[2025-07-29 10:53:09,328][05283] Saving configuration to /content/train_dir/default_experiment/config.json...
[2025-07-29 10:53:09,330][05283] Rollout worker 0 uses device cpu
[2025-07-29 10:53:09,330][05283] Rollout worker 1 uses device cpu
[2025-07-29 10:53:09,331][05283] Rollout worker 2 uses device cpu
[2025-07-29 10:53:09,332][05283] Rollout worker 3 uses device cpu
[2025-07-29 10:53:09,332][05283] Rollout worker 4 uses device cpu
[2025-07-29 10:53:09,334][05283] Rollout worker 5 uses device cpu
[2025-07-29 10:53:09,334][05283] Rollout worker 6 uses device cpu
[2025-07-29 10:53:09,335][05283] Rollout worker 7 uses device cpu
[2025-07-29 10:53:09,429][05283] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 10:53:09,430][05283] InferenceWorker_p0-w0: min num requests: 2
[2025-07-29 10:53:09,459][05283] Starting all processes...
[2025-07-29 10:53:09,460][05283] Starting process learner_proc0
[2025-07-29 10:53:09,462][05283] EvtLoop [Runner_EvtLoop, process=main process 5283] unhandled exception in slot='_on_start' connected to emitter=Emitter(object_id='Runner_EvtLoop', signal_name='start'), args=()
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 355, in _process_signal
    slot_callable(*args)
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/runners/runner_parallel.py", line 49, in _on_start
    self._start_processes()
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/runners/runner_parallel.py", line 56, in _start_processes
    p.start()
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 515, in start
    self._process.start()
  File "/usr/lib/python3.11/multiprocessing/process.py", line 121, in start
    self._popen = self._Popen(self)
                  ^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/multiprocessing/context.py", line 288, in _Popen
    return Popen(process_obj)
           ^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/multiprocessing/popen_spawn_posix.py", line 32, in __init__
    super().__init__(process_obj)
  File "/usr/lib/python3.11/multiprocessing/popen_fork.py", line 19, in __init__
    self._launch(process_obj)
  File "/usr/lib/python3.11/multiprocessing/popen_spawn_posix.py", line 47, in _launch
    reduction.dump(process_obj, fp)
  File "/usr/lib/python3.11/multiprocessing/reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)
TypeError: cannot pickle 'TLSBuffer' object
[2025-07-29 10:53:09,467][05283] Unhandled exception cannot pickle 'TLSBuffer' object in evt loop Runner_EvtLoop
[2025-07-29 10:53:09,468][05283] Uncaught exception in Runner evt loop
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/runners/runner.py", line 770, in run
    evt_loop_status = self.event_loop.exec()
                      ^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 403, in exec
    raise exc
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 399, in exec
    while self._loop_iteration():
          ^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 383, in _loop_iteration
    self._process_signal(s)
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 358, in _process_signal
    raise exc
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 355, in _process_signal
    slot_callable(*args)
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/runners/runner_parallel.py", line 49, in _on_start
    self._start_processes()
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/runners/runner_parallel.py", line 56, in _start_processes
    p.start()
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 515, in start
    self._process.start()
  File "/usr/lib/python3.11/multiprocessing/process.py", line 121, in start
    self._popen = self._Popen(self)
                  ^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/multiprocessing/context.py", line 288, in _Popen
    return Popen(process_obj)
           ^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/multiprocessing/popen_spawn_posix.py", line 32, in __init__
    super().__init__(process_obj)
  File "/usr/lib/python3.11/multiprocessing/popen_fork.py", line 19, in __init__
    self._launch(process_obj)
  File "/usr/lib/python3.11/multiprocessing/popen_spawn_posix.py", line 47, in _launch
    reduction.dump(process_obj, fp)
  File "/usr/lib/python3.11/multiprocessing/reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)
TypeError: cannot pickle 'TLSBuffer' object
[2025-07-29 10:53:09,470][05283] Runner profile tree view:
main_loop: 0.0113
[2025-07-29 10:53:09,471][05283] Collected {}, FPS: 0.0
[2025-07-29 10:53:31,275][05283] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 10:53:31,275][05283] Overriding arg 'num_workers' with value 1 passed from command line
[2025-07-29 10:53:31,276][05283] Adding new argument 'no_render'=True that is not in the saved config file!
[2025-07-29 10:53:31,277][05283] Adding new argument 'save_video'=True that is not in the saved config file!
[2025-07-29 10:53:31,277][05283] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 10:53:31,278][05283] Adding new argument 'video_name'=None that is not in the saved config file!
[2025-07-29 10:53:31,279][05283] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 10:53:31,280][05283] Adding new argument 'max_num_episodes'=10 that is not in the saved config file!
[2025-07-29 10:53:31,280][05283] Adding new argument 'push_to_hub'=False that is not in the saved config file!
[2025-07-29 10:53:31,281][05283] Adding new argument 'hf_repository'=None that is not in the saved config file!
[2025-07-29 10:53:31,281][05283] Adding new argument 'policy_index'=0 that is not in the saved config file!
[2025-07-29 10:53:31,282][05283] Adding new argument 'eval_deterministic'=False that is not in the saved config file!
[2025-07-29 10:53:31,282][05283] Adding new argument 'train_script'=None that is not in the saved config file!
[2025-07-29 10:53:31,283][05283] Adding new argument 'enjoy_script'=None that is not in the saved config file!
[2025-07-29 10:53:31,284][05283] Using frameskip 1 and render_action_repeat=4 for evaluation
[2025-07-29 10:53:31,311][05283] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 10:53:31,313][05283] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 10:53:31,316][05283] RunningMeanStd input shape: (1,)
[2025-07-29 10:53:31,329][05283] ConvEncoder: input_channels=3
[2025-07-29 10:53:31,591][05283] Conv encoder output size: 512
[2025-07-29 10:53:31,592][05283] Policy head output size: 512
[2025-07-29 10:53:31,915][05283] No checkpoints found
[2025-07-29 10:53:45,412][05283] Environment doom_basic already registered, overwriting...
[2025-07-29 10:53:45,413][05283] Environment doom_two_colors_easy already registered, overwriting...
[2025-07-29 10:53:45,414][05283] Environment doom_two_colors_hard already registered, overwriting...
[2025-07-29 10:53:45,415][05283] Environment doom_dm already registered, overwriting...
[2025-07-29 10:53:45,415][05283] Environment doom_dwango5 already registered, overwriting...
[2025-07-29 10:53:45,416][05283] Environment doom_my_way_home_flat_actions already registered, overwriting...
[2025-07-29 10:53:45,416][05283] Environment doom_defend_the_center_flat_actions already registered, overwriting...
[2025-07-29 10:53:45,417][05283] Environment doom_my_way_home already registered, overwriting...
[2025-07-29 10:53:45,417][05283] Environment doom_deadly_corridor already registered, overwriting...
[2025-07-29 10:53:45,418][05283] Environment doom_defend_the_center already registered, overwriting...
[2025-07-29 10:53:45,419][05283] Environment doom_defend_the_line already registered, overwriting...
[2025-07-29 10:53:45,419][05283] Environment doom_health_gathering already registered, overwriting...
[2025-07-29 10:53:45,420][05283] Environment doom_health_gathering_supreme already registered, overwriting...
[2025-07-29 10:53:45,420][05283] Environment doom_battle already registered, overwriting...
[2025-07-29 10:53:45,421][05283] Environment doom_battle2 already registered, overwriting...
[2025-07-29 10:53:45,422][05283] Environment doom_duel_bots already registered, overwriting...
[2025-07-29 10:53:45,422][05283] Environment doom_deathmatch_bots already registered, overwriting...
[2025-07-29 10:53:45,423][05283] Environment doom_duel already registered, overwriting...
[2025-07-29 10:53:45,423][05283] Environment doom_deathmatch_full already registered, overwriting...
[2025-07-29 10:53:45,424][05283] Environment doom_benchmark already registered, overwriting...
[2025-07-29 10:53:45,425][05283] register_encoder_factory: <function make_vizdoom_encoder at 0x7bcb34d2b1a0>
[2025-07-29 10:53:45,434][05283] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 10:53:45,438][05283] Experiment dir /content/train_dir/default_experiment already exists!
[2025-07-29 10:53:45,439][05283] Resuming existing experiment from /content/train_dir/default_experiment...
[2025-07-29 10:53:45,439][05283] Weights and Biases integration disabled
[2025-07-29 10:53:45,441][05283] Environment var CUDA_VISIBLE_DEVICES is 0

[2025-07-29 10:53:47,725][05283] Starting experiment with the following configuration:
help=False
algo=APPO
env=doom_health_gathering_supreme
experiment=default_experiment
train_dir=/content/train_dir
restart_behavior=resume
device=gpu
seed=None
num_policies=1
async_rl=True
serial_mode=False
batched_sampling=False
num_batches_to_accumulate=2
worker_num_splits=2
policy_workers_per_policy=1
max_policy_lag=1000
num_workers=8
num_envs_per_worker=4
batch_size=1024
num_batches_per_epoch=1
num_epochs=1
rollout=32
recurrence=32
shuffle_minibatches=False
gamma=0.99
reward_scale=1.0
reward_clip=1000.0
value_bootstrap=False
normalize_returns=True
exploration_loss_coeff=0.001
value_loss_coeff=0.5
kl_loss_coeff=0.0
exploration_loss=symmetric_kl
gae_lambda=0.95
ppo_clip_ratio=0.1
ppo_clip_value=0.2
with_vtrace=False
vtrace_rho=1.0
vtrace_c=1.0
optimizer=adam
adam_eps=1e-06
adam_beta1=0.9
adam_beta2=0.999
max_grad_norm=4.0
learning_rate=0.0001
lr_schedule=constant
lr_schedule_kl_threshold=0.008
lr_adaptive_min=1e-06
lr_adaptive_max=0.01
obs_subtract_mean=0.0
obs_scale=255.0
normalize_input=True
normalize_input_keys=None
decorrelate_experience_max_seconds=0
decorrelate_envs_on_one_worker=True
actor_worker_gpus=[]
set_workers_cpu_affinity=True
force_envs_single_thread=False
default_niceness=0
log_to_file=True
experiment_summaries_interval=10
flush_summaries_interval=30
stats_avg=100
summaries_use_frameskip=True
heartbeat_interval=20
heartbeat_reporting_interval=600
train_for_env_steps=4000000
train_for_seconds=10000000000
save_every_sec=120
keep_checkpoints=2
load_checkpoint_kind=latest
save_milestones_sec=-1
save_best_every_sec=5
save_best_metric=reward
save_best_after=100000
benchmark=False
encoder_mlp_layers=[512, 512]
encoder_conv_architecture=convnet_simple
encoder_conv_mlp_layers=[512]
use_rnn=True
rnn_size=512
rnn_type=gru
rnn_num_layers=1
decoder_mlp_layers=[]
nonlinearity=elu
policy_initialization=orthogonal
policy_init_gain=1.0
actor_critic_share_weights=True
adaptive_stddev=True
continuous_tanh_scale=0.0
initial_stddev=1.0
use_env_info_cache=False
env_gpu_actions=False
env_gpu_observations=True
env_frameskip=4
env_framestack=1
pixel_format=CHW
use_record_episode_statistics=False
with_wandb=False
wandb_user=None
wandb_project=sample_factory
wandb_group=None
wandb_job_type=SF
wandb_tags=[]
with_pbt=False
pbt_mix_policies_in_one_env=True
pbt_period_env_steps=5000000
pbt_start_mutation=20000000
pbt_replace_fraction=0.3
pbt_mutation_rate=0.15
pbt_replace_reward_gap=0.1
pbt_replace_reward_gap_absolute=1e-06
pbt_optimize_gamma=False
pbt_target_objective=true_objective
pbt_perturb_min=1.1
pbt_perturb_max=1.5
num_agents=-1
num_humans=0
num_bots=-1
start_bot_difficulty=None
timelimit=None
res_w=128
res_h=72
wide_aspect_ratio=False
eval_env_frameskip=1
fps=35
command_line=--env=doom_health_gathering_supreme --num_workers=8 --num_envs_per_worker=4 --train_for_env_steps=4000000
cli_args={'env': 'doom_health_gathering_supreme', 'num_workers': 8, 'num_envs_per_worker': 4, 'train_for_env_steps': 4000000}
git_hash=unknown
git_repo_name=not a git repository
[2025-07-29 10:53:47,726][05283] Saving configuration to /content/train_dir/default_experiment/config.json...
[2025-07-29 10:53:47,728][05283] Rollout worker 0 uses device cpu
[2025-07-29 10:53:47,728][05283] Rollout worker 1 uses device cpu
[2025-07-29 10:53:47,729][05283] Rollout worker 2 uses device cpu
[2025-07-29 10:53:47,730][05283] Rollout worker 3 uses device cpu
[2025-07-29 10:53:47,730][05283] Rollout worker 4 uses device cpu
[2025-07-29 10:53:47,731][05283] Rollout worker 5 uses device cpu
[2025-07-29 10:53:47,732][05283] Rollout worker 6 uses device cpu
[2025-07-29 10:53:47,733][05283] Rollout worker 7 uses device cpu
[2025-07-29 10:53:47,768][05283] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 10:53:47,769][05283] InferenceWorker_p0-w0: min num requests: 2
[2025-07-29 10:53:47,797][05283] Starting all processes...
[2025-07-29 10:53:47,797][05283] Starting process learner_proc0
[2025-07-29 10:53:47,800][05283] EvtLoop [Runner_EvtLoop, process=main process 5283] unhandled exception in slot='_on_start' connected to emitter=Emitter(object_id='Runner_EvtLoop', signal_name='start'), args=()
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 355, in _process_signal
    slot_callable(*args)
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/runners/runner_parallel.py", line 49, in _on_start
    self._start_processes()
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/runners/runner_parallel.py", line 56, in _start_processes
    p.start()
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 515, in start
    self._process.start()
  File "/usr/lib/python3.11/multiprocessing/process.py", line 121, in start
    self._popen = self._Popen(self)
                  ^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/multiprocessing/context.py", line 288, in _Popen
    return Popen(process_obj)
           ^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/multiprocessing/popen_spawn_posix.py", line 32, in __init__
    super().__init__(process_obj)
  File "/usr/lib/python3.11/multiprocessing/popen_fork.py", line 19, in __init__
    self._launch(process_obj)
  File "/usr/lib/python3.11/multiprocessing/popen_spawn_posix.py", line 47, in _launch
    reduction.dump(process_obj, fp)
  File "/usr/lib/python3.11/multiprocessing/reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)
TypeError: cannot pickle 'TLSBuffer' object
[2025-07-29 10:53:47,801][05283] Unhandled exception cannot pickle 'TLSBuffer' object in evt loop Runner_EvtLoop
[2025-07-29 10:53:47,801][05283] Uncaught exception in Runner evt loop
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/runners/runner.py", line 770, in run
    evt_loop_status = self.event_loop.exec()
                      ^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 403, in exec
    raise exc
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 399, in exec
    while self._loop_iteration():
          ^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 383, in _loop_iteration
    self._process_signal(s)
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 358, in _process_signal
    raise exc
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 355, in _process_signal
    slot_callable(*args)
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/runners/runner_parallel.py", line 49, in _on_start
    self._start_processes()
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/runners/runner_parallel.py", line 56, in _start_processes
    p.start()
  File "/usr/local/lib/python3.11/dist-packages/signal_slot/signal_slot.py", line 515, in start
    self._process.start()
  File "/usr/lib/python3.11/multiprocessing/process.py", line 121, in start
    self._popen = self._Popen(self)
                  ^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/multiprocessing/context.py", line 288, in _Popen
    return Popen(process_obj)
           ^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/multiprocessing/popen_spawn_posix.py", line 32, in __init__
    super().__init__(process_obj)
  File "/usr/lib/python3.11/multiprocessing/popen_fork.py", line 19, in __init__
    self._launch(process_obj)
  File "/usr/lib/python3.11/multiprocessing/popen_spawn_posix.py", line 47, in _launch
    reduction.dump(process_obj, fp)
  File "/usr/lib/python3.11/multiprocessing/reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)
TypeError: cannot pickle 'TLSBuffer' object
[2025-07-29 10:53:47,803][05283] Runner profile tree view:
main_loop: 0.0063
[2025-07-29 10:53:47,804][05283] Collected {}, FPS: 0.0
[2025-07-29 10:55:39,589][08356] Saving configuration to /content/train_dir/default_experiment/config.json...
[2025-07-29 10:55:39,591][08356] Rollout worker 0 uses device cpu
[2025-07-29 10:55:39,592][08356] Rollout worker 1 uses device cpu
[2025-07-29 10:55:39,593][08356] Rollout worker 2 uses device cpu
[2025-07-29 10:55:39,594][08356] Rollout worker 3 uses device cpu
[2025-07-29 10:55:39,594][08356] Rollout worker 4 uses device cpu
[2025-07-29 10:55:39,595][08356] Rollout worker 5 uses device cpu
[2025-07-29 10:55:39,596][08356] Rollout worker 6 uses device cpu
[2025-07-29 10:55:39,597][08356] Rollout worker 7 uses device cpu
[2025-07-29 10:55:39,635][08356] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 10:55:39,636][08356] InferenceWorker_p0-w0: min num requests: 2
[2025-07-29 10:55:39,669][08356] Starting all processes...
[2025-07-29 10:55:39,669][08356] Starting process learner_proc0
[2025-07-29 10:55:39,723][08356] Starting all processes...
[2025-07-29 10:55:39,727][08356] Starting process inference_proc0-0
[2025-07-29 10:55:39,728][08356] Starting process rollout_proc0
[2025-07-29 10:55:39,728][08356] Starting process rollout_proc1
[2025-07-29 10:55:39,728][08356] Starting process rollout_proc2
[2025-07-29 10:55:39,730][08356] Starting process rollout_proc3
[2025-07-29 10:55:39,730][08356] Starting process rollout_proc4
[2025-07-29 10:55:39,733][08356] Starting process rollout_proc5
[2025-07-29 10:55:39,734][08356] Starting process rollout_proc6
[2025-07-29 10:55:39,734][08356] Starting process rollout_proc7
[2025-07-29 10:55:42,283][08564] Worker 0 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 10:55:42,307][08550] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 10:55:42,308][08550] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for learning process 0
[2025-07-29 10:55:42,323][08550] Num visible devices: 1
[2025-07-29 10:55:42,324][08550] Starting seed is not provided
[2025-07-29 10:55:42,324][08550] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 10:55:42,324][08550] Initializing actor-critic model on device cuda:0
[2025-07-29 10:55:42,324][08550] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 10:55:42,326][08550] RunningMeanStd input shape: (1,)
[2025-07-29 10:55:42,337][08550] ConvEncoder: input_channels=3
[2025-07-29 10:55:42,443][08550] Conv encoder output size: 512
[2025-07-29 10:55:42,443][08550] Policy head output size: 512
[2025-07-29 10:55:42,459][08550] Created Actor Critic model with architecture:
[2025-07-29 10:55:42,459][08550] ActorCriticSharedWeights(
  (obs_normalizer): ObservationNormalizer(
    (running_mean_std): RunningMeanStdDictInPlace(
      (running_mean_std): ModuleDict(
        (obs): RunningMeanStdInPlace()
      )
    )
  )
  (returns_normalizer): RecursiveScriptModule(original_name=RunningMeanStdInPlace)
  (encoder): VizdoomEncoder(
    (basic_encoder): ConvEncoder(
      (enc): RecursiveScriptModule(
        original_name=ConvEncoderImpl
        (conv_head): RecursiveScriptModule(
          original_name=Sequential
          (0): RecursiveScriptModule(original_name=Conv2d)
          (1): RecursiveScriptModule(original_name=ELU)
          (2): RecursiveScriptModule(original_name=Conv2d)
          (3): RecursiveScriptModule(original_name=ELU)
          (4): RecursiveScriptModule(original_name=Conv2d)
          (5): RecursiveScriptModule(original_name=ELU)
        )
        (mlp_layers): RecursiveScriptModule(
          original_name=Sequential
          (0): RecursiveScriptModule(original_name=Linear)
          (1): RecursiveScriptModule(original_name=ELU)
        )
      )
    )
  )
  (core): ModelCoreRNN(
    (core): GRU(512, 512)
  )
  (decoder): MlpDecoder(
    (mlp): Identity()
  )
  (critic_linear): Linear(in_features=512, out_features=1, bias=True)
  (action_parameterization): ActionParameterizationDefault(
    (distribution_linear): Linear(in_features=512, out_features=5, bias=True)
  )
)
[2025-07-29 10:55:42,613][08550] Using optimizer <class 'torch.optim.adam.Adam'>
[2025-07-29 10:55:42,814][08565] Worker 1 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 10:55:42,838][08566] Worker 2 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 10:55:42,840][08569] Worker 5 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 10:55:42,864][08568] Worker 4 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 10:55:42,975][08563] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 10:55:42,975][08563] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for inference process 0
[2025-07-29 10:55:42,991][08563] Num visible devices: 1
[2025-07-29 10:55:43,006][08567] Worker 3 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 10:55:43,024][08570] Worker 7 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 10:55:43,065][08571] Worker 6 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 10:55:47,009][08550] No checkpoints found
[2025-07-29 10:55:47,009][08550] Did not load from checkpoint, starting from scratch!
[2025-07-29 10:55:47,009][08550] Initialized policy 0 weights for model version 0
[2025-07-29 10:55:47,011][08550] LearnerWorker_p0 finished initialization!
[2025-07-29 10:55:47,011][08550] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 10:55:47,121][08563] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 10:55:47,123][08563] RunningMeanStd input shape: (1,)
[2025-07-29 10:55:47,135][08563] ConvEncoder: input_channels=3
[2025-07-29 10:55:47,238][08563] Conv encoder output size: 512
[2025-07-29 10:55:47,238][08563] Policy head output size: 512
[2025-07-29 10:55:47,271][08356] Inference worker 0-0 is ready!
[2025-07-29 10:55:47,272][08356] All inference workers are ready! Signal rollout workers to start!
[2025-07-29 10:55:47,323][08568] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 10:55:47,323][08569] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 10:55:47,323][08566] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 10:55:47,323][08564] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 10:55:47,323][08567] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 10:55:47,324][08570] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 10:55:47,324][08571] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 10:55:47,324][08565] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 10:55:47,662][08567] Decorrelating experience for 0 frames...
[2025-07-29 10:55:47,662][08568] Decorrelating experience for 0 frames...
[2025-07-29 10:55:47,662][08566] Decorrelating experience for 0 frames...
[2025-07-29 10:55:47,662][08570] Decorrelating experience for 0 frames...
[2025-07-29 10:55:47,662][08564] Decorrelating experience for 0 frames...
[2025-07-29 10:55:47,915][08570] Decorrelating experience for 32 frames...
[2025-07-29 10:55:47,918][08567] Decorrelating experience for 32 frames...
[2025-07-29 10:55:47,919][08566] Decorrelating experience for 32 frames...
[2025-07-29 10:55:47,931][08565] Decorrelating experience for 0 frames...
[2025-07-29 10:55:47,932][08569] Decorrelating experience for 0 frames...
[2025-07-29 10:55:48,037][08564] Decorrelating experience for 32 frames...
[2025-07-29 10:55:48,055][08571] Decorrelating experience for 0 frames...
[2025-07-29 10:55:48,212][08568] Decorrelating experience for 32 frames...
[2025-07-29 10:55:48,212][08565] Decorrelating experience for 32 frames...
[2025-07-29 10:55:48,215][08569] Decorrelating experience for 32 frames...
[2025-07-29 10:55:48,244][08567] Decorrelating experience for 64 frames...
[2025-07-29 10:55:48,279][08570] Decorrelating experience for 64 frames...
[2025-07-29 10:55:48,305][08571] Decorrelating experience for 32 frames...
[2025-07-29 10:55:48,524][08566] Decorrelating experience for 64 frames...
[2025-07-29 10:55:48,564][08568] Decorrelating experience for 64 frames...
[2025-07-29 10:55:48,572][08564] Decorrelating experience for 64 frames...
[2025-07-29 10:55:48,798][08566] Decorrelating experience for 96 frames...
[2025-07-29 10:55:48,803][08567] Decorrelating experience for 96 frames...
[2025-07-29 10:55:48,840][08570] Decorrelating experience for 96 frames...
[2025-07-29 10:55:48,842][08568] Decorrelating experience for 96 frames...
[2025-07-29 10:55:49,076][08564] Decorrelating experience for 96 frames...
[2025-07-29 10:55:49,148][08571] Decorrelating experience for 64 frames...
[2025-07-29 10:55:49,357][08569] Decorrelating experience for 64 frames...
[2025-07-29 10:55:49,431][08571] Decorrelating experience for 96 frames...
[2025-07-29 10:55:49,642][08569] Decorrelating experience for 96 frames...
[2025-07-29 10:55:49,648][08565] Decorrelating experience for 64 frames...
[2025-07-29 10:55:50,005][08565] Decorrelating experience for 96 frames...
[2025-07-29 10:55:50,274][08356] Fps is (10 sec: nan, 60 sec: nan, 300 sec: nan). Total num frames: 0. Throughput: 0: nan. Samples: 0. Policy #0 lag: (min: -1.0, avg: -1.0, max: -1.0)
[2025-07-29 10:55:50,275][08356] Avg episode reward: [(0, '1.797')]
[2025-07-29 10:55:50,491][08550] Signal inference workers to stop experience collection...
