Code release for pix2pred paper

This is the version of the code used to obtain experimental results for our paper "From Pixels to Predicates: Learning Symbolic World
Models via Pretrained Vision-Language Models"