metadata
dataset_info:
features:
- name: speaker_embeddings
sequence: float32
- name: input_ids
sequence: int32
- name: labels
sequence:
sequence: float32
splits:
- name: train
num_bytes: 11280403594.14842
num_examples: 76392
- name: test
num_bytes: 1253525841.8515804
num_examples: 8489
download_size: 11731748054
dataset_size: 12533929436
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*