{ "dataset_repo_id": "rslxcvg/banana_act_direct_color_simple_v1_molmo_compat", "policy_path": "/mnt/vla_picknplace/outputs/molmoact2/molmoact2_overnight_frombase_prod_r128_c010_rw3_w8_gpu5_20260519_fullcoverage_v1/checkpoints/010000/pretrained_model", "strict_heldout": false, "strict_heldout_note": "posthoc split only: this checkpoint was trained on the full dataset", "copies_per_original": 12, "validation_groups": [ 0, 6, 8, 9, 19, 22 ], "num_samples": 6, "num_inference_steps": 10, "first_action_mae_mean": { "shoulder_pan": 0.4816676080226898, "shoulder_lift": 0.743005096912384, "elbow_flex": 1.3183645009994507, "wrist_flex": 0.256561279296875, "wrist_roll": 0.24771015346050262, "gripper": 1.5808461904525757 }, "first_action_mae_median": { "shoulder_pan": 0.3557915985584259, "shoulder_lift": 0.57275390625, "elbow_flex": 0.6577033996582031, "wrist_flex": 0.08750534057617188, "wrist_roll": 0.17719793319702148, "gripper": 0.22944116592407227 }, "horizon_mae_mean": { "shoulder_pan": 0.561618983745575, "shoulder_lift": 1.4108734130859375, "elbow_flex": 1.955206274986267, "wrist_flex": 0.7435843348503113, "wrist_roll": 0.9763356447219849, "gripper": 1.2048346996307373 }, "horizon_mae_median": { "shoulder_pan": 0.5513197183609009, "shoulder_lift": 1.0822944641113281, "elbow_flex": 0.8897666931152344, "wrist_flex": 0.41684722900390625, "wrist_roll": 0.5, "gripper": 0.24610137939453125 }, "predicted_action_outside_dataset_stats_pct": 0.0, "gripper_timing_error_mean_frames": 6.0, "gripper_timing_error_mean_seconds": 0.2, "sample_results": [ { "original_group": 6, "episode_index": 81, "global_index": 33517, "task": "Put the banana in the green bowl.", "valid_horizon": 30, "first_action_mae": { "shoulder_pan": 0.24605417251586914, "shoulder_lift": 1.5785598754882812, "elbow_flex": 0.323486328125, "wrist_flex": 0.07439422607421875, "wrist_roll": 0.5384616851806641, "gripper": 0.026277542114257812 }, "horizon_mae": { "shoulder_pan": 0.3311326205730438, "shoulder_lift": 1.0412631034851074, "elbow_flex": 0.29274240136146545, "wrist_flex": 0.44094720482826233, "wrist_roll": 0.46538493037223816, "gripper": 0.06022198870778084 }, "pred_first_action": { "shoulder_pan": -6.048252105712891, "shoulder_lift": 107.13899993896484, "elbow_flex": 133.75205993652344, "wrist_flex": 75.54692077636719, "wrist_roll": 11.10988998413086, "gripper": 26.11323356628418 }, "target_first_action": { "shoulder_pan": -5.8021979331970215, "shoulder_lift": 105.56044006347656, "elbow_flex": 133.42857360839844, "wrist_flex": 75.47252655029297, "wrist_roll": 11.648351669311523, "gripper": 26.086956024169922 }, "gripper_event_timing_error_frames": null }, { "original_group": 9, "episode_index": 117, "global_index": 48858, "task": "Put the banana in the green bowl.", "valid_horizon": 30, "first_action_mae": { "shoulder_pan": 0.8015332221984863, "shoulder_lift": 0.573486328125, "elbow_flex": 0.526092529296875, "wrist_flex": 0.04155731201171875, "wrist_roll": 0.14285659790039062, "gripper": 0.3431997299194336 }, "horizon_mae": { "shoulder_pan": 0.6587651968002319, "shoulder_lift": 0.5098983645439148, "elbow_flex": 1.3743138313293457, "wrist_flex": 1.54385244846344, "wrist_roll": 2.102930784225464, "gripper": 1.8509775400161743 }, "pred_first_action": { "shoulder_pan": -8.36197280883789, "shoulder_lift": 100.41552734375, "elbow_flex": 121.64697265625, "wrist_flex": 96.08551025390625, "wrist_roll": 27.263734817504883, "gripper": 11.99209213256836 }, "target_first_action": { "shoulder_pan": -7.560439586639404, "shoulder_lift": 100.989013671875, "elbow_flex": 121.12088012695312, "wrist_flex": 96.04395294189453, "wrist_roll": 27.120878219604492, "gripper": 11.648892402648926 }, "gripper_event_timing_error_frames": null }, { "original_group": 19, "episode_index": 237, "global_index": 95379, "task": "Put the banana in the blue bowl.", "valid_horizon": 30, "first_action_mae": { "shoulder_pan": 0.34705835580825806, "shoulder_lift": 0.16689300537109375, "elbow_flex": 1.0927200317382812, "wrist_flex": 