metadata
dataset_info:
features:
- name: item_name
dtype: string
- name: ctrspeech_dur
dtype: string
- name: ctrspeech_pitch
dtype: string
- name: energy
dtype: string
- name: gender
dtype: string
- name: emotion
dtype: string
- name: spk
dtype: string
- name: text
dtype: string
- name: style_prompt
dtype: string
- name: utterance_pitch_mean
dtype: float32
- name: utterance_pitch_std
dtype: float32
- name: snr
dtype: float64
- name: c50
dtype: float64
- name: speaking_rate
dtype: float64
- name: phonemes
dtype: string
splits:
- name: train
num_bytes: 113215819.49484062
num_examples: 283152
download_size: 61709917
dataset_size: 113215819.49484062
configs:
- config_name: default
data_files:
- split: train
path: data/train-*