video

[@ia_technojnoun]

très bonne vidéo explicative !

[@loneranger4441]

Really appreciate you sharing nuances and pitfalls that one may encounter. Love the details. Thanks again.

[@al3030]

Nicely edited and explained! I would just reduce the text a little bit or let it linger a bit more on the screen to give us enough time to read without rewinding multiple times.

[@xaviersykora]

Why the fuck aren't you even at 1K subs!!! This is an incredible tutorial. I've followed a different tutorial that explained just as much, but it had nowhere near the level of badass editing and well paced explanation with visual elements to support. Excited to see you blow up at some point!You'll definitely get there if you keep this up 🌄

[@fictitiousness]

Thanks so much! Yeah, i'm just getting started and need to get into a regular uploading schedule but will definitely keep uploading as much as i can without sacrificing editing quality ❤️

[@Staiks]

Man you are insane, crazy how cohesive your entire "lesson" was. If I were you i'd make a patreon for more in-depth stuff, while AI generated stuff is booming as you deserve it! Cheers

[@fictitiousness]

Thanks, appreciate it! I'm still not expert enough in this whole field to charge anyone for my "knowledge" :) i'm quite happy to grow a channel with an audience first and it also helps me figure out all that stuff while making the videos. Might add a ko-fi link, if someone wants to donate a buck.

[@Staiks]

@@fictitiousness A Ko-fi is a good idea! Completely should, make the patreon later on, good luck man!

[@charapoKKo]

​@@fictitiousness honestly, it has nothing to do with expertise and everything to do with effort. Nobody is an expert in this field. It's too new. You're spending time tweaking settings, and then spending even more time gathering what you've learned and presenting it in a digestible way. So others don't have to spend as much time as you did tweaking all the settings, to learn proficiency. That time is worth money. I would gladly subscribe to your Patreon if you generated even ONE tutorial of this quality every month.

[@charapoKKo]

​@@fictitiousness also another point to remember. You're at a challenging intersection of art and technology. Most folks are typically going to come from a background that's stronger in one area or the other. Your ability to communicate the available tools to a generalized audience is huge. Your technical explanation is clear enough that "artists" can follow along, but not overly descriptive so "technologists" are belabored by it. You're making this technology easier to adopt for large groups of people with separate and overlapping levels of experience. If you enjoy creating this kind of content, I highly recommend considering some kind of modest subscription based offering. While I won't complain about getting it for free, if compensation helps to motivate you to continue creating these and the quality is consistent (or improves) - I would be all for it.

[@loneranger4441]

Loved it. Excellent video. Please keep them coming. I have pressed the bell icon. Can’t wait for the next video to drop. Thanks so much for making this video.

[@dustin3118]

With the AI voice to keep the theme, nice.

[@ffxsam]

Quick note: you might have to add --enable-insecure-extension-access to your command line arguments to install Deforum.

[@bhblueberry]

Probably inserting keyframes for propagation control would help a lot. For drawers it is a big difference between drawing 1 or 30 frames :)

[@fictitiousness]

yup, i will introduce some tools for better visual keyframing in the next video. There's also a discussion on github about adding the function to the gradio UI, so maybe it will happen soon :)

[@endryu.creator]

Ти найкращій, я вже 5 днів сам тести проводжу, а тут твій канал, привіт тобі з України!)) You're the best, I've been doing tests myself for 5 days now, and here's your channel, greetings from Ukraine!!))

[@ogrekogrek]

thx.

[@agataLaeta]

thank you so much for this fantastic tutorial! could you please let me know how you interpolated between a 12fps video to 24fps video? did you do it with deforum interpolation mode and video_input or through some other method?

