hparameter.yaml

train:
  batch_size: 12
  adam:
    lr: 3e-4
    weight_decay: 1e-6
  decay:
    rate: 0.05
    start: 25000
    end: 100000
  num_workers: 16
  gpus: 2 #ddp
  loss_rate:
    dur: 1.0

data:
  lang: 'eng'
  text_cleaners: ['english_cleaners'] # korean_cleaners, english_cleaners, chinese_cleaners
  speakers: ['LJSpeech']
  train_dir: 'preprocessed_data/LJSpeech'
  train_meta: 'train.txt'  # relative path of metadata file from train_dir
  val_dir: 'preprocessed_data/LJSpeech'
  val_meta: 'val.txt'  # relative path of metadata file from val_dir'
  lexicon_path: 'lexicon/librispeech-lexicon.txt'

audio:
  n_mel_channels: 80
  filter_length: 1024
  hop_length: 256
  win_length: 1024
  sampling_rate: 22050
  mel_fmin: 0.0
  mel_fmax: 8000.0

encoder:
  channel: 512
  kernel: 5
  depth: 3
  dropout_rate: 0.5
  speaker_emb: 64

dur_predictor:
  dur_lstm_channel: 512
  range_lstm_channel: 512  

window:
  scale: 300
  length: 256

wavegrad:
  is_large: False #if False, Base
  encode_channel: 576 ##512+64
  scale_factors: [5,5,3,2,2]
  upsample:
    preconv_channel: 768
    out_channels: [512, 512, 256, 128, 128]
    #dilations: [[1,2,4,8],[1,2,4,8],[1,2,4,8],[1,2,4,8],[1,2,4,8]]
    dilations: [[1,2,4,8],[1,2,4,8],[1,2,4,8],[1,2,1,2],[1,2,1,2]]
  downsample:
    preconv_channel: 32
    out_channels: [128, 128, 256, 512]
    dilations: [[1,2,4],[1,2,4],[1,2,4],[1,2,4]]
  pos_emb_dim: 512 

ddpm:
  max_step: 1000
  noise_schedule: "torch.linspace(1e-6, 0.01, hparams.ddpm.max_step)"
  pos_emb_scale: 5000
  pos_emb_channels: 128 
  infer_step: 8 #TBD
  infer_schedule: "torch.tensor([1e-6,2e-6,1e-5,1e-4,1e-3,1e-2,1e-1,9e-1])" #TBD

log:
  name: 'wavegrad2'
  checkpoint_dir: 'checkpoint'
  tensorboard_dir: 'tensorboard'
  test_result_dir: 'test_sample/result'