sim2real sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z: sim/temporal/3

sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/published/sim/temporal/3.json

@@ -0,0 +1 @@
+{"stage": "sim/temporal/3", "files": ["sim/temporal/3/DONE", "sim/temporal/3/actions.jsonl", "sim/temporal/3/e2e_latencies.jsonl", "sim/temporal/3/episode_metrics.jsonl", "sim/temporal/3/infer_latencies.jsonl", "sim/temporal/3/latencies.jsonl", "sim/temporal/3/observation_attempts.jsonl", "sim/temporal/3/queue_eval_latency_profile_sample.json", "sim/temporal/3/queue_eval_results.jsonl", "sim/temporal/3/stdout.log", "sim/temporal/3/steps.jsonl"]}

sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/sim/temporal/3/episode_metrics.jsonl

@@ -0,0 +1,2 @@
+{"episode_id": 0, "episode_return": 2557.89211964274, "episode_return_env": 2557.89211964274, "game_score": null, "mean_latency_ms": 130.16120202317967, "metadata": {"action_prefix": {"mode": "none"}, "checkpoint_kind": "best", "checkpoint_path": "/data/ubuntu/lzj/cache/huggingface/hub/models--latency-sensitive-bench--benchmark-models/snapshots/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/ant/vla/starvla-qwengr00t-h1/ant_profile_gr00t_h1_5fps_2h100_20260917/checkpoints/model.pt", "config_name": "gr00t_ant_measured", "dropped_observation_count": 0, "env_fps": 4.029011771057316, "env_id": "LatencyBench/AntContinuous-v0", "episode_seed": 2000160, "frame_ms": 248.19982090486033, "gpu_class": "1x-rtx3090", "idle_worker_count": 0, "in_flight_count": 1, "inference_worker_count": 1, "instance_id": "instance_33621945a15d3b38", "latency_type": "profile_sample", "mode": "simulated", "model_id": "qwengr00t", "obs_fps": 4.029011771057316, "output_dir": "/data/ubuntu/lzj/sim2real-return-20260929/results/sim2real/replays/sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/4d554e4dcd73eb4074b3f7c819dac1abe263fd40/sim/temporal/3", "policy_id": "starvla", "profile_ref": null, "run_name": "gr00t_ant_measured", "sim_inference_pool_workers": 1, "simulated_worker_capacity": 1, "source_profile_path": "/data/ubuntu/lzj/sim2real-return-20260929/results/sim2real/replays/sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/4d554e4dcd73eb4074b3f7c819dac1abe263fd40/profile/profile.json", "source_run_id": "20260928T235512274259Z", "submitted_observation_frames": 1000, "workload_id": "ant"}, "num_actions": 1000, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 137.09109651265567, "p99_latency_ms": 149.68363650125036, "return_raw": null, "survival_steps": 1000}
+{"episode_id": 1, "episode_return": 2297.892448586378, "episode_return_env": 2297.892448586378, "game_score": null, "mean_latency_ms": 130.2271359596128, "metadata": {"action_prefix": {"mode": "none"}, "checkpoint_kind": "best", "checkpoint_path": "/data/ubuntu/lzj/cache/huggingface/hub/models--latency-sensitive-bench--benchmark-models/snapshots/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/ant/vla/starvla-qwengr00t-h1/ant_profile_gr00t_h1_5fps_2h100_20260917/checkpoints/model.pt", "config_name": "gr00t_ant_measured", "dropped_observation_count": 0, "env_fps": 4.029011771057316, "env_id": "LatencyBench/AntContinuous-v0", "episode_seed": 2000161, "frame_ms": 248.19982090486033, "gpu_class": "1x-rtx3090", "idle_worker_count": 0, "in_flight_count": 1, "inference_worker_count": 1, "instance_id": "instance_33621945a15d3b38", "latency_type": "profile_sample", "mode": "simulated", "model_id": "qwengr00t", "obs_fps": 4.029011771057316, "output_dir": "/data/ubuntu/lzj/sim2real-return-20260929/results/sim2real/replays/sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/4d554e4dcd73eb4074b3f7c819dac1abe263fd40/sim/temporal/3", "policy_id": "starvla", "profile_ref": null, "run_name": "gr00t_ant_measured", "sim_inference_pool_workers": 1, "simulated_worker_capacity": 1, "source_profile_path": "/data/ubuntu/lzj/sim2real-return-20260929/results/sim2real/replays/sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/4d554e4dcd73eb4074b3f7c819dac1abe263fd40/profile/profile.json", "source_run_id": "20260928T235512274259Z", "submitted_observation_frames": 1000, "workload_id": "ant"}, "num_actions": 1000, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 137.49333424067103, "p99_latency_ms": 151.0910033250068, "return_raw": null, "survival_steps": 1000}

sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/sim/temporal/3/queue_eval_latency_profile_sample.json

@@ -0,0 +1,22 @@
+{
+  "checkpoint_kind": "best",
+  "checkpoint_path": "/data/ubuntu/lzj/cache/huggingface/hub/models--latency-sensitive-bench--benchmark-models/snapshots/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/ant/vla/starvla-qwengr00t-h1/ant_profile_gr00t_h1_5fps_2h100_20260917/checkpoints/model.pt",
+  "experiment_name": "gr00t_ant_measured",
+  "latency": "profile_sample",
+  "latency_type": "profile_sample",
+  "lengths": [
+    1000,
+    1000
+  ],
+  "mean_length": 1000.0,
+  "mean_return": 2427.892284114559,
+  "returns": [
+    2557.89211964274,
+    2297.892448586378
+  ],
+  "seed": 2000160,
+  "source_profile_path": "/data/ubuntu/lzj/sim2real-return-20260929/results/sim2real/replays/sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/4d554e4dcd73eb4074b3f7c819dac1abe263fd40/profile/profile.json",
+  "std_return": 129.99983552818094,
+  "suite_name": "profile_sample",
+  "timestamp_utc": "2026-09-29T05:59:25.942014+00:00"
+}

sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/sim/temporal/3/queue_eval_results.jsonl

@@ -0,0 +1 @@
+{"checkpoint_kind": "best", "checkpoint_path": "/data/ubuntu/lzj/cache/huggingface/hub/models--latency-sensitive-bench--benchmark-models/snapshots/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/ant/vla/starvla-qwengr00t-h1/ant_profile_gr00t_h1_5fps_2h100_20260917/checkpoints/model.pt", "experiment_name": "gr00t_ant_measured", "latency": "profile_sample", "latency_type": "profile_sample", "lengths": [1000, 1000], "mean_length": 1000.0, "mean_return": 2427.892284114559, "returns": [2557.89211964274, 2297.892448586378], "seed": 2000160, "source_profile_path": "/data/ubuntu/lzj/sim2real-return-20260929/results/sim2real/replays/sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/4d554e4dcd73eb4074b3f7c819dac1abe263fd40/profile/profile.json", "std_return": 129.99983552818094, "suite_name": "profile_sample", "timestamp_utc": "2026-09-29T05:59:25.942014+00:00"}

sim2real/ant/qwengr00t/1x-rtx3090/instance_33621945a15d3b38/20260928T235512274259Z/sim/temporal/3/stdout.log

