diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,110034 @@
+{
+  "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": 3.6376631259918213,
+      "epoch": 0,
+      "step": 0
+    },
+    {
+      "epoch": 0,
+      "step": 0,
+      "torque_loss": 0.1310535967350006
+    },
+    {
+      "epoch": 0,
+      "force_loss": 0.02128283679485321,
+      "step": 0
+    },
+    {
+      "epoch": 0.008992805755395683,
+      "grad_norm": 9.891151428222656,
+      "learning_rate": 3.0000000000000004e-07,
+      "loss": 2.1385,
+      "step": 10
+    },
+    {
+      "action_loss": 1.5190473794937134,
+      "epoch": 0.008992805755395683,
+      "step": 10
+    },
+    {
+      "epoch": 0.008992805755395683,
+      "step": 10,
+      "torque_loss": 0.12258154153823853
+    },
+    {
+      "epoch": 0.008992805755395683,
+      "force_loss": 0.004153921268880367,
+      "step": 10
+    },
+    {
+      "epoch": 0.017985611510791366,
+      "grad_norm": 27.13469696044922,
+      "learning_rate": 6.333333333333333e-07,
+      "loss": 2.2744,
+      "step": 20
+    },
+    {
+      "action_loss": 1.9660362005233765,
+      "epoch": 0.017985611510791366,
+      "step": 20
+    },
+    {
+      "epoch": 0.017985611510791366,
+      "step": 20,
+      "torque_loss": 0.11698909848928452
+    },
+    {
+      "epoch": 0.017985611510791366,
+      "force_loss": 0.005360100418329239,
+      "step": 20
+    },
+    {
+      "epoch": 0.02697841726618705,
+      "grad_norm": 8.638763427734375,
+      "learning_rate": 9.666666666666668e-07,
+      "loss": 1.9938,
+      "step": 30
+    },
+    {
+      "action_loss": 2.620020866394043,
+      "epoch": 0.02697841726618705,
+      "step": 30
+    },
+    {
+      "epoch": 0.02697841726618705,
+      "step": 30,
+      "torque_loss": 0.1213768944144249
+    },
+    {
+      "epoch": 0.02697841726618705,
+      "force_loss": 0.005952438339591026,
+      "step": 30
+    },
+    {
+      "epoch": 0.03597122302158273,
+      "grad_norm": 7.296123504638672,
+      "learning_rate": 1.3e-06,
+      "loss": 1.8309,
+      "step": 40
+    },
+    {
+      "action_loss": 2.0353288650512695,
+      "epoch": 0.03597122302158273,
+      "step": 40
+    },
+    {
+      "epoch": 0.03597122302158273,
+      "step": 40,
+      "torque_loss": 0.10332077741622925
+    },
+    {
+      "epoch": 0.03597122302158273,
+      "force_loss": 0.004071989096701145,
+      "step": 40
+    },
+    {
+      "epoch": 0.044964028776978415,
+      "grad_norm": 9.959613800048828,
+      "learning_rate": 1.6333333333333333e-06,
+      "loss": 1.5416,
+      "step": 50
+    },
+    {
+      "action_loss": 1.45693838596344,
+      "epoch": 0.044964028776978415,
+      "step": 50
+    },
+    {
+      "epoch": 0.044964028776978415,
+      "step": 50,
+      "torque_loss": 0.15176773071289062
+    },
+    {
+      "epoch": 0.044964028776978415,
+      "force_loss": 0.0103449122980237,
+      "step": 50
+    },
+    {
+      "epoch": 0.0539568345323741,
+      "grad_norm": 9.581283569335938,
+      "learning_rate": 1.9666666666666668e-06,
+      "loss": 1.304,
+      "step": 60
+    },
+    {
+      "action_loss": 0.7612078785896301,
+      "epoch": 0.0539568345323741,
+      "step": 60
+    },
+    {
+      "epoch": 0.0539568345323741,
+      "step": 60,
+      "torque_loss": 0.12856867909431458
+    },
+    {
+      "epoch": 0.0539568345323741,
+      "force_loss": 0.008561134338378906,
+      "step": 60
+    },
+    {
+      "epoch": 0.06294964028776978,
+      "grad_norm": 2.9715723991394043,
+      "learning_rate": 2.3e-06,
+      "loss": 0.6737,
+      "step": 70
+    },
+    {
+      "action_loss": 0.5321075916290283,
+      "epoch": 0.06294964028776978,
+      "step": 70
+    },
+    {
+      "epoch": 0.06294964028776978,
+      "step": 70,
+      "torque_loss": 0.1495532989501953
+    },
+    {
+      "epoch": 0.06294964028776978,
+      "force_loss": 0.009209034964442253,
+      "step": 70
+    },
+    {
+      "epoch": 0.07194244604316546,
+      "grad_norm": 1.6039245128631592,
+      "learning_rate": 2.6333333333333337e-06,
+      "loss": 0.5001,
+      "step": 80
+    },
+    {
+      "action_loss": 0.3438389301300049,
+      "epoch": 0.07194244604316546,
+      "step": 80
+    },
+    {
+      "epoch": 0.07194244604316546,
+      "step": 80,
+      "torque_loss": 0.1306103616952896
+    },
+    {
+      "epoch": 0.07194244604316546,
+      "force_loss": 0.0066747204400599,
+      "step": 80
+    },
+    {
+      "epoch": 0.08093525179856115,
+      "grad_norm": 1.0983778238296509,
+      "learning_rate": 2.966666666666667e-06,
+      "loss": 0.3208,
+      "step": 90
+    },
+    {
+      "action_loss": 0.20547671616077423,
+      "epoch": 0.08093525179856115,
+      "step": 90
+    },
+    {
+      "epoch": 0.08093525179856115,
+      "step": 90,
+      "torque_loss": 0.0799812451004982
+    },
+    {
+      "epoch": 0.08093525179856115,
+      "force_loss": 0.010391616262495518,
+      "step": 90
+    },
+    {
+      "epoch": 0.08992805755395683,
+      "grad_norm": 1.26064133644104,
+      "learning_rate": 3.3e-06,
+      "loss": 0.2245,
+      "step": 100
+    },
+    {
+      "action_loss": 0.21050791442394257,
+      "epoch": 0.08992805755395683,
+      "step": 100
+    },
+    {
+      "epoch": 0.08992805755395683,
+      "step": 100,
+      "torque_loss": 0.12303809076547623
+    },
+    {
+      "epoch": 0.08992805755395683,
+      "force_loss": 0.03322964906692505,
+      "step": 100
+    },
+    {
+      "epoch": 0.09892086330935251,
+      "grad_norm": 1.037462830543518,
+      "learning_rate": 3.633333333333334e-06,
+      "loss": 0.1602,
+      "step": 110
+    },
+    {
+      "action_loss": 0.1163199171423912,
+      "epoch": 0.09892086330935251,
+      "step": 110
+    },
+    {
+      "epoch": 0.09892086330935251,
+      "step": 110,
+      "torque_loss": 0.13896340131759644
+    },
+    {
+      "epoch": 0.09892086330935251,
+      "force_loss": 0.012037482112646103,
+      "step": 110
+    },
+    {
+      "epoch": 0.1079136690647482,
+      "grad_norm": 0.969456136226654,
+      "learning_rate": 3.966666666666667e-06,
+      "loss": 0.141,
+      "step": 120
+    },
+    {
+      "action_loss": 0.08728659152984619,
+      "epoch": 0.1079136690647482,
+      "step": 120
+    },
+    {
+      "epoch": 0.1079136690647482,
+      "step": 120,
+      "torque_loss": 0.16430842876434326
+    },
+    {
+      "epoch": 0.1079136690647482,
+      "force_loss": 0.008811325766146183,
+      "step": 120
+    },
+    {
+      "epoch": 0.11690647482014388,
+      "grad_norm": 0.6086524724960327,
+      "learning_rate": 4.2999999999999995e-06,
+      "loss": 0.0974,
+      "step": 130
+    },
+    {
+      "action_loss": 0.0822930559515953,
+      "epoch": 0.11690647482014388,
+      "step": 130
+    },
+    {
+      "epoch": 0.11690647482014388,
+      "step": 130,
+      "torque_loss": 0.11346596479415894
+    },
+    {
+      "epoch": 0.11690647482014388,
+      "force_loss": 0.007961176335811615,
+      "step": 130
+    },
+    {
+      "epoch": 0.12589928057553956,
+      "grad_norm": 0.8939685225486755,
+      "learning_rate": 4.633333333333334e-06,
+      "loss": 0.098,
+      "step": 140
+    },
+    {
+      "action_loss": 0.0769037976861,
+      "epoch": 0.12589928057553956,
+      "step": 140
+    },
+    {
+      "epoch": 0.12589928057553956,
+      "step": 140,
+      "torque_loss": 0.09801127761602402
+    },
+    {
+      "epoch": 0.12589928057553956,
+      "force_loss": 0.012784930877387524,
+      "step": 140
+    },
+    {
+      "epoch": 0.13489208633093525,
+      "grad_norm": 0.6305805444717407,
+      "learning_rate": 4.966666666666667e-06,
+      "loss": 0.1025,
+      "step": 150
+    },
+    {
+      "action_loss": 0.1941053420305252,
+      "epoch": 0.13489208633093525,
+      "step": 150
+    },
+    {
+      "epoch": 0.13489208633093525,
+      "step": 150,
+      "torque_loss": 0.10263881832361221
+    },
+    {
+      "epoch": 0.13489208633093525,
+      "force_loss": 0.02107587456703186,
+      "step": 150
+    },
+    {
+      "epoch": 0.14388489208633093,
+      "grad_norm": 0.4861755669116974,
+      "learning_rate": 5.3e-06,
+      "loss": 0.0901,
+      "step": 160
+    },
+    {
+      "action_loss": 0.09890081733465195,
+      "epoch": 0.14388489208633093,
+      "step": 160
+    },
+    {
+      "epoch": 0.14388489208633093,
+      "step": 160,
+      "torque_loss": 0.17149204015731812
+    },
+    {
+      "epoch": 0.14388489208633093,
+      "force_loss": 0.022561684250831604,
+      "step": 160
+    },
+    {
+      "epoch": 0.1528776978417266,
+      "grad_norm": 0.6252902150154114,
+      "learning_rate": 5.633333333333333e-06,
+      "loss": 0.083,
+      "step": 170
+    },
+    {
+      "action_loss": 0.05647003650665283,
+      "epoch": 0.1528776978417266,
+      "step": 170
+    },
+    {
+      "epoch": 0.1528776978417266,
+      "step": 170,
+      "torque_loss": 0.08202075213193893
+    },
+    {
+      "epoch": 0.1528776978417266,
+      "force_loss": 0.006358828395605087,
+      "step": 170
+    },
+    {
+      "epoch": 0.1618705035971223,
+      "grad_norm": 0.5395506620407104,
+      "learning_rate": 5.9666666666666666e-06,
+      "loss": 0.0722,
+      "step": 180
+    },
+    {
+      "action_loss": 0.06473840028047562,
+      "epoch": 0.1618705035971223,
+      "step": 180
+    },
+    {
+      "epoch": 0.1618705035971223,
+      "step": 180,
+      "torque_loss": 0.15107816457748413
+    },
+    {
+      "epoch": 0.1618705035971223,
+      "force_loss": 0.008181636221706867,
+      "step": 180
+    },
+    {
+      "epoch": 0.17086330935251798,
+      "grad_norm": 0.5020890831947327,
+      "learning_rate": 6.300000000000001e-06,
+      "loss": 0.0903,
+      "step": 190
+    },
+    {
+      "action_loss": 0.04997554421424866,
+      "epoch": 0.17086330935251798,
+      "step": 190
+    },
+    {
+      "epoch": 0.17086330935251798,
+      "step": 190,
+      "torque_loss": 0.1321963667869568
+    },
+    {
+      "epoch": 0.17086330935251798,
+      "force_loss": 0.005873364862054586,
+      "step": 190
+    },
+    {
+      "epoch": 0.17985611510791366,
+      "grad_norm": 0.748670220375061,
+      "learning_rate": 6.633333333333333e-06,
+      "loss": 0.0817,
+      "step": 200
+    },
+    {
+      "action_loss": 0.05113643407821655,
+      "epoch": 0.17985611510791366,
+      "step": 200
+    },
+    {
+      "epoch": 0.17985611510791366,
+      "step": 200,
+      "torque_loss": 0.08320024609565735
+    },
+    {
+      "epoch": 0.17985611510791366,
+      "force_loss": 0.004743098746985197,
+      "step": 200
+    },
+    {
+      "epoch": 0.18884892086330934,
+      "grad_norm": 0.6618114113807678,
+      "learning_rate": 6.966666666666667e-06,
+      "loss": 0.0751,
+      "step": 210
+    },
+    {
+      "action_loss": 0.10174226760864258,
+      "epoch": 0.18884892086330934,
+      "step": 210
+    },
+    {
+      "epoch": 0.18884892086330934,
+      "step": 210,
+      "torque_loss": 0.14920921623706818
+    },
+    {
+      "epoch": 0.18884892086330934,
+      "force_loss": 0.026553340256214142,
+      "step": 210
+    },
+    {
+      "epoch": 0.19784172661870503,
+      "grad_norm": 0.9313657283782959,
+      "learning_rate": 7.2999999999999996e-06,
+      "loss": 0.0785,
+      "step": 220
+    },
+    {
+      "action_loss": 0.047935888171195984,
+      "epoch": 0.19784172661870503,
+      "step": 220
+    },
+    {
+      "epoch": 0.19784172661870503,
+      "step": 220,
+      "torque_loss": 0.10199981182813644
+    },
+    {
+      "epoch": 0.19784172661870503,
+      "force_loss": 0.006586765870451927,
+      "step": 220
+    },
+    {
+      "epoch": 0.2068345323741007,
+      "grad_norm": 0.6217069625854492,
+      "learning_rate": 7.633333333333334e-06,
+      "loss": 0.0726,
+      "step": 230
+    },
+    {
+      "action_loss": 0.06976015865802765,
+      "epoch": 0.2068345323741007,
+      "step": 230
+    },
+    {
+      "epoch": 0.2068345323741007,
+      "step": 230,
+      "torque_loss": 0.1708124876022339
+    },
+    {
+      "epoch": 0.2068345323741007,
+      "force_loss": 0.009130173362791538,
+      "step": 230
+    },
+    {
+      "epoch": 0.2158273381294964,
+      "grad_norm": 0.5605801939964294,
+      "learning_rate": 7.966666666666666e-06,
+      "loss": 0.0661,
+      "step": 240
+    },
+    {
+      "action_loss": 0.05596521869301796,
+      "epoch": 0.2158273381294964,
+      "step": 240
+    },
+    {
+      "epoch": 0.2158273381294964,
+      "step": 240,
+      "torque_loss": 0.1426011323928833
+    },
+    {
+      "epoch": 0.2158273381294964,
+      "force_loss": 0.02059830166399479,
+      "step": 240
+    },
+    {
+      "epoch": 0.22482014388489208,
+      "grad_norm": 0.8638727068901062,
+      "learning_rate": 8.3e-06,
+      "loss": 0.0681,
+      "step": 250
+    },
+    {
+      "action_loss": 0.040239691734313965,
+      "epoch": 0.22482014388489208,
+      "step": 250
+    },
+    {
+      "epoch": 0.22482014388489208,
+      "step": 250,
+      "torque_loss": 0.12120440602302551
+    },
+    {
+      "epoch": 0.22482014388489208,
+      "force_loss": 0.007772465702146292,
+      "step": 250
+    },
+    {
+      "epoch": 0.23381294964028776,
+      "grad_norm": 0.871470034122467,
+      "learning_rate": 8.633333333333334e-06,
+      "loss": 0.0692,
+      "step": 260
+    },
+    {
+      "action_loss": 0.03378848358988762,
+      "epoch": 0.23381294964028776,
+      "step": 260
+    },
+    {
+      "epoch": 0.23381294964028776,
+      "step": 260,
+      "torque_loss": 0.11716548353433609
+    },
+    {
+      "epoch": 0.23381294964028776,
+      "force_loss": 0.008904024958610535,
+      "step": 260
+    },
+    {
+      "epoch": 0.24280575539568344,
+      "grad_norm": 0.7010048031806946,
+      "learning_rate": 8.966666666666668e-06,
+      "loss": 0.0616,
+      "step": 270
+    },
+    {
+      "action_loss": 0.05553820729255676,
+      "epoch": 0.24280575539568344,
+      "step": 270
+    },
+    {
+      "epoch": 0.24280575539568344,
+      "step": 270,
+      "torque_loss": 0.12629489600658417
+    },
+    {
+      "epoch": 0.24280575539568344,
+      "force_loss": 0.011371352709829807,
+      "step": 270
+    },
+    {
+      "epoch": 0.2517985611510791,
+      "grad_norm": 0.967866063117981,
+      "learning_rate": 9.3e-06,
+      "loss": 0.0667,
+      "step": 280
+    },
+    {
+      "action_loss": 0.03285503759980202,
+      "epoch": 0.2517985611510791,
+      "step": 280
+    },
+    {
+      "epoch": 0.2517985611510791,
+      "step": 280,
+      "torque_loss": 0.13254724442958832
+    },
+    {
+      "epoch": 0.2517985611510791,
+      "force_loss": 0.007841754704713821,
+      "step": 280
+    },
+    {
+      "epoch": 0.2607913669064748,
+      "grad_norm": 0.917709231376648,
+      "learning_rate": 9.633333333333335e-06,
+      "loss": 0.0565,
+      "step": 290
+    },
+    {
+      "action_loss": 0.05310952663421631,
+      "epoch": 0.2607913669064748,
+      "step": 290
+    },
+    {
+      "epoch": 0.2607913669064748,
+      "step": 290,
+      "torque_loss": 0.12525951862335205
+    },
+    {
+      "epoch": 0.2607913669064748,
+      "force_loss": 0.015282098203897476,
+      "step": 290
+    },
+    {
+      "epoch": 0.2697841726618705,
+      "grad_norm": 0.8989470601081848,
+      "learning_rate": 9.966666666666667e-06,
+      "loss": 0.0683,
+      "step": 300
+    },
+    {
+      "action_loss": 0.05016603693366051,
+      "epoch": 0.2697841726618705,
+      "step": 300
+    },
+    {
+      "epoch": 0.2697841726618705,
+      "step": 300,
+      "torque_loss": 0.15589874982833862
+    },
+    {
+      "epoch": 0.2697841726618705,
+      "force_loss": 0.00809991080313921,
+      "step": 300
+    },
+    {
+      "epoch": 0.2787769784172662,
+      "grad_norm": 0.6818510890007019,
+      "learning_rate": 1.03e-05,
+      "loss": 0.0609,
+      "step": 310
+    },
+    {
+      "action_loss": 0.07677297294139862,
+      "epoch": 0.2787769784172662,
+      "step": 310
+    },
+    {
+      "epoch": 0.2787769784172662,
+      "step": 310,
+      "torque_loss": 0.14252005517482758
+    },
+    {
+      "epoch": 0.2787769784172662,
+      "force_loss": 0.016483528539538383,
+      "step": 310
+    },
+    {
+      "epoch": 0.28776978417266186,
+      "grad_norm": 0.7003334164619446,
+      "learning_rate": 1.0633333333333334e-05,
+      "loss": 0.0673,
+      "step": 320
+    },
+    {
+      "action_loss": 0.034718334674835205,
+      "epoch": 0.28776978417266186,
+      "step": 320
+    },
+    {
+      "epoch": 0.28776978417266186,
+      "step": 320,
+      "torque_loss": 0.1005847230553627
+    },
+    {
+      "epoch": 0.28776978417266186,
+      "force_loss": 0.007835912518203259,
+      "step": 320
+    },
+    {
+      "epoch": 0.29676258992805754,
+      "grad_norm": 1.7371386289596558,
+      "learning_rate": 1.0966666666666666e-05,
+      "loss": 0.0613,
+      "step": 330
+    },
+    {
+      "action_loss": 0.02670879103243351,
+      "epoch": 0.29676258992805754,
+      "step": 330
+    },
+    {
+      "epoch": 0.29676258992805754,
+      "step": 330,
+      "torque_loss": 0.12702833116054535
+    },
+    {
+      "epoch": 0.29676258992805754,
+      "force_loss": 0.006304746028035879,
+      "step": 330
+    },
+    {
+      "epoch": 0.3057553956834532,
+      "grad_norm": 0.8466010689735413,
+      "learning_rate": 1.13e-05,
+      "loss": 0.05,
+      "step": 340
+    },
+    {
+      "action_loss": 0.03459268435835838,
+      "epoch": 0.3057553956834532,
+      "step": 340
+    },
+    {
+      "epoch": 0.3057553956834532,
+      "step": 340,
+      "torque_loss": 0.16588880121707916
+    },
+    {
+      "epoch": 0.3057553956834532,
+      "force_loss": 0.005099931266158819,
+      "step": 340
+    },
+    {
+      "epoch": 0.3147482014388489,
+      "grad_norm": 0.8969944715499878,
+      "learning_rate": 1.1633333333333334e-05,
+      "loss": 0.0595,
+      "step": 350
+    },
+    {
+      "action_loss": 0.04767943546175957,
+      "epoch": 0.3147482014388489,
+      "step": 350
+    },
+    {
+      "epoch": 0.3147482014388489,
+      "step": 350,
+      "torque_loss": 0.11656659841537476
+    },
+    {
+      "epoch": 0.3147482014388489,
+      "force_loss": 0.010912281461060047,
+      "step": 350
+    },
+    {
+      "epoch": 0.3237410071942446,
+      "grad_norm": 1.1630769968032837,
+      "learning_rate": 1.1966666666666668e-05,
+      "loss": 0.0616,
+      "step": 360
+    },
+    {
+      "action_loss": 0.03671069070696831,
+      "epoch": 0.3237410071942446,
+      "step": 360
+    },
+    {
+      "epoch": 0.3237410071942446,
+      "step": 360,
+      "torque_loss": 0.1766238659620285
+    },
+    {
+      "epoch": 0.3237410071942446,
+      "force_loss": 0.005978018045425415,
+      "step": 360
+    },
+    {
+      "epoch": 0.3327338129496403,
+      "grad_norm": 1.4596916437149048,
+      "learning_rate": 1.23e-05,
+      "loss": 0.0484,
+      "step": 370
+    },
+    {
+      "action_loss": 0.051470864564180374,
+      "epoch": 0.3327338129496403,
+      "step": 370
+    },
+    {
+      "epoch": 0.3327338129496403,
+      "step": 370,
+      "torque_loss": 0.17661406099796295
+    },
+    {
+      "epoch": 0.3327338129496403,
+      "force_loss": 0.01126351673156023,
+      "step": 370
+    },
+    {
+      "epoch": 0.34172661870503596,
+      "grad_norm": 0.7545153498649597,
+      "learning_rate": 1.2633333333333333e-05,
+      "loss": 0.052,
+      "step": 380
+    },
+    {
+      "action_loss": 0.025191066786646843,
+      "epoch": 0.34172661870503596,
+      "step": 380
+    },
+    {
+      "epoch": 0.34172661870503596,
+      "step": 380,
+      "torque_loss": 0.1213822141289711
+    },
+    {
+      "epoch": 0.34172661870503596,
+      "force_loss": 0.004954246804118156,
+      "step": 380
+    },
+    {
+      "epoch": 0.35071942446043164,
+      "grad_norm": 0.7178869247436523,
+      "learning_rate": 1.2966666666666669e-05,
+      "loss": 0.0615,
+      "step": 390
+    },
+    {
+      "action_loss": 0.013403424061834812,
+      "epoch": 0.35071942446043164,
+      "step": 390
+    },
+    {
+      "epoch": 0.35071942446043164,
+      "step": 390,
+      "torque_loss": 0.11117445677518845
+    },
+    {
+      "epoch": 0.35071942446043164,
+      "force_loss": 0.004236349370330572,
+      "step": 390
+    },
+    {
+      "epoch": 0.3597122302158273,
+      "grad_norm": 0.997737467288971,
+      "learning_rate": 1.3300000000000001e-05,
+      "loss": 0.0447,
+      "step": 400
+    },
+    {
+      "action_loss": 0.045525435358285904,
+      "epoch": 0.3597122302158273,
+      "step": 400
+    },
+    {
+      "epoch": 0.3597122302158273,
+      "step": 400,
+      "torque_loss": 0.1176399514079094
+    },
+    {
+      "epoch": 0.3597122302158273,
+      "force_loss": 0.019651243463158607,
+      "step": 400
+    },
+    {
+      "epoch": 0.368705035971223,
+      "grad_norm": 0.7645737528800964,
+      "learning_rate": 1.3633333333333334e-05,
+      "loss": 0.0545,
+      "step": 410
+    },
+    {
+      "action_loss": 0.03362283110618591,
+      "epoch": 0.368705035971223,
+      "step": 410
+    },
+    {
+      "epoch": 0.368705035971223,
+      "step": 410,
+      "torque_loss": 0.16924066841602325
+    },
+    {
+      "epoch": 0.368705035971223,
+      "force_loss": 0.005955790635198355,
+      "step": 410
+    },
+    {
+      "epoch": 0.3776978417266187,
+      "grad_norm": 0.99172443151474,
+      "learning_rate": 1.3966666666666666e-05,
+      "loss": 0.0612,
+      "step": 420
+    },
+    {
+      "action_loss": 0.046323012560606,
+      "epoch": 0.3776978417266187,
+      "step": 420
+    },
+    {
+      "epoch": 0.3776978417266187,
+      "step": 420,
+      "torque_loss": 0.13669370114803314
+    },
+    {
+      "epoch": 0.3776978417266187,
+      "force_loss": 0.008809066377580166,
+      "step": 420
+    },
+    {
+      "epoch": 0.38669064748201437,
+      "grad_norm": 0.7131702899932861,
+      "learning_rate": 1.43e-05,
+      "loss": 0.0557,
+      "step": 430
+    },
+    {
+      "action_loss": 0.06484134495258331,
+      "epoch": 0.38669064748201437,
+      "step": 430
+    },
+    {
+      "epoch": 0.38669064748201437,
+      "step": 430,
+      "torque_loss": 0.1730773001909256
+    },
+    {
+      "epoch": 0.38669064748201437,
+      "force_loss": 0.022902347147464752,
+      "step": 430
+    },
+    {
+      "epoch": 0.39568345323741005,
+      "grad_norm": 0.9485829472541809,
+      "learning_rate": 1.4633333333333334e-05,
+      "loss": 0.0576,
+      "step": 440
+    },
+    {
+      "action_loss": 0.0656992644071579,
+      "epoch": 0.39568345323741005,
+      "step": 440
+    },
+    {
+      "epoch": 0.39568345323741005,
+      "step": 440,
+      "torque_loss": 0.14429962635040283
+    },
+    {
+      "epoch": 0.39568345323741005,
+      "force_loss": 0.020188216120004654,
+      "step": 440
+    },
+    {
+      "epoch": 0.40467625899280574,
+      "grad_norm": 1.7684407234191895,
+      "learning_rate": 1.4966666666666668e-05,
+      "loss": 0.0572,
+      "step": 450
+    },
+    {
+      "action_loss": 0.03498392179608345,
+      "epoch": 0.40467625899280574,
+      "step": 450
+    },
+    {
+      "epoch": 0.40467625899280574,
+      "step": 450,
+      "torque_loss": 0.12798847258090973
+    },
+    {
+      "epoch": 0.40467625899280574,
+      "force_loss": 0.007528254296630621,
+      "step": 450
+    },
+    {
+      "epoch": 0.4136690647482014,
+      "grad_norm": 1.1263477802276611,
+      "learning_rate": 1.53e-05,
+      "loss": 0.0498,
+      "step": 460
+    },
+    {
+      "action_loss": 0.03849872946739197,
+      "epoch": 0.4136690647482014,
+      "step": 460
+    },
+    {
+      "epoch": 0.4136690647482014,
+      "step": 460,
+      "torque_loss": 0.1360778659582138
+    },
+    {
+      "epoch": 0.4136690647482014,
+      "force_loss": 0.015374623239040375,
+      "step": 460
+    },
+    {
+      "epoch": 0.4226618705035971,
+      "grad_norm": 0.7700405716896057,
+      "learning_rate": 1.563333333333333e-05,
+      "loss": 0.0518,
+      "step": 470
+    },
+    {
+      "action_loss": 0.019634079188108444,
+      "epoch": 0.4226618705035971,
+      "step": 470
+    },
+    {
+      "epoch": 0.4226618705035971,
+      "step": 470,
+      "torque_loss": 0.09089621156454086
+    },
+    {
+      "epoch": 0.4226618705035971,
+      "force_loss": 0.004453516099601984,
+      "step": 470
+    },
+    {
+      "epoch": 0.4316546762589928,
+      "grad_norm": 1.2913683652877808,
+      "learning_rate": 1.5966666666666667e-05,
+      "loss": 0.0433,
+      "step": 480
+    },
+    {
+      "action_loss": 0.02762473188340664,
+      "epoch": 0.4316546762589928,
+      "step": 480
+    },
+    {
+      "epoch": 0.4316546762589928,
+      "step": 480,
+      "torque_loss": 0.14060725271701813
+    },
+    {
+      "epoch": 0.4316546762589928,
+      "force_loss": 0.007750004529953003,
+      "step": 480
+    },
+    {
+      "epoch": 0.44064748201438847,
+      "grad_norm": 1.2174705266952515,
+      "learning_rate": 1.63e-05,
+      "loss": 0.052,
+      "step": 490
+    },
+    {
+      "action_loss": 0.04802360013127327,
+      "epoch": 0.44064748201438847,
+      "step": 490
+    },
+    {
+      "epoch": 0.44064748201438847,
+      "step": 490,
+      "torque_loss": 0.14837199449539185
+    },
+    {
+      "epoch": 0.44064748201438847,
+      "force_loss": 0.016758328303694725,
+      "step": 490
+    },
+    {
+      "epoch": 0.44964028776978415,
+      "grad_norm": 1.0038282871246338,
+      "learning_rate": 1.6633333333333336e-05,
+      "loss": 0.0426,
+      "step": 500
+    },
+    {
+      "action_loss": 0.03698794171214104,
+      "epoch": 0.44964028776978415,
+      "step": 500
+    },
+    {
+      "epoch": 0.44964028776978415,
+      "step": 500,
+      "torque_loss": 0.11550172418355942
+    },
+    {
+      "epoch": 0.44964028776978415,
+      "force_loss": 0.009756727144122124,
+      "step": 500
+    },
+    {
+      "epoch": 0.45863309352517984,
+      "grad_norm": 1.243269443511963,
+      "learning_rate": 1.6966666666666668e-05,
+      "loss": 0.048,
+      "step": 510
+    },
+    {
+      "action_loss": 0.028438905254006386,
+      "epoch": 0.45863309352517984,
+      "step": 510
+    },
+    {
+      "epoch": 0.45863309352517984,
+      "step": 510,
+      "torque_loss": 0.10196426510810852
+    },
+    {
+      "epoch": 0.45863309352517984,
+      "force_loss": 0.005013458896428347,
+      "step": 510
+    },
+    {
+      "epoch": 0.4676258992805755,
+      "grad_norm": 1.2900280952453613,
+      "learning_rate": 1.73e-05,
+      "loss": 0.0418,
+      "step": 520
+    },
+    {
+      "action_loss": 0.034887660294771194,
+      "epoch": 0.4676258992805755,
+      "step": 520
+    },
+    {
+      "epoch": 0.4676258992805755,
+      "step": 520,
+      "torque_loss": 0.12301575392484665
+    },
+    {
+      "epoch": 0.4676258992805755,
+      "force_loss": 0.008527296595275402,
+      "step": 520
+    },
+    {
+      "epoch": 0.4766187050359712,
+      "grad_norm": 0.5747576951980591,
+      "learning_rate": 1.7633333333333336e-05,
+      "loss": 0.05,
+      "step": 530
+    },
+    {
+      "action_loss": 0.05602313578128815,
+      "epoch": 0.4766187050359712,
+      "step": 530
+    },
+    {
+      "epoch": 0.4766187050359712,
+      "step": 530,
+      "torque_loss": 0.15821845829486847
+    },
+    {
+      "epoch": 0.4766187050359712,
+      "force_loss": 0.01597823202610016,
+      "step": 530
+    },
+    {
+      "epoch": 0.4856115107913669,
+      "grad_norm": 0.7638629674911499,
+      "learning_rate": 1.796666666666667e-05,
+      "loss": 0.0564,
+      "step": 540
+    },
+    {
+      "action_loss": 0.04084310308098793,
+      "epoch": 0.4856115107913669,
+      "step": 540
+    },
+    {
+      "epoch": 0.4856115107913669,
+      "step": 540,
+      "torque_loss": 0.1660570353269577
+    },
+    {
+      "epoch": 0.4856115107913669,
+      "force_loss": 0.01596181094646454,
+      "step": 540
+    },
+    {
+      "epoch": 0.49460431654676257,
+      "grad_norm": 1.6049985885620117,
+      "learning_rate": 1.83e-05,
+      "loss": 0.0453,
+      "step": 550
+    },
+    {
+      "action_loss": 0.045941516757011414,
+      "epoch": 0.49460431654676257,
+      "step": 550
+    },
+    {
+      "epoch": 0.49460431654676257,
+      "step": 550,
+      "torque_loss": 0.13734887540340424
+    },
+    {
+      "epoch": 0.49460431654676257,
+      "force_loss": 0.02087695710361004,
+      "step": 550
+    },
+    {
+      "epoch": 0.5035971223021583,
+      "grad_norm": 0.9019855260848999,
+      "learning_rate": 1.8633333333333333e-05,
+      "loss": 0.048,
+      "step": 560
+    },
+    {
+      "action_loss": 0.017615145072340965,
+      "epoch": 0.5035971223021583,
+      "step": 560
+    },
+    {
+      "epoch": 0.5035971223021583,
+      "step": 560,
+      "torque_loss": 0.16029873490333557
+    },
+    {
+      "epoch": 0.5035971223021583,
+      "force_loss": 0.005696360487490892,
+      "step": 560
+    },
+    {
+      "epoch": 0.512589928057554,
+      "grad_norm": 0.8535636067390442,
+      "learning_rate": 1.896666666666667e-05,
+      "loss": 0.0563,
+      "step": 570
+    },
+    {
+      "action_loss": 0.031458910554647446,
+      "epoch": 0.512589928057554,
+      "step": 570
+    },
+    {
+      "epoch": 0.512589928057554,
+      "step": 570,
+      "torque_loss": 0.14906558394432068
+    },
+    {
+      "epoch": 0.512589928057554,
+      "force_loss": 0.010886810719966888,
+      "step": 570
+    },
+    {
+      "epoch": 0.5215827338129496,
+      "grad_norm": 0.8024583458900452,
+      "learning_rate": 1.93e-05,
+      "loss": 0.0499,
+      "step": 580
+    },
+    {
+      "action_loss": 0.03816935420036316,
+      "epoch": 0.5215827338129496,
+      "step": 580
+    },
+    {
+      "epoch": 0.5215827338129496,
+      "step": 580,
+      "torque_loss": 0.1175684705376625
+    },
+    {
+      "epoch": 0.5215827338129496,
+      "force_loss": 0.010934433899819851,
+      "step": 580
+    },
+    {
+      "epoch": 0.5305755395683454,
+      "grad_norm": 2.530550956726074,
+      "learning_rate": 1.9633333333333334e-05,
+      "loss": 0.0448,
+      "step": 590
+    },
+    {
+      "action_loss": 0.03544415533542633,
+      "epoch": 0.5305755395683454,
+      "step": 590
+    },
+    {
+      "epoch": 0.5305755395683454,
+      "step": 590,
+      "torque_loss": 0.18601089715957642
+    },
+    {
+      "epoch": 0.5305755395683454,
+      "force_loss": 0.007891523651778698,
+      "step": 590
+    },
+    {
+      "epoch": 0.539568345323741,
+      "grad_norm": 1.0873814821243286,
+      "learning_rate": 1.9966666666666666e-05,
+      "loss": 0.0524,
+      "step": 600
+    },
+    {
+      "action_loss": 0.0685107633471489,
+      "epoch": 0.539568345323741,
+      "step": 600
+    },
+    {
+      "epoch": 0.539568345323741,
+      "step": 600,
+      "torque_loss": 0.2041873335838318
+    },
+    {
+      "epoch": 0.539568345323741,
+      "force_loss": 0.051076740026474,
+      "step": 600
+    },
+    {
+      "epoch": 0.5485611510791367,
+      "grad_norm": 1.2049381732940674,
+      "learning_rate": 2.0300000000000002e-05,
+      "loss": 0.0524,
+      "step": 610
+    },
+    {
+      "action_loss": 0.045146096497774124,
+      "epoch": 0.5485611510791367,
+      "step": 610
+    },
+    {
+      "epoch": 0.5485611510791367,
+      "step": 610,
+      "torque_loss": 0.1632339060306549
+    },
+    {
+      "epoch": 0.5485611510791367,
+      "force_loss": 0.019419750198721886,
+      "step": 610
+    },
+    {
+      "epoch": 0.5575539568345323,
+      "grad_norm": 1.6513888835906982,
+      "learning_rate": 2.0633333333333335e-05,
+      "loss": 0.0495,
+      "step": 620
+    },
+    {
+      "action_loss": 0.023775087669491768,
+      "epoch": 0.5575539568345323,
+      "step": 620
+    },
+    {
+      "epoch": 0.5575539568345323,
+      "step": 620,
+      "torque_loss": 0.1636258214712143
+    },
+    {
+      "epoch": 0.5575539568345323,
+      "force_loss": 0.007806839887052774,
+      "step": 620
+    },
+    {
+      "epoch": 0.5665467625899281,
+      "grad_norm": 1.9310189485549927,
+      "learning_rate": 2.0966666666666667e-05,
+      "loss": 0.047,
+      "step": 630
+    },
+    {
+      "action_loss": 0.025371715426445007,
+      "epoch": 0.5665467625899281,
+      "step": 630
+    },
+    {
+      "epoch": 0.5665467625899281,
+      "step": 630,
+      "torque_loss": 0.11403391510248184
+    },
+    {
+      "epoch": 0.5665467625899281,
+      "force_loss": 0.008687383495271206,
+      "step": 630
+    },
+    {
+      "epoch": 0.5755395683453237,
+      "grad_norm": 0.7792282104492188,
+      "learning_rate": 2.13e-05,
+      "loss": 0.0448,
+      "step": 640
+    },
+    {
+      "action_loss": 0.04153499752283096,
+      "epoch": 0.5755395683453237,
+      "step": 640
+    },
+    {
+      "epoch": 0.5755395683453237,
+      "step": 640,
+      "torque_loss": 0.1256234049797058
+    },
+    {
+      "epoch": 0.5755395683453237,
+      "force_loss": 0.02491765283048153,
+      "step": 640
+    },
+    {
+      "epoch": 0.5845323741007195,
+      "grad_norm": 0.8404738903045654,
+      "learning_rate": 2.1633333333333332e-05,
+      "loss": 0.0429,
+      "step": 650
+    },
+    {
+      "action_loss": 0.012981973588466644,
+      "epoch": 0.5845323741007195,
+      "step": 650
+    },
+    {
+      "epoch": 0.5845323741007195,
+      "step": 650,
+      "torque_loss": 0.1623903065919876
+    },
+    {
+      "epoch": 0.5845323741007195,
+      "force_loss": 0.004864631220698357,
+      "step": 650
+    },
+    {
+      "epoch": 0.5935251798561151,
+      "grad_norm": 1.0602569580078125,
+      "learning_rate": 2.1966666666666668e-05,
+      "loss": 0.0444,
+      "step": 660
+    },
+    {
+      "action_loss": 0.02391764521598816,
+      "epoch": 0.5935251798561151,
+      "step": 660
+    },
+    {
+      "epoch": 0.5935251798561151,
+      "step": 660,
+      "torque_loss": 0.10528260469436646
+    },
+    {
+      "epoch": 0.5935251798561151,
+      "force_loss": 0.006534598767757416,
+      "step": 660
+    },
+    {
+      "epoch": 0.6025179856115108,
+      "grad_norm": 0.9499192833900452,
+      "learning_rate": 2.23e-05,
+      "loss": 0.0338,
+      "step": 670
+    },
+    {
+      "action_loss": 0.04245607927441597,
+      "epoch": 0.6025179856115108,
+      "step": 670
+    },
+    {
+      "epoch": 0.6025179856115108,
+      "step": 670,
+      "torque_loss": 0.21461541950702667
+    },
+    {
+      "epoch": 0.6025179856115108,
+      "force_loss": 0.01296437531709671,
+      "step": 670
+    },
+    {
+      "epoch": 0.6115107913669064,
+      "grad_norm": 2.2330446243286133,
+      "learning_rate": 2.2633333333333336e-05,
+      "loss": 0.0505,
+      "step": 680
+    },
+    {
+      "action_loss": 0.03410147503018379,
+      "epoch": 0.6115107913669064,
+      "step": 680
+    },
+    {
+      "epoch": 0.6115107913669064,
+      "step": 680,
+      "torque_loss": 0.1426069587469101
+    },
+    {
+      "epoch": 0.6115107913669064,
+      "force_loss": 0.016774063929915428,
+      "step": 680
+    },
+    {
+      "epoch": 0.6205035971223022,
+      "grad_norm": 1.304166555404663,
+      "learning_rate": 2.2966666666666668e-05,
+      "loss": 0.0391,
+      "step": 690
+    },
+    {
+      "action_loss": 0.030235351994633675,
+      "epoch": 0.6205035971223022,
+      "step": 690
+    },
+    {
+      "epoch": 0.6205035971223022,
+      "step": 690,
+      "torque_loss": 0.10690677165985107
+    },
+    {
+      "epoch": 0.6205035971223022,
+      "force_loss": 0.011885077692568302,
+      "step": 690
+    },
+    {
+      "epoch": 0.6294964028776978,
+      "grad_norm": 2.1882855892181396,
+      "learning_rate": 2.3300000000000004e-05,
+      "loss": 0.0415,
+      "step": 700
+    },
+    {
+      "action_loss": 0.014563643373548985,
+      "epoch": 0.6294964028776978,
+      "step": 700
+    },
+    {
+      "epoch": 0.6294964028776978,
+      "step": 700,
+      "torque_loss": 0.10989886522293091
+    },
+    {
+      "epoch": 0.6294964028776978,
+      "force_loss": 0.005410848185420036,
+      "step": 700
+    },
+    {
+      "epoch": 0.6384892086330936,
+      "grad_norm": 1.208815097808838,
+      "learning_rate": 2.3633333333333336e-05,
+      "loss": 0.0479,
+      "step": 710
+    },
+    {
+      "action_loss": 0.016269544139504433,
+      "epoch": 0.6384892086330936,
+      "step": 710
+    },
+    {
+      "epoch": 0.6384892086330936,
+      "step": 710,
+      "torque_loss": 0.10121416300535202
+    },
+    {
+      "epoch": 0.6384892086330936,
+      "force_loss": 0.005640504416078329,
+      "step": 710
+    },
+    {
+      "epoch": 0.6474820143884892,
+      "grad_norm": 3.213498830795288,
+      "learning_rate": 2.396666666666667e-05,
+      "loss": 0.0416,
+      "step": 720
+    },
+    {
+      "action_loss": 0.06165416166186333,
+      "epoch": 0.6474820143884892,
+      "step": 720
+    },
+    {
+      "epoch": 0.6474820143884892,
+      "step": 720,
+      "torque_loss": 0.22449012100696564
+    },
+    {
+      "epoch": 0.6474820143884892,
+      "force_loss": 0.012106549926102161,
+      "step": 720
+    },
+    {
+      "epoch": 0.6564748201438849,
+      "grad_norm": 1.0452207326889038,
+      "learning_rate": 2.43e-05,
+      "loss": 0.0515,
+      "step": 730
+    },
+    {
+      "action_loss": 0.043337106704711914,
+      "epoch": 0.6564748201438849,
+      "step": 730
+    },
+    {
+      "epoch": 0.6564748201438849,
+      "step": 730,
+      "torque_loss": 0.1278170943260193
+    },
+    {
+      "epoch": 0.6564748201438849,
+      "force_loss": 0.03837411105632782,
+      "step": 730
+    },
+    {
+      "epoch": 0.6654676258992805,
+      "grad_norm": 1.0941096544265747,
+      "learning_rate": 2.4633333333333334e-05,
+      "loss": 0.0446,
+      "step": 740
+    },
+    {
+      "action_loss": 0.02208028733730316,
+      "epoch": 0.6654676258992805,
+      "step": 740
+    },
+    {
+      "epoch": 0.6654676258992805,
+      "step": 740,
+      "torque_loss": 0.09663611650466919
+    },
+    {
+      "epoch": 0.6654676258992805,
+      "force_loss": 0.01026022806763649,
+      "step": 740
+    },
+    {
+      "epoch": 0.6744604316546763,
+      "grad_norm": 0.7206782698631287,
+      "learning_rate": 2.496666666666667e-05,
+      "loss": 0.0433,
+      "step": 750
+    },
+    {
+      "action_loss": 0.016235915943980217,
+      "epoch": 0.6744604316546763,
+      "step": 750
+    },
+    {
+      "epoch": 0.6744604316546763,
+      "step": 750,
+      "torque_loss": 0.12707258760929108
+    },
+    {
+      "epoch": 0.6744604316546763,
+      "force_loss": 0.004836099687963724,
+      "step": 750
+    },
+    {
+      "epoch": 0.6834532374100719,
+      "grad_norm": 1.1924729347229004,
+      "learning_rate": 2.5300000000000002e-05,
+      "loss": 0.0405,
+      "step": 760
+    },
+    {
+      "action_loss": 0.04032515361905098,
+      "epoch": 0.6834532374100719,
+      "step": 760
+    },
+    {
+      "epoch": 0.6834532374100719,
+      "step": 760,
+      "torque_loss": 0.12080534547567368
+    },
+    {
+      "epoch": 0.6834532374100719,