0.5554656982421875, "wrist_roll": 0.1318683624267578, "gripper": 0.10959285497665405 }, "horizon_mae": { "shoulder_pan": 0.8666812181472778, "shoulder_lift": 2.4227805137634277, "elbow_flex": 3.008012533187866, "wrist_flex": 0.7097671627998352, "wrist_roll": 0.11831547319889069, "gripper": 0.1705205738544464 }, "pred_first_action": { "shoulder_pan": 0.5320625305175781, "shoulder_lift": 77.96497344970703, "elbow_flex": 64.99519348144531, "wrist_flex": 81.77420806884766, "wrist_roll": 13.186813354492188, "gripper": 0.9299374222755432 }, "target_first_action": { "shoulder_pan": 0.8791208863258362, "shoulder_lift": 78.13186645507812, "elbow_flex": 66.0879135131836, "wrist_flex": 82.32967376708984, "wrist_roll": 13.05494499206543, "gripper": 0.8203445672988892 }, "gripper_event_timing_error_frames": null }, { "original_group": 22, "episode_index": 270, "global_index": 108629, "task": "Put the banana in the blue bowl.", "valid_horizon": 30, "first_action_mae": { "shoulder_pan": 0.9425656795501709, "shoulder_lift": 0.44420623779296875, "elbow_flex": 0.3433990478515625, "wrist_flex": 0.062408447265625, "wrist_roll": 0.21153926849365234, "gripper": 0.4922065734863281 }, "horizon_mae": { "shoulder_pan": 0.7259246706962585, "shoulder_lift": 1.4368120431900024, "elbow_flex": 0.5424626469612122, "wrist_flex": 0.5145024657249451, "wrist_roll": 0.18878218531608582, "gripper": 0.28711891174316406 }, "pred_first_action": { "shoulder_pan": -3.7557525634765625, "shoulder_lift": 99.67497253417969, "elbow_flex": 128.4258270263672, "wrist_flex": 74.74372863769531, "wrist_roll": 15.129119873046875, "gripper": 36.99753952026367 }, "target_first_action": { "shoulder_pan": -2.8131868839263916, "shoulder_lift": 99.23076629638672, "elbow_flex": 128.76922607421875, "wrist_flex": 74.68132019042969, "wrist_roll": 15.340659141540527, "gripper": 37.48974609375 }, "gripper_event_timing_error_frames": null }, { "original_group": 0, "episode_index": 4, "global_index": 2214, "task": "Put the banana in the blue bowl.", "valid_horizon": 30, "first_action_mae": { "shoulder_pan": 0.36452484130859375, "shoulder_lift": 0.572021484375, "elbow_flex": 0.7893142700195312, "wrist_flex": 0.704925537109375, "wrist_roll": 0.39560365676879883, "gripper": 0.11568260192871094 }, "horizon_mae": { "shoulder_pan": 0.679669976234436, "shoulder_lift": 1.5924034118652344, "elbow_flex": 1.930854082107544, "wrist_flex": 1.1070376634597778, "wrist_roll": 1.2397429943084717, "gripper": 0.2740412950515747 }, "pred_first_action": { "shoulder_pan": -40.69042205810547, "shoulder_lift": 107.80160522460938, "elbow_flex": 108.37173461914062, "wrist_flex": 75.11924743652344, "wrist_roll": 3.340658664703369, "gripper": 24.72601890563965 }, "target_first_action": { "shoulder_pan": -41.05494689941406, "shoulder_lift": 108.37362670898438, "elbow_flex": 107.5824203491211, "wrist_flex": 75.82417297363281, "wrist_roll": 2.9450550079345703, "gripper": 24.610336303710938 }, "gripper_event_timing_error_frames": null }, { "original_group": 8, "episode_index": 105, "global_index": 43686, "task": "Put the banana in the green bowl.", "valid_horizon": 30, "first_action_mae": { "shoulder_pan": 0.18826937675476074, "shoulder_lift": 1.12286376953125, "elbow_flex": 4.835174560546875, "wrist_flex": 0.100616455078125, "wrist_roll": 0.06593132019042969, "gripper": 8.398117065429688 }, "horizon_mae": { "shoulder_pan": 0.10753946006298065, "shoulder_lift": 1.4620821475982666, "elbow_flex": 4.582853317260742, "wrist_flex": 0.14539819955825806, "wrist_roll": 1.7428600788116455, "gripper": 4.586127281188965 }, "pred_first_action": { "shoulder_pan": -3.416126251220703, "shoulder_lift": 101.76022338867188, "elbow_flex": 143.27471923828125, "wrist_flex": 67.04566955566406, "wrist_roll": 28.417579650878906, "gripper": 15.637977600097656 }, "target_first_action": { "shoulder_pan": -3.604395627975464, "shoulder_lift": 100.63735961914062, "elbow_flex": 148.10989379882812, "wrist_flex": 66.94505310058594, "wrist_roll": 