SV3D IMG to 3D and multi-view synthesis with Stable Video 3D in ComfyUI AI

Рет қаралды 3,948

4 ай бұрын

In this tutorial I walk you through a basic SV3D workflow in ComfyUI. SV3D stands for Stable Video 3D and is now usable with ComfyUI. With SV3D in ComfyUI you can create multi-view sequences for now. In the near future there will be the possibility to create 3D models with support for defining a camera path. This will give a whole new world of possibility's.
Please comment if you have questions or tell me your suggestions for future videos.
What you need to follow the tutorial:
Be sure to update your ComfyUI to the newest version and install the needed models which are linked below.
SV3D Models:
huggingface.co/stabilityai/sv...
More infos SV3D:
sv3d.github.io/
arxiv.org/abs/2403.12008
Connect with me:
/ _neuron_ai
/ _neuron_ai
The equipment that I use (affiliate links):
GPU: ASUS ROG STRIX RTX 4090 amzn.to/42ptUTi
#comfyui #stablecascade #stablediffusion #tutorial #animatediff #automatic1111 #aiart #ai #inpaint

Пікірлер: 10

@aa-xn5hc 4 ай бұрын

thank you, great. please, keep us updated on the development of this!

@BigSaur 4 ай бұрын

Can you use this with a 2070 GTX?

@neuron_ai 4 ай бұрын

It seems some people use it with 8gb. But I think it might be not enough vram.

@EM-yc8tv 4 ай бұрын

I'm using a 3050 Ti with 4GB of VRAM, so same amount as the 2070 if I remember correctly...I was able to get up to 512x512 resolution, and 16 frames generated. It takes about 70-90 seconds with 4 GB....doing the default settings shown in the video threw an out of memory error.

@BigSaur 4 ай бұрын

@@EM-yc8tvThanks for this, really appreciate it.

@tryzu6462 4 ай бұрын

So this is not a 3d model per se. Because they use NERF, making it a 3d model obj file, its not yet possible with NERF. right?

@neuron_ai 4 ай бұрын

As far as I understand it, for now images are generated and then converted to 3d or multiview for now. So no direct 3D. But dont nail me to that. I am sure direct 3D will be possible soon.

@thevoid6756 4 ай бұрын

@@neuron_ai how do I go from an image sequence to a 3d model i can manipulate in Blender for example?

@neuron_ai 4 ай бұрын

@@thevoid6756 Hey, I am sure it will be supported soon with SV3D in ComfyUI until then you can go the TripoSR way. I am just uploading a new video on this. Or the CRM way. You can find my video on this here: kzfaq.info/get/bejne/er6ZnLmh0Muvl4U.html Hope that helps you.

@thevoid6756 4 ай бұрын

@@neuron_ai thats the next video I watched after leaving the comment :D thank you though