metadata
dataset_info:
features:
- name: hidden_state
sequence: float32
- name: label
dtype: int64
- name: id
dtype: int64
splits:
- name: train
num_bytes: 895707612
num_examples: 54603
- name: eval
num_bytes: 47145096
num_examples: 2874
download_size: 365734141
dataset_size: 942852708
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: eval
path: data/eval-*