[2025-08-26 15:23:01,662][89187] Saving configuration to /home/ubuntu/train_dir/default_experiment/config.json... [2025-08-26 15:23:01,663][89187] Rollout worker 0 uses device cpu [2025-08-26 15:23:01,663][89187] Rollout worker 1 uses device cpu [2025-08-26 15:23:01,664][89187] Rollout worker 2 uses device cpu [2025-08-26 15:23:01,664][89187] Rollout worker 3 uses device cpu [2025-08-26 15:23:01,665][89187] Rollout worker 4 uses device cpu [2025-08-26 15:23:01,665][89187] Rollout worker 5 uses device cpu [2025-08-26 15:23:01,666][89187] Rollout worker 6 uses device cpu [2025-08-26 15:23:01,666][89187] Rollout worker 7 uses device cpu [2025-08-26 15:23:01,707][89187] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:23:01,708][89187] InferenceWorker_p0-w0: min num requests: 2 [2025-08-26 15:23:01,738][89187] Starting all processes... [2025-08-26 15:23:01,739][89187] Starting process learner_proc0 [2025-08-26 15:23:01,788][89187] Starting all processes... [2025-08-26 15:23:01,793][89187] Starting process inference_proc0-0 [2025-08-26 15:23:01,794][89187] Starting process rollout_proc0 [2025-08-26 15:23:01,794][89187] Starting process rollout_proc1 [2025-08-26 15:23:01,794][89187] Starting process rollout_proc2 [2025-08-26 15:23:01,795][89187] Starting process rollout_proc3 [2025-08-26 15:23:01,795][89187] Starting process rollout_proc4 [2025-08-26 15:23:01,795][89187] Starting process rollout_proc5 [2025-08-26 15:23:01,795][89187] Starting process rollout_proc6 [2025-08-26 15:23:01,822][89187] Starting process rollout_proc7 [2025-08-26 15:23:03,996][89752] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:23:03,996][89752] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for learning process 0 [2025-08-26 15:23:04,010][89752] Num visible devices: 1 [2025-08-26 15:23:04,027][89752] Starting seed is not provided [2025-08-26 15:23:04,027][89752] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:23:04,027][89752] Initializing actor-critic model on device cuda:0 [2025-08-26 15:23:04,027][89752] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:23:04,028][89752] RunningMeanStd input shape: (1,) [2025-08-26 15:23:04,030][89770] Worker 4 uses CPU cores [4] [2025-08-26 15:23:04,040][89752] ConvEncoder: input_channels=3 [2025-08-26 15:23:04,070][89768] Worker 2 uses CPU cores [2] [2025-08-26 15:23:04,114][89767] Worker 0 uses CPU cores [0] [2025-08-26 15:23:04,158][89766] Worker 1 uses CPU cores [1] [2025-08-26 15:23:04,182][89769] Worker 3 uses CPU cores [3] [2025-08-26 15:23:04,200][89772] Worker 7 uses CPU cores [7] [2025-08-26 15:23:04,210][89752] Conv encoder output size: 512 [2025-08-26 15:23:04,210][89752] Policy head output size: 512 [2025-08-26 15:23:04,214][89774] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:23:04,214][89774] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for inference process 0 [2025-08-26 15:23:04,214][89771] Worker 5 uses CPU cores [5] [2025-08-26 15:23:04,218][89752] Created Actor Critic model with architecture: [2025-08-26 15:23:04,218][89752] 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) ) ) [2025-08-26 15:23:04,228][89774] Num visible devices: 1 [2025-08-26 15:23:04,360][89752] Using optimizer [2025-08-26 15:23:04,368][89773] Worker 6 uses CPU cores [6] [2025-08-26 15:23:04,897][89752] No checkpoints found [2025-08-26 15:23:04,897][89752] Did not load from checkpoint, starting from scratch! [2025-08-26 15:23:04,897][89752] Initialized policy 0 weights for model version 0 [2025-08-26 15:23:04,899][89752] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:23:04,901][89752] LearnerWorker_p0 finished initialization! [2025-08-26 15:23:04,951][89774] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:23:04,952][89774] RunningMeanStd input shape: (1,) [2025-08-26 15:23:04,959][89774] ConvEncoder: input_channels=3 [2025-08-26 15:23:05,010][89774] Conv encoder output size: 512 [2025-08-26 15:23:05,011][89774] Policy head output size: 512 [2025-08-26 15:23:05,034][89187] Inference worker 0-0 is ready! [2025-08-26 15:23:05,035][89187] All inference workers are ready! Signal rollout workers to start! [2025-08-26 15:23:05,062][89771] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:23:05,063][89768] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:23:05,063][89766] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:23:05,063][89770] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:23:05,063][89769] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:23:05,063][89767] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:23:05,063][89773] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:23:05,064][89772] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:23:05,225][89771] Decorrelating experience for 0 frames... [2025-08-26 15:23:05,229][89766] Decorrelating experience for 0 frames... [2025-08-26 15:23:05,280][89769] Decorrelating experience for 0 frames... [2025-08-26 15:23:05,281][89768] Decorrelating experience for 0 frames... [2025-08-26 15:23:05,282][89772] Decorrelating experience for 0 frames... [2025-08-26 15:23:05,284][89773] Decorrelating experience for 0 frames... [2025-08-26 15:23:05,363][89766] Decorrelating experience for 32 frames... [2025-08-26 15:23:05,414][89770] Decorrelating experience for 0 frames... [2025-08-26 15:23:05,450][89771] Decorrelating experience for 32 frames... [2025-08-26 15:23:05,472][89772] Decorrelating experience for 32 frames... [2025-08-26 15:23:05,474][89773] Decorrelating experience for 32 frames... [2025-08-26 15:23:05,552][89766] Decorrelating experience for 64 frames... [2025-08-26 15:23:05,562][89769] Decorrelating experience for 32 frames... [2025-08-26 15:23:05,668][89773] Decorrelating experience for 64 frames... [2025-08-26 15:23:05,688][89767] Decorrelating experience for 0 frames... [2025-08-26 15:23:05,769][89766] Decorrelating experience for 96 frames... [2025-08-26 15:23:05,773][89768] Decorrelating experience for 32 frames... [2025-08-26 15:23:05,800][89769] Decorrelating experience for 64 frames... [2025-08-26 15:23:05,835][89773] Decorrelating experience for 96 frames... [2025-08-26 15:23:05,851][89770] Decorrelating experience for 32 frames... [2025-08-26 15:23:05,881][89767] Decorrelating experience for 32 frames... [2025-08-26 15:23:05,949][89772] Decorrelating experience for 64 frames... [2025-08-26 15:23:05,965][89768] Decorrelating experience for 64 frames... [2025-08-26 15:23:06,026][89770] Decorrelating experience for 64 frames... [2025-08-26 15:23:06,102][89772] Decorrelating experience for 96 frames... [2025-08-26 15:23:06,170][89769] Decorrelating experience for 96 frames... [2025-08-26 15:23:06,179][89768] Decorrelating experience for 96 frames... [2025-08-26 15:23:06,281][89770] Decorrelating experience for 96 frames... [2025-08-26 15:23:06,314][89767] Decorrelating experience for 64 frames... [2025-08-26 15:23:06,468][89767] Decorrelating experience for 96 frames... [2025-08-26 15:23:06,482][89771] Decorrelating experience for 64 frames... [2025-08-26 15:23:06,635][89771] Decorrelating experience for 96 frames... [2025-08-26 15:23:07,143][89752] Signal inference workers to stop experience collection... [2025-08-26 15:23:07,145][89774] InferenceWorker_p0-w0: stopping experience collection [2025-08-26 15:23:08,362][89752] Signal inference workers to resume experience collection... [2025-08-26 15:23:08,363][89774] InferenceWorker_p0-w0: resuming experience collection [2025-08-26 15:23:09,362][89187] Fps is (10 sec: nan, 60 sec: nan, 300 sec: nan). Total num frames: 36864. Throughput: 0: nan. Samples: 2744. Policy #0 lag: (min: 0.0, avg: 0.0, max: 0.0) [2025-08-26 15:23:09,363][89187] Avg episode reward: [(0, '3.823')] [2025-08-26 15:23:09,425][89774] Updated weights for policy 0, policy_version 10 (0.0051) [2025-08-26 15:23:10,715][89774] Updated weights for policy 0, policy_version 20 (0.0007) [2025-08-26 15:23:11,999][89774] Updated weights for policy 0, policy_version 30 (0.0008) [2025-08-26 15:23:13,265][89774] Updated weights for policy 0, policy_version 40 (0.0006) [2025-08-26 15:23:14,362][89187] Fps is (10 sec: 31948.7, 60 sec: 31948.7, 300 sec: 31948.7). Total num frames: 196608. Throughput: 0: 8986.4. Samples: 47676. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:23:14,363][89187] Avg episode reward: [(0, '4.447')] [2025-08-26 15:23:14,363][89752] Saving new best policy, reward=4.447! [2025-08-26 15:23:14,557][89774] Updated weights for policy 0, policy_version 50 (0.0008) [2025-08-26 15:23:15,877][89774] Updated weights for policy 0, policy_version 60 (0.0006) [2025-08-26 15:23:17,164][89774] Updated weights for policy 0, policy_version 70 (0.0007) [2025-08-26 15:23:18,461][89774] Updated weights for policy 0, policy_version 80 (0.0006) [2025-08-26 15:23:19,362][89187] Fps is (10 sec: 31539.1, 60 sec: 31539.1, 300 sec: 31539.1). Total num frames: 352256. Throughput: 0: 6858.8. Samples: 71332. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0) [2025-08-26 15:23:19,363][89187] Avg episode reward: [(0, '4.396')] [2025-08-26 15:23:19,792][89774] Updated weights for policy 0, policy_version 90 (0.0006) [2025-08-26 15:23:21,062][89774] Updated weights for policy 0, policy_version 100 (0.0007) [2025-08-26 15:23:21,700][89187] Heartbeat connected on Batcher_0 [2025-08-26 15:23:21,703][89187] Heartbeat connected on LearnerWorker_p0 [2025-08-26 15:23:21,709][89187] Heartbeat connected on InferenceWorker_p0-w0 [2025-08-26 15:23:21,716][89187] Heartbeat connected on RolloutWorker_w0 [2025-08-26 15:23:21,720][89187] Heartbeat connected on RolloutWorker_w2 [2025-08-26 15:23:21,725][89187] Heartbeat connected on RolloutWorker_w3 [2025-08-26 15:23:21,729][89187] Heartbeat connected on RolloutWorker_w1 [2025-08-26 15:23:21,730][89187] Heartbeat connected on RolloutWorker_w4 [2025-08-26 15:23:21,733][89187] Heartbeat connected on RolloutWorker_w5 [2025-08-26 15:23:21,735][89187] Heartbeat connected on RolloutWorker_w6 [2025-08-26 15:23:21,737][89187] Heartbeat connected on RolloutWorker_w7 [2025-08-26 15:23:22,443][89774] Updated weights for policy 0, policy_version 110 (0.0008) [2025-08-26 15:23:23,770][89774] Updated weights for policy 0, policy_version 120 (0.0006) [2025-08-26 15:23:24,362][89187] Fps is (10 sec: 31129.5, 60 sec: 31402.5, 300 sec: 31402.5). Total num frames: 507904. Throughput: 0: 7673.2. Samples: 117842. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:23:24,363][89187] Avg episode reward: [(0, '4.591')] [2025-08-26 15:23:24,364][89752] Saving new best policy, reward=4.591! [2025-08-26 15:23:25,145][89774] Updated weights for policy 0, policy_version 130 (0.0006) [2025-08-26 15:23:26,468][89774] Updated weights for policy 0, policy_version 140 (0.0007) [2025-08-26 15:23:27,859][89774] Updated weights for policy 0, policy_version 150 (0.0008) [2025-08-26 15:23:29,362][89774] Updated weights for policy 0, policy_version 160 (0.0007) [2025-08-26 15:23:29,362][89187] Fps is (10 sec: 30310.5, 60 sec: 30924.8, 300 sec: 30924.8). Total num frames: 655360. Throughput: 0: 7952.1. Samples: 161786. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:23:29,363][89187] Avg episode reward: [(0, '4.584')] [2025-08-26 15:23:30,650][89774] Updated weights for policy 0, policy_version 170 (0.0006) [2025-08-26 15:23:31,927][89774] Updated weights for policy 0, policy_version 180 (0.0007) [2025-08-26 15:23:33,222][89774] Updated weights for policy 0, policy_version 190 (0.0006) [2025-08-26 15:23:34,362][89187] Fps is (10 sec: 30310.5, 60 sec: 30965.7, 300 sec: 30965.7). Total num frames: 811008. Throughput: 0: 7324.5. Samples: 185856. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0) [2025-08-26 15:23:34,363][89187] Avg episode reward: [(0, '5.248')] [2025-08-26 15:23:34,364][89752] Saving new best policy, reward=5.248! [2025-08-26 15:23:34,532][89774] Updated weights for policy 0, policy_version 200 (0.0008) [2025-08-26 15:23:35,808][89774] Updated weights for policy 0, policy_version 210 (0.0007) [2025-08-26 15:23:37,092][89774] Updated weights for policy 0, policy_version 220 (0.0008) [2025-08-26 15:23:38,386][89774] Updated weights for policy 0, policy_version 230 (0.0007) [2025-08-26 15:23:39,362][89187] Fps is (10 sec: 31539.1, 60 sec: 31129.6, 300 sec: 31129.6). Total num frames: 970752. Throughput: 0: 7687.1. Samples: 233356. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:23:39,363][89187] Avg episode reward: [(0, '4.654')] [2025-08-26 15:23:39,654][89774] Updated weights for policy 0, policy_version 240 (0.0006) [2025-08-26 15:23:40,914][89774] Updated weights for policy 0, policy_version 250 (0.0006) [2025-08-26 15:23:42,248][89774] Updated weights for policy 0, policy_version 260 (0.0006) [2025-08-26 15:23:43,546][89774] Updated weights for policy 0, policy_version 270 (0.0006) [2025-08-26 15:23:44,362][89187] Fps is (10 sec: 31948.6, 60 sec: 31246.6, 300 sec: 31246.6). Total num frames: 1130496. Throughput: 0: 7947.4. Samples: 280902. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:23:44,363][89187] Avg episode reward: [(0, '5.155')] [2025-08-26 15:23:44,879][89774] Updated weights for policy 0, policy_version 280 (0.0008) [2025-08-26 15:23:46,149][89774] Updated weights for policy 0, policy_version 290 (0.0006) [2025-08-26 15:23:47,454][89774] Updated weights for policy 0, policy_version 300 (0.0006) [2025-08-26 15:23:48,721][89774] Updated weights for policy 0, policy_version 310 (0.0007) [2025-08-26 15:23:49,362][89187] Fps is (10 sec: 31539.3, 60 sec: 31232.0, 300 sec: 31232.0). Total num frames: 1286144. Throughput: 0: 7544.5. Samples: 304524. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:23:49,362][89187] Avg episode reward: [(0, '5.855')] [2025-08-26 15:23:49,366][89752] Saving new best policy, reward=5.855! [2025-08-26 15:23:50,008][89774] Updated weights for policy 0, policy_version 320 (0.0006) [2025-08-26 15:23:51,295][89774] Updated weights for policy 0, policy_version 330 (0.0007) [2025-08-26 15:23:52,562][89774] Updated weights for policy 0, policy_version 340 (0.0007) [2025-08-26 15:23:53,860][89774] Updated weights for policy 0, policy_version 350 (0.0007) [2025-08-26 15:23:54,362][89187] Fps is (10 sec: 31948.9, 60 sec: 31402.6, 300 sec: 31402.6). Total num frames: 1449984. Throughput: 0: 7774.3. Samples: 352588. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:23:54,363][89187] Avg episode reward: [(0, '7.038')] [2025-08-26 15:23:54,363][89752] Saving new best policy, reward=7.038! [2025-08-26 15:23:55,134][89774] Updated weights for policy 0, policy_version 360 (0.0006) [2025-08-26 15:23:56,435][89774] Updated weights for policy 0, policy_version 370 (0.0006) [2025-08-26 15:23:57,688][89774] Updated weights for policy 0, policy_version 380 (0.0007) [2025-08-26 15:23:58,959][89774] Updated weights for policy 0, policy_version 390 (0.0007) [2025-08-26 15:23:59,362][89187] Fps is (10 sec: 32357.7, 60 sec: 31457.2, 300 sec: 31457.2). Total num frames: 1609728. Throughput: 0: 7844.1. Samples: 400660. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0) [2025-08-26 15:23:59,363][89187] Avg episode reward: [(0, '8.243')] [2025-08-26 15:23:59,367][89752] Saving new best policy, reward=8.243! [2025-08-26 15:24:00,258][89774] Updated weights for policy 0, policy_version 400 (0.0006) [2025-08-26 15:24:01,500][89774] Updated weights for policy 0, policy_version 410 (0.0006) [2025-08-26 15:24:02,731][89774] Updated weights for policy 0, policy_version 420 (0.0007) [2025-08-26 15:24:04,002][89774] Updated weights for policy 0, policy_version 430 (0.0007) [2025-08-26 15:24:04,362][89187] Fps is (10 sec: 31948.9, 60 sec: 31502.0, 300 sec: 31502.0). Total num frames: 1769472. Throughput: 0: 7857.0. Samples: 424898. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0) [2025-08-26 15:24:04,363][89187] Avg episode reward: [(0, '10.982')] [2025-08-26 15:24:04,373][89752] Saving new best policy, reward=10.982! [2025-08-26 15:24:05,250][89774] Updated weights for policy 0, policy_version 440 (0.0006) [2025-08-26 15:24:06,560][89774] Updated weights for policy 0, policy_version 450 (0.0007) [2025-08-26 15:24:07,850][89774] Updated weights for policy 0, policy_version 460 (0.0006) [2025-08-26 15:24:09,109][89774] Updated weights for policy 0, policy_version 470 (0.0005) [2025-08-26 15:24:09,362][89187] Fps is (10 sec: 32359.0, 60 sec: 31607.5, 300 sec: 31607.5). Total num frames: 1933312. Throughput: 0: 7898.4. Samples: 473270. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:24:09,363][89187] Avg episode reward: [(0, '12.293')] [2025-08-26 15:24:09,365][89752] Saving new best policy, reward=12.293! [2025-08-26 15:24:10,359][89774] Updated weights for policy 0, policy_version 480 (0.0006) [2025-08-26 15:24:11,622][89774] Updated weights for policy 0, policy_version 490 (0.0006) [2025-08-26 15:24:12,848][89774] Updated weights for policy 0, policy_version 500 (0.0007) [2025-08-26 15:24:14,133][89774] Updated weights for policy 0, policy_version 510 (0.0007) [2025-08-26 15:24:14,362][89187] Fps is (10 sec: 32358.3, 60 sec: 31607.5, 300 sec: 31633.7). Total num frames: 2093056. Throughput: 0: 8007.5. Samples: 522124. Policy #0 lag: (min: 0.0, avg: 0.4, max: 2.0) [2025-08-26 15:24:14,363][89187] Avg episode reward: [(0, '14.502')] [2025-08-26 15:24:14,364][89752] Saving new best policy, reward=14.502! [2025-08-26 15:24:15,400][89774] Updated weights for policy 0, policy_version 520 (0.0006) [2025-08-26 15:24:16,648][89774] Updated weights for policy 0, policy_version 530 (0.0006) [2025-08-26 15:24:17,976][89774] Updated weights for policy 0, policy_version 540 (0.0007) [2025-08-26 15:24:19,223][89774] Updated weights for policy 0, policy_version 550 (0.0006) [2025-08-26 15:24:19,362][89187] Fps is (10 sec: 32358.3, 60 sec: 31744.0, 300 sec: 31714.7). Total num frames: 2256896. Throughput: 0: 8009.6. Samples: 546286. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:24:19,363][89187] Avg episode reward: [(0, '15.657')] [2025-08-26 15:24:19,365][89752] Saving new best policy, reward=15.657! [2025-08-26 15:24:20,465][89774] Updated weights for policy 0, policy_version 560 (0.0007) [2025-08-26 15:24:21,734][89774] Updated weights for policy 0, policy_version 570 (0.0007) [2025-08-26 15:24:23,002][89774] Updated weights for policy 0, policy_version 580 (0.0006) [2025-08-26 15:24:24,250][89774] Updated weights for policy 0, policy_version 590 (0.0006) [2025-08-26 15:24:24,362][89187] Fps is (10 sec: 32358.3, 60 sec: 31812.3, 300 sec: 31730.3). Total num frames: 2416640. Throughput: 0: 8031.6. Samples: 594776. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0) [2025-08-26 15:24:24,363][89187] Avg episode reward: [(0, '17.068')] [2025-08-26 15:24:24,363][89752] Saving new best policy, reward=17.068! [2025-08-26 15:24:25,513][89774] Updated weights for policy 0, policy_version 600 (0.0006) [2025-08-26 15:24:26,804][89774] Updated weights for policy 0, policy_version 610 (0.0006) [2025-08-26 15:24:28,120][89774] Updated weights for policy 0, policy_version 620 (0.0006) [2025-08-26 15:24:29,362][89187] Fps is (10 sec: 31948.7, 60 sec: 32017.0, 300 sec: 31744.0). Total num frames: 2576384. Throughput: 0: 8038.0. Samples: 642614. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0) [2025-08-26 15:24:29,363][89187] Avg episode reward: [(0, '15.875')] [2025-08-26 15:24:29,420][89774] Updated weights for policy 0, policy_version 630 (0.0008) [2025-08-26 15:24:30,674][89774] Updated weights for policy 0, policy_version 640 (0.0006) [2025-08-26 15:24:31,961][89774] Updated weights for policy 0, policy_version 650 (0.0007) [2025-08-26 15:24:33,210][89774] Updated weights for policy 0, policy_version 660 (0.0006) [2025-08-26 15:24:34,362][89187] Fps is (10 sec: 32358.5, 60 sec: 32153.6, 300 sec: 31804.2). Total num frames: 2740224. Throughput: 0: 8052.8. Samples: 666902. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:24:34,363][89187] Avg episode reward: [(0, '15.814')] [2025-08-26 15:24:34,453][89774] Updated weights for policy 0, policy_version 670 (0.0006) [2025-08-26 15:24:35,717][89774] Updated weights for policy 0, policy_version 680 (0.0006) [2025-08-26 15:24:36,984][89774] Updated weights for policy 0, policy_version 690 (0.0006) [2025-08-26 15:24:38,230][89774] Updated weights for policy 0, policy_version 700 (0.0006) [2025-08-26 15:24:39,362][89187] Fps is (10 sec: 32358.0, 60 sec: 32153.5, 300 sec: 31812.2). Total num frames: 2899968. Throughput: 0: 8070.3. Samples: 715752. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:24:39,363][89187] Avg episode reward: [(0, '20.151')] [2025-08-26 15:24:39,366][89752] Saving new best policy, reward=20.151! [2025-08-26 15:24:39,551][89774] Updated weights for policy 0, policy_version 710 (0.0007) [2025-08-26 15:24:40,814][89774] Updated weights for policy 0, policy_version 720 (0.0007) [2025-08-26 15:24:42,077][89774] Updated weights for policy 0, policy_version 730 (0.0006) [2025-08-26 15:24:43,319][89774] Updated weights for policy 0, policy_version 740 (0.0007) [2025-08-26 15:24:44,362][89187] Fps is (10 sec: 32358.5, 60 sec: 32221.9, 300 sec: 31862.6). Total num frames: 3063808. Throughput: 0: 8078.9. Samples: 764208. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:24:44,363][89187] Avg episode reward: [(0, '19.523')] [2025-08-26 15:24:44,587][89774] Updated weights for policy 0, policy_version 750 (0.0006) [2025-08-26 15:24:45,831][89774] Updated weights for policy 0, policy_version 760 (0.0006) [2025-08-26 15:24:47,093][89774] Updated weights for policy 0, policy_version 770 (0.0006) [2025-08-26 15:24:48,361][89774] Updated weights for policy 0, policy_version 780 (0.0006) [2025-08-26 15:24:49,362][89187] Fps is (10 sec: 32768.4, 60 sec: 32358.4, 300 sec: 31907.8). Total num frames: 3227648. Throughput: 0: 8083.1. Samples: 788638. Policy #0 lag: (min: 0.0, avg: 0.4, max: 1.0) [2025-08-26 15:24:49,363][89187] Avg episode reward: [(0, '20.350')] [2025-08-26 15:24:49,366][89752] Saving new best policy, reward=20.350! [2025-08-26 15:24:49,608][89774] Updated weights for policy 0, policy_version 790 (0.0006) [2025-08-26 15:24:50,936][89774] Updated weights for policy 0, policy_version 800 (0.0006) [2025-08-26 15:24:52,181][89774] Updated weights for policy 0, policy_version 810 (0.0007) [2025-08-26 15:24:53,517][89774] Updated weights for policy 0, policy_version 820 (0.0006) [2025-08-26 15:24:54,362][89187] Fps is (10 sec: 31948.8, 60 sec: 32221.9, 300 sec: 31870.8). Total num frames: 3383296. Throughput: 0: 8073.2. Samples: 836564. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:24:54,363][89187] Avg episode reward: [(0, '19.603')] [2025-08-26 15:24:54,783][89774] Updated weights for policy 0, policy_version 830 (0.0008) [2025-08-26 15:24:56,026][89774] Updated weights for policy 0, policy_version 840 (0.0006) [2025-08-26 15:24:57,308][89774] Updated weights for policy 0, policy_version 850 (0.0007) [2025-08-26 15:24:58,701][89774] Updated weights for policy 0, policy_version 860 (0.0006) [2025-08-26 15:24:59,362][89187] Fps is (10 sec: 31539.2, 60 sec: 32221.9, 300 sec: 31874.3). Total num frames: 3543040. Throughput: 0: 8044.0. Samples: 884106. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:24:59,363][89187] Avg episode reward: [(0, '24.276')] [2025-08-26 15:24:59,367][89752] Saving /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000865_3543040.pth... [2025-08-26 15:24:59,416][89752] Saving new best policy, reward=24.276! [2025-08-26 15:24:59,959][89774] Updated weights for policy 0, policy_version 870 (0.0006) [2025-08-26 15:25:01,279][89774] Updated weights for policy 0, policy_version 880 (0.0008) [2025-08-26 15:25:02,568][89774] Updated weights for policy 0, policy_version 890 (0.0006) [2025-08-26 15:25:03,846][89774] Updated weights for policy 0, policy_version 900 (0.0006) [2025-08-26 15:25:04,362][89187] Fps is (10 sec: 31129.4, 60 sec: 32085.3, 300 sec: 31806.3). Total num frames: 3694592. Throughput: 0: 8031.7. Samples: 907714. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:25:04,363][89187] Avg episode reward: [(0, '20.412')] [2025-08-26 15:25:06,416][89774] Updated weights for policy 0, policy_version 910 (0.0009) [2025-08-26 15:25:08,489][89774] Updated weights for policy 0, policy_version 920 (0.0008) [2025-08-26 15:25:09,362][89187] Fps is (10 sec: 24985.6, 60 sec: 30993.0, 300 sec: 31300.3). Total num frames: 3792896. Throughput: 0: 7663.3. Samples: 939626. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:25:09,363][89187] Avg episode reward: [(0, '19.739')] [2025-08-26 15:25:09,946][89774] Updated weights for policy 0, policy_version 930 (0.0008) [2025-08-26 15:25:11,516][89774] Updated weights for policy 0, policy_version 940 (0.0006) [2025-08-26 15:25:13,022][89774] Updated weights for policy 0, policy_version 950 (0.0008) [2025-08-26 15:25:14,362][89187] Fps is (10 sec: 22937.6, 60 sec: 30515.2, 300 sec: 31096.8). Total num frames: 3923968. Throughput: 0: 7504.4. Samples: 980314. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:25:14,363][89187] Avg episode reward: [(0, '23.487')] [2025-08-26 15:25:14,513][89774] Updated weights for policy 0, policy_version 960 (0.0007) [2025-08-26 15:25:16,041][89774] Updated weights for policy 0, policy_version 970 (0.0007) [2025-08-26 15:25:17,182][89752] Saving /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:25:17,184][89752] Stopping Batcher_0... [2025-08-26 15:25:17,185][89752] Loop batcher_evt_loop terminating... [2025-08-26 15:25:17,182][89187] Component Batcher_0 stopped! [2025-08-26 15:25:17,199][89774] Weights refcount: 2 0 [2025-08-26 15:25:17,201][89774] Stopping InferenceWorker_p0-w0... [2025-08-26 15:25:17,201][89774] Loop inference_proc0-0_evt_loop terminating... [2025-08-26 15:25:17,201][89187] Component InferenceWorker_p0-w0 stopped! [2025-08-26 15:25:17,211][89772] Stopping RolloutWorker_w7... [2025-08-26 15:25:17,212][89772] Loop rollout_proc7_evt_loop terminating... [2025-08-26 15:25:17,212][89187] Component RolloutWorker_w7 stopped! [2025-08-26 15:25:17,220][89773] Stopping RolloutWorker_w6... [2025-08-26 15:25:17,221][89767] Stopping RolloutWorker_w0... [2025-08-26 15:25:17,220][89187] Component RolloutWorker_w6 stopped! [2025-08-26 15:25:17,221][89773] Loop rollout_proc6_evt_loop terminating... [2025-08-26 15:25:17,221][89767] Loop rollout_proc0_evt_loop terminating... [2025-08-26 15:25:17,221][89187] Component RolloutWorker_w0 stopped! [2025-08-26 15:25:17,227][89770] Stopping RolloutWorker_w4... [2025-08-26 15:25:17,228][89770] Loop rollout_proc4_evt_loop terminating... [2025-08-26 15:25:17,227][89187] Component RolloutWorker_w4 stopped! [2025-08-26 15:25:17,235][89768] Stopping RolloutWorker_w2... [2025-08-26 15:25:17,235][89187] Component RolloutWorker_w2 stopped! [2025-08-26 15:25:17,235][89768] Loop rollout_proc2_evt_loop terminating... [2025-08-26 15:25:17,235][89752] Saving /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:25:17,237][89771] Stopping RolloutWorker_w5... [2025-08-26 15:25:17,237][89771] Loop rollout_proc5_evt_loop terminating... [2025-08-26 15:25:17,237][89187] Component RolloutWorker_w5 stopped! [2025-08-26 15:25:17,248][89766] Stopping RolloutWorker_w1... [2025-08-26 15:25:17,248][89766] Loop rollout_proc1_evt_loop terminating... [2025-08-26 15:25:17,248][89187] Component RolloutWorker_w1 stopped! [2025-08-26 15:25:17,250][89769] Stopping RolloutWorker_w3... [2025-08-26 15:25:17,251][89769] Loop rollout_proc3_evt_loop terminating... [2025-08-26 15:25:17,254][89187] Component RolloutWorker_w3 stopped! [2025-08-26 15:25:17,325][89752] Stopping LearnerWorker_p0... [2025-08-26 15:25:17,325][89752] Loop learner_proc0_evt_loop terminating... [2025-08-26 15:25:17,325][89187] Component LearnerWorker_p0 stopped! [2025-08-26 15:25:17,326][89187] Waiting for process learner_proc0 to stop... [2025-08-26 15:25:18,162][89187] Waiting for process inference_proc0-0 to join... [2025-08-26 15:25:18,164][89187] Waiting for process rollout_proc0 to join... [2025-08-26 15:25:18,165][89187] Waiting for process rollout_proc1 to join... [2025-08-26 15:25:18,165][89187] Waiting for process rollout_proc2 to join... [2025-08-26 15:25:18,166][89187] Waiting for process rollout_proc3 to join... [2025-08-26 15:25:18,167][89187] Waiting for process rollout_proc4 to join... [2025-08-26 15:25:18,167][89187] Waiting for process rollout_proc5 to join... [2025-08-26 15:25:18,168][89187] Waiting for process rollout_proc6 to join... [2025-08-26 15:25:18,168][89187] Waiting for process rollout_proc7 to join... [2025-08-26 15:25:18,169][89187] Batcher 0 profile tree view: batching: 6.4040, releasing_batches: 0.0148 [2025-08-26 15:25:18,169][89187] InferenceWorker_p0-w0 profile tree view: wait_policy: 0.0000 wait_policy_total: 3.8458 update_model: 2.0387 weight_update: 0.0007 one_step: 0.0011 handle_policy_step: 118.4660 deserialize: 5.4686, stack: 0.7081, obs_to_device_normalize: 25.7784, forward: 56.2610, send_messages: 8.1610 prepare_outputs: 