metadata
dataset_info:
features:
- name: speaker_embeddings
sequence: float32
- name: input_ids
sequence: int32
- name: labels
sequence:
sequence: float32
splits:
- name: train
num_bytes: 367850467.6899811
num_examples: 1428
- name: test
num_bytes: 40958140.310018905
num_examples: 159
download_size: 406555793
dataset_size: 408808608
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*