add N015/sawyer_umi/dr_off/seed_5003

N015/sawyer_umi/dr_off/seed_5003/check.json

@@ -27,27 +27,27 @@
   },
   "rows = recorded control steps / 2 (= episode length at 25 Hz)": {
    "pass": true,
-   "detail": "3625 rows vs ceil(7250/2) = 3625"
+   "detail": "3603 rows vs ceil(7206/2) = 3603"
   },
   "rows are every 2nd control step (sim_step_idx)": {
    "pass": true,
-   "detail": "sim_step_idx 251..7499, step differences [2]"
+   "detail": "sim_step_idx 251..7455, step differences [2]"
   },
   "video observation.images.head decodes, 640x480, not black": {
    "pass": true,
-   "detail": "86.74 MB, 3625 frames at 25 fps, mean luma 84..147, std 35, new picture 3624x = 25.0 per s"
+   "detail": "85.11 MB, 3603 frames at 25 fps, mean luma 84..147, std 35, new picture 3602x = 25.0 per s"
   },
   "video observation.images.head frames = parquet rows": {
    "pass": true,
-   "detail": "3625 vs 3625"
+   "detail": "3603 vs 3603"
   },
   "video observation.images.wrist decodes, 640x480, not black": {
    "pass": true,
-   "detail": "73.35 MB, 3625 frames at 25 fps, mean luma 95..173, std 31, new picture 3624x = 25.0 per s"
+   "detail": "71.99 MB, 3603 frames at 25 fps, mean luma 95..176, std 31, new picture 3602x = 25.0 per s"
   },
   "video observation.images.wrist frames = parquet rows": {
    "pass": true,
-   "detail": "3625 vs 3625"
+   "detail": "3603 vs 3603"
   },
   "sim_info present": {
    "pass": true,

