{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 44.96402877697842, "eval_steps": 500, "global_step": 50000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "action_loss": 2.971473455429077, "epoch": 0, "step": 0 }, { "epoch": 0, "step": 0, "torque_loss": 0.21728301048278809 }, { "epoch": 0.008992805755395683, "grad_norm": 8.976465225219727, "learning_rate": 3.6e-07, "loss": 1.8469, "step": 10 }, { "action_loss": 1.3281450271606445, "epoch": 0.008992805755395683, "step": 10 }, { "epoch": 0.008992805755395683, "step": 10, "torque_loss": 0.15651147067546844 }, { "epoch": 0.017985611510791366, "grad_norm": 21.398143768310547, "learning_rate": 7.6e-07, "loss": 1.8354, "step": 20 }, { "action_loss": 1.6014081239700317, "epoch": 0.017985611510791366, "step": 20 }, { "epoch": 0.017985611510791366, "step": 20, "torque_loss": 0.15294383466243744 }, { "epoch": 0.02697841726618705, "grad_norm": 8.020687103271484, "learning_rate": 1.16e-06, "loss": 1.768, "step": 30 }, { "action_loss": 1.768692970275879, "epoch": 0.02697841726618705, "step": 30 }, { "epoch": 0.02697841726618705, "step": 30, "torque_loss": 0.17870111763477325 }, { "epoch": 0.03597122302158273, "grad_norm": 7.145173072814941, "learning_rate": 1.56e-06, "loss": 1.5625, "step": 40 }, { "action_loss": 1.8416624069213867, "epoch": 0.03597122302158273, "step": 40 }, { "epoch": 0.03597122302158273, "step": 40, "torque_loss": 0.1616661548614502 }, { "epoch": 0.044964028776978415, "grad_norm": 7.32466459274292, "learning_rate": 1.96e-06, "loss": 1.4602, "step": 50 }, { "action_loss": 1.3789459466934204, "epoch": 0.044964028776978415, "step": 50 }, { "epoch": 0.044964028776978415, "step": 50, "torque_loss": 0.19139915704727173 }, { "epoch": 0.0539568345323741, "grad_norm": 9.664463996887207, "learning_rate": 2.36e-06, "loss": 1.3059, "step": 60 }, { "action_loss": 0.7062929272651672, "epoch": 0.0539568345323741, "step": 60 }, { "epoch": 0.0539568345323741, "step": 60, "torque_loss": 0.14102481305599213 }, { "epoch": 0.06294964028776978, "grad_norm": 4.5777692794799805, "learning_rate": 2.7600000000000003e-06, "loss": 0.8098, "step": 70 }, { "action_loss": 0.7226126790046692, "epoch": 0.06294964028776978, "step": 70 }, { "epoch": 0.06294964028776978, "step": 70, "torque_loss": 0.17307864129543304 }, { "epoch": 0.07194244604316546, "grad_norm": 2.3523287773132324, "learning_rate": 3.1600000000000007e-06, "loss": 0.7456, "step": 80 }, { "action_loss": 0.6359246373176575, "epoch": 0.07194244604316546, "step": 80 }, { "epoch": 0.07194244604316546, "step": 80, "torque_loss": 0.20337752997875214 }, { "epoch": 0.08093525179856115, "grad_norm": 1.143373966217041, "learning_rate": 3.5600000000000002e-06, "loss": 0.5777, "step": 90 }, { "action_loss": 0.4115051329135895, "epoch": 0.08093525179856115, "step": 90 }, { "epoch": 0.08093525179856115, "step": 90, "torque_loss": 0.10164079815149307 }, { "epoch": 0.08992805755395683, "grad_norm": 2.410313844680786, "learning_rate": 3.96e-06, "loss": 0.4894, "step": 100 }, { "action_loss": 0.4393438398838043, "epoch": 0.08992805755395683, "step": 100 }, { "epoch": 0.08992805755395683, "step": 100, "torque_loss": 0.18885980546474457 }, { "epoch": 0.09892086330935251, "grad_norm": 1.24308180809021, "learning_rate": 4.360000000000001e-06, "loss": 0.4022, "step": 110 }, { "action_loss": 0.3656748831272125, "epoch": 0.09892086330935251, "step": 110 }, { "epoch": 0.09892086330935251, "step": 110, "torque_loss": 0.2100505381822586 }, { "epoch": 0.1079136690647482, "grad_norm": 2.561617612838745, "learning_rate": 4.76e-06, "loss": 0.3399, "step": 120 }, { "action_loss": 0.2629702389240265, "epoch": 0.1079136690647482, "step": 120 }, { "epoch": 0.1079136690647482, "step": 120, "torque_loss": 0.24143119156360626 }, { "epoch": 0.11690647482014388, "grad_norm": 0.7322325706481934, "learning_rate": 5.1600000000000006e-06, "loss": 0.2335, "step": 130 }, { "action_loss": 0.22311842441558838, "epoch": 0.11690647482014388, "step": 130 }, { "epoch": 0.11690647482014388, "step": 130, "torque_loss": 0.17302103340625763 }, { "epoch": 0.12589928057553956, "grad_norm": 0.9109329581260681, "learning_rate": 5.56e-06, "loss": 0.1819, "step": 140 }, { "action_loss": 0.14986270666122437, "epoch": 0.12589928057553956, "step": 140 }, { "epoch": 0.12589928057553956, "step": 140, "torque_loss": 0.16548623144626617 }, { "epoch": 0.13489208633093525, "grad_norm": 1.1349577903747559, "learning_rate": 5.9600000000000005e-06, "loss": 0.1784, "step": 150 }, { "action_loss": 0.25546330213546753, "epoch": 0.13489208633093525, "step": 150 }, { "epoch": 0.13489208633093525, "step": 150, "torque_loss": 0.22170548141002655 }, { "epoch": 0.14388489208633093, "grad_norm": 0.5102756023406982, "learning_rate": 6.360000000000001e-06, "loss": 0.1493, "step": 160 }, { "action_loss": 0.17076347768306732, "epoch": 0.14388489208633093, "step": 160 }, { "epoch": 0.14388489208633093, "step": 160, "torque_loss": 0.21280024945735931 }, { "epoch": 0.1528776978417266, "grad_norm": 0.8614314794540405, "learning_rate": 6.76e-06, "loss": 0.1314, "step": 170 }, { "action_loss": 0.08856940269470215, "epoch": 0.1528776978417266, "step": 170 }, { "epoch": 0.1528776978417266, "step": 170, "torque_loss": 0.1249815821647644 }, { "epoch": 0.1618705035971223, "grad_norm": 0.5325682759284973, "learning_rate": 7.16e-06, "loss": 0.1054, "step": 180 }, { "action_loss": 0.10207352787256241, "epoch": 0.1618705035971223, "step": 180 }, { "epoch": 0.1618705035971223, "step": 180, "torque_loss": 0.19481056928634644 }, { "epoch": 0.17086330935251798, "grad_norm": 0.703822672367096, "learning_rate": 7.5600000000000005e-06, "loss": 0.1205, "step": 190 }, { "action_loss": 0.0775100514292717, "epoch": 0.17086330935251798, "step": 190 }, { "epoch": 0.17086330935251798, "step": 190, "torque_loss": 0.17293672263622284 }, { "epoch": 0.17985611510791366, "grad_norm": 1.1393946409225464, "learning_rate": 7.96e-06, "loss": 0.1205, "step": 200 }, { "action_loss": 0.05081398785114288, "epoch": 0.17985611510791366, "step": 200 }, { "epoch": 0.17985611510791366, "step": 200, "torque_loss": 0.10418325662612915 }, { "epoch": 0.18884892086330934, "grad_norm": 0.8085319995880127, "learning_rate": 8.36e-06, "loss": 0.107, "step": 210 }, { "action_loss": 0.1355329304933548, "epoch": 0.18884892086330934, "step": 210 }, { "epoch": 0.18884892086330934, "step": 210, "torque_loss": 0.27171429991722107 }, { "epoch": 0.19784172661870503, "grad_norm": 0.6921412348747253, "learning_rate": 8.76e-06, "loss": 0.1082, "step": 220 }, { "action_loss": 0.06197875738143921, "epoch": 0.19784172661870503, "step": 220 }, { "epoch": 0.19784172661870503, "step": 220, "torque_loss": 0.18819157779216766 }, { "epoch": 0.2068345323741007, "grad_norm": 0.4388245940208435, "learning_rate": 9.16e-06, "loss": 0.0952, "step": 230 }, { "action_loss": 0.1012946143746376, "epoch": 0.2068345323741007, "step": 230 }, { "epoch": 0.2068345323741007, "step": 230, "torque_loss": 0.25180569291114807 }, { "epoch": 0.2158273381294964, "grad_norm": 0.8343818783760071, "learning_rate": 9.560000000000002e-06, "loss": 0.0967, "step": 240 }, { "action_loss": 0.11997269839048386, "epoch": 0.2158273381294964, "step": 240 }, { "epoch": 0.2158273381294964, "step": 240, "torque_loss": 0.1928635686635971 }, { "epoch": 0.22482014388489208, "grad_norm": 0.7615068554878235, "learning_rate": 9.96e-06, "loss": 0.0935, "step": 250 }, { "action_loss": 0.07618842273950577, "epoch": 0.22482014388489208, "step": 250 }, { "epoch": 0.22482014388489208, "step": 250, "torque_loss": 0.19423411786556244 }, { "epoch": 0.23381294964028776, "grad_norm": 0.9957984089851379, "learning_rate": 1.036e-05, "loss": 0.0944, "step": 260 }, { "action_loss": 0.05421711876988411, "epoch": 0.23381294964028776, "step": 260 }, { "epoch": 0.23381294964028776, "step": 260, "torque_loss": 0.15560798346996307 }, { "epoch": 0.24280575539568344, "grad_norm": 0.48495063185691833, "learning_rate": 1.076e-05, "loss": 0.0842, "step": 270 }, { "action_loss": 0.0696009024977684, "epoch": 0.24280575539568344, "step": 270 }, { "epoch": 0.24280575539568344, "step": 270, "torque_loss": 0.2156287431716919 }, { "epoch": 0.2517985611510791, "grad_norm": 0.7869534492492676, "learning_rate": 1.1160000000000002e-05, "loss": 0.0942, "step": 280 }, { "action_loss": 0.049339234828948975, "epoch": 0.2517985611510791, "step": 280 }, { "epoch": 0.2517985611510791, "step": 280, "torque_loss": 0.16389162838459015 }, { "epoch": 0.2607913669064748, "grad_norm": 0.6994296908378601, "learning_rate": 1.156e-05, "loss": 0.0747, "step": 290 }, { "action_loss": 0.06036926433444023, "epoch": 0.2607913669064748, "step": 290 }, { "epoch": 0.2607913669064748, "step": 290, "torque_loss": 0.2176186442375183 }, { "epoch": 0.2697841726618705, "grad_norm": 0.6179289221763611, "learning_rate": 1.196e-05, "loss": 0.0868, "step": 300 }, { "action_loss": 0.060035090893507004, "epoch": 0.2697841726618705, "step": 300 }, { "epoch": 0.2697841726618705, "step": 300, "torque_loss": 0.18307869136333466 }, { "epoch": 0.2787769784172662, "grad_norm": 0.7280067801475525, "learning_rate": 1.236e-05, "loss": 0.0771, "step": 310 }, { "action_loss": 0.10195493698120117, "epoch": 0.2787769784172662, "step": 310 }, { "epoch": 0.2787769784172662, "step": 310, "torque_loss": 0.21963100135326385 }, { "epoch": 0.28776978417266186, "grad_norm": 0.5775085687637329, "learning_rate": 1.276e-05, "loss": 0.0899, "step": 320 }, { "action_loss": 0.045851659029722214, "epoch": 0.28776978417266186, "step": 320 }, { "epoch": 0.28776978417266186, "step": 320, "torque_loss": 0.11048737913370132 }, { "epoch": 0.29676258992805754, "grad_norm": 0.7053504586219788, "learning_rate": 1.316e-05, "loss": 0.0793, "step": 330 }, { "action_loss": 0.052978258579969406, "epoch": 0.29676258992805754, "step": 330 }, { "epoch": 0.29676258992805754, "step": 330, "torque_loss": 0.11982706189155579 }, { "epoch": 0.3057553956834532, "grad_norm": 0.6267995238304138, "learning_rate": 1.356e-05, "loss": 0.0737, "step": 340 }, { "action_loss": 0.04208143427968025, "epoch": 0.3057553956834532, "step": 340 }, { "epoch": 0.3057553956834532, "step": 340, "torque_loss": 0.18034608662128448 }, { "epoch": 0.3147482014388489, "grad_norm": 0.7661734819412231, "learning_rate": 1.396e-05, "loss": 0.0797, "step": 350 }, { "action_loss": 0.050797004252672195, "epoch": 0.3147482014388489, "step": 350 }, { "epoch": 0.3147482014388489, "step": 350, "torque_loss": 0.17005570232868195 }, { "epoch": 0.3237410071942446, "grad_norm": 0.7015149593353271, "learning_rate": 1.4360000000000001e-05, "loss": 0.0774, "step": 360 }, { "action_loss": 0.04383273050189018, "epoch": 0.3237410071942446, "step": 360 }, { "epoch": 0.3237410071942446, "step": 360, "torque_loss": 0.19450370967388153 }, { "epoch": 0.3327338129496403, "grad_norm": 0.532209575176239, "learning_rate": 1.4760000000000001e-05, "loss": 0.0628, "step": 370 }, { "action_loss": 0.06142913177609444, "epoch": 0.3327338129496403, "step": 370 }, { "epoch": 0.3327338129496403, "step": 370, "torque_loss": 0.2043837308883667 }, { "epoch": 0.34172661870503596, "grad_norm": 0.3701375126838684, "learning_rate": 1.5160000000000002e-05, "loss": 0.0689, "step": 380 }, { "action_loss": 0.03120686672627926, "epoch": 0.34172661870503596, "step": 380 }, { "epoch": 0.34172661870503596, "step": 380, "torque_loss": 0.1804366558790207 }, { "epoch": 0.35071942446043164, "grad_norm": 0.6868853569030762, "learning_rate": 1.556e-05, "loss": 0.0855, "step": 390 }, { "action_loss": 0.025186220183968544, "epoch": 0.35071942446043164, "step": 390 }, { "epoch": 0.35071942446043164, "step": 390, "torque_loss": 0.15967240929603577 }, { "epoch": 0.3597122302158273, "grad_norm": 0.5015292763710022, "learning_rate": 1.596e-05, "loss": 0.0629, "step": 400 }, { "action_loss": 0.07168758660554886, "epoch": 0.3597122302158273, "step": 400 }, { "epoch": 0.3597122302158273, "step": 400, "torque_loss": 0.17054295539855957 }, { "epoch": 0.368705035971223, "grad_norm": 0.610634982585907, "learning_rate": 1.636e-05, "loss": 0.0741, "step": 410 }, { "action_loss": 0.04305843636393547, "epoch": 0.368705035971223, "step": 410 }, { "epoch": 0.368705035971223, "step": 410, "torque_loss": 0.2354564666748047 }, { "epoch": 0.3776978417266187, "grad_norm": 0.7810073494911194, "learning_rate": 1.6760000000000002e-05, "loss": 0.0783, "step": 420 }, { "action_loss": 0.059468742460012436, "epoch": 0.3776978417266187, "step": 420 }, { "epoch": 0.3776978417266187, "step": 420, "torque_loss": 0.18569517135620117 }, { "epoch": 0.38669064748201437, "grad_norm": 0.6635499596595764, "learning_rate": 1.7160000000000002e-05, "loss": 0.0712, "step": 430 }, { "action_loss": 0.10330591350793839, "epoch": 0.38669064748201437, "step": 430 }, { "epoch": 0.38669064748201437, "step": 430, "torque_loss": 0.28606978058815 }, { "epoch": 0.39568345323741005, "grad_norm": 0.4497431516647339, "learning_rate": 1.756e-05, "loss": 0.0792, "step": 440 }, { "action_loss": 0.08978798240423203, "epoch": 0.39568345323741005, "step": 440 }, { "epoch": 0.39568345323741005, "step": 440, "torque_loss": 0.2021617889404297 }, { "epoch": 0.40467625899280574, "grad_norm": 0.7599253058433533, "learning_rate": 1.796e-05, "loss": 0.0737, "step": 450 }, { "action_loss": 0.04219686985015869, "epoch": 0.40467625899280574, "step": 450 }, { "epoch": 0.40467625899280574, "step": 450, "torque_loss": 0.15015043318271637 }, { "epoch": 0.4136690647482014, "grad_norm": 0.5245656371116638, "learning_rate": 1.8360000000000004e-05, "loss": 0.067, "step": 460 }, { "action_loss": 0.045781876891851425, "epoch": 0.4136690647482014, "step": 460 }, { "epoch": 0.4136690647482014, "step": 460, "torque_loss": 0.19939370453357697 }, { "epoch": 0.4226618705035971, "grad_norm": 0.6752534508705139, "learning_rate": 1.876e-05, "loss": 0.0707, "step": 470 }, { "action_loss": 0.026874354109168053, "epoch": 0.4226618705035971, "step": 470 }, { "epoch": 0.4226618705035971, "step": 470, "torque_loss": 0.1312359720468521 }, { "epoch": 0.4316546762589928, "grad_norm": 0.4566693603992462, "learning_rate": 1.916e-05, "loss": 0.0573, "step": 480 }, { "action_loss": 0.044159263372421265, "epoch": 0.4316546762589928, "step": 480 }, { "epoch": 0.4316546762589928, "step": 480, "torque_loss": 0.19262342154979706 }, { "epoch": 0.44064748201438847, "grad_norm": 0.40263423323631287, "learning_rate": 1.956e-05, "loss": 0.0691, "step": 490 }, { "action_loss": 0.06504664570093155, "epoch": 0.44064748201438847, "step": 490 }, { "epoch": 0.44064748201438847, "step": 490, "torque_loss": 0.17867164313793182 }, { "epoch": 0.44964028776978415, "grad_norm": 0.46872779726982117, "learning_rate": 1.9960000000000002e-05, "loss": 0.0549, "step": 500 }, { "action_loss": 0.05051891505718231, "epoch": 0.44964028776978415, "step": 500 }, { "epoch": 0.44964028776978415, "step": 500, "torque_loss": 0.16960956156253815 }, { "epoch": 0.45863309352517984, "grad_norm": 0.718011200428009, "learning_rate": 2.036e-05, "loss": 0.0652, "step": 510 }, { "action_loss": 0.029472684487700462, "epoch": 0.45863309352517984, "step": 510 }, { "epoch": 0.45863309352517984, "step": 510, "torque_loss": 0.15747827291488647 }, { "epoch": 0.4676258992805755, "grad_norm": 0.5446356534957886, "learning_rate": 2.076e-05, "loss": 0.0533, "step": 520 }, { "action_loss": 0.05690797045826912, "epoch": 0.4676258992805755, "step": 520 }, { "epoch": 0.4676258992805755, "step": 520, "torque_loss": 0.20095066726207733 }, { "epoch": 0.4766187050359712, "grad_norm": 0.6428467631340027, "learning_rate": 2.116e-05, "loss": 0.0699, "step": 530 }, { "action_loss": 0.08365841954946518, "epoch": 0.4766187050359712, "step": 530 }, { "epoch": 0.4766187050359712, "step": 530, "torque_loss": 0.2774507403373718 }, { "epoch": 0.4856115107913669, "grad_norm": 0.4426825940608978, "learning_rate": 2.1560000000000004e-05, "loss": 0.0788, "step": 540 }, { "action_loss": 0.04415509104728699, "epoch": 0.4856115107913669, "step": 540 }, { "epoch": 0.4856115107913669, "step": 540, "torque_loss": 0.18824762105941772 }, { "epoch": 0.49460431654676257, "grad_norm": 0.7582957744598389, "learning_rate": 2.196e-05, "loss": 0.0608, "step": 550 }, { "action_loss": 0.05368471145629883, "epoch": 0.49460431654676257, "step": 550 }, { "epoch": 0.49460431654676257, "step": 550, "torque_loss": 0.24977219104766846 }, { "epoch": 0.5035971223021583, "grad_norm": 0.38401511311531067, "learning_rate": 2.236e-05, "loss": 0.0627, "step": 560 }, { "action_loss": 0.027193143963813782, "epoch": 0.5035971223021583, "step": 560 }, { "epoch": 0.5035971223021583, "step": 560, "torque_loss": 0.2041977196931839 }, { "epoch": 0.512589928057554, "grad_norm": 0.7763616442680359, "learning_rate": 2.2760000000000002e-05, "loss": 0.0822, "step": 570 }, { "action_loss": 0.04098575562238693, "epoch": 0.512589928057554, "step": 570 }, { "epoch": 0.512589928057554, "step": 570, "torque_loss": 0.22381407022476196 }, { "epoch": 0.5215827338129496, "grad_norm": 1.000070571899414, "learning_rate": 2.3160000000000002e-05, "loss": 0.0694, "step": 580 }, { "action_loss": 0.044417958706617355, "epoch": 0.5215827338129496, "step": 580 }, { "epoch": 0.5215827338129496, "step": 580, "torque_loss": 0.17623887956142426 }, { "epoch": 0.5305755395683454, "grad_norm": 0.5415641665458679, "learning_rate": 2.356e-05, "loss": 0.0608, "step": 590 }, { "action_loss": 0.05746312066912651, "epoch": 0.5305755395683454, "step": 590 }, { "epoch": 0.5305755395683454, "step": 590, "torque_loss": 0.23695926368236542 }, { "epoch": 0.539568345323741, "grad_norm": 0.6153416037559509, "learning_rate": 2.396e-05, "loss": 0.0691, "step": 600 }, { "action_loss": 0.10369882732629776, "epoch": 0.539568345323741, "step": 600 }, { "epoch": 0.539568345323741, "step": 600, "torque_loss": 0.24763722717761993 }, { "epoch": 0.5485611510791367, "grad_norm": 0.5401313900947571, "learning_rate": 2.4360000000000004e-05, "loss": 0.0722, "step": 610 }, { "action_loss": 0.06181579455733299, "epoch": 0.5485611510791367, "step": 610 }, { "epoch": 0.5485611510791367, "step": 610, "torque_loss": 0.2548808157444 }, { "epoch": 0.5575539568345323, "grad_norm": 0.3690989315509796, "learning_rate": 2.476e-05, "loss": 0.0668, "step": 620 }, { "action_loss": 0.03288828209042549, "epoch": 0.5575539568345323, "step": 620 }, { "epoch": 0.5575539568345323, "step": 620, "torque_loss": 0.2667144536972046 }, { "epoch": 0.5665467625899281, "grad_norm": 0.5731062293052673, "learning_rate": 2.516e-05, "loss": 0.0619, "step": 630 }, { "action_loss": 0.03954185172915459, "epoch": 0.5665467625899281, "step": 630 }, { "epoch": 0.5665467625899281, "step": 630, "torque_loss": 0.2396117001771927 }, { "epoch": 0.5755395683453237, "grad_norm": 0.43299898505210876, "learning_rate": 2.556e-05, "loss": 0.0586, "step": 640 }, { "action_loss": 0.06193164363503456, "epoch": 0.5755395683453237, "step": 640 }, { "epoch": 0.5755395683453237, "step": 640, "torque_loss": 0.24662399291992188 }, { "epoch": 0.5845323741007195, "grad_norm": 0.3966420292854309, "learning_rate": 2.5960000000000002e-05, "loss": 0.0576, "step": 650 }, { "action_loss": 0.016649065539240837, "epoch": 0.5845323741007195, "step": 650 }, { "epoch": 0.5845323741007195, "step": 650, "torque_loss": 0.10568499565124512 }, { "epoch": 0.5935251798561151, "grad_norm": 0.6101328730583191, "learning_rate": 2.6360000000000002e-05, "loss": 0.059, "step": 660 }, { "action_loss": 0.0313473679125309, "epoch": 0.5935251798561151, "step": 660 }, { "epoch": 0.5935251798561151, "step": 660, "torque_loss": 0.16035155951976776 }, { "epoch": 0.6025179856115108, "grad_norm": 0.5945420265197754, "learning_rate": 2.676e-05, "loss": 0.0441, "step": 670 }, { "action_loss": 0.05465022847056389, "epoch": 0.6025179856115108, "step": 670 }, { "epoch": 0.6025179856115108, "step": 670, "torque_loss": 0.26007410883903503 }, { "epoch": 0.6115107913669064, "grad_norm": 0.523535966873169, "learning_rate": 2.716e-05, "loss": 0.0692, "step": 680 }, { "action_loss": 0.05025331303477287, "epoch": 0.6115107913669064, "step": 680 }, { "epoch": 0.6115107913669064, "step": 680, "torque_loss": 0.20151560008525848 }, { "epoch": 0.6205035971223022, "grad_norm": 0.5459935665130615, "learning_rate": 2.7560000000000004e-05, "loss": 0.0522, "step": 690 }, { "action_loss": 0.027160009369254112, "epoch": 0.6205035971223022, "step": 690 }, { "epoch": 0.6205035971223022, "step": 690, "torque_loss": 0.15134750306606293 }, { "epoch": 0.6294964028776978, "grad_norm": 0.42754602432250977, "learning_rate": 2.7960000000000003e-05, "loss": 0.0526, "step": 700 }, { "action_loss": 0.019805898889899254, "epoch": 0.6294964028776978, "step": 700 }, { "epoch": 0.6294964028776978, "step": 700, "torque_loss": 0.10844121128320694 }, { "epoch": 0.6384892086330936, "grad_norm": 0.6223859786987305, "learning_rate": 2.8360000000000003e-05, "loss": 0.0628, "step": 710 }, { "action_loss": 0.01914464868605137, "epoch": 0.6384892086330936, "step": 710 }, { "epoch": 0.6384892086330936, "step": 710, "torque_loss": 0.16321997344493866 }, { "epoch": 0.6474820143884892, "grad_norm": 0.533725380897522, "learning_rate": 2.8760000000000002e-05, "loss": 0.0522, "step": 720 }, { "action_loss": 0.05486733093857765, "epoch": 0.6474820143884892, "step": 720 }, { "epoch": 0.6474820143884892, "step": 720, "torque_loss": 0.2750431299209595 }, { "epoch": 0.6564748201438849, "grad_norm": 0.5009811520576477, "learning_rate": 2.9160000000000005e-05, "loss": 0.0653, "step": 730 }, { "action_loss": 0.06562066823244095, "epoch": 0.6564748201438849, "step": 730 }, { "epoch": 0.6564748201438849, "step": 730, "torque_loss": 0.17505301535129547 }, { "epoch": 0.6654676258992805, "grad_norm": 0.6075042486190796, "learning_rate": 2.9559999999999998e-05, "loss": 0.0562, "step": 740 }, { "action_loss": 0.02212030254304409, "epoch": 0.6654676258992805, "step": 740 }, { "epoch": 0.6654676258992805, "step": 740, "torque_loss": 0.17404614388942719 }, { "epoch": 0.6744604316546763, "grad_norm": 0.25729435682296753, "learning_rate": 2.9959999999999998e-05, "loss": 0.0534, "step": 750 }, { "action_loss": 0.0175956878811121, "epoch": 0.6744604316546763, "step": 750 }, { "epoch": 0.6744604316546763, "step": 750, "torque_loss": 0.18320195376873016 }, { "epoch": 0.6834532374100719, "grad_norm": 0.4658522307872772, "learning_rate": 3.036e-05, "loss": 0.0521, "step": 760 }, { "action_loss": 0.03602911904454231, "epoch": 0.6834532374100719, "step": 760 }, { "epoch": 0.6834532374100719, "step": 760, "torque_loss": 0.15138091146945953 }, { "epoch": 0.6924460431654677, "grad_norm": 0.5302784442901611, "learning_rate": 3.076e-05, "loss": 0.0521, "step": 770 }, { "action_loss": 0.0202617384493351, "epoch": 0.6924460431654677, "step": 770 }, { "epoch": 0.6924460431654677, "step": 770, "torque_loss": 0.17004497349262238 }, { "epoch": 0.7014388489208633, "grad_norm": 0.3708288073539734, "learning_rate": 3.116e-05, "loss": 0.0576, "step": 780 }, { "action_loss": 0.017213476821780205, "epoch": 0.7014388489208633, "step": 780 }, { "epoch": 0.7014388489208633, "step": 780, "torque_loss": 0.1164519265294075 }, { "epoch": 0.710431654676259, "grad_norm": 0.4659675061702728, "learning_rate": 3.156e-05, "loss": 0.0458, "step": 790 }, { "action_loss": 0.03970985859632492, "epoch": 0.710431654676259, "step": 790 }, { "epoch": 0.710431654676259, "step": 790, "torque_loss": 0.20291666686534882 }, { "epoch": 0.7194244604316546, "grad_norm": 0.3958164155483246, "learning_rate": 3.196e-05, "loss": 0.0552, "step": 800 }, { "action_loss": 0.024934371933341026, "epoch": 0.7194244604316546, "step": 800 }, { "epoch": 0.7194244604316546, "step": 800, "torque_loss": 0.12873078882694244 }, { "epoch": 0.7284172661870504, "grad_norm": 0.4233374297618866, "learning_rate": 3.236e-05, "loss": 0.0501, "step": 810 }, { "action_loss": 0.058651864528656006, "epoch": 0.7284172661870504, "step": 810 }, { "epoch": 0.7284172661870504, "step": 810, "torque_loss": 0.2202492207288742 }, { "epoch": 0.737410071942446, "grad_norm": 0.4540588855743408, "learning_rate": 3.2760000000000005e-05, "loss": 0.0711, "step": 820 }, { "action_loss": 0.012392948381602764, "epoch": 0.737410071942446, "step": 820 }, { "epoch": 0.737410071942446, "step": 820, "torque_loss": 0.10931023210287094 }, { "epoch": 0.7464028776978417, "grad_norm": 0.39308398962020874, "learning_rate": 3.316e-05, "loss": 0.0549, "step": 830 }, { "action_loss": 0.020457815378904343, "epoch": 0.7464028776978417, "step": 830 }, { "epoch": 0.7464028776978417, "step": 830, "torque_loss": 0.1123281717300415 }, { "epoch": 0.7553956834532374, "grad_norm": 0.5928940773010254, "learning_rate": 3.3560000000000004e-05, "loss": 0.048, "step": 840 }, { "action_loss": 0.03166000917553902, "epoch": 0.7553956834532374, "step": 840 }, { "epoch": 0.7553956834532374, "step": 840, "torque_loss": 0.16852788627147675 }, { "epoch": 0.7643884892086331, "grad_norm": 0.4655683934688568, "learning_rate": 3.396e-05, "loss": 0.05, "step": 850 }, { "action_loss": 0.028250427916646004, "epoch": 0.7643884892086331, "step": 850 }, { "epoch": 0.7643884892086331, "step": 850, "torque_loss": 0.15471215546131134 }, { "epoch": 0.7733812949640287, "grad_norm": 0.7144094705581665, "learning_rate": 3.436e-05, "loss": 0.0604, "step": 860 }, { "action_loss": 0.03964892402291298, "epoch": 0.7733812949640287, "step": 860 }, { "epoch": 0.7733812949640287, "step": 860, "torque_loss": 0.30103400349617004 }, { "epoch": 0.7823741007194245, "grad_norm": 0.6225090026855469, "learning_rate": 3.4760000000000006e-05, "loss": 0.0508, "step": 870 }, { "action_loss": 0.03944830968976021, "epoch": 0.7823741007194245, "step": 870 }, { "epoch": 0.7823741007194245, "step": 870, "torque_loss": 0.14263294637203217 }, { "epoch": 0.7913669064748201, "grad_norm": 0.5022530555725098, "learning_rate": 3.516e-05, "loss": 0.0549, "step": 880 }, { "action_loss": 0.03458128497004509, "epoch": 0.7913669064748201, "step": 880 }, { "epoch": 0.7913669064748201, "step": 880, "torque_loss": 0.19621337950229645 }, { "epoch": 0.8003597122302158, "grad_norm": 0.33522191643714905, "learning_rate": 3.5560000000000005e-05, "loss": 0.0515, "step": 890 }, { "action_loss": 0.020155631005764008, "epoch": 0.8003597122302158, "step": 890 }, { "epoch": 0.8003597122302158, "step": 890, "torque_loss": 0.09344450384378433 }, { "epoch": 0.8093525179856115, "grad_norm": 0.38680991530418396, "learning_rate": 3.596e-05, "loss": 0.0509, "step": 900 }, { "action_loss": 0.05852341651916504, "epoch": 0.8093525179856115, "step": 900 }, { "epoch": 0.8093525179856115, "step": 900, "torque_loss": 0.27722081542015076 }, { "epoch": 0.8183453237410072, "grad_norm": 0.4631944000720978, "learning_rate": 3.636e-05, "loss": 0.0572, "step": 910 }, { "action_loss": 0.031825583428144455, "epoch": 0.8183453237410072, "step": 910 }, { "epoch": 0.8183453237410072, "step": 910, "torque_loss": 0.15583252906799316 }, { "epoch": 0.8273381294964028, "grad_norm": 0.4406748116016388, "learning_rate": 3.676e-05, "loss": 0.0588, "step": 920 }, { "action_loss": 0.028036067262291908, "epoch": 0.8273381294964028, "step": 920 }, { "epoch": 0.8273381294964028, "step": 920, "torque_loss": 0.1766703873872757 }, { "epoch": 0.8363309352517986, "grad_norm": 0.5505356192588806, "learning_rate": 3.716e-05, "loss": 0.0718, "step": 930 }, { "action_loss": 0.02862476371228695, "epoch": 0.8363309352517986, "step": 930 }, { "epoch": 0.8363309352517986, "step": 930, "torque_loss": 0.14178769290447235 }, { "epoch": 0.8453237410071942, "grad_norm": 0.47320014238357544, "learning_rate": 3.756e-05, "loss": 0.045, "step": 940 }, { "action_loss": 0.05340605974197388, "epoch": 0.8453237410071942, "step": 940 }, { "epoch": 0.8453237410071942, "step": 940, "torque_loss": 0.19522766768932343 }, { "epoch": 0.85431654676259, "grad_norm": 0.4562070667743683, "learning_rate": 3.796e-05, "loss": 0.0508, "step": 950 }, { "action_loss": 0.04853568598628044, "epoch": 0.85431654676259, "step": 950 }, { "epoch": 0.85431654676259, "step": 950, "torque_loss": 0.2683302164077759 }, { "epoch": 0.8633093525179856, "grad_norm": 0.49535953998565674, "learning_rate": 3.836e-05, "loss": 0.0586, "step": 960 }, { "action_loss": 0.02389262616634369, "epoch": 0.8633093525179856, "step": 960 }, { "epoch": 0.8633093525179856, "step": 960, "torque_loss": 0.22259783744812012 }, { "epoch": 0.8723021582733813, "grad_norm": 0.532315194606781, "learning_rate": 3.876e-05, "loss": 0.05, "step": 970 }, { "action_loss": 0.020066607743501663, "epoch": 0.8723021582733813, "step": 970 }, { "epoch": 0.8723021582733813, "step": 970, "torque_loss": 0.15999776124954224 }, { "epoch": 0.8812949640287769, "grad_norm": 0.41063129901885986, "learning_rate": 3.9160000000000005e-05, "loss": 0.0501, "step": 980 }, { "action_loss": 0.017572185024619102, "epoch": 0.8812949640287769, "step": 980 }, { "epoch": 0.8812949640287769, "step": 980, "torque_loss": 0.13796262443065643 }, { "epoch": 0.8902877697841727, "grad_norm": 0.35861673951148987, "learning_rate": 3.956e-05, "loss": 0.048, "step": 990 }, { "action_loss": 0.017416732385754585, "epoch": 0.8902877697841727, "step": 990 }, { "epoch": 0.8902877697841727, "step": 990, "torque_loss": 0.1409957855939865 }, { "epoch": 0.8992805755395683, "grad_norm": 0.4475710690021515, "learning_rate": 3.9960000000000004e-05, "loss": 0.0498, "step": 1000 }, { "action_loss": 0.0534302294254303, "epoch": 0.8992805755395683, "step": 1000 }, { "epoch": 0.8992805755395683, "step": 1000, "torque_loss": 0.22861206531524658 }, { "epoch": 0.908273381294964, "grad_norm": 0.3044046461582184, "learning_rate": 4.0360000000000007e-05, "loss": 0.0589, "step": 1010 }, { "action_loss": 0.01553951296955347, "epoch": 0.908273381294964, "step": 1010 }, { "epoch": 0.908273381294964, "step": 1010, "torque_loss": 0.1360899955034256 }, { "epoch": 0.9172661870503597, "grad_norm": 0.32783564925193787, "learning_rate": 4.076e-05, "loss": 0.0448, "step": 1020 }, { "action_loss": 0.03921547159552574, "epoch": 0.9172661870503597, "step": 1020 }, { "epoch": 0.9172661870503597, "step": 1020, "torque_loss": 0.1698317527770996 }, { "epoch": 0.9262589928057554, "grad_norm": 0.4188622236251831, "learning_rate": 4.1160000000000006e-05, "loss": 0.0479, "step": 1030 }, { "action_loss": 0.023148775100708008, "epoch": 0.9262589928057554, "step": 1030 }, { "epoch": 0.9262589928057554, "step": 1030, "torque_loss": 0.14986000955104828 }, { "epoch": 0.935251798561151, "grad_norm": 0.6149533987045288, "learning_rate": 4.156e-05, "loss": 0.0461, "step": 1040 }, { "action_loss": 0.024126717820763588, "epoch": 0.935251798561151, "step": 1040 }, { "epoch": 0.935251798561151, "step": 1040, "torque_loss": 0.1895468682050705 }, { "epoch": 0.9442446043165468, "grad_norm": 0.381228506565094, "learning_rate": 4.196e-05, "loss": 0.0498, "step": 1050 }, { "action_loss": 0.035189930349588394, "epoch": 0.9442446043165468, "step": 1050 }, { "epoch": 0.9442446043165468, "step": 1050, "torque_loss": 0.20947761833667755 }, { "epoch": 0.9532374100719424, "grad_norm": 0.37987107038497925, "learning_rate": 4.236e-05, "loss": 0.0588, "step": 1060 }, { "action_loss": 0.02284870482981205, "epoch": 0.9532374100719424, "step": 1060 }, { "epoch": 0.9532374100719424, "step": 1060, "torque_loss": 0.11762746423482895 }, { "epoch": 0.9622302158273381, "grad_norm": 0.2299790382385254, "learning_rate": 4.276e-05, "loss": 0.041, "step": 1070 }, { "action_loss": 0.05730075761675835, "epoch": 0.9622302158273381, "step": 1070 }, { "epoch": 0.9622302158273381, "step": 1070, "torque_loss": 0.27405789494514465 }, { "epoch": 0.9712230215827338, "grad_norm": 0.48734015226364136, "learning_rate": 4.316e-05, "loss": 0.0544, "step": 1080 }, { "action_loss": 0.01564769446849823, "epoch": 0.9712230215827338, "step": 1080 }, { "epoch": 0.9712230215827338, "step": 1080, "torque_loss": 0.1693197339773178 }, { "epoch": 0.9802158273381295, "grad_norm": 0.4651220440864563, "learning_rate": 4.356e-05, "loss": 0.0507, "step": 1090 }, { "action_loss": 0.04978068545460701, "epoch": 0.9802158273381295, "step": 1090 }, { "epoch": 0.9802158273381295, "step": 1090, "torque_loss": 0.22813434898853302 }, { "epoch": 0.9892086330935251, "grad_norm": 0.4476982057094574, "learning_rate": 4.396e-05, "loss": 0.0456, "step": 1100 }, { "action_loss": 0.049542635679244995, "epoch": 0.9892086330935251, "step": 1100 }, { "epoch": 0.9892086330935251, "step": 1100, "torque_loss": 0.2431374043226242 }, { "epoch": 0.9982014388489209, "grad_norm": 0.39087188243865967, "learning_rate": 4.436e-05, "loss": 0.0455, "step": 1110 }, { "action_loss": 0.021530643105506897, "epoch": 0.9982014388489209, "step": 1110 }, { "epoch": 0.9982014388489209, "step": 1110, "torque_loss": 0.1980867236852646 }, { "epoch": 1.0071942446043165, "grad_norm": 0.3524869978427887, "learning_rate": 4.4760000000000005e-05, "loss": 0.044, "step": 1120 }, { "action_loss": 0.01638363115489483, "epoch": 1.0071942446043165, "step": 1120 }, { "epoch": 1.0071942446043165, "step": 1120, "torque_loss": 0.14354729652404785 }, { "epoch": 1.0161870503597121, "grad_norm": 0.38940292596817017, "learning_rate": 4.516e-05, "loss": 0.046, "step": 1130 }, { "action_loss": 0.016111167147755623, "epoch": 1.0161870503597121, "step": 1130 }, { "epoch": 1.0161870503597121, "step": 1130, "torque_loss": 0.13946735858917236 }, { "epoch": 1.025179856115108, "grad_norm": 0.29563599824905396, "learning_rate": 4.5560000000000004e-05, "loss": 0.0429, "step": 1140 }, { "action_loss": 0.04645264148712158, "epoch": 1.025179856115108, "step": 1140 }, { "epoch": 1.025179856115108, "step": 1140, "torque_loss": 0.29726502299308777 }, { "epoch": 1.0341726618705036, "grad_norm": 0.3685888648033142, "learning_rate": 4.596e-05, "loss": 0.0502, "step": 1150 }, { "action_loss": 0.022467130795121193, "epoch": 1.0341726618705036, "step": 1150 }, { "epoch": 1.0341726618705036, "step": 1150, "torque_loss": 0.2026911824941635 }, { "epoch": 1.0431654676258992, "grad_norm": 0.2810155749320984, "learning_rate": 4.636e-05, "loss": 0.0445, "step": 1160 }, { "action_loss": 0.023996705189347267, "epoch": 1.0431654676258992, "step": 1160 }, { "epoch": 1.0431654676258992, "step": 1160, "torque_loss": 0.16748501360416412 }, { "epoch": 1.0521582733812949, "grad_norm": 0.4202764630317688, "learning_rate": 4.6760000000000006e-05, "loss": 0.0558, "step": 1170 }, { "action_loss": 0.04188133403658867, "epoch": 1.0521582733812949, "step": 1170 }, { "epoch": 1.0521582733812949, "step": 1170, "torque_loss": 0.24878065288066864 }, { "epoch": 1.0611510791366907, "grad_norm": 0.38165104389190674, "learning_rate": 4.716e-05, "loss": 0.0561, "step": 1180 }, { "action_loss": 0.028449922800064087, "epoch": 1.0611510791366907, "step": 1180 }, { "epoch": 1.0611510791366907, "step": 1180, "torque_loss": 0.2239365130662918 }, { "epoch": 1.0701438848920863, "grad_norm": 0.3462095260620117, "learning_rate": 4.7560000000000005e-05, "loss": 0.0514, "step": 1190 }, { "action_loss": 0.04515869915485382, "epoch": 1.0701438848920863, "step": 1190 }, { "epoch": 1.0701438848920863, "step": 1190, "torque_loss": 0.23675818741321564 }, { "epoch": 1.079136690647482, "grad_norm": 0.5111167430877686, "learning_rate": 4.796e-05, "loss": 0.0525, "step": 1200 }, { "action_loss": 0.017050890251994133, "epoch": 1.079136690647482, "step": 1200 }, { "epoch": 1.079136690647482, "step": 1200, "torque_loss": 0.1295180767774582 }, { "epoch": 1.0881294964028776, "grad_norm": 0.3835822343826294, "learning_rate": 4.836e-05, "loss": 0.0581, "step": 1210 }, { "action_loss": 0.03141402080655098, "epoch": 1.0881294964028776, "step": 1210 }, { "epoch": 1.0881294964028776, "step": 1210, "torque_loss": 0.19610624015331268 }, { "epoch": 1.0971223021582734, "grad_norm": 0.3829500377178192, "learning_rate": 4.876e-05, "loss": 0.0551, "step": 1220 }, { "action_loss": 0.03373543545603752, "epoch": 1.0971223021582734, "step": 1220 }, { "epoch": 1.0971223021582734, "step": 1220, "torque_loss": 0.20370018482208252 }, { "epoch": 1.106115107913669, "grad_norm": 0.391216903924942, "learning_rate": 4.9160000000000004e-05, "loss": 0.0462, "step": 1230 }, { "action_loss": 0.020509155467152596, "epoch": 1.106115107913669, "step": 1230 }, { "epoch": 1.106115107913669, "step": 1230, "torque_loss": 0.18182866275310516 }, { "epoch": 1.1151079136690647, "grad_norm": 0.5667259693145752, "learning_rate": 4.956e-05, "loss": 0.0475, "step": 1240 }, { "action_loss": 0.04859720543026924, "epoch": 1.1151079136690647, "step": 1240 }, { "epoch": 1.1151079136690647, "step": 1240, "torque_loss": 0.2351435422897339 }, { "epoch": 1.1241007194244603, "grad_norm": 0.4090467393398285, "learning_rate": 4.996e-05, "loss": 0.0446, "step": 1250 }, { "action_loss": 0.01935848779976368, "epoch": 1.1241007194244603, "step": 1250 }, { "epoch": 1.1241007194244603, "step": 1250, "torque_loss": 0.14848721027374268 }, { "epoch": 1.1330935251798562, "grad_norm": 0.344253271818161, "learning_rate": 5.0360000000000006e-05, "loss": 0.0413, "step": 1260 }, { "action_loss": 0.02568025141954422, "epoch": 1.1330935251798562, "step": 1260 }, { "epoch": 1.1330935251798562, "step": 1260, "torque_loss": 0.1360962837934494 }, { "epoch": 1.1420863309352518, "grad_norm": 0.3656346797943115, "learning_rate": 5.076000000000001e-05, "loss": 0.0388, "step": 1270 }, { "action_loss": 0.039098434150218964, "epoch": 1.1420863309352518, "step": 1270 }, { "epoch": 1.1420863309352518, "step": 1270, "torque_loss": 0.2793858051300049 }, { "epoch": 1.1510791366906474, "grad_norm": 0.670525848865509, "learning_rate": 5.1160000000000005e-05, "loss": 0.0497, "step": 1280 }, { "action_loss": 0.017696654424071312, "epoch": 1.1510791366906474, "step": 1280 }, { "epoch": 1.1510791366906474, "step": 1280, "torque_loss": 0.18088002502918243 }, { "epoch": 1.1600719424460433, "grad_norm": 0.4277499318122864, "learning_rate": 5.1559999999999994e-05, "loss": 0.0508, "step": 1290 }, { "action_loss": 0.02923022024333477, "epoch": 1.1600719424460433, "step": 1290 }, { "epoch": 1.1600719424460433, "step": 1290, "torque_loss": 0.18822503089904785 }, { "epoch": 1.169064748201439, "grad_norm": 0.5093660950660706, "learning_rate": 5.196e-05, "loss": 0.0489, "step": 1300 }, { "action_loss": 0.018707318231463432, "epoch": 1.169064748201439, "step": 1300 }, { "epoch": 1.169064748201439, "step": 1300, "torque_loss": 0.15995459258556366 }, { "epoch": 1.1780575539568345, "grad_norm": 0.4274079501628876, "learning_rate": 5.236e-05, "loss": 0.0454, "step": 1310 }, { "action_loss": 0.038242701441049576, "epoch": 1.1780575539568345, "step": 1310 }, { "epoch": 1.1780575539568345, "step": 1310, "torque_loss": 0.20593225955963135 }, { "epoch": 1.1870503597122302, "grad_norm": 0.38017386198043823, "learning_rate": 5.2759999999999996e-05, "loss": 0.0519, "step": 1320 }, { "action_loss": 0.03470459207892418, "epoch": 1.1870503597122302, "step": 1320 }, { "epoch": 1.1870503597122302, "step": 1320, "torque_loss": 0.2100237011909485 }, { "epoch": 1.1960431654676258, "grad_norm": 0.42054614424705505, "learning_rate": 5.316e-05, "loss": 0.0524, "step": 1330 }, { "action_loss": 0.017400534823536873, "epoch": 1.1960431654676258, "step": 1330 }, { "epoch": 1.1960431654676258, "step": 1330, "torque_loss": 0.15161898732185364 }, { "epoch": 1.2050359712230216, "grad_norm": 0.4100704789161682, "learning_rate": 5.356e-05, "loss": 0.0616, "step": 1340 }, { "action_loss": 0.024229541420936584, "epoch": 1.2050359712230216, "step": 1340 }, { "epoch": 1.2050359712230216, "step": 1340, "torque_loss": 0.19844985008239746 }, { "epoch": 1.2140287769784173, "grad_norm": 0.4476875364780426, "learning_rate": 5.396e-05, "loss": 0.043, "step": 1350 }, { "action_loss": 0.02864181064069271, "epoch": 1.2140287769784173, "step": 1350 }, { "epoch": 1.2140287769784173, "step": 1350, "torque_loss": 0.1767013818025589 }, { "epoch": 1.223021582733813, "grad_norm": 0.6143948435783386, "learning_rate": 5.436e-05, "loss": 0.0602, "step": 1360 }, { "action_loss": 0.022161653265357018, "epoch": 1.223021582733813, "step": 1360 }, { "epoch": 1.223021582733813, "step": 1360, "torque_loss": 0.16223780810832977 }, { "epoch": 1.2320143884892087, "grad_norm": 0.612689733505249, "learning_rate": 5.476e-05, "loss": 0.0513, "step": 1370 }, { "action_loss": 0.016723446547985077, "epoch": 1.2320143884892087, "step": 1370 }, { "epoch": 1.2320143884892087, "step": 1370, "torque_loss": 0.12243124097585678 }, { "epoch": 1.2410071942446044, "grad_norm": 0.4922971725463867, "learning_rate": 5.516e-05, "loss": 0.0507, "step": 1380 }, { "action_loss": 0.023983707651495934, "epoch": 1.2410071942446044, "step": 1380 }, { "epoch": 1.2410071942446044, "step": 1380, "torque_loss": 0.14170648157596588 }, { "epoch": 1.25, "grad_norm": 0.382380872964859, "learning_rate": 5.556e-05, "loss": 0.0484, "step": 1390 }, { "action_loss": 0.022591570392251015, "epoch": 1.25, "step": 1390 }, { "epoch": 1.25, "step": 1390, "torque_loss": 0.13047659397125244 }, { "epoch": 1.2589928057553956, "grad_norm": 0.3310081362724304, "learning_rate": 5.596e-05, "loss": 0.0459, "step": 1400 }, { "action_loss": 0.022876689210534096, "epoch": 1.2589928057553956, "step": 1400 }, { "epoch": 1.2589928057553956, "step": 1400, "torque_loss": 0.1372627168893814 }, { "epoch": 1.2679856115107913, "grad_norm": 0.36787641048431396, "learning_rate": 5.636e-05, "loss": 0.0627, "step": 1410 }, { "action_loss": 0.020593823865056038, "epoch": 1.2679856115107913, "step": 1410 }, { "epoch": 1.2679856115107913, "step": 1410, "torque_loss": 0.21719157695770264 }, { "epoch": 1.276978417266187, "grad_norm": 0.4183320105075836, "learning_rate": 5.6760000000000005e-05, "loss": 0.0515, "step": 1420 }, { "action_loss": 0.044062256813049316, "epoch": 1.276978417266187, "step": 1420 }, { "epoch": 1.276978417266187, "step": 1420, "torque_loss": 0.33146294951438904 }, { "epoch": 1.2859712230215827, "grad_norm": 0.509203314781189, "learning_rate": 5.716e-05, "loss": 0.0522, "step": 1430 }, { "action_loss": 0.04274165630340576, "epoch": 1.2859712230215827, "step": 1430 }, { "epoch": 1.2859712230215827, "step": 1430, "torque_loss": 0.23784594237804413 }, { "epoch": 1.2949640287769784, "grad_norm": 0.4527966380119324, "learning_rate": 5.7560000000000005e-05, "loss": 0.0434, "step": 1440 }, { "action_loss": 0.03770029917359352, "epoch": 1.2949640287769784, "step": 1440 }, { "epoch": 1.2949640287769784, "step": 1440, "torque_loss": 0.2693568170070648 }, { "epoch": 1.3039568345323742, "grad_norm": 0.36911994218826294, "learning_rate": 5.796e-05, "loss": 0.0519, "step": 1450 }, { "action_loss": 0.024354467168450356, "epoch": 1.3039568345323742, "step": 1450 }, { "epoch": 1.3039568345323742, "step": 1450, "torque_loss": 0.20227932929992676 }, { "epoch": 1.3129496402877698, "grad_norm": 0.4817804992198944, "learning_rate": 5.8360000000000004e-05, "loss": 0.0455, "step": 1460 }, { "action_loss": 0.015746375545859337, "epoch": 1.3129496402877698, "step": 1460 }, { "epoch": 1.3129496402877698, "step": 1460, "torque_loss": 0.1370859593153 }, { "epoch": 1.3219424460431655, "grad_norm": 0.4965204894542694, "learning_rate": 5.876000000000001e-05, "loss": 0.0388, "step": 1470 }, { "action_loss": 0.02545027993619442, "epoch": 1.3219424460431655, "step": 1470 }, { "epoch": 1.3219424460431655, "step": 1470, "torque_loss": 0.166796013712883 }, { "epoch": 1.330935251798561, "grad_norm": 0.33666500449180603, "learning_rate": 5.916e-05, "loss": 0.0453, "step": 1480 }, { "action_loss": 0.01567959599196911, "epoch": 1.330935251798561, "step": 1480 }, { "epoch": 1.330935251798561, "step": 1480, "torque_loss": 0.15042875707149506 }, { "epoch": 1.3399280575539567, "grad_norm": 0.43540501594543457, "learning_rate": 5.9560000000000006e-05, "loss": 0.0425, "step": 1490 }, { "action_loss": 0.034763142466545105, "epoch": 1.3399280575539567, "step": 1490 }, { "epoch": 1.3399280575539567, "step": 1490, "torque_loss": 0.17238204181194305 }, { "epoch": 1.3489208633093526, "grad_norm": 0.30634042620658875, "learning_rate": 5.996e-05, "loss": 0.052, "step": 1500 }, { "action_loss": 0.020850330591201782, "epoch": 1.3489208633093526, "step": 1500 }, { "epoch": 1.3489208633093526, "step": 1500, "torque_loss": 0.17308409512043 }, { "epoch": 1.3579136690647482, "grad_norm": 0.3596743643283844, "learning_rate": 6.0360000000000005e-05, "loss": 0.0369, "step": 1510 }, { "action_loss": 0.03814614936709404, "epoch": 1.3579136690647482, "step": 1510 }, { "epoch": 1.3579136690647482, "step": 1510, "torque_loss": 0.25480917096138 }, { "epoch": 1.3669064748201438, "grad_norm": 0.3382503092288971, "learning_rate": 6.076000000000001e-05, "loss": 0.0447, "step": 1520 }, { "action_loss": 0.027074245736002922, "epoch": 1.3669064748201438, "step": 1520 }, { "epoch": 1.3669064748201438, "step": 1520, "torque_loss": 0.1614661067724228 }, { "epoch": 1.3758992805755397, "grad_norm": 0.3323040008544922, "learning_rate": 6.116e-05, "loss": 0.05, "step": 1530 }, { "action_loss": 0.023789329454302788, "epoch": 1.3758992805755397, "step": 1530 }, { "epoch": 1.3758992805755397, "step": 1530, "torque_loss": 0.15351609885692596 }, { "epoch": 1.3848920863309353, "grad_norm": 0.376068115234375, "learning_rate": 6.156e-05, "loss": 0.0421, "step": 1540 }, { "action_loss": 0.01909157633781433, "epoch": 1.3848920863309353, "step": 1540 }, { "epoch": 1.3848920863309353, "step": 1540, "torque_loss": 0.14374971389770508 }, { "epoch": 1.393884892086331, "grad_norm": 0.40914031863212585, "learning_rate": 6.196000000000001e-05, "loss": 0.054, "step": 1550 }, { "action_loss": 0.023658379912376404, "epoch": 1.393884892086331, "step": 1550 }, { "epoch": 1.393884892086331, "step": 1550, "torque_loss": 0.18285103142261505 }, { "epoch": 1.4028776978417266, "grad_norm": 0.3349551260471344, "learning_rate": 6.236e-05, "loss": 0.0399, "step": 1560 }, { "action_loss": 0.019965916872024536, "epoch": 1.4028776978417266, "step": 1560 }, { "epoch": 1.4028776978417266, "step": 1560, "torque_loss": 0.14756068587303162 }, { "epoch": 1.4118705035971222, "grad_norm": 0.42672407627105713, "learning_rate": 6.276e-05, "loss": 0.0424, "step": 1570 }, { "action_loss": 0.0176348015666008, "epoch": 1.4118705035971222, "step": 1570 }, { "epoch": 1.4118705035971222, "step": 1570, "torque_loss": 0.14561396837234497 }, { "epoch": 1.420863309352518, "grad_norm": 0.37563541531562805, "learning_rate": 6.316000000000001e-05, "loss": 0.0448, "step": 1580 }, { "action_loss": 0.021297544240951538, "epoch": 1.420863309352518, "step": 1580 }, { "epoch": 1.420863309352518, "step": 1580, "torque_loss": 0.15647457540035248 }, { "epoch": 1.4298561151079137, "grad_norm": 0.424061119556427, "learning_rate": 6.356000000000001e-05, "loss": 0.0446, "step": 1590 }, { "action_loss": 0.02510392665863037, "epoch": 1.4298561151079137, "step": 1590 }, { "epoch": 1.4298561151079137, "step": 1590, "torque_loss": 0.19316184520721436 }, { "epoch": 1.4388489208633093, "grad_norm": 0.44568273425102234, "learning_rate": 6.396e-05, "loss": 0.0441, "step": 1600 }, { "action_loss": 0.0315643809735775, "epoch": 1.4388489208633093, "step": 1600 }, { "epoch": 1.4388489208633093, "step": 1600, "torque_loss": 0.24600310623645782 }, { "epoch": 1.4478417266187051, "grad_norm": 0.370071679353714, "learning_rate": 6.436e-05, "loss": 0.0529, "step": 1610 }, { "action_loss": 0.050967127084732056, "epoch": 1.4478417266187051, "step": 1610 }, { "epoch": 1.4478417266187051, "step": 1610, "torque_loss": 0.18963158130645752 }, { "epoch": 1.4568345323741008, "grad_norm": 0.28592362999916077, "learning_rate": 6.476e-05, "loss": 0.0537, "step": 1620 }, { "action_loss": 0.02606581151485443, "epoch": 1.4568345323741008, "step": 1620 }, { "epoch": 1.4568345323741008, "step": 1620, "torque_loss": 0.15517330169677734 }, { "epoch": 1.4658273381294964, "grad_norm": 0.5743815898895264, "learning_rate": 6.515999999999999e-05, "loss": 0.0551, "step": 1630 }, { "action_loss": 0.02338586002588272, "epoch": 1.4658273381294964, "step": 1630 }, { "epoch": 1.4658273381294964, "step": 1630, "torque_loss": 0.1256376951932907 }, { "epoch": 1.474820143884892, "grad_norm": 0.46561601758003235, "learning_rate": 6.556e-05, "loss": 0.0488, "step": 1640 }, { "action_loss": 0.03498927131295204, "epoch": 1.474820143884892, "step": 1640 }, { "epoch": 1.474820143884892, "step": 1640, "torque_loss": 0.19598519802093506 }, { "epoch": 1.4838129496402876, "grad_norm": 0.2959613800048828, "learning_rate": 6.596e-05, "loss": 0.0444, "step": 1650 }, { "action_loss": 0.017512520775198936, "epoch": 1.4838129496402876, "step": 1650 }, { "epoch": 1.4838129496402876, "step": 1650, "torque_loss": 0.14181210100650787 }, { "epoch": 1.4928057553956835, "grad_norm": 0.4936921298503876, "learning_rate": 6.636e-05, "loss": 0.0597, "step": 1660 }, { "action_loss": 0.030745545402169228, "epoch": 1.4928057553956835, "step": 1660 }, { "epoch": 1.4928057553956835, "step": 1660, "torque_loss": 0.17175215482711792 }, { "epoch": 1.5017985611510791, "grad_norm": 0.49330464005470276, "learning_rate": 6.676e-05, "loss": 0.0465, "step": 1670 }, { "action_loss": 0.04871199652552605, "epoch": 1.5017985611510791, "step": 1670 }, { "epoch": 1.5017985611510791, "step": 1670, "torque_loss": 0.17801575362682343 }, { "epoch": 1.5107913669064748, "grad_norm": 0.5171966552734375, "learning_rate": 6.716e-05, "loss": 0.0477, "step": 1680 }, { "action_loss": 0.020107023417949677, "epoch": 1.5107913669064748, "step": 1680 }, { "epoch": 1.5107913669064748, "step": 1680, "torque_loss": 0.17510217428207397 }, { "epoch": 1.5197841726618706, "grad_norm": 0.4598844647407532, "learning_rate": 6.756e-05, "loss": 0.0609, "step": 1690 }, { "action_loss": 0.01948166824877262, "epoch": 1.5197841726618706, "step": 1690 }, { "epoch": 1.5197841726618706, "step": 1690, "torque_loss": 0.15195266902446747 }, { "epoch": 1.5287769784172662, "grad_norm": 0.4174179136753082, "learning_rate": 6.796e-05, "loss": 0.0517, "step": 1700 }, { "action_loss": 0.03015150874853134, "epoch": 1.5287769784172662, "step": 1700 }, { "epoch": 1.5287769784172662, "step": 1700, "torque_loss": 0.20010225474834442 }, { "epoch": 1.5377697841726619, "grad_norm": 0.45165321230888367, "learning_rate": 6.836e-05, "loss": 0.0485, "step": 1710 }, { "action_loss": 0.02376771718263626, "epoch": 1.5377697841726619, "step": 1710 }, { "epoch": 1.5377697841726619, "step": 1710, "torque_loss": 0.18445666134357452 }, { "epoch": 1.5467625899280577, "grad_norm": 0.4547675848007202, "learning_rate": 6.876e-05, "loss": 0.0561, "step": 1720 }, { "action_loss": 0.02808285690844059, "epoch": 1.5467625899280577, "step": 1720 }, { "epoch": 1.5467625899280577, "step": 1720, "torque_loss": 0.11835739761590958 }, { "epoch": 1.5557553956834531, "grad_norm": 0.2547130286693573, "learning_rate": 6.916000000000001e-05, "loss": 0.0413, "step": 1730 }, { "action_loss": 0.06035536155104637, "epoch": 1.5557553956834531, "step": 1730 }, { "epoch": 1.5557553956834531, "step": 1730, "torque_loss": 0.38086238503456116 }, { "epoch": 1.564748201438849, "grad_norm": 0.3962048888206482, "learning_rate": 6.956e-05, "loss": 0.0545, "step": 1740 }, { "action_loss": 0.030291631817817688, "epoch": 1.564748201438849, "step": 1740 }, { "epoch": 1.564748201438849, "step": 1740, "torque_loss": 0.1999092549085617 }, { "epoch": 1.5737410071942446, "grad_norm": 0.25326114892959595, "learning_rate": 6.996e-05, "loss": 0.0452, "step": 1750 }, { "action_loss": 0.02310248464345932, "epoch": 1.5737410071942446, "step": 1750 }, { "epoch": 1.5737410071942446, "step": 1750, "torque_loss": 0.18782590329647064 }, { "epoch": 1.5827338129496402, "grad_norm": 0.4165748655796051, "learning_rate": 7.036e-05, "loss": 0.0459, "step": 1760 }, { "action_loss": 0.02599557675421238, "epoch": 1.5827338129496402, "step": 1760 }, { "epoch": 1.5827338129496402, "step": 1760, "torque_loss": 0.20621705055236816 }, { "epoch": 1.591726618705036, "grad_norm": 0.34173622727394104, "learning_rate": 7.076000000000001e-05, "loss": 0.0467, "step": 1770 }, { "action_loss": 0.013974405825138092, "epoch": 1.591726618705036, "step": 1770 }, { "epoch": 1.591726618705036, "step": 1770, "torque_loss": 0.19532422721385956 }, { "epoch": 1.6007194244604317, "grad_norm": 0.4493853747844696, "learning_rate": 7.116e-05, "loss": 0.0375, "step": 1780 }, { "action_loss": 0.019557425752282143, "epoch": 1.6007194244604317, "step": 1780 }, { "epoch": 1.6007194244604317, "step": 1780, "torque_loss": 0.20571713149547577 }, { "epoch": 1.6097122302158273, "grad_norm": 0.5231795907020569, "learning_rate": 7.156e-05, "loss": 0.0539, "step": 1790 }, { "action_loss": 0.018406517803668976, "epoch": 1.6097122302158273, "step": 1790 }, { "epoch": 1.6097122302158273, "step": 1790, "torque_loss": 0.15620791912078857 }, { "epoch": 1.6187050359712232, "grad_norm": 0.37757134437561035, "learning_rate": 7.196000000000001e-05, "loss": 0.0517, "step": 1800 }, { "action_loss": 0.02401515282690525, "epoch": 1.6187050359712232, "step": 1800 }, { "epoch": 1.6187050359712232, "step": 1800, "torque_loss": 0.1862899661064148 }, { "epoch": 1.6276978417266186, "grad_norm": 0.3540925979614258, "learning_rate": 7.236e-05, "loss": 0.044, "step": 1810 }, { "action_loss": 0.02475384809076786, "epoch": 1.6276978417266186, "step": 1810 }, { "epoch": 1.6276978417266186, "step": 1810, "torque_loss": 0.15303362905979156 }, { "epoch": 1.6366906474820144, "grad_norm": 0.40718966722488403, "learning_rate": 7.276e-05, "loss": 0.0396, "step": 1820 }, { "action_loss": 0.01457757968455553, "epoch": 1.6366906474820144, "step": 1820 }, { "epoch": 1.6366906474820144, "step": 1820, "torque_loss": 0.16081200540065765 }, { "epoch": 1.64568345323741, "grad_norm": 0.25310182571411133, "learning_rate": 7.316000000000001e-05, "loss": 0.0451, "step": 1830 }, { "action_loss": 0.02790679782629013, "epoch": 1.64568345323741, "step": 1830 }, { "epoch": 1.64568345323741, "step": 1830, "torque_loss": 0.154373899102211 }, { "epoch": 1.6546762589928057, "grad_norm": 0.3102841079235077, "learning_rate": 7.356000000000001e-05, "loss": 0.0486, "step": 1840 }, { "action_loss": 0.026622623205184937, "epoch": 1.6546762589928057, "step": 1840 }, { "epoch": 1.6546762589928057, "step": 1840, "torque_loss": 0.19630520045757294 }, { "epoch": 1.6636690647482015, "grad_norm": 0.45624664425849915, "learning_rate": 7.396e-05, "loss": 0.0435, "step": 1850 }, { "action_loss": 0.02250516414642334, "epoch": 1.6636690647482015, "step": 1850 }, { "epoch": 1.6636690647482015, "step": 1850, "torque_loss": 0.23746125400066376 }, { "epoch": 1.6726618705035972, "grad_norm": 0.46605783700942993, "learning_rate": 7.436000000000001e-05, "loss": 0.046, "step": 1860 }, { "action_loss": 0.02257947064936161, "epoch": 1.6726618705035972, "step": 1860 }, { "epoch": 1.6726618705035972, "step": 1860, "torque_loss": 0.1459151804447174 }, { "epoch": 1.6816546762589928, "grad_norm": 0.31663253903388977, "learning_rate": 7.476000000000001e-05, "loss": 0.039, "step": 1870 }, { "action_loss": 0.024729304015636444, "epoch": 1.6816546762589928, "step": 1870 }, { "epoch": 1.6816546762589928, "step": 1870, "torque_loss": 0.2858298122882843 }, { "epoch": 1.6906474820143886, "grad_norm": 0.3196669816970825, "learning_rate": 7.516e-05, "loss": 0.0429, "step": 1880 }, { "action_loss": 0.01059796754270792, "epoch": 1.6906474820143886, "step": 1880 }, { "epoch": 1.6906474820143886, "step": 1880, "torque_loss": 0.1258942186832428 }, { "epoch": 1.699640287769784, "grad_norm": 0.36749035120010376, "learning_rate": 7.556000000000002e-05, "loss": 0.0503, "step": 1890 }, { "action_loss": 0.03292349353432655, "epoch": 1.699640287769784, "step": 1890 }, { "epoch": 1.699640287769784, "step": 1890, "torque_loss": 0.2669077515602112 }, { "epoch": 1.70863309352518, "grad_norm": 0.36936649680137634, "learning_rate": 7.596000000000001e-05, "loss": 0.0521, "step": 1900 }, { "action_loss": 0.019254645332694054, "epoch": 1.70863309352518, "step": 1900 }, { "epoch": 1.70863309352518, "step": 1900, "torque_loss": 0.19512955844402313 }, { "epoch": 1.7176258992805755, "grad_norm": 0.3657597601413727, "learning_rate": 7.636e-05, "loss": 0.0421, "step": 1910 }, { "action_loss": 0.021825207397341728, "epoch": 1.7176258992805755, "step": 1910 }, { "epoch": 1.7176258992805755, "step": 1910, "torque_loss": 0.1333484649658203 }, { "epoch": 1.7266187050359711, "grad_norm": 0.36959418654441833, "learning_rate": 7.676e-05, "loss": 0.0385, "step": 1920 }, { "action_loss": 0.0544886477291584, "epoch": 1.7266187050359711, "step": 1920 }, { "epoch": 1.7266187050359711, "step": 1920, "torque_loss": 0.23721154034137726 }, { "epoch": 1.735611510791367, "grad_norm": 0.3930166959762573, "learning_rate": 7.716e-05, "loss": 0.0472, "step": 1930 }, { "action_loss": 0.027298495173454285, "epoch": 1.735611510791367, "step": 1930 }, { "epoch": 1.735611510791367, "step": 1930, "torque_loss": 0.21410979330539703 }, { "epoch": 1.7446043165467626, "grad_norm": 0.35163021087646484, "learning_rate": 7.756e-05, "loss": 0.0369, "step": 1940 }, { "action_loss": 0.03405458480119705, "epoch": 1.7446043165467626, "step": 1940 }, { "epoch": 1.7446043165467626, "step": 1940, "torque_loss": 0.2524723708629608 }, { "epoch": 1.7535971223021583, "grad_norm": 0.3515278995037079, "learning_rate": 7.796e-05, "loss": 0.0477, "step": 1950 }, { "action_loss": 0.019607674330472946, "epoch": 1.7535971223021583, "step": 1950 }, { "epoch": 1.7535971223021583, "step": 1950, "torque_loss": 0.2449243813753128 }, { "epoch": 1.762589928057554, "grad_norm": 0.43362104892730713, "learning_rate": 7.836e-05, "loss": 0.0442, "step": 1960 }, { "action_loss": 0.014032046310603619, "epoch": 1.762589928057554, "step": 1960 }, { "epoch": 1.762589928057554, "step": 1960, "torque_loss": 0.12898419797420502 }, { "epoch": 1.7715827338129495, "grad_norm": 0.39862072467803955, "learning_rate": 7.876e-05, "loss": 0.0474, "step": 1970 }, { "action_loss": 0.032457154244184494, "epoch": 1.7715827338129495, "step": 1970 }, { "epoch": 1.7715827338129495, "step": 1970, "torque_loss": 0.26165053248405457 }, { "epoch": 1.7805755395683454, "grad_norm": 0.4859599471092224, "learning_rate": 7.916e-05, "loss": 0.0499, "step": 1980 }, { "action_loss": 0.014843955636024475, "epoch": 1.7805755395683454, "step": 1980 }, { "epoch": 1.7805755395683454, "step": 1980, "torque_loss": 0.14862468838691711 }, { "epoch": 1.789568345323741, "grad_norm": 0.355523943901062, "learning_rate": 7.956e-05, "loss": 0.0456, "step": 1990 }, { "action_loss": 0.009541374631226063, "epoch": 1.789568345323741, "step": 1990 }, { "epoch": 1.789568345323741, "step": 1990, "torque_loss": 0.1603613644838333 }, { "epoch": 1.7985611510791366, "grad_norm": 0.4821268320083618, "learning_rate": 7.996e-05, "loss": 0.0472, "step": 2000 }, { "action_loss": 0.02052808366715908, "epoch": 1.7985611510791366, "step": 2000 }, { "epoch": 1.7985611510791366, "step": 2000, "torque_loss": 0.19009864330291748 }, { "epoch": 1.8075539568345325, "grad_norm": 0.3827817738056183, "learning_rate": 8.036e-05, "loss": 0.0407, "step": 2010 }, { "action_loss": 0.026843050494790077, "epoch": 1.8075539568345325, "step": 2010 }, { "epoch": 1.8075539568345325, "step": 2010, "torque_loss": 0.2169172763824463 }, { "epoch": 1.816546762589928, "grad_norm": 0.452923059463501, "learning_rate": 8.076e-05, "loss": 0.0411, "step": 2020 }, { "action_loss": 0.013223419897258282, "epoch": 1.816546762589928, "step": 2020 }, { "epoch": 1.816546762589928, "step": 2020, "torque_loss": 0.17841041088104248 }, { "epoch": 1.8255395683453237, "grad_norm": 0.2576029896736145, "learning_rate": 8.116e-05, "loss": 0.0453, "step": 2030 }, { "action_loss": 0.013421990908682346, "epoch": 1.8255395683453237, "step": 2030 }, { "epoch": 1.8255395683453237, "step": 2030, "torque_loss": 0.19068045914173126 }, { "epoch": 1.8345323741007196, "grad_norm": 0.24561616778373718, "learning_rate": 8.156e-05, "loss": 0.0426, "step": 2040 }, { "action_loss": 0.03313381224870682, "epoch": 1.8345323741007196, "step": 2040 }, { "epoch": 1.8345323741007196, "step": 2040, "torque_loss": 0.21115486323833466 }, { "epoch": 1.843525179856115, "grad_norm": 0.21678286790847778, "learning_rate": 8.196000000000001e-05, "loss": 0.0421, "step": 2050 }, { "action_loss": 0.015007619746029377, "epoch": 1.843525179856115, "step": 2050 }, { "epoch": 1.843525179856115, "step": 2050, "torque_loss": 0.1198016107082367 }, { "epoch": 1.8525179856115108, "grad_norm": 0.43975526094436646, "learning_rate": 8.236e-05, "loss": 0.0462, "step": 2060 }, { "action_loss": 0.015317643992602825, "epoch": 1.8525179856115108, "step": 2060 }, { "epoch": 1.8525179856115108, "step": 2060, "torque_loss": 0.1807159036397934 }, { "epoch": 1.8615107913669064, "grad_norm": 0.36634665727615356, "learning_rate": 8.276e-05, "loss": 0.0445, "step": 2070 }, { "action_loss": 0.017739886417984962, "epoch": 1.8615107913669064, "step": 2070 }, { "epoch": 1.8615107913669064, "step": 2070, "torque_loss": 0.20066165924072266 }, { "epoch": 1.870503597122302, "grad_norm": 0.40998631715774536, "learning_rate": 8.316000000000001e-05, "loss": 0.0489, "step": 2080 }, { "action_loss": 0.0288152564316988, "epoch": 1.870503597122302, "step": 2080 }, { "epoch": 1.870503597122302, "step": 2080, "torque_loss": 0.19291074573993683 }, { "epoch": 1.879496402877698, "grad_norm": 0.37377384305000305, "learning_rate": 8.356e-05, "loss": 0.0428, "step": 2090 }, { "action_loss": 0.024058997631072998, "epoch": 1.879496402877698, "step": 2090 }, { "epoch": 1.879496402877698, "step": 2090, "torque_loss": 0.24980060756206512 }, { "epoch": 1.8884892086330936, "grad_norm": 0.508649468421936, "learning_rate": 8.396e-05, "loss": 0.0456, "step": 2100 }, { "action_loss": 0.01327530387789011, "epoch": 1.8884892086330936, "step": 2100 }, { "epoch": 1.8884892086330936, "step": 2100, "torque_loss": 0.14350302517414093 }, { "epoch": 1.8974820143884892, "grad_norm": 0.28819912672042847, "learning_rate": 8.436000000000001e-05, "loss": 0.0524, "step": 2110 }, { "action_loss": 0.019350232556462288, "epoch": 1.8974820143884892, "step": 2110 }, { "epoch": 1.8974820143884892, "step": 2110, "torque_loss": 0.19250692427158356 }, { "epoch": 1.906474820143885, "grad_norm": 0.4516009986400604, "learning_rate": 8.476000000000001e-05, "loss": 0.0552, "step": 2120 }, { "action_loss": 0.02031048573553562, "epoch": 1.906474820143885, "step": 2120 }, { "epoch": 1.906474820143885, "step": 2120, "torque_loss": 0.14181549847126007 }, { "epoch": 1.9154676258992804, "grad_norm": 0.539149820804596, "learning_rate": 8.516e-05, "loss": 0.0571, "step": 2130 }, { "action_loss": 0.02683078497648239, "epoch": 1.9154676258992804, "step": 2130 }, { "epoch": 1.9154676258992804, "step": 2130, "torque_loss": 0.21588434278964996 }, { "epoch": 1.9244604316546763, "grad_norm": 0.3655242621898651, "learning_rate": 8.556e-05, "loss": 0.0454, "step": 2140 }, { "action_loss": 0.02539152465760708, "epoch": 1.9244604316546763, "step": 2140 }, { "epoch": 1.9244604316546763, "step": 2140, "torque_loss": 0.23690839111804962 }, { "epoch": 1.933453237410072, "grad_norm": 0.4050748348236084, "learning_rate": 8.596000000000001e-05, "loss": 0.0428, "step": 2150 }, { "action_loss": 0.013983297161757946, "epoch": 1.933453237410072, "step": 2150 }, { "epoch": 1.933453237410072, "step": 2150, "torque_loss": 0.1881575584411621 }, { "epoch": 1.9424460431654675, "grad_norm": 0.3659866154193878, "learning_rate": 8.636e-05, "loss": 0.0534, "step": 2160 }, { "action_loss": 0.028642624616622925, "epoch": 1.9424460431654675, "step": 2160 }, { "epoch": 1.9424460431654675, "step": 2160, "torque_loss": 0.1963462233543396 }, { "epoch": 1.9514388489208634, "grad_norm": 0.3649277985095978, "learning_rate": 8.676e-05, "loss": 0.0404, "step": 2170 }, { "action_loss": 0.03269346058368683, "epoch": 1.9514388489208634, "step": 2170 }, { "epoch": 1.9514388489208634, "step": 2170, "torque_loss": 0.1324346661567688 }, { "epoch": 1.960431654676259, "grad_norm": 0.24420559406280518, "learning_rate": 8.716000000000001e-05, "loss": 0.034, "step": 2180 }, { "action_loss": 0.029169246554374695, "epoch": 1.960431654676259, "step": 2180 }, { "epoch": 1.960431654676259, "step": 2180, "torque_loss": 0.2014407068490982 }, { "epoch": 1.9694244604316546, "grad_norm": 0.42443519830703735, "learning_rate": 8.756000000000001e-05, "loss": 0.0436, "step": 2190 }, { "action_loss": 0.020181622356176376, "epoch": 1.9694244604316546, "step": 2190 }, { "epoch": 1.9694244604316546, "step": 2190, "torque_loss": 0.19559967517852783 }, { "epoch": 1.9784172661870505, "grad_norm": 0.4118806719779968, "learning_rate": 8.796e-05, "loss": 0.043, "step": 2200 }, { "action_loss": 0.037703122943639755, "epoch": 1.9784172661870505, "step": 2200 }, { "epoch": 1.9784172661870505, "step": 2200, "torque_loss": 0.27716460824012756 }, { "epoch": 1.987410071942446, "grad_norm": 0.40948978066444397, "learning_rate": 8.836000000000001e-05, "loss": 0.0547, "step": 2210 }, { "action_loss": 0.0374760776758194, "epoch": 1.987410071942446, "step": 2210 }, { "epoch": 1.987410071942446, "step": 2210, "torque_loss": 0.22233889997005463 }, { "epoch": 1.9964028776978417, "grad_norm": 0.4391648471355438, "learning_rate": 8.876e-05, "loss": 0.0475, "step": 2220 }, { "action_loss": 0.022225258871912956, "epoch": 1.9964028776978417, "step": 2220 }, { "epoch": 1.9964028776978417, "step": 2220, "torque_loss": 0.1769689917564392 }, { "epoch": 2.0053956834532376, "grad_norm": 0.34823888540267944, "learning_rate": 8.916e-05, "loss": 0.052, "step": 2230 }, { "action_loss": 0.02990572713315487, "epoch": 2.0053956834532376, "step": 2230 }, { "epoch": 2.0053956834532376, "step": 2230, "torque_loss": 0.2613406777381897 }, { "epoch": 2.014388489208633, "grad_norm": 0.33220112323760986, "learning_rate": 8.956e-05, "loss": 0.0492, "step": 2240 }, { "action_loss": 0.03958668187260628, "epoch": 2.014388489208633, "step": 2240 }, { "epoch": 2.014388489208633, "step": 2240, "torque_loss": 0.2796221971511841 }, { "epoch": 2.023381294964029, "grad_norm": 0.45846596360206604, "learning_rate": 8.996e-05, "loss": 0.0489, "step": 2250 }, { "action_loss": 0.026067500934004784, "epoch": 2.023381294964029, "step": 2250 }, { "epoch": 2.023381294964029, "step": 2250, "torque_loss": 0.16077114641666412 }, { "epoch": 2.0323741007194243, "grad_norm": 0.39046573638916016, "learning_rate": 9.036e-05, "loss": 0.0582, "step": 2260 }, { "action_loss": 0.013568985275924206, "epoch": 2.0323741007194243, "step": 2260 }, { "epoch": 2.0323741007194243, "step": 2260, "torque_loss": 0.13078060746192932 }, { "epoch": 2.04136690647482, "grad_norm": 0.34572726488113403, "learning_rate": 9.076e-05, "loss": 0.0506, "step": 2270 }, { "action_loss": 0.05939916893839836, "epoch": 2.04136690647482, "step": 2270 }, { "epoch": 2.04136690647482, "step": 2270, "torque_loss": 0.2702166736125946 }, { "epoch": 2.050359712230216, "grad_norm": 0.35027891397476196, "learning_rate": 9.116e-05, "loss": 0.0473, "step": 2280 }, { "action_loss": 0.03312865272164345, "epoch": 2.050359712230216, "step": 2280 }, { "epoch": 2.050359712230216, "step": 2280, "torque_loss": 0.2224312573671341 }, { "epoch": 2.0593525179856114, "grad_norm": 0.34058988094329834, "learning_rate": 9.156e-05, "loss": 0.0498, "step": 2290 }, { "action_loss": 0.024109231308102608, "epoch": 2.0593525179856114, "step": 2290 }, { "epoch": 2.0593525179856114, "step": 2290, "torque_loss": 0.1873619556427002 }, { "epoch": 2.068345323741007, "grad_norm": 0.3226891756057739, "learning_rate": 9.196000000000001e-05, "loss": 0.0485, "step": 2300 }, { "action_loss": 0.014645648188889027, "epoch": 2.068345323741007, "step": 2300 }, { "epoch": 2.068345323741007, "step": 2300, "torque_loss": 0.16370071470737457 }, { "epoch": 2.077338129496403, "grad_norm": 0.31375652551651, "learning_rate": 9.236e-05, "loss": 0.0436, "step": 2310 }, { "action_loss": 0.017293060198426247, "epoch": 2.077338129496403, "step": 2310 }, { "epoch": 2.077338129496403, "step": 2310, "torque_loss": 0.14639590680599213 }, { "epoch": 2.0863309352517985, "grad_norm": 0.4027572572231293, "learning_rate": 9.276e-05, "loss": 0.0475, "step": 2320 }, { "action_loss": 0.017225148156285286, "epoch": 2.0863309352517985, "step": 2320 }, { "epoch": 2.0863309352517985, "step": 2320, "torque_loss": 0.11395815759897232 }, { "epoch": 2.0953237410071943, "grad_norm": 0.6365926265716553, "learning_rate": 9.316000000000001e-05, "loss": 0.0482, "step": 2330 }, { "action_loss": 0.01944718509912491, "epoch": 2.0953237410071943, "step": 2330 }, { "epoch": 2.0953237410071943, "step": 2330, "torque_loss": 0.15248285233974457 }, { "epoch": 2.1043165467625897, "grad_norm": 0.3276452124118805, "learning_rate": 9.356e-05, "loss": 0.0433, "step": 2340 }, { "action_loss": 0.012384705245494843, "epoch": 2.1043165467625897, "step": 2340 }, { "epoch": 2.1043165467625897, "step": 2340, "torque_loss": 0.21659469604492188 }, { "epoch": 2.1133093525179856, "grad_norm": 0.30678918957710266, "learning_rate": 9.396e-05, "loss": 0.0446, "step": 2350 }, { "action_loss": 0.018852153792977333, "epoch": 2.1133093525179856, "step": 2350 }, { "epoch": 2.1133093525179856, "step": 2350, "torque_loss": 0.1846916675567627 }, { "epoch": 2.1223021582733814, "grad_norm": 0.22968529164791107, "learning_rate": 9.436e-05, "loss": 0.0438, "step": 2360 }, { "action_loss": 0.022145768627524376, "epoch": 2.1223021582733814, "step": 2360 }, { "epoch": 2.1223021582733814, "step": 2360, "torque_loss": 0.1710815280675888 }, { "epoch": 2.131294964028777, "grad_norm": 0.37108904123306274, "learning_rate": 9.476000000000001e-05, "loss": 0.0479, "step": 2370 }, { "action_loss": 0.018300378695130348, "epoch": 2.131294964028777, "step": 2370 }, { "epoch": 2.131294964028777, "step": 2370, "torque_loss": 0.2471265345811844 }, { "epoch": 2.1402877697841727, "grad_norm": 0.5163068771362305, "learning_rate": 9.516e-05, "loss": 0.045, "step": 2380 }, { "action_loss": 0.013093468733131886, "epoch": 2.1402877697841727, "step": 2380 }, { "epoch": 2.1402877697841727, "step": 2380, "torque_loss": 0.18420164287090302 }, { "epoch": 2.1492805755395685, "grad_norm": 0.3325856626033783, "learning_rate": 9.556e-05, "loss": 0.0533, "step": 2390 }, { "action_loss": 0.025514045730233192, "epoch": 2.1492805755395685, "step": 2390 }, { "epoch": 2.1492805755395685, "step": 2390, "torque_loss": 0.1627814918756485 }, { "epoch": 2.158273381294964, "grad_norm": 0.46762335300445557, "learning_rate": 9.596000000000001e-05, "loss": 0.0443, "step": 2400 }, { "action_loss": 0.019961485639214516, "epoch": 2.158273381294964, "step": 2400 }, { "epoch": 2.158273381294964, "step": 2400, "torque_loss": 0.11923838406801224 }, { "epoch": 2.16726618705036, "grad_norm": 0.660432755947113, "learning_rate": 9.636e-05, "loss": 0.05, "step": 2410 }, { "action_loss": 0.02841023914515972, "epoch": 2.16726618705036, "step": 2410 }, { "epoch": 2.16726618705036, "step": 2410, "torque_loss": 0.18981443345546722 }, { "epoch": 2.176258992805755, "grad_norm": 0.4303589463233948, "learning_rate": 9.676e-05, "loss": 0.0503, "step": 2420 }, { "action_loss": 0.03526679798960686, "epoch": 2.176258992805755, "step": 2420 }, { "epoch": 2.176258992805755, "step": 2420, "torque_loss": 0.22785712778568268 }, { "epoch": 2.185251798561151, "grad_norm": 0.3792738616466522, "learning_rate": 9.716000000000001e-05, "loss": 0.0429, "step": 2430 }, { "action_loss": 0.01723029650747776, "epoch": 2.185251798561151, "step": 2430 }, { "epoch": 2.185251798561151, "step": 2430, "torque_loss": 0.1684281826019287 }, { "epoch": 2.194244604316547, "grad_norm": 0.2985127866268158, "learning_rate": 9.756000000000001e-05, "loss": 0.0427, "step": 2440 }, { "action_loss": 0.027929628267884254, "epoch": 2.194244604316547, "step": 2440 }, { "epoch": 2.194244604316547, "step": 2440, "torque_loss": 0.17521901428699493 }, { "epoch": 2.2032374100719423, "grad_norm": 0.24106746912002563, "learning_rate": 9.796e-05, "loss": 0.0487, "step": 2450 }, { "action_loss": 0.013967827893793583, "epoch": 2.2032374100719423, "step": 2450 }, { "epoch": 2.2032374100719423, "step": 2450, "torque_loss": 0.15822277963161469 }, { "epoch": 2.212230215827338, "grad_norm": 0.29848435521125793, "learning_rate": 9.836000000000001e-05, "loss": 0.0475, "step": 2460 }, { "action_loss": 0.020892038941383362, "epoch": 2.212230215827338, "step": 2460 }, { "epoch": 2.212230215827338, "step": 2460, "torque_loss": 0.21920152008533478 }, { "epoch": 2.221223021582734, "grad_norm": 0.2793034613132477, "learning_rate": 9.876000000000001e-05, "loss": 0.0404, "step": 2470 }, { "action_loss": 0.014425848610699177, "epoch": 2.221223021582734, "step": 2470 }, { "epoch": 2.221223021582734, "step": 2470, "torque_loss": 0.12352514266967773 }, { "epoch": 2.2302158273381294, "grad_norm": 0.4203096330165863, "learning_rate": 9.916e-05, "loss": 0.0458, "step": 2480 }, { "action_loss": 0.054599959403276443, "epoch": 2.2302158273381294, "step": 2480 }, { "epoch": 2.2302158273381294, "step": 2480, "torque_loss": 0.1926569938659668 }, { "epoch": 2.2392086330935252, "grad_norm": 0.41215425729751587, "learning_rate": 9.956e-05, "loss": 0.0508, "step": 2490 }, { "action_loss": 0.02414984069764614, "epoch": 2.2392086330935252, "step": 2490 }, { "epoch": 2.2392086330935252, "step": 2490, "torque_loss": 0.1802733689546585 }, { "epoch": 2.2482014388489207, "grad_norm": 0.26738131046295166, "learning_rate": 9.996000000000001e-05, "loss": 0.0507, "step": 2500 }, { "action_loss": 0.04011745750904083, "epoch": 2.2482014388489207, "step": 2500 }, { "epoch": 2.2482014388489207, "step": 2500, "torque_loss": 0.2768697142601013 }, { "epoch": 2.2571942446043165, "grad_norm": 0.3537469506263733, "learning_rate": 9.999999114196196e-05, "loss": 0.0472, "step": 2510 }, { "action_loss": 0.028092751279473305, "epoch": 2.2571942446043165, "step": 2510 }, { "epoch": 2.2571942446043165, "step": 2510, "torque_loss": 0.2762439548969269 }, { "epoch": 2.2661870503597124, "grad_norm": 0.40902188420295715, "learning_rate": 9.99999605215876e-05, "loss": 0.0433, "step": 2520 }, { "action_loss": 0.031259629875421524, "epoch": 2.2661870503597124, "step": 2520 }, { "epoch": 2.2661870503597124, "step": 2520, "torque_loss": 0.24397651851177216 }, { "epoch": 2.2751798561151078, "grad_norm": 0.5216095447540283, "learning_rate": 9.999990802953179e-05, "loss": 0.0438, "step": 2530 }, { "action_loss": 0.031690508127212524, "epoch": 2.2751798561151078, "step": 2530 }, { "epoch": 2.2751798561151078, "step": 2530, "torque_loss": 0.26015034317970276 }, { "epoch": 2.2841726618705036, "grad_norm": 0.28139474987983704, "learning_rate": 9.99998336658175e-05, "loss": 0.0576, "step": 2540 }, { "action_loss": 0.033493656665086746, "epoch": 2.2841726618705036, "step": 2540 }, { "epoch": 2.2841726618705036, "step": 2540, "torque_loss": 0.27626267075538635 }, { "epoch": 2.2931654676258995, "grad_norm": 0.4077867865562439, "learning_rate": 9.999973743047727e-05, "loss": 0.0433, "step": 2550 }, { "action_loss": 0.014439006336033344, "epoch": 2.2931654676258995, "step": 2550 }, { "epoch": 2.2931654676258995, "step": 2550, "torque_loss": 0.14382289350032806 }, { "epoch": 2.302158273381295, "grad_norm": 0.47085511684417725, "learning_rate": 9.999961932355319e-05, "loss": 0.0548, "step": 2560 }, { "action_loss": 0.02764672040939331, "epoch": 2.302158273381295, "step": 2560 }, { "epoch": 2.302158273381295, "step": 2560, "torque_loss": 0.21654850244522095 }, { "epoch": 2.3111510791366907, "grad_norm": 0.34810671210289, "learning_rate": 9.999947934509693e-05, "loss": 0.0512, "step": 2570 }, { "action_loss": 0.03147466853260994, "epoch": 2.3111510791366907, "step": 2570 }, { "epoch": 2.3111510791366907, "step": 2570, "torque_loss": 0.17439986765384674 }, { "epoch": 2.3201438848920866, "grad_norm": 0.3979521095752716, "learning_rate": 9.999931749516971e-05, "loss": 0.0512, "step": 2580 }, { "action_loss": 0.05042010545730591, "epoch": 2.3201438848920866, "step": 2580 }, { "epoch": 2.3201438848920866, "step": 2580, "torque_loss": 0.22158078849315643 }, { "epoch": 2.329136690647482, "grad_norm": 0.6061447262763977, "learning_rate": 9.999913377384233e-05, "loss": 0.0488, "step": 2590 }, { "action_loss": 0.02679470181465149, "epoch": 2.329136690647482, "step": 2590 }, { "epoch": 2.329136690647482, "step": 2590, "torque_loss": 0.19954311847686768 }, { "epoch": 2.338129496402878, "grad_norm": 0.553011417388916, "learning_rate": 9.999892818119517e-05, "loss": 0.0499, "step": 2600 }, { "action_loss": 0.02111467532813549, "epoch": 2.338129496402878, "step": 2600 }, { "epoch": 2.338129496402878, "step": 2600, "torque_loss": 0.13508151471614838 }, { "epoch": 2.347122302158273, "grad_norm": 0.3649100065231323, "learning_rate": 9.999870071731814e-05, "loss": 0.0508, "step": 2610 }, { "action_loss": 0.0232559684664011, "epoch": 2.347122302158273, "step": 2610 }, { "epoch": 2.347122302158273, "step": 2610, "torque_loss": 0.18895037472248077 }, { "epoch": 2.356115107913669, "grad_norm": 0.28968653082847595, "learning_rate": 9.999845138231076e-05, "loss": 0.045, "step": 2620 }, { "action_loss": 0.034350067377090454, "epoch": 2.356115107913669, "step": 2620 }, { "epoch": 2.356115107913669, "step": 2620, "torque_loss": 0.20595203340053558 }, { "epoch": 2.365107913669065, "grad_norm": 0.280923455953598, "learning_rate": 9.999818017628208e-05, "loss": 0.0468, "step": 2630 }, { "action_loss": 0.032667115330696106, "epoch": 2.365107913669065, "step": 2630 }, { "epoch": 2.365107913669065, "step": 2630, "torque_loss": 0.278799444437027 }, { "epoch": 2.3741007194244603, "grad_norm": 0.31720927357673645, "learning_rate": 9.999788709935078e-05, "loss": 0.0496, "step": 2640 }, { "action_loss": 0.057680532336235046, "epoch": 2.3741007194244603, "step": 2640 }, { "epoch": 2.3741007194244603, "step": 2640, "torque_loss": 0.28182926774024963 }, { "epoch": 2.383093525179856, "grad_norm": 0.27461472153663635, "learning_rate": 9.9997572151645e-05, "loss": 0.047, "step": 2650 }, { "action_loss": 0.018457312136888504, "epoch": 2.383093525179856, "step": 2650 }, { "epoch": 2.383093525179856, "step": 2650, "torque_loss": 0.18606983125209808 }, { "epoch": 2.3920863309352516, "grad_norm": 0.24027760326862335, "learning_rate": 9.999723533330254e-05, "loss": 0.0417, "step": 2660 }, { "action_loss": 0.06288313120603561, "epoch": 2.3920863309352516, "step": 2660 }, { "epoch": 2.3920863309352516, "step": 2660, "torque_loss": 0.24269211292266846 }, { "epoch": 2.4010791366906474, "grad_norm": 0.4183937907218933, "learning_rate": 9.999687664447074e-05, "loss": 0.0521, "step": 2670 }, { "action_loss": 0.020199408754706383, "epoch": 2.4010791366906474, "step": 2670 }, { "epoch": 2.4010791366906474, "step": 2670, "torque_loss": 0.15737378597259521 }, { "epoch": 2.4100719424460433, "grad_norm": 0.482683926820755, "learning_rate": 9.99964960853065e-05, "loss": 0.0456, "step": 2680 }, { "action_loss": 0.023698121309280396, "epoch": 2.4100719424460433, "step": 2680 }, { "epoch": 2.4100719424460433, "step": 2680, "torque_loss": 0.1470421701669693 }, { "epoch": 2.4190647482014387, "grad_norm": 0.3398328721523285, "learning_rate": 9.999609365597627e-05, "loss": 0.0404, "step": 2690 }, { "action_loss": 0.02120075561106205, "epoch": 2.4190647482014387, "step": 2690 }, { "epoch": 2.4190647482014387, "step": 2690, "torque_loss": 0.24876564741134644 }, { "epoch": 2.4280575539568345, "grad_norm": 0.25152382254600525, "learning_rate": 9.99956693566561e-05, "loss": 0.0421, "step": 2700 }, { "action_loss": 0.013934199698269367, "epoch": 2.4280575539568345, "step": 2700 }, { "epoch": 2.4280575539568345, "step": 2700, "torque_loss": 0.14443297684192657 }, { "epoch": 2.4370503597122304, "grad_norm": 0.3924514651298523, "learning_rate": 9.99952231875316e-05, "loss": 0.0418, "step": 2710 }, { "action_loss": 0.02063358575105667, "epoch": 2.4370503597122304, "step": 2710 }, { "epoch": 2.4370503597122304, "step": 2710, "torque_loss": 0.17202818393707275 }, { "epoch": 2.446043165467626, "grad_norm": 0.3540758490562439, "learning_rate": 9.999475514879795e-05, "loss": 0.0488, "step": 2720 }, { "action_loss": 0.011805173940956593, "epoch": 2.446043165467626, "step": 2720 }, { "epoch": 2.446043165467626, "step": 2720, "torque_loss": 0.11874985694885254 }, { "epoch": 2.4550359712230216, "grad_norm": 0.35547706484794617, "learning_rate": 9.999426524065984e-05, "loss": 0.0377, "step": 2730 }, { "action_loss": 0.020048359408974648, "epoch": 2.4550359712230216, "step": 2730 }, { "epoch": 2.4550359712230216, "step": 2730, "torque_loss": 0.21879935264587402 }, { "epoch": 2.4640287769784175, "grad_norm": 0.3920975923538208, "learning_rate": 9.999375346333162e-05, "loss": 0.0433, "step": 2740 }, { "action_loss": 0.02066042833030224, "epoch": 2.4640287769784175, "step": 2740 }, { "epoch": 2.4640287769784175, "step": 2740, "torque_loss": 0.20472292602062225 }, { "epoch": 2.473021582733813, "grad_norm": 0.32385510206222534, "learning_rate": 9.999321981703715e-05, "loss": 0.0452, "step": 2750 }, { "action_loss": 0.021466275677084923, "epoch": 2.473021582733813, "step": 2750 }, { "epoch": 2.473021582733813, "step": 2750, "torque_loss": 0.14635047316551208 }, { "epoch": 2.4820143884892087, "grad_norm": 0.4420037269592285, "learning_rate": 9.999266430200985e-05, "loss": 0.0428, "step": 2760 }, { "action_loss": 0.024416213855147362, "epoch": 2.4820143884892087, "step": 2760 }, { "epoch": 2.4820143884892087, "step": 2760, "torque_loss": 0.25987574458122253 }, { "epoch": 2.491007194244604, "grad_norm": 0.3671777546405792, "learning_rate": 9.999208691849271e-05, "loss": 0.0394, "step": 2770 }, { "action_loss": 0.028826376423239708, "epoch": 2.491007194244604, "step": 2770 }, { "epoch": 2.491007194244604, "step": 2770, "torque_loss": 0.2490323930978775 }, { "epoch": 2.5, "grad_norm": 0.37988388538360596, "learning_rate": 9.999148766673832e-05, "loss": 0.0452, "step": 2780 }, { "action_loss": 0.025146394968032837, "epoch": 2.5, "step": 2780 }, { "epoch": 2.5, "step": 2780, "torque_loss": 0.1857011914253235 }, { "epoch": 2.508992805755396, "grad_norm": 0.34698861837387085, "learning_rate": 9.999086654700881e-05, "loss": 0.0492, "step": 2790 }, { "action_loss": 0.015385262668132782, "epoch": 2.508992805755396, "step": 2790 }, { "epoch": 2.508992805755396, "step": 2790, "torque_loss": 0.14464743435382843 }, { "epoch": 2.5179856115107913, "grad_norm": 0.36120206117630005, "learning_rate": 9.999022355957588e-05, "loss": 0.0426, "step": 2800 }, { "action_loss": 0.029763976112008095, "epoch": 2.5179856115107913, "step": 2800 }, { "epoch": 2.5179856115107913, "step": 2800, "torque_loss": 0.20245511829853058 }, { "epoch": 2.526978417266187, "grad_norm": 0.42245301604270935, "learning_rate": 9.998955870472079e-05, "loss": 0.044, "step": 2810 }, { "action_loss": 0.026052730157971382, "epoch": 2.526978417266187, "step": 2810 }, { "epoch": 2.526978417266187, "step": 2810, "torque_loss": 0.19055677950382233 }, { "epoch": 2.5359712230215825, "grad_norm": 0.37630900740623474, "learning_rate": 9.998887198273437e-05, "loss": 0.0402, "step": 2820 }, { "action_loss": 0.016962936148047447, "epoch": 2.5359712230215825, "step": 2820 }, { "epoch": 2.5359712230215825, "step": 2820, "torque_loss": 0.1914600282907486 }, { "epoch": 2.5449640287769784, "grad_norm": 0.4241574704647064, "learning_rate": 9.998816339391701e-05, "loss": 0.0527, "step": 2830 }, { "action_loss": 0.032596468925476074, "epoch": 2.5449640287769784, "step": 2830 }, { "epoch": 2.5449640287769784, "step": 2830, "torque_loss": 0.23238764703273773 }, { "epoch": 2.553956834532374, "grad_norm": 0.47033342719078064, "learning_rate": 9.998743293857868e-05, "loss": 0.0481, "step": 2840 }, { "action_loss": 0.013140435330569744, "epoch": 2.553956834532374, "step": 2840 }, { "epoch": 2.553956834532374, "step": 2840, "torque_loss": 0.15100444853305817 }, { "epoch": 2.56294964028777, "grad_norm": 0.3070027530193329, "learning_rate": 9.998668061703891e-05, "loss": 0.0411, "step": 2850 }, { "action_loss": 0.022085009142756462, "epoch": 2.56294964028777, "step": 2850 }, { "epoch": 2.56294964028777, "step": 2850, "torque_loss": 0.2387438267469406 }, { "epoch": 2.5719424460431655, "grad_norm": 0.23611964285373688, "learning_rate": 9.998590642962679e-05, "loss": 0.0447, "step": 2860 }, { "action_loss": 0.00974121131002903, "epoch": 2.5719424460431655, "step": 2860 }, { "epoch": 2.5719424460431655, "step": 2860, "torque_loss": 0.20096363127231598 }, { "epoch": 2.5809352517985613, "grad_norm": 0.3491823971271515, "learning_rate": 9.998511037668095e-05, "loss": 0.0476, "step": 2870 }, { "action_loss": 0.015508655458688736, "epoch": 2.5809352517985613, "step": 2870 }, { "epoch": 2.5809352517985613, "step": 2870, "torque_loss": 0.22831940650939941 }, { "epoch": 2.5899280575539567, "grad_norm": 0.4288817346096039, "learning_rate": 9.998429245854964e-05, "loss": 0.0387, "step": 2880 }, { "action_loss": 0.03376004472374916, "epoch": 2.5899280575539567, "step": 2880 }, { "epoch": 2.5899280575539567, "step": 2880, "torque_loss": 0.19119231402873993 }, { "epoch": 2.5989208633093526, "grad_norm": 0.2516676187515259, "learning_rate": 9.998345267559064e-05, "loss": 0.0456, "step": 2890 }, { "action_loss": 0.020899498835206032, "epoch": 2.5989208633093526, "step": 2890 }, { "epoch": 2.5989208633093526, "step": 2890, "torque_loss": 0.22368502616882324 }, { "epoch": 2.6079136690647484, "grad_norm": 0.42055991291999817, "learning_rate": 9.998259102817129e-05, "loss": 0.0473, "step": 2900 }, { "action_loss": 0.021521294489502907, "epoch": 2.6079136690647484, "step": 2900 }, { "epoch": 2.6079136690647484, "step": 2900, "torque_loss": 0.14742891490459442 }, { "epoch": 2.616906474820144, "grad_norm": 0.35654258728027344, "learning_rate": 9.99817075166685e-05, "loss": 0.0445, "step": 2910 }, { "action_loss": 0.016283631324768066, "epoch": 2.616906474820144, "step": 2910 }, { "epoch": 2.616906474820144, "step": 2910, "torque_loss": 0.21817581355571747 }, { "epoch": 2.6258992805755397, "grad_norm": 0.31388017535209656, "learning_rate": 9.998080214146878e-05, "loss": 0.0436, "step": 2920 }, { "action_loss": 0.028618326410651207, "epoch": 2.6258992805755397, "step": 2920 }, { "epoch": 2.6258992805755397, "step": 2920, "torque_loss": 0.17718751728534698 }, { "epoch": 2.634892086330935, "grad_norm": 0.34708213806152344, "learning_rate": 9.997987490296813e-05, "loss": 0.0423, "step": 2930 }, { "action_loss": 0.01951897330582142, "epoch": 2.634892086330935, "step": 2930 }, { "epoch": 2.634892086330935, "step": 2930, "torque_loss": 0.2793252468109131 }, { "epoch": 2.643884892086331, "grad_norm": 0.3894035816192627, "learning_rate": 9.99789258015722e-05, "loss": 0.0489, "step": 2940 }, { "action_loss": 0.05399833247065544, "epoch": 2.643884892086331, "step": 2940 }, { "epoch": 2.643884892086331, "step": 2940, "torque_loss": 0.29225966334342957 }, { "epoch": 2.652877697841727, "grad_norm": 0.2990155518054962, "learning_rate": 9.997795483769611e-05, "loss": 0.0524, "step": 2950 }, { "action_loss": 0.03139448165893555, "epoch": 2.652877697841727, "step": 2950 }, { "epoch": 2.652877697841727, "step": 2950, "torque_loss": 0.26749280095100403 }, { "epoch": 2.661870503597122, "grad_norm": 0.38336148858070374, "learning_rate": 9.997696201176462e-05, "loss": 0.0487, "step": 2960 }, { "action_loss": 0.012963873334228992, "epoch": 2.661870503597122, "step": 2960 }, { "epoch": 2.661870503597122, "step": 2960, "torque_loss": 0.1689157336950302 }, { "epoch": 2.670863309352518, "grad_norm": 0.31800925731658936, "learning_rate": 9.997594732421203e-05, "loss": 0.0456, "step": 2970 }, { "action_loss": 0.028289735317230225, "epoch": 2.670863309352518, "step": 2970 }, { "epoch": 2.670863309352518, "step": 2970, "torque_loss": 0.21712636947631836 }, { "epoch": 2.6798561151079134, "grad_norm": 0.30736592411994934, "learning_rate": 9.997491077548217e-05, "loss": 0.0405, "step": 2980 }, { "action_loss": 0.014346574433147907, "epoch": 2.6798561151079134, "step": 2980 }, { "epoch": 2.6798561151079134, "step": 2980, "torque_loss": 0.1617565006017685 }, { "epoch": 2.6888489208633093, "grad_norm": 0.32628270983695984, "learning_rate": 9.997385236602851e-05, "loss": 0.0486, "step": 2990 }, { "action_loss": 0.020309695973992348, "epoch": 2.6888489208633093, "step": 2990 }, { "epoch": 2.6888489208633093, "step": 2990, "torque_loss": 0.32886746525764465 }, { "epoch": 2.697841726618705, "grad_norm": 0.34042027592658997, "learning_rate": 9.997277209631399e-05, "loss": 0.0541, "step": 3000 }, { "action_loss": 0.019967321306467056, "epoch": 2.697841726618705, "step": 3000 }, { "epoch": 2.697841726618705, "step": 3000, "torque_loss": 0.15111571550369263 }, { "epoch": 2.706834532374101, "grad_norm": 0.29430168867111206, "learning_rate": 9.997166996681118e-05, "loss": 0.0446, "step": 3010 }, { "action_loss": 0.03199227154254913, "epoch": 2.706834532374101, "step": 3010 }, { "epoch": 2.706834532374101, "step": 3010, "torque_loss": 0.19259095191955566 }, { "epoch": 2.7158273381294964, "grad_norm": 0.3936454653739929, "learning_rate": 9.997054597800218e-05, "loss": 0.0466, "step": 3020 }, { "action_loss": 0.02706374041736126, "epoch": 2.7158273381294964, "step": 3020 }, { "epoch": 2.7158273381294964, "step": 3020, "torque_loss": 0.22391079366207123 }, { "epoch": 2.7248201438848922, "grad_norm": 0.3603522777557373, "learning_rate": 9.996940013037866e-05, "loss": 0.0436, "step": 3030 }, { "action_loss": 0.030649809166789055, "epoch": 2.7248201438848922, "step": 3030 }, { "epoch": 2.7248201438848922, "step": 3030, "torque_loss": 0.23948407173156738 }, { "epoch": 2.7338129496402876, "grad_norm": 0.23252908885478973, "learning_rate": 9.996823242444186e-05, "loss": 0.0478, "step": 3040 }, { "action_loss": 0.01473324652761221, "epoch": 2.7338129496402876, "step": 3040 }, { "epoch": 2.7338129496402876, "step": 3040, "torque_loss": 0.1457158327102661 }, { "epoch": 2.7428057553956835, "grad_norm": 0.51251620054245, "learning_rate": 9.996704286070258e-05, "loss": 0.0525, "step": 3050 }, { "action_loss": 0.017851168289780617, "epoch": 2.7428057553956835, "step": 3050 }, { "epoch": 2.7428057553956835, "step": 3050, "torque_loss": 0.1922844648361206 }, { "epoch": 2.7517985611510793, "grad_norm": 0.29162734746932983, "learning_rate": 9.996583143968115e-05, "loss": 0.0413, "step": 3060 }, { "action_loss": 0.01625612936913967, "epoch": 2.7517985611510793, "step": 3060 }, { "epoch": 2.7517985611510793, "step": 3060, "torque_loss": 0.16007520258426666 }, { "epoch": 2.7607913669064748, "grad_norm": 0.30119791626930237, "learning_rate": 9.99645981619075e-05, "loss": 0.0411, "step": 3070 }, { "action_loss": 0.04256541654467583, "epoch": 2.7607913669064748, "step": 3070 }, { "epoch": 2.7607913669064748, "step": 3070, "torque_loss": 0.22109830379486084 }, { "epoch": 2.7697841726618706, "grad_norm": 0.36606770753860474, "learning_rate": 9.996334302792114e-05, "loss": 0.06, "step": 3080 }, { "action_loss": 0.04820352420210838, "epoch": 2.7697841726618706, "step": 3080 }, { "epoch": 2.7697841726618706, "step": 3080, "torque_loss": 0.23285503685474396 }, { "epoch": 2.778776978417266, "grad_norm": 0.2797081172466278, "learning_rate": 9.996206603827105e-05, "loss": 0.0468, "step": 3090 }, { "action_loss": 0.01822231337428093, "epoch": 2.778776978417266, "step": 3090 }, { "epoch": 2.778776978417266, "step": 3090, "torque_loss": 0.19264888763427734 }, { "epoch": 2.787769784172662, "grad_norm": 0.2850756347179413, "learning_rate": 9.996076719351587e-05, "loss": 0.0452, "step": 3100 }, { "action_loss": 0.027726342901587486, "epoch": 2.787769784172662, "step": 3100 }, { "epoch": 2.787769784172662, "step": 3100, "torque_loss": 0.16310372948646545 }, { "epoch": 2.7967625899280577, "grad_norm": 0.3288310766220093, "learning_rate": 9.995944649422374e-05, "loss": 0.0435, "step": 3110 }, { "action_loss": 0.03678765892982483, "epoch": 2.7967625899280577, "step": 3110 }, { "epoch": 2.7967625899280577, "step": 3110, "torque_loss": 0.2424965500831604 }, { "epoch": 2.805755395683453, "grad_norm": 0.3220342993736267, "learning_rate": 9.995810394097239e-05, "loss": 0.0455, "step": 3120 }, { "action_loss": 0.03402495011687279, "epoch": 2.805755395683453, "step": 3120 }, { "epoch": 2.805755395683453, "step": 3120, "torque_loss": 0.2815920114517212 }, { "epoch": 2.814748201438849, "grad_norm": 0.36193981766700745, "learning_rate": 9.995673953434909e-05, "loss": 0.0456, "step": 3130 }, { "action_loss": 0.012627596966922283, "epoch": 2.814748201438849, "step": 3130 }, { "epoch": 2.814748201438849, "step": 3130, "torque_loss": 0.21153785288333893 }, { "epoch": 2.8237410071942444, "grad_norm": 0.39594170451164246, "learning_rate": 9.995535327495068e-05, "loss": 0.0415, "step": 3140 }, { "action_loss": 0.0251734908670187, "epoch": 2.8237410071942444, "step": 3140 }, { "epoch": 2.8237410071942444, "step": 3140, "torque_loss": 0.2391372174024582 }, { "epoch": 2.83273381294964, "grad_norm": 0.27410489320755005, "learning_rate": 9.995394516338355e-05, "loss": 0.0433, "step": 3150 }, { "action_loss": 0.018191605806350708, "epoch": 2.83273381294964, "step": 3150 }, { "epoch": 2.83273381294964, "step": 3150, "torque_loss": 0.21286292374134064 }, { "epoch": 2.841726618705036, "grad_norm": 0.3401366174221039, "learning_rate": 9.995251520026367e-05, "loss": 0.04, "step": 3160 }, { "action_loss": 0.028230035677552223, "epoch": 2.841726618705036, "step": 3160 }, { "epoch": 2.841726618705036, "step": 3160, "torque_loss": 0.3039197325706482 }, { "epoch": 2.850719424460432, "grad_norm": 0.268216997385025, "learning_rate": 9.995106338621656e-05, "loss": 0.0447, "step": 3170 }, { "action_loss": 0.014690262265503407, "epoch": 2.850719424460432, "step": 3170 }, { "epoch": 2.850719424460432, "step": 3170, "torque_loss": 0.17608098685741425 }, { "epoch": 2.8597122302158273, "grad_norm": 0.3213290274143219, "learning_rate": 9.994958972187726e-05, "loss": 0.0524, "step": 3180 }, { "action_loss": 0.01623285748064518, "epoch": 2.8597122302158273, "step": 3180 }, { "epoch": 2.8597122302158273, "step": 3180, "torque_loss": 0.15132085978984833 }, { "epoch": 2.868705035971223, "grad_norm": 0.30003130435943604, "learning_rate": 9.994809420789044e-05, "loss": 0.0462, "step": 3190 }, { "action_loss": 0.03167622536420822, "epoch": 2.868705035971223, "step": 3190 }, { "epoch": 2.868705035971223, "step": 3190, "torque_loss": 0.2320825308561325 }, { "epoch": 2.8776978417266186, "grad_norm": 0.4774527847766876, "learning_rate": 9.994657684491027e-05, "loss": 0.0573, "step": 3200 }, { "action_loss": 0.019984053447842598, "epoch": 2.8776978417266186, "step": 3200 }, { "epoch": 2.8776978417266186, "step": 3200, "torque_loss": 0.11896345764398575 }, { "epoch": 2.8866906474820144, "grad_norm": 0.47682732343673706, "learning_rate": 9.994503763360048e-05, "loss": 0.0465, "step": 3210 }, { "action_loss": 0.02381005324423313, "epoch": 2.8866906474820144, "step": 3210 }, { "epoch": 2.8866906474820144, "step": 3210, "torque_loss": 0.18984897434711456 }, { "epoch": 2.8956834532374103, "grad_norm": 0.27741166949272156, "learning_rate": 9.99434765746344e-05, "loss": 0.0443, "step": 3220 }, { "action_loss": 0.02149048261344433, "epoch": 2.8956834532374103, "step": 3220 }, { "epoch": 2.8956834532374103, "step": 3220, "torque_loss": 0.11891814321279526 }, { "epoch": 2.9046762589928057, "grad_norm": 0.3487516939640045, "learning_rate": 9.994189366869488e-05, "loss": 0.0427, "step": 3230 }, { "action_loss": 0.02057950757443905, "epoch": 2.9046762589928057, "step": 3230 }, { "epoch": 2.9046762589928057, "step": 3230, "torque_loss": 0.19941584765911102 }, { "epoch": 2.9136690647482015, "grad_norm": 0.26850852370262146, "learning_rate": 9.994028891647433e-05, "loss": 0.0436, "step": 3240 }, { "action_loss": 0.029543833807110786, "epoch": 2.9136690647482015, "step": 3240 }, { "epoch": 2.9136690647482015, "step": 3240, "torque_loss": 0.18141476809978485 }, { "epoch": 2.922661870503597, "grad_norm": 0.1924889236688614, "learning_rate": 9.993866231867475e-05, "loss": 0.0469, "step": 3250 }, { "action_loss": 0.021875182166695595, "epoch": 2.922661870503597, "step": 3250 }, { "epoch": 2.922661870503597, "step": 3250, "torque_loss": 0.15827129781246185 }, { "epoch": 2.931654676258993, "grad_norm": 0.26236018538475037, "learning_rate": 9.993701387600762e-05, "loss": 0.0401, "step": 3260 }, { "action_loss": 0.029493289068341255, "epoch": 2.931654676258993, "step": 3260 }, { "epoch": 2.931654676258993, "step": 3260, "torque_loss": 0.25234535336494446 }, { "epoch": 2.9406474820143886, "grad_norm": 0.22226423025131226, "learning_rate": 9.993534358919408e-05, "loss": 0.0506, "step": 3270 }, { "action_loss": 0.019878238439559937, "epoch": 2.9406474820143886, "step": 3270 }, { "epoch": 2.9406474820143886, "step": 3270, "torque_loss": 0.14734746515750885 }, { "epoch": 2.949640287769784, "grad_norm": 0.40090757608413696, "learning_rate": 9.993365145896473e-05, "loss": 0.0489, "step": 3280 }, { "action_loss": 0.020166829228401184, "epoch": 2.949640287769784, "step": 3280 }, { "epoch": 2.949640287769784, "step": 3280, "torque_loss": 0.23923254013061523 }, { "epoch": 2.95863309352518, "grad_norm": 0.36292141675949097, "learning_rate": 9.993193748605977e-05, "loss": 0.0502, "step": 3290 }, { "action_loss": 0.02261490933597088, "epoch": 2.95863309352518, "step": 3290 }, { "epoch": 2.95863309352518, "step": 3290, "torque_loss": 0.2501518428325653 }, { "epoch": 2.9676258992805753, "grad_norm": 0.3994857370853424, "learning_rate": 9.993020167122898e-05, "loss": 0.0601, "step": 3300 }, { "action_loss": 0.02124435268342495, "epoch": 2.9676258992805753, "step": 3300 }, { "epoch": 2.9676258992805753, "step": 3300, "torque_loss": 0.19607341289520264 }, { "epoch": 2.976618705035971, "grad_norm": 0.22461672127246857, "learning_rate": 9.992844401523164e-05, "loss": 0.0412, "step": 3310 }, { "action_loss": 0.029191916808485985, "epoch": 2.976618705035971, "step": 3310 }, { "epoch": 2.976618705035971, "step": 3310, "torque_loss": 0.21270056068897247 }, { "epoch": 2.985611510791367, "grad_norm": 0.35049349069595337, "learning_rate": 9.992666451883661e-05, "loss": 0.0424, "step": 3320 }, { "action_loss": 0.018854429945349693, "epoch": 2.985611510791367, "step": 3320 }, { "epoch": 2.985611510791367, "step": 3320, "torque_loss": 0.15200971066951752 }, { "epoch": 2.994604316546763, "grad_norm": 0.3267016112804413, "learning_rate": 9.99248631828223e-05, "loss": 0.0482, "step": 3330 }, { "action_loss": 0.027130505070090294, "epoch": 2.994604316546763, "step": 3330 }, { "epoch": 2.994604316546763, "step": 3330, "torque_loss": 0.1646803766489029 }, { "epoch": 3.0035971223021583, "grad_norm": 0.3002015948295593, "learning_rate": 9.99230400079767e-05, "loss": 0.0392, "step": 3340 }, { "action_loss": 0.029561849310994148, "epoch": 3.0035971223021583, "step": 3340 }, { "epoch": 3.0035971223021583, "step": 3340, "torque_loss": 0.1991540640592575 }, { "epoch": 3.012589928057554, "grad_norm": 0.43256518244743347, "learning_rate": 9.992119499509728e-05, "loss": 0.0482, "step": 3350 }, { "action_loss": 0.01864507794380188, "epoch": 3.012589928057554, "step": 3350 }, { "epoch": 3.012589928057554, "step": 3350, "torque_loss": 0.22992654144763947 }, { "epoch": 3.0215827338129495, "grad_norm": 0.4205636978149414, "learning_rate": 9.991932814499114e-05, "loss": 0.0521, "step": 3360 }, { "action_loss": 0.022958016023039818, "epoch": 3.0215827338129495, "step": 3360 }, { "epoch": 3.0215827338129495, "step": 3360, "torque_loss": 0.20749163627624512 }, { "epoch": 3.0305755395683454, "grad_norm": 0.38024023175239563, "learning_rate": 9.991743945847493e-05, "loss": 0.0525, "step": 3370 }, { "action_loss": 0.01756373979151249, "epoch": 3.0305755395683454, "step": 3370 }, { "epoch": 3.0305755395683454, "step": 3370, "torque_loss": 0.15819860994815826 }, { "epoch": 3.039568345323741, "grad_norm": 0.27174460887908936, "learning_rate": 9.991552893637478e-05, "loss": 0.0424, "step": 3380 }, { "action_loss": 0.022966274991631508, "epoch": 3.039568345323741, "step": 3380 }, { "epoch": 3.039568345323741, "step": 3380, "torque_loss": 0.1728774756193161 }, { "epoch": 3.0485611510791366, "grad_norm": 0.41454774141311646, "learning_rate": 9.991359657952644e-05, "loss": 0.0463, "step": 3390 }, { "action_loss": 0.026225654408335686, "epoch": 3.0485611510791366, "step": 3390 }, { "epoch": 3.0485611510791366, "step": 3390, "torque_loss": 0.20603466033935547 }, { "epoch": 3.0575539568345325, "grad_norm": 0.2969230115413666, "learning_rate": 9.991164238877519e-05, "loss": 0.0441, "step": 3400 }, { "action_loss": 0.019959913566708565, "epoch": 3.0575539568345325, "step": 3400 }, { "epoch": 3.0575539568345325, "step": 3400, "torque_loss": 0.16015370190143585 }, { "epoch": 3.066546762589928, "grad_norm": 0.30241748690605164, "learning_rate": 9.990966636497585e-05, "loss": 0.0429, "step": 3410 }, { "action_loss": 0.025753362104296684, "epoch": 3.066546762589928, "step": 3410 }, { "epoch": 3.066546762589928, "step": 3410, "torque_loss": 0.2636500298976898 }, { "epoch": 3.0755395683453237, "grad_norm": 0.4391370713710785, "learning_rate": 9.99076685089928e-05, "loss": 0.0493, "step": 3420 }, { "action_loss": 0.013630550354719162, "epoch": 3.0755395683453237, "step": 3420 }, { "epoch": 3.0755395683453237, "step": 3420, "torque_loss": 0.11872907727956772 }, { "epoch": 3.0845323741007196, "grad_norm": 0.25410598516464233, "learning_rate": 9.990564882169998e-05, "loss": 0.0412, "step": 3430 }, { "action_loss": 0.042311668395996094, "epoch": 3.0845323741007196, "step": 3430 }, { "epoch": 3.0845323741007196, "step": 3430, "torque_loss": 0.27895215153694153 }, { "epoch": 3.093525179856115, "grad_norm": 0.504696786403656, "learning_rate": 9.990360730398088e-05, "loss": 0.0524, "step": 3440 }, { "action_loss": 0.026066089048981667, "epoch": 3.093525179856115, "step": 3440 }, { "epoch": 3.093525179856115, "step": 3440, "torque_loss": 0.1995856761932373 }, { "epoch": 3.102517985611511, "grad_norm": 0.27406635880470276, "learning_rate": 9.990154395672849e-05, "loss": 0.046, "step": 3450 }, { "action_loss": 0.033480580896139145, "epoch": 3.102517985611511, "step": 3450 }, { "epoch": 3.102517985611511, "step": 3450, "torque_loss": 0.258748859167099 }, { "epoch": 3.1115107913669067, "grad_norm": 0.2563491463661194, "learning_rate": 9.989945878084541e-05, "loss": 0.0449, "step": 3460 }, { "action_loss": 0.018306337296962738, "epoch": 3.1115107913669067, "step": 3460 }, { "epoch": 3.1115107913669067, "step": 3460, "torque_loss": 0.15698452293872833 }, { "epoch": 3.120503597122302, "grad_norm": 0.382094144821167, "learning_rate": 9.989735177724378e-05, "loss": 0.0457, "step": 3470 }, { "action_loss": 0.028970353305339813, "epoch": 3.120503597122302, "step": 3470 }, { "epoch": 3.120503597122302, "step": 3470, "torque_loss": 0.1712479144334793 }, { "epoch": 3.129496402877698, "grad_norm": 0.3397161364555359, "learning_rate": 9.989522294684526e-05, "loss": 0.0426, "step": 3480 }, { "action_loss": 0.042949020862579346, "epoch": 3.129496402877698, "step": 3480 }, { "epoch": 3.129496402877698, "step": 3480, "torque_loss": 0.23608164489269257 }, { "epoch": 3.1384892086330933, "grad_norm": 0.2763955593109131, "learning_rate": 9.989307229058107e-05, "loss": 0.0487, "step": 3490 }, { "action_loss": 0.019259966909885406, "epoch": 3.1384892086330933, "step": 3490 }, { "epoch": 3.1384892086330933, "step": 3490, "torque_loss": 0.16914553940296173 }, { "epoch": 3.147482014388489, "grad_norm": 0.3533303141593933, "learning_rate": 9.989089980939202e-05, "loss": 0.0408, "step": 3500 }, { "action_loss": 0.0341726690530777, "epoch": 3.147482014388489, "step": 3500 }, { "epoch": 3.147482014388489, "step": 3500, "torque_loss": 0.2427883893251419 }, { "epoch": 3.156474820143885, "grad_norm": 0.3304198086261749, "learning_rate": 9.988870550422835e-05, "loss": 0.0448, "step": 3510 }, { "action_loss": 0.016336947679519653, "epoch": 3.156474820143885, "step": 3510 }, { "epoch": 3.156474820143885, "step": 3510, "torque_loss": 0.23019611835479736 }, { "epoch": 3.1654676258992804, "grad_norm": 0.31918495893478394, "learning_rate": 9.988648937604999e-05, "loss": 0.0493, "step": 3520 }, { "action_loss": 0.02203187346458435, "epoch": 3.1654676258992804, "step": 3520 }, { "epoch": 3.1654676258992804, "step": 3520, "torque_loss": 0.24984531104564667 }, { "epoch": 3.1744604316546763, "grad_norm": 0.3204285502433777, "learning_rate": 9.988425142582632e-05, "loss": 0.0416, "step": 3530 }, { "action_loss": 0.031440746039152145, "epoch": 3.1744604316546763, "step": 3530 }, { "epoch": 3.1744604316546763, "step": 3530, "torque_loss": 0.20155327022075653 }, { "epoch": 3.183453237410072, "grad_norm": 0.31155112385749817, "learning_rate": 9.98819916545363e-05, "loss": 0.0531, "step": 3540 }, { "action_loss": 0.018734412267804146, "epoch": 3.183453237410072, "step": 3540 }, { "epoch": 3.183453237410072, "step": 3540, "torque_loss": 0.09512487798929214 }, { "epoch": 3.1924460431654675, "grad_norm": 0.31766363978385925, "learning_rate": 9.987971006316844e-05, "loss": 0.0425, "step": 3550 }, { "action_loss": 0.042372893542051315, "epoch": 3.1924460431654675, "step": 3550 }, { "epoch": 3.1924460431654675, "step": 3550, "torque_loss": 0.29289165139198303 }, { "epoch": 3.2014388489208634, "grad_norm": 0.31171464920043945, "learning_rate": 9.987740665272077e-05, "loss": 0.0447, "step": 3560 }, { "action_loss": 0.023512965068221092, "epoch": 3.2014388489208634, "step": 3560 }, { "epoch": 3.2014388489208634, "step": 3560, "torque_loss": 0.20632213354110718 }, { "epoch": 3.210431654676259, "grad_norm": 0.32148101925849915, "learning_rate": 9.98750814242009e-05, "loss": 0.0483, "step": 3570 }, { "action_loss": 0.03816453740000725, "epoch": 3.210431654676259, "step": 3570 }, { "epoch": 3.210431654676259, "step": 3570, "torque_loss": 0.31433022022247314 }, { "epoch": 3.2194244604316546, "grad_norm": 0.3312123715877533, "learning_rate": 9.987273437862594e-05, "loss": 0.0469, "step": 3580 }, { "action_loss": 0.014336742460727692, "epoch": 3.2194244604316546, "step": 3580 }, { "epoch": 3.2194244604316546, "step": 3580, "torque_loss": 0.1597456932067871 }, { "epoch": 3.2284172661870505, "grad_norm": 0.282716304063797, "learning_rate": 9.987036551702259e-05, "loss": 0.0424, "step": 3590 }, { "action_loss": 0.03010079450905323, "epoch": 3.2284172661870505, "step": 3590 }, { "epoch": 3.2284172661870505, "step": 3590, "torque_loss": 0.3509942591190338 }, { "epoch": 3.237410071942446, "grad_norm": 0.20494721829891205, "learning_rate": 9.986797484042706e-05, "loss": 0.0431, "step": 3600 }, { "action_loss": 0.013635491020977497, "epoch": 3.237410071942446, "step": 3600 }, { "epoch": 3.237410071942446, "step": 3600, "torque_loss": 0.1655367612838745 }, { "epoch": 3.2464028776978417, "grad_norm": 0.3820765018463135, "learning_rate": 9.986556234988512e-05, "loss": 0.0451, "step": 3610 }, { "action_loss": 0.014547268860042095, "epoch": 3.2464028776978417, "step": 3610 }, { "epoch": 3.2464028776978417, "step": 3610, "torque_loss": 0.2170739769935608 }, { "epoch": 3.2553956834532376, "grad_norm": 0.33779478073120117, "learning_rate": 9.986312804645205e-05, "loss": 0.0437, "step": 3620 }, { "action_loss": 0.01773812435567379, "epoch": 3.2553956834532376, "step": 3620 }, { "epoch": 3.2553956834532376, "step": 3620, "torque_loss": 0.1831478327512741 }, { "epoch": 3.264388489208633, "grad_norm": 0.3183571398258209, "learning_rate": 9.986067193119273e-05, "loss": 0.0463, "step": 3630 }, { "action_loss": 0.009773386642336845, "epoch": 3.264388489208633, "step": 3630 }, { "epoch": 3.264388489208633, "step": 3630, "torque_loss": 0.1374804526567459 }, { "epoch": 3.273381294964029, "grad_norm": 0.362811416387558, "learning_rate": 9.985819400518153e-05, "loss": 0.0342, "step": 3640 }, { "action_loss": 0.03478085622191429, "epoch": 3.273381294964029, "step": 3640 }, { "epoch": 3.273381294964029, "step": 3640, "torque_loss": 0.2138361930847168 }, { "epoch": 3.2823741007194247, "grad_norm": 0.20672979950904846, "learning_rate": 9.985569426950239e-05, "loss": 0.0437, "step": 3650 }, { "action_loss": 0.018741874024271965, "epoch": 3.2823741007194247, "step": 3650 }, { "epoch": 3.2823741007194247, "step": 3650, "torque_loss": 0.23941026628017426 }, { "epoch": 3.29136690647482, "grad_norm": 0.21847102046012878, "learning_rate": 9.985317272524876e-05, "loss": 0.0386, "step": 3660 }, { "action_loss": 0.014474342577159405, "epoch": 3.29136690647482, "step": 3660 }, { "epoch": 3.29136690647482, "step": 3660, "torque_loss": 0.21516089141368866 }, { "epoch": 3.300359712230216, "grad_norm": 0.3304024934768677, "learning_rate": 9.985062937352366e-05, "loss": 0.0425, "step": 3670 }, { "action_loss": 0.015983525663614273, "epoch": 3.300359712230216, "step": 3670 }, { "epoch": 3.300359712230216, "step": 3670, "torque_loss": 0.2722664773464203 }, { "epoch": 3.3093525179856114, "grad_norm": 0.3130881190299988, "learning_rate": 9.984806421543966e-05, "loss": 0.0378, "step": 3680 }, { "action_loss": 0.011246803216636181, "epoch": 3.3093525179856114, "step": 3680 }, { "epoch": 3.3093525179856114, "step": 3680, "torque_loss": 0.162871316075325 }, { "epoch": 3.318345323741007, "grad_norm": 0.3116949200630188, "learning_rate": 9.984547725211881e-05, "loss": 0.0397, "step": 3690 }, { "action_loss": 0.03490021824836731, "epoch": 3.318345323741007, "step": 3690 }, { "epoch": 3.318345323741007, "step": 3690, "torque_loss": 0.25693216919898987 }, { "epoch": 3.327338129496403, "grad_norm": 0.40892401337623596, "learning_rate": 9.984286848469276e-05, "loss": 0.0448, "step": 3700 }, { "action_loss": 0.02217724919319153, "epoch": 3.327338129496403, "step": 3700 }, { "epoch": 3.327338129496403, "step": 3700, "torque_loss": 0.13681383430957794 }, { "epoch": 3.3363309352517985, "grad_norm": 0.341425359249115, "learning_rate": 9.984023791430266e-05, "loss": 0.043, "step": 3710 }, { "action_loss": 0.029460808262228966, "epoch": 3.3363309352517985, "step": 3710 }, { "epoch": 3.3363309352517985, "step": 3710, "torque_loss": 0.29795631766319275 }, { "epoch": 3.3453237410071943, "grad_norm": 0.38855278491973877, "learning_rate": 9.983758554209924e-05, "loss": 0.0463, "step": 3720 }, { "action_loss": 0.025882313027977943, "epoch": 3.3453237410071943, "step": 3720 }, { "epoch": 3.3453237410071943, "step": 3720, "torque_loss": 0.22847847640514374 }, { "epoch": 3.3543165467625897, "grad_norm": 0.4016030430793762, "learning_rate": 9.983491136924268e-05, "loss": 0.0408, "step": 3730 }, { "action_loss": 0.03268524631857872, "epoch": 3.3543165467625897, "step": 3730 }, { "epoch": 3.3543165467625897, "step": 3730, "torque_loss": 0.1943405419588089 }, { "epoch": 3.3633093525179856, "grad_norm": 0.24671299755573273, "learning_rate": 9.983221539690282e-05, "loss": 0.0459, "step": 3740 }, { "action_loss": 0.02497261017560959, "epoch": 3.3633093525179856, "step": 3740 }, { "epoch": 3.3633093525179856, "step": 3740, "torque_loss": 0.2688029110431671 }, { "epoch": 3.3723021582733814, "grad_norm": 0.26968732476234436, "learning_rate": 9.982949762625892e-05, "loss": 0.0403, "step": 3750 }, { "action_loss": 0.014022731222212315, "epoch": 3.3723021582733814, "step": 3750 }, { "epoch": 3.3723021582733814, "step": 3750, "torque_loss": 0.12526936829090118 }, { "epoch": 3.381294964028777, "grad_norm": 0.25059086084365845, "learning_rate": 9.982675805849986e-05, "loss": 0.0325, "step": 3760 }, { "action_loss": 0.015992676839232445, "epoch": 3.381294964028777, "step": 3760 }, { "epoch": 3.381294964028777, "step": 3760, "torque_loss": 0.21207833290100098 }, { "epoch": 3.3902877697841727, "grad_norm": 0.24160955846309662, "learning_rate": 9.982399669482399e-05, "loss": 0.0359, "step": 3770 }, { "action_loss": 0.007107967045158148, "epoch": 3.3902877697841727, "step": 3770 }, { "epoch": 3.3902877697841727, "step": 3770, "torque_loss": 0.19582705199718475 }, { "epoch": 3.3992805755395685, "grad_norm": 0.18845851719379425, "learning_rate": 9.982121353643924e-05, "loss": 0.037, "step": 3780 }, { "action_loss": 0.050273824483156204, "epoch": 3.3992805755395685, "step": 3780 }, { "epoch": 3.3992805755395685, "step": 3780, "torque_loss": 0.24071545898914337 }, { "epoch": 3.408273381294964, "grad_norm": 0.3366371989250183, "learning_rate": 9.981840858456306e-05, "loss": 0.0469, "step": 3790 }, { "action_loss": 0.017290037125349045, "epoch": 3.408273381294964, "step": 3790 }, { "epoch": 3.408273381294964, "step": 3790, "torque_loss": 0.2787882387638092 }, { "epoch": 3.41726618705036, "grad_norm": 0.2709856629371643, "learning_rate": 9.981558184042243e-05, "loss": 0.0482, "step": 3800 }, { "action_loss": 0.023479975759983063, "epoch": 3.41726618705036, "step": 3800 }, { "epoch": 3.41726618705036, "step": 3800, "torque_loss": 0.2531757652759552 }, { "epoch": 3.4262589928057556, "grad_norm": 0.22275830805301666, "learning_rate": 9.981273330525387e-05, "loss": 0.0447, "step": 3810 }, { "action_loss": 0.03384237736463547, "epoch": 3.4262589928057556, "step": 3810 }, { "epoch": 3.4262589928057556, "step": 3810, "torque_loss": 0.2988356351852417 }, { "epoch": 3.435251798561151, "grad_norm": 0.3311561346054077, "learning_rate": 9.980986298030341e-05, "loss": 0.0439, "step": 3820 }, { "action_loss": 0.012458858080208302, "epoch": 3.435251798561151, "step": 3820 }, { "epoch": 3.435251798561151, "step": 3820, "torque_loss": 0.23402458429336548 }, { "epoch": 3.444244604316547, "grad_norm": 0.3056488037109375, "learning_rate": 9.980697086682662e-05, "loss": 0.0394, "step": 3830 }, { "action_loss": 0.03149498626589775, "epoch": 3.444244604316547, "step": 3830 }, { "epoch": 3.444244604316547, "step": 3830, "torque_loss": 0.3125287890434265 }, { "epoch": 3.4532374100719423, "grad_norm": 0.3795441687107086, "learning_rate": 9.980405696608866e-05, "loss": 0.0466, "step": 3840 }, { "action_loss": 0.03863132372498512, "epoch": 3.4532374100719423, "step": 3840 }, { "epoch": 3.4532374100719423, "step": 3840, "torque_loss": 0.2774350941181183 }, { "epoch": 3.462230215827338, "grad_norm": 0.36491337418556213, "learning_rate": 9.980112127936409e-05, "loss": 0.0402, "step": 3850 }, { "action_loss": 0.017395062372088432, "epoch": 3.462230215827338, "step": 3850 }, { "epoch": 3.462230215827338, "step": 3850, "torque_loss": 0.17966721951961517 }, { "epoch": 3.471223021582734, "grad_norm": 0.3790920376777649, "learning_rate": 9.979816380793717e-05, "loss": 0.0495, "step": 3860 }, { "action_loss": 0.03626688942313194, "epoch": 3.471223021582734, "step": 3860 }, { "epoch": 3.471223021582734, "step": 3860, "torque_loss": 0.24893446266651154 }, { "epoch": 3.4802158273381294, "grad_norm": 0.2964603900909424, "learning_rate": 9.979518455310151e-05, "loss": 0.0469, "step": 3870 }, { "action_loss": 0.01921301707625389, "epoch": 3.4802158273381294, "step": 3870 }, { "epoch": 3.4802158273381294, "step": 3870, "torque_loss": 0.1656799018383026 }, { "epoch": 3.4892086330935252, "grad_norm": 0.20576784014701843, "learning_rate": 9.97921835161604e-05, "loss": 0.0426, "step": 3880 }, { "action_loss": 0.02463994175195694, "epoch": 3.4892086330935252, "step": 3880 }, { "epoch": 3.4892086330935252, "step": 3880, "torque_loss": 0.20447230339050293 }, { "epoch": 3.4982014388489207, "grad_norm": 0.3219883441925049, "learning_rate": 9.978916069842656e-05, "loss": 0.0469, "step": 3890 }, { "action_loss": 0.024022618308663368, "epoch": 3.4982014388489207, "step": 3890 }, { "epoch": 3.4982014388489207, "step": 3890, "torque_loss": 0.23691006004810333 }, { "epoch": 3.5071942446043165, "grad_norm": 0.3491942286491394, "learning_rate": 9.97861161012223e-05, "loss": 0.043, "step": 3900 }, { "action_loss": 0.0229631457477808, "epoch": 3.5071942446043165, "step": 3900 }, { "epoch": 3.5071942446043165, "step": 3900, "torque_loss": 0.2682991921901703 }, { "epoch": 3.5161870503597124, "grad_norm": 0.2466835230588913, "learning_rate": 9.978304972587942e-05, "loss": 0.0383, "step": 3910 }, { "action_loss": 0.01629915088415146, "epoch": 3.5161870503597124, "step": 3910 }, { "epoch": 3.5161870503597124, "step": 3910, "torque_loss": 0.14422748982906342 }, { "epoch": 3.5251798561151078, "grad_norm": 0.28291934728622437, "learning_rate": 9.977996157373925e-05, "loss": 0.0368, "step": 3920 }, { "action_loss": 0.017947791144251823, "epoch": 3.5251798561151078, "step": 3920 }, { "epoch": 3.5251798561151078, "step": 3920, "torque_loss": 0.2248321771621704 }, { "epoch": 3.5341726618705036, "grad_norm": 0.33053886890411377, "learning_rate": 9.977685164615265e-05, "loss": 0.0424, "step": 3930 }, { "action_loss": 0.031047657132148743, "epoch": 3.5341726618705036, "step": 3930 }, { "epoch": 3.5341726618705036, "step": 3930, "torque_loss": 0.2453310489654541 }, { "epoch": 3.543165467625899, "grad_norm": 0.30067920684814453, "learning_rate": 9.977371994448002e-05, "loss": 0.0482, "step": 3940 }, { "action_loss": 0.01401116419583559, "epoch": 3.543165467625899, "step": 3940 }, { "epoch": 3.543165467625899, "step": 3940, "torque_loss": 0.214467391371727 }, { "epoch": 3.552158273381295, "grad_norm": 0.2870919704437256, "learning_rate": 9.977056647009127e-05, "loss": 0.0423, "step": 3950 }, { "action_loss": 0.03255582973361015, "epoch": 3.552158273381295, "step": 3950 }, { "epoch": 3.552158273381295, "step": 3950, "torque_loss": 0.2576692998409271 }, { "epoch": 3.5611510791366907, "grad_norm": 0.24000518023967743, "learning_rate": 9.976739122436582e-05, "loss": 0.0491, "step": 3960 }, { "action_loss": 0.01376141607761383, "epoch": 3.5611510791366907, "step": 3960 }, { "epoch": 3.5611510791366907, "step": 3960, "torque_loss": 0.2342214733362198 }, { "epoch": 3.5701438848920866, "grad_norm": 0.28710469603538513, "learning_rate": 9.976419420869265e-05, "loss": 0.0419, "step": 3970 }, { "action_loss": 0.03810037672519684, "epoch": 3.5701438848920866, "step": 3970 }, { "epoch": 3.5701438848920866, "step": 3970, "torque_loss": 0.21552295982837677 }, { "epoch": 3.579136690647482, "grad_norm": 0.33382025361061096, "learning_rate": 9.976097542447025e-05, "loss": 0.0378, "step": 3980 }, { "action_loss": 0.016968248412013054, "epoch": 3.579136690647482, "step": 3980 }, { "epoch": 3.579136690647482, "step": 3980, "torque_loss": 0.20608234405517578 }, { "epoch": 3.588129496402878, "grad_norm": 0.36003378033638, "learning_rate": 9.97577348731066e-05, "loss": 0.0378, "step": 3990 }, { "action_loss": 0.016921965405344963, "epoch": 3.588129496402878, "step": 3990 }, { "epoch": 3.588129496402878, "step": 3990, "torque_loss": 0.19685232639312744 }, { "epoch": 3.597122302158273, "grad_norm": 0.1994386613368988, "learning_rate": 9.975447255601927e-05, "loss": 0.0362, "step": 4000 }, { "action_loss": 0.018016261979937553, "epoch": 3.597122302158273, "step": 4000 }, { "epoch": 3.597122302158273, "step": 4000, "torque_loss": 0.14589910209178925 }, { "epoch": 3.606115107913669, "grad_norm": 0.27307766675949097, "learning_rate": 9.975118847463525e-05, "loss": 0.0377, "step": 4010 }, { "action_loss": 0.009845889173448086, "epoch": 3.606115107913669, "step": 4010 }, { "epoch": 3.606115107913669, "step": 4010, "torque_loss": 0.2073746919631958 }, { "epoch": 3.615107913669065, "grad_norm": 0.2994510233402252, "learning_rate": 9.974788263039114e-05, "loss": 0.0374, "step": 4020 }, { "action_loss": 0.025424256920814514, "epoch": 3.615107913669065, "step": 4020 }, { "epoch": 3.615107913669065, "step": 4020, "torque_loss": 0.2239735722541809 }, { "epoch": 3.6241007194244603, "grad_norm": 0.24163590371608734, "learning_rate": 9.974455502473303e-05, "loss": 0.0379, "step": 4030 }, { "action_loss": 0.016961175948381424, "epoch": 3.6241007194244603, "step": 4030 }, { "epoch": 3.6241007194244603, "step": 4030, "torque_loss": 0.18585102260112762 }, { "epoch": 3.633093525179856, "grad_norm": 0.3485456109046936, "learning_rate": 9.974120565911652e-05, "loss": 0.0434, "step": 4040 }, { "action_loss": 0.021618470549583435, "epoch": 3.633093525179856, "step": 4040 }, { "epoch": 3.633093525179856, "step": 4040, "torque_loss": 0.2638622522354126 }, { "epoch": 3.6420863309352516, "grad_norm": 0.40655243396759033, "learning_rate": 9.973783453500674e-05, "loss": 0.0464, "step": 4050 }, { "action_loss": 0.033294904977083206, "epoch": 3.6420863309352516, "step": 4050 }, { "epoch": 3.6420863309352516, "step": 4050, "torque_loss": 0.2239970713853836 }, { "epoch": 3.6510791366906474, "grad_norm": 0.3531898558139801, "learning_rate": 9.973444165387835e-05, "loss": 0.038, "step": 4060 }, { "action_loss": 0.017696179449558258, "epoch": 3.6510791366906474, "step": 4060 }, { "epoch": 3.6510791366906474, "step": 4060, "torque_loss": 0.23683805763721466 }, { "epoch": 3.6600719424460433, "grad_norm": 0.33781006932258606, "learning_rate": 9.973102701721549e-05, "loss": 0.053, "step": 4070 }, { "action_loss": 0.013308488763868809, "epoch": 3.6600719424460433, "step": 4070 }, { "epoch": 3.6600719424460433, "step": 4070, "torque_loss": 0.19608312845230103 }, { "epoch": 3.6690647482014387, "grad_norm": 0.41695186495780945, "learning_rate": 9.972759062651184e-05, "loss": 0.0451, "step": 4080 }, { "action_loss": 0.02931540459394455, "epoch": 3.6690647482014387, "step": 4080 }, { "epoch": 3.6690647482014387, "step": 4080, "torque_loss": 0.24683032929897308 }, { "epoch": 3.6780575539568345, "grad_norm": 0.25989270210266113, "learning_rate": 9.972413248327059e-05, "loss": 0.0411, "step": 4090 }, { "action_loss": 0.017929142341017723, "epoch": 3.6780575539568345, "step": 4090 }, { "epoch": 3.6780575539568345, "step": 4090, "torque_loss": 0.22172850370407104 }, { "epoch": 3.68705035971223, "grad_norm": 0.34006446599960327, "learning_rate": 9.972065258900447e-05, "loss": 0.0399, "step": 4100 }, { "action_loss": 0.011682406067848206, "epoch": 3.68705035971223, "step": 4100 }, { "epoch": 3.68705035971223, "step": 4100, "torque_loss": 0.11936268955469131 }, { "epoch": 3.696043165467626, "grad_norm": 0.3329427242279053, "learning_rate": 9.971715094523569e-05, "loss": 0.0353, "step": 4110 }, { "action_loss": 0.037147846072912216, "epoch": 3.696043165467626, "step": 4110 }, { "epoch": 3.696043165467626, "step": 4110, "torque_loss": 0.303457647562027 }, { "epoch": 3.7050359712230216, "grad_norm": 0.22225601971149445, "learning_rate": 9.971362755349598e-05, "loss": 0.0401, "step": 4120 }, { "action_loss": 0.018345266580581665, "epoch": 3.7050359712230216, "step": 4120 }, { "epoch": 3.7050359712230216, "step": 4120, "torque_loss": 0.23288697004318237 }, { "epoch": 3.7140287769784175, "grad_norm": 0.28345945477485657, "learning_rate": 9.971008241532662e-05, "loss": 0.0356, "step": 4130 }, { "action_loss": 0.018045639619231224, "epoch": 3.7140287769784175, "step": 4130 }, { "epoch": 3.7140287769784175, "step": 4130, "torque_loss": 0.1875085085630417 }, { "epoch": 3.723021582733813, "grad_norm": 0.31866249442100525, "learning_rate": 9.970651553227835e-05, "loss": 0.0415, "step": 4140 }, { "action_loss": 0.010034888051450253, "epoch": 3.723021582733813, "step": 4140 }, { "epoch": 3.723021582733813, "step": 4140, "torque_loss": 0.18685103952884674 }, { "epoch": 3.7320143884892087, "grad_norm": 0.3265850245952606, "learning_rate": 9.970292690591143e-05, "loss": 0.0387, "step": 4150 }, { "action_loss": 0.018624475225806236, "epoch": 3.7320143884892087, "step": 4150 }, { "epoch": 3.7320143884892087, "step": 4150, "torque_loss": 0.1676868051290512 }, { "epoch": 3.741007194244604, "grad_norm": 0.3032676875591278, "learning_rate": 9.969931653779569e-05, "loss": 0.0317, "step": 4160 }, { "action_loss": 0.015900155529379845, "epoch": 3.741007194244604, "step": 4160 }, { "epoch": 3.741007194244604, "step": 4160, "torque_loss": 0.1970764398574829 }, { "epoch": 3.75, "grad_norm": 0.16600197553634644, "learning_rate": 9.969568442951038e-05, "loss": 0.0329, "step": 4170 }, { "action_loss": 0.03165605664253235, "epoch": 3.75, "step": 4170 }, { "epoch": 3.75, "step": 4170, "torque_loss": 0.28911876678466797 }, { "epoch": 3.758992805755396, "grad_norm": 0.2434646189212799, "learning_rate": 9.969203058264436e-05, "loss": 0.0369, "step": 4180 }, { "action_loss": 0.010389798320829868, "epoch": 3.758992805755396, "step": 4180 }, { "epoch": 3.758992805755396, "step": 4180, "torque_loss": 0.15674376487731934 }, { "epoch": 3.7679856115107913, "grad_norm": 0.24791480600833893, "learning_rate": 9.96883549987959e-05, "loss": 0.0341, "step": 4190 }, { "action_loss": 0.009631063789129257, "epoch": 3.7679856115107913, "step": 4190 }, { "epoch": 3.7679856115107913, "step": 4190, "torque_loss": 0.1704435795545578 }, { "epoch": 3.776978417266187, "grad_norm": 0.2464461326599121, "learning_rate": 9.968465767957287e-05, "loss": 0.0329, "step": 4200 }, { "action_loss": 0.02811642549932003, "epoch": 3.776978417266187, "step": 4200 }, { "epoch": 3.776978417266187, "step": 4200, "torque_loss": 0.2642116844654083 }, { "epoch": 3.7859712230215825, "grad_norm": 0.26002582907676697, "learning_rate": 9.968093862659256e-05, "loss": 0.0365, "step": 4210 }, { "action_loss": 0.013177088461816311, "epoch": 3.7859712230215825, "step": 4210 }, { "epoch": 3.7859712230215825, "step": 4210, "torque_loss": 0.24155907332897186 }, { "epoch": 3.7949640287769784, "grad_norm": 0.21213382482528687, "learning_rate": 9.967719784148182e-05, "loss": 0.0394, "step": 4220 }, { "action_loss": 0.009521100670099258, "epoch": 3.7949640287769784, "step": 4220 }, { "epoch": 3.7949640287769784, "step": 4220, "torque_loss": 0.2157374769449234 }, { "epoch": 3.803956834532374, "grad_norm": 0.20634837448596954, "learning_rate": 9.967343532587702e-05, "loss": 0.0389, "step": 4230 }, { "action_loss": 0.008929560892283916, "epoch": 3.803956834532374, "step": 4230 }, { "epoch": 3.803956834532374, "step": 4230, "torque_loss": 0.16287462413311005 }, { "epoch": 3.81294964028777, "grad_norm": 0.3022758662700653, "learning_rate": 9.966965108142399e-05, "loss": 0.0373, "step": 4240 }, { "action_loss": 0.012736023403704166, "epoch": 3.81294964028777, "step": 4240 }, { "epoch": 3.81294964028777, "step": 4240, "torque_loss": 0.1861727237701416 }, { "epoch": 3.8219424460431655, "grad_norm": 0.4839564859867096, "learning_rate": 9.96658451097781e-05, "loss": 0.0488, "step": 4250 }, { "action_loss": 0.017311038449406624, "epoch": 3.8219424460431655, "step": 4250 }, { "epoch": 3.8219424460431655, "step": 4250, "torque_loss": 0.21867422759532928 }, { "epoch": 3.8309352517985613, "grad_norm": 0.3479961156845093, "learning_rate": 9.966201741260419e-05, "loss": 0.0387, "step": 4260 }, { "action_loss": 0.016204191371798515, "epoch": 3.8309352517985613, "step": 4260 }, { "epoch": 3.8309352517985613, "step": 4260, "torque_loss": 0.18519830703735352 }, { "epoch": 3.8399280575539567, "grad_norm": 0.3214625120162964, "learning_rate": 9.965816799157665e-05, "loss": 0.0435, "step": 4270 }, { "action_loss": 0.0068104988895356655, "epoch": 3.8399280575539567, "step": 4270 }, { "epoch": 3.8399280575539567, "step": 4270, "torque_loss": 0.16172698140144348 }, { "epoch": 3.8489208633093526, "grad_norm": 0.3448222875595093, "learning_rate": 9.965429684837935e-05, "loss": 0.0407, "step": 4280 }, { "action_loss": 0.0479632206261158, "epoch": 3.8489208633093526, "step": 4280 }, { "epoch": 3.8489208633093526, "step": 4280, "torque_loss": 0.34233418107032776 }, { "epoch": 3.8579136690647484, "grad_norm": 0.28621551394462585, "learning_rate": 9.965040398470562e-05, "loss": 0.0557, "step": 4290 }, { "action_loss": 0.01097599882632494, "epoch": 3.8579136690647484, "step": 4290 }, { "epoch": 3.8579136690647484, "step": 4290, "torque_loss": 0.18648040294647217 }, { "epoch": 3.866906474820144, "grad_norm": 0.24705705046653748, "learning_rate": 9.964648940225838e-05, "loss": 0.037, "step": 4300 }, { "action_loss": 0.017919253557920456, "epoch": 3.866906474820144, "step": 4300 }, { "epoch": 3.866906474820144, "step": 4300, "torque_loss": 0.1789391040802002 }, { "epoch": 3.8758992805755397, "grad_norm": 0.4225060045719147, "learning_rate": 9.964255310274997e-05, "loss": 0.0411, "step": 4310 }, { "action_loss": 0.008304022252559662, "epoch": 3.8758992805755397, "step": 4310 }, { "epoch": 3.8758992805755397, "step": 4310, "torque_loss": 0.1467752903699875 }, { "epoch": 3.884892086330935, "grad_norm": 0.4015888273715973, "learning_rate": 9.963859508790228e-05, "loss": 0.042, "step": 4320 }, { "action_loss": 0.01748173125088215, "epoch": 3.884892086330935, "step": 4320 }, { "epoch": 3.884892086330935, "step": 4320, "torque_loss": 0.20101027190685272 }, { "epoch": 3.893884892086331, "grad_norm": 0.3430492877960205, "learning_rate": 9.963461535944664e-05, "loss": 0.0397, "step": 4330 }, { "action_loss": 0.010893096216022968, "epoch": 3.893884892086331, "step": 4330 }, { "epoch": 3.893884892086331, "step": 4330, "torque_loss": 0.1456999033689499 }, { "epoch": 3.902877697841727, "grad_norm": 0.2527467608451843, "learning_rate": 9.963061391912399e-05, "loss": 0.0335, "step": 4340 }, { "action_loss": 0.02575250156223774, "epoch": 3.902877697841727, "step": 4340 }, { "epoch": 3.902877697841727, "step": 4340, "torque_loss": 0.16987954080104828 }, { "epoch": 3.911870503597122, "grad_norm": 0.3293278217315674, "learning_rate": 9.962659076868463e-05, "loss": 0.0358, "step": 4350 }, { "action_loss": 0.030439933761954308, "epoch": 3.911870503597122, "step": 4350 }, { "epoch": 3.911870503597122, "step": 4350, "torque_loss": 0.24096651375293732 }, { "epoch": 3.920863309352518, "grad_norm": 0.2990078628063202, "learning_rate": 9.962254590988846e-05, "loss": 0.051, "step": 4360 }, { "action_loss": 0.011168821714818478, "epoch": 3.920863309352518, "step": 4360 }, { "epoch": 3.920863309352518, "step": 4360, "torque_loss": 0.10494522005319595 }, { "epoch": 3.9298561151079134, "grad_norm": 0.29789650440216064, "learning_rate": 9.961847934450481e-05, "loss": 0.0341, "step": 4370 }, { "action_loss": 0.014692570082843304, "epoch": 3.9298561151079134, "step": 4370 }, { "epoch": 3.9298561151079134, "step": 4370, "torque_loss": 0.1903427243232727 }, { "epoch": 3.9388489208633093, "grad_norm": 0.3040759563446045, "learning_rate": 9.961439107431257e-05, "loss": 0.0407, "step": 4380 }, { "action_loss": 0.020068995654582977, "epoch": 3.9388489208633093, "step": 4380 }, { "epoch": 3.9388489208633093, "step": 4380, "torque_loss": 0.19864214956760406 }, { "epoch": 3.947841726618705, "grad_norm": 0.30093908309936523, "learning_rate": 9.961028110110006e-05, "loss": 0.0447, "step": 4390 }, { "action_loss": 0.011417507193982601, "epoch": 3.947841726618705, "step": 4390 }, { "epoch": 3.947841726618705, "step": 4390, "torque_loss": 0.1715865135192871 }, { "epoch": 3.956834532374101, "grad_norm": 0.4350472390651703, "learning_rate": 9.960614942666513e-05, "loss": 0.0333, "step": 4400 }, { "action_loss": 0.010899814777076244, "epoch": 3.956834532374101, "step": 4400 }, { "epoch": 3.956834532374101, "step": 4400, "torque_loss": 0.11360076814889908 }, { "epoch": 3.9658273381294964, "grad_norm": 0.33328384160995483, "learning_rate": 9.960199605281511e-05, "loss": 0.0372, "step": 4410 }, { "action_loss": 0.010983164422214031, "epoch": 3.9658273381294964, "step": 4410 }, { "epoch": 3.9658273381294964, "step": 4410, "torque_loss": 0.15459395945072174 }, { "epoch": 3.9748201438848922, "grad_norm": 0.25188544392585754, "learning_rate": 9.959782098136683e-05, "loss": 0.0364, "step": 4420 }, { "action_loss": 0.010903921909630299, "epoch": 3.9748201438848922, "step": 4420 }, { "epoch": 3.9748201438848922, "step": 4420, "torque_loss": 0.13019026815891266 }, { "epoch": 3.9838129496402876, "grad_norm": 0.23514485359191895, "learning_rate": 9.959362421414662e-05, "loss": 0.0333, "step": 4430 }, { "action_loss": 0.010918094776570797, "epoch": 3.9838129496402876, "step": 4430 }, { "epoch": 3.9838129496402876, "step": 4430, "torque_loss": 0.15667523443698883 }, { "epoch": 3.9928057553956835, "grad_norm": 0.2512267231941223, "learning_rate": 9.958940575299027e-05, "loss": 0.0387, "step": 4440 }, { "action_loss": 0.031005507335066795, "epoch": 3.9928057553956835, "step": 4440 }, { "epoch": 3.9928057553956835, "step": 4440, "torque_loss": 0.2892264425754547 }, { "epoch": 4.001798561151079, "grad_norm": 0.23990771174430847, "learning_rate": 9.95851655997431e-05, "loss": 0.0384, "step": 4450 }, { "action_loss": 0.021889247000217438, "epoch": 4.001798561151079, "step": 4450 }, { "epoch": 4.001798561151079, "step": 4450, "torque_loss": 0.2180912047624588 }, { "epoch": 4.010791366906475, "grad_norm": 0.2605549991130829, "learning_rate": 9.958090375625986e-05, "loss": 0.0416, "step": 4460 }, { "action_loss": 0.016613038256764412, "epoch": 4.010791366906475, "step": 4460 }, { "epoch": 4.010791366906475, "step": 4460, "torque_loss": 0.14605370163917542 }, { "epoch": 4.01978417266187, "grad_norm": 0.2168746441602707, "learning_rate": 9.957662022440486e-05, "loss": 0.0375, "step": 4470 }, { "action_loss": 0.022232340648770332, "epoch": 4.01978417266187, "step": 4470 }, { "epoch": 4.01978417266187, "step": 4470, "torque_loss": 0.21113240718841553 }, { "epoch": 4.028776978417266, "grad_norm": 0.17255952954292297, "learning_rate": 9.957231500605187e-05, "loss": 0.0327, "step": 4480 }, { "action_loss": 0.017909852787852287, "epoch": 4.028776978417266, "step": 4480 }, { "epoch": 4.028776978417266, "step": 4480, "torque_loss": 0.21224112808704376 }, { "epoch": 4.037769784172662, "grad_norm": 0.29133710265159607, "learning_rate": 9.95679881030841e-05, "loss": 0.0354, "step": 4490 }, { "action_loss": 0.03406090661883354, "epoch": 4.037769784172662, "step": 4490 }, { "epoch": 4.037769784172662, "step": 4490, "torque_loss": 0.19998252391815186 }, { "epoch": 4.046762589928058, "grad_norm": 0.18180306255817413, "learning_rate": 9.95636395173943e-05, "loss": 0.0443, "step": 4500 }, { "action_loss": 0.007999495603144169, "epoch": 4.046762589928058, "step": 4500 }, { "epoch": 4.046762589928058, "step": 4500, "torque_loss": 0.13817943632602692 }, { "epoch": 4.055755395683454, "grad_norm": 0.24657081067562103, "learning_rate": 9.95592692508847e-05, "loss": 0.0354, "step": 4510 }, { "action_loss": 0.05218636617064476, "epoch": 4.055755395683454, "step": 4510 }, { "epoch": 4.055755395683454, "step": 4510, "torque_loss": 0.26266059279441833 }, { "epoch": 4.0647482014388485, "grad_norm": 0.31246593594551086, "learning_rate": 9.9554877305467e-05, "loss": 0.0461, "step": 4520 }, { "action_loss": 0.02157926745712757, "epoch": 4.0647482014388485, "step": 4520 }, { "epoch": 4.0647482014388485, "step": 4520, "torque_loss": 0.19913919270038605 }, { "epoch": 4.073741007194244, "grad_norm": 0.28465744853019714, "learning_rate": 9.955046368306237e-05, "loss": 0.0372, "step": 4530 }, { "action_loss": 0.009676975198090076, "epoch": 4.073741007194244, "step": 4530 }, { "epoch": 4.073741007194244, "step": 4530, "torque_loss": 0.11468150466680527 }, { "epoch": 4.08273381294964, "grad_norm": 0.2803294360637665, "learning_rate": 9.954602838560153e-05, "loss": 0.0359, "step": 4540 }, { "action_loss": 0.021700626239180565, "epoch": 4.08273381294964, "step": 4540 }, { "epoch": 4.08273381294964, "step": 4540, "torque_loss": 0.24649162590503693 }, { "epoch": 4.091726618705036, "grad_norm": 0.30721941590309143, "learning_rate": 9.954157141502456e-05, "loss": 0.0411, "step": 4550 }, { "action_loss": 0.023979047313332558, "epoch": 4.091726618705036, "step": 4550 }, { "epoch": 4.091726618705036, "step": 4550, "torque_loss": 0.26188263297080994 }, { "epoch": 4.100719424460432, "grad_norm": 0.3338139057159424, "learning_rate": 9.953709277328112e-05, "loss": 0.0383, "step": 4560 }, { "action_loss": 0.0687512457370758, "epoch": 4.100719424460432, "step": 4560 }, { "epoch": 4.100719424460432, "step": 4560, "torque_loss": 0.32518520951271057 }, { "epoch": 4.109712230215828, "grad_norm": 0.21497409045696259, "learning_rate": 9.953259246233032e-05, "loss": 0.0379, "step": 4570 }, { "action_loss": 0.024254167452454567, "epoch": 4.109712230215828, "step": 4570 }, { "epoch": 4.109712230215828, "step": 4570, "torque_loss": 0.23311716318130493 }, { "epoch": 4.118705035971223, "grad_norm": 0.2930673360824585, "learning_rate": 9.952807048414077e-05, "loss": 0.0417, "step": 4580 }, { "action_loss": 0.018447676673531532, "epoch": 4.118705035971223, "step": 4580 }, { "epoch": 4.118705035971223, "step": 4580, "torque_loss": 0.2058137208223343 }, { "epoch": 4.127697841726619, "grad_norm": 0.20905113220214844, "learning_rate": 9.95235268406905e-05, "loss": 0.0357, "step": 4590 }, { "action_loss": 0.0148129528388381, "epoch": 4.127697841726619, "step": 4590 }, { "epoch": 4.127697841726619, "step": 4590, "torque_loss": 0.13457469642162323 }, { "epoch": 4.136690647482014, "grad_norm": 0.30189380049705505, "learning_rate": 9.951896153396708e-05, "loss": 0.033, "step": 4600 }, { "action_loss": 0.013195253908634186, "epoch": 4.136690647482014, "step": 4600 }, { "epoch": 4.136690647482014, "step": 4600, "torque_loss": 0.152057945728302 }, { "epoch": 4.14568345323741, "grad_norm": 0.2900134027004242, "learning_rate": 9.95143745659675e-05, "loss": 0.0453, "step": 4610 }, { "action_loss": 0.011930051259696484, "epoch": 4.14568345323741, "step": 4610 }, { "epoch": 4.14568345323741, "step": 4610, "torque_loss": 0.17968495190143585 }, { "epoch": 4.154676258992806, "grad_norm": 0.20539464056491852, "learning_rate": 9.95097659386983e-05, "loss": 0.0397, "step": 4620 }, { "action_loss": 0.018991360440850258, "epoch": 4.154676258992806, "step": 4620 }, { "epoch": 4.154676258992806, "step": 4620, "torque_loss": 0.14170686900615692 }, { "epoch": 4.163669064748201, "grad_norm": 0.228366419672966, "learning_rate": 9.950513565417542e-05, "loss": 0.0312, "step": 4630 }, { "action_loss": 0.004595672246068716, "epoch": 4.163669064748201, "step": 4630 }, { "epoch": 4.163669064748201, "step": 4630, "torque_loss": 0.16362564265727997 }, { "epoch": 4.172661870503597, "grad_norm": 0.27944257855415344, "learning_rate": 9.95004837144243e-05, "loss": 0.0417, "step": 4640 }, { "action_loss": 0.016754774376749992, "epoch": 4.172661870503597, "step": 4640 }, { "epoch": 4.172661870503597, "step": 4640, "torque_loss": 0.1277502030134201 }, { "epoch": 4.181654676258993, "grad_norm": 0.21322989463806152, "learning_rate": 9.949581012147988e-05, "loss": 0.042, "step": 4650 }, { "action_loss": 0.013768653385341167, "epoch": 4.181654676258993, "step": 4650 }, { "epoch": 4.181654676258993, "step": 4650, "torque_loss": 0.2350909560918808 }, { "epoch": 4.190647482014389, "grad_norm": 0.312701016664505, "learning_rate": 9.949111487738653e-05, "loss": 0.046, "step": 4660 }, { "action_loss": 0.043131355196237564, "epoch": 4.190647482014389, "step": 4660 }, { "epoch": 4.190647482014389, "step": 4660, "torque_loss": 0.25204771757125854 }, { "epoch": 4.1996402877697845, "grad_norm": 0.2677856385707855, "learning_rate": 9.948639798419813e-05, "loss": 0.0432, "step": 4670 }, { "action_loss": 0.030299022793769836, "epoch": 4.1996402877697845, "step": 4670 }, { "epoch": 4.1996402877697845, "step": 4670, "torque_loss": 0.19203495979309082 }, { "epoch": 4.2086330935251794, "grad_norm": 0.24676083028316498, "learning_rate": 9.948165944397799e-05, "loss": 0.0369, "step": 4680 }, { "action_loss": 0.012038360349833965, "epoch": 4.2086330935251794, "step": 4680 }, { "epoch": 4.2086330935251794, "step": 4680, "torque_loss": 0.19212262332439423 }, { "epoch": 4.217625899280575, "grad_norm": 0.20849820971488953, "learning_rate": 9.94768992587989e-05, "loss": 0.0433, "step": 4690 }, { "action_loss": 0.015020966529846191, "epoch": 4.217625899280575, "step": 4690 }, { "epoch": 4.217625899280575, "step": 4690, "torque_loss": 0.22710001468658447 }, { "epoch": 4.226618705035971, "grad_norm": 0.34599876403808594, "learning_rate": 9.947211743074313e-05, "loss": 0.0481, "step": 4700 }, { "action_loss": 0.016201509162783623, "epoch": 4.226618705035971, "step": 4700 }, { "epoch": 4.226618705035971, "step": 4700, "torque_loss": 0.27165767550468445 }, { "epoch": 4.235611510791367, "grad_norm": 0.28820520639419556, "learning_rate": 9.946731396190246e-05, "loss": 0.036, "step": 4710 }, { "action_loss": 0.015812909230589867, "epoch": 4.235611510791367, "step": 4710 }, { "epoch": 4.235611510791367, "step": 4710, "torque_loss": 0.23585756123065948 }, { "epoch": 4.244604316546763, "grad_norm": 0.4233056902885437, "learning_rate": 9.946248885437803e-05, "loss": 0.0353, "step": 4720 }, { "action_loss": 0.02318836748600006, "epoch": 4.244604316546763, "step": 4720 }, { "epoch": 4.244604316546763, "step": 4720, "torque_loss": 0.25306153297424316 }, { "epoch": 4.253597122302159, "grad_norm": 0.20028962194919586, "learning_rate": 9.945764211028053e-05, "loss": 0.0326, "step": 4730 }, { "action_loss": 0.02817024290561676, "epoch": 4.253597122302159, "step": 4730 }, { "epoch": 4.253597122302159, "step": 4730, "torque_loss": 0.2863808274269104 }, { "epoch": 4.262589928057554, "grad_norm": 0.44694581627845764, "learning_rate": 9.94527737317301e-05, "loss": 0.0455, "step": 4740 }, { "action_loss": 0.013712093234062195, "epoch": 4.262589928057554, "step": 4740 }, { "epoch": 4.262589928057554, "step": 4740, "torque_loss": 0.18142493069171906 }, { "epoch": 4.2715827338129495, "grad_norm": 0.2768239378929138, "learning_rate": 9.944788372085631e-05, "loss": 0.0329, "step": 4750 }, { "action_loss": 0.018681159242987633, "epoch": 4.2715827338129495, "step": 4750 }, { "epoch": 4.2715827338129495, "step": 4750, "torque_loss": 0.17181344330310822 }, { "epoch": 4.280575539568345, "grad_norm": 0.33729416131973267, "learning_rate": 9.944297207979825e-05, "loss": 0.037, "step": 4760 }, { "action_loss": 0.013559561222791672, "epoch": 4.280575539568345, "step": 4760 }, { "epoch": 4.280575539568345, "step": 4760, "torque_loss": 0.17120592296123505 }, { "epoch": 4.289568345323741, "grad_norm": 0.3703579902648926, "learning_rate": 9.943803881070441e-05, "loss": 0.0416, "step": 4770 }, { "action_loss": 0.017200050875544548, "epoch": 4.289568345323741, "step": 4770 }, { "epoch": 4.289568345323741, "step": 4770, "torque_loss": 0.20061971247196198 }, { "epoch": 4.298561151079137, "grad_norm": 0.19373881816864014, "learning_rate": 9.943308391573278e-05, "loss": 0.0384, "step": 4780 }, { "action_loss": 0.013790451921522617, "epoch": 4.298561151079137, "step": 4780 }, { "epoch": 4.298561151079137, "step": 4780, "torque_loss": 0.2750548720359802 }, { "epoch": 4.307553956834532, "grad_norm": 0.3851507008075714, "learning_rate": 9.942810739705079e-05, "loss": 0.0464, "step": 4790 }, { "action_loss": 0.008606557734310627, "epoch": 4.307553956834532, "step": 4790 }, { "epoch": 4.307553956834532, "step": 4790, "torque_loss": 0.1440414935350418 }, { "epoch": 4.316546762589928, "grad_norm": 0.2491762787103653, "learning_rate": 9.942310925683538e-05, "loss": 0.0334, "step": 4800 }, { "action_loss": 0.012978833168745041, "epoch": 4.316546762589928, "step": 4800 }, { "epoch": 4.316546762589928, "step": 4800, "torque_loss": 0.1608983725309372 }, { "epoch": 4.325539568345324, "grad_norm": 0.3544607162475586, "learning_rate": 9.941808949727285e-05, "loss": 0.0385, "step": 4810 }, { "action_loss": 0.011047515086829662, "epoch": 4.325539568345324, "step": 4810 }, { "epoch": 4.325539568345324, "step": 4810, "torque_loss": 0.1708141416311264 }, { "epoch": 4.33453237410072, "grad_norm": 0.25728684663772583, "learning_rate": 9.941304812055903e-05, "loss": 0.0318, "step": 4820 }, { "action_loss": 0.016430607065558434, "epoch": 4.33453237410072, "step": 4820 }, { "epoch": 4.33453237410072, "step": 4820, "torque_loss": 0.192008838057518 }, { "epoch": 4.343525179856115, "grad_norm": 0.23042146861553192, "learning_rate": 9.940798512889921e-05, "loss": 0.0359, "step": 4830 }, { "action_loss": 0.016978342086076736, "epoch": 4.343525179856115, "step": 4830 }, { "epoch": 4.343525179856115, "step": 4830, "torque_loss": 0.197518989443779 }, { "epoch": 4.35251798561151, "grad_norm": 0.30524536967277527, "learning_rate": 9.94029005245081e-05, "loss": 0.0358, "step": 4840 }, { "action_loss": 0.016361815854907036, "epoch": 4.35251798561151, "step": 4840 }, { "epoch": 4.35251798561151, "step": 4840, "torque_loss": 0.18858854472637177 }, { "epoch": 4.361510791366906, "grad_norm": 0.23033642768859863, "learning_rate": 9.939779430960988e-05, "loss": 0.0345, "step": 4850 }, { "action_loss": 0.022831425070762634, "epoch": 4.361510791366906, "step": 4850 }, { "epoch": 4.361510791366906, "step": 4850, "torque_loss": 0.20899562537670135 }, { "epoch": 4.370503597122302, "grad_norm": 0.24468231201171875, "learning_rate": 9.939266648643817e-05, "loss": 0.0397, "step": 4860 }, { "action_loss": 0.03888725861907005, "epoch": 4.370503597122302, "step": 4860 }, { "epoch": 4.370503597122302, "step": 4860, "torque_loss": 0.32564377784729004 }, { "epoch": 4.379496402877698, "grad_norm": 0.2965606153011322, "learning_rate": 9.938751705723607e-05, "loss": 0.0418, "step": 4870 }, { "action_loss": 0.00777336023747921, "epoch": 4.379496402877698, "step": 4870 }, { "epoch": 4.379496402877698, "step": 4870, "torque_loss": 0.12090053409337997 }, { "epoch": 4.388489208633094, "grad_norm": 0.3411363661289215, "learning_rate": 9.938234602425613e-05, "loss": 0.037, "step": 4880 }, { "action_loss": 0.02534198947250843, "epoch": 4.388489208633094, "step": 4880 }, { "epoch": 4.388489208633094, "step": 4880, "torque_loss": 0.20070414245128632 }, { "epoch": 4.39748201438849, "grad_norm": 0.2958827614784241, "learning_rate": 9.93771533897603e-05, "loss": 0.0418, "step": 4890 }, { "action_loss": 0.006607512477785349, "epoch": 4.39748201438849, "step": 4890 }, { "epoch": 4.39748201438849, "step": 4890, "torque_loss": 0.0975755974650383 }, { "epoch": 4.406474820143885, "grad_norm": 0.25562548637390137, "learning_rate": 9.937193915602004e-05, "loss": 0.0341, "step": 4900 }, { "action_loss": 0.022609109058976173, "epoch": 4.406474820143885, "step": 4900 }, { "epoch": 4.406474820143885, "step": 4900, "torque_loss": 0.22096388041973114 }, { "epoch": 4.41546762589928, "grad_norm": 0.18412433564662933, "learning_rate": 9.936670332531621e-05, "loss": 0.0409, "step": 4910 }, { "action_loss": 0.010052393190562725, "epoch": 4.41546762589928, "step": 4910 }, { "epoch": 4.41546762589928, "step": 4910, "torque_loss": 0.17219705879688263 }, { "epoch": 4.424460431654676, "grad_norm": 0.2874140739440918, "learning_rate": 9.936144589993916e-05, "loss": 0.0415, "step": 4920 }, { "action_loss": 0.006846677977591753, "epoch": 4.424460431654676, "step": 4920 }, { "epoch": 4.424460431654676, "step": 4920, "torque_loss": 0.1480386108160019 }, { "epoch": 4.433453237410072, "grad_norm": 0.26682335138320923, "learning_rate": 9.935616688218867e-05, "loss": 0.0327, "step": 4930 }, { "action_loss": 0.02375982515513897, "epoch": 4.433453237410072, "step": 4930 }, { "epoch": 4.433453237410072, "step": 4930, "torque_loss": 0.18113893270492554 }, { "epoch": 4.442446043165468, "grad_norm": 0.25523486733436584, "learning_rate": 9.935086627437395e-05, "loss": 0.0365, "step": 4940 }, { "action_loss": 0.010950892232358456, "epoch": 4.442446043165468, "step": 4940 }, { "epoch": 4.442446043165468, "step": 4940, "torque_loss": 0.13267768919467926 }, { "epoch": 4.451438848920863, "grad_norm": 0.23794423043727875, "learning_rate": 9.934554407881366e-05, "loss": 0.0349, "step": 4950 }, { "action_loss": 0.02736852318048477, "epoch": 4.451438848920863, "step": 4950 }, { "epoch": 4.451438848920863, "step": 4950, "torque_loss": 0.1894267350435257 }, { "epoch": 4.460431654676259, "grad_norm": 0.32557880878448486, "learning_rate": 9.934020029783593e-05, "loss": 0.0388, "step": 4960 }, { "action_loss": 0.030194422230124474, "epoch": 4.460431654676259, "step": 4960 }, { "epoch": 4.460431654676259, "step": 4960, "torque_loss": 0.3287641108036041 }, { "epoch": 4.469424460431655, "grad_norm": 0.2811258137226105, "learning_rate": 9.933483493377829e-05, "loss": 0.0334, "step": 4970 }, { "action_loss": 0.01878385990858078, "epoch": 4.469424460431655, "step": 4970 }, { "epoch": 4.469424460431655, "step": 4970, "torque_loss": 0.15842324495315552 }, { "epoch": 4.4784172661870505, "grad_norm": 0.3196343183517456, "learning_rate": 9.932944798898774e-05, "loss": 0.0329, "step": 4980 }, { "action_loss": 0.010743051767349243, "epoch": 4.4784172661870505, "step": 4980 }, { "epoch": 4.4784172661870505, "step": 4980, "torque_loss": 0.187040314078331 }, { "epoch": 4.487410071942446, "grad_norm": 0.30248549580574036, "learning_rate": 9.932403946582072e-05, "loss": 0.0363, "step": 4990 }, { "action_loss": 0.03616374358534813, "epoch": 4.487410071942446, "step": 4990 }, { "epoch": 4.487410071942446, "step": 4990, "torque_loss": 0.26096490025520325 }, { "epoch": 4.496402877697841, "grad_norm": 0.2692549526691437, "learning_rate": 9.93186093666431e-05, "loss": 0.0419, "step": 5000 }, { "action_loss": 0.03923150897026062, "epoch": 4.496402877697841, "step": 5000 }, { "epoch": 4.496402877697841, "step": 5000, "torque_loss": 0.2732076048851013 }, { "epoch": 4.505395683453237, "grad_norm": 0.21223613619804382, "learning_rate": 9.931315769383018e-05, "loss": 0.037, "step": 5010 }, { "action_loss": 0.013623333536088467, "epoch": 4.505395683453237, "step": 5010 }, { "epoch": 4.505395683453237, "step": 5010, "torque_loss": 0.19605588912963867 }, { "epoch": 4.514388489208633, "grad_norm": 0.26653775572776794, "learning_rate": 9.930768444976672e-05, "loss": 0.0378, "step": 5020 }, { "action_loss": 0.03638948127627373, "epoch": 4.514388489208633, "step": 5020 }, { "epoch": 4.514388489208633, "step": 5020, "torque_loss": 0.2871191203594208 }, { "epoch": 4.523381294964029, "grad_norm": 0.24549417197704315, "learning_rate": 9.93021896368469e-05, "loss": 0.0386, "step": 5030 }, { "action_loss": 0.012522836215794086, "epoch": 4.523381294964029, "step": 5030 }, { "epoch": 4.523381294964029, "step": 5030, "torque_loss": 0.1910237818956375 }, { "epoch": 4.532374100719425, "grad_norm": 0.24967892467975616, "learning_rate": 9.929667325747431e-05, "loss": 0.0326, "step": 5040 }, { "action_loss": 0.014402431435883045, "epoch": 4.532374100719425, "step": 5040 }, { "epoch": 4.532374100719425, "step": 5040, "torque_loss": 0.23713384568691254 }, { "epoch": 4.5413669064748206, "grad_norm": 0.3602423369884491, "learning_rate": 9.929113531406205e-05, "loss": 0.0378, "step": 5050 }, { "action_loss": 0.007328387349843979, "epoch": 4.5413669064748206, "step": 5050 }, { "epoch": 4.5413669064748206, "step": 5050, "torque_loss": 0.14906130731105804 }, { "epoch": 4.5503597122302155, "grad_norm": 0.2578965127468109, "learning_rate": 9.928557580903257e-05, "loss": 0.0303, "step": 5060 }, { "action_loss": 0.00766464089974761, "epoch": 4.5503597122302155, "step": 5060 }, { "epoch": 4.5503597122302155, "step": 5060, "torque_loss": 0.09736815840005875 }, { "epoch": 4.559352517985611, "grad_norm": 0.20581433176994324, "learning_rate": 9.927999474481779e-05, "loss": 0.0368, "step": 5070 }, { "action_loss": 0.01515835803002119, "epoch": 4.559352517985611, "step": 5070 }, { "epoch": 4.559352517985611, "step": 5070, "torque_loss": 0.19183532893657684 }, { "epoch": 4.568345323741007, "grad_norm": 0.31723883748054504, "learning_rate": 9.927439212385907e-05, "loss": 0.0382, "step": 5080 }, { "action_loss": 0.016854865476489067, "epoch": 4.568345323741007, "step": 5080 }, { "epoch": 4.568345323741007, "step": 5080, "torque_loss": 0.23176217079162598 }, { "epoch": 4.577338129496403, "grad_norm": 0.2346765697002411, "learning_rate": 9.926876794860718e-05, "loss": 0.0402, "step": 5090 }, { "action_loss": 0.007408909499645233, "epoch": 4.577338129496403, "step": 5090 }, { "epoch": 4.577338129496403, "step": 5090, "torque_loss": 0.21272313594818115 }, { "epoch": 4.586330935251799, "grad_norm": 0.3269188106060028, "learning_rate": 9.926312222152235e-05, "loss": 0.0407, "step": 5100 }, { "action_loss": 0.015699058771133423, "epoch": 4.586330935251799, "step": 5100 }, { "epoch": 4.586330935251799, "step": 5100, "torque_loss": 0.22285246849060059 }, { "epoch": 4.595323741007194, "grad_norm": 0.1920422911643982, "learning_rate": 9.925745494507414e-05, "loss": 0.0383, "step": 5110 }, { "action_loss": 0.022247666493058205, "epoch": 4.595323741007194, "step": 5110 }, { "epoch": 4.595323741007194, "step": 5110, "torque_loss": 0.19875027239322662 }, { "epoch": 4.60431654676259, "grad_norm": 0.3825668394565582, "learning_rate": 9.925176612174169e-05, "loss": 0.0374, "step": 5120 }, { "action_loss": 0.024428142234683037, "epoch": 4.60431654676259, "step": 5120 }, { "epoch": 4.60431654676259, "step": 5120, "torque_loss": 0.2510491609573364 }, { "epoch": 4.613309352517986, "grad_norm": 0.4073249399662018, "learning_rate": 9.924605575401346e-05, "loss": 0.0353, "step": 5130 }, { "action_loss": 0.03836296871304512, "epoch": 4.613309352517986, "step": 5130 }, { "epoch": 4.613309352517986, "step": 5130, "torque_loss": 0.2671622931957245 }, { "epoch": 4.622302158273381, "grad_norm": 0.2670999765396118, "learning_rate": 9.924032384438733e-05, "loss": 0.0453, "step": 5140 }, { "action_loss": 0.01523582637310028, "epoch": 4.622302158273381, "step": 5140 }, { "epoch": 4.622302158273381, "step": 5140, "torque_loss": 0.24175764620304108 }, { "epoch": 4.631294964028777, "grad_norm": 0.30314958095550537, "learning_rate": 9.923457039537066e-05, "loss": 0.0348, "step": 5150 }, { "action_loss": 0.014245983213186264, "epoch": 4.631294964028777, "step": 5150 }, { "epoch": 4.631294964028777, "step": 5150, "torque_loss": 0.18870048224925995 }, { "epoch": 4.640287769784173, "grad_norm": 0.2223777323961258, "learning_rate": 9.92287954094802e-05, "loss": 0.0375, "step": 5160 }, { "action_loss": 0.009851954877376556, "epoch": 4.640287769784173, "step": 5160 }, { "epoch": 4.640287769784173, "step": 5160, "torque_loss": 0.22172103822231293 }, { "epoch": 4.649280575539568, "grad_norm": 0.21119844913482666, "learning_rate": 9.922299888924212e-05, "loss": 0.0348, "step": 5170 }, { "action_loss": 0.020251648500561714, "epoch": 4.649280575539568, "step": 5170 }, { "epoch": 4.649280575539568, "step": 5170, "torque_loss": 0.2738679349422455 }, { "epoch": 4.658273381294964, "grad_norm": 0.23608602583408356, "learning_rate": 9.921718083719203e-05, "loss": 0.0401, "step": 5180 }, { "action_loss": 0.012969222851097584, "epoch": 4.658273381294964, "step": 5180 }, { "epoch": 4.658273381294964, "step": 5180, "torque_loss": 0.19744586944580078 }, { "epoch": 4.66726618705036, "grad_norm": 0.22203566133975983, "learning_rate": 9.921134125587491e-05, "loss": 0.0338, "step": 5190 }, { "action_loss": 0.028013430535793304, "epoch": 4.66726618705036, "step": 5190 }, { "epoch": 4.66726618705036, "step": 5190, "torque_loss": 0.26269009709358215 }, { "epoch": 4.676258992805756, "grad_norm": 0.19187013804912567, "learning_rate": 9.920548014784523e-05, "loss": 0.0385, "step": 5200 }, { "action_loss": 0.02507246844470501, "epoch": 4.676258992805756, "step": 5200 }, { "epoch": 4.676258992805756, "step": 5200, "torque_loss": 0.21576206386089325 }, { "epoch": 4.685251798561151, "grad_norm": 0.2274143099784851, "learning_rate": 9.919959751566681e-05, "loss": 0.0359, "step": 5210 }, { "action_loss": 0.009634428657591343, "epoch": 4.685251798561151, "step": 5210 }, { "epoch": 4.685251798561151, "step": 5210, "torque_loss": 0.21234752237796783 }, { "epoch": 4.694244604316546, "grad_norm": 0.25139209628105164, "learning_rate": 9.919369336191291e-05, "loss": 0.0305, "step": 5220 }, { "action_loss": 0.012007378041744232, "epoch": 4.694244604316546, "step": 5220 }, { "epoch": 4.694244604316546, "step": 5220, "torque_loss": 0.15313732624053955 }, { "epoch": 4.703237410071942, "grad_norm": 0.19148287177085876, "learning_rate": 9.918776768916625e-05, "loss": 0.0357, "step": 5230 }, { "action_loss": 0.010883837006986141, "epoch": 4.703237410071942, "step": 5230 }, { "epoch": 4.703237410071942, "step": 5230, "torque_loss": 0.19350405037403107 }, { "epoch": 4.712230215827338, "grad_norm": 0.21738046407699585, "learning_rate": 9.918182050001888e-05, "loss": 0.0329, "step": 5240 }, { "action_loss": 0.016322575509548187, "epoch": 4.712230215827338, "step": 5240 }, { "epoch": 4.712230215827338, "step": 5240, "torque_loss": 0.1797008067369461 }, { "epoch": 4.721223021582734, "grad_norm": 0.17685312032699585, "learning_rate": 9.91758517970723e-05, "loss": 0.0288, "step": 5250 }, { "action_loss": 0.014389521442353725, "epoch": 4.721223021582734, "step": 5250 }, { "epoch": 4.721223021582734, "step": 5250, "torque_loss": 0.20674942433834076 }, { "epoch": 4.73021582733813, "grad_norm": 0.2877127528190613, "learning_rate": 9.916986158293744e-05, "loss": 0.0345, "step": 5260 }, { "action_loss": 0.021448343992233276, "epoch": 4.73021582733813, "step": 5260 }, { "epoch": 4.73021582733813, "step": 5260, "torque_loss": 0.1973649263381958 }, { "epoch": 4.739208633093525, "grad_norm": 0.22940166294574738, "learning_rate": 9.916384986023463e-05, "loss": 0.0329, "step": 5270 }, { "action_loss": 0.00673726387321949, "epoch": 4.739208633093525, "step": 5270 }, { "epoch": 4.739208633093525, "step": 5270, "torque_loss": 0.2129761427640915 }, { "epoch": 4.748201438848921, "grad_norm": 0.2659098207950592, "learning_rate": 9.915781663159359e-05, "loss": 0.0275, "step": 5280 }, { "action_loss": 0.01317938044667244, "epoch": 4.748201438848921, "step": 5280 }, { "epoch": 4.748201438848921, "step": 5280, "torque_loss": 0.1981297731399536 }, { "epoch": 4.7571942446043165, "grad_norm": 0.22119496762752533, "learning_rate": 9.915176189965346e-05, "loss": 0.0332, "step": 5290 }, { "action_loss": 0.029652336612343788, "epoch": 4.7571942446043165, "step": 5290 }, { "epoch": 4.7571942446043165, "step": 5290, "torque_loss": 0.3250660300254822 }, { "epoch": 4.766187050359712, "grad_norm": 0.24052757024765015, "learning_rate": 9.914568566706279e-05, "loss": 0.043, "step": 5300 }, { "action_loss": 0.03818875923752785, "epoch": 4.766187050359712, "step": 5300 }, { "epoch": 4.766187050359712, "step": 5300, "torque_loss": 0.252105712890625 }, { "epoch": 4.775179856115108, "grad_norm": 0.33658015727996826, "learning_rate": 9.913958793647953e-05, "loss": 0.0439, "step": 5310 }, { "action_loss": 0.0071300603449344635, "epoch": 4.775179856115108, "step": 5310 }, { "epoch": 4.775179856115108, "step": 5310, "torque_loss": 0.12486160546541214 }, { "epoch": 4.784172661870503, "grad_norm": 0.2262764573097229, "learning_rate": 9.913346871057102e-05, "loss": 0.0341, "step": 5320 }, { "action_loss": 0.021540867164731026, "epoch": 4.784172661870503, "step": 5320 }, { "epoch": 4.784172661870503, "step": 5320, "torque_loss": 0.2454574704170227 }, { "epoch": 4.793165467625899, "grad_norm": 0.33543187379837036, "learning_rate": 9.912732799201403e-05, "loss": 0.0399, "step": 5330 }, { "action_loss": 0.017794286832213402, "epoch": 4.793165467625899, "step": 5330 }, { "epoch": 4.793165467625899, "step": 5330, "torque_loss": 0.19143564999103546 }, { "epoch": 4.802158273381295, "grad_norm": 0.21020378172397614, "learning_rate": 9.912116578349474e-05, "loss": 0.0344, "step": 5340 }, { "action_loss": 0.01549906749278307, "epoch": 4.802158273381295, "step": 5340 }, { "epoch": 4.802158273381295, "step": 5340, "torque_loss": 0.2074722796678543 }, { "epoch": 4.811151079136691, "grad_norm": 0.18370333313941956, "learning_rate": 9.911498208770867e-05, "loss": 0.0429, "step": 5350 }, { "action_loss": 0.037717949599027634, "epoch": 4.811151079136691, "step": 5350 }, { "epoch": 4.811151079136691, "step": 5350, "torque_loss": 0.29331305623054504 }, { "epoch": 4.820143884892087, "grad_norm": 0.3151189088821411, "learning_rate": 9.910877690736078e-05, "loss": 0.0388, "step": 5360 }, { "action_loss": 0.009055317379534245, "epoch": 4.820143884892087, "step": 5360 }, { "epoch": 4.820143884892087, "step": 5360, "torque_loss": 0.2094794511795044 }, { "epoch": 4.829136690647482, "grad_norm": 0.21977154910564423, "learning_rate": 9.910255024516546e-05, "loss": 0.0294, "step": 5370 }, { "action_loss": 0.011867512948811054, "epoch": 4.829136690647482, "step": 5370 }, { "epoch": 4.829136690647482, "step": 5370, "torque_loss": 0.21691320836544037 }, { "epoch": 4.838129496402877, "grad_norm": 0.27351272106170654, "learning_rate": 9.909630210384644e-05, "loss": 0.0399, "step": 5380 }, { "action_loss": 0.018553027883172035, "epoch": 4.838129496402877, "step": 5380 }, { "epoch": 4.838129496402877, "step": 5380, "torque_loss": 0.25806620717048645 }, { "epoch": 4.847122302158273, "grad_norm": 0.4768608510494232, "learning_rate": 9.909003248613688e-05, "loss": 0.0415, "step": 5390 }, { "action_loss": 0.013964137993752956, "epoch": 4.847122302158273, "step": 5390 }, { "epoch": 4.847122302158273, "step": 5390, "torque_loss": 0.17496563494205475 }, { "epoch": 4.856115107913669, "grad_norm": 0.3105069100856781, "learning_rate": 9.908374139477934e-05, "loss": 0.0401, "step": 5400 }, { "action_loss": 0.013288792222738266, "epoch": 4.856115107913669, "step": 5400 }, { "epoch": 4.856115107913669, "step": 5400, "torque_loss": 0.15699827671051025 }, { "epoch": 4.865107913669065, "grad_norm": 0.2533526122570038, "learning_rate": 9.907742883252571e-05, "loss": 0.0366, "step": 5410 }, { "action_loss": 0.014277520589530468, "epoch": 4.865107913669065, "step": 5410 }, { "epoch": 4.865107913669065, "step": 5410, "torque_loss": 0.13595828413963318 }, { "epoch": 4.874100719424461, "grad_norm": 0.2524736523628235, "learning_rate": 9.907109480213736e-05, "loss": 0.0378, "step": 5420 }, { "action_loss": 0.022553687915205956, "epoch": 4.874100719424461, "step": 5420 }, { "epoch": 4.874100719424461, "step": 5420, "torque_loss": 0.2544710338115692 }, { "epoch": 4.883093525179856, "grad_norm": 0.3551200032234192, "learning_rate": 9.906473930638498e-05, "loss": 0.0361, "step": 5430 }, { "action_loss": 0.009906820021569729, "epoch": 4.883093525179856, "step": 5430 }, { "epoch": 4.883093525179856, "step": 5430, "torque_loss": 0.21433468163013458 }, { "epoch": 4.892086330935252, "grad_norm": 0.20831958949565887, "learning_rate": 9.90583623480487e-05, "loss": 0.038, "step": 5440 }, { "action_loss": 0.019834229722619057, "epoch": 4.892086330935252, "step": 5440 }, { "epoch": 4.892086330935252, "step": 5440, "torque_loss": 0.17839163541793823 }, { "epoch": 4.901079136690647, "grad_norm": 0.26625367999076843, "learning_rate": 9.905196392991802e-05, "loss": 0.041, "step": 5450 }, { "action_loss": 0.01147453486919403, "epoch": 4.901079136690647, "step": 5450 }, { "epoch": 4.901079136690647, "step": 5450, "torque_loss": 0.20108072459697723 }, { "epoch": 4.910071942446043, "grad_norm": 0.3720373809337616, "learning_rate": 9.904554405479183e-05, "loss": 0.0378, "step": 5460 }, { "action_loss": 0.01712655834853649, "epoch": 4.910071942446043, "step": 5460 }, { "epoch": 4.910071942446043, "step": 5460, "torque_loss": 0.18690796196460724 }, { "epoch": 4.919064748201439, "grad_norm": 0.3345375061035156, "learning_rate": 9.903910272547838e-05, "loss": 0.0434, "step": 5470 }, { "action_loss": 0.01805746741592884, "epoch": 4.919064748201439, "step": 5470 }, { "epoch": 4.919064748201439, "step": 5470, "torque_loss": 0.22833694517612457 }, { "epoch": 4.928057553956835, "grad_norm": 0.33657339215278625, "learning_rate": 9.903263994479533e-05, "loss": 0.033, "step": 5480 }, { "action_loss": 0.013150329701602459, "epoch": 4.928057553956835, "step": 5480 }, { "epoch": 4.928057553956835, "step": 5480, "torque_loss": 0.18877319991588593 }, { "epoch": 4.93705035971223, "grad_norm": 0.22355948388576508, "learning_rate": 9.902615571556972e-05, "loss": 0.033, "step": 5490 }, { "action_loss": 0.014087931253015995, "epoch": 4.93705035971223, "step": 5490 }, { "epoch": 4.93705035971223, "step": 5490, "torque_loss": 0.16273115575313568 }, { "epoch": 4.946043165467626, "grad_norm": 0.17483201622962952, "learning_rate": 9.9019650040638e-05, "loss": 0.0334, "step": 5500 }, { "action_loss": 0.029800139367580414, "epoch": 4.946043165467626, "step": 5500 }, { "epoch": 4.946043165467626, "step": 5500, "torque_loss": 0.24753127992153168 }, { "epoch": 4.955035971223022, "grad_norm": 0.23372767865657806, "learning_rate": 9.901312292284594e-05, "loss": 0.0361, "step": 5510 }, { "action_loss": 0.024977819994091988, "epoch": 4.955035971223022, "step": 5510 }, { "epoch": 4.955035971223022, "step": 5510, "torque_loss": 0.14055036008358002 }, { "epoch": 4.9640287769784175, "grad_norm": 0.2815327048301697, "learning_rate": 9.900657436504875e-05, "loss": 0.0347, "step": 5520 }, { "action_loss": 0.02197909541428089, "epoch": 4.9640287769784175, "step": 5520 }, { "epoch": 4.9640287769784175, "step": 5520, "torque_loss": 0.1472618728876114 }, { "epoch": 4.9730215827338125, "grad_norm": 0.30424344539642334, "learning_rate": 9.900000437011097e-05, "loss": 0.0382, "step": 5530 }, { "action_loss": 0.010331119410693645, "epoch": 4.9730215827338125, "step": 5530 }, { "epoch": 4.9730215827338125, "step": 5530, "torque_loss": 0.15707720816135406 }, { "epoch": 4.982014388489208, "grad_norm": 0.2354409247636795, "learning_rate": 9.899341294090653e-05, "loss": 0.0306, "step": 5540 }, { "action_loss": 0.010295561514794827, "epoch": 4.982014388489208, "step": 5540 }, { "epoch": 4.982014388489208, "step": 5540, "torque_loss": 0.19228379428386688 }, { "epoch": 4.991007194244604, "grad_norm": 0.3393881916999817, "learning_rate": 9.898680008031877e-05, "loss": 0.0346, "step": 5550 }, { "action_loss": 0.015022027306258678, "epoch": 4.991007194244604, "step": 5550 }, { "epoch": 4.991007194244604, "step": 5550, "torque_loss": 0.15349937975406647 }, { "epoch": 5.0, "grad_norm": 0.30487877130508423, "learning_rate": 9.898016579124037e-05, "loss": 0.037, "step": 5560 }, { "action_loss": 0.018238382413983345, "epoch": 5.0, "step": 5560 }, { "epoch": 5.0, "step": 5560, "torque_loss": 0.16849128901958466 }, { "epoch": 5.008992805755396, "grad_norm": 0.22233229875564575, "learning_rate": 9.89735100765734e-05, "loss": 0.0368, "step": 5570 }, { "action_loss": 0.013041555881500244, "epoch": 5.008992805755396, "step": 5570 }, { "epoch": 5.008992805755396, "step": 5570, "torque_loss": 0.25215229392051697 }, { "epoch": 5.017985611510792, "grad_norm": 0.2698548436164856, "learning_rate": 9.89668329392293e-05, "loss": 0.0315, "step": 5580 }, { "action_loss": 0.01267301570624113, "epoch": 5.017985611510792, "step": 5580 }, { "epoch": 5.017985611510792, "step": 5580, "torque_loss": 0.1742863804101944 }, { "epoch": 5.026978417266187, "grad_norm": 0.2276596873998642, "learning_rate": 9.896013438212885e-05, "loss": 0.0336, "step": 5590 }, { "action_loss": 0.025819629430770874, "epoch": 5.026978417266187, "step": 5590 }, { "epoch": 5.026978417266187, "step": 5590, "torque_loss": 0.3056601285934448 }, { "epoch": 5.0359712230215825, "grad_norm": 0.284073144197464, "learning_rate": 9.895341440820225e-05, "loss": 0.0391, "step": 5600 }, { "action_loss": 0.016337258741259575, "epoch": 5.0359712230215825, "step": 5600 }, { "epoch": 5.0359712230215825, "step": 5600, "torque_loss": 0.1818888932466507 }, { "epoch": 5.044964028776978, "grad_norm": 0.19935961067676544, "learning_rate": 9.894667302038902e-05, "loss": 0.0359, "step": 5610 }, { "action_loss": 0.04154519364237785, "epoch": 5.044964028776978, "step": 5610 }, { "epoch": 5.044964028776978, "step": 5610, "torque_loss": 0.16438357532024384 }, { "epoch": 5.053956834532374, "grad_norm": 0.2262466698884964, "learning_rate": 9.893991022163812e-05, "loss": 0.0363, "step": 5620 }, { "action_loss": 0.012534874491393566, "epoch": 5.053956834532374, "step": 5620 }, { "epoch": 5.053956834532374, "step": 5620, "torque_loss": 0.16720782220363617 }, { "epoch": 5.06294964028777, "grad_norm": 0.22129759192466736, "learning_rate": 9.893312601490779e-05, "loss": 0.0279, "step": 5630 }, { "action_loss": 0.023487282916903496, "epoch": 5.06294964028777, "step": 5630 }, { "epoch": 5.06294964028777, "step": 5630, "torque_loss": 0.21747148036956787 }, { "epoch": 5.071942446043165, "grad_norm": 0.25412529706954956, "learning_rate": 9.892632040316568e-05, "loss": 0.039, "step": 5640 }, { "action_loss": 0.010521382093429565, "epoch": 5.071942446043165, "step": 5640 }, { "epoch": 5.071942446043165, "step": 5640, "torque_loss": 0.15664488077163696 }, { "epoch": 5.080935251798561, "grad_norm": 0.34960630536079407, "learning_rate": 9.891949338938877e-05, "loss": 0.0357, "step": 5650 }, { "action_loss": 0.015493582934141159, "epoch": 5.080935251798561, "step": 5650 }, { "epoch": 5.080935251798561, "step": 5650, "torque_loss": 0.20275382697582245 }, { "epoch": 5.089928057553957, "grad_norm": 0.33013591170310974, "learning_rate": 9.89126449765635e-05, "loss": 0.0383, "step": 5660 }, { "action_loss": 0.010965903289616108, "epoch": 5.089928057553957, "step": 5660 }, { "epoch": 5.089928057553957, "step": 5660, "torque_loss": 0.15598376095294952 }, { "epoch": 5.098920863309353, "grad_norm": 0.24977785348892212, "learning_rate": 9.890577516768551e-05, "loss": 0.0294, "step": 5670 }, { "action_loss": 0.09348752349615097, "epoch": 5.098920863309353, "step": 5670 }, { "epoch": 5.098920863309353, "step": 5670, "torque_loss": 0.2864314317703247 }, { "epoch": 5.107913669064748, "grad_norm": 0.1879677027463913, "learning_rate": 9.889888396575995e-05, "loss": 0.0436, "step": 5680 }, { "action_loss": 0.008056002669036388, "epoch": 5.107913669064748, "step": 5680 }, { "epoch": 5.107913669064748, "step": 5680, "torque_loss": 0.19880418479442596 }, { "epoch": 5.116906474820144, "grad_norm": 0.22320441901683807, "learning_rate": 9.889197137380125e-05, "loss": 0.0302, "step": 5690 }, { "action_loss": 0.030901586636900902, "epoch": 5.116906474820144, "step": 5690 }, { "epoch": 5.116906474820144, "step": 5690, "torque_loss": 0.335636168718338 }, { "epoch": 5.125899280575539, "grad_norm": 0.19828800857067108, "learning_rate": 9.888503739483321e-05, "loss": 0.0389, "step": 5700 }, { "action_loss": 0.006884100381284952, "epoch": 5.125899280575539, "step": 5700 }, { "epoch": 5.125899280575539, "step": 5700, "torque_loss": 0.13000597059726715 }, { "epoch": 5.134892086330935, "grad_norm": 0.21113580465316772, "learning_rate": 9.887808203188897e-05, "loss": 0.0337, "step": 5710 }, { "action_loss": 0.009440845809876919, "epoch": 5.134892086330935, "step": 5710 }, { "epoch": 5.134892086330935, "step": 5710, "torque_loss": 0.16059015691280365 }, { "epoch": 5.143884892086331, "grad_norm": 0.14249713718891144, "learning_rate": 9.887110528801108e-05, "loss": 0.0299, "step": 5720 }, { "action_loss": 0.010515662841498852, "epoch": 5.143884892086331, "step": 5720 }, { "epoch": 5.143884892086331, "step": 5720, "torque_loss": 0.1229720488190651 }, { "epoch": 5.152877697841727, "grad_norm": 0.3090737760066986, "learning_rate": 9.886410716625137e-05, "loss": 0.0357, "step": 5730 }, { "action_loss": 0.006768880877643824, "epoch": 5.152877697841727, "step": 5730 }, { "epoch": 5.152877697841727, "step": 5730, "torque_loss": 0.18319614231586456 }, { "epoch": 5.161870503597123, "grad_norm": 0.2382190078496933, "learning_rate": 9.885708766967108e-05, "loss": 0.0298, "step": 5740 }, { "action_loss": 0.007826067507266998, "epoch": 5.161870503597123, "step": 5740 }, { "epoch": 5.161870503597123, "step": 5740, "torque_loss": 0.17040376365184784 }, { "epoch": 5.170863309352518, "grad_norm": 0.4073828458786011, "learning_rate": 9.885004680134076e-05, "loss": 0.0374, "step": 5750 }, { "action_loss": 0.005990695208311081, "epoch": 5.170863309352518, "step": 5750 }, { "epoch": 5.170863309352518, "step": 5750, "torque_loss": 0.15421567857265472 }, { "epoch": 5.179856115107913, "grad_norm": 0.19894881546497345, "learning_rate": 9.884298456434033e-05, "loss": 0.0315, "step": 5760 }, { "action_loss": 0.01010716799646616, "epoch": 5.179856115107913, "step": 5760 }, { "epoch": 5.179856115107913, "step": 5760, "torque_loss": 0.2148268073797226 }, { "epoch": 5.188848920863309, "grad_norm": 0.2206190824508667, "learning_rate": 9.883590096175905e-05, "loss": 0.0332, "step": 5770 }, { "action_loss": 0.0337144136428833, "epoch": 5.188848920863309, "step": 5770 }, { "epoch": 5.188848920863309, "step": 5770, "torque_loss": 0.22723805904388428 }, { "epoch": 5.197841726618705, "grad_norm": 0.2759213447570801, "learning_rate": 9.882879599669554e-05, "loss": 0.0372, "step": 5780 }, { "action_loss": 0.012931185774505138, "epoch": 5.197841726618705, "step": 5780 }, { "epoch": 5.197841726618705, "step": 5780, "torque_loss": 0.1999068260192871 }, { "epoch": 5.206834532374101, "grad_norm": 0.4527994990348816, "learning_rate": 9.882166967225775e-05, "loss": 0.0381, "step": 5790 }, { "action_loss": 0.03465793654322624, "epoch": 5.206834532374101, "step": 5790 }, { "epoch": 5.206834532374101, "step": 5790, "torque_loss": 0.27744045853614807 }, { "epoch": 5.215827338129497, "grad_norm": 0.28949490189552307, "learning_rate": 9.881452199156296e-05, "loss": 0.0372, "step": 5800 }, { "action_loss": 0.022631680592894554, "epoch": 5.215827338129497, "step": 5800 }, { "epoch": 5.215827338129497, "step": 5800, "torque_loss": 0.23589615523815155 }, { "epoch": 5.224820143884892, "grad_norm": 0.24488933384418488, "learning_rate": 9.880735295773782e-05, "loss": 0.0492, "step": 5810 }, { "action_loss": 0.015045839361846447, "epoch": 5.224820143884892, "step": 5810 }, { "epoch": 5.224820143884892, "step": 5810, "torque_loss": 0.17539118230342865 }, { "epoch": 5.233812949640288, "grad_norm": 0.2522944509983063, "learning_rate": 9.88001625739183e-05, "loss": 0.0421, "step": 5820 }, { "action_loss": 0.01270270999521017, "epoch": 5.233812949640288, "step": 5820 }, { "epoch": 5.233812949640288, "step": 5820, "torque_loss": 0.13369286060333252 }, { "epoch": 5.2428057553956835, "grad_norm": 0.15548378229141235, "learning_rate": 9.879295084324976e-05, "loss": 0.0384, "step": 5830 }, { "action_loss": 0.0314728282392025, "epoch": 5.2428057553956835, "step": 5830 }, { "epoch": 5.2428057553956835, "step": 5830, "torque_loss": 0.2542954981327057 }, { "epoch": 5.251798561151079, "grad_norm": 0.24147416651248932, "learning_rate": 9.87857177688868e-05, "loss": 0.0344, "step": 5840 }, { "action_loss": 0.007940857671201229, "epoch": 5.251798561151079, "step": 5840 }, { "epoch": 5.251798561151079, "step": 5840, "torque_loss": 0.15240629017353058 }, { "epoch": 5.260791366906475, "grad_norm": 0.30958831310272217, "learning_rate": 9.877846335399343e-05, "loss": 0.036, "step": 5850 }, { "action_loss": 0.010167316533625126, "epoch": 5.260791366906475, "step": 5850 }, { "epoch": 5.260791366906475, "step": 5850, "torque_loss": 0.12473690509796143 }, { "epoch": 5.26978417266187, "grad_norm": 0.2307387739419937, "learning_rate": 9.8771187601743e-05, "loss": 0.0337, "step": 5860 }, { "action_loss": 0.01104799285531044, "epoch": 5.26978417266187, "step": 5860 }, { "epoch": 5.26978417266187, "step": 5860, "torque_loss": 0.1783452033996582 }, { "epoch": 5.278776978417266, "grad_norm": 0.27535372972488403, "learning_rate": 9.876389051531813e-05, "loss": 0.0324, "step": 5870 }, { "action_loss": 0.00719313882291317, "epoch": 5.278776978417266, "step": 5870 }, { "epoch": 5.278776978417266, "step": 5870, "torque_loss": 0.16083940863609314 }, { "epoch": 5.287769784172662, "grad_norm": 0.15380287170410156, "learning_rate": 9.875657209791088e-05, "loss": 0.0345, "step": 5880 }, { "action_loss": 0.028447752818465233, "epoch": 5.287769784172662, "step": 5880 }, { "epoch": 5.287769784172662, "step": 5880, "torque_loss": 0.236545130610466 }, { "epoch": 5.296762589928058, "grad_norm": 0.267253041267395, "learning_rate": 9.874923235272248e-05, "loss": 0.0313, "step": 5890 }, { "action_loss": 0.02301582135260105, "epoch": 5.296762589928058, "step": 5890 }, { "epoch": 5.296762589928058, "step": 5890, "torque_loss": 0.21220223605632782 }, { "epoch": 5.305755395683454, "grad_norm": 0.2017565369606018, "learning_rate": 9.874187128296369e-05, "loss": 0.0346, "step": 5900 }, { "action_loss": 0.028703520074486732, "epoch": 5.305755395683454, "step": 5900 }, { "epoch": 5.305755395683454, "step": 5900, "torque_loss": 0.24132847785949707 }, { "epoch": 5.3147482014388485, "grad_norm": 0.3197766840457916, "learning_rate": 9.873448889185439e-05, "loss": 0.0384, "step": 5910 }, { "action_loss": 0.026194147765636444, "epoch": 5.3147482014388485, "step": 5910 }, { "epoch": 5.3147482014388485, "step": 5910, "torque_loss": 0.23724035918712616 }, { "epoch": 5.323741007194244, "grad_norm": 0.31363382935523987, "learning_rate": 9.872708518262397e-05, "loss": 0.0406, "step": 5920 }, { "action_loss": 0.011259305290877819, "epoch": 5.323741007194244, "step": 5920 }, { "epoch": 5.323741007194244, "step": 5920, "torque_loss": 0.16486243903636932 }, { "epoch": 5.33273381294964, "grad_norm": 0.2246386855840683, "learning_rate": 9.871966015851101e-05, "loss": 0.0373, "step": 5930 }, { "action_loss": 0.030195334926247597, "epoch": 5.33273381294964, "step": 5930 }, { "epoch": 5.33273381294964, "step": 5930, "torque_loss": 0.2454845905303955 }, { "epoch": 5.341726618705036, "grad_norm": 0.23816409707069397, "learning_rate": 9.87122138227635e-05, "loss": 0.0417, "step": 5940 }, { "action_loss": 0.014693831093609333, "epoch": 5.341726618705036, "step": 5940 }, { "epoch": 5.341726618705036, "step": 5940, "torque_loss": 0.2752629220485687 }, { "epoch": 5.350719424460432, "grad_norm": 0.28694960474967957, "learning_rate": 9.87047461786387e-05, "loss": 0.0361, "step": 5950 }, { "action_loss": 0.007840428501367569, "epoch": 5.350719424460432, "step": 5950 }, { "epoch": 5.350719424460432, "step": 5950, "torque_loss": 0.2617259919643402 }, { "epoch": 5.359712230215827, "grad_norm": 0.23752400279045105, "learning_rate": 9.869725722940323e-05, "loss": 0.0335, "step": 5960 }, { "action_loss": 0.012761970050632954, "epoch": 5.359712230215827, "step": 5960 }, { "epoch": 5.359712230215827, "step": 5960, "torque_loss": 0.12883497774600983 }, { "epoch": 5.368705035971223, "grad_norm": 0.17245276272296906, "learning_rate": 9.868974697833299e-05, "loss": 0.0335, "step": 5970 }, { "action_loss": 0.009559202007949352, "epoch": 5.368705035971223, "step": 5970 }, { "epoch": 5.368705035971223, "step": 5970, "torque_loss": 0.13226477801799774 }, { "epoch": 5.377697841726619, "grad_norm": 0.2506467401981354, "learning_rate": 9.868221542871324e-05, "loss": 0.0286, "step": 5980 }, { "action_loss": 0.010122195817530155, "epoch": 5.377697841726619, "step": 5980 }, { "epoch": 5.377697841726619, "step": 5980, "torque_loss": 0.14424866437911987 }, { "epoch": 5.386690647482014, "grad_norm": 0.3165140450000763, "learning_rate": 9.867466258383853e-05, "loss": 0.0363, "step": 5990 }, { "action_loss": 0.015242546796798706, "epoch": 5.386690647482014, "step": 5990 }, { "epoch": 5.386690647482014, "step": 5990, "torque_loss": 0.28410807251930237 }, { "epoch": 5.39568345323741, "grad_norm": 0.34431999921798706, "learning_rate": 9.866708844701272e-05, "loss": 0.0413, "step": 6000 }, { "action_loss": 0.01898767799139023, "epoch": 5.39568345323741, "step": 6000 }, { "epoch": 5.39568345323741, "step": 6000, "torque_loss": 0.17796190083026886 }, { "epoch": 5.404676258992806, "grad_norm": 0.22980011999607086, "learning_rate": 9.865949302154899e-05, "loss": 0.0361, "step": 6010 }, { "action_loss": 0.007859327830374241, "epoch": 5.404676258992806, "step": 6010 }, { "epoch": 5.404676258992806, "step": 6010, "torque_loss": 0.12832000851631165 }, { "epoch": 5.413669064748201, "grad_norm": 0.35984542965888977, "learning_rate": 9.865187631076987e-05, "loss": 0.0323, "step": 6020 }, { "action_loss": 0.011755370534956455, "epoch": 5.413669064748201, "step": 6020 }, { "epoch": 5.413669064748201, "step": 6020, "torque_loss": 0.1702655553817749 }, { "epoch": 5.422661870503597, "grad_norm": 0.3096064031124115, "learning_rate": 9.864423831800712e-05, "loss": 0.0356, "step": 6030 }, { "action_loss": 0.0215452890843153, "epoch": 5.422661870503597, "step": 6030 }, { "epoch": 5.422661870503597, "step": 6030, "torque_loss": 0.2471461147069931 }, { "epoch": 5.431654676258993, "grad_norm": 0.23040026426315308, "learning_rate": 9.863657904660191e-05, "loss": 0.0415, "step": 6040 }, { "action_loss": 0.006796871777623892, "epoch": 5.431654676258993, "step": 6040 }, { "epoch": 5.431654676258993, "step": 6040, "torque_loss": 0.16031329333782196 }, { "epoch": 5.440647482014389, "grad_norm": 0.35656827688217163, "learning_rate": 9.862889849990462e-05, "loss": 0.0351, "step": 6050 }, { "action_loss": 0.00574294151738286, "epoch": 5.440647482014389, "step": 6050 }, { "epoch": 5.440647482014389, "step": 6050, "torque_loss": 0.0925755724310875 }, { "epoch": 5.4496402877697845, "grad_norm": 0.23308052122592926, "learning_rate": 9.8621196681275e-05, "loss": 0.032, "step": 6060 }, { "action_loss": 0.02592550776898861, "epoch": 5.4496402877697845, "step": 6060 }, { "epoch": 5.4496402877697845, "step": 6060, "torque_loss": 0.22472620010375977 }, { "epoch": 5.4586330935251794, "grad_norm": 0.1412021815776825, "learning_rate": 9.861347359408211e-05, "loss": 0.0325, "step": 6070 }, { "action_loss": 0.01827119290828705, "epoch": 5.4586330935251794, "step": 6070 }, { "epoch": 5.4586330935251794, "step": 6070, "torque_loss": 0.24097633361816406 }, { "epoch": 5.467625899280575, "grad_norm": 0.29278385639190674, "learning_rate": 9.860572924170426e-05, "loss": 0.03, "step": 6080 }, { "action_loss": 0.009143897332251072, "epoch": 5.467625899280575, "step": 6080 }, { "epoch": 5.467625899280575, "step": 6080, "torque_loss": 0.1588757038116455 }, { "epoch": 5.476618705035971, "grad_norm": 0.2924402356147766, "learning_rate": 9.85979636275291e-05, "loss": 0.0358, "step": 6090 }, { "action_loss": 0.01597907580435276, "epoch": 5.476618705035971, "step": 6090 }, { "epoch": 5.476618705035971, "step": 6090, "torque_loss": 0.19288839399814606 }, { "epoch": 5.485611510791367, "grad_norm": 0.18465612828731537, "learning_rate": 9.859017675495357e-05, "loss": 0.0358, "step": 6100 }, { "action_loss": 0.042671144008636475, "epoch": 5.485611510791367, "step": 6100 }, { "epoch": 5.485611510791367, "step": 6100, "torque_loss": 0.3236689865589142 }, { "epoch": 5.494604316546763, "grad_norm": 0.22965620458126068, "learning_rate": 9.858236862738392e-05, "loss": 0.042, "step": 6110 }, { "action_loss": 0.005922236945480108, "epoch": 5.494604316546763, "step": 6110 }, { "epoch": 5.494604316546763, "step": 6110, "torque_loss": 0.10983272641897202 }, { "epoch": 5.503597122302159, "grad_norm": 0.20989620685577393, "learning_rate": 9.85745392482357e-05, "loss": 0.0336, "step": 6120 }, { "action_loss": 0.02849365957081318, "epoch": 5.503597122302159, "step": 6120 }, { "epoch": 5.503597122302159, "step": 6120, "torque_loss": 0.3308640420436859 }, { "epoch": 5.512589928057554, "grad_norm": 0.21660150587558746, "learning_rate": 9.856668862093372e-05, "loss": 0.0428, "step": 6130 }, { "action_loss": 0.02609080635011196, "epoch": 5.512589928057554, "step": 6130 }, { "epoch": 5.512589928057554, "step": 6130, "torque_loss": 0.282927006483078 }, { "epoch": 5.5215827338129495, "grad_norm": 0.3830641806125641, "learning_rate": 9.855881674891214e-05, "loss": 0.0375, "step": 6140 }, { "action_loss": 0.025675924494862556, "epoch": 5.5215827338129495, "step": 6140 }, { "epoch": 5.5215827338129495, "step": 6140, "torque_loss": 0.2015385627746582 }, { "epoch": 5.530575539568345, "grad_norm": 0.189471036195755, "learning_rate": 9.855092363561437e-05, "loss": 0.0325, "step": 6150 }, { "action_loss": 0.022619498893618584, "epoch": 5.530575539568345, "step": 6150 }, { "epoch": 5.530575539568345, "step": 6150, "torque_loss": 0.2400531768798828 }, { "epoch": 5.539568345323741, "grad_norm": 0.2366422563791275, "learning_rate": 9.854300928449312e-05, "loss": 0.0387, "step": 6160 }, { "action_loss": 0.012370515614748001, "epoch": 5.539568345323741, "step": 6160 }, { "epoch": 5.539568345323741, "step": 6160, "torque_loss": 0.160190150141716 }, { "epoch": 5.548561151079137, "grad_norm": 0.17289139330387115, "learning_rate": 9.85350736990104e-05, "loss": 0.0345, "step": 6170 }, { "action_loss": 0.010832284577190876, "epoch": 5.548561151079137, "step": 6170 }, { "epoch": 5.548561151079137, "step": 6170, "torque_loss": 0.1754484921693802 }, { "epoch": 5.557553956834532, "grad_norm": 0.20715314149856567, "learning_rate": 9.852711688263752e-05, "loss": 0.0389, "step": 6180 }, { "action_loss": 0.013934895396232605, "epoch": 5.557553956834532, "step": 6180 }, { "epoch": 5.557553956834532, "step": 6180, "torque_loss": 0.19921953976154327 }, { "epoch": 5.566546762589928, "grad_norm": 0.3053532540798187, "learning_rate": 9.851913883885503e-05, "loss": 0.034, "step": 6190 }, { "action_loss": 0.008863204158842564, "epoch": 5.566546762589928, "step": 6190 }, { "epoch": 5.566546762589928, "step": 6190, "torque_loss": 0.2633378207683563 }, { "epoch": 5.575539568345324, "grad_norm": 0.22964707016944885, "learning_rate": 9.851113957115283e-05, "loss": 0.0416, "step": 6200 }, { "action_loss": 0.03379877656698227, "epoch": 5.575539568345324, "step": 6200 }, { "epoch": 5.575539568345324, "step": 6200, "torque_loss": 0.31093212962150574 }, { "epoch": 5.58453237410072, "grad_norm": 0.28499841690063477, "learning_rate": 9.850311908303005e-05, "loss": 0.0391, "step": 6210 }, { "action_loss": 0.013706929050385952, "epoch": 5.58453237410072, "step": 6210 }, { "epoch": 5.58453237410072, "step": 6210, "torque_loss": 0.19463665783405304 }, { "epoch": 5.593525179856115, "grad_norm": 0.2553088963031769, "learning_rate": 9.849507737799514e-05, "loss": 0.0449, "step": 6220 }, { "action_loss": 0.017022212967276573, "epoch": 5.593525179856115, "step": 6220 }, { "epoch": 5.593525179856115, "step": 6220, "torque_loss": 0.2042223960161209 }, { "epoch": 5.602517985611511, "grad_norm": 0.2357473075389862, "learning_rate": 9.84870144595658e-05, "loss": 0.0362, "step": 6230 }, { "action_loss": 0.02806907333433628, "epoch": 5.602517985611511, "step": 6230 }, { "epoch": 5.602517985611511, "step": 6230, "torque_loss": 0.26479998230934143 }, { "epoch": 5.611510791366906, "grad_norm": 0.2807994484901428, "learning_rate": 9.847893033126903e-05, "loss": 0.0325, "step": 6240 }, { "action_loss": 0.01606607623398304, "epoch": 5.611510791366906, "step": 6240 }, { "epoch": 5.611510791366906, "step": 6240, "torque_loss": 0.1594131737947464 }, { "epoch": 5.620503597122302, "grad_norm": 0.2125028371810913, "learning_rate": 9.847082499664111e-05, "loss": 0.0339, "step": 6250 }, { "action_loss": 0.009366370737552643, "epoch": 5.620503597122302, "step": 6250 }, { "epoch": 5.620503597122302, "step": 6250, "torque_loss": 0.13056109845638275 }, { "epoch": 5.629496402877698, "grad_norm": 0.2850125730037689, "learning_rate": 9.846269845922758e-05, "loss": 0.0409, "step": 6260 }, { "action_loss": 0.017916565760970116, "epoch": 5.629496402877698, "step": 6260 }, { "epoch": 5.629496402877698, "step": 6260, "torque_loss": 0.21308885514736176 }, { "epoch": 5.638489208633094, "grad_norm": 0.26307693123817444, "learning_rate": 9.845455072258326e-05, "loss": 0.0387, "step": 6270 }, { "action_loss": 0.015538153238594532, "epoch": 5.638489208633094, "step": 6270 }, { "epoch": 5.638489208633094, "step": 6270, "torque_loss": 0.29770758748054504 }, { "epoch": 5.647482014388489, "grad_norm": 0.2562432289123535, "learning_rate": 9.844638179027226e-05, "loss": 0.0356, "step": 6280 }, { "action_loss": 0.02134483866393566, "epoch": 5.647482014388489, "step": 6280 }, { "epoch": 5.647482014388489, "step": 6280, "torque_loss": 0.20977599918842316 }, { "epoch": 5.656474820143885, "grad_norm": 0.3038412928581238, "learning_rate": 9.843819166586795e-05, "loss": 0.0356, "step": 6290 }, { "action_loss": 0.019229769706726074, "epoch": 5.656474820143885, "step": 6290 }, { "epoch": 5.656474820143885, "step": 6290, "torque_loss": 0.19579149782657623 }, { "epoch": 5.66546762589928, "grad_norm": 0.37740108370780945, "learning_rate": 9.842998035295294e-05, "loss": 0.0377, "step": 6300 }, { "action_loss": 0.016417426988482475, "epoch": 5.66546762589928, "step": 6300 }, { "epoch": 5.66546762589928, "step": 6300, "torque_loss": 0.26300105452537537 }, { "epoch": 5.674460431654676, "grad_norm": 0.2538357973098755, "learning_rate": 9.842174785511918e-05, "loss": 0.0337, "step": 6310 }, { "action_loss": 0.01202788669615984, "epoch": 5.674460431654676, "step": 6310 }, { "epoch": 5.674460431654676, "step": 6310, "torque_loss": 0.1901797205209732 }, { "epoch": 5.683453237410072, "grad_norm": 0.31911417841911316, "learning_rate": 9.841349417596779e-05, "loss": 0.0389, "step": 6320 }, { "action_loss": 0.01173249538987875, "epoch": 5.683453237410072, "step": 6320 }, { "epoch": 5.683453237410072, "step": 6320, "torque_loss": 0.2098926454782486 }, { "epoch": 5.692446043165468, "grad_norm": 0.18730008602142334, "learning_rate": 9.840521931910926e-05, "loss": 0.0348, "step": 6330 }, { "action_loss": 0.01372226607054472, "epoch": 5.692446043165468, "step": 6330 }, { "epoch": 5.692446043165468, "step": 6330, "torque_loss": 0.20013739168643951 }, { "epoch": 5.701438848920863, "grad_norm": 0.19677115976810455, "learning_rate": 9.839692328816327e-05, "loss": 0.0368, "step": 6340 }, { "action_loss": 0.02195836789906025, "epoch": 5.701438848920863, "step": 6340 }, { "epoch": 5.701438848920863, "step": 6340, "torque_loss": 0.2360628992319107 }, { "epoch": 5.710431654676259, "grad_norm": 0.22019916772842407, "learning_rate": 9.838860608675879e-05, "loss": 0.0339, "step": 6350 }, { "action_loss": 0.012999337166547775, "epoch": 5.710431654676259, "step": 6350 }, { "epoch": 5.710431654676259, "step": 6350, "torque_loss": 0.1910635381937027 }, { "epoch": 5.719424460431655, "grad_norm": 0.19333107769489288, "learning_rate": 9.838026771853406e-05, "loss": 0.0308, "step": 6360 }, { "action_loss": 0.009762872941792011, "epoch": 5.719424460431655, "step": 6360 }, { "epoch": 5.719424460431655, "step": 6360, "torque_loss": 0.13476906716823578 }, { "epoch": 5.7284172661870505, "grad_norm": 0.19556564092636108, "learning_rate": 9.837190818713655e-05, "loss": 0.0347, "step": 6370 }, { "action_loss": 0.016138315200805664, "epoch": 5.7284172661870505, "step": 6370 }, { "epoch": 5.7284172661870505, "step": 6370, "torque_loss": 0.20072710514068604 }, { "epoch": 5.737410071942446, "grad_norm": 0.3013358414173126, "learning_rate": 9.836352749622299e-05, "loss": 0.0362, "step": 6380 }, { "action_loss": 0.01916511170566082, "epoch": 5.737410071942446, "step": 6380 }, { "epoch": 5.737410071942446, "step": 6380, "torque_loss": 0.23563013970851898 }, { "epoch": 