+      "force_loss": 0.009603076614439487,
+      "step": 760
+    },
+    {
+      "epoch": 0.6924460431654677,
+      "grad_norm": 1.8415714502334595,
+      "learning_rate": 2.5633333333333338e-05,
+      "loss": 0.0459,
+      "step": 770
+    },
+    {
+      "action_loss": 0.017865577712655067,
+      "epoch": 0.6924460431654677,
+      "step": 770
+    },
+    {
+      "epoch": 0.6924460431654677,
+      "step": 770,
+      "torque_loss": 0.15025244653224945
+    },
+    {
+      "epoch": 0.6924460431654677,
+      "force_loss": 0.0042395382188260555,
+      "step": 770
+    },
+    {
+      "epoch": 0.7014388489208633,
+      "grad_norm": 0.8931247591972351,
+      "learning_rate": 2.5966666666666667e-05,
+      "loss": 0.0507,
+      "step": 780
+    },
+    {
+      "action_loss": 0.014569099061191082,
+      "epoch": 0.7014388489208633,
+      "step": 780
+    },
+    {
+      "epoch": 0.7014388489208633,
+      "step": 780,
+      "torque_loss": 0.08025429397821426
+    },
+    {
+      "epoch": 0.7014388489208633,
+      "force_loss": 0.004249536897987127,
+      "step": 780
+    },
+    {
+      "epoch": 0.710431654676259,
+      "grad_norm": 1.091386318206787,
+      "learning_rate": 2.6300000000000002e-05,
+      "loss": 0.0392,
+      "step": 790
+    },
+    {
+      "action_loss": 0.03355321288108826,
+      "epoch": 0.710431654676259,
+      "step": 790
+    },
+    {
+      "epoch": 0.710431654676259,
+      "step": 790,
+      "torque_loss": 0.16912581026554108
+    },
+    {
+      "epoch": 0.710431654676259,
+      "force_loss": 0.010022146627306938,
+      "step": 790
+    },
+    {
+      "epoch": 0.7194244604316546,
+      "grad_norm": 0.7468305826187134,
+      "learning_rate": 2.663333333333333e-05,
+      "loss": 0.0461,
+      "step": 800
+    },
+    {
+      "action_loss": 0.01833420991897583,
+      "epoch": 0.7194244604316546,
+      "step": 800
+    },
+    {
+      "epoch": 0.7194244604316546,
+      "step": 800,
+      "torque_loss": 0.08790832757949829
+    },
+    {
+      "epoch": 0.7194244604316546,
+      "force_loss": 0.006794957909733057,
+      "step": 800
+    },
+    {
+      "epoch": 0.7284172661870504,
+      "grad_norm": 0.9255871772766113,
+      "learning_rate": 2.6966666666666667e-05,
+      "loss": 0.0393,
+      "step": 810
+    },
+    {
+      "action_loss": 0.05293245241045952,
+      "epoch": 0.7284172661870504,
+      "step": 810
+    },
+    {
+      "epoch": 0.7284172661870504,
+      "step": 810,
+      "torque_loss": 0.14482195675373077
+    },
+    {
+      "epoch": 0.7284172661870504,
+      "force_loss": 0.023813588544726372,
+      "step": 810
+    },
+    {
+      "epoch": 0.737410071942446,
+      "grad_norm": 1.567159652709961,
+      "learning_rate": 2.7300000000000003e-05,
+      "loss": 0.0581,
+      "step": 820
+    },
+    {
+      "action_loss": 0.009122147224843502,
+      "epoch": 0.737410071942446,
+      "step": 820
+    },
+    {
+      "epoch": 0.737410071942446,
+      "step": 820,
+      "torque_loss": 0.10559693723917007
+    },
+    {
+      "epoch": 0.737410071942446,
+      "force_loss": 0.004394026938825846,
+      "step": 820
+    },
+    {
+      "epoch": 0.7464028776978417,
+      "grad_norm": 0.9150029420852661,
+      "learning_rate": 2.7633333333333332e-05,
+      "loss": 0.0449,
+      "step": 830
+    },
+    {
+      "action_loss": 0.014827902428805828,
+      "epoch": 0.7464028776978417,
+      "step": 830
+    },
+    {
+      "epoch": 0.7464028776978417,
+      "step": 830,
+      "torque_loss": 0.13311408460140228
+    },
+    {
+      "epoch": 0.7464028776978417,
+      "force_loss": 0.005864080041646957,
+      "step": 830
+    },
+    {
+      "epoch": 0.7553956834532374,
+      "grad_norm": 1.084446668624878,
+      "learning_rate": 2.7966666666666668e-05,
+      "loss": 0.0367,
+      "step": 840
+    },
+    {
+      "action_loss": 0.032788172364234924,
+      "epoch": 0.7553956834532374,
+      "step": 840
+    },
+    {
+      "epoch": 0.7553956834532374,
+      "step": 840,
+      "torque_loss": 0.1256210207939148
+    },
+    {
+      "epoch": 0.7553956834532374,
+      "force_loss": 0.012543238699436188,
+      "step": 840
+    },
+    {
+      "epoch": 0.7643884892086331,
+      "grad_norm": 0.7480943202972412,
+      "learning_rate": 2.83e-05,
+      "loss": 0.0408,
+      "step": 850
+    },
+    {
+      "action_loss": 0.027371756732463837,
+      "epoch": 0.7643884892086331,
+      "step": 850
+    },
+    {
+      "epoch": 0.7643884892086331,
+      "step": 850,
+      "torque_loss": 0.11386056989431381
+    },
+    {
+      "epoch": 0.7643884892086331,
+      "force_loss": 0.022196611389517784,
+      "step": 850
+    },
+    {
+      "epoch": 0.7733812949640287,
+      "grad_norm": 1.3565993309020996,
+      "learning_rate": 2.8633333333333336e-05,
+      "loss": 0.0486,
+      "step": 860
+    },
+    {
+      "action_loss": 0.034494709223508835,
+      "epoch": 0.7733812949640287,
+      "step": 860
+    },
+    {
+      "epoch": 0.7733812949640287,
+      "step": 860,
+      "torque_loss": 0.13771109282970428
+    },
+    {
+      "epoch": 0.7733812949640287,
+      "force_loss": 0.01090602669864893,
+      "step": 860
+    },
+    {
+      "epoch": 0.7823741007194245,
+      "grad_norm": 1.561873435974121,
+      "learning_rate": 2.8966666666666668e-05,
+      "loss": 0.0377,
+      "step": 870
+    },
+    {
+      "action_loss": 0.039113353937864304,
+      "epoch": 0.7823741007194245,
+      "step": 870
+    },
+    {
+      "epoch": 0.7823741007194245,
+      "step": 870,
+      "torque_loss": 0.08401709794998169
+    },
+    {
+      "epoch": 0.7823741007194245,
+      "force_loss": 0.025932809337973595,
+      "step": 870
+    },
+    {
+      "epoch": 0.7913669064748201,
+      "grad_norm": 1.2967007160186768,
+      "learning_rate": 2.93e-05,
+      "loss": 0.0427,
+      "step": 880
+    },
+    {
+      "action_loss": 0.04003247246146202,
+      "epoch": 0.7913669064748201,
+      "step": 880
+    },
+    {
+      "epoch": 0.7913669064748201,
+      "step": 880,
+      "torque_loss": 0.10898923873901367
+    },
+    {
+      "epoch": 0.7913669064748201,
+      "force_loss": 0.008176526986062527,
+      "step": 880
+    },
+    {
+      "epoch": 0.8003597122302158,
+      "grad_norm": 1.5075591802597046,
+      "learning_rate": 2.9633333333333336e-05,
+      "loss": 0.0413,
+      "step": 890
+    },
+    {
+      "action_loss": 0.02105006016790867,
+      "epoch": 0.8003597122302158,
+      "step": 890
+    },
+    {
+      "epoch": 0.8003597122302158,
+      "step": 890,
+      "torque_loss": 0.09274104237556458
+    },
+    {
+      "epoch": 0.8003597122302158,
+      "force_loss": 0.004790377337485552,
+      "step": 890
+    },
+    {
+      "epoch": 0.8093525179856115,
+      "grad_norm": 1.1103328466415405,
+      "learning_rate": 2.9966666666666672e-05,
+      "loss": 0.0417,
+      "step": 900
+    },
+    {
+      "action_loss": 0.04317578300833702,
+      "epoch": 0.8093525179856115,
+      "step": 900
+    },
+    {
+      "epoch": 0.8093525179856115,
+      "step": 900,
+      "torque_loss": 0.1761828064918518
+    },
+    {
+      "epoch": 0.8093525179856115,
+      "force_loss": 0.017960799857974052,
+      "step": 900
+    },
+    {
+      "epoch": 0.8183453237410072,
+      "grad_norm": 0.800326943397522,
+      "learning_rate": 3.03e-05,
+      "loss": 0.0418,
+      "step": 910
+    },
+    {
+      "action_loss": 0.02699863910675049,
+      "epoch": 0.8183453237410072,
+      "step": 910
+    },
+    {
+      "epoch": 0.8183453237410072,
+      "step": 910,
+      "torque_loss": 0.14901010692119598
+    },
+    {
+      "epoch": 0.8183453237410072,
+      "force_loss": 0.012698336504399776,
+      "step": 910
+    },
+    {
+      "epoch": 0.8273381294964028,
+      "grad_norm": 0.9114958643913269,
+      "learning_rate": 3.063333333333334e-05,
+      "loss": 0.0504,
+      "step": 920
+    },
+    {
+      "action_loss": 0.0285883080214262,
+      "epoch": 0.8273381294964028,
+      "step": 920
+    },
+    {
+      "epoch": 0.8273381294964028,
+      "step": 920,
+      "torque_loss": 0.1757916361093521
+    },
+    {
+      "epoch": 0.8273381294964028,
+      "force_loss": 0.0068364874459803104,
+      "step": 920
+    },
+    {
+      "epoch": 0.8363309352517986,
+      "grad_norm": 1.1808520555496216,
+      "learning_rate": 3.096666666666666e-05,
+      "loss": 0.0542,
+      "step": 930
+    },
+    {
+      "action_loss": 0.02321232296526432,
+      "epoch": 0.8363309352517986,
+      "step": 930
+    },
+    {
+      "epoch": 0.8363309352517986,
+      "step": 930,
+      "torque_loss": 0.1254875212907791
+    },
+    {
+      "epoch": 0.8363309352517986,
+      "force_loss": 0.012459876947104931,
+      "step": 930
+    },
+    {
+      "epoch": 0.8453237410071942,
+      "grad_norm": 0.8187349438667297,
+      "learning_rate": 3.13e-05,
+      "loss": 0.0386,
+      "step": 940
+    },
+    {
+      "action_loss": 0.042479436844587326,
+      "epoch": 0.8453237410071942,
+      "step": 940
+    },
+    {
+      "epoch": 0.8453237410071942,
+      "step": 940,
+      "torque_loss": 0.1373445987701416
+    },
+    {
+      "epoch": 0.8453237410071942,
+      "force_loss": 0.02711704932153225,
+      "step": 940
+    },
+    {
+      "epoch": 0.85431654676259,
+      "grad_norm": 1.4400840997695923,
+      "learning_rate": 3.1633333333333334e-05,
+      "loss": 0.0377,
+      "step": 950
+    },
+    {
+      "action_loss": 0.039127472788095474,
+      "epoch": 0.85431654676259,
+      "step": 950
+    },
+    {
+      "epoch": 0.85431654676259,
+      "step": 950,
+      "torque_loss": 0.15974800288677216
+    },
+    {
+      "epoch": 0.85431654676259,
+      "force_loss": 0.01687513291835785,
+      "step": 950
+    },
+    {
+      "epoch": 0.8633093525179856,
+      "grad_norm": 1.0097832679748535,
+      "learning_rate": 3.196666666666667e-05,
+      "loss": 0.0455,
+      "step": 960
+    },
+    {
+      "action_loss": 0.02173447608947754,
+      "epoch": 0.8633093525179856,
+      "step": 960
+    },
+    {
+      "epoch": 0.8633093525179856,
+      "step": 960,
+      "torque_loss": 0.14499664306640625
+    },
+    {
+      "epoch": 0.8633093525179856,
+      "force_loss": 0.008056712336838245,
+      "step": 960
+    },
+    {
+      "epoch": 0.8723021582733813,
+      "grad_norm": 1.1613527536392212,
+      "learning_rate": 3.2300000000000006e-05,
+      "loss": 0.037,
+      "step": 970
+    },
+    {
+      "action_loss": 0.01486013364046812,
+      "epoch": 0.8723021582733813,
+      "step": 970
+    },
+    {
+      "epoch": 0.8723021582733813,
+      "step": 970,
+      "torque_loss": 0.13487625122070312
+    },
+    {
+      "epoch": 0.8723021582733813,
+      "force_loss": 0.006477018352597952,
+      "step": 970
+    },
+    {
+      "epoch": 0.8812949640287769,
+      "grad_norm": 1.1028674840927124,
+      "learning_rate": 3.263333333333333e-05,
+      "loss": 0.0403,
+      "step": 980
+    },
+    {
+      "action_loss": 0.015142365358769894,
+      "epoch": 0.8812949640287769,
+      "step": 980
+    },
+    {
+      "epoch": 0.8812949640287769,
+      "step": 980,
+      "torque_loss": 0.11137839406728745
+    },
+    {
+      "epoch": 0.8812949640287769,
+      "force_loss": 0.003947315271943808,
+      "step": 980
+    },
+    {
+      "epoch": 0.8902877697841727,
+      "grad_norm": 0.8612143993377686,
+      "learning_rate": 3.296666666666667e-05,
+      "loss": 0.041,
+      "step": 990
+    },
+    {
+      "action_loss": 0.01873861812055111,
+      "epoch": 0.8902877697841727,
+      "step": 990
+    },
+    {
+      "epoch": 0.8902877697841727,
+      "step": 990,
+      "torque_loss": 0.12820900976657867
+    },
+    {
+      "epoch": 0.8902877697841727,
+      "force_loss": 0.008802324533462524,
+      "step": 990
+    },
+    {
+      "epoch": 0.8992805755395683,
+      "grad_norm": 1.2536673545837402,
+      "learning_rate": 3.33e-05,
+      "loss": 0.0396,
+      "step": 1000
+    },
+    {
+      "action_loss": 0.043822210282087326,
+      "epoch": 0.8992805755395683,
+      "step": 1000
+    },
+    {
+      "epoch": 0.8992805755395683,
+      "step": 1000,
+      "torque_loss": 0.20643587410449982
+    },
+    {
+      "epoch": 0.8992805755395683,
+      "force_loss": 0.014320352114737034,
+      "step": 1000
+    },
+    {
+      "epoch": 0.908273381294964,
+      "grad_norm": 0.9552024602890015,
+      "learning_rate": 3.3633333333333335e-05,
+      "loss": 0.0469,
+      "step": 1010
+    },
+    {
+      "action_loss": 0.017567720264196396,
+      "epoch": 0.908273381294964,
+      "step": 1010
+    },
+    {
+      "epoch": 0.908273381294964,
+      "step": 1010,
+      "torque_loss": 0.11810340732336044
+    },
+    {
+      "epoch": 0.908273381294964,
+      "force_loss": 0.010300123132765293,
+      "step": 1010
+    },
+    {
+      "epoch": 0.9172661870503597,
+      "grad_norm": 0.49093523621559143,
+      "learning_rate": 3.396666666666667e-05,
+      "loss": 0.0375,
+      "step": 1020
+    },
+    {
+      "action_loss": 0.02955763228237629,
+      "epoch": 0.9172661870503597,
+      "step": 1020
+    },
+    {
+      "epoch": 0.9172661870503597,
+      "step": 1020,
+      "torque_loss": 0.11659219115972519
+    },
+    {
+      "epoch": 0.9172661870503597,
+      "force_loss": 0.009206675924360752,
+      "step": 1020
+    },
+    {
+      "epoch": 0.9262589928057554,
+      "grad_norm": 1.5677725076675415,
+      "learning_rate": 3.430000000000001e-05,
+      "loss": 0.0402,
+      "step": 1030
+    },
+    {
+      "action_loss": 0.014339379034936428,
+      "epoch": 0.9262589928057554,
+      "step": 1030
+    },
+    {
+      "epoch": 0.9262589928057554,
+      "step": 1030,
+      "torque_loss": 0.07817431539297104
+    },
+    {
+      "epoch": 0.9262589928057554,
+      "force_loss": 0.008804906159639359,
+      "step": 1030
+    },
+    {
+      "epoch": 0.935251798561151,
+      "grad_norm": 3.022254705429077,
+      "learning_rate": 3.463333333333333e-05,
+      "loss": 0.0382,
+      "step": 1040
+    },
+    {
+      "action_loss": 0.0204868633300066,
+      "epoch": 0.935251798561151,
+      "step": 1040
+    },
+    {
+      "epoch": 0.935251798561151,
+      "step": 1040,
+      "torque_loss": 0.10651201009750366
+    },
+    {
+      "epoch": 0.935251798561151,
+      "force_loss": 0.009721930138766766,
+      "step": 1040
+    },
+    {
+      "epoch": 0.9442446043165468,
+      "grad_norm": 1.391775369644165,
+      "learning_rate": 3.496666666666667e-05,
+      "loss": 0.0402,
+      "step": 1050
+    },
+    {
+      "action_loss": 0.024866757914423943,
+      "epoch": 0.9442446043165468,
+      "step": 1050
+    },
+    {
+      "epoch": 0.9442446043165468,
+      "step": 1050,
+      "torque_loss": 0.17966289818286896
+    },
+    {
+      "epoch": 0.9442446043165468,
+      "force_loss": 0.012933402322232723,
+      "step": 1050
+    },
+    {
+      "epoch": 0.9532374100719424,
+      "grad_norm": 0.8274032473564148,
+      "learning_rate": 3.53e-05,
+      "loss": 0.0477,
+      "step": 1060
+    },
+    {
+      "action_loss": 0.012262926436960697,
+      "epoch": 0.9532374100719424,
+      "step": 1060
+    },
+    {
+      "epoch": 0.9532374100719424,
+      "step": 1060,
+      "torque_loss": 0.09987610578536987
+    },
+    {
+      "epoch": 0.9532374100719424,
+      "force_loss": 0.004758056718856096,
+      "step": 1060
+    },
+    {
+      "epoch": 0.9622302158273381,
+      "grad_norm": 0.854238748550415,
+      "learning_rate": 3.563333333333334e-05,
+      "loss": 0.0356,
+      "step": 1070
+    },
+    {
+      "action_loss": 0.045313760638237,
+      "epoch": 0.9622302158273381,
+      "step": 1070
+    },
+    {
+      "epoch": 0.9622302158273381,
+      "step": 1070,
+      "torque_loss": 0.20544810593128204
+    },
+    {
+      "epoch": 0.9622302158273381,
+      "force_loss": 0.01412518322467804,
+      "step": 1070
+    },
+    {
+      "epoch": 0.9712230215827338,
+      "grad_norm": 1.3108383417129517,
+      "learning_rate": 3.596666666666667e-05,
+      "loss": 0.0462,
+      "step": 1080
+    },
+    {
+      "action_loss": 0.018526015803217888,
+      "epoch": 0.9712230215827338,
+      "step": 1080
+    },
+    {
+      "epoch": 0.9712230215827338,
+      "step": 1080,
+      "torque_loss": 0.12145505100488663
+    },
+    {
+      "epoch": 0.9712230215827338,
+      "force_loss": 0.006285226438194513,
+      "step": 1080
+    },
+    {
+      "epoch": 0.9802158273381295,
+      "grad_norm": 1.2655175924301147,
+      "learning_rate": 3.63e-05,
+      "loss": 0.0434,
+      "step": 1090
+    },
+    {
+      "action_loss": 0.03329155221581459,
+      "epoch": 0.9802158273381295,
+      "step": 1090
+    },
+    {
+      "epoch": 0.9802158273381295,
+      "step": 1090,
+      "torque_loss": 0.1332734078168869
+    },
+    {
+      "epoch": 0.9802158273381295,
+      "force_loss": 0.01068316400051117,
+      "step": 1090
+    },
+    {
+      "epoch": 0.9892086330935251,
+      "grad_norm": 0.9130733013153076,
+      "learning_rate": 3.6633333333333334e-05,
+      "loss": 0.0383,
+      "step": 1100
+    },
+    {
+      "action_loss": 0.0344303734600544,
+      "epoch": 0.9892086330935251,
+      "step": 1100
+    },
+    {
+      "epoch": 0.9892086330935251,
+      "step": 1100,
+      "torque_loss": 0.13855507969856262
+    },
+    {
+      "epoch": 0.9892086330935251,
+      "force_loss": 0.011614271439611912,
+      "step": 1100
+    },
+    {
+      "epoch": 0.9982014388489209,
+      "grad_norm": 0.6374790072441101,
+      "learning_rate": 3.6966666666666666e-05,
+      "loss": 0.0372,
+      "step": 1110
+    },
+    {
+      "action_loss": 0.01947695203125477,
+      "epoch": 0.9982014388489209,
+      "step": 1110
+    },
+    {
+      "epoch": 0.9982014388489209,
+      "step": 1110,
+      "torque_loss": 0.1781420260667801
+    },
+    {
+      "epoch": 0.9982014388489209,
+      "force_loss": 0.004689342807978392,
+      "step": 1110
+    },
+    {
+      "epoch": 1.0071942446043165,
+      "grad_norm": 1.2495883703231812,
+      "learning_rate": 3.73e-05,
+      "loss": 0.0364,
+      "step": 1120
+    },
+    {
+      "action_loss": 0.011348418891429901,
+      "epoch": 1.0071942446043165,
+      "step": 1120
+    },
+    {
+      "epoch": 1.0071942446043165,
+      "step": 1120,
+      "torque_loss": 0.12806175649166107
+    },
+    {
+      "epoch": 1.0071942446043165,
+      "force_loss": 0.005668466445058584,
+      "step": 1120
+    },
+    {
+      "epoch": 1.0161870503597121,
+      "grad_norm": 1.0541917085647583,
+      "learning_rate": 3.763333333333334e-05,
+      "loss": 0.0364,
+      "step": 1130
+    },
+    {
+      "action_loss": 0.017578043043613434,
+      "epoch": 1.0161870503597121,
+      "step": 1130
+    },
+    {
+      "epoch": 1.0161870503597121,
+      "step": 1130,
+      "torque_loss": 0.09509783983230591
+    },
+    {
+      "epoch": 1.0161870503597121,
+      "force_loss": 0.00580967403948307,
+      "step": 1130
+    },
+    {
+      "epoch": 1.025179856115108,
+      "grad_norm": 1.0330699682235718,
+      "learning_rate": 3.796666666666667e-05,
+      "loss": 0.0388,
+      "step": 1140
+    },
+    {
+      "action_loss": 0.039993640035390854,
+      "epoch": 1.025179856115108,
+      "step": 1140
+    },
+    {
+      "epoch": 1.025179856115108,
+      "step": 1140,
+      "torque_loss": 0.13236945867538452
+    },
+    {
+      "epoch": 1.025179856115108,
+      "force_loss": 0.0089992955327034,
+      "step": 1140
+    },
+    {
+      "epoch": 1.0341726618705036,
+      "grad_norm": 1.6210577487945557,
+      "learning_rate": 3.83e-05,
+      "loss": 0.0395,
+      "step": 1150
+    },
+    {
+      "action_loss": 0.020890899002552032,
+      "epoch": 1.0341726618705036,
+      "step": 1150
+    },
+    {
+      "epoch": 1.0341726618705036,
+      "step": 1150,
+      "torque_loss": 0.15326978266239166
+    },
+    {
+      "epoch": 1.0341726618705036,
+      "force_loss": 0.006767446640878916,
+      "step": 1150
+    },
+    {
+      "epoch": 1.0431654676258992,
+      "grad_norm": 1.3818142414093018,
+      "learning_rate": 3.8633333333333335e-05,
+      "loss": 0.0405,
+      "step": 1160
+    },
+    {
+      "action_loss": 0.02695208229124546,
+      "epoch": 1.0431654676258992,
+      "step": 1160
+    },
+    {
+      "epoch": 1.0431654676258992,
+      "step": 1160,
+      "torque_loss": 0.19758886098861694
+    },
+    {
+      "epoch": 1.0431654676258992,
+      "force_loss": 0.0074464925564825535,
+      "step": 1160
+    },
+    {
+      "epoch": 1.0521582733812949,
+      "grad_norm": 1.2562706470489502,
+      "learning_rate": 3.896666666666667e-05,
+      "loss": 0.0461,
+      "step": 1170
+    },
+    {
+      "action_loss": 0.02570805512368679,
+      "epoch": 1.0521582733812949,
+      "step": 1170
+    },
+    {
+      "epoch": 1.0521582733812949,
+      "step": 1170,
+      "torque_loss": 0.161972776055336
+    },
+    {
+      "epoch": 1.0521582733812949,
+      "force_loss": 0.012070474214851856,
+      "step": 1170
+    },
+    {
+      "epoch": 1.0611510791366907,
+      "grad_norm": 1.1769518852233887,
+      "learning_rate": 3.9300000000000007e-05,
+      "loss": 0.0424,
+      "step": 1180
+    },
+    {
+      "action_loss": 0.029856273904442787,
+      "epoch": 1.0611510791366907,
+      "step": 1180
+    },
+    {
+      "epoch": 1.0611510791366907,
+      "step": 1180,
+      "torque_loss": 0.12464544922113419
+    },
+    {
+      "epoch": 1.0611510791366907,
+      "force_loss": 0.010856080800294876,
+      "step": 1180
+    },
+    {
+      "epoch": 1.0701438848920863,
+      "grad_norm": 0.895358681678772,
+      "learning_rate": 3.963333333333333e-05,
+      "loss": 0.0383,
+      "step": 1190
+    },
+    {
+      "action_loss": 0.0406019426882267,
+      "epoch": 1.0701438848920863,
+      "step": 1190
+    },
+    {
+      "epoch": 1.0701438848920863,
+      "step": 1190,
+      "torque_loss": 0.16710002720355988
+    },
+    {
+      "epoch": 1.0701438848920863,
+      "force_loss": 0.016543073579669,
+      "step": 1190
+    },
+    {
+      "epoch": 1.079136690647482,
+      "grad_norm": 1.3069195747375488,
+      "learning_rate": 3.996666666666667e-05,
+      "loss": 0.0414,
+      "step": 1200
+    },
+    {
+      "action_loss": 0.01567600481212139,
+      "epoch": 1.079136690647482,
+      "step": 1200
+    },
+    {
+      "epoch": 1.079136690647482,
+      "step": 1200,
+      "torque_loss": 0.11840472370386124
+    },
+    {
+      "epoch": 1.079136690647482,
+      "force_loss": 0.005812687333673239,
+      "step": 1200
+    },
+    {
+      "epoch": 1.0881294964028776,
+      "grad_norm": 0.9081102609634399,
+      "learning_rate": 4.0300000000000004e-05,
+      "loss": 0.0483,
+      "step": 1210
+    },
+    {
+      "action_loss": 0.025998719036579132,
+      "epoch": 1.0881294964028776,
+      "step": 1210
+    },
+    {
+      "epoch": 1.0881294964028776,
+      "step": 1210,
+      "torque_loss": 0.16842909157276154
+    },
+    {
+      "epoch": 1.0881294964028776,
+      "force_loss": 0.015264018438756466,
+      "step": 1210
+    },
+    {
+      "epoch": 1.0971223021582734,
+      "grad_norm": 1.0810637474060059,
+      "learning_rate": 4.0633333333333336e-05,
+      "loss": 0.0456,
+      "step": 1220
+    },
+    {
+      "action_loss": 0.028165755793452263,
+      "epoch": 1.0971223021582734,
+      "step": 1220
+    },
+    {
+      "epoch": 1.0971223021582734,
+      "step": 1220,
+      "torque_loss": 0.14498767256736755
+    },
+    {
+      "epoch": 1.0971223021582734,
+      "force_loss": 0.010447487235069275,
+      "step": 1220
+    },
+    {
+      "epoch": 1.106115107913669,
+      "grad_norm": 0.7632102966308594,
+      "learning_rate": 4.096666666666667e-05,
+      "loss": 0.037,
+      "step": 1230
+    },
+    {
+      "action_loss": 0.01392114907503128,
+      "epoch": 1.106115107913669,
+      "step": 1230
+    },
+    {
+      "epoch": 1.106115107913669,
+      "step": 1230,
+      "torque_loss": 0.14419786632061005
+    },
+    {
+      "epoch": 1.106115107913669,
+      "force_loss": 0.005224850028753281,
+      "step": 1230
+    },
+    {
+      "epoch": 1.1151079136690647,
+      "grad_norm": 2.4905953407287598,
+      "learning_rate": 4.13e-05,
+      "loss": 0.0433,
+      "step": 1240
+    },
+    {
+      "action_loss": 0.037794772535562515,
+      "epoch": 1.1151079136690647,
+      "step": 1240
+    },
+    {
+      "epoch": 1.1151079136690647,
+      "step": 1240,
+      "torque_loss": 0.14677703380584717
+    },
+    {
+      "epoch": 1.1151079136690647,
+      "force_loss": 0.027140960097312927,
+      "step": 1240
+    },
+    {
+      "epoch": 1.1241007194244603,
+      "grad_norm": 1.1650500297546387,
+      "learning_rate": 4.1633333333333333e-05,
+      "loss": 0.0404,
+      "step": 1250
+    },
+    {
+      "action_loss": 0.02054390124976635,
+      "epoch": 1.1241007194244603,
+      "step": 1250
+    },
+    {
+      "epoch": 1.1241007194244603,
+      "step": 1250,
+      "torque_loss": 0.12222862243652344
+    },
+    {
+      "epoch": 1.1241007194244603,
+      "force_loss": 0.006180791649967432,
+      "step": 1250
+    },
+    {
+      "epoch": 1.1330935251798562,
+      "grad_norm": 1.9320732355117798,
+      "learning_rate": 4.196666666666667e-05,
+      "loss": 0.0337,
+      "step": 1260
+    },
+    {
+      "action_loss": 0.022450512275099754,
+      "epoch": 1.1330935251798562,
+      "step": 1260
+    },
+    {
+      "epoch": 1.1330935251798562,
+      "step": 1260,
+      "torque_loss": 0.11523070186376572
+    },
+    {
+      "epoch": 1.1330935251798562,
+      "force_loss": 0.006162143778055906,
+      "step": 1260
+    },
+    {
+      "epoch": 1.1420863309352518,
+      "grad_norm": 2.0869734287261963,
+      "learning_rate": 4.23e-05,
+      "loss": 0.0327,
+      "step": 1270
+    },
+    {
+      "action_loss": 0.034509140998125076,
+      "epoch": 1.1420863309352518,
+      "step": 1270
+    },
+    {
+      "epoch": 1.1420863309352518,
+      "step": 1270,
+      "torque_loss": 0.12448292225599289
+    },
+    {
+      "epoch": 1.1420863309352518,
+      "force_loss": 0.03167206421494484,
+      "step": 1270
+    },
+    {
+      "epoch": 1.1510791366906474,
+      "grad_norm": 0.9672078490257263,
+      "learning_rate": 4.263333333333334e-05,
+      "loss": 0.0393,
+      "step": 1280
+    },
+    {
+      "action_loss": 0.0149543983861804,
+      "epoch": 1.1510791366906474,
+      "step": 1280
+    },
+    {
+      "epoch": 1.1510791366906474,
+      "step": 1280,
+      "torque_loss": 0.12495412677526474
+    },
+    {
+      "epoch": 1.1510791366906474,
+      "force_loss": 0.00579708069562912,
+      "step": 1280
+    },
+    {
+      "epoch": 1.1600719424460433,
+      "grad_norm": 2.0741238594055176,
+      "learning_rate": 4.296666666666666e-05,
+      "loss": 0.041,
+      "step": 1290
+    },
+    {
+      "action_loss": 0.029177134856581688,
+      "epoch": 1.1600719424460433,
+      "step": 1290
+    },
+    {
+      "epoch": 1.1600719424460433,
+      "step": 1290,
+      "torque_loss": 0.18909180164337158
+    },
+    {
+      "epoch": 1.1600719424460433,
+      "force_loss": 0.013418339192867279,
+      "step": 1290
+    },
+    {
+      "epoch": 1.169064748201439,
+      "grad_norm": 0.8136199712753296,
+      "learning_rate": 4.33e-05,
+      "loss": 0.0413,
+      "step": 1300
+    },
+    {
+      "action_loss": 0.011396770365536213,
+      "epoch": 1.169064748201439,
+      "step": 1300
+    },
+    {
+      "epoch": 1.169064748201439,
+      "step": 1300,
+      "torque_loss": 0.09374099224805832
+    },
+    {
+      "epoch": 1.169064748201439,
+      "force_loss": 0.006107666064053774,
+      "step": 1300
+    },
+    {
+      "epoch": 1.1780575539568345,
+      "grad_norm": 0.7694397568702698,
+      "learning_rate": 4.3633333333333335e-05,
+      "loss": 0.0362,
+      "step": 1310
+    },
+    {
+      "action_loss": 0.036509785801172256,
+      "epoch": 1.1780575539568345,
+      "step": 1310
+    },
+    {
+      "epoch": 1.1780575539568345,
+      "step": 1310,
+      "torque_loss": 0.14868922531604767
+    },
+    {
+      "epoch": 1.1780575539568345,
+      "force_loss": 0.009908725507557392,
+      "step": 1310
+    },
+    {
+      "epoch": 1.1870503597122302,
+      "grad_norm": 1.3490244150161743,
+      "learning_rate": 4.396666666666667e-05,
+      "loss": 0.0422,
+      "step": 1320
+    },
+    {
+      "action_loss": 0.03237549960613251,
+      "epoch": 1.1870503597122302,
+      "step": 1320
+    },
+    {
+      "epoch": 1.1870503597122302,
+      "step": 1320,
+      "torque_loss": 0.14881716668605804
+    },
+    {
+      "epoch": 1.1870503597122302,
+      "force_loss": 0.016344182193279266,
+      "step": 1320
+    },
+    {
+      "epoch": 1.1960431654676258,
+      "grad_norm": 0.9727373719215393,
+      "learning_rate": 4.43e-05,
+      "loss": 0.0457,
+      "step": 1330
+    },
+    {
+      "action_loss": 0.015569333918392658,
+      "epoch": 1.1960431654676258,
+      "step": 1330
+    },
+    {
+      "epoch": 1.1960431654676258,
+      "step": 1330,
+      "torque_loss": 0.10138320922851562
+    },
+    {
+      "epoch": 1.1960431654676258,
+      "force_loss": 0.00933428667485714,
+      "step": 1330
+    },
+    {
+      "epoch": 1.2050359712230216,
+      "grad_norm": 1.123683214187622,
+      "learning_rate": 4.463333333333334e-05,
+      "loss": 0.0518,
+      "step": 1340
+    },
+    {
+      "action_loss": 0.03274621441960335,
+      "epoch": 1.2050359712230216,
+      "step": 1340
+    },
+    {
+      "epoch": 1.2050359712230216,
+      "step": 1340,
+      "torque_loss": 0.12057191133499146
+    },
+    {
+      "epoch": 1.2050359712230216,
+      "force_loss": 0.011477436870336533,
+      "step": 1340
+    },
+    {
+      "epoch": 1.2140287769784173,
+      "grad_norm": 0.5538128018379211,
+      "learning_rate": 4.496666666666667e-05,
+      "loss": 0.0405,
+      "step": 1350
+    },
+    {
+      "action_loss": 0.021062130108475685,
+      "epoch": 1.2140287769784173,
+      "step": 1350
+    },
+    {
+      "epoch": 1.2140287769784173,
+      "step": 1350,
+      "torque_loss": 0.146018847823143
+    },
+    {
+      "epoch": 1.2140287769784173,
+      "force_loss": 0.010402261279523373,
+      "step": 1350
+    },
+    {
+      "epoch": 1.223021582733813,
+      "grad_norm": 1.2760074138641357,
+      "learning_rate": 4.53e-05,
+      "loss": 0.0528,
+      "step": 1360
+    },
+    {
+      "action_loss": 0.032370369881391525,
+      "epoch": 1.223021582733813,
+      "step": 1360
+    },
+    {
+      "epoch": 1.223021582733813,
+      "step": 1360,
+      "torque_loss": 0.10648646950721741
+    },
+    {
+      "epoch": 1.223021582733813,
+      "force_loss": 0.01716136746108532,
+      "step": 1360
+    },
+    {
+      "epoch": 1.2320143884892087,
+      "grad_norm": 1.5575224161148071,
+      "learning_rate": 4.5633333333333336e-05,
+      "loss": 0.0412,
+      "step": 1370
+    },
+    {
+      "action_loss": 0.01791442185640335,
+      "epoch": 1.2320143884892087,
+      "step": 1370
+    },
+    {
+      "epoch": 1.2320143884892087,
+      "step": 1370,
+      "torque_loss": 0.11422237753868103
+    },
+    {
+      "epoch": 1.2320143884892087,
+      "force_loss": 0.007272792514413595,
+      "step": 1370
+    },
+    {
+      "epoch": 1.2410071942446044,
+      "grad_norm": 0.8581523895263672,
+      "learning_rate": 4.596666666666667e-05,
+      "loss": 0.0377,
+      "step": 1380
+    },
+    {
+      "action_loss": 0.018249275162816048,
+      "epoch": 1.2410071942446044,
+      "step": 1380
+    },
+    {
+      "epoch": 1.2410071942446044,
+      "step": 1380,
+      "torque_loss": 0.13104642927646637
+    },
+    {
+      "epoch": 1.2410071942446044,
+      "force_loss": 0.006711956113576889,
+      "step": 1380
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 1.0329571962356567,
+      "learning_rate": 4.630000000000001e-05,
+      "loss": 0.0397,
+      "step": 1390
+    },
+    {
+      "action_loss": 0.016140621155500412,
+      "epoch": 1.25,
+      "step": 1390
+    },
+    {
+      "epoch": 1.25,
+      "step": 1390,
+      "torque_loss": 0.11730929464101791
+    },
+    {
+      "epoch": 1.25,
+      "force_loss": 0.009129363112151623,
+      "step": 1390
+    },
+    {
+      "epoch": 1.2589928057553956,
+      "grad_norm": 1.5746816396713257,
+      "learning_rate": 4.663333333333333e-05,
+      "loss": 0.038,
+      "step": 1400
+    },
+    {
+      "action_loss": 0.02167298085987568,
+      "epoch": 1.2589928057553956,
+      "step": 1400
+    },
+    {
+      "epoch": 1.2589928057553956,
+      "step": 1400,
+      "torque_loss": 0.10270285606384277
+    },
+    {
+      "epoch": 1.2589928057553956,
+      "force_loss": 0.014194023795425892,
+      "step": 1400
+    },
+    {
+      "epoch": 1.2679856115107913,
+      "grad_norm": 1.1146483421325684,
+      "learning_rate": 4.696666666666667e-05,
+      "loss": 0.0567,
+      "step": 1410
+    },
+    {
+      "action_loss": 0.023223230615258217,
+      "epoch": 1.2679856115107913,
+      "step": 1410
+    },
+    {
+      "epoch": 1.2679856115107913,
+      "step": 1410,
+      "torque_loss": 0.19135254621505737
+    },
+    {
+      "epoch": 1.2679856115107913,
+      "force_loss": 0.011045348830521107,
+      "step": 1410
+    },
+    {
+      "epoch": 1.276978417266187,
+      "grad_norm": 1.0299912691116333,
+      "learning_rate": 4.73e-05,
+      "loss": 0.0457,
+      "step": 1420
+    },
+    {
+      "action_loss": 0.049068063497543335,
+      "epoch": 1.276978417266187,
+      "step": 1420
+    },
+    {
+      "epoch": 1.276978417266187,
+      "step": 1420,
+      "torque_loss": 0.14573934674263
+    },
+    {
+      "epoch": 1.276978417266187,
+      "force_loss": 0.029202574864029884,
+      "step": 1420
+    },
+    {
+      "epoch": 1.2859712230215827,
+      "grad_norm": 1.1865311861038208,
+      "learning_rate": 4.763333333333334e-05,
+      "loss": 0.0432,
+      "step": 1430
+    },
+    {
+      "action_loss": 0.030678488314151764,
+      "epoch": 1.2859712230215827,
+      "step": 1430
+    },
+    {
+      "epoch": 1.2859712230215827,
+      "step": 1430,
+      "torque_loss": 0.1374351531267166
+    },
+    {
+      "epoch": 1.2859712230215827,
+      "force_loss": 0.022890103980898857,
+      "step": 1430
+    },
+    {
+      "epoch": 1.2949640287769784,
+      "grad_norm": 1.7581628561019897,
+      "learning_rate": 4.796666666666667e-05,
+      "loss": 0.0341,
+      "step": 1440
+    },
+    {
+      "action_loss": 0.027712544426321983,
+      "epoch": 1.2949640287769784,
+      "step": 1440
+    },
+    {
+      "epoch": 1.2949640287769784,
+      "step": 1440,
+      "torque_loss": 0.13361425697803497
+    },
+    {
+      "epoch": 1.2949640287769784,
+      "force_loss": 0.014639712870121002,
+      "step": 1440
+    },
+    {
+      "epoch": 1.3039568345323742,
+      "grad_norm": 0.7644403576850891,
+      "learning_rate": 4.83e-05,
+      "loss": 0.0443,
+      "step": 1450
+    },
+    {
+      "action_loss": 0.02581537514925003,
+      "epoch": 1.3039568345323742,
+      "step": 1450
+    },
+    {
+      "epoch": 1.3039568345323742,
+      "step": 1450,
+      "torque_loss": 0.21452778577804565
+    },
+    {
+      "epoch": 1.3039568345323742,
+      "force_loss": 0.014838947914540768,
+      "step": 1450
+    },
+    {
+      "epoch": 1.3129496402877698,
+      "grad_norm": 0.8814418315887451,
+      "learning_rate": 4.8633333333333334e-05,
+      "loss": 0.0404,
+      "step": 1460
+    },
+    {
+      "action_loss": 0.01225553173571825,
+      "epoch": 1.3129496402877698,
+      "step": 1460
+    },
+    {
+      "epoch": 1.3129496402877698,
+      "step": 1460,
+      "torque_loss": 0.1041104793548584
+    },
+    {
+      "epoch": 1.3129496402877698,
+      "force_loss": 0.004430586937814951,
+      "step": 1460
+    },
+    {
+      "epoch": 1.3219424460431655,
+      "grad_norm": 1.0401335954666138,
+      "learning_rate": 4.8966666666666667e-05,
+      "loss": 0.0331,
+      "step": 1470
+    },
+    {
+      "action_loss": 0.014025881886482239,
+      "epoch": 1.3219424460431655,
+      "step": 1470
+    },
+    {
+      "epoch": 1.3219424460431655,
+      "step": 1470,
+      "torque_loss": 0.12877392768859863
+    },
+    {
+      "epoch": 1.3219424460431655,
+      "force_loss": 0.00756074720993638,
+      "step": 1470
+    },
+    {
+      "epoch": 1.330935251798561,
+      "grad_norm": 1.5328564643859863,
+      "learning_rate": 4.93e-05,
+      "loss": 0.0394,
+      "step": 1480
+    },
+    {
+      "action_loss": 0.016159093007445335,
+      "epoch": 1.330935251798561,
+      "step": 1480
+    },
+    {
+      "epoch": 1.330935251798561,
+      "step": 1480,
+      "torque_loss": 0.1405736654996872
+    },
+    {
+      "epoch": 1.330935251798561,
+      "force_loss": 0.004206780809909105,
+      "step": 1480
+    },
+    {
+      "epoch": 1.3399280575539567,
+      "grad_norm": 0.8034462332725525,
+      "learning_rate": 4.963333333333334e-05,
+      "loss": 0.0375,
+      "step": 1490
+    },
+    {
+      "action_loss": 0.03001769818365574,
+      "epoch": 1.3399280575539567,
+      "step": 1490
+    },
+    {
+      "epoch": 1.3399280575539567,
+      "step": 1490,
+      "torque_loss": 0.10380522161722183
+    },
+    {
+      "epoch": 1.3399280575539567,
+      "force_loss": 0.015295598655939102,
+      "step": 1490
+    },
+    {
+      "epoch": 1.3489208633093526,
+      "grad_norm": 0.6252776980400085,
+      "learning_rate": 4.996666666666667e-05,
+      "loss": 0.0436,
+      "step": 1500
+    },
+    {
+      "action_loss": 0.017488082870841026,
+      "epoch": 1.3489208633093526,
+      "step": 1500
+    },
+    {
+      "epoch": 1.3489208633093526,
+      "step": 1500,
+      "torque_loss": 0.13915669918060303
+    },
+    {
+      "epoch": 1.3489208633093526,
+      "force_loss": 0.008006691001355648,
+      "step": 1500
+    },
+    {
+      "epoch": 1.3579136690647482,
+      "grad_norm": 0.5311513543128967,
+      "learning_rate": 5.03e-05,
+      "loss": 0.0308,
+      "step": 1510
+    },
+    {
+      "action_loss": 0.0435921736061573,
+      "epoch": 1.3579136690647482,
+      "step": 1510
+    },
+    {
+      "epoch": 1.3579136690647482,
+      "step": 1510,
+      "torque_loss": 0.16664637625217438
+    },
+    {
+      "epoch": 1.3579136690647482,
+      "force_loss": 0.019571689888834953,
+      "step": 1510
+    },
+    {
+      "epoch": 1.3669064748201438,
+      "grad_norm": 1.3396062850952148,
+      "learning_rate": 5.0633333333333335e-05,
+      "loss": 0.038,
+      "step": 1520
+    },
+    {
+      "action_loss": 0.011165883392095566,
+      "epoch": 1.3669064748201438,
+      "step": 1520
+    },
+    {
+      "epoch": 1.3669064748201438,
+      "step": 1520,
+      "torque_loss": 0.12072101980447769
+    },
+    {
+      "epoch": 1.3669064748201438,
+      "force_loss": 0.0072516812942922115,
+      "step": 1520
+    },
+    {
+      "epoch": 1.3758992805755397,
+      "grad_norm": 1.793840765953064,
+      "learning_rate": 5.0966666666666674e-05,
+      "loss": 0.0385,
+      "step": 1530
+    },
+    {