[2025-07-29 10:55:50,511][08563] InferenceWorker_p0-w0: stopping experience collection
[2025-07-29 10:55:51,622][08550] Signal inference workers to resume experience collection...
[2025-07-29 10:55:51,623][08563] InferenceWorker_p0-w0: resuming experience collection
[2025-07-29 10:55:53,372][08563] Updated weights for policy 0, policy_version 10 (0.0089)
[2025-07-29 10:55:55,274][08356] Fps is (10 sec: 15564.7, 60 sec: 15564.7, 300 sec: 15564.7). Total num frames: 77824. Throughput: 0: 2153.2. Samples: 10766. Policy #0 lag: (min: 0.0, avg: 0.9, max: 2.0)
[2025-07-29 10:55:55,275][08356] Avg episode reward: [(0, '4.516')]
[2025-07-29 10:55:55,404][08563] Updated weights for policy 0, policy_version 20 (0.0012)
[2025-07-29 10:55:57,391][08563] Updated weights for policy 0, policy_version 30 (0.0012)
[2025-07-29 10:55:59,416][08563] Updated weights for policy 0, policy_version 40 (0.0011)
[2025-07-29 10:55:59,626][08356] Heartbeat connected on Batcher_0
[2025-07-29 10:55:59,629][08356] Heartbeat connected on LearnerWorker_p0
[2025-07-29 10:55:59,638][08356] Heartbeat connected on InferenceWorker_p0-w0
[2025-07-29 10:55:59,646][08356] Heartbeat connected on RolloutWorker_w0
[2025-07-29 10:55:59,648][08356] Heartbeat connected on RolloutWorker_w1
[2025-07-29 10:55:59,652][08356] Heartbeat connected on RolloutWorker_w2
[2025-07-29 10:55:59,658][08356] Heartbeat connected on RolloutWorker_w4
[2025-07-29 10:55:59,661][08356] Heartbeat connected on RolloutWorker_w3
[2025-07-29 10:55:59,664][08356] Heartbeat connected on RolloutWorker_w5
[2025-07-29 10:55:59,665][08356] Heartbeat connected on RolloutWorker_w6
[2025-07-29 10:55:59,674][08356] Heartbeat connected on RolloutWorker_w7
[2025-07-29 10:56:00,274][08356] Fps is (10 sec: 18022.5, 60 sec: 18022.5, 300 sec: 18022.5). Total num frames: 180224. Throughput: 0: 4124.6. Samples: 41246. Policy #0 lag: (min: 0.0, avg: 1.0, max: 2.0)
[2025-07-29 10:56:00,275][08356] Avg episode reward: [(0, '4.550')]
[2025-07-29 10:56:00,276][08550] Saving new best policy, reward=4.550!
[2025-07-29 10:56:01,506][08563] Updated weights for policy 0, policy_version 50 (0.0011)
[2025-07-29 10:56:03,634][08563] Updated weights for policy 0, policy_version 60 (0.0012)
[2025-07-29 10:56:05,274][08356] Fps is (10 sec: 20070.3, 60 sec: 18568.5, 300 sec: 18568.5). Total num frames: 278528. Throughput: 0: 3729.1. Samples: 55936. Policy #0 lag: (min: 0.0, avg: 0.9, max: 2.0)
[2025-07-29 10:56:05,275][08356] Avg episode reward: [(0, '4.571')]
[2025-07-29 10:56:05,280][08550] Saving new best policy, reward=4.571!
[2025-07-29 10:56:05,659][08563] Updated weights for policy 0, policy_version 70 (0.0011)
[2025-07-29 10:56:07,647][08563] Updated weights for policy 0, policy_version 80 (0.0012)
[2025-07-29 10:56:09,674][08563] Updated weights for policy 0, policy_version 90 (0.0011)
[2025-07-29 10:56:10,274][08356] Fps is (10 sec: 20070.4, 60 sec: 19046.4, 300 sec: 19046.4). Total num frames: 380928. Throughput: 0: 4303.7. Samples: 86074. Policy #0 lag: (min: 0.0, avg: 0.9, max: 2.0)
[2025-07-29 10:56:10,275][08356] Avg episode reward: [(0, '4.404')]
[2025-07-29 10:56:11,679][08563] Updated weights for policy 0, policy_version 100 (0.0011)
[2025-07-29 10:56:13,663][08563] Updated weights for policy 0, policy_version 110 (0.0011)
[2025-07-29 10:56:15,274][08356] Fps is (10 sec: 20070.6, 60 sec: 19169.3, 300 sec: 19169.3). Total num frames: 479232. Throughput: 0: 4664.7. Samples: 116618. Policy #0 lag: (min: 0.0, avg: 1.0, max: 2.0)
[2025-07-29 10:56:15,275][08356] Avg episode reward: [(0, '4.467')]
[2025-07-29 10:56:15,750][08563] Updated weights for policy 0, policy_version 120 (0.0012)
[2025-07-29 10:56:17,778][08563] Updated weights for policy 0, policy_version 130 (0.0011)
[2025-07-29 10:56:19,789][08563] Updated weights for policy 0, policy_version 140 (0.0012)
[2025-07-29 10:56:20,274][08356] Fps is (10 sec: 20070.5, 60 sec: 19387.8, 300 sec: 19387.8). Total num frames: 581632. Throughput: 0: 4379.7. Samples: 131390. Policy #0 lag: (min: 0.0, avg: 0.9, max: 2.0)
[2025-07-29 10:56:20,275][08356] Avg episode reward: [(0, '4.520')]
[2025-07-29 10:56:21,795][08563] Updated weights for policy 0, policy_version 150 (0.0011)
[2025-07-29 10:56:23,802][08563] Updated weights for policy 0, policy_version 160 (0.0011)
[2025-07-29 10:56:25,274][08356] Fps is (10 sec: 20480.0, 60 sec: 19543.8, 300 sec: 19543.8). Total num frames: 684032. Throughput: 0: 4629.0. Samples: 162016. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 10:56:25,275][08356] Avg episode reward: [(0, '4.459')]
[2025-07-29 10:56:25,804][08563] Updated weights for policy 0, policy_version 170 (0.0011)
[2025-07-29 10:56:27,847][08563] Updated weights for policy 0, policy_version 180 (0.0011)
[2025-07-29 10:56:29,893][08563] Updated weights for policy 0, policy_version 190 (0.0011)
[2025-07-29 10:56:30,274][08356] Fps is (10 sec: 20070.4, 60 sec: 19558.4, 300 sec: 19558.4). Total num frames: 782336. Throughput: 0: 4804.8. Samples: 192190. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 10:56:30,275][08356] Avg episode reward: [(0, '4.757')]
[2025-07-29 10:56:30,293][08550] Saving new best policy, reward=4.757!
[2025-07-29 10:56:31,910][08563] Updated weights for policy 0, policy_version 200 (0.0012)
[2025-07-29 10:56:33,906][08563] Updated weights for policy 0, policy_version 210 (0.0011)
[2025-07-29 10:56:35,274][08356] Fps is (10 sec: 20070.5, 60 sec: 19660.8, 300 sec: 19660.8). Total num frames: 884736. Throughput: 0: 4610.9. Samples: 207490. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 10:56:35,275][08356] Avg episode reward: [(0, '4.508')]
[2025-07-29 10:56:35,931][08563] Updated weights for policy 0, policy_version 220 (0.0011)
[2025-07-29 10:56:37,943][08563] Updated weights for policy 0, policy_version 230 (0.0011)
[2025-07-29 10:56:39,991][08563] Updated weights for policy 0, policy_version 240 (0.0011)
[2025-07-29 10:56:40,274][08356] Fps is (10 sec: 20480.0, 60 sec: 19742.8, 300 sec: 19742.8). Total num frames: 987136. Throughput: 0: 5051.0. Samples: 238062. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 10:56:40,275][08356] Avg episode reward: [(0, '4.751')]
[2025-07-29 10:56:42,103][08563] Updated weights for policy 0, policy_version 250 (0.0012)
[2025-07-29 10:56:44,094][08563] Updated weights for policy 0, policy_version 260 (0.0012)
[2025-07-29 10:56:45,273][08356] Fps is (10 sec: 20070.5, 60 sec: 19735.3, 300 sec: 19735.3). Total num frames: 1085440. Throughput: 0: 5037.5. Samples: 267934. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 10:56:45,275][08356] Avg episode reward: [(0, '4.370')]
[2025-07-29 10:56:46,099][08563] Updated weights for policy 0, policy_version 270 (0.0011)
[2025-07-29 10:56:48,100][08563] Updated weights for policy 0, policy_version 280 (0.0011)
[2025-07-29 10:56:50,133][08563] Updated weights for policy 0, policy_version 290 (0.0011)
[2025-07-29 10:56:50,274][08356] Fps is (10 sec: 20070.3, 60 sec: 19797.3, 300 sec: 19797.3). Total num frames: 1187840. Throughput: 0: 5052.9. Samples: 283316. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 10:56:50,275][08356] Avg episode reward: [(0, '4.721')]
[2025-07-29 10:56:52,161][08563] Updated weights for policy 0, policy_version 300 (0.0011)
[2025-07-29 10:56:54,240][08563] Updated weights for policy 0, policy_version 310 (0.0012)
[2025-07-29 10:56:55,274][08356] Fps is (10 sec: 20479.8, 60 sec: 20206.9, 300 sec: 19849.9). Total num frames: 1290240. Throughput: 0: 5056.0. Samples: 313594. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 10:56:55,275][08356] Avg episode reward: [(0, '4.902')]
[2025-07-29 10:56:55,280][08550] Saving new best policy, reward=4.902!
[2025-07-29 10:56:56,282][08563] Updated weights for policy 0, policy_version 320 (0.0011)
[2025-07-29 10:56:58,277][08563] Updated weights for policy 0, policy_version 330 (0.0011)
[2025-07-29 10:57:00,274][08356] Fps is (10 sec: 20070.6, 60 sec: 20138.7, 300 sec: 19836.4). Total num frames: 1388544. Throughput: 0: 5048.8. Samples: 343812. Policy #0 lag: (min: 0.0, avg: 0.7, max: 1.0)
[2025-07-29 10:57:00,275][08356] Avg episode reward: [(0, '5.191')]
[2025-07-29 10:57:00,284][08550] Saving new best policy, reward=5.191!
[2025-07-29 10:57:00,287][08563] Updated weights for policy 0, policy_version 340 (0.0012)
[2025-07-29 10:57:02,313][08563] Updated weights for policy 0, policy_version 350 (0.0011)
[2025-07-29 10:57:04,335][08563] Updated weights for policy 0, policy_version 360 (0.0012)
[2025-07-29 10:57:05,274][08356] Fps is (10 sec: 20070.5, 60 sec: 20207.0, 300 sec: 19879.3). Total num frames: 1490944. Throughput: 0: 5060.0. Samples: 359092. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 10:57:05,275][08356] Avg episode reward: [(0, '5.012')]
[2025-07-29 10:57:06,453][08563] Updated weights for policy 0, policy_version 370 (0.0012)
[2025-07-29 10:57:08,514][08563] Updated weights for policy 0, policy_version 380 (0.0011)
[2025-07-29 10:57:10,274][08356] Fps is (10 sec: 20070.3, 60 sec: 20138.7, 300 sec: 19865.6). Total num frames: 1589248. Throughput: 0: 5038.0. Samples: 388726. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 10:57:10,275][08356] Avg episode reward: [(0, '5.173')]
[2025-07-29 10:57:10,534][08563] Updated weights for policy 0, policy_version 390 (0.0011)
[2025-07-29 10:57:12,578][08563] Updated weights for policy 0, policy_version 400 (0.0011)
[2025-07-29 10:57:14,609][08563] Updated weights for policy 0, policy_version 410 (0.0011)
[2025-07-29 10:57:15,274][08356] Fps is (10 sec: 20070.5, 60 sec: 20207.0, 300 sec: 19901.8). Total num frames: 1691648. Throughput: 0: 5038.2. Samples: 418908. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 10:57:15,275][08356] Avg episode reward: [(0, '5.349')]
[2025-07-29 10:57:15,280][08550] Saving new best policy, reward=5.349!
[2025-07-29 10:57:16,630][08563] Updated weights for policy 0, policy_version 420 (0.0011)
[2025-07-29 10:57:18,679][08563] Updated weights for policy 0, policy_version 430 (0.0011)
[2025-07-29 10:57:20,274][08356] Fps is (10 sec: 20070.4, 60 sec: 20138.7, 300 sec: 19888.4). Total num frames: 1789952. Throughput: 0: 5034.3. Samples: 434034. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 10:57:20,275][08356] Avg episode reward: [(0, '6.162')]
[2025-07-29 10:57:20,276][08550] Saving new best policy, reward=6.162!
[2025-07-29 10:57:20,752][08563] Updated weights for policy 0, policy_version 440 (0.0012)
[2025-07-29 10:57:22,752][08563] Updated weights for policy 0, policy_version 450 (0.0012)
[2025-07-29 10:57:24,748][08563] Updated weights for policy 0, policy_version 460 (0.0011)
[2025-07-29 10:57:25,274][08356] Fps is (10 sec: 20070.2, 60 sec: 20138.6, 300 sec: 19919.5). Total num frames: 1892352. Throughput: 0: 5024.7. Samples: 464174. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 10:57:25,275][08356] Avg episode reward: [(0, '6.499')]
[2025-07-29 10:57:25,280][08550] Saving new best policy, reward=6.499!
[2025-07-29 10:57:26,729][08563] Updated weights for policy 0, policy_version 470 (0.0012)
[2025-07-29 10:57:28,736][08563] Updated weights for policy 0, policy_version 480 (0.0012)
[2025-07-29 10:57:30,274][08356] Fps is (10 sec: 20479.2, 60 sec: 20206.8, 300 sec: 19947.5). Total num frames: 1994752. Throughput: 0: 5044.4. Samples: 494934. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0)
[2025-07-29 10:57:30,275][08356] Avg episode reward: [(0, '6.789')]
[2025-07-29 10:57:30,277][08550] Saving new best policy, reward=6.789!
[2025-07-29 10:57:30,726][08563] Updated weights for policy 0, policy_version 490 (0.0011)
[2025-07-29 10:57:32,843][08563] Updated weights for policy 0, policy_version 500 (0.0012)
[2025-07-29 10:57:34,824][08563] Updated weights for policy 0, policy_version 510 (0.0012)
[2025-07-29 10:57:35,274][08356] Fps is (10 sec: 20480.2, 60 sec: 20206.9, 300 sec: 19972.9). Total num frames: 2097152. Throughput: 0: 5032.1. Samples: 509760. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0)
[2025-07-29 10:57:35,275][08356] Avg episode reward: [(0, '6.559')]
[2025-07-29 10:57:35,281][08550] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000512_2097152.pth...
[2025-07-29 10:57:36,832][08563] Updated weights for policy 0, policy_version 520 (0.0012)
[2025-07-29 10:57:38,851][08563] Updated weights for policy 0, policy_version 530 (0.0012)
[2025-07-29 10:57:40,274][08356] Fps is (10 sec: 20480.9, 60 sec: 20206.9, 300 sec: 19995.9). Total num frames: 2199552. Throughput: 0: 5039.7. Samples: 540382. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0)
[2025-07-29 10:57:40,275][08356] Avg episode reward: [(0, '6.527')]
[2025-07-29 10:57:40,873][08563] Updated weights for policy 0, policy_version 540 (0.0011)
[2025-07-29 10:57:42,855][08563] Updated weights for policy 0, policy_version 550 (0.0012)
[2025-07-29 10:57:44,901][08563] Updated weights for policy 0, policy_version 560 (0.0011)
[2025-07-29 10:57:45,274][08356] Fps is (10 sec: 20070.2, 60 sec: 20206.9, 300 sec: 19981.4). Total num frames: 2297856. Throughput: 0: 5047.5. Samples: 570952. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 10:57:45,275][08356] Avg episode reward: [(0, '8.490')]
[2025-07-29 10:57:45,280][08550] Saving new best policy, reward=8.490!
[2025-07-29 10:57:46,945][08563] Updated weights for policy 0, policy_version 570 (0.0011)
[2025-07-29 10:57:48,931][08563] Updated weights for policy 0, policy_version 580 (0.0012)
[2025-07-29 10:57:50,274][08356] Fps is (10 sec: 20070.3, 60 sec: 20206.9, 300 sec: 20002.1). Total num frames: 2400256. Throughput: 0: 5041.8. Samples: 585974. Policy #0 lag: (min: 0.0, avg: 0.5, max: 1.0)
[2025-07-29 10:57:50,275][08356] Avg episode reward: [(0, '8.160')]
[2025-07-29 10:57:50,924][08563] Updated weights for policy 0, policy_version 590 (0.0011)
[2025-07-29 10:57:52,900][08563] Updated weights for policy 0, policy_version 600 (0.0012)
[2025-07-29 10:57:54,902][08563] Updated weights for policy 0, policy_version 610 (0.0011)
[2025-07-29 10:57:55,274][08356] Fps is (10 sec: 20480.2, 60 sec: 20207.0, 300 sec: 20021.3). Total num frames: 2502656. Throughput: 0: 5068.9. Samples: 616826. Policy #0 lag: (min: 0.0, avg: 0.5, max: 1.0)
[2025-07-29 10:57:55,275][08356] Avg episode reward: [(0, '8.444')]
[2025-07-29 10:57:56,949][08563] Updated weights for policy 0, policy_version 620 (0.0012)
[2025-07-29 10:57:59,016][08563] Updated weights for policy 0, policy_version 630 (0.0011)
[2025-07-29 10:58:00,274][08356] Fps is (10 sec: 20479.9, 60 sec: 20275.2, 300 sec: 20038.9). Total num frames: 2605056. Throughput: 0: 5073.1. Samples: 647196. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 10:58:00,275][08356] Avg episode reward: [(0, '10.207')]
[2025-07-29 10:58:00,276][08550] Saving new best policy, reward=10.207!
[2025-07-29 10:58:01,028][08563] Updated weights for policy 0, policy_version 640 (0.0011)
[2025-07-29 10:58:03,022][08563] Updated weights for policy 0, policy_version 650 (0.0011)
[2025-07-29 10:58:04,990][08563] Updated weights for policy 0, policy_version 660 (0.0011)
[2025-07-29 10:58:05,274][08356] Fps is (10 sec: 20479.9, 60 sec: 20275.2, 300 sec: 20055.2). Total num frames: 2707456. Throughput: 0: 5076.4. Samples: 662472. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 10:58:05,275][08356] Avg episode reward: [(0, '11.219')]
[2025-07-29 10:58:05,279][08550] Saving new best policy, reward=11.219!
[2025-07-29 10:58:06,990][08563] Updated weights for policy 0, policy_version 670 (0.0011)
[2025-07-29 10:58:09,008][08563] Updated weights for policy 0, policy_version 680 (0.0012)
[2025-07-29 10:58:10,274][08356] Fps is (10 sec: 20480.0, 60 sec: 20343.5, 300 sec: 20070.4). Total num frames: 2809856. Throughput: 0: 5094.7. Samples: 693436. Policy #0 lag: (min: 0.0, avg: 0.8, max: 1.0)
[2025-07-29 10:58:10,275][08356] Avg episode reward: [(0, '11.927')]
[2025-07-29 10:58:10,277][08550] Saving new best policy, reward=11.927!
[2025-07-29 10:58:11,089][08563] Updated weights for policy 0, policy_version 690 (0.0012)
[2025-07-29 10:58:13,076][08563] Updated weights for policy 0, policy_version 700 (0.0011)
[2025-07-29 10:58:15,040][08563] Updated weights for policy 0, policy_version 710 (0.0011)
[2025-07-29 10:58:15,274][08356] Fps is (10 sec: 20479.9, 60 sec: 20343.4, 300 sec: 20084.5). Total num frames: 2912256. Throughput: 0: 5086.2. Samples: 723810. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 10:58:15,275][08356] Avg episode reward: [(0, '13.472')]
[2025-07-29 10:58:15,280][08550] Saving new best policy, reward=13.472!
[2025-07-29 10:58:17,025][08563] Updated weights for policy 0, policy_version 720 (0.0011)
[2025-07-29 10:58:19,021][08563] Updated weights for policy 0, policy_version 730 (0.0012)
[2025-07-29 10:58:20,274][08356] Fps is (10 sec: 20480.1, 60 sec: 20411.7, 300 sec: 20097.7). Total num frames: 3014656. Throughput: 0: 5098.8. Samples: 739208. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 10:58:20,275][08356] Avg episode reward: [(0, '14.176')]
[2025-07-29 10:58:20,276][08550] Saving new best policy, reward=14.176!
[2025-07-29 10:58:20,985][08563] Updated weights for policy 0, policy_version 740 (0.0011)
[2025-07-29 10:58:23,036][08563] Updated weights for policy 0, policy_version 750 (0.0012)
[2025-07-29 10:58:25,091][08563] Updated weights for policy 0, policy_version 760 (0.0012)
[2025-07-29 10:58:25,274][08356] Fps is (10 sec: 20070.5, 60 sec: 20343.5, 300 sec: 20083.6). Total num frames: 3112960. Throughput: 0: 5097.9. Samples: 769790. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 10:58:25,275][08356] Avg episode reward: [(0, '15.850')]
[2025-07-29 10:58:25,283][08550] Saving new best policy, reward=15.850!
[2025-07-29 10:58:27,069][08563] Updated weights for policy 0, policy_version 770 (0.0011)
[2025-07-29 10:58:29,052][08563] Updated weights for policy 0, policy_version 780 (0.0011)
[2025-07-29 10:58:30,274][08356] Fps is (10 sec: 20480.1, 60 sec: 20411.9, 300 sec: 20121.6). Total num frames: 3219456. Throughput: 0: 5104.4. Samples: 800650. Policy #0 lag: (min: 0.0, avg: 0.7, max: 1.0)
[2025-07-29 10:58:30,275][08356] Avg episode reward: [(0, '15.208')]
[2025-07-29 10:58:31,009][08563] Updated weights for policy 0, policy_version 790 (0.0011)
[2025-07-29 10:58:32,979][08563] Updated weights for policy 0, policy_version 800 (0.0011)
[2025-07-29 10:58:35,005][08563] Updated weights for policy 0, policy_version 810 (0.0012)
[2025-07-29 10:58:35,274][08356] Fps is (10 sec: 20889.5, 60 sec: 20411.7, 300 sec: 20132.5). Total num frames: 3321856. Throughput: 0: 5118.5. Samples: 816308. Policy #0 lag: (min: 0.0, avg: 0.7, max: 1.0)
[2025-07-29 10:58:35,276][08356] Avg episode reward: [(0, '17.220')]
[2025-07-29 10:58:35,282][08550] Saving new best policy, reward=17.220!
[2025-07-29 10:58:37,067][08563] Updated weights for policy 0, policy_version 820 (0.0012)
[2025-07-29 10:58:39,042][08563] Updated weights for policy 0, policy_version 830 (0.0012)
[2025-07-29 10:58:40,274][08356] Fps is (10 sec: 20479.9, 60 sec: 20411.7, 300 sec: 20142.7). Total num frames: 3424256. Throughput: 0: 5102.4. Samples: 846436. Policy #0 lag: (min: 0.0, avg: 0.5, max: 1.0)
[2025-07-29 10:58:40,275][08356] Avg episode reward: [(0, '17.562')]
[2025-07-29 10:58:40,276][08550] Saving new best policy, reward=17.562!
[2025-07-29 10:58:41,034][08563] Updated weights for policy 0, policy_version 840 (0.0011)
[2025-07-29 10:58:43,024][08563] Updated weights for policy 0, policy_version 850 (0.0011)
[2025-07-29 10:58:44,978][08563] Updated weights for policy 0, policy_version 860 (0.0011)
[2025-07-29 10:58:45,274][08356] Fps is (10 sec: 20480.1, 60 sec: 20480.0, 300 sec: 20152.3). Total num frames: 3526656. Throughput: 0: 5116.0. Samples: 877416. Policy #0 lag: (min: 0.0, avg: 0.4, max: 1.0)
[2025-07-29 10:58:45,275][08356] Avg episode reward: [(0, '19.748')]
[2025-07-29 10:58:45,281][08550] Saving new best policy, reward=19.748!
[2025-07-29 10:58:46,986][08563] Updated weights for policy 0, policy_version 870 (0.0011)
[2025-07-29 10:58:49,061][08563] Updated weights for policy 0, policy_version 880 (0.0012)
[2025-07-29 10:58:50,274][08356] Fps is (10 sec: 20480.0, 60 sec: 20480.0, 300 sec: 20161.4). Total num frames: 3629056. Throughput: 0: 5121.9. Samples: 892956. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 10:58:50,275][08356] Avg episode reward: [(0, '20.681')]
[2025-07-29 10:58:50,276][08550] Saving new best policy, reward=20.681!
