BFFusion (❁´◡`❁)

Integrating diffusers Functionalities within Blender

bffusion720_.mp4

GP+control_net_resize.mp4

diffusers_blender.2.mp4

Inference Time

• Less than 1 second

Features

-Rendered Image or Animation Conditioning

-Lora and CheckPoints

-ControlNet Pipeline

-Noise, CFG, Sampler, Schedulers ... the usual

just copy paste any path model from hf

BFFusion (●'◡'●)

Blender Scene

Img2Img

ctrl_net:

-canny / -depth / -normal

++

GPFFusion (❁´◡`❁)

Render

outputs with different Loras and combinations

AnimFFusion ᓚᘏᗢ

The addon supports also video generation from video inputs, but it´s still limited and currently in development.

AnimFFusion_saturn.mp4

kick_example.mp4

Setup

1. Download Blender:

   • Download Blender from the official website.
   • Use the zip version.

2. Prepare Blender:

   • Rename the downloaded folder to your preferred name and unzip it.
   • Navigate to the python folder within the unzipped directory and delete it.

3. Python Installation:

   • Ensure you have Python 3.10 installed on your device.

4. Install CUDA:

   • Download and install CUDA 12.1 (or newer) from the NVIDIA CUDA Archive.

5. Install Dependencies:

   • Open a terminal and navigate to your Blender folder.
   • Run the following command to install the required Python packages:

     pip install -r requirements.txt

6. Install PyTorch:

   • Visit the PyTorch website.
   • Select and install the PyTorch version compatible with CUDA 12.1 (tested version 2.2).

7. Start Blender:

   • Launch Blender from the executable within the unzipped folder.

8. Install the Addon:

   • Either install the Python script from the Addon Preferences menu in Blender or run it as a script directly. Both methods should work.

You're now ready to use BFFusion!

Performance

• Timing based on an NVIDIA RTX 4070 12GB GPU

Integrated using the diffusers docs

https://huggingface.co/docs/diffusers/v0.27.2/en/index