metadata
dataset_info:
features:
- name: speaker_embeddings
sequence: float32
- name: input_ids
sequence: int32
- name: labels
sequence:
sequence: float32
splits:
- name: train
num_bytes: 627391612
num_examples: 4944
download_size: 625741867
dataset_size: 627391612
configs:
- config_name: default
data_files:
- split: train
path: data/train-*