general:
  # Run mode. Options are 'prod' or 'dev'.
  run_mode: prod

dataset:
  #number of slices you want to split each recording into
  num_slices: 8

  #training/val split between the 2 data sets
  train_split: 0.8
  val_split : 0.2

  #used to initialize a random number generator.
  seed: 25

  #multiple modulations to contain in the dataset
  modulation_types: [bpsk, qpsk, qam16, qam64]

  # Rolloff factor for pulse shaping filter (0 < beta <= 1)
  beta: 0.3

  # Samples per symbol (determines bandwidth of the digital signal)
  sps: 4

  # SNR sweep range: start, stop (exclusive), and step (in dB)
  snr_start: -6     # Start value of SNR sweep (in dB)
  snr_stop: 13      # Stop value (exclusive) of SNR sweep (in dB)
  snr_step: 3       # Step size for SNR sweep (in dB)

  # Number of iterations (samples) per modulation and SNR combination
  num_iterations: 3

  # Number of samples per generated recording
  recording_length: 1024

  # Settings for each modulation scheme
  # Keys must match entries in `modulation_types`
  #   - `num_bits_per_symbol`: how many bits each symbol encodes (e.g., 1 for BPSK, 4 for 16-QAM)
  #   - `constellation_type`: type of modulation (e.g., "psk", "qam", "fsk", "ofdm")
  modulation_settings:
    bpsk:
      num_bits_per_symbol: 1
      constellation_type: psk
    qpsk:
      num_bits_per_symbol: 2
      constellation_type: psk
    qam16:
      num_bits_per_symbol: 4
      constellation_type: qam
    qam64:
      num_bits_per_symbol: 6
      constellation_type: qam



training:
  # Number of training samples processed together before the model updates its weights
  batch_size: 256

  # Number of complete passes through the training dataset during training
  epochs: 5

  # Learning rate: how much weights are updated after every batch
  # Suggested range for fine-tuning: 1e-6 to 1e-4
  learning_rate: 1e-4

  # Whether to use GPU acceleration for training (if available)
  use_gpu: true

  # Dropout rate for individual neurons/layers (probability of dropping out a unit)
  drop_rate: 0.5

  # Drop path rate: probability of dropping entire residual paths (stochastic depth)
  drop_path_rate: 0.2

  # Weight decay (L2 regularization) to help prevent overfitting
  wd: 0.01


app:
  # Optimization style for ORT conversion. Options: 'Fixed', 'None'
  optimization_style: Fixed

  # Target platform architecture. Common options: 'amd64', 'arm64'
  target_platform: amd64