freQuensy23's picture
Update README.md
165d8c4
metadata
license: apache-2.0
dataset_info:
  features:
    - name: date
      dtype: timestamp[ns]
    - name: rebellion_info
      dtype: int64
  splits:
    - name: train
      num_bytes: 140256
      num_examples: 8766
  download_size: 79236
  dataset_size: 140256
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
task_categories:
  - token-classification
language:
  - ru
  - en
tags:
  - seq2seq
  - historical
  - russia
size_categories:
  - 10K<n<100K

FULLY OPENSOURCE NOW (apache license)

Dataset short info

This dataset contains all the information about the mutinies or military coups that took place in the 21st century. The dataset has about 10 thousand records and has been compiled for a long time. In addition to the archival value, this information can be used to train models to work with sequences (RNN, LSTM etc) - for example, seq2seq models.

Data format

file: data/train-....parquet

  - date: str - date string in default python datetime format (ex. "2000-01-01T00:00:00")
  
  - rebellion_info: int64 - did a rebellion happen then (1 yes, 0 - no) (ex 0) [other 18,446,744,073,709,551,615 int64 values can be use as labels in the future]