program: training/run_sweep.py
method: bayes
metric:
  name: val_accuracy
  goal: maximize
parameters:
  dataset:
    value: EmnistDataset
  model:
    value: CharacterModel
  network:
    value: ResidualNetwork
  network_args.block_sizes:
    distribution: q_uniform
    min: 16
    max: 256
    q: 8
  network_args.depths:
    distribution: int_uniform
    min: 1
    max: 3
  network_args.levels:
      distribution: int_uniform
      min: 1
      max: 2
  network_args.activation:
    distribution: categorical
    values:
      - gelu
      - leaky_relu
      - relu
      - selu
  optimizer_args.lr:
    distribution: uniform
    min: 1.e-5
    max: 1.e-1
  lr_scheduler_args.max_lr:
    distribution: uniform
    min: 1.e-5
    max: 1.e-1
  train_args.batch_size:
    distribution: q_uniform
    min: 32
    max: 256
    q: 8
  train_args.epochs:
    value: 5
early_terminate:
   type: hyperband
   min_iter: 2