text_to_speech/egs/datasets/audio/wenetspeech/base_mel2wav.yaml

base_config:
  - egs/egs_bases/tts/base_zh.yaml

raw_data_dir: 'data/raw/AISHELL-3'
processed_data_dir: 'data/processed/aishell3'
binary_data_dir: 'data/binary/aishell3_fs2s'

audio_sample_rate: 16000
fmin: 0
fmax: 8000
pre_align_cls: egs.datasets.audio.aishell3.pre_align.AiShell3PreAlign

use_spk_id: true
num_spk: 220
# test_num: 200
binarization_args:
  shuffle: true
  trim_eos_bos: true
use_pitch_embed: false
vocoder: vocoders.hifigan.HifiGAN
vocoder_ckpt: 'pretrained/0707_hifigan_as3_ori_1'