Single image to animation update with ComfyUI !
Finally only using sdxl to keep faces coherence, but still the mess haha. Promise I will share soon the workflow, still want to make it better ! Still using LLM to describe the image and instantid ( also why i dont use SD 1.5 )
#AI
Starting from a sketch of 1900 to transform it using ComfyUI ( AI ) and integrate it automaticly.
I am presenting this workflow Friday 19th at 3pm PST on the twitch of
@HelloCivitai
. Hope you will be their !
I am using ComfyUI to do sketch2img, imgto3D, 3D2img, integration
Yes Yes Yes ! 3D is coming also in ComfyUI ;) !
Very interesting library from
@MrForExample
(amazing job!)
That is just an example of image generation from single view, but there is also tool to manage directly 3d objects and texture.
I manage to make
With
@Dogstudio
comfyUi crazy 3D animation workflow One prompt > SDXL image génération > instantmesh and zero123 to generate 3D > upscale of texture > animation of orbit caméra > vid2vid with SD 1.5 LCM > latent upscale > boummm haha
With Amazing node of
@MrForExample
Always more ComfyUI workflow, here something that I create to integrate elements easy in a isometric game fully automaticly. It convert a single image to a 3D object, allow to select the good view, and integrate it inside an image. Its working for everything from a single prompt
AI VIDEO animation on Game Boy Color ;).
I did that quickly using an old Ai animation made with
#comfyui
and convert it for GBC.
Lets make it interactive haha
How to create this 3D animation for free in 10min without knowledge in 3D ?
1- Generate an image of character using the prompt "[SUBJECT], T-Pose 3D, full body" in
@midjourney
or other
2- Generate your 3D character using from
@tripoai
, export in custom
With
@Dogstudio
, I created this 3D animation only using AI and automatic tools !
ComfyUI with SDXL, Comfy-3D-Pack and TripoSR to generate the images references and the 3D.
Blender Plugin to batch vertex color to texture
Mixamo to rig and animate.
C4D and Octane for the little
🚀 Dive into the future with my blend of 3D sound & AI in animation, a passion project at
@dogstudio
! 🎧🤖
Crafted with Cinema 4D, finessed with ComfyUI & AnimateDiff.
A huge shoutout to those who provided invaluable learning resources! 📖
@PurzBeats
,
@8bit_e
and
@c0nsumption_
First time using Magic Animate at
@dogstudio
! I created my own Dense Pose using Cinema 4D, I used ComfyUI , SDXL and controlnet, to generate visual. And finally MagicAnimate for the result.
I think face are not good yet, but body are pretty interesting ;) !
How to create a 3D model from a sketch and how to integrate it directly fully inside ComfyUI ?
Its what I will present tomorrow ;) with Civitai ( link coming soon ) at 3pm PST time.
Here more result for people that follow me yesterday :)
I think Relight of is very powerfull.
I wanted to show it was also possible for video inside ComfyUI ;) . Its not perfect, but its one little night of work.
I created this workflow that is fully automated, only tracking camera for background is
So I tested for me the 3 best 3D AI generator from single image :
-
#Rodin
Gen 1 from
@DeemosTech
-
#Tripo3D
fromn
@tripoai
-
#3Dcsmai
from
@CSM_ai
Notice that this is just two examples, and each tool can be configured or use different inputs to create a better result.
With
@dogstudio
, I created this new workflow in ComfyUI, that generate for you 3D object from Single prompt and tried to make it better than before. I use Comfyui-3d-pack from
@MrForExample
and I modify the workflow to latent upscale and bring "correction" to the image
From Single input image to cool animation with ComfyUI.
Transform your friend into an animate !
Its combining a lot of node, its using LLM (Gemini), SDXL, SD1.5, animatediff etc...
yes its a mess haha, but didnt have time to clean up yet, and its not perfect yet as a workflow
Really really impress by the quality of
@tripoai
! The first image is generated with midjourney, and next its result with Tripo without any retouch.. Congrat to the team !
I made it work TripoSR inside Comfy-3D-Pack from
@MrForExample
. And all that directly in Google Colab !
Generation take half a second with a A100 !
@StabilityAI
@c0nsumption_
Lets see if I can create animation from it now !
Trying at
@Dogstudio
, to create a Gpts with the new chatgpt from
@OpenAI
that would do "WebGL". I provided the full
@threejs
doc to make it work.
And the result its just after only one message !
Not bad, but we still need our work to make a 3D webgl website great :) . Let's try
🎵🎵 Exciting news! My 1/5 SHOWREEL about WebGL Project at
@dogstudio
, made in 2 min through AI & automation, is out 😅!
