metadata
dataset_info:
features:
- name: cls_embeds
sequence: float32
splits:
- name: train
num_bytes: 200210688
num_examples: 65088
download_size: 240628473
dataset_size: 200210688
configs:
- config_name: default
data_files:
- split: train
path: data/train-*