28.351648330688477, "gripper": 24.036094665527344 }, "gripper_event_timing_error_frames": 6 } ], "prompt_sensitivity": [ { "original_group": 6, "episode_index": 81, "global_index": 33517, "dataset_task": "Put the banana in the green bowl.", "num_prompts": 7, "pairwise_first_action_l2_mean": 0.7708305560407185, "pairwise_horizon_l2_mean": 1.9917896077746438, "pairs": [ { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana in the red bowl.", "first_action_l2": 0.7291299104690552, "horizon_mean_l2": 3.081355333328247, "mean_abs_by_joint": { "shoulder_pan": 1.1950699090957642, "shoulder_lift": 0.8464434146881104, "elbow_flex": 2.3241770267486572, "wrist_flex": 0.29658710956573486, "wrist_roll": 1.1891027688980103, "gripper": 0.17310543358325958 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana in the green bowl.", "first_action_l2": 0.0, "horizon_mean_l2": 0.0, "mean_abs_by_joint": { "shoulder_pan": 0.0, "shoulder_lift": 0.0, "elbow_flex": 0.0, "wrist_flex": 0.0, "wrist_roll": 0.0, "gripper": 0.0 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana in the blue bowl.", "first_action_l2": 0.29509198665618896, "horizon_mean_l2": 1.8509770631790161, "mean_abs_by_joint": { "shoulder_pan": 0.07995427399873734, "shoulder_lift": 0.8425447940826416, "elbow_flex": 1.2535492181777954, "wrist_flex": 0.19088593125343323, "wrist_roll": 0.9868590831756592, "gripper": 0.0782531127333641 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 0.86460280418396, "horizon_mean_l2": 1.351377248764038, "mean_abs_by_joint": { "shoulder_pan": 0.7691190242767334, "shoulder_lift": 0.44725722074508667, "elbow_flex": 0.7962422966957092, "wrist_flex": 0.4301035702228546, "wrist_roll": 0.13974355161190033, "gripper": 0.16836312413215637 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 0.9566856026649475, "horizon_mean_l2": 1.266802191734314, "mean_abs_by_joint": { "shoulder_pan": 0.3155718445777893, "shoulder_lift": 0.5268343687057495, "elbow_flex": 0.8608057498931885, "wrist_flex": 0.3330955505371094, "wrist_roll": 0.42179495096206665, "gripper": 0.24424482882022858 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 0.799412727355957, "horizon_mean_l2": 1.5466362237930298, "mean_abs_by_joint": { "shoulder_pan": 0.41816824674606323, "shoulder_lift": 0.32935282588005066, "elbow_flex": 0.8016230463981628, "wrist_flex": 1.015282154083252, "wrist_roll": 0.1487179398536682, "gripper": 0.16006343066692352 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana in the green bowl.", "first_action_l2": 0.7291299104690552, "horizon_mean_l2": 3.081355333328247, "mean_abs_by_joint": { "shoulder_pan": 1.1950699090957642, "shoulder_lift": 0.8464434146881104, "elbow_flex": 2.3241770267486572, "wrist_flex": 0.29658710956573486, "wrist_roll": 1.1891027688980103, "gripper": 0.17310543358325958 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana in the blue bowl.", "first_action_l2": 0.8239096403121948, "horizon_mean_l2": 1.8333712816238403, "mean_abs_by_joint": { "shoulder_pan": 1.2127588987350464, "shoulder_lift": 0.16240234673023224, "elbow_flex": 1.0921467542648315, "wrist_flex": 0.12656250596046448, "wrist_roll": 0.2810901701450348, "gripper": 0.09485232084989548 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 1.469856858253479, "horizon_mean_l2": 3.7783288955688477, "mean_abs_by_joint": { "shoulder_pan": 0.6070864200592041, "shoulder_lift": 1.2937006950378418, "elbow_flex": 3.1204192638397217, "wrist_flex": 0.37968647480010986, "wrist_roll": 1.3121795654296875, "gripper": 0.026084264740347862 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 1.600048542022705, "horizon_mean_l2": 3.027641773223877, "mean_abs_by_joint": { "shoulder_pan": 1.5106416940689087, "shoulder_lift": 0.7652419805526733, "elbow_flex": 1.9690943956375122, "wrist_flex": 0.252081036567688, "wrist_roll": 0.7685899138450623, "gripper": 0.07113940268754959 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 1.3973205089569092, "horizon_mean_l2": 