Skip to content

Latest commit

 

History

History
17 lines (13 loc) · 489 Bytes

File metadata and controls

17 lines (13 loc) · 489 Bytes

generative-physics-inference

Slip or Not? Unsupervised Learning to Understand Physical Scene Using Multimodal Variational Physics Inference Network Zhengxuan Wu; Xiyu Zhang Stanford University

install

pytorch

run

python trainMVAE.py --visualize

dataset

You can generative datasets using this simulation engine. https://github.com/frankaging/2d-static-friction-simulation

multimodal

This model is based on MVAE. It could be used to do multimodal physics scene understandings.