[2025-07-29 10:58:51,043][08563] Updated weights for policy 0, policy_version 890 (0.0011)
[2025-07-29 10:58:53,004][08563] Updated weights for policy 0, policy_version 900 (0.0011)
[2025-07-29 10:58:54,984][08563] Updated weights for policy 0, policy_version 910 (0.0011)
[2025-07-29 10:58:55,274][08356] Fps is (10 sec: 20480.0, 60 sec: 20480.0, 300 sec: 20170.0). Total num frames: 3731456. Throughput: 0: 5114.8. Samples: 923600. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 10:58:55,275][08356] Avg episode reward: [(0, '22.012')]
[2025-07-29 10:58:55,281][08550] Saving new best policy, reward=22.012!
[2025-07-29 10:58:56,978][08563] Updated weights for policy 0, policy_version 920 (0.0012)
[2025-07-29 10:58:58,950][08563] Updated weights for policy 0, policy_version 930 (0.0011)
[2025-07-29 10:59:00,274][08356] Fps is (10 sec: 20480.0, 60 sec: 20480.0, 300 sec: 20178.2). Total num frames: 3833856. Throughput: 0: 5128.0. Samples: 954570. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 10:59:00,275][08356] Avg episode reward: [(0, '21.860')]
[2025-07-29 10:59:01,008][08563] Updated weights for policy 0, policy_version 940 (0.0011)
[2025-07-29 10:59:03,040][08563] Updated weights for policy 0, policy_version 950 (0.0012)
[2025-07-29 10:59:05,015][08563] Updated weights for policy 0, policy_version 960 (0.0011)
[2025-07-29 10:59:05,274][08356] Fps is (10 sec: 20480.1, 60 sec: 20480.0, 300 sec: 20185.9). Total num frames: 3936256. Throughput: 0: 5116.3. Samples: 969442. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 10:59:05,275][08356] Avg episode reward: [(0, '24.220')]
[2025-07-29 10:59:05,281][08550] Saving new best policy, reward=24.220!
[2025-07-29 10:59:06,977][08563] Updated weights for policy 0, policy_version 970 (0.0011)
[2025-07-29 10:59:08,542][08550] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:08,543][08356] Component Batcher_0 stopped!
[2025-07-29 10:59:08,543][08550] Stopping Batcher_0...
[2025-07-29 10:59:08,548][08550] Loop batcher_evt_loop terminating...
[2025-07-29 10:59:08,565][08563] Weights refcount: 2 0
[2025-07-29 10:59:08,567][08563] Stopping InferenceWorker_p0-w0...
[2025-07-29 10:59:08,567][08563] Loop inference_proc0-0_evt_loop terminating...
[2025-07-29 10:59:08,567][08356] Component InferenceWorker_p0-w0 stopped!
[2025-07-29 10:59:08,584][08569] Stopping RolloutWorker_w5...
[2025-07-29 10:59:08,584][08571] Stopping RolloutWorker_w6...
[2025-07-29 10:59:08,585][08571] Loop rollout_proc6_evt_loop terminating...
[2025-07-29 10:59:08,584][08356] Component RolloutWorker_w5 stopped!
[2025-07-29 10:59:08,585][08569] Loop rollout_proc5_evt_loop terminating...
[2025-07-29 10:59:08,586][08566] Stopping RolloutWorker_w2...
[2025-07-29 10:59:08,586][08568] Stopping RolloutWorker_w4...
[2025-07-29 10:59:08,586][08566] Loop rollout_proc2_evt_loop terminating...
[2025-07-29 10:59:08,587][08568] Loop rollout_proc4_evt_loop terminating...
[2025-07-29 10:59:08,587][08567] Stopping RolloutWorker_w3...
[2025-07-29 10:59:08,586][08356] Component RolloutWorker_w6 stopped!
[2025-07-29 10:59:08,588][08567] Loop rollout_proc3_evt_loop terminating...
[2025-07-29 10:59:08,588][08570] Stopping RolloutWorker_w7...
[2025-07-29 10:59:08,588][08356] Component RolloutWorker_w2 stopped!
[2025-07-29 10:59:08,589][08570] Loop rollout_proc7_evt_loop terminating...
[2025-07-29 10:59:08,589][08564] Stopping RolloutWorker_w0...
[2025-07-29 10:59:08,590][08564] Loop rollout_proc0_evt_loop terminating...
[2025-07-29 10:59:08,589][08356] Component RolloutWorker_w4 stopped!
[2025-07-29 10:59:08,590][08565] Stopping RolloutWorker_w1...
[2025-07-29 10:59:08,591][08565] Loop rollout_proc1_evt_loop terminating...
[2025-07-29 10:59:08,590][08356] Component RolloutWorker_w3 stopped!
[2025-07-29 10:59:08,591][08356] Component RolloutWorker_w7 stopped!
[2025-07-29 10:59:08,592][08356] Component RolloutWorker_w0 stopped!
[2025-07-29 10:59:08,593][08356] Component RolloutWorker_w1 stopped!
[2025-07-29 10:59:08,616][08550] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:08,720][08550] Stopping LearnerWorker_p0...
[2025-07-29 10:59:08,721][08550] Loop learner_proc0_evt_loop terminating...
[2025-07-29 10:59:08,720][08356] Component LearnerWorker_p0 stopped!
[2025-07-29 10:59:08,721][08356] Waiting for process learner_proc0 to stop...
[2025-07-29 10:59:09,675][08356] Waiting for process inference_proc0-0 to join...
[2025-07-29 10:59:09,676][08356] Waiting for process rollout_proc0 to join...
[2025-07-29 10:59:09,677][08356] Waiting for process rollout_proc1 to join...
[2025-07-29 10:59:09,678][08356] Waiting for process rollout_proc2 to join...
[2025-07-29 10:59:09,678][08356] Waiting for process rollout_proc3 to join...
[2025-07-29 10:59:09,679][08356] Waiting for process rollout_proc4 to join...
[2025-07-29 10:59:09,680][08356] Waiting for process rollout_proc5 to join...
[2025-07-29 10:59:09,680][08356] Waiting for process rollout_proc6 to join...
[2025-07-29 10:59:09,681][08356] Waiting for process rollout_proc7 to join...
[2025-07-29 10:59:09,682][08356] Batcher 0 profile tree view:
batching: 12.3062, releasing_batches: 0.0226
[2025-07-29 10:59:09,683][08356] InferenceWorker_p0-w0 profile tree view:
wait_policy: 0.0001
  wait_policy_total: 3.8979
update_model: 3.2072
  weight_update: 0.0011
one_step: 0.0028
  handle_policy_step: 185.2466
    deserialize: 7.5720, stack: 1.2933, obs_to_device_normalize: 45.3519, forward: 88.9795, send_messages: 12.7549
    prepare_outputs: 22.0606
      to_cpu: 14.0958
[2025-07-29 10:59:09,684][08356] Learner 0 profile tree view:
misc: 0.0036, prepare_batch: 6.5850
train: 18.8969
  epoch_init: 0.0041, minibatch_init: 0.0056, losses_postprocess: 0.3338, kl_divergence: 0.3881, after_optimizer: 1.9327
  calculate_losses: 8.5966
    losses_init: 0.0031, forward_head: 0.6340, bptt_initial: 4.6393, tail: 0.6094, advantages_returns: 0.1555, losses: 1.2006
    bptt: 1.2132
      bptt_forward_core: 1.1620
  update: 7.3269
    clip: 0.7956
[2025-07-29 10:59:09,684][08356] RolloutWorker_w0 profile tree view:
wait_for_trajectories: 0.1235, enqueue_policy_requests: 8.8549, env_step: 127.0001, overhead: 5.4236, complete_rollouts: 0.2088
save_policy_outputs: 7.9631
  split_output_tensors: 3.0557
[2025-07-29 10:59:09,685][08356] RolloutWorker_w7 profile tree view:
wait_for_trajectories: 0.1222, enqueue_policy_requests: 8.8302, env_step: 127.1710, overhead: 5.5206, complete_rollouts: 0.2087
save_policy_outputs: 7.9508
  split_output_tensors: 3.0303
[2025-07-29 10:59:09,686][08356] Loop Runner_EvtLoop terminating...
[2025-07-29 10:59:09,687][08356] Runner profile tree view:
main_loop: 210.0184
[2025-07-29 10:59:09,687][08356] Collected {0: 4005888}, FPS: 19074.0
[2025-07-29 10:59:21,569][08356] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 10:59:21,570][08356] Overriding arg 'num_workers' with value 1 passed from command line
[2025-07-29 10:59:21,570][08356] Adding new argument 'no_render'=True that is not in the saved config file!
[2025-07-29 10:59:21,571][08356] Adding new argument 'save_video'=True that is not in the saved config file!
[2025-07-29 10:59:21,572][08356] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 10:59:21,572][08356] Adding new argument 'video_name'=None that is not in the saved config file!
[2025-07-29 10:59:21,573][08356] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 10:59:21,574][08356] Adding new argument 'max_num_episodes'=10 that is not in the saved config file!
[2025-07-29 10:59:21,575][08356] Adding new argument 'push_to_hub'=False that is not in the saved config file!
[2025-07-29 10:59:21,575][08356] Adding new argument 'hf_repository'=None that is not in the saved config file!
[2025-07-29 10:59:21,576][08356] Adding new argument 'policy_index'=0 that is not in the saved config file!
[2025-07-29 10:59:21,576][08356] Adding new argument 'eval_deterministic'=False that is not in the saved config file!
[2025-07-29 10:59:21,577][08356] Adding new argument 'train_script'=None that is not in the saved config file!
[2025-07-29 10:59:21,578][08356] Adding new argument 'enjoy_script'=None that is not in the saved config file!
[2025-07-29 10:59:21,578][08356] Using frameskip 1 and render_action_repeat=4 for evaluation
[2025-07-29 10:59:21,608][08356] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 10:59:21,611][08356] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 10:59:21,612][08356] RunningMeanStd input shape: (1,)
[2025-07-29 10:59:21,625][08356] ConvEncoder: input_channels=3
[2025-07-29 10:59:21,733][08356] Conv encoder output size: 512
[2025-07-29 10:59:21,734][08356] Policy head output size: 512
[2025-07-29 10:59:21,918][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:21,920][08356] Could not load from checkpoint, attempt 0
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 10:59:21,922][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:21,923][08356] Could not load from checkpoint, attempt 1
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 10:59:21,924][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:21,925][08356] Could not load from checkpoint, attempt 2
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 10:59:43,523][08356] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 10:59:43,523][08356] Overriding arg 'num_workers' with value 1 passed from command line
[2025-07-29 10:59:43,524][08356] Adding new argument 'no_render'=True that is not in the saved config file!
[2025-07-29 10:59:43,525][08356] Adding new argument 'save_video'=True that is not in the saved config file!
[2025-07-29 10:59:43,526][08356] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 10:59:43,527][08356] Adding new argument 'video_name'=None that is not in the saved config file!
[2025-07-29 10:59:43,527][08356] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 10:59:43,528][08356] Adding new argument 'max_num_episodes'=10 that is not in the saved config file!
[2025-07-29 10:59:43,529][08356] Adding new argument 'push_to_hub'=False that is not in the saved config file!
[2025-07-29 10:59:43,530][08356] Adding new argument 'hf_repository'=None that is not in the saved config file!
[2025-07-29 10:59:43,531][08356] Adding new argument 'policy_index'=0 that is not in the saved config file!
[2025-07-29 10:59:43,532][08356] Adding new argument 'eval_deterministic'=False that is not in the saved config file!
[2025-07-29 10:59:43,532][08356] Adding new argument 'train_script'=None that is not in the saved config file!
[2025-07-29 10:59:43,533][08356] Adding new argument 'enjoy_script'=None that is not in the saved config file!
[2025-07-29 10:59:43,533][08356] Using frameskip 1 and render_action_repeat=4 for evaluation
[2025-07-29 10:59:43,562][08356] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 10:59:43,563][08356] RunningMeanStd input shape: (1,)
[2025-07-29 10:59:43,572][08356] ConvEncoder: input_channels=3
[2025-07-29 10:59:43,608][08356] Conv encoder output size: 512
[2025-07-29 10:59:43,609][08356] Policy head output size: 512
[2025-07-29 10:59:43,627][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:43,628][08356] Could not load from checkpoint, attempt 0
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 10:59:43,629][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:43,630][08356] Could not load from checkpoint, attempt 1
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 10:59:43,631][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:43,633][08356] Could not load from checkpoint, attempt 2
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 10:59:49,480][08356] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 10:59:49,481][08356] Overriding arg 'num_workers' with value 1 passed from command line
[2025-07-29 10:59:49,481][08356] Adding new argument 'no_render'=True that is not in the saved config file!
[2025-07-29 10:59:49,482][08356] Adding new argument 'save_video'=True that is not in the saved config file!
[2025-07-29 10:59:49,483][08356] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 10:59:49,483][08356] Adding new argument 'video_name'=None that is not in the saved config file!
[2025-07-29 10:59:49,484][08356] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 10:59:49,485][08356] Adding new argument 'max_num_episodes'=10 that is not in the saved config file!
[2025-07-29 10:59:49,485][08356] Adding new argument 'push_to_hub'=False that is not in the saved config file!
[2025-07-29 10:59:49,486][08356] Adding new argument 'hf_repository'=None that is not in the saved config file!
[2025-07-29 10:59:49,487][08356] Adding new argument 'policy_index'=0 that is not in the saved config file!
[2025-07-29 10:59:49,487][08356] Adding new argument 'eval_deterministic'=False that is not in the saved config file!
[2025-07-29 10:59:49,488][08356] Adding new argument 'train_script'=None that is not in the saved config file!
[2025-07-29 10:59:49,488][08356] Adding new argument 'enjoy_script'=None that is not in the saved config file!
[2025-07-29 10:59:49,490][08356] Using frameskip 1 and render_action_repeat=4 for evaluation
[2025-07-29 10:59:49,513][08356] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 10:59:49,514][08356] RunningMeanStd input shape: (1,)
[2025-07-29 10:59:49,523][08356] ConvEncoder: input_channels=3
[2025-07-29 10:59:49,556][08356] Conv encoder output size: 512
[2025-07-29 10:59:49,557][08356] Policy head output size: 512
[2025-07-29 10:59:49,574][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:49,577][08356] Could not load from checkpoint, attempt 0
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 10:59:49,578][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:49,579][08356] Could not load from checkpoint, attempt 1
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 10:59:49,580][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 10:59:49,581][08356] Could not load from checkpoint, attempt 2
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:00:31,147][08356] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 11:00:31,148][08356] Overriding arg 'num_workers' with value 1 passed from command line
[2025-07-29 11:00:31,149][08356] Adding new argument 'no_render'=True that is not in the saved config file!
[2025-07-29 11:00:31,149][08356] Adding new argument 'save_video'=True that is not in the saved config file!
[2025-07-29 11:00:31,150][08356] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:00:31,151][08356] Adding new argument 'video_name'=None that is not in the saved config file!
[2025-07-29 11:00:31,151][08356] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:00:31,152][08356] Adding new argument 'max_num_episodes'=10 that is not in the saved config file!
[2025-07-29 11:00:31,152][08356] Adding new argument 'push_to_hub'=False that is not in the saved config file!
[2025-07-29 11:00:31,154][08356] Adding new argument 'hf_repository'=None that is not in the saved config file!
[2025-07-29 11:00:31,154][08356] Adding new argument 'policy_index'=0 that is not in the saved config file!
[2025-07-29 11:00:31,155][08356] Adding new argument 'eval_deterministic'=False that is not in the saved config file!
[2025-07-29 11:00:31,155][08356] Adding new argument 'train_script'=None that is not in the saved config file!
[2025-07-29 11:00:31,156][08356] Adding new argument 'enjoy_script'=None that is not in the saved config file!
[2025-07-29 11:00:31,157][08356] Using frameskip 1 and render_action_repeat=4 for evaluation
[2025-07-29 11:00:31,182][08356] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:00:31,183][08356] RunningMeanStd input shape: (1,)
[2025-07-29 11:00:31,192][08356] ConvEncoder: input_channels=3
[2025-07-29 11:00:31,227][08356] Conv encoder output size: 512
[2025-07-29 11:00:31,228][08356] Policy head output size: 512
[2025-07-29 11:00:31,247][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:00:31,249][08356] Could not load from checkpoint, attempt 0
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:00:31,250][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:00:31,251][08356] Could not load from checkpoint, attempt 1
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:00:31,251][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:00:31,252][08356] Could not load from checkpoint, attempt 2
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:00:50,477][08356] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 11:00:50,478][08356] Overriding arg 'num_workers' with value 1 passed from command line
[2025-07-29 11:00:50,478][08356] Adding new argument 'no_render'=True that is not in the saved config file!
[2025-07-29 11:00:50,479][08356] Adding new argument 'save_video'=True that is not in the saved config file!
[2025-07-29 11:00:50,480][08356] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:00:50,481][08356] Adding new argument 'video_name'=None that is not in the saved config file!
[2025-07-29 11:00:50,481][08356] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:00:50,482][08356] Adding new argument 'max_num_episodes'=5 that is not in the saved config file!
[2025-07-29 11:00:50,482][08356] Adding new argument 'push_to_hub'=False that is not in the saved config file!
[2025-07-29 11:00:50,483][08356] Adding new argument 'hf_repository'=None that is not in the saved config file!
[2025-07-29 11:00:50,484][08356] Adding new argument 'policy_index'=0 that is not in the saved config file!
[2025-07-29 11:00:50,484][08356] Adding new argument 'eval_deterministic'=False that is not in the saved config file!
[2025-07-29 11:00:50,485][08356] Adding new argument 'train_script'=None that is not in the saved config file!
[2025-07-29 11:00:50,486][08356] Adding new argument 'enjoy_script'=None that is not in the saved config file!
[2025-07-29 11:00:50,487][08356] Using frameskip 1 and render_action_repeat=4 for evaluation
[2025-07-29 11:00:50,512][08356] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:00:50,514][08356] RunningMeanStd input shape: (1,)
[2025-07-29 11:00:50,523][08356] ConvEncoder: input_channels=3
[2025-07-29 11:00:50,558][08356] Conv encoder output size: 512
[2025-07-29 11:00:50,559][08356] Policy head output size: 512
[2025-07-29 11:00:50,579][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:00:50,580][08356] Could not load from checkpoint, attempt 0
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:00:50,581][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:00:50,582][08356] Could not load from checkpoint, attempt 1
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:00:50,583][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:00:50,584][08356] Could not load from checkpoint, attempt 2
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:05:34,005][08356] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 11:05:34,006][08356] Overriding arg 'num_workers' with value 1 passed from command line
[2025-07-29 11:05:34,007][08356] Adding new argument 'no_render'=True that is not in the saved config file!
[2025-07-29 11:05:34,008][08356] Adding new argument 'save_video'=True that is not in the saved config file!
[2025-07-29 11:05:34,008][08356] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:05:34,009][08356] Adding new argument 'video_name'=None that is not in the saved config file!
[2025-07-29 11:05:34,009][08356] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:05:34,010][08356] Adding new argument 'max_num_episodes'=10 that is not in the saved config file!
[2025-07-29 11:05:34,010][08356] Adding new argument 'push_to_hub'=False that is not in the saved config file!
[2025-07-29 11:05:34,012][08356] Adding new argument 'hf_repository'=None that is not in the saved config file!
[2025-07-29 11:05:34,012][08356] Adding new argument 'policy_index'=0 that is not in the saved config file!
[2025-07-29 11:05:34,012][08356] Adding new argument 'eval_deterministic'=False that is not in the saved config file!
[2025-07-29 11:05:34,013][08356] Adding new argument 'train_script'=None that is not in the saved config file!
[2025-07-29 11:05:34,014][08356] Adding new argument 'enjoy_script'=None that is not in the saved config file!
[2025-07-29 11:05:34,015][08356] Using frameskip 1 and render_action_repeat=4 for evaluation
[2025-07-29 11:05:34,041][08356] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:05:34,042][08356] RunningMeanStd input shape: (1,)
[2025-07-29 11:05:34,052][08356] ConvEncoder: input_channels=3
[2025-07-29 11:05:34,087][08356] Conv encoder output size: 512
[2025-07-29 11:05:34,088][08356] Policy head output size: 512
[2025-07-29 11:05:34,108][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:05:34,109][08356] Could not load from checkpoint, attempt 0
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:05:34,110][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:05:34,111][08356] Could not load from checkpoint, attempt 1
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:05:34,112][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:05:34,113][08356] Could not load from checkpoint, attempt 2
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:06:40,036][08356] Environment doom_basic already registered, overwriting...
[2025-07-29 11:06:40,037][08356] Environment doom_two_colors_easy already registered, overwriting...
[2025-07-29 11:06:40,037][08356] Environment doom_two_colors_hard already registered, overwriting...
[2025-07-29 11:06:40,038][08356] Environment doom_dm already registered, overwriting...
[2025-07-29 11:06:40,039][08356] Environment doom_dwango5 already registered, overwriting...
[2025-07-29 11:06:40,039][08356] Environment doom_my_way_home_flat_actions already registered, overwriting...
[2025-07-29 11:06:40,040][08356] Environment doom_defend_the_center_flat_actions already registered, overwriting...
[2025-07-29 11:06:40,041][08356] Environment doom_my_way_home already registered, overwriting...
[2025-07-29 11:06:40,041][08356] Environment doom_deadly_corridor already registered, overwriting...
[2025-07-29 11:06:40,042][08356] Environment doom_defend_the_center already registered, overwriting...
[2025-07-29 11:06:40,043][08356] Environment doom_defend_the_line already registered, overwriting...
[2025-07-29 11:06:40,043][08356] Environment doom_health_gathering already registered, overwriting...
[2025-07-29 11:06:40,044][08356] Environment doom_health_gathering_supreme already registered, overwriting...
[2025-07-29 11:06:40,045][08356] Environment doom_battle already registered, overwriting...
[2025-07-29 11:06:40,045][08356] Environment doom_battle2 already registered, overwriting...
[2025-07-29 11:06:40,046][08356] Environment doom_duel_bots already registered, overwriting...
[2025-07-29 11:06:40,046][08356] Environment doom_deathmatch_bots already registered, overwriting...
[2025-07-29 11:06:40,047][08356] Environment doom_duel already registered, overwriting...
[2025-07-29 11:06:40,048][08356] Environment doom_deathmatch_full already registered, overwriting...
[2025-07-29 11:06:40,049][08356] Environment doom_benchmark already registered, overwriting...
[2025-07-29 11:06:40,049][08356] register_encoder_factory: <function make_vizdoom_encoder at 0x7b3a471e7880>
[2025-07-29 11:06:40,057][08356] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 11:06:40,061][08356] Experiment dir /content/train_dir/default_experiment already exists!
[2025-07-29 11:06:40,062][08356] Resuming existing experiment from /content/train_dir/default_experiment...
[2025-07-29 11:06:40,062][08356] Weights and Biases integration disabled
[2025-07-29 11:06:40,064][08356] Environment var CUDA_VISIBLE_DEVICES is 0

[2025-07-29 11:06:42,354][08356] Starting experiment with the following configuration:
help=False
algo=APPO
env=doom_health_gathering_supreme
experiment=default_experiment
train_dir=/content/train_dir
restart_behavior=resume
device=gpu
seed=None
num_policies=1
async_rl=True
serial_mode=False
batched_sampling=False
num_batches_to_accumulate=2
worker_num_splits=2
policy_workers_per_policy=1
max_policy_lag=1000
num_workers=8
num_envs_per_worker=4
batch_size=1024
num_batches_per_epoch=1
num_epochs=1
rollout=32
recurrence=32
shuffle_minibatches=False
gamma=0.99
reward_scale=1.0
reward_clip=1000.0
value_bootstrap=False
normalize_returns=True
exploration_loss_coeff=0.001
value_loss_coeff=0.5
kl_loss_coeff=0.0
exploration_loss=symmetric_kl
gae_lambda=0.95
ppo_clip_ratio=0.1
ppo_clip_value=0.2
with_vtrace=False
vtrace_rho=1.0
vtrace_c=1.0
optimizer=adam
adam_eps=1e-06
adam_beta1=0.9
adam_beta2=0.999
max_grad_norm=4.0
learning_rate=0.0001
lr_schedule=constant
lr_schedule_kl_threshold=0.008
lr_adaptive_min=1e-06
lr_adaptive_max=0.01
obs_subtract_mean=0.0
obs_scale=255.0
normalize_input=True
normalize_input_keys=None
decorrelate_experience_max_seconds=0
decorrelate_envs_on_one_worker=True
actor_worker_gpus=[]
set_workers_cpu_affinity=True
force_envs_single_thread=False
default_niceness=0
log_to_file=True
experiment_summaries_interval=10
flush_summaries_interval=30
stats_avg=100
summaries_use_frameskip=True
heartbeat_interval=20
heartbeat_reporting_interval=600
train_for_env_steps=4000000
train_for_seconds=10000000000
save_every_sec=120
keep_checkpoints=2
load_checkpoint_kind=latest
save_milestones_sec=-1
save_best_every_sec=5
save_best_metric=reward
save_best_after=100000
benchmark=False
encoder_mlp_layers=[512, 512]
encoder_conv_architecture=convnet_simple
encoder_conv_mlp_layers=[512]
use_rnn=True
rnn_size=512
rnn_type=gru
rnn_num_layers=1
decoder_mlp_layers=[]
nonlinearity=elu
policy_initialization=orthogonal
policy_init_gain=1.0
actor_critic_share_weights=True
adaptive_stddev=True
continuous_tanh_scale=0.0
initial_stddev=1.0
use_env_info_cache=False
env_gpu_actions=False
env_gpu_observations=True
env_frameskip=4
env_framestack=1
pixel_format=CHW
use_record_episode_statistics=False
with_wandb=False
wandb_user=None
wandb_project=sample_factory
wandb_group=None
wandb_job_type=SF
wandb_tags=[]
with_pbt=False
pbt_mix_policies_in_one_env=True
pbt_period_env_steps=5000000
pbt_start_mutation=20000000
pbt_replace_fraction=0.3
pbt_mutation_rate=0.15
pbt_replace_reward_gap=0.1
pbt_replace_reward_gap_absolute=1e-06
pbt_optimize_gamma=False
pbt_target_objective=true_objective
pbt_perturb_min=1.1
pbt_perturb_max=1.5
num_agents=-1
num_humans=0
num_bots=-1
start_bot_difficulty=None
timelimit=None
res_w=128
res_h=72
wide_aspect_ratio=False
eval_env_frameskip=1
fps=35
command_line=--env=doom_health_gathering_supreme --num_workers=8 --num_envs_per_worker=4 --train_for_env_steps=4000000
cli_args={'env': 'doom_health_gathering_supreme', 'num_workers': 8, 'num_envs_per_worker': 4, 'train_for_env_steps': 4000000}
git_hash=unknown
git_repo_name=not a git repository
[2025-07-29 11:06:42,355][08356] Saving configuration to /content/train_dir/default_experiment/config.json...
