metadata
dataset_info:
features:
- name: frame_dir
dtype: string
- name: index
dtype: int64
- name: clip
sequence:
sequence:
sequence: float32
- name: label
dtype: int64
- name: subject_name
dtype: int64
splits:
- name: train
num_bytes: 11837062684
num_examples: 4478
- name: test
num_bytes: 14250930296
num_examples: 5356
download_size: 8240985584
dataset_size: 26087992980
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*