16.4081 to_cpu: 10.2349 [2025-08-26 15:25:18,170][89187] Learner 0 profile tree view: misc: 0.0033, prepare_batch: 4.1677 train: 10.1227 epoch_init: 0.0031, minibatch_init: 0.0032, losses_postprocess: 0.1358, kl_divergence: 0.1723, after_optimizer: 1.4787 calculate_losses: 3.9020 losses_init: 0.0017, forward_head: 0.3282, bptt_initial: 2.0125, tail: 0.3131, advantages_returns: 0.0783, losses: 0.5483 bptt: 0.5381 bptt_forward_core: 0.5099 update: 4.2198 clip: 0.4218 [2025-08-26 15:25:18,170][89187] RolloutWorker_w0 profile tree view: wait_for_trajectories: 0.0856, enqueue_policy_requests: 4.2614, env_step: 61.2950, overhead: 5.1159, complete_rollouts: 0.2467 save_policy_outputs: 4.9281 split_output_tensors: 2.4055 [2025-08-26 15:25:18,170][89187] RolloutWorker_w7 profile tree view: wait_for_trajectories: 0.0830, enqueue_policy_requests: 4.2969, env_step: 61.4537, overhead: 5.0304, complete_rollouts: 0.3091 save_policy_outputs: 4.8909 split_output_tensors: 2.3940 [2025-08-26 15:25:18,171][89187] Loop Runner_EvtLoop terminating... [2025-08-26 15:25:18,171][89187] Runner profile tree view: main_loop: 136.4337 [2025-08-26 15:25:18,172][89187] Collected {0: 4005888}, FPS: 29361.4 [2025-08-26 15:25:18,258][89187] Loading existing experiment configuration from /home/ubuntu/train_dir/default_experiment/config.json [2025-08-26 15:25:18,259][89187] Overriding arg 'num_workers' with value 1 passed from command line [2025-08-26 15:25:18,259][89187] Adding new argument 'no_render'=True that is not in the saved config file! [2025-08-26 15:25:18,259][89187] Adding new argument 'save_video'=True that is not in the saved config file! [2025-08-26 15:25:18,260][89187] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:25:18,260][89187] Adding new argument 'video_name'=None that is not in the saved config file! [2025-08-26 15:25:18,260][89187] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:25:18,261][89187] Adding new argument 'max_num_episodes'=10 that is not in the saved config file! [2025-08-26 15:25:18,261][89187] Adding new argument 'push_to_hub'=False that is not in the saved config file! [2025-08-26 15:25:18,261][89187] Adding new argument 'hf_repository'=None that is not in the saved config file! [2025-08-26 15:25:18,262][89187] Adding new argument 'policy_index'=0 that is not in the saved config file! [2025-08-26 15:25:18,263][89187] Adding new argument 'eval_deterministic'=False that is not in the saved config file! [2025-08-26 15:25:18,263][89187] Adding new argument 'train_script'=None that is not in the saved config file! [2025-08-26 15:25:18,263][89187] Adding new argument 'enjoy_script'=None that is not in the saved config file! [2025-08-26 15:25:18,264][89187] Using frameskip 1 and render_action_repeat=4 for evaluation [2025-08-26 15:25:18,279][89187] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:25:18,280][89187] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:25:18,281][89187] RunningMeanStd input shape: (1,) [2025-08-26 15:25:18,287][89187] ConvEncoder: input_channels=3 [2025-08-26 15:25:18,339][89187] Conv encoder output size: 512 [2025-08-26 15:25:18,339][89187] Policy head output size: 512 [2025-08-26 15:25:18,436][89187] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:25:18,438][89187] Could not load from checkpoint, attempt 0 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:25:18,440][89187] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:25:18,440][89187] Could not load from checkpoint, attempt 1 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:25:18,441][89187] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:25:18,442][89187] Could not load from checkpoint, attempt 2 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:25:46,325][89187] Loading existing experiment configuration from /home/ubuntu/train_dir/default_experiment/config.json [2025-08-26 15:25:46,326][89187] Overriding arg 'num_workers' with value 1 passed from command line [2025-08-26 15:25:46,326][89187] Adding new argument 'no_render'=True that is not in the saved config file! [2025-08-26 15:25:46,327][89187] Adding new argument 'save_video'=True that is not in the saved config file! [2025-08-26 15:25:46,327][89187] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:25:46,327][89187] Adding new argument 'video_name'=None that is not in the saved config file! [2025-08-26 15:25:46,328][89187] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:25:46,328][89187] Adding new argument 'max_num_episodes'=10 that is not in the saved config file! [2025-08-26 15:25:46,329][89187] Adding new argument 'push_to_hub'=False that is not in the saved config file! [2025-08-26 15:25:46,329][89187] Adding new argument 'hf_repository'=None that is not in the saved config file! [2025-08-26 15:25:46,329][89187] Adding new argument 'policy_index'=0 that is not in the saved config file! [2025-08-26 15:25:46,330][89187] Adding new argument 'eval_deterministic'=False that is not in the saved config file! [2025-08-26 15:25:46,330][89187] Adding new argument 'train_script'=None that is not in the saved config file! [2025-08-26 15:25:46,330][89187] Adding new argument 'enjoy_script'=None that is not in the saved config file! [2025-08-26 15:25:46,331][89187] Using frameskip 1 and render_action_repeat=4 for evaluation [2025-08-26 15:25:46,346][89187] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:25:46,347][89187] RunningMeanStd input shape: (1,) [2025-08-26 15:25:46,352][89187] ConvEncoder: input_channels=3 [2025-08-26 15:25:46,379][89187] Conv encoder output size: 512 [2025-08-26 15:25:46,379][89187] Policy head output size: 512 [2025-08-26 15:25:46,391][89187] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:25:46,392][89187] Could not load from checkpoint, attempt 0 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:25:46,392][89187] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:25:46,393][89187] Could not load from checkpoint, attempt 1 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:25:46,394][89187] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:25:46,394][89187] Could not load from checkpoint, attempt 2 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:38:04,937][104163] Saving configuration to /home/ubuntu/train_dir/default_experiment/config.json... [2025-08-26 15:38:04,938][104163] Rollout worker 0 uses device cpu [2025-08-26 15:38:04,939][104163] Rollout worker 1 uses device cpu [2025-08-26 15:38:04,939][104163] Rollout worker 2 uses device cpu [2025-08-26 15:38:04,940][104163] Rollout worker 3 uses device cpu [2025-08-26 15:38:04,940][104163] Rollout worker 4 uses device cpu [2025-08-26 15:38:04,940][104163] Rollout worker 5 uses device cpu [2025-08-26 15:38:04,941][104163] Rollout worker 6 uses device cpu [2025-08-26 15:38:04,941][104163] Rollout worker 7 uses device cpu [2025-08-26 15:38:04,980][104163] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:38:04,981][104163] InferenceWorker_p0-w0: min num requests: 2 [2025-08-26 15:38:05,010][104163] Starting all processes... [2025-08-26 15:38:05,010][104163] Starting process learner_proc0 [2025-08-26 15:38:05,060][104163] Starting all processes... [2025-08-26 15:38:05,063][104163] Starting process inference_proc0-0 [2025-08-26 15:38:05,064][104163] Starting process rollout_proc0 [2025-08-26 15:38:05,064][104163] Starting process rollout_proc1 [2025-08-26 15:38:05,064][104163] Starting process rollout_proc2 [2025-08-26 15:38:05,065][104163] Starting process rollout_proc3 [2025-08-26 15:38:05,066][104163] Starting process rollout_proc4 [2025-08-26 15:38:05,066][104163] Starting process rollout_proc5 [2025-08-26 15:38:05,069][104163] Starting process rollout_proc6 [2025-08-26 15:38:05,073][104163] Starting process rollout_proc7 [2025-08-26 15:38:07,205][104382] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:38:07,205][104382] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for inference process 0 [2025-08-26 15:38:07,210][104383] Worker 2 uses CPU cores [2] [2025-08-26 15:38:07,210][104368] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:38:07,210][104368] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for learning process 0 [2025-08-26 15:38:07,218][104382] Num visible devices: 1 [2025-08-26 15:38:07,223][104368] Num visible devices: 1 [2025-08-26 15:38:07,224][104368] Starting seed is not provided [2025-08-26 15:38:07,224][104368] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:38:07,224][104368] Initializing actor-critic model on device cuda:0 [2025-08-26 15:38:07,224][104368] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:38:07,225][104368] RunningMeanStd input shape: (1,) [2025-08-26 15:38:07,235][104368] ConvEncoder: input_channels=3 [2025-08-26 15:38:07,236][104387] Worker 5 uses CPU cores [5] [2025-08-26 15:38:07,254][104384] Worker 3 uses CPU cores [3] [2025-08-26 15:38:07,278][104385] Worker 1 uses CPU cores [1] [2025-08-26 15:38:07,291][104395] Worker 6 uses CPU cores [6] [2025-08-26 15:38:07,302][104386] Worker 4 uses CPU cores [4] [2025-08-26 15:38:07,312][104368] Conv encoder output size: 512 [2025-08-26 15:38:07,312][104368] Policy head output size: 512 [2025-08-26 15:38:07,313][104381] Worker 0 uses CPU cores [0] [2025-08-26 15:38:07,320][104368] Created Actor Critic model with architecture: [2025-08-26 15:38:07,320][104368] 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) ) ) [2025-08-26 15:38:07,368][104396] Worker 7 uses CPU cores [7] [2025-08-26 15:38:07,388][104368] Using optimizer [2025-08-26 15:38:07,937][104368] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:38:07,939][104368] Could not load from checkpoint, attempt 0 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:38:07,939][104368] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:38:07,939][104368] Could not load from checkpoint, attempt 1 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:38:07,940][104368] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:38:07,940][104368] Could not load from checkpoint, attempt 2 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:38:07,940][104368] Did not load from checkpoint, starting from scratch! [2025-08-26 15:38:07,940][104368] Initialized policy 0 weights for model version 0 [2025-08-26 15:38:07,942][104368] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:38:07,942][104368] LearnerWorker_p0 finished initialization! [2025-08-26 15:38:07,994][104382] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:38:07,995][104382] RunningMeanStd input shape: (1,) [2025-08-26 15:38:08,003][104382] ConvEncoder: input_channels=3 [2025-08-26 15:38:08,054][104382] Conv encoder output size: 512 [2025-08-26 15:38:08,055][104382] Policy head output size: 512 [2025-08-26 15:38:08,078][104163] Inference worker 0-0 is ready! [2025-08-26 15:38:08,079][104163] All inference workers are ready! Signal rollout workers to start! [2025-08-26 15:38:08,088][104396] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:38:08,088][104395] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:38:08,088][104384] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:38:08,088][104387] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:38:08,088][104386] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:38:08,088][104383] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:38:08,088][104385] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:38:08,088][104381] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:38:08,233][104383] Decorrelating experience for 0 frames... [2025-08-26 15:38:08,234][104385] Decorrelating experience for 0 frames... [2025-08-26 15:38:08,289][104396] Decorrelating experience for 0 frames... [2025-08-26 15:38:08,289][104386] Decorrelating experience for 0 frames... [2025-08-26 15:38:08,289][104387] Decorrelating experience for 0 frames... [2025-08-26 15:38:08,291][104395] Decorrelating experience for 0 frames... [2025-08-26 15:38:08,390][104383] Decorrelating experience for 32 frames... [2025-08-26 15:38:08,422][104396] Decorrelating experience for 32 frames... [2025-08-26 15:38:08,435][104386] Decorrelating experience for 32 frames... [2025-08-26 15:38:08,469][104384] Decorrelating experience for 0 frames... [2025-08-26 15:38:08,469][104385] Decorrelating experience for 32 frames... [2025-08-26 15:38:08,558][104387] Decorrelating experience for 32 frames... [2025-08-26 15:38:08,575][104396] Decorrelating experience for 64 frames... [2025-08-26 15:38:08,616][104384] Decorrelating experience for 32 frames... [2025-08-26 15:38:08,633][104385] Decorrelating experience for 64 frames... [2025-08-26 15:38:08,681][104386] Decorrelating experience for 64 frames... [2025-08-26 15:38:08,725][104383] Decorrelating experience for 64 frames... [2025-08-26 15:38:08,728][104396] Decorrelating experience for 96 frames... [2025-08-26 15:38:08,791][104387] Decorrelating experience for 64 frames... [2025-08-26 15:38:08,816][104381] Decorrelating experience for 0 frames... [2025-08-26 15:38:08,849][104385] Decorrelating experience for 96 frames... [2025-08-26 15:38:08,928][104395] Decorrelating experience for 32 frames... [2025-08-26 15:38:08,941][104383] Decorrelating experience for 96 frames... [2025-08-26 15:38:08,945][104384] Decorrelating experience for 64 frames... [2025-08-26 15:38:08,975][104381] Decorrelating experience for 32 frames... [2025-08-26 15:38:08,975][104387] Decorrelating experience for 96 frames... [2025-08-26 15:38:09,038][104386] Decorrelating experience for 96 frames... [2025-08-26 15:38:09,113][104384] Decorrelating experience for 96 frames... [2025-08-26 15:38:09,144][104395] Decorrelating experience for 64 frames... [2025-08-26 15:38:09,303][104381] Decorrelating experience for 64 frames... [2025-08-26 15:38:09,329][104395] Decorrelating experience for 96 frames... [2025-08-26 15:38:09,502][104381] Decorrelating experience for 96 frames... [2025-08-26 15:38:09,670][104368] Signal inference workers to stop experience collection... [2025-08-26 15:38:09,672][104382] InferenceWorker_p0-w0: stopping experience collection [2025-08-26 15:38:10,387][104368] Signal inference workers to resume experience collection... [2025-08-26 15:38:10,388][104382] InferenceWorker_p0-w0: resuming experience collection [2025-08-26 15:38:11,482][104382] Updated weights for policy 0, policy_version 10 (0.0044) [2025-08-26 15:38:12,634][104163] Fps is (10 sec: nan, 60 sec: nan, 300 sec: nan). Total num frames: 77824. Throughput: 0: nan. Samples: 7656. Policy #0 lag: (min: 0.0, avg: 0.4, max: 1.0) [2025-08-26 15:38:12,635][104163] Avg episode reward: [(0, '4.254')] [2025-08-26 15:38:12,762][104382] Updated weights for policy 0, policy_version 20 (0.0007) [2025-08-26 15:38:14,040][104382] Updated weights for policy 0, policy_version 30 (0.0007) [2025-08-26 15:38:15,334][104382] Updated weights for policy 0, policy_version 40 (0.0006) [2025-08-26 15:38:16,644][104382] Updated weights for policy 0, policy_version 50 (0.0006) [2025-08-26 15:38:17,634][104163] Fps is (10 sec: 31129.6, 60 sec: 31129.6, 300 sec: 31129.6). Total num frames: 233472. Throughput: 0: 9558.8. Samples: 55450. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:38:17,635][104163] Avg episode reward: [(0, '4.372')] [2025-08-26 15:38:17,646][104368] Saving new best policy, reward=4.372! [2025-08-26 15:38:17,912][104382] Updated weights for policy 0, policy_version 60 (0.0006) [2025-08-26 15:38:19,186][104382] Updated weights for policy 0, policy_version 70 (0.0006) [2025-08-26 15:38:20,454][104382] Updated weights for policy 0, policy_version 80 (0.0006) [2025-08-26 15:38:21,690][104382] Updated weights for policy 0, policy_version 90 (0.0006) [2025-08-26 15:38:22,634][104163] Fps is (10 sec: 31948.7, 60 sec: 31948.7, 300 sec: 31948.7). Total num frames: 397312. Throughput: 0: 7190.0. Samples: 79556. Policy #0 lag: (min: 0.0, avg: 0.4, max: 2.0) [2025-08-26 15:38:22,635][104163] Avg episode reward: [(0, '4.791')] [2025-08-26 15:38:22,638][104368] Saving new best policy, reward=4.791! [2025-08-26 15:38:23,018][104382] Updated weights for policy 0, policy_version 100 (0.0006) [2025-08-26 15:38:24,266][104382] Updated weights for policy 0, policy_version 110 (0.0007) [2025-08-26 15:38:24,972][104163] Heartbeat connected on Batcher_0 [2025-08-26 15:38:24,976][104163] Heartbeat connected on LearnerWorker_p0 [2025-08-26 15:38:24,982][104163] Heartbeat connected on InferenceWorker_p0-w0 [2025-08-26 15:38:24,985][104163] Heartbeat connected on RolloutWorker_w0 [2025-08-26 15:38:24,993][104163] Heartbeat connected on RolloutWorker_w2 [2025-08-26 15:38:24,997][104163] Heartbeat connected on RolloutWorker_w1 [2025-08-26 15:38:24,999][104163] Heartbeat connected on RolloutWorker_w3 [2025-08-26 15:38:25,000][104163] Heartbeat connected on RolloutWorker_w4 [2025-08-26 15:38:25,003][104163] Heartbeat connected on RolloutWorker_w5 [2025-08-26 15:38:25,007][104163] Heartbeat connected on RolloutWorker_w6 [2025-08-26 15:38:25,009][104163] Heartbeat connected on RolloutWorker_w7 [2025-08-26 15:38:25,574][104382] Updated weights for policy 0, policy_version 120 (0.0006) [2025-08-26 15:38:26,840][104382] Updated weights for policy 0, policy_version 130 (0.0008) [2025-08-26 15:38:27,634][104163] Fps is (10 sec: 32358.5, 60 sec: 31948.9, 300 sec: 31948.9). Total num frames: 557056. Throughput: 0: 7992.1. Samples: 127538. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:38:27,635][104163] Avg episode reward: [(0, '4.507')] [2025-08-26 15:38:28,117][104382] Updated weights for policy 0, policy_version 140 (0.0008) [2025-08-26 15:38:29,377][104382] Updated weights for policy 0, policy_version 150 (0.0006) [2025-08-26 15:38:30,652][104382] Updated weights for policy 0, policy_version 160 (0.0006) [2025-08-26 15:38:31,912][104382] Updated weights for policy 0, policy_version 170 (0.0006) [2025-08-26 15:38:32,634][104163] Fps is (10 sec: 31948.9, 60 sec: 31948.8, 300 sec: 31948.8). Total num frames: 716800. Throughput: 0: 8420.7. Samples: 176070. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:38:32,635][104163] Avg episode reward: [(0, '4.768')] [2025-08-26 15:38:33,192][104382] Updated weights for policy 0, policy_version 180 (0.0007) [2025-08-26 15:38:34,427][104382] Updated weights for policy 0, policy_version 190 (0.0006) [2025-08-26 15:38:35,749][104382] Updated weights for policy 0, policy_version 200 (0.0009) [2025-08-26 15:38:37,057][104382] Updated weights for policy 0, policy_version 210 (0.0006) [2025-08-26 15:38:37,634][104163] Fps is (10 sec: 31948.6, 60 sec: 31948.8, 300 sec: 31948.8). Total num frames: 876544. Throughput: 0: 7704.1. Samples: 200258. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:38:37,635][104163] Avg episode reward: [(0, '5.803')] [2025-08-26 15:38:37,636][104368] Saving new best policy, reward=5.803! [2025-08-26 15:38:38,333][104382] Updated weights for policy 0, policy_version 220 (0.0007) [2025-08-26 15:38:39,601][104382] Updated weights for policy 0, policy_version 230 (0.0006) [2025-08-26 15:38:40,881][104382] Updated weights for policy 0, policy_version 240 (0.0007) [2025-08-26 15:38:42,136][104382] Updated weights for policy 0, policy_version 250 (0.0008) [2025-08-26 15:38:42,634][104163] Fps is (10 sec: 31948.9, 60 sec: 31948.8, 300 sec: 31948.8). Total num frames: 1036288. Throughput: 0: 8015.1. Samples: 248110. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0) [2025-08-26 15:38:42,635][104163] Avg episode reward: [(0, '7.211')] [2025-08-26 15:38:42,654][104368] Saving new best policy, reward=7.211! [2025-08-26 15:38:43,412][104382] Updated weights for policy 0, policy_version 260 (0.0006) [2025-08-26 15:38:44,649][104382] Updated weights for policy 0, policy_version 270 (0.0007) [2025-08-26 15:38:45,934][104382] Updated weights for policy 0, policy_version 280 (0.0006) [2025-08-26 15:38:47,213][104382] Updated weights for policy 0, policy_version 290 (0.0007) [2025-08-26 15:38:47,634][104163] Fps is (10 sec: 32358.5, 60 sec: 32065.8, 300 sec: 32065.8). Total num frames: 1200128. Throughput: 0: 8253.6. Samples: 296532. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0) [2025-08-26 15:38:47,636][104163] Avg episode reward: [(0, '8.603')] [2025-08-26 15:38:47,637][104368] Saving new best policy, reward=8.603! [2025-08-26 15:38:48,516][104382] Updated weights for policy 0, policy_version 300 (0.0006) [2025-08-26 15:38:49,760][104382] Updated weights for policy 0, policy_version 310 (0.0007) [2025-08-26 15:38:51,053][104382] Updated weights for policy 0, policy_version 320 (0.0006) [2025-08-26 15:38:52,322][104382] Updated weights for policy 0, policy_version 330 (0.0006) [2025-08-26 15:38:52,634][104163] Fps is (10 sec: 32358.1, 60 sec: 32051.1, 300 sec: 32051.1). Total num frames: 1359872. Throughput: 0: 7820.9. Samples: 320492. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:38:52,636][104163] Avg episode reward: [(0, '10.093')] [2025-08-26 15:38:52,638][104368] Saving new best policy, reward=10.093! [2025-08-26 15:38:53,601][104382] Updated weights for policy 0, policy_version 340 (0.0007) [2025-08-26 15:38:54,865][104382] Updated weights for policy 0, policy_version 350 (0.0007) [2025-08-26 15:38:56,138][104382] Updated weights for policy 0, policy_version 360 (0.0007) [2025-08-26 15:38:57,392][104382] Updated weights for policy 0, policy_version 370 (0.0006) [2025-08-26 15:38:57,634][104163] Fps is (10 sec: 31948.9, 60 sec: 32039.9, 300 sec: 32039.9). Total num frames: 1519616. Throughput: 0: 8028.1. Samples: 368922. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:38:57,635][104163] Avg episode reward: [(0, '12.715')] [2025-08-26 15:38:57,645][104368] Saving new best policy, reward=12.715! [2025-08-26 15:38:58,691][104382] Updated weights for policy 0, policy_version 380 (0.0006) [2025-08-26 15:38:59,966][104382] Updated weights for policy 0, policy_version 390 (0.0007) [2025-08-26 15:39:01,246][104382] Updated weights for policy 0, policy_version 400 (0.0006) [2025-08-26 15:39:02,501][104382] Updated weights for policy 0, policy_version 410 (0.0006) [2025-08-26 15:39:02,634][104163] Fps is (10 sec: 32358.7, 60 sec: 32112.6, 300 sec: 32112.6). Total num frames: 1683456. Throughput: 0: 8035.7. Samples: 417058. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:39:02,635][104163] Avg episode reward: [(0, '14.037')] [2025-08-26 15:39:02,638][104368] Saving new best policy, reward=14.037! [2025-08-26 15:39:03,776][104382] Updated weights for policy 0, policy_version 420 (0.0006) [2025-08-26 15:39:05,041][104382] Updated weights for policy 0, policy_version 430 (0.0006) [2025-08-26 15:39:06,293][104382] Updated weights for policy 0, policy_version 440 (0.0006) [2025-08-26 15:39:07,574][104382] Updated weights for policy 0, policy_version 450 (0.0007) [2025-08-26 15:39:07,634][104163] Fps is (10 sec: 32358.0, 60 sec: 32097.7, 300 sec: 32097.7). Total num frames: 1843200. Throughput: 0: 8039.5. Samples: 441334. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:39:07,635][104163] Avg episode reward: [(0, '15.014')] [2025-08-26 15:39:07,636][104368] Saving new best policy, reward=15.014! [2025-08-26 15:39:08,859][104382] Updated weights for policy 0, policy_version 460 (0.0006) [2025-08-26 15:39:10,173][104382] Updated weights for policy 0, policy_version 470 (0.0006) [2025-08-26 15:39:11,439][104382] Updated weights for policy 0, policy_version 480 (0.0007) [2025-08-26 15:39:12,634][104163] Fps is (10 sec: 31948.6, 60 sec: 32085.3, 300 sec: 32085.3). Total num frames: 2002944. Throughput: 0: 8039.2. Samples: 489302. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:39:12,636][104163] Avg episode reward: [(0, '17.896')] [2025-08-26 15:39:12,639][104368] Saving new best policy, reward=17.896! [2025-08-26 15:39:12,720][104382] Updated weights for policy 0, policy_version 490 (0.0006) [2025-08-26 15:39:13,967][104382] Updated weights for policy 0, policy_version 500 (0.0008) [2025-08-26 15:39:15,242][104382] Updated weights for policy 0, policy_version 510 (0.0006) [2025-08-26 15:39:16,497][104382] Updated weights for policy 0, policy_version 520 (0.0006) [2025-08-26 15:39:17,634][104163] Fps is (10 sec: 31949.1, 60 sec: 32153.6, 300 sec: 32074.8). Total num frames: 2162688. Throughput: 0: 8038.4. Samples: 537796. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:39:17,635][104163] Avg episode reward: [(0, '18.098')] [2025-08-26 15:39:17,636][104368] Saving new best policy, reward=18.098! [2025-08-26 15:39:17,778][104382] Updated weights for policy 0, policy_version 530 (0.0007) [2025-08-26 15:39:19,064][104382] Updated weights for policy 0, policy_version 540 (0.0007) [2025-08-26 15:39:20,351][104382] Updated weights for policy 0, policy_version 550 (0.0006) [2025-08-26 15:39:21,645][104382] Updated weights for policy 0, policy_version 560 (0.0007) [2025-08-26 15:39:22,634][104163] Fps is (10 sec: 31948.9, 60 sec: 32085.4, 300 sec: 32065.8). Total num frames: 2322432. Throughput: 0: 8032.7. Samples: 561728. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0) [2025-08-26 15:39:22,635][104163] Avg episode reward: [(0, '17.572')] [2025-08-26 15:39:22,900][104382] Updated weights for policy 0, policy_version 570 (0.0006) [2025-08-26 15:39:24,175][104382] Updated weights for policy 0, policy_version 580 (0.0007) [2025-08-26 15:39:25,438][104382] Updated weights for policy 0, policy_version 590 (0.0007) [2025-08-26 15:39:26,719][104382] Updated weights for policy 0, policy_version 600 (0.0006) [2025-08-26 15:39:27,634][104163] Fps is (10 sec: 32358.3, 60 sec: 32153.6, 300 sec: 32112.6). Total num frames: 2486272. Throughput: 0: 8038.7. Samples: 609850. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:39:27,635][104163] Avg episode reward: [(0, '18.465')] [2025-08-26 15:39:27,636][104368] Saving new best policy, reward=18.465! [2025-08-26 15:39:28,011][104382] Updated weights for policy 0, policy_version 610 (0.0007) [2025-08-26 15:39:29,306][104382] Updated weights for policy 0, policy_version 620 (0.0007) [2025-08-26 15:39:30,598][104382] Updated weights for policy 0, policy_version 630 (0.0006) [2025-08-26 15:39:31,889][104382] Updated weights for policy 0, policy_version 640 (0.0006) [2025-08-26 15:39:32,634][104163] Fps is (10 sec: 31948.8, 60 sec: 32085.3, 300 sec: 32051.2). Total num frames: 2641920. Throughput: 0: 8021.4. Samples: 657494. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0) [2025-08-26 15:39:32,635][104163] Avg episode reward: [(0, '20.794')] [2025-08-26 15:39:32,638][104368] Saving new best policy, reward=20.794! [2025-08-26 15:39:33,174][104382] Updated weights for policy 0, policy_version 650 (0.0008) [2025-08-26 15:39:34,441][104382] Updated weights for policy 0, policy_version 660 (0.0006) [2025-08-26 15:39:35,688][104382] Updated weights for policy 0, policy_version 670 (0.0006) [2025-08-26 15:39:36,989][104382] Updated weights for policy 0, policy_version 680 (0.0006) [2025-08-26 15:39:37,634][104163] Fps is (10 sec: 31948.5, 60 sec: 32153.6, 300 sec: 32093.3). Total num frames: 2805760. Throughput: 0: 8024.4. Samples: 681592. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:39:37,635][104163] Avg episode reward: [(0, '24.101')] [2025-08-26 15:39:37,636][104368] Saving new best policy, reward=24.101! [2025-08-26 15:39:38,265][104382] Updated weights for policy 0, policy_version 690 (0.0006) [2025-08-26 15:39:39,534][104382] Updated weights for policy 0, policy_version 700 (0.0007) [2025-08-26 15:39:40,801][104382] Updated weights for policy 0, policy_version 710 (0.0007) [2025-08-26 15:39:42,100][104382] Updated weights for policy 0, policy_version 720 (0.0006) [2025-08-26 15:39:42,634][104163] Fps is (10 sec: 32358.4, 60 sec: 32153.6, 300 sec: 32085.3). Total num frames: 2965504. Throughput: 0: 8017.8. Samples: 729722. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:39:42,635][104163] Avg episode reward: [(0, '19.267')] [2025-08-26 15:39:43,393][104382] Updated weights for policy 0, policy_version 730 (0.0006) [2025-08-26 15:39:44,674][104382] Updated weights for policy 0, policy_version 740 (0.0006) [2025-08-26 15:39:45,937][104382] Updated weights for policy 0, policy_version 750 (0.0008) [2025-08-26 15:39:47,222][104382] Updated weights for policy 0, policy_version 760 (0.0006) [2025-08-26 15:39:47,634][104163] Fps is (10 sec: 31949.1, 60 sec: 32085.3, 300 sec: 32078.1). Total num frames: 3125248. Throughput: 0: 8016.0. Samples: 777778. Policy #0 lag: (min: 0.0, avg: 0.5, max: 2.0) [2025-08-26 15:39:47,635][104163] Avg episode reward: [(0, '18.555')] [2025-08-26 15:39:48,492][104382] Updated weights for policy 0, policy_version 770 (0.0007) [2025-08-26 15:39:49,757][104382] Updated weights for policy 0, policy_version 780 (0.0007) [2025-08-26 15:39:51,032][104382] Updated weights for policy 0, policy_version 790 (0.0008) [2025-08-26 15:39:52,333][104382] Updated weights for policy 0, policy_version 800 (0.0007) [2025-08-26 15:39:52,634][104163] Fps is (10 sec: 31948.7, 60 sec: 32085.4, 300 sec: 32071.7). Total num frames: 3284992. Throughput: 0: 8011.7. Samples: 801858. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:39:52,635][104163] Avg episode reward: [(0, '22.071')] [2025-08-26 15:39:53,626][104382] Updated weights for policy 0, policy_version 810 (0.0006) [2025-08-26 15:39:54,893][104382] Updated weights for policy 0, policy_version 820 (0.0008) [2025-08-26 15:39:56,197][104382] Updated weights for policy 0, policy_version 830 (0.0006) [2025-08-26 15:39:57,522][104382] Updated weights for policy 0, policy_version 840 (0.0007) [2025-08-26 15:39:57,634][104163] Fps is (10 sec: 31539.1, 60 sec: 32017.0, 300 sec: 32026.8). Total num frames: 3440640. Throughput: 0: 8011.5. Samples: 849818. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:39:57,635][104163] Avg episode reward: [(0, '22.008')] [2025-08-26 15:39:58,764][104382] Updated weights for policy 0, policy_version 850 (0.0007) [2025-08-26 15:40:00,025][104382] Updated weights for policy 0, policy_version 860 (0.0007) [2025-08-26 15:40:01,297][104382] Updated weights for policy 0, policy_version 870 (0.0006) [2025-08-26 15:40:02,541][104382] Updated weights for policy 0, policy_version 880 (0.0006) [2025-08-26 15:40:02,634][104163] Fps is (10 sec: 31948.7, 60 sec: 32017.0, 300 sec: 32060.5). Total num frames: 3604480. Throughput: 0: 8003.2. Samples: 897940. Policy #0 lag: (min: 0.0, avg: 0.6, max: 2.0) [2025-08-26 15:40:02,635][104163] Avg episode reward: [(0, '21.884')] [2025-08-26 15:40:02,639][104368] Saving /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000880_3604480.pth... [2025-08-26 15:40:02,690][104368] Removing /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000865_3543040.pth [2025-08-26 15:40:03,850][104382] Updated weights for policy 0, policy_version 890 (0.0006) [2025-08-26 15:40:05,133][104382] Updated weights for policy 0, policy_version 900 (0.0006) [2025-08-26 15:40:06,397][104382] Updated weights for policy 0, policy_version 910 (0.0006) [2025-08-26 15:40:07,634][104163] Fps is (10 sec: 32358.6, 60 sec: 32017.1, 300 sec: 32055.7). Total num frames: 3764224. Throughput: 0: 8002.6. Samples: 921846. Policy #0 lag: (min: 0.0, avg: 0.7, max: 2.0) [2025-08-26 15:40:07,635][104163] Avg episode reward: [(0, '22.428')] [2025-08-26 15:40:07,656][104382] Updated weights for policy 0, policy_version 920 (0.0007) [2025-08-26 15:40:08,909][104382] Updated weights for policy 0, policy_version 930 (0.0006) [2025-08-26 15:40:10,187][104382] Updated weights for policy 0, policy_version 940 (0.0006) [2025-08-26 15:40:11,440][104382] Updated weights for policy 0, policy_version 950 (0.0006) [2025-08-26 15:40:12,634][104163] Fps is (10 sec: 32358.4, 60 sec: 32085.3, 300 sec: 32085.3). Total num frames: 3928064. Throughput: 0: 8014.8. Samples: 970518. Policy #0 lag: (min: 0.0, avg: 0.6, max: 1.0) [2025-08-26 15:40:12,635][104163] Avg episode reward: [(0, '21.595')] [2025-08-26 15:40:12,723][104382] Updated weights for policy 0, policy_version 960 (0.0006) [2025-08-26 15:40:13,972][104382] Updated weights for policy 0, policy_version 970 (0.0006) [2025-08-26 15:40:14,967][104368] Stopping Batcher_0... [2025-08-26 15:40:14,968][104368] Saving /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:40:14,968][104368] Loop batcher_evt_loop terminating... [2025-08-26 15:40:14,972][104163] Component Batcher_0 stopped! [2025-08-26 15:40:14,991][104395] Stopping RolloutWorker_w6... [2025-08-26 15:40:14,991][104382] Weights refcount: 2 0 [2025-08-26 15:40:14,992][104395] Loop rollout_proc6_evt_loop terminating... [2025-08-26 15:40:14,992][104382] Stopping InferenceWorker_p0-w0... [2025-08-26 15:40:14,993][104382] Loop inference_proc0-0_evt_loop terminating... [2025-08-26 15:40:14,993][104384] Stopping RolloutWorker_w3... [2025-08-26 15:40:14,993][104384] Loop rollout_proc3_evt_loop terminating... [2025-08-26 15:40:14,993][104383] Stopping RolloutWorker_w2... [2025-08-26 15:40:14,993][104383] Loop rollout_proc2_evt_loop terminating... [2025-08-26 15:40:14,995][104381] Stopping RolloutWorker_w0... [2025-08-26 15:40:14,995][104381] Loop rollout_proc0_evt_loop terminating... [2025-08-26 15:40:14,997][104396] Stopping RolloutWorker_w7... [2025-08-26 15:40:14,997][104396] Loop rollout_proc7_evt_loop terminating... [2025-08-26 15:40:14,995][104163] Component RolloutWorker_w6 stopped! [2025-08-26 15:40:14,999][104163] Component InferenceWorker_p0-w0 stopped! [2025-08-26 15:40:15,000][104163] Component RolloutWorker_w3 stopped! [2025-08-26 15:40:15,000][104163] Component RolloutWorker_w2 stopped! [2025-08-26 15:40:15,001][104386] Stopping RolloutWorker_w4... [2025-08-26 15:40:15,001][104386] Loop rollout_proc4_evt_loop terminating... [2025-08-26 15:40:15,001][104163] Component RolloutWorker_w0 stopped! [2025-08-26 15:40:15,001][104163] Component RolloutWorker_w7 stopped! [2025-08-26 15:40:15,002][104163] Component RolloutWorker_w4 stopped! [2025-08-26 15:40:15,030][104387] Stopping RolloutWorker_w5... [2025-08-26 15:40:15,030][104387] Loop rollout_proc5_evt_loop terminating... [2025-08-26 15:40:15,030][104163] Component RolloutWorker_w5 stopped! [2025-08-26 15:40:15,048][104368] Saving new best policy, reward=24.907! [2025-08-26 15:40:15,112][104385] Stopping RolloutWorker_w1... [2025-08-26 15:40:15,113][104385] Loop rollout_proc1_evt_loop terminating... [2025-08-26 15:40:15,112][104163] Component RolloutWorker_w1 stopped! [2025-08-26 15:40:15,109][104368] Saving /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:40:15,182][104368] Stopping LearnerWorker_p0... [2025-08-26 15:40:15,182][104368] Loop learner_proc0_evt_loop terminating... [2025-08-26 15:40:15,182][104163] Component LearnerWorker_p0 stopped! [2025-08-26 15:40:15,184][104163] Waiting for process learner_proc0 to stop... [2025-08-26 15:40:15,946][104163] Waiting for process inference_proc0-0 to join... [2025-08-26 15:40:15,947][104163] Waiting for process rollout_proc0 to join... [2025-08-26 15:40:15,948][104163] Waiting for process rollout_proc1 to join... [2025-08-26 15:40:15,949][104163] Waiting for process rollout_proc2 to join... [2025-08-26 15:40:15,950][104163] Waiting for process rollout_proc3 to join... [2025-08-26 15:40:15,950][104163] Waiting for process rollout_proc4 to join... [2025-08-26 15:40:15,951][104163] Waiting for process rollout_proc5 to join... [2025-08-26 15:40:15,951][104163] Waiting for process rollout_proc6 to join... [2025-08-26 15:40:15,952][104163] Waiting for process rollout_proc7 to join... [2025-08-26 15:40:15,952][104163] Batcher 0 profile tree view: batching: 6.5070, releasing_batches: 0.0151 [2025-08-26 15:40:15,953][104163] InferenceWorker_p0-w0 profile tree view: wait_policy: 0.0000 wait_policy_total: 2.8934 update_model: 1.9419 weight_update: 0.0006 one_step: 0.0012 handle_policy_step: 114.8844 deserialize: 5.4740, stack: 0.6782, obs_to_device_normalize: 25.2069, forward: 54.2009, send_messages: 8.0795 prepare_outputs: 15.7530 to_cpu: 9.7958 [2025-08-26 15:40:15,953][104163] Learner 0 profile tree view: misc: 0.0034, prepare_batch: 3.9240 train: 9.6528 epoch_init: 0.0031, minibatch_init: 0.0033, losses_postprocess: 0.1340, kl_divergence: 0.1665, after_optimizer: 1.3548 calculate_losses: 3.8251 losses_init: 0.0018, forward_head: 0.3181, bptt_initial: 1.9405, tail: 0.3149, advantages_returns: 0.0797, losses: 0.5454 bptt: 0.5430 bptt_forward_core: 0.5146 update: 3.9640 clip: 0.4258 [2025-08-26 15:40:15,954][104163] RolloutWorker_w0 profile tree view: wait_for_trajectories: 0.0902, enqueue_policy_requests: 4.0846, env_step: 57.9990, overhead: 4.8994, complete_rollouts: 0.2964 save_policy_outputs: 4.7429 split_output_tensors: 2.3279 [2025-08-26 15:40:15,954][104163] RolloutWorker_w7 profile tree view: wait_for_trajectories: 0.1038, enqueue_policy_requests: 4.1550, env_step: 59.5275, overhead: 4.9723, complete_rollouts: 0.2782 save_policy_outputs: 4.8431 split_output_tensors: 2.3471 [2025-08-26 15:40:15,955][104163] Loop Runner_EvtLoop terminating... [2025-08-26 15:40:15,956][104163] Runner profile tree view: main_loop: 130.9461 [2025-08-26 15:40:15,956][104163] Collected {0: 4005888}, FPS: 30591.9 [2025-08-26 15:40:16,035][104163] Loading existing experiment configuration from /home/ubuntu/train_dir/default_experiment/config.json [2025-08-26 15:40:16,035][104163] Overriding arg 'num_workers' with value 1 passed from command line [2025-08-26 15:40:16,036][104163] Adding new argument 'no_render'=True that is not in the saved config file! [2025-08-26 15:40:16,036][104163] Adding new argument 'save_video'=True that is not in the saved config file! [2025-08-26 15:40:16,037][104163] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:40:16,037][104163] Adding new argument 'video_name'=None that is not in the saved config file! [2025-08-26 15:40:16,038][104163] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:40:16,039][104163] Adding new argument 'max_num_episodes'=10 that is not in the saved config file! [2025-08-26 15:40:16,039][104163] Adding new argument 'push_to_hub'=False that is not in the saved config file! [2025-08-26 15:40:16,039][104163] Adding new argument 'hf_repository'=None that is not in the saved config file! [2025-08-26 15:40:16,040][104163] Adding new argument 'policy_index'=0 that is not in the saved config file! [2025-08-26 15:40:16,040][104163] Adding new argument 'eval_deterministic'=False that is not in the saved config file! [2025-08-26 15:40:16,041][104163] Adding new argument 'train_script'=None that is not in the saved config file! [2025-08-26 15:40:16,041][104163] Adding new argument 'enjoy_script'=None that is not in the saved config file! [2025-08-26 15:40:16,041][104163] Using frameskip 1 and render_action_repeat=4 for evaluation [2025-08-26 15:40:16,048][104163] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:40:16,049][104163] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:40:16,049][104163] RunningMeanStd input shape: (1,) [2025-08-26 15:40:16,056][104163] ConvEncoder: input_channels=3 [2025-08-26 15:40:16,114][104163] Conv encoder output size: 512 [2025-08-26 15:40:16,116][104163] Policy head output size: 512 [2025-08-26 15:40:16,204][104163] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:40:16,206][104163] Could not load from checkpoint, attempt 0 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:40:16,207][104163] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:40:16,208][104163] Could not load from checkpoint, attempt 1 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:40:16,208][104163] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:40:16,209][104163] Could not load from checkpoint, attempt 2 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:40:21,806][104163] Loading existing experiment configuration from /home/ubuntu/train_dir/default_experiment/config.json [2025-08-26 15:40:21,807][104163] Overriding arg 'num_workers' with value 1 passed from command line [2025-08-26 15:40:21,807][104163] Adding new argument 'no_render'=True that is not in the saved config file! [2025-08-26 15:40:21,808][104163] Adding new argument 'save_video'=True that is not in the saved config file! [2025-08-26 15:40:21,808][104163] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:40:21,809][104163] Adding new argument 'video_name'=None that is not in the saved config file! [2025-08-26 15:40:21,809][104163] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:40:21,810][104163] Adding new argument 'max_num_episodes'=10 that is not in the saved config file! [2025-08-26 15:40:21,810][104163] Adding new argument 'push_to_hub'=False that is not in the saved config file! [2025-08-26 15:40:21,810][104163] Adding new argument 'hf_repository'=None that is not in the saved config file! [2025-08-26 15:40:21,811][104163] Adding new argument 'policy_index'=0 that is not in the saved config file! [2025-08-26 15:40:21,812][104163] Adding new argument 'eval_deterministic'=False that is not in the saved config file! [2025-08-26 15:40:21,812][104163] Adding new argument 'train_script'=None that is not in the saved config file! [2025-08-26 15:40:21,812][104163] Adding new argument 'enjoy_script'=None that is not in the saved config file! [2025-08-26 15:40:21,813][104163] Using frameskip 1 and render_action_repeat=4 for evaluation [2025-08-26 15:40:21,819][104163] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:40:21,820][104163] RunningMeanStd input shape: (1,) [2025-08-26 15:40:21,825][104163] ConvEncoder: input_channels=3 [2025-08-26 15:40:21,848][104163] Conv encoder output size: 512 [2025-08-26 15:40:21,849][104163] Policy head output size: 512 [2025-08-26 15:40:21,856][104163] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:40:21,857][104163] Could not load from checkpoint, attempt 0 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:40:21,858][104163] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:40:21,858][104163] Could not load from checkpoint, attempt 1 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:40:21,859][104163] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:40:21,859][104163] Could not load from checkpoint, attempt 2 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:42:27,844][104163] Loading existing experiment configuration from /home/ubuntu/train_dir/default_experiment/config.json [2025-08-26 15:42:27,845][104163] Overriding arg 'num_workers' with value 1 passed from command line [2025-08-26 15:42:27,845][104163] Adding new argument 'no_render'=True that is not in the saved config file! [2025-08-26 15:42:27,846][104163] Adding new argument 'save_video'=True that is not in the saved config file! [2025-08-26 15:42:27,846][104163] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:42:27,847][104163] Adding new argument 'video_name'=None that is not in the saved config file! [2025-08-26 15:42:27,847][104163] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:42:27,848][104163] Adding new argument 'max_num_episodes'=10 that is not in the saved config file! [2025-08-26 15:42:27,848][104163] Adding new argument 'push_to_hub'=False that is not in the saved config file! [2025-08-26 15:42:27,849][104163] Adding new argument 'hf_repository'=None that is not in the saved config file! [2025-08-26 15:42:27,849][104163] Adding new argument 'policy_index'=0 that is not in the saved config file! [2025-08-26 15:42:27,849][104163] Adding new argument 'eval_deterministic'=False that is not in the saved config file! [2025-08-26 15:42:27,850][104163] Adding new argument 'train_script'=None that is not in the saved config file! [2025-08-26 15:42:27,850][104163] Adding new argument 'enjoy_script'=None that is not in the saved config file! [2025-08-26 15:42:27,851][104163] Using frameskip 1 and render_action_repeat=4 for evaluation [2025-08-26 15:42:27,857][104163] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:42:27,857][104163] RunningMeanStd input shape: (1,) [2025-08-26 15:42:27,864][104163] ConvEncoder: input_channels=3 [2025-08-26 15:42:27,886][104163] Conv encoder output size: 512 [2025-08-26 15:42:27,887][104163] Policy head output size: 512 [2025-08-26 15:42:27,900][104163] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:42:27,901][104163] Could not load from checkpoint, attempt 0 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:42:27,902][104163] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:42:27,903][104163] Could not load from checkpoint, attempt 1 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:42:27,903][104163] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:42:27,904][104163] Could not load from checkpoint, attempt 2 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, 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([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.