5.746402877697841, "grad_norm": 0.3764853775501251, "learning_rate": 9.835512564945941e-05, "loss": 0.0385, "step": 6390 }, { "action_loss": 0.03921766206622124, "epoch": 5.746402877697841, "step": 6390 }, { "epoch": 5.746402877697841, "step": 6390, "torque_loss": 0.17733311653137207 }, { "epoch": 5.755395683453237, "grad_norm": 0.25252747535705566, "learning_rate": 9.834670265052104e-05, "loss": 0.0391, "step": 6400 }, { "action_loss": 0.013442532159388065, "epoch": 5.755395683453237, "step": 6400 }, { "epoch": 5.755395683453237, "step": 6400, "torque_loss": 0.1677142232656479 }, { "epoch": 5.764388489208633, "grad_norm": 0.28526580333709717, "learning_rate": 9.833825850309239e-05, "loss": 0.0369, "step": 6410 }, { "action_loss": 0.025135336443781853, "epoch": 5.764388489208633, "step": 6410 }, { "epoch": 5.764388489208633, "step": 6410, "torque_loss": 0.2571388781070709 }, { "epoch": 5.773381294964029, "grad_norm": 0.22975799441337585, "learning_rate": 9.832979321086723e-05, "loss": 0.0342, "step": 6420 }, { "action_loss": 0.0070021809078752995, "epoch": 5.773381294964029, "step": 6420 }, { "epoch": 5.773381294964029, "step": 6420, "torque_loss": 0.1768457293510437 }, { "epoch": 5.782374100719425, "grad_norm": 0.2983244061470032, "learning_rate": 9.832130677754854e-05, "loss": 0.0335, "step": 6430 }, { "action_loss": 0.008453206159174442, "epoch": 5.782374100719425, "step": 6430 }, { "epoch": 5.782374100719425, "step": 6430, "torque_loss": 0.1400304138660431 }, { "epoch": 5.7913669064748206, "grad_norm": 0.2520173192024231, "learning_rate": 9.831279920684861e-05, "loss": 0.0302, "step": 6440 }, { "action_loss": 0.00663608917966485, "epoch": 5.7913669064748206, "step": 6440 }, { "epoch": 5.7913669064748206, "step": 6440, "torque_loss": 0.1341293901205063 }, { "epoch": 5.8003597122302155, "grad_norm": 0.38055849075317383, "learning_rate": 9.830427050248891e-05, "loss": 0.0399, "step": 6450 }, { "action_loss": 0.012866075150668621, "epoch": 5.8003597122302155, "step": 6450 }, { "epoch": 5.8003597122302155, "step": 6450, "torque_loss": 0.17372600734233856 }, { "epoch": 5.809352517985611, "grad_norm": 0.20285749435424805, "learning_rate": 9.829572066820021e-05, "loss": 0.0288, "step": 6460 }, { "action_loss": 0.011196225881576538, "epoch": 5.809352517985611, "step": 6460 }, { "epoch": 5.809352517985611, "step": 6460, "torque_loss": 0.15682204067707062 }, { "epoch": 5.818345323741007, "grad_norm": 0.2434333711862564, "learning_rate": 9.828714970772247e-05, "loss": 0.0362, "step": 6470 }, { "action_loss": 0.01235190499573946, "epoch": 5.818345323741007, "step": 6470 }, { "epoch": 5.818345323741007, "step": 6470, "torque_loss": 0.16389800608158112 }, { "epoch": 5.827338129496403, "grad_norm": 0.2407158464193344, "learning_rate": 9.827855762480493e-05, "loss": 0.0271, "step": 6480 }, { "action_loss": 0.01089573185890913, "epoch": 5.827338129496403, "step": 6480 }, { "epoch": 5.827338129496403, "step": 6480, "torque_loss": 0.16889095306396484 }, { "epoch": 5.836330935251799, "grad_norm": 0.24765080213546753, "learning_rate": 9.826994442320608e-05, "loss": 0.0293, "step": 6490 }, { "action_loss": 0.017446845769882202, "epoch": 5.836330935251799, "step": 6490 }, { "epoch": 5.836330935251799, "step": 6490, "torque_loss": 0.18721015751361847 }, { "epoch": 5.845323741007194, "grad_norm": 0.1934436857700348, "learning_rate": 9.82613101066936e-05, "loss": 0.0302, "step": 6500 }, { "action_loss": 0.021243492141366005, "epoch": 5.845323741007194, "step": 6500 }, { "epoch": 5.845323741007194, "step": 6500, "torque_loss": 0.22437231242656708 }, { "epoch": 5.85431654676259, "grad_norm": 0.3296975791454315, "learning_rate": 9.825265467904445e-05, "loss": 0.0424, "step": 6510 }, { "action_loss": 0.03134140372276306, "epoch": 5.85431654676259, "step": 6510 }, { "epoch": 5.85431654676259, "step": 6510, "torque_loss": 0.26659002900123596 }, { "epoch": 5.863309352517986, "grad_norm": 0.3010760247707367, "learning_rate": 9.82439781440448e-05, "loss": 0.0374, "step": 6520 }, { "action_loss": 0.012870613485574722, "epoch": 5.863309352517986, "step": 6520 }, { "epoch": 5.863309352517986, "step": 6520, "torque_loss": 0.24101369082927704 }, { "epoch": 5.872302158273381, "grad_norm": 0.2284010946750641, "learning_rate": 9.823528050549006e-05, "loss": 0.0316, "step": 6530 }, { "action_loss": 0.012109239585697651, "epoch": 5.872302158273381, "step": 6530 }, { "epoch": 5.872302158273381, "step": 6530, "torque_loss": 0.17640036344528198 }, { "epoch": 5.881294964028777, "grad_norm": 0.24685916304588318, "learning_rate": 9.822656176718487e-05, "loss": 0.0385, "step": 6540 }, { "action_loss": 0.011038929224014282, "epoch": 5.881294964028777, "step": 6540 }, { "epoch": 5.881294964028777, "step": 6540, "torque_loss": 0.13955003023147583 }, { "epoch": 5.890287769784173, "grad_norm": 0.19282841682434082, "learning_rate": 9.82178219329431e-05, "loss": 0.0357, "step": 6550 }, { "action_loss": 0.011592199094593525, "epoch": 5.890287769784173, "step": 6550 }, { "epoch": 5.890287769784173, "step": 6550, "torque_loss": 0.15974178910255432 }, { "epoch": 5.899280575539568, "grad_norm": 0.15481148660182953, "learning_rate": 9.820906100658789e-05, "loss": 0.0282, "step": 6560 }, { "action_loss": 0.025478770956397057, "epoch": 5.899280575539568, "step": 6560 }, { "epoch": 5.899280575539568, "step": 6560, "torque_loss": 0.21129103004932404 }, { "epoch": 5.908273381294964, "grad_norm": 0.2365194708108902, "learning_rate": 9.82002789919515e-05, "loss": 0.0392, "step": 6570 }, { "action_loss": 0.010436092503368855, "epoch": 5.908273381294964, "step": 6570 }, { "epoch": 5.908273381294964, "step": 6570, "torque_loss": 0.21714550256729126 }, { "epoch": 5.91726618705036, "grad_norm": 0.25832316279411316, "learning_rate": 9.819147589287554e-05, "loss": 0.0333, "step": 6580 }, { "action_loss": 0.016511736437678337, "epoch": 5.91726618705036, "step": 6580 }, { "epoch": 5.91726618705036, "step": 6580, "torque_loss": 0.15938712656497955 }, { "epoch": 5.926258992805756, "grad_norm": 0.16358362138271332, "learning_rate": 9.818265171321074e-05, "loss": 0.0318, "step": 6590 }, { "action_loss": 0.010049316100776196, "epoch": 5.926258992805756, "step": 6590 }, { "epoch": 5.926258992805756, "step": 6590, "torque_loss": 0.189083531498909 }, { "epoch": 5.935251798561151, "grad_norm": 0.26976731419563293, "learning_rate": 9.817380645681714e-05, "loss": 0.0309, "step": 6600 }, { "action_loss": 0.005697149783372879, "epoch": 5.935251798561151, "step": 6600 }, { "epoch": 5.935251798561151, "step": 6600, "torque_loss": 0.12493282556533813 }, { "epoch": 5.944244604316546, "grad_norm": 0.20806150138378143, "learning_rate": 9.816494012756392e-05, "loss": 0.0264, "step": 6610 }, { "action_loss": 0.014299802482128143, "epoch": 5.944244604316546, "step": 6610 }, { "epoch": 5.944244604316546, "step": 6610, "torque_loss": 0.25041821599006653 }, { "epoch": 5.953237410071942, "grad_norm": 0.24630863964557648, "learning_rate": 9.815605272932953e-05, "loss": 0.0347, "step": 6620 }, { "action_loss": 0.01078491285443306, "epoch": 5.953237410071942, "step": 6620 }, { "epoch": 5.953237410071942, "step": 6620, "torque_loss": 0.19720041751861572 }, { "epoch": 5.962230215827338, "grad_norm": 0.23373550176620483, "learning_rate": 9.814714426600162e-05, "loss": 0.0339, "step": 6630 }, { "action_loss": 0.023071570321917534, "epoch": 5.962230215827338, "step": 6630 }, { "epoch": 5.962230215827338, "step": 6630, "torque_loss": 0.2802812159061432 }, { "epoch": 5.971223021582734, "grad_norm": 0.2173418402671814, "learning_rate": 9.813821474147704e-05, "loss": 0.0317, "step": 6640 }, { "action_loss": 0.006887845229357481, "epoch": 5.971223021582734, "step": 6640 }, { "epoch": 5.971223021582734, "step": 6640, "torque_loss": 0.12592139840126038 }, { "epoch": 5.98021582733813, "grad_norm": 0.31486061215400696, "learning_rate": 9.812926415966189e-05, "loss": 0.0303, "step": 6650 }, { "action_loss": 0.01891106180846691, "epoch": 5.98021582733813, "step": 6650 }, { "epoch": 5.98021582733813, "step": 6650, "torque_loss": 0.1875094324350357 }, { "epoch": 5.989208633093525, "grad_norm": 0.2152397483587265, "learning_rate": 9.812029252447143e-05, "loss": 0.0432, "step": 6660 }, { "action_loss": 0.020827248692512512, "epoch": 5.989208633093525, "step": 6660 }, { "epoch": 5.989208633093525, "step": 6660, "torque_loss": 0.3003087341785431 }, { "epoch": 5.998201438848921, "grad_norm": 0.31491222977638245, "learning_rate": 9.811129983983018e-05, "loss": 0.036, "step": 6670 }, { "action_loss": 0.013580936007201672, "epoch": 5.998201438848921, "step": 6670 }, { "epoch": 5.998201438848921, "step": 6670, "torque_loss": 0.15839630365371704 }, { "epoch": 6.0071942446043165, "grad_norm": 0.152921661734581, "learning_rate": 9.810228610967185e-05, "loss": 0.0306, "step": 6680 }, { "action_loss": 0.022246167063713074, "epoch": 6.0071942446043165, "step": 6680 }, { "epoch": 6.0071942446043165, "step": 6680, "torque_loss": 0.2334902137517929 }, { "epoch": 6.016187050359712, "grad_norm": 0.2377561777830124, "learning_rate": 9.809325133793934e-05, "loss": 0.0334, "step": 6690 }, { "action_loss": 0.018999189138412476, "epoch": 6.016187050359712, "step": 6690 }, { "epoch": 6.016187050359712, "step": 6690, "torque_loss": 0.2168780416250229 }, { "epoch": 6.025179856115108, "grad_norm": 0.21448127925395966, "learning_rate": 9.808419552858477e-05, "loss": 0.0391, "step": 6700 }, { "action_loss": 0.013624235987663269, "epoch": 6.025179856115108, "step": 6700 }, { "epoch": 6.025179856115108, "step": 6700, "torque_loss": 0.17065413296222687 }, { "epoch": 6.034172661870503, "grad_norm": 0.2016364485025406, "learning_rate": 9.807511868556944e-05, "loss": 0.0349, "step": 6710 }, { "action_loss": 0.00607558386400342, "epoch": 6.034172661870503, "step": 6710 }, { "epoch": 6.034172661870503, "step": 6710, "torque_loss": 0.1676161289215088 }, { "epoch": 6.043165467625899, "grad_norm": 0.24924781918525696, "learning_rate": 9.806602081286392e-05, "loss": 0.0318, "step": 6720 }, { "action_loss": 0.010024218820035458, "epoch": 6.043165467625899, "step": 6720 }, { "epoch": 6.043165467625899, "step": 6720, "torque_loss": 0.17806710302829742 }, { "epoch": 6.052158273381295, "grad_norm": 0.3049251437187195, "learning_rate": 9.805690191444785e-05, "loss": 0.0459, "step": 6730 }, { "action_loss": 0.01185252983123064, "epoch": 6.052158273381295, "step": 6730 }, { "epoch": 6.052158273381295, "step": 6730, "torque_loss": 0.2454712837934494 }, { "epoch": 6.061151079136691, "grad_norm": 0.24671323597431183, "learning_rate": 9.804776199431022e-05, "loss": 0.0404, "step": 6740 }, { "action_loss": 0.03595711663365364, "epoch": 6.061151079136691, "step": 6740 }, { "epoch": 6.061151079136691, "step": 6740, "torque_loss": 0.3225799798965454 }, { "epoch": 6.070143884892087, "grad_norm": 0.308902770280838, "learning_rate": 9.80386010564491e-05, "loss": 0.0336, "step": 6750 }, { "action_loss": 0.014053617604076862, "epoch": 6.070143884892087, "step": 6750 }, { "epoch": 6.070143884892087, "step": 6750, "torque_loss": 0.19354188442230225 }, { "epoch": 6.079136690647482, "grad_norm": 0.26206347346305847, "learning_rate": 9.802941910487183e-05, "loss": 0.0401, "step": 6760 }, { "action_loss": 0.011641654185950756, "epoch": 6.079136690647482, "step": 6760 }, { "epoch": 6.079136690647482, "step": 6760, "torque_loss": 0.11774853616952896 }, { "epoch": 6.088129496402877, "grad_norm": 0.25050726532936096, "learning_rate": 9.802021614359487e-05, "loss": 0.0396, "step": 6770 }, { "action_loss": 0.011354968883097172, "epoch": 6.088129496402877, "step": 6770 }, { "epoch": 6.088129496402877, "step": 6770, "torque_loss": 0.17078645527362823 }, { "epoch": 6.097122302158273, "grad_norm": 0.31096312403678894, "learning_rate": 9.801099217664394e-05, "loss": 0.0347, "step": 6780 }, { "action_loss": 0.024810845032334328, "epoch": 6.097122302158273, "step": 6780 }, { "epoch": 6.097122302158273, "step": 6780, "torque_loss": 0.19471339881420135 }, { "epoch": 6.106115107913669, "grad_norm": 0.2786165475845337, "learning_rate": 9.800174720805388e-05, "loss": 0.0355, "step": 6790 }, { "action_loss": 0.022046895697712898, "epoch": 6.106115107913669, "step": 6790 }, { "epoch": 6.106115107913669, "step": 6790, "torque_loss": 0.2254621982574463 }, { "epoch": 6.115107913669065, "grad_norm": 0.1707524061203003, "learning_rate": 9.799248124186878e-05, "loss": 0.0436, "step": 6800 }, { "action_loss": 0.022634213790297508, "epoch": 6.115107913669065, "step": 6800 }, { "epoch": 6.115107913669065, "step": 6800, "torque_loss": 0.19380205869674683 }, { "epoch": 6.124100719424461, "grad_norm": 0.3239249289035797, "learning_rate": 9.798319428214187e-05, "loss": 0.0399, "step": 6810 }, { "action_loss": 0.010940228588879108, "epoch": 6.124100719424461, "step": 6810 }, { "epoch": 6.124100719424461, "step": 6810, "torque_loss": 0.18419568240642548 }, { "epoch": 6.133093525179856, "grad_norm": 0.2275323122739792, "learning_rate": 9.79738863329356e-05, "loss": 0.0327, "step": 6820 }, { "action_loss": 0.01217444147914648, "epoch": 6.133093525179856, "step": 6820 }, { "epoch": 6.133093525179856, "step": 6820, "torque_loss": 0.12891501188278198 }, { "epoch": 6.142086330935252, "grad_norm": 0.2781737148761749, "learning_rate": 9.79645573983216e-05, "loss": 0.0415, "step": 6830 }, { "action_loss": 0.007080879062414169, "epoch": 6.142086330935252, "step": 6830 }, { "epoch": 6.142086330935252, "step": 6830, "torque_loss": 0.11797899007797241 }, { "epoch": 6.151079136690647, "grad_norm": 0.30807051062583923, "learning_rate": 9.795520748238062e-05, "loss": 0.0328, "step": 6840 }, { "action_loss": 0.010175878182053566, "epoch": 6.151079136690647, "step": 6840 }, { "epoch": 6.151079136690647, "step": 6840, "torque_loss": 0.20219020545482635 }, { "epoch": 6.160071942446043, "grad_norm": 0.2237953096628189, "learning_rate": 9.794583658920264e-05, "loss": 0.035, "step": 6850 }, { "action_loss": 0.005680173635482788, "epoch": 6.160071942446043, "step": 6850 }, { "epoch": 6.160071942446043, "step": 6850, "torque_loss": 0.144404798746109 }, { "epoch": 6.169064748201439, "grad_norm": 0.2527371942996979, "learning_rate": 9.793644472288683e-05, "loss": 0.0311, "step": 6860 }, { "action_loss": 0.011660064570605755, "epoch": 6.169064748201439, "step": 6860 }, { "epoch": 6.169064748201439, "step": 6860, "torque_loss": 0.22680704295635223 }, { "epoch": 6.178057553956835, "grad_norm": 0.2709013521671295, "learning_rate": 9.79270318875415e-05, "loss": 0.0298, "step": 6870 }, { "action_loss": 0.014253144152462482, "epoch": 6.178057553956835, "step": 6870 }, { "epoch": 6.178057553956835, "step": 6870, "torque_loss": 0.1706206351518631 }, { "epoch": 6.18705035971223, "grad_norm": 0.2470679134130478, "learning_rate": 9.791759808728416e-05, "loss": 0.0423, "step": 6880 }, { "action_loss": 0.01840740628540516, "epoch": 6.18705035971223, "step": 6880 }, { "epoch": 6.18705035971223, "step": 6880, "torque_loss": 0.24777770042419434 }, { "epoch": 6.196043165467626, "grad_norm": 0.1315324753522873, "learning_rate": 9.790814332624143e-05, "loss": 0.0317, "step": 6890 }, { "action_loss": 0.0186751838773489, "epoch": 6.196043165467626, "step": 6890 }, { "epoch": 6.196043165467626, "step": 6890, "torque_loss": 0.2048245221376419 }, { "epoch": 6.205035971223022, "grad_norm": 0.32870444655418396, "learning_rate": 9.789866760854919e-05, "loss": 0.0339, "step": 6900 }, { "action_loss": 0.005981553345918655, "epoch": 6.205035971223022, "step": 6900 }, { "epoch": 6.205035971223022, "step": 6900, "torque_loss": 0.09379657357931137 }, { "epoch": 6.2140287769784175, "grad_norm": 0.2953180968761444, "learning_rate": 9.788917093835242e-05, "loss": 0.0364, "step": 6910 }, { "action_loss": 0.029477184638381004, "epoch": 6.2140287769784175, "step": 6910 }, { "epoch": 6.2140287769784175, "step": 6910, "torque_loss": 0.20334994792938232 }, { "epoch": 6.223021582733813, "grad_norm": 0.21550510823726654, "learning_rate": 9.787965331980528e-05, "loss": 0.0372, "step": 6920 }, { "action_loss": 0.011757652275264263, "epoch": 6.223021582733813, "step": 6920 }, { "epoch": 6.223021582733813, "step": 6920, "torque_loss": 0.13022732734680176 }, { "epoch": 6.232014388489208, "grad_norm": 0.2242659330368042, "learning_rate": 9.787011475707111e-05, "loss": 0.032, "step": 6930 }, { "action_loss": 0.01843089796602726, "epoch": 6.232014388489208, "step": 6930 }, { "epoch": 6.232014388489208, "step": 6930, "torque_loss": 0.21318864822387695 }, { "epoch": 6.241007194244604, "grad_norm": 0.17778171598911285, "learning_rate": 9.786055525432241e-05, "loss": 0.0315, "step": 6940 }, { "action_loss": 0.01030788104981184, "epoch": 6.241007194244604, "step": 6940 }, { "epoch": 6.241007194244604, "step": 6940, "torque_loss": 0.16531981527805328 }, { "epoch": 6.25, "grad_norm": 0.19120973348617554, "learning_rate": 9.78509748157408e-05, "loss": 0.0337, "step": 6950 }, { "action_loss": 0.019776834174990654, "epoch": 6.25, "step": 6950 }, { "epoch": 6.25, "step": 6950, "torque_loss": 0.2525666654109955 }, { "epoch": 6.258992805755396, "grad_norm": 0.26516351103782654, "learning_rate": 9.784137344551713e-05, "loss": 0.0361, "step": 6960 }, { "action_loss": 0.012833558954298496, "epoch": 6.258992805755396, "step": 6960 }, { "epoch": 6.258992805755396, "step": 6960, "torque_loss": 0.14385680854320526 }, { "epoch": 6.267985611510792, "grad_norm": 0.2052464783191681, "learning_rate": 9.783175114785134e-05, "loss": 0.0355, "step": 6970 }, { "action_loss": 0.013637754134833813, "epoch": 6.267985611510792, "step": 6970 }, { "epoch": 6.267985611510792, "step": 6970, "torque_loss": 0.16237320005893707 }, { "epoch": 6.276978417266187, "grad_norm": 0.19418570399284363, "learning_rate": 9.782210792695254e-05, "loss": 0.0347, "step": 6980 }, { "action_loss": 0.00835245568305254, "epoch": 6.276978417266187, "step": 6980 }, { "epoch": 6.276978417266187, "step": 6980, "torque_loss": 0.17030037939548492 }, { "epoch": 6.2859712230215825, "grad_norm": 0.3131643533706665, "learning_rate": 9.781244378703904e-05, "loss": 0.0466, "step": 6990 }, { "action_loss": 0.023105978965759277, "epoch": 6.2859712230215825, "step": 6990 }, { "epoch": 6.2859712230215825, "step": 6990, "torque_loss": 0.2551632225513458 }, { "epoch": 6.294964028776978, "grad_norm": 0.1972609907388687, "learning_rate": 9.780275873233824e-05, "loss": 0.0376, "step": 7000 }, { "action_loss": 0.012625806033611298, "epoch": 6.294964028776978, "step": 7000 }, { "epoch": 6.294964028776978, "step": 7000, "torque_loss": 0.17118032276630402 }, { "epoch": 6.303956834532374, "grad_norm": 0.22481662034988403, "learning_rate": 9.77930527670867e-05, "loss": 0.0369, "step": 7010 }, { "action_loss": 0.0165741965174675, "epoch": 6.303956834532374, "step": 7010 }, { "epoch": 6.303956834532374, "step": 7010, "torque_loss": 0.25457867980003357 }, { "epoch": 6.31294964028777, "grad_norm": 0.1925717443227768, "learning_rate": 9.778332589553016e-05, "loss": 0.0346, "step": 7020 }, { "action_loss": 0.006092716008424759, "epoch": 6.31294964028777, "step": 7020 }, { "epoch": 6.31294964028777, "step": 7020, "torque_loss": 0.21338050067424774 }, { "epoch": 6.321942446043165, "grad_norm": 0.26968276500701904, "learning_rate": 9.777357812192349e-05, "loss": 0.0344, "step": 7030 }, { "action_loss": 0.011580315418541431, "epoch": 6.321942446043165, "step": 7030 }, { "epoch": 6.321942446043165, "step": 7030, "torque_loss": 0.20355486869812012 }, { "epoch": 6.330935251798561, "grad_norm": 0.2048341929912567, "learning_rate": 9.776380945053066e-05, "loss": 0.0333, "step": 7040 }, { "action_loss": 0.01429398637264967, "epoch": 6.330935251798561, "step": 7040 }, { "epoch": 6.330935251798561, "step": 7040, "torque_loss": 0.24594277143478394 }, { "epoch": 6.339928057553957, "grad_norm": 0.22794359922409058, "learning_rate": 9.775401988562486e-05, "loss": 0.0331, "step": 7050 }, { "action_loss": 0.00692318519577384, "epoch": 6.339928057553957, "step": 7050 }, { "epoch": 6.339928057553957, "step": 7050, "torque_loss": 0.13860298693180084 }, { "epoch": 6.348920863309353, "grad_norm": 0.2114047110080719, "learning_rate": 9.774420943148835e-05, "loss": 0.0334, "step": 7060 }, { "action_loss": 0.020445117726922035, "epoch": 6.348920863309353, "step": 7060 }, { "epoch": 6.348920863309353, "step": 7060, "torque_loss": 0.17392690479755402 }, { "epoch": 6.357913669064748, "grad_norm": 0.22371390461921692, "learning_rate": 9.773437809241257e-05, "loss": 0.037, "step": 7070 }, { "action_loss": 0.005433314014226198, "epoch": 6.357913669064748, "step": 7070 }, { "epoch": 6.357913669064748, "step": 7070, "torque_loss": 0.10813897103071213 }, { "epoch": 6.366906474820144, "grad_norm": 0.34734076261520386, "learning_rate": 9.772452587269808e-05, "loss": 0.0402, "step": 7080 }, { "action_loss": 0.016093621030449867, "epoch": 6.366906474820144, "step": 7080 }, { "epoch": 6.366906474820144, "step": 7080, "torque_loss": 0.2080608457326889 }, { "epoch": 6.375899280575539, "grad_norm": 0.3264528512954712, "learning_rate": 9.771465277665457e-05, "loss": 0.038, "step": 7090 }, { "action_loss": 0.016190355643630028, "epoch": 6.375899280575539, "step": 7090 }, { "epoch": 6.375899280575539, "step": 7090, "torque_loss": 0.14148792624473572 }, { "epoch": 6.384892086330935, "grad_norm": 0.28365087509155273, "learning_rate": 9.770475880860089e-05, "loss": 0.0363, "step": 7100 }, { "action_loss": 0.01656288094818592, "epoch": 6.384892086330935, "step": 7100 }, { "epoch": 6.384892086330935, "step": 7100, "torque_loss": 0.20512966811656952 }, { "epoch": 6.393884892086331, "grad_norm": 0.1983071267604828, "learning_rate": 9.769484397286495e-05, "loss": 0.037, "step": 7110 }, { "action_loss": 0.017752448096871376, "epoch": 6.393884892086331, "step": 7110 }, { "epoch": 6.393884892086331, "step": 7110, "torque_loss": 0.20597811043262482 }, { "epoch": 6.402877697841727, "grad_norm": 0.32953304052352905, "learning_rate": 9.768490827378388e-05, "loss": 0.036, "step": 7120 }, { "action_loss": 0.020886242389678955, "epoch": 6.402877697841727, "step": 7120 }, { "epoch": 6.402877697841727, "step": 7120, "torque_loss": 0.23246753215789795 }, { "epoch": 6.411870503597123, "grad_norm": 0.23068207502365112, "learning_rate": 9.767495171570387e-05, "loss": 0.0364, "step": 7130 }, { "action_loss": 0.03366747498512268, "epoch": 6.411870503597123, "step": 7130 }, { "epoch": 6.411870503597123, "step": 7130, "torque_loss": 0.22627921402454376 }, { "epoch": 6.420863309352518, "grad_norm": 0.3193269670009613, "learning_rate": 9.766497430298027e-05, "loss": 0.0437, "step": 7140 }, { "action_loss": 0.015697361901402473, "epoch": 6.420863309352518, "step": 7140 }, { "epoch": 6.420863309352518, "step": 7140, "torque_loss": 0.13621215522289276 }, { "epoch": 6.429856115107913, "grad_norm": 0.43821853399276733, "learning_rate": 9.765497603997754e-05, "loss": 0.0346, "step": 7150 }, { "action_loss": 0.008440941572189331, "epoch": 6.429856115107913, "step": 7150 }, { "epoch": 6.429856115107913, "step": 7150, "torque_loss": 0.14730389416217804 }, { "epoch": 6.438848920863309, "grad_norm": 0.22015342116355896, "learning_rate": 9.764495693106926e-05, "loss": 0.0421, "step": 7160 }, { "action_loss": 0.012973722070455551, "epoch": 6.438848920863309, "step": 7160 }, { "epoch": 6.438848920863309, "step": 7160, "torque_loss": 0.1456461101770401 }, { "epoch": 6.447841726618705, "grad_norm": 0.338163286447525, "learning_rate": 9.76349169806381e-05, "loss": 0.0336, "step": 7170 }, { "action_loss": 0.017031658440828323, "epoch": 6.447841726618705, "step": 7170 }, { "epoch": 6.447841726618705, "step": 7170, "torque_loss": 0.1435295194387436 }, { "epoch": 6.456834532374101, "grad_norm": 0.274263858795166, "learning_rate": 9.762485619307591e-05, "loss": 0.0344, "step": 7180 }, { "action_loss": 0.02178359590470791, "epoch": 6.456834532374101, "step": 7180 }, { "epoch": 6.456834532374101, "step": 7180, "torque_loss": 0.21631459891796112 }, { "epoch": 6.465827338129497, "grad_norm": 0.262514591217041, "learning_rate": 9.76147745727836e-05, "loss": 0.035, "step": 7190 }, { "action_loss": 0.014337177388370037, "epoch": 6.465827338129497, "step": 7190 }, { "epoch": 6.465827338129497, "step": 7190, "torque_loss": 0.1725512593984604 }, { "epoch": 6.474820143884892, "grad_norm": 0.17837198078632355, "learning_rate": 9.760467212417124e-05, "loss": 0.0406, "step": 7200 }, { "action_loss": 0.020267128944396973, "epoch": 6.474820143884892, "step": 7200 }, { "epoch": 6.474820143884892, "step": 7200, "torque_loss": 0.19617600739002228 }, { "epoch": 6.483812949640288, "grad_norm": 0.2238658219575882, "learning_rate": 9.759454885165796e-05, "loss": 0.0367, "step": 7210 }, { "action_loss": 0.007339919451624155, "epoch": 6.483812949640288, "step": 7210 }, { "epoch": 6.483812949640288, "step": 7210, "torque_loss": 0.1701149344444275 }, { "epoch": 6.4928057553956835, "grad_norm": 0.23517997562885284, "learning_rate": 9.7584404759672e-05, "loss": 0.0368, "step": 7220 }, { "action_loss": 0.024365602061152458, "epoch": 6.4928057553956835, "step": 7220 }, { "epoch": 6.4928057553956835, "step": 7220, "torque_loss": 0.1599692553281784 }, { "epoch": 6.501798561151079, "grad_norm": 0.22002707421779633, "learning_rate": 9.75742398526508e-05, "loss": 0.0328, "step": 7230 }, { "action_loss": 0.012206639163196087, "epoch": 6.501798561151079, "step": 7230 }, { "epoch": 6.501798561151079, "step": 7230, "torque_loss": 0.1511208564043045 }, { "epoch": 6.510791366906475, "grad_norm": 0.2810378670692444, "learning_rate": 9.756405413504077e-05, "loss": 0.0337, "step": 7240 }, { "action_loss": 0.011043502949178219, "epoch": 6.510791366906475, "step": 7240 }, { "epoch": 6.510791366906475, "step": 7240, "torque_loss": 0.16482536494731903 }, { "epoch": 6.51978417266187, "grad_norm": 0.2201722413301468, "learning_rate": 9.755384761129752e-05, "loss": 0.0384, "step": 7250 }, { "action_loss": 0.017181163653731346, "epoch": 6.51978417266187, "step": 7250 }, { "epoch": 6.51978417266187, "step": 7250, "torque_loss": 0.1515629142522812 }, { "epoch": 6.528776978417266, "grad_norm": 0.2324623167514801, "learning_rate": 9.754362028588572e-05, "loss": 0.0318, "step": 7260 }, { "action_loss": 0.011309110559523106, "epoch": 6.528776978417266, "step": 7260 }, { "epoch": 6.528776978417266, "step": 7260, "torque_loss": 0.1453501135110855 }, { "epoch": 6.537769784172662, "grad_norm": 0.26380446553230286, "learning_rate": 9.753337216327917e-05, "loss": 0.0401, "step": 7270 }, { "action_loss": 0.005604044068604708, "epoch": 6.537769784172662, "step": 7270 }, { "epoch": 6.537769784172662, "step": 7270, "torque_loss": 0.15165197849273682 }, { "epoch": 6.546762589928058, "grad_norm": 0.24602115154266357, "learning_rate": 9.752310324796073e-05, "loss": 0.0331, "step": 7280 }, { "action_loss": 0.007842514663934708, "epoch": 6.546762589928058, "step": 7280 }, { "epoch": 6.546762589928058, "step": 7280, "torque_loss": 0.10499471426010132 }, { "epoch": 6.555755395683454, "grad_norm": 0.21746893227100372, "learning_rate": 9.751281354442235e-05, "loss": 0.0312, "step": 7290 }, { "action_loss": 0.05249454081058502, "epoch": 6.555755395683454, "step": 7290 }, { "epoch": 6.555755395683454, "step": 7290, "torque_loss": 0.3044299781322479 }, { "epoch": 6.564748201438849, "grad_norm": 0.1529739797115326, "learning_rate": 9.750250305716513e-05, "loss": 0.0399, "step": 7300 }, { "action_loss": 0.00989918876439333, "epoch": 6.564748201438849, "step": 7300 }, { "epoch": 6.564748201438849, "step": 7300, "torque_loss": 0.1533505767583847 }, { "epoch": 6.573741007194244, "grad_norm": 0.1763128638267517, "learning_rate": 9.749217179069923e-05, "loss": 0.0301, "step": 7310 }, { "action_loss": 0.008123058825731277, "epoch": 6.573741007194244, "step": 7310 }, { "epoch": 6.573741007194244, "step": 7310, "torque_loss": 0.13116313517093658 }, { "epoch": 6.58273381294964, "grad_norm": 0.40412548184394836, "learning_rate": 9.748181974954386e-05, "loss": 0.034, "step": 7320 }, { "action_loss": 0.007752410601824522, "epoch": 6.58273381294964, "step": 7320 }, { "epoch": 6.58273381294964, "step": 7320, "torque_loss": 0.15210823714733124 }, { "epoch": 6.591726618705036, "grad_norm": 0.2884897291660309, "learning_rate": 9.74714469382274e-05, "loss": 0.0347, "step": 7330 }, { "action_loss": 0.02771294116973877, "epoch": 6.591726618705036, "step": 7330 }, { "epoch": 6.591726618705036, "step": 7330, "torque_loss": 0.3314042389392853 }, { "epoch": 6.600719424460432, "grad_norm": 0.204076886177063, "learning_rate": 9.746105336128724e-05, "loss": 0.0358, "step": 7340 }, { "action_loss": 0.009010565467178822, "epoch": 6.600719424460432, "step": 7340 }, { "epoch": 6.600719424460432, "step": 7340, "torque_loss": 0.14603997766971588 }, { "epoch": 6.609712230215827, "grad_norm": 0.23956270515918732, "learning_rate": 9.745063902326988e-05, "loss": 0.0387, "step": 7350 }, { "action_loss": 0.015893803909420967, "epoch": 6.609712230215827, "step": 7350 }, { "epoch": 6.609712230215827, "step": 7350, "torque_loss": 0.1979053020477295 }, { "epoch": 6.618705035971223, "grad_norm": 0.19055277109146118, "learning_rate": 9.744020392873093e-05, "loss": 0.0395, "step": 7360 }, { "action_loss": 0.01996254362165928, "epoch": 6.618705035971223, "step": 7360 }, { "epoch": 6.618705035971223, "step": 7360, "torque_loss": 0.2446606159210205 }, { "epoch": 6.627697841726619, "grad_norm": 0.25372999906539917, "learning_rate": 9.742974808223504e-05, "loss": 0.0404, "step": 7370 }, { "action_loss": 0.006861125584691763, "epoch": 6.627697841726619, "step": 7370 }, { "epoch": 6.627697841726619, "step": 7370, "torque_loss": 0.15606223046779633 }, { "epoch": 6.636690647482014, "grad_norm": 0.32625240087509155, "learning_rate": 9.741927148835596e-05, "loss": 0.0394, "step": 7380 }, { "action_loss": 0.01455233246088028, "epoch": 6.636690647482014, "step": 7380 }, { "epoch": 6.636690647482014, "step": 7380, "torque_loss": 0.21166370809078217 }, { "epoch": 6.64568345323741, "grad_norm": 0.21436667442321777, "learning_rate": 9.740877415167648e-05, "loss": 0.0329, "step": 7390 }, { "action_loss": 0.0171328317373991, "epoch": 6.64568345323741, "step": 7390 }, { "epoch": 6.64568345323741, "step": 7390, "torque_loss": 0.17293782532215118 }, { "epoch": 6.654676258992806, "grad_norm": 0.24924977123737335, "learning_rate": 9.739825607678854e-05, "loss": 0.0331, "step": 7400 }, { "action_loss": 0.01512714009732008, "epoch": 6.654676258992806, "step": 7400 }, { "epoch": 6.654676258992806, "step": 7400, "torque_loss": 0.15224385261535645 }, { "epoch": 6.663669064748201, "grad_norm": 0.2044312059879303, "learning_rate": 9.738771726829308e-05, "loss": 0.0329, "step": 7410 }, { "action_loss": 0.019331006333231926, "epoch": 6.663669064748201, "step": 7410 }, { "epoch": 6.663669064748201, "step": 7410, "torque_loss": 0.2419968992471695 }, { "epoch": 6.672661870503597, "grad_norm": 0.236807182431221, "learning_rate": 9.737715773080011e-05, "loss": 0.0455, "step": 7420 }, { "action_loss": 0.007557944860309362, "epoch": 6.672661870503597, "step": 7420 }, { "epoch": 6.672661870503597, "step": 7420, "torque_loss": 0.14710232615470886 }, { "epoch": 6.681654676258993, "grad_norm": 0.24394632875919342, "learning_rate": 9.736657746892877e-05, "loss": 0.0431, "step": 7430 }, { "action_loss": 0.009445984847843647, "epoch": 6.681654676258993, "step": 7430 }, { "epoch": 6.681654676258993, "step": 7430, "torque_loss": 0.1976945847272873 }, { "epoch": 6.690647482014389, "grad_norm": 0.28178995847702026, "learning_rate": 9.735597648730718e-05, "loss": 0.0326, "step": 7440 }, { "action_loss": 0.011553645133972168, "epoch": 6.690647482014389, "step": 7440 }, { "epoch": 6.690647482014389, "step": 7440, "torque_loss": 0.14683173596858978 }, { "epoch": 6.6996402877697845, "grad_norm": 0.20389620959758759, "learning_rate": 9.734535479057262e-05, "loss": 0.0322, "step": 7450 }, { "action_loss": 0.013862833380699158, "epoch": 6.6996402877697845, "step": 7450 }, { "epoch": 6.6996402877697845, "step": 7450, "torque_loss": 0.1791970282793045 }, { "epoch": 6.7086330935251794, "grad_norm": 0.21621543169021606, "learning_rate": 9.733471238337136e-05, "loss": 0.0342, "step": 7460 }, { "action_loss": 0.008912877179682255, "epoch": 6.7086330935251794, "step": 7460 }, { "epoch": 6.7086330935251794, "step": 7460, "torque_loss": 0.1802537590265274 }, { "epoch": 6.717625899280575, "grad_norm": 0.20147642493247986, "learning_rate": 9.732404927035871e-05, "loss": 0.0352, "step": 7470 }, { "action_loss": 0.011099833995103836, "epoch": 6.717625899280575, "step": 7470 }, { "epoch": 6.717625899280575, "step": 7470, "torque_loss": 0.20700538158416748 }, { "epoch": 6.726618705035971, "grad_norm": 0.18419833481311798, "learning_rate": 9.731336545619915e-05, "loss": 0.039, "step": 7480 }, { "action_loss": 0.005099974572658539, "epoch": 6.726618705035971, "step": 7480 }, { "epoch": 6.726618705035971, "step": 7480, "torque_loss": 0.1244763657450676 }, { "epoch": 6.735611510791367, "grad_norm": 0.25352004170417786, "learning_rate": 9.73026609455661e-05, "loss": 0.0352, "step": 7490 }, { "action_loss": 0.010369685478508472, "epoch": 6.735611510791367, "step": 7490 }, { "epoch": 6.735611510791367, "step": 7490, "torque_loss": 0.1680436134338379 }, { "epoch": 6.744604316546763, "grad_norm": 0.28053006529808044, "learning_rate": 9.72919357431421e-05, "loss": 0.0351, "step": 7500 }, { "action_loss": 0.00642326520755887, "epoch": 6.744604316546763, "step": 7500 }, { "epoch": 6.744604316546763, "step": 7500, "torque_loss": 0.13427145779132843 }, { "epoch": 6.753597122302159, "grad_norm": 0.3128982186317444, "learning_rate": 9.72811898536187e-05, "loss": 0.0295, "step": 7510 }, { "action_loss": 0.010074051097035408, "epoch": 6.753597122302159, "step": 7510 }, { "epoch": 6.753597122302159, "step": 7510, "torque_loss": 0.19938921928405762 }, { "epoch": 6.762589928057554, "grad_norm": 0.2687797248363495, "learning_rate": 9.72704232816965e-05, "loss": 0.0424, "step": 7520 }, { "action_loss": 0.005761066917330027, "epoch": 6.762589928057554, "step": 7520 }, { "epoch": 6.762589928057554, "step": 7520, "torque_loss": 0.12792430818080902 }, { "epoch": 6.7715827338129495, "grad_norm": 0.24854443967342377, "learning_rate": 9.725963603208519e-05, "loss": 0.0437, "step": 7530 }, { "action_loss": 0.00909495074301958, "epoch": 6.7715827338129495, "step": 7530 }, { "epoch": 6.7715827338129495, "step": 7530, "torque_loss": 0.1428878754377365 }, { "epoch": 6.780575539568345, "grad_norm": 0.20811226963996887, "learning_rate": 9.724882810950348e-05, "loss": 0.0353, "step": 7540 }, { "action_loss": 0.02131967432796955, "epoch": 6.780575539568345, "step": 7540 }, { "epoch": 6.780575539568345, "step": 7540, "torque_loss": 0.267506867647171 }, { "epoch": 6.789568345323741, "grad_norm": 0.2780385911464691, "learning_rate": 9.723799951867913e-05, "loss": 0.0405, "step": 7550 }, { "action_loss": 0.010906483046710491, "epoch": 6.789568345323741, "step": 7550 }, { "epoch": 6.789568345323741, "step": 7550, "torque_loss": 0.20233629643917084 }, { "epoch": 6.798561151079137, "grad_norm": 0.2690713107585907, "learning_rate": 9.722715026434889e-05, "loss": 0.033, "step": 7560 }, { "action_loss": 0.01866484247148037, "epoch": 6.798561151079137, "step": 7560 }, { "epoch": 6.798561151079137, "step": 7560, "torque_loss": 0.2940963804721832 }, { "epoch": 6.807553956834532, "grad_norm": 0.3023234009742737, "learning_rate": 9.721628035125866e-05, "loss": 0.0279, "step": 7570 }, { "action_loss": 0.01079652551561594, "epoch": 6.807553956834532, "step": 7570 }, { "epoch": 6.807553956834532, "step": 7570, "torque_loss": 0.22203151881694794 }, { "epoch": 6.816546762589928, "grad_norm": 0.24089418351650238, "learning_rate": 9.720538978416325e-05, "loss": 0.0363, "step": 7580 }, { "action_loss": 0.00943953637033701, "epoch": 6.816546762589928, "step": 7580 }, { "epoch": 6.816546762589928, "step": 7580, "torque_loss": 0.2643302083015442 }, { "epoch": 6.825539568345324, "grad_norm": 0.29046183824539185, "learning_rate": 9.71944785678266e-05, "loss": 0.0377, "step": 7590 }, { "action_loss": 0.011076603084802628, "epoch": 6.825539568345324, "step": 7590 }, { "epoch": 6.825539568345324, "step": 7590, "torque_loss": 0.2480284720659256 }, { "epoch": 6.83453237410072, "grad_norm": 0.318910151720047, "learning_rate": 9.718354670702161e-05, "loss": 0.0359, "step": 7600 }, { "action_loss": 0.025784553959965706, "epoch": 6.83453237410072, "step": 7600 }, { "epoch": 6.83453237410072, "step": 7600, "torque_loss": 0.2004326730966568 }, { "epoch": 6.843525179856115, "grad_norm": 0.25673672556877136, "learning_rate": 9.717259420653029e-05, "loss": 0.0325, "step": 7610 }, { "action_loss": 0.009198442101478577, "epoch": 6.843525179856115, "step": 7610 }, { "epoch": 6.843525179856115, "step": 7610, "torque_loss": 0.18364739418029785 }, { "epoch": 6.852517985611511, "grad_norm": 0.1281818449497223, "learning_rate": 9.716162107114361e-05, "loss": 0.027, "step": 7620 }, { "action_loss": 0.018043439835309982, "epoch": 6.852517985611511, "step": 7620 }, { "epoch": 6.852517985611511, "step": 7620, "torque_loss": 0.18027164041996002 }, { "epoch": 6.861510791366906, "grad_norm": 0.2363055795431137, "learning_rate": 9.71506273056616e-05, "loss": 0.0426, "step": 7630 }, { "action_loss": 0.01393075380474329, "epoch": 6.861510791366906, "step": 7630 }, { "epoch": 6.861510791366906, "step": 7630, "torque_loss": 0.24413366615772247 }, { "epoch": 6.870503597122302, "grad_norm": 0.27332988381385803, "learning_rate": 9.713961291489331e-05, "loss": 0.0308, "step": 7640 }, { "action_loss": 0.01969669759273529, "epoch": 6.870503597122302, "step": 7640 }, { "epoch": 6.870503597122302, "step": 7640, "torque_loss": 0.20344926416873932 }, { "epoch": 6.879496402877698, "grad_norm": 0.39337897300720215, "learning_rate": 9.71285779036568e-05, "loss": 0.039, "step": 7650 }, { "action_loss": 0.015961002558469772, "epoch": 6.879496402877698, "step": 7650 }, { "epoch": 6.879496402877698, "step": 7650, "torque_loss": 0.2519373297691345 }, { "epoch": 6.888489208633094, "grad_norm": 0.2754145860671997, "learning_rate": 9.711752227677916e-05, "loss": 0.0358, "step": 7660 }, { "action_loss": 0.006699572782963514, "epoch": 6.888489208633094, "step": 7660 }, { "epoch": 6.888489208633094, "step": 7660, "torque_loss": 0.16727757453918457 }, { "epoch": 6.897482014388489, "grad_norm": 0.2699446976184845, "learning_rate": 9.710644603909649e-05, "loss": 0.0341, "step": 7670 }, { "action_loss": 0.013674709014594555, "epoch": 6.897482014388489, "step": 7670 }, { "epoch": 6.897482014388489, "step": 7670, "torque_loss": 0.22164447605609894 }, { "epoch": 6.906474820143885, "grad_norm": 0.18383492529392242, "learning_rate": 9.709534919545393e-05, "loss": 0.0362, "step": 7680 }, { "action_loss": 0.021071961149573326, "epoch": 6.906474820143885, "step": 7680 }, { "epoch": 6.906474820143885, "step": 7680, "torque_loss": 0.21320052444934845 }, { "epoch": 6.91546762589928, "grad_norm": 0.24798434972763062, "learning_rate": 9.708423175070563e-05, "loss": 0.0351, "step": 7690 }, { "action_loss": 0.007988282479345798, "epoch": 6.91546762589928, "step": 7690 }, { "epoch": 6.91546762589928, "step": 7690, "torque_loss": 0.13824856281280518 }, { "epoch": 6.924460431654676, "grad_norm": 0.2850552797317505, "learning_rate": 9.707309370971468e-05, "loss": 0.0317, "step": 7700 }, { "action_loss": 0.008167085237801075, "epoch": 6.924460431654676, "step": 7700 }, { "epoch": 6.924460431654676, "step": 7700, "torque_loss": 0.19973933696746826 }, { "epoch": 6.933453237410072, "grad_norm": 0.19419805705547333, "learning_rate": 9.70619350773533e-05, "loss": 0.0333, "step": 7710 }, { "action_loss": 0.01679009571671486, "epoch": 6.933453237410072, "step": 7710 }, { "epoch": 6.933453237410072, "step": 7710, "torque_loss": 0.31064900755882263 }, { "epoch": 6.942446043165468, "grad_norm": 0.22843217849731445, "learning_rate": 9.70507558585026e-05, "loss": 0.0413, "step": 7720 }, { "action_loss": 0.028345605358481407, "epoch": 6.942446043165468, "step": 7720 }, { "epoch": 6.942446043165468, "step": 7720, "torque_loss": 0.23012633621692657 }, { "epoch": 6.951438848920863, "grad_norm": 0.24229955673217773, "learning_rate": 9.703955605805281e-05, "loss": 0.0333, "step": 7730 }, { "action_loss": 0.026893222704529762, "epoch": 6.951438848920863, "step": 7730 }, { "epoch": 6.951438848920863, "step": 7730, "torque_loss": 0.16960346698760986 }, { "epoch": 6.960431654676259, "grad_norm": 0.22038985788822174, "learning_rate": 9.702833568090306e-05, "loss": 0.0328, "step": 7740 }, { "action_loss": 0.022112702950835228, "epoch": 6.960431654676259, "step": 7740 }, { "epoch": 6.960431654676259, "step": 7740, "torque_loss": 0.2108824998140335 }, { "epoch": 6.969424460431655, "grad_norm": 0.36356741189956665, "learning_rate": 9.701709473196154e-05, "loss": 0.04, "step": 7750 }, { "action_loss": 0.013074073009192944, "epoch": 6.969424460431655, "step": 7750 }, { "epoch": 6.969424460431655, "step": 7750, "torque_loss": 0.18839621543884277 }, { "epoch": 6.9784172661870505, "grad_norm": 0.26319462060928345, "learning_rate": 9.700583321614541e-05, "loss": 0.0308, "step": 7760 }, { "action_loss": 0.005845933686941862, "epoch": 6.9784172661870505, "step": 7760 }, { "epoch": 6.9784172661870505, "step": 7760, "torque_loss": 0.15782374143600464 }, { "epoch": 6.987410071942446, "grad_norm": 0.2162293940782547, "learning_rate": 9.699455113838085e-05, "loss": 0.0319, "step": 7770 }, { "action_loss": 0.011911284178495407, "epoch": 6.987410071942446, "step": 7770 }, { "epoch": 6.987410071942446, "step": 7770, "torque_loss": 0.20553816854953766 }, { "epoch": 6.996402877697841, "grad_norm": 0.3822821080684662, "learning_rate": 9.698324850360303e-05, "loss": 0.0336, "step": 7780 }, { "action_loss": 0.0110965920612216, "epoch": 6.996402877697841, "step": 7780 }, { "epoch": 6.996402877697841, "step": 7780, "torque_loss": 0.24490495026111603 }, { "epoch": 7.005395683453237, "grad_norm": 0.21143105626106262, "learning_rate": 9.69719253167561e-05, "loss": 0.0317, "step": 7790 }, { "action_loss": 0.007044622674584389, "epoch": 7.005395683453237, "step": 7790 }, { "epoch": 7.005395683453237, "step": 7790, "torque_loss": 0.14932368695735931 }, { "epoch": 7.014388489208633, "grad_norm": 0.19464042782783508, "learning_rate": 9.69605815827932e-05, "loss": 0.0316, "step": 7800 }, { "action_loss": 0.013103079050779343, "epoch": 7.014388489208633, "step": 7800 }, { "epoch": 7.014388489208633, "step": 7800, "torque_loss": 0.2855842113494873 }, { "epoch": 7.023381294964029, "grad_norm": 0.18678387999534607, "learning_rate": 9.694921730667647e-05, "loss": 0.0344, "step": 7810 }, { "action_loss": 0.022480012848973274, "epoch": 7.023381294964029, "step": 7810 }, { "epoch": 7.023381294964029, "step": 7810, "torque_loss": 0.2454233169555664 }, { "epoch": 7.032374100719425, "grad_norm": 0.2135317474603653, "learning_rate": 9.693783249337705e-05, "loss": 0.0342, "step": 7820 }, { "action_loss": 0.00990782305598259, "epoch": 7.032374100719425, "step": 7820 }, { "epoch": 7.032374100719425, "step": 7820, "torque_loss": 0.1814882755279541 }, { "epoch": 7.0413669064748206, "grad_norm": 0.38954171538352966, "learning_rate": 9.692642714787501e-05, "loss": 0.037, "step": 7830 }, { "action_loss": 0.007656545843929052, "epoch": 7.0413669064748206, "step": 7830 }, { "epoch": 7.0413669064748206, "step": 7830, "torque_loss": 0.16832704842090607 }, { "epoch": 7.0503597122302155, "grad_norm": 0.17942845821380615, "learning_rate": 9.691500127515945e-05, "loss": 0.0312, "step": 7840 }, { "action_loss": 0.017248444259166718, "epoch": 7.0503597122302155, "step": 7840 }, { "epoch": 7.0503597122302155, "step": 7840, "torque_loss": 0.24778304994106293 }, { "epoch": 7.059352517985611, "grad_norm": 0.27985844016075134, "learning_rate": 9.690355488022844e-05, "loss": 0.0369, "step": 7850 }, { "action_loss": 0.011058385483920574, "epoch": 7.059352517985611, "step": 7850 }, { "epoch": 7.059352517985611, "step": 7850, "torque_loss": 0.21027004718780518 }, { "epoch": 7.068345323741007, "grad_norm": 0.19915609061717987, "learning_rate": 9.689208796808902e-05, "loss": 0.0341, "step": 7860 }, { "action_loss": 0.016458574682474136, "epoch": 7.068345323741007, "step": 7860 }, { "epoch": 7.068345323741007, "step": 7860, "torque_loss": 0.21000342071056366 }, { "epoch": 7.077338129496403, "grad_norm": 0.2614104449748993, "learning_rate": 9.68806005437572e-05, "loss": 0.0379, "step": 7870 }, { "action_loss": 0.014139116741716862, "epoch": 7.077338129496403, "step": 7870 }, { "epoch": 7.077338129496403, "step": 7870, "torque_loss": 0.1844182163476944 }, { "epoch": 7.086330935251799, "grad_norm": 0.29580584168434143, "learning_rate": 9.686909261225796e-05, "loss": 0.0346, "step": 7880 }, { "action_loss": 0.005617355462163687, "epoch": 7.086330935251799, "step": 7880 }, { "epoch": 7.086330935251799, "step": 7880, "torque_loss": 0.12551754713058472 }, { "epoch": 7.095323741007194, "grad_norm": 0.2591843903064728, "learning_rate": 9.685756417862531e-05, "loss": 0.0341, "step": 7890 }, { "action_loss": 0.005938439164310694, "epoch": 7.095323741007194, "step": 7890 }, { "epoch": 7.095323741007194, "step": 7890, "torque_loss": 0.14310914278030396 }, { "epoch": 7.10431654676259, "grad_norm": 0.3004903793334961, "learning_rate": 9.684601524790212e-05, "loss": 0.0408, "step": 7900 }, { "action_loss": 0.01320075523108244, "epoch": 7.10431654676259, "step": 7900 }, { "epoch": 7.10431654676259, "step": 7900, "torque_loss": 0.1949516087770462 }, { "epoch": 7.113309352517986, "grad_norm": 0.2633335292339325, "learning_rate": 9.68344458251403e-05, "loss": 0.0347, "step": 7910 }, { "action_loss": 0.012264844961464405, "epoch": 7.113309352517986, "step": 7910 }, { "epoch": 7.113309352517986, "step": 7910, "torque_loss": 0.2251339554786682 }, { "epoch": 7.122302158273381, "grad_norm": 0.3024846315383911, "learning_rate": 9.682285591540072e-05, "loss": 0.034, "step": 7920 }, { "action_loss": 0.008617785759270191, "epoch": 7.122302158273381, "step": 7920 }, { "epoch": 7.122302158273381, "step": 7920, "torque_loss": 0.17405278980731964 }, { "epoch": 7.131294964028777, "grad_norm": 0.1792706698179245, "learning_rate": 9.681124552375322e-05, "loss": 0.0322, "step": 7930 }, { "action_loss": 0.01015088614076376, "epoch": 7.131294964028777, "step": 7930 }, { "epoch": 7.131294964028777, "step": 7930, "torque_loss": 0.20823077857494354 }, { "epoch": 7.140287769784172, "grad_norm": 0.21062767505645752, "learning_rate": 9.679961465527654e-05, "loss": 0.0384, "step": 7940 }, { "action_loss": 0.016100822016596794, "epoch": 7.140287769784172, "step": 7940 }, { "epoch": 7.140287769784172, "step": 7940, "torque_loss": 0.2058030366897583 }, { "epoch": 7.149280575539568, "grad_norm": 0.2780526876449585, "learning_rate": 9.678796331505843e-05, "loss": 0.0309, "step": 7950 }, { "action_loss": 0.016414668411016464, "epoch": 7.149280575539568, "step": 7950 }, { "epoch": 7.149280575539568, "step": 7950, "torque_loss": 0.19334618747234344 }, { "epoch": 7.158273381294964, "grad_norm": 0.20677366852760315, "learning_rate": 9.677629150819558e-05, "loss": 0.0315, "step": 7960 }, { "action_loss": 0.032675851136446, "epoch": 7.158273381294964, "step": 7960 }, { "epoch": 7.158273381294964, "step": 7960, "torque_loss": 0.212631955742836 }, { "epoch": 7.16726618705036, "grad_norm": 0.2512941062450409, "learning_rate": 9.676459923979366e-05, "loss": 0.0356, "step": 7970 }, { "action_loss": 0.010680650360882282, "epoch": 7.16726618705036, "step": 7970 }, { "epoch": 7.16726618705036, "step": 7970, "torque_loss": 0.17450399696826935 }, { "epoch": 7.176258992805756, "grad_norm": 0.15898439288139343, "learning_rate": 9.675288651496723e-05, "loss": 0.0311, "step": 7980 }, { "action_loss": 0.014774791896343231, "epoch": 7.176258992805756, "step": 7980 }, { "epoch": 7.176258992805756, "step": 7980, "torque_loss": 0.20560704171657562 }, { "epoch": 7.1852517985611515, "grad_norm": 0.3281826078891754, "learning_rate": 9.674115333883986e-05, "loss": 0.0375, "step": 7990 }, { "action_loss": 0.016302896663546562, "epoch": 7.1852517985611515, "step": 7990 }, { "epoch": 7.1852517985611515, "step": 7990, "torque_loss": 0.21732540428638458 }, { "epoch": 7.194244604316546, "grad_norm": 0.28530779480934143, "learning_rate": 9.672939971654402e-05, "loss": 0.0376, "step": 8000 }, { "action_loss": 0.02483583241701126, "epoch": 7.194244604316546, "step": 8000 }, { "epoch": 7.194244604316546, "step": 8000, "torque_loss": 0.21372336149215698 }, { "epoch": 7.203237410071942, "grad_norm": 0.31884950399398804, "learning_rate": 9.671762565322117e-05, "loss": 0.0336, "step": 8010 }, { "action_loss": 0.020027825608849525, "epoch": 7.203237410071942, "step": 8010 }, { "epoch": 7.203237410071942, "step": 8010, "torque_loss": 0.20337945222854614 }, { "epoch": 7.212230215827338, "grad_norm": 0.15755747258663177, "learning_rate": 9.670583115402166e-05, "loss": 0.0334, "step": 8020 }, { "action_loss": 0.012202315032482147, "epoch": 7.212230215827338, "step": 8020 }, { "epoch": 7.212230215827338, "step": 8020, "torque_loss": 0.18008540570735931 }, { "epoch": 7.221223021582734, "grad_norm": 0.3766329288482666, "learning_rate": 9.669401622410482e-05, "loss": 0.0305, "step": 8030 }, { "action_loss": 0.008078057318925858, "epoch": 7.221223021582734, "step": 8030 }, { "epoch": 7.221223021582734, "step": 8030, "torque_loss": 0.1285731941461563 }, { "epoch": 7.23021582733813, "grad_norm": 0.27575868368148804, "learning_rate": 9.668218086863887e-05, "loss": 0.0351, "step": 8040 }, { "action_loss": 0.009649393148720264, "epoch": 7.23021582733813, "step": 8040 }, { "epoch": 7.23021582733813, "step": 8040, "torque_loss": 0.19371606409549713 }, { "epoch": 7.239208633093525, "grad_norm": 0.18650777637958527, "learning_rate": 9.667032509280103e-05, "loss": 0.034, "step": 8050 }, { "action_loss": 0.01893775351345539, "epoch": 7.239208633093525, "step": 8050 }, { "epoch": 7.239208633093525, "step": 8050, "torque_loss": 0.1918845772743225 }, { "epoch": 7.248201438848921, "grad_norm": 0.19822032749652863, "learning_rate": 9.665844890177743e-05, "loss": 0.0329, "step": 8060 }, { "action_loss": 0.027953391894698143, "epoch": 7.248201438848921, "step": 8060 }, { "epoch": 7.248201438848921, "step": 8060, "torque_loss": 0.18490123748779297 }, { "epoch": 7.2571942446043165, "grad_norm": 0.12689220905303955, "learning_rate": 9.664655230076309e-05, "loss": 0.0323, "step": 8070 }, { "action_loss": 0.014615654945373535, "epoch": 7.2571942446043165, "step": 8070 }, { "epoch": 7.2571942446043165, "step": 8070, "torque_loss": 0.19516535103321075 }, { "epoch": 7.266187050359712, "grad_norm": 0.1828797608613968, "learning_rate": 9.663463529496199e-05, "loss": 0.038, "step": 8080 }, { "action_loss": 0.027008576318621635, "epoch": 7.266187050359712, "step": 8080 }, { "epoch": 7.266187050359712, "step": 8080, "torque_loss": 0.2954448163509369 }, { "epoch": 7.275179856115108, "grad_norm": 0.210269495844841, "learning_rate": 9.662269788958705e-05, "loss": 0.0329, "step": 8090 }, { "action_loss": 0.005695914383977652, "epoch": 7.275179856115108, "step": 8090 }, { "epoch": 7.275179856115108, "step": 8090, "torque_loss": 0.1543056219816208 }, { "epoch": 7.284172661870503, "grad_norm": 0.3476890027523041, "learning_rate": 9.661074008986008e-05, "loss": 0.0371, "step": 8100 }, { "action_loss": 0.01865754835307598, "epoch": 7.284172661870503, "step": 8100 }, { "epoch": 7.284172661870503, "step": 8100, "torque_loss": 0.22520901262760162 }, { "epoch": 7.293165467625899, "grad_norm": 0.2870640158653259, "learning_rate": 9.659876190101184e-05, "loss": 0.0332, "step": 8110 }, { "action_loss": 0.016189076006412506, "epoch": 7.293165467625899, "step": 8110 }, { "epoch": 7.293165467625899, "step": 8110, "torque_loss": 0.2061740905046463 }, { "epoch": 7.302158273381295, "grad_norm": 0.30750155448913574, "learning_rate": 9.6586763328282e-05, "loss": 0.0416, "step": 8120 }, { "action_loss": 0.03247161582112312, "epoch": 7.302158273381295, "step": 8120 }, { "epoch": 7.302158273381295, "step": 8120, "torque_loss": 0.19880561530590057 }, { "epoch": 7.311151079136691, "grad_norm": 0.3182356357574463, "learning_rate": 9.657474437691913e-05, "loss": 0.0387, "step": 8130 }, { "action_loss": 0.015767274424433708, "epoch": 7.311151079136691, "step": 8130 }, { "epoch": 7.311151079136691, "step": 8130, "torque_loss": 0.18229232728481293 }, { "epoch": 7.320143884892087, "grad_norm": 0.30099862813949585, "learning_rate": 9.656270505218073e-05, "loss": 0.0347, "step": 8140 }, { "action_loss": 0.023168841376900673, "epoch": 7.320143884892087, "step": 8140 }, { "epoch": 7.320143884892087, "step": 8140, "torque_loss": 0.2264184206724167 }, { "epoch": 7.329136690647482, "grad_norm": 0.27383026480674744, "learning_rate": 9.655064535933321e-05, "loss": 0.0338, "step": 8150 }, { "action_loss": 0.015593434683978558, "epoch": 7.329136690647482, "step": 8150 }, { "epoch": 7.329136690647482, "step": 8150, "torque_loss": 0.2207316905260086 }, { "epoch": 7.338129496402877, "grad_norm": 0.244410440325737, "learning_rate": 9.653856530365189e-05, "loss": 0.0368, "step": 8160 }, { "action_loss": 0.016926631331443787, "epoch": 7.338129496402877, "step": 8160 }, { "epoch": 7.338129496402877, "step": 8160, "torque_loss": 0.172541543841362 }, { "epoch": 7.347122302158273, "grad_norm": 0.17890721559524536, "learning_rate": 9.6526464890421e-05, "loss": 0.0322, "step": 8170 }, { "action_loss": 0.010293482802808285, "epoch": 7.347122302158273, "step": 8170 }, { "epoch": 7.347122302158273, "step": 8170, "torque_loss": 0.20459724962711334 }, { "epoch": 7.356115107913669, "grad_norm": 0.1767326444387436, "learning_rate": 9.651434412493367e-05, "loss": 0.0299, "step": 8180 }, { "action_loss": 0.01561713870614767, "epoch": 7.356115107913669, "step": 8180 }, { "epoch": 7.356115107913669, "step": 8180, "torque_loss": 0.19990195333957672 }, { "epoch": 7.365107913669065, "grad_norm": 0.22605395317077637, "learning_rate": 9.650220301249195e-05, "loss": 0.0349, "step": 8190 }, { "action_loss": 0.01282206829637289, "epoch": 7.365107913669065, "step": 8190 }, { "epoch": 7.365107913669065, "step": 8190, "torque_loss": 0.1780342310667038 }, { "epoch": 7.374100719424461, "grad_norm": 0.24138274788856506, "learning_rate": 9.649004155840675e-05, "loss": 0.0378, "step": 8200 }, { "action_loss": 0.015732524916529655, "epoch": 7.374100719424461, "step": 8200 }, { "epoch": 7.374100719424461, "step": 8200, "torque_loss": 0.20720303058624268 }, { "epoch": 7.383093525179856, "grad_norm": 0.1903763860464096, "learning_rate": 9.647785976799792e-05, "loss": 0.0351, "step": 8210 }, { "action_loss": 0.010661535896360874, "epoch": 7.383093525179856, "step": 8210 }, { "epoch": 7.383093525179856, "step": 8210, "torque_loss": 0.17766590416431427 }, { "epoch": 7.392086330935252, "grad_norm": 0.3407178223133087, "learning_rate": 9.646565764659417e-05, "loss": 0.0361, "step": 8220 }, { "action_loss": 0.011777136474847794, "epoch": 7.392086330935252, "step": 8220 }, { "epoch": 7.392086330935252, "step": 8220, "torque_loss": 0.15041334927082062 }, { "epoch": 7.401079136690647, "grad_norm": 0.14842362701892853, "learning_rate": 9.645343519953314e-05, "loss": 0.0307, "step": 8230 }, { "action_loss": 0.009190740995109081, "epoch": 7.401079136690647, "step": 8230 }, { "epoch": 7.401079136690647, "step": 8230, "torque_loss": 0.16744650900363922 }, { "epoch": 7.410071942446043, "grad_norm": 0.24655140936374664, "learning_rate": 9.644119243216135e-05, "loss": 0.0364, "step": 8240 }, { "action_loss": 0.005820287857204676, "epoch": 7.410071942446043, "step": 8240 }, { "epoch": 7.410071942446043, "step": 8240, "torque_loss": 0.1246153712272644 }, { "epoch": 7.419064748201439, "grad_norm": 0.266490638256073, "learning_rate": 9.642892934983417e-05, "loss": 0.0335, "step": 8250 }, { "action_loss": 0.010870248079299927, "epoch": 7.419064748201439, "step": 8250 }, { "epoch": 7.419064748201439, "step": 8250, "torque_loss": 0.22012068331241608 }, { "epoch": 7.428057553956835, "grad_norm": 0.25841522216796875, "learning_rate": 9.641664595791593e-05, "loss": 0.0334, "step": 8260 }, { "action_loss": 0.025610826909542084, "epoch": 7.428057553956835, "step": 8260 }, { "epoch": 7.428057553956835, "step": 8260, "torque_loss": 0.18311883509159088 }, { "epoch": 7.43705035971223, "grad_norm": 0.17928124964237213, "learning_rate": 9.640434226177977e-05, "loss": 0.0328, "step": 8270 }, { "action_loss": 0.016335228458046913, "epoch": 7.43705035971223, "step": 8270 }, { "epoch": 7.43705035971223, "step": 8270, "torque_loss": 0.18634001910686493 }, { "epoch": 7.446043165467626, "grad_norm": 0.2508355677127838, "learning_rate": 9.639201826680777e-05, "loss": 0.0402, "step": 8280 }, { "action_loss": 0.01144059281796217, "epoch": 7.446043165467626, "step": 8280 }, { "epoch": 7.446043165467626, "step": 8280, "torque_loss": 0.22922773659229279 }, { "epoch": 7.455035971223022, "grad_norm": 0.2021593302488327, "learning_rate": 9.637967397839083e-05, "loss": 0.0345, "step": 8290 }, { "action_loss": 0.011430948972702026, "epoch": 7.455035971223022, "step": 8290 }, { "epoch": 7.455035971223022, "step": 8290, "torque_loss": 0.19275449216365814 }, { "epoch": 7.4640287769784175, "grad_norm": 0.23334908485412598, "learning_rate": 9.63673094019288e-05, "loss": 0.0281, "step": 8300 }, { "action_loss": 0.009379236958920956, "epoch": 7.4640287769784175, "step": 8300 }, { "epoch": 7.4640287769784175, "step": 8300, "torque_loss": 0.21247637271881104 }, { "epoch": 7.473021582733813, "grad_norm": 0.1854831427335739, "learning_rate": 9.635492454283035e-05, "loss": 0.0273, "step": 8310 }, { "action_loss": 0.01031899731606245, "epoch": 7.473021582733813, "step": 8310 }, { "epoch": 7.473021582733813, "step": 8310, "torque_loss": 0.10674581676721573 }, { "epoch": 7.482014388489208, "grad_norm": 0.2426842451095581, "learning_rate": 9.634251940651302e-05, "loss": 0.0354, "step": 8320 }, { "action_loss": 0.008602027781307697, "epoch": 7.482014388489208, "step": 8320 }, { "epoch": 7.482014388489208, "step": 8320, "torque_loss": 0.18987195193767548 }, { "epoch": 7.491007194244604, "grad_norm": 0.2265763133764267, "learning_rate": 9.633009399840327e-05, "loss": 0.0326, "step": 8330 }, { "action_loss": 0.014680899679660797, "epoch": 7.491007194244604, "step": 8330 }, { "epoch": 7.491007194244604, "step": 8330, "torque_loss": 0.20563183724880219 }, { "epoch": 7.5, "grad_norm": 0.22845737636089325, "learning_rate": 9.631764832393639e-05, "loss": 0.0394, "step": 8340 }, { "action_loss": 0.009452690370380878, "epoch": 7.5, "step": 8340 }, { "epoch": 7.5, "step": 8340, "torque_loss": 0.22714637219905853 }, { "epoch": 7.508992805755396, "grad_norm": 0.27022337913513184, "learning_rate": 9.630518238855651e-05, "loss": 0.0342, "step": 8350 }, { "action_loss": 0.014300968497991562, "epoch": 7.508992805755396, "step": 8350 }, { "epoch": 7.508992805755396, "step": 8350, "torque_loss": 0.21503937244415283 }, { "epoch": 7.517985611510792, "grad_norm": 0.3810560405254364, "learning_rate": 9.629269619771668e-05, "loss": 0.0398, "step": 8360 }, { "action_loss": 0.011350934393703938, "epoch": 7.517985611510792, "step": 8360 }, { "epoch": 7.517985611510792, "step": 8360, "torque_loss": 0.17682601511478424 }, { "epoch": 7.5269784172661875, "grad_norm": 0.23303671181201935, "learning_rate": 9.628018975687878e-05, "loss": 0.0392, "step": 8370 }, { "action_loss": 0.00950328167527914, "epoch": 7.5269784172661875, "step": 8370 }, { "epoch": 7.5269784172661875, "step": 8370, "torque_loss": 0.13524849712848663 }, { "epoch": 7.5359712230215825, "grad_norm": 0.25787612795829773, "learning_rate": 9.626766307151355e-05, "loss": 0.036, "step": 8380 }, { "action_loss": 0.013693120330572128, "epoch": 7.5359712230215825, "step": 8380 }, { "epoch": 7.5359712230215825, "step": 8380, "torque_loss": 0.18043793737888336 }, { "epoch": 7.544964028776978, "grad_norm": 0.1805596649646759, "learning_rate": 9.62551161471006e-05, "loss": 0.0322, "step": 8390 }, { "action_loss": 0.01003164704889059, "epoch": 7.544964028776978, "step": 8390 }, { "epoch": 7.544964028776978, "step": 8390, "torque_loss": 0.20959655940532684 }, { "epoch": 7.553956834532374, "grad_norm": 0.15548758208751678, "learning_rate": 9.624254898912837e-05, "loss": 0.0299, "step": 8400 }, { "action_loss": 0.0068396530114114285, "epoch": 7.553956834532374, "step": 8400 }, { "epoch": 7.553956834532374, "step": 8400, "torque_loss": 0.12478544563055038 }, { "epoch": 7.56294964028777, "grad_norm": 0.20482492446899414, "learning_rate": 9.622996160309414e-05, "loss": 0.0295, "step": 8410 }, { "action_loss": 0.01279856264591217, "epoch": 7.56294964028777, "step": 8410 }, { "epoch": 7.56294964028777, "step": 8410, "torque_loss": 0.24536454677581787 }, { "epoch": 7.571942446043165, "grad_norm": 0.29561489820480347, "learning_rate": 9.62173539945041e-05, "loss": 0.031, "step": 8420 }, { "action_loss": 0.007749082054942846, "epoch": 7.571942446043165, "step": 8420 }, { "epoch": 7.571942446043165, "step": 8420, "torque_loss": 0.23283743858337402 }, { "epoch": 7.580935251798561, "grad_norm": 0.22502467036247253, "learning_rate": 9.620472616887323e-05, "loss": 0.0349, "step": 8430 }, { "action_loss": 0.009010125882923603, "epoch": 7.580935251798561, "step": 8430 }, { "epoch": 7.580935251798561, "step": 8430, "torque_loss": 0.2593877613544464 }, { "epoch": 7.589928057553957, "grad_norm": 0.21326887607574463, "learning_rate": 9.619207813172536e-05, "loss": 0.0316, "step": 8440 }, { "action_loss": 0.020570838823914528, "epoch": 7.589928057553957, "step": 8440 }, { "epoch": 7.589928057553957, "step": 8440, "torque_loss": 0.24425768852233887 }, { "epoch": 7.598920863309353, "grad_norm": 0.29244279861450195, "learning_rate": 9.617940988859318e-05, "loss": 0.0289, "step": 8450 }, { "action_loss": 0.012233230285346508, "epoch": 7.598920863309353, "step": 8450 }, { "epoch": 7.598920863309353, "step": 8450, "torque_loss": 0.2457510232925415 }, { "epoch": 7.607913669064748, "grad_norm": 0.15376687049865723, "learning_rate": 9.616672144501821e-05, "loss": 0.0306, "step": 8460 }, { "action_loss": 0.013181056827306747, "epoch": 7.607913669064748, "step": 8460 }, { "epoch": 7.607913669064748, "step": 8460, "torque_loss": 0.22066496312618256 }, { "epoch": 7.616906474820144, "grad_norm": 0.20861715078353882, "learning_rate": 9.615401280655081e-05, "loss": 0.0346, "step": 8470 }, { "action_loss": 0.007283002603799105, "epoch": 7.616906474820144, "step": 8470 }, { "epoch": 7.616906474820144, "step": 8470, "torque_loss": 0.1575421690940857 }, { "epoch": 7.625899280575539, "grad_norm": 0.285529762506485, "learning_rate": 9.614128397875017e-05, "loss": 0.0367, "step": 8480 }, { "action_loss": 0.012693043798208237, "epoch": 7.625899280575539, "step": 8480 }, { "epoch": 7.625899280575539, "step": 8480, "torque_loss": 0.1971394270658493 }, { "epoch": 7.634892086330935, "grad_norm": 0.2195603996515274, "learning_rate": 9.612853496718429e-05, "loss": 0.0362, "step": 8490 }, { "action_loss": 0.013487418182194233, "epoch": 7.634892086330935, "step": 8490 }, { "epoch": 7.634892086330935, "step": 8490, "torque_loss": 0.23296666145324707 }, { "epoch": 7.643884892086331, "grad_norm": 0.4223857522010803, "learning_rate": 9.611576577743004e-05, "loss": 0.0398, "step": 8500 }, { "action_loss": 0.01399412751197815, "epoch": 7.643884892086331, "step": 8500 }, { "epoch": 7.643884892086331, "step": 8500, "torque_loss": 0.1942361742258072 }, { "epoch": 7.652877697841727, "grad_norm": 0.3399066627025604, "learning_rate": 9.610297641507313e-05, "loss": 0.0431, "step": 8510 }, { "action_loss": 0.008916366845369339, "epoch": 7.652877697841727, "step": 8510 }, { "epoch": 7.652877697841727, "step": 8510, "torque_loss": 0.17110209167003632 }, { "epoch": 7.661870503597123, "grad_norm": 0.3511617183685303, "learning_rate": 9.6090166885708e-05, "loss": 0.0311, "step": 8520 }, { "action_loss": 0.006589203607290983, "epoch": 7.661870503597123, "step": 8520 }, { "epoch": 7.661870503597123, "step": 8520, "torque_loss": 0.12997086346149445 }, { "epoch": 7.670863309352518, "grad_norm": 0.3991003930568695, "learning_rate": 9.607733719493798e-05, "loss": 0.0373, "step": 8530 }, { "action_loss": 0.009811045601963997, "epoch": 7.670863309352518, "step": 8530 }, { "epoch": 7.670863309352518, "step": 8530, "torque_loss": 0.1656200885772705 }, { "epoch": 7.679856115107913, "grad_norm": 0.21513712406158447, "learning_rate": 9.606448734837527e-05, "loss": 0.0293, "step": 8540 }, { "action_loss": 0.011705639772117138, "epoch": 7.679856115107913, "step": 8540 }, { "epoch": 7.679856115107913, "step": 8540, "torque_loss": 0.15745770931243896 }, { "epoch": 7.688848920863309, "grad_norm": 0.2059655785560608, "learning_rate": 9.605161735164079e-05, "loss": 0.0322, "step": 8550 }, { "action_loss": 0.006816213950514793, "epoch": 7.688848920863309, "step": 8550 }, { "epoch": 7.688848920863309, "step": 8550, "torque_loss": 0.20677673816680908 }, { "epoch": 7.697841726618705, "grad_norm": 0.24987934529781342, "learning_rate": 9.60387272103643e-05, "loss": 0.0336, "step": 8560 }, { "action_loss": 0.017456693574786186, "epoch": 7.697841726618705, "step": 8560 }, { "epoch": 7.697841726618705, "step": 8560, "torque_loss": 0.23026704788208008 }, { "epoch": 7.706834532374101, "grad_norm": 0.1722753494977951, "learning_rate": 9.60258169301844e-05, "loss": 0.0289, "step": 8570 }, { "action_loss": 0.015503396280109882, "epoch": 7.706834532374101, "step": 8570 }, { "epoch": 7.706834532374101, "step": 8570, "torque_loss": 0.24945016205310822 }, { "epoch": 7.715827338129497, "grad_norm": 0.24144640564918518, "learning_rate": 9.601288651674851e-05, "loss": 0.036, "step": 8580 }, { "action_loss": 0.02303829789161682, "epoch": 7.715827338129497, "step": 8580 }, { "epoch": 7.715827338129497, "step": 8580, "torque_loss": 0.24646174907684326 }, { "epoch": 7.724820143884892, "grad_norm": 0.24054798483848572, "learning_rate": 9.59999359757128e-05, "loss": 0.0348, "step": 8590 }, { "action_loss": 0.022518714889883995, "epoch": 7.724820143884892, "step": 8590 }, { "epoch": 7.724820143884892, "step": 8590, "torque_loss": 0.20453090965747833 }, { "epoch": 7.733812949640288, "grad_norm": 0.21568138897418976, "learning_rate": 9.598696531274227e-05, "loss": 0.0334, "step": 8600 }, { "action_loss": 0.01782199554145336, "epoch": 7.733812949640288, "step": 8600 }, { "epoch": 7.733812949640288, "step": 8600, "torque_loss": 0.20949172973632812 }, { "epoch": 7.7428057553956835, "grad_norm": 0.31863802671432495, "learning_rate": 9.597397453351076e-05, "loss": 0.0326, "step": 8610 }, { "action_loss": 0.010348950512707233, "epoch": 7.7428057553956835, "step": 8610 }, { "epoch": 7.7428057553956835, "step": 8610, "torque_loss": 0.18151575326919556 }, { "epoch": 7.751798561151079, "grad_norm": 0.30161434412002563, "learning_rate": 9.596096364370087e-05, "loss": 0.0321, "step": 8620 }, { "action_loss": 0.062096964567899704, "epoch": 7.751798561151079, "step": 8620 }, { "epoch": 7.751798561151079, "step": 8620, "torque_loss": 0.23695261776447296 }, { "epoch": 7.760791366906475, "grad_norm": 0.24387876689434052, "learning_rate": 9.594793264900399e-05, "loss": 0.0367, "step": 8630 }, { "action_loss": 0.013692453503608704, "epoch": 7.760791366906475, "step": 8630 }, { "epoch": 7.760791366906475, "step": 8630, "torque_loss": 0.16989247500896454 }, { "epoch": 7.76978417266187, "grad_norm": 0.23933689296245575, "learning_rate": 9.593488155512032e-05, "loss": 0.0311, "step": 8640 }, { "action_loss": 0.006363391876220703, "epoch": 7.76978417266187, "step": 8640 }, { "epoch": 7.76978417266187, "step": 8640, "torque_loss": 0.18580962717533112 }, { "epoch": 7.778776978417266, "grad_norm": 0.20835010707378387, "learning_rate": 9.592181036775886e-05, "loss": 0.0333, "step": 8650 }, { "action_loss": 0.01548890769481659, "epoch": 7.778776978417266, "step": 8650 }, { "epoch": 7.778776978417266, "step": 8650, "torque_loss": 0.2271490842103958 }, { "epoch": 7.787769784172662, "grad_norm": 0.15793628990650177, "learning_rate": 9.590871909263741e-05, "loss": 0.0275, "step": 8660 }, { "action_loss": 0.013975817710161209, "epoch": 7.787769784172662, "step": 8660 }, { "epoch": 7.787769784172662, "step": 8660, "torque_loss": 0.1603052169084549 }, { "epoch": 7.796762589928058, "grad_norm": 0.306657612323761, "learning_rate": 9.589560773548252e-05, "loss": 0.0295, "step": 8670 }, { "action_loss": 0.022066868841648102, "epoch": 7.796762589928058, "step": 8670 }, { "epoch": 7.796762589928058, "step": 8670, "torque_loss": 0.24621210992336273 }, { "epoch": 7.805755395683454, "grad_norm": 0.23913083970546722, "learning_rate": 9.588247630202952e-05, "loss": 0.0328, "step": 8680 }, { "action_loss": 0.016041062772274017, "epoch": 7.805755395683454, "step": 8680 }, { "epoch": 7.805755395683454, "step": 8680, "torque_loss": 0.18289054930210114 }, { "epoch": 7.814748201438849, "grad_norm": 0.24877707660198212, "learning_rate": 9.586932479802258e-05, "loss": 0.0341, "step": 8690 }, { "action_loss": 0.01022382453083992, "epoch": 7.814748201438849, "step": 8690 }, { "epoch": 7.814748201438849, "step": 8690, "torque_loss": 0.20218081772327423 }, { "epoch": 7.823741007194244, "grad_norm": 0.2411748468875885, "learning_rate": 9.585615322921462e-05, "loss": 0.0348, "step": 8700 }, { "action_loss": 0.012298822402954102, "epoch": 7.823741007194244, "step": 8700 }, { "epoch": 7.823741007194244, "step": 8700, "torque_loss": 0.20801346004009247 }, { "epoch": 7.83273381294964, "grad_norm": 0.19498229026794434, "learning_rate": 9.58429616013673e-05, "loss": 0.0289, "step": 8710 }, { "action_loss": 0.01258592028170824, "epoch": 7.83273381294964, "step": 8710 }, { "epoch": 7.83273381294964, "step": 8710, "torque_loss": 0.19402217864990234 }, { "epoch": 7.841726618705036, "grad_norm": 0.23386725783348083, "learning_rate": 9.58297499202511e-05, "loss": 0.0333, "step": 8720 }, { "action_loss": 0.009129581041634083, "epoch": 7.841726618705036, "step": 8720 }, { "epoch": 7.841726618705036, "step": 8720, "torque_loss": 0.21926438808441162 }, { "epoch": 7.850719424460432, "grad_norm": 0.21630914509296417, "learning_rate": 9.581651819164525e-05, "loss": 0.0325, "step": 8730 }, { "action_loss": 0.014978591352701187, "epoch": 7.850719424460432, "step": 8730 }, { "epoch": 7.850719424460432, "step": 8730, "torque_loss": 0.23881103098392487 }, { "epoch": 7.859712230215827, "grad_norm": 0.20738057792186737, "learning_rate": 9.580326642133777e-05, "loss": 0.0297, "step": 8740 }, { "action_loss": 0.010281139053404331, "epoch": 7.859712230215827, "step": 8740 }, { "epoch": 7.859712230215827, "step": 8740, "torque_loss": 0.1385318785905838 }, { "epoch": 7.868705035971223, "grad_norm": 0.22646532952785492, "learning_rate": 9.578999461512544e-05, "loss": 0.0302, "step": 8750 }, { "action_loss": 0.005905198398977518, "epoch": 7.868705035971223, "step": 8750 }, { "epoch": 7.868705035971223, "step": 8750, "torque_loss": 0.19305109977722168 }, { "epoch": 7.877697841726619, "grad_norm": 0.23357421159744263, "learning_rate": 9.577670277881378e-05, "loss": 0.0326, "step": 8760 }, { "action_loss": 0.01911504752933979, "epoch": 7.877697841726619, "step": 8760 }, { "epoch": 7.877697841726619, "step": 8760, "torque_loss": 0.19560764729976654 }, { "epoch": 7.886690647482014, "grad_norm": 0.24940824508666992, "learning_rate": 9.57633909182171e-05, "loss": 0.0344, "step": 8770 }, { "action_loss": 0.008569804020226002, "epoch": 7.886690647482014, "step": 8770 }, { "epoch": 7.886690647482014, "step": 8770, "torque_loss": 0.14988677203655243 }, { "epoch": 7.89568345323741, "grad_norm": 0.13588757812976837, "learning_rate": 9.575005903915847e-05, "loss": 0.0339, "step": 8780 }, { "action_loss": 0.018409017473459244, "epoch": 7.89568345323741, "step": 8780 }, { "epoch": 7.89568345323741, "step": 8780, "torque_loss": 0.2720673382282257 }, { "epoch": 7.904676258992806, "grad_norm": 0.26678162813186646, "learning_rate": 9.573670714746972e-05, "loss": 0.0379, "step": 8790 }, { "action_loss": 0.021250100806355476, "epoch": 7.904676258992806, "step": 8790 }, { "epoch": 7.904676258992806, "step": 8790, "torque_loss": 0.18902647495269775 }, { "epoch": 7.913669064748201, "grad_norm": 0.26934707164764404, "learning_rate": 9.572333524899136e-05, "loss": 0.0426, "step": 8800 }, { "action_loss": 0.03147335350513458, "epoch": 7.913669064748201, "step": 8800 }, { "epoch": 7.913669064748201, "step": 8800, "torque_loss": 0.20067428052425385 }, { "epoch": 7.922661870503597, "grad_norm": 0.25989556312561035, "learning_rate": 9.570994334957278e-05, "loss": 0.0366, "step": 8810 }, { "action_loss": 0.00897859688848257, "epoch": 7.922661870503597, "step": 8810 }, { "epoch": 7.922661870503597, "step": 8810, "torque_loss": 0.1293591409921646 }, { "epoch": 7.931654676258993, "grad_norm": 0.20775410532951355, "learning_rate": 9.569653145507201e-05, "loss": 0.0338, "step": 8820 }, { "action_loss": 0.013996862806379795, "epoch": 7.931654676258993, "step": 8820 }, { "epoch": 7.931654676258993, "step": 8820, "torque_loss": 0.2332824468612671 }, { "epoch": 7.940647482014389, "grad_norm": 0.19653087854385376, "learning_rate": 9.568309957135587e-05, "loss": 0.0348, "step": 8830 }, { "action_loss": 0.0068814815022051334, "epoch": 7.940647482014389, "step": 8830 }, { "epoch": 7.940647482014389, "step": 8830, "torque_loss": 0.15921874344348907 }, { "epoch": 7.9496402877697845, "grad_norm": 0.3249801993370056, "learning_rate": 9.566964770429995e-05, "loss": 0.0356, "step": 8840 }, { "action_loss": 0.026550954207777977, "epoch": 7.9496402877697845, "step": 8840 }, { "epoch": 7.9496402877697845, "step": 8840, "torque_loss": 0.2644028961658478 }, { "epoch": 7.9586330935251794, "grad_norm": 0.2737503945827484, "learning_rate": 9.565617585978853e-05, "loss": 0.0316, "step": 8850 }, { "action_loss": 0.005166737828403711, "epoch": 7.9586330935251794, "step": 8850 }, { "epoch": 7.9586330935251794, "step": 8850, "torque_loss": 0.191722571849823 }, { "epoch": 7.967625899280575, "grad_norm": 0.12927016615867615, "learning_rate": 9.564268404371466e-05, "loss": 0.0312, "step": 8860 }, { "action_loss": 0.00994812697172165, "epoch": 7.967625899280575, "step": 8860 }, { "epoch": 7.967625899280575, "step": 8860, "torque_loss": 0.22137205302715302 }, { "epoch": 7.976618705035971, "grad_norm": 0.1863330602645874, "learning_rate": 9.562917226198013e-05, "loss": 0.0351, "step": 8870 }, { "action_loss": 0.012736499309539795, "epoch": 7.976618705035971, "step": 8870 }, { "epoch": 7.976618705035971, "step": 8870, "torque_loss": 0.279283732175827 }, { "epoch": 7.985611510791367, "grad_norm": 0.24956724047660828, "learning_rate": 9.561564052049543e-05, "loss": 0.042, "step": 8880 }, { "action_loss": 0.010394817218184471, "epoch": 7.985611510791367, "step": 8880 }, { "epoch": 7.985611510791367, "step": 8880, "torque_loss": 0.27338293194770813 }, { "epoch": 7.994604316546763, "grad_norm": 0.20164383947849274, "learning_rate": 9.560208882517982e-05, "loss": 0.0361, "step": 8890 }, { "action_loss": 0.005562593694776297, "epoch": 7.994604316546763, "step": 8890 }, { "epoch": 7.994604316546763, "step": 8890, "torque_loss": 0.08609280735254288 }, { "epoch": 8.003597122302159, "grad_norm": 0.14579297602176666, "learning_rate": 9.558851718196127e-05, "loss": 0.0248, "step": 8900 }, { "action_loss": 0.007749459240585566, "epoch": 8.003597122302159, "step": 8900 }, { "epoch": 8.003597122302159, "step": 8900, "torque_loss": 0.11794918030500412 }, { "epoch": 8.012589928057555, "grad_norm": 0.3905242085456848, "learning_rate": 9.557492559677646e-05, "loss": 0.038, "step": 8910 }, { "action_loss": 0.01324125099927187, "epoch": 8.012589928057555, "step": 8910 }, { "epoch": 8.012589928057555, "step": 8910, "torque_loss": 0.20988421142101288 }, { "epoch": 8.02158273381295, "grad_norm": 0.24176478385925293, "learning_rate": 9.556131407557082e-05, "loss": 0.032, "step": 8920 }, { "action_loss": 0.015691310167312622, "epoch": 8.02158273381295, "step": 8920 }, { "epoch": 8.02158273381295, "step": 8920, "torque_loss": 0.1987942010164261 }, { "epoch": 8.030575539568344, "grad_norm": 0.20899009704589844, "learning_rate": 9.554768262429853e-05, "loss": 0.034, "step": 8930 }, { "action_loss": 0.009550796821713448, "epoch": 8.030575539568344, "step": 8930 }, { "epoch": 8.030575539568344, "step": 8930, "torque_loss": 0.17148785293102264 }, { "epoch": 8.03956834532374, "grad_norm": 0.20250782370567322, "learning_rate": 9.553403124892239e-05, "loss": 0.032, "step": 8940 }, { "action_loss": 0.010284815914928913, "epoch": 8.03956834532374, "step": 8940 }, { "epoch": 8.03956834532374, "step": 8940, "torque_loss": 0.17091698944568634 }, { "epoch": 8.048561151079136, "grad_norm": 0.20848608016967773, "learning_rate": 9.552035995541402e-05, "loss": 0.0301, "step": 8950 }, { "action_loss": 0.01712142862379551, "epoch": 8.048561151079136, "step": 8950 }, { "epoch": 8.048561151079136, "step": 8950, "torque_loss": 0.1858498603105545 }, { "epoch": 8.057553956834532, "grad_norm": 0.17823579907417297, "learning_rate": 9.550666874975368e-05, "loss": 0.0344, "step": 8960 }, { "action_loss": 0.021199295297265053, "epoch": 8.057553956834532, "step": 8960 }, { "epoch": 8.057553956834532, "step": 8960, "torque_loss": 0.2790888845920563 }, { "epoch": 8.066546762589928, "grad_norm": 0.17919528484344482, "learning_rate": 9.549295763793038e-05, "loss": 0.0308, "step": 8970 }, { "action_loss": 0.04262082278728485, "epoch": 8.066546762589928, "step": 8970 }, { "epoch": 8.066546762589928, "step": 8970, "torque_loss": 0.24509359896183014 }, { "epoch": 8.075539568345324, "grad_norm": 0.21801340579986572, "learning_rate": 9.547922662594183e-05, "loss": 0.0346, "step": 8980 }, { "action_loss": 0.004592904821038246, "epoch": 8.075539568345324, "step": 8980 }, { "epoch": 8.075539568345324, "step": 8980, "torque_loss": 0.13046418130397797 }, { "epoch": 8.08453237410072, "grad_norm": 0.20201760530471802, "learning_rate": 9.546547571979443e-05, "loss": 0.0331, "step": 8990 }, { "action_loss": 0.008230033330619335, "epoch": 8.08453237410072, "step": 8990 }, { "epoch": 8.08453237410072, "step": 8990, "torque_loss": 0.1564786583185196 }, { "epoch": 8.093525179856115, "grad_norm": 0.2154601663351059, "learning_rate": 9.545170492550331e-05, "loss": 0.0308, "step": 9000 }, { "action_loss": 0.012577638030052185, "epoch": 8.093525179856115, "step": 9000 }, { "epoch": 8.093525179856115, "step": 9000, "torque_loss": 0.19014179706573486 }, { "epoch": 8.102517985611511, "grad_norm": 0.32803159952163696, "learning_rate": 9.543791424909226e-05, "loss": 0.0376, "step": 9010 }, { "action_loss": 0.035452935844659805, "epoch": 8.102517985611511, "step": 9010 }, { "epoch": 8.102517985611511, "step": 9010, "torque_loss": 0.23770791292190552 }, { "epoch": 8.111510791366907, "grad_norm": 0.1573616862297058, "learning_rate": 9.542410369659382e-05, "loss": 0.0381, "step": 9020 }, { "action_loss": 0.018018722534179688, "epoch": 8.111510791366907, "step": 9020 }, { "epoch": 8.111510791366907, "step": 9020, "torque_loss": 0.18817168474197388 }, { "epoch": 8.120503597122303, "grad_norm": 0.2724946141242981, "learning_rate": 9.541027327404916e-05, "loss": 0.037, "step": 9030 }, { "action_loss": 0.008988060988485813, "epoch": 8.120503597122303, "step": 9030 }, { "epoch": 8.120503597122303, "step": 9030, "torque_loss": 0.11920662969350815 }, { "epoch": 8.129496402877697, "grad_norm": 0.20779722929000854, "learning_rate": 9.539642298750821e-05, "loss": 0.0318, "step": 9040 }, { "action_loss": 0.03262241184711456, "epoch": 8.129496402877697, "step": 9040 }, { "epoch": 8.129496402877697, "step": 9040, "torque_loss": 0.29601338505744934 }, { "epoch": 8.138489208633093, "grad_norm": 0.21889305114746094, "learning_rate": 9.538255284302954e-05, "loss": 0.0337, "step": 9050 }, { "action_loss": 0.008097046054899693, "epoch": 8.138489208633093, "step": 9050 }, { "epoch": 8.138489208633093, "step": 9050, "torque_loss": 0.1534663438796997 }, { "epoch": 8.147482014388489, "grad_norm": 0.29646506905555725, "learning_rate": 9.536866284668043e-05, "loss": 0.0325, "step": 9060 }, { "action_loss": 0.006912857294082642, "epoch": 8.147482014388489, "step": 9060 }, { "epoch": 8.147482014388489, "step": 9060, "torque_loss": 0.1404895931482315 }, { "epoch": 8.156474820143885, "grad_norm": 0.2141975611448288, "learning_rate": 9.53547530045368e-05, "loss": 0.0295, "step": 9070 }, { "action_loss": 0.011797166429460049, "epoch": 8.156474820143885, "step": 9070 }, { "epoch": 8.156474820143885, "step": 9070, "torque_loss": 0.2944466173648834 }, { "epoch": 8.16546762589928, "grad_norm": 0.2026107758283615, "learning_rate": 9.534082332268335e-05, "loss": 0.0347, "step": 9080 }, { "action_loss": 0.008533322252333164, "epoch": 8.16546762589928, "step": 9080 }, { "epoch": 8.16546762589928, "step": 9080, "torque_loss": 0.15596476197242737 }, { "epoch": 8.174460431654676, "grad_norm": 0.2761709988117218, "learning_rate": 9.532687380721334e-05, "loss": 0.0297, "step": 9090 }, { "action_loss": 0.007767761591821909, "epoch": 8.174460431654676, "step": 9090 }, { "epoch": 8.174460431654676, "step": 9090, "torque_loss": 0.12849830090999603 }, { "epoch": 8.183453237410072, "grad_norm": 0.24079664051532745, "learning_rate": 9.53129044642288e-05, "loss": 0.0286, "step": 9100 }, { "action_loss": 0.022633960470557213, "epoch": 8.183453237410072, "step": 9100 }, { "epoch": 8.183453237410072, "step": 9100, "torque_loss": 0.20255047082901 }, { "epoch": 8.192446043165468, "grad_norm": 0.19201241433620453, "learning_rate": 9.529891529984039e-05, "loss": 0.028, "step": 9110 }, { "action_loss": 0.01579468324780464, "epoch": 8.192446043165468, "step": 9110 }, { "epoch": 8.192446043165468, "step": 9110, "torque_loss": 0.2050972431898117 }, { "epoch": 8.201438848920864, "grad_norm": 0.24770013988018036, "learning_rate": 9.528490632016743e-05, "loss": 0.0351, "step": 9120 }, { "action_loss": 0.01279588881880045, "epoch": 8.201438848920864, "step": 9120 }, { "epoch": 8.201438848920864, "step": 9120, "torque_loss": 0.16052399575710297 }, { "epoch": 8.21043165467626, "grad_norm": 0.2814595401287079, "learning_rate": 9.52708775313379e-05, "loss": 0.0318, "step": 9130 }, { "action_loss": 0.015461468137800694, "epoch": 8.21043165467626, "step": 9130 }, { "epoch": 8.21043165467626, "step": 9130, "torque_loss": 0.19719941914081573 }, { "epoch": 8.219424460431656, "grad_norm": 0.217536062002182, "learning_rate": 9.525682893948853e-05, "loss": 0.0311, "step": 9140 }, { "action_loss": 0.016205640509724617, "epoch": 8.219424460431656, "step": 9140 }, { "epoch": 8.219424460431656, "step": 9140, "torque_loss": 0.23096077144145966 }, { "epoch": 8.22841726618705, "grad_norm": 0.21743372082710266, "learning_rate": 9.524276055076463e-05, "loss": 0.029, "step": 9150 }, { "action_loss": 0.009623491205275059, "epoch": 8.22841726618705, "step": 9150 }, { "epoch": 8.22841726618705, "step": 9150, "torque_loss": 0.18568545579910278 }, { "epoch": 8.237410071942445, "grad_norm": 0.3182494342327118, "learning_rate": 9.522867237132016e-05, "loss": 0.0345, "step": 9160 }, { "action_loss": 0.01133361179381609, "epoch": 8.237410071942445, "step": 9160 }, { "epoch": 8.237410071942445, "step": 9160, "torque_loss": 0.2061631679534912 }, { "epoch": 8.246402877697841, "grad_norm": 0.2327340990304947, "learning_rate": 9.52145644073178e-05, "loss": 0.0313, "step": 9170 }, { "action_loss": 0.01834051124751568, "epoch": 8.246402877697841, "step": 9170 }, { "epoch": 8.246402877697841, "step": 9170, "torque_loss": 0.18843336403369904 }, { "epoch": 8.255395683453237, "grad_norm": 0.22034458816051483, "learning_rate": 9.520043666492884e-05, "loss": 0.0311, "step": 9180 }, { "action_loss": 0.01288488507270813, "epoch": 8.255395683453237, "step": 9180 }, { "epoch": 8.255395683453237, "step": 9180, "torque_loss": 0.23487909138202667 }, { "epoch": 8.264388489208633, "grad_norm": 0.1997852623462677, "learning_rate": 9.518628915033323e-05, "loss": 0.0321, "step": 9190 }, { "action_loss": 0.006719779223203659, "epoch": 8.264388489208633, "step": 9190 }, { "epoch": 8.264388489208633, "step": 9190, "torque_loss": 0.19678103923797607 }, { "epoch": 8.273381294964029, "grad_norm": 0.26754164695739746, "learning_rate": 9.517212186971957e-05, "loss": 0.0331, "step": 9200 }, { "action_loss": 0.013803966343402863, "epoch": 8.273381294964029, "step": 9200 }, { "epoch": 8.273381294964029, "step": 9200, "torque_loss": 0.20217116177082062 }, { "epoch": 8.282374100719425, "grad_norm": 0.2017534077167511, "learning_rate": 9.515793482928515e-05, "loss": 0.0412, "step": 9210 }, { "action_loss": 0.02174907922744751, "epoch": 8.282374100719425, "step": 9210 }, { "epoch": 8.282374100719425, "step": 9210, "torque_loss": 0.246564581990242 }, { "epoch": 8.29136690647482, "grad_norm": 0.3030499517917633, "learning_rate": 9.514372803523582e-05, "loss": 0.0298, "step": 9220 }, { "action_loss": 0.009235230274498463, "epoch": 8.29136690647482, "step": 9220 }, { "epoch": 8.29136690647482, "step": 9220, "torque_loss": 0.16625751554965973 }, { "epoch": 8.300359712230216, "grad_norm": 0.27933257818222046, "learning_rate": 9.512950149378613e-05, "loss": 0.0342, "step": 9230 }, { "action_loss": 0.011213235557079315, "epoch": 8.300359712230216, "step": 9230 }, { "epoch": 8.300359712230216, "step": 9230, "torque_loss": 0.23161321878433228 }, { "epoch": 8.309352517985612, "grad_norm": 0.2194683998823166, "learning_rate": 9.511525521115925e-05, "loss": 0.0334, "step": 9240 }, { "action_loss": 0.006594844628125429, "epoch": 8.309352517985612, "step": 9240 }, { "epoch": 8.309352517985612, "step": 9240, "torque_loss": 0.12788136303424835 }, { "epoch": 8.318345323741006, "grad_norm": 0.21604207158088684, "learning_rate": 9.510098919358698e-05, "loss": 0.0314, "step": 9250 }, { "action_loss": 0.019207242876291275, "epoch": 8.318345323741006, "step": 9250 }, { "epoch": 8.318345323741006, "step": 9250, "torque_loss": 0.2455517202615738 }, { "epoch": 8.327338129496402, "grad_norm": 0.1921977549791336, "learning_rate": 9.508670344730979e-05, "loss": 0.0323, "step": 9260 }, { "action_loss": 0.027315175160765648, "epoch": 8.327338129496402, "step": 9260 }, { "epoch": 8.327338129496402, "step": 9260, "torque_loss": 0.22330975532531738 }, { "epoch": 8.336330935251798, "grad_norm": 0.2262052595615387, "learning_rate": 9.507239797857672e-05, "loss": 0.0289, "step": 9270 }, { "action_loss": 0.013392443768680096, "epoch": 8.336330935251798, "step": 9270 }, { "epoch": 8.336330935251798, "step": 9270, "torque_loss": 0.2337317019701004 }, { "epoch": 8.345323741007194, "grad_norm": 0.17264635860919952, "learning_rate": 9.505807279364548e-05, "loss": 0.0364, "step": 9280 }, { "action_loss": 0.010998625308275223, "epoch": 8.345323741007194, "step": 9280 }, { "epoch": 8.345323741007194, "step": 9280, "torque_loss": 0.1584264487028122 }, { "epoch": 8.35431654676259, "grad_norm": 0.2728192210197449, "learning_rate": 9.504372789878239e-05, "loss": 0.0322, "step": 9290 }, { "action_loss": 0.01890534907579422, "epoch": 8.35431654676259, "step": 9290 }, { "epoch": 8.35431654676259, "step": 9290, "torque_loss": 0.19627141952514648 }, { "epoch": 8.363309352517986, "grad_norm": 0.28376224637031555, "learning_rate": 9.502936330026239e-05, "loss": 0.0389, "step": 9300 }, { "action_loss": 0.002791597740724683, "epoch": 8.363309352517986, "step": 9300 }, { "epoch": 8.363309352517986, "step": 9300, "torque_loss": 0.09172096103429794 }, { "epoch": 8.372302158273381, "grad_norm": 0.26374828815460205, "learning_rate": 9.501497900436906e-05, "loss": 0.0341, "step": 9310 }, { "action_loss": 0.07753735780715942, "epoch": 8.372302158273381, "step": 9310 }, { "epoch": 8.372302158273381, "step": 9310, "torque_loss": 0.20749075710773468 }, { "epoch": 8.381294964028777, "grad_norm": 0.23102137446403503, "learning_rate": 9.500057501739457e-05, "loss": 0.0478, "step": 9320 }, { "action_loss": 0.020763156935572624, "epoch": 8.381294964028777, "step": 9320 }, { "epoch": 8.381294964028777, "step": 9320, "torque_loss": 0.3501322269439697 }, { "epoch": 8.390287769784173, "grad_norm": 0.25343987345695496, "learning_rate": 9.49861513456397e-05, "loss": 0.0368, "step": 9330 }, { "action_loss": 0.004633811302483082, "epoch": 8.390287769784173, "step": 9330 }, { "epoch": 8.390287769784173, "step": 9330, "torque_loss": 0.13975343108177185 }, { "epoch": 8.399280575539569, "grad_norm": 0.22005675733089447, "learning_rate": 9.497170799541388e-05, "loss": 0.034, "step": 9340 }, { "action_loss": 0.016890445724129677, "epoch": 8.399280575539569, "step": 9340 }, { "epoch": 8.399280575539569, "step": 9340, "torque_loss": 0.18466798961162567 }, { "epoch": 8.408273381294965, "grad_norm": 0.28564584255218506, "learning_rate": 9.495724497303508e-05, "loss": 0.0302, "step": 9350 }, { "action_loss": 0.004129704553633928, "epoch": 8.408273381294965, "step": 9350 }, { "epoch": 8.408273381294965, "step": 9350, "torque_loss": 0.2001556158065796 }, { "epoch": 8.417266187050359, "grad_norm": 0.18240807950496674, "learning_rate": 9.494276228482998e-05, "loss": 0.0284, "step": 9360 }, { "action_loss": 0.005544432904571295, "epoch": 8.417266187050359, "step": 9360 }, { "epoch": 8.417266187050359, "step": 9360, "torque_loss": 0.1618836671113968 }, { "epoch": 8.426258992805755, "grad_norm": 0.2250901162624359, "learning_rate": 9.492825993713374e-05, "loss": 0.0385, "step": 9370 }, { "action_loss": 0.004289452917873859, "epoch": 8.426258992805755, "step": 9370 }, { "epoch": 8.426258992805755, "step": 9370, "torque_loss": 0.1827588826417923 }, { "epoch": 8.43525179856115, "grad_norm": 0.20547355711460114, "learning_rate": 9.491373793629023e-05, "loss": 0.029, "step": 9380 }, { "action_loss": 0.008865582756698132, "epoch": 8.43525179856115, "step": 9380 }, { "epoch": 8.43525179856115, "step": 9380, "torque_loss": 0.18096347153186798 }, { "epoch": 8.444244604316546, "grad_norm": 0.26721471548080444, "learning_rate": 9.489919628865182e-05, "loss": 0.0414, "step": 9390 }, { "action_loss": 0.017180761322379112, "epoch": 8.444244604316546, "step": 9390 }, { "epoch": 8.444244604316546, "step": 9390, "torque_loss": 0.280579537153244 }, { "epoch": 8.453237410071942, "grad_norm": 0.22225865721702576, "learning_rate": 9.488463500057955e-05, "loss": 0.0325, "step": 9400 }, { "action_loss": 0.011215906590223312, "epoch": 8.453237410071942, "step": 9400 }, { "epoch": 8.453237410071942, "step": 9400, "torque_loss": 0.15685392916202545 }, { "epoch": 8.462230215827338, "grad_norm": 0.2510843276977539, "learning_rate": 9.487005407844302e-05, "loss": 0.0326, "step": 9410 }, { "action_loss": 0.011384663172066212, "epoch": 8.462230215827338, "step": 9410 }, { "epoch": 8.462230215827338, "step": 9410, "torque_loss": 0.16092202067375183 }, { "epoch": 8.471223021582734, "grad_norm": 0.19549211859703064, "learning_rate": 9.485545352862039e-05, "loss": 0.033, "step": 9420 }, { "action_loss": 0.005874400958418846, "epoch": 8.471223021582734, "step": 9420 }, { "epoch": 8.471223021582734, "step": 9420, "torque_loss": 0.17132945358753204 }, { "epoch": 8.48021582733813, "grad_norm": 0.253324955701828, "learning_rate": 9.48408333574985e-05, "loss": 0.0351, "step": 9430 }, { "action_loss": 0.013113193213939667, "epoch": 8.48021582733813, "step": 9430 }, { "epoch": 8.48021582733813, "step": 9430, "torque_loss": 0.24848294258117676 }, { "epoch": 8.489208633093526, "grad_norm": 0.23459428548812866, "learning_rate": 9.482619357147264e-05, "loss": 0.0349, "step": 9440 }, { "action_loss": 0.009507402777671814, "epoch": 8.489208633093526, "step": 9440 }, { "epoch": 8.489208633093526, "step": 9440, "torque_loss": 0.22732675075531006 }, { "epoch": 8.498201438848922, "grad_norm": 0.2612297236919403, "learning_rate": 9.481153417694679e-05, "loss": 0.0388, "step": 9450 }, { "action_loss": 0.0067267208360135555, "epoch": 8.498201438848922, "step": 9450 }, { "epoch": 8.498201438848922, "step": 9450, "torque_loss": 0.18959498405456543 }, { "epoch": 8.507194244604317, "grad_norm": 0.24052949249744415, "learning_rate": 9.479685518033347e-05, "loss": 0.0303, "step": 9460 }, { "action_loss": 0.009385772980749607, "epoch": 8.507194244604317, "step": 9460 }, { "epoch": 8.507194244604317, "step": 9460, "torque_loss": 0.17726635932922363 }, { "epoch": 8.516187050359711, "grad_norm": 0.3221879303455353, "learning_rate": 9.478215658805375e-05, "loss": 0.0301, "step": 9470 }, { "action_loss": 0.006748522166162729, "epoch": 8.516187050359711, "step": 9470 }, { "epoch": 8.516187050359711, "step": 9470, "torque_loss": 0.1655224859714508 }, { "epoch": 8.525179856115107, "grad_norm": 0.18909989297389984, "learning_rate": 9.47674384065373e-05, "loss": 0.032, "step": 9480 }, { "action_loss": 0.013896775431931019, "epoch": 8.525179856115107, "step": 9480 }, { "epoch": 8.525179856115107, "step": 9480, "torque_loss": 0.23310458660125732 }, { "epoch": 8.534172661870503, "grad_norm": 0.33004119992256165, "learning_rate": 9.475270064222237e-05, "loss": 0.0338, "step": 9490 }, { "action_loss": 0.009924544021487236, "epoch": 8.534172661870503, "step": 9490 }, { "epoch": 8.534172661870503, "step": 9490, "torque_loss": 0.1602601408958435 }, { "epoch": 8.543165467625899, "grad_norm": 0.2736348509788513, "learning_rate": 9.473794330155572e-05, "loss": 0.0352, "step": 9500 }, { "action_loss": 0.018747840076684952, "epoch": 8.543165467625899, "step": 9500 }, { "epoch": 8.543165467625899, "step": 9500, "torque_loss": 0.2131328135728836 }, { "epoch": 8.552158273381295, "grad_norm": 0.23312871158123016, "learning_rate": 9.472316639099276e-05, "loss": 0.0301, "step": 9510 }, { "action_loss": 0.008278075605630875, "epoch": 8.552158273381295, "step": 9510 }, { "epoch": 8.552158273381295, "step": 9510, "torque_loss": 0.1323881894350052 }, { "epoch": 8.56115107913669, "grad_norm": 0.3072591722011566, "learning_rate": 9.470836991699739e-05, "loss": 0.0352, "step": 9520 }, { "action_loss": 0.006909554824233055, "epoch": 8.56115107913669, "step": 9520 }, { "epoch": 8.56115107913669, "step": 9520, "torque_loss": 0.1829233318567276 }, { "epoch": 8.570143884892087, "grad_norm": 0.25605571269989014, "learning_rate": 9.469355388604208e-05, "loss": 0.0329, "step": 9530 }, { "action_loss": 0.014309708960354328, "epoch": 8.570143884892087, "step": 9530 }, { "epoch": 8.570143884892087, "step": 9530, "torque_loss": 0.20180664956569672 }, { "epoch": 8.579136690647482, "grad_norm": 0.21852368116378784, "learning_rate": 9.467871830460787e-05, "loss": 0.0354, "step": 9540 }, { "action_loss": 0.016414666548371315, "epoch": 8.579136690647482, "step": 9540 }, { "epoch": 8.579136690647482, "step": 9540, "torque_loss": 0.25274530053138733 }, { "epoch": 8.588129496402878, "grad_norm": 0.25352397561073303, "learning_rate": 9.466386317918436e-05, "loss": 0.0406, "step": 9550 }, { "action_loss": 0.011177766136825085, "epoch": 8.588129496402878, "step": 9550 }, { "epoch": 8.588129496402878, "step": 9550, "torque_loss": 0.19724953174591064 }, { "epoch": 8.597122302158274, "grad_norm": 0.1589156985282898, "learning_rate": 9.464898851626969e-05, "loss": 0.0302, "step": 9560 }, { "action_loss": 0.027082661166787148, "epoch": 8.597122302158274, "step": 9560 }, { "epoch": 8.597122302158274, "step": 9560, "torque_loss": 0.29302629828453064 }, { "epoch": 8.60611510791367, "grad_norm": 0.3287903964519501, "learning_rate": 9.463409432237051e-05, "loss": 0.0341, "step": 9570 }, { "action_loss": 0.005883027333766222, "epoch": 8.60611510791367, "step": 9570 }, { "epoch": 8.60611510791367, "step": 9570, "torque_loss": 0.1717609316110611 }, { "epoch": 8.615107913669064, "grad_norm": 0.25668665766716003, "learning_rate": 9.461918060400209e-05, "loss": 0.0289, "step": 9580 }, { "action_loss": 0.0038599977269768715, "epoch": 8.615107913669064, "step": 9580 }, { "epoch": 8.615107913669064, "step": 9580, "torque_loss": 0.08856336027383804 }, { "epoch": 8.62410071942446, "grad_norm": 0.2788194417953491, "learning_rate": 9.460424736768816e-05, "loss": 0.0298, "step": 9590 }, { "action_loss": 0.019450830295681953, "epoch": 8.62410071942446, "step": 9590 }, { "epoch": 8.62410071942446, "step": 9590, "torque_loss": 0.20992068946361542 }, { "epoch": 8.633093525179856, "grad_norm": 0.19769906997680664, "learning_rate": 9.458929461996105e-05, "loss": 0.0399, "step": 9600 }, { "action_loss": 0.020525531843304634, "epoch": 8.633093525179856, "step": 9600 }, { "epoch": 8.633093525179856, "step": 9600, "torque_loss": 0.2827661335468292 }, { "epoch": 8.642086330935252, "grad_norm": 0.22531037032604218, "learning_rate": 9.457432236736158e-05, "loss": 0.0281, "step": 9610 }, { "action_loss": 0.0252432469278574, "epoch": 8.642086330935252, "step": 9610 }, { "epoch": 8.642086330935252, "step": 9610, "torque_loss": 0.19401396811008453 }, { "epoch": 8.651079136690647, "grad_norm": 0.2507822513580322, "learning_rate": 9.455933061643916e-05, "loss": 0.0351, "step": 9620 }, { "action_loss": 0.011465179733932018, "epoch": 8.651079136690647, "step": 9620 }, { "epoch": 8.651079136690647, "step": 9620, "torque_loss": 0.19177095592021942 }, { "epoch": 8.660071942446043, "grad_norm": 0.27144983410835266, "learning_rate": 9.454431937375164e-05, "loss": 0.0304, "step": 9630 }, { "action_loss": 0.005199111998081207, "epoch": 8.660071942446043, "step": 9630 }, { "epoch": 8.660071942446043, "step": 9630, "torque_loss": 0.15957452356815338 }, { "epoch": 8.66906474820144, "grad_norm": 0.21695546805858612, "learning_rate": 9.452928864586547e-05, "loss": 0.027, "step": 9640 }, { "action_loss": 0.007833280600607395, "epoch": 8.66906474820144, "step": 9640 }, { "epoch": 8.66906474820144, "step": 9640, "torque_loss": 0.14782781898975372 }, { "epoch": 8.678057553956835, "grad_norm": 0.16947869956493378, "learning_rate": 9.451423843935563e-05, "loss": 0.0354, "step": 9650 }, { "action_loss": 0.006929731462150812, "epoch": 8.678057553956835, "step": 9650 }, { "epoch": 8.678057553956835, "step": 9650, "torque_loss": 0.18131910264492035 }, { "epoch": 8.68705035971223, "grad_norm": 0.21195925772190094, "learning_rate": 9.449916876080553e-05, "loss": 0.0351, "step": 9660 }, { "action_loss": 0.012510131113231182, "epoch": 8.68705035971223, "step": 9660 }, { "epoch": 8.68705035971223, "step": 9660, "torque_loss": 0.21863628923892975 }, { "epoch": 8.696043165467627, "grad_norm": 0.2643700838088989, "learning_rate": 9.44840796168072e-05, "loss": 0.046, "step": 9670 }, { "action_loss": 0.009424169547855854, "epoch": 8.696043165467627, "step": 9670 }, { "epoch": 8.696043165467627, "step": 9670, "torque_loss": 0.22595594823360443 }, { "epoch": 8.70503597122302, "grad_norm": 0.19569142162799835, "learning_rate": 9.446897101396115e-05, "loss": 0.0317, "step": 9680 }, { "action_loss": 0.008691067807376385, "epoch": 8.70503597122302, "step": 9680 }, { "epoch": 8.70503597122302, "step": 9680, "torque_loss": 0.18836958706378937 }, { "epoch": 8.714028776978417, "grad_norm": 0.280246376991272, "learning_rate": 9.445384295887638e-05, "loss": 0.0321, "step": 9690 }, { "action_loss": 0.006904292851686478, "epoch": 8.714028776978417, "step": 9690 }, { "epoch": 8.714028776978417, "step": 9690, "torque_loss": 0.1022673025727272 }, { "epoch": 8.723021582733812, "grad_norm": 0.1809791922569275, "learning_rate": 9.443869545817043e-05, "loss": 0.0317, "step": 9700 }, { "action_loss": 0.00733586959540844, "epoch": 8.723021582733812, "step": 9700 }, { "epoch": 8.723021582733812, "step": 9700, "torque_loss": 0.18251699209213257 }, { "epoch": 8.732014388489208, "grad_norm": 0.18556571006774902, "learning_rate": 9.442352851846929e-05, "loss": 0.0281, "step": 9710 }, { "action_loss": 0.018562471494078636, "epoch": 8.732014388489208, "step": 9710 }, { "epoch": 8.732014388489208, "step": 9710, "torque_loss": 0.2451024204492569 }, { "epoch": 8.741007194244604, "grad_norm": 0.1554565131664276, "learning_rate": 9.440834214640755e-05, "loss": 0.0385, "step": 9720 }, { "action_loss": 0.00997657235711813, "epoch": 8.741007194244604, "step": 9720 }, { "epoch": 8.741007194244604, "step": 9720, "torque_loss": 0.1495974063873291 }, { "epoch": 8.75, "grad_norm": 0.16415755450725555, "learning_rate": 9.439313634862823e-05, "loss": 0.0291, "step": 9730 }, { "action_loss": 0.018862811848521233, "epoch": 8.75, "step": 9730 }, { "epoch": 8.75, "step": 9730, "torque_loss": 0.1854207068681717 }, { "epoch": 8.758992805755396, "grad_norm": 0.22966349124908447, "learning_rate": 9.437791113178282e-05, "loss": 0.0303, "step": 9740 }, { "action_loss": 0.01588273234665394, "epoch": 8.758992805755396, "step": 9740 }, { "epoch": 8.758992805755396, "step": 9740, "torque_loss": 0.16966795921325684 }, { "epoch": 8.767985611510792, "grad_norm": 0.18431130051612854, "learning_rate": 9.43626665025314e-05, "loss": 0.0362, "step": 9750 }, { "action_loss": 0.00794563814997673, "epoch": 8.767985611510792, "step": 9750 }, { "epoch": 8.767985611510792, "step": 9750, "torque_loss": 0.16801369190216064 }, { "epoch": 8.776978417266188, "grad_norm": 0.19922985136508942, "learning_rate": 9.434740246754248e-05, "loss": 0.0345, "step": 9760 }, { "action_loss": 0.009754304774105549, "epoch": 8.776978417266188, "step": 9760 }, { "epoch": 8.776978417266188, "step": 9760, "torque_loss": 0.19485688209533691 }, { "epoch": 8.785971223021583, "grad_norm": 0.25072625279426575, "learning_rate": 9.433211903349304e-05, "loss": 0.0304, "step": 9770 }, { "action_loss": 0.012740758247673512, "epoch": 8.785971223021583, "step": 9770 }, { "epoch": 8.785971223021583, "step": 9770, "torque_loss": 0.24931462109088898 }, { "epoch": 8.79496402877698, "grad_norm": 0.23990750312805176, "learning_rate": 9.431681620706858e-05, "loss": 0.0374, "step": 9780 }, { "action_loss": 0.008768956176936626, "epoch": 8.79496402877698, "step": 9780 }, { "epoch": 8.79496402877698, "step": 9780, "torque_loss": 0.21515683829784393 }, { "epoch": 8.803956834532373, "grad_norm": 0.3316521942615509, "learning_rate": 9.43014939949631e-05, "loss": 0.0354, "step": 9790 }, { "action_loss": 0.031964123249053955, "epoch": 8.803956834532373, "step": 9790 }, { "epoch": 8.803956834532373, "step": 9790, "torque_loss": 0.3195500671863556 }, { "epoch": 8.81294964028777, "grad_norm": 0.3705693185329437, "learning_rate": 9.428615240387904e-05, "loss": 0.0366, "step": 9800 }, { "action_loss": 0.013704146258533001, "epoch": 8.81294964028777, "step": 9800 }, { "epoch": 8.81294964028777, "step": 9800, "torque_loss": 0.19660258293151855 }, { "epoch": 8.821942446043165, "grad_norm": 0.2952157258987427, "learning_rate": 9.427079144052732e-05, "loss": 0.0341, "step": 9810 }, { "action_loss": 0.035437002778053284, "epoch": 8.821942446043165, "step": 9810 }, { "epoch": 8.821942446043165, "step": 9810, "torque_loss": 0.2831108272075653 }, { "epoch": 8.83093525179856, "grad_norm": 0.17751209437847137, "learning_rate": 9.425541111162739e-05, "loss": 0.029, "step": 9820 }, { "action_loss": 0.006327638402581215, "epoch": 8.83093525179856, "step": 9820 }, { "epoch": 8.83093525179856, "step": 9820, "torque_loss": 0.12184473127126694 }, { "epoch": 8.839928057553957, "grad_norm": 0.1445959508419037, "learning_rate": 9.424001142390709e-05, "loss": 0.0312, "step": 9830 }, { "action_loss": 0.007143082562834024, "epoch": 8.839928057553957, "step": 9830 }, { "epoch": 8.839928057553957, "step": 9830, "torque_loss": 0.17052555084228516 }, { "epoch": 8.848920863309353, "grad_norm": 0.16754254698753357, "learning_rate": 9.422459238410277e-05, "loss": 0.0279, "step": 9840 }, { "action_loss": 0.03915095701813698, "epoch": 8.848920863309353, "step": 9840 }, { "epoch": 8.848920863309353, "step": 9840, "torque_loss": 0.3830861747264862 }, { "epoch": 8.857913669064748, "grad_norm": 0.20404845476150513, "learning_rate": 9.420915399895926e-05, "loss": 0.0361, "step": 9850 }, { "action_loss": 0.014672771096229553, "epoch": 8.857913669064748, "step": 9850 }, { "epoch": 8.857913669064748, "step": 9850, "torque_loss": 0.2591046988964081 }, { "epoch": 8.866906474820144, "grad_norm": 0.18080642819404602, "learning_rate": 9.419369627522981e-05, "loss": 0.0354, "step": 9860 }, { "action_loss": 0.009361518546938896, "epoch": 8.866906474820144, "step": 9860 }, { "epoch": 8.866906474820144, "step": 9860, "torque_loss": 0.20863735675811768 }, { "epoch": 8.87589928057554, "grad_norm": 0.18278440833091736, "learning_rate": 9.417821921967618e-05, "loss": 0.0297, "step": 9870 }, { "action_loss": 0.002510546473786235, "epoch": 8.87589928057554, "step": 9870 }, { "epoch": 8.87589928057554, "step": 9870, "torque_loss": 0.09862566739320755 }, { "epoch": 8.884892086330936, "grad_norm": 0.2612435221672058, "learning_rate": 9.416272283906855e-05, "loss": 0.0326, "step": 9880 }, { "action_loss": 0.009028936736285686, "epoch": 8.884892086330936, "step": 9880 }, { "epoch": 8.884892086330936, "step": 9880, "torque_loss": 0.15105801820755005 }, { "epoch": 8.89388489208633, "grad_norm": 0.2415289729833603, "learning_rate": 9.414720714018554e-05, "loss": 0.0303, "step": 9890 }, { "action_loss": 0.03566320613026619, "epoch": 8.89388489208633, "step": 9890 }, { "epoch": 8.89388489208633, "step": 9890, "torque_loss": 0.1937968134880066 }, { "epoch": 8.902877697841726, "grad_norm": 0.23371197283267975, "learning_rate": 9.413167212981427e-05, "loss": 0.0372, "step": 9900 }, { "action_loss": 0.013769612647593021, "epoch": 8.902877697841726, "step": 9900 }, { "epoch": 8.902877697841726, "step": 9900, "torque_loss": 0.1676376610994339 }, { "epoch": 8.911870503597122, "grad_norm": 0.22912655770778656, "learning_rate": 9.41161178147503e-05, "loss": 0.0324, "step": 9910 }, { "action_loss": 0.00573281804099679, "epoch": 8.911870503597122, "step": 9910 }, { "epoch": 8.911870503597122, "step": 9910, "torque_loss": 0.1542486548423767 }, { "epoch": 8.920863309352518, "grad_norm": 0.16112713515758514, "learning_rate": 9.410054420179755e-05, "loss": 0.0255, "step": 9920 }, { "action_loss": 0.007712062448263168, "epoch": 8.920863309352518, "step": 9920 }, { "epoch": 8.920863309352518, "step": 9920, "torque_loss": 0.16079328954219818 }, { "epoch": 8.929856115107913, "grad_norm": 0.2572729289531708, "learning_rate": 9.408495129776851e-05, "loss": 0.0402, "step": 9930 }, { "action_loss": 0.006605897098779678, "epoch": 8.929856115107913, "step": 9930 }, { "epoch": 8.929856115107913, "step": 9930, "torque_loss": 0.18125416338443756 }, { "epoch": 8.93884892086331, "grad_norm": 0.17791199684143066, "learning_rate": 9.406933910948403e-05, "loss": 0.0335, "step": 9940 }, { "action_loss": 0.005604291334748268, "epoch": 8.93884892086331, "step": 9940 }, { "epoch": 8.93884892086331, "step": 9940, "torque_loss": 0.1516488790512085 }, { "epoch": 8.947841726618705, "grad_norm": 0.32619208097457886, "learning_rate": 9.40537076437734e-05, "loss": 0.0416, "step": 9950 }, { "action_loss": 0.011389710009098053, "epoch": 8.947841726618705, "step": 9950 }, { "epoch": 8.947841726618705, "step": 9950, "torque_loss": 0.17810480296611786 }, { "epoch": 8.956834532374101, "grad_norm": 0.1609949916601181, "learning_rate": 9.403805690747436e-05, "loss": 0.0295, "step": 9960 }, { "action_loss": 0.009692792780697346, "epoch": 8.956834532374101, "step": 9960 }, { "epoch": 8.956834532374101, "step": 9960, "torque_loss": 0.15659791231155396 }, { "epoch": 8.965827338129497, "grad_norm": 0.29904234409332275, "learning_rate": 9.402238690743308e-05, "loss": 0.0352, "step": 9970 }, { "action_loss": 0.007928424514830112, "epoch": 8.965827338129497, "step": 9970 }, { "epoch": 8.965827338129497, "step": 9970, "torque_loss": 0.1280156373977661 }, { "epoch": 8.974820143884893, "grad_norm": 0.19847488403320312, "learning_rate": 9.400669765050413e-05, "loss": 0.0362, "step": 9980 }, { "action_loss": 0.03046298772096634, "epoch": 8.974820143884893, "step": 9980 }, { "epoch": 8.974820143884893, "step": 9980, "torque_loss": 0.21040700376033783 }, { "epoch": 8.983812949640289, "grad_norm": 0.34056350588798523, "learning_rate": 9.399098914355055e-05, "loss": 0.0351, "step": 9990 }, { "action_loss": 0.0157011691480875, "epoch": 8.983812949640289, "step": 9990 }, { "epoch": 8.983812949640289, "step": 9990, "torque_loss": 0.16027142107486725 }, { "epoch": 8.992805755395683, "grad_norm": 0.24084602296352386, "learning_rate": 9.397526139344378e-05, "loss": 0.0341, "step": 10000 }, { "action_loss": 0.012702672742307186, "epoch": 8.992805755395683, "step": 10000 }, { "epoch": 8.992805755395683, "step": 10000, "torque_loss": 0.17381401360034943 }, { "epoch": 9.001798561151078, "grad_norm": 0.19891168177127838, "learning_rate": 9.395951440706362e-05, "loss": 0.0305, "step": 10010 }, { "action_loss": 0.008842985145747662, "epoch": 9.001798561151078, "step": 10010 }, { "epoch": 9.001798561151078, "step": 10010, "torque_loss": 0.14742881059646606 }, { "epoch": 9.010791366906474, "grad_norm": 0.2538318336009979, "learning_rate": 9.394374819129839e-05, "loss": 0.0311, "step": 10020 }, { "action_loss": 0.01400744915008545, "epoch": 9.010791366906474, "step": 10020 }, { "epoch": 9.010791366906474, "step": 10020, "torque_loss": 0.1485472470521927 }, { "epoch": 9.01978417266187, "grad_norm": 0.14946694672107697, "learning_rate": 9.392796275304474e-05, "loss": 0.0341, "step": 10030 }, { "action_loss": 0.015457329340279102, "epoch": 9.01978417266187, "step": 10030 }, { "epoch": 9.01978417266187, "step": 10030, "torque_loss": 0.18822526931762695 }, { "epoch": 9.028776978417266, "grad_norm": 0.2079620063304901, "learning_rate": 9.391215809920778e-05, "loss": 0.0335, "step": 10040 }, { "action_loss": 0.03309183567762375, "epoch": 9.028776978417266, "step": 10040 }, { "epoch": 9.028776978417266, "step": 10040, "torque_loss": 0.2438611388206482 }, { "epoch": 9.037769784172662, "grad_norm": 0.2895292341709137, "learning_rate": 9.389633423670099e-05, "loss": 0.0329, "step": 10050 }, { "action_loss": 0.013985581696033478, "epoch": 9.037769784172662, "step": 10050 }, { "epoch": 9.037769784172662, "step": 10050, "torque_loss": 0.21381942927837372 }, { "epoch": 9.046762589928058, "grad_norm": 0.21295231580734253, "learning_rate": 9.388049117244626e-05, "loss": 0.032, "step": 10060 }, { "action_loss": 0.00931435264647007, "epoch": 9.046762589928058, "step": 10060 }, { "epoch": 9.046762589928058, "step": 10060, "torque_loss": 0.22409997880458832 }, { "epoch": 9.055755395683454, "grad_norm": 0.21336399018764496, "learning_rate": 9.386462891337389e-05, "loss": 0.0356, "step": 10070 }, { "action_loss": 0.009143088944256306, "epoch": 9.055755395683454, "step": 10070 }, { "epoch": 9.055755395683454, "step": 10070, "torque_loss": 0.15038739144802094 }, { "epoch": 9.06474820143885, "grad_norm": 0.16419632732868195, "learning_rate": 9.384874746642257e-05, "loss": 0.0303, "step": 10080 }, { "action_loss": 0.009269064292311668, "epoch": 9.06474820143885, "step": 10080 }, { "epoch": 9.06474820143885, "step": 10080, "torque_loss": 0.15190324187278748 }, { "epoch": 9.073741007194245, "grad_norm": 0.1613650768995285, "learning_rate": 9.383284683853937e-05, "loss": 0.0347, "step": 10090 }, { "action_loss": 0.008314737118780613, "epoch": 9.073741007194245, "step": 10090 }, { "epoch": 9.073741007194245, "step": 10090, "torque_loss": 0.15567505359649658 }, { "epoch": 9.082733812949641, "grad_norm": 0.15526673197746277, "learning_rate": 9.381692703667981e-05, "loss": 0.0296, "step": 10100 }, { "action_loss": 0.008786574937403202, "epoch": 9.082733812949641, "step": 10100 }, { "epoch": 9.082733812949641, "step": 10100, "torque_loss": 0.16680443286895752 }, { "epoch": 9.091726618705035, "grad_norm": 0.17640849947929382, "learning_rate": 9.380098806780771e-05, "loss": 0.03, "step": 10110 }, { "action_loss": 0.004173473920673132, "epoch": 9.091726618705035, "step": 10110 }, { "epoch": 9.091726618705035, "step": 10110, "torque_loss": 0.2025788426399231 }, { "epoch": 9.100719424460431, "grad_norm": 0.1565021574497223, "learning_rate": 9.378502993889533e-05, "loss": 0.034, "step": 10120 }, { "action_loss": 0.015746431425213814, "epoch": 9.100719424460431, "step": 10120 }, { "epoch": 9.100719424460431, "step": 10120, "torque_loss": 0.2603205740451813 }, { "epoch": 9.109712230215827, "grad_norm": 0.1621607095003128, "learning_rate": 9.376905265692329e-05, "loss": 0.0293, "step": 10130 }, { "action_loss": 0.006315015256404877, "epoch": 9.109712230215827, "step": 10130 }, { "epoch": 9.109712230215827, "step": 10130, "torque_loss": 0.2110251933336258 }, { "epoch": 9.118705035971223, "grad_norm": 0.13114936649799347, "learning_rate": 9.37530562288806e-05, "loss": 0.0262, "step": 10140 }, { "action_loss": 0.025886690244078636, "epoch": 9.118705035971223, "step": 10140 }, { "epoch": 9.118705035971223, "step": 10140, "torque_loss": 0.26178938150405884 }, { "epoch": 9.127697841726619, "grad_norm": 0.2051761895418167, "learning_rate": 9.373704066176465e-05, "loss": 0.0347, "step": 10150 }, { "action_loss": 0.015345588326454163, "epoch": 9.127697841726619, "step": 10150 }, { "epoch": 9.127697841726619, "step": 10150, "torque_loss": 0.20698332786560059 }, { "epoch": 9.136690647482014, "grad_norm": 0.15391358733177185, "learning_rate": 9.372100596258118e-05, "loss": 0.0319, "step": 10160 }, { "action_loss": 0.01594315655529499, "epoch": 9.136690647482014, "step": 10160 }, { "epoch": 9.136690647482014, "step": 10160, "torque_loss": 0.24263687431812286 }, { "epoch": 9.14568345323741, "grad_norm": 0.2642539143562317, "learning_rate": 9.370495213834433e-05, "loss": 0.0309, "step": 10170 }, { "action_loss": 0.02183375507593155, "epoch": 9.14568345323741, "step": 10170 }, { "epoch": 9.14568345323741, "step": 10170, "torque_loss": 0.2322831153869629 }, { "epoch": 9.154676258992806, "grad_norm": 0.26160943508148193, "learning_rate": 9.368887919607657e-05, "loss": 0.0359, "step": 10180 }, { "action_loss": 0.027252063155174255, "epoch": 9.154676258992806, "step": 10180 }, { "epoch": 9.154676258992806, "step": 10180, "torque_loss": 0.22891223430633545 }, { "epoch": 9.163669064748202, "grad_norm": 0.2325047105550766, "learning_rate": 9.367278714280876e-05, "loss": 0.0414, "step": 10190 }, { "action_loss": 0.007583634927868843, "epoch": 9.163669064748202, "step": 10190 }, { "epoch": 9.163669064748202, "step": 10190, "torque_loss": 0.16966044902801514 }, { "epoch": 9.172661870503598, "grad_norm": 0.23205257952213287, "learning_rate": 9.36566759855801e-05, "loss": 0.0308, "step": 10200 }, { "action_loss": 0.014871266670525074, "epoch": 9.172661870503598, "step": 10200 }, { "epoch": 9.172661870503598, "step": 10200, "torque_loss": 0.19320599734783173 }, { "epoch": 9.181654676258994, "grad_norm": 0.23064212501049042, "learning_rate": 9.36405457314382e-05, "loss": 0.0327, "step": 10210 }, { "action_loss": 0.007223496213555336, "epoch": 9.181654676258994, "step": 10210 }, { "epoch": 9.181654676258994, "step": 10210, "torque_loss": 0.15896828472614288 }, { "epoch": 9.190647482014388, "grad_norm": 0.2588762044906616, "learning_rate": 9.36243963874389e-05, "loss": 0.0309, "step": 10220 }, { "action_loss": 0.01795051433146, "epoch": 9.190647482014388, "step": 10220 }, { "epoch": 9.190647482014388, "step": 10220, "torque_loss": 0.1948198825120926 }, { "epoch": 9.199640287769784, "grad_norm": 0.2969806492328644, "learning_rate": 9.360822796064655e-05, "loss": 0.0401, "step": 10230 }, { "action_loss": 0.0088943000882864, "epoch": 9.199640287769784, "step": 10230 }, { "epoch": 9.199640287769784, "step": 10230, "torque_loss": 0.11188807338476181 }, { "epoch": 9.20863309352518, "grad_norm": 0.30555981397628784, "learning_rate": 9.359204045813372e-05, "loss": 0.0356, "step": 10240 }, { "action_loss": 0.010719220153987408, "epoch": 9.20863309352518, "step": 10240 }, { "epoch": 9.20863309352518, "step": 10240, "torque_loss": 0.18189625442028046 }, { "epoch": 9.217625899280575, "grad_norm": 0.16889891028404236, "learning_rate": 9.357583388698141e-05, "loss": 0.0257, "step": 10250 }, { "action_loss": 0.012678106315433979, "epoch": 9.217625899280575, "step": 10250 }, { "epoch": 9.217625899280575, "step": 10250, "torque_loss": 0.17006880044937134 }, { "epoch": 9.226618705035971, "grad_norm": 0.18725800514221191, "learning_rate": 9.35596082542789e-05, "loss": 0.0314, "step": 10260 }, { "action_loss": 0.0055060312151908875, "epoch": 9.226618705035971, "step": 10260 }, { "epoch": 9.226618705035971, "step": 10260, "torque_loss": 0.17573018372058868 }, { "epoch": 9.235611510791367, "grad_norm": 0.23934952914714813, "learning_rate": 9.354336356712383e-05, "loss": 0.0278, "step": 10270 }, { "action_loss": 0.01660672388970852, "epoch": 9.235611510791367, "step": 10270 }, { "epoch": 9.235611510791367, "step": 10270, "torque_loss": 0.23869991302490234 }, { "epoch": 9.244604316546763, "grad_norm": 0.32584214210510254, "learning_rate": 9.35270998326222e-05, "loss": 0.0333, "step": 10280 }, { "action_loss": 0.0075563895516097546, "epoch": 9.244604316546763, "step": 10280 }, { "epoch": 9.244604316546763, "step": 10280, "torque_loss": 0.17103849351406097 }, { "epoch": 9.253597122302159, "grad_norm": 0.23612827062606812, "learning_rate": 9.351081705788831e-05, "loss": 0.0258, "step": 10290 }, { "action_loss": 0.013762940652668476, "epoch": 9.253597122302159, "step": 10290 }, { "epoch": 9.253597122302159, "step": 10290, "torque_loss": 0.25113368034362793 }, { "epoch": 9.262589928057555, "grad_norm": 0.24035349488258362, "learning_rate": 9.349451525004477e-05, "loss": 0.0355, "step": 10300 }, { "action_loss": 0.005381789058446884, "epoch": 9.262589928057555, "step": 10300 }, { "epoch": 9.262589928057555, "step": 10300, "torque_loss": 0.11768152564764023 }, { "epoch": 9.27158273381295, "grad_norm": 0.20524492859840393, "learning_rate": 9.347819441622261e-05, "loss": 0.0339, "step": 10310 }, { "action_loss": 0.020206434652209282, "epoch": 9.27158273381295, "step": 10310 }, { "epoch": 9.27158273381295, "step": 10310, "torque_loss": 0.18269865214824677 }, { "epoch": 9.280575539568344, "grad_norm": 0.25002044439315796, "learning_rate": 9.346185456356105e-05, "loss": 0.0367, "step": 10320 }, { "action_loss": 0.005240834783762693, "epoch": 9.280575539568344, "step": 10320 }, { "epoch": 9.280575539568344, "step": 10320, "torque_loss": 0.14054971933364868 }, { "epoch": 9.28956834532374, "grad_norm": 0.2139863818883896, "learning_rate": 9.344549569920774e-05, "loss": 0.0256, "step": 10330 }, { "action_loss": 0.0044484431855380535, "epoch": 9.28956834532374, "step": 10330 }, { "epoch": 9.28956834532374, "step": 10330, "torque_loss": 0.14488165080547333 }, { "epoch": 9.298561151079136, "grad_norm": 0.1985069215297699, "learning_rate": 9.342911783031858e-05, "loss": 0.027, "step": 10340 }, { "action_loss": 0.00493738055229187, "epoch": 9.298561151079136, "step": 10340 }, { "epoch": 9.298561151079136, "step": 10340, "torque_loss": 0.140042245388031 }, { "epoch": 9.307553956834532, "grad_norm": 0.19116157293319702, "learning_rate": 9.341272096405782e-05, "loss": 0.0313, "step": 10350 }, { "action_loss": 0.010305076837539673, "epoch": 9.307553956834532, "step": 10350 }, { "epoch": 9.307553956834532, "step": 10350, "torque_loss": 0.20405077934265137 }, { "epoch": 9.316546762589928, "grad_norm": 0.22001159191131592, "learning_rate": 9.3396305107598e-05, "loss": 0.033, "step": 10360 }, { "action_loss": 0.009003647603094578, "epoch": 9.316546762589928, "step": 10360 }, { "epoch": 9.316546762589928, "step": 10360, "torque_loss": 0.1769474893808365 }, { "epoch": 9.325539568345324, "grad_norm": 0.22024506330490112, "learning_rate": 9.337987026811998e-05, "loss": 0.0376, "step": 10370 }, { "action_loss": 0.04089425876736641, "epoch": 9.325539568345324, "step": 10370 }, { "epoch": 9.325539568345324, "step": 10370, "torque_loss": 0.2294379472732544 }, { "epoch": 9.33453237410072, "grad_norm": 0.1845734417438507, "learning_rate": 9.33634164528129e-05, "loss": 0.0377, "step": 10380 }, { "action_loss": 0.02686726301908493, "epoch": 9.33453237410072, "step": 10380 }, { "epoch": 9.33453237410072, "step": 10380, "torque_loss": 0.2332073450088501 }, { "epoch": 9.343525179856115, "grad_norm": 0.2563718259334564, "learning_rate": 9.334694366887424e-05, "loss": 0.0423, "step": 10390 }, { "action_loss": 0.014007225632667542, "epoch": 9.343525179856115, "step": 10390 }, { "epoch": 9.343525179856115, "step": 10390, "torque_loss": 0.19536931812763214 }, { "epoch": 9.352517985611511, "grad_norm": 0.22102369368076324, "learning_rate": 9.333045192350973e-05, "loss": 0.0343, "step": 10400 }, { "action_loss": 0.00820402055978775, "epoch": 9.352517985611511, "step": 10400 }, { "epoch": 9.352517985611511, "step": 10400, "torque_loss": 0.1395239681005478 }, { "epoch": 9.361510791366907, "grad_norm": 0.21081696450710297, "learning_rate": 9.331394122393345e-05, "loss": 0.034, "step": 10410 }, { "action_loss": 0.010164129547774792, "epoch": 9.361510791366907, "step": 10410 }, { "epoch": 9.361510791366907, "step": 10410, "torque_loss": 0.1435527354478836 }, { "epoch": 9.370503597122303, "grad_norm": 0.36962124705314636, "learning_rate": 9.329741157736771e-05, "loss": 0.0336, "step": 10420 }, { "action_loss": 0.029474372044205666, "epoch": 9.370503597122303, "step": 10420 }, { "epoch": 9.370503597122303, "step": 10420, "torque_loss": 0.2629508972167969 }, { "epoch": 9.379496402877697, "grad_norm": 0.23455674946308136, "learning_rate": 9.328086299104317e-05, "loss": 0.0336, "step": 10430 }, { "action_loss": 0.006527598947286606, "epoch": 9.379496402877697, "step": 10430 }, { "epoch": 9.379496402877697, "step": 10430, "torque_loss": 0.14954955875873566 }, { "epoch": 9.388489208633093, "grad_norm": 0.23919692635536194, "learning_rate": 9.326429547219872e-05, "loss": 0.0343, "step": 10440 }, { "action_loss": 0.013456483371555805, "epoch": 9.388489208633093, "step": 10440 }, { "epoch": 9.388489208633093, "step": 10440, "torque_loss": 0.21129775047302246 }, { "epoch": 9.397482014388489, "grad_norm": 0.16621363162994385, "learning_rate": 9.324770902808155e-05, "loss": 0.0347, "step": 10450 }, { "action_loss": 0.007199177052825689, "epoch": 9.397482014388489, "step": 10450 }, { "epoch": 9.397482014388489, "step": 10450, "torque_loss": 0.10497885942459106 }, { "epoch": 9.406474820143885, "grad_norm": 0.11329523473978043, "learning_rate": 9.323110366594717e-05, "loss": 0.0309, "step": 10460 }, { "action_loss": 0.013992187567055225, "epoch": 9.406474820143885, "step": 10460 }, { "epoch": 9.406474820143885, "step": 10460, "torque_loss": 0.17016947269439697 }, { "epoch": 9.41546762589928, "grad_norm": 0.11841721832752228, "learning_rate": 9.32144793930593e-05, "loss": 0.0293, "step": 10470 }, { "action_loss": 0.00928895641118288, "epoch": 9.41546762589928, "step": 10470 }, { "epoch": 9.41546762589928, "step": 10470, "torque_loss": 0.1760941594839096 }, { "epoch": 9.424460431654676, "grad_norm": 0.15925189852714539, "learning_rate": 9.319783621668996e-05, "loss": 0.0305, "step": 10480 }, { "action_loss": 0.004199052695184946, "epoch": 9.424460431654676, "step": 10480 }, { "epoch": 9.424460431654676, "step": 10480, "torque_loss": 0.1613842397928238 }, { "epoch": 9.433453237410072, "grad_norm": 0.1911863088607788, "learning_rate": 9.318117414411947e-05, "loss": 0.0268, "step": 10490 }, { "action_loss": 0.014030832797288895, "epoch": 9.433453237410072, "step": 10490 }, { "epoch": 9.433453237410072, "step": 10490, "torque_loss": 0.23953068256378174 }, { "epoch": 9.442446043165468, "grad_norm": 0.1596827208995819, "learning_rate": 9.316449318263635e-05, "loss": 0.0288, "step": 10500 }, { "action_loss": 0.007970732636749744, "epoch": 9.442446043165468, "step": 10500 }, { "epoch": 9.442446043165468, "step": 10500, "torque_loss": 0.16086392104625702 }, { "epoch": 9.451438848920864, "grad_norm": 0.25032564997673035, "learning_rate": 9.314779333953744e-05, "loss": 0.0334, "step": 10510 }, { "action_loss": 0.012625637464225292, "epoch": 9.451438848920864, "step": 10510 }, { "epoch": 9.451438848920864, "step": 10510, "torque_loss": 0.1949591189622879 }, { "epoch": 9.46043165467626, "grad_norm": 0.2120397537946701, "learning_rate": 9.313107462212781e-05, "loss": 0.0324, "step": 10520 }, { "action_loss": 0.01344725489616394, "epoch": 9.46043165467626, "step": 10520 }, { "epoch": 9.46043165467626, "step": 10520, "torque_loss": 0.24303068220615387 }, { "epoch": 9.469424460431656, "grad_norm": 0.30767256021499634, "learning_rate": 9.311433703772082e-05, "loss": 0.0403, "step": 10530 }, { "action_loss": 0.008041768334805965, "epoch": 9.469424460431656, "step": 10530 }, { "epoch": 9.469424460431656, "step": 10530, "torque_loss": 0.12805776298046112 }, { "epoch": 9.47841726618705, "grad_norm": 0.20313069224357605, "learning_rate": 9.3097580593638e-05, "loss": 0.0273, "step": 10540 }, { "action_loss": 0.013714556582272053, "epoch": 9.47841726618705, "step": 10540 }, { "epoch": 9.47841726618705, "step": 10540, "torque_loss": 0.24988853931427002 }, { "epoch": 9.487410071942445, "grad_norm": 0.23206846415996552, "learning_rate": 9.308080529720926e-05, "loss": 0.0385, "step": 10550 }, { "action_loss": 0.007039204239845276, "epoch": 9.487410071942445, "step": 10550 }, { "epoch": 9.487410071942445, "step": 10550, "torque_loss": 0.1845177412033081 }, { "epoch": 9.496402877697841, "grad_norm": 0.22799797356128693, "learning_rate": 9.306401115577264e-05, "loss": 0.0318, "step": 10560 }, { "action_loss": 0.023681076243519783, "epoch": 9.496402877697841, "step": 10560 }, { "epoch": 9.496402877697841, "step": 10560, "torque_loss": 0.2168584018945694 }, { "epoch": 9.505395683453237, "grad_norm": 0.18423861265182495, "learning_rate": 9.304719817667447e-05, "loss": 0.0359, "step": 10570 }, { "action_loss": 0.010229700244963169, "epoch": 9.505395683453237, "step": 10570 }, { "epoch": 9.505395683453237, "step": 10570, "torque_loss": 0.2204524129629135 }, { "epoch": 9.514388489208633, "grad_norm": 0.24747991561889648, "learning_rate": 9.303036636726934e-05, "loss": 0.0342, "step": 10580 }, { "action_loss": 0.010357875376939774, "epoch": 9.514388489208633, "step": 10580 }, { "epoch": 9.514388489208633, "step": 10580, "torque_loss": 0.16393837332725525 }, { "epoch": 9.523381294964029, "grad_norm": 0.23201262950897217, "learning_rate": 9.301351573492003e-05, "loss": 0.0301, "step": 10590 }, { "action_loss": 0.01952570676803589, "epoch": 9.523381294964029, "step": 10590 }, { "epoch": 9.523381294964029, "step": 10590, "torque_loss": 0.24653291702270508 }, { "epoch": 9.532374100719425, "grad_norm": 0.18701651692390442, "learning_rate": 9.299664628699758e-05, "loss": 0.0339, "step": 10600 }, { "action_loss": 0.00848198588937521, "epoch": 9.532374100719425, "step": 10600 }, { "epoch": 9.532374100719425, "step": 10600, "torque_loss": 0.19585375487804413 }, { "epoch": 9.54136690647482, "grad_norm": 0.31593310832977295, "learning_rate": 9.297975803088129e-05, "loss": 0.033, "step": 10610 }, { "action_loss": 0.016485029831528664, "epoch": 9.54136690647482, "step": 10610 }, { "epoch": 9.54136690647482, "step": 10610, "torque_loss": 0.24552643299102783 }, { "epoch": 9.550359712230216, "grad_norm": 0.18127229809761047, "learning_rate": 9.296285097395864e-05, "loss": 0.0296, "step": 10620 }, { "action_loss": 0.012261074967682362, "epoch": 9.550359712230216, "step": 10620 }, { "epoch": 9.550359712230216, "step": 10620, "torque_loss": 0.19605503976345062 }, { "epoch": 9.559352517985612, "grad_norm": 0.2619415819644928, "learning_rate": 9.294592512362533e-05, "loss": 0.0348, "step": 10630 }, { "action_loss": 0.009627426974475384, "epoch": 9.559352517985612, "step": 10630 }, { "epoch": 9.559352517985612, "step": 10630, "torque_loss": 0.1804397851228714 }, { "epoch": 9.568345323741006, "grad_norm": 0.17299653589725494, "learning_rate": 9.292898048728533e-05, "loss": 0.0303, "step": 10640 }, { "action_loss": 0.01931062340736389, "epoch": 9.568345323741006, "step": 10640 }, { "epoch": 9.568345323741006, "step": 10640, "torque_loss": 0.23825521767139435 }, { "epoch": 9.577338129496402, "grad_norm": 0.19766458868980408, "learning_rate": 9.29120170723508e-05, "loss": 0.0315, "step": 10650 }, { "action_loss": 0.010250705294311047, "epoch": 9.577338129496402, "step": 10650 }, { "epoch": 9.577338129496402, "step": 10650, "torque_loss": 0.20014826953411102 }, { "epoch": 9.586330935251798, "grad_norm": 0.18078845739364624, "learning_rate": 9.28950348862421e-05, "loss": 0.0281, "step": 10660 }, { "action_loss": 0.024618377909064293, "epoch": 9.586330935251798, "step": 10660 }, { "epoch": 9.586330935251798, "step": 10660, "torque_loss": 0.2373911291360855 }, { "epoch": 9.595323741007194, "grad_norm": 0.23547016084194183, "learning_rate": 9.287803393638781e-05, "loss": 0.0352, "step": 10670 }, { "action_loss": 0.010205190628767014, "epoch": 9.595323741007194, "step": 10670 }, { "epoch": 9.595323741007194, "step": 10670, "torque_loss": 0.15266333520412445 }, { "epoch": 9.60431654676259, "grad_norm": 0.29557761549949646, "learning_rate": 9.286101423022474e-05, "loss": 0.0315, "step": 10680 }, { "action_loss": 0.008626552298665047, "epoch": 9.60431654676259, "step": 10680 }, { "epoch": 9.60431654676259, "step": 10680, "torque_loss": 0.19111968576908112 }, { "epoch": 9.613309352517986, "grad_norm": 0.21164695918560028, "learning_rate": 9.284397577519788e-05, "loss": 0.0308, "step": 10690 }, { "action_loss": 0.010635477490723133, "epoch": 9.613309352517986, "step": 10690 }, { "epoch": 9.613309352517986, "step": 10690, "torque_loss": 0.15761934220790863 }, { "epoch": 9.622302158273381, "grad_norm": 0.32315072417259216, "learning_rate": 9.282691857876043e-05, "loss": 0.0323, "step": 10700 }, { "action_loss": 0.011383924633264542, "epoch": 9.622302158273381, "step": 10700 }, { "epoch": 9.622302158273381, "step": 10700, "torque_loss": 0.16746918857097626 }, { "epoch": 9.631294964028777, "grad_norm": 0.3276017904281616, "learning_rate": 9.280984264837377e-05, "loss": 0.0371, "step": 10710 }, { "action_loss": 0.014147710986435413, "epoch": 9.631294964028777, "step": 10710 }, { "epoch": 9.631294964028777, "step": 10710, "torque_loss": 0.19477863609790802 }, { "epoch": 9.640287769784173, "grad_norm": 0.20161785185337067, "learning_rate": 9.279274799150752e-05, "loss": 0.0316, "step": 10720 }, { "action_loss": 0.02210789918899536, "epoch": 9.640287769784173, "step": 10720 }, { "epoch": 9.640287769784173, "step": 10720, "torque_loss": 0.29860928654670715 }, { "epoch": 9.649280575539569, "grad_norm": 0.1988583654165268, "learning_rate": 9.277563461563945e-05, "loss": 0.0347, "step": 10730 }, { "action_loss": 0.013669158332049847, "epoch": 9.649280575539569, "step": 10730 }, { "epoch": 9.649280575539569, "step": 10730, "torque_loss": 0.20277631282806396 }, { "epoch": 9.658273381294965, "grad_norm": 0.19804486632347107, "learning_rate": 9.275850252825555e-05, "loss": 0.036, "step": 10740 }, { "action_loss": 0.005411358550190926, "epoch": 9.658273381294965, "step": 10740 }, { "epoch": 9.658273381294965, "step": 10740, "torque_loss": 0.1391807347536087 }, { "epoch": 9.667266187050359, "grad_norm": 0.2694489657878876, "learning_rate": 9.274135173684994e-05, "loss": 0.0322, "step": 10750 }, { "action_loss": 0.009477950632572174, "epoch": 9.667266187050359, "step": 10750 }, { "epoch": 9.667266187050359, "step": 10750, "torque_loss": 0.19637279212474823 }, { "epoch": 9.676258992805755, "grad_norm": 0.22720816731452942, "learning_rate": 9.272418224892498e-05, "loss": 0.0297, "step": 10760 }, { "action_loss": 0.010234690271317959, "epoch": 9.676258992805755, "step": 10760 }, { "epoch": 9.676258992805755, "step": 10760, "torque_loss": 0.21078014373779297 }, { "epoch": 9.68525179856115, "grad_norm": 0.2102486789226532, "learning_rate": 9.27069940719912e-05, "loss": 0.0303, "step": 10770 }, { "action_loss": 0.005710203666239977, "epoch": 9.68525179856115, "step": 10770 }, { "epoch": 9.68525179856115, "step": 10770, "torque_loss": 0.21761102974414825 }, { "epoch": 9.694244604316546, "grad_norm": 0.22987216711044312, "learning_rate": 9.268978721356727e-05, "loss": 0.0306, "step": 10780 }, { "action_loss": 0.006621500942856073, "epoch": 9.694244604316546, "step": 10780 }, { "epoch": 9.694244604316546, "step": 10780, "torque_loss": 0.1370505541563034 }, { "epoch": 9.703237410071942, "grad_norm": 0.134994775056839, "learning_rate": 9.267256168118008e-05, "loss": 0.0291, "step": 10790 }, { "action_loss": 0.04004029929637909, "epoch": 9.703237410071942, "step": 10790 }, { "epoch": 9.703237410071942, "step": 10790, "torque_loss": 0.32727450132369995 }, { "epoch": 9.712230215827338, "grad_norm": 0.22943344712257385, "learning_rate": 9.265531748236463e-05, "loss": 0.0363, "step": 10800 }, { "action_loss": 0.013966056518256664, "epoch": 9.712230215827338, "step": 10800 }, { "epoch": 9.712230215827338, "step": 10800, "torque_loss": 0.23224413394927979 }, { "epoch": 9.721223021582734, "grad_norm": 0.3392222225666046, "learning_rate": 9.263805462466416e-05, "loss": 0.0347, "step": 10810 }, { "action_loss": 0.014413051307201385, "epoch": 9.721223021582734, "step": 10810 }, { "epoch": 9.721223021582734, "step": 10810, "torque_loss": 0.192440465092659 }, { "epoch": 9.73021582733813, "grad_norm": 0.2733033001422882, "learning_rate": 9.262077311562998e-05, "loss": 0.0306, "step": 10820 }, { "action_loss": 0.020604193210601807, "epoch": 9.73021582733813, "step": 10820 }, { "epoch": 9.73021582733813, "step": 10820, "torque_loss": 0.1659264713525772 }, { "epoch": 9.739208633093526, "grad_norm": 0.20234142243862152, "learning_rate": 9.260347296282165e-05, "loss": 0.0312, "step": 10830 }, { "action_loss": 0.009529420174658298, "epoch": 9.739208633093526, "step": 10830 }, { "epoch": 9.739208633093526, "step": 10830, "torque_loss": 0.20799100399017334 }, { "epoch": 9.748201438848922, "grad_norm": 0.20966073870658875, "learning_rate": 9.258615417380683e-05, "loss": 0.0285, "step": 10840 }, { "action_loss": 0.013147074729204178, "epoch": 9.748201438848922, "step": 10840 }, { "epoch": 9.748201438848922, "step": 10840, "torque_loss": 0.19398315250873566 }, { "epoch": 9.757194244604317, "grad_norm": 0.2283264547586441, "learning_rate": 9.256881675616133e-05, "loss": 0.0326, "step": 10850 }, { "action_loss": 0.01806856319308281, "epoch": 9.757194244604317, "step": 10850 }, { "epoch": 9.757194244604317, "step": 10850, "torque_loss": 0.1984369307756424 }, { "epoch": 9.766187050359711, "grad_norm": 0.29245203733444214, "learning_rate": 9.255146071746917e-05, "loss": 0.0331, "step": 10860 }, { "action_loss": 0.009318777360022068, "epoch": 9.766187050359711, "step": 10860 }, { "epoch": 9.766187050359711, "step": 10860, "torque_loss": 0.16363315284252167 }, { "epoch": 9.775179856115107, "grad_norm": 0.23157228529453278, "learning_rate": 9.253408606532241e-05, "loss": 0.0271, "step": 10870 }, { "action_loss": 0.012517611496150494, "epoch": 9.775179856115107, "step": 10870 }, { "epoch": 9.775179856115107, "step": 10870, "torque_loss": 0.213754341006279 }, { "epoch": 9.784172661870503, "grad_norm": 0.24977344274520874, "learning_rate": 9.251669280732137e-05, "loss": 0.0326, "step": 10880 }, { "action_loss": 0.018543237820267677, "epoch": 9.784172661870503, "step": 10880 }, { "epoch": 9.784172661870503, "step": 10880, "torque_loss": 0.22417975962162018 }, { "epoch": 9.793165467625899, "grad_norm": 0.2554338276386261, "learning_rate": 9.249928095107441e-05, "loss": 0.0398, "step": 10890 }, { "action_loss": 0.006139622535556555, "epoch": 9.793165467625899, "step": 10890 }, { "epoch": 9.793165467625899, "step": 10890, "torque_loss": 0.19343920052051544 }, { "epoch": 9.802158273381295, "grad_norm": 0.16925951838493347, "learning_rate": 9.248185050419811e-05, "loss": 0.033, "step": 10900 }, { "action_loss": 0.010589093901216984, "epoch": 9.802158273381295, "step": 10900 }, { "epoch": 9.802158273381295, "step": 10900, "torque_loss": 0.173154816031456 }, { "epoch": 9.81115107913669, "grad_norm": 0.17546811699867249, "learning_rate": 9.24644014743171e-05, "loss": 0.0256, "step": 10910 }, { "action_loss": 0.013025929220020771, "epoch": 9.81115107913669, "step": 10910 }, { "epoch": 9.81115107913669, "step": 10910, "torque_loss": 0.25308966636657715 }, { "epoch": 9.820143884892087, "grad_norm": 0.19243353605270386, "learning_rate": 9.24469338690642e-05, "loss": 0.0332, "step": 10920 }, { "action_loss": 0.035935260355472565, "epoch": 9.820143884892087, "step": 10920 }, { "epoch": 9.820143884892087, "step": 10920, "torque_loss": 0.20603513717651367 }, { "epoch": 9.829136690647482, "grad_norm": 0.2178477793931961, "learning_rate": 9.242944769608033e-05, "loss": 0.0355, "step": 10930 }, { "action_loss": 0.01067620050162077, "epoch": 9.829136690647482, "step": 10930 }, { "epoch": 9.829136690647482, "step": 10930, "torque_loss": 0.2384815216064453 }, { "epoch": 9.838129496402878, "grad_norm": 0.17295458912849426, "learning_rate": 9.241194296301454e-05, "loss": 0.0308, "step": 10940 }, { "action_loss": 0.014240026473999023, "epoch": 9.838129496402878, "step": 10940 }, { "epoch": 9.838129496402878, "step": 10940, "torque_loss": 0.21927326917648315 }, { "epoch": 9.847122302158274, "grad_norm": 0.16965655982494354, "learning_rate": 9.239441967752397e-05, "loss": 0.026, "step": 10950 }, { "action_loss": 0.01024958398193121, "epoch": 9.847122302158274, "step": 10950 }, { "epoch": 9.847122302158274, "step": 10950, "torque_loss": 0.2021789401769638 }, { "epoch": 9.85611510791367, "grad_norm": 0.2175796777009964, "learning_rate": 9.237687784727393e-05, "loss": 0.0312, "step": 10960 }, { "action_loss": 0.020551374182105064, "epoch": 9.85611510791367, "step": 10960 }, { "epoch": 9.85611510791367, "step": 10960, "torque_loss": 0.21819107234477997 }, { "epoch": 9.865107913669064, "grad_norm": 0.23902353644371033, "learning_rate": 9.235931747993781e-05, "loss": 0.0305, "step": 10970 }, { "action_loss": 0.004576109815388918, "epoch": 9.865107913669064, "step": 10970 }, { "epoch": 9.865107913669064, "step": 10970, "torque_loss": 0.11481503397226334 }, { "epoch": 9.87410071942446, "grad_norm": 0.1859700083732605, "learning_rate": 9.234173858319707e-05, "loss": 0.0265, "step": 10980 }, { "action_loss": 0.01033331174403429, "epoch": 9.87410071942446, "step": 10980 }, { "epoch": 9.87410071942446, "step": 10980, "torque_loss": 0.16619426012039185 }, { "epoch": 9.883093525179856, "grad_norm": 0.19040335714817047, "learning_rate": 9.23241411647414e-05, "loss": 0.029, "step": 10990 }, { "action_loss": 0.005928190890699625, "epoch": 9.883093525179856, "step": 10990 }, { "epoch": 9.883093525179856, "step": 10990, "torque_loss": 0.17878150939941406 }, { "epoch": 9.892086330935252, "grad_norm": 0.2998950481414795, "learning_rate": 9.230652523226841e-05, "loss": 0.0294, "step": 11000 }, { "action_loss": 0.01300948765128851, "epoch": 9.892086330935252, "step": 11000 }, { "epoch": 9.892086330935252, "step": 11000, "torque_loss": 0.1847091168165207 }, { "epoch": 9.901079136690647, "grad_norm": 0.24902814626693726, "learning_rate": 9.2288890793484e-05, "loss": 0.0311, "step": 11010 }, { "action_loss": 0.015351121313869953, "epoch": 9.901079136690647, "step": 11010 }, { "epoch": 9.901079136690647, "step": 11010, "torque_loss": 0.1914665549993515 }, { "epoch": 9.910071942446043, "grad_norm": 0.23827841877937317, "learning_rate": 9.227123785610199e-05, "loss": 0.0303, "step": 11020 }, { "action_loss": 0.014660920016467571, "epoch": 9.910071942446043, "step": 11020 }, { "epoch": 9.910071942446043, "step": 11020, "torque_loss": 0.17296351492404938 }, { "epoch": 9.91906474820144, "grad_norm": 0.17522397637367249, "learning_rate": 9.225356642784443e-05, "loss": 0.0286, "step": 11030 }, { "action_loss": 0.009580276906490326, "epoch": 9.91906474820144, "step": 11030 }, { "epoch": 9.91906474820144, "step": 11030, "torque_loss": 0.15926691889762878 }, { "epoch": 9.928057553956835, "grad_norm": 0.25978922843933105, "learning_rate": 9.223587651644139e-05, "loss": 0.027, "step": 11040 }, { "action_loss": 0.015358668752014637, "epoch": 9.928057553956835, "step": 11040 }, { "epoch": 9.928057553956835, "step": 11040, "torque_loss": 0.183932363986969 }, { "epoch": 9.93705035971223, "grad_norm": 0.19201567769050598, "learning_rate": 9.221816812963104e-05, "loss": 0.0354, "step": 11050 }, { "action_loss": 0.01300725620239973, "epoch": 9.93705035971223, "step": 11050 }, { "epoch": 9.93705035971223, "step": 11050, "torque_loss": 0.20409636199474335 }, { "epoch": 9.946043165467627, "grad_norm": 0.22361761331558228, "learning_rate": 9.22004412751596e-05, "loss": 0.0398, "step": 11060 }, { "action_loss": 0.008597888052463531, "epoch": 9.946043165467627, "step": 11060 }, { "epoch": 9.946043165467627, "step": 11060, "torque_loss": 0.18423445522785187 }, { "epoch": 9.95503597122302, "grad_norm": 0.1839527040719986, "learning_rate": 9.218269596078146e-05, "loss": 0.033, "step": 11070 }, { "action_loss": 0.009964783675968647, "epoch": 9.95503597122302, "step": 11070 }, { "epoch": 9.95503597122302, "step": 11070, "torque_loss": 0.1689341813325882 }, { "epoch": 9.964028776978417, "grad_norm": 0.17804288864135742, "learning_rate": 9.216493219425895e-05, "loss": 0.025, "step": 11080 }, { "action_loss": 0.015768956393003464, "epoch": 9.964028776978417, "step": 11080 }, { "epoch": 9.964028776978417, "step": 11080, "torque_loss": 0.1511099636554718 }, { "epoch": 9.973021582733812, "grad_norm": 0.24265249073505402, "learning_rate": 9.21471499833626e-05, "loss": 0.026, "step": 11090 }, { "action_loss": 0.011338951997458935, "epoch": 9.973021582733812, "step": 11090 }, { "epoch": 9.973021582733812, "step": 11090, "torque_loss": 0.16293112933635712 }, { "epoch": 9.982014388489208, "grad_norm": 0.1529238075017929, "learning_rate": 9.212934933587094e-05, "loss": 0.028, "step": 11100 }, { "action_loss": 0.012545387260615826, "epoch": 9.982014388489208, "step": 11100 }, { "epoch": 9.982014388489208, "step": 11100, "torque_loss": 0.33018749952316284 }, { "epoch": 9.991007194244604, "grad_norm": 0.2609499990940094, "learning_rate": 9.211153025957056e-05, "loss": 0.033, "step": 11110 }, { "action_loss": 0.00816583912819624, "epoch": 9.991007194244604, "step": 11110 }, { "epoch": 9.991007194244604, "step": 11110, "torque_loss": 0.18148262798786163 }, { "epoch": 10.0, "grad_norm": 0.24665586650371552, "learning_rate": 9.209369276225614e-05, "loss": 0.032, "step": 11120 }, { "action_loss": 0.0073966034688055515, "epoch": 10.0, "step": 11120 }, { "epoch": 10.0, "step": 11120, "torque_loss": 0.11560068279504776 }, { "epoch": 10.008992805755396, "grad_norm": 0.2268957793712616, "learning_rate": 9.207583685173042e-05, "loss": 0.036, "step": 11130 }, { "action_loss": 0.005369236227124929, "epoch": 10.008992805755396, "step": 11130 }, { "epoch": 10.008992805755396, "step": 11130, "torque_loss": 0.17771124839782715 }, { "epoch": 10.017985611510792, "grad_norm": 0.26946526765823364, "learning_rate": 9.205796253580417e-05, "loss": 0.0284, "step": 11140 }, { "action_loss": 0.015743374824523926, "epoch": 10.017985611510792, "step": 11140 }, { "epoch": 10.017985611510792, "step": 11140, "torque_loss": 0.2233002632856369 }, { "epoch": 10.026978417266188, "grad_norm": 0.1742691546678543, "learning_rate": 9.204006982229621e-05, "loss": 0.03, "step": 11150 }, { "action_loss": 0.005234666634351015, "epoch": 10.026978417266188, "step": 11150 }, { "epoch": 10.026978417266188, "step": 11150, "torque_loss": 0.14537359774112701 }, { "epoch": 10.035971223021583, "grad_norm": 0.1902252733707428, "learning_rate": 9.202215871903346e-05, "loss": 0.0327, "step": 11160 }, { "action_loss": 0.0065641156397759914, "epoch": 10.035971223021583, "step": 11160 }, { "epoch": 10.035971223021583, "step": 11160, "torque_loss": 0.14304500818252563 }, { "epoch": 10.04496402877698, "grad_norm": 0.43965375423431396, "learning_rate": 9.20042292338508e-05, "loss": 0.0336, "step": 11170 }, { "action_loss": 0.005340630654245615, "epoch": 10.04496402877698, "step": 11170 }, { "epoch": 10.04496402877698, "step": 11170, "torque_loss": 0.16460628807544708 }, { "epoch": 10.053956834532373, "grad_norm": 0.16906937956809998, "learning_rate": 9.198628137459123e-05, "loss": 0.0317, "step": 11180 }, { "action_loss": 0.02517998218536377, "epoch": 10.053956834532373, "step": 11180 }, { "epoch": 10.053956834532373, "step": 11180, "torque_loss": 0.24255172908306122 }, { "epoch": 10.06294964028777, "grad_norm": 0.2363465130329132, "learning_rate": 9.196831514910572e-05, "loss": 0.0373, "step": 11190 }, { "action_loss": 0.0062674120999872684, "epoch": 10.06294964028777, "step": 11190 }, { "epoch": 10.06294964028777, "step": 11190, "torque_loss": 0.11568494886159897 }, { "epoch": 10.071942446043165, "grad_norm": 0.2829124629497528, "learning_rate": 9.195033056525332e-05, "loss": 0.0303, "step": 11200 }, { "action_loss": 0.015992294996976852, "epoch": 10.071942446043165, "step": 11200 }, { "epoch": 10.071942446043165, "step": 11200, "torque_loss": 0.21790771186351776 }, { "epoch": 10.08093525179856, "grad_norm": 0.3041268587112427, "learning_rate": 9.193232763090114e-05, "loss": 0.0426, "step": 11210 }, { "action_loss": 0.018581440672278404, "epoch": 10.08093525179856, "step": 11210 }, { "epoch": 10.08093525179856, "step": 11210, "torque_loss": 0.2233474999666214 }, { "epoch": 10.089928057553957, "grad_norm": 0.24881288409233093, "learning_rate": 9.191430635392422e-05, "loss": 0.0308, "step": 11220 }, { "action_loss": 0.018752897158265114, "epoch": 10.089928057553957, "step": 11220 }, { "epoch": 10.089928057553957, "step": 11220, "torque_loss": 0.20647652447223663 }, { "epoch": 10.098920863309353, "grad_norm": 0.3097681403160095, "learning_rate": 9.18962667422057e-05, "loss": 0.0349, "step": 11230 }, { "action_loss": 0.008549368008971214, "epoch": 10.098920863309353, "step": 11230 }, { "epoch": 10.098920863309353, "step": 11230, "torque_loss": 0.15989887714385986 }, { "epoch": 10.107913669064748, "grad_norm": 0.16858474910259247, "learning_rate": 9.187820880363671e-05, "loss": 0.0315, "step": 11240 }, { "action_loss": 0.015586517751216888, "epoch": 10.107913669064748, "step": 11240 }, { "epoch": 10.107913669064748, "step": 11240, "torque_loss": 0.24829596281051636 }, { "epoch": 10.116906474820144, "grad_norm": 0.23566663265228271, "learning_rate": 9.186013254611643e-05, "loss": 0.0301, "step": 11250 }, { "action_loss": 0.010969811119139194, "epoch": 10.116906474820144, "step": 11250 }, { "epoch": 10.116906474820144, "step": 11250, "torque_loss": 0.1609896719455719 }, { "epoch": 10.12589928057554, "grad_norm": 0.24014912545681, "learning_rate": 9.1842037977552e-05, "loss": 0.0269, "step": 11260 }, { "action_loss": 0.049974825233221054, "epoch": 10.12589928057554, "step": 11260 }, { "epoch": 10.12589928057554, "step": 11260, "torque_loss": 0.29085829854011536 }, { "epoch": 10.134892086330936, "grad_norm": 0.1743062287569046, "learning_rate": 9.182392510585862e-05, "loss": 0.0314, "step": 11270 }, { "action_loss": 0.028341123834252357, "epoch": 10.134892086330936, "step": 11270 }, { "epoch": 10.134892086330936, "step": 11270, "torque_loss": 0.16735957562923431 }, { "epoch": 10.14388489208633, "grad_norm": 0.2503890097141266, "learning_rate": 9.180579393895946e-05, "loss": 0.0401, "step": 11280 }, { "action_loss": 0.014138315804302692, "epoch": 10.14388489208633, "step": 11280 }, { "epoch": 10.14388489208633, "step": 11280, "torque_loss": 0.15501496195793152 }, { "epoch": 10.152877697841726, "grad_norm": 0.21406103670597076, "learning_rate": 9.178764448478572e-05, "loss": 0.0263, "step": 11290 }, { "action_loss": 0.013413398526608944, "epoch": 10.152877697841726, "step": 11290 }, { "epoch": 10.152877697841726, "step": 11290, "torque_loss": 0.19575028121471405 }, { "epoch": 10.161870503597122, "grad_norm": 0.2366679459810257, "learning_rate": 9.176947675127658e-05, "loss": 0.0292, "step": 11300 }, { "action_loss": 0.012792673893272877, "epoch": 10.161870503597122, "step": 11300 }, { "epoch": 10.161870503597122, "step": 11300, "torque_loss": 0.1665315181016922 }, { "epoch": 10.170863309352518, "grad_norm": 0.24282009899616241, "learning_rate": 9.175129074637924e-05, "loss": 0.0297, "step": 11310 }, { "action_loss": 0.004698914010077715, "epoch": 10.170863309352518, "step": 11310 }, { "epoch": 10.170863309352518, "step": 11310, "torque_loss": 0.11599647253751755 }, { "epoch": 10.179856115107913, "grad_norm": 0.1524960994720459, "learning_rate": 9.173308647804884e-05, "loss": 0.0361, "step": 11320 }, { "action_loss": 0.014696460217237473, "epoch": 10.179856115107913, "step": 11320 }, { "epoch": 10.179856115107913, "step": 11320, "torque_loss": 0.18439753353595734 }, { "epoch": 10.18884892086331, "grad_norm": 0.3444632589817047, "learning_rate": 9.171486395424859e-05, "loss": 0.0333, "step": 11330 }, { "action_loss": 0.004543851129710674, "epoch": 10.18884892086331, "step": 11330 }, { "epoch": 10.18884892086331, "step": 11330, "torque_loss": 0.14349831640720367 }, { "epoch": 10.197841726618705, "grad_norm": 0.15004302561283112, "learning_rate": 9.16966231829496e-05, "loss": 0.0303, "step": 11340 }, { "action_loss": 0.008103818632662296, "epoch": 10.197841726618705, "step": 11340 }, { "epoch": 10.197841726618705, "step": 11340, "torque_loss": 0.23614972829818726 }, { "epoch": 10.206834532374101, "grad_norm": 0.21022354066371918, "learning_rate": 9.167836417213105e-05, "loss": 0.0304, "step": 11350 }, { "action_loss": 0.015602094121277332, "epoch": 10.206834532374101, "step": 11350 }, { "epoch": 10.206834532374101, "step": 11350, "torque_loss": 0.20716679096221924 }, { "epoch": 10.215827338129497, "grad_norm": 0.20711909234523773, "learning_rate": 9.166008692978001e-05, "loss": 0.0319, "step": 11360 }, { "action_loss": 0.009044968523085117, "epoch": 10.215827338129497, "step": 11360 }, { "epoch": 10.215827338129497, "step": 11360, "torque_loss": 0.1473022699356079 }, { "epoch": 10.224820143884893, "grad_norm": 0.26360559463500977, "learning_rate": 9.164179146389158e-05, "loss": 0.0286, "step": 11370 }, { "action_loss": 0.005122673232108355, "epoch": 10.224820143884893, "step": 11370 }, { "epoch": 10.224820143884893, "step": 11370, "torque_loss": 0.19620276987552643 }, { "epoch": 10.233812949640289, "grad_norm": 0.3266734182834625, "learning_rate": 9.162347778246882e-05, "loss": 0.0387, "step": 11380 }, { "action_loss": 0.006670856848359108, "epoch": 10.233812949640289, "step": 11380 }, { "epoch": 10.233812949640289, "step": 11380, "torque_loss": 0.14149601757526398 }, { "epoch": 10.242805755395683, "grad_norm": 0.17200349271297455, "learning_rate": 9.160514589352276e-05, "loss": 0.0299, "step": 11390 }, { "action_loss": 0.004332894925028086, "epoch": 10.242805755395683, "step": 11390 }, { "epoch": 10.242805755395683, "step": 11390, "torque_loss": 0.10667292028665543 }, { "epoch": 10.251798561151078, "grad_norm": 0.1906873881816864, "learning_rate": 9.15867958050724e-05, "loss": 0.0243, "step": 11400 }, { "action_loss": 0.004334140568971634, "epoch": 10.251798561151078, "step": 11400 }, { "epoch": 10.251798561151078, "step": 11400, "torque_loss": 0.11579582840204239 }, { "epoch": 10.260791366906474, "grad_norm": 0.2574847638607025, "learning_rate": 9.156842752514466e-05, "loss": 0.0372, "step": 11410 }, { "action_loss": 0.005920920521020889, "epoch": 10.260791366906474, "step": 11410 }, { "epoch": 10.260791366906474, "step": 11410, "torque_loss": 0.1688777208328247 }, { "epoch": 10.26978417266187, "grad_norm": 0.2477315068244934, "learning_rate": 9.155004106177447e-05, "loss": 0.0339, "step": 11420 }, { "action_loss": 0.00458971643820405, "epoch": 10.26978417266187, "step": 11420 }, { "epoch": 10.26978417266187, "step": 11420, "torque_loss": 0.1281627118587494 }, { "epoch": 10.278776978417266, "grad_norm": 0.2616393566131592, "learning_rate": 9.153163642300471e-05, "loss": 0.0397, "step": 11430 }, { "action_loss": 0.009611072950065136, "epoch": 10.278776978417266, "step": 11430 }, { "epoch": 10.278776978417266, "step": 11430, "torque_loss": 0.20855343341827393 }, { "epoch": 10.287769784172662, "grad_norm": 0.21691273152828217, "learning_rate": 9.151321361688616e-05, "loss": 0.0354, "step": 11440 }, { "action_loss": 0.01678404211997986, "epoch": 10.287769784172662, "step": 11440 }, { "epoch": 10.287769784172662, "step": 11440, "torque_loss": 0.26270294189453125 }, { "epoch": 10.296762589928058, "grad_norm": 0.2666130065917969, "learning_rate": 9.149477265147762e-05, "loss": 0.0384, "step": 11450 }, { "action_loss": 0.010591189377009869, "epoch": 10.296762589928058, "step": 11450 }, { "epoch": 10.296762589928058, "step": 11450, "torque_loss": 0.1888049840927124 }, { "epoch": 10.305755395683454, "grad_norm": 0.3075964152812958, "learning_rate": 9.147631353484574e-05, "loss": 0.0331, "step": 11460 }, { "action_loss": 0.008815390057861805, "epoch": 10.305755395683454, "step": 11460 }, { "epoch": 10.305755395683454, "step": 11460, "torque_loss": 0.19690842926502228 }, { "epoch": 10.31474820143885, "grad_norm": 0.19090311229228973, "learning_rate": 9.145783627506522e-05, "loss": 0.0278, "step": 11470 }, { "action_loss": 0.017304787412285805, "epoch": 10.31474820143885, "step": 11470 }, { "epoch": 10.31474820143885, "step": 11470, "torque_loss": 0.2757319509983063 }, { "epoch": 10.323741007194245, "grad_norm": 0.157246395945549, "learning_rate": 9.143934088021861e-05, "loss": 0.0352, "step": 11480 }, { "action_loss": 0.0111984983086586, "epoch": 10.323741007194245, "step": 11480 }, { "epoch": 10.323741007194245, "step": 11480, "torque_loss": 0.18273812532424927 }, { "epoch": 10.332733812949641, "grad_norm": 0.2680390775203705, "learning_rate": 9.142082735839645e-05, "loss": 0.0303, "step": 11490 }, { "action_loss": 0.012378417886793613, "epoch": 10.332733812949641, "step": 11490 }, { "epoch": 10.332733812949641, "step": 11490, "torque_loss": 0.1873340755701065 }, { "epoch": 10.341726618705035, "grad_norm": 0.37505897879600525, "learning_rate": 9.140229571769715e-05, "loss": 0.0315, "step": 11500 }, { "action_loss": 0.006648375187069178, "epoch": 10.341726618705035, "step": 11500 }, { "epoch": 10.341726618705035, "step": 11500, "torque_loss": 0.19182133674621582 }, { "epoch": 10.350719424460431, "grad_norm": 0.25060340762138367, "learning_rate": 9.138374596622709e-05, "loss": 0.0284, "step": 11510 }, { "action_loss": 0.012444700114428997, "epoch": 10.350719424460431, "step": 11510 }, { "epoch": 10.350719424460431, "step": 11510, "torque_loss": 0.2819845974445343 }, { "epoch": 10.359712230215827, "grad_norm": 0.2614997625350952, "learning_rate": 9.136517811210059e-05, "loss": 0.034, "step": 11520 }, { "action_loss": 0.01818975992500782, "epoch": 10.359712230215827, "step": 11520 }, { "epoch": 10.359712230215827, "step": 11520, "torque_loss": 0.21032845973968506 }, { "epoch": 10.368705035971223, "grad_norm": 0.17489291727542877, "learning_rate": 9.134659216343984e-05, "loss": 0.034, "step": 11530 }, { "action_loss": 0.00534864654764533, "epoch": 10.368705035971223, "step": 11530 }, { "epoch": 10.368705035971223, "step": 11530, "torque_loss": 0.16666646301746368 }, { "epoch": 10.377697841726619, "grad_norm": 0.14208434522151947, "learning_rate": 9.132798812837494e-05, "loss": 0.0284, "step": 11540 }, { "action_loss": 0.013838149607181549, "epoch": 10.377697841726619, "step": 11540 }, { "epoch": 10.377697841726619, "step": 11540, "torque_loss": 0.2177634984254837 }, { "epoch": 10.386690647482014, "grad_norm": 0.22956329584121704, "learning_rate": 9.130936601504396e-05, "loss": 0.0313, "step": 11550 }, { "action_loss": 0.02291925996541977, "epoch": 10.386690647482014, "step": 11550 }, { "epoch": 10.386690647482014, "step": 11550, "torque_loss": 0.30214667320251465 }, { "epoch": 10.39568345323741, "grad_norm": 0.21822713315486908, "learning_rate": 9.129072583159284e-05, "loss": 0.034, "step": 11560 }, { "action_loss": 0.011776305735111237, "epoch": 10.39568345323741, "step": 11560 }, { "epoch": 10.39568345323741, "step": 11560, "torque_loss": 0.14217759668827057 }, { "epoch": 10.404676258992806, "grad_norm": 0.21938499808311462, "learning_rate": 9.127206758617542e-05, "loss": 0.0263, "step": 11570 }, { "action_loss": 0.04047005623579025, "epoch": 10.404676258992806, "step": 11570 }, { "epoch": 10.404676258992806, "step": 11570, "torque_loss": 0.19423367083072662 }, { "epoch": 10.413669064748202, "grad_norm": 0.13436774909496307, "learning_rate": 9.125339128695346e-05, "loss": 0.0331, "step": 11580 }, { "action_loss": 0.011626961641013622, "epoch": 10.413669064748202, "step": 11580 }, { "epoch": 10.413669064748202, "step": 11580, "torque_loss": 0.18577957153320312 }, { "epoch": 10.422661870503598, "grad_norm": 0.21789926290512085, "learning_rate": 9.123469694209659e-05, "loss": 0.0296, "step": 11590 }, { "action_loss": 0.011910882778465748, "epoch": 10.422661870503598, "step": 11590 }, { "epoch": 10.422661870503598, "step": 11590, "torque_loss": 0.22530417144298553 }, { "epoch": 10.431654676258994, "grad_norm": 0.2054622620344162, "learning_rate": 9.121598455978239e-05, "loss": 0.0274, "step": 11600 }, { "action_loss": 0.010285927914083004, "epoch": 10.431654676258994, "step": 11600 }, { "epoch": 10.431654676258994, "step": 11600, "torque_loss": 0.19846202433109283 }, { "epoch": 10.440647482014388, "grad_norm": 0.24638892710208893, "learning_rate": 9.119725414819624e-05, "loss": 0.0279, "step": 11610 }, { "action_loss": 0.009298774413764477, "epoch": 10.440647482014388, "step": 11610 }, { "epoch": 10.440647482014388, "step": 11610, "torque_loss": 0.18136537075042725 }, { "epoch": 10.449640287769784, "grad_norm": 0.20627032220363617, "learning_rate": 9.117850571553149e-05, "loss": 0.0289, "step": 11620 }, { "action_loss": 0.0115048261359334, "epoch": 10.449640287769784, "step": 11620 }, { "epoch": 10.449640287769784, "step": 11620, "torque_loss": 0.15564045310020447 }, { "epoch": 10.45863309352518, "grad_norm": 0.14988666772842407, "learning_rate": 9.115973926998935e-05, "loss": 0.0263, "step": 11630 }, { "action_loss": 0.004541305359452963, "epoch": 10.45863309352518, "step": 11630 }, { "epoch": 10.45863309352518, "step": 11630, "torque_loss": 0.16438166797161102 }, { "epoch": 10.467625899280575, "grad_norm": 0.19959604740142822, "learning_rate": 9.114095481977888e-05, "loss": 0.0304, "step": 11640 }, { "action_loss": 0.009829209186136723, "epoch": 10.467625899280575, "step": 11640 }, { "epoch": 10.467625899280575, "step": 11640, "torque_loss": 0.19216632843017578 }, { "epoch": 10.476618705035971, "grad_norm": 0.21539774537086487, "learning_rate": 9.112215237311703e-05, "loss": 0.0357, "step": 11650 }, { "action_loss": 0.010566015727818012, "epoch": 10.476618705035971, "step": 11650 }, { "epoch": 10.476618705035971, "step": 11650, "torque_loss": 0.20906563103199005 }, { "epoch": 10.485611510791367, "grad_norm": 0.2886096239089966, "learning_rate": 9.110333193822867e-05, "loss": 0.037, "step": 11660 }, { "action_loss": 0.009523848071694374, "epoch": 10.485611510791367, "step": 11660 }, { "epoch": 10.485611510791367, "step": 11660, "torque_loss": 0.17853958904743195 }, { "epoch": 10.494604316546763, "grad_norm": 0.2941327393054962, "learning_rate": 9.108449352334645e-05, "loss": 0.0363, "step": 11670 }, { "action_loss": 0.007248438894748688, "epoch": 10.494604316546763, "step": 11670 }, { "epoch": 10.494604316546763, "step": 11670, "torque_loss": 0.18417108058929443 }, { "epoch": 10.503597122302159, "grad_norm": 0.22126518189907074, "learning_rate": 9.106563713671094e-05, "loss": 0.034, "step": 11680 }, { "action_loss": 0.006717660930007696, "epoch": 10.503597122302159, "step": 11680 }, { "epoch": 10.503597122302159, "step": 11680, "torque_loss": 0.2794835865497589 }, { "epoch": 10.512589928057555, "grad_norm": 0.23414549231529236, "learning_rate": 9.104676278657061e-05, "loss": 0.0353, "step": 11690 }, { "action_loss": 0.010129923932254314, "epoch": 10.512589928057555, "step": 11690 }, { "epoch": 10.512589928057555, "step": 11690, "torque_loss": 0.23317675292491913 }, { "epoch": 10.52158273381295, "grad_norm": 0.15562132000923157, "learning_rate": 9.102787048118169e-05, "loss": 0.0369, "step": 11700 }, { "action_loss": 0.03635971620678902, "epoch": 10.52158273381295, "step": 11700 }, { "epoch": 10.52158273381295, "step": 11700, "torque_loss": 0.21885406970977783 }, { "epoch": 10.530575539568346, "grad_norm": 0.23160989582538605, "learning_rate": 9.100896022880834e-05, "loss": 0.0319, "step": 11710 }, { "action_loss": 0.005920311436057091, "epoch": 10.530575539568346, "step": 11710 }, { "epoch": 10.530575539568346, "step": 11710, "torque_loss": 0.221637561917305 }, { "epoch": 10.53956834532374, "grad_norm": 0.26543840765953064, "learning_rate": 9.099003203772254e-05, "loss": 0.0311, "step": 11720 }, { "action_loss": 0.019814683124423027, "epoch": 10.53956834532374, "step": 11720 }, { "epoch": 10.53956834532374, "step": 11720, "torque_loss": 0.15575270354747772 }, { "epoch": 10.548561151079136, "grad_norm": 0.1426060050725937, "learning_rate": 9.097108591620413e-05, "loss": 0.0311, "step": 11730 }, { "action_loss": 0.010805311612784863, "epoch": 10.548561151079136, "step": 11730 }, { "epoch": 10.548561151079136, "step": 11730, "torque_loss": 0.17313463985919952 }, { "epoch": 10.557553956834532, "grad_norm": 0.21247488260269165, "learning_rate": 9.095212187254078e-05, "loss": 0.0294, "step": 11740 }, { "action_loss": 0.028592268005013466, "epoch": 10.557553956834532, "step": 11740 }, { "epoch": 10.557553956834532, "step": 11740, "torque_loss": 0.2598828971385956 }, { "epoch": 10.566546762589928, "grad_norm": 0.21937918663024902, "learning_rate": 9.093313991502801e-05, "loss": 0.0389, "step": 11750 }, { "action_loss": 0.011208348907530308, "epoch": 10.566546762589928, "step": 11750 }, { "epoch": 10.566546762589928, "step": 11750, "torque_loss": 0.1866885870695114 }, { "epoch": 10.575539568345324, "grad_norm": 0.3010329604148865, "learning_rate": 9.091414005196917e-05, "loss": 0.0433, "step": 11760 }, { "action_loss": 0.013312087394297123, "epoch": 10.575539568345324, "step": 11760 }, { "epoch": 10.575539568345324, "step": 11760, "torque_loss": 0.2491127997636795 }, { "epoch": 10.58453237410072, "grad_norm": 0.1962144672870636, "learning_rate": 9.089512229167545e-05, "loss": 0.0323, "step": 11770 }, { "action_loss": 0.007089593913406134, "epoch": 10.58453237410072, "step": 11770 }, { "epoch": 10.58453237410072, "step": 11770, "torque_loss": 0.20661990344524384 }, { "epoch": 10.593525179856115, "grad_norm": 0.19528934359550476, "learning_rate": 9.087608664246587e-05, "loss": 0.0317, "step": 11780 }, { "action_loss": 0.008773746900260448, "epoch": 10.593525179856115, "step": 11780 }, { "epoch": 10.593525179856115, "step": 11780, "torque_loss": 0.1335018426179886 }, { "epoch": 10.602517985611511, "grad_norm": 0.28924569487571716, "learning_rate": 9.085703311266727e-05, "loss": 0.0289, "step": 11790 }, { "action_loss": 0.009674842469394207, "epoch": 10.602517985611511, "step": 11790 }, { "epoch": 10.602517985611511, "step": 11790, "torque_loss": 0.19239003956317902 }, { "epoch": 10.611510791366907, "grad_norm": 0.23402845859527588, "learning_rate": 9.083796171061429e-05, "loss": 0.0342, "step": 11800 }, { "action_loss": 0.03525560721755028, "epoch": 10.611510791366907, "step": 11800 }, { "epoch": 10.611510791366907, "step": 11800, "torque_loss": 0.2701485753059387 }, { "epoch": 10.620503597122303, "grad_norm": 0.16814948618412018, "learning_rate": 9.081887244464941e-05, "loss": 0.0314, "step": 11810 }, { "action_loss": 0.006032746750861406, "epoch": 10.620503597122303, "step": 11810 }, { "epoch": 10.620503597122303, "step": 11810, "torque_loss": 0.15590624511241913 }, { "epoch": 10.629496402877697, "grad_norm": 0.22721467912197113, "learning_rate": 9.079976532312297e-05, "loss": 0.0324, "step": 11820 }, { "action_loss": 0.006485124584287405, "epoch": 10.629496402877697, "step": 11820 }, { "epoch": 10.629496402877697, "step": 11820, "torque_loss": 0.12401092052459717 }, { "epoch": 10.638489208633093, "grad_norm": 0.1827595829963684, "learning_rate": 9.078064035439301e-05, "loss": 0.031, "step": 11830 }, { "action_loss": 0.009495007805526257, "epoch": 10.638489208633093, "step": 11830 }, { "epoch": 10.638489208633093, "step": 11830, "torque_loss": 0.2732420861721039 }, { "epoch": 10.647482014388489, "grad_norm": 0.21364907920360565, "learning_rate": 9.07614975468255e-05, "loss": 0.0351, "step": 11840 }, { "action_loss": 0.0068366252817213535, "epoch": 10.647482014388489, "step": 11840 }, { "epoch": 10.647482014388489, "step": 11840, "torque_loss": 0.10751303285360336 }, { "epoch": 10.656474820143885, "grad_norm": 0.31251001358032227, "learning_rate": 9.074233690879412e-05, "loss": 0.0311, "step": 11850 }, { "action_loss": 0.015909498557448387, "epoch": 10.656474820143885, "step": 11850 }, { "epoch": 10.656474820143885, "step": 11850, "torque_loss": 0.18710333108901978 }, { "epoch": 10.66546762589928, "grad_norm": 0.21151630580425262, "learning_rate": 9.072315844868038e-05, "loss": 0.0292, "step": 11860 }, { "action_loss": 0.019820602610707283, "epoch": 10.66546762589928, "step": 11860 }, { "epoch": 10.66546762589928, "step": 11860, "torque_loss": 0.21648059785366058 }, { "epoch": 10.674460431654676, "grad_norm": 0.13465149700641632, "learning_rate": 9.07039621748736e-05, "loss": 0.0315, "step": 11870 }, { "action_loss": 0.020386971533298492, "epoch": 10.674460431654676, "step": 11870 }, { "epoch": 10.674460431654676, "step": 11870, "torque_loss": 0.2298038750886917 }, { "epoch": 10.683453237410072, "grad_norm": 0.2024124711751938, "learning_rate": 9.06847480957709e-05, "loss": 0.0313, "step": 11880 }, { "action_loss": 0.01095876470208168, "epoch": 10.683453237410072, "step": 11880 }, { "epoch": 10.683453237410072, "step": 11880, "torque_loss": 0.1909073144197464 }, { "epoch": 10.692446043165468, "grad_norm": 0.19694097340106964, "learning_rate": 9.066551621977713e-05, "loss": 0.0301, "step": 11890 }, { "action_loss": 0.0034543685615062714, "epoch": 10.692446043165468, "step": 11890 }, { "epoch": 10.692446043165468, "step": 11890, "torque_loss": 0.14498786628246307 }, { "epoch": 10.701438848920864, "grad_norm": 0.2216789275407791, "learning_rate": 9.064626655530501e-05, "loss": 0.0304, "step": 11900 }, { "action_loss": 0.007404852658510208, "epoch": 10.701438848920864, "step": 11900 }, { "epoch": 10.701438848920864, "step": 11900, "torque_loss": 0.16433240473270416 }, { "epoch": 10.71043165467626, "grad_norm": 0.23905102908611298, "learning_rate": 9.062699911077497e-05, "loss": 0.0282, "step": 11910 }, { "action_loss": 0.005924290511757135, "epoch": 10.71043165467626, "step": 11910 }, { "epoch": 10.71043165467626, "step": 11910, "torque_loss": 0.18131816387176514 }, { "epoch": 10.719424460431654, "grad_norm": 0.13007010519504547, "learning_rate": 9.060771389461524e-05, "loss": 0.0295, "step": 11920 }, { "action_loss": 0.006051788572221994, "epoch": 10.719424460431654, "step": 11920 }, { "epoch": 10.719424460431654, "step": 11920, "torque_loss": 0.13918490707874298 }, { "epoch": 10.72841726618705, "grad_norm": 0.16358411312103271, "learning_rate": 9.058841091526187e-05, "loss": 0.0263, "step": 11930 }, { "action_loss": 0.009965035133063793, "epoch": 10.72841726618705, "step": 11930 }, { "epoch": 10.72841726618705, "step": 11930, "torque_loss": 0.2163391262292862 }, { "epoch": 10.737410071942445, "grad_norm": 0.14393463730812073, "learning_rate": 9.056909018115858e-05, "loss": 0.0285, "step": 11940 }, { "action_loss": 0.0032086323481053114, "epoch": 10.737410071942445, "step": 11940 }, { "epoch": 10.737410071942445, "step": 11940, "torque_loss": 0.13067378103733063 }, { "epoch": 10.746402877697841, "grad_norm": 0.17969419062137604, "learning_rate": 9.054975170075697e-05, "loss": 0.0282, "step": 11950 }, { "action_loss": 0.01721990667283535, "epoch": 10.746402877697841, "step": 11950 }, { "epoch": 10.746402877697841, "step": 11950, "torque_loss": 0.21659404039382935 }, { "epoch": 10.755395683453237, "grad_norm": 0.20930370688438416, "learning_rate": 9.053039548251631e-05, "loss": 0.0316, "step": 11960 }, { "action_loss": 0.00482887402176857, "epoch": 10.755395683453237, "step": 11960 }, { "epoch": 10.755395683453237, "step": 11960, "torque_loss": 0.1315644234418869 }, { "epoch": 10.764388489208633, "grad_norm": 0.16728225350379944, "learning_rate": 9.051102153490368e-05, "loss": 0.0248, "step": 11970 }, { "action_loss": 0.0062143378891050816, "epoch": 10.764388489208633, "step": 11970 }, { "epoch": 10.764388489208633, "step": 11970, "torque_loss": 0.18559300899505615 }, { "epoch": 10.773381294964029, "grad_norm": 0.1417398750782013, "learning_rate": 9.04916298663939e-05, "loss": 0.0291, "step": 11980 }, { "action_loss": 0.011587683111429214, "epoch": 10.773381294964029, "step": 11980 }, { "epoch": 10.773381294964029, "step": 11980, "torque_loss": 0.15461741387844086 }, { "epoch": 10.782374100719425, "grad_norm": 0.19082997739315033, "learning_rate": 9.047222048546955e-05, "loss": 0.0282, "step": 11990 }, { "action_loss": 0.015971291810274124, "epoch": 10.782374100719425, "step": 11990 }, { "epoch": 10.782374100719425, "step": 11990, "torque_loss": 0.24414849281311035 }, { "epoch": 10.79136690647482, "grad_norm": 0.15882135927677155, "learning_rate": 9.045279340062097e-05, "loss": 0.031, "step": 12000 }, { "action_loss": 0.005504039581865072, "epoch": 10.79136690647482, "step": 12000 }, { "epoch": 10.79136690647482, "step": 12000, "torque_loss": 0.16693031787872314 }, { "epoch": 10.800359712230216, "grad_norm": 0.21189679205417633, "learning_rate": 9.043334862034618e-05, "loss": 0.0261, "step": 12010 }, { "action_loss": 0.007999527268111706, "epoch": 10.800359712230216, "step": 12010 }, { "epoch": 10.800359712230216, "step": 12010, "torque_loss": 0.15788006782531738 }, { "epoch": 10.809352517985612, "grad_norm": 0.20057405531406403, "learning_rate": 9.041388615315102e-05, "loss": 0.03, "step": 12020 }, { "action_loss": 0.011706340126693249, "epoch": 10.809352517985612, "step": 12020 }, { "epoch": 10.809352517985612, "step": 12020, "torque_loss": 0.22653865814208984 }, { "epoch": 10.818345323741006, "grad_norm": 0.2308758795261383, "learning_rate": 9.039440600754905e-05, "loss": 0.0309, "step": 12030 }, { "action_loss": 0.01140350941568613, "epoch": 10.818345323741006, "step": 12030 }, { "epoch": 10.818345323741006, "step": 12030, "torque_loss": 0.1760893017053604 }, { "epoch": 10.827338129496402, "grad_norm": 0.2453826665878296, "learning_rate": 9.037490819206151e-05, "loss": 0.0278, "step": 12040 }, { "action_loss": 0.006058104801923037, "epoch": 10.827338129496402, "step": 12040 }, { "epoch": 10.827338129496402, "step": 12040, "torque_loss": 0.15073992311954498 }, { "epoch": 10.836330935251798, "grad_norm": 0.1952904313802719, "learning_rate": 9.035539271521744e-05, "loss": 0.0294, "step": 12050 }, { "action_loss": 0.010088863782584667, "epoch": 10.836330935251798, "step": 12050 }, { "epoch": 10.836330935251798, "step": 12050, "torque_loss": 0.1884118765592575 }, { "epoch": 10.845323741007194, "grad_norm": 0.15845289826393127, "learning_rate": 9.033585958555356e-05, "loss": 0.0243, "step": 12060 }, { "action_loss": 0.014636886306107044, "epoch": 10.845323741007194, "step": 12060 }, { "epoch": 10.845323741007194, "step": 12060, "torque_loss": 0.16515123844146729 }, { "epoch": 10.85431654676259, "grad_norm": 0.22495681047439575, "learning_rate": 9.031630881161431e-05, "loss": 0.034, "step": 12070 }, { "action_loss": 0.008091024123132229, "epoch": 10.85431654676259, "step": 12070 }, { "epoch": 10.85431654676259, "step": 12070, "torque_loss": 0.14628948271274567 }, { "epoch": 10.863309352517986, "grad_norm": 0.15040798485279083, "learning_rate": 9.029674040195186e-05, "loss": 0.0274, "step": 12080 }, { "action_loss": 0.006142331752926111, "epoch": 10.863309352517986, "step": 12080 }, { "epoch": 10.863309352517986, "step": 12080, "torque_loss": 0.2245601862668991 }, { "epoch": 10.872302158273381, "grad_norm": 0.20534174144268036, "learning_rate": 9.027715436512613e-05, "loss": 0.0267, "step": 12090 }, { "action_loss": 0.0030439055990427732, "epoch": 10.872302158273381, "step": 12090 }, { "epoch": 10.872302158273381, "step": 12090, "torque_loss": 0.10250061750411987 }, { "epoch": 10.881294964028777, "grad_norm": 0.15108062326908112, "learning_rate": 9.02575507097047e-05, "loss": 0.0248, "step": 12100 }, { "action_loss": 0.005271879490464926, "epoch": 10.881294964028777, "step": 12100 }, { "epoch": 10.881294964028777, "step": 12100, "torque_loss": 0.15467973053455353 }, { "epoch": 10.890287769784173, "grad_norm": 0.21634534001350403, "learning_rate": 9.023792944426286e-05, "loss": 0.0281, "step": 12110 }, { "action_loss": 0.012054219841957092, "epoch": 10.890287769784173, "step": 12110 }, { "epoch": 10.890287769784173, "step": 12110, "torque_loss": 0.19748163223266602 }, { "epoch": 10.899280575539569, "grad_norm": 0.14518967270851135, "learning_rate": 9.021829057738364e-05, "loss": 0.0301, "step": 12120 }, { "action_loss": 0.009313699789345264, "epoch": 10.899280575539569, "step": 12120 }, { "epoch": 10.899280575539569, "step": 12120, "torque_loss": 0.17234991490840912 }, { "epoch": 10.908273381294965, "grad_norm": 0.17789609730243683, "learning_rate": 9.019863411765775e-05, "loss": 0.0305, "step": 12130 }, { "action_loss": 0.017619656398892403, "epoch": 10.908273381294965, "step": 12130 }, { "epoch": 10.908273381294965, "step": 12130, "torque_loss": 0.2812999486923218 }, { "epoch": 10.917266187050359, "grad_norm": 0.207818865776062, "learning_rate": 9.017896007368357e-05, "loss": 0.0314, "step": 12140 }, { "action_loss": 0.006562154274433851, "epoch": 10.917266187050359, "step": 12140 }, { "epoch": 10.917266187050359, "step": 12140, "torque_loss": 0.13263390958309174 }, { "epoch": 10.926258992805755, "grad_norm": 0.09818840026855469, "learning_rate": 9.015926845406722e-05, "loss": 0.0332, "step": 12150 }, { "action_loss": 0.011802549473941326, "epoch": 10.926258992805755, "step": 12150 }, { "epoch": 10.926258992805755, "step": 12150, "torque_loss": 0.2240474373102188 }, { "epoch": 10.93525179856115, "grad_norm": 0.22229865193367004, "learning_rate": 9.013955926742245e-05, "loss": 0.0312, "step": 12160 }, { "action_loss": 0.01042988896369934, "epoch": 10.93525179856115, "step": 12160 }, { "epoch": 10.93525179856115, "step": 12160, "torque_loss": 0.2015032023191452 }, { "epoch": 10.944244604316546, "grad_norm": 0.1779651641845703, "learning_rate": 9.011983252237077e-05, "loss": 0.0273, "step": 12170 }, { "action_loss": 0.008561884984374046, "epoch": 10.944244604316546, "step": 12170 }, { "epoch": 10.944244604316546, "step": 12170, "torque_loss": 0.21692736446857452 }, { "epoch": 10.953237410071942, "grad_norm": 0.1727658212184906, "learning_rate": 9.01000882275413e-05, "loss": 0.0293, "step": 12180 }, { "action_loss": 0.012925028800964355, "epoch": 10.953237410071942, "step": 12180 }, { "epoch": 10.953237410071942, "step": 12180, "torque_loss": 0.21620295941829681 }, { "epoch": 10.962230215827338, "grad_norm": 0.1585124433040619, "learning_rate": 9.008032639157088e-05, "loss": 0.0249, "step": 12190 }, { "action_loss": 0.00497041642665863, "epoch": 10.962230215827338, "step": 12190 }, { "epoch": 10.962230215827338, "step": 12190, "torque_loss": 0.18305431306362152 }, { "epoch": 10.971223021582734, "grad_norm": 0.16875554621219635, "learning_rate": 9.006054702310401e-05, "loss": 0.028, "step": 12200 }, { "action_loss": 0.004173215012997389, "epoch": 10.971223021582734, "step": 12200 }, { "epoch": 10.971223021582734, "step": 12200, "torque_loss": 0.12373337894678116 }, { "epoch": 10.98021582733813, "grad_norm": 0.17881320416927338, "learning_rate": 9.004075013079283e-05, "loss": 0.0304, "step": 12210 }, { "action_loss": 0.018285296857357025, "epoch": 10.98021582733813, "step": 12210 }, { "epoch": 10.98021582733813, "step": 12210, "torque_loss": 0.2549385130405426 }, { "epoch": 10.989208633093526, "grad_norm": 0.15815487504005432, "learning_rate": 9.00209357232972e-05, "loss": 0.0277, "step": 12220 }, { "action_loss": 0.026053346693515778, "epoch": 10.989208633093526, "step": 12220 }, { "epoch": 10.989208633093526, "step": 12220, "torque_loss": 0.23926587402820587 }, { "epoch": 10.998201438848922, "grad_norm": 0.2309231460094452, "learning_rate": 9.000110380928461e-05, "loss": 0.0344, "step": 12230 }, { "action_loss": 0.0053879148326814175, "epoch": 10.998201438848922, "step": 12230 }, { "epoch": 10.998201438848922, "step": 12230, "torque_loss": 0.1339779496192932 }, { "epoch": 11.007194244604317, "grad_norm": 0.2882443368434906, "learning_rate": 8.998125439743021e-05, "loss": 0.0393, "step": 12240 }, { "action_loss": 0.010147511027753353, "epoch": 11.007194244604317, "step": 12240 }, { "epoch": 11.007194244604317, "step": 12240, "torque_loss": 0.1840289682149887 }, { "epoch": 11.016187050359711, "grad_norm": 0.2403547763824463, "learning_rate": 8.996138749641682e-05, "loss": 0.0386, "step": 12250 }, { "action_loss": 0.004677586257457733, "epoch": 11.016187050359711, "step": 12250 }, { "epoch": 11.016187050359711, "step": 12250, "torque_loss": 0.16063284873962402 }, { "epoch": 11.025179856115107, "grad_norm": 0.26001423597335815, "learning_rate": 8.994150311493488e-05, "loss": 0.0329, "step": 12260 }, { "action_loss": 0.009920960292220116, "epoch": 11.025179856115107, "step": 12260 }, { "epoch": 11.025179856115107, "step": 12260, "torque_loss": 0.15886269509792328 }, { "epoch": 11.034172661870503, "grad_norm": 0.21665340662002563, "learning_rate": 8.992160126168247e-05, "loss": 0.0269, "step": 12270 }, { "action_loss": 0.006042433902621269, "epoch": 11.034172661870503, "step": 12270 }, { "epoch": 11.034172661870503, "step": 12270, "torque_loss": 0.20556549727916718 }, { "epoch": 11.043165467625899, "grad_norm": 0.20047657191753387, "learning_rate": 8.99016819453654e-05, "loss": 0.0259, "step": 12280 }, { "action_loss": 0.009127664379775524, "epoch": 11.043165467625899, "step": 12280 }, { "epoch": 11.043165467625899, "step": 12280, "torque_loss": 0.19060200452804565 }, { "epoch": 11.052158273381295, "grad_norm": 0.19225309789180756, "learning_rate": 8.988174517469702e-05, "loss": 0.0269, "step": 12290 }, { "action_loss": 0.003619343740865588, "epoch": 11.052158273381295, "step": 12290 }, { "epoch": 11.052158273381295, "step": 12290, "torque_loss": 0.18174685537815094 }, { "epoch": 11.06115107913669, "grad_norm": 0.15561477839946747, "learning_rate": 8.986179095839835e-05, "loss": 0.0245, "step": 12300 }, { "action_loss": 0.003642042400315404, "epoch": 11.06115107913669, "step": 12300 }, { "epoch": 11.06115107913669, "step": 12300, "torque_loss": 0.14531786739826202 }, { "epoch": 11.070143884892087, "grad_norm": 0.21000909805297852, "learning_rate": 8.984181930519804e-05, "loss": 0.0275, "step": 12310 }, { "action_loss": 0.02838161773979664, "epoch": 11.070143884892087, "step": 12310 }, { "epoch": 11.070143884892087, "step": 12310, "torque_loss": 0.19220590591430664 }, { "epoch": 11.079136690647482, "grad_norm": 0.15154945850372314, "learning_rate": 8.982183022383237e-05, "loss": 0.0333, "step": 12320 }, { "action_loss": 0.00846054870635271, "epoch": 11.079136690647482, "step": 12320 }, { "epoch": 11.079136690647482, "step": 12320, "torque_loss": 0.14809934794902802 }, { "epoch": 11.088129496402878, "grad_norm": 0.14139828085899353, "learning_rate": 8.980182372304525e-05, "loss": 0.0288, "step": 12330 }, { "action_loss": 0.013421404175460339, "epoch": 11.088129496402878, "step": 12330 }, { "epoch": 11.088129496402878, "step": 12330, "torque_loss": 0.2041407972574234 }, { "epoch": 11.097122302158274, "grad_norm": 0.1950264573097229, "learning_rate": 8.97817998115882e-05, "loss": 0.0324, "step": 12340 }, { "action_loss": 0.020213937386870384, "epoch": 11.097122302158274, "step": 12340 }, { "epoch": 11.097122302158274, "step": 12340, "torque_loss": 0.2502474784851074 }, { "epoch": 11.10611510791367, "grad_norm": 0.20089569687843323, "learning_rate": 8.976175849822038e-05, "loss": 0.0345, "step": 12350 }, { "action_loss": 0.007261296734213829, "epoch": 11.10611510791367, "step": 12350 }, { "epoch": 11.10611510791367, "step": 12350, "torque_loss": 0.17184709012508392 }, { "epoch": 11.115107913669064, "grad_norm": 0.28043437004089355, "learning_rate": 8.97416997917085e-05, "loss": 0.0292, "step": 12360 }, { "action_loss": 0.010377935133874416, "epoch": 11.115107913669064, "step": 12360 }, { "epoch": 11.115107913669064, "step": 12360, "torque_loss": 0.19486767053604126 }, { "epoch": 11.12410071942446, "grad_norm": 0.14281968772411346, "learning_rate": 8.972162370082695e-05, "loss": 0.0329, "step": 12370 }, { "action_loss": 0.01857764460146427, "epoch": 11.12410071942446, "step": 12370 }, { "epoch": 11.12410071942446, "step": 12370, "torque_loss": 0.25543323159217834 }, { "epoch": 11.133093525179856, "grad_norm": 0.2670236825942993, "learning_rate": 8.97015302343577e-05, "loss": 0.0356, "step": 12380 }, { "action_loss": 0.012139253318309784, "epoch": 11.133093525179856, "step": 12380 }, { "epoch": 11.133093525179856, "step": 12380, "torque_loss": 0.2430310845375061 }, { "epoch": 11.142086330935252, "grad_norm": 0.2232159674167633, "learning_rate": 8.968141940109027e-05, "loss": 0.0357, "step": 12390 }, { "action_loss": 0.005596252158284187, "epoch": 11.142086330935252, "step": 12390 }, { "epoch": 11.142086330935252, "step": 12390, "torque_loss": 0.15600310266017914 }, { "epoch": 11.151079136690647, "grad_norm": 0.1737404465675354, "learning_rate": 8.966129120982188e-05, "loss": 0.0257, "step": 12400 }, { "action_loss": 0.007898437790572643, "epoch": 11.151079136690647, "step": 12400 }, { "epoch": 11.151079136690647, "step": 12400, "torque_loss": 0.18580566346645355 }, { "epoch": 11.160071942446043, "grad_norm": 0.17360268533229828, "learning_rate": 8.964114566935724e-05, "loss": 0.0288, "step": 12410 }, { "action_loss": 0.014894269406795502, "epoch": 11.160071942446043, "step": 12410 }, { "epoch": 11.160071942446043, "step": 12410, "torque_loss": 0.2016017585992813 }, { "epoch": 11.16906474820144, "grad_norm": 0.20608431100845337, "learning_rate": 8.962098278850871e-05, "loss": 0.0293, "step": 12420 }, { "action_loss": 0.016461726278066635, "epoch": 11.16906474820144, "step": 12420 }, { "epoch": 11.16906474820144, "step": 12420, "torque_loss": 0.23517580330371857 }, { "epoch": 11.178057553956835, "grad_norm": 0.2372955083847046, "learning_rate": 8.960080257609622e-05, "loss": 0.026, "step": 12430 }, { "action_loss": 0.006553641054779291, "epoch": 11.178057553956835, "step": 12430 }, { "epoch": 11.178057553956835, "step": 12430, "torque_loss": 0.1827816367149353 }, { "epoch": 11.18705035971223, "grad_norm": 0.20906750857830048, "learning_rate": 8.95806050409473e-05, "loss": 0.0325, "step": 12440 }, { "action_loss": 0.01918232999742031, "epoch": 11.18705035971223, "step": 12440 }, { "epoch": 11.18705035971223, "step": 12440, "torque_loss": 0.1909012645483017 }, { "epoch": 11.196043165467627, "grad_norm": 0.3020627498626709, "learning_rate": 8.9560390191897e-05, "loss": 0.0337, "step": 12450 }, { "action_loss": 0.006363318767398596, "epoch": 11.196043165467627, "step": 12450 }, { "epoch": 11.196043165467627, "step": 12450, "torque_loss": 0.19660596549510956 }, { "epoch": 11.20503597122302, "grad_norm": 0.25864332914352417, "learning_rate": 8.9540158037788e-05, "loss": 0.0268, "step": 12460 }, { "action_loss": 0.006315045524388552, "epoch": 11.20503597122302, "step": 12460 }, { "epoch": 11.20503597122302, "step": 12460, "torque_loss": 0.1681615710258484 }, { "epoch": 11.214028776978417, "grad_norm": 0.13613726198673248, "learning_rate": 8.951990858747054e-05, "loss": 0.0284, "step": 12470 }, { "action_loss": 0.014465626329183578, "epoch": 11.214028776978417, "step": 12470 }, { "epoch": 11.214028776978417, "step": 12470, "torque_loss": 0.13783244788646698 }, { "epoch": 11.223021582733812, "grad_norm": 0.3034631311893463, "learning_rate": 8.94996418498024e-05, "loss": 0.0317, "step": 12480 }, { "action_loss": 0.0052797808311879635, "epoch": 11.223021582733812, "step": 12480 }, { "epoch": 11.223021582733812, "step": 12480, "torque_loss": 0.13776862621307373 }, { "epoch": 11.232014388489208, "grad_norm": 0.1802556961774826, "learning_rate": 8.947935783364896e-05, "loss": 0.0282, "step": 12490 }, { "action_loss": 0.016093432903289795, "epoch": 11.232014388489208, "step": 12490 }, { "epoch": 11.232014388489208, "step": 12490, "torque_loss": 0.19321735203266144 }, { "epoch": 11.241007194244604, "grad_norm": 0.2624744176864624, "learning_rate": 8.945905654788311e-05, "loss": 0.0358, "step": 12500 }, { "action_loss": 0.022084886208176613, "epoch": 11.241007194244604, "step": 12500 }, { "epoch": 11.241007194244604, "step": 12500, "torque_loss": 0.19918997585773468 }, { "epoch": 11.25, "grad_norm": 0.28228986263275146, "learning_rate": 8.943873800138535e-05, "loss": 0.0306, "step": 12510 }, { "action_loss": 0.011752462945878506, "epoch": 11.25, "step": 12510 }, { "epoch": 11.25, "step": 12510, "torque_loss": 0.2081861048936844 }, { "epoch": 11.258992805755396, "grad_norm": 0.18978872895240784, "learning_rate": 8.94184022030437e-05, "loss": 0.0333, "step": 12520 }, { "action_loss": 0.012568988837301731, "epoch": 11.258992805755396, "step": 12520 }, { "epoch": 11.258992805755396, "step": 12520, "torque_loss": 0.18206687271595 }, { "epoch": 11.267985611510792, "grad_norm": 0.19253014028072357, "learning_rate": 8.939804916175372e-05, "loss": 0.0318, "step": 12530 }, { "action_loss": 0.005972782615572214, "epoch": 11.267985611510792, "step": 12530 }, { "epoch": 11.267985611510792, "step": 12530, "torque_loss": 0.10189008712768555 }, { "epoch": 11.276978417266188, "grad_norm": 0.1603737324476242, "learning_rate": 8.93776788864185e-05, "loss": 0.0292, "step": 12540 }, { "action_loss": 0.01510569453239441, "epoch": 11.276978417266188, "step": 12540 }, { "epoch": 11.276978417266188, "step": 12540, "torque_loss": 0.22126738727092743 }, { "epoch": 11.285971223021583, "grad_norm": 0.19593943655490875, "learning_rate": 8.935729138594873e-05, "loss": 0.0348, "step": 12550 }, { "action_loss": 0.01672619953751564, "epoch": 11.285971223021583, "step": 12550 }, { "epoch": 11.285971223021583, "step": 12550, "torque_loss": 0.2160668522119522 }, { "epoch": 11.29496402877698, "grad_norm": 0.1976185292005539, "learning_rate": 8.933688666926258e-05, "loss": 0.0301, "step": 12560 }, { "action_loss": 0.00834492314606905, "epoch": 11.29496402877698, "step": 12560 }, { "epoch": 11.29496402877698, "step": 12560, "torque_loss": 0.2060023695230484 }, { "epoch": 11.303956834532373, "grad_norm": 0.31694531440734863, "learning_rate": 8.931646474528575e-05, "loss": 0.0384, "step": 12570 }, { "action_loss": 0.020635908469557762, "epoch": 11.303956834532373, "step": 12570 }, { "epoch": 11.303956834532373, "step": 12570, "torque_loss": 0.327355295419693 }, { "epoch": 11.31294964028777, "grad_norm": 0.28113749623298645, "learning_rate": 8.929602562295151e-05, "loss": 0.0336, "step": 12580 }, { "action_loss": 0.00551646389067173, "epoch": 11.31294964028777, "step": 12580 }, { "epoch": 11.31294964028777, "step": 12580, "torque_loss": 0.16490690410137177 }, { "epoch": 11.321942446043165, "grad_norm": 0.2821654975414276, "learning_rate": 8.92755693112006e-05, "loss": 0.0296, "step": 12590 }, { "action_loss": 0.015587673522531986, "epoch": 11.321942446043165, "step": 12590 }, { "epoch": 11.321942446043165, "step": 12590, "torque_loss": 0.25854283571243286 }, { "epoch": 11.33093525179856, "grad_norm": 0.15455025434494019, "learning_rate": 8.925509581898136e-05, "loss": 0.0301, "step": 12600 }, { "action_loss": 0.004966836888343096, "epoch": 11.33093525179856, "step": 12600 }, { "epoch": 11.33093525179856, "step": 12600, "torque_loss": 0.11901575326919556 }, { "epoch": 11.339928057553957, "grad_norm": 0.1357034593820572, "learning_rate": 8.923460515524951e-05, "loss": 0.0285, "step": 12610 }, { "action_loss": 0.00818824302405119, "epoch": 11.339928057553957, "step": 12610 }, { "epoch": 11.339928057553957, "step": 12610, "torque_loss": 0.1897115707397461 }, { "epoch": 11.348920863309353, "grad_norm": 0.1955556571483612, "learning_rate": 8.921409732896842e-05, "loss": 0.0307, "step": 12620 }, { "action_loss": 0.009309977293014526, "epoch": 11.348920863309353, "step": 12620 }, { "epoch": 11.348920863309353, "step": 12620, "torque_loss": 0.15425772964954376 }, { "epoch": 11.357913669064748, "grad_norm": 0.14854736626148224, "learning_rate": 8.919357234910887e-05, "loss": 0.0246, "step": 12630 }, { "action_loss": 0.00890857819467783, "epoch": 11.357913669064748, "step": 12630 }, { "epoch": 11.357913669064748, "step": 12630, "torque_loss": 0.16162316501140594 }, { "epoch": 11.366906474820144, "grad_norm": 0.17914903163909912, "learning_rate": 8.917303022464923e-05, "loss": 0.0311, "step": 12640 }, { "action_loss": 0.009389854967594147, "epoch": 11.366906474820144, "step": 12640 }, { "epoch": 11.366906474820144, "step": 12640, "torque_loss": 0.2158380001783371 }, { "epoch": 11.37589928057554, "grad_norm": 0.16493237018585205, "learning_rate": 8.915247096457531e-05, "loss": 0.0302, "step": 12650 }, { "action_loss": 0.008026530034840107, "epoch": 11.37589928057554, "step": 12650 }, { "epoch": 11.37589928057554, "step": 12650, "torque_loss": 0.14493274688720703 }, { "epoch": 11.384892086330936, "grad_norm": 0.20292361080646515, "learning_rate": 8.91318945778804e-05, "loss": 0.028, "step": 12660 }, { "action_loss": 0.02778881974518299, "epoch": 11.384892086330936, "step": 12660 }, { "epoch": 11.384892086330936, "step": 12660, "torque_loss": 0.2293180227279663 }, { "epoch": 11.39388489208633, "grad_norm": 0.1928991973400116, "learning_rate": 8.911130107356534e-05, "loss": 0.0303, "step": 12670 }, { "action_loss": 0.003207719186320901, "epoch": 11.39388489208633, "step": 12670 }, { "epoch": 11.39388489208633, "step": 12670, "torque_loss": 0.09077239781618118 }, { "epoch": 11.402877697841726, "grad_norm": 0.2781136631965637, "learning_rate": 8.90906904606384e-05, "loss": 0.0272, "step": 12680 }, { "action_loss": 0.007122432347387075, "epoch": 11.402877697841726, "step": 12680 }, { "epoch": 11.402877697841726, "step": 12680, "torque_loss": 0.1642494648694992 }, { "epoch": 11.411870503597122, "grad_norm": 0.16621044278144836, "learning_rate": 8.90700627481154e-05, "loss": 0.0264, "step": 12690 }, { "action_loss": 