# @package _group_

common:
  fp16: true
  log_format: json
  log_interval: 200
  tensorboard_logdir: tblog
  seed: 1337

checkpoint:
  save_interval: 1000000
  keep_last_epochs: 5
  save_interval_updates: 1000
  keep_interval_updates_pattern: 10000
  keep_interval_updates: 5
  best_checkpoint_metric: accuracy
  maximize_best_checkpoint_metric: true

distributed_training:
  ddp_backend: c10d
  find_unused_parameters: true
  distributed_world_size: 1
  nprocs_per_node: 8


criterion:
  _name: "label_smoothed_cross_entropy"


task:
  _name: "translation_from_jst"

dataset:
  num_workers: 0
  max_tokens: 4096
  skip_invalid_size_inputs_valid_test: true
  validate_after_updates: ${model.freeze_finetune_updates}
  validate_interval: ${checkpoint.save_interval}
  validate_interval_updates: ${checkpoint.save_interval_updates}
  train_subset: train_clean_100
  valid_subset: dev_clean
  required_batch_size_multiple: 1

optimizer:
  _name: adam
  adam_betas: (0.9,0.98)
  adam_eps: 1e-06
  weight_decay: 0.0

lr_scheduler:
  _name: tri_stage
  phase_ratio: [0.1, 0.4, 0.5]
  final_lr_scale: 0.05

model:
  _name: hubert_t2c
  w2v_path: ???
  layerdrop: 0.1
  decoder_layerdrop: 0.1
  activation_dropout: 0.1
  feature_grad_mult: 0.0
  freeze_finetune_updates: 0

hydra:
  job:
    config:
      override_dirname:
        kv_sep: '-'
        item_sep: '__'
        exclude_keys:
          - run
          - task.data
          - task.label_dir
          - model.w2v_path
          - dataset.train_subset
          - dataset.valid_subset
  run:
    dir: ???
  sweep:
    dir: ???
    subdir: ${hydra.job.config_name}__${hydra.job.override_dirname}