+      "action_loss": 0.020988086238503456,
+      "epoch": 1.3758992805755397,
+      "step": 1530
+    },
+    {
+      "epoch": 1.3758992805755397,
+      "step": 1530,
+      "torque_loss": 0.14598026871681213
+    },
+    {
+      "epoch": 1.3758992805755397,
+      "force_loss": 0.008936233818531036,
+      "step": 1530
+    },
+    {
+      "epoch": 1.3848920863309353,
+      "grad_norm": 1.572195291519165,
+      "learning_rate": 5.130000000000001e-05,
+      "loss": 0.0363,
+      "step": 1540
+    },
+    {
+      "action_loss": 0.01592131517827511,
+      "epoch": 1.3848920863309353,
+      "step": 1540
+    },
+    {
+      "epoch": 1.3848920863309353,
+      "step": 1540,
+      "torque_loss": 0.1095358356833458
+    },
+    {
+      "epoch": 1.3848920863309353,
+      "force_loss": 0.006344737019389868,
+      "step": 1540
+    },
+    {
+      "epoch": 1.393884892086331,
+      "grad_norm": 1.034877061843872,
+      "learning_rate": 5.163333333333333e-05,
+      "loss": 0.0449,
+      "step": 1550
+    },
+    {
+      "action_loss": 0.016093989834189415,
+      "epoch": 1.393884892086331,
+      "step": 1550
+    },
+    {
+      "epoch": 1.393884892086331,
+      "step": 1550,
+      "torque_loss": 0.1271398961544037
+    },
+    {
+      "epoch": 1.393884892086331,
+      "force_loss": 0.008789844810962677,
+      "step": 1550
+    },
+    {
+      "epoch": 1.4028776978417266,
+      "grad_norm": 0.6551756262779236,
+      "learning_rate": 5.196666666666667e-05,
+      "loss": 0.0325,
+      "step": 1560
+    },
+    {
+      "action_loss": 0.013529728166759014,
+      "epoch": 1.4028776978417266,
+      "step": 1560
+    },
+    {
+      "epoch": 1.4028776978417266,
+      "step": 1560,
+      "torque_loss": 0.12017623335123062
+    },
+    {
+      "epoch": 1.4028776978417266,
+      "force_loss": 0.008128629997372627,
+      "step": 1560
+    },
+    {
+      "epoch": 1.4118705035971222,
+      "grad_norm": 0.8998020887374878,
+      "learning_rate": 5.2300000000000004e-05,
+      "loss": 0.0348,
+      "step": 1570
+    },
+    {
+      "action_loss": 0.022844159975647926,
+      "epoch": 1.4118705035971222,
+      "step": 1570
+    },
+    {
+      "epoch": 1.4118705035971222,
+      "step": 1570,
+      "torque_loss": 0.1208563819527626
+    },
+    {
+      "epoch": 1.4118705035971222,
+      "force_loss": 0.005417071282863617,
+      "step": 1570
+    },
+    {
+      "epoch": 1.420863309352518,
+      "grad_norm": 1.5591024160385132,
+      "learning_rate": 5.2633333333333336e-05,
+      "loss": 0.0384,
+      "step": 1580
+    },
+    {
+      "action_loss": 0.01836899295449257,
+      "epoch": 1.420863309352518,
+      "step": 1580
+    },
+    {
+      "epoch": 1.420863309352518,
+      "step": 1580,
+      "torque_loss": 0.13783617317676544
+    },
+    {
+      "epoch": 1.420863309352518,
+      "force_loss": 0.007158016320317984,
+      "step": 1580
+    },
+    {
+      "epoch": 1.4298561151079137,
+      "grad_norm": 1.5710378885269165,
+      "learning_rate": 5.296666666666666e-05,
+      "loss": 0.0417,
+      "step": 1590
+    },
+    {
+      "action_loss": 0.020005570724606514,
+      "epoch": 1.4298561151079137,
+      "step": 1590
+    },
+    {
+      "epoch": 1.4298561151079137,
+      "step": 1590,
+      "torque_loss": 0.155879944562912
+    },
+    {
+      "epoch": 1.4298561151079137,
+      "force_loss": 0.010471027344465256,
+      "step": 1590
+    },
+    {
+      "epoch": 1.4388489208633093,
+      "grad_norm": 0.8876373171806335,
+      "learning_rate": 5.330000000000001e-05,
+      "loss": 0.0403,
+      "step": 1600
+    },
+    {
+      "action_loss": 0.01961771585047245,
+      "epoch": 1.4388489208633093,
+      "step": 1600
+    },
+    {
+      "epoch": 1.4388489208633093,
+      "step": 1600,
+      "torque_loss": 0.1969851702451706
+    },
+    {
+      "epoch": 1.4388489208633093,
+      "force_loss": 0.010652844794094563,
+      "step": 1600
+    },
+    {
+      "epoch": 1.4478417266187051,
+      "grad_norm": 1.0628294944763184,
+      "learning_rate": 5.3633333333333334e-05,
+      "loss": 0.0416,
+      "step": 1610
+    },
+    {
+      "action_loss": 0.038969844579696655,
+      "epoch": 1.4478417266187051,
+      "step": 1610
+    },
+    {
+      "epoch": 1.4478417266187051,
+      "step": 1610,
+      "torque_loss": 0.13635770976543427
+    },
+    {
+      "epoch": 1.4478417266187051,
+      "force_loss": 0.020792650058865547,
+      "step": 1610
+    },
+    {
+      "epoch": 1.4568345323741008,
+      "grad_norm": 1.399761438369751,
+      "learning_rate": 5.3966666666666666e-05,
+      "loss": 0.0386,
+      "step": 1620
+    },
+    {
+      "action_loss": 0.01814407669007778,
+      "epoch": 1.4568345323741008,
+      "step": 1620
+    },
+    {
+      "epoch": 1.4568345323741008,
+      "step": 1620,
+      "torque_loss": 0.12877829372882843
+    },
+    {
+      "epoch": 1.4568345323741008,
+      "force_loss": 0.007057117763906717,
+      "step": 1620
+    },
+    {
+      "epoch": 1.4658273381294964,
+      "grad_norm": 1.0869120359420776,
+      "learning_rate": 5.4300000000000005e-05,
+      "loss": 0.0456,
+      "step": 1630
+    },
+    {
+      "action_loss": 0.017783258110284805,
+      "epoch": 1.4658273381294964,
+      "step": 1630
+    },
+    {
+      "epoch": 1.4658273381294964,
+      "step": 1630,
+      "torque_loss": 0.11481910943984985
+    },
+    {
+      "epoch": 1.4658273381294964,
+      "force_loss": 0.008570303209125996,
+      "step": 1630
+    },
+    {
+      "epoch": 1.474820143884892,
+      "grad_norm": 2.3219895362854004,
+      "learning_rate": 5.463333333333334e-05,
+      "loss": 0.0406,
+      "step": 1640
+    },
+    {
+      "action_loss": 0.025417575612664223,
+      "epoch": 1.474820143884892,
+      "step": 1640
+    },
+    {
+      "epoch": 1.474820143884892,
+      "step": 1640,
+      "torque_loss": 0.14974556863307953
+    },
+    {
+      "epoch": 1.474820143884892,
+      "force_loss": 0.021351585164666176,
+      "step": 1640
+    },
+    {
+      "epoch": 1.4838129496402876,
+      "grad_norm": 0.5720558166503906,
+      "learning_rate": 5.496666666666666e-05,
+      "loss": 0.0356,
+      "step": 1650
+    },
+    {
+      "action_loss": 0.01504466962069273,
+      "epoch": 1.4838129496402876,
+      "step": 1650
+    },
+    {
+      "epoch": 1.4838129496402876,
+      "step": 1650,
+      "torque_loss": 0.11802510172128677
+    },
+    {
+      "epoch": 1.4838129496402876,
+      "force_loss": 0.006154266651719809,
+      "step": 1650
+    },
+    {
+      "epoch": 1.4928057553956835,
+      "grad_norm": 0.8320698142051697,
+      "learning_rate": 5.530000000000001e-05,
+      "loss": 0.0472,
+      "step": 1660
+    },
+    {
+      "action_loss": 0.028241241350769997,
+      "epoch": 1.4928057553956835,
+      "step": 1660
+    },
+    {
+      "epoch": 1.4928057553956835,
+      "step": 1660,
+      "torque_loss": 0.12319370359182358
+    },
+    {
+      "epoch": 1.4928057553956835,
+      "force_loss": 0.010020174086093903,
+      "step": 1660
+    },
+    {
+      "epoch": 1.5017985611510791,
+      "grad_norm": 2.339298963546753,
+      "learning_rate": 5.5633333333333335e-05,
+      "loss": 0.0451,
+      "step": 1670
+    },
+    {
+      "action_loss": 0.054126013070344925,
+      "epoch": 1.5017985611510791,
+      "step": 1670
+    },
+    {
+      "epoch": 1.5017985611510791,
+      "step": 1670,
+      "torque_loss": 0.16401435434818268
+    },
+    {
+      "epoch": 1.5017985611510791,
+      "force_loss": 0.015665413811802864,
+      "step": 1670
+    },
+    {
+      "epoch": 1.5107913669064748,
+      "grad_norm": 1.5541661977767944,
+      "learning_rate": 5.596666666666667e-05,
+      "loss": 0.0484,
+      "step": 1680
+    },
+    {
+      "action_loss": 0.02101924829185009,
+      "epoch": 1.5107913669064748,
+      "step": 1680
+    },
+    {
+      "epoch": 1.5107913669064748,
+      "step": 1680,
+      "torque_loss": 0.12515272200107574
+    },
+    {
+      "epoch": 1.5107913669064748,
+      "force_loss": 0.007671245839446783,
+      "step": 1680
+    },
+    {
+      "epoch": 1.5197841726618706,
+      "grad_norm": 1.5375518798828125,
+      "learning_rate": 5.63e-05,
+      "loss": 0.0632,
+      "step": 1690
+    },
+    {
+      "action_loss": 0.020116286352276802,
+      "epoch": 1.5197841726618706,
+      "step": 1690
+    },
+    {
+      "epoch": 1.5197841726618706,
+      "step": 1690,
+      "torque_loss": 0.1242685541510582
+    },
+    {
+      "epoch": 1.5197841726618706,
+      "force_loss": 0.012467063032090664,
+      "step": 1690
+    },
+    {
+      "epoch": 1.5287769784172662,
+      "grad_norm": 1.3778345584869385,
+      "learning_rate": 5.663333333333334e-05,
+      "loss": 0.0457,
+      "step": 1700
+    },
+    {
+      "action_loss": 0.030835585668683052,
+      "epoch": 1.5287769784172662,
+      "step": 1700
+    },
+    {
+      "epoch": 1.5287769784172662,
+      "step": 1700,
+      "torque_loss": 0.12443309277296066
+    },
+    {
+      "epoch": 1.5287769784172662,
+      "force_loss": 0.013805332593619823,
+      "step": 1700
+    },
+    {
+      "epoch": 1.5377697841726619,
+      "grad_norm": 1.384293556213379,
+      "learning_rate": 5.696666666666667e-05,
+      "loss": 0.0411,
+      "step": 1710
+    },
+    {
+      "action_loss": 0.02501256950199604,
+      "epoch": 1.5377697841726619,
+      "step": 1710
+    },
+    {
+      "epoch": 1.5377697841726619,
+      "step": 1710,
+      "torque_loss": 0.13218729197978973
+    },
+    {
+      "epoch": 1.5377697841726619,
+      "force_loss": 0.008499592542648315,
+      "step": 1710
+    },
+    {
+      "epoch": 1.5467625899280577,
+      "grad_norm": 1.01209557056427,
+      "learning_rate": 5.73e-05,
+      "loss": 0.0453,
+      "step": 1720
+    },
+    {
+      "action_loss": 0.027362314984202385,
+      "epoch": 1.5467625899280577,
+      "step": 1720
+    },
+    {
+      "epoch": 1.5467625899280577,
+      "step": 1720,
+      "torque_loss": 0.1004047766327858
+    },
+    {
+      "epoch": 1.5467625899280577,
+      "force_loss": 0.01234254240989685,
+      "step": 1720
+    },
+    {
+      "epoch": 1.5557553956834531,
+      "grad_norm": 1.0071674585342407,
+      "learning_rate": 5.7633333333333336e-05,
+      "loss": 0.0384,
+      "step": 1730
+    },
+    {
+      "action_loss": 0.05230461433529854,
+      "epoch": 1.5557553956834531,
+      "step": 1730
+    },
+    {
+      "epoch": 1.5557553956834531,
+      "step": 1730,
+      "torque_loss": 0.22703103721141815
+    },
+    {
+      "epoch": 1.5557553956834531,
+      "force_loss": 0.026303531602025032,
+      "step": 1730
+    },
+    {
+      "epoch": 1.564748201438849,
+      "grad_norm": 1.1060887575149536,
+      "learning_rate": 5.796666666666667e-05,
+      "loss": 0.0437,
+      "step": 1740
+    },
+    {
+      "action_loss": 0.027820652350783348,
+      "epoch": 1.564748201438849,
+      "step": 1740
+    },
+    {
+      "epoch": 1.564748201438849,
+      "step": 1740,
+      "torque_loss": 0.11444851011037827
+    },
+    {
+      "epoch": 1.564748201438849,
+      "force_loss": 0.009924407117068768,
+      "step": 1740
+    },
+    {
+      "epoch": 1.5737410071942446,
+      "grad_norm": 1.1261448860168457,
+      "learning_rate": 5.83e-05,
+      "loss": 0.0391,
+      "step": 1750
+    },
+    {
+      "action_loss": 0.023280911147594452,
+      "epoch": 1.5737410071942446,
+      "step": 1750
+    },
+    {
+      "epoch": 1.5737410071942446,
+      "step": 1750,
+      "torque_loss": 0.16219337284564972
+    },
+    {
+      "epoch": 1.5737410071942446,
+      "force_loss": 0.011544729582965374,
+      "step": 1750
+    },
+    {
+      "epoch": 1.5827338129496402,
+      "grad_norm": 2.5638270378112793,
+      "learning_rate": 5.863333333333334e-05,
+      "loss": 0.0399,
+      "step": 1760
+    },
+    {
+      "action_loss": 0.037021443247795105,
+      "epoch": 1.5827338129496402,
+      "step": 1760
+    },
+    {
+      "epoch": 1.5827338129496402,
+      "step": 1760,
+      "torque_loss": 0.12499305605888367
+    },
+    {
+      "epoch": 1.5827338129496402,
+      "force_loss": 0.01578935794532299,
+      "step": 1760
+    },
+    {
+      "epoch": 1.591726618705036,
+      "grad_norm": 0.9227173924446106,
+      "learning_rate": 5.896666666666667e-05,
+      "loss": 0.0405,
+      "step": 1770
+    },
+    {
+      "action_loss": 0.010805748403072357,
+      "epoch": 1.591726618705036,
+      "step": 1770
+    },
+    {
+      "epoch": 1.591726618705036,
+      "step": 1770,
+      "torque_loss": 0.1569158285856247
+    },
+    {
+      "epoch": 1.591726618705036,
+      "force_loss": 0.007633899804204702,
+      "step": 1770
+    },
+    {
+      "epoch": 1.6007194244604317,
+      "grad_norm": 0.7877851724624634,
+      "learning_rate": 5.93e-05,
+      "loss": 0.0311,
+      "step": 1780
+    },
+    {
+      "action_loss": 0.019057104364037514,
+      "epoch": 1.6007194244604317,
+      "step": 1780
+    },
+    {
+      "epoch": 1.6007194244604317,
+      "step": 1780,
+      "torque_loss": 0.13095499575138092
+    },
+    {
+      "epoch": 1.6007194244604317,
+      "force_loss": 0.007099824491888285,
+      "step": 1780
+    },
+    {
+      "epoch": 1.6097122302158273,
+      "grad_norm": 1.6048319339752197,
+      "learning_rate": 5.9633333333333344e-05,
+      "loss": 0.0518,
+      "step": 1790
+    },
+    {
+      "action_loss": 0.022755520418286324,
+      "epoch": 1.6097122302158273,
+      "step": 1790
+    },
+    {
+      "epoch": 1.6097122302158273,
+      "step": 1790,
+      "torque_loss": 0.12178144603967667
+    },
+    {
+      "epoch": 1.6097122302158273,
+      "force_loss": 0.008644222281873226,
+      "step": 1790
+    },
+    {
+      "epoch": 1.6187050359712232,
+      "grad_norm": 2.2193334102630615,
+      "learning_rate": 5.996666666666667e-05,
+      "loss": 0.0496,
+      "step": 1800
+    },
+    {
+      "action_loss": 0.01914903335273266,
+      "epoch": 1.6187050359712232,
+      "step": 1800
+    },
+    {
+      "epoch": 1.6187050359712232,
+      "step": 1800,
+      "torque_loss": 0.13406257331371307
+    },
+    {
+      "epoch": 1.6187050359712232,
+      "force_loss": 0.009918020106852055,
+      "step": 1800
+    },
+    {
+      "epoch": 1.6276978417266186,
+      "grad_norm": 1.2487359046936035,
+      "learning_rate": 6.03e-05,
+      "loss": 0.0385,
+      "step": 1810
+    },
+    {
+      "action_loss": 0.020536141470074654,
+      "epoch": 1.6276978417266186,
+      "step": 1810
+    },
+    {
+      "epoch": 1.6276978417266186,
+      "step": 1810,
+      "torque_loss": 0.10585105419158936
+    },
+    {
+      "epoch": 1.6276978417266186,
+      "force_loss": 0.007804557215422392,
+      "step": 1810
+    },
+    {
+      "epoch": 1.6366906474820144,
+      "grad_norm": 0.9121411442756653,
+      "learning_rate": 6.063333333333333e-05,
+      "loss": 0.0321,
+      "step": 1820
+    },
+    {
+      "action_loss": 0.012331143952906132,
+      "epoch": 1.6366906474820144,
+      "step": 1820
+    },
+    {
+      "epoch": 1.6366906474820144,
+      "step": 1820,
+      "torque_loss": 0.16718094050884247
+    },
+    {
+      "epoch": 1.6366906474820144,
+      "force_loss": 0.00467007327824831,
+      "step": 1820
+    },
+    {
+      "epoch": 1.64568345323741,
+      "grad_norm": 1.7599858045578003,
+      "learning_rate": 6.0966666666666674e-05,
+      "loss": 0.0384,
+      "step": 1830
+    },
+    {
+      "action_loss": 0.028700992465019226,
+      "epoch": 1.64568345323741,
+      "step": 1830
+    },
+    {
+      "epoch": 1.64568345323741,
+      "step": 1830,
+      "torque_loss": 0.13857920467853546
+    },
+    {
+      "epoch": 1.64568345323741,
+      "force_loss": 0.015253528952598572,
+      "step": 1830
+    },
+    {
+      "epoch": 1.6546762589928057,
+      "grad_norm": 1.1731390953063965,
+      "learning_rate": 6.13e-05,
+      "loss": 0.0441,
+      "step": 1840
+    },
+    {
+      "action_loss": 0.02267463319003582,
+      "epoch": 1.6546762589928057,
+      "step": 1840
+    },
+    {
+      "epoch": 1.6546762589928057,
+      "step": 1840,
+      "torque_loss": 0.12330130487680435
+    },
+    {
+      "epoch": 1.6546762589928057,
+      "force_loss": 0.012602976523339748,
+      "step": 1840
+    },
+    {
+      "epoch": 1.6636690647482015,
+      "grad_norm": 0.6124871373176575,
+      "learning_rate": 6.163333333333333e-05,
+      "loss": 0.0366,
+      "step": 1850
+    },
+    {
+      "action_loss": 0.019395431503653526,
+      "epoch": 1.6636690647482015,
+      "step": 1850
+    },
+    {
+      "epoch": 1.6636690647482015,
+      "step": 1850,
+      "torque_loss": 0.12141523510217667
+    },
+    {
+      "epoch": 1.6636690647482015,
+      "force_loss": 0.007657719310373068,
+      "step": 1850
+    },
+    {
+      "epoch": 1.6726618705035972,
+      "grad_norm": 1.1004042625427246,
+      "learning_rate": 6.196666666666668e-05,
+      "loss": 0.0398,
+      "step": 1860
+    },
+    {
+      "action_loss": 0.021091600880026817,
+      "epoch": 1.6726618705035972,
+      "step": 1860
+    },
+    {
+      "epoch": 1.6726618705035972,
+      "step": 1860,
+      "torque_loss": 0.12089512497186661
+    },
+    {
+      "epoch": 1.6726618705035972,
+      "force_loss": 0.016227731481194496,
+      "step": 1860
+    },
+    {
+      "epoch": 1.6816546762589928,
+      "grad_norm": 1.5495952367782593,
+      "learning_rate": 6.23e-05,
+      "loss": 0.0331,
+      "step": 1870
+    },
+    {
+      "action_loss": 0.022025009617209435,
+      "epoch": 1.6816546762589928,
+      "step": 1870
+    },
+    {
+      "epoch": 1.6816546762589928,
+      "step": 1870,
+      "torque_loss": 0.13842914998531342
+    },
+    {
+      "epoch": 1.6816546762589928,
+      "force_loss": 0.012225431390106678,
+      "step": 1870
+    },
+    {
+      "epoch": 1.6906474820143886,
+      "grad_norm": 0.9583244323730469,
+      "learning_rate": 6.263333333333333e-05,
+      "loss": 0.0362,
+      "step": 1880
+    },
+    {
+      "action_loss": 0.011240922845900059,
+      "epoch": 1.6906474820143886,
+      "step": 1880
+    },
+    {
+      "epoch": 1.6906474820143886,
+      "step": 1880,
+      "torque_loss": 0.14442689716815948
+    },
+    {
+      "epoch": 1.6906474820143886,
+      "force_loss": 0.005097482819110155,
+      "step": 1880
+    },
+    {
+      "epoch": 1.699640287769784,
+      "grad_norm": 0.9254547953605652,
+      "learning_rate": 6.296666666666667e-05,
+      "loss": 0.0426,
+      "step": 1890
+    },
+    {
+      "action_loss": 0.03405832126736641,
+      "epoch": 1.699640287769784,
+      "step": 1890
+    },
+    {
+      "epoch": 1.699640287769784,
+      "step": 1890,
+      "torque_loss": 0.151302307844162
+    },
+    {
+      "epoch": 1.699640287769784,
+      "force_loss": 0.017275933176279068,
+      "step": 1890
+    },
+    {
+      "epoch": 1.70863309352518,
+      "grad_norm": 1.3274579048156738,
+      "learning_rate": 6.330000000000001e-05,
+      "loss": 0.0473,
+      "step": 1900
+    },
+    {
+      "action_loss": 0.026476452127099037,
+      "epoch": 1.70863309352518,
+      "step": 1900
+    },
+    {
+      "epoch": 1.70863309352518,
+      "step": 1900,
+      "torque_loss": 0.14379094541072845
+    },
+    {
+      "epoch": 1.70863309352518,
+      "force_loss": 0.00846109539270401,
+      "step": 1900
+    },
+    {
+      "epoch": 1.7176258992805755,
+      "grad_norm": 2.355393648147583,
+      "learning_rate": 6.363333333333334e-05,
+      "loss": 0.0359,
+      "step": 1910
+    },
+    {
+      "action_loss": 0.016802387312054634,
+      "epoch": 1.7176258992805755,
+      "step": 1910
+    },
+    {
+      "epoch": 1.7176258992805755,
+      "step": 1910,
+      "torque_loss": 0.1140458956360817
+    },
+    {
+      "epoch": 1.7176258992805755,
+      "force_loss": 0.007565314415842295,
+      "step": 1910
+    },
+    {
+      "epoch": 1.7266187050359711,
+      "grad_norm": 1.7675760984420776,
+      "learning_rate": 6.396666666666667e-05,
+      "loss": 0.0324,
+      "step": 1920
+    },
+    {
+      "action_loss": 0.06040431559085846,
+      "epoch": 1.7266187050359711,
+      "step": 1920
+    },
+    {
+      "epoch": 1.7266187050359711,
+      "step": 1920,
+      "torque_loss": 0.1805545836687088
+    },
+    {
+      "epoch": 1.7266187050359711,
+      "force_loss": 0.03229805454611778,
+      "step": 1920
+    },
+    {
+      "epoch": 1.735611510791367,
+      "grad_norm": 1.8652795553207397,
+      "learning_rate": 6.43e-05,
+      "loss": 0.0441,
+      "step": 1930
+    },
+    {
+      "action_loss": 0.030441507697105408,
+      "epoch": 1.735611510791367,
+      "step": 1930
+    },
+    {
+      "epoch": 1.735611510791367,
+      "step": 1930,
+      "torque_loss": 0.13247232139110565
+    },
+    {
+      "epoch": 1.735611510791367,
+      "force_loss": 0.00965114776045084,
+      "step": 1930
+    },
+    {
+      "epoch": 1.7446043165467626,
+      "grad_norm": 1.0935837030410767,
+      "learning_rate": 6.463333333333334e-05,
+      "loss": 0.0319,
+      "step": 1940
+    },
+    {
+      "action_loss": 0.030394727364182472,
+      "epoch": 1.7446043165467626,
+      "step": 1940
+    },
+    {
+      "epoch": 1.7446043165467626,
+      "step": 1940,
+      "torque_loss": 0.14992812275886536
+    },
+    {
+      "epoch": 1.7446043165467626,
+      "force_loss": 0.011297706514596939,
+      "step": 1940
+    },
+    {
+      "epoch": 1.7535971223021583,
+      "grad_norm": 1.1542115211486816,
+      "learning_rate": 6.496666666666667e-05,
+      "loss": 0.0407,
+      "step": 1950
+    },
+    {
+      "action_loss": 0.02443859539926052,
+      "epoch": 1.7535971223021583,
+      "step": 1950
+    },
+    {
+      "epoch": 1.7535971223021583,
+      "step": 1950,
+      "torque_loss": 0.14138247072696686
+    },
+    {
+      "epoch": 1.7535971223021583,
+      "force_loss": 0.012840240262448788,
+      "step": 1950
+    },
+    {
+      "epoch": 1.762589928057554,
+      "grad_norm": 0.8995655179023743,
+      "learning_rate": 6.53e-05,
+      "loss": 0.0394,
+      "step": 1960
+    },
+    {
+      "action_loss": 0.014818400144577026,
+      "epoch": 1.762589928057554,
+      "step": 1960
+    },
+    {
+      "epoch": 1.762589928057554,
+      "step": 1960,
+      "torque_loss": 0.08133085817098618
+    },
+    {
+      "epoch": 1.762589928057554,
+      "force_loss": 0.005469355266541243,
+      "step": 1960
+    },
+    {
+      "epoch": 1.7715827338129495,
+      "grad_norm": 1.2325860261917114,
+      "learning_rate": 6.563333333333333e-05,
+      "loss": 0.0428,
+      "step": 1970
+    },
+    {
+      "action_loss": 0.05178338661789894,
+      "epoch": 1.7715827338129495,
+      "step": 1970
+    },
+    {
+      "epoch": 1.7715827338129495,
+      "step": 1970,
+      "torque_loss": 0.11794144660234451
+    },
+    {
+      "epoch": 1.7715827338129495,
+      "force_loss": 0.03476705029606819,
+      "step": 1970
+    },
+    {
+      "epoch": 1.7805755395683454,
+      "grad_norm": 1.9361802339553833,
+      "learning_rate": 6.596666666666667e-05,
+      "loss": 0.047,
+      "step": 1980
+    },
+    {
+      "action_loss": 0.015835566446185112,
+      "epoch": 1.7805755395683454,
+      "step": 1980
+    },
+    {
+      "epoch": 1.7805755395683454,
+      "step": 1980,
+      "torque_loss": 0.11120063066482544
+    },
+    {
+      "epoch": 1.7805755395683454,
+      "force_loss": 0.007940416224300861,
+      "step": 1980
+    },
+    {
+      "epoch": 1.789568345323741,
+      "grad_norm": 1.2197179794311523,
+      "learning_rate": 6.630000000000001e-05,
+      "loss": 0.0407,
+      "step": 1990
+    },
+    {
+      "action_loss": 0.009139064699411392,
+      "epoch": 1.789568345323741,
+      "step": 1990
+    },
+    {
+      "epoch": 1.789568345323741,
+      "step": 1990,
+      "torque_loss": 0.10894829034805298
+    },
+    {
+      "epoch": 1.789568345323741,
+      "force_loss": 0.006585546303540468,
+      "step": 1990
+    },
+    {
+      "epoch": 1.7985611510791366,
+      "grad_norm": 1.0811537504196167,
+      "learning_rate": 6.663333333333333e-05,
+      "loss": 0.0376,
+      "step": 2000
+    },
+    {
+      "action_loss": 0.01651059091091156,
+      "epoch": 1.7985611510791366,
+      "step": 2000
+    },
+    {
+      "epoch": 1.7985611510791366,
+      "step": 2000,
+      "torque_loss": 0.1283266544342041
+    },
+    {
+      "epoch": 1.7985611510791366,
+      "force_loss": 0.008875512517988682,
+      "step": 2000
+    },
+    {
+      "epoch": 1.8075539568345325,
+      "grad_norm": 2.1722841262817383,
+      "learning_rate": 6.696666666666666e-05,
+      "loss": 0.0382,
+      "step": 2010
+    },
+    {
+      "action_loss": 0.0240608062595129,
+      "epoch": 1.8075539568345325,
+      "step": 2010
+    },
+    {
+      "epoch": 1.8075539568345325,
+      "step": 2010,
+      "torque_loss": 0.12283951044082642
+    },
+    {
+      "epoch": 1.8075539568345325,
+      "force_loss": 0.012677472084760666,
+      "step": 2010
+    },
+    {
+      "epoch": 1.816546762589928,
+      "grad_norm": 1.5948004722595215,
+      "learning_rate": 6.730000000000001e-05,
+      "loss": 0.0372,
+      "step": 2020
+    },
+    {
+      "action_loss": 0.017273152247071266,
+      "epoch": 1.816546762589928,
+      "step": 2020
+    },
+    {
+      "epoch": 1.816546762589928,
+      "step": 2020,
+      "torque_loss": 0.1408601701259613
+    },
+    {
+      "epoch": 1.816546762589928,
+      "force_loss": 0.009422182105481625,
+      "step": 2020
+    },
+    {
+      "epoch": 1.8255395683453237,
+      "grad_norm": 1.5452513694763184,
+      "learning_rate": 6.763333333333334e-05,
+      "loss": 0.0389,
+      "step": 2030
+    },
+    {
+      "action_loss": 0.012955605052411556,
+      "epoch": 1.8255395683453237,
+      "step": 2030
+    },
+    {
+      "epoch": 1.8255395683453237,
+      "step": 2030,
+      "torque_loss": 0.13550597429275513
+    },
+    {
+      "epoch": 1.8255395683453237,
+      "force_loss": 0.00767833786085248,
+      "step": 2030
+    },
+    {
+      "epoch": 1.8345323741007196,
+      "grad_norm": 1.8239887952804565,
+      "learning_rate": 6.796666666666666e-05,
+      "loss": 0.0347,
+      "step": 2040
+    },
+    {
+      "action_loss": 0.016966264694929123,
+      "epoch": 1.8345323741007196,
+      "step": 2040
+    },
+    {
+      "epoch": 1.8345323741007196,
+      "step": 2040,
+      "torque_loss": 0.1556432992219925
+    },
+    {
+      "epoch": 1.8345323741007196,
+      "force_loss": 0.008319436572492123,
+      "step": 2040
+    },
+    {
+      "epoch": 1.843525179856115,
+      "grad_norm": 0.7103005051612854,
+      "learning_rate": 6.83e-05,
+      "loss": 0.0374,
+      "step": 2050
+    },
+    {
+      "action_loss": 0.010314648039638996,
+      "epoch": 1.843525179856115,
+      "step": 2050
+    },
+    {
+      "epoch": 1.843525179856115,
+      "step": 2050,
+      "torque_loss": 0.10243510454893112
+    },
+    {
+      "epoch": 1.843525179856115,
+      "force_loss": 0.006999240722507238,
+      "step": 2050
+    },
+    {
+      "epoch": 1.8525179856115108,
+      "grad_norm": 1.2091851234436035,
+      "learning_rate": 6.863333333333334e-05,
+      "loss": 0.041,
+      "step": 2060
+    },
+    {
+      "action_loss": 0.01964065432548523,
+      "epoch": 1.8525179856115108,
+      "step": 2060
+    },
+    {
+      "epoch": 1.8525179856115108,
+      "step": 2060,
+      "torque_loss": 0.2040068656206131
+    },
+    {
+      "epoch": 1.8525179856115108,
+      "force_loss": 0.011498930864036083,
+      "step": 2060
+    },
+    {
+      "epoch": 1.8615107913669064,
+      "grad_norm": 0.8997297883033752,
+      "learning_rate": 6.896666666666667e-05,
+      "loss": 0.041,
+      "step": 2070
+    },
+    {
+      "action_loss": 0.014001811854541302,
+      "epoch": 1.8615107913669064,
+      "step": 2070
+    },
+    {
+      "epoch": 1.8615107913669064,
+      "step": 2070,
+      "torque_loss": 0.14261697232723236
+    },
+    {
+      "epoch": 1.8615107913669064,
+      "force_loss": 0.013104937970638275,
+      "step": 2070
+    },
+    {
+      "epoch": 1.870503597122302,
+      "grad_norm": 1.457098364830017,
+      "learning_rate": 6.93e-05,
+      "loss": 0.0413,
+      "step": 2080
+    },
+    {
+      "action_loss": 0.02490692026913166,
+      "epoch": 1.870503597122302,
+      "step": 2080
+    },
+    {
+      "epoch": 1.870503597122302,
+      "step": 2080,
+      "torque_loss": 0.13433009386062622
+    },
+    {
+      "epoch": 1.870503597122302,
+      "force_loss": 0.020111190155148506,
+      "step": 2080
+    },
+    {
+      "epoch": 1.879496402877698,
+      "grad_norm": 1.1311707496643066,
+      "learning_rate": 6.963333333333334e-05,
+      "loss": 0.0345,
+      "step": 2090
+    },
+    {
+      "action_loss": 0.016696559265255928,
+      "epoch": 1.879496402877698,
+      "step": 2090
+    },
+    {
+      "epoch": 1.879496402877698,
+      "step": 2090,
+      "torque_loss": 0.12842392921447754
+    },
+    {
+      "epoch": 1.879496402877698,
+      "force_loss": 0.011607293970882893,
+      "step": 2090
+    },
+    {
+      "epoch": 1.8884892086330936,
+      "grad_norm": 1.4410679340362549,
+      "learning_rate": 6.996666666666667e-05,
+      "loss": 0.0369,
+      "step": 2100
+    },
+    {
+      "action_loss": 0.013090457767248154,
+      "epoch": 1.8884892086330936,
+      "step": 2100
+    },
+    {
+      "epoch": 1.8884892086330936,
+      "step": 2100,
+      "torque_loss": 0.11940749734640121
+    },
+    {
+      "epoch": 1.8884892086330936,
+      "force_loss": 0.005699618253856897,
+      "step": 2100
+    },
+    {
+      "epoch": 1.8974820143884892,
+      "grad_norm": 0.9415628910064697,
+      "learning_rate": 7.03e-05,
+      "loss": 0.0432,
+      "step": 2110
+    },
+    {
+      "action_loss": 0.020249422639608383,
+      "epoch": 1.8974820143884892,
+      "step": 2110
+    },
+    {
+      "epoch": 1.8974820143884892,
+      "step": 2110,
+      "torque_loss": 0.14470814168453217
+    },
+    {
+      "epoch": 1.8974820143884892,
+      "force_loss": 0.007382591720670462,
+      "step": 2110
+    },
+    {
+      "epoch": 1.906474820143885,
+      "grad_norm": 1.7874438762664795,
+      "learning_rate": 7.063333333333333e-05,
+      "loss": 0.051,
+      "step": 2120
+    },
+    {
+      "action_loss": 0.016154659911990166,
+      "epoch": 1.906474820143885,
+      "step": 2120
+    },
+    {
+      "epoch": 1.906474820143885,
+      "step": 2120,
+      "torque_loss": 0.18611186742782593
+    },
+    {
+      "epoch": 1.906474820143885,
+      "force_loss": 0.013002969324588776,
+      "step": 2120
+    },
+    {
+      "epoch": 1.9154676258992804,
+      "grad_norm": 1.8646923303604126,
+      "learning_rate": 7.096666666666667e-05,
+      "loss": 0.0476,
+      "step": 2130
+    },
+    {
+      "action_loss": 0.03132406249642372,
+      "epoch": 1.9154676258992804,
+      "step": 2130
+    },
+    {
+      "epoch": 1.9154676258992804,
+      "step": 2130,
+      "torque_loss": 0.1391480565071106
+    },
+    {
+      "epoch": 1.9154676258992804,
+      "force_loss": 0.014118481427431107,
+      "step": 2130
+    },
+    {
+      "epoch": 1.9244604316546763,
+      "grad_norm": 0.6361132860183716,
+      "learning_rate": 7.13e-05,
+      "loss": 0.0414,
+      "step": 2140
+    },
+    {
+      "action_loss": 0.020025895908474922,
+      "epoch": 1.9244604316546763,
+      "step": 2140
+    },
+    {
+      "epoch": 1.9244604316546763,
+      "step": 2140,
+      "torque_loss": 0.17662544548511505
+    },
+    {
+      "epoch": 1.9244604316546763,
+      "force_loss": 0.023267442360520363,
+      "step": 2140
+    },
+    {
+      "epoch": 1.933453237410072,
+      "grad_norm": 0.46818807721138,
+      "learning_rate": 7.163333333333334e-05,
+      "loss": 0.0366,
+      "step": 2150
+    },
+    {
+      "action_loss": 0.0124775106087327,
+      "epoch": 1.933453237410072,
+      "step": 2150
+    },
+    {
+      "epoch": 1.933453237410072,
+      "step": 2150,
+      "torque_loss": 0.13043367862701416
+    },
+    {
+      "epoch": 1.933453237410072,
+      "force_loss": 0.0071367863565683365,
+      "step": 2150
+    },
+    {
+      "epoch": 1.9424460431654675,
+      "grad_norm": 0.7108094096183777,
+      "learning_rate": 7.196666666666668e-05,
+      "loss": 0.0459,
+      "step": 2160
+    },
+    {
+      "action_loss": 0.024643266573548317,
+      "epoch": 1.9424460431654675,
+      "step": 2160
+    },
+    {
+      "epoch": 1.9424460431654675,
+      "step": 2160,
+      "torque_loss": 0.16619999706745148
+    },
+    {
+      "epoch": 1.9424460431654675,
+      "force_loss": 0.009283944964408875,
+      "step": 2160
+    },
+    {
+      "epoch": 1.9514388489208634,
+      "grad_norm": 0.7339408993721008,
+      "learning_rate": 7.23e-05,
+      "loss": 0.0328,
+      "step": 2170
+    },
+    {
+      "action_loss": 0.02632194198668003,
+      "epoch": 1.9514388489208634,
+      "step": 2170
+    },
+    {
+      "epoch": 1.9514388489208634,
+      "step": 2170,
+      "torque_loss": 0.10700962692499161
+    },
+    {
+      "epoch": 1.9514388489208634,
+      "force_loss": 0.022282054647803307,
+      "step": 2170
+    },
+    {
+      "epoch": 1.960431654676259,
+      "grad_norm": 1.891281008720398,
+      "learning_rate": 7.263333333333334e-05,
+      "loss": 0.0284,
+      "step": 2180
+    },
+    {
+      "action_loss": 0.019738411530852318,
+      "epoch": 1.960431654676259,
+      "step": 2180
+    },
+    {
+      "epoch": 1.960431654676259,
+      "step": 2180,
+      "torque_loss": 0.1465553194284439
+    },
+    {
+      "epoch": 1.960431654676259,
+      "force_loss": 0.020632626488804817,
+      "step": 2180
+    },
+    {
+      "epoch": 1.9694244604316546,
+      "grad_norm": 1.102104663848877,
+      "learning_rate": 7.296666666666667e-05,
+      "loss": 0.0365,
+      "step": 2190
+    },
+    {
+      "action_loss": 0.02356117218732834,
+      "epoch": 1.9694244604316546,
+      "step": 2190
+    },
+    {
+      "epoch": 1.9694244604316546,
+      "step": 2190,
+      "torque_loss": 0.14908255636692047
+    },
+    {
+      "epoch": 1.9694244604316546,
+      "force_loss": 0.012325418181717396,
+      "step": 2190
+    },
+    {
+      "epoch": 1.9784172661870505,
+      "grad_norm": 2.1154210567474365,
+      "learning_rate": 7.33e-05,
+      "loss": 0.0421,
+      "step": 2200
+    },
+    {
+      "action_loss": 0.04876772686839104,
+      "epoch": 1.9784172661870505,
+      "step": 2200
+    },
+    {
+      "epoch": 1.9784172661870505,
+      "step": 2200,
+      "torque_loss": 0.15356247127056122
+    },
+    {
+      "epoch": 1.9784172661870505,
+      "force_loss": 0.020470868796110153,
+      "step": 2200
+    },
+    {
+      "epoch": 1.987410071942446,
+      "grad_norm": 1.734969973564148,
+      "learning_rate": 7.363333333333334e-05,
+      "loss": 0.0464,
+      "step": 2210
+    },
+    {
+      "action_loss": 0.029785921797156334,
+      "epoch": 1.987410071942446,
+      "step": 2210
+    },
+    {
+      "epoch": 1.987410071942446,
+      "step": 2210,
+      "torque_loss": 0.18497033417224884
+    },
+    {
+      "epoch": 1.987410071942446,
+      "force_loss": 0.01955021359026432,
+      "step": 2210
+    },
+    {
+      "epoch": 1.9964028776978417,
+      "grad_norm": 1.0125097036361694,
+      "learning_rate": 7.396666666666667e-05,
+      "loss": 0.04,
+      "step": 2220
+    },
+    {
+      "action_loss": 0.02043708600103855,
+      "epoch": 1.9964028776978417,
+      "step": 2220
+    },
+    {
+      "epoch": 1.9964028776978417,
+      "step": 2220,
+      "torque_loss": 0.172343909740448
+    },
+    {
+      "epoch": 1.9964028776978417,
+      "force_loss": 0.016166238114237785,
+      "step": 2220
+    },
+    {
+      "epoch": 2.0053956834532376,
+      "grad_norm": 0.709406852722168,
+      "learning_rate": 7.43e-05,
+      "loss": 0.045,
+      "step": 2230
+    },
+    {
+      "action_loss": 0.03095175139605999,
+      "epoch": 2.0053956834532376,
+      "step": 2230
+    },
+    {
+      "epoch": 2.0053956834532376,
+      "step": 2230,
+      "torque_loss": 0.19164550304412842
+    },
+    {
+      "epoch": 2.0053956834532376,
+      "force_loss": 0.020785285159945488,
+      "step": 2230
+    },
+    {
+      "epoch": 2.014388489208633,
+      "grad_norm": 0.7892938256263733,
+      "learning_rate": 7.463333333333334e-05,
+      "loss": 0.0433,
+      "step": 2240
+    },
+    {
+      "action_loss": 0.034813642501831055,
+      "epoch": 2.014388489208633,
+      "step": 2240
+    },
+    {
+      "epoch": 2.014388489208633,
+      "step": 2240,
+      "torque_loss": 0.17479079961776733
+    },
+    {
+      "epoch": 2.014388489208633,
+      "force_loss": 0.018014518544077873,
+      "step": 2240
+    },
+    {
+      "epoch": 2.023381294964029,
+      "grad_norm": 0.5668711066246033,
+      "learning_rate": 7.496666666666667e-05,
+      "loss": 0.0409,
+      "step": 2250
+    },
+    {
+      "action_loss": 0.021270493045449257,
+      "epoch": 2.023381294964029,
+      "step": 2250
+    },
+    {
+      "epoch": 2.023381294964029,
+      "step": 2250,
+      "torque_loss": 0.10206601768732071
+    },
+    {
+      "epoch": 2.023381294964029,
+      "force_loss": 0.011215499602258205,
+      "step": 2250
+    },
+    {
+      "epoch": 2.0323741007194243,
+      "grad_norm": 1.3354475498199463,
+      "learning_rate": 7.53e-05,
+      "loss": 0.0515,
+      "step": 2260
+    },
+    {
+      "action_loss": 0.016435161232948303,
+      "epoch": 2.0323741007194243,
+      "step": 2260
+    },
+    {
+      "epoch": 2.0323741007194243,
+      "step": 2260,
+      "torque_loss": 0.16192492842674255
+    },
+    {
+      "epoch": 2.0323741007194243,
+      "force_loss": 0.008937232196331024,
+      "step": 2260
+    },
+    {
+      "epoch": 2.04136690647482,
+      "grad_norm": 1.3260750770568848,
+      "learning_rate": 7.563333333333333e-05,
+      "loss": 0.0438,
+      "step": 2270
+    },
+    {
+      "action_loss": 0.05003343150019646,
+      "epoch": 2.04136690647482,
+      "step": 2270
+    },
+    {
+      "epoch": 2.04136690647482,
+      "step": 2270,
+      "torque_loss": 0.15236492455005646
+    },
+    {
+      "epoch": 2.04136690647482,
+      "force_loss": 0.028767839074134827,
+      "step": 2270
+    },
+    {
+      "epoch": 2.050359712230216,
+      "grad_norm": 1.6289191246032715,
+      "learning_rate": 7.596666666666668e-05,
+      "loss": 0.0442,
+      "step": 2280
+    },
+    {
+      "action_loss": 0.032095033675432205,
+      "epoch": 2.050359712230216,
+      "step": 2280
+    },
+    {
+      "epoch": 2.050359712230216,
+      "step": 2280,
+      "torque_loss": 0.15725067257881165
+    },
+    {
+      "epoch": 2.050359712230216,
+      "force_loss": 0.01950998790562153,
+      "step": 2280
+    },
+    {
+      "epoch": 2.0593525179856114,
+      "grad_norm": 1.321658968925476,
+      "learning_rate": 7.630000000000001e-05,
+      "loss": 0.0451,
+      "step": 2290
+    },
+    {
+      "action_loss": 0.01688532717525959,
+      "epoch": 2.0593525179856114,
+      "step": 2290
+    },
+    {
+      "epoch": 2.0593525179856114,
+      "step": 2290,
