metadata
dataset_info:
features:
- name: llama_decoder
sequence:
sequence:
sequence: float32
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 16263406076
num_examples: 7753
download_size: 6568746989
dataset_size: 16263406076
configs:
- config_name: default
data_files:
- split: train
path: data/train-*