# lightning.pytorch==2.2.5
seed_everything: 0
trainer:
  accelerator: gpu
  strategy: ddp_find_unused_parameters_true
  devices: 8
  num_nodes: 1
  precision: 32
  logger:
    class_path: lightning.pytorch.loggers.WandbLogger
    init_args:
      name: mel_big
      save_dir: vq_audio_simvq_bert_mel/8k_ration_20_loss
      version: null
      offline: false
      dir: null
      id: null
      anonymous: null
      project: endresult_tmp
      log_model: false
      experiment: null
      prefix: ''
      checkpoint_name: null
      job_type: null
      config: null
      entity: null
      reinit: null
      tags: null
      group: null
      notes: null
      magic: null
      config_exclude_keys: null
      config_include_keys: null
      mode: null
      allow_val_change: null
      resume: null
      force: null
      tensorboard: null
      sync_tensorboard: null
      monitor_gym: null
      save_code: null
      fork_from: null
      resume_from: null
      settings: null
  callbacks:
  - class_path: lightning.pytorch.callbacks.ModelCheckpoint
    init_args:
      dirpath: vq_audio_simvq_bert_mel/8k_ration_20_loss
      filename: null
      monitor: null
      verbose: false
      save_last: null
      save_top_k: -1
      save_weights_only: false
      mode: min
      auto_insert_metric_name: true
      every_n_train_steps: null
      train_time_interval: null
      every_n_epochs: null
      save_on_train_epoch_end: null
      enable_version_counter: true
  - class_path: lightning.pytorch.callbacks.LearningRateMonitor
    init_args:
      logging_interval: step
      log_momentum: false
      log_weight_decay: false
  fast_dev_run: false
  max_epochs: 50
  min_epochs: null
  max_steps: -1
  min_steps: null
  max_time: null
  limit_train_batches: null
  limit_val_batches: null
  limit_test_batches: null
  limit_predict_batches: null
  overfit_batches: 0.0
  val_check_interval: null
  check_val_every_n_epoch: 1
  num_sanity_val_steps: 0
  log_every_n_steps: 100
  enable_checkpointing: null
  enable_progress_bar: null
  enable_model_summary: null
  accumulate_grad_batches: 1
  gradient_clip_val: null
  gradient_clip_algorithm: null
  deterministic: null
  benchmark: null
  inference_mode: true
  use_distributed_sampler: true
  profiler: null
  detect_anomaly: false
  barebones: false
  plugins: null
  sync_batchnorm: false
  reload_dataloaders_every_n_epochs: 0
  default_root_dir: null
model:
  class_path: taming.models.vq_audio_simvq_mel.VQModel
  init_args:
    ddconfig:
      causal: true
      dimension: 512
      ratios:
      - 8
      - 8
      - 4
      - 4
    lossconfig:
      target: taming.modules.losses.stft_simvq_mel.VQSTFTWithDiscriminator
      params:
        disc_conditional: false
        disc_in_channels: 1
        disc_start: 0
        codebook_enlarge_ratio: 0
        codebook_enlarge_steps: 2000
        sample_rate: 24000
        commit_weight: 1000.0
        gen_loss_weight: 1.0
        mel_loss_coeff: 45.0
        mrd_loss_coeff: 1.0
    quantconfig: null
    sample_rate: 24000
    target_bandwidths: null
    audio_normalize: false
    segment: None
    ckpt_path: null
    ignore_keys: []
    colorize_nlabels: null
    monitor: null
    learning_rate: 0.0001
    warmup_epochs: 1.0
    scheduler_type: None
    min_learning_rate: 0
    use_ema: true
    stage: null
data:
  class_path: taming.data.speechtokenizer_24k.SpeechTokenizerDataModule
  init_args:
    batch_size: 6
    num_workers: 8
    train_path:
    - /mnt/nfs3/zhangjinouwen/dataset/rep/rep_small_mel_hubert_train.txt
    - /mnt/nfs3/zhangjinouwen/dataset/rep/rep_middle_mel_hubert_train.txt
    - /mnt/nfs3/zhangjinouwen/dataset/rep/rep_Emila1_mel_hubert_train.txt
    - /mnt/nfs3/zhangjinouwen/dataset/rep/rep_vc_mel_hubert_train.txt
    val_path: /mnt/nfs3/zhangjinouwen/dataset/rep/rep_small_wav_eval.txt
optimizer: null
lr_scheduler: null
ckpt_path: null