behaviors:
  Tennis:
    trainer: ppo
    batch_size: 1024
    beta: 0.005
    buffer_size: 10240
    epsilon: 0.2
    hidden_units: 256
    lambd: 0.95
    learning_rate: 0.0003
    learning_rate_schedule: constant
    max_steps: 5.0e7
    memory_size: 128
    normalize: true
    num_epoch: 3
    num_layers: 2
    time_horizon: 1000
    sequence_length: 64
    summary_freq: 10000
    use_recurrent: false
    vis_encode_type: simple
    reward_signals:
      extrinsic:
        strength: 1.0
        gamma: 0.99
    self_play:
      window: 10
      play_against_latest_model_ratio: 0.5
      save_steps: 50000
      swap_steps: 50000
      team_change: 100000