[2025-07-29 11:06:42,357][08356] Rollout worker 0 uses device cpu
[2025-07-29 11:06:42,357][08356] Rollout worker 1 uses device cpu
[2025-07-29 11:06:42,358][08356] Rollout worker 2 uses device cpu
[2025-07-29 11:06:42,359][08356] Rollout worker 3 uses device cpu
[2025-07-29 11:06:42,360][08356] Rollout worker 4 uses device cpu
[2025-07-29 11:06:42,360][08356] Rollout worker 5 uses device cpu
[2025-07-29 11:06:42,361][08356] Rollout worker 6 uses device cpu
[2025-07-29 11:06:42,362][08356] Rollout worker 7 uses device cpu
[2025-07-29 11:06:42,401][08356] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:06:42,402][08356] InferenceWorker_p0-w0: min num requests: 2
[2025-07-29 11:06:42,431][08356] Starting all processes...
[2025-07-29 11:06:42,431][08356] Starting process learner_proc0
[2025-07-29 11:06:42,483][08356] Starting all processes...
[2025-07-29 11:06:42,488][08356] Starting process inference_proc0-0
[2025-07-29 11:06:42,488][08356] Starting process rollout_proc0
[2025-07-29 11:06:42,489][08356] Starting process rollout_proc1
[2025-07-29 11:06:42,490][08356] Starting process rollout_proc2
[2025-07-29 11:06:42,491][08356] Starting process rollout_proc3
[2025-07-29 11:06:42,491][08356] Starting process rollout_proc4
[2025-07-29 11:06:42,492][08356] Starting process rollout_proc5
[2025-07-29 11:06:42,492][08356] Starting process rollout_proc6
[2025-07-29 11:06:42,499][08356] Starting process rollout_proc7
[2025-07-29 11:06:45,331][12434] Worker 5 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:06:45,410][12436] Worker 7 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:06:45,435][12430] Worker 2 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:06:45,475][12432] Worker 3 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:06:45,575][12431] Worker 1 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:06:45,590][12415] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:06:45,591][12415] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for learning process 0
[2025-07-29 11:06:45,606][12415] Num visible devices: 1
[2025-07-29 11:06:45,606][12415] Starting seed is not provided
[2025-07-29 11:06:45,607][12415] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:06:45,607][12415] Initializing actor-critic model on device cuda:0
[2025-07-29 11:06:45,607][12415] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:06:45,608][12415] RunningMeanStd input shape: (1,)
[2025-07-29 11:06:45,620][12415] ConvEncoder: input_channels=3
[2025-07-29 11:06:45,684][12435] Worker 6 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:06:45,704][12429] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:06:45,704][12429] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for inference process 0
[2025-07-29 11:06:45,709][12428] Worker 0 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:06:45,719][12429] Num visible devices: 1
[2025-07-29 11:06:45,727][12415] Conv encoder output size: 512
[2025-07-29 11:06:45,727][12415] Policy head output size: 512
[2025-07-29 11:06:45,736][12433] Worker 4 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:06:45,742][12415] Created Actor Critic model with architecture:
[2025-07-29 11:06:45,742][12415] ActorCriticSharedWeights(
  (obs_normalizer): ObservationNormalizer(
    (running_mean_std): RunningMeanStdDictInPlace(
      (running_mean_std): ModuleDict(
        (obs): RunningMeanStdInPlace()
      )
    )
  )
  (returns_normalizer): RecursiveScriptModule(original_name=RunningMeanStdInPlace)
  (encoder): VizdoomEncoder(
    (basic_encoder): ConvEncoder(
      (enc): RecursiveScriptModule(
        original_name=ConvEncoderImpl
        (conv_head): RecursiveScriptModule(
          original_name=Sequential
          (0): RecursiveScriptModule(original_name=Conv2d)
          (1): RecursiveScriptModule(original_name=ELU)
          (2): RecursiveScriptModule(original_name=Conv2d)
          (3): RecursiveScriptModule(original_name=ELU)
          (4): RecursiveScriptModule(original_name=Conv2d)
          (5): RecursiveScriptModule(original_name=ELU)
        )
        (mlp_layers): RecursiveScriptModule(
          original_name=Sequential
          (0): RecursiveScriptModule(original_name=Linear)
          (1): RecursiveScriptModule(original_name=ELU)
        )
      )
    )
  )
  (core): ModelCoreRNN(
    (core): GRU(512, 512)
  )
  (decoder): MlpDecoder(
    (mlp): Identity()
  )
  (critic_linear): Linear(in_features=512, out_features=1, bias=True)
  (action_parameterization): ActionParameterizationDefault(
    (distribution_linear): Linear(in_features=512, out_features=5, bias=True)
  )
)
[2025-07-29 11:06:45,869][12415] Using optimizer <class 'torch.optim.adam.Adam'>
[2025-07-29 11:06:46,791][12415] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:06:46,792][12415] Could not load from checkpoint, attempt 0
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:06:46,793][12415] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:06:46,794][12415] Could not load from checkpoint, attempt 1
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:06:46,794][12415] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:06:46,795][12415] Could not load from checkpoint, attempt 2
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:06:46,795][12415] Did not load from checkpoint, starting from scratch!
[2025-07-29 11:06:46,796][12415] Initialized policy 0 weights for model version 0
[2025-07-29 11:06:46,798][12415] LearnerWorker_p0 finished initialization!
[2025-07-29 11:06:46,798][12415] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:06:46,910][12429] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:06:46,911][12429] RunningMeanStd input shape: (1,)
[2025-07-29 11:06:46,923][12429] ConvEncoder: input_channels=3
[2025-07-29 11:06:47,026][12429] Conv encoder output size: 512
[2025-07-29 11:06:47,026][12429] Policy head output size: 512
[2025-07-29 11:06:47,059][08356] Inference worker 0-0 is ready!
[2025-07-29 11:06:47,060][08356] All inference workers are ready! Signal rollout workers to start!
[2025-07-29 11:06:47,092][12428] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:06:47,093][12434] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:06:47,094][12433] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:06:47,098][12430] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:06:47,111][12432] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:06:47,111][12431] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:06:47,112][12435] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:06:47,112][12436] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:06:47,367][12428] Decorrelating experience for 0 frames...
[2025-07-29 11:06:47,371][12433] Decorrelating experience for 0 frames...
[2025-07-29 11:06:47,375][12430] Decorrelating experience for 0 frames...
[2025-07-29 11:06:47,402][12435] Decorrelating experience for 0 frames...
[2025-07-29 11:06:47,404][12432] Decorrelating experience for 0 frames...
[2025-07-29 11:06:47,606][12428] Decorrelating experience for 32 frames...
[2025-07-29 11:06:47,610][12430] Decorrelating experience for 32 frames...
[2025-07-29 11:06:47,637][12435] Decorrelating experience for 32 frames...
[2025-07-29 11:06:47,851][12436] Decorrelating experience for 0 frames...
[2025-07-29 11:06:47,877][12434] Decorrelating experience for 0 frames...
[2025-07-29 11:06:47,908][12431] Decorrelating experience for 0 frames...
[2025-07-29 11:06:47,928][12430] Decorrelating experience for 64 frames...
[2025-07-29 11:06:47,963][12435] Decorrelating experience for 64 frames...
[2025-07-29 11:06:48,089][12436] Decorrelating experience for 32 frames...
[2025-07-29 11:06:48,115][12432] Decorrelating experience for 32 frames...
[2025-07-29 11:06:48,131][12434] Decorrelating experience for 32 frames...
[2025-07-29 11:06:48,156][12428] Decorrelating experience for 64 frames...
[2025-07-29 11:06:48,333][12435] Decorrelating experience for 96 frames...
[2025-07-29 11:06:48,368][12433] Decorrelating experience for 32 frames...
[2025-07-29 11:06:48,420][12431] Decorrelating experience for 32 frames...
[2025-07-29 11:06:48,468][12432] Decorrelating experience for 64 frames...
[2025-07-29 11:06:48,469][12428] Decorrelating experience for 96 frames...
[2025-07-29 11:06:48,496][12430] Decorrelating experience for 96 frames...
[2025-07-29 11:06:48,505][12434] Decorrelating experience for 64 frames...
[2025-07-29 11:06:48,646][12436] Decorrelating experience for 64 frames...
[2025-07-29 11:06:48,774][12432] Decorrelating experience for 96 frames...
[2025-07-29 11:06:48,785][12433] Decorrelating experience for 64 frames...
[2025-07-29 11:06:48,792][12431] Decorrelating experience for 64 frames...
[2025-07-29 11:06:48,906][12434] Decorrelating experience for 96 frames...
[2025-07-29 11:06:49,069][12436] Decorrelating experience for 96 frames...
[2025-07-29 11:06:49,102][12433] Decorrelating experience for 96 frames...
[2025-07-29 11:06:49,456][12431] Decorrelating experience for 96 frames...
[2025-07-29 11:06:49,650][12415] Signal inference workers to stop experience collection...
[2025-07-29 11:06:49,668][12429] InferenceWorker_p0-w0: stopping experience collection
[2025-07-29 11:06:50,065][08356] Fps is (10 sec: nan, 60 sec: nan, 300 sec: nan). Total num frames: 0. Throughput: 0: nan. Samples: 2588. Policy #0 lag: (min: -1.0, avg: -1.0, max: -1.0)
[2025-07-29 11:06:50,066][08356] Avg episode reward: [(0, '2.322')]
[2025-07-29 11:06:50,489][12415] Signal inference workers to resume experience collection...
[2025-07-29 11:06:50,490][12429] InferenceWorker_p0-w0: resuming experience collection
[2025-07-29 11:06:52,276][12429] Updated weights for policy 0, policy_version 10 (0.0090)
[2025-07-29 11:06:54,296][12429] Updated weights for policy 0, policy_version 20 (0.0011)
[2025-07-29 11:06:55,065][08356] Fps is (10 sec: 18841.2, 60 sec: 18841.2, 300 sec: 18841.2). Total num frames: 94208. Throughput: 0: 2463.5. Samples: 14906. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:06:55,066][08356] Avg episode reward: [(0, '4.433')]
[2025-07-29 11:06:56,284][12429] Updated weights for policy 0, policy_version 30 (0.0011)
[2025-07-29 11:06:58,307][12429] Updated weights for policy 0, policy_version 40 (0.0012)
[2025-07-29 11:07:00,065][08356] Fps is (10 sec: 19660.4, 60 sec: 19660.4, 300 sec: 19660.4). Total num frames: 196608. Throughput: 0: 4291.5. Samples: 45504. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:07:00,066][08356] Avg episode reward: [(0, '4.493')]
[2025-07-29 11:07:00,073][12415] Saving new best policy, reward=4.493!
[2025-07-29 11:07:00,320][12429] Updated weights for policy 0, policy_version 50 (0.0012)
[2025-07-29 11:07:02,364][12429] Updated weights for policy 0, policy_version 60 (0.0012)
[2025-07-29 11:07:02,394][08356] Heartbeat connected on Batcher_0
[2025-07-29 11:07:02,397][08356] Heartbeat connected on LearnerWorker_p0
[2025-07-29 11:07:02,406][08356] Heartbeat connected on InferenceWorker_p0-w0
[2025-07-29 11:07:02,409][08356] Heartbeat connected on RolloutWorker_w0
[2025-07-29 11:07:02,412][08356] Heartbeat connected on RolloutWorker_w1
[2025-07-29 11:07:02,415][08356] Heartbeat connected on RolloutWorker_w2
[2025-07-29 11:07:02,418][08356] Heartbeat connected on RolloutWorker_w3
[2025-07-29 11:07:02,424][08356] Heartbeat connected on RolloutWorker_w5
[2025-07-29 11:07:02,425][08356] Heartbeat connected on RolloutWorker_w4
[2025-07-29 11:07:02,427][08356] Heartbeat connected on RolloutWorker_w6
[2025-07-29 11:07:02,430][08356] Heartbeat connected on RolloutWorker_w7
[2025-07-29 11:07:04,450][12429] Updated weights for policy 0, policy_version 70 (0.0012)
[2025-07-29 11:07:05,065][08356] Fps is (10 sec: 20480.1, 60 sec: 19933.8, 300 sec: 19933.8). Total num frames: 299008. Throughput: 0: 3868.1. Samples: 60610. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:07:05,066][08356] Avg episode reward: [(0, '4.623')]
[2025-07-29 11:07:05,068][12415] Saving new best policy, reward=4.623!
[2025-07-29 11:07:06,435][12429] Updated weights for policy 0, policy_version 80 (0.0011)
[2025-07-29 11:07:08,483][12429] Updated weights for policy 0, policy_version 90 (0.0012)
[2025-07-29 11:07:10,065][08356] Fps is (10 sec: 20070.5, 60 sec: 19865.5, 300 sec: 19865.5). Total num frames: 397312. Throughput: 0: 4409.4. Samples: 90776. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:07:10,066][08356] Avg episode reward: [(0, '4.477')]
[2025-07-29 11:07:10,520][12429] Updated weights for policy 0, policy_version 100 (0.0011)
[2025-07-29 11:07:12,509][12429] Updated weights for policy 0, policy_version 110 (0.0012)
[2025-07-29 11:07:14,543][12429] Updated weights for policy 0, policy_version 120 (0.0012)
[2025-07-29 11:07:15,065][08356] Fps is (10 sec: 20070.1, 60 sec: 19988.3, 300 sec: 19988.3). Total num frames: 499712. Throughput: 0: 4745.3. Samples: 121222. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:07:15,066][08356] Avg episode reward: [(0, '4.683')]
[2025-07-29 11:07:15,067][12415] Saving new best policy, reward=4.683!
[2025-07-29 11:07:16,609][12429] Updated weights for policy 0, policy_version 130 (0.0012)
[2025-07-29 11:07:18,619][12429] Updated weights for policy 0, policy_version 140 (0.0012)
[2025-07-29 11:07:20,065][08356] Fps is (10 sec: 20480.2, 60 sec: 20070.4, 300 sec: 20070.4). Total num frames: 602112. Throughput: 0: 4959.8. Samples: 151382. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:07:20,066][08356] Avg episode reward: [(0, '4.708')]
[2025-07-29 11:07:20,071][12415] Saving new best policy, reward=4.708!
[2025-07-29 11:07:20,606][12429] Updated weights for policy 0, policy_version 150 (0.0011)
[2025-07-29 11:07:22,607][12429] Updated weights for policy 0, policy_version 160 (0.0011)
[2025-07-29 11:07:24,612][12429] Updated weights for policy 0, policy_version 170 (0.0012)
[2025-07-29 11:07:25,065][08356] Fps is (10 sec: 20480.1, 60 sec: 20128.8, 300 sec: 20128.8). Total num frames: 704512. Throughput: 0: 4688.2. Samples: 166674. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0)
[2025-07-29 11:07:25,066][08356] Avg episode reward: [(0, '4.567')]
[2025-07-29 11:07:26,621][12429] Updated weights for policy 0, policy_version 180 (0.0011)
[2025-07-29 11:07:28,746][12429] Updated weights for policy 0, policy_version 190 (0.0012)
[2025-07-29 11:07:30,065][08356] Fps is (10 sec: 20070.3, 60 sec: 20070.3, 300 sec: 20070.3). Total num frames: 802816. Throughput: 0: 4851.6. Samples: 196654. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:07:30,066][08356] Avg episode reward: [(0, '4.901')]
[2025-07-29 11:07:30,071][12415] Saving new best policy, reward=4.901!
[2025-07-29 11:07:30,841][12429] Updated weights for policy 0, policy_version 200 (0.0012)
[2025-07-29 11:07:32,875][12429] Updated weights for policy 0, policy_version 210 (0.0012)
[2025-07-29 11:07:34,897][12429] Updated weights for policy 0, policy_version 220 (0.0012)
[2025-07-29 11:07:35,065][08356] Fps is (10 sec: 19660.8, 60 sec: 20024.8, 300 sec: 20024.8). Total num frames: 901120. Throughput: 0: 4645.8. Samples: 211648. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:07:35,066][08356] Avg episode reward: [(0, '5.587')]
[2025-07-29 11:07:35,067][12415] Saving new best policy, reward=5.587!
[2025-07-29 11:07:36,933][12429] Updated weights for policy 0, policy_version 230 (0.0012)
[2025-07-29 11:07:38,927][12429] Updated weights for policy 0, policy_version 240 (0.0011)
[2025-07-29 11:07:40,065][08356] Fps is (10 sec: 20070.4, 60 sec: 20070.4, 300 sec: 20070.4). Total num frames: 1003520. Throughput: 0: 5047.9. Samples: 242060. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:07:40,066][08356] Avg episode reward: [(0, '6.193')]
[2025-07-29 11:07:40,072][12415] Saving new best policy, reward=6.193!
[2025-07-29 11:07:41,001][12429] Updated weights for policy 0, policy_version 250 (0.0012)
[2025-07-29 11:07:43,086][12429] Updated weights for policy 0, policy_version 260 (0.0012)
[2025-07-29 11:07:45,065][08356] Fps is (10 sec: 20070.3, 60 sec: 20033.1, 300 sec: 20033.1). Total num frames: 1101824. Throughput: 0: 5033.3. Samples: 272004. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:07:45,066][08356] Avg episode reward: [(0, '6.077')]
[2025-07-29 11:07:45,087][12429] Updated weights for policy 0, policy_version 270 (0.0012)
[2025-07-29 11:07:47,079][12429] Updated weights for policy 0, policy_version 280 (0.0012)
[2025-07-29 11:07:49,054][12429] Updated weights for policy 0, policy_version 290 (0.0011)
[2025-07-29 11:07:50,065][08356] Fps is (10 sec: 20480.1, 60 sec: 20138.6, 300 sec: 20138.6). Total num frames: 1208320. Throughput: 0: 5382.9. Samples: 302842. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:07:50,066][08356] Avg episode reward: [(0, '8.214')]
[2025-07-29 11:07:50,071][12415] Saving new best policy, reward=8.214!
[2025-07-29 11:07:51,050][12429] Updated weights for policy 0, policy_version 300 (0.0012)
[2025-07-29 11:07:53,055][12429] Updated weights for policy 0, policy_version 310 (0.0012)
[2025-07-29 11:07:55,065][08356] Fps is (10 sec: 20480.0, 60 sec: 20206.9, 300 sec: 20101.9). Total num frames: 1306624. Throughput: 0: 5051.3. Samples: 318084. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:07:55,066][08356] Avg episode reward: [(0, '10.397')]
[2025-07-29 11:07:55,068][12415] Saving new best policy, reward=10.397!
[2025-07-29 11:07:55,172][12429] Updated weights for policy 0, policy_version 320 (0.0012)
[2025-07-29 11:07:57,158][12429] Updated weights for policy 0, policy_version 330 (0.0012)
[2025-07-29 11:07:59,127][12429] Updated weights for policy 0, policy_version 340 (0.0012)
[2025-07-29 11:08:00,065][08356] Fps is (10 sec: 20070.3, 60 sec: 20207.0, 300 sec: 20128.9). Total num frames: 1409024. Throughput: 0: 5050.6. Samples: 348498. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:08:00,066][08356] Avg episode reward: [(0, '12.791')]
[2025-07-29 11:08:00,071][12415] Saving new best policy, reward=12.791!
[2025-07-29 11:08:01,103][12429] Updated weights for policy 0, policy_version 350 (0.0012)
[2025-07-29 11:08:03,116][12429] Updated weights for policy 0, policy_version 360 (0.0011)
[2025-07-29 11:08:05,065][08356] Fps is (10 sec: 20480.3, 60 sec: 20206.9, 300 sec: 20152.3). Total num frames: 1511424. Throughput: 0: 5065.2. Samples: 379314. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:08:05,066][08356] Avg episode reward: [(0, '11.732')]
[2025-07-29 11:08:05,109][12429] Updated weights for policy 0, policy_version 370 (0.0012)
[2025-07-29 11:08:07,201][12429] Updated weights for policy 0, policy_version 380 (0.0012)
[2025-07-29 11:08:09,215][12429] Updated weights for policy 0, policy_version 390 (0.0012)
[2025-07-29 11:08:10,065][08356] Fps is (10 sec: 20479.9, 60 sec: 20275.2, 300 sec: 20172.8). Total num frames: 1613824. Throughput: 0: 5052.9. Samples: 394056. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 11:08:10,066][08356] Avg episode reward: [(0, '12.746')]
[2025-07-29 11:08:11,225][12429] Updated weights for policy 0, policy_version 400 (0.0011)
[2025-07-29 11:08:13,209][12429] Updated weights for policy 0, policy_version 410 (0.0011)
[2025-07-29 11:08:15,065][08356] Fps is (10 sec: 20479.8, 60 sec: 20275.2, 300 sec: 20190.8). Total num frames: 1716224. Throughput: 0: 5072.8. Samples: 424928. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:08:15,066][08356] Avg episode reward: [(0, '13.065')]
[2025-07-29 11:08:15,068][12415] Saving new best policy, reward=13.065!
[2025-07-29 11:08:15,184][12429] Updated weights for policy 0, policy_version 420 (0.0012)
[2025-07-29 11:08:17,153][12429] Updated weights for policy 0, policy_version 430 (0.0012)
[2025-07-29 11:08:19,189][12429] Updated weights for policy 0, policy_version 440 (0.0012)
[2025-07-29 11:08:20,065][08356] Fps is (10 sec: 20480.1, 60 sec: 20275.2, 300 sec: 20206.9). Total num frames: 1818624. Throughput: 0: 5419.5. Samples: 455524. Policy #0 lag: (min: 0.0, avg: 0.5, max: 1.0)
[2025-07-29 11:08:20,066][08356] Avg episode reward: [(0, '13.716')]
[2025-07-29 11:08:20,070][12415] Saving new best policy, reward=13.716!
[2025-07-29 11:08:21,229][12429] Updated weights for policy 0, policy_version 450 (0.0012)
[2025-07-29 11:08:23,202][12429] Updated weights for policy 0, policy_version 460 (0.0012)
[2025-07-29 11:08:25,065][08356] Fps is (10 sec: 20480.0, 60 sec: 20275.2, 300 sec: 20221.3). Total num frames: 1921024. Throughput: 0: 5085.4. Samples: 470904. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 11:08:25,066][08356] Avg episode reward: [(0, '15.275')]
[2025-07-29 11:08:25,068][12415] Saving new best policy, reward=15.275!
[2025-07-29 11:08:25,208][12429] Updated weights for policy 0, policy_version 470 (0.0012)
[2025-07-29 11:08:27,178][12429] Updated weights for policy 0, policy_version 480 (0.0012)
[2025-07-29 11:08:29,171][12429] Updated weights for policy 0, policy_version 490 (0.0012)
[2025-07-29 11:08:30,065][08356] Fps is (10 sec: 20480.0, 60 sec: 20343.5, 300 sec: 20234.2). Total num frames: 2023424. Throughput: 0: 5107.9. Samples: 501860. Policy #0 lag: (min: 0.0, avg: 0.5, max: 1.0)
[2025-07-29 11:08:30,066][08356] Avg episode reward: [(0, '14.421')]
[2025-07-29 11:08:31,175][12429] Updated weights for policy 0, policy_version 500 (0.0012)
[2025-07-29 11:08:33,255][12429] Updated weights for policy 0, policy_version 510 (0.0011)
[2025-07-29 11:08:35,065][08356] Fps is (10 sec: 20480.2, 60 sec: 20411.8, 300 sec: 20245.9). Total num frames: 2125824. Throughput: 0: 5096.0. Samples: 532164. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 11:08:35,066][08356] Avg episode reward: [(0, '17.210')]
[2025-07-29 11:08:35,067][12415] Saving new best policy, reward=17.210!
