metadata
dataset_info:
features:
- name: file_path
dtype: string
- name: transcription
dtype: string
- name: instruction
dtype: string
- name: discrete_unit
sequence: int64
- name: response_interleaf
dtype: string
- name: response_text
dtype: string
- name: response_tokens
sequence: int64
- name: TTS_speech
dtype: audio
- name: TTS_speech_trans
dtype: string
- name: wer-res
dtype: float64
- name: cer-res
dtype: float64
- name: wer
dtype: float64
- name: cer
dtype: float64
- name: mos_score
dtype: float64
splits:
- name: train
num_bytes: 1717858412.125
num_examples: 5559
download_size: 1485427329
dataset_size: 1717858412.125
configs:
- config_name: default
data_files:
- split: train
path: data/train-*