You are lazy like me ? Just use my script that automate editing synchro to audio BPM in After Effect ! 🔔🔔
🔗 Download the After-Effect script .jsx:
Sora is it magic ?
I didnt see the video here, so I decided to repost it from instagram
@OpenAI
.
Its the first time in a video that you can see that Sora can also produce bug. Really happy to see that, as its mean more how the tech would be than totally perfect rework videos :)
Really impress by the render and smoothness of
@sketchfab
inside
#Quest2
.
#WebVR
is clearly here.
Here in the video , a photogrammetry model that I made a few month ago,
You can even jump on table, and that is cool !
Sora is it magic ?
I didnt see the video here, so I decided to repost it from instagram
@OpenAI
.
Its the first time in a video that you can see that Sora can also produce bug. Really happy to see that, as its mean more how the tech would be than totally perfect rework videos :)
You want to test ComfyUI-3D-pack from
@MrForExample
but its a bit long to install ?
I rewrite a Google Colab from
@dreammachineai
to totally work with this amazing node !
have a look here
And thanks again
@MrForExample
for this
Procedural Building generator from single image. Still not perfect but pretty surprise by the quality of this tool and the impact it could have on real estate. The source is here :
Pretty hard to make it work, don't hesitate if you need help to contact me.
I am looking for a dev unity in freelance in Paris for all august. Its for a big project of an immersive formation for Caretaker. If you dont like holiday in August contact me :)
@Skydeunnn
J'ai mon entreprise, je paye des impôts en France. Je pourrais payer 0 en allant à Malte. Mais ethniquement je trouve ça triste et irrespectueux des institutions françaises qui m'ont aidé à être là aujourd'hui. C'est normal qu'il soit critiqué pour ça.
I worked during 6 months on the 3D part with the great team
@Dogstudio
! Its finnaly out ! I will share you soon more render of my own work for the project ! Glad i can finnaly speak about it !
After a year of intense collaboration with the
@expo2020dubai
teams, the most ambitious and challenging project we have been lucky to ever collaborate on is now public: an interactive content platform paired with a
#webgl
3D replica of the actual expo site in Dubaï.
[ thread👇]
The best way that I found to Make animation for webgl is to upload any type of 3d file and animations to sketchfab and export it from it in gltf. Its not conventionnal. But never found any other convert that dont have bugs for some type of files or animations.
Because since COVID remote work was very good for my own company. I will travel to Barbados for one year starting in April. It will be a good test to see if I can manage everything from the beach :). Barbados give remote work visa if you Want to go there as well ;)
Laval Virtual will be between the 22th and 24th april in this virtual place , it's simple and free. I am also very happy to speak at a round-table about cross-tech in XR, Wednesday 22th at 11:55am. Let's meet there :). by
@lavalvirtual
&
@VirBELA1
I went to so Many event in San Francisco about AI. And its crazy how 99% of companies only focus on LLM. I rarelly even meet people or companies that know what is ComfyUi or what is visual AI. It shows also all the potential of it that people dont even expect or imagine.
@Dogstudio
@PurzBeats
@8bit_e
@c0nsumption_
Here my little workflow :
- Cinema 4D with mograph sound effector animation
- Octane render for depth map
- Comfy Ui and animatediff
- Topaz AI VIdeo
- Final Editing with After effect ( just changing a bit the color )
As I am 3D designer, Nomad Digital and sometimes teacher :) , I decided to give free 3d lesson to public university in every country i am travelling to. Exchange is life :) !
#ComfyUI
#Comfy3D
Comfy3D Update:
- Integrated Era3D multi-view diffusion model
(Work with InstantMesh 3D reconstruction model, see video below)
Cat image from: well, I forgot😸
Project GitHub page:
🎊 🎊 Super happy to announce that I will be juror at Dubai’s inaugural Artificial Intelligence Film Festival ! 🎬🌐
I made this from the logo of
#AIFF2024
with
#ComfyUI
&
#AnimateDiff
!
You've got 15 days to submit your AI films :
@expocitydubai
Heeey, I found this funny workflow using
#stablediffusion
#ebsynth
and
#realitycapture
to generate a 3D model.
I think it could be useful and there is a lot of cool way to use it !
I am very proud to show one of my work realise with
@makemepulse
in freelance. I mix facial and body mobcap, clean them, and prepare them to render for
@assassinscreed
. It was a great team work !
Merry Christmas ! With
@Dogstudio
, I create this new AI video workflow interacting with sound .