[2025-07-29 11:08:35,241][12429] Updated weights for policy 0, policy_version 520 (0.0012)
[2025-07-29 11:08:37,227][12429] Updated weights for policy 0, policy_version 530 (0.0011)
[2025-07-29 11:08:39,211][12429] Updated weights for policy 0, policy_version 540 (0.0012)
[2025-07-29 11:08:40,065][08356] Fps is (10 sec: 20480.2, 60 sec: 20411.7, 300 sec: 20256.6). Total num frames: 2228224. Throughput: 0: 5100.9. Samples: 547624. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 11:08:40,066][08356] Avg episode reward: [(0, '20.572')]
[2025-07-29 11:08:40,071][12415] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000544_2228224.pth...
[2025-07-29 11:08:40,136][12415] Removing /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000512_2097152.pth
[2025-07-29 11:08:40,143][12415] Saving new best policy, reward=20.572!
[2025-07-29 11:08:41,192][12429] Updated weights for policy 0, policy_version 550 (0.0012)
[2025-07-29 11:08:43,170][12429] Updated weights for policy 0, policy_version 560 (0.0011)
[2025-07-29 11:08:45,065][08356] Fps is (10 sec: 20479.9, 60 sec: 20480.0, 300 sec: 20266.3). Total num frames: 2330624. Throughput: 0: 5110.4. Samples: 578466. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:08:45,066][08356] Avg episode reward: [(0, '17.952')]
[2025-07-29 11:08:45,261][12429] Updated weights for policy 0, policy_version 570 (0.0012)
[2025-07-29 11:08:47,293][12429] Updated weights for policy 0, policy_version 580 (0.0012)
[2025-07-29 11:08:49,269][12429] Updated weights for policy 0, policy_version 590 (0.0012)
[2025-07-29 11:08:50,065][08356] Fps is (10 sec: 20479.9, 60 sec: 20411.7, 300 sec: 20275.2). Total num frames: 2433024. Throughput: 0: 4759.1. Samples: 593472. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:08:50,066][08356] Avg episode reward: [(0, '19.162')]
[2025-07-29 11:08:51,260][12429] Updated weights for policy 0, policy_version 600 (0.0012)
[2025-07-29 11:08:53,247][12429] Updated weights for policy 0, policy_version 610 (0.0012)
[2025-07-29 11:08:55,065][08356] Fps is (10 sec: 20479.9, 60 sec: 20480.0, 300 sec: 20283.4). Total num frames: 2535424. Throughput: 0: 5117.5. Samples: 624342. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:08:55,066][08356] Avg episode reward: [(0, '19.000')]
[2025-07-29 11:08:55,238][12429] Updated weights for policy 0, policy_version 620 (0.0011)
[2025-07-29 11:08:57,290][12429] Updated weights for policy 0, policy_version 630 (0.0012)
[2025-07-29 11:08:59,355][12429] Updated weights for policy 0, policy_version 640 (0.0012)
[2025-07-29 11:09:00,065][08356] Fps is (10 sec: 20070.5, 60 sec: 20411.7, 300 sec: 20259.4). Total num frames: 2633728. Throughput: 0: 5105.3. Samples: 654666. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:09:00,066][08356] Avg episode reward: [(0, '23.100')]
[2025-07-29 11:09:00,071][12415] Saving new best policy, reward=23.100!
[2025-07-29 11:09:01,334][12429] Updated weights for policy 0, policy_version 650 (0.0011)
[2025-07-29 11:09:03,336][12429] Updated weights for policy 0, policy_version 660 (0.0012)
[2025-07-29 11:09:05,065][08356] Fps is (10 sec: 20070.4, 60 sec: 20411.7, 300 sec: 20267.6). Total num frames: 2736128. Throughput: 0: 4771.2. Samples: 670230. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:09:05,066][08356] Avg episode reward: [(0, '24.874')]
[2025-07-29 11:09:05,068][12415] Saving new best policy, reward=24.874!
[2025-07-29 11:09:05,303][12429] Updated weights for policy 0, policy_version 670 (0.0012)
[2025-07-29 11:09:07,288][12429] Updated weights for policy 0, policy_version 680 (0.0011)
[2025-07-29 11:09:09,281][12429] Updated weights for policy 0, policy_version 690 (0.0011)
[2025-07-29 11:09:10,065][08356] Fps is (10 sec: 20479.9, 60 sec: 20411.7, 300 sec: 20275.2). Total num frames: 2838528. Throughput: 0: 5118.0. Samples: 701214. Policy #0 lag: (min: 0.0, avg: 0.8, max: 1.0)
[2025-07-29 11:09:10,066][08356] Avg episode reward: [(0, '22.482')]
[2025-07-29 11:09:11,357][12429] Updated weights for policy 0, policy_version 700 (0.0012)
[2025-07-29 11:09:13,361][12429] Updated weights for policy 0, policy_version 710 (0.0012)
[2025-07-29 11:09:15,065][08356] Fps is (10 sec: 20480.0, 60 sec: 20411.7, 300 sec: 20282.2). Total num frames: 2940928. Throughput: 0: 5104.3. Samples: 731552. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:09:15,066][08356] Avg episode reward: [(0, '20.169')]
[2025-07-29 11:09:15,338][12429] Updated weights for policy 0, policy_version 720 (0.0011)
[2025-07-29 11:09:17,326][12429] Updated weights for policy 0, policy_version 730 (0.0012)
[2025-07-29 11:09:19,329][12429] Updated weights for policy 0, policy_version 740 (0.0012)
[2025-07-29 11:09:20,065][08356] Fps is (10 sec: 20480.0, 60 sec: 20411.7, 300 sec: 20288.8). Total num frames: 3043328. Throughput: 0: 4776.2. Samples: 747092. Policy #0 lag: (min: 0.0, avg: 0.9, max: 2.0)
[2025-07-29 11:09:20,066][08356] Avg episode reward: [(0, '22.839')]
[2025-07-29 11:09:21,300][12429] Updated weights for policy 0, policy_version 750 (0.0012)
[2025-07-29 11:09:23,348][12429] Updated weights for policy 0, policy_version 760 (0.0012)
[2025-07-29 11:09:25,065][08356] Fps is (10 sec: 20480.0, 60 sec: 20411.7, 300 sec: 20295.0). Total num frames: 3145728. Throughput: 0: 5113.9. Samples: 777750. Policy #0 lag: (min: 0.0, avg: 0.9, max: 2.0)
[2025-07-29 11:09:25,066][08356] Avg episode reward: [(0, '22.733')]
[2025-07-29 11:09:25,379][12429] Updated weights for policy 0, policy_version 770 (0.0011)
[2025-07-29 11:09:27,352][12429] Updated weights for policy 0, policy_version 780 (0.0011)
[2025-07-29 11:09:29,328][12429] Updated weights for policy 0, policy_version 790 (0.0012)
[2025-07-29 11:09:30,065][08356] Fps is (10 sec: 20480.1, 60 sec: 20411.8, 300 sec: 20300.8). Total num frames: 3248128. Throughput: 0: 5112.9. Samples: 808546. Policy #0 lag: (min: 0.0, avg: 0.9, max: 2.0)
[2025-07-29 11:09:30,066][08356] Avg episode reward: [(0, '19.832')]
[2025-07-29 11:09:31,310][12429] Updated weights for policy 0, policy_version 800 (0.0011)
[2025-07-29 11:09:33,286][12429] Updated weights for policy 0, policy_version 810 (0.0012)
[2025-07-29 11:09:35,065][08356] Fps is (10 sec: 20480.1, 60 sec: 20411.7, 300 sec: 20306.2). Total num frames: 3350528. Throughput: 0: 5466.2. Samples: 839450. Policy #0 lag: (min: 0.0, avg: 0.9, max: 2.0)
[2025-07-29 11:09:35,066][08356] Avg episode reward: [(0, '22.250')]
[2025-07-29 11:09:35,286][12429] Updated weights for policy 0, policy_version 820 (0.0011)
[2025-07-29 11:09:37,361][12429] Updated weights for policy 0, policy_version 830 (0.0012)
[2025-07-29 11:09:39,338][12429] Updated weights for policy 0, policy_version 840 (0.0011)
[2025-07-29 11:09:40,065][08356] Fps is (10 sec: 20479.9, 60 sec: 20411.7, 300 sec: 20311.3). Total num frames: 3452928. Throughput: 0: 5112.9. Samples: 854422. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:09:40,066][08356] Avg episode reward: [(0, '24.782')]
[2025-07-29 11:09:41,308][12429] Updated weights for policy 0, policy_version 850 (0.0011)
[2025-07-29 11:09:43,284][12429] Updated weights for policy 0, policy_version 860 (0.0012)
[2025-07-29 11:09:45,065][08356] Fps is (10 sec: 20889.6, 60 sec: 20480.0, 300 sec: 20339.6). Total num frames: 3559424. Throughput: 0: 5129.9. Samples: 885510. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:09:45,066][08356] Avg episode reward: [(0, '23.621')]
[2025-07-29 11:09:45,276][12429] Updated weights for policy 0, policy_version 870 (0.0011)
[2025-07-29 11:09:47,410][12429] Updated weights for policy 0, policy_version 880 (0.0012)
[2025-07-29 11:09:49,514][12429] Updated weights for policy 0, policy_version 890 (0.0012)
[2025-07-29 11:09:50,065][08356] Fps is (10 sec: 20070.5, 60 sec: 20343.5, 300 sec: 20297.9). Total num frames: 3653632. Throughput: 0: 5433.1. Samples: 914720. Policy #0 lag: (min: 0.0, avg: 0.4, max: 1.0)
[2025-07-29 11:09:50,066][08356] Avg episode reward: [(0, '22.036')]
[2025-07-29 11:09:51,501][12429] Updated weights for policy 0, policy_version 900 (0.0012)
[2025-07-29 11:09:53,482][12429] Updated weights for policy 0, policy_version 910 (0.0012)
[2025-07-29 11:09:55,065][08356] Fps is (10 sec: 20070.4, 60 sec: 20411.7, 300 sec: 20325.0). Total num frames: 3760128. Throughput: 0: 5090.2. Samples: 930274. Policy #0 lag: (min: 0.0, avg: 0.5, max: 1.0)
[2025-07-29 11:09:55,066][08356] Avg episode reward: [(0, '23.609')]
[2025-07-29 11:09:55,455][12429] Updated weights for policy 0, policy_version 920 (0.0012)
[2025-07-29 11:09:57,431][12429] Updated weights for policy 0, policy_version 930 (0.0011)
[2025-07-29 11:09:59,413][12429] Updated weights for policy 0, policy_version 940 (0.0012)
[2025-07-29 11:10:00,065][08356] Fps is (10 sec: 20889.5, 60 sec: 20480.0, 300 sec: 20329.1). Total num frames: 3862528. Throughput: 0: 5105.7. Samples: 961310. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 11:10:00,066][08356] Avg episode reward: [(0, '26.307')]
[2025-07-29 11:10:00,072][12415] Saving new best policy, reward=26.307!
[2025-07-29 11:10:01,481][12429] Updated weights for policy 0, policy_version 950 (0.0012)
[2025-07-29 11:10:03,535][12429] Updated weights for policy 0, policy_version 960 (0.0012)
[2025-07-29 11:10:05,065][08356] Fps is (10 sec: 20070.1, 60 sec: 20411.7, 300 sec: 20311.9). Total num frames: 3960832. Throughput: 0: 5432.7. Samples: 991562. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0)
[2025-07-29 11:10:05,066][08356] Avg episode reward: [(0, '24.505')]
[2025-07-29 11:10:05,539][12429] Updated weights for policy 0, policy_version 970 (0.0011)
[2025-07-29 11:10:07,138][12415] Stopping Batcher_0...
[2025-07-29 11:10:07,138][12415] Loop batcher_evt_loop terminating...
[2025-07-29 11:10:07,138][08356] Component Batcher_0 stopped!
[2025-07-29 11:10:07,139][12415] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:10:07,161][12429] Weights refcount: 2 0
[2025-07-29 11:10:07,162][12429] Stopping InferenceWorker_p0-w0...
[2025-07-29 11:10:07,163][12429] Loop inference_proc0-0_evt_loop terminating...
[2025-07-29 11:10:07,163][08356] Component InferenceWorker_p0-w0 stopped!
[2025-07-29 11:10:07,174][12434] Stopping RolloutWorker_w5...
[2025-07-29 11:10:07,175][12434] Loop rollout_proc5_evt_loop terminating...
[2025-07-29 11:10:07,175][12432] Stopping RolloutWorker_w3...
[2025-07-29 11:10:07,174][08356] Component RolloutWorker_w5 stopped!
[2025-07-29 11:10:07,175][12432] Loop rollout_proc3_evt_loop terminating...
[2025-07-29 11:10:07,176][08356] Component RolloutWorker_w3 stopped!
[2025-07-29 11:10:07,177][12430] Stopping RolloutWorker_w2...
[2025-07-29 11:10:07,177][12433] Stopping RolloutWorker_w4...
[2025-07-29 11:10:07,178][12430] Loop rollout_proc2_evt_loop terminating...
[2025-07-29 11:10:07,178][12433] Loop rollout_proc4_evt_loop terminating...
[2025-07-29 11:10:07,177][08356] Component RolloutWorker_w2 stopped!
[2025-07-29 11:10:07,178][08356] Component RolloutWorker_w4 stopped!
[2025-07-29 11:10:07,179][12431] Stopping RolloutWorker_w1...
[2025-07-29 11:10:07,180][08356] Component RolloutWorker_w1 stopped!
[2025-07-29 11:10:07,180][12436] Stopping RolloutWorker_w7...
[2025-07-29 11:10:07,180][12431] Loop rollout_proc1_evt_loop terminating...
[2025-07-29 11:10:07,181][12436] Loop rollout_proc7_evt_loop terminating...
[2025-07-29 11:10:07,181][08356] Component RolloutWorker_w7 stopped!
[2025-07-29 11:10:07,181][12428] Stopping RolloutWorker_w0...
[2025-07-29 11:10:07,182][12435] Stopping RolloutWorker_w6...
[2025-07-29 11:10:07,182][12428] Loop rollout_proc0_evt_loop terminating...
[2025-07-29 11:10:07,181][08356] Component RolloutWorker_w0 stopped!
[2025-07-29 11:10:07,183][12435] Loop rollout_proc6_evt_loop terminating...
[2025-07-29 11:10:07,183][08356] Component RolloutWorker_w6 stopped!
[2025-07-29 11:10:07,230][12415] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:10:07,304][12415] Stopping LearnerWorker_p0...
[2025-07-29 11:10:07,304][12415] Loop learner_proc0_evt_loop terminating...
[2025-07-29 11:10:07,304][08356] Component LearnerWorker_p0 stopped!
[2025-07-29 11:10:07,306][08356] Waiting for process learner_proc0 to stop...
[2025-07-29 11:10:08,149][08356] Waiting for process inference_proc0-0 to join...
[2025-07-29 11:10:08,150][08356] Waiting for process rollout_proc0 to join...
[2025-07-29 11:10:08,151][08356] Waiting for process rollout_proc1 to join...
[2025-07-29 11:10:08,152][08356] Waiting for process rollout_proc2 to join...
[2025-07-29 11:10:08,153][08356] Waiting for process rollout_proc3 to join...
[2025-07-29 11:10:08,153][08356] Waiting for process rollout_proc4 to join...
[2025-07-29 11:10:08,154][08356] Waiting for process rollout_proc5 to join...
[2025-07-29 11:10:08,155][08356] Waiting for process rollout_proc6 to join...
[2025-07-29 11:10:08,155][08356] Waiting for process rollout_proc7 to join...
[2025-07-29 11:10:08,156][08356] Batcher 0 profile tree view:
batching: 16.6304, releasing_batches: 0.0214
[2025-07-29 11:10:08,157][08356] InferenceWorker_p0-w0 profile tree view:
wait_policy: 0.0000
  wait_policy_total: 3.6828
update_model: 3.2133
  weight_update: 0.0011
one_step: 0.0029
  handle_policy_step: 184.6307
    deserialize: 7.6461, stack: 1.2934, obs_to_device_normalize: 45.3181, forward: 88.3579, send_messages: 12.7485
    prepare_outputs: 22.0962
      to_cpu: 14.2240
[2025-07-29 11:10:08,157][08356] Learner 0 profile tree view:
misc: 0.0036, prepare_batch: 6.5193
train: 18.4938
  epoch_init: 0.0041, minibatch_init: 0.0051, losses_postprocess: 0.3442, kl_divergence: 0.3916, after_optimizer: 1.9269
  calculate_losses: 8.3696
    losses_init: 0.0035, forward_head: 0.6279, bptt_initial: 4.4161, tail: 0.6188, advantages_returns: 0.1577, losses: 1.1792
    bptt: 1.2235
      bptt_forward_core: 1.1725
  update: 7.1419
    clip: 0.7660
[2025-07-29 11:10:08,158][08356] RolloutWorker_w0 profile tree view:
wait_for_trajectories: 0.1227, enqueue_policy_requests: 8.8482, env_step: 126.6760, overhead: 5.4741, complete_rollouts: 0.2066
save_policy_outputs: 7.9668
  split_output_tensors: 3.0468
[2025-07-29 11:10:08,159][08356] RolloutWorker_w7 profile tree view:
wait_for_trajectories: 0.1191, enqueue_policy_requests: 8.8728, env_step: 126.4309, overhead: 5.5495, complete_rollouts: 0.2074
save_policy_outputs: 8.0131
  split_output_tensors: 3.0342
[2025-07-29 11:10:08,160][08356] Loop Runner_EvtLoop terminating...
[2025-07-29 11:10:08,161][08356] Runner profile tree view:
main_loop: 205.7305
[2025-07-29 11:10:08,161][08356] Collected {0: 4005888}, FPS: 19471.5
[2025-07-29 11:10:58,192][08356] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 11:10:58,193][08356] Overriding arg 'num_workers' with value 1 passed from command line
[2025-07-29 11:10:58,194][08356] Adding new argument 'no_render'=True that is not in the saved config file!
[2025-07-29 11:10:58,195][08356] Adding new argument 'save_video'=True that is not in the saved config file!
[2025-07-29 11:10:58,195][08356] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:10:58,196][08356] Adding new argument 'video_name'=None that is not in the saved config file!
[2025-07-29 11:10:58,197][08356] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:10:58,197][08356] Adding new argument 'max_num_episodes'=10 that is not in the saved config file!
[2025-07-29 11:10:58,198][08356] Adding new argument 'push_to_hub'=False that is not in the saved config file!
[2025-07-29 11:10:58,199][08356] Adding new argument 'hf_repository'=None that is not in the saved config file!
[2025-07-29 11:10:58,199][08356] Adding new argument 'policy_index'=0 that is not in the saved config file!
[2025-07-29 11:10:58,200][08356] Adding new argument 'eval_deterministic'=False that is not in the saved config file!
[2025-07-29 11:10:58,200][08356] Adding new argument 'train_script'=None that is not in the saved config file!
[2025-07-29 11:10:58,201][08356] Adding new argument 'enjoy_script'=None that is not in the saved config file!
[2025-07-29 11:10:58,202][08356] Using frameskip 1 and render_action_repeat=4 for evaluation
[2025-07-29 11:10:58,230][08356] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:10:58,231][08356] RunningMeanStd input shape: (1,)
[2025-07-29 11:10:58,241][08356] ConvEncoder: input_channels=3
[2025-07-29 11:10:58,277][08356] Conv encoder output size: 512
[2025-07-29 11:10:58,277][08356] Policy head output size: 512
[2025-07-29 11:10:58,298][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:10:58,299][08356] Could not load from checkpoint, attempt 0
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:10:58,300][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:10:58,301][08356] Could not load from checkpoint, attempt 1
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:10:58,302][08356] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:10:58,303][08356] Could not load from checkpoint, attempt 2
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:16:06,923][14877] Saving configuration to /content/train_dir/default_experiment/config.json...
[2025-07-29 11:16:06,925][14877] Rollout worker 0 uses device cpu
[2025-07-29 11:16:06,926][14877] Rollout worker 1 uses device cpu
[2025-07-29 11:16:06,926][14877] Rollout worker 2 uses device cpu
[2025-07-29 11:16:06,927][14877] Rollout worker 3 uses device cpu
[2025-07-29 11:16:06,928][14877] Rollout worker 4 uses device cpu
[2025-07-29 11:16:06,928][14877] Rollout worker 5 uses device cpu
[2025-07-29 11:16:06,929][14877] Rollout worker 6 uses device cpu
[2025-07-29 11:16:06,929][14877] Rollout worker 7 uses device cpu
[2025-07-29 11:16:06,967][14877] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:16:06,968][14877] InferenceWorker_p0-w0: min num requests: 2
[2025-07-29 11:16:06,997][14877] Starting all processes...
[2025-07-29 11:16:06,998][14877] Starting process learner_proc0
[2025-07-29 11:16:07,051][14877] Starting all processes...
[2025-07-29 11:16:07,056][14877] Starting process inference_proc0-0
[2025-07-29 11:16:07,056][14877] Starting process rollout_proc0
[2025-07-29 11:16:07,056][14877] Starting process rollout_proc1
[2025-07-29 11:16:07,057][14877] Starting process rollout_proc2
[2025-07-29 11:16:07,057][14877] Starting process rollout_proc3
[2025-07-29 11:16:07,057][14877] Starting process rollout_proc4
[2025-07-29 11:16:07,058][14877] Starting process rollout_proc5
[2025-07-29 11:16:07,058][14877] Starting process rollout_proc6
[2025-07-29 11:16:07,058][14877] Starting process rollout_proc7
[2025-07-29 11:16:09,944][15381] Worker 6 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:16:10,092][15382] Worker 5 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:16:10,243][15380] Worker 1 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:16:10,256][15362] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:16:10,256][15362] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for learning process 0
[2025-07-29 11:16:10,274][15362] Num visible devices: 1
[2025-07-29 11:16:10,275][15362] Starting seed is not provided
[2025-07-29 11:16:10,276][15362] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:16:10,276][15362] Initializing actor-critic model on device cuda:0
[2025-07-29 11:16:10,276][15362] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:16:10,278][15362] RunningMeanStd input shape: (1,)
[2025-07-29 11:16:10,293][15375] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:16:10,293][15375] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for inference process 0
[2025-07-29 11:16:10,297][15362] ConvEncoder: input_channels=3
[2025-07-29 11:16:10,300][15378] Worker 2 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:16:10,309][15375] Num visible devices: 1
[2025-07-29 11:16:10,369][15377] Worker 3 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:16:10,413][15376] Worker 0 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:16:10,422][15362] Conv encoder output size: 512
[2025-07-29 11:16:10,422][15362] Policy head output size: 512
[2025-07-29 11:16:10,437][15362] Created Actor Critic model with architecture:
[2025-07-29 11:16:10,437][15362] ActorCriticSharedWeights(
  (obs_normalizer): ObservationNormalizer(
    (running_mean_std): RunningMeanStdDictInPlace(
      (running_mean_std): ModuleDict(
        (obs): RunningMeanStdInPlace()
      )
    )
  )
  (returns_normalizer): RecursiveScriptModule(original_name=RunningMeanStdInPlace)
  (encoder): VizdoomEncoder(
    (basic_encoder): ConvEncoder(
      (enc): RecursiveScriptModule(
        original_name=ConvEncoderImpl
        (conv_head): RecursiveScriptModule(
          original_name=Sequential
          (0): RecursiveScriptModule(original_name=Conv2d)
          (1): RecursiveScriptModule(original_name=ELU)
          (2): RecursiveScriptModule(original_name=Conv2d)
          (3): RecursiveScriptModule(original_name=ELU)
          (4): RecursiveScriptModule(original_name=Conv2d)
          (5): RecursiveScriptModule(original_name=ELU)
        )
        (mlp_layers): RecursiveScriptModule(
          original_name=Sequential
          (0): RecursiveScriptModule(original_name=Linear)
          (1): RecursiveScriptModule(original_name=ELU)
        )
      )
    )
  )
  (core): ModelCoreRNN(
    (core): GRU(512, 512)
  )
  (decoder): MlpDecoder(
    (mlp): Identity()
  )
  (critic_linear): Linear(in_features=512, out_features=1, bias=True)
  (action_parameterization): ActionParameterizationDefault(
    (distribution_linear): Linear(in_features=512, out_features=5, bias=True)
  )
)
[2025-07-29 11:16:10,465][15379] Worker 4 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:16:10,506][15383] Worker 7 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:16:10,589][15362] Using optimizer <class 'torch.optim.adam.Adam'>
[2025-07-29 11:16:11,512][15362] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:16:11,513][15362] Could not load from checkpoint, attempt 0
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:16:11,514][15362] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:16:11,515][15362] Could not load from checkpoint, attempt 1
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:16:11,515][15362] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:16:11,516][15362] Could not load from checkpoint, attempt 2
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/dist-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint
    checkpoint_dict = torch.load(latest_checkpoint, map_location=device)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/torch/serialization.py", line 1470, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
[2025-07-29 11:16:11,516][15362] Did not load from checkpoint, starting from scratch!
[2025-07-29 11:16:11,517][15362] Initialized policy 0 weights for model version 0
[2025-07-29 11:16:11,519][15362] LearnerWorker_p0 finished initialization!
