0:00
/3:25

In the world of 3d art, having the right platform integration can make all the difference.

Today we'll be using Hyper Stable Diffusion in ComfyUI and Blender. Blender is renowned for its capabilities in creating 3D models and images. You might wonder how to integrate ComfyUI and Blender. By the end, you'll have a clear understanding of how to leverage these two in real-time. Lets get started.

What is Hyper Stable Diffusion? 

ByteDance has demonstrated its dedication to both speed and innovation with the introduction of Hyper-SD. It is designed to speed up generation time significantly.

0:00
/1:34

With increased speed, it also ensures that images are sharper, more detailed and visually appealing.

ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows the comparison with SDXL Lightning
Credits: ByteDance Hyper-SD

The image above is a comparison between Hyper SD and SDXL Lightning using 1 step. Tests show that Hyper-SD has better quality and works faster than earlier models such as SDXL-Lightning.

What is Blender?

ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows the interface of the blender

If you're looking for 3D software, one option is Blender. It is a free and open-source software for creating 3D models, animations and images. It allows users to build, shape and paint objects in three dimensions, making it a powerful tool for artists and designers. With Blender, you can create detailed characters, elaborate scenes and stunning effects for movies, video games and digital art. The software is supported by a passionate community that shares tips, tutorials and plugins to help each other improve their skills.

Blender and Hyper SD in ComfyUI

ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows the comfyui that uses a screenshare node which connects to the 3d software

Ever wondered how we can integrate Blender into ComfyUI? ComfyUI can connect to Blender with the custom node MixLab. This seamless integration ensures consistent and reliable results, which are essential for any professional projects. We would like to credit Andrea Baioni for his outstanding work in the development of this ComfyUI workflow where he combines the Hyper SD base model, Blender and ComfyUI.


How to Use

ComfyUI - One-Time Setup

💡
Download the workflow and drag & drop it into your ComfyUI window, whether locally or on ThinkDiffusion. If you're using ThinkDiffusion, it's necessary to use at minimum the Turbo 24gb machine, but we do recommend the Ultra 48gb machine.

Custom Nodes

If there are red nodes in the workflow, it means that the workflow lacks the certain required nodes. Install the custom nodes in order for the workflow to work.

  1. Go to ComfyUI Manager  > Click Install Missing Custom Nodes
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows where to install missing custom nodes
  1. Check the list below if there's a list of custom nodes that needs to be installed and click the install.
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows the area where the list of missing custom nodes appear

Models

Download the recommended models (see list below) using the ComfyUI manager and go to Install models. Refresh or restart the machine after the files have downloaded.

  1. Go to ComfyUI Manager  > Click Install Models
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows where to install models
  1. When you find the exact model that you're looking for, click install and make sure to press refresh when you are finished.
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows where to search the list of models

Model Path Source

Use the model path source if you prefer to install the models using model's link address and paste into ThinkDiffusion MyFiles using upload URL.

Model Name Model Link Address
Hyper-SDXL-1step-Unet-Comfyui.fp16.safetensors
📋 Copy Path
depth_anything_vitl14.pth Pre-loaded
OpenPoseXL2.safetensors
📋 Copy Path
sai_xl_depth_128lora.safetensors
📋 Copy Path
sai_xl_canny_128lora.safetensors
📋 Copy Path
t2i-adapter_diffusers_xl_lineart.safetensors
📋 Copy Path

Guide Table for Upload

Recommended Models

Node’s Value Name

Node

ThinkDiffusion Upload File Directory

Hyper-SDXL-1step-Unet-Comfyui.fp16.safetensors

MODEL

Load Checkpoint

…comfyui/models/checkpoint

depth_anything_vitl14.pth

ckpt_name

Depth Anything

preloaded

OpenPoseXL2.safetensors

CONTROL_NET

Load ControlNet Model

…comfyui/models/controlnet

sai_xl_depth_128lora.safetensors

CONTROL_NET

Load ControlNet Model

…comfyui/models/controlnet

sai_xl_canny_128lora.safetensors

CONTROL_NET

Load ControlNet Model

…comfyui/models/controlnet

t2i-adapter_diffusers_xl_lineart.safetensors

CONTROL_NET

Load ControlNet Model

…comfyui/models/controlnet

Reminder
💡
Refresh or restart the machine after uploading the files in ThinkDiffusion My Files.
💡
You can upload models by copying the link address of download button/icon from Civitai or Huggingface and paste into the Upload section of ThinkDiffusion My Files using the copied URL.

