This dataset can be used for translation or token classification tasks. There are two versions: 'nl-st' contains over 1.2 million records and 'nl-st-lg' contains over 9.8 million records. Each record has 6 features:
sentence (string) - natural language (English) sentence that describes the state.
state (string) - state information consisting of percept value pairs stored as a string (percept value)
ner_tags (string[]) - NER tags for… See the full description on the dataset page:
https://huggingface.co/datasets/cw1521/nl-st.
