behaviors:
  FoodCollector:
    trainer: ppo
    batch_size: 64
    beta: 0.005
    buffer_size: 10240
    epsilon: 0.2
    hidden_units: 128
    lambd: 0.95
    learning_rate: 0.0003
    max_steps: 2.0e6
    memory_size: 256
    normalize: false
    num_epoch: 3
    num_layers: 2
    time_horizon: 64
    sequence_length: 32
    summary_freq: 10000
    use_recurrent: false
    reward_signals:
      gail:
        strength: 0.1
        gamma: 0.99
        encoding_size: 128
        demo_path: Project/Assets/ML-Agents/Examples/FoodCollector/Demos/ExpertFood.demo
    behavioral_cloning:
      demo_path: Project/Assets/ML-Agents/Examples/FoodCollector/Demos/ExpertFood.demo
      strength: 1.0
      steps: 0