N015/sawyer_umi/dr_off/seed_5003/check.txt

@@ -1,21 +1,21 @@
 collect dataset check  cell=/mnt/scratch/sz850_atomic/n015_saw_off_5003/out  task=N015  at verify_collect_dataset.py
-dataset root /mnt/scratch/sz850_atomic/n015_saw_off_5003/out/tiltmaze/000/lerobot_v21  codebase v2.1  fps 25  episodes 1  frames 3625  videos 2  robot_type RidgebackSawyerUmiCfg
+dataset root /mnt/scratch/sz850_atomic/n015_saw_off_5003/out/tiltmaze/000/lerobot_v21  codebase v2.1  fps 25  episodes 1  frames 3603  videos 2  robot_type RidgebackSawyerUmiCfg
 [PASS] codebase v2.1: v2.1
 [PASS] dataset fps = 25: info.fps 25
 (no `lerobot` python given (--lerobot-python / LEROBOT_PY): loading with pyarrow + meta/info.json only)
 [PASS] parquet episodes: 1 files, info.total_episodes=1
---- episode file episode_000000.parquet: 3625 rows, columns ['action', 'action_is_hold', 'observation.state', 'observation.joint_vel', 'observation.eef_pose', 'observation.environment_state', 'action.command', 'action.applied', 'narration', 'collect_info', 'extra.reward', 'extra.action_info', 'next.truncated', 'sim_step_idx', 'timestamp', 'frame_index', 'episode_index', 'index', 'task_index', 'next.done']
+--- episode file episode_000000.parquet: 3603 rows, columns ['action', 'action_is_hold', 'observation.state', 'observation.joint_vel', 'observation.eef_pose', 'observation.environment_state', 'action.command', 'action.applied', 'narration', 'collect_info', 'extra.reward', 'extra.action_info', 'next.truncated', 'sim_step_idx', 'timestamp', 'frame_index', 'episode_index', 'index', 'task_index', 'next.done']
 [PASS] observation.state shape: parquet width 15, info shape [15], 15 names first: ['dummy_base_prismatic_x_joint', 'dummy_base_prismatic_y_joint', 'dummy_base_revolute_z_joint', 'right_j0', 'head_pan', 'right_j1']
 [PASS] action shape: parquet width 11, info shape [11], 11 names first: ['dummy_base_prismatic_x_joint', 'dummy_base_prismatic_y_joint', 'dummy_base_revolute_z_joint', 'frame0.pos_x', 'frame0.pos_y', 'frame0.pos_z']
 [PASS] task text = MobileBench instruction: task_index 0: 'Go to the rack, pick up the tray with the water in it, carry it to the work table, set it on the holder, and tilt it so the water flows around the barrier into the collecting pocket.'; expected 'Go to the rack, pick up the tray with the water in it, carry it to the work table, set it on the holder, and tilt it so the water flows around the barrier into the collecting pocket.'
-rows 3625 at dataset fps 25 = 145.00 s
-[PASS] rows = recorded control steps / 2 (= episode length at 25 Hz): 3625 rows vs ceil(7250/2) = 3625
-[PASS] rows are every 2nd control step (sim_step_idx): sim_step_idx 251..7499, step differences [2]
-[PASS] video observation.images.head decodes, 640x480, not black: 86.74 MB, 3625 frames at 25 fps, mean luma 84..147, std 35, new picture 3624x = 25.0 per s
-[PASS] video observation.images.head frames = parquet rows: 3625 vs 3625
-[PASS] video observation.images.wrist decodes, 640x480, not black: 73.35 MB, 3625 frames at 25 fps, mean luma 95..173, std 31, new picture 3624x = 25.0 per s
-[PASS] video observation.images.wrist frames = parquet rows: 3625 vs 3625
-parquet size 33.62 MB
+rows 3603 at dataset fps 25 = 144.12 s
+[PASS] rows = recorded control steps / 2 (= episode length at 25 Hz): 3603 rows vs ceil(7206/2) = 3603
+[PASS] rows are every 2nd control step (sim_step_idx): sim_step_idx 251..7455, step differences [2]
+[PASS] video observation.images.head decodes, 640x480, not black: 85.11 MB, 3603 frames at 25 fps, mean luma 84..147, std 35, new picture 3602x = 25.0 per s
+[PASS] video observation.images.head frames = parquet rows: 3603 vs 3603
+[PASS] video observation.images.wrist decodes, 640x480, not black: 71.99 MB, 3603 frames at 25 fps, mean luma 95..176, std 31, new picture 3602x = 25.0 per s
+[PASS] video observation.images.wrist frames = parquet rows: 3603 vs 3603
+parquet size 33.39 MB
 [PASS] sim_info present: tiltmaze/000/sim_info
-dataset size 194.41 MB; whole cell dir 244.4 MB
+dataset size 191.20 MB; whole cell dir 240.9 MB
 RESULT: ALL PASS

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/lerobot_v21/data/chunk-000/episode_000000.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:50fae845bf18595d20a29de1d561c8a8862c66524ad411d961158cd9f9dfaa37
-size 33622556
+oid sha256:f3651408dfe991b349c6267255f705cfd3433a20eee549d2b4ac2961c13afe43
+size 33394686

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/lerobot_v21/meta/episodes.jsonl

@@ -1 +1 @@
-{"episode_index": 0, "tasks": ["Go to the rack, pick up the tray with the water in it, carry it to the work table, set it on the holder, and tilt it so the water flows around the barrier into the collecting pocket."], "length": 3625, "magicsim/control_steps": 7250, "magicsim/subsample_every": 2, "magicsim/env_id": 0, "magicsim/trajectory_id": 0, "magicsim/scene_caption": "A Ridgeback-Sawyer mobile manipulator with a UMI gripper is positioned at the workspace, facing the brown rack a0. There are a large gray table a, a large gray table b, a small tray, a large brown rack a0 and a small cradle. The gray table a is to the right of the brown rack a0. The gray table b is to the right of the cradle. The tray is on top of the brown rack a0. The tray is to the right of the brown rack a0. The brown rack a0 is to the left of the tray. The cradle is on top of the gray table b. The cradle is to the left of the gray table b. The scene is lit by a cool light from the front. The scene is viewed from front, at table height."}
+{"episode_index": 0, "tasks": ["Go to the rack, pick up the tray with the water in it, carry it to the work table, set it on the holder, and tilt it so the water flows around the barrier into the collecting pocket."], "length": 3603, "magicsim/control_steps": 7206, "magicsim/subsample_every": 2, "magicsim/env_id": 0, "magicsim/trajectory_id": 0, "magicsim/scene_caption": "A Ridgeback-Sawyer mobile manipulator with a UMI gripper is positioned at the workspace, facing the brown rack a0. There are a large gray table a, a large gray table b, a small tray, a large brown rack a0 and a small cradle. The gray table a is to the right of the brown rack a0. The gray table b is to the right of the cradle. The tray is on top of the brown rack a0. The tray is to the right of the brown rack a0. The brown rack a0 is to the left of the tray. The cradle is on top of the gray table b. The cradle is to the left of the gray table b. The scene is lit by a cool light from the front. The scene is viewed from front, at table height."}