3.50516939163208, "mean_abs_by_joint": { "shoulder_pan": 0.8023733496665955, "shoulder_lift": 0.6405169367790222, "elbow_flex": 2.5985560417175293, "wrist_flex": 0.8619471192359924, "wrist_roll": 1.2750002145767212, "gripper": 0.039126649498939514 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana in the blue bowl.", "first_action_l2": 0.29509198665618896, "horizon_mean_l2": 1.8509770631790161, "mean_abs_by_joint": { "shoulder_pan": 0.07995427399873734, "shoulder_lift": 0.8425447940826416, "elbow_flex": 1.2535492181777954, "wrist_flex": 0.19088593125343323, "wrist_roll": 0.9868590831756592, "gripper": 0.0782531127333641 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 0.86460280418396, "horizon_mean_l2": 1.351377248764038, "mean_abs_by_joint": { "shoulder_pan": 0.7691190242767334, "shoulder_lift": 0.44725722074508667, "elbow_flex": 0.7962422966957092, "wrist_flex": 0.4301035702228546, "wrist_roll": 0.13974355161190033, "gripper": 0.16836312413215637 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 0.9566856026649475, "horizon_mean_l2": 1.266802191734314, "mean_abs_by_joint": { "shoulder_pan": 0.3155718445777893, "shoulder_lift": 0.5268343687057495, "elbow_flex": 0.8608057498931885, "wrist_flex": 0.3330955505371094, "wrist_roll": 0.42179495096206665, "gripper": 0.24424482882022858 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 0.799412727355957, "horizon_mean_l2": 1.5466362237930298, "mean_abs_by_joint": { "shoulder_pan": 0.41816824674606323, "shoulder_lift": 0.32935282588005066, "elbow_flex": 0.8016230463981628, "wrist_flex": 1.015282154083252, "wrist_roll": 0.1487179398536682, "gripper": 0.16006343066692352 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 0.8180261254310608, "horizon_mean_l2": 2.8715333938598633, "mean_abs_by_joint": { "shoulder_pan": 0.7868080139160156, "shoulder_lift": 1.289802074432373, "elbow_flex": 2.0282723903656006, "wrist_flex": 0.40750324726104736, "wrist_roll": 1.109935998916626, "gripper": 0.09722404181957245 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 0.842451274394989, "horizon_mean_l2": 1.5129761695861816, "mean_abs_by_joint": { "shoulder_pan": 0.2978828549385071, "shoulder_lift": 0.6833907961845398, "elbow_flex": 0.8984664678573608, "wrist_flex": 0.2729443907737732, "wrist_roll": 0.5663462281227112, "gripper": 0.16599172353744507 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 0.6663229465484619, "horizon_mean_l2": 2.462028741836548, "mean_abs_by_joint": { "shoulder_pan": 0.444348007440567, "shoulder_lift": 0.6210273504257202, "elbow_flex": 1.5064091682434082, "wrist_flex": 0.8994989991188049, "wrist_roll": 1.0727565288543701, "gripper": 0.09129568934440613 } }, { "prompt_a": "Put the banana into the leftmost bowl from the robot perspective.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 0.475327730178833, "horizon_mean_l2": 1.950879454612732, "mean_abs_by_joint": { "shoulder_pan": 1.084690809249878, "shoulder_lift": 0.6668248772621155, "elbow_flex": 1.205127477645874, "wrist_flex": 0.16585209965705872, "wrist_roll": 0.5435898303985596, "gripper": 0.07825291901826859 } }, { "prompt_a": "Put the banana into the leftmost bowl from the robot perspective.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 0.4021580219268799, "horizon_mean_l2": 1.2282354831695557, "mean_abs_by_joint": { "shoulder_pan": 0.3509507477283478, "shoulder_lift": 0.6687746644020081, "elbow_flex": 0.5971826910972595, "wrist_flex": 0.6463735103607178, "wrist_roll": 0.050000064074993134, "gripper": 0.02727019041776657 } }, { "prompt_a": "Put the banana into the middle bowl from the robot perspective.