[2025-07-29 11:16:11,519][15362] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:16:11,626][15375] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:16:11,627][15375] RunningMeanStd input shape: (1,)
[2025-07-29 11:16:11,639][15375] ConvEncoder: input_channels=3
[2025-07-29 11:16:11,743][15375] Conv encoder output size: 512
[2025-07-29 11:16:11,744][15375] Policy head output size: 512
[2025-07-29 11:16:11,776][14877] Inference worker 0-0 is ready!
[2025-07-29 11:16:11,778][14877] All inference workers are ready! Signal rollout workers to start!
[2025-07-29 11:16:11,810][15383] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:16:11,810][15379] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:16:11,811][15376] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:16:11,811][15380] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:16:11,831][15377] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:16:11,831][15382] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:16:11,831][15381] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:16:11,831][15378] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:16:12,102][15376] Decorrelating experience for 0 frames...
[2025-07-29 11:16:12,106][15380] Decorrelating experience for 0 frames...
[2025-07-29 11:16:12,107][15379] Decorrelating experience for 0 frames...
[2025-07-29 11:16:12,114][15383] Decorrelating experience for 0 frames...
[2025-07-29 11:16:12,130][15378] Decorrelating experience for 0 frames...
[2025-07-29 11:16:12,133][15382] Decorrelating experience for 0 frames...
[2025-07-29 11:16:12,360][15376] Decorrelating experience for 32 frames...
[2025-07-29 11:16:12,362][15380] Decorrelating experience for 32 frames...
[2025-07-29 11:16:12,364][15379] Decorrelating experience for 32 frames...
[2025-07-29 11:16:12,369][15383] Decorrelating experience for 32 frames...
[2025-07-29 11:16:12,384][15377] Decorrelating experience for 0 frames...
[2025-07-29 11:16:12,388][15378] Decorrelating experience for 32 frames...
[2025-07-29 11:16:12,581][14877] Fps is (10 sec: nan, 60 sec: nan, 300 sec: nan). Total num frames: 0. Throughput: 0: nan. Samples: 0. Policy #0 lag: (min: -1.0, avg: -1.0, max: -1.0)
[2025-07-29 11:16:12,627][15381] Decorrelating experience for 0 frames...
[2025-07-29 11:16:12,643][15377] Decorrelating experience for 32 frames...
[2025-07-29 11:16:12,684][15376] Decorrelating experience for 64 frames...
[2025-07-29 11:16:12,727][15383] Decorrelating experience for 64 frames...
[2025-07-29 11:16:12,730][15379] Decorrelating experience for 64 frames...
[2025-07-29 11:16:12,875][15381] Decorrelating experience for 32 frames...
[2025-07-29 11:16:12,924][15380] Decorrelating experience for 64 frames...
[2025-07-29 11:16:12,934][15382] Decorrelating experience for 32 frames...
[2025-07-29 11:16:13,023][15379] Decorrelating experience for 96 frames...
[2025-07-29 11:16:13,057][15383] Decorrelating experience for 96 frames...
[2025-07-29 11:16:13,192][15378] Decorrelating experience for 64 frames...
[2025-07-29 11:16:13,201][15381] Decorrelating experience for 64 frames...
[2025-07-29 11:16:13,221][15380] Decorrelating experience for 96 frames...
[2025-07-29 11:16:13,265][15377] Decorrelating experience for 64 frames...
[2025-07-29 11:16:13,436][15376] Decorrelating experience for 96 frames...
[2025-07-29 11:16:13,476][15378] Decorrelating experience for 96 frames...
[2025-07-29 11:16:13,540][15381] Decorrelating experience for 96 frames...
[2025-07-29 11:16:13,704][15377] Decorrelating experience for 96 frames...
[2025-07-29 11:16:14,036][15382] Decorrelating experience for 64 frames...
[2025-07-29 11:16:14,318][15362] Signal inference workers to stop experience collection...
[2025-07-29 11:16:14,323][15375] InferenceWorker_p0-w0: stopping experience collection
[2025-07-29 11:16:14,395][15382] Decorrelating experience for 96 frames...
[2025-07-29 11:16:15,154][15362] Signal inference workers to resume experience collection...
[2025-07-29 11:16:15,155][15375] InferenceWorker_p0-w0: resuming experience collection
[2025-07-29 11:16:16,922][15375] Updated weights for policy 0, policy_version 10 (0.0088)
[2025-07-29 11:16:17,581][14877] Fps is (10 sec: 10649.7, 60 sec: 10649.7, 300 sec: 10649.7). Total num frames: 53248. Throughput: 0: 543.6. Samples: 2718. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:16:17,582][14877] Avg episode reward: [(0, '4.451')]
[2025-07-29 11:16:18,956][15375] Updated weights for policy 0, policy_version 20 (0.0011)
[2025-07-29 11:16:21,073][15375] Updated weights for policy 0, policy_version 30 (0.0012)
[2025-07-29 11:16:22,581][14877] Fps is (10 sec: 15155.2, 60 sec: 15155.2, 300 sec: 15155.2). Total num frames: 151552. Throughput: 0: 3010.4. Samples: 30104. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:16:22,582][14877] Avg episode reward: [(0, '4.403')]
[2025-07-29 11:16:22,587][15362] Saving new best policy, reward=4.403!
[2025-07-29 11:16:23,196][15375] Updated weights for policy 0, policy_version 40 (0.0012)
[2025-07-29 11:16:25,189][15375] Updated weights for policy 0, policy_version 50 (0.0011)
[2025-07-29 11:16:26,959][14877] Heartbeat connected on Batcher_0
[2025-07-29 11:16:26,972][14877] Heartbeat connected on InferenceWorker_p0-w0
[2025-07-29 11:16:26,973][14877] Heartbeat connected on LearnerWorker_p0
[2025-07-29 11:16:26,975][14877] Heartbeat connected on RolloutWorker_w0
[2025-07-29 11:16:26,980][14877] Heartbeat connected on RolloutWorker_w1
[2025-07-29 11:16:26,982][14877] Heartbeat connected on RolloutWorker_w2
[2025-07-29 11:16:26,988][14877] Heartbeat connected on RolloutWorker_w4
[2025-07-29 11:16:26,989][14877] Heartbeat connected on RolloutWorker_w3
[2025-07-29 11:16:26,991][14877] Heartbeat connected on RolloutWorker_w5
[2025-07-29 11:16:26,994][14877] Heartbeat connected on RolloutWorker_w6
[2025-07-29 11:16:26,997][14877] Heartbeat connected on RolloutWorker_w7
[2025-07-29 11:16:27,180][15375] Updated weights for policy 0, policy_version 60 (0.0012)
[2025-07-29 11:16:27,581][14877] Fps is (10 sec: 20070.3, 60 sec: 16930.2, 300 sec: 16930.2). Total num frames: 253952. Throughput: 0: 4012.5. Samples: 60188. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:16:27,582][14877] Avg episode reward: [(0, '4.371')]
[2025-07-29 11:16:29,166][15375] Updated weights for policy 0, policy_version 70 (0.0011)
[2025-07-29 11:16:31,174][15375] Updated weights for policy 0, policy_version 80 (0.0012)
[2025-07-29 11:16:32,581][14877] Fps is (10 sec: 20480.0, 60 sec: 17817.6, 300 sec: 17817.6). Total num frames: 356352. Throughput: 0: 3779.3. Samples: 75586. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:16:32,582][14877] Avg episode reward: [(0, '4.305')]
[2025-07-29 11:16:33,197][15375] Updated weights for policy 0, policy_version 90 (0.0011)
[2025-07-29 11:16:35,302][15375] Updated weights for policy 0, policy_version 100 (0.0012)
[2025-07-29 11:16:37,318][15375] Updated weights for policy 0, policy_version 110 (0.0012)
[2025-07-29 11:16:37,581][14877] Fps is (10 sec: 20070.5, 60 sec: 18186.3, 300 sec: 18186.3). Total num frames: 454656. Throughput: 0: 4224.8. Samples: 105620. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:16:37,582][14877] Avg episode reward: [(0, '4.606')]
[2025-07-29 11:16:37,584][15362] Saving new best policy, reward=4.606!
[2025-07-29 11:16:39,326][15375] Updated weights for policy 0, policy_version 120 (0.0012)
[2025-07-29 11:16:41,320][15375] Updated weights for policy 0, policy_version 130 (0.0011)
[2025-07-29 11:16:42,581][14877] Fps is (10 sec: 20070.5, 60 sec: 18568.6, 300 sec: 18568.6). Total num frames: 557056. Throughput: 0: 4540.5. Samples: 136216. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:16:42,582][14877] Avg episode reward: [(0, '4.683')]
[2025-07-29 11:16:42,587][15362] Saving new best policy, reward=4.683!
[2025-07-29 11:16:43,334][15375] Updated weights for policy 0, policy_version 140 (0.0011)
[2025-07-29 11:16:45,354][15375] Updated weights for policy 0, policy_version 150 (0.0012)
[2025-07-29 11:16:47,446][15375] Updated weights for policy 0, policy_version 160 (0.0012)
[2025-07-29 11:16:47,581][14877] Fps is (10 sec: 20070.3, 60 sec: 18724.6, 300 sec: 18724.6). Total num frames: 655360. Throughput: 0: 4331.3. Samples: 151594. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:16:47,582][14877] Avg episode reward: [(0, '4.935')]
[2025-07-29 11:16:47,583][15362] Saving new best policy, reward=4.935!
[2025-07-29 11:16:49,452][15375] Updated weights for policy 0, policy_version 170 (0.0012)
[2025-07-29 11:16:51,459][15375] Updated weights for policy 0, policy_version 180 (0.0012)
[2025-07-29 11:16:52,581][14877] Fps is (10 sec: 20070.3, 60 sec: 18944.0, 300 sec: 18944.0). Total num frames: 757760. Throughput: 0: 4540.3. Samples: 181610. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:16:52,582][14877] Avg episode reward: [(0, '5.278')]
[2025-07-29 11:16:52,588][15362] Saving new best policy, reward=5.278!
[2025-07-29 11:16:53,476][15375] Updated weights for policy 0, policy_version 190 (0.0012)
[2025-07-29 11:16:55,471][15375] Updated weights for policy 0, policy_version 200 (0.0012)
[2025-07-29 11:16:57,448][15375] Updated weights for policy 0, policy_version 210 (0.0012)
[2025-07-29 11:16:57,581][14877] Fps is (10 sec: 20480.1, 60 sec: 19114.7, 300 sec: 19114.7). Total num frames: 860160. Throughput: 0: 4718.5. Samples: 212330. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:16:57,582][14877] Avg episode reward: [(0, '5.428')]
[2025-07-29 11:16:57,584][15362] Saving new best policy, reward=5.428!
[2025-07-29 11:16:59,542][15375] Updated weights for policy 0, policy_version 220 (0.0012)
[2025-07-29 11:17:01,636][15375] Updated weights for policy 0, policy_version 230 (0.0012)
[2025-07-29 11:17:02,581][14877] Fps is (10 sec: 20070.5, 60 sec: 19169.3, 300 sec: 19169.3). Total num frames: 958464. Throughput: 0: 4988.3. Samples: 227192. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:17:02,582][14877] Avg episode reward: [(0, '5.942')]
[2025-07-29 11:17:02,586][15362] Saving new best policy, reward=5.942!
[2025-07-29 11:17:03,662][15375] Updated weights for policy 0, policy_version 240 (0.0012)
[2025-07-29 11:17:05,786][15375] Updated weights for policy 0, policy_version 250 (0.0012)
[2025-07-29 11:17:07,581][14877] Fps is (10 sec: 19660.8, 60 sec: 19214.0, 300 sec: 19214.0). Total num frames: 1056768. Throughput: 0: 5035.9. Samples: 256720. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:17:07,582][14877] Avg episode reward: [(0, '6.500')]
[2025-07-29 11:17:07,584][15362] Saving new best policy, reward=6.500!
[2025-07-29 11:17:07,890][15375] Updated weights for policy 0, policy_version 260 (0.0011)
[2025-07-29 11:17:09,897][15375] Updated weights for policy 0, policy_version 270 (0.0012)
[2025-07-29 11:17:11,987][15375] Updated weights for policy 0, policy_version 280 (0.0012)
[2025-07-29 11:17:12,581][14877] Fps is (10 sec: 19660.8, 60 sec: 19251.2, 300 sec: 19251.2). Total num frames: 1155072. Throughput: 0: 5031.8. Samples: 286618. Policy #0 lag: (min: 0.0, avg: 0.9, max: 2.0)
[2025-07-29 11:17:12,582][14877] Avg episode reward: [(0, '7.833')]
[2025-07-29 11:17:12,587][15362] Saving new best policy, reward=7.833!
[2025-07-29 11:17:14,063][15375] Updated weights for policy 0, policy_version 290 (0.0012)
[2025-07-29 11:17:16,126][15375] Updated weights for policy 0, policy_version 300 (0.0011)
[2025-07-29 11:17:17,581][14877] Fps is (10 sec: 20070.4, 60 sec: 20070.4, 300 sec: 19345.7). Total num frames: 1257472. Throughput: 0: 5019.6. Samples: 301468. Policy #0 lag: (min: 0.0, avg: 0.9, max: 2.0)
[2025-07-29 11:17:17,582][14877] Avg episode reward: [(0, '8.027')]
[2025-07-29 11:17:17,584][15362] Saving new best policy, reward=8.027!
[2025-07-29 11:17:18,176][15375] Updated weights for policy 0, policy_version 310 (0.0012)
[2025-07-29 11:17:20,217][15375] Updated weights for policy 0, policy_version 320 (0.0012)
[2025-07-29 11:17:22,213][15375] Updated weights for policy 0, policy_version 330 (0.0012)
[2025-07-29 11:17:22,581][14877] Fps is (10 sec: 20070.2, 60 sec: 20070.4, 300 sec: 19368.2). Total num frames: 1355776. Throughput: 0: 5018.5. Samples: 331452. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:17:22,582][14877] Avg episode reward: [(0, '11.054')]
[2025-07-29 11:17:22,587][15362] Saving new best policy, reward=11.054!
[2025-07-29 11:17:24,265][15375] Updated weights for policy 0, policy_version 340 (0.0012)
[2025-07-29 11:17:26,310][15375] Updated weights for policy 0, policy_version 350 (0.0011)
[2025-07-29 11:17:27,581][14877] Fps is (10 sec: 20070.4, 60 sec: 20070.4, 300 sec: 19442.4). Total num frames: 1458176. Throughput: 0: 5007.7. Samples: 361564. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:17:27,582][14877] Avg episode reward: [(0, '11.577')]
[2025-07-29 11:17:27,584][15362] Saving new best policy, reward=11.577!
[2025-07-29 11:17:28,313][15375] Updated weights for policy 0, policy_version 360 (0.0011)
[2025-07-29 11:17:30,340][15375] Updated weights for policy 0, policy_version 370 (0.0012)
[2025-07-29 11:17:32,340][15375] Updated weights for policy 0, policy_version 380 (0.0011)
[2025-07-29 11:17:32,581][14877] Fps is (10 sec: 20480.2, 60 sec: 20070.4, 300 sec: 19507.2). Total num frames: 1560576. Throughput: 0: 5007.5. Samples: 376932. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:17:32,583][14877] Avg episode reward: [(0, '13.717')]
[2025-07-29 11:17:32,589][15362] Saving new best policy, reward=13.717!
[2025-07-29 11:17:34,324][15375] Updated weights for policy 0, policy_version 390 (0.0012)
[2025-07-29 11:17:36,323][15375] Updated weights for policy 0, policy_version 400 (0.0012)
[2025-07-29 11:17:37,581][14877] Fps is (10 sec: 20480.0, 60 sec: 20138.7, 300 sec: 19564.4). Total num frames: 1662976. Throughput: 0: 5024.2. Samples: 407700. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:17:37,583][14877] Avg episode reward: [(0, '17.373')]
[2025-07-29 11:17:37,584][15362] Saving new best policy, reward=17.373!
[2025-07-29 11:17:38,398][15375] Updated weights for policy 0, policy_version 410 (0.0012)
[2025-07-29 11:17:40,396][15375] Updated weights for policy 0, policy_version 420 (0.0011)
[2025-07-29 11:17:42,384][15375] Updated weights for policy 0, policy_version 430 (0.0012)
[2025-07-29 11:17:42,581][14877] Fps is (10 sec: 20480.0, 60 sec: 20138.7, 300 sec: 19615.3). Total num frames: 1765376. Throughput: 0: 5015.0. Samples: 438006. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:17:42,582][14877] Avg episode reward: [(0, '15.654')]
[2025-07-29 11:17:44,384][15375] Updated weights for policy 0, policy_version 440 (0.0011)
[2025-07-29 11:17:46,366][15375] Updated weights for policy 0, policy_version 450 (0.0012)
[2025-07-29 11:17:47,581][14877] Fps is (10 sec: 20070.5, 60 sec: 20138.7, 300 sec: 19617.7). Total num frames: 1863680. Throughput: 0: 5028.2. Samples: 453462. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:17:47,582][14877] Avg episode reward: [(0, '14.627')]
[2025-07-29 11:17:48,399][15375] Updated weights for policy 0, policy_version 460 (0.0012)
[2025-07-29 11:17:50,552][15375] Updated weights for policy 0, policy_version 470 (0.0013)
[2025-07-29 11:17:52,581][14877] Fps is (10 sec: 19660.8, 60 sec: 20070.4, 300 sec: 19619.8). Total num frames: 1961984. Throughput: 0: 5033.5. Samples: 483226. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:17:52,582][14877] Avg episode reward: [(0, '17.532')]
[2025-07-29 11:17:52,603][15362] Saving new best policy, reward=17.532!
[2025-07-29 11:17:52,606][15375] Updated weights for policy 0, policy_version 480 (0.0012)
[2025-07-29 11:17:54,649][15375] Updated weights for policy 0, policy_version 490 (0.0012)
[2025-07-29 11:17:56,672][15375] Updated weights for policy 0, policy_version 500 (0.0011)
[2025-07-29 11:17:57,581][14877] Fps is (10 sec: 20070.3, 60 sec: 20070.4, 300 sec: 19660.8). Total num frames: 2064384. Throughput: 0: 5037.6. Samples: 513312. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:17:57,582][14877] Avg episode reward: [(0, '16.624')]
[2025-07-29 11:17:58,682][15375] Updated weights for policy 0, policy_version 510 (0.0012)
[2025-07-29 11:18:00,683][15375] Updated weights for policy 0, policy_version 520 (0.0012)
[2025-07-29 11:18:02,581][14877] Fps is (10 sec: 20479.9, 60 sec: 20138.7, 300 sec: 19698.0). Total num frames: 2166784. Throughput: 0: 5048.4. Samples: 528644. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:18:02,582][14877] Avg episode reward: [(0, '16.520')]
[2025-07-29 11:18:02,588][15362] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000529_2166784.pth...
[2025-07-29 11:18:02,653][15362] Removing /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000529_2166784.pth
[2025-07-29 11:18:02,782][15375] Updated weights for policy 0, policy_version 530 (0.0012)
[2025-07-29 11:18:04,841][15375] Updated weights for policy 0, policy_version 540 (0.0012)
[2025-07-29 11:18:06,830][15375] Updated weights for policy 0, policy_version 550 (0.0011)
[2025-07-29 11:18:07,581][14877] Fps is (10 sec: 20070.4, 60 sec: 20138.7, 300 sec: 19696.4). Total num frames: 2265088. Throughput: 0: 5048.1. Samples: 558618. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:18:07,582][14877] Avg episode reward: [(0, '18.642')]
[2025-07-29 11:18:07,583][15362] Saving new best policy, reward=18.642!
[2025-07-29 11:18:08,855][15375] Updated weights for policy 0, policy_version 560 (0.0012)
[2025-07-29 11:18:10,853][15375] Updated weights for policy 0, policy_version 570 (0.0012)
[2025-07-29 11:18:12,581][14877] Fps is (10 sec: 20070.5, 60 sec: 20206.9, 300 sec: 19729.1). Total num frames: 2367488. Throughput: 0: 5061.8. Samples: 589346. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:18:12,582][14877] Avg episode reward: [(0, '20.052')]
[2025-07-29 11:18:12,586][15362] Saving new best policy, reward=20.052!
[2025-07-29 11:18:12,849][15375] Updated weights for policy 0, policy_version 580 (0.0012)
[2025-07-29 11:18:14,846][15375] Updated weights for policy 0, policy_version 590 (0.0012)
[2025-07-29 11:18:17,085][15375] Updated weights for policy 0, policy_version 600 (0.0012)
[2025-07-29 11:18:17,581][14877] Fps is (10 sec: 20070.3, 60 sec: 20138.7, 300 sec: 19726.3). Total num frames: 2465792. Throughput: 0: 5059.0. Samples: 604586. Policy #0 lag: (min: 0.0, avg: 0.7, max: 1.0)
[2025-07-29 11:18:17,582][14877] Avg episode reward: [(0, '20.239')]
[2025-07-29 11:18:17,584][15362] Saving new best policy, reward=20.239!
[2025-07-29 11:18:19,094][15375] Updated weights for policy 0, policy_version 610 (0.0011)
[2025-07-29 11:18:21,091][15375] Updated weights for policy 0, policy_version 620 (0.0012)
[2025-07-29 11:18:22,581][14877] Fps is (10 sec: 20070.4, 60 sec: 20207.0, 300 sec: 19755.3). Total num frames: 2568192. Throughput: 0: 5032.0. Samples: 634140. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:18:22,582][14877] Avg episode reward: [(0, '19.108')]
[2025-07-29 11:18:23,088][15375] Updated weights for policy 0, policy_version 630 (0.0011)
[2025-07-29 11:18:25,114][15375] Updated weights for policy 0, policy_version 640 (0.0011)
[2025-07-29 11:18:27,122][15375] Updated weights for policy 0, policy_version 650 (0.0012)
[2025-07-29 11:18:27,581][14877] Fps is (10 sec: 20480.1, 60 sec: 20206.9, 300 sec: 19782.2). Total num frames: 2670592. Throughput: 0: 5037.0. Samples: 664670. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:18:27,582][14877] Avg episode reward: [(0, '19.611')]
[2025-07-29 11:18:29,241][15375] Updated weights for policy 0, policy_version 660 (0.0013)
[2025-07-29 11:18:31,258][15375] Updated weights for policy 0, policy_version 670 (0.0012)
[2025-07-29 11:18:32,581][14877] Fps is (10 sec: 20070.4, 60 sec: 20138.7, 300 sec: 19777.8). Total num frames: 2768896. Throughput: 0: 5017.4. Samples: 679246. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0)
[2025-07-29 11:18:32,582][14877] Avg episode reward: [(0, '18.892')]
[2025-07-29 11:18:33,246][15375] Updated weights for policy 0, policy_version 680 (0.0012)
[2025-07-29 11:18:35,234][15375] Updated weights for policy 0, policy_version 690 (0.0011)
[2025-07-29 11:18:37,236][15375] Updated weights for policy 0, policy_version 700 (0.0012)
[2025-07-29 11:18:37,581][14877] Fps is (10 sec: 20070.4, 60 sec: 20138.7, 300 sec: 19802.0). Total num frames: 2871296. Throughput: 0: 5040.0. Samples: 710026. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:18:37,582][14877] Avg episode reward: [(0, '20.674')]
[2025-07-29 11:18:37,583][15362] Saving new best policy, reward=20.674!
[2025-07-29 11:18:39,242][15375] Updated weights for policy 0, policy_version 710 (0.0012)
[2025-07-29 11:18:41,296][15375] Updated weights for policy 0, policy_version 720 (0.0012)
[2025-07-29 11:18:42,581][14877] Fps is (10 sec: 20479.9, 60 sec: 20138.6, 300 sec: 19824.6). Total num frames: 2973696. Throughput: 0: 5044.2. Samples: 740302. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:18:42,582][14877] Avg episode reward: [(0, '21.549')]
[2025-07-29 11:18:42,588][15362] Saving new best policy, reward=21.549!
[2025-07-29 11:18:43,352][15375] Updated weights for policy 0, policy_version 730 (0.0012)
[2025-07-29 11:18:45,329][15375] Updated weights for policy 0, policy_version 740 (0.0012)
[2025-07-29 11:18:47,327][15375] Updated weights for policy 0, policy_version 750 (0.0012)
[2025-07-29 11:18:47,581][14877] Fps is (10 sec: 20480.1, 60 sec: 20206.9, 300 sec: 19845.8). Total num frames: 3076096. Throughput: 0: 5044.3. Samples: 755636. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:18:47,582][14877] Avg episode reward: [(0, '22.523')]
[2025-07-29 11:18:47,583][15362] Saving new best policy, reward=22.523!