I took first the beautiful song of "introduction in the snow" by "Miracle Musical". I used touchdesigner to create an animation in black and white that is generated from this music.
I will be at
@lavalvirtual
between the 20 and 24. And for the first time with my own stand to show you all my creation and my future company.. keep in touch ;)
@FranckPicabia
@lemondefr
Haha ça me rassure quand même ! Même aujourd'hui il y a quand même toujours des gros c** qui sortent lexcuuse de la grippe. Je pense que vous serez poursuivis pour votre tweet et quand vous aurez choper le virus vous aurez l'air malin.
I'm sharing a tool that we made a few month ago to train caretaker.
I worked on the design and technical part with thewalkingnerds, recruting and managing a team of 5.
The best is that the entire tool (mini-game, users etc..) is configurable using a simple Wordpress BackOffice.
Hey everyone!
Tomorrow, Friday, May 24th at 10 AM PST (7 PM CET), I will be presenting my first Masterclass on ComfyUI: GenAI, Animation, and 3D, in partnership with
@Dogstudio
( that always support me whatever weird AI test I am doing 😅 ) and
@future__spaces
( with the
@bilawalsidhu
But i imagine they did balloon head because there IS not consistency yet in character with Sora. So its also you Need to take into account the tool you have
Probably the best quality projection that I ever seen !. It's at
@ArtechHouse
in
#NewYorkCity
.
Really beautiful fractal art / voxel
The artist is Julius Horsthuis and He use the software
#mandelbulb3d
My first ever photogrammetry that I made fully myself. And to make it crazy, its a full room.. More than 2000 photos and retopo of it. Pretty happy of the result for a first try :)
🚀 Just released a presentation on using Open Source tools like comfyUi & Stable Diffusion for AI animations! Unlike Runway or Pikalab, these offer full control 🎛️. Big thanks to
@Dogstudio
for their support in my AI journey. Check it out for unique AI insights! 🌟
#AI
#Animation
@Rahll
@javilopen
You will be sad soon. Now Almost all New model on Stable Diffusion are not trained with Real visual anymore. But they are trained with the result from différent other models ;). And soon it will be even more complex. So it will be impossible to know exactly what is the source
After create multi script for C4D, create a trading bot for crypto :
I am starting to create a full
#WebGL
website in
#threejs
only using
#midjourney
and
#chatgpt
! More news about it soon ;)
I recreate fake 3D on single 2D image automatically thanks to some machine learning code and depthy for the mix of depth map. Not perfect, but no human needed.
After some test with
@Dogstudio
, pretty impressed by the quality output of 360 people from a single photo with this .
Lets try to use Nerf to transform it in a proper 3D.
It was my first discover of animatediff , it was not even available in comfyui AT the Time ;) , and I didnt know it will be a tool that I will use almost everyday in my work
The exposition
#behindthegame
about
@assassinscreed
start today in Paris ! Very proud that I work with
@Soixanteci
in freelance for an interactive installation where I made the 3D model.
Today, we’re thrilled to announce the open weights for Stable Diffusion 3 Medium, the latest and most advanced text-to-image AI model in our Stable Diffusion 3 series!
This new release represents a major milestone in the evolution of generative AI and continues our commitment to
The year start good ! I am really happy to be invite by Chinese Academy of Press and Publication to do a presentation for World Conference on VR Industry in Nanchang, China. It will be the 20th October . I will let you know soon for me info
Wowowo I try it a bit earlier and I am so happy that my friends create a so nice aventure. So beauty , and I am happy to discover all of it in one Time !. Amazing ;)
@alexlesubmarine
@FabMotte
@Pat_Hg
and others :)
🌟 Excited to join the jury at the AI Film Festival AIFF
@expocitydubai
! 🎥🤖 Mark your calendars for 29/02/2024. Beyond thrilled to showcase a quick reel of my AI, coding, and 3D work combined. 🚀✨
Though a tad disappointed to miss out on the panel talk 🗣️💔, being part of
For the last 2 months, I have been patiently waiting for a chance to reveal two AI-assisted films we created
@Dogstudio
. Hopefully will get clearance soon.
This one was 600 hours of production for 2 minutes, so who said AI was just pressing a button?
Meanwhile, here's a GIF.
As everyone , I am pretty impressed by Dream Machine from
@LumaLabsAI
. I guess the training data are also coming from 3D data as NERF, and as they are specialized in that, I tried to reverse a shot generate from Dream Fusion and use it inside LumaLabs capture generator for a