0.006922371685504913, "epoch": 11.411870503597122, "step": 12690 }, { "epoch": 11.411870503597122, "step": 12690, "torque_loss": 0.13141286373138428 }, { "epoch": 11.420863309352518, "grad_norm": 0.2633991241455078, "learning_rate": 8.904941794501957e-05, "loss": 0.0256, "step": 12700 }, { "action_loss": 0.0073656849563121796, "epoch": 11.420863309352518, "step": 12700 }, { "epoch": 11.420863309352518, "step": 12700, "torque_loss": 0.182292178273201 }, { "epoch": 11.429856115107913, "grad_norm": 0.2301812320947647, "learning_rate": 8.902875606038166e-05, "loss": 0.0273, "step": 12710 }, { "action_loss": 0.009195040911436081, "epoch": 11.429856115107913, "step": 12710 }, { "epoch": 11.429856115107913, "step": 12710, "torque_loss": 0.19657953083515167 }, { "epoch": 11.43884892086331, "grad_norm": 0.16223342716693878, "learning_rate": 8.900807710323989e-05, "loss": 0.0337, "step": 12720 }, { "action_loss": 0.039480868726968765, "epoch": 11.43884892086331, "step": 12720 }, { "epoch": 11.43884892086331, "step": 12720, "torque_loss": 0.17983682453632355 }, { "epoch": 11.447841726618705, "grad_norm": 0.19976426661014557, "learning_rate": 8.898738108263993e-05, "loss": 0.0311, "step": 12730 }, { "action_loss": 0.004560575354844332, "epoch": 11.447841726618705, "step": 12730 }, { "epoch": 11.447841726618705, "step": 12730, "torque_loss": 0.07577246427536011 }, { "epoch": 11.456834532374101, "grad_norm": 0.234747976064682, "learning_rate": 8.896666800763491e-05, "loss": 0.0322, "step": 12740 }, { "action_loss": 0.009100507013499737, "epoch": 11.456834532374101, "step": 12740 }, { "epoch": 11.456834532374101, "step": 12740, "torque_loss": 0.1312837153673172 }, { "epoch": 11.465827338129497, "grad_norm": 0.17032139003276825, "learning_rate": 8.894593788728546e-05, "loss": 0.0288, "step": 12750 }, { "action_loss": 0.009091665036976337, "epoch": 11.465827338129497, "step": 12750 }, { "epoch": 11.465827338129497, "step": 12750, "torque_loss": 0.15731678903102875 }, { "epoch": 11.474820143884893, "grad_norm": 0.16683335602283478, "learning_rate": 8.892519073065961e-05, "loss": 0.0281, "step": 12760 }, { "action_loss": 0.009793524630367756, "epoch": 11.474820143884893, "step": 12760 }, { "epoch": 11.474820143884893, "step": 12760, "torque_loss": 0.2169938087463379 }, { "epoch": 11.483812949640289, "grad_norm": 0.2066175788640976, "learning_rate": 8.89044265468329e-05, "loss": 0.0319, "step": 12770 }, { "action_loss": 0.006877157837152481, "epoch": 11.483812949640289, "step": 12770 }, { "epoch": 11.483812949640289, "step": 12770, "torque_loss": 0.1776363104581833 }, { "epoch": 11.492805755395683, "grad_norm": 0.24511893093585968, "learning_rate": 8.888364534488827e-05, "loss": 0.0302, "step": 12780 }, { "action_loss": 0.015540887601673603, "epoch": 11.492805755395683, "step": 12780 }, { "epoch": 11.492805755395683, "step": 12780, "torque_loss": 0.25520047545433044 }, { "epoch": 11.501798561151078, "grad_norm": 0.18352137506008148, "learning_rate": 8.886284713391613e-05, "loss": 0.0299, "step": 12790 }, { "action_loss": 0.014775601215660572, "epoch": 11.501798561151078, "step": 12790 }, { "epoch": 11.501798561151078, "step": 12790, "torque_loss": 0.20654095709323883 }, { "epoch": 11.510791366906474, "grad_norm": 0.20450425148010254, "learning_rate": 8.884203192301431e-05, "loss": 0.0292, "step": 12800 }, { "action_loss": 0.004174581728875637, "epoch": 11.510791366906474, "step": 12800 }, { "epoch": 11.510791366906474, "step": 12800, "torque_loss": 0.10478103160858154 }, { "epoch": 11.51978417266187, "grad_norm": 0.2365919053554535, "learning_rate": 8.88211997212881e-05, "loss": 0.0331, "step": 12810 }, { "action_loss": 0.006401780992746353, "epoch": 11.51978417266187, "step": 12810 }, { "epoch": 11.51978417266187, "step": 12810, "torque_loss": 0.1361834853887558 }, { "epoch": 11.528776978417266, "grad_norm": 0.1701410859823227, "learning_rate": 8.880035053785023e-05, "loss": 0.03, "step": 12820 }, { "action_loss": 0.005734918639063835, "epoch": 11.528776978417266, "step": 12820 }, { "epoch": 11.528776978417266, "step": 12820, "torque_loss": 0.1156456246972084 }, { "epoch": 11.537769784172662, "grad_norm": 0.16715745627880096, "learning_rate": 8.877948438182083e-05, "loss": 0.0269, "step": 12830 }, { "action_loss": 0.018632911145687103, "epoch": 11.537769784172662, "step": 12830 }, { "epoch": 11.537769784172662, "step": 12830, "torque_loss": 0.24985499680042267 }, { "epoch": 11.546762589928058, "grad_norm": 0.2820403575897217, "learning_rate": 8.875860126232745e-05, "loss": 0.0339, "step": 12840 }, { "action_loss": 0.006001768633723259, "epoch": 11.546762589928058, "step": 12840 }, { "epoch": 11.546762589928058, "step": 12840, "torque_loss": 0.10628543049097061 }, { "epoch": 11.555755395683454, "grad_norm": 0.23473921418190002, "learning_rate": 8.87377011885051e-05, "loss": 0.027, "step": 12850 }, { "action_loss": 0.012101888656616211, "epoch": 11.555755395683454, "step": 12850 }, { "epoch": 11.555755395683454, "step": 12850, "torque_loss": 0.22083406150341034 }, { "epoch": 11.56474820143885, "grad_norm": 0.18862634897232056, "learning_rate": 8.871678416949617e-05, "loss": 0.0279, "step": 12860 }, { "action_loss": 0.01664895750582218, "epoch": 11.56474820143885, "step": 12860 }, { "epoch": 11.56474820143885, "step": 12860, "torque_loss": 0.268710732460022 }, { "epoch": 11.573741007194245, "grad_norm": 0.18011480569839478, "learning_rate": 8.869585021445046e-05, "loss": 0.0324, "step": 12870 }, { "action_loss": 0.0065832785330712795, "epoch": 11.573741007194245, "step": 12870 }, { "epoch": 11.573741007194245, "step": 12870, "torque_loss": 0.1165023148059845 }, { "epoch": 11.582733812949641, "grad_norm": 0.2230781763792038, "learning_rate": 8.867489933252521e-05, "loss": 0.03, "step": 12880 }, { "action_loss": 0.011390864849090576, "epoch": 11.582733812949641, "step": 12880 }, { "epoch": 11.582733812949641, "step": 12880, "torque_loss": 0.21631748974323273 }, { "epoch": 11.591726618705035, "grad_norm": 0.17765098810195923, "learning_rate": 8.865393153288504e-05, "loss": 0.0248, "step": 12890 }, { "action_loss": 0.00843681488186121, "epoch": 11.591726618705035, "step": 12890 }, { "epoch": 11.591726618705035, "step": 12890, "torque_loss": 0.1680113524198532 }, { "epoch": 11.600719424460431, "grad_norm": 0.11012658476829529, "learning_rate": 8.8632946824702e-05, "loss": 0.0261, "step": 12900 }, { "action_loss": 0.009584318846464157, "epoch": 11.600719424460431, "step": 12900 }, { "epoch": 11.600719424460431, "step": 12900, "torque_loss": 0.1804989129304886 }, { "epoch": 11.609712230215827, "grad_norm": 0.15850213170051575, "learning_rate": 8.86119452171555e-05, "loss": 0.0248, "step": 12910 }, { "action_loss": 0.011187344789505005, "epoch": 11.609712230215827, "step": 12910 }, { "epoch": 11.609712230215827, "step": 12910, "torque_loss": 0.18966734409332275 }, { "epoch": 11.618705035971223, "grad_norm": 0.15422673523426056, "learning_rate": 8.859092671943234e-05, "loss": 0.0286, "step": 12920 }, { "action_loss": 0.00873828399926424, "epoch": 11.618705035971223, "step": 12920 }, { "epoch": 11.618705035971223, "step": 12920, "torque_loss": 0.20111067593097687 }, { "epoch": 11.627697841726619, "grad_norm": 0.24172499775886536, "learning_rate": 8.856989134072676e-05, "loss": 0.0282, "step": 12930 }, { "action_loss": 0.00626682722941041, "epoch": 11.627697841726619, "step": 12930 }, { "epoch": 11.627697841726619, "step": 12930, "torque_loss": 0.1820818930864334 }, { "epoch": 11.636690647482014, "grad_norm": 0.23784632980823517, "learning_rate": 8.85488390902403e-05, "loss": 0.0345, "step": 12940 }, { "action_loss": 0.019077060744166374, "epoch": 11.636690647482014, "step": 12940 }, { "epoch": 11.636690647482014, "step": 12940, "torque_loss": 0.269060343503952 }, { "epoch": 11.64568345323741, "grad_norm": 0.2542358636856079, "learning_rate": 8.852776997718199e-05, "loss": 0.0267, "step": 12950 }, { "action_loss": 0.022175610065460205, "epoch": 11.64568345323741, "step": 12950 }, { "epoch": 11.64568345323741, "step": 12950, "torque_loss": 0.21385924518108368 }, { "epoch": 11.654676258992806, "grad_norm": 0.21773430705070496, "learning_rate": 8.850668401076812e-05, "loss": 0.0274, "step": 12960 }, { "action_loss": 0.003777679754421115, "epoch": 11.654676258992806, "step": 12960 }, { "epoch": 11.654676258992806, "step": 12960, "torque_loss": 0.10807100683450699 }, { "epoch": 11.663669064748202, "grad_norm": 0.17255106568336487, "learning_rate": 8.848558120022246e-05, "loss": 0.024, "step": 12970 }, { "action_loss": 0.005991819780319929, "epoch": 11.663669064748202, "step": 12970 }, { "epoch": 11.663669064748202, "step": 12970, "torque_loss": 0.15884993970394135 }, { "epoch": 11.672661870503598, "grad_norm": 0.20121388137340546, "learning_rate": 8.846446155477603e-05, "loss": 0.033, "step": 12980 }, { "action_loss": 0.004754465073347092, "epoch": 11.672661870503598, "step": 12980 }, { "epoch": 11.672661870503598, "step": 12980, "torque_loss": 0.13508401811122894 }, { "epoch": 11.681654676258994, "grad_norm": 0.2935260534286499, "learning_rate": 8.844332508366735e-05, "loss": 0.0293, "step": 12990 }, { "action_loss": 0.017871858552098274, "epoch": 11.681654676258994, "step": 12990 }, { "epoch": 11.681654676258994, "step": 12990, "torque_loss": 0.2422192096710205 }, { "epoch": 11.690647482014388, "grad_norm": 0.1912057250738144, "learning_rate": 8.84221717961422e-05, "loss": 0.0282, "step": 13000 }, { "action_loss": 0.008534647524356842, "epoch": 11.690647482014388, "step": 13000 }, { "epoch": 11.690647482014388, "step": 13000, "torque_loss": 0.17504847049713135 }, { "epoch": 11.699640287769784, "grad_norm": 0.2205926924943924, "learning_rate": 8.840100170145374e-05, "loss": 0.0298, "step": 13010 }, { "action_loss": 0.0057526156306266785, "epoch": 11.699640287769784, "step": 13010 }, { "epoch": 11.699640287769784, "step": 13010, "torque_loss": 0.16436602175235748 }, { "epoch": 11.70863309352518, "grad_norm": 0.3740343451499939, "learning_rate": 8.837981480886249e-05, "loss": 0.0312, "step": 13020 }, { "action_loss": 0.015074525028467178, "epoch": 11.70863309352518, "step": 13020 }, { "epoch": 11.70863309352518, "step": 13020, "torque_loss": 0.1497437208890915 }, { "epoch": 11.717625899280575, "grad_norm": 0.21704895794391632, "learning_rate": 8.835861112763633e-05, "loss": 0.0261, "step": 13030 }, { "action_loss": 0.00927565898746252, "epoch": 11.717625899280575, "step": 13030 }, { "epoch": 11.717625899280575, "step": 13030, "torque_loss": 0.19547586143016815 }, { "epoch": 11.726618705035971, "grad_norm": 0.2548276484012604, "learning_rate": 8.833739066705044e-05, "loss": 0.0394, "step": 13040 }, { "action_loss": 0.010368031449615955, "epoch": 11.726618705035971, "step": 13040 }, { "epoch": 11.726618705035971, "step": 13040, "torque_loss": 0.1925855129957199 }, { "epoch": 11.735611510791367, "grad_norm": 0.19146443903446198, "learning_rate": 8.831615343638742e-05, "loss": 0.035, "step": 13050 }, { "action_loss": 0.009020619094371796, "epoch": 11.735611510791367, "step": 13050 }, { "epoch": 11.735611510791367, "step": 13050, "torque_loss": 0.2048758864402771 }, { "epoch": 11.744604316546763, "grad_norm": 0.1380242556333542, "learning_rate": 8.829489944493711e-05, "loss": 0.0258, "step": 13060 }, { "action_loss": 0.029064776375889778, "epoch": 11.744604316546763, "step": 13060 }, { "epoch": 11.744604316546763, "step": 13060, "torque_loss": 0.2597084045410156 }, { "epoch": 11.753597122302159, "grad_norm": 0.22611233592033386, "learning_rate": 8.827362870199675e-05, "loss": 0.0326, "step": 13070 }, { "action_loss": 0.00797505397349596, "epoch": 11.753597122302159, "step": 13070 }, { "epoch": 11.753597122302159, "step": 13070, "torque_loss": 0.22591108083724976 }, { "epoch": 11.762589928057555, "grad_norm": 0.14971740543842316, "learning_rate": 8.825234121687089e-05, "loss": 0.0235, "step": 13080 }, { "action_loss": 0.007795120123773813, "epoch": 11.762589928057555, "step": 13080 }, { "epoch": 11.762589928057555, "step": 13080, "torque_loss": 0.1958048790693283 }, { "epoch": 11.77158273381295, "grad_norm": 0.2113044112920761, "learning_rate": 8.823103699887139e-05, "loss": 0.0291, "step": 13090 }, { "action_loss": 0.013447259552776814, "epoch": 11.77158273381295, "step": 13090 }, { "epoch": 11.77158273381295, "step": 13090, "torque_loss": 0.34842994809150696 }, { "epoch": 11.780575539568346, "grad_norm": 0.2796385884284973, "learning_rate": 8.820971605731745e-05, "loss": 0.0302, "step": 13100 }, { "action_loss": 0.010369664989411831, "epoch": 11.780575539568346, "step": 13100 }, { "epoch": 11.780575539568346, "step": 13100, "torque_loss": 0.16782458126544952 }, { "epoch": 11.78956834532374, "grad_norm": 0.3464469313621521, "learning_rate": 8.818837840153556e-05, "loss": 0.0342, "step": 13110 }, { "action_loss": 0.01899600587785244, "epoch": 11.78956834532374, "step": 13110 }, { "epoch": 11.78956834532374, "step": 13110, "torque_loss": 0.28297629952430725 }, { "epoch": 11.798561151079136, "grad_norm": 0.2458200305700302, "learning_rate": 8.816702404085952e-05, "loss": 0.0301, "step": 13120 }, { "action_loss": 0.00908240769058466, "epoch": 11.798561151079136, "step": 13120 }, { "epoch": 11.798561151079136, "step": 13120, "torque_loss": 0.12565983831882477 }, { "epoch": 11.807553956834532, "grad_norm": 0.28967660665512085, "learning_rate": 8.814565298463048e-05, "loss": 0.0368, "step": 13130 }, { "action_loss": 0.005491857882589102, "epoch": 11.807553956834532, "step": 13130 }, { "epoch": 11.807553956834532, "step": 13130, "torque_loss": 0.11507780104875565 }, { "epoch": 11.816546762589928, "grad_norm": 0.20069074630737305, "learning_rate": 8.812426524219688e-05, "loss": 0.0347, "step": 13140 }, { "action_loss": 0.015966398641467094, "epoch": 11.816546762589928, "step": 13140 }, { "epoch": 11.816546762589928, "step": 13140, "torque_loss": 0.3160371780395508 }, { "epoch": 11.825539568345324, "grad_norm": 0.22031104564666748, "learning_rate": 8.81028608229144e-05, "loss": 0.0367, "step": 13150 }, { "action_loss": 0.013406774960458279, "epoch": 11.825539568345324, "step": 13150 }, { "epoch": 11.825539568345324, "step": 13150, "torque_loss": 0.19510968029499054 }, { "epoch": 11.83453237410072, "grad_norm": 0.218713641166687, "learning_rate": 8.808143973614611e-05, "loss": 0.0361, "step": 13160 }, { "action_loss": 0.006318250205367804, "epoch": 11.83453237410072, "step": 13160 }, { "epoch": 11.83453237410072, "step": 13160, "torque_loss": 0.15496516227722168 }, { "epoch": 11.843525179856115, "grad_norm": 0.15494956076145172, "learning_rate": 8.806000199126228e-05, "loss": 0.0294, "step": 13170 }, { "action_loss": 0.007812475319951773, "epoch": 11.843525179856115, "step": 13170 }, { "epoch": 11.843525179856115, "step": 13170, "torque_loss": 0.15702210366725922 }, { "epoch": 11.852517985611511, "grad_norm": 0.19239163398742676, "learning_rate": 8.803854759764052e-05, "loss": 0.0294, "step": 13180 }, { "action_loss": 0.004298389423638582, "epoch": 11.852517985611511, "step": 13180 }, { "epoch": 11.852517985611511, "step": 13180, "torque_loss": 0.14222092926502228 }, { "epoch": 11.861510791366907, "grad_norm": 0.22586558759212494, "learning_rate": 8.801707656466572e-05, "loss": 0.0264, "step": 13190 }, { "action_loss": 0.007657547947019339, "epoch": 11.861510791366907, "step": 13190 }, { "epoch": 11.861510791366907, "step": 13190, "torque_loss": 0.19022731482982635 }, { "epoch": 11.870503597122303, "grad_norm": 0.2778726816177368, "learning_rate": 8.799558890173003e-05, "loss": 0.0296, "step": 13200 }, { "action_loss": 0.009671909734606743, "epoch": 11.870503597122303, "step": 13200 }, { "epoch": 11.870503597122303, "step": 13200, "torque_loss": 0.17718635499477386 }, { "epoch": 11.879496402877697, "grad_norm": 0.14290344715118408, "learning_rate": 8.79740846182329e-05, "loss": 0.028, "step": 13210 }, { "action_loss": 0.012000802904367447, "epoch": 11.879496402877697, "step": 13210 }, { "epoch": 11.879496402877697, "step": 13210, "torque_loss": 0.2561923563480377 }, { "epoch": 11.888489208633093, "grad_norm": 0.14727021753787994, "learning_rate": 8.7952563723581e-05, "loss": 0.0337, "step": 13220 }, { "action_loss": 0.020062079653143883, "epoch": 11.888489208633093, "step": 13220 }, { "epoch": 11.888489208633093, "step": 13220, "torque_loss": 0.18929481506347656 }, { "epoch": 11.897482014388489, "grad_norm": 0.19571225345134735, "learning_rate": 8.793102622718834e-05, "loss": 0.0289, "step": 13230 }, { "action_loss": 0.007154311519116163, "epoch": 11.897482014388489, "step": 13230 }, { "epoch": 11.897482014388489, "step": 13230, "torque_loss": 0.1355283558368683 }, { "epoch": 11.906474820143885, "grad_norm": 0.20224164426326752, "learning_rate": 8.790947213847613e-05, "loss": 0.0269, "step": 13240 }, { "action_loss": 0.01838960312306881, "epoch": 11.906474820143885, "step": 13240 }, { "epoch": 11.906474820143885, "step": 13240, "torque_loss": 0.19085264205932617 }, { "epoch": 11.91546762589928, "grad_norm": 0.20597146451473236, "learning_rate": 8.788790146687286e-05, "loss": 0.0342, "step": 13250 }, { "action_loss": 0.0136630954220891, "epoch": 11.91546762589928, "step": 13250 }, { "epoch": 11.91546762589928, "step": 13250, "torque_loss": 0.25419890880584717 }, { "epoch": 11.924460431654676, "grad_norm": 0.21837638318538666, "learning_rate": 8.786631422181429e-05, "loss": 0.0305, "step": 13260 }, { "action_loss": 0.006800854112952948, "epoch": 11.924460431654676, "step": 13260 }, { "epoch": 11.924460431654676, "step": 13260, "torque_loss": 0.2194669246673584 }, { "epoch": 11.933453237410072, "grad_norm": 0.17411084473133087, "learning_rate": 8.78447104127434e-05, "loss": 0.0327, "step": 13270 }, { "action_loss": 0.010021612048149109, "epoch": 11.933453237410072, "step": 13270 }, { "epoch": 11.933453237410072, "step": 13270, "torque_loss": 0.15113182365894318 }, { "epoch": 11.942446043165468, "grad_norm": 0.19853894412517548, "learning_rate": 8.782309004911042e-05, "loss": 0.0307, "step": 13280 }, { "action_loss": 0.006821583956480026, "epoch": 11.942446043165468, "step": 13280 }, { "epoch": 11.942446043165468, "step": 13280, "torque_loss": 0.1586562544107437 }, { "epoch": 11.951438848920864, "grad_norm": 0.17550790309906006, "learning_rate": 8.780145314037286e-05, "loss": 0.0238, "step": 13290 }, { "action_loss": 0.020739855244755745, "epoch": 11.951438848920864, "step": 13290 }, { "epoch": 11.951438848920864, "step": 13290, "torque_loss": 0.2443382740020752 }, { "epoch": 11.96043165467626, "grad_norm": 0.187747061252594, "learning_rate": 8.777979969599542e-05, "loss": 0.0299, "step": 13300 }, { "action_loss": 0.003966301213949919, "epoch": 11.96043165467626, "step": 13300 }, { "epoch": 11.96043165467626, "step": 13300, "torque_loss": 0.09474916011095047 }, { "epoch": 11.969424460431654, "grad_norm": 0.19022296369075775, "learning_rate": 8.775812972545006e-05, "loss": 0.0296, "step": 13310 }, { "action_loss": 0.00971760880202055, "epoch": 11.969424460431654, "step": 13310 }, { "epoch": 11.969424460431654, "step": 13310, "torque_loss": 0.19722360372543335 }, { "epoch": 11.97841726618705, "grad_norm": 0.15819807350635529, "learning_rate": 8.773644323821596e-05, "loss": 0.03, "step": 13320 }, { "action_loss": 0.004943752195686102, "epoch": 11.97841726618705, "step": 13320 }, { "epoch": 11.97841726618705, "step": 13320, "torque_loss": 0.2347731590270996 }, { "epoch": 11.987410071942445, "grad_norm": 0.23026011884212494, "learning_rate": 8.771474024377953e-05, "loss": 0.0285, "step": 13330 }, { "action_loss": 0.005482815206050873, "epoch": 11.987410071942445, "step": 13330 }, { "epoch": 11.987410071942445, "step": 13330, "torque_loss": 0.2239314317703247 }, { "epoch": 11.996402877697841, "grad_norm": 0.2017931044101715, "learning_rate": 8.769302075163438e-05, "loss": 0.0275, "step": 13340 }, { "action_loss": 0.0031294787768274546, "epoch": 11.996402877697841, "step": 13340 }, { "epoch": 11.996402877697841, "step": 13340, "torque_loss": 0.12656830251216888 }, { "epoch": 12.005395683453237, "grad_norm": 0.16484378278255463, "learning_rate": 8.767128477128137e-05, "loss": 0.0241, "step": 13350 }, { "action_loss": 0.014991614036262035, "epoch": 12.005395683453237, "step": 13350 }, { "epoch": 12.005395683453237, "step": 13350, "torque_loss": 0.2388380765914917 }, { "epoch": 12.014388489208633, "grad_norm": 0.15282633900642395, "learning_rate": 8.764953231222854e-05, "loss": 0.0299, "step": 13360 }, { "action_loss": 0.005582134705036879, "epoch": 12.014388489208633, "step": 13360 }, { "epoch": 12.014388489208633, "step": 13360, "torque_loss": 0.13705702126026154 }, { "epoch": 12.023381294964029, "grad_norm": 0.1365174949169159, "learning_rate": 8.762776338399119e-05, "loss": 0.0253, "step": 13370 }, { "action_loss": 0.012231819331645966, "epoch": 12.023381294964029, "step": 13370 }, { "epoch": 12.023381294964029, "step": 13370, "torque_loss": 0.24758408963680267 }, { "epoch": 12.032374100719425, "grad_norm": 0.1794433891773224, "learning_rate": 8.760597799609176e-05, "loss": 0.0325, "step": 13380 }, { "action_loss": 0.008624141104519367, "epoch": 12.032374100719425, "step": 13380 }, { "epoch": 12.032374100719425, "step": 13380, "torque_loss": 0.208155557513237 }, { "epoch": 12.04136690647482, "grad_norm": 0.1969967931509018, "learning_rate": 8.758417615805992e-05, "loss": 0.0286, "step": 13390 }, { "action_loss": 0.015437210910022259, "epoch": 12.04136690647482, "step": 13390 }, { "epoch": 12.04136690647482, "step": 13390, "torque_loss": 0.21850208938121796 }, { "epoch": 12.050359712230216, "grad_norm": 0.2400890290737152, "learning_rate": 8.756235787943254e-05, "loss": 0.0306, "step": 13400 }, { "action_loss": 0.008588382042944431, "epoch": 12.050359712230216, "step": 13400 }, { "epoch": 12.050359712230216, "step": 13400, "torque_loss": 0.1782732456922531 }, { "epoch": 12.059352517985612, "grad_norm": 0.21354088187217712, "learning_rate": 8.754052316975367e-05, "loss": 0.0274, "step": 13410 }, { "action_loss": 0.005849506240338087, "epoch": 12.059352517985612, "step": 13410 }, { "epoch": 12.059352517985612, "step": 13410, "torque_loss": 0.16433827579021454 }, { "epoch": 12.068345323741006, "grad_norm": 0.21063202619552612, "learning_rate": 8.751867203857455e-05, "loss": 0.031, "step": 13420 }, { "action_loss": 0.007084265351295471, "epoch": 12.068345323741006, "step": 13420 }, { "epoch": 12.068345323741006, "step": 13420, "torque_loss": 0.1567583829164505 }, { "epoch": 12.077338129496402, "grad_norm": 0.21516399085521698, "learning_rate": 8.749680449545363e-05, "loss": 0.0278, "step": 13430 }, { "action_loss": 0.006202444434165955, "epoch": 12.077338129496402, "step": 13430 }, { "epoch": 12.077338129496402, "step": 13430, "torque_loss": 0.11177494376897812 }, { "epoch": 12.086330935251798, "grad_norm": 0.26193246245384216, "learning_rate": 8.747492054995649e-05, "loss": 0.0293, "step": 13440 }, { "action_loss": 0.020284678786993027, "epoch": 12.086330935251798, "step": 13440 }, { "epoch": 12.086330935251798, "step": 13440, "torque_loss": 0.21813379228115082 }, { "epoch": 12.095323741007194, "grad_norm": 0.18548575043678284, "learning_rate": 8.745302021165595e-05, "loss": 0.027, "step": 13450 }, { "action_loss": 0.005594077054411173, "epoch": 12.095323741007194, "step": 13450 }, { "epoch": 12.095323741007194, "step": 13450, "torque_loss": 0.1924503594636917 }, { "epoch": 12.10431654676259, "grad_norm": 0.25332528352737427, "learning_rate": 8.743110349013192e-05, "loss": 0.0334, "step": 13460 }, { "action_loss": 0.014192253351211548, "epoch": 12.10431654676259, "step": 13460 }, { "epoch": 12.10431654676259, "step": 13460, "torque_loss": 0.24333934485912323 }, { "epoch": 12.113309352517986, "grad_norm": 0.25553032755851746, "learning_rate": 8.740917039497153e-05, "loss": 0.0367, "step": 13470 }, { "action_loss": 0.008681324310600758, "epoch": 12.113309352517986, "step": 13470 }, { "epoch": 12.113309352517986, "step": 13470, "torque_loss": 0.13937032222747803 }, { "epoch": 12.122302158273381, "grad_norm": 0.19732677936553955, "learning_rate": 8.738722093576906e-05, "loss": 0.0275, "step": 13480 }, { "action_loss": 0.00623747706413269, "epoch": 12.122302158273381, "step": 13480 }, { "epoch": 12.122302158273381, "step": 13480, "torque_loss": 0.1394103765487671 }, { "epoch": 12.131294964028777, "grad_norm": 0.12501166760921478, "learning_rate": 8.736525512212597e-05, "loss": 0.0263, "step": 13490 }, { "action_loss": 0.003224798245355487, "epoch": 12.131294964028777, "step": 13490 }, { "epoch": 12.131294964028777, "step": 13490, "torque_loss": 0.20911644399166107 }, { "epoch": 12.140287769784173, "grad_norm": 0.15259595215320587, "learning_rate": 8.734327296365084e-05, "loss": 0.0258, "step": 13500 }, { "action_loss": 0.0035129731986671686, "epoch": 12.140287769784173, "step": 13500 }, { "epoch": 12.140287769784173, "step": 13500, "torque_loss": 0.13137580454349518 }, { "epoch": 12.149280575539569, "grad_norm": 0.12937882542610168, "learning_rate": 8.732127446995939e-05, "loss": 0.0237, "step": 13510 }, { "action_loss": 0.00416044844314456, "epoch": 12.149280575539569, "step": 13510 }, { "epoch": 12.149280575539569, "step": 13510, "torque_loss": 0.13530878722667694 }, { "epoch": 12.158273381294965, "grad_norm": 0.16132386028766632, "learning_rate": 8.729925965067454e-05, "loss": 0.0281, "step": 13520 }, { "action_loss": 0.00907626748085022, "epoch": 12.158273381294965, "step": 13520 }, { "epoch": 12.158273381294965, "step": 13520, "torque_loss": 0.13862015306949615 }, { "epoch": 12.167266187050359, "grad_norm": 0.1387118101119995, "learning_rate": 8.72772285154263e-05, "loss": 0.0226, "step": 13530 }, { "action_loss": 0.011052281595766544, "epoch": 12.167266187050359, "step": 13530 }, { "epoch": 12.167266187050359, "step": 13530, "torque_loss": 0.1944781392812729 }, { "epoch": 12.176258992805755, "grad_norm": 0.16718530654907227, "learning_rate": 8.725518107385187e-05, "loss": 0.0264, "step": 13540 }, { "action_loss": 0.00865952204912901, "epoch": 12.176258992805755, "step": 13540 }, { "epoch": 12.176258992805755, "step": 13540, "torque_loss": 0.1949598789215088 }, { "epoch": 12.18525179856115, "grad_norm": 0.17221274971961975, "learning_rate": 8.72331173355955e-05, "loss": 0.0255, "step": 13550 }, { "action_loss": 0.004157043993473053, "epoch": 12.18525179856115, "step": 13550 }, { "epoch": 12.18525179856115, "step": 13550, "torque_loss": 0.1445056051015854 }, { "epoch": 12.194244604316546, "grad_norm": 0.15135227143764496, "learning_rate": 8.721103731030867e-05, "loss": 0.0305, "step": 13560 }, { "action_loss": 0.006421221420168877, "epoch": 12.194244604316546, "step": 13560 }, { "epoch": 12.194244604316546, "step": 13560, "torque_loss": 0.13704736530780792 }, { "epoch": 12.203237410071942, "grad_norm": 0.1336110383272171, "learning_rate": 8.718894100764989e-05, "loss": 0.03, "step": 13570 }, { "action_loss": 0.005444832146167755, "epoch": 12.203237410071942, "step": 13570 }, { "epoch": 12.203237410071942, "step": 13570, "torque_loss": 0.12585489451885223 }, { "epoch": 12.212230215827338, "grad_norm": 0.19279047846794128, "learning_rate": 8.716682843728485e-05, "loss": 0.0268, "step": 13580 }, { "action_loss": 0.005671229213476181, "epoch": 12.212230215827338, "step": 13580 }, { "epoch": 12.212230215827338, "step": 13580, "torque_loss": 0.21320730447769165 }, { "epoch": 12.221223021582734, "grad_norm": 0.2297142744064331, "learning_rate": 8.714469960888634e-05, "loss": 0.0336, "step": 13590 }, { "action_loss": 0.011373278684914112, "epoch": 12.221223021582734, "step": 13590 }, { "epoch": 12.221223021582734, "step": 13590, "torque_loss": 0.1786758154630661 }, { "epoch": 12.23021582733813, "grad_norm": 0.17900219559669495, "learning_rate": 8.712255453213427e-05, "loss": 0.0366, "step": 13600 }, { "action_loss": 0.004334051162004471, "epoch": 12.23021582733813, "step": 13600 }, { "epoch": 12.23021582733813, "step": 13600, "torque_loss": 0.14664481580257416 }, { "epoch": 12.239208633093526, "grad_norm": 0.21416576206684113, "learning_rate": 8.710039321671563e-05, "loss": 0.0279, "step": 13610 }, { "action_loss": 0.011809024959802628, "epoch": 12.239208633093526, "step": 13610 }, { "epoch": 12.239208633093526, "step": 13610, "torque_loss": 0.17496351897716522 }, { "epoch": 12.248201438848922, "grad_norm": 0.20426622033119202, "learning_rate": 8.707821567232456e-05, "loss": 0.0252, "step": 13620 }, { "action_loss": 0.019772714003920555, "epoch": 12.248201438848922, "step": 13620 }, { "epoch": 12.248201438848922, "step": 13620, "torque_loss": 0.1827607899904251 }, { "epoch": 12.257194244604317, "grad_norm": 0.20221176743507385, "learning_rate": 8.705602190866225e-05, "loss": 0.0262, "step": 13630 }, { "action_loss": 0.0059941355139017105, "epoch": 12.257194244604317, "step": 13630 }, { "epoch": 12.257194244604317, "step": 13630, "torque_loss": 0.17777514457702637 }, { "epoch": 12.266187050359711, "grad_norm": 0.17171527445316315, "learning_rate": 8.703381193543701e-05, "loss": 0.0246, "step": 13640 }, { "action_loss": 0.006695365533232689, "epoch": 12.266187050359711, "step": 13640 }, { "epoch": 12.266187050359711, "step": 13640, "torque_loss": 0.17444372177124023 }, { "epoch": 12.275179856115107, "grad_norm": 0.1129479631781578, "learning_rate": 8.701158576236423e-05, "loss": 0.024, "step": 13650 }, { "action_loss": 0.005662817042320967, "epoch": 12.275179856115107, "step": 13650 }, { "epoch": 12.275179856115107, "step": 13650, "torque_loss": 0.16009199619293213 }, { "epoch": 12.284172661870503, "grad_norm": 0.2028488665819168, "learning_rate": 8.69893433991664e-05, "loss": 0.028, "step": 13660 }, { "action_loss": 0.008642747066915035, "epoch": 12.284172661870503, "step": 13660 }, { "epoch": 12.284172661870503, "step": 13660, "torque_loss": 0.1456352323293686 }, { "epoch": 12.293165467625899, "grad_norm": 0.177302286028862, "learning_rate": 8.69670848555731e-05, "loss": 0.0227, "step": 13670 }, { "action_loss": 0.002919982885941863, "epoch": 12.293165467625899, "step": 13670 }, { "epoch": 12.293165467625899, "step": 13670, "torque_loss": 0.12210426479578018 }, { "epoch": 12.302158273381295, "grad_norm": 0.1609041541814804, "learning_rate": 8.694481014132096e-05, "loss": 0.0291, "step": 13680 }, { "action_loss": 0.005989426746964455, "epoch": 12.302158273381295, "step": 13680 }, { "epoch": 12.302158273381295, "step": 13680, "torque_loss": 0.20446713268756866 }, { "epoch": 12.31115107913669, "grad_norm": 0.18425804376602173, "learning_rate": 8.69225192661537e-05, "loss": 0.0256, "step": 13690 }, { "action_loss": 0.016211483627557755, "epoch": 12.31115107913669, "step": 13690 }, { "epoch": 12.31115107913669, "step": 13690, "torque_loss": 0.18407924473285675 }, { "epoch": 12.320143884892087, "grad_norm": 0.20495326817035675, "learning_rate": 8.690021223982208e-05, "loss": 0.0258, "step": 13700 }, { "action_loss": 0.008879588916897774, "epoch": 12.320143884892087, "step": 13700 }, { "epoch": 12.320143884892087, "step": 13700, "torque_loss": 0.2368558794260025 }, { "epoch": 12.329136690647482, "grad_norm": 0.2585136890411377, "learning_rate": 8.687788907208398e-05, "loss": 0.0315, "step": 13710 }, { "action_loss": 0.006447577383369207, "epoch": 12.329136690647482, "step": 13710 }, { "epoch": 12.329136690647482, "step": 13710, "torque_loss": 0.1522107869386673 }, { "epoch": 12.338129496402878, "grad_norm": 0.2420971691608429, "learning_rate": 8.685554977270431e-05, "loss": 0.0348, "step": 13720 }, { "action_loss": 0.011868362314999104, "epoch": 12.338129496402878, "step": 13720 }, { "epoch": 12.338129496402878, "step": 13720, "torque_loss": 0.1899990439414978 }, { "epoch": 12.347122302158274, "grad_norm": 0.3158075213432312, "learning_rate": 8.683319435145503e-05, "loss": 0.0312, "step": 13730 }, { "action_loss": 0.009543884545564651, "epoch": 12.347122302158274, "step": 13730 }, { "epoch": 12.347122302158274, "step": 13730, "torque_loss": 0.23458606004714966 }, { "epoch": 12.35611510791367, "grad_norm": 0.21426863968372345, "learning_rate": 8.681082281811517e-05, "loss": 0.0347, "step": 13740 }, { "action_loss": 0.004482394549995661, "epoch": 12.35611510791367, "step": 13740 }, { "epoch": 12.35611510791367, "step": 13740, "torque_loss": 0.11032015830278397 }, { "epoch": 12.365107913669064, "grad_norm": 0.22547325491905212, "learning_rate": 8.67884351824708e-05, "loss": 0.0306, "step": 13750 }, { "action_loss": 0.03389107808470726, "epoch": 12.365107913669064, "step": 13750 }, { "epoch": 12.365107913669064, "step": 13750, "torque_loss": 0.29089948534965515 }, { "epoch": 12.37410071942446, "grad_norm": 0.17954598367214203, "learning_rate": 8.676603145431501e-05, "loss": 0.0353, "step": 13760 }, { "action_loss": 0.017749479040503502, "epoch": 12.37410071942446, "step": 13760 }, { "epoch": 12.37410071942446, "step": 13760, "torque_loss": 0.2003048211336136 }, { "epoch": 12.383093525179856, "grad_norm": 0.1984575092792511, "learning_rate": 8.674361164344799e-05, "loss": 0.0317, "step": 13770 }, { "action_loss": 0.01362281572073698, "epoch": 12.383093525179856, "step": 13770 }, { "epoch": 12.383093525179856, "step": 13770, "torque_loss": 0.2103000432252884 }, { "epoch": 12.392086330935252, "grad_norm": 0.18536479771137238, "learning_rate": 8.672117575967688e-05, "loss": 0.0305, "step": 13780 }, { "action_loss": 0.007576656993478537, "epoch": 12.392086330935252, "step": 13780 }, { "epoch": 12.392086330935252, "step": 13780, "torque_loss": 0.21629302203655243 }, { "epoch": 12.401079136690647, "grad_norm": 0.17437700927257538, "learning_rate": 8.669872381281595e-05, "loss": 0.0297, "step": 13790 }, { "action_loss": 0.009451230056583881, "epoch": 12.401079136690647, "step": 13790 }, { "epoch": 12.401079136690647, "step": 13790, "torque_loss": 0.17411558330059052 }, { "epoch": 12.410071942446043, "grad_norm": 0.2558109164237976, "learning_rate": 8.667625581268639e-05, "loss": 0.0279, "step": 13800 }, { "action_loss": 0.014797155745327473, "epoch": 12.410071942446043, "step": 13800 }, { "epoch": 12.410071942446043, "step": 13800, "torque_loss": 0.20863260328769684 }, { "epoch": 12.41906474820144, "grad_norm": 0.25887173414230347, "learning_rate": 8.665377176911651e-05, "loss": 0.0351, "step": 13810 }, { "action_loss": 0.002787972567602992, "epoch": 12.41906474820144, "step": 13810 }, { "epoch": 12.41906474820144, "step": 13810, "torque_loss": 0.1083017885684967 }, { "epoch": 12.428057553956835, "grad_norm": 0.1929418295621872, "learning_rate": 8.663127169194159e-05, "loss": 0.025, "step": 13820 }, { "action_loss": 0.010388625785708427, "epoch": 12.428057553956835, "step": 13820 }, { "epoch": 12.428057553956835, "step": 13820, "torque_loss": 0.18371157348155975 }, { "epoch": 12.43705035971223, "grad_norm": 0.3769432604312897, "learning_rate": 8.660875559100389e-05, "loss": 0.0358, "step": 13830 }, { "action_loss": 0.00624109199270606, "epoch": 12.43705035971223, "step": 13830 }, { "epoch": 12.43705035971223, "step": 13830, "torque_loss": 0.12006125599145889 }, { "epoch": 12.446043165467627, "grad_norm": 0.2790631055831909, "learning_rate": 8.658622347615274e-05, "loss": 0.0321, "step": 13840 }, { "action_loss": 0.0030693253502249718, "epoch": 12.446043165467627, "step": 13840 }, { "epoch": 12.446043165467627, "step": 13840, "torque_loss": 0.10295349359512329 }, { "epoch": 12.45503597122302, "grad_norm": 0.21114811301231384, "learning_rate": 8.656367535724448e-05, "loss": 0.0301, "step": 13850 }, { "action_loss": 0.0069730584509670734, "epoch": 12.45503597122302, "step": 13850 }, { "epoch": 12.45503597122302, "step": 13850, "torque_loss": 0.19449037313461304 }, { "epoch": 12.464028776978417, "grad_norm": 0.16922609508037567, "learning_rate": 8.65411112441424e-05, "loss": 0.0249, "step": 13860 }, { "action_loss": 0.009300419129431248, "epoch": 12.464028776978417, "step": 13860 }, { "epoch": 12.464028776978417, "step": 13860, "torque_loss": 0.19749347865581512 }, { "epoch": 12.473021582733812, "grad_norm": 0.2075595259666443, "learning_rate": 8.651853114671679e-05, "loss": 0.0267, "step": 13870 }, { "action_loss": 0.009459600783884525, "epoch": 12.473021582733812, "step": 13870 }, { "epoch": 12.473021582733812, "step": 13870, "torque_loss": 0.15329697728157043 }, { "epoch": 12.482014388489208, "grad_norm": 0.20303109288215637, "learning_rate": 8.649593507484499e-05, "loss": 0.0261, "step": 13880 }, { "action_loss": 0.03017239458858967, "epoch": 12.482014388489208, "step": 13880 }, { "epoch": 12.482014388489208, "step": 13880, "torque_loss": 0.22510559856891632 }, { "epoch": 12.491007194244604, "grad_norm": 0.293861985206604, "learning_rate": 8.647332303841126e-05, "loss": 0.0324, "step": 13890 }, { "action_loss": 0.008938915096223354, "epoch": 12.491007194244604, "step": 13890 }, { "epoch": 12.491007194244604, "step": 13890, "torque_loss": 0.17778806388378143 }, { "epoch": 12.5, "grad_norm": 0.31910720467567444, "learning_rate": 8.645069504730689e-05, "loss": 0.0339, "step": 13900 }, { "action_loss": 0.007730050478130579, "epoch": 12.5, "step": 13900 }, { "epoch": 12.5, "step": 13900, "torque_loss": 0.22709965705871582 }, { "epoch": 12.508992805755396, "grad_norm": 0.31126874685287476, "learning_rate": 8.64280511114301e-05, "loss": 0.0318, "step": 13910 }, { "action_loss": 0.0074292621575295925, "epoch": 12.508992805755396, "step": 13910 }, { "epoch": 12.508992805755396, "step": 13910, "torque_loss": 0.14535504579544067 }, { "epoch": 12.517985611510792, "grad_norm": 0.3557453453540802, "learning_rate": 8.640539124068617e-05, "loss": 0.0317, "step": 13920 }, { "action_loss": 0.006586879026144743, "epoch": 12.517985611510792, "step": 13920 }, { "epoch": 12.517985611510792, "step": 13920, "torque_loss": 0.14251063764095306 }, { "epoch": 12.526978417266188, "grad_norm": 0.18607614934444427, "learning_rate": 8.638271544498727e-05, "loss": 0.0282, "step": 13930 }, { "action_loss": 0.004078706260770559, "epoch": 12.526978417266188, "step": 13930 }, { "epoch": 12.526978417266188, "step": 13930, "torque_loss": 0.14117054641246796 }, { "epoch": 12.535971223021583, "grad_norm": 0.26420363783836365, "learning_rate": 8.636002373425257e-05, "loss": 0.036, "step": 13940 }, { "action_loss": 0.021406589075922966, "epoch": 12.535971223021583, "step": 13940 }, { "epoch": 12.535971223021583, "step": 13940, "torque_loss": 0.19551688432693481 }, { "epoch": 12.54496402877698, "grad_norm": 0.23907963931560516, "learning_rate": 8.633731611840817e-05, "loss": 0.0293, "step": 13950 }, { "action_loss": 0.007059094030410051, "epoch": 12.54496402877698, "step": 13950 }, { "epoch": 12.54496402877698, "step": 13950, "torque_loss": 0.17851687967777252 }, { "epoch": 12.553956834532373, "grad_norm": 0.2058265060186386, "learning_rate": 8.631459260738717e-05, "loss": 0.0298, "step": 13960 }, { "action_loss": 0.013033946044743061, "epoch": 12.553956834532373, "step": 13960 }, { "epoch": 12.553956834532373, "step": 13960, "torque_loss": 0.18960046768188477 }, { "epoch": 12.56294964028777, "grad_norm": 0.1577250212430954, "learning_rate": 8.62918532111296e-05, "loss": 0.0305, "step": 13970 }, { "action_loss": 0.008079376071691513, "epoch": 12.56294964028777, "step": 13970 }, { "epoch": 12.56294964028777, "step": 13970, "torque_loss": 0.22447027266025543 }, { "epoch": 12.571942446043165, "grad_norm": 0.25627270340919495, "learning_rate": 8.626909793958248e-05, "loss": 0.0322, "step": 13980 }, { "action_loss": 0.005876054521650076, "epoch": 12.571942446043165, "step": 13980 }, { "epoch": 12.571942446043165, "step": 13980, "torque_loss": 0.23657183349132538 }, { "epoch": 12.58093525179856, "grad_norm": 0.1503450572490692, "learning_rate": 8.624632680269969e-05, "loss": 0.0244, "step": 13990 }, { "action_loss": 0.024331025779247284, "epoch": 12.58093525179856, "step": 13990 }, { "epoch": 12.58093525179856, "step": 13990, "torque_loss": 0.25837817788124084 }, { "epoch": 12.589928057553957, "grad_norm": 0.20421604812145233, "learning_rate": 8.622353981044212e-05, "loss": 0.0327, "step": 14000 }, { "action_loss": 0.01031468529254198, "epoch": 12.589928057553957, "step": 14000 }, { "epoch": 12.589928057553957, "step": 14000, "torque_loss": 0.20022381842136383 }, { "epoch": 12.598920863309353, "grad_norm": 0.17306531965732574, "learning_rate": 8.620073697277757e-05, "loss": 0.0294, "step": 14010 }, { "action_loss": 0.015550724230706692, "epoch": 12.598920863309353, "step": 14010 }, { "epoch": 12.598920863309353, "step": 14010, "torque_loss": 0.22854821383953094 }, { "epoch": 12.607913669064748, "grad_norm": 0.2369096875190735, "learning_rate": 8.617791829968079e-05, "loss": 0.0316, "step": 14020 }, { "action_loss": 0.007500786334276199, "epoch": 12.607913669064748, "step": 14020 }, { "epoch": 12.607913669064748, "step": 14020, "torque_loss": 0.14340105652809143 }, { "epoch": 12.616906474820144, "grad_norm": 0.15291820466518402, "learning_rate": 8.615508380113344e-05, "loss": 0.0272, "step": 14030 }, { "action_loss": 0.005747800227254629, "epoch": 12.616906474820144, "step": 14030 }, { "epoch": 12.616906474820144, "step": 14030, "torque_loss": 0.17200352251529694 }, { "epoch": 12.62589928057554, "grad_norm": 0.13673777878284454, "learning_rate": 8.613223348712408e-05, "loss": 0.0251, "step": 14040 }, { "action_loss": 0.008844892494380474, "epoch": 12.62589928057554, "step": 14040 }, { "epoch": 12.62589928057554, "step": 14040, "torque_loss": 0.13180887699127197 }, { "epoch": 12.634892086330936, "grad_norm": 0.19318139553070068, "learning_rate": 8.610936736764824e-05, "loss": 0.0293, "step": 14050 }, { "action_loss": 0.011744599789381027, "epoch": 12.634892086330936, "step": 14050 }, { "epoch": 12.634892086330936, "step": 14050, "torque_loss": 0.19017760455608368 }, { "epoch": 12.64388489208633, "grad_norm": 0.29291272163391113, "learning_rate": 8.608648545270833e-05, "loss": 0.0334, "step": 14060 }, { "action_loss": 0.007258343044668436, "epoch": 12.64388489208633, "step": 14060 }, { "epoch": 12.64388489208633, "step": 14060, "torque_loss": 0.19921047985553741 }, { "epoch": 12.652877697841726, "grad_norm": 0.16635005176067352, "learning_rate": 8.606358775231366e-05, "loss": 0.0269, "step": 14070 }, { "action_loss": 0.008396687917411327, "epoch": 12.652877697841726, "step": 14070 }, { "epoch": 12.652877697841726, "step": 14070, "torque_loss": 0.14025096595287323 }, { "epoch": 12.661870503597122, "grad_norm": 0.18981537222862244, "learning_rate": 8.60406742764805e-05, "loss": 0.0373, "step": 14080 }, { "action_loss": 0.004061806481331587, "epoch": 12.661870503597122, "step": 14080 }, { "epoch": 12.661870503597122, "step": 14080, "torque_loss": 0.14928661286830902 }, { "epoch": 12.670863309352518, "grad_norm": 0.1671130210161209, "learning_rate": 8.601774503523195e-05, "loss": 0.0321, "step": 14090 }, { "action_loss": 0.05101076140999794, "epoch": 12.670863309352518, "step": 14090 }, { "epoch": 12.670863309352518, "step": 14090, "torque_loss": 0.265736848115921 }, { "epoch": 12.679856115107913, "grad_norm": 0.29594436287879944, "learning_rate": 8.599480003859805e-05, "loss": 0.0376, "step": 14100 }, { "action_loss": 0.008639012463390827, "epoch": 12.679856115107913, "step": 14100 }, { "epoch": 12.679856115107913, "step": 14100, "torque_loss": 0.15968520939350128 }, { "epoch": 12.68884892086331, "grad_norm": 0.19418510794639587, "learning_rate": 8.597183929661573e-05, "loss": 0.0286, "step": 14110 }, { "action_loss": 0.004509430844336748, "epoch": 12.68884892086331, "step": 14110 }, { "epoch": 12.68884892086331, "step": 14110, "torque_loss": 0.19787995517253876 }, { "epoch": 12.697841726618705, "grad_norm": 0.18790556490421295, "learning_rate": 8.594886281932879e-05, "loss": 0.0281, "step": 14120 }, { "action_loss": 0.019316477701067924, "epoch": 12.697841726618705, "step": 14120 }, { "epoch": 12.697841726618705, "step": 14120, "torque_loss": 0.2895883023738861 }, { "epoch": 12.706834532374101, "grad_norm": 0.24451220035552979, "learning_rate": 8.59258706167879e-05, "loss": 0.0352, "step": 14130 }, { "action_loss": 0.014689509756863117, "epoch": 12.706834532374101, "step": 14130 }, { "epoch": 12.706834532374101, "step": 14130, "torque_loss": 0.20996792614459991 }, { "epoch": 12.715827338129497, "grad_norm": 0.21057312190532684, "learning_rate": 8.590286269905068e-05, "loss": 0.0266, "step": 14140 }, { "action_loss": 0.007599760312587023, "epoch": 12.715827338129497, "step": 14140 }, { "epoch": 12.715827338129497, "step": 14140, "torque_loss": 0.20964264869689941 }, { "epoch": 12.724820143884893, "grad_norm": 0.12359791249036789, "learning_rate": 8.587983907618154e-05, "loss": 0.0258, "step": 14150 }, { "action_loss": 0.011631835252046585, "epoch": 12.724820143884893, "step": 14150 }, { "epoch": 12.724820143884893, "step": 14150, "torque_loss": 0.16758382320404053 }, { "epoch": 12.733812949640289, "grad_norm": 0.12577933073043823, "learning_rate": 8.585679975825178e-05, "loss": 0.0278, "step": 14160 }, { "action_loss": 0.005627541337162256, "epoch": 12.733812949640289, "step": 14160 }, { "epoch": 12.733812949640289, "step": 14160, "torque_loss": 0.12847115099430084 }, { "epoch": 12.742805755395683, "grad_norm": 0.1574677675962448, "learning_rate": 8.583374475533962e-05, "loss": 0.0305, "step": 14170 }, { "action_loss": 0.015127304010093212, "epoch": 12.742805755395683, "step": 14170 }, { "epoch": 12.742805755395683, "step": 14170, "torque_loss": 0.22763170301914215 }, { "epoch": 12.751798561151078, "grad_norm": 0.15239888429641724, "learning_rate": 8.581067407753009e-05, "loss": 0.0258, "step": 14180 }, { "action_loss": 0.01812293566763401, "epoch": 12.751798561151078, "step": 14180 }, { "epoch": 12.751798561151078, "step": 14180, "torque_loss": 0.2816791832447052 }, { "epoch": 12.760791366906474, "grad_norm": 0.1701103299856186, "learning_rate": 8.578758773491507e-05, "loss": 0.0298, "step": 14190 }, { "action_loss": 0.008589521050453186, "epoch": 12.760791366906474, "step": 14190 }, { "epoch": 12.760791366906474, "step": 14190, "torque_loss": 0.12062601000070572 }, { "epoch": 12.76978417266187, "grad_norm": 0.1520964354276657, "learning_rate": 8.576448573759332e-05, "loss": 0.0284, "step": 14200 }, { "action_loss": 0.00448227534070611, "epoch": 12.76978417266187, "step": 14200 }, { "epoch": 12.76978417266187, "step": 14200, "torque_loss": 0.18326221406459808 }, { "epoch": 12.778776978417266, "grad_norm": 0.18156994879245758, "learning_rate": 8.574136809567044e-05, "loss": 0.0314, "step": 14210 }, { "action_loss": 0.005484896246343851, "epoch": 12.778776978417266, "step": 14210 }, { "epoch": 12.778776978417266, "step": 14210, "torque_loss": 0.17273831367492676 }, { "epoch": 12.787769784172662, "grad_norm": 0.1314326524734497, "learning_rate": 8.57182348192589e-05, "loss": 0.0278, "step": 14220 }, { "action_loss": 0.008402419276535511, "epoch": 12.787769784172662, "step": 14220 }, { "epoch": 12.787769784172662, "step": 14220, "torque_loss": 0.15353767573833466 }, { "epoch": 12.796762589928058, "grad_norm": 0.134739950299263, "learning_rate": 8.569508591847792e-05, "loss": 0.0276, "step": 14230 }, { "action_loss": 0.005975306499749422, "epoch": 12.796762589928058, "step": 14230 }, { "epoch": 12.796762589928058, "step": 14230, "torque_loss": 0.13206849992275238 }, { "epoch": 12.805755395683454, "grad_norm": 0.2611387073993683, "learning_rate": 8.567192140345367e-05, "loss": 0.0321, "step": 14240 }, { "action_loss": 0.007509020622819662, "epoch": 12.805755395683454, "step": 14240 }, { "epoch": 12.805755395683454, "step": 14240, "torque_loss": 0.1398455649614334 }, { "epoch": 12.81474820143885, "grad_norm": 0.14935310184955597, "learning_rate": 8.564874128431906e-05, "loss": 0.0255, "step": 14250 }, { "action_loss": 0.0050607905723154545, "epoch": 12.81474820143885, "step": 14250 }, { "epoch": 12.81474820143885, "step": 14250, "torque_loss": 0.17700429260730743 }, { "epoch": 12.823741007194245, "grad_norm": 0.19804422557353973, "learning_rate": 8.562554557121389e-05, "loss": 0.0295, "step": 14260 }, { "action_loss": 0.005357346031814814, "epoch": 12.823741007194245, "step": 14260 }, { "epoch": 12.823741007194245, "step": 14260, "torque_loss": 0.1496833860874176 }, { "epoch": 12.832733812949641, "grad_norm": 0.20404033362865448, "learning_rate": 8.560233427428475e-05, "loss": 0.0298, "step": 14270 }, { "action_loss": 0.008582540787756443, "epoch": 12.832733812949641, "step": 14270 }, { "epoch": 12.832733812949641, "step": 14270, "torque_loss": 0.14672265946865082 }, { "epoch": 12.841726618705035, "grad_norm": 0.15363521873950958, "learning_rate": 8.557910740368503e-05, "loss": 0.0324, "step": 14280 }, { "action_loss": 0.015489146113395691, "epoch": 12.841726618705035, "step": 14280 }, { "epoch": 12.841726618705035, "step": 14280, "torque_loss": 0.2799747884273529 }, { "epoch": 12.850719424460431, "grad_norm": 0.26343733072280884, "learning_rate": 8.555586496957495e-05, "loss": 0.0355, "step": 14290 }, { "action_loss": 0.008443809114396572, "epoch": 12.850719424460431, "step": 14290 }, { "epoch": 12.850719424460431, "step": 14290, "torque_loss": 0.19550581276416779 }, { "epoch": 12.859712230215827, "grad_norm": 0.21162909269332886, "learning_rate": 8.553260698212155e-05, "loss": 0.0255, "step": 14300 }, { "action_loss": 0.03489687666296959, "epoch": 12.859712230215827, "step": 14300 }, { "epoch": 12.859712230215827, "step": 14300, "torque_loss": 0.22943419218063354 }, { "epoch": 12.868705035971223, "grad_norm": 0.15795579552650452, "learning_rate": 8.550933345149868e-05, "loss": 0.0333, "step": 14310 }, { "action_loss": 0.005489621311426163, "epoch": 12.868705035971223, "step": 14310 }, { "epoch": 12.868705035971223, "step": 14310, "torque_loss": 0.1625135838985443 }, { "epoch": 12.877697841726619, "grad_norm": 0.2337634414434433, "learning_rate": 8.548604438788696e-05, "loss": 0.0291, "step": 14320 }, { "action_loss": 0.012372669763863087, "epoch": 12.877697841726619, "step": 14320 }, { "epoch": 12.877697841726619, "step": 14320, "torque_loss": 0.2051544338464737 }, { "epoch": 12.886690647482014, "grad_norm": 0.1917862594127655, "learning_rate": 8.546273980147383e-05, "loss": 0.0324, "step": 14330 }, { "action_loss": 0.004897570237517357, "epoch": 12.886690647482014, "step": 14330 }, { "epoch": 12.886690647482014, "step": 14330, "torque_loss": 0.18191592395305634 }, { "epoch": 12.89568345323741, "grad_norm": 0.22361145913600922, "learning_rate": 8.543941970245348e-05, "loss": 0.0275, "step": 14340 }, { "action_loss": 0.02309895120561123, "epoch": 12.89568345323741, "step": 14340 }, { "epoch": 12.89568345323741, "step": 14340, "torque_loss": 0.17901627719402313 }, { "epoch": 12.904676258992806, "grad_norm": 0.23548567295074463, "learning_rate": 8.541608410102693e-05, "loss": 0.0299, "step": 14350 }, { "action_loss": 0.006832033395767212, "epoch": 12.904676258992806, "step": 14350 }, { "epoch": 12.904676258992806, "step": 14350, "torque_loss": 0.14190083742141724 }, { "epoch": 12.913669064748202, "grad_norm": 0.16329969465732574, "learning_rate": 8.539273300740195e-05, "loss": 0.0295, "step": 14360 }, { "action_loss": 0.013045174069702625, "epoch": 12.913669064748202, "step": 14360 }, { "epoch": 12.913669064748202, "step": 14360, "torque_loss": 0.14924898743629456 }, { "epoch": 12.922661870503598, "grad_norm": 0.12262631952762604, "learning_rate": 8.536936643179313e-05, "loss": 0.0329, "step": 14370 }, { "action_loss": 0.013122498989105225, "epoch": 12.922661870503598, "step": 14370 }, { "epoch": 12.922661870503598, "step": 14370, "torque_loss": 0.20336513221263885 }, { "epoch": 12.931654676258994, "grad_norm": 0.17296084761619568, "learning_rate": 8.534598438442179e-05, "loss": 0.0339, "step": 14380 }, { "action_loss": 0.011918255127966404, "epoch": 12.931654676258994, "step": 14380 }, { "epoch": 12.931654676258994, "step": 14380, "torque_loss": 0.20961900055408478 }, { "epoch": 12.940647482014388, "grad_norm": 0.227578803896904, "learning_rate": 8.532258687551603e-05, "loss": 0.027, "step": 14390 }, { "action_loss": 0.011316847987473011, "epoch": 12.940647482014388, "step": 14390 }, { "epoch": 12.940647482014388, "step": 14390, "torque_loss": 0.2462724894285202 }, { "epoch": 12.949640287769784, "grad_norm": 0.15177172422409058, "learning_rate": 8.529917391531071e-05, "loss": 0.03, "step": 14400 }, { "action_loss": 0.004767971113324165, "epoch": 12.949640287769784, "step": 14400 }, { "epoch": 12.949640287769784, "step": 14400, "torque_loss": 0.16986016929149628 }, { "epoch": 12.95863309352518, "grad_norm": 0.1448248028755188, "learning_rate": 8.527574551404747e-05, "loss": 0.0255, "step": 14410 }, { "action_loss": 0.0036569281946867704, "epoch": 12.95863309352518, "step": 14410 }, { "epoch": 12.95863309352518, "step": 14410, "torque_loss": 0.19845564663410187 }, { "epoch": 12.967625899280575, "grad_norm": 0.22834846377372742, "learning_rate": 8.525230168197468e-05, "loss": 0.0259, "step": 14420 }, { "action_loss": 0.0032851677387952805, "epoch": 12.967625899280575, "step": 14420 }, { "epoch": 12.967625899280575, "step": 14420, "torque_loss": 0.12870970368385315 }, { "epoch": 12.976618705035971, "grad_norm": 0.2209164798259735, "learning_rate": 8.522884242934745e-05, "loss": 0.0356, "step": 14430 }, { "action_loss": 0.005844240542501211, "epoch": 12.976618705035971, "step": 14430 }, { "epoch": 12.976618705035971, "step": 14430, "torque_loss": 0.16976690292358398 }, { "epoch": 12.985611510791367, "grad_norm": 0.1764727532863617, "learning_rate": 8.520536776642768e-05, "loss": 0.0369, "step": 14440 }, { "action_loss": 0.00698113115504384, "epoch": 12.985611510791367, "step": 14440 }, { "epoch": 12.985611510791367, "step": 14440, "torque_loss": 0.20558901131153107 }, { "epoch": 12.994604316546763, "grad_norm": 0.1983625441789627, "learning_rate": 8.5181877703484e-05, "loss": 0.0298, "step": 14450 }, { "action_loss": 0.01406185608357191, "epoch": 12.994604316546763, "step": 14450 }, { "epoch": 12.994604316546763, "step": 14450, "torque_loss": 0.1783262938261032 }, { "epoch": 13.003597122302159, "grad_norm": 0.24503201246261597, "learning_rate": 8.51583722507917e-05, "loss": 0.0264, "step": 14460 }, { "action_loss": 0.0038624757435172796, "epoch": 13.003597122302159, "step": 14460 }, { "epoch": 13.003597122302159, "step": 14460, "torque_loss": 0.12693606317043304 }, { "epoch": 13.012589928057555, "grad_norm": 0.23146101832389832, "learning_rate": 8.513485141863293e-05, "loss": 0.0287, "step": 14470 }, { "action_loss": 0.008080572821199894, "epoch": 13.012589928057555, "step": 14470 }, { "epoch": 13.012589928057555, "step": 14470, "torque_loss": 0.22116689383983612 }, { "epoch": 13.02158273381295, "grad_norm": 0.19682364165782928, "learning_rate": 8.511131521729647e-05, "loss": 0.0316, "step": 14480 }, { "action_loss": 0.005304162856191397, "epoch": 13.02158273381295, "step": 14480 }, { "epoch": 13.02158273381295, "step": 14480, "torque_loss": 0.1255100816488266 }, { "epoch": 13.030575539568344, "grad_norm": 0.1617492288351059, "learning_rate": 8.508776365707787e-05, "loss": 0.0266, "step": 14490 }, { "action_loss": 0.006797282490879297, "epoch": 13.030575539568344, "step": 14490 }, { "epoch": 13.030575539568344, "step": 14490, "torque_loss": 0.1731594353914261 }, { "epoch": 13.03956834532374, "grad_norm": 0.18877705931663513, "learning_rate": 8.506419674827934e-05, "loss": 0.0282, "step": 14500 }, { "action_loss": 0.010000627487897873, "epoch": 13.03956834532374, "step": 14500 }, { "epoch": 13.03956834532374, "step": 14500, "torque_loss": 0.21487902104854584 }, { "epoch": 13.048561151079136, "grad_norm": 0.3361363410949707, "learning_rate": 8.50406145012099e-05, "loss": 0.03, "step": 14510 }, { "action_loss": 0.005209268536418676, "epoch": 13.048561151079136, "step": 14510 }, { "epoch": 13.048561151079136, "step": 14510, "torque_loss": 0.09855590015649796 }, { "epoch": 13.057553956834532, "grad_norm": 0.24827119708061218, "learning_rate": 8.501701692618519e-05, "loss": 0.0304, "step": 14520 }, { "action_loss": 0.01102709025144577, "epoch": 13.057553956834532, "step": 14520 }, { "epoch": 13.057553956834532, "step": 14520, "torque_loss": 0.17111073434352875 }, { "epoch": 13.066546762589928, "grad_norm": 0.15930984914302826, "learning_rate": 8.499340403352761e-05, "loss": 0.031, "step": 14530 }, { "action_loss": 0.008980818092823029, "epoch": 13.066546762589928, "step": 14530 }, { "epoch": 13.066546762589928, "step": 14530, "torque_loss": 0.16246762871742249 }, { "epoch": 13.075539568345324, "grad_norm": 0.21571609377861023, "learning_rate": 8.496977583356623e-05, "loss": 0.0289, "step": 14540 }, { "action_loss": 0.0037248830776661634, "epoch": 13.075539568345324, "step": 14540 }, { "epoch": 13.075539568345324, "step": 14540, "torque_loss": 0.15840628743171692 }, { "epoch": 13.08453237410072, "grad_norm": 0.1982259601354599, "learning_rate": 8.494613233663684e-05, "loss": 0.0259, "step": 14550 }, { "action_loss": 0.004489060956984758, "epoch": 13.08453237410072, "step": 14550 }, { "epoch": 13.08453237410072, "step": 14550, "torque_loss": 0.14810886979103088 }, { "epoch": 13.093525179856115, "grad_norm": 0.16830973327159882, "learning_rate": 8.492247355308189e-05, "loss": 0.0253, "step": 14560 }, { "action_loss": 0.006629097741097212, "epoch": 13.093525179856115, "step": 14560 }, { "epoch": 13.093525179856115, "step": 14560, "torque_loss": 0.16877801716327667 }, { "epoch": 13.102517985611511, "grad_norm": 0.2571273148059845, "learning_rate": 8.489879949325056e-05, "loss": 0.0245, "step": 14570 }, { "action_loss": 0.019234435632824898, "epoch": 13.102517985611511, "step": 14570 }, { "epoch": 13.102517985611511, "step": 14570, "torque_loss": 0.2640039026737213 }, { "epoch": 13.111510791366907, "grad_norm": 0.20412921905517578, "learning_rate": 8.487511016749868e-05, "loss": 0.0313, "step": 14580 }, { "action_loss": 0.018253536894917488, "epoch": 13.111510791366907, "step": 14580 }, { "epoch": 13.111510791366907, "step": 14580, "torque_loss": 0.25172486901283264 }, { "epoch": 13.120503597122303, "grad_norm": 0.2752811908721924, "learning_rate": 8.485140558618874e-05, "loss": 0.0342, "step": 14590 }, { "action_loss": 0.008097507990896702, "epoch": 13.120503597122303, "step": 14590 }, { "epoch": 13.120503597122303, "step": 14590, "torque_loss": 0.12355483323335648 }, { "epoch": 13.129496402877697, "grad_norm": 0.13315004110336304, "learning_rate": 8.482768575968995e-05, "loss": 0.027, "step": 14600 }, { "action_loss": 0.009922287426888943, "epoch": 13.129496402877697, "step": 14600 }, { "epoch": 13.129496402877697, "step": 14600, "torque_loss": 0.13269223272800446 }, { "epoch": 13.138489208633093, "grad_norm": 0.1301357001066208, "learning_rate": 8.480395069837818e-05, "loss": 0.0325, "step": 14610 }, { "action_loss": 0.024008750915527344, "epoch": 13.138489208633093, "step": 14610 }, { "epoch": 13.138489208633093, "step": 14610, "torque_loss": 0.27067771553993225 }, { "epoch": 13.147482014388489, "grad_norm": 0.15632355213165283, "learning_rate": 8.478020041263595e-05, "loss": 0.0317, "step": 14620 }, { "action_loss": 0.0070924521423876286, "epoch": 13.147482014388489, "step": 14620 }, { "epoch": 13.147482014388489, "step": 14620, "torque_loss": 0.16045419871807098 }, { "epoch": 13.156474820143885, "grad_norm": 0.14577145874500275, "learning_rate": 8.475643491285242e-05, "loss": 0.0258, "step": 14630 }, { "action_loss": 0.0056858607567846775, "epoch": 13.156474820143885, "step": 14630 }, { "epoch": 13.156474820143885, "step": 14630, "torque_loss": 0.16915206611156464 }, { "epoch": 13.16546762589928, "grad_norm": 0.2226993888616562, "learning_rate": 8.473265420942345e-05, "loss": 0.027, "step": 14640 }, { "action_loss": 0.01239150669425726, "epoch": 13.16546762589928, "step": 14640 }, { "epoch": 13.16546762589928, "step": 14640, "torque_loss": 0.2702014148235321 }, { "epoch": 13.174460431654676, "grad_norm": 0.2424813061952591, "learning_rate": 8.470885831275151e-05, "loss": 0.0282, "step": 14650 }, { "action_loss": 0.013514656573534012, "epoch": 13.174460431654676, "step": 14650 }, { "epoch": 13.174460431654676, "step": 14650, "torque_loss": 0.1984085887670517 }, { "epoch": 13.183453237410072, "grad_norm": 0.21910570561885834, "learning_rate": 8.468504723324574e-05, "loss": 0.031, "step": 14660 }, { "action_loss": 0.007414929568767548, "epoch": 13.183453237410072, "step": 14660 }, { "epoch": 13.183453237410072, "step": 14660, "torque_loss": 0.13029487431049347 }, { "epoch": 13.192446043165468, "grad_norm": 0.15684321522712708, "learning_rate": 8.466122098132193e-05, "loss": 0.0248, "step": 14670 }, { "action_loss": 0.01091399509459734, "epoch": 13.192446043165468, "step": 14670 }, { "epoch": 13.192446043165468, "step": 14670, "torque_loss": 0.13976258039474487 }, { "epoch": 13.201438848920864, "grad_norm": 0.1669568419456482, "learning_rate": 8.463737956740245e-05, "loss": 0.0325, "step": 14680 }, { "action_loss": 0.01623978279531002, "epoch": 13.201438848920864, "step": 14680 }, { "epoch": 13.201438848920864, "step": 14680, "torque_loss": 0.19226162135601044 }, { "epoch": 13.21043165467626, "grad_norm": 0.30810075998306274, "learning_rate": 8.461352300191639e-05, "loss": 0.0318, "step": 14690 }, { "action_loss": 0.010056703351438046, "epoch": 13.21043165467626, "step": 14690 }, { "epoch": 13.21043165467626, "step": 14690, "torque_loss": 0.1737358570098877 }, { "epoch": 13.219424460431656, "grad_norm": 0.19646821916103363, "learning_rate": 8.45896512952994e-05, "loss": 0.0264, "step": 14700 }, { "action_loss": 0.010671086609363556, "epoch": 13.219424460431656, "step": 14700 }, { "epoch": 13.219424460431656, "step": 14700, "torque_loss": 0.17081184685230255 }, { "epoch": 13.22841726618705, "grad_norm": 0.16733834147453308, "learning_rate": 8.456576445799377e-05, "loss": 0.0276, "step": 14710 }, { "action_loss": 0.007681468967348337, "epoch": 13.22841726618705, "step": 14710 }, { "epoch": 13.22841726618705, "step": 14710, "torque_loss": 0.18353891372680664 }, { "epoch": 13.237410071942445, "grad_norm": 0.17833010852336884, "learning_rate": 8.454186250044844e-05, "loss": 0.0275, "step": 14720 }, { "action_loss": 0.009167798794806004, "epoch": 13.237410071942445, "step": 14720 }, { "epoch": 13.237410071942445, "step": 14720, "torque_loss": 0.17262141406536102 }, { "epoch": 13.246402877697841, "grad_norm": 0.1754665970802307, "learning_rate": 8.451794543311892e-05, "loss": 0.0245, "step": 14730 }, { "action_loss": 0.0058291577734053135, "epoch": 13.246402877697841, "step": 14730 }, { "epoch": 13.246402877697841, "step": 14730, "torque_loss": 0.144082710146904 }, { "epoch": 13.255395683453237, "grad_norm": 0.15958864986896515, "learning_rate": 8.449401326646736e-05, "loss": 0.0298, "step": 14740 }, { "action_loss": 0.013366654515266418, "epoch": 13.255395683453237, "step": 14740 }, { "epoch": 13.255395683453237, "step": 14740, "torque_loss": 0.19165027141571045 }, { "epoch": 13.264388489208633, "grad_norm": 0.16917292773723602, "learning_rate": 8.447006601096248e-05, "loss": 0.0341, "step": 14750 }, { "action_loss": 0.008074288256466389, "epoch": 13.264388489208633, "step": 14750 }, { "epoch": 13.264388489208633, "step": 14750, "torque_loss": 0.22471050918102264 }, { "epoch": 13.273381294964029, "grad_norm": 0.22097310423851013, "learning_rate": 8.444610367707964e-05, "loss": 0.0257, "step": 14760 }, { "action_loss": 0.02045709826052189, "epoch": 13.273381294964029, "step": 14760 }, { "epoch": 13.273381294964029, "step": 14760, "torque_loss": 0.22345715761184692 }, { "epoch": 13.282374100719425, "grad_norm": 0.262327641248703, "learning_rate": 8.442212627530078e-05, "loss": 0.0395, "step": 14770 }, { "action_loss": 0.010301532223820686, "epoch": 13.282374100719425, "step": 14770 }, { "epoch": 13.282374100719425, "step": 14770, "torque_loss": 0.19923050701618195 }, { "epoch": 13.29136690647482, "grad_norm": 0.17741473019123077, "learning_rate": 8.439813381611441e-05, "loss": 0.0263, "step": 14780 }, { "action_loss": 0.005328765604645014, "epoch": 13.29136690647482, "step": 14780 }, { "epoch": 13.29136690647482, "step": 14780, "torque_loss": 0.13162636756896973 }, { "epoch": 13.300359712230216, "grad_norm": 0.17519335448741913, "learning_rate": 8.437412631001567e-05, "loss": 0.031, "step": 14790 }, { "action_loss": 0.007286790758371353, "epoch": 13.300359712230216, "step": 14790 }, { "epoch": 13.300359712230216, "step": 14790, "torque_loss": 0.21255941689014435 }, { "epoch": 13.309352517985612, "grad_norm": 0.16458463668823242, "learning_rate": 8.435010376750626e-05, "loss": 0.026, "step": 14800 }, { "action_loss": 0.005067877005785704, "epoch": 13.309352517985612, "step": 14800 }, { "epoch": 13.309352517985612, "step": 14800, "torque_loss": 0.1584402471780777 }, { "epoch": 13.318345323741006, "grad_norm": 0.2572159469127655, "learning_rate": 8.432606619909442e-05, "loss": 0.0228, "step": 14810 }, { "action_loss": 0.035323817282915115, "epoch": 13.318345323741006, "step": 14810 }, { "epoch": 13.318345323741006, "step": 14810, "torque_loss": 0.23175446689128876 }, { "epoch": 13.327338129496402, "grad_norm": 0.13273318111896515, "learning_rate": 8.430201361529506e-05, "loss": 0.0291, "step": 14820 }, { "action_loss": 0.007183941546827555, "epoch": 13.327338129496402, "step": 14820 }, { "epoch": 13.327338129496402, "step": 14820, "torque_loss": 0.15050463378429413 }, { "epoch": 13.336330935251798, "grad_norm": 0.16368672251701355, "learning_rate": 8.427794602662954e-05, "loss": 0.0276, "step": 14830 }, { "action_loss": 0.02285233698785305, "epoch": 13.336330935251798, "step": 14830 }, { "epoch": 13.336330935251798, "step": 14830, "torque_loss": 0.20539319515228271 }, { "epoch": 13.345323741007194, "grad_norm": 0.19519126415252686, "learning_rate": 8.425386344362586e-05, "loss": 0.0312, "step": 14840 }, { "action_loss": 0.005631196778267622, "epoch": 13.345323741007194, "step": 14840 }, { "epoch": 13.345323741007194, "step": 14840, "torque_loss": 0.14546506106853485 }, { "epoch": 13.35431654676259, "grad_norm": 0.16537103056907654, "learning_rate": 8.422976587681859e-05, "loss": 0.0269, "step": 14850 }, { "action_loss": 0.0072525800205767155, "epoch": 13.35431654676259, "step": 14850 }, { "epoch": 13.35431654676259, "step": 14850, "torque_loss": 0.19681744277477264 }, { "epoch": 13.363309352517986, "grad_norm": 0.22983983159065247, "learning_rate": 8.42056533367488e-05, "loss": 0.0309, "step": 14860 }, { "action_loss": 0.011476523242890835, "epoch": 13.363309352517986, "step": 14860 }, { "epoch": 13.363309352517986, "step": 14860, "torque_loss": 0.1831885427236557 }, { "epoch": 13.372302158273381, "grad_norm": 0.24088549613952637, "learning_rate": 8.41815258339641e-05, "loss": 0.0319, "step": 14870 }, { "action_loss": 0.010860812850296497, "epoch": 13.372302158273381, "step": 14870 }, { "epoch": 13.372302158273381, "step": 14870, "torque_loss": 0.17623616755008698 }, { "epoch": 13.381294964028777, "grad_norm": 0.14103935658931732, "learning_rate": 8.415738337901874e-05, "loss": 0.0264, "step": 14880 }, { "action_loss": 0.0038755263667553663, "epoch": 13.381294964028777, "step": 14880 }, { "epoch": 13.381294964028777, "step": 14880, "torque_loss": 0.18165616691112518 }, { "epoch": 13.390287769784173, "grad_norm": 0.1781671792268753, "learning_rate": 8.413322598247342e-05, "loss": 0.0296, "step": 14890 }, { "action_loss": 0.011688471771776676, "epoch": 13.390287769784173, "step": 14890 }, { "epoch": 13.390287769784173, "step": 14890, "torque_loss": 0.1531115621328354 }, { "epoch": 13.399280575539569, "grad_norm": 0.16274413466453552, "learning_rate": 8.41090536548954e-05, "loss": 0.0224, "step": 14900 }, { "action_loss": 0.0024776088539510965, "epoch": 13.399280575539569, "step": 14900 }, { "epoch": 13.399280575539569, "step": 14900, "torque_loss": 0.06799498200416565 }, { "epoch": 13.408273381294965, "grad_norm": 0.09280410408973694, "learning_rate": 8.408486640685849e-05, "loss": 0.0211, "step": 14910 }, { "action_loss": 0.030977316200733185, "epoch": 13.408273381294965, "step": 14910 }, { "epoch": 13.408273381294965, "step": 14910, "torque_loss": 0.3241472542285919 }, { "epoch": 13.417266187050359, "grad_norm": 0.13736110925674438, "learning_rate": 8.4060664248943e-05, "loss": 0.0324, "step": 14920 }, { "action_loss": 0.005020348355174065, "epoch": 13.417266187050359, "step": 14920 }, { "epoch": 13.417266187050359, "step": 14920, "torque_loss": 0.17478500306606293 }, { "epoch": 13.426258992805755, "grad_norm": 0.14139464497566223, "learning_rate": 8.40364471917358e-05, "loss": 0.0294, "step": 14930 }, { "action_loss": 0.007305748760700226, "epoch": 13.426258992805755, "step": 14930 }, { "epoch": 13.426258992805755, "step": 14930, "torque_loss": 0.19416797161102295 }, { "epoch": 13.43525179856115, "grad_norm": 0.26027965545654297, "learning_rate": 8.401221524583024e-05, "loss": 0.0271, "step": 14940 }, { "action_loss": 0.004550204146653414, "epoch": 13.43525179856115, "step": 14940 }, { "epoch": 13.43525179856115, "step": 14940, "torque_loss": 0.20384447276592255 }, { "epoch": 13.444244604316546, "grad_norm": 0.2563116252422333, "learning_rate": 8.398796842182619e-05, "loss": 0.0308, "step": 14950 }, { "action_loss": 0.012186425738036633, "epoch": 13.444244604316546, "step": 14950 }, { "epoch": 13.444244604316546, "step": 14950, "torque_loss": 0.21431533992290497 }, { "epoch": 13.453237410071942, "grad_norm": 0.2518470883369446, "learning_rate": 8.396370673033006e-05, "loss": 0.0286, "step": 14960 }, { "action_loss": 0.00923926755785942, "epoch": 13.453237410071942, "step": 14960 }, { "epoch": 13.453237410071942, "step": 14960, "torque_loss": 0.16263365745544434 }, { "epoch": 13.462230215827338, "grad_norm": 0.1338990181684494, "learning_rate": 8.39394301819547e-05, "loss": 0.0346, "step": 14970 }, { "action_loss": 0.01104278489947319, "epoch": 13.462230215827338, "step": 14970 }, { "epoch": 13.462230215827338, "step": 14970, "torque_loss": 0.1751183271408081 }, { "epoch": 13.471223021582734, "grad_norm": 0.18603111803531647, "learning_rate": 8.391513878731949e-05, "loss": 0.0292, "step": 14980 }, { "action_loss": 0.007716303691267967, "epoch": 13.471223021582734, "step": 14980 }, { "epoch": 13.471223021582734, "step": 14980, "torque_loss": 0.22590310871601105 }, { "epoch": 13.48021582733813, "grad_norm": 0.25837817788124084, "learning_rate": 8.389083255705037e-05, "loss": 0.03, "step": 14990 }, { "action_loss": 0.008414355106651783, "epoch": 13.48021582733813, "step": 14990 }, { "epoch": 13.48021582733813, "step": 14990, "torque_loss": 0.21487663686275482 }, { "epoch": 13.489208633093526, "grad_norm": 0.2049047350883484, "learning_rate": 8.386651150177968e-05, "loss": 0.0301, "step": 15000 }, { "action_loss": 0.007454264909029007, "epoch": 13.489208633093526, "step": 15000 }, { "epoch": 13.489208633093526, "step": 15000, "torque_loss": 0.261800080537796 }, { "epoch": 13.498201438848922, "grad_norm": 0.12070826441049576, "learning_rate": 8.384217563214627e-05, "loss": 0.0267, "step": 15010 }, { "action_loss": 0.018477745354175568, "epoch": 13.498201438848922, "step": 15010 }, { "epoch": 13.498201438848922, "step": 15010, "torque_loss": 0.2102290242910385 }, { "epoch": 13.507194244604317, "grad_norm": 0.1691012680530548, "learning_rate": 8.381782495879546e-05, "loss": 0.0236, "step": 15020 }, { "action_loss": 0.006040126085281372, "epoch": 13.507194244604317, "step": 15020 }, { "epoch": 13.507194244604317, "step": 15020, "torque_loss": 0.22561068832874298 }, { "epoch": 13.516187050359711, "grad_norm": 0.17865873873233795, "learning_rate": 8.37934594923791e-05, "loss": 0.0323, "step": 15030 }, { "action_loss": 0.004458179231733084, "epoch": 13.516187050359711, "step": 15030 }, { "epoch": 13.516187050359711, "step": 15030, "torque_loss": 0.1269637793302536 }, { "epoch": 13.525179856115107, "grad_norm": 0.14457544684410095, "learning_rate": 8.376907924355546e-05, "loss": 0.0291, "step": 15040 }, { "action_loss": 0.0052194963209331036, "epoch": 13.525179856115107, "step": 15040 }, { "epoch": 13.525179856115107, "step": 15040, "torque_loss": 0.22602993249893188 }, { "epoch": 13.534172661870503, "grad_norm": 0.17838545143604279, "learning_rate": 8.374468422298926e-05, "loss": 0.027, "step": 15050 }, { "action_loss": 0.009122132323682308, "epoch": 13.534172661870503, "step": 15050 }, { "epoch": 13.534172661870503, "step": 15050, "torque_loss": 0.20560558140277863 }, { "epoch": 13.543165467625899, "grad_norm": 0.23026098310947418, "learning_rate": 8.372027444135176e-05, "loss": 0.0314, "step": 15060 }, { "action_loss": 0.015016310848295689, "epoch": 13.543165467625899, "step": 15060 }, { "epoch": 13.543165467625899, "step": 15060, "torque_loss": 0.22769713401794434 }, { "epoch": 13.552158273381295, "grad_norm": 0.22403854131698608, "learning_rate": 8.36958499093206e-05, "loss": 0.0269, "step": 15070 }, { "action_loss": 0.004735040944069624, "epoch": 13.552158273381295, "step": 15070 }, { "epoch": 13.552158273381295, "step": 15070, "torque_loss": 0.14384539425373077 }, { "epoch": 13.56115107913669, "grad_norm": 0.1701783984899521, "learning_rate": 8.367141063757988e-05, "loss": 0.0265, "step": 15080 }, { "action_loss": 0.006069012451916933, "epoch": 13.56115107913669, "step": 15080 }, { "epoch": 13.56115107913669, "step": 15080, "torque_loss": 0.1704898476600647 }, { "epoch": 13.570143884892087, "grad_norm": 0.16661737859249115, "learning_rate": 8.364695663682022e-05, "loss": 0.0279, "step": 15090 }, { "action_loss": 0.009942124597728252, "epoch": 13.570143884892087, "step": 15090 }, { "epoch": 13.570143884892087, "step": 15090, "torque_loss": 0.15727750957012177 }, { "epoch": 13.579136690647482, "grad_norm": 0.19281336665153503, "learning_rate": 8.362248791773858e-05, "loss": 0.0279, "step": 15100 }, { "action_loss": 0.006407874170690775, "epoch": 13.579136690647482, "step": 15100 }, { "epoch": 13.579136690647482, "step": 15100, "torque_loss": 0.21607427299022675 }, { "epoch": 13.588129496402878, "grad_norm": 0.13875405490398407, "learning_rate": 8.359800449103842e-05, "loss": 0.0245, "step": 15110 }, { "action_loss": 0.007619692478328943, "epoch": 13.588129496402878, "step": 15110 }, { "epoch": 13.588129496402878, "step": 15110, "torque_loss": 0.16798466444015503 }, { "epoch": 13.597122302158274, "grad_norm": 0.18619704246520996, "learning_rate": 8.357350636742967e-05, "loss": 0.0281, "step": 15120 }, { "action_loss": 0.009690687991678715, "epoch": 13.597122302158274, "step": 15120 }, { "epoch": 13.597122302158274, "step": 15120, "torque_loss": 0.1302601546049118 }, { "epoch": 13.60611510791367, "grad_norm": 0.11425930261611938, "learning_rate": 8.354899355762858e-05, "loss": 0.0225, "step": 15130 }, { "action_loss": 0.005365763325244188, "epoch": 13.60611510791367, "step": 15130 }, { "epoch": 13.60611510791367, "step": 15130, "torque_loss": 0.17414361238479614 }, { "epoch": 13.615107913669064, "grad_norm": 0.185357928276062, "learning_rate": 8.352446607235791e-05, "loss": 0.0345, "step": 15140 }, { "action_loss": 0.0030274074524641037, "epoch": 13.615107913669064, "step": 15140 }, { "epoch": 13.615107913669064, "step": 15140, "torque_loss": 0.12594188749790192 }, { "epoch": 13.62410071942446, "grad_norm": 0.20031195878982544, "learning_rate": 8.349992392234683e-05, "loss": 0.0269, "step": 15150 }, { "action_loss": 0.006930793169885874, "epoch": 13.62410071942446, "step": 15150 }, { "epoch": 13.62410071942446, "step": 15150, "torque_loss": 0.18283069133758545 }, { "epoch": 13.633093525179856, "grad_norm": 0.28024429082870483, "learning_rate": 8.347536711833088e-05, "loss": 0.0354, "step": 15160 }, { "action_loss": 0.006271719932556152, "epoch": 13.633093525179856, "step": 15160 }, { "epoch": 13.633093525179856, "step": 15160, "torque_loss": 0.1496305763721466 }, { "epoch": 13.642086330935252, "grad_norm": 0.2723066806793213, "learning_rate": 8.345079567105206e-05, "loss": 0.0299, "step": 15170 }, { "action_loss": 0.0061704739928245544, "epoch": 13.642086330935252, "step": 15170 }, { "epoch": 13.642086330935252, "step": 15170, "torque_loss": 0.12738679349422455 }, { "epoch": 13.651079136690647, "grad_norm": 0.16911423206329346, "learning_rate": 8.342620959125875e-05, "loss": 0.0277, "step": 15180 }, { "action_loss": 0.005297248717397451, "epoch": 13.651079136690647, "step": 15180 }, { "epoch": 13.651079136690647, "step": 15180, "torque_loss": 0.18191631138324738 }, { "epoch": 13.660071942446043, "grad_norm": 0.19064965844154358, "learning_rate": 8.340160888970574e-05, "loss": 0.0307, "step": 15190 }, { "action_loss": 0.013848207890987396, "epoch": 13.660071942446043, "step": 15190 }, { "epoch": 13.660071942446043, "step": 15190, "torque_loss": 0.18009133636951447 }, { "epoch": 13.66906474820144, "grad_norm": 0.21504659950733185, "learning_rate": 8.337699357715422e-05, "loss": 0.0299, "step": 15200 }, { "action_loss": 0.005623711738735437, "epoch": 13.66906474820144, "step": 15200 }, { "epoch": 13.66906474820144, "step": 15200, "torque_loss": 0.1421443372964859 }, { "epoch": 13.678057553956835, "grad_norm": 0.18499258160591125, "learning_rate": 8.335236366437174e-05, "loss": 0.0246, "step": 15210 }, { "action_loss": 0.005163074936717749, "epoch": 13.678057553956835, "step": 15210 }, { "epoch": 13.678057553956835, "step": 15210, "torque_loss": 0.14711187779903412 }, { "epoch": 13.68705035971223, "grad_norm": 0.2858845889568329, "learning_rate": 8.332771916213227e-05, "loss": 0.0264, "step": 15220 }, { "action_loss": 0.008339453488588333, "epoch": 13.68705035971223, "step": 15220 }, { "epoch": 13.68705035971223, "step": 15220, "torque_loss": 0.13850398361682892 }, { "epoch": 13.696043165467627, "grad_norm": 0.165247842669487, "learning_rate": 8.330306008121618e-05, "loss": 0.0262, "step": 15230 }, { "action_loss": 0.00511687807738781, "epoch": 13.696043165467627, "step": 15230 }, { "epoch": 13.696043165467627, "step": 15230, "torque_loss": 0.18347012996673584 }, { "epoch": 13.70503597122302, "grad_norm": 0.15709181129932404, "learning_rate": 8.327838643241014e-05, "loss": 0.025, "step": 15240 }, { "action_loss": 0.006151461973786354, "epoch": 13.70503597122302, "step": 15240 }, { "epoch": 13.70503597122302, "step": 15240, "torque_loss": 0.15336830914020538 }, { "epoch": 13.714028776978417, "grad_norm": 0.19233566522598267, "learning_rate": 8.32536982265073e-05, "loss": 0.029, "step": 15250 }, { "action_loss": 0.0032078928779810667, "epoch": 13.714028776978417, "step": 15250 }, { "epoch": 13.714028776978417, "step": 15250, "torque_loss": 0.08413007110357285 }, { "epoch": 13.723021582733812, "grad_norm": 0.2303439825773239, "learning_rate": 8.32289954743071e-05, "loss": 0.0262, "step": 15260 }, { "action_loss": 0.02366062067449093, "epoch": 13.723021582733812, "step": 15260 }, { "epoch": 13.723021582733812, "step": 15260, "torque_loss": 0.2028161734342575 }, { "epoch": 13.732014388489208, "grad_norm": 0.30171409249305725, "learning_rate": 8.320427818661532e-05, "loss": 0.0308, "step": 15270 }, { "action_loss": 0.004693672526627779, "epoch": 13.732014388489208, "step": 15270 }, { "epoch": 13.732014388489208, "step": 15270, "torque_loss": 0.1463063359260559 }, { "epoch": 13.741007194244604, "grad_norm": 0.24207988381385803, "learning_rate": 8.317954637424419e-05, "loss": 0.0234, "step": 15280 }, { "action_loss": 0.029937343671917915, "epoch": 13.741007194244604, "step": 15280 }, { "epoch": 13.741007194244604, "step": 15280, "torque_loss": 0.1700475811958313 }, { "epoch": 13.75, "grad_norm": 0.16039183735847473, "learning_rate": 8.315480004801221e-05, "loss": 0.0306, "step": 15290 }, { "action_loss": 0.004989855457097292, "epoch": 13.75, "step": 15290 }, { "epoch": 13.75, "step": 15290, "torque_loss": 0.15788237750530243 }, { "epoch": 13.758992805755396, "grad_norm": 0.22785283625125885, "learning_rate": 8.31300392187443e-05, "loss": 0.0274, "step": 15300 }, { "action_loss": 0.020298494026064873, "epoch": 13.758992805755396, "step": 15300 }, { "epoch": 13.758992805755396, "step": 15300, "torque_loss": 0.2193116694688797 }, { "epoch": 13.767985611510792, "grad_norm": 0.21620965003967285, "learning_rate": 8.31052638972717e-05, "loss": 0.0282, "step": 15310 }, { "action_loss": 0.006743166595697403, "epoch": 13.767985611510792, "step": 15310 }, { "epoch": 13.767985611510792, "step": 15310, "torque_loss": 0.19963575899600983 }, { "epoch": 13.776978417266188, "grad_norm": 0.11667854338884354, "learning_rate": 8.30804740944319e-05, "loss": 0.0253, "step": 15320 }, { "action_loss": 0.010897506028413773, "epoch": 13.776978417266188, "step": 15320 }, { "epoch": 13.776978417266188, "step": 15320, "torque_loss": 0.19620662927627563 }, { "epoch": 13.785971223021583, "grad_norm": 0.2114720344543457, "learning_rate": 8.305566982106887e-05, "loss": 0.03, "step": 15330 }, { "action_loss": 0.005128500983119011, "epoch": 13.785971223021583, "step": 15330 }, { "epoch": 13.785971223021583, "step": 15330, "torque_loss": 0.22651426494121552 }, { "epoch": 13.79496402877698, "grad_norm": 0.16168460249900818, "learning_rate": 8.303085108803282e-05, "loss": 0.025, "step": 15340 }, { "action_loss": 0.0038001176435500383, "epoch": 13.79496402877698, "step": 15340 }, { "epoch": 13.79496402877698, "step": 15340, "torque_loss": 0.18439610302448273 }, { "epoch": 13.803956834532373, "grad_norm": 0.12572243809700012, "learning_rate": 8.300601790618032e-05, "loss": 0.026, "step": 15350 }, { "action_loss": 0.006317974999547005, "epoch": 13.803956834532373, "step": 15350 }, { "epoch": 13.803956834532373, "step": 15350, "torque_loss": 0.18193064630031586 }, { "epoch": 13.81294964028777, "grad_norm": 0.1691419631242752, "learning_rate": 8.298117028637423e-05, "loss": 0.0269, "step": 15360 }, { "action_loss": 0.005181512329727411, "epoch": 13.81294964028777, "step": 15360 }, { "epoch": 13.81294964028777, "step": 15360, "torque_loss": 0.17988042533397675 }, { "epoch": 13.821942446043165, "grad_norm": 0.19126656651496887, "learning_rate": 8.295630823948375e-05, "loss": 0.0289, "step": 15370 }, { "action_loss": 0.012444384396076202, "epoch": 13.821942446043165, "step": 15370 }, { "epoch": 13.821942446043165, "step": 15370, "torque_loss": 0.23776547610759735 }, { "epoch": 13.83093525179856, "grad_norm": 0.20654746890068054, "learning_rate": 8.293143177638438e-05, "loss": 0.0266, "step": 15380 }, { "action_loss": 0.0026967469602823257, "epoch": 13.83093525179856, "step": 15380 }, { "epoch": 13.83093525179856, "step": 15380, "torque_loss": 0.10126584768295288 }, { "epoch": 13.839928057553957, "grad_norm": 0.3066680133342743, "learning_rate": 8.290654090795795e-05, "loss": 0.025, "step": 15390 }, { "action_loss": 0.010547250509262085, "epoch": 13.839928057553957, "step": 15390 }, { "epoch": 13.839928057553957, "step": 15390, "torque_loss": 0.1611822545528412 }, { "epoch": 13.848920863309353, "grad_norm": 0.19107399880886078, "learning_rate": 8.288163564509253e-05, "loss": 0.0273, "step": 15400 }, { "action_loss": 0.003785765962675214, "epoch": 13.848920863309353, "step": 15400 }, { "epoch": 13.848920863309353, "step": 15400, "torque_loss": 0.17067570984363556 }, { "epoch": 13.857913669064748, "grad_norm": 0.1921701431274414, "learning_rate": 8.285671599868257e-05, "loss": 0.0271, "step": 15410 }, { "action_loss": 0.008458774536848068, "epoch": 13.857913669064748, "step": 15410 }, { "epoch": 13.857913669064748, "step": 15410, "torque_loss": 0.21028001606464386 }, { "epoch": 13.866906474820144, "grad_norm": 0.20366641879081726, "learning_rate": 8.283178197962875e-05, "loss": 0.0295, "step": 15420 }, { "action_loss": 0.007012511137872934, "epoch": 13.866906474820144, "step": 15420 }, { "epoch": 13.866906474820144, "step": 15420, "torque_loss": 0.14455051720142365 }, { "epoch": 13.87589928057554, "grad_norm": 0.20931357145309448, "learning_rate": 8.280683359883808e-05, "loss": 0.0252, "step": 15430 }, { "action_loss": 0.006559660192579031, "epoch": 13.87589928057554, "step": 15430 }, { "epoch": 13.87589928057554, "step": 15430, "torque_loss": 0.20309264957904816 }, { "epoch": 13.884892086330936, "grad_norm": 0.2010437250137329, "learning_rate": 8.278187086722377e-05, "loss": 0.0265, "step": 15440 }, { "action_loss": 0.00827688816934824, "epoch": 13.884892086330936, "step": 15440 }, { "epoch": 13.884892086330936, "step": 15440, "torque_loss": 0.178869366645813 }, { "epoch": 13.89388489208633, "grad_norm": 0.1754104048013687, "learning_rate": 8.275689379570544e-05, "loss": 0.0294, "step": 15450 }, { "action_loss": 0.012202979065477848, "epoch": 13.89388489208633, "step": 15450 }, { "epoch": 13.89388489208633, "step": 15450, "torque_loss": 0.2069903165102005 }, { "epoch": 13.902877697841726, "grad_norm": 0.21486322581768036, "learning_rate": 8.273190239520887e-05, "loss": 0.0329, "step": 15460 }, { "action_loss": 0.006497527007013559, "epoch": 13.902877697841726, "step": 15460 }, { "epoch": 13.902877697841726, "step": 15460, "torque_loss": 0.12501882016658783 }, { "epoch": 13.911870503597122, "grad_norm": 0.215187668800354, "learning_rate": 8.270689667666616e-05, "loss": 0.0257, "step": 15470 }, { "action_loss": 0.02547474391758442, "epoch": 13.911870503597122, "step": 15470 }, { "epoch": 13.911870503597122, "step": 15470, "torque_loss": 0.2840932607650757 }, { "epoch": 13.920863309352518, "grad_norm": 0.21794728934764862, "learning_rate": 8.268187665101563e-05, "loss": 0.0301, "step": 15480 }, { "action_loss": 0.014895924367010593, "epoch": 13.920863309352518, "step": 15480 }, { "epoch": 13.920863309352518, "step": 15480, "torque_loss": 0.19484849274158478 }, { "epoch": 13.929856115107913, "grad_norm": 0.22063829004764557, "learning_rate": 8.265684232920194e-05, "loss": 0.0273, "step": 15490 }, { "action_loss": 0.013136650435626507, "epoch": 13.929856115107913, "step": 15490 }, { "epoch": 13.929856115107913, "step": 15490, "torque_loss": 0.21910949051380157 }, { "epoch": 13.93884892086331, "grad_norm": 0.16773544251918793, "learning_rate": 8.263179372217592e-05, "loss": 0.0293, "step": 15500 }, { "action_loss": 0.006448735948652029, "epoch": 13.93884892086331, "step": 15500 }, { "epoch": 13.93884892086331, "step": 15500, "torque_loss": 0.18178415298461914 }, { "epoch": 13.947841726618705, "grad_norm": 0.1852494180202484, "learning_rate": 8.260673084089468e-05, "loss": 0.023, "step": 15510 }, { "action_loss": 0.007075134664773941, "epoch": 13.947841726618705, "step": 15510 }, { "epoch": 13.947841726618705, "step": 15510, "torque_loss": 0.19029419124126434 }, { "epoch": 13.956834532374101, "grad_norm": 0.1810329109430313, "learning_rate": 8.25816536963216e-05, "loss": 0.0278, "step": 15520 }, { "action_loss": 0.004522424656897783, "epoch": 13.956834532374101, "step": 15520 }, { "epoch": 13.956834532374101, "step": 15520, "torque_loss": 0.16532452404499054 }, { "epoch": 13.965827338129497, "grad_norm": 0.11406462639570236, "learning_rate": 8.255656229942626e-05, "loss": 0.0211, "step": 15530 }, { "action_loss": 0.010583451949059963, "epoch": 13.965827338129497, "step": 15530 }, { "epoch": 13.965827338129497, "step": 15530, "torque_loss": 0.2706197500228882 }, { "epoch": 13.974820143884893, "grad_norm": 0.18943293392658234, "learning_rate": 8.253145666118447e-05, "loss": 0.0258, "step": 15540 }, { "action_loss": 0.012651890516281128, "epoch": 13.974820143884893, "step": 15540 }, { "epoch": 13.974820143884893, "step": 15540, "torque_loss": 0.14964303374290466 }, { "epoch": 13.983812949640289, "grad_norm": 0.1733827143907547, "learning_rate": 8.250633679257834e-05, "loss": 0.0275, "step": 15550 }, { "action_loss": 0.007921230047941208, "epoch": 13.983812949640289, "step": 15550 }, { "epoch": 13.983812949640289, "step": 15550, "torque_loss": 0.15648913383483887 }, { "epoch": 13.992805755395683, "grad_norm": 0.2748901844024658, "learning_rate": 8.24812027045961e-05, "loss": 0.0268, "step": 15560 }, { "action_loss": 0.006391818169504404, "epoch": 13.992805755395683, "step": 15560 }, { "epoch": 13.992805755395683, "step": 15560, "torque_loss": 0.16883768141269684 }, { "epoch": 14.001798561151078, "grad_norm": 0.20515406131744385, "learning_rate": 8.245605440823226e-05, "loss": 0.0251, "step": 15570 }, { "action_loss": 0.004098861012607813, "epoch": 14.001798561151078, "step": 15570 }, { "epoch": 14.001798561151078, "step": 15570, "torque_loss": 0.15846402943134308 }, { "epoch": 14.010791366906474, "grad_norm": 0.19241735339164734, "learning_rate": 8.243089191448759e-05, "loss": 0.023, "step": 15580 }, { "action_loss": 0.00879135075956583, "epoch": 14.010791366906474, "step": 15580 }, { "epoch": 14.010791366906474, "step": 15580, "torque_loss": 0.1252160221338272 }, { "epoch": 14.01978417266187, "grad_norm": 0.17180278897285461, "learning_rate": 8.240571523436896e-05, "loss": 0.0333, "step": 15590 }, { "action_loss": 0.005445382092148066, "epoch": 14.01978417266187, "step": 15590 }, { "epoch": 14.01978417266187, "step": 15590, "torque_loss": 0.14672134816646576 }, { "epoch": 14.028776978417266, "grad_norm": 0.15661008656024933, "learning_rate": 8.238052437888955e-05, "loss": 0.02, "step": 15600 }, { "action_loss": 0.007185249123722315, "epoch": 14.028776978417266, "step": 15600 }, { "epoch": 14.028776978417266, "step": 15600, "torque_loss": 0.2166367620229721 }, { "epoch": 14.037769784172662, "grad_norm": 0.151626855134964, "learning_rate": 8.235531935906866e-05, "loss": 0.0258, "step": 15610 }, { "action_loss": 0.0052911159582436085, "epoch": 14.037769784172662, "step": 15610 }, { "epoch": 14.037769784172662, "step": 15610, "torque_loss": 0.19016532599925995 }, { "epoch": 14.046762589928058, "grad_norm": 0.198669895529747, "learning_rate": 8.233010018593185e-05, "loss": 0.0298, "step": 15620 }, { "action_loss": 0.004222550895065069, "epoch": 14.046762589928058, "step": 15620 }, { "epoch": 14.046762589928058, "step": 15620, "torque_loss": 0.195447638630867 }, { "epoch": 14.055755395683454, "grad_norm": 0.2624839246273041, "learning_rate": 8.230486687051082e-05, "loss": 0.0364, "step": 15630 }, { "action_loss": 0.0021780033130198717, "epoch": 14.055755395683454, "step": 15630 }, { "epoch": 14.055755395683454, "step": 15630, "torque_loss": 0.0823260024189949 }, { "epoch": 14.06474820143885, "grad_norm": 0.21699167788028717, "learning_rate": 8.22796194238435e-05, "loss": 0.0277, "step": 15640 }, { "action_loss": 0.00867017824202776, "epoch": 14.06474820143885, "step": 15640 }, { "epoch": 14.06474820143885, "step": 15640, "torque_loss": 0.19374680519104004 }, { "epoch": 14.073741007194245, "grad_norm": 0.2084931880235672, "learning_rate": 8.225435785697397e-05, "loss": 0.0345, "step": 15650 }, { "action_loss": 0.010082476772367954, "epoch": 14.073741007194245, "step": 15650 }, { "epoch": 14.073741007194245, "step": 15650, "torque_loss": 0.1525116115808487 }, { "epoch": 14.082733812949641, "grad_norm": 0.20280709862709045, "learning_rate": 8.222908218095252e-05, "loss": 0.0351, "step": 15660 }, { "action_loss": 0.0037487123627215624, "epoch": 14.082733812949641, "step": 15660 }, { "epoch": 14.082733812949641, "step": 15660, "torque_loss": 0.14748883247375488 }, { "epoch": 14.091726618705035, "grad_norm": 0.23783917725086212, "learning_rate": 8.220379240683553e-05, "loss": 0.0259, "step": 15670 }, { "action_loss": 0.006717061158269644, "epoch": 14.091726618705035, "step": 15670 }, { "epoch": 14.091726618705035, "step": 15670, "torque_loss": 0.17947852611541748 }, { "epoch": 14.100719424460431, "grad_norm": 0.14756682515144348, "learning_rate": 8.217848854568566e-05, "loss": 0.0305, "step": 15680 }, { "action_loss": 0.005621691700071096, "epoch": 14.100719424460431, "step": 15680 }, { "epoch": 14.100719424460431, "step": 15680, "torque_loss": 0.1534814089536667 }, { "epoch": 14.109712230215827, "grad_norm": 0.1166827604174614, "learning_rate": 8.215317060857167e-05, "loss": 0.0271, "step": 15690 }, { "action_loss": 0.0038011956494301558, "epoch": 14.109712230215827, "step": 15690 }, { "epoch": 14.109712230215827, "step": 15690, "torque_loss": 0.230473592877388 }, { "epoch": 14.118705035971223, "grad_norm": 0.17950080335140228, "learning_rate": 8.212783860656847e-05, "loss": 0.0244, "step": 15700 }, { "action_loss": 0.01665867678821087, "epoch": 14.118705035971223, "step": 15700 }, { "epoch": 14.118705035971223, "step": 15700, "torque_loss": 0.20346109569072723 }, { "epoch": 14.127697841726619, "grad_norm": 0.20927277207374573, "learning_rate": 8.210249255075717e-05, "loss": 0.0241, "step": 15710 }, { "action_loss": 0.013618972152471542, "epoch": 14.127697841726619, "step": 15710 }, { "epoch": 14.127697841726619, "step": 15710, "torque_loss": 0.19861452281475067 }, { "epoch": 14.136690647482014, "grad_norm": 0.2539014518260956, "learning_rate": 8.207713245222496e-05, "loss": 0.0278, "step": 15720 }, { "action_loss": 0.004692788701504469, "epoch": 14.136690647482014, "step": 15720 }, { "epoch": 14.136690647482014, "step": 15720, "torque_loss": 0.1921657770872116 }, { "epoch": 14.14568345323741, "grad_norm": 0.21955765783786774, "learning_rate": 8.205175832206522e-05, "loss": 0.0277, "step": 15730 }, { "action_loss": 0.0043145823292434216, "epoch": 14.14568345323741, "step": 15730 }, { "epoch": 14.14568345323741, "step": 15730, "torque_loss": 0.16494427621364594 }, { "epoch": 14.154676258992806, "grad_norm": 0.20109796524047852, "learning_rate": 8.202637017137747e-05, "loss": 0.0258, "step": 15740 }, { "action_loss": 0.009229659102857113, "epoch": 14.154676258992806, "step": 15740 }, { "epoch": 14.154676258992806, "step": 15740, "torque_loss": 0.13556157052516937 }, { "epoch": 14.163669064748202, "grad_norm": 0.15270638465881348, "learning_rate": 8.200096801126734e-05, "loss": 0.0311, "step": 15750 }, { "action_loss": 0.004583155270665884, "epoch": 14.163669064748202, "step": 15750 }, { "epoch": 14.163669064748202, "step": 15750, "torque_loss": 0.17109711468219757 }, { "epoch": 14.172661870503598, "grad_norm": 0.1784757673740387, "learning_rate": 8.197555185284658e-05, "loss": 0.0282, "step": 15760 }, { "action_loss": 0.01332582626491785, "epoch": 14.172661870503598, "step": 15760 }, { "epoch": 14.172661870503598, "step": 15760, "torque_loss": 0.16426993906497955 }, { "epoch": 14.181654676258994, "grad_norm": 0.19809339940547943, "learning_rate": 8.19501217072331e-05, "loss": 0.03, "step": 15770 }, { "action_loss": 0.009185263887047768, "epoch": 14.181654676258994, "step": 15770 }, { "epoch": 14.181654676258994, "step": 15770, "torque_loss": 0.3045293092727661 }, { "epoch": 14.190647482014388, "grad_norm": 0.21773044764995575, "learning_rate": 8.192467758555091e-05, "loss": 0.0294, "step": 15780 }, { "action_loss": 0.011332002468407154, "epoch": 14.190647482014388, "step": 15780 }, { "epoch": 14.190647482014388, "step": 15780, "torque_loss": 0.12916521728038788 }, { "epoch": 14.199640287769784, "grad_norm": 0.17262805998325348, "learning_rate": 8.189921949893014e-05, "loss": 0.028, "step": 15790 }, { "action_loss": 0.019819587469100952, "epoch": 14.199640287769784, "step": 15790 }, { "epoch": 14.199640287769784, "step": 15790, "torque_loss": 0.2079230695962906 }, { "epoch": 14.20863309352518, "grad_norm": 0.15081563591957092, "learning_rate": 8.187374745850702e-05, "loss": 0.0244, "step": 15800 }, { "action_loss": 0.007641105446964502, "epoch": 14.20863309352518, "step": 15800 }, { "epoch": 14.20863309352518, "step": 15800, "torque_loss": 0.15649999678134918 }, { "epoch": 14.217625899280575, "grad_norm": 0.20587511360645294, "learning_rate": 8.184826147542387e-05, "loss": 0.032, "step": 15810 }, { "action_loss": 0.007752681151032448, "epoch": 14.217625899280575, "step": 15810 }, { "epoch": 14.217625899280575, "step": 15810, "torque_loss": 0.19670434296131134 }, { "epoch": 14.226618705035971, "grad_norm": 0.2682218551635742, "learning_rate": 8.18227615608291e-05, "loss": 0.0297, "step": 15820 }, { "action_loss": 0.00643535889685154, "epoch": 14.226618705035971, "step": 15820 }, { "epoch": 14.226618705035971, "step": 15820, "torque_loss": 0.26389166712760925 }, { "epoch": 14.235611510791367, "grad_norm": 0.17242789268493652, "learning_rate": 8.179724772587732e-05, "loss": 0.0268, "step": 15830 }, { "action_loss": 0.00896189734339714, "epoch": 14.235611510791367, "step": 15830 }, { "epoch": 14.235611510791367, "step": 15830, "torque_loss": 0.18638072907924652 }, { "epoch": 14.244604316546763, "grad_norm": 0.17232726514339447, "learning_rate": 8.177171998172908e-05, "loss": 0.0256, "step": 15840 }, { "action_loss": 0.005999261047691107, "epoch": 14.244604316546763, "step": 15840 }, { "epoch": 14.244604316546763, "step": 15840, "torque_loss": 0.22385184466838837 }, { "epoch": 14.253597122302159, "grad_norm": 0.15359458327293396, "learning_rate": 8.174617833955112e-05, "loss": 0.0228, "step": 15850 }, { "action_loss": 0.02795347012579441, "epoch": 14.253597122302159, "step": 15850 }, { "epoch": 14.253597122302159, "step": 15850, "torque_loss": 0.2790805995464325 }, { "epoch": 14.262589928057555, "grad_norm": 0.21190424263477325, "learning_rate": 8.172062281051618e-05, "loss": 0.0339, "step": 15860 }, { "action_loss": 0.005878280848264694, "epoch": 14.262589928057555, "step": 15860 }, { "epoch": 14.262589928057555, "step": 15860, "torque_loss": 0.15571118891239166 }, { "epoch": 14.27158273381295, "grad_norm": 0.1717182993888855, "learning_rate": 8.169505340580316e-05, "loss": 0.0283, "step": 15870 }, { "action_loss": 0.009972182102501392, "epoch": 14.27158273381295, "step": 15870 }, { "epoch": 14.27158273381295, "step": 15870, "torque_loss": 0.1707373857498169 }, { "epoch": 14.280575539568344, "grad_norm": 0.1659136861562729, "learning_rate": 8.166947013659696e-05, "loss": 0.0286, "step": 15880 }, { "action_loss": 0.034774910658597946, "epoch": 14.280575539568344, "step": 15880 }, { "epoch": 14.280575539568344, "step": 15880, "torque_loss": 0.24643348157405853 }, { "epoch": 14.28956834532374, "grad_norm": 0.23443163931369781, "learning_rate": 8.164387301408858e-05, "loss": 0.0343, "step": 15890 }, { "action_loss": 0.017139948904514313, "epoch": 14.28956834532374, "step": 15890 }, { "epoch": 14.28956834532374, "step": 15890, "torque_loss": 0.21677441895008087 }, { "epoch": 14.298561151079136, "grad_norm": 0.1326141506433487, "learning_rate": 8.161826204947507e-05, "loss": 0.0334, "step": 15900 }, { "action_loss": 0.034860819578170776, "epoch": 14.298561151079136, "step": 15900 }, { "epoch": 14.298561151079136, "step": 15900, "torque_loss": 0.2449798583984375 }, { "epoch": 14.307553956834532, "grad_norm": 0.18587900698184967, "learning_rate": 8.159263725395954e-05, "loss": 0.033, "step": 15910 }, { "action_loss": 0.005927134305238724, "epoch": 14.307553956834532, "step": 15910 }, { "epoch": 14.307553956834532, "step": 15910, "torque_loss": 0.1728048324584961 }, { "epoch": 14.316546762589928, "grad_norm": 0.19386008381843567, "learning_rate": 8.156699863875115e-05, "loss": 0.0377, "step": 15920 }, { "action_loss": 0.008695930242538452, "epoch": 14.316546762589928, "step": 15920 }, { "epoch": 14.316546762589928, "step": 15920, "torque_loss": 0.19905667006969452 }, { "epoch": 14.325539568345324, "grad_norm": 0.13857348263263702, "learning_rate": 8.154134621506508e-05, "loss": 0.0281, "step": 15930 }, { "action_loss": 0.01036036852747202, "epoch": 14.325539568345324, "step": 15930 }, { "epoch": 14.325539568345324, "step": 15930, "torque_loss": 0.1331731379032135 }, { "epoch": 14.33453237410072, "grad_norm": 0.2607772946357727, "learning_rate": 8.15156799941226e-05, "loss": 0.0292, "step": 15940 }, { "action_loss": 0.007870401255786419, "epoch": 14.33453237410072, "step": 15940 }, { "epoch": 14.33453237410072, "step": 15940, "torque_loss": 0.20437605679035187 }, { "epoch": 14.343525179856115, "grad_norm": 0.22847919166088104, "learning_rate": 8.148999998715096e-05, "loss": 0.0322, "step": 15950 }, { "action_loss": 0.005662532057613134, "epoch": 14.343525179856115, "step": 15950 }, { "epoch": 14.343525179856115, "step": 15950, "torque_loss": 0.20124448835849762 }, { "epoch": 14.352517985611511, "grad_norm": 0.19503864645957947, "learning_rate": 8.146430620538348e-05, "loss": 0.0363, "step": 15960 }, { "action_loss": 0.005449378862977028, "epoch": 14.352517985611511, "step": 15960 }, { "epoch": 14.352517985611511, "step": 15960, "torque_loss": 0.19747893512248993 }, { "epoch": 14.361510791366907, "grad_norm": 0.17886225879192352, "learning_rate": 8.143859866005949e-05, "loss": 0.0251, "step": 15970 }, { "action_loss": 0.00957407895475626, "epoch": 14.361510791366907, "step": 15970 }, { "epoch": 14.361510791366907, "step": 15970, "torque_loss": 0.18741042912006378 }, { "epoch": 14.370503597122303, "grad_norm": 0.15553033351898193, "learning_rate": 8.141287736242438e-05, "loss": 0.0308, "step": 15980 }, { "action_loss": 0.01142653077840805, "epoch": 14.370503597122303, "step": 15980 }, { "epoch": 14.370503597122303, "step": 15980, "torque_loss": 0.1981809288263321 }, { "epoch": 14.379496402877697, "grad_norm": 0.2796368896961212, "learning_rate": 8.138714232372945e-05, "loss": 0.0252, "step": 15990 }, { "action_loss": 0.014036453329026699, "epoch": 14.379496402877697, "step": 15990 }, { "epoch": 14.379496402877697, "step": 15990, "torque_loss": 0.17951953411102295 }, { "epoch": 14.388489208633093, "grad_norm": 0.2602987587451935, "learning_rate": 8.136139355523213e-05, "loss": 0.0338, "step": 16000 }, { "action_loss": 0.01521371304988861, "epoch": 14.388489208633093, "step": 16000 }, { "epoch": 14.388489208633093, "step": 16000, "torque_loss": 0.2047508955001831 }, { "epoch": 14.397482014388489, "grad_norm": 0.17693477869033813, "learning_rate": 8.13356310681958e-05, "loss": 0.0302, "step": 16010 }, { "action_loss": 0.014133471064269543, "epoch": 14.397482014388489, "step": 16010 }, { "epoch": 14.397482014388489, "step": 16010, "torque_loss": 0.1503065824508667 }, { "epoch": 14.406474820143885, "grad_norm": 0.21315455436706543, "learning_rate": 8.130985487388982e-05, "loss": 0.0309, "step": 16020 }, { "action_loss": 0.019918128848075867, "epoch": 14.406474820143885, "step": 16020 }, { "epoch": 14.406474820143885, "step": 16020, "torque_loss": 0.22401177883148193 }, { "epoch": 14.41546762589928, "grad_norm": 0.18081782758235931, "learning_rate": 8.12840649835896e-05, "loss": 0.0329, "step": 16030 }, { "action_loss": 0.004505643155425787, "epoch": 14.41546762589928, "step": 16030 }, { "epoch": 14.41546762589928, "step": 16030, "torque_loss": 0.13162828981876373 }, { "epoch": 14.424460431654676, "grad_norm": 0.1981586366891861, "learning_rate": 8.125826140857653e-05, "loss": 0.0261, "step": 16040 }, { "action_loss": 0.026930788531899452, "epoch": 14.424460431654676, "step": 16040 }, { "epoch": 14.424460431654676, "step": 16040, "torque_loss": 0.22157998383045197 }, { "epoch": 14.433453237410072, "grad_norm": 0.2590559124946594, "learning_rate": 8.123244416013792e-05, "loss": 0.0301, "step": 16050 }, { "action_loss": 0.00943481083959341, "epoch": 14.433453237410072, "step": 16050 }, { "epoch": 14.433453237410072, "step": 16050, "torque_loss": 0.21496085822582245 }, { "epoch": 14.442446043165468, "grad_norm": 0.14687497913837433, "learning_rate": 8.120661324956716e-05, "loss": 0.0268, "step": 16060 }, { "action_loss": 0.004284620750695467, "epoch": 14.442446043165468, "step": 16060 }, { "epoch": 14.442446043165468, "step": 16060, "torque_loss": 0.141185462474823 }, { "epoch": 14.451438848920864, "grad_norm": 0.24839434027671814, "learning_rate": 8.118076868816353e-05, "loss": 0.0311, "step": 16070 }, { "action_loss": 0.008901023305952549, "epoch": 14.451438848920864, "step": 16070 }, { "epoch": 14.451438848920864, "step": 16070, "torque_loss": 0.16202501952648163 }, { "epoch": 14.46043165467626, "grad_norm": 0.26850685477256775, "learning_rate": 8.115491048723234e-05, "loss": 0.0257, "step": 16080 }, { "action_loss": 0.005215059500187635, "epoch": 14.46043165467626, "step": 16080 }, { "epoch": 14.46043165467626, "step": 16080, "torque_loss": 0.1476864218711853 }, { "epoch": 14.469424460431656, "grad_norm": 0.19155076146125793, "learning_rate": 8.112903865808486e-05, "loss": 0.0261, "step": 16090 }, { "action_loss": 0.008088864386081696, "epoch": 14.469424460431656, "step": 16090 }, { "epoch": 14.469424460431656, "step": 16090, "torque_loss": 0.16434603929519653 }, { "epoch": 14.47841726618705, "grad_norm": 0.17354179918766022, "learning_rate": 8.110315321203826e-05, "loss": 0.0265, "step": 16100 }, { "action_loss": 0.011113796383142471, "epoch": 14.47841726618705, "step": 16100 }, { "epoch": 14.47841726618705, "step": 16100, "torque_loss": 0.2502990663051605 }, { "epoch": 14.487410071942445, "grad_norm": 0.19399385154247284, "learning_rate": 8.107725416041575e-05, "loss": 0.0301, "step": 16110 }, { "action_loss": 0.004911335650831461, "epoch": 14.487410071942445, "step": 16110 }, { "epoch": 14.487410071942445, "step": 16110, "torque_loss": 0.1552998274564743 }, { "epoch": 14.496402877697841, "grad_norm": 0.22929739952087402, "learning_rate": 8.105134151454645e-05, "loss": 0.0302, "step": 16120 }, { "action_loss": 0.007425554562360048, "epoch": 14.496402877697841, "step": 16120 }, { "epoch": 14.496402877697841, "step": 16120, "torque_loss": 0.1637578159570694 }, { "epoch": 14.505395683453237, "grad_norm": 0.25756463408470154, "learning_rate": 8.102541528576544e-05, "loss": 0.03, "step": 16130 }, { "action_loss": 0.0028275782242417336, "epoch": 14.505395683453237, "step": 16130 }, { "epoch": 14.505395683453237, "step": 16130, "torque_loss": 0.09122394770383835 }, { "epoch": 14.514388489208633, "grad_norm": 0.17728452384471893, "learning_rate": 8.099947548541371e-05, "loss": 0.0224, "step": 16140 }, { "action_loss": 0.006580611225217581, "epoch": 14.514388489208633, "step": 16140 }, { "epoch": 14.514388489208633, "step": 16140, "torque_loss": 0.16860537230968475 }, { "epoch": 14.523381294964029, "grad_norm": 0.21240060031414032, "learning_rate": 8.097352212483824e-05, "loss": 0.0288, "step": 16150 }, { "action_loss": 0.006486766505986452, "epoch": 14.523381294964029, "step": 16150 }, { "epoch": 14.523381294964029, "step": 16150, "torque_loss": 0.24762468039989471 }, { "epoch": 14.532374100719425, "grad_norm": 0.1926598995923996, "learning_rate": 8.094755521539187e-05, "loss": 0.0273, "step": 16160 }, { "action_loss": 0.007804935332387686, "epoch": 14.532374100719425, "step": 16160 }, { "epoch": 14.532374100719425, "step": 16160, "torque_loss": 0.1845872849225998 }, { "epoch": 14.54136690647482, "grad_norm": 0.15254667401313782, "learning_rate": 8.092157476843346e-05, "loss": 0.0277, "step": 16170 }, { "action_loss": 0.006795607972890139, "epoch": 14.54136690647482, "step": 16170 }, { "epoch": 14.54136690647482, "step": 16170, "torque_loss": 0.12098955363035202 }, { "epoch": 14.550359712230216, "grad_norm": 0.16911807656288147, "learning_rate": 8.089558079532769e-05, "loss": 0.0249, "step": 16180 }, { "action_loss": 0.01125127449631691, "epoch": 14.550359712230216, "step": 16180 }, { "epoch": 14.550359712230216, "step": 16180, "torque_loss": 0.13711516559123993 }, { "epoch": 14.559352517985612, "grad_norm": 0.1294543296098709, "learning_rate": 8.086957330744522e-05, "loss": 0.0273, "step": 16190 }, { "action_loss": 0.003908399026840925, "epoch": 14.559352517985612, "step": 16190 }, { "epoch": 14.559352517985612, "step": 16190, "torque_loss": 0.16650836169719696 }, { "epoch": 14.568345323741006, "grad_norm": 0.3415554165840149, "learning_rate": 8.084355231616265e-05, "loss": 0.0217, "step": 16200 }, { "action_loss": 0.005322450306266546, "epoch": 14.568345323741006, "step": 16200 }, { "epoch": 14.568345323741006, "step": 16200, "torque_loss": 0.15937013924121857 }, { "epoch": 14.577338129496402, "grad_norm": 0.16280849277973175, "learning_rate": 8.08175178328624e-05, "loss": 0.0277, "step": 16210 }, { "action_loss": 0.0079047167673707, "epoch": 14.577338129496402, "step": 16210 }, { "epoch": 14.577338129496402, "step": 16210, "torque_loss": 0.20597775280475616 }, { "epoch": 14.586330935251798, "grad_norm": 0.2107785940170288, "learning_rate": 8.079146986893284e-05, "loss": 0.0315, "step": 16220 }, { "action_loss": 0.011037341319024563, "epoch": 14.586330935251798, "step": 16220 }, { "epoch": 14.586330935251798, "step": 16220, "torque_loss": 0.1534397453069687 }, { "epoch": 14.595323741007194, "grad_norm": 0.21100926399230957, "learning_rate": 8.076540843576826e-05, "loss": 0.0271, "step": 16230 }, { "action_loss": 0.005840206518769264, "epoch": 14.595323741007194, "step": 16230 }, { "epoch": 14.595323741007194, "step": 16230, "torque_loss": 0.10322091728448868 }, { "epoch": 14.60431654676259, "grad_norm": 0.1950031816959381, "learning_rate": 8.073933354476879e-05, "loss": 0.0311, "step": 16240 }, { "action_loss": 0.006514185573905706, "epoch": 14.60431654676259, "step": 16240 }, { "epoch": 14.60431654676259, "step": 16240, "torque_loss": 0.14690665900707245 }, { "epoch": 14.613309352517986, "grad_norm": 0.3067255914211273, "learning_rate": 8.07132452073405e-05, "loss": 0.0297, "step": 16250 }, { "action_loss": 0.009699739515781403, "epoch": 14.613309352517986, "step": 16250 }, { "epoch": 14.613309352517986, "step": 16250, "torque_loss": 0.19101740419864655 }, { "epoch": 14.622302158273381, "grad_norm": 0.240230992436409, "learning_rate": 8.068714343489529e-05, "loss": 0.0246, "step": 16260 }, { "action_loss": 0.005423149559646845, "epoch": 14.622302158273381, "step": 16260 }, { "epoch": 14.622302158273381, "step": 16260, "torque_loss": 0.1446235030889511 }, { "epoch": 14.631294964028777, "grad_norm": 0.19996391236782074, "learning_rate": 8.066102823885097e-05, "loss": 0.0263, "step": 16270 }, { "action_loss": 0.015090077184140682, "epoch": 14.631294964028777, "step": 16270 }, { "epoch": 14.631294964028777, "step": 16270, "torque_loss": 0.2547461986541748 }, { "epoch": 14.640287769784173, "grad_norm": 0.17355835437774658, "learning_rate": 8.063489963063125e-05, "loss": 0.0344, "step": 16280 }, { "action_loss": 0.01415166910737753, "epoch": 14.640287769784173, "step": 16280 }, { "epoch": 14.640287769784173, "step": 16280, "torque_loss": 0.197776198387146 }, { "epoch": 14.649280575539569, "grad_norm": 0.20094314217567444, "learning_rate": 8.06087576216656e-05, "loss": 0.0283, "step": 16290 }, { "action_loss": 0.011817525140941143, "epoch": 14.649280575539569, "step": 16290 }, { "epoch": 14.649280575539569, "step": 16290, "torque_loss": 0.19134621322155 }, { "epoch": 14.658273381294965, "grad_norm": 0.19729380309581757, "learning_rate": 8.058260222338946e-05, "loss": 0.0299, "step": 16300 }, { "action_loss": 0.018209129571914673, "epoch": 14.658273381294965, "step": 16300 }, { "epoch": 14.658273381294965, "step": 16300, "torque_loss": 0.2561369836330414 }, { "epoch": 14.667266187050359, "grad_norm": 0.1501963883638382, "learning_rate": 8.05564334472441e-05, "loss": 0.0299, "step": 16310 }, { "action_loss": 0.009623055346310139, "epoch": 14.667266187050359, "step": 16310 }, { "epoch": 14.667266187050359, "step": 16310, "torque_loss": 0.18497247993946075 }, { "epoch": 14.676258992805755, "grad_norm": 0.13716070353984833, "learning_rate": 8.053025130467664e-05, "loss": 0.0256, "step": 16320 }, { "action_loss": 0.008113403804600239, "epoch": 14.676258992805755, "step": 16320 }, { "epoch": 14.676258992805755, "step": 16320, "torque_loss": 0.2286670207977295 }, { "epoch": 14.68525179856115, "grad_norm": 0.15270496904850006, "learning_rate": 8.050405580714e-05, "loss": 0.0268, "step": 16330 }, { "action_loss": 0.019032880663871765, "epoch": 14.68525179856115, "step": 16330 }, { "epoch": 14.68525179856115, "step": 16330, "torque_loss": 0.2635438144207001 }, { "epoch": 14.694244604316546, "grad_norm": 0.16145238280296326, "learning_rate": 8.047784696609301e-05, "loss": 0.0273, "step": 16340 }, { "action_loss": 0.01583469845354557, "epoch": 14.694244604316546, "step": 16340 }, { "epoch": 14.694244604316546, "step": 16340, "torque_loss": 0.22028841078281403 }, { "epoch": 14.703237410071942, "grad_norm": 0.20646846294403076, "learning_rate": 8.045162479300029e-05, "loss": 0.0317, "step": 16350 }, { "action_loss": 0.011974782682955265, "epoch": 14.703237410071942, "step": 16350 }, { "epoch": 14.703237410071942, "step": 16350, "torque_loss": 0.26716718077659607 }, { "epoch": 14.712230215827338, "grad_norm": 0.18182162940502167, "learning_rate": 8.042538929933232e-05, "loss": 0.0295, "step": 16360 }, { "action_loss": 0.012068215757608414, "epoch": 14.712230215827338, "step": 16360 }, { "epoch": 14.712230215827338, "step": 16360, "torque_loss": 0.24682007730007172 }, { "epoch": 14.721223021582734, "grad_norm": 0.16946950554847717, "learning_rate": 8.039914049656539e-05, "loss": 0.0305, "step": 16370 }, { "action_loss": 0.01123885065317154, "epoch": 14.721223021582734, "step": 16370 }, { "epoch": 14.721223021582734, "step": 16370, "torque_loss": 0.15651269257068634 }, { "epoch": 14.73021582733813, "grad_norm": 0.17630265653133392, "learning_rate": 8.037287839618163e-05, "loss": 0.0291, "step": 16380 }, { "action_loss": 0.010587519966065884, "epoch": 14.73021582733813, "step": 16380 }, { "epoch": 14.73021582733813, "step": 16380, "torque_loss": 0.18120145797729492 }, { "epoch": 14.739208633093526, "grad_norm": 0.17956063151359558, "learning_rate": 8.034660300966898e-05, "loss": 0.0314, "step": 16390 }, { "action_loss": 0.003402011701837182, "epoch": 14.739208633093526, "step": 16390 }, { "epoch": 14.739208633093526, "step": 16390, "torque_loss": 0.15529344975948334 }, { "epoch": 14.748201438848922, "grad_norm": 0.12901650369167328, "learning_rate": 8.032031434852115e-05, "loss": 0.0246, "step": 16400 }, { "action_loss": 0.010789514519274235, "epoch": 14.748201438848922, "step": 16400 }, { "epoch": 14.748201438848922, "step": 16400, "torque_loss": 0.23860780894756317 }, { "epoch": 14.757194244604317, "grad_norm": 0.2518579959869385, "learning_rate": 8.029401242423773e-05, "loss": 0.0267, "step": 16410 }, { "action_loss": 0.00801508966833353, "epoch": 14.757194244604317, "step": 16410 }, { "epoch": 14.757194244604317, "step": 16410, "torque_loss": 0.16531135141849518 }, { "epoch": 14.766187050359711, "grad_norm": 0.16671262681484222, "learning_rate": 8.026769724832406e-05, "loss": 0.0287, "step": 16420 }, { "action_loss": 0.016939541324973106, "epoch": 14.766187050359711, "step": 16420 }, { "epoch": 14.766187050359711, "step": 16420, "torque_loss": 0.2505215108394623 }, { "epoch": 14.775179856115107, "grad_norm": 0.2310776710510254, "learning_rate": 8.02413688322913e-05, "loss": 0.0304, "step": 16430 }, { "action_loss": 0.006482100114226341, "epoch": 14.775179856115107, "step": 16430 }, { "epoch": 14.775179856115107, "step": 16430, "torque_loss": 0.12301721423864365 }, { "epoch": 14.784172661870503, "grad_norm": 0.18220920860767365, "learning_rate": 8.021502718765639e-05, "loss": 0.0263, "step": 16440 }, { "action_loss": 0.004125228151679039, "epoch": 14.784172661870503, "step": 16440 }, { "epoch": 14.784172661870503, "step": 16440, "torque_loss": 0.19747072458267212 }, { "epoch": 14.793165467625899, "grad_norm": 0.2004413604736328, "learning_rate": 8.018867232594204e-05, "loss": 0.027, "step": 16450 }, { "action_loss": 0.004599942825734615, "epoch": 14.793165467625899, "step": 16450 }, { "epoch": 14.793165467625899, "step": 16450, "torque_loss": 0.15352307260036469 }, { "epoch": 14.802158273381295, "grad_norm": 0.2808452248573303, "learning_rate": 8.016230425867679e-05, "loss": 0.0309, "step": 16460 }, { "action_loss": 0.006825400050729513, "epoch": 14.802158273381295, "step": 16460 }, { "epoch": 14.802158273381295, "step": 16460, "torque_loss": 0.17247605323791504 }, { "epoch": 14.81115107913669, "grad_norm": 0.23744294047355652, "learning_rate": 8.013592299739494e-05, "loss": 0.0257, "step": 16470 }, { "action_loss": 0.004501888062804937, "epoch": 14.81115107913669, "step": 16470 }, { "epoch": 14.81115107913669, "step": 16470, "torque_loss": 0.13375966250896454 }, { "epoch": 14.820143884892087, "grad_norm": 0.21841861307621002, "learning_rate": 8.01095285536365e-05, "loss": 0.0282, "step": 16480 }, { "action_loss": 0.007353923749178648, "epoch": 14.820143884892087, "step": 16480 }, { "epoch": 14.820143884892087, "step": 16480, "torque_loss": 0.16802380979061127 }, { "epoch": 14.829136690647482, "grad_norm": 0.16225075721740723, "learning_rate": 8.008312093894735e-05, "loss": 0.0238, "step": 16490 }, { "action_loss": 0.006540358066558838, "epoch": 14.829136690647482, "step": 16490 }, { "epoch": 14.829136690647482, "step": 16490, "torque_loss": 0.2050585001707077 }, { "epoch": 14.838129496402878, "grad_norm": 0.15284936130046844, "learning_rate": 8.005670016487903e-05, "loss": 0.0265, "step": 16500 }, { "action_loss": 0.009608536958694458, "epoch": 14.838129496402878, "step": 16500 }, { "epoch": 14.838129496402878, "step": 16500, "torque_loss": 0.19422276318073273 }, { "epoch": 14.847122302158274, "grad_norm": 0.11127635836601257, "learning_rate": 8.003026624298892e-05, "loss": 0.0244, "step": 16510 }, { "action_loss": 0.021026423200964928, "epoch": 14.847122302158274, "step": 16510 }, { "epoch": 14.847122302158274, "step": 16510, "torque_loss": 0.21476761996746063 }, { "epoch": 14.85611510791367, "grad_norm": 0.10332544147968292, "learning_rate": 8.00038191848401e-05, "loss": 0.0244, "step": 16520 }, { "action_loss": 0.01698770932853222, "epoch": 14.85611510791367, "step": 16520 }, { "epoch": 14.85611510791367, "step": 16520, "torque_loss": 0.25956299901008606 }, { "epoch": 14.865107913669064, "grad_norm": 0.1898094266653061, "learning_rate": 7.997735900200142e-05, "loss": 0.0337, "step": 16530 }, { "action_loss": 0.013318966142833233, "epoch": 14.865107913669064, "step": 16530 }, { "epoch": 14.865107913669064, "step": 16530, "torque_loss": 0.17763499915599823 }, { "epoch": 14.87410071942446, "grad_norm": 0.201179638504982, "learning_rate": 7.995088570604745e-05, "loss": 0.0344, "step": 16540 }, { "action_loss": 0.011952902190387249, "epoch": 14.87410071942446, "step": 16540 }, { "epoch": 14.87410071942446, "step": 16540, "torque_loss": 0.23567664623260498 }, { "epoch": 14.883093525179856, "grad_norm": 0.15088587999343872, "learning_rate": 7.992439930855852e-05, "loss": 0.0301, "step": 16550 }, { "action_loss": 0.008318829350173473, "epoch": 14.883093525179856, "step": 16550 }, { "epoch": 14.883093525179856, "step": 16550, "torque_loss": 0.27194246649742126 }, { "epoch": 14.892086330935252, "grad_norm": 0.18864353001117706, "learning_rate": 7.989789982112067e-05, "loss": 0.0272, "step": 16560 }, { "action_loss": 0.03834978863596916, "epoch": 14.892086330935252, "step": 16560 }, { "epoch": 14.892086330935252, "step": 16560, "torque_loss": 0.2727101147174835 }, { "epoch": 14.901079136690647, "grad_norm": 0.2678527534008026, "learning_rate": 7.987138725532572e-05, "loss": 0.0368, "step": 16570 }, { "action_loss": 0.011138432659208775, "epoch": 14.901079136690647, "step": 16570 }, { "epoch": 14.901079136690647, "step": 16570, "torque_loss": 0.1408112496137619 }, { "epoch": 14.910071942446043, "grad_norm": 0.17998702824115753, "learning_rate": 7.98448616227711e-05, "loss": 0.0261, "step": 16580 }, { "action_loss": 0.007497669663280249, "epoch": 14.910071942446043, "step": 16580 }, { "epoch": 14.910071942446043, "step": 16580, "torque_loss": 0.12993483245372772 }, { "epoch": 14.91906474820144, "grad_norm": 0.2515982389450073, "learning_rate": 7.981832293506004e-05, "loss": 0.0302, "step": 16590 }, { "action_loss": 0.020560413599014282, "epoch": 14.91906474820144, "step": 16590 }, { "epoch": 14.91906474820144, "step": 16590, "torque_loss": 0.19264046847820282 }, { "epoch": 14.928057553956835, "grad_norm": 0.15714536607265472, "learning_rate": 7.97917712038015e-05, "loss": 0.0246, "step": 16600 }, { "action_loss": 0.011079397052526474, "epoch": 14.928057553956835, "step": 16600 }, { "epoch": 14.928057553956835, "step": 16600, "torque_loss": 0.20469750463962555 }, { "epoch": 14.93705035971223, "grad_norm": 0.1592392772436142, "learning_rate": 7.976520644061007e-05, "loss": 0.0302, "step": 16610 }, { "action_loss": 0.01055964082479477, "epoch": 14.93705035971223, "step": 16610 }, { "epoch": 14.93705035971223, "step": 16610, "torque_loss": 0.2375333458185196 }, { "epoch": 14.946043165467627, "grad_norm": 0.15750092267990112, "learning_rate": 7.973862865710611e-05, "loss": 0.0339, "step": 16620 }, { "action_loss": 0.00518082408234477, "epoch": 14.946043165467627, "step": 16620 }, { "epoch": 14.946043165467627, "step": 16620, "torque_loss": 0.16011890769004822 }, { "epoch": 14.95503597122302, "grad_norm": 0.14998318254947662, "learning_rate": 7.971203786491563e-05, "loss": 0.0293, "step": 16630 }, { "action_loss": 0.013434316031634808, "epoch": 14.95503597122302, "step": 16630 }, { "epoch": 14.95503597122302, "step": 16630, "torque_loss": 0.20295433700084686 }, { "epoch": 14.964028776978417, "grad_norm": 0.18460765480995178, "learning_rate": 7.968543407567034e-05, "loss": 0.0296, "step": 16640 }, { "action_loss": 0.004712393973022699, "epoch": 14.964028776978417, "step": 16640 }, { "epoch": 14.964028776978417, "step": 16640, "torque_loss": 0.12423651665449142 }, { "epoch": 14.973021582733812, "grad_norm": 0.30065611004829407, "learning_rate": 7.965881730100765e-05, "loss": 0.0265, "step": 16650 }, { "action_loss": 0.01350222248584032, "epoch": 14.973021582733812, "step": 16650 }, { "epoch": 14.973021582733812, "step": 16650, "torque_loss": 0.17097030580043793 }, { "epoch": 14.982014388489208, "grad_norm": 0.1866050511598587, "learning_rate": 7.963218755257064e-05, "loss": 0.0295, "step": 16660 }, { "action_loss": 0.003779676742851734, "epoch": 14.982014388489208, "step": 16660 }, { "epoch": 14.982014388489208, "step": 16660, "torque_loss": 0.15606601536273956 }, { "epoch": 14.991007194244604, "grad_norm": 0.24543991684913635, "learning_rate": 7.960554484200807e-05, "loss": 0.029, "step": 16670 }, { "action_loss": 0.005265433806926012, "epoch": 14.991007194244604, "step": 16670 }, { "epoch": 14.991007194244604, "step": 16670, "torque_loss": 0.11596450954675674 }, { "epoch": 15.0, "grad_norm": 0.17027366161346436, "learning_rate": 7.957888918097436e-05, "loss": 0.0244, "step": 16680 }, { "action_loss": 0.014637407846748829, "epoch": 15.0, "step": 16680 }, { "epoch": 15.0, "step": 16680, "torque_loss": 0.16715364158153534 }, { "epoch": 15.008992805755396, "grad_norm": 0.19685706496238708, "learning_rate": 7.955222058112962e-05, "loss": 0.0276, "step": 16690 }, { "action_loss": 0.007213737349957228, "epoch": 15.008992805755396, "step": 16690 }, { "epoch": 15.008992805755396, "step": 16690, "torque_loss": 0.2183525115251541 }, { "epoch": 15.017985611510792, "grad_norm": 0.15622608363628387, "learning_rate": 7.952553905413955e-05, "loss": 0.0261, "step": 16700 }, { "action_loss": 0.006470248103141785, "epoch": 15.017985611510792, "step": 16700 }, { "epoch": 15.017985611510792, "step": 16700, "torque_loss": 0.15820403397083282 }, { "epoch": 15.026978417266188, "grad_norm": 0.2049107551574707, "learning_rate": 7.949884461167563e-05, "loss": 0.0272, "step": 16710 }, { "action_loss": 0.00949773471802473, "epoch": 15.026978417266188, "step": 16710 }, { "epoch": 15.026978417266188, "step": 16710, "torque_loss": 0.21390819549560547 }, { "epoch": 15.035971223021583, "grad_norm": 0.1730986088514328, "learning_rate": 7.947213726541484e-05, "loss": 0.029, "step": 16720 }, { "action_loss": 0.005850309040397406, "epoch": 15.035971223021583, "step": 16720 }, { "epoch": 15.035971223021583, "step": 16720, "torque_loss": 0.11821410804986954 }, { "epoch": 15.04496402877698, "grad_norm": 0.12899768352508545, "learning_rate": 7.944541702703996e-05, "loss": 0.0232, "step": 16730 }, { "action_loss": 0.006912915501743555, "epoch": 15.04496402877698, "step": 16730 }, { "epoch": 15.04496402877698, "step": 16730, "torque_loss": 0.17071150243282318 }, { "epoch": 15.053956834532373, "grad_norm": 0.18456853926181793, "learning_rate": 7.941868390823928e-05, "loss": 0.0263, "step": 16740 }, { "action_loss": 0.005980206187814474, "epoch": 15.053956834532373, "step": 16740 }, { "epoch": 15.053956834532373, "step": 16740, "torque_loss": 0.13181786239147186 }, { "epoch": 15.06294964028777, "grad_norm": 0.2128789722919464, "learning_rate": 7.939193792070677e-05, "loss": 0.0264, "step": 16750 }, { "action_loss": 0.008025581948459148, "epoch": 15.06294964028777, "step": 16750 }, { "epoch": 15.06294964028777, "step": 16750, "torque_loss": 0.15090493857860565 }, { "epoch": 15.071942446043165, "grad_norm": 0.1597091555595398, "learning_rate": 7.936517907614207e-05, "loss": 0.0288, "step": 16760 }, { "action_loss": 0.007464766036719084, "epoch": 15.071942446043165, "step": 16760 }, { "epoch": 15.071942446043165, "step": 16760, "torque_loss": 0.16146127879619598 }, { "epoch": 15.08093525179856, "grad_norm": 0.12844452261924744, "learning_rate": 7.933840738625036e-05, "loss": 0.0274, "step": 16770 }, { "action_loss": 0.005562277976423502, "epoch": 15.08093525179856, "step": 16770 }, { "epoch": 15.08093525179856, "step": 16770, "torque_loss": 0.10406067222356796 }, { "epoch": 15.089928057553957, "grad_norm": 0.17104899883270264, "learning_rate": 7.931162286274252e-05, "loss": 0.0275, "step": 16780 }, { "action_loss": 0.004674242343753576, "epoch": 15.089928057553957, "step": 16780 }, { "epoch": 15.089928057553957, "step": 16780, "torque_loss": 0.11188779026269913 }, { "epoch": 15.098920863309353, "grad_norm": 0.19153420627117157, "learning_rate": 7.928482551733503e-05, "loss": 0.0252, "step": 16790 }, { "action_loss": 0.003707529976963997, "epoch": 15.098920863309353, "step": 16790 }, { "epoch": 15.098920863309353, "step": 16790, "torque_loss": 0.13715806603431702 }, { "epoch": 15.107913669064748, "grad_norm": 0.12855465710163116, "learning_rate": 7.925801536174991e-05, "loss": 0.0207, "step": 16800 }, { "action_loss": 0.004145676735788584, "epoch": 15.107913669064748, "step": 16800 }, { "epoch": 15.107913669064748, "step": 16800, "torque_loss": 0.18184328079223633 }, { "epoch": 15.116906474820144, "grad_norm": 0.1645287424325943, "learning_rate": 7.923119240771484e-05, "loss": 0.0269, "step": 16810 }, { "action_loss": 0.008074761368334293, "epoch": 15.116906474820144, "step": 16810 }, { "epoch": 15.116906474820144, "step": 16810, "torque_loss": 0.14668825268745422 }, { "epoch": 15.12589928057554, "grad_norm": 0.21783845126628876, "learning_rate": 7.920435666696313e-05, "loss": 0.0278, "step": 16820 }, { "action_loss": 0.008541309274733067, "epoch": 15.12589928057554, "step": 16820 }, { "epoch": 15.12589928057554, "step": 16820, "torque_loss": 0.15872043371200562 }, { "epoch": 15.134892086330936, "grad_norm": 0.1701868176460266, "learning_rate": 7.917750815123362e-05, "loss": 0.0252, "step": 16830 }, { "action_loss": 0.011033457703888416, "epoch": 15.134892086330936, "step": 16830 }, { "epoch": 15.134892086330936, "step": 16830, "torque_loss": 0.21271054446697235 }, { "epoch": 15.14388489208633, "grad_norm": 0.144536554813385, "learning_rate": 7.915064687227075e-05, "loss": 0.0269, "step": 16840 }, { "action_loss": 0.009486610069870949, "epoch": 15.14388489208633, "step": 16840 }, { "epoch": 15.14388489208633, "step": 16840, "torque_loss": 0.17999260127544403 }, { "epoch": 15.152877697841726, "grad_norm": 0.21975582838058472, "learning_rate": 7.912377284182458e-05, "loss": 0.0244, "step": 16850 }, { "action_loss": 0.014323006384074688, "epoch": 15.152877697841726, "step": 16850 }, { "epoch": 15.152877697841726, "step": 16850, "torque_loss": 0.20107297599315643 }, { "epoch": 15.161870503597122, "grad_norm": 0.15889103710651398, "learning_rate": 7.90968860716507e-05, "loss": 0.0266, "step": 16860 }, { "action_loss": 0.007023878861218691, "epoch": 15.161870503597122, "step": 16860 }, { "epoch": 15.161870503597122, "step": 16860, "torque_loss": 0.12972335517406464 }, { "epoch": 15.170863309352518, "grad_norm": 0.1598789095878601, "learning_rate": 7.906998657351033e-05, "loss": 0.0206, "step": 16870 }, { "action_loss": 0.016504483297467232, "epoch": 15.170863309352518, "step": 16870 }, { "epoch": 15.170863309352518, "step": 16870, "torque_loss": 0.24182850122451782 }, { "epoch": 15.179856115107913, "grad_norm": 0.17418430745601654, "learning_rate": 7.90430743591702e-05, "loss": 0.027, "step": 16880 }, { "action_loss": 0.015493113547563553, "epoch": 15.179856115107913, "step": 16880 }, { "epoch": 15.179856115107913, "step": 16880, "torque_loss": 0.1861128956079483 }, { "epoch": 15.18884892086331, "grad_norm": 0.16429878771305084, "learning_rate": 7.901614944040263e-05, "loss": 0.0251, "step": 16890 }, { "action_loss": 0.010259880684316158, "epoch": 15.18884892086331, "step": 16890 }, { "epoch": 15.18884892086331, "step": 16890, "torque_loss": 0.24201172590255737 }, { "epoch": 15.197841726618705, "grad_norm": 0.2049872875213623, "learning_rate": 7.89892118289855e-05, "loss": 0.0271, "step": 16900 }, { "action_loss": 0.006805474404245615, "epoch": 15.197841726618705, "step": 16900 }, { "epoch": 15.197841726618705, "step": 16900, "torque_loss": 0.18415069580078125 }, { "epoch": 15.206834532374101, "grad_norm": 0.15558432042598724, "learning_rate": 7.896226153670223e-05, "loss": 0.0228, "step": 16910 }, { "action_loss": 0.008261079899966717, "epoch": 15.206834532374101, "step": 16910 }, { "epoch": 15.206834532374101, "step": 16910, "torque_loss": 0.1764654666185379 }, { "epoch": 15.215827338129497, "grad_norm": 0.0795050859451294, "learning_rate": 7.89352985753418e-05, "loss": 0.0235, "step": 16920 }, { "action_loss": 0.011078259907662868, "epoch": 15.215827338129497, "step": 16920 }, { "epoch": 15.215827338129497, "step": 16920, "torque_loss": 0.17886191606521606 }, { "epoch": 15.224820143884893, "grad_norm": 0.17470425367355347, "learning_rate": 7.890832295669871e-05, "loss": 0.026, "step": 16930 }, { "action_loss": 0.0038725845515727997, "epoch": 15.224820143884893, "step": 16930 }, { "epoch": 15.224820143884893, "step": 16930, "torque_loss": 0.1721639782190323 }, { "epoch": 15.233812949640289, "grad_norm": 0.15127217769622803, "learning_rate": 7.888133469257303e-05, "loss": 0.0241, "step": 16940 }, { "action_loss": 0.0031395088881254196, "epoch": 15.233812949640289, "step": 16940 }, { "epoch": 15.233812949640289, "step": 16940, "torque_loss": 0.09847740083932877 }, { "epoch": 15.242805755395683, "grad_norm": 0.18090787529945374, "learning_rate": 7.885433379477034e-05, "loss": 0.0239, "step": 16950 }, { "action_loss": 0.004407352302223444, "epoch": 15.242805755395683, "step": 16950 }, { "epoch": 15.242805755395683, "step": 16950, "torque_loss": 0.15461425483226776 }, { "epoch": 15.251798561151078, "grad_norm": 0.17700165510177612, "learning_rate": 7.882732027510174e-05, "loss": 0.0359, "step": 16960 }, { "action_loss": 0.0030405966099351645, "epoch": 15.251798561151078, "step": 16960 }, { "epoch": 15.251798561151078, "step": 16960, "torque_loss": 0.1268129199743271 }, { "epoch": 15.260791366906474, "grad_norm": 0.1724492460489273, "learning_rate": 7.880029414538386e-05, "loss": 0.0245, "step": 16970 }, { "action_loss": 0.028412863612174988, "epoch": 15.260791366906474, "step": 16970 }, { "epoch": 15.260791366906474, "step": 16970, "torque_loss": 0.2450650930404663 }, { "epoch": 15.26978417266187, "grad_norm": 0.2746950685977936, "learning_rate": 7.877325541743888e-05, "loss": 0.0366, "step": 16980 }, { "action_loss": 0.007531909737735987, "epoch": 15.26978417266187, "step": 16980 }, { "epoch": 15.26978417266187, "step": 16980, "torque_loss": 0.17459452152252197 }, { "epoch": 15.278776978417266, "grad_norm": 0.19538675248622894, "learning_rate": 7.874620410309443e-05, "loss": 0.0243, "step": 16990 }, { "action_loss": 0.00930294580757618, "epoch": 15.278776978417266, "step": 16990 }, { "epoch": 15.278776978417266, "step": 16990, "torque_loss": 0.17002296447753906 }, { "epoch": 15.287769784172662, "grad_norm": 0.2843603491783142, "learning_rate": 7.871914021418368e-05, "loss": 0.0278, "step": 17000 }, { "action_loss": 0.012195144779980183, "epoch": 15.287769784172662, "step": 17000 }, { "epoch": 15.287769784172662, "step": 17000, "torque_loss": 0.16904783248901367 }, { "epoch": 15.296762589928058, "grad_norm": 0.3319053053855896, "learning_rate": 7.869206376254528e-05, "loss": 0.0318, "step": 17010 }, { "action_loss": 0.019314926117658615, "epoch": 15.296762589928058, "step": 17010 }, { "epoch": 15.296762589928058, "step": 17010, "torque_loss": 0.16755427420139313 }, { "epoch": 15.305755395683454, "grad_norm": 0.21892587840557098, "learning_rate": 7.86649747600234e-05, "loss": 0.0327, "step": 17020 }, { "action_loss": 0.012537998147308826, "epoch": 15.305755395683454, "step": 17020 }, { "epoch": 15.305755395683454, "step": 17020, "torque_loss": 0.22251899540424347 }, { "epoch": 15.31474820143885, "grad_norm": 0.37954041361808777, "learning_rate": 7.86378732184677e-05, "loss": 0.0386, "step": 17030 }, { "action_loss": 0.02209329605102539, "epoch": 15.31474820143885, "step": 17030 }, { "epoch": 15.31474820143885, "step": 17030, "torque_loss": 0.1867138147354126 }, { "epoch": 15.323741007194245, "grad_norm": 0.2662128806114197, "learning_rate": 7.86107591497333e-05, "loss": 0.0349, "step": 17040 }, { "action_loss": 0.01006261259317398, "epoch": 15.323741007194245, "step": 17040 }, { "epoch": 15.323741007194245, "step": 17040, "torque_loss": 0.18505656719207764 }, { "epoch": 15.332733812949641, "grad_norm": 0.24868550896644592, "learning_rate": 7.858363256568083e-05, "loss": 0.0308, "step": 17050 }, { "action_loss": 0.004058347083628178, "epoch": 15.332733812949641, "step": 17050 }, { "epoch": 15.332733812949641, "step": 17050, "torque_loss": 0.11333142966032028 }, { "epoch": 15.341726618705035, "grad_norm": 0.16137456893920898, "learning_rate": 7.855649347817637e-05, "loss": 0.0259, "step": 17060 }, { "action_loss": 0.005950575228780508, "epoch": 15.341726618705035, "step": 17060 }, { "epoch": 15.341726618705035, "step": 17060, "torque_loss": 0.16318635642528534 }, { "epoch": 15.350719424460431, "grad_norm": 0.21709822118282318, "learning_rate": 7.85293418990915e-05, "loss": 0.0256, "step": 17070 }, { "action_loss": 0.0034819103311747313, "epoch": 15.350719424460431, "step": 17070 }, { "epoch": 15.350719424460431, "step": 17070, "torque_loss": 0.19635792076587677 }, { "epoch": 15.359712230215827, "grad_norm": 0.16478808224201202, "learning_rate": 7.85021778403032e-05, "loss": 0.0245, "step": 17080 }, { "action_loss": 0.03306150436401367, "epoch": 15.359712230215827, "step": 17080 }, { "epoch": 15.359712230215827, "step": 17080, "torque_loss": 0.29718315601348877 }, { "epoch": 15.368705035971223, "grad_norm": 0.22329308092594147, "learning_rate": 7.847500131369397e-05, "loss": 0.0332, "step": 17090 }, { "action_loss": 0.003995938692241907, "epoch": 15.368705035971223, "step": 17090 }, { "epoch": 15.368705035971223, "step": 17090, "torque_loss": 0.13106007874011993 }, { "epoch": 15.377697841726619, "grad_norm": 0.22407498955726624, "learning_rate": 7.844781233115176e-05, "loss": 0.0289, "step": 17100 }, { "action_loss": 0.004494953900575638, "epoch": 15.377697841726619, "step": 17100 }, { "epoch": 15.377697841726619, "step": 17100, "torque_loss": 0.14494039118289948 }, { "epoch": 15.386690647482014, "grad_norm": 0.2083798497915268, "learning_rate": 7.842061090456996e-05, "loss": 0.0274, "step": 17110 }, { "action_loss": 0.00518846558406949, "epoch": 15.386690647482014, "step": 17110 }, { "epoch": 15.386690647482014, "step": 17110, "torque_loss": 0.16543732583522797 }, { "epoch": 15.39568345323741, "grad_norm": 0.14527781307697296, "learning_rate": 7.839339704584736e-05, "loss": 0.0262, "step": 17120 }, { "action_loss": 0.00839036051183939, "epoch": 15.39568345323741, "step": 17120 }, { "epoch": 15.39568345323741, "step": 17120, "torque_loss": 0.16587235033512115 }, { "epoch": 15.404676258992806, "grad_norm": 0.23340296745300293, "learning_rate": 7.836617076688828e-05, "loss": 0.0279, "step": 17130 }, { "action_loss": 0.0026397863402962685, "epoch": 15.404676258992806, "step": 17130 }, { "epoch": 15.404676258992806, "step": 17130, "torque_loss": 0.12792496383190155 }, { "epoch": 15.413669064748202, "grad_norm": 0.13055990636348724, "learning_rate": 7.833893207960236e-05, "loss": 0.0311, "step": 17140 }, { "action_loss": 0.00746953347697854, "epoch": 15.413669064748202, "step": 17140 }, { "epoch": 15.413669064748202, "step": 17140, "torque_loss": 0.18954916298389435 }, { "epoch": 15.422661870503598, "grad_norm": 0.15580390393733978, "learning_rate": 7.831168099590478e-05, "loss": 0.0244, "step": 17150 }, { "action_loss": 0.011275912635028362, "epoch": 15.422661870503598, "step": 17150 }, { "epoch": 15.422661870503598, "step": 17150, "torque_loss": 0.18528799712657928 }, { "epoch": 15.431654676258994, "grad_norm": 0.19770509004592896, "learning_rate": 7.828441752771607e-05, "loss": 0.0258, "step": 17160 }, { "action_loss": 0.003943413496017456, "epoch": 15.431654676258994, "step": 17160 }, { "epoch": 15.431654676258994, "step": 17160, "torque_loss": 0.1647362858057022 }, { "epoch": 15.440647482014388, "grad_norm": 0.16569502651691437, "learning_rate": 7.825714168696219e-05, "loss": 0.0224, "step": 17170 }, { "action_loss": 0.010842494666576385, "epoch": 15.440647482014388, "step": 17170 }, { "epoch": 15.440647482014388, "step": 17170, "torque_loss": 0.16758835315704346 }, { "epoch": 15.449640287769784, "grad_norm": 0.23225724697113037, "learning_rate": 7.82298534855745e-05, "loss": 0.0277, "step": 17180 }, { "action_loss": 0.01105294469743967, "epoch": 15.449640287769784, "step": 17180 }, { "epoch": 15.449640287769784, "step": 17180, "torque_loss": 0.20038552582263947 }, { "epoch": 15.45863309352518, "grad_norm": 0.19433696568012238, "learning_rate": 7.820255293548984e-05, "loss": 0.0323, "step": 17190 }, { "action_loss": 0.019267311319708824, "epoch": 15.45863309352518, "step": 17190 }, { "epoch": 15.45863309352518, "step": 17190, "torque_loss": 0.1991293579339981 }, { "epoch": 15.467625899280575, "grad_norm": 0.17419637739658356, "learning_rate": 7.817524004865035e-05, "loss": 0.0297, "step": 17200 }, { "action_loss": 0.007497832179069519, "epoch": 15.467625899280575, "step": 17200 }, { "epoch": 15.467625899280575, "step": 17200, "torque_loss": 0.11092785745859146 }, { "epoch": 15.476618705035971, "grad_norm": 0.1935954988002777, "learning_rate": 7.814791483700366e-05, "loss": 0.0272, "step": 17210 }, { "action_loss": 0.0062278336845338345, "epoch": 15.476618705035971, "step": 17210 }, { "epoch": 15.476618705035971, "step": 17210, "torque_loss": 0.25232818722724915 }, { "epoch": 15.485611510791367, "grad_norm": 0.23848220705986023, "learning_rate": 7.81205773125027e-05, "loss": 0.032, "step": 17220 }, { "action_loss": 0.009303661063313484, "epoch": 15.485611510791367, "step": 17220 }, { "epoch": 15.485611510791367, "step": 17220, "torque_loss": 0.15493494272232056 }, { "epoch": 15.494604316546763, "grad_norm": 0.2086597979068756, "learning_rate": 7.809322748710585e-05, "loss": 0.0283, "step": 17230 }, { "action_loss": 0.03798645734786987, "epoch": 15.494604316546763, "step": 17230 }, { "epoch": 15.494604316546763, "step": 17230, "torque_loss": 0.2328011393547058 }, { "epoch": 15.503597122302159, "grad_norm": 0.11721385270357132, "learning_rate": 7.806586537277685e-05, "loss": 0.029, "step": 17240 }, { "action_loss": 0.004505268298089504, "epoch": 15.503597122302159, "step": 17240 }, { "epoch": 15.503597122302159, "step": 17240, "torque_loss": 0.23974210023880005 }, { "epoch": 15.512589928057555, "grad_norm": 0.17027859389781952, "learning_rate": 7.803849098148484e-05, "loss": 0.0259, "step": 17250 }, { "action_loss": 0.00544849643483758, "epoch": 15.512589928057555, "step": 17250 }, { "epoch": 15.512589928057555, "step": 17250, "torque_loss": 0.1599426120519638 }, { "epoch": 15.52158273381295, "grad_norm": 0.16610272228717804, "learning_rate": 7.801110432520427e-05, "loss": 0.0247, "step": 17260 }, { "action_loss": 0.024087557569146156, "epoch": 15.52158273381295, "step": 17260 }, { "epoch": 15.52158273381295, "step": 17260, "torque_loss": 0.256637841463089 }, { "epoch": 15.530575539568346, "grad_norm": 0.17576058208942413, "learning_rate": 7.798370541591504e-05, "loss": 0.0283, "step": 17270 }, { "action_loss": 0.004418588243424892, "epoch": 15.530575539568346, "step": 17270 }, { "epoch": 15.530575539568346, "step": 17270, "torque_loss": 0.21899037063121796 }, { "epoch": 15.53956834532374, "grad_norm": 0.13293059170246124, "learning_rate": 7.795629426560232e-05, "loss": 0.0241, "step": 17280 }, { "action_loss": 0.0041373842395842075, "epoch": 15.53956834532374, "step": 17280 }, { "epoch": 15.53956834532374, "step": 17280, "torque_loss": 0.14301425218582153 }, { "epoch": 15.548561151079136, "grad_norm": 0.15366168320178986, "learning_rate": 7.792887088625671e-05, "loss": 0.0281, "step": 17290 }, { "action_loss": 0.004809594247490168, "epoch": 15.548561151079136, "step": 17290 }, { "epoch": 15.548561151079136, "step": 17290, "torque_loss": 0.09904583543539047 }, { "epoch": 15.557553956834532, "grad_norm": 0.19738300144672394, "learning_rate": 7.790143528987412e-05, "loss": 0.0269, "step": 17300 }, { "action_loss": 0.003020266769453883, "epoch": 15.557553956834532, "step": 17300 }, { "epoch": 15.557553956834532, "step": 17300, "torque_loss": 0.18931835889816284 }, { "epoch": 15.566546762589928, "grad_norm": 0.11879749596118927, "learning_rate": 7.787398748845578e-05, "loss": 0.0243, "step": 17310 }, { "action_loss": 0.006498158443719149, "epoch": 15.566546762589928, "step": 17310 }, { "epoch": 15.566546762589928, "step": 17310, "torque_loss": 0.18805868923664093 }, { "epoch": 15.575539568345324, "grad_norm": 0.19066229462623596, "learning_rate": 7.784652749400834e-05, "loss": 0.0279, "step": 17320 }, { "action_loss": 0.007478540763258934, "epoch": 15.575539568345324, "step": 17320 }, { "epoch": 15.575539568345324, "step": 17320, "torque_loss": 0.2128412276506424 }, { "epoch": 15.58453237410072, "grad_norm": 0.14774727821350098, "learning_rate": 7.78190553185437e-05, "loss": 0.0261, "step": 17330 }, { "action_loss": 0.005801381077617407, "epoch": 15.58453237410072, "step": 17330 }, { "epoch": 15.58453237410072, "step": 17330, "torque_loss": 0.16133026778697968 }, { "epoch": 15.593525179856115, "grad_norm": 0.16804605722427368, "learning_rate": 7.779157097407915e-05, "loss": 0.0271, "step": 17340 }, { "action_loss": 0.006722456309944391, "epoch": 15.593525179856115, "step": 17340 }, { "epoch": 15.593525179856115, "step": 17340, "torque_loss": 0.19297949969768524 }, { "epoch": 15.602517985611511, "grad_norm": 0.24750524759292603, "learning_rate": 7.776407447263725e-05, "loss": 0.0301, "step": 17350 }, { "action_loss": 0.008713485673069954, "epoch": 15.602517985611511, "step": 17350 }, { "epoch": 15.602517985611511, "step": 17350, "torque_loss": 0.22022847831249237 }, { "epoch": 15.611510791366907, "grad_norm": 0.2249387800693512, "learning_rate": 7.773656582624593e-05, "loss": 0.0327, "step": 17360 }, { "action_loss": 0.012541916221380234, "epoch": 15.611510791366907, "step": 17360 }, { "epoch": 15.611510791366907, "step": 17360, "torque_loss": 0.16617673635482788 }, { "epoch": 15.620503597122303, "grad_norm": 0.18361398577690125, "learning_rate": 7.770904504693837e-05, "loss": 0.0267, "step": 17370 }, { "action_loss": 0.01614127866923809, "epoch": 15.620503597122303, "step": 17370 }, { "epoch": 15.620503597122303, "step": 17370, "torque_loss": 0.18775300681591034 }, { "epoch": 15.629496402877697, "grad_norm": 0.2888108789920807, "learning_rate": 7.768151214675314e-05, "loss": 0.027, "step": 17380 }, { "action_loss": 0.003575295442715287, "epoch": 15.629496402877697, "step": 17380 }, { "epoch": 15.629496402877697, "step": 17380, "torque_loss": 0.11386126279830933 }, { "epoch": 15.638489208633093, "grad_norm": 0.15972256660461426, "learning_rate": 7.765396713773402e-05, "loss": 0.0276, "step": 17390 }, { "action_loss": 0.006615676451474428, "epoch": 15.638489208633093, "step": 17390 }, { "epoch": 15.638489208633093, "step": 17390, "torque_loss": 0.20771940052509308 }, { "epoch": 15.647482014388489, "grad_norm": 0.18826600909233093, "learning_rate": 7.762641003193017e-05, "loss": 0.0243, "step": 17400 }, { "action_loss": 0.0046968962997198105, "epoch": 15.647482014388489, "step": 17400 }, { "epoch": 15.647482014388489, "step": 17400, "torque_loss": 0.1336163729429245 }, { "epoch": 15.656474820143885, "grad_norm": 0.16549979150295258, "learning_rate": 7.759884084139597e-05, "loss": 0.0253, "step": 17410 }, { "action_loss": 0.003986987750977278, "epoch": 15.656474820143885, "step": 17410 }, { "epoch": 15.656474820143885, "step": 17410, "torque_loss": 0.11599623411893845 }, { "epoch": 15.66546762589928, "grad_norm": 0.13686694204807281, "learning_rate": 7.757125957819116e-05, "loss": 0.0222, "step": 17420 }, { "action_loss": 0.00244709849357605, "epoch": 15.66546762589928, "step": 17420 }, { "epoch": 15.66546762589928, "step": 17420, "torque_loss": 0.15691602230072021 }, { "epoch": 15.674460431654676, "grad_norm": 0.18310461938381195, "learning_rate": 7.754366625438069e-05, "loss": 0.0286, "step": 17430 }, { "action_loss": 0.0076816692017018795, "epoch": 15.674460431654676, "step": 17430 }, { "epoch": 15.674460431654676, "step": 17430, "torque_loss": 0.19836264848709106 }, { "epoch": 15.683453237410072, "grad_norm": 0.19082532823085785, "learning_rate": 7.751606088203483e-05, "loss": 0.0274, "step": 17440 }, { "action_loss": 0.012254956178367138, "epoch": 15.683453237410072, "step": 17440 }, { "epoch": 15.683453237410072, "step": 17440, "torque_loss": 0.21271686255931854 }, { "epoch": 15.692446043165468, "grad_norm": 0.18013322353363037, "learning_rate": 7.74884434732291e-05, "loss": 0.028, "step": 17450 }, { "action_loss": 0.004813971929252148, "epoch": 15.692446043165468, "step": 17450 }, { "epoch": 15.692446043165468, "step": 17450, "torque_loss": 0.1569044142961502 }, { "epoch": 15.701438848920864, "grad_norm": 0.15595124661922455, "learning_rate": 7.746081404004431e-05, "loss": 0.0269, "step": 17460 }, { "action_loss": 0.005562575999647379, "epoch": 15.701438848920864, "step": 17460 }, { "epoch": 15.701438848920864, "step": 17460, "torque_loss": 0.19156873226165771 }, { "epoch": 15.71043165467626, "grad_norm": 0.13949178159236908, "learning_rate": 7.743317259456649e-05, "loss": 0.023, "step": 17470 }, { "action_loss": 0.019850894808769226, "epoch": 15.71043165467626, "step": 17470 }, { "epoch": 15.71043165467626, "step": 17470, "torque_loss": 0.27617591619491577 }, { "epoch": 15.719424460431654, "grad_norm": 0.14280159771442413, "learning_rate": 7.740551914888696e-05, "loss": 0.0291, "step": 17480 }, { "action_loss": 0.010317365638911724, "epoch": 15.719424460431654, "step": 17480 }, { "epoch": 15.719424460431654, "step": 17480, "torque_loss": 0.2168177366256714 }, { "epoch": 15.72841726618705, "grad_norm": 0.17036430537700653, "learning_rate": 7.737785371510229e-05, "loss": 0.0275, "step": 17490 }, { "action_loss": 0.006094533484429121, "epoch": 15.72841726618705, "step": 17490 }, { "epoch": 15.72841726618705, "step": 17490, "torque_loss": 0.15257857739925385 }, { "epoch": 15.737410071942445, "grad_norm": 0.1765010803937912, "learning_rate": 7.735017630531426e-05, "loss": 0.0248, "step": 17500 }, { "action_loss": 0.004612719640135765, "epoch": 15.737410071942445, "step": 17500 }, { "epoch": 15.737410071942445, "step": 17500, "torque_loss": 0.205169215798378 }, { "epoch": 15.746402877697841, "grad_norm": 0.18342803418636322, "learning_rate": 7.732248693162991e-05, "loss": 0.026, "step": 17510 }, { "action_loss": 0.010777863673865795, "epoch": 15.746402877697841, "step": 17510 }, { "epoch": 15.746402877697841, "step": 17510, "torque_loss": 0.1779286116361618 }, { "epoch": 15.755395683453237, "grad_norm": 0.17703701555728912, "learning_rate": 7.729478560616153e-05, "loss": 0.0309, "step": 17520 }, { "action_loss": 0.004772601183503866, "epoch": 15.755395683453237, "step": 17520 }, { "epoch": 15.755395683453237, "step": 17520, "torque_loss": 0.12997475266456604 }, { "epoch": 15.764388489208633, "grad_norm": 0.1423044204711914, "learning_rate": 7.726707234102659e-05, "loss": 0.0239, "step": 17530 }, { "action_loss": 0.016569199040532112, "epoch": 15.764388489208633, "step": 17530 }, { "epoch": 15.764388489208633, "step": 17530, "torque_loss": 0.20620156824588776 }, { "epoch": 15.773381294964029, "grad_norm": 0.20897264778614044, "learning_rate": 7.723934714834785e-05, "loss": 0.0271, "step": 17540 }, { "action_loss": 0.007279558107256889, "epoch": 15.773381294964029, "step": 17540 }, { "epoch": 15.773381294964029, "step": 17540, "torque_loss": 0.18744225800037384 }, { "epoch": 15.782374100719425, "grad_norm": 0.2310231477022171, "learning_rate": 7.721161004025323e-05, "loss": 0.0306, "step": 17550 }, { "action_loss": 0.006632741540670395, "epoch": 15.782374100719425, "step": 17550 }, { "epoch": 15.782374100719425, "step": 17550, "torque_loss": 0.1559918373823166 }, { "epoch": 15.79136690647482, "grad_norm": 0.1885027438402176, "learning_rate": 7.718386102887588e-05, "loss": 0.0273, "step": 17560 }, { "action_loss": 0.008109447546303272, "epoch": 15.79136690647482, "step": 17560 }, { "epoch": 15.79136690647482, "step": 17560, "torque_loss": 0.18828792870044708 }, { "epoch": 15.800359712230216, "grad_norm": 0.2615581750869751, "learning_rate": 7.715610012635418e-05, "loss": 0.0355, "step": 17570 }, { "action_loss": 0.006829819176346064, "epoch": 15.800359712230216, "step": 17570 }, { "epoch": 15.800359712230216, "step": 17570, "torque_loss": 0.11181183904409409 }, { "epoch": 15.809352517985612, "grad_norm": 0.28485235571861267, "learning_rate": 7.712832734483166e-05, "loss": 0.0249, "step": 17580 }, { "action_loss": 0.007888606749475002, "epoch": 15.809352517985612, "step": 17580 }, { "epoch": 15.809352517985612, "step": 17580, "torque_loss": 0.140815868973732 }, { "epoch": 15.818345323741006, "grad_norm": 0.15966500341892242, "learning_rate": 7.710054269645713e-05, "loss": 0.0274, "step": 17590 }, { "action_loss": 0.014945689588785172, "epoch": 15.818345323741006, "step": 17590 }, { "epoch": 15.818345323741006, "step": 17590, "torque_loss": 0.22487397491931915 }, { "epoch": 15.827338129496402, "grad_norm": 0.21007616817951202, "learning_rate": 7.707274619338449e-05, "loss": 0.0281, "step": 17600 }, { "action_loss": 0.05538554862141609, "epoch": 15.827338129496402, "step": 17600 }, { "epoch": 15.827338129496402, "step": 17600, "torque_loss": 0.28700220584869385 }, { "epoch": 15.836330935251798, "grad_norm": 0.18211005628108978, "learning_rate": 7.704493784777289e-05, "loss": 0.0335, "step": 17610 }, { "action_loss": 0.004011752083897591, "epoch": 15.836330935251798, "step": 17610 }, { "epoch": 15.836330935251798, "step": 17610, "torque_loss": 0.15195654332637787 }, { "epoch": 15.845323741007194, "grad_norm": 0.18728812038898468, "learning_rate": 7.701711767178665e-05, "loss": 0.0243, "step": 17620 }, { "action_loss": 0.0061521450988948345, "epoch": 15.845323741007194, "step": 17620 }, { "epoch": 15.845323741007194, "step": 17620, "torque_loss": 0.12411549687385559 }, { "epoch": 15.85431654676259, "grad_norm": 0.14531806111335754, "learning_rate": 7.698928567759525e-05, "loss": 0.0241, "step": 17630 }, { "action_loss": 0.008756113238632679, "epoch": 15.85431654676259, "step": 17630 }, { "epoch": 15.85431654676259, "step": 17630, "torque_loss": 0.11698045581579208 }, { "epoch": 15.863309352517986, "grad_norm": 0.24201951920986176, "learning_rate": 7.696144187737335e-05, "loss": 0.0353, "step": 17640 }, { "action_loss": 0.004573378711938858, "epoch": 15.863309352517986, "step": 17640 }, { "epoch": 15.863309352517986, "step": 17640, "torque_loss": 0.20190303027629852 }, { "epoch": 15.872302158273381, "grad_norm": 0.17075695097446442, "learning_rate": 7.69335862833008e-05, "loss": 0.0263, "step": 17650 }, { "action_loss": 0.005956029053777456, "epoch": 15.872302158273381, "step": 17650 }, { "epoch": 15.872302158273381, "step": 17650, "torque_loss": 0.15296508371829987 }, { "epoch": 15.881294964028777, "grad_norm": 0.2540319859981537, "learning_rate": 7.690571890756256e-05, "loss": 0.029, "step": 17660 }, { "action_loss": 0.012282018549740314, "epoch": 15.881294964028777, "step": 17660 }, { "epoch": 15.881294964028777, "step": 17660, "torque_loss": 0.2505040168762207 }, { "epoch": 15.890287769784173, "grad_norm": 0.2053036242723465, "learning_rate": 7.687783976234877e-05, "loss": 0.0229, "step": 17670 }, { "action_loss": 0.007773584220558405, "epoch": 15.890287769784173, "step": 17670 }, { "epoch": 15.890287769784173, "step": 17670, "torque_loss": 0.17392611503601074 }, { "epoch": 15.899280575539569, "grad_norm": 0.1486809104681015, "learning_rate": 7.684994885985472e-05, "loss": 0.0258, "step": 17680 }, { "action_loss": 0.005349487531930208, "epoch": 15.899280575539569, "step": 17680 }, { "epoch": 15.899280575539569, "step": 17680, "torque_loss": 0.12213150411844254 }, { "epoch": 15.908273381294965, "grad_norm": 0.2295757234096527, "learning_rate": 7.682204621228083e-05, "loss": 0.0251, "step": 17690 }, { "action_loss": 0.02856607735157013, "epoch": 15.908273381294965, "step": 17690 }, { "epoch": 15.908273381294965, "step": 17690, "torque_loss": 0.3534543514251709 }, { "epoch": 15.917266187050359, "grad_norm": 0.14381155371665955, "learning_rate": 7.679413183183267e-05, "loss": 0.0335, "step": 17700 }, { "action_loss": 0.006156770512461662, "epoch": 15.917266187050359, "step": 17700 }, { "epoch": 15.917266187050359, "step": 17700, "torque_loss": 0.13632802665233612 }, { "epoch": 15.926258992805755, "grad_norm": 0.15874457359313965, "learning_rate": 7.676620573072093e-05, "loss": 0.0299, "step": 17710 }, { "action_loss": 0.022072480991482735, "epoch": 15.926258992805755, "step": 17710 }, { "epoch": 15.926258992805755, "step": 17710, "torque_loss": 0.2933259904384613 }, { "epoch": 15.93525179856115, "grad_norm": 0.18747590482234955, "learning_rate": 7.673826792116145e-05, "loss": 0.0286, "step": 17720 }, { "action_loss": 0.012891985476016998, "epoch": 15.93525179856115, "step": 17720 }, { "epoch": 15.93525179856115, "step": 17720, "torque_loss": 0.21111977100372314 }, { "epoch": 15.944244604316546, "grad_norm": 0.17870353162288666, "learning_rate": 7.671031841537519e-05, "loss": 0.027, "step": 17730 }, { "action_loss": 0.003557722084224224, "epoch": 15.944244604316546, "step": 17730 }, { "epoch": 15.944244604316546, "step": 17730, "torque_loss": 0.14018182456493378 }, { "epoch": 15.953237410071942, "grad_norm": 0.17597417533397675, "learning_rate": 7.668235722558819e-05, "loss": 0.0215, "step": 17740 }, { "action_loss": 0.008532124571502209, "epoch": 15.953237410071942, "step": 17740 }, { "epoch": 15.953237410071942, "step": 17740, "torque_loss": 0.184936061501503 }, { "epoch": 15.962230215827338, "grad_norm": 0.1765197515487671, "learning_rate": 7.665438436403161e-05, "loss": 0.0254, "step": 17750 }, { "action_loss": 0.004796160850673914, "epoch": 15.962230215827338, "step": 17750 }, { "epoch": 15.962230215827338, "step": 17750, "torque_loss": 0.15961404144763947 }, { "epoch": 15.971223021582734, "grad_norm": 0.20839260518550873, "learning_rate": 7.662639984294178e-05, "loss": 0.0282, "step": 17760 }, { "action_loss": 0.005502055864781141, "epoch": 15.971223021582734, "step": 17760 }, { "epoch": 15.971223021582734, "step": 17760, "torque_loss": 0.14036910235881805 }, { "epoch": 15.98021582733813, "grad_norm": 0.11567670106887817, "learning_rate": 7.659840367456002e-05, "loss": 0.0251, "step": 17770 }, { "action_loss": 0.009714819490909576, "epoch": 15.98021582733813, "step": 17770 }, { "epoch": 15.98021582733813, "step": 17770, "torque_loss": 0.1820969581604004 }, { "epoch": 15.989208633093526, "grad_norm": 0.23549482226371765, "learning_rate": 7.657039587113287e-05, "loss": 0.0241, "step": 17780 }, { "action_loss": 0.005093837156891823, "epoch": 15.989208633093526, "step": 17780 }, { "epoch": 15.989208633093526, "step": 17780, "torque_loss": 0.21882180869579315 }, { "epoch": 15.998201438848922, "grad_norm": 0.14245517551898956, "learning_rate": 7.654237644491182e-05, "loss": 0.0234, "step": 17790 }, { "action_loss": 0.006996897980570793, "epoch": 15.998201438848922, "step": 17790 }, { "epoch": 15.998201438848922, "step": 17790, "torque_loss": 0.19860891997814178 }, { "epoch": 16.007194244604317, "grad_norm": 0.19679756462574005, "learning_rate": 7.651434540815358e-05, "loss": 0.0329, "step": 17800 }, { "action_loss": 0.012803296558558941, "epoch": 16.007194244604317, "step": 17800 }, { "epoch": 16.007194244604317, "step": 17800, "torque_loss": 0.21719639003276825 }, { "epoch": 16.01618705035971, "grad_norm": 0.12019194662570953, "learning_rate": 7.648630277311986e-05, "loss": 0.0228, "step": 17810 }, { "action_loss": 0.006259353365749121, "epoch": 16.01618705035971, "step": 17810 }, { "epoch": 16.01618705035971, "step": 17810, "torque_loss": 0.16849608719348907 }, { "epoch": 16.02517985611511, "grad_norm": 0.18905629217624664, "learning_rate": 7.645824855207744e-05, "loss": 0.0261, "step": 17820 }, { "action_loss": 0.00819416344165802, "epoch": 16.02517985611511, "step": 17820 }, { "epoch": 16.02517985611511, "step": 17820, "torque_loss": 0.20236007869243622 }, { "epoch": 16.034172661870503, "grad_norm": 0.14535625278949738, "learning_rate": 7.643018275729821e-05, "loss": 0.0247, "step": 17830 }, { "action_loss": 0.012366066686809063, "epoch": 16.034172661870503, "step": 17830 }, { "epoch": 16.034172661870503, "step": 17830, "torque_loss": 0.28804805874824524 }, { "epoch": 16.0431654676259, "grad_norm": 0.17488634586334229, "learning_rate": 7.640210540105911e-05, "loss": 0.0259, "step": 17840 }, { "action_loss": 0.006570710334926844, "epoch": 16.0431654676259, "step": 17840 }, { "epoch": 16.0431654676259, "step": 17840, "torque_loss": 0.11502385884523392 }, { "epoch": 16.052158273381295, "grad_norm": 0.16769777238368988, "learning_rate": 7.637401649564213e-05, "loss": 0.0251, "step": 17850 }, { "action_loss": 0.0048112887889146805, "epoch": 16.052158273381295, "step": 17850 }, { "epoch": 16.052158273381295, "step": 17850, "torque_loss": 0.1865035742521286 }, { "epoch": 16.06115107913669, "grad_norm": 0.18136055767536163, "learning_rate": 7.63459160533343e-05, "loss": 0.0234, "step": 17860 }, { "action_loss": 0.010174510069191456, "epoch": 16.06115107913669, "step": 17860 }, { "epoch": 16.06115107913669, "step": 17860, "torque_loss": 0.2261877804994583 }, { "epoch": 16.070143884892087, "grad_norm": 0.17103439569473267, "learning_rate": 7.631780408642769e-05, "loss": 0.0262, "step": 17870 }, { "action_loss": 0.007114643696695566, "epoch": 16.070143884892087, "step": 17870 }, { "epoch": 16.070143884892087, "step": 17870, "torque_loss": 0.1638062447309494 }, { "epoch": 16.07913669064748, "grad_norm": 0.17381764948368073, "learning_rate": 7.628968060721946e-05, "loss": 0.0257, "step": 17880 }, { "action_loss": 0.0044964211992919445, "epoch": 16.07913669064748, "step": 17880 }, { "epoch": 16.07913669064748, "step": 17880, "torque_loss": 0.11267802864313126 }, { "epoch": 16.08812949640288, "grad_norm": 0.23288743197917938, "learning_rate": 7.626154562801176e-05, "loss": 0.0279, "step": 17890 }, { "action_loss": 0.012663707137107849, "epoch": 16.08812949640288, "step": 17890 }, { "epoch": 16.08812949640288, "step": 17890, "torque_loss": 0.21894896030426025 }, { "epoch": 16.097122302158272, "grad_norm": 0.19557444751262665, "learning_rate": 7.623339916111177e-05, "loss": 0.0266, "step": 17900 }, { "action_loss": 0.014821202494204044, "epoch": 16.097122302158272, "step": 17900 }, { "epoch": 16.097122302158272, "step": 17900, "torque_loss": 0.19775839149951935 }, { "epoch": 16.10611510791367, "grad_norm": 0.2092047780752182, "learning_rate": 7.620524121883174e-05, "loss": 0.0289, "step": 17910 }, { "action_loss": 0.003911059815436602, "epoch": 16.10611510791367, "step": 17910 }, { "epoch": 16.10611510791367, "step": 17910, "torque_loss": 0.17873036861419678 }, { "epoch": 16.115107913669064, "grad_norm": 0.20797742903232574, "learning_rate": 7.617707181348892e-05, "loss": 0.0283, "step": 17920 }, { "action_loss": 0.010250955820083618, "epoch": 16.115107913669064, "step": 17920 }, { "epoch": 16.115107913669064, "step": 17920, "torque_loss": 0.17218150198459625 }, { "epoch": 16.12410071942446, "grad_norm": 0.14154793322086334, "learning_rate": 7.614889095740554e-05, "loss": 0.0244, "step": 17930 }, { "action_loss": 0.009300144389271736, "epoch": 16.12410071942446, "step": 17930 }, { "epoch": 16.12410071942446, "step": 17930, "torque_loss": 0.16572053730487823 }, { "epoch": 16.133093525179856, "grad_norm": 0.14865589141845703, "learning_rate": 7.612069866290885e-05, "loss": 0.0247, "step": 17940 }, { "action_loss": 0.00316786696203053, "epoch": 16.133093525179856, "step": 17940 }, { "epoch": 16.133093525179856, "step": 17940, "torque_loss": 0.1434975117444992 }, { "epoch": 16.142086330935253, "grad_norm": 0.12318126112222672, "learning_rate": 7.609249494233115e-05, "loss": 0.0224, "step": 17950 }, { "action_loss": 0.00637756148353219, "epoch": 16.142086330935253, "step": 17950 }, { "epoch": 16.142086330935253, "step": 17950, "torque_loss": 0.17607474327087402 }, { "epoch": 16.151079136690647, "grad_norm": 0.16750183701515198, "learning_rate": 7.60642798080097e-05, "loss": 0.0323, "step": 17960 }, { "action_loss": 0.0032882934901863337, "epoch": 16.151079136690647, "step": 17960 }, { "epoch": 16.151079136690647, "step": 17960, "torque_loss": 0.1324407160282135 }, { "epoch": 16.16007194244604, "grad_norm": 0.17791958153247833, "learning_rate": 7.603605327228673e-05, "loss": 0.0272, "step": 17970 }, { "action_loss": 0.002998570678755641, "epoch": 16.16007194244604, "step": 17970 }, { "epoch": 16.16007194244604, "step": 17970, "torque_loss": 0.1393435001373291 }, { "epoch": 16.16906474820144, "grad_norm": 0.19023218750953674, "learning_rate": 7.600781534750953e-05, "loss": 0.0256, "step": 17980 }, { "action_loss": 0.008287985809147358, "epoch": 16.16906474820144, "step": 17980 }, { "epoch": 16.16906474820144, "step": 17980, "torque_loss": 0.23996149003505707 }, { "epoch": 16.178057553956833, "grad_norm": 0.27441152930259705, "learning_rate": 7.597956604603029e-05, "loss": 0.0327, "step": 17990 }, { "action_loss": 0.009474624879658222, "epoch": 16.178057553956833, "step": 17990 }, { "epoch": 16.178057553956833, "step": 17990, "torque_loss": 0.21199457347393036 }, { "epoch": 16.18705035971223, "grad_norm": 0.21010243892669678, "learning_rate": 7.595130538020622e-05, "loss": 0.0285, "step": 18000 }, { "action_loss": 0.0037202928215265274, "epoch": 16.18705035971223, "step": 18000 }, { "epoch": 16.18705035971223, "step": 18000, "torque_loss": 0.08550474792718887 }, { "epoch": 16.196043165467625, "grad_norm": 0.21125400066375732, "learning_rate": 7.592303336239952e-05, "loss": 0.0244, "step": 18010 }, { "action_loss": 0.012638733722269535, "epoch": 16.196043165467625, "step": 18010 }, { "epoch": 16.196043165467625, "step": 18010, "torque_loss": 0.1943662315607071 }, { "epoch": 16.205035971223023, "grad_norm": 0.20621782541275024, "learning_rate": 7.589475000497733e-05, "loss": 0.0239, "step": 18020 }, { "action_loss": 0.011806018650531769, "epoch": 16.205035971223023, "step": 18020 }, { "epoch": 16.205035971223023, "step": 18020, "torque_loss": 0.1397971659898758 }, { "epoch": 16.214028776978417, "grad_norm": 0.14142552018165588, "learning_rate": 7.586645532031171e-05, "loss": 0.0236, "step": 18030 }, { "action_loss": 0.004602537490427494, "epoch": 16.214028776978417, "step": 18030 }, { "epoch": 16.214028776978417, "step": 18030, "torque_loss": 0.11392756551504135 }, { "epoch": 16.223021582733814, "grad_norm": 0.17165783047676086, "learning_rate": 7.583814932077975e-05, "loss": 0.0235, "step": 18040 }, { "action_loss": 0.00439723813906312, "epoch": 16.223021582733814, "step": 18040 }, { "epoch": 16.223021582733814, "step": 18040, "torque_loss": 0.11797577142715454 }, { "epoch": 16.23201438848921, "grad_norm": 0.2556707262992859, "learning_rate": 7.580983201876343e-05, "loss": 0.0292, "step": 18050 }, { "action_loss": 0.00804500188678503, "epoch": 16.23201438848921, "step": 18050 }, { "epoch": 16.23201438848921, "step": 18050, "torque_loss": 0.11639169603586197 }, { "epoch": 16.241007194244606, "grad_norm": 0.14992490410804749, "learning_rate": 7.578150342664973e-05, "loss": 0.0212, "step": 18060 }, { "action_loss": 0.0038039765786379576, "epoch": 16.241007194244606, "step": 18060 }, { "epoch": 16.241007194244606, "step": 18060, "torque_loss": 0.10899501293897629 }, { "epoch": 16.25, "grad_norm": 0.13805744051933289, "learning_rate": 7.575316355683055e-05, "loss": 0.0229, "step": 18070 }, { "action_loss": 0.00496792234480381, "epoch": 16.25, "step": 18070 }, { "epoch": 16.25, "step": 18070, "torque_loss": 0.14491936564445496 }, { "epoch": 16.258992805755394, "grad_norm": 0.15197338163852692, "learning_rate": 7.572481242170266e-05, "loss": 0.0242, "step": 18080 }, { "action_loss": 0.005189352668821812, "epoch": 16.258992805755394, "step": 18080 }, { "epoch": 16.258992805755394, "step": 18080, "torque_loss": 0.13090737164020538 }, { "epoch": 16.26798561151079, "grad_norm": 0.18560338020324707, "learning_rate": 7.569645003366786e-05, "loss": 0.0227, "step": 18090 }, { "action_loss": 0.013069731183350086, "epoch": 16.26798561151079, "step": 18090 }, { "epoch": 16.26798561151079, "step": 18090, "torque_loss": 0.20149068534374237 }, { "epoch": 16.276978417266186, "grad_norm": 0.20448260009288788, "learning_rate": 7.566807640513278e-05, "loss": 0.0268, "step": 18100 }, { "action_loss": 0.003941286355257034, "epoch": 16.276978417266186, "step": 18100 }, { "epoch": 16.276978417266186, "step": 18100, "torque_loss": 0.1900024265050888 }, { "epoch": 16.285971223021583, "grad_norm": 0.16104057431221008, "learning_rate": 7.563969154850905e-05, "loss": 0.0299, "step": 18110 }, { "action_loss": 0.011571814306080341, "epoch": 16.285971223021583, "step": 18110 }, { "epoch": 16.285971223021583, "step": 18110, "torque_loss": 0.1524675488471985 }, { "epoch": 16.294964028776977, "grad_norm": 0.10985659807920456, "learning_rate": 7.561129547621314e-05, "loss": 0.021, "step": 18120 }, { "action_loss": 0.004704346880316734, "epoch": 16.294964028776977, "step": 18120 }, { "epoch": 16.294964028776977, "step": 18120, "torque_loss": 0.17088651657104492 }, { "epoch": 16.303956834532375, "grad_norm": 0.1742449551820755, "learning_rate": 7.558288820066648e-05, "loss": 0.0234, "step": 18130 }, { "action_loss": 0.00886985007673502, "epoch": 16.303956834532375, "step": 18130 }, { "epoch": 16.303956834532375, "step": 18130, "torque_loss": 0.16490839421749115 }, { "epoch": 16.31294964028777, "grad_norm": 0.18544407188892365, "learning_rate": 7.555446973429534e-05, "loss": 0.0288, "step": 18140 }, { "action_loss": 0.0073993452824652195, "epoch": 16.31294964028777, "step": 18140 }, { "epoch": 16.31294964028777, "step": 18140, "torque_loss": 0.22173388302326202 }, { "epoch": 16.321942446043167, "grad_norm": 0.15972398221492767, "learning_rate": 7.552604008953097e-05, "loss": 0.0292, "step": 18150 }, { "action_loss": 0.004761522635817528, "epoch": 16.321942446043167, "step": 18150 }, { "epoch": 16.321942446043167, "step": 18150, "torque_loss": 0.13457345962524414 }, { "epoch": 16.33093525179856, "grad_norm": 0.20439720153808594, "learning_rate": 7.549759927880942e-05, "loss": 0.0239, "step": 18160 }, { "action_loss": 0.0025094670709222555, "epoch": 16.33093525179856, "step": 18160 }, { "epoch": 16.33093525179856, "step": 18160, "torque_loss": 0.14216218888759613 }, { "epoch": 16.33992805755396, "grad_norm": 0.1541639268398285, "learning_rate": 7.54691473145717e-05, "loss": 0.0232, "step": 18170 }, { "action_loss": 0.005941762123256922, "epoch": 16.33992805755396, "step": 18170 }, { "epoch": 16.33992805755396, "step": 18170, "torque_loss": 0.1520761251449585 }, { "epoch": 16.348920863309353, "grad_norm": 0.1871304214000702, "learning_rate": 7.544068420926365e-05, "loss": 0.0275, "step": 18180 }, { "action_loss": 0.003345802426338196, "epoch": 16.348920863309353, "step": 18180 }, { "epoch": 16.348920863309353, "step": 18180, "torque_loss": 0.10805019736289978 }, { "epoch": 16.357913669064747, "grad_norm": 0.16050568222999573, "learning_rate": 7.5412209975336e-05, "loss": 0.0243, "step": 18190 }, { "action_loss": 0.005326725076884031, "epoch": 16.357913669064747, "step": 18190 }, { "epoch": 16.357913669064747, "step": 18190, "torque_loss": 0.19686861336231232 }, { "epoch": 16.366906474820144, "grad_norm": 0.1778029501438141, "learning_rate": 7.538372462524435e-05, "loss": 0.0277, "step": 18200 }, { "action_loss": 0.009341910481452942, "epoch": 16.366906474820144, "step": 18200 }, { "epoch": 16.366906474820144, "step": 18200, "torque_loss": 0.19948823750019073 }, { "epoch": 16.37589928057554, "grad_norm": 0.13005395233631134, "learning_rate": 7.535522817144915e-05, "loss": 0.0258, "step": 18210 }, { "action_loss": 0.008734260685741901, "epoch": 16.37589928057554, "step": 18210 }, { "epoch": 16.37589928057554, "step": 18210, "torque_loss": 0.2183040976524353 }, { "epoch": 16.384892086330936, "grad_norm": 0.1839977502822876, "learning_rate": 7.532672062641575e-05, "loss": 0.0336, "step": 18220 }, { "action_loss": 0.009408458136022091, "epoch": 16.384892086330936, "step": 18220 }, { "epoch": 16.384892086330936, "step": 18220, "torque_loss": 0.1721487045288086 }, { "epoch": 16.39388489208633, "grad_norm": 0.19940851628780365, "learning_rate": 7.529820200261427e-05, "loss": 0.0305, "step": 18230 }, { "action_loss": 0.006654661148786545, "epoch": 16.39388489208633, "step": 18230 }, { "epoch": 16.39388489208633, "step": 18230, "torque_loss": 0.15105263888835907 }, { "epoch": 16.402877697841728, "grad_norm": 0.2318461388349533, "learning_rate": 7.526967231251977e-05, "loss": 0.0243, "step": 18240 }, { "action_loss": 0.004604358226060867, "epoch": 16.402877697841728, "step": 18240 }, { "epoch": 16.402877697841728, "step": 18240, "torque_loss": 0.09833366423845291 }, { "epoch": 16.41187050359712, "grad_norm": 0.12737949192523956, "learning_rate": 7.524113156861206e-05, "loss": 0.0205, "step": 18250 }, { "action_loss": 0.011750292964279652, "epoch": 16.41187050359712, "step": 18250 }, { "epoch": 16.41187050359712, "step": 18250, "torque_loss": 0.12092309445142746 }, { "epoch": 16.42086330935252, "grad_norm": 0.1400890052318573, "learning_rate": 7.521257978337587e-05, "loss": 0.0227, "step": 18260 }, { "action_loss": 0.0036512520164251328, "epoch": 16.42086330935252, "step": 18260 }, { "epoch": 16.42086330935252, "step": 18260, "torque_loss": 0.1726120561361313 }, { "epoch": 16.429856115107913, "grad_norm": 0.168061763048172, "learning_rate": 7.518401696930072e-05, "loss": 0.0252, "step": 18270 }, { "action_loss": 0.006329696625471115, "epoch": 16.429856115107913, "step": 18270 }, { "epoch": 16.429856115107913, "step": 18270, "torque_loss": 0.18545491993427277 }, { "epoch": 16.43884892086331, "grad_norm": 0.12935742735862732, "learning_rate": 7.515544313888092e-05, "loss": 0.0284, "step": 18280 }, { "action_loss": 0.0029496243223547935, "epoch": 16.43884892086331, "step": 18280 }, { "epoch": 16.43884892086331, "step": 18280, "torque_loss": 0.10036798566579819 }, { "epoch": 16.447841726618705, "grad_norm": 0.15768662095069885, "learning_rate": 7.512685830461568e-05, "loss": 0.0267, "step": 18290 }, { "action_loss": 0.02265513874590397, "epoch": 16.447841726618705, "step": 18290 }, { "epoch": 16.447841726618705, "step": 18290, "torque_loss": 0.21844488382339478 }, { "epoch": 16.4568345323741, "grad_norm": 0.18949052691459656, "learning_rate": 7.509826247900892e-05, "loss": 0.0326, "step": 18300 }, { "action_loss": 0.01358390599489212, "epoch": 16.4568345323741, "step": 18300 }, { "epoch": 16.4568345323741, "step": 18300, "torque_loss": 0.23019075393676758 }, { "epoch": 16.465827338129497, "grad_norm": 0.29081445932388306, "learning_rate": 7.50696556745695e-05, "loss": 0.027, "step": 18310 }, { "action_loss": 0.016152398660779, "epoch": 16.465827338129497, "step": 18310 }, { "epoch": 16.465827338129497, "step": 18310, "torque_loss": 0.25664785504341125 }, { "epoch": 16.47482014388489, "grad_norm": 0.18473103642463684, "learning_rate": 7.504103790381095e-05, "loss": 0.0287, "step": 18320 }, { "action_loss": 0.0057393629103899, "epoch": 16.47482014388489, "step": 18320 }, { "epoch": 16.47482014388489, "step": 18320, "torque_loss": 0.13667233288288116 }, { "epoch": 16.48381294964029, "grad_norm": 0.18138159811496735, "learning_rate": 7.501240917925167e-05, "loss": 0.0282, "step": 18330 }, { "action_loss": 0.0032766994554549456, "epoch": 16.48381294964029, "step": 18330 }, { "epoch": 16.48381294964029, "step": 18330, "torque_loss": 0.12194094806909561 }, { "epoch": 16.492805755395683, "grad_norm": 0.22869063913822174, "learning_rate": 7.498376951341483e-05, "loss": 0.035, "step": 18340 }, { "action_loss": 0.011488559655845165, "epoch": 16.492805755395683, "step": 18340 }, { "epoch": 16.492805755395683, "step": 18340, "torque_loss": 0.19534547626972198 }, { "epoch": 16.50179856115108, "grad_norm": 0.14215272665023804, "learning_rate": 7.495511891882844e-05, "loss": 0.0275, "step": 18350 }, { "action_loss": 0.005431142169982195, "epoch": 16.50179856115108, "step": 18350 }, { "epoch": 16.50179856115108, "step": 18350, "torque_loss": 0.1135612204670906 }, { "epoch": 16.510791366906474, "grad_norm": 0.20555372536182404, "learning_rate": 7.492645740802517e-05, "loss": 0.0272, "step": 18360 }, { "action_loss": 0.009431847371160984, "epoch": 16.510791366906474, "step": 18360 }, { "epoch": 16.510791366906474, "step": 18360, "torque_loss": 0.13881820440292358 }, { "epoch": 16.519784172661872, "grad_norm": 0.25450965762138367, "learning_rate": 7.489778499354259e-05, "loss": 0.0237, "step": 18370 }, { "action_loss": 0.0063600726425647736, "epoch": 16.519784172661872, "step": 18370 }, { "epoch": 16.519784172661872, "step": 18370, "torque_loss": 0.15433402359485626 }, { "epoch": 16.528776978417266, "grad_norm": 0.19664040207862854, "learning_rate": 7.486910168792298e-05, "loss": 0.0254, "step": 18380 }, { "action_loss": 0.005094888620078564, "epoch": 16.528776978417266, "step": 18380 }, { "epoch": 16.528776978417266, "step": 18380, "torque_loss": 0.12526445090770721 }, { "epoch": 16.53776978417266, "grad_norm": 0.12017615884542465, "learning_rate": 7.484040750371337e-05, "loss": 0.0293, "step": 18390 }, { "action_loss": 0.006440289318561554, "epoch": 16.53776978417266, "step": 18390 }, { "epoch": 16.53776978417266, "step": 18390, "torque_loss": 0.22523467242717743 }, { "epoch": 16.546762589928058, "grad_norm": 0.17284122109413147, "learning_rate": 7.48117024534656e-05, "loss": 0.0256, "step": 18400 }, { "action_loss": 0.0057229953818023205, "epoch": 16.546762589928058, "step": 18400 }, { "epoch": 16.546762589928058, "step": 18400, "torque_loss": 0.11033066362142563 }, { "epoch": 16.555755395683452, "grad_norm": 0.15533237159252167, "learning_rate": 7.478298654973621e-05, "loss": 0.0275, "step": 18410 }, { "action_loss": 0.0033234190195798874, "epoch": 16.555755395683452, "step": 18410 }, { "epoch": 16.555755395683452, "step": 18410, "torque_loss": 0.15884171426296234 }, { "epoch": 16.56474820143885, "grad_norm": 0.22943174839019775, "learning_rate": 7.475425980508654e-05, "loss": 0.0268, "step": 18420 }, { "action_loss": 0.00635510915890336, "epoch": 16.56474820143885, "step": 18420 }, { "epoch": 16.56474820143885, "step": 18420, "torque_loss": 0.197931706905365 }, { "epoch": 16.573741007194243, "grad_norm": 0.1561257690191269, "learning_rate": 7.47255222320826e-05, "loss": 0.0303, "step": 18430 }, { "action_loss": 0.003953635226935148, "epoch": 16.573741007194243, "step": 18430 }, { "epoch": 16.573741007194243, "step": 18430, "torque_loss": 0.09540310502052307 }, { "epoch": 16.58273381294964, "grad_norm": 0.1097845509648323, "learning_rate": 7.469677384329522e-05, "loss": 0.0292, "step": 18440 }, { "action_loss": 0.004112576600164175, "epoch": 16.58273381294964, "step": 18440 }, { "epoch": 16.58273381294964, "step": 18440, "torque_loss": 0.1744411736726761 }, { "epoch": 16.591726618705035, "grad_norm": 0.15889239311218262, "learning_rate": 7.46680146512999e-05, "loss": 0.0259, "step": 18450 }, { "action_loss": 0.004128531087189913, "epoch": 16.591726618705035, "step": 18450 }, { "epoch": 16.591726618705035, "step": 18450, "torque_loss": 0.19088202714920044 }, { "epoch": 16.600719424460433, "grad_norm": 0.14640529453754425, "learning_rate": 7.463924466867692e-05, "loss": 0.0267, "step": 18460 }, { "action_loss": 0.007191802840679884, "epoch": 16.600719424460433, "step": 18460 }, { "epoch": 16.600719424460433, "step": 18460, "torque_loss": 0.19217412173748016 }, { "epoch": 16.609712230215827, "grad_norm": 0.18546687066555023, "learning_rate": 7.461046390801121e-05, "loss": 0.0272, "step": 18470 }, { "action_loss": 0.004565414506942034, "epoch": 16.609712230215827, "step": 18470 }, { "epoch": 16.609712230215827, "step": 18470, "torque_loss": 0.16002371907234192 }, { "epoch": 16.618705035971225, "grad_norm": 0.1531180739402771, "learning_rate": 7.458167238189248e-05, "loss": 0.0225, "step": 18480 }, { "action_loss": 0.005718864966183901, "epoch": 16.618705035971225, "step": 18480 }, { "epoch": 16.618705035971225, "step": 18480, "torque_loss": 0.17797459661960602 }, { "epoch": 16.62769784172662, "grad_norm": 0.2373874932527542, "learning_rate": 7.455287010291509e-05, "loss": 0.0309, "step": 18490 }, { "action_loss": 0.007092036306858063, "epoch": 16.62769784172662, "step": 18490 }, { "epoch": 16.62769784172662, "step": 18490, "torque_loss": 0.16608156263828278 }, { "epoch": 16.636690647482013, "grad_norm": 0.16346952319145203, "learning_rate": 7.452405708367818e-05, "loss": 0.027, "step": 18500 }, { "action_loss": 0.0066217719577252865, "epoch": 16.636690647482013, "step": 18500 }, { "epoch": 16.636690647482013, "step": 18500, "torque_loss": 0.21949392557144165 }, { "epoch": 16.64568345323741, "grad_norm": 0.22210067510604858, "learning_rate": 7.449523333678549e-05, "loss": 0.0273, "step": 18510 }, { "action_loss": 0.00905030220746994, "epoch": 16.64568345323741, "step": 18510 }, { "epoch": 16.64568345323741, "step": 18510, "torque_loss": 0.18421514332294464 }, { "epoch": 16.654676258992804, "grad_norm": 0.20543192327022552, "learning_rate": 7.446639887484555e-05, "loss": 0.024, "step": 18520 }, { "action_loss": 0.01170292031019926, "epoch": 16.654676258992804, "step": 18520 }, { "epoch": 16.654676258992804, "step": 18520, "torque_loss": 0.23470771312713623 }, { "epoch": 16.663669064748202, "grad_norm": 0.1772155910730362, "learning_rate": 7.443755371047152e-05, "loss": 0.0285, "step": 18530 }, { "action_loss": 0.005259264260530472, "epoch": 16.663669064748202, "step": 18530 }, { "epoch": 16.663669064748202, "step": 18530, "torque_loss": 0.15692083537578583 }, { "epoch": 16.672661870503596, "grad_norm": 0.13260012865066528, "learning_rate": 7.440869785628124e-05, "loss": 0.0266, "step": 18540 }, { "action_loss": 0.004011450335383415, "epoch": 16.672661870503596, "step": 18540 }, { "epoch": 16.672661870503596, "step": 18540, "torque_loss": 0.13455736637115479 }, { "epoch": 16.681654676258994, "grad_norm": 0.17277446389198303, "learning_rate": 7.437983132489727e-05, "loss": 0.024, "step": 18550 }, { "action_loss": 0.008724978193640709, "epoch": 16.681654676258994, "step": 18550 }, { "epoch": 16.681654676258994, "step": 18550, "torque_loss": 0.18267077207565308 }, { "epoch": 16.690647482014388, "grad_norm": 0.14089395105838776, "learning_rate": 7.435095412894678e-05, "loss": 0.0242, "step": 18560 }, { "action_loss": 0.006566143129020929, "epoch": 16.690647482014388, "step": 18560 }, { "epoch": 16.690647482014388, "step": 18560, "torque_loss": 0.22331804037094116 }, { "epoch": 16.699640287769785, "grad_norm": 0.12983338534832, "learning_rate": 7.432206628106165e-05, "loss": 0.0225, "step": 18570 }, { "action_loss": 0.00410258024930954, "epoch": 16.699640287769785, "step": 18570 }, { "epoch": 16.699640287769785, "step": 18570, "torque_loss": 0.13766251504421234 }, { "epoch": 16.70863309352518, "grad_norm": 0.10957200080156326, "learning_rate": 7.429316779387842e-05, "loss": 0.0235, "step": 18580 }, { "action_loss": 0.003995430190116167, "epoch": 16.70863309352518, "step": 18580 }, { "epoch": 16.70863309352518, "step": 18580, "torque_loss": 0.145485520362854 }, { "epoch": 16.717625899280577, "grad_norm": 0.24192236363887787, "learning_rate": 7.426425868003825e-05, "loss": 0.0289, "step": 18590 }, { "action_loss": 0.00523468479514122, "epoch": 16.717625899280577, "step": 18590 }, { "epoch": 16.717625899280577, "step": 18590, "torque_loss": 0.17754323780536652 }, { "epoch": 16.72661870503597, "grad_norm": 0.198989599943161, "learning_rate": 7.423533895218698e-05, "loss": 0.0264, "step": 18600 }, { "action_loss": 0.0032333750277757645, "epoch": 16.72661870503597, "step": 18600 }, { "epoch": 16.72661870503597, "step": 18600, "torque_loss": 0.14364756643772125 }, { "epoch": 16.735611510791365, "grad_norm": 0.1893932968378067, "learning_rate": 7.420640862297507e-05, "loss": 0.0238, "step": 18610 }, { "action_loss": 0.0035887062549591064, "epoch": 16.735611510791365, "step": 18610 }, { "epoch": 16.735611510791365, "step": 18610, "torque_loss": 0.129144087433815 }, { "epoch": 16.744604316546763, "grad_norm": 0.1843235194683075, "learning_rate": 7.417746770505766e-05, "loss": 0.0286, "step": 18620 }, { "action_loss": 0.01180723961442709, "epoch": 16.744604316546763, "step": 18620 }, { "epoch": 16.744604316546763, "step": 18620, "torque_loss": 0.21802359819412231 }, { "epoch": 16.753597122302157, "grad_norm": 0.20086729526519775, "learning_rate": 7.414851621109444e-05, "loss": 0.0243, "step": 18630 }, { "action_loss": 0.01818993128836155, "epoch": 16.753597122302157, "step": 18630 }, { "epoch": 16.753597122302157, "step": 18630, "torque_loss": 0.2569330632686615 }, { "epoch": 16.762589928057555, "grad_norm": 0.1725822538137436, "learning_rate": 7.411955415374982e-05, "loss": 0.027, "step": 18640 }, { "action_loss": 0.002520430600270629, "epoch": 16.762589928057555, "step": 18640 }, { "epoch": 16.762589928057555, "step": 18640, "torque_loss": 0.14170299470424652 }, { "epoch": 16.77158273381295, "grad_norm": 0.12697875499725342, "learning_rate": 7.409058154569278e-05, "loss": 0.0246, "step": 18650 }, { "action_loss": 0.012167845852673054, "epoch": 16.77158273381295, "step": 18650 }, { "epoch": 16.77158273381295, "step": 18650, "torque_loss": 0.24139106273651123 }, { "epoch": 16.780575539568346, "grad_norm": 0.21686483919620514, "learning_rate": 7.406159839959689e-05, "loss": 0.0316, "step": 18660 }, { "action_loss": 0.007673734799027443, "epoch": 16.780575539568346, "step": 18660 }, { "epoch": 16.780575539568346, "step": 18660, "torque_loss": 0.2169157862663269 }, { "epoch": 16.78956834532374, "grad_norm": 0.21652692556381226, "learning_rate": 7.403260472814039e-05, "loss": 0.0294, "step": 18670 }, { "action_loss": 0.004239364061504602, "epoch": 16.78956834532374, "step": 18670 }, { "epoch": 16.78956834532374, "step": 18670, "torque_loss": 0.24540305137634277 }, { "epoch": 16.798561151079138, "grad_norm": 0.12017788738012314, "learning_rate": 7.400360054400611e-05, "loss": 0.0269, "step": 18680 }, { "action_loss": 0.008088127709925175, "epoch": 16.798561151079138, "step": 18680 }, { "epoch": 16.798561151079138, "step": 18680, "torque_loss": 0.1714942455291748 }, { "epoch": 16.807553956834532, "grad_norm": 0.14970213174819946, "learning_rate": 7.397458585988143e-05, "loss": 0.0253, "step": 18690 }, { "action_loss": 0.00584146985784173, "epoch": 16.807553956834532, "step": 18690 }, { "epoch": 16.807553956834532, "step": 18690, "torque_loss": 0.14723612368106842 }, { "epoch": 16.81654676258993, "grad_norm": 0.11477617919445038, "learning_rate": 7.394556068845841e-05, "loss": 0.0237, "step": 18700 }, { "action_loss": 0.004099185112863779, "epoch": 16.81654676258993, "step": 18700 }, { "epoch": 16.81654676258993, "step": 18700, "torque_loss": 0.21481244266033173 }, { "epoch": 16.825539568345324, "grad_norm": 0.14808602631092072, "learning_rate": 7.39165250424336e-05, "loss": 0.0275, "step": 18710 }, { "action_loss": 0.004383049439638853, "epoch": 16.825539568345324, "step": 18710 }, { "epoch": 16.825539568345324, "step": 18710, "torque_loss": 0.14601463079452515 }, { "epoch": 16.834532374100718, "grad_norm": 0.21575608849525452, "learning_rate": 7.38874789345082e-05, "loss": 0.0299, "step": 18720 }, { "action_loss": 0.011743874289095402, "epoch": 16.834532374100718, "step": 18720 }, { "epoch": 16.834532374100718, "step": 18720, "torque_loss": 0.19752247631549835 }, { "epoch": 16.843525179856115, "grad_norm": 0.16041381657123566, "learning_rate": 7.385842237738794e-05, "loss": 0.0288, "step": 18730 }, { "action_loss": 0.020290417596697807, "epoch": 16.843525179856115, "step": 18730 }, { "epoch": 16.843525179856115, "step": 18730, "torque_loss": 0.16667063534259796 }, { "epoch": 16.85251798561151, "grad_norm": 0.1394721418619156, "learning_rate": 7.382935538378318e-05, "loss": 0.0283, "step": 18740 }, { "action_loss": 0.005998073611408472, "epoch": 16.85251798561151, "step": 18740 }, { "epoch": 16.85251798561151, "step": 18740, "torque_loss": 0.15122656524181366 }, { "epoch": 16.861510791366907, "grad_norm": 0.145687997341156, "learning_rate": 7.38002779664088e-05, "loss": 0.0245, "step": 18750 }, { "action_loss": 0.0065108356066048145, "epoch": 16.861510791366907, "step": 18750 }, { "epoch": 16.861510791366907, "step": 18750, "torque_loss": 0.23853857815265656 }, { "epoch": 16.8705035971223, "grad_norm": 0.1582309603691101, "learning_rate": 7.377119013798422e-05, "loss": 0.0304, "step": 18760 }, { "action_loss": 0.0037261508405208588, "epoch": 16.8705035971223, "step": 18760 }, { "epoch": 16.8705035971223, "step": 18760, "torque_loss": 0.09815946221351624 }, { "epoch": 16.8794964028777, "grad_norm": 0.23737287521362305, "learning_rate": 7.374209191123349e-05, "loss": 0.0265, "step": 18770 }, { "action_loss": 0.011521848849952221, "epoch": 16.8794964028777, "step": 18770 }, { "epoch": 16.8794964028777, "step": 18770, "torque_loss": 0.29483550786972046 }, { "epoch": 16.888489208633093, "grad_norm": 0.1939205825328827, "learning_rate": 7.371298329888512e-05, "loss": 0.0298, "step": 18780 }, { "action_loss": 0.00568612664937973, "epoch": 16.888489208633093, "step": 18780 }, { "epoch": 16.888489208633093, "step": 18780, "torque_loss": 0.1487354189157486 }, { "epoch": 16.89748201438849, "grad_norm": 0.12327494472265244, "learning_rate": 7.368386431367223e-05, "loss": 0.0218, "step": 18790 }, { "action_loss": 0.005887256935238838, "epoch": 16.89748201438849, "step": 18790 }, { "epoch": 16.89748201438849, "step": 18790, "torque_loss": 0.14240767061710358 }, { "epoch": 16.906474820143885, "grad_norm": 0.11580915749073029, "learning_rate": 7.365473496833246e-05, "loss": 0.0249, "step": 18800 }, { "action_loss": 0.0033399772364646196, "epoch": 16.906474820143885, "step": 18800 }, { "epoch": 16.906474820143885, "step": 18800, "torque_loss": 0.08425765484571457 }, { "epoch": 16.915467625899282, "grad_norm": 0.12494596838951111, "learning_rate": 7.362559527560797e-05, "loss": 0.0242, "step": 18810 }, { "action_loss": 0.005311096087098122, "epoch": 16.915467625899282, "step": 18810 }, { "epoch": 16.915467625899282, "step": 18810, "torque_loss": 0.18086813390254974 }, { "epoch": 16.924460431654676, "grad_norm": 0.14357179403305054, "learning_rate": 7.359644524824545e-05, "loss": 0.0241, "step": 18820 }, { "action_loss": 0.0029133555945008993, "epoch": 16.924460431654676, "step": 18820 }, { "epoch": 16.924460431654676, "step": 18820, "torque_loss": 0.17879247665405273 }, { "epoch": 16.93345323741007, "grad_norm": 0.24765393137931824, "learning_rate": 7.35672848989961e-05, "loss": 0.0284, "step": 18830 }, { "action_loss": 0.004151089582592249, "epoch": 16.93345323741007, "step": 18830 }, { "epoch": 16.93345323741007, "step": 18830, "torque_loss": 0.20418083667755127 }, { "epoch": 16.942446043165468, "grad_norm": 0.18767191469669342, "learning_rate": 7.353811424061565e-05, "loss": 0.0235, "step": 18840 }, { "action_loss": 0.006805283483117819, "epoch": 16.942446043165468, "step": 18840 }, { "epoch": 16.942446043165468, "step": 18840, "torque_loss": 0.1689748764038086 }, { "epoch": 16.951438848920862, "grad_norm": 0.18691879510879517, "learning_rate": 7.350893328586435e-05, "loss": 0.0251, "step": 18850 }, { "action_loss": 0.0053267874754965305, "epoch": 16.951438848920862, "step": 18850 }, { "epoch": 16.951438848920862, "step": 18850, "torque_loss": 0.13641990721225739 }, { "epoch": 16.96043165467626, "grad_norm": 0.24113285541534424, "learning_rate": 7.347974204750696e-05, "loss": 0.0259, "step": 18860 }, { "action_loss": 0.005568645894527435, "epoch": 16.96043165467626, "step": 18860 }, { "epoch": 16.96043165467626, "step": 18860, "torque_loss": 0.13273917138576508 }, { "epoch": 16.969424460431654, "grad_norm": 0.13457167148590088, "learning_rate": 7.345054053831266e-05, "loss": 0.0221, "step": 18870 }, { "action_loss": 0.008917546831071377, "epoch": 16.969424460431654, "step": 18870 }, { "epoch": 16.969424460431654, "step": 18870, "torque_loss": 0.16134054958820343 }, { "epoch": 16.97841726618705, "grad_norm": 0.28665733337402344, "learning_rate": 7.342132877105523e-05, "loss": 0.0356, "step": 18880 }, { "action_loss": 0.004711789544671774, "epoch": 16.97841726618705, "step": 18880 }, { "epoch": 16.97841726618705, "step": 18880, "torque_loss": 0.20705847442150116 }, { "epoch": 16.987410071942445, "grad_norm": 0.249512180685997, "learning_rate": 7.339210675851288e-05, "loss": 0.0265, "step": 18890 }, { "action_loss": 0.00257282261736691, "epoch": 16.987410071942445, "step": 18890 }, { "epoch": 16.987410071942445, "step": 18890, "torque_loss": 0.09867548942565918 }, { "epoch": 16.996402877697843, "grad_norm": 0.12468027323484421, "learning_rate": 7.336287451346831e-05, "loss": 0.0214, "step": 18900 }, { "action_loss": 0.002184713026508689, "epoch": 16.996402877697843, "step": 18900 }, { "epoch": 16.996402877697843, "step": 18900, "torque_loss": 0.12911677360534668 }, { "epoch": 17.005395683453237, "grad_norm": 0.13015203177928925, "learning_rate": 7.333363204870868e-05, "loss": 0.0247, "step": 18910 }, { "action_loss": 0.008119267411530018, "epoch": 17.005395683453237, "step": 18910 }, { "epoch": 17.005395683453237, "step": 18910, "torque_loss": 0.20803172886371613 }, { "epoch": 17.014388489208635, "grad_norm": 0.15351541340351105, "learning_rate": 7.330437937702566e-05, "loss": 0.0305, "step": 18920 }, { "action_loss": 0.0040747420862317085, "epoch": 17.014388489208635, "step": 18920 }, { "epoch": 17.014388489208635, "step": 18920, "torque_loss": 0.19129176437854767 }, { "epoch": 17.02338129496403, "grad_norm": 0.12733995914459229, "learning_rate": 7.327511651121534e-05, "loss": 0.0247, "step": 18930 }, { "action_loss": 0.006288781762123108, "epoch": 17.02338129496403, "step": 18930 }, { "epoch": 17.02338129496403, "step": 18930, "torque_loss": 0.22835950553417206 }, { "epoch": 17.032374100719423, "grad_norm": 0.1848829686641693, "learning_rate": 7.324584346407833e-05, "loss": 0.0279, "step": 18940 }, { "action_loss": 0.008811168372631073, "epoch": 17.032374100719423, "step": 18940 }, { "epoch": 17.032374100719423, "step": 18940, "torque_loss": 0.15470127761363983 }, { "epoch": 17.04136690647482, "grad_norm": 0.24806001782417297, "learning_rate": 7.321656024841962e-05, "loss": 0.0244, "step": 18950 }, { "action_loss": 0.008369199000298977, "epoch": 17.04136690647482, "step": 18950 }, { "epoch": 17.04136690647482, "step": 18950, "torque_loss": 0.1662229299545288 }, { "epoch": 17.050359712230215, "grad_norm": 0.18254354596138, "learning_rate": 7.31872668770487e-05, "loss": 0.0256, "step": 18960 }, { "action_loss": 0.014205140061676502, "epoch": 17.050359712230215, "step": 18960 }, { "epoch": 17.050359712230215, "step": 18960, "torque_loss": 0.1872081607580185 }, { "epoch": 17.059352517985612, "grad_norm": 0.15173116326332092, "learning_rate": 7.315796336277949e-05, "loss": 0.0274, "step": 18970 }, { "action_loss": 0.004219681490212679, "epoch": 17.059352517985612, "step": 18970 }, { "epoch": 17.059352517985612, "step": 18970, "torque_loss": 0.22281216084957123 }, { "epoch": 17.068345323741006, "grad_norm": 0.1139591857790947, "learning_rate": 7.312864971843031e-05, "loss": 0.0306, "step": 18980 }, { "action_loss": 0.007682659197598696, "epoch": 17.068345323741006, "step": 18980 }, { "epoch": 17.068345323741006, "step": 18980, "torque_loss": 0.23462408781051636 }, { "epoch": 17.077338129496404, "grad_norm": 0.22774487733840942, "learning_rate": 7.309932595682398e-05, "loss": 0.0249, "step": 18990 }, { "action_loss": 0.0028612397145479918, "epoch": 17.077338129496404, "step": 18990 }, { "epoch": 17.077338129496404, "step": 18990, "torque_loss": 0.1307472586631775 }, { "epoch": 17.086330935251798, "grad_norm": 0.2528056204319, "learning_rate": 7.30699920907877e-05, "loss": 0.0321, "step": 19000 }, { "action_loss": 0.006969435606151819, "epoch": 17.086330935251798, "step": 19000 }, { "epoch": 17.086330935251798, "step": 19000, "torque_loss": 0.1746336668729782 }, { "epoch": 17.095323741007196, "grad_norm": 0.20991967618465424, "learning_rate": 7.304064813315308e-05, "loss": 0.0276, "step": 19010 }, { "action_loss": 0.01825275830924511, "epoch": 17.095323741007196, "step": 19010 }, { "epoch": 17.095323741007196, "step": 19010, "torque_loss": 0.2142220288515091 }, { "epoch": 17.10431654676259, "grad_norm": 0.210128054022789, "learning_rate": 7.30112940967562e-05, "loss": 0.025, "step": 19020 }, { "action_loss": 0.032472480088472366, "epoch": 17.10431654676259, "step": 19020 }, { "epoch": 17.10431654676259, "step": 19020, "torque_loss": 0.2762405276298523 }, { "epoch": 17.113309352517987, "grad_norm": 0.1509125828742981, "learning_rate": 7.298192999443747e-05, "loss": 0.0313, "step": 19030 }, { "action_loss": 0.004466900136321783, "epoch": 17.113309352517987, "step": 19030 }, { "epoch": 17.113309352517987, "step": 19030, "torque_loss": 0.1621396690607071 }, { "epoch": 17.12230215827338, "grad_norm": 0.14873577654361725, "learning_rate": 7.295255583904179e-05, "loss": 0.0271, "step": 19040 }, { "action_loss": 0.011399631388485432, "epoch": 17.12230215827338, "step": 19040 }, { "epoch": 17.12230215827338, "step": 19040, "torque_loss": 0.1646312028169632 }, { "epoch": 17.131294964028775, "grad_norm": 0.17810940742492676, "learning_rate": 7.29231716434184e-05, "loss": 0.0259, "step": 19050 }, { "action_loss": 0.00659271702170372, "epoch": 17.131294964028775, "step": 19050 }, { "epoch": 17.131294964028775, "step": 19050, "torque_loss": 0.25040972232818604 }, { "epoch": 17.140287769784173, "grad_norm": 0.2764738202095032, "learning_rate": 7.289377742042092e-05, "loss": 0.0278, "step": 19060 }, { "action_loss": 0.004033213946968317, "epoch": 17.140287769784173, "step": 19060 }, { "epoch": 17.140287769784173, "step": 19060, "torque_loss": 0.18174612522125244 }, { "epoch": 17.149280575539567, "grad_norm": 0.1819467693567276, "learning_rate": 7.286437318290741e-05, "loss": 0.0313, "step": 19070 }, { "action_loss": 0.0037011851090937853, "epoch": 17.149280575539567, "step": 19070 }, { "epoch": 17.149280575539567, "step": 19070, "torque_loss": 0.1286700814962387 }, { "epoch": 17.158273381294965, "grad_norm": 0.1500064879655838, "learning_rate": 7.283495894374027e-05, "loss": 0.0244, "step": 19080 }, { "action_loss": 0.025055592879652977, "epoch": 17.158273381294965, "step": 19080 }, { "epoch": 17.158273381294965, "step": 19080, "torque_loss": 0.22115467488765717 }, { "epoch": 17.16726618705036, "grad_norm": 0.2452566772699356, "learning_rate": 7.280553471578629e-05, "loss": 0.0298, "step": 19090 }, { "action_loss": 0.004465942736715078, "epoch": 17.16726618705036, "step": 19090 }, { "epoch": 17.16726618705036, "step": 19090, "torque_loss": 0.1256551295518875 }, { "epoch": 17.176258992805757, "grad_norm": 0.207931786775589, "learning_rate": 7.277610051191663e-05, "loss": 0.0291, "step": 19100 }, { "action_loss": 0.007272310554981232, "epoch": 17.176258992805757, "step": 19100 }, { "epoch": 17.176258992805757, "step": 19100, "torque_loss": 0.2172609120607376 }, { "epoch": 17.18525179856115, "grad_norm": 0.18279758095741272, "learning_rate": 7.27466563450068e-05, "loss": 0.0276, "step": 19110 }, { "action_loss": 0.002924456959590316, "epoch": 17.18525179856115, "step": 19110 }, { "epoch": 17.18525179856115, "step": 19110, "torque_loss": 0.1869306117296219 }, { "epoch": 17.194244604316548, "grad_norm": 0.18412843346595764, "learning_rate": 7.271720222793668e-05, "loss": 0.0306, "step": 19120 }, { "action_loss": 0.010884410701692104, "epoch": 17.194244604316548, "step": 19120 }, { "epoch": 17.194244604316548, "step": 19120, "torque_loss": 0.23292726278305054 }, { "epoch": 17.203237410071942, "grad_norm": 0.19559861719608307, "learning_rate": 7.268773817359054e-05, "loss": 0.0244, "step": 19130 }, { "action_loss": 0.006547515746206045, "epoch": 17.203237410071942, "step": 19130 }, { "epoch": 17.203237410071942, "step": 19130, "torque_loss": 0.18103699386119843 }, { "epoch": 17.21223021582734, "grad_norm": 0.17627805471420288, "learning_rate": 7.265826419485689e-05, "loss": 0.0297, "step": 19140 }, { "action_loss": 0.006307186093181372, "epoch": 17.21223021582734, "step": 19140 }, { "epoch": 17.21223021582734, "step": 19140, "torque_loss": 0.22095464169979095 }, { "epoch": 17.221223021582734, "grad_norm": 0.14431864023208618, "learning_rate": 7.26287803046287e-05, "loss": 0.025, "step": 19150 }, { "action_loss": 0.0027952890377491713, "epoch": 17.221223021582734, "step": 19150 }, { "epoch": 17.221223021582734, "step": 19150, "torque_loss": 0.08681433647871017 }, { "epoch": 17.230215827338128, "grad_norm": 0.14483557641506195, "learning_rate": 7.259928651580322e-05, "loss": 0.0278, "step": 19160 }, { "action_loss": 0.0023236721754074097, "epoch": 17.230215827338128, "step": 19160 }, { "epoch": 17.230215827338128, "step": 19160, "torque_loss": 0.12375151365995407 }, { "epoch": 17.239208633093526, "grad_norm": 0.24076047539710999, "learning_rate": 7.256978284128202e-05, "loss": 0.0248, "step": 19170 }, { "action_loss": 0.005025820806622505, "epoch": 17.239208633093526, "step": 19170 }, { "epoch": 17.239208633093526, "step": 19170, "torque_loss": 0.1550057977437973 }, { "epoch": 17.24820143884892, "grad_norm": 0.12846864759922028, "learning_rate": 7.254026929397101e-05, "loss": 0.0187, "step": 19180 }, { "action_loss": 0.013144913129508495, "epoch": 17.24820143884892, "step": 19180 }, { "epoch": 17.24820143884892, "step": 19180, "torque_loss": 0.1902444213628769 }, { "epoch": 17.257194244604317, "grad_norm": 0.11881018429994583, "learning_rate": 7.251074588678044e-05, "loss": 0.023, "step": 19190 }, { "action_loss": 0.003111265366896987, "epoch": 17.257194244604317, "step": 19190 }, { "epoch": 17.257194244604317, "step": 19190, "torque_loss": 0.18269972503185272 }, { "epoch": 17.26618705035971, "grad_norm": 0.1816643476486206, "learning_rate": 7.248121263262482e-05, "loss": 0.0279, "step": 19200 }, { "action_loss": 0.004374123644083738, "epoch": 17.26618705035971, "step": 19200 }, { "epoch": 17.26618705035971, "step": 19200, "torque_loss": 0.2020426243543625 }, { "epoch": 17.27517985611511, "grad_norm": 0.20439617335796356, "learning_rate": 7.245166954442304e-05, "loss": 0.0294, "step": 19210 }, { "action_loss": 0.007668237667530775, "epoch": 17.27517985611511, "step": 19210 }, { "epoch": 17.27517985611511, "step": 19210, "torque_loss": 0.1769089251756668 }, { "epoch": 17.284172661870503, "grad_norm": 0.18851500749588013, "learning_rate": 7.24221166350982e-05, "loss": 0.0298, "step": 19220 }, { "action_loss": 0.004457330796867609, "epoch": 17.284172661870503, "step": 19220 }, { "epoch": 17.284172661870503, "step": 19220, "torque_loss": 0.16253601014614105 }, { "epoch": 17.2931654676259, "grad_norm": 0.12898537516593933, "learning_rate": 7.239255391757777e-05, "loss": 0.0274, "step": 19230 }, { "action_loss": 0.0063797286711633205, "epoch": 17.2931654676259, "step": 19230 }, { "epoch": 17.2931654676259, "step": 19230, "torque_loss": 0.16626910865306854 }, { "epoch": 17.302158273381295, "grad_norm": 0.13400304317474365, "learning_rate": 7.236298140479351e-05, "loss": 0.0287, "step": 19240 }, { "action_loss": 0.014592297375202179, "epoch": 17.302158273381295, "step": 19240 }, { "epoch": 17.302158273381295, "step": 19240, "torque_loss": 0.2798449695110321 }, { "epoch": 17.31115107913669, "grad_norm": 0.17936840653419495, "learning_rate": 7.233339910968141e-05, "loss": 0.0339, "step": 19250 }, { "action_loss": 0.0047820513136684895, "epoch": 17.31115107913669, "step": 19250 }, { "epoch": 17.31115107913669, "step": 19250, "torque_loss": 0.09975913912057877 }, { "epoch": 17.320143884892087, "grad_norm": 0.23523779213428497, "learning_rate": 7.230380704518178e-05, "loss": 0.0269, "step": 19260 }, { "action_loss": 0.003071452723816037, "epoch": 17.320143884892087, "step": 19260 }, { "epoch": 17.320143884892087, "step": 19260, "torque_loss": 0.13973425328731537 }, { "epoch": 17.32913669064748, "grad_norm": 0.14996105432510376, "learning_rate": 7.227420522423919e-05, "loss": 0.0272, "step": 19270 }, { "action_loss": 0.003758189035579562, "epoch": 17.32913669064748, "step": 19270 }, { "epoch": 17.32913669064748, "step": 19270, "torque_loss": 0.12302418798208237 }, { "epoch": 17.33812949640288, "grad_norm": 0.11268094182014465, "learning_rate": 7.224459365980249e-05, "loss": 0.0212, "step": 19280 }, { "action_loss": 0.01073405984789133, "epoch": 17.33812949640288, "step": 19280 }, { "epoch": 17.33812949640288, "step": 19280, "torque_loss": 0.15786966681480408 }, { "epoch": 17.347122302158272, "grad_norm": 0.15189243853092194, "learning_rate": 7.221497236482479e-05, "loss": 0.0279, "step": 19290 }, { "action_loss": 0.004551757127046585, "epoch": 17.347122302158272, "step": 19290 }, { "epoch": 17.347122302158272, "step": 19290, "torque_loss": 0.1822315901517868 }, { "epoch": 17.35611510791367, "grad_norm": 0.23895305395126343, "learning_rate": 7.218534135226345e-05, "loss": 0.0271, "step": 19300 }, { "action_loss": 0.003391003469005227, "epoch": 17.35611510791367, "step": 19300 }, { "epoch": 17.35611510791367, "step": 19300, "torque_loss": 0.12641367316246033 }, { "epoch": 17.365107913669064, "grad_norm": 0.10930794477462769, "learning_rate": 7.215570063508005e-05, "loss": 0.024, "step": 19310 }, { "action_loss": 0.030795888975262642, "epoch": 17.365107913669064, "step": 19310 }, { "epoch": 17.365107913669064, "step": 19310, "torque_loss": 0.32069820165634155 }, { "epoch": 17.37410071942446, "grad_norm": 0.1517961174249649, "learning_rate": 7.21260502262405e-05, "loss": 0.0282, "step": 19320 }, { "action_loss": 0.007207495626062155, "epoch": 17.37410071942446, "step": 19320 }, { "epoch": 17.37410071942446, "step": 19320, "torque_loss": 0.1860639452934265 }, { "epoch": 17.383093525179856, "grad_norm": 0.1354009211063385, "learning_rate": 7.209639013871487e-05, "loss": 0.0253, "step": 19330 }, { "action_loss": 0.013259525410830975, "epoch": 17.383093525179856, "step": 19330 }, { "epoch": 17.383093525179856, "step": 19330, "torque_loss": 0.23232567310333252 }, { "epoch": 17.392086330935253, "grad_norm": 0.18129022419452667, "learning_rate": 7.206672038547746e-05, "loss": 0.0263, "step": 19340 }, { "action_loss": 0.004748337436467409, "epoch": 17.392086330935253, "step": 19340 }, { "epoch": 17.392086330935253, "step": 19340, "torque_loss": 0.1451169103384018 }, { "epoch": 17.401079136690647, "grad_norm": 0.15022119879722595, "learning_rate": 7.203704097950687e-05, "loss": 0.0243, "step": 19350 }, { "action_loss": 0.03294539824128151, "epoch": 17.401079136690647, "step": 19350 }, { "epoch": 17.401079136690647, "step": 19350, "torque_loss": 0.2516399323940277 }, { "epoch": 17.41007194244604, "grad_norm": 0.16620178520679474, "learning_rate": 7.200735193378587e-05, "loss": 0.0301, "step": 19360 }, { "action_loss": 0.00643984554335475, "epoch": 17.41007194244604, "step": 19360 }, { "epoch": 17.41007194244604, "step": 19360, "torque_loss": 0.15467604994773865 }, { "epoch": 17.41906474820144, "grad_norm": 0.14660100638866425, "learning_rate": 7.197765326130145e-05, "loss": 0.0226, "step": 19370 }, { "action_loss": 0.007062908262014389, "epoch": 17.41906474820144, "step": 19370 }, { "epoch": 17.41906474820144, "step": 19370, "torque_loss": 0.1701626032590866 }, { "epoch": 17.428057553956833, "grad_norm": 0.1616678088903427, "learning_rate": 7.194794497504481e-05, "loss": 0.0246, "step": 19380 }, { "action_loss": 0.0025909633841365576, "epoch": 17.428057553956833, "step": 19380 }, { "epoch": 17.428057553956833, "step": 19380, "torque_loss": 0.09858983755111694 }, { "epoch": 17.43705035971223, "grad_norm": 0.18758323788642883, "learning_rate": 7.19182270880114e-05, "loss": 0.0255, "step": 19390 }, { "action_loss": 0.004313016310334206, "epoch": 17.43705035971223, "step": 19390 }, { "epoch": 17.43705035971223, "step": 19390, "torque_loss": 0.11303204298019409 }, { "epoch": 17.446043165467625, "grad_norm": 0.1511133760213852, "learning_rate": 7.188849961320079e-05, "loss": 0.0226, "step": 19400 }, { "action_loss": 0.006958421785384417, "epoch": 17.446043165467625, "step": 19400 }, { "epoch": 17.446043165467625, "step": 19400, "torque_loss": 0.22536510229110718 }, { "epoch": 17.455035971223023, "grad_norm": 0.10430648177862167, "learning_rate": 7.185876256361681e-05, "loss": 0.0273, "step": 19410 }, { "action_loss": 0.0030553124379366636, "epoch": 17.455035971223023, "step": 19410 }, { "epoch": 17.455035971223023, "step": 19410, "torque_loss": 0.13001614809036255 }, { "epoch": 17.464028776978417, "grad_norm": 0.12355191260576248, "learning_rate": 7.182901595226745e-05, "loss": 0.0295, "step": 19420 }, { "action_loss": 0.019860675558447838, "epoch": 17.464028776978417, "step": 19420 }, { "epoch": 17.464028776978417, "step": 19420, "torque_loss": 0.24365536868572235 }, { "epoch": 17.473021582733814, "grad_norm": 0.2019413262605667, "learning_rate": 7.179925979216491e-05, "loss": 0.0277, "step": 19430 }, { "action_loss": 0.0026305969804525375, "epoch": 17.473021582733814, "step": 19430 }, { "epoch": 17.473021582733814, "step": 19430, "torque_loss": 0.04779420793056488 }, { "epoch": 17.48201438848921, "grad_norm": 0.19627638161182404, "learning_rate": 7.17694940963255e-05, "loss": 0.0248, "step": 19440 }, { "action_loss": 0.0037216851487755775, "epoch": 17.48201438848921, "step": 19440 }, { "epoch": 17.48201438848921, "step": 19440, "torque_loss": 0.16339796781539917 }, { "epoch": 17.491007194244606, "grad_norm": 0.19363991916179657, "learning_rate": 7.17397188777698e-05, "loss": 0.0256, "step": 19450 }, { "action_loss": 0.01447672862559557, "epoch": 17.491007194244606, "step": 19450 }, { "epoch": 17.491007194244606, "step": 19450, "torque_loss": 0.20429527759552002 }, { "epoch": 17.5, "grad_norm": 0.1479211449623108, "learning_rate": 7.170993414952245e-05, "loss": 0.029, "step": 19460 }, { "action_loss": 0.0043165418319404125, "epoch": 17.5, "step": 19460 }, { "epoch": 17.5, "step": 19460, "torque_loss": 0.21941149234771729 }, { "epoch": 17.508992805755394, "grad_norm": 0.12966762483119965, "learning_rate": 7.168013992461233e-05, "loss": 0.0288, "step": 19470 }, { "action_loss": 0.00463486323133111, "epoch": 17.508992805755394, "step": 19470 }, { "epoch": 17.508992805755394, "step": 19470, "torque_loss": 0.16918136179447174 }, { "epoch": 17.51798561151079, "grad_norm": 0.09688174724578857, "learning_rate": 7.165033621607245e-05, "loss": 0.0233, "step": 19480 }, { "action_loss": 0.0031083442736417055, "epoch": 17.51798561151079, "step": 19480 }, { "epoch": 17.51798561151079, "step": 19480, "torque_loss": 0.1532091349363327 }, { "epoch": 17.526978417266186, "grad_norm": 0.24893540143966675, "learning_rate": 7.162052303693995e-05, "loss": 0.0284, "step": 19490 }, { "action_loss": 0.005469114053994417, "epoch": 17.526978417266186, "step": 19490 }, { "epoch": 17.526978417266186, "step": 19490, "torque_loss": 0.18617798388004303 }, { "epoch": 17.535971223021583, "grad_norm": 0.15590089559555054, "learning_rate": 7.159070040025614e-05, "loss": 0.0235, "step": 19500 }, { "action_loss": 0.003484289161860943, "epoch": 17.535971223021583, "step": 19500 }, { "epoch": 17.535971223021583, "step": 19500, "torque_loss": 0.13710562884807587 }, { "epoch": 17.544964028776977, "grad_norm": 0.2231217473745346, "learning_rate": 7.156086831906646e-05, "loss": 0.0252, "step": 19510 }, { "action_loss": 0.008287115953862667, "epoch": 17.544964028776977, "step": 19510 }, { "epoch": 17.544964028776977, "step": 19510, "torque_loss": 0.204880490899086 }, { "epoch": 17.553956834532375, "grad_norm": 0.14768680930137634, "learning_rate": 7.153102680642044e-05, "loss": 0.0255, "step": 19520 }, { "action_loss": 0.007479535415768623, "epoch": 17.553956834532375, "step": 19520 }, { "epoch": 17.553956834532375, "step": 19520, "torque_loss": 0.14911650121212006 }, { "epoch": 17.56294964028777, "grad_norm": 0.15385816991329193, "learning_rate": 7.150117587537182e-05, "loss": 0.0251, "step": 19530 }, { "action_loss": 0.013875964097678661, "epoch": 17.56294964028777, "step": 19530 }, { "epoch": 17.56294964028777, "step": 19530, "torque_loss": 0.2951521575450897 }, { "epoch": 17.571942446043167, "grad_norm": 0.1736459583044052, "learning_rate": 7.147131553897838e-05, "loss": 0.0249, "step": 19540 }, { "action_loss": 0.0060715205036103725, "epoch": 17.571942446043167, "step": 19540 }, { "epoch": 17.571942446043167, "step": 19540, "torque_loss": 0.26593732833862305 }, { "epoch": 17.58093525179856, "grad_norm": 0.16400375962257385, "learning_rate": 7.144144581030205e-05, "loss": 0.0282, "step": 19550 }, { "action_loss": 0.005866901483386755, "epoch": 17.58093525179856, "step": 19550 }, { "epoch": 17.58093525179856, "step": 19550, "torque_loss": 0.13077552616596222 }, { "epoch": 17.58992805755396, "grad_norm": 0.11010793596506119, "learning_rate": 7.141156670240889e-05, "loss": 0.0219, "step": 19560 }, { "action_loss": 0.010278341360390186, "epoch": 17.58992805755396, "step": 19560 }, { "epoch": 17.58992805755396, "step": 19560, "torque_loss": 0.17668354511260986 }, { "epoch": 17.598920863309353, "grad_norm": 0.14604321122169495, "learning_rate": 7.138167822836899e-05, "loss": 0.023, "step": 19570 }, { "action_loss": 0.006113534327596426, "epoch": 17.598920863309353, "step": 19570 }, { "epoch": 17.598920863309353, "step": 19570, "torque_loss": 0.1774229258298874 }, { "epoch": 17.607913669064747, "grad_norm": 0.14306199550628662, "learning_rate": 7.135178040125661e-05, "loss": 0.0303, "step": 19580 }, { "action_loss": 0.003959578927606344, "epoch": 17.607913669064747, "step": 19580 }, { "epoch": 17.607913669064747, "step": 19580, "torque_loss": 0.16014067828655243 }, { "epoch": 17.616906474820144, "grad_norm": 0.165462464094162, "learning_rate": 7.13218732341501e-05, "loss": 0.0215, "step": 19590 }, { "action_loss": 0.004540224093943834, "epoch": 17.616906474820144, "step": 19590 }, { "epoch": 17.616906474820144, "step": 19590, "torque_loss": 0.1878345012664795 }, { "epoch": 17.62589928057554, "grad_norm": 0.13097257912158966, "learning_rate": 7.129195674013181e-05, "loss": 0.0247, "step": 19600 }, { "action_loss": 0.010841414332389832, "epoch": 17.62589928057554, "step": 19600 }, { "epoch": 17.62589928057554, "step": 19600, "torque_loss": 0.24670064449310303 }, { "epoch": 17.634892086330936, "grad_norm": 0.1583622246980667, "learning_rate": 7.126203093228829e-05, "loss": 0.0275, "step": 19610 }, { "action_loss": 0.0044606695882976055, "epoch": 17.634892086330936, "step": 19610 }, { "epoch": 17.634892086330936, "step": 19610, "torque_loss": 0.1811656504869461 }, { "epoch": 17.64388489208633, "grad_norm": 0.20288777351379395, "learning_rate": 7.123209582371006e-05, "loss": 0.0281, "step": 19620 }, { "action_loss": 0.0023536996450275183, "epoch": 17.64388489208633, "step": 19620 }, { "epoch": 17.64388489208633, "step": 19620, "torque_loss": 0.09810514003038406 }, { "epoch": 17.652877697841728, "grad_norm": 0.13700339198112488, "learning_rate": 7.120215142749179e-05, "loss": 0.0199, "step": 19630 }, { "action_loss": 0.004915942437946796, "epoch": 17.652877697841728, "step": 19630 }, { "epoch": 17.652877697841728, "step": 19630, "torque_loss": 0.1895238757133484 }, { "epoch": 17.66187050359712, "grad_norm": 0.2296268790960312, "learning_rate": 7.117219775673218e-05, "loss": 0.0236, "step": 19640 }, { "action_loss": 0.003147497773170471, "epoch": 17.66187050359712, "step": 19640 }, { "epoch": 17.66187050359712, "step": 19640, "torque_loss": 0.1538042575120926 }, { "epoch": 17.67086330935252, "grad_norm": 0.16519813239574432, "learning_rate": 7.114223482453394e-05, "loss": 0.0341, "step": 19650 }, { "action_loss": 0.009263877756893635, "epoch": 17.67086330935252, "step": 19650 }, { "epoch": 17.67086330935252, "step": 19650, "torque_loss": 0.25253501534461975 }, { "epoch": 17.679856115107913, "grad_norm": 0.16854725778102875, "learning_rate": 7.11122626440039e-05, "loss": 0.025, "step": 19660 }, { "action_loss": 0.015910087153315544, "epoch": 17.679856115107913, "step": 19660 }, { "epoch": 17.679856115107913, "step": 19660, "torque_loss": 0.1648203730583191 }, { "epoch": 17.68884892086331, "grad_norm": 0.15290264785289764, "learning_rate": 7.10822812282529e-05, "loss": 0.0247, "step": 19670 }, { "action_loss": 0.006588177755475044, "epoch": 17.68884892086331, "step": 19670 }, { "epoch": 17.68884892086331, "step": 19670, "torque_loss": 0.17125976085662842 }, { "epoch": 17.697841726618705, "grad_norm": 0.2031092494726181, "learning_rate": 7.105229059039586e-05, "loss": 0.0305, "step": 19680 }, { "action_loss": 0.002964943880215287, "epoch": 17.697841726618705, "step": 19680 }, { "epoch": 17.697841726618705, "step": 19680, "torque_loss": 0.12727221846580505 }, { "epoch": 17.7068345323741, "grad_norm": 0.2028103470802307, "learning_rate": 7.102229074355168e-05, "loss": 0.027, "step": 19690 }, { "action_loss": 0.00854964554309845, "epoch": 17.7068345323741, "step": 19690 }, { "epoch": 17.7068345323741, "step": 19690, "torque_loss": 0.2237056940793991 }, { "epoch": 17.715827338129497, "grad_norm": 0.17994579672813416, "learning_rate": 7.099228170084332e-05, "loss": 0.0332, "step": 19700 }, { "action_loss": 0.004732249304652214, "epoch": 17.715827338129497, "step": 19700 }, { "epoch": 17.715827338129497, "step": 19700, "torque_loss": 0.1732974499464035 }, { "epoch": 17.72482014388489, "grad_norm": 0.18711218237876892, "learning_rate": 7.096226347539771e-05, "loss": 0.0299, "step": 19710 }, { "action_loss": 0.0051733399741351604, "epoch": 17.72482014388489, "step": 19710 }, { "epoch": 17.72482014388489, "step": 19710, "torque_loss": 0.1657908409833908 }, { "epoch": 17.73381294964029, "grad_norm": 0.1905994415283203, "learning_rate": 7.093223608034592e-05, "loss": 0.0269, "step": 19720 }, { "action_loss": 0.0047072418965399265, "epoch": 17.73381294964029, "step": 19720 }, { "epoch": 17.73381294964029, "step": 19720, "torque_loss": 0.1661970615386963 }, { "epoch": 17.742805755395683, "grad_norm": 0.1807229369878769, "learning_rate": 7.090219952882291e-05, "loss": 0.0227, "step": 19730 }, { "action_loss": 0.0028475208673626184, "epoch": 17.742805755395683, "step": 19730 }, { "epoch": 17.742805755395683, "step": 19730, "torque_loss": 0.14144782721996307 }, { "epoch": 17.75179856115108, "grad_norm": 0.22854049503803253, "learning_rate": 7.087215383396768e-05, "loss": 0.0234, "step": 19740 }, { "action_loss": 0.011112756095826626, "epoch": 17.75179856115108, "step": 19740 }, { "epoch": 17.75179856115108, "step": 19740, "torque_loss": 0.2363177090883255 }, { "epoch": 17.760791366906474, "grad_norm": 0.14479106664657593, "learning_rate": 7.084209900892325e-05, "loss": 0.0287, "step": 19750 }, { "action_loss": 0.00855351984500885, "epoch": 17.760791366906474, "step": 19750 }, { "epoch": 17.760791366906474, "step": 19750, "torque_loss": 0.13060736656188965 }, { "epoch": 17.769784172661872, "grad_norm": 0.16195783019065857, "learning_rate": 7.081203506683663e-05, "loss": 0.0233, "step": 19760 }, { "action_loss": 0.011315964162349701, "epoch": 17.769784172661872, "step": 19760 }, { "epoch": 17.769784172661872, "step": 19760, "torque_loss": 0.22485513985157013 }, { "epoch": 17.778776978417266, "grad_norm": 0.14783456921577454, "learning_rate": 7.078196202085879e-05, "loss": 0.0256, "step": 19770 }, { "action_loss": 0.011994617991149426, "epoch": 17.778776978417266, "step": 19770 }, { "epoch": 17.778776978417266, "step": 19770, "torque_loss": 0.18349285423755646 }, { "epoch": 17.78776978417266, "grad_norm": 0.24817322194576263, "learning_rate": 7.075187988414473e-05, "loss": 0.0337, "step": 19780 }, { "action_loss": 0.005460911896079779, "epoch": 17.78776978417266, "step": 19780 }, { "epoch": 17.78776978417266, "step": 19780, "torque_loss": 0.1778661459684372 }, { "epoch": 17.796762589928058, "grad_norm": 0.25188109278678894, "learning_rate": 7.072178866985334e-05, "loss": 0.0233, "step": 19790 }, { "action_loss": 0.02437114529311657, "epoch": 17.796762589928058, "step": 19790 }, { "epoch": 17.796762589928058, "step": 19790, "torque_loss": 0.24691443145275116 }, { "epoch": 17.805755395683452, "grad_norm": 0.15192408859729767, "learning_rate": 7.069168839114764e-05, "loss": 0.0299, "step": 19800 }, { "action_loss": 0.003710186807438731, "epoch": 17.805755395683452, "step": 19800 }, { "epoch": 17.805755395683452, "step": 19800, "torque_loss": 0.16596601903438568 }, { "epoch": 17.81474820143885, "grad_norm": 0.11166971176862717, "learning_rate": 7.066157906119441e-05, "loss": 0.0238, "step": 19810 }, { "action_loss": 0.0032899111974984407, "epoch": 17.81474820143885, "step": 19810 }, { "epoch": 17.81474820143885, "step": 19810, "torque_loss": 0.23374319076538086 }, { "epoch": 17.823741007194243, "grad_norm": 0.12746866047382355, "learning_rate": 7.063146069316456e-05, "loss": 0.0316, "step": 19820 }, { "action_loss": 0.00594719173386693, "epoch": 17.823741007194243, "step": 19820 }, { "epoch": 17.823741007194243, "step": 19820, "torque_loss": 0.21670882403850555 }, { "epoch": 17.83273381294964, "grad_norm": 0.19465947151184082, "learning_rate": 7.060133330023287e-05, "loss": 0.0284, "step": 19830 }, { "action_loss": 0.006766466423869133, "epoch": 17.83273381294964, "step": 19830 }, { "epoch": 17.83273381294964, "step": 19830, "torque_loss": 0.13354714214801788 }, { "epoch": 17.841726618705035, "grad_norm": 0.15424075722694397, "learning_rate": 7.057119689557807e-05, "loss": 0.0224, "step": 19840 }, { "action_loss": 0.019078565761446953, "epoch": 17.841726618705035, "step": 19840 }, { "epoch": 17.841726618705035, "step": 19840, "torque_loss": 0.2733074724674225 }, { "epoch": 17.850719424460433, "grad_norm": 0.15521247684955597, "learning_rate": 7.054105149238287e-05, "loss": 0.0223, "step": 19850 }, { "action_loss": 0.011305666528642178, "epoch": 17.850719424460433, "step": 19850 }, { "epoch": 17.850719424460433, "step": 19850, "torque_loss": 0.2748826742172241 }, { "epoch": 17.859712230215827, "grad_norm": 0.15741735696792603, "learning_rate": 7.051089710383387e-05, "loss": 0.023, "step": 19860 }, { "action_loss": 0.005914783570915461, "epoch": 17.859712230215827, "step": 19860 }, { "epoch": 17.859712230215827, "step": 19860, "torque_loss": 0.15348537266254425 }, { "epoch": 17.868705035971225, "grad_norm": 0.1761241853237152, "learning_rate": 7.048073374312166e-05, "loss": 0.0331, "step": 19870 }, { "action_loss": 0.0142129585146904, "epoch": 17.868705035971225, "step": 19870 }, { "epoch": 17.868705035971225, "step": 19870, "torque_loss": 0.23035310208797455 }, { "epoch": 17.87769784172662, "grad_norm": 0.1950933188199997, "learning_rate": 7.045056142344067e-05, "loss": 0.0299, "step": 19880 }, { "action_loss": 0.009975403547286987, "epoch": 17.87769784172662, "step": 19880 }, { "epoch": 17.87769784172662, "step": 19880, "torque_loss": 0.23752684891223907 }, { "epoch": 17.886690647482013, "grad_norm": 0.13616009056568146, "learning_rate": 7.042038015798934e-05, "loss": 0.0258, "step": 19890 }, { "action_loss": 0.010083253495395184, "epoch": 17.886690647482013, "step": 19890 }, { "epoch": 17.886690647482013, "step": 19890, "torque_loss": 0.19111263751983643 }, { "epoch": 17.89568345323741, "grad_norm": 0.1591976135969162, "learning_rate": 7.039018995996995e-05, "loss": 0.0297, "step": 19900 }, { "action_loss": 0.02357087843120098, "epoch": 17.89568345323741, "step": 19900 }, { "epoch": 17.89568345323741, "step": 19900, "torque_loss": 0.23316161334514618 }, { "epoch": 17.904676258992804, "grad_norm": 0.22128549218177795, "learning_rate": 7.035999084258876e-05, "loss": 0.0288, "step": 19910 }, { "action_loss": 0.012417751364409924, "epoch": 17.904676258992804, "step": 19910 }, { "epoch": 17.904676258992804, "step": 19910, "torque_loss": 0.24676163494586945 }, { "epoch": 17.913669064748202, "grad_norm": 0.2453598827123642, "learning_rate": 7.032978281905583e-05, "loss": 0.0253, "step": 19920 }, { "action_loss": 0.01522892341017723, "epoch": 17.913669064748202, "step": 19920 }, { "epoch": 17.913669064748202, "step": 19920, "torque_loss": 0.1805073469877243 }, { "epoch": 17.922661870503596, "grad_norm": 0.19113744795322418, "learning_rate": 7.029956590258522e-05, "loss": 0.0235, "step": 19930 }, { "action_loss": 0.005699444562196732, "epoch": 17.922661870503596, "step": 19930 }, { "epoch": 17.922661870503596, "step": 19930, "torque_loss": 0.12359610944986343 }, { "epoch": 17.931654676258994, "grad_norm": 0.1865319311618805, "learning_rate": 7.026934010639483e-05, "loss": 0.0265, "step": 19940 }, { "action_loss": 0.008674785494804382, "epoch": 17.931654676258994, "step": 19940 }, { "epoch": 17.931654676258994, "step": 19940, "torque_loss": 0.20358265936374664 }, { "epoch": 17.940647482014388, "grad_norm": 0.20492039620876312, "learning_rate": 7.023910544370645e-05, "loss": 0.0229, "step": 19950 }, { "action_loss": 0.0035684898030012846, "epoch": 17.940647482014388, "step": 19950 }, { "epoch": 17.940647482014388, "step": 19950, "torque_loss": 0.156870499253273 }, { "epoch": 17.949640287769785, "grad_norm": 0.19884514808654785, "learning_rate": 7.020886192774573e-05, "loss": 0.0274, "step": 19960 }, { "action_loss": 0.007199096027761698, "epoch": 17.949640287769785, "step": 19960 }, { "epoch": 17.949640287769785, "step": 19960, "torque_loss": 0.2328612357378006 }, { "epoch": 17.95863309352518, "grad_norm": 0.11190813034772873, "learning_rate": 7.017860957174222e-05, "loss": 0.0277, "step": 19970 }, { "action_loss": 0.003324056277051568, "epoch": 17.95863309352518, "step": 19970 }, { "epoch": 17.95863309352518, "step": 19970, "torque_loss": 0.1957581490278244 }, { "epoch": 17.967625899280577, "grad_norm": 0.21633417904376984, "learning_rate": 7.014834838892934e-05, "loss": 0.0256, "step": 19980 }, { "action_loss": 0.006781875621527433, "epoch": 17.967625899280577, "step": 19980 }, { "epoch": 17.967625899280577, "step": 19980, "torque_loss": 0.16320152580738068 }, { "epoch": 17.97661870503597, "grad_norm": 0.13900385797023773, "learning_rate": 7.011807839254435e-05, "loss": 0.0269, "step": 19990 }, { "action_loss": 0.01020299643278122, "epoch": 17.97661870503597, "step": 19990 }, { "epoch": 17.97661870503597, "step": 19990, "torque_loss": 0.21523261070251465 }, { "epoch": 17.985611510791365, "grad_norm": 0.21153445541858673, "learning_rate": 7.008779959582837e-05, "loss": 0.0284, "step": 20000 }, { "action_loss": 0.0035449285060167313, "epoch": 17.985611510791365, "step": 20000 }, { "epoch": 17.985611510791365, "step": 20000, "torque_loss": 0.1367781162261963 }, { "epoch": 17.994604316546763, "grad_norm": 0.14305150508880615, "learning_rate": 7.005751201202639e-05, "loss": 0.0271, "step": 20010 }, { "action_loss": 0.004365788772702217, "epoch": 17.994604316546763, "step": 20010 }, { "epoch": 17.994604316546763, "step": 20010, "torque_loss": 0.10742858797311783 }, { "epoch": 18.003597122302157, "grad_norm": 0.20541517436504364, "learning_rate": 7.002721565438724e-05, "loss": 0.0274, "step": 20020 }, { "action_loss": 0.01080611627548933, "epoch": 18.003597122302157, "step": 20020 }, { "epoch": 18.003597122302157, "step": 20020, "torque_loss": 0.20701098442077637 }, { "epoch": 18.012589928057555, "grad_norm": 0.1827513426542282, "learning_rate": 6.999691053616353e-05, "loss": 0.0286, "step": 20030 }, { "action_loss": 0.00931230653077364, "epoch": 18.012589928057555, "step": 20030 }, { "epoch": 18.012589928057555, "step": 20030, "torque_loss": 0.2308022826910019 }, { "epoch": 18.02158273381295, "grad_norm": 0.14823488891124725, "learning_rate": 6.996659667061178e-05, "loss": 0.0279, "step": 20040 }, { "action_loss": 0.004389928188174963, "epoch": 18.02158273381295, "step": 20040 }, { "epoch": 18.02158273381295, "step": 20040, "torque_loss": 0.11925303936004639 }, { "epoch": 18.030575539568346, "grad_norm": 0.1488521844148636, "learning_rate": 6.993627407099229e-05, "loss": 0.0258, "step": 20050 }, { "action_loss": 0.003335254266858101, "epoch": 18.030575539568346, "step": 20050 }, { "epoch": 18.030575539568346, "step": 20050, "torque_loss": 0.16114366054534912 }, { "epoch": 18.03956834532374, "grad_norm": 0.20224657654762268, "learning_rate": 6.990594275056921e-05, "loss": 0.0242, "step": 20060 }, { "action_loss": 0.003977966029196978, "epoch": 18.03956834532374, "step": 20060 }, { "epoch": 18.03956834532374, "step": 20060, "torque_loss": 0.14079934358596802 }, { "epoch": 18.048561151079138, "grad_norm": 0.23600871860980988, "learning_rate": 6.987560272261049e-05, "loss": 0.0294, "step": 20070 }, { "action_loss": 0.013994085602462292, "epoch": 18.048561151079138, "step": 20070 }, { "epoch": 18.048561151079138, "step": 20070, "torque_loss": 0.18048854172229767 }, { "epoch": 18.057553956834532, "grad_norm": 0.20483630895614624, "learning_rate": 6.984525400038788e-05, "loss": 0.0324, "step": 20080 }, { "action_loss": 0.003217655001208186, "epoch": 18.057553956834532, "step": 20080 }, { "epoch": 18.057553956834532, "step": 20080, "torque_loss": 0.14331412315368652 }, { "epoch": 18.06654676258993, "grad_norm": 0.1550559103488922, "learning_rate": 6.981489659717693e-05, "loss": 0.0243, "step": 20090 }, { "action_loss": 0.02648422308266163, "epoch": 18.06654676258993, "step": 20090 }, { "epoch": 18.06654676258993, "step": 20090, "torque_loss": 0.3163195550441742 }, { "epoch": 18.075539568345324, "grad_norm": 0.16358892619609833, "learning_rate": 6.978453052625703e-05, "loss": 0.0245, "step": 20100 }, { "action_loss": 0.010220148600637913, "epoch": 18.075539568345324, "step": 20100 }, { "epoch": 18.075539568345324, "step": 20100, "torque_loss": 0.23264992237091064 }, { "epoch": 18.084532374100718, "grad_norm": 0.13628998398780823, "learning_rate": 6.975415580091131e-05, "loss": 0.0303, "step": 20110 }, { "action_loss": 0.004250491503626108, "epoch": 18.084532374100718, "step": 20110 }, { "epoch": 18.084532374100718, "step": 20110, "torque_loss": 0.1684853434562683 }, { "epoch": 18.093525179856115, "grad_norm": 0.14744457602500916, "learning_rate": 6.972377243442672e-05, "loss": 0.0248, "step": 20120 }, { "action_loss": 0.0036731779109686613, "epoch": 18.093525179856115, "step": 20120 }, { "epoch": 18.093525179856115, "step": 20120, "torque_loss": 0.11406629532575607 }, { "epoch": 18.10251798561151, "grad_norm": 0.1498994529247284, "learning_rate": 6.969338044009396e-05, "loss": 0.0235, "step": 20130 }, { "action_loss": 0.05747652426362038, "epoch": 18.10251798561151, "step": 20130 }, { "epoch": 18.10251798561151, "step": 20130, "torque_loss": 0.31429341435432434 }, { "epoch": 18.111510791366907, "grad_norm": 0.12178940325975418, "learning_rate": 6.966297983120753e-05, "loss": 0.0266, "step": 20140 }, { "action_loss": 0.012614336796104908, "epoch": 18.111510791366907, "step": 20140 }, { "epoch": 18.111510791366907, "step": 20140, "torque_loss": 0.20242293179035187 }, { "epoch": 18.1205035971223, "grad_norm": 0.14095020294189453, "learning_rate": 6.963257062106571e-05, "loss": 0.0244, "step": 20150 }, { "action_loss": 0.005033973138779402, "epoch": 18.1205035971223, "step": 20150 }, { "epoch": 18.1205035971223, "step": 20150, "torque_loss": 0.18483151495456696 }, { "epoch": 18.1294964028777, "grad_norm": 0.10566337406635284, "learning_rate": 6.960215282297048e-05, "loss": 0.0239, "step": 20160 }, { "action_loss": 0.00978111568838358, "epoch": 18.1294964028777, "step": 20160 }, { "epoch": 18.1294964028777, "step": 20160, "torque_loss": 0.21172916889190674 }, { "epoch": 18.138489208633093, "grad_norm": 0.15124592185020447, "learning_rate": 6.957172645022765e-05, "loss": 0.0232, "step": 20170 }, { "action_loss": 0.023398032411932945, "epoch": 18.138489208633093, "step": 20170 }, { "epoch": 18.138489208633093, "step": 20170, "torque_loss": 0.19242149591445923 }, { "epoch": 18.14748201438849, "grad_norm": 0.13650822639465332, "learning_rate": 6.954129151614673e-05, "loss": 0.0236, "step": 20180 }, { "action_loss": 0.004834024701267481, "epoch": 18.14748201438849, "step": 20180 }, { "epoch": 18.14748201438849, "step": 20180, "torque_loss": 0.11479083448648453 }, { "epoch": 18.156474820143885, "grad_norm": 0.2662452459335327, "learning_rate": 6.9510848034041e-05, "loss": 0.025, "step": 20190 }, { "action_loss": 0.003763460787013173, "epoch": 18.156474820143885, "step": 20190 }, { "epoch": 18.156474820143885, "step": 20190, "torque_loss": 0.14653295278549194 }, { "epoch": 18.165467625899282, "grad_norm": 0.21676287055015564, "learning_rate": 6.94803960172275e-05, "loss": 0.0237, "step": 20200 }, { "action_loss": 0.0024849448818713427, "epoch": 18.165467625899282, "step": 20200 }, { "epoch": 18.165467625899282, "step": 20200, "torque_loss": 0.16826653480529785 }, { "epoch": 18.174460431654676, "grad_norm": 0.2394707202911377, "learning_rate": 6.944993547902692e-05, "loss": 0.0257, "step": 20210 }, { "action_loss": 0.016539348289370537, "epoch": 18.174460431654676, "step": 20210 }, { "epoch": 18.174460431654676, "step": 20210, "torque_loss": 0.22499966621398926 }, { "epoch": 18.18345323741007, "grad_norm": 0.14513570070266724, "learning_rate": 6.941946643276379e-05, "loss": 0.0281, "step": 20220 }, { "action_loss": 0.014591419138014317, "epoch": 18.18345323741007, "step": 20220 }, { "epoch": 18.18345323741007, "step": 20220, "torque_loss": 0.15907178819179535 }, { "epoch": 18.192446043165468, "grad_norm": 0.18101316690444946, "learning_rate": 6.938898889176626e-05, "loss": 0.0236, "step": 20230 }, { "action_loss": 0.010054406709969044, "epoch": 18.192446043165468, "step": 20230 }, { "epoch": 18.192446043165468, "step": 20230, "torque_loss": 0.2012980431318283 }, { "epoch": 18.201438848920862, "grad_norm": 0.2181405872106552, "learning_rate": 6.935850286936627e-05, "loss": 0.0268, "step": 20240 }, { "action_loss": 0.004881935194134712, "epoch": 18.201438848920862, "step": 20240 }, { "epoch": 18.201438848920862, "step": 20240, "torque_loss": 0.1282065361738205 }, { "epoch": 18.21043165467626, "grad_norm": 0.1784363090991974, "learning_rate": 6.932800837889943e-05, "loss": 0.0236, "step": 20250 }, { "action_loss": 0.007745554205030203, "epoch": 18.21043165467626, "step": 20250 }, { "epoch": 18.21043165467626, "step": 20250, "torque_loss": 0.20371614396572113 }, { "epoch": 18.219424460431654, "grad_norm": 0.10905595868825912, "learning_rate": 6.92975054337051e-05, "loss": 0.0253, "step": 20260 }, { "action_loss": 0.009665998630225658, "epoch": 18.219424460431654, "step": 20260 }, { "epoch": 18.219424460431654, "step": 20260, "torque_loss": 0.2336927205324173 }, { "epoch": 18.22841726618705, "grad_norm": 0.14442507922649384, "learning_rate": 6.926699404712627e-05, "loss": 0.0225, "step": 20270 }, { "action_loss": 0.002028797287493944, "epoch": 18.22841726618705, "step": 20270 }, { "epoch": 18.22841726618705, "step": 20270, "torque_loss": 0.155457004904747 }, { "epoch": 18.237410071942445, "grad_norm": 0.1639198660850525, "learning_rate": 6.923647423250967e-05, "loss": 0.0262, "step": 20280 }, { "action_loss": 0.004330373834818602, "epoch": 18.237410071942445, "step": 20280 }, { "epoch": 18.237410071942445, "step": 20280, "torque_loss": 0.15789315104484558 }, { "epoch": 18.246402877697843, "grad_norm": 0.14901599287986755, "learning_rate": 6.920594600320567e-05, "loss": 0.0223, "step": 20290 }, { "action_loss": 0.01046821940690279, "epoch": 18.246402877697843, "step": 20290 }, { "epoch": 18.246402877697843, "step": 20290, "torque_loss": 0.21615850925445557 }, { "epoch": 18.255395683453237, "grad_norm": 0.17891862988471985, "learning_rate": 6.91754093725684e-05, "loss": 0.0248, "step": 20300 }, { "action_loss": 0.008543655276298523, "epoch": 18.255395683453237, "step": 20300 }, { "epoch": 18.255395683453237, "step": 20300, "torque_loss": 0.20403039455413818 }, { "epoch": 18.264388489208635, "grad_norm": 0.135762020945549, "learning_rate": 6.914486435395561e-05, "loss": 0.023, "step": 20310 }, { "action_loss": 0.003191316733136773, "epoch": 18.264388489208635, "step": 20310 }, { "epoch": 18.264388489208635, "step": 20310, "torque_loss": 0.1331763118505478 }, { "epoch": 18.27338129496403, "grad_norm": 0.12252966314554214, "learning_rate": 6.911431096072871e-05, "loss": 0.0234, "step": 20320 }, { "action_loss": 0.012102216482162476, "epoch": 18.27338129496403, "step": 20320 }, { "epoch": 18.27338129496403, "step": 20320, "torque_loss": 0.21017885208129883 }, { "epoch": 18.282374100719423, "grad_norm": 0.2100948989391327, "learning_rate": 6.90837492062528e-05, "loss": 0.0268, "step": 20330 }, { "action_loss": 0.006849188357591629, "epoch": 18.282374100719423, "step": 20330 }, { "epoch": 18.282374100719423, "step": 20330, "torque_loss": 0.20555461943149567 }, { "epoch": 18.29136690647482, "grad_norm": 0.14799924194812775, "learning_rate": 6.905317910389664e-05, "loss": 0.0234, "step": 20340 }, { "action_loss": 0.006437352392822504, "epoch": 18.29136690647482, "step": 20340 }, { "epoch": 18.29136690647482, "step": 20340, "torque_loss": 0.17774450778961182 }, { "epoch": 18.300359712230215, "grad_norm": 0.13645288348197937, "learning_rate": 6.902260066703262e-05, "loss": 0.0237, "step": 20350 }, { "action_loss": 0.006203215569257736, "epoch": 18.300359712230215, "step": 20350 }, { "epoch": 18.300359712230215, "step": 20350, "torque_loss": 0.20327939093112946 }, { "epoch": 18.309352517985612, "grad_norm": 0.13451671600341797, "learning_rate": 6.89920139090368e-05, "loss": 0.025, "step": 20360 }, { "action_loss": 0.021598555147647858, "epoch": 18.309352517985612, "step": 20360 }, { "epoch": 18.309352517985612, "step": 20360, "torque_loss": 0.2582719326019287 }, { "epoch": 18.318345323741006, "grad_norm": 0.15358704328536987, "learning_rate": 6.896141884328886e-05, "loss": 0.026, "step": 20370 }, { "action_loss": 0.012873257510364056, "epoch": 18.318345323741006, "step": 20370 }, { "epoch": 18.318345323741006, "step": 20370, "torque_loss": 0.20089779794216156 }, { "epoch": 18.327338129496404, "grad_norm": 0.08816166967153549, "learning_rate": 6.893081548317211e-05, "loss": 0.0286, "step": 20380 }, { "action_loss": 0.010433492250740528, "epoch": 18.327338129496404, "step": 20380 }, { "epoch": 18.327338129496404, "step": 20380, "torque_loss": 0.21136574447155 }, { "epoch": 18.336330935251798, "grad_norm": 0.21928486227989197, "learning_rate": 6.890020384207353e-05, "loss": 0.0279, "step": 20390 }, { "action_loss": 0.019586702808737755, "epoch": 18.336330935251798, "step": 20390 }, { "epoch": 18.336330935251798, "step": 20390, "torque_loss": 0.14504201710224152 }, { "epoch": 18.345323741007196, "grad_norm": 0.1442275196313858, "learning_rate": 6.886958393338366e-05, "loss": 0.0244, "step": 20400 }, { "action_loss": 0.019323796033859253, "epoch": 18.345323741007196, "step": 20400 }, { "epoch": 18.345323741007196, "step": 20400, "torque_loss": 0.23294289410114288 }, { "epoch": 18.35431654676259, "grad_norm": 0.11972532421350479, "learning_rate": 6.883895577049668e-05, "loss": 0.0284, "step": 20410 }, { "action_loss": 0.004034596029669046, "epoch": 18.35431654676259, "step": 20410 }, { "epoch": 18.35431654676259, "step": 20410, "torque_loss": 0.1834542602300644 }, { "epoch": 18.363309352517987, "grad_norm": 0.12932534515857697, "learning_rate": 6.880831936681045e-05, "loss": 0.0265, "step": 20420 }, { "action_loss": 0.006005550269037485, "epoch": 18.363309352517987, "step": 20420 }, { "epoch": 18.363309352517987, "step": 20420, "torque_loss": 0.16225384175777435 }, { "epoch": 18.37230215827338, "grad_norm": 0.13535460829734802, "learning_rate": 6.87776747357263e-05, "loss": 0.0222, "step": 20430 }, { "action_loss": 0.025806432589888573, "epoch": 18.37230215827338, "step": 20430 }, { "epoch": 18.37230215827338, "step": 20430, "torque_loss": 0.2429707646369934 }, { "epoch": 18.381294964028775, "grad_norm": 0.16648931801319122, "learning_rate": 6.874702189064927e-05, "loss": 0.0276, "step": 20440 }, { "action_loss": 0.018279964104294777, "epoch": 18.381294964028775, "step": 20440 }, { "epoch": 18.381294964028775, "step": 20440, "torque_loss": 0.23844511806964874 }, { "epoch": 18.390287769784173, "grad_norm": 0.131710946559906, "learning_rate": 6.871636084498796e-05, "loss": 0.0278, "step": 20450 }, { "action_loss": 0.011895246803760529, "epoch": 18.390287769784173, "step": 20450 }, { "epoch": 18.390287769784173, "step": 20450, "torque_loss": 0.21047978103160858 }, { "epoch": 18.399280575539567, "grad_norm": 0.19896382093429565, "learning_rate": 6.868569161215453e-05, "loss": 0.0253, "step": 20460 }, { "action_loss": 0.004945473279803991, "epoch": 18.399280575539567, "step": 20460 }, { "epoch": 18.399280575539567, "step": 20460, "torque_loss": 0.16239582002162933 }, { "epoch": 18.408273381294965, "grad_norm": 0.11227654665708542, "learning_rate": 6.865501420556477e-05, "loss": 0.0226, "step": 20470 }, { "action_loss": 0.007866100408136845, "epoch": 18.408273381294965, "step": 20470 }, { "epoch": 18.408273381294965, "step": 20470, "torque_loss": 0.2618071734905243 }, { "epoch": 18.41726618705036, "grad_norm": 0.15741026401519775, "learning_rate": 6.862432863863802e-05, "loss": 0.0255, "step": 20480 }, { "action_loss": 0.005110687110573053, "epoch": 18.41726618705036, "step": 20480 }, { "epoch": 18.41726618705036, "step": 20480, "torque_loss": 0.16580964624881744 }, { "epoch": 18.426258992805757, "grad_norm": 0.101328045129776, "learning_rate": 6.859363492479716e-05, "loss": 0.0251, "step": 20490 }, { "action_loss": 0.002448503626510501, "epoch": 18.426258992805757, "step": 20490 }, { "epoch": 18.426258992805757, "step": 20490, "torque_loss": 0.13670679926872253 }, { "epoch": 18.43525179856115, "grad_norm": 0.21810545027256012, "learning_rate": 6.856293307746868e-05, "loss": 0.0234, "step": 20500 }, { "action_loss": 0.002453358145430684, "epoch": 18.43525179856115, "step": 20500 }, { "epoch": 18.43525179856115, "step": 20500, "torque_loss": 0.177724227309227 }, { "epoch": 18.444244604316548, "grad_norm": 0.171026811003685, "learning_rate": 6.853222311008263e-05, "loss": 0.0204, "step": 20510 }, { "action_loss": 0.003674576058983803, "epoch": 18.444244604316548, "step": 20510 }, { "epoch": 18.444244604316548, "step": 20510, "torque_loss": 0.17298376560211182 }, { "epoch": 18.453237410071942, "grad_norm": 0.19983065128326416, "learning_rate": 6.850150503607256e-05, "loss": 0.0316, "step": 20520 }, { "action_loss": 0.03135925158858299, "epoch": 18.453237410071942, "step": 20520 }, { "epoch": 18.453237410071942, "step": 20520, "torque_loss": 0.20285719633102417 }, { "epoch": 18.46223021582734, "grad_norm": 0.14997811615467072, "learning_rate": 6.847077886887562e-05, "loss": 0.0264, "step": 20530 }, { "action_loss": 0.005747733637690544, "epoch": 18.46223021582734, "step": 20530 }, { "epoch": 18.46223021582734, "step": 20530, "torque_loss": 0.1144978404045105 }, { "epoch": 18.471223021582734, "grad_norm": 0.16265207529067993, "learning_rate": 6.844004462193248e-05, "loss": 0.0269, "step": 20540 }, { "action_loss": 0.012616773135960102, "epoch": 18.471223021582734, "step": 20540 }, { "epoch": 18.471223021582734, "step": 20540, "torque_loss": 0.2559221684932709 }, { "epoch": 18.480215827338128, "grad_norm": 0.21908777952194214, "learning_rate": 6.840930230868734e-05, "loss": 0.0334, "step": 20550 }, { "action_loss": 0.003984017297625542, "epoch": 18.480215827338128, "step": 20550 }, { "epoch": 18.480215827338128, "step": 20550, "torque_loss": 0.15024133026599884 }, { "epoch": 18.489208633093526, "grad_norm": 0.2986522614955902, "learning_rate": 6.837855194258791e-05, "loss": 0.0289, "step": 20560 }, { "action_loss": 0.016637174412608147, "epoch": 18.489208633093526, "step": 20560 }, { "epoch": 18.489208633093526, "step": 20560, "torque_loss": 0.19325228035449982 }, { "epoch": 18.49820143884892, "grad_norm": 0.2299172431230545, "learning_rate": 6.834779353708548e-05, "loss": 0.0311, "step": 20570 }, { "action_loss": 0.0035855385940521955, "epoch": 18.49820143884892, "step": 20570 }, { "epoch": 18.49820143884892, "step": 20570, "torque_loss": 0.1501588672399521 }, { "epoch": 18.507194244604317, "grad_norm": 0.15601594746112823, "learning_rate": 6.831702710563478e-05, "loss": 0.026, "step": 20580 }, { "action_loss": 0.0029783144127577543, "epoch": 18.507194244604317, "step": 20580 }, { "epoch": 18.507194244604317, "step": 20580, "torque_loss": 0.16146616637706757 }, { "epoch": 18.51618705035971, "grad_norm": 0.13733558356761932, "learning_rate": 6.828625266169416e-05, "loss": 0.025, "step": 20590 }, { "action_loss": 0.0034104965161532164, "epoch": 18.51618705035971, "step": 20590 }, { "epoch": 18.51618705035971, "step": 20590, "torque_loss": 0.14267030358314514 }, { "epoch": 18.52517985611511, "grad_norm": 0.1859523206949234, "learning_rate": 6.825547021872534e-05, "loss": 0.0227, "step": 20600 }, { "action_loss": 0.004381501581519842, "epoch": 18.52517985611511, "step": 20600 }, { "epoch": 18.52517985611511, "step": 20600, "torque_loss": 0.23140935599803925 }, { "epoch": 18.534172661870503, "grad_norm": 0.15529638528823853, "learning_rate": 6.822467979019365e-05, "loss": 0.0204, "step": 20610 }, { "action_loss": 0.013464272022247314, "epoch": 18.534172661870503, "step": 20610 }, { "epoch": 18.534172661870503, "step": 20610, "torque_loss": 0.25630778074264526 }, { "epoch": 18.5431654676259, "grad_norm": 0.15808257460594177, "learning_rate": 6.819388138956783e-05, "loss": 0.0292, "step": 20620 }, { "action_loss": 0.007558638695627451, "epoch": 18.5431654676259, "step": 20620 }, { "epoch": 18.5431654676259, "step": 20620, "torque_loss": 0.18304188549518585 }, { "epoch": 18.552158273381295, "grad_norm": 0.12555937469005585, "learning_rate": 6.816307503032019e-05, "loss": 0.027, "step": 20630 }, { "action_loss": 0.0038219222333282232, "epoch": 18.552158273381295, "step": 20630 }, { "epoch": 18.552158273381295, "step": 20630, "torque_loss": 0.13792893290519714 }, { "epoch": 18.56115107913669, "grad_norm": 0.1171749085187912, "learning_rate": 6.813226072592645e-05, "loss": 0.0257, "step": 20640 }, { "action_loss": 0.003546901745721698, "epoch": 18.56115107913669, "step": 20640 }, { "epoch": 18.56115107913669, "step": 20640, "torque_loss": 0.16807764768600464 }, { "epoch": 18.570143884892087, "grad_norm": 0.14039944112300873, "learning_rate": 6.810143848986583e-05, "loss": 0.0205, "step": 20650 }, { "action_loss": 0.018164770677685738, "epoch": 18.570143884892087, "step": 20650 }, { "epoch": 18.570143884892087, "step": 20650, "torque_loss": 0.2592806816101074 }, { "epoch": 18.57913669064748, "grad_norm": 0.22907902300357819, "learning_rate": 6.807060833562106e-05, "loss": 0.0285, "step": 20660 }, { "action_loss": 0.003477317513898015, "epoch": 18.57913669064748, "step": 20660 }, { "epoch": 18.57913669064748, "step": 20660, "torque_loss": 0.1885877102613449 }, { "epoch": 18.58812949640288, "grad_norm": 0.15569178760051727, "learning_rate": 6.803977027667825e-05, "loss": 0.0225, "step": 20670 }, { "action_loss": 0.006491851527243853, "epoch": 18.58812949640288, "step": 20670 }, { "epoch": 18.58812949640288, "step": 20670, "torque_loss": 0.1532955765724182 }, { "epoch": 18.597122302158272, "grad_norm": 0.15832549333572388, "learning_rate": 6.800892432652704e-05, "loss": 0.0232, "step": 20680 }, { "action_loss": 0.004773515742272139, "epoch": 18.597122302158272, "step": 20680 }, { "epoch": 18.597122302158272, "step": 20680, "torque_loss": 0.16648052632808685 }, { "epoch": 18.60611510791367, "grad_norm": 0.1774715632200241, "learning_rate": 6.797807049866049e-05, "loss": 0.025, "step": 20690 }, { "action_loss": 0.0025985485408455133, "epoch": 18.60611510791367, "step": 20690 }, { "epoch": 18.60611510791367, "step": 20690, "torque_loss": 0.17065297067165375 }, { "epoch": 18.615107913669064, "grad_norm": 0.1499963104724884, "learning_rate": 6.794720880657512e-05, "loss": 0.0221, "step": 20700 }, { "action_loss": 0.0092219403013587, "epoch": 18.615107913669064, "step": 20700 }, { "epoch": 18.615107913669064, "step": 20700, "torque_loss": 0.17825736105442047 }, { "epoch": 18.62410071942446, "grad_norm": 0.20726075768470764, "learning_rate": 6.791633926377085e-05, "loss": 0.0217, "step": 20710 }, { "action_loss": 0.004844908136874437, "epoch": 18.62410071942446, "step": 20710 }, { "epoch": 18.62410071942446, "step": 20710, "torque_loss": 0.10726379603147507 }, { "epoch": 18.633093525179856, "grad_norm": 0.11242514848709106, "learning_rate": 6.788546188375114e-05, "loss": 0.0217, "step": 20720 }, { "action_loss": 0.009675734676420689, "epoch": 18.633093525179856, "step": 20720 }, { "epoch": 18.633093525179856, "step": 20720, "torque_loss": 0.24555246531963348 }, { "epoch": 18.642086330935253, "grad_norm": 0.1682382971048355, "learning_rate": 6.785457668002274e-05, "loss": 0.0301, "step": 20730 }, { "action_loss": 0.0032551269978284836, "epoch": 18.642086330935253, "step": 20730 }, { "epoch": 18.642086330935253, "step": 20730, "torque_loss": 0.10750150680541992 }, { "epoch": 18.651079136690647, "grad_norm": 0.25300973653793335, "learning_rate": 6.78236836660959e-05, "loss": 0.0284, "step": 20740 }, { "action_loss": 0.007439013104885817, "epoch": 18.651079136690647, "step": 20740 }, { "epoch": 18.651079136690647, "step": 20740, "torque_loss": 0.1540493667125702 }, { "epoch": 18.66007194244604, "grad_norm": 0.14739002287387848, "learning_rate": 6.779278285548427e-05, "loss": 0.026, "step": 20750 }, { "action_loss": 0.00866441149264574, "epoch": 18.66007194244604, "step": 20750 }, { "epoch": 18.66007194244604, "step": 20750, "torque_loss": 0.183208629488945 }, { "epoch": 18.66906474820144, "grad_norm": 0.18579845130443573, "learning_rate": 6.776187426170493e-05, "loss": 0.0266, "step": 20760 }, { "action_loss": 0.017494574189186096, "epoch": 18.66906474820144, "step": 20760 }, { "epoch": 18.66906474820144, "step": 20760, "torque_loss": 0.2684972584247589 }, { "epoch": 18.678057553956833, "grad_norm": 0.1207680031657219, "learning_rate": 6.773095789827837e-05, "loss": 0.0242, "step": 20770 }, { "action_loss": 0.011039621196687222, "epoch": 18.678057553956833, "step": 20770 }, { "epoch": 18.678057553956833, "step": 20770, "torque_loss": 0.24157828092575073 }, { "epoch": 18.68705035971223, "grad_norm": 0.12559086084365845, "learning_rate": 6.770003377872841e-05, "loss": 0.023, "step": 20780 }, { "action_loss": 0.004702722653746605, "epoch": 18.68705035971223, "step": 20780 }, { "epoch": 18.68705035971223, "step": 20780, "torque_loss": 0.09543908387422562 }, { "epoch": 18.696043165467625, "grad_norm": 0.12135283648967743, "learning_rate": 6.766910191658233e-05, "loss": 0.0304, "step": 20790 }, { "action_loss": 0.003513273550197482, "epoch": 18.696043165467625, "step": 20790 }, { "epoch": 18.696043165467625, "step": 20790, "torque_loss": 0.16447724401950836 }, { "epoch": 18.705035971223023, "grad_norm": 0.11438947170972824, "learning_rate": 6.763816232537079e-05, "loss": 0.0208, "step": 20800 }, { "action_loss": 0.004264415707439184, "epoch": 18.705035971223023, "step": 20800 }, { "epoch": 18.705035971223023, "step": 20800, "torque_loss": 0.15879477560520172 }, { "epoch": 18.714028776978417, "grad_norm": 0.107841357588768, "learning_rate": 6.76072150186278e-05, "loss": 0.025, "step": 20810 }, { "action_loss": 0.0036803698167204857, "epoch": 18.714028776978417, "step": 20810 }, { "epoch": 18.714028776978417, "step": 20810, "torque_loss": 0.11733261495828629 }, { "epoch": 18.723021582733814, "grad_norm": 0.19099965691566467, "learning_rate": 6.757626000989077e-05, "loss": 0.0213, "step": 20820 }, { "action_loss": 0.0035440598148852587, "epoch": 18.723021582733814, "step": 20820 }, { "epoch": 18.723021582733814, "step": 20820, "torque_loss": 0.15382374823093414 }, { "epoch": 18.73201438848921, "grad_norm": 0.17101454734802246, "learning_rate": 6.754529731270048e-05, "loss": 0.0253, "step": 20830 }, { "action_loss": 0.0029233170207589865, "epoch": 18.73201438848921, "step": 20830 }, { "epoch": 18.73201438848921, "step": 20830, "torque_loss": 0.14880222082138062 }, { "epoch": 18.741007194244606, "grad_norm": 0.12845352292060852, "learning_rate": 6.751432694060105e-05, "loss": 0.023, "step": 20840 }, { "action_loss": 0.004210290033370256, "epoch": 18.741007194244606, "step": 20840 }, { "epoch": 18.741007194244606, "step": 20840, "torque_loss": 0.16542208194732666 }, { "epoch": 18.75, "grad_norm": 0.15347662568092346, "learning_rate": 6.748334890713999e-05, "loss": 0.0219, "step": 20850 }, { "action_loss": 0.017639288678765297, "epoch": 18.75, "step": 20850 }, { "epoch": 18.75, "step": 20850, "torque_loss": 0.24087609350681305 }, { "epoch": 18.758992805755394, "grad_norm": 0.22063027322292328, "learning_rate": 6.745236322586813e-05, "loss": 0.0242, "step": 20860 }, { "action_loss": 0.012254844419658184, "epoch": 18.758992805755394, "step": 20860 }, { "epoch": 18.758992805755394, "step": 20860, "torque_loss": 0.208878755569458 }, { "epoch": 18.76798561151079, "grad_norm": 0.20849177241325378, "learning_rate": 6.742136991033967e-05, "loss": 0.0279, "step": 20870 }, { "action_loss": 0.007658079266548157, "epoch": 18.76798561151079, "step": 20870 }, { "epoch": 18.76798561151079, "step": 20870, "torque_loss": 0.2178516387939453 }, { "epoch": 18.776978417266186, "grad_norm": 0.1416793018579483, "learning_rate": 6.739036897411215e-05, "loss": 0.0266, "step": 20880 }, { "action_loss": 0.017978603020310402, "epoch": 18.776978417266186, "step": 20880 }, { "epoch": 18.776978417266186, "step": 20880, "torque_loss": 0.2974969744682312 }, { "epoch": 18.785971223021583, "grad_norm": 0.1782495677471161, "learning_rate": 6.73593604307464e-05, "loss": 0.0263, "step": 20890 }, { "action_loss": 0.0047957804054021835, "epoch": 18.785971223021583, "step": 20890 }, { "epoch": 18.785971223021583, "step": 20890, "torque_loss": 0.2114909291267395 }, { "epoch": 18.794964028776977, "grad_norm": 0.184695303440094, "learning_rate": 6.732834429380663e-05, "loss": 0.027, "step": 20900 }, { "action_loss": 0.00502544641494751, "epoch": 18.794964028776977, "step": 20900 }, { "epoch": 18.794964028776977, "step": 20900, "torque_loss": 0.17837531864643097 }, { "epoch": 18.803956834532375, "grad_norm": 0.12044169753789902, "learning_rate": 6.729732057686036e-05, "loss": 0.0195, "step": 20910 }, { "action_loss": 0.0032422479707747698, "epoch": 18.803956834532375, "step": 20910 }, { "epoch": 18.803956834532375, "step": 20910, "torque_loss": 0.1657613217830658 }, { "epoch": 18.81294964028777, "grad_norm": 0.18058590590953827, "learning_rate": 6.726628929347839e-05, "loss": 0.0287, "step": 20920 }, { "action_loss": 0.003288896521553397, "epoch": 18.81294964028777, "step": 20920 }, { "epoch": 18.81294964028777, "step": 20920, "torque_loss": 0.18499666452407837 }, { "epoch": 18.821942446043167, "grad_norm": 0.22041529417037964, "learning_rate": 6.723525045723488e-05, "loss": 0.0282, "step": 20930 }, { "action_loss": 0.010015743784606457, "epoch": 18.821942446043167, "step": 20930 }, { "epoch": 18.821942446043167, "step": 20930, "torque_loss": 0.175650954246521 }, { "epoch": 18.83093525179856, "grad_norm": 0.11766204237937927, "learning_rate": 6.720420408170728e-05, "loss": 0.0214, "step": 20940 }, { "action_loss": 0.004308870527893305, "epoch": 18.83093525179856, "step": 20940 }, { "epoch": 18.83093525179856, "step": 20940, "torque_loss": 0.13457418978214264 }, { "epoch": 18.83992805755396, "grad_norm": 0.18825693428516388, "learning_rate": 6.71731501804763e-05, "loss": 0.0237, "step": 20950 }, { "action_loss": 0.006237228866666555, "epoch": 18.83992805755396, "step": 20950 }, { "epoch": 18.83992805755396, "step": 20950, "torque_loss": 0.22777558863162994 }, { "epoch": 18.848920863309353, "grad_norm": 0.17888407409191132, "learning_rate": 6.714208876712602e-05, "loss": 0.024, "step": 20960 }, { "action_loss": 0.004437034949660301, "epoch": 18.848920863309353, "step": 20960 }, { "epoch": 18.848920863309353, "step": 20960, "torque_loss": 0.18519838154315948 }, { "epoch": 18.857913669064747, "grad_norm": 0.24445879459381104, "learning_rate": 6.71110198552437e-05, "loss": 0.024, "step": 20970 }, { "action_loss": 0.009461636655032635, "epoch": 18.857913669064747, "step": 20970 }, { "epoch": 18.857913669064747, "step": 20970, "torque_loss": 0.23043812811374664 }, { "epoch": 18.866906474820144, "grad_norm": 0.18353337049484253, "learning_rate": 6.707994345841996e-05, "loss": 0.0258, "step": 20980 }, { "action_loss": 0.017886927351355553, "epoch": 18.866906474820144, "step": 20980 }, { "epoch": 18.866906474820144, "step": 20980, "torque_loss": 0.25259992480278015 }, { "epoch": 18.87589928057554, "grad_norm": 0.19281990826129913, "learning_rate": 6.70488595902487e-05, "loss": 0.0251, "step": 20990 }, { "action_loss": 0.005773074924945831, "epoch": 18.87589928057554, "step": 20990 }, { "epoch": 18.87589928057554, "step": 20990, "torque_loss": 0.2055639624595642 }, { "epoch": 18.884892086330936, "grad_norm": 0.1353333294391632, "learning_rate": 6.7017768264327e-05, "loss": 0.0244, "step": 21000 }, { "action_loss": 0.003047056496143341, "epoch": 18.884892086330936, "step": 21000 }, { "epoch": 18.884892086330936, "step": 21000, "torque_loss": 0.11156494170427322 }, { "epoch": 18.89388489208633, "grad_norm": 0.13580413162708282, "learning_rate": 6.698666949425532e-05, "loss": 0.0241, "step": 21010 }, { "action_loss": 0.007554126903414726, "epoch": 18.89388489208633, "step": 21010 }, { "epoch": 18.89388489208633, "step": 21010, "torque_loss": 0.10391855984926224 }, { "epoch": 18.902877697841728, "grad_norm": 0.18231232464313507, "learning_rate": 6.69555632936373e-05, "loss": 0.0335, "step": 21020 }, { "action_loss": 0.017048629000782967, "epoch": 18.902877697841728, "step": 21020 }, { "epoch": 18.902877697841728, "step": 21020, "torque_loss": 0.1844208687543869 }, { "epoch": 18.91187050359712, "grad_norm": 0.1798350065946579, "learning_rate": 6.692444967607983e-05, "loss": 0.0317, "step": 21030 }, { "action_loss": 0.005828669294714928, "epoch": 18.91187050359712, "step": 21030 }, { "epoch": 18.91187050359712, "step": 21030, "torque_loss": 0.17416536808013916 }, { "epoch": 18.92086330935252, "grad_norm": 0.23330983519554138, "learning_rate": 6.689332865519309e-05, "loss": 0.0284, "step": 21040 }, { "action_loss": 0.010224304161965847, "epoch": 18.92086330935252, "step": 21040 }, { "epoch": 18.92086330935252, "step": 21040, "torque_loss": 0.14523600041866302 }, { "epoch": 18.929856115107913, "grad_norm": 0.2270248383283615, "learning_rate": 6.686220024459045e-05, "loss": 0.0278, "step": 21050 }, { "action_loss": 0.013964599929749966, "epoch": 18.929856115107913, "step": 21050 }, { "epoch": 18.929856115107913, "step": 21050, "torque_loss": 0.1793643683195114 }, { "epoch": 18.93884892086331, "grad_norm": 0.22213371098041534, "learning_rate": 6.683106445788856e-05, "loss": 0.0291, "step": 21060 }, { "action_loss": 0.0076513648964464664, "epoch": 18.93884892086331, "step": 21060 }, { "epoch": 18.93884892086331, "step": 21060, "torque_loss": 0.18685555458068848 }, { "epoch": 18.947841726618705, "grad_norm": 0.1348399817943573, "learning_rate": 6.679992130870723e-05, "loss": 0.0232, "step": 21070 }, { "action_loss": 0.009474613703787327, "epoch": 18.947841726618705, "step": 21070 }, { "epoch": 18.947841726618705, "step": 21070, "torque_loss": 0.17449752986431122 }, { "epoch": 18.9568345323741, "grad_norm": 0.14760923385620117, "learning_rate": 6.676877081066957e-05, "loss": 0.0234, "step": 21080 }, { "action_loss": 0.007580104749649763, "epoch": 18.9568345323741, "step": 21080 }, { "epoch": 18.9568345323741, "step": 21080, "torque_loss": 0.1757577806711197 }, { "epoch": 18.965827338129497, "grad_norm": 0.19994507730007172, "learning_rate": 6.673761297740186e-05, "loss": 0.0223, "step": 21090 }, { "action_loss": 0.0032024383544921875, "epoch": 18.965827338129497, "step": 21090 }, { "epoch": 18.965827338129497, "step": 21090, "torque_loss": 0.16187803447246552 }, { "epoch": 18.97482014388489, "grad_norm": 0.1732955127954483, "learning_rate": 6.670644782253358e-05, "loss": 0.0239, "step": 21100 }, { "action_loss": 0.007451870013028383, "epoch": 18.97482014388489, "step": 21100 }, { "epoch": 18.97482014388489, "step": 21100, "torque_loss": 0.1705426722764969 }, { "epoch": 18.98381294964029, "grad_norm": 0.19745033979415894, "learning_rate": 6.667527535969744e-05, "loss": 0.0249, "step": 21110 }, { "action_loss": 0.004398660268634558, "epoch": 18.98381294964029, "step": 21110 }, { "epoch": 18.98381294964029, "step": 21110, "torque_loss": 0.11727386713027954 }, { "epoch": 18.992805755395683, "grad_norm": 0.16542772948741913, "learning_rate": 6.664409560252933e-05, "loss": 0.0259, "step": 21120 }, { "action_loss": 0.0037083879578858614, "epoch": 18.992805755395683, "step": 21120 }, { "epoch": 18.992805755395683, "step": 21120, "torque_loss": 0.18636643886566162 }, { "epoch": 19.00179856115108, "grad_norm": 0.159934401512146, "learning_rate": 6.661290856466833e-05, "loss": 0.025, "step": 21130 }, { "action_loss": 0.008173966780304909, "epoch": 19.00179856115108, "step": 21130 }, { "epoch": 19.00179856115108, "step": 21130, "torque_loss": 0.2474142163991928 }, { "epoch": 19.010791366906474, "grad_norm": 0.17664913833141327, "learning_rate": 6.658171425975672e-05, "loss": 0.0276, "step": 21140 }, { "action_loss": 0.005904544144868851, "epoch": 19.010791366906474, "step": 21140 }, { "epoch": 19.010791366906474, "step": 21140, "torque_loss": 0.18570275604724884 }, { "epoch": 19.019784172661872, "grad_norm": 0.14797627925872803, "learning_rate": 6.655051270143994e-05, "loss": 0.0227, "step": 21150 }, { "action_loss": 0.005433402955532074, "epoch": 19.019784172661872, "step": 21150 }, { "epoch": 19.019784172661872, "step": 21150, "torque_loss": 0.14920365810394287 }, { "epoch": 19.028776978417266, "grad_norm": 0.12988349795341492, "learning_rate": 6.651930390336662e-05, "loss": 0.0218, "step": 21160 }, { "action_loss": 0.005471998360008001, "epoch": 19.028776978417266, "step": 21160 }, { "epoch": 19.028776978417266, "step": 21160, "torque_loss": 0.1468564122915268 }, { "epoch": 19.037769784172664, "grad_norm": 0.15534694492816925, "learning_rate": 6.648808787918854e-05, "loss": 0.0282, "step": 21170 }, { "action_loss": 0.005859410855919123, "epoch": 19.037769784172664, "step": 21170 }, { "epoch": 19.037769784172664, "step": 21170, "torque_loss": 0.17661362886428833 }, { "epoch": 19.046762589928058, "grad_norm": 0.1787232905626297, "learning_rate": 6.645686464256065e-05, "loss": 0.0251, "step": 21180 }, { "action_loss": 0.004480494651943445, "epoch": 19.046762589928058, "step": 21180 }, { "epoch": 19.046762589928058, "step": 21180, "torque_loss": 0.19824548065662384 }, { "epoch": 19.055755395683452, "grad_norm": 0.13856546580791473, "learning_rate": 6.642563420714106e-05, "loss": 0.0233, "step": 21190 }, { "action_loss": 0.00844952929764986, "epoch": 19.055755395683452, "step": 21190 }, { "epoch": 19.055755395683452, "step": 21190, "torque_loss": 0.1897115260362625 }, { "epoch": 19.06474820143885, "grad_norm": 0.23064187169075012, "learning_rate": 6.639439658659102e-05, "loss": 0.0264, "step": 21200 }, { "action_loss": 0.006483687553554773, "epoch": 19.06474820143885, "step": 21200 }, { "epoch": 19.06474820143885, "step": 21200, "torque_loss": 0.15296627581119537 }, { "epoch": 19.073741007194243, "grad_norm": 0.10523490607738495, "learning_rate": 6.636315179457492e-05, "loss": 0.0215, "step": 21210 }, { "action_loss": 0.0036980232689529657, "epoch": 19.073741007194243, "step": 21210 }, { "epoch": 19.073741007194243, "step": 21210, "torque_loss": 0.1289188712835312 }, { "epoch": 19.08273381294964, "grad_norm": 0.15550898015499115, "learning_rate": 6.633189984476031e-05, "loss": 0.0276, "step": 21220 }, { "action_loss": 0.002529613673686981, "epoch": 19.08273381294964, "step": 21220 }, { "epoch": 19.08273381294964, "step": 21220, "torque_loss": 0.11808160692453384 }, { "epoch": 19.091726618705035, "grad_norm": 0.21047517657279968, "learning_rate": 6.630064075081785e-05, "loss": 0.0265, "step": 21230 }, { "action_loss": 0.0050361305475234985, "epoch": 19.091726618705035, "step": 21230 }, { "epoch": 19.091726618705035, "step": 21230, "torque_loss": 0.1845494657754898 }, { "epoch": 19.100719424460433, "grad_norm": 0.08930861949920654, "learning_rate": 6.626937452642132e-05, "loss": 0.0207, "step": 21240 }, { "action_loss": 0.03422587737441063, "epoch": 19.100719424460433, "step": 21240 }, { "epoch": 19.100719424460433, "step": 21240, "torque_loss": 0.24984383583068848 }, { "epoch": 19.109712230215827, "grad_norm": 0.1808987408876419, "learning_rate": 6.623810118524765e-05, "loss": 0.0275, "step": 21250 }, { "action_loss": 0.01679813675582409, "epoch": 19.109712230215827, "step": 21250 }, { "epoch": 19.109712230215827, "step": 21250, "torque_loss": 0.29417794942855835 }, { "epoch": 19.118705035971225, "grad_norm": 0.15285788476467133, "learning_rate": 6.620682074097682e-05, "loss": 0.0281, "step": 21260 }, { "action_loss": 0.007115019951015711, "epoch": 19.118705035971225, "step": 21260 }, { "epoch": 19.118705035971225, "step": 21260, "torque_loss": 0.2042456418275833 }, { "epoch": 19.12769784172662, "grad_norm": 0.1393643021583557, "learning_rate": 6.6175533207292e-05, "loss": 0.0235, "step": 21270 }, { "action_loss": 0.006425156723707914, "epoch": 19.12769784172662, "step": 21270 }, { "epoch": 19.12769784172662, "step": 21270, "torque_loss": 0.16557590663433075 }, { "epoch": 19.136690647482013, "grad_norm": 0.08940165489912033, "learning_rate": 6.61442385978794e-05, "loss": 0.0229, "step": 21280 }, { "action_loss": 0.0019334416138008237, "epoch": 19.136690647482013, "step": 21280 }, { "epoch": 19.136690647482013, "step": 21280, "torque_loss": 0.11233479529619217 }, { "epoch": 19.14568345323741, "grad_norm": 0.15218935906887054, "learning_rate": 6.611293692642836e-05, "loss": 0.02, "step": 21290 }, { "action_loss": 0.00222808332182467, "epoch": 19.14568345323741, "step": 21290 }, { "epoch": 19.14568345323741, "step": 21290, "torque_loss": 0.10092685371637344 }, { "epoch": 19.154676258992804, "grad_norm": 0.1579880267381668, "learning_rate": 6.608162820663128e-05, "loss": 0.0241, "step": 21300 }, { "action_loss": 0.011801861226558685, "epoch": 19.154676258992804, "step": 21300 }, { "epoch": 19.154676258992804, "step": 21300, "torque_loss": 0.17739231884479523 }, { "epoch": 19.163669064748202, "grad_norm": 0.17435283958911896, "learning_rate": 6.605031245218365e-05, "loss": 0.0252, "step": 21310 }, { "action_loss": 0.004816838074475527, "epoch": 19.163669064748202, "step": 21310 }, { "epoch": 19.163669064748202, "step": 21310, "torque_loss": 0.20902693271636963 }, { "epoch": 19.172661870503596, "grad_norm": 0.1386164277791977, "learning_rate": 6.601898967678408e-05, "loss": 0.0236, "step": 21320 }, { "action_loss": 0.025547659024596214, "epoch": 19.172661870503596, "step": 21320 }, { "epoch": 19.172661870503596, "step": 21320, "torque_loss": 0.27740103006362915 }, { "epoch": 19.181654676258994, "grad_norm": 0.1801876574754715, "learning_rate": 6.598765989413419e-05, "loss": 0.0272, "step": 21330 }, { "action_loss": 0.009009399451315403, "epoch": 19.181654676258994, "step": 21330 }, { "epoch": 19.181654676258994, "step": 21330, "torque_loss": 0.2012113332748413 }, { "epoch": 19.190647482014388, "grad_norm": 0.1610928326845169, "learning_rate": 6.595632311793867e-05, "loss": 0.0222, "step": 21340 }, { "action_loss": 0.010915224440395832, "epoch": 19.190647482014388, "step": 21340 }, { "epoch": 19.190647482014388, "step": 21340, "torque_loss": 0.15487007796764374 }, { "epoch": 19.199640287769785, "grad_norm": 0.11209916323423386, "learning_rate": 6.592497936190534e-05, "loss": 0.0241, "step": 21350 }, { "action_loss": 0.0035044553223997355, "epoch": 19.199640287769785, "step": 21350 }, { "epoch": 19.199640287769785, "step": 21350, "torque_loss": 0.1664838045835495 }, { "epoch": 19.20863309352518, "grad_norm": 0.1596292108297348, "learning_rate": 6.5893628639745e-05, "loss": 0.0277, "step": 21360 }, { "action_loss": 0.002535823965445161, "epoch": 19.20863309352518, "step": 21360 }, { "epoch": 19.20863309352518, "step": 21360, "torque_loss": 0.1392560750246048 }, { "epoch": 19.217625899280577, "grad_norm": 0.11903524398803711, "learning_rate": 6.586227096517151e-05, "loss": 0.0215, "step": 21370 }, { "action_loss": 0.006178554613143206, "epoch": 19.217625899280577, "step": 21370 }, { "epoch": 19.217625899280577, "step": 21370, "torque_loss": 0.12748537957668304 }, { "epoch": 19.22661870503597, "grad_norm": 0.1429099291563034, "learning_rate": 6.58309063519018e-05, "loss": 0.0248, "step": 21380 }, { "action_loss": 0.0031833285465836525, "epoch": 19.22661870503597, "step": 21380 }, { "epoch": 19.22661870503597, "step": 21380, "torque_loss": 0.1900458186864853 }, { "epoch": 19.235611510791365, "grad_norm": 0.13029512763023376, "learning_rate": 6.579953481365582e-05, "loss": 0.0234, "step": 21390 }, { "action_loss": 0.004599392879754305, "epoch": 19.235611510791365, "step": 21390 }, { "epoch": 19.235611510791365, "step": 21390, "torque_loss": 0.20037706196308136 }, { "epoch": 19.244604316546763, "grad_norm": 0.20429491996765137, "learning_rate": 6.576815636415652e-05, "loss": 0.0251, "step": 21400 }, { "action_loss": 0.004603391047567129, "epoch": 19.244604316546763, "step": 21400 }, { "epoch": 19.244604316546763, "step": 21400, "torque_loss": 0.1761099100112915 }, { "epoch": 19.253597122302157, "grad_norm": 0.15671156346797943, "learning_rate": 6.573677101712991e-05, "loss": 0.0216, "step": 21410 }, { "action_loss": 0.013576905243098736, "epoch": 19.253597122302157, "step": 21410 }, { "epoch": 19.253597122302157, "step": 21410, "torque_loss": 0.18755872547626495 }, { "epoch": 19.262589928057555, "grad_norm": 0.16194374859333038, "learning_rate": 6.570537878630502e-05, "loss": 0.0335, "step": 21420 }, { "action_loss": 0.0026706140488386154, "epoch": 19.262589928057555, "step": 21420 }, { "epoch": 19.262589928057555, "step": 21420, "torque_loss": 0.10148439556360245 }, { "epoch": 19.27158273381295, "grad_norm": 0.1531413346529007, "learning_rate": 6.56739796854139e-05, "loss": 0.0196, "step": 21430 }, { "action_loss": 0.004591602366417646, "epoch": 19.27158273381295, "step": 21430 }, { "epoch": 19.27158273381295, "step": 21430, "torque_loss": 0.12798815965652466 }, { "epoch": 19.280575539568346, "grad_norm": 0.10811510682106018, "learning_rate": 6.564257372819153e-05, "loss": 0.0224, "step": 21440 }, { "action_loss": 0.004677170421928167, "epoch": 19.280575539568346, "step": 21440 }, { "epoch": 19.280575539568346, "step": 21440, "torque_loss": 0.1387873888015747 }, { "epoch": 19.28956834532374, "grad_norm": 0.11221326142549515, "learning_rate": 6.561116092837596e-05, "loss": 0.0252, "step": 21450 }, { "action_loss": 0.0027894105296581984, "epoch": 19.28956834532374, "step": 21450 }, { "epoch": 19.28956834532374, "step": 21450, "torque_loss": 0.14971475303173065 }, { "epoch": 19.298561151079138, "grad_norm": 0.10466010123491287, "learning_rate": 6.557974129970825e-05, "loss": 0.0205, "step": 21460 }, { "action_loss": 0.007587359752506018, "epoch": 19.298561151079138, "step": 21460 }, { "epoch": 19.298561151079138, "step": 21460, "torque_loss": 0.23531603813171387 }, { "epoch": 19.307553956834532, "grad_norm": 0.15808506309986115, "learning_rate": 6.554831485593239e-05, "loss": 0.0257, "step": 21470 }, { "action_loss": 0.004269487224519253, "epoch": 19.307553956834532, "step": 21470 }, { "epoch": 19.307553956834532, "step": 21470, "torque_loss": 0.17839284241199493 }, { "epoch": 19.31654676258993, "grad_norm": 0.1350494772195816, "learning_rate": 6.551688161079534e-05, "loss": 0.0245, "step": 21480 }, { "action_loss": 0.0073910970240831375, "epoch": 19.31654676258993, "step": 21480 }, { "epoch": 19.31654676258993, "step": 21480, "torque_loss": 0.2000276893377304 }, { "epoch": 19.325539568345324, "grad_norm": 0.13395801186561584, "learning_rate": 6.548544157804712e-05, "loss": 0.0238, "step": 21490 }, { "action_loss": 0.00987231358885765, "epoch": 19.325539568345324, "step": 21490 }, { "epoch": 19.325539568345324, "step": 21490, "torque_loss": 0.22860103845596313 }, { "epoch": 19.334532374100718, "grad_norm": 0.17408886551856995, "learning_rate": 6.545399477144065e-05, "loss": 0.0233, "step": 21500 }, { "action_loss": 0.0057996902614831924, "epoch": 19.334532374100718, "step": 21500 }, { "epoch": 19.334532374100718, "step": 21500, "torque_loss": 0.14056311547756195 }, { "epoch": 19.343525179856115, "grad_norm": 0.18387256562709808, "learning_rate": 6.542254120473184e-05, "loss": 0.0224, "step": 21510 }, { "action_loss": 0.004401488229632378, "epoch": 19.343525179856115, "step": 21510 }, { "epoch": 19.343525179856115, "step": 21510, "torque_loss": 0.168506920337677 }, { "epoch": 19.35251798561151, "grad_norm": 0.23970137536525726, "learning_rate": 6.539108089167953e-05, "loss": 0.0303, "step": 21520 }, { "action_loss": 0.00491791358217597, "epoch": 19.35251798561151, "step": 21520 }, { "epoch": 19.35251798561151, "step": 21520, "torque_loss": 0.1526906043291092 }, { "epoch": 19.361510791366907, "grad_norm": 0.14816467463970184, "learning_rate": 6.535961384604554e-05, "loss": 0.0237, "step": 21530 }, { "action_loss": 0.007762577384710312, "epoch": 19.361510791366907, "step": 21530 }, { "epoch": 19.361510791366907, "step": 21530, "torque_loss": 0.20972484350204468 }, { "epoch": 19.3705035971223, "grad_norm": 0.16172634065151215, "learning_rate": 6.532814008159461e-05, "loss": 0.028, "step": 21540 }, { "action_loss": 0.013098192401230335, "epoch": 19.3705035971223, "step": 21540 }, { "epoch": 19.3705035971223, "step": 21540, "torque_loss": 0.22673726081848145 }, { "epoch": 19.3794964028777, "grad_norm": 0.15047496557235718, "learning_rate": 6.529665961209446e-05, "loss": 0.0269, "step": 21550 }, { "action_loss": 0.006684416439384222, "epoch": 19.3794964028777, "step": 21550 }, { "epoch": 19.3794964028777, "step": 21550, "torque_loss": 0.18457551300525665 }, { "epoch": 19.388489208633093, "grad_norm": 0.13703468441963196, "learning_rate": 6.52651724513157e-05, "loss": 0.0246, "step": 21560 }, { "action_loss": 0.008012373931705952, "epoch": 19.388489208633093, "step": 21560 }, { "epoch": 19.388489208633093, "step": 21560, "torque_loss": 0.2232627272605896 }, { "epoch": 19.39748201438849, "grad_norm": 0.12062965333461761, "learning_rate": 6.523367861303189e-05, "loss": 0.0239, "step": 21570 }, { "action_loss": 0.002322352956980467, "epoch": 19.39748201438849, "step": 21570 }, { "epoch": 19.39748201438849, "step": 21570, "torque_loss": 0.0763939842581749 }, { "epoch": 19.406474820143885, "grad_norm": 0.11819237470626831, "learning_rate": 6.520217811101951e-05, "loss": 0.0299, "step": 21580 }, { "action_loss": 0.003441259264945984, "epoch": 19.406474820143885, "step": 21580 }, { "epoch": 19.406474820143885, "step": 21580, "torque_loss": 0.1442926675081253 }, { "epoch": 19.415467625899282, "grad_norm": 0.16779999434947968, "learning_rate": 6.517067095905793e-05, "loss": 0.0257, "step": 21590 }, { "action_loss": 0.007759074214845896, "epoch": 19.415467625899282, "step": 21590 }, { "epoch": 19.415467625899282, "step": 21590, "torque_loss": 0.1799723356962204 }, { "epoch": 19.424460431654676, "grad_norm": 0.16498583555221558, "learning_rate": 6.513915717092948e-05, "loss": 0.023, "step": 21600 }, { "action_loss": 0.0042844065465033054, "epoch": 19.424460431654676, "step": 21600 }, { "epoch": 19.424460431654676, "step": 21600, "torque_loss": 0.16567952930927277 }, { "epoch": 19.43345323741007, "grad_norm": 0.17910659313201904, "learning_rate": 6.510763676041935e-05, "loss": 0.021, "step": 21610 }, { "action_loss": 0.01934056170284748, "epoch": 19.43345323741007, "step": 21610 }, { "epoch": 19.43345323741007, "step": 21610, "torque_loss": 0.19526951014995575 }, { "epoch": 19.442446043165468, "grad_norm": 0.15878432989120483, "learning_rate": 6.50761097413156e-05, "loss": 0.0258, "step": 21620 }, { "action_loss": 0.003504617838189006, "epoch": 19.442446043165468, "step": 21620 }, { "epoch": 19.442446043165468, "step": 21620, "torque_loss": 0.13385845720767975 }, { "epoch": 19.451438848920862, "grad_norm": 0.21515311300754547, "learning_rate": 6.504457612740928e-05, "loss": 0.0332, "step": 21630 }, { "action_loss": 0.005041537340730429, "epoch": 19.451438848920862, "step": 21630 }, { "epoch": 19.451438848920862, "step": 21630, "torque_loss": 0.13154321908950806 }, { "epoch": 19.46043165467626, "grad_norm": 0.15729883313179016, "learning_rate": 6.501303593249423e-05, "loss": 0.0251, "step": 21640 }, { "action_loss": 0.007281400263309479, "epoch": 19.46043165467626, "step": 21640 }, { "epoch": 19.46043165467626, "step": 21640, "torque_loss": 0.15662086009979248 }, { "epoch": 19.469424460431654, "grad_norm": 0.15577417612075806, "learning_rate": 6.49814891703672e-05, "loss": 0.0262, "step": 21650 }, { "action_loss": 0.005913580302149057, "epoch": 19.469424460431654, "step": 21650 }, { "epoch": 19.469424460431654, "step": 21650, "torque_loss": 0.1456102579832077 }, { "epoch": 19.47841726618705, "grad_norm": 0.198501318693161, "learning_rate": 6.494993585482783e-05, "loss": 0.0289, "step": 21660 }, { "action_loss": 0.012256864458322525, "epoch": 19.47841726618705, "step": 21660 }, { "epoch": 19.47841726618705, "step": 21660, "torque_loss": 0.21375970542430878 }, { "epoch": 19.487410071942445, "grad_norm": 0.1321963220834732, "learning_rate": 6.49183759996786e-05, "loss": 0.0276, "step": 21670 }, { "action_loss": 0.004475229885429144, "epoch": 19.487410071942445, "step": 21670 }, { "epoch": 19.487410071942445, "step": 21670, "torque_loss": 0.17328466475009918 }, { "epoch": 19.496402877697843, "grad_norm": 0.1530921310186386, "learning_rate": 6.488680961872486e-05, "loss": 0.0258, "step": 21680 }, { "action_loss": 0.02035091631114483, "epoch": 19.496402877697843, "step": 21680 }, { "epoch": 19.496402877697843, "step": 21680, "torque_loss": 0.2344917207956314 }, { "epoch": 19.505395683453237, "grad_norm": 0.1743939369916916, "learning_rate": 6.485523672577485e-05, "loss": 0.0269, "step": 21690 }, { "action_loss": 0.006293881218880415, "epoch": 19.505395683453237, "step": 21690 }, { "epoch": 19.505395683453237, "step": 21690, "torque_loss": 0.20441073179244995 }, { "epoch": 19.514388489208635, "grad_norm": 0.0867101177573204, "learning_rate": 6.482365733463959e-05, "loss": 0.0202, "step": 21700 }, { "action_loss": 0.00857460405677557, "epoch": 19.514388489208635, "step": 21700 }, { "epoch": 19.514388489208635, "step": 21700, "torque_loss": 0.1316225379705429 }, { "epoch": 19.52338129496403, "grad_norm": 0.16349267959594727, "learning_rate": 6.4792071459133e-05, "loss": 0.0212, "step": 21710 }, { "action_loss": 0.007478183601051569, "epoch": 19.52338129496403, "step": 21710 }, { "epoch": 19.52338129496403, "step": 21710, "torque_loss": 0.19985522329807281 }, { "epoch": 19.532374100719423, "grad_norm": 0.21141427755355835, "learning_rate": 6.476047911307179e-05, "loss": 0.0259, "step": 21720 }, { "action_loss": 0.003148999297991395, "epoch": 19.532374100719423, "step": 21720 }, { "epoch": 19.532374100719423, "step": 21720, "torque_loss": 0.1828974038362503 }, { "epoch": 19.54136690647482, "grad_norm": 0.2784110903739929, "learning_rate": 6.472888031027556e-05, "loss": 0.029, "step": 21730 }, { "action_loss": 0.0030961830634623766, "epoch": 19.54136690647482, "step": 21730 }, { "epoch": 19.54136690647482, "step": 21730, "torque_loss": 0.09329160302877426 }, { "epoch": 19.550359712230215, "grad_norm": 0.12765203416347504, "learning_rate": 6.469727506456666e-05, "loss": 0.0203, "step": 21740 }, { "action_loss": 0.009923648089170456, "epoch": 19.550359712230215, "step": 21740 }, { "epoch": 19.550359712230215, "step": 21740, "torque_loss": 0.22140055894851685 }, { "epoch": 19.559352517985612, "grad_norm": 0.19547533988952637, "learning_rate": 6.466566338977034e-05, "loss": 0.0244, "step": 21750 }, { "action_loss": 0.010665670037269592, "epoch": 19.559352517985612, "step": 21750 }, { "epoch": 19.559352517985612, "step": 21750, "torque_loss": 0.16786162555217743 }, { "epoch": 19.568345323741006, "grad_norm": 0.1215410977602005, "learning_rate": 6.46340452997146e-05, "loss": 0.0239, "step": 21760 }, { "action_loss": 0.0030324384570121765, "epoch": 19.568345323741006, "step": 21760 }, { "epoch": 19.568345323741006, "step": 21760, "torque_loss": 0.1555875986814499 }, { "epoch": 19.577338129496404, "grad_norm": 0.16622142493724823, "learning_rate": 6.460242080823025e-05, "loss": 0.0237, "step": 21770 }, { "action_loss": 0.005294675473123789, "epoch": 19.577338129496404, "step": 21770 }, { "epoch": 19.577338129496404, "step": 21770, "torque_loss": 0.18601767718791962 }, { "epoch": 19.586330935251798, "grad_norm": 0.18698054552078247, "learning_rate": 6.457078992915093e-05, "loss": 0.0236, "step": 21780 }, { "action_loss": 0.008670964278280735, "epoch": 19.586330935251798, "step": 21780 }, { "epoch": 19.586330935251798, "step": 21780, "torque_loss": 0.1974949985742569 }, { "epoch": 19.595323741007196, "grad_norm": 0.1718345731496811, "learning_rate": 6.453915267631308e-05, "loss": 0.0274, "step": 21790 }, { "action_loss": 0.002556554740294814, "epoch": 19.595323741007196, "step": 21790 }, { "epoch": 19.595323741007196, "step": 21790, "torque_loss": 0.10123666375875473 }, { "epoch": 19.60431654676259, "grad_norm": 0.10166063159704208, "learning_rate": 6.450750906355589e-05, "loss": 0.0202, "step": 21800 }, { "action_loss": 0.007008934393525124, "epoch": 19.60431654676259, "step": 21800 }, { "epoch": 19.60431654676259, "step": 21800, "torque_loss": 0.23254258930683136 }, { "epoch": 19.613309352517987, "grad_norm": 0.18448127806186676, "learning_rate": 6.447585910472134e-05, "loss": 0.0274, "step": 21810 }, { "action_loss": 0.007626285310834646, "epoch": 19.613309352517987, "step": 21810 }, { "epoch": 19.613309352517987, "step": 21810, "torque_loss": 0.19382689893245697 }, { "epoch": 19.62230215827338, "grad_norm": 0.1715136021375656, "learning_rate": 6.444420281365424e-05, "loss": 0.0218, "step": 21820 }, { "action_loss": 0.00263276812620461, "epoch": 19.62230215827338, "step": 21820 }, { "epoch": 19.62230215827338, "step": 21820, "torque_loss": 0.13239628076553345 }, { "epoch": 19.631294964028775, "grad_norm": 0.09370624274015427, "learning_rate": 6.441254020420208e-05, "loss": 0.02, "step": 21830 }, { "action_loss": 0.0038018198683857918, "epoch": 19.631294964028775, "step": 21830 }, { "epoch": 19.631294964028775, "step": 21830, "torque_loss": 0.2058856040239334 }, { "epoch": 19.640287769784173, "grad_norm": 0.08575434982776642, "learning_rate": 6.438087129021522e-05, "loss": 0.0239, "step": 21840 }, { "action_loss": 0.029341667890548706, "epoch": 19.640287769784173, "step": 21840 }, { "epoch": 19.640287769784173, "step": 21840, "torque_loss": 0.25212836265563965 }, { "epoch": 19.649280575539567, "grad_norm": 0.1915718913078308, "learning_rate": 6.434919608554663e-05, "loss": 0.0278, "step": 21850 }, { "action_loss": 0.0017958451062440872, "epoch": 19.649280575539567, "step": 21850 }, { "epoch": 19.649280575539567, "step": 21850, "torque_loss": 0.10312455892562866 }, { "epoch": 19.658273381294965, "grad_norm": 0.1752156764268875, "learning_rate": 6.43175146040522e-05, "loss": 0.0256, "step": 21860 }, { "action_loss": 0.0023562812712043524, "epoch": 19.658273381294965, "step": 21860 }, { "epoch": 19.658273381294965, "step": 21860, "torque_loss": 0.15006299316883087 }, { "epoch": 19.66726618705036, "grad_norm": 0.12584246695041656, "learning_rate": 6.428582685959045e-05, "loss": 0.0215, "step": 21870 }, { "action_loss": 0.005390464793890715, "epoch": 19.66726618705036, "step": 21870 }, { "epoch": 19.66726618705036, "step": 21870, "torque_loss": 0.2194172888994217 }, { "epoch": 19.676258992805757, "grad_norm": 0.11595011502504349, "learning_rate": 6.42541328660227e-05, "loss": 0.0255, "step": 21880 }, { "action_loss": 0.0072502330876886845, "epoch": 19.676258992805757, "step": 21880 }, { "epoch": 19.676258992805757, "step": 21880, "torque_loss": 0.1723819524049759 }, { "epoch": 19.68525179856115, "grad_norm": 0.182925283908844, "learning_rate": 6.422243263721293e-05, "loss": 0.0279, "step": 21890 }, { "action_loss": 0.006741801276803017, "epoch": 19.68525179856115, "step": 21890 }, { "epoch": 19.68525179856115, "step": 21890, "torque_loss": 0.2127838283777237 }, { "epoch": 19.694244604316548, "grad_norm": 0.19795434176921844, "learning_rate": 6.419072618702794e-05, "loss": 0.0236, "step": 21900 }, { "action_loss": 0.003528502769768238, "epoch": 19.694244604316548, "step": 21900 }, { "epoch": 19.694244604316548, "step": 21900, "torque_loss": 0.11716582626104355 }, { "epoch": 19.703237410071942, "grad_norm": 0.12061955034732819, "learning_rate": 6.41590135293372e-05, "loss": 0.0223, "step": 21910 }, { "action_loss": 0.006452583242207766, "epoch": 19.703237410071942, "step": 21910 }, { "epoch": 19.703237410071942, "step": 21910, "torque_loss": 0.18809115886688232 }, { "epoch": 19.71223021582734, "grad_norm": 0.17159536480903625, "learning_rate": 6.41272946780129e-05, "loss": 0.0248, "step": 21920 }, { "action_loss": 0.00555334473028779, "epoch": 19.71223021582734, "step": 21920 }, { "epoch": 19.71223021582734, "step": 21920, "torque_loss": 0.2128783017396927 }, { "epoch": 19.721223021582734, "grad_norm": 0.17030739784240723, "learning_rate": 6.409556964692996e-05, "loss": 0.0233, "step": 21930 }, { "action_loss": 0.00977045763283968, "epoch": 19.721223021582734, "step": 21930 }, { "epoch": 19.721223021582734, "step": 21930, "torque_loss": 0.21219931542873383 }, { "epoch": 19.730215827338128, "grad_norm": 0.18941724300384521, "learning_rate": 6.406383844996595e-05, "loss": 0.0249, "step": 21940 }, { "action_loss": 0.004954869393259287, "epoch": 19.730215827338128, "step": 21940 }, { "epoch": 19.730215827338128, "step": 21940, "torque_loss": 0.14636178314685822 }, { "epoch": 19.739208633093526, "grad_norm": 0.1519356220960617, "learning_rate": 6.403210110100121e-05, "loss": 0.0261, "step": 21950 }, { "action_loss": 0.003382258117198944, "epoch": 19.739208633093526, "step": 21950 }, { "epoch": 19.739208633093526, "step": 21950, "torque_loss": 0.14398345351219177 }, { "epoch": 19.74820143884892, "grad_norm": 0.17838826775550842, "learning_rate": 6.400035761391871e-05, "loss": 0.0232, "step": 21960 }, { "action_loss": 0.0037145812530070543, "epoch": 19.74820143884892, "step": 21960 }, { "epoch": 19.74820143884892, "step": 21960, "torque_loss": 0.10427781194448471 }, { "epoch": 19.757194244604317, "grad_norm": 0.15501810610294342, "learning_rate": 6.396860800260416e-05, "loss": 0.0235, "step": 21970 }, { "action_loss": 0.004880235064774752, "epoch": 19.757194244604317, "step": 21970 }, { "epoch": 19.757194244604317, "step": 21970, "torque_loss": 0.10143474489450455 }, { "epoch": 19.76618705035971, "grad_norm": 0.10031376779079437, "learning_rate": 6.393685228094588e-05, "loss": 0.0267, "step": 21980 }, { "action_loss": 0.0036112547386437654, "epoch": 19.76618705035971, "step": 21980 }, { "epoch": 19.76618705035971, "step": 21980, "torque_loss": 0.1501360982656479 }, { "epoch": 19.77517985611511, "grad_norm": 0.17547838389873505, "learning_rate": 6.390509046283493e-05, "loss": 0.0281, "step": 21990 }, { "action_loss": 0.01227665413171053, "epoch": 19.77517985611511, "step": 21990 }, { "epoch": 19.77517985611511, "step": 21990, "torque_loss": 0.19415272772312164 }, { "epoch": 19.784172661870503, "grad_norm": 0.2159641683101654, "learning_rate": 6.387332256216498e-05, "loss": 0.0294, "step": 22000 }, { "action_loss": 0.005144659895449877, "epoch": 19.784172661870503, "step": 22000 }, { "epoch": 19.784172661870503, "step": 22000, "torque_loss": 0.19127897918224335 }, { "epoch": 19.7931654676259, "grad_norm": 0.12627151608467102, "learning_rate": 6.384154859283243e-05, "loss": 0.0246, "step": 22010 }, { "action_loss": 0.0043901801109313965, "epoch": 19.7931654676259, "step": 22010 }, { "epoch": 19.7931654676259, "step": 22010, "torque_loss": 0.14552658796310425 }, { "epoch": 19.802158273381295, "grad_norm": 0.1610877513885498, "learning_rate": 6.380976856873627e-05, "loss": 0.0228, "step": 22020 }, { "action_loss": 0.006590474396944046, "epoch": 19.802158273381295, "step": 22020 }, { "epoch": 19.802158273381295, "step": 22020, "torque_loss": 0.12984204292297363 }, { "epoch": 19.81115107913669, "grad_norm": 0.20834708213806152, "learning_rate": 6.377798250377815e-05, "loss": 0.0254, "step": 22030 }, { "action_loss": 0.004012411925941706, "epoch": 19.81115107913669, "step": 22030 }, { "epoch": 19.81115107913669, "step": 22030, "torque_loss": 0.16872085630893707 }, { "epoch": 19.820143884892087, "grad_norm": 0.1951146423816681, "learning_rate": 6.374619041186238e-05, "loss": 0.0249, "step": 22040 }, { "action_loss": 0.004391305614262819, "epoch": 19.820143884892087, "step": 22040 }, { "epoch": 19.820143884892087, "step": 22040, "torque_loss": 0.2116963118314743 }, { "epoch": 19.82913669064748, "grad_norm": 0.2660455107688904, "learning_rate": 6.371439230689592e-05, "loss": 0.0324, "step": 22050 }, { "action_loss": 0.006005011964589357, "epoch": 19.82913669064748, "step": 22050 }, { "epoch": 19.82913669064748, "step": 22050, "torque_loss": 0.15918861329555511 }, { "epoch": 19.83812949640288, "grad_norm": 0.12193452566862106, "learning_rate": 6.368258820278833e-05, "loss": 0.0246, "step": 22060 }, { "action_loss": 0.004215416964143515, "epoch": 19.83812949640288, "step": 22060 }, { "epoch": 19.83812949640288, "step": 22060, "torque_loss": 0.1768992394208908 }, { "epoch": 19.847122302158272, "grad_norm": 0.1520417034626007, "learning_rate": 6.365077811345182e-05, "loss": 0.0206, "step": 22070 }, { "action_loss": 0.012235354632139206, "epoch": 19.847122302158272, "step": 22070 }, { "epoch": 19.847122302158272, "step": 22070, "torque_loss": 0.1685224324464798 }, { "epoch": 19.85611510791367, "grad_norm": 0.17348502576351166, "learning_rate": 6.361896205280117e-05, "loss": 0.0219, "step": 22080 }, { "action_loss": 0.0036134098190814257, "epoch": 19.85611510791367, "step": 22080 }, { "epoch": 19.85611510791367, "step": 22080, "torque_loss": 0.1856507658958435 }, { "epoch": 19.865107913669064, "grad_norm": 0.09298979490995407, "learning_rate": 6.358714003475384e-05, "loss": 0.021, "step": 22090 }, { "action_loss": 0.011348079890012741, "epoch": 19.865107913669064, "step": 22090 }, { "epoch": 19.865107913669064, "step": 22090, "torque_loss": 0.20221669971942902 }, { "epoch": 19.87410071942446, "grad_norm": 0.14724554121494293, "learning_rate": 6.355531207322983e-05, "loss": 0.0277, "step": 22100 }, { "action_loss": 0.0028407692443579435, "epoch": 19.87410071942446, "step": 22100 }, { "epoch": 19.87410071942446, "step": 22100, "torque_loss": 0.1838970184326172 }, { "epoch": 19.883093525179856, "grad_norm": 0.1368079036474228, "learning_rate": 6.35234781821518e-05, "loss": 0.0219, "step": 22110 }, { "action_loss": 0.002305930946022272, "epoch": 19.883093525179856, "step": 22110 }, { "epoch": 19.883093525179856, "step": 22110, "torque_loss": 0.249501034617424 }, { "epoch": 19.892086330935253, "grad_norm": 0.14284944534301758, "learning_rate": 6.349163837544497e-05, "loss": 0.0211, "step": 22120 }, { "action_loss": 0.020829832181334496, "epoch": 19.892086330935253, "step": 22120 }, { "epoch": 19.892086330935253, "step": 22120, "torque_loss": 0.24605388939380646 }, { "epoch": 19.901079136690647, "grad_norm": 0.13260520994663239, "learning_rate": 6.345979266703714e-05, "loss": 0.0288, "step": 22130 }, { "action_loss": 0.021589281037449837, "epoch": 19.901079136690647, "step": 22130 }, { "epoch": 19.901079136690647, "step": 22130, "torque_loss": 0.2809416353702545 }, { "epoch": 19.91007194244604, "grad_norm": 0.15983082354068756, "learning_rate": 6.342794107085872e-05, "loss": 0.0271, "step": 22140 }, { "action_loss": 0.0038079414516687393, "epoch": 19.91007194244604, "step": 22140 }, { "epoch": 19.91007194244604, "step": 22140, "torque_loss": 0.11794238537549973 }, { "epoch": 19.91906474820144, "grad_norm": 0.10546845197677612, "learning_rate": 6.339608360084265e-05, "loss": 0.0263, "step": 22150 }, { "action_loss": 0.003086519194766879, "epoch": 19.91906474820144, "step": 22150 }, { "epoch": 19.91906474820144, "step": 22150, "torque_loss": 0.14482279121875763 }, { "epoch": 19.928057553956833, "grad_norm": 0.11547217518091202, "learning_rate": 6.336422027092452e-05, "loss": 0.0252, "step": 22160 }, { "action_loss": 0.0031780730932950974, "epoch": 19.928057553956833, "step": 22160 }, { "epoch": 19.928057553956833, "step": 22160, "torque_loss": 0.15256066620349884 }, { "epoch": 19.93705035971223, "grad_norm": 0.16514235734939575, "learning_rate": 6.333235109504241e-05, "loss": 0.0244, "step": 22170 }, { "action_loss": 0.004174226429313421, "epoch": 19.93705035971223, "step": 22170 }, { "epoch": 19.93705035971223, "step": 22170, "torque_loss": 0.12811052799224854 }, { "epoch": 19.946043165467625, "grad_norm": 0.11082803457975388, "learning_rate": 6.330047608713695e-05, "loss": 0.0235, "step": 22180 }, { "action_loss": 0.006254516541957855, "epoch": 19.946043165467625, "step": 22180 }, { "epoch": 19.946043165467625, "step": 22180, "torque_loss": 0.1899874210357666 }, { "epoch": 19.955035971223023, "grad_norm": 0.20897474884986877, "learning_rate": 6.32685952611514e-05, "loss": 0.0268, "step": 22190 }, { "action_loss": 0.004259452689439058, "epoch": 19.955035971223023, "step": 22190 }, { "epoch": 19.955035971223023, "step": 22190, "torque_loss": 0.16896891593933105 }, { "epoch": 19.964028776978417, "grad_norm": 0.12383235991001129, "learning_rate": 6.323670863103149e-05, "loss": 0.0253, "step": 22200 }, { "action_loss": 0.049418855458498, "epoch": 19.964028776978417, "step": 22200 }, { "epoch": 19.964028776978417, "step": 22200, "torque_loss": 0.25973203778266907 }, { "epoch": 19.973021582733814, "grad_norm": 0.21549640595912933, "learning_rate": 6.32048162107255e-05, "loss": 0.0339, "step": 22210 }, { "action_loss": 0.014711973257362843, "epoch": 19.973021582733814, "step": 22210 }, { "epoch": 19.973021582733814, "step": 22210, "torque_loss": 0.19644515216350555 }, { "epoch": 19.98201438848921, "grad_norm": 0.14071746170520782, "learning_rate": 6.317291801418431e-05, "loss": 0.0231, "step": 22220 }, { "action_loss": 0.009564233012497425, "epoch": 19.98201438848921, "step": 22220 }, { "epoch": 19.98201438848921, "step": 22220, "torque_loss": 0.23022647202014923 }, { "epoch": 19.991007194244606, "grad_norm": 0.10695449262857437, "learning_rate": 6.31410140553612e-05, "loss": 0.0255, "step": 22230 }, { "action_loss": 0.005892835557460785, "epoch": 19.991007194244606, "step": 22230 }, { "epoch": 19.991007194244606, "step": 22230, "torque_loss": 0.15533889830112457 }, { "epoch": 20.0, "grad_norm": 0.1174347922205925, "learning_rate": 6.310910434821211e-05, "loss": 0.023, "step": 22240 }, { "action_loss": 0.01189504936337471, "epoch": 20.0, "step": 22240 }, { "epoch": 20.0, "step": 22240, "torque_loss": 0.2832203209400177 }, { "epoch": 20.008992805755394, "grad_norm": 0.15607349574565887, "learning_rate": 6.307718890669541e-05, "loss": 0.0255, "step": 22250 }, { "action_loss": 0.005972022656351328, "epoch": 20.008992805755394, "step": 22250 }, { "epoch": 20.008992805755394, "step": 22250, "torque_loss": 0.1345062553882599 }, { "epoch": 20.01798561151079, "grad_norm": 0.1599549800157547, "learning_rate": 6.304526774477198e-05, "loss": 0.0249, "step": 22260 }, { "action_loss": 0.004900231491774321, "epoch": 20.01798561151079, "step": 22260 }, { "epoch": 20.01798561151079, "step": 22260, "torque_loss": 0.11838529258966446 }, { "epoch": 20.026978417266186, "grad_norm": 0.15363891422748566, "learning_rate": 6.301334087640526e-05, "loss": 0.0248, "step": 22270 }, { "action_loss": 0.003281977726146579, "epoch": 20.026978417266186, "step": 22270 }, { "epoch": 20.026978417266186, "step": 22270, "torque_loss": 0.1942298263311386 }, { "epoch": 20.035971223021583, "grad_norm": 0.17812328040599823, "learning_rate": 6.298140831556112e-05, "loss": 0.0292, "step": 22280 }, { "action_loss": 0.02454301156103611, "epoch": 20.035971223021583, "step": 22280 }, { "epoch": 20.035971223021583, "step": 22280, "torque_loss": 0.1847788542509079 }, { "epoch": 20.044964028776977, "grad_norm": 0.11873629689216614, "learning_rate": 6.294947007620794e-05, "loss": 0.0373, "step": 22290 }, { "action_loss": 0.002774987369775772, "epoch": 20.044964028776977, "step": 22290 }, { "epoch": 20.044964028776977, "step": 22290, "torque_loss": 0.1410352736711502 }, { "epoch": 20.053956834532375, "grad_norm": 0.12291037291288376, "learning_rate": 6.291752617231661e-05, "loss": 0.0252, "step": 22300 }, { "action_loss": 0.007650184910744429, "epoch": 20.053956834532375, "step": 22300 }, { "epoch": 20.053956834532375, "step": 22300, "torque_loss": 0.18832014501094818 }, { "epoch": 20.06294964028777, "grad_norm": 0.09950955212116241, "learning_rate": 6.288557661786047e-05, "loss": 0.0268, "step": 22310 }, { "action_loss": 0.0122847780585289, "epoch": 20.06294964028777, "step": 22310 }, { "epoch": 20.06294964028777, "step": 22310, "torque_loss": 0.19465744495391846 }, { "epoch": 20.071942446043167, "grad_norm": 0.13288074731826782, "learning_rate": 6.285362142681534e-05, "loss": 0.0276, "step": 22320 }, { "action_loss": 0.008164136670529842, "epoch": 20.071942446043167, "step": 22320 }, { "epoch": 20.071942446043167, "step": 22320, "torque_loss": 0.18949942290782928 }, { "epoch": 20.08093525179856, "grad_norm": 0.21409198641777039, "learning_rate": 6.28216606131595e-05, "loss": 0.027, "step": 22330 }, { "action_loss": 0.013006952591240406, "epoch": 20.08093525179856, "step": 22330 }, { "epoch": 20.08093525179856, "step": 22330, "torque_loss": 0.2042529582977295 }, { "epoch": 20.08992805755396, "grad_norm": 0.16876132786273956, "learning_rate": 6.27896941908737e-05, "loss": 0.0282, "step": 22340 }, { "action_loss": 0.007269782945513725, "epoch": 20.08992805755396, "step": 22340 }, { "epoch": 20.08992805755396, "step": 22340, "torque_loss": 0.1799641102552414 }, { "epoch": 20.098920863309353, "grad_norm": 0.14469113945960999, "learning_rate": 6.275772217394115e-05, "loss": 0.0256, "step": 22350 }, { "action_loss": 0.004008785355836153, "epoch": 20.098920863309353, "step": 22350 }, { "epoch": 20.098920863309353, "step": 22350, "torque_loss": 0.15314018726348877 }, { "epoch": 20.107913669064747, "grad_norm": 0.11093519628047943, "learning_rate": 6.27257445763475e-05, "loss": 0.0203, "step": 22360 }, { "action_loss": 0.004407496191561222, "epoch": 20.107913669064747, "step": 22360 }, { "epoch": 20.107913669064747, "step": 22360, "torque_loss": 0.1758395880460739 }, { "epoch": 20.116906474820144, "grad_norm": 0.14188045263290405, "learning_rate": 6.269376141208082e-05, "loss": 0.0244, "step": 22370 }, { "action_loss": 0.00513448566198349, "epoch": 20.116906474820144, "step": 22370 }, { "epoch": 20.116906474820144, "step": 22370, "torque_loss": 0.18527734279632568 }, { "epoch": 20.12589928057554, "grad_norm": 0.18335206806659698, "learning_rate": 6.266177269513162e-05, "loss": 0.0318, "step": 22380 }, { "action_loss": 0.011967998929321766, "epoch": 20.12589928057554, "step": 22380 }, { "epoch": 20.12589928057554, "step": 22380, "torque_loss": 0.22926442325115204 }, { "epoch": 20.134892086330936, "grad_norm": 0.21279941499233246, "learning_rate": 6.262977843949289e-05, "loss": 0.0297, "step": 22390 }, { "action_loss": 0.04303504154086113, "epoch": 20.134892086330936, "step": 22390 }, { "epoch": 20.134892086330936, "step": 22390, "torque_loss": 0.3043879568576813 }, { "epoch": 20.14388489208633, "grad_norm": 0.09215684235095978, "learning_rate": 6.259777865915997e-05, "loss": 0.0276, "step": 22400 }, { "action_loss": 0.0053776465356349945, "epoch": 20.14388489208633, "step": 22400 }, { "epoch": 20.14388489208633, "step": 22400, "torque_loss": 0.1870919018983841 }, { "epoch": 20.152877697841728, "grad_norm": 0.12018004059791565, "learning_rate": 6.256577336813069e-05, "loss": 0.0291, "step": 22410 }, { "action_loss": 0.006327465176582336, "epoch": 20.152877697841728, "step": 22410 }, { "epoch": 20.152877697841728, "step": 22410, "torque_loss": 0.15498381853103638 }, { "epoch": 20.16187050359712, "grad_norm": 0.13793061673641205, "learning_rate": 6.253376258040521e-05, "loss": 0.0257, "step": 22420 }, { "action_loss": 0.004843676462769508, "epoch": 20.16187050359712, "step": 22420 }, { "epoch": 20.16187050359712, "step": 22420, "torque_loss": 0.18403209745883942 }, { "epoch": 20.17086330935252, "grad_norm": 0.15668714046478271, "learning_rate": 6.250174630998616e-05, "loss": 0.022, "step": 22430 }, { "action_loss": 0.007489507552236319, "epoch": 20.17086330935252, "step": 22430 }, { "epoch": 20.17086330935252, "step": 22430, "torque_loss": 0.1886400729417801 }, { "epoch": 20.179856115107913, "grad_norm": 0.1487397402524948, "learning_rate": 6.246972457087854e-05, "loss": 0.0271, "step": 22440 }, { "action_loss": 0.004254410043358803, "epoch": 20.179856115107913, "step": 22440 }, { "epoch": 20.179856115107913, "step": 22440, "torque_loss": 0.1205201968550682 }, { "epoch": 20.18884892086331, "grad_norm": 0.15444578230381012, "learning_rate": 6.243769737708974e-05, "loss": 0.0237, "step": 22450 }, { "action_loss": 0.005754539277404547, "epoch": 20.18884892086331, "step": 22450 }, { "epoch": 20.18884892086331, "step": 22450, "torque_loss": 0.16427770256996155 }, { "epoch": 20.197841726618705, "grad_norm": 0.2503502368927002, "learning_rate": 6.240566474262955e-05, "loss": 0.0293, "step": 22460 }, { "action_loss": 0.0035250838845968246, "epoch": 20.197841726618705, "step": 22460 }, { "epoch": 20.197841726618705, "step": 22460, "torque_loss": 0.1401272714138031 }, { "epoch": 20.2068345323741, "grad_norm": 0.13652649521827698, "learning_rate": 6.237362668151012e-05, "loss": 0.0233, "step": 22470 }, { "action_loss": 0.016140326857566833, "epoch": 20.2068345323741, "step": 22470 }, { "epoch": 20.2068345323741, "step": 22470, "torque_loss": 0.20272450149059296 }, { "epoch": 20.215827338129497, "grad_norm": 0.10822010785341263, "learning_rate": 6.2341583207746e-05, "loss": 0.0222, "step": 22480 }, { "action_loss": 0.004651625175029039, "epoch": 20.215827338129497, "step": 22480 }, { "epoch": 20.215827338129497, "step": 22480, "torque_loss": 0.16234944760799408 }, { "epoch": 20.22482014388489, "grad_norm": 0.16735567152500153, "learning_rate": 6.23095343353541e-05, "loss": 0.0317, "step": 22490 }, { "action_loss": 0.0044649336487054825, "epoch": 20.22482014388489, "step": 22490 }, { "epoch": 20.22482014388489, "step": 22490, "torque_loss": 0.22306329011917114 }, { "epoch": 20.23381294964029, "grad_norm": 0.1730186492204666, "learning_rate": 6.227748007835366e-05, "loss": 0.0267, "step": 22500 }, { "action_loss": 0.0037607632111757994, "epoch": 20.23381294964029, "step": 22500 }, { "epoch": 20.23381294964029, "step": 22500, "torque_loss": 0.14649519324302673 }, { "epoch": 20.242805755395683, "grad_norm": 0.14310143887996674, "learning_rate": 6.224542045076632e-05, "loss": 0.0244, "step": 22510 }, { "action_loss": 0.0062174410559237, "epoch": 20.242805755395683, "step": 22510 }, { "epoch": 20.242805755395683, "step": 22510, "torque_loss": 0.2283797711133957 }, { "epoch": 20.25179856115108, "grad_norm": 0.10064697265625, "learning_rate": 6.221335546661606e-05, "loss": 0.0201, "step": 22520 }, { "action_loss": 0.002043825341388583, "epoch": 20.25179856115108, "step": 22520 }, { "epoch": 20.25179856115108, "step": 22520, "torque_loss": 0.11760652810335159 }, { "epoch": 20.260791366906474, "grad_norm": 0.14807412028312683, "learning_rate": 6.218128513992917e-05, "loss": 0.0249, "step": 22530 }, { "action_loss": 0.004139122553169727, "epoch": 20.260791366906474, "step": 22530 }, { "epoch": 20.260791366906474, "step": 22530, "torque_loss": 0.14678379893302917 }, { "epoch": 20.269784172661872, "grad_norm": 0.14277423918247223, "learning_rate": 6.214920948473433e-05, "loss": 0.0227, "step": 22540 }, { "action_loss": 0.008325730450451374, "epoch": 20.269784172661872, "step": 22540 }, { "epoch": 20.269784172661872, "step": 22540, "torque_loss": 0.2465149611234665 }, { "epoch": 20.278776978417266, "grad_norm": 0.15212717652320862, "learning_rate": 6.211712851506249e-05, "loss": 0.0232, "step": 22550 }, { "action_loss": 0.007390917744487524, "epoch": 20.278776978417266, "step": 22550 }, { "epoch": 20.278776978417266, "step": 22550, "torque_loss": 0.1703399419784546 }, { "epoch": 20.28776978417266, "grad_norm": 0.10650083422660828, "learning_rate": 6.208504224494696e-05, "loss": 0.0259, "step": 22560 }, { "action_loss": 0.013565249741077423, "epoch": 20.28776978417266, "step": 22560 }, { "epoch": 20.28776978417266, "step": 22560, "torque_loss": 0.2269352674484253 }, { "epoch": 20.296762589928058, "grad_norm": 0.17941363155841827, "learning_rate": 6.205295068842338e-05, "loss": 0.0249, "step": 22570 }, { "action_loss": 0.0023244319017976522, "epoch": 20.296762589928058, "step": 22570 }, { "epoch": 20.296762589928058, "step": 22570, "torque_loss": 0.12614117562770844 }, { "epoch": 20.305755395683452, "grad_norm": 0.14636829495429993, "learning_rate": 6.202085385952968e-05, "loss": 0.0262, "step": 22580 }, { "action_loss": 0.008873228915035725, "epoch": 20.305755395683452, "step": 22580 }, { "epoch": 20.305755395683452, "step": 22580, "torque_loss": 0.21948234736919403 }, { "epoch": 20.31474820143885, "grad_norm": 0.19798427820205688, "learning_rate": 6.19887517723061e-05, "loss": 0.0273, "step": 22590 }, { "action_loss": 0.0037120163906365633, "epoch": 20.31474820143885, "step": 22590 }, { "epoch": 20.31474820143885, "step": 22590, "torque_loss": 0.15730684995651245 }, { "epoch": 20.323741007194243, "grad_norm": 0.13506022095680237, "learning_rate": 6.19566444407952e-05, "loss": 0.0247, "step": 22600 }, { "action_loss": 0.006663397420197725, "epoch": 20.323741007194243, "step": 22600 }, { "epoch": 20.323741007194243, "step": 22600, "torque_loss": 0.17911481857299805 }, { "epoch": 20.33273381294964, "grad_norm": 0.18792752921581268, "learning_rate": 6.192453187904178e-05, "loss": 0.0202, "step": 22610 }, { "action_loss": 0.01207656878978014, "epoch": 20.33273381294964, "step": 22610 }, { "epoch": 20.33273381294964, "step": 22610, "torque_loss": 0.2289881855249405 }, { "epoch": 20.341726618705035, "grad_norm": 0.13003921508789062, "learning_rate": 6.189241410109303e-05, "loss": 0.0268, "step": 22620 }, { "action_loss": 0.00840776413679123, "epoch": 20.341726618705035, "step": 22620 }, { "epoch": 20.341726618705035, "step": 22620, "torque_loss": 0.23797555267810822 }, { "epoch": 20.350719424460433, "grad_norm": 0.14720571041107178, "learning_rate": 6.186029112099829e-05, "loss": 0.0232, "step": 22630 }, { "action_loss": 0.005166820716112852, "epoch": 20.350719424460433, "step": 22630 }, { "epoch": 20.350719424460433, "step": 22630, "torque_loss": 0.20828747749328613 }, { "epoch": 20.359712230215827, "grad_norm": 0.17791371047496796, "learning_rate": 6.182816295280926e-05, "loss": 0.0223, "step": 22640 }, { "action_loss": 0.009519174695014954, "epoch": 20.359712230215827, "step": 22640 }, { "epoch": 20.359712230215827, "step": 22640, "torque_loss": 0.12080679088830948 }, { "epoch": 20.368705035971225, "grad_norm": 0.10336935520172119, "learning_rate": 6.179602961057993e-05, "loss": 0.029, "step": 22650 }, { "action_loss": 0.0069413576275110245, "epoch": 20.368705035971225, "step": 22650 }, { "epoch": 20.368705035971225, "step": 22650, "torque_loss": 0.23011399805545807 }, { "epoch": 20.37769784172662, "grad_norm": 0.17865590751171112, "learning_rate": 6.176389110836647e-05, "loss": 0.0253, "step": 22660 }, { "action_loss": 0.0054345689713954926, "epoch": 20.37769784172662, "step": 22660 }, { "epoch": 20.37769784172662, "step": 22660, "torque_loss": 0.15023894608020782 }, { "epoch": 20.386690647482013, "grad_norm": 0.21268542110919952, "learning_rate": 6.173174746022738e-05, "loss": 0.0286, "step": 22670 }, { "action_loss": 0.004015697631984949, "epoch": 20.386690647482013, "step": 22670 }, { "epoch": 20.386690647482013, "step": 22670, "torque_loss": 0.1694745421409607 }, { "epoch": 20.39568345323741, "grad_norm": 0.2358068823814392, "learning_rate": 6.169959868022338e-05, "loss": 0.0293, "step": 22680 }, { "action_loss": 0.010677143931388855, "epoch": 20.39568345323741, "step": 22680 }, { "epoch": 20.39568345323741, "step": 22680, "torque_loss": 0.16102667152881622 }, { "epoch": 20.404676258992804, "grad_norm": 0.15690474212169647, "learning_rate": 6.166744478241742e-05, "loss": 0.0227, "step": 22690 }, { "action_loss": 0.007162459194660187, "epoch": 20.404676258992804, "step": 22690 }, { "epoch": 20.404676258992804, "step": 22690, "torque_loss": 0.19726012647151947 }, { "epoch": 20.413669064748202, "grad_norm": 0.07338397949934006, "learning_rate": 6.163528578087473e-05, "loss": 0.0225, "step": 22700 }, { "action_loss": 0.005805560853332281, "epoch": 20.413669064748202, "step": 22700 }, { "epoch": 20.413669064748202, "step": 22700, "torque_loss": 0.12870778143405914 }, { "epoch": 20.422661870503596, "grad_norm": 0.11413939297199249, "learning_rate": 6.160312168966274e-05, "loss": 0.0226, "step": 22710 }, { "action_loss": 0.0044333431869745255, "epoch": 20.422661870503596, "step": 22710 }, { "epoch": 20.422661870503596, "step": 22710, "torque_loss": 0.17297489941120148 }, { "epoch": 20.431654676258994, "grad_norm": 0.12727141380310059, "learning_rate": 6.157095252285112e-05, "loss": 0.0215, "step": 22720 }, { "action_loss": 0.006600415799766779, "epoch": 20.431654676258994, "step": 22720 }, { "epoch": 20.431654676258994, "step": 22720, "torque_loss": 0.19785577058792114 }, { "epoch": 20.440647482014388, "grad_norm": 0.16329990327358246, "learning_rate": 6.153877829451174e-05, "loss": 0.0286, "step": 22730 }, { "action_loss": 0.0024769396986812353, "epoch": 20.440647482014388, "step": 22730 }, { "epoch": 20.440647482014388, "step": 22730, "torque_loss": 0.12161407619714737 }, { "epoch": 20.449640287769785, "grad_norm": 0.16467441618442535, "learning_rate": 6.150659901871872e-05, "loss": 0.0245, "step": 22740 }, { "action_loss": 0.00931615475565195, "epoch": 20.449640287769785, "step": 22740 }, { "epoch": 20.449640287769785, "step": 22740, "torque_loss": 0.2041863650083542 }, { "epoch": 20.45863309352518, "grad_norm": 0.2190363109111786, "learning_rate": 6.147441470954836e-05, "loss": 0.0249, "step": 22750 }, { "action_loss": 0.00767575204372406, "epoch": 20.45863309352518, "step": 22750 }, { "epoch": 20.45863309352518, "step": 22750, "torque_loss": 0.19007658958435059 }, { "epoch": 20.467625899280577, "grad_norm": 0.1774035543203354, "learning_rate": 6.144222538107918e-05, "loss": 0.0221, "step": 22760 }, { "action_loss": 0.003190611256286502, "epoch": 20.467625899280577, "step": 22760 }, { "epoch": 20.467625899280577, "step": 22760, "torque_loss": 0.11039058119058609 }, { "epoch": 20.47661870503597, "grad_norm": 0.12860041856765747, "learning_rate": 6.141003104739188e-05, "loss": 0.0196, "step": 22770 }, { "action_loss": 0.021924281492829323, "epoch": 20.47661870503597, "step": 22770 }, { "epoch": 20.47661870503597, "step": 22770, "torque_loss": 0.28860557079315186 }, { "epoch": 20.485611510791365, "grad_norm": 0.12812483310699463, "learning_rate": 6.137783172256935e-05, "loss": 0.0251, "step": 22780 }, { "action_loss": 0.003717406652867794, "epoch": 20.485611510791365, "step": 22780 }, { "epoch": 20.485611510791365, "step": 22780, "torque_loss": 0.1417907029390335 }, { "epoch": 20.494604316546763, "grad_norm": 0.18511831760406494, "learning_rate": 6.134562742069669e-05, "loss": 0.0238, "step": 22790 }, { "action_loss": 0.0022998519707471132, "epoch": 20.494604316546763, "step": 22790 }, { "epoch": 20.494604316546763, "step": 22790, "torque_loss": 0.11742868274450302 }, { "epoch": 20.503597122302157, "grad_norm": 0.1911991834640503, "learning_rate": 6.131341815586111e-05, "loss": 0.0202, "step": 22800 }, { "action_loss": 0.003701435402035713, "epoch": 20.503597122302157, "step": 22800 }, { "epoch": 20.503597122302157, "step": 22800, "torque_loss": 0.18033230304718018 }, { "epoch": 20.512589928057555, "grad_norm": 0.17437732219696045, "learning_rate": 6.128120394215209e-05, "loss": 0.0283, "step": 22810 }, { "action_loss": 0.007806075736880302, "epoch": 20.512589928057555, "step": 22810 }, { "epoch": 20.512589928057555, "step": 22810, "torque_loss": 0.22815756499767303 }, { "epoch": 20.52158273381295, "grad_norm": 0.13209058344364166, "learning_rate": 6.124898479366116e-05, "loss": 0.0227, "step": 22820 }, { "action_loss": 0.005471537355333567, "epoch": 20.52158273381295, "step": 22820 }, { "epoch": 20.52158273381295, "step": 22820, "torque_loss": 0.17229948937892914 }, { "epoch": 20.530575539568346, "grad_norm": 0.16829873621463776, "learning_rate": 6.121676072448214e-05, "loss": 0.0239, "step": 22830 }, { "action_loss": 0.0024006774183362722, "epoch": 20.530575539568346, "step": 22830 }, { "epoch": 20.530575539568346, "step": 22830, "torque_loss": 0.19364865124225616 }, { "epoch": 20.53956834532374, "grad_norm": 0.11621227115392685, "learning_rate": 6.118453174871089e-05, "loss": 0.0256, "step": 22840 }, { "action_loss": 0.005171626806259155, "epoch": 20.53956834532374, "step": 22840 }, { "epoch": 20.53956834532374, "step": 22840, "torque_loss": 0.20931105315685272 }, { "epoch": 20.548561151079138, "grad_norm": 0.1446719914674759, "learning_rate": 6.115229788044548e-05, "loss": 0.0232, "step": 22850 }, { "action_loss": 0.011339779943227768, "epoch": 20.548561151079138, "step": 22850 }, { "epoch": 20.548561151079138, "step": 22850, "torque_loss": 0.2280806303024292 }, { "epoch": 20.557553956834532, "grad_norm": 0.11504968255758286, "learning_rate": 6.112005913378607e-05, "loss": 0.0241, "step": 22860 }, { "action_loss": 0.0028791495133191347, "epoch": 20.557553956834532, "step": 22860 }, { "epoch": 20.557553956834532, "step": 22860, "torque_loss": 0.2184343785047531 }, { "epoch": 20.56654676258993, "grad_norm": 0.15106505155563354, "learning_rate": 6.1087815522835e-05, "loss": 0.0219, "step": 22870 }, { "action_loss": 0.007836989127099514, "epoch": 20.56654676258993, "step": 22870 }, { "epoch": 20.56654676258993, "step": 22870, "torque_loss": 0.22255070507526398 }, { "epoch": 20.575539568345324, "grad_norm": 0.13125106692314148, "learning_rate": 6.105556706169673e-05, "loss": 0.0241, "step": 22880 }, { "action_loss": 0.0043595633469522, "epoch": 20.575539568345324, "step": 22880 }, { "epoch": 20.575539568345324, "step": 22880, "torque_loss": 0.22591345012187958 }, { "epoch": 20.584532374100718, "grad_norm": 0.10613292455673218, "learning_rate": 6.102331376447782e-05, "loss": 0.0227, "step": 22890 }, { "action_loss": 0.01937740296125412, "epoch": 20.584532374100718, "step": 22890 }, { "epoch": 20.584532374100718, "step": 22890, "torque_loss": 0.27396419644355774 }, { "epoch": 20.593525179856115, "grad_norm": 0.11622392386198044, "learning_rate": 6.099105564528697e-05, "loss": 0.0278, "step": 22900 }, { "action_loss": 0.004292618483304977, "epoch": 20.593525179856115, "step": 22900 }, { "epoch": 20.593525179856115, "step": 22900, "torque_loss": 0.1651124209165573 }, { "epoch": 20.60251798561151, "grad_norm": 0.10792898386716843, "learning_rate": 6.095879271823495e-05, "loss": 0.0245, "step": 22910 }, { "action_loss": 0.030478045344352722, "epoch": 20.60251798561151, "step": 22910 }, { "epoch": 20.60251798561151, "step": 22910, "torque_loss": 0.23456184566020966 }, { "epoch": 20.611510791366907, "grad_norm": 0.1667008101940155, "learning_rate": 6.09265249974347e-05, "loss": 0.0353, "step": 22920 }, { "action_loss": 0.01146495807915926, "epoch": 20.611510791366907, "step": 22920 }, { "epoch": 20.611510791366907, "step": 22920, "torque_loss": 0.17383591830730438 }, { "epoch": 20.6205035971223, "grad_norm": 0.17408771812915802, "learning_rate": 6.089425249700118e-05, "loss": 0.0251, "step": 22930 }, { "action_loss": 0.0042907726019620895, "epoch": 20.6205035971223, "step": 22930 }, { "epoch": 20.6205035971223, "step": 22930, "torque_loss": 0.18756286799907684 }, { "epoch": 20.6294964028777, "grad_norm": 0.16495320200920105, "learning_rate": 6.0861975231051505e-05, "loss": 0.0244, "step": 22940 }, { "action_loss": 0.023721009492874146, "epoch": 20.6294964028777, "step": 22940 }, { "epoch": 20.6294964028777, "step": 22940, "torque_loss": 0.22363872826099396 }, { "epoch": 20.638489208633093, "grad_norm": 0.20523135364055634, "learning_rate": 6.082969321370484e-05, "loss": 0.029, "step": 22950 }, { "action_loss": 0.0043367356993258, "epoch": 20.638489208633093, "step": 22950 }, { "epoch": 20.638489208633093, "step": 22950, "torque_loss": 0.20268352329730988 }, { "epoch": 20.64748201438849, "grad_norm": 0.10281065851449966, "learning_rate": 6.079740645908243e-05, "loss": 0.0196, "step": 22960 }, { "action_loss": 0.006355665624141693, "epoch": 20.64748201438849, "step": 22960 }, { "epoch": 20.64748201438849, "step": 22960, "torque_loss": 0.16951163113117218 }, { "epoch": 20.656474820143885, "grad_norm": 0.083852119743824, "learning_rate": 6.07651149813076e-05, "loss": 0.0198, "step": 22970 }, { "action_loss": 0.00886655692011118, "epoch": 20.656474820143885, "step": 22970 }, { "epoch": 20.656474820143885, "step": 22970, "torque_loss": 0.1709216833114624 }, { "epoch": 20.665467625899282, "grad_norm": 0.13441118597984314, "learning_rate": 6.073281879450574e-05, "loss": 0.0222, "step": 22980 }, { "action_loss": 0.00547057157382369, "epoch": 20.665467625899282, "step": 22980 }, { "epoch": 20.665467625899282, "step": 22980, "torque_loss": 0.14564137160778046 }, { "epoch": 20.674460431654676, "grad_norm": 0.12962153553962708, "learning_rate": 6.070051791280431e-05, "loss": 0.0238, "step": 22990 }, { "action_loss": 0.004570978228002787, "epoch": 20.674460431654676, "step": 22990 }, { "epoch": 20.674460431654676, "step": 22990, "torque_loss": 0.1575692594051361 }, { "epoch": 20.68345323741007, "grad_norm": 0.11820270121097565, "learning_rate": 6.066821235033281e-05, "loss": 0.0269, "step": 23000 }, { "action_loss": 0.008250545710325241, "epoch": 20.68345323741007, "step": 23000 }, { "epoch": 20.68345323741007, "step": 23000, "torque_loss": 0.1505172699689865 }, { "epoch": 20.692446043165468, "grad_norm": 0.14849214255809784, "learning_rate": 6.063590212122277e-05, "loss": 0.0202, "step": 23010 }, { "action_loss": 0.003666885895654559, "epoch": 20.692446043165468, "step": 23010 }, { "epoch": 20.692446043165468, "step": 23010, "torque_loss": 0.17148785293102264 }, { "epoch": 20.701438848920862, "grad_norm": 0.1331247240304947, "learning_rate": 6.0603587239607794e-05, "loss": 0.0244, "step": 23020 }, { "action_loss": 0.002964280778542161, "epoch": 20.701438848920862, "step": 23020 }, { "epoch": 20.701438848920862, "step": 23020, "torque_loss": 0.12954063713550568 }, { "epoch": 20.71043165467626, "grad_norm": 0.22749526798725128, "learning_rate": 6.057126771962353e-05, "loss": 0.0214, "step": 23030 }, { "action_loss": 0.009571000933647156, "epoch": 20.71043165467626, "step": 23030 }, { "epoch": 20.71043165467626, "step": 23030, "torque_loss": 0.25056588649749756 }, { "epoch": 20.719424460431654, "grad_norm": 0.12530307471752167, "learning_rate": 6.05389435754076e-05, "loss": 0.0265, "step": 23040 }, { "action_loss": 0.012158684432506561, "epoch": 20.719424460431654, "step": 23040 }, { "epoch": 20.719424460431654, "step": 23040, "torque_loss": 0.2516556680202484 }, { "epoch": 20.72841726618705, "grad_norm": 0.1338571459054947, "learning_rate": 6.0506614821099705e-05, "loss": 0.0253, "step": 23050 }, { "action_loss": 0.009259522892534733, "epoch": 20.72841726618705, "step": 23050 }, { "epoch": 20.72841726618705, "step": 23050, "torque_loss": 0.21669237315654755 }, { "epoch": 20.737410071942445, "grad_norm": 0.1097947284579277, "learning_rate": 6.047428147084153e-05, "loss": 0.0257, "step": 23060 }, { "action_loss": 0.0017050523310899734, "epoch": 20.737410071942445, "step": 23060 }, { "epoch": 20.737410071942445, "step": 23060, "torque_loss": 0.14601261913776398 }, { "epoch": 20.746402877697843, "grad_norm": 0.16584816575050354, "learning_rate": 6.044194353877679e-05, "loss": 0.0223, "step": 23070 }, { "action_loss": 0.0021196026355028152, "epoch": 20.746402877697843, "step": 23070 }, { "epoch": 20.746402877697843, "step": 23070, "torque_loss": 0.0919923409819603 }, { "epoch": 20.755395683453237, "grad_norm": 0.13618101179599762, "learning_rate": 6.04096010390512e-05, "loss": 0.0212, "step": 23080 }, { "action_loss": 0.0061862594448029995, "epoch": 20.755395683453237, "step": 23080 }, { "epoch": 20.755395683453237, "step": 23080, "torque_loss": 0.12493109703063965 }, { "epoch": 20.764388489208635, "grad_norm": 0.16691409051418304, "learning_rate": 6.0377253985812445e-05, "loss": 0.0259, "step": 23090 }, { "action_loss": 0.007197209168225527, "epoch": 20.764388489208635, "step": 23090 }, { "epoch": 20.764388489208635, "step": 23090, "torque_loss": 0.142685666680336 }, { "epoch": 20.77338129496403, "grad_norm": 0.15179936587810516, "learning_rate": 6.0344902393210235e-05, "loss": 0.0242, "step": 23100 }, { "action_loss": 0.00930699147284031, "epoch": 20.77338129496403, "step": 23100 }, { "epoch": 20.77338129496403, "step": 23100, "torque_loss": 0.23952220380306244 }, { "epoch": 20.782374100719423, "grad_norm": 0.14766766130924225, "learning_rate": 6.031254627539625e-05, "loss": 0.0246, "step": 23110 }, { "action_loss": 0.012741215527057648, "epoch": 20.782374100719423, "step": 23110 }, { "epoch": 20.782374100719423, "step": 23110, "torque_loss": 0.17937994003295898 }, { "epoch": 20.79136690647482, "grad_norm": 0.17724449932575226, "learning_rate": 6.028018564652418e-05, "loss": 0.025, "step": 23120 }, { "action_loss": 0.003829159075394273, "epoch": 20.79136690647482, "step": 23120 }, { "epoch": 20.79136690647482, "step": 23120, "torque_loss": 0.19346682727336884 }, { "epoch": 20.800359712230215, "grad_norm": 0.18216465413570404, "learning_rate": 6.024782052074964e-05, "loss": 0.027, "step": 23130 }, { "action_loss": 0.00555550679564476, "epoch": 20.800359712230215, "step": 23130 }, { "epoch": 20.800359712230215, "step": 23130, "torque_loss": 0.13755038380622864 }, { "epoch": 20.809352517985612, "grad_norm": 0.13654601573944092, "learning_rate": 6.021545091223026e-05, "loss": 0.0212, "step": 23140 }, { "action_loss": 0.0048015909269452095, "epoch": 20.809352517985612, "step": 23140 }, { "epoch": 20.809352517985612, "step": 23140, "torque_loss": 0.14798979461193085 }, { "epoch": 20.818345323741006, "grad_norm": 0.16502246260643005, "learning_rate": 6.0183076835125586e-05, "loss": 0.0228, "step": 23150 }, { "action_loss": 0.004487965255975723, "epoch": 20.818345323741006, "step": 23150 }, { "epoch": 20.818345323741006, "step": 23150, "torque_loss": 0.13624094426631927 }, { "epoch": 20.827338129496404, "grad_norm": 0.10001427680253983, "learning_rate": 6.015069830359714e-05, "loss": 0.0218, "step": 23160 }, { "action_loss": 0.006113726645708084, "epoch": 20.827338129496404, "step": 23160 }, { "epoch": 20.827338129496404, "step": 23160, "torque_loss": 0.12702864408493042 }, { "epoch": 20.836330935251798, "grad_norm": 0.1595495343208313, "learning_rate": 6.0118315331808384e-05, "loss": 0.0264, "step": 23170 }, { "action_loss": 0.01597784273326397, "epoch": 20.836330935251798, "step": 23170 }, { "epoch": 20.836330935251798, "step": 23170, "torque_loss": 0.22745056450366974 }, { "epoch": 20.845323741007196, "grad_norm": 0.15152300894260406, "learning_rate": 6.0085927933924745e-05, "loss": 0.0238, "step": 23180 }, { "action_loss": 0.00474746897816658, "epoch": 20.845323741007196, "step": 23180 }, { "epoch": 20.845323741007196, "step": 23180, "torque_loss": 0.1542697548866272 }, { "epoch": 20.85431654676259, "grad_norm": 0.10632342100143433, "learning_rate": 6.0053536124113585e-05, "loss": 0.0249, "step": 23190 }, { "action_loss": 0.004599768202751875, "epoch": 20.85431654676259, "step": 23190 }, { "epoch": 20.85431654676259, "step": 23190, "torque_loss": 0.1478145271539688 }, { "epoch": 20.863309352517987, "grad_norm": 0.12291907519102097, "learning_rate": 6.0021139916544144e-05, "loss": 0.0244, "step": 23200 }, { "action_loss": 0.0026483230758458376, "epoch": 20.863309352517987, "step": 23200 }, { "epoch": 20.863309352517987, "step": 23200, "torque_loss": 0.1592559516429901 }, { "epoch": 20.87230215827338, "grad_norm": 0.10727442055940628, "learning_rate": 5.998873932538764e-05, "loss": 0.0238, "step": 23210 }, { "action_loss": 0.012379147112369537, "epoch": 20.87230215827338, "step": 23210 }, { "epoch": 20.87230215827338, "step": 23210, "torque_loss": 0.14638763666152954 }, { "epoch": 20.881294964028775, "grad_norm": 0.08437880873680115, "learning_rate": 5.995633436481721e-05, "loss": 0.0214, "step": 23220 }, { "action_loss": 0.005857991054654121, "epoch": 20.881294964028775, "step": 23220 }, { "epoch": 20.881294964028775, "step": 23220, "torque_loss": 0.16527344286441803 }, { "epoch": 20.890287769784173, "grad_norm": 0.18109773099422455, "learning_rate": 5.9923925049007855e-05, "loss": 0.024, "step": 23230 }, { "action_loss": 0.005063416436314583, "epoch": 20.890287769784173, "step": 23230 }, { "epoch": 20.890287769784173, "step": 23230, "torque_loss": 0.16421402990818024 }, { "epoch": 20.899280575539567, "grad_norm": 0.12381438165903091, "learning_rate": 5.989151139213656e-05, "loss": 0.0271, "step": 23240 }, { "action_loss": 0.014426521956920624, "epoch": 20.899280575539567, "step": 23240 }, { "epoch": 20.899280575539567, "step": 23240, "torque_loss": 0.19015438854694366 }, { "epoch": 20.908273381294965, "grad_norm": 0.17024797201156616, "learning_rate": 5.9859093408382094e-05, "loss": 0.0226, "step": 23250 }, { "action_loss": 0.01918930374085903, "epoch": 20.908273381294965, "step": 23250 }, { "epoch": 20.908273381294965, "step": 23250, "torque_loss": 0.2839520275592804 }, { "epoch": 20.91726618705036, "grad_norm": 0.13499805331230164, "learning_rate": 5.982667111192523e-05, "loss": 0.0227, "step": 23260 }, { "action_loss": 0.003947723191231489, "epoch": 20.91726618705036, "step": 23260 }, { "epoch": 20.91726618705036, "step": 23260, "torque_loss": 0.12676756083965302 }, { "epoch": 20.926258992805757, "grad_norm": 0.15917271375656128, "learning_rate": 5.979424451694857e-05, "loss": 0.0303, "step": 23270 }, { "action_loss": 0.00267926137894392, "epoch": 20.926258992805757, "step": 23270 }, { "epoch": 20.926258992805757, "step": 23270, "torque_loss": 0.12172795087099075 }, { "epoch": 20.93525179856115, "grad_norm": 0.15289340913295746, "learning_rate": 5.97618136376366e-05, "loss": 0.0189, "step": 23280 }, { "action_loss": 0.0021075664553791285, "epoch": 20.93525179856115, "step": 23280 }, { "epoch": 20.93525179856115, "step": 23280, "torque_loss": 0.15689803659915924 }, { "epoch": 20.944244604316548, "grad_norm": 0.1833505779504776, "learning_rate": 5.97293784881757e-05, "loss": 0.0257, "step": 23290 }, { "action_loss": 0.0026568931061774492, "epoch": 20.944244604316548, "step": 23290 }, { "epoch": 20.944244604316548, "step": 23290, "torque_loss": 0.14128634333610535 }, { "epoch": 20.953237410071942, "grad_norm": 0.14904280006885529, "learning_rate": 5.969693908275411e-05, "loss": 0.0207, "step": 23300 }, { "action_loss": 0.008910269476473331, "epoch": 20.953237410071942, "step": 23300 }, { "epoch": 20.953237410071942, "step": 23300, "torque_loss": 0.21021194756031036 }, { "epoch": 20.96223021582734, "grad_norm": 0.1025594025850296, "learning_rate": 5.9664495435561916e-05, "loss": 0.0236, "step": 23310 }, { "action_loss": 0.00630685081705451, "epoch": 20.96223021582734, "step": 23310 }, { "epoch": 20.96223021582734, "step": 23310, "torque_loss": 0.1683763712644577 }, { "epoch": 20.971223021582734, "grad_norm": 0.1149328202009201, "learning_rate": 5.963204756079108e-05, "loss": 0.0202, "step": 23320 }, { "action_loss": 0.01018888596445322, "epoch": 20.971223021582734, "step": 23320 }, { "epoch": 20.971223021582734, "step": 23320, "torque_loss": 0.24541430175304413 }, { "epoch": 20.980215827338128, "grad_norm": 0.15490199625492096, "learning_rate": 5.959959547263538e-05, "loss": 0.0252, "step": 23330 }, { "action_loss": 0.002386584412306547, "epoch": 20.980215827338128, "step": 23330 }, { "epoch": 20.980215827338128, "step": 23330, "torque_loss": 0.17145191133022308 }, { "epoch": 20.989208633093526, "grad_norm": 0.15050609409809113, "learning_rate": 5.95671391852905e-05, "loss": 0.0262, "step": 23340 }, { "action_loss": 0.0029779269825667143, "epoch": 20.989208633093526, "step": 23340 }, { "epoch": 20.989208633093526, "step": 23340, "torque_loss": 0.15757562220096588 }, { "epoch": 20.99820143884892, "grad_norm": 0.15403766930103302, "learning_rate": 5.95346787129539e-05, "loss": 0.0216, "step": 23350 }, { "action_loss": 0.010426200926303864, "epoch": 20.99820143884892, "step": 23350 }, { "epoch": 20.99820143884892, "step": 23350, "torque_loss": 0.15847110748291016 }, { "epoch": 21.007194244604317, "grad_norm": 0.144560307264328, "learning_rate": 5.950221406982489e-05, "loss": 0.0205, "step": 23360 }, { "action_loss": 0.0048403069376945496, "epoch": 21.007194244604317, "step": 23360 }, { "epoch": 21.007194244604317, "step": 23360, "torque_loss": 0.173487588763237 }, { "epoch": 21.01618705035971, "grad_norm": 0.0981268361210823, "learning_rate": 5.9469745270104625e-05, "loss": 0.0231, "step": 23370 }, { "action_loss": 0.009549039416015148, "epoch": 21.01618705035971, "step": 23370 }, { "epoch": 21.01618705035971, "step": 23370, "torque_loss": 0.23567552864551544 }, { "epoch": 21.02517985611511, "grad_norm": 0.1625726968050003, "learning_rate": 5.9437272327996064e-05, "loss": 0.0255, "step": 23380 }, { "action_loss": 0.004539032932370901, "epoch": 21.02517985611511, "step": 23380 }, { "epoch": 21.02517985611511, "step": 23380, "torque_loss": 0.1672266274690628 }, { "epoch": 21.034172661870503, "grad_norm": 0.16199791431427002, "learning_rate": 5.940479525770395e-05, "loss": 0.0261, "step": 23390 }, { "action_loss": 0.0038268628995865583, "epoch": 21.034172661870503, "step": 23390 }, { "epoch": 21.034172661870503, "step": 23390, "torque_loss": 0.135786771774292 }, { "epoch": 21.0431654676259, "grad_norm": 0.12012537568807602, "learning_rate": 5.9372314073434876e-05, "loss": 0.0227, "step": 23400 }, { "action_loss": 0.011312887072563171, "epoch": 21.0431654676259, "step": 23400 }, { "epoch": 21.0431654676259, "step": 23400, "torque_loss": 0.22103101015090942 }, { "epoch": 21.052158273381295, "grad_norm": 0.1955961436033249, "learning_rate": 5.93398287893972e-05, "loss": 0.0316, "step": 23410 }, { "action_loss": 0.004007836803793907, "epoch": 21.052158273381295, "step": 23410 }, { "epoch": 21.052158273381295, "step": 23410, "torque_loss": 0.1624465435743332 }, { "epoch": 21.06115107913669, "grad_norm": 0.12047040462493896, "learning_rate": 5.930733941980111e-05, "loss": 0.024, "step": 23420 }, { "action_loss": 0.002375669777393341, "epoch": 21.06115107913669, "step": 23420 }, { "epoch": 21.06115107913669, "step": 23420, "torque_loss": 0.15578687191009521 }, { "epoch": 21.070143884892087, "grad_norm": 0.2559376358985901, "learning_rate": 5.927484597885855e-05, "loss": 0.0212, "step": 23430 }, { "action_loss": 0.005694929510354996, "epoch": 21.070143884892087, "step": 23430 }, { "epoch": 21.070143884892087, "step": 23430, "torque_loss": 0.1535157412290573 }, { "epoch": 21.07913669064748, "grad_norm": 0.17024393379688263, "learning_rate": 5.924234848078325e-05, "loss": 0.0229, "step": 23440 }, { "action_loss": 0.01118276733905077, "epoch": 21.07913669064748, "step": 23440 }, { "epoch": 21.07913669064748, "step": 23440, "torque_loss": 0.12892170250415802 }, { "epoch": 21.08812949640288, "grad_norm": 0.13820859789848328, "learning_rate": 5.9209846939790715e-05, "loss": 0.0215, "step": 23450 }, { "action_loss": 0.01937904953956604, "epoch": 21.08812949640288, "step": 23450 }, { "epoch": 21.08812949640288, "step": 23450, "torque_loss": 0.16467136144638062 }, { "epoch": 21.097122302158272, "grad_norm": 0.22963455319404602, "learning_rate": 5.917734137009824e-05, "loss": 0.0249, "step": 23460 }, { "action_loss": 0.006705934181809425, "epoch": 21.097122302158272, "step": 23460 }, { "epoch": 21.097122302158272, "step": 23460, "torque_loss": 0.15295736491680145 }, { "epoch": 21.10611510791367, "grad_norm": 0.17012079060077667, "learning_rate": 5.914483178592485e-05, "loss": 0.0241, "step": 23470 }, { "action_loss": 0.007287175860255957, "epoch": 21.10611510791367, "step": 23470 }, { "epoch": 21.10611510791367, "step": 23470, "torque_loss": 0.21325039863586426 }, { "epoch": 21.115107913669064, "grad_norm": 0.15567393600940704, "learning_rate": 5.911231820149136e-05, "loss": 0.0213, "step": 23480 }, { "action_loss": 0.025484154000878334, "epoch": 21.115107913669064, "step": 23480 }, { "epoch": 21.115107913669064, "step": 23480, "torque_loss": 0.20464001595973969 }, { "epoch": 21.12410071942446, "grad_norm": 0.16541145741939545, "learning_rate": 5.907980063102031e-05, "loss": 0.0244, "step": 23490 }, { "action_loss": 0.0025985215324908495, "epoch": 21.12410071942446, "step": 23490 }, { "epoch": 21.12410071942446, "step": 23490, "torque_loss": 0.16275236010551453 }, { "epoch": 21.133093525179856, "grad_norm": 0.12449093908071518, "learning_rate": 5.904727908873597e-05, "loss": 0.0234, "step": 23500 }, { "action_loss": 0.007382199168205261, "epoch": 21.133093525179856, "step": 23500 }, { "epoch": 21.133093525179856, "step": 23500, "torque_loss": 0.1535424143075943 }, { "epoch": 21.142086330935253, "grad_norm": 0.13892626762390137, "learning_rate": 5.90147535888644e-05, "loss": 0.0259, "step": 23510 }, { "action_loss": 0.0022944873198866844, "epoch": 21.142086330935253, "step": 23510 }, { "epoch": 21.142086330935253, "step": 23510, "torque_loss": 0.10345756262540817 }, { "epoch": 21.151079136690647, "grad_norm": 0.18563874065876007, "learning_rate": 5.898222414563333e-05, "loss": 0.0239, "step": 23520 }, { "action_loss": 0.0027720818761736155, "epoch": 21.151079136690647, "step": 23520 }, { "epoch": 21.151079136690647, "step": 23520, "torque_loss": 0.1293373703956604 }, { "epoch": 21.16007194244604, "grad_norm": 0.12266092747449875, "learning_rate": 5.8949690773272256e-05, "loss": 0.0182, "step": 23530 }, { "action_loss": 0.003752313554286957, "epoch": 21.16007194244604, "step": 23530 }, { "epoch": 21.16007194244604, "step": 23530, "torque_loss": 0.17867903411388397 }, { "epoch": 21.16906474820144, "grad_norm": 0.12894463539123535, "learning_rate": 5.891715348601239e-05, "loss": 0.0228, "step": 23540 }, { "action_loss": 0.00586392916738987, "epoch": 21.16906474820144, "step": 23540 }, { "epoch": 21.16906474820144, "step": 23540, "torque_loss": 0.21285806596279144 }, { "epoch": 21.178057553956833, "grad_norm": 0.14364883303642273, "learning_rate": 5.888461229808663e-05, "loss": 0.0244, "step": 23550 }, { "action_loss": 0.00298278103582561, "epoch": 21.178057553956833, "step": 23550 }, { "epoch": 21.178057553956833, "step": 23550, "torque_loss": 0.16807322204113007 }, { "epoch": 21.18705035971223, "grad_norm": 0.09249569475650787, "learning_rate": 5.885206722372959e-05, "loss": 0.019, "step": 23560 }, { "action_loss": 0.0023651099763810635, "epoch": 21.18705035971223, "step": 23560 }, { "epoch": 21.18705035971223, "step": 23560, "torque_loss": 0.10341819375753403 }, { "epoch": 21.196043165467625, "grad_norm": 0.19515012204647064, "learning_rate": 5.881951827717763e-05, "loss": 0.0246, "step": 23570 }, { "action_loss": 0.0035823441576212645, "epoch": 21.196043165467625, "step": 23570 }, { "epoch": 21.196043165467625, "step": 23570, "torque_loss": 0.16291171312332153 }, { "epoch": 21.205035971223023, "grad_norm": 0.17994627356529236, "learning_rate": 5.878696547266872e-05, "loss": 0.0254, "step": 23580 }, { "action_loss": 0.0045920745469629765, "epoch": 21.205035971223023, "step": 23580 }, { "epoch": 21.205035971223023, "step": 23580, "torque_loss": 0.12688709795475006 }, { "epoch": 21.214028776978417, "grad_norm": 0.10504645109176636, "learning_rate": 5.875440882444259e-05, "loss": 0.0257, "step": 23590 }, { "action_loss": 0.010038464330136776, "epoch": 21.214028776978417, "step": 23590 }, { "epoch": 21.214028776978417, "step": 23590, "torque_loss": 0.189666286110878 }, { "epoch": 21.223021582733814, "grad_norm": 0.14001160860061646, "learning_rate": 5.8721848346740617e-05, "loss": 0.0206, "step": 23600 }, { "action_loss": 0.0019722983706742525, "epoch": 21.223021582733814, "step": 23600 }, { "epoch": 21.223021582733814, "step": 23600, "torque_loss": 0.0747261568903923 }, { "epoch": 21.23201438848921, "grad_norm": 0.111606165766716, "learning_rate": 5.868928405380585e-05, "loss": 0.0194, "step": 23610 }, { "action_loss": 0.006870871875435114, "epoch": 21.23201438848921, "step": 23610 }, { "epoch": 21.23201438848921, "step": 23610, "torque_loss": 0.1412166804075241 }, { "epoch": 21.241007194244606, "grad_norm": 0.17168965935707092, "learning_rate": 5.865671595988304e-05, "loss": 0.0239, "step": 23620 }, { "action_loss": 0.00394378462806344, "epoch": 21.241007194244606, "step": 23620 }, { "epoch": 21.241007194244606, "step": 23620, "torque_loss": 0.14852845668792725 }, { "epoch": 21.25, "grad_norm": 0.18136465549468994, "learning_rate": 5.862414407921857e-05, "loss": 0.0265, "step": 23630 }, { "action_loss": 0.0028832333628088236, "epoch": 21.25, "step": 23630 }, { "epoch": 21.25, "step": 23630, "torque_loss": 0.07935059815645218 }, { "epoch": 21.258992805755394, "grad_norm": 0.12265216559171677, "learning_rate": 5.8591568426060475e-05, "loss": 0.0238, "step": 23640 }, { "action_loss": 0.01620212011039257, "epoch": 21.258992805755394, "step": 23640 }, { "epoch": 21.258992805755394, "step": 23640, "torque_loss": 0.1667066216468811 }, { "epoch": 21.26798561151079, "grad_norm": 0.13068757951259613, "learning_rate": 5.8558989014658464e-05, "loss": 0.0249, "step": 23650 }, { "action_loss": 0.010519745759665966, "epoch": 21.26798561151079, "step": 23650 }, { "epoch": 21.26798561151079, "step": 23650, "torque_loss": 0.19620545208454132 }, { "epoch": 21.276978417266186, "grad_norm": 0.15888257324695587, "learning_rate": 5.8526405859263855e-05, "loss": 0.0232, "step": 23660 }, { "action_loss": 0.005383369047194719, "epoch": 21.276978417266186, "step": 23660 }, { "epoch": 21.276978417266186, "step": 23660, "torque_loss": 0.22165946662425995 }, { "epoch": 21.285971223021583, "grad_norm": 0.20768994092941284, "learning_rate": 5.8493818974129667e-05, "loss": 0.0208, "step": 23670 }, { "action_loss": 0.002960286336019635, "epoch": 21.285971223021583, "step": 23670 }, { "epoch": 21.285971223021583, "step": 23670, "torque_loss": 0.14397163689136505 }, { "epoch": 21.294964028776977, "grad_norm": 0.16980376839637756, "learning_rate": 5.846122837351048e-05, "loss": 0.0235, "step": 23680 }, { "action_loss": 0.0036443686112761497, "epoch": 21.294964028776977, "step": 23680 }, { "epoch": 21.294964028776977, "step": 23680, "torque_loss": 0.18063108623027802 }, { "epoch": 21.303956834532375, "grad_norm": 0.09694986045360565, "learning_rate": 5.842863407166255e-05, "loss": 0.0283, "step": 23690 }, { "action_loss": 0.004560304805636406, "epoch": 21.303956834532375, "step": 23690 }, { "epoch": 21.303956834532375, "step": 23690, "torque_loss": 0.14390456676483154 }, { "epoch": 21.31294964028777, "grad_norm": 0.10198498517274857, "learning_rate": 5.83960360828437e-05, "loss": 0.0218, "step": 23700 }, { "action_loss": 0.0025603699032217264, "epoch": 21.31294964028777, "step": 23700 }, { "epoch": 21.31294964028777, "step": 23700, "torque_loss": 0.15129895508289337 }, { "epoch": 21.321942446043167, "grad_norm": 0.15937840938568115, "learning_rate": 5.836343442131343e-05, "loss": 0.027, "step": 23710 }, { "action_loss": 0.0019201274262741208, "epoch": 21.321942446043167, "step": 23710 }, { "epoch": 21.321942446043167, "step": 23710, "torque_loss": 0.1300998479127884 }, { "epoch": 21.33093525179856, "grad_norm": 0.16774845123291016, "learning_rate": 5.833082910133281e-05, "loss": 0.0276, "step": 23720 }, { "action_loss": 0.00445624440908432, "epoch": 21.33093525179856, "step": 23720 }, { "epoch": 21.33093525179856, "step": 23720, "torque_loss": 0.1443278044462204 }, { "epoch": 21.33992805755396, "grad_norm": 0.15537510812282562, "learning_rate": 5.82982201371645e-05, "loss": 0.027, "step": 23730 }, { "action_loss": 0.002993350848555565, "epoch": 21.33992805755396, "step": 23730 }, { "epoch": 21.33992805755396, "step": 23730, "torque_loss": 0.11670312285423279 }, { "epoch": 21.348920863309353, "grad_norm": 0.18917757272720337, "learning_rate": 5.826560754307279e-05, "loss": 0.027, "step": 23740 }, { "action_loss": 0.003800373524427414, "epoch": 21.348920863309353, "step": 23740 }, { "epoch": 21.348920863309353, "step": 23740, "torque_loss": 0.13696976006031036 }, { "epoch": 21.357913669064747, "grad_norm": 0.1193547397851944, "learning_rate": 5.8232991333323516e-05, "loss": 0.0218, "step": 23750 }, { "action_loss": 0.004455262795090675, "epoch": 21.357913669064747, "step": 23750 }, { "epoch": 21.357913669064747, "step": 23750, "torque_loss": 0.1618349403142929 }, { "epoch": 21.366906474820144, "grad_norm": 0.20883043110370636, "learning_rate": 5.8200371522184116e-05, "loss": 0.022, "step": 23760 }, { "action_loss": 0.003306623548269272, "epoch": 21.366906474820144, "step": 23760 }, { "epoch": 21.366906474820144, "step": 23760, "torque_loss": 0.14213980734348297 }, { "epoch": 21.37589928057554, "grad_norm": 0.14434406161308289, "learning_rate": 5.816774812392362e-05, "loss": 0.0209, "step": 23770 }, { "action_loss": 0.009905329905450344, "epoch": 21.37589928057554, "step": 23770 }, { "epoch": 21.37589928057554, "step": 23770, "torque_loss": 0.1547560691833496 }, { "epoch": 21.384892086330936, "grad_norm": 0.15602831542491913, "learning_rate": 5.8135121152812614e-05, "loss": 0.0265, "step": 23780 }, { "action_loss": 0.00550634041428566, "epoch": 21.384892086330936, "step": 23780 }, { "epoch": 21.384892086330936, "step": 23780, "torque_loss": 0.134693905711174 }, { "epoch": 21.39388489208633, "grad_norm": 0.25151991844177246, "learning_rate": 5.8102490623123204e-05, "loss": 0.0262, "step": 23790 }, { "action_loss": 0.002505992306396365, "epoch": 21.39388489208633, "step": 23790 }, { "epoch": 21.39388489208633, "step": 23790, "torque_loss": 0.14255641400814056 }, { "epoch": 21.402877697841728, "grad_norm": 0.13087087869644165, "learning_rate": 5.806985654912914e-05, "loss": 0.0284, "step": 23800 }, { "action_loss": 0.0022140059154480696, "epoch": 21.402877697841728, "step": 23800 }, { "epoch": 21.402877697841728, "step": 23800, "torque_loss": 0.12839670479297638 }, { "epoch": 21.41187050359712, "grad_norm": 0.20188204944133759, "learning_rate": 5.8037218945105644e-05, "loss": 0.0224, "step": 23810 }, { "action_loss": 0.010439674369990826, "epoch": 21.41187050359712, "step": 23810 }, { "epoch": 21.41187050359712, "step": 23810, "torque_loss": 0.1830003708600998 }, { "epoch": 21.42086330935252, "grad_norm": 0.11154816299676895, "learning_rate": 5.800457782532953e-05, "loss": 0.0255, "step": 23820 }, { "action_loss": 0.0038856479804962873, "epoch": 21.42086330935252, "step": 23820 }, { "epoch": 21.42086330935252, "step": 23820, "torque_loss": 0.17679722607135773 }, { "epoch": 21.429856115107913, "grad_norm": 0.11156576126813889, "learning_rate": 5.797193320407912e-05, "loss": 0.0214, "step": 23830 }, { "action_loss": 0.0042680236510932446, "epoch": 21.429856115107913, "step": 23830 }, { "epoch": 21.429856115107913, "step": 23830, "torque_loss": 0.15654172003269196 }, { "epoch": 21.43884892086331, "grad_norm": 0.14647425711154938, "learning_rate": 5.793928509563431e-05, "loss": 0.0235, "step": 23840 }, { "action_loss": 0.005311912391334772, "epoch": 21.43884892086331, "step": 23840 }, { "epoch": 21.43884892086331, "step": 23840, "torque_loss": 0.11826664209365845 }, { "epoch": 21.447841726618705, "grad_norm": 0.09505723416805267, "learning_rate": 5.790663351427644e-05, "loss": 0.0198, "step": 23850 }, { "action_loss": 0.005293553229421377, "epoch": 21.447841726618705, "step": 23850 }, { "epoch": 21.447841726618705, "step": 23850, "torque_loss": 0.15332360565662384 }, { "epoch": 21.4568345323741, "grad_norm": 0.13271242380142212, "learning_rate": 5.7873978474288484e-05, "loss": 0.024, "step": 23860 }, { "action_loss": 0.009559138678014278, "epoch": 21.4568345323741, "step": 23860 }, { "epoch": 21.4568345323741, "step": 23860, "torque_loss": 0.16174830496311188 }, { "epoch": 21.465827338129497, "grad_norm": 0.14878560602664948, "learning_rate": 5.78413199899548e-05, "loss": 0.0268, "step": 23870 }, { "action_loss": 0.004546660464257002, "epoch": 21.465827338129497, "step": 23870 }, { "epoch": 21.465827338129497, "step": 23870, "torque_loss": 0.1669650822877884 }, { "epoch": 21.47482014388489, "grad_norm": 0.13853363692760468, "learning_rate": 5.780865807556136e-05, "loss": 0.0236, "step": 23880 }, { "action_loss": 0.010708369314670563, "epoch": 21.47482014388489, "step": 23880 }, { "epoch": 21.47482014388489, "step": 23880, "torque_loss": 0.2143252044916153 }, { "epoch": 21.48381294964029, "grad_norm": 0.14068253338336945, "learning_rate": 5.777599274539559e-05, "loss": 0.0277, "step": 23890 }, { "action_loss": 0.004271382000297308, "epoch": 21.48381294964029, "step": 23890 }, { "epoch": 21.48381294964029, "step": 23890, "torque_loss": 0.17225880920886993 }, { "epoch": 21.492805755395683, "grad_norm": 0.1599103808403015, "learning_rate": 5.77433240137464e-05, "loss": 0.0213, "step": 23900 }, { "action_loss": 0.007970544509589672, "epoch": 21.492805755395683, "step": 23900 }, { "epoch": 21.492805755395683, "step": 23900, "torque_loss": 0.2105865329504013 }, { "epoch": 21.50179856115108, "grad_norm": 0.19597604870796204, "learning_rate": 5.771065189490422e-05, "loss": 0.023, "step": 23910 }, { "action_loss": 0.005685098003596067, "epoch": 21.50179856115108, "step": 23910 }, { "epoch": 21.50179856115108, "step": 23910, "torque_loss": 0.17952638864517212 }, { "epoch": 21.510791366906474, "grad_norm": 0.15390437841415405, "learning_rate": 5.767797640316095e-05, "loss": 0.0223, "step": 23920 }, { "action_loss": 0.05594316124916077, "epoch": 21.510791366906474, "step": 23920 }, { "epoch": 21.510791366906474, "step": 23920, "torque_loss": 0.21778249740600586 }, { "epoch": 21.519784172661872, "grad_norm": 0.1484500914812088, "learning_rate": 5.764529755280994e-05, "loss": 0.0302, "step": 23930 }, { "action_loss": 0.007375185843557119, "epoch": 21.519784172661872, "step": 23930 }, { "epoch": 21.519784172661872, "step": 23930, "torque_loss": 0.1689317226409912 }, { "epoch": 21.528776978417266, "grad_norm": 0.14435017108917236, "learning_rate": 5.761261535814603e-05, "loss": 0.0204, "step": 23940 }, { "action_loss": 0.015085329301655293, "epoch": 21.528776978417266, "step": 23940 }, { "epoch": 21.528776978417266, "step": 23940, "torque_loss": 0.1964707225561142 }, { "epoch": 21.53776978417266, "grad_norm": 0.1598413735628128, "learning_rate": 5.757992983346554e-05, "loss": 0.0223, "step": 23950 }, { "action_loss": 0.0020660681184381247, "epoch": 21.53776978417266, "step": 23950 }, { "epoch": 21.53776978417266, "step": 23950, "torque_loss": 0.15529726445674896 }, { "epoch": 21.546762589928058, "grad_norm": 0.2100176066160202, "learning_rate": 5.754724099306621e-05, "loss": 0.0242, "step": 23960 }, { "action_loss": 0.0041373781859874725, "epoch": 21.546762589928058, "step": 23960 }, { "epoch": 21.546762589928058, "step": 23960, "torque_loss": 0.13742339611053467 }, { "epoch": 21.555755395683452, "grad_norm": 0.17785467207431793, "learning_rate": 5.751454885124727e-05, "loss": 0.0209, "step": 23970 }, { "action_loss": 0.002887089503929019, "epoch": 21.555755395683452, "step": 23970 }, { "epoch": 21.555755395683452, "step": 23970, "torque_loss": 0.17301416397094727 }, { "epoch": 21.56474820143885, "grad_norm": 0.14950363337993622, "learning_rate": 5.748185342230935e-05, "loss": 0.0238, "step": 23980 }, { "action_loss": 0.007823565043509007, "epoch": 21.56474820143885, "step": 23980 }, { "epoch": 21.56474820143885, "step": 23980, "torque_loss": 0.19255630671977997 }, { "epoch": 21.573741007194243, "grad_norm": 0.1451520472764969, "learning_rate": 5.744915472055456e-05, "loss": 0.0209, "step": 23990 }, { "action_loss": 0.009865554980933666, "epoch": 21.573741007194243, "step": 23990 }, { "epoch": 21.573741007194243, "step": 23990, "torque_loss": 0.24624872207641602 }, { "epoch": 21.58273381294964, "grad_norm": 0.10126594454050064, "learning_rate": 5.741645276028642e-05, "loss": 0.0245, "step": 24000 }, { "action_loss": 0.003991777077317238, "epoch": 21.58273381294964, "step": 24000 }, { "epoch": 21.58273381294964, "step": 24000, "torque_loss": 0.17960144579410553 }, { "epoch": 21.591726618705035, "grad_norm": 0.15734632313251495, "learning_rate": 5.7383747555809873e-05, "loss": 0.0262, "step": 24010 }, { "action_loss": 0.0025603780522942543, "epoch": 21.591726618705035, "step": 24010 }, { "epoch": 21.591726618705035, "step": 24010, "torque_loss": 0.1340937465429306 }, { "epoch": 21.600719424460433, "grad_norm": 0.1295587122440338, "learning_rate": 5.7351039121431294e-05, "loss": 0.0229, "step": 24020 }, { "action_loss": 0.0024284187238663435, "epoch": 21.600719424460433, "step": 24020 }, { "epoch": 21.600719424460433, "step": 24020, "torque_loss": 0.1478341668844223 }, { "epoch": 21.609712230215827, "grad_norm": 0.17668791115283966, "learning_rate": 5.7318327471458454e-05, "loss": 0.0246, "step": 24030 }, { "action_loss": 0.018357977271080017, "epoch": 21.609712230215827, "step": 24030 }, { "epoch": 21.609712230215827, "step": 24030, "torque_loss": 0.222134068608284 }, { "epoch": 21.618705035971225, "grad_norm": 0.11191485077142715, "learning_rate": 5.728561262020054e-05, "loss": 0.028, "step": 24040 }, { "action_loss": 0.006211543921381235, "epoch": 21.618705035971225, "step": 24040 }, { "epoch": 21.618705035971225, "step": 24040, "torque_loss": 0.16402579843997955 }, { "epoch": 21.62769784172662, "grad_norm": 0.15158860385417938, "learning_rate": 5.725289458196814e-05, "loss": 0.0242, "step": 24050 }, { "action_loss": 0.011757668107748032, "epoch": 21.62769784172662, "step": 24050 }, { "epoch": 21.62769784172662, "step": 24050, "torque_loss": 0.20744411647319794 }, { "epoch": 21.636690647482013, "grad_norm": 0.16050173342227936, "learning_rate": 5.722017337107324e-05, "loss": 0.0273, "step": 24060 }, { "action_loss": 0.00793642457574606, "epoch": 21.636690647482013, "step": 24060 }, { "epoch": 21.636690647482013, "step": 24060, "torque_loss": 0.17719906568527222 }, { "epoch": 21.64568345323741, "grad_norm": 0.11381406337022781, "learning_rate": 5.7187449001829206e-05, "loss": 0.0214, "step": 24070 }, { "action_loss": 0.004795247223228216, "epoch": 21.64568345323741, "step": 24070 }, { "epoch": 21.64568345323741, "step": 24070, "torque_loss": 0.17630831897258759 }, { "epoch": 21.654676258992804, "grad_norm": 0.1811693012714386, "learning_rate": 5.715472148855081e-05, "loss": 0.0253, "step": 24080 }, { "action_loss": 0.011160480789840221, "epoch": 21.654676258992804, "step": 24080 }, { "epoch": 21.654676258992804, "step": 24080, "torque_loss": 0.22756077349185944 }, { "epoch": 21.663669064748202, "grad_norm": 0.15784014761447906, "learning_rate": 5.712199084555414e-05, "loss": 0.0272, "step": 24090 }, { "action_loss": 0.0028677741065621376, "epoch": 21.663669064748202, "step": 24090 }, { "epoch": 21.663669064748202, "step": 24090, "torque_loss": 0.1411949098110199 }, { "epoch": 21.672661870503596, "grad_norm": 0.20205210149288177, "learning_rate": 5.7089257087156734e-05, "loss": 0.0253, "step": 24100 }, { "action_loss": 0.0038904219400137663, "epoch": 21.672661870503596, "step": 24100 }, { "epoch": 21.672661870503596, "step": 24100, "torque_loss": 0.15724597871303558 }, { "epoch": 21.681654676258994, "grad_norm": 0.12250746041536331, "learning_rate": 5.70565202276774e-05, "loss": 0.0189, "step": 24110 }, { "action_loss": 0.00944593921303749, "epoch": 21.681654676258994, "step": 24110 }, { "epoch": 21.681654676258994, "step": 24110, "torque_loss": 0.17082113027572632 }, { "epoch": 21.690647482014388, "grad_norm": 0.2273031622171402, "learning_rate": 5.70237802814364e-05, "loss": 0.0314, "step": 24120 }, { "action_loss": 0.00680103525519371, "epoch": 21.690647482014388, "step": 24120 }, { "epoch": 21.690647482014388, "step": 24120, "torque_loss": 0.1557566374540329 }, { "epoch": 21.699640287769785, "grad_norm": 0.14771486818790436, "learning_rate": 5.6991037262755277e-05, "loss": 0.027, "step": 24130 }, { "action_loss": 0.00311366468667984, "epoch": 21.699640287769785, "step": 24130 }, { "epoch": 21.699640287769785, "step": 24130, "torque_loss": 0.09529444575309753 }, { "epoch": 21.70863309352518, "grad_norm": 0.19200846552848816, "learning_rate": 5.6958291185956924e-05, "loss": 0.0222, "step": 24140 }, { "action_loss": 0.0033917620312422514, "epoch": 21.70863309352518, "step": 24140 }, { "epoch": 21.70863309352518, "step": 24140, "torque_loss": 0.17136645317077637 }, { "epoch": 21.717625899280577, "grad_norm": 0.19615669548511505, "learning_rate": 5.692554206536561e-05, "loss": 0.0211, "step": 24150 }, { "action_loss": 0.0028456116560846567, "epoch": 21.717625899280577, "step": 24150 }, { "epoch": 21.717625899280577, "step": 24150, "torque_loss": 0.13093197345733643 }, { "epoch": 21.72661870503597, "grad_norm": 0.19161124527454376, "learning_rate": 5.689278991530692e-05, "loss": 0.0238, "step": 24160 }, { "action_loss": 0.002549146767705679, "epoch": 21.72661870503597, "step": 24160 }, { "epoch": 21.72661870503597, "step": 24160, "torque_loss": 0.13401739299297333 }, { "epoch": 21.735611510791365, "grad_norm": 0.15796969830989838, "learning_rate": 5.686003475010773e-05, "loss": 0.0195, "step": 24170 }, { "action_loss": 0.0029932884499430656, "epoch": 21.735611510791365, "step": 24170 }, { "epoch": 21.735611510791365, "step": 24170, "torque_loss": 0.11494358628988266 }, { "epoch": 21.744604316546763, "grad_norm": 0.0853123739361763, "learning_rate": 5.682727658409628e-05, "loss": 0.0198, "step": 24180 }, { "action_loss": 0.006542128045111895, "epoch": 21.744604316546763, "step": 24180 }, { "epoch": 21.744604316546763, "step": 24180, "torque_loss": 0.2032637745141983 }, { "epoch": 21.753597122302157, "grad_norm": 0.13528156280517578, "learning_rate": 5.67945154316021e-05, "loss": 0.0264, "step": 24190 }, { "action_loss": 0.024319661781191826, "epoch": 21.753597122302157, "step": 24190 }, { "epoch": 21.753597122302157, "step": 24190, "torque_loss": 0.2149476259946823 }, { "epoch": 21.762589928057555, "grad_norm": 0.1339154690504074, "learning_rate": 5.676175130695602e-05, "loss": 0.0232, "step": 24200 }, { "action_loss": 0.0016722161090001464, "epoch": 21.762589928057555, "step": 24200 }, { "epoch": 21.762589928057555, "step": 24200, "torque_loss": 0.10185282677412033 }, { "epoch": 21.77158273381295, "grad_norm": 0.12566420435905457, "learning_rate": 5.672898422449022e-05, "loss": 0.0218, "step": 24210 }, { "action_loss": 0.00481782341375947, "epoch": 21.77158273381295, "step": 24210 }, { "epoch": 21.77158273381295, "step": 24210, "torque_loss": 0.18417923152446747 }, { "epoch": 21.780575539568346, "grad_norm": 0.11434269696474075, "learning_rate": 5.66962141985381e-05, "loss": 0.0227, "step": 24220 }, { "action_loss": 0.0074643041007220745, "epoch": 21.780575539568346, "step": 24220 }, { "epoch": 21.780575539568346, "step": 24220, "torque_loss": 0.17827598750591278 }, { "epoch": 21.78956834532374, "grad_norm": 0.15639279782772064, "learning_rate": 5.666344124343439e-05, "loss": 0.023, "step": 24230 }, { "action_loss": 0.002923124236986041, "epoch": 21.78956834532374, "step": 24230 }, { "epoch": 21.78956834532374, "step": 24230, "torque_loss": 0.18402965366840363 }, { "epoch": 21.798561151079138, "grad_norm": 0.13211524486541748, "learning_rate": 5.66306653735151e-05, "loss": 0.0269, "step": 24240 }, { "action_loss": 0.00227209716103971, "epoch": 21.798561151079138, "step": 24240 }, { "epoch": 21.798561151079138, "step": 24240, "torque_loss": 0.1522873193025589 }, { "epoch": 21.807553956834532, "grad_norm": 0.1728944182395935, "learning_rate": 5.659788660311749e-05, "loss": 0.0226, "step": 24250 }, { "action_loss": 0.006192414555698633, "epoch": 21.807553956834532, "step": 24250 }, { "epoch": 21.807553956834532, "step": 24250, "torque_loss": 0.17843294143676758 }, { "epoch": 21.81654676258993, "grad_norm": 0.1931765377521515, "learning_rate": 5.656510494658014e-05, "loss": 0.0237, "step": 24260 }, { "action_loss": 0.006392145063728094, "epoch": 21.81654676258993, "step": 24260 }, { "epoch": 21.81654676258993, "step": 24260, "torque_loss": 0.18986207246780396 }, { "epoch": 21.825539568345324, "grad_norm": 0.1137361079454422, "learning_rate": 5.6532320418242835e-05, "loss": 0.0231, "step": 24270 }, { "action_loss": 0.0077096507884562016, "epoch": 21.825539568345324, "step": 24270 }, { "epoch": 21.825539568345324, "step": 24270, "torque_loss": 0.21814382076263428 }, { "epoch": 21.834532374100718, "grad_norm": 0.11662101000547409, "learning_rate": 5.649953303244665e-05, "loss": 0.0234, "step": 24280 }, { "action_loss": 0.005463178735226393, "epoch": 21.834532374100718, "step": 24280 }, { "epoch": 21.834532374100718, "step": 24280, "torque_loss": 0.13920438289642334 }, { "epoch": 21.843525179856115, "grad_norm": 0.17978104948997498, "learning_rate": 5.646674280353389e-05, "loss": 0.0269, "step": 24290 }, { "action_loss": 0.01454926747828722, "epoch": 21.843525179856115, "step": 24290 }, { "epoch": 21.843525179856115, "step": 24290, "torque_loss": 0.22819215059280396 }, { "epoch": 21.85251798561151, "grad_norm": 0.16362853348255157, "learning_rate": 5.6433949745848135e-05, "loss": 0.0202, "step": 24300 }, { "action_loss": 0.007829271256923676, "epoch": 21.85251798561151, "step": 24300 }, { "epoch": 21.85251798561151, "step": 24300, "torque_loss": 0.15054833889007568 }, { "epoch": 21.861510791366907, "grad_norm": 0.11148873716592789, "learning_rate": 5.6401153873734166e-05, "loss": 0.0366, "step": 24310 }, { "action_loss": 0.0073571596294641495, "epoch": 21.861510791366907, "step": 24310 }, { "epoch": 21.861510791366907, "step": 24310, "torque_loss": 0.2287823110818863 }, { "epoch": 21.8705035971223, "grad_norm": 0.1410575807094574, "learning_rate": 5.6368355201538036e-05, "loss": 0.0225, "step": 24320 }, { "action_loss": 0.007226552814245224, "epoch": 21.8705035971223, "step": 24320 }, { "epoch": 21.8705035971223, "step": 24320, "torque_loss": 0.2107999175786972 }, { "epoch": 21.8794964028777, "grad_norm": 0.15885011851787567, "learning_rate": 5.633555374360697e-05, "loss": 0.0229, "step": 24330 }, { "action_loss": 0.004065298475325108, "epoch": 21.8794964028777, "step": 24330 }, { "epoch": 21.8794964028777, "step": 24330, "torque_loss": 0.1676393300294876 }, { "epoch": 21.888489208633093, "grad_norm": 0.09202226996421814, "learning_rate": 5.630274951428944e-05, "loss": 0.023, "step": 24340 }, { "action_loss": 0.010539223439991474, "epoch": 21.888489208633093, "step": 24340 }, { "epoch": 21.888489208633093, "step": 24340, "torque_loss": 0.24180348217487335 }, { "epoch": 21.89748201438849, "grad_norm": 0.14649586379528046, "learning_rate": 5.626994252793517e-05, "loss": 0.0269, "step": 24350 }, { "action_loss": 0.00533549627289176, "epoch": 21.89748201438849, "step": 24350 }, { "epoch": 21.89748201438849, "step": 24350, "torque_loss": 0.15455184876918793 }, { "epoch": 21.906474820143885, "grad_norm": 0.12330912798643112, "learning_rate": 5.6237132798895006e-05, "loss": 0.0263, "step": 24360 }, { "action_loss": 0.027584699913859367, "epoch": 21.906474820143885, "step": 24360 }, { "epoch": 21.906474820143885, "step": 24360, "torque_loss": 0.24795083701610565 }, { "epoch": 21.915467625899282, "grad_norm": 0.2375085949897766, "learning_rate": 5.620432034152106e-05, "loss": 0.0319, "step": 24370 }, { "action_loss": 0.005914218258112669, "epoch": 21.915467625899282, "step": 24370 }, { "epoch": 21.915467625899282, "step": 24370, "torque_loss": 0.16908098757266998 }, { "epoch": 21.924460431654676, "grad_norm": 0.2259691059589386, "learning_rate": 5.6171505170166604e-05, "loss": 0.0243, "step": 24380 }, { "action_loss": 0.003288605948910117, "epoch": 21.924460431654676, "step": 24380 }, { "epoch": 21.924460431654676, "step": 24380, "torque_loss": 0.18969941139221191 }, { "epoch": 21.93345323741007, "grad_norm": 0.13166023790836334, "learning_rate": 5.613868729918612e-05, "loss": 0.0205, "step": 24390 }, { "action_loss": 0.0044005634263157845, "epoch": 21.93345323741007, "step": 24390 }, { "epoch": 21.93345323741007, "step": 24390, "torque_loss": 0.16333670914173126 }, { "epoch": 21.942446043165468, "grad_norm": 0.16918525099754333, "learning_rate": 5.6105866742935256e-05, "loss": 0.0238, "step": 24400 }, { "action_loss": 0.004440667573362589, "epoch": 21.942446043165468, "step": 24400 }, { "epoch": 21.942446043165468, "step": 24400, "torque_loss": 0.1293322890996933 }, { "epoch": 21.951438848920862, "grad_norm": 0.12962043285369873, "learning_rate": 5.607304351577083e-05, "loss": 0.0252, "step": 24410 }, { "action_loss": 0.009187176823616028, "epoch": 21.951438848920862, "step": 24410 }, { "epoch": 21.951438848920862, "step": 24410, "torque_loss": 0.12588687241077423 }, { "epoch": 21.96043165467626, "grad_norm": 0.18044984340667725, "learning_rate": 5.604021763205085e-05, "loss": 0.0283, "step": 24420 }, { "action_loss": 0.00696048466488719, "epoch": 21.96043165467626, "step": 24420 }, { "epoch": 21.96043165467626, "step": 24420, "torque_loss": 0.1588822603225708 }, { "epoch": 21.969424460431654, "grad_norm": 0.17786149680614471, "learning_rate": 5.600738910613445e-05, "loss": 0.0242, "step": 24430 }, { "action_loss": 0.0031557241454720497, "epoch": 21.969424460431654, "step": 24430 }, { "epoch": 21.969424460431654, "step": 24430, "torque_loss": 0.11153527349233627 }, { "epoch": 21.97841726618705, "grad_norm": 0.15278500318527222, "learning_rate": 5.597455795238198e-05, "loss": 0.024, "step": 24440 }, { "action_loss": 0.006267073098570108, "epoch": 21.97841726618705, "step": 24440 }, { "epoch": 21.97841726618705, "step": 24440, "torque_loss": 0.21322579681873322 }, { "epoch": 21.987410071942445, "grad_norm": 0.1849873960018158, "learning_rate": 5.5941724185154876e-05, "loss": 0.0255, "step": 24450 }, { "action_loss": 0.0016437700251117349, "epoch": 21.987410071942445, "step": 24450 }, { "epoch": 21.987410071942445, "step": 24450, "torque_loss": 0.10084164887666702 }, { "epoch": 21.996402877697843, "grad_norm": 0.17235422134399414, "learning_rate": 5.5908887818815746e-05, "loss": 0.0305, "step": 24460 }, { "action_loss": 0.0028640853706747293, "epoch": 21.996402877697843, "step": 24460 }, { "epoch": 21.996402877697843, "step": 24460, "torque_loss": 0.19432811439037323 }, { "epoch": 22.005395683453237, "grad_norm": 0.15964937210083008, "learning_rate": 5.587604886772834e-05, "loss": 0.022, "step": 24470 }, { "action_loss": 0.005022621247917414, "epoch": 22.005395683453237, "step": 24470 }, { "epoch": 22.005395683453237, "step": 24470, "torque_loss": 0.17137236893177032 }, { "epoch": 22.014388489208635, "grad_norm": 0.22207622230052948, "learning_rate": 5.584320734625752e-05, "loss": 0.0233, "step": 24480 }, { "action_loss": 0.004546090494841337, "epoch": 22.014388489208635, "step": 24480 }, { "epoch": 22.014388489208635, "step": 24480, "torque_loss": 0.1996380090713501 }, { "epoch": 22.02338129496403, "grad_norm": 0.14097072184085846, "learning_rate": 5.581036326876928e-05, "loss": 0.0214, "step": 24490 }, { "action_loss": 0.0025727502070367336, "epoch": 22.02338129496403, "step": 24490 }, { "epoch": 22.02338129496403, "step": 24490, "torque_loss": 0.158044695854187 }, { "epoch": 22.032374100719423, "grad_norm": 0.21069210767745972, "learning_rate": 5.5777516649630756e-05, "loss": 0.0227, "step": 24500 }, { "action_loss": 0.0029548201709985733, "epoch": 22.032374100719423, "step": 24500 }, { "epoch": 22.032374100719423, "step": 24500, "torque_loss": 0.15024498105049133 }, { "epoch": 22.04136690647482, "grad_norm": 0.1503758579492569, "learning_rate": 5.574466750321017e-05, "loss": 0.0245, "step": 24510 }, { "action_loss": 0.005083165597170591, "epoch": 22.04136690647482, "step": 24510 }, { "epoch": 22.04136690647482, "step": 24510, "torque_loss": 0.13949769735336304 }, { "epoch": 22.050359712230215, "grad_norm": 0.13550932705402374, "learning_rate": 5.571181584387683e-05, "loss": 0.0247, "step": 24520 }, { "action_loss": 0.01852327026426792, "epoch": 22.050359712230215, "step": 24520 }, { "epoch": 22.050359712230215, "step": 24520, "torque_loss": 0.22127850353717804 }, { "epoch": 22.059352517985612, "grad_norm": 0.1855095773935318, "learning_rate": 5.5678961686001196e-05, "loss": 0.0254, "step": 24530 }, { "action_loss": 0.008152628317475319, "epoch": 22.059352517985612, "step": 24530 }, { "epoch": 22.059352517985612, "step": 24530, "torque_loss": 0.2266409993171692 }, { "epoch": 22.068345323741006, "grad_norm": 0.15228554606437683, "learning_rate": 5.5646105043954785e-05, "loss": 0.0254, "step": 24540 }, { "action_loss": 0.0020510407630354166, "epoch": 22.068345323741006, "step": 24540 }, { "epoch": 22.068345323741006, "step": 24540, "torque_loss": 0.1309548169374466 }, { "epoch": 22.077338129496404, "grad_norm": 0.13346761465072632, "learning_rate": 5.5613245932110205e-05, "loss": 0.0217, "step": 24550 }, { "action_loss": 0.012908682227134705, "epoch": 22.077338129496404, "step": 24550 }, { "epoch": 22.077338129496404, "step": 24550, "torque_loss": 0.24389608204364777 }, { "epoch": 22.086330935251798, "grad_norm": 0.12421543151140213, "learning_rate": 5.5580384364841157e-05, "loss": 0.0266, "step": 24560 }, { "action_loss": 0.007695457432419062, "epoch": 22.086330935251798, "step": 24560 }, { "epoch": 22.086330935251798, "step": 24560, "torque_loss": 0.16236071288585663 }, { "epoch": 22.095323741007196, "grad_norm": 0.12151192873716354, "learning_rate": 5.554752035652237e-05, "loss": 0.0226, "step": 24570 }, { "action_loss": 0.005820412188768387, "epoch": 22.095323741007196, "step": 24570 }, { "epoch": 22.095323741007196, "step": 24570, "torque_loss": 0.16814327239990234 }, { "epoch": 22.10431654676259, "grad_norm": 0.15803925693035126, "learning_rate": 5.5514653921529736e-05, "loss": 0.0214, "step": 24580 }, { "action_loss": 0.004435693379491568, "epoch": 22.10431654676259, "step": 24580 }, { "epoch": 22.10431654676259, "step": 24580, "torque_loss": 0.16194143891334534 }, { "epoch": 22.113309352517987, "grad_norm": 0.14214497804641724, "learning_rate": 5.5481785074240114e-05, "loss": 0.0212, "step": 24590 }, { "action_loss": 0.006233118008822203, "epoch": 22.113309352517987, "step": 24590 }, { "epoch": 22.113309352517987, "step": 24590, "torque_loss": 0.20257587730884552 }, { "epoch": 22.12230215827338, "grad_norm": 0.12182547152042389, "learning_rate": 5.544891382903146e-05, "loss": 0.0228, "step": 24600 }, { "action_loss": 0.0056479438208043575, "epoch": 22.12230215827338, "step": 24600 }, { "epoch": 22.12230215827338, "step": 24600, "torque_loss": 0.13119888305664062 }, { "epoch": 22.131294964028775, "grad_norm": 0.1332070529460907, "learning_rate": 5.541604020028277e-05, "loss": 0.0215, "step": 24610 }, { "action_loss": 0.00284171593375504, "epoch": 22.131294964028775, "step": 24610 }, { "epoch": 22.131294964028775, "step": 24610, "torque_loss": 0.2059619426727295 }, { "epoch": 22.140287769784173, "grad_norm": 0.26309332251548767, "learning_rate": 5.5383164202374105e-05, "loss": 0.0228, "step": 24620 }, { "action_loss": 0.013848252594470978, "epoch": 22.140287769784173, "step": 24620 }, { "epoch": 22.140287769784173, "step": 24620, "torque_loss": 0.21839487552642822 }, { "epoch": 22.149280575539567, "grad_norm": 0.11114880442619324, "learning_rate": 5.5350285849686515e-05, "loss": 0.0235, "step": 24630 }, { "action_loss": 0.006417728960514069, "epoch": 22.149280575539567, "step": 24630 }, { "epoch": 22.149280575539567, "step": 24630, "torque_loss": 0.16121910512447357 }, { "epoch": 22.158273381294965, "grad_norm": 0.11531244218349457, "learning_rate": 5.531740515660213e-05, "loss": 0.0242, "step": 24640 }, { "action_loss": 0.005977736320346594, "epoch": 22.158273381294965, "step": 24640 }, { "epoch": 22.158273381294965, "step": 24640, "torque_loss": 0.09957949072122574 }, { "epoch": 22.16726618705036, "grad_norm": 0.08154119551181793, "learning_rate": 5.528452213750408e-05, "loss": 0.0196, "step": 24650 }, { "action_loss": 0.0026358989998698235, "epoch": 22.16726618705036, "step": 24650 }, { "epoch": 22.16726618705036, "step": 24650, "torque_loss": 0.11036231368780136 }, { "epoch": 22.176258992805757, "grad_norm": 0.20108018815517426, "learning_rate": 5.52516368067765e-05, "loss": 0.0234, "step": 24660 }, { "action_loss": 0.004828148987144232, "epoch": 22.176258992805757, "step": 24660 }, { "epoch": 22.176258992805757, "step": 24660, "torque_loss": 0.15047167241573334 }, { "epoch": 22.18525179856115, "grad_norm": 0.25339803099632263, "learning_rate": 5.5218749178804563e-05, "loss": 0.0239, "step": 24670 }, { "action_loss": 0.003805997781455517, "epoch": 22.18525179856115, "step": 24670 }, { "epoch": 22.18525179856115, "step": 24670, "torque_loss": 0.1686844825744629 }, { "epoch": 22.194244604316548, "grad_norm": 0.11515042930841446, "learning_rate": 5.518585926797443e-05, "loss": 0.0205, "step": 24680 }, { "action_loss": 0.011961720883846283, "epoch": 22.194244604316548, "step": 24680 }, { "epoch": 22.194244604316548, "step": 24680, "torque_loss": 0.17996914684772491 }, { "epoch": 22.203237410071942, "grad_norm": 0.1885412037372589, "learning_rate": 5.5152967088673265e-05, "loss": 0.0242, "step": 24690 }, { "action_loss": 0.005072618369013071, "epoch": 22.203237410071942, "step": 24690 }, { "epoch": 22.203237410071942, "step": 24690, "torque_loss": 0.14319775998592377 }, { "epoch": 22.21223021582734, "grad_norm": 0.06771261245012283, "learning_rate": 5.512007265528924e-05, "loss": 0.0222, "step": 24700 }, { "action_loss": 0.005441195797175169, "epoch": 22.21223021582734, "step": 24700 }, { "epoch": 22.21223021582734, "step": 24700, "torque_loss": 0.20202100276947021 }, { "epoch": 22.221223021582734, "grad_norm": 0.11035338789224625, "learning_rate": 5.508717598221148e-05, "loss": 0.0203, "step": 24710 }, { "action_loss": 0.0049959709867835045, "epoch": 22.221223021582734, "step": 24710 }, { "epoch": 22.221223021582734, "step": 24710, "torque_loss": 0.15892350673675537 }, { "epoch": 22.230215827338128, "grad_norm": 0.1517138034105301, "learning_rate": 5.505427708383012e-05, "loss": 0.0222, "step": 24720 }, { "action_loss": 0.004690217319875956, "epoch": 22.230215827338128, "step": 24720 }, { "epoch": 22.230215827338128, "step": 24720, "torque_loss": 0.20128262042999268 }, { "epoch": 22.239208633093526, "grad_norm": 0.11171448975801468, "learning_rate": 5.502137597453624e-05, "loss": 0.0196, "step": 24730 }, { "action_loss": 0.0037811249494552612, "epoch": 22.239208633093526, "step": 24730 }, { "epoch": 22.239208633093526, "step": 24730, "torque_loss": 0.08843883872032166 }, { "epoch": 22.24820143884892, "grad_norm": 0.12603534758090973, "learning_rate": 5.4988472668721914e-05, "loss": 0.0192, "step": 24740 }, { "action_loss": 0.011832957156002522, "epoch": 22.24820143884892, "step": 24740 }, { "epoch": 22.24820143884892, "step": 24740, "torque_loss": 0.194475457072258 }, { "epoch": 22.257194244604317, "grad_norm": 0.15904612839221954, "learning_rate": 5.4955567180780164e-05, "loss": 0.0248, "step": 24750 }, { "action_loss": 0.007061919197440147, "epoch": 22.257194244604317, "step": 24750 }, { "epoch": 22.257194244604317, "step": 24750, "torque_loss": 0.21828918159008026 }, { "epoch": 22.26618705035971, "grad_norm": 0.16176308691501617, "learning_rate": 5.492265952510497e-05, "loss": 0.0275, "step": 24760 }, { "action_loss": 0.006581681314855814, "epoch": 22.26618705035971, "step": 24760 }, { "epoch": 22.26618705035971, "step": 24760, "torque_loss": 0.13271714746952057 }, { "epoch": 22.27517985611511, "grad_norm": 0.1438520848751068, "learning_rate": 5.4889749716091254e-05, "loss": 0.0218, "step": 24770 }, { "action_loss": 0.003152300603687763, "epoch": 22.27517985611511, "step": 24770 }, { "epoch": 22.27517985611511, "step": 24770, "torque_loss": 0.12965397536754608 }, { "epoch": 22.284172661870503, "grad_norm": 0.09169828146696091, "learning_rate": 5.4856837768134885e-05, "loss": 0.0245, "step": 24780 }, { "action_loss": 0.0062849316745996475, "epoch": 22.284172661870503, "step": 24780 }, { "epoch": 22.284172661870503, "step": 24780, "torque_loss": 0.19084198772907257 }, { "epoch": 22.2931654676259, "grad_norm": 0.1308412402868271, "learning_rate": 5.482392369563267e-05, "loss": 0.0253, "step": 24790 }, { "action_loss": 0.013857029378414154, "epoch": 22.2931654676259, "step": 24790 }, { "epoch": 22.2931654676259, "step": 24790, "torque_loss": 0.22623217105865479 }, { "epoch": 22.302158273381295, "grad_norm": 0.170404314994812, "learning_rate": 5.4791007512982354e-05, "loss": 0.0286, "step": 24800 }, { "action_loss": 0.007277810946106911, "epoch": 22.302158273381295, "step": 24800 }, { "epoch": 22.302158273381295, "step": 24800, "torque_loss": 0.19964104890823364 }, { "epoch": 22.31115107913669, "grad_norm": 0.12136830389499664, "learning_rate": 5.4758089234582566e-05, "loss": 0.0253, "step": 24810 }, { "action_loss": 0.015154997818171978, "epoch": 22.31115107913669, "step": 24810 }, { "epoch": 22.31115107913669, "step": 24810, "torque_loss": 0.2643536329269409 }, { "epoch": 22.320143884892087, "grad_norm": 0.15998420119285583, "learning_rate": 5.47251688748329e-05, "loss": 0.0262, "step": 24820 }, { "action_loss": 0.009238668717443943, "epoch": 22.320143884892087, "step": 24820 }, { "epoch": 22.320143884892087, "step": 24820, "torque_loss": 0.21799087524414062 }, { "epoch": 22.32913669064748, "grad_norm": 0.16697855293750763, "learning_rate": 5.4692246448133843e-05, "loss": 0.0247, "step": 24830 }, { "action_loss": 0.005967763718217611, "epoch": 22.32913669064748, "step": 24830 }, { "epoch": 22.32913669064748, "step": 24830, "torque_loss": 0.18311245739459991 }, { "epoch": 22.33812949640288, "grad_norm": 0.18535968661308289, "learning_rate": 5.465932196888676e-05, "loss": 0.0265, "step": 24840 }, { "action_loss": 0.0033128985669463873, "epoch": 22.33812949640288, "step": 24840 }, { "epoch": 22.33812949640288, "step": 24840, "torque_loss": 0.09521421790122986 }, { "epoch": 22.347122302158272, "grad_norm": 0.12994052469730377, "learning_rate": 5.462639545149395e-05, "loss": 0.0224, "step": 24850 }, { "action_loss": 0.0020264284685254097, "epoch": 22.347122302158272, "step": 24850 }, { "epoch": 22.347122302158272, "step": 24850, "torque_loss": 0.19146089255809784 }, { "epoch": 22.35611510791367, "grad_norm": 0.17417950928211212, "learning_rate": 5.459346691035861e-05, "loss": 0.0257, "step": 24860 }, { "action_loss": 0.0038116525392979383, "epoch": 22.35611510791367, "step": 24860 }, { "epoch": 22.35611510791367, "step": 24860, "torque_loss": 0.12889593839645386 }, { "epoch": 22.365107913669064, "grad_norm": 0.13815060257911682, "learning_rate": 5.4560536359884776e-05, "loss": 0.0205, "step": 24870 }, { "action_loss": 0.013800610788166523, "epoch": 22.365107913669064, "step": 24870 }, { "epoch": 22.365107913669064, "step": 24870, "torque_loss": 0.22764933109283447 }, { "epoch": 22.37410071942446, "grad_norm": 0.1270178109407425, "learning_rate": 5.45276038144774e-05, "loss": 0.0255, "step": 24880 }, { "action_loss": 0.006845576222985983, "epoch": 22.37410071942446, "step": 24880 }, { "epoch": 22.37410071942446, "step": 24880, "torque_loss": 0.20313559472560883 }, { "epoch": 22.383093525179856, "grad_norm": 0.11127965897321701, "learning_rate": 5.449466928854229e-05, "loss": 0.0207, "step": 24890 }, { "action_loss": 0.006485458463430405, "epoch": 22.383093525179856, "step": 24890 }, { "epoch": 22.383093525179856, "step": 24890, "torque_loss": 0.24268598854541779 }, { "epoch": 22.392086330935253, "grad_norm": 0.16079416871070862, "learning_rate": 5.446173279648613e-05, "loss": 0.0237, "step": 24900 }, { "action_loss": 0.011824525892734528, "epoch": 22.392086330935253, "step": 24900 }, { "epoch": 22.392086330935253, "step": 24900, "torque_loss": 0.19755826890468597 }, { "epoch": 22.401079136690647, "grad_norm": 0.1400780826807022, "learning_rate": 5.442879435271646e-05, "loss": 0.0256, "step": 24910 }, { "action_loss": 0.005275214556604624, "epoch": 22.401079136690647, "step": 24910 }, { "epoch": 22.401079136690647, "step": 24910, "torque_loss": 0.23795248568058014 }, { "epoch": 22.41007194244604, "grad_norm": 0.1991378366947174, "learning_rate": 5.4395853971641675e-05, "loss": 0.0257, "step": 24920 }, { "action_loss": 0.010468658059835434, "epoch": 22.41007194244604, "step": 24920 }, { "epoch": 22.41007194244604, "step": 24920, "torque_loss": 0.15180319547653198 }, { "epoch": 22.41906474820144, "grad_norm": 0.13602708280086517, "learning_rate": 5.4362911667671015e-05, "loss": 0.0241, "step": 24930 }, { "action_loss": 0.0028790393844246864, "epoch": 22.41906474820144, "step": 24930 }, { "epoch": 22.41906474820144, "step": 24930, "torque_loss": 0.14825089275836945 }, { "epoch": 22.428057553956833, "grad_norm": 0.13203607499599457, "learning_rate": 5.432996745521458e-05, "loss": 0.0217, "step": 24940 }, { "action_loss": 0.043567705899477005, "epoch": 22.428057553956833, "step": 24940 }, { "epoch": 22.428057553956833, "step": 24940, "torque_loss": 0.2324451208114624 }, { "epoch": 22.43705035971223, "grad_norm": 0.11169055104255676, "learning_rate": 5.4297021348683264e-05, "loss": 0.0282, "step": 24950 }, { "action_loss": 0.003374951658770442, "epoch": 22.43705035971223, "step": 24950 }, { "epoch": 22.43705035971223, "step": 24950, "torque_loss": 0.12394073605537415 }, { "epoch": 22.446043165467625, "grad_norm": 0.13291658461093903, "learning_rate": 5.426407336248882e-05, "loss": 0.0221, "step": 24960 }, { "action_loss": 0.003150752978399396, "epoch": 22.446043165467625, "step": 24960 }, { "epoch": 22.446043165467625, "step": 24960, "torque_loss": 0.21036970615386963 }, { "epoch": 22.455035971223023, "grad_norm": 0.10400117188692093, "learning_rate": 5.423112351104382e-05, "loss": 0.0212, "step": 24970 }, { "action_loss": 0.011177725158631802, "epoch": 22.455035971223023, "step": 24970 }, { "epoch": 22.455035971223023, "step": 24970, "torque_loss": 0.23840957880020142 }, { "epoch": 22.464028776978417, "grad_norm": 0.1750277876853943, "learning_rate": 5.4198171808761654e-05, "loss": 0.0245, "step": 24980 }, { "action_loss": 0.004575620871037245, "epoch": 22.464028776978417, "step": 24980 }, { "epoch": 22.464028776978417, "step": 24980, "torque_loss": 0.2002529501914978 }, { "epoch": 22.473021582733814, "grad_norm": 0.12737776339054108, "learning_rate": 5.41652182700565e-05, "loss": 0.0224, "step": 24990 }, { "action_loss": 0.006294598802924156, "epoch": 22.473021582733814, "step": 24990 }, { "epoch": 22.473021582733814, "step": 24990, "torque_loss": 0.1081562265753746 }, { "epoch": 22.48201438848921, "grad_norm": 0.19495901465415955, "learning_rate": 5.413226290934338e-05, "loss": 0.0249, "step": 25000 }, { "action_loss": 0.0020835271570831537, "epoch": 22.48201438848921, "step": 25000 }, { "epoch": 22.48201438848921, "step": 25000, "torque_loss": 0.1360519528388977 }, { "epoch": 22.491007194244606, "grad_norm": 0.13298127055168152, "learning_rate": 5.4099305741038056e-05, "loss": 0.0208, "step": 25010 }, { "action_loss": 0.0036703653167933226, "epoch": 22.491007194244606, "step": 25010 }, { "epoch": 22.491007194244606, "step": 25010, "torque_loss": 0.13748522102832794 }, { "epoch": 22.5, "grad_norm": 0.16108660399913788, "learning_rate": 5.406634677955713e-05, "loss": 0.0249, "step": 25020 }, { "action_loss": 0.002186000579968095, "epoch": 22.5, "step": 25020 }, { "epoch": 22.5, "step": 25020, "torque_loss": 0.19376514852046967 }, { "epoch": 22.508992805755394, "grad_norm": 0.13511033356189728, "learning_rate": 5.403338603931798e-05, "loss": 0.0215, "step": 25030 }, { "action_loss": 0.0017886239802464843, "epoch": 22.508992805755394, "step": 25030 }, { "epoch": 22.508992805755394, "step": 25030, "torque_loss": 0.13630972802639008 }, { "epoch": 22.51798561151079, "grad_norm": 0.1358717978000641, "learning_rate": 5.4000423534738745e-05, "loss": 0.0208, "step": 25040 }, { "action_loss": 0.0044100224040448666, "epoch": 22.51798561151079, "step": 25040 }, { "epoch": 22.51798561151079, "step": 25040, "torque_loss": 0.14715953171253204 }, { "epoch": 22.526978417266186, "grad_norm": 0.12301664054393768, "learning_rate": 5.396745928023835e-05, "loss": 0.0225, "step": 25050 }, { "action_loss": 0.007063267286866903, "epoch": 22.526978417266186, "step": 25050 }, { "epoch": 22.526978417266186, "step": 25050, "torque_loss": 0.19640354812145233 }, { "epoch": 22.535971223021583, "grad_norm": 0.18340817093849182, "learning_rate": 5.3934493290236474e-05, "loss": 0.0209, "step": 25060 }, { "action_loss": 0.0063055469654500484, "epoch": 22.535971223021583, "step": 25060 }, { "epoch": 22.535971223021583, "step": 25060, "torque_loss": 0.24158990383148193 }, { "epoch": 22.544964028776977, "grad_norm": 0.11677511781454086, "learning_rate": 5.390152557915357e-05, "loss": 0.023, "step": 25070 }, { "action_loss": 0.0025417471770197153, "epoch": 22.544964028776977, "step": 25070 }, { "epoch": 22.544964028776977, "step": 25070, "torque_loss": 0.127644881606102 }, { "epoch": 22.553956834532375, "grad_norm": 0.1633521467447281, "learning_rate": 5.3868556161410796e-05, "loss": 0.0203, "step": 25080 }, { "action_loss": 0.005591684486716986, "epoch": 22.553956834532375, "step": 25080 }, { "epoch": 22.553956834532375, "step": 25080, "torque_loss": 0.14006567001342773 }, { "epoch": 22.56294964028777, "grad_norm": 0.1385440230369568, "learning_rate": 5.383558505143015e-05, "loss": 0.0209, "step": 25090 }, { "action_loss": 0.004367899149656296, "epoch": 22.56294964028777, "step": 25090 }, { "epoch": 22.56294964028777, "step": 25090, "torque_loss": 0.20020723342895508 }, { "epoch": 22.571942446043167, "grad_norm": 0.1359289586544037, "learning_rate": 5.38026122636343e-05, "loss": 0.0268, "step": 25100 }, { "action_loss": 0.009880579076707363, "epoch": 22.571942446043167, "step": 25100 }, { "epoch": 22.571942446043167, "step": 25100, "torque_loss": 0.2626066505908966 }, { "epoch": 22.58093525179856, "grad_norm": 0.18494759500026703, "learning_rate": 5.3769637812446636e-05, "loss": 0.0225, "step": 25110 }, { "action_loss": 0.0064749098382890224, "epoch": 22.58093525179856, "step": 25110 }, { "epoch": 22.58093525179856, "step": 25110, "torque_loss": 0.1782163381576538 }, { "epoch": 22.58992805755396, "grad_norm": 0.20037710666656494, "learning_rate": 5.373666171229133e-05, "loss": 0.0208, "step": 25120 }, { "action_loss": 0.005767613183706999, "epoch": 22.58992805755396, "step": 25120 }, { "epoch": 22.58992805755396, "step": 25120, "torque_loss": 0.13761569559574127 }, { "epoch": 22.598920863309353, "grad_norm": 0.09623949229717255, "learning_rate": 5.370368397759324e-05, "loss": 0.0202, "step": 25130 }, { "action_loss": 0.0036416279617697, "epoch": 22.598920863309353, "step": 25130 }, { "epoch": 22.598920863309353, "step": 25130, "torque_loss": 0.16640301048755646 }, { "epoch": 22.607913669064747, "grad_norm": 0.18285290896892548, "learning_rate": 5.3670704622777944e-05, "loss": 0.0299, "step": 25140 }, { "action_loss": 