metadata
dataset_info:
features:
- name: data_id
dtype: string
- name: prompt_ids
sequence: int64
- name: complete_training_sequence_ids
sequence: int64
- name: prompt_ids_len
dtype: int64
- name: traj_position_indices
sequence: int64
splits:
- name: train
num_bytes: 3461344541
num_examples: 250619
download_size: 703970632
dataset_size: 3461344541
configs:
- config_name: default
data_files:
- split: train
path: data/train-*