@@ -0,0 +1,80 @@
+[bench] run=gr00t_ant_measured sweeps=1 episodes_per_sweep=2 total_episode_runs=2
+[bench] sweep 1/1: eval_latency=profile_sample
+09/29 [05:56:20] INFO     | >> No OpenGL_accelerate module loaded: No module named 'OpenGL_accelerate'                                                                                             acceleratesupport.py:24
+/data/ubuntu/lzj/conda/envs/qwenoft/lib/python3.10/site-packages/albumentations/__init__.py:13: UserWarning: A new version of Albumentations is available: 2.0.8 (you have 1.4.18). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1.
+  check_for_updates()
+09/29 [05:56:24] INFO     | >>  Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']                                                                                                        registry.py:113
+                 INFO     | >>  Loaded data_config from DOMINO: ['robotwin']                                                                                                                               registry.py:107
+                 INFO     | >>  Loaded embodiment_tags from DOMINO (data_config): []                                                                                                                       registry.py:110
+                 INFO     | >>  Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']                                                                   registry.py:113
+                 INFO     | >>  Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']                                                                           registry.py:107
+                 INFO     | >>  Loaded embodiment_tags from Franka (data_config): []                                                                                                                       registry.py:110
+                 INFO     | >>  Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']                                                     registry.py:113
+                 INFO     | >>  Loaded data_config from LIBERO: ['libero_franka']                                                                                                                          registry.py:107
+                 INFO     | >>  Loaded embodiment_tags from LIBERO (data_config): []                                                                                                                       registry.py:110
+                 INFO     | >>  Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']                                                                                    registry.py:113
+                 INFO     | >>  Loaded data_config from MIKASA: ['mikasa_franka_h1']                                                                                                                       registry.py:107
+                 INFO     | >>  Loaded embodiment_tags from MIKASA (data_config): ['mikasa_franka_h1']                                                                                                     registry.py:110
+                 INFO     | >>  Loaded mixtures from MIKASA (data_config): ['local/intercept_grab_fast_vla_v0_h1_train']                                                                                   registry.py:113
+                 INFO     | >>  Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']                                                        registry.py:107
+                 INFO     | >>  Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']                                      registry.py:110
+                 INFO     | >>  Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', registry.py:113
+                          'robochallenge_table30v2_dosw1_all']                                                                                                                                                            
+                 INFO     | >>  Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']                                                                                                          registry.py:107
+                 INFO     | >>  Loaded embodiment_tags from Robocasa_365 (data_config): []                                                                                                                 registry.py:110
+                 INFO     | >>  Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', registry.py:113
+                          'robocasa365_target_human_all']                                                                                                                                                                 
+                 INFO     | >>  Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']                                                                                                      registry.py:107
+                 INFO     | >>  Loaded embodiment_tags from Robocasa_tabletop (data_config): []                                                                                                            registry.py:110
+                 INFO     | >>  Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']                                                                                         registry.py:113
+                 INFO     | >>  Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']                                                                                                     registry.py:107
+                 INFO     | >>  Loaded embodiment_tags from Robotwin (data_config): []                                                                                                                     registry.py:110
+                 INFO     | >>  Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']                                 registry.py:113
+                 INFO     | >>  Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']                                                                                                 registry.py:107
+                 INFO     | >>  Loaded embodiment_tags from SimplerEnv (data_config): []                                                                                                                   registry.py:110
+                 INFO     | >>  Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']                                                                                                   registry.py:113
+                 INFO     | >>  Loaded data_config from VLA-Arena: ['vla_arena_franka']                                                                                                                    registry.py:107
+                 INFO     | >>  Loaded embodiment_tags from VLA-Arena (data_config): []                                                                                                                    registry.py:110
+                 INFO     | >>  Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']                                                                       registry.py:113
+                 INFO     | >>  Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_defend_the_line', 'rl_games_deadly_corridor', 'rl_games_asterix',                 registry.py:107