[@fictitiousness]

Thanks! Sure, the interpolation app is called Flowframes and it's free to use

[@suzanazzz]

Hi! Thanks for the awesome tutorial, it's probably the best and most detailed I Found! I have been running Deforum collab Notebook and I am now looking into Deforum in Auto1111. It's still confusing as there are new parameters to experiment with. I need some guidance for the following if that's easy for you: At 3:23 you talk about importing the settings from a .txt file. I cant find a way to do that.. 1- the notebook is updated so that tab is not there any more 2- I am running Fast Auto1111 with google collab so the .txt is in my Outputs Folder - when I right click I do not have the option -> copy PATH Would very much appreciate your help on how to load my settings. Happy holidays and thanks again ❤

[@edoardoottone6548]

Great Video, finally a good resource to look at. May I ask you which tool are you showing at 6:45. Thank you

[@fictitiousness]

which one do you mean, the math graph? :) It's graphtoy.com

[@edoardoottone6548]

@@fictitiousness yes was at 6:56. thank you. I would like to ask more question. If there is a place where I can contact you would be great but is ok for now.thanks for replying

[@fictitiousness]

Hit me up on discord: AffenBrot#0292

[@hopps4117]

Great video thank you. Do you have any tips as to why when I use a init image, I can't get it to be part of my animation? I've tried a high strength schedule but no matter what I do it seems like I cant get my init image to have an effect, any help would be appreciated!

[@fictitiousness]

The init image is being used as the first frame of the animation and doesn't play a role after that. If you want it to influence the video for the whole time, you could make a video out of the still image and use video input mode for your animation.

[@niatro]

Hi great video, looks like zoom and angle are just 2d operators and no have effect in 3d animation. right?

[@fictitiousness]

Yes, you are right! This is being wrongly displayed in the UI, but i will be sure to explain that in the next video. Thanks!

[@museinchaos3014]

Thank you for an amazing video. I am having trouble adding the deforum extension. I get an error message every time I try. Can anyone here help me with it? Thank you.

[@Xishnik94]

is there a way to reduce blur after the first frame?

[@studio85amsterdam]

hi thanks, i can not get it to find the input movie i have placed for "video input" even after I placed it in a folder in the stable diffusion webui folder and gave the location respective to this folder. any ideas?

[@findingforeverness8569]

amazing stuff, cheers for this, I see you were using runpod, do you know how to link audio/init images, I tried to copy the path from my contents folder as well as copying the web link but no luck?

[@fictitiousness]

should be working, if you link the right path (rooting from the sd-webui folder), check the other comments where i gave an example how to correctly link the source path :)

[@jourdanbiziou853]

This is really great! thank you so much for the great info:) Could you please explain the : Seed_Scheduler a little more. The Syntax and how to add Seeds over time ? been trying for a week to get this to work:( thanks again:)

[@fictitiousness]

Sure, basically just like in the example. Set the seed behavior in the Run settings to "schedule" and set your seeds you want to use in the "seed_schedule"-input field like this: (0) 555, (10) 666, (40) 888. Then the animation will these as a fixed seed on these points. Frame 0-9 = 555, 10-39 = 666, 40-... = 888. Hope that makes it more clear.

[@DeliciousBrain12]

@@fictitiousness Thanks a lot dude! I'm at the moment writing my bachelors thesis about AI generated art and you couldn't help me more with your videos

[@fictitiousness]

@@DeliciousBrain12 Thanks, really appreciate it! There's actually a section about "t%" in the official deforum quick guide explaining it. This would be diving into math expressions which i wanted to cover in a future video :)

[@pablo6570]

where do I specify path to sequence of images or the video file for video mode?

[@3pm479]

where can i check the deform animation u created and used in the video. They are so cool. thankx for this series.

[@fictitiousness]

Sure, the prompt is "photography of a brutalist building, black and white, symmetric, grainy, detailed, volumetric lighting" with seed 666 :P You might get a slightly different starting frame, because i've used a custom hypernetwork for these, should be close though.

[@3pm479]

@@fictitiousness thank you 😊, so kind of you, please also share the prompt for animation at 2:50

[@fictitiousness]

@@3pm479 It's not my generation, so no idea. Best check with the user in discord (credited right below the video)

[@3pm479]

@@fictitiousness I see, thank you.

[@urisrssfeeds]

how do you make it smoother w/ interpolation later?

[@LouisJDF]

fantastic video, thanks. One question I have is that my automatic1111 UI looks very different, with a lot less options. for example there is no seed behaviour option in my run tab. would b much appreciated if u could help.

[@fictitiousness]

Sure thing, you can find seed behavior in the "Seed" tab now :)

[@LouisJDF]

@@fictitiousness ah ur a legend, thanks mate

[@LouisJDF]

@@fictitiousness also forgot to ask, how do u turn on dark mode?

