metadata
dataset_info:
features:
- name: completion
dtype: string
- name: source
dtype: string
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 37996032128.4282
num_examples: 229092
- name: validation
num_bytes: 4221836632.571804
num_examples: 25455
download_size: 630062871
dataset_size: 42217868761
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*