metadata
dataset_info:
features:
- name: text
dtype: string
- name: meta
struct:
- name: perplexity_score
dtype: float64
- name: pile_set_name
dtype: string
splits:
- name: train
num_bytes: 484845334.4
num_examples: 80000
- name: validation
num_bytes: 60605666.8
num_examples: 10000
- name: test
num_bytes: 60605666.8
num_examples: 10000
download_size: 329390472
dataset_size: 606056667.9999999
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*