+      "torque_loss": 0.1754584163427353
+    },
+    {
+      "epoch": 2.0593525179856114,
+      "force_loss": 0.00883524026721716,
+      "step": 2290
+    },
+    {
+      "epoch": 2.068345323741007,
+      "grad_norm": 1.4887975454330444,
+      "learning_rate": 7.663333333333333e-05,
+      "loss": 0.0428,
+      "step": 2300
+    },
+    {
+      "action_loss": 0.010168353095650673,
+      "epoch": 2.068345323741007,
+      "step": 2300
+    },
+    {
+      "epoch": 2.068345323741007,
+      "step": 2300,
+      "torque_loss": 0.13163544237613678
+    },
+    {
+      "epoch": 2.068345323741007,
+      "force_loss": 0.009628458879888058,
+      "step": 2300
+    },
+    {
+      "epoch": 2.077338129496403,
+      "grad_norm": 2.233642578125,
+      "learning_rate": 7.696666666666668e-05,
+      "loss": 0.0363,
+      "step": 2310
+    },
+    {
+      "action_loss": 0.02015838585793972,
+      "epoch": 2.077338129496403,
+      "step": 2310
+    },
+    {
+      "epoch": 2.077338129496403,
+      "step": 2310,
+      "torque_loss": 0.12331744283437729
+    },
+    {
+      "epoch": 2.077338129496403,
+      "force_loss": 0.01339507382363081,
+      "step": 2310
+    },
+    {
+      "epoch": 2.0863309352517985,
+      "grad_norm": 1.5707989931106567,
+      "learning_rate": 7.730000000000001e-05,
+      "loss": 0.0444,
+      "step": 2320
+    },
+    {
+      "action_loss": 0.016280699521303177,
+      "epoch": 2.0863309352517985,
+      "step": 2320
+    },
+    {
+      "epoch": 2.0863309352517985,
+      "step": 2320,
+      "torque_loss": 0.10791879892349243
+    },
+    {
+      "epoch": 2.0863309352517985,
+      "force_loss": 0.007361226249486208,
+      "step": 2320
+    },
+    {
+      "epoch": 2.0953237410071943,
+      "grad_norm": 1.6029785871505737,
+      "learning_rate": 7.763333333333334e-05,
+      "loss": 0.0426,
+      "step": 2330
+    },
+    {
+      "action_loss": 0.02585158683359623,
+      "epoch": 2.0953237410071943,
+      "step": 2330
+    },
+    {
+      "epoch": 2.0953237410071943,
+      "step": 2330,
+      "torque_loss": 0.11488571017980576
+    },
+    {
+      "epoch": 2.0953237410071943,
+      "force_loss": 0.008152122609317303,
+      "step": 2330
+    },
+    {
+      "epoch": 2.1043165467625897,
+      "grad_norm": 1.4889627695083618,
+      "learning_rate": 7.796666666666666e-05,
+      "loss": 0.0398,
+      "step": 2340
+    },
+    {
+      "action_loss": 0.011856448836624622,
+      "epoch": 2.1043165467625897,
+      "step": 2340
+    },
+    {
+      "epoch": 2.1043165467625897,
+      "step": 2340,
+      "torque_loss": 0.18074285984039307
+    },
+    {
+      "epoch": 2.1043165467625897,
+      "force_loss": 0.008535501547157764,
+      "step": 2340
+    },
+    {
+      "epoch": 2.1133093525179856,
+      "grad_norm": 1.0595521926879883,
+      "learning_rate": 7.83e-05,
+      "loss": 0.0382,
+      "step": 2350
+    },
+    {
+      "action_loss": 0.015816258266568184,
+      "epoch": 2.1133093525179856,
+      "step": 2350
+    },
+    {
+      "epoch": 2.1133093525179856,
+      "step": 2350,
+      "torque_loss": 0.1994006186723709
+    },
+    {
+      "epoch": 2.1133093525179856,
+      "force_loss": 0.008162776939570904,
+      "step": 2350
+    },
+    {
+      "epoch": 2.1223021582733814,
+      "grad_norm": 2.5009734630584717,
+      "learning_rate": 7.863333333333334e-05,
+      "loss": 0.0371,
+      "step": 2360
+    },
+    {
+      "action_loss": 0.019855961203575134,
+      "epoch": 2.1223021582733814,
+      "step": 2360
+    },
+    {
+      "epoch": 2.1223021582733814,
+      "step": 2360,
+      "torque_loss": 0.09618862718343735
+    },
+    {
+      "epoch": 2.1223021582733814,
+      "force_loss": 0.027574585750699043,
+      "step": 2360
+    },
+    {
+      "epoch": 2.131294964028777,
+      "grad_norm": 1.38224196434021,
+      "learning_rate": 7.896666666666667e-05,
+      "loss": 0.0413,
+      "step": 2370
+    },
+    {
+      "action_loss": 0.012773166410624981,
+      "epoch": 2.131294964028777,
+      "step": 2370
+    },
+    {
+      "epoch": 2.131294964028777,
+      "step": 2370,
+      "torque_loss": 0.14643804728984833
+    },
+    {
+      "epoch": 2.131294964028777,
+      "force_loss": 0.00693405419588089,
+      "step": 2370
+    },
+    {
+      "epoch": 2.1402877697841727,
+      "grad_norm": 2.8662095069885254,
+      "learning_rate": 7.93e-05,
+      "loss": 0.0382,
+      "step": 2380
+    },
+    {
+      "action_loss": 0.019567472860217094,
+      "epoch": 2.1402877697841727,
+      "step": 2380
+    },
+    {
+      "epoch": 2.1402877697841727,
+      "step": 2380,
+      "torque_loss": 0.10939296334981918
+    },
+    {
+      "epoch": 2.1402877697841727,
+      "force_loss": 0.011409125290811062,
+      "step": 2380
+    },
+    {
+      "epoch": 2.1492805755395685,
+      "grad_norm": 0.5628333687782288,
+      "learning_rate": 7.963333333333334e-05,
+      "loss": 0.0443,
+      "step": 2390
+    },
+    {
+      "action_loss": 0.015659721568226814,
+      "epoch": 2.1492805755395685,
+      "step": 2390
+    },
+    {
+      "epoch": 2.1492805755395685,
+      "step": 2390,
+      "torque_loss": 0.1429305225610733
+    },
+    {
+      "epoch": 2.1492805755395685,
+      "force_loss": 0.013227318413555622,
+      "step": 2390
+    },
+    {
+      "epoch": 2.158273381294964,
+      "grad_norm": 1.2696839570999146,
+      "learning_rate": 7.996666666666667e-05,
+      "loss": 0.0361,
+      "step": 2400
+    },
+    {
+      "action_loss": 0.014651473611593246,
+      "epoch": 2.158273381294964,
+      "step": 2400
+    },
+    {
+      "epoch": 2.158273381294964,
+      "step": 2400,
+      "torque_loss": 0.12815704941749573
+    },
+    {
+      "epoch": 2.158273381294964,
+      "force_loss": 0.010313412174582481,
+      "step": 2400
+    },
+    {
+      "epoch": 2.16726618705036,
+      "grad_norm": 1.1857028007507324,
+      "learning_rate": 8.030000000000001e-05,
+      "loss": 0.0399,
+      "step": 2410
+    },
+    {
+      "action_loss": 0.020552780479192734,
+      "epoch": 2.16726618705036,
+      "step": 2410
+    },
+    {
+      "epoch": 2.16726618705036,
+      "step": 2410,
+      "torque_loss": 0.08756649494171143
+    },
+    {
+      "epoch": 2.16726618705036,
+      "force_loss": 0.01110069826245308,
+      "step": 2410
+    },
+    {
+      "epoch": 2.176258992805755,
+      "grad_norm": 1.1977838277816772,
+      "learning_rate": 8.063333333333333e-05,
+      "loss": 0.0384,
+      "step": 2420
+    },
+    {
+      "action_loss": 0.03344753757119179,
+      "epoch": 2.176258992805755,
+      "step": 2420
+    },
+    {
+      "epoch": 2.176258992805755,
+      "step": 2420,
+      "torque_loss": 0.18582697212696075
+    },
+    {
+      "epoch": 2.176258992805755,
+      "force_loss": 0.020389042794704437,
+      "step": 2420
+    },
+    {
+      "epoch": 2.185251798561151,
+      "grad_norm": 0.6409598588943481,
+      "learning_rate": 8.096666666666667e-05,
+      "loss": 0.0352,
+      "step": 2430
+    },
+    {
+      "action_loss": 0.021592723205685616,
+      "epoch": 2.185251798561151,
+      "step": 2430
+    },
+    {
+      "epoch": 2.185251798561151,
+      "step": 2430,
+      "torque_loss": 0.11224746704101562
+    },
+    {
+      "epoch": 2.185251798561151,
+      "force_loss": 0.029283076524734497,
+      "step": 2430
+    },
+    {
+      "epoch": 2.194244604316547,
+      "grad_norm": 1.0430068969726562,
+      "learning_rate": 8.13e-05,
+      "loss": 0.0359,
+      "step": 2440
+    },
+    {
+      "action_loss": 0.021318459883332253,
+      "epoch": 2.194244604316547,
+      "step": 2440
+    },
+    {
+      "epoch": 2.194244604316547,
+      "step": 2440,
+      "torque_loss": 0.14711694419384003
+    },
+    {
+      "epoch": 2.194244604316547,
+      "force_loss": 0.020500699058175087,
+      "step": 2440
+    },
+    {
+      "epoch": 2.2032374100719423,
+      "grad_norm": 0.7934525012969971,
+      "learning_rate": 8.163333333333334e-05,
+      "loss": 0.0459,
+      "step": 2450
+    },
+    {
+      "action_loss": 0.014271169900894165,
+      "epoch": 2.2032374100719423,
+      "step": 2450
+    },
+    {
+      "epoch": 2.2032374100719423,
+      "step": 2450,
+      "torque_loss": 0.10306980460882187
+    },
+    {
+      "epoch": 2.2032374100719423,
+      "force_loss": 0.007952618412673473,
+      "step": 2450
+    },
+    {
+      "epoch": 2.212230215827338,
+      "grad_norm": 0.858775794506073,
+      "learning_rate": 8.196666666666668e-05,
+      "loss": 0.0423,
+      "step": 2460
+    },
+    {
+      "action_loss": 0.02108023501932621,
+      "epoch": 2.212230215827338,
+      "step": 2460
+    },
+    {
+      "epoch": 2.212230215827338,
+      "step": 2460,
+      "torque_loss": 0.19167710840702057
+    },
+    {
+      "epoch": 2.212230215827338,
+      "force_loss": 0.012655667029321194,
+      "step": 2460
+    },
+    {
+      "epoch": 2.221223021582734,
+      "grad_norm": 0.8969601988792419,
+      "learning_rate": 8.23e-05,
+      "loss": 0.0351,
+      "step": 2470
+    },
+    {
+      "action_loss": 0.012416555546224117,
+      "epoch": 2.221223021582734,
+      "step": 2470
+    },
+    {
+      "epoch": 2.221223021582734,
+      "step": 2470,
+      "torque_loss": 0.1196342408657074
+    },
+    {
+      "epoch": 2.221223021582734,
+      "force_loss": 0.008173524402081966,
+      "step": 2470
+    },
+    {
+      "epoch": 2.2302158273381294,
+      "grad_norm": 1.3362475633621216,
+      "learning_rate": 8.263333333333334e-05,
+      "loss": 0.0419,
+      "step": 2480
+    },
+    {
+      "action_loss": 0.048754189163446426,
+      "epoch": 2.2302158273381294,
+      "step": 2480
+    },
+    {
+      "epoch": 2.2302158273381294,
+      "step": 2480,
+      "torque_loss": 0.20010678470134735
+    },
+    {
+      "epoch": 2.2302158273381294,
+      "force_loss": 0.06305330246686935,
+      "step": 2480
+    },
+    {
+      "epoch": 2.2392086330935252,
+      "grad_norm": 1.9638789892196655,
+      "learning_rate": 8.296666666666667e-05,
+      "loss": 0.0458,
+      "step": 2490
+    },
+    {
+      "action_loss": 0.021870682016015053,
+      "epoch": 2.2392086330935252,
+      "step": 2490
+    },
+    {
+      "epoch": 2.2392086330935252,
+      "step": 2490,
+      "torque_loss": 0.11235183477401733
+    },
+    {
+      "epoch": 2.2392086330935252,
+      "force_loss": 0.011008105240762234,
+      "step": 2490
+    },
+    {
+      "epoch": 2.2482014388489207,
+      "grad_norm": 1.2399091720581055,
+      "learning_rate": 8.33e-05,
+      "loss": 0.0437,
+      "step": 2500
+    },
+    {
+      "action_loss": 0.036117155104875565,
+      "epoch": 2.2482014388489207,
+      "step": 2500
+    },
+    {
+      "epoch": 2.2482014388489207,
+      "step": 2500,
+      "torque_loss": 0.1411147117614746
+    },
+    {
+      "epoch": 2.2482014388489207,
+      "force_loss": 0.027891939505934715,
+      "step": 2500
+    },
+    {
+      "epoch": 2.2571942446043165,
+      "grad_norm": 1.6755642890930176,
+      "learning_rate": 8.363333333333334e-05,
+      "loss": 0.0407,
+      "step": 2510
+    },
+    {
+      "action_loss": 0.024394290521740913,
+      "epoch": 2.2571942446043165,
+      "step": 2510
+    },
+    {
+      "epoch": 2.2571942446043165,
+      "step": 2510,
+      "torque_loss": 0.1530729979276657
+    },
+    {
+      "epoch": 2.2571942446043165,
+      "force_loss": 0.016769492998719215,
+      "step": 2510
+    },
+    {
+      "epoch": 2.2661870503597124,
+      "grad_norm": 0.5203109979629517,
+      "learning_rate": 8.396666666666667e-05,
+      "loss": 0.0356,
+      "step": 2520
+    },
+    {
+      "action_loss": 0.029405226930975914,
+      "epoch": 2.2661870503597124,
+      "step": 2520
+    },
+    {
+      "epoch": 2.2661870503597124,
+      "step": 2520,
+      "torque_loss": 0.12017150968313217
+    },
+    {
+      "epoch": 2.2661870503597124,
+      "force_loss": 0.019859282299876213,
+      "step": 2520
+    },
+    {
+      "epoch": 2.2751798561151078,
+      "grad_norm": 1.0105420351028442,
+      "learning_rate": 8.43e-05,
+      "loss": 0.0331,
+      "step": 2530
+    },
+    {
+      "action_loss": 0.03128426522016525,
+      "epoch": 2.2751798561151078,
+      "step": 2530
+    },
+    {
+      "epoch": 2.2751798561151078,
+      "step": 2530,
+      "torque_loss": 0.1860082596540451
+    },
+    {
+      "epoch": 2.2751798561151078,
+      "force_loss": 0.021311545744538307,
+      "step": 2530
+    },
+    {
+      "epoch": 2.2841726618705036,
+      "grad_norm": 0.5562395453453064,
+      "learning_rate": 8.463333333333335e-05,
+      "loss": 0.0478,
+      "step": 2540
+    },
+    {
+      "action_loss": 0.030484000220894814,
+      "epoch": 2.2841726618705036,
+      "step": 2540
+    },
+    {
+      "epoch": 2.2841726618705036,
+      "step": 2540,
+      "torque_loss": 0.14260172843933105
+    },
+    {
+      "epoch": 2.2841726618705036,
+      "force_loss": 0.026872314512729645,
+      "step": 2540
+    },
+    {
+      "epoch": 2.2931654676258995,
+      "grad_norm": 1.1357439756393433,
+      "learning_rate": 8.496666666666667e-05,
+      "loss": 0.0358,
+      "step": 2550
+    },
+    {
+      "action_loss": 0.011914070695638657,
+      "epoch": 2.2931654676258995,
+      "step": 2550
+    },
+    {
+      "epoch": 2.2931654676258995,
+      "step": 2550,
+      "torque_loss": 0.13292258977890015
+    },
+    {
+      "epoch": 2.2931654676258995,
+      "force_loss": 0.009125075303018093,
+      "step": 2550
+    },
+    {
+      "epoch": 2.302158273381295,
+      "grad_norm": 1.1095728874206543,
+      "learning_rate": 8.53e-05,
+      "loss": 0.0458,
+      "step": 2560
+    },
+    {
+      "action_loss": 0.026217376813292503,
+      "epoch": 2.302158273381295,
+      "step": 2560
+    },
+    {
+      "epoch": 2.302158273381295,
+      "step": 2560,
+      "torque_loss": 0.1510026752948761
+    },
+    {
+      "epoch": 2.302158273381295,
+      "force_loss": 0.031473398208618164,
+      "step": 2560
+    },
+    {
+      "epoch": 2.3111510791366907,
+      "grad_norm": 1.512447476387024,
+      "learning_rate": 8.563333333333333e-05,
+      "loss": 0.0442,
+      "step": 2570
+    },
+    {
+      "action_loss": 0.02989950217306614,
+      "epoch": 2.3111510791366907,
+      "step": 2570
+    },
+    {
+      "epoch": 2.3111510791366907,
+      "step": 2570,
+      "torque_loss": 0.11103609949350357
+    },
+    {
+      "epoch": 2.3111510791366907,
+      "force_loss": 0.015186742879450321,
+      "step": 2570
+    },
+    {
+      "epoch": 2.3201438848920866,
+      "grad_norm": 0.44876179099082947,
+      "learning_rate": 8.596666666666668e-05,
+      "loss": 0.0451,
+      "step": 2580
+    },
+    {
+      "action_loss": 0.0363553948700428,
+      "epoch": 2.3201438848920866,
+      "step": 2580
+    },
+    {
+      "epoch": 2.3201438848920866,
+      "step": 2580,
+      "torque_loss": 0.16425319015979767
+    },
+    {
+      "epoch": 2.3201438848920866,
+      "force_loss": 0.02153819240629673,
+      "step": 2580
+    },
+    {
+      "epoch": 2.329136690647482,
+      "grad_norm": 1.0684629678726196,
+      "learning_rate": 8.63e-05,
+      "loss": 0.0386,
+      "step": 2590
+    },
+    {
+      "action_loss": 0.01931898109614849,
+      "epoch": 2.329136690647482,
+      "step": 2590
+    },
+    {
+      "epoch": 2.329136690647482,
+      "step": 2590,
+      "torque_loss": 0.13259916007518768
+    },
+    {
+      "epoch": 2.329136690647482,
+      "force_loss": 0.013353735208511353,
+      "step": 2590
+    },
+    {
+      "epoch": 2.338129496402878,
+      "grad_norm": 2.289297580718994,
+      "learning_rate": 8.663333333333333e-05,
+      "loss": 0.04,
+      "step": 2600
+    },
+    {
+      "action_loss": 0.020353885367512703,
+      "epoch": 2.338129496402878,
+      "step": 2600
+    },
+    {
+      "epoch": 2.338129496402878,
+      "step": 2600,
+      "torque_loss": 0.12929637730121613
+    },
+    {
+      "epoch": 2.338129496402878,
+      "force_loss": 0.015994930639863014,
+      "step": 2600
+    },
+    {
+      "epoch": 2.347122302158273,
+      "grad_norm": 1.6838279962539673,
+      "learning_rate": 8.696666666666668e-05,
+      "loss": 0.0422,
+      "step": 2610
+    },
+    {
+      "action_loss": 0.01813647337257862,
+      "epoch": 2.347122302158273,
+      "step": 2610
+    },
+    {
+      "epoch": 2.347122302158273,
+      "step": 2610,
+      "torque_loss": 0.10282281041145325
+    },
+    {
+      "epoch": 2.347122302158273,
+      "force_loss": 0.010300427675247192,
+      "step": 2610
+    },
+    {
+      "epoch": 2.356115107913669,
+      "grad_norm": 0.9916293621063232,
+      "learning_rate": 8.730000000000001e-05,
+      "loss": 0.0379,
+      "step": 2620
+    },
+    {
+      "action_loss": 0.038559798151254654,
+      "epoch": 2.356115107913669,
+      "step": 2620
+    },
+    {
+      "epoch": 2.356115107913669,
+      "step": 2620,
+      "torque_loss": 0.15211163461208344
+    },
+    {
+      "epoch": 2.356115107913669,
+      "force_loss": 0.020414920523762703,
+      "step": 2620
+    },
+    {
+      "epoch": 2.365107913669065,
+      "grad_norm": 0.9382750391960144,
+      "learning_rate": 8.763333333333334e-05,
+      "loss": 0.0407,
+      "step": 2630
+    },
+    {
+      "action_loss": 0.03999155014753342,
+      "epoch": 2.365107913669065,
+      "step": 2630
+    },
+    {
+      "epoch": 2.365107913669065,
+      "step": 2630,
+      "torque_loss": 0.1645486205816269
+    },
+    {
+      "epoch": 2.365107913669065,
+      "force_loss": 0.023098522797226906,
+      "step": 2630
+    },
+    {
+      "epoch": 2.3741007194244603,
+      "grad_norm": 0.8879019618034363,
+      "learning_rate": 8.796666666666667e-05,
+      "loss": 0.0409,
+      "step": 2640
+    },
+    {
+      "action_loss": 0.04192976653575897,
+      "epoch": 2.3741007194244603,
+      "step": 2640
+    },
+    {
+      "epoch": 2.3741007194244603,
+      "step": 2640,
+      "torque_loss": 0.1851922869682312
+    },
+    {
+      "epoch": 2.3741007194244603,
+      "force_loss": 0.02950838953256607,
+      "step": 2640
+    },
+    {
+      "epoch": 2.383093525179856,
+      "grad_norm": 0.6435950398445129,
+      "learning_rate": 8.83e-05,
+      "loss": 0.0391,
+      "step": 2650
+    },
+    {
+      "action_loss": 0.015594280324876308,
+      "epoch": 2.383093525179856,
+      "step": 2650
+    },
+    {
+      "epoch": 2.383093525179856,
+      "step": 2650,
+      "torque_loss": 0.18892382085323334
+    },
+    {
+      "epoch": 2.383093525179856,
+      "force_loss": 0.010612313635647297,
+      "step": 2650
+    },
+    {
+      "epoch": 2.3920863309352516,
+      "grad_norm": 1.2399665117263794,
+      "learning_rate": 8.863333333333334e-05,
+      "loss": 0.0405,
+      "step": 2660
+    },
+    {
+      "action_loss": 0.06340255588293076,
+      "epoch": 2.3920863309352516,
+      "step": 2660
+    },
+    {
+      "epoch": 2.3920863309352516,
+      "step": 2660,
+      "torque_loss": 0.181207537651062
+    },
+    {
+      "epoch": 2.3920863309352516,
+      "force_loss": 0.05579102039337158,
+      "step": 2660
+    },
+    {
+      "epoch": 2.4010791366906474,
+      "grad_norm": 0.8568428158760071,
+      "learning_rate": 8.896666666666667e-05,
+      "loss": 0.0457,
+      "step": 2670
+    },
+    {
+      "action_loss": 0.019339805468916893,
+      "epoch": 2.4010791366906474,
+      "step": 2670
+    },
+    {
+      "epoch": 2.4010791366906474,
+      "step": 2670,
+      "torque_loss": 0.13437801599502563
+    },
+    {
+      "epoch": 2.4010791366906474,
+      "force_loss": 0.008796132169663906,
+      "step": 2670
+    },
+    {
+      "epoch": 2.4100719424460433,
+      "grad_norm": 0.705787718296051,
+      "learning_rate": 8.93e-05,
+      "loss": 0.0337,
+      "step": 2680
+    },
+    {
+      "action_loss": 0.016918500885367393,
+      "epoch": 2.4100719424460433,
+      "step": 2680
+    },
+    {
+      "epoch": 2.4100719424460433,
+      "step": 2680,
+      "torque_loss": 0.09708541631698608
+    },
+    {
+      "epoch": 2.4100719424460433,
+      "force_loss": 0.009710412472486496,
+      "step": 2680
+    },
+    {
+      "epoch": 2.4190647482014387,
+      "grad_norm": 0.9660235047340393,
+      "learning_rate": 8.963333333333333e-05,
+      "loss": 0.0326,
+      "step": 2690
+    },
+    {
+      "action_loss": 0.01665041781961918,
+      "epoch": 2.4190647482014387,
+      "step": 2690
+    },
+    {
+      "epoch": 2.4190647482014387,
+      "step": 2690,
+      "torque_loss": 0.18289561569690704
+    },
+    {
+      "epoch": 2.4190647482014387,
+      "force_loss": 0.012846916913986206,
+      "step": 2690
+    },
+    {
+      "epoch": 2.4280575539568345,
+      "grad_norm": 1.450756549835205,
+      "learning_rate": 8.996666666666667e-05,
+      "loss": 0.0353,
+      "step": 2700
+    },
+    {
+      "action_loss": 0.012606129981577396,
+      "epoch": 2.4280575539568345,
+      "step": 2700
+    },
+    {
+      "epoch": 2.4280575539568345,
+      "step": 2700,
+      "torque_loss": 0.10766936093568802
+    },
+    {
+      "epoch": 2.4280575539568345,
+      "force_loss": 0.0076093655079603195,
+      "step": 2700
+    },
+    {
+      "epoch": 2.4370503597122304,
+      "grad_norm": 1.1657781600952148,
+      "learning_rate": 9.030000000000001e-05,
+      "loss": 0.0362,
+      "step": 2710
+    },
+    {
+      "action_loss": 0.02287411130964756,
+      "epoch": 2.4370503597122304,
+      "step": 2710
+    },
+    {
+      "epoch": 2.4370503597122304,
+      "step": 2710,
+      "torque_loss": 0.122210793197155
+    },
+    {
+      "epoch": 2.4370503597122304,
+      "force_loss": 0.015399813652038574,
+      "step": 2710
+    },
+    {
+      "epoch": 2.446043165467626,
+      "grad_norm": 1.0665206909179688,
+      "learning_rate": 9.063333333333333e-05,
+      "loss": 0.0421,
+      "step": 2720
+    },
+    {
+      "action_loss": 0.009829004295170307,
+      "epoch": 2.446043165467626,
+      "step": 2720
+    },
+    {
+      "epoch": 2.446043165467626,
+      "step": 2720,
+      "torque_loss": 0.11033079028129578
+    },
+    {
+      "epoch": 2.446043165467626,
+      "force_loss": 0.007507554721087217,
+      "step": 2720
+    },
+    {
+      "epoch": 2.4550359712230216,
+      "grad_norm": 0.8438506126403809,
+      "learning_rate": 9.096666666666666e-05,
+      "loss": 0.0314,
+      "step": 2730
+    },
+    {
+      "action_loss": 0.01683415099978447,
+      "epoch": 2.4550359712230216,
+      "step": 2730
+    },
+    {
+      "epoch": 2.4550359712230216,
+      "step": 2730,
+      "torque_loss": 0.15783236920833588
+    },
+    {
+      "epoch": 2.4550359712230216,
+      "force_loss": 0.024075910449028015,
+      "step": 2730
+    },
+    {
+      "epoch": 2.4640287769784175,
+      "grad_norm": 2.404555082321167,
+      "learning_rate": 9.130000000000001e-05,
+      "loss": 0.0345,
+      "step": 2740
+    },
+    {
+      "action_loss": 0.02086368203163147,
+      "epoch": 2.4640287769784175,
+      "step": 2740
+    },
+    {
+      "epoch": 2.4640287769784175,
+      "step": 2740,
+      "torque_loss": 0.16255338490009308
+    },
+    {
+      "epoch": 2.4640287769784175,
+      "force_loss": 0.011286056600511074,
+      "step": 2740
+    },
+    {
+      "epoch": 2.473021582733813,
+      "grad_norm": 1.7568706274032593,
+      "learning_rate": 9.163333333333334e-05,
+      "loss": 0.0456,
+      "step": 2750
+    },
+    {
+      "action_loss": 0.026560992002487183,
+      "epoch": 2.473021582733813,
+      "step": 2750
+    },
+    {
+      "epoch": 2.473021582733813,
+      "step": 2750,
+      "torque_loss": 0.15280811488628387
+    },
+    {
+      "epoch": 2.473021582733813,
+      "force_loss": 0.01929291896522045,
+      "step": 2750
+    },
+    {
+      "epoch": 2.4820143884892087,
+      "grad_norm": 0.8318678736686707,
+      "learning_rate": 9.196666666666666e-05,
+      "loss": 0.0404,
+      "step": 2760
+    },
+    {
+      "action_loss": 0.01715129427611828,
+      "epoch": 2.4820143884892087,
+      "step": 2760
+    },
+    {
+      "epoch": 2.4820143884892087,
+      "step": 2760,
+      "torque_loss": 0.12959398329257965
+    },
+    {
+      "epoch": 2.4820143884892087,
+      "force_loss": 0.011284488253295422,
+      "step": 2760
+    },
+    {
+      "epoch": 2.491007194244604,
+      "grad_norm": 1.5090669393539429,
+      "learning_rate": 9.230000000000001e-05,
+      "loss": 0.0334,
+      "step": 2770
+    },
+    {
+      "action_loss": 0.03781120851635933,
+      "epoch": 2.491007194244604,
+      "step": 2770
+    },
+    {
+      "epoch": 2.491007194244604,
+      "step": 2770,
+      "torque_loss": 0.14982609450817108
+    },
+    {
+      "epoch": 2.491007194244604,
+      "force_loss": 0.01831093803048134,
+      "step": 2770
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 2.4214015007019043,
+      "learning_rate": 9.263333333333334e-05,
+      "loss": 0.0355,
+      "step": 2780
+    },
+    {
+      "action_loss": 0.02332303486764431,
+      "epoch": 2.5,
+      "step": 2780
+    },
+    {
+      "epoch": 2.5,
+      "step": 2780,
+      "torque_loss": 0.11137477308511734
+    },
+    {
+      "epoch": 2.5,
+      "force_loss": 0.013371962122619152,
+      "step": 2780
+    },
+    {
+      "epoch": 2.508992805755396,
+      "grad_norm": 1.7286295890808105,
+      "learning_rate": 9.296666666666667e-05,
+      "loss": 0.0421,
+      "step": 2790
+    },
+    {
+      "action_loss": 0.01142234355211258,
+      "epoch": 2.508992805755396,
+      "step": 2790
+    },
+    {
+      "epoch": 2.508992805755396,
+      "step": 2790,
+      "torque_loss": 0.09397289156913757
+    },
+    {
+      "epoch": 2.508992805755396,
+      "force_loss": 0.009112835861742496,
+      "step": 2790
+    },
+    {
+      "epoch": 2.5179856115107913,
+      "grad_norm": 0.696000337600708,
+      "learning_rate": 9.33e-05,
+      "loss": 0.0371,
+      "step": 2800
+    },
+    {
+      "action_loss": 0.02488979883491993,
+      "epoch": 2.5179856115107913,
+      "step": 2800
+    },
+    {
+      "epoch": 2.5179856115107913,
+      "step": 2800,
+      "torque_loss": 0.17817990481853485
+    },
+    {
+      "epoch": 2.5179856115107913,
+      "force_loss": 0.02367742918431759,
+      "step": 2800
+    },
+    {
+      "epoch": 2.526978417266187,
+      "grad_norm": 2.2079663276672363,
+      "learning_rate": 9.363333333333334e-05,
+      "loss": 0.0344,
+      "step": 2810
+    },
+    {
+      "action_loss": 0.016026629135012627,
+      "epoch": 2.526978417266187,
+      "step": 2810
+    },
+    {
+      "epoch": 2.526978417266187,
+      "step": 2810,
+      "torque_loss": 0.13567723333835602
+    },
+    {
+      "epoch": 2.526978417266187,
+      "force_loss": 0.01275847852230072,
+      "step": 2810
+    },
+    {
+      "epoch": 2.5359712230215825,
+      "grad_norm": 1.2285265922546387,
+      "learning_rate": 9.396666666666667e-05,
+      "loss": 0.0332,
+      "step": 2820
+    },
+    {
+      "action_loss": 0.014813491143286228,
+      "epoch": 2.5359712230215825,
+      "step": 2820
+    },
+    {
+      "epoch": 2.5359712230215825,
+      "step": 2820,
+      "torque_loss": 0.12180599570274353
+    },
+    {
+      "epoch": 2.5359712230215825,
+      "force_loss": 0.010316254571080208,
+      "step": 2820
+    },
+    {
+      "epoch": 2.5449640287769784,
+      "grad_norm": 1.1561086177825928,
+      "learning_rate": 9.43e-05,
+      "loss": 0.0453,
+      "step": 2830
+    },
+    {
+      "action_loss": 0.025606466457247734,
+      "epoch": 2.5449640287769784,
+      "step": 2830
+    },
+    {
+      "epoch": 2.5449640287769784,
+      "step": 2830,
+      "torque_loss": 0.11959757655858994
+    },
+    {
+      "epoch": 2.5449640287769784,
+      "force_loss": 0.02054104395210743,
+      "step": 2830
+    },
+    {
+      "epoch": 2.553956834532374,
+      "grad_norm": 0.8689306378364563,
+      "learning_rate": 9.463333333333333e-05,
+      "loss": 0.0383,
+      "step": 2840
+    },
+    {
+      "action_loss": 0.01377404760569334,
+      "epoch": 2.553956834532374,
+      "step": 2840
+    },
+    {
+      "epoch": 2.553956834532374,
+      "step": 2840,
+      "torque_loss": 0.13466019928455353
+    },
+    {
+      "epoch": 2.553956834532374,
+      "force_loss": 0.010488275438547134,
+      "step": 2840
+    },
+    {
+      "epoch": 2.56294964028777,
+      "grad_norm": 0.4736124873161316,
+      "learning_rate": 9.496666666666667e-05,
+      "loss": 0.0377,
+      "step": 2850
+    },
+    {
+      "action_loss": 0.019194582477211952,
+      "epoch": 2.56294964028777,
+      "step": 2850
+    },
+    {
+      "epoch": 2.56294964028777,
+      "step": 2850,
+      "torque_loss": 0.1544509381055832
+    },
+    {
+      "epoch": 2.56294964028777,
+      "force_loss": 0.011677307076752186,
+      "step": 2850
+    },
+    {
+      "epoch": 2.5719424460431655,
+      "grad_norm": 1.1699731349945068,
+      "learning_rate": 9.53e-05,
+      "loss": 0.0431,
+      "step": 2860
+    },
+    {
+      "action_loss": 0.015994906425476074,
+      "epoch": 2.5719424460431655,
+      "step": 2860
+    },
+    {
+      "epoch": 2.5719424460431655,
+      "step": 2860,
+      "torque_loss": 0.13511356711387634
+    },
+    {
+      "epoch": 2.5719424460431655,
+      "force_loss": 0.011160057969391346,
+      "step": 2860
+    },
+    {
+      "epoch": 2.5809352517985613,
+      "grad_norm": 2.9630091190338135,
+      "learning_rate": 9.563333333333334e-05,
+      "loss": 0.0434,
+      "step": 2870
+    },
+    {
+      "action_loss": 0.02055782824754715,
+      "epoch": 2.5809352517985613,
+      "step": 2870
+    },
+    {
+      "epoch": 2.5809352517985613,
+      "step": 2870,
+      "torque_loss": 0.16927486658096313
+    },
+    {
+      "epoch": 2.5809352517985613,
+      "force_loss": 0.012406043708324432,
+      "step": 2870
+    },
+    {
+      "epoch": 2.5899280575539567,
+      "grad_norm": 0.8095554113388062,
+      "learning_rate": 9.596666666666668e-05,
+      "loss": 0.0338,
+      "step": 2880
+    },
+    {
+      "action_loss": 0.022499138489365578,
+      "epoch": 2.5899280575539567,
+      "step": 2880
+    },
+    {
+      "epoch": 2.5899280575539567,
+      "step": 2880,
+      "torque_loss": 0.12810824811458588
+    },
+    {
+      "epoch": 2.5899280575539567,
+      "force_loss": 0.016522439196705818,
+      "step": 2880
+    },
+    {
+      "epoch": 2.5989208633093526,
+      "grad_norm": 1.1473280191421509,
+      "learning_rate": 9.63e-05,
+      "loss": 0.0396,
+      "step": 2890
+    },
+    {
+      "action_loss": 0.026943305507302284,
+      "epoch": 2.5989208633093526,
+      "step": 2890
+    },
+    {
+      "epoch": 2.5989208633093526,
+      "step": 2890,
+      "torque_loss": 0.1600051373243332
+    },
+    {
+      "epoch": 2.5989208633093526,
+      "force_loss": 0.016423838213086128,
+      "step": 2890
+    },
+    {
+      "epoch": 2.6079136690647484,
+      "grad_norm": 1.2999749183654785,
+      "learning_rate": 9.663333333333334e-05,
+      "loss": 0.0442,
+      "step": 2900
+    },
+    {
+      "action_loss": 0.017461325973272324,
+      "epoch": 2.6079136690647484,
+      "step": 2900
+    },
+    {
+      "epoch": 2.6079136690647484,
+      "step": 2900,
+      "torque_loss": 0.11378884315490723
+    },
+    {
+      "epoch": 2.6079136690647484,
+      "force_loss": 0.010942693799734116,
+      "step": 2900
+    },
+    {
+      "epoch": 2.616906474820144,
+      "grad_norm": 0.8496351838111877,
+      "learning_rate": 9.696666666666667e-05,
+      "loss": 0.0454,
+      "step": 2910
+    },
+    {
+      "action_loss": 0.018060246482491493,
+      "epoch": 2.616906474820144,
+      "step": 2910
+    },
+    {
+      "epoch": 2.616906474820144,
+      "step": 2910,
+      "torque_loss": 0.16233943402767181
+    },
+    {
+      "epoch": 2.616906474820144,
+      "force_loss": 0.013489101082086563,
+      "step": 2910
+    },
+    {
+      "epoch": 2.6258992805755397,
+      "grad_norm": 1.1536489725112915,
+      "learning_rate": 9.730000000000001e-05,
+      "loss": 0.0407,
+      "step": 2920
+    },
+    {
+      "action_loss": 0.0315416194498539,
+      "epoch": 2.6258992805755397,
+      "step": 2920
+    },
+    {
+      "epoch": 2.6258992805755397,
+      "step": 2920,
+      "torque_loss": 0.12135192006826401
+    },
+    {
+      "epoch": 2.6258992805755397,
+      "force_loss": 0.018383117392659187,
+      "step": 2920
+    },
+    {
+      "epoch": 2.634892086330935,
+      "grad_norm": 0.6894007325172424,
+      "learning_rate": 9.763333333333334e-05,
+      "loss": 0.0374,
+      "step": 2930
+    },
+    {
+      "action_loss": 0.022216318175196648,
+      "epoch": 2.634892086330935,
+      "step": 2930
+    },
+    {
+      "epoch": 2.634892086330935,
+      "step": 2930,
+      "torque_loss": 0.15010647475719452
+    },
+    {
+      "epoch": 2.634892086330935,
+      "force_loss": 0.018463822081685066,
+      "step": 2930
+    },
+    {
+      "epoch": 2.643884892086331,
+      "grad_norm": 2.491455554962158,
+      "learning_rate": 9.796666666666667e-05,
+      "loss": 0.0439,
+      "step": 2940
+    },
+    {
+      "action_loss": 0.039160724729299545,
+      "epoch": 2.643884892086331,
+      "step": 2940
+    },
+    {
+      "epoch": 2.643884892086331,
+      "step": 2940,
+      "torque_loss": 0.2687141001224518
+    },
+    {
+      "epoch": 2.643884892086331,
+      "force_loss": 0.024796312674880028,
+      "step": 2940
+    },
+    {
+      "epoch": 2.652877697841727,
+      "grad_norm": 1.2095692157745361,
+      "learning_rate": 9.83e-05,
+      "loss": 0.0505,
+      "step": 2950
+    },
+    {
+      "action_loss": 0.02901189960539341,
+      "epoch": 2.652877697841727,
+      "step": 2950
+    },
+    {
+      "epoch": 2.652877697841727,
+      "step": 2950,
+      "torque_loss": 0.18363797664642334
+    },
+    {
+      "epoch": 2.652877697841727,
+      "force_loss": 0.014832672663033009,
+      "step": 2950
+    },
+    {
+      "epoch": 2.661870503597122,
+      "grad_norm": 1.1525763273239136,
+      "learning_rate": 9.863333333333334e-05,
+      "loss": 0.0428,
+      "step": 2960
+    },
+    {
+      "action_loss": 0.01046085823327303,
+      "epoch": 2.661870503597122,
+      "step": 2960
+    },
+    {
+      "epoch": 2.661870503597122,
+      "step": 2960,
+      "torque_loss": 0.11623861640691757
+    },
+    {
+      "epoch": 2.661870503597122,
+      "force_loss": 0.007993306033313274,
+      "step": 2960
+    },
+    {
+      "epoch": 2.670863309352518,
+      "grad_norm": 2.065459728240967,
+      "learning_rate": 9.896666666666667e-05,
+      "loss": 0.0374,
+      "step": 2970
+    },
+    {
+      "action_loss": 0.02344103343784809,
+      "epoch": 2.670863309352518,
+      "step": 2970
+    },
+    {
+      "epoch": 2.670863309352518,
+      "step": 2970,
+      "torque_loss": 0.15691407024860382
+    },
+    {
+      "epoch": 2.670863309352518,
+      "force_loss": 0.015483230352401733,
+      "step": 2970
+    },
+    {
+      "epoch": 2.6798561151079134,
+      "grad_norm": 1.7671762704849243,
+      "learning_rate": 9.93e-05,
+      "loss": 0.0386,
+      "step": 2980
+    },
+    {
+      "action_loss": 0.020954905077815056,
+      "epoch": 2.6798561151079134,
+      "step": 2980
+    },
+    {
+      "epoch": 2.6798561151079134,
+      "step": 2980,
+      "torque_loss": 0.12083371728658676
+    },
+    {
+      "epoch": 2.6798561151079134,
+      "force_loss": 0.009894735179841518,
+      "step": 2980
+    },
+    {
+      "epoch": 2.6888489208633093,
+      "grad_norm": 1.8522980213165283,
+      "learning_rate": 9.963333333333333e-05,
+      "loss": 0.0451,
+      "step": 2990
+    },
+    {
+      "action_loss": 0.02034606784582138,
+      "epoch": 2.6888489208633093,
+      "step": 2990
+    },
+    {
+      "epoch": 2.6888489208633093,
+      "step": 2990,
+      "torque_loss": 0.19639647006988525
+    },
+    {
+      "epoch": 2.6888489208633093,
+      "force_loss": 0.016407953575253487,
+      "step": 2990
+    },
+    {
+      "epoch": 2.697841726618705,
+      "grad_norm": 0.999684751033783,
+      "learning_rate": 9.996666666666668e-05,
+      "loss": 0.051,
+      "step": 3000
+    },
+    {
+      "action_loss": 0.018691303208470345,
+      "epoch": 2.697841726618705,
+      "step": 3000
+    },
+    {
+      "epoch": 2.697841726618705,
+      "step": 3000,
+      "torque_loss": 0.10870843380689621
+    },
+    {
+      "epoch": 2.697841726618705,
+      "force_loss": 0.01539594680070877,
+      "step": 3000
+    },
+    {
+      "epoch": 2.706834532374101,
+      "grad_norm": 0.994063675403595,
+      "learning_rate": 9.999999384858465e-05,
+      "loss": 0.038,
+      "step": 3010
+    },
+    {
+      "action_loss": 0.016244731843471527,
+      "epoch": 2.706834532374101,
+      "step": 3010
+    },
+    {
+      "epoch": 2.706834532374101,
+      "step": 3010,
+      "torque_loss": 0.10822829604148865
+    },
+    {
+      "epoch": 2.706834532374101,
+      "force_loss": 0.011545299552381039,
+      "step": 3010
+    },
+    {
+      "epoch": 2.7158273381294964,
+      "grad_norm": 1.5519417524337769,
+      "learning_rate": 9.999997258443473e-05,
+      "loss": 0.0401,
+      "step": 3020
+    },
+    {
+      "action_loss": 0.033651623874902725,
+      "epoch": 2.7158273381294964,
+      "step": 3020
+    },
+    {
+      "epoch": 2.7158273381294964,
+      "step": 3020,
+      "torque_loss": 0.16406987607479095
+    },
+    {
+      "epoch": 2.7158273381294964,
+      "force_loss": 0.02589082531630993,
+      "step": 3020
+    },
+    {
+      "epoch": 2.7248201438848922,
+      "grad_norm": 1.5173958539962769,
+      "learning_rate": 9.999993613161331e-05,
+      "loss": 0.042,
+      "step": 3030
+    },
+    {
+      "action_loss": 0.03164878487586975,
+      "epoch": 2.7248201438848922,
+      "step": 3030
+    },
+    {
+      "epoch": 2.7248201438848922,
+      "step": 3030,
+      "torque_loss": 0.15949127078056335
+    },
+    {
+      "epoch": 2.7248201438848922,
+      "force_loss": 0.021960146725177765,
+      "step": 3030
+    },
+    {
+      "epoch": 2.7338129496402876,
+      "grad_norm": 1.3673168420791626,
+      "learning_rate": 9.999988449013146e-05,
+      "loss": 0.0433,
+      "step": 3040
+    },
+    {
+      "action_loss": 0.026378417387604713,
+      "epoch": 2.7338129496402876,
+      "step": 3040
+    },
+    {
+      "epoch": 2.7338129496402876,
+      "step": 3040,
+      "torque_loss": 0.16195188462734222
+    },
+    {
+      "epoch": 2.7338129496402876,
+      "force_loss": 0.01337435096502304,
+      "step": 3040
+    },
+    {
+      "epoch": 2.7428057553956835,
+      "grad_norm": 0.9938601851463318,
+      "learning_rate": 9.99998176600049e-05,
+      "loss": 0.0489,
+      "step": 3050
+    },
+    {
+      "action_loss": 0.011601743288338184,
+      "epoch": 2.7428057553956835,
+      "step": 3050
+    },
+    {
+      "epoch": 2.7428057553956835,
+      "step": 3050,
+      "torque_loss": 0.15897555649280548
+    },
+    {
+      "epoch": 2.7428057553956835,
+      "force_loss": 0.01026233658194542,
+      "step": 3050
+    },
+    {
+      "epoch": 2.7517985611510793,
+      "grad_norm": 1.2809144258499146,