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. [2025-08-26 15:45:09,707][104163] Loading existing experiment configuration from /home/ubuntu/train_dir/default_experiment/config.json [2025-08-26 15:45:09,708][104163] Overriding arg 'num_workers' with value 1 passed from command line [2025-08-26 15:45:09,708][104163] Adding new argument 'no_render'=True that is not in the saved config file! [2025-08-26 15:45:09,709][104163] Adding new argument 'save_video'=True that is not in the saved config file! [2025-08-26 15:45:09,709][104163] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:45:09,710][104163] Adding new argument 'video_name'=None that is not in the saved config file! [2025-08-26 15:45:09,710][104163] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:45:09,710][104163] Adding new argument 'max_num_episodes'=10 that is not in the saved config file! [2025-08-26 15:45:09,711][104163] Adding new argument 'push_to_hub'=False that is not in the saved config file! [2025-08-26 15:45:09,711][104163] Adding new argument 'hf_repository'=None that is not in the saved config file! [2025-08-26 15:45:09,712][104163] Adding new argument 'policy_index'=0 that is not in the saved config file! [2025-08-26 15:45:09,712][104163] Adding new argument 'eval_deterministic'=False that is not in the saved config file! [2025-08-26 15:45:09,713][104163] Adding new argument 'train_script'=None that is not in the saved config file! [2025-08-26 15:45:09,713][104163] Adding new argument 'enjoy_script'=None that is not in the saved config file! [2025-08-26 15:45:09,714][104163] Using frameskip 1 and render_action_repeat=4 for evaluation [2025-08-26 15:45:09,718][104163] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:45:09,719][104163] RunningMeanStd input shape: (1,) [2025-08-26 15:45:09,725][104163] ConvEncoder: input_channels=3 [2025-08-26 15:45:09,751][104163] Conv encoder output size: 512 [2025-08-26 15:45:09,752][104163] Policy head output size: 512 [2025-08-26 15:45:09,763][104163] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:45:09,764][104163] Could not load from checkpoint, attempt 0 Traceback (most recent call last): File "/home/ubuntu/.local/lib/python3.10/site-packages/sample_factory/algo/learning/learner.py", line 281, in load_checkpoint checkpoint_dict = torch.load(latest_checkpoint, map_location=device) File "/home/ubuntu/.local/lib/python3.10/site-packages/torch/serialization.py", line 1529, in load (storage_offset,) = struct.unpack(" [2025-08-26 15:45:29,589][120868] Worker 7 uses CPU cores [7] [2025-08-26 15:45:30,058][120848] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000978_4005888.pth... [2025-08-26 15:45:30,075][120848] Loading model from checkpoint [2025-08-26 15:45:30,076][120848] Loaded experiment state at self.train_step=978, self.env_steps=4005888 [2025-08-26 15:45:30,076][120848] Initialized policy 0 weights for model version 978 [2025-08-26 15:45:30,077][120848] LearnerWorker_p0 finished initialization! [2025-08-26 15:45:30,077][120848] Using GPUs [0] for process 0 (actually maps to GPUs [0]) [2025-08-26 15:45:30,145][120862] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:45:30,146][120862] RunningMeanStd input shape: (1,) [2025-08-26 15:45:30,152][120862] ConvEncoder: input_channels=3 [2025-08-26 15:45:30,206][120862] Conv encoder output size: 512 [2025-08-26 15:45:30,206][120862] Policy head output size: 512 [2025-08-26 15:45:30,229][120654] Inference worker 0-0 is ready! [2025-08-26 15:45:30,230][120654] All inference workers are ready! Signal rollout workers to start! [2025-08-26 15:45:30,240][120867] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:45:30,240][120866] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:45:30,240][120861] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:45:30,240][120864] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:45:30,240][120863] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:45:30,240][120865] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:45:30,240][120876] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:45:30,241][120868] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:45:30,393][120863] Decorrelating experience for 0 frames... [2025-08-26 15:45:30,445][120866] Decorrelating experience for 0 frames... [2025-08-26 15:45:30,445][120867] Decorrelating experience for 0 frames... [2025-08-26 15:45:30,446][120864] Decorrelating experience for 0 frames... [2025-08-26 15:45:30,446][120865] Decorrelating experience for 0 frames... [2025-08-26 15:45:30,641][120865] Decorrelating experience for 32 frames... [2025-08-26 15:45:30,641][120864] Decorrelating experience for 32 frames... [2025-08-26 15:45:30,681][120863] Decorrelating experience for 32 frames... [2025-08-26 15:45:30,697][120866] Decorrelating experience for 32 frames... [2025-08-26 15:45:30,706][120867] Decorrelating experience for 32 frames... [2025-08-26 15:45:30,718][120868] Decorrelating experience for 0 frames... [2025-08-26 15:45:30,809][120865] Decorrelating experience for 64 frames... [2025-08-26 15:45:30,858][120868] Decorrelating experience for 32 frames... [2025-08-26 15:45:30,864][120866] Decorrelating experience for 64 frames... [2025-08-26 15:45:30,893][120863] Decorrelating experience for 64 frames... [2025-08-26 15:45:30,910][120861] Decorrelating experience for 0 frames... [2025-08-26 15:45:31,003][120865] Decorrelating experience for 96 frames... [2025-08-26 15:45:31,019][120866] Decorrelating experience for 96 frames... [2025-08-26 15:45:31,064][120864] Decorrelating experience for 64 frames... [2025-08-26 15:45:31,111][120863] Decorrelating experience for 96 frames... [2025-08-26 15:45:31,111][120861] Decorrelating experience for 32 frames... [2025-08-26 15:45:31,231][120868] Decorrelating experience for 64 frames... [2025-08-26 15:45:31,272][120864] Decorrelating experience for 96 frames... [2025-08-26 15:45:31,327][120867] Decorrelating experience for 64 frames... [2025-08-26 15:45:31,333][120861] Decorrelating experience for 64 frames... [2025-08-26 15:45:31,533][120861] Decorrelating experience for 96 frames... [2025-08-26 15:45:31,534][120867] Decorrelating experience for 96 frames... [2025-08-26 15:45:31,557][120868] Decorrelating experience for 96 frames... [2025-08-26 15:45:31,618][120876] Decorrelating experience for 0 frames... [2025-08-26 15:45:31,803][120876] Decorrelating experience for 32 frames... [2025-08-26 15:45:31,926][120848] Signal inference workers to stop experience collection... [2025-08-26 15:45:31,928][120862] InferenceWorker_p0-w0: stopping experience collection [2025-08-26 15:45:32,000][120876] Decorrelating experience for 64 frames... [2025-08-26 15:45:32,155][120876] Decorrelating experience for 96 frames... [2025-08-26 15:45:32,696][120848] Signal inference workers to resume experience collection... [2025-08-26 15:45:32,697][120848] Stopping Batcher_0... [2025-08-26 15:45:32,697][120848] Loop batcher_evt_loop terminating... [2025-08-26 15:45:32,704][120654] Component Batcher_0 stopped! [2025-08-26 15:45:32,710][120862] Weights refcount: 2 0 [2025-08-26 15:45:32,714][120862] Stopping InferenceWorker_p0-w0... [2025-08-26 15:45:32,714][120862] Loop inference_proc0-0_evt_loop terminating... [2025-08-26 15:45:32,714][120654] Component InferenceWorker_p0-w0 stopped! [2025-08-26 15:45:32,722][120861] Stopping RolloutWorker_w0... [2025-08-26 15:45:32,723][120861] Loop rollout_proc0_evt_loop terminating... [2025-08-26 15:45:32,722][120654] Component RolloutWorker_w0 stopped! [2025-08-26 15:45:32,724][120876] Stopping RolloutWorker_w6... [2025-08-26 15:45:32,725][120876] Loop rollout_proc6_evt_loop terminating... [2025-08-26 15:45:32,727][120654] Component RolloutWorker_w6 stopped! [2025-08-26 15:45:32,727][120868] Stopping RolloutWorker_w7... [2025-08-26 15:45:32,727][120868] Loop rollout_proc7_evt_loop terminating... [2025-08-26 15:45:32,727][120654] Component RolloutWorker_w7 stopped! [2025-08-26 15:45:32,729][120863] Stopping RolloutWorker_w1... [2025-08-26 15:45:32,729][120863] Loop rollout_proc1_evt_loop terminating... [2025-08-26 15:45:32,729][120654] Component RolloutWorker_w1 stopped! [2025-08-26 15:45:32,729][120864] Stopping RolloutWorker_w2... [2025-08-26 15:45:32,729][120864] Loop rollout_proc2_evt_loop terminating... [2025-08-26 15:45:32,729][120654] Component RolloutWorker_w2 stopped! [2025-08-26 15:45:32,738][120865] Stopping RolloutWorker_w3... [2025-08-26 15:45:32,738][120865] Loop rollout_proc3_evt_loop terminating... [2025-08-26 15:45:32,738][120866] Stopping RolloutWorker_w4... [2025-08-26 15:45:32,738][120654] Component RolloutWorker_w3 stopped! [2025-08-26 15:45:32,738][120866] Loop rollout_proc4_evt_loop terminating... [2025-08-26 15:45:32,739][120654] Component RolloutWorker_w4 stopped! [2025-08-26 15:45:32,839][120867] Stopping RolloutWorker_w5... [2025-08-26 15:45:32,839][120867] Loop rollout_proc5_evt_loop terminating... [2025-08-26 15:45:32,839][120654] Component RolloutWorker_w5 stopped! [2025-08-26 15:45:32,861][120848] Saving /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000980_4014080.pth... [2025-08-26 15:45:32,917][120848] Removing /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000880_3604480.pth [2025-08-26 15:45:32,929][120848] Saving /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000980_4014080.pth... [2025-08-26 15:45:32,996][120848] Stopping LearnerWorker_p0... [2025-08-26 15:45:32,997][120848] Loop learner_proc0_evt_loop terminating... [2025-08-26 15:45:32,996][120654] Component LearnerWorker_p0 stopped! [2025-08-26 15:45:32,998][120654] Waiting for process learner_proc0 to stop... [2025-08-26 15:45:33,607][120654] Waiting for process inference_proc0-0 to join... [2025-08-26 15:45:33,608][120654] Waiting for process rollout_proc0 to join... [2025-08-26 15:45:33,609][120654] Waiting for process rollout_proc1 to join... [2025-08-26 15:45:33,609][120654] Waiting for process rollout_proc2 to join... [2025-08-26 15:45:33,610][120654] Waiting for process rollout_proc3 to join... [2025-08-26 15:45:33,610][120654] Waiting for process rollout_proc4 to join... [2025-08-26 15:45:33,611][120654] Waiting for process rollout_proc5 to join... [2025-08-26 15:45:33,612][120654] Waiting for process rollout_proc6 to join... [2025-08-26 15:45:33,612][120654] Waiting for process rollout_proc7 to join... [2025-08-26 15:45:33,613][120654] Batcher 0 profile tree view: batching: 0.0169, releasing_batches: 0.0003 [2025-08-26 15:45:33,613][120654] InferenceWorker_p0-w0 profile tree view: update_model: 0.0036 wait_policy: 0.0000 wait_policy_total: 0.8228 one_step: 0.0139 handle_policy_step: 0.8544 deserialize: 0.0249, stack: 0.0021, obs_to_device_normalize: 0.1393, forward: 0.5971, send_messages: 0.0208 prepare_outputs: 0.0467 to_cpu: 0.0282 [2025-08-26 15:45:33,614][120654] Learner 0 profile tree view: misc: 0.0000, prepare_batch: 0.5212 train: 0.5603 epoch_init: 0.0000, minibatch_init: 0.0000, losses_postprocess: 0.0002, kl_divergence: 0.0058, after_optimizer: 0.0317 calculate_losses: 0.2745 losses_init: 0.0000, forward_head: 0.0395, bptt_initial: 0.1919, tail: 0.0256, advantages_returns: 0.0005, losses: 0.0154 bptt: 0.0013 bptt_forward_core: 0.0013 update: 0.2475 clip: 0.0241 [2025-08-26 15:45:33,614][120654] RolloutWorker_w0 profile tree view: wait_for_trajectories: 0.0003, enqueue_policy_requests: 0.0129, env_step: 0.1365, overhead: 0.0116, complete_rollouts: 0.0005 save_policy_outputs: 0.0114 split_output_tensors: 0.0057 [2025-08-26 15:45:33,615][120654] RolloutWorker_w7 profile tree view: wait_for_trajectories: 0.0003, enqueue_policy_requests: 0.0108, env_step: 0.1146, overhead: 0.0092, complete_rollouts: 0.0003 save_policy_outputs: 0.0090 split_output_tensors: 0.0044 [2025-08-26 15:45:33,615][120654] Loop Runner_EvtLoop terminating... [2025-08-26 15:45:33,616][120654] Runner profile tree view: main_loop: 6.6058 [2025-08-26 15:45:33,616][120654] Collected {0: 4014080}, FPS: 1240.1 [2025-08-26 15:45:33,696][120654] Loading existing experiment configuration from /home/ubuntu/train_dir/default_experiment/config.json [2025-08-26 15:45:33,697][120654] Overriding arg 'num_workers' with value 1 passed from command line [2025-08-26 15:45:33,698][120654] Adding new argument 'no_render'=True that is not in the saved config file! [2025-08-26 15:45:33,698][120654] Adding new argument 'save_video'=True that is not in the saved config file! [2025-08-26 15:45:33,699][120654] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:45:33,699][120654] Adding new argument 'video_name'=None that is not in the saved config file! [2025-08-26 15:45:33,700][120654] Adding new argument 'max_num_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:45:33,700][120654] Adding new argument 'max_num_episodes'=10 that is not in the saved config file! [2025-08-26 15:45:33,701][120654] Adding new argument 'push_to_hub'=False that is not in the saved config file! [2025-08-26 15:45:33,701][120654] Adding new argument 'hf_repository'=None that is not in the saved config file! [2025-08-26 15:45:33,701][120654] Adding new argument 'policy_index'=0 that is not in the saved config file! [2025-08-26 