general_cfg:
  algo_name: BCQ
  device: cpu
  env_name: gym
  eval_eps: 10
  eval_per_episode: 5
  load_checkpoint: false
  load_path: tasks
  max_steps: 200
  mode: train
  mp_backend: mp
  new_step_api: true
  render: false
  render_mode: human
  save_fig: true
  seed: 0
  show_fig: false
  test_eps: 20
  train_eps: 1000
  wrapper: null
algo_cfg:
  actor_hidden_dims:
  - 400
  - 300
  actor_lr: 0.001
  batch_size: 128
  behavior_agent_name: DDPG
  behavior_agent_parameters_path: /behave_param.yaml
  behavior_policy_path: /behaviour_models
  buffer_size: 100000
  collect_eps: 500
  collect_explore_data: true
  critic_hidden_dims:
  - 400
  - 300
  critic_lr: 0.001
  gamma: 0.99
  iters_per_ep: 10
  lmbda: 0.75
  phi: 0.05
  start_learn_buffer_size: 1000.0
  tau: 0.005
  vae_hidden_dims:
  - 750
  - 750
  vae_lr: 0.001
env_cfg:
  id: HalfCheetah-v3
  new_step_api: true
  render_mode: null