metadata
dataset_info:
features:
- name: speaker_embeddings
sequence: float32
- name: input_ids
sequence: int32
- name: labels
sequence:
sequence: float32
splits:
- name: train
num_bytes: 376533427.98253274
num_examples: 1442
- name: test
num_bytes: 42040140.017467245
num_examples: 161
download_size: 416197032
dataset_size: 418573568
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*