Skip to content

Latest commit

 

History

History
17 lines (12 loc) · 1000 Bytes

README.md

File metadata and controls

17 lines (12 loc) · 1000 Bytes

Code for "Generating by Understanding: Neural Visual Generation with Logical Symbol Groundings" https://arxiv.org/abs/2310.17451

The program works for conducting rule learning on the Mario dataset. PyTorch and Swi-Prolog are required for code running.

Dataset Preparation:

Conduct Learning:

  • Running "main.py" in the root folder.
  • Path configurations need to be set properly based on the parser arguments therein.
  • More learning configurations can be set in the "exp_config/rule_learning_config.py" file.