[2025-07-29 11:18:49,323][15375] Updated weights for policy 0, policy_version 760 (0.0012)
[2025-07-29 11:18:51,323][15375] Updated weights for policy 0, policy_version 770 (0.0012)
[2025-07-29 11:18:52,581][14877] Fps is (10 sec: 20480.2, 60 sec: 20275.2, 300 sec: 19865.6). Total num frames: 3178496. Throughput: 0: 5062.2. Samples: 786416. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:18:52,582][14877] Avg episode reward: [(0, '19.452')]
[2025-07-29 11:18:53,362][15375] Updated weights for policy 0, policy_version 780 (0.0012)
[2025-07-29 11:18:55,433][15375] Updated weights for policy 0, policy_version 790 (0.0011)
[2025-07-29 11:18:57,435][15375] Updated weights for policy 0, policy_version 800 (0.0011)
[2025-07-29 11:18:57,581][14877] Fps is (10 sec: 20070.5, 60 sec: 20206.9, 300 sec: 19859.4). Total num frames: 3276800. Throughput: 0: 5050.4. Samples: 816612. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:18:57,582][14877] Avg episode reward: [(0, '21.263')]
[2025-07-29 11:18:59,448][15375] Updated weights for policy 0, policy_version 810 (0.0012)
[2025-07-29 11:19:01,447][15375] Updated weights for policy 0, policy_version 820 (0.0012)
[2025-07-29 11:19:02,581][14877] Fps is (10 sec: 20070.2, 60 sec: 20206.9, 300 sec: 19877.6). Total num frames: 3379200. Throughput: 0: 5051.4. Samples: 831900. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:19:02,582][14877] Avg episode reward: [(0, '19.952')]
[2025-07-29 11:19:03,463][15375] Updated weights for policy 0, policy_version 830 (0.0012)
[2025-07-29 11:19:05,460][15375] Updated weights for policy 0, policy_version 840 (0.0012)
[2025-07-29 11:19:07,514][15375] Updated weights for policy 0, policy_version 850 (0.0011)
[2025-07-29 11:19:07,581][14877] Fps is (10 sec: 20480.0, 60 sec: 20275.2, 300 sec: 19894.9). Total num frames: 3481600. Throughput: 0: 5076.5. Samples: 862584. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:19:07,582][14877] Avg episode reward: [(0, '23.121')]
[2025-07-29 11:19:07,583][15362] Saving new best policy, reward=23.121!
[2025-07-29 11:19:09,550][15375] Updated weights for policy 0, policy_version 860 (0.0012)
[2025-07-29 11:19:11,531][15375] Updated weights for policy 0, policy_version 870 (0.0012)
[2025-07-29 11:19:12,581][14877] Fps is (10 sec: 20480.0, 60 sec: 20275.2, 300 sec: 19911.1). Total num frames: 3584000. Throughput: 0: 5072.0. Samples: 892912. Policy #0 lag: (min: 0.0, avg: 0.7, max: 1.0)
[2025-07-29 11:19:12,582][14877] Avg episode reward: [(0, '20.525')]
[2025-07-29 11:19:13,524][15375] Updated weights for policy 0, policy_version 880 (0.0011)
[2025-07-29 11:19:15,507][15375] Updated weights for policy 0, policy_version 890 (0.0012)
[2025-07-29 11:19:17,486][15375] Updated weights for policy 0, policy_version 900 (0.0012)
[2025-07-29 11:19:17,581][14877] Fps is (10 sec: 20479.9, 60 sec: 20343.5, 300 sec: 19926.5). Total num frames: 3686400. Throughput: 0: 5092.1. Samples: 908392. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0)
[2025-07-29 11:19:17,582][14877] Avg episode reward: [(0, '23.864')]
[2025-07-29 11:19:17,583][15362] Saving new best policy, reward=23.864!
[2025-07-29 11:19:19,558][15375] Updated weights for policy 0, policy_version 910 (0.0012)
[2025-07-29 11:19:21,620][15375] Updated weights for policy 0, policy_version 920 (0.0011)
[2025-07-29 11:19:22,581][14877] Fps is (10 sec: 20070.5, 60 sec: 20275.2, 300 sec: 19919.5). Total num frames: 3784704. Throughput: 0: 5080.4. Samples: 938642. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:19:22,582][14877] Avg episode reward: [(0, '22.878')]
[2025-07-29 11:19:23,628][15375] Updated weights for policy 0, policy_version 930 (0.0012)
[2025-07-29 11:19:25,630][15375] Updated weights for policy 0, policy_version 940 (0.0012)
[2025-07-29 11:19:27,581][14877] Fps is (10 sec: 20070.3, 60 sec: 20275.2, 300 sec: 19933.9). Total num frames: 3887104. Throughput: 0: 5087.1. Samples: 969222. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:19:27,582][14877] Avg episode reward: [(0, '22.381')]
[2025-07-29 11:19:27,655][15375] Updated weights for policy 0, policy_version 950 (0.0011)
[2025-07-29 11:19:29,638][15375] Updated weights for policy 0, policy_version 960 (0.0012)
[2025-07-29 11:19:31,652][15375] Updated weights for policy 0, policy_version 970 (0.0011)
[2025-07-29 11:19:32,581][14877] Fps is (10 sec: 20479.9, 60 sec: 20343.4, 300 sec: 19947.5). Total num frames: 3989504. Throughput: 0: 5089.3. Samples: 984654. Policy #0 lag: (min: 0.0, avg: 0.8, max: 2.0)
[2025-07-29 11:19:32,582][14877] Avg episode reward: [(0, '21.898')]
[2025-07-29 11:19:33,283][15362] Stopping Batcher_0...
[2025-07-29 11:19:33,283][14877] Component Batcher_0 stopped!
[2025-07-29 11:19:33,283][15362] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:19:33,284][15362] Loop batcher_evt_loop terminating...
[2025-07-29 11:19:33,309][15375] Weights refcount: 2 0
[2025-07-29 11:19:33,311][15375] Stopping InferenceWorker_p0-w0...
[2025-07-29 11:19:33,311][15375] Loop inference_proc0-0_evt_loop terminating...
[2025-07-29 11:19:33,311][14877] Component InferenceWorker_p0-w0 stopped!
[2025-07-29 11:19:33,322][15382] Stopping RolloutWorker_w5...
[2025-07-29 11:19:33,323][14877] Component RolloutWorker_w5 stopped!
[2025-07-29 11:19:33,323][15382] Loop rollout_proc5_evt_loop terminating...
[2025-07-29 11:19:33,325][15379] Stopping RolloutWorker_w4...
[2025-07-29 11:19:33,325][15379] Loop rollout_proc4_evt_loop terminating...
[2025-07-29 11:19:33,326][15383] Stopping RolloutWorker_w7...
[2025-07-29 11:19:33,325][14877] Component RolloutWorker_w4 stopped!
[2025-07-29 11:19:33,326][15378] Stopping RolloutWorker_w2...
[2025-07-29 11:19:33,326][15383] Loop rollout_proc7_evt_loop terminating...
[2025-07-29 11:19:33,327][15378] Loop rollout_proc2_evt_loop terminating...
[2025-07-29 11:19:33,327][14877] Component RolloutWorker_w7 stopped!
[2025-07-29 11:19:33,327][15381] Stopping RolloutWorker_w6...
[2025-07-29 11:19:33,328][14877] Component RolloutWorker_w2 stopped!
[2025-07-29 11:19:33,328][15377] Stopping RolloutWorker_w3...
[2025-07-29 11:19:33,329][15377] Loop rollout_proc3_evt_loop terminating...
[2025-07-29 11:19:33,328][14877] Component RolloutWorker_w6 stopped!
[2025-07-29 11:19:33,329][15381] Loop rollout_proc6_evt_loop terminating...
[2025-07-29 11:19:33,330][15376] Stopping RolloutWorker_w0...
[2025-07-29 11:19:33,329][14877] Component RolloutWorker_w3 stopped!
[2025-07-29 11:19:33,330][15376] Loop rollout_proc0_evt_loop terminating...
[2025-07-29 11:19:33,330][14877] Component RolloutWorker_w0 stopped!
[2025-07-29 11:19:33,331][15380] Stopping RolloutWorker_w1...
[2025-07-29 11:19:33,332][15380] Loop rollout_proc1_evt_loop terminating...
[2025-07-29 11:19:33,331][14877] Component RolloutWorker_w1 stopped!
[2025-07-29 11:19:33,387][15362] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:19:33,497][15362] Stopping LearnerWorker_p0...
[2025-07-29 11:19:33,498][15362] Loop learner_proc0_evt_loop terminating...
[2025-07-29 11:19:33,497][14877] Component LearnerWorker_p0 stopped!
[2025-07-29 11:19:33,499][14877] Waiting for process learner_proc0 to stop...
[2025-07-29 11:19:34,390][14877] Waiting for process inference_proc0-0 to join...
[2025-07-29 11:19:34,391][14877] Waiting for process rollout_proc0 to join...
[2025-07-29 11:19:34,392][14877] Waiting for process rollout_proc1 to join...
[2025-07-29 11:19:34,393][14877] Waiting for process rollout_proc2 to join...
[2025-07-29 11:19:34,393][14877] Waiting for process rollout_proc3 to join...
[2025-07-29 11:19:34,394][14877] Waiting for process rollout_proc4 to join...
[2025-07-29 11:19:34,395][14877] Waiting for process rollout_proc5 to join...
[2025-07-29 11:19:34,395][14877] Waiting for process rollout_proc6 to join...
[2025-07-29 11:19:34,396][14877] Waiting for process rollout_proc7 to join...
[2025-07-29 11:19:34,397][14877] Batcher 0 profile tree view:
batching: 15.9394, releasing_batches: 0.0226
[2025-07-29 11:19:34,397][14877] InferenceWorker_p0-w0 profile tree view:
wait_policy: 0.0000
  wait_policy_total: 3.6368
update_model: 3.2563
  weight_update: 0.0012
one_step: 0.0028
  handle_policy_step: 185.7811
    deserialize: 7.7116, stack: 1.3444, obs_to_device_normalize: 45.7883, forward: 88.8487, send_messages: 12.8386
    prepare_outputs: 22.0814
      to_cpu: 14.1540
[2025-07-29 11:19:34,398][14877] Learner 0 profile tree view:
misc: 0.0036, prepare_batch: 6.5224
train: 18.3334
  epoch_init: 0.0043, minibatch_init: 0.0052, losses_postprocess: 0.3326, kl_divergence: 0.3777, after_optimizer: 1.8839
  calculate_losses: 8.2978
    losses_init: 0.0031, forward_head: 0.6269, bptt_initial: 4.3473, tail: 0.6268, advantages_returns: 0.1559, losses: 1.1816
    bptt: 1.2128
      bptt_forward_core: 1.1619
  update: 7.1203
    clip: 0.7673
[2025-07-29 11:19:34,399][14877] RolloutWorker_w0 profile tree view:
wait_for_trajectories: 0.1222, enqueue_policy_requests: 8.9219, env_step: 127.3378, overhead: 5.4915, complete_rollouts: 0.2142
save_policy_outputs: 7.9989
  split_output_tensors: 3.0649
[2025-07-29 11:19:34,400][14877] RolloutWorker_w7 profile tree view:
wait_for_trajectories: 0.1216, enqueue_policy_requests: 8.8727, env_step: 127.3302, overhead: 5.5457, complete_rollouts: 0.2093
save_policy_outputs: 8.0453
  split_output_tensors: 3.0660
[2025-07-29 11:19:34,400][14877] Loop Runner_EvtLoop terminating...
[2025-07-29 11:19:34,401][14877] Runner profile tree view:
main_loop: 207.4044
[2025-07-29 11:19:34,402][14877] Collected {0: 4005888}, FPS: 19314.4
[2025-07-29 11:25:37,292][17818] Saving configuration to /content/train_dir/default_experiment/config.json...
[2025-07-29 11:25:37,294][17818] Rollout worker 0 uses device cpu
[2025-07-29 11:25:37,294][17818] Rollout worker 1 uses device cpu
[2025-07-29 11:25:37,295][17818] Rollout worker 2 uses device cpu
[2025-07-29 11:25:37,296][17818] Rollout worker 3 uses device cpu
[2025-07-29 11:25:37,296][17818] Rollout worker 4 uses device cpu
[2025-07-29 11:25:37,297][17818] Rollout worker 5 uses device cpu
[2025-07-29 11:25:37,298][17818] Rollout worker 6 uses device cpu
[2025-07-29 11:25:37,299][17818] Rollout worker 7 uses device cpu
[2025-07-29 11:25:37,341][17818] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:25:37,342][17818] InferenceWorker_p0-w0: min num requests: 2
[2025-07-29 11:25:37,371][17818] Starting all processes...
[2025-07-29 11:25:37,372][17818] Starting process learner_proc0
[2025-07-29 11:25:37,424][17818] Starting all processes...
[2025-07-29 11:25:37,428][17818] Starting process inference_proc0-0
[2025-07-29 11:25:37,428][17818] Starting process rollout_proc0
[2025-07-29 11:25:37,429][17818] Starting process rollout_proc1
[2025-07-29 11:25:37,429][17818] Starting process rollout_proc2
[2025-07-29 11:25:37,430][17818] Starting process rollout_proc3
[2025-07-29 11:25:37,430][17818] Starting process rollout_proc4
[2025-07-29 11:25:37,430][17818] Starting process rollout_proc5
[2025-07-29 11:25:37,433][17818] Starting process rollout_proc6
[2025-07-29 11:25:37,434][17818] Starting process rollout_proc7
[2025-07-29 11:25:39,421][18484] Worker 3 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:25:39,498][18477] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:25:39,499][18477] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for inference process 0
[2025-07-29 11:25:39,517][18477] Num visible devices: 1
[2025-07-29 11:25:39,535][18478] Worker 0 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:25:39,578][18480] Worker 2 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:25:39,650][18464] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:25:39,651][18464] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for learning process 0
[2025-07-29 11:25:39,667][18483] Worker 7 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:25:39,668][18464] Num visible devices: 1
[2025-07-29 11:25:39,671][18485] Worker 4 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:25:39,708][18481] Worker 6 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:25:39,710][18479] Worker 1 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:25:39,713][18464] Starting seed is not provided
[2025-07-29 11:25:39,713][18464] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:25:39,713][18464] Initializing actor-critic model on device cuda:0
[2025-07-29 11:25:39,713][18464] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:25:39,714][18464] RunningMeanStd input shape: (1,)
[2025-07-29 11:25:39,728][18464] ConvEncoder: input_channels=3
[2025-07-29 11:25:39,758][18482] Worker 5 uses CPU cores [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
[2025-07-29 11:25:39,858][18464] Conv encoder output size: 512
[2025-07-29 11:25:39,859][18464] Policy head output size: 512
[2025-07-29 11:25:39,874][18464] Created Actor Critic model with architecture:
[2025-07-29 11:25:39,874][18464] ActorCriticSharedWeights(
  (obs_normalizer): ObservationNormalizer(
    (running_mean_std): RunningMeanStdDictInPlace(
      (running_mean_std): ModuleDict(
        (obs): RunningMeanStdInPlace()
      )
    )
  )
  (returns_normalizer): RecursiveScriptModule(original_name=RunningMeanStdInPlace)
  (encoder): VizdoomEncoder(
    (basic_encoder): ConvEncoder(
      (enc): RecursiveScriptModule(
        original_name=ConvEncoderImpl
        (conv_head): RecursiveScriptModule(
          original_name=Sequential
          (0): RecursiveScriptModule(original_name=Conv2d)
          (1): RecursiveScriptModule(original_name=ELU)
          (2): RecursiveScriptModule(original_name=Conv2d)
          (3): RecursiveScriptModule(original_name=ELU)
          (4): RecursiveScriptModule(original_name=Conv2d)
          (5): RecursiveScriptModule(original_name=ELU)
        )
        (mlp_layers): RecursiveScriptModule(
          original_name=Sequential
          (0): RecursiveScriptModule(original_name=Linear)
          (1): RecursiveScriptModule(original_name=ELU)
        )
      )
    )
  )
  (core): ModelCoreRNN(
    (core): GRU(512, 512)
  )
  (decoder): MlpDecoder(
    (mlp): Identity()
  )
  (critic_linear): Linear(in_features=512, out_features=1, bias=True)
  (action_parameterization): ActionParameterizationDefault(
    (distribution_linear): Linear(in_features=512, out_features=5, bias=True)
  )
)
[2025-07-29 11:25:41,670][18464] Using optimizer <class 'torch.optim.adam.Adam'>
[2025-07-29 11:25:41,670][18464] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth...
[2025-07-29 11:25:41,702][18464] Loading model from checkpoint
[2025-07-29 11:25:41,706][18464] Loaded experiment state at self.train_step=978, self.env_steps=4005888
[2025-07-29 11:25:41,707][18464] Initialized policy 0 weights for model version 978
[2025-07-29 11:25:41,709][18464] LearnerWorker_p0 finished initialization!
[2025-07-29 11:25:41,709][18464] Using GPUs [0] for process 0 (actually maps to GPUs [0])
[2025-07-29 11:25:41,806][18477] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:25:41,807][18477] RunningMeanStd input shape: (1,)
[2025-07-29 11:25:41,818][18477] ConvEncoder: input_channels=3
[2025-07-29 11:25:41,924][18477] Conv encoder output size: 512
[2025-07-29 11:25:41,925][18477] Policy head output size: 512
[2025-07-29 11:25:43,599][17818] Inference worker 0-0 is ready!
[2025-07-29 11:25:43,600][17818] All inference workers are ready! Signal rollout workers to start!
[2025-07-29 11:25:43,613][18480] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:25:43,613][18479] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:25:43,618][18484] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:25:43,618][18478] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:25:43,618][18483] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:25:43,618][18481] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:25:43,619][18482] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:25:43,619][18485] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:25:43,877][18479] Decorrelating experience for 0 frames...
[2025-07-29 11:25:43,896][18485] Decorrelating experience for 0 frames...
[2025-07-29 11:25:43,896][18484] Decorrelating experience for 0 frames...
[2025-07-29 11:25:43,897][18483] Decorrelating experience for 0 frames...
[2025-07-29 11:25:43,897][18478] Decorrelating experience for 0 frames...
[2025-07-29 11:25:44,120][18480] Decorrelating experience for 0 frames...
[2025-07-29 11:25:44,142][18484] Decorrelating experience for 32 frames...
[2025-07-29 11:25:44,142][18478] Decorrelating experience for 32 frames...
[2025-07-29 11:25:44,142][18483] Decorrelating experience for 32 frames...
[2025-07-29 11:25:44,164][18479] Decorrelating experience for 32 frames...
[2025-07-29 11:25:44,391][18481] Decorrelating experience for 0 frames...
[2025-07-29 11:25:44,435][18483] Decorrelating experience for 64 frames...
[2025-07-29 11:25:44,435][18482] Decorrelating experience for 0 frames...
[2025-07-29 11:25:44,439][18484] Decorrelating experience for 64 frames...
[2025-07-29 11:25:44,451][18480] Decorrelating experience for 32 frames...
[2025-07-29 11:25:44,473][18478] Decorrelating experience for 64 frames...
[2025-07-29 11:25:44,548][17818] Fps is (10 sec: nan, 60 sec: nan, 300 sec: nan). Total num frames: 4005888. Throughput: 0: nan. Samples: 0. Policy #0 lag: (min: -1.0, avg: -1.0, max: -1.0)
[2025-07-29 11:25:44,646][18485] Decorrelating experience for 32 frames...
[2025-07-29 11:25:44,687][18482] Decorrelating experience for 32 frames...
[2025-07-29 11:25:44,738][18481] Decorrelating experience for 32 frames...
[2025-07-29 11:25:44,744][18484] Decorrelating experience for 96 frames...
[2025-07-29 11:25:44,746][18483] Decorrelating experience for 96 frames...
[2025-07-29 11:25:44,931][18478] Decorrelating experience for 96 frames...
[2025-07-29 11:25:44,948][18485] Decorrelating experience for 64 frames...
[2025-07-29 11:25:44,997][18480] Decorrelating experience for 64 frames...
[2025-07-29 11:25:45,003][18482] Decorrelating experience for 64 frames...
[2025-07-29 11:25:45,007][18479] Decorrelating experience for 64 frames...
[2025-07-29 11:25:45,289][18482] Decorrelating experience for 96 frames...
[2025-07-29 11:25:45,290][18480] Decorrelating experience for 96 frames...
[2025-07-29 11:25:45,297][18485] Decorrelating experience for 96 frames...
[2025-07-29 11:25:45,304][18479] Decorrelating experience for 96 frames...
[2025-07-29 11:25:45,605][18481] Decorrelating experience for 64 frames...
[2025-07-29 11:25:45,983][18464] Signal inference workers to stop experience collection...
[2025-07-29 11:25:45,993][18477] InferenceWorker_p0-w0: stopping experience collection
[2025-07-29 11:25:46,007][18481] Decorrelating experience for 96 frames...
[2025-07-29 11:25:47,020][18464] Signal inference workers to resume experience collection...
[2025-07-29 11:25:47,021][18477] InferenceWorker_p0-w0: resuming experience collection
[2025-07-29 11:25:47,022][18464] Stopping Batcher_0...
[2025-07-29 11:25:47,022][18464] Loop batcher_evt_loop terminating...
[2025-07-29 11:25:47,027][17818] Component Batcher_0 stopped!
[2025-07-29 11:25:47,034][18477] Weights refcount: 2 0
[2025-07-29 11:25:47,035][18477] Stopping InferenceWorker_p0-w0...
[2025-07-29 11:25:47,036][18477] Loop inference_proc0-0_evt_loop terminating...
[2025-07-29 11:25:47,036][17818] Component InferenceWorker_p0-w0 stopped!
[2025-07-29 11:25:47,055][18480] Stopping RolloutWorker_w2...
[2025-07-29 11:25:47,055][18480] Loop rollout_proc2_evt_loop terminating...
[2025-07-29 11:25:47,055][17818] Component RolloutWorker_w2 stopped!
[2025-07-29 11:25:47,056][18484] Stopping RolloutWorker_w3...
[2025-07-29 11:25:47,057][17818] Component RolloutWorker_w3 stopped!
[2025-07-29 11:25:47,057][18482] Stopping RolloutWorker_w5...
[2025-07-29 11:25:47,057][18485] Stopping RolloutWorker_w4...
[2025-07-29 11:25:47,057][18484] Loop rollout_proc3_evt_loop terminating...
[2025-07-29 11:25:47,058][18482] Loop rollout_proc5_evt_loop terminating...
[2025-07-29 11:25:47,057][18479] Stopping RolloutWorker_w1...
[2025-07-29 11:25:47,058][18485] Loop rollout_proc4_evt_loop terminating...
[2025-07-29 11:25:47,057][18481] Stopping RolloutWorker_w6...
[2025-07-29 11:25:47,058][17818] Component RolloutWorker_w6 stopped!
[2025-07-29 11:25:47,058][18479] Loop rollout_proc1_evt_loop terminating...
[2025-07-29 11:25:47,058][17818] Component RolloutWorker_w5 stopped!
[2025-07-29 11:25:47,059][18481] Loop rollout_proc6_evt_loop terminating...
[2025-07-29 11:25:47,059][17818] Component RolloutWorker_w4 stopped!
[2025-07-29 11:25:47,060][18478] Stopping RolloutWorker_w0...
[2025-07-29 11:25:47,060][17818] Component RolloutWorker_w1 stopped!
[2025-07-29 11:25:47,061][18478] Loop rollout_proc0_evt_loop terminating...
[2025-07-29 11:25:47,061][17818] Component RolloutWorker_w0 stopped!
[2025-07-29 11:25:47,061][18483] Stopping RolloutWorker_w7...
[2025-07-29 11:25:47,062][17818] Component RolloutWorker_w7 stopped!
[2025-07-29 11:25:47,062][18483] Loop rollout_proc7_evt_loop terminating...
[2025-07-29 11:25:48,071][18464] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000980_4014080.pth...
[2025-07-29 11:25:48,116][18464] Removing /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000544_2228224.pth
[2025-07-29 11:25:48,121][18464] Saving /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000980_4014080.pth...
[2025-07-29 11:25:48,184][18464] Stopping LearnerWorker_p0...
[2025-07-29 11:25:48,185][18464] Loop learner_proc0_evt_loop terminating...
[2025-07-29 11:25:48,185][17818] Component LearnerWorker_p0 stopped!
[2025-07-29 11:25:48,186][17818] Waiting for process learner_proc0 to stop...
[2025-07-29 11:25:48,739][17818] Waiting for process inference_proc0-0 to join...
[2025-07-29 11:25:48,740][17818] Waiting for process rollout_proc0 to join...
[2025-07-29 11:25:48,741][17818] Waiting for process rollout_proc1 to join...
[2025-07-29 11:25:48,742][17818] Waiting for process rollout_proc2 to join...
[2025-07-29 11:25:48,743][17818] Waiting for process rollout_proc3 to join...
[2025-07-29 11:25:48,744][17818] Waiting for process rollout_proc4 to join...
[2025-07-29 11:25:48,744][17818] Waiting for process rollout_proc5 to join...
[2025-07-29 11:25:48,745][17818] Waiting for process rollout_proc6 to join...
[2025-07-29 11:25:48,746][17818] Waiting for process rollout_proc7 to join...
