This is a simple ComfyUI workflow to generate a new EPIC animated profile picture based on the input profile picture and the prompts.

Inputs :

• Profile picture to imitate

• Resolution (512 by default)

• Profile picture seed

• Profile iterations

• Background animation seed

• Background iterations

• Number of frames for the animation

• Frame rate

Models used for the example :

• 3DAnimationDiffusion : https://civarchive.com/models/118086?modelVersionId=128046

• LORA 3DMM_v12 : https://civarchive.com/models/73756?modelVersionId=107366

• ControlNet : https://huggingface.co/lllyasviel/ControlNet-v1-1/tree/main

• AnimeDiff motion v15v2 : https://civarchive.com/models/108836?modelVersionId=159987

A few notes :

• I suggest you put the seeds to "fixed" during generation and increment it manually to get the result you want without having to regenerate everything each times

• Feel free to change the sampling methods of the picture and background generation to fit your needs

• Change the CFG for diverse results

• You can adjust the background remover algorithm and threshold if it does not work for your profile picture

• You can change the upscaling easily for the output

• Select WEBP, GIF, WEBM or any other format in Video Combine at the output