metadata
dataset_info:
features:
- name: sample_number
dtype: int64
- name: abstract
dtype: string
- name: article
dtype: string
- name: compressed_article
dtype: string
- name: compression_time
dtype: float64
- name: original_tokens
dtype: int64
- name: compressed_tokens
dtype: int64
- name: token_saving
dtype: string
splits:
- name: train
num_bytes: 454265077
num_examples: 28257
download_size: 231070985
dataset_size: 454265077
configs:
- config_name: default
data_files:
- split: train
path: data/train-*