+      "learning_rate": 9.999973564125389e-05,
+      "loss": 0.0361,
+      "step": 3060
+    },
+    {
+      "action_loss": 0.01920217089354992,
+      "epoch": 2.7517985611510793,
+      "step": 3060
+    },
+    {
+      "epoch": 2.7517985611510793,
+      "step": 3060,
+      "torque_loss": 0.13005442917346954
+    },
+    {
+      "epoch": 2.7517985611510793,
+      "force_loss": 0.012239661999046803,
+      "step": 3060
+    },
+    {
+      "epoch": 2.7607913669064748,
+      "grad_norm": 1.192671298980713,
+      "learning_rate": 9.999963843390335e-05,
+      "loss": 0.0373,
+      "step": 3070
+    },
+    {
+      "action_loss": 0.04148460552096367,
+      "epoch": 2.7607913669064748,
+      "step": 3070
+    },
+    {
+      "epoch": 2.7607913669064748,
+      "step": 3070,
+      "torque_loss": 0.1638404279947281
+    },
+    {
+      "epoch": 2.7607913669064748,
+      "force_loss": 0.025303175672888756,
+      "step": 3070
+    },
+    {
+      "epoch": 2.7697841726618706,
+      "grad_norm": 2.369666814804077,
+      "learning_rate": 9.999952603798282e-05,
+      "loss": 0.0562,
+      "step": 3080
+    },
+    {
+      "action_loss": 0.047801535576581955,
+      "epoch": 2.7697841726618706,
+      "step": 3080
+    },
+    {
+      "epoch": 2.7697841726618706,
+      "step": 3080,
+      "torque_loss": 0.20009560883045197
+    },
+    {
+      "epoch": 2.7697841726618706,
+      "force_loss": 0.06334620714187622,
+      "step": 3080
+    },
+    {
+      "epoch": 2.778776978417266,
+      "grad_norm": 1.7056851387023926,
+      "learning_rate": 9.999939845352646e-05,
+      "loss": 0.0467,
+      "step": 3090
+    },
+    {
+      "action_loss": 0.024913745000958443,
+      "epoch": 2.778776978417266,
+      "step": 3090
+    },
+    {
+      "epoch": 2.778776978417266,
+      "step": 3090,
+      "torque_loss": 0.19030427932739258
+    },
+    {
+      "epoch": 2.778776978417266,
+      "force_loss": 0.014128959737718105,
+      "step": 3090
+    },
+    {
+      "epoch": 2.787769784172662,
+      "grad_norm": 1.7222480773925781,
+      "learning_rate": 9.999925568057298e-05,
+      "loss": 0.0455,
+      "step": 3100
+    },
+    {
+      "action_loss": 0.04152792692184448,
+      "epoch": 2.787769784172662,
+      "step": 3100
+    },
+    {
+      "epoch": 2.787769784172662,
+      "step": 3100,
+      "torque_loss": 0.15055231750011444
+    },
+    {
+      "epoch": 2.787769784172662,
+      "force_loss": 0.01842602528631687,
+      "step": 3100
+    },
+    {
+      "epoch": 2.7967625899280577,
+      "grad_norm": 0.839058518409729,
+      "learning_rate": 9.999909771916578e-05,
+      "loss": 0.0426,
+      "step": 3110
+    },
+    {
+      "action_loss": 0.03791915252804756,
+      "epoch": 2.7967625899280577,
+      "step": 3110
+    },
+    {
+      "epoch": 2.7967625899280577,
+      "step": 3110,
+      "torque_loss": 0.13088136911392212
+    },
+    {
+      "epoch": 2.7967625899280577,
+      "force_loss": 0.019136101007461548,
+      "step": 3110
+    },
+    {
+      "epoch": 2.805755395683453,
+      "grad_norm": 1.6002094745635986,
+      "learning_rate": 9.999892456935285e-05,
+      "loss": 0.0456,
+      "step": 3120
+    },
+    {
+      "action_loss": 0.05152573063969612,
+      "epoch": 2.805755395683453,
+      "step": 3120
+    },
+    {
+      "epoch": 2.805755395683453,
+      "step": 3120,
+      "torque_loss": 0.1504654586315155
+    },
+    {
+      "epoch": 2.805755395683453,
+      "force_loss": 0.03299170359969139,
+      "step": 3120
+    },
+    {
+      "epoch": 2.814748201438849,
+      "grad_norm": 0.798694908618927,
+      "learning_rate": 9.999873623118679e-05,
+      "loss": 0.0433,
+      "step": 3130
+    },
+    {
+      "action_loss": 0.013080894947052002,
+      "epoch": 2.814748201438849,
+      "step": 3130
+    },
+    {
+      "epoch": 2.814748201438849,
+      "step": 3130,
+      "torque_loss": 0.16791932284832
+    },
+    {
+      "epoch": 2.814748201438849,
+      "force_loss": 0.012379822321236134,
+      "step": 3130
+    },
+    {
+      "epoch": 2.8237410071942444,
+      "grad_norm": 0.9472343921661377,
+      "learning_rate": 9.999853270472479e-05,
+      "loss": 0.0408,
+      "step": 3140
+    },
+    {
+      "action_loss": 0.02110356092453003,
+      "epoch": 2.8237410071942444,
+      "step": 3140
+    },
+    {
+      "epoch": 2.8237410071942444,
+      "step": 3140,
+      "torque_loss": 0.15463107824325562
+    },
+    {
+      "epoch": 2.8237410071942444,
+      "force_loss": 0.02173072099685669,
+      "step": 3140
+    },
+    {
+      "epoch": 2.83273381294964,
+      "grad_norm": 1.3234020471572876,
+      "learning_rate": 9.999831399002871e-05,
+      "loss": 0.0421,
+      "step": 3150
+    },
+    {
+      "action_loss": 0.01979002356529236,
+      "epoch": 2.83273381294964,
+      "step": 3150
+    },
+    {
+      "epoch": 2.83273381294964,
+      "step": 3150,
+      "torque_loss": 0.12477517127990723
+    },
+    {
+      "epoch": 2.83273381294964,
+      "force_loss": 0.012610559351742268,
+      "step": 3150
+    },
+    {
+      "epoch": 2.841726618705036,
+      "grad_norm": 1.0811312198638916,
+      "learning_rate": 9.999808008716494e-05,
+      "loss": 0.0384,
+      "step": 3160
+    },
+    {
+      "action_loss": 0.03057975135743618,
+      "epoch": 2.841726618705036,
+      "step": 3160
+    },
+    {
+      "epoch": 2.841726618705036,
+      "step": 3160,
+      "torque_loss": 0.2436930537223816
+    },
+    {
+      "epoch": 2.841726618705036,
+      "force_loss": 0.014245248399674892,
+      "step": 3160
+    },
+    {
+      "epoch": 2.850719424460432,
+      "grad_norm": 0.8396542072296143,
+      "learning_rate": 9.999783099620459e-05,
+      "loss": 0.0432,
+      "step": 3170
+    },
+    {
+      "action_loss": 0.013761981390416622,
+      "epoch": 2.850719424460432,
+      "step": 3170
+    },
+    {
+      "epoch": 2.850719424460432,
+      "step": 3170,
+      "torque_loss": 0.12116435915231705
+    },
+    {
+      "epoch": 2.850719424460432,
+      "force_loss": 0.010504312813282013,
+      "step": 3170
+    },
+    {
+      "epoch": 2.8597122302158273,
+      "grad_norm": 0.9482388496398926,
+      "learning_rate": 9.999756671722328e-05,
+      "loss": 0.0393,
+      "step": 3180
+    },
+    {
+      "action_loss": 0.012676995247602463,
+      "epoch": 2.8597122302158273,
+      "step": 3180
+    },
+    {
+      "epoch": 2.8597122302158273,
+      "step": 3180,
+      "torque_loss": 0.1399737149477005
+    },
+    {
+      "epoch": 2.8597122302158273,
+      "force_loss": 0.009577300399541855,
+      "step": 3180
+    },
+    {
+      "epoch": 2.868705035971223,
+      "grad_norm": 2.415886878967285,
+      "learning_rate": 9.99972872503013e-05,
+      "loss": 0.0388,
+      "step": 3190
+    },
+    {
+      "action_loss": 0.029606467112898827,
+      "epoch": 2.868705035971223,
+      "step": 3190
+    },
+    {
+      "epoch": 2.868705035971223,
+      "step": 3190,
+      "torque_loss": 0.15825851261615753
+    },
+    {
+      "epoch": 2.868705035971223,
+      "force_loss": 0.018408508971333504,
+      "step": 3190
+    },
+    {
+      "epoch": 2.8776978417266186,
+      "grad_norm": 0.8147604465484619,
+      "learning_rate": 9.999699259552359e-05,
+      "loss": 0.0422,
+      "step": 3200
+    },
+    {
+      "action_loss": 0.010048252530395985,
+      "epoch": 2.8776978417266186,
+      "step": 3200
+    },
+    {
+      "epoch": 2.8776978417266186,
+      "step": 3200,
+      "torque_loss": 0.12385464459657669
+    },
+    {
+      "epoch": 2.8776978417266186,
+      "force_loss": 0.0070101716555655,
+      "step": 3200
+    },
+    {
+      "epoch": 2.8866906474820144,
+      "grad_norm": 0.8463675379753113,
+      "learning_rate": 9.99966827529796e-05,
+      "loss": 0.0362,
+      "step": 3210
+    },
+    {
+      "action_loss": 0.020685601979494095,
+      "epoch": 2.8866906474820144,
+      "step": 3210
+    },
+    {
+      "epoch": 2.8866906474820144,
+      "step": 3210,
+      "torque_loss": 0.18193601071834564
+    },
+    {
+      "epoch": 2.8866906474820144,
+      "force_loss": 0.012167420238256454,
+      "step": 3210
+    },
+    {
+      "epoch": 2.8956834532374103,
+      "grad_norm": 1.0764975547790527,
+      "learning_rate": 9.999635772276348e-05,
+      "loss": 0.0377,
+      "step": 3220
+    },
+    {
+      "action_loss": 0.010956761427223682,
+      "epoch": 2.8956834532374103,
+      "step": 3220
+    },
+    {
+      "epoch": 2.8956834532374103,
+      "step": 3220,
+      "torque_loss": 0.12044543772935867
+    },
+    {
+      "epoch": 2.8956834532374103,
+      "force_loss": 0.0089794360101223,
+      "step": 3220
+    },
+    {
+      "epoch": 2.9046762589928057,
+      "grad_norm": 0.772395670413971,
+      "learning_rate": 9.999601750497396e-05,
+      "loss": 0.0387,
+      "step": 3230
+    },
+    {
+      "action_loss": 0.017194250598549843,
+      "epoch": 2.9046762589928057,
+      "step": 3230
+    },
+    {
+      "epoch": 2.9046762589928057,
+      "step": 3230,
+      "torque_loss": 0.16221556067466736
+    },
+    {
+      "epoch": 2.9046762589928057,
+      "force_loss": 0.013018782250583172,
+      "step": 3230
+    },
+    {
+      "epoch": 2.9136690647482015,
+      "grad_norm": 0.7906728386878967,
+      "learning_rate": 9.99956620997144e-05,
+      "loss": 0.0349,
+      "step": 3240
+    },
+    {
+      "action_loss": 0.0322260856628418,
+      "epoch": 2.9136690647482015,
+      "step": 3240
+    },
+    {
+      "epoch": 2.9136690647482015,
+      "step": 3240,
+      "torque_loss": 0.14162257313728333
+    },
+    {
+      "epoch": 2.9136690647482015,
+      "force_loss": 0.017613641917705536,
+      "step": 3240
+    },
+    {
+      "epoch": 2.922661870503597,
+      "grad_norm": 1.1173397302627563,
+      "learning_rate": 9.999529150709275e-05,
+      "loss": 0.0415,
+      "step": 3250
+    },
+    {
+      "action_loss": 0.020265748724341393,
+      "epoch": 2.922661870503597,
+      "step": 3250
+    },
+    {
+      "epoch": 2.922661870503597,
+      "step": 3250,
+      "torque_loss": 0.1406986564397812
+    },
+    {
+      "epoch": 2.922661870503597,
+      "force_loss": 0.009460187517106533,
+      "step": 3250
+    },
+    {
+      "epoch": 2.931654676258993,
+      "grad_norm": 0.6254609823226929,
+      "learning_rate": 9.999490572722158e-05,
+      "loss": 0.0372,
+      "step": 3260
+    },
+    {
+      "action_loss": 0.025432869791984558,
+      "epoch": 2.931654676258993,
+      "step": 3260
+    },
+    {
+      "epoch": 2.931654676258993,
+      "step": 3260,
+      "torque_loss": 0.14959515631198883
+    },
+    {
+      "epoch": 2.931654676258993,
+      "force_loss": 0.014176805503666401,
+      "step": 3260
+    },
+    {
+      "epoch": 2.9406474820143886,
+      "grad_norm": 0.8070449233055115,
+      "learning_rate": 9.99945047602181e-05,
+      "loss": 0.0408,
+      "step": 3270
+    },
+    {
+      "action_loss": 0.01896510273218155,
+      "epoch": 2.9406474820143886,
+      "step": 3270
+    },
+    {
+      "epoch": 2.9406474820143886,
+      "step": 3270,
+      "torque_loss": 0.11664503812789917
+    },
+    {
+      "epoch": 2.9406474820143886,
+      "force_loss": 0.011725406162440777,
+      "step": 3270
+    },
+    {
+      "epoch": 2.949640287769784,
+      "grad_norm": 1.269662857055664,
+      "learning_rate": 9.99940886062041e-05,
+      "loss": 0.0389,
+      "step": 3280
+    },
+    {
+      "action_loss": 0.016527635976672173,
+      "epoch": 2.949640287769784,
+      "step": 3280
+    },
+    {
+      "epoch": 2.949640287769784,
+      "step": 3280,
+      "torque_loss": 0.12627343833446503
+    },
+    {
+      "epoch": 2.949640287769784,
+      "force_loss": 0.011279787868261337,
+      "step": 3280
+    },
+    {
+      "epoch": 2.95863309352518,
+      "grad_norm": 1.1411021947860718,
+      "learning_rate": 9.999365726530599e-05,
+      "loss": 0.04,
+      "step": 3290
+    },
+    {
+      "action_loss": 0.01918291114270687,
+      "epoch": 2.95863309352518,
+      "step": 3290
+    },
+    {
+      "epoch": 2.95863309352518,
+      "step": 3290,
+      "torque_loss": 0.16686545312404633
+    },
+    {
+      "epoch": 2.95863309352518,
+      "force_loss": 0.012200385332107544,
+      "step": 3290
+    },
+    {
+      "epoch": 2.9676258992805753,
+      "grad_norm": 0.8965969085693359,
+      "learning_rate": 9.999321073765481e-05,
+      "loss": 0.0462,
+      "step": 3300
+    },
+    {
+      "action_loss": 0.017377378419041634,
+      "epoch": 2.9676258992805753,
+      "step": 3300
+    },
+    {
+      "epoch": 2.9676258992805753,
+      "step": 3300,
+      "torque_loss": 0.1282060742378235
+    },
+    {
+      "epoch": 2.9676258992805753,
+      "force_loss": 0.017754575237631798,
+      "step": 3300
+    },
+    {
+      "epoch": 2.976618705035971,
+      "grad_norm": 0.48440292477607727,
+      "learning_rate": 9.99927490233862e-05,
+      "loss": 0.0331,
+      "step": 3310
+    },
+    {
+      "action_loss": 0.02285308577120304,
+      "epoch": 2.976618705035971,
+      "step": 3310
+    },
+    {
+      "epoch": 2.976618705035971,
+      "step": 3310,
+      "torque_loss": 0.22604024410247803
+    },
+    {
+      "epoch": 2.976618705035971,
+      "force_loss": 0.019879691302776337,
+      "step": 3310
+    },
+    {
+      "epoch": 2.985611510791367,
+      "grad_norm": 1.7107858657836914,
+      "learning_rate": 9.999227212264043e-05,
+      "loss": 0.0364,
+      "step": 3320
+    },
+    {
+      "action_loss": 0.01286298781633377,
+      "epoch": 2.985611510791367,
+      "step": 3320
+    },
+    {
+      "epoch": 2.985611510791367,
+      "step": 3320,
+      "torque_loss": 0.0966697558760643
+    },
+    {
+      "epoch": 2.985611510791367,
+      "force_loss": 0.01715177856385708,
+      "step": 3320
+    },
+    {
+      "epoch": 2.994604316546763,
+      "grad_norm": 0.9540174007415771,
+      "learning_rate": 9.999178003556236e-05,
+      "loss": 0.0386,
+      "step": 3330
+    },
+    {
+      "action_loss": 0.020100312307476997,
+      "epoch": 2.994604316546763,
+      "step": 3330
+    },
+    {
+      "epoch": 2.994604316546763,
+      "step": 3330,
+      "torque_loss": 0.11698877811431885
+    },
+    {
+      "epoch": 2.994604316546763,
+      "force_loss": 0.019933760166168213,
+      "step": 3330
+    },
+    {
+      "epoch": 3.0035971223021583,
+      "grad_norm": 0.68222975730896,
+      "learning_rate": 9.999127276230146e-05,
+      "loss": 0.0321,
+      "step": 3340
+    },
+    {
+      "action_loss": 0.020617427304387093,
+      "epoch": 3.0035971223021583,
+      "step": 3340
+    },
+    {
+      "epoch": 3.0035971223021583,
+      "step": 3340,
+      "torque_loss": 0.12337452173233032
+    },
+    {
+      "epoch": 3.0035971223021583,
+      "force_loss": 0.01989404298365116,
+      "step": 3340
+    },
+    {
+      "epoch": 3.012589928057554,
+      "grad_norm": 2.0390818119049072,
+      "learning_rate": 9.999075030301184e-05,
+      "loss": 0.0391,
+      "step": 3350
+    },
+    {
+      "action_loss": 0.016074737533926964,
+      "epoch": 3.012589928057554,
+      "step": 3350
+    },
+    {
+      "epoch": 3.012589928057554,
+      "step": 3350,
+      "torque_loss": 0.13831709325313568
+    },
+    {
+      "epoch": 3.012589928057554,
+      "force_loss": 0.015261041931807995,
+      "step": 3350
+    },
+    {
+      "epoch": 3.0215827338129495,
+      "grad_norm": 0.5331881642341614,
+      "learning_rate": 9.999021265785221e-05,
+      "loss": 0.043,
+      "step": 3360
+    },
+    {
+      "action_loss": 0.020711304619908333,
+      "epoch": 3.0215827338129495,
+      "step": 3360
+    },
+    {
+      "epoch": 3.0215827338129495,
+      "step": 3360,
+      "torque_loss": 0.215559720993042
+    },
+    {
+      "epoch": 3.0215827338129495,
+      "force_loss": 0.01075026299804449,
+      "step": 3360
+    },
+    {
+      "epoch": 3.0305755395683454,
+      "grad_norm": 2.314973831176758,
+      "learning_rate": 9.998965982698589e-05,
+      "loss": 0.045,
+      "step": 3370
+    },
+    {
+      "action_loss": 0.02324887178838253,
+      "epoch": 3.0305755395683454,
+      "step": 3370
+    },
+    {
+      "epoch": 3.0305755395683454,
+      "step": 3370,
+      "torque_loss": 0.13435320556163788
+    },
+    {
+      "epoch": 3.0305755395683454,
+      "force_loss": 0.01448883581906557,
+      "step": 3370
+    },
+    {
+      "epoch": 3.039568345323741,
+      "grad_norm": 1.7778946161270142,
+      "learning_rate": 9.998909181058082e-05,
+      "loss": 0.0364,
+      "step": 3380
+    },
+    {
+      "action_loss": 0.023925187066197395,
+      "epoch": 3.039568345323741,
+      "step": 3380
+    },
+    {
+      "epoch": 3.039568345323741,
+      "step": 3380,
+      "torque_loss": 0.12586790323257446
+    },
+    {
+      "epoch": 3.039568345323741,
+      "force_loss": 0.026657452806830406,
+      "step": 3380
+    },
+    {
+      "epoch": 3.0485611510791366,
+      "grad_norm": 1.2528610229492188,
+      "learning_rate": 9.998850860880953e-05,
+      "loss": 0.0421,
+      "step": 3390
+    },
+    {
+      "action_loss": 0.02990814298391342,
+      "epoch": 3.0485611510791366,
+      "step": 3390
+    },
+    {
+      "epoch": 3.0485611510791366,
+      "step": 3390,
+      "torque_loss": 0.21017348766326904
+    },
+    {
+      "epoch": 3.0485611510791366,
+      "force_loss": 0.012373861856758595,
+      "step": 3390
+    },
+    {
+      "epoch": 3.0575539568345325,
+      "grad_norm": 1.3309173583984375,
+      "learning_rate": 9.998791022184922e-05,
+      "loss": 0.0418,
+      "step": 3400
+    },
+    {
+      "action_loss": 0.014552815817296505,
+      "epoch": 3.0575539568345325,
+      "step": 3400
+    },
+    {
+      "epoch": 3.0575539568345325,
+      "step": 3400,
+      "torque_loss": 0.11067936569452286
+    },
+    {
+      "epoch": 3.0575539568345325,
+      "force_loss": 0.009075048379600048,
+      "step": 3400
+    },
+    {
+      "epoch": 3.066546762589928,
+      "grad_norm": 2.535957098007202,
+      "learning_rate": 9.99872966498816e-05,
+      "loss": 0.0373,
+      "step": 3410
+    },
+    {
+      "action_loss": 0.03187643364071846,
+      "epoch": 3.066546762589928,
+      "step": 3410
+    },
+    {
+      "epoch": 3.066546762589928,
+      "step": 3410,
+      "torque_loss": 0.1743755340576172
+    },
+    {
+      "epoch": 3.066546762589928,
+      "force_loss": 0.0168390404433012,
+      "step": 3410
+    },
+    {
+      "epoch": 3.0755395683453237,
+      "grad_norm": 0.8553867340087891,
+      "learning_rate": 9.998666789309313e-05,
+      "loss": 0.0424,
+      "step": 3420
+    },
+    {
+      "action_loss": 0.007862004451453686,
+      "epoch": 3.0755395683453237,
+      "step": 3420
+    },
+    {
+      "epoch": 3.0755395683453237,
+      "step": 3420,
+      "torque_loss": 0.14355872571468353
+    },
+    {
+      "epoch": 3.0755395683453237,
+      "force_loss": 0.007574799004942179,
+      "step": 3420
+    },
+    {
+      "epoch": 3.0845323741007196,
+      "grad_norm": 0.7984017729759216,
+      "learning_rate": 9.998602395167475e-05,
+      "loss": 0.0343,
+      "step": 3430
+    },
+    {
+      "action_loss": 0.03197667375206947,
+      "epoch": 3.0845323741007196,
+      "step": 3430
+    },
+    {
+      "epoch": 3.0845323741007196,
+      "step": 3430,
+      "torque_loss": 0.15288887917995453
+    },
+    {
+      "epoch": 3.0845323741007196,
+      "force_loss": 0.02210446633398533,
+      "step": 3430
+    },
+    {
+      "epoch": 3.093525179856115,
+      "grad_norm": 0.7792773842811584,
+      "learning_rate": 9.998536482582213e-05,
+      "loss": 0.0443,
+      "step": 3440
+    },
+    {
+      "action_loss": 0.02684091031551361,
+      "epoch": 3.093525179856115,
+      "step": 3440
+    },
+    {
+      "epoch": 3.093525179856115,
+      "step": 3440,
+      "torque_loss": 0.14769357442855835
+    },
+    {
+      "epoch": 3.093525179856115,
+      "force_loss": 0.013106010854244232,
+      "step": 3440
+    },
+    {
+      "epoch": 3.102517985611511,
+      "grad_norm": 0.9977942109107971,
+      "learning_rate": 9.998469051573544e-05,
+      "loss": 0.0436,
+      "step": 3450
+    },
+    {
+      "action_loss": 0.03798346221446991,
+      "epoch": 3.102517985611511,
+      "step": 3450
+    },
+    {
+      "epoch": 3.102517985611511,
+      "step": 3450,
+      "torque_loss": 0.23463428020477295
+    },
+    {
+      "epoch": 3.102517985611511,
+      "force_loss": 0.0281215887516737,
+      "step": 3450
+    },
+    {
+      "epoch": 3.1115107913669067,
+      "grad_norm": 1.1203683614730835,
+      "learning_rate": 9.998400102161954e-05,
+      "loss": 0.0419,
+      "step": 3460
+    },
+    {
+      "action_loss": 0.01838778890669346,
+      "epoch": 3.1115107913669067,
+      "step": 3460
+    },
+    {
+      "epoch": 3.1115107913669067,
+      "step": 3460,
+      "torque_loss": 0.169475719332695
+    },
+    {
+      "epoch": 3.1115107913669067,
+      "force_loss": 0.010304216295480728,
+      "step": 3460
+    },
+    {
+      "epoch": 3.120503597122302,
+      "grad_norm": 1.2310011386871338,
+      "learning_rate": 9.998329634368388e-05,
+      "loss": 0.0401,
+      "step": 3470
+    },
+    {
+      "action_loss": 0.02858925424516201,
+      "epoch": 3.120503597122302,
+      "step": 3470
+    },
+    {
+      "epoch": 3.120503597122302,
+      "step": 3470,
+      "torque_loss": 0.13004864752292633
+    },
+    {
+      "epoch": 3.120503597122302,
+      "force_loss": 0.014659442007541656,
+      "step": 3470
+    },
+    {
+      "epoch": 3.129496402877698,
+      "grad_norm": 0.9335276484489441,
+      "learning_rate": 9.998257648214253e-05,
+      "loss": 0.0355,
+      "step": 3480
+    },
+    {
+      "action_loss": 0.039215151220560074,
+      "epoch": 3.129496402877698,
+      "step": 3480
+    },
+    {
+      "epoch": 3.129496402877698,
+      "step": 3480,
+      "torque_loss": 0.22563539445400238
+    },
+    {
+      "epoch": 3.129496402877698,
+      "force_loss": 0.028997493907809258,
+      "step": 3480
+    },
+    {
+      "epoch": 3.1384892086330933,
+      "grad_norm": 1.2678771018981934,
+      "learning_rate": 9.998184143721417e-05,
+      "loss": 0.0467,
+      "step": 3490
+    },
+    {
+      "action_loss": 0.016788633540272713,
+      "epoch": 3.1384892086330933,
+      "step": 3490
+    },
+    {
+      "epoch": 3.1384892086330933,
+      "step": 3490,
+      "torque_loss": 0.11538002640008926
+    },
+    {
+      "epoch": 3.1384892086330933,
+      "force_loss": 0.016048191115260124,
+      "step": 3490
+    },
+    {
+      "epoch": 3.147482014388489,
+      "grad_norm": 0.9799803495407104,
+      "learning_rate": 9.998109120912206e-05,
+      "loss": 0.0353,
+      "step": 3500
+    },
+    {
+      "action_loss": 0.02866787649691105,
+      "epoch": 3.147482014388489,
+      "step": 3500
+    },
+    {
+      "epoch": 3.147482014388489,
+      "step": 3500,
+      "torque_loss": 0.15908962488174438
+    },
+    {
+      "epoch": 3.147482014388489,
+      "force_loss": 0.020994337275624275,
+      "step": 3500
+    },
+    {
+      "epoch": 3.156474820143885,
+      "grad_norm": 1.8498404026031494,
+      "learning_rate": 9.998032579809411e-05,
+      "loss": 0.0427,
+      "step": 3510
+    },
+    {
+      "action_loss": 0.014884372241795063,
+      "epoch": 3.156474820143885,
+      "step": 3510
+    },
+    {
+      "epoch": 3.156474820143885,
+      "step": 3510,
+      "torque_loss": 0.20385859906673431
+    },
+    {
+      "epoch": 3.156474820143885,
+      "force_loss": 0.014859344810247421,
+      "step": 3510
+    },
+    {
+      "epoch": 3.1654676258992804,
+      "grad_norm": 0.9601037502288818,
+      "learning_rate": 9.997954520436286e-05,
+      "loss": 0.0447,
+      "step": 3520
+    },
+    {
+      "action_loss": 0.021648511290550232,
+      "epoch": 3.1654676258992804,
+      "step": 3520
+    },
+    {
+      "epoch": 3.1654676258992804,
+      "step": 3520,
+      "torque_loss": 0.1645248681306839
+    },
+    {
+      "epoch": 3.1654676258992804,
+      "force_loss": 0.016823599115014076,
+      "step": 3520
+    },
+    {
+      "epoch": 3.1744604316546763,
+      "grad_norm": 1.0865265130996704,
+      "learning_rate": 9.997874942816538e-05,
+      "loss": 0.0363,
+      "step": 3530
+    },
+    {
+      "action_loss": 0.034875307232141495,
+      "epoch": 3.1744604316546763,
+      "step": 3530
+    },
+    {
+      "epoch": 3.1744604316546763,
+      "step": 3530,
+      "torque_loss": 0.13772688806056976
+    },
+    {
+      "epoch": 3.1744604316546763,
+      "force_loss": 0.02724367380142212,
+      "step": 3530
+    },
+    {
+      "epoch": 3.183453237410072,
+      "grad_norm": 0.9987258911132812,
+      "learning_rate": 9.997793846974345e-05,
+      "loss": 0.0481,
+      "step": 3540
+    },
+    {
+      "action_loss": 0.013754952698946,
+      "epoch": 3.183453237410072,
+      "step": 3540
+    },
+    {
+      "epoch": 3.183453237410072,
+      "step": 3540,
+      "torque_loss": 0.11794767528772354
+    },
+    {
+      "epoch": 3.183453237410072,
+      "force_loss": 0.008873325772583485,
+      "step": 3540
+    },
+    {
+      "epoch": 3.1924460431654675,
+      "grad_norm": 1.811292052268982,
+      "learning_rate": 9.997711232934341e-05,
+      "loss": 0.0422,
+      "step": 3550
+    },
+    {
+      "action_loss": 0.04341394826769829,
+      "epoch": 3.1924460431654675,
+      "step": 3550
+    },
+    {
+      "epoch": 3.1924460431654675,
+      "step": 3550,
+      "torque_loss": 0.17486751079559326
+    },
+    {
+      "epoch": 3.1924460431654675,
+      "force_loss": 0.03140212595462799,
+      "step": 3550
+    },
+    {
+      "epoch": 3.2014388489208634,
+      "grad_norm": 0.49703332781791687,
+      "learning_rate": 9.99762710072162e-05,
+      "loss": 0.0408,
+      "step": 3560
+    },
+    {
+      "action_loss": 0.023574858903884888,
+      "epoch": 3.2014388489208634,
+      "step": 3560
+    },
+    {
+      "epoch": 3.2014388489208634,
+      "step": 3560,
+      "torque_loss": 0.17619562149047852
+    },
+    {
+      "epoch": 3.2014388489208634,
+      "force_loss": 0.02313327044248581,
+      "step": 3560
+    },
+    {
+      "epoch": 3.210431654676259,
+      "grad_norm": 0.9942551255226135,
+      "learning_rate": 9.997541450361743e-05,
+      "loss": 0.0471,
+      "step": 3570
+    },
+    {
+      "action_loss": 0.0332251638174057,
+      "epoch": 3.210431654676259,
+      "step": 3570
+    },
+    {
+      "epoch": 3.210431654676259,
+      "step": 3570,
+      "torque_loss": 0.2173522710800171
+    },
+    {
+      "epoch": 3.210431654676259,
+      "force_loss": 0.03791753575205803,
+      "step": 3570
+    },
+    {
+      "epoch": 3.2194244604316546,
+      "grad_norm": 0.9259369373321533,
+      "learning_rate": 9.997454281880723e-05,
+      "loss": 0.0405,
+      "step": 3580
+    },
+    {
+      "action_loss": 0.014682799577713013,
+      "epoch": 3.2194244604316546,
+      "step": 3580
+    },
+    {
+      "epoch": 3.2194244604316546,
+      "step": 3580,
+      "torque_loss": 0.12946762144565582
+    },
+    {
+      "epoch": 3.2194244604316546,
+      "force_loss": 0.007857296615839005,
+      "step": 3580
+    },
+    {
+      "epoch": 3.2284172661870505,
+      "grad_norm": 1.395731806755066,
+      "learning_rate": 9.997365595305044e-05,
+      "loss": 0.0395,
+      "step": 3590
+    },
+    {
+      "action_loss": 0.03349921479821205,
+      "epoch": 3.2284172661870505,
+      "step": 3590
+    },
+    {
+      "epoch": 3.2284172661870505,
+      "step": 3590,
+      "torque_loss": 0.22216928005218506
+    },
+    {
+      "epoch": 3.2284172661870505,
+      "force_loss": 0.029241399839520454,
+      "step": 3590
+    },
+    {
+      "epoch": 3.237410071942446,
+      "grad_norm": 1.07401442527771,
+      "learning_rate": 9.997275390661644e-05,
+      "loss": 0.0382,
+      "step": 3600
+    },
+    {
+      "action_loss": 0.013655010610818863,
+      "epoch": 3.237410071942446,
+      "step": 3600
+    },
+    {
+      "epoch": 3.237410071942446,
+      "step": 3600,
+      "torque_loss": 0.1119321957230568
+    },
+    {
+      "epoch": 3.237410071942446,
+      "force_loss": 0.01100183930248022,
+      "step": 3600
+    },
+    {
+      "epoch": 3.2464028776978417,
+      "grad_norm": 1.507638692855835,
+      "learning_rate": 9.997183667977926e-05,
+      "loss": 0.0392,
+      "step": 3610
+    },
+    {
+      "action_loss": 0.01712990552186966,
+      "epoch": 3.2464028776978417,
+      "step": 3610
+    },
+    {
+      "epoch": 3.2464028776978417,
+      "step": 3610,
+      "torque_loss": 0.15502284467220306
+    },
+    {
+      "epoch": 3.2464028776978417,
+      "force_loss": 0.011407186277210712,
+      "step": 3610
+    },
+    {
+      "epoch": 3.2553956834532376,
+      "grad_norm": 1.8317286968231201,
+      "learning_rate": 9.997090427281752e-05,
+      "loss": 0.0379,
+      "step": 3620
+    },
+    {
+      "action_loss": 0.014633041806519032,
+      "epoch": 3.2553956834532376,
+      "step": 3620
+    },
+    {
+      "epoch": 3.2553956834532376,
+      "step": 3620,
+      "torque_loss": 0.10888916999101639
+    },
+    {
+      "epoch": 3.2553956834532376,
+      "force_loss": 0.014046691358089447,
+      "step": 3620
+    },
+    {
+      "epoch": 3.264388489208633,
+      "grad_norm": 0.5321849584579468,
+      "learning_rate": 9.996995668601448e-05,
+      "loss": 0.0357,
+      "step": 3630
+    },
+    {
+      "action_loss": 0.008210469037294388,
+      "epoch": 3.264388489208633,
+      "step": 3630
+    },
+    {
+      "epoch": 3.264388489208633,
+      "step": 3630,
+      "torque_loss": 0.07508663088083267
+    },
+    {
+      "epoch": 3.264388489208633,
+      "force_loss": 0.006780738476663828,
+      "step": 3630
+    },
+    {
+      "epoch": 3.273381294964029,
+      "grad_norm": 0.8848887085914612,
+      "learning_rate": 9.996899391965798e-05,
+      "loss": 0.0296,
+      "step": 3640
+    },
+    {
+      "action_loss": 0.028002088889479637,
+      "epoch": 3.273381294964029,
+      "step": 3640
+    },
+    {
+      "epoch": 3.273381294964029,
+      "step": 3640,
+      "torque_loss": 0.18522708117961884
+    },
+    {
+      "epoch": 3.273381294964029,
+      "force_loss": 0.024815693497657776,
+      "step": 3640
+    },
+    {
+      "epoch": 3.2823741007194247,
+      "grad_norm": 1.1505377292633057,
+      "learning_rate": 9.996801597404048e-05,
+      "loss": 0.042,
+      "step": 3650
+    },
+    {
+      "action_loss": 0.016801102086901665,
+      "epoch": 3.2823741007194247,
+      "step": 3650
+    },
+    {
+      "epoch": 3.2823741007194247,
+      "step": 3650,
+      "torque_loss": 0.16668272018432617
+    },
+    {
+      "epoch": 3.2823741007194247,
+      "force_loss": 0.012274055741727352,
+      "step": 3650
+    },
+    {
+      "epoch": 3.29136690647482,
+      "grad_norm": 1.0085569620132446,
+      "learning_rate": 9.996702284945905e-05,
+      "loss": 0.0342,
+      "step": 3660
+    },
+    {
+      "action_loss": 0.023353412747383118,
+      "epoch": 3.29136690647482,
+      "step": 3660
+    },
+    {
+      "epoch": 3.29136690647482,
+      "step": 3660,
+      "torque_loss": 0.12883375585079193
+    },
+    {
+      "epoch": 3.29136690647482,
+      "force_loss": 0.021165847778320312,
+      "step": 3660
+    },
+    {
+      "epoch": 3.300359712230216,
+      "grad_norm": 1.063624620437622,
+      "learning_rate": 9.996601454621539e-05,
+      "loss": 0.0396,
+      "step": 3670
+    },
+    {
+      "action_loss": 0.01499168947339058,
+      "epoch": 3.300359712230216,
+      "step": 3670
+    },
+    {
+      "epoch": 3.300359712230216,
+      "step": 3670,
+      "torque_loss": 0.15017037093639374
+    },
+    {
+      "epoch": 3.300359712230216,
+      "force_loss": 0.011859290301799774,
+      "step": 3670
+    },
+    {
+      "epoch": 3.3093525179856114,
+      "grad_norm": 0.9896354079246521,
+      "learning_rate": 9.996499106461577e-05,
+      "loss": 0.0348,
+      "step": 3680
+    },
+    {
+      "action_loss": 0.01218866091221571,
+      "epoch": 3.3093525179856114,
+      "step": 3680
+    },
+    {
+      "epoch": 3.3093525179856114,
+      "step": 3680,
+      "torque_loss": 0.1389753669500351
+    },
+    {
+      "epoch": 3.3093525179856114,
+      "force_loss": 0.011266622692346573,
+      "step": 3680
+    },
+    {
+      "epoch": 3.318345323741007,
+      "grad_norm": 1.788463830947876,
+      "learning_rate": 9.996395240497112e-05,
+      "loss": 0.0368,
+      "step": 3690
+    },
+    {
+      "action_loss": 0.03510531410574913,
+      "epoch": 3.318345323741007,
+      "step": 3690
+    },
+    {
+      "epoch": 3.318345323741007,
+      "step": 3690,
+      "torque_loss": 0.1701621562242508
+    },
+    {
+      "epoch": 3.318345323741007,
+      "force_loss": 0.02181488275527954,
+      "step": 3690
+    },
+    {
+      "epoch": 3.327338129496403,
+      "grad_norm": 1.2068116664886475,
+      "learning_rate": 9.996289856759696e-05,
+      "loss": 0.0397,
+      "step": 3700
+    },
+    {
+      "action_loss": 0.019518474116921425,
+      "epoch": 3.327338129496403,
+      "step": 3700
+    },
+    {
+      "epoch": 3.327338129496403,
+      "step": 3700,
+      "torque_loss": 0.11307797580957413
+    },
+    {
+      "epoch": 3.327338129496403,
+      "force_loss": 0.013637223280966282,
+      "step": 3700
+    },
+    {
+      "epoch": 3.3363309352517985,
+      "grad_norm": 1.6817306280136108,
+      "learning_rate": 9.996182955281342e-05,
+      "loss": 0.0374,
+      "step": 3710
+    },
+    {
+      "action_loss": 0.03645653650164604,
+      "epoch": 3.3363309352517985,
+      "step": 3710
+    },
+    {
+      "epoch": 3.3363309352517985,
+      "step": 3710,
+      "torque_loss": 0.18721990287303925
+    },
+    {
+      "epoch": 3.3363309352517985,
+      "force_loss": 0.023440003395080566,
+      "step": 3710
+    },
+    {
+      "epoch": 3.3453237410071943,
+      "grad_norm": 1.4603242874145508,
+      "learning_rate": 9.996074536094519e-05,
+      "loss": 0.0406,
+      "step": 3720
+    },
+    {
+      "action_loss": 0.03198450058698654,
+      "epoch": 3.3453237410071943,
+      "step": 3720
+    },
+    {
+      "epoch": 3.3453237410071943,
+      "step": 3720,
+      "torque_loss": 0.1488967388868332
+    },
+    {
+      "epoch": 3.3453237410071943,
+      "force_loss": 0.02351897396147251,
+      "step": 3720
+    },
+    {
+      "epoch": 3.3543165467625897,
+      "grad_norm": 0.5646918416023254,
+      "learning_rate": 9.995964599232168e-05,
+      "loss": 0.0365,
+      "step": 3730
+    },
+    {
+      "action_loss": 0.03587939590215683,
+      "epoch": 3.3543165467625897,
+      "step": 3730
+    },
+    {
+      "epoch": 3.3543165467625897,
+      "step": 3730,
+      "torque_loss": 0.14359591901302338
+    },
+    {
+      "epoch": 3.3543165467625897,
+      "force_loss": 0.026941703632473946,
+      "step": 3730
+    },
+    {
+      "epoch": 3.3633093525179856,
+      "grad_norm": 0.924182116985321,
+      "learning_rate": 9.995853144727683e-05,
+      "loss": 0.0409,
+      "step": 3740
+    },
+    {
+      "action_loss": 0.029336510226130486,
+      "epoch": 3.3633093525179856,
+      "step": 3740
+    },
+    {
+      "epoch": 3.3633093525179856,
+      "step": 3740,
+      "torque_loss": 0.16399581730365753
+    },
+    {
+      "epoch": 3.3633093525179856,
+      "force_loss": 0.018774529919028282,
+      "step": 3740
+    },
+    {
+      "epoch": 3.3723021582733814,
+      "grad_norm": 2.190215587615967,
+      "learning_rate": 9.99574017261492e-05,
+      "loss": 0.0464,
+      "step": 3750
+    },
+    {
+      "action_loss": 0.01895342394709587,
+      "epoch": 3.3723021582733814,
+      "step": 3750
+    },
+    {
+      "epoch": 3.3723021582733814,
+      "step": 3750,
+      "torque_loss": 0.09289685636758804
+    },
+    {
+      "epoch": 3.3723021582733814,
+      "force_loss": 0.015037271194159985,
+      "step": 3750
+    },
+    {
+      "epoch": 3.381294964028777,
+      "grad_norm": 0.9864612221717834,
+      "learning_rate": 9.995625682928198e-05,
+      "loss": 0.0378,
+      "step": 3760
+    },
+    {
+      "action_loss": 0.01902839168906212,
+      "epoch": 3.381294964028777,
+      "step": 3760
+    },
+    {
+      "epoch": 3.381294964028777,
+      "step": 3760,
+      "torque_loss": 0.15829142928123474
+    },
+    {
+      "epoch": 3.381294964028777,
+      "force_loss": 0.013129468075931072,
+      "step": 3760
+    },
+    {
+      "epoch": 3.3902877697841727,
+      "grad_norm": 1.2213791608810425,
+      "learning_rate": 9.995509675702295e-05,
+      "loss": 0.045,
+      "step": 3770
+    },
+    {
+      "action_loss": 0.01340919453650713,
+      "epoch": 3.3902877697841727,
+      "step": 3770
+    },
+    {
+      "epoch": 3.3902877697841727,
+      "step": 3770,
+      "torque_loss": 0.14504216611385345
+    },
+    {
+      "epoch": 3.3902877697841727,
+      "force_loss": 0.009189505130052567,
+      "step": 3770
+    },
+    {
+      "epoch": 3.3992805755395685,
+      "grad_norm": 0.8528039455413818,
+      "learning_rate": 9.995392150972451e-05,
+      "loss": 0.0436,
+      "step": 3780
+    },
+    {
+      "action_loss": 0.07091962546110153,
+      "epoch": 3.3992805755395685,
+      "step": 3780
+    },
+    {
+      "epoch": 3.3992805755395685,
+      "step": 3780,
+      "torque_loss": 0.14519180357456207
+    },
+    {
+      "epoch": 3.3992805755395685,
+      "force_loss": 0.0425095297396183,
+      "step": 3780
+    },
+    {
+      "epoch": 3.408273381294964,
+      "grad_norm": 0.9560016393661499,
+      "learning_rate": 9.995273108774366e-05,
+      "loss": 0.0528,
+      "step": 3790
+    },
+    {
+      "action_loss": 0.016869833692908287,
+      "epoch": 3.408273381294964,
+      "step": 3790
+    },
+    {
+      "epoch": 3.408273381294964,
+      "step": 3790,
+      "torque_loss": 0.12144602090120316
+    },
+    {
+      "epoch": 3.408273381294964,
+      "force_loss": 0.00925482902675867,
+      "step": 3790
+    },
+    {
+      "epoch": 3.41726618705036,
+      "grad_norm": 1.135102391242981,
+      "learning_rate": 9.995152549144205e-05,
+      "loss": 0.0495,
+      "step": 3800
+    },
+    {
+      "action_loss": 0.020546669140458107,
+      "epoch": 3.41726618705036,
+      "step": 3800
+    },
+    {
+      "epoch": 3.41726618705036,
+      "step": 3800,
+      "torque_loss": 0.20160607993602753
+    },
+    {
+      "epoch": 3.41726618705036,
+      "force_loss": 0.015021462924778461,
+      "step": 3800
+    },
+    {
+      "epoch": 3.4262589928057556,
+      "grad_norm": 0.9326383471488953,
+      "learning_rate": 9.995030472118587e-05,
+      "loss": 0.0426,
+      "step": 3810
+    },
+    {
+      "action_loss": 0.04457582160830498,
+      "epoch": 3.4262589928057556,
+      "step": 3810
+    },
+    {
+      "epoch": 3.4262589928057556,
+      "step": 3810,
+      "torque_loss": 0.17118744552135468
+    },
+    {
+      "epoch": 3.4262589928057556,
+      "force_loss": 0.026535334065556526,
+      "step": 3810
+    },
+    {
+      "epoch": 3.435251798561151,
+      "grad_norm": 1.203792691230774,
+      "learning_rate": 9.9949068777346e-05,
