config.ar_nar_double.yaml

dataset:
  training: [
  ]
  validation: [
  ]
  noise: [
  ]
  
  speaker_name_getter: "lambda p: f'{p.parts[-3]}_{p.parts[-2]}'"
  
  use_hdf5: True
  use_metadata: True
  hdf5_flag: r
  validate: True

  workers: 4
  cache: True

  phones_range: [4, 256]
  duration_range: [1.0, 16.0]

  random_utterance: 1.0
  max_prompts: 3
  prompt_duration: 3.0

  sample_type: speaker

  tasks_list: ["tts"] # , "ns", "sr", "tse", "cse", "nse", "tts"]

models:
  _prom_levels: 4
  _max_levels: 8

  _models:
  - name: "ar+nar"
    size: "double"
    resp_levels: 4
    prom_levels: 4
    tasks: 8
    arch_type: "retnet"
    training: True

hyperparameters:
  batch_size: 8
  gradient_accumulation_steps: 1
  gradient_clipping: 100
  
  optimizer: AdamW
  learning_rate: 1.0e-5
  
  scheduler_type: ""

evaluation:
  batch_size: 16
  frequency: 500
  size: 16
  
  steps: 300
  ar_temperature: 0.95
  nar_temperature: 0.25
  load_disabled_engines: True

trainer:
  iterations: 1_000_000
  
  save_tag: step
  save_on_oom: True
  save_on_quit: True
  save_frequency: 500
  export_on_save: True

  keep_last_checkpoints: 4

  aggressive_optimizations: False
  load_disabled_engines: False

  load_state_dict: True
  
  gc_mode: None # "global_step"

  weight_dtype: float32
  amp: False

  backend: local
  deepspeed:
    zero_optimization_level: 0
    use_compression_training: True

inference:
  weight_dtype: float32
  amp: False

  use_vocos: True
  normalize: False

  recurrent_chunk_size: 0
  recurrent_forward: False

bitsandbytes:
  enabled: False
  injects: True
  linear: True
  embedding: True

device: cpu