autoencoder/symAD_vctk_48000_hop300/config.yml

adv_batch_length: 9600
adv_train_max_steps: 700000
batch_length: 9600
batch_size: 16
config: config/autoencoder/symAD_vctk_48000_hop300.yaml
data:
  path: /mnt/home/yichiaowu/datasets/vctk_noisy/48000
  subset:
    test: clean_testset_wav
    train: clean_trainset_84spk_wav
    valid: clean_validset_84spk_wav
disable_cudnn: 'False'
discriminator_adv_loss_params:
  average_by_discriminators: false
discriminator_grad_norm: -1
discriminator_optimizer_params:
  betas:
  - 0.5
  - 0.9
  lr: 0.0002
  weight_decay: 0.0
discriminator_optimizer_type: Adam
discriminator_params:
  follow_official_norm: true
  period_discriminator_params:
    bias: true
    channels: 32
    downsample_scales:
    - 3
    - 3
    - 3
    - 3
    - 1
    in_channels: 1
    kernel_sizes:
    - 5
    - 3
    max_downsample_channels: 1024
    nonlinear_activation: LeakyReLU
    nonlinear_activation_params:
      negative_slope: 0.1
    out_channels: 1
    use_spectral_norm: false
    use_weight_norm: true
  periods:
  - 2
  - 3
  - 5
  - 7
  - 11
  scale_discriminator_params:
    bias: true
    channels: 128
    downsample_scales:
    - 4
    - 4
    - 4
    - 4
    - 1
    in_channels: 1
    kernel_sizes:
    - 15
    - 41
    - 5
    - 3
    max_downsample_channels: 1024
    max_groups: 16
    nonlinear_activation: LeakyReLU
    nonlinear_activation_params:
      negative_slope: 0.1
    out_channels: 1
  scale_downsample_pooling: AvgPool1d
  scale_downsample_pooling_params:
    kernel_size: 4
    padding: 2
    stride: 2
  scales: 3
discriminator_scheduler_params:
  gamma: 0.5
  milestones:
  - 200000
  - 400000
  - 600000
  - 800000
discriminator_scheduler_type: MultiStepLR
eval_interval_steps: 1000
exp_root: exp
feat_match_loss_params:
  average_by_discriminators: false
  average_by_layers: false
  include_final_outputs: false
generator_adv_loss_params:
  average_by_discriminators: false
generator_grad_norm: -1
generator_optimizer_params:
  betas:
  - 0.5
  - 0.9
  lr: 0.0001
  weight_decay: 0.0
generator_optimizer_type: Adam
generator_params:
  bias: true
  code_dim: 64
  codebook_num: 8
  codebook_size: 1024
  codec: audiodec
  dec_ratios:
  - 16
  - 8
  - 4
  - 2
  dec_strides:
  - 5
  - 5
  - 4
  - 3
  decode_channels: 32
  enc_ratios:
  - 2
  - 4
  - 8
  - 16
  enc_strides:
  - 3
  - 4
  - 5
  - 5
  encode_channels: 32
  input_channels: 1
  mode: causal
  output_channels: 1
  projector: conv1d
  quantier: residual_vq
generator_scheduler_params:
  gamma: 1.0
  step_size: 200000
generator_scheduler_type: StepLR
lambda_adv: 1.0
lambda_feat_match: 2.0
lambda_mel_loss: 45.0
lambda_shape_loss: 45.0
lambda_stft_loss: 45.0
lambda_vq_loss: 1.0
log_interval_steps: 100
mel_loss_params:
  fft_sizes:
  - 2048
  fmax: 24000
  fmin: 0
  fs: 48000
  hop_sizes:
  - 300
  log_base: null
  num_mels: 80
  win_lengths:
  - 2048
  window: hann_window
model_type: symAudioDec
num_workers: 2
outdir: exp/autoencoder/symAD_vctk_48000_hop300
paradigm: efficient
pin_memory: true
resume: exp/autoencoder/symAD_vctk_48000_hop300/checkpoint-200000steps.pkl
sampling_rate: 48000
save_interval_steps: 100000
seed: 1337
shape_loss_params:
  winlen:
  - 300
start_steps:
  discriminator: 200000
  generator: 0
stft_loss_params:
  fft_sizes:
  - 1024
  - 2048
  - 512
  hop_sizes:
  - 120
  - 240
  - 50
  win_lengths:
  - 600
  - 1200
  - 240
  window: hann_window
tag: autoencoder/symAD_vctk_48000_hop300
train_max_steps: 200000
train_mode: autoencoder
use_feat_match_loss: true
use_mel_loss: true
use_shape_loss: false
use_stft_loss: false