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/lerobot_v21/meta/extra/episode_000000.json

@@ -4,7 +4,7 @@
  "env_id": 0,
  "task_type": "TiltMaze",
  "subsample_every": 2,
- "control_steps": 7250,
+ "control_steps": 7206,
  "language": {
   "scene": {
    "objects": [
@@ -72,20 +72,20 @@
      "material": null,
      "category": "rigid",
      "pos": [
-      4.0623546660645395,
-      3.4363015998022535,
-      0.6600000494609577
+      4.0623546820833,
+      3.436301865880574,
+      0.6599999898563133
      ],
      "bbox": {
       "extent": [
-       0.3150417791430822,
-       0.09843732589641087,
-       0.06200051328340228
+       0.31504177693613755,
+       0.09843749271283697,
+       0.06200036716218538
       ],
       "center": [
-       0.09523187500533548,
-       -0.003394714367880858,
-       -0.020999935369483747
+       0.09523188092357042,
+       -0.0033947363371045515,
+       -0.02099995780354391
       ]
      }
     },

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/lerobot_v21/meta/info.json

@@ -2,7 +2,7 @@
     "codebase_version": "v2.1",
     "robot_type": "RidgebackSawyerUmiCfg",
     "total_episodes": 1,
-    "total_frames": 3625,
+    "total_frames": 3603,
     "total_tasks": 1,
     "total_videos": 2,
     "total_chunks": 1,

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/lerobot_v21/meta/magicsim_manifest.json

@@ -1,7 +1,7 @@
 {
     "schema_version": 1,
     "generated": {
-        "time": "2026-11-08 23:56:45"
+        "time": "2026-11-09 18:43:07"
     },
     "embodiment": {
         "robots": [

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/lerobot_v21/videos/chunk-000/observation.images.head/episode_000000.mp4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:433350b6bd6c09e7f8f19fd441017ab9ad69d6ad0997b792595f804c6f22dca5
-size 86738662
+oid sha256:e4bd19fbcd73f4801fcdb7934ad4feda8c06fff881ef1ea5d86b5df6abb91662
+size 85114847

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/lerobot_v21/videos/chunk-000/observation.images.wrist/episode_000000.mp4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8ae635cc68369dd4446f84e7601a995fd40f10708a1b336153427f1f45a388b8
-size 73347523
+oid sha256:351e9b439b849aeb9d421f05043a14347f428754fa1bdc052188e8f5ac9c54fd
+size 71985464

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/run_meta.json

@@ -1,6 +1,6 @@
 {
     "run_id": "tiltmaze/000",
-    "created": "2026-11-08 23:34:39",
+    "created": "2026-11-09 18:25:13",
     "num_envs": 1,
     "session_id": null,
     "record_config": {

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/sim_info/0/actions.npz

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9ee6955de6591e4ffc01e5ff5a9a1e77841870c62e20300d2b608ed3bb49dc29
-size 785799
+oid sha256:9dcf11a08fc4e7f9cc69bc5a17faa409c980b33e8547c86762f80fb0263ddc94
+size 785442

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/sim_info/0/sim_states.npz

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:35854d774c4634902ccb7ba55629704a3572c631b6843088b2a1833a907ec23c
-size 47550103
+oid sha256:2a2eb4e0e97774492343485f01c03f6422e7ca08ebd9ad9c9af1af5e9ee2b620
+size 47208610

N015/sawyer_umi/dr_off/seed_5003/tiltmaze/000/sim_info/magicsim_manifest.json

@@ -1,7 +1,7 @@
 {
     "schema_version": 1,
     "generated": {
-        "time": "2026-11-08 23:57:19"
+        "time": "2026-11-09 18:43:49"
     },
     "embodiment": {
         "robots": [