metadata
configs:
- config_name: default
data_files:
- split: train
path: stackoverflow/train.csv
StackExchange Dataset
Working doc: https://docs.google.com/document/d/1h585bH5sYcQW4pkHzqWyQqA4ape2Bq6o1Cya0TkMOQc/edit?usp=sharing
The code to create this dataset is here.
File structure
Each folder represents the different StackExchange sites (~182). The largest one is StackOverflow.