Stayin' Alive with RAFT and ControlNet

Made with Deforum Stable Diffusion. RealisticVision 1.3 model.

- Two ControlNets, one for Pose, one for Depth
- RAFT Optical Flow from Hybrid Video
- No hybrid compositing.
- Video mask made with humanmasks feature in Hybrid Video tab.
- I also used my new Redo Generation feature at 2x. So, each generation with both controlnets was fed back on itself 3x each generation. This has the byproduct of reinforcing the controlnet conditioning if using controlnet.

"Original Stayin Alive dance - Aubrey Fisher x Vik"
https://www.youtube.com/shorts/F0LpaHcaRR8

#aiart #aiartist #deforum #stablediffusion

Deforum Stable Diffusion
http://deforum.github.io

🌲⭐🌃
Forest Star Walz
https://www.youtube.com/reallybigname
https://www.facebook.com/reallybigname
https://www.instagram.com/reallybigname
https://www.twitter.com/reallybigname
https://www.soundcloud.com/forest-star-walz