Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+replay.mp4 filter=lfs diff=lfs merge=lfs -text

.summary/0/events.out.tfevents.1755486737.arun-GF63-Thin-11UC

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14d868681c221a4ee768507fe4b6f2d503c4e76305d731644d7c1b5d3bfc5efc
+size 40

.summary/0/events.out.tfevents.1755486808.arun-GF63-Thin-11UC

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec39fd10fc248820d691382018fb0733077a2ca264b2b9bdb3b23caa388ec854
+size 40

.summary/0/events.out.tfevents.1755487116.arun-GF63-Thin-11UC

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf813355ccbb06d3968f89788b14d23332cc7bb9f3cc173fe111a799d5b0ee7d
+size 40

.summary/0/events.out.tfevents.1755487117.arun-GF63-Thin-11UC

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1588d63aad315d0585c66d533a9f8a5c25e2b46b26eac89633903ad93d6e2a51
+size 40

.summary/0/events.out.tfevents.1755487241.arun-GF63-Thin-11UC

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92c48427498b7193be1d80dbdca7650bdf557711785b023d2f12a39ddddc8ea6
+size 40

.summary/0/events.out.tfevents.1755487432.arun-GF63-Thin-11UC

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f0c64c3bb384fd40b06a80be8af1f85d9c55eaf9d30c56f41fb9b10c4adec72e
+size 40

.summary/0/events.out.tfevents.1755487646.arun-GF63-Thin-11UC

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:416b9ca1ec94499912dfe9176ef98cb33840b382a2bac76f8e5fc408d9bf73f7
+size 40

.summary/0/events.out.tfevents.1755487705.arun-GF63-Thin-11UC

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29a0234b1d590c9f5ce96050d0c112aba0678d1591a23c1079406becb42ef08b
+size 40

.summary/0/events.out.tfevents.1755487793.arun-GF63-Thin-11UC

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c09b64e7280b4f7a7293139b35ca65ac64ad775c56ace06e327f8129df47ed80
+size 40

.summary/0/events.out.tfevents.1755487854.arun-GF63-Thin-11UC

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a833d523a13168d986b3fd759b4d2e3fbd08aa62d23da7db17c6b4b0f1a13b1
+size 4198

README.md

@@ -0,0 +1,56 @@
+---
+library_name: sample-factory
+tags:
+- deep-reinforcement-learning
+- reinforcement-learning
+- sample-factory
+model-index:
+- name: APPO
+  results:
+  - task:
+      type: reinforcement-learning
+      name: reinforcement-learning
+    dataset:
+      name: doom_health_gathering_supreme
+      type: doom_health_gathering_supreme
+    metrics:
+    - type: mean_reward
+      value: 4.03 +/- 0.42
+      name: mean_reward
+      verified: false
+---
+
+A(n) **APPO** model trained on the **doom_health_gathering_supreme** environment.
+
+This model was trained using Sample-Factory 2.0: https://github.com/alex-petrenko/sample-factory.
+Documentation for how to use Sample-Factory can be found at https://www.samplefactory.dev/
+
+
+## Downloading the model
+
+After installing Sample-Factory, download the model with:
+```
+python -m sample_factory.huggingface.load_from_hub -r ArunKr/rl_course_vizdoom_health_gathering_supreme
+```
+
+    
+## Using the model
+
+To run the model after download, use the `enjoy` script corresponding to this environment:
+```
+python -m <path.to.enjoy.module> --algo=APPO --env=doom_health_gathering_supreme --train_dir=./train_dir --experiment=rl_course_vizdoom_health_gathering_supreme
+```
+
+
+You can also upload models to the Hugging Face Hub using the same script with the `--push_to_hub` flag.
+See https://www.samplefactory.dev/10-huggingface/huggingface/ for more details
+    
+## Training with this model
+
+To continue training with this model, use the `train` script corresponding to this environment:
+```
+python -m <path.to.train.module> --algo=APPO --env=doom_health_gathering_supreme --train_dir=./train_dir --experiment=rl_course_vizdoom_health_gathering_supreme --restart_behavior=resume --train_for_env_steps=10000000000
+```
+
+Note, you may have to adjust `--train_for_env_steps` to a suitably high number as the experiment will resume at the number of steps it concluded at.
+    

checkpoint_p0/checkpoint_000000003_12288.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43cab3c26df6ba9434c82505e7f4cc4573afc0e829d7006900b8f693267d2b19
+size 34929457

config.json

@@ -0,0 +1,142 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_health_gathering_supreme",
+  "experiment": "default_experiment",
+  "train_dir": "/home/arun/workspace-2/hf_drl_course/train_dir",
+  "restart_behavior": "resume",
+  "device": "cpu",
+  "seed": null,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": false,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 2,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 1000,
+  "num_workers": 4,
+  "num_envs_per_worker": 8,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 1,
+  "num_epochs": 1,
+  "rollout": 32,
+  "recurrence": 32,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.001,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "symmetric_kl",
+  "gae_lambda": 0.95,
+  "ppo_clip_ratio": 0.1,
+  "ppo_clip_value": 0.2,
+  "with_vtrace": false,
+  "vtrace_rho": 1.0,
+  "vtrace_c": 1.0,
+  "optimizer": "adam",
+  "adam_eps": 1e-06,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 4.0,
+  "learning_rate": 0.0001,
+  "lr_schedule": "constant",
+  "lr_schedule_kl_threshold": 0.008,
+  "lr_adaptive_min": 1e-06,
+  "lr_adaptive_max": 0.01,
+  "obs_subtract_mean": 0.0,
+  "obs_scale": 255.0,
+  "normalize_input": true,
+  "normalize_input_keys": null,
+  "decorrelate_experience_max_seconds": 0,
+  "decorrelate_envs_on_one_worker": true,
+  "actor_worker_gpus": [],
+  "set_workers_cpu_affinity": true,
+  "force_envs_single_thread": false,
+  "default_niceness": 0,
+  "log_to_file": true,
+  "experiment_summaries_interval": 10,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 4000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 120,
+  "keep_checkpoints": 2,
+  "load_checkpoint_kind": "latest",
+  "save_milestones_sec": -1,
+  "save_best_every_sec": 5,
+  "save_best_metric": "reward",
+  "save_best_after": 100000,
+  "benchmark": false,
+  "encoder_mlp_layers": [
+    512,
+    512
+  ],
+  "encoder_conv_architecture": "convnet_simple",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": true,
+  "rnn_size": 512,
+  "rnn_type": "gru",
+  "rnn_num_layers": 1,
+  "decoder_mlp_layers": [],
+  "nonlinearity": "elu",
+  "policy_initialization": "orthogonal",
+  "policy_init_gain": 1.0,
+  "actor_critic_share_weights": true,
+  "adaptive_stddev": true,
+  "continuous_tanh_scale": 0.0,
+  "initial_stddev": 1.0,
+  "use_env_info_cache": false,
+  "env_gpu_actions": false,
+  "env_gpu_observations": true,
+  "env_frameskip": 4,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": false,
+  "wandb_user": null,
+  "wandb_project": "sample_factory",
+  "wandb_group": null,
+  "wandb_job_type": "SF",
+  "wandb_tags": [],
+  "with_pbt": false,
+  "pbt_mix_policies_in_one_env": true,
+  "pbt_period_env_steps": 5000000,
+  "pbt_start_mutation": 20000000,
+  "pbt_replace_fraction": 0.3,
+  "pbt_mutation_rate": 0.15,
+  "pbt_replace_reward_gap": 0.1,
+  "pbt_replace_reward_gap_absolute": 1e-06,
+  "pbt_optimize_gamma": false,
+  "pbt_target_objective": "true_objective",
+  "pbt_perturb_min": 1.1,
+  "pbt_perturb_max": 1.5,
+  "num_agents": -1,
+  "num_humans": 0,
+  "num_bots": -1,
+  "start_bot_difficulty": null,
+  "timelimit": null,
+  "res_w": 128,
+  "res_h": 72,
+  "wide_aspect_ratio": false,
+  "eval_env_frameskip": 1,
+  "fps": 35,
+  "command_line": "--env=doom_health_gathering_supreme --num_workers=8 --num_envs_per_worker=8 --train_for_env_steps=4000",
+  "cli_args": {
+    "env": "doom_health_gathering_supreme",
+    "num_workers": 8,
+    "num_envs_per_worker": 8,
+    "train_for_env_steps": 4000
+  },
+  "git_hash": "unknown",
+  "git_repo_name": "not a git repository"
+}

replay.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0aef9adc9d5e44740b2c9bf3ad7a4569e7d8e975699ca960533a0c13d78cd5ec
+size 5831874

sf_log.txt

@@ -0,0 +1,424 @@
+[2025-08-18 08:58:29,873][791384] Saving configuration to /home/arun/workspace-2/hf_drl_course/train_dir/default_experiment/config.json...
+[2025-08-18 08:58:29,874][791384] Rollout worker 0 uses device cpu
+[2025-08-18 08:58:29,874][791384] Rollout worker 1 uses device cpu
+[2025-08-18 08:58:29,874][791384] Rollout worker 2 uses device cpu
+[2025-08-18 08:58:29,874][791384] Rollout worker 3 uses device cpu
+[2025-08-18 08:58:29,874][791384] Rollout worker 4 uses device cpu
+[2025-08-18 08:58:29,875][791384] Rollout worker 5 uses device cpu
+[2025-08-18 08:58:29,875][791384] Rollout worker 6 uses device cpu
+[2025-08-18 08:58:29,875][791384] Rollout worker 7 uses device cpu
+[2025-08-18 08:58:29,956][791384] Using GPUs [0] for process 0 (actually maps to GPUs [0])
+[2025-08-18 08:58:29,956][791384] InferenceWorker_p0-w0: min num requests: 2
+[2025-08-18 08:58:30,033][791384] Starting all processes...
+[2025-08-18 08:58:30,034][791384] Starting process learner_proc0
+[2025-08-18 08:58:31,879][791384] Starting all processes...
+[2025-08-18 08:58:31,884][791730] Using GPUs [0] for process 0 (actually maps to GPUs [0])
+[2025-08-18 08:58:31,884][791730] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for learning process 0
+[2025-08-18 08:58:31,886][791384] Starting process inference_proc0-0
+[2025-08-18 08:58:31,887][791384] Starting process rollout_proc0
+[2025-08-18 08:58:31,887][791384] Starting process rollout_proc1
+[2025-08-18 08:58:31,887][791384] Starting process rollout_proc2
+[2025-08-18 08:58:31,890][791384] Starting process rollout_proc4
+[2025-08-18 08:58:31,887][791384] Starting process rollout_proc3
+[2025-08-18 08:58:31,893][791384] Starting process rollout_proc5
+[2025-08-18 08:58:31,896][791384] Starting process rollout_proc6
+[2025-08-18 08:58:31,898][791730] Num visible devices: 1
+[2025-08-18 08:58:31,899][791730] Starting seed is not provided
+[2025-08-18 08:58:31,899][791730] Using GPUs [0] for process 0 (actually maps to GPUs [0])
+[2025-08-18 08:58:31,899][791730] Initializing actor-critic model on device cuda:0
+[2025-08-18 08:58:31,899][791730] RunningMeanStd input shape: (3, 72, 128)
+[2025-08-18 08:58:31,896][791384] Starting process rollout_proc7
+[2025-08-18 08:58:31,900][791730] RunningMeanStd input shape: (1,)
+[2025-08-18 08:58:31,920][791730] ConvEncoder: input_channels=3
+[2025-08-18 08:58:32,122][791730] Conv encoder output size: 512
+[2025-08-18 08:58:32,123][791730] Policy head output size: 512
+[2025-08-18 08:58:32,144][791730] Created Actor Critic model with architecture:
+[2025-08-18 08:58:32,145][791730] 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)
+  )
+)
+[2025-08-18 08:58:32,159][791730] EvtLoop [learner_proc0_evt_loop, process=learner_proc0] unhandled exception in slot='init' connected to emitter=Emitter(object_id='Runner_EvtLoop', signal_name='start'), args=()
+Traceback (most recent call last):
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/signal_slot/signal_slot.py", line 355, in _process_signal
+    slot_callable(*args)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/sample_factory/algo/learning/learner_worker.py", line 139, in init
+    init_model_data = self.learner.init()
+                      ^^^^^^^^^^^^^^^^^^^
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/sample_factory/algo/learning/learner.py", line 215, in init
+    self.actor_critic.model_to_device(self.device)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/sample_factory/model/actor_critic.py", line 60, in model_to_device
+    module.to(device)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1369, in to
+    return self._apply(convert)
+           ^^^^^^^^^^^^^^^^^^^^
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 928, in _apply
+    module._apply(fn)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 928, in _apply
+    module._apply(fn)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 928, in _apply
+    module._apply(fn)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1020, in _apply
+    self._buffers[key] = fn(buf)
+                         ^^^^^^^
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1355, in convert
+    return t.to(
+           ^^^^^
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/cuda/__init__.py", line 412, in _lazy_init
+    torch._C._cuda_init()
+RuntimeError: CUDA unknown error - this may be due to an incorrectly set up environment, e.g. changing env variable CUDA_VISIBLE_DEVICES after program start. Setting the available devices to be zero.
+[2025-08-18 08:58:32,162][791730] Unhandled exception CUDA unknown error - this may be due to an incorrectly set up environment, e.g. changing env variable CUDA_VISIBLE_DEVICES after program start. Setting the available devices to be zero. in evt loop learner_proc0_evt_loop
+[2025-08-18 08:58:34,355][791876] Worker 7 uses CPU cores [14, 15]
+[2025-08-18 08:58:34,369][791877] Worker 2 uses CPU cores [4, 5]
+[2025-08-18 08:58:34,371][791871] Worker 1 uses CPU cores [2, 3]
+[2025-08-18 08:58:34,502][791874] Worker 3 uses CPU cores [6, 7]
+[2025-08-18 08:58:34,538][791878] Worker 6 uses CPU cores [12, 13]
+[2025-08-18 08:58:34,569][791869] Worker 0 uses CPU cores [0, 1]
+[2025-08-18 08:58:34,577][791875] Worker 5 uses CPU cores [10, 11]
+[2025-08-18 08:58:34,695][791870] Using GPUs [0] for process 0 (actually maps to GPUs [0])
+[2025-08-18 08:58:34,695][791870] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for inference process 0
+[2025-08-18 08:58:34,709][791870] Num visible devices: 1
+[2025-08-18 08:58:34,758][791873] Worker 4 uses CPU cores [8, 9]
+[2025-08-18 08:58:49,945][791384] Heartbeat connected on Batcher_0
+[2025-08-18 08:58:49,956][791384] Heartbeat connected on InferenceWorker_p0-w0
+[2025-08-18 08:58:49,960][791384] Heartbeat connected on RolloutWorker_w0
+[2025-08-18 08:58:49,964][791384] Heartbeat connected on RolloutWorker_w1
+[2025-08-18 08:58:49,967][791384] Heartbeat connected on RolloutWorker_w2
+[2025-08-18 08:58:49,971][791384] Heartbeat connected on RolloutWorker_w3
+[2025-08-18 08:58:49,973][791384] Heartbeat connected on RolloutWorker_w4
+[2025-08-18 08:58:49,976][791384] Heartbeat connected on RolloutWorker_w5
+[2025-08-18 08:58:50,030][791384] Heartbeat connected on RolloutWorker_w6
+[2025-08-18 08:58:50,033][791384] Heartbeat connected on RolloutWorker_w7
+[2025-08-18 08:59:56,330][796369] Saving configuration to /home/arun/workspace-2/hf_drl_course/train_dir/default_experiment/config.json...
+[2025-08-18 08:59:56,330][796369] Rollout worker 0 uses device cpu
+[2025-08-18 08:59:56,330][796369] Rollout worker 1 uses device cpu
+[2025-08-18 08:59:56,331][796369] Rollout worker 2 uses device cpu
+[2025-08-18 08:59:56,331][796369] Rollout worker 3 uses device cpu
+[2025-08-18 08:59:56,361][796369] Using GPUs [0] for process 0 (actually maps to GPUs [0])
+[2025-08-18 08:59:56,361][796369] InferenceWorker_p0-w0: min num requests: 1
+[2025-08-18 08:59:56,371][796369] Starting all processes...
+[2025-08-18 08:59:56,371][796369] Starting process learner_proc0
+[2025-08-18 08:59:57,824][796369] Starting all processes...
+[2025-08-18 08:59:57,827][796369] Starting process inference_proc0-0
+[2025-08-18 08:59:57,829][796703] Using GPUs [0] for process 0 (actually maps to GPUs [0])
+[2025-08-18 08:59:57,829][796703] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for learning process 0
+[2025-08-18 08:59:57,828][796369] Starting process rollout_proc0
+[2025-08-18 08:59:57,828][796369] Starting process rollout_proc1
+[2025-08-18 08:59:57,828][796369] Starting process rollout_proc2
+[2025-08-18 08:59:57,830][796369] Starting process rollout_proc3
+[2025-08-18 08:59:57,842][796703] Num visible devices: 1
+[2025-08-18 08:59:57,843][796703] Starting seed is not provided
+[2025-08-18 08:59:57,843][796703] Using GPUs [0] for process 0 (actually maps to GPUs [0])
+[2025-08-18 08:59:57,843][796703] Initializing actor-critic model on device cuda:0
+[2025-08-18 08:59:57,844][796703] RunningMeanStd input shape: (3, 72, 128)
+[2025-08-18 08:59:57,845][796703] RunningMeanStd input shape: (1,)
+[2025-08-18 08:59:57,860][796703] ConvEncoder: input_channels=3
+[2025-08-18 08:59:58,022][796703] Conv encoder output size: 512
+[2025-08-18 08:59:58,022][796703] Policy head output size: 512
+[2025-08-18 08:59:58,038][796703] Created Actor Critic model with architecture:
+[2025-08-18 08:59:58,038][796703] 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)
+  )
+)
+[2025-08-18 08:59:58,043][796703] EvtLoop [learner_proc0_evt_loop, process=learner_proc0] unhandled exception in slot='init' connected to emitter=Emitter(object_id='Runner_EvtLoop', signal_name='start'), args=()
+Traceback (most recent call last):
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/signal_slot/signal_slot.py", line 355, in _process_signal
+    slot_callable(*args)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/sample_factory/algo/learning/learner_worker.py", line 139, in init
+    init_model_data = self.learner.init()
+                      ^^^^^^^^^^^^^^^^^^^
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/sample_factory/algo/learning/learner.py", line 215, in init
+    self.actor_critic.model_to_device(self.device)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/sample_factory/model/actor_critic.py", line 60, in model_to_device
+    module.to(device)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1369, in to
+    return self._apply(convert)
+           ^^^^^^^^^^^^^^^^^^^^
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 928, in _apply
+    module._apply(fn)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 928, in _apply
+    module._apply(fn)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 928, in _apply
+    module._apply(fn)
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1020, in _apply
+    self._buffers[key] = fn(buf)
+                         ^^^^^^^
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1355, in convert
+    return t.to(
+           ^^^^^
+  File "/home/arun/workspace-2/hf_drl_course/env/lib/python3.11/site-packages/torch/cuda/__init__.py", line 412, in _lazy_init
+    torch._C._cuda_init()
+RuntimeError: CUDA unknown error - this may be due to an incorrectly set up environment, e.g. changing env variable CUDA_VISIBLE_DEVICES after program start. Setting the available devices to be zero.
+[2025-08-18 08:59:58,046][796703] Unhandled exception CUDA unknown error - this may be due to an incorrectly set up environment, e.g. changing env variable CUDA_VISIBLE_DEVICES after program start. Setting the available devices to be zero. in evt loop learner_proc0_evt_loop
+[2025-08-18 08:59:59,736][796837] Worker 3 uses CPU cores [12, 13, 14, 15]
+[2025-08-18 08:59:59,738][796834] Worker 0 uses CPU cores [0, 1, 2, 3]
+[2025-08-18 08:59:59,742][796835] Worker 1 uses CPU cores [4, 5, 6, 7]
+[2025-08-18 08:59:59,760][796836] Worker 2 uses CPU cores [8, 9, 10, 11]
+[2025-08-18 08:59:59,778][796833] Using GPUs [0] for process 0 (actually maps to GPUs [0])
+[2025-08-18 08:59:59,778][796833] Set environment var CUDA_VISIBLE_DEVICES to '0' (GPU indices [0]) for inference process 0
+[2025-08-18 08:59:59,791][796833] Num visible devices: 1
+[2025-08-18 09:00:16,354][796369] Heartbeat connected on Batcher_0
+[2025-08-18 09:00:16,362][796369] Heartbeat connected on InferenceWorker_p0-w0
+[2025-08-18 09:00:16,364][796369] Heartbeat connected on RolloutWorker_w0
+[2025-08-18 09:00:16,367][796369] Heartbeat connected on RolloutWorker_w1
+[2025-08-18 09:00:16,369][796369] Heartbeat connected on RolloutWorker_w2
+[2025-08-18 09:00:16,371][796369] Heartbeat connected on RolloutWorker_w3
+[2025-08-18 09:00:56,746][800154] Saving configuration to /home/arun/workspace-2/hf_drl_course/train_dir/default_experiment/config.json...
+[2025-08-18 09:00:56,747][800154] Rollout worker 0 uses device cpu
+[2025-08-18 09:00:56,747][800154] Rollout worker 1 uses device cpu
+[2025-08-18 09:00:56,748][800154] Rollout worker 2 uses device cpu
+[2025-08-18 09:00:56,748][800154] Rollout worker 3 uses device cpu
+[2025-08-18 09:00:56,822][800154] InferenceWorker_p0-w0: min num requests: 1
+[2025-08-18 09:00:56,837][800154] Starting all processes...
+[2025-08-18 09:00:56,837][800154] Starting process learner_proc0
+[2025-08-18 09:00:58,499][800154] Starting all processes...
+[2025-08-18 09:00:58,503][800154] Starting process inference_proc0-0
+[2025-08-18 09:00:58,503][800154] Starting process rollout_proc0
+[2025-08-18 09:00:58,504][800491] Starting seed is not provided
+[2025-08-18 09:00:58,504][800491] Initializing actor-critic model on device cpu
+[2025-08-18 09:00:58,505][800491] RunningMeanStd input shape: (3, 72, 128)
+[2025-08-18 09:00:58,506][800491] RunningMeanStd input shape: (1,)
+[2025-08-18 09:00:58,503][800154] Starting process rollout_proc1
+[2025-08-18 09:00:58,503][800154] Starting process rollout_proc2
+[2025-08-18 09:00:58,503][800154] Starting process rollout_proc3
+[2025-08-18 09:00:58,519][800491] ConvEncoder: input_channels=3
+[2025-08-18 09:00:58,645][800491] Conv encoder output size: 512
+[2025-08-18 09:00:58,646][800491] Policy head output size: 512
+[2025-08-18 09:00:58,664][800491] Created Actor Critic model with architecture:
+[2025-08-18 09:00:58,665][800491] 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)
+  )
+)
+[2025-08-18 09:00:59,005][800491] Using optimizer <class 'torch.optim.adam.Adam'>
+[2025-08-18 09:01:00,705][800578] Worker 1 uses CPU cores [4, 5, 6, 7]
+[2025-08-18 09:01:00,729][800577] Worker 3 uses CPU cores [12, 13, 14, 15]
+[2025-08-18 09:01:00,753][800579] Worker 2 uses CPU cores [8, 9, 10, 11]
+[2025-08-18 09:01:00,885][800575] Worker 0 uses CPU cores [0, 1, 2, 3]
+[2025-08-18 09:01:03,490][800491] No checkpoints found
+[2025-08-18 09:01:03,490][800491] Did not load from checkpoint, starting from scratch!
+[2025-08-18 09:01:03,490][800491] Initialized policy 0 weights for model version 0
+[2025-08-18 09:01:03,492][800491] LearnerWorker_p0 finished initialization!
+[2025-08-18 09:01:03,493][800576] RunningMeanStd input shape: (3, 72, 128)
+[2025-08-18 09:01:03,494][800576] RunningMeanStd input shape: (1,)
+[2025-08-18 09:01:03,506][800576] ConvEncoder: input_channels=3
+[2025-08-18 09:01:03,658][800576] Conv encoder output size: 512
+[2025-08-18 09:01:03,658][800576] Policy head output size: 512
+[2025-08-18 09:01:03,674][800154] Inference worker 0-0 is ready!
+[2025-08-18 09:01:03,675][800154] All inference workers are ready! Signal rollout workers to start!
+[2025-08-18 09:01:03,704][800575] Doom resolution: 160x120, resize resolution: (128, 72)
+[2025-08-18 09:01:03,708][800578] Doom resolution: 160x120, resize resolution: (128, 72)
+[2025-08-18 09:01:03,708][800577] Doom resolution: 160x120, resize resolution: (128, 72)
+[2025-08-18 09:01:03,708][800579] Doom resolution: 160x120, resize resolution: (128, 72)
+[2025-08-18 09:01:04,082][800575] Decorrelating experience for 0 frames...
+[2025-08-18 09:01:04,106][800578] Decorrelating experience for 0 frames...
+[2025-08-18 09:01:04,108][800577] Decorrelating experience for 0 frames...
+[2025-08-18 09:01:04,355][800575] Decorrelating experience for 32 frames...
+[2025-08-18 09:01:04,364][800579] Decorrelating experience for 0 frames...
+[2025-08-18 09:01:04,381][800578] Decorrelating experience for 32 frames...
+[2025-08-18 09:01:04,383][800577] Decorrelating experience for 32 frames...
+[2025-08-18 09:01:04,536][800154] Fps is (10 sec: nan, 60 sec: nan, 300 sec: nan). Total num frames: 0. Throughput: 0: nan. Samples: 0. Policy #0 lag: (min: -1.0, avg: -1.0, max: -1.0)
+[2025-08-18 09:01:04,641][800579] Decorrelating experience for 32 frames...
+[2025-08-18 09:01:04,680][800578] Decorrelating experience for 64 frames...
+[2025-08-18 09:01:04,681][800577] Decorrelating experience for 64 frames...
+[2025-08-18 09:01:04,985][800577] Decorrelating experience for 96 frames...
+[2025-08-18 09:01:04,991][800579] Decorrelating experience for 64 frames...
+[2025-08-18 09:01:05,011][800575] Decorrelating experience for 64 frames...
+[2025-08-18 09:01:05,289][800578] Decorrelating experience for 96 frames...
+[2025-08-18 09:01:05,327][800575] Decorrelating experience for 96 frames...
+[2025-08-18 09:01:05,413][800577] Decorrelating experience for 128 frames...
+[2025-08-18 09:01:05,610][800579] Decorrelating experience for 96 frames...
+[2025-08-18 09:01:05,739][800578] Decorrelating experience for 128 frames...
+[2025-08-18 09:01:05,778][800577] Decorrelating experience for 160 frames...
+[2025-08-18 09:01:05,797][800575] Decorrelating experience for 128 frames...
+[2025-08-18 09:01:06,085][800579] Decorrelating experience for 128 frames...
+[2025-08-18 09:01:06,104][800578] Decorrelating experience for 160 frames...
+[2025-08-18 09:01:06,159][800577] Decorrelating experience for 192 frames...
+[2025-08-18 09:01:06,432][800575] Decorrelating experience for 160 frames...
+[2025-08-18 09:01:06,463][800579] Decorrelating experience for 160 frames...
+[2025-08-18 09:01:06,511][800578] Decorrelating experience for 192 frames...
+[2025-08-18 09:01:06,782][800577] Decorrelating experience for 224 frames...
+[2025-08-18 09:01:06,895][800575] Decorrelating experience for 192 frames...
+[2025-08-18 09:01:06,911][800578] Decorrelating experience for 224 frames...
+[2025-08-18 09:01:07,125][800579] Decorrelating experience for 192 frames...
+[2025-08-18 09:01:07,344][800575] Decorrelating experience for 224 frames...
+[2025-08-18 09:01:07,542][800579] Decorrelating experience for 224 frames...
+[2025-08-18 09:01:08,866][800491] Signal inference workers to stop experience collection...
+[2025-08-18 09:01:08,883][800576] InferenceWorker_p0-w0: stopping experience collection
+[2025-08-18 09:01:09,536][800154] Fps is (10 sec: 0.0, 60 sec: 0.0, 300 sec: 0.0). Total num frames: 0. Throughput: 0: 0.0. Samples: 0. Policy #0 lag: (min: -1.0, avg: -1.0, max: -1.0)
+[2025-08-18 09:01:09,536][800154] Avg episode reward: [(0, '2.199')]
+[2025-08-18 09:01:10,148][800491] Signal inference workers to resume experience collection...
+[2025-08-18 09:01:10,148][800576] InferenceWorker_p0-w0: resuming experience collection
+[2025-08-18 09:01:12,186][800154] Component Batcher_0 stopped!
+[2025-08-18 09:01:12,186][800491] Stopping Batcher_0...
+[2025-08-18 09:01:12,186][800491] Loop batcher_evt_loop terminating...
+[2025-08-18 09:01:12,200][800576] Weights refcount: 2 0
+[2025-08-18 09:01:12,202][800576] Stopping InferenceWorker_p0-w0...
