leo_rbdl_zmq_drl_real.yaml

experiment: 
  type: experiment/online_learning
  runs: 1
  trials: 0
  steps: 0
  rate: 0
  test_interval: 0
  output: test_on_pertubed_model
  environment: 
    type: environment/sandbox
    model: 
      type: sandbox_model/leo_walk
      control_step: 0.0003
      integration_steps: 1
      dynamics: 
        type: dynamics/rbdl
        file: leo_ff_dl_perturbed.lua
        points: tip_left, tip_right, heel_left, heel_right
        auxiliary: com
      target_dof: 9
      animation: full
      condition: [0.01, 0.01]
      mode: vc
      sub_state_drl: 
        type: signal/vector
    task: 
      type: task/leo_walking
      timeout: 20
      randomize: 0
      measurement_noise: 0
  agent: 
    type: agent/zmq_action_state
    communicator: 
      type: communicator/zeromq/request_reply
      pattern: ZMQ_REQ
      sync: 0
      addr: tcp://localhost:5555
    observation_dims: experiment/environment/task/observation_dims
    action_dims: experiment/environment/task/action_dims
    action_min: experiment/environment/task/action_min
    action_max: experiment/environment/task/action_max
    test: 0
    pub_state_drl: experiment/environment/model/sub_state_drl
  test_agent: 
    type: agent/zmq_action_state
    communicator: 
      type: communicator/zeromq/request_reply
      pattern: ZMQ_REQ
      sync: 0
      addr: tcp://localhost:5555
    observation_dims: experiment/environment/task/observation_dims
    action_dims: experiment/environment/task/action_dims
    action_min: experiment/environment/task/action_min
    action_max: experiment/environment/task/action_max
    test: 1
    pub_state_drl: experiment/environment/model/sub_state_drl
  load_file: model-leo-rbdl-with-diff-1.ckpt
  save_every: never