
==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 133.1857841014862 seconds
inference_precision: fp32

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 131.6325900554657 seconds
inference_precision: fp32

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 119.98270344734192 seconds
inference_precision: fp32

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 121.47896695137024 seconds
inference_precision: fp32

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: bf16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 84.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True, False,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 97.36299586296082 seconds
inference_precision: bf16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 94.01083040237427 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 89.50024104118347 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 86.20240807533264 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt
profile:
  enabled: true
  export_tensorboard: false
  export_chrome_trace: false

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 518.512722492218 seconds
inference_precision: fp16
profile_dir: /mnt/ASC1637/lewm_baseline/le-wm/torch_profile
profile_summary: /mnt/ASC1637/lewm_baseline/le-wm/torch_profile/key_averages.txt

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 86.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True, False,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 89.49835586547852 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 86.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True, False,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 105.07861399650574 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 152.31250739097595 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 122.81560277938843 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 101.30036067962646 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 122.01387643814087 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 86.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True, False,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 110.37948775291443 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 94.35640263557434 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 98.5384590625763 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 93.3659656047821 seconds
inference_precision: fp16

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 67.28308987617493 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 46.01574730873108 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 44.974061727523804 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 102.31317353248596 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 45.355348110198975 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt
profile:
  enabled: true
  export_tensorboard: false
  export_chrome_trace: false

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 110.91939687728882 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead
profile_dir: /mnt/ASC1637/lewm_baseline/le-wm/torch_profile
profile_summary: /mnt/ASC1637/lewm_baseline/le-wm/torch_profile/key_averages.txt

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 54.21496343612671 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 43.69562244415283 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 42.99847435951233 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 43.14276576042175 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 43.71034002304077 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 47.23623466491699 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 57.10417580604553 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 51.94328594207764 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 46.037922620773315 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 40.61683630943298 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 41.09517192840576 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 41.62089252471924 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 49.26965045928955 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 1
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 40.442394495010376 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: ${eval.num_eval}
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False]), 'seeds': None}
evaluation_time: 103.50640678405762 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: ${eval.num_eval}
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False]), 'seeds': None}
evaluation_time: 101.80308318138123 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 60.68150067329407 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 31.417109727859497 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 16
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True, False,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 59.96041440963745 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 16
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True, False,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 30.851833820343018 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 12
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True, False,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 68.07364082336426 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 12
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: 100
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 90.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True, False,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 90.14458179473877 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: ???
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt
multi_gpu:
  enabled: true
  devices:
  - 0
  - 1

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 71.5921995639801 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: ???
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt
multi_gpu:
  enabled: true
  devices:
  - 0
  - 1

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 38.652626514434814 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: ???
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt
multi_gpu:
  enabled: true
  devices:
  - 0
  - 1

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 38.510936975479126 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: ???
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt
multi_gpu:
  enabled: true
  devices:
  - 0
  - 1
  - 2
  - 3

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 80.60203051567078 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: ???
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 31.07235813140869 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: ???
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True, False, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 30.950740575790405 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: ???
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt
multi_gpu:
  enabled: true
  devices:
  - 0
  - 1
  - 2

==== RESULTS ====
metrics: {'success_rate': 92.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 191.0513756275177 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: ???
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt
multi_gpu:
  enabled: true
  devices:
  - 0
  - 1
  - 2

==== RESULTS ====
metrics: {'success_rate': 92.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 133.4684717655182 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead

==== CONFIG ====
cache_dir: null
solver:
  _target_: stable_worldmodel.solver.CEMSolver
  model: ???
  batch_size: 8
  num_samples: 300
  var_scale: 1.0
  n_steps: 30
  topk: 30
  device: cuda
  seed: ${seed}
world:
  env_name: swm/TwoRoom-v1
  num_envs: ${eval.num_eval}
  max_episode_steps: ???
  history_size: 1
  frame_skip: 1
seed: 42
policy: two-room/tworoom/lejepa
inference_precision: fp16
dataset:
  stats: ${eval.dataset_name}
  keys_to_cache:
  - action
  - proprio
plan_config:
  horizon: 5
  receding_horizon: 5
  action_block: 5
eval:
  num_eval: 50
  goal_offset_steps: 25
  eval_budget: 50
  img_size: 224
  dataset_name: tworoom
  callables:
  - method: _set_state
    args:
      state:
        value: proprio
  - method: _set_goal_state
    args:
      goal_state:
        value: goal_proprio
output:
  filename: tworoom_results.txt
multi_gpu:
  enabled: true
  devices:
  - 0
  - 1

==== RESULTS ====
metrics: {'success_rate': 88.0, 'episode_successes': array([ True, False,  True, False,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True, False,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True, False,  True,  True,  True,  True,
        True,  True,  True, False,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True]), 'seeds': None}
evaluation_time: 38.12551426887512 seconds
inference_precision: fp16
inference_compile_target: predictor
inference_compile_mode: reduce-overhead