0.0037279303651303053, "epoch": 22.607913669064747, "step": 25140 }, { "epoch": 22.607913669064747, "step": 25140, "torque_loss": 0.18048661947250366 }, { "epoch": 22.616906474820144, "grad_norm": 0.13420526683330536, "learning_rate": 5.3637723662271744e-05, "loss": 0.0219, "step": 25150 }, { "action_loss": 0.00828558485955, "epoch": 22.616906474820144, "step": 25150 }, { "epoch": 22.616906474820144, "step": 25150, "torque_loss": 0.23222021758556366 }, { "epoch": 22.62589928057554, "grad_norm": 0.12542574107646942, "learning_rate": 5.360474111050162e-05, "loss": 0.0258, "step": 25160 }, { "action_loss": 0.0034175908658653498, "epoch": 22.62589928057554, "step": 25160 }, { "epoch": 22.62589928057554, "step": 25160, "torque_loss": 0.12904882431030273 }, { "epoch": 22.634892086330936, "grad_norm": 0.14534123241901398, "learning_rate": 5.357175698189527e-05, "loss": 0.0226, "step": 25170 }, { "action_loss": 0.003703384893015027, "epoch": 22.634892086330936, "step": 25170 }, { "epoch": 22.634892086330936, "step": 25170, "torque_loss": 0.16835761070251465 }, { "epoch": 22.64388489208633, "grad_norm": 0.12861134111881256, "learning_rate": 5.353877129088107e-05, "loss": 0.0215, "step": 25180 }, { "action_loss": 0.01007279846817255, "epoch": 22.64388489208633, "step": 25180 }, { "epoch": 22.64388489208633, "step": 25180, "torque_loss": 0.16919822990894318 }, { "epoch": 22.652877697841728, "grad_norm": 0.10902013629674911, "learning_rate": 5.350578405188809e-05, "loss": 0.0219, "step": 25190 }, { "action_loss": 0.00475480780005455, "epoch": 22.652877697841728, "step": 25190 }, { "epoch": 22.652877697841728, "step": 25190, "torque_loss": 0.19542278349399567 }, { "epoch": 22.66187050359712, "grad_norm": 0.1971260905265808, "learning_rate": 5.347279527934603e-05, "loss": 0.0268, "step": 25200 }, { "action_loss": 0.005096907261759043, "epoch": 22.66187050359712, "step": 25200 }, { "epoch": 22.66187050359712, "step": 25200, "torque_loss": 0.17275528609752655 }, { "epoch": 22.67086330935252, "grad_norm": 0.19193746149539948, "learning_rate": 5.343980498768536e-05, "loss": 0.0196, "step": 25210 }, { "action_loss": 0.002406742423772812, "epoch": 22.67086330935252, "step": 25210 }, { "epoch": 22.67086330935252, "step": 25210, "torque_loss": 0.14266318082809448 }, { "epoch": 22.679856115107913, "grad_norm": 0.10882573574781418, "learning_rate": 5.340681319133711e-05, "loss": 0.0211, "step": 25220 }, { "action_loss": 0.0033563680481165648, "epoch": 22.679856115107913, "step": 25220 }, { "epoch": 22.679856115107913, "step": 25220, "torque_loss": 0.139910027384758 }, { "epoch": 22.68884892086331, "grad_norm": 0.14031830430030823, "learning_rate": 5.337381990473305e-05, "loss": 0.0255, "step": 25230 }, { "action_loss": 0.0020542130805552006, "epoch": 22.68884892086331, "step": 25230 }, { "epoch": 22.68884892086331, "step": 25230, "torque_loss": 0.22107143700122833 }, { "epoch": 22.697841726618705, "grad_norm": 0.15872125327587128, "learning_rate": 5.3340825142305504e-05, "loss": 0.0212, "step": 25240 }, { "action_loss": 0.005888367537409067, "epoch": 22.697841726618705, "step": 25240 }, { "epoch": 22.697841726618705, "step": 25240, "torque_loss": 0.13613416254520416 }, { "epoch": 22.7068345323741, "grad_norm": 0.13879311084747314, "learning_rate": 5.330782891848756e-05, "loss": 0.0216, "step": 25250 }, { "action_loss": 0.004587147384881973, "epoch": 22.7068345323741, "step": 25250 }, { "epoch": 22.7068345323741, "step": 25250, "torque_loss": 0.19867944717407227 }, { "epoch": 22.715827338129497, "grad_norm": 0.09635479003190994, "learning_rate": 5.327483124771285e-05, "loss": 0.0247, "step": 25260 }, { "action_loss": 0.003984099719673395, "epoch": 22.715827338129497, "step": 25260 }, { "epoch": 22.715827338129497, "step": 25260, "torque_loss": 0.1691085696220398 }, { "epoch": 22.72482014388489, "grad_norm": 0.09793604910373688, "learning_rate": 5.3241832144415706e-05, "loss": 0.0236, "step": 25270 }, { "action_loss": 0.002333814511075616, "epoch": 22.72482014388489, "step": 25270 }, { "epoch": 22.72482014388489, "step": 25270, "torque_loss": 0.12268368154764175 }, { "epoch": 22.73381294964029, "grad_norm": 0.0995931327342987, "learning_rate": 5.320883162303104e-05, "loss": 0.0224, "step": 25280 }, { "action_loss": 0.0023210274521261454, "epoch": 22.73381294964029, "step": 25280 }, { "epoch": 22.73381294964029, "step": 25280, "torque_loss": 0.1571170687675476 }, { "epoch": 22.742805755395683, "grad_norm": 0.10586139559745789, "learning_rate": 5.3175829697994416e-05, "loss": 0.0237, "step": 25290 }, { "action_loss": 0.005923504009842873, "epoch": 22.742805755395683, "step": 25290 }, { "epoch": 22.742805755395683, "step": 25290, "torque_loss": 0.17514441907405853 }, { "epoch": 22.75179856115108, "grad_norm": 0.08688942342996597, "learning_rate": 5.3142826383741975e-05, "loss": 0.0203, "step": 25300 }, { "action_loss": 0.002308244351297617, "epoch": 22.75179856115108, "step": 25300 }, { "epoch": 22.75179856115108, "step": 25300, "torque_loss": 0.12322612851858139 }, { "epoch": 22.760791366906474, "grad_norm": 0.13557827472686768, "learning_rate": 5.3109821694710506e-05, "loss": 0.0221, "step": 25310 }, { "action_loss": 0.003773240139707923, "epoch": 22.760791366906474, "step": 25310 }, { "epoch": 22.760791366906474, "step": 25310, "torque_loss": 0.15425537526607513 }, { "epoch": 22.769784172661872, "grad_norm": 0.09995083510875702, "learning_rate": 5.307681564533736e-05, "loss": 0.0248, "step": 25320 }, { "action_loss": 0.009609272703528404, "epoch": 22.769784172661872, "step": 25320 }, { "epoch": 22.769784172661872, "step": 25320, "torque_loss": 0.198014497756958 }, { "epoch": 22.778776978417266, "grad_norm": 0.21460092067718506, "learning_rate": 5.3043808250060535e-05, "loss": 0.0231, "step": 25330 }, { "action_loss": 0.002328372560441494, "epoch": 22.778776978417266, "step": 25330 }, { "epoch": 22.778776978417266, "step": 25330, "torque_loss": 0.13984860479831696 }, { "epoch": 22.78776978417266, "grad_norm": 0.13254739344120026, "learning_rate": 5.3010799523318566e-05, "loss": 0.026, "step": 25340 }, { "action_loss": 0.0070233759470283985, "epoch": 22.78776978417266, "step": 25340 }, { "epoch": 22.78776978417266, "step": 25340, "torque_loss": 0.1586199551820755 }, { "epoch": 22.796762589928058, "grad_norm": 0.15601493418216705, "learning_rate": 5.29777894795506e-05, "loss": 0.0213, "step": 25350 }, { "action_loss": 0.003634588560089469, "epoch": 22.796762589928058, "step": 25350 }, { "epoch": 22.796762589928058, "step": 25350, "torque_loss": 0.09343317896127701 }, { "epoch": 22.805755395683452, "grad_norm": 0.12755295634269714, "learning_rate": 5.294477813319634e-05, "loss": 0.0204, "step": 25360 }, { "action_loss": 0.007208446506410837, "epoch": 22.805755395683452, "step": 25360 }, { "epoch": 22.805755395683452, "step": 25360, "torque_loss": 0.19211263954639435 }, { "epoch": 22.81474820143885, "grad_norm": 0.12367548048496246, "learning_rate": 5.291176549869608e-05, "loss": 0.0255, "step": 25370 }, { "action_loss": 0.0047527155838906765, "epoch": 22.81474820143885, "step": 25370 }, { "epoch": 22.81474820143885, "step": 25370, "torque_loss": 0.20653212070465088 }, { "epoch": 22.823741007194243, "grad_norm": 0.19360660016536713, "learning_rate": 5.287875159049065e-05, "loss": 0.0314, "step": 25380 }, { "action_loss": 0.014799867756664753, "epoch": 22.823741007194243, "step": 25380 }, { "epoch": 22.823741007194243, "step": 25380, "torque_loss": 0.23008005321025848 }, { "epoch": 22.83273381294964, "grad_norm": 0.20703405141830444, "learning_rate": 5.284573642302149e-05, "loss": 0.0274, "step": 25390 }, { "action_loss": 0.006117114331573248, "epoch": 22.83273381294964, "step": 25390 }, { "epoch": 22.83273381294964, "step": 25390, "torque_loss": 0.13083697855472565 }, { "epoch": 22.841726618705035, "grad_norm": 0.16118504106998444, "learning_rate": 5.2812720010730543e-05, "loss": 0.0231, "step": 25400 }, { "action_loss": 0.007739573251456022, "epoch": 22.841726618705035, "step": 25400 }, { "epoch": 22.841726618705035, "step": 25400, "torque_loss": 0.1770210862159729 }, { "epoch": 22.850719424460433, "grad_norm": 0.12802095711231232, "learning_rate": 5.277970236806029e-05, "loss": 0.0203, "step": 25410 }, { "action_loss": 0.010395831428468227, "epoch": 22.850719424460433, "step": 25410 }, { "epoch": 22.850719424460433, "step": 25410, "torque_loss": 0.19535644352436066 }, { "epoch": 22.859712230215827, "grad_norm": 0.11885704845190048, "learning_rate": 5.2746683509453795e-05, "loss": 0.0207, "step": 25420 }, { "action_loss": 0.00461698928847909, "epoch": 22.859712230215827, "step": 25420 }, { "epoch": 22.859712230215827, "step": 25420, "torque_loss": 0.1878272145986557 }, { "epoch": 22.868705035971225, "grad_norm": 0.09257720410823822, "learning_rate": 5.271366344935461e-05, "loss": 0.0232, "step": 25430 }, { "action_loss": 0.002797633409500122, "epoch": 22.868705035971225, "step": 25430 }, { "epoch": 22.868705035971225, "step": 25430, "torque_loss": 0.14320485293865204 }, { "epoch": 22.87769784172662, "grad_norm": 0.11995459347963333, "learning_rate": 5.268064220220683e-05, "loss": 0.0245, "step": 25440 }, { "action_loss": 0.002526690950617194, "epoch": 22.87769784172662, "step": 25440 }, { "epoch": 22.87769784172662, "step": 25440, "torque_loss": 0.08949947357177734 }, { "epoch": 22.886690647482013, "grad_norm": 0.15516656637191772, "learning_rate": 5.2647619782455094e-05, "loss": 0.0232, "step": 25450 }, { "action_loss": 0.0017795321764424443, "epoch": 22.886690647482013, "step": 25450 }, { "epoch": 22.886690647482013, "step": 25450, "torque_loss": 0.1123981848359108 }, { "epoch": 22.89568345323741, "grad_norm": 0.10655581206083298, "learning_rate": 5.261459620454451e-05, "loss": 0.0226, "step": 25460 }, { "action_loss": 0.0035901369992643595, "epoch": 22.89568345323741, "step": 25460 }, { "epoch": 22.89568345323741, "step": 25460, "torque_loss": 0.17003358900547028 }, { "epoch": 22.904676258992804, "grad_norm": 0.10739203542470932, "learning_rate": 5.258157148292071e-05, "loss": 0.0206, "step": 25470 }, { "action_loss": 0.02539128065109253, "epoch": 22.904676258992804, "step": 25470 }, { "epoch": 22.904676258992804, "step": 25470, "torque_loss": 0.187272310256958 }, { "epoch": 22.913669064748202, "grad_norm": 0.1082937940955162, "learning_rate": 5.2548545632029846e-05, "loss": 0.0255, "step": 25480 }, { "action_loss": 0.0052339802496135235, "epoch": 22.913669064748202, "step": 25480 }, { "epoch": 22.913669064748202, "step": 25480, "torque_loss": 0.169399693608284 }, { "epoch": 22.922661870503596, "grad_norm": 0.1108526736497879, "learning_rate": 5.2515518666318534e-05, "loss": 0.0216, "step": 25490 }, { "action_loss": 0.008445404469966888, "epoch": 22.922661870503596, "step": 25490 }, { "epoch": 22.922661870503596, "step": 25490, "torque_loss": 0.1798778921365738 }, { "epoch": 22.931654676258994, "grad_norm": 0.174293652176857, "learning_rate": 5.248249060023389e-05, "loss": 0.0224, "step": 25500 }, { "action_loss": 0.005164081696420908, "epoch": 22.931654676258994, "step": 25500 }, { "epoch": 22.931654676258994, "step": 25500, "torque_loss": 0.19280190765857697 }, { "epoch": 22.940647482014388, "grad_norm": 0.11022061854600906, "learning_rate": 5.2449461448223517e-05, "loss": 0.023, "step": 25510 }, { "action_loss": 0.015785910189151764, "epoch": 22.940647482014388, "step": 25510 }, { "epoch": 22.940647482014388, "step": 25510, "torque_loss": 0.18951328098773956 }, { "epoch": 22.949640287769785, "grad_norm": 0.08959149569272995, "learning_rate": 5.2416431224735486e-05, "loss": 0.0219, "step": 25520 }, { "action_loss": 0.002587687922641635, "epoch": 22.949640287769785, "step": 25520 }, { "epoch": 22.949640287769785, "step": 25520, "torque_loss": 0.12629075348377228 }, { "epoch": 22.95863309352518, "grad_norm": 0.16631384193897247, "learning_rate": 5.238339994421836e-05, "loss": 0.0231, "step": 25530 }, { "action_loss": 0.002288093790411949, "epoch": 22.95863309352518, "step": 25530 }, { "epoch": 22.95863309352518, "step": 25530, "torque_loss": 0.13596951961517334 }, { "epoch": 22.967625899280577, "grad_norm": 0.19683244824409485, "learning_rate": 5.235036762112111e-05, "loss": 0.0192, "step": 25540 }, { "action_loss": 0.006514893379062414, "epoch": 22.967625899280577, "step": 25540 }, { "epoch": 22.967625899280577, "step": 25540, "torque_loss": 0.19561167061328888 }, { "epoch": 22.97661870503597, "grad_norm": 0.22327181696891785, "learning_rate": 5.2317334269893206e-05, "loss": 0.0325, "step": 25550 }, { "action_loss": 0.0020510684698820114, "epoch": 22.97661870503597, "step": 25550 }, { "epoch": 22.97661870503597, "step": 25550, "torque_loss": 0.14429402351379395 }, { "epoch": 22.985611510791365, "grad_norm": 0.17885106801986694, "learning_rate": 5.228429990498458e-05, "loss": 0.022, "step": 25560 }, { "action_loss": 0.0021295102778822184, "epoch": 22.985611510791365, "step": 25560 }, { "epoch": 22.985611510791365, "step": 25560, "torque_loss": 0.11854513734579086 }, { "epoch": 22.994604316546763, "grad_norm": 0.16415326297283173, "learning_rate": 5.225126454084556e-05, "loss": 0.0248, "step": 25570 }, { "action_loss": 0.003812268376350403, "epoch": 22.994604316546763, "step": 25570 }, { "epoch": 22.994604316546763, "step": 25570, "torque_loss": 0.13177931308746338 }, { "epoch": 23.003597122302157, "grad_norm": 0.14216938614845276, "learning_rate": 5.221822819192694e-05, "loss": 0.0267, "step": 25580 }, { "action_loss": 0.0075640142895281315, "epoch": 23.003597122302157, "step": 25580 }, { "epoch": 23.003597122302157, "step": 25580, "torque_loss": 0.15461139380931854 }, { "epoch": 23.012589928057555, "grad_norm": 0.2581377923488617, "learning_rate": 5.218519087267995e-05, "loss": 0.0237, "step": 25590 }, { "action_loss": 0.0025711441412568092, "epoch": 23.012589928057555, "step": 25590 }, { "epoch": 23.012589928057555, "step": 25590, "torque_loss": 0.20277635753154755 }, { "epoch": 23.02158273381295, "grad_norm": 0.15646064281463623, "learning_rate": 5.2152152597556225e-05, "loss": 0.0209, "step": 25600 }, { "action_loss": 0.009281493723392487, "epoch": 23.02158273381295, "step": 25600 }, { "epoch": 23.02158273381295, "step": 25600, "torque_loss": 0.16589273512363434 }, { "epoch": 23.030575539568346, "grad_norm": 0.13767769932746887, "learning_rate": 5.211911338100784e-05, "loss": 0.0227, "step": 25610 }, { "action_loss": 0.007379112299531698, "epoch": 23.030575539568346, "step": 25610 }, { "epoch": 23.030575539568346, "step": 25610, "torque_loss": 0.20587550103664398 }, { "epoch": 23.03956834532374, "grad_norm": 0.1943831443786621, "learning_rate": 5.2086073237487264e-05, "loss": 0.026, "step": 25620 }, { "action_loss": 0.002615562407299876, "epoch": 23.03956834532374, "step": 25620 }, { "epoch": 23.03956834532374, "step": 25620, "torque_loss": 0.1494436115026474 }, { "epoch": 23.048561151079138, "grad_norm": 0.12590284645557404, "learning_rate": 5.205303218144739e-05, "loss": 0.0205, "step": 25630 }, { "action_loss": 0.005026188213378191, "epoch": 23.048561151079138, "step": 25630 }, { "epoch": 23.048561151079138, "step": 25630, "torque_loss": 0.17518067359924316 }, { "epoch": 23.057553956834532, "grad_norm": 0.19361512362957, "learning_rate": 5.2019990227341494e-05, "loss": 0.0264, "step": 25640 }, { "action_loss": 0.0025798806454986334, "epoch": 23.057553956834532, "step": 25640 }, { "epoch": 23.057553956834532, "step": 25640, "torque_loss": 0.11975566297769547 }, { "epoch": 23.06654676258993, "grad_norm": 0.1538919061422348, "learning_rate": 5.198694738962324e-05, "loss": 0.0224, "step": 25650 }, { "action_loss": 0.005295734852552414, "epoch": 23.06654676258993, "step": 25650 }, { "epoch": 23.06654676258993, "step": 25650, "torque_loss": 0.11691313982009888 }, { "epoch": 23.075539568345324, "grad_norm": 0.07791341841220856, "learning_rate": 5.19539036827467e-05, "loss": 0.0238, "step": 25660 }, { "action_loss": 0.0033219456672668457, "epoch": 23.075539568345324, "step": 25660 }, { "epoch": 23.075539568345324, "step": 25660, "torque_loss": 0.1826685667037964 }, { "epoch": 23.084532374100718, "grad_norm": 0.15807032585144043, "learning_rate": 5.192085912116629e-05, "loss": 0.0248, "step": 25670 }, { "action_loss": 0.011696930974721909, "epoch": 23.084532374100718, "step": 25670 }, { "epoch": 23.084532374100718, "step": 25670, "torque_loss": 0.1938273310661316 }, { "epoch": 23.093525179856115, "grad_norm": 0.14208592474460602, "learning_rate": 5.188781371933685e-05, "loss": 0.0233, "step": 25680 }, { "action_loss": 0.0030097949784249067, "epoch": 23.093525179856115, "step": 25680 }, { "epoch": 23.093525179856115, "step": 25680, "torque_loss": 0.12753210961818695 }, { "epoch": 23.10251798561151, "grad_norm": 0.11003544926643372, "learning_rate": 5.185476749171356e-05, "loss": 0.0196, "step": 25690 }, { "action_loss": 0.008272401057183743, "epoch": 23.10251798561151, "step": 25690 }, { "epoch": 23.10251798561151, "step": 25690, "torque_loss": 0.2177422046661377 }, { "epoch": 23.111510791366907, "grad_norm": 0.08421832323074341, "learning_rate": 5.1821720452751945e-05, "loss": 0.0223, "step": 25700 }, { "action_loss": 0.0019896519370377064, "epoch": 23.111510791366907, "step": 25700 }, { "epoch": 23.111510791366907, "step": 25700, "torque_loss": 0.12420288473367691 }, { "epoch": 23.1205035971223, "grad_norm": 0.12091498076915741, "learning_rate": 5.178867261690791e-05, "loss": 0.0215, "step": 25710 }, { "action_loss": 0.012821841053664684, "epoch": 23.1205035971223, "step": 25710 }, { "epoch": 23.1205035971223, "step": 25710, "torque_loss": 0.2092607468366623 }, { "epoch": 23.1294964028777, "grad_norm": 0.1625688672065735, "learning_rate": 5.1755623998637715e-05, "loss": 0.0228, "step": 25720 }, { "action_loss": 0.0033441931009292603, "epoch": 23.1294964028777, "step": 25720 }, { "epoch": 23.1294964028777, "step": 25720, "torque_loss": 0.12249796837568283 }, { "epoch": 23.138489208633093, "grad_norm": 0.15344645082950592, "learning_rate": 5.172257461239794e-05, "loss": 0.0221, "step": 25730 }, { "action_loss": 0.0043710279278457165, "epoch": 23.138489208633093, "step": 25730 }, { "epoch": 23.138489208633093, "step": 25730, "torque_loss": 0.17560331523418427 }, { "epoch": 23.14748201438849, "grad_norm": 0.19342714548110962, "learning_rate": 5.168952447264548e-05, "loss": 0.0262, "step": 25740 }, { "action_loss": 0.009180142544209957, "epoch": 23.14748201438849, "step": 25740 }, { "epoch": 23.14748201438849, "step": 25740, "torque_loss": 0.16096144914627075 }, { "epoch": 23.156474820143885, "grad_norm": 0.1814686506986618, "learning_rate": 5.165647359383763e-05, "loss": 0.0224, "step": 25750 }, { "action_loss": 0.002860763343051076, "epoch": 23.156474820143885, "step": 25750 }, { "epoch": 23.156474820143885, "step": 25750, "torque_loss": 0.1601506620645523 }, { "epoch": 23.165467625899282, "grad_norm": 0.12067956477403641, "learning_rate": 5.1623421990431954e-05, "loss": 0.0231, "step": 25760 }, { "action_loss": 0.004386862274259329, "epoch": 23.165467625899282, "step": 25760 }, { "epoch": 23.165467625899282, "step": 25760, "torque_loss": 0.13497000932693481 }, { "epoch": 23.174460431654676, "grad_norm": 0.13159358501434326, "learning_rate": 5.159036967688635e-05, "loss": 0.0201, "step": 25770 }, { "action_loss": 0.008638809435069561, "epoch": 23.174460431654676, "step": 25770 }, { "epoch": 23.174460431654676, "step": 25770, "torque_loss": 0.1810109168291092 }, { "epoch": 23.18345323741007, "grad_norm": 0.12117709219455719, "learning_rate": 5.1557316667659015e-05, "loss": 0.0256, "step": 25780 }, { "action_loss": 0.015291646122932434, "epoch": 23.18345323741007, "step": 25780 }, { "epoch": 23.18345323741007, "step": 25780, "torque_loss": 0.18754178285598755 }, { "epoch": 23.192446043165468, "grad_norm": 0.1425536870956421, "learning_rate": 5.152426297720846e-05, "loss": 0.0247, "step": 25790 }, { "action_loss": 0.005521934013813734, "epoch": 23.192446043165468, "step": 25790 }, { "epoch": 23.192446043165468, "step": 25790, "torque_loss": 0.20278821885585785 }, { "epoch": 23.201438848920862, "grad_norm": 0.11354301124811172, "learning_rate": 5.1491208619993505e-05, "loss": 0.0191, "step": 25800 }, { "action_loss": 0.015547972172498703, "epoch": 23.201438848920862, "step": 25800 }, { "epoch": 23.201438848920862, "step": 25800, "torque_loss": 0.19481830298900604 }, { "epoch": 23.21043165467626, "grad_norm": 0.14759418368339539, "learning_rate": 5.145815361047321e-05, "loss": 0.022, "step": 25810 }, { "action_loss": 0.004502683412283659, "epoch": 23.21043165467626, "step": 25810 }, { "epoch": 23.21043165467626, "step": 25810, "torque_loss": 0.14928124845027924 }, { "epoch": 23.219424460431654, "grad_norm": 0.10549262166023254, "learning_rate": 5.1425097963107003e-05, "loss": 0.0202, "step": 25820 }, { "action_loss": 0.0026811379939317703, "epoch": 23.219424460431654, "step": 25820 }, { "epoch": 23.219424460431654, "step": 25820, "torque_loss": 0.17278771102428436 }, { "epoch": 23.22841726618705, "grad_norm": 0.1654650866985321, "learning_rate": 5.1392041692354544e-05, "loss": 0.0254, "step": 25830 }, { "action_loss": 0.0023883385583758354, "epoch": 23.22841726618705, "step": 25830 }, { "epoch": 23.22841726618705, "step": 25830, "torque_loss": 0.16833500564098358 }, { "epoch": 23.237410071942445, "grad_norm": 0.21816183626651764, "learning_rate": 5.1358984812675736e-05, "loss": 0.0248, "step": 25840 }, { "action_loss": 0.013809273950755596, "epoch": 23.237410071942445, "step": 25840 }, { "epoch": 23.237410071942445, "step": 25840, "torque_loss": 0.26302388310432434 }, { "epoch": 23.246402877697843, "grad_norm": 0.16212952136993408, "learning_rate": 5.13259273385308e-05, "loss": 0.0224, "step": 25850 }, { "action_loss": 0.022540554404258728, "epoch": 23.246402877697843, "step": 25850 }, { "epoch": 23.246402877697843, "step": 25850, "torque_loss": 0.15127579867839813 }, { "epoch": 23.255395683453237, "grad_norm": 0.2006453275680542, "learning_rate": 5.12928692843802e-05, "loss": 0.0227, "step": 25860 }, { "action_loss": 0.0026791831478476524, "epoch": 23.255395683453237, "step": 25860 }, { "epoch": 23.255395683453237, "step": 25860, "torque_loss": 0.1192520335316658 }, { "epoch": 23.264388489208635, "grad_norm": 0.09900075942277908, "learning_rate": 5.125981066468465e-05, "loss": 0.0188, "step": 25870 }, { "action_loss": 0.0055341534316539764, "epoch": 23.264388489208635, "step": 25870 }, { "epoch": 23.264388489208635, "step": 25870, "torque_loss": 0.2099742442369461 }, { "epoch": 23.27338129496403, "grad_norm": 0.12532402575016022, "learning_rate": 5.122675149390514e-05, "loss": 0.0215, "step": 25880 }, { "action_loss": 0.00721973180770874, "epoch": 23.27338129496403, "step": 25880 }, { "epoch": 23.27338129496403, "step": 25880, "torque_loss": 0.20489752292633057 }, { "epoch": 23.282374100719423, "grad_norm": 0.11951443552970886, "learning_rate": 5.119369178650282e-05, "loss": 0.0241, "step": 25890 }, { "action_loss": 0.004246692638844252, "epoch": 23.282374100719423, "step": 25890 }, { "epoch": 23.282374100719423, "step": 25890, "torque_loss": 0.16140736639499664 }, { "epoch": 23.29136690647482, "grad_norm": 0.11923055350780487, "learning_rate": 5.1160631556939156e-05, "loss": 0.0216, "step": 25900 }, { "action_loss": 0.008979380130767822, "epoch": 23.29136690647482, "step": 25900 }, { "epoch": 23.29136690647482, "step": 25900, "torque_loss": 0.16876240074634552 }, { "epoch": 23.300359712230215, "grad_norm": 0.18254195153713226, "learning_rate": 5.112757081967584e-05, "loss": 0.0226, "step": 25910 }, { "action_loss": 0.0038351838011294603, "epoch": 23.300359712230215, "step": 25910 }, { "epoch": 23.300359712230215, "step": 25910, "torque_loss": 0.1879989355802536 }, { "epoch": 23.309352517985612, "grad_norm": 0.08216993510723114, "learning_rate": 5.109450958917472e-05, "loss": 0.0213, "step": 25920 }, { "action_loss": 0.0051038130186498165, "epoch": 23.309352517985612, "step": 25920 }, { "epoch": 23.309352517985612, "step": 25920, "torque_loss": 0.2021799087524414 }, { "epoch": 23.318345323741006, "grad_norm": 0.12007123231887817, "learning_rate": 5.106144787989794e-05, "loss": 0.0227, "step": 25930 }, { "action_loss": 0.009072327055037022, "epoch": 23.318345323741006, "step": 25930 }, { "epoch": 23.318345323741006, "step": 25930, "torque_loss": 0.14080488681793213 }, { "epoch": 23.327338129496404, "grad_norm": 0.10927730053663254, "learning_rate": 5.102838570630778e-05, "loss": 0.0259, "step": 25940 }, { "action_loss": 0.0016529159620404243, "epoch": 23.327338129496404, "step": 25940 }, { "epoch": 23.327338129496404, "step": 25940, "torque_loss": 0.10848996788263321 }, { "epoch": 23.336330935251798, "grad_norm": 0.1866844892501831, "learning_rate": 5.099532308286678e-05, "loss": 0.0243, "step": 25950 }, { "action_loss": 0.0022612044122070074, "epoch": 23.336330935251798, "step": 25950 }, { "epoch": 23.336330935251798, "step": 25950, "torque_loss": 0.12649333477020264 }, { "epoch": 23.345323741007196, "grad_norm": 0.12826865911483765, "learning_rate": 5.0962260024037645e-05, "loss": 0.0222, "step": 25960 }, { "action_loss": 0.0029491104651242495, "epoch": 23.345323741007196, "step": 25960 }, { "epoch": 23.345323741007196, "step": 25960, "torque_loss": 0.14969240128993988 }, { "epoch": 23.35431654676259, "grad_norm": 0.13946382701396942, "learning_rate": 5.0929196544283276e-05, "loss": 0.0235, "step": 25970 }, { "action_loss": 0.008710924535989761, "epoch": 23.35431654676259, "step": 25970 }, { "epoch": 23.35431654676259, "step": 25970, "torque_loss": 0.23404519259929657 }, { "epoch": 23.363309352517987, "grad_norm": 0.19065837562084198, "learning_rate": 5.089613265806678e-05, "loss": 0.0237, "step": 25980 }, { "action_loss": 0.0035727175418287516, "epoch": 23.363309352517987, "step": 25980 }, { "epoch": 23.363309352517987, "step": 25980, "torque_loss": 0.1379915326833725 }, { "epoch": 23.37230215827338, "grad_norm": 0.11607183516025543, "learning_rate": 5.086306837985141e-05, "loss": 0.0214, "step": 25990 }, { "action_loss": 0.010335169732570648, "epoch": 23.37230215827338, "step": 25990 }, { "epoch": 23.37230215827338, "step": 25990, "torque_loss": 0.16358986496925354 }, { "epoch": 23.381294964028775, "grad_norm": 0.18087837100028992, "learning_rate": 5.0830003724100604e-05, "loss": 0.026, "step": 26000 }, { "action_loss": 0.0028483327478170395, "epoch": 23.381294964028775, "step": 26000 }, { "epoch": 23.381294964028775, "step": 26000, "torque_loss": 0.15708626806735992 }, { "epoch": 23.390287769784173, "grad_norm": 0.15104742348194122, "learning_rate": 5.0796938705277984e-05, "loss": 0.0252, "step": 26010 }, { "action_loss": 0.0018918405985459685, "epoch": 23.390287769784173, "step": 26010 }, { "epoch": 23.390287769784173, "step": 26010, "torque_loss": 0.13728433847427368 }, { "epoch": 23.399280575539567, "grad_norm": 0.11089194566011429, "learning_rate": 5.076387333784729e-05, "loss": 0.0237, "step": 26020 }, { "action_loss": 0.006259243935346603, "epoch": 23.399280575539567, "step": 26020 }, { "epoch": 23.399280575539567, "step": 26020, "torque_loss": 0.139131560921669 }, { "epoch": 23.408273381294965, "grad_norm": 0.08042974770069122, "learning_rate": 5.0730807636272445e-05, "loss": 0.0225, "step": 26030 }, { "action_loss": 0.003405290422961116, "epoch": 23.408273381294965, "step": 26030 }, { "epoch": 23.408273381294965, "step": 26030, "torque_loss": 0.19358880817890167 }, { "epoch": 23.41726618705036, "grad_norm": 0.1570676565170288, "learning_rate": 5.0697741615017524e-05, "loss": 0.0234, "step": 26040 }, { "action_loss": 0.0017390744760632515, "epoch": 23.41726618705036, "step": 26040 }, { "epoch": 23.41726618705036, "step": 26040, "torque_loss": 0.14632220566272736 }, { "epoch": 23.426258992805757, "grad_norm": 0.15077567100524902, "learning_rate": 5.06646752885467e-05, "loss": 0.0252, "step": 26050 }, { "action_loss": 0.0024331232998520136, "epoch": 23.426258992805757, "step": 26050 }, { "epoch": 23.426258992805757, "step": 26050, "torque_loss": 0.1819942146539688 }, { "epoch": 23.43525179856115, "grad_norm": 0.10671349614858627, "learning_rate": 5.0631608671324324e-05, "loss": 0.021, "step": 26060 }, { "action_loss": 0.00417243642732501, "epoch": 23.43525179856115, "step": 26060 }, { "epoch": 23.43525179856115, "step": 26060, "torque_loss": 0.1233770027756691 }, { "epoch": 23.444244604316548, "grad_norm": 0.13804306089878082, "learning_rate": 5.059854177781487e-05, "loss": 0.0229, "step": 26070 }, { "action_loss": 0.0033368095755577087, "epoch": 23.444244604316548, "step": 26070 }, { "epoch": 23.444244604316548, "step": 26070, "torque_loss": 0.16108550131320953 }, { "epoch": 23.453237410071942, "grad_norm": 0.11896970123052597, "learning_rate": 5.05654746224829e-05, "loss": 0.0204, "step": 26080 }, { "action_loss": 0.0061957999132573605, "epoch": 23.453237410071942, "step": 26080 }, { "epoch": 23.453237410071942, "step": 26080, "torque_loss": 0.14518703520298004 }, { "epoch": 23.46223021582734, "grad_norm": 0.1532876044511795, "learning_rate": 5.053240721979311e-05, "loss": 0.0218, "step": 26090 }, { "action_loss": 0.004344310611486435, "epoch": 23.46223021582734, "step": 26090 }, { "epoch": 23.46223021582734, "step": 26090, "torque_loss": 0.16016538441181183 }, { "epoch": 23.471223021582734, "grad_norm": 0.10081236064434052, "learning_rate": 5.049933958421031e-05, "loss": 0.0232, "step": 26100 }, { "action_loss": 0.002877580700442195, "epoch": 23.471223021582734, "step": 26100 }, { "epoch": 23.471223021582734, "step": 26100, "torque_loss": 0.1992940455675125 }, { "epoch": 23.480215827338128, "grad_norm": 0.1294994205236435, "learning_rate": 5.046627173019943e-05, "loss": 0.0213, "step": 26110 }, { "action_loss": 0.0018512429669499397, "epoch": 23.480215827338128, "step": 26110 }, { "epoch": 23.480215827338128, "step": 26110, "torque_loss": 0.13618165254592896 }, { "epoch": 23.489208633093526, "grad_norm": 0.08837675303220749, "learning_rate": 5.043320367222545e-05, "loss": 0.0235, "step": 26120 }, { "action_loss": 0.002692128298804164, "epoch": 23.489208633093526, "step": 26120 }, { "epoch": 23.489208633093526, "step": 26120, "torque_loss": 0.12883630394935608 }, { "epoch": 23.49820143884892, "grad_norm": 0.16701149940490723, "learning_rate": 5.040013542475346e-05, "loss": 0.0187, "step": 26130 }, { "action_loss": 0.0022566523402929306, "epoch": 23.49820143884892, "step": 26130 }, { "epoch": 23.49820143884892, "step": 26130, "torque_loss": 0.14948658645153046 }, { "epoch": 23.507194244604317, "grad_norm": 0.09675898402929306, "learning_rate": 5.0367067002248646e-05, "loss": 0.0187, "step": 26140 }, { "action_loss": 0.004869204480201006, "epoch": 23.507194244604317, "step": 26140 }, { "epoch": 23.507194244604317, "step": 26140, "torque_loss": 0.18425090610980988 }, { "epoch": 23.51618705035971, "grad_norm": 0.10076414048671722, "learning_rate": 5.0333998419176254e-05, "loss": 0.0208, "step": 26150 }, { "action_loss": 0.004113116767257452, "epoch": 23.51618705035971, "step": 26150 }, { "epoch": 23.51618705035971, "step": 26150, "torque_loss": 0.18480725586414337 }, { "epoch": 23.52517985611511, "grad_norm": 0.20769789814949036, "learning_rate": 5.030092969000161e-05, "loss": 0.0225, "step": 26160 }, { "action_loss": 0.0018326876452192664, "epoch": 23.52517985611511, "step": 26160 }, { "epoch": 23.52517985611511, "step": 26160, "torque_loss": 0.12907367944717407 }, { "epoch": 23.534172661870503, "grad_norm": 0.15760579705238342, "learning_rate": 5.026786082919008e-05, "loss": 0.0256, "step": 26170 }, { "action_loss": 0.004649663344025612, "epoch": 23.534172661870503, "step": 26170 }, { "epoch": 23.534172661870503, "step": 26170, "torque_loss": 0.15821242332458496 }, { "epoch": 23.5431654676259, "grad_norm": 0.15988075733184814, "learning_rate": 5.0234791851207144e-05, "loss": 0.0209, "step": 26180 }, { "action_loss": 0.003332264954224229, "epoch": 23.5431654676259, "step": 26180 }, { "epoch": 23.5431654676259, "step": 26180, "torque_loss": 0.13674281537532806 }, { "epoch": 23.552158273381295, "grad_norm": 0.1739148646593094, "learning_rate": 5.0201722770518256e-05, "loss": 0.0232, "step": 26190 }, { "action_loss": 0.004171749111264944, "epoch": 23.552158273381295, "step": 26190 }, { "epoch": 23.552158273381295, "step": 26190, "torque_loss": 0.12168864160776138 }, { "epoch": 23.56115107913669, "grad_norm": 0.15991154313087463, "learning_rate": 5.0168653601589e-05, "loss": 0.0233, "step": 26200 }, { "action_loss": 0.0026139551773667336, "epoch": 23.56115107913669, "step": 26200 }, { "epoch": 23.56115107913669, "step": 26200, "torque_loss": 0.16004213690757751 }, { "epoch": 23.570143884892087, "grad_norm": 0.1631985306739807, "learning_rate": 5.01355843588849e-05, "loss": 0.0247, "step": 26210 }, { "action_loss": 0.004537239670753479, "epoch": 23.570143884892087, "step": 26210 }, { "epoch": 23.570143884892087, "step": 26210, "torque_loss": 0.16817128658294678 }, { "epoch": 23.57913669064748, "grad_norm": 0.08685801178216934, "learning_rate": 5.010251505687161e-05, "loss": 0.02, "step": 26220 }, { "action_loss": 0.0023348238319158554, "epoch": 23.57913669064748, "step": 26220 }, { "epoch": 23.57913669064748, "step": 26220, "torque_loss": 0.13803496956825256 }, { "epoch": 23.58812949640288, "grad_norm": 0.1828140914440155, "learning_rate": 5.006944571001474e-05, "loss": 0.0244, "step": 26230 }, { "action_loss": 0.00433203624561429, "epoch": 23.58812949640288, "step": 26230 }, { "epoch": 23.58812949640288, "step": 26230, "torque_loss": 0.15460197627544403 }, { "epoch": 23.597122302158272, "grad_norm": 0.12189728766679764, "learning_rate": 5.003637633277997e-05, "loss": 0.0233, "step": 26240 }, { "action_loss": 0.08403456956148148, "epoch": 23.597122302158272, "step": 26240 }, { "epoch": 23.597122302158272, "step": 26240, "torque_loss": 0.21641166508197784 }, { "epoch": 23.60611510791367, "grad_norm": 0.12461790442466736, "learning_rate": 5.0003306939632946e-05, "loss": 0.031, "step": 26250 }, { "action_loss": 0.005019210744649172, "epoch": 23.60611510791367, "step": 26250 }, { "epoch": 23.60611510791367, "step": 26250, "torque_loss": 0.1690114140510559 }, { "epoch": 23.615107913669064, "grad_norm": 0.13439294695854187, "learning_rate": 4.997023754503937e-05, "loss": 0.0227, "step": 26260 }, { "action_loss": 0.0015032918890938163, "epoch": 23.615107913669064, "step": 26260 }, { "epoch": 23.615107913669064, "step": 26260, "torque_loss": 0.131398543715477 }, { "epoch": 23.62410071942446, "grad_norm": 0.14568272233009338, "learning_rate": 4.9937168163464897e-05, "loss": 0.0235, "step": 26270 }, { "action_loss": 0.0027591511607170105, "epoch": 23.62410071942446, "step": 26270 }, { "epoch": 23.62410071942446, "step": 26270, "torque_loss": 0.15505939722061157 }, { "epoch": 23.633093525179856, "grad_norm": 0.11902061849832535, "learning_rate": 4.990409880937519e-05, "loss": 0.0214, "step": 26280 }, { "action_loss": 0.0049167172983288765, "epoch": 23.633093525179856, "step": 26280 }, { "epoch": 23.633093525179856, "step": 26280, "torque_loss": 0.17178501188755035 }, { "epoch": 23.642086330935253, "grad_norm": 0.18084515631198883, "learning_rate": 4.9871029497235936e-05, "loss": 0.023, "step": 26290 }, { "action_loss": 0.00260282214730978, "epoch": 23.642086330935253, "step": 26290 }, { "epoch": 23.642086330935253, "step": 26290, "torque_loss": 0.08040245622396469 }, { "epoch": 23.651079136690647, "grad_norm": 0.13365943729877472, "learning_rate": 4.9837960241512766e-05, "loss": 0.0192, "step": 26300 }, { "action_loss": 0.006863368209451437, "epoch": 23.651079136690647, "step": 26300 }, { "epoch": 23.651079136690647, "step": 26300, "torque_loss": 0.15737344324588776 }, { "epoch": 23.66007194244604, "grad_norm": 0.14302489161491394, "learning_rate": 4.9804891056671265e-05, "loss": 0.0194, "step": 26310 }, { "action_loss": 0.004139429423958063, "epoch": 23.66007194244604, "step": 26310 }, { "epoch": 23.66007194244604, "step": 26310, "torque_loss": 0.1482887715101242 }, { "epoch": 23.66906474820144, "grad_norm": 0.13444365561008453, "learning_rate": 4.9771821957177074e-05, "loss": 0.0193, "step": 26320 }, { "action_loss": 0.004179096315056086, "epoch": 23.66906474820144, "step": 26320 }, { "epoch": 23.66906474820144, "step": 26320, "torque_loss": 0.13614065945148468 }, { "epoch": 23.678057553956833, "grad_norm": 0.11411213129758835, "learning_rate": 4.9738752957495694e-05, "loss": 0.0252, "step": 26330 }, { "action_loss": 0.0019289868650957942, "epoch": 23.678057553956833, "step": 26330 }, { "epoch": 23.678057553956833, "step": 26330, "torque_loss": 0.13258109986782074 }, { "epoch": 23.68705035971223, "grad_norm": 0.12558086216449738, "learning_rate": 4.970568407209265e-05, "loss": 0.0234, "step": 26340 }, { "action_loss": 0.004622471984475851, "epoch": 23.68705035971223, "step": 26340 }, { "epoch": 23.68705035971223, "step": 26340, "torque_loss": 0.14884252846240997 }, { "epoch": 23.696043165467625, "grad_norm": 0.10169040411710739, "learning_rate": 4.967261531543339e-05, "loss": 0.0225, "step": 26350 }, { "action_loss": 0.002300550928339362, "epoch": 23.696043165467625, "step": 26350 }, { "epoch": 23.696043165467625, "step": 26350, "torque_loss": 0.16660378873348236 }, { "epoch": 23.705035971223023, "grad_norm": 0.08591862767934799, "learning_rate": 4.963954670198332e-05, "loss": 0.0195, "step": 26360 }, { "action_loss": 0.004208078142255545, "epoch": 23.705035971223023, "step": 26360 }, { "epoch": 23.705035971223023, "step": 26360, "torque_loss": 0.14932848513126373 }, { "epoch": 23.714028776978417, "grad_norm": 0.11926119774580002, "learning_rate": 4.960647824620776e-05, "loss": 0.0194, "step": 26370 }, { "action_loss": 0.008385672233998775, "epoch": 23.714028776978417, "step": 26370 }, { "epoch": 23.714028776978417, "step": 26370, "torque_loss": 0.16803723573684692 }, { "epoch": 23.723021582733814, "grad_norm": 0.1332128494977951, "learning_rate": 4.9573409962571995e-05, "loss": 0.0192, "step": 26380 }, { "action_loss": 0.010380402207374573, "epoch": 23.723021582733814, "step": 26380 }, { "epoch": 23.723021582733814, "step": 26380, "torque_loss": 0.20952554047107697 }, { "epoch": 23.73201438848921, "grad_norm": 0.11685872822999954, "learning_rate": 4.954034186554121e-05, "loss": 0.0243, "step": 26390 }, { "action_loss": 0.0033560243900865316, "epoch": 23.73201438848921, "step": 26390 }, { "epoch": 23.73201438848921, "step": 26390, "torque_loss": 0.17750489711761475 }, { "epoch": 23.741007194244606, "grad_norm": 0.1998709887266159, "learning_rate": 4.950727396958051e-05, "loss": 0.0322, "step": 26400 }, { "action_loss": 0.002899646759033203, "epoch": 23.741007194244606, "step": 26400 }, { "epoch": 23.741007194244606, "step": 26400, "torque_loss": 0.11562615633010864 }, { "epoch": 23.75, "grad_norm": 0.08749103546142578, "learning_rate": 4.947420628915489e-05, "loss": 0.0231, "step": 26410 }, { "action_loss": 0.007048906292766333, "epoch": 23.75, "step": 26410 }, { "epoch": 23.75, "step": 26410, "torque_loss": 0.2024591714143753 }, { "epoch": 23.758992805755394, "grad_norm": 0.1500033736228943, "learning_rate": 4.944113883872932e-05, "loss": 0.0248, "step": 26420 }, { "action_loss": 0.0021078127902001143, "epoch": 23.758992805755394, "step": 26420 }, { "epoch": 23.758992805755394, "step": 26420, "torque_loss": 0.11161182075738907 }, { "epoch": 23.76798561151079, "grad_norm": 0.11774825304746628, "learning_rate": 4.9408071632768624e-05, "loss": 0.0231, "step": 26430 }, { "action_loss": 0.005939201917499304, "epoch": 23.76798561151079, "step": 26430 }, { "epoch": 23.76798561151079, "step": 26430, "torque_loss": 0.18040145933628082 }, { "epoch": 23.776978417266186, "grad_norm": 0.13871917128562927, "learning_rate": 4.937500468573747e-05, "loss": 0.0232, "step": 26440 }, { "action_loss": 0.00913160014897585, "epoch": 23.776978417266186, "step": 26440 }, { "epoch": 23.776978417266186, "step": 26440, "torque_loss": 0.22929199039936066 }, { "epoch": 23.785971223021583, "grad_norm": 0.09882478415966034, "learning_rate": 4.934193801210053e-05, "loss": 0.024, "step": 26450 }, { "action_loss": 0.007660467177629471, "epoch": 23.785971223021583, "step": 26450 }, { "epoch": 23.785971223021583, "step": 26450, "torque_loss": 0.19053463637828827 }, { "epoch": 23.794964028776977, "grad_norm": 0.12285587191581726, "learning_rate": 4.930887162632224e-05, "loss": 0.0232, "step": 26460 }, { "action_loss": 0.005700769368559122, "epoch": 23.794964028776977, "step": 26460 }, { "epoch": 23.794964028776977, "step": 26460, "torque_loss": 0.22370369732379913 }, { "epoch": 23.803956834532375, "grad_norm": 0.09472943842411041, "learning_rate": 4.9275805542866994e-05, "loss": 0.0265, "step": 26470 }, { "action_loss": 0.004183455370366573, "epoch": 23.803956834532375, "step": 26470 }, { "epoch": 23.803956834532375, "step": 26470, "torque_loss": 0.12401920557022095 }, { "epoch": 23.81294964028777, "grad_norm": 0.079226553440094, "learning_rate": 4.9242739776199006e-05, "loss": 0.0217, "step": 26480 }, { "action_loss": 0.003338893875479698, "epoch": 23.81294964028777, "step": 26480 }, { "epoch": 23.81294964028777, "step": 26480, "torque_loss": 0.18054623901844025 }, { "epoch": 23.821942446043167, "grad_norm": 0.15228629112243652, "learning_rate": 4.920967434078237e-05, "loss": 0.0284, "step": 26490 }, { "action_loss": 0.007680969312787056, "epoch": 23.821942446043167, "step": 26490 }, { "epoch": 23.821942446043167, "step": 26490, "torque_loss": 0.1862647533416748 }, { "epoch": 23.83093525179856, "grad_norm": 0.14881467819213867, "learning_rate": 4.917660925108101e-05, "loss": 0.0216, "step": 26500 }, { "action_loss": 0.0053643956780433655, "epoch": 23.83093525179856, "step": 26500 }, { "epoch": 23.83093525179856, "step": 26500, "torque_loss": 0.1426694244146347 }, { "epoch": 23.83992805755396, "grad_norm": 0.17119775712490082, "learning_rate": 4.914354452155876e-05, "loss": 0.0251, "step": 26510 }, { "action_loss": 0.002480622148141265, "epoch": 23.83992805755396, "step": 26510 }, { "epoch": 23.83992805755396, "step": 26510, "torque_loss": 0.17063303291797638 }, { "epoch": 23.848920863309353, "grad_norm": 0.16129398345947266, "learning_rate": 4.911048016667923e-05, "loss": 0.0236, "step": 26520 }, { "action_loss": 0.005083829630166292, "epoch": 23.848920863309353, "step": 26520 }, { "epoch": 23.848920863309353, "step": 26520, "torque_loss": 0.18037985265254974 }, { "epoch": 23.857913669064747, "grad_norm": 0.14344274997711182, "learning_rate": 4.907741620090588e-05, "loss": 0.0218, "step": 26530 }, { "action_loss": 0.010014726780354977, "epoch": 23.857913669064747, "step": 26530 }, { "epoch": 23.857913669064747, "step": 26530, "torque_loss": 0.2582177221775055 }, { "epoch": 23.866906474820144, "grad_norm": 0.15179352462291718, "learning_rate": 4.9044352638702046e-05, "loss": 0.0296, "step": 26540 }, { "action_loss": 0.012859261594712734, "epoch": 23.866906474820144, "step": 26540 }, { "epoch": 23.866906474820144, "step": 26540, "torque_loss": 0.1739397495985031 }, { "epoch": 23.87589928057554, "grad_norm": 0.1403455138206482, "learning_rate": 4.9011289494530834e-05, "loss": 0.0242, "step": 26550 }, { "action_loss": 0.0019434549612924457, "epoch": 23.87589928057554, "step": 26550 }, { "epoch": 23.87589928057554, "step": 26550, "torque_loss": 0.12086769193410873 }, { "epoch": 23.884892086330936, "grad_norm": 0.16395649313926697, "learning_rate": 4.897822678285516e-05, "loss": 0.0218, "step": 26560 }, { "action_loss": 0.00344690284691751, "epoch": 23.884892086330936, "step": 26560 }, { "epoch": 23.884892086330936, "step": 26560, "torque_loss": 0.13244779407978058 }, { "epoch": 23.89388489208633, "grad_norm": 0.1426006406545639, "learning_rate": 4.8945164518137835e-05, "loss": 0.0218, "step": 26570 }, { "action_loss": 0.007928355596959591, "epoch": 23.89388489208633, "step": 26570 }, { "epoch": 23.89388489208633, "step": 26570, "torque_loss": 0.15911729633808136 }, { "epoch": 23.902877697841728, "grad_norm": 0.1310887336730957, "learning_rate": 4.891210271484138e-05, "loss": 0.0296, "step": 26580 }, { "action_loss": 0.006006209179759026, "epoch": 23.902877697841728, "step": 26580 }, { "epoch": 23.902877697841728, "step": 26580, "torque_loss": 0.1520402431488037 }, { "epoch": 23.91187050359712, "grad_norm": 0.15357927978038788, "learning_rate": 4.8879041387428164e-05, "loss": 0.0273, "step": 26590 }, { "action_loss": 0.024286672472953796, "epoch": 23.91187050359712, "step": 26590 }, { "epoch": 23.91187050359712, "step": 26590, "torque_loss": 0.25956133008003235 }, { "epoch": 23.92086330935252, "grad_norm": 0.2298935204744339, "learning_rate": 4.8845980550360306e-05, "loss": 0.0268, "step": 26600 }, { "action_loss": 0.0030073849484324455, "epoch": 23.92086330935252, "step": 26600 }, { "epoch": 23.92086330935252, "step": 26600, "torque_loss": 0.07734090834856033 }, { "epoch": 23.929856115107913, "grad_norm": 0.0932997614145279, "learning_rate": 4.8812920218099786e-05, "loss": 0.0197, "step": 26610 }, { "action_loss": 0.008038226515054703, "epoch": 23.929856115107913, "step": 26610 }, { "epoch": 23.929856115107913, "step": 26610, "torque_loss": 0.17693953216075897 }, { "epoch": 23.93884892086331, "grad_norm": 0.11694427579641342, "learning_rate": 4.8779860405108294e-05, "loss": 0.02, "step": 26620 }, { "action_loss": 0.0025053909048438072, "epoch": 23.93884892086331, "step": 26620 }, { "epoch": 23.93884892086331, "step": 26620, "torque_loss": 0.1516590267419815 }, { "epoch": 23.947841726618705, "grad_norm": 0.0884697288274765, "learning_rate": 4.8746801125847285e-05, "loss": 0.0213, "step": 26630 }, { "action_loss": 0.005148632917553186, "epoch": 23.947841726618705, "step": 26630 }, { "epoch": 23.947841726618705, "step": 26630, "torque_loss": 0.16363196074962616 }, { "epoch": 23.9568345323741, "grad_norm": 0.11278025805950165, "learning_rate": 4.8713742394778064e-05, "loss": 0.0244, "step": 26640 }, { "action_loss": 0.0029926488641649485, "epoch": 23.9568345323741, "step": 26640 }, { "epoch": 23.9568345323741, "step": 26640, "torque_loss": 0.10129974037408829 }, { "epoch": 23.965827338129497, "grad_norm": 0.17556484043598175, "learning_rate": 4.868068422636162e-05, "loss": 0.0222, "step": 26650 }, { "action_loss": 0.009272074326872826, "epoch": 23.965827338129497, "step": 26650 }, { "epoch": 23.965827338129497, "step": 26650, "torque_loss": 0.25187167525291443 }, { "epoch": 23.97482014388489, "grad_norm": 0.16949014365673065, "learning_rate": 4.86476266350587e-05, "loss": 0.0299, "step": 26660 }, { "action_loss": 0.00487523665651679, "epoch": 23.97482014388489, "step": 26660 }, { "epoch": 23.97482014388489, "step": 26660, "torque_loss": 0.14316754043102264 }, { "epoch": 23.98381294964029, "grad_norm": 0.16737371683120728, "learning_rate": 4.861456963532985e-05, "loss": 0.0223, "step": 26670 }, { "action_loss": 0.0025757630355656147, "epoch": 23.98381294964029, "step": 26670 }, { "epoch": 23.98381294964029, "step": 26670, "torque_loss": 0.11509964615106583 }, { "epoch": 23.992805755395683, "grad_norm": 0.11965494602918625, "learning_rate": 4.8581513241635316e-05, "loss": 0.0208, "step": 26680 }, { "action_loss": 0.004285641014575958, "epoch": 23.992805755395683, "step": 26680 }, { "epoch": 23.992805755395683, "step": 26680, "torque_loss": 0.17739911377429962 }, { "epoch": 24.00179856115108, "grad_norm": 0.10174641758203506, "learning_rate": 4.8548457468435065e-05, "loss": 0.0203, "step": 26690 }, { "action_loss": 0.0014743931824341416, "epoch": 24.00179856115108, "step": 26690 }, { "epoch": 24.00179856115108, "step": 26690, "torque_loss": 0.11501789838075638 }, { "epoch": 24.010791366906474, "grad_norm": 0.12525318562984467, "learning_rate": 4.851540233018884e-05, "loss": 0.0194, "step": 26700 }, { "action_loss": 0.00446607731282711, "epoch": 24.010791366906474, "step": 26700 }, { "epoch": 24.010791366906474, "step": 26700, "torque_loss": 0.1104874387383461 }, { "epoch": 24.019784172661872, "grad_norm": 0.1385360062122345, "learning_rate": 4.848234784135608e-05, "loss": 0.0242, "step": 26710 }, { "action_loss": 0.0027518931310623884, "epoch": 24.019784172661872, "step": 26710 }, { "epoch": 24.019784172661872, "step": 26710, "torque_loss": 0.13690966367721558 }, { "epoch": 24.028776978417266, "grad_norm": 0.13381418585777283, "learning_rate": 4.8449294016395916e-05, "loss": 0.0202, "step": 26720 }, { "action_loss": 0.01531184185296297, "epoch": 24.028776978417266, "step": 26720 }, { "epoch": 24.028776978417266, "step": 26720, "torque_loss": 0.19639605283737183 }, { "epoch": 24.037769784172664, "grad_norm": 0.13910481333732605, "learning_rate": 4.8416240869767246e-05, "loss": 0.0239, "step": 26730 }, { "action_loss": 0.0026123712304979563, "epoch": 24.037769784172664, "step": 26730 }, { "epoch": 24.037769784172664, "step": 26730, "torque_loss": 0.11859352141618729 }, { "epoch": 24.046762589928058, "grad_norm": 0.180426687002182, "learning_rate": 4.838318841592864e-05, "loss": 0.0202, "step": 26740 }, { "action_loss": 0.008129849098622799, "epoch": 24.046762589928058, "step": 26740 }, { "epoch": 24.046762589928058, "step": 26740, "torque_loss": 0.16472427546977997 }, { "epoch": 24.055755395683452, "grad_norm": 0.14158602058887482, "learning_rate": 4.835013666933834e-05, "loss": 0.0229, "step": 26750 }, { "action_loss": 0.002992800436913967, "epoch": 24.055755395683452, "step": 26750 }, { "epoch": 24.055755395683452, "step": 26750, "torque_loss": 0.15286381542682648 }, { "epoch": 24.06474820143885, "grad_norm": 0.09793687611818314, "learning_rate": 4.83170856444543e-05, "loss": 0.0223, "step": 26760 }, { "action_loss": 0.004911135416477919, "epoch": 24.06474820143885, "step": 26760 }, { "epoch": 24.06474820143885, "step": 26760, "torque_loss": 0.10956591367721558 }, { "epoch": 24.073741007194243, "grad_norm": 0.12796097993850708, "learning_rate": 4.828403535573419e-05, "loss": 0.0239, "step": 26770 }, { "action_loss": 0.003034821944311261, "epoch": 24.073741007194243, "step": 26770 }, { "epoch": 24.073741007194243, "step": 26770, "torque_loss": 0.13766932487487793 }, { "epoch": 24.08273381294964, "grad_norm": 0.10752294957637787, "learning_rate": 4.825098581763532e-05, "loss": 0.0222, "step": 26780 }, { "action_loss": 0.0019424440106377006, "epoch": 24.08273381294964, "step": 26780 }, { "epoch": 24.08273381294964, "step": 26780, "torque_loss": 0.1354731321334839 }, { "epoch": 24.091726618705035, "grad_norm": 0.06600068509578705, "learning_rate": 4.821793704461465e-05, "loss": 0.0195, "step": 26790 }, { "action_loss": 0.0038731638342142105, "epoch": 24.091726618705035, "step": 26790 }, { "epoch": 24.091726618705035, "step": 26790, "torque_loss": 0.2219896763563156 }, { "epoch": 24.100719424460433, "grad_norm": 0.14445777237415314, "learning_rate": 4.8184889051128884e-05, "loss": 0.0216, "step": 26800 }, { "action_loss": 0.006610054522752762, "epoch": 24.100719424460433, "step": 26800 }, { "epoch": 24.100719424460433, "step": 26800, "torque_loss": 0.1830577403306961 }, { "epoch": 24.109712230215827, "grad_norm": 0.12537525594234467, "learning_rate": 4.81518418516343e-05, "loss": 0.024, "step": 26810 }, { "action_loss": 0.0031895998399704695, "epoch": 24.109712230215827, "step": 26810 }, { "epoch": 24.109712230215827, "step": 26810, "torque_loss": 0.22940288484096527 }, { "epoch": 24.118705035971225, "grad_norm": 0.18636691570281982, "learning_rate": 4.811879546058688e-05, "loss": 0.0235, "step": 26820 }, { "action_loss": 0.006193190813064575, "epoch": 24.118705035971225, "step": 26820 }, { "epoch": 24.118705035971225, "step": 26820, "torque_loss": 0.16094054281711578 }, { "epoch": 24.12769784172662, "grad_norm": 0.15140286087989807, "learning_rate": 4.808574989244224e-05, "loss": 0.0222, "step": 26830 }, { "action_loss": 0.008937719278037548, "epoch": 24.12769784172662, "step": 26830 }, { "epoch": 24.12769784172662, "step": 26830, "torque_loss": 0.19074831902980804 }, { "epoch": 24.136690647482013, "grad_norm": 0.12297545373439789, "learning_rate": 4.8052705161655644e-05, "loss": 0.0234, "step": 26840 }, { "action_loss": 0.006861841771751642, "epoch": 24.136690647482013, "step": 26840 }, { "epoch": 24.136690647482013, "step": 26840, "torque_loss": 0.18485920131206512 }, { "epoch": 24.14568345323741, "grad_norm": 0.19806544482707977, "learning_rate": 4.8019661282681936e-05, "loss": 0.0272, "step": 26850 }, { "action_loss": 0.00328815053217113, "epoch": 24.14568345323741, "step": 26850 }, { "epoch": 24.14568345323741, "step": 26850, "torque_loss": 0.16219639778137207 }, { "epoch": 24.154676258992804, "grad_norm": 0.11493869870901108, "learning_rate": 4.79866182699757e-05, "loss": 0.0234, "step": 26860 }, { "action_loss": 0.001931338687427342, "epoch": 24.154676258992804, "step": 26860 }, { "epoch": 24.154676258992804, "step": 26860, "torque_loss": 0.12423322349786758 }, { "epoch": 24.163669064748202, "grad_norm": 0.13752472400665283, "learning_rate": 4.795357613799103e-05, "loss": 0.0252, "step": 26870 }, { "action_loss": 0.002801531692966819, "epoch": 24.163669064748202, "step": 26870 }, { "epoch": 24.163669064748202, "step": 26870, "torque_loss": 0.16240514814853668 }, { "epoch": 24.172661870503596, "grad_norm": 0.12072455137968063, "learning_rate": 4.792053490118166e-05, "loss": 0.0261, "step": 26880 }, { "action_loss": 0.0021683278027921915, "epoch": 24.172661870503596, "step": 26880 }, { "epoch": 24.172661870503596, "step": 26880, "torque_loss": 0.10938204079866409 }, { "epoch": 24.181654676258994, "grad_norm": 0.2103017419576645, "learning_rate": 4.788749457400099e-05, "loss": 0.0237, "step": 26890 }, { "action_loss": 0.00446072593331337, "epoch": 24.181654676258994, "step": 26890 }, { "epoch": 24.181654676258994, "step": 26890, "torque_loss": 0.17834343016147614 }, { "epoch": 24.190647482014388, "grad_norm": 0.12669554352760315, "learning_rate": 4.7854455170901966e-05, "loss": 0.0238, "step": 26900 }, { "action_loss": 0.011377627961337566, "epoch": 24.190647482014388, "step": 26900 }, { "epoch": 24.190647482014388, "step": 26900, "torque_loss": 0.17070530354976654 }, { "epoch": 24.199640287769785, "grad_norm": 0.16047996282577515, "learning_rate": 4.7821416706337136e-05, "loss": 0.024, "step": 26910 }, { "action_loss": 0.00336249265819788, "epoch": 24.199640287769785, "step": 26910 }, { "epoch": 24.199640287769785, "step": 26910, "torque_loss": 0.1318318247795105 }, { "epoch": 24.20863309352518, "grad_norm": 0.1179412379860878, "learning_rate": 4.7788379194758645e-05, "loss": 0.0216, "step": 26920 }, { "action_loss": 0.014476582407951355, "epoch": 24.20863309352518, "step": 26920 }, { "epoch": 24.20863309352518, "step": 26920, "torque_loss": 0.2497655749320984 }, { "epoch": 24.217625899280577, "grad_norm": 0.1509823501110077, "learning_rate": 4.7755342650618246e-05, "loss": 0.0221, "step": 26930 }, { "action_loss": 0.006336517166346312, "epoch": 24.217625899280577, "step": 26930 }, { "epoch": 24.217625899280577, "step": 26930, "torque_loss": 0.1898011565208435 }, { "epoch": 24.22661870503597, "grad_norm": 0.119803786277771, "learning_rate": 4.7722307088367235e-05, "loss": 0.024, "step": 26940 }, { "action_loss": 0.0018637818284332752, "epoch": 24.22661870503597, "step": 26940 }, { "epoch": 24.22661870503597, "step": 26940, "torque_loss": 0.17751669883728027 }, { "epoch": 24.235611510791365, "grad_norm": 0.11943319439888, "learning_rate": 4.768927252245646e-05, "loss": 0.0173, "step": 26950 }, { "action_loss": 0.0022880297619849443, "epoch": 24.235611510791365, "step": 26950 }, { "epoch": 24.235611510791365, "step": 26950, "torque_loss": 0.1694178432226181 }, { "epoch": 24.244604316546763, "grad_norm": 0.10776650160551071, "learning_rate": 4.765623896733642e-05, "loss": 0.0229, "step": 26960 }, { "action_loss": 0.005111176986247301, "epoch": 24.244604316546763, "step": 26960 }, { "epoch": 24.244604316546763, "step": 26960, "torque_loss": 0.18665213882923126 }, { "epoch": 24.253597122302157, "grad_norm": 0.1379559338092804, "learning_rate": 4.762320643745708e-05, "loss": 0.0303, "step": 26970 }, { "action_loss": 0.0031291854102164507, "epoch": 24.253597122302157, "step": 26970 }, { "epoch": 24.253597122302157, "step": 26970, "torque_loss": 0.15354298055171967 }, { "epoch": 24.262589928057555, "grad_norm": 0.10181007534265518, "learning_rate": 4.7590174947267966e-05, "loss": 0.0268, "step": 26980 }, { "action_loss": 0.02529303729534149, "epoch": 24.262589928057555, "step": 26980 }, { "epoch": 24.262589928057555, "step": 26980, "torque_loss": 0.1622001975774765 }, { "epoch": 24.27158273381295, "grad_norm": 0.15257012844085693, "learning_rate": 4.755714451121823e-05, "loss": 0.0239, "step": 26990 }, { "action_loss": 0.00595500273630023, "epoch": 24.27158273381295, "step": 26990 }, { "epoch": 24.27158273381295, "step": 26990, "torque_loss": 0.12720917165279388 }, { "epoch": 24.280575539568346, "grad_norm": 0.11154939234256744, "learning_rate": 4.7524115143756476e-05, "loss": 0.0224, "step": 27000 }, { "action_loss": 0.011271080933511257, "epoch": 24.280575539568346, "step": 27000 }, { "epoch": 24.280575539568346, "step": 27000, "torque_loss": 0.25480344891548157 }, { "epoch": 24.28956834532374, "grad_norm": 0.10428525507450104, "learning_rate": 4.7491086859330854e-05, "loss": 0.0218, "step": 27010 }, { "action_loss": 0.020943688228726387, "epoch": 24.28956834532374, "step": 27010 }, { "epoch": 24.28956834532374, "step": 27010, "torque_loss": 0.2429385930299759 }, { "epoch": 24.298561151079138, "grad_norm": 0.1629381775856018, "learning_rate": 4.74580596723891e-05, "loss": 0.0276, "step": 27020 }, { "action_loss": 0.010488796979188919, "epoch": 24.298561151079138, "step": 27020 }, { "epoch": 24.298561151079138, "step": 27020, "torque_loss": 0.23662392795085907 }, { "epoch": 24.307553956834532, "grad_norm": 0.18169212341308594, "learning_rate": 4.742503359737841e-05, "loss": 0.0213, "step": 27030 }, { "action_loss": 0.0023668641224503517, "epoch": 24.307553956834532, "step": 27030 }, { "epoch": 24.307553956834532, "step": 27030, "torque_loss": 0.20623524487018585 }, { "epoch": 24.31654676258993, "grad_norm": 0.16496948897838593, "learning_rate": 4.739200864874547e-05, "loss": 0.02, "step": 27040 }, { "action_loss": 0.014316685497760773, "epoch": 24.31654676258993, "step": 27040 }, { "epoch": 24.31654676258993, "step": 27040, "torque_loss": 0.16324563324451447 }, { "epoch": 24.325539568345324, "grad_norm": 0.16741475462913513, "learning_rate": 4.735898484093658e-05, "loss": 0.0204, "step": 27050 }, { "action_loss": 0.0034943793434649706, "epoch": 24.325539568345324, "step": 27050 }, { "epoch": 24.325539568345324, "step": 27050, "torque_loss": 0.14189405739307404 }, { "epoch": 24.334532374100718, "grad_norm": 0.11132337898015976, "learning_rate": 4.7325962188397444e-05, "loss": 0.0208, "step": 27060 }, { "action_loss": 0.021028459072113037, "epoch": 24.334532374100718, "step": 27060 }, { "epoch": 24.334532374100718, "step": 27060, "torque_loss": 0.20869894325733185 }, { "epoch": 24.343525179856115, "grad_norm": 0.14347071945667267, "learning_rate": 4.729294070557329e-05, "loss": 0.0235, "step": 27070 }, { "action_loss": 0.014729191549122334, "epoch": 24.343525179856115, "step": 27070 }, { "epoch": 24.343525179856115, "step": 27070, "torque_loss": 0.19674114882946014 }, { "epoch": 24.35251798561151, "grad_norm": 0.1463773548603058, "learning_rate": 4.7259920406908833e-05, "loss": 0.0268, "step": 27080 }, { "action_loss": 0.006251323968172073, "epoch": 24.35251798561151, "step": 27080 }, { "epoch": 24.35251798561151, "step": 27080, "torque_loss": 0.22120316326618195 }, { "epoch": 24.361510791366907, "grad_norm": 0.15834872424602509, "learning_rate": 4.7226901306848306e-05, "loss": 0.0225, "step": 27090 }, { "action_loss": 0.016511525958776474, "epoch": 24.361510791366907, "step": 27090 }, { "epoch": 24.361510791366907, "step": 27090, "torque_loss": 0.24720223248004913 }, { "epoch": 24.3705035971223, "grad_norm": 0.10555365681648254, "learning_rate": 4.719388341983535e-05, "loss": 0.025, "step": 27100 }, { "action_loss": 0.009141774848103523, "epoch": 24.3705035971223, "step": 27100 }, { "epoch": 24.3705035971223, "step": 27100, "torque_loss": 0.22052128612995148 }, { "epoch": 24.3794964028777, "grad_norm": 0.12273430079221725, "learning_rate": 4.71608667603131e-05, "loss": 0.0245, "step": 27110 }, { "action_loss": 0.005010738503187895, "epoch": 24.3794964028777, "step": 27110 }, { "epoch": 24.3794964028777, "step": 27110, "torque_loss": 0.17877967655658722 }, { "epoch": 24.388489208633093, "grad_norm": 0.11372718960046768, "learning_rate": 4.712785134272422e-05, "loss": 0.0252, "step": 27120 }, { "action_loss": 0.0031124132219702005, "epoch": 24.388489208633093, "step": 27120 }, { "epoch": 24.388489208633093, "step": 27120, "torque_loss": 0.1574135571718216 }, { "epoch": 24.39748201438849, "grad_norm": 0.1319555938243866, "learning_rate": 4.7094837181510734e-05, "loss": 0.0178, "step": 27130 }, { "action_loss": 0.0032843798398971558, "epoch": 24.39748201438849, "step": 27130 }, { "epoch": 24.39748201438849, "step": 27130, "torque_loss": 0.10249635577201843 }, { "epoch": 24.406474820143885, "grad_norm": 0.09592676907777786, "learning_rate": 4.706182429111414e-05, "loss": 0.0208, "step": 27140 }, { "action_loss": 0.05772559717297554, "epoch": 24.406474820143885, "step": 27140 }, { "epoch": 24.406474820143885, "step": 27140, "torque_loss": 0.21163170039653778 }, { "epoch": 24.415467625899282, "grad_norm": 0.15498970448970795, "learning_rate": 4.702881268597545e-05, "loss": 0.028, "step": 27150 }, { "action_loss": 0.0014200586592778563, "epoch": 24.415467625899282, "step": 27150 }, { "epoch": 24.415467625899282, "step": 27150, "torque_loss": 0.11208400875329971 }, { "epoch": 24.424460431654676, "grad_norm": 0.11015353351831436, "learning_rate": 4.699580238053503e-05, "loss": 0.0222, "step": 27160 }, { "action_loss": 0.018035491928458214, "epoch": 24.424460431654676, "step": 27160 }, { "epoch": 24.424460431654676, "step": 27160, "torque_loss": 0.19876784086227417 }, { "epoch": 24.43345323741007, "grad_norm": 0.09317728132009506, "learning_rate": 4.69627933892327e-05, "loss": 0.0239, "step": 27170 }, { "action_loss": 0.0031177245546132326, "epoch": 24.43345323741007, "step": 27170 }, { "epoch": 24.43345323741007, "step": 27170, "torque_loss": 0.15651638805866241 }, { "epoch": 24.442446043165468, "grad_norm": 0.12089903652667999, "learning_rate": 4.692978572650773e-05, "loss": 0.0196, "step": 27180 }, { "action_loss": 0.0032018572092056274, "epoch": 24.442446043165468, "step": 27180 }, { "epoch": 24.442446043165468, "step": 27180, "torque_loss": 0.15516728162765503 }, { "epoch": 24.451438848920862, "grad_norm": 0.11059103161096573, "learning_rate": 4.689677940679881e-05, "loss": 0.0252, "step": 27190 }, { "action_loss": 0.007235704455524683, "epoch": 24.451438848920862, "step": 27190 }, { "epoch": 24.451438848920862, "step": 27190, "torque_loss": 0.19588983058929443 }, { "epoch": 24.46043165467626, "grad_norm": 0.13607922196388245, "learning_rate": 4.6863774444543984e-05, "loss": 0.0227, "step": 27200 }, { "action_loss": 0.0028133466839790344, "epoch": 24.46043165467626, "step": 27200 }, { "epoch": 24.46043165467626, "step": 27200, "torque_loss": 0.18671256303787231 }, { "epoch": 24.469424460431654, "grad_norm": 0.10003270953893661, "learning_rate": 4.6830770854180784e-05, "loss": 0.0211, "step": 27210 }, { "action_loss": 0.0023756902664899826, "epoch": 24.469424460431654, "step": 27210 }, { "epoch": 24.469424460431654, "step": 27210, "torque_loss": 0.11151387542486191 }, { "epoch": 24.47841726618705, "grad_norm": 0.08567650616168976, "learning_rate": 4.679776865014608e-05, "loss": 0.0239, "step": 27220 }, { "action_loss": 0.004873606376349926, "epoch": 24.47841726618705, "step": 27220 }, { "epoch": 24.47841726618705, "step": 27220, "torque_loss": 0.23614509403705597 }, { "epoch": 24.487410071942445, "grad_norm": 0.14769697189331055, "learning_rate": 4.676476784687615e-05, "loss": 0.0253, "step": 27230 }, { "action_loss": 0.0033025210723280907, "epoch": 24.487410071942445, "step": 27230 }, { "epoch": 24.487410071942445, "step": 27230, "torque_loss": 0.17041485011577606 }, { "epoch": 24.496402877697843, "grad_norm": 0.15999074280261993, "learning_rate": 4.673176845880669e-05, "loss": 0.0229, "step": 27240 }, { "action_loss": 0.010381183587014675, "epoch": 24.496402877697843, "step": 27240 }, { "epoch": 24.496402877697843, "step": 27240, "torque_loss": 0.21486584842205048 }, { "epoch": 24.505395683453237, "grad_norm": 0.15448035299777985, "learning_rate": 4.669877050037275e-05, "loss": 0.0263, "step": 27250 }, { "action_loss": 0.002341276267543435, "epoch": 24.505395683453237, "step": 27250 }, { "epoch": 24.505395683453237, "step": 27250, "torque_loss": 0.13715308904647827 }, { "epoch": 24.514388489208635, "grad_norm": 0.16360194981098175, "learning_rate": 4.666577398600874e-05, "loss": 0.0242, "step": 27260 }, { "action_loss": 0.003105379408225417, "epoch": 24.514388489208635, "step": 27260 }, { "epoch": 24.514388489208635, "step": 27260, "torque_loss": 0.14195516705513 }, { "epoch": 24.52338129496403, "grad_norm": 0.1675805002450943, "learning_rate": 4.663277893014846e-05, "loss": 0.0226, "step": 27270 }, { "action_loss": 0.0028803665190935135, "epoch": 24.52338129496403, "step": 27270 }, { "epoch": 24.52338129496403, "step": 27270, "torque_loss": 0.16959714889526367 }, { "epoch": 24.532374100719423, "grad_norm": 0.13525408506393433, "learning_rate": 4.659978534722508e-05, "loss": 0.0247, "step": 27280 }, { "action_loss": 0.0037877503782510757, "epoch": 24.532374100719423, "step": 27280 }, { "epoch": 24.532374100719423, "step": 27280, "torque_loss": 0.13841618597507477 }, { "epoch": 24.54136690647482, "grad_norm": 0.10213790833950043, "learning_rate": 4.6566793251671104e-05, "loss": 0.0216, "step": 27290 }, { "action_loss": 0.003498675301671028, "epoch": 24.54136690647482, "step": 27290 }, { "epoch": 24.54136690647482, "step": 27290, "torque_loss": 0.13535262644290924 }, { "epoch": 24.550359712230215, "grad_norm": 0.17160356044769287, "learning_rate": 4.65338026579184e-05, "loss": 0.0233, "step": 27300 }, { "action_loss": 0.011626475490629673, "epoch": 24.550359712230215, "step": 27300 }, { "epoch": 24.550359712230215, "step": 27300, "torque_loss": 0.22349822521209717 }, { "epoch": 24.559352517985612, "grad_norm": 0.19819793105125427, "learning_rate": 4.650081358039817e-05, "loss": 0.027, "step": 27310 }, { "action_loss": 0.004815492779016495, "epoch": 24.559352517985612, "step": 27310 }, { "epoch": 24.559352517985612, "step": 27310, "torque_loss": 0.17725424468517303 }, { "epoch": 24.568345323741006, "grad_norm": 0.13799160718917847, "learning_rate": 4.646782603354098e-05, "loss": 0.0203, "step": 27320 }, { "action_loss": 0.008510948158800602, "epoch": 24.568345323741006, "step": 27320 }, { "epoch": 24.568345323741006, "step": 27320, "torque_loss": 0.19002318382263184 }, { "epoch": 24.577338129496404, "grad_norm": 0.11106012016534805, "learning_rate": 4.643484003177666e-05, "loss": 0.0225, "step": 27330 }, { "action_loss": 0.004957386292517185, "epoch": 24.577338129496404, "step": 27330 }, { "epoch": 24.577338129496404, "step": 27330, "torque_loss": 0.1478305459022522 }, { "epoch": 24.586330935251798, "grad_norm": 0.1419718712568283, "learning_rate": 4.6401855589534454e-05, "loss": 0.0187, "step": 27340 }, { "action_loss": 0.002918557496741414, "epoch": 24.586330935251798, "step": 27340 }, { "epoch": 24.586330935251798, "step": 27340, "torque_loss": 0.11257755756378174 }, { "epoch": 24.595323741007196, "grad_norm": 0.11783216148614883, "learning_rate": 4.636887272124286e-05, "loss": 0.0268, "step": 27350 }, { "action_loss": 0.0021545588970184326, "epoch": 24.595323741007196, "step": 27350 }, { "epoch": 24.595323741007196, "step": 27350, "torque_loss": 0.1040005311369896 }, { "epoch": 24.60431654676259, "grad_norm": 0.15413735806941986, "learning_rate": 4.6335891441329685e-05, "loss": 0.025, "step": 27360 }, { "action_loss": 0.010469019412994385, "epoch": 24.60431654676259, "step": 27360 }, { "epoch": 24.60431654676259, "step": 27360, "torque_loss": 0.25111958384513855 }, { "epoch": 24.613309352517987, "grad_norm": 0.15614508092403412, "learning_rate": 4.630291176422209e-05, "loss": 0.0247, "step": 27370 }, { "action_loss": 0.005835400428622961, "epoch": 24.613309352517987, "step": 27370 }, { "epoch": 24.613309352517987, "step": 27370, "torque_loss": 0.15023942291736603 }, { "epoch": 24.62230215827338, "grad_norm": 0.09509376436471939, "learning_rate": 4.62699337043465e-05, "loss": 0.0205, "step": 27380 }, { "action_loss": 0.010941880755126476, "epoch": 24.62230215827338, "step": 27380 }, { "epoch": 24.62230215827338, "step": 27380, "torque_loss": 0.2559201717376709 }, { "epoch": 24.631294964028775, "grad_norm": 0.12065871804952621, "learning_rate": 4.623695727612861e-05, "loss": 0.0256, "step": 27390 }, { "action_loss": 0.0018352697370573878, "epoch": 24.631294964028775, "step": 27390 }, { "epoch": 24.631294964028775, "step": 27390, "torque_loss": 0.12601828575134277 }, { "epoch": 24.640287769784173, "grad_norm": 0.09421553462743759, "learning_rate": 4.620398249399347e-05, "loss": 0.0209, "step": 27400 }, { "action_loss": 0.006105158478021622, "epoch": 24.640287769784173, "step": 27400 }, { "epoch": 24.640287769784173, "step": 27400, "torque_loss": 0.2544954717159271 }, { "epoch": 24.649280575539567, "grad_norm": 0.25355497002601624, "learning_rate": 4.617100937236535e-05, "loss": 0.0227, "step": 27410 }, { "action_loss": 0.0032501574605703354, "epoch": 24.649280575539567, "step": 27410 }, { "epoch": 24.649280575539567, "step": 27410, "torque_loss": 0.149628147482872 }, { "epoch": 24.658273381294965, "grad_norm": 0.16567669808864594, "learning_rate": 4.613803792566781e-05, "loss": 0.0267, "step": 27420 }, { "action_loss": 0.0021990155801177025, "epoch": 24.658273381294965, "step": 27420 }, { "epoch": 24.658273381294965, "step": 27420, "torque_loss": 0.1523781269788742 }, { "epoch": 24.66726618705036, "grad_norm": 0.13975991308689117, "learning_rate": 4.61050681683237e-05, "loss": 0.0172, "step": 27430 }, { "action_loss": 0.0025836594868451357, "epoch": 24.66726618705036, "step": 27430 }, { "epoch": 24.66726618705036, "step": 27430, "torque_loss": 0.15645861625671387 }, { "epoch": 24.676258992805757, "grad_norm": 0.15658169984817505, "learning_rate": 4.6072100114755104e-05, "loss": 0.0228, "step": 27440 }, { "action_loss": 0.004140668548643589, "epoch": 24.676258992805757, "step": 27440 }, { "epoch": 24.676258992805757, "step": 27440, "torque_loss": 0.2158084362745285 }, { "epoch": 24.68525179856115, "grad_norm": 0.1333329826593399, "learning_rate": 4.603913377938336e-05, "loss": 0.0238, "step": 27450 }, { "action_loss": 0.006493519991636276, "epoch": 24.68525179856115, "step": 27450 }, { "epoch": 24.68525179856115, "step": 27450, "torque_loss": 0.18629200756549835 }, { "epoch": 24.694244604316548, "grad_norm": 0.14299806952476501, "learning_rate": 4.600616917662905e-05, "loss": 0.0239, "step": 27460 }, { "action_loss": 0.002036165213212371, "epoch": 24.694244604316548, "step": 27460 }, { "epoch": 24.694244604316548, "step": 27460, "torque_loss": 0.11383044719696045 }, { "epoch": 24.703237410071942, "grad_norm": 0.13086697459220886, "learning_rate": 4.597320632091206e-05, "loss": 0.0188, "step": 27470 }, { "action_loss": 0.008502217940986156, "epoch": 24.703237410071942, "step": 27470 }, { "epoch": 24.703237410071942, "step": 27470, "torque_loss": 0.23281006515026093 }, { "epoch": 24.71223021582734, "grad_norm": 0.1095411404967308, "learning_rate": 4.594024522665142e-05, "loss": 0.0247, "step": 27480 }, { "action_loss": 0.06051628291606903, "epoch": 24.71223021582734, "step": 27480 }, { "epoch": 24.71223021582734, "step": 27480, "torque_loss": 0.2987652122974396 }, { "epoch": 24.721223021582734, "grad_norm": 0.1486675888299942, "learning_rate": 4.590728590826543e-05, "loss": 0.0266, "step": 27490 }, { "action_loss": 0.003267914755269885, "epoch": 24.721223021582734, "step": 27490 }, { "epoch": 24.721223021582734, "step": 27490, "torque_loss": 0.20352625846862793 }, { "epoch": 24.730215827338128, "grad_norm": 0.14439904689788818, "learning_rate": 4.587432838017165e-05, "loss": 0.0267, "step": 27500 }, { "action_loss": 0.002776360372081399, "epoch": 24.730215827338128, "step": 27500 }, { "epoch": 24.730215827338128, "step": 27500, "torque_loss": 0.12453710287809372 }, { "epoch": 24.739208633093526, "grad_norm": 0.14095830917358398, "learning_rate": 4.58413726567868e-05, "loss": 0.0202, "step": 27510 }, { "action_loss": 0.006524639204144478, "epoch": 24.739208633093526, "step": 27510 }, { "epoch": 24.739208633093526, "step": 27510, "torque_loss": 0.14111439883708954 }, { "epoch": 24.74820143884892, "grad_norm": 0.19374331831932068, "learning_rate": 4.5808418752526814e-05, "loss": 0.0228, "step": 27520 }, { "action_loss": 0.007144020404666662, "epoch": 24.74820143884892, "step": 27520 }, { "epoch": 24.74820143884892, "step": 27520, "torque_loss": 0.22002585232257843 }, { "epoch": 24.757194244604317, "grad_norm": 0.16827693581581116, "learning_rate": 4.5775466681806895e-05, "loss": 0.0232, "step": 27530 }, { "action_loss": 0.010897154919803143, "epoch": 24.757194244604317, "step": 27530 }, { "epoch": 24.757194244604317, "step": 27530, "torque_loss": 0.22319185733795166 }, { "epoch": 24.76618705035971, "grad_norm": 0.2211882323026657, "learning_rate": 4.574251645904136e-05, "loss": 0.0247, "step": 27540 }, { "action_loss": 0.004429883323609829, "epoch": 24.76618705035971, "step": 27540 }, { "epoch": 24.76618705035971, "step": 27540, "torque_loss": 0.1693265438079834 }, { "epoch": 24.77517985611511, "grad_norm": 0.1506786346435547, "learning_rate": 4.5709568098643776e-05, "loss": 0.0216, "step": 27550 }, { "action_loss": 0.012045949697494507, "epoch": 24.77517985611511, "step": 27550 }, { "epoch": 24.77517985611511, "step": 27550, "torque_loss": 0.22262780368328094 }, { "epoch": 24.784172661870503, "grad_norm": 0.13078148663043976, "learning_rate": 4.567662161502687e-05, "loss": 0.0209, "step": 27560 }, { "action_loss": 0.003929849248379469, "epoch": 24.784172661870503, "step": 27560 }, { "epoch": 24.784172661870503, "step": 27560, "torque_loss": 0.15130607783794403 }, { "epoch": 24.7931654676259, "grad_norm": 0.13625560700893402, "learning_rate": 4.5643677022602554e-05, "loss": 0.026, "step": 27570 }, { "action_loss": 0.003192029194906354, "epoch": 24.7931654676259, "step": 27570 }, { "epoch": 24.7931654676259, "step": 27570, "torque_loss": 0.12553943693637848 }, { "epoch": 24.802158273381295, "grad_norm": 0.09755675494670868, "learning_rate": 4.56107343357819e-05, "loss": 0.0216, "step": 27580 }, { "action_loss": 0.003872040892019868, "epoch": 24.802158273381295, "step": 27580 }, { "epoch": 24.802158273381295, "step": 27580, "torque_loss": 0.19287008047103882 }, { "epoch": 24.81115107913669, "grad_norm": 0.1095246896147728, "learning_rate": 4.557779356897518e-05, "loss": 0.0255, "step": 27590 }, { "action_loss": 0.0033716789912432432, "epoch": 24.81115107913669, "step": 27590 }, { "epoch": 24.81115107913669, "step": 27590, "torque_loss": 0.19370782375335693 }, { "epoch": 24.820143884892087, "grad_norm": 0.19932590425014496, "learning_rate": 4.55448547365918e-05, "loss": 0.0234, "step": 27600 }, { "action_loss": 0.011215589940547943, "epoch": 24.820143884892087, "step": 27600 }, { "epoch": 24.820143884892087, "step": 27600, "torque_loss": 0.30139705538749695 }, { "epoch": 24.82913669064748, "grad_norm": 0.12591733038425446, "learning_rate": 4.5511917853040315e-05, "loss": 0.0243, "step": 27610 }, { "action_loss": 0.005622644443064928, "epoch": 24.82913669064748, "step": 27610 }, { "epoch": 24.82913669064748, "step": 27610, "torque_loss": 0.21403782069683075 }, { "epoch": 24.83812949640288, "grad_norm": 0.09089886397123337, "learning_rate": 4.547898293272842e-05, "loss": 0.0261, "step": 27620 }, { "action_loss": 0.006969816982746124, "epoch": 24.83812949640288, "step": 27620 }, { "epoch": 24.83812949640288, "step": 27620, "torque_loss": 0.26555225253105164 }, { "epoch": 24.847122302158272, "grad_norm": 0.09924180805683136, "learning_rate": 4.544604999006301e-05, "loss": 0.0199, "step": 27630 }, { "action_loss": 0.0045326161198318005, "epoch": 24.847122302158272, "step": 27630 }, { "epoch": 24.847122302158272, "step": 27630, "torque_loss": 0.2096373587846756 }, { "epoch": 24.85611510791367, "grad_norm": 0.24549025297164917, "learning_rate": 4.541311903945004e-05, "loss": 0.0273, "step": 27640 }, { "action_loss": 0.027210479602217674, "epoch": 24.85611510791367, "step": 27640 }, { "epoch": 24.85611510791367, "step": 27640, "torque_loss": 0.19164127111434937 }, { "epoch": 24.865107913669064, "grad_norm": 0.16041001677513123, "learning_rate": 4.538019009529463e-05, "loss": 0.0227, "step": 27650 }, { "action_loss": 0.001636698842048645, "epoch": 24.865107913669064, "step": 27650 }, { "epoch": 24.865107913669064, "step": 27650, "torque_loss": 0.10260676592588425 }, { "epoch": 24.87410071942446, "grad_norm": 0.11807884275913239, "learning_rate": 4.5347263172001035e-05, "loss": 0.0213, "step": 27660 }, { "action_loss": 0.002742191543802619, "epoch": 24.87410071942446, "step": 27660 }, { "epoch": 24.87410071942446, "step": 27660, "torque_loss": 0.23657715320587158 }, { "epoch": 24.883093525179856, "grad_norm": 0.1131807491183281, "learning_rate": 4.531433828397261e-05, "loss": 0.0233, "step": 27670 }, { "action_loss": 0.0024291444569826126, "epoch": 24.883093525179856, "step": 27670 }, { "epoch": 24.883093525179856, "step": 27670, "torque_loss": 0.15829795598983765 }, { "epoch": 24.892086330935253, "grad_norm": 0.19062285125255585, "learning_rate": 4.528141544561179e-05, "loss": 0.0214, "step": 27680 }, { "action_loss": 0.005026606842875481, "epoch": 24.892086330935253, "step": 27680 }, { "epoch": 24.892086330935253, "step": 27680, "torque_loss": 0.13803784549236298 }, { "epoch": 24.901079136690647, "grad_norm": 0.14234161376953125, "learning_rate": 4.5248494671320186e-05, "loss": 0.0204, "step": 27690 }, { "action_loss": 0.002902928041294217, "epoch": 24.901079136690647, "step": 27690 }, { "epoch": 24.901079136690647, "step": 27690, "torque_loss": 0.15030789375305176 }, { "epoch": 24.91007194244604, "grad_norm": 0.10047154873609543, "learning_rate": 4.521557597549844e-05, "loss": 0.0203, "step": 27700 }, { "action_loss": 0.03849456086754799, "epoch": 24.91007194244604, "step": 27700 }, { "epoch": 24.91007194244604, "step": 27700, "torque_loss": 0.1811387538909912 }, { "epoch": 24.91906474820144, "grad_norm": 0.1674501746892929, "learning_rate": 4.5182659372546284e-05, "loss": 0.0273, "step": 27710 }, { "action_loss": 0.002584988484159112, "epoch": 24.91906474820144, "step": 27710 }, { "epoch": 24.91906474820144, "step": 27710, "torque_loss": 0.12857280671596527 }, { "epoch": 24.928057553956833, "grad_norm": 0.10465491563081741, "learning_rate": 4.514974487686261e-05, "loss": 0.0215, "step": 27720 }, { "action_loss": 0.001344221062026918, "epoch": 24.928057553956833, "step": 27720 }, { "epoch": 24.928057553956833, "step": 27720, "torque_loss": 0.16132043302059174 }, { "epoch": 24.93705035971223, "grad_norm": 0.1367192417383194, "learning_rate": 4.5116832502845306e-05, "loss": 0.0208, "step": 27730 }, { "action_loss": 0.007130756974220276, "epoch": 24.93705035971223, "step": 27730 }, { "epoch": 24.93705035971223, "step": 27730, "torque_loss": 0.12777823209762573 }, { "epoch": 24.946043165467625, "grad_norm": 0.11539722979068756, "learning_rate": 4.508392226489136e-05, "loss": 0.0195, "step": 27740 }, { "action_loss": 0.0015659675700590014, "epoch": 24.946043165467625, "step": 27740 }, { "epoch": 24.946043165467625, "step": 27740, "torque_loss": 0.16185010969638824 }, { "epoch": 24.955035971223023, "grad_norm": 0.12134676426649094, "learning_rate": 4.505101417739683e-05, "loss": 0.0238, "step": 27750 }, { "action_loss": 0.002062130719423294, "epoch": 24.955035971223023, "step": 27750 }, { "epoch": 24.955035971223023, "step": 27750, "torque_loss": 0.10668975114822388 }, { "epoch": 24.964028776978417, "grad_norm": 0.12052172422409058, "learning_rate": 4.501810825475685e-05, "loss": 0.0197, "step": 27760 }, { "action_loss": 0.004086682107299566, "epoch": 24.964028776978417, "step": 27760 }, { "epoch": 24.964028776978417, "step": 27760, "torque_loss": 0.12366127222776413 }, { "epoch": 24.973021582733814, "grad_norm": 0.10713327676057816, "learning_rate": 4.4985204511365556e-05, "loss": 0.0231, "step": 27770 }, { "action_loss": 0.0020233073737472296, "epoch": 24.973021582733814, "step": 27770 }, { "epoch": 24.973021582733814, "step": 27770, "torque_loss": 0.11464069038629532 }, { "epoch": 24.98201438848921, "grad_norm": 0.1205105111002922, "learning_rate": 4.49523029616162e-05, "loss": 0.0233, "step": 27780 }, { "action_loss": 0.005808198358863592, "epoch": 24.98201438848921, "step": 27780 }, { "epoch": 24.98201438848921, "step": 27780, "torque_loss": 0.1837887018918991 }, { "epoch": 24.991007194244606, "grad_norm": 0.2148618847131729, "learning_rate": 4.491940361990101e-05, "loss": 0.0243, "step": 27790 }, { "action_loss": 0.002606813795864582, "epoch": 24.991007194244606, "step": 27790 }, { "epoch": 24.991007194244606, "step": 27790, "torque_loss": 0.15685728192329407 }, { "epoch": 25.0, "grad_norm": 0.1446496993303299, "learning_rate": 4.48865065006113e-05, "loss": 0.0255, "step": 27800 }, { "action_loss": 0.004512217361479998, "epoch": 25.0, "step": 27800 }, { "epoch": 25.0, "step": 27800, "torque_loss": 0.12046683579683304 }, { "epoch": 25.008992805755394, "grad_norm": 0.14950087666511536, "learning_rate": 4.485361161813735e-05, "loss": 0.0217, "step": 27810 }, { "action_loss": 0.007476126775145531, "epoch": 25.008992805755394, "step": 27810 }, { "epoch": 25.008992805755394, "step": 27810, "torque_loss": 0.19265496730804443 }, { "epoch": 25.01798561151079, "grad_norm": 0.1768384873867035, "learning_rate": 4.482071898686854e-05, "loss": 0.0224, "step": 27820 }, { "action_loss": 0.0022253075148910284, "epoch": 25.01798561151079, "step": 27820 }, { "epoch": 25.01798561151079, "step": 27820, "torque_loss": 0.1258188933134079 }, { "epoch": 25.026978417266186, "grad_norm": 0.10103258490562439, "learning_rate": 4.478782862119321e-05, "loss": 0.0213, "step": 27830 }, { "action_loss": 0.0016376575222238898, "epoch": 25.026978417266186, "step": 27830 }, { "epoch": 25.026978417266186, "step": 27830, "torque_loss": 0.13858692348003387 }, { "epoch": 25.035971223021583, "grad_norm": 0.10260771214962006, "learning_rate": 4.475494053549869e-05, "loss": 0.0275, "step": 27840 }, { "action_loss": 0.004479337949305773, "epoch": 25.035971223021583, "step": 27840 }, { "epoch": 25.035971223021583, "step": 27840, "torque_loss": 0.15828172862529755 }, { "epoch": 25.044964028776977, "grad_norm": 0.189952090382576, "learning_rate": 4.4722054744171405e-05, "loss": 0.023, "step": 27850 }, { "action_loss": 0.0032615249510854483, "epoch": 25.044964028776977, "step": 27850 }, { "epoch": 25.044964028776977, "step": 27850, "torque_loss": 0.14147506654262543 }, { "epoch": 25.053956834532375, "grad_norm": 0.09739566594362259, "learning_rate": 4.468917126159669e-05, "loss": 0.0189, "step": 27860 }, { "action_loss": 0.008305385708808899, "epoch": 25.053956834532375, "step": 27860 }, { "epoch": 25.053956834532375, "step": 27860, "torque_loss": 0.19906610250473022 }, { "epoch": 25.06294964028777, "grad_norm": 0.09743918478488922, "learning_rate": 4.465629010215887e-05, "loss": 0.0243, "step": 27870 }, { "action_loss": 0.008783289231359959, "epoch": 25.06294964028777, "step": 27870 }, { "epoch": 25.06294964028777, "step": 27870, "torque_loss": 0.15576417744159698 }, { "epoch": 25.071942446043167, "grad_norm": 0.11603810638189316, "learning_rate": 4.462341128024133e-05, "loss": 0.0184, "step": 27880 }, { "action_loss": 0.013216051273047924, "epoch": 25.071942446043167, "step": 27880 }, { "epoch": 25.071942446043167, "step": 27880, "torque_loss": 0.22297227382659912 }, { "epoch": 25.08093525179856, "grad_norm": 0.12549829483032227, "learning_rate": 4.4590534810226356e-05, "loss": 0.0237, "step": 27890 }, { "action_loss": 0.009483401663601398, "epoch": 25.08093525179856, "step": 27890 }, { "epoch": 25.08093525179856, "step": 27890, "torque_loss": 0.24888913333415985 }, { "epoch": 25.08992805755396, "grad_norm": 0.15513356029987335, "learning_rate": 4.455766070649523e-05, "loss": 0.0276, "step": 27900 }, { "action_loss": 0.008529405109584332, "epoch": 25.08992805755396, "step": 27900 }, { "epoch": 25.08992805755396, "step": 27900, "torque_loss": 0.15145882964134216 }, { "epoch": 25.098920863309353, "grad_norm": 0.15166372060775757, "learning_rate": 4.4524788983428215e-05, "loss": 0.0199, "step": 27910 }, { "action_loss": 0.003106704680249095, "epoch": 25.098920863309353, "step": 27910 }, { "epoch": 25.098920863309353, "step": 27910, "torque_loss": 0.16179277002811432 }, { "epoch": 25.107913669064747, "grad_norm": 0.118801549077034, "learning_rate": 4.4491919655404515e-05, "loss": 0.0216, "step": 27920 }, { "action_loss": 0.002424175152555108, "epoch": 25.107913669064747, "step": 27920 }, { "epoch": 25.107913669064747, "step": 27920, "torque_loss": 0.1419239193201065 }, { "epoch": 25.116906474820144, "grad_norm": 0.07315818220376968, "learning_rate": 4.445905273680228e-05, "loss": 0.0199, "step": 27930 }, { "action_loss": 0.0038855865132063627, "epoch": 25.116906474820144, "step": 27930 }, { "epoch": 25.116906474820144, "step": 27930, "torque_loss": 0.16703300178050995 }, { "epoch": 25.12589928057554, "grad_norm": 0.11063680797815323, "learning_rate": 4.442618824199865e-05, "loss": 0.0216, "step": 27940 }, { "action_loss": 0.020262083038687706, "epoch": 25.12589928057554, "step": 27940 }, { "epoch": 25.12589928057554, "step": 27940, "torque_loss": 0.17791052162647247 }, { "epoch": 25.134892086330936, "grad_norm": 0.15532134473323822, "learning_rate": 4.4393326185369635e-05, "loss": 0.023, "step": 27950 }, { "action_loss": 0.006409835070371628, "epoch": 25.134892086330936, "step": 27950 }, { "epoch": 25.134892086330936, "step": 27950, "torque_loss": 0.19334352016448975 }, { "epoch": 25.14388489208633, "grad_norm": 0.11116760224103928, "learning_rate": 4.436046658129024e-05, "loss": 0.0208, "step": 27960 }, { "action_loss": 0.012684267945587635, "epoch": 25.14388489208633, "step": 27960 }, { "epoch": 25.14388489208633, "step": 27960, "torque_loss": 0.20225636661052704 }, { "epoch": 25.152877697841728, "grad_norm": 0.09643512964248657, "learning_rate": 4.432760944413433e-05, "loss": 0.023, "step": 27970 }, { "action_loss": 0.00592966889962554, "epoch": 25.152877697841728, "step": 27970 }, { "epoch": 25.152877697841728, "step": 27970, "torque_loss": 0.17179228365421295 }, { "epoch": 25.16187050359712, "grad_norm": 0.11754798144102097, "learning_rate": 4.4294754788274796e-05, "loss": 0.0201, "step": 27980 }, { "action_loss": 0.0015518688596785069, "epoch": 25.16187050359712, "step": 27980 }, { "epoch": 25.16187050359712, "step": 27980, "torque_loss": 0.11845464259386063 }, { "epoch": 25.17086330935252, "grad_norm": 0.1611069142818451, "learning_rate": 4.426190262808334e-05, "loss": 0.0194, "step": 27990 }, { "action_loss": 0.013110603205859661, "epoch": 25.17086330935252, "step": 27990 }, { "epoch": 25.17086330935252, "step": 27990, "torque_loss": 0.20870955288410187 }, { "epoch": 25.179856115107913, "grad_norm": 0.08146590739488602, "learning_rate": 4.42290529779306e-05, "loss": 0.0187, "step": 28000 }, { "action_loss": 0.004406276624649763, "epoch": 25.179856115107913, "step": 28000 }, { "epoch": 25.179856115107913, "step": 28000, "torque_loss": 0.1498972326517105 }, { "epoch": 25.18884892086331, "grad_norm": 0.144245907664299, "learning_rate": 4.4196205852186176e-05, "loss": 0.0221, "step": 28010 }, { "action_loss": 0.0028640683740377426, "epoch": 25.18884892086331, "step": 28010 }, { "epoch": 25.18884892086331, "step": 28010, "torque_loss": 0.15541774034500122 }, { "epoch": 25.197841726618705, "grad_norm": 0.07920895516872406, "learning_rate": 4.416336126521848e-05, "loss": 0.0225, "step": 28020 }, { "action_loss": 0.003595103742554784, "epoch": 25.197841726618705, "step": 28020 }, { "epoch": 25.197841726618705, "step": 28020, "torque_loss": 0.16557441651821136 }, { "epoch": 25.2068345323741, "grad_norm": 0.10872795432806015, "learning_rate": 4.4130519231394865e-05, "loss": 0.0226, "step": 28030 }, { "action_loss": 0.002046237699687481, "epoch": 25.2068345323741, "step": 28030 }, { "epoch": 25.2068345323741, "step": 28030, "torque_loss": 0.13420550525188446 }, { "epoch": 25.215827338129497, "grad_norm": 0.14615988731384277, "learning_rate": 4.409767976508154e-05, "loss": 0.024, "step": 28040 }, { "action_loss": 0.007284549530595541, "epoch": 25.215827338129497, "step": 28040 }, { "epoch": 25.215827338129497, "step": 28040, "torque_loss": 0.1715562492609024 }, { "epoch": 25.22482014388489, "grad_norm": 0.1082322970032692, "learning_rate": 4.406484288064363e-05, "loss": 0.0195, "step": 28050 }, { "action_loss": 0.0027299318462610245, "epoch": 25.22482014388489, "step": 28050 }, { "epoch": 25.22482014388489, "step": 28050, "torque_loss": 0.15275011956691742 }, { "epoch": 25.23381294964029, "grad_norm": 0.14253783226013184, "learning_rate": 4.4032008592445055e-05, "loss": 0.0221, "step": 28060 }, { "action_loss": 0.007131289690732956, "epoch": 25.23381294964029, "step": 28060 }, { "epoch": 25.23381294964029, "step": 28060, "torque_loss": 0.17833000421524048 }, { "epoch": 25.242805755395683, "grad_norm": 0.22384706139564514, "learning_rate": 4.3999176914848714e-05, "loss": 0.0232, "step": 28070 }, { "action_loss": 0.0014928249875083566, "epoch": 25.242805755395683, "step": 28070 }, { "epoch": 25.242805755395683, "step": 28070, "torque_loss": 0.1469803899526596 }, { "epoch": 25.25179856115108, "grad_norm": 0.12846295535564423, "learning_rate": 4.396634786221625e-05, "loss": 0.0198, "step": 28080 }, { "action_loss": 0.0037724778521806, "epoch": 25.25179856115108, "step": 28080 }, { "epoch": 25.25179856115108, "step": 28080, "torque_loss": 0.11389940977096558 }, { "epoch": 25.260791366906474, "grad_norm": 0.16501262784004211, "learning_rate": 4.3933521448908203e-05, "loss": 0.0194, "step": 28090 }, { "action_loss": 0.002503708703443408, "epoch": 25.260791366906474, "step": 28090 }, { "epoch": 25.260791366906474, "step": 28090, "torque_loss": 0.15453918278217316 }, { "epoch": 25.269784172661872, "grad_norm": 0.09631391614675522, "learning_rate": 4.3900697689284e-05, "loss": 0.0212, "step": 28100 }, { "action_loss": 0.0026545876171439886, "epoch": 25.269784172661872, "step": 28100 }, { "epoch": 25.269784172661872, "step": 28100, "torque_loss": 0.20481355488300323 }, { "epoch": 25.278776978417266, "grad_norm": 0.113540418446064, "learning_rate": 4.386787659770185e-05, "loss": 0.0212, "step": 28110 }, { "action_loss": 0.003227679990231991, "epoch": 25.278776978417266, "step": 28110 }, { "epoch": 25.278776978417266, "step": 28110, "torque_loss": 0.14522285759449005 }, { "epoch": 25.28776978417266, "grad_norm": 0.1850171536207199, "learning_rate": 4.383505818851878e-05, "loss": 0.024, "step": 28120 }, { "action_loss": 0.003825524589046836, "epoch": 25.28776978417266, "step": 28120 }, { "epoch": 25.28776978417266, "step": 28120, "torque_loss": 0.1553659588098526 }, { "epoch": 25.296762589928058, "grad_norm": 0.16736236214637756, "learning_rate": 4.380224247609073e-05, "loss": 0.0214, "step": 28130 }, { "action_loss": 0.00573207950219512, "epoch": 25.296762589928058, "step": 28130 }, { "epoch": 25.296762589928058, "step": 28130, "torque_loss": 0.19999738037586212 }, { "epoch": 25.305755395683452, "grad_norm": 0.2347375601530075, "learning_rate": 4.3769429474772376e-05, "loss": 0.0226, "step": 28140 }, { "action_loss": 0.01694120280444622, "epoch": 25.305755395683452, "step": 28140 }, { "epoch": 25.305755395683452, "step": 28140, "torque_loss": 0.2766306698322296 }, { "epoch": 25.31474820143885, "grad_norm": 0.12758556008338928, "learning_rate": 4.373661919891726e-05, "loss": 0.0226, "step": 28150 }, { "action_loss": 0.007324086036533117, "epoch": 25.31474820143885, "step": 28150 }, { "epoch": 25.31474820143885, "step": 28150, "torque_loss": 0.1718282699584961 }, { "epoch": 25.323741007194243, "grad_norm": 0.08969476819038391, "learning_rate": 4.3703811662877674e-05, "loss": 0.0201, "step": 28160 }, { "action_loss": 0.004058807622641325, "epoch": 25.323741007194243, "step": 28160 }, { "epoch": 25.323741007194243, "step": 28160, "torque_loss": 0.14388054609298706 }, { "epoch": 25.33273381294964, "grad_norm": 0.1129944771528244, "learning_rate": 4.36710068810048e-05, "loss": 0.0188, "step": 28170 }, { "action_loss": 0.002745836740359664, "epoch": 25.33273381294964, "step": 28170 }, { "epoch": 25.33273381294964, "step": 28170, "torque_loss": 0.12477225065231323 }, { "epoch": 25.341726618705035, "grad_norm": 0.07400233298540115, "learning_rate": 4.363820486764853e-05, "loss": 0.0223, "step": 28180 }, { "action_loss": 0.0014700364554300904, "epoch": 25.341726618705035, "step": 28180 }, { "epoch": 25.341726618705035, "step": 28180, "torque_loss": 0.14163924753665924 }, { "epoch": 25.350719424460433, "grad_norm": 0.09190073609352112, "learning_rate": 4.360540563715758e-05, "loss": 0.0185, "step": 28190 }, { "action_loss": 0.0019950640853494406, "epoch": 25.350719424460433, "step": 28190 }, { "epoch": 25.350719424460433, "step": 28190, "torque_loss": 0.20684003829956055 }, { "epoch": 25.359712230215827, "grad_norm": 0.1622389703989029, "learning_rate": 4.357260920387946e-05, "loss": 0.0244, "step": 28200 }, { "action_loss": 0.008915973827242851, "epoch": 25.359712230215827, "step": 28200 }, { "epoch": 25.359712230215827, "step": 28200, "torque_loss": 0.11956175416707993 }, { "epoch": 25.368705035971225, "grad_norm": 0.09985308349132538, "learning_rate": 4.353981558216045e-05, "loss": 0.0195, "step": 28210 }, { "action_loss": 0.0015346179716289043, "epoch": 25.368705035971225, "step": 28210 }, { "epoch": 25.368705035971225, "step": 28210, "torque_loss": 0.13376984000205994 }, { "epoch": 25.37769784172662, "grad_norm": 0.13577821850776672, "learning_rate": 4.3507024786345565e-05, "loss": 0.0195, "step": 28220 }, { "action_loss": 0.005264170467853546, "epoch": 25.37769784172662, "step": 28220 }, { "epoch": 25.37769784172662, "step": 28220, "torque_loss": 0.128196582198143 }, { "epoch": 25.386690647482013, "grad_norm": 0.10726791620254517, "learning_rate": 4.3474236830778646e-05, "loss": 0.0198, "step": 28230 }, { "action_loss": 0.002361505525186658, "epoch": 25.386690647482013, "step": 28230 }, { "epoch": 25.386690647482013, "step": 28230, "torque_loss": 0.1807606965303421 }, { "epoch": 25.39568345323741, "grad_norm": 0.09854821860790253, "learning_rate": 4.344145172980223e-05, "loss": 0.0216, "step": 28240 }, { "action_loss": 0.0027708124835044146, "epoch": 25.39568345323741, "step": 28240 }, { "epoch": 25.39568345323741, "step": 28240, "torque_loss": 0.1233232244849205 }, { "epoch": 25.404676258992804, "grad_norm": 0.16670218110084534, "learning_rate": 4.3408669497757656e-05, "loss": 0.024, "step": 28250 }, { "action_loss": 0.0019782879389822483, "epoch": 25.404676258992804, "step": 28250 }, { "epoch": 25.404676258992804, "step": 28250, "torque_loss": 0.11636938899755478 }, { "epoch": 25.413669064748202, "grad_norm": 0.11536840349435806, "learning_rate": 4.3375890148984976e-05, "loss": 0.0226, "step": 28260 }, { "action_loss": 0.002507193712517619, "epoch": 25.413669064748202, "step": 28260 }, { "epoch": 25.413669064748202, "step": 28260, "torque_loss": 0.10790161043405533 }, { "epoch": 25.422661870503596, "grad_norm": 0.14416952431201935, "learning_rate": 4.334311369782301e-05, "loss": 0.0233, "step": 28270 }, { "action_loss": 0.008342674002051353, "epoch": 25.422661870503596, "step": 28270 }, { "epoch": 25.422661870503596, "step": 28270, "torque_loss": 0.19621175527572632 }, { "epoch": 25.431654676258994, "grad_norm": 0.14454805850982666, "learning_rate": 4.3310340158609244e-05, "loss": 0.0223, "step": 28280 }, { "action_loss": 0.002403712598606944, "epoch": 25.431654676258994, "step": 28280 }, { "epoch": 25.431654676258994, "step": 28280, "torque_loss": 0.13489273190498352 }, { "epoch": 25.440647482014388, "grad_norm": 0.0793890729546547, "learning_rate": 4.327756954568e-05, "loss": 0.0207, "step": 28290 }, { "action_loss": 0.0024748770520091057, "epoch": 25.440647482014388, "step": 28290 }, { "epoch": 25.440647482014388, "step": 28290, "torque_loss": 0.11818430572748184 }, { "epoch": 25.449640287769785, "grad_norm": 0.12489881366491318, "learning_rate": 4.324480187337021e-05, "loss": 0.0201, "step": 28300 }, { "action_loss": 0.00379113364033401, "epoch": 25.449640287769785, "step": 28300 }, { "epoch": 25.449640287769785, "step": 28300, "torque_loss": 0.1885492354631424 }, { "epoch": 25.45863309352518, "grad_norm": 0.13403350114822388, "learning_rate": 4.3212037156013596e-05, "loss": 0.0246, "step": 28310 }, { "action_loss": 0.004515785723924637, "epoch": 25.45863309352518, "step": 28310 }, { "epoch": 25.45863309352518, "step": 28310, "torque_loss": 0.17501074075698853 }, { "epoch": 25.467625899280577, "grad_norm": 0.12589506804943085, "learning_rate": 4.317927540794251e-05, "loss": 0.0208, "step": 28320 }, { "action_loss": 0.012424024753272533, "epoch": 25.467625899280577, "step": 28320 }, { "epoch": 25.467625899280577, "step": 28320, "torque_loss": 0.20594970881938934 }, { "epoch": 25.47661870503597, "grad_norm": 0.0998511016368866, "learning_rate": 4.314651664348812e-05, "loss": 0.0264, "step": 28330 }, { "action_loss": 0.004497608635574579, "epoch": 25.47661870503597, "step": 28330 }, { "epoch": 25.47661870503597, "step": 28330, "torque_loss": 0.19287694990634918 }, { "epoch": 25.485611510791365, "grad_norm": 0.07340522110462189, "learning_rate": 4.3113760876980195e-05, "loss": 0.0221, "step": 28340 }, { "action_loss": 0.0034726494923233986, "epoch": 25.485611510791365, "step": 28340 }, { "epoch": 25.485611510791365, "step": 28340, "torque_loss": 0.18374527990818024 }, { "epoch": 25.494604316546763, "grad_norm": 0.124787338078022, "learning_rate": 4.308100812274719e-05, "loss": 0.0239, "step": 28350 }, { "action_loss": 0.0020148202311247587, "epoch": 25.494604316546763, "step": 28350 }, { "epoch": 25.494604316546763, "step": 28350, "torque_loss": 0.14961375296115875 }, { "epoch": 25.503597122302157, "grad_norm": 0.1415979415178299, "learning_rate": 4.304825839511632e-05, "loss": 0.0193, "step": 28360 }, { "action_loss": 0.0014589334605261683, "epoch": 25.503597122302157, "step": 28360 }, { "epoch": 25.503597122302157, "step": 28360, "torque_loss": 0.15681682527065277 }, { "epoch": 25.512589928057555, "grad_norm": 0.13793952763080597, "learning_rate": 4.3015511708413395e-05, "loss": 0.0214, "step": 28370 }, { "action_loss": 0.010196559131145477, "epoch": 25.512589928057555, "step": 28370 }, { "epoch": 25.512589928057555, "step": 28370, "torque_loss": 0.24453039467334747 }, { "epoch": 25.52158273381295, "grad_norm": 0.08100342750549316, "learning_rate": 4.298276807696294e-05, "loss": 0.0219, "step": 28380 }, { "action_loss": 0.0096847889944911, "epoch": 25.52158273381295, "step": 28380 }, { "epoch": 25.52158273381295, "step": 28380, "torque_loss": 0.16664113104343414 }, { "epoch": 25.530575539568346, "grad_norm": 0.11165519058704376, "learning_rate": 4.2950027515088136e-05, "loss": 0.0213, "step": 28390 }, { "action_loss": 0.004431576002389193, "epoch": 25.530575539568346, "step": 28390 }, { "epoch": 25.530575539568346, "step": 28390, "torque_loss": 0.2663022577762604 }, { "epoch": 25.53956834532374, "grad_norm": 0.14001323282718658, "learning_rate": 4.291729003711083e-05, "loss": 0.0225, "step": 28400 }, { "action_loss": 0.012286772020161152, "epoch": 25.53956834532374, "step": 28400 }, { "epoch": 25.53956834532374, "step": 28400, "torque_loss": 0.21420539915561676 }, { "epoch": 25.548561151079138, "grad_norm": 0.14982256293296814, "learning_rate": 4.2884555657351455e-05, "loss": 0.0221, "step": 28410 }, { "action_loss": 0.0030690934509038925, "epoch": 25.548561151079138, "step": 28410 }, { "epoch": 25.548561151079138, "step": 28410, "torque_loss": 0.19886834919452667 }, { "epoch": 25.557553956834532, "grad_norm": 0.15076768398284912, "learning_rate": 4.285182439012921e-05, "loss": 0.0213, "step": 28420 }, { "action_loss": 0.005456799175590277, "epoch": 25.557553956834532, "step": 28420 }, { "epoch": 25.557553956834532, "step": 28420, "torque_loss": 0.24081633985042572 }, { "epoch": 25.56654676258993, "grad_norm": 0.21796947717666626, "learning_rate": 4.281909624976183e-05, "loss": 0.0281, "step": 28430 }, { "action_loss": 0.002496053697541356, "epoch": 25.56654676258993, "step": 28430 }, { "epoch": 25.56654676258993, "step": 28430, "torque_loss": 0.15099646151065826 }, { "epoch": 25.575539568345324, "grad_norm": 0.1650753617286682, "learning_rate": 4.278637125056568e-05, "loss": 0.0203, "step": 28440 }, { "action_loss": 0.009498156607151031, "epoch": 25.575539568345324, "step": 28440 }, { "epoch": 25.575539568345324, "step": 28440, "torque_loss": 0.26972511410713196 }, { "epoch": 25.584532374100718, "grad_norm": 0.1358889788389206, "learning_rate": 4.2753649406855855e-05, "loss": 0.0242, "step": 28450 }, { "action_loss": 0.0035133585333824158, "epoch": 25.584532374100718, "step": 28450 }, { "epoch": 25.584532374100718, "step": 28450, "torque_loss": 0.10769341140985489 }, { "epoch": 25.593525179856115, "grad_norm": 0.13213832676410675, "learning_rate": 4.2720930732945966e-05, "loss": 0.0232, "step": 28460 }, { "action_loss": 0.006198357790708542, "epoch": 25.593525179856115, "step": 28460 }, { "epoch": 25.593525179856115, "step": 28460, "torque_loss": 0.19281822443008423 }, { "epoch": 25.60251798561151, "grad_norm": 0.09352119266986847, "learning_rate": 4.268821524314826e-05, "loss": 0.0198, "step": 28470 }, { "action_loss": 0.001705045229755342, "epoch": 25.60251798561151, "step": 28470 }, { "epoch": 25.60251798561151, "step": 28470, "torque_loss": 0.13860346376895905 }, { "epoch": 25.611510791366907, "grad_norm": 0.19211946427822113, "learning_rate": 4.2655502951773623e-05, "loss": 0.0206, "step": 28480 }, { "action_loss": 0.017754940316081047, "epoch": 25.611510791366907, "step": 28480 }, { "epoch": 25.611510791366907, "step": 28480, "torque_loss": 0.24274297058582306 }, { "epoch": 25.6205035971223, "grad_norm": 0.19121716916561127, "learning_rate": 4.262279387313152e-05, "loss": 0.0247, "step": 28490 }, { "action_loss": 0.004910919349640608, "epoch": 25.6205035971223, "step": 28490 }, { "epoch": 25.6205035971223, "step": 28490, "torque_loss": 0.13457806408405304 }, { "epoch": 25.6294964028777, "grad_norm": 0.07863488793373108, "learning_rate": 4.259008802153001e-05, "loss": 0.022, "step": 28500 }, { "action_loss": 0.003372481092810631, "epoch": 25.6294964028777, "step": 28500 }, { "epoch": 25.6294964028777, "step": 28500, "torque_loss": 0.16014142334461212 }, { "epoch": 25.638489208633093, "grad_norm": 0.14450344443321228, "learning_rate": 4.2557385411275735e-05, "loss": 0.0291, "step": 28510 }, { "action_loss": 0.001416989485733211, "epoch": 25.638489208633093, "step": 28510 }, { "epoch": 25.638489208633093, "step": 28510, "torque_loss": 0.14001105725765228 }, { "epoch": 25.64748201438849, "grad_norm": 0.1580708920955658, "learning_rate": 4.252468605667395e-05, "loss": 0.0199, "step": 28520 }, { "action_loss": 0.003805305575951934, "epoch": 25.64748201438849, "step": 28520 }, { "epoch": 25.64748201438849, "step": 28520, "torque_loss": 0.19701939821243286 }, { "epoch": 25.656474820143885, "grad_norm": 0.07489155977964401, "learning_rate": 4.2491989972028456e-05, "loss": 0.0217, "step": 28530 }, { "action_loss": 0.0021108221262693405, "epoch": 25.656474820143885, "step": 28530 }, { "epoch": 25.656474820143885, "step": 28530, "torque_loss": 0.12676851451396942 }, { "epoch": 25.665467625899282, "grad_norm": 0.11346941441297531, "learning_rate": 4.245929717164161e-05, "loss": 0.018, "step": 28540 }, { "action_loss": 0.00235982914455235, "epoch": 25.665467625899282, "step": 28540 }, { "epoch": 25.665467625899282, "step": 28540, "torque_loss": 0.14250707626342773 }, { "epoch": 25.674460431654676, "grad_norm": 0.11580733954906464, "learning_rate": 4.242660766981438e-05, "loss": 0.0214, "step": 28550 }, { "action_loss": 0.006159099284559488, "epoch": 25.674460431654676, "step": 28550 }, { "epoch": 25.674460431654676, "step": 28550, "torque_loss": 0.20388364791870117 }, { "epoch": 25.68345323741007, "grad_norm": 0.11056851595640182, "learning_rate": 4.2393921480846264e-05, "loss": 0.0219, "step": 28560 }, { "action_loss": 0.0017268186202272773, "epoch": 25.68345323741007, "step": 28560 }, { "epoch": 25.68345323741007, "step": 28560, "torque_loss": 0.12680411338806152 }, { "epoch": 25.692446043165468, "grad_norm": 0.1020786464214325, "learning_rate": 4.236123861903528e-05, "loss": 0.0176, "step": 28570 }, { "action_loss": 0.00944980327039957, "epoch": 25.692446043165468, "step": 28570 }, { "epoch": 25.692446043165468, "step": 28570, "torque_loss": 0.2171207219362259 }, { "epoch": 25.701438848920862, "grad_norm": 0.10418865829706192, "learning_rate": 4.232855909867806e-05, "loss": 0.0187, "step": 28580 }, { "action_loss": 0.004788719117641449, "epoch": 25.701438848920862, "step": 28580 }, { "epoch": 25.701438848920862, "step": 28580, "torque_loss": 0.15700815618038177 }, { "epoch": 25.71043165467626, "grad_norm": 0.1245153620839119, "learning_rate": 4.229588293406972e-05, "loss": 0.0208, "step": 28590 }, { "action_loss": 0.0054487488232553005, "epoch": 25.71043165467626, "step": 28590 }, { "epoch": 25.71043165467626, "step": 28590, "torque_loss": 0.17177818715572357 }, { "epoch": 25.719424460431654, "grad_norm": 0.08446118980646133, "learning_rate": 4.226321013950392e-05, "loss": 0.0202, "step": 28600 }, { "action_loss": 0.004016272258013487, "epoch": 25.719424460431654, "step": 28600 }, { "epoch": 25.719424460431654, "step": 28600, "torque_loss": 0.1649494171142578 }, { "epoch": 25.72841726618705, "grad_norm": 0.10520760715007782, "learning_rate": 4.223054072927286e-05, "loss": 0.0191, "step": 28610 }, { "action_loss": 0.003237060271203518, "epoch": 25.72841726618705, "step": 28610 }, { "epoch": 25.72841726618705, "step": 28610, "torque_loss": 0.18610511720180511 }, { "epoch": 25.737410071942445, "grad_norm": 0.1317649632692337, "learning_rate": 4.219787471766724e-05, "loss": 0.0205, "step": 28620 }, { "action_loss": 0.0022436014842242002, "epoch": 25.737410071942445, "step": 28620 }, { "epoch": 25.737410071942445, "step": 28620, "torque_loss": 0.17078958451747894 }, { "epoch": 25.746402877697843, "grad_norm": 0.08526501059532166, "learning_rate": 4.216521211897629e-05, "loss": 0.0256, "step": 28630 }, { "action_loss": 0.006675835233181715, "epoch": 25.746402877697843, "step": 28630 }, { "epoch": 25.746402877697843, "step": 28630, "torque_loss": 0.15970580279827118 }, { "epoch": 25.755395683453237, "grad_norm": 0.11559554934501648, "learning_rate": 4.2132552947487734e-05, "loss": 0.0228, "step": 28640 }, { "action_loss": 0.002921121893450618, "epoch": 25.755395683453237, "step": 28640 }, { "epoch": 25.755395683453237, "step": 28640, "torque_loss": 0.15714608132839203 }, { "epoch": 25.764388489208635, "grad_norm": 0.15411534905433655, "learning_rate": 4.209989721748783e-05, "loss": 0.0256, "step": 28650 }, { "action_loss": 0.004831206519156694, "epoch": 25.764388489208635, "step": 28650 }, { "epoch": 25.764388489208635, "step": 28650, "torque_loss": 0.18338799476623535 }, { "epoch": 25.77338129496403, "grad_norm": 0.18259376287460327, "learning_rate": 4.206724494326127e-05, "loss": 0.0235, "step": 28660 }, { "action_loss": 0.004398395773023367, "epoch": 25.77338129496403, "step": 28660 }, { "epoch": 25.77338129496403, "step": 28660, "torque_loss": 0.10508159548044205 }, { "epoch": 25.782374100719423, "grad_norm": 0.11063174903392792, "learning_rate": 4.203459613909126e-05, "loss": 0.0201, "step": 28670 }, { "action_loss": 0.002701404271647334, "epoch": 25.782374100719423, "step": 28670 }, { "epoch": 25.782374100719423, "step": 28670, "torque_loss": 0.10292718559503555 }, { "epoch": 25.79136690647482, "grad_norm": 0.1567777395248413, "learning_rate": 4.200195081925955e-05, "loss": 0.0274, "step": 28680 }, { "action_loss": 0.007820982486009598, "epoch": 25.79136690647482, "step": 28680 }, { "epoch": 25.79136690647482, "step": 28680, "torque_loss": 0.22760678827762604 }, { "epoch": 25.800359712230215, "grad_norm": 0.08581152558326721, "learning_rate": 4.196930899804627e-05, "loss": 0.0224, "step": 28690 }, { "action_loss": 0.005272306501865387, "epoch": 25.800359712230215, "step": 28690 }, { "epoch": 25.800359712230215, "step": 28690, "torque_loss": 0.21074247360229492 }, { "epoch": 25.809352517985612, "grad_norm": 0.1733335554599762, "learning_rate": 4.1936670689730044e-05, "loss": 0.0229, "step": 28700 }, { "action_loss": 0.004036616533994675, "epoch": 25.809352517985612, "step": 28700 }, { "epoch": 25.809352517985612, "step": 28700, "torque_loss": 0.16361841559410095 }, { "epoch": 25.818345323741006, "grad_norm": 0.1404062658548355, "learning_rate": 4.190403590858804e-05, "loss": 0.0236, "step": 28710 }, { "action_loss": 0.006963518913835287, "epoch": 25.818345323741006, "step": 28710 }, { "epoch": 25.818345323741006, "step": 28710, "torque_loss": 0.16162829101085663 }, { "epoch": 25.827338129496404, "grad_norm": 0.17357084155082703, "learning_rate": 4.187140466889575e-05, "loss": 0.0267, "step": 28720 }, { "action_loss": 0.004593187477439642, "epoch": 25.827338129496404, "step": 28720 }, { "epoch": 25.827338129496404, "step": 28720, "torque_loss": 0.1999182552099228 }, { "epoch": 25.836330935251798, "grad_norm": 0.08258546888828278, "learning_rate": 4.183877698492724e-05, "loss": 0.0242, "step": 28730 }, { "action_loss": 0.0035086290445178747, "epoch": 25.836330935251798, "step": 28730 }, { "epoch": 25.836330935251798, "step": 28730, "torque_loss": 0.1472090631723404 }, { "epoch": 25.845323741007196, "grad_norm": 0.13652856647968292, "learning_rate": 4.1806152870954935e-05, "loss": 0.0205, "step": 28740 }, { "action_loss": 0.0052121845073997974, "epoch": 25.845323741007196, "step": 28740 }, { "epoch": 25.845323741007196, "step": 28740, "torque_loss": 0.1465599536895752 }, { "epoch": 25.85431654676259, "grad_norm": 0.10011620074510574, "learning_rate": 4.177353234124976e-05, "loss": 0.0233, "step": 28750 }, { "action_loss": 0.004026100970804691, "epoch": 25.85431654676259, "step": 28750 }, { "epoch": 25.85431654676259, "step": 28750, "torque_loss": 0.1578672081232071 }, { "epoch": 25.863309352517987, "grad_norm": 0.12123135477304459, "learning_rate": 4.1740915410081e-05, "loss": 0.0256, "step": 28760 }, { "action_loss": 0.0032627389300614595, "epoch": 25.863309352517987, "step": 28760 }, { "epoch": 25.863309352517987, "step": 28760, "torque_loss": 0.14970523118972778 }, { "epoch": 25.87230215827338, "grad_norm": 0.16715578734874725, "learning_rate": 4.170830209171646e-05, "loss": 0.0255, "step": 28770 }, { "action_loss": 0.005778031423687935, "epoch": 25.87230215827338, "step": 28770 }, { "epoch": 25.87230215827338, "step": 28770, "torque_loss": 0.15064562857151031 }, { "epoch": 25.881294964028775, "grad_norm": 0.13449867069721222, "learning_rate": 4.1675692400422286e-05, "loss": 0.0169, "step": 28780 }, { "action_loss": 0.0018326104618608952, "epoch": 25.881294964028775, "step": 28780 }, { "epoch": 25.881294964028775, "step": 28780, "torque_loss": 0.10478680580854416 }, { "epoch": 25.890287769784173, "grad_norm": 0.15678630769252777, "learning_rate": 4.164308635046304e-05, "loss": 0.0197, "step": 28790 }, { "action_loss": 0.01152393501251936, "epoch": 25.890287769784173, "step": 28790 }, { "epoch": 25.890287769784173, "step": 28790, "torque_loss": 0.21437211334705353 }, { "epoch": 25.899280575539567, "grad_norm": 0.14973974227905273, "learning_rate": 4.161048395610177e-05, "loss": 0.0239, "step": 28800 }, { "action_loss": 0.006735544186085463, "epoch": 25.899280575539567, "step": 28800 }, { "epoch": 25.899280575539567, "step": 28800, "torque_loss": 0.146545872092247 }, { "epoch": 25.908273381294965, "grad_norm": 0.17768187820911407, "learning_rate": 4.157788523159985e-05, "loss": 0.0225, "step": 28810 }, { "action_loss": 0.006389678921550512, "epoch": 25.908273381294965, "step": 28810 }, { "epoch": 25.908273381294965, "step": 28810, "torque_loss": 0.1945662945508957 }, { "epoch": 25.91726618705036, "grad_norm": 0.09089110791683197, "learning_rate": 4.154529019121706e-05, "loss": 0.0236, "step": 28820 }, { "action_loss": 0.031031429767608643, "epoch": 25.91726618705036, "step": 28820 }, { "epoch": 25.91726618705036, "step": 28820, "torque_loss": 0.21911419928073883 }, { "epoch": 25.926258992805757, "grad_norm": 0.11740414053201675, "learning_rate": 4.1512698849211564e-05, "loss": 0.023, "step": 28830 }, { "action_loss": 0.001910169143229723, "epoch": 25.926258992805757, "step": 28830 }, { "epoch": 25.926258992805757, "step": 28830, "torque_loss": 0.14140081405639648 }, { "epoch": 25.93525179856115, "grad_norm": 0.10531394928693771, "learning_rate": 4.1480111219839954e-05, "loss": 0.0193, "step": 28840 }, { "action_loss": 0.0022395679261535406, "epoch": 25.93525179856115, "step": 28840 }, { "epoch": 25.93525179856115, "step": 28840, "torque_loss": 0.1551615446805954 }, { "epoch": 25.944244604316548, "grad_norm": 0.0947795882821083, "learning_rate": 4.144752731735714e-05, "loss": 0.0193, "step": 28850 }, { "action_loss": 0.003322310047224164, "epoch": 25.944244604316548, "step": 28850 }, { "epoch": 25.944244604316548, "step": 28850, "torque_loss": 0.14300738275051117 }, { "epoch": 25.953237410071942, "grad_norm": 0.09266085922718048, "learning_rate": 4.1414947156016446e-05, "loss": 0.0166, "step": 28860 }, { "action_loss": 0.0007353471009992063, "epoch": 25.953237410071942, "step": 28860 }, { "epoch": 25.953237410071942, "step": 28860, "torque_loss": 0.0897320881485939 }, { "epoch": 25.96223021582734, "grad_norm": 0.16091670095920563, "learning_rate": 4.138237075006953e-05, "loss": 0.0184, "step": 28870 }, { "action_loss": 0.005946214776486158, "epoch": 25.96223021582734, "step": 28870 }, { "epoch": 25.96223021582734, "step": 28870, "torque_loss": 0.2110566645860672 }, { "epoch": 25.971223021582734, "grad_norm": 0.10500064492225647, "learning_rate": 4.134979811376645e-05, "loss": 0.0206, "step": 28880 }, { "action_loss": 0.006917253136634827, "epoch": 25.971223021582734, "step": 28880 }, { "epoch": 25.971223021582734, "step": 28880, "torque_loss": 0.14202642440795898 }, { "epoch": 25.980215827338128, "grad_norm": 0.10998032242059708, "learning_rate": 4.131722926135552e-05, "loss": 0.022, "step": 28890 }, { "action_loss": 0.001641315990127623, "epoch": 25.980215827338128, "step": 28890 }, { "epoch": 25.980215827338128, "step": 28890, "torque_loss": 0.13504478335380554 }, { "epoch": 25.989208633093526, "grad_norm": 0.10791103541851044, "learning_rate": 4.1284664207083526e-05, "loss": 0.0296, "step": 28900 }, { "action_loss": 0.002712105168029666, "epoch": 25.989208633093526, "step": 28900 }, { "epoch": 25.989208633093526, "step": 28900, "torque_loss": 0.11908525228500366 }, { "epoch": 25.99820143884892, "grad_norm": 0.18382038176059723, "learning_rate": 4.125210296519551e-05, "loss": 0.0206, "step": 28910 }, { "action_loss": 0.0026197193656116724, "epoch": 25.99820143884892, "step": 28910 }, { "epoch": 25.99820143884892, "step": 28910, "torque_loss": 0.1706715077161789 }, { "epoch": 26.007194244604317, "grad_norm": 0.12210680544376373, "learning_rate": 4.121954554993484e-05, "loss": 0.023, "step": 28920 }, { "action_loss": 0.0033911988139152527, "epoch": 26.007194244604317, "step": 28920 }, { "epoch": 26.007194244604317, "step": 28920, "torque_loss": 0.12769602239131927 }, { "epoch": 26.01618705035971, "grad_norm": 0.12031745165586472, "learning_rate": 4.118699197554327e-05, "loss": 0.0233, "step": 28930 }, { "action_loss": 0.005229547619819641, "epoch": 26.01618705035971, "step": 28930 }, { "epoch": 26.01618705035971, "step": 28930, "torque_loss": 0.15248648822307587 }, { "epoch": 26.02517985611511, "grad_norm": 0.15384982526302338, "learning_rate": 4.1154442256260814e-05, "loss": 0.0216, "step": 28940 }, { "action_loss": 0.012994606979191303, "epoch": 26.02517985611511, "step": 28940 }, { "epoch": 26.02517985611511, "step": 28940, "torque_loss": 0.22612206637859344 }, { "epoch": 26.034172661870503, "grad_norm": 0.164109468460083, "learning_rate": 4.1121896406325824e-05, "loss": 0.0225, "step": 28950 }, { "action_loss": 0.0043679578229784966, "epoch": 26.034172661870503, "step": 28950 }, { "epoch": 26.034172661870503, "step": 28950, "torque_loss": 0.17692343890666962 }, { "epoch": 26.0431654676259, "grad_norm": 0.18014094233512878, "learning_rate": 4.108935443997499e-05, "loss": 0.0284, "step": 28960 }, { "action_loss": 0.0022657145746052265, "epoch": 26.0431654676259, "step": 28960 }, { "epoch": 26.0431654676259, "step": 28960, "torque_loss": 0.1193445548415184 }, { "epoch": 26.052158273381295, "grad_norm": 0.11593759804964066, "learning_rate": 4.105681637144325e-05, "loss": 0.0232, "step": 28970 }, { "action_loss": 0.0016256101662293077, "epoch": 26.052158273381295, "step": 28970 }, { "epoch": 26.052158273381295, "step": 28970, "torque_loss": 0.13431380689144135 }, { "epoch": 26.06115107913669, "grad_norm": 0.17102880775928497, "learning_rate": 4.102428221496387e-05, "loss": 0.022, "step": 28980 }, { "action_loss": 0.0016963517991825938, "epoch": 26.06115107913669, "step": 28980 }, { "epoch": 26.06115107913669, "step": 28980, "torque_loss": 0.1291065216064453 }, { "epoch": 26.070143884892087, "grad_norm": 0.14771504700183868, "learning_rate": 4.099175198476838e-05, "loss": 0.0174, "step": 28990 }, { "action_loss": 0.0015340446261689067, "epoch": 26.070143884892087, "step": 28990 }, { "epoch": 26.070143884892087, "step": 28990, "torque_loss": 0.09162331372499466 }, { "epoch": 26.07913669064748, "grad_norm": 0.08395427465438843, "learning_rate": 4.0959225695086635e-05, "loss": 0.0179, "step": 29000 }, { "action_loss": 0.002866266993805766, "epoch": 26.07913669064748, "step": 29000 }, { "epoch": 26.07913669064748, "step": 29000, "torque_loss": 0.13674671947956085 }, { "epoch": 26.08812949640288, "grad_norm": 0.08217231184244156, "learning_rate": 4.0926703360146724e-05, "loss": 0.022, "step": 29010 }, { "action_loss": 0.002416537841781974, "epoch": 26.08812949640288, "step": 29010 }, { "epoch": 26.08812949640288, "step": 29010, "torque_loss": 0.16063399612903595 }, { "epoch": 26.097122302158272, "grad_norm": 0.13781143724918365, "learning_rate": 4.0894184994175e-05, "loss": 0.0202, "step": 29020 }, { "action_loss": 0.0047187223099172115, "epoch": 26.097122302158272, "step": 29020 }, { "epoch": 26.097122302158272, "step": 29020, "torque_loss": 0.11781498044729233 }, { "epoch": 26.10611510791367, "grad_norm": 0.12680257856845856, "learning_rate": 4.086167061139614e-05, "loss": 0.0187, "step": 29030 }, { "action_loss": 0.0015049698995426297, "epoch": 26.10611510791367, "step": 29030 }, { "epoch": 26.10611510791367, "step": 29030, "torque_loss": 0.1261035054922104 }, { "epoch": 26.115107913669064, "grad_norm": 0.13888460397720337, "learning_rate": 4.082916022603303e-05, "loss": 0.02, "step": 29040 }, { "action_loss": 0.00834045372903347, "epoch": 26.115107913669064, "step": 29040 }, { "epoch": 26.115107913669064, "step": 29040, "torque_loss": 0.16171269118785858 }, { "epoch": 26.12410071942446, "grad_norm": 0.14320512115955353, "learning_rate": 4.079665385230678e-05, "loss": 0.0217, "step": 29050 }, { "action_loss": 0.0024459578562527895, "epoch": 26.12410071942446, "step": 29050 }, { "epoch": 26.12410071942446, "step": 29050, "torque_loss": 0.15577422082424164 }, { "epoch": 26.133093525179856, "grad_norm": 0.11623814702033997, "learning_rate": 4.076415150443683e-05, "loss": 0.0199, "step": 29060 }, { "action_loss": 0.014290407299995422, "epoch": 26.133093525179856, "step": 29060 }, { "epoch": 26.133093525179856, "step": 29060, "torque_loss": 0.20742686092853546 }, { "epoch": 26.142086330935253, "grad_norm": 0.12829123437404633, "learning_rate": 4.073165319664079e-05, "loss": 0.0286, "step": 29070 }, { "action_loss": 0.002963111735880375, "epoch": 26.142086330935253, "step": 29070 }, { "epoch": 26.142086330935253, "step": 29070, "torque_loss": 0.14111335575580597 }, { "epoch": 26.151079136690647, "grad_norm": 0.1316927671432495, "learning_rate": 4.069915894313451e-05, "loss": 0.0214, "step": 29080 }, { "action_loss": 0.0046196202747523785, "epoch": 26.151079136690647, "step": 29080 }, { "epoch": 26.151079136690647, "step": 29080, "torque_loss": 0.21188221871852875 }, { "epoch": 26.16007194244604, "grad_norm": 0.13375438749790192, "learning_rate": 4.066666875813212e-05, "loss": 0.0185, "step": 29090 }, { "action_loss": 0.002686726162210107, "epoch": 26.16007194244604, "step": 29090 }, { "epoch": 26.16007194244604, "step": 29090, "torque_loss": 0.11054139584302902 }, { "epoch": 26.16906474820144, "grad_norm": 0.10907522588968277, "learning_rate": 4.063418265584589e-05, "loss": 0.0255, "step": 29100 }, { "action_loss": 0.0015314511256292462, "epoch": 26.16906474820144, "step": 29100 }, { "epoch": 26.16906474820144, "step": 29100, "torque_loss": 0.15379402041435242 }, { "epoch": 26.178057553956833, "grad_norm": 0.12138248980045319, "learning_rate": 4.060170065048636e-05, "loss": 0.0233, "step": 29110 }, { "action_loss": 0.0027119789738208055, "epoch": 26.178057553956833, "step": 29110 }, { "epoch": 26.178057553956833, "step": 29110, "torque_loss": 0.14489583671092987 }, { "epoch": 26.18705035971223, "grad_norm": 0.2056860327720642, "learning_rate": 4.056922275626227e-05, "loss": 0.0212, "step": 29120 }, { "action_loss": 0.005991655867546797, "epoch": 26.18705035971223, "step": 29120 }, { "epoch": 26.18705035971223, "step": 29120, "torque_loss": 0.13666976988315582 }, { "epoch": 26.196043165467625, "grad_norm": 0.10840976238250732, "learning_rate": 4.053674898738056e-05, "loss": 0.0239, "step": 29130 }, { "action_loss": 0.0068871378898620605, "epoch": 26.196043165467625, "step": 29130 }, { "epoch": 26.196043165467625, "step": 29130, "torque_loss": 0.1886642426252365 }, { "epoch": 26.205035971223023, "grad_norm": 0.2080901563167572, "learning_rate": 4.050427935804633e-05, "loss": 0.0209, "step": 29140 }, { "action_loss": 0.0022719749249517918, "epoch": 26.205035971223023, "step": 29140 }, { "epoch": 26.205035971223023, "step": 29140, "torque_loss": 0.14381802082061768 }, { "epoch": 26.214028776978417, "grad_norm": 0.13270334899425507, "learning_rate": 4.047181388246293e-05, "loss": 0.0259, "step": 29150 }, { "action_loss": 0.008023891597986221, "epoch": 26.214028776978417, "step": 29150 }, { "epoch": 26.214028776978417, "step": 29150, "torque_loss": 0.17024385929107666 }, { "epoch": 26.223021582733814, "grad_norm": 0.11345487833023071, "learning_rate": 4.043935257483185e-05, "loss": 0.0205, "step": 29160 }, { "action_loss": 0.0024701340589672327, "epoch": 26.223021582733814, "step": 29160 }, { "epoch": 26.223021582733814, "step": 29160, "torque_loss": 0.11511067301034927 }, { "epoch": 26.23201438848921, "grad_norm": 0.16037702560424805, "learning_rate": 4.0406895449352775e-05, "loss": 0.0213, "step": 29170 }, { "action_loss": 0.0033449570182710886, "epoch": 26.23201438848921, "step": 29170 }, { "epoch": 26.23201438848921, "step": 29170, "torque_loss": 0.13680826127529144 }, { "epoch": 26.241007194244606, "grad_norm": 0.08767462521791458, "learning_rate": 4.0374442520223525e-05, "loss": 0.0182, "step": 29180 }, { "action_loss": 0.004048902541399002, "epoch": 26.241007194244606, "step": 29180 }, { "epoch": 26.241007194244606, "step": 29180, "torque_loss": 0.12752637267112732 }, { "epoch": 26.25, "grad_norm": 0.09748853743076324, "learning_rate": 4.0341993801640155e-05, "loss": 0.0201, "step": 29190 }, { "action_loss": 0.0014096404192969203, "epoch": 26.25, "step": 29190 }, { "epoch": 26.25, "step": 29190, "torque_loss": 0.1082082986831665 }, { "epoch": 26.258992805755394, "grad_norm": 0.1529197096824646, "learning_rate": 4.030954930779681e-05, "loss": 0.0208, "step": 29200 }, { "action_loss": 0.014633804559707642, "epoch": 26.258992805755394, "step": 29200 }, { "epoch": 26.258992805755394, "step": 29200, "torque_loss": 0.15411341190338135 }, { "epoch": 26.26798561151079, "grad_norm": 0.11110325902700424, "learning_rate": 4.027710905288582e-05, "loss": 0.0208, "step": 29210 }, { "action_loss": 0.0028416209388524294, "epoch": 26.26798561151079, "step": 29210 }, { "epoch": 26.26798561151079, "step": 29210, "torque_loss": 0.1697186827659607 }, { "epoch": 26.276978417266186, "grad_norm": 0.12982390820980072, "learning_rate": 4.0244673051097666e-05, "loss": 0.0239, "step": 29220 }, { "action_loss": 0.006358036305755377, "epoch": 26.276978417266186, "step": 29220 }, { "epoch": 26.276978417266186, "step": 29220, "torque_loss": 0.14274068176746368 }, { "epoch": 26.285971223021583, "grad_norm": 0.11101784557104111, "learning_rate": 4.021224131662095e-05, "loss": 0.0204, "step": 29230 }, { "action_loss": 0.003676583059132099, "epoch": 26.285971223021583, "step": 29230 }, { "epoch": 26.285971223021583, "step": 29230, "torque_loss": 0.1956387758255005 }, { "epoch": 26.294964028776977, "grad_norm": 0.19710513949394226, "learning_rate": 4.0179813863642396e-05, "loss": 0.0213, "step": 29240 }, { "action_loss": 0.0034501480404287577, "epoch": 26.294964028776977, "step": 29240 }, { "epoch": 26.294964028776977, "step": 29240, "torque_loss": 0.18516544997692108 }, { "epoch": 26.303956834532375, "grad_norm": 0.16247282922267914, "learning_rate": 4.014739070634691e-05, "loss": 0.0281, "step": 29250 }, { "action_loss": 0.0050180028192698956, "epoch": 26.303956834532375, "step": 29250 }, { "epoch": 26.303956834532375, "step": 29250, "torque_loss": 0.16443650424480438 }, { "epoch": 26.31294964028777, "grad_norm": 0.1296796351671219, "learning_rate": 4.011497185891746e-05, "loss": 0.0235, "step": 29260 }, { "action_loss": 0.002068557543680072, "epoch": 26.31294964028777, "step": 29260 }, { "epoch": 26.31294964028777, "step": 29260, "torque_loss": 0.15981249511241913 }, { "epoch": 26.321942446043167, "grad_norm": 0.14297544956207275, "learning_rate": 4.008255733553513e-05, "loss": 0.023, "step": 29270 }, { "action_loss": 0.0019318587146699429, "epoch": 26.321942446043167, "step": 29270 }, { "epoch": 26.321942446043167, "step": 29270, "torque_loss": 0.1317761391401291 }, { "epoch": 26.33093525179856, "grad_norm": 0.15520355105400085, "learning_rate": 4.0050147150379194e-05, "loss": 0.0215, "step": 29280 }, { "action_loss": 0.0018034171080216765, "epoch": 26.33093525179856, "step": 29280 }, { "epoch": 26.33093525179856, "step": 29280, "torque_loss": 0.18010203540325165 }, { "epoch": 26.33992805755396, "grad_norm": 0.14644335210323334, "learning_rate": 4.001774131762692e-05, "loss": 0.0207, "step": 29290 }, { "action_loss": 0.002523066010326147, "epoch": 26.33992805755396, "step": 29290 }, { "epoch": 26.33992805755396, "step": 29290, "torque_loss": 0.13494601845741272 }, { "epoch": 26.348920863309353, "grad_norm": 0.11155841499567032, "learning_rate": 3.9985339851453705e-05, "loss": 0.0202, "step": 29300 }, { "action_loss": 0.0012708542635664344, "epoch": 26.348920863309353, "step": 29300 }, { "epoch": 26.348920863309353, "step": 29300, "torque_loss": 0.09489384293556213 }, { "epoch": 26.357913669064747, "grad_norm": 0.14213672280311584, "learning_rate": 3.9952942766033114e-05, "loss": 0.0209, "step": 29310 }, { "action_loss": 0.013935687951743603, "epoch": 26.357913669064747, "step": 29310 }, { "epoch": 26.357913669064747, "step": 29310, "torque_loss": 0.1786225289106369 }, { "epoch": 26.366906474820144, "grad_norm": 0.15532416105270386, "learning_rate": 3.9920550075536686e-05, "loss": 0.0207, "step": 29320 }, { "action_loss": 0.0015708663268014789, "epoch": 26.366906474820144, "step": 29320 }, { "epoch": 26.366906474820144, "step": 29320, "torque_loss": 0.1509076952934265 }, { "epoch": 26.37589928057554, "grad_norm": 0.10929926484823227, "learning_rate": 3.988816179413408e-05, "loss": 0.021, "step": 29330 }, { "action_loss": 0.0015877034747973084, "epoch": 26.37589928057554, "step": 29330 }, { "epoch": 26.37589928057554, "step": 29330, "torque_loss": 0.10712689906358719 }, { "epoch": 26.384892086330936, "grad_norm": 0.08386145532131195, "learning_rate": 3.985577793599307e-05, "loss": 0.0185, "step": 29340 }, { "action_loss": 0.009817731566727161, "epoch": 26.384892086330936, "step": 29340 }, { "epoch": 26.384892086330936, "step": 29340, "torque_loss": 0.18335890769958496 }, { "epoch": 26.39388489208633, "grad_norm": 0.10256163030862808, "learning_rate": 3.982339851527942e-05, "loss": 0.0227, "step": 29350 }, { "action_loss": 0.009002487175166607, "epoch": 26.39388489208633, "step": 29350 }, { "epoch": 26.39388489208633, "step": 29350, "torque_loss": 0.15804266929626465 }, { "epoch": 26.402877697841728, "grad_norm": 0.09664078801870346, "learning_rate": 3.979102354615701e-05, "loss": 0.0214, "step": 29360 }, { "action_loss": 0.0038939164951443672, "epoch": 26.402877697841728, "step": 29360 }, { "epoch": 26.402877697841728, "step": 29360, "torque_loss": 0.18424272537231445 }, { "epoch": 26.41187050359712, "grad_norm": 0.07921268790960312, "learning_rate": 3.975865304278772e-05, "loss": 0.0181, "step": 29370 }, { "action_loss": 0.0021675967145711184, "epoch": 26.41187050359712, "step": 29370 }, { "epoch": 26.41187050359712, "step": 29370, "torque_loss": 0.1284850388765335 }, { "epoch": 26.42086330935252, "grad_norm": 0.07034535706043243, "learning_rate": 3.9726287019331546e-05, "loss": 0.0214, "step": 29380 }, { "action_loss": 0.01370676327496767, "epoch": 26.42086330935252, "step": 29380 }, { "epoch": 26.42086330935252, "step": 29380, "torque_loss": 0.21433019638061523 }, { "epoch": 26.429856115107913, "grad_norm": 0.1334947943687439, "learning_rate": 3.9693925489946474e-05, "loss": 0.023, "step": 29390 }, { "action_loss": 0.0043615675531327724, "epoch": 26.429856115107913, "step": 29390 }, { "epoch": 26.429856115107913, "step": 29390, "torque_loss": 0.14497463405132294 }, { "epoch": 26.43884892086331, "grad_norm": 0.11402691900730133, "learning_rate": 3.966156846878852e-05, "loss": 0.0208, "step": 29400 }, { "action_loss": 0.004450885578989983, "epoch": 26.43884892086331, "step": 29400 }, { "epoch": 26.43884892086331, "step": 29400, "torque_loss": 0.18300761282444 }, { "epoch": 26.447841726618705, "grad_norm": 0.08880218118429184, "learning_rate": 3.962921597001177e-05, "loss": 0.0334, "step": 29410 }, { "action_loss": 0.0022582656238228083, "epoch": 26.447841726618705, "step": 29410 }, { "epoch": 26.447841726618705, "step": 29410, "torque_loss": 0.08253326267004013 }, { "epoch": 26.4568345323741, "grad_norm": 0.0893537774682045, "learning_rate": 3.9596868007768294e-05, "loss": 0.0216, "step": 29420 }, { "action_loss": 0.002326074056327343, "epoch": 26.4568345323741, "step": 29420 }, { "epoch": 26.4568345323741, "step": 29420, "torque_loss": 0.10567759722471237 }, { "epoch": 26.465827338129497, "grad_norm": 0.15856477618217468, "learning_rate": 3.9564524596208165e-05, "loss": 0.0222, "step": 29430 }, { "action_loss": 0.017577754333615303, "epoch": 26.465827338129497, "step": 29430 }, { "epoch": 26.465827338129497, "step": 29430, "torque_loss": 0.23068690299987793 }, { "epoch": 26.47482014388489, "grad_norm": 0.17027349770069122, "learning_rate": 3.953218574947954e-05, "loss": 0.0259, "step": 29440 }, { "action_loss": 0.001563845667988062, "epoch": 26.47482014388489, "step": 29440 }, { "epoch": 26.47482014388489, "step": 29440, "torque_loss": 0.17453134059906006 }, { "epoch": 26.48381294964029, "grad_norm": 0.08813691139221191, "learning_rate": 3.949985148172849e-05, "loss": 0.0193, "step": 29450 }, { "action_loss": 0.005141554865986109, "epoch": 26.48381294964029, "step": 29450 }, { "epoch": 26.48381294964029, "step": 29450, "torque_loss": 0.13639968633651733 }, { "epoch": 26.492805755395683, "grad_norm": 0.16780446469783783, "learning_rate": 3.946752180709914e-05, "loss": 0.0251, "step": 29460 }, { "action_loss": 0.003245672909542918, "epoch": 26.492805755395683, "step": 29460 }, { "epoch": 26.492805755395683, "step": 29460, "torque_loss": 0.1413494348526001 }, { "epoch": 26.50179856115108, "grad_norm": 0.14374876022338867, "learning_rate": 3.943519673973357e-05, "loss": 0.0249, "step": 29470 }, { "action_loss": 0.006208312232047319, "epoch": 26.50179856115108, "step": 29470 }, { "epoch": 26.50179856115108, "step": 29470, "torque_loss": 0.16210058331489563 }, { "epoch": 26.510791366906474, "grad_norm": 0.15464477241039276, "learning_rate": 3.940287629377189e-05, "loss": 0.0271, "step": 29480 }, { "action_loss": 0.00249760621227324, "epoch": 26.510791366906474, "step": 29480 }, { "epoch": 26.510791366906474, "step": 29480, "torque_loss": 0.1291167289018631 }, { "epoch": 26.519784172661872, "grad_norm": 0.12932471930980682, "learning_rate": 3.937056048335214e-05, "loss": 0.0193, "step": 29490 }, { "action_loss": 0.0026702519971877337, "epoch": 26.519784172661872, "step": 29490 }, { "epoch": 26.519784172661872, "step": 29490, "torque_loss": 0.08566858619451523 }, { "epoch": 26.528776978417266, "grad_norm": 0.14514368772506714, "learning_rate": 3.933824932261037e-05, "loss": 0.0257, "step": 29500 }, { "action_loss": 0.0021793250925838947, "epoch": 26.528776978417266, "step": 29500 }, { "epoch": 26.528776978417266, "step": 29500, "torque_loss": 0.10143949836492538 }, { "epoch": 26.53776978417266, "grad_norm": 0.15660449862480164, "learning_rate": 3.9305942825680576e-05, "loss": 0.022, "step": 29510 }, { "action_loss": 0.003972330596297979, "epoch": 26.53776978417266, "step": 29510 }, { "epoch": 26.53776978417266, "step": 29510, "torque_loss": 0.1863747090101242 }, { "epoch": 26.546762589928058, "grad_norm": 0.08154560625553131, "learning_rate": 3.9273641006694706e-05, "loss": 0.0223, "step": 29520 }, { "action_loss": 0.015473082661628723, "epoch": 26.546762589928058, "step": 29520 }, { "epoch": 26.546762589928058, "step": 29520, "torque_loss": 0.21365995705127716 }, { "epoch": 26.555755395683452, "grad_norm": 0.14202281832695007, "learning_rate": 3.924134387978266e-05, "loss": 0.0258, "step": 29530 }, { "action_loss": 0.010766989551484585, "epoch": 26.555755395683452, "step": 29530 }, { "epoch": 26.555755395683452, "step": 29530, "torque_loss": 0.22599463164806366 }, { "epoch": 26.56474820143885, "grad_norm": 0.15240263938903809, "learning_rate": 3.920905145907234e-05, "loss": 0.0264, "step": 29540 }, { "action_loss": 0.0028259754180908203, "epoch": 26.56474820143885, "step": 29540 }, { "epoch": 26.56474820143885, "step": 29540, "torque_loss": 0.13665802776813507 }, { "epoch": 26.573741007194243, "grad_norm": 0.11786355078220367, "learning_rate": 3.9176763758689525e-05, "loss": 0.0259, "step": 29550 }, { "action_loss": 0.0021181851625442505, "epoch": 26.573741007194243, "step": 29550 }, { "epoch": 26.573741007194243, "step": 29550, "torque_loss": 0.08919768780469894 }, { "epoch": 26.58273381294964, "grad_norm": 0.15337027609348297, "learning_rate": 3.9144480792757925e-05, "loss": 0.0206, "step": 29560 }, { "action_loss": 0.004834370221942663, "epoch": 26.58273381294964, "step": 29560 }, { "epoch": 26.58273381294964, "step": 29560, "torque_loss": 0.20410199463367462 }, { "epoch": 26.591726618705035, "grad_norm": 0.1048530712723732, "learning_rate": 3.911220257539926e-05, "loss": 0.0238, "step": 29570 }, { "action_loss": 0.0015198095934465528, "epoch": 26.591726618705035, "step": 29570 }, { "epoch": 26.591726618705035, "step": 29570, "torque_loss": 0.1116572842001915 }, { "epoch": 26.600719424460433, "grad_norm": 0.09145135432481766, "learning_rate": 3.907992912073308e-05, "loss": 0.0218, "step": 29580 }, { "action_loss": 0.00512356823310256, "epoch": 26.600719424460433, "step": 29580 }, { "epoch": 26.600719424460433, "step": 29580, "torque_loss": 0.2456936240196228 }, { "epoch": 26.609712230215827, "grad_norm": 0.13316620886325836, "learning_rate": 3.90476604428769e-05, "loss": 0.0201, "step": 29590 }, { "action_loss": 0.0021036292891949415, "epoch": 26.609712230215827, "step": 29590 }, { "epoch": 26.609712230215827, "step": 29590, "torque_loss": 0.16230474412441254 }, { "epoch": 26.618705035971225, "grad_norm": 0.071772500872612, "learning_rate": 3.901539655594614e-05, "loss": 0.0158, "step": 29600 }, { "action_loss": 0.0039002876728773117, "epoch": 26.618705035971225, "step": 29600 }, { "epoch": 26.618705035971225, "step": 29600, "torque_loss": 0.1847655326128006 }, { "epoch": 26.62769784172662, "grad_norm": 0.11372648179531097, "learning_rate": 3.898313747405413e-05, "loss": 0.02, "step": 29610 }, { "action_loss": 0.0018602170748636127, "epoch": 26.62769784172662, "step": 29610 }, { "epoch": 26.62769784172662, "step": 29610, "torque_loss": 0.15225671231746674 }, { "epoch": 26.636690647482013, "grad_norm": 0.0894283801317215, "learning_rate": 3.895088321131205e-05, "loss": 0.0212, "step": 29620 }, { "action_loss": 0.003570633940398693, "epoch": 26.636690647482013, "step": 29620 }, { "epoch": 26.636690647482013, "step": 29620, "torque_loss": 0.18005287647247314 }, { "epoch": 26.64568345323741, "grad_norm": 0.10229866206645966, "learning_rate": 3.891863378182907e-05, "loss": 0.0226, "step": 29630 }, { "action_loss": 0.005159385967999697, "epoch": 26.64568345323741, "step": 29630 }, { "epoch": 26.64568345323741, "step": 29630, "torque_loss": 0.14074036478996277 }, { "epoch": 26.654676258992804, "grad_norm": 0.13335265219211578, "learning_rate": 3.8886389199712145e-05, "loss": 0.0197, "step": 29640 }, { "action_loss": 0.015562723390758038, "epoch": 26.654676258992804, "step": 29640 }, { "epoch": 26.654676258992804, "step": 29640, "torque_loss": 0.25224098563194275 }, { "epoch": 26.663669064748202, "grad_norm": 0.12274406105279922, "learning_rate": 3.885414947906614e-05, "loss": 0.0207, "step": 29650 }, { "action_loss": 0.013524177484214306, "epoch": 26.663669064748202, "step": 29650 }, { "epoch": 26.663669064748202, "step": 29650, "torque_loss": 0.20087963342666626 }, { "epoch": 26.672661870503596, "grad_norm": 0.11502240598201752, "learning_rate": 3.882191463399385e-05, "loss": 0.0259, "step": 29660 }, { "action_loss": 0.004258813802152872, "epoch": 26.672661870503596, "step": 29660 }, { "epoch": 26.672661870503596, "step": 29660, "torque_loss": 0.12587721645832062 }, { "epoch": 26.681654676258994, "grad_norm": 0.10791315883398056, "learning_rate": 3.878968467859585e-05, "loss": 0.0207, "step": 29670 }, { "action_loss": 0.003753396449610591, "epoch": 26.681654676258994, "step": 29670 }, { "epoch": 26.681654676258994, "step": 29670, "torque_loss": 0.21767081320285797 }, { "epoch": 26.690647482014388, "grad_norm": 0.10078836977481842, "learning_rate": 3.875745962697063e-05, "loss": 0.0225, "step": 29680 }, { "action_loss": 0.0024113378021866083, "epoch": 26.690647482014388, "step": 29680 }, { "epoch": 26.690647482014388, "step": 29680, "torque_loss": 0.10961667448282242 }, { "epoch": 26.699640287769785, "grad_norm": 0.09588268399238586, "learning_rate": 3.872523949321454e-05, "loss": 0.0205, "step": 29690 }, { "action_loss": 0.003374338848516345, "epoch": 26.699640287769785, "step": 29690 }, { "epoch": 26.699640287769785, "step": 29690, "torque_loss": 0.1575063019990921 }, { "epoch": 26.70863309352518, "grad_norm": 0.09865562617778778, "learning_rate": 3.869302429142173e-05, "loss": 0.0197, "step": 29700 }, { "action_loss": 0.0031361058354377747, "epoch": 26.70863309352518, "step": 29700 }, { "epoch": 26.70863309352518, "step": 29700, "torque_loss": 0.15402764081954956 }, { "epoch": 26.717625899280577, "grad_norm": 0.08347349613904953, "learning_rate": 3.866081403568425e-05, "loss": 0.0199, "step": 29710 }, { "action_loss": 0.0039377412758767605, "epoch": 26.717625899280577, "step": 29710 }, { "epoch": 26.717625899280577, "step": 29710, "torque_loss": 0.18666569888591766 }, { "epoch": 26.72661870503597, "grad_norm": 0.10133419930934906, "learning_rate": 3.8628608740091934e-05, "loss": 0.0222, "step": 29720 }, { "action_loss": 0.0064175184816122055, "epoch": 26.72661870503597, "step": 29720 }, { "epoch": 26.72661870503597, "step": 29720, "torque_loss": 0.15784616768360138 }, { "epoch": 26.735611510791365, "grad_norm": 0.08610817044973373, "learning_rate": 3.85964084187325e-05, "loss": 0.0211, "step": 29730 }, { "action_loss": 0.00886071752756834, "epoch": 26.735611510791365, "step": 29730 }, { "epoch": 26.735611510791365, "step": 29730, "torque_loss": 0.18900257349014282 }, { "epoch": 26.744604316546763, "grad_norm": 0.15631181001663208, "learning_rate": 3.856421308569146e-05, "loss": 0.0216, "step": 29740 }, { "action_loss": 0.002163437195122242, "epoch": 26.744604316546763, "step": 29740 }, { "epoch": 26.744604316546763, "step": 29740, "torque_loss": 0.17830891907215118 }, { "epoch": 26.753597122302157, "grad_norm": 0.13163891434669495, "learning_rate": 3.853202275505212e-05, "loss": 0.022, "step": 29750 }, { "action_loss": 0.0033812858164310455, "epoch": 26.753597122302157, "step": 29750 }, { "epoch": 26.753597122302157, "step": 29750, "torque_loss": 0.1356896609067917 }, { "epoch": 26.762589928057555, "grad_norm": 0.14675822854042053, "learning_rate": 3.849983744089565e-05, "loss": 0.0209, "step": 29760 }, { "action_loss": 0.0014690393581986427, "epoch": 26.762589928057555, "step": 29760 }, { "epoch": 26.762589928057555, "step": 29760, "torque_loss": 0.1377881020307541 }, { "epoch": 26.77158273381295, "grad_norm": 0.13707120716571808, "learning_rate": 3.8467657157301e-05, "loss": 0.0191, "step": 29770 }, { "action_loss": 0.003671988844871521, "epoch": 26.77158273381295, "step": 29770 }, { "epoch": 26.77158273381295, "step": 29770, "torque_loss": 0.22373239696025848 }, { "epoch": 26.780575539568346, "grad_norm": 0.08144668489694595, "learning_rate": 3.8435481918344884e-05, "loss": 0.021, "step": 29780 }, { "action_loss": 0.006177651230245829, "epoch": 26.780575539568346, "step": 29780 }, { "epoch": 26.780575539568346, "step": 29780, "torque_loss": 0.19329531490802765 }, { "epoch": 26.78956834532374, "grad_norm": 0.1699010282754898, "learning_rate": 3.84033117381019e-05, "loss": 0.0227, "step": 29790 }, { "action_loss": 0.0010709055932238698, "epoch": 26.78956834532374, "step": 29790 }, { "epoch": 26.78956834532374, "step": 29790, "torque_loss": 0.13098450005054474 }, { "epoch": 26.798561151079138, "grad_norm": 0.06336239725351334, "learning_rate": 3.8371146630644335e-05, "loss": 0.0167, "step": 29800 }, { "action_loss": 0.0028517795726656914, "epoch": 26.798561151079138, "step": 29800 }, { "epoch": 26.798561151079138, "step": 29800, "torque_loss": 0.20193322002887726 }, { "epoch": 26.807553956834532, "grad_norm": 0.10949410498142242, "learning_rate": 3.833898661004231e-05, "loss": 0.0198, "step": 29810 }, { "action_loss": 0.0028849272057414055, "epoch": 26.807553956834532, "step": 29810 }, { "epoch": 26.807553956834532, "step": 29810, "torque_loss": 0.11933223158121109 }, { "epoch": 26.81654676258993, "grad_norm": 0.09295832365751266, "learning_rate": 3.830683169036372e-05, "loss": 0.0195, "step": 29820 }, { "action_loss": 0.0013932079309597611, "epoch": 26.81654676258993, "step": 29820 }, { "epoch": 26.81654676258993, "step": 29820, "torque_loss": 0.11626984924077988 }, { "epoch": 26.825539568345324, "grad_norm": 0.13079966604709625, "learning_rate": 3.827468188567422e-05, "loss": 0.0261, "step": 29830 }, { "action_loss": 0.04076644778251648, "epoch": 26.825539568345324, "step": 29830 }, { "epoch": 26.825539568345324, "step": 29830, "torque_loss": 0.2220831662416458 }, { "epoch": 26.834532374100718, "grad_norm": 0.18308484554290771, "learning_rate": 3.82425372100372e-05, "loss": 0.0289, "step": 29840 }, { "action_loss": 0.002392185851931572, "epoch": 26.834532374100718, "step": 29840 }, { "epoch": 26.834532374100718, "step": 29840, "torque_loss": 0.1109570637345314 }, { "epoch": 26.843525179856115, "grad_norm": 0.12420032173395157, "learning_rate": 3.821039767751387e-05, "loss": 0.02, "step": 29850 }, { "action_loss": 0.002721346216276288, "epoch": 26.843525179856115, "step": 29850 }, { "epoch": 26.843525179856115, "step": 29850, "torque_loss": 0.15678824484348297 }, { "epoch": 26.85251798561151, "grad_norm": 0.06861255317926407, "learning_rate": 3.817826330216315e-05, "loss": 0.0199, "step": 29860 }, { "action_loss": 0.002428071340546012, "epoch": 26.85251798561151, "step": 29860 }, { "epoch": 26.85251798561151, "step": 29860, "torque_loss": 0.12276742607355118 }, { "epoch": 26.861510791366907, "grad_norm": 0.1554756909608841, "learning_rate": 3.814613409804169e-05, "loss": 0.023, "step": 29870 }, { "action_loss": 0.005038144998252392, "epoch": 26.861510791366907, "step": 29870 }, { "epoch": 26.861510791366907, "step": 29870, "torque_loss": 0.1866605281829834 }, { "epoch": 26.8705035971223, "grad_norm": 0.14501215517520905, "learning_rate": 3.811401007920388e-05, "loss": 0.0209, "step": 29880 }, { "action_loss": 0.020537598058581352, "epoch": 26.8705035971223, "step": 29880 }, { "epoch": 26.8705035971223, "step": 29880, "torque_loss": 0.22492855787277222 }, { "epoch": 26.8794964028777, "grad_norm": 0.11159006506204605, "learning_rate": 3.8081891259701895e-05, "loss": 0.0232, "step": 29890 }, { "action_loss": 0.0026637904811650515, "epoch": 26.8794964028777, "step": 29890 }, { "epoch": 26.8794964028777, "step": 29890, "torque_loss": 0.17815802991390228 }, { "epoch": 26.888489208633093, "grad_norm": 0.12173932045698166, "learning_rate": 3.8049777653585584e-05, "loss": 0.0197, "step": 29900 }, { "action_loss": 0.000946228567045182, "epoch": 26.888489208633093, "step": 29900 }, { "epoch": 26.888489208633093, "step": 29900, "torque_loss": 0.12001687288284302 }, { "epoch": 26.89748201438849, "grad_norm": 0.11624958366155624, "learning_rate": 3.80176692749025e-05, "loss": 0.0185, "step": 29910 }, { "action_loss": 0.002948896260932088, "epoch": 26.89748201438849, "step": 29910 }, { "epoch": 26.89748201438849, "step": 29910, "torque_loss": 0.18622331321239471 }, { "epoch": 26.906474820143885, "grad_norm": 0.11667143553495407, "learning_rate": 3.798556613769799e-05, "loss": 0.0183, "step": 29920 }, { "action_loss": 0.002895432524383068, "epoch": 26.906474820143885, "step": 29920 }, { "epoch": 26.906474820143885, "step": 29920, "torque_loss": 0.12632466852664948 }, { "epoch": 26.915467625899282, "grad_norm": 0.1278277337551117, "learning_rate": 3.795346825601502e-05, "loss": 0.0192, "step": 29930 }, { "action_loss": 0.003628704696893692, "epoch": 26.915467625899282, "step": 29930 }, { "epoch": 26.915467625899282, "step": 29930, "torque_loss": 0.1762712001800537 }, { "epoch": 26.924460431654676, "grad_norm": 0.08643265813589096, "learning_rate": 3.7921375643894305e-05, "loss": 0.0178, "step": 29940 }, { "action_loss": 0.013024684973061085, "epoch": 26.924460431654676, "step": 29940 }, { "epoch": 26.924460431654676, "step": 29940, "torque_loss": 0.20929430425167084 }, { "epoch": 26.93345323741007, "grad_norm": 0.07497149705886841, "learning_rate": 3.788928831537424e-05, "loss": 0.0214, "step": 29950 }, { "action_loss": 0.0018611421110108495, "epoch": 26.93345323741007, "step": 29950 }, { "epoch": 26.93345323741007, "step": 29950, "torque_loss": 0.15519475936889648 }, { "epoch": 26.942446043165468, "grad_norm": 0.10070818662643433, "learning_rate": 3.7857206284490934e-05, "loss": 0.0175, "step": 29960 }, { "action_loss": 0.0059706843458116055, "epoch": 26.942446043165468, "step": 29960 }, { "epoch": 26.942446043165468, "step": 29960, "torque_loss": 0.17020170390605927 }, { "epoch": 26.951438848920862, "grad_norm": 0.11001425236463547, "learning_rate": 3.7825129565278114e-05, "loss": 0.0235, "step": 29970 }, { "action_loss": 0.01993541046977043, "epoch": 26.951438848920862, "step": 29970 }, { "epoch": 26.951438848920862, "step": 29970, "torque_loss": 0.1707025170326233 }, { "epoch": 26.96043165467626, "grad_norm": 0.15760090947151184, "learning_rate": 3.779305817176728e-05, "loss": 0.0264, "step": 29980 }, { "action_loss": 0.004085892345756292, "epoch": 26.96043165467626, "step": 29980 }, { "epoch": 26.96043165467626, "step": 29980, "torque_loss": 0.11197134107351303 }, { "epoch": 26.969424460431654, "grad_norm": 0.1385548859834671, "learning_rate": 3.776099211798751e-05, "loss": 0.0186, "step": 29990 }, { "action_loss": 0.0015754193300381303, "epoch": 26.969424460431654, "step": 29990 }, { "epoch": 26.969424460431654, "step": 29990, "torque_loss": 0.10547974705696106 }, { "epoch": 26.97841726618705, "grad_norm": 0.09917159378528595, "learning_rate": 3.7728931417965584e-05, "loss": 0.0229, "step": 30000 }, { "action_loss": 0.00414919713512063, "epoch": 26.97841726618705, "step": 30000 }, { "epoch": 26.97841726618705, "step": 30000, "torque_loss": 0.2277369499206543 }, { "epoch": 26.987410071942445, "grad_norm": 0.1312890648841858, "learning_rate": 3.769687608572598e-05, "loss": 0.021, "step": 30010 }, { "action_loss": 0.002335860626772046, "epoch": 26.987410071942445, "step": 30010 }, { "epoch": 26.987410071942445, "step": 30010, "torque_loss": 0.1274641901254654 }, { "epoch": 26.996402877697843, "grad_norm": 0.1641484647989273, "learning_rate": 3.766482613529076e-05, "loss": 0.0209, "step": 30020 }, { "action_loss": 0.005232540424913168, "epoch": 26.996402877697843, "step": 30020 }, { "epoch": 26.996402877697843, "step": 30020, "torque_loss": 0.16724956035614014 }, { "epoch": 27.005395683453237, "grad_norm": 0.1709304004907608, "learning_rate": 3.763278158067964e-05, "loss": 0.0203, "step": 30030 }, { "action_loss": 0.004101251717656851, "epoch": 27.005395683453237, "step": 30030 }, { "epoch": 27.005395683453237, "step": 30030, "torque_loss": 0.1799580454826355 }, { "epoch": 27.014388489208635, "grad_norm": 0.13012553751468658, "learning_rate": 3.7600742435910045e-05, "loss": 0.0251, "step": 30040 }, { "action_loss": 0.001996435457840562, "epoch": 27.014388489208635, "step": 30040 }, { "epoch": 27.014388489208635, "step": 30040, "torque_loss": 0.12711051106452942 }, { "epoch": 27.02338129496403, "grad_norm": 0.10580137372016907, "learning_rate": 3.756870871499696e-05, "loss": 0.0232, "step": 30050 }, { "action_loss": 0.005930574610829353, "epoch": 27.02338129496403, "step": 30050 }, { "epoch": 27.02338129496403, "step": 30050, "torque_loss": 0.25652849674224854 }, { "epoch": 27.032374100719423, "grad_norm": 0.14800575375556946, "learning_rate": 3.7536680431953025e-05, "loss": 0.0227, "step": 30060 }, { "action_loss": 0.001735243946313858, "epoch": 27.032374100719423, "step": 30060 }, { "epoch": 27.032374100719423, "step": 30060, "torque_loss": 0.14548881351947784 }, { "epoch": 27.04136690647482, "grad_norm": 0.10228632390499115, "learning_rate": 3.7504657600788485e-05, "loss": 0.0217, "step": 30070 }, { "action_loss": 0.0018825592705979943, "epoch": 27.04136690647482, "step": 30070 }, { "epoch": 27.04136690647482, "step": 30070, "torque_loss": 0.15734855830669403 }, { "epoch": 27.050359712230215, "grad_norm": 0.16703476011753082, "learning_rate": 3.747264023551125e-05, "loss": 0.0232, "step": 30080 }, { "action_loss": 0.006909611169248819, "epoch": 27.050359712230215, "step": 30080 }, { "epoch": 27.050359712230215, "step": 30080, "torque_loss": 0.12670685350894928 }, { "epoch": 27.059352517985612, "grad_norm": 0.1721789687871933, "learning_rate": 3.744062835012678e-05, "loss": 0.0231, "step": 30090 }, { "action_loss": 0.004833681043237448, "epoch": 27.059352517985612, "step": 30090 }, { "epoch": 27.059352517985612, "step": 30090, "torque_loss": 0.1914224475622177 }, { "epoch": 27.068345323741006, "grad_norm": 0.06798239052295685, "learning_rate": 3.740862195863814e-05, "loss": 0.0201, "step": 30100 }, { "action_loss": 0.004093765746802092, "epoch": 27.068345323741006, "step": 30100 }, { "epoch": 27.068345323741006, "step": 30100, "torque_loss": 0.20746450126171112 }, { "epoch": 27.077338129496404, "grad_norm": 0.0913657546043396, "learning_rate": 3.737662107504606e-05, "loss": 0.0187, "step": 30110 }, { "action_loss": 0.0018942983588203788, "epoch": 27.077338129496404, "step": 30110 }, { "epoch": 27.077338129496404, "step": 30110, "torque_loss": 0.15788353979587555 }, { "epoch": 27.086330935251798, "grad_norm": 0.15912683308124542, "learning_rate": 3.734462571334879e-05, "loss": 0.0217, "step": 30120 }, { "action_loss": 0.0027012682985514402, "epoch": 27.086330935251798, "step": 30120 }, { "epoch": 27.086330935251798, "step": 30120, "torque_loss": 0.142137810587883 }, { "epoch": 27.095323741007196, "grad_norm": 0.09154969453811646, "learning_rate": 3.731263588754217e-05, "loss": 0.022, "step": 30130 }, { "action_loss": 0.006669348105788231, "epoch": 27.095323741007196, "step": 30130 }, { "epoch": 27.095323741007196, "step": 30130, "torque_loss": 0.16398482024669647 }, { "epoch": 27.10431654676259, "grad_norm": 0.14638619124889374, "learning_rate": 3.7280651611619665e-05, "loss": 0.0216, "step": 30140 }, { "action_loss": 0.0044488427229225636, "epoch": 27.10431654676259, "step": 30140 }, { "epoch": 27.10431654676259, "step": 30140, "torque_loss": 0.17754685878753662 }, { "epoch": 27.113309352517987, "grad_norm": 0.11123962700366974, "learning_rate": 3.724867289957229e-05, "loss": 0.0201, "step": 30150 }, { "action_loss": 0.0037718017119914293, "epoch": 27.113309352517987, "step": 30150 }, { "epoch": 27.113309352517987, "step": 30150, "torque_loss": 0.1783781498670578 }, { "epoch": 27.12230215827338, "grad_norm": 0.12605679035186768, "learning_rate": 3.7216699765388576e-05, "loss": 0.0186, "step": 30160 }, { "action_loss": 0.004741601180285215, "epoch": 27.12230215827338, "step": 30160 }, { "epoch": 27.12230215827338, "step": 30160, "torque_loss": 0.13735617697238922 }, { "epoch": 27.131294964028775, "grad_norm": 0.07766731083393097, "learning_rate": 3.71847322230547e-05, "loss": 0.0209, "step": 30170 }, { "action_loss": 0.005227468907833099, "epoch": 27.131294964028775, "step": 30170 }, { "epoch": 27.131294964028775, "step": 30170, "torque_loss": 0.24344803392887115 }, { "epoch": 27.140287769784173, "grad_norm": 0.15271060168743134, "learning_rate": 3.7152770286554335e-05, "loss": 0.0209, "step": 30180 }, { "action_loss": 0.001768327783793211, "epoch": 27.140287769784173, "step": 30180 }, { "epoch": 27.140287769784173, "step": 30180, "torque_loss": 0.12722830474376678 }, { "epoch": 27.149280575539567, "grad_norm": 0.1323515623807907, "learning_rate": 3.712081396986872e-05, "loss": 0.02, "step": 30190 }, { "action_loss": 0.0050520021468400955, "epoch": 27.149280575539567, "step": 30190 }, { "epoch": 27.149280575539567, "step": 30190, "torque_loss": 0.15034973621368408 }, { "epoch": 27.158273381294965, "grad_norm": 0.08908218145370483, "learning_rate": 3.708886328697663e-05, "loss": 0.0195, "step": 30200 }, { "action_loss": 0.013033236376941204, "epoch": 27.158273381294965, "step": 30200 }, { "epoch": 27.158273381294965, "step": 30200, "torque_loss": 0.1949307918548584 }, { "epoch": 27.16726618705036, "grad_norm": 0.0759287104010582, "learning_rate": 3.7056918251854395e-05, "loss": 0.0185, "step": 30210 }, { "action_loss": 0.030347460880875587, "epoch": 27.16726618705036, "step": 30210 }, { "epoch": 27.16726618705036, "step": 30210, "torque_loss": 0.28099682927131653 }, { "epoch": 27.176258992805757, "grad_norm": 0.09923818707466125, "learning_rate": 3.7024978878475846e-05, "loss": 0.0288, "step": 30220 }, { "action_loss": 0.035565588623285294, "epoch": 27.176258992805757, "step": 30220 }, { "epoch": 27.176258992805757, "step": 30220, "torque_loss": 0.2999943494796753 }, { "epoch": 27.18525179856115, "grad_norm": 0.12378242611885071, "learning_rate": 3.699304518081233e-05, "loss": 0.0285, "step": 30230 }, { "action_loss": 0.006802581716328859, "epoch": 27.18525179856115, "step": 30230 }, { "epoch": 27.18525179856115, "step": 30230, "torque_loss": 0.28226438164711 }, { "epoch": 27.194244604316548, "grad_norm": 0.11680993437767029, "learning_rate": 3.696111717283277e-05, "loss": 0.0205, "step": 30240 }, { "action_loss": 0.0049086278304457664, "epoch": 27.194244604316548, "step": 30240 }, { "epoch": 27.194244604316548, "step": 30240, "torque_loss": 0.1417580544948578 }, { "epoch": 27.203237410071942, "grad_norm": 0.08121365308761597, "learning_rate": 3.692919486850355e-05, "loss": 0.0224, "step": 30250 }, { "action_loss": 0.003642237512394786, "epoch": 27.203237410071942, "step": 30250 }, { "epoch": 27.203237410071942, "step": 30250, "torque_loss": 0.17425024509429932 }, { "epoch": 27.21223021582734, "grad_norm": 0.12918336689472198, "learning_rate": 3.689727828178854e-05, "loss": 0.0215, "step": 30260 }, { "action_loss": 0.001822876394726336, "epoch": 27.21223021582734, "step": 30260 }, { "epoch": 27.21223021582734, "step": 30260, "torque_loss": 0.17137600481510162 }, { "epoch": 27.221223021582734, "grad_norm": 0.138112872838974, "learning_rate": 3.6865367426649185e-05, "loss": 0.0201, "step": 30270 }, { "action_loss": 0.016114600002765656, "epoch": 27.221223021582734, "step": 30270 }, { "epoch": 27.221223021582734, "step": 30270, "torque_loss": 0.14598239958286285 }, { "epoch": 27.230215827338128, "grad_norm": 0.06393221020698547, "learning_rate": 3.6833462317044346e-05, "loss": 0.0194, "step": 30280 }, { "action_loss": 0.0014722524210810661, "epoch": 27.230215827338128, "step": 30280 }, { "epoch": 27.230215827338128, "step": 30280, "torque_loss": 0.11226323992013931 }, { "epoch": 27.239208633093526, "grad_norm": 0.07667608559131622, "learning_rate": 3.6801562966930406e-05, "loss": 0.0213, "step": 30290 }, { "action_loss": 0.001149900839664042, "epoch": 27.239208633093526, "step": 30290 }, { "epoch": 27.239208633093526, "step": 30290, "torque_loss": 0.11722948402166367 }, { "epoch": 27.24820143884892, "grad_norm": 0.0813024714589119, "learning_rate": 3.676966939026123e-05, "loss": 0.0174, "step": 30300 }, { "action_loss": 0.0027656902093440294, "epoch": 27.24820143884892, "step": 30300 }, { "epoch": 27.24820143884892, "step": 30300, "torque_loss": 0.11862236261367798 }, { "epoch": 27.257194244604317, "grad_norm": 0.0952511727809906, "learning_rate": 3.6737781600988166e-05, "loss": 0.0195, "step": 30310 }, { "action_loss": 0.002607795177027583, "epoch": 27.257194244604317, "step": 30310 }, { "epoch": 27.257194244604317, "step": 30310, "torque_loss": 0.17439652979373932 }, { "epoch": 27.26618705035971, "grad_norm": 0.09308844059705734, "learning_rate": 3.670589961305998e-05, "loss": 0.0203, "step": 30320 }, { "action_loss": 0.004616000223904848, "epoch": 27.26618705035971, "step": 30320 }, { "epoch": 27.26618705035971, "step": 30320, "torque_loss": 0.18243081867694855 }, { "epoch": 27.27517985611511, "grad_norm": 0.07460649311542511, "learning_rate": 3.667402344042298e-05, "loss": 0.0172, "step": 30330 }, { "action_loss": 0.00870160385966301, "epoch": 27.27517985611511, "step": 30330 }, { "epoch": 27.27517985611511, "step": 30330, "torque_loss": 0.1823231726884842 }, { "epoch": 27.284172661870503, "grad_norm": 0.11573625355958939, "learning_rate": 3.664215309702086e-05, "loss": 0.025, "step": 30340 }, { "action_loss": 0.0029515577480196953, "epoch": 27.284172661870503, "step": 30340 }, { "epoch": 27.284172661870503, "step": 30340, "torque_loss": 0.18525294959545135 }, { "epoch": 27.2931654676259, "grad_norm": 0.11612632125616074, "learning_rate": 3.661028859679478e-05, "loss": 0.0186, "step": 30350 }, { "action_loss": 0.0048974305391311646, "epoch": 27.2931654676259, "step": 30350 }, { "epoch": 27.2931654676259, "step": 30350, "torque_loss": 0.1998310089111328 }, { "epoch": 27.302158273381295, "grad_norm": 0.15211233496665955, "learning_rate": 3.657842995368339e-05, "loss": 0.021, "step": 30360 }, { "action_loss": 0.004024719353765249, "epoch": 27.302158273381295, "step": 30360 }, { "epoch": 27.302158273381295, "step": 30360, "torque_loss": 0.17019283771514893 }, { "epoch": 27.31115107913669, "grad_norm": 0.11175265163183212, "learning_rate": 3.654657718162273e-05, "loss": 0.0184, "step": 30370 }, { "action_loss": 0.0022862886544317007, "epoch": 27.31115107913669, "step": 30370 }, { "epoch": 27.31115107913669, "step": 30370, "torque_loss": 0.14149974286556244 }, { "epoch": 27.320143884892087, "grad_norm": 0.1257227510213852, "learning_rate": 3.6514730294546275e-05, "loss": 0.0222, "step": 30380 }, { "action_loss": 0.001236400450579822, "epoch": 27.320143884892087, "step": 30380 }, { "epoch": 27.320143884892087, "step": 30380, "torque_loss": 0.15501759946346283 }, { "epoch": 27.32913669064748, "grad_norm": 0.09603379666805267, "learning_rate": 3.648288930638492e-05, "loss": 0.0244, "step": 30390 }, { "action_loss": 0.014597046189010143, "epoch": 27.32913669064748, "step": 30390 }, { "epoch": 27.32913669064748, "step": 30390, "torque_loss": 0.28388291597366333 }, { "epoch": 27.33812949640288, "grad_norm": 0.12001395970582962, "learning_rate": 3.645105423106704e-05, "loss": 0.0212, "step": 30400 }, { "action_loss": 0.0036466270685195923, "epoch": 27.33812949640288, "step": 30400 }, { "epoch": 27.33812949640288, "step": 30400, "torque_loss": 0.1387311816215515 }, { "epoch": 27.347122302158272, "grad_norm": 0.11049757152795792, "learning_rate": 3.6419225082518336e-05, "loss": 0.0191, "step": 30410 }, { "action_loss": 0.002692457986995578, "epoch": 27.347122302158272, "step": 30410 }, { "epoch": 27.347122302158272, "step": 30410, "torque_loss": 0.21266190707683563 }, { "epoch": 27.35611510791367, "grad_norm": 0.10802708566188812, "learning_rate": 3.6387401874661975e-05, "loss": 0.0205, "step": 30420 }, { "action_loss": 0.008490991778671741, "epoch": 27.35611510791367, "step": 30420 }, { "epoch": 27.35611510791367, "step": 30420, "torque_loss": 0.18116426467895508 }, { "epoch": 27.365107913669064, "grad_norm": 0.0745926946401596, "learning_rate": 3.6355584621418516e-05, "loss": 0.0214, "step": 30430 }, { "action_loss": 0.0020168712362647057, "epoch": 27.365107913669064, "step": 30430 }, { "epoch": 27.365107913669064, "step": 30430, "torque_loss": 0.1457200050354004 }, { "epoch": 27.37410071942446, "grad_norm": 0.14922663569450378, "learning_rate": 3.632377333670591e-05, "loss": 0.0192, "step": 30440 }, { "action_loss": 0.0022780185099691153, "epoch": 27.37410071942446, "step": 30440 }, { "epoch": 27.37410071942446, "step": 30440, "torque_loss": 0.14391778409481049 }, { "epoch": 27.383093525179856, "grad_norm": 0.08650868386030197, "learning_rate": 3.629196803443946e-05, "loss": 0.019, "step": 30450 }, { "action_loss": 0.0034122082870453596, "epoch": 27.383093525179856, "step": 30450 }, { "epoch": 27.383093525179856, "step": 30450, "torque_loss": 0.1309633105993271 }, { "epoch": 27.392086330935253, "grad_norm": 0.11541128903627396, "learning_rate": 3.626016872853193e-05, "loss": 0.022, "step": 30460 }, { "action_loss": 0.00281718373298645, "epoch": 27.392086330935253, "step": 30460 }, { "epoch": 27.392086330935253, "step": 30460, "torque_loss": 0.13642103970050812 }, { "epoch": 27.401079136690647, "grad_norm": 0.13355515897274017, "learning_rate": 3.6228375432893405e-05, "loss": 0.0208, "step": 30470 }, { "action_loss": 0.0033257624600082636, "epoch": 27.401079136690647, "step": 30470 }, { "epoch": 27.401079136690647, "step": 30470, "torque_loss": 0.18312613666057587 }, { "epoch": 27.41007194244604, "grad_norm": 0.12111479789018631, "learning_rate": 3.619658816143131e-05, "loss": 0.0231, "step": 30480 }, { "action_loss": 0.006908678915351629, "epoch": 27.41007194244604, "step": 30480 }, { "epoch": 27.41007194244604, "step": 30480, "torque_loss": 0.15271468460559845 }, { "epoch": 27.41906474820144, "grad_norm": 0.15278272330760956, "learning_rate": 3.6164806928050544e-05, "loss": 0.0269, "step": 30490 }, { "action_loss": 0.0019426789367571473, "epoch": 27.41906474820144, "step": 30490 }, { "epoch": 27.41906474820144, "step": 30490, "torque_loss": 0.1550852209329605 }, { "epoch": 27.428057553956833, "grad_norm": 0.10447946935892105, "learning_rate": 3.613303174665327e-05, "loss": 0.0198, "step": 30500 }, { "action_loss": 0.003345302538946271, "epoch": 27.428057553956833, "step": 30500 }, { "epoch": 27.428057553956833, "step": 30500, "torque_loss": 0.16664110124111176 }, { "epoch": 27.43705035971223, "grad_norm": 0.11211450397968292, "learning_rate": 3.6101262631138996e-05, "loss": 0.0187, "step": 30510 }, { "action_loss": 0.004019649233669043, "epoch": 27.43705035971223, "step": 30510 }, { "epoch": 27.43705035971223, "step": 30510, "torque_loss": 0.1322459727525711 }, { "epoch": 27.446043165467625, "grad_norm": 0.12226039916276932, "learning_rate": 3.606949959540467e-05, "loss": 0.0219, "step": 30520 }, { "action_loss": 0.007938220165669918, "epoch": 27.446043165467625, "step": 30520 }, { "epoch": 27.446043165467625, "step": 30520, "torque_loss": 0.20638065040111542 }, { "epoch": 27.455035971223023, "grad_norm": 0.07794566452503204, "learning_rate": 3.603774265334449e-05, "loss": 0.0227, "step": 30530 }, { "action_loss": 0.013236157596111298, "epoch": 27.455035971223023, "step": 30530 }, { "epoch": 27.455035971223023, "step": 30530, "torque_loss": 0.2614925801753998 }, { "epoch": 27.464028776978417, "grad_norm": 0.10820599645376205, "learning_rate": 3.600599181885003e-05, "loss": 0.0224, "step": 30540 }, { "action_loss": 0.0035738374572247267, "epoch": 27.464028776978417, "step": 30540 }, { "epoch": 27.464028776978417, "step": 30540, "torque_loss": 0.1981535702943802 }, { "epoch": 27.473021582733814, "grad_norm": 0.13177178800106049, "learning_rate": 3.597424710581018e-05, "loss": 0.0212, "step": 30550 }, { "action_loss": 0.0018218777840957046, "epoch": 27.473021582733814, "step": 30550 }, { "epoch": 27.473021582733814, "step": 30550, "torque_loss": 0.07454241812229156 }, { "epoch": 27.48201438848921, "grad_norm": 0.11496081203222275, "learning_rate": 3.5942508528111164e-05, "loss": 0.0209, "step": 30560 }, { "action_loss": 0.011581361293792725, "epoch": 27.48201438848921, "step": 30560 }, { "epoch": 27.48201438848921, "step": 30560, "torque_loss": 0.2115193009376526 }, { "epoch": 27.491007194244606, "grad_norm": 0.21508054435253143, "learning_rate": 3.591077609963651e-05, "loss": 0.0269, "step": 30570 }, { "action_loss": 0.0016177943907678127, "epoch": 27.491007194244606, "step": 30570 }, { "epoch": 27.491007194244606, "step": 30570, "torque_loss": 0.13878341019153595 }, { "epoch": 27.5, "grad_norm": 0.11458247154951096, "learning_rate": 3.587904983426704e-05, "loss": 0.0214, "step": 30580 }, { "action_loss": 0.002299673156812787, "epoch": 27.5, "step": 30580 }, { "epoch": 27.5, "step": 30580, "torque_loss": 0.1995590478181839 }, { "epoch": 27.508992805755394, "grad_norm": 0.08266418427228928, "learning_rate": 3.584732974588094e-05, "loss": 0.0228, "step": 30590 }, { "action_loss": 0.006121322046965361, "epoch": 27.508992805755394, "step": 30590 }, { "epoch": 27.508992805755394, "step": 30590, "torque_loss": 0.18923282623291016 }, { "epoch": 27.51798561151079, "grad_norm": 0.08974456042051315, "learning_rate": 3.5815615848353645e-05, "loss": 0.0199, "step": 30600 }, { "action_loss": 0.008363104425370693, "epoch": 27.51798561151079, "step": 30600 }, { "epoch": 27.51798561151079, "step": 30600, "torque_loss": 0.1992030292749405 }, { "epoch": 27.526978417266186, "grad_norm": 0.20254695415496826, "learning_rate": 3.578390815555784e-05, "loss": 0.0227, "step": 30610 }, { "action_loss": 0.003180965781211853, "epoch": 27.526978417266186, "step": 30610 }, { "epoch": 27.526978417266186, "step": 30610, "torque_loss": 0.1986948400735855 }, { "epoch": 27.535971223021583, "grad_norm": 0.2396429181098938, "learning_rate": 3.5752206681363626e-05, "loss": 0.0241, "step": 30620 }, { "action_loss": 0.0018866792088374496, "epoch": 27.535971223021583, "step": 30620 }, { "epoch": 27.535971223021583, "step": 30620, "torque_loss": 0.18051616847515106 }, { "epoch": 27.544964028776977, "grad_norm": 0.1937987059354782, "learning_rate": 3.572051143963826e-05, "loss": 0.0278, "step": 30630 }, { "action_loss": 0.0021406712476164103, "epoch": 27.544964028776977, "step": 30630 }, { "epoch": 27.544964028776977, "step": 30630, "torque_loss": 0.1431119740009308 }, { "epoch": 27.553956834532375, "grad_norm": 0.15023306012153625, "learning_rate": 3.5688822444246294e-05, "loss": 0.0241, "step": 30640 }, { "action_loss": 0.006277123931795359, "epoch": 27.553956834532375, "step": 30640 }, { "epoch": 27.553956834532375, "step": 30640, "torque_loss": 0.15162907540798187 }, { "epoch": 27.56294964028777, "grad_norm": 0.11435942351818085, "learning_rate": 3.565713970904963e-05, "loss": 0.0253, "step": 30650 }, { "action_loss": 0.007067311555147171, "epoch": 27.56294964028777, "step": 30650 }, { "epoch": 27.56294964028777, "step": 30650, "torque_loss": 0.25219032168388367 }, { "epoch": 27.571942446043167, "grad_norm": 0.09238531440496445, "learning_rate": 3.562546324790733e-05, "loss": 0.0219, "step": 30660 }, { "action_loss": 0.011749744415283203, "epoch": 27.571942446043167, "step": 30660 }, { "epoch": 27.571942446043167, "step": 30660, "torque_loss": 0.2518843412399292 }, { "epoch": 27.58093525179856, "grad_norm": 0.08451806753873825, "learning_rate": 3.559379307467577e-05, "loss": 0.0228, "step": 30670 }, { "action_loss": 0.0017994296504184604, "epoch": 27.58093525179856, "step": 30670 }, { "epoch": 27.58093525179856, "step": 30670, "torque_loss": 0.2324695587158203 }, { "epoch": 27.58992805755396, "grad_norm": 0.07311127334833145, "learning_rate": 3.556212920320856e-05, "loss": 0.0214, "step": 30680 }, { "action_loss": 0.006469568703323603, "epoch": 27.58992805755396, "step": 30680 }, { "epoch": 27.58992805755396, "step": 30680, "torque_loss": 0.15603585541248322 }, { "epoch": 27.598920863309353, "grad_norm": 0.09845016151666641, "learning_rate": 3.553047164735655e-05, "loss": 0.0214, "step": 30690 }, { "action_loss": 0.004484797362238169, "epoch": 27.598920863309353, "step": 30690 }, { "epoch": 27.598920863309353, "step": 30690, "torque_loss": 0.17361466586589813 }, { "epoch": 27.607913669064747, "grad_norm": 0.15865296125411987, "learning_rate": 3.549882042096783e-05, "loss": 0.0211, "step": 30700 }, { "action_loss": 0.0035786789376288652, "epoch": 27.607913669064747, "step": 30700 }, { "epoch": 27.607913669064747, "step": 30700, "torque_loss": 0.16720406711101532 }, { "epoch": 27.616906474820144, "grad_norm": 0.20286814868450165, "learning_rate": 3.546717553788772e-05, "loss": 0.0196, "step": 30710 }, { "action_loss": 0.0031113324221223593, "epoch": 27.616906474820144, "step": 30710 }, { "epoch": 27.616906474820144, "step": 30710, "torque_loss": 0.15975429117679596 }, { "epoch": 27.62589928057554, "grad_norm": 0.1481788158416748, "learning_rate": 3.543553701195879e-05, "loss": 0.0219, "step": 30720 }, { "action_loss": 0.0024048974737524986, "epoch": 27.62589928057554, "step": 30720 }, { "epoch": 27.62589928057554, "step": 30720, "torque_loss": 0.13395042717456818 }, { "epoch": 27.634892086330936, "grad_norm": 0.09027183055877686, "learning_rate": 3.540390485702077e-05, "loss": 0.0184, "step": 30730 }, { "action_loss": 0.0017569070914760232, "epoch": 27.634892086330936, "step": 30730 }, { "epoch": 27.634892086330936, "step": 30730, "torque_loss": 0.1683279424905777 }, { "epoch": 27.64388489208633, "grad_norm": 0.06272830069065094, "learning_rate": 3.5372279086910645e-05, "loss": 0.0201, "step": 30740 }, { "action_loss": 0.00499735027551651, "epoch": 27.64388489208633, "step": 30740 }, { "epoch": 27.64388489208633, "step": 30740, "torque_loss": 0.20607821643352509 }, { "epoch": 27.652877697841728, "grad_norm": 0.12352719902992249, "learning_rate": 3.5340659715462635e-05, "loss": 0.0202, "step": 30750 }, { "action_loss": 0.004995288327336311, "epoch": 27.652877697841728, "step": 30750 }, { "epoch": 27.652877697841728, "step": 30750, "torque_loss": 0.1524285525083542 }, { "epoch": 27.66187050359712, "grad_norm": 0.08540654927492142, "learning_rate": 3.530904675650808e-05, "loss": 0.0218, "step": 30760 }, { "action_loss": 0.0017092494526878, "epoch": 27.66187050359712, "step": 30760 }, { "epoch": 27.66187050359712, "step": 30760, "torque_loss": 0.12594352662563324 }, { "epoch": 27.67086330935252, "grad_norm": 0.15088915824890137, "learning_rate": 3.527744022387559e-05, "loss": 0.0215, "step": 30770 }, { "action_loss": 0.0076436265371739864, "epoch": 27.67086330935252, "step": 30770 }, { "epoch": 27.67086330935252, "step": 30770, "torque_loss": 0.23265717923641205 }, { "epoch": 27.679856115107913, "grad_norm": 0.14234358072280884, "learning_rate": 3.524584013139094e-05, "loss": 0.0287, "step": 30780 }, { "action_loss": 0.005656864959746599, "epoch": 27.679856115107913, "step": 30780 }, { "epoch": 27.679856115107913, "step": 30780, "torque_loss": 0.21428661048412323 }, { "epoch": 27.68884892086331, "grad_norm": 0.11006513983011246, "learning_rate": 3.521424649287708e-05, "loss": 0.022, "step": 30790 }, { "action_loss": 0.008721526712179184, "epoch": 27.68884892086331, "step": 30790 }, { "epoch": 27.68884892086331, "step": 30790, "torque_loss": 0.285540372133255 }, { "epoch": 27.697841726618705, "grad_norm": 0.18682831525802612, "learning_rate": 3.518265932215412e-05, "loss": 0.0267, "step": 30800 }, { "action_loss": 0.0024861348792910576, "epoch": 27.697841726618705, "step": 30800 }, { "epoch": 27.697841726618705, "step": 30800, "torque_loss": 0.16646160185337067 }, { "epoch": 27.7068345323741, "grad_norm": 0.13193848729133606, "learning_rate": 3.51510786330394e-05, "loss": 0.0216, "step": 30810 }, { "action_loss": 0.0015533588593825698, "epoch": 27.7068345323741, "step": 30810 }, { "epoch": 27.7068345323741, "step": 30810, "torque_loss": 0.1099221333861351 }, { "epoch": 27.715827338129497, "grad_norm": 0.10659679770469666, "learning_rate": 3.511950443934736e-05, "loss": 0.0175, "step": 30820 }, { "action_loss": 0.00555955246090889, "epoch": 27.715827338129497, "step": 30820 }, { "epoch": 27.715827338129497, "step": 30820, "torque_loss": 0.14945900440216064 }, { "epoch": 27.72482014388489, "grad_norm": 0.19705326855182648, "learning_rate": 3.508793675488961e-05, "loss": 0.0256, "step": 30830 }, { "action_loss": 0.005785790737718344, "epoch": 27.72482014388489, "step": 30830 }, { "epoch": 27.72482014388489, "step": 30830, "torque_loss": 0.14337153732776642 }, { "epoch": 27.73381294964029, "grad_norm": 0.09710820019245148, "learning_rate": 3.505637559347498e-05, "loss": 0.0192, "step": 30840 }, { "action_loss": 0.005642393138259649, "epoch": 27.73381294964029, "step": 30840 }, { "epoch": 27.73381294964029, "step": 30840, "torque_loss": 0.20311033725738525 }, { "epoch": 27.742805755395683, "grad_norm": 0.08483481407165527, "learning_rate": 3.5024820968909345e-05, "loss": 0.02, "step": 30850 }, { "action_loss": 0.0028583004605025053, "epoch": 27.742805755395683, "step": 30850 }, { "epoch": 27.742805755395683, "step": 30850, "torque_loss": 0.16424739360809326 }, { "epoch": 27.75179856115108, "grad_norm": 0.053949907422065735, "learning_rate": 3.4993272894995774e-05, "loss": 0.0193, "step": 30860 }, { "action_loss": 0.003262605518102646, "epoch": 27.75179856115108, "step": 30860 }, { "epoch": 27.75179856115108, "step": 30860, "torque_loss": 0.17756839096546173 }, { "epoch": 27.760791366906474, "grad_norm": 0.10908671468496323, "learning_rate": 3.49617313855345e-05, "loss": 0.0215, "step": 30870 }, { "action_loss": 0.0029529056046158075, "epoch": 27.760791366906474, "step": 30870 }, { "epoch": 27.760791366906474, "step": 30870, "torque_loss": 0.1552206426858902 }, { "epoch": 27.769784172661872, "grad_norm": 0.14875933527946472, "learning_rate": 3.4930196454322834e-05, "loss": 0.0192, "step": 30880 }, { "action_loss": 0.01298731192946434, "epoch": 27.769784172661872, "step": 30880 }, { "epoch": 27.769784172661872, "step": 30880, "torque_loss": 0.1601102352142334 }, { "epoch": 27.778776978417266, "grad_norm": 0.12585504353046417, "learning_rate": 3.48986681151552e-05, "loss": 0.0251, "step": 30890 }, { "action_loss": 0.0020617784466594458, "epoch": 27.778776978417266, "step": 30890 }, { "epoch": 27.778776978417266, "step": 30890, "torque_loss": 0.14722870290279388 }, { "epoch": 27.78776978417266, "grad_norm": 0.11347538977861404, "learning_rate": 3.4867146381823204e-05, "loss": 0.0186, "step": 30900 }, { "action_loss": 0.005854612682014704, "epoch": 27.78776978417266, "step": 30900 }, { "epoch": 27.78776978417266, "step": 30900, "torque_loss": 0.1289881318807602 }, { "epoch": 27.796762589928058, "grad_norm": 0.16114260256290436, "learning_rate": 3.48356312681155e-05, "loss": 0.0213, "step": 30910 }, { "action_loss": 0.002375322161242366, "epoch": 27.796762589928058, "step": 30910 }, { "epoch": 27.796762589928058, "step": 30910, "torque_loss": 0.08687955141067505 }, { "epoch": 27.805755395683452, "grad_norm": 0.12303773313760757, "learning_rate": 3.480412278781788e-05, "loss": 0.0209, "step": 30920 }, { "action_loss": 0.004155390430241823, "epoch": 27.805755395683452, "step": 30920 }, { "epoch": 27.805755395683452, "step": 30920, "torque_loss": 0.11544676870107651 }, { "epoch": 27.81474820143885, "grad_norm": 0.13601957261562347, "learning_rate": 3.47726209547132e-05, "loss": 0.0205, "step": 30930 }, { "action_loss": 0.0026810194831341505, "epoch": 27.81474820143885, "step": 30930 }, { "epoch": 27.81474820143885, "step": 30930, "torque_loss": 0.20043718814849854 }, { "epoch": 27.823741007194243, "grad_norm": 0.10501456260681152, "learning_rate": 3.474112578258147e-05, "loss": 0.0227, "step": 30940 }, { "action_loss": 0.0014737105229869485, "epoch": 27.823741007194243, "step": 30940 }, { "epoch": 27.823741007194243, "step": 30940, "torque_loss": 0.11393038183450699 }, { "epoch": 27.83273381294964, "grad_norm": 0.09810516983270645, "learning_rate": 3.470963728519974e-05, "loss": 0.0193, "step": 30950 }, { "action_loss": 0.001882367767393589, "epoch": 27.83273381294964, "step": 30950 }, { "epoch": 27.83273381294964, "step": 30950, "torque_loss": 0.11295130103826523 }, { "epoch": 27.841726618705035, "grad_norm": 0.1340315341949463, "learning_rate": 3.467815547634211e-05, "loss": 0.0179, "step": 30960 }, { "action_loss": 0.007497523445636034, "epoch": 27.841726618705035, "step": 30960 }, { "epoch": 27.841726618705035, "step": 30960, "torque_loss": 0.21056704223155975 }, { "epoch": 27.850719424460433, "grad_norm": 0.17776808142662048, "learning_rate": 3.4646680369779836e-05, "loss": 0.0232, "step": 30970 }, { "action_loss": 0.004540067631751299, "epoch": 27.850719424460433, "step": 30970 }, { "epoch": 27.850719424460433, "step": 30970, "torque_loss": 0.17641572654247284 }, { "epoch": 27.859712230215827, "grad_norm": 0.09370633214712143, "learning_rate": 3.46152119792812e-05, "loss": 0.0219, "step": 30980 }, { "action_loss": 0.004903512075543404, "epoch": 27.859712230215827, "step": 30980 }, { "epoch": 27.859712230215827, "step": 30980, "torque_loss": 0.1937316656112671 }, { "epoch": 27.868705035971225, "grad_norm": 0.15889839828014374, "learning_rate": 3.45837503186115e-05, "loss": 0.0189, "step": 30990 }, { "action_loss": 0.0054432726465165615, "epoch": 27.868705035971225, "step": 30990 }, { "epoch": 27.868705035971225, "step": 30990, "torque_loss": 0.18454326689243317 }, { "epoch": 27.87769784172662, "grad_norm": 0.12468957155942917, "learning_rate": 3.455229540153318e-05, "loss": 0.0216, "step": 31000 }, { "action_loss": 0.003326557809486985, "epoch": 27.87769784172662, "step": 31000 }, { "epoch": 27.87769784172662, "step": 31000, "torque_loss": 0.13975267112255096 }, { "epoch": 27.886690647482013, "grad_norm": 0.09265166521072388, "learning_rate": 3.4520847241805665e-05, "loss": 0.0203, "step": 31010 }, { "action_loss": 0.0014915341744199395, "epoch": 27.886690647482013, "step": 31010 }, { "epoch": 27.886690647482013, "step": 31010, "torque_loss": 0.11919999122619629 }, { "epoch": 27.89568345323741, "grad_norm": 0.12375739961862564, "learning_rate": 3.4489405853185465e-05, "loss": 0.0182, "step": 31020 }, { "action_loss": 0.007669379469007254, "epoch": 27.89568345323741, "step": 31020 }, { "epoch": 27.89568345323741, "step": 31020, "torque_loss": 0.21082596480846405 }, { "epoch": 27.904676258992804, "grad_norm": 0.09111649543046951, "learning_rate": 3.44579712494261e-05, "loss": 0.0226, "step": 31030 }, { "action_loss": 0.0019618209917098284, "epoch": 27.904676258992804, "step": 31030 }, { "epoch": 27.904676258992804, "step": 31030, "torque_loss": 0.14306621253490448 }, { "epoch": 27.913669064748202, "grad_norm": 0.10283534228801727, "learning_rate": 3.442654344427814e-05, "loss": 0.0217, "step": 31040 }, { "action_loss": 0.001797600300051272, "epoch": 27.913669064748202, "step": 31040 }, { "epoch": 27.913669064748202, "step": 31040, "torque_loss": 0.12337583303451538 }, { "epoch": 27.922661870503596, "grad_norm": 0.10325551778078079, "learning_rate": 3.439512245148916e-05, "loss": 0.0289, "step": 31050 }, { "action_loss": 0.0031581695657223463, "epoch": 27.922661870503596, "step": 31050 }, { "epoch": 27.922661870503596, "step": 31050, "torque_loss": 0.1564221829175949 }, { "epoch": 27.931654676258994, "grad_norm": 0.1353825181722641, "learning_rate": 3.4363708284803796e-05, "loss": 0.0208, "step": 31060 }, { "action_loss": 0.002294059144333005, "epoch": 27.931654676258994, "step": 31060 }, { "epoch": 27.931654676258994, "step": 31060, "torque_loss": 0.12020477652549744 }, { "epoch": 27.940647482014388, "grad_norm": 0.08597324043512344, "learning_rate": 3.433230095796367e-05, "loss": 0.0225, "step": 31070 }, { "action_loss": 0.009162437170743942, "epoch": 27.940647482014388, "step": 31070 }, { "epoch": 27.940647482014388, "step": 31070, "torque_loss": 0.19342462718486786 }, { "epoch": 27.949640287769785, "grad_norm": 0.10118081420660019, "learning_rate": 3.430090048470739e-05, "loss": 0.0238, "step": 31080 }, { "action_loss": 0.0014718960737809539, "epoch": 27.949640287769785, "step": 31080 }, { "epoch": 27.949640287769785, "step": 31080, "torque_loss": 0.17385601997375488 }, { "epoch": 27.95863309352518, "grad_norm": 0.09352227300405502, "learning_rate": 3.4269506878770585e-05, "loss": 0.0206, "step": 31090 }, { "action_loss": 0.0014075598446652293, "epoch": 27.95863309352518, "step": 31090 }, { "epoch": 27.95863309352518, "step": 31090, "torque_loss": 0.179682657122612 }, { "epoch": 27.967625899280577, "grad_norm": 0.14982910454273224, "learning_rate": 3.4238120153885925e-05, "loss": 0.0204, "step": 31100 }, { "action_loss": 0.001088688033632934, "epoch": 27.967625899280577, "step": 31100 }, { "epoch": 27.967625899280577, "step": 31100, "torque_loss": 0.10022753477096558 }, { "epoch": 27.97661870503597, "grad_norm": 0.18497413396835327, "learning_rate": 3.4206740323782994e-05, "loss": 0.0213, "step": 31110 }, { "action_loss": 0.0026485559064894915, "epoch": 27.97661870503597, "step": 31110 }, { "epoch": 27.97661870503597, "step": 31110, "torque_loss": 0.19103388488292694 }, { "epoch": 27.985611510791365, "grad_norm": 0.08537789434194565, "learning_rate": 3.4175367402188395e-05, "loss": 0.021, "step": 31120 }, { "action_loss": 0.004301587585359812, "epoch": 27.985611510791365, "step": 31120 }, { "epoch": 27.985611510791365, "step": 31120, "torque_loss": 0.1434025913476944 }, { "epoch": 27.994604316546763, "grad_norm": 0.08168845623731613, "learning_rate": 3.414400140282574e-05, "loss": 0.0203, "step": 31130 }, { "action_loss": 0.001361040398478508, "epoch": 27.994604316546763, "step": 31130 }, { "epoch": 27.994604316546763, "step": 31130, "torque_loss": 0.07770238071680069 }, { "epoch": 28.003597122302157, "grad_norm": 0.10081585496664047, "learning_rate": 3.411264233941556e-05, "loss": 0.0195, "step": 31140 }, { "action_loss": 0.0032522620167583227, "epoch": 28.003597122302157, "step": 31140 }, { "epoch": 28.003597122302157, "step": 31140, "torque_loss": 0.12054529041051865 }, { "epoch": 28.012589928057555, "grad_norm": 0.1336572915315628, "learning_rate": 3.408129022567537e-05, "loss": 0.0204, "step": 31150 }, { "action_loss": 0.018990719690918922, "epoch": 28.012589928057555, "step": 31150 }, { "epoch": 28.012589928057555, "step": 31150, "torque_loss": 0.2673269510269165 }, { "epoch": 28.02158273381295, "grad_norm": 0.11125960201025009, "learning_rate": 3.404994507531968e-05, "loss": 0.023, "step": 31160 }, { "action_loss": 0.0027312266174703836, "epoch": 28.02158273381295, "step": 31160 }, { "epoch": 28.02158273381295, "step": 31160, "torque_loss": 0.11192747950553894 }, { "epoch": 28.030575539568346, "grad_norm": 0.10602489113807678, "learning_rate": 3.40186069020599e-05, "loss": 0.0198, "step": 31170 }, { "action_loss": 0.002167949452996254, "epoch": 28.030575539568346, "step": 31170 }, { "epoch": 28.030575539568346, "step": 31170, "torque_loss": 0.1374727487564087 }, { "epoch": 28.03956834532374, "grad_norm": 0.09430484473705292, "learning_rate": 3.3987275719604386e-05, "loss": 0.0224, "step": 31180 }, { "action_loss": 0.010488140396773815, "epoch": 28.03956834532374, "step": 31180 }, { "epoch": 28.03956834532374, "step": 31180, "torque_loss": 0.1456911712884903 }, { "epoch": 28.048561151079138, "grad_norm": 0.06694001704454422, "learning_rate": 3.3955951541658526e-05, "loss": 0.0201, "step": 31190 }, { "action_loss": 0.005537931341677904, "epoch": 28.048561151079138, "step": 31190 }, { "epoch": 28.048561151079138, "step": 31190, "torque_loss": 0.21414989233016968 }, { "epoch": 28.057553956834532, "grad_norm": 0.06730049848556519, "learning_rate": 3.392463438192454e-05, "loss": 0.019, "step": 31200 }, { "action_loss": 0.005116456653922796, "epoch": 28.057553956834532, "step": 31200 }, { "epoch": 28.057553956834532, "step": 31200, "torque_loss": 0.17919428646564484 }, { "epoch": 28.06654676258993, "grad_norm": 0.09413666278123856, "learning_rate": 3.38933242541016e-05, "loss": 0.0185, "step": 31210 }, { "action_loss": 0.00200983346439898, "epoch": 28.06654676258993, "step": 31210 }, { "epoch": 28.06654676258993, "step": 31210, "torque_loss": 0.1262333244085312 }, { "epoch": 28.075539568345324, "grad_norm": 0.1230282336473465, "learning_rate": 3.386202117188586e-05, "loss": 0.0203, "step": 31220 }, { "action_loss": 0.0016026148805394769, "epoch": 28.075539568345324, "step": 31220 }, { "epoch": 28.075539568345324, "step": 31220, "torque_loss": 0.1259116679430008 }, { "epoch": 28.084532374100718, "grad_norm": 0.09970535337924957, "learning_rate": 3.383072514897033e-05, "loss": 0.0206, "step": 31230 }, { "action_loss": 0.0021829979959875345, "epoch": 28.084532374100718, "step": 31230 }, { "epoch": 28.084532374100718, "step": 31230, "torque_loss": 0.15743890404701233 }, { "epoch": 28.093525179856115, "grad_norm": 0.15346063673496246, "learning_rate": 3.379943619904494e-05, "loss": 0.0204, "step": 31240 }, { "action_loss": 0.004266119096428156, "epoch": 28.093525179856115, "step": 31240 }, { "epoch": 28.093525179856115, "step": 31240, "torque_loss": 0.16131757199764252 }, { "epoch": 28.10251798561151, "grad_norm": 0.10691185295581818, "learning_rate": 3.3768154335796583e-05, "loss": 0.0162, "step": 31250 }, { "action_loss": 0.0033982235472649336, "epoch": 28.10251798561151, "step": 31250 }, { "epoch": 28.10251798561151, "step": 31250, "torque_loss": 0.188104510307312 }, { "epoch": 28.111510791366907, "grad_norm": 0.10417689383029938, "learning_rate": 3.3736879572908963e-05, "loss": 0.0189, "step": 31260 }, { "action_loss": 0.0011839530197903514, "epoch": 28.111510791366907, "step": 31260 }, { "epoch": 28.111510791366907, "step": 31260, "torque_loss": 0.12255824357271194 }, { "epoch": 28.1205035971223, "grad_norm": 0.16482071578502655, "learning_rate": 3.370561192406276e-05, "loss": 0.018, "step": 31270 }, { "action_loss": 0.0036599962040781975, "epoch": 28.1205035971223, "step": 31270 }, { "epoch": 28.1205035971223, "step": 31270, "torque_loss": 0.12764307856559753 }, { "epoch": 28.1294964028777, "grad_norm": 0.09218344837427139, "learning_rate": 3.367435140293547e-05, "loss": 0.0203, "step": 31280 }, { "action_loss": 0.00857881922274828, "epoch": 28.1294964028777, "step": 31280 }, { "epoch": 28.1294964028777, "step": 31280, "torque_loss": 0.16739505529403687 }, { "epoch": 28.138489208633093, "grad_norm": 0.17725645005702972, "learning_rate": 3.3643098023201545e-05, "loss": 0.0268, "step": 31290 }, { "action_loss": 0.0045646489597857, "epoch": 28.138489208633093, "step": 31290 }, { "epoch": 28.138489208633093, "step": 31290, "torque_loss": 0.13186727464199066 }, { "epoch": 28.14748201438849, "grad_norm": 0.11425267904996872, "learning_rate": 3.361185179853225e-05, "loss": 0.0201, "step": 31300 }, { "action_loss": 0.0015209442935883999, "epoch": 28.14748201438849, "step": 31300 }, { "epoch": 28.14748201438849, "step": 31300, "torque_loss": 0.08943650126457214 }, { "epoch": 28.156474820143885, "grad_norm": 0.08275480568408966, "learning_rate": 3.358061274259574e-05, "loss": 0.0189, "step": 31310 }, { "action_loss": 0.0028736002277582884, "epoch": 28.156474820143885, "step": 31310 }, { "epoch": 28.156474820143885, "step": 31310, "torque_loss": 0.18325643241405487 }, { "epoch": 28.165467625899282, "grad_norm": 0.1277395486831665, "learning_rate": 3.3549380869057064e-05, "loss": 0.0228, "step": 31320 }, { "action_loss": 0.0014967983588576317, "epoch": 28.165467625899282, "step": 31320 }, { "epoch": 28.165467625899282, "step": 31320, "torque_loss": 0.1441316157579422 }, { "epoch": 28.174460431654676, "grad_norm": 0.0847783014178276, "learning_rate": 3.351815619157809e-05, "loss": 0.0209, "step": 31330 }, { "action_loss": 0.0023775214795023203, "epoch": 28.174460431654676, "step": 31330 }, { "epoch": 28.174460431654676, "step": 31330, "torque_loss": 0.13019883632659912 }, { "epoch": 28.18345323741007, "grad_norm": 0.0801793560385704, "learning_rate": 3.348693872381754e-05, "loss": 0.0211, "step": 31340 }, { "action_loss": 0.0034168416168540716, "epoch": 28.18345323741007, "step": 31340 }, { "epoch": 28.18345323741007, "step": 31340, "torque_loss": 0.08068560063838959 }, { "epoch": 28.192446043165468, "grad_norm": 0.08449673652648926, "learning_rate": 3.3455728479431025e-05, "loss": 0.0184, "step": 31350 }, { "action_loss": 0.007304813712835312, "epoch": 28.192446043165468, "step": 31350 }, { "epoch": 28.192446043165468, "step": 31350, "torque_loss": 