stick-figure
dancing-stick-figures
Dancing Stick Figures — v0.2
A small, fully-labelled synthetic video dataset for learning (and teaching) video diffusion on one consumer GPU.
1,340 clips · 6 s @ 20 fps · 128×128 RGBA · 482,400 frames · 134 text prompts × 10 seeds × 3 cameras ·
every frame carries the 3D skeleton, camera and G-buffer (depth, normals, part segmentation) that produced it.
Think of it as an MNIST for video generation: small enough that a 64² video diffusion model trains from scratch
in a few… See the full description on the dataset page: https://huggingface.co/datasets/sprited/dancing-stick-figures.humanml3d_stick_figureshumanml3d_stick_figures_5_frameshumanml3d_stick_figures_simplehumanml3d_stick_figures