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 0.4021739661693573, "horizon_mean_l2": 1.4631210565567017, "mean_abs_by_joint": { "shoulder_pan": 0.7337400913238525, "shoulder_lift": 0.2078758180141449, "elbow_flex": 0.672503650188446, "wrist_flex": 0.687749981880188, "wrist_roll": 0.5076924562454224, "gripper": 0.09129530936479568 } } ] }, { "original_group": 9, "episode_index": 117, "global_index": 48858, "dataset_task": "Put the banana in the green bowl.", "num_prompts": 7, "pairwise_first_action_l2_mean": 1.1516095002492268, "pairwise_horizon_l2_mean": 3.2861017215819586, "pairs": [ { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana in the red bowl.", "first_action_l2": 0.42190471291542053, "horizon_mean_l2": 2.222421169281006, "mean_abs_by_joint": { "shoulder_pan": 1.3236693143844604, "shoulder_lift": 0.3163599669933319, "elbow_flex": 0.6240832209587097, "wrist_flex": 0.8400408625602722, "wrist_roll": 0.6538460850715637, "gripper": 1.1189610958099365 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana in the green bowl.", "first_action_l2": 0.0, "horizon_mean_l2": 0.0, "mean_abs_by_joint": { "shoulder_pan": 0.0, "shoulder_lift": 0.0, "elbow_flex": 0.0, "wrist_flex": 0.0, "wrist_roll": 0.0, "gripper": 0.0 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana in the blue bowl.", "first_action_l2": 0.9664211869239807, "horizon_mean_l2": 2.9833662509918213, "mean_abs_by_joint": { "shoulder_pan": 0.8563252687454224, "shoulder_lift": 0.36183294653892517, "elbow_flex": 1.7646515369415283, "wrist_flex": 2.0333445072174072, "wrist_roll": 0.21282044053077698, "gripper": 0.460182249546051 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 1.7762058973312378, "horizon_mean_l2": 4.269879341125488, "mean_abs_by_joint": { "shoulder_pan": 0.17052243649959564, "shoulder_lift": 1.0172892808914185, "elbow_flex": 2.173537254333496, "wrist_flex": 2.714834451675415, "wrist_roll": 0.6615384221076965, "gripper": 1.930096983909607 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 1.7758368253707886, "horizon_mean_l2": 4.703806400299072, "mean_abs_by_joint": { "shoulder_pan": 0.6847422122955322, "shoulder_lift": 0.45602670311927795, "elbow_flex": 2.595869302749634, "wrist_flex": 3.3490383625030518, "wrist_roll": 1.4564100503921509, "gripper": 0.8078753352165222 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 1.6803284883499146, "horizon_mean_l2": 5.498891353607178, "mean_abs_by_joint": { "shoulder_pan": 0.9444168210029602, "shoulder_lift": 0.8055175542831421, "elbow_flex": 2.7868590354919434, "wrist_flex": 4.0458269119262695, "wrist_roll": 0.9794867634773254, "gripper": 1.5277156829833984 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana in the green bowl.", "first_action_l2": 0.42190471291542053, "horizon_mean_l2": 2.222421169281006, "mean_abs_by_joint": { "shoulder_pan": 1.3236693143844604, "shoulder_lift": 0.3163599669933319, "elbow_flex": 0.6240832209587097, "wrist_flex": 0.8400408625602722, "wrist_roll": 0.6538460850715637, "gripper": 1.1189610958099365 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana in the blue bowl.", "first_action_l2": 0.9940071702003479, "horizon_mean_l2": 3.030785083770752, "mean_abs_by_joint": { "shoulder_pan": 0.4807872772216797, "shoulder_lift": 0.6366187930107117, "elbow_flex": 2.3887345790863037, "wrist_flex": 1.1933037042617798, "wrist_roll": 0.861538290977478, "gripper": 0.6587788462638855 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 1.8483214378356934, "horizon_mean_l2": 4.234662055969238, "mean_abs_by_joint": { "shoulder_pan": 1.2348698377609253, "shoulder_lift": 1.315460205078125, "elbow_flex": 2.7976202964782715, "wrist_flex": 1.874793529510498, "wrist_roll": 1.3153845071792603, "gripper": 0.8991705775260925 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 1.9467982053756714, "horizon_mean_l2": 4.781572341918945, "mean_abs_by_joint": { "shoulder_pan": 0.6792578101158142, "shoulder_lift": 0.733410120010376, "elbow_flex": 3.219952344894409, "wrist_flex": 2.508997678756714, "wrist_roll": 2.1102561950683594, "gripper": 0.45751431584358215 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 1.8196462392807007, "horizon_mean_l2": 5.253678798675537, "mean_abs_by_joint": { "shoulder_pan": 0.37925249338150024, "shoulder_lift": 1.1036885976791382, "elbow_flex": 3.410942316055298, "wrist_flex": 3.2057862281799316, "wrist_roll": 1.6333329677581787, "gripper": 