15:45:33,702][120654] Adding new argument 'eval_deterministic'=False that is not in the saved config file! [2025-08-26 15:45:33,702][120654] Adding new argument 'train_script'=None that is not in the saved config file! [2025-08-26 15:45:33,703][120654] Adding new argument 'enjoy_script'=None that is not in the saved config file! [2025-08-26 15:45:33,703][120654] Using frameskip 1 and render_action_repeat=4 for evaluation [2025-08-26 15:45:33,709][120654] Doom resolution: 160x120, resize resolution: (128, 72) [2025-08-26 15:45:33,710][120654] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:45:33,711][120654] RunningMeanStd input shape: (1,) [2025-08-26 15:45:33,719][120654] ConvEncoder: input_channels=3 [2025-08-26 15:45:33,777][120654] Conv encoder output size: 512 [2025-08-26 15:45:33,778][120654] Policy head output size: 512 [2025-08-26 15:45:33,867][120654] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000980_4014080.pth... [2025-08-26 15:45:34,393][120654] Num frames 100... [2025-08-26 15:45:34,459][120654] Num frames 200... [2025-08-26 15:45:34,524][120654] Num frames 300... [2025-08-26 15:45:34,588][120654] Num frames 400... [2025-08-26 15:45:34,654][120654] Num frames 500... [2025-08-26 15:45:34,718][120654] Num frames 600... [2025-08-26 15:45:34,781][120654] Num frames 700... [2025-08-26 15:45:34,852][120654] Num frames 800... [2025-08-26 15:45:34,919][120654] Num frames 900... [2025-08-26 15:45:34,974][120654] Avg episode rewards: #0: 21.010, true rewards: #0: 9.010 [2025-08-26 15:45:34,974][120654] Avg episode reward: 21.010, avg true_objective: 9.010 [2025-08-26 15:45:35,038][120654] Num frames 1000... [2025-08-26 15:45:35,104][120654] Num frames 1100... [2025-08-26 15:45:35,165][120654] Num frames 1200... [2025-08-26 15:45:35,232][120654] Num frames 1300... [2025-08-26 15:45:35,302][120654] Num frames 1400... [2025-08-26 15:45:35,370][120654] Num frames 1500... [2025-08-26 15:45:35,435][120654] Num frames 1600... [2025-08-26 15:45:35,499][120654] Num frames 1700... [2025-08-26 15:45:35,566][120654] Num frames 1800... [2025-08-26 15:45:35,634][120654] Num frames 1900... [2025-08-26 15:45:35,699][120654] Num frames 2000... [2025-08-26 15:45:35,764][120654] Num frames 2100... [2025-08-26 15:45:35,849][120654] Avg episode rewards: #0: 26.245, true rewards: #0: 10.745 [2025-08-26 15:45:35,849][120654] Avg episode reward: 26.245, avg true_objective: 10.745 [2025-08-26 15:45:35,884][120654] Num frames 2200... [2025-08-26 15:45:35,952][120654] Num frames 2300... [2025-08-26 15:45:36,018][120654] Num frames 2400... [2025-08-26 15:45:36,083][120654] Num frames 2500... [2025-08-26 15:45:36,147][120654] Num frames 2600... [2025-08-26 15:45:36,220][120654] Avg episode rewards: #0: 20.097, true rewards: #0: 8.763 [2025-08-26 15:45:36,221][120654] Avg episode reward: 20.097, avg true_objective: 8.763 [2025-08-26 15:45:36,265][120654] Num frames 2700... [2025-08-26 15:45:36,331][120654] Num frames 2800... [2025-08-26 15:45:36,392][120654] Num frames 2900... [2025-08-26 15:45:36,458][120654] Num frames 3000... [2025-08-26 15:45:36,526][120654] Num frames 3100... [2025-08-26 15:45:36,593][120654] Num frames 3200... [2025-08-26 15:45:36,658][120654] Num frames 3300... [2025-08-26 15:45:36,727][120654] Num frames 3400... [2025-08-26 15:45:36,791][120654] Num frames 3500... [2025-08-26 15:45:36,859][120654] Num frames 3600... [2025-08-26 15:45:36,922][120654] Num frames 3700... [2025-08-26 15:45:36,990][120654] Num frames 3800... [2025-08-26 15:45:37,055][120654] Num frames 3900... [2025-08-26 15:45:37,157][120654] Avg episode rewards: #0: 23.933, true rewards: #0: 9.932 [2025-08-26 15:45:37,157][120654] Avg episode reward: 23.933, avg true_objective: 9.932 [2025-08-26 15:45:37,175][120654] Num frames 4000... [2025-08-26 15:45:37,239][120654] Num frames 4100... [2025-08-26 15:45:37,301][120654] Num frames 4200... [2025-08-26 15:45:37,361][120654] Num frames 4300... [2025-08-26 15:45:37,427][120654] Num frames 4400... [2025-08-26 15:45:37,491][120654] Num frames 4500... [2025-08-26 15:45:37,554][120654] Num frames 4600... [2025-08-26 15:45:37,620][120654] Num frames 4700... [2025-08-26 15:45:37,689][120654] Num frames 4800... [2025-08-26 15:45:37,758][120654] Num frames 4900... [2025-08-26 15:45:37,823][120654] Num frames 5000... [2025-08-26 15:45:37,887][120654] Num frames 5100... [2025-08-26 15:45:37,954][120654] Num frames 5200... [2025-08-26 15:45:38,041][120654] Avg episode rewards: #0: 25.706, true rewards: #0: 10.506 [2025-08-26 15:45:38,042][120654] Avg episode reward: 25.706, avg true_objective: 10.506 [2025-08-26 15:45:38,076][120654] Num frames 5300... [2025-08-26 15:45:38,140][120654] Num frames 5400... [2025-08-26 15:45:38,202][120654] Num frames 5500... [2025-08-26 15:45:38,264][120654] Num frames 5600... [2025-08-26 15:45:38,329][120654] Num frames 5700... [2025-08-26 15:45:38,393][120654] Num frames 5800... [2025-08-26 15:45:38,456][120654] Num frames 5900... [2025-08-26 15:45:38,517][120654] Num frames 6000... [2025-08-26 15:45:38,578][120654] Num frames 6100... [2025-08-26 15:45:38,641][120654] Num frames 6200... [2025-08-26 15:45:38,726][120654] Avg episode rewards: #0: 25.242, true rewards: #0: 10.408 [2025-08-26 15:45:38,726][120654] Avg episode reward: 25.242, avg true_objective: 10.408 [2025-08-26 15:45:38,762][120654] Num frames 6300... [2025-08-26 15:45:38,828][120654] Num frames 6400... [2025-08-26 15:45:38,892][120654] Num frames 6500... [2025-08-26 15:45:38,956][120654] Num frames 6600... [2025-08-26 15:45:39,018][120654] Num frames 6700... [2025-08-26 15:45:39,083][120654] Num frames 6800... [2025-08-26 15:45:39,150][120654] Num frames 6900... [2025-08-26 15:45:39,216][120654] Num frames 7000... [2025-08-26 15:45:39,282][120654] Num frames 7100... [2025-08-26 15:45:39,341][120654] Avg episode rewards: #0: 24.442, true rewards: #0: 10.156 [2025-08-26 15:45:39,342][120654] Avg episode reward: 24.442, avg true_objective: 10.156 [2025-08-26 15:45:39,400][120654] Num frames 7200... [2025-08-26 15:45:39,464][120654] Num frames 7300... [2025-08-26 15:45:39,526][120654] Num frames 7400... [2025-08-26 15:45:39,591][120654] Num frames 7500... [2025-08-26 15:45:39,654][120654] Num frames 7600... [2025-08-26 15:45:39,716][120654] Num frames 7700... [2025-08-26 15:45:39,780][120654] Num frames 7800... [2025-08-26 15:45:39,844][120654] Num frames 7900... [2025-08-26 15:45:39,908][120654] Num frames 8000... [2025-08-26 15:45:39,977][120654] Num frames 8100... [2025-08-26 15:45:40,042][120654] Num frames 8200... [2025-08-26 15:45:40,107][120654] Num frames 8300... [2025-08-26 15:45:40,171][120654] Num frames 8400... [2025-08-26 15:45:40,237][120654] Num frames 8500... [2025-08-26 15:45:40,340][120654] Avg episode rewards: #0: 25.849, true rewards: #0: 10.724 [2025-08-26 15:45:40,340][120654] Avg episode reward: 25.849, avg true_objective: 10.724 [2025-08-26 15:45:40,355][120654] Num frames 8600... [2025-08-26 15:45:40,415][120654] Num frames 8700... [2025-08-26 15:45:40,476][120654] Num frames 8800... [2025-08-26 15:45:40,541][120654] Num frames 8900... [2025-08-26 15:45:40,607][120654] Num frames 9000... [2025-08-26 15:45:40,672][120654] Num frames 9100... [2025-08-26 15:45:40,738][120654] Num frames 9200... [2025-08-26 15:45:40,803][120654] Num frames 9300... [2025-08-26 15:45:40,869][120654] Num frames 9400... [2025-08-26 15:45:40,936][120654] Num frames 9500... [2025-08-26 15:45:40,999][120654] Num frames 9600... [2025-08-26 15:45:41,067][120654] Num frames 9700... [2025-08-26 15:45:41,134][120654] Num frames 9800... [2025-08-26 15:45:41,200][120654] Num frames 9900... [2025-08-26 15:45:41,268][120654] Num frames 10000... [2025-08-26 15:45:41,333][120654] Num frames 10100... [2025-08-26 15:45:41,398][120654] Num frames 10200... [2025-08-26 15:45:41,461][120654] Num frames 10300... [2025-08-26 15:45:41,526][120654] Num frames 10400... [2025-08-26 15:45:41,592][120654] Avg episode rewards: #0: 28.021, true rewards: #0: 11.577 [2025-08-26 15:45:41,593][120654] Avg episode reward: 28.021, avg true_objective: 11.577 [2025-08-26 15:45:41,645][120654] Num frames 10500... [2025-08-26 15:45:41,710][120654] Num frames 10600... [2025-08-26 15:45:41,772][120654] Num frames 10700... [2025-08-26 15:45:41,836][120654] Num frames 10800... [2025-08-26 15:45:41,901][120654] Num frames 10900... [2025-08-26 15:45:41,963][120654] Num frames 11000... [2025-08-26 15:45:42,024][120654] Num frames 11100... [2025-08-26 15:45:42,088][120654] Num frames 11200... [2025-08-26 15:45:42,154][120654] Num frames 11300... [2025-08-26 15:45:42,236][120654] Avg episode rewards: #0: 27.244, true rewards: #0: 11.344 [2025-08-26 15:45:42,237][120654] Avg episode reward: 27.244, avg true_objective: 11.344 [2025-08-26 15:45:58,295][120654] Replay video saved to /home/ubuntu/train_dir/default_experiment/replay.mp4! [2025-08-26 15:47:13,203][120654] Loading existing experiment configuration from /home/ubuntu/train_dir/default_experiment/config.json [2025-08-26 15:47:13,204][120654] Overriding arg 'num_workers' with value 1 passed from command line [2025-08-26 15:47:13,204][120654] Adding new argument 'no_render'=True that is not in the saved config file! [2025-08-26 15:47:13,205][120654] Adding new argument 'save_video'=True that is not in the saved config file! [2025-08-26 15:47:13,206][120654] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:47:13,206][120654] Adding new argument 'video_name'=None that is not in the saved config file! [2025-08-26 15:47:13,207][120654] Adding new argument 'max_num_frames'=100000 that is not in the saved config file! [2025-08-26 15:47:13,207][120654] Adding new argument 'max_num_episodes'=10 that is not in the saved config file! [2025-08-26 15:47:13,208][120654] Adding new argument 'push_to_hub'=True that is not in the saved config file! [2025-08-26 15:47:13,208][120654] Adding new argument 'hf_repository'='igzi/rl_course_vizdoom_health_gathering_supreme' that is not in the saved config file! [2025-08-26 15:47:13,209][120654] Adding new argument 'policy_index'=0 that is not in the saved config file! [2025-08-26 15:47:13,209][120654] Adding new argument 'eval_deterministic'=False that is not in the saved config file! [2025-08-26 15:47:13,209][120654] Adding new argument 'train_script'=None that is not in the saved config file! [2025-08-26 15:47:13,210][120654] Adding new argument 'enjoy_script'=None that is not in the saved config file! [2025-08-26 15:47:13,210][120654] Using frameskip 1 and render_action_repeat=4 for evaluation [2025-08-26 15:47:13,217][120654] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:47:13,218][120654] RunningMeanStd input shape: (1,) [2025-08-26 15:47:13,225][120654] ConvEncoder: input_channels=3 [2025-08-26 15:47:13,247][120654] Conv encoder output size: 512 [2025-08-26 15:47:13,248][120654] Policy head output size: 512 [2025-08-26 15:47:13,260][120654] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000980_4014080.pth... [2025-08-26 15:47:13,517][120654] Num frames 100... [2025-08-26 15:47:13,582][120654] Num frames 200... [2025-08-26 15:47:13,642][120654] Num frames 300... [2025-08-26 15:47:13,707][120654] Num frames 400... [2025-08-26 15:47:13,776][120654] Num frames 500... [2025-08-26 15:47:13,841][120654] Num frames 600... [2025-08-26 15:47:13,909][120654] Num frames 700... [2025-08-26 15:47:13,972][120654] Num frames 800... [2025-08-26 15:47:14,035][120654] Num frames 900... [2025-08-26 15:47:14,104][120654] Avg episode rewards: #0: 19.250, true rewards: #0: 9.250 [2025-08-26 15:47:14,105][120654] Avg episode reward: 19.250, avg true_objective: 9.250 [2025-08-26 15:47:14,155][120654] Num frames 1000... [2025-08-26 15:47:14,218][120654] Num frames 1100... [2025-08-26 15:47:14,284][120654] Num frames 1200... [2025-08-26 15:47:14,344][120654] Num frames 1300... [2025-08-26 15:47:14,408][120654] Num frames 1400... [2025-08-26 15:47:14,470][120654] Num frames 1500... [2025-08-26 15:47:14,534][120654] Num frames 1600... [2025-08-26 15:47:14,598][120654] Num frames 1700... [2025-08-26 15:47:14,660][120654] Num frames 1800... [2025-08-26 15:47:14,725][120654] Num frames 1900... [2025-08-26 15:47:14,790][120654] Num frames 2000... [2025-08-26 15:47:14,853][120654] Num frames 2100... [2025-08-26 15:47:14,918][120654] Num frames 2200... [2025-08-26 15:47:14,985][120654] Num frames 2300... [2025-08-26 15:47:15,047][120654] Num frames 2400... [2025-08-26 15:47:15,121][120654] Avg episode rewards: #0: 28.175, true rewards: #0: 12.175 [2025-08-26 15:47:15,122][120654] Avg episode reward: 28.175, avg true_objective: 12.175 [2025-08-26 15:47:15,164][120654] Num frames 2500... [2025-08-26 15:47:15,229][120654] Num frames 2600... [2025-08-26 15:47:15,293][120654] Num frames 2700... [2025-08-26 15:47:15,362][120654] Num frames 2800... [2025-08-26 15:47:15,427][120654] Num frames 2900... [2025-08-26 15:47:15,490][120654] Num frames 3000... [2025-08-26 15:47:15,556][120654] Num frames 3100... [2025-08-26 15:47:15,623][120654] Num frames 3200... [2025-08-26 15:47:15,687][120654] Num frames 3300... [2025-08-26 15:47:15,755][120654] Num frames 3400... [2025-08-26 15:47:15,826][120654] Num frames 3500... [2025-08-26 15:47:15,891][120654] Num frames 3600... [2025-08-26 15:47:15,955][120654] Num frames 3700... [2025-08-26 15:47:16,021][120654] Avg episode rewards: #0: 27.737, true rewards: #0: 12.403 [2025-08-26 15:47:16,022][120654] Avg episode reward: 27.737, avg true_objective: 12.403 [2025-08-26 15:47:16,076][120654] Num frames 3800... [2025-08-26 15:47:16,144][120654] Num frames 3900... [2025-08-26 15:47:16,208][120654] Num frames 4000... [2025-08-26 15:47:16,277][120654] Num frames 4100... [2025-08-26 15:47:16,345][120654] Num frames 4200... [2025-08-26 15:47:16,407][120654] Num frames 4300... [2025-08-26 15:47:16,470][120654] Num frames 4400... [2025-08-26 15:47:16,537][120654] Num frames 4500... [2025-08-26 15:47:16,646][120654] Avg episode rewards: #0: 24.963, true rewards: #0: 11.462 [2025-08-26 15:47:16,647][120654] Avg episode reward: 24.963, avg true_objective: 11.462 [2025-08-26 15:47:16,658][120654] Num frames 4600... [2025-08-26 15:47:16,723][120654] Num frames 4700... [2025-08-26 15:47:16,793][120654] Num frames 4800... [2025-08-26 15:47:16,865][120654] Num frames 4900... [2025-08-26 15:47:16,932][120654] Num frames 5000... [2025-08-26 15:47:17,000][120654] Num frames 5100... [2025-08-26 15:47:17,065][120654] Num frames 5200... [2025-08-26 