metadata
dataset_info:
features:
- name: input_codes
sequence:
sequence: int64
- name: labels
sequence:
sequence: int64
- name: speaker_embedding
sequence:
sequence: float32
- name: espeak
dtype: string
- name: emotion
sequence:
sequence:
sequence: float32
- name: fmax
sequence:
sequence:
sequence: float32
- name: pitch_std
sequence:
sequence:
sequence: float32
- name: speaking_rate
sequence:
sequence:
sequence: float32
- name: language_id
sequence:
sequence:
sequence: int64
- name: ctc_loss
sequence:
sequence:
sequence: float32
- name: speaker_noised
sequence:
sequence:
sequence: int64
- name: metadata
dtype: string
splits:
- name: train
num_bytes: 20747123909
num_examples: 300000
download_size: 3579815116
dataset_size: 20747123909
configs:
- config_name: default
data_files:
- split: train
path: data/train-*