1.0270729064941406 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana in the blue bowl.", "first_action_l2": 0.9664211869239807, "horizon_mean_l2": 2.9833662509918213, "mean_abs_by_joint": { "shoulder_pan": 0.8563252687454224, "shoulder_lift": 0.36183294653892517, "elbow_flex": 1.7646515369415283, "wrist_flex": 2.0333445072174072, "wrist_roll": 0.21282044053077698, "gripper": 0.460182249546051 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 1.7762058973312378, "horizon_mean_l2": 4.269879341125488, "mean_abs_by_joint": { "shoulder_pan": 0.17052243649959564, "shoulder_lift": 1.0172892808914185, "elbow_flex": 2.173537254333496, "wrist_flex": 2.714834451675415, "wrist_roll": 0.6615384221076965, "gripper": 1.930096983909607 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 1.7758368253707886, "horizon_mean_l2": 4.703806400299072, "mean_abs_by_joint": { "shoulder_pan": 0.6847422122955322, "shoulder_lift": 0.45602670311927795, "elbow_flex": 2.595869302749634, "wrist_flex": 3.3490383625030518, "wrist_roll": 1.4564100503921509, "gripper": 0.8078753352165222 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 1.6803284883499146, "horizon_mean_l2": 5.498891353607178, "mean_abs_by_joint": { "shoulder_pan": 0.9444168210029602, "shoulder_lift": 0.8055175542831421, "elbow_flex": 2.7868590354919434, "wrist_flex": 4.0458269119262695, "wrist_roll": 0.9794867634773254, "gripper": 1.5277156829833984 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 0.8892030119895935, "horizon_mean_l2": 2.1461997032165527, "mean_abs_by_joint": { "shoulder_pan": 0.7700023651123047, "shoulder_lift": 0.6788414120674133, "elbow_flex": 0.4088856875896454, "wrist_flex": 0.6814898252487183, "wrist_roll": 0.4589744508266449, "gripper": 1.484142541885376 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 1.105545163154602, "horizon_mean_l2": 2.1361703872680664, "mean_abs_by_joint": { "shoulder_pan": 0.20979157090187073, "shoulder_lift": 0.10588658601045609, "elbow_flex": 0.8312177062034607, "wrist_flex": 1.315693974494934, "wrist_roll": 1.2487179040908813, "gripper": 0.5341373085975647 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 0.9491774439811707, "horizon_mean_l2": 2.8086187839508057, "mean_abs_by_joint": { "shoulder_pan": 0.18361206352710724, "shoulder_lift": 0.46706974506378174, "elbow_flex": 1.0222076177597046, "wrist_flex": 2.0124824047088623, "wrist_roll": 0.7717945575714111, "gripper": 1.264648675918579 } }, { "prompt_a": "Put the banana into the leftmost bowl from the robot perspective.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 0.5812988877296448, "horizon_mean_l2": 1.8941317796707153, "mean_abs_by_joint": { "shoulder_pan": 0.5955890417098999, "shoulder_lift": 0.582050085067749, "elbow_flex": 0.4223319888114929, "wrist_flex": 0.634204089641571, "wrist_roll": 0.7948716282844543, "gripper": 1.2686500549316406 } }, { "prompt_a": "Put the banana into the leftmost bowl from the robot perspective.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 0.43362265825271606, "horizon_mean_l2": 1.9810305833816528, "mean_abs_by_joint": { "shoulder_pan": 0.8718916773796082, "shoulder_lift": 0.2117716521024704, "elbow_flex": 0.6240819096565247, "wrist_flex": 1.330992579460144, "wrist_roll": 0.3179483413696289, "gripper": 0.6383266448974609 } }, { "prompt_a": "Put the banana into the middle bowl from the robot perspective.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 0.37478506565093994, "horizon_mean_l2": 1.3845576047897339, "mean_abs_by_joint": { "shoulder_pan": 0.33043086528778076, "shoulder_lift": 0.37027841806411743, "elbow_flex": 0.22865015268325806, "wrist_flex": 0.6967885494232178, "wrist_roll": 0.4769233167171478, "gripper": 0.7681558132171631 } } ] }, { "original_group": 19, "episode_index": 237, "global_index": 95379, "dataset_task": "Put the banana in the blue bowl.", "num_prompts": 7, "pairwise_first_action_l2_mean": 9.246774928910392, "pairwise_horizon_l2_mean": 21.601163563274202, "pairs": [ { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana in the red bowl.", "first_action_l2": 16.408475875854492, "horizon_mean_l2": 33.13117599487305, "mean_abs_by_joint": { "shoulder_pan": 18.49422264099121, "shoulder_lift": 13.979626655578613, "elbow_flex": 23.204099655151367, "wrist_flex": 1.6198869943618774, "wrist_roll": 2.067307710647583, "gripper": 0.8347006440162659 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana in the green bowl.", "first_action_l2": 11.759034156799316, "horizon_mean_l2": 31.617958068847656, "mean_abs_by_joint": { "shoulder_pan": 25.301660537719727, "shoulder_lift": 13.86269474029541, "elbow_flex": 6.82726526260376, "wrist_flex": 8.200093269348145, "wrist_roll": 5.788460731506348, "gripper": 0.10433757305145264 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana in the blue bowl.", "first_action_l2": 0.0, "horizon_mean_l2": 0.0, "mean_abs_by_joint": { "shoulder_pan": 0.0, "shoulder_lift": 0.0, "elbow_flex": 0.0, "wrist_flex": 0.0, "wrist_roll": 0.0, "gripper": 0.0 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 1.8155051469802856, "horizon_mean_l2": 2.1692354679107666, "mean_abs_by_joint": { "shoulder_pan": 1.354271411895752, "shoulder_lift": 1.0601651668548584, "elbow_flex": 0.39274317026138306, "wrist_flex": 0.2578623592853546, "wrist_roll": 0.15897433459758759, "gripper": 0.0972236841917038 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 0.8035499453544617, "horizon_mean_l2": 1.5898617506027222, "mean_abs_by_joint": { "shoulder_pan": 1.494368076324463, "shoulder_lift": 0.12472584843635559, "elbow_flex": 0.03766059875488281, "wrist_flex": 0.45135751366615295, "wrist_roll": 0.11666679382324219, "gripper": 0.047426190227270126 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 6.591071128845215, "horizon_mean_l2": 17.99409294128418, "mean_abs_by_joint": { "shoulder_pan": 15.525296211242676, "shoulder_lift": 5.030585289001465, "elbow_flex": 6.235463619232178, "wrist_flex": 2.8844690322875977, "wrist_roll": 1.8884618282318115, "gripper": 0.0853670984506607 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana in the green bowl.", "first_action_l2": 18.345474243164062, "horizon_mean_l2": 47.74506759643555, "mean_abs_by_joint": { "shoulder_pan": 43.79587936401367, "shoulder_lift": 0.7405560612678528, "elbow_flex": 16.3768310546875, "wrist_flex": 7.541248798370361, "wrist_roll": 3.721153974533081, "gripper": 0.9390382170677185 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana in the blue bowl.", "first_action_l2": 16.408475875854492, "horizon_mean_l2": 33.13117599487305, "mean_abs_by_joint": { "shoulder_pan": 18.49422264099121, "shoulder_lift": 13.979626655578613, "elbow_flex": 23.204099655151367, "wrist_flex": 1.6198869943618774, "wrist_roll": 2.067307710647583, "gripper": 0.8347006440162659 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 18.07362174987793, "horizon_mean_l2": 34.6060791015625, "mean_abs_by_joint": { "shoulder_pan": 19.84848976135254, "shoulder_lift": 14.665614128112793, "elbow_flex": 23.596839904785156, "wrist_flex": 1.777611494064331, "wrist_roll": 1.9083333015441895, "gripper": 0.7374768853187561 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 16.867815017700195, "horizon_mean_l2": 34.09500503540039, "mean_abs_by_joint": { "shoulder_pan": 19.988588333129883, "shoulder_lift": 14.104351997375488, "elbow_flex": 23.230998992919922, "wrist_flex": 1.8389809131622314, "wrist_roll": 2.046794891357422, "gripper": 0.7872743010520935 } }, { "prompt_a": "Put the banana in the red bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 11.46985149383545, "horizon_mean_l2": 19.616174697875977, "mean_abs_by_joint": { "shoulder_pan": 2.9689245223999023, "shoulder_lift": 8.9490385055542, "elbow_flex": 16.9686336517334, "wrist_flex": 2.2256228923797607, "wrist_roll": 0.8211538195610046, "gripper": 0.9200678467750549 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana in the blue bowl.", "first_action_l2": 11.759034156799316, "horizon_mean_l2": 