general_cfg:
  algo_name: DDPG
  collect_traj: false
  device: cuda
  env_name: gym
  load_checkpoint: false
  load_model_step: best
  load_path: Train_single_CartPole-v1_DQN_20230515-211721
  max_episode: 200
  max_step: 200
  mode: train
  model_save_fre: 2000
  mp_backend: ray
  n_learners: 1
  n_workers: 4
  online_eval: true
  online_eval_episode: 20
  seed: 10
  share_buffer: true
algo_cfg:
  action_type: dpg
  actor_layers:
  - activation: relu
    layer_size:
    - 256
    layer_type: linear
  - activation: relu
    layer_size:
    - 256
    layer_type: linear
  actor_lr: 0.0001
  batch_size: 128
  buffer_size: 8000
  buffer_type: REPLAY_QUE
  critic_layers:
  - activation: relu
    layer_size:
    - 256
    layer_type: linear
  - activation: relu
    layer_size:
    - 256
    layer_type: linear
  critic_lr: 0.001
  gamma: 0.99
  policy_loss_weight: 0.002
  tau: 0.001
  value_max: .inf
  value_min: -.inf
env_cfg:
  id: Pendulum-v1
  ignore_params:
  - wrapper
  - ignore_params
  render_mode: null
  wrapper: null