15:47:17,129][120654] Num frames 5300... [2025-08-26 15:47:17,194][120654] Num frames 5400... [2025-08-26 15:47:17,257][120654] Num frames 5500... [2025-08-26 15:47:17,317][120654] Num frames 5600... [2025-08-26 15:47:17,384][120654] Num frames 5700... [2025-08-26 15:47:17,450][120654] Num frames 5800... [2025-08-26 15:47:17,515][120654] Num frames 5900... [2025-08-26 15:47:17,582][120654] Num frames 6000... [2025-08-26 15:47:17,644][120654] Num frames 6100... [2025-08-26 15:47:17,705][120654] Num frames 6200... [2025-08-26 15:47:17,772][120654] Num frames 6300... [2025-08-26 15:47:17,841][120654] Num frames 6400... [2025-08-26 15:47:17,910][120654] Num frames 6500... [2025-08-26 15:47:17,965][120654] Avg episode rewards: #0: 29.004, true rewards: #0: 13.004 [2025-08-26 15:47:17,965][120654] Avg episode reward: 29.004, avg true_objective: 13.004 [2025-08-26 15:47:18,031][120654] Num frames 6600... [2025-08-26 15:47:18,094][120654] Num frames 6700... [2025-08-26 15:47:18,161][120654] Num frames 6800... [2025-08-26 15:47:18,224][120654] Num frames 6900... [2025-08-26 15:47:18,289][120654] Num frames 7000... [2025-08-26 15:47:18,353][120654] Num frames 7100... [2025-08-26 15:47:18,461][120654] Avg episode rewards: #0: 25.978, true rewards: #0: 11.978 [2025-08-26 15:47:18,461][120654] Avg episode reward: 25.978, avg true_objective: 11.978 [2025-08-26 15:47:18,470][120654] Num frames 7200... [2025-08-26 15:47:18,537][120654] Num frames 7300... [2025-08-26 15:47:18,600][120654] Num frames 7400... [2025-08-26 15:47:18,666][120654] Num frames 7500... [2025-08-26 15:47:18,732][120654] Num frames 7600... [2025-08-26 15:47:18,797][120654] Num frames 7700... [2025-08-26 15:47:18,864][120654] Num frames 7800... [2025-08-26 15:47:18,932][120654] Num frames 7900... [2025-08-26 15:47:19,011][120654] Num frames 8000... [2025-08-26 15:47:19,079][120654] Num frames 8100... [2025-08-26 15:47:19,142][120654] Num frames 8200... [2025-08-26 15:47:19,206][120654] Num frames 8300... [2025-08-26 15:47:19,270][120654] Num frames 8400... [2025-08-26 15:47:19,335][120654] Num frames 8500... [2025-08-26 15:47:19,402][120654] Num frames 8600... [2025-08-26 15:47:19,499][120654] Avg episode rewards: #0: 27.669, true rewards: #0: 12.383 [2025-08-26 15:47:19,500][120654] Avg episode reward: 27.669, avg true_objective: 12.383 [2025-08-26 15:47:19,521][120654] Num frames 8700... [2025-08-26 15:47:19,585][120654] Num frames 8800... [2025-08-26 15:47:19,648][120654] Num frames 8900... [2025-08-26 15:47:19,714][120654] Num frames 9000... [2025-08-26 15:47:19,777][120654] Num frames 9100... [2025-08-26 15:47:19,844][120654] Num frames 9200... [2025-08-26 15:47:19,909][120654] Num frames 9300... [2025-08-26 15:47:19,994][120654] Avg episode rewards: #0: 26.059, true rewards: #0: 11.684 [2025-08-26 15:47:19,995][120654] Avg episode reward: 26.059, avg true_objective: 11.684 [2025-08-26 15:47:20,030][120654] Num frames 9400... [2025-08-26 15:47:20,095][120654] Num frames 9500... [2025-08-26 15:47:20,161][120654] Num frames 9600... [2025-08-26 15:47:20,224][120654] Num frames 9700... [2025-08-26 15:47:20,290][120654] Num frames 9800... [2025-08-26 15:47:20,360][120654] Num frames 9900... [2025-08-26 15:47:20,424][120654] Num frames 10000... [2025-08-26 15:47:20,490][120654] Num frames 10100... [2025-08-26 15:47:20,555][120654] Num frames 10200... [2025-08-26 15:47:20,620][120654] Num frames 10300... [2025-08-26 15:47:20,687][120654] Num frames 10400... [2025-08-26 15:47:20,753][120654] Num frames 10500... [2025-08-26 15:47:20,820][120654] Num frames 10600... [2025-08-26 15:47:20,886][120654] Num frames 10700... [2025-08-26 15:47:20,954][120654] Num frames 10800... [2025-08-26 15:47:21,020][120654] Num frames 10900... [2025-08-26 15:47:21,088][120654] Num frames 11000... [2025-08-26 15:47:21,160][120654] Num frames 11100... [2025-08-26 15:47:21,228][120654] Num frames 11200... [2025-08-26 15:47:21,297][120654] Num frames 11300... [2025-08-26 15:47:21,365][120654] Num frames 11400... [2025-08-26 15:47:21,436][120654] Avg episode rewards: #0: 29.697, true rewards: #0: 12.697 [2025-08-26 15:47:21,437][120654] Avg episode reward: 29.697, avg true_objective: 12.697 [2025-08-26 15:47:21,484][120654] Num frames 11500... [2025-08-26 15:47:21,546][120654] Num frames 11600... [2025-08-26 15:47:21,610][120654] Num frames 11700... [2025-08-26 15:47:21,675][120654] Num frames 11800... [2025-08-26 15:47:21,736][120654] Num frames 11900... [2025-08-26 15:47:21,798][120654] Num frames 12000... [2025-08-26 15:47:21,860][120654] Num frames 12100... [2025-08-26 15:47:21,929][120654] Num frames 12200... [2025-08-26 15:47:22,021][120654] Avg episode rewards: #0: 28.559, true rewards: #0: 12.259 [2025-08-26 15:47:22,022][120654] Avg episode reward: 28.559, avg true_objective: 12.259 [2025-08-26 15:47:39,373][120654] Replay video saved to /home/ubuntu/train_dir/default_experiment/replay.mp4! [2025-08-26 15:47:42,691][120654] The model has been pushed to https://huggingface.co/igzi/rl_course_vizdoom_health_gathering_supreme [2025-08-26 15:47:42,703][120654] Loading existing experiment configuration from /home/ubuntu/train_dir/default_experiment/config.json [2025-08-26 15:47:42,703][120654] Overriding arg 'num_workers' with value 1 passed from command line [2025-08-26 15:47:42,703][120654] Adding new argument 'no_render'=True that is not in the saved config file! [2025-08-26 15:47:42,704][120654] Adding new argument 'save_video'=True that is not in the saved config file! [2025-08-26 15:47:42,704][120654] Adding new argument 'video_frames'=1000000000.0 that is not in the saved config file! [2025-08-26 15:47:42,704][120654] Adding new argument 'video_name'=None that is not in the saved config file! [2025-08-26 15:47:42,705][120654] Adding new argument 'max_num_frames'=100000 that is not in the saved config file! [2025-08-26 15:47:42,705][120654] Adding new argument 'max_num_episodes'=10 that is not in the saved config file! [2025-08-26 15:47:42,705][120654] Adding new argument 'push_to_hub'=True that is not in the saved config file! [2025-08-26 15:47:42,706][120654] Adding new argument 'hf_repository'='igzi/rl_course_vizdoom_health_gathering_supreme' that is not in the saved config file! [2025-08-26 15:47:42,706][120654] Adding new argument 'policy_index'=0 that is not in the saved config file! [2025-08-26 15:47:42,707][120654] Adding new argument 'eval_deterministic'=False that is not in the saved config file! [2025-08-26 15:47:42,707][120654] Adding new argument 'train_script'=None that is not in the saved config file! [2025-08-26 15:47:42,708][120654] Adding new argument 'enjoy_script'=None that is not in the saved config file! [2025-08-26 15:47:42,709][120654] Using frameskip 1 and render_action_repeat=4 for evaluation [2025-08-26 15:47:42,711][120654] RunningMeanStd input shape: (3, 72, 128) [2025-08-26 15:47:42,712][120654] RunningMeanStd input shape: (1,) [2025-08-26 15:47:42,718][120654] ConvEncoder: input_channels=3 [2025-08-26 15:47:42,747][120654] Conv encoder output size: 512 [2025-08-26 15:47:42,748][120654] Policy head output size: 512 [2025-08-26 15:47:42,756][120654] Loading state from checkpoint /home/ubuntu/train_dir/default_experiment/checkpoint_p0/checkpoint_000000980_4014080.pth... [2025-08-26 15:47:43,000][120654] Num frames 100... [2025-08-26 15:47:43,061][120654] Num frames 200... [2025-08-26 15:47:43,124][120654] Num frames 300... [2025-08-26 15:47:43,189][120654] Num frames 400... [2025-08-26 15:47:43,291][120654] Avg episode rewards: #0: 7.800, true rewards: #0: 4.800 [2025-08-26 15:47:43,292][120654] Avg episode reward: 7.800, avg true_objective: 4.800 [2025-08-26 15:47:43,306][120654] Num frames 500... [2025-08-26 15:47:43,371][120654] Num frames 600... [2025-08-26 15:47:43,438][120654] Num frames 700... [2025-08-26 15:47:43,505][120654] Num frames 800... [2025-08-26 15:47:43,567][120654] Num frames 900... [2025-08-26 15:47:43,632][120654] Num frames 1000... [2025-08-26 15:47:43,697][120654] Num frames 1100... [2025-08-26 15:47:43,760][120654] Num frames 1200... [2025-08-26 15:47:43,821][120654] Num frames 1300... [2025-08-26 15:47:43,882][120654] Num frames 1400... [2025-08-26 15:47:43,948][120654] Num frames 1500... [2025-08-26 15:47:44,014][120654] Num frames 1600... [2025-08-26 15:47:44,078][120654] Num frames 1700... [2025-08-26 15:47:44,141][120654] Num frames 1800... [2025-08-26 15:47:44,203][120654] Num frames 1900... [2025-08-26 15:47:44,268][120654] Num frames 2000... [2025-08-26 15:47:44,338][120654] Avg episode rewards: #0: 23.140, true rewards: #0: 10.140 [2025-08-26 15:47:44,339][120654] Avg episode reward: 23.140, avg true_objective: 10.140 [2025-08-26 15:47:44,382][120654] Num frames 2100... [2025-08-26 15:47:44,444][120654] Num frames 2200... [2025-08-26 15:47:44,505][120654] Num frames 2300... [2025-08-26 15:47:44,565][120654] Num frames 2400... [2025-08-26 15:47:44,625][120654] Num frames 2500... [2025-08-26 15:47:44,690][120654] Num frames 2600... [2025-08-26 15:47:44,754][120654] Num frames 2700... [2025-08-26 15:47:44,815][120654] Num frames 2800... [2025-08-26 15:47:44,877][120654] Num frames 2900... [2025-08-26 15:47:44,940][120654] Num frames 3000... [2025-08-26 15:47:44,999][120654] Num frames 3100... [2025-08-26 15:47:45,064][120654] Num frames 3200... [2025-08-26 15:47:45,127][120654] Num frames 3300... [2025-08-26 15:47:45,191][120654] Num frames 3400... [2025-08-26 15:47:45,255][120654] Num frames 3500... [2025-08-26 15:47:45,306][120654] Avg episode rewards: #0: 26.333, true rewards: #0: 11.667 [2025-08-26 15:47:45,307][120654] Avg episode reward: 26.333, avg true_objective: 11.667 [2025-08-26 15:47:45,372][120654] Num frames 3600... [2025-08-26 15:47:45,440][120654] Num frames 3700... [2025-08-26 15:47:45,512][120654] Num frames 3800... [2025-08-26 15:47:45,575][120654] Num frames 3900... [2025-08-26 15:47:45,638][120654] Num frames 4000... [2025-08-26 15:47:45,699][120654] Num frames 4100... [2025-08-26 15:47:45,758][120654] Num frames 4200... [2025-08-26 15:47:45,819][120654] Num frames 4300... [2025-08-26 15:47:45,883][120654] Num frames 4400... [2025-08-26 15:47:45,951][120654] Num frames 4500... [2025-08-26 15:47:46,017][120654] Num frames 4600... [2025-08-26 15:47:46,080][120654] Num frames 4700... [2025-08-26 15:47:46,145][120654] Num frames 4800... [2025-08-26 15:47:46,209][120654] Num frames 4900... [2025-08-26 15:47:46,287][120654] Avg episode rewards: #0: 27.850, true rewards: #0: 12.350 [2025-08-26 15:47:46,288][120654] Avg episode reward: 27.850, avg true_objective: 12.350 [2025-08-26 15:47:46,328][120654] Num frames 5000... [2025-08-26 15:47:46,391][120654] Num frames 5100... [2025-08-26 15:47:46,455][120654] Num frames 5200... [2025-08-26 15:47:46,519][120654] Num frames 5300... [2025-08-26 15:47:46,580][120654] Num frames 5400... [2025-08-26 15:47:46,646][120654] Num frames 5500... [2025-08-26 15:47:46,710][120654] Num frames 5600... [2025-08-26 15:47:46,791][120654] Avg episode rewards: #0: 25.088, true rewards: #0: 11.288 [2025-08-26 15:47:46,792][120654] Avg episode reward: 25.088, avg true_objective: 11.288 [2025-08-26 15:47:46,828][120654] Num frames 5700... [2025-08-26 15:47:46,893][120654] Num frames 5800... [2025-08-26 15:47:46,957][120654] Num frames 5900... [2025-08-26 15:47:47,021][120654] Num frames 6000... [2025-08-26 15:47:47,086][120654] Num frames 6100... [2025-08-26 15:47:47,164][120654] Avg episode rewards: #0: 22.072, true rewards: #0: 10.238 [2025-08-26 15:47:47,165][120654] Avg episode reward: 22.072, avg true_objective: 10.238 [2025-08-26 15:47:47,201][120654] Num frames 6200... [2025-08-26 15:47:47,262][120654] Num frames 6300... [2025-08-26 15:47:47,320][120654] Num frames 6400... [2025-08-26 15:47:47,384][120654] Num frames 6500... [2025-08-26 15:47:47,443][120654] Num frames 6600... [2025-08-26 15:47:47,547][120654] Avg episode rewards: #0: 19.981, true rewards: #0: 9.553 [2025-08-26 15:47:47,548][120654] Avg episode reward: 19.981, avg true_objective: 9.553 [2025-08-26 15:47:47,557][120654] Num frames 6700... [2025-08-26 15:47:47,619][120654] Num frames 6800... [2025-08-26 15:47:47,679][120654] Num frames 6900... [2025-08-26 15:47:47,741][120654] Num frames 7000... [2025-08-26 15:47:47,807][120654] Num frames 7100... [2025-08-26 15:47:47,870][120654] Num frames 7200... [2025-08-26 15:47:47,933][120654] Num frames 7300... [2025-08-26 15:47:47,996][120654] Num frames 7400... [2025-08-26 15:47:48,062][120654] Num frames 7500... [2025-08-26 15:47:48,131][120654] Num frames 7600... [2025-08-26 15:47:48,199][120654] Num frames 7700... [2025-08-26 15:47:48,263][120654] Num frames 7800... [2025-08-26 15:47:48,328][120654] Num frames 7900... [2025-08-26 15:47:48,395][120654] Num frames 8000... [2025-08-26 15:47:48,457][120654] Num frames 8100... [2025-08-26 15:47:48,520][120654] Num frames 8200... [2025-08-26 15:47:48,576][120654] Avg episode rewards: #0: 22.256, true rewards: #0: 10.256 [2025-08-26 15:47:48,577][120654] Avg episode reward: 22.256, avg true_objective: 10.256 [2025-08-26 15:47:48,638][120654] Num frames 8300... [2025-08-26 15:47:48,705][120654] Num frames 8400... [2025-08-26 15:47:48,765][120654] Num frames 8500... [2025-08-26 15:47:48,831][120654] Num frames 8600... [2025-08-26 15:47:48,897][120654] Num frames 8700... [2025-08-26 15:47:48,965][120654] Num frames 8800... [2025-08-26 15:47:49,033][120654] Num frames 8900... [2025-08-26 15:47:49,098][120654] Num frames 9000... [2025-08-26 15:47:49,159][120654] Num frames 9100... [2025-08-26 15:47:49,213][120654] Avg episode rewards: #0: 21.668, true rewards: #0: 10.112 [2025-08-26 15:47:49,213][120654] Avg episode reward: 21.668, avg true_objective: 10.112 [2025-08-26 15:47:49,272][120654] Num frames 9200... [2025-08-26 15:47:49,331][120654] Num frames 9300... [2025-08-26 15:47:49,389][120654] Num frames 9400... [2025-08-26 15:47:49,452][120654] Num frames 9500... [2025-08-26 15:47:49,518][120654] Num frames 9600... [2025-08-26 15:47:49,584][120654] Num frames 9700... [2025-08-26 15:47:49,649][120654] Num frames 9800... [2025-08-26 15:47:49,715][120654] Num frames 9900... [2025-08-26 15:47:49,780][120654] Num frames 10000... [2025-08-26 15:47:49,872][120654] Avg episode rewards: #0: 21.661, true rewards: #0: 10.061 [2025-08-26 15:47:49,873][120654] Avg episode reward: 21.661, avg true_objective: 10.061 [2025-08-26 15:48:04,257][120654] Replay video saved to /home/ubuntu/train_dir/default_experiment/replay.mp4!