31.617958068847656, "mean_abs_by_joint": { "shoulder_pan": 25.301660537719727, "shoulder_lift": 13.86269474029541, "elbow_flex": 6.82726526260376, "wrist_flex": 8.200093269348145, "wrist_roll": 5.788460731506348, "gripper": 0.10433757305145264 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 12.74959945678711, "horizon_mean_l2": 31.08345603942871, "mean_abs_by_joint": { "shoulder_pan": 23.947389602661133, "shoulder_lift": 14.548685073852539, "elbow_flex": 7.2200093269348145, "wrist_flex": 8.382157325744629, "wrist_roll": 5.629487037658691, "gripper": 0.20156127214431763 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 11.297257423400879, "horizon_mean_l2": 30.63994598388672, "mean_abs_by_joint": { "shoulder_pan": 23.80729103088379, "shoulder_lift": 13.987420082092285, "elbow_flex": 6.854165554046631, "wrist_flex": 8.65145206451416, "wrist_roll": 5.767947196960449, "gripper": 0.15176376700401306 } }, { "prompt_a": "Put the banana in the green bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 13.778914451599121, "horizon_mean_l2": 42.333595275878906, "mean_abs_by_joint": { "shoulder_pan": 40.826961517333984, "shoulder_lift": 8.832108497619629, "elbow_flex": 0.7639650702476501, "wrist_flex": 5.315625190734863, "wrist_roll": 3.8999998569488525, "gripper": 0.06165401265025139 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the leftmost bowl from the robot perspective.", "first_action_l2": 1.8155051469802856, "horizon_mean_l2": 2.1692354679107666, "mean_abs_by_joint": { "shoulder_pan": 1.354271411895752, "shoulder_lift": 1.0601651668548584, "elbow_flex": 0.39274317026138306, "wrist_flex": 0.2578623592853546, "wrist_roll": 0.15897433459758759, "gripper": 0.0972236841917038 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 0.8035499453544617, "horizon_mean_l2": 1.5898617506027222, "mean_abs_by_joint": { "shoulder_pan": 1.494368076324463, "shoulder_lift": 0.12472584843635559, "elbow_flex": 0.03766059875488281, "wrist_flex": 0.45135751366615295, "wrist_roll": 0.11666679382324219, "gripper": 0.047426190227270126 } }, { "prompt_a": "Put the banana in the blue bowl.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 6.591071128845215, "horizon_mean_l2": 17.99409294128418, "mean_abs_by_joint": { "shoulder_pan": 15.525296211242676, "shoulder_lift": 5.030585289001465, "elbow_flex": 6.235463619232178, "wrist_flex": 2.8844690322875977, "wrist_roll": 1.8884618282318115, "gripper": 0.0853670984506607 } }, { "prompt_a": "Put the banana into the leftmost bowl from the robot perspective.", "prompt_b": "Put the banana into the middle bowl from the robot perspective.", "first_action_l2": 1.7892165184020996, "horizon_mean_l2": 1.3498412370681763, "mean_abs_by_joint": { "shoulder_pan": 0.5278403162956238, "shoulder_lift": 0.977014422416687, "elbow_flex": 0.3658425509929657, "wrist_flex": 0.30232492089271545, "wrist_roll": 0.1384614259004593, "gripper": 0.049797482788562775 } }, { "prompt_a": "Put the banana into the leftmost bowl from the robot perspective.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 7.822325706481934, "horizon_mean_l2": 19.723468780517578, "mean_abs_by_joint": { "shoulder_pan": 16.879568099975586, "shoulder_lift": 5.716574668884277, "elbow_flex": 6.628204822540283, "wrist_flex": 3.0665323734283447, "wrist_roll": 1.7294870615005493, "gripper": 0.1825907826423645 } }, { "prompt_a": "Put the banana into the middle bowl from the robot perspective.", "prompt_b": "Put the banana into the rightmost bowl from the robot perspective.", "first_action_l2": 7.232924938201904, "horizon_mean_l2": 19.427152633666992, "mean_abs_by_joint": { "shoulder_pan": 17.01966094970703, "shoulder_lift": 5.155311584472656, "elbow_flex": 6.262363910675049, "wrist_flex": 3.3358266353607178, "wrist_roll": 1.8679486513137817, "gripper": 0.13279329240322113 } } ] } ], "policy_config": { "type": "molmoact2", "chunk_size": 30, "n_action_steps": 30, "image_keys": [ "observation.images.front" ], "norm_tag": "so100_so101_molmoact2", "action_mode": "continuous" } }