0.28821149468421936 }, { "epoch": 28.201438848920862, "grad_norm": 0.08336354047060013, "learning_rate": 3.342452547207094e-05, "loss": 0.02, "step": 31360 }, { "action_loss": 0.006134776398539543, "epoch": 28.201438848920862, "step": 31360 }, { "epoch": 28.201438848920862, "step": 31360, "torque_loss": 0.18074189126491547 }, { "epoch": 28.21043165467626, "grad_norm": 0.1536443829536438, "learning_rate": 3.3393329715386555e-05, "loss": 0.0206, "step": 31370 }, { "action_loss": 0.0014227397041395307, "epoch": 28.21043165467626, "step": 31370 }, { "epoch": 28.21043165467626, "step": 31370, "torque_loss": 0.12711748480796814 }, { "epoch": 28.219424460431654, "grad_norm": 0.09778077155351639, "learning_rate": 3.336214122302395e-05, "loss": 0.0225, "step": 31380 }, { "action_loss": 0.0023821862414479256, "epoch": 28.219424460431654, "step": 31380 }, { "epoch": 28.219424460431654, "step": 31380, "torque_loss": 0.20195479691028595 }, { "epoch": 28.22841726618705, "grad_norm": 0.08796828985214233, "learning_rate": 3.333096000862604e-05, "loss": 0.0191, "step": 31390 }, { "action_loss": 0.002850495046004653, "epoch": 28.22841726618705, "step": 31390 }, { "epoch": 28.22841726618705, "step": 31390, "torque_loss": 0.1334175020456314 }, { "epoch": 28.237410071942445, "grad_norm": 0.11104850471019745, "learning_rate": 3.3299786085832516e-05, "loss": 0.0178, "step": 31400 }, { "action_loss": 0.002860400825738907, "epoch": 28.237410071942445, "step": 31400 }, { "epoch": 28.237410071942445, "step": 31400, "torque_loss": 0.19894623756408691 }, { "epoch": 28.246402877697843, "grad_norm": 0.09065321832895279, "learning_rate": 3.3268619468279956e-05, "loss": 0.0245, "step": 31410 }, { "action_loss": 0.001223485916852951, "epoch": 28.246402877697843, "step": 31410 }, { "epoch": 28.246402877697843, "step": 31410, "torque_loss": 0.1094021201133728 }, { "epoch": 28.255395683453237, "grad_norm": 0.08587372303009033, "learning_rate": 3.323746016960167e-05, "loss": 0.0175, "step": 31420 }, { "action_loss": 0.015518649481236935, "epoch": 28.255395683453237, "step": 31420 }, { "epoch": 28.255395683453237, "step": 31420, "torque_loss": 0.1958705335855484 }, { "epoch": 28.264388489208635, "grad_norm": 0.14131848514080048, "learning_rate": 3.3206308203427806e-05, "loss": 0.0238, "step": 31430 }, { "action_loss": 0.0021933906245976686, "epoch": 28.264388489208635, "step": 31430 }, { "epoch": 28.264388489208635, "step": 31430, "torque_loss": 0.13611368834972382 }, { "epoch": 28.27338129496403, "grad_norm": 0.08683064579963684, "learning_rate": 3.3175163583385254e-05, "loss": 0.0206, "step": 31440 }, { "action_loss": 0.0019082920625805855, "epoch": 28.27338129496403, "step": 31440 }, { "epoch": 28.27338129496403, "step": 31440, "torque_loss": 0.13989071547985077 }, { "epoch": 28.282374100719423, "grad_norm": 0.16418884694576263, "learning_rate": 3.314402632309779e-05, "loss": 0.0198, "step": 31450 }, { "action_loss": 0.002224990399554372, "epoch": 28.282374100719423, "step": 31450 }, { "epoch": 28.282374100719423, "step": 31450, "torque_loss": 0.17244593799114227 }, { "epoch": 28.29136690647482, "grad_norm": 0.11428585648536682, "learning_rate": 3.311289643618587e-05, "loss": 0.0177, "step": 31460 }, { "action_loss": 0.019518202170729637, "epoch": 28.29136690647482, "step": 31460 }, { "epoch": 28.29136690647482, "step": 31460, "torque_loss": 0.2159142941236496 }, { "epoch": 28.300359712230215, "grad_norm": 0.08562730997800827, "learning_rate": 3.308177393626676e-05, "loss": 0.0215, "step": 31470 }, { "action_loss": 0.0034897755831480026, "epoch": 28.300359712230215, "step": 31470 }, { "epoch": 28.300359712230215, "step": 31470, "torque_loss": 0.17795205116271973 }, { "epoch": 28.309352517985612, "grad_norm": 0.06975605338811874, "learning_rate": 3.305065883695453e-05, "loss": 0.0222, "step": 31480 }, { "action_loss": 0.004819151014089584, "epoch": 28.309352517985612, "step": 31480 }, { "epoch": 28.309352517985612, "step": 31480, "torque_loss": 0.16515950858592987 }, { "epoch": 28.318345323741006, "grad_norm": 0.126082181930542, "learning_rate": 3.301955115185996e-05, "loss": 0.0209, "step": 31490 }, { "action_loss": 0.0031302517745643854, "epoch": 28.318345323741006, "step": 31490 }, { "epoch": 28.318345323741006, "step": 31490, "torque_loss": 0.1768290400505066 }, { "epoch": 28.327338129496404, "grad_norm": 0.11345570534467697, "learning_rate": 3.29884508945906e-05, "loss": 0.0196, "step": 31500 }, { "action_loss": 0.009390055201947689, "epoch": 28.327338129496404, "step": 31500 }, { "epoch": 28.327338129496404, "step": 31500, "torque_loss": 0.1832987666130066 }, { "epoch": 28.336330935251798, "grad_norm": 0.1792803704738617, "learning_rate": 3.295735807875079e-05, "loss": 0.0213, "step": 31510 }, { "action_loss": 0.0021693771705031395, "epoch": 28.336330935251798, "step": 31510 }, { "epoch": 28.336330935251798, "step": 31510, "torque_loss": 0.1653016209602356 }, { "epoch": 28.345323741007196, "grad_norm": 0.11212191730737686, "learning_rate": 3.292627271794156e-05, "loss": 0.0208, "step": 31520 }, { "action_loss": 0.0076895300298929214, "epoch": 28.345323741007196, "step": 31520 }, { "epoch": 28.345323741007196, "step": 31520, "torque_loss": 0.23436814546585083 }, { "epoch": 28.35431654676259, "grad_norm": 0.1571553498506546, "learning_rate": 3.289519482576069e-05, "loss": 0.0208, "step": 31530 }, { "action_loss": 0.007069619372487068, "epoch": 28.35431654676259, "step": 31530 }, { "epoch": 28.35431654676259, "step": 31530, "torque_loss": 0.1648261994123459 }, { "epoch": 28.363309352517987, "grad_norm": 0.16213572025299072, "learning_rate": 3.286412441580274e-05, "loss": 0.0232, "step": 31540 }, { "action_loss": 0.00859085377305746, "epoch": 28.363309352517987, "step": 31540 }, { "epoch": 28.363309352517987, "step": 31540, "torque_loss": 0.18275976181030273 }, { "epoch": 28.37230215827338, "grad_norm": 0.12878035008907318, "learning_rate": 3.283306150165896e-05, "loss": 0.0235, "step": 31550 }, { "action_loss": 0.010539297014474869, "epoch": 28.37230215827338, "step": 31550 }, { "epoch": 28.37230215827338, "step": 31550, "torque_loss": 0.2060130387544632 }, { "epoch": 28.381294964028775, "grad_norm": 0.14221642911434174, "learning_rate": 3.2802006096917285e-05, "loss": 0.0229, "step": 31560 }, { "action_loss": 0.0017771568382158875, "epoch": 28.381294964028775, "step": 31560 }, { "epoch": 28.381294964028775, "step": 31560, "torque_loss": 0.17578403651714325 }, { "epoch": 28.390287769784173, "grad_norm": 0.13526460528373718, "learning_rate": 3.277095821516245e-05, "loss": 0.0191, "step": 31570 }, { "action_loss": 0.002272867364808917, "epoch": 28.390287769784173, "step": 31570 }, { "epoch": 28.390287769784173, "step": 31570, "torque_loss": 0.18258611857891083 }, { "epoch": 28.399280575539567, "grad_norm": 0.10106790065765381, "learning_rate": 3.273991786997583e-05, "loss": 0.0233, "step": 31580 }, { "action_loss": 0.0023840491194278, "epoch": 28.399280575539567, "step": 31580 }, { "epoch": 28.399280575539567, "step": 31580, "torque_loss": 0.08619040250778198 }, { "epoch": 28.408273381294965, "grad_norm": 0.0839749276638031, "learning_rate": 3.2708885074935514e-05, "loss": 0.0187, "step": 31590 }, { "action_loss": 0.0040284814313054085, "epoch": 28.408273381294965, "step": 31590 }, { "epoch": 28.408273381294965, "step": 31590, "torque_loss": 0.2532986104488373 }, { "epoch": 28.41726618705036, "grad_norm": 0.17244893312454224, "learning_rate": 3.267785984361633e-05, "loss": 0.0248, "step": 31600 }, { "action_loss": 0.0017630374059081078, "epoch": 28.41726618705036, "step": 31600 }, { "epoch": 28.41726618705036, "step": 31600, "torque_loss": 0.1364748626947403 }, { "epoch": 28.426258992805757, "grad_norm": 0.12804298102855682, "learning_rate": 3.2646842189589754e-05, "loss": 0.0211, "step": 31610 }, { "action_loss": 0.0015797233209013939, "epoch": 28.426258992805757, "step": 31610 }, { "epoch": 28.426258992805757, "step": 31610, "torque_loss": 0.1078765019774437 }, { "epoch": 28.43525179856115, "grad_norm": 0.15445230901241302, "learning_rate": 3.261583212642397e-05, "loss": 0.0191, "step": 31620 }, { "action_loss": 0.0014333397848531604, "epoch": 28.43525179856115, "step": 31620 }, { "epoch": 28.43525179856115, "step": 31620, "torque_loss": 0.12916041910648346 }, { "epoch": 28.444244604316548, "grad_norm": 0.17140518128871918, "learning_rate": 3.25848296676838e-05, "loss": 0.0257, "step": 31630 }, { "action_loss": 0.0019960133358836174, "epoch": 28.444244604316548, "step": 31630 }, { "epoch": 28.444244604316548, "step": 31630, "torque_loss": 0.1491376757621765 }, { "epoch": 28.453237410071942, "grad_norm": 0.1594138890504837, "learning_rate": 3.255383482693083e-05, "loss": 0.0209, "step": 31640 }, { "action_loss": 0.0021318315993994474, "epoch": 28.453237410071942, "step": 31640 }, { "epoch": 28.453237410071942, "step": 31640, "torque_loss": 0.0908394381403923 }, { "epoch": 28.46223021582734, "grad_norm": 0.09033747017383575, "learning_rate": 3.252284761772322e-05, "loss": 0.0222, "step": 31650 }, { "action_loss": 0.003006397048011422, "epoch": 28.46223021582734, "step": 31650 }, { "epoch": 28.46223021582734, "step": 31650, "torque_loss": 0.11005163192749023 }, { "epoch": 28.471223021582734, "grad_norm": 0.1844639927148819, "learning_rate": 3.249186805361581e-05, "loss": 0.0225, "step": 31660 }, { "action_loss": 0.00434848153963685, "epoch": 28.471223021582734, "step": 31660 }, { "epoch": 28.471223021582734, "step": 31660, "torque_loss": 0.15793690085411072 }, { "epoch": 28.480215827338128, "grad_norm": 0.1408555954694748, "learning_rate": 3.246089614816016e-05, "loss": 0.0189, "step": 31670 }, { "action_loss": 0.017871325835585594, "epoch": 28.480215827338128, "step": 31670 }, { "epoch": 28.480215827338128, "step": 31670, "torque_loss": 0.20009076595306396 }, { "epoch": 28.489208633093526, "grad_norm": 0.05831293389201164, "learning_rate": 3.242993191490441e-05, "loss": 0.0206, "step": 31680 }, { "action_loss": 0.0035017505288124084, "epoch": 28.489208633093526, "step": 31680 }, { "epoch": 28.489208633093526, "step": 31680, "torque_loss": 0.17129535973072052 }, { "epoch": 28.49820143884892, "grad_norm": 0.15574558079242706, "learning_rate": 3.239897536739335e-05, "loss": 0.0283, "step": 31690 }, { "action_loss": 0.0027207685634493828, "epoch": 28.49820143884892, "step": 31690 }, { "epoch": 28.49820143884892, "step": 31690, "torque_loss": 0.18165405094623566 }, { "epoch": 28.507194244604317, "grad_norm": 0.06086970493197441, "learning_rate": 3.236802651916845e-05, "loss": 0.0179, "step": 31700 }, { "action_loss": 0.004797490779310465, "epoch": 28.507194244604317, "step": 31700 }, { "epoch": 28.507194244604317, "step": 31700, "torque_loss": 0.16957835853099823 }, { "epoch": 28.51618705035971, "grad_norm": 0.1803838312625885, "learning_rate": 3.23370853837678e-05, "loss": 0.0194, "step": 31710 }, { "action_loss": 0.002451079897582531, "epoch": 28.51618705035971, "step": 31710 }, { "epoch": 28.51618705035971, "step": 31710, "torque_loss": 0.14559338986873627 }, { "epoch": 28.52517985611511, "grad_norm": 0.08614952117204666, "learning_rate": 3.2306151974726056e-05, "loss": 0.02, "step": 31720 }, { "action_loss": 0.003116508014500141, "epoch": 28.52517985611511, "step": 31720 }, { "epoch": 28.52517985611511, "step": 31720, "torque_loss": 0.16574490070343018 }, { "epoch": 28.534172661870503, "grad_norm": 0.1315573751926422, "learning_rate": 3.227522630557459e-05, "loss": 0.0187, "step": 31730 }, { "action_loss": 0.004052350763231516, "epoch": 28.534172661870503, "step": 31730 }, { "epoch": 28.534172661870503, "step": 31730, "torque_loss": 0.26096248626708984 }, { "epoch": 28.5431654676259, "grad_norm": 0.07235080003738403, "learning_rate": 3.224430838984132e-05, "loss": 0.0212, "step": 31740 }, { "action_loss": 0.01633908413350582, "epoch": 28.5431654676259, "step": 31740 }, { "epoch": 28.5431654676259, "step": 31740, "torque_loss": 0.22580939531326294 }, { "epoch": 28.552158273381295, "grad_norm": 0.12708014249801636, "learning_rate": 3.2213398241050784e-05, "loss": 0.0263, "step": 31750 }, { "action_loss": 0.012322376482188702, "epoch": 28.552158273381295, "step": 31750 }, { "epoch": 28.552158273381295, "step": 31750, "torque_loss": 0.19965839385986328 }, { "epoch": 28.56115107913669, "grad_norm": 0.10529501736164093, "learning_rate": 3.218249587272415e-05, "loss": 0.0232, "step": 31760 }, { "action_loss": 0.0008971720817498863, "epoch": 28.56115107913669, "step": 31760 }, { "epoch": 28.56115107913669, "step": 31760, "torque_loss": 0.10747227817773819 }, { "epoch": 28.570143884892087, "grad_norm": 0.07047266513109207, "learning_rate": 3.215160129837916e-05, "loss": 0.0182, "step": 31770 }, { "action_loss": 0.0017767191166058183, "epoch": 28.570143884892087, "step": 31770 }, { "epoch": 28.570143884892087, "step": 31770, "torque_loss": 0.1110009178519249 }, { "epoch": 28.57913669064748, "grad_norm": 0.12397440522909164, "learning_rate": 3.2120714531530146e-05, "loss": 0.0217, "step": 31780 }, { "action_loss": 0.00830161664634943, "epoch": 28.57913669064748, "step": 31780 }, { "epoch": 28.57913669064748, "step": 31780, "torque_loss": 0.17061424255371094 }, { "epoch": 28.58812949640288, "grad_norm": 0.06659765541553497, "learning_rate": 3.2089835585688014e-05, "loss": 0.0217, "step": 31790 }, { "action_loss": 0.005224976688623428, "epoch": 28.58812949640288, "step": 31790 }, { "epoch": 28.58812949640288, "step": 31790, "torque_loss": 0.16645973920822144 }, { "epoch": 28.597122302158272, "grad_norm": 0.07916451245546341, "learning_rate": 3.20589644743603e-05, "loss": 0.0214, "step": 31800 }, { "action_loss": 0.0066188327036798, "epoch": 28.597122302158272, "step": 31800 }, { "epoch": 28.597122302158272, "step": 31800, "torque_loss": 0.17959624528884888 }, { "epoch": 28.60611510791367, "grad_norm": 0.0956561267375946, "learning_rate": 3.2028101211051054e-05, "loss": 0.0245, "step": 31810 }, { "action_loss": 0.0018236382165923715, "epoch": 28.60611510791367, "step": 31810 }, { "epoch": 28.60611510791367, "step": 31810, "torque_loss": 0.19249732792377472 }, { "epoch": 28.615107913669064, "grad_norm": 0.10467629134654999, "learning_rate": 3.19972458092609e-05, "loss": 0.023, "step": 31820 }, { "action_loss": 0.0034542439971119165, "epoch": 28.615107913669064, "step": 31820 }, { "epoch": 28.615107913669064, "step": 31820, "torque_loss": 0.16831856966018677 }, { "epoch": 28.62410071942446, "grad_norm": 0.10525064915418625, "learning_rate": 3.196639828248707e-05, "loss": 0.0224, "step": 31830 }, { "action_loss": 0.004958712961524725, "epoch": 28.62410071942446, "step": 31830 }, { "epoch": 28.62410071942446, "step": 31830, "torque_loss": 0.1664915680885315 }, { "epoch": 28.633093525179856, "grad_norm": 0.07330106198787689, "learning_rate": 3.193555864422329e-05, "loss": 0.0205, "step": 31840 }, { "action_loss": 0.011655117385089397, "epoch": 28.633093525179856, "step": 31840 }, { "epoch": 28.633093525179856, "step": 31840, "torque_loss": 0.20839405059814453 }, { "epoch": 28.642086330935253, "grad_norm": 0.17743226885795593, "learning_rate": 3.19047269079599e-05, "loss": 0.0304, "step": 31850 }, { "action_loss": 0.0017776374006643891, "epoch": 28.642086330935253, "step": 31850 }, { "epoch": 28.642086330935253, "step": 31850, "torque_loss": 0.14631278812885284 }, { "epoch": 28.651079136690647, "grad_norm": 0.1256626397371292, "learning_rate": 3.1873903087183734e-05, "loss": 0.0217, "step": 31860 }, { "action_loss": 0.0062319389544427395, "epoch": 28.651079136690647, "step": 31860 }, { "epoch": 28.651079136690647, "step": 31860, "torque_loss": 0.22491534054279327 }, { "epoch": 28.66007194244604, "grad_norm": 0.10475005954504013, "learning_rate": 3.1843087195378176e-05, "loss": 0.0183, "step": 31870 }, { "action_loss": 0.006071853917092085, "epoch": 28.66007194244604, "step": 31870 }, { "epoch": 28.66007194244604, "step": 31870, "torque_loss": 0.18864862620830536 }, { "epoch": 28.66906474820144, "grad_norm": 0.16048474609851837, "learning_rate": 3.181227924602312e-05, "loss": 0.0216, "step": 31880 }, { "action_loss": 0.006966667715460062, "epoch": 28.66906474820144, "step": 31880 }, { "epoch": 28.66906474820144, "step": 31880, "torque_loss": 0.1572665572166443 }, { "epoch": 28.678057553956833, "grad_norm": 0.14149990677833557, "learning_rate": 3.178147925259506e-05, "loss": 0.0223, "step": 31890 }, { "action_loss": 0.0024468740448355675, "epoch": 28.678057553956833, "step": 31890 }, { "epoch": 28.678057553956833, "step": 31890, "torque_loss": 0.1969764679670334 }, { "epoch": 28.68705035971223, "grad_norm": 0.15002650022506714, "learning_rate": 3.175068722856693e-05, "loss": 0.0219, "step": 31900 }, { "action_loss": 0.009818772785365582, "epoch": 28.68705035971223, "step": 31900 }, { "epoch": 28.68705035971223, "step": 31900, "torque_loss": 0.16745346784591675 }, { "epoch": 28.696043165467625, "grad_norm": 0.10018232464790344, "learning_rate": 3.17199031874082e-05, "loss": 0.0248, "step": 31910 }, { "action_loss": 0.0020968790631741285, "epoch": 28.696043165467625, "step": 31910 }, { "epoch": 28.696043165467625, "step": 31910, "torque_loss": 0.1471724957227707 }, { "epoch": 28.705035971223023, "grad_norm": 0.13150444626808167, "learning_rate": 3.168912714258487e-05, "loss": 0.0236, "step": 31920 }, { "action_loss": 0.0071662371046841145, "epoch": 28.705035971223023, "step": 31920 }, { "epoch": 28.705035971223023, "step": 31920, "torque_loss": 0.22965092957019806 }, { "epoch": 28.714028776978417, "grad_norm": 0.11821627616882324, "learning_rate": 3.165835910755944e-05, "loss": 0.0233, "step": 31930 }, { "action_loss": 0.004787010606378317, "epoch": 28.714028776978417, "step": 31930 }, { "epoch": 28.714028776978417, "step": 31930, "torque_loss": 0.11111640930175781 }, { "epoch": 28.723021582733814, "grad_norm": 0.17807334661483765, "learning_rate": 3.1627599095790874e-05, "loss": 0.0198, "step": 31940 }, { "action_loss": 0.0021072078961879015, "epoch": 28.723021582733814, "step": 31940 }, { "epoch": 28.723021582733814, "step": 31940, "torque_loss": 0.1985653191804886 }, { "epoch": 28.73201438848921, "grad_norm": 0.05385740473866463, "learning_rate": 3.159684712073462e-05, "loss": 0.0188, "step": 31950 }, { "action_loss": 0.006990883965045214, "epoch": 28.73201438848921, "step": 31950 }, { "epoch": 28.73201438848921, "step": 31950, "torque_loss": 0.25535377860069275 }, { "epoch": 28.741007194244606, "grad_norm": 0.10229743272066116, "learning_rate": 3.1566103195842695e-05, "loss": 0.0233, "step": 31960 }, { "action_loss": 0.0023460660595446825, "epoch": 28.741007194244606, "step": 31960 }, { "epoch": 28.741007194244606, "step": 31960, "torque_loss": 0.14913193881511688 }, { "epoch": 28.75, "grad_norm": 0.07627954334020615, "learning_rate": 3.153536733456349e-05, "loss": 0.0189, "step": 31970 }, { "action_loss": 0.00317993201315403, "epoch": 28.75, "step": 31970 }, { "epoch": 28.75, "step": 31970, "torque_loss": 0.12622642517089844 }, { "epoch": 28.758992805755394, "grad_norm": 0.08063183724880219, "learning_rate": 3.150463955034194e-05, "loss": 0.0208, "step": 31980 }, { "action_loss": 0.003154151141643524, "epoch": 28.758992805755394, "step": 31980 }, { "epoch": 28.758992805755394, "step": 31980, "torque_loss": 0.1989678293466568 }, { "epoch": 28.76798561151079, "grad_norm": 0.17098680138587952, "learning_rate": 3.147391985661942e-05, "loss": 0.0196, "step": 31990 }, { "action_loss": 0.0020737105514854193, "epoch": 28.76798561151079, "step": 31990 }, { "epoch": 28.76798561151079, "step": 31990, "torque_loss": 0.13289552927017212 }, { "epoch": 28.776978417266186, "grad_norm": 0.08052986115217209, "learning_rate": 3.1443208266833755e-05, "loss": 0.0197, "step": 32000 }, { "action_loss": 0.008520659059286118, "epoch": 28.776978417266186, "step": 32000 }, { "epoch": 28.776978417266186, "step": 32000, "torque_loss": 0.23236209154129028 }, { "epoch": 28.785971223021583, "grad_norm": 0.08078768849372864, "learning_rate": 3.141250479441924e-05, "loss": 0.0191, "step": 32010 }, { "action_loss": 0.0030669542029500008, "epoch": 28.785971223021583, "step": 32010 }, { "epoch": 28.785971223021583, "step": 32010, "torque_loss": 0.18595372140407562 }, { "epoch": 28.794964028776977, "grad_norm": 0.07277786731719971, "learning_rate": 3.138180945280664e-05, "loss": 0.0202, "step": 32020 }, { "action_loss": 0.0021955145057290792, "epoch": 28.794964028776977, "step": 32020 }, { "epoch": 28.794964028776977, "step": 32020, "torque_loss": 0.17667287588119507 }, { "epoch": 28.803956834532375, "grad_norm": 0.0765414610505104, "learning_rate": 3.1351122255423116e-05, "loss": 0.0166, "step": 32030 }, { "action_loss": 0.0014743093634024262, "epoch": 28.803956834532375, "step": 32030 }, { "epoch": 28.803956834532375, "step": 32030, "torque_loss": 0.13356781005859375 }, { "epoch": 28.81294964028777, "grad_norm": 0.10441063344478607, "learning_rate": 3.132044321569228e-05, "loss": 0.0189, "step": 32040 }, { "action_loss": 0.0011345226084813476, "epoch": 28.81294964028777, "step": 32040 }, { "epoch": 28.81294964028777, "step": 32040, "torque_loss": 0.07393792271614075 }, { "epoch": 28.821942446043167, "grad_norm": 0.09046515822410583, "learning_rate": 3.1289772347034226e-05, "loss": 0.0207, "step": 32050 }, { "action_loss": 0.00659811869263649, "epoch": 28.821942446043167, "step": 32050 }, { "epoch": 28.821942446043167, "step": 32050, "torque_loss": 0.22895009815692902 }, { "epoch": 28.83093525179856, "grad_norm": 0.1221834048628807, "learning_rate": 3.1259109662865404e-05, "loss": 0.0187, "step": 32060 }, { "action_loss": 0.0018987784860655665, "epoch": 28.83093525179856, "step": 32060 }, { "epoch": 28.83093525179856, "step": 32060, "torque_loss": 0.14758996665477753 }, { "epoch": 28.83992805755396, "grad_norm": 0.08075769990682602, "learning_rate": 3.122845517659871e-05, "loss": 0.0187, "step": 32070 }, { "action_loss": 0.0033621962647885084, "epoch": 28.83992805755396, "step": 32070 }, { "epoch": 28.83992805755396, "step": 32070, "torque_loss": 0.12209158390760422 }, { "epoch": 28.848920863309353, "grad_norm": 0.12162955105304718, "learning_rate": 3.119780890164348e-05, "loss": 0.0217, "step": 32080 }, { "action_loss": 0.0015116498107090592, "epoch": 28.848920863309353, "step": 32080 }, { "epoch": 28.848920863309353, "step": 32080, "torque_loss": 0.1595485359430313 }, { "epoch": 28.857913669064747, "grad_norm": 0.10727350413799286, "learning_rate": 3.116717085140542e-05, "loss": 0.0222, "step": 32090 }, { "action_loss": 0.003850676817819476, "epoch": 28.857913669064747, "step": 32090 }, { "epoch": 28.857913669064747, "step": 32090, "torque_loss": 0.19333255290985107 }, { "epoch": 28.866906474820144, "grad_norm": 0.08193277567625046, "learning_rate": 3.1136541039286657e-05, "loss": 0.0205, "step": 32100 }, { "action_loss": 0.004155144095420837, "epoch": 28.866906474820144, "step": 32100 }, { "epoch": 28.866906474820144, "step": 32100, "torque_loss": 0.14650800824165344 }, { "epoch": 28.87589928057554, "grad_norm": 0.0659315288066864, "learning_rate": 3.110591947868571e-05, "loss": 0.0194, "step": 32110 }, { "action_loss": 0.0008439512457698584, "epoch": 28.87589928057554, "step": 32110 }, { "epoch": 28.87589928057554, "step": 32110, "torque_loss": 0.12629328668117523 }, { "epoch": 28.884892086330936, "grad_norm": 0.09662966430187225, "learning_rate": 3.107530618299751e-05, "loss": 0.0204, "step": 32120 }, { "action_loss": 0.001013195957057178, "epoch": 28.884892086330936, "step": 32120 }, { "epoch": 28.884892086330936, "step": 32120, "torque_loss": 0.08498212695121765 }, { "epoch": 28.89388489208633, "grad_norm": 0.11788192391395569, "learning_rate": 3.104470116561333e-05, "loss": 0.0182, "step": 32130 }, { "action_loss": 0.0017374572344124317, "epoch": 28.89388489208633, "step": 32130 }, { "epoch": 28.89388489208633, "step": 32130, "torque_loss": 0.1218065619468689 }, { "epoch": 28.902877697841728, "grad_norm": 0.13601109385490417, "learning_rate": 3.1014104439920835e-05, "loss": 0.0204, "step": 32140 }, { "action_loss": 0.010837105102837086, "epoch": 28.902877697841728, "step": 32140 }, { "epoch": 28.902877697841728, "step": 32140, "torque_loss": 0.19214856624603271 }, { "epoch": 28.91187050359712, "grad_norm": 0.1397956907749176, "learning_rate": 3.098351601930411e-05, "loss": 0.0258, "step": 32150 }, { "action_loss": 0.0033748254645615816, "epoch": 28.91187050359712, "step": 32150 }, { "epoch": 28.91187050359712, "step": 32150, "torque_loss": 0.14768366515636444 }, { "epoch": 28.92086330935252, "grad_norm": 0.18616142868995667, "learning_rate": 3.0952935917143533e-05, "loss": 0.0214, "step": 32160 }, { "action_loss": 0.013582897372543812, "epoch": 28.92086330935252, "step": 32160 }, { "epoch": 28.92086330935252, "step": 32160, "torque_loss": 0.1613825559616089 }, { "epoch": 28.929856115107913, "grad_norm": 0.1501222550868988, "learning_rate": 3.092236414681588e-05, "loss": 0.0235, "step": 32170 }, { "action_loss": 0.0019848928786814213, "epoch": 28.929856115107913, "step": 32170 }, { "epoch": 28.929856115107913, "step": 32170, "torque_loss": 0.1389930248260498 }, { "epoch": 28.93884892086331, "grad_norm": 0.09440736472606659, "learning_rate": 3.089180072169431e-05, "loss": 0.0186, "step": 32180 }, { "action_loss": 0.0007443958893418312, "epoch": 28.93884892086331, "step": 32180 }, { "epoch": 28.93884892086331, "step": 32180, "torque_loss": 0.08345947414636612 }, { "epoch": 28.947841726618705, "grad_norm": 0.12977519631385803, "learning_rate": 3.08612456551483e-05, "loss": 0.0205, "step": 32190 }, { "action_loss": 0.01674823649227619, "epoch": 28.947841726618705, "step": 32190 }, { "epoch": 28.947841726618705, "step": 32190, "torque_loss": 0.2047959715127945 }, { "epoch": 28.9568345323741, "grad_norm": 0.09737298637628555, "learning_rate": 3.083069896054363e-05, "loss": 0.0189, "step": 32200 }, { "action_loss": 0.003066855249926448, "epoch": 28.9568345323741, "step": 32200 }, { "epoch": 28.9568345323741, "step": 32200, "torque_loss": 0.16558003425598145 }, { "epoch": 28.965827338129497, "grad_norm": 0.08448009938001633, "learning_rate": 3.080016065124253e-05, "loss": 0.0223, "step": 32210 }, { "action_loss": 0.002165993442758918, "epoch": 28.965827338129497, "step": 32210 }, { "epoch": 28.965827338129497, "step": 32210, "torque_loss": 0.13518662750720978 }, { "epoch": 28.97482014388489, "grad_norm": 0.14776700735092163, "learning_rate": 3.076963074060344e-05, "loss": 0.0201, "step": 32220 }, { "action_loss": 0.0025829970836639404, "epoch": 28.97482014388489, "step": 32220 }, { "epoch": 28.97482014388489, "step": 32220, "torque_loss": 0.12155959755182266 }, { "epoch": 28.98381294964029, "grad_norm": 0.15462705492973328, "learning_rate": 3.07391092419812e-05, "loss": 0.0196, "step": 32230 }, { "action_loss": 0.004213893320411444, "epoch": 28.98381294964029, "step": 32230 }, { "epoch": 28.98381294964029, "step": 32230, "torque_loss": 0.15966087579727173 }, { "epoch": 28.992805755395683, "grad_norm": 0.07563593983650208, "learning_rate": 3.070859616872695e-05, "loss": 0.0225, "step": 32240 }, { "action_loss": 0.0012414386728778481, "epoch": 28.992805755395683, "step": 32240 }, { "epoch": 28.992805755395683, "step": 32240, "torque_loss": 0.14162355661392212 }, { "epoch": 29.00179856115108, "grad_norm": 0.10727914422750473, "learning_rate": 3.067809153418816e-05, "loss": 0.0199, "step": 32250 }, { "action_loss": 0.0007799608283676207, "epoch": 29.00179856115108, "step": 32250 }, { "epoch": 29.00179856115108, "step": 32250, "torque_loss": 0.10433361679315567 }, { "epoch": 29.010791366906474, "grad_norm": 0.10342073440551758, "learning_rate": 3.0647595351708545e-05, "loss": 0.021, "step": 32260 }, { "action_loss": 0.005120972171425819, "epoch": 29.010791366906474, "step": 32260 }, { "epoch": 29.010791366906474, "step": 32260, "torque_loss": 0.1504223495721817 }, { "epoch": 29.019784172661872, "grad_norm": 0.07757530361413956, "learning_rate": 3.061710763462824e-05, "loss": 0.0198, "step": 32270 }, { "action_loss": 0.001555678783915937, "epoch": 29.019784172661872, "step": 32270 }, { "epoch": 29.019784172661872, "step": 32270, "torque_loss": 0.13847623765468597 }, { "epoch": 29.028776978417266, "grad_norm": 0.1023135557770729, "learning_rate": 3.0586628396283564e-05, "loss": 0.0231, "step": 32280 }, { "action_loss": 0.0013907141983509064, "epoch": 29.028776978417266, "step": 32280 }, { "epoch": 29.028776978417266, "step": 32280, "torque_loss": 0.15575478971004486 }, { "epoch": 29.037769784172664, "grad_norm": 0.05776931345462799, "learning_rate": 3.0556157650007174e-05, "loss": 0.0208, "step": 32290 }, { "action_loss": 0.0024874273221939802, "epoch": 29.037769784172664, "step": 32290 }, { "epoch": 29.037769784172664, "step": 32290, "torque_loss": 0.11904299259185791 }, { "epoch": 29.046762589928058, "grad_norm": 0.10125965625047684, "learning_rate": 3.052569540912801e-05, "loss": 0.0186, "step": 32300 }, { "action_loss": 0.003214112250134349, "epoch": 29.046762589928058, "step": 32300 }, { "epoch": 29.046762589928058, "step": 32300, "torque_loss": 0.11788830161094666 }, { "epoch": 29.055755395683452, "grad_norm": 0.09393616020679474, "learning_rate": 3.049524168697131e-05, "loss": 0.0271, "step": 32310 }, { "action_loss": 0.002415404887869954, "epoch": 29.055755395683452, "step": 32310 }, { "epoch": 29.055755395683452, "step": 32310, "torque_loss": 0.15493616461753845 }, { "epoch": 29.06474820143885, "grad_norm": 0.06855583190917969, "learning_rate": 3.0464796496858538e-05, "loss": 0.0206, "step": 32320 }, { "action_loss": 0.007415202911943197, "epoch": 29.06474820143885, "step": 32320 }, { "epoch": 29.06474820143885, "step": 32320, "torque_loss": 0.17279426753520966 }, { "epoch": 29.073741007194243, "grad_norm": 0.11001640558242798, "learning_rate": 3.0434359852107473e-05, "loss": 0.0184, "step": 32330 }, { "action_loss": 0.005415622610598803, "epoch": 29.073741007194243, "step": 32330 }, { "epoch": 29.073741007194243, "step": 32330, "torque_loss": 0.16044674813747406 }, { "epoch": 29.08273381294964, "grad_norm": 0.0731629803776741, "learning_rate": 3.0403931766032135e-05, "loss": 0.0206, "step": 32340 }, { "action_loss": 0.007014767732471228, "epoch": 29.08273381294964, "step": 32340 }, { "epoch": 29.08273381294964, "step": 32340, "torque_loss": 0.1984502077102661 }, { "epoch": 29.091726618705035, "grad_norm": 0.10585106909275055, "learning_rate": 3.0373512251942814e-05, "loss": 0.0281, "step": 32350 }, { "action_loss": 0.0014418052742257714, "epoch": 29.091726618705035, "step": 32350 }, { "epoch": 29.091726618705035, "step": 32350, "torque_loss": 0.13658009469509125 }, { "epoch": 29.100719424460433, "grad_norm": 0.08495144546031952, "learning_rate": 3.0343101323146007e-05, "loss": 0.0204, "step": 32360 }, { "action_loss": 0.004114379640668631, "epoch": 29.100719424460433, "step": 32360 }, { "epoch": 29.100719424460433, "step": 32360, "torque_loss": 0.17303548753261566 }, { "epoch": 29.109712230215827, "grad_norm": 0.17426933348178864, "learning_rate": 3.0312698992944523e-05, "loss": 0.0234, "step": 32370 }, { "action_loss": 0.0019718266557902098, "epoch": 29.109712230215827, "step": 32370 }, { "epoch": 29.109712230215827, "step": 32370, "torque_loss": 0.2073371410369873 }, { "epoch": 29.118705035971225, "grad_norm": 0.10669472813606262, "learning_rate": 3.028230527463736e-05, "loss": 0.0188, "step": 32380 }, { "action_loss": 0.003556271316483617, "epoch": 29.118705035971225, "step": 32380 }, { "epoch": 29.118705035971225, "step": 32380, "torque_loss": 0.205423042178154 }, { "epoch": 29.12769784172662, "grad_norm": 0.10277862846851349, "learning_rate": 3.0251920181519746e-05, "loss": 0.0184, "step": 32390 }, { "action_loss": 0.007585123647004366, "epoch": 29.12769784172662, "step": 32390 }, { "epoch": 29.12769784172662, "step": 32390, "torque_loss": 0.181723952293396 }, { "epoch": 29.136690647482013, "grad_norm": 0.071595698595047, "learning_rate": 3.0221543726883183e-05, "loss": 0.0255, "step": 32400 }, { "action_loss": 0.002286534057930112, "epoch": 29.136690647482013, "step": 32400 }, { "epoch": 29.136690647482013, "step": 32400, "torque_loss": 0.17180819809436798 }, { "epoch": 29.14568345323741, "grad_norm": 0.09397833794355392, "learning_rate": 3.019117592401535e-05, "loss": 0.0179, "step": 32410 }, { "action_loss": 0.0014754043659195304, "epoch": 29.14568345323741, "step": 32410 }, { "epoch": 29.14568345323741, "step": 32410, "torque_loss": 0.11980164796113968 }, { "epoch": 29.154676258992804, "grad_norm": 0.09027447551488876, "learning_rate": 3.0160816786200137e-05, "loss": 0.0174, "step": 32420 }, { "action_loss": 0.0022771544754505157, "epoch": 29.154676258992804, "step": 32420 }, { "epoch": 29.154676258992804, "step": 32420, "torque_loss": 0.09799134731292725 }, { "epoch": 29.163669064748202, "grad_norm": 0.13370440900325775, "learning_rate": 3.01304663267177e-05, "loss": 0.0187, "step": 32430 }, { "action_loss": 0.007694751489907503, "epoch": 29.163669064748202, "step": 32430 }, { "epoch": 29.163669064748202, "step": 32430, "torque_loss": 0.33179059624671936 }, { "epoch": 29.172661870503596, "grad_norm": 0.10094071924686432, "learning_rate": 3.0100124558844338e-05, "loss": 0.0224, "step": 32440 }, { "action_loss": 0.0028460777830332518, "epoch": 29.172661870503596, "step": 32440 }, { "epoch": 29.172661870503596, "step": 32440, "torque_loss": 0.11028891801834106 }, { "epoch": 29.181654676258994, "grad_norm": 0.08496449887752533, "learning_rate": 3.0069791495852562e-05, "loss": 0.0195, "step": 32450 }, { "action_loss": 0.005375606473535299, "epoch": 29.181654676258994, "step": 32450 }, { "epoch": 29.181654676258994, "step": 32450, "torque_loss": 0.17082905769348145 }, { "epoch": 29.190647482014388, "grad_norm": 0.1428866982460022, "learning_rate": 3.0039467151011113e-05, "loss": 0.0245, "step": 32460 }, { "action_loss": 0.007459214422851801, "epoch": 29.190647482014388, "step": 32460 }, { "epoch": 29.190647482014388, "step": 32460, "torque_loss": 0.13349147140979767 }, { "epoch": 29.199640287769785, "grad_norm": 0.14222104847431183, "learning_rate": 3.000915153758487e-05, "loss": 0.0209, "step": 32470 }, { "action_loss": 0.0035169136244803667, "epoch": 29.199640287769785, "step": 32470 }, { "epoch": 29.199640287769785, "step": 32470, "torque_loss": 0.18279998004436493 }, { "epoch": 29.20863309352518, "grad_norm": 0.09663254767656326, "learning_rate": 2.9978844668834917e-05, "loss": 0.0238, "step": 32480 }, { "action_loss": 0.004967643413692713, "epoch": 29.20863309352518, "step": 32480 }, { "epoch": 29.20863309352518, "step": 32480, "torque_loss": 0.1658596247434616 }, { "epoch": 29.217625899280577, "grad_norm": 0.12206155806779861, "learning_rate": 2.994854655801849e-05, "loss": 0.0219, "step": 32490 }, { "action_loss": 0.002261396963149309, "epoch": 29.217625899280577, "step": 32490 }, { "epoch": 29.217625899280577, "step": 32490, "torque_loss": 0.11570433527231216 }, { "epoch": 29.22661870503597, "grad_norm": 0.09513985365629196, "learning_rate": 2.991825721838905e-05, "loss": 0.0176, "step": 32500 }, { "action_loss": 0.008077294565737247, "epoch": 29.22661870503597, "step": 32500 }, { "epoch": 29.22661870503597, "step": 32500, "torque_loss": 0.2739437520503998 }, { "epoch": 29.235611510791365, "grad_norm": 0.09817604720592499, "learning_rate": 2.9887976663196166e-05, "loss": 0.0207, "step": 32510 }, { "action_loss": 0.0041628205217421055, "epoch": 29.235611510791365, "step": 32510 }, { "epoch": 29.235611510791365, "step": 32510, "torque_loss": 0.20030204951763153 }, { "epoch": 29.244604316546763, "grad_norm": 0.10451491177082062, "learning_rate": 2.985770490568556e-05, "loss": 0.0221, "step": 32520 }, { "action_loss": 0.004692921414971352, "epoch": 29.244604316546763, "step": 32520 }, { "epoch": 29.244604316546763, "step": 32520, "torque_loss": 0.18791024386882782 }, { "epoch": 29.253597122302157, "grad_norm": 0.14771102368831635, "learning_rate": 2.9827441959099157e-05, "loss": 0.023, "step": 32530 }, { "action_loss": 0.007193736732006073, "epoch": 29.253597122302157, "step": 32530 }, { "epoch": 29.253597122302157, "step": 32530, "torque_loss": 0.1592717170715332 }, { "epoch": 29.262589928057555, "grad_norm": 0.0856170505285263, "learning_rate": 2.979718783667499e-05, "loss": 0.0194, "step": 32540 }, { "action_loss": 0.0010500472271814942, "epoch": 29.262589928057555, "step": 32540 }, { "epoch": 29.262589928057555, "step": 32540, "torque_loss": 0.10309294611215591 }, { "epoch": 29.27158273381295, "grad_norm": 0.07473571598529816, "learning_rate": 2.9766942551647215e-05, "loss": 0.0182, "step": 32550 }, { "action_loss": 0.006443655584007502, "epoch": 29.27158273381295, "step": 32550 }, { "epoch": 29.27158273381295, "step": 32550, "torque_loss": 0.1327536553144455 }, { "epoch": 29.280575539568346, "grad_norm": 0.15634630620479584, "learning_rate": 2.9736706117246188e-05, "loss": 0.0194, "step": 32560 }, { "action_loss": 0.007348746061325073, "epoch": 29.280575539568346, "step": 32560 }, { "epoch": 29.280575539568346, "step": 32560, "torque_loss": 0.2023712545633316 }, { "epoch": 29.28956834532374, "grad_norm": 0.1172395721077919, "learning_rate": 2.9706478546698324e-05, "loss": 0.0255, "step": 32570 }, { "action_loss": 0.0033929653000086546, "epoch": 29.28956834532374, "step": 32570 }, { "epoch": 29.28956834532374, "step": 32570, "torque_loss": 0.17285166680812836 }, { "epoch": 29.298561151079138, "grad_norm": 0.07643081992864609, "learning_rate": 2.9676259853226195e-05, "loss": 0.0178, "step": 32580 }, { "action_loss": 0.004020860884338617, "epoch": 29.298561151079138, "step": 32580 }, { "epoch": 29.298561151079138, "step": 32580, "torque_loss": 0.2086765021085739 }, { "epoch": 29.307553956834532, "grad_norm": 0.11604268103837967, "learning_rate": 2.964605005004849e-05, "loss": 0.0202, "step": 32590 }, { "action_loss": 0.0013401126489043236, "epoch": 29.307553956834532, "step": 32590 }, { "epoch": 29.307553956834532, "step": 32590, "torque_loss": 0.1493382304906845 }, { "epoch": 29.31654676258993, "grad_norm": 0.1184636577963829, "learning_rate": 2.9615849150380016e-05, "loss": 0.0212, "step": 32600 }, { "action_loss": 0.0013551292940974236, "epoch": 29.31654676258993, "step": 32600 }, { "epoch": 29.31654676258993, "step": 32600, "torque_loss": 0.0965518057346344 }, { "epoch": 29.325539568345324, "grad_norm": 0.18724480271339417, "learning_rate": 2.9585657167431637e-05, "loss": 0.0226, "step": 32610 }, { "action_loss": 0.0012842287542298436, "epoch": 29.325539568345324, "step": 32610 }, { "epoch": 29.325539568345324, "step": 32610, "torque_loss": 0.09295233339071274 }, { "epoch": 29.334532374100718, "grad_norm": 0.15558847784996033, "learning_rate": 2.9555474114410388e-05, "loss": 0.019, "step": 32620 }, { "action_loss": 0.0031334739178419113, "epoch": 29.334532374100718, "step": 32620 }, { "epoch": 29.334532374100718, "step": 32620, "torque_loss": 0.14667890965938568 }, { "epoch": 29.343525179856115, "grad_norm": 0.1708192676305771, "learning_rate": 2.9525300004519348e-05, "loss": 0.0206, "step": 32630 }, { "action_loss": 0.003243977902457118, "epoch": 29.343525179856115, "step": 32630 }, { "epoch": 29.343525179856115, "step": 32630, "torque_loss": 0.17801521718502045 }, { "epoch": 29.35251798561151, "grad_norm": 0.11398147791624069, "learning_rate": 2.9495134850957696e-05, "loss": 0.0176, "step": 32640 }, { "action_loss": 0.002649519592523575, "epoch": 29.35251798561151, "step": 32640 }, { "epoch": 29.35251798561151, "step": 32640, "torque_loss": 0.16649363934993744 }, { "epoch": 29.361510791366907, "grad_norm": 0.07478804886341095, "learning_rate": 2.9464978666920674e-05, "loss": 0.0193, "step": 32650 }, { "action_loss": 0.001402780064381659, "epoch": 29.361510791366907, "step": 32650 }, { "epoch": 29.361510791366907, "step": 32650, "torque_loss": 0.1016933023929596 }, { "epoch": 29.3705035971223, "grad_norm": 0.08095639944076538, "learning_rate": 2.9434831465599662e-05, "loss": 0.023, "step": 32660 }, { "action_loss": 0.013748443685472012, "epoch": 29.3705035971223, "step": 32660 }, { "epoch": 29.3705035971223, "step": 32660, "torque_loss": 0.1780717521905899 }, { "epoch": 29.3794964028777, "grad_norm": 0.08225543797016144, "learning_rate": 2.9404693260182047e-05, "loss": 0.0231, "step": 32670 }, { "action_loss": 0.0034840998705476522, "epoch": 29.3794964028777, "step": 32670 }, { "epoch": 29.3794964028777, "step": 32670, "torque_loss": 0.15794174373149872 }, { "epoch": 29.388489208633093, "grad_norm": 0.044121965765953064, "learning_rate": 2.9374564063851295e-05, "loss": 0.0178, "step": 32680 }, { "action_loss": 0.002138974377885461, "epoch": 29.388489208633093, "step": 32680 }, { "epoch": 29.388489208633093, "step": 32680, "torque_loss": 0.15739266574382782 }, { "epoch": 29.39748201438849, "grad_norm": 0.10166259855031967, "learning_rate": 2.9344443889786967e-05, "loss": 0.0168, "step": 32690 }, { "action_loss": 0.02320968173444271, "epoch": 29.39748201438849, "step": 32690 }, { "epoch": 29.39748201438849, "step": 32690, "torque_loss": 0.12846624851226807 }, { "epoch": 29.406474820143885, "grad_norm": 0.07982988655567169, "learning_rate": 2.931433275116463e-05, "loss": 0.0218, "step": 32700 }, { "action_loss": 0.014276781119406223, "epoch": 29.406474820143885, "step": 32700 }, { "epoch": 29.406474820143885, "step": 32700, "torque_loss": 0.17282706499099731 }, { "epoch": 29.415467625899282, "grad_norm": 0.12348821759223938, "learning_rate": 2.928423066115592e-05, "loss": 0.0195, "step": 32710 }, { "action_loss": 0.009239115752279758, "epoch": 29.415467625899282, "step": 32710 }, { "epoch": 29.415467625899282, "step": 32710, "torque_loss": 0.14344088733196259 }, { "epoch": 29.424460431654676, "grad_norm": 0.10468842834234238, "learning_rate": 2.9254137632928537e-05, "loss": 0.0231, "step": 32720 }, { "action_loss": 0.0024840908590704203, "epoch": 29.424460431654676, "step": 32720 }, { "epoch": 29.424460431654676, "step": 32720, "torque_loss": 0.18224655091762543 }, { "epoch": 29.43345323741007, "grad_norm": 0.15204884111881256, "learning_rate": 2.9224053679646167e-05, "loss": 0.0206, "step": 32730 }, { "action_loss": 0.002465967321768403, "epoch": 29.43345323741007, "step": 32730 }, { "epoch": 29.43345323741007, "step": 32730, "torque_loss": 0.15037675201892853 }, { "epoch": 29.442446043165468, "grad_norm": 0.09932439029216766, "learning_rate": 2.9193978814468547e-05, "loss": 0.0231, "step": 32740 }, { "action_loss": 0.003687510499730706, "epoch": 29.442446043165468, "step": 32740 }, { "epoch": 29.442446043165468, "step": 32740, "torque_loss": 0.12291783094406128 }, { "epoch": 29.451438848920862, "grad_norm": 0.21968024969100952, "learning_rate": 2.9163913050551485e-05, "loss": 0.0232, "step": 32750 }, { "action_loss": 0.003120528534054756, "epoch": 29.451438848920862, "step": 32750 }, { "epoch": 29.451438848920862, "step": 32750, "torque_loss": 0.11133842915296555 }, { "epoch": 29.46043165467626, "grad_norm": 0.12502114474773407, "learning_rate": 2.9133856401046735e-05, "loss": 0.0231, "step": 32760 }, { "action_loss": 0.0013453682186082006, "epoch": 29.46043165467626, "step": 32760 }, { "epoch": 29.46043165467626, "step": 32760, "torque_loss": 0.12620650231838226 }, { "epoch": 29.469424460431654, "grad_norm": 0.12524518370628357, "learning_rate": 2.910380887910209e-05, "loss": 0.0206, "step": 32770 }, { "action_loss": 0.003983702044934034, "epoch": 29.469424460431654, "step": 32770 }, { "epoch": 29.469424460431654, "step": 32770, "torque_loss": 0.14788605272769928 }, { "epoch": 29.47841726618705, "grad_norm": 0.12891528010368347, "learning_rate": 2.9073770497861386e-05, "loss": 0.0215, "step": 32780 }, { "action_loss": 0.008069592528045177, "epoch": 29.47841726618705, "step": 32780 }, { "epoch": 29.47841726618705, "step": 32780, "torque_loss": 0.24694009125232697 }, { "epoch": 29.487410071942445, "grad_norm": 0.15813350677490234, "learning_rate": 2.904374127046442e-05, "loss": 0.0234, "step": 32790 }, { "action_loss": 0.002809709869325161, "epoch": 29.487410071942445, "step": 32790 }, { "epoch": 29.487410071942445, "step": 32790, "torque_loss": 0.18726052343845367 }, { "epoch": 29.496402877697843, "grad_norm": 0.11121579259634018, "learning_rate": 2.9013721210046973e-05, "loss": 0.0199, "step": 32800 }, { "action_loss": 0.00394628057256341, "epoch": 29.496402877697843, "step": 32800 }, { "epoch": 29.496402877697843, "step": 32800, "torque_loss": 0.13594447076320648 }, { "epoch": 29.505395683453237, "grad_norm": 0.12713362276554108, "learning_rate": 2.8983710329740882e-05, "loss": 0.0203, "step": 32810 }, { "action_loss": 0.005276167765259743, "epoch": 29.505395683453237, "step": 32810 }, { "epoch": 29.505395683453237, "step": 32810, "torque_loss": 0.21277093887329102 }, { "epoch": 29.514388489208635, "grad_norm": 0.11035461723804474, "learning_rate": 2.8953708642673898e-05, "loss": 0.0225, "step": 32820 }, { "action_loss": 0.002452853135764599, "epoch": 29.514388489208635, "step": 32820 }, { "epoch": 29.514388489208635, "step": 32820, "torque_loss": 0.15263092517852783 }, { "epoch": 29.52338129496403, "grad_norm": 0.13299627602100372, "learning_rate": 2.8923716161969782e-05, "loss": 0.017, "step": 32830 }, { "action_loss": 0.0011883607367053628, "epoch": 29.52338129496403, "step": 32830 }, { "epoch": 29.52338129496403, "step": 32830, "torque_loss": 0.10445577651262283 }, { "epoch": 29.532374100719423, "grad_norm": 0.14399267733097076, "learning_rate": 2.8893732900748248e-05, "loss": 0.0194, "step": 32840 }, { "action_loss": 0.00438841013237834, "epoch": 29.532374100719423, "step": 32840 }, { "epoch": 29.532374100719423, "step": 32840, "torque_loss": 0.16822046041488647 }, { "epoch": 29.54136690647482, "grad_norm": 0.07396921515464783, "learning_rate": 2.8863758872125036e-05, "loss": 0.0192, "step": 32850 }, { "action_loss": 0.002539534354582429, "epoch": 29.54136690647482, "step": 32850 }, { "epoch": 29.54136690647482, "step": 32850, "torque_loss": 0.15771406888961792 }, { "epoch": 29.550359712230215, "grad_norm": 0.08415376394987106, "learning_rate": 2.883379408921175e-05, "loss": 0.018, "step": 32860 }, { "action_loss": 0.0012938237050548196, "epoch": 29.550359712230215, "step": 32860 }, { "epoch": 29.550359712230215, "step": 32860, "torque_loss": 0.10659661144018173 }, { "epoch": 29.559352517985612, "grad_norm": 0.06051163002848625, "learning_rate": 2.880383856511606e-05, "loss": 0.0193, "step": 32870 }, { "action_loss": 0.003010194981470704, "epoch": 29.559352517985612, "step": 32870 }, { "epoch": 29.559352517985612, "step": 32870, "torque_loss": 0.1773587465286255 }, { "epoch": 29.568345323741006, "grad_norm": 0.1507188230752945, "learning_rate": 2.8773892312941476e-05, "loss": 0.0218, "step": 32880 }, { "action_loss": 0.002917367732152343, "epoch": 29.568345323741006, "step": 32880 }, { "epoch": 29.568345323741006, "step": 32880, "torque_loss": 0.14958234131336212 }, { "epoch": 29.577338129496404, "grad_norm": 0.2045745849609375, "learning_rate": 2.8743955345787555e-05, "loss": 0.021, "step": 32890 }, { "action_loss": 0.004910058807581663, "epoch": 29.577338129496404, "step": 32890 }, { "epoch": 29.577338129496404, "step": 32890, "torque_loss": 0.19014815986156464 }, { "epoch": 29.586330935251798, "grad_norm": 0.10188382863998413, "learning_rate": 2.87140276767497e-05, "loss": 0.0175, "step": 32900 }, { "action_loss": 0.004029459785670042, "epoch": 29.586330935251798, "step": 32900 }, { "epoch": 29.586330935251798, "step": 32900, "torque_loss": 0.15749406814575195 }, { "epoch": 29.595323741007196, "grad_norm": 0.10423313826322556, "learning_rate": 2.868410931891934e-05, "loss": 0.0186, "step": 32910 }, { "action_loss": 0.0017243041656911373, "epoch": 29.595323741007196, "step": 32910 }, { "epoch": 29.595323741007196, "step": 32910, "torque_loss": 0.13707908987998962 }, { "epoch": 29.60431654676259, "grad_norm": 0.1208294928073883, "learning_rate": 2.8654200285383748e-05, "loss": 0.0218, "step": 32920 }, { "action_loss": 0.0024101317394524813, "epoch": 29.60431654676259, "step": 32920 }, { "epoch": 29.60431654676259, "step": 32920, "torque_loss": 0.13834606111049652 }, { "epoch": 29.613309352517987, "grad_norm": 0.07834932208061218, "learning_rate": 2.8624300589226126e-05, "loss": 0.0192, "step": 32930 }, { "action_loss": 0.0019124280661344528, "epoch": 29.613309352517987, "step": 32930 }, { "epoch": 29.613309352517987, "step": 32930, "torque_loss": 0.13302208483219147 }, { "epoch": 29.62230215827338, "grad_norm": 0.10141357779502869, "learning_rate": 2.859441024352568e-05, "loss": 0.0219, "step": 32940 }, { "action_loss": 0.0031148178968578577, "epoch": 29.62230215827338, "step": 32940 }, { "epoch": 29.62230215827338, "step": 32940, "torque_loss": 0.11587735265493393 }, { "epoch": 29.631294964028775, "grad_norm": 0.09118271619081497, "learning_rate": 2.8564529261357427e-05, "loss": 0.0213, "step": 32950 }, { "action_loss": 0.006421081256121397, "epoch": 29.631294964028775, "step": 32950 }, { "epoch": 29.631294964028775, "step": 32950, "torque_loss": 0.26957085728645325 }, { "epoch": 29.640287769784173, "grad_norm": 0.07335694879293442, "learning_rate": 2.8534657655792307e-05, "loss": 0.022, "step": 32960 }, { "action_loss": 0.003969993442296982, "epoch": 29.640287769784173, "step": 32960 }, { "epoch": 29.640287769784173, "step": 32960, "torque_loss": 0.23805873095989227 }, { "epoch": 29.649280575539567, "grad_norm": 0.12138570100069046, "learning_rate": 2.8504795439897208e-05, "loss": 0.0212, "step": 32970 }, { "action_loss": 0.010994494892656803, "epoch": 29.649280575539567, "step": 32970 }, { "epoch": 29.649280575539567, "step": 32970, "torque_loss": 0.1601657122373581 }, { "epoch": 29.658273381294965, "grad_norm": 0.1708589792251587, "learning_rate": 2.847494262673487e-05, "loss": 0.023, "step": 32980 }, { "action_loss": 0.0019740862771868706, "epoch": 29.658273381294965, "step": 32980 }, { "epoch": 29.658273381294965, "step": 32980, "torque_loss": 0.1644071340560913 }, { "epoch": 29.66726618705036, "grad_norm": 0.09818920493125916, "learning_rate": 2.8445099229363925e-05, "loss": 0.0203, "step": 32990 }, { "action_loss": 0.003441849024966359, "epoch": 29.66726618705036, "step": 32990 }, { "epoch": 29.66726618705036, "step": 32990, "torque_loss": 0.19568109512329102 }, { "epoch": 29.676258992805757, "grad_norm": 0.15913470089435577, "learning_rate": 2.841526526083886e-05, "loss": 0.0189, "step": 33000 }, { "action_loss": 0.0022762783337384462, "epoch": 29.676258992805757, "step": 33000 }, { "epoch": 29.676258992805757, "step": 33000, "torque_loss": 0.15563859045505524 }, { "epoch": 29.68525179856115, "grad_norm": 0.0897207111120224, "learning_rate": 2.838544073421011e-05, "loss": 0.018, "step": 33010 }, { "action_loss": 0.0017021667445078492, "epoch": 29.68525179856115, "step": 33010 }, { "epoch": 29.68525179856115, "step": 33010, "torque_loss": 0.12828175723552704 }, { "epoch": 29.694244604316548, "grad_norm": 0.10953191667795181, "learning_rate": 2.8355625662523928e-05, "loss": 0.0202, "step": 33020 }, { "action_loss": 0.008038862608373165, "epoch": 29.694244604316548, "step": 33020 }, { "epoch": 29.694244604316548, "step": 33020, "torque_loss": 0.16764886677265167 }, { "epoch": 29.703237410071942, "grad_norm": 0.09561242163181305, "learning_rate": 2.8325820058822407e-05, "loss": 0.0197, "step": 33030 }, { "action_loss": 0.008803114295005798, "epoch": 29.703237410071942, "step": 33030 }, { "epoch": 29.703237410071942, "step": 33030, "torque_loss": 0.2446274310350418 }, { "epoch": 29.71223021582734, "grad_norm": 0.11602343618869781, "learning_rate": 2.829602393614357e-05, "loss": 0.0203, "step": 33040 }, { "action_loss": 0.00267770211212337, "epoch": 29.71223021582734, "step": 33040 }, { "epoch": 29.71223021582734, "step": 33040, "torque_loss": 0.1425379365682602 }, { "epoch": 29.721223021582734, "grad_norm": 0.09162773936986923, "learning_rate": 2.8266237307521236e-05, "loss": 0.0229, "step": 33050 }, { "action_loss": 0.0015087742358446121, "epoch": 29.721223021582734, "step": 33050 }, { "epoch": 29.721223021582734, "step": 33050, "torque_loss": 0.14593161642551422 }, { "epoch": 29.730215827338128, "grad_norm": 0.09409111738204956, "learning_rate": 2.8236460185985074e-05, "loss": 0.0164, "step": 33060 }, { "action_loss": 0.002251578262075782, "epoch": 29.730215827338128, "step": 33060 }, { "epoch": 29.730215827338128, "step": 33060, "torque_loss": 0.11582118272781372 }, { "epoch": 29.739208633093526, "grad_norm": 0.1917630434036255, "learning_rate": 2.8206692584560646e-05, "loss": 0.0203, "step": 33070 }, { "action_loss": 0.0022316565737128258, "epoch": 29.739208633093526, "step": 33070 }, { "epoch": 29.739208633093526, "step": 33070, "torque_loss": 0.13140319287776947 }, { "epoch": 29.74820143884892, "grad_norm": 0.09761732816696167, "learning_rate": 2.8176934516269283e-05, "loss": 0.0191, "step": 33080 }, { "action_loss": 0.0012494641123339534, "epoch": 29.74820143884892, "step": 33080 }, { "epoch": 29.74820143884892, "step": 33080, "torque_loss": 0.13107044994831085 }, { "epoch": 29.757194244604317, "grad_norm": 0.09114094823598862, "learning_rate": 2.8147185994128167e-05, "loss": 0.024, "step": 33090 }, { "action_loss": 0.0017028320580720901, "epoch": 29.757194244604317, "step": 33090 }, { "epoch": 29.757194244604317, "step": 33090, "torque_loss": 0.1393912434577942 }, { "epoch": 29.76618705035971, "grad_norm": 0.11633317917585373, "learning_rate": 2.8117447031150346e-05, "loss": 0.0186, "step": 33100 }, { "action_loss": 0.0012363591231405735, "epoch": 29.76618705035971, "step": 33100 }, { "epoch": 29.76618705035971, "step": 33100, "torque_loss": 0.10289403796195984 }, { "epoch": 29.77517985611511, "grad_norm": 0.16955967247486115, "learning_rate": 2.8087717640344618e-05, "loss": 0.0202, "step": 33110 }, { "action_loss": 0.005187015514820814, "epoch": 29.77517985611511, "step": 33110 }, { "epoch": 29.77517985611511, "step": 33110, "torque_loss": 0.1542149931192398 }, { "epoch": 29.784172661870503, "grad_norm": 0.15285900235176086, "learning_rate": 2.805799783471566e-05, "loss": 0.0182, "step": 33120 }, { "action_loss": 0.003585250349715352, "epoch": 29.784172661870503, "step": 33120 }, { "epoch": 29.784172661870503, "step": 33120, "torque_loss": 0.1899162083864212 }, { "epoch": 29.7931654676259, "grad_norm": 0.1430606096982956, "learning_rate": 2.8028287627263895e-05, "loss": 0.0195, "step": 33130 }, { "action_loss": 0.004816384520381689, "epoch": 29.7931654676259, "step": 33130 }, { "epoch": 29.7931654676259, "step": 33130, "torque_loss": 0.16467206180095673 }, { "epoch": 29.802158273381295, "grad_norm": 0.11108581721782684, "learning_rate": 2.7998587030985608e-05, "loss": 0.0208, "step": 33140 }, { "action_loss": 0.0021806098520755768, "epoch": 29.802158273381295, "step": 33140 }, { "epoch": 29.802158273381295, "step": 33140, "torque_loss": 0.17641842365264893 }, { "epoch": 29.81115107913669, "grad_norm": 0.10165586322546005, "learning_rate": 2.7968896058872806e-05, "loss": 0.0219, "step": 33150 }, { "action_loss": 0.0012387675233185291, "epoch": 29.81115107913669, "step": 33150 }, { "epoch": 29.81115107913669, "step": 33150, "torque_loss": 0.15540367364883423 }, { "epoch": 29.820143884892087, "grad_norm": 0.10510080307722092, "learning_rate": 2.7939214723913388e-05, "loss": 0.0213, "step": 33160 }, { "action_loss": 0.008400652557611465, "epoch": 29.820143884892087, "step": 33160 }, { "epoch": 29.820143884892087, "step": 33160, "torque_loss": 0.20774690806865692 }, { "epoch": 29.82913669064748, "grad_norm": 0.14816568791866302, "learning_rate": 2.7909543039090935e-05, "loss": 0.0226, "step": 33170 }, { "action_loss": 0.001780212507583201, "epoch": 29.82913669064748, "step": 33170 }, { "epoch": 29.82913669064748, "step": 33170, "torque_loss": 0.10752472281455994 }, { "epoch": 29.83812949640288, "grad_norm": 0.14909568428993225, "learning_rate": 2.787988101738487e-05, "loss": 0.0181, "step": 33180 }, { "action_loss": 0.0016297047259286046, "epoch": 29.83812949640288, "step": 33180 }, { "epoch": 29.83812949640288, "step": 33180, "torque_loss": 0.1621939092874527 }, { "epoch": 29.847122302158272, "grad_norm": 0.1632760465145111, "learning_rate": 2.7850228671770334e-05, "loss": 0.0221, "step": 33190 }, { "action_loss": 0.002422734396532178, "epoch": 29.847122302158272, "step": 33190 }, { "epoch": 29.847122302158272, "step": 33190, "torque_loss": 0.19653242826461792 }, { "epoch": 29.85611510791367, "grad_norm": 0.12774218618869781, "learning_rate": 2.782058601521832e-05, "loss": 0.0186, "step": 33200 }, { "action_loss": 0.002084554173052311, "epoch": 29.85611510791367, "step": 33200 }, { "epoch": 29.85611510791367, "step": 33200, "torque_loss": 0.16613174974918365 }, { "epoch": 29.865107913669064, "grad_norm": 0.10930493474006653, "learning_rate": 2.77909530606955e-05, "loss": 0.0215, "step": 33210 }, { "action_loss": 0.001908290316350758, "epoch": 29.865107913669064, "step": 33210 }, { "epoch": 29.865107913669064, "step": 33210, "torque_loss": 0.14206081628799438 }, { "epoch": 29.87410071942446, "grad_norm": 0.07270121574401855, "learning_rate": 2.7761329821164335e-05, "loss": 0.0208, "step": 33220 }, { "action_loss": 0.002216146793216467, "epoch": 29.87410071942446, "step": 33220 }, { "epoch": 29.87410071942446, "step": 33220, "torque_loss": 0.18850164115428925 }, { "epoch": 29.883093525179856, "grad_norm": 0.169539675116539, "learning_rate": 2.7731716309583062e-05, "loss": 0.0212, "step": 33230 }, { "action_loss": 0.009325223043560982, "epoch": 29.883093525179856, "step": 33230 }, { "epoch": 29.883093525179856, "step": 33230, "torque_loss": 0.21138782799243927 }, { "epoch": 29.892086330935253, "grad_norm": 0.1449701488018036, "learning_rate": 2.7702112538905624e-05, "loss": 0.023, "step": 33240 }, { "action_loss": 0.004166687373071909, "epoch": 29.892086330935253, "step": 33240 }, { "epoch": 29.892086330935253, "step": 33240, "torque_loss": 0.1679486483335495 }, { "epoch": 29.901079136690647, "grad_norm": 0.10132138431072235, "learning_rate": 2.7672518522081682e-05, "loss": 0.0211, "step": 33250 }, { "action_loss": 0.010995694436132908, "epoch": 29.901079136690647, "step": 33250 }, { "epoch": 29.901079136690647, "step": 33250, "torque_loss": 0.17369444668293 }, { "epoch": 29.91007194244604, "grad_norm": 0.10514941811561584, "learning_rate": 2.7642934272056726e-05, "loss": 0.0187, "step": 33260 }, { "action_loss": 0.01647227443754673, "epoch": 29.91007194244604, "step": 33260 }, { "epoch": 29.91007194244604, "step": 33260, "torque_loss": 0.1619759351015091 }, { "epoch": 29.91906474820144, "grad_norm": 0.05109230801463127, "learning_rate": 2.7613359801771877e-05, "loss": 0.0207, "step": 33270 }, { "action_loss": 0.0015528799267485738, "epoch": 29.91906474820144, "step": 33270 }, { "epoch": 29.91906474820144, "step": 33270, "torque_loss": 0.11957808583974838 }, { "epoch": 29.928057553956833, "grad_norm": 0.08432646095752716, "learning_rate": 2.7583795124163996e-05, "loss": 0.0187, "step": 33280 }, { "action_loss": 0.005499735940247774, "epoch": 29.928057553956833, "step": 33280 }, { "epoch": 29.928057553956833, "step": 33280, "torque_loss": 0.20017175376415253 }, { "epoch": 29.93705035971223, "grad_norm": 0.10528919845819473, "learning_rate": 2.755424025216572e-05, "loss": 0.0204, "step": 33290 }, { "action_loss": 0.0014203698374330997, "epoch": 29.93705035971223, "step": 33290 }, { "epoch": 29.93705035971223, "step": 33290, "torque_loss": 0.16313259303569794 }, { "epoch": 29.946043165467625, "grad_norm": 0.10986977070569992, "learning_rate": 2.7524695198705334e-05, "loss": 0.0239, "step": 33300 }, { "action_loss": 0.008802227675914764, "epoch": 29.946043165467625, "step": 33300 }, { "epoch": 29.946043165467625, "step": 33300, "torque_loss": 0.18967382609844208 }, { "epoch": 29.955035971223023, "grad_norm": 0.11494746059179306, "learning_rate": 2.749515997670683e-05, "loss": 0.0221, "step": 33310 }, { "action_loss": 0.00174515915568918, "epoch": 29.955035971223023, "step": 33310 }, { "epoch": 29.955035971223023, "step": 33310, "torque_loss": 0.0869208350777626 }, { "epoch": 29.964028776978417, "grad_norm": 0.1639346480369568, "learning_rate": 2.746563459908995e-05, "loss": 0.0164, "step": 33320 }, { "action_loss": 0.006910183932632208, "epoch": 29.964028776978417, "step": 33320 }, { "epoch": 29.964028776978417, "step": 33320, "torque_loss": 0.14043770730495453 }, { "epoch": 29.973021582733814, "grad_norm": 0.06526437401771545, "learning_rate": 2.7436119078770085e-05, "loss": 0.0201, "step": 33330 }, { "action_loss": 0.005844520404934883, "epoch": 29.973021582733814, "step": 33330 }, { "epoch": 29.973021582733814, "step": 33330, "torque_loss": 0.21030592918395996 }, { "epoch": 29.98201438848921, "grad_norm": 0.1594824194908142, "learning_rate": 2.74066134286583e-05, "loss": 0.0209, "step": 33340 }, { "action_loss": 0.001712875789962709, "epoch": 29.98201438848921, "step": 33340 }, { "epoch": 29.98201438848921, "step": 33340, "torque_loss": 0.16559571027755737 }, { "epoch": 29.991007194244606, "grad_norm": 0.10976922512054443, "learning_rate": 2.7377117661661412e-05, "loss": 0.0208, "step": 33350 }, { "action_loss": 0.003720579668879509, "epoch": 29.991007194244606, "step": 33350 }, { "epoch": 29.991007194244606, "step": 33350, "torque_loss": 0.14863690733909607 }, { "epoch": 30.0, "grad_norm": 0.08077903836965561, "learning_rate": 2.734763179068183e-05, "loss": 0.0179, "step": 33360 }, { "action_loss": 0.04700217768549919, "epoch": 30.0, "step": 33360 }, { "epoch": 30.0, "step": 33360, "torque_loss": 0.2823943793773651 }, { "epoch": 30.008992805755394, "grad_norm": 0.09225159138441086, "learning_rate": 2.7318155828617718e-05, "loss": 0.0236, "step": 33370 }, { "action_loss": 0.020236531272530556, "epoch": 30.008992805755394, "step": 33370 }, { "epoch": 30.008992805755394, "step": 33370, "torque_loss": 0.27664491534233093 }, { "epoch": 30.01798561151079, "grad_norm": 0.1380845159292221, "learning_rate": 2.7288689788362832e-05, "loss": 0.0218, "step": 33380 }, { "action_loss": 0.005405212286859751, "epoch": 30.01798561151079, "step": 33380 }, { "epoch": 30.01798561151079, "step": 33380, "torque_loss": 0.188145712018013 }, { "epoch": 30.026978417266186, "grad_norm": 0.09748855233192444, "learning_rate": 2.725923368280664e-05, "loss": 0.0203, "step": 33390 }, { "action_loss": 0.0011589544592425227, "epoch": 30.026978417266186, "step": 33390 }, { "epoch": 30.026978417266186, "step": 33390, "torque_loss": 0.13116352260112762 }, { "epoch": 30.035971223021583, "grad_norm": 0.17581795156002045, "learning_rate": 2.7229787524834237e-05, "loss": 0.0233, "step": 33400 }, { "action_loss": 0.00831079762428999, "epoch": 30.035971223021583, "step": 33400 }, { "epoch": 30.035971223021583, "step": 33400, "torque_loss": 0.17700912058353424 }, { "epoch": 30.044964028776977, "grad_norm": 0.09665466099977493, "learning_rate": 2.7200351327326345e-05, "loss": 0.0191, "step": 33410 }, { "action_loss": 0.0011187291238456964, "epoch": 30.044964028776977, "step": 33410 }, { "epoch": 30.044964028776977, "step": 33410, "torque_loss": 0.1405772715806961 }, { "epoch": 30.053956834532375, "grad_norm": 0.10146012902259827, "learning_rate": 2.7170925103159407e-05, "loss": 0.0176, "step": 33420 }, { "action_loss": 0.0035397671163082123, "epoch": 30.053956834532375, "step": 33420 }, { "epoch": 30.053956834532375, "step": 33420, "torque_loss": 0.19431078433990479 }, { "epoch": 30.06294964028777, "grad_norm": 0.1139940544962883, "learning_rate": 2.714150886520542e-05, "loss": 0.0165, "step": 33430 }, { "action_loss": 0.0015771797625347972, "epoch": 30.06294964028777, "step": 33430 }, { "epoch": 30.06294964028777, "step": 33430, "torque_loss": 0.12130958586931229 }, { "epoch": 30.071942446043167, "grad_norm": 0.11621126532554626, "learning_rate": 2.7112102626332038e-05, "loss": 0.0268, "step": 33440 }, { "action_loss": 0.0011244164779782295, "epoch": 30.071942446043167, "step": 33440 }, { "epoch": 30.071942446043167, "step": 33440, "torque_loss": 0.08166234940290451 }, { "epoch": 30.08093525179856, "grad_norm": 0.0740848034620285, "learning_rate": 2.7082706399402574e-05, "loss": 0.02, "step": 33450 }, { "action_loss": 0.0010566925629973412, "epoch": 30.08093525179856, "step": 33450 }, { "epoch": 30.08093525179856, "step": 33450, "torque_loss": 0.10865703970193863 }, { "epoch": 30.08992805755396, "grad_norm": 0.09320288896560669, "learning_rate": 2.7053320197275934e-05, "loss": 0.0172, "step": 33460 }, { "action_loss": 0.0032616329845041037, "epoch": 30.08992805755396, "step": 33460 }, { "epoch": 30.08992805755396, "step": 33460, "torque_loss": 0.13540780544281006 }, { "epoch": 30.098920863309353, "grad_norm": 0.08572821319103241, "learning_rate": 2.7023944032806614e-05, "loss": 0.0183, "step": 33470 }, { "action_loss": 0.0009642569930292666, "epoch": 30.098920863309353, "step": 33470 }, { "epoch": 30.098920863309353, "step": 33470, "torque_loss": 0.10577783733606339 }, { "epoch": 30.107913669064747, "grad_norm": 0.1051703467965126, "learning_rate": 2.699457791884478e-05, "loss": 0.0167, "step": 33480 }, { "action_loss": 0.005725198891013861, "epoch": 30.107913669064747, "step": 33480 }, { "epoch": 30.107913669064747, "step": 33480, "torque_loss": 0.18201954662799835 }, { "epoch": 30.116906474820144, "grad_norm": 0.09626594930887222, "learning_rate": 2.6965221868236155e-05, "loss": 0.0192, "step": 33490 }, { "action_loss": 0.002612986834719777, "epoch": 30.116906474820144, "step": 33490 }, { "epoch": 30.116906474820144, "step": 33490, "torque_loss": 0.15690913796424866 }, { "epoch": 30.12589928057554, "grad_norm": 0.10360583662986755, "learning_rate": 2.6935875893822077e-05, "loss": 0.0173, "step": 33500 }, { "action_loss": 0.009002628736197948, "epoch": 30.12589928057554, "step": 33500 }, { "epoch": 30.12589928057554, "step": 33500, "torque_loss": 0.17370611429214478 }, { "epoch": 30.134892086330936, "grad_norm": 0.10546093434095383, "learning_rate": 2.690654000843944e-05, "loss": 0.02, "step": 33510 }, { "action_loss": 0.014143005013465881, "epoch": 30.134892086330936, "step": 33510 }, { "epoch": 30.134892086330936, "step": 33510, "torque_loss": 0.21141456067562103 }, { "epoch": 30.14388489208633, "grad_norm": 0.09220226854085922, "learning_rate": 2.687721422492081e-05, "loss": 0.022, "step": 33520 }, { "action_loss": 0.0017459234222769737, "epoch": 30.14388489208633, "step": 33520 }, { "epoch": 30.14388489208633, "step": 33520, "torque_loss": 0.1486617475748062 }, { "epoch": 30.152877697841728, "grad_norm": 0.1703742891550064, "learning_rate": 2.6847898556094252e-05, "loss": 0.0253, "step": 33530 }, { "action_loss": 0.00806442927569151, "epoch": 30.152877697841728, "step": 33530 }, { "epoch": 30.152877697841728, "step": 33530, "torque_loss": 0.1560528725385666 }, { "epoch": 30.16187050359712, "grad_norm": 0.1014469563961029, "learning_rate": 2.6818593014783417e-05, "loss": 0.0186, "step": 33540 }, { "action_loss": 0.00162221968639642, "epoch": 30.16187050359712, "step": 33540 }, { "epoch": 30.16187050359712, "step": 33540, "torque_loss": 0.10522293299436569 }, { "epoch": 30.17086330935252, "grad_norm": 0.10283380001783371, "learning_rate": 2.6789297613807585e-05, "loss": 0.0179, "step": 33550 }, { "action_loss": 0.0024860159028321505, "epoch": 30.17086330935252, "step": 33550 }, { "epoch": 30.17086330935252, "step": 33550, "torque_loss": 0.12942767143249512 }, { "epoch": 30.179856115107913, "grad_norm": 0.09723129868507385, "learning_rate": 2.676001236598154e-05, "loss": 0.0205, "step": 33560 }, { "action_loss": 0.002922952175140381, "epoch": 30.179856115107913, "step": 33560 }, { "epoch": 30.179856115107913, "step": 33560, "torque_loss": 0.14944472908973694 }, { "epoch": 30.18884892086331, "grad_norm": 0.14774593710899353, "learning_rate": 2.6730737284115626e-05, "loss": 0.0207, "step": 33570 }, { "action_loss": 0.0020626625046133995, "epoch": 30.18884892086331, "step": 33570 }, { "epoch": 30.18884892086331, "step": 33570, "torque_loss": 0.18920695781707764 }, { "epoch": 30.197841726618705, "grad_norm": 0.15143121778964996, "learning_rate": 2.670147238101579e-05, "loss": 0.0187, "step": 33580 }, { "action_loss": 0.0028989408165216446, "epoch": 30.197841726618705, "step": 33580 }, { "epoch": 30.197841726618705, "step": 33580, "torque_loss": 0.14709939062595367 }, { "epoch": 30.2068345323741, "grad_norm": 0.1943870484828949, "learning_rate": 2.6672217669483473e-05, "loss": 0.0224, "step": 33590 }, { "action_loss": 0.0027664639055728912, "epoch": 30.2068345323741, "step": 33590 }, { "epoch": 30.2068345323741, "step": 33590, "torque_loss": 0.1834307461977005 }, { "epoch": 30.215827338129497, "grad_norm": 0.08426954597234726, "learning_rate": 2.6642973162315706e-05, "loss": 0.018, "step": 33600 }, { "action_loss": 0.005642725620418787, "epoch": 30.215827338129497, "step": 33600 }, { "epoch": 30.215827338129497, "step": 33600, "torque_loss": 0.22850246727466583 }, { "epoch": 30.22482014388489, "grad_norm": 0.1289300173521042, "learning_rate": 2.6613738872304988e-05, "loss": 0.0215, "step": 33610 }, { "action_loss": 0.003927105106413364, "epoch": 30.22482014388489, "step": 33610 }, { "epoch": 30.22482014388489, "step": 33610, "torque_loss": 0.25300663709640503 }, { "epoch": 30.23381294964029, "grad_norm": 0.09613407403230667, "learning_rate": 2.6584514812239437e-05, "loss": 0.0224, "step": 33620 }, { "action_loss": 0.0025257705710828304, "epoch": 30.23381294964029, "step": 33620 }, { "epoch": 30.23381294964029, "step": 33620, "torque_loss": 0.14254404604434967 }, { "epoch": 30.242805755395683, "grad_norm": 0.11786649376153946, "learning_rate": 2.655530099490261e-05, "loss": 0.0213, "step": 33630 }, { "action_loss": 0.0015421848511323333, "epoch": 30.242805755395683, "step": 33630 }, { "epoch": 30.242805755395683, "step": 33630, "torque_loss": 0.13608282804489136 }, { "epoch": 30.25179856115108, "grad_norm": 0.12129846215248108, "learning_rate": 2.6526097433073653e-05, "loss": 0.0191, "step": 33640 }, { "action_loss": 0.0014421831583604217, "epoch": 30.25179856115108, "step": 33640 }, { "epoch": 30.25179856115108, "step": 33640, "torque_loss": 0.16831763088703156 }, { "epoch": 30.260791366906474, "grad_norm": 0.12531360983848572, "learning_rate": 2.649690413952719e-05, "loss": 0.0194, "step": 33650 }, { "action_loss": 0.0016438706079497933, "epoch": 30.260791366906474, "step": 33650 }, { "epoch": 30.260791366906474, "step": 33650, "torque_loss": 0.1360214501619339 }, { "epoch": 30.269784172661872, "grad_norm": 0.12550345063209534, "learning_rate": 2.646772112703334e-05, "loss": 0.0206, "step": 33660 }, { "action_loss": 0.006869344040751457, "epoch": 30.269784172661872, "step": 33660 }, { "epoch": 30.269784172661872, "step": 33660, "torque_loss": 0.1897404044866562 }, { "epoch": 30.278776978417266, "grad_norm": 0.1366465836763382, "learning_rate": 2.6438548408357768e-05, "loss": 0.0216, "step": 33670 }, { "action_loss": 0.002034419449046254, "epoch": 30.278776978417266, "step": 33670 }, { "epoch": 30.278776978417266, "step": 33670, "torque_loss": 0.16300980746746063 }, { "epoch": 30.28776978417266, "grad_norm": 0.06380903720855713, "learning_rate": 2.6409385996261603e-05, "loss": 0.0203, "step": 33680 }, { "action_loss": 0.004901159554719925, "epoch": 30.28776978417266, "step": 33680 }, { "epoch": 30.28776978417266, "step": 33680, "torque_loss": 0.15965145826339722 }, { "epoch": 30.296762589928058, "grad_norm": 0.12740732729434967, "learning_rate": 2.638023390350148e-05, "loss": 0.0195, "step": 33690 }, { "action_loss": 0.005158859770745039, "epoch": 30.296762589928058, "step": 33690 }, { "epoch": 30.296762589928058, "step": 33690, "torque_loss": 0.18155010044574738 }, { "epoch": 30.305755395683452, "grad_norm": 0.13021783530712128, "learning_rate": 2.635109214282948e-05, "loss": 0.0184, "step": 33700 }, { "action_loss": 0.00302179460413754, "epoch": 30.305755395683452, "step": 33700 }, { "epoch": 30.305755395683452, "step": 33700, "torque_loss": 0.1255984902381897 }, { "epoch": 30.31474820143885, "grad_norm": 0.07150348275899887, "learning_rate": 2.6321960726993254e-05, "loss": 0.0199, "step": 33710 }, { "action_loss": 0.0021024958696216345, "epoch": 30.31474820143885, "step": 33710 }, { "epoch": 30.31474820143885, "step": 33710, "torque_loss": 0.1476706862449646 }, { "epoch": 30.323741007194243, "grad_norm": 0.0963139533996582, "learning_rate": 2.6292839668735846e-05, "loss": 0.018, "step": 33720 }, { "action_loss": 0.0013327108463272452, "epoch": 30.323741007194243, "step": 33720 }, { "epoch": 30.323741007194243, "step": 33720, "torque_loss": 0.14326973259449005 }, { "epoch": 30.33273381294964, "grad_norm": 0.05594521760940552, "learning_rate": 2.6263728980795778e-05, "loss": 0.0188, "step": 33730 }, { "action_loss": 0.00643345108255744, "epoch": 30.33273381294964, "step": 33730 }, { "epoch": 30.33273381294964, "step": 33730, "torque_loss": 0.18229681253433228 }, { "epoch": 30.341726618705035, "grad_norm": 0.050538524985313416, "learning_rate": 2.6234628675907092e-05, "loss": 0.0193, "step": 33740 }, { "action_loss": 0.0031699726823717356, "epoch": 30.341726618705035, "step": 33740 }, { "epoch": 30.341726618705035, "step": 33740, "torque_loss": 0.24716179072856903 }, { "epoch": 30.350719424460433, "grad_norm": 0.12294463813304901, "learning_rate": 2.620553876679923e-05, "loss": 0.0236, "step": 33750 }, { "action_loss": 0.011652366258203983, "epoch": 30.350719424460433, "step": 33750 }, { "epoch": 30.350719424460433, "step": 33750, "torque_loss": 0.25826895236968994 }, { "epoch": 30.359712230215827, "grad_norm": 0.0778680071234703, "learning_rate": 2.617645926619709e-05, "loss": 0.0205, "step": 33760 }, { "action_loss": 0.004992046859115362, "epoch": 30.359712230215827, "step": 33760 }, { "epoch": 30.359712230215827, "step": 33760, "torque_loss": 0.13597817718982697 }, { "epoch": 30.368705035971225, "grad_norm": 0.2233171910047531, "learning_rate": 2.6147390186821064e-05, "loss": 0.0247, "step": 33770 }, { "action_loss": 0.0013789753429591656, "epoch": 30.368705035971225, "step": 33770 }, { "epoch": 30.368705035971225, "step": 33770, "torque_loss": 0.1334240883588791 }, { "epoch": 30.37769784172662, "grad_norm": 0.07752396166324615, "learning_rate": 2.6118331541386955e-05, "loss": 0.0218, "step": 33780 }, { "action_loss": 0.003683392656967044, "epoch": 30.37769784172662, "step": 33780 }, { "epoch": 30.37769784172662, "step": 33780, "torque_loss": 0.23130935430526733 }, { "epoch": 30.386690647482013, "grad_norm": 0.15704050660133362, "learning_rate": 2.6089283342605953e-05, "loss": 0.0274, "step": 33790 }, { "action_loss": 0.006544125732034445, "epoch": 30.386690647482013, "step": 33790 }, { "epoch": 30.386690647482013, "step": 33790, "torque_loss": 0.2232423573732376 }, { "epoch": 30.39568345323741, "grad_norm": 0.12513676285743713, "learning_rate": 2.6060245603184797e-05, "loss": 0.0247, "step": 33800 }, { "action_loss": 0.004006055649369955, "epoch": 30.39568345323741, "step": 33800 }, { "epoch": 30.39568345323741, "step": 33800, "torque_loss": 0.13656418025493622 }, { "epoch": 30.404676258992804, "grad_norm": 0.17561165988445282, "learning_rate": 2.6031218335825536e-05, "loss": 0.0214, "step": 33810 }, { "action_loss": 0.006778651848435402, "epoch": 30.404676258992804, "step": 33810 }, { "epoch": 30.404676258992804, "step": 33810, "torque_loss": 0.19991903007030487 }, { "epoch": 30.413669064748202, "grad_norm": 0.08232305198907852, "learning_rate": 2.600220155322568e-05, "loss": 0.018, "step": 33820 }, { "action_loss": 0.002433456713333726, "epoch": 30.413669064748202, "step": 33820 }, { "epoch": 30.413669064748202, "step": 33820, "torque_loss": 0.07751625776290894 }, { "epoch": 30.422661870503596, "grad_norm": 0.057543374598026276, "learning_rate": 2.5973195268078186e-05, "loss": 0.0178, "step": 33830 }, { "action_loss": 0.002128718188032508, "epoch": 30.422661870503596, "step": 33830 }, { "epoch": 30.422661870503596, "step": 33830, "torque_loss": 0.19289511442184448 }, { "epoch": 30.431654676258994, "grad_norm": 0.06352401524782181, "learning_rate": 2.5944199493071352e-05, "loss": 0.0206, "step": 33840 }, { "action_loss": 0.0008624077308923006, "epoch": 30.431654676258994, "step": 33840 }, { "epoch": 30.431654676258994, "step": 33840, "torque_loss": 0.12320742756128311 }, { "epoch": 30.440647482014388, "grad_norm": 0.09368792921304703, "learning_rate": 2.591521424088895e-05, "loss": 0.0178, "step": 33850 }, { "action_loss": 0.00814792886376381, "epoch": 30.440647482014388, "step": 33850 }, { "epoch": 30.440647482014388, "step": 33850, "torque_loss": 0.13587628304958344 }, { "epoch": 30.449640287769785, "grad_norm": 0.16858676075935364, "learning_rate": 2.5886239524210086e-05, "loss": 0.0219, "step": 33860 }, { "action_loss": 0.0035856759641319513, "epoch": 30.449640287769785, "step": 33860 }, { "epoch": 30.449640287769785, "step": 33860, "torque_loss": 0.1642872393131256 }, { "epoch": 30.45863309352518, "grad_norm": 0.09453180432319641, "learning_rate": 2.5857275355709317e-05, "loss": 0.0181, "step": 33870 }, { "action_loss": 0.0014351438730955124, "epoch": 30.45863309352518, "step": 33870 }, { "epoch": 30.45863309352518, "step": 33870, "torque_loss": 0.12294063717126846 }, { "epoch": 30.467625899280577, "grad_norm": 0.09183092415332794, "learning_rate": 2.5828321748056533e-05, "loss": 0.0214, "step": 33880 }, { "action_loss": 0.0026031031738966703, "epoch": 30.467625899280577, "step": 33880 }, { "epoch": 30.467625899280577, "step": 33880, "torque_loss": 0.14232014119625092 }, { "epoch": 30.47661870503597, "grad_norm": 0.10824121534824371, "learning_rate": 2.5799378713917012e-05, "loss": 0.0173, "step": 33890 }, { "action_loss": 0.003212655195966363, "epoch": 30.47661870503597, "step": 33890 }, { "epoch": 30.47661870503597, "step": 33890, "torque_loss": 0.17220725119113922 }, { "epoch": 30.485611510791365, "grad_norm": 0.08400178700685501, "learning_rate": 2.5770446265951455e-05, "loss": 0.0219, "step": 33900 }, { "action_loss": 0.003994186874479055, "epoch": 30.485611510791365, "step": 33900 }, { "epoch": 30.485611510791365, "step": 33900, "torque_loss": 0.1861283928155899 }, { "epoch": 30.494604316546763, "grad_norm": 0.21598219871520996, "learning_rate": 2.5741524416815877e-05, "loss": 0.0234, "step": 33910 }, { "action_loss": 0.004364007618278265, "epoch": 30.494604316546763, "step": 33910 }, { "epoch": 30.494604316546763, "step": 33910, "torque_loss": 0.16831903159618378 }, { "epoch": 30.503597122302157, "grad_norm": 0.09567581117153168, "learning_rate": 2.571261317916166e-05, "loss": 0.0186, "step": 33920 }, { "action_loss": 0.0021863009314984083, "epoch": 30.503597122302157, "step": 33920 }, { "epoch": 30.503597122302157, "step": 33920, "torque_loss": 0.1704670935869217 }, { "epoch": 30.512589928057555, "grad_norm": 0.14178244769573212, "learning_rate": 2.5683712565635597e-05, "loss": 0.0243, "step": 33930 }, { "action_loss": 0.0018485118634998798, "epoch": 30.512589928057555, "step": 33930 }, { "epoch": 30.512589928057555, "step": 33930, "torque_loss": 0.15970924496650696 }, { "epoch": 30.52158273381295, "grad_norm": 0.0751936212182045, "learning_rate": 2.565482258887979e-05, "loss": 0.0201, "step": 33940 }, { "action_loss": 0.003079652087762952, "epoch": 30.52158273381295, "step": 33940 }, { "epoch": 30.52158273381295, "step": 33940, "torque_loss": 0.11301524192094803 }, { "epoch": 30.530575539568346, "grad_norm": 0.08187395334243774, "learning_rate": 2.5625943261531664e-05, "loss": 0.0175, "step": 33950 }, { "action_loss": 0.0010211453773081303, "epoch": 30.530575539568346, "step": 33950 }, { "epoch": 30.530575539568346, "step": 33950, "torque_loss": 0.16672784090042114 }, { "epoch": 30.53956834532374, "grad_norm": 0.102367103099823, "learning_rate": 2.5597074596224067e-05, "loss": 0.0223, "step": 33960 }, { "action_loss": 0.004819296766072512, "epoch": 30.53956834532374, "step": 33960 }, { "epoch": 30.53956834532374, "step": 33960, "torque_loss": 0.1495039016008377 }, { "epoch": 30.548561151079138, "grad_norm": 0.09447304159402847, "learning_rate": 2.5568216605585105e-05, "loss": 0.0208, "step": 33970 }, { "action_loss": 0.0025545803364366293, "epoch": 30.548561151079138, "step": 33970 }, { "epoch": 30.548561151079138, "step": 33970, "torque_loss": 0.10548561811447144 }, { "epoch": 30.557553956834532, "grad_norm": 0.092860646545887, "learning_rate": 2.5539369302238235e-05, "loss": 0.0241, "step": 33980 }, { "action_loss": 0.009647613391280174, "epoch": 30.557553956834532, "step": 33980 }, { "epoch": 30.557553956834532, "step": 33980, "torque_loss": 0.1720690280199051 }, { "epoch": 30.56654676258993, "grad_norm": 0.10429108142852783, "learning_rate": 2.5510532698802276e-05, "loss": 0.0216, "step": 33990 }, { "action_loss": 0.00284187332727015, "epoch": 30.56654676258993, "step": 33990 }, { "epoch": 30.56654676258993, "step": 33990, "torque_loss": 0.1903630644083023 }, { "epoch": 30.575539568345324, "grad_norm": 0.1269548088312149, "learning_rate": 2.5481706807891324e-05, "loss": 0.0249, "step": 34000 }, { "action_loss": 0.0020409568678587675, "epoch": 30.575539568345324, "step": 34000 }, { "epoch": 30.575539568345324, "step": 34000, "torque_loss": 0.17130924761295319 }, { "epoch": 30.584532374100718, "grad_norm": 0.12907467782497406, "learning_rate": 2.5452891642114783e-05, "loss": 0.0214, "step": 34010 }, { "action_loss": 0.002300378866493702, "epoch": 30.584532374100718, "step": 34010 }, { "epoch": 30.584532374100718, "step": 34010, "torque_loss": 0.1566399335861206 }, { "epoch": 30.593525179856115, "grad_norm": 0.09744244068861008, "learning_rate": 2.542408721407742e-05, "loss": 0.0188, "step": 34020 }, { "action_loss": 0.002021976513788104, "epoch": 30.593525179856115, "step": 34020 }, { "epoch": 30.593525179856115, "step": 34020, "torque_loss": 0.12979383766651154 }, { "epoch": 30.60251798561151, "grad_norm": 0.06963662803173065, "learning_rate": 2.539529353637926e-05, "loss": 0.0204, "step": 34030 }, { "action_loss": 0.0010059825144708157, "epoch": 30.60251798561151, "step": 34030 }, { "epoch": 30.60251798561151, "step": 34030, "torque_loss": 0.13367041945457458 }, { "epoch": 30.611510791366907, "grad_norm": 0.06940454989671707, "learning_rate": 2.5366510621615626e-05, "loss": 0.0196, "step": 34040 }, { "action_loss": 0.003764127613976598, "epoch": 30.611510791366907, "step": 34040 }, { "epoch": 30.611510791366907, "step": 34040, "torque_loss": 0.162023663520813 }, { "epoch": 30.6205035971223, "grad_norm": 0.1382577270269394, "learning_rate": 2.5337738482377132e-05, "loss": 0.024, "step": 34050 }, { "action_loss": 0.002603432396426797, "epoch": 30.6205035971223, "step": 34050 }, { "epoch": 30.6205035971223, "step": 34050, "torque_loss": 0.12707136571407318 }, { "epoch": 30.6294964028777, "grad_norm": 0.1444888412952423, "learning_rate": 2.5308977131249722e-05, "loss": 0.0178, "step": 34060 }, { "action_loss": 0.011363279074430466, "epoch": 30.6294964028777, "step": 34060 }, { "epoch": 30.6294964028777, "step": 34060, "torque_loss": 0.20389215648174286 }, { "epoch": 30.638489208633093, "grad_norm": 0.11934559047222137, "learning_rate": 2.5280226580814554e-05, "loss": 0.0178, "step": 34070 }, { "action_loss": 0.003261146368458867, "epoch": 30.638489208633093, "step": 34070 }, { "epoch": 30.638489208633093, "step": 34070, "torque_loss": 0.2585909962654114 }, { "epoch": 30.64748201438849, "grad_norm": 0.07265567779541016, "learning_rate": 2.525148684364813e-05, "loss": 0.0199, "step": 34080 }, { "action_loss": 0.0015829623444005847, "epoch": 30.64748201438849, "step": 34080 }, { "epoch": 30.64748201438849, "step": 34080, "torque_loss": 0.13030196726322174 }, { "epoch": 30.656474820143885, "grad_norm": 0.07037771493196487, "learning_rate": 2.522275793232215e-05, "loss": 0.0184, "step": 34090 }, { "action_loss": 0.001642380258999765, "epoch": 30.656474820143885, "step": 34090 }, { "epoch": 30.656474820143885, "step": 34090, "torque_loss": 0.13071317970752716 }, { "epoch": 30.665467625899282, "grad_norm": 0.10533572733402252, "learning_rate": 2.519403985940365e-05, "loss": 0.0201, "step": 34100 }, { "action_loss": 0.004447475541383028, "epoch": 30.665467625899282, "step": 34100 }, { "epoch": 30.665467625899282, "step": 34100, "torque_loss": 0.19062761962413788 }, { "epoch": 30.674460431654676, "grad_norm": 0.1132902204990387, "learning_rate": 2.5165332637454863e-05, "loss": 0.0201, "step": 34110 }, { "action_loss": 0.00470720836892724, "epoch": 30.674460431654676, "step": 34110 }, { "epoch": 30.674460431654676, "step": 34110, "torque_loss": 0.20476244390010834 }, { "epoch": 30.68345323741007, "grad_norm": 0.11384543031454086, "learning_rate": 2.513663627903334e-05, "loss": 0.0206, "step": 34120 }, { "action_loss": 0.001450957846827805, "epoch": 30.68345323741007, "step": 34120 }, { "epoch": 30.68345323741007, "step": 34120, "torque_loss": 0.15185189247131348 }, { "epoch": 30.692446043165468, "grad_norm": 0.10649235546588898, "learning_rate": 2.5107950796691815e-05, "loss": 0.0178, "step": 34130 }, { "action_loss": 0.0027149952948093414, "epoch": 30.692446043165468, "step": 34130 }, { "epoch": 30.692446043165468, "step": 34130, "torque_loss": 0.19178636372089386 }, { "epoch": 30.701438848920862, "grad_norm": 0.08798311650753021, "learning_rate": 2.5079276202978276e-05, "loss": 0.0245, "step": 34140 }, { "action_loss": 0.029008671641349792, "epoch": 30.701438848920862, "step": 34140 }, { "epoch": 30.701438848920862, "step": 34140, "torque_loss": 0.3429086208343506 }, { "epoch": 30.71043165467626, "grad_norm": 0.1288924217224121, "learning_rate": 2.505061251043601e-05, "loss": 0.0252, "step": 34150 }, { "action_loss": 0.0013240628177300096, "epoch": 30.71043165467626, "step": 34150 }, { "epoch": 30.71043165467626, "step": 34150, "torque_loss": 0.1674036979675293 }, { "epoch": 30.719424460431654, "grad_norm": 0.08068011701107025, "learning_rate": 2.502195973160346e-05, "loss": 0.0197, "step": 34160 }, { "action_loss": 0.005302051547914743, "epoch": 30.719424460431654, "step": 34160 }, { "epoch": 30.719424460431654, "step": 34160, "torque_loss": 0.1989191323518753 }, { "epoch": 30.72841726618705, "grad_norm": 0.08408977091312408, "learning_rate": 2.4993317879014316e-05, "loss": 0.0172, "step": 34170 }, { "action_loss": 0.0034778888802975416, "epoch": 30.72841726618705, "step": 34170 }, { "epoch": 30.72841726618705, "step": 34170, "torque_loss": 0.15632887184619904 }, { "epoch": 30.737410071942445, "grad_norm": 0.06271220743656158, "learning_rate": 2.4964686965197526e-05, "loss": 0.0176, "step": 34180 }, { "action_loss": 0.0006990178371779621, "epoch": 30.737410071942445, "step": 34180 }, { "epoch": 30.737410071942445, "step": 34180, "torque_loss": 0.14930413663387299 }, { "epoch": 30.746402877697843, "grad_norm": 0.14772704243659973, "learning_rate": 2.4936067002677217e-05, "loss": 0.024, "step": 34190 }, { "action_loss": 0.001436710706911981, "epoch": 30.746402877697843, "step": 34190 }, { "epoch": 30.746402877697843, "step": 34190, "torque_loss": 0.08970198780298233 }, { "epoch": 30.755395683453237, "grad_norm": 0.06639686226844788, "learning_rate": 2.4907458003972724e-05, "loss": 0.0167, "step": 34200 }, { "action_loss": 0.003114434890449047, "epoch": 30.755395683453237, "step": 34200 }, { "epoch": 30.755395683453237, "step": 34200, "torque_loss": 0.18980558216571808 }, { "epoch": 30.764388489208635, "grad_norm": 0.1351892650127411, "learning_rate": 2.4878859981598572e-05, "loss": 0.019, "step": 34210 }, { "action_loss": 0.001973865320906043, "epoch": 30.764388489208635, "step": 34210 }, { "epoch": 30.764388489208635, "step": 34210, "torque_loss": 0.1560981124639511 }, { "epoch": 30.77338129496403, "grad_norm": 0.09459756314754486, "learning_rate": 2.4850272948064555e-05, "loss": 0.0223, "step": 34220 }, { "action_loss": 0.004207255784422159, "epoch": 30.77338129496403, "step": 34220 }, { "epoch": 30.77338129496403, "step": 34220, "torque_loss": 0.12792332470417023 }, { "epoch": 30.782374100719423, "grad_norm": 0.136164590716362, "learning_rate": 2.4821696915875596e-05, "loss": 0.0221, "step": 34230 }, { "action_loss": 0.0013605663552880287, "epoch": 30.782374100719423, "step": 34230 }, { "epoch": 30.782374100719423, "step": 34230, "torque_loss": 0.08262664824724197 }, { "epoch": 30.79136690647482, "grad_norm": 0.10528832674026489, "learning_rate": 2.4793131897531798e-05, "loss": 0.0205, "step": 34240 }, { "action_loss": 0.0033839147072285414, "epoch": 30.79136690647482, "step": 34240 }, { "epoch": 30.79136690647482, "step": 34240, "torque_loss": 0.15222248435020447 }, { "epoch": 30.800359712230215, "grad_norm": 0.11472193896770477, "learning_rate": 2.47645779055285e-05, "loss": 0.0213, "step": 34250 }, { "action_loss": 0.013224899768829346, "epoch": 30.800359712230215, "step": 34250 }, { "epoch": 30.800359712230215, "step": 34250, "torque_loss": 0.277692586183548 }, { "epoch": 30.809352517985612, "grad_norm": 0.11138035356998444, "learning_rate": 2.4736034952356186e-05, "loss": 0.0271, "step": 34260 }, { "action_loss": 0.00750555656850338, "epoch": 30.809352517985612, "step": 34260 }, { "epoch": 30.809352517985612, "step": 34260, "torque_loss": 0.15692700445652008 }, { "epoch": 30.818345323741006, "grad_norm": 0.09886226058006287, "learning_rate": 2.470750305050048e-05, "loss": 0.026, "step": 34270 }, { "action_loss": 0.00330236554145813, "epoch": 30.818345323741006, "step": 34270 }, { "epoch": 30.818345323741006, "step": 34270, "torque_loss": 0.15161408483982086 }, { "epoch": 30.827338129496404, "grad_norm": 0.07174279540777206, "learning_rate": 2.4678982212442253e-05, "loss": 0.018, "step": 34280 }, { "action_loss": 0.0013512568548321724, "epoch": 30.827338129496404, "step": 34280 }, { "epoch": 30.827338129496404, "step": 34280, "torque_loss": 0.1344536393880844 }, { "epoch": 30.836330935251798, "grad_norm": 0.0940091535449028, "learning_rate": 2.4650472450657463e-05, "loss": 0.0185, "step": 34290 }, { "action_loss": 0.004038678947836161, "epoch": 30.836330935251798, "step": 34290 }, { "epoch": 30.836330935251798, "step": 34290, "torque_loss": 0.2389117032289505 }, { "epoch": 30.845323741007196, "grad_norm": 0.08534209430217743, "learning_rate": 2.4621973777617237e-05, "loss": 0.0214, "step": 34300 }, { "action_loss": 0.0018738344078883529, "epoch": 30.845323741007196, "step": 34300 }, { "epoch": 30.845323741007196, "step": 34300, "torque_loss": 0.11073143035173416 }, { "epoch": 30.85431654676259, "grad_norm": 0.07355383783578873, "learning_rate": 2.459348620578789e-05, "loss": 0.0193, "step": 34310 }, { "action_loss": 0.0011263369815424085, "epoch": 30.85431654676259, "step": 34310 }, { "epoch": 30.85431654676259, "step": 34310, "torque_loss": 0.12759153544902802 }, { "epoch": 30.863309352517987, "grad_norm": 0.1261092871427536, "learning_rate": 2.456500974763083e-05, "loss": 0.021, "step": 34320 }, { "action_loss": 0.0006512566469609737, "epoch": 30.863309352517987, "step": 34320 }, { "epoch": 30.863309352517987, "step": 34320, "torque_loss": 0.12857066094875336 }, { "epoch": 30.87230215827338, "grad_norm": 0.10019045323133469, "learning_rate": 2.4536544415602653e-05, "loss": 0.0181, "step": 34330 }, { "action_loss": 0.003240247955545783, "epoch": 30.87230215827338, "step": 34330 }, { "epoch": 30.87230215827338, "step": 34330, "torque_loss": 0.13284270465373993 }, { "epoch": 30.881294964028775, "grad_norm": 0.10965247452259064, "learning_rate": 2.4508090222155027e-05, "loss": 0.019, "step": 34340 }, { "action_loss": 0.001839691773056984, "epoch": 30.881294964028775, "step": 34340 }, { "epoch": 30.881294964028775, "step": 34340, "torque_loss": 0.11405777931213379 }, { "epoch": 30.890287769784173, "grad_norm": 0.07406434416770935, "learning_rate": 2.4479647179734822e-05, "loss": 0.0183, "step": 34350 }, { "action_loss": 0.004433223977684975, "epoch": 30.890287769784173, "step": 34350 }, { "epoch": 30.890287769784173, "step": 34350, "torque_loss": 0.18962806463241577 }, { "epoch": 30.899280575539567, "grad_norm": 0.07580983638763428, "learning_rate": 2.4451215300783953e-05, "loss": 0.0225, "step": 34360 }, { "action_loss": 0.0020610967185348272, "epoch": 30.899280575539567, "step": 34360 }, { "epoch": 30.899280575539567, "step": 34360, "torque_loss": 0.14914734661579132 }, { "epoch": 30.908273381294965, "grad_norm": 0.07682088762521744, "learning_rate": 2.442279459773953e-05, "loss": 0.0193, "step": 34370 }, { "action_loss": 0.0011288844980299473, "epoch": 30.908273381294965, "step": 34370 }, { "epoch": 30.908273381294965, "step": 34370, "torque_loss": 0.13550572097301483 }, { "epoch": 30.91726618705036, "grad_norm": 0.16463957726955414, "learning_rate": 2.4394385083033712e-05, "loss": 0.0188, "step": 34380 }, { "action_loss": 0.0030624738428741693, "epoch": 30.91726618705036, "step": 34380 }, { "epoch": 30.91726618705036, "step": 34380, "torque_loss": 0.12809337675571442 }, { "epoch": 30.926258992805757, "grad_norm": 0.1127704530954361, "learning_rate": 2.4365986769093792e-05, "loss": 0.0178, "step": 34390 }, { "action_loss": 0.0009578629396855831, "epoch": 30.926258992805757, "step": 34390 }, { "epoch": 30.926258992805757, "step": 34390, "torque_loss": 0.12573909759521484 }, { "epoch": 30.93525179856115, "grad_norm": 0.07150290161371231, "learning_rate": 2.4337599668342127e-05, "loss": 0.0235, "step": 34400 }, { "action_loss": 0.04022134467959404, "epoch": 30.93525179856115, "step": 34400 }, { "epoch": 30.93525179856115, "step": 34400, "torque_loss": 0.23492753505706787 }, { "epoch": 30.944244604316548, "grad_norm": 0.10518238693475723, "learning_rate": 2.4309223793196245e-05, "loss": 0.0227, "step": 34410 }, { "action_loss": 0.0018149851821362972, "epoch": 30.944244604316548, "step": 34410 }, { "epoch": 30.944244604316548, "step": 34410, "torque_loss": 0.1790589839220047 }, { "epoch": 30.953237410071942, "grad_norm": 0.12821035087108612, "learning_rate": 2.4280859156068698e-05, "loss": 0.0208, "step": 34420 }, { "action_loss": 0.0058220867067575455, "epoch": 30.953237410071942, "step": 34420 }, { "epoch": 30.953237410071942, "step": 34420, "torque_loss": 0.14638561010360718 }, { "epoch": 30.96223021582734, "grad_norm": 0.1297077089548111, "learning_rate": 2.4252505769367123e-05, "loss": 0.0202, "step": 34430 }, { "action_loss": 0.0010568975703790784, "epoch": 30.96223021582734, "step": 34430 }, { "epoch": 30.96223021582734, "step": 34430, "torque_loss": 0.12366481870412827 }, { "epoch": 30.971223021582734, "grad_norm": 0.08166655153036118, "learning_rate": 2.422416364549429e-05, "loss": 0.0189, "step": 34440 }, { "action_loss": 0.002560846507549286, "epoch": 30.971223021582734, "step": 34440 }, { "epoch": 30.971223021582734, "step": 34440, "torque_loss": 0.1638130098581314 }, { "epoch": 30.980215827338128, "grad_norm": 0.06123783811926842, "learning_rate": 2.4195832796847994e-05, "loss": 0.0209, "step": 34450 }, { "action_loss": 0.004625857342034578, "epoch": 30.980215827338128, "step": 34450 }, { "epoch": 30.980215827338128, "step": 34450, "torque_loss": 0.21164502203464508 }, { "epoch": 30.989208633093526, "grad_norm": 0.10935582965612411, "learning_rate": 2.416751323582109e-05, "loss": 0.0193, "step": 34460 }, { "action_loss": 0.0012962780892848969, "epoch": 30.989208633093526, "step": 34460 }, { "epoch": 30.989208633093526, "step": 34460, "torque_loss": 0.13711941242218018 }, { "epoch": 30.99820143884892, "grad_norm": 0.06455239653587341, "learning_rate": 2.4139204974801544e-05, "loss": 0.0166, "step": 34470 }, { "action_loss": 0.0015378798125311732, "epoch": 30.99820143884892, "step": 34470 }, { "epoch": 30.99820143884892, "step": 34470, "torque_loss": 0.13710293173789978 }, { "epoch": 31.007194244604317, "grad_norm": 0.0789744034409523, "learning_rate": 2.4110908026172347e-05, "loss": 0.0177, "step": 34480 }, { "action_loss": 0.0020598305854946375, "epoch": 31.007194244604317, "step": 34480 }, { "epoch": 31.007194244604317, "step": 34480, "torque_loss": 0.1498877853155136 }, { "epoch": 31.01618705035971, "grad_norm": 0.14841333031654358, "learning_rate": 2.4082622402311518e-05, "loss": 0.0292, "step": 34490 }, { "action_loss": 0.004936228971928358, "epoch": 31.01618705035971, "step": 34490 }, { "epoch": 31.01618705035971, "step": 34490, "torque_loss": 0.15294449031352997 }, { "epoch": 31.02517985611511, "grad_norm": 0.1322138011455536, "learning_rate": 2.4054348115592183e-05, "loss": 0.0172, "step": 34500 }, { "action_loss": 0.0019173339242115617, "epoch": 31.02517985611511, "step": 34500 }, { "epoch": 31.02517985611511, "step": 34500, "torque_loss": 0.12719659507274628 }, { "epoch": 31.034172661870503, "grad_norm": 0.09887433797121048, "learning_rate": 2.4026085178382475e-05, "loss": 0.0208, "step": 34510 }, { "action_loss": 0.001340744667686522, "epoch": 31.034172661870503, "step": 34510 }, { "epoch": 31.034172661870503, "step": 34510, "torque_loss": 0.1595817655324936 }, { "epoch": 31.0431654676259, "grad_norm": 0.11805813014507294, "learning_rate": 2.3997833603045523e-05, "loss": 0.0201, "step": 34520 }, { "action_loss": 0.003718761960044503, "epoch": 31.0431654676259, "step": 34520 }, { "epoch": 31.0431654676259, "step": 34520, "torque_loss": 0.14180059731006622 }, { "epoch": 31.052158273381295, "grad_norm": 0.06823715567588806, "learning_rate": 2.396959340193958e-05, "loss": 0.0209, "step": 34530 }, { "action_loss": 0.0014618203276768327, "epoch": 31.052158273381295, "step": 34530 }, { "epoch": 31.052158273381295, "step": 34530, "torque_loss": 0.20703424513339996 }, { "epoch": 31.06115107913669, "grad_norm": 0.1128980740904808, "learning_rate": 2.3941364587417836e-05, "loss": 0.0223, "step": 34540 }, { "action_loss": 0.0018120108870789409, "epoch": 31.06115107913669, "step": 34540 }, { "epoch": 31.06115107913669, "step": 34540, "torque_loss": 0.14472854137420654 }, { "epoch": 31.070143884892087, "grad_norm": 0.09955873340368271, "learning_rate": 2.3913147171828526e-05, "loss": 0.0178, "step": 34550 }, { "action_loss": 0.0039430526085197926, "epoch": 31.070143884892087, "step": 34550 }, { "epoch": 31.070143884892087, "step": 34550, "torque_loss": 0.11918941885232925 }, { "epoch": 31.07913669064748, "grad_norm": 0.12947870790958405, "learning_rate": 2.388494116751494e-05, "loss": 0.0257, "step": 34560 }, { "action_loss": 0.002953346585854888, "epoch": 31.07913669064748, "step": 34560 }, { "epoch": 31.07913669064748, "step": 34560, "torque_loss": 0.17923490703105927 }, { "epoch": 31.08812949640288, "grad_norm": 0.06938618421554565, "learning_rate": 2.3856746586815302e-05, "loss": 0.02, "step": 34570 }, { "action_loss": 0.002551676006987691, "epoch": 31.08812949640288, "step": 34570 }, { "epoch": 31.08812949640288, "step": 34570, "torque_loss": 0.1510433703660965 }, { "epoch": 31.097122302158272, "grad_norm": 0.14664165675640106, "learning_rate": 2.3828563442062924e-05, "loss": 0.0194, "step": 34580 }, { "action_loss": 0.0012387667084112763, "epoch": 31.097122302158272, "step": 34580 }, { "epoch": 31.097122302158272, "step": 34580, "torque_loss": 0.15789075195789337 }, { "epoch": 31.10611510791367, "grad_norm": 0.149921253323555, "learning_rate": 2.3800391745586026e-05, "loss": 0.0235, "step": 34590 }, { "action_loss": 0.0009527886286377907, "epoch": 31.10611510791367, "step": 34590 }, { "epoch": 31.10611510791367, "step": 34590, "torque_loss": 0.14651863276958466 }, { "epoch": 31.115107913669064, "grad_norm": 0.09929375350475311, "learning_rate": 2.377223150970791e-05, "loss": 0.0216, "step": 34600 }, { "action_loss": 0.0056782071478664875, "epoch": 31.115107913669064, "step": 34600 }, { "epoch": 31.115107913669064, "step": 34600, "torque_loss": 0.1859421581029892 }, { "epoch": 31.12410071942446, "grad_norm": 0.12905140221118927, "learning_rate": 2.3744082746746793e-05, "loss": 0.022, "step": 34610 }, { "action_loss": 0.0039091589860618114, "epoch": 31.12410071942446, "step": 34610 }, { "epoch": 31.12410071942446, "step": 34610, "torque_loss": 0.1804490089416504 }, { "epoch": 31.133093525179856, "grad_norm": 0.2190348505973816, "learning_rate": 2.3715945469015893e-05, "loss": 0.0229, "step": 34620 }, { "action_loss": 0.0024058744311332703, "epoch": 31.133093525179856, "step": 34620 }, { "epoch": 31.133093525179856, "step": 34620, "torque_loss": 0.1539030820131302 }, { "epoch": 31.142086330935253, "grad_norm": 0.0567525252699852, "learning_rate": 2.368781968882343e-05, "loss": 0.0176, "step": 34630 }, { "action_loss": 0.0015139365568757057, "epoch": 31.142086330935253, "step": 34630 }, { "epoch": 31.142086330935253, "step": 34630, "torque_loss": 0.13654662668704987 }, { "epoch": 31.151079136690647, "grad_norm": 0.11937715858221054, "learning_rate": 2.365970541847258e-05, "loss": 0.0177, "step": 34640 }, { "action_loss": 0.0017084144055843353, "epoch": 31.151079136690647, "step": 34640 }, { "epoch": 31.151079136690647, "step": 34640, "torque_loss": 0.13339585065841675 }, { "epoch": 31.16007194244604, "grad_norm": 0.09219534695148468, "learning_rate": 2.3631602670261445e-05, "loss": 0.0179, "step": 34650 }, { "action_loss": 0.003229718655347824, "epoch": 31.16007194244604, "step": 34650 }, { "epoch": 31.16007194244604, "step": 34650, "torque_loss": 0.16176225244998932 }, { "epoch": 31.16906474820144, "grad_norm": 0.11315011233091354, "learning_rate": 2.3603511456483164e-05, "loss": 0.0247, "step": 34660 }, { "action_loss": 0.012769113294780254, "epoch": 31.16906474820144, "step": 34660 }, { "epoch": 31.16906474820144, "step": 34660, "torque_loss": 0.2559875547885895 }, { "epoch": 31.178057553956833, "grad_norm": 0.1056421548128128, "learning_rate": 2.357543178942577e-05, "loss": 0.022, "step": 34670 }, { "action_loss": 0.004601588472723961, "epoch": 31.178057553956833, "step": 34670 }, { "epoch": 31.178057553956833, "step": 34670, "torque_loss": 0.23496557772159576 }, { "epoch": 31.18705035971223, "grad_norm": 0.07892335951328278, "learning_rate": 2.3547363681372242e-05, "loss": 0.0207, "step": 34680 }, { "action_loss": 0.0035982641857117414, "epoch": 31.18705035971223, "step": 34680 }, { "epoch": 31.18705035971223, "step": 34680, "torque_loss": 0.11688532680273056 }, { "epoch": 31.196043165467625, "grad_norm": 0.19918371737003326, "learning_rate": 2.3519307144600565e-05, "loss": 0.0215, "step": 34690 }, { "action_loss": 0.0031824230682104826, "epoch": 31.196043165467625, "step": 34690 }, { "epoch": 31.196043165467625, "step": 34690, "torque_loss": 0.21146805584430695 }, { "epoch": 31.205035971223023, "grad_norm": 0.13810057938098907, "learning_rate": 2.3491262191383607e-05, "loss": 0.0216, "step": 34700 }, { "action_loss": 0.0037801507860422134, "epoch": 31.205035971223023, "step": 34700 }, { "epoch": 31.205035971223023, "step": 34700, "torque_loss": 0.17492710053920746 }, { "epoch": 31.214028776978417, "grad_norm": 0.10581222176551819, "learning_rate": 2.3463228833989154e-05, "loss": 0.0221, "step": 34710 }, { "action_loss": 0.006150847766548395, "epoch": 31.214028776978417, "step": 34710 }, { "epoch": 31.214028776978417, "step": 34710, "torque_loss": 0.18530254065990448 }, { "epoch": 31.223021582733814, "grad_norm": 0.10975227504968643, "learning_rate": 2.343520708467999e-05, "loss": 0.0194, "step": 34720 }, { "action_loss": 0.002248611068353057, "epoch": 31.223021582733814, "step": 34720 }, { "epoch": 31.223021582733814, "step": 34720, "torque_loss": 0.15766160190105438 }, { "epoch": 31.23201438848921, "grad_norm": 0.06902801245450974, "learning_rate": 2.3407196955713767e-05, "loss": 0.0209, "step": 34730 }, { "action_loss": 0.004064986482262611, "epoch": 31.23201438848921, "step": 34730 }, { "epoch": 31.23201438848921, "step": 34730, "torque_loss": 0.13647951185703278 }, { "epoch": 31.241007194244606, "grad_norm": 0.05442330613732338, "learning_rate": 2.337919845934306e-05, "loss": 0.0175, "step": 34740 }, { "action_loss": 0.0028225609567016363, "epoch": 31.241007194244606, "step": 34740 }, { "epoch": 31.241007194244606, "step": 34740, "torque_loss": 0.12988698482513428 }, { "epoch": 31.25, "grad_norm": 0.07657089084386826, "learning_rate": 2.335121160781535e-05, "loss": 0.0156, "step": 34750 }, { "action_loss": 0.0031330445781350136, "epoch": 31.25, "step": 34750 }, { "epoch": 31.25, "step": 34750, "torque_loss": 0.18154418468475342 }, { "epoch": 31.258992805755394, "grad_norm": 0.06976009905338287, "learning_rate": 2.3323236413373063e-05, "loss": 0.0198, "step": 34760 }, { "action_loss": 0.0018902028677985072, "epoch": 31.258992805755394, "step": 34760 }, { "epoch": 31.258992805755394, "step": 34760, "torque_loss": 0.11975022405385971 }, { "epoch": 31.26798561151079, "grad_norm": 0.08759357035160065, "learning_rate": 2.3295272888253494e-05, "loss": 0.0187, "step": 34770 }, { "action_loss": 0.0015193348517641425, "epoch": 31.26798561151079, "step": 34770 }, { "epoch": 31.26798561151079, "step": 34770, "torque_loss": 0.14068782329559326 }, { "epoch": 31.276978417266186, "grad_norm": 0.1048693135380745, "learning_rate": 2.3267321044688815e-05, "loss": 0.0169, "step": 34780 }, { "action_loss": 0.0012706443667411804, "epoch": 31.276978417266186, "step": 34780 }, { "epoch": 31.276978417266186, "step": 34780, "torque_loss": 0.13121049106121063 }, { "epoch": 31.285971223021583, "grad_norm": 0.15383011102676392, "learning_rate": 2.323938089490615e-05, "loss": 0.0191, "step": 34790 }, { "action_loss": 0.0013484313385561109, "epoch": 31.285971223021583, "step": 34790 }, { "epoch": 31.285971223021583, "step": 34790, "torque_loss": 0.13562364876270294 }, { "epoch": 31.294964028776977, "grad_norm": 0.16121439635753632, "learning_rate": 2.321145245112743e-05, "loss": 0.0172, "step": 34800 }, { "action_loss": 0.00232702842913568, "epoch": 31.294964028776977, "step": 34800 }, { "epoch": 31.294964028776977, "step": 34800, "torque_loss": 0.21514266729354858 }, { "epoch": 31.303956834532375, "grad_norm": 0.11476306617259979, "learning_rate": 2.318353572556953e-05, "loss": 0.0196, "step": 34810 }, { "action_loss": 0.004786956589668989, "epoch": 31.303956834532375, "step": 34810 }, { "epoch": 31.303956834532375, "step": 34810, "torque_loss": 0.19242887198925018 }, { "epoch": 31.31294964028777, "grad_norm": 0.08842013031244278, "learning_rate": 2.3155630730444183e-05, "loss": 0.0219, "step": 34820 }, { "action_loss": 0.0018067910568788648, "epoch": 31.31294964028777, "step": 34820 }, { "epoch": 31.31294964028777, "step": 34820, "torque_loss": 0.15711994469165802 }, { "epoch": 31.321942446043167, "grad_norm": 0.07142828404903412, "learning_rate": 2.3127737477957968e-05, "loss": 0.0199, "step": 34830 }, { "action_loss": 0.005499780178070068, "epoch": 31.321942446043167, "step": 34830 }, { "epoch": 31.321942446043167, "step": 34830, "torque_loss": 0.22311373054981232 }, { "epoch": 31.33093525179856, "grad_norm": 0.2346479892730713, "learning_rate": 2.3099855980312324e-05, "loss": 0.024, "step": 34840 }, { "action_loss": 0.0009483764879405499, "epoch": 31.33093525179856, "step": 34840 }, { "epoch": 31.33093525179856, "step": 34840, "torque_loss": 0.12075670808553696 }, { "epoch": 31.33992805755396, "grad_norm": 0.1399787813425064, "learning_rate": 2.3071986249703602e-05, "loss": 0.0207, "step": 34850 }, { "action_loss": 0.002283180132508278, "epoch": 31.33992805755396, "step": 34850 }, { "epoch": 31.33992805755396, "step": 34850, "torque_loss": 0.13253095746040344 }, { "epoch": 31.348920863309353, "grad_norm": 0.09132447838783264, "learning_rate": 2.3044128298322953e-05, "loss": 0.0243, "step": 34860 }, { "action_loss": 0.005173944868147373, "epoch": 31.348920863309353, "step": 34860 }, { "epoch": 31.348920863309353, "step": 34860, "torque_loss": 0.17076772451400757 }, { "epoch": 31.357913669064747, "grad_norm": 0.12819825112819672, "learning_rate": 2.301628213835637e-05, "loss": 0.0192, "step": 34870 }, { "action_loss": 0.0027618298772722483, "epoch": 31.357913669064747, "step": 34870 }, { "epoch": 31.357913669064747, "step": 34870, "torque_loss": 0.13930131494998932 }, { "epoch": 31.366906474820144, "grad_norm": 0.12085374444723129, "learning_rate": 2.2988447781984744e-05, "loss": 0.0181, "step": 34880 }, { "action_loss": 0.001301614916883409, "epoch": 31.366906474820144, "step": 34880 }, { "epoch": 31.366906474820144, "step": 34880, "torque_loss": 0.1846768856048584 }, { "epoch": 31.37589928057554, "grad_norm": 0.15166422724723816, "learning_rate": 2.2960625241383753e-05, "loss": 0.0189, "step": 34890 }, { "action_loss": 0.005907545331865549, "epoch": 31.37589928057554, "step": 34890 }, { "epoch": 31.37589928057554, "step": 34890, "torque_loss": 0.1490657925605774 }, { "epoch": 31.384892086330936, "grad_norm": 0.1074652299284935, "learning_rate": 2.2932814528723922e-05, "loss": 0.0192, "step": 34900 }, { "action_loss": 0.0011766611132770777, "epoch": 31.384892086330936, "step": 34900 }, { "epoch": 31.384892086330936, "step": 34900, "torque_loss": 0.09122773259878159 }, { "epoch": 31.39388489208633, "grad_norm": 0.14670997858047485, "learning_rate": 2.2905015656170582e-05, "loss": 0.018, "step": 34910 }, { "action_loss": 0.0035661172587424517, "epoch": 31.39388489208633, "step": 34910 }, { "epoch": 31.39388489208633, "step": 34910, "torque_loss": 0.2302415370941162 }, { "epoch": 31.402877697841728, "grad_norm": 0.1122933104634285, "learning_rate": 2.287722863588394e-05, "loss": 0.0229, "step": 34920 }, { "action_loss": 0.0013180729001760483, "epoch": 31.402877697841728, "step": 34920 }, { "epoch": 31.402877697841728, "step": 34920, "torque_loss": 0.09575729817152023 }, { "epoch": 31.41187050359712, "grad_norm": 0.07942786812782288, "learning_rate": 2.2849453480018962e-05, "loss": 0.0172, "step": 34930 }, { "action_loss": 0.001048261416144669, "epoch": 31.41187050359712, "step": 34930 }, { "epoch": 31.41187050359712, "step": 34930, "torque_loss": 0.10852700471878052 }, { "epoch": 31.42086330935252, "grad_norm": 0.11689916253089905, "learning_rate": 2.282169020072543e-05, "loss": 0.0212, "step": 34940 }, { "action_loss": 0.0022403569892048836, "epoch": 31.42086330935252, "step": 34940 }, { "epoch": 31.42086330935252, "step": 34940, "torque_loss": 0.17711134254932404 }, { "epoch": 31.429856115107913, "grad_norm": 0.14771591126918793, "learning_rate": 2.2793938810147986e-05, "loss": 0.0207, "step": 34950 }, { "action_loss": 0.0015909395879134536, "epoch": 31.429856115107913, "step": 34950 }, { "epoch": 31.429856115107913, "step": 34950, "torque_loss": 0.10090557485818863 }, { "epoch": 31.43884892086331, "grad_norm": 0.07175612449645996, "learning_rate": 2.2766199320426007e-05, "loss": 0.0206, "step": 34960 }, { "action_loss": 0.009523162618279457, "epoch": 31.43884892086331, "step": 34960 }, { "epoch": 31.43884892086331, "step": 34960, "torque_loss": 0.18854935467243195 }, { "epoch": 31.447841726618705, "grad_norm": 0.1196250319480896, "learning_rate": 2.2738471743693674e-05, "loss": 0.0201, "step": 34970 }, { "action_loss": 0.0025753520894795656, "epoch": 31.447841726618705, "step": 34970 }, { "epoch": 31.447841726618705, "step": 34970, "torque_loss": 0.11368593573570251 }, { "epoch": 31.4568345323741, "grad_norm": 0.1054653748869896, "learning_rate": 2.2710756092080006e-05, "loss": 0.0211, "step": 34980 }, { "action_loss": 0.0023319043684750795, "epoch": 31.4568345323741, "step": 34980 }, { "epoch": 31.4568345323741, "step": 34980, "torque_loss": 0.18059229850769043 }, { "epoch": 31.465827338129497, "grad_norm": 0.07258525490760803, "learning_rate": 2.2683052377708747e-05, "loss": 0.0177, "step": 34990 }, { "action_loss": 0.0017964104190468788, "epoch": 31.465827338129497, "step": 34990 }, { "epoch": 31.465827338129497, "step": 34990, "torque_loss": 0.14929789304733276 }, { "epoch": 31.47482014388489, "grad_norm": 0.05562737584114075, "learning_rate": 2.265536061269843e-05, "loss": 0.0166, "step": 35000 }, { "action_loss": 0.0024581437464803457, "epoch": 31.47482014388489, "step": 35000 }, { "epoch": 31.47482014388489, "step": 35000, "torque_loss": 0.16166239976882935 }, { "epoch": 31.48381294964029, "grad_norm": 0.08788628876209259, "learning_rate": 2.262768080916241e-05, "loss": 0.0199, "step": 35010 }, { "action_loss": 0.0024294175673276186, "epoch": 31.48381294964029, "step": 35010 }, { "epoch": 31.48381294964029, "step": 35010, "torque_loss": 0.10649312287569046 }, { "epoch": 31.492805755395683, "grad_norm": 0.0646236389875412, "learning_rate": 2.260001297920874e-05, "loss": 0.0159, "step": 35020 }, { "action_loss": 0.003379674395546317, "epoch": 31.492805755395683, "step": 35020 }, { "epoch": 31.492805755395683, "step": 35020, "torque_loss": 0.18474836647510529 }, { "epoch": 31.50179856115108, "grad_norm": 0.10504340380430222, "learning_rate": 2.2572357134940297e-05, "loss": 0.0225, "step": 35030 }, { "action_loss": 0.002449335530400276, "epoch": 31.50179856115108, "step": 35030 }, { "epoch": 31.50179856115108, "step": 35030, "torque_loss": 0.14162582159042358 }, { "epoch": 31.510791366906474, "grad_norm": 0.10714162886142731, "learning_rate": 2.2544713288454655e-05, "loss": 0.0185, "step": 35040 }, { "action_loss": 0.009792315773665905, "epoch": 31.510791366906474, "step": 35040 }, { "epoch": 31.510791366906474, "step": 35040, "torque_loss": 0.1962791532278061 }, { "epoch": 31.519784172661872, "grad_norm": 0.04921826720237732, "learning_rate": 2.2517081451844213e-05, "loss": 0.0193, "step": 35050 }, { "action_loss": 0.002387396292760968, "epoch": 31.519784172661872, "step": 35050 }, { "epoch": 31.519784172661872, "step": 35050, "torque_loss": 0.1394648551940918 }, { "epoch": 31.528776978417266, "grad_norm": 0.10170610249042511, "learning_rate": 2.248946163719605e-05, "loss": 0.0224, "step": 35060 }, { "action_loss": 0.0020427836570888758, "epoch": 31.528776978417266, "step": 35060 }, { "epoch": 31.528776978417266, "step": 35060, "torque_loss": 0.1190112829208374 }, { "epoch": 31.53776978417266, "grad_norm": 0.09520536661148071, "learning_rate": 2.2461853856591996e-05, "loss": 0.0269, "step": 35070 }, { "action_loss": 0.001315962872467935, "epoch": 31.53776978417266, "step": 35070 }, { "epoch": 31.53776978417266, "step": 35070, "torque_loss": 0.15537697076797485 }, { "epoch": 31.546762589928058, "grad_norm": 0.08330997079610825, "learning_rate": 2.2434258122108664e-05, "loss": 0.0201, "step": 35080 }, { "action_loss": 0.006089413072913885, "epoch": 31.546762589928058, "step": 35080 }, { "epoch": 31.546762589928058, "step": 35080, "torque_loss": 0.12946756184101105 }, { "epoch": 31.555755395683452, "grad_norm": 0.07494467496871948, "learning_rate": 2.2406674445817356e-05, "loss": 0.0262, "step": 35090 }, { "action_loss": 0.0009950444800779223, "epoch": 31.555755395683452, "step": 35090 }, { "epoch": 31.555755395683452, "step": 35090, "torque_loss": 0.1407192498445511 }, { "epoch": 31.56474820143885, "grad_norm": 0.17957063019275665, "learning_rate": 2.2379102839784083e-05, "loss": 0.0181, "step": 35100 }, { "action_loss": 0.0038129196036607027, "epoch": 31.56474820143885, "step": 35100 }, { "epoch": 31.56474820143885, "step": 35100, "torque_loss": 0.18542219698429108 }, { "epoch": 31.573741007194243, "grad_norm": 0.05581500381231308, "learning_rate": 2.2351543316069645e-05, "loss": 0.0193, "step": 35110 }, { "action_loss": 0.005910432431846857, "epoch": 31.573741007194243, "step": 35110 }, { "epoch": 31.573741007194243, "step": 35110, "torque_loss": 0.1704595535993576 }, { "epoch": 31.58273381294964, "grad_norm": 0.08067623525857925, "learning_rate": 2.2323995886729493e-05, "loss": 0.0247, "step": 35120 }, { "action_loss": 0.0015942234313115478, "epoch": 31.58273381294964, "step": 35120 }, { "epoch": 31.58273381294964, "step": 35120, "torque_loss": 0.11598517745733261 }, { "epoch": 31.591726618705035, "grad_norm": 0.12149898707866669, "learning_rate": 2.2296460563813786e-05, "loss": 0.0186, "step": 35130 }, { "action_loss": 0.0018361978000029922, "epoch": 31.591726618705035, "step": 35130 }, { "epoch": 31.591726618705035, "step": 35130, "torque_loss": 0.1486889123916626 }, { "epoch": 31.600719424460433, "grad_norm": 0.10517563670873642, "learning_rate": 2.2268937359367457e-05, "loss": 0.0203, "step": 35140 }, { "action_loss": 0.0025915270671248436, "epoch": 31.600719424460433, "step": 35140 }, { "epoch": 31.600719424460433, "step": 35140, "torque_loss": 0.13088743388652802 }, { "epoch": 31.609712230215827, "grad_norm": 0.1873469352722168, "learning_rate": 2.224142628543006e-05, "loss": 0.0233, "step": 35150 }, { "action_loss": 0.00253126653842628, "epoch": 31.609712230215827, "step": 35150 }, { "epoch": 31.609712230215827, "step": 35150, "torque_loss": 0.1482718139886856 }, { "epoch": 31.618705035971225, "grad_norm": 0.11183354258537292, "learning_rate": 2.2213927354035863e-05, "loss": 0.0208, "step": 35160 }, { "action_loss": 0.0018958878936246037, "epoch": 31.618705035971225, "step": 35160 }, { "epoch": 31.618705035971225, "step": 35160, "torque_loss": 0.1619616001844406 }, { "epoch": 31.62769784172662, "grad_norm": 0.11349520832300186, "learning_rate": 2.218644057721387e-05, "loss": 0.0187, "step": 35170 }, { "action_loss": 0.0030774192418903112, "epoch": 31.62769784172662, "step": 35170 }, { "epoch": 31.62769784172662, "step": 35170, "torque_loss": 0.19453953206539154 }, { "epoch": 31.636690647482013, "grad_norm": 0.07192272692918777, "learning_rate": 2.2158965966987703e-05, "loss": 0.0186, "step": 35180 }, { "action_loss": 0.0013721389696002007, "epoch": 31.636690647482013, "step": 35180 }, { "epoch": 31.636690647482013, "step": 35180, "torque_loss": 0.1300308257341385 }, { "epoch": 31.64568345323741, "grad_norm": 0.05899902060627937, "learning_rate": 2.2131503535375687e-05, "loss": 0.0166, "step": 35190 }, { "action_loss": 0.0056028966791927814, "epoch": 31.64568345323741, "step": 35190 }, { "epoch": 31.64568345323741, "step": 35190, "torque_loss": 0.17848962545394897 }, { "epoch": 31.654676258992804, "grad_norm": 0.05461748316884041, "learning_rate": 2.2104053294390846e-05, "loss": 0.0184, "step": 35200 }, { "action_loss": 0.004253041464835405, "epoch": 31.654676258992804, "step": 35200 }, { "epoch": 31.654676258992804, "step": 35200, "torque_loss": 0.23904024064540863 }, { "epoch": 31.663669064748202, "grad_norm": 0.04468195140361786, "learning_rate": 2.2076615256040834e-05, "loss": 0.0166, "step": 35210 }, { "action_loss": 0.0014536784728989005, "epoch": 31.663669064748202, "step": 35210 }, { "epoch": 31.663669064748202, "step": 35210, "torque_loss": 0.15709367394447327 }, { "epoch": 31.672661870503596, "grad_norm": 0.06596878916025162, "learning_rate": 2.2049189432327966e-05, "loss": 0.0205, "step": 35220 }, { "action_loss": 0.003055400447919965, "epoch": 31.672661870503596, "step": 35220 }, { "epoch": 31.672661870503596, "step": 35220, "torque_loss": 0.17718331515789032 }, { "epoch": 31.681654676258994, "grad_norm": 0.11315882951021194, "learning_rate": 2.202177583524926e-05, "loss": 0.0196, "step": 35230 }, { "action_loss": 0.012637472711503506, "epoch": 31.681654676258994, "step": 35230 }, { "epoch": 31.681654676258994, "step": 35230, "torque_loss": 0.16247910261154175 }, { "epoch": 31.690647482014388, "grad_norm": 0.0705241933465004, "learning_rate": 2.1994374476796343e-05, "loss": 0.0194, "step": 35240 }, { "action_loss": 0.0035930683370679617, "epoch": 31.690647482014388, "step": 35240 }, { "epoch": 31.690647482014388, "step": 35240, "torque_loss": 0.16095249354839325 }, { "epoch": 31.699640287769785, "grad_norm": 0.0823090597987175, "learning_rate": 2.1966985368955477e-05, "loss": 0.021, "step": 35250 }, { "action_loss": 0.001982971793040633, "epoch": 31.699640287769785, "step": 35250 }, { "epoch": 31.699640287769785, "step": 35250, "torque_loss": 0.19303488731384277 }, { "epoch": 31.70863309352518, "grad_norm": 0.07798737287521362, "learning_rate": 2.193960852370762e-05, "loss": 0.0181, "step": 35260 }, { "action_loss": 0.001811112859286368, "epoch": 31.70863309352518, "step": 35260 }, { "epoch": 31.70863309352518, "step": 35260, "torque_loss": 0.15502811968326569 }, { "epoch": 31.717625899280577, "grad_norm": 0.14613249897956848, "learning_rate": 2.1912243953028317e-05, "loss": 0.0213, "step": 35270 }, { "action_loss": 0.0038573257625102997, "epoch": 31.717625899280577, "step": 35270 }, { "epoch": 31.717625899280577, "step": 35270, "torque_loss": 0.23083281517028809 }, { "epoch": 31.72661870503597, "grad_norm": 0.08717263489961624, "learning_rate": 2.1884891668887776e-05, "loss": 0.0203, "step": 35280 }, { "action_loss": 0.002756731351837516, "epoch": 31.72661870503597, "step": 35280 }, { "epoch": 31.72661870503597, "step": 35280, "torque_loss": 0.1776779145002365 }, { "epoch": 31.735611510791365, "grad_norm": 0.08979557454586029, "learning_rate": 2.1857551683250787e-05, "loss": 0.0185, "step": 35290 }, { "action_loss": 0.0030697763431817293, "epoch": 31.735611510791365, "step": 35290 }, { "epoch": 31.735611510791365, "step": 35290, "torque_loss": 0.19168995320796967 }, { "epoch": 31.744604316546763, "grad_norm": 0.06861703842878342, "learning_rate": 2.1830224008076827e-05, "loss": 0.0182, "step": 35300 }, { "action_loss": 0.004198869690299034, "epoch": 31.744604316546763, "step": 35300 }, { "epoch": 31.744604316546763, "step": 35300, "torque_loss": 0.17760664224624634 }, { "epoch": 31.753597122302157, "grad_norm": 0.10514809936285019, "learning_rate": 2.1802908655319924e-05, "loss": 0.0216, "step": 35310 }, { "action_loss": 0.005137852858752012, "epoch": 31.753597122302157, "step": 35310 }, { "epoch": 31.753597122302157, "step": 35310, "torque_loss": 0.20563292503356934 }, { "epoch": 31.762589928057555, "grad_norm": 0.09060496091842651, "learning_rate": 2.1775605636928738e-05, "loss": 0.0198, "step": 35320 }, { "action_loss": 0.00165898900013417, "epoch": 31.762589928057555, "step": 35320 }, { "epoch": 31.762589928057555, "step": 35320, "torque_loss": 0.1310795396566391 }, { "epoch": 31.77158273381295, "grad_norm": 0.2052532434463501, "learning_rate": 2.1748314964846555e-05, "loss": 0.026, "step": 35330 }, { "action_loss": 0.0023333572316914797, "epoch": 31.77158273381295, "step": 35330 }, { "epoch": 31.77158273381295, "step": 35330, "torque_loss": 0.1348271518945694 }, { "epoch": 31.780575539568346, "grad_norm": 0.10218067467212677, "learning_rate": 2.172103665101124e-05, "loss": 0.0209, "step": 35340 }, { "action_loss": 0.0023191303480416536, "epoch": 31.780575539568346, "step": 35340 }, { "epoch": 31.780575539568346, "step": 35340, "torque_loss": 0.1952252835035324 }, { "epoch": 31.78956834532374, "grad_norm": 0.11441787332296371, "learning_rate": 2.1693770707355226e-05, "loss": 0.0169, "step": 35350 }, { "action_loss": 0.004345593973994255, "epoch": 31.78956834532374, "step": 35350 }, { "epoch": 31.78956834532374, "step": 35350, "torque_loss": 0.1666562408208847 }, { "epoch": 31.798561151079138, "grad_norm": 0.06445194780826569, "learning_rate": 2.1666517145805605e-05, "loss": 0.0199, "step": 35360 }, { "action_loss": 0.0013416189467534423, "epoch": 31.798561151079138, "step": 35360 }, { "epoch": 31.798561151079138, "step": 35360, "torque_loss": 0.14308874309062958 }, { "epoch": 31.807553956834532, "grad_norm": 0.0664673000574112, "learning_rate": 2.1639275978283985e-05, "loss": 0.0188, "step": 35370 }, { "action_loss": 0.008593783713877201, "epoch": 31.807553956834532, "step": 35370 }, { "epoch": 31.807553956834532, "step": 35370, "torque_loss": 0.23036940395832062 }, { "epoch": 31.81654676258993, "grad_norm": 0.09671834856271744, "learning_rate": 2.1612047216706567e-05, "loss": 0.0186, "step": 35380 }, { "action_loss": 0.009752633981406689, "epoch": 31.81654676258993, "step": 35380 }, { "epoch": 31.81654676258993, "step": 35380, "torque_loss": 0.2659834921360016 }, { "epoch": 31.825539568345324, "grad_norm": 0.0740460529923439, "learning_rate": 2.158483087298417e-05, "loss": 0.0257, "step": 35390 }, { "action_loss": 0.0034651977475732565, "epoch": 31.825539568345324, "step": 35390 }, { "epoch": 31.825539568345324, "step": 35390, "torque_loss": 0.19768555462360382 }, { "epoch": 31.834532374100718, "grad_norm": 0.11850806325674057, "learning_rate": 2.155762695902213e-05, "loss": 0.0262, "step": 35400 }, { "action_loss": 0.004696036223322153, "epoch": 31.834532374100718, "step": 35400 }, { "epoch": 31.834532374100718, "step": 35400, "torque_loss": 0.1884813755750656 }, { "epoch": 31.843525179856115, "grad_norm": 0.1414499133825302, "learning_rate": 2.1530435486720363e-05, "loss": 0.0213, "step": 35410 }, { "action_loss": 0.002084974432364106, "epoch": 31.843525179856115, "step": 35410 }, { "epoch": 31.843525179856115, "step": 35410, "torque_loss": 0.14180804789066315 }, { "epoch": 31.85251798561151, "grad_norm": 0.08461251854896545, "learning_rate": 2.1503256467973327e-05, "loss": 0.0174, "step": 35420 }, { "action_loss": 0.002300904830917716, "epoch": 31.85251798561151, "step": 35420 }, { "epoch": 31.85251798561151, "step": 35420, "torque_loss": 0.17610788345336914 }, { "epoch": 31.861510791366907, "grad_norm": 0.0828874334692955, "learning_rate": 2.147608991467008e-05, "loss": 0.0179, "step": 35430 }, { "action_loss": 0.008799468167126179, "epoch": 31.861510791366907, "step": 35430 }, { "epoch": 31.861510791366907, "step": 35430, "torque_loss": 0.26415637135505676 }, { "epoch": 31.8705035971223, "grad_norm": 0.08730320632457733, "learning_rate": 2.144893583869418e-05, "loss": 0.0211, "step": 35440 }, { "action_loss": 0.0027593474369496107, "epoch": 31.8705035971223, "step": 35440 }, { "epoch": 31.8705035971223, "step": 35440, "torque_loss": 0.1390617936849594 }, { "epoch": 31.8794964028777, "grad_norm": 0.10835523158311844, "learning_rate": 2.142179425192372e-05, "loss": 0.0186, "step": 35450 }, { "action_loss": 0.0010499021736904979, "epoch": 31.8794964028777, "step": 35450 }, { "epoch": 31.8794964028777, "step": 35450, "torque_loss": 0.14701375365257263 }, { "epoch": 31.888489208633093, "grad_norm": 0.10900828242301941, "learning_rate": 2.1394665166231393e-05, "loss": 0.0251, "step": 35460 }, { "action_loss": 0.002589271403849125, "epoch": 31.888489208633093, "step": 35460 }, { "epoch": 31.888489208633093, "step": 35460, "torque_loss": 0.16850005090236664 }, { "epoch": 31.89748201438849, "grad_norm": 0.08019507676362991, "learning_rate": 2.1367548593484364e-05, "loss": 0.02, "step": 35470 }, { "action_loss": 0.008389608934521675, "epoch": 31.89748201438849, "step": 35470 }, { "epoch": 31.89748201438849, "step": 35470, "torque_loss": 0.16289012134075165 }, { "epoch": 31.906474820143885, "grad_norm": 0.11460325866937637, "learning_rate": 2.134044454554433e-05, "loss": 0.0216, "step": 35480 }, { "action_loss": 0.0007337657152675092, "epoch": 31.906474820143885, "step": 35480 }, { "epoch": 31.906474820143885, "step": 35480, "torque_loss": 0.10754179209470749 }, { "epoch": 31.915467625899282, "grad_norm": 0.09579305350780487, "learning_rate": 2.1313353034267548e-05, "loss": 0.0191, "step": 35490 }, { "action_loss": 0.0029969383031129837, "epoch": 31.915467625899282, "step": 35490 }, { "epoch": 31.915467625899282, "step": 35490, "torque_loss": 0.15243178606033325 }, { "epoch": 31.924460431654676, "grad_norm": 0.08013154566287994, "learning_rate": 2.1286274071504735e-05, "loss": 0.0166, "step": 35500 }, { "action_loss": 0.001628347672522068, "epoch": 31.924460431654676, "step": 35500 }, { "epoch": 31.924460431654676, "step": 35500, "torque_loss": 0.1512541025876999 }, { "epoch": 31.93345323741007, "grad_norm": 0.07348043471574783, "learning_rate": 2.125920766910119e-05, "loss": 0.0179, "step": 35510 }, { "action_loss": 0.0013734353706240654, "epoch": 31.93345323741007, "step": 35510 }, { "epoch": 31.93345323741007, "step": 35510, "torque_loss": 0.185282900929451 }, { "epoch": 31.942446043165468, "grad_norm": 0.0906180664896965, "learning_rate": 2.1232153838896628e-05, "loss": 0.0215, "step": 35520 }, { "action_loss": 0.002101943828165531, "epoch": 31.942446043165468, "step": 35520 }, { "epoch": 31.942446043165468, "step": 35520, "torque_loss": 0.1454625129699707 }, { "epoch": 31.951438848920862, "grad_norm": 0.07581029832363129, "learning_rate": 2.1205112592725366e-05, "loss": 0.0244, "step": 35530 }, { "action_loss": 0.0017169221537187696, "epoch": 31.951438848920862, "step": 35530 }, { "epoch": 31.951438848920862, "step": 35530, "torque_loss": 0.1362989991903305 }, { "epoch": 31.96043165467626, "grad_norm": 0.10771419107913971, "learning_rate": 2.1178083942416114e-05, "loss": 0.019, "step": 35540 }, { "action_loss": 0.002014454221352935, "epoch": 31.96043165467626, "step": 35540 }, { "epoch": 31.96043165467626, "step": 35540, "torque_loss": 0.07997659593820572 }, { "epoch": 31.969424460431654, "grad_norm": 0.08137757331132889, "learning_rate": 2.1151067899792164e-05, "loss": 0.0165, "step": 35550 }, { "action_loss": 0.006542498711496592, "epoch": 31.969424460431654, "step": 35550 }, { "epoch": 31.969424460431654, "step": 35550, "torque_loss": 0.1695971041917801 }, { "epoch": 31.97841726618705, "grad_norm": 0.07746504247188568, "learning_rate": 2.1124064476671235e-05, "loss": 0.023, "step": 35560 }, { "action_loss": 0.009572614915668964, "epoch": 31.97841726618705, "step": 35560 }, { "epoch": 31.97841726618705, "step": 35560, "torque_loss": 0.30909231305122375 }, { "epoch": 31.987410071942445, "grad_norm": 0.08917418122291565, "learning_rate": 2.1097073684865516e-05, "loss": 0.0198, "step": 35570 }, { "action_loss": 0.015904782339930534, "epoch": 31.987410071942445, "step": 35570 }, { "epoch": 31.987410071942445, "step": 35570, "torque_loss": 0.19290578365325928 }, { "epoch": 31.996402877697843, "grad_norm": 0.10590720176696777, "learning_rate": 2.107009553618174e-05, "loss": 0.0185, "step": 35580 }, { "action_loss": 0.001773991622030735, "epoch": 31.996402877697843, "step": 35580 }, { "epoch": 31.996402877697843, "step": 35580, "torque_loss": 0.18522532284259796 }, { "epoch": 32.00539568345324, "grad_norm": 0.09394945204257965, "learning_rate": 2.1043130042421044e-05, "loss": 0.0196, "step": 35590 }, { "action_loss": 0.0014529902255162597, "epoch": 32.00539568345324, "step": 35590 }, { "epoch": 32.00539568345324, "step": 35590, "torque_loss": 0.14075016975402832 }, { "epoch": 32.014388489208635, "grad_norm": 0.10643348842859268, "learning_rate": 2.1016177215379053e-05, "loss": 0.0233, "step": 35600 }, { "action_loss": 0.014232625253498554, "epoch": 32.014388489208635, "step": 35600 }, { "epoch": 32.014388489208635, "step": 35600, "torque_loss": 0.19081532955169678 }, { "epoch": 32.023381294964025, "grad_norm": 0.09501466155052185, "learning_rate": 2.0989237066845824e-05, "loss": 0.0214, "step": 35610 }, { "action_loss": 0.006270830985158682, "epoch": 32.023381294964025, "step": 35610 }, { "epoch": 32.023381294964025, "step": 35610, "torque_loss": 0.2705206573009491 }, { "epoch": 32.03237410071942, "grad_norm": 0.08715463429689407, "learning_rate": 2.096230960860594e-05, "loss": 0.0201, "step": 35620 }, { "action_loss": 0.0024168917443603277, "epoch": 32.03237410071942, "step": 35620 }, { "epoch": 32.03237410071942, "step": 35620, "torque_loss": 0.17964893579483032 }, { "epoch": 32.04136690647482, "grad_norm": 0.06326192617416382, "learning_rate": 2.0935394852438366e-05, "loss": 0.0171, "step": 35630 }, { "action_loss": 0.003561466932296753, "epoch": 32.04136690647482, "step": 35630 }, { "epoch": 32.04136690647482, "step": 35630, "torque_loss": 0.19042102992534637 }, { "epoch": 32.05035971223022, "grad_norm": 0.16086876392364502, "learning_rate": 2.090849281011651e-05, "loss": 0.0216, "step": 35640 }, { "action_loss": 0.005281759891659021, "epoch": 32.05035971223022, "step": 35640 }, { "epoch": 32.05035971223022, "step": 35640, "torque_loss": 0.16515947878360748 }, { "epoch": 32.05935251798561, "grad_norm": 0.09084933996200562, "learning_rate": 2.088160349340827e-05, "loss": 0.0193, "step": 35650 }, { "action_loss": 0.0034646561834961176, "epoch": 32.05935251798561, "step": 35650 }, { "epoch": 32.05935251798561, "step": 35650, "torque_loss": 0.12200654298067093 }, { "epoch": 32.068345323741006, "grad_norm": 0.09949246793985367, "learning_rate": 2.085472691407594e-05, "loss": 0.0183, "step": 35660 }, { "action_loss": 0.0018418632680550218, "epoch": 32.068345323741006, "step": 35660 }, { "epoch": 32.068345323741006, "step": 35660, "torque_loss": 0.13123483955860138 }, { "epoch": 32.077338129496404, "grad_norm": 0.08428632467985153, "learning_rate": 2.082786308387622e-05, "loss": 0.0211, "step": 35670 }, { "action_loss": 0.004351743962615728, "epoch": 32.077338129496404, "step": 35670 }, { "epoch": 32.077338129496404, "step": 35670, "torque_loss": 0.12575159966945648 }, { "epoch": 32.0863309352518, "grad_norm": 0.11507375538349152, "learning_rate": 2.0801012014560306e-05, "loss": 0.022, "step": 35680 }, { "action_loss": 0.0027493902016431093, "epoch": 32.0863309352518, "step": 35680 }, { "epoch": 32.0863309352518, "step": 35680, "torque_loss": 0.16173619031906128 }, { "epoch": 32.09532374100719, "grad_norm": 0.14594657719135284, "learning_rate": 2.0774173717873746e-05, "loss": 0.032, "step": 35690 }, { "action_loss": 0.002678284188732505, "epoch": 32.09532374100719, "step": 35690 }, { "epoch": 32.09532374100719, "step": 35690, "torque_loss": 0.22461090981960297 }, { "epoch": 32.10431654676259, "grad_norm": 0.09703799337148666, "learning_rate": 2.074734820555651e-05, "loss": 0.024, "step": 35700 }, { "action_loss": 0.002340447623282671, "epoch": 32.10431654676259, "step": 35700 }, { "epoch": 32.10431654676259, "step": 35700, "torque_loss": 0.15057100355625153 }, { "epoch": 32.11330935251799, "grad_norm": 0.0674981027841568, "learning_rate": 2.0720535489343014e-05, "loss": 0.0207, "step": 35710 }, { "action_loss": 0.002846966264769435, "epoch": 32.11330935251799, "step": 35710 }, { "epoch": 32.11330935251799, "step": 35710, "torque_loss": 0.16718168556690216 }, { "epoch": 32.12230215827338, "grad_norm": 0.08680958300828934, "learning_rate": 2.0693735580962044e-05, "loss": 0.0171, "step": 35720 }, { "action_loss": 0.002814929001033306, "epoch": 32.12230215827338, "step": 35720 }, { "epoch": 32.12230215827338, "step": 35720, "torque_loss": 0.1280939131975174 }, { "epoch": 32.131294964028775, "grad_norm": 0.07098838686943054, "learning_rate": 2.0666948492136766e-05, "loss": 0.0192, "step": 35730 }, { "action_loss": 0.002007890259847045, "epoch": 32.131294964028775, "step": 35730 }, { "epoch": 32.131294964028775, "step": 35730, "torque_loss": 0.18793897330760956 }, { "epoch": 32.14028776978417, "grad_norm": 0.053605783730745316, "learning_rate": 2.0640174234584797e-05, "loss": 0.0187, "step": 35740 }, { "action_loss": 0.0046053738333284855, "epoch": 32.14028776978417, "step": 35740 }, { "epoch": 32.14028776978417, "step": 35740, "torque_loss": 0.15811601281166077 }, { "epoch": 32.14928057553957, "grad_norm": 0.08963663130998611, "learning_rate": 2.061341282001808e-05, "loss": 0.0213, "step": 35750 }, { "action_loss": 0.0028725385200232267, "epoch": 32.14928057553957, "step": 35750 }, { "epoch": 32.14928057553957, "step": 35750, "torque_loss": 0.19680137932300568 }, { "epoch": 32.15827338129496, "grad_norm": 0.11483994871377945, "learning_rate": 2.0586664260142997e-05, "loss": 0.0214, "step": 35760 }, { "action_loss": 0.0033529922366142273, "epoch": 32.15827338129496, "step": 35760 }, { "epoch": 32.15827338129496, "step": 35760, "torque_loss": 0.1760561615228653 }, { "epoch": 32.16726618705036, "grad_norm": 0.0875113233923912, "learning_rate": 2.0559928566660237e-05, "loss": 0.018, "step": 35770 }, { "action_loss": 0.003018447430804372, "epoch": 32.16726618705036, "step": 35770 }, { "epoch": 32.16726618705036, "step": 35770, "torque_loss": 0.17268072068691254 }, { "epoch": 32.17625899280576, "grad_norm": 0.06346404552459717, "learning_rate": 2.0533205751264945e-05, "loss": 0.0176, "step": 35780 }, { "action_loss": 0.0024785541463643312, "epoch": 32.17625899280576, "step": 35780 }, { "epoch": 32.17625899280576, "step": 35780, "torque_loss": 0.1460900753736496 }, { "epoch": 32.185251798561154, "grad_norm": 0.06770112365484238, "learning_rate": 2.050649582564656e-05, "loss": 0.0248, "step": 35790 }, { "action_loss": 0.0013388745719566941, "epoch": 32.185251798561154, "step": 35790 }, { "epoch": 32.185251798561154, "step": 35790, "torque_loss": 0.16853666305541992 }, { "epoch": 32.194244604316545, "grad_norm": 0.07015237212181091, "learning_rate": 2.04797988014889e-05, "loss": 0.0161, "step": 35800 }, { "action_loss": 0.003295597853139043, "epoch": 32.194244604316545, "step": 35800 }, { "epoch": 32.194244604316545, "step": 35800, "torque_loss": 0.1546686291694641 }, { "epoch": 32.20323741007194, "grad_norm": 0.09006325155496597, "learning_rate": 2.0453114690470194e-05, "loss": 0.0183, "step": 35810 }, { "action_loss": 0.004702422767877579, "epoch": 32.20323741007194, "step": 35810 }, { "epoch": 32.20323741007194, "step": 35810, "torque_loss": 0.19254070520401 }, { "epoch": 32.21223021582734, "grad_norm": 0.05317559465765953, "learning_rate": 2.0426443504262944e-05, "loss": 0.0195, "step": 35820 }, { "action_loss": 0.0047776959836483, "epoch": 32.21223021582734, "step": 35820 }, { "epoch": 32.21223021582734, "step": 35820, "torque_loss": 0.23107969760894775 }, { "epoch": 32.22122302158273, "grad_norm": 0.0913740023970604, "learning_rate": 2.0399785254534027e-05, "loss": 0.0175, "step": 35830 }, { "action_loss": 0.058605119585990906, "epoch": 32.22122302158273, "step": 35830 }, { "epoch": 32.22122302158273, "step": 35830, "torque_loss": 0.18897907435894012 }, { "epoch": 32.23021582733813, "grad_norm": 0.08727133274078369, "learning_rate": 2.0373139952944697e-05, "loss": 0.0281, "step": 35840 }, { "action_loss": 0.004404375795274973, "epoch": 32.23021582733813, "step": 35840 }, { "epoch": 32.23021582733813, "step": 35840, "torque_loss": 0.18527357280254364 }, { "epoch": 32.239208633093526, "grad_norm": 0.09433799982070923, "learning_rate": 2.03465076111505e-05, "loss": 0.0209, "step": 35850 }, { "action_loss": 0.0010543399257585406, "epoch": 32.239208633093526, "step": 35850 }, { "epoch": 32.239208633093526, "step": 35850, "torque_loss": 0.09813620895147324 }, { "epoch": 32.24820143884892, "grad_norm": 0.1647006869316101, "learning_rate": 2.0319888240801305e-05, "loss": 0.0183, "step": 35860 }, { "action_loss": 0.0016328789060935378, "epoch": 32.24820143884892, "step": 35860 }, { "epoch": 32.24820143884892, "step": 35860, "torque_loss": 0.16654296219348907 }, { "epoch": 32.257194244604314, "grad_norm": 0.14974819123744965, "learning_rate": 2.0293281853541375e-05, "loss": 0.0215, "step": 35870 }, { "action_loss": 0.0023451971355825663, "epoch": 32.257194244604314, "step": 35870 }, { "epoch": 32.257194244604314, "step": 35870, "torque_loss": 0.17375890910625458 }, { "epoch": 32.26618705035971, "grad_norm": 0.16470250487327576, "learning_rate": 2.0266688461009216e-05, "loss": 0.0208, "step": 35880 }, { "action_loss": 0.0016946991672739387, "epoch": 32.26618705035971, "step": 35880 }, { "epoch": 32.26618705035971, "step": 35880, "torque_loss": 0.16120116412639618 }, { "epoch": 32.27517985611511, "grad_norm": 0.07682036608457565, "learning_rate": 2.0240108074837666e-05, "loss": 0.0194, "step": 35890 }, { "action_loss": 0.0016223435522988439, "epoch": 32.27517985611511, "step": 35890 }, { "epoch": 32.27517985611511, "step": 35890, "torque_loss": 0.12318050861358643 }, { "epoch": 32.28417266187051, "grad_norm": 0.05689557269215584, "learning_rate": 2.021354070665393e-05, "loss": 0.0182, "step": 35900 }, { "action_loss": 0.008662051521241665, "epoch": 32.28417266187051, "step": 35900 }, { "epoch": 32.28417266187051, "step": 35900, "torque_loss": 0.1891642063856125 }, { "epoch": 32.2931654676259, "grad_norm": 0.0600326769053936, "learning_rate": 2.0186986368079465e-05, "loss": 0.0227, "step": 35910 }, { "action_loss": 0.0017056390643119812, "epoch": 32.2931654676259, "step": 35910 }, { "epoch": 32.2931654676259, "step": 35910, "torque_loss": 0.12060508877038956 }, { "epoch": 32.302158273381295, "grad_norm": 0.06366968154907227, "learning_rate": 2.016044507073001e-05, "loss": 0.0196, "step": 35920 }, { "action_loss": 0.0012896762927994132, "epoch": 32.302158273381295, "step": 35920 }, { "epoch": 32.302158273381295, "step": 35920, "torque_loss": 0.1327023059129715 }, { "epoch": 32.31115107913669, "grad_norm": 0.169472798705101, "learning_rate": 2.013391682621567e-05, "loss": 0.0217, "step": 35930 }, { "action_loss": 0.002294153207913041, "epoch": 32.31115107913669, "step": 35930 }, { "epoch": 32.31115107913669, "step": 35930, "torque_loss": 0.12693215906620026 }, { "epoch": 32.32014388489208, "grad_norm": 0.10500852763652802, "learning_rate": 2.0107401646140787e-05, "loss": 0.0264, "step": 35940 }, { "action_loss": 0.004095152020454407, "epoch": 32.32014388489208, "step": 35940 }, { "epoch": 32.32014388489208, "step": 35940, "torque_loss": 0.20633743703365326 }, { "epoch": 32.32913669064748, "grad_norm": 0.0813286229968071, "learning_rate": 2.0080899542104003e-05, "loss": 0.0192, "step": 35950 }, { "action_loss": 0.002834334271028638, "epoch": 32.32913669064748, "step": 35950 }, { "epoch": 32.32913669064748, "step": 35950, "torque_loss": 0.1333792805671692 }, { "epoch": 32.33812949640288, "grad_norm": 0.08297336101531982, "learning_rate": 2.0054410525698215e-05, "loss": 0.0183, "step": 35960 }, { "action_loss": 0.002521367510780692, "epoch": 32.33812949640288, "step": 35960 }, { "epoch": 32.33812949640288, "step": 35960, "torque_loss": 0.11520075798034668 }, { "epoch": 32.347122302158276, "grad_norm": 0.10467672348022461, "learning_rate": 2.0027934608510653e-05, "loss": 0.0225, "step": 35970 }, { "action_loss": 0.0014474695781245828, "epoch": 32.347122302158276, "step": 35970 }, { "epoch": 32.347122302158276, "step": 35970, "torque_loss": 0.16167844831943512 }, { "epoch": 32.356115107913666, "grad_norm": 0.07194022089242935, "learning_rate": 2.000147180212275e-05, "loss": 0.0193, "step": 35980 }, { "action_loss": 0.0028595279436558485, "epoch": 32.356115107913666, "step": 35980 }, { "epoch": 32.356115107913666, "step": 35980, "torque_loss": 0.13912741839885712 }, { "epoch": 32.365107913669064, "grad_norm": 0.08645035326480865, "learning_rate": 1.9975022118110275e-05, "loss": 0.02, "step": 35990 }, { "action_loss": 0.0008596935658715665, "epoch": 32.365107913669064, "step": 35990 }, { "epoch": 32.365107913669064, "step": 35990, "torque_loss": 0.1540491282939911 }, { "epoch": 32.37410071942446, "grad_norm": 0.0673464983701706, "learning_rate": 1.994858556804318e-05, "loss": 0.0235, "step": 36000 }, { "action_loss": 0.0063211568631231785, "epoch": 32.37410071942446, "step": 36000 }, { "epoch": 32.37410071942446, "step": 36000, "torque_loss": 0.22212953865528107 }, { "epoch": 32.38309352517986, "grad_norm": 0.14424638450145721, "learning_rate": 1.9922162163485748e-05, "loss": 0.0212, "step": 36010 }, { "action_loss": 0.0036635694559663534, "epoch": 32.38309352517986, "step": 36010 }, { "epoch": 32.38309352517986, "step": 36010, "torque_loss": 0.13614024221897125 }, { "epoch": 32.39208633093525, "grad_norm": 0.07825211435556412, "learning_rate": 1.989575191599643e-05, "loss": 0.0182, "step": 36020 }, { "action_loss": 0.025416573509573936, "epoch": 32.39208633093525, "step": 36020 }, { "epoch": 32.39208633093525, "step": 36020, "torque_loss": 0.204695463180542 }, { "epoch": 32.40107913669065, "grad_norm": 0.10382894426584244, "learning_rate": 1.9869354837128012e-05, "loss": 0.0256, "step": 36030 }, { "action_loss": 0.0022435651626437902, "epoch": 32.40107913669065, "step": 36030 }, { "epoch": 32.40107913669065, "step": 36030, "torque_loss": 0.1864575743675232 }, { "epoch": 32.410071942446045, "grad_norm": 0.08147107064723969, "learning_rate": 1.984297093842745e-05, "loss": 0.0178, "step": 36040 }, { "action_loss": 0.0016167787835001945, "epoch": 32.410071942446045, "step": 36040 }, { "epoch": 32.410071942446045, "step": 36040, "torque_loss": 0.15592803061008453 }, { "epoch": 32.419064748201436, "grad_norm": 0.14806152880191803, "learning_rate": 1.981660023143594e-05, "loss": 0.0237, "step": 36050 }, { "action_loss": 0.0034720657858997583, "epoch": 32.419064748201436, "step": 36050 }, { "epoch": 32.419064748201436, "step": 36050, "torque_loss": 0.1748344749212265 }, { "epoch": 32.42805755395683, "grad_norm": 0.12232770025730133, "learning_rate": 1.9790242727688963e-05, "loss": 0.0212, "step": 36060 }, { "action_loss": 0.010845921002328396, "epoch": 32.42805755395683, "step": 36060 }, { "epoch": 32.42805755395683, "step": 36060, "torque_loss": 0.19396817684173584 }, { "epoch": 32.43705035971223, "grad_norm": 0.10710000991821289, "learning_rate": 1.9763898438716165e-05, "loss": 0.0192, "step": 36070 }, { "action_loss": 0.0011836111079901457, "epoch": 32.43705035971223, "step": 36070 }, { "epoch": 32.43705035971223, "step": 36070, "torque_loss": 0.09452768415212631 }, { "epoch": 32.44604316546763, "grad_norm": 0.10473062843084335, "learning_rate": 1.973756737604142e-05, "loss": 0.0211, "step": 36080 }, { "action_loss": 0.002305548870936036, "epoch": 32.44604316546763, "step": 36080 }, { "epoch": 32.44604316546763, "step": 36080, "torque_loss": 0.1699189692735672 }, { "epoch": 32.45503597122302, "grad_norm": 0.11287882924079895, "learning_rate": 1.9711249551182858e-05, "loss": 0.0211, "step": 36090 }, { "action_loss": 0.0007846414227969944, "epoch": 32.45503597122302, "step": 36090 }, { "epoch": 32.45503597122302, "step": 36090, "torque_loss": 0.17547418177127838 }, { "epoch": 32.46402877697842, "grad_norm": 0.08426173031330109, "learning_rate": 1.968494497565278e-05, "loss": 0.0218, "step": 36100 }, { "action_loss": 0.008796638809144497, "epoch": 32.46402877697842, "step": 36100 }, { "epoch": 32.46402877697842, "step": 36100, "torque_loss": 0.24511952698230743 }, { "epoch": 32.473021582733814, "grad_norm": 0.05462004616856575, "learning_rate": 1.9658653660957698e-05, "loss": 0.0226, "step": 36110 }, { "action_loss": 0.003009397303685546, "epoch": 32.473021582733814, "step": 36110 }, { "epoch": 32.473021582733814, "step": 36110, "torque_loss": 0.11500024050474167 }, { "epoch": 32.48201438848921, "grad_norm": 0.0770864188671112, "learning_rate": 1.9632375618598302e-05, "loss": 0.0211, "step": 36120 }, { "action_loss": 0.0013586374698206782, "epoch": 32.48201438848921, "step": 36120 }, { "epoch": 32.48201438848921, "step": 36120, "torque_loss": 0.21131356060504913 }, { "epoch": 32.4910071942446, "grad_norm": 0.07352428138256073, "learning_rate": 1.960611086006955e-05, "loss": 0.0208, "step": 36130 }, { "action_loss": 0.003522912971675396, "epoch": 32.4910071942446, "step": 36130 }, { "epoch": 32.4910071942446, "step": 36130, "torque_loss": 0.1592748910188675 }, { "epoch": 32.5, "grad_norm": 0.0874905064702034, "learning_rate": 1.9579859396860513e-05, "loss": 0.0185, "step": 36140 }, { "action_loss": 0.0029997602105140686, "epoch": 32.5, "step": 36140 }, { "epoch": 32.5, "step": 36140, "torque_loss": 0.20047813653945923 }, { "epoch": 32.5089928057554, "grad_norm": 0.09504545480012894, "learning_rate": 1.9553621240454452e-05, "loss": 0.0193, "step": 36150 }, { "action_loss": 0.0007456480525434017, "epoch": 32.5089928057554, "step": 36150 }, { "epoch": 32.5089928057554, "step": 36150, "torque_loss": 0.09887956827878952 }, { "epoch": 32.51798561151079, "grad_norm": 0.11435726284980774, "learning_rate": 1.9527396402328864e-05, "loss": 0.0211, "step": 36160 }, { "action_loss": 0.013724944554269314, "epoch": 32.51798561151079, "step": 36160 }, { "epoch": 32.51798561151079, "step": 36160, "torque_loss": 0.1963847130537033 }, { "epoch": 32.526978417266186, "grad_norm": 0.07502736896276474, "learning_rate": 1.9501184893955376e-05, "loss": 0.0193, "step": 36170 }, { "action_loss": 0.0015575034776702523, "epoch": 32.526978417266186, "step": 36170 }, { "epoch": 32.526978417266186, "step": 36170, "torque_loss": 0.14176611602306366 }, { "epoch": 32.53597122302158, "grad_norm": 0.11188478767871857, "learning_rate": 1.9474986726799765e-05, "loss": 0.0185, "step": 36180 }, { "action_loss": 0.021549366414546967, "epoch": 32.53597122302158, "step": 36180 }, { "epoch": 32.53597122302158, "step": 36180, "torque_loss": 0.21154038608074188 }, { "epoch": 32.54496402877698, "grad_norm": 0.09710349887609482, "learning_rate": 1.944880191232204e-05, "loss": 0.0202, "step": 36190 }, { "action_loss": 0.0009203816298395395, "epoch": 32.54496402877698, "step": 36190 }, { "epoch": 32.54496402877698, "step": 36190, "torque_loss": 0.10807874798774719 }, { "epoch": 32.55395683453237, "grad_norm": 0.0727107897400856, "learning_rate": 1.9422630461976304e-05, "loss": 0.0168, "step": 36200 }, { "action_loss": 0.004384372849017382, "epoch": 32.55395683453237, "step": 36200 }, { "epoch": 32.55395683453237, "step": 36200, "torque_loss": 0.15472058951854706 }, { "epoch": 32.56294964028777, "grad_norm": 0.1190573051571846, "learning_rate": 1.9396472387210835e-05, "loss": 0.0224, "step": 36210 }, { "action_loss": 0.001956377411261201, "epoch": 32.56294964028777, "step": 36210 }, { "epoch": 32.56294964028777, "step": 36210, "torque_loss": 0.2076510637998581 }, { "epoch": 32.57194244604317, "grad_norm": 0.07006721943616867, "learning_rate": 1.9370327699468082e-05, "loss": 0.0211, "step": 36220 }, { "action_loss": 0.0016811443492770195, "epoch": 32.57194244604317, "step": 36220 }, { "epoch": 32.57194244604317, "step": 36220, "torque_loss": 0.1652502417564392 }, { "epoch": 32.580935251798564, "grad_norm": 0.05727817118167877, "learning_rate": 1.9344196410184594e-05, "loss": 0.0174, "step": 36230 }, { "action_loss": 0.0007219502585940063, "epoch": 32.580935251798564, "step": 36230 }, { "epoch": 32.580935251798564, "step": 36230, "torque_loss": 0.11833473294973373 }, { "epoch": 32.589928057553955, "grad_norm": 0.08565147966146469, "learning_rate": 1.9318078530791123e-05, "loss": 0.0229, "step": 36240 }, { "action_loss": 0.0008476909715682268, "epoch": 32.589928057553955, "step": 36240 }, { "epoch": 32.589928057553955, "step": 36240, "torque_loss": 0.0643095150589943 }, { "epoch": 32.59892086330935, "grad_norm": 0.06346631050109863, "learning_rate": 1.9291974072712473e-05, "loss": 0.0149, "step": 36250 }, { "action_loss": 0.010798630304634571, "epoch": 32.59892086330935, "step": 36250 }, { "epoch": 32.59892086330935, "step": 36250, "torque_loss": 0.2391314059495926 }, { "epoch": 32.60791366906475, "grad_norm": 0.072246253490448, "learning_rate": 1.9265883047367656e-05, "loss": 0.02, "step": 36260 }, { "action_loss": 0.0015649175038561225, "epoch": 32.60791366906475, "step": 36260 }, { "epoch": 32.60791366906475, "step": 36260, "torque_loss": 0.1806374341249466 }, { "epoch": 32.61690647482014, "grad_norm": 0.11028102040290833, "learning_rate": 1.9239805466169748e-05, "loss": 0.0206, "step": 36270 }, { "action_loss": 0.0035409412812441587, "epoch": 32.61690647482014, "step": 36270 }, { "epoch": 32.61690647482014, "step": 36270, "torque_loss": 0.22289632260799408 }, { "epoch": 32.62589928057554, "grad_norm": 0.13077889382839203, "learning_rate": 1.9213741340525992e-05, "loss": 0.0254, "step": 36280 }, { "action_loss": 0.005709500517696142, "epoch": 32.62589928057554, "step": 36280 }, { "epoch": 32.62589928057554, "step": 36280, "torque_loss": 0.19990338385105133 }, { "epoch": 32.634892086330936, "grad_norm": 0.09825698286294937, "learning_rate": 1.9187690681837713e-05, "loss": 0.018, "step": 36290 }, { "action_loss": 0.00536083010956645, "epoch": 32.634892086330936, "step": 36290 }, { "epoch": 32.634892086330936, "step": 36290, "torque_loss": 0.2580653727054596 }, { "epoch": 32.643884892086334, "grad_norm": 0.15290983021259308, "learning_rate": 1.916165350150035e-05, "loss": 0.0185, "step": 36300 }, { "action_loss": 0.001500776968896389, "epoch": 32.643884892086334, "step": 36300 }, { "epoch": 32.643884892086334, "step": 36300, "torque_loss": 0.09950516372919083 }, { "epoch": 32.652877697841724, "grad_norm": 0.0695805549621582, "learning_rate": 1.9135629810903434e-05, "loss": 0.0182, "step": 36310 }, { "action_loss": 0.0062024760991334915, "epoch": 32.652877697841724, "step": 36310 }, { "epoch": 32.652877697841724, "step": 36310, "torque_loss": 0.18199540674686432 }, { "epoch": 32.66187050359712, "grad_norm": 0.11475037783384323, "learning_rate": 1.9109619621430648e-05, "loss": 0.0245, "step": 36320 }, { "action_loss": 0.0007527169655077159, "epoch": 32.66187050359712, "step": 36320 }, { "epoch": 32.66187050359712, "step": 36320, "torque_loss": 0.08482670783996582 }, { "epoch": 32.67086330935252, "grad_norm": 0.11214625835418701, "learning_rate": 1.9083622944459712e-05, "loss": 0.0154, "step": 36330 }, { "action_loss": 0.00401442451402545, "epoch": 32.67086330935252, "step": 36330 }, { "epoch": 32.67086330935252, "step": 36330, "torque_loss": 0.19549775123596191 }, { "epoch": 32.67985611510792, "grad_norm": 0.08411579579114914, "learning_rate": 1.9057639791362437e-05, "loss": 0.0188, "step": 36340 }, { "action_loss": 0.0009330532629974186, "epoch": 32.67985611510792, "step": 36340 }, { "epoch": 32.67985611510792, "step": 36340, "torque_loss": 0.12906239926815033 }, { "epoch": 32.68884892086331, "grad_norm": 0.07354149222373962, "learning_rate": 1.903167017350478e-05, "loss": 0.0182, "step": 36350 }, { "action_loss": 0.0009261040831916034, "epoch": 32.68884892086331, "step": 36350 }, { "epoch": 32.68884892086331, "step": 36350, "torque_loss": 0.15490998327732086 }, { "epoch": 32.697841726618705, "grad_norm": 0.14789697527885437, "learning_rate": 1.9005714102246707e-05, "loss": 0.023, "step": 36360 }, { "action_loss": 0.0018919106805697083, "epoch": 32.697841726618705, "step": 36360 }, { "epoch": 32.697841726618705, "step": 36360, "torque_loss": 0.11545442789793015 }, { "epoch": 32.7068345323741, "grad_norm": 0.06897977739572525, "learning_rate": 1.897977158894227e-05, "loss": 0.0214, "step": 36370 }, { "action_loss": 0.002668382367119193, "epoch": 32.7068345323741, "step": 36370 }, { "epoch": 32.7068345323741, "step": 36370, "torque_loss": 0.18526311218738556 }, { "epoch": 32.71582733812949, "grad_norm": 0.19017943739891052, "learning_rate": 1.8953842644939644e-05, "loss": 0.0189, "step": 36380 }, { "action_loss": 0.001665702904574573, "epoch": 32.71582733812949, "step": 36380 }, { "epoch": 32.71582733812949, "step": 36380, "torque_loss": 0.11420151591300964 }, { "epoch": 32.72482014388489, "grad_norm": 0.07308471202850342, "learning_rate": 1.8927927281581015e-05, "loss": 0.0163, "step": 36390 }, { "action_loss": 0.008680100552737713, "epoch": 32.72482014388489, "step": 36390 }, { "epoch": 32.72482014388489, "step": 36390, "torque_loss": 0.19102852046489716 }, { "epoch": 32.73381294964029, "grad_norm": 0.05895889177918434, "learning_rate": 1.890202551020262e-05, "loss": 0.0177, "step": 36400 }, { "action_loss": 0.0011807478731498122, "epoch": 32.73381294964029, "step": 36400 }, { "epoch": 32.73381294964029, "step": 36400, "torque_loss": 0.14112354815006256 }, { "epoch": 32.742805755395686, "grad_norm": 0.0465964674949646, "learning_rate": 1.8876137342134813e-05, "loss": 0.0163, "step": 36410 }, { "action_loss": 0.0021074069663882256, "epoch": 32.742805755395686, "step": 36410 }, { "epoch": 32.742805755395686, "step": 36410, "torque_loss": 0.16278283298015594 }, { "epoch": 32.75179856115108, "grad_norm": 0.06672060489654541, "learning_rate": 1.885026278870194e-05, "loss": 0.0188, "step": 36420 }, { "action_loss": 0.0015693440800532699, "epoch": 32.75179856115108, "step": 36420 }, { "epoch": 32.75179856115108, "step": 36420, "torque_loss": 0.12718592584133148 }, { "epoch": 32.760791366906474, "grad_norm": 0.09254632890224457, "learning_rate": 1.8824401861222384e-05, "loss": 0.0181, "step": 36430 }, { "action_loss": 0.001159978099167347, "epoch": 32.760791366906474, "step": 36430 }, { "epoch": 32.760791366906474, "step": 36430, "torque_loss": 0.09709224849939346 }, { "epoch": 32.76978417266187, "grad_norm": 0.06876827031373978, "learning_rate": 1.879855457100864e-05, "loss": 0.0213, "step": 36440 }, { "action_loss": 0.000995486625470221, "epoch": 32.76978417266187, "step": 36440 }, { "epoch": 32.76978417266187, "step": 36440, "torque_loss": 0.14130625128746033 }, { "epoch": 32.77877697841727, "grad_norm": 0.057219550013542175, "learning_rate": 1.877272092936717e-05, "loss": 0.0175, "step": 36450 }, { "action_loss": 0.0033733586315065622, "epoch": 32.77877697841727, "step": 36450 }, { "epoch": 32.77877697841727, "step": 36450, "torque_loss": 0.18551643192768097 }, { "epoch": 32.78776978417266, "grad_norm": 0.05733375623822212, "learning_rate": 1.8746900947598467e-05, "loss": 0.0232, "step": 36460 }, { "action_loss": 0.0031862102914601564, "epoch": 32.78776978417266, "step": 36460 }, { "epoch": 32.78776978417266, "step": 36460, "torque_loss": 0.13739515841007233 }, { "epoch": 32.79676258992806, "grad_norm": 0.09109096974134445, "learning_rate": 1.87210946369971e-05, "loss": 0.0205, "step": 36470 }, { "action_loss": 0.005097666289657354, "epoch": 32.79676258992806, "step": 36470 }, { "epoch": 32.79676258992806, "step": 36470, "torque_loss": 0.17342901229858398 }, { "epoch": 32.805755395683455, "grad_norm": 0.10270794481039047, "learning_rate": 1.8695302008851602e-05, "loss": 0.0207, "step": 36480 }, { "action_loss": 0.008462046273052692, "epoch": 32.805755395683455, "step": 36480 }, { "epoch": 32.805755395683455, "step": 36480, "torque_loss": 0.18644340336322784 }, { "epoch": 32.814748201438846, "grad_norm": 0.0624672994017601, "learning_rate": 1.8669523074444568e-05, "loss": 0.021, "step": 36490 }, { "action_loss": 0.0015753010520711541, "epoch": 32.814748201438846, "step": 36490 }, { "epoch": 32.814748201438846, "step": 36490, "torque_loss": 0.18851858377456665 }, { "epoch": 32.82374100719424, "grad_norm": 0.12072528898715973, "learning_rate": 1.8643757845052545e-05, "loss": 0.0222, "step": 36500 }, { "action_loss": 0.003434262238442898, "epoch": 32.82374100719424, "step": 36500 }, { "epoch": 32.82374100719424, "step": 36500, "torque_loss": 0.12422550469636917 }, { "epoch": 32.83273381294964, "grad_norm": 0.07149874418973923, "learning_rate": 1.8618006331946164e-05, "loss": 0.019, "step": 36510 }, { "action_loss": 0.001641064533032477, "epoch": 32.83273381294964, "step": 36510 }, { "epoch": 32.83273381294964, "step": 36510, "torque_loss": 0.09685685485601425 }, { "epoch": 32.84172661870504, "grad_norm": 0.07226046174764633, "learning_rate": 1.859226854638999e-05, "loss": 0.016, "step": 36520 }, { "action_loss": 0.0020266342908143997, "epoch": 32.84172661870504, "step": 36520 }, { "epoch": 32.84172661870504, "step": 36520, "torque_loss": 0.17410512268543243 }, { "epoch": 32.85071942446043, "grad_norm": 0.09949751943349838, "learning_rate": 1.8566544499642587e-05, "loss": 0.0213, "step": 36530 }, { "action_loss": 0.0012045842595398426, "epoch": 32.85071942446043, "step": 36530 }, { "epoch": 32.85071942446043, "step": 36530, "torque_loss": 0.17665092647075653 }, { "epoch": 32.85971223021583, "grad_norm": 0.0713317021727562, "learning_rate": 1.854083420295656e-05, "loss": 0.018, "step": 36540 }, { "action_loss": 0.0021777048241347075, "epoch": 32.85971223021583, "step": 36540 }, { "epoch": 32.85971223021583, "step": 36540, "torque_loss": 0.13701654970645905 }, { "epoch": 32.868705035971225, "grad_norm": 0.11089528352022171, "learning_rate": 1.8515137667578453e-05, "loss": 0.0197, "step": 36550 }, { "action_loss": 0.001238802564330399, "epoch": 32.868705035971225, "step": 36550 }, { "epoch": 32.868705035971225, "step": 36550, "torque_loss": 0.14154265820980072 }, { "epoch": 32.87769784172662, "grad_norm": 0.06316062062978745, "learning_rate": 1.8489454904748784e-05, "loss": 0.0148, "step": 36560 }, { "action_loss": 0.0012261865194886923, "epoch": 32.87769784172662, "step": 36560 }, { "epoch": 32.87769784172662, "step": 36560, "torque_loss": 0.10627686977386475 }, { "epoch": 32.88669064748201, "grad_norm": 0.10069926828145981, "learning_rate": 1.8463785925702094e-05, "loss": 0.0201, "step": 36570 }, { "action_loss": 0.003463905304670334, "epoch": 32.88669064748201, "step": 36570 }, { "epoch": 32.88669064748201, "step": 36570, "torque_loss": 0.16760586202144623 }, { "epoch": 32.89568345323741, "grad_norm": 0.08628041297197342, "learning_rate": 1.843813074166686e-05, "loss": 0.0223, "step": 36580 }, { "action_loss": 0.003751936601474881, "epoch": 32.89568345323741, "step": 36580 }, { "epoch": 32.89568345323741, "step": 36580, "torque_loss": 0.16206969320774078 }, { "epoch": 32.90467625899281, "grad_norm": 0.06974107027053833, "learning_rate": 1.841248936386551e-05, "loss": 0.0212, "step": 36590 }, { "action_loss": 0.006412104237824678, "epoch": 32.90467625899281, "step": 36590 }, { "epoch": 32.90467625899281, "step": 36590, "torque_loss": 0.20836083590984344 }, { "epoch": 32.9136690647482, "grad_norm": 0.06989284604787827, "learning_rate": 1.8386861803514483e-05, "loss": 0.0218, "step": 36600 }, { "action_loss": 0.0024074276443570852, "epoch": 32.9136690647482, "step": 36600 }, { "epoch": 32.9136690647482, "step": 36600, "torque_loss": 0.11295616626739502 }, { "epoch": 32.922661870503596, "grad_norm": 0.08954238891601562, "learning_rate": 1.836124807182414e-05, "loss": 0.0185, "step": 36610 }, { "action_loss": 0.0017289072275161743, "epoch": 32.922661870503596, "step": 36610 }, { "epoch": 32.922661870503596, "step": 36610, "torque_loss": 0.127568319439888 }, { "epoch": 32.931654676258994, "grad_norm": 0.07537220418453217, "learning_rate": 1.8335648179998783e-05, "loss": 0.0191, "step": 36620 }, { "action_loss": 0.002276795916259289, "epoch": 32.931654676258994, "step": 36620 }, { "epoch": 32.931654676258994, "step": 36620, "torque_loss": 0.17983084917068481 }, { "epoch": 32.94064748201439, "grad_norm": 0.05469014123082161, "learning_rate": 1.8310062139236667e-05, "loss": 0.0169, "step": 36630 }, { "action_loss": 0.0027126662898808718, "epoch": 32.94064748201439, "step": 36630 }, { "epoch": 32.94064748201439, "step": 36630, "torque_loss": 0.21271200478076935 }, { "epoch": 32.94964028776978, "grad_norm": 0.05765775591135025, "learning_rate": 1.828448996073002e-05, "loss": 0.0166, "step": 36640 }, { "action_loss": 0.0015948130749166012, "epoch": 32.94964028776978, "step": 36640 }, { "epoch": 32.94964028776978, "step": 36640, "torque_loss": 0.18010596930980682 }, { "epoch": 32.95863309352518, "grad_norm": 0.1063721626996994, "learning_rate": 1.825893165566498e-05, "loss": 0.0184, "step": 36650 }, { "action_loss": 0.0007900021155364811, "epoch": 32.95863309352518, "step": 36650 }, { "epoch": 32.95863309352518, "step": 36650, "torque_loss": 0.17204158008098602 }, { "epoch": 32.96762589928058, "grad_norm": 0.11845213174819946, "learning_rate": 1.823338723522159e-05, "loss": 0.0203, "step": 36660 }, { "action_loss": 0.0007847709930501878, "epoch": 32.96762589928058, "step": 36660 }, { "epoch": 32.96762589928058, "step": 36660, "torque_loss": 0.09973279386758804 }, { "epoch": 32.976618705035975, "grad_norm": 0.06562856584787369, "learning_rate": 1.8207856710573878e-05, "loss": 0.0177, "step": 36670 }, { "action_loss": 0.0010565707925707102, "epoch": 32.976618705035975, "step": 36670 }, { "epoch": 32.976618705035975, "step": 36670, "torque_loss": 0.16654138267040253 }, { "epoch": 32.985611510791365, "grad_norm": 0.08274702727794647, "learning_rate": 1.8182340092889756e-05, "loss": 0.0185, "step": 36680 }, { "action_loss": 0.0006667642737738788, "epoch": 32.985611510791365, "step": 36680 }, { "epoch": 32.985611510791365, "step": 36680, "torque_loss": 0.1189330592751503 }, { "epoch": 32.99460431654676, "grad_norm": 0.07869885116815567, "learning_rate": 1.8156837393331038e-05, "loss": 0.0195, "step": 36690 }, { "action_loss": 0.002245419891551137, "epoch": 32.99460431654676, "step": 36690 }, { "epoch": 32.99460431654676, "step": 36690, "torque_loss": 0.11832670122385025 }, { "epoch": 33.00359712230216, "grad_norm": 0.15725499391555786, "learning_rate": 1.81313486230535e-05, "loss": 0.0182, "step": 36700 }, { "action_loss": 0.004658807534724474, "epoch": 33.00359712230216, "step": 36700 }, { "epoch": 33.00359712230216, "step": 36700, "torque_loss": 0.19889603555202484 }, { "epoch": 33.01258992805755, "grad_norm": 0.10122749209403992, "learning_rate": 1.8105873793206773e-05, "loss": 0.0188, "step": 36710 }, { "action_loss": 0.0013945760438218713, "epoch": 33.01258992805755, "step": 36710 }, { "epoch": 33.01258992805755, "step": 36710, "torque_loss": 0.11745864152908325 }, { "epoch": 33.02158273381295, "grad_norm": 0.09415537863969803, "learning_rate": 1.8080412914934436e-05, "loss": 0.0219, "step": 36720 }, { "action_loss": 0.006458505988121033, "epoch": 33.02158273381295, "step": 36720 }, { "epoch": 33.02158273381295, "step": 36720, "torque_loss": 0.18400216102600098 }, { "epoch": 33.030575539568346, "grad_norm": 0.07915939390659332, "learning_rate": 1.805496599937392e-05, "loss": 0.0184, "step": 36730 }, { "action_loss": 0.0054591563530266285, "epoch": 33.030575539568346, "step": 36730 }, { "epoch": 33.030575539568346, "step": 36730, "torque_loss": 0.15212951600551605 }, { "epoch": 33.039568345323744, "grad_norm": 0.06773534417152405, "learning_rate": 1.802953305765659e-05, "loss": 0.0256, "step": 36740 }, { "action_loss": 0.0027074571698904037, "epoch": 33.039568345323744, "step": 36740 }, { "epoch": 33.039568345323744, "step": 36740, "torque_loss": 0.11004626750946045 }, { "epoch": 33.048561151079134, "grad_norm": 0.08843197673559189, "learning_rate": 1.8004114100907653e-05, "loss": 0.0175, "step": 36750 }, { "action_loss": 0.006959015969187021, "epoch": 33.048561151079134, "step": 36750 }, { "epoch": 33.048561151079134, "step": 36750, "torque_loss": 0.2118672877550125 }, { "epoch": 33.05755395683453, "grad_norm": 0.12264720350503922, "learning_rate": 1.7978709140246263e-05, "loss": 0.0217, "step": 36760 }, { "action_loss": 0.0011374199530109763, "epoch": 33.05755395683453, "step": 36760 }, { "epoch": 33.05755395683453, "step": 36760, "torque_loss": 0.14387647807598114 }, { "epoch": 33.06654676258993, "grad_norm": 0.11571577191352844, "learning_rate": 1.795331818678539e-05, "loss": 0.0178, "step": 36770 }, { "action_loss": 0.0012992293341085315, "epoch": 33.06654676258993, "step": 36770 }, { "epoch": 33.06654676258993, "step": 36770, "torque_loss": 0.11185828596353531 }, { "epoch": 33.07553956834533, "grad_norm": 0.05895185098052025, "learning_rate": 1.792794125163188e-05, "loss": 0.0185, "step": 36780 }, { "action_loss": 0.0019678848329931498, "epoch": 33.07553956834533, "step": 36780 }, { "epoch": 33.07553956834533, "step": 36780, "torque_loss": 0.13763023912906647 }, { "epoch": 33.08453237410072, "grad_norm": 0.10362949967384338, "learning_rate": 1.7902578345886507e-05, "loss": 0.0242, "step": 36790 }, { "action_loss": 0.0020110716577619314, "epoch": 33.08453237410072, "step": 36790 }, { "epoch": 33.08453237410072, "step": 36790, "torque_loss": 0.16955487430095673 }, { "epoch": 33.093525179856115, "grad_norm": 0.0755833089351654, "learning_rate": 1.7877229480643843e-05, "loss": 0.0178, "step": 36800 }, { "action_loss": 0.06603357940912247, "epoch": 33.093525179856115, "step": 36800 }, { "epoch": 33.093525179856115, "step": 36800, "torque_loss": 0.2053154855966568 }, { "epoch": 33.10251798561151, "grad_norm": 0.05768866837024689, "learning_rate": 1.785189466699235e-05, "loss": 0.024, "step": 36810 }, { "action_loss": 0.0015378614189103246, "epoch": 33.10251798561151, "step": 36810 }, { "epoch": 33.10251798561151, "step": 36810, "torque_loss": 0.13801364600658417 }, { "epoch": 33.111510791366904, "grad_norm": 0.11623333394527435, "learning_rate": 1.7826573916014317e-05, "loss": 0.0202, "step": 36820 }, { "action_loss": 0.004335507284849882, "epoch": 33.111510791366904, "step": 36820 }, { "epoch": 33.111510791366904, "step": 36820, "torque_loss": 0.16872815787792206 }, { "epoch": 33.1205035971223, "grad_norm": 0.13951340317726135, "learning_rate": 1.7801267238785935e-05, "loss": 0.021, "step": 36830 }, { "action_loss": 0.0027979512233287096, "epoch": 33.1205035971223, "step": 36830 }, { "epoch": 33.1205035971223, "step": 36830, "torque_loss": 0.1251920461654663 }, { "epoch": 33.1294964028777, "grad_norm": 0.09098456799983978, "learning_rate": 1.7775974646377187e-05, "loss": 0.0178, "step": 36840 }, { "action_loss": 0.0017414394533261657, "epoch": 33.1294964028777, "step": 36840 }, { "epoch": 33.1294964028777, "step": 36840, "torque_loss": 0.1475883275270462 }, { "epoch": 33.138489208633096, "grad_norm": 0.14173200726509094, "learning_rate": 1.775069614985189e-05, "loss": 0.0205, "step": 36850 }, { "action_loss": 0.013014880008995533, "epoch": 33.138489208633096, "step": 36850 }, { "epoch": 33.138489208633096, "step": 36850, "torque_loss": 0.19355343282222748 }, { "epoch": 33.14748201438849, "grad_norm": 0.09647049754858017, "learning_rate": 1.7725431760267757e-05, "loss": 0.0204, "step": 36860 }, { "action_loss": 0.0016385390190407634, "epoch": 33.14748201438849, "step": 36860 }, { "epoch": 33.14748201438849, "step": 36860, "torque_loss": 0.1712220162153244 }, { "epoch": 33.156474820143885, "grad_norm": 0.0633731409907341, "learning_rate": 1.7700181488676277e-05, "loss": 0.0182, "step": 36870 }, { "action_loss": 0.0014817602932453156, "epoch": 33.156474820143885, "step": 36870 }, { "epoch": 33.156474820143885, "step": 36870, "torque_loss": 0.18013028800487518 }, { "epoch": 33.16546762589928, "grad_norm": 0.10208921134471893, "learning_rate": 1.767494534612275e-05, "loss": 0.0201, "step": 36880 }, { "action_loss": 0.008546584285795689, "epoch": 33.16546762589928, "step": 36880 }, { "epoch": 33.16546762589928, "step": 36880, "torque_loss": 0.22605770826339722 }, { "epoch": 33.17446043165467, "grad_norm": 0.0973130539059639, "learning_rate": 1.7649723343646367e-05, "loss": 0.019, "step": 36890 }, { "action_loss": 0.009601856581866741, "epoch": 33.17446043165467, "step": 36890 }, { "epoch": 33.17446043165467, "step": 36890, "torque_loss": 0.19981533288955688 }, { "epoch": 33.18345323741007, "grad_norm": 0.08395054936408997, "learning_rate": 1.7624515492280063e-05, "loss": 0.019, "step": 36900 }, { "action_loss": 0.023742442950606346, "epoch": 33.18345323741007, "step": 36900 }, { "epoch": 33.18345323741007, "step": 36900, "torque_loss": 0.1995820552110672 }, { "epoch": 33.19244604316547, "grad_norm": 0.1025146096944809, "learning_rate": 1.7599321803050596e-05, "loss": 0.0218, "step": 36910 }, { "action_loss": 0.002794584957882762, "epoch": 33.19244604316547, "step": 36910 }, { "epoch": 33.19244604316547, "step": 36910, "torque_loss": 0.16859792172908783 }, { "epoch": 33.201438848920866, "grad_norm": 0.07475409656763077, "learning_rate": 1.7574142286978574e-05, "loss": 0.0202, "step": 36920 }, { "action_loss": 0.0029258409049361944, "epoch": 33.201438848920866, "step": 36920 }, { "epoch": 33.201438848920866, "step": 36920, "torque_loss": 0.13163602352142334 }, { "epoch": 33.210431654676256, "grad_norm": 0.11417053639888763, "learning_rate": 1.7548976955078354e-05, "loss": 0.0192, "step": 36930 }, { "action_loss": 0.0012018346460536122, "epoch": 33.210431654676256, "step": 36930 }, { "epoch": 33.210431654676256, "step": 36930, "torque_loss": 0.16980160772800446 }, { "epoch": 33.219424460431654, "grad_norm": 0.08601208031177521, "learning_rate": 1.7523825818358092e-05, "loss": 0.0183, "step": 36940 }, { "action_loss": 0.0011593123199418187, "epoch": 33.219424460431654, "step": 36940 }, { "epoch": 33.219424460431654, "step": 36940, "torque_loss": 0.2094997614622116 }, { "epoch": 33.22841726618705, "grad_norm": 0.045143984258174896, "learning_rate": 1.7498688887819785e-05, "loss": 0.0162, "step": 36950 }, { "action_loss": 0.0010874910512939095, "epoch": 33.22841726618705, "step": 36950 }, { "epoch": 33.22841726618705, "step": 36950, "torque_loss": 0.12693153321743011 }, { "epoch": 33.23741007194245, "grad_norm": 0.10156385600566864, "learning_rate": 1.7473566174459143e-05, "loss": 0.0179, "step": 36960 }, { "action_loss": 0.0037702813278883696, "epoch": 33.23741007194245, "step": 36960 }, { "epoch": 33.23741007194245, "step": 36960, "torque_loss": 0.15969443321228027 }, { "epoch": 33.24640287769784, "grad_norm": 0.132468119263649, "learning_rate": 1.7448457689265735e-05, "loss": 0.0212, "step": 36970 }, { "action_loss": 0.007933579385280609, "epoch": 33.24640287769784, "step": 36970 }, { "epoch": 33.24640287769784, "step": 36970, "torque_loss": 0.17577986419200897 }, { "epoch": 33.25539568345324, "grad_norm": 0.1046718880534172, "learning_rate": 1.7423363443222823e-05, "loss": 0.0182, "step": 36980 }, { "action_loss": 0.004219569731503725, "epoch": 33.25539568345324, "step": 36980 }, { "epoch": 33.25539568345324, "step": 36980, "torque_loss": 0.18709838390350342 }, { "epoch": 33.264388489208635, "grad_norm": 0.11962137371301651, "learning_rate": 1.7398283447307524e-05, "loss": 0.0202, "step": 36990 }, { "action_loss": 0.003466907190158963, "epoch": 33.264388489208635, "step": 36990 }, { "epoch": 33.264388489208635, "step": 36990, "torque_loss": 0.17106784880161285 }, { "epoch": 33.273381294964025, "grad_norm": 0.06545446068048477, "learning_rate": 1.737321771249066e-05, "loss": 0.0186, "step": 37000 }, { "action_loss": 0.0014804662205278873, "epoch": 33.273381294964025, "step": 37000 }, { "epoch": 33.273381294964025, "step": 37000, "torque_loss": 0.19448888301849365 }, { "epoch": 33.28237410071942, "grad_norm": 0.1166982352733612, "learning_rate": 1.7348166249736814e-05, "loss": 0.0224, "step": 37010 }, { "action_loss": 0.002159177092835307, "epoch": 33.28237410071942, "step": 37010 }, { "epoch": 33.28237410071942, "step": 37010, "torque_loss": 0.14498640596866608 }, { "epoch": 33.29136690647482, "grad_norm": 0.09712051600217819, "learning_rate": 1.732312907000439e-05, "loss": 0.0235, "step": 37020 }, { "action_loss": 0.0018170783296227455, "epoch": 33.29136690647482, "step": 37020 }, { "epoch": 33.29136690647482, "step": 37020, "torque_loss": 0.15684598684310913 }, { "epoch": 33.30035971223022, "grad_norm": 0.06600018590688705, "learning_rate": 1.7298106184245478e-05, "loss": 0.0196, "step": 37030 }, { "action_loss": 0.0007807340589351952, "epoch": 33.30035971223022, "step": 37030 }, { "epoch": 33.30035971223022, "step": 37030, "torque_loss": 0.13268275558948517 }, { "epoch": 33.30935251798561, "grad_norm": 0.12401936203241348, "learning_rate": 1.727309760340592e-05, "loss": 0.0175, "step": 37040 }, { "action_loss": 0.0020140926353633404, "epoch": 33.30935251798561, "step": 37040 }, { "epoch": 33.30935251798561, "step": 37040, "torque_loss": 0.16995126008987427 }, { "epoch": 33.318345323741006, "grad_norm": 0.07438317686319351, "learning_rate": 1.724810333842536e-05, "loss": 0.0186, "step": 37050 }, { "action_loss": 0.0026403069496154785, "epoch": 33.318345323741006, "step": 37050 }, { "epoch": 33.318345323741006, "step": 37050, "torque_loss": 0.18740206956863403 }, { "epoch": 33.327338129496404, "grad_norm": 0.13722747564315796, "learning_rate": 1.7223123400237112e-05, "loss": 0.0218, "step": 37060 }, { "action_loss": 0.0015337580116465688, "epoch": 33.327338129496404, "step": 37060 }, { "epoch": 33.327338129496404, "step": 37060, "torque_loss": 0.1360255628824234 }, { "epoch": 33.3363309352518, "grad_norm": 0.11531494557857513, "learning_rate": 1.7198157799768238e-05, "loss": 0.0244, "step": 37070 }, { "action_loss": 0.006489959079772234, "epoch": 33.3363309352518, "step": 37070 }, { "epoch": 33.3363309352518, "step": 37070, "torque_loss": 0.19970698654651642 }, { "epoch": 33.34532374100719, "grad_norm": 0.0920199379324913, "learning_rate": 1.7173206547939576e-05, "loss": 0.0233, "step": 37080 }, { "action_loss": 0.0013743456220254302, "epoch": 33.34532374100719, "step": 37080 }, { "epoch": 33.34532374100719, "step": 37080, "torque_loss": 0.143948033452034 }, { "epoch": 33.35431654676259, "grad_norm": 0.1270473599433899, "learning_rate": 1.7148269655665628e-05, "loss": 0.0193, "step": 37090 }, { "action_loss": 0.008460345678031445, "epoch": 33.35431654676259, "step": 37090 }, { "epoch": 33.35431654676259, "step": 37090, "torque_loss": 0.18803291022777557 }, { "epoch": 33.36330935251799, "grad_norm": 0.11461339890956879, "learning_rate": 1.712334713385463e-05, "loss": 0.0166, "step": 37100 }, { "action_loss": 0.0005513170617632568, "epoch": 33.36330935251799, "step": 37100 }, { "epoch": 33.36330935251799, "step": 37100, "torque_loss": 0.10444110631942749 }, { "epoch": 33.37230215827338, "grad_norm": 0.07465025037527084, "learning_rate": 1.709843899340856e-05, "loss": 0.0221, "step": 37110 }, { "action_loss": 0.003719265339896083, "epoch": 33.37230215827338, "step": 37110 }, { "epoch": 33.37230215827338, "step": 37110, "torque_loss": 0.13774190843105316 }, { "epoch": 33.381294964028775, "grad_norm": 0.12658821046352386, "learning_rate": 1.7073545245223084e-05, "loss": 0.0212, "step": 37120 }, { "action_loss": 0.0011840014485642314, "epoch": 33.381294964028775, "step": 37120 }, { "epoch": 33.381294964028775, "step": 37120, "torque_loss": 0.1735852211713791 }, { "epoch": 33.39028776978417, "grad_norm": 0.11020264029502869, "learning_rate": 1.704866590018755e-05, "loss": 0.0196, "step": 37130 }, { "action_loss": 0.002191381761804223, "epoch": 33.39028776978417, "step": 37130 }, { "epoch": 33.39028776978417, "step": 37130, "torque_loss": 0.14424680173397064 }, { "epoch": 33.39928057553957, "grad_norm": 0.09365174919366837, "learning_rate": 1.7023800969185054e-05, "loss": 0.0202, "step": 37140 }, { "action_loss": 0.001041674637235701, "epoch": 33.39928057553957, "step": 37140 }, { "epoch": 33.39928057553957, "step": 37140, "torque_loss": 0.09457748383283615 }, { "epoch": 33.40827338129496, "grad_norm": 0.05832100287079811, "learning_rate": 1.6998950463092354e-05, "loss": 0.0134, "step": 37150 }, { "action_loss": 0.0023935965728014708, "epoch": 33.40827338129496, "step": 37150 }, { "epoch": 33.40827338129496, "step": 37150, "torque_loss": 0.2516269385814667 }, { "epoch": 33.41726618705036, "grad_norm": 0.08347023278474808, "learning_rate": 1.6974114392779907e-05, "loss": 0.0229, "step": 37160 }, { "action_loss": 0.004984689876437187, "epoch": 33.41726618705036, "step": 37160 }, { "epoch": 33.41726618705036, "step": 37160, "torque_loss": 0.20474527776241302 }, { "epoch": 33.42625899280576, "grad_norm": 0.07835441082715988, "learning_rate": 1.694929276911183e-05, "loss": 0.0207, "step": 37170 }, { "action_loss": 0.0032929799053817987, "epoch": 33.42625899280576, "step": 37170 }, { "epoch": 33.42625899280576, "step": 37170, "torque_loss": 0.22164969146251678 }, { "epoch": 33.435251798561154, "grad_norm": 0.0988234281539917, "learning_rate": 1.6924485602945983e-05, "loss": 0.0212, "step": 37180 }, { "action_loss": 0.00349435955286026, "epoch": 33.435251798561154, "step": 37180 }, { "epoch": 33.435251798561154, "step": 37180, "torque_loss": 0.2059825211763382 }, { "epoch": 33.444244604316545, "grad_norm": 0.10243342816829681, "learning_rate": 1.6899692905133825e-05, "loss": 0.0203, "step": 37190 }, { "action_loss": 0.0011820938671007752, "epoch": 33.444244604316545, "step": 37190 }, { "epoch": 33.444244604316545, "step": 37190, "torque_loss": 0.11919966340065002 }, { "epoch": 33.45323741007194, "grad_norm": 0.15820275247097015, "learning_rate": 1.6874914686520565e-05, "loss": 0.0205, "step": 37200 }, { "action_loss": 0.001725311973132193, "epoch": 33.45323741007194, "step": 37200 }, { "epoch": 33.45323741007194, "step": 37200, "torque_loss": 0.12694954872131348 }, { "epoch": 33.46223021582734, "grad_norm": 0.17697706818580627, "learning_rate": 1.6850150957945e-05, "loss": 0.0222, "step": 37210 }, { "action_loss": 0.001566090271808207, "epoch": 33.46223021582734, "step": 37210 }, { "epoch": 33.46223021582734, "step": 37210, "torque_loss": 0.1174796000123024 }, { "epoch": 33.47122302158273, "grad_norm": 0.15338116884231567, "learning_rate": 1.682540173023966e-05, "loss": 0.0191, "step": 37220 }, { "action_loss": 0.0008998709381558001, "epoch": 33.47122302158273, "step": 37220 }, { "epoch": 33.47122302158273, "step": 37220, "torque_loss": 0.13248474895954132 }, { "epoch": 33.48021582733813, "grad_norm": 0.07847344130277634, "learning_rate": 1.6800667014230665e-05, "loss": 0.0166, "step": 37230 }, { "action_loss": 0.001247177249751985, "epoch": 33.48021582733813, "step": 37230 }, { "epoch": 33.48021582733813, "step": 37230, "torque_loss": 0.15848080813884735 }, { "epoch": 33.489208633093526, "grad_norm": 0.12960998713970184, "learning_rate": 1.677594682073785e-05, "loss": 0.018, "step": 37240 }, { "action_loss": 0.002195250242948532, "epoch": 33.489208633093526, "step": 37240 }, { "epoch": 33.489208633093526, "step": 37240, "torque_loss": 0.19278855621814728 }, { "epoch": 33.49820143884892, "grad_norm": 0.10542070865631104, "learning_rate": 1.6751241160574653e-05, "loss": 0.0194, "step": 37250 }, { "action_loss": 0.0022449989337474108, "epoch": 33.49820143884892, "step": 37250 }, { "epoch": 33.49820143884892, "step": 37250, "torque_loss": 0.16256965696811676 }, { "epoch": 33.507194244604314, "grad_norm": 0.05987897515296936, "learning_rate": 1.6726550044548156e-05, "loss": 0.0195, "step": 37260 }, { "action_loss": 0.002443272853270173, "epoch": 33.507194244604314, "step": 37260 }, { "epoch": 33.507194244604314, "step": 37260, "torque_loss": 0.18772895634174347 }, { "epoch": 33.51618705035971, "grad_norm": 0.12851965427398682, "learning_rate": 1.6701873483459125e-05, "loss": 0.0176, "step": 37270 }, { "action_loss": 0.0011806521797552705, "epoch": 33.51618705035971, "step": 37270 }, { "epoch": 33.51618705035971, "step": 37270, "torque_loss": 0.12234029173851013 }, { "epoch": 33.52517985611511, "grad_norm": 0.05005881190299988, "learning_rate": 1.6677211488101906e-05, "loss": 0.0185, "step": 37280 }, { "action_loss": 0.009731988422572613, "epoch": 33.52517985611511, "step": 37280 }, { "epoch": 33.52517985611511, "step": 37280, "torque_loss": 0.18419446051120758 }, { "epoch": 33.53417266187051, "grad_norm": 0.06814000755548477, "learning_rate": 1.6652564069264475e-05, "loss": 0.0196, "step": 37290 }, { "action_loss": 0.004277957137674093, "epoch": 33.53417266187051, "step": 37290 }, { "epoch": 33.53417266187051, "step": 37290, "torque_loss": 0.178883895277977 }, { "epoch": 33.5431654676259, "grad_norm": 0.07967919856309891, "learning_rate": 1.662793123772849e-05, "loss": 0.0184, "step": 37300 }, { "action_loss": 0.0012736866483464837, "epoch": 33.5431654676259, "step": 37300 }, { "epoch": 33.5431654676259, "step": 37300, "torque_loss": 0.1290474832057953 }, { "epoch": 33.552158273381295, "grad_norm": 0.12318191677331924, "learning_rate": 1.6603313004269168e-05, "loss": 0.0174, "step": 37310 }, { "action_loss": 0.002686013700440526, "epoch": 33.552158273381295, "step": 37310 }, { "epoch": 33.552158273381295, "step": 37310, "torque_loss": 0.13959567248821259 }, { "epoch": 33.56115107913669, "grad_norm": 0.11135230213403702, "learning_rate": 1.657870937965536e-05, "loss": 0.0198, "step": 37320 }, { "action_loss": 0.0008711711852811277, "epoch": 33.56115107913669, "step": 37320 }, { "epoch": 33.56115107913669, "step": 37320, "torque_loss": 0.09471788257360458 }, { "epoch": 33.57014388489208, "grad_norm": 0.12792359292507172, "learning_rate": 1.655412037464952e-05, "loss": 0.0233, "step": 37330 }, { "action_loss": 0.006293805781751871, "epoch": 33.57014388489208, "step": 37330 }, { "epoch": 33.57014388489208, "step": 37330, "torque_loss": 0.17746619880199432 }, { "epoch": 33.57913669064748, "grad_norm": 0.11652892827987671, "learning_rate": 1.652954600000773e-05, "loss": 0.0295, "step": 37340 }, { "action_loss": 0.0027557078283280134, "epoch": 33.57913669064748, "step": 37340 }, { "epoch": 33.57913669064748, "step": 37340, "torque_loss": 0.17052601277828217 }, { "epoch": 33.58812949640288, "grad_norm": 0.07793822139501572, "learning_rate": 1.650498626647965e-05, "loss": 0.0226, "step": 37350 }, { "action_loss": 0.0015995943685993552, "epoch": 33.58812949640288, "step": 37350 }, { "epoch": 33.58812949640288, "step": 37350, "torque_loss": 0.14258073270320892 }, { "epoch": 33.597122302158276, "grad_norm": 0.07535295188426971, "learning_rate": 1.6480441184808526e-05, "loss": 0.0157, "step": 37360 }, { "action_loss": 0.004378974437713623, "epoch": 33.597122302158276, "step": 37360 }, { "epoch": 33.597122302158276, "step": 37360, "torque_loss": 0.16032655537128448 }, { "epoch": 33.606115107913666, "grad_norm": 0.05543661490082741, "learning_rate": 1.6455910765731242e-05, "loss": 0.0203, "step": 37370 }, { "action_loss": 0.0031044490169733763, "epoch": 33.606115107913666, "step": 37370 }, { "epoch": 33.606115107913666, "step": 37370, "torque_loss": 0.16597478091716766 }, { "epoch": 33.615107913669064, "grad_norm": 0.11878691613674164, "learning_rate": 1.6431395019978217e-05, "loss": 0.0199, "step": 37380 }, { "action_loss": 0.003121551824733615, "epoch": 33.615107913669064, "step": 37380 }, { "epoch": 33.615107913669064, "step": 37380, "torque_loss": 0.14505301415920258 }, { "epoch": 33.62410071942446, "grad_norm": 0.11024442315101624, "learning_rate": 1.6406893958273467e-05, "loss": 0.0192, "step": 37390 }, { "action_loss": 0.002206997247412801, "epoch": 33.62410071942446, "step": 37390 }, { "epoch": 33.62410071942446, "step": 37390, "torque_loss": 0.15314684808254242 }, { "epoch": 33.63309352517986, "grad_norm": 0.06914316117763519, "learning_rate": 1.6382407591334602e-05, "loss": 0.0227, "step": 37400 }, { "action_loss": 0.0055460333824157715, "epoch": 33.63309352517986, "step": 37400 }, { "epoch": 33.63309352517986, "step": 37400, "torque_loss": 0.17152923345565796 }, { "epoch": 33.64208633093525, "grad_norm": 0.06548137217760086, "learning_rate": 1.635793592987279e-05, "loss": 0.0219, "step": 37410 }, { "action_loss": 0.001982274930924177, "epoch": 33.64208633093525, "step": 37410 }, { "epoch": 33.64208633093525, "step": 37410, "torque_loss": 0.13558794558048248 }, { "epoch": 33.65107913669065, "grad_norm": 0.11413434892892838, "learning_rate": 1.633347898459275e-05, "loss": 0.0176, "step": 37420 }, { "action_loss": 0.002193828346207738, "epoch": 33.65107913669065, "step": 37420 }, { "epoch": 33.65107913669065, "step": 37420, "torque_loss": 0.18776647746562958 }, { "epoch": 33.660071942446045, "grad_norm": 0.13055267930030823, "learning_rate": 1.6309036766192805e-05, "loss": 0.0196, "step": 37430 }, { "action_loss": 0.0018005819292739034, "epoch": 33.660071942446045, "step": 37430 }, { "epoch": 33.660071942446045, "step": 37430, "torque_loss": 0.14616835117340088 }, { "epoch": 33.669064748201436, "grad_norm": 0.07816451787948608, "learning_rate": 1.628460928536478e-05, "loss": 0.0176, "step": 37440 }, { "action_loss": 0.0024602848570793867, "epoch": 33.669064748201436, "step": 37440 }, { "epoch": 33.669064748201436, "step": 37440, "torque_loss": 0.17779600620269775 }, { "epoch": 33.67805755395683, "grad_norm": 0.06523854285478592, "learning_rate": 1.6260196552794128e-05, "loss": 0.0177, "step": 37450 }, { "action_loss": 0.0016644721617922187, "epoch": 33.67805755395683, "step": 37450 }, { "epoch": 33.67805755395683, "step": 37450, "torque_loss": 0.13347004354000092 }, { "epoch": 33.68705035971223, "grad_norm": 0.061941538006067276, "learning_rate": 1.6235798579159766e-05, "loss": 0.0172, "step": 37460 }, { "action_loss": 0.002025514841079712, "epoch": 33.68705035971223, "step": 37460 }, { "epoch": 33.68705035971223, "step": 37460, "torque_loss": 0.11380680650472641 }, { "epoch": 33.69604316546763, "grad_norm": 0.09505584836006165, "learning_rate": 1.6211415375134233e-05, "loss": 0.0199, "step": 37470 }, { "action_loss": 0.011529806070029736, "epoch": 33.69604316546763, "step": 37470 }, { "epoch": 33.69604316546763, "step": 37470, "torque_loss": 0.24294525384902954 }, { "epoch": 33.70503597122302, "grad_norm": 0.10085511207580566, "learning_rate": 1.618704695138353e-05, "loss": 0.0173, "step": 37480 }, { "action_loss": 0.0020968422759324312, "epoch": 33.70503597122302, "step": 37480 }, { "epoch": 33.70503597122302, "step": 37480, "torque_loss": 0.13023541867733002 }, { "epoch": 33.71402877697842, "grad_norm": 0.04492134600877762, "learning_rate": 1.6162693318567273e-05, "loss": 0.0167, "step": 37490 }, { "action_loss": 0.0018775895005092025, "epoch": 33.71402877697842, "step": 37490 }, { "epoch": 33.71402877697842, "step": 37490, "torque_loss": 0.11357460170984268 }, { "epoch": 33.723021582733814, "grad_norm": 0.06299616396427155, "learning_rate": 1.613835448733856e-05, "loss": 0.0173, "step": 37500 }, { "action_loss": 0.0023816099856048822, "epoch": 33.723021582733814, "step": 37500 }, { "epoch": 33.723021582733814, "step": 37500, "torque_loss": 0.17596890032291412 }, { "epoch": 33.73201438848921, "grad_norm": 0.04245999455451965, "learning_rate": 1.6114030468344006e-05, "loss": 0.0185, "step": 37510 }, { "action_loss": 0.006618810351938009, "epoch": 33.73201438848921, "step": 37510 }, { "epoch": 33.73201438848921, "step": 37510, "torque_loss": 0.19283480942249298 }, { "epoch": 33.7410071942446, "grad_norm": 0.07177361100912094, "learning_rate": 1.6089721272223755e-05, "loss": 0.0176, "step": 37520 }, { "action_loss": 0.0065265134908258915, "epoch": 33.7410071942446, "step": 37520 }, { "epoch": 33.7410071942446, "step": 37520, "torque_loss": 0.2401614636182785 }, { "epoch": 33.75, "grad_norm": 0.0786796510219574, "learning_rate": 1.606542690961151e-05, "loss": 0.0192, "step": 37530 }, { "action_loss": 0.0049143903888762, "epoch": 33.75, "step": 37530 }, { "epoch": 33.75, "step": 37530, "torque_loss": 0.16944777965545654 }, { "epoch": 33.7589928057554, "grad_norm": 0.11859776824712753, "learning_rate": 1.6041147391134438e-05, "loss": 0.0189, "step": 37540 }, { "action_loss": 0.0022998081985861063, "epoch": 33.7589928057554, "step": 37540 }, { "epoch": 33.7589928057554, "step": 37540, "torque_loss": 0.15952493250370026 }, { "epoch": 33.76798561151079, "grad_norm": 0.14837095141410828, "learning_rate": 1.6016882727413196e-05, "loss": 0.0177, "step": 37550 }, { "action_loss": 0.0009316588402725756, "epoch": 33.76798561151079, "step": 37550 }, { "epoch": 33.76798561151079, "step": 37550, "torque_loss": 0.1458263397216797 }, { "epoch": 33.776978417266186, "grad_norm": 0.10533616691827774, "learning_rate": 1.5992632929062014e-05, "loss": 0.0191, "step": 37560 }, { "action_loss": 0.0024989440571516752, "epoch": 33.776978417266186, "step": 37560 }, { "epoch": 33.776978417266186, "step": 37560, "torque_loss": 0.15000219643115997 }, { "epoch": 33.78597122302158, "grad_norm": 