[@fictitiousness]

@@LouisJDF add "/?__theme=dark" after the IP address in the url

[@ene_77]

excelent video any black margic sorcery to resume an input video animation?

[@fictitiousness]

Thanks! If you have already created the video with deforum, you have a timestring value in the settings file, that you can input with the same settings and it should resume the video. If it's a completely new video, i would just input the last frame as an init image and work with different settings from there.

[@PlayerGamesOtaku]

hello, could you redo the tutorial? Because I noticed that the version I have is very different from the one in the video, and I'm having trouble since it's the first time I've used such a program

[@oboy9090]

Hi, just installed deforum into SD and I am wondering if there is a way to upload personal videos. When doing via colab there is an upload video given but, not in the local install.

[@fictitiousness]

Yes, you can set a link to a video in the "Init" tab in deforum

[@hitlab]

Does anyone know if this version of Deforum is accessible via an API? I have a project that I’m working on currently that uses another version of a local Deforum install and I can integrate it into my own Python code but the version I’m using is older.

[@ModRot]

thank you for this amazing tutorial! im having a simple issue that i cant solve - how do i load previous settings? i tried to write the folder and the file name in the load setting textbox in many different ways but no matter what i tried it gives me an error ..

[@fictitiousness]

Hey, it may be that the settings file you're trying to upload is from an older version and that gives you an error if there are new settings included in the new file. Try generating a new file and compare it with the one you're trying to upload.

[@cyndydee5302]

Can i ask why Automatic1111 logo have picture of Hochiminh on it?

[@enilenis]

If anyone needs a camera keyframe exporter for Maya to Deforum, I've written one, though I wrote it under standalone installer, not the Auto1111 revision, and I think the rate of 3D transformation doesn't match on the Z axis. I'm collecting feedback on how to improve the exporter.

[@Xandr_Nekomata]

how to get this dark theme in SD 1111 ?

[@AllTheCuties]

hey, loved the video! I have a question tho. When switching to 3d it downloads some stuff and then I get an error render_animation(args, anim_args, root.animation_prompts, root) File "C:\Users\xxxx\stable-diffusion-webui\extensions\deforum-for-automatic1111-webui\scripts\deforum_helpers ender.py", line 75, in render_animation depth_model.load_adabins(root.models_path) File "C:\Users\xxxx\stable-diffusion-webui\extensions\deforum-for-automatic1111-webui\scripts\deforum_helpers\depth.py", line 29, in load_adabins self.adabins_helper = InferenceHelper(models_path=models_path, dataset='nyu', device=self.device) File "C:\Users\xxxx\stable-diffusion-webui/extensions/deforum-for-automatic1111-webui/scripts/deforum_helpers/src\infer.py", line 85, in __init__ model, _, _ = model_io.load_checkpoint(pretrained_path, model) File "C:\Users\xxxx\stable-diffusion-webui/extensions/deforum-for-automatic1111-webui/scripts/deforum_helpers/src\model_io.py", line 39, in load_checkpoint optimizer = ckpt.get('optimizer', None) AttributeError: 'NoneType' object has no attribute 'get' Did anyone else experience this or know a sulution? Any help is appreciated!

[@fictitiousness]

solved? ;)

[@massive-duck]

@@fictitiousness yep, had to manually download it. For some reason it didnt work with aoutomatic download through the extentions tab

[@elbedouirayen7938]

hi please can u help me the? i can't dfefine the video input directory for the deforum thank u

[@fictitiousness]

Sure, just create a folder in your root sd-webui called "input" and name your video "input.mp4". Then you just enter the path in the "video_init_path" field like this: input/input.mp4. This should work, i think the main issue is the first "/" in the default path example which will give you the error.

[@TroytCoburn12345]

Do you Have any links for "Kewk Lightworker"? You featured their work in this post. Cant find on Discord. Thanks!

[@fictitiousness]

Which one do you mean? Exact timestamp please :)

[@TroytCoburn12345]

@@fictitiousness 5:23

[@wingleemanny]

what AI voice do you use to narrate your videos?

[@fictitiousness]

descript.com

[@dazlava]

strenght_schedule makes difference using video input?

[@fictitiousness]

It should, because it's the only strength setting you have to determine the influence of the image. I will do some more testing though and correct, if i should be wrong.