Blender - One-Time Setup

  1. Download Blender here https://www.blender.org/download/. Read about the Blender system requirements here  https://www.blender.org/download/requirements/
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows the download section of blender website
  1. Install Blender onto your device
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows the installation instruction
  1. Go here https://www.mixamo.com/#/?page=1&type=Character Choose 3D characters in Characters tab then Click the Animations tab to choose what kind of movement you want for your character.
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows on where to download 3D character.
  1. Download the Animated Character as FBX for Unity or you may choose any FBX version to test the compatibility of your version of Blender.
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows  on what file that you'll need for
  1. Open Blender and delete the default object Cube. Click the Select tool and click the cube and hit delete.
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows the default layout of the blender
  1. Import the .fbx file to Blender
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows on how to import a file
  1. After importing, you can see the Blender interface with its imported 3D model that will be used for the HyperSD workflow. For more info about Blender, you can read here Getting Started with Blender manual.
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows the uploaded 3d character
  1. The ideal perspective view of a 3D model character that will be used for the workflow.
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows ideal perspective view of character that will be used in this workflow
💡
Aside from Mixamo, you can use other sources for your 3d model such as:
1. free3d.com
2. cgtrader.com
3. sketchfab.com
4. 3dexport.com
and more...
💡
3D file types that are compatible for Blender and good to be used for this tutorial guide: ABC, USD, OBJ, FBX, PLY, STL
💡
This tutorial guide is not limited to human 3D character only. You can use other 3D models too such as architecture, vehicle, animals, sports, etc.

Procedures

Now that the hard work is out of the way, let's get creative. You need to follow the steps from top to bottom. The workflow is a one-click process after everything has been set up.

Steps Detailed Steps Default Nodes
Start Screen Share 1. Click Share Screen

2. In Pop-up dialog window, select Blender screen in its window tab.

3. Click the Set Area and create a box

4. Click Live Run to establish Live connection
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui shows how to screen share node
Set the Image Size ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui set the image size
Set the Base Model ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui set the base model as seen on the image
Write a Prompt Write a prompt of what you want to be in your 3D model ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui write a prompt
Set the Model Set the model as seen on the image. ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui set the model as seen on the image 2
Check the Settings and Run the Queue Prompt Set the step only in 1 and sampler name to lcm only. ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui check the sampler settings and run the prompt
Check the Image Preview Check the preview and you can test the 3d live preview by moving or rotating the view of the 3D model in blender. ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui check the image preview

Reminders

💡
If you get deformed generation results, just go back to Screenshare node to Set Area box again.
💡
If you don't use a character model and prefer other model such as nature, architecture, shapes, etc, disconnect the openpose controlnet node and use depth, lineart or canny.

Hyper SD + Blender Image Examples

Blender Text Prompt ComfyUI
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui superman input 3d model Superman, background is city buildings ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui superman output
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui spiderman input 3d model Spiderman, walking, background scary dungeon ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui spiderman output
ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui assassin input 3d model Assassin, jumping, background is forest ThinkDiffusion StableDiffusion ComfyUI img2img hypersd blender comfyui assassin output

Hyper SD + Blender Real-Time Examples

0:00
/1:56
0:00
/1:01

If you’re having issues with installation or slow hardware, you can try any of these workflows on a more powerful GPU in your browser with ThinkDiffusion.

If you enjoy ComfyUI and you want to test out creating awesome animations, then feel free to check out this AnimateDiff tutorial here. Happy memory creating!