TempoFunk

Make-a-Video like model using stable diffusion

Samples:

Version .1.8.19*:

Prompt: "A man dancing in the middle of the scene, coreography, beautiful, motion, adequate lighting, professional" Negative Prompt: "Disgusting, Bad lighting, Too bright, disco lights, Bad quality, Shaky"

We add extra temporal attention and freeze the rest of the unet, similar to make-a-video (?)

Models

Some models are available at chavinlo/TempoFunk/starry_pop

Credits

RedmondAI for compute donation

Based on lucidrains/make-a-video-pytorch and lxj616's make-a-stable-diffusion-video

Archive:

Version ???: