Import factor-sweep assets 401-527

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs6:obs30:stride1:seed11/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs6_seed11",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 11,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs6-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs6",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 6,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 11,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed11/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs6_seed11 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 11 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed11/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 11 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs6-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs6 seed11 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 6 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs6_seed11",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 11,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs6-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs6",
+      "seed11"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 6,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 11,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed11/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs6/seed_11",
+  "wandb_unique_id": "flappy-fs6-uniform_u1_3-s11"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs6:obs30:stride1:seed12/checkpoint_p0/checkpoint_000048848_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:014d403a5012a06b8ed53817c88c7782e75d955e72241f0c2c98fd6bd33dedfd
+size 8897853

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs6:obs30:stride1:seed12/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs6_seed12",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 12,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs6-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs6",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 6,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 12,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed12/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs6_seed12 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 12 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed12/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 12 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs6-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs6 seed12 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 6 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs6_seed12",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 12,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs6-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs6",
+      "seed12"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 6,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 12,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed12/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs6/seed_12",
+  "wandb_unique_id": "flappy-fs6-uniform_u1_3-s12"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs6:obs30:stride1:seed13/checkpoint_p0/checkpoint_000048848_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09e078d30189840492551e1cbefaf0c572f759824e34a7876a4ab3851ea72507
+size 8897853

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs6:obs30:stride1:seed13/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs6_seed13",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 13,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs6-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs6",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 6,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 13,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed13/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs6_seed13 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 13 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed13/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 13 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs6-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs6 seed13 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 6 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs6_seed13",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 13,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs6-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs6",
+      "seed13"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 6,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 13,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed13/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs6/seed_13",
+  "wandb_unique_id": "flappy-fs6-uniform_u1_3-s13"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs6:obs30:stride1:seed14/checkpoint_p0/checkpoint_000048840_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34e4783a1c7b69598cd725a16df7768b9d4e00e3f6a13dbe31f247a7566b1229
+size 8897853

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs6:obs30:stride1:seed14/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs6_seed14",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 14,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs6-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs6",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 6,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 14,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed14/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs6_seed14 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 14 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed14/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 14 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs6-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs6 seed14 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 6 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs6_seed14",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 14,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs6-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs6",
+      "seed14"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 6,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 14,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs6_seed14/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs6/seed_14",
+  "wandb_unique_id": "flappy-fs6-uniform_u1_3-s14"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs7:obs30:stride1:seed10/checkpoint_p0/checkpoint_000048848_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9fe8ed7ea2c351a614930f14bb93ba604a5d2c9a15dd7546cdaf1d452bb7ed0f
+size 9261117

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs7:obs30:stride1:seed10/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs7_seed10",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 10,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs7-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs7",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 7,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 10,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed10/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs7_seed10 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 10 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed10/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 10 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs7-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs7 seed10 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 7 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs7_seed10",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 10,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs7-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs7",
+      "seed10"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 7,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 10,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed10/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs7/seed_10",
+  "wandb_unique_id": "flappy-fs7-uniform_u1_3-s10"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs7:obs30:stride1:seed11/checkpoint_p0/checkpoint_000048840_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88ede8f6a75e888994331766085f9d9acf444383be1762de9636c618350895a9
+size 9261117

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs7:obs30:stride1:seed11/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs7_seed11",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 11,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs7-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs7",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 7,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 11,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed11/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs7_seed11 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 11 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed11/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 11 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs7-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs7 seed11 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 7 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs7_seed11",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 11,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs7-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs7",
+      "seed11"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 7,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 11,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed11/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs7/seed_11",
+  "wandb_unique_id": "flappy-fs7-uniform_u1_3-s11"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs7:obs30:stride1:seed12/checkpoint_p0/checkpoint_000048808_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b7d5c3b8b2da8ff4174b5410eb7079035754d4fc180dcd2088a8268c7b65168
+size 9261117

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs7:obs30:stride1:seed12/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs7_seed12",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 12,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs7-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs7",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 7,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 12,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed12/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs7_seed12 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 12 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed12/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 12 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs7-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs7 seed12 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 7 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs7_seed12",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 12,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs7-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs7",
+      "seed12"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 7,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 12,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed12/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs7/seed_12",
+  "wandb_unique_id": "flappy-fs7-uniform_u1_3-s12"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs7:obs30:stride1:seed13/checkpoint_p0/checkpoint_000048840_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c80cb2340a2c82d711f9c016e6ebe347487bedba4b479323097e860601d5e8f5
+size 9261117

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs7:obs30:stride1:seed13/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs7_seed13",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 13,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs7-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs7",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 7,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 13,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed13/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs7_seed13 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 13 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed13/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 13 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs7-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs7 seed13 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 7 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs7_seed13",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 13,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs7-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs7",
+      "seed13"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 7,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 13,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed13/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs7/seed_13",
+  "wandb_unique_id": "flappy-fs7-uniform_u1_3-s13"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs7:obs30:stride1:seed14/checkpoint_p0/checkpoint_000048848_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38e9f24e78d02e53e1e1a2eabb7f1660eff289ac3a0cbc1df3f10cc47d5ffa5f
+size 9261117

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs7:obs30:stride1:seed14/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs7_seed14",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 14,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs7-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs7",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 7,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 14,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed14/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs7_seed14 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 14 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed14/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 14 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs7-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs7 seed14 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 7 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs7_seed14",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 14,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs7-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs7",
+      "seed14"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 7,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 14,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs7_seed14/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs7/seed_14",
+  "wandb_unique_id": "flappy-fs7-uniform_u1_3-s14"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs8:obs30:stride1:seed10/checkpoint_p0/checkpoint_000048840_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8311d598c265a12101d23606d506b55ec20167e24e2917d0772588550afc37f8
+size 9624381

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs8:obs30:stride1:seed10/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs8_seed10",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 10,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs8-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs8",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 8,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 10,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed10/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs8_seed10 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 10 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed10/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 10 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs8-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs8 seed10 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 8 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs8_seed10",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 10,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs8-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs8",
+      "seed10"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 8,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 10,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed10/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs8/seed_10",
+  "wandb_unique_id": "flappy-fs8-uniform_u1_3-s10"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs8:obs30:stride1:seed11/checkpoint_p0/checkpoint_000048824_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6970c95d9cb37bb37a21020c8773c653ff295fceb08921650b6a2cd3a21559fc
+size 9624381

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs8:obs30:stride1:seed11/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs8_seed11",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 11,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs8-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs8",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 8,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 11,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed11/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs8_seed11 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 11 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed11/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 11 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs8-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs8 seed11 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 8 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs8_seed11",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 11,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs8-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs8",
+      "seed11"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 8,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 11,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed11/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs8/seed_11",
+  "wandb_unique_id": "flappy-fs8-uniform_u1_3-s11"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs8:obs30:stride1:seed12/checkpoint_p0/checkpoint_000048840_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:854336b294909de067402556b8c958150bfc6fb57c702589167f96d4a80f15ab
+size 9624381

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs8:obs30:stride1:seed12/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs8_seed12",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 12,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs8-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs8",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 8,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 12,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed12/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs8_seed12 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 12 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed12/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 12 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs8-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs8 seed12 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 8 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs8_seed12",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 12,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs8-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs8",
+      "seed12"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 8,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 12,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed12/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs8/seed_12",
+  "wandb_unique_id": "flappy-fs8-uniform_u1_3-s12"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs8:obs30:stride1:seed13/checkpoint_p0/checkpoint_000048848_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee18312bdeba74a9d05d0e044b795363fd75dca95e49d0f2aa1b0f48cb345e19
+size 9624381

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs8:obs30:stride1:seed13/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs8_seed13",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 13,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs8-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs8",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 8,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 13,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed13/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs8_seed13 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 13 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed13/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 13 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs8-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs8 seed13 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 8 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs8_seed13",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 13,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs8-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs8",
+      "seed13"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 8,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 13,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed13/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs8/seed_13",
+  "wandb_unique_id": "flappy-fs8-uniform_u1_3-s13"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs8:obs30:stride1:seed14/checkpoint_p0/checkpoint_000048776_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c218fe4eb2aaeaa80b53bd0642a4ecf6bd70e0349e957bac9ad756b58be9b0fe
+size 9624381

latency-aware/flappy/small-policy/factor_sweep:flappy:frame_stack:uniform_u1_3:fs8:obs30:stride1:seed14/config.json

@@ -0,0 +1,268 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_frame_stack_uniform_u1_3_fs8_seed14",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 14,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-fs8-uniform_u1_3",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "frame_stack",
+    "uniform",
+    "uniform_u1_3",
+    "fs8",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 8,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "uniform_distribution",
+  "fixed_latency_ms": null,
+  "mean_latency_ms": null,
+  "std_latency_ms": null,
+  "min_latency_ms": 33.333333333333336,
+  "max_latency_ms": 100.0,
+  "latency_seed": 14,
+  "add_latency_info": false,
+  "max_pending_actions": null,
+  "hold_policy": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed14/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_frame_stack_uniform_u1_3_fs8_seed14 --train_dir results/checkpoints_factor_sweeps/flappy/context_window --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 14 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed14/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type uniform_distribution --min-latency-ms 33.333333333333336 --max-latency-ms 100.0 --latency-seed 14 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-fs8-uniform_u1_3 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy frame_stack uniform uniform_u1_3 fs8 seed14 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 8 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_frame_stack_uniform_u1_3_fs8_seed14",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/context_window",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 14,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-fs8-uniform_u1_3",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "frame_stack",
+      "uniform",
+      "uniform_u1_3",
+      "fs8",
+      "seed14"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 8,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "uniform_distribution",
+    "min_latency_ms": 33.333333333333336,
+    "max_latency_ms": 100.0,
+    "latency_seed": 14,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/context_window/flappy_frame_stack_uniform_u1_3_fs8_seed14/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/context_window/train/frame_stack/uniform_u1_3/fs8/seed_14",
+  "wandb_unique_id": "flappy-fs8-uniform_u1_3-s14"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs10:stride3:seed10/checkpoint_p0/checkpoint_000048696_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8624af5403d24dfa14b12fbc6919c58dab8e0943ec5cbbef1fb80ba5626d9f30
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs10:stride3:seed10/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride3_seed10",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 10,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride3-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride3",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 10.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed10/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride3_seed10 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 10 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed10/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride3-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride3 seed10 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 10.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride3_seed10",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 10,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride3-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride3",
+      "seed10"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 10.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed10/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride3/seed_10",
+  "wandb_unique_id": "flappy-stride3-fixed_l2-s10"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs10:stride3:seed11/checkpoint_p0/checkpoint_000048688_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36a38ef1642cd8aa112165be7dc2536abc30ad5c897849bd84e90c8a639ac73c
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs10:stride3:seed11/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride3_seed11",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 11,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride3-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride3",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 10.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed11/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride3_seed11 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 11 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed11/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride3-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride3 seed11 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 10.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride3_seed11",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 11,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride3-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride3",
+      "seed11"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 10.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed11/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride3/seed_11",
+  "wandb_unique_id": "flappy-stride3-fixed_l2-s11"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs10:stride3:seed12/checkpoint_p0/checkpoint_000048680_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7d5269a88f9247efe7d67e10f997a96f1592420dffbd6c8154c19b23f46e82f
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs10:stride3:seed12/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride3_seed12",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 12,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride3-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride3",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 10.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed12/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride3_seed12 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 12 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed12/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride3-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride3 seed12 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 10.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride3_seed12",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 12,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride3-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride3",
+      "seed12"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 10.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed12/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride3/seed_12",
+  "wandb_unique_id": "flappy-stride3-fixed_l2-s12"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs10:stride3:seed13/checkpoint_p0/checkpoint_000048672_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27d563f62240408e3146d4684fc397ba63fdc06a5f05883bf58ffaf528a2f419
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs10:stride3:seed13/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride3_seed13",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 13,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride3-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride3",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 10.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed13/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride3_seed13 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 13 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed13/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride3-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride3 seed13 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 10.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride3_seed13",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 13,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride3-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride3",
+      "seed13"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 10.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed13/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride3/seed_13",
+  "wandb_unique_id": "flappy-stride3-fixed_l2-s13"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs10:stride3:seed14/checkpoint_p0/checkpoint_000048696_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff65aeb8a51fcf625537d84a4dab66c0cebd037601c997cc7f979f93d76e322f
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs10:stride3:seed14/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride3_seed14",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 14,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride3-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride3",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 10.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed14/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride3_seed14 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 14 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed14/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride3-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride3 seed14 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 10.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride3_seed14",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 14,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride3-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride3",
+      "seed14"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 10.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride3_seed14/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride3/seed_14",
+  "wandb_unique_id": "flappy-stride3-fixed_l2-s14"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs15:stride2:seed10/checkpoint_p0/checkpoint_000048760_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03da43bc7b18c0f2a9c47371107c208521664cf1e9a27c1057049399293fcac8
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs15:stride2:seed10/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride2_seed10",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 10,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride2-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride2",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 15.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed10/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride2_seed10 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 10 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed10/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride2-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride2 seed10 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 15.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride2_seed10",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 10,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride2-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride2",
+      "seed10"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 15.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed10/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride2/seed_10",
+  "wandb_unique_id": "flappy-stride2-fixed_l2-s10"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs15:stride2:seed11/checkpoint_p0/checkpoint_000048736_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0269fb4dc79d32e416d8dd1c9dbd88bf55e060e8fddcd1d1f997238f5b0bdffa
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs15:stride2:seed11/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride2_seed11",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 11,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride2-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride2",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 15.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed11/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride2_seed11 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 11 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed11/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride2-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride2 seed11 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 15.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride2_seed11",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 11,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride2-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride2",
+      "seed11"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 15.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed11/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride2/seed_11",
+  "wandb_unique_id": "flappy-stride2-fixed_l2-s11"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs15:stride2:seed12/checkpoint_p0/checkpoint_000048832_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3663804ec7080b9629dd765a5578cc346a4a9620c4644dd9578a2e67e32645a
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs15:stride2:seed12/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride2_seed12",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 12,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride2-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride2",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 15.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed12/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride2_seed12 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 12 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed12/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride2-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride2 seed12 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 15.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride2_seed12",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 12,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride2-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride2",
+      "seed12"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 15.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed12/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride2/seed_12",
+  "wandb_unique_id": "flappy-stride2-fixed_l2-s12"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs15:stride2:seed13/checkpoint_p0/checkpoint_000048672_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5485a87742e1f4c5379dd71c0ebd6679e29ae3d2f911baa8bdf8092f391353ca
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs15:stride2:seed13/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride2_seed13",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 13,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride2-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride2",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 15.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed13/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride2_seed13 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 13 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed13/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride2-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride2 seed13 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 15.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride2_seed13",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 13,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride2-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride2",
+      "seed13"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 15.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed13/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride2/seed_13",
+  "wandb_unique_id": "flappy-stride2-fixed_l2-s13"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs15:stride2:seed14/checkpoint_p0/checkpoint_000048744_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6293d94d68bbd2908b11ea4eb3787e79f96f5d8083c556fe91250eca9666a2a2
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs15:stride2:seed14/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride2_seed14",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 14,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride2-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride2",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 15.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed14/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride2_seed14 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 14 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed14/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride2-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride2 seed14 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 15.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride2_seed14",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 14,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride2-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride2",
+      "seed14"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 15.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride2_seed14/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride2/seed_14",
+  "wandb_unique_id": "flappy-stride2-fixed_l2-s14"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs30:stride1:seed10/checkpoint_p0/checkpoint_000048824_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a74f54e0eafcbb778eb596f0d5735b489e327932ef3208ac3f93932af092830
+size 8171325

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs30:stride1:seed10/config.json

@@ -0,0 +1,266 @@
+{
+  "help": false,
+  "algo": "APPO",
+  "env": "latency_flappy",
+  "experiment": "flappy_observation_stride_fixed_l2_stride1_seed10",
+  "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+  "restart_behavior": "resume",
+  "device": "gpu",
+  "seed": 10,
+  "num_policies": 1,
+  "async_rl": true,
+  "serial_mode": false,
+  "batched_sampling": true,
+  "num_batches_to_accumulate": 2,
+  "worker_num_splits": 1,
+  "policy_workers_per_policy": 1,
+  "max_policy_lag": 400,
+  "num_workers": 2,
+  "num_envs_per_worker": 1,
+  "batch_size": 4096,
+  "num_batches_per_epoch": 8,
+  "num_epochs": 8,
+  "rollout": 128,
+  "recurrence": 1,
+  "shuffle_minibatches": false,
+  "gamma": 0.99,
+  "reward_scale": 1.0,
+  "reward_clip": 1000.0,
+  "value_bootstrap": false,
+  "normalize_returns": true,
+  "exploration_loss_coeff": 0.003,
+  "value_loss_coeff": 0.5,
+  "kl_loss_coeff": 0.0,
+  "exploration_loss": "entropy",
+  "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-05,
+  "adam_beta1": 0.9,
+  "adam_beta2": 0.999,
+  "max_grad_norm": 0.5,
+  "learning_rate": 0.00025,
+  "lr_schedule": "linear_decay",
+  "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": [
+    0
+  ],
+  "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": 180,
+  "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_atari",
+  "encoder_conv_mlp_layers": [
+    512
+  ],
+  "use_rnn": false,
+  "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": true,
+  "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": "flappy-stride1-fixed_l2",
+  "wandb_job_type": "sample_factory",
+  "wandb_tags": [
+    "factor_sweep",
+    "flappy",
+    "observation_stride",
+    "fixed",
+    "fixed_l2",
+    "stride1",
+    "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,
+  "gym_id": "FlappyBird-v0",
+  "env_fps": 30.0,
+  "obs_fps": 30.0,
+  "use_lidar": false,
+  "normalize_obs": true,
+  "audio_on": false,
+  "screen_size": "",
+  "obs_resize": "84,84",
+  "use_gpu_render": true,
+  "simulator": "gpu",
+  "gpu_render_device": "auto",
+  "gpu_render_batch_size": 128,
+  "gpu_render_profile": false,
+  "gpu_render_profile_interval": 200,
+  "pipe_gap": 100,
+  "bird_color": "yellow",
+  "pipe_color": "green",
+  "background": "day",
+  "score_limit": -1,
+  "frame_stack": 4,
+  "debug": false,
+  "debug_timelimit_diagnostics": false,
+  "max_episode_steps": 0,
+  "mode": "train",
+  "latency_type": "fixed",
+  "fixed_latency_ms": 66.66666666666667,
+  "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": "one_frame_then_noop",
+  "ordering_policy": "latest_ready",
+  "eval_episodes": 100,
+  "eval_parallel_envs": 100,
+  "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+  "eval_max_steps": 3600,
+  "eval_deterministic": true,
+  "eval_raw_reward": false,
+  "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride1_seed10/episode_metrics.jsonl",
+  "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_observation_stride_fixed_l2_stride1_seed10 --train_dir results/checkpoints_factor_sweeps/flappy/observation_stride --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 10 --episode_metrics_path results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride1_seed10/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 66.66666666666667 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 100 --eval-max-steps 3600 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-stride1-fixed_l2 --wandb_job_type sample_factory --wandb_tags factor_sweep flappy observation_stride fixed fixed_l2 stride1 seed10 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+  "cli_args": {
+    "algo": "APPO",
+    "env": "latency_flappy",
+    "experiment": "flappy_observation_stride_fixed_l2_stride1_seed10",
+    "train_dir": "results/checkpoints_factor_sweeps/flappy/observation_stride",
+    "restart_behavior": "resume",
+    "device": "gpu",
+    "seed": 10,
+    "num_policies": 1,
+    "async_rl": true,
+    "batched_sampling": true,
+    "worker_num_splits": 1,
+    "max_policy_lag": 400,
+    "num_workers": 2,
+    "num_envs_per_worker": 1,
+    "batch_size": 4096,
+    "num_batches_per_epoch": 8,
+    "num_epochs": 8,
+    "rollout": 128,
+    "recurrence": 1,
+    "gamma": 0.99,
+    "normalize_returns": true,
+    "value_loss_coeff": 0.5,
+    "gae_lambda": 0.95,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "max_grad_norm": 0.5,
+    "learning_rate": 0.00025,
+    "normalize_input": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "experiment_summaries_interval": 1,
+    "stats_avg": 100,
+    "train_for_env_steps": 25000000,
+    "save_every_sec": 600,
+    "keep_checkpoints": 5,
+    "use_rnn": false,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "with_wandb": true,
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_group": "flappy-stride1-fixed_l2",
+    "wandb_job_type": "sample_factory",
+    "wandb_tags": [
+      "factor_sweep",
+      "flappy",
+      "observation_stride",
+      "fixed",
+      "fixed_l2",
+      "stride1",
+      "seed10"
+    ],
+    "gym_id": "FlappyBird-v0",
+    "env_fps": 30.0,
+    "obs_fps": 30.0,
+    "use_lidar": false,
+    "normalize_obs": true,
+    "audio_on": false,
+    "obs_resize": "84,84",
+    "use_gpu_render": true,
+    "simulator": "gpu",
+    "gpu_render_device": "auto",
+    "gpu_render_batch_size": 128,
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "pipe_gap": 100,
+    "bird_color": "yellow",
+    "pipe_color": "green",
+    "background": "day",
+    "frame_stack": 4,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "mode": "train",
+    "latency_type": "fixed",
+    "fixed_latency_ms": 66.66666666666667,
+    "add_latency_info": false,
+    "hold_policy": "one_frame_then_noop",
+    "ordering_policy": "latest_ready",
+    "eval_episodes": 100,
+    "eval_parallel_envs": 100,
+    "eval_max_steps": 3600,
+    "eval_deterministic": true,
+    "episode_metrics_path": "results/checkpoints_factor_sweeps/flappy/observation_stride/flappy_observation_stride_fixed_l2_stride1_seed10/episode_metrics.jsonl"
+  },
+  "git_hash": "284fe8ace24f0e8a40c03c5b559969abd7caeb29",
+  "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
+  "eval_env_frameskip": 1,
+  "output_dir": "outputs/factor_sweeps/flappy/observation_stride/train/observation_stride/fixed_l2/stride1/seed_10",
+  "wandb_unique_id": "flappy-stride1-fixed_l2-s10"
+}

latency-aware/flappy/small-policy/factor_sweep:flappy:observation_stride:fixed_l2:fs4:obs30:stride1:seed11/checkpoint_p0/checkpoint_000048792_25034752.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:630e186030b0da152abde0784b4d24d6b55d2f9d88a98bfffaf8172cdf6cef1d
+size 8171325