+                          'rl_games_atlantis', 'rl_games_gymnasium', 'rl_games_gymnasium_discrete', 'rl_games_gymnasium_native']                                                                                          
+                 INFO     | >>  Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_defend_the_line', 'rl_games_deadly_corridor',                   registry.py:110
+                          'rl_games_asterix', 'rl_games_atlantis', 'rl_games_gymnasium', 'rl_games_gymnasium_discrete', 'rl_games_gymnasium_native']                                                                      
+                 INFO     | >>  Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge',                  registry.py:113
+                          'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'defend_the_line_train',                                    
+                          'defend_the_line_train__bridge', 'defend_the_line_mixed_latency_train', 'defend_the_line_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge',                
+                          'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge', 'asterix_train', 'asterix_train__bridge', 'asterix_mixed_latency_train',                                  
+                          'asterix_mixed_latency_train__bridge', 'atlantis_train', 'atlantis_train__bridge', 'atlantis_mixed_latency_train', 'atlantis_mixed_latency_train__bridge',                                      
+                          'h1hand_balance_hard']                                                                                                                                                                          
+                 INFO     | >> PolicyServerWrapper: loading framework from                                                                                                                            policy_wrapper.py:73
+                          /data/ubuntu/lzj/cache/huggingface/hub/models--latency-sensitive-bench--benchmark-models/snapshots/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/ant/vla/starvla-q                     
+                          wengr00t-h1/ant_profile_gr00t_h1_5fps_2h100_20260917/checkpoints/model.pt                                                                                                                       
+                 INFO     | >> [*] Loading from local checkpoint path                                                                                                                                   share_tools.py:418
+                          `/data/ubuntu/lzj/cache/huggingface/hub/models--latency-sensitive-bench--benchmark-models/snapshots/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/ant/vla/starvla-qw                   
+                          engr00t-h1/ant_profile_gr00t_h1_5fps_2h100_20260917/checkpoints/model.pt`                                                                                                                       
+[QWen3] loading /data/ubuntu/lzj/cache/huggingface/hub/models--Qwen--Qwen3-VL-4B-Instruct/snapshots/ebb281ec70b05090aa6165b016eac8ec08e71b17 with gradient_checkpointing=True
+
+[QWen3] gradient_checkpointing ENABLED (use_reentrant=False, active=True, text_use_cache=False)
+Total number of DiT parameters:  157032448
+09/29 [05:56:38] INFO     | >> [*] Loading from local checkpoint path                                                                                                                                   share_tools.py:418
+                          `/data/ubuntu/lzj/cache/huggingface/hub/models--latency-sensitive-bench--benchmark-models/snapshots/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/ant/vla/starvla-qw                   
+                          engr00t-h1/ant_profile_gr00t_h1_5fps_2h100_20260917/checkpoints/model.pt`                                                                                                                       
+                 INFO     | >> [*] Loading from local checkpoint path                                                                                                                                   share_tools.py:418
+                          `/data/ubuntu/lzj/cache/huggingface/hub/models--latency-sensitive-bench--benchmark-models/snapshots/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/ant/vla/starvla-qw                   
+                          engr00t-h1/ant_profile_gr00t_h1_5fps_2h100_20260917/checkpoints/model.pt`                                                                                                                       
+                 INFO     | >> [*] Loading from local checkpoint path                                                                                                                                   share_tools.py:418
+                          `/data/ubuntu/lzj/cache/huggingface/hub/models--latency-sensitive-bench--benchmark-models/snapshots/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/ant/vla/starvla-qw                   
+                          engr00t-h1/ant_profile_gr00t_h1_5fps_2h100_20260917/checkpoints/model.pt`                                                                                                                       
+                 INFO     | >> PolicyNormProcessor ready: robot_type=rl_games_gymnasium, unnorm_key=new_embodiment, action_keys=['action.button'] (dims=[8]), state_keys=['state.game_state'] policy_norm_processor.py:333
+                 INFO     | >> PolicyServerWrapper ready: action_chunk_size=1, default_unnorm_key=new_embodiment, available_unnorm_keys=['new_embodiment'], action_keys=['action.button'],           policy_wrapper.py:126
+                          state_keys=['state.game_state']                                                                                                                                                                 
+                 INFO     | >> [*] Loading from local checkpoint path                                                                                                                                   share_tools.py:418
+                          `/data/ubuntu/lzj/cache/huggingface/hub/models--latency-sensitive-bench--benchmark-models/snapshots/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/ant/vla/starvla-qw                   
+                          engr00t-h1/ant_profile_gr00t_h1_5fps_2h100_20260917/checkpoints/model.pt`                                                                                                                       
+                 INFO     | >> PolicyNormProcessor ready: robot_type=rl_games_gymnasium, unnorm_key=new_embodiment, action_keys=['action.button'] (dims=[8]), state_keys=['state.game_state'] policy_norm_processor.py:333
+[bench] sweep 1/1 episode 1/2 done
+[bench] sweep 1/1 episode 2/2 done
+[bench] sweep 1/1 complete elapsed=186.3s
+episode=0 return=2557.892 steps=1000 mean_latency_ms=130.16120202317967
+episode=1 return=2297.892 steps=1000 mean_latency_ms=130.2271359596128
+summary latency=profile_sample episodes=2 mean_return=2427.892 std_return=130.000 mean_length=1000.0