This folder contains all data and code for the NeurIPS submission 'A Benchmark for Systematic Generalization in Grounded Language Understanding'

* Data
All data can be found compressed in the folder 'Data Used in Experiments (compressed)'
Find all data additionally in an anonymous github repo: https://github.com/groundedSCAN/gSCAN_data and in https://github.com/lauraruis/groundedSCAN in the folder data in the .zip files.

* Model code
All model code + detailed README on how to run can be found in folder 'Model Code'
Find the same model code in a github repo: https://github.com/lauraruis/multimodal_seq2seq_gSCAN

* Appendix PDF
The pdf of the appendix can be found in the folder 'Appendix'

* Data examples
Animated data examples per split and adverb can be found in folder 'Animated Data Examples per Split and per Adverb'

* Data Generation Code
Code to generate new data, do error analysis on predicted data, or execute predicted commands to visualize them can be found in 'Dataset Generation'. This folder also contains a detailed README with documentation.
Find all data generation code in a github repo: https://github.com/lauraruis/groundedSCAN