+      "loss": 0.0442,
+      "step": 3820
+    },
+    {
+      "action_loss": 0.0204255823045969,
+      "epoch": 3.435251798561151,
+      "step": 3820
+    },
+    {
+      "epoch": 3.435251798561151,
+      "step": 3820,
+      "torque_loss": 0.18842995166778564
+    },
+    {
+      "epoch": 3.435251798561151,
+      "force_loss": 0.01134722400456667,
+      "step": 3820
+    },
+    {
+      "epoch": 3.444244604316547,
+      "grad_norm": 1.5818146467208862,
+      "learning_rate": 9.994781766029786e-05,
+      "loss": 0.0392,
+      "step": 3830
+    },
+    {
+      "action_loss": 0.041504666209220886,
+      "epoch": 3.444244604316547,
+      "step": 3830
+    },
+    {
+      "epoch": 3.444244604316547,
+      "step": 3830,
+      "torque_loss": 0.18697874248027802
+    },
+    {
+      "epoch": 3.444244604316547,
+      "force_loss": 0.03712286055088043,
+      "step": 3830
+    },
+    {
+      "epoch": 3.4532374100719423,
+      "grad_norm": 0.9084063768386841,
+      "learning_rate": 9.994655137042151e-05,
+      "loss": 0.0465,
+      "step": 3840
+    },
+    {
+      "action_loss": 0.04241631552577019,
+      "epoch": 3.4532374100719423,
+      "step": 3840
+    },
+    {
+      "epoch": 3.4532374100719423,
+      "step": 3840,
+      "torque_loss": 0.16232672333717346
+    },
+    {
+      "epoch": 3.4532374100719423,
+      "force_loss": 0.022606873884797096,
+      "step": 3840
+    },
+    {
+      "epoch": 3.462230215827338,
+      "grad_norm": 1.1572521924972534,
+      "learning_rate": 9.99452699081016e-05,
+      "loss": 0.0365,
+      "step": 3850
+    },
+    {
+      "action_loss": 0.02277149260044098,
+      "epoch": 3.462230215827338,
+      "step": 3850
+    },
+    {
+      "epoch": 3.462230215827338,
+      "step": 3850,
+      "torque_loss": 0.17342841625213623
+    },
+    {
+      "epoch": 3.462230215827338,
+      "force_loss": 0.011866346001625061,
+      "step": 3850
+    },
+    {
+      "epoch": 3.471223021582734,
+      "grad_norm": 1.7111402750015259,
+      "learning_rate": 9.994397327372743e-05,
+      "loss": 0.0443,
+      "step": 3860
+    },
+    {
+      "action_loss": 0.033042121678590775,
+      "epoch": 3.471223021582734,
+      "step": 3860
+    },
+    {
+      "epoch": 3.471223021582734,
+      "step": 3860,
+      "torque_loss": 0.21560950577259064
+    },
+    {
+      "epoch": 3.471223021582734,
+      "force_loss": 0.035754211246967316,
+      "step": 3860
+    },
+    {
+      "epoch": 3.4802158273381294,
+      "grad_norm": 1.9151628017425537,
+      "learning_rate": 9.994266146769286e-05,
+      "loss": 0.045,
+      "step": 3870
+    },
+    {
+      "action_loss": 0.01448530051857233,
+      "epoch": 3.4802158273381294,
+      "step": 3870
+    },
+    {
+      "epoch": 3.4802158273381294,
+      "step": 3870,
+      "torque_loss": 0.15126045048236847
+    },
+    {
+      "epoch": 3.4802158273381294,
+      "force_loss": 0.012448366731405258,
+      "step": 3870
+    },
+    {
+      "epoch": 3.4892086330935252,
+      "grad_norm": 1.1126052141189575,
+      "learning_rate": 9.994133449039642e-05,
+      "loss": 0.041,
+      "step": 3880
+    },
+    {
+      "action_loss": 0.02015937864780426,
+      "epoch": 3.4892086330935252,
+      "step": 3880
+    },
+    {
+      "epoch": 3.4892086330935252,
+      "step": 3880,
+      "torque_loss": 0.1298368126153946
+    },
+    {
+      "epoch": 3.4892086330935252,
+      "force_loss": 0.012545089237391949,
+      "step": 3880
+    },
+    {
+      "epoch": 3.4982014388489207,
+      "grad_norm": 1.0569645166397095,
+      "learning_rate": 9.993999234224118e-05,
+      "loss": 0.0429,
+      "step": 3890
+    },
+    {
+      "action_loss": 0.026209423318505287,
+      "epoch": 3.4982014388489207,
+      "step": 3890
+    },
+    {
+      "epoch": 3.4982014388489207,
+      "step": 3890,
+      "torque_loss": 0.1408567875623703
+    },
+    {
+      "epoch": 3.4982014388489207,
+      "force_loss": 0.022115493193268776,
+      "step": 3890
+    },
+    {
+      "epoch": 3.5071942446043165,
+      "grad_norm": 1.0926611423492432,
+      "learning_rate": 9.993863502363485e-05,
+      "loss": 0.0374,
+      "step": 3900
+    },
+    {
+      "action_loss": 0.02498788945376873,
+      "epoch": 3.5071942446043165,
+      "step": 3900
+    },
+    {
+      "epoch": 3.5071942446043165,
+      "step": 3900,
+      "torque_loss": 0.15101857483386993
+    },
+    {
+      "epoch": 3.5071942446043165,
+      "force_loss": 0.02080947905778885,
+      "step": 3900
+    },
+    {
+      "epoch": 3.5161870503597124,
+      "grad_norm": 0.7063518762588501,
+      "learning_rate": 9.993726253498976e-05,
+      "loss": 0.034,
+      "step": 3910
+    },
+    {
+      "action_loss": 0.02012750692665577,
+      "epoch": 3.5161870503597124,
+      "step": 3910
+    },
+    {
+      "epoch": 3.5161870503597124,
+      "step": 3910,
+      "torque_loss": 0.11606822162866592
+    },
+    {
+      "epoch": 3.5161870503597124,
+      "force_loss": 0.017499500885605812,
+      "step": 3910
+    },
+    {
+      "epoch": 3.5251798561151078,
+      "grad_norm": 1.7970179319381714,
+      "learning_rate": 9.993587487672282e-05,
+      "loss": 0.0349,
+      "step": 3920
+    },
+    {
+      "action_loss": 0.025140687823295593,
+      "epoch": 3.5251798561151078,
+      "step": 3920
+    },
+    {
+      "epoch": 3.5251798561151078,
+      "step": 3920,
+      "torque_loss": 0.12630312144756317
+    },
+    {
+      "epoch": 3.5251798561151078,
+      "force_loss": 0.01934153400361538,
+      "step": 3920
+    },
+    {
+      "epoch": 3.5341726618705036,
+      "grad_norm": 1.2401013374328613,
+      "learning_rate": 9.993447204925558e-05,
+      "loss": 0.0422,
+      "step": 3930
+    },
+    {
+      "action_loss": 0.03816840425133705,
+      "epoch": 3.5341726618705036,
+      "step": 3930
+    },
+    {
+      "epoch": 3.5341726618705036,
+      "step": 3930,
+      "torque_loss": 0.22597800195217133
+    },
+    {
+      "epoch": 3.5341726618705036,
+      "force_loss": 0.0375300869345665,
+      "step": 3930
+    },
+    {
+      "epoch": 3.543165467625899,
+      "grad_norm": 0.7885603308677673,
+      "learning_rate": 9.993305405301416e-05,
+      "loss": 0.0448,
+      "step": 3940
+    },
+    {
+      "action_loss": 0.020705319941043854,
+      "epoch": 3.543165467625899,
+      "step": 3940
+    },
+    {
+      "epoch": 3.543165467625899,
+      "step": 3940,
+      "torque_loss": 0.11626722663640976
+    },
+    {
+      "epoch": 3.543165467625899,
+      "force_loss": 0.02008006162941456,
+      "step": 3940
+    },
+    {
+      "epoch": 3.552158273381295,
+      "grad_norm": 0.9565855264663696,
+      "learning_rate": 9.993162088842935e-05,
+      "loss": 0.0412,
+      "step": 3950
+    },
+    {
+      "action_loss": 0.04231289029121399,
+      "epoch": 3.552158273381295,
+      "step": 3950
+    },
+    {
+      "epoch": 3.552158273381295,
+      "step": 3950,
+      "torque_loss": 0.18785852193832397
+    },
+    {
+      "epoch": 3.552158273381295,
+      "force_loss": 0.022834552451968193,
+      "step": 3950
+    },
+    {
+      "epoch": 3.5611510791366907,
+      "grad_norm": 0.9896112084388733,
+      "learning_rate": 9.993017255593646e-05,
+      "loss": 0.048,
+      "step": 3960
+    },
+    {
+      "action_loss": 0.015152017585933208,
+      "epoch": 3.5611510791366907,
+      "step": 3960
+    },
+    {
+      "epoch": 3.5611510791366907,
+      "step": 3960,
+      "torque_loss": 0.16741900146007538
+    },
+    {
+      "epoch": 3.5611510791366907,
+      "force_loss": 0.009725605137646198,
+      "step": 3960
+    },
+    {
+      "epoch": 3.5701438848920866,
+      "grad_norm": 0.6375613808631897,
+      "learning_rate": 9.992870905597548e-05,
+      "loss": 0.0382,
+      "step": 3970
+    },
+    {
+      "action_loss": 0.0428573377430439,
+      "epoch": 3.5701438848920866,
+      "step": 3970
+    },
+    {
+      "epoch": 3.5701438848920866,
+      "step": 3970,
+      "torque_loss": 0.11328927427530289
+    },
+    {
+      "epoch": 3.5701438848920866,
+      "force_loss": 0.03325273469090462,
+      "step": 3970
+    },
+    {
+      "epoch": 3.579136690647482,
+      "grad_norm": 0.8191521763801575,
+      "learning_rate": 9.9927230388991e-05,
+      "loss": 0.0361,
+      "step": 3980
+    },
+    {
+      "action_loss": 0.018099667504429817,
+      "epoch": 3.579136690647482,
+      "step": 3980
+    },
+    {
+      "epoch": 3.579136690647482,
+      "step": 3980,
+      "torque_loss": 0.12714602053165436
+    },
+    {
+      "epoch": 3.579136690647482,
+      "force_loss": 0.015349042601883411,
+      "step": 3980
+    },
+    {
+      "epoch": 3.588129496402878,
+      "grad_norm": 1.7662588357925415,
+      "learning_rate": 9.992573655543215e-05,
+      "loss": 0.0349,
+      "step": 3990
+    },
+    {
+      "action_loss": 0.018912767991423607,
+      "epoch": 3.588129496402878,
+      "step": 3990
+    },
+    {
+      "epoch": 3.588129496402878,
+      "step": 3990,
+      "torque_loss": 0.1362752616405487
+    },
+    {
+      "epoch": 3.588129496402878,
+      "force_loss": 0.011249770410358906,
+      "step": 3990
+    },
+    {
+      "epoch": 3.597122302158273,
+      "grad_norm": 0.8619410991668701,
+      "learning_rate": 9.992422755575277e-05,
+      "loss": 0.034,
+      "step": 4000
+    },
+    {
+      "action_loss": 0.022844454273581505,
+      "epoch": 3.597122302158273,
+      "step": 4000
+    },
+    {
+      "epoch": 3.597122302158273,
+      "step": 4000,
+      "torque_loss": 0.14633597433567047
+    },
+    {
+      "epoch": 3.597122302158273,
+      "force_loss": 0.011328269727528095,
+      "step": 4000
+    },
+    {
+      "epoch": 3.606115107913669,
+      "grad_norm": 0.6428923606872559,
+      "learning_rate": 9.992270339041123e-05,
+      "loss": 0.0381,
+      "step": 4010
+    },
+    {
+      "action_loss": 0.015509418211877346,
+      "epoch": 3.606115107913669,
+      "step": 4010
+    },
+    {
+      "epoch": 3.606115107913669,
+      "step": 4010,
+      "torque_loss": 0.20643214881420135
+    },
+    {
+      "epoch": 3.606115107913669,
+      "force_loss": 0.007344987243413925,
+      "step": 4010
+    },
+    {
+      "epoch": 3.615107913669065,
+      "grad_norm": 2.0724446773529053,
+      "learning_rate": 9.992116405987053e-05,
+      "loss": 0.0369,
+      "step": 4020
+    },
+    {
+      "action_loss": 0.025622619315981865,
+      "epoch": 3.615107913669065,
+      "step": 4020
+    },
+    {
+      "epoch": 3.615107913669065,
+      "step": 4020,
+      "torque_loss": 0.16435176134109497
+    },
+    {
+      "epoch": 3.615107913669065,
+      "force_loss": 0.023499146103858948,
+      "step": 4020
+    },
+    {
+      "epoch": 3.6241007194244603,
+      "grad_norm": 0.7356473803520203,
+      "learning_rate": 9.991960956459828e-05,
+      "loss": 0.0386,
+      "step": 4030
+    },
+    {
+      "action_loss": 0.023059682920575142,
+      "epoch": 3.6241007194244603,
+      "step": 4030
+    },
+    {
+      "epoch": 3.6241007194244603,
+      "step": 4030,
+      "torque_loss": 0.15098251402378082
+    },
+    {
+      "epoch": 3.6241007194244603,
+      "force_loss": 0.01126853097230196,
+      "step": 4030
+    },
+    {
+      "epoch": 3.633093525179856,
+      "grad_norm": 0.8907721042633057,
+      "learning_rate": 9.991803990506669e-05,
+      "loss": 0.0403,
+      "step": 4040
+    },
+    {
+      "action_loss": 0.02048400230705738,
+      "epoch": 3.633093525179856,
+      "step": 4040
+    },
+    {
+      "epoch": 3.633093525179856,
+      "step": 4040,
+      "torque_loss": 0.16468499600887299
+    },
+    {
+      "epoch": 3.633093525179856,
+      "force_loss": 0.021904809400439262,
+      "step": 4040
+    },
+    {
+      "epoch": 3.6420863309352516,
+      "grad_norm": 1.6261680126190186,
+      "learning_rate": 9.991645508175258e-05,
+      "loss": 0.0453,
+      "step": 4050
+    },
+    {
+      "action_loss": 0.03897305205464363,
+      "epoch": 3.6420863309352516,
+      "step": 4050
+    },
+    {
+      "epoch": 3.6420863309352516,
+      "step": 4050,
+      "torque_loss": 0.1823081225156784
+    },
+    {
+      "epoch": 3.6420863309352516,
+      "force_loss": 0.022277308627963066,
+      "step": 4050
+    },
+    {
+      "epoch": 3.6510791366906474,
+      "grad_norm": 0.8403676152229309,
+      "learning_rate": 9.99148550951374e-05,
+      "loss": 0.0349,
+      "step": 4060
+    },
+    {
+      "action_loss": 0.019982775673270226,
+      "epoch": 3.6510791366906474,
+      "step": 4060
+    },
+    {
+      "epoch": 3.6510791366906474,
+      "step": 4060,
+      "torque_loss": 0.14510613679885864
+    },
+    {
+      "epoch": 3.6510791366906474,
+      "force_loss": 0.016272230073809624,
+      "step": 4060
+    },
+    {
+      "epoch": 3.6600719424460433,
+      "grad_norm": 0.7153682112693787,
+      "learning_rate": 9.991323994570716e-05,
+      "loss": 0.0481,
+      "step": 4070
+    },
+    {
+      "action_loss": 0.012759196572005749,
+      "epoch": 3.6600719424460433,
+      "step": 4070
+    },
+    {
+      "epoch": 3.6600719424460433,
+      "step": 4070,
+      "torque_loss": 0.1196979507803917
+    },
+    {
+      "epoch": 3.6600719424460433,
+      "force_loss": 0.011219008825719357,
+      "step": 4070
+    },
+    {
+      "epoch": 3.6690647482014387,
+      "grad_norm": 1.0684508085250854,
+      "learning_rate": 9.99116096339525e-05,
+      "loss": 0.038,
+      "step": 4080
+    },
+    {
+      "action_loss": 0.03012736700475216,
+      "epoch": 3.6690647482014387,
+      "step": 4080
+    },
+    {
+      "epoch": 3.6690647482014387,
+      "step": 4080,
+      "torque_loss": 0.13003218173980713
+    },
+    {
+      "epoch": 3.6690647482014387,
+      "force_loss": 0.020892785862088203,
+      "step": 4080
+    },
+    {
+      "epoch": 3.6780575539568345,
+      "grad_norm": 1.5951957702636719,
+      "learning_rate": 9.990996416036869e-05,
+      "loss": 0.0351,
+      "step": 4090
+    },
+    {
+      "action_loss": 0.01665342226624489,
+      "epoch": 3.6780575539568345,
+      "step": 4090
+    },
+    {
+      "epoch": 3.6780575539568345,
+      "step": 4090,
+      "torque_loss": 0.1305442452430725
+    },
+    {
+      "epoch": 3.6780575539568345,
+      "force_loss": 0.01119800191372633,
+      "step": 4090
+    },
+    {
+      "epoch": 3.68705035971223,
+      "grad_norm": 1.4955958127975464,
+      "learning_rate": 9.990830352545555e-05,
+      "loss": 0.0366,
+      "step": 4100
+    },
+    {
+      "action_loss": 0.013869650661945343,
+      "epoch": 3.68705035971223,
+      "step": 4100
+    },
+    {
+      "epoch": 3.68705035971223,
+      "step": 4100,
+      "torque_loss": 0.10206267982721329
+    },
+    {
+      "epoch": 3.68705035971223,
+      "force_loss": 0.008940272964537144,
+      "step": 4100
+    },
+    {
+      "epoch": 3.696043165467626,
+      "grad_norm": 0.483769029378891,
+      "learning_rate": 9.990662772971756e-05,
+      "loss": 0.0326,
+      "step": 4110
+    },
+    {
+      "action_loss": 0.03507047891616821,
+      "epoch": 3.696043165467626,
+      "step": 4110
+    },
+    {
+      "epoch": 3.696043165467626,
+      "step": 4110,
+      "torque_loss": 0.19386164844036102
+    },
+    {
+      "epoch": 3.696043165467626,
+      "force_loss": 0.02307868003845215,
+      "step": 4110
+    },
+    {
+      "epoch": 3.7050359712230216,
+      "grad_norm": 1.2861037254333496,
+      "learning_rate": 9.990493677366376e-05,
+      "loss": 0.0356,
+      "step": 4120
+    },
+    {
+      "action_loss": 0.018631864339113235,
+      "epoch": 3.7050359712230216,
+      "step": 4120
+    },
+    {
+      "epoch": 3.7050359712230216,
+      "step": 4120,
+      "torque_loss": 0.18866489827632904
+    },
+    {
+      "epoch": 3.7050359712230216,
+      "force_loss": 0.01411637756973505,
+      "step": 4120
+    },
+    {
+      "epoch": 3.7140287769784175,
+      "grad_norm": 0.7408118844032288,
+      "learning_rate": 9.990323065780786e-05,
+      "loss": 0.032,
+      "step": 4130
+    },
+    {
+      "action_loss": 0.022415978834033012,
+      "epoch": 3.7140287769784175,
+      "step": 4130
+    },
+    {
+      "epoch": 3.7140287769784175,
+      "step": 4130,
+      "torque_loss": 0.09953228384256363
+    },
+    {
+      "epoch": 3.7140287769784175,
+      "force_loss": 0.011655923910439014,
+      "step": 4130
+    },
+    {
+      "epoch": 3.723021582733813,
+      "grad_norm": 1.2279568910598755,
+      "learning_rate": 9.990150938266808e-05,
+      "loss": 0.0389,
+      "step": 4140
+    },
+    {
+      "action_loss": 0.012769377790391445,
+      "epoch": 3.723021582733813,
+      "step": 4140
+    },
+    {
+      "epoch": 3.723021582733813,
+      "step": 4140,
+      "torque_loss": 0.1138978898525238
+    },
+    {
+      "epoch": 3.723021582733813,
+      "force_loss": 0.011995532549917698,
+      "step": 4140
+    },
+    {
+      "epoch": 3.7320143884892087,
+      "grad_norm": 1.4583615064620972,
+      "learning_rate": 9.989977294876733e-05,
+      "loss": 0.0369,
+      "step": 4150
+    },
+    {
+      "action_loss": 0.015207845717668533,
+      "epoch": 3.7320143884892087,
+      "step": 4150
+    },
+    {
+      "epoch": 3.7320143884892087,
+      "step": 4150,
+      "torque_loss": 0.11968394368886948
+    },
+    {
+      "epoch": 3.7320143884892087,
+      "force_loss": 0.009818506427109241,
+      "step": 4150
+    },
+    {
+      "epoch": 3.741007194244604,
+      "grad_norm": 0.9236806631088257,
+      "learning_rate": 9.989802135663308e-05,
+      "loss": 0.0294,
+      "step": 4160
+    },
+    {
+      "action_loss": 0.02421175129711628,
+      "epoch": 3.741007194244604,
+      "step": 4160
+    },
+    {
+      "epoch": 3.741007194244604,
+      "step": 4160,
+      "torque_loss": 0.17359507083892822
+    },
+    {
+      "epoch": 3.741007194244604,
+      "force_loss": 0.014731325209140778,
+      "step": 4160
+    },
+    {
+      "epoch": 3.75,
+      "grad_norm": 0.9404249787330627,
+      "learning_rate": 9.989625460679743e-05,
+      "loss": 0.0325,
+      "step": 4170
+    },
+    {
+      "action_loss": 0.031850870698690414,
+      "epoch": 3.75,
+      "step": 4170
+    },
+    {
+      "epoch": 3.75,
+      "step": 4170,
+      "torque_loss": 0.196450337767601
+    },
+    {
+      "epoch": 3.75,
+      "force_loss": 0.036668237298727036,
+      "step": 4170
+    },
+    {
+      "epoch": 3.758992805755396,
+      "grad_norm": 0.9139700531959534,
+      "learning_rate": 9.989447269979706e-05,
+      "loss": 0.0343,
+      "step": 4180
+    },
+    {
+      "action_loss": 0.012918020598590374,
+      "epoch": 3.758992805755396,
+      "step": 4180
+    },
+    {
+      "epoch": 3.758992805755396,
+      "step": 4180,
+      "torque_loss": 0.1268961876630783
+    },
+    {
+      "epoch": 3.758992805755396,
+      "force_loss": 0.008404101245105267,
+      "step": 4180
+    },
+    {
+      "epoch": 3.7679856115107913,
+      "grad_norm": 0.8955261707305908,
+      "learning_rate": 9.989267563617328e-05,
+      "loss": 0.0357,
+      "step": 4190
+    },
+    {
+      "action_loss": 0.009634359739720821,
+      "epoch": 3.7679856115107913,
+      "step": 4190
+    },
+    {
+      "epoch": 3.7679856115107913,
+      "step": 4190,
+      "torque_loss": 0.12917061150074005
+    },
+    {
+      "epoch": 3.7679856115107913,
+      "force_loss": 0.007190661504864693,
+      "step": 4190
+    },
+    {
+      "epoch": 3.776978417266187,
+      "grad_norm": 1.0073851346969604,
+      "learning_rate": 9.989086341647198e-05,
+      "loss": 0.0317,
+      "step": 4200
+    },
+    {
+      "action_loss": 0.033623646944761276,
+      "epoch": 3.776978417266187,
+      "step": 4200
+    },
+    {
+      "epoch": 3.776978417266187,
+      "step": 4200,
+      "torque_loss": 0.12459053844213486
+    },
+    {
+      "epoch": 3.776978417266187,
+      "force_loss": 0.02177286706864834,
+      "step": 4200
+    },
+    {
+      "epoch": 3.7859712230215825,
+      "grad_norm": 0.9160759449005127,
+      "learning_rate": 9.988903604124366e-05,
+      "loss": 0.0321,
+      "step": 4210
+    },
+    {
+      "action_loss": 0.009711216203868389,
+      "epoch": 3.7859712230215825,
+      "step": 4210
+    },
+    {
+      "epoch": 3.7859712230215825,
+      "step": 4210,
+      "torque_loss": 0.174884632229805
+    },
+    {
+      "epoch": 3.7859712230215825,
+      "force_loss": 0.0061207967810332775,
+      "step": 4210
+    },
+    {
+      "epoch": 3.7949640287769784,
+      "grad_norm": 0.9023813605308533,
+      "learning_rate": 9.988719351104343e-05,
+      "loss": 0.0331,
+      "step": 4220
+    },
+    {
+      "action_loss": 0.009800736792385578,
+      "epoch": 3.7949640287769784,
+      "step": 4220
+    },
+    {
+      "epoch": 3.7949640287769784,
+      "step": 4220,
+      "torque_loss": 0.14536048471927643
+    },
+    {
+      "epoch": 3.7949640287769784,
+      "force_loss": 0.008176678791642189,
+      "step": 4220
+    },
+    {
+      "epoch": 3.803956834532374,
+      "grad_norm": 1.2028447389602661,
+      "learning_rate": 9.9885335826431e-05,
+      "loss": 0.0324,
+      "step": 4230
+    },
+    {
+      "action_loss": 0.009005378931760788,
+      "epoch": 3.803956834532374,
+      "step": 4230
+    },
+    {
+      "epoch": 3.803956834532374,
+      "step": 4230,
+      "torque_loss": 0.15913419425487518
+    },
+    {
+      "epoch": 3.803956834532374,
+      "force_loss": 0.006897900719195604,
+      "step": 4230
+    },
+    {
+      "epoch": 3.81294964028777,
+      "grad_norm": 0.7526970505714417,
+      "learning_rate": 9.988346298797071e-05,
+      "loss": 0.03,
+      "step": 4240
+    },
+    {
+      "action_loss": 0.014618687331676483,
+      "epoch": 3.81294964028777,
+      "step": 4240
+    },
+    {
+      "epoch": 3.81294964028777,
+      "step": 4240,
+      "torque_loss": 0.13127528131008148
+    },
+    {
+      "epoch": 3.81294964028777,
+      "force_loss": 0.006682276260107756,
+      "step": 4240
+    },
+    {
+      "epoch": 3.8219424460431655,
+      "grad_norm": 2.0396652221679688,
+      "learning_rate": 9.988157499623146e-05,
+      "loss": 0.0416,
+      "step": 4250
+    },
+    {
+      "action_loss": 0.020371615886688232,
+      "epoch": 3.8219424460431655,
+      "step": 4250
+    },
+    {
+      "epoch": 3.8219424460431655,
+      "step": 4250,
+      "torque_loss": 0.1893899291753769
+    },
+    {
+      "epoch": 3.8219424460431655,
+      "force_loss": 0.012332126498222351,
+      "step": 4250
+    },
+    {
+      "epoch": 3.8309352517985613,
+      "grad_norm": 1.1619176864624023,
+      "learning_rate": 9.987967185178677e-05,
+      "loss": 0.0327,
+      "step": 4260
+    },
+    {
+      "action_loss": 0.01781490258872509,
+      "epoch": 3.8309352517985613,
+      "step": 4260
+    },
+    {
+      "epoch": 3.8309352517985613,
+      "step": 4260,
+      "torque_loss": 0.1966101974248886
+    },
+    {
+      "epoch": 3.8309352517985613,
+      "force_loss": 0.015079325996339321,
+      "step": 4260
+    },
+    {
+      "epoch": 3.8399280575539567,
+      "grad_norm": 0.9576718807220459,
+      "learning_rate": 9.987775355521476e-05,
+      "loss": 0.0387,
+      "step": 4270
+    },
+    {
+      "action_loss": 0.007879335433244705,
+      "epoch": 3.8399280575539567,
+      "step": 4270
+    },
+    {
+      "epoch": 3.8399280575539567,
+      "step": 4270,
+      "torque_loss": 0.12961077690124512
+    },
+    {
+      "epoch": 3.8399280575539567,
+      "force_loss": 0.008596367202699184,
+      "step": 4270
+    },
+    {
+      "epoch": 3.8489208633093526,
+      "grad_norm": 1.0311295986175537,
+      "learning_rate": 9.987582010709817e-05,
+      "loss": 0.0368,
+      "step": 4280
+    },
+    {
+      "action_loss": 0.056443434208631516,
+      "epoch": 3.8489208633093526,
+      "step": 4280
+    },
+    {
+      "epoch": 3.8489208633093526,
+      "step": 4280,
+      "torque_loss": 0.23857808113098145
+    },
+    {
+      "epoch": 3.8489208633093526,
+      "force_loss": 0.051936257630586624,
+      "step": 4280
+    },
+    {
+      "epoch": 3.8579136690647484,
+      "grad_norm": 0.3889097571372986,
+      "learning_rate": 9.987387150802431e-05,
+      "loss": 0.0491,
+      "step": 4290
+    },
+    {
+      "action_loss": 0.010186254046857357,
+      "epoch": 3.8579136690647484,
+      "step": 4290
+    },
+    {
+      "epoch": 3.8579136690647484,
+      "step": 4290,
+      "torque_loss": 0.11526137590408325
+    },
+    {
+      "epoch": 3.8579136690647484,
+      "force_loss": 0.00836224015802145,
+      "step": 4290
+    },
+    {
+      "epoch": 3.866906474820144,
+      "grad_norm": 0.7624739408493042,
+      "learning_rate": 9.987190775858517e-05,
+      "loss": 0.0328,
+      "step": 4300
+    },
+    {
+      "action_loss": 0.021824024617671967,
+      "epoch": 3.866906474820144,
+      "step": 4300
+    },
+    {
+      "epoch": 3.866906474820144,
+      "step": 4300,
+      "torque_loss": 0.11139372736215591
+    },
+    {
+      "epoch": 3.866906474820144,
+      "force_loss": 0.011579248122870922,
+      "step": 4300
+    },
+    {
+      "epoch": 3.8758992805755397,
+      "grad_norm": 0.6951278448104858,
+      "learning_rate": 9.98699288593772e-05,
+      "loss": 0.0351,
+      "step": 4310
+    },
+    {
+      "action_loss": 0.009745229966938496,
+      "epoch": 3.8758992805755397,
+      "step": 4310
+    },
+    {
+      "epoch": 3.8758992805755397,
+      "step": 4310,
+      "torque_loss": 0.1391468197107315
+    },
+    {
+      "epoch": 3.8758992805755397,
+      "force_loss": 0.01467133779078722,
+      "step": 4310
+    },
+    {
+      "epoch": 3.884892086330935,
+      "grad_norm": 1.9014315605163574,
+      "learning_rate": 9.986793481100161e-05,
+      "loss": 0.0376,
+      "step": 4320
+    },
+    {
+      "action_loss": 0.024018773809075356,
+      "epoch": 3.884892086330935,
+      "step": 4320
+    },
+    {
+      "epoch": 3.884892086330935,
+      "step": 4320,
+      "torque_loss": 0.13993610441684723
+    },
+    {
+      "epoch": 3.884892086330935,
+      "force_loss": 0.021737635135650635,
+      "step": 4320
+    },
+    {
+      "epoch": 3.893884892086331,
+      "grad_norm": 0.5066997408866882,
+      "learning_rate": 9.986592561406412e-05,
+      "loss": 0.035,
+      "step": 4330
+    },
+    {
+      "action_loss": 0.01049784291535616,
+      "epoch": 3.893884892086331,
+      "step": 4330
+    },
+    {
+      "epoch": 3.893884892086331,
+      "step": 4330,
+      "torque_loss": 0.16271261870861053
+    },
+    {
+      "epoch": 3.893884892086331,
+      "force_loss": 0.007432944606989622,
+      "step": 4330
+    },
+    {
+      "epoch": 3.902877697841727,
+      "grad_norm": 0.7239689826965332,
+      "learning_rate": 9.986390126917503e-05,
+      "loss": 0.0298,
+      "step": 4340
+    },
+    {
+      "action_loss": 0.01687888614833355,
+      "epoch": 3.902877697841727,
+      "step": 4340
+    },
+    {
+      "epoch": 3.902877697841727,
+      "step": 4340,
+      "torque_loss": 0.20288710296154022
+    },
+    {
+      "epoch": 3.902877697841727,
+      "force_loss": 0.010263652540743351,
+      "step": 4340
+    },
+    {
+      "epoch": 3.911870503597122,
+      "grad_norm": 0.8635382056236267,
+      "learning_rate": 9.986186177694933e-05,
+      "loss": 0.0307,
+      "step": 4350
+    },
+    {
+      "action_loss": 0.02397182583808899,
+      "epoch": 3.911870503597122,
+      "step": 4350
+    },
+    {
+      "epoch": 3.911870503597122,
+      "step": 4350,
+      "torque_loss": 0.14415116608142853
+    },
+    {
+      "epoch": 3.911870503597122,
+      "force_loss": 0.021520474925637245,
+      "step": 4350
+    },
+    {
+      "epoch": 3.920863309352518,
+      "grad_norm": 1.4948008060455322,
+      "learning_rate": 9.985980713800656e-05,
+      "loss": 0.0429,
+      "step": 4360
+    },
+    {
+      "action_loss": 0.008482703007757664,
+      "epoch": 3.920863309352518,
+      "step": 4360
+    },
+    {
+      "epoch": 3.920863309352518,
+      "step": 4360,
+      "torque_loss": 0.11190765351057053
+    },
+    {
+      "epoch": 3.920863309352518,
+      "force_loss": 0.008000750094652176,
+      "step": 4360
+    },
+    {
+      "epoch": 3.9298561151079134,
+      "grad_norm": 0.8221027255058289,
+      "learning_rate": 9.985773735297084e-05,
+      "loss": 0.0305,
+      "step": 4370
+    },
+    {
+      "action_loss": 0.014182105660438538,
+      "epoch": 3.9298561151079134,
+      "step": 4370
+    },
+    {
+      "epoch": 3.9298561151079134,
+      "step": 4370,
+      "torque_loss": 0.179058238863945
+    },
+    {
+      "epoch": 3.9298561151079134,
+      "force_loss": 0.010382688604295254,
+      "step": 4370
+    },
+    {
+      "epoch": 3.9388489208633093,
+      "grad_norm": 0.8402854204177856,
+      "learning_rate": 9.985565242247092e-05,
+      "loss": 0.0347,
+      "step": 4380
+    },
+    {
+      "action_loss": 0.01737954281270504,
+      "epoch": 3.9388489208633093,
+      "step": 4380
+    },
+    {
+      "epoch": 3.9388489208633093,
+      "step": 4380,
+      "torque_loss": 0.16029177606105804
+    },
+    {
+      "epoch": 3.9388489208633093,
+      "force_loss": 0.01366510521620512,
+      "step": 4380
+    },
+    {
+      "epoch": 3.947841726618705,
+      "grad_norm": 1.3627126216888428,
+      "learning_rate": 9.985355234714016e-05,
+      "loss": 0.0421,
+      "step": 4390
+    },
+    {
+      "action_loss": 0.009394989348948002,
+      "epoch": 3.947841726618705,
+      "step": 4390
+    },
+    {
+      "epoch": 3.947841726618705,
+      "step": 4390,
+      "torque_loss": 0.14305204153060913
+    },
+    {
+      "epoch": 3.947841726618705,
+      "force_loss": 0.009510352276265621,
+      "step": 4390
+    },
+    {
+      "epoch": 3.956834532374101,
+      "grad_norm": 1.1128345727920532,
+      "learning_rate": 9.985143712761652e-05,
+      "loss": 0.029,
+      "step": 4400
+    },
+    {
+      "action_loss": 0.011551998555660248,
+      "epoch": 3.956834532374101,
+      "step": 4400
+    },
+    {
+      "epoch": 3.956834532374101,
+      "step": 4400,
+      "torque_loss": 0.07601193338632584
+    },
+    {
+      "epoch": 3.956834532374101,
+      "force_loss": 0.007855924777686596,
+      "step": 4400
+    },
+    {
+      "epoch": 3.9658273381294964,
+      "grad_norm": 0.510873019695282,
+      "learning_rate": 9.984930676454252e-05,
+      "loss": 0.0339,
+      "step": 4410
+    },
+    {
+      "action_loss": 0.012548827566206455,
+      "epoch": 3.9658273381294964,
+      "step": 4410
+    },
+    {
+      "epoch": 3.9658273381294964,
+      "step": 4410,
+      "torque_loss": 0.10738100856542587
+    },
+    {
+      "epoch": 3.9658273381294964,
+      "force_loss": 0.007798769976943731,
+      "step": 4410
+    },
+    {
+      "epoch": 3.9748201438848922,
+      "grad_norm": 0.6574655771255493,
+      "learning_rate": 9.984716125856532e-05,
+      "loss": 0.0325,
+      "step": 4420
+    },
+    {
+      "action_loss": 0.00918615236878395,
+      "epoch": 3.9748201438848922,
+      "step": 4420
+    },
+    {
+      "epoch": 3.9748201438848922,
+      "step": 4420,
+      "torque_loss": 0.12600557506084442
+    },
+    {
+      "epoch": 3.9748201438848922,
+      "force_loss": 0.007819893769919872,
+      "step": 4420
+    },
+    {
+      "epoch": 3.9838129496402876,
+      "grad_norm": 0.6425493955612183,
+      "learning_rate": 9.984500061033667e-05,
+      "loss": 0.0293,
+      "step": 4430
+    },
+    {
+      "action_loss": 0.011316080577671528,
+      "epoch": 3.9838129496402876,
+      "step": 4430
+    },
+    {
+      "epoch": 3.9838129496402876,
+      "step": 4430,
+      "torque_loss": 0.12432371824979782
+    },
+    {
+      "epoch": 3.9838129496402876,
+      "force_loss": 0.007674422115087509,
+      "step": 4430
+    },
+    {
+      "epoch": 3.9928057553956835,
+      "grad_norm": 1.767102599143982,
+      "learning_rate": 9.984282482051293e-05,
+      "loss": 0.0342,
+      "step": 4440
+    },
+    {
+      "action_loss": 0.029515380039811134,
+      "epoch": 3.9928057553956835,
+      "step": 4440
+    },
+    {
+      "epoch": 3.9928057553956835,
+      "step": 4440,
+      "torque_loss": 0.20241832733154297
+    },
+    {
+      "epoch": 3.9928057553956835,
+      "force_loss": 0.03437669947743416,
+      "step": 4440
+    },
+    {
+      "epoch": 4.001798561151079,
+      "grad_norm": 1.1519076824188232,
+      "learning_rate": 9.9840633889755e-05,
+      "loss": 0.0349,
+      "step": 4450
+    },
+    {
+      "action_loss": 0.02478351630270481,
+      "epoch": 4.001798561151079,
+      "step": 4450
+    },
+    {
+      "epoch": 4.001798561151079,
+      "step": 4450,
+      "torque_loss": 0.22265885770320892
+    },
+    {
+      "epoch": 4.001798561151079,
+      "force_loss": 0.023469364270567894,
+      "step": 4450
+    },
+    {
+      "epoch": 4.010791366906475,
+      "grad_norm": 0.9709997177124023,
+      "learning_rate": 9.983842781872848e-05,
+      "loss": 0.0393,
+      "step": 4460
+    },
+    {
+      "action_loss": 0.022763893008232117,
+      "epoch": 4.010791366906475,
+      "step": 4460
+    },
+    {
+      "epoch": 4.010791366906475,
+      "step": 4460,
+      "torque_loss": 0.156441792845726
+    },
+    {
+      "epoch": 4.010791366906475,
+      "force_loss": 0.012009362690150738,
+      "step": 4460
+    },
+    {
+      "epoch": 4.01978417266187,
+      "grad_norm": 0.885554850101471,
+      "learning_rate": 9.98362066081035e-05,
+      "loss": 0.0374,
+      "step": 4470
+    },
+    {
+      "action_loss": 0.025641275569796562,
+      "epoch": 4.01978417266187,
+      "step": 4470
+    },
+    {
+      "epoch": 4.01978417266187,
+      "step": 4470,
+      "torque_loss": 0.16256771981716156
+    },
+    {
+      "epoch": 4.01978417266187,
+      "force_loss": 0.02282743901014328,
+      "step": 4470
+    },
+    {
+      "epoch": 4.028776978417266,
+      "grad_norm": 1.4418059587478638,
+      "learning_rate": 9.983397025855479e-05,
+      "loss": 0.032,
+      "step": 4480
+    },
+    {
+      "action_loss": 0.02729133330285549,
+      "epoch": 4.028776978417266,
+      "step": 4480
+    },
+    {
+      "epoch": 4.028776978417266,
+      "step": 4480,
+      "torque_loss": 0.15347303450107574
+    },
+    {
+      "epoch": 4.028776978417266,
+      "force_loss": 0.02187216840684414,
+      "step": 4480
+    },
+    {
+      "epoch": 4.037769784172662,
+      "grad_norm": 1.0959330797195435,
+      "learning_rate": 9.983171877076171e-05,
+      "loss": 0.034,
+      "step": 4490
+    },
+    {
+      "action_loss": 0.052611928433179855,
+      "epoch": 4.037769784172662,
+      "step": 4490
+    },
+    {
+      "epoch": 4.037769784172662,
+      "step": 4490,
+      "torque_loss": 0.18150770664215088
+    },
+    {
+      "epoch": 4.037769784172662,
+      "force_loss": 0.044015612453222275,
+      "step": 4490
+    },
+    {
+      "epoch": 4.046762589928058,
+      "grad_norm": 0.8289868831634521,
+      "learning_rate": 9.98294521454082e-05,
+      "loss": 0.0457,
+      "step": 4500
+    },
+    {
+      "action_loss": 0.008363536559045315,
+      "epoch": 4.046762589928058,
+      "step": 4500
+    },
+    {
+      "epoch": 4.046762589928058,
+      "step": 4500,
+      "torque_loss": 0.14328737556934357
+    },
+    {
+      "epoch": 4.046762589928058,
+      "force_loss": 0.010264570824801922,
+      "step": 4500
+    },
+    {
+      "epoch": 4.055755395683454,
+      "grad_norm": 1.583045482635498,
+      "learning_rate": 9.98271703831828e-05,
+      "loss": 0.0347,
+      "step": 4510
+    },
+    {
+      "action_loss": 0.045835670083761215,
+      "epoch": 4.055755395683454,
+      "step": 4510
+    },
+    {
+      "epoch": 4.055755395683454,
+      "step": 4510,
+      "torque_loss": 0.19038699567317963
+    },
+    {
+      "epoch": 4.055755395683454,
+      "force_loss": 0.032266441732645035,
+      "step": 4510
+    },
+    {
+      "epoch": 4.0647482014388485,
+      "grad_norm": 1.173993706703186,
+      "learning_rate": 9.982487348477865e-05,
+      "loss": 0.0466,
+      "step": 4520
+    },
+    {
+      "action_loss": 0.02522791363298893,
+      "epoch": 4.0647482014388485,
+      "step": 4520
+    },
+    {
+      "epoch": 4.0647482014388485,
+      "step": 4520,
+      "torque_loss": 0.12414098531007767
+    },
+    {
+      "epoch": 4.0647482014388485,
+      "force_loss": 0.02000529319047928,
+      "step": 4520
+    },
+    {
+      "epoch": 4.073741007194244,
+      "grad_norm": 0.7844947576522827,
+      "learning_rate": 9.982256145089347e-05,
+      "loss": 0.0371,
+      "step": 4530
+    },
+    {
+      "action_loss": 0.009618385694921017,
+      "epoch": 4.073741007194244,
+      "step": 4530
+    },
+    {
+      "epoch": 4.073741007194244,
+      "step": 4530,
+      "torque_loss": 0.10716839879751205
+    },
+    {
+      "epoch": 4.073741007194244,
+      "force_loss": 0.008798129856586456,
+      "step": 4530
+    },
+    {
+      "epoch": 4.08273381294964,
+      "grad_norm": 1.02362859249115,
+      "learning_rate": 9.982023428222962e-05,
+      "loss": 0.0323,
+      "step": 4540
+    },
+    {
+      "action_loss": 0.023549756035208702,
+      "epoch": 4.08273381294964,
+      "step": 4540
+    },
+    {
+      "epoch": 4.08273381294964,
+      "step": 4540,
+      "torque_loss": 0.19426637887954712
+    },
+    {
+      "epoch": 4.08273381294964,
+      "force_loss": 0.014226471073925495,
+      "step": 4540
+    },
+    {
+      "epoch": 4.091726618705036,
+      "grad_norm": 0.9992864727973938,
+      "learning_rate": 9.981789197949403e-05,
+      "loss": 0.0407,
+      "step": 4550
+    },
+    {
+      "action_loss": 0.02116341143846512,
+      "epoch": 4.091726618705036,
+      "step": 4550
+    },
+    {
+      "epoch": 4.091726618705036,
+      "step": 4550,
+      "torque_loss": 0.15985850989818573
+    },
+    {
+      "epoch": 4.091726618705036,
+      "force_loss": 0.02765163965523243,
+      "step": 4550
+    },
+    {
+      "epoch": 4.100719424460432,
+      "grad_norm": 0.8838006854057312,
+      "learning_rate": 9.98155345433982e-05,
+      "loss": 0.0352,
+      "step": 4560
+    },
+    {
+      "action_loss": 0.0556286983191967,
+      "epoch": 4.100719424460432,
+      "step": 4560
+    },
+    {
+      "epoch": 4.100719424460432,
+      "step": 4560,
+      "torque_loss": 0.17061519622802734
+    },
+    {
+      "epoch": 4.100719424460432,
+      "force_loss": 0.03908475860953331,
+      "step": 4560
+    },
+    {
+      "epoch": 4.109712230215828,
+      "grad_norm": 0.9535768628120422,
+      "learning_rate": 9.981316197465831e-05,
+      "loss": 0.0341,
+      "step": 4570
+    },
+    {
+      "action_loss": 0.018756555393338203,
+      "epoch": 4.109712230215828,
+      "step": 4570
+    },
+    {
+      "epoch": 4.109712230215828,
+      "step": 4570,
+      "torque_loss": 0.14983071386814117
+    },
+    {
+      "epoch": 4.109712230215828,
+      "force_loss": 0.018597392365336418,
+      "step": 4570
+    },
+    {
+      "epoch": 4.118705035971223,
+      "grad_norm": 0.9701385498046875,
+      "learning_rate": 9.981077427399504e-05,
+      "loss": 0.0369,
+      "step": 4580
+    },
+    {
