[Pipeline] MimicMotion

[reallyigor]

Model/Pipeline/Scheduler description

MimicMotion is a recent paper on Personalised Image2Video generation. An input is a reference image and a driving video, motion from the driving video is taken to animate reference image.

The results are really great!:

Project page License is Apache 2.0

I would like to add this pipeline to the diffusers.

Open source status

• [X] The model implementation is available.
• [X] The model weights are available (Only relevant if addition is not a scheduler).

Provide useful links for the implementation

https://github.com/tencent/MimicMotion Authors: @gujiaxi

cc. @DN6 @a-r-r-o-w

[a-r-r-o-w]

I believe they already provide a Diffusers pipeline which means you can already load it with custom pipelines or directly copy and use. Usually, if a project already builds on top of Diffusers, we don't need to support it as a core pipeline. Maybe since this looks very promising and high usage (1.5k stars), it could be considered but better for @DN6 to decide on that

[reallyigor]

Usually, if a project already builds on top of Diffusers, we don't need to support it as a core pipelin

Got it 👍