+[2025-08-18 09:01:12,202][800154] Component InferenceWorker_p0-w0 stopped!
+[2025-08-18 09:01:12,202][800576] Loop inference_proc0-0_evt_loop terminating...
+[2025-08-18 09:01:12,254][800577] Stopping RolloutWorker_w3...
+[2025-08-18 09:01:12,255][800577] Loop rollout_proc3_evt_loop terminating...
+[2025-08-18 09:01:12,254][800154] Component RolloutWorker_w3 stopped!
+[2025-08-18 09:01:12,260][800578] Stopping RolloutWorker_w1...
+[2025-08-18 09:01:12,260][800154] Component RolloutWorker_w1 stopped!
+[2025-08-18 09:01:12,261][800578] Loop rollout_proc1_evt_loop terminating...
+[2025-08-18 09:01:12,274][800154] Component RolloutWorker_w0 stopped!
+[2025-08-18 09:01:12,273][800575] Stopping RolloutWorker_w0...
+[2025-08-18 09:01:12,274][800575] Loop rollout_proc0_evt_loop terminating...
+[2025-08-18 09:01:12,294][800579] Stopping RolloutWorker_w2...
+[2025-08-18 09:01:12,295][800579] Loop rollout_proc2_evt_loop terminating...
+[2025-08-18 09:01:12,296][800154] Component RolloutWorker_w2 stopped!
+[2025-08-18 09:01:14,231][800491] Saving /home/arun/workspace-2/hf_drl_course/train_dir/default_experiment/checkpoint_p0/checkpoint_000000003_12288.pth...
+[2025-08-18 09:01:14,268][800491] Saving /home/arun/workspace-2/hf_drl_course/train_dir/default_experiment/checkpoint_p0/checkpoint_000000003_12288.pth...
+[2025-08-18 09:01:14,325][800491] Stopping LearnerWorker_p0...
+[2025-08-18 09:01:14,325][800154] Component LearnerWorker_p0 stopped!
+[2025-08-18 09:01:14,326][800491] Loop learner_proc0_evt_loop terminating...
+[2025-08-18 09:01:14,326][800154] Waiting for process learner_proc0 to stop...
+[2025-08-18 09:01:14,933][800154] Waiting for process inference_proc0-0 to join...
+[2025-08-18 09:01:14,934][800154] Waiting for process rollout_proc0 to join...
+[2025-08-18 09:01:14,934][800154] Waiting for process rollout_proc1 to join...
+[2025-08-18 09:01:14,934][800154] Waiting for process rollout_proc2 to join...
+[2025-08-18 09:01:14,934][800154] Waiting for process rollout_proc3 to join...
+[2025-08-18 09:01:14,934][800154] Batcher 0 profile tree view:
+batching: 0.0248, releasing_batches: 0.0007
+[2025-08-18 09:01:14,935][800154] InferenceWorker_p0-w0 profile tree view:
+wait_policy: 0.0000
+  wait_policy_total: 3.2258
+update_model: 0.3717
+  weight_update: 0.3635
+one_step: 0.0147
+  handle_policy_step: 2.6186
+    deserialize: 0.0311, stack: 0.0046, obs_to_device_normalize: 0.2297, forward: 2.2344, send_messages: 0.0343
+    prepare_outputs: 0.0459
+      to_cpu: 0.0052
+[2025-08-18 09:01:14,935][800154] Learner 0 profile tree view:
+misc: 0.0000, prepare_batch: 1.5144
+train: 4.5077
+  epoch_init: 0.0000, minibatch_init: 0.0000, losses_postprocess: 0.0005, kl_divergence: 0.0007, after_optimizer: 0.0085
+  calculate_losses: 1.7299
+    losses_init: 0.0000, forward_head: 1.5384, bptt_initial: 0.0047, tail: 0.0034, advantages_returns: 0.0008, losses: 0.0025
+    bptt: 0.1795
+      bptt_forward_core: 0.1785
+  update: 2.7669
+    clip: 0.0073
+[2025-08-18 09:01:14,935][800154] RolloutWorker_w0 profile tree view:
+wait_for_trajectories: 0.0011, enqueue_policy_requests: 0.0424, env_step: 0.6034, overhead: 0.0336, complete_rollouts: 0.0008
+save_policy_outputs: 0.0517
+  split_output_tensors: 0.0168
+[2025-08-18 09:01:14,935][800154] RolloutWorker_w3 profile tree view:
+wait_for_trajectories: 0.0011, enqueue_policy_requests: 0.0616, env_step: 0.8382, overhead: 0.0458, complete_rollouts: 0.0018
+save_policy_outputs: 0.0694
+  split_output_tensors: 0.0223
+[2025-08-18 09:01:14,935][800154] Loop Runner_EvtLoop terminating...
+[2025-08-18 09:01:14,936][800154] Runner profile tree view:
+main_loop: 18.0988
+[2025-08-18 09:01:14,936][800154] Collected {0: 12288}, FPS: 678.9