general_cfg:
  algo_name: PER_DQN
  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: 100
  max_step: 200
  mode: train
  model_save_fre: 500
  mp_backend: single
  n_learners: 1
  n_workers: 2
  online_eval: true
  online_eval_episode: 10
  seed: 1
  share_buffer: true
algo_cfg:
  batch_size: 64
  buffer_size: 100000
  buffer_type: PER_QUE
  epsilon_decay: 1000
  epsilon_end: 0.01
  epsilon_start: 0.95
  gamma: 0.99
  lr: 0.0001
  per_alpha: 0.6
  per_beta: 0.4
  per_beta_annealing: 0.001
  per_epsilon: 0.01
  target_update: 4
  value_layers:
  - activation: relu
    layer_size:
    - 256
    layer_type: linear
  - activation: relu
    layer_size:
    - 256
    layer_type: linear
env_cfg:
  id: CartPole-v1
  ignore_params:
  - wrapper
  - ignore_params
  render_mode: null
  wrapper: null