Finetune Model for WAV 27db Sample Rate 22050/finetune_archived_230424-020826/train_louise_230424-020648.log

23-04-24 02:06:48.155 - INFO:   name: louise
  model: extensibletrainer
  scale: 1
  gpu_ids: [0]
  start_step: 0
  checkpointing_enabled: True
  fp16: True
  bitsandbytes: True
  gpus: 1
  datasets:[
    train:[
      name: training
      n_workers: 2
      batch_size: 34
      mode: paired_voice_audio
      path: ./training/louise/train.txt
      fetcher_mode: ['lj']
      phase: train
      max_wav_length: 255995
      max_text_length: 200
      sample_rate: 22050
      load_conditioning: True
      num_conditioning_candidates: 2
      conditioning_length: 44000
      use_bpe_tokenizer: True
      tokenizer_vocab: ./modules/tortoise-tts/tortoise/data/tokenizer.json
      load_aligned_codes: False
      data_type: img
    ]
    val:[
      name: validation
      n_workers: 2
      batch_size: 8
      mode: paired_voice_audio
      path: ./training/louise/validation.txt
      fetcher_mode: ['lj']
      phase: val
      max_wav_length: 255995
      max_text_length: 200
      sample_rate: 22050
      load_conditioning: True
      num_conditioning_candidates: 2
      conditioning_length: 44000
      use_bpe_tokenizer: True
      tokenizer_vocab: ./modules/tortoise-tts/tortoise/data/tokenizer.json
      load_aligned_codes: False
      data_type: img
    ]
  ]
  steps:[
    gpt_train:[
      training: gpt
      loss_log_buffer: 500
      optimizer: adamw
      optimizer_params:[
        lr: 1e-05
        weight_decay: 0.01
        beta1: 0.9
        beta2: 0.96
      ]
      clip_grad_eps: 4
      injectors:[
        paired_to_mel:[
          type: torch_mel_spectrogram
          mel_norm_file: ./modules/tortoise-tts/tortoise/data/mel_norms.pth
          in: wav
          out: paired_mel
        ]
        paired_cond_to_mel:[
          type: for_each
          subtype: torch_mel_spectrogram
          mel_norm_file: ./modules/tortoise-tts/tortoise/data/mel_norms.pth
          in: conditioning
          out: paired_conditioning_mel
        ]
        to_codes:[
          type: discrete_token
          in: paired_mel
          out: paired_mel_codes
          dvae_config: ./models/tortoise/train_diffusion_vocoder_22k_level.yml
        ]
        paired_fwd_text:[
          type: generator
          generator: gpt
          in: ['paired_conditioning_mel', 'padded_text', 'text_lengths', 'paired_mel_codes', 'wav_lengths']
          out: ['loss_text_ce', 'loss_mel_ce', 'logits']
        ]
      ]
      losses:[
        text_ce:[
          type: direct
          weight: 0.01
          key: loss_text_ce
        ]
        mel_ce:[
          type: direct
          weight: 1
          key: loss_mel_ce
        ]
      ]
    ]
  ]
  networks:[
    gpt:[
      type: generator
      which_model_G: unified_voice2
      kwargs:[
        layers: 30
        model_dim: 1024
        heads: 16
        max_text_tokens: 402
        max_mel_tokens: 604
        max_conditioning_inputs: 2
        mel_length_compression: 1024
        number_text_tokens: 256
        number_mel_codes: 8194
        start_mel_token: 8192
        stop_mel_token: 8193
        start_text_token: 255
        train_solo_embeddings: False
        use_mel_codes_as_input: True
        checkpointing: True
        tortoise_compat: True
      ]
    ]
  ]
  path:[
    strict_load: True
    pretrain_model_gpt: ./models/tortoise/autoregressive.pth
    root: ./
    experiments_root: ./training/louise/finetune
    models: ./training/louise/finetune/models
    training_state: ./training/louise/finetune/training_state
    log: ./training/louise/finetune
    val_images: ./training/louise/finetune/val_images
  ]
  train:[
    niter: 4700
    warmup_iter: -1
    mega_batch_factor: 4
    val_freq: 100
    ema_enabled: False
    default_lr_scheme: MultiStepLR
    gen_lr_steps: [2, 4, 9, 18, 25, 33, 50, 59]
    lr_gamma: 0.5
  ]
  eval:[
    pure: False
    output_state: gen
  ]
  logger:[
    save_checkpoint_freq: 100
    visuals: ['gen', 'mel']
    visual_debug_rate: 900
    is_mel_spectrogram: True
  ]
  is_train: True
  dist: False

23-04-24 02:06:48.156 - INFO: Random seed: 9424
23-04-24 02:06:48.891 - INFO: Number of training data elements: 293, iters: 9
23-04-24 02:06:48.891 - INFO: Total epochs needed: 523 for iters 4,700
23-04-24 02:06:56.504 - INFO: Loading model for [./models/tortoise/autoregressive.pth]
23-04-24 02:07:00.586 - INFO: Start training from epoch: 0, iter: 0