Import factor-sweep assets 1-400

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs10:obs8p75:stride4:seed10/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec8f71129e88c590a6a6a873527c722e4384ed0839cd2ba70d0c00a02577712a
+size 16154320

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs10:obs8p75:stride4:seed10/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs10_seed10",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 10,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs10-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs10",
+    "seed10"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 10,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed10/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs10_seed10 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 10 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed10/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs10-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs10 seed10 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 10 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs10_seed10",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 10,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs10-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs10",
+      "seed10"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 10,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed10/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs10/seed_10",
+  "wandb_unique_id": "deadly-corridor-fs10-fixed_l4-s10"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs10:obs8p75:stride4:seed11/checkpoint_p0/checkpoint_000048840_25006080.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a02a80129d36d625629e199825eb93d2d011ad8495745863e27befdde3e737f9
+size 16154320

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs10:obs8p75:stride4:seed11/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs10_seed11",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 11,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs10-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs10",
+    "seed11"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 10,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed11/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs10_seed11 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 11 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed11/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs10-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs10 seed11 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 10 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs10_seed11",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 11,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs10-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs10",
+      "seed11"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 10,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed11/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs10/seed_11",
+  "wandb_unique_id": "deadly-corridor-fs10-fixed_l4-s11"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs10:obs8p75:stride4:seed12/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfa6dbb065291bca85c70a27281fe5213e38def8bdb240d75c867a073d5a6dde
+size 16154320

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs10:obs8p75:stride4:seed12/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs10_seed12",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 12,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs10-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs10",
+    "seed12"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 10,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed12/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs10_seed12 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 12 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed12/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs10-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs10 seed12 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 10 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs10_seed12",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 12,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs10-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs10",
+      "seed12"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 10,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed12/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs10/seed_12",
+  "wandb_unique_id": "deadly-corridor-fs10-fixed_l4-s12"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs10:obs8p75:stride4:seed13/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29605984642c87a4d0e807074e2f0fe9d3e02f557932945925619525eb68f748
+size 16154953

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs10:obs8p75:stride4:seed13/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs10_seed13",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 13,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs10-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs10",
+    "seed13"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 10,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed13/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs10_seed13 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 13 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed13/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs10-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs10 seed13 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 10 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs10_seed13",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 13,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs10-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs10",
+      "seed13"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 10,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed13/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs10/seed_13",
+  "wandb_unique_id": "deadly-corridor-fs10-fixed_l4-s13"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs10:obs8p75:stride4:seed14/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0112a01cc9461ff455982651e797006ab12e22b3c178fadb7501eaca7c4bb586
+size 16154320

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs10:obs8p75:stride4:seed14/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs10_seed14",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 14,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs10-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs10",
+    "seed14"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 10,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed14/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs10_seed14 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 14 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed14/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs10-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs10 seed14 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 10 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs10_seed14",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 14,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs10-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs10",
+      "seed14"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 10,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs10_seed14/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs10/seed_14",
+  "wandb_unique_id": "deadly-corridor-fs10-fixed_l4-s14"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs12:obs8p75:stride4:seed10/checkpoint_p0/checkpoint_000048844_25008128.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a06c6977e4a7da6e4cf705258d92f600215eba8e6c2bef6d39cd1af05ff10f4b
+size 17088208

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs12:obs8p75:stride4:seed10/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs12_seed10",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 10,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs12-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs12",
+    "seed10"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 12,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed10/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs12_seed10 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 10 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed10/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs12-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs12 seed10 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 12 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs12_seed10",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 10,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs12-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs12",
+      "seed10"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 12,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed10/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs12/seed_10",
+  "wandb_unique_id": "deadly-corridor-fs12-fixed_l4-s10"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs12:obs8p75:stride4:seed11/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ddc234f92784fef3ccad3a3f64d8ac8855a0ea299de4da5149badca5e9ec40b
+size 17088208

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs12:obs8p75:stride4:seed11/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs12_seed11",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 11,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs12-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs12",
+    "seed11"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 12,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed11/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs12_seed11 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 11 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed11/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs12-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs12 seed11 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 12 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs12_seed11",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 11,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs12-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs12",
+      "seed11"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 12,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed11/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs12/seed_11",
+  "wandb_unique_id": "deadly-corridor-fs12-fixed_l4-s11"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs12:obs8p75:stride4:seed12/checkpoint_p0/checkpoint_000048844_25008128.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20f363a39b66146808560995595465b7a1e4b9a424d988fe9741ce1cca81931d
+size 17088208

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs12:obs8p75:stride4:seed12/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs12_seed12",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 12,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs12-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs12",
+    "seed12"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 12,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed12/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs12_seed12 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 12 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed12/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs12-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs12 seed12 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 12 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs12_seed12",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 12,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs12-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs12",
+      "seed12"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 12,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed12/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs12/seed_12",
+  "wandb_unique_id": "deadly-corridor-fs12-fixed_l4-s12"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs12:obs8p75:stride4:seed13/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:605f91872196a05cd9debfc0488955a120c17f72b55c449af97aaf1a984eb31e
+size 17088208

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs12:obs8p75:stride4:seed13/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs12_seed13",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 13,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs12-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs12",
+    "seed13"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 12,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed13/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs12_seed13 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 13 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed13/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs12-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs12 seed13 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 12 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs12_seed13",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 13,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs12-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs12",
+      "seed13"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 12,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed13/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs12/seed_13",
+  "wandb_unique_id": "deadly-corridor-fs12-fixed_l4-s13"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs12:obs8p75:stride4:seed14/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4f3e45c09a1e621713278d7f6b2f132b6faad58549a317c281f274cbb468840
+size 17088208

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs12:obs8p75:stride4:seed14/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs12_seed14",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 14,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs12-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs12",
+    "seed14"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 12,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed14/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs12_seed14 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 14 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed14/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs12-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs12 seed14 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 12 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs12_seed14",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 14,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs12-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs12",
+      "seed14"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 12,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs12_seed14/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs12/seed_14",
+  "wandb_unique_id": "deadly-corridor-fs12-fixed_l4-s14"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs14:obs8p75:stride4:seed10/checkpoint_p0/checkpoint_000048844_25008128.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23a03c39dd7bc5ccfbc29cfe37040a0ef68a64148310017313f9cafb7e0558f8
+size 18022096

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs14:obs8p75:stride4:seed10/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs14_seed10",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 10,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs14-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs14",
+    "seed10"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 14,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed10/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs14_seed10 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 10 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed10/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs14-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs14 seed10 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 14 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs14_seed10",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 10,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs14-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs14",
+      "seed10"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 14,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed10/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs14/seed_10",
+  "wandb_unique_id": "deadly-corridor-fs14-fixed_l4-s10"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs14:obs8p75:stride4:seed11/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19dca664c0267c3bb292d5584c6165893ed6210e1de2b171dc96e91b11485520
+size 18022096

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs14:obs8p75:stride4:seed11/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs14_seed11",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 11,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs14-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs14",
+    "seed11"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 14,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed11/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs14_seed11 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 11 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed11/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs14-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs14 seed11 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 14 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs14_seed11",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 11,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs14-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs14",
+      "seed11"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 14,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed11/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs14/seed_11",
+  "wandb_unique_id": "deadly-corridor-fs14-fixed_l4-s11"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs14:obs8p75:stride4:seed12/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e166f307d72aacb0b690aea3283142a7c69a13fe560d1344f048340f2f2851f6
+size 18022096

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs14:obs8p75:stride4:seed12/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs14_seed12",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 12,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs14-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs14",
+    "seed12"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 14,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed12/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs14_seed12 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 12 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed12/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs14-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs14 seed12 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 14 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs14_seed12",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 12,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs14-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs14",
+      "seed12"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 14,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed12/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs14/seed_12",
+  "wandb_unique_id": "deadly-corridor-fs14-fixed_l4-s12"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs14:obs8p75:stride4:seed13/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73dd071c395344b13e4ccaae4ac439ef05012e344d2e3ff0f72b8d2fc233d26a
+size 18022096

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs14:obs8p75:stride4:seed13/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs14_seed13",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 13,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs14-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs14",
+    "seed13"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 14,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed13/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs14_seed13 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 13 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed13/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs14-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs14 seed13 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 14 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs14_seed13",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 13,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs14-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs14",
+      "seed13"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 14,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed13/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs14/seed_13",
+  "wandb_unique_id": "deadly-corridor-fs14-fixed_l4-s13"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs14:obs8p75:stride4:seed14/checkpoint_p0/checkpoint_000048844_25008128.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b454e2f1402c4766658e1b636a6ca0654371b07de284113de92068fa79b30fdb
+size 18022096

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs14:obs8p75:stride4:seed14/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs14_seed14",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 14,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs14-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs14",
+    "seed14"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 14,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed14/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs14_seed14 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 14 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed14/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs14-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs14 seed14 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 14 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs14_seed14",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 14,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs14-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs14",
+      "seed14"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 14,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs14_seed14/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs14/seed_14",
+  "wandb_unique_id": "deadly-corridor-fs14-fixed_l4-s14"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs16:obs8p75:stride4:seed10/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e968bde655391dec6b99d4ed4963a9ebcc54e6055f92e7bf8abf011bc47ade6
+size 18955984

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs16:obs8p75:stride4:seed10/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs16_seed10",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 10,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs16-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs16",
+    "seed10"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 16,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed10/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs16_seed10 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 10 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed10/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs16-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs16 seed10 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 16 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs16_seed10",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 10,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs16-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs16",
+      "seed10"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 16,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed10/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs16/seed_10",
+  "wandb_unique_id": "deadly-corridor-fs16-fixed_l4-s10"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs16:obs8p75:stride4:seed11/checkpoint_p0/checkpoint_000048840_25006080.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:decc9a1e23880ea0c2a7fd259fa0581d21bc47a09fee668e7e94a6da64991749
+size 18955984

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs16:obs8p75:stride4:seed11/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs16_seed11",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 11,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs16-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs16",
+    "seed11"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 16,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed11/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs16_seed11 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 11 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed11/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs16-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs16 seed11 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 16 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs16_seed11",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 11,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs16-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs16",
+      "seed11"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 16,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed11/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs16/seed_11",
+  "wandb_unique_id": "deadly-corridor-fs16-fixed_l4-s11"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs16:obs8p75:stride4:seed12/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2f39f6ec34845d010adacdccd44ebe9edb07125288dab3d3f059a9346fda545
+size 18955984

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs16:obs8p75:stride4:seed12/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs16_seed12",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 12,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs16-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs16",
+    "seed12"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 16,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed12/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs16_seed12 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 12 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed12/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs16-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs16 seed12 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 16 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs16_seed12",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 12,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs16-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs16",
+      "seed12"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 16,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed12/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs16/seed_12",
+  "wandb_unique_id": "deadly-corridor-fs16-fixed_l4-s12"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs16:obs8p75:stride4:seed13/checkpoint_p0/checkpoint_000048844_25008128.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3188561d018e71610e5b622f4fa12a43f3d418e29aeede3f16903fc6d01880d1
+size 18955984

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs16:obs8p75:stride4:seed13/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs16_seed13",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 13,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs16-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs16",
+    "seed13"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 16,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed13/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs16_seed13 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 13 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed13/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs16-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs16 seed13 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 16 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs16_seed13",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 13,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs16-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs16",
+      "seed13"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 16,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed13/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs16/seed_13",
+  "wandb_unique_id": "deadly-corridor-fs16-fixed_l4-s13"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs16:obs8p75:stride4:seed14/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:466efa09fd60e489e3f1040e7e5f4ac0efc364817bb01c285feb136ea9e5f593
+size 18955984

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs16:obs8p75:stride4:seed14/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs16_seed14",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 14,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs16-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs16",
+    "seed14"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 16,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed14/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs16_seed14 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 14 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed14/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs16-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs16 seed14 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 16 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs16_seed14",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 14,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs16-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs16",
+      "seed14"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 16,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs16_seed14/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs16/seed_14",
+  "wandb_unique_id": "deadly-corridor-fs16-fixed_l4-s14"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs1:obs8p75:stride4:seed10/checkpoint_p0/checkpoint_000048840_25006080.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e71a5a21fb853427f796a3ed20881b6f4b779b194c979713a60237b36acd14b6
+size 11951824

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs1:obs8p75:stride4:seed10/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs1_seed10",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 10,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs1-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs1",
+    "seed10"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 1,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed10/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs1_seed10 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 10 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed10/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs1-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs1 seed10 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 1 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs1_seed10",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 10,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs1-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs1",
+      "seed10"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 1,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed10/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs1/seed_10",
+  "wandb_unique_id": "deadly-corridor-fs1-fixed_l4-s10"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs1:obs8p75:stride4:seed11/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c90d9a0315318d2c6292fb6ffb54a45565ae8a4581b22ff0a691401de96b538
+size 11951824

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs1:obs8p75:stride4:seed11/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs1_seed11",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 11,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs1-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs1",
+    "seed11"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 1,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed11/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs1_seed11 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 11 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed11/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs1-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs1 seed11 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 1 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs1_seed11",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 11,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs1-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs1",
+      "seed11"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 1,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed11/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs1/seed_11",
+  "wandb_unique_id": "deadly-corridor-fs1-fixed_l4-s11"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs1:obs8p75:stride4:seed12/checkpoint_p0/checkpoint_000048840_25006080.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee59ee2f63868924a065e88171819ef4dba85cfd83b267728410f53b39336927
+size 11951824

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs1:obs8p75:stride4:seed12/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs1_seed12",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 12,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs1-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs1",
+    "seed12"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 1,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed12/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs1_seed12 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 12 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed12/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs1-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs1 seed12 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 1 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs1_seed12",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 12,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs1-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs1",
+      "seed12"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 1,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed12/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs1/seed_12",
+  "wandb_unique_id": "deadly-corridor-fs1-fixed_l4-s12"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs1:obs8p75:stride4:seed13/checkpoint_p0/checkpoint_000048836_25004032.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e48ed781f044c7a2cadaecdf3bdccedfd1c7674f4b54a2edcde0898f89e2fb2
+size 11951824

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs1:obs8p75:stride4:seed13/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs1_seed13",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 13,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs1-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs1",
+    "seed13"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 1,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed13/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs1_seed13 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 13 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed13/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs1-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs1 seed13 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 1 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs1_seed13",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 13,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs1-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs1",
+      "seed13"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 1,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed13/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs1/seed_13",
+  "wandb_unique_id": "deadly-corridor-fs1-fixed_l4-s13"
+}

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs1:obs8p75:stride4:seed14/checkpoint_p0/checkpoint_000048844_25008128.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d4132fecedb74e8670b2199b1f68c536bf76c9298a6396e1e7df94de53427f4
+size 11951824

latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:frame_stack:fixed_l4:fs1:obs8p75:stride4:seed14/config.json

@@ -0,0 +1,261 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "doom_deadly_corridor",
+  "experiment": "deadly_corridor_frame_stack_fixed_l4_fs1_seed14",
+  "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 14,
+  "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": 32,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
+  "num_batches_per_epoch": 2,
+  "num_epochs": 2,
+  "rollout": 128,
+  "recurrence": 128,
+  "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": 1,
+  "flush_summaries_interval": 30,
+  "stats_avg": 100,
+  "summaries_use_frameskip": true,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 25000000,
+  "train_for_seconds": 10000000000,
+  "save_every_sec": 600,
+  "keep_checkpoints": 5,
+  "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": 1,
+  "env_framestack": 1,
+  "pixel_format": "CHW",
+  "use_record_episode_statistics": false,
+  "with_wandb": true,
+  "wandb_user": null,
+  "wandb_project": "latency-sensitive-bench",
+  "wandb_group": "deadly-corridor-fs1-fixed_l4",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "deadly-corridor",
+    "frame_stack",
+    "fixed",
+    "fixed_l4",
+    "fs1",
+    "seed14"
+  ],
+  "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,
+  "fps": 35,
+  "eval_env_frameskip": 1,
+  "no_render": false,
+  "save_video": false,
+  "video_frames": 1000000000.0,
+  "video_name": null,
+  "max_num_frames": 1000000000.0,
+  "max_num_episodes": 1000000000.0,
+  "push_to_hub": false,
+  "hf_repository": null,
+  "policy_index": 0,
+  "eval_deterministic": true,
+  "train_script": null,
+  "enjoy_script": null,
+  "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,
+  "record_to": null,
+  "mode": "train",
+  "env_fps": 35.0,
+  "obs_fps": 8.75,
+  "latency_type": "fixed",
+  "fixed_latency_ms": 114.28571428571429,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": null,
+  "max_latency_ms": null,
+  "latency_seed": null,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "hold",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 32,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40",
+  "eval_max_steps": 30000,
+  "eval_raw_reward": true,
+  "disable_reward_scaling": true,
+  "frame_stack": 1,
+  "max_episode_steps": 0,
+  "simulator": "cpu",
+  "envpool_batch_size": 1,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed14/episode_metrics.jsonl",
+  "vizdoom_init_max_parallel": 10,
+  "vizdoom_init_lock_timeout_sec": 5.0,
+  "vizdoom_init_lock_max_wait_sec": 60.0,
+  "vizdoom_init_retry_jitter_sec": 0.25,
+  "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_fixed_l4_fs1_seed14 --train_dir results/checkpoints_factor_sweeps/deadly_corridor/context_window --restart_behavior resume --device gpu --seed 14 --episode_metrics_path results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed14/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 114.28571428571429 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 32 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-fs1-fixed_l4 --wandb_job_type sample_factory --wandb_tags factor_sweep deadly-corridor frame_stack fixed fixed_l4 fs1 seed14 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 1 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "doom_deadly_corridor",
+    "experiment": "deadly_corridor_frame_stack_fixed_l4_fs1_seed14",
+    "train_dir": "results/checkpoints_factor_sweeps/deadly_corridor/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 14,
+    "num_policies": 1,
+    "async_rl": true,
+    "worker_num_splits": 2,
+    "num_workers": 32,
+    "num_envs_per_worker": 4,
+    "batch_size": 1024,
+    "num_batches_per_epoch": 2,
+    "num_epochs": 2,
+    "rollout": 128,
+    "recurrence": 128,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "exploration_loss_coeff": 0.001,
+    "value_loss_coeff": 0.5,
+    "exploration_loss": "symmetric_kl",
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 4.0,
+    "learning_rate": 0.0001,
+    "normalize_input": true,
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": true,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "deadly-corridor-fs1-fixed_l4",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "deadly-corridor",
+      "frame_stack",
+      "fixed",
+      "fixed_l4",
+      "fs1",
+      "seed14"
+    ],
+    "eval_deterministic": true,
+    "res_w": 128,
+    "res_h": 72,
+    "wide_aspect_ratio": false,
+    "mode": "train",
+    "env_fps": 35.0,
+    "obs_fps": 8.75,
+    "latency_type": "fixed",
+    "fixed_latency_ms": 114.28571428571429,
+    "add_latency_info": false,
+    "hold_policy": "hold",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 32,
+    "eval_max_steps": 30000,
+    "eval_raw_reward": true,
+    "disable_reward_scaling": true,
+    "frame_stack": 1,
+    "simulator": "cpu",
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/deadly_corridor/context_window/deadly_corridor_frame_stack_fixed_l4_fs1_seed14/episode_metrics.jsonl",
+    "env_frameskip": 1,
+    "eval_env_frameskip": 1
+  },
+  "git_hash": "eb3a2e1efbd2aa03a60d7f44f5e18d8fdd0f5a2d",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "output_dir": "outputs/factor_sweeps/deadly_corridor/context_window/train/frame_stack/fixed_l4/fs1/seed_14",
+  "wandb_unique_id": "deadly-corridor-fs1-fixed_l4-s14"
+}