[2025-07-29 11:25:48,747][17818] Batcher 0 profile tree view:
batching: 0.0806, releasing_batches: 0.0004
[2025-07-29 11:25:48,748][17818] InferenceWorker_p0-w0 profile tree view:
update_model: 0.0057
wait_policy: 0.0000
  wait_policy_total: 1.2174
one_step: 0.0029
  handle_policy_step: 1.1303
    deserialize: 0.0287, stack: 0.0038, obs_to_device_normalize: 0.1447, forward: 0.8114, send_messages: 0.0405
    prepare_outputs: 0.0768
      to_cpu: 0.0473
[2025-07-29 11:25:48,748][17818] Learner 0 profile tree view:
misc: 0.0000, prepare_batch: 2.0204
train: 0.3211
  epoch_init: 0.0000, minibatch_init: 0.0000, losses_postprocess: 0.0003, kl_divergence: 0.0004, after_optimizer: 0.0037
  calculate_losses: 0.0610
    losses_init: 0.0000, forward_head: 0.0493, bptt_initial: 0.0044, tail: 0.0013, advantages_returns: 0.0009, losses: 0.0025
    bptt: 0.0022
      bptt_forward_core: 0.0022
  update: 0.2550
    clip: 0.0023
[2025-07-29 11:25:48,749][17818] RolloutWorker_w0 profile tree view:
wait_for_trajectories: 0.0007, enqueue_policy_requests: 0.0368, env_step: 0.4303, overhead: 0.0194, complete_rollouts: 0.0007
save_policy_outputs: 0.0320
  split_output_tensors: 0.0107
[2025-07-29 11:25:48,750][17818] RolloutWorker_w7 profile tree view:
wait_for_trajectories: 0.0007, enqueue_policy_requests: 0.0368, env_step: 0.4288, overhead: 0.0198, complete_rollouts: 0.0007
save_policy_outputs: 0.0326
  split_output_tensors: 0.0108
[2025-07-29 11:25:48,751][17818] Loop Runner_EvtLoop terminating...
[2025-07-29 11:25:48,752][17818] Runner profile tree view:
main_loop: 11.3815
[2025-07-29 11:25:48,753][17818] Collected {0: 4014080}, FPS: 719.8
[2025-07-29 11:25:59,169][17818] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 11:25:59,170][17818] Overriding arg 'num_workers' with value 1 passed from command line
[2025-07-29 11:25:59,170][17818] Adding new argument 'no_render'=True that is not in the saved config file!
[2025-07-29 11:25:59,171][17818] Adding new argument 'save_video'=True that is not in the saved config file!
[2025-07-29 11:25:59,172][17818] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:25:59,172][17818] Adding new argument 'video_name'=None that is not in the saved config file!
[2025-07-29 11:25:59,173][17818] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:25:59,173][17818] Adding new argument 'max_num_episodes'=10 that is not in the saved config file!
[2025-07-29 11:25:59,175][17818] Adding new argument 'push_to_hub'=False that is not in the saved config file!
[2025-07-29 11:25:59,175][17818] Adding new argument 'hf_repository'=None that is not in the saved config file!
[2025-07-29 11:25:59,176][17818] Adding new argument 'policy_index'=0 that is not in the saved config file!
[2025-07-29 11:25:59,176][17818] Adding new argument 'eval_deterministic'=False that is not in the saved config file!
[2025-07-29 11:25:59,177][17818] Adding new argument 'train_script'=None that is not in the saved config file!
[2025-07-29 11:25:59,177][17818] Adding new argument 'enjoy_script'=None that is not in the saved config file!
[2025-07-29 11:25:59,178][17818] Using frameskip 1 and render_action_repeat=4 for evaluation
[2025-07-29 11:25:59,189][17818] Doom resolution: 160x120, resize resolution: (128, 72)
[2025-07-29 11:25:59,190][17818] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:25:59,191][17818] RunningMeanStd input shape: (1,)
[2025-07-29 11:25:59,203][17818] ConvEncoder: input_channels=3
[2025-07-29 11:25:59,335][17818] Conv encoder output size: 512
[2025-07-29 11:25:59,336][17818] Policy head output size: 512
[2025-07-29 11:26:01,153][17818] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000980_4014080.pth...
[2025-07-29 11:26:02,083][17818] Num frames 100...
[2025-07-29 11:26:02,203][17818] Num frames 200...
[2025-07-29 11:26:02,325][17818] Num frames 300...
[2025-07-29 11:26:02,448][17818] Num frames 400...
[2025-07-29 11:26:02,560][17818] Avg episode rewards: #0: 5.480, true rewards: #0: 4.480
[2025-07-29 11:26:02,561][17818] Avg episode reward: 5.480, avg true_objective: 4.480
[2025-07-29 11:26:02,629][17818] Num frames 500...
[2025-07-29 11:26:02,759][17818] Num frames 600...
[2025-07-29 11:26:02,891][17818] Num frames 700...
[2025-07-29 11:26:03,017][17818] Num frames 800...
[2025-07-29 11:26:03,140][17818] Num frames 900...
[2025-07-29 11:26:03,264][17818] Num frames 1000...
[2025-07-29 11:26:03,386][17818] Num frames 1100...
[2025-07-29 11:26:03,508][17818] Num frames 1200...
[2025-07-29 11:26:03,630][17818] Num frames 1300...
[2025-07-29 11:26:03,753][17818] Num frames 1400...
[2025-07-29 11:26:03,874][17818] Num frames 1500...
[2025-07-29 11:26:03,999][17818] Num frames 1600...
[2025-07-29 11:26:04,119][17818] Num frames 1700...
[2025-07-29 11:26:04,240][17818] Num frames 1800...
[2025-07-29 11:26:04,364][17818] Num frames 1900...
[2025-07-29 11:26:04,489][17818] Num frames 2000...
[2025-07-29 11:26:04,610][17818] Num frames 2100...
[2025-07-29 11:26:04,733][17818] Num frames 2200...
[2025-07-29 11:26:04,857][17818] Num frames 2300...
[2025-07-29 11:26:04,982][17818] Num frames 2400...
[2025-07-29 11:26:05,107][17818] Num frames 2500...
[2025-07-29 11:26:05,218][17818] Avg episode rewards: #0: 29.739, true rewards: #0: 12.740
[2025-07-29 11:26:05,219][17818] Avg episode reward: 29.739, avg true_objective: 12.740
[2025-07-29 11:26:05,284][17818] Num frames 2600...
[2025-07-29 11:26:05,404][17818] Num frames 2700...
[2025-07-29 11:26:05,525][17818] Num frames 2800...
[2025-07-29 11:26:05,646][17818] Num frames 2900...
[2025-07-29 11:26:05,767][17818] Num frames 3000...
[2025-07-29 11:26:05,888][17818] Num frames 3100...
[2025-07-29 11:26:06,010][17818] Avg episode rewards: #0: 23.520, true rewards: #0: 10.520
[2025-07-29 11:26:06,010][17818] Avg episode reward: 23.520, avg true_objective: 10.520
[2025-07-29 11:26:06,064][17818] Num frames 3200...
[2025-07-29 11:26:06,185][17818] Num frames 3300...
[2025-07-29 11:26:06,303][17818] Num frames 3400...
[2025-07-29 11:26:06,422][17818] Num frames 3500...
[2025-07-29 11:26:06,544][17818] Num frames 3600...
[2025-07-29 11:26:06,693][17818] Avg episode rewards: #0: 19.942, true rewards: #0: 9.193
[2025-07-29 11:26:06,694][17818] Avg episode reward: 19.942, avg true_objective: 9.193
[2025-07-29 11:26:06,722][17818] Num frames 3700...
[2025-07-29 11:26:06,841][17818] Num frames 3800...
[2025-07-29 11:26:06,962][17818] Num frames 3900...
[2025-07-29 11:26:07,083][17818] Num frames 4000...
[2025-07-29 11:26:07,201][17818] Num frames 4100...
[2025-07-29 11:26:07,320][17818] Num frames 4200...
[2025-07-29 11:26:07,442][17818] Num frames 4300...
[2025-07-29 11:26:07,595][17818] Avg episode rewards: #0: 18.762, true rewards: #0: 8.762
[2025-07-29 11:26:07,596][17818] Avg episode reward: 18.762, avg true_objective: 8.762
[2025-07-29 11:26:07,619][17818] Num frames 4400...
[2025-07-29 11:26:07,739][17818] Num frames 4500...
[2025-07-29 11:26:07,863][17818] Num frames 4600...
[2025-07-29 11:26:07,985][17818] Num frames 4700...
[2025-07-29 11:26:08,106][17818] Num frames 4800...
[2025-07-29 11:26:08,227][17818] Num frames 4900...
[2025-07-29 11:26:08,351][17818] Num frames 5000...
[2025-07-29 11:26:08,473][17818] Num frames 5100...
[2025-07-29 11:26:08,594][17818] Num frames 5200...
[2025-07-29 11:26:08,716][17818] Num frames 5300...
[2025-07-29 11:26:08,836][17818] Num frames 5400...
[2025-07-29 11:26:08,935][17818] Avg episode rewards: #0: 19.895, true rewards: #0: 9.062
[2025-07-29 11:26:08,936][17818] Avg episode reward: 19.895, avg true_objective: 9.062
[2025-07-29 11:26:09,011][17818] Num frames 5500...
[2025-07-29 11:26:09,134][17818] Num frames 5600...
[2025-07-29 11:26:09,254][17818] Num frames 5700...
[2025-07-29 11:26:09,374][17818] Num frames 5800...
[2025-07-29 11:26:09,496][17818] Num frames 5900...
[2025-07-29 11:26:09,617][17818] Num frames 6000...
[2025-07-29 11:26:09,737][17818] Num frames 6100...
[2025-07-29 11:26:09,859][17818] Num frames 6200...
[2025-07-29 11:26:09,980][17818] Num frames 6300...
[2025-07-29 11:26:10,101][17818] Num frames 6400...
[2025-07-29 11:26:10,230][17818] Avg episode rewards: #0: 20.516, true rewards: #0: 9.230
[2025-07-29 11:26:10,231][17818] Avg episode reward: 20.516, avg true_objective: 9.230
[2025-07-29 11:26:10,280][17818] Num frames 6500...
[2025-07-29 11:26:10,399][17818] Num frames 6600...
[2025-07-29 11:26:10,521][17818] Num frames 6700...
[2025-07-29 11:26:10,644][17818] Num frames 6800...
[2025-07-29 11:26:10,765][17818] Num frames 6900...
[2025-07-29 11:26:10,886][17818] Num frames 7000...
[2025-07-29 11:26:11,007][17818] Num frames 7100...
[2025-07-29 11:26:11,127][17818] Num frames 7200...
[2025-07-29 11:26:11,248][17818] Num frames 7300...
[2025-07-29 11:26:11,370][17818] Num frames 7400...
[2025-07-29 11:26:11,494][17818] Num frames 7500...
[2025-07-29 11:26:11,617][17818] Num frames 7600...
[2025-07-29 11:26:11,741][17818] Num frames 7700...
[2025-07-29 11:26:11,864][17818] Num frames 7800...
[2025-07-29 11:26:11,987][17818] Num frames 7900...
[2025-07-29 11:26:12,111][17818] Num frames 8000...
[2025-07-29 11:26:12,236][17818] Num frames 8100...
[2025-07-29 11:26:12,360][17818] Num frames 8200...
[2025-07-29 11:26:12,485][17818] Num frames 8300...
[2025-07-29 11:26:12,609][17818] Num frames 8400...
[2025-07-29 11:26:12,680][17818] Avg episode rewards: #0: 24.391, true rewards: #0: 10.516
[2025-07-29 11:26:12,681][17818] Avg episode reward: 24.391, avg true_objective: 10.516
[2025-07-29 11:26:12,784][17818] Num frames 8500...
[2025-07-29 11:26:12,907][17818] Num frames 8600...
[2025-07-29 11:26:13,027][17818] Num frames 8700...
[2025-07-29 11:26:13,148][17818] Num frames 8800...
[2025-07-29 11:26:13,273][17818] Num frames 8900...
[2025-07-29 11:26:13,395][17818] Num frames 9000...
[2025-07-29 11:26:13,524][17818] Num frames 9100...
[2025-07-29 11:26:13,638][17818] Avg episode rewards: #0: 23.165, true rewards: #0: 10.166
[2025-07-29 11:26:13,639][17818] Avg episode reward: 23.165, avg true_objective: 10.166
[2025-07-29 11:26:13,701][17818] Num frames 9200...
[2025-07-29 11:26:13,824][17818] Num frames 9300...
[2025-07-29 11:26:13,944][17818] Num frames 9400...
[2025-07-29 11:26:14,067][17818] Num frames 9500...
[2025-07-29 11:26:14,163][17818] Avg episode rewards: #0: 21.733, true rewards: #0: 9.533
[2025-07-29 11:26:14,164][17818] Avg episode reward: 21.733, avg true_objective: 9.533
[2025-07-29 11:26:37,015][17818] Replay video saved to /content/train_dir/default_experiment/replay.mp4!
[2025-07-29 11:30:35,093][17818] Loading existing experiment configuration from /content/train_dir/default_experiment/config.json
[2025-07-29 11:30:35,094][17818] Overriding arg 'num_workers' with value 1 passed from command line
[2025-07-29 11:30:35,095][17818] Adding new argument 'no_render'=True that is not in the saved config file!
[2025-07-29 11:30:35,095][17818] Adding new argument 'save_video'=True that is not in the saved config file!
[2025-07-29 11:30:35,096][17818] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file!
[2025-07-29 11:30:35,096][17818] Adding new argument 'video_name'=None that is not in the saved config file!
[2025-07-29 11:30:35,097][17818] Adding new argument 'max_num_frames'=100000 that is not in the saved config file!
[2025-07-29 11:30:35,097][17818] Adding new argument 'max_num_episodes'=10 that is not in the saved config file!
[2025-07-29 11:30:35,098][17818] Adding new argument 'push_to_hub'=True that is not in the saved config file!
[2025-07-29 11:30:35,099][17818] Adding new argument 'hf_repository'='Dumoura/rl_vizdoom_health_gathering_supreme' that is not in the saved config file!
[2025-07-29 11:30:35,100][17818] Adding new argument 'policy_index'=0 that is not in the saved config file!
[2025-07-29 11:30:35,100][17818] Adding new argument 'eval_deterministic'=False that is not in the saved config file!
[2025-07-29 11:30:35,101][17818] Adding new argument 'train_script'=None that is not in the saved config file!
[2025-07-29 11:30:35,101][17818] Adding new argument 'enjoy_script'=None that is not in the saved config file!
[2025-07-29 11:30:35,102][17818] Using frameskip 1 and render_action_repeat=4 for evaluation
[2025-07-29 11:30:35,108][17818] RunningMeanStd input shape: (3, 72, 128)
[2025-07-29 11:30:35,109][17818] RunningMeanStd input shape: (1,)
[2025-07-29 11:30:35,118][17818] ConvEncoder: input_channels=3
[2025-07-29 11:30:35,151][17818] Conv encoder output size: 512
[2025-07-29 11:30:35,152][17818] Policy head output size: 512
[2025-07-29 11:30:35,170][17818] Loading state from checkpoint /content/train_dir/default_experiment/checkpoint_p0/checkpoint_000000980_4014080.pth...
[2025-07-29 11:30:35,654][17818] Num frames 100...
[2025-07-29 11:30:35,776][17818] Num frames 200...
[2025-07-29 11:30:35,898][17818] Num frames 300...
[2025-07-29 11:30:36,019][17818] Num frames 400...
[2025-07-29 11:30:36,142][17818] Num frames 500...
[2025-07-29 11:30:36,262][17818] Num frames 600...
[2025-07-29 11:30:36,380][17818] Num frames 700...
[2025-07-29 11:30:36,501][17818] Num frames 800...
[2025-07-29 11:30:36,626][17818] Num frames 900...
[2025-07-29 11:30:36,723][17818] Avg episode rewards: #0: 20.350, true rewards: #0: 9.350
[2025-07-29 11:30:36,724][17818] Avg episode reward: 20.350, avg true_objective: 9.350
[2025-07-29 11:30:36,803][17818] Num frames 1000...
[2025-07-29 11:30:36,922][17818] Num frames 1100...
[2025-07-29 11:30:37,049][17818] Num frames 1200...
[2025-07-29 11:30:37,174][17818] Num frames 1300...
[2025-07-29 11:30:37,301][17818] Num frames 1400...
[2025-07-29 11:30:37,421][17818] Num frames 1500...
[2025-07-29 11:30:37,489][17818] Avg episode rewards: #0: 16.555, true rewards: #0: 7.555
[2025-07-29 11:30:37,490][17818] Avg episode reward: 16.555, avg true_objective: 7.555
[2025-07-29 11:30:37,600][17818] Num frames 1600...
[2025-07-29 11:30:37,719][17818] Num frames 1700...
[2025-07-29 11:30:37,842][17818] Num frames 1800...
[2025-07-29 11:30:37,967][17818] Num frames 1900...
[2025-07-29 11:30:38,094][17818] Num frames 2000...
[2025-07-29 11:30:38,225][17818] Num frames 2100...
[2025-07-29 11:30:38,357][17818] Num frames 2200...
[2025-07-29 11:30:38,488][17818] Num frames 2300...
[2025-07-29 11:30:38,617][17818] Num frames 2400...
[2025-07-29 11:30:38,748][17818] Num frames 2500...
[2025-07-29 11:30:38,872][17818] Num frames 2600...
[2025-07-29 11:30:38,991][17818] Num frames 2700...
[2025-07-29 11:30:39,113][17818] Num frames 2800...
[2025-07-29 11:30:39,235][17818] Num frames 2900...
[2025-07-29 11:30:39,359][17818] Num frames 3000...
[2025-07-29 11:30:39,480][17818] Num frames 3100...
[2025-07-29 11:30:39,606][17818] Num frames 3200...
[2025-07-29 11:30:39,729][17818] Num frames 3300...
[2025-07-29 11:30:39,850][17818] Num frames 3400...
[2025-07-29 11:30:39,972][17818] Num frames 3500...
[2025-07-29 11:30:40,095][17818] Num frames 3600...
[2025-07-29 11:30:40,163][17818] Avg episode rewards: #0: 30.036, true rewards: #0: 12.037
[2025-07-29 11:30:40,164][17818] Avg episode reward: 30.036, avg true_objective: 12.037
[2025-07-29 11:30:40,272][17818] Num frames 3700...
[2025-07-29 11:30:40,392][17818] Num frames 3800...
[2025-07-29 11:30:40,513][17818] Num frames 3900...
[2025-07-29 11:30:40,634][17818] Num frames 4000...
[2025-07-29 11:30:40,754][17818] Num frames 4100...
[2025-07-29 11:30:40,874][17818] Num frames 4200...
[2025-07-29 11:30:40,997][17818] Num frames 4300...
[2025-07-29 11:30:41,118][17818] Num frames 4400...
[2025-07-29 11:30:41,241][17818] Num frames 4500...
[2025-07-29 11:30:41,363][17818] Num frames 4600...
[2025-07-29 11:30:41,484][17818] Num frames 4700...
[2025-07-29 11:30:41,607][17818] Num frames 4800...
[2025-07-29 11:30:41,727][17818] Num frames 4900...
[2025-07-29 11:30:41,849][17818] Num frames 5000...
[2025-07-29 11:30:41,969][17818] Num frames 5100...
[2025-07-29 11:30:42,090][17818] Num frames 5200...
[2025-07-29 11:30:42,213][17818] Num frames 5300...
[2025-07-29 11:30:42,336][17818] Num frames 5400...
[2025-07-29 11:30:42,462][17818] Num frames 5500...
[2025-07-29 11:30:42,585][17818] Num frames 5600...
[2025-07-29 11:30:42,709][17818] Num frames 5700...
[2025-07-29 11:30:42,777][17818] Avg episode rewards: #0: 35.777, true rewards: #0: 14.278
[2025-07-29 11:30:42,778][17818] Avg episode reward: 35.777, avg true_objective: 14.278
[2025-07-29 11:30:42,887][17818] Num frames 5800...
[2025-07-29 11:30:43,008][17818] Num frames 5900...
[2025-07-29 11:30:43,130][17818] Num frames 6000...
[2025-07-29 11:30:43,254][17818] Num frames 6100...
[2025-07-29 11:30:43,375][17818] Num frames 6200...
[2025-07-29 11:30:43,496][17818] Num frames 6300...
[2025-07-29 11:30:43,618][17818] Num frames 6400...
[2025-07-29 11:30:43,737][17818] Num frames 6500...
[2025-07-29 11:30:43,862][17818] Num frames 6600...
[2025-07-29 11:30:43,985][17818] Num frames 6700...
[2025-07-29 11:30:44,107][17818] Num frames 6800...
[2025-07-29 11:30:44,282][17818] Avg episode rewards: #0: 34.196, true rewards: #0: 13.796
[2025-07-29 11:30:44,283][17818] Avg episode reward: 34.196, avg true_objective: 13.796
[2025-07-29 11:30:44,286][17818] Num frames 6900...
[2025-07-29 11:30:44,408][17818] Num frames 7000...
[2025-07-29 11:30:44,530][17818] Num frames 7100...
[2025-07-29 11:30:44,652][17818] Num frames 7200...
[2025-07-29 11:30:44,776][17818] Num frames 7300...
[2025-07-29 11:30:44,898][17818] Num frames 7400...
[2025-07-29 11:30:45,021][17818] Num frames 7500...
[2025-07-29 11:30:45,142][17818] Num frames 7600...
[2025-07-29 11:30:45,262][17818] Num frames 7700...
[2025-07-29 11:30:45,384][17818] Num frames 7800...
[2025-07-29 11:30:45,509][17818] Num frames 7900...
[2025-07-29 11:30:45,631][17818] Num frames 8000...
[2025-07-29 11:30:45,752][17818] Num frames 8100...
[2025-07-29 11:30:45,877][17818] Num frames 8200...
[2025-07-29 11:30:45,996][17818] Num frames 8300...
[2025-07-29 11:30:46,120][17818] Num frames 8400...
[2025-07-29 11:30:46,240][17818] Num frames 8500...
[2025-07-29 11:30:46,364][17818] Num frames 8600...
[2025-07-29 11:30:46,488][17818] Num frames 8700...
[2025-07-29 11:30:46,612][17818] Num frames 8800...
[2025-07-29 11:30:46,736][17818] Num frames 8900...
[2025-07-29 11:30:46,912][17818] Avg episode rewards: #0: 37.829, true rewards: #0: 14.997
[2025-07-29 11:30:46,913][17818] Avg episode reward: 37.829, avg true_objective: 14.997
[2025-07-29 11:30:46,915][17818] Num frames 9000...
[2025-07-29 11:30:47,036][17818] Num frames 9100...
[2025-07-29 11:30:47,157][17818] Num frames 9200...
[2025-07-29 11:30:47,282][17818] Num frames 9300...
[2025-07-29 11:30:47,404][17818] Num frames 9400...
[2025-07-29 11:30:47,554][17818] Avg episode rewards: #0: 33.824, true rewards: #0: 13.539
[2025-07-29 11:30:47,555][17818] Avg episode reward: 33.824, avg true_objective: 13.539
[2025-07-29 11:30:47,583][17818] Num frames 9500...
[2025-07-29 11:30:47,703][17818] Num frames 9600...
[2025-07-29 11:30:47,826][17818] Num frames 9700...
[2025-07-29 11:30:47,949][17818] Num frames 9800...
[2025-07-29 11:30:48,070][17818] Num frames 9900...
[2025-07-29 11:30:48,193][17818] Num frames 10000...
[2025-07-29 11:30:48,316][17818] Num frames 10100...
[2025-07-29 11:30:48,441][17818] Num frames 10200...
[2025-07-29 11:30:48,569][17818] Num frames 10300...
[2025-07-29 11:30:48,697][17818] Num frames 10400...
[2025-07-29 11:30:48,822][17818] Num frames 10500...
[2025-07-29 11:30:48,945][17818] Num frames 10600...
[2025-07-29 11:30:49,118][17818] Avg episode rewards: #0: 32.745, true rewards: #0: 13.370
[2025-07-29 11:30:49,119][17818] Avg episode reward: 32.745, avg true_objective: 13.370
[2025-07-29 11:30:49,126][17818] Num frames 10700...
[2025-07-29 11:30:49,244][17818] Num frames 10800...
[2025-07-29 11:30:49,368][17818] Num frames 10900...
[2025-07-29 11:30:49,494][17818] Num frames 11000...
[2025-07-29 11:30:49,624][17818] Num frames 11100...
[2025-07-29 11:30:49,755][17818] Num frames 11200...
[2025-07-29 11:30:49,859][17818] Avg episode rewards: #0: 30.153, true rewards: #0: 12.487
[2025-07-29 11:30:49,860][17818] Avg episode reward: 30.153, avg true_objective: 12.487
[2025-07-29 11:30:49,941][17818] Num frames 11300...
[2025-07-29 11:30:50,072][17818] Num frames 11400...
[2025-07-29 11:30:50,204][17818] Num frames 11500...
[2025-07-29 11:30:50,335][17818] Num frames 11600...
[2025-07-29 11:30:50,464][17818] Num frames 11700...
[2025-07-29 11:30:50,591][17818] Num frames 11800...
[2025-07-29 11:30:50,721][17818] Num frames 11900...
[2025-07-29 11:30:50,852][17818] Num frames 12000...
[2025-07-29 11:30:50,983][17818] Num frames 12100...
[2025-07-29 11:30:51,114][17818] Num frames 12200...
[2025-07-29 11:30:51,243][17818] Num frames 12300...
[2025-07-29 11:30:51,374][17818] Avg episode rewards: #0: 29.558, true rewards: #0: 12.358
[2025-07-29 11:30:51,375][17818] Avg episode reward: 29.558, avg true_objective: 12.358
[2025-07-29 11:31:20,688][17818] Replay video saved to /content/train_dir/default_experiment/replay.mp4!