A simulator-generated, multimodal benchmark for world models in contact-rich, frictional robotic manipulation.
DreamerBench provides time-synchronized RGB observations, camera-aligned contact-splat visualizations, proprioception, low-level control actions, and dense physics annotations (contact forces, friction coefficients, and contact-mode flags), together with precomputed discrete visual latents for token-based world models. It is the dataset used to train and… See the full description on the dataset page:
https://huggingface.co/datasets/zzhou292/DreamerBench.