0.04814053699374199, "learning_rate": 1.5968398006688557e-05, "loss": 0.0196, "step": 37570 }, { "action_loss": 0.0006105579086579382, "epoch": 33.78597122302158, "step": 37570 }, { "epoch": 33.78597122302158, "step": 37570, "torque_loss": 0.146027609705925 }, { "epoch": 33.79496402877698, "grad_norm": 0.08774349838495255, "learning_rate": 1.594417797089399e-05, "loss": 0.0196, "step": 37580 }, { "action_loss": 0.0010253824293613434, "epoch": 33.79496402877698, "step": 37580 }, { "epoch": 33.79496402877698, "step": 37580, "torque_loss": 0.15333224833011627 }, { "epoch": 33.80395683453237, "grad_norm": 0.06515316665172577, "learning_rate": 1.5919972832273015e-05, "loss": 0.0226, "step": 37590 }, { "action_loss": 0.0009146946831606328, "epoch": 33.80395683453237, "step": 37590 }, { "epoch": 33.80395683453237, "step": 37590, "torque_loss": 0.1238727867603302 }, { "epoch": 33.81294964028777, "grad_norm": 0.12062844634056091, "learning_rate": 1.5895782601413767e-05, "loss": 0.0177, "step": 37600 }, { "action_loss": 0.00991820264607668, "epoch": 33.81294964028777, "step": 37600 }, { "epoch": 33.81294964028777, "step": 37600, "torque_loss": 0.16855458915233612 }, { "epoch": 33.82194244604317, "grad_norm": 0.05156318470835686, "learning_rate": 1.587160728889785e-05, "loss": 0.0228, "step": 37610 }, { "action_loss": 0.0070867217145860195, "epoch": 33.82194244604317, "step": 37610 }, { "epoch": 33.82194244604317, "step": 37610, "torque_loss": 0.18526458740234375 }, { "epoch": 33.830935251798564, "grad_norm": 0.0481836162507534, "learning_rate": 1.58474469053004e-05, "loss": 0.0174, "step": 37620 }, { "action_loss": 0.0008785412064753473, "epoch": 33.830935251798564, "step": 37620 }, { "epoch": 33.830935251798564, "step": 37620, "torque_loss": 0.17327530682086945 }, { "epoch": 33.839928057553955, "grad_norm": 0.06896165013313293, "learning_rate": 1.5823301461189976e-05, "loss": 0.0171, "step": 37630 }, { "action_loss": 0.0018434199737384915, "epoch": 33.839928057553955, "step": 37630 }, { "epoch": 33.839928057553955, "step": 37630, "torque_loss": 0.15218497812747955 }, { "epoch": 33.84892086330935, "grad_norm": 0.0841773971915245, "learning_rate": 1.5799170967128606e-05, "loss": 0.0197, "step": 37640 }, { "action_loss": 0.0009482040186412632, "epoch": 33.84892086330935, "step": 37640 }, { "epoch": 33.84892086330935, "step": 37640, "torque_loss": 0.1173444390296936 }, { "epoch": 33.85791366906475, "grad_norm": 0.10791154205799103, "learning_rate": 1.5775055433671814e-05, "loss": 0.0179, "step": 37650 }, { "action_loss": 0.0018952926620841026, "epoch": 33.85791366906475, "step": 37650 }, { "epoch": 33.85791366906475, "step": 37650, "torque_loss": 0.21103815734386444 }, { "epoch": 33.86690647482014, "grad_norm": 0.06429515033960342, "learning_rate": 1.5750954871368535e-05, "loss": 0.0185, "step": 37660 }, { "action_loss": 0.0018994776764884591, "epoch": 33.86690647482014, "step": 37660 }, { "epoch": 33.86690647482014, "step": 37660, "torque_loss": 0.11946368217468262 }, { "epoch": 33.87589928057554, "grad_norm": 0.11940155178308487, "learning_rate": 1.5726869290761158e-05, "loss": 0.0175, "step": 37670 }, { "action_loss": 0.0005386331467889249, "epoch": 33.87589928057554, "step": 37670 }, { "epoch": 33.87589928057554, "step": 37670, "torque_loss": 0.14595310389995575 }, { "epoch": 33.884892086330936, "grad_norm": 0.07047353684902191, "learning_rate": 1.5702798702385574e-05, "loss": 0.0173, "step": 37680 }, { "action_loss": 0.005623879376798868, "epoch": 33.884892086330936, "step": 37680 }, { "epoch": 33.884892086330936, "step": 37680, "torque_loss": 0.11292990297079086 }, { "epoch": 33.893884892086334, "grad_norm": 0.1049802228808403, "learning_rate": 1.5678743116771038e-05, "loss": 0.02, "step": 37690 }, { "action_loss": 0.0013001972110942006, "epoch": 33.893884892086334, "step": 37690 }, { "epoch": 33.893884892086334, "step": 37690, "torque_loss": 0.09745850414037704 }, { "epoch": 33.902877697841724, "grad_norm": 0.1432323455810547, "learning_rate": 1.5654702544440313e-05, "loss": 0.019, "step": 37700 }, { "action_loss": 0.0014815792674198747, "epoch": 33.902877697841724, "step": 37700 }, { "epoch": 33.902877697841724, "step": 37700, "torque_loss": 0.10183290392160416 }, { "epoch": 33.91187050359712, "grad_norm": 0.08361122757196426, "learning_rate": 1.5630676995909532e-05, "loss": 0.0167, "step": 37710 }, { "action_loss": 0.005430722143501043, "epoch": 33.91187050359712, "step": 37710 }, { "epoch": 33.91187050359712, "step": 37710, "torque_loss": 0.2271084189414978 }, { "epoch": 33.92086330935252, "grad_norm": 0.07197365909814835, "learning_rate": 1.560666648168832e-05, "loss": 0.0186, "step": 37720 }, { "action_loss": 0.0012073480756953359, "epoch": 33.92086330935252, "step": 37720 }, { "epoch": 33.92086330935252, "step": 37720, "torque_loss": 0.08672010898590088 }, { "epoch": 33.92985611510792, "grad_norm": 0.08140970021486282, "learning_rate": 1.5582671012279663e-05, "loss": 0.0169, "step": 37730 }, { "action_loss": 0.0008148884517140687, "epoch": 33.92985611510792, "step": 37730 }, { "epoch": 33.92985611510792, "step": 37730, "torque_loss": 0.12816663086414337 }, { "epoch": 33.93884892086331, "grad_norm": 0.08312903344631195, "learning_rate": 1.5558690598179997e-05, "loss": 0.0201, "step": 37740 }, { "action_loss": 0.0033869475591927767, "epoch": 33.93884892086331, "step": 37740 }, { "epoch": 33.93884892086331, "step": 37740, "torque_loss": 0.17401540279388428 }, { "epoch": 33.947841726618705, "grad_norm": 0.11405637860298157, "learning_rate": 1.553472524987919e-05, "loss": 0.0173, "step": 37750 }, { "action_loss": 0.0015119625022634864, "epoch": 33.947841726618705, "step": 37750 }, { "epoch": 33.947841726618705, "step": 37750, "torque_loss": 0.14093133807182312 }, { "epoch": 33.9568345323741, "grad_norm": 0.06328603625297546, "learning_rate": 1.5510774977860485e-05, "loss": 0.021, "step": 37760 }, { "action_loss": 0.0029961427208036184, "epoch": 33.9568345323741, "step": 37760 }, { "epoch": 33.9568345323741, "step": 37760, "torque_loss": 0.12030506134033203 }, { "epoch": 33.96582733812949, "grad_norm": 0.06111197918653488, "learning_rate": 1.5486839792600532e-05, "loss": 0.0165, "step": 37770 }, { "action_loss": 0.006418503820896149, "epoch": 33.96582733812949, "step": 37770 }, { "epoch": 33.96582733812949, "step": 37770, "torque_loss": 0.23632626235485077 }, { "epoch": 33.97482014388489, "grad_norm": 0.10056407749652863, "learning_rate": 1.546291970456942e-05, "loss": 0.0192, "step": 37780 }, { "action_loss": 0.0028194691985845566, "epoch": 33.97482014388489, "step": 37780 }, { "epoch": 33.97482014388489, "step": 37780, "torque_loss": 0.16016168892383575 }, { "epoch": 33.98381294964029, "grad_norm": 0.07196898758411407, "learning_rate": 1.5439014724230606e-05, "loss": 0.0172, "step": 37790 }, { "action_loss": 0.0015111728571355343, "epoch": 33.98381294964029, "step": 37790 }, { "epoch": 33.98381294964029, "step": 37790, "torque_loss": 0.10431236028671265 }, { "epoch": 33.992805755395686, "grad_norm": 0.04775509983301163, "learning_rate": 1.5415124862040907e-05, "loss": 0.0228, "step": 37800 }, { "action_loss": 0.0005882842815481126, "epoch": 33.992805755395686, "step": 37800 }, { "epoch": 33.992805755395686, "step": 37800, "torque_loss": 0.14208592474460602 }, { "epoch": 34.00179856115108, "grad_norm": 0.053798217326402664, "learning_rate": 1.53912501284506e-05, "loss": 0.017, "step": 37810 }, { "action_loss": 0.0076754130423069, "epoch": 34.00179856115108, "step": 37810 }, { "epoch": 34.00179856115108, "step": 37810, "torque_loss": 0.2262921780347824 }, { "epoch": 34.010791366906474, "grad_norm": 0.09615228325128555, "learning_rate": 1.5367390533903292e-05, "loss": 0.0207, "step": 37820 }, { "action_loss": 0.0014374632155522704, "epoch": 34.010791366906474, "step": 37820 }, { "epoch": 34.010791366906474, "step": 37820, "torque_loss": 0.115237295627594 }, { "epoch": 34.01978417266187, "grad_norm": 0.09545709937810898, "learning_rate": 1.5343546088835964e-05, "loss": 0.0219, "step": 37830 }, { "action_loss": 0.011948962695896626, "epoch": 34.01978417266187, "step": 37830 }, { "epoch": 34.01978417266187, "step": 37830, "torque_loss": 0.2006690353155136 }, { "epoch": 34.02877697841727, "grad_norm": 0.11042989790439606, "learning_rate": 1.531971680367901e-05, "loss": 0.0214, "step": 37840 }, { "action_loss": 0.006692370865494013, "epoch": 34.02877697841727, "step": 37840 }, { "epoch": 34.02877697841727, "step": 37840, "torque_loss": 0.20059943199157715 }, { "epoch": 34.03776978417266, "grad_norm": 0.06969524174928665, "learning_rate": 1.529590268885616e-05, "loss": 0.0219, "step": 37850 }, { "action_loss": 0.0025147877167910337, "epoch": 34.03776978417266, "step": 37850 }, { "epoch": 34.03776978417266, "step": 37850, "torque_loss": 0.1493871659040451 }, { "epoch": 34.04676258992806, "grad_norm": 0.07186007499694824, "learning_rate": 1.5272103754784517e-05, "loss": 0.016, "step": 37860 }, { "action_loss": 0.004223404917865992, "epoch": 34.04676258992806, "step": 37860 }, { "epoch": 34.04676258992806, "step": 37860, "torque_loss": 0.1960792988538742 }, { "epoch": 34.055755395683455, "grad_norm": 0.09153104573488235, "learning_rate": 1.5248320011874523e-05, "loss": 0.0201, "step": 37870 }, { "action_loss": 0.0028013151604682207, "epoch": 34.055755395683455, "step": 37870 }, { "epoch": 34.055755395683455, "step": 37870, "torque_loss": 0.21996742486953735 }, { "epoch": 34.064748201438846, "grad_norm": 0.05555305629968643, "learning_rate": 1.5224551470530035e-05, "loss": 0.0191, "step": 37880 }, { "action_loss": 0.0010397806763648987, "epoch": 34.064748201438846, "step": 37880 }, { "epoch": 34.064748201438846, "step": 37880, "torque_loss": 0.10109037905931473 }, { "epoch": 34.07374100719424, "grad_norm": 0.06929381191730499, "learning_rate": 1.5200798141148193e-05, "loss": 0.0167, "step": 37890 }, { "action_loss": 0.00690482696518302, "epoch": 34.07374100719424, "step": 37890 }, { "epoch": 34.07374100719424, "step": 37890, "torque_loss": 0.1498149335384369 }, { "epoch": 34.08273381294964, "grad_norm": 0.07216275483369827, "learning_rate": 1.5177060034119505e-05, "loss": 0.017, "step": 37900 }, { "action_loss": 0.00177294434979558, "epoch": 34.08273381294964, "step": 37900 }, { "epoch": 34.08273381294964, "step": 37900, "torque_loss": 0.18981988728046417 }, { "epoch": 34.09172661870504, "grad_norm": 0.09351886808872223, "learning_rate": 1.5153337159827835e-05, "loss": 0.0221, "step": 37910 }, { "action_loss": 0.015705345198512077, "epoch": 34.09172661870504, "step": 37910 }, { "epoch": 34.09172661870504, "step": 37910, "torque_loss": 0.2754689157009125 }, { "epoch": 34.10071942446043, "grad_norm": 0.08991079032421112, "learning_rate": 1.5129629528650397e-05, "loss": 0.0198, "step": 37920 }, { "action_loss": 0.0011218303116038442, "epoch": 34.10071942446043, "step": 37920 }, { "epoch": 34.10071942446043, "step": 37920, "torque_loss": 0.16816158592700958 }, { "epoch": 34.10971223021583, "grad_norm": 0.1667451709508896, "learning_rate": 1.5105937150957671e-05, "loss": 0.0209, "step": 37930 }, { "action_loss": 0.0006043437751941383, "epoch": 34.10971223021583, "step": 37930 }, { "epoch": 34.10971223021583, "step": 37930, "torque_loss": 0.15901033580303192 }, { "epoch": 34.118705035971225, "grad_norm": 0.06593692302703857, "learning_rate": 1.5082260037113548e-05, "loss": 0.016, "step": 37940 }, { "action_loss": 0.0008741689962334931, "epoch": 34.118705035971225, "step": 37940 }, { "epoch": 34.118705035971225, "step": 37940, "torque_loss": 0.1323806494474411 }, { "epoch": 34.12769784172662, "grad_norm": 0.10087300091981888, "learning_rate": 1.5058598197475187e-05, "loss": 0.0185, "step": 37950 }, { "action_loss": 0.002368246205151081, "epoch": 34.12769784172662, "step": 37950 }, { "epoch": 34.12769784172662, "step": 37950, "torque_loss": 0.1046668067574501 }, { "epoch": 34.13669064748201, "grad_norm": 0.06445340812206268, "learning_rate": 1.5034951642393064e-05, "loss": 0.0176, "step": 37960 }, { "action_loss": 0.002251628553494811, "epoch": 34.13669064748201, "step": 37960 }, { "epoch": 34.13669064748201, "step": 37960, "torque_loss": 0.1891235113143921 }, { "epoch": 34.14568345323741, "grad_norm": 0.06257492303848267, "learning_rate": 1.5011320382211013e-05, "loss": 0.016, "step": 37970 }, { "action_loss": 0.0008241583709605038, "epoch": 34.14568345323741, "step": 37970 }, { "epoch": 34.14568345323741, "step": 37970, "torque_loss": 0.15481580793857574 }, { "epoch": 34.15467625899281, "grad_norm": 0.04384126141667366, "learning_rate": 1.4987704427266136e-05, "loss": 0.0156, "step": 37980 }, { "action_loss": 0.0016614248743280768, "epoch": 34.15467625899281, "step": 37980 }, { "epoch": 34.15467625899281, "step": 37980, "torque_loss": 0.10713764280080795 }, { "epoch": 34.1636690647482, "grad_norm": 0.09792515635490417, "learning_rate": 1.4964103787888839e-05, "loss": 0.0183, "step": 37990 }, { "action_loss": 0.0005090991035103798, "epoch": 34.1636690647482, "step": 37990 }, { "epoch": 34.1636690647482, "step": 37990, "torque_loss": 0.14026404917240143 }, { "epoch": 34.172661870503596, "grad_norm": 0.03727690875530243, "learning_rate": 1.4940518474402882e-05, "loss": 0.0149, "step": 38000 }, { "action_loss": 0.009728694334626198, "epoch": 34.172661870503596, "step": 38000 }, { "epoch": 34.172661870503596, "step": 38000, "torque_loss": 0.19219918549060822 }, { "epoch": 34.181654676258994, "grad_norm": 0.044409945607185364, "learning_rate": 1.4916948497125249e-05, "loss": 0.0189, "step": 38010 }, { "action_loss": 0.0008516584639437497, "epoch": 34.181654676258994, "step": 38010 }, { "epoch": 34.181654676258994, "step": 38010, "torque_loss": 0.11524339765310287 }, { "epoch": 34.19064748201439, "grad_norm": 0.07368732243776321, "learning_rate": 1.4893393866366267e-05, "loss": 0.0175, "step": 38020 }, { "action_loss": 0.003338295267894864, "epoch": 34.19064748201439, "step": 38020 }, { "epoch": 34.19064748201439, "step": 38020, "torque_loss": 0.11805311590433121 }, { "epoch": 34.19964028776978, "grad_norm": 0.05130127817392349, "learning_rate": 1.4869854592429506e-05, "loss": 0.0211, "step": 38030 }, { "action_loss": 0.009153703227639198, "epoch": 34.19964028776978, "step": 38030 }, { "epoch": 34.19964028776978, "step": 38030, "torque_loss": 0.1760467141866684 }, { "epoch": 34.20863309352518, "grad_norm": 0.06753101944923401, "learning_rate": 1.4846330685611875e-05, "loss": 0.0196, "step": 38040 }, { "action_loss": 0.0011793534504249692, "epoch": 34.20863309352518, "step": 38040 }, { "epoch": 34.20863309352518, "step": 38040, "torque_loss": 0.15072451531887054 }, { "epoch": 34.21762589928058, "grad_norm": 0.09450419992208481, "learning_rate": 1.482282215620352e-05, "loss": 0.0226, "step": 38050 }, { "action_loss": 0.0014222568133845925, "epoch": 34.21762589928058, "step": 38050 }, { "epoch": 34.21762589928058, "step": 38050, "torque_loss": 0.16651590168476105 }, { "epoch": 34.226618705035975, "grad_norm": 0.10059326142072678, "learning_rate": 1.4799329014487857e-05, "loss": 0.0198, "step": 38060 }, { "action_loss": 0.002468032529577613, "epoch": 34.226618705035975, "step": 38060 }, { "epoch": 34.226618705035975, "step": 38060, "torque_loss": 0.23268604278564453 }, { "epoch": 34.235611510791365, "grad_norm": 0.07080873101949692, "learning_rate": 1.4775851270741603e-05, "loss": 0.0174, "step": 38070 }, { "action_loss": 0.0044439309276640415, "epoch": 34.235611510791365, "step": 38070 }, { "epoch": 34.235611510791365, "step": 38070, "torque_loss": 0.2083311676979065 }, { "epoch": 34.24460431654676, "grad_norm": 0.13401897251605988, "learning_rate": 1.4752388935234719e-05, "loss": 0.0181, "step": 38080 }, { "action_loss": 0.0029440049547702074, "epoch": 34.24460431654676, "step": 38080 }, { "epoch": 34.24460431654676, "step": 38080, "torque_loss": 0.17046602070331573 }, { "epoch": 34.25359712230216, "grad_norm": 0.05380817875266075, "learning_rate": 1.4728942018230401e-05, "loss": 0.0178, "step": 38090 }, { "action_loss": 0.001739873201586306, "epoch": 34.25359712230216, "step": 38090 }, { "epoch": 34.25359712230216, "step": 38090, "torque_loss": 0.1251876801252365 }, { "epoch": 34.26258992805755, "grad_norm": 0.06296231597661972, "learning_rate": 1.4705510529985167e-05, "loss": 0.0163, "step": 38100 }, { "action_loss": 0.0006689238944090903, "epoch": 34.26258992805755, "step": 38100 }, { "epoch": 34.26258992805755, "step": 38100, "torque_loss": 0.12300732731819153 }, { "epoch": 34.27158273381295, "grad_norm": 0.04602948576211929, "learning_rate": 1.4682094480748726e-05, "loss": 0.0158, "step": 38110 }, { "action_loss": 0.005195119883865118, "epoch": 34.27158273381295, "step": 38110 }, { "epoch": 34.27158273381295, "step": 38110, "torque_loss": 0.16365166008472443 }, { "epoch": 34.280575539568346, "grad_norm": 0.05157650634646416, "learning_rate": 1.4658693880764035e-05, "loss": 0.019, "step": 38120 }, { "action_loss": 0.0012376467930153012, "epoch": 34.280575539568346, "step": 38120 }, { "epoch": 34.280575539568346, "step": 38120, "torque_loss": 0.15214507281780243 }, { "epoch": 34.289568345323744, "grad_norm": 0.09748195111751556, "learning_rate": 1.463530874026735e-05, "loss": 0.0186, "step": 38130 }, { "action_loss": 0.001633903943002224, "epoch": 34.289568345323744, "step": 38130 }, { "epoch": 34.289568345323744, "step": 38130, "torque_loss": 0.19075751304626465 }, { "epoch": 34.298561151079134, "grad_norm": 0.09542537480592728, "learning_rate": 1.461193906948809e-05, "loss": 0.0183, "step": 38140 }, { "action_loss": 0.0008566421456634998, "epoch": 34.298561151079134, "step": 38140 }, { "epoch": 34.298561151079134, "step": 38140, "torque_loss": 0.13199754059314728 }, { "epoch": 34.30755395683453, "grad_norm": 0.07362113893032074, "learning_rate": 1.4588584878648975e-05, "loss": 0.0177, "step": 38150 }, { "action_loss": 0.003926007077097893, "epoch": 34.30755395683453, "step": 38150 }, { "epoch": 34.30755395683453, "step": 38150, "torque_loss": 0.19494004547595978 }, { "epoch": 34.31654676258993, "grad_norm": 0.06966660916805267, "learning_rate": 1.4565246177965896e-05, "loss": 0.0204, "step": 38160 }, { "action_loss": 0.0017892220057547092, "epoch": 34.31654676258993, "step": 38160 }, { "epoch": 34.31654676258993, "step": 38160, "torque_loss": 0.15476006269454956 }, { "epoch": 34.32553956834532, "grad_norm": 0.07647073268890381, "learning_rate": 1.4541922977648015e-05, "loss": 0.0167, "step": 38170 }, { "action_loss": 0.006575560662895441, "epoch": 34.32553956834532, "step": 38170 }, { "epoch": 34.32553956834532, "step": 38170, "torque_loss": 0.19949598610401154 }, { "epoch": 34.33453237410072, "grad_norm": 0.12340639531612396, "learning_rate": 1.451861528789768e-05, "loss": 0.0189, "step": 38180 }, { "action_loss": 0.0009049242362380028, "epoch": 34.33453237410072, "step": 38180 }, { "epoch": 34.33453237410072, "step": 38180, "torque_loss": 0.1289643496274948 }, { "epoch": 34.343525179856115, "grad_norm": 0.08107496052980423, "learning_rate": 1.4495323118910454e-05, "loss": 0.0189, "step": 38190 }, { "action_loss": 0.000927973713260144, "epoch": 34.343525179856115, "step": 38190 }, { "epoch": 34.343525179856115, "step": 38190, "torque_loss": 0.12418901920318604 }, { "epoch": 34.35251798561151, "grad_norm": 0.06655401736497879, "learning_rate": 1.4472046480875145e-05, "loss": 0.0193, "step": 38200 }, { "action_loss": 0.002001926302909851, "epoch": 34.35251798561151, "step": 38200 }, { "epoch": 34.35251798561151, "step": 38200, "torque_loss": 0.13955886662006378 }, { "epoch": 34.361510791366904, "grad_norm": 0.0707448348402977, "learning_rate": 1.4448785383973745e-05, "loss": 0.0175, "step": 38210 }, { "action_loss": 0.007633227854967117, "epoch": 34.361510791366904, "step": 38210 }, { "epoch": 34.361510791366904, "step": 38210, "torque_loss": 0.1440444141626358 }, { "epoch": 34.3705035971223, "grad_norm": 0.10509834438562393, "learning_rate": 1.4425539838381418e-05, "loss": 0.0189, "step": 38220 }, { "action_loss": 0.0013935304014012218, "epoch": 34.3705035971223, "step": 38220 }, { "epoch": 34.3705035971223, "step": 38220, "torque_loss": 0.15371640026569366 }, { "epoch": 34.3794964028777, "grad_norm": 0.05981909856200218, "learning_rate": 1.4402309854266594e-05, "loss": 0.0194, "step": 38230 }, { "action_loss": 0.0010028014658018947, "epoch": 34.3794964028777, "step": 38230 }, { "epoch": 34.3794964028777, "step": 38230, "torque_loss": 0.14388059079647064 }, { "epoch": 34.388489208633096, "grad_norm": 0.06152966618537903, "learning_rate": 1.4379095441790846e-05, "loss": 0.0174, "step": 38240 }, { "action_loss": 0.0011738265166059136, "epoch": 34.388489208633096, "step": 38240 }, { "epoch": 34.388489208633096, "step": 38240, "torque_loss": 0.13445429503917694 }, { "epoch": 34.39748201438849, "grad_norm": 0.05836695805191994, "learning_rate": 1.4355896611108921e-05, "loss": 0.0193, "step": 38250 }, { "action_loss": 0.0011303307255730033, "epoch": 34.39748201438849, "step": 38250 }, { "epoch": 34.39748201438849, "step": 38250, "torque_loss": 0.10114140063524246 }, { "epoch": 34.406474820143885, "grad_norm": 0.08347189426422119, "learning_rate": 1.433271337236881e-05, "loss": 0.0204, "step": 38260 }, { "action_loss": 0.0010911583667621017, "epoch": 34.406474820143885, "step": 38260 }, { "epoch": 34.406474820143885, "step": 38260, "torque_loss": 0.14340922236442566 }, { "epoch": 34.41546762589928, "grad_norm": 0.10821869224309921, "learning_rate": 1.4309545735711638e-05, "loss": 0.0185, "step": 38270 }, { "action_loss": 0.0009471275843679905, "epoch": 34.41546762589928, "step": 38270 }, { "epoch": 34.41546762589928, "step": 38270, "torque_loss": 0.10737812519073486 }, { "epoch": 34.42446043165468, "grad_norm": 0.12591159343719482, "learning_rate": 1.4286393711271696e-05, "loss": 0.0158, "step": 38280 }, { "action_loss": 0.0016887084348127246, "epoch": 34.42446043165468, "step": 38280 }, { "epoch": 34.42446043165468, "step": 38280, "torque_loss": 0.12520118057727814 }, { "epoch": 34.43345323741007, "grad_norm": 0.08249036222696304, "learning_rate": 1.4263257309176497e-05, "loss": 0.0192, "step": 38290 }, { "action_loss": 0.002001750050112605, "epoch": 34.43345323741007, "step": 38290 }, { "epoch": 34.43345323741007, "step": 38290, "torque_loss": 0.16276328265666962 }, { "epoch": 34.44244604316547, "grad_norm": 0.0759679526090622, "learning_rate": 1.4240136539546678e-05, "loss": 0.0206, "step": 38300 }, { "action_loss": 0.01612114906311035, "epoch": 34.44244604316547, "step": 38300 }, { "epoch": 34.44244604316547, "step": 38300, "torque_loss": 0.1709708422422409 }, { "epoch": 34.451438848920866, "grad_norm": 0.10740894079208374, "learning_rate": 1.4217031412496029e-05, "loss": 0.0219, "step": 38310 }, { "action_loss": 0.0008726793457753956, "epoch": 34.451438848920866, "step": 38310 }, { "epoch": 34.451438848920866, "step": 38310, "torque_loss": 0.11150071024894714 }, { "epoch": 34.460431654676256, "grad_norm": 0.06637918949127197, "learning_rate": 1.4193941938131554e-05, "loss": 0.0195, "step": 38320 }, { "action_loss": 0.0020323798526078463, "epoch": 34.460431654676256, "step": 38320 }, { "epoch": 34.460431654676256, "step": 38320, "torque_loss": 0.1461600512266159 }, { "epoch": 34.469424460431654, "grad_norm": 0.07345297187566757, "learning_rate": 1.4170868126553356e-05, "loss": 0.0177, "step": 38330 }, { "action_loss": 0.0023338987957686186, "epoch": 34.469424460431654, "step": 38330 }, { "epoch": 34.469424460431654, "step": 38330, "torque_loss": 0.1835598349571228 }, { "epoch": 34.47841726618705, "grad_norm": 0.10346246510744095, "learning_rate": 1.4147809987854682e-05, "loss": 0.024, "step": 38340 }, { "action_loss": 0.001656743697822094, "epoch": 34.47841726618705, "step": 38340 }, { "epoch": 34.47841726618705, "step": 38340, "torque_loss": 0.26698431372642517 }, { "epoch": 34.48741007194245, "grad_norm": 0.06331776827573776, "learning_rate": 1.4124767532121985e-05, "loss": 0.0184, "step": 38350 }, { "action_loss": 0.0011494419304654002, "epoch": 34.48741007194245, "step": 38350 }, { "epoch": 34.48741007194245, "step": 38350, "torque_loss": 0.14915968477725983 }, { "epoch": 34.49640287769784, "grad_norm": 0.06732575595378876, "learning_rate": 1.41017407694348e-05, "loss": 0.0192, "step": 38360 }, { "action_loss": 0.002721318742260337, "epoch": 34.49640287769784, "step": 38360 }, { "epoch": 34.49640287769784, "step": 38360, "torque_loss": 0.1765460968017578 }, { "epoch": 34.50539568345324, "grad_norm": 0.0881899893283844, "learning_rate": 1.4078729709865801e-05, "loss": 0.0173, "step": 38370 }, { "action_loss": 0.0012300423113629222, "epoch": 34.50539568345324, "step": 38370 }, { "epoch": 34.50539568345324, "step": 38370, "torque_loss": 0.1865534782409668 }, { "epoch": 34.514388489208635, "grad_norm": 0.0730958953499794, "learning_rate": 1.4055734363480833e-05, "loss": 0.0215, "step": 38380 }, { "action_loss": 0.0022835957352072, "epoch": 34.514388489208635, "step": 38380 }, { "epoch": 34.514388489208635, "step": 38380, "torque_loss": 0.18698932230472565 }, { "epoch": 34.523381294964025, "grad_norm": 0.10040915757417679, "learning_rate": 1.4032754740338815e-05, "loss": 0.0196, "step": 38390 }, { "action_loss": 0.0010627557057887316, "epoch": 34.523381294964025, "step": 38390 }, { "epoch": 34.523381294964025, "step": 38390, "torque_loss": 0.13039588928222656 }, { "epoch": 34.53237410071942, "grad_norm": 0.07548246532678604, "learning_rate": 1.4009790850491844e-05, "loss": 0.0176, "step": 38400 }, { "action_loss": 0.0009267593850381672, "epoch": 34.53237410071942, "step": 38400 }, { "epoch": 34.53237410071942, "step": 38400, "torque_loss": 0.16736267507076263 }, { "epoch": 34.54136690647482, "grad_norm": 0.07763104140758514, "learning_rate": 1.3986842703985076e-05, "loss": 0.0172, "step": 38410 }, { "action_loss": 0.0006841846625320613, "epoch": 34.54136690647482, "step": 38410 }, { "epoch": 34.54136690647482, "step": 38410, "torque_loss": 0.11194366216659546 }, { "epoch": 34.55035971223022, "grad_norm": 0.07719670981168747, "learning_rate": 1.3963910310856842e-05, "loss": 0.0169, "step": 38420 }, { "action_loss": 0.0021749555598944426, "epoch": 34.55035971223022, "step": 38420 }, { "epoch": 34.55035971223022, "step": 38420, "torque_loss": 0.12046913057565689 }, { "epoch": 34.55935251798561, "grad_norm": 0.10570816695690155, "learning_rate": 1.394099368113853e-05, "loss": 0.0198, "step": 38430 }, { "action_loss": 0.0012199541088193655, "epoch": 34.55935251798561, "step": 38430 }, { "epoch": 34.55935251798561, "step": 38430, "torque_loss": 0.13678419589996338 }, { "epoch": 34.568345323741006, "grad_norm": 0.09743461012840271, "learning_rate": 1.3918092824854644e-05, "loss": 0.0177, "step": 38440 }, { "action_loss": 0.002223086543381214, "epoch": 34.568345323741006, "step": 38440 }, { "epoch": 34.568345323741006, "step": 38440, "torque_loss": 0.11787442117929459 }, { "epoch": 34.577338129496404, "grad_norm": 0.09319806098937988, "learning_rate": 1.3895207752022821e-05, "loss": 0.0215, "step": 38450 }, { "action_loss": 0.0011186398332938552, "epoch": 34.577338129496404, "step": 38450 }, { "epoch": 34.577338129496404, "step": 38450, "torque_loss": 0.16720950603485107 }, { "epoch": 34.5863309352518, "grad_norm": 0.0615699365735054, "learning_rate": 1.3872338472653756e-05, "loss": 0.0201, "step": 38460 }, { "action_loss": 0.000901529157999903, "epoch": 34.5863309352518, "step": 38460 }, { "epoch": 34.5863309352518, "step": 38460, "torque_loss": 0.11219793558120728 }, { "epoch": 34.59532374100719, "grad_norm": 0.03496710583567619, "learning_rate": 1.3849484996751233e-05, "loss": 0.0155, "step": 38470 }, { "action_loss": 0.004992044996470213, "epoch": 34.59532374100719, "step": 38470 }, { "epoch": 34.59532374100719, "step": 38470, "torque_loss": 0.1857418566942215 }, { "epoch": 34.60431654676259, "grad_norm": 0.09042414277791977, "learning_rate": 1.382664733431217e-05, "loss": 0.0217, "step": 38480 }, { "action_loss": 0.001659360365010798, "epoch": 34.60431654676259, "step": 38480 }, { "epoch": 34.60431654676259, "step": 38480, "torque_loss": 0.18174679577350616 }, { "epoch": 34.61330935251799, "grad_norm": 0.10923931747674942, "learning_rate": 1.3803825495326522e-05, "loss": 0.0188, "step": 38490 }, { "action_loss": 0.0038762514013797045, "epoch": 34.61330935251799, "step": 38490 }, { "epoch": 34.61330935251799, "step": 38490, "torque_loss": 0.17718636989593506 }, { "epoch": 34.62230215827338, "grad_norm": 0.10103274136781693, "learning_rate": 1.3781019489777324e-05, "loss": 0.0182, "step": 38500 }, { "action_loss": 0.00158797949552536, "epoch": 34.62230215827338, "step": 38500 }, { "epoch": 34.62230215827338, "step": 38500, "torque_loss": 0.14552536606788635 }, { "epoch": 34.631294964028775, "grad_norm": 0.09145431220531464, "learning_rate": 1.375822932764072e-05, "loss": 0.0202, "step": 38510 }, { "action_loss": 0.0008690033573657274, "epoch": 34.631294964028775, "step": 38510 }, { "epoch": 34.631294964028775, "step": 38510, "torque_loss": 0.1257411241531372 }, { "epoch": 34.64028776978417, "grad_norm": 0.06362023949623108, "learning_rate": 1.373545501888589e-05, "loss": 0.0189, "step": 38520 }, { "action_loss": 0.002359890379011631, "epoch": 34.64028776978417, "step": 38520 }, { "epoch": 34.64028776978417, "step": 38520, "torque_loss": 0.15513287484645844 }, { "epoch": 34.64928057553957, "grad_norm": 0.1254902333021164, "learning_rate": 1.3712696573475092e-05, "loss": 0.0182, "step": 38530 }, { "action_loss": 0.0009768279269337654, "epoch": 34.64928057553957, "step": 38530 }, { "epoch": 34.64928057553957, "step": 38530, "torque_loss": 0.1857972890138626 }, { "epoch": 34.65827338129496, "grad_norm": 0.14493954181671143, "learning_rate": 1.368995400136363e-05, "loss": 0.0212, "step": 38540 }, { "action_loss": 0.0008147154585458338, "epoch": 34.65827338129496, "step": 38540 }, { "epoch": 34.65827338129496, "step": 38540, "torque_loss": 0.107089102268219 }, { "epoch": 34.66726618705036, "grad_norm": 0.08253894746303558, "learning_rate": 1.3667227312499903e-05, "loss": 0.0181, "step": 38550 }, { "action_loss": 0.015248785726726055, "epoch": 34.66726618705036, "step": 38550 }, { "epoch": 34.66726618705036, "step": 38550, "torque_loss": 0.2999379336833954 }, { "epoch": 34.67625899280576, "grad_norm": 0.17509208619594574, "learning_rate": 1.3644516516825323e-05, "loss": 0.0205, "step": 38560 }, { "action_loss": 0.005957507994025946, "epoch": 34.67625899280576, "step": 38560 }, { "epoch": 34.67625899280576, "step": 38560, "torque_loss": 0.22564227879047394 }, { "epoch": 34.685251798561154, "grad_norm": 0.10516833513975143, "learning_rate": 1.3621821624274356e-05, "loss": 0.0225, "step": 38570 }, { "action_loss": 0.0010612790938466787, "epoch": 34.685251798561154, "step": 38570 }, { "epoch": 34.685251798561154, "step": 38570, "torque_loss": 0.11569731682538986 }, { "epoch": 34.694244604316545, "grad_norm": 0.1255018711090088, "learning_rate": 1.3599142644774537e-05, "loss": 0.0185, "step": 38580 }, { "action_loss": 0.0034214959014207125, "epoch": 34.694244604316545, "step": 38580 }, { "epoch": 34.694244604316545, "step": 38580, "torque_loss": 0.18557065725326538 }, { "epoch": 34.70323741007194, "grad_norm": 0.08696429431438446, "learning_rate": 1.357647958824641e-05, "loss": 0.0195, "step": 38590 }, { "action_loss": 0.000592725642491132, "epoch": 34.70323741007194, "step": 38590 }, { "epoch": 34.70323741007194, "step": 38590, "torque_loss": 0.12591630220413208 }, { "epoch": 34.71223021582734, "grad_norm": 0.09051788598299026, "learning_rate": 1.3553832464603555e-05, "loss": 0.0175, "step": 38600 }, { "action_loss": 0.0013313529780134559, "epoch": 34.71223021582734, "step": 38600 }, { "epoch": 34.71223021582734, "step": 38600, "torque_loss": 0.1275702565908432 }, { "epoch": 34.72122302158273, "grad_norm": 0.06524856388568878, "learning_rate": 1.3531201283752625e-05, "loss": 0.0172, "step": 38610 }, { "action_loss": 0.0032066914718598127, "epoch": 34.72122302158273, "step": 38610 }, { "epoch": 34.72122302158273, "step": 38610, "torque_loss": 0.15157894790172577 }, { "epoch": 34.73021582733813, "grad_norm": 0.03755180165171623, "learning_rate": 1.350858605559323e-05, "loss": 0.0181, "step": 38620 }, { "action_loss": 0.002324140863493085, "epoch": 34.73021582733813, "step": 38620 }, { "epoch": 34.73021582733813, "step": 38620, "torque_loss": 0.15796305239200592 }, { "epoch": 34.739208633093526, "grad_norm": 0.06541687995195389, "learning_rate": 1.348598679001808e-05, "loss": 0.0176, "step": 38630 }, { "action_loss": 0.0008264803327620029, "epoch": 34.739208633093526, "step": 38630 }, { "epoch": 34.739208633093526, "step": 38630, "torque_loss": 0.14837555587291718 }, { "epoch": 34.74820143884892, "grad_norm": 0.04431714862585068, "learning_rate": 1.3463403496912819e-05, "loss": 0.0156, "step": 38640 }, { "action_loss": 0.0011160016292706132, "epoch": 34.74820143884892, "step": 38640 }, { "epoch": 34.74820143884892, "step": 38640, "torque_loss": 0.13942791521549225 }, { "epoch": 34.757194244604314, "grad_norm": 0.10182619094848633, "learning_rate": 1.3440836186156187e-05, "loss": 0.018, "step": 38650 }, { "action_loss": 0.0015604691579937935, "epoch": 34.757194244604314, "step": 38650 }, { "epoch": 34.757194244604314, "step": 38650, "torque_loss": 0.13671475648880005 }, { "epoch": 34.76618705035971, "grad_norm": 0.07375392317771912, "learning_rate": 1.3418284867619852e-05, "loss": 0.0175, "step": 38660 }, { "action_loss": 0.0025640460662543774, "epoch": 34.76618705035971, "step": 38660 }, { "epoch": 34.76618705035971, "step": 38660, "torque_loss": 0.1832483559846878 }, { "epoch": 34.77517985611511, "grad_norm": 0.06290231645107269, "learning_rate": 1.3395749551168567e-05, "loss": 0.0199, "step": 38670 }, { "action_loss": 0.006796950940042734, "epoch": 34.77517985611511, "step": 38670 }, { "epoch": 34.77517985611511, "step": 38670, "torque_loss": 0.1644372195005417 }, { "epoch": 34.78417266187051, "grad_norm": 0.053544387221336365, "learning_rate": 1.3373230246660024e-05, "loss": 0.0165, "step": 38680 }, { "action_loss": 0.0009390080231241882, "epoch": 34.78417266187051, "step": 38680 }, { "epoch": 34.78417266187051, "step": 38680, "torque_loss": 0.11690616607666016 }, { "epoch": 34.7931654676259, "grad_norm": 0.08537854254245758, "learning_rate": 1.335072696394492e-05, "loss": 0.0177, "step": 38690 }, { "action_loss": 0.0020989617332816124, "epoch": 34.7931654676259, "step": 38690 }, { "epoch": 34.7931654676259, "step": 38690, "torque_loss": 0.11615058034658432 }, { "epoch": 34.802158273381295, "grad_norm": 0.16067864000797272, "learning_rate": 1.3328239712866986e-05, "loss": 0.0236, "step": 38700 }, { "action_loss": 0.0010781362652778625, "epoch": 34.802158273381295, "step": 38700 }, { "epoch": 34.802158273381295, "step": 38700, "torque_loss": 0.13183490931987762 }, { "epoch": 34.81115107913669, "grad_norm": 0.08009597659111023, "learning_rate": 1.3305768503262889e-05, "loss": 0.0164, "step": 38710 }, { "action_loss": 0.0016414799029007554, "epoch": 34.81115107913669, "step": 38710 }, { "epoch": 34.81115107913669, "step": 38710, "torque_loss": 0.167210653424263 }, { "epoch": 34.82014388489208, "grad_norm": 0.08795179426670074, "learning_rate": 1.3283313344962305e-05, "loss": 0.0171, "step": 38720 }, { "action_loss": 0.003612378379330039, "epoch": 34.82014388489208, "step": 38720 }, { "epoch": 34.82014388489208, "step": 38720, "torque_loss": 0.14370258152484894 }, { "epoch": 34.82913669064748, "grad_norm": 0.08271797746419907, "learning_rate": 1.3260874247787863e-05, "loss": 0.0169, "step": 38730 }, { "action_loss": 0.0006484823534265161, "epoch": 34.82913669064748, "step": 38730 }, { "epoch": 34.82913669064748, "step": 38730, "torque_loss": 0.10517614334821701 }, { "epoch": 34.83812949640288, "grad_norm": 0.07381057739257812, "learning_rate": 1.3238451221555226e-05, "loss": 0.0185, "step": 38740 }, { "action_loss": 0.003511979943141341, "epoch": 34.83812949640288, "step": 38740 }, { "epoch": 34.83812949640288, "step": 38740, "torque_loss": 0.19135935604572296 }, { "epoch": 34.847122302158276, "grad_norm": 0.10984306037425995, "learning_rate": 1.3216044276072965e-05, "loss": 0.0195, "step": 38750 }, { "action_loss": 0.0031176842749118805, "epoch": 34.847122302158276, "step": 38750 }, { "epoch": 34.847122302158276, "step": 38750, "torque_loss": 0.131856307387352 }, { "epoch": 34.856115107913666, "grad_norm": 0.04863303527235985, "learning_rate": 1.3193653421142622e-05, "loss": 0.0185, "step": 38760 }, { "action_loss": 0.0016559241339564323, "epoch": 34.856115107913666, "step": 38760 }, { "epoch": 34.856115107913666, "step": 38760, "torque_loss": 0.20143479108810425 }, { "epoch": 34.865107913669064, "grad_norm": 0.08344557136297226, "learning_rate": 1.3171278666558756e-05, "loss": 0.0173, "step": 38770 }, { "action_loss": 0.006649853195995092, "epoch": 34.865107913669064, "step": 38770 }, { "epoch": 34.865107913669064, "step": 38770, "torque_loss": 0.26248469948768616 }, { "epoch": 34.87410071942446, "grad_norm": 0.11142486333847046, "learning_rate": 1.3148920022108819e-05, "loss": 0.0212, "step": 38780 }, { "action_loss": 0.0011095762019976974, "epoch": 34.87410071942446, "step": 38780 }, { "epoch": 34.87410071942446, "step": 38780, "torque_loss": 0.20346780121326447 }, { "epoch": 34.88309352517986, "grad_norm": 0.07518848031759262, "learning_rate": 1.312657749757324e-05, "loss": 0.0193, "step": 38790 }, { "action_loss": 0.0006887055933475494, "epoch": 34.88309352517986, "step": 38790 }, { "epoch": 34.88309352517986, "step": 38790, "torque_loss": 0.12600590288639069 }, { "epoch": 34.89208633093525, "grad_norm": 0.07225699722766876, "learning_rate": 1.3104251102725412e-05, "loss": 0.0168, "step": 38800 }, { "action_loss": 0.0010668489849194884, "epoch": 34.89208633093525, "step": 38800 }, { "epoch": 34.89208633093525, "step": 38800, "torque_loss": 0.1242656335234642 }, { "epoch": 34.90107913669065, "grad_norm": 0.07877486199140549, "learning_rate": 1.3081940847331659e-05, "loss": 0.0187, "step": 38810 }, { "action_loss": 0.0017977299867197871, "epoch": 34.90107913669065, "step": 38810 }, { "epoch": 34.90107913669065, "step": 38810, "torque_loss": 0.12885713577270508 }, { "epoch": 34.910071942446045, "grad_norm": 0.14181628823280334, "learning_rate": 1.3059646741151222e-05, "loss": 0.0226, "step": 38820 }, { "action_loss": 0.003824833780527115, "epoch": 34.910071942446045, "step": 38820 }, { "epoch": 34.910071942446045, "step": 38820, "torque_loss": 0.13296088576316833 }, { "epoch": 34.919064748201436, "grad_norm": 0.11626174300909042, "learning_rate": 1.3037368793936334e-05, "loss": 0.0197, "step": 38830 }, { "action_loss": 0.000810096797067672, "epoch": 34.919064748201436, "step": 38830 }, { "epoch": 34.919064748201436, "step": 38830, "torque_loss": 0.10483639687299728 }, { "epoch": 34.92805755395683, "grad_norm": 0.09148307889699936, "learning_rate": 1.3015107015432104e-05, "loss": 0.0164, "step": 38840 }, { "action_loss": 0.0033492252696305513, "epoch": 34.92805755395683, "step": 38840 }, { "epoch": 34.92805755395683, "step": 38840, "torque_loss": 0.164734348654747 }, { "epoch": 34.93705035971223, "grad_norm": 0.13486847281455994, "learning_rate": 1.2992861415376584e-05, "loss": 0.0187, "step": 38850 }, { "action_loss": 0.002185351448133588, "epoch": 34.93705035971223, "step": 38850 }, { "epoch": 34.93705035971223, "step": 38850, "torque_loss": 0.14462679624557495 }, { "epoch": 34.94604316546763, "grad_norm": 0.07203659415245056, "learning_rate": 1.2970632003500782e-05, "loss": 0.0186, "step": 38860 }, { "action_loss": 0.0011589756468310952, "epoch": 34.94604316546763, "step": 38860 }, { "epoch": 34.94604316546763, "step": 38860, "torque_loss": 0.10664992779493332 }, { "epoch": 34.95503597122302, "grad_norm": 0.06942428648471832, "learning_rate": 1.294841878952856e-05, "loss": 0.0182, "step": 38870 }, { "action_loss": 0.0016190381720662117, "epoch": 34.95503597122302, "step": 38870 }, { "epoch": 34.95503597122302, "step": 38870, "torque_loss": 0.13088040053844452 }, { "epoch": 34.96402877697842, "grad_norm": 0.07913756370544434, "learning_rate": 1.2926221783176779e-05, "loss": 0.0209, "step": 38880 }, { "action_loss": 0.0014578509144484997, "epoch": 34.96402877697842, "step": 38880 }, { "epoch": 34.96402877697842, "step": 38880, "torque_loss": 0.14049683511257172 }, { "epoch": 34.973021582733814, "grad_norm": 0.1411549150943756, "learning_rate": 1.290404099415512e-05, "loss": 0.0207, "step": 38890 }, { "action_loss": 0.002757009118795395, "epoch": 34.973021582733814, "step": 38890 }, { "epoch": 34.973021582733814, "step": 38890, "torque_loss": 0.1880366951227188 }, { "epoch": 34.98201438848921, "grad_norm": 0.10464098304510117, "learning_rate": 1.2881876432166246e-05, "loss": 0.0166, "step": 38900 }, { "action_loss": 0.0024035435635596514, "epoch": 34.98201438848921, "step": 38900 }, { "epoch": 34.98201438848921, "step": 38900, "torque_loss": 0.1464916318655014 }, { "epoch": 34.9910071942446, "grad_norm": 0.05913747847080231, "learning_rate": 1.2859728106905678e-05, "loss": 0.0163, "step": 38910 }, { "action_loss": 0.0012188036926090717, "epoch": 34.9910071942446, "step": 38910 }, { "epoch": 34.9910071942446, "step": 38910, "torque_loss": 0.15003834664821625 }, { "epoch": 35.0, "grad_norm": 0.04663366079330444, "learning_rate": 1.2837596028061832e-05, "loss": 0.017, "step": 38920 }, { "action_loss": 0.0020296317525207996, "epoch": 35.0, "step": 38920 }, { "epoch": 35.0, "step": 38920, "torque_loss": 0.1223723292350769 }, { "epoch": 35.0089928057554, "grad_norm": 0.16619937121868134, "learning_rate": 1.2815480205316055e-05, "loss": 0.0233, "step": 38930 }, { "action_loss": 0.00692763039842248, "epoch": 35.0089928057554, "step": 38930 }, { "epoch": 35.0089928057554, "step": 38930, "torque_loss": 0.17397098243236542 }, { "epoch": 35.01798561151079, "grad_norm": 0.059125207364559174, "learning_rate": 1.279338064834255e-05, "loss": 0.0196, "step": 38940 }, { "action_loss": 0.0026053711771965027, "epoch": 35.01798561151079, "step": 38940 }, { "epoch": 35.01798561151079, "step": 38940, "torque_loss": 0.1725262999534607 }, { "epoch": 35.026978417266186, "grad_norm": 0.07287316024303436, "learning_rate": 1.27712973668084e-05, "loss": 0.0228, "step": 38950 }, { "action_loss": 0.008013460785150528, "epoch": 35.026978417266186, "step": 38950 }, { "epoch": 35.026978417266186, "step": 38950, "torque_loss": 0.21159698069095612 }, { "epoch": 35.03597122302158, "grad_norm": 0.06016039475798607, "learning_rate": 1.2749230370373605e-05, "loss": 0.021, "step": 38960 }, { "action_loss": 0.020713115110993385, "epoch": 35.03597122302158, "step": 38960 }, { "epoch": 35.03597122302158, "step": 38960, "torque_loss": 0.28654730319976807 }, { "epoch": 35.04496402877698, "grad_norm": 0.08395566791296005, "learning_rate": 1.2727179668691009e-05, "loss": 0.027, "step": 38970 }, { "action_loss": 0.0007832972332835197, "epoch": 35.04496402877698, "step": 38970 }, { "epoch": 35.04496402877698, "step": 38970, "torque_loss": 0.09823677688837051 }, { "epoch": 35.05395683453237, "grad_norm": 0.09839576482772827, "learning_rate": 1.2705145271406327e-05, "loss": 0.0183, "step": 38980 }, { "action_loss": 0.002088825451210141, "epoch": 35.05395683453237, "step": 38980 }, { "epoch": 35.05395683453237, "step": 38980, "torque_loss": 0.16507388651371002 }, { "epoch": 35.06294964028777, "grad_norm": 0.07937440276145935, "learning_rate": 1.2683127188158172e-05, "loss": 0.0176, "step": 38990 }, { "action_loss": 0.0007591703906655312, "epoch": 35.06294964028777, "step": 38990 }, { "epoch": 35.06294964028777, "step": 38990, "torque_loss": 0.10414717346429825 }, { "epoch": 35.07194244604317, "grad_norm": 0.11020886152982712, "learning_rate": 1.2661125428577997e-05, "loss": 0.0179, "step": 39000 }, { "action_loss": 0.0013296805555000901, "epoch": 35.07194244604317, "step": 39000 }, { "epoch": 35.07194244604317, "step": 39000, "torque_loss": 0.10512913018465042 }, { "epoch": 35.080935251798564, "grad_norm": 0.08065380156040192, "learning_rate": 1.2639140002290101e-05, "loss": 0.018, "step": 39010 }, { "action_loss": 0.0032179690897464752, "epoch": 35.080935251798564, "step": 39010 }, { "epoch": 35.080935251798564, "step": 39010, "torque_loss": 0.12040373682975769 }, { "epoch": 35.089928057553955, "grad_norm": 0.0788675993680954, "learning_rate": 1.261717091891168e-05, "loss": 0.0215, "step": 39020 }, { "action_loss": 0.004071805160492659, "epoch": 35.089928057553955, "step": 39020 }, { "epoch": 35.089928057553955, "step": 39020, "torque_loss": 0.17950622737407684 }, { "epoch": 35.09892086330935, "grad_norm": 0.051993343979120255, "learning_rate": 1.2595218188052755e-05, "loss": 0.0188, "step": 39030 }, { "action_loss": 0.0011992636136710644, "epoch": 35.09892086330935, "step": 39030 }, { "epoch": 35.09892086330935, "step": 39030, "torque_loss": 0.12281490117311478 }, { "epoch": 35.10791366906475, "grad_norm": 0.054909344762563705, "learning_rate": 1.2573281819316174e-05, "loss": 0.0162, "step": 39040 }, { "action_loss": 0.0016963655361905694, "epoch": 35.10791366906475, "step": 39040 }, { "epoch": 35.10791366906475, "step": 39040, "torque_loss": 0.15489953756332397 }, { "epoch": 35.11690647482014, "grad_norm": 0.06085480377078056, "learning_rate": 1.2551361822297674e-05, "loss": 0.0187, "step": 39050 }, { "action_loss": 0.018709376454353333, "epoch": 35.11690647482014, "step": 39050 }, { "epoch": 35.11690647482014, "step": 39050, "torque_loss": 0.1931246519088745 }, { "epoch": 35.12589928057554, "grad_norm": 0.08175470679998398, "learning_rate": 1.2529458206585804e-05, "loss": 0.019, "step": 39060 }, { "action_loss": 0.002057716716080904, "epoch": 35.12589928057554, "step": 39060 }, { "epoch": 35.12589928057554, "step": 39060, "torque_loss": 0.12409508228302002 }, { "epoch": 35.134892086330936, "grad_norm": 0.13135571777820587, "learning_rate": 1.250757098176194e-05, "loss": 0.0172, "step": 39070 }, { "action_loss": 0.0030177701264619827, "epoch": 35.134892086330936, "step": 39070 }, { "epoch": 35.134892086330936, "step": 39070, "torque_loss": 0.16153277456760406 }, { "epoch": 35.143884892086334, "grad_norm": 0.08827921748161316, "learning_rate": 1.2485700157400294e-05, "loss": 0.0184, "step": 39080 }, { "action_loss": 0.0033284726087003946, "epoch": 35.143884892086334, "step": 39080 }, { "epoch": 35.143884892086334, "step": 39080, "torque_loss": 0.0988030806183815 }, { "epoch": 35.152877697841724, "grad_norm": 0.164914071559906, "learning_rate": 1.246384574306792e-05, "loss": 0.0176, "step": 39090 }, { "action_loss": 0.004197588190436363, "epoch": 35.152877697841724, "step": 39090 }, { "epoch": 35.152877697841724, "step": 39090, "torque_loss": 0.15719358623027802 }, { "epoch": 35.16187050359712, "grad_norm": 0.05866703391075134, "learning_rate": 1.2442007748324664e-05, "loss": 0.0266, "step": 39100 }, { "action_loss": 0.0008360845968127251, "epoch": 35.16187050359712, "step": 39100 }, { "epoch": 35.16187050359712, "step": 39100, "torque_loss": 0.12080981582403183 }, { "epoch": 35.17086330935252, "grad_norm": 0.16506914794445038, "learning_rate": 1.2420186182723237e-05, "loss": 0.0196, "step": 39110 }, { "action_loss": 0.0009586228989064693, "epoch": 35.17086330935252, "step": 39110 }, { "epoch": 35.17086330935252, "step": 39110, "torque_loss": 0.14286303520202637 }, { "epoch": 35.17985611510792, "grad_norm": 0.11569065600633621, "learning_rate": 1.23983810558091e-05, "loss": 0.0194, "step": 39120 }, { "action_loss": 0.0037979769986122847, "epoch": 35.17985611510792, "step": 39120 }, { "epoch": 35.17985611510792, "step": 39120, "torque_loss": 0.14963027834892273 }, { "epoch": 35.18884892086331, "grad_norm": 0.06031525880098343, "learning_rate": 1.2376592377120582e-05, "loss": 0.0213, "step": 39130 }, { "action_loss": 0.004966945853084326, "epoch": 35.18884892086331, "step": 39130 }, { "epoch": 35.18884892086331, "step": 39130, "torque_loss": 0.15146224200725555 }, { "epoch": 35.197841726618705, "grad_norm": 0.13343390822410583, "learning_rate": 1.2354820156188768e-05, "loss": 0.0214, "step": 39140 }, { "action_loss": 0.0024062080774456263, "epoch": 35.197841726618705, "step": 39140 }, { "epoch": 35.197841726618705, "step": 39140, "torque_loss": 0.15906254947185516 }, { "epoch": 35.2068345323741, "grad_norm": 0.15424257516860962, "learning_rate": 1.2333064402537597e-05, "loss": 0.0186, "step": 39150 }, { "action_loss": 0.0015146381920203567, "epoch": 35.2068345323741, "step": 39150 }, { "epoch": 35.2068345323741, "step": 39150, "torque_loss": 0.15214969217777252 }, { "epoch": 35.21582733812949, "grad_norm": 0.05971822142601013, "learning_rate": 1.2311325125683754e-05, "loss": 0.0204, "step": 39160 }, { "action_loss": 0.002437076298519969, "epoch": 35.21582733812949, "step": 39160 }, { "epoch": 35.21582733812949, "step": 39160, "torque_loss": 0.17855584621429443 }, { "epoch": 35.22482014388489, "grad_norm": 0.08001008629798889, "learning_rate": 1.228960233513673e-05, "loss": 0.0153, "step": 39170 }, { "action_loss": 0.0006992164999246597, "epoch": 35.22482014388489, "step": 39170 }, { "epoch": 35.22482014388489, "step": 39170, "torque_loss": 0.1394682079553604 }, { "epoch": 35.23381294964029, "grad_norm": 0.054770391434431076, "learning_rate": 1.2267896040398836e-05, "loss": 0.0192, "step": 39180 }, { "action_loss": 0.0019952179864048958, "epoch": 35.23381294964029, "step": 39180 }, { "epoch": 35.23381294964029, "step": 39180, "torque_loss": 0.1579218953847885 }, { "epoch": 35.242805755395686, "grad_norm": 0.058269329369068146, "learning_rate": 1.2246206250965125e-05, "loss": 0.0153, "step": 39190 }, { "action_loss": 0.001708907657302916, "epoch": 35.242805755395686, "step": 39190 }, { "epoch": 35.242805755395686, "step": 39190, "torque_loss": 0.17791931331157684 }, { "epoch": 35.25179856115108, "grad_norm": 0.08410835266113281, "learning_rate": 1.2224532976323444e-05, "loss": 0.0174, "step": 39200 }, { "action_loss": 0.0014238854637369514, "epoch": 35.25179856115108, "step": 39200 }, { "epoch": 35.25179856115108, "step": 39200, "torque_loss": 0.1479945182800293 }, { "epoch": 35.260791366906474, "grad_norm": 0.05284886434674263, "learning_rate": 1.220287622595444e-05, "loss": 0.0176, "step": 39210 }, { "action_loss": 0.002872987650334835, "epoch": 35.260791366906474, "step": 39210 }, { "epoch": 35.260791366906474, "step": 39210, "torque_loss": 0.18116004765033722 }, { "epoch": 35.26978417266187, "grad_norm": 0.07230570912361145, "learning_rate": 1.2181236009331497e-05, "loss": 0.0153, "step": 39220 }, { "action_loss": 0.0007834804127924144, "epoch": 35.26978417266187, "step": 39220 }, { "epoch": 35.26978417266187, "step": 39220, "torque_loss": 0.12926992774009705 }, { "epoch": 35.27877697841727, "grad_norm": 0.06377770751714706, "learning_rate": 1.2159612335920784e-05, "loss": 0.0176, "step": 39230 }, { "action_loss": 0.0030368308071047068, "epoch": 35.27877697841727, "step": 39230 }, { "epoch": 35.27877697841727, "step": 39230, "torque_loss": 0.18355463445186615 }, { "epoch": 35.28776978417266, "grad_norm": 0.07403595000505447, "learning_rate": 1.2138005215181219e-05, "loss": 0.0203, "step": 39240 }, { "action_loss": 0.005997072905302048, "epoch": 35.28776978417266, "step": 39240 }, { "epoch": 35.28776978417266, "step": 39240, "torque_loss": 0.2006559818983078 }, { "epoch": 35.29676258992806, "grad_norm": 0.08650003373622894, "learning_rate": 1.2116414656564512e-05, "loss": 0.0187, "step": 39250 }, { "action_loss": 0.0008419298101216555, "epoch": 35.29676258992806, "step": 39250 }, { "epoch": 35.29676258992806, "step": 39250, "torque_loss": 0.10458674281835556 }, { "epoch": 35.305755395683455, "grad_norm": 0.09734980016946793, "learning_rate": 1.2094840669515095e-05, "loss": 0.0171, "step": 39260 }, { "action_loss": 0.0014140862040221691, "epoch": 35.305755395683455, "step": 39260 }, { "epoch": 35.305755395683455, "step": 39260, "torque_loss": 0.14885927736759186 }, { "epoch": 35.314748201438846, "grad_norm": 0.040815237909555435, "learning_rate": 1.2073283263470152e-05, "loss": 0.0185, "step": 39270 }, { "action_loss": 0.0008710599504411221, "epoch": 35.314748201438846, "step": 39270 }, { "epoch": 35.314748201438846, "step": 39270, "torque_loss": 0.10111608356237411 }, { "epoch": 35.32374100719424, "grad_norm": 0.038054246455430984, "learning_rate": 1.2051742447859653e-05, "loss": 0.0157, "step": 39280 }, { "action_loss": 0.001173189259134233, "epoch": 35.32374100719424, "step": 39280 }, { "epoch": 35.32374100719424, "step": 39280, "torque_loss": 0.13713304698467255 }, { "epoch": 35.33273381294964, "grad_norm": 0.052500586956739426, "learning_rate": 1.203021823210626e-05, "loss": 0.015, "step": 39290 }, { "action_loss": 0.0008782254881225526, "epoch": 35.33273381294964, "step": 39290 }, { "epoch": 35.33273381294964, "step": 39290, "torque_loss": 0.14786791801452637 }, { "epoch": 35.34172661870504, "grad_norm": 0.059975918382406235, "learning_rate": 1.2008710625625397e-05, "loss": 0.0176, "step": 39300 }, { "action_loss": 0.0010056113824248314, "epoch": 35.34172661870504, "step": 39300 }, { "epoch": 35.34172661870504, "step": 39300, "torque_loss": 0.18832510709762573 }, { "epoch": 35.35071942446043, "grad_norm": 0.06534845381975174, "learning_rate": 1.1987219637825236e-05, "loss": 0.0208, "step": 39310 }, { "action_loss": 0.00045877366210334003, "epoch": 35.35071942446043, "step": 39310 }, { "epoch": 35.35071942446043, "step": 39310, "torque_loss": 0.13842764496803284 }, { "epoch": 35.35971223021583, "grad_norm": 0.054303862154483795, "learning_rate": 1.1965745278106656e-05, "loss": 0.0206, "step": 39320 }, { "action_loss": 0.0016070250421762466, "epoch": 35.35971223021583, "step": 39320 }, { "epoch": 35.35971223021583, "step": 39320, "torque_loss": 0.12977053225040436 }, { "epoch": 35.368705035971225, "grad_norm": 0.08790136873722076, "learning_rate": 1.1944287555863259e-05, "loss": 0.0207, "step": 39330 }, { "action_loss": 0.00504640955477953, "epoch": 35.368705035971225, "step": 39330 }, { "epoch": 35.368705035971225, "step": 39330, "torque_loss": 0.23245175182819366 }, { "epoch": 35.37769784172662, "grad_norm": 0.1355903446674347, "learning_rate": 1.1922846480481409e-05, "loss": 0.0194, "step": 39340 }, { "action_loss": 0.0012302137911319733, "epoch": 35.37769784172662, "step": 39340 }, { "epoch": 35.37769784172662, "step": 39340, "torque_loss": 0.10167726129293442 }, { "epoch": 35.38669064748201, "grad_norm": 0.09253852814435959, "learning_rate": 1.1901422061340133e-05, "loss": 0.0161, "step": 39350 }, { "action_loss": 0.003083476098254323, "epoch": 35.38669064748201, "step": 39350 }, { "epoch": 35.38669064748201, "step": 39350, "torque_loss": 0.2218901365995407 }, { "epoch": 35.39568345323741, "grad_norm": 0.05885617434978485, "learning_rate": 1.1880014307811227e-05, "loss": 0.0188, "step": 39360 }, { "action_loss": 0.0009717806242406368, "epoch": 35.39568345323741, "step": 39360 }, { "epoch": 35.39568345323741, "step": 39360, "torque_loss": 0.12031609565019608 }, { "epoch": 35.40467625899281, "grad_norm": 0.07856699079275131, "learning_rate": 1.1858623229259142e-05, "loss": 0.0171, "step": 39370 }, { "action_loss": 0.0006585614755749702, "epoch": 35.40467625899281, "step": 39370 }, { "epoch": 35.40467625899281, "step": 39370, "torque_loss": 0.11528901010751724 }, { "epoch": 35.4136690647482, "grad_norm": 0.07185006141662598, "learning_rate": 1.1837248835041092e-05, "loss": 0.0179, "step": 39380 }, { "action_loss": 0.011827322654426098, "epoch": 35.4136690647482, "step": 39380 }, { "epoch": 35.4136690647482, "step": 39380, "torque_loss": 0.15854613482952118 }, { "epoch": 35.422661870503596, "grad_norm": 0.039713144302368164, "learning_rate": 1.1815891134506935e-05, "loss": 0.0223, "step": 39390 }, { "action_loss": 0.0035747757647186518, "epoch": 35.422661870503596, "step": 39390 }, { "epoch": 35.422661870503596, "step": 39390, "torque_loss": 0.18444472551345825 }, { "epoch": 35.431654676258994, "grad_norm": 0.10939273238182068, "learning_rate": 1.1794550136999277e-05, "loss": 0.0239, "step": 39400 }, { "action_loss": 0.003716015489771962, "epoch": 35.431654676258994, "step": 39400 }, { "epoch": 35.431654676258994, "step": 39400, "torque_loss": 0.18204011023044586 }, { "epoch": 35.44064748201439, "grad_norm": 0.041753873229026794, "learning_rate": 1.1773225851853387e-05, "loss": 0.0167, "step": 39410 }, { "action_loss": 0.00235662586055696, "epoch": 35.44064748201439, "step": 39410 }, { "epoch": 35.44064748201439, "step": 39410, "torque_loss": 0.11385053396224976 }, { "epoch": 35.44964028776978, "grad_norm": 0.06256745755672455, "learning_rate": 1.1751918288397235e-05, "loss": 0.019, "step": 39420 }, { "action_loss": 0.0020198095589876175, "epoch": 35.44964028776978, "step": 39420 }, { "epoch": 35.44964028776978, "step": 39420, "torque_loss": 0.15916325151920319 }, { "epoch": 35.45863309352518, "grad_norm": 0.09466572105884552, "learning_rate": 1.1730627455951449e-05, "loss": 0.0199, "step": 39430 }, { "action_loss": 0.0031406779307872057, "epoch": 35.45863309352518, "step": 39430 }, { "epoch": 35.45863309352518, "step": 39430, "torque_loss": 0.230691596865654 }, { "epoch": 35.46762589928058, "grad_norm": 0.06946948915719986, "learning_rate": 1.17093533638294e-05, "loss": 0.0167, "step": 39440 }, { "action_loss": 0.0010719302808865905, "epoch": 35.46762589928058, "step": 39440 }, { "epoch": 35.46762589928058, "step": 39440, "torque_loss": 0.14999715983867645 }, { "epoch": 35.476618705035975, "grad_norm": 0.192119762301445, "learning_rate": 1.168809602133708e-05, "loss": 0.0247, "step": 39450 }, { "action_loss": 0.0016671818448230624, "epoch": 35.476618705035975, "step": 39450 }, { "epoch": 35.476618705035975, "step": 39450, "torque_loss": 0.17012231051921844 }, { "epoch": 35.485611510791365, "grad_norm": 0.06358414143323898, "learning_rate": 1.1666855437773162e-05, "loss": 0.017, "step": 39460 }, { "action_loss": 0.002038917737081647, "epoch": 35.485611510791365, "step": 39460 }, { "epoch": 35.485611510791365, "step": 39460, "torque_loss": 0.14426153898239136 }, { "epoch": 35.49460431654676, "grad_norm": 0.041971106082201004, "learning_rate": 1.1645631622429026e-05, "loss": 0.0214, "step": 39470 }, { "action_loss": 0.0020663219038397074, "epoch": 35.49460431654676, "step": 39470 }, { "epoch": 35.49460431654676, "step": 39470, "torque_loss": 0.15660113096237183 }, { "epoch": 35.50359712230216, "grad_norm": 0.13788428902626038, "learning_rate": 1.1624424584588673e-05, "loss": 0.0188, "step": 39480 }, { "action_loss": 0.009845822118222713, "epoch": 35.50359712230216, "step": 39480 }, { "epoch": 35.50359712230216, "step": 39480, "torque_loss": 0.2776040732860565 }, { "epoch": 35.51258992805755, "grad_norm": 0.14317534863948822, "learning_rate": 1.1603234333528768e-05, "loss": 0.0226, "step": 39490 }, { "action_loss": 0.003274890361353755, "epoch": 35.51258992805755, "step": 39490 }, { "epoch": 35.51258992805755, "step": 39490, "torque_loss": 0.23547308146953583 }, { "epoch": 35.52158273381295, "grad_norm": 0.13849882781505585, "learning_rate": 1.1582060878518674e-05, "loss": 0.0184, "step": 39500 }, { "action_loss": 0.0023412087466567755, "epoch": 35.52158273381295, "step": 39500 }, { "epoch": 35.52158273381295, "step": 39500, "torque_loss": 0.14104504883289337 }, { "epoch": 35.530575539568346, "grad_norm": 0.05034421756863594, "learning_rate": 1.1560904228820375e-05, "loss": 0.0214, "step": 39510 }, { "action_loss": 0.0012747488217428327, "epoch": 35.530575539568346, "step": 39510 }, { "epoch": 35.530575539568346, "step": 39510, "torque_loss": 0.1303667277097702 }, { "epoch": 35.539568345323744, "grad_norm": 0.09827246516942978, "learning_rate": 1.1539764393688478e-05, "loss": 0.0167, "step": 39520 }, { "action_loss": 0.02021893300116062, "epoch": 35.539568345323744, "step": 39520 }, { "epoch": 35.539568345323744, "step": 39520, "torque_loss": 0.26354917883872986 }, { "epoch": 35.548561151079134, "grad_norm": 0.08893363177776337, "learning_rate": 1.151864138237031e-05, "loss": 0.0206, "step": 39530 }, { "action_loss": 0.005044812802225351, "epoch": 35.548561151079134, "step": 39530 }, { "epoch": 35.548561151079134, "step": 39530, "torque_loss": 0.14649830758571625 }, { "epoch": 35.55755395683453, "grad_norm": 0.06624574959278107, "learning_rate": 1.1497535204105769e-05, "loss": 0.0196, "step": 39540 }, { "action_loss": 0.0017108069732785225, "epoch": 35.55755395683453, "step": 39540 }, { "epoch": 35.55755395683453, "step": 39540, "torque_loss": 0.22920094430446625 }, { "epoch": 35.56654676258993, "grad_norm": 0.09247153252363205, "learning_rate": 1.14764458681274e-05, "loss": 0.0198, "step": 39550 }, { "action_loss": 0.0022171002347022295, "epoch": 35.56654676258993, "step": 39550 }, { "epoch": 35.56654676258993, "step": 39550, "torque_loss": 0.1980665773153305 }, { "epoch": 35.57553956834532, "grad_norm": 0.10862989723682404, "learning_rate": 1.1455373383660434e-05, "loss": 0.0209, "step": 39560 }, { "action_loss": 0.002364456420764327, "epoch": 35.57553956834532, "step": 39560 }, { "epoch": 35.57553956834532, "step": 39560, "torque_loss": 0.239026740193367 }, { "epoch": 35.58453237410072, "grad_norm": 0.0805584266781807, "learning_rate": 1.1434317759922664e-05, "loss": 0.0196, "step": 39570 }, { "action_loss": 0.0009240301442332566, "epoch": 35.58453237410072, "step": 39570 }, { "epoch": 35.58453237410072, "step": 39570, "torque_loss": 0.09621953964233398 }, { "epoch": 35.593525179856115, "grad_norm": 0.10640450567007065, "learning_rate": 1.1413279006124528e-05, "loss": 0.0211, "step": 39580 }, { "action_loss": 0.003612608416005969, "epoch": 35.593525179856115, "step": 39580 }, { "epoch": 35.593525179856115, "step": 39580, "torque_loss": 0.15319015085697174 }, { "epoch": 35.60251798561151, "grad_norm": 0.054962072521448135, "learning_rate": 1.1392257131469119e-05, "loss": 0.018, "step": 39590 }, { "action_loss": 0.003509873291477561, "epoch": 35.60251798561151, "step": 39590 }, { "epoch": 35.60251798561151, "step": 39590, "torque_loss": 0.1868697553873062 }, { "epoch": 35.611510791366904, "grad_norm": 0.04518958926200867, "learning_rate": 1.1371252145152095e-05, "loss": 0.0191, "step": 39600 }, { "action_loss": 0.0032444961834698915, "epoch": 35.611510791366904, "step": 39600 }, { "epoch": 35.611510791366904, "step": 39600, "torque_loss": 0.12318428605794907 }, { "epoch": 35.6205035971223, "grad_norm": 0.1072598248720169, "learning_rate": 1.1350264056361776e-05, "loss": 0.021, "step": 39610 }, { "action_loss": 0.0025497155729681253, "epoch": 35.6205035971223, "step": 39610 }, { "epoch": 35.6205035971223, "step": 39610, "torque_loss": 0.1906116008758545 }, { "epoch": 35.6294964028777, "grad_norm": 0.07885411381721497, "learning_rate": 1.1329292874279029e-05, "loss": 0.0183, "step": 39620 }, { "action_loss": 0.001029249862767756, "epoch": 35.6294964028777, "step": 39620 }, { "epoch": 35.6294964028777, "step": 39620, "torque_loss": 0.1072642132639885 }, { "epoch": 35.638489208633096, "grad_norm": 0.05739763751626015, "learning_rate": 1.1308338608077401e-05, "loss": 0.0163, "step": 39630 }, { "action_loss": 0.0031547124963253736, "epoch": 35.638489208633096, "step": 39630 }, { "epoch": 35.638489208633096, "step": 39630, "torque_loss": 0.19463485479354858 }, { "epoch": 35.64748201438849, "grad_norm": 0.11190634965896606, "learning_rate": 1.1287401266922981e-05, "loss": 0.0206, "step": 39640 }, { "action_loss": 0.0028232391923666, "epoch": 35.64748201438849, "step": 39640 }, { "epoch": 35.64748201438849, "step": 39640, "torque_loss": 0.13005989789962769 }, { "epoch": 35.656474820143885, "grad_norm": 0.07783986628055573, "learning_rate": 1.126648085997446e-05, "loss": 0.0215, "step": 39650 }, { "action_loss": 0.0023863029200583696, "epoch": 35.656474820143885, "step": 39650 }, { "epoch": 35.656474820143885, "step": 39650, "torque_loss": 0.1635221689939499 }, { "epoch": 35.66546762589928, "grad_norm": 0.07601258903741837, "learning_rate": 1.124557739638316e-05, "loss": 0.0168, "step": 39660 }, { "action_loss": 0.0017852289602160454, "epoch": 35.66546762589928, "step": 39660 }, { "epoch": 35.66546762589928, "step": 39660, "torque_loss": 0.1607973575592041 }, { "epoch": 35.67446043165468, "grad_norm": 0.053558599203825, "learning_rate": 1.1224690885292955e-05, "loss": 0.0218, "step": 39670 }, { "action_loss": 0.0006309084710665047, "epoch": 35.67446043165468, "step": 39670 }, { "epoch": 35.67446043165468, "step": 39670, "torque_loss": 0.1440623700618744 }, { "epoch": 35.68345323741007, "grad_norm": 0.11147014796733856, "learning_rate": 1.1203821335840303e-05, "loss": 0.0205, "step": 39680 }, { "action_loss": 0.0022620062809437513, "epoch": 35.68345323741007, "step": 39680 }, { "epoch": 35.68345323741007, "step": 39680, "torque_loss": 0.18588991463184357 }, { "epoch": 35.69244604316547, "grad_norm": 0.06986290216445923, "learning_rate": 1.1182968757154278e-05, "loss": 0.0189, "step": 39690 }, { "action_loss": 0.004676717799156904, "epoch": 35.69244604316547, "step": 39690 }, { "epoch": 35.69244604316547, "step": 39690, "torque_loss": 0.18100230395793915 }, { "epoch": 35.701438848920866, "grad_norm": 0.06297983974218369, "learning_rate": 1.1162133158356492e-05, "loss": 0.022, "step": 39700 }, { "action_loss": 0.002204609103500843, "epoch": 35.701438848920866, "step": 39700 }, { "epoch": 35.701438848920866, "step": 39700, "torque_loss": 0.16077536344528198 }, { "epoch": 35.710431654676256, "grad_norm": 0.08147146552801132, "learning_rate": 1.1141314548561133e-05, "loss": 0.022, "step": 39710 }, { "action_loss": 0.001988602802157402, "epoch": 35.710431654676256, "step": 39710 }, { "epoch": 35.710431654676256, "step": 39710, "torque_loss": 0.16300463676452637 }, { "epoch": 35.719424460431654, "grad_norm": 0.14188434183597565, "learning_rate": 1.1120512936874994e-05, "loss": 0.02, "step": 39720 }, { "action_loss": 0.0012708690483123064, "epoch": 35.719424460431654, "step": 39720 }, { "epoch": 35.719424460431654, "step": 39720, "torque_loss": 0.1240919753909111 }, { "epoch": 35.72841726618705, "grad_norm": 0.08475551754236221, "learning_rate": 1.109972833239739e-05, "loss": 0.0184, "step": 39730 }, { "action_loss": 0.0021963119506835938, "epoch": 35.72841726618705, "step": 39730 }, { "epoch": 35.72841726618705, "step": 39730, "torque_loss": 0.13194502890110016 }, { "epoch": 35.73741007194245, "grad_norm": 0.12713085114955902, "learning_rate": 1.1078960744220218e-05, "loss": 0.0207, "step": 39740 }, { "action_loss": 0.0021947703789919615, "epoch": 35.73741007194245, "step": 39740 }, { "epoch": 35.73741007194245, "step": 39740, "torque_loss": 0.23871736228466034 }, { "epoch": 35.74640287769784, "grad_norm": 0.12905898690223694, "learning_rate": 1.1058210181427914e-05, "loss": 0.0176, "step": 39750 }, { "action_loss": 0.003979666158556938, "epoch": 35.74640287769784, "step": 39750 }, { "epoch": 35.74640287769784, "step": 39750, "torque_loss": 0.20033545792102814 }, { "epoch": 35.75539568345324, "grad_norm": 0.053538452833890915, "learning_rate": 1.10374766530975e-05, "loss": 0.0175, "step": 39760 }, { "action_loss": 0.002967281499877572, "epoch": 35.75539568345324, "step": 39760 }, { "epoch": 35.75539568345324, "step": 39760, "torque_loss": 0.14780502021312714 }, { "epoch": 35.764388489208635, "grad_norm": 0.04613998532295227, "learning_rate": 1.1016760168298518e-05, "loss": 0.019, "step": 39770 }, { "action_loss": 0.0027772728353738785, "epoch": 35.764388489208635, "step": 39770 }, { "epoch": 35.764388489208635, "step": 39770, "torque_loss": 0.12991635501384735 }, { "epoch": 35.773381294964025, "grad_norm": 0.04901827871799469, "learning_rate": 1.0996060736093044e-05, "loss": 0.0188, "step": 39780 }, { "action_loss": 0.0056060864590108395, "epoch": 35.773381294964025, "step": 39780 }, { "epoch": 35.773381294964025, "step": 39780, "torque_loss": 0.1947043389081955 }, { "epoch": 35.78237410071942, "grad_norm": 0.03701317682862282, "learning_rate": 1.0975378365535749e-05, "loss": 0.0176, "step": 39790 }, { "action_loss": 0.0018783222185447812, "epoch": 35.78237410071942, "step": 39790 }, { "epoch": 35.78237410071942, "step": 39790, "torque_loss": 0.13147252798080444 }, { "epoch": 35.79136690647482, "grad_norm": 0.045410510152578354, "learning_rate": 1.0954713065673777e-05, "loss": 0.0188, "step": 39800 }, { "action_loss": 0.0008452872280031443, "epoch": 35.79136690647482, "step": 39800 }, { "epoch": 35.79136690647482, "step": 39800, "torque_loss": 0.16039526462554932 }, { "epoch": 35.80035971223022, "grad_norm": 0.06660544127225876, "learning_rate": 1.0934064845546832e-05, "loss": 0.0159, "step": 39810 }, { "action_loss": 0.01950039528310299, "epoch": 35.80035971223022, "step": 39810 }, { "epoch": 35.80035971223022, "step": 39810, "torque_loss": 0.18943250179290771 }, { "epoch": 35.80935251798561, "grad_norm": 0.11757684499025345, "learning_rate": 1.0913433714187166e-05, "loss": 0.0202, "step": 39820 }, { "action_loss": 0.0009459996945224702, "epoch": 35.80935251798561, "step": 39820 }, { "epoch": 35.80935251798561, "step": 39820, "torque_loss": 0.11733973771333694 }, { "epoch": 35.818345323741006, "grad_norm": 0.05428013205528259, "learning_rate": 1.089281968061951e-05, "loss": 0.0192, "step": 39830 }, { "action_loss": 0.0014503203565254807, "epoch": 35.818345323741006, "step": 39830 }, { "epoch": 35.818345323741006, "step": 39830, "torque_loss": 0.1560855358839035 }, { "epoch": 35.827338129496404, "grad_norm": 0.04442354664206505, "learning_rate": 1.0872222753861172e-05, "loss": 0.018, "step": 39840 }, { "action_loss": 0.003899435279890895, "epoch": 35.827338129496404, "step": 39840 }, { "epoch": 35.827338129496404, "step": 39840, "torque_loss": 0.18760107457637787 }, { "epoch": 35.8363309352518, "grad_norm": 0.061576589941978455, "learning_rate": 1.0851642942921925e-05, "loss": 0.0163, "step": 39850 }, { "action_loss": 0.00881543755531311, "epoch": 35.8363309352518, "step": 39850 }, { "epoch": 35.8363309352518, "step": 39850, "torque_loss": 0.18093134462833405 }, { "epoch": 35.84532374100719, "grad_norm": 0.08678477257490158, "learning_rate": 1.0831080256804093e-05, "loss": 0.0175, "step": 39860 }, { "action_loss": 0.0012589830439537764, "epoch": 35.84532374100719, "step": 39860 }, { "epoch": 35.84532374100719, "step": 39860, "torque_loss": 0.10891128331422806 }, { "epoch": 35.85431654676259, "grad_norm": 0.0592028945684433, "learning_rate": 1.0810534704502478e-05, "loss": 0.0171, "step": 39870 }, { "action_loss": 0.0017991522327065468, "epoch": 35.85431654676259, "step": 39870 }, { "epoch": 35.85431654676259, "step": 39870, "torque_loss": 0.14371564984321594 }, { "epoch": 35.86330935251799, "grad_norm": 0.07568463683128357, "learning_rate": 1.0790006295004423e-05, "loss": 0.0197, "step": 39880 }, { "action_loss": 0.002098644617944956, "epoch": 35.86330935251799, "step": 39880 }, { "epoch": 35.86330935251799, "step": 39880, "torque_loss": 0.24698756635189056 }, { "epoch": 35.87230215827338, "grad_norm": 0.07035006582736969, "learning_rate": 1.076949503728974e-05, "loss": 0.0175, "step": 39890 }, { "action_loss": 0.0041990564204752445, "epoch": 35.87230215827338, "step": 39890 }, { "epoch": 35.87230215827338, "step": 39890, "torque_loss": 0.16593647003173828 }, { "epoch": 35.881294964028775, "grad_norm": 0.04291374981403351, "learning_rate": 1.074900094033074e-05, "loss": 0.0219, "step": 39900 }, { "action_loss": 0.0007594072376377881, "epoch": 35.881294964028775, "step": 39900 }, { "epoch": 35.881294964028775, "step": 39900, "torque_loss": 0.08204134553670883 }, { "epoch": 35.89028776978417, "grad_norm": 0.0811246708035469, "learning_rate": 1.0728524013092256e-05, "loss": 0.0162, "step": 39910 }, { "action_loss": 0.00135989545378834, "epoch": 35.89028776978417, "step": 39910 }, { "epoch": 35.89028776978417, "step": 39910, "torque_loss": 0.11029419302940369 }, { "epoch": 35.89928057553957, "grad_norm": 0.10351167619228363, "learning_rate": 1.0708064264531581e-05, "loss": 0.0176, "step": 39920 }, { "action_loss": 0.0035610359627753496, "epoch": 35.89928057553957, "step": 39920 }, { "epoch": 35.89928057553957, "step": 39920, "torque_loss": 0.18977807462215424 }, { "epoch": 35.90827338129496, "grad_norm": 0.07573878765106201, "learning_rate": 1.068762170359851e-05, "loss": 0.0179, "step": 39930 }, { "action_loss": 0.007748974487185478, "epoch": 35.90827338129496, "step": 39930 }, { "epoch": 35.90827338129496, "step": 39930, "torque_loss": 0.22742073237895966 }, { "epoch": 35.91726618705036, "grad_norm": 0.1335897147655487, "learning_rate": 1.0667196339235297e-05, "loss": 0.0205, "step": 39940 }, { "action_loss": 0.001160156331025064, "epoch": 35.91726618705036, "step": 39940 }, { "epoch": 35.91726618705036, "step": 39940, "torque_loss": 0.13546110689640045 }, { "epoch": 35.92625899280576, "grad_norm": 0.1038166955113411, "learning_rate": 1.0646788180376716e-05, "loss": 0.0156, "step": 39950 }, { "action_loss": 0.0013579869410023093, "epoch": 35.92625899280576, "step": 39950 }, { "epoch": 35.92625899280576, "step": 39950, "torque_loss": 0.15731869637966156 }, { "epoch": 35.935251798561154, "grad_norm": 0.0379953607916832, "learning_rate": 1.0626397235949976e-05, "loss": 0.0184, "step": 39960 }, { "action_loss": 0.0010266390163451433, "epoch": 35.935251798561154, "step": 39960 }, { "epoch": 35.935251798561154, "step": 39960, "torque_loss": 0.14089886844158173 }, { "epoch": 35.944244604316545, "grad_norm": 0.08034412562847137, "learning_rate": 1.0606023514874751e-05, "loss": 0.0203, "step": 39970 }, { "action_loss": 0.00436632614582777, "epoch": 35.944244604316545, "step": 39970 }, { "epoch": 35.944244604316545, "step": 39970, "torque_loss": 0.17826427519321442 }, { "epoch": 35.95323741007194, "grad_norm": 0.13995970785617828, "learning_rate": 1.0585667026063233e-05, "loss": 0.0201, "step": 39980 }, { "action_loss": 0.0007700046407990158, "epoch": 35.95323741007194, "step": 39980 }, { "epoch": 35.95323741007194, "step": 39980, "torque_loss": 0.0781567171216011 }, { "epoch": 35.96223021582734, "grad_norm": 0.05317391827702522, "learning_rate": 1.0565327778420025e-05, "loss": 0.015, "step": 39990 }, { "action_loss": 0.0017427640268579125, "epoch": 35.96223021582734, "step": 39990 }, { "epoch": 35.96223021582734, "step": 39990, "torque_loss": 0.178803488612175 }, { "epoch": 35.97122302158273, "grad_norm": 0.06977489590644836, "learning_rate": 1.0545005780842193e-05, "loss": 0.0173, "step": 40000 }, { "action_loss": 0.0007345918565988541, "epoch": 35.97122302158273, "step": 40000 }, { "epoch": 35.97122302158273, "step": 40000, "torque_loss": 0.09197504073381424 }, { "epoch": 35.98021582733813, "grad_norm": 0.06809783726930618, "learning_rate": 1.0524701042219299e-05, "loss": 0.0174, "step": 40010 }, { "action_loss": 0.0008511199266649783, "epoch": 35.98021582733813, "step": 40010 }, { "epoch": 35.98021582733813, "step": 40010, "torque_loss": 0.13152897357940674 }, { "epoch": 35.989208633093526, "grad_norm": 0.10137112438678741, "learning_rate": 1.0504413571433303e-05, "loss": 0.0193, "step": 40020 }, { "action_loss": 0.0018825792940333486, "epoch": 35.989208633093526, "step": 40020 }, { "epoch": 35.989208633093526, "step": 40020, "torque_loss": 0.2008628398180008 }, { "epoch": 35.99820143884892, "grad_norm": 0.03879229724407196, "learning_rate": 1.0484143377358624e-05, "loss": 0.0185, "step": 40030 }, { "action_loss": 0.0011088162427768111, "epoch": 35.99820143884892, "step": 40030 }, { "epoch": 35.99820143884892, "step": 40030, "torque_loss": 0.16796667873859406 }, { "epoch": 36.007194244604314, "grad_norm": 0.05679899454116821, "learning_rate": 1.046389046886217e-05, "loss": 0.0164, "step": 40040 }, { "action_loss": 0.001401111134327948, "epoch": 36.007194244604314, "step": 40040 }, { "epoch": 36.007194244604314, "step": 40040, "torque_loss": 0.14154386520385742 }, { "epoch": 36.01618705035971, "grad_norm": 0.0755242332816124, "learning_rate": 1.044365485480322e-05, "loss": 0.0171, "step": 40050 }, { "action_loss": 0.004859426524490118, "epoch": 36.01618705035971, "step": 40050 }, { "epoch": 36.01618705035971, "step": 40050, "torque_loss": 0.17595036327838898 }, { "epoch": 36.02517985611511, "grad_norm": 0.1360095739364624, "learning_rate": 1.0423436544033522e-05, "loss": 0.0194, "step": 40060 }, { "action_loss": 0.000715860805939883, "epoch": 36.02517985611511, "step": 40060 }, { "epoch": 36.02517985611511, "step": 40060, "torque_loss": 0.0974721685051918 }, { "epoch": 36.03417266187051, "grad_norm": 0.0959271788597107, "learning_rate": 1.0403235545397272e-05, "loss": 0.0169, "step": 40070 }, { "action_loss": 0.0009603537037037313, "epoch": 36.03417266187051, "step": 40070 }, { "epoch": 36.03417266187051, "step": 40070, "torque_loss": 0.15385277569293976 }, { "epoch": 36.0431654676259, "grad_norm": 0.08439351618289948, "learning_rate": 1.0383051867731043e-05, "loss": 0.0213, "step": 40080 }, { "action_loss": 0.0014979118714109063, "epoch": 36.0431654676259, "step": 40080 }, { "epoch": 36.0431654676259, "step": 40080, "torque_loss": 0.1351558268070221 }, { "epoch": 36.052158273381295, "grad_norm": 0.05832209438085556, "learning_rate": 1.0362885519863896e-05, "loss": 0.0164, "step": 40090 }, { "action_loss": 0.0014581605792045593, "epoch": 36.052158273381295, "step": 40090 }, { "epoch": 36.052158273381295, "step": 40090, "torque_loss": 0.1011371985077858 }, { "epoch": 36.06115107913669, "grad_norm": 0.09453587234020233, "learning_rate": 1.034273651061724e-05, "loss": 0.0189, "step": 40100 }, { "action_loss": 0.0011546831810846925, "epoch": 36.06115107913669, "step": 40100 }, { "epoch": 36.06115107913669, "step": 40100, "torque_loss": 0.15925608575344086 }, { "epoch": 36.07014388489208, "grad_norm": 0.06702969968318939, "learning_rate": 1.0322604848804973e-05, "loss": 0.0221, "step": 40110 }, { "action_loss": 0.0029998281970620155, "epoch": 36.07014388489208, "step": 40110 }, { "epoch": 36.07014388489208, "step": 40110, "torque_loss": 0.1532248854637146 }, { "epoch": 36.07913669064748, "grad_norm": 0.04648956283926964, "learning_rate": 1.030249054323335e-05, "loss": 0.0171, "step": 40120 }, { "action_loss": 0.0008248367230407894, "epoch": 36.07913669064748, "step": 40120 }, { "epoch": 36.07913669064748, "step": 40120, "torque_loss": 0.10535504668951035 }, { "epoch": 36.08812949640288, "grad_norm": 0.06437236815690994, "learning_rate": 1.0282393602701041e-05, "loss": 0.019, "step": 40130 }, { "action_loss": 0.0011240496532991529, "epoch": 36.08812949640288, "step": 40130 }, { "epoch": 36.08812949640288, "step": 40130, "torque_loss": 0.12464926391839981 }, { "epoch": 36.097122302158276, "grad_norm": 0.09780384600162506, "learning_rate": 1.026231403599915e-05, "loss": 0.0179, "step": 40140 }, { "action_loss": 0.004050246439874172, "epoch": 36.097122302158276, "step": 40140 }, { "epoch": 36.097122302158276, "step": 40140, "torque_loss": 0.16782468557357788 }, { "epoch": 36.106115107913666, "grad_norm": 0.09070917963981628, "learning_rate": 1.0242251851911155e-05, "loss": 0.0192, "step": 40150 }, { "action_loss": 0.0008819131180644035, "epoch": 36.106115107913666, "step": 40150 }, { "epoch": 36.106115107913666, "step": 40150, "torque_loss": 0.10035437345504761 }, { "epoch": 36.115107913669064, "grad_norm": 0.04397730156779289, "learning_rate": 1.0222207059212918e-05, "loss": 0.0155, "step": 40160 }, { "action_loss": 0.0007970877340994775, "epoch": 36.115107913669064, "step": 40160 }, { "epoch": 36.115107913669064, "step": 40160, "torque_loss": 0.11025583744049072 }, { "epoch": 36.12410071942446, "grad_norm": 0.0942092165350914, "learning_rate": 1.0202179666672746e-05, "loss": 0.0163, "step": 40170 }, { "action_loss": 0.002500703325495124, "epoch": 36.12410071942446, "step": 40170 }, { "epoch": 36.12410071942446, "step": 40170, "torque_loss": 0.23583857715129852 }, { "epoch": 36.13309352517986, "grad_norm": 0.05334809049963951, "learning_rate": 1.0182169683051284e-05, "loss": 0.0195, "step": 40180 }, { "action_loss": 0.0006878346321173012, "epoch": 36.13309352517986, "step": 40180 }, { "epoch": 36.13309352517986, "step": 40180, "torque_loss": 0.08487850427627563 }, { "epoch": 36.14208633093525, "grad_norm": 0.08197461813688278, "learning_rate": 1.0162177117101567e-05, "loss": 0.0195, "step": 40190 }, { "action_loss": 0.0008092137868516147, "epoch": 36.14208633093525, "step": 40190 }, { "epoch": 36.14208633093525, "step": 40190, "torque_loss": 0.1075054407119751 }, { "epoch": 36.15107913669065, "grad_norm": 0.11852174252271652, "learning_rate": 1.0142201977569044e-05, "loss": 0.0162, "step": 40200 }, { "action_loss": 0.001782472594641149, "epoch": 36.15107913669065, "step": 40200 }, { "epoch": 36.15107913669065, "step": 40200, "torque_loss": 0.13796444237232208 }, { "epoch": 36.160071942446045, "grad_norm": 0.06246250495314598, "learning_rate": 1.0122244273191517e-05, "loss": 0.0199, "step": 40210 }, { "action_loss": 0.0008548746700398624, "epoch": 36.160071942446045, "step": 40210 }, { "epoch": 36.160071942446045, "step": 40210, "torque_loss": 0.14070983231067657 }, { "epoch": 36.169064748201436, "grad_norm": 0.05701175704598427, "learning_rate": 1.0102304012699137e-05, "loss": 0.0136, "step": 40220 }, { "action_loss": 0.0009531136602163315, "epoch": 36.169064748201436, "step": 40220 }, { "epoch": 36.169064748201436, "step": 40220, "torque_loss": 0.1805793046951294 }, { "epoch": 36.17805755395683, "grad_norm": 0.03623141348361969, "learning_rate": 1.0082381204814484e-05, "loss": 0.0165, "step": 40230 }, { "action_loss": 0.0017861599335446954, "epoch": 36.17805755395683, "step": 40230 }, { "epoch": 36.17805755395683, "step": 40230, "torque_loss": 0.18298578262329102 }, { "epoch": 36.18705035971223, "grad_norm": 0.051190465688705444, "learning_rate": 1.0062475858252463e-05, "loss": 0.0179, "step": 40240 }, { "action_loss": 0.0035618122201412916, "epoch": 36.18705035971223, "step": 40240 }, { "epoch": 36.18705035971223, "step": 40240, "torque_loss": 0.11424124240875244 }, { "epoch": 36.19604316546763, "grad_norm": 0.07086443156003952, "learning_rate": 1.0042587981720326e-05, "loss": 0.0161, "step": 40250 }, { "action_loss": 0.0010626636212691665, "epoch": 36.19604316546763, "step": 40250 }, { "epoch": 36.19604316546763, "step": 40250, "torque_loss": 0.12697486579418182 }, { "epoch": 36.20503597122302, "grad_norm": 0.11181472986936569, "learning_rate": 1.002271758391774e-05, "loss": 0.0177, "step": 40260 }, { "action_loss": 0.0041012242436409, "epoch": 36.20503597122302, "step": 40260 }, { "epoch": 36.20503597122302, "step": 40260, "torque_loss": 0.14911524951457977 }, { "epoch": 36.21402877697842, "grad_norm": 0.1135302484035492, "learning_rate": 1.0002864673536676e-05, "loss": 0.0202, "step": 40270 }, { "action_loss": 0.0009501318563707173, "epoch": 36.21402877697842, "step": 40270 }, { "epoch": 36.21402877697842, "step": 40270, "torque_loss": 0.10493654012680054 }, { "epoch": 36.223021582733814, "grad_norm": 0.10824653506278992, "learning_rate": 9.983029259261473e-06, "loss": 0.0199, "step": 40280 }, { "action_loss": 0.0026902572717517614, "epoch": 36.223021582733814, "step": 40280 }, { "epoch": 36.223021582733814, "step": 40280, "torque_loss": 0.14473769068717957 }, { "epoch": 36.23201438848921, "grad_norm": 0.07024800032377243, "learning_rate": 9.963211349768798e-06, "loss": 0.0197, "step": 40290 }, { "action_loss": 0.001129840617068112, "epoch": 36.23201438848921, "step": 40290 }, { "epoch": 36.23201438848921, "step": 40290, "torque_loss": 0.13286978006362915 }, { "epoch": 36.2410071942446, "grad_norm": 0.09734983742237091, "learning_rate": 9.94341095372771e-06, "loss": 0.0187, "step": 40300 }, { "action_loss": 0.0015948059735819697, "epoch": 36.2410071942446, "step": 40300 }, { "epoch": 36.2410071942446, "step": 40300, "torque_loss": 0.12468812614679337 }, { "epoch": 36.25, "grad_norm": 0.12328631430864334, "learning_rate": 9.923628079799546e-06, "loss": 0.0205, "step": 40310 }, { "action_loss": 0.003462405176833272, "epoch": 36.25, "step": 40310 }, { "epoch": 36.25, "step": 40310, "torque_loss": 0.15464796125888824 }, { "epoch": 36.2589928057554, "grad_norm": 0.05570133030414581, "learning_rate": 9.903862736638031e-06, "loss": 0.0206, "step": 40320 }, { "action_loss": 0.0015662070363759995, "epoch": 36.2589928057554, "step": 40320 }, { "epoch": 36.2589928057554, "step": 40320, "torque_loss": 0.1256508231163025 }, { "epoch": 36.26798561151079, "grad_norm": 0.089267298579216, "learning_rate": 9.884114932889171e-06, "loss": 0.0168, "step": 40330 }, { "action_loss": 0.0026596852112561464, "epoch": 36.26798561151079, "step": 40330 }, { "epoch": 36.26798561151079, "step": 40330, "torque_loss": 0.18719907104969025 }, { "epoch": 36.276978417266186, "grad_norm": 0.0997847318649292, "learning_rate": 9.864384677191357e-06, "loss": 0.0182, "step": 40340 }, { "action_loss": 0.001047169091179967, "epoch": 36.276978417266186, "step": 40340 }, { "epoch": 36.276978417266186, "step": 40340, "torque_loss": 0.16348202526569366 }, { "epoch": 36.28597122302158, "grad_norm": 0.10165919363498688, "learning_rate": 9.844671978175247e-06, "loss": 0.0216, "step": 40350 }, { "action_loss": 0.002221146132797003, "epoch": 36.28597122302158, "step": 40350 }, { "epoch": 36.28597122302158, "step": 40350, "torque_loss": 0.14897333085536957 }, { "epoch": 36.29496402877698, "grad_norm": 0.0795971155166626, "learning_rate": 9.824976844463862e-06, "loss": 0.0194, "step": 40360 }, { "action_loss": 0.0020545555744320154, "epoch": 36.29496402877698, "step": 40360 }, { "epoch": 36.29496402877698, "step": 40360, "torque_loss": 0.21863989531993866 }, { "epoch": 36.30395683453237, "grad_norm": 0.037828121334314346, "learning_rate": 9.805299284672525e-06, "loss": 0.0181, "step": 40370 }, { "action_loss": 0.0010303640738129616, "epoch": 36.30395683453237, "step": 40370 }, { "epoch": 36.30395683453237, "step": 40370, "torque_loss": 0.16111436486244202 }, { "epoch": 36.31294964028777, "grad_norm": 0.09328702092170715, "learning_rate": 9.785639307408845e-06, "loss": 0.0193, "step": 40380 }, { "action_loss": 0.0013219774700701237, "epoch": 36.31294964028777, "step": 40380 }, { "epoch": 36.31294964028777, "step": 40380, "torque_loss": 0.15946799516677856 }, { "epoch": 36.32194244604317, "grad_norm": 0.0498848482966423, "learning_rate": 9.765996921272792e-06, "loss": 0.016, "step": 40390 }, { "action_loss": 0.011749278753995895, "epoch": 36.32194244604317, "step": 40390 }, { "epoch": 36.32194244604317, "step": 40390, "torque_loss": 0.23398633301258087 }, { "epoch": 36.330935251798564, "grad_norm": 0.04225704073905945, "learning_rate": 9.746372134856595e-06, "loss": 0.0318, "step": 40400 }, { "action_loss": 0.001433639437891543, "epoch": 36.330935251798564, "step": 40400 }, { "epoch": 36.330935251798564, "step": 40400, "torque_loss": 0.14855940639972687 }, { "epoch": 36.339928057553955, "grad_norm": 0.07003213465213776, "learning_rate": 9.726764956744805e-06, "loss": 0.0193, "step": 40410 }, { "action_loss": 0.0016303147422149777, "epoch": 36.339928057553955, "step": 40410 }, { "epoch": 36.339928057553955, "step": 40410, "torque_loss": 0.16185852885246277 }, { "epoch": 36.34892086330935, "grad_norm": 0.08341187238693237, "learning_rate": 9.707175395514278e-06, "loss": 0.0179, "step": 40420 }, { "action_loss": 0.0009093406260944903, "epoch": 36.34892086330935, "step": 40420 }, { "epoch": 36.34892086330935, "step": 40420, "torque_loss": 0.09284073114395142 }, { "epoch": 36.35791366906475, "grad_norm": 0.0698714330792427, "learning_rate": 9.68760345973415e-06, "loss": 0.0172, "step": 40430 }, { "action_loss": 0.0016066305106505752, "epoch": 36.35791366906475, "step": 40430 }, { "epoch": 36.35791366906475, "step": 40430, "torque_loss": 0.13102972507476807 }, { "epoch": 36.36690647482014, "grad_norm": 0.08843961358070374, "learning_rate": 9.668049157965848e-06, "loss": 0.021, "step": 40440 }, { "action_loss": 0.0006687199347652495, "epoch": 36.36690647482014, "step": 40440 }, { "epoch": 36.36690647482014, "step": 40440, "torque_loss": 0.18001538515090942 }, { "epoch": 36.37589928057554, "grad_norm": 0.09183326363563538, "learning_rate": 9.648512498763074e-06, "loss": 0.0175, "step": 40450 }, { "action_loss": 0.004699326120316982, "epoch": 36.37589928057554, "step": 40450 }, { "epoch": 36.37589928057554, "step": 40450, "torque_loss": 0.1413063406944275 }, { "epoch": 36.384892086330936, "grad_norm": 0.0789666622877121, "learning_rate": 9.62899349067185e-06, "loss": 0.0164, "step": 40460 }, { "action_loss": 0.000646314409095794, "epoch": 36.384892086330936, "step": 40460 }, { "epoch": 36.384892086330936, "step": 40460, "torque_loss": 0.10635096579790115 }, { "epoch": 36.393884892086334, "grad_norm": 0.05648427456617355, "learning_rate": 9.609492142230442e-06, "loss": 0.0183, "step": 40470 }, { "action_loss": 0.0010652769124135375, "epoch": 36.393884892086334, "step": 40470 }, { "epoch": 36.393884892086334, "step": 40470, "torque_loss": 0.15429387986660004 }, { "epoch": 36.402877697841724, "grad_norm": 0.04936891049146652, "learning_rate": 9.59000846196939e-06, "loss": 0.0211, "step": 40480 }, { "action_loss": 0.0011174647370353341, "epoch": 36.402877697841724, "step": 40480 }, { "epoch": 36.402877697841724, "step": 40480, "torque_loss": 0.10114756971597672 }, { "epoch": 36.41187050359712, "grad_norm": 0.10939277708530426, "learning_rate": 9.570542458411552e-06, "loss": 0.0163, "step": 40490 }, { "action_loss": 0.0016367457574233413, "epoch": 36.41187050359712, "step": 40490 }, { "epoch": 36.41187050359712, "step": 40490, "torque_loss": 0.11559329181909561 }, { "epoch": 36.42086330935252, "grad_norm": 0.038395557552576065, "learning_rate": 9.551094140071993e-06, "loss": 0.015, "step": 40500 }, { "action_loss": 0.0022012796252965927, "epoch": 36.42086330935252, "step": 40500 }, { "epoch": 36.42086330935252, "step": 40500, "torque_loss": 0.13404320180416107 }, { "epoch": 36.42985611510792, "grad_norm": 0.04773832485079765, "learning_rate": 9.53166351545806e-06, "loss": 0.02, "step": 40510 }, { "action_loss": 0.00097272964194417, "epoch": 36.42985611510792, "step": 40510 }, { "epoch": 36.42985611510792, "step": 40510, "torque_loss": 0.1720360964536667 }, { "epoch": 36.43884892086331, "grad_norm": 0.07244818657636642, "learning_rate": 9.512250593069393e-06, "loss": 0.0178, "step": 40520 }, { "action_loss": 0.002809508005157113, "epoch": 36.43884892086331, "step": 40520 }, { "epoch": 36.43884892086331, "step": 40520, "torque_loss": 0.1297025829553604 }, { "epoch": 36.447841726618705, "grad_norm": 0.07645133137702942, "learning_rate": 9.492855381397859e-06, "loss": 0.0202, "step": 40530 }, { "action_loss": 0.0032468989957123995, "epoch": 36.447841726618705, "step": 40530 }, { "epoch": 36.447841726618705, "step": 40530, "torque_loss": 0.12541572749614716 }, { "epoch": 36.4568345323741, "grad_norm": 0.053344160318374634, "learning_rate": 9.473477888927556e-06, "loss": 0.0176, "step": 40540 }, { "action_loss": 0.002234363229945302, "epoch": 36.4568345323741, "step": 40540 }, { "epoch": 36.4568345323741, "step": 40540, "torque_loss": 0.1519530564546585 }, { "epoch": 36.46582733812949, "grad_norm": 0.03861289843916893, "learning_rate": 9.454118124134892e-06, "loss": 0.0176, "step": 40550 }, { "action_loss": 0.0009085524943657219, "epoch": 36.46582733812949, "step": 40550 }, { "epoch": 36.46582733812949, "step": 40550, "torque_loss": 0.12516851723194122 }, { "epoch": 36.47482014388489, "grad_norm": 0.06776393204927444, "learning_rate": 9.434776095488463e-06, "loss": 0.0193, "step": 40560 }, { "action_loss": 0.0007719279383309186, "epoch": 36.47482014388489, "step": 40560 }, { "epoch": 36.47482014388489, "step": 40560, "torque_loss": 0.12374768406152725 }, { "epoch": 36.48381294964029, "grad_norm": 0.052394550293684006, "learning_rate": 9.41545181144915e-06, "loss": 0.0156, "step": 40570 }, { "action_loss": 0.0006594588048756123, "epoch": 36.48381294964029, "step": 40570 }, { "epoch": 36.48381294964029, "step": 40570, "torque_loss": 0.1086977943778038 }, { "epoch": 36.492805755395686, "grad_norm": 0.049731869250535965, "learning_rate": 9.396145280470025e-06, "loss": 0.0174, "step": 40580 }, { "action_loss": 0.0012135821161791682, "epoch": 36.492805755395686, "step": 40580 }, { "epoch": 36.492805755395686, "step": 40580, "torque_loss": 0.14752954244613647 }, { "epoch": 36.50179856115108, "grad_norm": 0.08847640454769135, "learning_rate": 9.376856510996445e-06, "loss": 0.02, "step": 40590 }, { "action_loss": 0.001268157851882279, "epoch": 36.50179856115108, "step": 40590 }, { "epoch": 36.50179856115108, "step": 40590, "torque_loss": 0.1190325990319252 }, { "epoch": 36.510791366906474, "grad_norm": 0.04350990802049637, "learning_rate": 9.357585511465955e-06, "loss": 0.0151, "step": 40600 }, { "action_loss": 0.0016322076553478837, "epoch": 36.510791366906474, "step": 40600 }, { "epoch": 36.510791366906474, "step": 40600, "torque_loss": 0.2206132858991623 }, { "epoch": 36.51978417266187, "grad_norm": 0.07641981542110443, "learning_rate": 9.338332290308367e-06, "loss": 0.0172, "step": 40610 }, { "action_loss": 0.002882507862523198, "epoch": 36.51978417266187, "step": 40610 }, { "epoch": 36.51978417266187, "step": 40610, "torque_loss": 0.1344791203737259 }, { "epoch": 36.52877697841727, "grad_norm": 0.05562519654631615, "learning_rate": 9.319096855945675e-06, "loss": 0.0183, "step": 40620 }, { "action_loss": 0.0014981994172558188, "epoch": 36.52877697841727, "step": 40620 }, { "epoch": 36.52877697841727, "step": 40620, "torque_loss": 0.1092572808265686 }, { "epoch": 36.53776978417266, "grad_norm": 0.05495310202240944, "learning_rate": 9.299879216792118e-06, "loss": 0.0193, "step": 40630 }, { "action_loss": 0.0011133478255942464, "epoch": 36.53776978417266, "step": 40630 }, { "epoch": 36.53776978417266, "step": 40630, "torque_loss": 0.15398262441158295 }, { "epoch": 36.54676258992806, "grad_norm": 0.05816878750920296, "learning_rate": 9.280679381254132e-06, "loss": 0.0168, "step": 40640 }, { "action_loss": 0.0014378043124452233, "epoch": 36.54676258992806, "step": 40640 }, { "epoch": 36.54676258992806, "step": 40640, "torque_loss": 0.11539391428232193 }, { "epoch": 36.555755395683455, "grad_norm": 0.08362750709056854, "learning_rate": 9.261497357730397e-06, "loss": 0.0191, "step": 40650 }, { "action_loss": 0.001474191783927381, "epoch": 36.555755395683455, "step": 40650 }, { "epoch": 36.555755395683455, "step": 40650, "torque_loss": 0.15562665462493896 }, { "epoch": 36.564748201438846, "grad_norm": 0.10648524761199951, "learning_rate": 9.242333154611777e-06, "loss": 0.0179, "step": 40660 }, { "action_loss": 0.008594222366809845, "epoch": 36.564748201438846, "step": 40660 }, { "epoch": 36.564748201438846, "step": 40660, "torque_loss": 0.169036865234375 }, { "epoch": 36.57374100719424, "grad_norm": 0.10800755769014359, "learning_rate": 9.223186780281323e-06, "loss": 0.0212, "step": 40670 }, { "action_loss": 0.0021760875824838877, "epoch": 36.57374100719424, "step": 40670 }, { "epoch": 36.57374100719424, "step": 40670, "torque_loss": 0.17178048193454742 }, { "epoch": 36.58273381294964, "grad_norm": 0.12814471125602722, "learning_rate": 9.20405824311435e-06, "loss": 0.0251, "step": 40680 }, { "action_loss": 0.0009761604014784098, "epoch": 36.58273381294964, "step": 40680 }, { "epoch": 36.58273381294964, "step": 40680, "torque_loss": 0.1259402483701706 }, { "epoch": 36.59172661870504, "grad_norm": 0.09460262954235077, "learning_rate": 9.184947551478301e-06, "loss": 0.0179, "step": 40690 }, { "action_loss": 0.00040639968938194215, "epoch": 36.59172661870504, "step": 40690 }, { "epoch": 36.59172661870504, "step": 40690, "torque_loss": 0.09348567575216293 }, { "epoch": 36.60071942446043, "grad_norm": 0.08909200131893158, "learning_rate": 9.165854713732835e-06, "loss": 0.018, "step": 40700 }, { "action_loss": 0.004309512209147215, "epoch": 36.60071942446043, "step": 40700 }, { "epoch": 36.60071942446043, "step": 40700, "torque_loss": 0.16621243953704834 }, { "epoch": 36.60971223021583, "grad_norm": 0.08464545756578445, "learning_rate": 9.146779738229838e-06, "loss": 0.0197, "step": 40710 }, { "action_loss": 0.0008640022133477032, "epoch": 36.60971223021583, "step": 40710 }, { "epoch": 36.60971223021583, "step": 40710, "torque_loss": 0.12719950079917908 }, { "epoch": 36.618705035971225, "grad_norm": 0.0490155965089798, "learning_rate": 9.12772263331334e-06, "loss": 0.0175, "step": 40720 }, { "action_loss": 0.003619137918576598, "epoch": 36.618705035971225, "step": 40720 }, { "epoch": 36.618705035971225, "step": 40720, "torque_loss": 0.1706429272890091 }, { "epoch": 36.62769784172662, "grad_norm": 0.06518664956092834, "learning_rate": 9.10868340731954e-06, "loss": 0.0161, "step": 40730 }, { "action_loss": 0.0016164440894499421, "epoch": 36.62769784172662, "step": 40730 }, { "epoch": 36.62769784172662, "step": 40730, "torque_loss": 0.17559675872325897 }, { "epoch": 36.63669064748201, "grad_norm": 0.04798539727926254, "learning_rate": 9.089662068576887e-06, "loss": 0.0185, "step": 40740 }, { "action_loss": 0.004245240241289139, "epoch": 36.63669064748201, "step": 40740 }, { "epoch": 36.63669064748201, "step": 40740, "torque_loss": 0.16073429584503174 }, { "epoch": 36.64568345323741, "grad_norm": 0.06613031774759293, "learning_rate": 9.07065862540593e-06, "loss": 0.018, "step": 40750 }, { "action_loss": 0.002742457902058959, "epoch": 36.64568345323741, "step": 40750 }, { "epoch": 36.64568345323741, "step": 40750, "torque_loss": 0.13680194318294525 }, { "epoch": 36.65467625899281, "grad_norm": 0.040189824998378754, "learning_rate": 9.051673086119422e-06, "loss": 0.0167, "step": 40760 }, { "action_loss": 0.0026670172810554504, "epoch": 36.65467625899281, "step": 40760 }, { "epoch": 36.65467625899281, "step": 40760, "torque_loss": 0.13049589097499847 }, { "epoch": 36.6636690647482, "grad_norm": 0.16413459181785583, "learning_rate": 9.0327054590223e-06, "loss": 0.0213, "step": 40770 }, { "action_loss": 0.0072829811833798885, "epoch": 36.6636690647482, "step": 40770 }, { "epoch": 36.6636690647482, "step": 40770, "torque_loss": 0.2607351243495941 }, { "epoch": 36.672661870503596, "grad_norm": 0.062062423676252365, "learning_rate": 9.01375575241164e-06, "loss": 0.0173, "step": 40780 }, { "action_loss": 0.006691840942949057, "epoch": 36.672661870503596, "step": 40780 }, { "epoch": 36.672661870503596, "step": 40780, "torque_loss": 0.16132615506649017 }, { "epoch": 36.681654676258994, "grad_norm": 0.05685938149690628, "learning_rate": 8.994823974576672e-06, "loss": 0.0161, "step": 40790 }, { "action_loss": 0.0006602267385460436, "epoch": 36.681654676258994, "step": 40790 }, { "epoch": 36.681654676258994, "step": 40790, "torque_loss": 0.16081617772579193 }, { "epoch": 36.69064748201439, "grad_norm": 0.051564235240221024, "learning_rate": 8.975910133798816e-06, "loss": 0.0171, "step": 40800 }, { "action_loss": 0.0009801125852391124, "epoch": 36.69064748201439, "step": 40800 }, { "epoch": 36.69064748201439, "step": 40800, "torque_loss": 0.12072873115539551 }, { "epoch": 36.69964028776978, "grad_norm": 0.07492532581090927, "learning_rate": 8.957014238351635e-06, "loss": 0.0161, "step": 40810 }, { "action_loss": 0.004106612876057625, "epoch": 36.69964028776978, "step": 40810 }, { "epoch": 36.69964028776978, "step": 40810, "torque_loss": 0.201177716255188 }, { "epoch": 36.70863309352518, "grad_norm": 0.04824964702129364, "learning_rate": 8.938136296500826e-06, "loss": 0.0235, "step": 40820 }, { "action_loss": 0.0005430878954939544, "epoch": 36.70863309352518, "step": 40820 }, { "epoch": 36.70863309352518, "step": 40820, "torque_loss": 0.10285601764917374 }, { "epoch": 36.71762589928058, "grad_norm": 0.11050429940223694, "learning_rate": 8.919276316504227e-06, "loss": 0.0178, "step": 40830 }, { "action_loss": 0.0024871928617358208, "epoch": 36.71762589928058, "step": 40830 }, { "epoch": 36.71762589928058, "step": 40830, "torque_loss": 0.17047341167926788 }, { "epoch": 36.726618705035975, "grad_norm": 0.11209029704332352, "learning_rate": 8.900434306611866e-06, "loss": 0.016, "step": 40840 }, { "action_loss": 0.001960193505510688, "epoch": 36.726618705035975, "step": 40840 }, { "epoch": 36.726618705035975, "step": 40840, "torque_loss": 0.18515749275684357 }, { "epoch": 36.735611510791365, "grad_norm": 0.07373429089784622, "learning_rate": 8.881610275065861e-06, "loss": 0.0203, "step": 40850 }, { "action_loss": 0.0011853569885715842, "epoch": 36.735611510791365, "step": 40850 }, { "epoch": 36.735611510791365, "step": 40850, "torque_loss": 0.13337965309619904 }, { "epoch": 36.74460431654676, "grad_norm": 0.041803885251283646, "learning_rate": 8.862804230100463e-06, "loss": 0.0175, "step": 40860 }, { "action_loss": 0.0008431951864622533, "epoch": 36.74460431654676, "step": 40860 }, { "epoch": 36.74460431654676, "step": 40860, "torque_loss": 0.20533175766468048 }, { "epoch": 36.75359712230216, "grad_norm": 0.06081048771739006, "learning_rate": 8.844016179942116e-06, "loss": 0.0206, "step": 40870 }, { "action_loss": 0.0035609055776149035, "epoch": 36.75359712230216, "step": 40870 }, { "epoch": 36.75359712230216, "step": 40870, "torque_loss": 0.16033880412578583 }, { "epoch": 36.76258992805755, "grad_norm": 0.05571701377630234, "learning_rate": 8.825246132809329e-06, "loss": 0.0196, "step": 40880 }, { "action_loss": 0.0032510722521692514, "epoch": 36.76258992805755, "step": 40880 }, { "epoch": 36.76258992805755, "step": 40880, "torque_loss": 0.1256343275308609 }, { "epoch": 36.77158273381295, "grad_norm": 0.0487775094807148, "learning_rate": 8.80649409691275e-06, "loss": 0.0143, "step": 40890 }, { "action_loss": 0.0009038467542268336, "epoch": 36.77158273381295, "step": 40890 }, { "epoch": 36.77158273381295, "step": 40890, "torque_loss": 0.13035134971141815 }, { "epoch": 36.780575539568346, "grad_norm": 0.05197848379611969, "learning_rate": 8.78776008045517e-06, "loss": 0.0187, "step": 40900 }, { "action_loss": 0.0020594836678355932, "epoch": 36.780575539568346, "step": 40900 }, { "epoch": 36.780575539568346, "step": 40900, "torque_loss": 0.18672089278697968 }, { "epoch": 36.789568345323744, "grad_norm": 0.04880190268158913, "learning_rate": 8.76904409163149e-06, "loss": 0.0184, "step": 40910 }, { "action_loss": 0.00047212120261974633, "epoch": 36.789568345323744, "step": 40910 }, { "epoch": 36.789568345323744, "step": 40910, "torque_loss": 0.1398942768573761 }, { "epoch": 36.798561151079134, "grad_norm": 0.03651731461286545, "learning_rate": 8.750346138628696e-06, "loss": 0.0168, "step": 40920 }, { "action_loss": 0.0008405509288422763, "epoch": 36.798561151079134, "step": 40920 }, { "epoch": 36.798561151079134, "step": 40920, "torque_loss": 0.1821574717760086 }, { "epoch": 36.80755395683453, "grad_norm": 0.06353026628494263, "learning_rate": 8.73166622962593e-06, "loss": 0.0191, "step": 40930 }, { "action_loss": 0.0007066301186569035, "epoch": 36.80755395683453, "step": 40930 }, { "epoch": 36.80755395683453, "step": 40930, "torque_loss": 0.09452412277460098 }, { "epoch": 36.81654676258993, "grad_norm": 0.048573028296232224, "learning_rate": 8.713004372794415e-06, "loss": 0.0185, "step": 40940 }, { "action_loss": 0.0010891767451539636, "epoch": 36.81654676258993, "step": 40940 }, { "epoch": 36.81654676258993, "step": 40940, "torque_loss": 0.15600235760211945 }, { "epoch": 36.82553956834532, "grad_norm": 0.052911948412656784, "learning_rate": 8.694360576297461e-06, "loss": 0.0172, "step": 40950 }, { "action_loss": 0.0012243466917425394, "epoch": 36.82553956834532, "step": 40950 }, { "epoch": 36.82553956834532, "step": 40950, "torque_loss": 0.18533281981945038 }, { "epoch": 36.83453237410072, "grad_norm": 0.05702388286590576, "learning_rate": 8.675734848290529e-06, "loss": 0.0176, "step": 40960 }, { "action_loss": 0.0017843203386291862, "epoch": 36.83453237410072, "step": 40960 }, { "epoch": 36.83453237410072, "step": 40960, "torque_loss": 0.1838752180337906 }, { "epoch": 36.843525179856115, "grad_norm": 0.05132812634110451, "learning_rate": 8.657127196921133e-06, "loss": 0.0179, "step": 40970 }, { "action_loss": 0.0006225207471288741, "epoch": 36.843525179856115, "step": 40970 }, { "epoch": 36.843525179856115, "step": 40970, "torque_loss": 0.11212480068206787 }, { "epoch": 36.85251798561151, "grad_norm": 0.06655820459127426, "learning_rate": 8.638537630328885e-06, "loss": 0.0178, "step": 40980 }, { "action_loss": 0.0011566512985154986, "epoch": 36.85251798561151, "step": 40980 }, { "epoch": 36.85251798561151, "step": 40980, "torque_loss": 0.15187151730060577 }, { "epoch": 36.861510791366904, "grad_norm": 0.058029137551784515, "learning_rate": 8.619966156645487e-06, "loss": 0.0143, "step": 40990 }, { "action_loss": 0.0012775881914421916, "epoch": 36.861510791366904, "step": 40990 }, { "epoch": 36.861510791366904, "step": 40990, "torque_loss": 0.19591902196407318 }, { "epoch": 36.8705035971223, "grad_norm": 0.04310765117406845, "learning_rate": 8.601412783994756e-06, "loss": 0.017, "step": 41000 }, { "action_loss": 0.0006743862177245319, "epoch": 36.8705035971223, "step": 41000 }, { "epoch": 36.8705035971223, "step": 41000, "torque_loss": 0.12459840625524521 }, { "epoch": 36.8794964028777, "grad_norm": 0.0498870387673378, "learning_rate": 8.582877520492538e-06, "loss": 0.0252, "step": 41010 }, { "action_loss": 0.00143511313945055, "epoch": 36.8794964028777, "step": 41010 }, { "epoch": 36.8794964028777, "step": 41010, "torque_loss": 0.14242589473724365 }, { "epoch": 36.888489208633096, "grad_norm": 0.04407629743218422, "learning_rate": 8.564360374246821e-06, "loss": 0.0173, "step": 41020 }, { "action_loss": 0.00456085754558444, "epoch": 36.888489208633096, "step": 41020 }, { "epoch": 36.888489208633096, "step": 41020, "torque_loss": 0.22840003669261932 }, { "epoch": 36.89748201438849, "grad_norm": 0.04850015044212341, "learning_rate": 8.545861353357597e-06, "loss": 0.0175, "step": 41030 }, { "action_loss": 0.007934115827083588, "epoch": 36.89748201438849, "step": 41030 }, { "epoch": 36.89748201438849, "step": 41030, "torque_loss": 0.1951369196176529 }, { "epoch": 36.906474820143885, "grad_norm": 0.061977602541446686, "learning_rate": 8.527380465917001e-06, "loss": 0.0182, "step": 41040 }, { "action_loss": 0.0019425712525844574, "epoch": 36.906474820143885, "step": 41040 }, { "epoch": 36.906474820143885, "step": 41040, "torque_loss": 0.15448513627052307 }, { "epoch": 36.91546762589928, "grad_norm": 0.08293847739696503, "learning_rate": 8.508917720009175e-06, "loss": 0.0199, "step": 41050 }, { "action_loss": 0.0013245722511783242, "epoch": 36.91546762589928, "step": 41050 }, { "epoch": 36.91546762589928, "step": 41050, "torque_loss": 0.14431869983673096 }, { "epoch": 36.92446043165468, "grad_norm": 0.06896130740642548, "learning_rate": 8.490473123710374e-06, "loss": 0.0163, "step": 41060 }, { "action_loss": 0.0006410995847545564, "epoch": 36.92446043165468, "step": 41060 }, { "epoch": 36.92446043165468, "step": 41060, "torque_loss": 0.16589771211147308 }, { "epoch": 36.93345323741007, "grad_norm": 0.0626501813530922, "learning_rate": 8.472046685088874e-06, "loss": 0.0217, "step": 41070 }, { "action_loss": 0.0023082615807652473, "epoch": 36.93345323741007, "step": 41070 }, { "epoch": 36.93345323741007, "step": 41070, "torque_loss": 0.15949426591396332 }, { "epoch": 36.94244604316547, "grad_norm": 0.05052231252193451, "learning_rate": 8.453638412205023e-06, "loss": 0.0182, "step": 41080 }, { "action_loss": 0.0062398226000368595, "epoch": 36.94244604316547, "step": 41080 }, { "epoch": 36.94244604316547, "step": 41080, "torque_loss": 0.18685674667358398 }, { "epoch": 36.951438848920866, "grad_norm": 0.03512560576200485, "learning_rate": 8.435248313111243e-06, "loss": 0.0161, "step": 41090 }, { "action_loss": 0.0019252719357609749, "epoch": 36.951438848920866, "step": 41090 }, { "epoch": 36.951438848920866, "step": 41090, "torque_loss": 0.13759030401706696 }, { "epoch": 36.960431654676256, "grad_norm": 0.057365916669368744, "learning_rate": 8.416876395851974e-06, "loss": 0.0205, "step": 41100 }, { "action_loss": 0.0009348850580863655, "epoch": 36.960431654676256, "step": 41100 }, { "epoch": 36.960431654676256, "step": 41100, "torque_loss": 0.08750862628221512 }, { "epoch": 36.969424460431654, "grad_norm": 0.050619762390851974, "learning_rate": 8.39852266846371e-06, "loss": 0.0188, "step": 41110 }, { "action_loss": 0.013197305612266064, "epoch": 36.969424460431654, "step": 41110 }, { "epoch": 36.969424460431654, "step": 41110, "torque_loss": 0.19185693562030792 }, { "epoch": 36.97841726618705, "grad_norm": 0.05303014814853668, "learning_rate": 8.380187138975021e-06, "loss": 0.0222, "step": 41120 }, { "action_loss": 0.004722967278212309, "epoch": 36.97841726618705, "step": 41120 }, { "epoch": 36.97841726618705, "step": 41120, "torque_loss": 0.2862480580806732 }, { "epoch": 36.98741007194245, "grad_norm": 0.057424481958150864, "learning_rate": 8.36186981540647e-06, "loss": 0.0188, "step": 41130 }, { "action_loss": 0.000874580058734864, "epoch": 36.98741007194245, "step": 41130 }, { "epoch": 36.98741007194245, "step": 41130, "torque_loss": 0.14175288379192352 }, { "epoch": 36.99640287769784, "grad_norm": 0.1802574098110199, "learning_rate": 8.343570705770681e-06, "loss": 0.0219, "step": 41140 }, { "action_loss": 0.0015472974628210068, "epoch": 36.99640287769784, "step": 41140 }, { "epoch": 36.99640287769784, "step": 41140, "torque_loss": 0.23626548051834106 }, { "epoch": 37.00539568345324, "grad_norm": 0.06913451850414276, "learning_rate": 8.325289818072296e-06, "loss": 0.0183, "step": 41150 }, { "action_loss": 0.0016083167865872383, "epoch": 37.00539568345324, "step": 41150 }, { "epoch": 37.00539568345324, "step": 41150, "torque_loss": 0.17762720584869385 }, { "epoch": 37.014388489208635, "grad_norm": 0.04643669351935387, "learning_rate": 8.307027160308018e-06, "loss": 0.0205, "step": 41160 }, { "action_loss": 0.002363982144743204, "epoch": 37.014388489208635, "step": 41160 }, { "epoch": 37.014388489208635, "step": 41160, "torque_loss": 0.15592242777347565 }, { "epoch": 37.023381294964025, "grad_norm": 0.05144321173429489, "learning_rate": 8.288782740466538e-06, "loss": 0.0178, "step": 41170 }, { "action_loss": 0.0023600950371474028, "epoch": 37.023381294964025, "step": 41170 }, { "epoch": 37.023381294964025, "step": 41170, "torque_loss": 0.19270408153533936 }, { "epoch": 37.03237410071942, "grad_norm": 0.06225834786891937, "learning_rate": 8.270556566528581e-06, "loss": 0.017, "step": 41180 }, { "action_loss": 0.001442202366888523, "epoch": 37.03237410071942, "step": 41180 }, { "epoch": 37.03237410071942, "step": 41180, "torque_loss": 0.1627493053674698 }, { "epoch": 37.04136690647482, "grad_norm": 0.06636302173137665, "learning_rate": 8.252348646466918e-06, "loss": 0.0186, "step": 41190 }, { "action_loss": 0.0032663068268448114, "epoch": 37.04136690647482, "step": 41190 }, { "epoch": 37.04136690647482, "step": 41190, "torque_loss": 0.17182032763957977 }, { "epoch": 37.05035971223022, "grad_norm": 0.054050691425800323, "learning_rate": 8.234158988246294e-06, "loss": 0.0203, "step": 41200 }, { "action_loss": 0.001572919194586575, "epoch": 37.05035971223022, "step": 41200 }, { "epoch": 37.05035971223022, "step": 41200, "torque_loss": 0.12908340990543365 }, { "epoch": 37.05935251798561, "grad_norm": 0.08131720125675201, "learning_rate": 8.215987599823476e-06, "loss": 0.0216, "step": 41210 }, { "action_loss": 0.0011701307957991958, "epoch": 37.05935251798561, "step": 41210 }, { "epoch": 37.05935251798561, "step": 41210, "torque_loss": 0.14059440791606903 }, { "epoch": 37.068345323741006, "grad_norm": 0.13811594247817993, "learning_rate": 8.19783448914727e-06, "loss": 0.0178, "step": 41220 }, { "action_loss": 0.001530173816718161, "epoch": 37.068345323741006, "step": 41220 }, { "epoch": 37.068345323741006, "step": 41220, "torque_loss": 0.1842443197965622 }, { "epoch": 37.077338129496404, "grad_norm": 0.055566102266311646, "learning_rate": 8.179699664158452e-06, "loss": 0.018, "step": 41230 }, { "action_loss": 0.0010501811048015952, "epoch": 37.077338129496404, "step": 41230 }, { "epoch": 37.077338129496404, "step": 41230, "torque_loss": 0.17498497664928436 }, { "epoch": 37.0863309352518, "grad_norm": 0.1110939309000969, "learning_rate": 8.16158313278979e-06, "loss": 0.0169, "step": 41240 }, { "action_loss": 0.002007025294005871, "epoch": 37.0863309352518, "step": 41240 }, { "epoch": 37.0863309352518, "step": 41240, "torque_loss": 0.1303231120109558 }, { "epoch": 37.09532374100719, "grad_norm": 0.0634908527135849, "learning_rate": 8.14348490296611e-06, "loss": 0.0155, "step": 41250 }, { "action_loss": 0.0007906975224614143, "epoch": 37.09532374100719, "step": 41250 }, { "epoch": 37.09532374100719, "step": 41250, "torque_loss": 0.1488173007965088 }, { "epoch": 37.10431654676259, "grad_norm": 0.05402609705924988, "learning_rate": 8.125404982604156e-06, "loss": 0.0179, "step": 41260 }, { "action_loss": 0.0005208333022892475, "epoch": 37.10431654676259, "step": 41260 }, { "epoch": 37.10431654676259, "step": 41260, "torque_loss": 0.10944489389657974 }, { "epoch": 37.11330935251799, "grad_norm": 0.08556127548217773, "learning_rate": 8.107343379612725e-06, "loss": 0.0172, "step": 41270 }, { "action_loss": 0.0010815717978402972, "epoch": 37.11330935251799, "step": 41270 }, { "epoch": 37.11330935251799, "step": 41270, "torque_loss": 0.11538573354482651 }, { "epoch": 37.12230215827338, "grad_norm": 0.05253635346889496, "learning_rate": 8.08930010189256e-06, "loss": 0.016, "step": 41280 }, { "action_loss": 0.0012614879524335265, "epoch": 37.12230215827338, "step": 41280 }, { "epoch": 37.12230215827338, "step": 41280, "torque_loss": 0.13480131328105927 }, { "epoch": 37.131294964028775, "grad_norm": 0.08075811713933945, "learning_rate": 8.071275157336423e-06, "loss": 0.0168, "step": 41290 }, { "action_loss": 0.0006448647473007441, "epoch": 37.131294964028775, "step": 41290 }, { "epoch": 37.131294964028775, "step": 41290, "torque_loss": 0.1651608794927597 }, { "epoch": 37.14028776978417, "grad_norm": 0.05424968898296356, "learning_rate": 8.053268553829018e-06, "loss": 0.0192, "step": 41300 }, { "action_loss": 0.0025377923157066107, "epoch": 37.14028776978417, "step": 41300 }, { "epoch": 37.14028776978417, "step": 41300, "torque_loss": 0.1651812046766281 }, { "epoch": 37.14928057553957, "grad_norm": 0.061135999858379364, "learning_rate": 8.03528029924704e-06, "loss": 0.0186, "step": 41310 }, { "action_loss": 0.0016526104882359505, "epoch": 37.14928057553957, "step": 41310 }, { "epoch": 37.14928057553957, "step": 41310, "torque_loss": 0.14905968308448792 }, { "epoch": 37.15827338129496, "grad_norm": 0.0796515941619873, "learning_rate": 8.017310401459188e-06, "loss": 0.0211, "step": 41320 }, { "action_loss": 0.0014460167149081826, "epoch": 37.15827338129496, "step": 41320 }, { "epoch": 37.15827338129496, "step": 41320, "torque_loss": 0.11195822805166245 }, { "epoch": 37.16726618705036, "grad_norm": 0.059572137892246246, "learning_rate": 7.999358868326095e-06, "loss": 0.0186, "step": 41330 }, { "action_loss": 0.004018616396933794, "epoch": 37.16726618705036, "step": 41330 }, { "epoch": 37.16726618705036, "step": 41330, "torque_loss": 0.14455486834049225 }, { "epoch": 37.17625899280576, "grad_norm": 0.10190143436193466, "learning_rate": 7.981425707700352e-06, "loss": 0.0167, "step": 41340 }, { "action_loss": 0.0011913361959159374, "epoch": 37.17625899280576, "step": 41340 }, { "epoch": 37.17625899280576, "step": 41340, "torque_loss": 0.1278386116027832 }, { "epoch": 37.185251798561154, "grad_norm": 0.11824408918619156, "learning_rate": 7.96351092742656e-06, "loss": 0.0179, "step": 41350 }, { "action_loss": 0.004384189378470182, "epoch": 37.185251798561154, "step": 41350 }, { "epoch": 37.185251798561154, "step": 41350, "torque_loss": 0.16131387650966644 }, { "epoch": 37.194244604316545, "grad_norm": 0.05246562510728836, "learning_rate": 7.945614535341246e-06, "loss": 0.0181, "step": 41360 }, { "action_loss": 0.0007997037027962506, "epoch": 37.194244604316545, "step": 41360 }, { "epoch": 37.194244604316545, "step": 41360, "torque_loss": 0.13940943777561188 }, { "epoch": 37.20323741007194, "grad_norm": 0.07435371726751328, "learning_rate": 7.92773653927288e-06, "loss": 0.0173, "step": 41370 }, { "action_loss": 0.0012962570181116462, "epoch": 37.20323741007194, "step": 41370 }, { "epoch": 37.20323741007194, "step": 41370, "torque_loss": 0.19201534986495972 }, { "epoch": 37.21223021582734, "grad_norm": 0.04030809924006462, "learning_rate": 7.909876947041934e-06, "loss": 0.0187, "step": 41380 }, { "action_loss": 0.0028442901093512774, "epoch": 37.21223021582734, "step": 41380 }, { "epoch": 37.21223021582734, "step": 41380, "torque_loss": 0.1818339079618454 }, { "epoch": 37.22122302158273, "grad_norm": 0.07267946749925613, "learning_rate": 7.892035766460782e-06, "loss": 0.0176, "step": 41390 }, { "action_loss": 0.0008776795584708452, "epoch": 37.22122302158273, "step": 41390 }, { "epoch": 37.22122302158273, "step": 41390, "torque_loss": 0.11842313408851624 }, { "epoch": 37.23021582733813, "grad_norm": 0.052413392812013626, "learning_rate": 7.874213005333764e-06, "loss": 0.0181, "step": 41400 }, { "action_loss": 0.001457682461477816, "epoch": 37.23021582733813, "step": 41400 }, { "epoch": 37.23021582733813, "step": 41400, "torque_loss": 0.11596717685461044 }, { "epoch": 37.239208633093526, "grad_norm": 0.04600180312991142, "learning_rate": 7.856408671457166e-06, "loss": 0.0174, "step": 41410 }, { "action_loss": 0.0022860760800540447, "epoch": 37.239208633093526, "step": 41410 }, { "epoch": 37.239208633093526, "step": 41410, "torque_loss": 0.16262368857860565 }, { "epoch": 37.24820143884892, "grad_norm": 0.05367524176836014, "learning_rate": 7.838622772619214e-06, "loss": 0.0176, "step": 41420 }, { "action_loss": 0.001030491548590362, "epoch": 37.24820143884892, "step": 41420 }, { "epoch": 37.24820143884892, "step": 41420, "torque_loss": 0.14536641538143158 }, { "epoch": 37.257194244604314, "grad_norm": 0.04205944016575813, "learning_rate": 7.820855316600045e-06, "loss": 0.0196, "step": 41430 }, { "action_loss": 0.0017862984677776694, "epoch": 37.257194244604314, "step": 41430 }, { "epoch": 37.257194244604314, "step": 41430, "torque_loss": 0.15263913571834564 }, { "epoch": 37.26618705035971, "grad_norm": 0.1057993695139885, "learning_rate": 7.803106311171771e-06, "loss": 0.0181, "step": 41440 }, { "action_loss": 0.0010924008674919605, "epoch": 37.26618705035971, "step": 41440 }, { "epoch": 37.26618705035971, "step": 41440, "torque_loss": 0.14417220652103424 }, { "epoch": 37.27517985611511, "grad_norm": 0.0750778317451477, "learning_rate": 7.785375764098401e-06, "loss": 0.0178, "step": 41450 }, { "action_loss": 0.008085902780294418, "epoch": 37.27517985611511, "step": 41450 }, { "epoch": 37.27517985611511, "step": 41450, "torque_loss": 0.23285458981990814 }, { "epoch": 37.28417266187051, "grad_norm": 0.07048002630472183, "learning_rate": 7.767663683135862e-06, "loss": 0.02, "step": 41460 }, { "action_loss": 0.0009067127830348909, "epoch": 37.28417266187051, "step": 41460 }, { "epoch": 37.28417266187051, "step": 41460, "torque_loss": 0.1382031887769699 }, { "epoch": 37.2931654676259, "grad_norm": 0.041492313146591187, "learning_rate": 7.749970076032049e-06, "loss": 0.0155, "step": 41470 }, { "action_loss": 0.0040754214860498905, "epoch": 37.2931654676259, "step": 41470 }, { "epoch": 37.2931654676259, "step": 41470, "torque_loss": 0.1857265681028366 }, { "epoch": 37.302158273381295, "grad_norm": 0.04225022718310356, "learning_rate": 7.73229495052673e-06, "loss": 0.0179, "step": 41480 }, { "action_loss": 0.0009064057958312333, "epoch": 37.302158273381295, "step": 41480 }, { "epoch": 37.302158273381295, "step": 41480, "torque_loss": 0.14294540882110596 }, { "epoch": 37.31115107913669, "grad_norm": 0.05750800669193268, "learning_rate": 7.714638314351597e-06, "loss": 0.017, "step": 41490 }, { "action_loss": 0.003850663313642144, "epoch": 37.31115107913669, "step": 41490 }, { "epoch": 37.31115107913669, "step": 41490, "torque_loss": 0.1507067233324051 }, { "epoch": 37.32014388489208, "grad_norm": 0.044912032783031464, "learning_rate": 7.69700017523029e-06, "loss": 0.0173, "step": 41500 }, { "action_loss": 0.0005862153484486043, "epoch": 37.32014388489208, "step": 41500 }, { "epoch": 37.32014388489208, "step": 41500, "torque_loss": 0.13069111108779907 }, { "epoch": 37.32913669064748, "grad_norm": 0.05862947553396225, "learning_rate": 7.679380540878294e-06, "loss": 0.0165, "step": 41510 }, { "action_loss": 0.002699708566069603, "epoch": 37.32913669064748, "step": 41510 }, { "epoch": 37.32913669064748, "step": 41510, "torque_loss": 0.20182685554027557 }, { "epoch": 37.33812949640288, "grad_norm": 0.051522038877010345, "learning_rate": 7.661779419003073e-06, "loss": 0.0173, "step": 41520 }, { "action_loss": 0.004238147754222155, "epoch": 37.33812949640288, "step": 41520 }, { "epoch": 37.33812949640288, "step": 41520, "torque_loss": 0.14103730022907257 }, { "epoch": 37.347122302158276, "grad_norm": 0.05010287091135979, "learning_rate": 7.644196817303928e-06, "loss": 0.0182, "step": 41530 }, { "action_loss": 0.0010781255550682545, "epoch": 37.347122302158276, "step": 41530 }, { "epoch": 37.347122302158276, "step": 41530, "torque_loss": 0.16469156742095947 }, { "epoch": 37.356115107913666, "grad_norm": 0.11898995190858841, "learning_rate": 7.626632743472106e-06, "loss": 0.0192, "step": 41540 }, { "action_loss": 0.0010494141606613994, "epoch": 37.356115107913666, "step": 41540 }, { "epoch": 37.356115107913666, "step": 41540, "torque_loss": 0.12180707603693008 }, { "epoch": 37.365107913669064, "grad_norm": 0.11914090812206268, "learning_rate": 7.609087205190724e-06, "loss": 0.0213, "step": 41550 }, { "action_loss": 0.002760716015473008, "epoch": 37.365107913669064, "step": 41550 }, { "epoch": 37.365107913669064, "step": 41550, "torque_loss": 0.18612785637378693 }, { "epoch": 37.37410071942446, "grad_norm": 0.056344568729400635, "learning_rate": 7.591560210134774e-06, "loss": 0.0201, "step": 41560 }, { "action_loss": 0.012709264643490314, "epoch": 37.37410071942446, "step": 41560 }, { "epoch": 37.37410071942446, "step": 41560, "torque_loss": 0.24489283561706543 }, { "epoch": 37.38309352517986, "grad_norm": 0.10056792944669724, "learning_rate": 7.574051765971196e-06, "loss": 0.0204, "step": 41570 }, { "action_loss": 0.0017751418054103851, "epoch": 37.38309352517986, "step": 41570 }, { "epoch": 37.38309352517986, "step": 41570, "torque_loss": 0.13663645088672638 }, { "epoch": 37.39208633093525, "grad_norm": 0.12523074448108673, "learning_rate": 7.556561880358754e-06, "loss": 0.02, "step": 41580 }, { "action_loss": 0.009172559715807438, "epoch": 37.39208633093525, "step": 41580 }, { "epoch": 37.39208633093525, "step": 41580, "torque_loss": 0.2053375244140625 }, { "epoch": 37.40107913669065, "grad_norm": 0.12866505980491638, "learning_rate": 7.5390905609481196e-06, "loss": 0.0203, "step": 41590 }, { "action_loss": 0.0008790221181698143, "epoch": 37.40107913669065, "step": 41590 }, { "epoch": 37.40107913669065, "step": 41590, "torque_loss": 0.16331914067268372 }, { "epoch": 37.410071942446045, "grad_norm": 0.055299147963523865, "learning_rate": 7.521637815381855e-06, "loss": 0.0166, "step": 41600 }, { "action_loss": 0.0009659889037720859, "epoch": 37.410071942446045, "step": 41600 }, { "epoch": 37.410071942446045, "step": 41600, "torque_loss": 0.16806791722774506 }, { "epoch": 37.419064748201436, "grad_norm": 0.05165699124336243, "learning_rate": 7.504203651294373e-06, "loss": 0.0156, "step": 41610 }, { "action_loss": 0.0028678912203758955, "epoch": 37.419064748201436, "step": 41610 }, { "epoch": 37.419064748201436, "step": 41610, "torque_loss": 0.2364652156829834 }, { "epoch": 37.42805755395683, "grad_norm": 0.05518099665641785, "learning_rate": 7.486788076311963e-06, "loss": 0.0191, "step": 41620 }, { "action_loss": 0.0026763516943901777, "epoch": 37.42805755395683, "step": 41620 }, { "epoch": 37.42805755395683, "step": 41620, "torque_loss": 0.1896640509366989 }, { "epoch": 37.43705035971223, "grad_norm": 0.043188173323869705, "learning_rate": 7.469391098052808e-06, "loss": 0.0172, "step": 41630 }, { "action_loss": 0.002962797647342086, "epoch": 37.43705035971223, "step": 41630 }, { "epoch": 37.43705035971223, "step": 41630, "torque_loss": 0.17432059347629547 }, { "epoch": 37.44604316546763, "grad_norm": 0.0535515658557415, "learning_rate": 7.452012724126928e-06, "loss": 0.0187, "step": 41640 }, { "action_loss": 0.0005644952179864049, "epoch": 37.44604316546763, "step": 41640 }, { "epoch": 37.44604316546763, "step": 41640, "torque_loss": 0.12827670574188232 }, { "epoch": 37.45503597122302, "grad_norm": 0.08655522763729095, "learning_rate": 7.434652962136213e-06, "loss": 0.017, "step": 41650 }, { "action_loss": 0.0009734313935041428, "epoch": 37.45503597122302, "step": 41650 }, { "epoch": 37.45503597122302, "step": 41650, "torque_loss": 0.20131780207157135 }, { "epoch": 37.46402877697842, "grad_norm": 0.04782893881201744, "learning_rate": 7.4173118196744e-06, "loss": 0.0175, "step": 41660 }, { "action_loss": 0.0008208765648305416, "epoch": 37.46402877697842, "step": 41660 }, { "epoch": 37.46402877697842, "step": 41660, "torque_loss": 0.1391148865222931 }, { "epoch": 37.473021582733814, "grad_norm": 0.04602248966693878, "learning_rate": 7.39998930432712e-06, "loss": 0.0169, "step": 41670 }, { "action_loss": 0.0028204030822962523, "epoch": 37.473021582733814, "step": 41670 }, { "epoch": 37.473021582733814, "step": 41670, "torque_loss": 0.1425851434469223 }, { "epoch": 37.48201438848921, "grad_norm": 0.08300581574440002, "learning_rate": 7.38268542367182e-06, "loss": 0.0202, "step": 41680 }, { "action_loss": 0.005045368801802397, "epoch": 37.48201438848921, "step": 41680 }, { "epoch": 37.48201438848921, "step": 41680, "torque_loss": 0.19815297424793243 }, { "epoch": 37.4910071942446, "grad_norm": 0.052948638796806335, "learning_rate": 7.365400185277782e-06, "loss": 0.0204, "step": 41690 }, { "action_loss": 0.0016523479716852307, "epoch": 37.4910071942446, "step": 41690 }, { "epoch": 37.4910071942446, "step": 41690, "torque_loss": 0.10986437648534775 }, { "epoch": 37.5, "grad_norm": 0.055919747799634933, "learning_rate": 7.348133596706197e-06, "loss": 0.0166, "step": 41700 }, { "action_loss": 0.00415696669369936, "epoch": 37.5, "step": 41700 }, { "epoch": 37.5, "step": 41700, "torque_loss": 0.162169948220253 }, { "epoch": 37.5089928057554, "grad_norm": 0.08049921691417694, "learning_rate": 7.3308856655100286e-06, "loss": 0.0184, "step": 41710 }, { "action_loss": 0.0006040320149622858, "epoch": 37.5089928057554, "step": 41710 }, { "epoch": 37.5089928057554, "step": 41710, "torque_loss": 0.10543444752693176 }, { "epoch": 37.51798561151079, "grad_norm": 0.036027178168296814, "learning_rate": 7.313656399234109e-06, "loss": 0.0143, "step": 41720 }, { "action_loss": 0.0007058786577545106, "epoch": 37.51798561151079, "step": 41720 }, { "epoch": 37.51798561151079, "step": 41720, "torque_loss": 0.1694130301475525 }, { "epoch": 37.526978417266186, "grad_norm": 0.05010506883263588, "learning_rate": 7.296445805415125e-06, "loss": 0.0174, "step": 41730 }, { "action_loss": 0.0008469028398394585, "epoch": 37.526978417266186, "step": 41730 }, { "epoch": 37.526978417266186, "step": 41730, "torque_loss": 0.13741548359394073 }, { "epoch": 37.53597122302158, "grad_norm": 0.06304850429296494, "learning_rate": 7.2792538915815504e-06, "loss": 0.0163, "step": 41740 }, { "action_loss": 0.0008351723081432283, "epoch": 37.53597122302158, "step": 41740 }, { "epoch": 37.53597122302158, "step": 41740, "torque_loss": 0.13151462376117706 }, { "epoch": 37.54496402877698, "grad_norm": 0.047735556960105896, "learning_rate": 7.262080665253729e-06, "loss": 0.0187, "step": 41750 }, { "action_loss": 0.0031507767271250486, "epoch": 37.54496402877698, "step": 41750 }, { "epoch": 37.54496402877698, "step": 41750, "torque_loss": 0.18892236053943634 }, { "epoch": 37.55395683453237, "grad_norm": 0.12919078767299652, "learning_rate": 7.244926133943803e-06, "loss": 0.0191, "step": 41760 }, { "action_loss": 0.0006205799872986972, "epoch": 37.55395683453237, "step": 41760 }, { "epoch": 37.55395683453237, "step": 41760, "torque_loss": 0.13463403284549713 }, { "epoch": 37.56294964028777, "grad_norm": 0.0434875413775444, "learning_rate": 7.227790305155763e-06, "loss": 0.0162, "step": 41770 }, { "action_loss": 0.001411812729202211, "epoch": 37.56294964028777, "step": 41770 }, { "epoch": 37.56294964028777, "step": 41770, "torque_loss": 0.20453254878520966 }, { "epoch": 37.57194244604317, "grad_norm": 0.07006972283124924, "learning_rate": 7.210673186385381e-06, "loss": 0.0183, "step": 41780 }, { "action_loss": 0.0010819312883540988, "epoch": 37.57194244604317, "step": 41780 }, { "epoch": 37.57194244604317, "step": 41780, "torque_loss": 0.1996721625328064 }, { "epoch": 37.580935251798564, "grad_norm": 0.060370396822690964, "learning_rate": 7.193574785120288e-06, "loss": 0.02, "step": 41790 }, { "action_loss": 0.008038978092372417, "epoch": 37.580935251798564, "step": 41790 }, { "epoch": 37.580935251798564, "step": 41790, "torque_loss": 0.20422965288162231 }, { "epoch": 37.589928057553955, "grad_norm": 0.04975264146924019, "learning_rate": 7.176495108839893e-06, "loss": 0.017, "step": 41800 }, { "action_loss": 0.0015544475754722953, "epoch": 37.589928057553955, "step": 41800 }, { "epoch": 37.589928057553955, "step": 41800, "torque_loss": 0.11568903923034668 }, { "epoch": 37.59892086330935, "grad_norm": 0.1226324588060379, "learning_rate": 7.159434165015417e-06, "loss": 0.018, "step": 41810 }, { "action_loss": 0.0008940251427702606, "epoch": 37.59892086330935, "step": 41810 }, { "epoch": 37.59892086330935, "step": 41810, "torque_loss": 0.15518413484096527 }, { "epoch": 37.60791366906475, "grad_norm": 0.06182819977402687, "learning_rate": 7.142391961109923e-06, "loss": 0.017, "step": 41820 }, { "action_loss": 0.0016003047348931432, "epoch": 37.60791366906475, "step": 41820 }, { "epoch": 37.60791366906475, "step": 41820, "torque_loss": 0.22056174278259277 }, { "epoch": 37.61690647482014, "grad_norm": 0.049388859421014786, "learning_rate": 7.1253685045782335e-06, "loss": 0.0184, "step": 41830 }, { "action_loss": 0.001769349561072886, "epoch": 37.61690647482014, "step": 41830 }, { "epoch": 37.61690647482014, "step": 41830, "torque_loss": 0.1873747706413269 }, { "epoch": 37.62589928057554, "grad_norm": 0.14011216163635254, "learning_rate": 7.108363802866991e-06, "loss": 0.0242, "step": 41840 }, { "action_loss": 0.0011920315446332097, "epoch": 37.62589928057554, "step": 41840 }, { "epoch": 37.62589928057554, "step": 41840, "torque_loss": 0.15809385478496552 }, { "epoch": 37.634892086330936, "grad_norm": 0.1152060329914093, "learning_rate": 7.091377863414611e-06, "loss": 0.0179, "step": 41850 }, { "action_loss": 0.0012411260977387428, "epoch": 37.634892086330936, "step": 41850 }, { "epoch": 37.634892086330936, "step": 41850, "torque_loss": 0.17296600341796875 }, { "epoch": 37.643884892086334, "grad_norm": 0.03858078643679619, "learning_rate": 7.074410693651351e-06, "loss": 0.0166, "step": 41860 }, { "action_loss": 0.0007775977719575167, "epoch": 37.643884892086334, "step": 41860 }, { "epoch": 37.643884892086334, "step": 41860, "torque_loss": 0.14954949915409088 }, { "epoch": 37.652877697841724, "grad_norm": 0.07276845723390579, "learning_rate": 7.057462300999212e-06, "loss": 0.017, "step": 41870 }, { "action_loss": 0.0010307506890967488, "epoch": 37.652877697841724, "step": 41870 }, { "epoch": 37.652877697841724, "step": 41870, "torque_loss": 0.17110483348369598 }, { "epoch": 37.66187050359712, "grad_norm": 0.10008863359689713, "learning_rate": 7.0405326928719876e-06, "loss": 0.0198, "step": 41880 }, { "action_loss": 0.0034784500021487474, "epoch": 37.66187050359712, "step": 41880 }, { "epoch": 37.66187050359712, "step": 41880, "torque_loss": 0.14947405457496643 }, { "epoch": 37.67086330935252, "grad_norm": 0.044276442378759384, "learning_rate": 7.0236218766752844e-06, "loss": 0.0178, "step": 41890 }, { "action_loss": 0.0017161131836473942, "epoch": 37.67086330935252, "step": 41890 }, { "epoch": 37.67086330935252, "step": 41890, "torque_loss": 0.17656536400318146 }, { "epoch": 37.67985611510792, "grad_norm": 0.07043138891458511, "learning_rate": 7.006729859806454e-06, "loss": 0.0176, "step": 41900 }, { "action_loss": 0.0010209575993940234, "epoch": 37.67985611510792, "step": 41900 }, { "epoch": 37.67985611510792, "step": 41900, "torque_loss": 0.15676939487457275 }, { "epoch": 37.68884892086331, "grad_norm": 0.11231251806020737, "learning_rate": 6.989856649654636e-06, "loss": 0.02, "step": 41910 }, { "action_loss": 0.0007988444995135069, "epoch": 37.68884892086331, "step": 41910 }, { "epoch": 37.68884892086331, "step": 41910, "torque_loss": 0.17180980741977692 }, { "epoch": 37.697841726618705, "grad_norm": 0.06315597146749496, "learning_rate": 6.973002253600764e-06, "loss": 0.0179, "step": 41920 }, { "action_loss": 0.004526956472545862, "epoch": 37.697841726618705, "step": 41920 }, { "epoch": 37.697841726618705, "step": 41920, "torque_loss": 0.1916590929031372 }, { "epoch": 37.7068345323741, "grad_norm": 0.06368633359670639, "learning_rate": 6.956166679017506e-06, "loss": 0.0213, "step": 41930 }, { "action_loss": 0.02472725510597229, "epoch": 37.7068345323741, "step": 41930 }, { "epoch": 37.7068345323741, "step": 41930, "torque_loss": 0.1806018352508545 }, { "epoch": 37.71582733812949, "grad_norm": 0.07155721634626389, "learning_rate": 6.939349933269312e-06, "loss": 0.0187, "step": 41940 }, { "action_loss": 0.0027145566418766975, "epoch": 37.71582733812949, "step": 41940 }, { "epoch": 37.71582733812949, "step": 41940, "torque_loss": 0.2029862254858017 }, { "epoch": 37.72482014388489, "grad_norm": 0.1385761797428131, "learning_rate": 6.9225520237124146e-06, "loss": 0.0197, "step": 41950 }, { "action_loss": 0.0011153826490044594, "epoch": 37.72482014388489, "step": 41950 }, { "epoch": 37.72482014388489, "step": 41950, "torque_loss": 0.13079802691936493 }, { "epoch": 37.73381294964029, "grad_norm": 0.06343908607959747, "learning_rate": 6.905772957694784e-06, "loss": 0.0176, "step": 41960 }, { "action_loss": 0.0024309600703418255, "epoch": 37.73381294964029, "step": 41960 }, { "epoch": 37.73381294964029, "step": 41960, "torque_loss": 0.18831531703472137 }, { "epoch": 37.742805755395686, "grad_norm": 0.06354968994855881, "learning_rate": 6.889012742556133e-06, "loss": 0.0214, "step": 41970 }, { "action_loss": 0.001032356289215386, "epoch": 37.742805755395686, "step": 41970 }, { "epoch": 37.742805755395686, "step": 41970, "torque_loss": 0.1122000440955162 }, { "epoch": 37.75179856115108, "grad_norm": 0.0985625684261322, "learning_rate": 6.872271385627982e-06, "loss": 0.0207, "step": 41980 }, { "action_loss": 0.0006567759555764496, "epoch": 37.75179856115108, "step": 41980 }, { "epoch": 37.75179856115108, "step": 41980, "torque_loss": 0.10264953970909119 }, { "epoch": 37.760791366906474, "grad_norm": 0.05033744499087334, "learning_rate": 6.8555488942335435e-06, "loss": 0.0225, "step": 41990 }, { "action_loss": 0.007731765508651733, "epoch": 37.760791366906474, "step": 41990 }, { "epoch": 37.760791366906474, "step": 41990, "torque_loss": 0.168293759226799 }, { "epoch": 37.76978417266187, "grad_norm": 0.07659100741147995, "learning_rate": 6.838845275687822e-06, "loss": 0.0197, "step": 42000 }, { "action_loss": 0.0011730510741472244, "epoch": 37.76978417266187, "step": 42000 }, { "epoch": 37.76978417266187, "step": 42000, "torque_loss": 0.1481899917125702 }, { "epoch": 37.77877697841727, "grad_norm": 0.044095005840063095, "learning_rate": 6.822160537297534e-06, "loss": 0.0152, "step": 42010 }, { "action_loss": 0.0013240650296211243, "epoch": 37.77877697841727, "step": 42010 }, { "epoch": 37.77877697841727, "step": 42010, "torque_loss": 0.1405470371246338 }, { "epoch": 37.78776978417266, "grad_norm": 0.07065676152706146, "learning_rate": 6.805494686361164e-06, "loss": 0.0169, "step": 42020 }, { "action_loss": 0.0034514309372752905, "epoch": 37.78776978417266, "step": 42020 }, { "epoch": 37.78776978417266, "step": 42020, "torque_loss": 0.1594809889793396 }, { "epoch": 37.79676258992806, "grad_norm": 0.05504222214221954, "learning_rate": 6.788847730168912e-06, "loss": 0.0177, "step": 42030 }, { "action_loss": 0.0016430557006970048, "epoch": 37.79676258992806, "step": 42030 }, { "epoch": 37.79676258992806, "step": 42030, "torque_loss": 0.11183745414018631 }, { "epoch": 37.805755395683455, "grad_norm": 0.04062315449118614, "learning_rate": 6.772219676002716e-06, "loss": 0.0169, "step": 42040 }, { "action_loss": 0.02232939563691616, "epoch": 37.805755395683455, "step": 42040 }, { "epoch": 37.805755395683455, "step": 42040, "torque_loss": 0.20347900688648224 }, { "epoch": 37.814748201438846, "grad_norm": 0.03238654509186745, "learning_rate": 6.755610531136269e-06, "loss": 0.0206, "step": 42050 }, { "action_loss": 0.008734532631933689, "epoch": 37.814748201438846, "step": 42050 }, { "epoch": 37.814748201438846, "step": 42050, "torque_loss": 0.21925435960292816 }, { "epoch": 37.82374100719424, "grad_norm": 0.036656297743320465, "learning_rate": 6.73902030283497e-06, "loss": 0.0173, "step": 42060 }, { "action_loss": 0.0014255210990086198, "epoch": 37.82374100719424, "step": 42060 }, { "epoch": 37.82374100719424, "step": 42060, "torque_loss": 0.16688697040081024 }, { "epoch": 37.83273381294964, "grad_norm": 0.06974110007286072, "learning_rate": 6.7224489983559295e-06, "loss": 0.0163, "step": 42070 }, { "action_loss": 0.001996403792873025, "epoch": 37.83273381294964, "step": 42070 }, { "epoch": 37.83273381294964, "step": 42070, "torque_loss": 0.16818857192993164 }, { "epoch": 37.84172661870504, "grad_norm": 0.06523439288139343, "learning_rate": 6.705896624948027e-06, "loss": 0.017, "step": 42080 }, { "action_loss": 0.0014513606438413262, "epoch": 37.84172661870504, "step": 42080 }, { "epoch": 37.84172661870504, "step": 42080, "torque_loss": 0.18753431737422943 }, { "epoch": 37.85071942446043, "grad_norm": 0.04328520968556404, "learning_rate": 6.689363189851816e-06, "loss": 0.0188, "step": 42090 }, { "action_loss": 0.0005697804735973477, "epoch": 37.85071942446043, "step": 42090 }, { "epoch": 37.85071942446043, "step": 42090, "torque_loss": 0.12591849267482758 }, { "epoch": 37.85971223021583, "grad_norm": 0.06680206209421158, "learning_rate": 6.672848700299583e-06, "loss": 0.0178, "step": 42100 }, { "action_loss": 0.0019082174403592944, "epoch": 37.85971223021583, "step": 42100 }, { "epoch": 37.85971223021583, "step": 42100, "torque_loss": 0.1261916607618332 }, { "epoch": 37.868705035971225, "grad_norm": 0.05153803899884224, "learning_rate": 6.656353163515339e-06, "loss": 0.0175, "step": 42110 }, { "action_loss": 0.0009268067660741508, "epoch": 37.868705035971225, "step": 42110 }, { "epoch": 37.868705035971225, "step": 42110, "torque_loss": 0.16846956312656403 }, { "epoch": 37.87769784172662, "grad_norm": 0.030789121985435486, "learning_rate": 6.6398765867147824e-06, "loss": 0.0213, "step": 42120 }, { "action_loss": 0.001814066432416439, "epoch": 37.87769784172662, "step": 42120 }, { "epoch": 37.87769784172662, "step": 42120, "torque_loss": 0.15192651748657227 }, { "epoch": 37.88669064748201, "grad_norm": 0.035193756222724915, "learning_rate": 6.623418977105322e-06, "loss": 0.0161, "step": 42130 }, { "action_loss": 0.0005820958758704364, "epoch": 37.88669064748201, "step": 42130 }, { "epoch": 37.88669064748201, "step": 42130, "torque_loss": 0.08498462289571762 }, { "epoch": 37.89568345323741, "grad_norm": 0.05089793726801872, "learning_rate": 6.606980341886087e-06, "loss": 0.0175, "step": 42140 }, { "action_loss": 0.0047450740821659565, "epoch": 37.89568345323741, "step": 42140 }, { "epoch": 37.89568345323741, "step": 42140, "torque_loss": 0.1051316037774086 }, { "epoch": 37.90467625899281, "grad_norm": 0.055578265339136124, "learning_rate": 6.590560688247898e-06, "loss": 0.0222, "step": 42150 }, { "action_loss": 0.01658688485622406, "epoch": 37.90467625899281, "step": 42150 }, { "epoch": 37.90467625899281, "step": 42150, "torque_loss": 0.16796386241912842 }, { "epoch": 37.9136690647482, "grad_norm": 0.047047942876815796, "learning_rate": 6.574160023373249e-06, "loss": 0.0212, "step": 42160 }, { "action_loss": 0.0012036184780299664, "epoch": 37.9136690647482, "step": 42160 }, { "epoch": 37.9136690647482, "step": 42160, "torque_loss": 0.15432047843933105 }, { "epoch": 37.922661870503596, "grad_norm": 0.04727201908826828, "learning_rate": 6.557778354436367e-06, "loss": 0.0183, "step": 42170 }, { "action_loss": 0.0010545665863901377, "epoch": 37.922661870503596, "step": 42170 }, { "epoch": 37.922661870503596, "step": 42170, "torque_loss": 0.11122532933950424 }, { "epoch": 37.931654676258994, "grad_norm": 0.0440584197640419, "learning_rate": 6.5414156886031475e-06, "loss": 0.0165, "step": 42180 }, { "action_loss": 0.009377182461321354, "epoch": 37.931654676258994, "step": 42180 }, { "epoch": 37.931654676258994, "step": 42180, "torque_loss": 0.19254271686077118 }, { "epoch": 37.94064748201439, "grad_norm": 0.08771151304244995, "learning_rate": 6.525072033031177e-06, "loss": 0.0207, "step": 42190 }, { "action_loss": 0.0015422633150592446, "epoch": 37.94064748201439, "step": 42190 }, { "epoch": 37.94064748201439, "step": 42190, "torque_loss": 0.15007255971431732 }, { "epoch": 37.94964028776978, "grad_norm": 0.0918329730629921, "learning_rate": 6.508747394869708e-06, "loss": 0.0206, "step": 42200 }, { "action_loss": 0.0013633131748065352, "epoch": 37.94964028776978, "step": 42200 }, { "epoch": 37.94964028776978, "step": 42200, "torque_loss": 0.22320519387722015 }, { "epoch": 37.95863309352518, "grad_norm": 0.10298048704862595, "learning_rate": 6.492441781259717e-06, "loss": 0.0178, "step": 42210 }, { "action_loss": 0.0005553353112190962, "epoch": 37.95863309352518, "step": 42210 }, { "epoch": 37.95863309352518, "step": 42210, "torque_loss": 0.16677720844745636 }, { "epoch": 37.96762589928058, "grad_norm": 0.06874558329582214, "learning_rate": 6.476155199333811e-06, "loss": 0.0171, "step": 42220 }, { "action_loss": 0.00123967498075217, "epoch": 37.96762589928058, "step": 42220 }, { "epoch": 37.96762589928058, "step": 42220, "torque_loss": 0.13846834003925323 }, { "epoch": 37.976618705035975, "grad_norm": 0.037086326628923416, "learning_rate": 6.459887656216318e-06, "loss": 0.0195, "step": 42230 }, { "action_loss": 0.00042172582470811903, "epoch": 37.976618705035975, "step": 42230 }, { "epoch": 37.976618705035975, "step": 42230, "torque_loss": 0.15614429116249084 }, { "epoch": 37.985611510791365, "grad_norm": 0.04142042621970177, "learning_rate": 6.443639159023191e-06, "loss": 0.017, "step": 42240 }, { "action_loss": 0.0006715712952427566, "epoch": 37.985611510791365, "step": 42240 }, { "epoch": 37.985611510791365, "step": 42240, "torque_loss": 0.12324490398168564 }, { "epoch": 37.99460431654676, "grad_norm": 0.0473322868347168, "learning_rate": 6.427409714862092e-06, "loss": 0.0155, "step": 42250 }, { "action_loss": 0.0007219809922389686, "epoch": 37.99460431654676, "step": 42250 }, { "epoch": 37.99460431654676, "step": 42250, "torque_loss": 0.08869022130966187 }, { "epoch": 38.00359712230216, "grad_norm": 0.10457423329353333, "learning_rate": 6.411199330832318e-06, "loss": 0.0192, "step": 42260 }, { "action_loss": 0.000772673636674881, "epoch": 38.00359712230216, "step": 42260 }, { "epoch": 38.00359712230216, "step": 42260, "torque_loss": 0.12518858909606934 }, { "epoch": 38.01258992805755, "grad_norm": 0.07355096191167831, "learning_rate": 6.3950080140248556e-06, "loss": 0.02, "step": 42270 }, { "action_loss": 0.0030356713104993105, "epoch": 38.01258992805755, "step": 42270 }, { "epoch": 38.01258992805755, "step": 42270, "torque_loss": 0.2200181484222412 }, { "epoch": 38.02158273381295, "grad_norm": 0.09078072011470795, "learning_rate": 6.378835771522335e-06, "loss": 0.0174, "step": 42280 }, { "action_loss": 0.0010565532138571143, "epoch": 38.02158273381295, "step": 42280 }, { "epoch": 38.02158273381295, "step": 42280, "torque_loss": 0.15266047418117523 }, { "epoch": 38.030575539568346, "grad_norm": 0.034941017627716064, "learning_rate": 6.362682610399023e-06, "loss": 0.0191, "step": 42290 }, { "action_loss": 0.0008900340180844069, "epoch": 38.030575539568346, "step": 42290 }, { "epoch": 38.030575539568346, "step": 42290, "torque_loss": 0.1500856876373291 }, { "epoch": 38.039568345323744, "grad_norm": 0.048112623393535614, "learning_rate": 6.346548537720886e-06, "loss": 0.0184, "step": 42300 }, { "action_loss": 0.0011595945106819272, "epoch": 38.039568345323744, "step": 42300 }, { "epoch": 38.039568345323744, "step": 42300, "torque_loss": 0.10285637527704239 }, { "epoch": 38.048561151079134, "grad_norm": 0.1124735102057457, "learning_rate": 6.330433560545512e-06, "loss": 0.019, "step": 42310 }, { "action_loss": 0.0032328504603356123, "epoch": 38.048561151079134, "step": 42310 }, { "epoch": 38.048561151079134, "step": 42310, "torque_loss": 0.15850794315338135 }, { "epoch": 38.05755395683453, "grad_norm": 0.0999177023768425, "learning_rate": 6.3143376859221225e-06, "loss": 0.0152, "step": 42320 }, { "action_loss": 0.003180508268997073, "epoch": 38.05755395683453, "step": 42320 }, { "epoch": 38.05755395683453, "step": 42320, "torque_loss": 0.19414418935775757 }, { "epoch": 38.06654676258993, "grad_norm": 0.045444123446941376, "learning_rate": 6.298260920891624e-06, "loss": 0.0181, "step": 42330 }, { "action_loss": 0.010428208857774734, "epoch": 38.06654676258993, "step": 42330 }, { "epoch": 38.06654676258993, "step": 42330, "torque_loss": 0.22553765773773193 }, { "epoch": 38.07553956834533, "grad_norm": 0.056217972189188004, "learning_rate": 6.282203272486525e-06, "loss": 0.0188, "step": 42340 }, { "action_loss": 0.0011824524262920022, "epoch": 38.07553956834533, "step": 42340 }, { "epoch": 38.07553956834533, "step": 42340, "torque_loss": 0.15862905979156494 }, { "epoch": 38.08453237410072, "grad_norm": 0.06606738269329071, "learning_rate": 6.266164747730985e-06, "loss": 0.0178, "step": 42350 }, { "action_loss": 0.001091175014153123, "epoch": 38.08453237410072, "step": 42350 }, { "epoch": 38.08453237410072, "step": 42350, "torque_loss": 0.12030888348817825 }, { "epoch": 38.093525179856115, "grad_norm": 0.07709832489490509, "learning_rate": 6.250145353640796e-06, "loss": 0.0159, "step": 42360 }, { "action_loss": 0.0006537222652696073, "epoch": 38.093525179856115, "step": 42360 }, { "epoch": 38.093525179856115, "step": 42360, "torque_loss": 0.11570852994918823 }, { "epoch": 38.10251798561151, "grad_norm": 0.08579786121845245, "learning_rate": 6.234145097223404e-06, "loss": 0.014, "step": 42370 }, { "action_loss": 0.0013832791009917855, "epoch": 38.10251798561151, "step": 42370 }, { "epoch": 38.10251798561151, "step": 42370, "torque_loss": 0.15272296965122223 }, { "epoch": 38.111510791366904, "grad_norm": 0.08787189424037933, "learning_rate": 6.218163985477854e-06, "loss": 0.0181, "step": 42380 }, { "action_loss": 0.001343550975434482, "epoch": 38.111510791366904, "step": 42380 }, { "epoch": 38.111510791366904, "step": 42380, "torque_loss": 0.1928831785917282 }, { "epoch": 38.1205035971223, "grad_norm": 0.09174984693527222, "learning_rate": 6.202202025394815e-06, "loss": 0.0184, "step": 42390 }, { "action_loss": 0.0027111044619232416, "epoch": 38.1205035971223, "step": 42390 }, { "epoch": 38.1205035971223, "step": 42390, "torque_loss": 0.09604659676551819 }, { "epoch": 38.1294964028777, "grad_norm": 0.03715192899107933, "learning_rate": 6.186259223956614e-06, "loss": 0.0163, "step": 42400 }, { "action_loss": 0.002611171454191208, "epoch": 38.1294964028777, "step": 42400 }, { "epoch": 38.1294964028777, "step": 42400, "torque_loss": 0.15158437192440033 }, { "epoch": 38.138489208633096, "grad_norm": 0.04749196395277977, "learning_rate": 6.170335588137161e-06, "loss": 0.0191, "step": 42410 }, { "action_loss": 0.0007071160362102091, "epoch": 38.138489208633096, "step": 42410 }, { "epoch": 38.138489208633096, "step": 42410, "torque_loss": 0.12049204111099243 }, { "epoch": 38.14748201438849, "grad_norm": 0.044367481023073196, "learning_rate": 6.154431124901982e-06, "loss": 0.0165, "step": 42420 }, { "action_loss": 0.001471402239985764, "epoch": 38.14748201438849, "step": 42420 }, { "epoch": 38.14748201438849, "step": 42420, "torque_loss": 0.17719805240631104 }, { "epoch": 38.156474820143885, "grad_norm": 0.05139533802866936, "learning_rate": 6.1385458412082595e-06, "loss": 0.0185, "step": 42430 }, { "action_loss": 0.0008163747261278331, "epoch": 38.156474820143885, "step": 42430 }, { "epoch": 38.156474820143885, "step": 42430, "torque_loss": 0.17266809940338135 }, { "epoch": 38.16546762589928, "grad_norm": 0.058822356164455414, "learning_rate": 6.122679744004739e-06, "loss": 0.0173, "step": 42440 }, { "action_loss": 0.003254394978284836, "epoch": 38.16546762589928, "step": 42440 }, { "epoch": 38.16546762589928, "step": 42440, "torque_loss": 0.15180425345897675 }, { "epoch": 38.17446043165467, "grad_norm": 0.12171584367752075, "learning_rate": 6.106832840231786e-06, "loss": 0.0189, "step": 42450 }, { "action_loss": 0.012254313565790653, "epoch": 38.17446043165467, "step": 42450 }, { "epoch": 38.17446043165467, "step": 42450, "torque_loss": 0.19823722541332245 }, { "epoch": 38.18345323741007, "grad_norm": 0.11576937139034271, "learning_rate": 6.091005136821387e-06, "loss": 0.0223, "step": 42460 }, { "action_loss": 0.0028847737703472376, "epoch": 38.18345323741007, "step": 42460 }, { "epoch": 38.18345323741007, "step": 42460, "torque_loss": 0.09474947303533554 }, { "epoch": 38.19244604316547, "grad_norm": 0.06536700576543808, "learning_rate": 6.075196640697101e-06, "loss": 0.0182, "step": 42470 }, { "action_loss": 0.0022572220768779516, "epoch": 38.19244604316547, "step": 42470 }, { "epoch": 38.19244604316547, "step": 42470, "torque_loss": 0.17061568796634674 }, { "epoch": 38.201438848920866, "grad_norm": 0.06310677528381348, "learning_rate": 6.059407358774122e-06, "loss": 0.0176, "step": 42480 }, { "action_loss": 0.0015911298105493188, "epoch": 38.201438848920866, "step": 42480 }, { "epoch": 38.201438848920866, "step": 42480, "torque_loss": 0.16271890699863434 }, { "epoch": 38.210431654676256, "grad_norm": 0.07247700542211533, "learning_rate": 6.043637297959198e-06, "loss": 0.0207, "step": 42490 }, { "action_loss": 0.0010569331934675574, "epoch": 38.210431654676256, "step": 42490 }, { "epoch": 38.210431654676256, "step": 42490, "torque_loss": 0.1627376824617386 }, { "epoch": 38.219424460431654, "grad_norm": 0.047220226377248764, "learning_rate": 6.0278864651507065e-06, "loss": 0.0163, "step": 42500 }, { "action_loss": 0.0009295556810684502, "epoch": 38.219424460431654, "step": 42500 }, { "epoch": 38.219424460431654, "step": 42500, "torque_loss": 0.1823459267616272 }, { "epoch": 38.22841726618705, "grad_norm": 0.053867656737565994, "learning_rate": 6.012154867238584e-06, "loss": 0.0191, "step": 42510 }, { "action_loss": 0.001087130163796246, "epoch": 38.22841726618705, "step": 42510 }, { "epoch": 38.22841726618705, "step": 42510, "torque_loss": 0.1258636862039566 }, { "epoch": 38.23741007194245, "grad_norm": 0.04741089046001434, "learning_rate": 5.996442511104378e-06, "loss": 0.0217, "step": 42520 }, { "action_loss": 0.004075563047081232, "epoch": 38.23741007194245, "step": 42520 }, { "epoch": 38.23741007194245, "step": 42520, "torque_loss": 0.21327979862689972 }, { "epoch": 38.24640287769784, "grad_norm": 0.11042828857898712, "learning_rate": 5.9807494036211995e-06, "loss": 0.022, "step": 42530 }, { "action_loss": 0.0006907818024046719, "epoch": 38.24640287769784, "step": 42530 }, { "epoch": 38.24640287769784, "step": 42530, "torque_loss": 0.12060344219207764 }, { "epoch": 38.25539568345324, "grad_norm": 0.06385179609060287, "learning_rate": 5.965075551653748e-06, "loss": 0.0202, "step": 42540 }, { "action_loss": 0.019396504387259483, "epoch": 38.25539568345324, "step": 42540 }, { "epoch": 38.25539568345324, "step": 42540, "torque_loss": 0.17844410240650177 }, { "epoch": 38.264388489208635, "grad_norm": 0.0874159187078476, "learning_rate": 5.949420962058294e-06, "loss": 0.0218, "step": 42550 }, { "action_loss": 0.0069309622049331665, "epoch": 38.264388489208635, "step": 42550 }, { "epoch": 38.264388489208635, "step": 42550, "torque_loss": 0.1922663450241089 }, { "epoch": 38.273381294964025, "grad_norm": 0.0484338216483593, "learning_rate": 5.933785641682693e-06, "loss": 0.0171, "step": 42560 }, { "action_loss": 0.0010920260101556778, "epoch": 38.273381294964025, "step": 42560 }, { "epoch": 38.273381294964025, "step": 42560, "torque_loss": 0.1487373560667038 }, { "epoch": 38.28237410071942, "grad_norm": 0.05505296587944031, "learning_rate": 5.918169597366369e-06, "loss": 0.0175, "step": 42570 }, { "action_loss": 0.003010520711541176, "epoch": 38.28237410071942, "step": 42570 }, { "epoch": 38.28237410071942, "step": 42570, "torque_loss": 0.16532345116138458 }, { "epoch": 38.29136690647482, "grad_norm": 0.03453463688492775, "learning_rate": 5.902572835940295e-06, "loss": 0.0194, "step": 42580 }, { "action_loss": 0.0031671959441155195, "epoch": 38.29136690647482, "step": 42580 }, { "epoch": 38.29136690647482, "step": 42580, "torque_loss": 0.15892496705055237 }, { "epoch": 38.30035971223022, "grad_norm": 0.04700566828250885, "learning_rate": 5.886995364227044e-06, "loss": 0.0222, "step": 42590 }, { "action_loss": 0.0008882309193722904, "epoch": 38.30035971223022, "step": 42590 }, { "epoch": 38.30035971223022, "step": 42590, "torque_loss": 0.1500418782234192 }, { "epoch": 38.30935251798561, "grad_norm": 0.050925351679325104, "learning_rate": 5.871437189040718e-06, "loss": 0.0187, "step": 42600 }, { "action_loss": 0.0012022062437608838, "epoch": 38.30935251798561, "step": 42600 }, { "epoch": 38.30935251798561, "step": 42600, "torque_loss": 0.17660807073116302 }, { "epoch": 38.318345323741006, "grad_norm": 0.03986629098653793, "learning_rate": 5.855898317186992e-06, "loss": 0.0149, "step": 42610 }, { "action_loss": 0.00102273712400347, "epoch": 38.318345323741006, "step": 42610 }, { "epoch": 38.318345323741006, "step": 42610, "torque_loss": 0.1304580420255661 }, { "epoch": 38.327338129496404, "grad_norm": 0.055414311587810516, "learning_rate": 5.840378755463105e-06, "loss": 0.0171, "step": 42620 }, { "action_loss": 0.0014808349078521132, "epoch": 38.327338129496404, "step": 42620 }, { "epoch": 38.327338129496404, "step": 42620, "torque_loss": 0.13674820959568024 }, { "epoch": 38.3363309352518, "grad_norm": 0.06948042660951614, "learning_rate": 5.824878510657833e-06, "loss": 0.0154, "step": 42630 }, { "action_loss": 0.004254118073731661, "epoch": 38.3363309352518, "step": 42630 }, { "epoch": 38.3363309352518, "step": 42630, "torque_loss": 0.1556008905172348 }, { "epoch": 38.34532374100719, "grad_norm": 0.06258565932512283, "learning_rate": 5.809397589551502e-06, "loss": 0.0187, "step": 42640 }, { "action_loss": 0.012012436985969543, "epoch": 38.34532374100719, "step": 42640 }, { "epoch": 38.34532374100719, "step": 42640, "torque_loss": 0.15699057281017303 }, { "epoch": 38.35431654676259, "grad_norm": 0.053338464349508286, "learning_rate": 5.793935998916017e-06, "loss": 0.0178, "step": 42650 }, { "action_loss": 0.0008913329802453518, "epoch": 38.35431654676259, "step": 42650 }, { "epoch": 38.35431654676259, "step": 42650, "torque_loss": 0.117613285779953 }, { "epoch": 38.36330935251799, "grad_norm": 0.06669580936431885, "learning_rate": 5.778493745514785e-06, "loss": 0.0164, "step": 42660 }, { "action_loss": 0.0013203799026086926, "epoch": 38.36330935251799, "step": 42660 }, { "epoch": 38.36330935251799, "step": 42660, "torque_loss": 0.1603761911392212 }, { "epoch": 38.37230215827338, "grad_norm": 0.12507401406764984, "learning_rate": 5.763070836102768e-06, "loss": 0.0191, "step": 42670 }, { "action_loss": 0.0024367955047637224, "epoch": 38.37230215827338, "step": 42670 }, { "epoch": 38.37230215827338, "step": 42670, "torque_loss": 0.1597738265991211 }, { "epoch": 38.381294964028775, "grad_norm": 0.05913960933685303, "learning_rate": 5.7476672774264885e-06, "loss": 0.0164, "step": 42680 }, { "action_loss": 0.0005909926258027554, "epoch": 38.381294964028775, "step": 42680 }, { "epoch": 38.381294964028775, "step": 42680, "torque_loss": 0.11134696006774902 }, { "epoch": 38.39028776978417, "grad_norm": 0.06533736735582352, "learning_rate": 5.732283076223977e-06, "loss": 0.0177, "step": 42690 }, { "action_loss": 0.0024145550560206175, "epoch": 38.39028776978417, "step": 42690 }, { "epoch": 38.39028776978417, "step": 42690, "torque_loss": 0.1617317646741867 }, { "epoch": 38.39928057553957, "grad_norm": 0.04527072235941887, "learning_rate": 5.716918239224794e-06, "loss": 0.0171, "step": 42700 }, { "action_loss": 0.003486666129902005, "epoch": 38.39928057553957, "step": 42700 }, { "epoch": 38.39928057553957, "step": 42700, "torque_loss": 0.18362009525299072 }, { "epoch": 38.40827338129496, "grad_norm": 0.11521030217409134, "learning_rate": 5.70157277315006e-06, "loss": 0.0244, "step": 42710 }, { "action_loss": 0.0007191989570856094, "epoch": 38.40827338129496, "step": 42710 }, { "epoch": 38.40827338129496, "step": 42710, "torque_loss": 0.11952918767929077 }, { "epoch": 38.41726618705036, "grad_norm": 0.06396497786045074, "learning_rate": 5.686246684712387e-06, "loss": 0.0187, "step": 42720 }, { "action_loss": 0.009719615802168846, "epoch": 38.41726618705036, "step": 42720 }, { "epoch": 38.41726618705036, "step": 42720, "torque_loss": 0.17976336181163788 }, { "epoch": 38.42625899280576, "grad_norm": 0.08930161595344543, "learning_rate": 5.670939980615947e-06, "loss": 0.0199, "step": 42730 }, { "action_loss": 0.004325489979237318, "epoch": 38.42625899280576, "step": 42730 }, { "epoch": 38.42625899280576, "step": 42730, "torque_loss": 0.1676226258277893 }, { "epoch": 38.435251798561154, "grad_norm": 0.04573126882314682, "learning_rate": 5.655652667556388e-06, "loss": 0.0179, "step": 42740 }, { "action_loss": 0.0011401997180655599, "epoch": 38.435251798561154, "step": 42740 }, { "epoch": 38.435251798561154, "step": 42740, "torque_loss": 0.17784719169139862 }, { "epoch": 38.444244604316545, "grad_norm": 0.052137743681669235, "learning_rate": 5.640384752220928e-06, "loss": 0.0177, "step": 42750 }, { "action_loss": 0.0021822191774845123, "epoch": 38.444244604316545, "step": 42750 }, { "epoch": 38.444244604316545, "step": 42750, "torque_loss": 0.21540577709674835 }, { "epoch": 38.45323741007194, "grad_norm": 0.06659938395023346, "learning_rate": 5.625136241288253e-06, "loss": 0.0187, "step": 42760 }, { "action_loss": 0.009790059179067612, "epoch": 38.45323741007194, "step": 42760 }, { "epoch": 38.45323741007194, "step": 42760, "torque_loss": 0.1897241324186325 }, { "epoch": 38.46223021582734, "grad_norm": 0.053375985473394394, "learning_rate": 5.609907141428572e-06, "loss": 0.0174, "step": 42770 }, { "action_loss": 0.0008927487651817501, "epoch": 38.46223021582734, "step": 42770 }, { "epoch": 38.46223021582734, "step": 42770, "torque_loss": 0.11549878120422363 }, { "epoch": 38.47122302158273, "grad_norm": 0.07940716296434402, "learning_rate": 5.594697459303633e-06, "loss": 0.0178, "step": 42780 }, { "action_loss": 0.001970327226445079, "epoch": 38.47122302158273, "step": 42780 }, { "epoch": 38.47122302158273, "step": 42780, "torque_loss": 0.18757522106170654 }, { "epoch": 38.48021582733813, "grad_norm": 0.08611907809972763, "learning_rate": 5.579507201566658e-06, "loss": 0.0212, "step": 42790 }, { "action_loss": 0.0033203375060111284, "epoch": 38.48021582733813, "step": 42790 }, { "epoch": 38.48021582733813, "step": 42790, "torque_loss": 0.21169961988925934 }, { "epoch": 38.489208633093526, "grad_norm": 0.03607136383652687, "learning_rate": 5.564336374862372e-06, "loss": 0.0167, "step": 42800 }, { "action_loss": 0.0012360610999166965, "epoch": 38.489208633093526, "step": 42800 }, { "epoch": 38.489208633093526, "step": 42800, "torque_loss": 0.1334528774023056 }, { "epoch": 38.49820143884892, "grad_norm": 0.053152430802583694, "learning_rate": 5.549184985827027e-06, "loss": 0.0188, "step": 42810 }, { "action_loss": 0.0012954773847013712, "epoch": 38.49820143884892, "step": 42810 }, { "epoch": 38.49820143884892, "step": 42810, "torque_loss": 0.12386459112167358 }, { "epoch": 38.507194244604314, "grad_norm": 0.074305959045887, "learning_rate": 5.534053041088349e-06, "loss": 0.0177, "step": 42820 }, { "action_loss": 0.00130509736482054, "epoch": 38.507194244604314, "step": 42820 }, { "epoch": 38.507194244604314, "step": 42820, "torque_loss": 0.26216256618499756 }, { "epoch": 38.51618705035971, "grad_norm": 0.06702377647161484, "learning_rate": 5.5189405472655495e-06, "loss": 0.0206, "step": 42830 }, { "action_loss": 0.002671522321179509, "epoch": 38.51618705035971, "step": 42830 }, { "epoch": 38.51618705035971, "step": 42830, "torque_loss": 0.14986933767795563 }, { "epoch": 38.52517985611511, "grad_norm": 0.05815712362527847, "learning_rate": 5.503847510969374e-06, "loss": 0.0174, "step": 42840 }, { "action_loss": 0.0017714300192892551, "epoch": 38.52517985611511, "step": 42840 }, { "epoch": 38.52517985611511, "step": 42840, "torque_loss": 0.12525397539138794 }, { "epoch": 38.53417266187051, "grad_norm": 0.040361929684877396, "learning_rate": 5.488773938802011e-06, "loss": 0.0185, "step": 42850 }, { "action_loss": 0.0009710534359328449, "epoch": 38.53417266187051, "step": 42850 }, { "epoch": 38.53417266187051, "step": 42850, "torque_loss": 0.1569676548242569 }, { "epoch": 38.5431654676259, "grad_norm": 0.07397483289241791, "learning_rate": 5.473719837357155e-06, "loss": 0.0202, "step": 42860 }, { "action_loss": 0.0008757147588767111, "epoch": 38.5431654676259, "step": 42860 }, { "epoch": 38.5431654676259, "step": 42860, "torque_loss": 0.15326298773288727 }, { "epoch": 38.552158273381295, "grad_norm": 0.05686574429273605, "learning_rate": 5.4586852132199825e-06, "loss": 0.0175, "step": 42870 }, { "action_loss": 0.0015029212227091193, "epoch": 38.552158273381295, "step": 42870 }, { "epoch": 38.552158273381295, "step": 42870, "torque_loss": 0.13402868807315826 }, { "epoch": 38.56115107913669, "grad_norm": 0.07099306583404541, "learning_rate": 5.443670072967155e-06, "loss": 0.0222, "step": 42880 }, { "action_loss": 0.004476054571568966, "epoch": 38.56115107913669, "step": 42880 }, { "epoch": 38.56115107913669, "step": 42880, "torque_loss": 0.21039797365665436 }, { "epoch": 38.57014388489208, "grad_norm": 0.05339775234460831, "learning_rate": 5.428674423166796e-06, "loss": 0.0238, "step": 42890 }, { "action_loss": 0.0013891999842599034, "epoch": 38.57014388489208, "step": 42890 }, { "epoch": 38.57014388489208, "step": 42890, "torque_loss": 0.11700282245874405 }, { "epoch": 38.57913669064748, "grad_norm": 0.043130699545145035, "learning_rate": 5.413698270378515e-06, "loss": 0.0189, "step": 42900 }, { "action_loss": 0.0014488723827525973, "epoch": 38.57913669064748, "step": 42900 }, { "epoch": 38.57913669064748, "step": 42900, "torque_loss": 0.1130807027220726 }, { "epoch": 38.58812949640288, "grad_norm": 0.051837258040905, "learning_rate": 5.398741621153397e-06, "loss": 0.0184, "step": 42910 }, { "action_loss": 0.002246892312541604, "epoch": 38.58812949640288, "step": 42910 }, { "epoch": 38.58812949640288, "step": 42910, "torque_loss": 0.1667078584432602 }, { "epoch": 38.597122302158276, "grad_norm": 0.06566240638494492, "learning_rate": 5.383804482033983e-06, "loss": 0.0168, "step": 42920 }, { "action_loss": 0.004622059408575296, "epoch": 38.597122302158276, "step": 42920 }, { "epoch": 38.597122302158276, "step": 42920, "torque_loss": 0.1721739023923874 }, { "epoch": 38.606115107913666, "grad_norm": 0.07643205672502518, "learning_rate": 5.368886859554273e-06, "loss": 0.0198, "step": 42930 }, { "action_loss": 0.0007801707834005356, "epoch": 38.606115107913666, "step": 42930 }, { "epoch": 38.606115107913666, "step": 42930, "torque_loss": 0.16039584577083588 }, { "epoch": 38.615107913669064, "grad_norm": 0.07050401717424393, "learning_rate": 5.353988760239764e-06, "loss": 0.0193, "step": 42940 }, { "action_loss": 0.0008953860378824174, "epoch": 38.615107913669064, "step": 42940 }, { "epoch": 38.615107913669064, "step": 42940, "torque_loss": 0.10242652893066406 }, { "epoch": 38.62410071942446, "grad_norm": 0.057998672127723694, "learning_rate": 5.339110190607377e-06, "loss": 0.0167, "step": 42950 }, { "action_loss": 0.002482575597241521, "epoch": 38.62410071942446, "step": 42950 }, { "epoch": 38.62410071942446, "step": 42950, "torque_loss": 0.13250966370105743 }, { "epoch": 38.63309352517986, "grad_norm": 0.13040538132190704, "learning_rate": 5.324251157165511e-06, "loss": 0.0206, "step": 42960 }, { "action_loss": 0.002552793361246586, "epoch": 38.63309352517986, "step": 42960 }, { "epoch": 38.63309352517986, "step": 42960, "torque_loss": 0.16683627665042877 }, { "epoch": 38.64208633093525, "grad_norm": 0.052463237196207047, "learning_rate": 5.309411666413999e-06, "loss": 0.0197, "step": 42970 }, { "action_loss": 0.0013189790770411491, "epoch": 38.64208633093525, "step": 42970 }, { "epoch": 38.64208633093525, "step": 42970, "torque_loss": 0.08760831505060196 }, { "epoch": 38.65107913669065, "grad_norm": 0.11566832661628723, "learning_rate": 5.294591724844161e-06, "loss": 0.017, "step": 42980 }, { "action_loss": 0.0005271651898510754, "epoch": 38.65107913669065, "step": 42980 }, { "epoch": 38.65107913669065, "step": 42980, "torque_loss": 0.14176033437252045 }, { "epoch": 38.660071942446045, "grad_norm": 0.12864789366722107, "learning_rate": 5.279791338938716e-06, "loss": 0.0198, "step": 42990 }, { "action_loss": 0.002600444946438074, "epoch": 38.660071942446045, "step": 42990 }, { "epoch": 38.660071942446045, "step": 42990, "torque_loss": 0.17624831199645996 }, { "epoch": 38.669064748201436, "grad_norm": 0.06030598655343056, "learning_rate": 5.265010515171886e-06, "loss": 0.0174, "step": 43000 }, { "action_loss": 0.001616306253708899, "epoch": 38.669064748201436, "step": 43000 }, { "epoch": 38.669064748201436, "step": 43000, "torque_loss": 0.14188487827777863 }, { "epoch": 38.67805755395683, "grad_norm": 0.10340328514575958, "learning_rate": 5.250249260009282e-06, "loss": 0.019, "step": 43010 }, { "action_loss": 0.004163767211139202, "epoch": 38.67805755395683, "step": 43010 }, { "epoch": 38.67805755395683, "step": 43010, "torque_loss": 0.19215834140777588 }, { "epoch": 38.68705035971223, "grad_norm": 0.0586693175137043, "learning_rate": 5.2355075799079835e-06, "loss": 0.0172, "step": 43020 }, { "action_loss": 0.004296261351555586, "epoch": 38.68705035971223, "step": 43020 }, { "epoch": 38.68705035971223, "step": 43020, "torque_loss": 0.12914639711380005 }, { "epoch": 38.69604316546763, "grad_norm": 0.06820395588874817, "learning_rate": 5.220785481316509e-06, "loss": 0.018, "step": 43030 }, { "action_loss": 0.0028647631406784058, "epoch": 38.69604316546763, "step": 43030 }, { "epoch": 38.69604316546763, "step": 43030, "torque_loss": 0.2016553431749344 }, { "epoch": 38.70503597122302, "grad_norm": 0.05956597626209259, "learning_rate": 5.206082970674797e-06, "loss": 0.0194, "step": 43040 }, { "action_loss": 0.0007786068017594516, "epoch": 38.70503597122302, "step": 43040 }, { "epoch": 38.70503597122302, "step": 43040, "torque_loss": 0.12030748277902603 }, { "epoch": 38.71402877697842, "grad_norm": 0.08960975706577301, "learning_rate": 5.191400054414231e-06, "loss": 0.0156, "step": 43050 }, { "action_loss": 0.0043816519901156425, "epoch": 38.71402877697842, "step": 43050 }, { "epoch": 38.71402877697842, "step": 43050, "torque_loss": 0.1784217357635498 }, { "epoch": 38.723021582733814, "grad_norm": 0.04463832080364227, "learning_rate": 5.176736738957605e-06, "loss": 0.0167, "step": 43060 }, { "action_loss": 0.004694166127592325, "epoch": 38.723021582733814, "step": 43060 }, { "epoch": 38.723021582733814, "step": 43060, "torque_loss": 0.20430447161197662 }, { "epoch": 38.73201438848921, "grad_norm": 0.04671015962958336, "learning_rate": 5.162093030719173e-06, "loss": 0.0201, "step": 43070 }, { "action_loss": 0.001239377656020224, "epoch": 38.73201438848921, "step": 43070 }, { "epoch": 38.73201438848921, "step": 43070, "torque_loss": 0.16384252905845642 }, { "epoch": 38.7410071942446, "grad_norm": 0.029468519613146782, "learning_rate": 5.1474689361045715e-06, "loss": 0.0164, "step": 43080 }, { "action_loss": 0.00646629324182868, "epoch": 38.7410071942446, "step": 43080 }, { "epoch": 38.7410071942446, "step": 43080, "torque_loss": 0.15984027087688446 }, { "epoch": 38.75, "grad_norm": 0.028798328712582588, "learning_rate": 5.132864461510878e-06, "loss": 0.0146, "step": 43090 }, { "action_loss": 0.0005318688345141709, "epoch": 38.75, "step": 43090 }, { "epoch": 38.75, "step": 43090, "torque_loss": 0.12632156908512115 }, { "epoch": 38.7589928057554, "grad_norm": 0.057669397443532944, "learning_rate": 5.118279613326604e-06, "loss": 0.0178, "step": 43100 }, { "action_loss": 0.0016536841867491603, "epoch": 38.7589928057554, "step": 43100 }, { "epoch": 38.7589928057554, "step": 43100, "torque_loss": 0.18986725807189941 }, { "epoch": 38.76798561151079, "grad_norm": 0.12905395030975342, "learning_rate": 5.103714397931636e-06, "loss": 0.0179, "step": 43110 }, { "action_loss": 0.004349039867520332, "epoch": 38.76798561151079, "step": 43110 }, { "epoch": 38.76798561151079, "step": 43110, "torque_loss": 0.20633113384246826 }, { "epoch": 38.776978417266186, "grad_norm": 0.06453090161085129, "learning_rate": 5.0891688216972985e-06, "loss": 0.0155, "step": 43120 }, { "action_loss": 0.006912784185260534, "epoch": 38.776978417266186, "step": 43120 }, { "epoch": 38.776978417266186, "step": 43120, "torque_loss": 0.21426473557949066 }, { "epoch": 38.78597122302158, "grad_norm": 0.04606693238019943, "learning_rate": 5.074642890986336e-06, "loss": 0.0192, "step": 43130 }, { "action_loss": 0.0007393377018161118, "epoch": 38.78597122302158, "step": 43130 }, { "epoch": 38.78597122302158, "step": 43130, "torque_loss": 0.16828711330890656 }, { "epoch": 38.79496402877698, "grad_norm": 0.0455792173743248, "learning_rate": 5.060136612152871e-06, "loss": 0.0165, "step": 43140 }, { "action_loss": 0.0006514799897558987, "epoch": 38.79496402877698, "step": 43140 }, { "epoch": 38.79496402877698, "step": 43140, "torque_loss": 0.137227401137352 }, { "epoch": 38.80395683453237, "grad_norm": 0.046251874417066574, "learning_rate": 5.045649991542434e-06, "loss": 0.021, "step": 43150 }, { "action_loss": 0.0010042650392279029, "epoch": 38.80395683453237, "step": 43150 }, { "epoch": 38.80395683453237, "step": 43150, "torque_loss": 0.13460706174373627 }, { "epoch": 38.81294964028777, "grad_norm": 0.0710509642958641, "learning_rate": 5.031183035491982e-06, "loss": 0.016, "step": 43160 }, { "action_loss": 0.0006269235746003687, "epoch": 38.81294964028777, "step": 43160 }, { "epoch": 38.81294964028777, "step": 43160, "torque_loss": 0.1457054167985916 }, { "epoch": 38.82194244604317, "grad_norm": 0.04875088855624199, "learning_rate": 5.016735750329854e-06, "loss": 0.0227, "step": 43170 }, { "action_loss": 0.003421621397137642, "epoch": 38.82194244604317, "step": 43170 }, { "epoch": 38.82194244604317, "step": 43170, "torque_loss": 0.17173318564891815 }, { "epoch": 38.830935251798564, "grad_norm": 0.11596594005823135, "learning_rate": 5.002308142375761e-06, "loss": 0.0207, "step": 43180 }, { "action_loss": 0.009155008010566235, "epoch": 38.830935251798564, "step": 43180 }, { "epoch": 38.830935251798564, "step": 43180, "torque_loss": 0.16262286901474 }, { "epoch": 38.839928057553955, "grad_norm": 0.09018560498952866, "learning_rate": 4.987900217940855e-06, "loss": 0.0174, "step": 43190 }, { "action_loss": 0.0006899679428897798, "epoch": 38.839928057553955, "step": 43190 }, { "epoch": 38.839928057553955, "step": 43190, "torque_loss": 0.1376054286956787 }, { "epoch": 38.84892086330935, "grad_norm": 0.08918216824531555, "learning_rate": 4.973511983327633e-06, "loss": 0.0158, "step": 43200 }, { "action_loss": 0.0019729037303477526, "epoch": 38.84892086330935, "step": 43200 }, { "epoch": 38.84892086330935, "step": 43200, "torque_loss": 0.21833555400371552 }, { "epoch": 38.85791366906475, "grad_norm": 0.032423462718725204, "learning_rate": 4.959143444830011e-06, "loss": 0.0191, "step": 43210 }, { "action_loss": 0.0033932628575712442, "epoch": 38.85791366906475, "step": 43210 }, { "epoch": 38.85791366906475, "step": 43210, "torque_loss": 0.1790115088224411 }, { "epoch": 38.86690647482014, "grad_norm": 0.09675239771604538, "learning_rate": 4.944794608733267e-06, "loss": 0.0205, "step": 43220 }, { "action_loss": 0.0075646378099918365, "epoch": 38.86690647482014, "step": 43220 }, { "epoch": 38.86690647482014, "step": 43220, "torque_loss": 0.15841154754161835 }, { "epoch": 38.87589928057554, "grad_norm": 0.06125636398792267, "learning_rate": 4.930465481314078e-06, "loss": 0.0191, "step": 43230 }, { "action_loss": 0.00047685267054475844, "epoch": 38.87589928057554, "step": 43230 }, { "epoch": 38.87589928057554, "step": 43230, "torque_loss": 0.08932632207870483 }, { "epoch": 38.884892086330936, "grad_norm": 0.06549660116434097, "learning_rate": 4.916156068840483e-06, "loss": 0.0172, "step": 43240 }, { "action_loss": 0.0015617158496752381, "epoch": 38.884892086330936, "step": 43240 }, { "epoch": 38.884892086330936, "step": 43240, "torque_loss": 0.09784158319234848 }, { "epoch": 38.893884892086334, "grad_norm": 0.05799908936023712, "learning_rate": 4.901866377571901e-06, "loss": 0.0197, "step": 43250 }, { "action_loss": 0.0017475169152021408, "epoch": 38.893884892086334, "step": 43250 }, { "epoch": 38.893884892086334, "step": 43250, "torque_loss": 0.11713046580553055 }, { "epoch": 38.902877697841724, "grad_norm": 0.08880750834941864, "learning_rate": 4.887596413759143e-06, "loss": 0.0169, "step": 43260 }, { "action_loss": 0.0005755724268965423, "epoch": 38.902877697841724, "step": 43260 }, { "epoch": 38.902877697841724, "step": 43260, "torque_loss": 0.10366632789373398 }, { "epoch": 38.91187050359712, "grad_norm": 0.04275050386786461, "learning_rate": 4.873346183644373e-06, "loss": 0.017, "step": 43270 }, { "action_loss": 0.0015872785588726401, "epoch": 38.91187050359712, "step": 43270 }, { "epoch": 38.91187050359712, "step": 43270, "torque_loss": 0.14625783264636993 }, { "epoch": 38.92086330935252, "grad_norm": 0.09521574527025223, "learning_rate": 4.8591156934611085e-06, "loss": 0.0186, "step": 43280 }, { "action_loss": 0.001718660700134933, "epoch": 38.92086330935252, "step": 43280 }, { "epoch": 38.92086330935252, "step": 43280, "torque_loss": 0.13419412076473236 }, { "epoch": 38.92985611510792, "grad_norm": 0.05122873932123184, "learning_rate": 4.8449049494342704e-06, "loss": 0.0157, "step": 43290 }, { "action_loss": 0.004321198910474777, "epoch": 38.92985611510792, "step": 43290 }, { "epoch": 38.92985611510792, "step": 43290, "torque_loss": 0.2079758495092392 }, { "epoch": 38.93884892086331, "grad_norm": 0.04664567485451698, "learning_rate": 4.830713957780114e-06, "loss": 0.018, "step": 43300 }, { "action_loss": 0.0018066061893478036, "epoch": 38.93884892086331, "step": 43300 }, { "epoch": 38.93884892086331, "step": 43300, "torque_loss": 0.18751977384090424 }, { "epoch": 38.947841726618705, "grad_norm": 0.037274084985256195, "learning_rate": 4.816542724706253e-06, "loss": 0.0166, "step": 43310 }, { "action_loss": 0.005923821125179529, "epoch": 38.947841726618705, "step": 43310 }, { "epoch": 38.947841726618705, "step": 43310, "torque_loss": 0.15845780074596405 }, { "epoch": 38.9568345323741, "grad_norm": 0.029321787878870964, "learning_rate": 4.802391256411675e-06, "loss": 0.0167, "step": 43320 }, { "action_loss": 0.0011026826687157154, "epoch": 38.9568345323741, "step": 43320 }, { "epoch": 38.9568345323741, "step": 43320, "torque_loss": 0.1328093260526657 }, { "epoch": 38.96582733812949, "grad_norm": 0.13990920782089233, "learning_rate": 4.788259559086717e-06, "loss": 0.0161, "step": 43330 }, { "action_loss": 0.0006338268867693841, "epoch": 38.96582733812949, "step": 43330 }, { "epoch": 38.96582733812949, "step": 43330, "torque_loss": 0.10499467700719833 }, { "epoch": 38.97482014388489, "grad_norm": 0.07085034996271133, "learning_rate": 4.774147638913046e-06, "loss": 0.0198, "step": 43340 }, { "action_loss": 0.001612056978046894, "epoch": 38.97482014388489, "step": 43340 }, { "epoch": 38.97482014388489, "step": 43340, "torque_loss": 0.13982772827148438 }, { "epoch": 38.98381294964029, "grad_norm": 0.05936288461089134, "learning_rate": 4.760055502063721e-06, "loss": 0.018, "step": 43350 }, { "action_loss": 0.0016678167739883065, "epoch": 38.98381294964029, "step": 43350 }, { "epoch": 38.98381294964029, "step": 43350, "torque_loss": 0.12063447386026382 }, { "epoch": 38.992805755395686, "grad_norm": 0.059623055160045624, "learning_rate": 4.745983154703099e-06, "loss": 0.0171, "step": 43360 }, { "action_loss": 0.0022073877044022083, "epoch": 38.992805755395686, "step": 43360 }, { "epoch": 38.992805755395686, "step": 43360, "torque_loss": 0.22586627304553986 }, { "epoch": 39.00179856115108, "grad_norm": 0.03137751296162605, "learning_rate": 4.731930602986906e-06, "loss": 0.0192, "step": 43370 }, { "action_loss": 0.000511264952365309, "epoch": 39.00179856115108, "step": 43370 }, { "epoch": 39.00179856115108, "step": 43370, "torque_loss": 0.1426137089729309 }, { "epoch": 39.010791366906474, "grad_norm": 0.15716542303562164, "learning_rate": 4.7178978530622196e-06, "loss": 0.0212, "step": 43380 }, { "action_loss": 0.000761491886805743, "epoch": 39.010791366906474, "step": 43380 }, { "epoch": 39.010791366906474, "step": 43380, "torque_loss": 0.10746441036462784 }, { "epoch": 39.01978417266187, "grad_norm": 0.04597942903637886, "learning_rate": 4.70388491106743e-06, "loss": 0.0216, "step": 43390 }, { "action_loss": 0.003213794669136405, "epoch": 39.01978417266187, "step": 43390 }, { "epoch": 39.01978417266187, "step": 43390, "torque_loss": 0.08720684051513672 }, { "epoch": 39.02877697841727, "grad_norm": 0.06924919784069061, "learning_rate": 4.689891783132278e-06, "loss": 0.0189, "step": 43400 }, { "action_loss": 0.002309758448973298, "epoch": 39.02877697841727, "step": 43400 }, { "epoch": 39.02877697841727, "step": 43400, "torque_loss": 0.16734719276428223 }, { "epoch": 39.03776978417266, "grad_norm": 0.06114083528518677, "learning_rate": 4.675918475377822e-06, "loss": 0.0172, "step": 43410 }, { "action_loss": 0.003273393725976348, "epoch": 39.03776978417266, "step": 43410 }, { "epoch": 39.03776978417266, "step": 43410, "torque_loss": 0.15677697956562042 }, { "epoch": 39.04676258992806, "grad_norm": 0.04313582926988602, "learning_rate": 4.6619649939164845e-06, "loss": 0.017, "step": 43420 }, { "action_loss": 0.005581945646554232, "epoch": 39.04676258992806, "step": 43420 }, { "epoch": 39.04676258992806, "step": 43420, "torque_loss": 0.22785501182079315 }, { "epoch": 39.055755395683455, "grad_norm": 0.08957116305828094, "learning_rate": 4.6480313448519665e-06, "loss": 0.0201, "step": 43430 }, { "action_loss": 0.0011057330993935466, "epoch": 39.055755395683455, "step": 43430 }, { "epoch": 39.055755395683455, "step": 43430, "torque_loss": 0.09812530875205994 }, { "epoch": 39.064748201438846, "grad_norm": 0.09114480018615723, "learning_rate": 4.634117534279337e-06, "loss": 0.0161, "step": 43440 }, { "action_loss": 0.0008162555168382823, "epoch": 39.064748201438846, "step": 43440 }, { "epoch": 39.064748201438846, "step": 43440, "torque_loss": 0.12627200782299042 }, { "epoch": 39.07374100719424, "grad_norm": 0.08800221234560013, "learning_rate": 4.620223568284959e-06, "loss": 0.0166, "step": 43450 }, { "action_loss": 0.0023439296055585146, "epoch": 39.07374100719424, "step": 43450 }, { "epoch": 39.07374100719424, "step": 43450, "torque_loss": 0.17204521596431732 }, { "epoch": 39.08273381294964, "grad_norm": 0.14949104189872742, "learning_rate": 4.60634945294654e-06, "loss": 0.0199, "step": 43460 }, { "action_loss": 0.0007191440090537071, "epoch": 39.08273381294964, "step": 43460 }, { "epoch": 39.08273381294964, "step": 43460, "torque_loss": 0.12994267046451569 }, { "epoch": 39.09172661870504, "grad_norm": 0.0420435406267643, "learning_rate": 4.592495194333074e-06, "loss": 0.0175, "step": 43470 }, { "action_loss": 0.0009315079660154879, "epoch": 39.09172661870504, "step": 43470 }, { "epoch": 39.09172661870504, "step": 43470, "torque_loss": 0.1477973908185959 }, { "epoch": 39.10071942446043, "grad_norm": 0.042586151510477066, "learning_rate": 4.578660798504897e-06, "loss": 0.0156, "step": 43480 }, { "action_loss": 0.0006007030024193227, "epoch": 39.10071942446043, "step": 43480 }, { "epoch": 39.10071942446043, "step": 43480, "torque_loss": 0.13439543545246124 }, { "epoch": 39.10971223021583, "grad_norm": 0.04142146185040474, "learning_rate": 4.5648462715136346e-06, "loss": 0.0187, "step": 43490 }, { "action_loss": 0.0014357101172208786, "epoch": 39.10971223021583, "step": 43490 }, { "epoch": 39.10971223021583, "step": 43490, "torque_loss": 0.22627873718738556 }, { "epoch": 39.118705035971225, "grad_norm": 0.05122925341129303, "learning_rate": 4.551051619402225e-06, "loss": 0.0192, "step": 43500 }, { "action_loss": 0.021620502695441246, "epoch": 39.118705035971225, "step": 43500 }, { "epoch": 39.118705035971225, "step": 43500, "torque_loss": 0.2136193960905075 }, { "epoch": 39.12769784172662, "grad_norm": 0.0383622944355011, "learning_rate": 4.537276848204936e-06, "loss": 0.0208, "step": 43510 }, { "action_loss": 0.001035218476317823, "epoch": 39.12769784172662, "step": 43510 }, { "epoch": 39.12769784172662, "step": 43510, "torque_loss": 0.14062272012233734 }, { "epoch": 39.13669064748201, "grad_norm": 0.04973725229501724, "learning_rate": 4.523521963947308e-06, "loss": 0.0171, "step": 43520 }, { "action_loss": 0.000395251321606338, "epoch": 39.13669064748201, "step": 43520 }, { "epoch": 39.13669064748201, "step": 43520, "torque_loss": 0.11277355998754501 }, { "epoch": 39.14568345323741, "grad_norm": 0.06405878067016602, "learning_rate": 4.509786972646191e-06, "loss": 0.0188, "step": 43530 }, { "action_loss": 0.0007396815344691277, "epoch": 39.14568345323741, "step": 43530 }, { "epoch": 39.14568345323741, "step": 43530, "torque_loss": 0.13114476203918457 }, { "epoch": 39.15467625899281, "grad_norm": 0.02542947791516781, "learning_rate": 4.496071880309749e-06, "loss": 0.0151, "step": 43540 }, { "action_loss": 0.000864419445861131, "epoch": 39.15467625899281, "step": 43540 }, { "epoch": 39.15467625899281, "step": 43540, "torque_loss": 0.11282992362976074 }, { "epoch": 39.1636690647482, "grad_norm": 0.10063283890485764, "learning_rate": 4.4823766929374234e-06, "loss": 0.0167, "step": 43550 }, { "action_loss": 0.0006018150597810745, "epoch": 39.1636690647482, "step": 43550 }, { "epoch": 39.1636690647482, "step": 43550, "torque_loss": 0.15637683868408203 }, { "epoch": 39.172661870503596, "grad_norm": 0.07024732232093811, "learning_rate": 4.468701416519955e-06, "loss": 0.0196, "step": 43560 }, { "action_loss": 0.0010167057625949383, "epoch": 39.172661870503596, "step": 43560 }, { "epoch": 39.172661870503596, "step": 43560, "torque_loss": 0.09906811267137527 }, { "epoch": 39.181654676258994, "grad_norm": 0.057678088545799255, "learning_rate": 4.4550460570393635e-06, "loss": 0.0173, "step": 43570 }, { "action_loss": 0.010527096688747406, "epoch": 39.181654676258994, "step": 43570 }, { "epoch": 39.181654676258994, "step": 43570, "torque_loss": 0.22979247570037842 }, { "epoch": 39.19064748201439, "grad_norm": 0.09272950142621994, "learning_rate": 4.441410620468989e-06, "loss": 0.0176, "step": 43580 }, { "action_loss": 0.0013321031583473086, "epoch": 39.19064748201439, "step": 43580 }, { "epoch": 39.19064748201439, "step": 43580, "torque_loss": 0.1021658182144165 }, { "epoch": 39.19964028776978, "grad_norm": 0.07246216386556625, "learning_rate": 4.427795112773419e-06, "loss": 0.0154, "step": 43590 }, { "action_loss": 0.002635716227814555, "epoch": 39.19964028776978, "step": 43590 }, { "epoch": 39.19964028776978, "step": 43590, "torque_loss": 0.16005586087703705 }, { "epoch": 39.20863309352518, "grad_norm": 0.04647119343280792, "learning_rate": 4.414199539908531e-06, "loss": 0.0174, "step": 43600 }, { "action_loss": 0.0004846658557653427, "epoch": 39.20863309352518, "step": 43600 }, { "epoch": 39.20863309352518, "step": 43600, "torque_loss": 0.11144167184829712 }, { "epoch": 39.21762589928058, "grad_norm": 0.10463470220565796, "learning_rate": 4.400623907821516e-06, "loss": 0.0169, "step": 43610 }, { "action_loss": 0.0003897062561009079, "epoch": 39.21762589928058, "step": 43610 }, { "epoch": 39.21762589928058, "step": 43610, "torque_loss": 0.1096026673913002 }, { "epoch": 39.226618705035975, "grad_norm": 0.10006248950958252, "learning_rate": 4.387068222450802e-06, "loss": 0.0233, "step": 43620 }, { "action_loss": 0.0004934371099807322, "epoch": 39.226618705035975, "step": 43620 }, { "epoch": 39.226618705035975, "step": 43620, "torque_loss": 0.20420245826244354 }, { "epoch": 39.235611510791365, "grad_norm": 0.04286454617977142, "learning_rate": 4.3735324897260954e-06, "loss": 0.0157, "step": 43630 }, { "action_loss": 0.0009009846835397184, "epoch": 39.235611510791365, "step": 43630 }, { "epoch": 39.235611510791365, "step": 43630, "torque_loss": 0.17079244554042816 }, { "epoch": 39.24460431654676, "grad_norm": 0.03750170022249222, "learning_rate": 4.36001671556841e-06, "loss": 0.0145, "step": 43640 }, { "action_loss": 0.002480751136317849, "epoch": 39.24460431654676, "step": 43640 }, { "epoch": 39.24460431654676, "step": 43640, "torque_loss": 0.16170521080493927 }, { "epoch": 39.25359712230216, "grad_norm": 0.11116047203540802, "learning_rate": 4.346520905889989e-06, "loss": 0.0174, "step": 43650 }, { "action_loss": 0.002735388232395053, "epoch": 39.25359712230216, "step": 43650 }, { "epoch": 39.25359712230216, "step": 43650, "torque_loss": 0.13635338842868805 }, { "epoch": 39.26258992805755, "grad_norm": 0.06906181573867798, "learning_rate": 4.333045066594354e-06, "loss": 0.0192, "step": 43660 }, { "action_loss": 0.0008527730242349207, "epoch": 39.26258992805755, "step": 43660 }, { "epoch": 39.26258992805755, "step": 43660, "torque_loss": 0.15758661925792694 }, { "epoch": 39.27158273381295, "grad_norm": 0.06205800548195839, "learning_rate": 4.319589203576313e-06, "loss": 0.0163, "step": 43670 }, { "action_loss": 0.0009774916106835008, "epoch": 39.27158273381295, "step": 43670 }, { "epoch": 39.27158273381295, "step": 43670, "torque_loss": 0.13790196180343628 }, { "epoch": 39.280575539568346, "grad_norm": 0.13186298310756683, "learning_rate": 4.306153322721895e-06, "loss": 0.0177, "step": 43680 }, { "action_loss": 0.001528030727058649, "epoch": 39.280575539568346, "step": 43680 }, { "epoch": 39.280575539568346, "step": 43680, "torque_loss": 0.13404661417007446 }, { "epoch": 39.289568345323744, "grad_norm": 0.0460977666079998, "learning_rate": 4.292737429908422e-06, "loss": 0.018, "step": 43690 }, { "action_loss": 0.0016195668140426278, "epoch": 39.289568345323744, "step": 43690 }, { "epoch": 39.289568345323744, "step": 43690, "torque_loss": 0.17575900256633759 }, { "epoch": 39.298561151079134, "grad_norm": 0.04561508074402809, "learning_rate": 4.279341531004471e-06, "loss": 0.02, "step": 43700 }, { "action_loss": 0.00053079315694049, "epoch": 39.298561151079134, "step": 43700 }, { "epoch": 39.298561151079134, "step": 43700, "torque_loss": 0.14880986511707306 }, { "epoch": 39.30755395683453, "grad_norm": 0.10481049120426178, "learning_rate": 4.26596563186985e-06, "loss": 0.0173, "step": 43710 }, { "action_loss": 0.0026141817215830088, "epoch": 39.30755395683453, "step": 43710 }, { "epoch": 39.30755395683453, "step": 43710, "torque_loss": 0.2087504118680954 }, { "epoch": 39.31654676258993, "grad_norm": 0.04149629548192024, "learning_rate": 4.252609738355634e-06, "loss": 0.0189, "step": 43720 }, { "action_loss": 0.0015714679611846805, "epoch": 39.31654676258993, "step": 43720 }, { "epoch": 39.31654676258993, "step": 43720, "torque_loss": 0.17751234769821167 }, { "epoch": 39.32553956834532, "grad_norm": 0.10454102605581284, "learning_rate": 4.239273856304149e-06, "loss": 0.0177, "step": 43730 }, { "action_loss": 0.001390215358696878, "epoch": 39.32553956834532, "step": 43730 }, { "epoch": 39.32553956834532, "step": 43730, "torque_loss": 0.1373143047094345 }, { "epoch": 39.33453237410072, "grad_norm": 0.10542908310890198, "learning_rate": 4.2259579915489625e-06, "loss": 0.0187, "step": 43740 }, { "action_loss": 0.0011235737474635243, "epoch": 39.33453237410072, "step": 43740 }, { "epoch": 39.33453237410072, "step": 43740, "torque_loss": 0.14400041103363037 }, { "epoch": 39.343525179856115, "grad_norm": 0.12765854597091675, "learning_rate": 4.212662149914886e-06, "loss": 0.0169, "step": 43750 }, { "action_loss": 0.0020834181923419237, "epoch": 39.343525179856115, "step": 43750 }, { "epoch": 39.343525179856115, "step": 43750, "torque_loss": 0.15433400869369507 }, { "epoch": 39.35251798561151, "grad_norm": 0.055199552327394485, "learning_rate": 4.199386337217958e-06, "loss": 0.0188, "step": 43760 }, { "action_loss": 0.0008949429611675441, "epoch": 39.35251798561151, "step": 43760 }, { "epoch": 39.35251798561151, "step": 43760, "torque_loss": 0.10628703236579895 }, { "epoch": 39.361510791366904, "grad_norm": 0.06375440210103989, "learning_rate": 4.186130559265494e-06, "loss": 0.017, "step": 43770 }, { "action_loss": 0.0012934299884364009, "epoch": 39.361510791366904, "step": 43770 }, { "epoch": 39.361510791366904, "step": 43770, "torque_loss": 0.15247052907943726 }, { "epoch": 39.3705035971223, "grad_norm": 0.03863959759473801, "learning_rate": 4.172894821856011e-06, "loss": 0.0169, "step": 43780 }, { "action_loss": 0.0013376515125855803, "epoch": 39.3705035971223, "step": 43780 }, { "epoch": 39.3705035971223, "step": 43780, "torque_loss": 0.16986866295337677 }, { "epoch": 39.3794964028777, "grad_norm": 0.044768549501895905, "learning_rate": 4.159679130779254e-06, "loss": 0.0168, "step": 43790 }, { "action_loss": 0.017790712416172028, "epoch": 39.3794964028777, "step": 43790 }, { "epoch": 39.3794964028777, "step": 43790, "torque_loss": 0.1831369400024414 }, { "epoch": 39.388489208633096, "grad_norm": 0.08695049583911896, "learning_rate": 4.1464834918162445e-06, "loss": 0.0189, "step": 43800 }, { "action_loss": 0.0004648704780265689, "epoch": 39.388489208633096, "step": 43800 }, { "epoch": 39.388489208633096, "step": 43800, "torque_loss": 0.11211779713630676 }, { "epoch": 39.39748201438849, "grad_norm": 0.06574749946594238, "learning_rate": 4.133307910739187e-06, "loss": 0.0205, "step": 43810 }, { "action_loss": 0.001625177450478077, "epoch": 39.39748201438849, "step": 43810 }, { "epoch": 39.39748201438849, "step": 43810, "torque_loss": 0.12638652324676514 }, { "epoch": 39.406474820143885, "grad_norm": 0.026813426986336708, "learning_rate": 4.120152393311521e-06, "loss": 0.017, "step": 43820 }, { "action_loss": 0.0017236362909898162, "epoch": 39.406474820143885, "step": 43820 }, { "epoch": 39.406474820143885, "step": 43820, "torque_loss": 0.13749364018440247 }, { "epoch": 39.41546762589928, "grad_norm": 0.042580995708703995, "learning_rate": 4.107016945287933e-06, "loss": 0.016, "step": 43830 }, { "action_loss": 0.00048301686183549464, "epoch": 39.41546762589928, "step": 43830 }, { "epoch": 39.41546762589928, "step": 43830, "torque_loss": 0.10997983813285828 }, { "epoch": 39.42446043165468, "grad_norm": 0.050933752208948135, "learning_rate": 4.09390157241431e-06, "loss": 0.0164, "step": 43840 }, { "action_loss": 0.0009923955658450723, "epoch": 39.42446043165468, "step": 43840 }, { "epoch": 39.42446043165468, "step": 43840, "torque_loss": 0.09941098093986511 }, { "epoch": 39.43345323741007, "grad_norm": 0.10544679313898087, "learning_rate": 4.0808062804277515e-06, "loss": 0.0204, "step": 43850 }, { "action_loss": 0.008495043031871319, "epoch": 39.43345323741007, "step": 43850 }, { "epoch": 39.43345323741007, "step": 43850, "torque_loss": 0.20722921192646027 }, { "epoch": 39.44244604316547, "grad_norm": 0.08135449141263962, "learning_rate": 4.067731075056596e-06, "loss": 0.0226, "step": 43860 }, { "action_loss": 0.012836635112762451, "epoch": 39.44244604316547, "step": 43860 }, { "epoch": 39.44244604316547, "step": 43860, "torque_loss": 0.27102208137512207 }, { "epoch": 39.451438848920866, "grad_norm": 0.1029772013425827, "learning_rate": 4.054675962020376e-06, "loss": 0.0202, "step": 43870 }, { "action_loss": 0.0026610230561345816, "epoch": 39.451438848920866, "step": 43870 }, { "epoch": 39.451438848920866, "step": 43870, "torque_loss": 0.1934422105550766 }, { "epoch": 39.460431654676256, "grad_norm": 0.15189875662326813, "learning_rate": 4.04164094702984e-06, "loss": 0.02, "step": 43880 }, { "action_loss": 0.0011148704215884209, "epoch": 39.460431654676256, "step": 43880 }, { "epoch": 39.460431654676256, "step": 43880, "torque_loss": 0.13304246962070465 }, { "epoch": 39.469424460431654, "grad_norm": 0.054010193794965744, "learning_rate": 4.028626035786953e-06, "loss": 0.0161, "step": 43890 }, { "action_loss": 0.007334461901336908, "epoch": 39.469424460431654, "step": 43890 }, { "epoch": 39.469424460431654, "step": 43890, "torque_loss": 0.18421483039855957 }, { "epoch": 39.47841726618705, "grad_norm": 0.05498647317290306, "learning_rate": 4.0156312339848675e-06, "loss": 0.0163, "step": 43900 }, { "action_loss": 0.008033383637666702, "epoch": 39.47841726618705, "step": 43900 }, { "epoch": 39.47841726618705, "step": 43900, "torque_loss": 0.22317402064800262 }, { "epoch": 39.48741007194245, "grad_norm": 0.0976153016090393, "learning_rate": 4.00265654730797e-06, "loss": 0.0197, "step": 43910 }, { "action_loss": 0.0015953012043610215, "epoch": 39.48741007194245, "step": 43910 }, { "epoch": 39.48741007194245, "step": 43910, "torque_loss": 0.1661381870508194 }, { "epoch": 39.49640287769784, "grad_norm": 0.05054456368088722, "learning_rate": 3.989701981431809e-06, "loss": 0.015, "step": 43920 }, { "action_loss": 0.001620753319002688, "epoch": 39.49640287769784, "step": 43920 }, { "epoch": 39.49640287769784, "step": 43920, "torque_loss": 0.11082214117050171 }, { "epoch": 39.50539568345324, "grad_norm": 0.053519681096076965, "learning_rate": 3.976767542023169e-06, "loss": 0.0143, "step": 43930 }, { "action_loss": 0.001369109028019011, "epoch": 39.50539568345324, "step": 43930 }, { "epoch": 39.50539568345324, "step": 43930, "torque_loss": 0.15753225982189178 }, { "epoch": 39.514388489208635, "grad_norm": 0.0703023374080658, "learning_rate": 3.963853234740006e-06, "loss": 0.0171, "step": 43940 }, { "action_loss": 0.0010858335299417377, "epoch": 39.514388489208635, "step": 43940 }, { "epoch": 39.514388489208635, "step": 43940, "torque_loss": 0.1568060666322708 }, { "epoch": 39.523381294964025, "grad_norm": 0.09448989480733871, "learning_rate": 3.950959065231464e-06, "loss": 0.0196, "step": 43950 }, { "action_loss": 0.004417553078383207, "epoch": 39.523381294964025, "step": 43950 }, { "epoch": 39.523381294964025, "step": 43950, "torque_loss": 0.20116274058818817 }, { "epoch": 39.53237410071942, "grad_norm": 0.0571850910782814, "learning_rate": 3.938085039137917e-06, "loss": 0.0264, "step": 43960 }, { "action_loss": 0.0015125606441870332, "epoch": 39.53237410071942, "step": 43960 }, { "epoch": 39.53237410071942, "step": 43960, "torque_loss": 0.15785318613052368 }, { "epoch": 39.54136690647482, "grad_norm": 0.0463075265288353, "learning_rate": 3.925231162090881e-06, "loss": 0.0186, "step": 43970 }, { "action_loss": 0.0020482053514569998, "epoch": 39.54136690647482, "step": 43970 }, { "epoch": 39.54136690647482, "step": 43970, "torque_loss": 0.13240893185138702 }, { "epoch": 39.55035971223022, "grad_norm": 0.030954916030168533, "learning_rate": 3.912397439713084e-06, "loss": 0.016, "step": 43980 }, { "action_loss": 0.0017186993500217795, "epoch": 39.55035971223022, "step": 43980 }, { "epoch": 39.55035971223022, "step": 43980, "torque_loss": 0.16786639392375946 }, { "epoch": 39.55935251798561, "grad_norm": 0.10649067908525467, "learning_rate": 3.899583877618435e-06, "loss": 0.0184, "step": 43990 }, { "action_loss": 0.007743338588625193, "epoch": 39.55935251798561, "step": 43990 }, { "epoch": 39.55935251798561, "step": 43990, "torque_loss": 0.16244535148143768 }, { "epoch": 39.568345323741006, "grad_norm": 0.023966263979673386, "learning_rate": 3.8867904814120225e-06, "loss": 0.0177, "step": 44000 }, { "action_loss": 0.0019518430344760418, "epoch": 39.568345323741006, "step": 44000 }, { "epoch": 39.568345323741006, "step": 44000, "torque_loss": 0.1355932205915451 }, { "epoch": 39.577338129496404, "grad_norm": 0.07360388338565826, "learning_rate": 3.8740172566901044e-06, "loss": 0.0179, "step": 44010 }, { "action_loss": 0.0017113123321905732, "epoch": 39.577338129496404, "step": 44010 }, { "epoch": 39.577338129496404, "step": 44010, "torque_loss": 0.12792950868606567 }, { "epoch": 39.5863309352518, "grad_norm": 0.03407083451747894, "learning_rate": 3.861264209040133e-06, "loss": 0.0182, "step": 44020 }, { "action_loss": 0.0013723570154979825, "epoch": 39.5863309352518, "step": 44020 }, { "epoch": 39.5863309352518, "step": 44020, "torque_loss": 0.16261081397533417 }, { "epoch": 39.59532374100719, "grad_norm": 0.06424294412136078, "learning_rate": 3.848531344040718e-06, "loss": 0.0192, "step": 44030 }, { "action_loss": 0.0014316214947029948, "epoch": 39.59532374100719, "step": 44030 }, { "epoch": 39.59532374100719, "step": 44030, "torque_loss": 0.1284208744764328 }, { "epoch": 39.60431654676259, "grad_norm": 0.04112186282873154, "learning_rate": 3.835818667261642e-06, "loss": 0.0163, "step": 44040 }, { "action_loss": 0.005739850457757711, "epoch": 39.60431654676259, "step": 44040 }, { "epoch": 39.60431654676259, "step": 44040, "torque_loss": 0.1951051950454712 }, { "epoch": 39.61330935251799, "grad_norm": 0.15012089908123016, "learning_rate": 3.823126184263881e-06, "loss": 0.0194, "step": 44050 }, { "action_loss": 0.0024324816185981035, "epoch": 39.61330935251799, "step": 44050 }, { "epoch": 39.61330935251799, "step": 44050, "torque_loss": 0.1445113867521286 }, { "epoch": 39.62230215827338, "grad_norm": 0.038335151970386505, "learning_rate": 3.810453900599542e-06, "loss": 0.0166, "step": 44060 }, { "action_loss": 0.0006359452963806689, "epoch": 39.62230215827338, "step": 44060 }, { "epoch": 39.62230215827338, "step": 44060, "torque_loss": 0.10342053323984146 }, { "epoch": 39.631294964028775, "grad_norm": 0.0662231519818306, "learning_rate": 3.7978018218119137e-06, "loss": 0.017, "step": 44070 }, { "action_loss": 0.00220387801527977, "epoch": 39.631294964028775, "step": 44070 }, { "epoch": 39.631294964028775, "step": 44070, "torque_loss": 0.19889028370380402 }, { "epoch": 39.64028776978417, "grad_norm": 0.03292842581868172, "learning_rate": 3.7851699534354524e-06, "loss": 0.0194, "step": 44080 }, { "action_loss": 0.0005478335660882294, "epoch": 39.64028776978417, "step": 44080 }, { "epoch": 39.64028776978417, "step": 44080, "torque_loss": 0.14793364703655243 }, { "epoch": 39.64928057553957, "grad_norm": 0.08965746313333511, "learning_rate": 3.77255830099576e-06, "loss": 0.017, "step": 44090 }, { "action_loss": 0.0017973753856495023, "epoch": 39.64928057553957, "step": 44090 }, { "epoch": 39.64928057553957, "step": 44090, "torque_loss": 0.15050891041755676 }, { "epoch": 39.65827338129496, "grad_norm": 0.04319548234343529, "learning_rate": 3.759966870009607e-06, "loss": 0.0178, "step": 44100 }, { "action_loss": 0.002303807297721505, "epoch": 39.65827338129496, "step": 44100 }, { "epoch": 39.65827338129496, "step": 44100, "torque_loss": 0.19139324128627777 }, { "epoch": 39.66726618705036, "grad_norm": 0.04941382259130478, "learning_rate": 3.7473956659848987e-06, "loss": 0.0183, "step": 44110 }, { "action_loss": 0.001525206957012415, "epoch": 39.66726618705036, "step": 44110 }, { "epoch": 39.66726618705036, "step": 44110, "torque_loss": 0.1506592184305191 }, { "epoch": 39.67625899280576, "grad_norm": 0.0527416430413723, "learning_rate": 3.7348446944207305e-06, "loss": 0.015, "step": 44120 }, { "action_loss": 0.0005847452557645738, "epoch": 39.67625899280576, "step": 44120 }, { "epoch": 39.67625899280576, "step": 44120, "torque_loss": 0.099855937063694 }, { "epoch": 39.685251798561154, "grad_norm": 0.05197472497820854, "learning_rate": 3.7223139608073e-06, "loss": 0.0152, "step": 44130 }, { "action_loss": 0.0010514762252569199, "epoch": 39.685251798561154, "step": 44130 }, { "epoch": 39.685251798561154, "step": 44130, "torque_loss": 0.16574572026729584 }, { "epoch": 39.694244604316545, "grad_norm": 0.08615020662546158, "learning_rate": 3.7098034706259943e-06, "loss": 0.0189, "step": 44140 }, { "action_loss": 0.0005724687944166362, "epoch": 39.694244604316545, "step": 44140 }, { "epoch": 39.694244604316545, "step": 44140, "torque_loss": 0.13075442612171173 }, { "epoch": 39.70323741007194, "grad_norm": 0.03731941431760788, "learning_rate": 3.697313229349314e-06, "loss": 0.0184, "step": 44150 }, { "action_loss": 0.00045228670933283865, "epoch": 39.70323741007194, "step": 44150 }, { "epoch": 39.70323741007194, "step": 44150, "torque_loss": 0.13552965223789215 }, { "epoch": 39.71223021582734, "grad_norm": 0.050404731184244156, "learning_rate": 3.684843242440933e-06, "loss": 0.0169, "step": 44160 }, { "action_loss": 0.004115581978112459, "epoch": 39.71223021582734, "step": 44160 }, { "epoch": 39.71223021582734, "step": 44160, "torque_loss": 0.17179244756698608 }, { "epoch": 39.72122302158273, "grad_norm": 0.059925515204668045, "learning_rate": 3.672393515355621e-06, "loss": 0.0184, "step": 44170 }, { "action_loss": 0.0008848612778820097, "epoch": 39.72122302158273, "step": 44170 }, { "epoch": 39.72122302158273, "step": 44170, "torque_loss": 0.16507945954799652 }, { "epoch": 39.73021582733813, "grad_norm": 0.07223214209079742, "learning_rate": 3.659964053539333e-06, "loss": 0.0177, "step": 44180 }, { "action_loss": 0.0006589386612176895, "epoch": 39.73021582733813, "step": 44180 }, { "epoch": 39.73021582733813, "step": 44180, "torque_loss": 0.13468492031097412 }, { "epoch": 39.739208633093526, "grad_norm": 0.04475221037864685, "learning_rate": 3.647554862429131e-06, "loss": 0.0172, "step": 44190 }, { "action_loss": 0.00347028742544353, "epoch": 39.739208633093526, "step": 44190 }, { "epoch": 39.739208633093526, "step": 44190, "torque_loss": 0.19471687078475952 }, { "epoch": 39.74820143884892, "grad_norm": 0.10606010258197784, "learning_rate": 3.635165947453212e-06, "loss": 0.0178, "step": 44200 }, { "action_loss": 0.0027402909472584724, "epoch": 39.74820143884892, "step": 44200 }, { "epoch": 39.74820143884892, "step": 44200, "torque_loss": 0.16140900552272797 }, { "epoch": 39.757194244604314, "grad_norm": 0.06411716341972351, "learning_rate": 3.622797314030918e-06, "loss": 0.0179, "step": 44210 }, { "action_loss": 0.007177095394581556, "epoch": 39.757194244604314, "step": 44210 }, { "epoch": 39.757194244604314, "step": 44210, "torque_loss": 0.21833841502666473 }, { "epoch": 39.76618705035971, "grad_norm": 0.033046089112758636, "learning_rate": 3.610448967572705e-06, "loss": 0.0208, "step": 44220 }, { "action_loss": 0.0032425050158053637, "epoch": 39.76618705035971, "step": 44220 }, { "epoch": 39.76618705035971, "step": 44220, "torque_loss": 0.1987760066986084 }, { "epoch": 39.77517985611511, "grad_norm": 0.06172819435596466, "learning_rate": 3.5981209134801464e-06, "loss": 0.0183, "step": 44230 }, { "action_loss": 0.004376609344035387, "epoch": 39.77517985611511, "step": 44230 }, { "epoch": 39.77517985611511, "step": 44230, "torque_loss": 0.1007232666015625 }, { "epoch": 39.78417266187051, "grad_norm": 0.06228501349687576, "learning_rate": 3.585813157145973e-06, "loss": 0.016, "step": 44240 }, { "action_loss": 0.0009430335485376418, "epoch": 39.78417266187051, "step": 44240 }, { "epoch": 39.78417266187051, "step": 44240, "torque_loss": 0.14538244903087616 }, { "epoch": 39.7931654676259, "grad_norm": 0.053814228624105453, "learning_rate": 3.5735257039540114e-06, "loss": 0.0189, "step": 44250 }, { "action_loss": 0.007368437945842743, "epoch": 39.7931654676259, "step": 44250 }, { "epoch": 39.7931654676259, "step": 44250, "torque_loss": 0.19996969401836395 }, { "epoch": 39.802158273381295, "grad_norm": 0.04747457429766655, "learning_rate": 3.5612585592792014e-06, "loss": 0.0161, "step": 44260 }, { "action_loss": 0.002387766493484378, "epoch": 39.802158273381295, "step": 44260 }, { "epoch": 39.802158273381295, "step": 44260, "torque_loss": 0.1558082401752472 }, { "epoch": 39.81115107913669, "grad_norm": 0.030126653611660004, "learning_rate": 3.5490117284876055e-06, "loss": 0.016, "step": 44270 }, { "action_loss": 0.005655436310917139, "epoch": 39.81115107913669, "step": 44270 }, { "epoch": 39.81115107913669, "step": 44270, "torque_loss": 0.18635351955890656 }, { "epoch": 39.82014388489208, "grad_norm": 0.042566101998090744, "learning_rate": 3.5367852169364224e-06, "loss": 0.0189, "step": 44280 }, { "action_loss": 0.0014046753058210015, "epoch": 39.82014388489208, "step": 44280 }, { "epoch": 39.82014388489208, "step": 44280, "torque_loss": 0.1795816570520401 }, { "epoch": 39.82913669064748, "grad_norm": 0.05928388610482216, "learning_rate": 3.524579029973929e-06, "loss": 0.0176, "step": 44290 }, { "action_loss": 0.001953840022906661, "epoch": 39.82913669064748, "step": 44290 }, { "epoch": 39.82913669064748, "step": 44290, "torque_loss": 0.21578781306743622 }, { "epoch": 39.83812949640288, "grad_norm": 0.04175931215286255, "learning_rate": 3.5123931729395267e-06, "loss": 0.0203, "step": 44300 }, { "action_loss": 0.005344205070286989, "epoch": 39.83812949640288, "step": 44300 }, { "epoch": 39.83812949640288, "step": 44300, "torque_loss": 0.22133545577526093 }, { "epoch": 39.847122302158276, "grad_norm": 0.0466756634414196, "learning_rate": 3.500227651163729e-06, "loss": 0.0167, "step": 44310 }, { "action_loss": 0.0038555841892957687, "epoch": 39.847122302158276, "step": 44310 }, { "epoch": 39.847122302158276, "step": 44310, "torque_loss": 0.1865941286087036 }, { "epoch": 39.856115107913666, "grad_norm": 0.046361710876226425, "learning_rate": 3.488082469968146e-06, "loss": 0.0195, "step": 44320 }, { "action_loss": 0.002213576575741172, "epoch": 39.856115107913666, "step": 44320 }, { "epoch": 39.856115107913666, "step": 44320, "torque_loss": 0.19541902840137482 }, { "epoch": 39.865107913669064, "grad_norm": 0.03657254949212074, "learning_rate": 3.475957634665489e-06, "loss": 0.0141, "step": 44330 }, { "action_loss": 0.0019613229669630527, "epoch": 39.865107913669064, "step": 44330 }, { "epoch": 39.865107913669064, "step": 44330, "torque_loss": 0.19279640913009644 }, { "epoch": 39.87410071942446, "grad_norm": 0.04695335775613785, "learning_rate": 3.463853150559582e-06, "loss": 0.017, "step": 44340 }, { "action_loss": 0.0008924933499656618, "epoch": 39.87410071942446, "step": 44340 }, { "epoch": 39.87410071942446, "step": 44340, "torque_loss": 0.17348627746105194 }, { "epoch": 39.88309352517986, "grad_norm": 0.10763643682003021, "learning_rate": 3.451769022945334e-06, "loss": 0.0192, "step": 44350 }, { "action_loss": 0.0010131866438314319, "epoch": 39.88309352517986, "step": 44350 }, { "epoch": 39.88309352517986, "step": 44350, "torque_loss": 0.21534425020217896 }, { "epoch": 39.89208633093525, "grad_norm": 0.15003468096256256, "learning_rate": 3.4397052571087385e-06, "loss": 0.0194, "step": 44360 }, { "action_loss": 0.006287405267357826, "epoch": 39.89208633093525, "step": 44360 }, { "epoch": 39.89208633093525, "step": 44360, "torque_loss": 0.21880953013896942 }, { "epoch": 39.90107913669065, "grad_norm": 0.12204865366220474, "learning_rate": 3.427661858326925e-06, "loss": 0.0208, "step": 44370 }, { "action_loss": 0.0008201432065106928, "epoch": 39.90107913669065, "step": 44370 }, { "epoch": 39.90107913669065, "step": 44370, "torque_loss": 0.12102118879556656 }, { "epoch": 39.910071942446045, "grad_norm": 0.03434750437736511, "learning_rate": 3.4156388318680566e-06, "loss": 0.0145, "step": 44380 }, { "action_loss": 0.00122206995729357, "epoch": 39.910071942446045, "step": 44380 }, { "epoch": 39.910071942446045, "step": 44380, "torque_loss": 0.115489162504673 }, { "epoch": 39.919064748201436, "grad_norm": 0.04022282361984253, "learning_rate": 3.4036361829914322e-06, "loss": 0.0182, "step": 44390 }, { "action_loss": 0.0011347898980602622, "epoch": 39.919064748201436, "step": 44390 }, { "epoch": 39.919064748201436, "step": 44390, "torque_loss": 0.13613532483577728 }, { "epoch": 39.92805755395683, "grad_norm": 0.04931551218032837, "learning_rate": 3.391653916947407e-06, "loss": 0.0171, "step": 44400 }, { "action_loss": 0.0003969090466853231, "epoch": 39.92805755395683, "step": 44400 }, { "epoch": 39.92805755395683, "step": 44400, "torque_loss": 0.1314137727022171 }, { "epoch": 39.93705035971223, "grad_norm": 0.04117388650774956, "learning_rate": 3.3796920389774444e-06, "loss": 0.0153, "step": 44410 }, { "action_loss": 0.0012747684959322214, "epoch": 39.93705035971223, "step": 44410 }, { "epoch": 39.93705035971223, "step": 44410, "torque_loss": 0.15617822110652924 }, { "epoch": 39.94604316546763, "grad_norm": 0.04495853930711746, "learning_rate": 3.3677505543140643e-06, "loss": 0.0173, "step": 44420 }, { "action_loss": 0.002715957583859563, "epoch": 39.94604316546763, "step": 44420 }, { "epoch": 39.94604316546763, "step": 44420, "torque_loss": 0.160085067152977 }, { "epoch": 39.95503597122302, "grad_norm": 0.02982201986014843, "learning_rate": 3.355829468180882e-06, "loss": 0.0192, "step": 44430 }, { "action_loss": 0.001248729880899191, "epoch": 39.95503597122302, "step": 44430 }, { "epoch": 39.95503597122302, "step": 44430, "torque_loss": 0.14148344099521637 }, { "epoch": 39.96402877697842, "grad_norm": 0.04472898319363594, "learning_rate": 3.343928785792588e-06, "loss": 0.0192, "step": 44440 }, { "action_loss": 0.0011714557185769081, "epoch": 39.96402877697842, "step": 44440 }, { "epoch": 39.96402877697842, "step": 44440, "torque_loss": 0.12212011963129044 }, { "epoch": 39.973021582733814, "grad_norm": 0.0699080303311348, "learning_rate": 3.3320485123549405e-06, "loss": 0.018, "step": 44450 }, { "action_loss": 0.00616893544793129, "epoch": 39.973021582733814, "step": 44450 }, { "epoch": 39.973021582733814, "step": 44450, "torque_loss": 0.17010442912578583 }, { "epoch": 39.98201438848921, "grad_norm": 0.11357060074806213, "learning_rate": 3.3201886530647707e-06, "loss": 0.0175, "step": 44460 }, { "action_loss": 0.000997263821773231, "epoch": 39.98201438848921, "step": 44460 }, { "epoch": 39.98201438848921, "step": 44460, "torque_loss": 0.19333051145076752 }, { "epoch": 39.9910071942446, "grad_norm": 0.062299374490976334, "learning_rate": 3.3083492131100015e-06, "loss": 0.016, "step": 44470 }, { "action_loss": 0.004044957924634218, "epoch": 39.9910071942446, "step": 44470 }, { "epoch": 39.9910071942446, "step": 44470, "torque_loss": 0.17810791730880737 }, { "epoch": 40.0, "grad_norm": 0.08093071728944778, "learning_rate": 3.2965301976695896e-06, "loss": 0.0197, "step": 44480 }, { "action_loss": 0.0009915768168866634, "epoch": 40.0, "step": 44480 }, { "epoch": 40.0, "step": 44480, "torque_loss": 0.15584833920001984 }, { "epoch": 40.0089928057554, "grad_norm": 0.05866309627890587, "learning_rate": 3.284731611913577e-06, "loss": 0.018, "step": 44490 }, { "action_loss": 0.0007351686363108456, "epoch": 40.0089928057554, "step": 44490 }, { "epoch": 40.0089928057554, "step": 44490, "torque_loss": 0.11539775133132935 }, { "epoch": 40.01798561151079, "grad_norm": 0.06049003452062607, "learning_rate": 3.27295346100307e-06, "loss": 0.0191, "step": 44500 }, { "action_loss": 0.0025369746144860983, "epoch": 40.01798561151079, "step": 44500 }, { "epoch": 40.01798561151079, "step": 44500, "torque_loss": 0.16331566870212555 }, { "epoch": 40.026978417266186, "grad_norm": 0.14884941279888153, "learning_rate": 3.2611957500902347e-06, "loss": 0.0226, "step": 44510 }, { "action_loss": 0.00045271121780388057, "epoch": 40.026978417266186, "step": 44510 }, { "epoch": 40.026978417266186, "step": 44510, "torque_loss": 0.1072947308421135 }, { "epoch": 40.03597122302158, "grad_norm": 0.04254307597875595, "learning_rate": 3.2494584843182796e-06, "loss": 0.0148, "step": 44520 }, { "action_loss": 0.0020905036944895983, "epoch": 40.03597122302158, "step": 44520 }, { "epoch": 40.03597122302158, "step": 44520, "torque_loss": 0.16505546867847443 }, { "epoch": 40.04496402877698, "grad_norm": 0.14128737151622772, "learning_rate": 3.2377416688214978e-06, "loss": 0.0175, "step": 44530 }, { "action_loss": 0.000856280792504549, "epoch": 40.04496402877698, "step": 44530 }, { "epoch": 40.04496402877698, "step": 44530, "torque_loss": 0.1378452628850937 }, { "epoch": 40.05395683453237, "grad_norm": 0.056241847574710846, "learning_rate": 3.2260453087252175e-06, "loss": 0.0164, "step": 44540 }, { "action_loss": 0.0038554652128368616, "epoch": 40.05395683453237, "step": 44540 }, { "epoch": 40.05395683453237, "step": 44540, "torque_loss": 0.2432764768600464 }, { "epoch": 40.06294964028777, "grad_norm": 0.05273684486746788, "learning_rate": 3.2143694091458175e-06, "loss": 0.0191, "step": 44550 }, { "action_loss": 0.002176050329580903, "epoch": 40.06294964028777, "step": 44550 }, { "epoch": 40.06294964028777, "step": 44550, "torque_loss": 0.12870670855045319 }, { "epoch": 40.07194244604317, "grad_norm": 0.059277161955833435, "learning_rate": 3.2027139751907466e-06, "loss": 0.0145, "step": 44560 }, { "action_loss": 0.0012505956692621112, "epoch": 40.07194244604317, "step": 44560 }, { "epoch": 40.07194244604317, "step": 44560, "torque_loss": 0.10893835872411728 }, { "epoch": 40.080935251798564, "grad_norm": 0.07225599884986877, "learning_rate": 3.191079011958481e-06, "loss": 0.0165, "step": 44570 }, { "action_loss": 0.0005420779925771058, "epoch": 40.080935251798564, "step": 44570 }, { "epoch": 40.080935251798564, "step": 44570, "torque_loss": 0.10582629591226578 }, { "epoch": 40.089928057553955, "grad_norm": 0.03028920292854309, "learning_rate": 3.1794645245385337e-06, "loss": 0.0153, "step": 44580 }, { "action_loss": 0.00340887438505888, "epoch": 40.089928057553955, "step": 44580 }, { "epoch": 40.089928057553955, "step": 44580, "torque_loss": 0.20491480827331543 }, { "epoch": 40.09892086330935, "grad_norm": 0.055685173720121384, "learning_rate": 3.1678705180115063e-06, "loss": 0.0167, "step": 44590 }, { "action_loss": 0.0009248948190361261, "epoch": 40.09892086330935, "step": 44590 }, { "epoch": 40.09892086330935, "step": 44590, "torque_loss": 0.1238255575299263 }, { "epoch": 40.10791366906475, "grad_norm": 0.09785160422325134, "learning_rate": 3.1562969974489933e-06, "loss": 0.0167, "step": 44600 }, { "action_loss": 0.002068291651085019, "epoch": 40.10791366906475, "step": 44600 }, { "epoch": 40.10791366906475, "step": 44600, "torque_loss": 0.2046235054731369 }, { "epoch": 40.11690647482014, "grad_norm": 0.030459605157375336, "learning_rate": 3.144743967913638e-06, "loss": 0.0173, "step": 44610 }, { "action_loss": 0.00221127737313509, "epoch": 40.11690647482014, "step": 44610 }, { "epoch": 40.11690647482014, "step": 44610, "torque_loss": 0.13900452852249146 }, { "epoch": 40.12589928057554, "grad_norm": 0.03993735462427139, "learning_rate": 3.1332114344591436e-06, "loss": 0.0176, "step": 44620 }, { "action_loss": 0.0024986565113067627, "epoch": 40.12589928057554, "step": 44620 }, { "epoch": 40.12589928057554, "step": 44620, "torque_loss": 0.12477871775627136 }, { "epoch": 40.134892086330936, "grad_norm": 0.06606079638004303, "learning_rate": 3.1216994021302235e-06, "loss": 0.0189, "step": 44630 }, { "action_loss": 0.0041197799146175385, "epoch": 40.134892086330936, "step": 44630 }, { "epoch": 40.134892086330936, "step": 44630, "torque_loss": 0.1798877716064453 }, { "epoch": 40.143884892086334, "grad_norm": 0.09939391911029816, "learning_rate": 3.1102078759626396e-06, "loss": 0.0212, "step": 44640 }, { "action_loss": 0.0016411951510235667, "epoch": 40.143884892086334, "step": 44640 }, { "epoch": 40.143884892086334, "step": 44640, "torque_loss": 0.15364688634872437 }, { "epoch": 40.152877697841724, "grad_norm": 0.13608282804489136, "learning_rate": 3.098736860983159e-06, "loss": 0.0175, "step": 44650 }, { "action_loss": 0.0008834800100885332, "epoch": 40.152877697841724, "step": 44650 }, { "epoch": 40.152877697841724, "step": 44650, "torque_loss": 0.13175782561302185 }, { "epoch": 40.16187050359712, "grad_norm": 0.09642522037029266, "learning_rate": 3.087286362209618e-06, "loss": 0.0186, "step": 44660 }, { "action_loss": 0.0008799534407444298, "epoch": 40.16187050359712, "step": 44660 }, { "epoch": 40.16187050359712, "step": 44660, "torque_loss": 0.18098598718643188 }, { "epoch": 40.17086330935252, "grad_norm": 0.09784408658742905, "learning_rate": 3.0758563846508383e-06, "loss": 0.0183, "step": 44670 }, { "action_loss": 0.0017474511405453086, "epoch": 40.17086330935252, "step": 44670 }, { "epoch": 40.17086330935252, "step": 44670, "torque_loss": 0.17897774279117584 }, { "epoch": 40.17985611510792, "grad_norm": 0.06754494458436966, "learning_rate": 3.0644469333066748e-06, "loss": 0.0179, "step": 44680 }, { "action_loss": 0.004412876907736063, "epoch": 40.17985611510792, "step": 44680 }, { "epoch": 40.17985611510792, "step": 44680, "torque_loss": 0.16341352462768555 }, { "epoch": 40.18884892086331, "grad_norm": 0.03970944881439209, "learning_rate": 3.0530580131680307e-06, "loss": 0.0185, "step": 44690 }, { "action_loss": 0.0005347169935703278, "epoch": 40.18884892086331, "step": 44690 }, { "epoch": 40.18884892086331, "step": 44690, "torque_loss": 0.10396354645490646 }, { "epoch": 40.197841726618705, "grad_norm": 0.06000059098005295, "learning_rate": 3.0416896292167875e-06, "loss": 0.0152, "step": 44700 }, { "action_loss": 0.0005234706914052367, "epoch": 40.197841726618705, "step": 44700 }, { "epoch": 40.197841726618705, "step": 44700, "torque_loss": 0.10843700170516968 }, { "epoch": 40.2068345323741, "grad_norm": 0.11019933223724365, "learning_rate": 3.030341786425861e-06, "loss": 0.0154, "step": 44710 }, { "action_loss": 0.002218622015789151, "epoch": 40.2068345323741, "step": 44710 }, { "epoch": 40.2068345323741, "step": 44710, "torque_loss": 0.16652028262615204 }, { "epoch": 40.21582733812949, "grad_norm": 0.02413218654692173, "learning_rate": 3.0190144897591975e-06, "loss": 0.017, "step": 44720 }, { "action_loss": 0.001972752623260021, "epoch": 40.21582733812949, "step": 44720 }, { "epoch": 40.21582733812949, "step": 44720, "torque_loss": 0.18362893164157867 }, { "epoch": 40.22482014388489, "grad_norm": 0.11671900749206543, "learning_rate": 3.0077077441717395e-06, "loss": 0.0186, "step": 44730 }, { "action_loss": 0.010712094604969025, "epoch": 40.22482014388489, "step": 44730 }, { "epoch": 40.22482014388489, "step": 44730, "torque_loss": 0.19557583332061768 }, { "epoch": 40.23381294964029, "grad_norm": 0.055931538343429565, "learning_rate": 2.9964215546094244e-06, "loss": 0.0195, "step": 44740 }, { "action_loss": 0.005792643874883652, "epoch": 40.23381294964029, "step": 44740 }, { "epoch": 40.23381294964029, "step": 44740, "torque_loss": 0.14508384466171265 }, { "epoch": 40.242805755395686, "grad_norm": 0.04492677375674248, "learning_rate": 2.9851559260092367e-06, "loss": 0.0218, "step": 44750 }, { "action_loss": 0.0047550261951982975, "epoch": 40.242805755395686, "step": 44750 }, { "epoch": 40.242805755395686, "step": 44750, "torque_loss": 0.1944429725408554 }, { "epoch": 40.25179856115108, "grad_norm": 0.0420168936252594, "learning_rate": 2.973910863299134e-06, "loss": 0.0164, "step": 44760 }, { "action_loss": 0.004731254186481237, "epoch": 40.25179856115108, "step": 44760 }, { "epoch": 40.25179856115108, "step": 44760, "torque_loss": 0.16743701696395874 }, { "epoch": 40.260791366906474, "grad_norm": 0.0350896455347538, "learning_rate": 2.962686371398088e-06, "loss": 0.0184, "step": 44770 }, { "action_loss": 0.001661851885728538, "epoch": 40.260791366906474, "step": 44770 }, { "epoch": 40.260791366906474, "step": 44770, "torque_loss": 0.14244671165943146 }, { "epoch": 40.26978417266187, "grad_norm": 0.031932320445775986, "learning_rate": 2.95148245521607e-06, "loss": 0.0184, "step": 44780 }, { "action_loss": 0.0007857335731387138, "epoch": 40.26978417266187, "step": 44780 }, { "epoch": 40.26978417266187, "step": 44780, "torque_loss": 0.16185222566127777 }, { "epoch": 40.27877697841727, "grad_norm": 0.0432160347700119, "learning_rate": 2.9402991196540665e-06, "loss": 0.0195, "step": 44790 }, { "action_loss": 0.0005125013994984329, "epoch": 40.27877697841727, "step": 44790 }, { "epoch": 40.27877697841727, "step": 44790, "torque_loss": 0.12128376960754395 }, { "epoch": 40.28776978417266, "grad_norm": 0.05489025264978409, "learning_rate": 2.929136369604041e-06, "loss": 0.0185, "step": 44800 }, { "action_loss": 0.0006014798418618739, "epoch": 40.28776978417266, "step": 44800 }, { "epoch": 40.28776978417266, "step": 44800, "torque_loss": 0.10062748193740845 }, { "epoch": 40.29676258992806, "grad_norm": 0.053421881049871445, "learning_rate": 2.9179942099489487e-06, "loss": 0.0195, "step": 44810 }, { "action_loss": 0.0016852751141414046, "epoch": 40.29676258992806, "step": 44810 }, { "epoch": 40.29676258992806, "step": 44810, "torque_loss": 0.1398448497056961 }, { "epoch": 40.305755395683455, "grad_norm": 0.07851473242044449, "learning_rate": 2.9068726455627694e-06, "loss": 0.0207, "step": 44820 }, { "action_loss": 0.0007749342476017773, "epoch": 40.305755395683455, "step": 44820 }, { "epoch": 40.305755395683455, "step": 44820, "torque_loss": 0.1368171125650406 }, { "epoch": 40.314748201438846, "grad_norm": 0.12788952887058258, "learning_rate": 2.8957716813104384e-06, "loss": 0.0179, "step": 44830 }, { "action_loss": 0.0016527032712474465, "epoch": 40.314748201438846, "step": 44830 }, { "epoch": 40.314748201438846, "step": 44830, "torque_loss": 0.1890193670988083 }, { "epoch": 40.32374100719424, "grad_norm": 0.053863611072301865, "learning_rate": 2.8846913220479e-06, "loss": 0.0174, "step": 44840 }, { "action_loss": 0.0015477630076929927, "epoch": 40.32374100719424, "step": 44840 }, { "epoch": 40.32374100719424, "step": 44840, "torque_loss": 0.1535254567861557 }, { "epoch": 40.33273381294964, "grad_norm": 0.03785260394215584, "learning_rate": 2.873631572622082e-06, "loss": 0.0174, "step": 44850 }, { "action_loss": 0.0026511510368436575, "epoch": 40.33273381294964, "step": 44850 }, { "epoch": 40.33273381294964, "step": 44850, "torque_loss": 0.1349181979894638 }, { "epoch": 40.34172661870504, "grad_norm": 0.028245359659194946, "learning_rate": 2.862592437870881e-06, "loss": 0.0178, "step": 44860 }, { "action_loss": 0.000650558911729604, "epoch": 40.34172661870504, "step": 44860 }, { "epoch": 40.34172661870504, "step": 44860, "torque_loss": 0.16778302192687988 }, { "epoch": 40.35071942446043, "grad_norm": 0.10998382419347763, "learning_rate": 2.8515739226232075e-06, "loss": 0.0177, "step": 44870 }, { "action_loss": 0.0010101619409397244, "epoch": 40.35071942446043, "step": 44870 }, { "epoch": 40.35071942446043, "step": 44870, "torque_loss": 0.14620983600616455 }, { "epoch": 40.35971223021583, "grad_norm": 0.038670606911182404, "learning_rate": 2.840576031698927e-06, "loss": 0.0136, "step": 44880 }, { "action_loss": 0.003919862676411867, "epoch": 40.35971223021583, "step": 44880 }, { "epoch": 40.35971223021583, "step": 44880, "torque_loss": 0.15822143852710724 }, { "epoch": 40.368705035971225, "grad_norm": 0.04212317243218422, "learning_rate": 2.8295987699088923e-06, "loss": 0.0264, "step": 44890 }, { "action_loss": 0.0015465611359104514, "epoch": 40.368705035971225, "step": 44890 }, { "epoch": 40.368705035971225, "step": 44890, "torque_loss": 0.20927058160305023 }, { "epoch": 40.37769784172662, "grad_norm": 0.0782947689294815, "learning_rate": 2.8186421420549227e-06, "loss": 0.0227, "step": 44900 }, { "action_loss": 0.002973083406686783, "epoch": 40.37769784172662, "step": 44900 }, { "epoch": 40.37769784172662, "step": 44900, "torque_loss": 0.1549772024154663 }, { "epoch": 40.38669064748201, "grad_norm": 0.039490390568971634, "learning_rate": 2.8077061529298343e-06, "loss": 0.0188, "step": 44910 }, { "action_loss": 0.0006783371209166944, "epoch": 40.38669064748201, "step": 44910 }, { "epoch": 40.38669064748201, "step": 44910, "torque_loss": 0.16121433675289154 }, { "epoch": 40.39568345323741, "grad_norm": 0.05836275964975357, "learning_rate": 2.79679080731739e-06, "loss": 0.0195, "step": 44920 }, { "action_loss": 0.0019029452232643962, "epoch": 40.39568345323741, "step": 44920 }, { "epoch": 40.39568345323741, "step": 44920, "torque_loss": 0.1315392106771469 }, { "epoch": 40.40467625899281, "grad_norm": 0.06265176832675934, "learning_rate": 2.7858961099923254e-06, "loss": 0.0159, "step": 44930 }, { "action_loss": 0.0008278791210614145, "epoch": 40.40467625899281, "step": 44930 }, { "epoch": 40.40467625899281, "step": 44930, "torque_loss": 0.14392395317554474 }, { "epoch": 40.4136690647482, "grad_norm": 0.04210640490055084, "learning_rate": 2.7750220657203727e-06, "loss": 0.0174, "step": 44940 }, { "action_loss": 0.0009340645628981292, "epoch": 40.4136690647482, "step": 44940 }, { "epoch": 40.4136690647482, "step": 44940, "torque_loss": 0.12245085090398788 }, { "epoch": 40.422661870503596, "grad_norm": 0.03519757464528084, "learning_rate": 2.764168679258189e-06, "loss": 0.0178, "step": 44950 }, { "action_loss": 0.0010542654199525714, "epoch": 40.422661870503596, "step": 44950 }, { "epoch": 40.422661870503596, "step": 44950, "torque_loss": 0.2012682408094406 }, { "epoch": 40.431654676258994, "grad_norm": 0.06721373647451401, "learning_rate": 2.7533359553534254e-06, "loss": 0.0168, "step": 44960 }, { "action_loss": 0.0005302878562361002, "epoch": 40.431654676258994, "step": 44960 }, { "epoch": 40.431654676258994, "step": 44960, "torque_loss": 0.11894140392541885 }, { "epoch": 40.44064748201439, "grad_norm": 0.04374073073267937, "learning_rate": 2.7425238987446754e-06, "loss": 0.0169, "step": 44970 }, { "action_loss": 0.0007580690435133874, "epoch": 40.44064748201439, "step": 44970 }, { "epoch": 40.44064748201439, "step": 44970, "torque_loss": 0.08558335900306702 }, { "epoch": 40.44964028776978, "grad_norm": 0.10397753864526749, "learning_rate": 2.731732514161506e-06, "loss": 0.0231, "step": 44980 }, { "action_loss": 0.004929992835968733, "epoch": 40.44964028776978, "step": 44980 }, { "epoch": 40.44964028776978, "step": 44980, "torque_loss": 0.1503092497587204 }, { "epoch": 40.45863309352518, "grad_norm": 0.06036553531885147, "learning_rate": 2.720961806324435e-06, "loss": 0.018, "step": 44990 }, { "action_loss": 0.0005715120933018625, "epoch": 40.45863309352518, "step": 44990 }, { "epoch": 40.45863309352518, "step": 44990, "torque_loss": 0.11023926734924316 }, { "epoch": 40.46762589928058, "grad_norm": 0.03747275099158287, "learning_rate": 2.7102117799449266e-06, "loss": 0.0173, "step": 45000 }, { "action_loss": 0.0008999897981993854, "epoch": 40.46762589928058, "step": 45000 }, { "epoch": 40.46762589928058, "step": 45000, "torque_loss": 0.10747257620096207 }, { "epoch": 40.476618705035975, "grad_norm": 0.08223747462034225, "learning_rate": 2.6994824397254258e-06, "loss": 0.0157, "step": 45010 }, { "action_loss": 0.0021790212485939264, "epoch": 40.476618705035975, "step": 45010 }, { "epoch": 40.476618705035975, "step": 45010, "torque_loss": 0.19574613869190216 }, { "epoch": 40.485611510791365, "grad_norm": 0.09797104448080063, "learning_rate": 2.688773790359295e-06, "loss": 0.0188, "step": 45020 }, { "action_loss": 0.0007654931396245956, "epoch": 40.485611510791365, "step": 45020 }, { "epoch": 40.485611510791365, "step": 45020, "torque_loss": 0.12471884489059448 }, { "epoch": 40.49460431654676, "grad_norm": 0.06767489016056061, "learning_rate": 2.678085836530858e-06, "loss": 0.0186, "step": 45030 }, { "action_loss": 0.0030222341883927584, "epoch": 40.49460431654676, "step": 45030 }, { "epoch": 40.49460431654676, "step": 45030, "torque_loss": 0.17917799949645996 }, { "epoch": 40.50359712230216, "grad_norm": 0.03459380194544792, "learning_rate": 2.6674185829154096e-06, "loss": 0.0163, "step": 45040 }, { "action_loss": 0.003140262560918927, "epoch": 40.50359712230216, "step": 45040 }, { "epoch": 40.50359712230216, "step": 45040, "torque_loss": 0.247935488820076 }, { "epoch": 40.51258992805755, "grad_norm": 0.04401544854044914, "learning_rate": 2.6567720341791556e-06, "loss": 0.0199, "step": 45050 }, { "action_loss": 0.0012926219496876001, "epoch": 40.51258992805755, "step": 45050 }, { "epoch": 40.51258992805755, "step": 45050, "torque_loss": 0.13548649847507477 }, { "epoch": 40.52158273381295, "grad_norm": 0.034875426441431046, "learning_rate": 2.646146194979249e-06, "loss": 0.0167, "step": 45060 }, { "action_loss": 0.0010384079068899155, "epoch": 40.52158273381295, "step": 45060 }, { "epoch": 40.52158273381295, "step": 45060, "torque_loss": 0.11100991815328598 }, { "epoch": 40.530575539568346, "grad_norm": 0.051521625369787216, "learning_rate": 2.6355410699638094e-06, "loss": 0.0166, "step": 45070 }, { "action_loss": 0.0009226088877767324, "epoch": 40.530575539568346, "step": 45070 }, { "epoch": 40.530575539568346, "step": 45070, "torque_loss": 0.15452082455158234 }, { "epoch": 40.539568345323744, "grad_norm": 0.12612341344356537, "learning_rate": 2.6249566637718714e-06, "loss": 0.0168, "step": 45080 }, { "action_loss": 0.0017418060451745987, "epoch": 40.539568345323744, "step": 45080 }, { "epoch": 40.539568345323744, "step": 45080, "torque_loss": 0.10851436853408813 }, { "epoch": 40.548561151079134, "grad_norm": 0.03563077375292778, "learning_rate": 2.614392981033409e-06, "loss": 0.0172, "step": 45090 }, { "action_loss": 0.001133995014242828, "epoch": 40.548561151079134, "step": 45090 }, { "epoch": 40.548561151079134, "step": 45090, "torque_loss": 0.13585233688354492 }, { "epoch": 40.55755395683453, "grad_norm": 0.0358416922390461, "learning_rate": 2.603850026369348e-06, "loss": 0.0158, "step": 45100 }, { "action_loss": 0.0010711923241615295, "epoch": 40.55755395683453, "step": 45100 }, { "epoch": 40.55755395683453, "step": 45100, "torque_loss": 0.16481684148311615 }, { "epoch": 40.56654676258993, "grad_norm": 0.12105964124202728, "learning_rate": 2.593327804391521e-06, "loss": 0.0169, "step": 45110 }, { "action_loss": 0.0013169239973649383, "epoch": 40.56654676258993, "step": 45110 }, { "epoch": 40.56654676258993, "step": 45110, "torque_loss": 0.19250178337097168 }, { "epoch": 40.57553956834532, "grad_norm": 0.04623909667134285, "learning_rate": 2.5828263197027247e-06, "loss": 0.0163, "step": 45120 }, { "action_loss": 0.0009452486992813647, "epoch": 40.57553956834532, "step": 45120 }, { "epoch": 40.57553956834532, "step": 45120, "torque_loss": 0.12845540046691895 }, { "epoch": 40.58453237410072, "grad_norm": 0.1054903194308281, "learning_rate": 2.57234557689664e-06, "loss": 0.0166, "step": 45130 }, { "action_loss": 0.0013933294685557485, "epoch": 40.58453237410072, "step": 45130 }, { "epoch": 40.58453237410072, "step": 45130, "torque_loss": 0.12394380569458008 }, { "epoch": 40.593525179856115, "grad_norm": 0.11834952235221863, "learning_rate": 2.5618855805579213e-06, "loss": 0.0205, "step": 45140 }, { "action_loss": 0.0033924842718988657, "epoch": 40.593525179856115, "step": 45140 }, { "epoch": 40.593525179856115, "step": 45140, "torque_loss": 0.13580207526683807 }, { "epoch": 40.60251798561151, "grad_norm": 0.0793999657034874, "learning_rate": 2.5514463352621197e-06, "loss": 0.0157, "step": 45150 }, { "action_loss": 0.0006159344920888543, "epoch": 40.60251798561151, "step": 45150 }, { "epoch": 40.60251798561151, "step": 45150, "torque_loss": 0.1202593445777893 }, { "epoch": 40.611510791366904, "grad_norm": 0.03708802908658981, "learning_rate": 2.541027845575711e-06, "loss": 0.0168, "step": 45160 }, { "action_loss": 0.0017790874699130654, "epoch": 40.611510791366904, "step": 45160 }, { "epoch": 40.611510791366904, "step": 45160, "torque_loss": 0.13989050686359406 }, { "epoch": 40.6205035971223, "grad_norm": 0.02832549624145031, "learning_rate": 2.530630116056104e-06, "loss": 0.0157, "step": 45170 }, { "action_loss": 0.0004959315410815179, "epoch": 40.6205035971223, "step": 45170 }, { "epoch": 40.6205035971223, "step": 45170, "torque_loss": 0.0955076813697815 }, { "epoch": 40.6294964028777, "grad_norm": 0.0634092465043068, "learning_rate": 2.5202531512516114e-06, "loss": 0.016, "step": 45180 }, { "action_loss": 0.0005399350775405765, "epoch": 40.6294964028777, "step": 45180 }, { "epoch": 40.6294964028777, "step": 45180, "torque_loss": 0.0918605700135231 }, { "epoch": 40.638489208633096, "grad_norm": 0.05990077927708626, "learning_rate": 2.509896955701474e-06, "loss": 0.0155, "step": 45190 }, { "action_loss": 0.011347775347530842, "epoch": 40.638489208633096, "step": 45190 }, { "epoch": 40.638489208633096, "step": 45190, "torque_loss": 0.2616772949695587 }, { "epoch": 40.64748201438849, "grad_norm": 0.03440946713089943, "learning_rate": 2.4995615339358404e-06, "loss": 0.019, "step": 45200 }, { "action_loss": 0.0012890971265733242, "epoch": 40.64748201438849, "step": 45200 }, { "epoch": 40.64748201438849, "step": 45200, "torque_loss": 0.12863998115062714 }, { "epoch": 40.656474820143885, "grad_norm": 0.06806843727827072, "learning_rate": 2.489246890475788e-06, "loss": 0.0163, "step": 45210 }, { "action_loss": 0.0018028700724244118, "epoch": 40.656474820143885, "step": 45210 }, { "epoch": 40.656474820143885, "step": 45210, "torque_loss": 0.13881927728652954 }, { "epoch": 40.66546762589928, "grad_norm": 0.1237320527434349, "learning_rate": 2.4789530298332696e-06, "loss": 0.0192, "step": 45220 }, { "action_loss": 0.000803604896645993, "epoch": 40.66546762589928, "step": 45220 }, { "epoch": 40.66546762589928, "step": 45220, "torque_loss": 0.16354985535144806 }, { "epoch": 40.67446043165468, "grad_norm": 0.1252484917640686, "learning_rate": 2.4686799565111885e-06, "loss": 0.0181, "step": 45230 }, { "action_loss": 0.001761385821737349, "epoch": 40.67446043165468, "step": 45230 }, { "epoch": 40.67446043165468, "step": 45230, "torque_loss": 0.19322554767131805 }, { "epoch": 40.68345323741007, "grad_norm": 0.04021700099110603, "learning_rate": 2.4584276750033332e-06, "loss": 0.017, "step": 45240 }, { "action_loss": 0.003746550530195236, "epoch": 40.68345323741007, "step": 45240 }, { "epoch": 40.68345323741007, "step": 45240, "torque_loss": 0.18246233463287354 }, { "epoch": 40.69244604316547, "grad_norm": 0.11822677403688431, "learning_rate": 2.448196189794383e-06, "loss": 0.0168, "step": 45250 }, { "action_loss": 0.0014499998651444912, "epoch": 40.69244604316547, "step": 45250 }, { "epoch": 40.69244604316547, "step": 45250, "torque_loss": 0.20088203251361847 }, { "epoch": 40.701438848920866, "grad_norm": 0.042929522693157196, "learning_rate": 2.437985505359963e-06, "loss": 0.0148, "step": 45260 }, { "action_loss": 0.0060035898350179195, "epoch": 40.701438848920866, "step": 45260 }, { "epoch": 40.701438848920866, "step": 45260, "torque_loss": 0.2017371505498886 }, { "epoch": 40.710431654676256, "grad_norm": 0.04112852364778519, "learning_rate": 2.4277956261665626e-06, "loss": 0.0202, "step": 45270 }, { "action_loss": 0.001485279412008822, "epoch": 40.710431654676256, "step": 45270 }, { "epoch": 40.710431654676256, "step": 45270, "torque_loss": 0.19818924367427826 }, { "epoch": 40.719424460431654, "grad_norm": 0.041085369884967804, "learning_rate": 2.4176265566715706e-06, "loss": 0.0168, "step": 45280 }, { "action_loss": 0.0012084353948011994, "epoch": 40.719424460431654, "step": 45280 }, { "epoch": 40.719424460431654, "step": 45280, "torque_loss": 0.1449568271636963 }, { "epoch": 40.72841726618705, "grad_norm": 0.031857769936323166, "learning_rate": 2.407478301323296e-06, "loss": 0.0185, "step": 45290 }, { "action_loss": 0.0006597451865673065, "epoch": 40.72841726618705, "step": 45290 }, { "epoch": 40.72841726618705, "step": 45290, "torque_loss": 0.12100466340780258 }, { "epoch": 40.73741007194245, "grad_norm": 0.1355614960193634, "learning_rate": 2.397350864560932e-06, "loss": 0.0162, "step": 45300 }, { "action_loss": 0.0032256923150271177, "epoch": 40.73741007194245, "step": 45300 }, { "epoch": 40.73741007194245, "step": 45300, "torque_loss": 0.2195979356765747 }, { "epoch": 40.74640287769784, "grad_norm": 0.03638632968068123, "learning_rate": 2.387244250814552e-06, "loss": 0.0166, "step": 45310 }, { "action_loss": 0.0006789359613321722, "epoch": 40.74640287769784, "step": 45310 }, { "epoch": 40.74640287769784, "step": 45310, "torque_loss": 0.1043342649936676 }, { "epoch": 40.75539568345324, "grad_norm": 0.043176449835300446, "learning_rate": 2.377158464505136e-06, "loss": 0.0179, "step": 45320 }, { "action_loss": 0.0016959545901045203, "epoch": 40.75539568345324, "step": 45320 }, { "epoch": 40.75539568345324, "step": 45320, "torque_loss": 0.12337278574705124 }, { "epoch": 40.764388489208635, "grad_norm": 0.10392025113105774, "learning_rate": 2.36709351004456e-06, "loss": 0.0201, "step": 45330 }, { "action_loss": 0.002141400473192334, "epoch": 40.764388489208635, "step": 45330 }, { "epoch": 40.764388489208635, "step": 45330, "torque_loss": 0.1558098942041397 }, { "epoch": 40.773381294964025, "grad_norm": 0.0687347874045372, "learning_rate": 2.357049391835564e-06, "loss": 0.0156, "step": 45340 }, { "action_loss": 0.0012867581099271774, "epoch": 40.773381294964025, "step": 45340 }, { "epoch": 40.773381294964025, "step": 45340, "torque_loss": 0.17520572245121002 }, { "epoch": 40.78237410071942, "grad_norm": 0.04294181242585182, "learning_rate": 2.347026114271794e-06, "loss": 0.016, "step": 45350 }, { "action_loss": 0.0010041806381195784, "epoch": 40.78237410071942, "step": 45350 }, { "epoch": 40.78237410071942, "step": 45350, "torque_loss": 0.09862671047449112 }, { "epoch": 40.79136690647482, "grad_norm": 0.04402753338217735, "learning_rate": 2.3370236817377653e-06, "loss": 0.0213, "step": 45360 }, { "action_loss": 0.0011980218114331365, "epoch": 40.79136690647482, "step": 45360 }, { "epoch": 40.79136690647482, "step": 45360, "torque_loss": 0.18620876967906952 }, { "epoch": 40.80035971223022, "grad_norm": 0.041554175317287445, "learning_rate": 2.3270420986088947e-06, "loss": 0.0194, "step": 45370 }, { "action_loss": 0.0023867664858698845, "epoch": 40.80035971223022, "step": 45370 }, { "epoch": 40.80035971223022, "step": 45370, "torque_loss": 0.18594145774841309 }, { "epoch": 40.80935251798561, "grad_norm": 0.08017437160015106, "learning_rate": 2.3170813692514505e-06, "loss": 0.0204, "step": 45380 }, { "action_loss": 0.0010283696465194225, "epoch": 40.80935251798561, "step": 45380 }, { "epoch": 40.80935251798561, "step": 45380, "torque_loss": 0.1539141684770584 }, { "epoch": 40.818345323741006, "grad_norm": 0.03200472146272659, "learning_rate": 2.3071414980226025e-06, "loss": 0.0169, "step": 45390 }, { "action_loss": 0.005783845204859972, "epoch": 40.818345323741006, "step": 45390 }, { "epoch": 40.818345323741006, "step": 45390, "torque_loss": 0.15639866888523102 }, { "epoch": 40.827338129496404, "grad_norm": 0.10449035465717316, "learning_rate": 2.29722248927039e-06, "loss": 0.0223, "step": 45400 }, { "action_loss": 0.001041750074364245, "epoch": 40.827338129496404, "step": 45400 }, { "epoch": 40.827338129496404, "step": 45400, "torque_loss": 0.10312532633543015 }, { "epoch": 40.8363309352518, "grad_norm": 0.05574957653880119, "learning_rate": 2.287324347333719e-06, "loss": 0.0178, "step": 45410 }, { "action_loss": 0.0015564793720841408, "epoch": 40.8363309352518, "step": 45410 }, { "epoch": 40.8363309352518, "step": 45410, "torque_loss": 0.17319250106811523 }, { "epoch": 40.84532374100719, "grad_norm": 0.04165513813495636, "learning_rate": 2.2774470765423773e-06, "loss": 0.0171, "step": 45420 }, { "action_loss": 0.0035401468630880117, "epoch": 40.84532374100719, "step": 45420 }, { "epoch": 40.84532374100719, "step": 45420, "torque_loss": 0.1426815241575241 }, { "epoch": 40.85431654676259, "grad_norm": 0.0382222943007946, "learning_rate": 2.2675906812170133e-06, "loss": 0.0181, "step": 45430 }, { "action_loss": 0.0020782428327947855, "epoch": 40.85431654676259, "step": 45430 }, { "epoch": 40.85431654676259, "step": 45430, "torque_loss": 0.17641305923461914 }, { "epoch": 40.86330935251799, "grad_norm": 0.05197184532880783, "learning_rate": 2.257755165669151e-06, "loss": 0.0194, "step": 45440 }, { "action_loss": 0.0010405275970697403, "epoch": 40.86330935251799, "step": 45440 }, { "epoch": 40.86330935251799, "step": 45440, "torque_loss": 0.16732710599899292 }, { "epoch": 40.87230215827338, "grad_norm": 0.03161049261689186, "learning_rate": 2.2479405342011872e-06, "loss": 0.0199, "step": 45450 }, { "action_loss": 0.0004963353276252747, "epoch": 40.87230215827338, "step": 45450 }, { "epoch": 40.87230215827338, "step": 45450, "torque_loss": 0.12858976423740387 }, { "epoch": 40.881294964028775, "grad_norm": 0.09586868435144424, "learning_rate": 2.238146791106366e-06, "loss": 0.0178, "step": 45460 }, { "action_loss": 0.01036209985613823, "epoch": 40.881294964028775, "step": 45460 }, { "epoch": 40.881294964028775, "step": 45460, "torque_loss": 0.2254522293806076 }, { "epoch": 40.89028776978417, "grad_norm": 0.06437923014163971, "learning_rate": 2.2283739406688096e-06, "loss": 0.0211, "step": 45470 }, { "action_loss": 0.009387576021254063, "epoch": 40.89028776978417, "step": 45470 }, { "epoch": 40.89028776978417, "step": 45470, "torque_loss": 0.15324240922927856 }, { "epoch": 40.89928057553957, "grad_norm": 0.02934378944337368, "learning_rate": 2.2186219871634774e-06, "loss": 0.0156, "step": 45480 }, { "action_loss": 0.0008592049707658589, "epoch": 40.89928057553957, "step": 45480 }, { "epoch": 40.89928057553957, "step": 45480, "torque_loss": 0.1149805411696434 }, { "epoch": 40.90827338129496, "grad_norm": 0.0367867536842823, "learning_rate": 2.2088909348562335e-06, "loss": 0.0181, "step": 45490 }, { "action_loss": 0.0030390294268727303, "epoch": 40.90827338129496, "step": 45490 }, { "epoch": 40.90827338129496, "step": 45490, "torque_loss": 0.19575421512126923 }, { "epoch": 40.91726618705036, "grad_norm": 0.03528987243771553, "learning_rate": 2.1991807880037517e-06, "loss": 0.02, "step": 45500 }, { "action_loss": 0.002616998739540577, "epoch": 40.91726618705036, "step": 45500 }, { "epoch": 40.91726618705036, "step": 45500, "torque_loss": 0.20649534463882446 }, { "epoch": 40.92625899280576, "grad_norm": 0.05065111443400383, "learning_rate": 2.1894915508535784e-06, "loss": 0.0176, "step": 45510 }, { "action_loss": 0.0012332272017374635, "epoch": 40.92625899280576, "step": 45510 }, { "epoch": 40.92625899280576, "step": 45510, "torque_loss": 0.1633538454771042 }, { "epoch": 40.935251798561154, "grad_norm": 0.09776406735181808, "learning_rate": 2.179823227644123e-06, "loss": 0.0183, "step": 45520 }, { "action_loss": 0.004611039534211159, "epoch": 40.935251798561154, "step": 45520 }, { "epoch": 40.935251798561154, "step": 45520, "torque_loss": 0.18349076807498932 }, { "epoch": 40.944244604316545, "grad_norm": 0.030646272003650665, "learning_rate": 2.170175822604642e-06, "loss": 0.0165, "step": 45530 }, { "action_loss": 0.0006353706703521311, "epoch": 40.944244604316545, "step": 45530 }, { "epoch": 40.944244604316545, "step": 45530, "torque_loss": 0.16816185414791107 }, { "epoch": 40.95323741007194, "grad_norm": 0.05105588585138321, "learning_rate": 2.1605493399552256e-06, "loss": 0.0181, "step": 45540 }, { "action_loss": 0.0012793125351890922, "epoch": 40.95323741007194, "step": 45540 }, { "epoch": 40.95323741007194, "step": 45540, "torque_loss": 0.14230643212795258 }, { "epoch": 40.96223021582734, "grad_norm": 0.05496403947472572, "learning_rate": 2.150943783906839e-06, "loss": 0.0164, "step": 45550 }, { "action_loss": 0.0008652556571178138, "epoch": 40.96223021582734, "step": 45550 }, { "epoch": 40.96223021582734, "step": 45550, "torque_loss": 0.13015596568584442 }, { "epoch": 40.97122302158273, "grad_norm": 0.041103798896074295, "learning_rate": 2.1413591586612712e-06, "loss": 0.0186, "step": 45560 }, { "action_loss": 0.015763135626912117, "epoch": 40.97122302158273, "step": 45560 }, { "epoch": 40.97122302158273, "step": 45560, "torque_loss": 0.2051790952682495 }, { "epoch": 40.98021582733813, "grad_norm": 0.059373047202825546, "learning_rate": 2.1317954684111576e-06, "loss": 0.0238, "step": 45570 }, { "action_loss": 0.0015008639311417937, "epoch": 40.98021582733813, "step": 45570 }, { "epoch": 40.98021582733813, "step": 45570, "torque_loss": 0.18616099655628204 }, { "epoch": 40.989208633093526, "grad_norm": 0.037947673350572586, "learning_rate": 2.1222527173399896e-06, "loss": 0.016, "step": 45580 }, { "action_loss": 0.0014323345385491848, "epoch": 40.989208633093526, "step": 45580 }, { "epoch": 40.989208633093526, "step": 45580, "torque_loss": 0.16083253920078278 }, { "epoch": 40.99820143884892, "grad_norm": 0.03578939661383629, "learning_rate": 2.112730909622085e-06, "loss": 0.018, "step": 45590 }, { "action_loss": 0.0025925887748599052, "epoch": 40.99820143884892, "step": 45590 }, { "epoch": 40.99820143884892, "step": 45590, "torque_loss": 0.23078693449497223 }, { "epoch": 41.007194244604314, "grad_norm": 0.024307243525981903, "learning_rate": 2.103230049422611e-06, "loss": 0.0154, "step": 45600 }, { "action_loss": 0.0006644045934081078, "epoch": 41.007194244604314, "step": 45600 }, { "epoch": 41.007194244604314, "step": 45600, "torque_loss": 0.1475115269422531 }, { "epoch": 41.01618705035971, "grad_norm": 0.0294047761708498, "learning_rate": 2.0937501408975603e-06, "loss": 0.0208, "step": 45610 }, { "action_loss": 0.0007857850869186223, "epoch": 41.01618705035971, "step": 45610 }, { "epoch": 41.01618705035971, "step": 45610, "torque_loss": 0.15996432304382324 }, { "epoch": 41.02517985611511, "grad_norm": 0.08591058105230331, "learning_rate": 2.084291188193782e-06, "loss": 0.0191, "step": 45620 }, { "action_loss": 0.0005806683329865336, "epoch": 41.02517985611511, "step": 45620 }, { "epoch": 41.02517985611511, "step": 45620, "torque_loss": 0.1224791407585144 }, { "epoch": 41.03417266187051, "grad_norm": 0.044995732605457306, "learning_rate": 2.0748531954489224e-06, "loss": 0.0176, "step": 45630 }, { "action_loss": 0.001566546969115734, "epoch": 41.03417266187051, "step": 45630 }, { "epoch": 41.03417266187051, "step": 45630, "torque_loss": 0.18465833365917206 }, { "epoch": 41.0431654676259, "grad_norm": 0.03930361568927765, "learning_rate": 2.065436166791496e-06, "loss": 0.0174, "step": 45640 }, { "action_loss": 0.0011634601978585124, "epoch": 41.0431654676259, "step": 45640 }, { "epoch": 41.0431654676259, "step": 45640, "torque_loss": 0.0898890495300293 }, { "epoch": 41.052158273381295, "grad_norm": 0.11348725855350494, "learning_rate": 2.056040106340823e-06, "loss": 0.0158, "step": 45650 }, { "action_loss": 0.000591907009948045, "epoch": 41.052158273381295, "step": 45650 }, { "epoch": 41.052158273381295, "step": 45650, "torque_loss": 0.1336694210767746 }, { "epoch": 41.06115107913669, "grad_norm": 0.08602424710988998, "learning_rate": 2.0466650182070667e-06, "loss": 0.0168, "step": 45660 }, { "action_loss": 0.0005168432253412902, "epoch": 41.06115107913669, "step": 45660 }, { "epoch": 41.06115107913669, "step": 45660, "torque_loss": 0.134135439991951 }, { "epoch": 41.07014388489208, "grad_norm": 0.03776926547288895, "learning_rate": 2.0373109064911964e-06, "loss": 0.0169, "step": 45670 }, { "action_loss": 0.0005143474554643035, "epoch": 41.07014388489208, "step": 45670 }, { "epoch": 41.07014388489208, "step": 45670, "torque_loss": 0.11474374681711197 }, { "epoch": 41.07913669064748, "grad_norm": 0.04040950536727905, "learning_rate": 2.027977775285028e-06, "loss": 0.0176, "step": 45680 }, { "action_loss": 0.0006297287181951106, "epoch": 41.07913669064748, "step": 45680 }, { "epoch": 41.07913669064748, "step": 45680, "torque_loss": 0.12072622776031494 }, { "epoch": 41.08812949640288, "grad_norm": 0.029596181586384773, "learning_rate": 2.0186656286711958e-06, "loss": 0.0158, "step": 45690 }, { "action_loss": 0.00330696371383965, "epoch": 41.08812949640288, "step": 45690 }, { "epoch": 41.08812949640288, "step": 45690, "torque_loss": 0.16518574953079224 }, { "epoch": 41.097122302158276, "grad_norm": 0.034919653087854385, "learning_rate": 2.00937447072313e-06, "loss": 0.0179, "step": 45700 }, { "action_loss": 0.0004700031131505966, "epoch": 41.097122302158276, "step": 45700 }, { "epoch": 41.097122302158276, "step": 45700, "torque_loss": 0.11863724142313004 }, { "epoch": 41.106115107913666, "grad_norm": 0.1674702912569046, "learning_rate": 2.0001043055051183e-06, "loss": 0.0204, "step": 45710 }, { "action_loss": 0.0003464439359959215, "epoch": 41.106115107913666, "step": 45710 }, { "epoch": 41.106115107913666, "step": 45710, "torque_loss": 0.09888085722923279 }, { "epoch": 41.115107913669064, "grad_norm": 0.03551172837615013, "learning_rate": 1.9908551370722393e-06, "loss": 0.0169, "step": 45720 }, { "action_loss": 0.0022344058379530907, "epoch": 41.115107913669064, "step": 45720 }, { "epoch": 41.115107913669064, "step": 45720, "torque_loss": 0.11991176754236221 }, { "epoch": 41.12410071942446, "grad_norm": 0.034492626786231995, "learning_rate": 1.981626969470385e-06, "loss": 0.0161, "step": 45730 }, { "action_loss": 0.002031820360571146, "epoch": 41.12410071942446, "step": 45730 }, { "epoch": 41.12410071942446, "step": 45730, "torque_loss": 0.17248423397541046 }, { "epoch": 41.13309352517986, "grad_norm": 0.13561096787452698, "learning_rate": 1.9724198067362864e-06, "loss": 0.0157, "step": 45740 }, { "action_loss": 0.0019420851022005081, "epoch": 41.13309352517986, "step": 45740 }, { "epoch": 41.13309352517986, "step": 45740, "torque_loss": 0.1536230444908142 }, { "epoch": 41.14208633093525, "grad_norm": 0.029044104740023613, "learning_rate": 1.963233652897456e-06, "loss": 0.0171, "step": 45750 }, { "action_loss": 0.004073444288223982, "epoch": 41.14208633093525, "step": 45750 }, { "epoch": 41.14208633093525, "step": 45750, "torque_loss": 0.195884570479393 }, { "epoch": 41.15107913669065, "grad_norm": 0.04305809363722801, "learning_rate": 1.9540685119722234e-06, "loss": 0.016, "step": 45760 }, { "action_loss": 0.002833815524354577, "epoch": 41.15107913669065, "step": 45760 }, { "epoch": 41.15107913669065, "step": 45760, "torque_loss": 0.15263451635837555 }, { "epoch": 41.160071942446045, "grad_norm": 0.09773742407560349, "learning_rate": 1.9449243879697543e-06, "loss": 0.017, "step": 45770 }, { "action_loss": 0.0006103370687924325, "epoch": 41.160071942446045, "step": 45770 }, { "epoch": 41.160071942446045, "step": 45770, "torque_loss": 0.17234057188034058 }, { "epoch": 41.169064748201436, "grad_norm": 0.04225276783108711, "learning_rate": 1.9358012848899874e-06, "loss": 0.0214, "step": 45780 }, { "action_loss": 0.0007094574975781143, "epoch": 41.169064748201436, "step": 45780 }, { "epoch": 41.169064748201436, "step": 45780, "torque_loss": 0.2017115205526352 }, { "epoch": 41.17805755395683, "grad_norm": 0.06486285477876663, "learning_rate": 1.92669920672367e-06, "loss": 0.0166, "step": 45790 }, { "action_loss": 0.001916468609124422, "epoch": 41.17805755395683, "step": 45790 }, { "epoch": 41.17805755395683, "step": 45790, "torque_loss": 0.1499018669128418 }, { "epoch": 41.18705035971223, "grad_norm": 0.05309562385082245, "learning_rate": 1.917618157452372e-06, "loss": 0.0174, "step": 45800 }, { "action_loss": 0.002018343424424529, "epoch": 41.18705035971223, "step": 45800 }, { "epoch": 41.18705035971223, "step": 45800, "torque_loss": 0.1572277843952179 }, { "epoch": 41.19604316546763, "grad_norm": 0.02951306477189064, "learning_rate": 1.9085581410484498e-06, "loss": 0.0175, "step": 45810 }, { "action_loss": 0.0012296180939301848, "epoch": 41.19604316546763, "step": 45810 }, { "epoch": 41.19604316546763, "step": 45810, "torque_loss": 0.14706094563007355 }, { "epoch": 41.20503597122302, "grad_norm": 0.047769155353307724, "learning_rate": 1.8995191614750607e-06, "loss": 0.0201, "step": 45820 }, { "action_loss": 0.004277164116501808, "epoch": 41.20503597122302, "step": 45820 }, { "epoch": 41.20503597122302, "step": 45820, "torque_loss": 0.1532793641090393 }, { "epoch": 41.21402877697842, "grad_norm": 0.14454486966133118, "learning_rate": 1.890501222686164e-06, "loss": 0.0202, "step": 45830 }, { "action_loss": 0.0013272515498101711, "epoch": 41.21402877697842, "step": 45830 }, { "epoch": 41.21402877697842, "step": 45830, "torque_loss": 0.14130020141601562 }, { "epoch": 41.223021582733814, "grad_norm": 0.06672444194555283, "learning_rate": 1.8815043286265045e-06, "loss": 0.0159, "step": 45840 }, { "action_loss": 0.00312092830426991, "epoch": 41.223021582733814, "step": 45840 }, { "epoch": 41.223021582733814, "step": 45840, "torque_loss": 0.19035963714122772 }, { "epoch": 41.23201438848921, "grad_norm": 0.12603707611560822, "learning_rate": 1.8725284832316448e-06, "loss": 0.018, "step": 45850 }, { "action_loss": 0.0008341858629137278, "epoch": 41.23201438848921, "step": 45850 }, { "epoch": 41.23201438848921, "step": 45850, "torque_loss": 0.12623460590839386 }, { "epoch": 41.2410071942446, "grad_norm": 0.12560194730758667, "learning_rate": 1.8635736904279055e-06, "loss": 0.0166, "step": 45860 }, { "action_loss": 0.00159002048894763, "epoch": 41.2410071942446, "step": 45860 }, { "epoch": 41.2410071942446, "step": 45860, "torque_loss": 0.09848115593194962 }, { "epoch": 41.25, "grad_norm": 0.04895864799618721, "learning_rate": 1.8546399541324367e-06, "loss": 0.0148, "step": 45870 }, { "action_loss": 0.0004677496908698231, "epoch": 41.25, "step": 45870 }, { "epoch": 41.25, "step": 45870, "torque_loss": 0.10056862235069275 }, { "epoch": 41.2589928057554, "grad_norm": 0.05205276980996132, "learning_rate": 1.8457272782531453e-06, "loss": 0.0158, "step": 45880 }, { "action_loss": 0.0009570395923219621, "epoch": 41.2589928057554, "step": 45880 }, { "epoch": 41.2589928057554, "step": 45880, "torque_loss": 0.12252835184335709 }, { "epoch": 41.26798561151079, "grad_norm": 0.10347089916467667, "learning_rate": 1.836835666688741e-06, "loss": 0.0189, "step": 45890 }, { "action_loss": 0.0014783154474571347, "epoch": 41.26798561151079, "step": 45890 }, { "epoch": 41.26798561151079, "step": 45890, "torque_loss": 0.1623109132051468 }, { "epoch": 41.276978417266186, "grad_norm": 0.08301058411598206, "learning_rate": 1.8279651233287175e-06, "loss": 0.0158, "step": 45900 }, { "action_loss": 0.0005455096252262592, "epoch": 41.276978417266186, "step": 45900 }, { "epoch": 41.276978417266186, "step": 45900, "torque_loss": 0.10667093843221664 }, { "epoch": 41.28597122302158, "grad_norm": 0.04662534222006798, "learning_rate": 1.8191156520533547e-06, "loss": 0.0173, "step": 45910 }, { "action_loss": 0.0037494630087167025, "epoch": 41.28597122302158, "step": 45910 }, { "epoch": 41.28597122302158, "step": 45910, "torque_loss": 0.19141936302185059 }, { "epoch": 41.29496402877698, "grad_norm": 0.035348981618881226, "learning_rate": 1.8102872567337059e-06, "loss": 0.0167, "step": 45920 }, { "action_loss": 0.0008197265560738742, "epoch": 41.29496402877698, "step": 45920 }, { "epoch": 41.29496402877698, "step": 45920, "torque_loss": 0.1445455104112625 }, { "epoch": 41.30395683453237, "grad_norm": 0.04340989515185356, "learning_rate": 1.8014799412316152e-06, "loss": 0.0147, "step": 45930 }, { "action_loss": 0.0012601438211277127, "epoch": 41.30395683453237, "step": 45930 }, { "epoch": 41.30395683453237, "step": 45930, "torque_loss": 0.1659848839044571 }, { "epoch": 41.31294964028777, "grad_norm": 0.03543716296553612, "learning_rate": 1.7926937093997065e-06, "loss": 0.0168, "step": 45940 }, { "action_loss": 0.0012667811242863536, "epoch": 41.31294964028777, "step": 45940 }, { "epoch": 41.31294964028777, "step": 45940, "torque_loss": 0.1462322324514389 }, { "epoch": 41.32194244604317, "grad_norm": 0.08430071175098419, "learning_rate": 1.7839285650813608e-06, "loss": 0.0201, "step": 45950 }, { "action_loss": 0.0015754246851429343, "epoch": 41.32194244604317, "step": 45950 }, { "epoch": 41.32194244604317, "step": 45950, "torque_loss": 0.21824799478054047 }, { "epoch": 41.330935251798564, "grad_norm": 0.05783989652991295, "learning_rate": 1.7751845121107668e-06, "loss": 0.0191, "step": 45960 }, { "action_loss": 0.0009781831176951528, "epoch": 41.330935251798564, "step": 45960 }, { "epoch": 41.330935251798564, "step": 45960, "torque_loss": 0.12386366724967957 }, { "epoch": 41.339928057553955, "grad_norm": 0.026109756901860237, "learning_rate": 1.7664615543128648e-06, "loss": 0.0152, "step": 45970 }, { "action_loss": 0.005852010566741228, "epoch": 41.339928057553955, "step": 45970 }, { "epoch": 41.339928057553955, "step": 45970, "torque_loss": 0.18911834061145782 }, { "epoch": 41.34892086330935, "grad_norm": 0.0604536347091198, "learning_rate": 1.7577596955033692e-06, "loss": 0.0175, "step": 45980 }, { "action_loss": 0.001250453875400126, "epoch": 41.34892086330935, "step": 45980 }, { "epoch": 41.34892086330935, "step": 45980, "torque_loss": 0.10123749822378159 }, { "epoch": 41.35791366906475, "grad_norm": 0.06943168491125107, "learning_rate": 1.7490789394887686e-06, "loss": 0.017, "step": 45990 }, { "action_loss": 0.0019423641497269273, "epoch": 41.35791366906475, "step": 45990 }, { "epoch": 41.35791366906475, "step": 45990, "torque_loss": 0.17328709363937378 }, { "epoch": 41.36690647482014, "grad_norm": 0.12389127910137177, "learning_rate": 1.7404192900663252e-06, "loss": 0.0178, "step": 46000 }, { "action_loss": 0.00073953066021204, "epoch": 41.36690647482014, "step": 46000 }, { "epoch": 41.36690647482014, "step": 46000, "torque_loss": 0.167486310005188 }, { "epoch": 41.37589928057554, "grad_norm": 0.1284342110157013, "learning_rate": 1.7317807510240647e-06, "loss": 0.0185, "step": 46010 }, { "action_loss": 0.002797364955767989, "epoch": 41.37589928057554, "step": 46010 }, { "epoch": 41.37589928057554, "step": 46010, "torque_loss": 0.09773794561624527 }, { "epoch": 41.384892086330936, "grad_norm": 0.07010067254304886, "learning_rate": 1.7231633261407642e-06, "loss": 0.0176, "step": 46020 }, { "action_loss": 0.0011897679651156068, "epoch": 41.384892086330936, "step": 46020 }, { "epoch": 41.384892086330936, "step": 46020, "torque_loss": 0.13536040484905243 }, { "epoch": 41.393884892086334, "grad_norm": 0.03221447765827179, "learning_rate": 1.7145670191859974e-06, "loss": 0.0181, "step": 46030 }, { "action_loss": 0.0008733072900213301, "epoch": 41.393884892086334, "step": 46030 }, { "epoch": 41.393884892086334, "step": 46030, "torque_loss": 0.14972329139709473 }, { "epoch": 41.402877697841724, "grad_norm": 0.07037355750799179, "learning_rate": 1.7059918339200732e-06, "loss": 0.0178, "step": 46040 }, { "action_loss": 0.004625646863132715, "epoch": 41.402877697841724, "step": 46040 }, { "epoch": 41.402877697841724, "step": 46040, "torque_loss": 0.18920688331127167 }, { "epoch": 41.41187050359712, "grad_norm": 0.04527951776981354, "learning_rate": 1.6974377740940629e-06, "loss": 0.0199, "step": 46050 }, { "action_loss": 0.0008807972189970315, "epoch": 41.41187050359712, "step": 46050 }, { "epoch": 41.41187050359712, "step": 46050, "torque_loss": 0.1184941828250885 }, { "epoch": 41.42086330935252, "grad_norm": 0.034539688378572464, "learning_rate": 1.6889048434498066e-06, "loss": 0.0188, "step": 46060 }, { "action_loss": 0.0005654340493492782, "epoch": 41.42086330935252, "step": 46060 }, { "epoch": 41.42086330935252, "step": 46060, "torque_loss": 0.11776119470596313 }, { "epoch": 41.42985611510792, "grad_norm": 0.036748483777046204, "learning_rate": 1.6803930457199023e-06, "loss": 0.0178, "step": 46070 }, { "action_loss": 0.008463594131171703, "epoch": 41.42985611510792, "step": 46070 }, { "epoch": 41.42985611510792, "step": 46070, "torque_loss": 0.22539113461971283 }, { "epoch": 41.43884892086331, "grad_norm": 0.04159962758421898, "learning_rate": 1.6719023846276938e-06, "loss": 0.0208, "step": 46080 }, { "action_loss": 0.002441044431179762, "epoch": 41.43884892086331, "step": 46080 }, { "epoch": 41.43884892086331, "step": 46080, "torque_loss": 0.1748627871274948 }, { "epoch": 41.447841726618705, "grad_norm": 0.11649847775697708, "learning_rate": 1.6634328638872877e-06, "loss": 0.0197, "step": 46090 }, { "action_loss": 0.0012179111363366246, "epoch": 41.447841726618705, "step": 46090 }, { "epoch": 41.447841726618705, "step": 46090, "torque_loss": 0.09930034726858139 }, { "epoch": 41.4568345323741, "grad_norm": 0.03536606952548027, "learning_rate": 1.6549844872035424e-06, "loss": 0.0242, "step": 46100 }, { "action_loss": 0.0021087026689201593, "epoch": 41.4568345323741, "step": 46100 }, { "epoch": 41.4568345323741, "step": 46100, "torque_loss": 0.15497301518917084 }, { "epoch": 41.46582733812949, "grad_norm": 0.06068553403019905, "learning_rate": 1.6465572582720579e-06, "loss": 0.015, "step": 46110 }, { "action_loss": 0.0024524538312107325, "epoch": 41.46582733812949, "step": 46110 }, { "epoch": 41.46582733812949, "step": 46110, "torque_loss": 0.18456630408763885 }, { "epoch": 41.47482014388489, "grad_norm": 0.053121648728847504, "learning_rate": 1.6381511807792017e-06, "loss": 0.0208, "step": 46120 }, { "action_loss": 0.0010405428474768996, "epoch": 41.47482014388489, "step": 46120 }, { "epoch": 41.47482014388489, "step": 46120, "torque_loss": 0.1323406994342804 }, { "epoch": 41.48381294964029, "grad_norm": 0.09722535312175751, "learning_rate": 1.6297662584020656e-06, "loss": 0.0174, "step": 46130 }, { "action_loss": 0.0008905662107281387, "epoch": 41.48381294964029, "step": 46130 }, { "epoch": 41.48381294964029, "step": 46130, "torque_loss": 0.129548579454422 }, { "epoch": 41.492805755395686, "grad_norm": 0.13828764855861664, "learning_rate": 1.621402494808505e-06, "loss": 0.0248, "step": 46140 }, { "action_loss": 0.000951309863012284, "epoch": 41.492805755395686, "step": 46140 }, { "epoch": 41.492805755395686, "step": 46140, "torque_loss": 0.15290389955043793 }, { "epoch": 41.50179856115108, "grad_norm": 0.08861977607011795, "learning_rate": 1.613059893657115e-06, "loss": 0.0174, "step": 46150 }, { "action_loss": 0.010328011587262154, "epoch": 41.50179856115108, "step": 46150 }, { "epoch": 41.50179856115108, "step": 46150, "torque_loss": 0.2558341324329376 }, { "epoch": 41.510791366906474, "grad_norm": 0.0596337616443634, "learning_rate": 1.604738458597238e-06, "loss": 0.0211, "step": 46160 }, { "action_loss": 0.0014392592711374164, "epoch": 41.510791366906474, "step": 46160 }, { "epoch": 41.510791366906474, "step": 46160, "torque_loss": 0.1628272533416748 }, { "epoch": 41.51978417266187, "grad_norm": 0.042536262422800064, "learning_rate": 1.5964381932689398e-06, "loss": 0.0163, "step": 46170 }, { "action_loss": 0.0013233805075287819, "epoch": 41.51978417266187, "step": 46170 }, { "epoch": 41.51978417266187, "step": 46170, "torque_loss": 0.116231270134449 }, { "epoch": 41.52877697841727, "grad_norm": 0.07434417307376862, "learning_rate": 1.5881591013030384e-06, "loss": 0.0198, "step": 46180 }, { "action_loss": 0.0012008969206362963, "epoch": 41.52877697841727, "step": 46180 }, { "epoch": 41.52877697841727, "step": 46180, "torque_loss": 0.10907591134309769 }, { "epoch": 41.53776978417266, "grad_norm": 0.06204238533973694, "learning_rate": 1.5799011863211089e-06, "loss": 0.0168, "step": 46190 }, { "action_loss": 0.014848294667899609, "epoch": 41.53776978417266, "step": 46190 }, { "epoch": 41.53776978417266, "step": 46190, "torque_loss": 0.21597151458263397 }, { "epoch": 41.54676258992806, "grad_norm": 0.06118246167898178, "learning_rate": 1.5716644519354285e-06, "loss": 0.0237, "step": 46200 }, { "action_loss": 0.0009129985119216144, "epoch": 41.54676258992806, "step": 46200 }, { "epoch": 41.54676258992806, "step": 46200, "torque_loss": 0.10402816534042358 }, { "epoch": 41.555755395683455, "grad_norm": 0.027253994718194008, "learning_rate": 1.5634489017490151e-06, "loss": 0.0173, "step": 46210 }, { "action_loss": 0.001071442966349423, "epoch": 41.555755395683455, "step": 46210 }, { "epoch": 41.555755395683455, "step": 46210, "torque_loss": 0.1794067621231079 }, { "epoch": 41.564748201438846, "grad_norm": 0.03742016479372978, "learning_rate": 1.555254539355655e-06, "loss": 0.0183, "step": 46220 }, { "action_loss": 0.006270707119256258, "epoch": 41.564748201438846, "step": 46220 }, { "epoch": 41.564748201438846, "step": 46220, "torque_loss": 0.23208022117614746 }, { "epoch": 41.57374100719424, "grad_norm": 0.07659360766410828, "learning_rate": 1.5470813683398255e-06, "loss": 0.02, "step": 46230 }, { "action_loss": 0.0007782878819853067, "epoch": 41.57374100719424, "step": 46230 }, { "epoch": 41.57374100719424, "step": 46230, "torque_loss": 0.11729240417480469 }, { "epoch": 41.58273381294964, "grad_norm": 0.03147014230489731, "learning_rate": 1.5389293922767444e-06, "loss": 0.0183, "step": 46240 }, { "action_loss": 0.001355412881821394, "epoch": 41.58273381294964, "step": 46240 }, { "epoch": 41.58273381294964, "step": 46240, "torque_loss": 0.17227546870708466 }, { "epoch": 41.59172661870504, "grad_norm": 0.03327574208378792, "learning_rate": 1.530798614732376e-06, "loss": 0.0179, "step": 46250 }, { "action_loss": 0.0017148703336715698, "epoch": 41.59172661870504, "step": 46250 }, { "epoch": 41.59172661870504, "step": 46250, "torque_loss": 0.1428215056657791 }, { "epoch": 41.60071942446043, "grad_norm": 0.05682683736085892, "learning_rate": 1.5226890392633919e-06, "loss": 0.0174, "step": 46260 }, { "action_loss": 0.0007365791243501008, "epoch": 41.60071942446043, "step": 46260 }, { "epoch": 41.60071942446043, "step": 46260, "torque_loss": 0.14968326687812805 }, { "epoch": 41.60971223021583, "grad_norm": 0.039525046944618225, "learning_rate": 1.5146006694171932e-06, "loss": 0.0156, "step": 46270 }, { "action_loss": 0.0008615698316134512, "epoch": 41.60971223021583, "step": 46270 }, { "epoch": 41.60971223021583, "step": 46270, "torque_loss": 0.15084461867809296 }, { "epoch": 41.618705035971225, "grad_norm": 0.0565735399723053, "learning_rate": 1.5065335087319166e-06, "loss": 0.0198, "step": 46280 }, { "action_loss": 0.0004790286475326866, "epoch": 41.618705035971225, "step": 46280 }, { "epoch": 41.618705035971225, "step": 46280, "torque_loss": 0.1670101433992386 }, { "epoch": 41.62769784172662, "grad_norm": 0.02553567849099636, "learning_rate": 1.4984875607364057e-06, "loss": 0.022, "step": 46290 }, { "action_loss": 0.001504051499068737, "epoch": 41.62769784172662, "step": 46290 }, { "epoch": 41.62769784172662, "step": 46290, "torque_loss": 0.19162581861019135 }, { "epoch": 41.63669064748201, "grad_norm": 0.040425676852464676, "learning_rate": 1.4904628289502287e-06, "loss": 0.0179, "step": 46300 }, { "action_loss": 0.0004860105400439352, "epoch": 41.63669064748201, "step": 46300 }, { "epoch": 41.63669064748201, "step": 46300, "torque_loss": 0.13453586399555206 }, { "epoch": 41.64568345323741, "grad_norm": 0.04982108250260353, "learning_rate": 1.482459316883683e-06, "loss": 0.0228, "step": 46310 }, { "action_loss": 0.0009523858316242695, "epoch": 41.64568345323741, "step": 46310 }, { "epoch": 41.64568345323741, "step": 46310, "torque_loss": 0.08928487449884415 }, { "epoch": 41.65467625899281, "grad_norm": 0.06185057759284973, "learning_rate": 1.474477028037763e-06, "loss": 0.0172, "step": 46320 }, { "action_loss": 0.008204912766814232, "epoch": 41.65467625899281, "step": 46320 }, { "epoch": 41.65467625899281, "step": 46320, "torque_loss": 0.19008217751979828 }, { "epoch": 41.6636690647482, "grad_norm": 0.05735287442803383, "learning_rate": 1.4665159659042139e-06, "loss": 0.0193, "step": 46330 }, { "action_loss": 0.0008205363410525024, "epoch": 41.6636690647482, "step": 46330 }, { "epoch": 41.6636690647482, "step": 46330, "torque_loss": 0.16368190944194794 }, { "epoch": 41.672661870503596, "grad_norm": 0.09231332689523697, "learning_rate": 1.4585761339654613e-06, "loss": 0.0178, "step": 46340 }, { "action_loss": 0.001977955223992467, "epoch": 41.672661870503596, "step": 46340 }, { "epoch": 41.672661870503596, "step": 46340, "torque_loss": 0.14272934198379517 }, { "epoch": 41.681654676258994, "grad_norm": 0.10745901614427567, "learning_rate": 1.4506575356946606e-06, "loss": 0.0178, "step": 46350 }, { "action_loss": 0.0009283426334150136, "epoch": 41.681654676258994, "step": 46350 }, { "epoch": 41.681654676258994, "step": 46350, "torque_loss": 0.1282581090927124 }, { "epoch": 41.69064748201439, "grad_norm": 0.03947827219963074, "learning_rate": 1.4427601745556796e-06, "loss": 0.0222, "step": 46360 }, { "action_loss": 0.0007445023511536419, "epoch": 41.69064748201439, "step": 46360 }, { "epoch": 41.69064748201439, "step": 46360, "torque_loss": 0.1171979308128357 }, { "epoch": 41.69964028776978, "grad_norm": 0.03042510151863098, "learning_rate": 1.4348840540030828e-06, "loss": 0.0176, "step": 46370 }, { "action_loss": 0.0008838625508360565, "epoch": 41.69964028776978, "step": 46370 }, { "epoch": 41.69964028776978, "step": 46370, "torque_loss": 0.18375857174396515 }, { "epoch": 41.70863309352518, "grad_norm": 0.08514200150966644, "learning_rate": 1.4270291774821586e-06, "loss": 0.0148, "step": 46380 }, { "action_loss": 0.0006140273180790246, "epoch": 41.70863309352518, "step": 46380 }, { "epoch": 41.70863309352518, "step": 46380, "torque_loss": 0.13688741624355316 }, { "epoch": 41.71762589928058, "grad_norm": 0.10704721510410309, "learning_rate": 1.4191955484289032e-06, "loss": 0.0191, "step": 46390 }, { "action_loss": 0.0021548199001699686, "epoch": 41.71762589928058, "step": 46390 }, { "epoch": 41.71762589928058, "step": 46390, "torque_loss": 0.18908517062664032 }, { "epoch": 41.726618705035975, "grad_norm": 0.04931522533297539, "learning_rate": 1.4113831702699976e-06, "loss": 0.0175, "step": 46400 }, { "action_loss": 0.0006987331435084343, "epoch": 41.726618705035975, "step": 46400 }, { "epoch": 41.726618705035975, "step": 46400, "torque_loss": 0.17924658954143524 }, { "epoch": 41.735611510791365, "grad_norm": 0.08417188376188278, "learning_rate": 1.4035920464228526e-06, "loss": 0.0212, "step": 46410 }, { "action_loss": 0.0032582220155745745, "epoch": 41.735611510791365, "step": 46410 }, { "epoch": 41.735611510791365, "step": 46410, "torque_loss": 0.16875046491622925 }, { "epoch": 41.74460431654676, "grad_norm": 0.058326274156570435, "learning_rate": 1.3958221802955696e-06, "loss": 0.0187, "step": 46420 }, { "action_loss": 0.001501624472439289, "epoch": 41.74460431654676, "step": 46420 }, { "epoch": 41.74460431654676, "step": 46420, "torque_loss": 0.17350272834300995 }, { "epoch": 41.75359712230216, "grad_norm": 0.05141650140285492, "learning_rate": 1.3880735752869412e-06, "loss": 0.0192, "step": 46430 }, { "action_loss": 0.0005713627324439585, "epoch": 41.75359712230216, "step": 46430 }, { "epoch": 41.75359712230216, "step": 46430, "torque_loss": 0.11617898941040039 }, { "epoch": 41.76258992805755, "grad_norm": 0.11761302500963211, "learning_rate": 1.3803462347864893e-06, "loss": 0.0182, "step": 46440 }, { "action_loss": 0.0024542673490941525, "epoch": 41.76258992805755, "step": 46440 }, { "epoch": 41.76258992805755, "step": 46440, "torque_loss": 0.1723630428314209 }, { "epoch": 41.77158273381295, "grad_norm": 0.053326427936553955, "learning_rate": 1.372640162174399e-06, "loss": 0.0236, "step": 46450 }, { "action_loss": 0.0005520866834558547, "epoch": 41.77158273381295, "step": 46450 }, { "epoch": 41.77158273381295, "step": 46450, "torque_loss": 0.12341942638158798 }, { "epoch": 41.780575539568346, "grad_norm": 0.0482429601252079, "learning_rate": 1.3649553608215738e-06, "loss": 0.02, "step": 46460 }, { "action_loss": 0.008065737783908844, "epoch": 41.780575539568346, "step": 46460 }, { "epoch": 41.780575539568346, "step": 46460, "torque_loss": 0.2836357653141022 }, { "epoch": 41.789568345323744, "grad_norm": 0.04144630208611488, "learning_rate": 1.357291834089608e-06, "loss": 0.0182, "step": 46470 }, { "action_loss": 0.0013520546490326524, "epoch": 41.789568345323744, "step": 46470 }, { "epoch": 41.789568345323744, "step": 46470, "torque_loss": 0.23985423147678375 }, { "epoch": 41.798561151079134, "grad_norm": 0.034578125923871994, "learning_rate": 1.3496495853307921e-06, "loss": 0.0185, "step": 46480 }, { "action_loss": 0.0008121027494780719, "epoch": 41.798561151079134, "step": 46480 }, { "epoch": 41.798561151079134, "step": 46480, "torque_loss": 0.1343890279531479 }, { "epoch": 41.80755395683453, "grad_norm": 0.02459840103983879, "learning_rate": 1.3420286178880914e-06, "loss": 0.0213, "step": 46490 }, { "action_loss": 0.0021042192820459604, "epoch": 41.80755395683453, "step": 46490 }, { "epoch": 41.80755395683453, "step": 46490, "torque_loss": 0.17861777544021606 }, { "epoch": 41.81654676258993, "grad_norm": 0.04093730449676514, "learning_rate": 1.3344289350951943e-06, "loss": 0.0209, "step": 46500 }, { "action_loss": 0.0008588752243667841, "epoch": 41.81654676258993, "step": 46500 }, { "epoch": 41.81654676258993, "step": 46500, "torque_loss": 0.19344301521778107 }, { "epoch": 41.82553956834532, "grad_norm": 0.038643669337034225, "learning_rate": 1.3268505402764475e-06, "loss": 0.0157, "step": 46510 }, { "action_loss": 0.0011981161078438163, "epoch": 41.82553956834532, "step": 46510 }, { "epoch": 41.82553956834532, "step": 46510, "torque_loss": 0.1483754962682724 }, { "epoch": 41.83453237410072, "grad_norm": 0.09448421746492386, "learning_rate": 1.3192934367469045e-06, "loss": 0.0228, "step": 46520 }, { "action_loss": 0.003336001420393586, "epoch": 41.83453237410072, "step": 46520 }, { "epoch": 41.83453237410072, "step": 46520, "torque_loss": 0.2585517466068268 }, { "epoch": 41.843525179856115, "grad_norm": 0.04741101339459419, "learning_rate": 1.3117576278122934e-06, "loss": 0.0169, "step": 46530 }, { "action_loss": 0.0009436767431907356, "epoch": 41.843525179856115, "step": 46530 }, { "epoch": 41.843525179856115, "step": 46530, "torque_loss": 0.18315714597702026 }, { "epoch": 41.85251798561151, "grad_norm": 0.037213750183582306, "learning_rate": 1.3042431167690382e-06, "loss": 0.0184, "step": 46540 }, { "action_loss": 0.0007450736593455076, "epoch": 41.85251798561151, "step": 46540 }, { "epoch": 41.85251798561151, "step": 46540, "torque_loss": 0.177223801612854 }, { "epoch": 41.861510791366904, "grad_norm": 0.030786044895648956, "learning_rate": 1.2967499069042377e-06, "loss": 0.0173, "step": 46550 }, { "action_loss": 0.0006660696235485375, "epoch": 41.861510791366904, "step": 46550 }, { "epoch": 41.861510791366904, "step": 46550, "torque_loss": 0.14239485561847687 }, { "epoch": 41.8705035971223, "grad_norm": 0.12794044613838196, "learning_rate": 1.2892780014956862e-06, "loss": 0.019, "step": 46560 }, { "action_loss": 0.03158560395240784, "epoch": 41.8705035971223, "step": 46560 }, { "epoch": 41.8705035971223, "step": 46560, "torque_loss": 0.21978963911533356 }, { "epoch": 41.8794964028777, "grad_norm": 0.05427245423197746, "learning_rate": 1.2818274038118306e-06, "loss": 0.0208, "step": 46570 }, { "action_loss": 0.0005991921643726528, "epoch": 41.8794964028777, "step": 46570 }, { "epoch": 41.8794964028777, "step": 46570, "torque_loss": 0.13713055849075317 }, { "epoch": 41.888489208633096, "grad_norm": 0.04216175153851509, "learning_rate": 1.2743981171118357e-06, "loss": 0.0174, "step": 46580 }, { "action_loss": 0.0027462404686957598, "epoch": 41.888489208633096, "step": 46580 }, { "epoch": 41.888489208633096, "step": 46580, "torque_loss": 0.1774054765701294 }, { "epoch": 41.89748201438849, "grad_norm": 0.07932965457439423, "learning_rate": 1.2669901446455024e-06, "loss": 0.0177, "step": 46590 }, { "action_loss": 0.006868436466902494, "epoch": 41.89748201438849, "step": 46590 }, { "epoch": 41.89748201438849, "step": 46590, "torque_loss": 0.2392359972000122 }, { "epoch": 41.906474820143885, "grad_norm": 0.037359777837991714, "learning_rate": 1.2596034896533549e-06, "loss": 0.0183, "step": 46600 }, { "action_loss": 0.00039478353573940694, "epoch": 41.906474820143885, "step": 46600 }, { "epoch": 41.906474820143885, "step": 46600, "torque_loss": 0.08697102218866348 }, { "epoch": 41.91546762589928, "grad_norm": 0.07775095105171204, "learning_rate": 1.2522381553665474e-06, "loss": 0.0179, "step": 46610 }, { "action_loss": 0.000752136402297765, "epoch": 41.91546762589928, "step": 46610 }, { "epoch": 41.91546762589928, "step": 46610, "torque_loss": 0.12390343099832535 }, { "epoch": 41.92446043165468, "grad_norm": 0.09543216973543167, "learning_rate": 1.2448941450069306e-06, "loss": 0.0154, "step": 46620 }, { "action_loss": 0.0026069164741784334, "epoch": 41.92446043165468, "step": 46620 }, { "epoch": 41.92446043165468, "step": 46620, "torque_loss": 0.23427660763263702 }, { "epoch": 41.93345323741007, "grad_norm": 0.048261478543281555, "learning_rate": 1.2375714617870293e-06, "loss": 0.019, "step": 46630 }, { "action_loss": 0.0008114916854538023, "epoch": 41.93345323741007, "step": 46630 }, { "epoch": 41.93345323741007, "step": 46630, "torque_loss": 0.16181181371212006 }, { "epoch": 41.94244604316547, "grad_norm": 0.04962342977523804, "learning_rate": 1.2302701089100366e-06, "loss": 0.0161, "step": 46640 }, { "action_loss": 0.0024961628951132298, "epoch": 41.94244604316547, "step": 46640 }, { "epoch": 41.94244604316547, "step": 46640, "torque_loss": 0.18028675019741058 }, { "epoch": 41.951438848920866, "grad_norm": 0.11037667095661163, "learning_rate": 1.2229900895698032e-06, "loss": 0.0195, "step": 46650 }, { "action_loss": 0.0016290112398564816, "epoch": 41.951438848920866, "step": 46650 }, { "epoch": 41.951438848920866, "step": 46650, "torque_loss": 0.16029736399650574 }, { "epoch": 41.960431654676256, "grad_norm": 0.05710646137595177, "learning_rate": 1.2157314069508596e-06, "loss": 0.0167, "step": 46660 }, { "action_loss": 0.0005048543680459261, "epoch": 41.960431654676256, "step": 46660 }, { "epoch": 41.960431654676256, "step": 46660, "torque_loss": 0.12700651586055756 }, { "epoch": 41.969424460431654, "grad_norm": 0.0238661952316761, "learning_rate": 1.2084940642284048e-06, "loss": 0.0174, "step": 46670 }, { "action_loss": 0.0030760460067540407, "epoch": 41.969424460431654, "step": 46670 }, { "epoch": 41.969424460431654, "step": 46670, "torque_loss": 0.1511586457490921 }, { "epoch": 41.97841726618705, "grad_norm": 0.04621659219264984, "learning_rate": 1.2012780645682952e-06, "loss": 0.018, "step": 46680 }, { "action_loss": 0.0005133256199769676, "epoch": 41.97841726618705, "step": 46680 }, { "epoch": 41.97841726618705, "step": 46680, "torque_loss": 0.15244166553020477 }, { "epoch": 41.98741007194245, "grad_norm": 0.025856997817754745, "learning_rate": 1.1940834111270505e-06, "loss": 0.0156, "step": 46690 }, { "action_loss": 0.001103671151213348, "epoch": 41.98741007194245, "step": 46690 }, { "epoch": 41.98741007194245, "step": 46690, "torque_loss": 0.1460888683795929 }, { "epoch": 41.99640287769784, "grad_norm": 0.111343614757061, "learning_rate": 1.1869101070518695e-06, "loss": 0.022, "step": 46700 }, { "action_loss": 0.0008050985634326935, "epoch": 41.99640287769784, "step": 46700 }, { "epoch": 41.99640287769784, "step": 46700, "torque_loss": 0.16740679740905762 }, { "epoch": 42.00539568345324, "grad_norm": 0.052390098571777344, "learning_rate": 1.1797581554805813e-06, "loss": 0.0164, "step": 46710 }, { "action_loss": 0.006598193198442459, "epoch": 42.00539568345324, "step": 46710 }, { "epoch": 42.00539568345324, "step": 46710, "torque_loss": 0.1669522523880005 }, { "epoch": 42.014388489208635, "grad_norm": 0.08760137856006622, "learning_rate": 1.1726275595417001e-06, "loss": 0.0151, "step": 46720 }, { "action_loss": 0.0008030316676013172, "epoch": 42.014388489208635, "step": 46720 }, { "epoch": 42.014388489208635, "step": 46720, "torque_loss": 0.12343326210975647 }, { "epoch": 42.023381294964025, "grad_norm": 0.029501009732484818, "learning_rate": 1.1655183223543976e-06, "loss": 0.0168, "step": 46730 }, { "action_loss": 0.0011983518488705158, "epoch": 42.023381294964025, "step": 46730 }, { "epoch": 42.023381294964025, "step": 46730, "torque_loss": 0.17675264179706573 }, { "epoch": 42.03237410071942, "grad_norm": 0.06446801126003265, "learning_rate": 1.1584304470284912e-06, "loss": 0.0182, "step": 46740 }, { "action_loss": 0.0006791629712097347, "epoch": 42.03237410071942, "step": 46740 }, { "epoch": 42.03237410071942, "step": 46740, "torque_loss": 0.13715024292469025 }, { "epoch": 42.04136690647482, "grad_norm": 0.05522613972425461, "learning_rate": 1.1513639366644513e-06, "loss": 0.0184, "step": 46750 }, { "action_loss": 0.0009971782565116882, "epoch": 42.04136690647482, "step": 46750 }, { "epoch": 42.04136690647482, "step": 46750, "torque_loss": 0.15661068260669708 }, { "epoch": 42.05035971223022, "grad_norm": 0.025218676775693893, "learning_rate": 1.1443187943534162e-06, "loss": 0.0178, "step": 46760 }, { "action_loss": 0.004524456337094307, "epoch": 42.05035971223022, "step": 46760 }, { "epoch": 42.05035971223022, "step": 46760, "torque_loss": 0.18747298419475555 }, { "epoch": 42.05935251798561, "grad_norm": 0.03193159028887749, "learning_rate": 1.1372950231771707e-06, "loss": 0.0187, "step": 46770 }, { "action_loss": 0.000522095535416156, "epoch": 42.05935251798561, "step": 46770 }, { "epoch": 42.05935251798561, "step": 46770, "torque_loss": 0.11719825118780136 }, { "epoch": 42.068345323741006, "grad_norm": 0.03310796990990639, "learning_rate": 1.1302926262081515e-06, "loss": 0.0177, "step": 46780 }, { "action_loss": 0.008161385543644428, "epoch": 42.068345323741006, "step": 46780 }, { "epoch": 42.068345323741006, "step": 46780, "torque_loss": 0.2504183053970337 }, { "epoch": 42.077338129496404, "grad_norm": 0.05770005285739899, "learning_rate": 1.1233116065094362e-06, "loss": 0.0174, "step": 46790 }, { "action_loss": 0.0033632952254265547, "epoch": 42.077338129496404, "step": 46790 }, { "epoch": 42.077338129496404, "step": 46790, "torque_loss": 0.21022482216358185 }, { "epoch": 42.0863309352518, "grad_norm": 0.06667614728212357, "learning_rate": 1.1163519671347767e-06, "loss": 0.0165, "step": 46800 }, { "action_loss": 0.0015812820056453347, "epoch": 42.0863309352518, "step": 46800 }, { "epoch": 42.0863309352518, "step": 46800, "torque_loss": 0.18317438662052155 }, { "epoch": 42.09532374100719, "grad_norm": 0.0631331354379654, "learning_rate": 1.1094137111285375e-06, "loss": 0.0204, "step": 46810 }, { "action_loss": 0.0011994056403636932, "epoch": 42.09532374100719, "step": 46810 }, { "epoch": 42.09532374100719, "step": 46810, "torque_loss": 0.15644074976444244 }, { "epoch": 42.10431654676259, "grad_norm": 0.03776656463742256, "learning_rate": 1.102496841525763e-06, "loss": 0.0155, "step": 46820 }, { "action_loss": 0.009776419028639793, "epoch": 42.10431654676259, "step": 46820 }, { "epoch": 42.10431654676259, "step": 46820, "torque_loss": 0.18424952030181885 }, { "epoch": 42.11330935251799, "grad_norm": 0.06259895116090775, "learning_rate": 1.095601361352122e-06, "loss": 0.0187, "step": 46830 }, { "action_loss": 0.0014466851716861129, "epoch": 42.11330935251799, "step": 46830 }, { "epoch": 42.11330935251799, "step": 46830, "torque_loss": 0.11891093105077744 }, { "epoch": 42.12230215827338, "grad_norm": 0.03489401564002037, "learning_rate": 1.088727273623924e-06, "loss": 0.0158, "step": 46840 }, { "action_loss": 0.0246401596814394, "epoch": 42.12230215827338, "step": 46840 }, { "epoch": 42.12230215827338, "step": 46840, "torque_loss": 0.2051825076341629 }, { "epoch": 42.131294964028775, "grad_norm": 0.03770482912659645, "learning_rate": 1.0818745813481356e-06, "loss": 0.0203, "step": 46850 }, { "action_loss": 0.0005022140685468912, "epoch": 42.131294964028775, "step": 46850 }, { "epoch": 42.131294964028775, "step": 46850, "torque_loss": 0.15586256980895996 }, { "epoch": 42.14028776978417, "grad_norm": 0.021493397653102875, "learning_rate": 1.0750432875223594e-06, "loss": 0.0148, "step": 46860 }, { "action_loss": 0.000673114147502929, "epoch": 42.14028776978417, "step": 46860 }, { "epoch": 42.14028776978417, "step": 46860, "torque_loss": 0.13748209178447723 }, { "epoch": 42.14928057553957, "grad_norm": 0.07116604596376419, "learning_rate": 1.0682333951348323e-06, "loss": 0.0191, "step": 46870 }, { "action_loss": 0.0017522796988487244, "epoch": 42.14928057553957, "step": 46870 }, { "epoch": 42.14928057553957, "step": 46870, "torque_loss": 0.1792955994606018 }, { "epoch": 42.15827338129496, "grad_norm": 0.053214747458696365, "learning_rate": 1.0614449071644272e-06, "loss": 0.0168, "step": 46880 }, { "action_loss": 0.0016748033231124282, "epoch": 42.15827338129496, "step": 46880 }, { "epoch": 42.15827338129496, "step": 46880, "torque_loss": 0.12030744552612305 }, { "epoch": 42.16726618705036, "grad_norm": 0.03556716442108154, "learning_rate": 1.054677826580669e-06, "loss": 0.018, "step": 46890 }, { "action_loss": 0.0012432554503902793, "epoch": 42.16726618705036, "step": 46890 }, { "epoch": 42.16726618705036, "step": 46890, "torque_loss": 0.1417139172554016 }, { "epoch": 42.17625899280576, "grad_norm": 0.034174609929323196, "learning_rate": 1.0479321563437062e-06, "loss": 0.0165, "step": 46900 }, { "action_loss": 0.000766729936003685, "epoch": 42.17625899280576, "step": 46900 }, { "epoch": 42.17625899280576, "step": 46900, "torque_loss": 0.13172318041324615 }, { "epoch": 42.185251798561154, "grad_norm": 0.10231893509626389, "learning_rate": 1.0412078994043172e-06, "loss": 0.0204, "step": 46910 }, { "action_loss": 0.0009914793772622943, "epoch": 42.185251798561154, "step": 46910 }, { "epoch": 42.185251798561154, "step": 46910, "torque_loss": 0.13838781416416168 }, { "epoch": 42.194244604316545, "grad_norm": 0.030078956857323647, "learning_rate": 1.0345050587039273e-06, "loss": 0.018, "step": 46920 }, { "action_loss": 0.0008496935479342937, "epoch": 42.194244604316545, "step": 46920 }, { "epoch": 42.194244604316545, "step": 46920, "torque_loss": 0.152640700340271 }, { "epoch": 42.20323741007194, "grad_norm": 0.030644584447145462, "learning_rate": 1.0278236371745854e-06, "loss": 0.0161, "step": 46930 }, { "action_loss": 0.005384494084864855, "epoch": 42.20323741007194, "step": 46930 }, { "epoch": 42.20323741007194, "step": 46930, "torque_loss": 0.21834023296833038 }, { "epoch": 42.21223021582734, "grad_norm": 0.030260369181632996, "learning_rate": 1.02116363773897e-06, "loss": 0.0186, "step": 46940 }, { "action_loss": 0.01453391183167696, "epoch": 42.21223021582734, "step": 46940 }, { "epoch": 42.21223021582734, "step": 46940, "torque_loss": 0.21346421539783478 }, { "epoch": 42.22122302158273, "grad_norm": 0.0626397505402565, "learning_rate": 1.0145250633103954e-06, "loss": 0.0189, "step": 46950 }, { "action_loss": 0.0013940612552687526, "epoch": 42.22122302158273, "step": 46950 }, { "epoch": 42.22122302158273, "step": 46950, "torque_loss": 0.10242140293121338 }, { "epoch": 42.23021582733813, "grad_norm": 0.04490174353122711, "learning_rate": 1.007907916792794e-06, "loss": 0.0205, "step": 46960 }, { "action_loss": 0.0009623367222957313, "epoch": 42.23021582733813, "step": 46960 }, { "epoch": 42.23021582733813, "step": 46960, "torque_loss": 0.14711223542690277 }, { "epoch": 42.239208633093526, "grad_norm": 0.047330763190984726, "learning_rate": 1.001312201080734e-06, "loss": 0.0187, "step": 46970 }, { "action_loss": 0.0005305373924784362, "epoch": 42.239208633093526, "step": 46970 }, { "epoch": 42.239208633093526, "step": 46970, "torque_loss": 0.1252501755952835 }, { "epoch": 42.24820143884892, "grad_norm": 0.07287359237670898, "learning_rate": 9.947379190594075e-07, "loss": 0.0178, "step": 46980 }, { "action_loss": 0.0023587741889059544, "epoch": 42.24820143884892, "step": 46980 }, { "epoch": 42.24820143884892, "step": 46980, "torque_loss": 0.136091947555542 }, { "epoch": 42.257194244604314, "grad_norm": 0.10181210935115814, "learning_rate": 9.88185073604625e-07, "loss": 0.0179, "step": 46990 }, { "action_loss": 0.00603514164686203, "epoch": 42.257194244604314, "step": 46990 }, { "epoch": 42.257194244604314, "step": 46990, "torque_loss": 0.17534656822681427 }, { "epoch": 42.26618705035971, "grad_norm": 0.035501908510923386, "learning_rate": 9.816536675828214e-07, "loss": 0.0186, "step": 47000 }, { "action_loss": 0.001357876230031252, "epoch": 42.26618705035971, "step": 47000 }, { "epoch": 42.26618705035971, "step": 47000, "torque_loss": 0.17393611371517181 }, { "epoch": 42.27517985611511, "grad_norm": 0.03217051178216934, "learning_rate": 9.751437038510615e-07, "loss": 0.0169, "step": 47010 }, { "action_loss": 0.0004643871507141739, "epoch": 42.27517985611511, "step": 47010 }, { "epoch": 42.27517985611511, "step": 47010, "torque_loss": 0.14181330800056458 }, { "epoch": 42.28417266187051, "grad_norm": 0.10946110635995865, "learning_rate": 9.686551852570225e-07, "loss": 0.017, "step": 47020 }, { "action_loss": 0.0009374544024467468, "epoch": 42.28417266187051, "step": 47020 }, { "epoch": 42.28417266187051, "step": 47020, "torque_loss": 0.14336508512496948 }, { "epoch": 42.2931654676259, "grad_norm": 0.040090784430503845, "learning_rate": 9.621881146390067e-07, "loss": 0.0184, "step": 47030 }, { "action_loss": 0.003020514501258731, "epoch": 42.2931654676259, "step": 47030 }, { "epoch": 42.2931654676259, "step": 47030, "torque_loss": 0.12268929928541183 }, { "epoch": 42.302158273381295, "grad_norm": 0.043631915003061295, "learning_rate": 9.557424948259175e-07, "loss": 0.0174, "step": 47040 }, { "action_loss": 0.004585853312164545, "epoch": 42.302158273381295, "step": 47040 }, { "epoch": 42.302158273381295, "step": 47040, "torque_loss": 0.17542581260204315 }, { "epoch": 42.31115107913669, "grad_norm": 0.1166117861866951, "learning_rate": 9.493183286373053e-07, "loss": 0.0183, "step": 47050 }, { "action_loss": 0.0035418858751654625, "epoch": 42.31115107913669, "step": 47050 }, { "epoch": 42.31115107913669, "step": 47050, "torque_loss": 0.1495177298784256 }, { "epoch": 42.32014388489208, "grad_norm": 0.07916712015867233, "learning_rate": 9.429156188833111e-07, "loss": 0.0157, "step": 47060 }, { "action_loss": 0.0013381773605942726, "epoch": 42.32014388489208, "step": 47060 }, { "epoch": 42.32014388489208, "step": 47060, "torque_loss": 0.15171168744564056 }, { "epoch": 42.32913669064748, "grad_norm": 0.03867485374212265, "learning_rate": 9.365343683646944e-07, "loss": 0.0185, "step": 47070 }, { "action_loss": 0.003134099766612053, "epoch": 42.32913669064748, "step": 47070 }, { "epoch": 42.32913669064748, "step": 47070, "torque_loss": 0.1481509953737259 }, { "epoch": 42.33812949640288, "grad_norm": 0.037911299616098404, "learning_rate": 9.301745798728334e-07, "loss": 0.0166, "step": 47080 }, { "action_loss": 0.0018640533089637756, "epoch": 42.33812949640288, "step": 47080 }, { "epoch": 42.33812949640288, "step": 47080, "torque_loss": 0.16590215265750885 }, { "epoch": 42.347122302158276, "grad_norm": 0.029441392049193382, "learning_rate": 9.238362561897141e-07, "loss": 0.0157, "step": 47090 }, { "action_loss": 0.0005234747659415007, "epoch": 42.347122302158276, "step": 47090 }, { "epoch": 42.347122302158276, "step": 47090, "torque_loss": 0.12652085721492767 }, { "epoch": 42.356115107913666, "grad_norm": 0.036503348499536514, "learning_rate": 9.17519400087935e-07, "loss": 0.0141, "step": 47100 }, { "action_loss": 0.0009836130775511265, "epoch": 42.356115107913666, "step": 47100 }, { "epoch": 42.356115107913666, "step": 47100, "torque_loss": 0.14767156541347504 }, { "epoch": 42.365107913669064, "grad_norm": 0.02915215492248535, "learning_rate": 9.112240143307138e-07, "loss": 0.0171, "step": 47110 }, { "action_loss": 0.003351312130689621, "epoch": 42.365107913669064, "step": 47110 }, { "epoch": 42.365107913669064, "step": 47110, "torque_loss": 0.19928598403930664 }, { "epoch": 42.37410071942446, "grad_norm": 0.04485519230365753, "learning_rate": 9.049501016718475e-07, "loss": 0.0178, "step": 47120 }, { "action_loss": 0.0006480918382294476, "epoch": 42.37410071942446, "step": 47120 }, { "epoch": 42.37410071942446, "step": 47120, "torque_loss": 0.14394600689411163 }, { "epoch": 42.38309352517986, "grad_norm": 0.05368902161717415, "learning_rate": 8.986976648557743e-07, "loss": 0.0192, "step": 47130 }, { "action_loss": 0.0006656521582044661, "epoch": 42.38309352517986, "step": 47130 }, { "epoch": 42.38309352517986, "step": 47130, "torque_loss": 0.10549262911081314 }, { "epoch": 42.39208633093525, "grad_norm": 0.04340372607111931, "learning_rate": 8.924667066175119e-07, "loss": 0.0174, "step": 47140 }, { "action_loss": 0.0015663960948586464, "epoch": 42.39208633093525, "step": 47140 }, { "epoch": 42.39208633093525, "step": 47140, "torque_loss": 0.12531165778636932 }, { "epoch": 42.40107913669065, "grad_norm": 0.025330442935228348, "learning_rate": 8.86257229682702e-07, "loss": 0.0178, "step": 47150 }, { "action_loss": 0.001039800699800253, "epoch": 42.40107913669065, "step": 47150 }, { "epoch": 42.40107913669065, "step": 47150, "torque_loss": 0.17329365015029907 }, { "epoch": 42.410071942446045, "grad_norm": 0.07471010088920593, "learning_rate": 8.800692367675722e-07, "loss": 0.0187, "step": 47160 }, { "action_loss": 0.003734142752364278, "epoch": 42.410071942446045, "step": 47160 }, { "epoch": 42.410071942446045, "step": 47160, "torque_loss": 0.13822750747203827 }, { "epoch": 42.419064748201436, "grad_norm": 0.06790333986282349, "learning_rate": 8.739027305789683e-07, "loss": 0.0153, "step": 47170 }, { "action_loss": 0.0005854289047420025, "epoch": 42.419064748201436, "step": 47170 }, { "epoch": 42.419064748201436, "step": 47170, "torque_loss": 0.1236548200249672 }, { "epoch": 42.42805755395683, "grad_norm": 0.0777805745601654, "learning_rate": 8.677577138143267e-07, "loss": 0.0198, "step": 47180 }, { "action_loss": 0.004280204884707928, "epoch": 42.42805755395683, "step": 47180 }, { "epoch": 42.42805755395683, "step": 47180, "torque_loss": 0.2095683068037033 }, { "epoch": 42.43705035971223, "grad_norm": 0.09523981809616089, "learning_rate": 8.616341891616864e-07, "loss": 0.0188, "step": 47190 }, { "action_loss": 0.0006304702255874872, "epoch": 42.43705035971223, "step": 47190 }, { "epoch": 42.43705035971223, "step": 47190, "torque_loss": 0.07150332629680634 }, { "epoch": 42.44604316546763, "grad_norm": 0.07041090726852417, "learning_rate": 8.555321592996824e-07, "loss": 0.0173, "step": 47200 }, { "action_loss": 0.0007813306874595582, "epoch": 42.44604316546763, "step": 47200 }, { "epoch": 42.44604316546763, "step": 47200, "torque_loss": 0.15661762654781342 }, { "epoch": 42.45503597122302, "grad_norm": 0.031056400388479233, "learning_rate": 8.494516268975572e-07, "loss": 0.0165, "step": 47210 }, { "action_loss": 0.0017905962886288762, "epoch": 42.45503597122302, "step": 47210 }, { "epoch": 42.45503597122302, "step": 47210, "torque_loss": 0.1703433245420456 }, { "epoch": 42.46402877697842, "grad_norm": 0.10515554994344711, "learning_rate": 8.433925946151333e-07, "loss": 0.0214, "step": 47220 }, { "action_loss": 0.003612136235460639, "epoch": 42.46402877697842, "step": 47220 }, { "epoch": 42.46402877697842, "step": 47220, "torque_loss": 0.17726387083530426 }, { "epoch": 42.473021582733814, "grad_norm": 0.033007413148880005, "learning_rate": 8.373550651028406e-07, "loss": 0.0173, "step": 47230 }, { "action_loss": 0.0009496815036982298, "epoch": 42.473021582733814, "step": 47230 }, { "epoch": 42.473021582733814, "step": 47230, "torque_loss": 0.08797717839479446 }, { "epoch": 42.48201438848921, "grad_norm": 0.13961336016654968, "learning_rate": 8.313390410017052e-07, "loss": 0.0207, "step": 47240 }, { "action_loss": 0.0005061663105152547, "epoch": 42.48201438848921, "step": 47240 }, { "epoch": 42.48201438848921, "step": 47240, "torque_loss": 0.11201699823141098 }, { "epoch": 42.4910071942446, "grad_norm": 0.04189762845635414, "learning_rate": 8.253445249433334e-07, "loss": 0.0186, "step": 47250 }, { "action_loss": 0.0009125465876422822, "epoch": 42.4910071942446, "step": 47250 }, { "epoch": 42.4910071942446, "step": 47250, "torque_loss": 0.1338486671447754 }, { "epoch": 42.5, "grad_norm": 0.04007969796657562, "learning_rate": 8.193715195499329e-07, "loss": 0.0167, "step": 47260 }, { "action_loss": 0.0009841141290962696, "epoch": 42.5, "step": 47260 }, { "epoch": 42.5, "step": 47260, "torque_loss": 0.17630882561206818 }, { "epoch": 42.5089928057554, "grad_norm": 0.02953289821743965, "learning_rate": 8.134200274342918e-07, "loss": 0.0172, "step": 47270 }, { "action_loss": 0.001301749492995441, "epoch": 42.5089928057554, "step": 47270 }, { "epoch": 42.5089928057554, "step": 47270, "torque_loss": 0.16996198892593384 }, { "epoch": 42.51798561151079, "grad_norm": 0.019814690575003624, "learning_rate": 8.074900511998107e-07, "loss": 0.0249, "step": 47280 }, { "action_loss": 0.002863869071006775, "epoch": 42.51798561151079, "step": 47280 }, { "epoch": 42.51798561151079, "step": 47280, "torque_loss": 0.15299619734287262 }, { "epoch": 42.526978417266186, "grad_norm": 0.08844815194606781, "learning_rate": 8.01581593440448e-07, "loss": 0.0156, "step": 47290 }, { "action_loss": 0.012838907539844513, "epoch": 42.526978417266186, "step": 47290 }, { "epoch": 42.526978417266186, "step": 47290, "torque_loss": 0.22119511663913727 }, { "epoch": 42.53597122302158, "grad_norm": 0.027599235996603966, "learning_rate": 7.956946567407753e-07, "loss": 0.0208, "step": 47300 }, { "action_loss": 0.0004991675377823412, "epoch": 42.53597122302158, "step": 47300 }, { "epoch": 42.53597122302158, "step": 47300, "torque_loss": 0.11084584146738052 }, { "epoch": 42.54496402877698, "grad_norm": 0.04414723441004753, "learning_rate": 7.898292436759324e-07, "loss": 0.0208, "step": 47310 }, { "action_loss": 0.002126897219568491, "epoch": 42.54496402877698, "step": 47310 }, { "epoch": 42.54496402877698, "step": 47310, "torque_loss": 0.16549570858478546 }, { "epoch": 42.55395683453237, "grad_norm": 0.03478928655385971, "learning_rate": 7.839853568116506e-07, "loss": 0.0151, "step": 47320 }, { "action_loss": 0.0010206860024482012, "epoch": 42.55395683453237, "step": 47320 }, { "epoch": 42.55395683453237, "step": 47320, "torque_loss": 0.12860800325870514 }, { "epoch": 42.56294964028777, "grad_norm": 0.06136491894721985, "learning_rate": 7.78162998704246e-07, "loss": 0.0186, "step": 47330 }, { "action_loss": 0.009515383280813694, "epoch": 42.56294964028777, "step": 47330 }, { "epoch": 42.56294964028777, "step": 47330, "torque_loss": 0.2807491421699524 }, { "epoch": 42.57194244604317, "grad_norm": 0.029560502618551254, "learning_rate": 7.723621719006147e-07, "loss": 0.0204, "step": 47340 }, { "action_loss": 0.00042540751746855676, "epoch": 42.57194244604317, "step": 47340 }, { "epoch": 42.57194244604317, "step": 47340, "torque_loss": 0.1250023990869522 }, { "epoch": 42.580935251798564, "grad_norm": 0.03350398689508438, "learning_rate": 7.665828789382323e-07, "loss": 0.017, "step": 47350 }, { "action_loss": 0.0020659074652940035, "epoch": 42.580935251798564, "step": 47350 }, { "epoch": 42.580935251798564, "step": 47350, "torque_loss": 0.21816504001617432 }, { "epoch": 42.589928057553955, "grad_norm": 0.111976258456707, "learning_rate": 7.608251223451602e-07, "loss": 0.0225, "step": 47360 }, { "action_loss": 0.001509068883024156, "epoch": 42.589928057553955, "step": 47360 }, { "epoch": 42.589928057553955, "step": 47360, "torque_loss": 0.13544170558452606 }, { "epoch": 42.59892086330935, "grad_norm": 0.05605824664235115, "learning_rate": 7.550889046400444e-07, "loss": 0.0158, "step": 47370 }, { "action_loss": 0.0023058478254824877, "epoch": 42.59892086330935, "step": 47370 }, { "epoch": 42.59892086330935, "step": 47370, "torque_loss": 0.22646625339984894 }, { "epoch": 42.60791366906475, "grad_norm": 0.04083345830440521, "learning_rate": 7.493742283320893e-07, "loss": 0.0185, "step": 47380 }, { "action_loss": 0.010015767067670822, "epoch": 42.60791366906475, "step": 47380 }, { "epoch": 42.60791366906475, "step": 47380, "torque_loss": 0.20994555950164795 }, { "epoch": 42.61690647482014, "grad_norm": 0.06260521709918976, "learning_rate": 7.43681095921095e-07, "loss": 0.0184, "step": 47390 }, { "action_loss": 0.0031583011150360107, "epoch": 42.61690647482014, "step": 47390 }, { "epoch": 42.61690647482014, "step": 47390, "torque_loss": 0.1582719385623932 }, { "epoch": 42.62589928057554, "grad_norm": 0.02384527400135994, "learning_rate": 7.380095098974305e-07, "loss": 0.0155, "step": 47400 }, { "action_loss": 0.0014924347633495927, "epoch": 42.62589928057554, "step": 47400 }, { "epoch": 42.62589928057554, "step": 47400, "torque_loss": 0.12955044209957123 }, { "epoch": 42.634892086330936, "grad_norm": 0.03963591158390045, "learning_rate": 7.323594727420446e-07, "loss": 0.0165, "step": 47410 }, { "action_loss": 0.0013576395576819777, "epoch": 42.634892086330936, "step": 47410 }, { "epoch": 42.634892086330936, "step": 47410, "torque_loss": 0.19447767734527588 }, { "epoch": 42.643884892086334, "grad_norm": 0.026276178658008575, "learning_rate": 7.267309869264438e-07, "loss": 0.0169, "step": 47420 }, { "action_loss": 0.0006411370704881847, "epoch": 42.643884892086334, "step": 47420 }, { "epoch": 42.643884892086334, "step": 47420, "torque_loss": 0.150314599275589 }, { "epoch": 42.652877697841724, "grad_norm": 0.03978276252746582, "learning_rate": 7.211240549127252e-07, "loss": 0.0185, "step": 47430 }, { "action_loss": 0.0008743090438656509, "epoch": 42.652877697841724, "step": 47430 }, { "epoch": 42.652877697841724, "step": 47430, "torque_loss": 0.1309385746717453 }, { "epoch": 42.66187050359712, "grad_norm": 0.05082136020064354, "learning_rate": 7.155386791535546e-07, "loss": 0.0179, "step": 47440 }, { "action_loss": 0.03103605844080448, "epoch": 42.66187050359712, "step": 47440 }, { "epoch": 42.66187050359712, "step": 47440, "torque_loss": 0.16127635538578033 }, { "epoch": 42.67086330935252, "grad_norm": 0.03678048402070999, "learning_rate": 7.09974862092161e-07, "loss": 0.0256, "step": 47450 }, { "action_loss": 0.005368551705032587, "epoch": 42.67086330935252, "step": 47450 }, { "epoch": 42.67086330935252, "step": 47450, "torque_loss": 0.2260650247335434 }, { "epoch": 42.67985611510792, "grad_norm": 0.04437785968184471, "learning_rate": 7.044326061623529e-07, "loss": 0.0155, "step": 47460 }, { "action_loss": 0.0006068962975405157, "epoch": 42.67985611510792, "step": 47460 }, { "epoch": 42.67985611510792, "step": 47460, "torque_loss": 0.08740528672933578 }, { "epoch": 42.68884892086331, "grad_norm": 0.03082597628235817, "learning_rate": 6.989119137884914e-07, "loss": 0.0172, "step": 47470 }, { "action_loss": 0.0005421272944658995, "epoch": 42.68884892086331, "step": 47470 }, { "epoch": 42.68884892086331, "step": 47470, "torque_loss": 0.17903117835521698 }, { "epoch": 42.697841726618705, "grad_norm": 0.028605835512280464, "learning_rate": 6.934127873855223e-07, "loss": 0.016, "step": 47480 }, { "action_loss": 0.0009121696348302066, "epoch": 42.697841726618705, "step": 47480 }, { "epoch": 42.697841726618705, "step": 47480, "torque_loss": 0.15829800069332123 }, { "epoch": 42.7068345323741, "grad_norm": 0.04404692351818085, "learning_rate": 6.879352293589492e-07, "loss": 0.0157, "step": 47490 }, { "action_loss": 0.0020772006828337908, "epoch": 42.7068345323741, "step": 47490 }, { "epoch": 42.7068345323741, "step": 47490, "torque_loss": 0.16722674667835236 }, { "epoch": 42.71582733812949, "grad_norm": 0.04051554948091507, "learning_rate": 6.824792421048388e-07, "loss": 0.0229, "step": 47500 }, { "action_loss": 0.00043166568502783775, "epoch": 42.71582733812949, "step": 47500 }, { "epoch": 42.71582733812949, "step": 47500, "torque_loss": 0.10944336652755737 }, { "epoch": 42.72482014388489, "grad_norm": 0.04127798229455948, "learning_rate": 6.770448280098319e-07, "loss": 0.0174, "step": 47510 }, { "action_loss": 0.0008933879435062408, "epoch": 42.72482014388489, "step": 47510 }, { "epoch": 42.72482014388489, "step": 47510, "torque_loss": 0.10303286463022232 }, { "epoch": 42.73381294964029, "grad_norm": 0.024491753429174423, "learning_rate": 6.716319894511158e-07, "loss": 0.0181, "step": 47520 }, { "action_loss": 0.0008199385483749211, "epoch": 42.73381294964029, "step": 47520 }, { "epoch": 42.73381294964029, "step": 47520, "torque_loss": 0.12704892456531525 }, { "epoch": 42.742805755395686, "grad_norm": 0.026178529486060143, "learning_rate": 6.662407287964634e-07, "loss": 0.0182, "step": 47530 }, { "action_loss": 0.00045187282375991344, "epoch": 42.742805755395686, "step": 47530 }, { "epoch": 42.742805755395686, "step": 47530, "torque_loss": 0.10840103775262833 }, { "epoch": 42.75179856115108, "grad_norm": 0.031164752319455147, "learning_rate": 6.608710484041824e-07, "loss": 0.0155, "step": 47540 }, { "action_loss": 0.0011267256923019886, "epoch": 42.75179856115108, "step": 47540 }, { "epoch": 42.75179856115108, "step": 47540, "torque_loss": 0.15228047966957092 }, { "epoch": 42.760791366906474, "grad_norm": 0.04692116007208824, "learning_rate": 6.555229506231608e-07, "loss": 0.0164, "step": 47550 }, { "action_loss": 0.0011702032061293721, "epoch": 42.760791366906474, "step": 47550 }, { "epoch": 42.760791366906474, "step": 47550, "torque_loss": 0.1112443208694458 }, { "epoch": 42.76978417266187, "grad_norm": 0.030441127717494965, "learning_rate": 6.501964377928382e-07, "loss": 0.0184, "step": 47560 }, { "action_loss": 0.0015076400013640523, "epoch": 42.76978417266187, "step": 47560 }, { "epoch": 42.76978417266187, "step": 47560, "torque_loss": 0.14589889347553253 }, { "epoch": 42.77877697841727, "grad_norm": 0.11555515229701996, "learning_rate": 6.448915122432064e-07, "loss": 0.0191, "step": 47570 }, { "action_loss": 0.00045646014041267335, "epoch": 42.77877697841727, "step": 47570 }, { "epoch": 42.77877697841727, "step": 47570, "torque_loss": 0.13099803030490875 }, { "epoch": 42.78776978417266, "grad_norm": 0.09893430769443512, "learning_rate": 6.396081762948258e-07, "loss": 0.0192, "step": 47580 }, { "action_loss": 0.0005818729405291378, "epoch": 42.78776978417266, "step": 47580 }, { "epoch": 42.78776978417266, "step": 47580, "torque_loss": 0.08853762596845627 }, { "epoch": 42.79676258992806, "grad_norm": 0.02596547082066536, "learning_rate": 6.343464322588033e-07, "loss": 0.0183, "step": 47590 }, { "action_loss": 0.0011691171675920486, "epoch": 42.79676258992806, "step": 47590 }, { "epoch": 42.79676258992806, "step": 47590, "torque_loss": 0.14267989993095398 }, { "epoch": 42.805755395683455, "grad_norm": 0.02359190210700035, "learning_rate": 6.291062824368143e-07, "loss": 0.0164, "step": 47600 }, { "action_loss": 0.005304525140672922, "epoch": 42.805755395683455, "step": 47600 }, { "epoch": 42.805755395683455, "step": 47600, "torque_loss": 0.15095144510269165 }, { "epoch": 42.814748201438846, "grad_norm": 0.03027995675802231, "learning_rate": 6.238877291210587e-07, "loss": 0.017, "step": 47610 }, { "action_loss": 0.000492954917717725, "epoch": 42.814748201438846, "step": 47610 }, { "epoch": 42.814748201438846, "step": 47610, "torque_loss": 0.15368609130382538 }, { "epoch": 42.82374100719424, "grad_norm": 0.13457319140434265, "learning_rate": 6.186907745943271e-07, "loss": 0.0221, "step": 47620 }, { "action_loss": 0.0007983970572240651, "epoch": 42.82374100719424, "step": 47620 }, { "epoch": 42.82374100719424, "step": 47620, "torque_loss": 0.14697761833667755 }, { "epoch": 42.83273381294964, "grad_norm": 0.0264405719935894, "learning_rate": 6.135154211299399e-07, "loss": 0.0174, "step": 47630 }, { "action_loss": 0.0012300109956413507, "epoch": 42.83273381294964, "step": 47630 }, { "epoch": 42.83273381294964, "step": 47630, "torque_loss": 0.11403024196624756 }, { "epoch": 42.84172661870504, "grad_norm": 0.04176852107048035, "learning_rate": 6.083616709917639e-07, "loss": 0.0157, "step": 47640 }, { "action_loss": 0.0035921011585742235, "epoch": 42.84172661870504, "step": 47640 }, { "epoch": 42.84172661870504, "step": 47640, "torque_loss": 0.20122845470905304 }, { "epoch": 42.85071942446043, "grad_norm": 0.04224685952067375, "learning_rate": 6.032295264342403e-07, "loss": 0.0192, "step": 47650 }, { "action_loss": 0.0024473227094858885, "epoch": 42.85071942446043, "step": 47650 }, { "epoch": 42.85071942446043, "step": 47650, "torque_loss": 0.19142551720142365 }, { "epoch": 42.85971223021583, "grad_norm": 0.035518523305654526, "learning_rate": 5.981189897023287e-07, "loss": 0.0194, "step": 47660 }, { "action_loss": 0.0005453568301163614, "epoch": 42.85971223021583, "step": 47660 }, { "epoch": 42.85971223021583, "step": 47660, "torque_loss": 0.10283675044775009 }, { "epoch": 42.868705035971225, "grad_norm": 0.035268787294626236, "learning_rate": 5.930300630315522e-07, "loss": 0.0168, "step": 47670 }, { "action_loss": 0.0019035536097362638, "epoch": 42.868705035971225, "step": 47670 }, { "epoch": 42.868705035971225, "step": 47670, "torque_loss": 0.16989056766033173 }, { "epoch": 42.87769784172662, "grad_norm": 0.058224692940711975, "learning_rate": 5.879627486479911e-07, "loss": 0.0172, "step": 47680 }, { "action_loss": 0.001923002884723246, "epoch": 42.87769784172662, "step": 47680 }, { "epoch": 42.87769784172662, "step": 47680, "torque_loss": 0.16992513835430145 }, { "epoch": 42.88669064748201, "grad_norm": 0.09302951395511627, "learning_rate": 5.829170487682556e-07, "loss": 0.0222, "step": 47690 }, { "action_loss": 0.00212832516990602, "epoch": 42.88669064748201, "step": 47690 }, { "epoch": 42.88669064748201, "step": 47690, "torque_loss": 0.18614345788955688 }, { "epoch": 42.89568345323741, "grad_norm": 0.04738694429397583, "learning_rate": 5.778929655995025e-07, "loss": 0.0195, "step": 47700 }, { "action_loss": 0.0024436728563159704, "epoch": 42.89568345323741, "step": 47700 }, { "epoch": 42.89568345323741, "step": 47700, "torque_loss": 0.1635904759168625 }, { "epoch": 42.90467625899281, "grad_norm": 0.08641044050455093, "learning_rate": 5.728905013394403e-07, "loss": 0.0185, "step": 47710 }, { "action_loss": 0.0007271857466548681, "epoch": 42.90467625899281, "step": 47710 }, { "epoch": 42.90467625899281, "step": 47710, "torque_loss": 0.13316848874092102 }, { "epoch": 42.9136690647482, "grad_norm": 0.02761390060186386, "learning_rate": 5.679096581763188e-07, "loss": 0.02, "step": 47720 }, { "action_loss": 0.002065597102046013, "epoch": 42.9136690647482, "step": 47720 }, { "epoch": 42.9136690647482, "step": 47720, "torque_loss": 0.16196106374263763 }, { "epoch": 42.922661870503596, "grad_norm": 0.046861015260219574, "learning_rate": 5.629504382889173e-07, "loss": 0.018, "step": 47730 }, { "action_loss": 0.0004606572911143303, "epoch": 42.922661870503596, "step": 47730 }, { "epoch": 42.922661870503596, "step": 47730, "torque_loss": 0.08646509796380997 }, { "epoch": 42.931654676258994, "grad_norm": 0.03345511108636856, "learning_rate": 5.580128438465837e-07, "loss": 0.0188, "step": 47740 }, { "action_loss": 0.011776584200561047, "epoch": 42.931654676258994, "step": 47740 }, { "epoch": 42.931654676258994, "step": 47740, "torque_loss": 0.21613240242004395 }, { "epoch": 42.94064748201439, "grad_norm": 0.03891528397798538, "learning_rate": 5.530968770091737e-07, "loss": 0.022, "step": 47750 }, { "action_loss": 0.0036437201779335737, "epoch": 42.94064748201439, "step": 47750 }, { "epoch": 42.94064748201439, "step": 47750, "torque_loss": 0.23435823619365692 }, { "epoch": 42.94964028776978, "grad_norm": 0.15161119401454926, "learning_rate": 5.482025399271063e-07, "loss": 0.0221, "step": 47760 }, { "action_loss": 0.0009181545465253294, "epoch": 42.94964028776978, "step": 47760 }, { "epoch": 42.94964028776978, "step": 47760, "torque_loss": 0.19641327857971191 }, { "epoch": 42.95863309352518, "grad_norm": 0.052150122821331024, "learning_rate": 5.433298347413296e-07, "loss": 0.019, "step": 47770 }, { "action_loss": 0.0004265161696821451, "epoch": 42.95863309352518, "step": 47770 }, { "epoch": 42.95863309352518, "step": 47770, "torque_loss": 0.16746412217617035 }, { "epoch": 42.96762589928058, "grad_norm": 0.023786330595612526, "learning_rate": 5.384787635833332e-07, "loss": 0.0168, "step": 47780 }, { "action_loss": 0.0008694872376509011, "epoch": 42.96762589928058, "step": 47780 }, { "epoch": 42.96762589928058, "step": 47780, "torque_loss": 0.12419164925813675 }, { "epoch": 42.976618705035975, "grad_norm": 0.03918415680527687, "learning_rate": 5.33649328575142e-07, "loss": 0.0169, "step": 47790 }, { "action_loss": 0.001176691148430109, "epoch": 42.976618705035975, "step": 47790 }, { "epoch": 42.976618705035975, "step": 47790, "torque_loss": 0.18030281364917755 }, { "epoch": 42.985611510791365, "grad_norm": 0.049085646867752075, "learning_rate": 5.288415318292994e-07, "loss": 0.0167, "step": 47800 }, { "action_loss": 0.0010067867115139961, "epoch": 42.985611510791365, "step": 47800 }, { "epoch": 42.985611510791365, "step": 47800, "torque_loss": 0.16146445274353027 }, { "epoch": 42.99460431654676, "grad_norm": 0.082651786506176, "learning_rate": 5.24055375448923e-07, "loss": 0.0176, "step": 47810 }, { "action_loss": 0.0006994024734012783, "epoch": 42.99460431654676, "step": 47810 }, { "epoch": 42.99460431654676, "step": 47810, "torque_loss": 0.09114549309015274 }, { "epoch": 43.00359712230216, "grad_norm": 0.023561453446745872, "learning_rate": 5.192908615276215e-07, "loss": 0.0141, "step": 47820 }, { "action_loss": 0.0018968874355778098, "epoch": 43.00359712230216, "step": 47820 }, { "epoch": 43.00359712230216, "step": 47820, "torque_loss": 0.13607211410999298 }, { "epoch": 43.01258992805755, "grad_norm": 0.04875337705016136, "learning_rate": 5.145479921495556e-07, "loss": 0.0178, "step": 47830 }, { "action_loss": 0.009544781409204006, "epoch": 43.01258992805755, "step": 47830 }, { "epoch": 43.01258992805755, "step": 47830, "torque_loss": 0.22543865442276 }, { "epoch": 43.02158273381295, "grad_norm": 0.04227053001523018, "learning_rate": 5.098267693894321e-07, "loss": 0.0235, "step": 47840 }, { "action_loss": 0.0007961521041579545, "epoch": 43.02158273381295, "step": 47840 }, { "epoch": 43.02158273381295, "step": 47840, "torque_loss": 0.08778093010187149 }, { "epoch": 43.030575539568346, "grad_norm": 0.12689907848834991, "learning_rate": 5.051271953124603e-07, "loss": 0.0184, "step": 47850 }, { "action_loss": 0.0020037421490997076, "epoch": 43.030575539568346, "step": 47850 }, { "epoch": 43.030575539568346, "step": 47850, "torque_loss": 0.12191643565893173 }, { "epoch": 43.039568345323744, "grad_norm": 0.09605760872364044, "learning_rate": 5.004492719744014e-07, "loss": 0.018, "step": 47860 }, { "action_loss": 0.01009285356849432, "epoch": 43.039568345323744, "step": 47860 }, { "epoch": 43.039568345323744, "step": 47860, "torque_loss": 0.2804848253726959 }, { "epoch": 43.048561151079134, "grad_norm": 0.061013706028461456, "learning_rate": 4.957930014215351e-07, "loss": 0.0173, "step": 47870 }, { "action_loss": 0.0016770838992670178, "epoch": 43.048561151079134, "step": 47870 }, { "epoch": 43.048561151079134, "step": 47870, "torque_loss": 0.16159754991531372 }, { "epoch": 43.05755395683453, "grad_norm": 0.03759870305657387, "learning_rate": 4.91158385690671e-07, "loss": 0.0181, "step": 47880 }, { "action_loss": 0.0005182063905522227, "epoch": 43.05755395683453, "step": 47880 }, { "epoch": 43.05755395683453, "step": 47880, "torque_loss": 0.07004943490028381 }, { "epoch": 43.06654676258993, "grad_norm": 0.025336094200611115, "learning_rate": 4.865454268091485e-07, "loss": 0.0171, "step": 47890 }, { "action_loss": 0.0011444945121183991, "epoch": 43.06654676258993, "step": 47890 }, { "epoch": 43.06654676258993, "step": 47890, "torque_loss": 0.15201306343078613 }, { "epoch": 43.07553956834533, "grad_norm": 0.03226742520928383, "learning_rate": 4.819541267948313e-07, "loss": 0.0177, "step": 47900 }, { "action_loss": 0.001532218069769442, "epoch": 43.07553956834533, "step": 47900 }, { "epoch": 43.07553956834533, "step": 47900, "torque_loss": 0.19413460791110992 }, { "epoch": 43.08453237410072, "grad_norm": 0.041396364569664, "learning_rate": 4.773844876561129e-07, "loss": 0.0162, "step": 47910 }, { "action_loss": 0.0008676841971464455, "epoch": 43.08453237410072, "step": 47910 }, { "epoch": 43.08453237410072, "step": 47910, "torque_loss": 0.11434301733970642 }, { "epoch": 43.093525179856115, "grad_norm": 0.037382904440164566, "learning_rate": 4.7283651139191086e-07, "loss": 0.0186, "step": 47920 }, { "action_loss": 0.0008622154709883034, "epoch": 43.093525179856115, "step": 47920 }, { "epoch": 43.093525179856115, "step": 47920, "torque_loss": 0.16869527101516724 }, { "epoch": 43.10251798561151, "grad_norm": 0.04409138858318329, "learning_rate": 4.6831019999165617e-07, "loss": 0.0168, "step": 47930 }, { "action_loss": 0.0011966516030952334, "epoch": 43.10251798561151, "step": 47930 }, { "epoch": 43.10251798561151, "step": 47930, "torque_loss": 0.16078825294971466 }, { "epoch": 43.111510791366904, "grad_norm": 0.05183008313179016, "learning_rate": 4.6380555543532045e-07, "loss": 0.0162, "step": 47940 }, { "action_loss": 0.00037061702460050583, "epoch": 43.111510791366904, "step": 47940 }, { "epoch": 43.111510791366904, "step": 47940, "torque_loss": 0.09414634853601456 }, { "epoch": 43.1205035971223, "grad_norm": 0.032611239701509476, "learning_rate": 4.5932257969338313e-07, "loss": 0.0175, "step": 47950 }, { "action_loss": 0.0032834492158144712, "epoch": 43.1205035971223, "step": 47950 }, { "epoch": 43.1205035971223, "step": 47950, "torque_loss": 0.13807223737239838 }, { "epoch": 43.1294964028777, "grad_norm": 0.02632395550608635, "learning_rate": 4.5486127472684215e-07, "loss": 0.0165, "step": 47960 }, { "action_loss": 0.000839745334815234, "epoch": 43.1294964028777, "step": 47960 }, { "epoch": 43.1294964028777, "step": 47960, "torque_loss": 0.14528189599514008 }, { "epoch": 43.138489208633096, "grad_norm": 0.09842406958341599, "learning_rate": 4.5042164248724205e-07, "loss": 0.0192, "step": 47970 }, { "action_loss": 0.0009692104067653418, "epoch": 43.138489208633096, "step": 47970 }, { "epoch": 43.138489208633096, "step": 47970, "torque_loss": 0.1734147071838379 }, { "epoch": 43.14748201438849, "grad_norm": 0.0772879421710968, "learning_rate": 4.4600368491661826e-07, "loss": 0.0198, "step": 47980 }, { "action_loss": 0.000576717488002032, "epoch": 43.14748201438849, "step": 47980 }, { "epoch": 43.14748201438849, "step": 47980, "torque_loss": 0.09333185106515884 }, { "epoch": 43.156474820143885, "grad_norm": 0.03448014706373215, "learning_rate": 4.4160740394753595e-07, "loss": 0.0187, "step": 47990 }, { "action_loss": 0.0005668638623319566, "epoch": 43.156474820143885, "step": 47990 }, { "epoch": 43.156474820143885, "step": 47990, "torque_loss": 0.17332549393177032 }, { "epoch": 43.16546762589928, "grad_norm": 0.020259426906704903, "learning_rate": 4.3723280150307357e-07, "loss": 0.0188, "step": 48000 }, { "action_loss": 0.007072234991937876, "epoch": 43.16546762589928, "step": 48000 }, { "epoch": 43.16546762589928, "step": 48000, "torque_loss": 0.24280939996242523 }, { "epoch": 43.17446043165467, "grad_norm": 0.029815975576639175, "learning_rate": 4.3287987949683917e-07, "loss": 0.0183, "step": 48010 }, { "action_loss": 0.003222050843760371, "epoch": 43.17446043165467, "step": 48010 }, { "epoch": 43.17446043165467, "step": 48010, "torque_loss": 0.18823380768299103 }, { "epoch": 43.18345323741007, "grad_norm": 0.040959682315588, "learning_rate": 4.285486398329486e-07, "loss": 0.0184, "step": 48020 }, { "action_loss": 0.0017634726827964187, "epoch": 43.18345323741007, "step": 48020 }, { "epoch": 43.18345323741007, "step": 48020, "torque_loss": 0.1379438191652298 }, { "epoch": 43.19244604316547, "grad_norm": 0.03962277993559837, "learning_rate": 4.242390844060251e-07, "loss": 0.017, "step": 48030 }, { "action_loss": 0.0008538918918929994, "epoch": 43.19244604316547, "step": 48030 }, { "epoch": 43.19244604316547, "step": 48030, "torque_loss": 0.15558548271656036 }, { "epoch": 43.201438848920866, "grad_norm": 0.15824949741363525, "learning_rate": 4.1995121510122193e-07, "loss": 0.0234, "step": 48040 }, { "action_loss": 0.0009954865090548992, "epoch": 43.201438848920866, "step": 48040 }, { "epoch": 43.201438848920866, "step": 48040, "torque_loss": 0.10510730743408203 }, { "epoch": 43.210431654676256, "grad_norm": 0.03201127424836159, "learning_rate": 4.156850337941942e-07, "loss": 0.022, "step": 48050 }, { "action_loss": 0.001279507065191865, "epoch": 43.210431654676256, "step": 48050 }, { "epoch": 43.210431654676256, "step": 48050, "torque_loss": 0.13946181535720825 }, { "epoch": 43.219424460431654, "grad_norm": 0.022770432755351067, "learning_rate": 4.114405423511214e-07, "loss": 0.0128, "step": 48060 }, { "action_loss": 0.00278556440025568, "epoch": 43.219424460431654, "step": 48060 }, { "epoch": 43.219424460431654, "step": 48060, "torque_loss": 0.15976743400096893 }, { "epoch": 43.22841726618705, "grad_norm": 0.037578683346509933, "learning_rate": 4.072177426286794e-07, "loss": 0.0207, "step": 48070 }, { "action_loss": 0.0007809217204339802, "epoch": 43.22841726618705, "step": 48070 }, { "epoch": 43.22841726618705, "step": 48070, "torque_loss": 0.16807220876216888 }, { "epoch": 43.23741007194245, "grad_norm": 0.037433356046676636, "learning_rate": 4.0301663647407396e-07, "loss": 0.015, "step": 48080 }, { "action_loss": 0.0008946121670305729, "epoch": 43.23741007194245, "step": 48080 }, { "epoch": 43.23741007194245, "step": 48080, "torque_loss": 0.17750954627990723 }, { "epoch": 43.24640287769784, "grad_norm": 0.02057844027876854, "learning_rate": 3.988372257249962e-07, "loss": 0.0172, "step": 48090 }, { "action_loss": 0.004839766770601273, "epoch": 43.24640287769784, "step": 48090 }, { "epoch": 43.24640287769784, "step": 48090, "torque_loss": 0.19890253245830536 }, { "epoch": 43.25539568345324, "grad_norm": 0.024019910022616386, "learning_rate": 3.946795122096836e-07, "loss": 0.0174, "step": 48100 }, { "action_loss": 0.002002333989366889, "epoch": 43.25539568345324, "step": 48100 }, { "epoch": 43.25539568345324, "step": 48100, "torque_loss": 0.16349905729293823 }, { "epoch": 43.264388489208635, "grad_norm": 0.023377006873488426, "learning_rate": 3.9054349774684254e-07, "loss": 0.0189, "step": 48110 }, { "action_loss": 0.003176972968503833, "epoch": 43.264388489208635, "step": 48110 }, { "epoch": 43.264388489208635, "step": 48110, "torque_loss": 0.1807045191526413 }, { "epoch": 43.273381294964025, "grad_norm": 0.036625053733587265, "learning_rate": 3.864291841457146e-07, "loss": 0.0156, "step": 48120 }, { "action_loss": 0.0016562985256314278, "epoch": 43.273381294964025, "step": 48120 }, { "epoch": 43.273381294964025, "step": 48120, "torque_loss": 0.19835810363292694 }, { "epoch": 43.28237410071942, "grad_norm": 0.0688944011926651, "learning_rate": 3.8233657320604355e-07, "loss": 0.0159, "step": 48130 }, { "action_loss": 0.0016519906930625439, "epoch": 43.28237410071942, "step": 48130 }, { "epoch": 43.28237410071942, "step": 48130, "torque_loss": 0.15055201947689056 }, { "epoch": 43.29136690647482, "grad_norm": 0.021146269515156746, "learning_rate": 3.782656667180695e-07, "loss": 0.0197, "step": 48140 }, { "action_loss": 0.0027665328234434128, "epoch": 43.29136690647482, "step": 48140 }, { "epoch": 43.29136690647482, "step": 48140, "torque_loss": 0.22693419456481934 }, { "epoch": 43.30035971223022, "grad_norm": 0.036260634660720825, "learning_rate": 3.742164664625458e-07, "loss": 0.0189, "step": 48150 }, { "action_loss": 0.00042287760879844427, "epoch": 43.30035971223022, "step": 48150 }, { "epoch": 43.30035971223022, "step": 48150, "torque_loss": 0.11670053005218506 }, { "epoch": 43.30935251798561, "grad_norm": 0.05012710019946098, "learning_rate": 3.7018897421073337e-07, "loss": 0.0159, "step": 48160 }, { "action_loss": 0.0005586407496593893, "epoch": 43.30935251798561, "step": 48160 }, { "epoch": 43.30935251798561, "step": 48160, "torque_loss": 0.142322838306427 }, { "epoch": 43.318345323741006, "grad_norm": 0.02448764443397522, "learning_rate": 3.6618319172439517e-07, "loss": 0.0228, "step": 48170 }, { "action_loss": 0.0021612958516925573, "epoch": 43.318345323741006, "step": 48170 }, { "epoch": 43.318345323741006, "step": 48170, "torque_loss": 0.1451757699251175 }, { "epoch": 43.327338129496404, "grad_norm": 0.04355032369494438, "learning_rate": 3.6219912075579065e-07, "loss": 0.019, "step": 48180 }, { "action_loss": 0.001675742561928928, "epoch": 43.327338129496404, "step": 48180 }, { "epoch": 43.327338129496404, "step": 48180, "torque_loss": 0.13658709824085236 }, { "epoch": 43.3363309352518, "grad_norm": 0.026354677975177765, "learning_rate": 3.5823676304769793e-07, "loss": 0.0158, "step": 48190 }, { "action_loss": 0.0007285904139280319, "epoch": 43.3363309352518, "step": 48190 }, { "epoch": 43.3363309352518, "step": 48190, "torque_loss": 0.13501127064228058 }, { "epoch": 43.34532374100719, "grad_norm": 0.06375373899936676, "learning_rate": 3.5429612033338057e-07, "loss": 0.0152, "step": 48200 }, { "action_loss": 0.0010319051798433065, "epoch": 43.34532374100719, "step": 48200 }, { "epoch": 43.34532374100719, "step": 48200, "torque_loss": 0.13555152714252472 }, { "epoch": 43.35431654676259, "grad_norm": 0.056850966066122055, "learning_rate": 3.503771943366041e-07, "loss": 0.0186, "step": 48210 }, { "action_loss": 0.0017708525992929935, "epoch": 43.35431654676259, "step": 48210 }, { "epoch": 43.35431654676259, "step": 48210, "torque_loss": 0.162538081407547 }, { "epoch": 43.36330935251799, "grad_norm": 0.03321891650557518, "learning_rate": 3.464799867716473e-07, "loss": 0.0184, "step": 48220 }, { "action_loss": 0.0008986329776234925, "epoch": 43.36330935251799, "step": 48220 }, { "epoch": 43.36330935251799, "step": 48220, "torque_loss": 0.17580783367156982 }, { "epoch": 43.37230215827338, "grad_norm": 0.11469372361898422, "learning_rate": 3.4260449934327977e-07, "loss": 0.021, "step": 48230 }, { "action_loss": 0.000647804350592196, "epoch": 43.37230215827338, "step": 48230 }, { "epoch": 43.37230215827338, "step": 48230, "torque_loss": 0.13489829003810883 }, { "epoch": 43.381294964028775, "grad_norm": 0.12632498145103455, "learning_rate": 3.3875073374677326e-07, "loss": 0.0171, "step": 48240 }, { "action_loss": 0.0006500830058939755, "epoch": 43.381294964028775, "step": 48240 }, { "epoch": 43.381294964028775, "step": 48240, "torque_loss": 0.1429004967212677 }, { "epoch": 43.39028776978417, "grad_norm": 0.03823905810713768, "learning_rate": 3.3491869166789035e-07, "loss": 0.0167, "step": 48250 }, { "action_loss": 0.0008642282919026911, "epoch": 43.39028776978417, "step": 48250 }, { "epoch": 43.39028776978417, "step": 48250, "torque_loss": 0.13564267754554749 }, { "epoch": 43.39928057553957, "grad_norm": 0.06258780509233475, "learning_rate": 3.3110837478290137e-07, "loss": 0.0159, "step": 48260 }, { "action_loss": 0.0024061885196715593, "epoch": 43.39928057553957, "step": 48260 }, { "epoch": 43.39928057553957, "step": 48260, "torque_loss": 0.13933421671390533 }, { "epoch": 43.40827338129496, "grad_norm": 0.06374562531709671, "learning_rate": 3.273197847585674e-07, "loss": 0.0175, "step": 48270 }, { "action_loss": 0.000597058271523565, "epoch": 43.40827338129496, "step": 48270 }, { "epoch": 43.40827338129496, "step": 48270, "torque_loss": 0.18858130276203156 }, { "epoch": 43.41726618705036, "grad_norm": 0.027679700404405594, "learning_rate": 3.2355292325214617e-07, "loss": 0.0174, "step": 48280 }, { "action_loss": 0.0016535362228751183, "epoch": 43.41726618705036, "step": 48280 }, { "epoch": 43.41726618705036, "step": 48280, "torque_loss": 0.15965554118156433 }, { "epoch": 43.42625899280576, "grad_norm": 0.02406228706240654, "learning_rate": 3.198077919113862e-07, "loss": 0.0152, "step": 48290 }, { "action_loss": 0.0033042205031961203, "epoch": 43.42625899280576, "step": 48290 }, { "epoch": 43.42625899280576, "step": 48290, "torque_loss": 0.1560760736465454 }, { "epoch": 43.435251798561154, "grad_norm": 0.07839212566614151, "learning_rate": 3.160843923745438e-07, "loss": 0.0199, "step": 48300 }, { "action_loss": 0.0006115049473010004, "epoch": 43.435251798561154, "step": 48300 }, { "epoch": 43.435251798561154, "step": 48300, "torque_loss": 0.16528268158435822 }, { "epoch": 43.444244604316545, "grad_norm": 0.10512612760066986, "learning_rate": 3.123827262703549e-07, "loss": 0.018, "step": 48310 }, { "action_loss": 0.0003128579701296985, "epoch": 43.444244604316545, "step": 48310 }, { "epoch": 43.444244604316545, "step": 48310, "torque_loss": 0.10011399537324905 }, { "epoch": 43.45323741007194, "grad_norm": 0.05370231345295906, "learning_rate": 3.0870279521805765e-07, "loss": 0.0159, "step": 48320 }, { "action_loss": 0.0016103563830256462, "epoch": 43.45323741007194, "step": 48320 }, { "epoch": 43.45323741007194, "step": 48320, "torque_loss": 0.12003996223211288 }, { "epoch": 43.46223021582734, "grad_norm": 0.042107779532670975, "learning_rate": 3.0504460082737553e-07, "loss": 0.0171, "step": 48330 }, { "action_loss": 0.0005174425314180553, "epoch": 43.46223021582734, "step": 48330 }, { "epoch": 43.46223021582734, "step": 48330, "torque_loss": 0.10105825215578079 }, { "epoch": 43.47122302158273, "grad_norm": 0.05875881016254425, "learning_rate": 3.0140814469852286e-07, "loss": 0.0193, "step": 48340 }, { "action_loss": 0.0009997854940593243, "epoch": 43.47122302158273, "step": 48340 }, { "epoch": 43.47122302158273, "step": 48340, "torque_loss": 0.12845736742019653 }, { "epoch": 43.48021582733813, "grad_norm": 0.051279906183481216, "learning_rate": 2.9779342842221614e-07, "loss": 0.0165, "step": 48350 }, { "action_loss": 0.0027727957349270582, "epoch": 43.48021582733813, "step": 48350 }, { "epoch": 43.48021582733813, "step": 48350, "torque_loss": 0.17058944702148438 }, { "epoch": 43.489208633093526, "grad_norm": 0.051795843988657, "learning_rate": 2.942004535796572e-07, "loss": 0.017, "step": 48360 }, { "action_loss": 0.0007466888055205345, "epoch": 43.489208633093526, "step": 48360 }, { "epoch": 43.489208633093526, "step": 48360, "torque_loss": 0.11449148505926132 }, { "epoch": 43.49820143884892, "grad_norm": 0.03425915911793709, "learning_rate": 2.906292217425277e-07, "loss": 0.0187, "step": 48370 }, { "action_loss": 0.0013038469478487968, "epoch": 43.49820143884892, "step": 48370 }, { "epoch": 43.49820143884892, "step": 48370, "torque_loss": 0.19151334464550018 }, { "epoch": 43.507194244604314, "grad_norm": 0.03540847823023796, "learning_rate": 2.8707973447300584e-07, "loss": 0.0211, "step": 48380 }, { "action_loss": 0.002292243530973792, "epoch": 43.507194244604314, "step": 48380 }, { "epoch": 43.507194244604314, "step": 48380, "torque_loss": 0.22976309061050415 }, { "epoch": 43.51618705035971, "grad_norm": 0.033358264714479446, "learning_rate": 2.8355199332376627e-07, "loss": 0.0171, "step": 48390 }, { "action_loss": 0.0009651659056544304, "epoch": 43.51618705035971, "step": 48390 }, { "epoch": 43.51618705035971, "step": 48390, "torque_loss": 0.1636713594198227 }, { "epoch": 43.52517985611511, "grad_norm": 0.08725760877132416, "learning_rate": 2.8004599983795785e-07, "loss": 0.0182, "step": 48400 }, { "action_loss": 0.0012782354606315494, "epoch": 43.52517985611511, "step": 48400 }, { "epoch": 43.52517985611511, "step": 48400, "torque_loss": 0.16124409437179565 }, { "epoch": 43.53417266187051, "grad_norm": 0.0407826229929924, "learning_rate": 2.765617555492206e-07, "loss": 0.0165, "step": 48410 }, { "action_loss": 0.002572224708274007, "epoch": 43.53417266187051, "step": 48410 }, { "epoch": 43.53417266187051, "step": 48410, "torque_loss": 0.17923732101917267 }, { "epoch": 43.5431654676259, "grad_norm": 0.039023809134960175, "learning_rate": 2.7309926198168524e-07, "loss": 0.0178, "step": 48420 }, { "action_loss": 0.0016282383585348725, "epoch": 43.5431654676259, "step": 48420 }, { "epoch": 43.5431654676259, "step": 48420, "torque_loss": 0.11466997116804123 }, { "epoch": 43.552158273381295, "grad_norm": 0.07570040971040726, "learning_rate": 2.696585206499569e-07, "loss": 0.0178, "step": 48430 }, { "action_loss": 0.0010025331284850836, "epoch": 43.552158273381295, "step": 48430 }, { "epoch": 43.552158273381295, "step": 48430, "torque_loss": 0.1666746586561203 }, { "epoch": 43.56115107913669, "grad_norm": 0.08492486923933029, "learning_rate": 2.662395330591372e-07, "loss": 0.0173, "step": 48440 }, { "action_loss": 0.0008182378369383514, "epoch": 43.56115107913669, "step": 48440 }, { "epoch": 43.56115107913669, "step": 48440, "torque_loss": 0.13330964744091034 }, { "epoch": 43.57014388489208, "grad_norm": 0.03648685663938522, "learning_rate": 2.628423007048075e-07, "loss": 0.0164, "step": 48450 }, { "action_loss": 0.0017483490519225597, "epoch": 43.57014388489208, "step": 48450 }, { "epoch": 43.57014388489208, "step": 48450, "torque_loss": 0.12569718062877655 }, { "epoch": 43.57913669064748, "grad_norm": 0.07037028670310974, "learning_rate": 2.594668250730292e-07, "loss": 0.0184, "step": 48460 }, { "action_loss": 0.0023634780663996935, "epoch": 43.57913669064748, "step": 48460 }, { "epoch": 43.57913669064748, "step": 48460, "torque_loss": 0.2320653200149536 }, { "epoch": 43.58812949640288, "grad_norm": 0.0330241322517395, "learning_rate": 2.5611310764035445e-07, "loss": 0.0178, "step": 48470 }, { "action_loss": 0.0024211653508245945, "epoch": 43.58812949640288, "step": 48470 }, { "epoch": 43.58812949640288, "step": 48470, "torque_loss": 0.2053179293870926 }, { "epoch": 43.597122302158276, "grad_norm": 0.06093104928731918, "learning_rate": 2.5278114987380973e-07, "loss": 0.0187, "step": 48480 }, { "action_loss": 0.0033814848866313696, "epoch": 43.597122302158276, "step": 48480 }, { "epoch": 43.597122302158276, "step": 48480, "torque_loss": 0.15262965857982635 }, { "epoch": 43.606115107913666, "grad_norm": 0.03146669268608093, "learning_rate": 2.494709532309125e-07, "loss": 0.0179, "step": 48490 }, { "action_loss": 0.00038629499613307416, "epoch": 43.606115107913666, "step": 48490 }, { "epoch": 43.606115107913666, "step": 48490, "torque_loss": 0.11055835336446762 }, { "epoch": 43.615107913669064, "grad_norm": 0.10244312882423401, "learning_rate": 2.4618251915964896e-07, "loss": 0.0169, "step": 48500 }, { "action_loss": 0.0005045586149208248, "epoch": 43.615107913669064, "step": 48500 }, { "epoch": 43.615107913669064, "step": 48500, "torque_loss": 0.11037340015172958 }, { "epoch": 43.62410071942446, "grad_norm": 0.031118514016270638, "learning_rate": 2.429158490984962e-07, "loss": 0.0162, "step": 48510 }, { "action_loss": 0.0010728690540418029, "epoch": 43.62410071942446, "step": 48510 }, { "epoch": 43.62410071942446, "step": 48510, "torque_loss": 0.11333155632019043 }, { "epoch": 43.63309352517986, "grad_norm": 0.07780785858631134, "learning_rate": 2.396709444764056e-07, "loss": 0.0201, "step": 48520 }, { "action_loss": 0.0006769543397240341, "epoch": 43.63309352517986, "step": 48520 }, { "epoch": 43.63309352517986, "step": 48520, "torque_loss": 0.10692103952169418 }, { "epoch": 43.64208633093525, "grad_norm": 0.10582876205444336, "learning_rate": 2.364478067128084e-07, "loss": 0.0157, "step": 48530 }, { "action_loss": 0.0005029463791288435, "epoch": 43.64208633093525, "step": 48530 }, { "epoch": 43.64208633093525, "step": 48530, "torque_loss": 0.07060987502336502 }, { "epoch": 43.65107913669065, "grad_norm": 0.049480944871902466, "learning_rate": 2.332464372176102e-07, "loss": 0.0172, "step": 48540 }, { "action_loss": 0.0008752801804803312, "epoch": 43.65107913669065, "step": 48540 }, { "epoch": 43.65107913669065, "step": 48540, "torque_loss": 0.16318471729755402 }, { "epoch": 43.660071942446045, "grad_norm": 0.10750694572925568, "learning_rate": 2.300668373912074e-07, "loss": 0.0177, "step": 48550 }, { "action_loss": 0.00798408966511488, "epoch": 43.660071942446045, "step": 48550 }, { "epoch": 43.660071942446045, "step": 48550, "torque_loss": 0.2192131131887436 }, { "epoch": 43.669064748201436, "grad_norm": 0.034789878875017166, "learning_rate": 2.269090086244541e-07, "loss": 0.0177, "step": 48560 }, { "action_loss": 0.0008199980948120356, "epoch": 43.669064748201436, "step": 48560 }, { "epoch": 43.669064748201436, "step": 48560, "torque_loss": 0.14148001372814178 }, { "epoch": 43.67805755395683, "grad_norm": 0.046904370188713074, "learning_rate": 2.237729522987009e-07, "loss": 0.0173, "step": 48570 }, { "action_loss": 0.009100731462240219, "epoch": 43.67805755395683, "step": 48570 }, { "epoch": 43.67805755395683, "step": 48570, "torque_loss": 0.2136242389678955 }, { "epoch": 43.68705035971223, "grad_norm": 0.05661512166261673, "learning_rate": 2.2065866978576156e-07, "loss": 0.0166, "step": 48580 }, { "action_loss": 0.0008353324956260622, "epoch": 43.68705035971223, "step": 48580 }, { "epoch": 43.68705035971223, "step": 48580, "torque_loss": 0.08882912248373032 }, { "epoch": 43.69604316546763, "grad_norm": 0.05165771022439003, "learning_rate": 2.1756616244792972e-07, "loss": 0.0177, "step": 48590 }, { "action_loss": 0.0008091195486485958, "epoch": 43.69604316546763, "step": 48590 }, { "epoch": 43.69604316546763, "step": 48590, "torque_loss": 0.1489698737859726 }, { "epoch": 43.70503597122302, "grad_norm": 0.04091545194387436, "learning_rate": 2.1449543163797324e-07, "loss": 0.0169, "step": 48600 }, { "action_loss": 0.0019205905264243484, "epoch": 43.70503597122302, "step": 48600 }, { "epoch": 43.70503597122302, "step": 48600, "torque_loss": 0.1975274533033371 }, { "epoch": 43.71402877697842, "grad_norm": 0.0430917926132679, "learning_rate": 2.1144647869913992e-07, "loss": 0.0168, "step": 48610 }, { "action_loss": 0.004224300850182772, "epoch": 43.71402877697842, "step": 48610 }, { "epoch": 43.71402877697842, "step": 48610, "torque_loss": 0.20188254117965698 }, { "epoch": 43.723021582733814, "grad_norm": 0.027506856247782707, "learning_rate": 2.0841930496513507e-07, "loss": 0.0169, "step": 48620 }, { "action_loss": 0.015200448222458363, "epoch": 43.723021582733814, "step": 48620 }, { "epoch": 43.723021582733814, "step": 48620, "torque_loss": 0.25679337978363037 }, { "epoch": 43.73201438848921, "grad_norm": 0.02716761641204357, "learning_rate": 2.0541391176014945e-07, "loss": 0.0192, "step": 48630 }, { "action_loss": 0.004742088261991739, "epoch": 43.73201438848921, "step": 48630 }, { "epoch": 43.73201438848921, "step": 48630, "torque_loss": 0.19576053321361542 }, { "epoch": 43.7410071942446, "grad_norm": 0.07716891169548035, "learning_rate": 2.0243030039884814e-07, "loss": 0.0206, "step": 48640 }, { "action_loss": 0.0013661282137036324, "epoch": 43.7410071942446, "step": 48640 }, { "epoch": 43.7410071942446, "step": 48640, "torque_loss": 0.1606309860944748 }, { "epoch": 43.75, "grad_norm": 0.055557239800691605, "learning_rate": 1.994684721863649e-07, "loss": 0.0189, "step": 48650 }, { "action_loss": 0.0010676681995391846, "epoch": 43.75, "step": 48650 }, { "epoch": 43.75, "step": 48650, "torque_loss": 0.16242524981498718 }, { "epoch": 43.7589928057554, "grad_norm": 0.09796971082687378, "learning_rate": 1.9652842841829667e-07, "loss": 0.017, "step": 48660 }, { "action_loss": 0.0014504962600767612, "epoch": 43.7589928057554, "step": 48660 }, { "epoch": 43.7589928057554, "step": 48660, "torque_loss": 0.13297320902347565 }, { "epoch": 43.76798561151079, "grad_norm": 0.0743151530623436, "learning_rate": 1.9361017038072583e-07, "loss": 0.0168, "step": 48670 }, { "action_loss": 0.0006852769292891026, "epoch": 43.76798561151079, "step": 48670 }, { "epoch": 43.76798561151079, "step": 48670, "torque_loss": 0.2111491560935974 }, { "epoch": 43.776978417266186, "grad_norm": 0.02887517400085926, "learning_rate": 1.907136993501979e-07, "loss": 0.0167, "step": 48680 }, { "action_loss": 0.0007232643547467887, "epoch": 43.776978417266186, "step": 48680 }, { "epoch": 43.776978417266186, "step": 48680, "torque_loss": 0.15303361415863037 }, { "epoch": 43.78597122302158, "grad_norm": 0.046648651361465454, "learning_rate": 1.878390165937216e-07, "loss": 0.0178, "step": 48690 }, { "action_loss": 0.0036630157846957445, "epoch": 43.78597122302158, "step": 48690 }, { "epoch": 43.78597122302158, "step": 48690, "torque_loss": 0.19405074417591095 }, { "epoch": 43.79496402877698, "grad_norm": 0.08671586215496063, "learning_rate": 1.849861233687855e-07, "loss": 0.018, "step": 48700 }, { "action_loss": 0.004318962339311838, "epoch": 43.79496402877698, "step": 48700 }, { "epoch": 43.79496402877698, "step": 48700, "torque_loss": 0.21175964176654816 }, { "epoch": 43.80395683453237, "grad_norm": 0.04676042124629021, "learning_rate": 1.8215502092334135e-07, "loss": 0.0204, "step": 48710 }, { "action_loss": 0.0005883145495317876, "epoch": 43.80395683453237, "step": 48710 }, { "epoch": 43.80395683453237, "step": 48710, "torque_loss": 0.1726757287979126 }, { "epoch": 43.81294964028777, "grad_norm": 0.0875447541475296, "learning_rate": 1.7934571049580963e-07, "loss": 0.0195, "step": 48720 }, { "action_loss": 0.0006656036712229252, "epoch": 43.81294964028777, "step": 48720 }, { "epoch": 43.81294964028777, "step": 48720, "torque_loss": 0.12407486885786057 }, { "epoch": 43.82194244604317, "grad_norm": 0.03257829695940018, "learning_rate": 1.7655819331507394e-07, "loss": 0.0175, "step": 48730 }, { "action_loss": 0.0011031790636479855, "epoch": 43.82194244604317, "step": 48730 }, { "epoch": 43.82194244604317, "step": 48730, "torque_loss": 0.14357192814350128 }, { "epoch": 43.830935251798564, "grad_norm": 0.07743673771619797, "learning_rate": 1.7379247060049785e-07, "loss": 0.0175, "step": 48740 }, { "action_loss": 0.0011478894157335162, "epoch": 43.830935251798564, "step": 48740 }, { "epoch": 43.830935251798564, "step": 48740, "torque_loss": 0.14738309383392334 }, { "epoch": 43.839928057553955, "grad_norm": 0.028913777321577072, "learning_rate": 1.7104854356189693e-07, "loss": 0.0136, "step": 48750 }, { "action_loss": 0.0031108844559639692, "epoch": 43.839928057553955, "step": 48750 }, { "epoch": 43.839928057553955, "step": 48750, "torque_loss": 0.1854148507118225 }, { "epoch": 43.84892086330935, "grad_norm": 0.06987588852643967, "learning_rate": 1.6832641339955546e-07, "loss": 0.0193, "step": 48760 }, { "action_loss": 0.0016172860050573945, "epoch": 43.84892086330935, "step": 48760 }, { "epoch": 43.84892086330935, "step": 48760, "torque_loss": 0.18398360908031464 }, { "epoch": 43.85791366906475, "grad_norm": 0.034450601786375046, "learning_rate": 1.6562608130422653e-07, "loss": 0.0176, "step": 48770 }, { "action_loss": 0.05847353860735893, "epoch": 43.85791366906475, "step": 48770 }, { "epoch": 43.85791366906475, "step": 48770, "torque_loss": 0.19559873640537262 }, { "epoch": 43.86690647482014, "grad_norm": 0.025537297129631042, "learning_rate": 1.6294754845713745e-07, "loss": 0.0248, "step": 48780 }, { "action_loss": 0.0006439680582843721, "epoch": 43.86690647482014, "step": 48780 }, { "epoch": 43.86690647482014, "step": 48780, "torque_loss": 0.12123071402311325 }, { "epoch": 43.87589928057554, "grad_norm": 0.03029554896056652, "learning_rate": 1.60290816029951e-07, "loss": 0.017, "step": 48790 }, { "action_loss": 0.0005296625895425677, "epoch": 43.87589928057554, "step": 48790 }, { "epoch": 43.87589928057554, "step": 48790, "torque_loss": 0.14854253828525543 }, { "epoch": 43.884892086330936, "grad_norm": 0.1557793766260147, "learning_rate": 1.5765588518482643e-07, "loss": 0.0186, "step": 48800 }, { "action_loss": 0.00159444659948349, "epoch": 43.884892086330936, "step": 48800 }, { "epoch": 43.884892086330936, "step": 48800, "torque_loss": 0.16642236709594727 }, { "epoch": 43.893884892086334, "grad_norm": 0.04529412463307381, "learning_rate": 1.5504275707436956e-07, "loss": 0.0153, "step": 48810 }, { "action_loss": 0.001213751034811139, "epoch": 43.893884892086334, "step": 48810 }, { "epoch": 43.893884892086334, "step": 48810, "torque_loss": 0.13281677663326263 }, { "epoch": 43.902877697841724, "grad_norm": 0.04557750001549721, "learning_rate": 1.524514328416493e-07, "loss": 0.0185, "step": 48820 }, { "action_loss": 0.0008453487535007298, "epoch": 43.902877697841724, "step": 48820 }, { "epoch": 43.902877697841724, "step": 48820, "torque_loss": 0.10774753242731094 }, { "epoch": 43.91187050359712, "grad_norm": 0.041855812072753906, "learning_rate": 1.498819136201979e-07, "loss": 0.0195, "step": 48830 }, { "action_loss": 0.004135924857109785, "epoch": 43.91187050359712, "step": 48830 }, { "epoch": 43.91187050359712, "step": 48830, "torque_loss": 0.2345551997423172 }, { "epoch": 43.92086330935252, "grad_norm": 0.10138747096061707, "learning_rate": 1.4733420053401616e-07, "loss": 0.0255, "step": 48840 }, { "action_loss": 0.0006244787946343422, "epoch": 43.92086330935252, "step": 48840 }, { "epoch": 43.92086330935252, "step": 48840, "torque_loss": 0.0742010846734047 }, { "epoch": 43.92985611510792, "grad_norm": 0.045811865478754044, "learning_rate": 1.4480829469755153e-07, "loss": 0.0155, "step": 48850 }, { "action_loss": 0.0006577321910299361, "epoch": 43.92985611510792, "step": 48850 }, { "epoch": 43.92985611510792, "step": 48850, "torque_loss": 0.17429308593273163 }, { "epoch": 43.93884892086331, "grad_norm": 0.0524599626660347, "learning_rate": 1.4230419721572574e-07, "loss": 0.0165, "step": 48860 }, { "action_loss": 0.0013955304166302085, "epoch": 43.93884892086331, "step": 48860 }, { "epoch": 43.93884892086331, "step": 48860, "torque_loss": 0.2001289576292038 }, { "epoch": 43.947841726618705, "grad_norm": 0.023561380803585052, "learning_rate": 1.3982190918391812e-07, "loss": 0.0157, "step": 48870 }, { "action_loss": 0.0009029004722833633, "epoch": 43.947841726618705, "step": 48870 }, { "epoch": 43.947841726618705, "step": 48870, "torque_loss": 0.1611425280570984 }, { "epoch": 43.9568345323741, "grad_norm": 0.04490578547120094, "learning_rate": 1.3736143168796012e-07, "loss": 0.0171, "step": 48880 }, { "action_loss": 0.0022063979413360357, "epoch": 43.9568345323741, "step": 48880 }, { "epoch": 43.9568345323741, "step": 48880, "torque_loss": 0.11891114711761475 }, { "epoch": 43.96582733812949, "grad_norm": 0.12764273583889008, "learning_rate": 1.349227658041574e-07, "loss": 0.0191, "step": 48890 }, { "action_loss": 0.003114465856924653, "epoch": 43.96582733812949, "step": 48890 }, { "epoch": 43.96582733812949, "step": 48890, "torque_loss": 0.2286706566810608 }, { "epoch": 43.97482014388489, "grad_norm": 0.08649265021085739, "learning_rate": 1.325059125992567e-07, "loss": 0.0195, "step": 48900 }, { "action_loss": 0.006246721837669611, "epoch": 43.97482014388489, "step": 48900 }, { "epoch": 43.97482014388489, "step": 48900, "torque_loss": 0.16662345826625824 }, { "epoch": 43.98381294964029, "grad_norm": 0.07394905388355255, "learning_rate": 1.3011087313047344e-07, "loss": 0.0181, "step": 48910 }, { "action_loss": 0.011719703674316406, "epoch": 43.98381294964029, "step": 48910 }, { "epoch": 43.98381294964029, "step": 48910, "torque_loss": 0.16518600285053253 }, { "epoch": 43.992805755395686, "grad_norm": 0.02672315016388893, "learning_rate": 1.2773764844548064e-07, "loss": 0.016, "step": 48920 }, { "action_loss": 0.0014684194466099143, "epoch": 43.992805755395686, "step": 48920 }, { "epoch": 43.992805755395686, "step": 48920, "torque_loss": 0.15052103996276855 }, { "epoch": 44.00179856115108, "grad_norm": 0.08966319262981415, "learning_rate": 1.2538623958240903e-07, "loss": 0.021, "step": 48930 }, { "action_loss": 0.0006343098357319832, "epoch": 44.00179856115108, "step": 48930 }, { "epoch": 44.00179856115108, "step": 48930, "torque_loss": 0.14657068252563477 }, { "epoch": 44.010791366906474, "grad_norm": 0.05192401632666588, "learning_rate": 1.2305664756984137e-07, "loss": 0.0184, "step": 48940 }, { "action_loss": 0.0007079007918946445, "epoch": 44.010791366906474, "step": 48940 }, { "epoch": 44.010791366906474, "step": 48940, "torque_loss": 0.09121644496917725 }, { "epoch": 44.01978417266187, "grad_norm": 0.09910505264997482, "learning_rate": 1.2074887342681806e-07, "loss": 0.0196, "step": 48950 }, { "action_loss": 0.0010124589316546917, "epoch": 44.01978417266187, "step": 48950 }, { "epoch": 44.01978417266187, "step": 48950, "torque_loss": 0.15354324877262115 }, { "epoch": 44.02877697841727, "grad_norm": 0.029317405074834824, "learning_rate": 1.1846291816284827e-07, "loss": 0.0153, "step": 48960 }, { "action_loss": 0.0010191771434620023, "epoch": 44.02877697841727, "step": 48960 }, { "epoch": 44.02877697841727, "step": 48960, "torque_loss": 0.2018103450536728 }, { "epoch": 44.03776978417266, "grad_norm": 0.019503558054566383, "learning_rate": 1.1619878277787654e-07, "loss": 0.0188, "step": 48970 }, { "action_loss": 0.0007811528630554676, "epoch": 44.03776978417266, "step": 48970 }, { "epoch": 44.03776978417266, "step": 48970, "torque_loss": 0.13435731828212738 }, { "epoch": 44.04676258992806, "grad_norm": 0.05934366583824158, "learning_rate": 1.1395646826231621e-07, "loss": 0.017, "step": 48980 }, { "action_loss": 0.0013470702106133103, "epoch": 44.04676258992806, "step": 48980 }, { "epoch": 44.04676258992806, "step": 48980, "torque_loss": 0.19930176436901093 }, { "epoch": 44.055755395683455, "grad_norm": 0.02559617906808853, "learning_rate": 1.1173597559703263e-07, "loss": 0.0168, "step": 48990 }, { "action_loss": 0.0006705443374812603, "epoch": 44.055755395683455, "step": 48990 }, { "epoch": 44.055755395683455, "step": 48990, "torque_loss": 0.15838654339313507 }, { "epoch": 44.064748201438846, "grad_norm": 0.022162022069096565, "learning_rate": 1.0953730575334331e-07, "loss": 0.0158, "step": 49000 }, { "action_loss": 0.0006071893149055541, "epoch": 44.064748201438846, "step": 49000 }, { "epoch": 44.064748201438846, "step": 49000, "torque_loss": 0.11947248131036758 }, { "epoch": 44.07374100719424, "grad_norm": 0.06408954411745071, "learning_rate": 1.0736045969301778e-07, "loss": 0.0156, "step": 49010 }, { "action_loss": 0.0008323523215949535, "epoch": 44.07374100719424, "step": 49010 }, { "epoch": 44.07374100719424, "step": 49010, "torque_loss": 0.12947550415992737 }, { "epoch": 44.08273381294964, "grad_norm": 0.03330742195248604, "learning_rate": 1.0520543836829433e-07, "loss": 0.0159, "step": 49020 }, { "action_loss": 0.00800991989672184, "epoch": 44.08273381294964, "step": 49020 }, { "epoch": 44.08273381294964, "step": 49020, "torque_loss": 0.22114457190036774 }, { "epoch": 44.09172661870504, "grad_norm": 0.0426843985915184, "learning_rate": 1.0307224272183558e-07, "loss": 0.0192, "step": 49030 }, { "action_loss": 0.0006521962932310998, "epoch": 44.09172661870504, "step": 49030 }, { "epoch": 44.09172661870504, "step": 49030, "torque_loss": 0.1187044009566307 }, { "epoch": 44.10071942446043, "grad_norm": 0.03994366154074669, "learning_rate": 1.0096087368678952e-07, "loss": 0.0166, "step": 49040 }, { "action_loss": 0.0023804467637091875, "epoch": 44.10071942446043, "step": 49040 }, { "epoch": 44.10071942446043, "step": 49040, "torque_loss": 0.2042096108198166 }, { "epoch": 44.10971223021583, "grad_norm": 0.03742614760994911, "learning_rate": 9.88713321867285e-08, "loss": 0.0195, "step": 49050 }, { "action_loss": 0.013500381261110306, "epoch": 44.10971223021583, "step": 49050 }, { "epoch": 44.10971223021583, "step": 49050, "torque_loss": 0.2061520367860794 }, { "epoch": 44.118705035971225, "grad_norm": 0.07119819521903992, "learning_rate": 9.680361913569913e-08, "loss": 0.0221, "step": 49060 }, { "action_loss": 0.0012046502670273185, "epoch": 44.118705035971225, "step": 49060 }, { "epoch": 44.118705035971225, "step": 49060, "torque_loss": 0.17440657317638397 }, { "epoch": 44.12769784172662, "grad_norm": 0.05352238938212395, "learning_rate": 9.475773543818344e-08, "loss": 0.0164, "step": 49070 }, { "action_loss": 0.0003970424004364759, "epoch": 44.12769784172662, "step": 49070 }, { "epoch": 44.12769784172662, "step": 49070, "torque_loss": 0.07852771878242493 }, { "epoch": 44.13669064748201, "grad_norm": 0.03659896180033684, "learning_rate": 9.273368198911559e-08, "loss": 0.0192, "step": 49080 }, { "action_loss": 0.005734902340918779, "epoch": 44.13669064748201, "step": 49080 }, { "epoch": 44.13669064748201, "step": 49080, "torque_loss": 0.18320651352405548 }, { "epoch": 44.14568345323741, "grad_norm": 0.053277164697647095, "learning_rate": 9.07314596738984e-08, "loss": 0.0177, "step": 49090 }, { "action_loss": 0.0007481514476239681, "epoch": 44.14568345323741, "step": 49090 }, { "epoch": 44.14568345323741, "step": 49090, "torque_loss": 0.12288928031921387 }, { "epoch": 44.15467625899281, "grad_norm": 0.048427775502204895, "learning_rate": 8.875106936836463e-08, "loss": 0.0165, "step": 49100 }, { "action_loss": 0.0010850965045392513, "epoch": 44.15467625899281, "step": 49100 }, { "epoch": 44.15467625899281, "step": 49100, "torque_loss": 0.11718564480543137 }, { "epoch": 44.1636690647482, "grad_norm": 0.03345099464058876, "learning_rate": 8.679251193879911e-08, "loss": 0.0139, "step": 49110 }, { "action_loss": 0.0025968728587031364, "epoch": 44.1636690647482, "step": 49110 }, { "epoch": 44.1636690647482, "step": 49110, "torque_loss": 0.25076282024383545 }, { "epoch": 44.172661870503596, "grad_norm": 0.043705981224775314, "learning_rate": 8.485578824194984e-08, "loss": 0.0169, "step": 49120 }, { "action_loss": 0.001394097344018519, "epoch": 44.172661870503596, "step": 49120 }, { "epoch": 44.172661870503596, "step": 49120, "torque_loss": 0.11137769371271133 }, { "epoch": 44.181654676258994, "grad_norm": 0.03766937926411629, "learning_rate": 8.294089912500024e-08, "loss": 0.0156, "step": 49130 }, { "action_loss": 0.0026693406980484724, "epoch": 44.181654676258994, "step": 49130 }, { "epoch": 44.181654676258994, "step": 49130, "torque_loss": 0.1805208921432495 }, { "epoch": 44.19064748201439, "grad_norm": 0.02978658676147461, "learning_rate": 8.10478454255914e-08, "loss": 0.0166, "step": 49140 }, { "action_loss": 0.0009693014435470104, "epoch": 44.19064748201439, "step": 49140 }, { "epoch": 44.19064748201439, "step": 49140, "torque_loss": 0.060772184282541275 }, { "epoch": 44.19964028776978, "grad_norm": 0.05623413994908333, "learning_rate": 7.917662797181091e-08, "loss": 0.0157, "step": 49150 }, { "action_loss": 0.0015024080639705062, "epoch": 44.19964028776978, "step": 49150 }, { "epoch": 44.19964028776978, "step": 49150, "torque_loss": 0.16131548583507538 }, { "epoch": 44.20863309352518, "grad_norm": 0.03610723838210106, "learning_rate": 7.732724758218734e-08, "loss": 0.0139, "step": 49160 }, { "action_loss": 0.004081004299223423, "epoch": 44.20863309352518, "step": 49160 }, { "epoch": 44.20863309352518, "step": 49160, "torque_loss": 0.15963254868984222 }, { "epoch": 44.21762589928058, "grad_norm": 0.021574167534708977, "learning_rate": 7.549970506570691e-08, "loss": 0.0198, "step": 49170 }, { "action_loss": 0.000834158796351403, "epoch": 44.21762589928058, "step": 49170 }, { "epoch": 44.21762589928058, "step": 49170, "torque_loss": 0.12775489687919617 }, { "epoch": 44.226618705035975, "grad_norm": 0.055128563195466995, "learning_rate": 7.369400122179126e-08, "loss": 0.021, "step": 49180 }, { "action_loss": 0.002550410805270076, "epoch": 44.226618705035975, "step": 49180 }, { "epoch": 44.226618705035975, "step": 49180, "torque_loss": 0.18629658222198486 }, { "epoch": 44.235611510791365, "grad_norm": 0.04501991719007492, "learning_rate": 7.191013684032522e-08, "loss": 0.0191, "step": 49190 }, { "action_loss": 0.001416692859493196, "epoch": 44.235611510791365, "step": 49190 }, { "epoch": 44.235611510791365, "step": 49190, "torque_loss": 0.1740909069776535 }, { "epoch": 44.24460431654676, "grad_norm": 0.04877220466732979, "learning_rate": 7.014811270163458e-08, "loss": 0.0184, "step": 49200 }, { "action_loss": 0.002311553107574582, "epoch": 44.24460431654676, "step": 49200 }, { "epoch": 44.24460431654676, "step": 49200, "torque_loss": 0.2051295042037964 }, { "epoch": 44.25359712230216, "grad_norm": 0.10750783234834671, "learning_rate": 6.840792957647501e-08, "loss": 0.0179, "step": 49210 }, { "action_loss": 0.0008674776763655245, "epoch": 44.25359712230216, "step": 49210 }, { "epoch": 44.25359712230216, "step": 49210, "torque_loss": 0.11892776936292648 }, { "epoch": 44.26258992805755, "grad_norm": 0.042452991008758545, "learning_rate": 6.668958822607096e-08, "loss": 0.0187, "step": 49220 }, { "action_loss": 0.0004800403257831931, "epoch": 44.26258992805755, "step": 49220 }, { "epoch": 44.26258992805755, "step": 49220, "torque_loss": 0.09398152679204941 }, { "epoch": 44.27158273381295, "grad_norm": 0.14278487861156464, "learning_rate": 6.499308940208226e-08, "loss": 0.0189, "step": 49230 }, { "action_loss": 0.00579003244638443, "epoch": 44.27158273381295, "step": 49230 }, { "epoch": 44.27158273381295, "step": 49230, "torque_loss": 0.11836133152246475 }, { "epoch": 44.280575539568346, "grad_norm": 0.03330826759338379, "learning_rate": 6.331843384661529e-08, "loss": 0.0184, "step": 49240 }, { "action_loss": 0.0008863750845193863, "epoch": 44.280575539568346, "step": 49240 }, { "epoch": 44.280575539568346, "step": 49240, "torque_loss": 0.13608194887638092 }, { "epoch": 44.289568345323744, "grad_norm": 0.05498507618904114, "learning_rate": 6.166562229222295e-08, "loss": 0.0177, "step": 49250 }, { "action_loss": 0.0010268178302794695, "epoch": 44.289568345323744, "step": 49250 }, { "epoch": 44.289568345323744, "step": 49250, "torque_loss": 0.18162469565868378 }, { "epoch": 44.298561151079134, "grad_norm": 0.028960874304175377, "learning_rate": 6.003465546189358e-08, "loss": 0.0196, "step": 49260 }, { "action_loss": 0.0003961642214562744, "epoch": 44.298561151079134, "step": 49260 }, { "epoch": 44.298561151079134, "step": 49260, "torque_loss": 0.15054656565189362 }, { "epoch": 44.30755395683453, "grad_norm": 0.0791701003909111, "learning_rate": 5.8425534069078694e-08, "loss": 0.0146, "step": 49270 }, { "action_loss": 0.00293677463196218, "epoch": 44.30755395683453, "step": 49270 }, { "epoch": 44.30755395683453, "step": 49270, "torque_loss": 0.16007429361343384 }, { "epoch": 44.31654676258993, "grad_norm": 0.026838185265660286, "learning_rate": 5.6838258817654144e-08, "loss": 0.017, "step": 49280 }, { "action_loss": 0.004245828837156296, "epoch": 44.31654676258993, "step": 49280 }, { "epoch": 44.31654676258993, "step": 49280, "torque_loss": 0.2325952649116516 }, { "epoch": 44.32553956834532, "grad_norm": 0.06083565950393677, "learning_rate": 5.527283040195341e-08, "loss": 0.018, "step": 49290 }, { "action_loss": 0.0024348243605345488, "epoch": 44.32553956834532, "step": 49290 }, { "epoch": 44.32553956834532, "step": 49290, "torque_loss": 0.15513615310192108 }, { "epoch": 44.33453237410072, "grad_norm": 0.11258852481842041, "learning_rate": 5.372924950673986e-08, "loss": 0.0158, "step": 49300 }, { "action_loss": 0.0008351849392056465, "epoch": 44.33453237410072, "step": 49300 }, { "epoch": 44.33453237410072, "step": 49300, "torque_loss": 0.1333675980567932 }, { "epoch": 44.343525179856115, "grad_norm": 0.032906562089920044, "learning_rate": 5.2207516807240006e-08, "loss": 0.0152, "step": 49310 }, { "action_loss": 0.00045547462650574744, "epoch": 44.343525179856115, "step": 49310 }, { "epoch": 44.343525179856115, "step": 49310, "torque_loss": 0.10722843557596207 }, { "epoch": 44.35251798561151, "grad_norm": 0.03350178152322769, "learning_rate": 5.07076329690992e-08, "loss": 0.0175, "step": 49320 }, { "action_loss": 0.0029930954333394766, "epoch": 44.35251798561151, "step": 49320 }, { "epoch": 44.35251798561151, "step": 49320, "torque_loss": 0.16314391791820526 }, { "epoch": 44.361510791366904, "grad_norm": 0.03772290050983429, "learning_rate": 4.922959864842591e-08, "loss": 0.0177, "step": 49330 }, { "action_loss": 0.0012678547063842416, "epoch": 44.361510791366904, "step": 49330 }, { "epoch": 44.361510791366904, "step": 49330, "torque_loss": 0.16419416666030884 }, { "epoch": 44.3705035971223, "grad_norm": 0.15726347267627716, "learning_rate": 4.7773414491758537e-08, "loss": 0.0199, "step": 49340 }, { "action_loss": 0.0015700329095125198, "epoch": 44.3705035971223, "step": 49340 }, { "epoch": 44.3705035971223, "step": 49340, "torque_loss": 0.20968584716320038 }, { "epoch": 44.3794964028777, "grad_norm": 0.02928970381617546, "learning_rate": 4.6339081136087534e-08, "loss": 0.0183, "step": 49350 }, { "action_loss": 0.0014421994565054774, "epoch": 44.3794964028777, "step": 49350 }, { "epoch": 44.3794964028777, "step": 49350, "torque_loss": 0.1179339811205864 }, { "epoch": 44.388489208633096, "grad_norm": 0.11818565428256989, "learning_rate": 4.492659920882769e-08, "loss": 0.0178, "step": 49360 }, { "action_loss": 0.0037712857592850924, "epoch": 44.388489208633096, "step": 49360 }, { "epoch": 44.388489208633096, "step": 49360, "torque_loss": 0.1552792340517044 }, { "epoch": 44.39748201438849, "grad_norm": 0.08019427955150604, "learning_rate": 4.3535969327862525e-08, "loss": 0.0212, "step": 49370 }, { "action_loss": 0.001598857343196869, "epoch": 44.39748201438849, "step": 49370 }, { "epoch": 44.39748201438849, "step": 49370, "torque_loss": 0.19322383403778076 }, { "epoch": 44.406474820143885, "grad_norm": 0.2403063327074051, "learning_rate": 4.2167192101483234e-08, "loss": 0.0222, "step": 49380 }, { "action_loss": 0.001384497620165348, "epoch": 44.406474820143885, "step": 49380 }, { "epoch": 44.406474820143885, "step": 49380, "torque_loss": 0.1744784563779831 }, { "epoch": 44.41546762589928, "grad_norm": 0.04543573781847954, "learning_rate": 4.0820268128449744e-08, "loss": 0.0175, "step": 49390 }, { "action_loss": 0.0019091898575425148, "epoch": 44.41546762589928, "step": 49390 }, { "epoch": 44.41546762589928, "step": 49390, "torque_loss": 0.18810993432998657 }, { "epoch": 44.42446043165468, "grad_norm": 0.05422678962349892, "learning_rate": 3.949519799795742e-08, "loss": 0.017, "step": 49400 }, { "action_loss": 0.003269718959927559, "epoch": 44.42446043165468, "step": 49400 }, { "epoch": 44.42446043165468, "step": 49400, "torque_loss": 0.17657677829265594 }, { "epoch": 44.43345323741007, "grad_norm": 0.043214101344347, "learning_rate": 3.819198228962595e-08, "loss": 0.0138, "step": 49410 }, { "action_loss": 0.0030668892432004213, "epoch": 44.43345323741007, "step": 49410 }, { "epoch": 44.43345323741007, "step": 49410, "torque_loss": 0.12128561735153198 }, { "epoch": 44.44244604316547, "grad_norm": 0.025035375729203224, "learning_rate": 3.6910621573527094e-08, "loss": 0.0161, "step": 49420 }, { "action_loss": 0.006083955988287926, "epoch": 44.44244604316547, "step": 49420 }, { "epoch": 44.44244604316547, "step": 49420, "torque_loss": 0.22146421670913696 }, { "epoch": 44.451438848920866, "grad_norm": 0.054411567747592926, "learning_rate": 3.5651116410179156e-08, "loss": 0.0187, "step": 49430 }, { "action_loss": 0.0020879304502159357, "epoch": 44.451438848920866, "step": 49430 }, { "epoch": 44.451438848920866, "step": 49430, "torque_loss": 0.1689508706331253 }, { "epoch": 44.460431654676256, "grad_norm": 0.04714209586381912, "learning_rate": 3.441346735052475e-08, "loss": 0.0165, "step": 49440 }, { "action_loss": 0.0012200138298794627, "epoch": 44.460431654676256, "step": 49440 }, { "epoch": 44.460431654676256, "step": 49440, "torque_loss": 0.16071394085884094 }, { "epoch": 44.469424460431654, "grad_norm": 0.039116863161325455, "learning_rate": 3.31976749359586e-08, "loss": 0.0194, "step": 49450 }, { "action_loss": 0.0028816789854317904, "epoch": 44.469424460431654, "step": 49450 }, { "epoch": 44.469424460431654, "step": 49450, "torque_loss": 0.16831238567829132 }, { "epoch": 44.47841726618705, "grad_norm": 0.027294106781482697, "learning_rate": 3.200373969831083e-08, "loss": 0.015, "step": 49460 }, { "action_loss": 0.0005799782229587436, "epoch": 44.47841726618705, "step": 49460 }, { "epoch": 44.47841726618705, "step": 49460, "torque_loss": 0.14454872906208038 }, { "epoch": 44.48741007194245, "grad_norm": 0.06131836771965027, "learning_rate": 3.0831662159841456e-08, "loss": 0.0158, "step": 49470 }, { "action_loss": 0.0016874974826350808, "epoch": 44.48741007194245, "step": 49470 }, { "epoch": 44.48741007194245, "step": 49470, "torque_loss": 0.17827631533145905 }, { "epoch": 44.49640287769784, "grad_norm": 0.04275878518819809, "learning_rate": 2.9681442833268124e-08, "loss": 0.0167, "step": 49480 }, { "action_loss": 0.0026356943417340517, "epoch": 44.49640287769784, "step": 49480 }, { "epoch": 44.49640287769784, "step": 49480, "torque_loss": 0.12289778143167496 }, { "epoch": 44.50539568345324, "grad_norm": 0.028406355530023575, "learning_rate": 2.8553082221721704e-08, "loss": 0.0189, "step": 49490 }, { "action_loss": 0.00112581392750144, "epoch": 44.50539568345324, "step": 49490 }, { "epoch": 44.50539568345324, "step": 49490, "torque_loss": 0.18210966885089874 }, { "epoch": 44.514388489208635, "grad_norm": 0.028278114274144173, "learning_rate": 2.7446580818801805e-08, "loss": 0.0181, "step": 49500 }, { "action_loss": 0.0012140815379098058, "epoch": 44.514388489208635, "step": 49500 }, { "epoch": 44.514388489208635, "step": 49500, "torque_loss": 0.1361469328403473 }, { "epoch": 44.523381294964025, "grad_norm": 0.05388869717717171, "learning_rate": 2.63619391085157e-08, "loss": 0.0149, "step": 49510 }, { "action_loss": 0.0008922933484427631, "epoch": 44.523381294964025, "step": 49510 }, { "epoch": 44.523381294964025, "step": 49510, "torque_loss": 0.15504871308803558 }, { "epoch": 44.53237410071942, "grad_norm": 0.029241476207971573, "learning_rate": 2.5299157565328302e-08, "loss": 0.0151, "step": 49520 }, { "action_loss": 0.0018130609532818198, "epoch": 44.53237410071942, "step": 49520 }, { "epoch": 44.53237410071942, "step": 49520, "torque_loss": 0.18221603333950043 }, { "epoch": 44.54136690647482, "grad_norm": 0.02777266502380371, "learning_rate": 2.4258236654139954e-08, "loss": 0.017, "step": 49530 }, { "action_loss": 0.0010777092538774014, "epoch": 44.54136690647482, "step": 49530 }, { "epoch": 44.54136690647482, "step": 49530, "torque_loss": 0.1416083574295044 }, { "epoch": 44.55035971223022, "grad_norm": 0.04220537096261978, "learning_rate": 2.3239176830280873e-08, "loss": 0.0155, "step": 49540 }, { "action_loss": 0.001250771339982748, "epoch": 44.55035971223022, "step": 49540 }, { "epoch": 44.55035971223022, "step": 49540, "torque_loss": 0.12377363443374634 }, { "epoch": 44.55935251798561, "grad_norm": 0.02053026109933853, "learning_rate": 2.2241978539522256e-08, "loss": 0.0171, "step": 49550 }, { "action_loss": 0.0007498925551772118, "epoch": 44.55935251798561, "step": 49550 }, { "epoch": 44.55935251798561, "step": 49550, "torque_loss": 0.14334887266159058 }, { "epoch": 44.568345323741006, "grad_norm": 0.09817753732204437, "learning_rate": 2.1266642218070733e-08, "loss": 0.019, "step": 49560 }, { "action_loss": 0.0029025112744420767, "epoch": 44.568345323741006, "step": 49560 }, { "epoch": 44.568345323741006, "step": 49560, "torque_loss": 0.16057445108890533 }, { "epoch": 44.577338129496404, "grad_norm": 0.03802315890789032, "learning_rate": 2.0313168292579456e-08, "loss": 0.0173, "step": 49570 }, { "action_loss": 0.0007273534429259598, "epoch": 44.577338129496404, "step": 49570 }, { "epoch": 44.577338129496404, "step": 49570, "torque_loss": 0.13027383387088776 }, { "epoch": 44.5863309352518, "grad_norm": 0.0552452877163887, "learning_rate": 1.9381557180120357e-08, "loss": 0.0185, "step": 49580 }, { "action_loss": 0.0010018311440944672, "epoch": 44.5863309352518, "step": 49580 }, { "epoch": 44.5863309352518, "step": 49580, "torque_loss": 0.17405645549297333 }, { "epoch": 44.59532374100719, "grad_norm": 0.055365223437547684, "learning_rate": 1.8471809288217456e-08, "loss": 0.0174, "step": 49590 }, { "action_loss": 0.007323647383600473, "epoch": 44.59532374100719, "step": 49590 }, { "epoch": 44.59532374100719, "step": 49590, "torque_loss": 0.13495634496212006 }, { "epoch": 44.60431654676259, "grad_norm": 0.10795816779136658, "learning_rate": 1.758392501482464e-08, "loss": 0.0173, "step": 49600 }, { "action_loss": 0.0005775532335974276, "epoch": 44.60431654676259, "step": 49600 }, { "epoch": 44.60431654676259, "step": 49600, "torque_loss": 0.13618777692317963 }, { "epoch": 44.61330935251799, "grad_norm": 0.029840299859642982, "learning_rate": 1.6717904748331236e-08, "loss": 0.0175, "step": 49610 }, { "action_loss": 0.002700721612200141, "epoch": 44.61330935251799, "step": 49610 }, { "epoch": 44.61330935251799, "step": 49610, "torque_loss": 0.1933966428041458 }, { "epoch": 44.62230215827338, "grad_norm": 0.02993468940258026, "learning_rate": 1.5873748867561988e-08, "loss": 0.0176, "step": 49620 }, { "action_loss": 0.0019199323141947389, "epoch": 44.62230215827338, "step": 49620 }, { "epoch": 44.62230215827338, "step": 49620, "torque_loss": 0.14300720393657684 }, { "epoch": 44.631294964028775, "grad_norm": 0.05294740945100784, "learning_rate": 1.5051457741788177e-08, "loss": 0.0158, "step": 49630 }, { "action_loss": 0.001095421495847404, "epoch": 44.631294964028775, "step": 49630 }, { "epoch": 44.631294964028775, "step": 49630, "torque_loss": 0.16756343841552734 }, { "epoch": 44.64028776978417, "grad_norm": 0.04459712281823158, "learning_rate": 1.425103173069986e-08, "loss": 0.0158, "step": 49640 }, { "action_loss": 0.008231129497289658, "epoch": 44.64028776978417, "step": 49640 }, { "epoch": 44.64028776978417, "step": 49640, "torque_loss": 0.2072393149137497 }, { "epoch": 44.64928057553957, "grad_norm": 0.09873118996620178, "learning_rate": 1.347247118442807e-08, "loss": 0.0169, "step": 49650 }, { "action_loss": 0.00137148832436651, "epoch": 44.64928057553957, "step": 49650 }, { "epoch": 44.64928057553957, "step": 49650, "torque_loss": 0.18795961141586304 }, { "epoch": 44.65827338129496, "grad_norm": 0.022340334951877594, "learning_rate": 1.2715776443550375e-08, "loss": 0.0189, "step": 49660 }, { "action_loss": 0.0012258124770596623, "epoch": 44.65827338129496, "step": 49660 }, { "epoch": 44.65827338129496, "step": 49660, "torque_loss": 0.10819771140813828 }, { "epoch": 44.66726618705036, "grad_norm": 0.06973244994878769, "learning_rate": 1.1980947839063116e-08, "loss": 0.0181, "step": 49670 }, { "action_loss": 0.0006048788200132549, "epoch": 44.66726618705036, "step": 49670 }, { "epoch": 44.66726618705036, "step": 49670, "torque_loss": 0.15863870084285736 }, { "epoch": 44.67625899280576, "grad_norm": 0.05309416726231575, "learning_rate": 1.1267985692414718e-08, "loss": 0.0172, "step": 49680 }, { "action_loss": 0.001091939164325595, "epoch": 44.67625899280576, "step": 49680 }, { "epoch": 44.67625899280576, "step": 49680, "torque_loss": 0.13710780441761017 }, { "epoch": 44.685251798561154, "grad_norm": 0.031249146908521652, "learning_rate": 1.0576890315466826e-08, "loss": 0.0225, "step": 49690 }, { "action_loss": 0.0011114258086308837, "epoch": 44.685251798561154, "step": 49690 }, { "epoch": 44.685251798561154, "step": 49690, "torque_loss": 0.12164144963026047 }, { "epoch": 44.694244604316545, "grad_norm": 0.028013046830892563, "learning_rate": 9.907662010538721e-09, "loss": 0.019, "step": 49700 }, { "action_loss": 0.0017975919181481004, "epoch": 44.694244604316545, "step": 49700 }, { "epoch": 44.694244604316545, "step": 49700, "torque_loss": 0.15165533125400543 }, { "epoch": 44.70323741007194, "grad_norm": 0.16710971295833588, "learning_rate": 9.26030107036846e-09, "loss": 0.0201, "step": 49710 }, { "action_loss": 0.003002611920237541, "epoch": 44.70323741007194, "step": 49710 }, { "epoch": 44.70323741007194, "step": 49710, "torque_loss": 0.1970147341489792 }, { "epoch": 44.71223021582734, "grad_norm": 0.023029277101159096, "learning_rate": 8.63480777813508e-09, "loss": 0.0185, "step": 49720 }, { "action_loss": 0.0004511277948040515, "epoch": 44.71223021582734, "step": 49720 }, { "epoch": 44.71223021582734, "step": 49720, "torque_loss": 0.13696284592151642 }, { "epoch": 44.72122302158273, "grad_norm": 0.02830575779080391, "learning_rate": 8.03118240744749e-09, "loss": 0.0138, "step": 49730 }, { "action_loss": 0.001287761959247291, "epoch": 44.72122302158273, "step": 49730 }, { "epoch": 44.72122302158273, "step": 49730, "torque_loss": 0.14908134937286377 }, { "epoch": 44.73021582733813, "grad_norm": 0.038482315838336945, "learning_rate": 7.4494252223555885e-09, "loss": 0.0153, "step": 49740 }, { "action_loss": 0.0005988159100525081, "epoch": 44.73021582733813, "step": 49740 }, { "epoch": 44.73021582733813, "step": 49740, "torque_loss": 0.09916585683822632 }, { "epoch": 44.739208633093526, "grad_norm": 0.08004981279373169, "learning_rate": 6.8895364773391426e-09, "loss": 0.0178, "step": 49750 }, { "action_loss": 0.001388698467053473, "epoch": 44.739208633093526, "step": 49750 }, { "epoch": 44.739208633093526, "step": 49750, "torque_loss": 0.1706920713186264 }, { "epoch": 44.74820143884892, "grad_norm": 0.03309183567762375, "learning_rate": 6.3515164173078015e-09, "loss": 0.0163, "step": 49760 }, { "action_loss": 0.0015013240044936538, "epoch": 44.74820143884892, "step": 49760 }, { "epoch": 44.74820143884892, "step": 49760, "torque_loss": 0.18516790866851807 }, { "epoch": 44.757194244604314, "grad_norm": 0.0386442169547081, "learning_rate": 5.835365277617744e-09, "loss": 0.0156, "step": 49770 }, { "action_loss": 0.004380650352686644, "epoch": 44.757194244604314, "step": 49770 }, { "epoch": 44.757194244604314, "step": 49770, "torque_loss": 0.21934755146503448 }, { "epoch": 44.76618705035971, "grad_norm": 0.029859501868486404, "learning_rate": 5.341083284049475e-09, "loss": 0.0181, "step": 49780 }, { "action_loss": 0.0007863110513426363, "epoch": 44.76618705035971, "step": 49780 }, { "epoch": 44.76618705035971, "step": 49780, "torque_loss": 0.183440163731575 }, { "epoch": 44.77517985611511, "grad_norm": 0.04555516317486763, "learning_rate": 4.868670652813378e-09, "loss": 0.0186, "step": 49790 }, { "action_loss": 0.002390054753050208, "epoch": 44.77517985611511, "step": 49790 }, { "epoch": 44.77517985611511, "step": 49790, "torque_loss": 0.11251243948936462 }, { "epoch": 44.78417266187051, "grad_norm": 0.03547353297472, "learning_rate": 4.418127590560817e-09, "loss": 0.0174, "step": 49800 }, { "action_loss": 0.0011847895802929997, "epoch": 44.78417266187051, "step": 49800 }, { "epoch": 44.78417266187051, "step": 49800, "torque_loss": 0.17376156151294708 }, { "epoch": 44.7931654676259, "grad_norm": 0.07507643848657608, "learning_rate": 3.989454294378581e-09, "loss": 0.0163, "step": 49810 }, { "action_loss": 0.0003351694904267788, "epoch": 44.7931654676259, "step": 49810 }, { "epoch": 44.7931654676259, "step": 49810, "torque_loss": 0.1223229393362999 }, { "epoch": 44.802158273381295, "grad_norm": 0.07103197276592255, "learning_rate": 3.5826509517777883e-09, "loss": 0.0188, "step": 49820 }, { "action_loss": 0.0007612052722834051, "epoch": 44.802158273381295, "step": 49820 }, { "epoch": 44.802158273381295, "step": 49820, "torque_loss": 0.17625679075717926 }, { "epoch": 44.81115107913669, "grad_norm": 0.02628691866993904, "learning_rate": 3.1977177407105372e-09, "loss": 0.0158, "step": 49830 }, { "action_loss": 0.0006918418221175671, "epoch": 44.81115107913669, "step": 49830 }, { "epoch": 44.81115107913669, "step": 49830, "torque_loss": 0.18588225543498993 }, { "epoch": 44.82014388489208, "grad_norm": 0.02599191479384899, "learning_rate": 2.8346548295588027e-09, "loss": 0.0173, "step": 49840 }, { "action_loss": 0.0011274437420070171, "epoch": 44.82014388489208, "step": 49840 }, { "epoch": 44.82014388489208, "step": 49840, "torque_loss": 0.12820890545845032 }, { "epoch": 44.82913669064748, "grad_norm": 0.05443376675248146, "learning_rate": 2.493462377139988e-09, "loss": 0.0164, "step": 49850 }, { "action_loss": 0.0008843921241350472, "epoch": 44.82913669064748, "step": 49850 }, { "epoch": 44.82913669064748, "step": 49850, "torque_loss": 0.1924421638250351 }, { "epoch": 44.83812949640288, "grad_norm": 0.14708028733730316, "learning_rate": 2.1741405327013744e-09, "loss": 0.023, "step": 49860 }, { "action_loss": 0.0013591571478173137, "epoch": 44.83812949640288, "step": 49860 }, { "epoch": 44.83812949640288, "step": 49860, "torque_loss": 0.15991292893886566 }, { "epoch": 44.847122302158276, "grad_norm": 0.03899678960442543, "learning_rate": 1.8766894359312227e-09, "loss": 0.0175, "step": 49870 }, { "action_loss": 0.0008393071475438774, "epoch": 44.847122302158276, "step": 49870 }, { "epoch": 44.847122302158276, "step": 49870, "torque_loss": 0.17972707748413086 }, { "epoch": 44.856115107913666, "grad_norm": 0.06337025761604309, "learning_rate": 1.6011092169310182e-09, "loss": 0.0202, "step": 49880 }, { "action_loss": 0.0014504585415124893, "epoch": 44.856115107913666, "step": 49880 }, { "epoch": 44.856115107913666, "step": 49880, "torque_loss": 0.14060144126415253 }, { "epoch": 44.865107913669064, "grad_norm": 0.026791315525770187, "learning_rate": 1.3473999962654304e-09, "loss": 0.0158, "step": 49890 }, { "action_loss": 0.002879668027162552, "epoch": 44.865107913669064, "step": 49890 }, { "epoch": 44.865107913669064, "step": 49890, "torque_loss": 0.15785866975784302 }, { "epoch": 44.87410071942446, "grad_norm": 0.04345202445983887, "learning_rate": 1.1155618849068016e-09, "loss": 0.0172, "step": 49900 }, { "action_loss": 0.0009296455536969006, "epoch": 44.87410071942446, "step": 49900 }, { "epoch": 44.87410071942446, "step": 49900, "torque_loss": 0.2029813677072525 }, { "epoch": 44.88309352517986, "grad_norm": 0.08159800618886948, "learning_rate": 9.055949842684541e-10, "loss": 0.0186, "step": 49910 }, { "action_loss": 0.0009604948572814465, "epoch": 44.88309352517986, "step": 49910 }, { "epoch": 44.88309352517986, "step": 49910, "torque_loss": 0.1409691721200943 }, { "epoch": 44.89208633093525, "grad_norm": 0.0584058053791523, "learning_rate": 7.174993861991386e-10, "loss": 0.0152, "step": 49920 }, { "action_loss": 0.003340619383379817, "epoch": 44.89208633093525, "step": 49920 }, { "epoch": 44.89208633093525, "step": 49920, "torque_loss": 0.17112302780151367 }, { "epoch": 44.90107913669065, "grad_norm": 0.03570042923092842, "learning_rate": 5.512751729774834e-10, "loss": 0.0166, "step": 49930 }, { "action_loss": 0.0011381992371752858, "epoch": 44.90107913669065, "step": 49930 }, { "epoch": 44.90107913669065, "step": 49930, "torque_loss": 0.15109242498874664 }, { "epoch": 44.910071942446045, "grad_norm": 0.02333991229534149, "learning_rate": 4.0692241731199456e-10, "loss": 0.0252, "step": 49940 }, { "action_loss": 0.0008633295074105263, "epoch": 44.910071942446045, "step": 49940 }, { "epoch": 44.910071942446045, "step": 49940, "torque_loss": 0.09125860780477524 }, { "epoch": 44.919064748201436, "grad_norm": 0.11495416611433029, "learning_rate": 2.8444118235770865e-10, "loss": 0.0174, "step": 49950 }, { "action_loss": 0.004448993131518364, "epoch": 44.919064748201436, "step": 49950 }, { "epoch": 44.919064748201436, "step": 49950, "torque_loss": 0.1918921321630478 }, { "epoch": 44.92805755395683, "grad_norm": 0.06512721627950668, "learning_rate": 1.8383152168288675e-10, "loss": 0.0166, "step": 49960 }, { "action_loss": 0.003352251835167408, "epoch": 44.92805755395683, "step": 49960 }, { "epoch": 44.92805755395683, "step": 49960, "torque_loss": 0.17401151359081268 }, { "epoch": 44.93705035971223, "grad_norm": 0.05201731622219086, "learning_rate": 1.050934793023206e-10, "loss": 0.0167, "step": 49970 }, { "action_loss": 0.0047438195906579494, "epoch": 44.93705035971223, "step": 49970 }, { "epoch": 44.93705035971223, "step": 49970, "torque_loss": 0.1985289603471756 }, { "epoch": 44.94604316546763, "grad_norm": 0.033551327884197235, "learning_rate": 4.822708965512845e-11, "loss": 0.018, "step": 49980 }, { "action_loss": 0.004069116432219744, "epoch": 44.94604316546763, "step": 49980 }, { "epoch": 44.94604316546763, "step": 49980, "torque_loss": 0.2385614663362503 }, { "epoch": 44.95503597122302, "grad_norm": 0.03538154810667038, "learning_rate": 1.323237761585716e-11, "loss": 0.0167, "step": 49990 }, { "action_loss": 0.0015367678133770823, "epoch": 44.95503597122302, "step": 49990 }, { "epoch": 44.95503597122302, "step": 49990, "torque_loss": 0.13076509535312653 }, { "epoch": 44.96402877697842, "grad_norm": 0.2261248677968979, "learning_rate": 1.0935850003335191e-13, "loss": 0.0206, "step": 50000 } ], "logging_steps": 10, "max_steps": 50000, "num_input_tokens_seen": 0, "num_train_epochs": 45, "save_steps": 10000, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 16, "trial_name": null, "trial_params": null }