metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: embedding
sequence: float64
splits:
- name: train
num_bytes: 6284845810
num_examples: 871292
download_size: 5195767476
dataset_size: 6284845810
configs:
- config_name: default
data_files:
- split: train
path: data/train-*