+      "action_loss": 0.016903484240174294,
+      "epoch": 4.118705035971223,
+      "step": 4580
+    },
+    {
+      "epoch": 4.118705035971223,
+      "step": 4580,
+      "torque_loss": 0.13766567409038544
+    },
+    {
+      "epoch": 4.118705035971223,
+      "force_loss": 0.009743332862854004,
+      "step": 4580
+    },
+    {
+      "epoch": 4.127697841726619,
+      "grad_norm": 0.6447258591651917,
+      "learning_rate": 9.980837144213371e-05,
+      "loss": 0.0316,
+      "step": 4590
+    },
+    {
+      "action_loss": 0.01526044774800539,
+      "epoch": 4.127697841726619,
+      "step": 4590
+    },
+    {
+      "epoch": 4.127697841726619,
+      "step": 4590,
+      "torque_loss": 0.12720173597335815
+    },
+    {
+      "epoch": 4.127697841726619,
+      "force_loss": 0.013995117507874966,
+      "step": 4590
+    },
+    {
+      "epoch": 4.136690647482014,
+      "grad_norm": 0.5802412033081055,
+      "learning_rate": 9.980595347980426e-05,
+      "loss": 0.0311,
+      "step": 4600
+    },
+    {
+      "action_loss": 0.02217998541891575,
+      "epoch": 4.136690647482014,
+      "step": 4600
+    },
+    {
+      "epoch": 4.136690647482014,
+      "step": 4600,
+      "torque_loss": 0.09590903669595718
+    },
+    {
+      "epoch": 4.136690647482014,
+      "force_loss": 0.016717659309506416,
+      "step": 4600
+    },
+    {
+      "epoch": 4.14568345323741,
+      "grad_norm": 0.8760179281234741,
+      "learning_rate": 9.980352038774119e-05,
+      "loss": 0.0419,
+      "step": 4610
+    },
+    {
+      "action_loss": 0.014122466556727886,
+      "epoch": 4.14568345323741,
+      "step": 4610
+    },
+    {
+      "epoch": 4.14568345323741,
+      "step": 4610,
+      "torque_loss": 0.11563873291015625
+    },
+    {
+      "epoch": 4.14568345323741,
+      "force_loss": 0.007175022270530462,
+      "step": 4610
+    },
+    {
+      "epoch": 4.154676258992806,
+      "grad_norm": 0.5086389183998108,
+      "learning_rate": 9.98010721666836e-05,
+      "loss": 0.0368,
+      "step": 4620
+    },
+    {
+      "action_loss": 0.010292047634720802,
+      "epoch": 4.154676258992806,
+      "step": 4620
+    },
+    {
+      "epoch": 4.154676258992806,
+      "step": 4620,
+      "torque_loss": 0.14402835071086884
+    },
+    {
+      "epoch": 4.154676258992806,
+      "force_loss": 0.018386662006378174,
+      "step": 4620
+    },
+    {
+      "epoch": 4.163669064748201,
+      "grad_norm": 0.776138186454773,
+      "learning_rate": 9.979860881737523e-05,
+      "loss": 0.0281,
+      "step": 4630
+    },
+    {
+      "action_loss": 0.007989101111888885,
+      "epoch": 4.163669064748201,
+      "step": 4630
+    },
+    {
+      "epoch": 4.163669064748201,
+      "step": 4630,
+      "torque_loss": 0.12181323766708374
+    },
+    {
+      "epoch": 4.163669064748201,
+      "force_loss": 0.005654403939843178,
+      "step": 4630
+    },
+    {
+      "epoch": 4.172661870503597,
+      "grad_norm": 0.9099301099777222,
+      "learning_rate": 9.979613034056434e-05,
+      "loss": 0.0417,
+      "step": 4640
+    },
+    {
+      "action_loss": 0.012895102612674236,
+      "epoch": 4.172661870503597,
+      "step": 4640
+    },
+    {
+      "epoch": 4.172661870503597,
+      "step": 4640,
+      "torque_loss": 0.12507624924182892
+    },
+    {
+      "epoch": 4.172661870503597,
+      "force_loss": 0.00915098562836647,
+      "step": 4640
+    },
+    {
+      "epoch": 4.181654676258993,
+      "grad_norm": 0.9602915644645691,
+      "learning_rate": 9.979363673700386e-05,
+      "loss": 0.0369,
+      "step": 4650
+    },
+    {
+      "action_loss": 0.01851240172982216,
+      "epoch": 4.181654676258993,
+      "step": 4650
+    },
+    {
+      "epoch": 4.181654676258993,
+      "step": 4650,
+      "torque_loss": 0.12963305413722992
+    },
+    {
+      "epoch": 4.181654676258993,
+      "force_loss": 0.0167622659355402,
+      "step": 4650
+    },
+    {
+      "epoch": 4.190647482014389,
+      "grad_norm": 1.073164701461792,
+      "learning_rate": 9.979112800745124e-05,
+      "loss": 0.0441,
+      "step": 4660
+    },
+    {
+      "action_loss": 0.03333734720945358,
+      "epoch": 4.190647482014389,
+      "step": 4660
+    },
+    {
+      "epoch": 4.190647482014389,
+      "step": 4660,
+      "torque_loss": 0.14169570803642273
+    },
+    {
+      "epoch": 4.190647482014389,
+      "force_loss": 0.022027716040611267,
+      "step": 4660
+    },
+    {
+      "epoch": 4.1996402877697845,
+      "grad_norm": 0.4481082558631897,
+      "learning_rate": 9.978860415266861e-05,
+      "loss": 0.0384,
+      "step": 4670
+    },
+    {
+      "action_loss": 0.020381757989525795,
+      "epoch": 4.1996402877697845,
+      "step": 4670
+    },
+    {
+      "epoch": 4.1996402877697845,
+      "step": 4670,
+      "torque_loss": 0.14372999966144562
+    },
+    {
+      "epoch": 4.1996402877697845,
+      "force_loss": 0.015198520384728909,
+      "step": 4670
+    },
+    {
+      "epoch": 4.2086330935251794,
+      "grad_norm": 0.9679866433143616,
+      "learning_rate": 9.978606517342262e-05,
+      "loss": 0.0357,
+      "step": 4680
+    },
+    {
+      "action_loss": 0.01251304429024458,
+      "epoch": 4.2086330935251794,
+      "step": 4680
+    },
+    {
+      "epoch": 4.2086330935251794,
+      "step": 4680,
+      "torque_loss": 0.1499743014574051
+    },
+    {
+      "epoch": 4.2086330935251794,
+      "force_loss": 0.010061842389404774,
+      "step": 4680
+    },
+    {
+      "epoch": 4.217625899280575,
+      "grad_norm": 0.41633597016334534,
+      "learning_rate": 9.978351107048456e-05,
+      "loss": 0.0399,
+      "step": 4690
+    },
+    {
+      "action_loss": 0.01703997142612934,
+      "epoch": 4.217625899280575,
+      "step": 4690
+    },
+    {
+      "epoch": 4.217625899280575,
+      "step": 4690,
+      "torque_loss": 0.16055503487586975
+    },
+    {
+      "epoch": 4.217625899280575,
+      "force_loss": 0.02300129272043705,
+      "step": 4690
+    },
+    {
+      "epoch": 4.226618705035971,
+      "grad_norm": 1.6063692569732666,
+      "learning_rate": 9.978094184463029e-05,
+      "loss": 0.0459,
+      "step": 4700
+    },
+    {
+      "action_loss": 0.015734244138002396,
+      "epoch": 4.226618705035971,
+      "step": 4700
+    },
+    {
+      "epoch": 4.226618705035971,
+      "step": 4700,
+      "torque_loss": 0.21210801601409912
+    },
+    {
+      "epoch": 4.226618705035971,
+      "force_loss": 0.016246922314167023,
+      "step": 4700
+    },
+    {
+      "epoch": 4.235611510791367,
+      "grad_norm": 0.6652126908302307,
+      "learning_rate": 9.977835749664029e-05,
+      "loss": 0.0305,
+      "step": 4710
+    },
+    {
+      "action_loss": 0.017241256311535835,
+      "epoch": 4.235611510791367,
+      "step": 4710
+    },
+    {
+      "epoch": 4.235611510791367,
+      "step": 4710,
+      "torque_loss": 0.13138684630393982
+    },
+    {
+      "epoch": 4.235611510791367,
+      "force_loss": 0.009017406962811947,
+      "step": 4710
+    },
+    {
+      "epoch": 4.244604316546763,
+      "grad_norm": 0.9250365495681763,
+      "learning_rate": 9.97757580272996e-05,
+      "loss": 0.0323,
+      "step": 4720
+    },
+    {
+      "action_loss": 0.01591983251273632,
+      "epoch": 4.244604316546763,
+      "step": 4720
+    },
+    {
+      "epoch": 4.244604316546763,
+      "step": 4720,
+      "torque_loss": 0.15921808779239655
+    },
+    {
+      "epoch": 4.244604316546763,
+      "force_loss": 0.00837093684822321,
+      "step": 4720
+    },
+    {
+      "epoch": 4.253597122302159,
+      "grad_norm": 0.5056507587432861,
+      "learning_rate": 9.977314343739786e-05,
+      "loss": 0.0292,
+      "step": 4730
+    },
+    {
+      "action_loss": 0.032968662679195404,
+      "epoch": 4.253597122302159,
+      "step": 4730
+    },
+    {
+      "epoch": 4.253597122302159,
+      "step": 4730,
+      "torque_loss": 0.1851831078529358
+    },
+    {
+      "epoch": 4.253597122302159,
+      "force_loss": 0.03455134853720665,
+      "step": 4730
+    },
+    {
+      "epoch": 4.262589928057554,
+      "grad_norm": 0.6013080477714539,
+      "learning_rate": 9.977051372772934e-05,
+      "loss": 0.0392,
+      "step": 4740
+    },
+    {
+      "action_loss": 0.014050872065126896,
+      "epoch": 4.262589928057554,
+      "step": 4740
+    },
+    {
+      "epoch": 4.262589928057554,
+      "step": 4740,
+      "torque_loss": 0.1577189713716507
+    },
+    {
+      "epoch": 4.262589928057554,
+      "force_loss": 0.012093194760382175,
+      "step": 4740
+    },
+    {
+      "epoch": 4.2715827338129495,
+      "grad_norm": 1.4672819375991821,
+      "learning_rate": 9.976786889909286e-05,
+      "loss": 0.0288,
+      "step": 4750
+    },
+    {
+      "action_loss": 0.015345960855484009,
+      "epoch": 4.2715827338129495,
+      "step": 4750
+    },
+    {
+      "epoch": 4.2715827338129495,
+      "step": 4750,
+      "torque_loss": 0.14214837551116943
+    },
+    {
+      "epoch": 4.2715827338129495,
+      "force_loss": 0.012158502824604511,
+      "step": 4750
+    },
+    {
+      "epoch": 4.280575539568345,
+      "grad_norm": 0.8512645363807678,
+      "learning_rate": 9.976520895229185e-05,
+      "loss": 0.0337,
+      "step": 4760
+    },
+    {
+      "action_loss": 0.018188422545790672,
+      "epoch": 4.280575539568345,
+      "step": 4760
+    },
+    {
+      "epoch": 4.280575539568345,
+      "step": 4760,
+      "torque_loss": 0.13265514373779297
+    },
+    {
+      "epoch": 4.280575539568345,
+      "force_loss": 0.012605768628418446,
+      "step": 4760
+    },
+    {
+      "epoch": 4.289568345323741,
+      "grad_norm": 0.6212011575698853,
+      "learning_rate": 9.976253388813433e-05,
+      "loss": 0.0386,
+      "step": 4770
+    },
+    {
+      "action_loss": 0.014844626188278198,
+      "epoch": 4.289568345323741,
+      "step": 4770
+    },
+    {
+      "epoch": 4.289568345323741,
+      "step": 4770,
+      "torque_loss": 0.13482101261615753
+    },
+    {
+      "epoch": 4.289568345323741,
+      "force_loss": 0.015131277032196522,
+      "step": 4770
+    },
+    {
+      "epoch": 4.298561151079137,
+      "grad_norm": 0.7584494352340698,
+      "learning_rate": 9.975984370743293e-05,
+      "loss": 0.0381,
+      "step": 4780
+    },
+    {
+      "action_loss": 0.014859539456665516,
+      "epoch": 4.298561151079137,
+      "step": 4780
+    },
+    {
+      "epoch": 4.298561151079137,
+      "step": 4780,
+      "torque_loss": 0.17865495383739471
+    },
+    {
+      "epoch": 4.298561151079137,
+      "force_loss": 0.014067028649151325,
+      "step": 4780
+    },
+    {
+      "epoch": 4.307553956834532,
+      "grad_norm": 1.149087905883789,
+      "learning_rate": 9.975713841100485e-05,
+      "loss": 0.0468,
+      "step": 4790
+    },
+    {
+      "action_loss": 0.010169833898544312,
+      "epoch": 4.307553956834532,
+      "step": 4790
+    },
+    {
+      "epoch": 4.307553956834532,
+      "step": 4790,
+      "torque_loss": 0.14036309719085693
+    },
+    {
+      "epoch": 4.307553956834532,
+      "force_loss": 0.010213927365839481,
+      "step": 4790
+    },
+    {
+      "epoch": 4.316546762589928,
+      "grad_norm": 0.916983962059021,
+      "learning_rate": 9.975441799967187e-05,
+      "loss": 0.0324,
+      "step": 4800
+    },
+    {
+      "action_loss": 0.016165580600500107,
+      "epoch": 4.316546762589928,
+      "step": 4800
+    },
+    {
+      "epoch": 4.316546762589928,
+      "step": 4800,
+      "torque_loss": 0.10830014944076538
+    },
+    {
+      "epoch": 4.316546762589928,
+      "force_loss": 0.01386170368641615,
+      "step": 4800
+    },
+    {
+      "epoch": 4.325539568345324,
+      "grad_norm": 0.9218096137046814,
+      "learning_rate": 9.975168247426039e-05,
+      "loss": 0.0344,
+      "step": 4810
+    },
+    {
+      "action_loss": 0.016445549204945564,
+      "epoch": 4.325539568345324,
+      "step": 4810
+    },
+    {
+      "epoch": 4.325539568345324,
+      "step": 4810,
+      "torque_loss": 0.10059714317321777
+    },
+    {
+      "epoch": 4.325539568345324,
+      "force_loss": 0.012037478387355804,
+      "step": 4810
+    },
+    {
+      "epoch": 4.33453237410072,
+      "grad_norm": 1.8911771774291992,
+      "learning_rate": 9.974893183560139e-05,
+      "loss": 0.0306,
+      "step": 4820
+    },
+    {
+      "action_loss": 0.017487460747361183,
+      "epoch": 4.33453237410072,
+      "step": 4820
+    },
+    {
+      "epoch": 4.33453237410072,
+      "step": 4820,
+      "torque_loss": 0.1674678772687912
+    },
+    {
+      "epoch": 4.33453237410072,
+      "force_loss": 0.021262051537632942,
+      "step": 4820
+    },
+    {
+      "epoch": 4.343525179856115,
+      "grad_norm": 0.8325040340423584,
+      "learning_rate": 9.974616608453045e-05,
+      "loss": 0.0337,
+      "step": 4830
+    },
+    {
+      "action_loss": 0.01841500587761402,
+      "epoch": 4.343525179856115,
+      "step": 4830
+    },
+    {
+      "epoch": 4.343525179856115,
+      "step": 4830,
+      "torque_loss": 0.14614729583263397
+    },
+    {
+      "epoch": 4.343525179856115,
+      "force_loss": 0.022032881155610085,
+      "step": 4830
+    },
+    {
+      "epoch": 4.35251798561151,
+      "grad_norm": 1.0409008264541626,
+      "learning_rate": 9.974338522188772e-05,
+      "loss": 0.0332,
+      "step": 4840
+    },
+    {
+      "action_loss": 0.014770823530852795,
+      "epoch": 4.35251798561151,
+      "step": 4840
+    },
+    {
+      "epoch": 4.35251798561151,
+      "step": 4840,
+      "torque_loss": 0.1954604834318161
+    },
+    {
+      "epoch": 4.35251798561151,
+      "force_loss": 0.01183202862739563,
+      "step": 4840
+    },
+    {
+      "epoch": 4.361510791366906,
+      "grad_norm": 1.3001381158828735,
+      "learning_rate": 9.974058924851797e-05,
+      "loss": 0.0319,
+      "step": 4850
+    },
+    {
+      "action_loss": 0.02393859624862671,
+      "epoch": 4.361510791366906,
+      "step": 4850
+    },
+    {
+      "epoch": 4.361510791366906,
+      "step": 4850,
+      "torque_loss": 0.16679538786411285
+    },
+    {
+      "epoch": 4.361510791366906,
+      "force_loss": 0.02487037144601345,
+      "step": 4850
+    },
+    {
+      "epoch": 4.370503597122302,
+      "grad_norm": 1.2244654893875122,
+      "learning_rate": 9.973777816527051e-05,
+      "loss": 0.0361,
+      "step": 4860
+    },
+    {
+      "action_loss": 0.04007313400506973,
+      "epoch": 4.370503597122302,
+      "step": 4860
+    },
+    {
+      "epoch": 4.370503597122302,
+      "step": 4860,
+      "torque_loss": 0.1552257388830185
+    },
+    {
+      "epoch": 4.370503597122302,
+      "force_loss": 0.03166845440864563,
+      "step": 4860
+    },
+    {
+      "epoch": 4.379496402877698,
+      "grad_norm": 0.8373429179191589,
+      "learning_rate": 9.973495197299931e-05,
+      "loss": 0.0366,
+      "step": 4870
+    },
+    {
+      "action_loss": 0.007736609783023596,
+      "epoch": 4.379496402877698,
+      "step": 4870
+    },
+    {
+      "epoch": 4.379496402877698,
+      "step": 4870,
+      "torque_loss": 0.09443923085927963
+    },
+    {
+      "epoch": 4.379496402877698,
+      "force_loss": 0.01111941784620285,
+      "step": 4870
+    },
+    {
+      "epoch": 4.388489208633094,
+      "grad_norm": 0.5578615069389343,
+      "learning_rate": 9.973211067256287e-05,
+      "loss": 0.0338,
+      "step": 4880
+    },
+    {
+      "action_loss": 0.01750328578054905,
+      "epoch": 4.388489208633094,
+      "step": 4880
+    },
+    {
+      "epoch": 4.388489208633094,
+      "step": 4880,
+      "torque_loss": 0.18886463344097137
+    },
+    {
+      "epoch": 4.388489208633094,
+      "force_loss": 0.016042115166783333,
+      "step": 4880
+    },
+    {
+      "epoch": 4.39748201438849,
+      "grad_norm": 1.8016473054885864,
+      "learning_rate": 9.97292542648243e-05,
+      "loss": 0.0381,
+      "step": 4890
+    },
+    {
+      "action_loss": 0.010596145875751972,
+      "epoch": 4.39748201438849,
+      "step": 4890
+    },
+    {
+      "epoch": 4.39748201438849,
+      "step": 4890,
+      "torque_loss": 0.10433429479598999
+    },
+    {
+      "epoch": 4.39748201438849,
+      "force_loss": 0.006710140500217676,
+      "step": 4890
+    },
+    {
+      "epoch": 4.406474820143885,
+      "grad_norm": 1.21533203125,
+      "learning_rate": 9.972638275065131e-05,
+      "loss": 0.0311,
+      "step": 4900
+    },
+    {
+      "action_loss": 0.03270336613059044,
+      "epoch": 4.406474820143885,
+      "step": 4900
+    },
+    {
+      "epoch": 4.406474820143885,
+      "step": 4900,
+      "torque_loss": 0.1301143318414688
+    },
+    {
+      "epoch": 4.406474820143885,
+      "force_loss": 0.0368109829723835,
+      "step": 4900
+    },
+    {
+      "epoch": 4.41546762589928,
+      "grad_norm": 0.7246043682098389,
+      "learning_rate": 9.972349613091621e-05,
+      "loss": 0.0358,
+      "step": 4910
+    },
+    {
+      "action_loss": 0.01927509531378746,
+      "epoch": 4.41546762589928,
+      "step": 4910
+    },
+    {
+      "epoch": 4.41546762589928,
+      "step": 4910,
+      "torque_loss": 0.1635270118713379
+    },
+    {
+      "epoch": 4.41546762589928,
+      "force_loss": 0.011132900603115559,
+      "step": 4910
+    },
+    {
+      "epoch": 4.424460431654676,
+      "grad_norm": 1.5251998901367188,
+      "learning_rate": 9.972059440649584e-05,
+      "loss": 0.0402,
+      "step": 4920
+    },
+    {
+      "action_loss": 0.01019121054559946,
+      "epoch": 4.424460431654676,
+      "step": 4920
+    },
+    {
+      "epoch": 4.424460431654676,
+      "step": 4920,
+      "torque_loss": 0.12546344101428986
+    },
+    {
+      "epoch": 4.424460431654676,
+      "force_loss": 0.008014569990336895,
+      "step": 4920
+    },
+    {
+      "epoch": 4.433453237410072,
+      "grad_norm": 0.8194612264633179,
+      "learning_rate": 9.971767757827168e-05,
+      "loss": 0.032,
+      "step": 4930
+    },
+    {
+      "action_loss": 0.024361327290534973,
+      "epoch": 4.433453237410072,
+      "step": 4930
+    },
+    {
+      "epoch": 4.433453237410072,
+      "step": 4930,
+      "torque_loss": 0.11441129446029663
+    },
+    {
+      "epoch": 4.433453237410072,
+      "force_loss": 0.022045262157917023,
+      "step": 4930
+    },
+    {
+      "epoch": 4.442446043165468,
+      "grad_norm": 1.2900441884994507,
+      "learning_rate": 9.971474564712982e-05,
+      "loss": 0.0359,
+      "step": 4940
+    },
+    {
+      "action_loss": 0.016533298417925835,
+      "epoch": 4.442446043165468,
+      "step": 4940
+    },
+    {
+      "epoch": 4.442446043165468,
+      "step": 4940,
+      "torque_loss": 0.15621978044509888
+    },
+    {
+      "epoch": 4.442446043165468,
+      "force_loss": 0.011242390610277653,
+      "step": 4940
+    },
+    {
+      "epoch": 4.451438848920863,
+      "grad_norm": 0.8452239632606506,
+      "learning_rate": 9.971179861396084e-05,
+      "loss": 0.0329,
+      "step": 4950
+    },
+    {
+      "action_loss": 0.024112576618790627,
+      "epoch": 4.451438848920863,
+      "step": 4950
+    },
+    {
+      "epoch": 4.451438848920863,
+      "step": 4950,
+      "torque_loss": 0.12866103649139404
+    },
+    {
+      "epoch": 4.451438848920863,
+      "force_loss": 0.018411686643958092,
+      "step": 4950
+    },
+    {
+      "epoch": 4.460431654676259,
+      "grad_norm": 0.6208613514900208,
+      "learning_rate": 9.970883647966003e-05,
+      "loss": 0.0326,
+      "step": 4960
+    },
+    {
+      "action_loss": 0.036774277687072754,
+      "epoch": 4.460431654676259,
+      "step": 4960
+    },
+    {
+      "epoch": 4.460431654676259,
+      "step": 4960,
+      "torque_loss": 0.23626530170440674
+    },
+    {
+      "epoch": 4.460431654676259,
+      "force_loss": 0.01701635681092739,
+      "step": 4960
+    },
+    {
+      "epoch": 4.469424460431655,
+      "grad_norm": 0.982664167881012,
+      "learning_rate": 9.970585924512717e-05,
+      "loss": 0.0345,
+      "step": 4970
+    },
+    {
+      "action_loss": 0.016286931931972504,
+      "epoch": 4.469424460431655,
+      "step": 4970
+    },
+    {
+      "epoch": 4.469424460431655,
+      "step": 4970,
+      "torque_loss": 0.1323186755180359
+    },
+    {
+      "epoch": 4.469424460431655,
+      "force_loss": 0.017630256712436676,
+      "step": 4970
+    },
+    {
+      "epoch": 4.4784172661870505,
+      "grad_norm": 0.6811734437942505,
+      "learning_rate": 9.970286691126669e-05,
+      "loss": 0.033,
+      "step": 4980
+    },
+    {
+      "action_loss": 0.0160505548119545,
+      "epoch": 4.4784172661870505,
+      "step": 4980
+    },
+    {
+      "epoch": 4.4784172661870505,
+      "step": 4980,
+      "torque_loss": 0.12203121185302734
+    },
+    {
+      "epoch": 4.4784172661870505,
+      "force_loss": 0.007708249147981405,
+      "step": 4980
+    },
+    {
+      "epoch": 4.487410071942446,
+      "grad_norm": 0.8383340239524841,
+      "learning_rate": 9.969985947898756e-05,
+      "loss": 0.0386,
+      "step": 4990
+    },
+    {
+      "action_loss": 0.03993428871035576,
+      "epoch": 4.487410071942446,
+      "step": 4990
+    },
+    {
+      "epoch": 4.487410071942446,
+      "step": 4990,
+      "torque_loss": 0.20964229106903076
+    },
+    {
+      "epoch": 4.487410071942446,
+      "force_loss": 0.014576214365661144,
+      "step": 4990
+    },
+    {
+      "epoch": 4.496402877697841,
+      "grad_norm": 0.8964179754257202,
+      "learning_rate": 9.969683694920337e-05,
+      "loss": 0.0396,
+      "step": 5000
+    },
+    {
+      "action_loss": 0.03832182660698891,
+      "epoch": 4.496402877697841,
+      "step": 5000
+    },
+    {
+      "epoch": 4.496402877697841,
+      "step": 5000,
+      "torque_loss": 0.1555413454771042
+    },
+    {
+      "epoch": 4.496402877697841,
+      "force_loss": 0.02865142934024334,
+      "step": 5000
+    },
+    {
+      "epoch": 4.505395683453237,
+      "grad_norm": 0.6202208399772644,
+      "learning_rate": 9.969379932283228e-05,
+      "loss": 0.0387,
+      "step": 5010
+    },
+    {
+      "action_loss": 0.012801670469343662,
+      "epoch": 4.505395683453237,
+      "step": 5010
+    },
+    {
+      "epoch": 4.505395683453237,
+      "step": 5010,
+      "torque_loss": 0.15284575521945953
+    },
+    {
+      "epoch": 4.505395683453237,
+      "force_loss": 0.01691793091595173,
+      "step": 5010
+    },
+    {
+      "epoch": 4.514388489208633,
+      "grad_norm": 0.9778483510017395,
+      "learning_rate": 9.969074660079704e-05,
+      "loss": 0.0351,
+      "step": 5020
+    },
+    {
+      "action_loss": 0.045336682349443436,
+      "epoch": 4.514388489208633,
+      "step": 5020
+    },
+    {
+      "epoch": 4.514388489208633,
+      "step": 5020,
+      "torque_loss": 0.21425507962703705
+    },
+    {
+      "epoch": 4.514388489208633,
+      "force_loss": 0.04962506890296936,
+      "step": 5020
+    },
+    {
+      "epoch": 4.523381294964029,
+      "grad_norm": 1.2425585985183716,
+      "learning_rate": 9.968767878402501e-05,
+      "loss": 0.0371,
+      "step": 5030
+    },
+    {
+      "action_loss": 0.011635159142315388,
+      "epoch": 4.523381294964029,
+      "step": 5030
+    },
+    {
+      "epoch": 4.523381294964029,
+      "step": 5030,
+      "torque_loss": 0.11857563257217407
+    },
+    {
+      "epoch": 4.523381294964029,
+      "force_loss": 0.012137492187321186,
+      "step": 5030
+    },
+    {
+      "epoch": 4.532374100719425,
+      "grad_norm": 0.9975318312644958,
+      "learning_rate": 9.968459587344808e-05,
+      "loss": 0.0305,
+      "step": 5040
+    },
+    {
+      "action_loss": 0.012168730609118938,
+      "epoch": 4.532374100719425,
+      "step": 5040
+    },
+    {
+      "epoch": 4.532374100719425,
+      "step": 5040,
+      "torque_loss": 0.1289518177509308
+    },
+    {
+      "epoch": 4.532374100719425,
+      "force_loss": 0.007105143740773201,
+      "step": 5040
+    },
+    {
+      "epoch": 4.5413669064748206,
+      "grad_norm": 0.7822269201278687,
+      "learning_rate": 9.968149787000278e-05,
+      "loss": 0.0339,
+      "step": 5050
+    },
+    {
+      "action_loss": 0.010288389399647713,
+      "epoch": 4.5413669064748206,
+      "step": 5050
+    },
+    {
+      "epoch": 4.5413669064748206,
+      "step": 5050,
+      "torque_loss": 0.16448652744293213
+    },
+    {
+      "epoch": 4.5413669064748206,
+      "force_loss": 0.007236870471388102,
+      "step": 5050
+    },
+    {
+      "epoch": 4.5503597122302155,
+      "grad_norm": 0.6134659647941589,
+      "learning_rate": 9.967838477463018e-05,
+      "loss": 0.0282,
+      "step": 5060
+    },
+    {
+      "action_loss": 0.008654690347611904,
+      "epoch": 4.5503597122302155,
+      "step": 5060
+    },
+    {
+      "epoch": 4.5503597122302155,
+      "step": 5060,
+      "torque_loss": 0.1068754494190216
+    },
+    {
+      "epoch": 4.5503597122302155,
+      "force_loss": 0.00835444126278162,
+      "step": 5060
+    },
+    {
+      "epoch": 4.559352517985611,
+      "grad_norm": 1.0853147506713867,
+      "learning_rate": 9.967525658827597e-05,
+      "loss": 0.0328,
+      "step": 5070
+    },
+    {
+      "action_loss": 0.017858320847153664,
+      "epoch": 4.559352517985611,
+      "step": 5070
+    },
+    {
+      "epoch": 4.559352517985611,
+      "step": 5070,
+      "torque_loss": 0.10284539312124252
+    },
+    {
+      "epoch": 4.559352517985611,
+      "force_loss": 0.019653117284178734,
+      "step": 5070
+    },
+    {
+      "epoch": 4.568345323741007,
+      "grad_norm": 1.014072299003601,
+      "learning_rate": 9.967211331189042e-05,
+      "loss": 0.0349,
+      "step": 5080
+    },
+    {
+      "action_loss": 0.01476411521434784,
+      "epoch": 4.568345323741007,
+      "step": 5080
+    },
+    {
+      "epoch": 4.568345323741007,
+      "step": 5080,
+      "torque_loss": 0.12568867206573486
+    },
+    {
+      "epoch": 4.568345323741007,
+      "force_loss": 0.010045386850833893,
+      "step": 5080
+    },
+    {
+      "epoch": 4.577338129496403,
+      "grad_norm": 1.1342768669128418,
+      "learning_rate": 9.966895494642834e-05,
+      "loss": 0.0357,
+      "step": 5090
+    },
+    {
+      "action_loss": 0.015453745611011982,
+      "epoch": 4.577338129496403,
+      "step": 5090
+    },
+    {
+      "epoch": 4.577338129496403,
+      "step": 5090,
+      "torque_loss": 0.1701519638299942
+    },
+    {
+      "epoch": 4.577338129496403,
+      "force_loss": 0.009401858784258366,
+      "step": 5090
+    },
+    {
+      "epoch": 4.586330935251799,
+      "grad_norm": 0.6681179404258728,
+      "learning_rate": 9.96657814928492e-05,
+      "loss": 0.0402,
+      "step": 5100
+    },
+    {
+      "action_loss": 0.01565312221646309,
+      "epoch": 4.586330935251799,
+      "step": 5100
+    },
+    {
+      "epoch": 4.586330935251799,
+      "step": 5100,
+      "torque_loss": 0.13380496203899384
+    },
+    {
+      "epoch": 4.586330935251799,
+      "force_loss": 0.011605246923863888,
+      "step": 5100
+    },
+    {
+      "epoch": 4.595323741007194,
+      "grad_norm": 0.6765364408493042,
+      "learning_rate": 9.966259295211697e-05,
+      "loss": 0.0325,
+      "step": 5110
+    },
+    {
+      "action_loss": 0.01968207024037838,
+      "epoch": 4.595323741007194,
+      "step": 5110
+    },
+    {
+      "epoch": 4.595323741007194,
+      "step": 5110,
+      "torque_loss": 0.1369035392999649
+    },
+    {
+      "epoch": 4.595323741007194,
+      "force_loss": 0.014389949850738049,
+      "step": 5110
+    },
+    {
+      "epoch": 4.60431654676259,
+      "grad_norm": 0.7684283256530762,
+      "learning_rate": 9.965938932520028e-05,
+      "loss": 0.036,
+      "step": 5120
+    },
+    {
+      "action_loss": 0.023936502635478973,
+      "epoch": 4.60431654676259,
+      "step": 5120
+    },
+    {
+      "epoch": 4.60431654676259,
+      "step": 5120,
+      "torque_loss": 0.12168872356414795
+    },
+    {
+      "epoch": 4.60431654676259,
+      "force_loss": 0.01676768623292446,
+      "step": 5120
+    },
+    {
+      "epoch": 4.613309352517986,
+      "grad_norm": 0.764106273651123,
+      "learning_rate": 9.965617061307229e-05,
+      "loss": 0.0314,
+      "step": 5130
+    },
+    {
+      "action_loss": 0.02953227050602436,
+      "epoch": 4.613309352517986,
+      "step": 5130
+    },
+    {
+      "epoch": 4.613309352517986,
+      "step": 5130,
+      "torque_loss": 0.1810380071401596
+    },
+    {
+      "epoch": 4.613309352517986,
+      "force_loss": 0.024404026567935944,
+      "step": 5130
+    },
+    {
+      "epoch": 4.622302158273381,
+      "grad_norm": 0.874100387096405,
+      "learning_rate": 9.965293681671077e-05,
+      "loss": 0.0401,
+      "step": 5140
+    },
+    {
+      "action_loss": 0.01979588158428669,
+      "epoch": 4.622302158273381,
+      "step": 5140
+    },
+    {
+      "epoch": 4.622302158273381,
+      "step": 5140,
+      "torque_loss": 0.12462959438562393
+    },
+    {
+      "epoch": 4.622302158273381,
+      "force_loss": 0.016895165666937828,
+      "step": 5140
+    },
+    {
+      "epoch": 4.631294964028777,
+      "grad_norm": 0.6007933020591736,
+      "learning_rate": 9.964968793709804e-05,
+      "loss": 0.0313,
+      "step": 5150
+    },
+    {
+      "action_loss": 0.012876530177891254,
+      "epoch": 4.631294964028777,
+      "step": 5150
+    },
+    {
+      "epoch": 4.631294964028777,
+      "step": 5150,
+      "torque_loss": 0.12673906981945038
+    },
+    {
+      "epoch": 4.631294964028777,
+      "force_loss": 0.009997191838920116,
+      "step": 5150
+    },
+    {
+      "epoch": 4.640287769784173,
+      "grad_norm": 0.6168749332427979,
+      "learning_rate": 9.964642397522106e-05,
+      "loss": 0.0355,
+      "step": 5160
+    },
+    {
+      "action_loss": 0.015389413572847843,
+      "epoch": 4.640287769784173,
+      "step": 5160
+    },
+    {
+      "epoch": 4.640287769784173,
+      "step": 5160,
+      "torque_loss": 0.1390451341867447
+    },
+    {
+      "epoch": 4.640287769784173,
+      "force_loss": 0.01586478017270565,
+      "step": 5160
+    },
+    {
+      "epoch": 4.649280575539568,
+      "grad_norm": 0.7872642278671265,
+      "learning_rate": 9.96431449320713e-05,
+      "loss": 0.0325,
+      "step": 5170
+    },
+    {
+      "action_loss": 0.025892110541462898,
+      "epoch": 4.649280575539568,
+      "step": 5170
+    },
+    {
+      "epoch": 4.649280575539568,
+      "step": 5170,
+      "torque_loss": 0.1645011156797409
+    },
+    {
+      "epoch": 4.649280575539568,
+      "force_loss": 0.026243895292282104,
+      "step": 5170
+    },
+    {
+      "epoch": 4.658273381294964,
+      "grad_norm": 0.8102500438690186,
+      "learning_rate": 9.963985080864486e-05,
+      "loss": 0.0398,
+      "step": 5180
+    },
+    {
+      "action_loss": 0.014946692623198032,
+      "epoch": 4.658273381294964,
+      "step": 5180
+    },
+    {
+      "epoch": 4.658273381294964,
+      "step": 5180,
+      "torque_loss": 0.1599656194448471
+    },
+    {
+      "epoch": 4.658273381294964,
+      "force_loss": 0.018800029531121254,
+      "step": 5180
+    },
+    {
+      "epoch": 4.66726618705036,
+      "grad_norm": 0.8238599300384521,
+      "learning_rate": 9.96365416059424e-05,
+      "loss": 0.0333,
+      "step": 5190
+    },
+    {
+      "action_loss": 0.022762248292565346,
+      "epoch": 4.66726618705036,
+      "step": 5190
+    },
+    {
+      "epoch": 4.66726618705036,
+      "step": 5190,
+      "torque_loss": 0.13875947892665863
+    },
+    {
+      "epoch": 4.66726618705036,
+      "force_loss": 0.01594657450914383,
+      "step": 5190
+    },
+    {
+      "epoch": 4.676258992805756,
+      "grad_norm": 1.2515461444854736,
+      "learning_rate": 9.963321732496919e-05,
+      "loss": 0.0351,
+      "step": 5200
+    },
+    {
+      "action_loss": 0.02745429426431656,
+      "epoch": 4.676258992805756,
+      "step": 5200
+    },
+    {
+      "epoch": 4.676258992805756,
+      "step": 5200,
+      "torque_loss": 0.1591407209634781
+    },
+    {
+      "epoch": 4.676258992805756,
+      "force_loss": 0.016723915934562683,
+      "step": 5200
+    },
+    {
+      "epoch": 4.685251798561151,
+      "grad_norm": 0.5526142716407776,
+      "learning_rate": 9.962987796673506e-05,
+      "loss": 0.0363,
+      "step": 5210
+    },
+    {
+      "action_loss": 0.01568163000047207,
+      "epoch": 4.685251798561151,
+      "step": 5210
+    },
+    {
+      "epoch": 4.685251798561151,
+      "step": 5210,
+      "torque_loss": 0.13518263399600983
+    },
+    {
+      "epoch": 4.685251798561151,
+      "force_loss": 0.010564791969954967,
+      "step": 5210
+    },
+    {
+      "epoch": 4.694244604316546,
+      "grad_norm": 1.299788236618042,
+      "learning_rate": 9.962652353225438e-05,
+      "loss": 0.0301,
+      "step": 5220
+    },
+    {
+      "action_loss": 0.015987521037459373,
+      "epoch": 4.694244604316546,
+      "step": 5220
+    },
+    {
+      "epoch": 4.694244604316546,
+      "step": 5220,
+      "torque_loss": 0.12131834030151367
+    },
+    {
+      "epoch": 4.694244604316546,
+      "force_loss": 0.012146626599133015,
+      "step": 5220
+    },
+    {
+      "epoch": 4.703237410071942,
+      "grad_norm": 1.3127905130386353,
+      "learning_rate": 9.962315402254619e-05,
+      "loss": 0.0356,
+      "step": 5230
+    },
+    {
+      "action_loss": 0.01628502644598484,
+      "epoch": 4.703237410071942,
+      "step": 5230
+    },
+    {
+      "epoch": 4.703237410071942,
+      "step": 5230,
+      "torque_loss": 0.15971876680850983
+    },
+    {
+      "epoch": 4.703237410071942,
+      "force_loss": 0.00842433050274849,
+      "step": 5230
+    },
+    {
+      "epoch": 4.712230215827338,
+      "grad_norm": 0.564807653427124,
+      "learning_rate": 9.9619769438634e-05,
+      "loss": 0.0334,
+      "step": 5240
+    },
+    {
+      "action_loss": 0.023909149691462517,
+      "epoch": 4.712230215827338,
+      "step": 5240
+    },
+    {
+      "epoch": 4.712230215827338,
+      "step": 5240,
+      "torque_loss": 0.15270943939685822
+    },
+    {
+      "epoch": 4.712230215827338,
+      "force_loss": 0.014199353754520416,
+      "step": 5240
+    },
+    {
+      "epoch": 4.721223021582734,
+      "grad_norm": 0.8288257122039795,
+      "learning_rate": 9.9616369781546e-05,
+      "loss": 0.0296,
+      "step": 5250
+    },
+    {
+      "action_loss": 0.018591200932860374,
+      "epoch": 4.721223021582734,
+      "step": 5250
+    },
+    {
+      "epoch": 4.721223021582734,
+      "step": 5250,
+      "torque_loss": 0.15318669378757477
+    },
+    {
+      "epoch": 4.721223021582734,
+      "force_loss": 0.015463444404304028,
+      "step": 5250
+    },
+    {
+      "epoch": 4.73021582733813,
+      "grad_norm": 0.5895083546638489,
+      "learning_rate": 9.961295505231491e-05,
+      "loss": 0.0348,
+      "step": 5260
+    },
+    {
+      "action_loss": 0.023209886625409126,
+      "epoch": 4.73021582733813,
+      "step": 5260
+    },
+    {
+      "epoch": 4.73021582733813,
+      "step": 5260,
+      "torque_loss": 0.13728447258472443
+    },
+    {
+      "epoch": 4.73021582733813,
+      "force_loss": 0.019676925614476204,
+      "step": 5260
+    },
+    {
+      "epoch": 4.739208633093525,
+      "grad_norm": 1.345011591911316,
+      "learning_rate": 9.960952525197804e-05,
+      "loss": 0.0319,
+      "step": 5270
+    },
+    {
+      "action_loss": 0.005507905501872301,
+      "epoch": 4.739208633093525,
+      "step": 5270
+    },
+    {
+      "epoch": 4.739208633093525,
+      "step": 5270,
+      "torque_loss": 0.13759738206863403
+    },
+    {
+      "epoch": 4.739208633093525,
+      "force_loss": 0.006951266899704933,
+      "step": 5270
+    },
+    {
+      "epoch": 4.748201438848921,
+      "grad_norm": 0.7037013173103333,
+      "learning_rate": 9.960608038157724e-05,
+      "loss": 0.0262,
+      "step": 5280
+    },
+    {
+      "action_loss": 0.01759648509323597,
+      "epoch": 4.748201438848921,
+      "step": 5280
+    },
+    {
+      "epoch": 4.748201438848921,
+      "step": 5280,
+      "torque_loss": 0.12934188544750214
+    },
+    {
+      "epoch": 4.748201438848921,
+      "force_loss": 0.021389560773968697,
+      "step": 5280
+    },
+    {
+      "epoch": 4.7571942446043165,
+      "grad_norm": 1.2372910976409912,
+      "learning_rate": 9.960262044215901e-05,
+      "loss": 0.0297,
+      "step": 5290
+    },
+    {
+      "action_loss": 0.027033813297748566,
+      "epoch": 4.7571942446043165,
+      "step": 5290
+    },
+    {
+      "epoch": 4.7571942446043165,
+      "step": 5290,
+      "torque_loss": 0.22342686355113983
+    },
+    {
+      "epoch": 4.7571942446043165,
+      "force_loss": 0.03127235919237137,
+      "step": 5290
+    },
+    {
+      "epoch": 4.766187050359712,
+      "grad_norm": 0.9708763360977173,
+      "learning_rate": 9.959914543477435e-05,
+      "loss": 0.0408,
+      "step": 5300
+    },
+    {
+      "action_loss": 0.053166795521974564,
+      "epoch": 4.766187050359712,
+      "step": 5300
+    },
+    {
+      "epoch": 4.766187050359712,
+      "step": 5300,
+      "torque_loss": 0.18424074351787567
+    },
+    {
+      "epoch": 4.766187050359712,
+      "force_loss": 0.06984049826860428,
+      "step": 5300
+    },
+    {
+      "epoch": 4.775179856115108,
+      "grad_norm": 1.3754055500030518,
+      "learning_rate": 9.959565536047892e-05,
+      "loss": 0.0418,
+      "step": 5310
+    },
+    {
+      "action_loss": 0.00839131698012352,
+      "epoch": 4.775179856115108,
+      "step": 5310
+    },
+    {
+      "epoch": 4.775179856115108,
+      "step": 5310,
+      "torque_loss": 0.10514402389526367
+    },
+    {
+      "epoch": 4.775179856115108,
+      "force_loss": 0.0082511892542243,
+      "step": 5310
+    },
+    {
+      "epoch": 4.784172661870503,
+      "grad_norm": 0.5044088363647461,
+      "learning_rate": 9.959215022033288e-05,
+      "loss": 0.0321,
+      "step": 5320
+    },
+    {
+      "action_loss": 0.02304108999669552,
+      "epoch": 4.784172661870503,
+      "step": 5320
+    },
+    {
+      "epoch": 4.784172661870503,
+      "step": 5320,
+      "torque_loss": 0.18510663509368896
+    },
+    {
+      "epoch": 4.784172661870503,
+      "force_loss": 0.021340645849704742,
+      "step": 5320
+    },
+    {
+      "epoch": 4.793165467625899,
+      "grad_norm": 0.6199601292610168,
+      "learning_rate": 9.9588630015401e-05,
+      "loss": 0.0359,
+      "step": 5330
+    },
+    {
+      "action_loss": 0.016309387981891632,
+      "epoch": 4.793165467625899,
+      "step": 5330
+    },
+    {
+      "epoch": 4.793165467625899,
+      "step": 5330,
+      "torque_loss": 0.1842377781867981
+    },
+    {
+      "epoch": 4.793165467625899,
+      "force_loss": 0.017711469903588295,
+      "step": 5330
+    },
+    {
+      "epoch": 4.802158273381295,
+      "grad_norm": 0.6002806425094604,
+      "learning_rate": 9.958509474675264e-05,
+      "loss": 0.033,
+      "step": 5340
+    },
+    {
+      "action_loss": 0.013878691010177135,
+      "epoch": 4.802158273381295,
+      "step": 5340
+    },
+    {
+      "epoch": 4.802158273381295,
+      "step": 5340,
+      "torque_loss": 0.15046750009059906
