Skip to content

lil-lab/scone

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 

Repository files navigation

scone

Data

All data is available in the data/ directory, whose subdirectories correspond to the three domains in SCONE (Alchemy, Scene, and Tangrams). Each subdirectory contains JSON files storing the processed train, dev, and test data. We also provide a script that generates action sequences for supervised learning (action_sequences.py).

Structure of the JSON files:

  • One line per file, containing a single JSON object with a list of examples.
  • Each example is a dictionary with a few properties:
    • identifier: A unique identifier for the interaction.
    • initial_env: The initial world state for the interaction.
    • utterances: Sequences of instructions. The instruction is the natural language instruction; after_env is the gold environment after executing the instruction according to the original data; and actions is a sequence of low-level actions generated from action_sequences.py and only used during the supervised learning setting.

Model

See model/. There are a few necessary steps to start running the code.

Requirements

You could install the requirements simply by

pip install -r requirements.txt

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published