+    },
+    {
+      "epoch": 4.802158273381295,
+      "force_loss": 0.011715096421539783,
+      "step": 5340
+    },
+    {
+      "epoch": 4.811151079136691,
+      "grad_norm": 0.9952367544174194,
+      "learning_rate": 9.958154441546171e-05,
+      "loss": 0.0397,
+      "step": 5350
+    },
+    {
+      "action_loss": 0.034337714314460754,
+      "epoch": 4.811151079136691,
+      "step": 5350
+    },
+    {
+      "epoch": 4.811151079136691,
+      "step": 5350,
+      "torque_loss": 0.161305770277977
+    },
+    {
+      "epoch": 4.811151079136691,
+      "force_loss": 0.027293890714645386,
+      "step": 5350
+    },
+    {
+      "epoch": 4.820143884892087,
+      "grad_norm": 1.044042706489563,
+      "learning_rate": 9.957797902260673e-05,
+      "loss": 0.0374,
+      "step": 5360
+    },
+    {
+      "action_loss": 0.011409183032810688,
+      "epoch": 4.820143884892087,
+      "step": 5360
+    },
+    {
+      "epoch": 4.820143884892087,
+      "step": 5360,
+      "torque_loss": 0.13266690075397491
+    },
+    {
+      "epoch": 4.820143884892087,
+      "force_loss": 0.007539886515587568,
+      "step": 5360
+    },
+    {
+      "epoch": 4.829136690647482,
+      "grad_norm": 0.7261578440666199,
+      "learning_rate": 9.957439856927073e-05,
+      "loss": 0.0314,
+      "step": 5370
+    },
+    {
+      "action_loss": 0.017129620537161827,
+      "epoch": 4.829136690647482,
+      "step": 5370
+    },
+    {
+      "epoch": 4.829136690647482,
+      "step": 5370,
+      "torque_loss": 0.11524128913879395
+    },
+    {
+      "epoch": 4.829136690647482,
+      "force_loss": 0.017683295533061028,
+      "step": 5370
+    },
+    {
+      "epoch": 4.838129496402877,
+      "grad_norm": 1.042865514755249,
+      "learning_rate": 9.957080305654139e-05,
+      "loss": 0.0392,
+      "step": 5380
+    },
+    {
+      "action_loss": 0.017216136679053307,
+      "epoch": 4.838129496402877,
+      "step": 5380
+    },
+    {
+      "epoch": 4.838129496402877,
+      "step": 5380,
+      "torque_loss": 0.16361086070537567
+    },
+    {
+      "epoch": 4.838129496402877,
+      "force_loss": 0.013957299292087555,
+      "step": 5380
+    },
+    {
+      "epoch": 4.847122302158273,
+      "grad_norm": 1.5993086099624634,
+      "learning_rate": 9.956719248551092e-05,
+      "loss": 0.0388,
+      "step": 5390
+    },
+    {
+      "action_loss": 0.014758334495127201,
+      "epoch": 4.847122302158273,
+      "step": 5390
+    },
+    {
+      "epoch": 4.847122302158273,
+      "step": 5390,
+      "torque_loss": 0.1057029739022255
+    },
+    {
+      "epoch": 4.847122302158273,
+      "force_loss": 0.01089871022850275,
+      "step": 5390
+    },
+    {
+      "epoch": 4.856115107913669,
+      "grad_norm": 0.75738924741745,
+      "learning_rate": 9.956356685727612e-05,
+      "loss": 0.0362,
+      "step": 5400
+    },
+    {
+      "action_loss": 0.010722232051193714,
+      "epoch": 4.856115107913669,
+      "step": 5400
+    },
+    {
+      "epoch": 4.856115107913669,
+      "step": 5400,
+      "torque_loss": 0.12964071333408356
+    },
+    {
+      "epoch": 4.856115107913669,
+      "force_loss": 0.009252565912902355,
+      "step": 5400
+    },
+    {
+      "epoch": 4.865107913669065,
+      "grad_norm": 0.45988819003105164,
+      "learning_rate": 9.955992617293836e-05,
+      "loss": 0.0329,
+      "step": 5410
+    },
+    {
+      "action_loss": 0.01255980134010315,
+      "epoch": 4.865107913669065,
+      "step": 5410
+    },
+    {
+      "epoch": 4.865107913669065,
+      "step": 5410,
+      "torque_loss": 0.17839623987674713
+    },
+    {
+      "epoch": 4.865107913669065,
+      "force_loss": 0.011090818792581558,
+      "step": 5410
+    },
+    {
+      "epoch": 4.874100719424461,
+      "grad_norm": 0.3930070996284485,
+      "learning_rate": 9.955627043360358e-05,
+      "loss": 0.0355,
+      "step": 5420
+    },
+    {
+      "action_loss": 0.01895631104707718,
+      "epoch": 4.874100719424461,
+      "step": 5420
+    },
+    {
+      "epoch": 4.874100719424461,
+      "step": 5420,
+      "torque_loss": 0.17519693076610565
+    },
+    {
+      "epoch": 4.874100719424461,
+      "force_loss": 0.017761051654815674,
+      "step": 5420
+    },
+    {
+      "epoch": 4.883093525179856,
+      "grad_norm": 1.7386761903762817,
+      "learning_rate": 9.955259964038231e-05,
+      "loss": 0.0335,
+      "step": 5430
+    },
+    {
+      "action_loss": 0.01452229917049408,
+      "epoch": 4.883093525179856,
+      "step": 5430
+    },
+    {
+      "epoch": 4.883093525179856,
+      "step": 5430,
+      "torque_loss": 0.15585090219974518
+    },
+    {
+      "epoch": 4.883093525179856,
+      "force_loss": 0.016538606956601143,
+      "step": 5430
+    },
+    {
+      "epoch": 4.892086330935252,
+      "grad_norm": 1.6594011783599854,
+      "learning_rate": 9.954891379438962e-05,
+      "loss": 0.0363,
+      "step": 5440
+    },
+    {
+      "action_loss": 0.018527643755078316,
+      "epoch": 4.892086330935252,
+      "step": 5440
+    },
+    {
+      "epoch": 4.892086330935252,
+      "step": 5440,
+      "torque_loss": 0.16896848380565643
+    },
+    {
+      "epoch": 4.892086330935252,
+      "force_loss": 0.022833295166492462,
+      "step": 5440
+    },
+    {
+      "epoch": 4.901079136690647,
+      "grad_norm": 1.4442071914672852,
+      "learning_rate": 9.954521289674519e-05,
+      "loss": 0.0374,
+      "step": 5450
+    },
+    {
+      "action_loss": 0.015118967741727829,
+      "epoch": 4.901079136690647,
+      "step": 5450
+    },
+    {
+      "epoch": 4.901079136690647,
+      "step": 5450,
+      "torque_loss": 0.11366186290979385
+    },
+    {
+      "epoch": 4.901079136690647,
+      "force_loss": 0.013646438717842102,
+      "step": 5450
+    },
+    {
+      "epoch": 4.910071942446043,
+      "grad_norm": 0.9077381491661072,
+      "learning_rate": 9.954149694857325e-05,
+      "loss": 0.0344,
+      "step": 5460
+    },
+    {
+      "action_loss": 0.014302697964012623,
+      "epoch": 4.910071942446043,
+      "step": 5460
+    },
+    {
+      "epoch": 4.910071942446043,
+      "step": 5460,
+      "torque_loss": 0.17495770752429962
+    },
+    {
+      "epoch": 4.910071942446043,
+      "force_loss": 0.018926559016108513,
+      "step": 5460
+    },
+    {
+      "epoch": 4.919064748201439,
+      "grad_norm": 0.5356278419494629,
+      "learning_rate": 9.953776595100258e-05,
+      "loss": 0.0386,
+      "step": 5470
+    },
+    {
+      "action_loss": 0.016110757365822792,
+      "epoch": 4.919064748201439,
+      "step": 5470
+    },
+    {
+      "epoch": 4.919064748201439,
+      "step": 5470,
+      "torque_loss": 0.17800450325012207
+    },
+    {
+      "epoch": 4.919064748201439,
+      "force_loss": 0.015125256963074207,
+      "step": 5470
+    },
+    {
+      "epoch": 4.928057553956835,
+      "grad_norm": 1.376873254776001,
+      "learning_rate": 9.95340199051666e-05,
+      "loss": 0.0282,
+      "step": 5480
+    },
+    {
+      "action_loss": 0.011282348074018955,
+      "epoch": 4.928057553956835,
+      "step": 5480
+    },
+    {
+      "epoch": 4.928057553956835,
+      "step": 5480,
+      "torque_loss": 0.14414973556995392
+    },
+    {
+      "epoch": 4.928057553956835,
+      "force_loss": 0.007896043360233307,
+      "step": 5480
+    },
+    {
+      "epoch": 4.93705035971223,
+      "grad_norm": 0.8240759968757629,
+      "learning_rate": 9.953025881220325e-05,
+      "loss": 0.0294,
+      "step": 5490
+    },
+    {
+      "action_loss": 0.01085557509213686,
+      "epoch": 4.93705035971223,
+      "step": 5490
+    },
+    {
+      "epoch": 4.93705035971223,
+      "step": 5490,
+      "torque_loss": 0.13415758311748505
+    },
+    {
+      "epoch": 4.93705035971223,
+      "force_loss": 0.009497441351413727,
+      "step": 5490
+    },
+    {
+      "epoch": 4.946043165467626,
+      "grad_norm": 0.7632565498352051,
+      "learning_rate": 9.952648267325504e-05,
+      "loss": 0.0299,
+      "step": 5500
+    },
+    {
+      "action_loss": 0.023374313488602638,
+      "epoch": 4.946043165467626,
+      "step": 5500
+    },
+    {
+      "epoch": 4.946043165467626,
+      "step": 5500,
+      "torque_loss": 0.14962723851203918
+    },
+    {
+      "epoch": 4.946043165467626,
+      "force_loss": 0.025252118706703186,
+      "step": 5500
+    },
+    {
+      "epoch": 4.955035971223022,
+      "grad_norm": 0.9009926319122314,
+      "learning_rate": 9.952269148946905e-05,
+      "loss": 0.0324,
+      "step": 5510
+    },
+    {
+      "action_loss": 0.022167472168803215,
+      "epoch": 4.955035971223022,
+      "step": 5510
+    },
+    {
+      "epoch": 4.955035971223022,
+      "step": 5510,
+      "torque_loss": 0.1405017226934433
+    },
+    {
+      "epoch": 4.955035971223022,
+      "force_loss": 0.03133165463805199,
+      "step": 5510
+    },
+    {
+      "epoch": 4.9640287769784175,
+      "grad_norm": 0.883463978767395,
+      "learning_rate": 9.951888526199697e-05,
+      "loss": 0.0337,
+      "step": 5520
+    },
+    {
+      "action_loss": 0.0214130487293005,
+      "epoch": 4.9640287769784175,
+      "step": 5520
+    },
+    {
+      "epoch": 4.9640287769784175,
+      "step": 5520,
+      "torque_loss": 0.14747439324855804
+    },
+    {
+      "epoch": 4.9640287769784175,
+      "force_loss": 0.020815743133425713,
+      "step": 5520
+    },
+    {
+      "epoch": 4.9730215827338125,
+      "grad_norm": 1.318161129951477,
+      "learning_rate": 9.951506399199501e-05,
+      "loss": 0.0347,
+      "step": 5530
+    },
+    {
+      "action_loss": 0.013590328395366669,
+      "epoch": 4.9730215827338125,
+      "step": 5530
+    },
+    {
+      "epoch": 4.9730215827338125,
+      "step": 5530,
+      "torque_loss": 0.12350980192422867
+    },
+    {
+      "epoch": 4.9730215827338125,
+      "force_loss": 0.009893982671201229,
+      "step": 5530
+    },
+    {
+      "epoch": 4.982014388489208,
+      "grad_norm": 1.0562297105789185,
+      "learning_rate": 9.951122768062399e-05,
+      "loss": 0.0323,
+      "step": 5540
+    },
+    {
+      "action_loss": 0.006967715919017792,
+      "epoch": 4.982014388489208,
+      "step": 5540
+    },
+    {
+      "epoch": 4.982014388489208,
+      "step": 5540,
+      "torque_loss": 0.12151825428009033
+    },
+    {
+      "epoch": 4.982014388489208,
+      "force_loss": 0.0068406532518565655,
+      "step": 5540
+    },
+    {
+      "epoch": 4.991007194244604,
+      "grad_norm": 0.664605438709259,
+      "learning_rate": 9.950737632904927e-05,
+      "loss": 0.03,
+      "step": 5550
+    },
+    {
+      "action_loss": 0.014268617145717144,
+      "epoch": 4.991007194244604,
+      "step": 5550
+    },
+    {
+      "epoch": 4.991007194244604,
+      "step": 5550,
+      "torque_loss": 0.12223458290100098
+    },
+    {
+      "epoch": 4.991007194244604,
+      "force_loss": 0.010322910733520985,
+      "step": 5550
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.986358642578125,
+      "learning_rate": 9.950350993844077e-05,
+      "loss": 0.0333,
+      "step": 5560
+    },
+    {
+      "action_loss": 0.02257212996482849,
+      "epoch": 5.0,
+      "step": 5560
+    },
+    {
+      "epoch": 5.0,
+      "step": 5560,
+      "torque_loss": 0.15743350982666016
+    },
+    {
+      "epoch": 5.0,
+      "force_loss": 0.013492533005774021,
+      "step": 5560
+    },
+    {
+      "epoch": 5.008992805755396,
+      "grad_norm": 0.6122710704803467,
+      "learning_rate": 9.949962850997303e-05,
+      "loss": 0.0362,
+      "step": 5570
+    },
+    {
+      "action_loss": 0.017204903066158295,
+      "epoch": 5.008992805755396,
+      "step": 5570
+    },
+    {
+      "epoch": 5.008992805755396,
+      "step": 5570,
+      "torque_loss": 0.1751583367586136
+    },
+    {
+      "epoch": 5.008992805755396,
+      "force_loss": 0.010946311056613922,
+      "step": 5570
+    },
+    {
+      "epoch": 5.017985611510792,
+      "grad_norm": 1.5294653177261353,
+      "learning_rate": 9.949573204482512e-05,
+      "loss": 0.0296,
+      "step": 5580
+    },
+    {
+      "action_loss": 0.016943540424108505,
+      "epoch": 5.017985611510792,
+      "step": 5580
+    },
+    {
+      "epoch": 5.017985611510792,
+      "step": 5580,
+      "torque_loss": 0.11341521888971329
+    },
+    {
+      "epoch": 5.017985611510792,
+      "force_loss": 0.010504749603569508,
+      "step": 5580
+    },
+    {
+      "epoch": 5.026978417266187,
+      "grad_norm": 1.120571255683899,
+      "learning_rate": 9.949182054418064e-05,
+      "loss": 0.0327,
+      "step": 5590
+    },
+    {
+      "action_loss": 0.02788245677947998,
+      "epoch": 5.026978417266187,
+      "step": 5590
+    },
+    {
+      "epoch": 5.026978417266187,
+      "step": 5590,
+      "torque_loss": 0.23911577463150024
+    },
+    {
+      "epoch": 5.026978417266187,
+      "force_loss": 0.030254175886511803,
+      "step": 5590
+    },
+    {
+      "epoch": 5.0359712230215825,
+      "grad_norm": 1.5035184621810913,
+      "learning_rate": 9.948789400922787e-05,
+      "loss": 0.0367,
+      "step": 5600
+    },
+    {
+      "action_loss": 0.024786515161395073,
+      "epoch": 5.0359712230215825,
+      "step": 5600
+    },
+    {
+      "epoch": 5.0359712230215825,
+      "step": 5600,
+      "torque_loss": 0.1330796480178833
+    },
+    {
+      "epoch": 5.0359712230215825,
+      "force_loss": 0.010250705294311047,
+      "step": 5600
+    },
+    {
+      "epoch": 5.044964028776978,
+      "grad_norm": 1.7596006393432617,
+      "learning_rate": 9.948395244115953e-05,
+      "loss": 0.0327,
+      "step": 5610
+    },
+    {
+      "action_loss": 0.04590732976794243,
+      "epoch": 5.044964028776978,
+      "step": 5610
+    },
+    {
+      "epoch": 5.044964028776978,
+      "step": 5610,
+      "torque_loss": 0.1576809585094452
+    },
+    {
+      "epoch": 5.044964028776978,
+      "force_loss": 0.046087637543678284,
+      "step": 5610
+    },
+    {
+      "epoch": 5.053956834532374,
+      "grad_norm": 0.893988311290741,
+      "learning_rate": 9.9479995841173e-05,
+      "loss": 0.0335,
+      "step": 5620
+    },
+    {
+      "action_loss": 0.012712088413536549,
+      "epoch": 5.053956834532374,
+      "step": 5620
+    },
+    {
+      "epoch": 5.053956834532374,
+      "step": 5620,
+      "torque_loss": 0.12414965033531189
+    },
+    {
+      "epoch": 5.053956834532374,
+      "force_loss": 0.015016489662230015,
+      "step": 5620
+    },
+    {
+      "epoch": 5.06294964028777,
+      "grad_norm": 0.6276310086250305,
+      "learning_rate": 9.947602421047017e-05,
+      "loss": 0.0247,
+      "step": 5630
+    },
+    {
+      "action_loss": 0.027676789090037346,
+      "epoch": 5.06294964028777,
+      "step": 5630
+    },
+    {
+      "epoch": 5.06294964028777,
+      "step": 5630,
+      "torque_loss": 0.1542523056268692
+    },
+    {
+      "epoch": 5.06294964028777,
+      "force_loss": 0.019151249900460243,
+      "step": 5630
+    },
+    {
+      "epoch": 5.071942446043165,
+      "grad_norm": 1.2098196744918823,
+      "learning_rate": 9.947203755025753e-05,
+      "loss": 0.0347,
+      "step": 5640
+    },
+    {
+      "action_loss": 0.011504284106194973,
+      "epoch": 5.071942446043165,
+      "step": 5640
+    },
+    {
+      "epoch": 5.071942446043165,
+      "step": 5640,
+      "torque_loss": 0.11376959085464478
+    },
+    {
+      "epoch": 5.071942446043165,
+      "force_loss": 0.024473929777741432,
+      "step": 5640
+    },
+    {
+      "epoch": 5.080935251798561,
+      "grad_norm": 1.16839599609375,
+      "learning_rate": 9.946803586174611e-05,
+      "loss": 0.033,
+      "step": 5650
+    },
+    {
+      "action_loss": 0.021279791370034218,
+      "epoch": 5.080935251798561,
+      "step": 5650
+    },
+    {
+      "epoch": 5.080935251798561,
+      "step": 5650,
+      "torque_loss": 0.1337127685546875
+    },
+    {
+      "epoch": 5.080935251798561,
+      "force_loss": 0.01976182870566845,
+      "step": 5650
+    },
+    {
+      "epoch": 5.089928057553957,
+      "grad_norm": 0.5171191692352295,
+      "learning_rate": 9.946401914615151e-05,
+      "loss": 0.0341,
+      "step": 5660
+    },
+    {
+      "action_loss": 0.009707505814731121,
+      "epoch": 5.089928057553957,
+      "step": 5660
+    },
+    {
+      "epoch": 5.089928057553957,
+      "step": 5660,
+      "torque_loss": 0.0918097272515297
+    },
+    {
+      "epoch": 5.089928057553957,
+      "force_loss": 0.00940024945884943,
+      "step": 5660
+    },
+    {
+      "epoch": 5.098920863309353,
+      "grad_norm": 0.841602087020874,
+      "learning_rate": 9.945998740469394e-05,
+      "loss": 0.0259,
+      "step": 5670
+    },
+    {
+      "action_loss": 0.07884933799505234,
+      "epoch": 5.098920863309353,
+      "step": 5670
+    },
+    {
+      "epoch": 5.098920863309353,
+      "step": 5670,
+      "torque_loss": 0.20755648612976074
+    },
+    {
+      "epoch": 5.098920863309353,
+      "force_loss": 0.07677451521158218,
+      "step": 5670
+    },
+    {
+      "epoch": 5.107913669064748,
+      "grad_norm": 1.2909084558486938,
+      "learning_rate": 9.945594063859809e-05,
+      "loss": 0.041,
+      "step": 5680
+    },
+    {
+      "action_loss": 0.008488588966429234,
+      "epoch": 5.107913669064748,
+      "step": 5680
+    },
+    {
+      "epoch": 5.107913669064748,
+      "step": 5680,
+      "torque_loss": 0.14961610734462738
+    },
+    {
+      "epoch": 5.107913669064748,
+      "force_loss": 0.008059262298047543,
+      "step": 5680
+    },
+    {
+      "epoch": 5.116906474820144,
+      "grad_norm": 1.4823791980743408,
+      "learning_rate": 9.94518788490933e-05,
+      "loss": 0.0297,
+      "step": 5690
+    },
+    {
+      "action_loss": 0.030524766072630882,
+      "epoch": 5.116906474820144,
+      "step": 5690
+    },
+    {
+      "epoch": 5.116906474820144,
+      "step": 5690,
+      "torque_loss": 0.19970446825027466
+    },
+    {
+      "epoch": 5.116906474820144,
+      "force_loss": 0.022799404338002205,
+      "step": 5690
+    },
+    {
+      "epoch": 5.125899280575539,
+      "grad_norm": 0.34172824025154114,
+      "learning_rate": 9.944780203741341e-05,
+      "loss": 0.0379,
+      "step": 5700
+    },
+    {
+      "action_loss": 0.010089091025292873,
+      "epoch": 5.125899280575539,
+      "step": 5700
+    },
+    {
+      "epoch": 5.125899280575539,
+      "step": 5700,
+      "torque_loss": 0.09576954692602158
+    },
+    {
+      "epoch": 5.125899280575539,
+      "force_loss": 0.005920028779655695,
+      "step": 5700
+    },
+    {
+      "epoch": 5.134892086330935,
+      "grad_norm": 0.9330561757087708,
+      "learning_rate": 9.944371020479686e-05,
+      "loss": 0.0324,
+      "step": 5710
+    },
+    {
+      "action_loss": 0.0141999376937747,
+      "epoch": 5.134892086330935,
+      "step": 5710
+    },
+    {
+      "epoch": 5.134892086330935,
+      "step": 5710,
+      "torque_loss": 0.14227472245693207
+    },
+    {
+      "epoch": 5.134892086330935,
+      "force_loss": 0.00944321695715189,
+      "step": 5710
+    },
+    {
+      "epoch": 5.143884892086331,
+      "grad_norm": 0.49037691950798035,
+      "learning_rate": 9.943960335248662e-05,
+      "loss": 0.0298,
+      "step": 5720
+    },
+    {
+      "action_loss": 0.009946809150278568,
+      "epoch": 5.143884892086331,
+      "step": 5720
+    },
+    {
+      "epoch": 5.143884892086331,
+      "step": 5720,
+      "torque_loss": 0.11354728788137436
+    },
+    {
+      "epoch": 5.143884892086331,
+      "force_loss": 0.00917693879455328,
+      "step": 5720
+    },
+    {
+      "epoch": 5.152877697841727,
+      "grad_norm": 1.005693793296814,
+      "learning_rate": 9.943548148173027e-05,
+      "loss": 0.0339,
+      "step": 5730
+    },
+    {
+      "action_loss": 0.009529300034046173,
+      "epoch": 5.152877697841727,
+      "step": 5730
+    },
+    {
+      "epoch": 5.152877697841727,
+      "step": 5730,
+      "torque_loss": 0.10242322087287903
+    },
+    {
+      "epoch": 5.152877697841727,
+      "force_loss": 0.008441810496151447,
+      "step": 5730
+    },
+    {
+      "epoch": 5.161870503597123,
+      "grad_norm": 0.7240970730781555,
+      "learning_rate": 9.943134459377992e-05,
+      "loss": 0.0274,
+      "step": 5740
+    },
+    {
+      "action_loss": 0.009209486655890942,
+      "epoch": 5.161870503597123,
+      "step": 5740
+    },
+    {
+      "epoch": 5.161870503597123,
+      "step": 5740,
+      "torque_loss": 0.1518832892179489
+    },
+    {
+      "epoch": 5.161870503597123,
+      "force_loss": 0.008651758544147015,
+      "step": 5740
+    },
+    {
+      "epoch": 5.170863309352518,
+      "grad_norm": 0.6877978444099426,
+      "learning_rate": 9.942719268989222e-05,
+      "loss": 0.036,
+      "step": 5750
+    },
+    {
+      "action_loss": 0.008205019868910313,
+      "epoch": 5.170863309352518,
+      "step": 5750
+    },
+    {
+      "epoch": 5.170863309352518,
+      "step": 5750,
+      "torque_loss": 0.14600127935409546
+    },
+    {
+      "epoch": 5.170863309352518,
+      "force_loss": 0.008497755043208599,
+      "step": 5750
+    },
+    {
+      "epoch": 5.179856115107913,
+      "grad_norm": 0.7438530325889587,
+      "learning_rate": 9.942302577132844e-05,
+      "loss": 0.0308,
+      "step": 5760
+    },
+    {
+      "action_loss": 0.01791953109204769,
+      "epoch": 5.179856115107913,
+      "step": 5760
+    },
+    {
+      "epoch": 5.179856115107913,
+      "step": 5760,
+      "torque_loss": 0.12526437640190125
+    },
+    {
+      "epoch": 5.179856115107913,
+      "force_loss": 0.01129516214132309,
+      "step": 5760
+    },
+    {
+      "epoch": 5.188848920863309,
+      "grad_norm": 0.9081936478614807,
+      "learning_rate": 9.941884383935438e-05,
+      "loss": 0.0291,
+      "step": 5770
+    },
+    {
+      "action_loss": 0.03710727021098137,
+      "epoch": 5.188848920863309,
+      "step": 5770
+    },
+    {
+      "epoch": 5.188848920863309,
+      "step": 5770,
+      "torque_loss": 0.18137772381305695
+    },
+    {
+      "epoch": 5.188848920863309,
+      "force_loss": 0.041631344705820084,
+      "step": 5770
+    },
+    {
+      "epoch": 5.197841726618705,
+      "grad_norm": 0.7799079418182373,
+      "learning_rate": 9.941464689524039e-05,
+      "loss": 0.034,
+      "step": 5780
+    },
+    {
+      "action_loss": 0.008697784505784512,
+      "epoch": 5.197841726618705,
+      "step": 5780
+    },
+    {
+      "epoch": 5.197841726618705,
+      "step": 5780,
+      "torque_loss": 0.11888748407363892
+    },
+    {
+      "epoch": 5.197841726618705,
+      "force_loss": 0.007591801229864359,
+      "step": 5780
+    },
+    {
+      "epoch": 5.206834532374101,
+      "grad_norm": 1.5161101818084717,
+      "learning_rate": 9.941043494026139e-05,
+      "loss": 0.0344,
+      "step": 5790
+    },
+    {
+      "action_loss": 0.03356954827904701,
+      "epoch": 5.206834532374101,
+      "step": 5790
+    },
+    {
+      "epoch": 5.206834532374101,
+      "step": 5790,
+      "torque_loss": 0.1888788938522339
+    },
+    {
+      "epoch": 5.206834532374101,
+      "force_loss": 0.030938828364014626,
+      "step": 5790
+    },
+    {
+      "epoch": 5.215827338129497,
+      "grad_norm": 1.2998768091201782,
+      "learning_rate": 9.940620797569685e-05,
+      "loss": 0.0321,
+      "step": 5800
+    },
+    {
+      "action_loss": 0.02752203680574894,
+      "epoch": 5.215827338129497,
+      "step": 5800
+    },
+    {
+      "epoch": 5.215827338129497,
+      "step": 5800,
+      "torque_loss": 0.1494486778974533
+    },
+    {
+      "epoch": 5.215827338129497,
+      "force_loss": 0.020826246589422226,
+      "step": 5800
+    },
+    {
+      "epoch": 5.224820143884892,
+      "grad_norm": 0.6310189366340637,
+      "learning_rate": 9.940196600283082e-05,
+      "loss": 0.0421,
+      "step": 5810
+    },
+    {
+      "action_loss": 0.016576392576098442,
+      "epoch": 5.224820143884892,
+      "step": 5810
+    },
+    {
+      "epoch": 5.224820143884892,
+      "step": 5810,
+      "torque_loss": 0.1569250077009201
+    },
+    {
+      "epoch": 5.224820143884892,
+      "force_loss": 0.012351572513580322,
+      "step": 5810
+    },
+    {
+      "epoch": 5.233812949640288,
+      "grad_norm": 0.8556134700775146,
+      "learning_rate": 9.939770902295192e-05,
+      "loss": 0.0398,
+      "step": 5820
+    },
+    {
+      "action_loss": 0.010393361561000347,
+      "epoch": 5.233812949640288,
+      "step": 5820
+    },
+    {
+      "epoch": 5.233812949640288,
+      "step": 5820,
+      "torque_loss": 0.12201127409934998
+    },
+    {
+      "epoch": 5.233812949640288,
+      "force_loss": 0.011214190162718296,
+      "step": 5820
+    },
+    {
+      "epoch": 5.2428057553956835,
+      "grad_norm": 0.8266036510467529,
+      "learning_rate": 9.939343703735329e-05,
+      "loss": 0.0347,
+      "step": 5830
+    },
+    {
+      "action_loss": 0.02877046726644039,
+      "epoch": 5.2428057553956835,
+      "step": 5830
+    },
+    {
+      "epoch": 5.2428057553956835,
+      "step": 5830,
+      "torque_loss": 0.1741381287574768
+    },
+    {
+      "epoch": 5.2428057553956835,
+      "force_loss": 0.02859148569405079,
+      "step": 5830
+    },
+    {
+      "epoch": 5.251798561151079,
+      "grad_norm": 1.0230200290679932,
+      "learning_rate": 9.938915004733264e-05,
+      "loss": 0.0326,
+      "step": 5840
+    },
+    {
+      "action_loss": 0.01110775861889124,
+      "epoch": 5.251798561151079,
+      "step": 5840
+    },
+    {
+      "epoch": 5.251798561151079,
+      "step": 5840,
+      "torque_loss": 0.1227857694029808
+    },
+    {
+      "epoch": 5.251798561151079,
+      "force_loss": 0.008381695486605167,
+      "step": 5840
+    },
+    {
+      "epoch": 5.260791366906475,
+      "grad_norm": 0.6586215496063232,
+      "learning_rate": 9.938484805419224e-05,
+      "loss": 0.0321,
+      "step": 5850
+    },
+    {
+      "action_loss": 0.009457248263061047,
+      "epoch": 5.260791366906475,
+      "step": 5850
+    },
+    {
+      "epoch": 5.260791366906475,
+      "step": 5850,
+      "torque_loss": 0.1029616966843605
+    },
+    {
+      "epoch": 5.260791366906475,
+      "force_loss": 0.008344221860170364,
+      "step": 5850
+    },
+    {
+      "epoch": 5.26978417266187,
+      "grad_norm": 1.0250165462493896,
+      "learning_rate": 9.938053105923894e-05,
+      "loss": 0.0305,
+      "step": 5860
+    },
+    {
+      "action_loss": 0.01243405882269144,
+      "epoch": 5.26978417266187,
+      "step": 5860
+    },
+    {
+      "epoch": 5.26978417266187,
+      "step": 5860,
+      "torque_loss": 0.12254921346902847
+    },
+    {
+      "epoch": 5.26978417266187,
+      "force_loss": 0.007495278958231211,
+      "step": 5860
+    },
+    {
+      "epoch": 5.278776978417266,
+      "grad_norm": 0.8344842195510864,
+      "learning_rate": 9.937619906378413e-05,
+      "loss": 0.0304,
+      "step": 5870
+    },
+    {
+      "action_loss": 0.014139537699520588,
+      "epoch": 5.278776978417266,
+      "step": 5870
+    },
+    {
+      "epoch": 5.278776978417266,
+      "step": 5870,
+      "torque_loss": 0.12871983647346497
+    },
+    {
+      "epoch": 5.278776978417266,
+      "force_loss": 0.008537528105080128,
+      "step": 5870
+    },
+    {
+      "epoch": 5.287769784172662,
+      "grad_norm": 1.210252046585083,
+      "learning_rate": 9.937185206914374e-05,
+      "loss": 0.0337,
+      "step": 5880
+    },
+    {
+      "action_loss": 0.032883673906326294,
+      "epoch": 5.287769784172662,
+      "step": 5880
+    },
+    {
+      "epoch": 5.287769784172662,
+      "step": 5880,
+      "torque_loss": 0.1661003828048706
+    },
+    {
+      "epoch": 5.287769784172662,
+      "force_loss": 0.03350367024540901,
+      "step": 5880
+    },
+    {
+      "epoch": 5.296762589928058,
+      "grad_norm": 1.44407057762146,
+      "learning_rate": 9.936749007663829e-05,
+      "loss": 0.0304,
+      "step": 5890
+    },
+    {
+      "action_loss": 0.01643591932952404,
+      "epoch": 5.296762589928058,
+      "step": 5890
+    },
+    {
+      "epoch": 5.296762589928058,
+      "step": 5890,
+      "torque_loss": 0.14750412106513977
+    },
+    {
+      "epoch": 5.296762589928058,
+      "force_loss": 0.030642440542578697,
+      "step": 5890
+    },
+    {
+      "epoch": 5.305755395683454,
+      "grad_norm": 0.7410878539085388,
+      "learning_rate": 9.93631130875928e-05,
+      "loss": 0.033,
+      "step": 5900
+    },
+    {
+      "action_loss": 0.02585108019411564,
+      "epoch": 5.305755395683454,
+      "step": 5900
+    },
+    {
+      "epoch": 5.305755395683454,
+      "step": 5900,
+      "torque_loss": 0.15912772715091705
+    },
+    {
+      "epoch": 5.305755395683454,
+      "force_loss": 0.026567354798316956,
+      "step": 5900
+    },
+    {
+      "epoch": 5.3147482014388485,
+      "grad_norm": 1.1565377712249756,
+      "learning_rate": 9.935872110333692e-05,
+      "loss": 0.0373,
+      "step": 5910
+    },
+    {
+      "action_loss": 0.02192777208983898,
+      "epoch": 5.3147482014388485,
+      "step": 5910
+    },
+    {
+      "epoch": 5.3147482014388485,
+      "step": 5910,
+      "torque_loss": 0.1286812275648117
+    },
+    {
+      "epoch": 5.3147482014388485,
+      "force_loss": 0.019034305587410927,
+      "step": 5910
+    },
+    {
+      "epoch": 5.323741007194244,
+      "grad_norm": 2.197715997695923,
+      "learning_rate": 9.935431412520484e-05,
+      "loss": 0.0362,
+      "step": 5920
+    },
+    {
+      "action_loss": 0.010740249417722225,
+      "epoch": 5.323741007194244,
+      "step": 5920
+    },
+    {
+      "epoch": 5.323741007194244,
+      "step": 5920,
+      "torque_loss": 0.18354232609272003
+    },
+    {
+      "epoch": 5.323741007194244,
+      "force_loss": 0.011155351996421814,
+      "step": 5920
+    },
+    {
+      "epoch": 5.33273381294964,
+      "grad_norm": 0.7228177189826965,
+      "learning_rate": 9.934989215453523e-05,
+      "loss": 0.0343,
+      "step": 5930
+    },
+    {
+      "action_loss": 0.023066185414791107,
+      "epoch": 5.33273381294964,
+      "step": 5930
+    },
+    {
+      "epoch": 5.33273381294964,
+      "step": 5930,
+      "torque_loss": 0.19458584487438202
+    },
+    {
+      "epoch": 5.33273381294964,
+      "force_loss": 0.034828025847673416,
+      "step": 5930
+    },
+    {
+      "epoch": 5.341726618705036,
+      "grad_norm": 0.5792932510375977,
+      "learning_rate": 9.934545519267139e-05,
+      "loss": 0.037,
+      "step": 5940
+    },
+    {
+      "action_loss": 0.024221688508987427,
+      "epoch": 5.341726618705036,
+      "step": 5940
+    },
+    {
+      "epoch": 5.341726618705036,
+      "step": 5940,
+      "torque_loss": 0.19522254168987274
+    },
+    {
+      "epoch": 5.341726618705036,
+      "force_loss": 0.012741059064865112,
+      "step": 5940
+    },
+    {
+      "epoch": 5.350719424460432,
+      "grad_norm": 0.9161556363105774,
+      "learning_rate": 9.934100324096117e-05,
+      "loss": 0.0347,
+      "step": 5950
+    },
+    {
+      "action_loss": 0.01206832006573677,
+      "epoch": 5.350719424460432,
+      "step": 5950
+    },
+    {
+      "epoch": 5.350719424460432,
+      "step": 5950,
+      "torque_loss": 0.15647456049919128
+    },
+    {
+      "epoch": 5.350719424460432,
+      "force_loss": 0.009277389384806156,
+      "step": 5950
+    },
+    {
+      "epoch": 5.359712230215827,
+      "grad_norm": 0.8598365783691406,
+      "learning_rate": 9.933653630075692e-05,
+      "loss": 0.0308,
+      "step": 5960
+    },
+    {
+      "action_loss": 0.013124617747962475,
+      "epoch": 5.359712230215827,
+      "step": 5960
+    },
+    {
+      "epoch": 5.359712230215827,
+      "step": 5960,
+      "torque_loss": 0.14300768077373505
+    },
+    {
+      "epoch": 5.359712230215827,
+      "force_loss": 0.01134935487061739,
+      "step": 5960
+    },
+    {
+      "epoch": 5.368705035971223,
+      "grad_norm": 1.8468637466430664,
+      "learning_rate": 9.93320543734156e-05,
+      "loss": 0.03,
+      "step": 5970
+    },
+    {
+      "action_loss": 0.01102599035948515,
+      "epoch": 5.368705035971223,
+      "step": 5970
+    },
+    {
+      "epoch": 5.368705035971223,
+      "step": 5970,
+      "torque_loss": 0.1412300020456314
+    },
+    {
+      "epoch": 5.368705035971223,
+      "force_loss": 0.008286123163998127,
+      "step": 5970
+    },
+    {
+      "epoch": 5.377697841726619,
+      "grad_norm": 1.8177632093429565,
+      "learning_rate": 9.932755746029871e-05,
+      "loss": 0.0295,
+      "step": 5980
+    },
+    {
+      "action_loss": 0.013801603578031063,
+      "epoch": 5.377697841726619,
+      "step": 5980
+    },
+    {
+      "epoch": 5.377697841726619,
+      "step": 5980,
+      "torque_loss": 0.10726621001958847
+    },
+    {
+      "epoch": 5.377697841726619,
+      "force_loss": 0.012256343849003315,
+      "step": 5980
+    },
+    {
+      "epoch": 5.386690647482014,
+      "grad_norm": 1.0670894384384155,
+      "learning_rate": 9.932304556277228e-05,
+      "loss": 0.0343,
+      "step": 5990
+    },
+    {
+      "action_loss": 0.018760504201054573,
+      "epoch": 5.386690647482014,
+      "step": 5990
+    },
+    {
+      "epoch": 5.386690647482014,
+      "step": 5990,
+      "torque_loss": 0.20835180580615997
+    },
+    {
+      "epoch": 5.386690647482014,
+      "force_loss": 0.017465850338339806,
+      "step": 5990
+    },
+    {
+      "epoch": 5.39568345323741,
+      "grad_norm": 1.05755615234375,
+      "learning_rate": 9.93185186822069e-05,
+      "loss": 0.0377,
+      "step": 6000
+    },
+    {
+      "action_loss": 0.02312903106212616,
+      "epoch": 5.39568345323741,
+      "step": 6000
+    },
+    {
+      "epoch": 5.39568345323741,
+      "step": 6000,
+      "torque_loss": 0.13508160412311554
+    },
+    {
+      "epoch": 5.39568345323741,
+      "force_loss": 0.0179379191249609,
+      "step": 6000
+    },
+    {
+      "epoch": 5.404676258992806,
+      "grad_norm": 1.0471394062042236,
+      "learning_rate": 9.931397681997773e-05,
+      "loss": 0.0331,
+      "step": 6010
+    },
+    {
+      "action_loss": 0.0077772438526153564,
+      "epoch": 5.404676258992806,
+      "step": 6010
+    },
+    {
+      "epoch": 5.404676258992806,
+      "step": 6010,
+      "torque_loss": 0.12868870794773102
+    },
+    {
+      "epoch": 5.404676258992806,
+      "force_loss": 0.008187138475477695,
+      "step": 6010
+    },
+    {
+      "epoch": 5.413669064748201,
+      "grad_norm": 0.59952712059021,
+      "learning_rate": 9.930941997746446e-05,
+      "loss": 0.0294,
+      "step": 6020
+    },
+    {
+      "action_loss": 0.014202840626239777,
+      "epoch": 5.413669064748201,
+      "step": 6020
+    },
+    {
+      "epoch": 5.413669064748201,
+      "step": 6020,
+      "torque_loss": 0.13312187790870667
+    },
+    {
+      "epoch": 5.413669064748201,
+      "force_loss": 0.01386262010782957,
+      "step": 6020
+    },
+    {
+      "epoch": 5.422661870503597,
+      "grad_norm": 0.424985408782959,
+      "learning_rate": 9.930484815605134e-05,
+      "loss": 0.0329,
+      "step": 6030
+    },
+    {
+      "action_loss": 0.01719067431986332,
+      "epoch": 5.422661870503597,
+      "step": 6030
+    },
+    {
+      "epoch": 5.422661870503597,
+      "step": 6030,
+      "torque_loss": 0.1441829800605774
+    },
+    {
+      "epoch": 5.422661870503597,
+      "force_loss": 0.014642399735748768,
+      "step": 6030
+    },
+    {
+      "epoch": 5.431654676258993,
+      "grad_norm": 0.7542428374290466,
+      "learning_rate": 9.930026135712717e-05,
+      "loss": 0.0365,
+      "step": 6040
+    },
+    {
+      "action_loss": 0.00925524067133665,
+      "epoch": 5.431654676258993,
+      "step": 6040
+    },
+    {
+      "epoch": 5.431654676258993,
+      "step": 6040,
+      "torque_loss": 0.1308252066373825
+    },
+    {
+      "epoch": 5.431654676258993,
+      "force_loss": 0.007872766815125942,
+      "step": 6040
+    },
+    {
+      "epoch": 5.440647482014389,
+      "grad_norm": 0.7442678213119507,
+      "learning_rate": 9.92956595820853e-05,
+      "loss": 0.0347,
+      "step": 6050
+    },
+    {
+      "action_loss": 0.006951652467250824,
+      "epoch": 5.440647482014389,
+      "step": 6050
+    },
+    {
+      "epoch": 5.440647482014389,
+      "step": 6050,
+      "torque_loss": 0.121300108730793
+    },
+    {
+      "epoch": 5.440647482014389,
+      "force_loss": 0.007886544801294804,
+      "step": 6050
+    },
+    {
+      "epoch": 5.4496402877697845,
+      "grad_norm": 0.8208723068237305,
+      "learning_rate": 9.929104283232362e-05,
+      "loss": 0.0302,
+      "step": 6060
+    },
+    {
+      "action_loss": 0.024792775511741638,
+      "epoch": 5.4496402877697845,
+      "step": 6060
+    },
+    {
+      "epoch": 5.4496402877697845,
+      "step": 6060,
+      "torque_loss": 0.17975036799907684
+    },
+    {
+      "epoch": 5.4496402877697845,
+      "force_loss": 0.036326829344034195,
+      "step": 6060
+    },
+    {
+      "epoch": 5.4586330935251794,
+      "grad_norm": 0.7009875774383545,
+      "learning_rate": 9.92864111092446e-05,
+      "loss": 0.0302,
+      "step": 6070
+    },
+    {
+      "action_loss": 0.015496629290282726,
+      "epoch": 5.4586330935251794,
+      "step": 6070
+    },
+    {
+      "epoch": 5.4586330935251794,
+      "step": 6070,
+      "torque_loss": 0.12660561501979828
+    },
+    {
+      "epoch": 5.4586330935251794,
+      "force_loss": 0.022680183872580528,
+      "step": 6070
+    },
+    {
+      "epoch": 5.467625899280575,
+      "grad_norm": 1.0963332653045654,
+      "learning_rate": 9.92817644142552e-05,
+      "loss": 0.0288,
+      "step": 6080
+    },
+    {
+      "action_loss": 0.01309314277023077,
+      "epoch": 5.467625899280575,
+      "step": 6080
+    },
+    {
+      "epoch": 5.467625899280575,
+      "step": 6080,
+      "torque_loss": 0.12677361071109772
+    },
+    {
+      "epoch": 5.467625899280575,
+      "force_loss": 0.00979540217667818,
+      "step": 6080
+    },
+    {
+      "epoch": 5.476618705035971,
+      "grad_norm": 1.43858802318573,
+      "learning_rate": 9.927710274876698e-05,
+      "loss": 0.035,
+      "step": 6090
+    },
+    {
+      "action_loss": 0.014266262762248516,
+      "epoch": 5.476618705035971,
+      "step": 6090
+    },
+    {
+      "epoch": 5.476618705035971,
+      "step": 6090,
+      "torque_loss": 0.14075371623039246
+    },
+    {
+      "epoch": 5.476618705035971,
+      "force_loss": 0.009991207160055637,
+      "step": 6090
+    },
+    {
+      "epoch": 5.485611510791367,
+      "grad_norm": 0.41903799772262573,
+      "learning_rate": 9.927242611419603e-05,
+      "loss": 0.0334,
+      "step": 6100
+    },
+    {
+      "action_loss": 0.04521968960762024,
+      "epoch": 5.485611510791367,
+      "step": 6100
+    },
+    {
+      "epoch": 5.485611510791367,
+      "step": 6100,
+      "torque_loss": 0.15832291543483734
+    },
+    {
+      "epoch": 5.485611510791367,
+      "force_loss": 0.03506319597363472,
+      "step": 6100
+    },
+    {
+      "epoch": 5.494604316546763,
+      "grad_norm": 0.8055093884468079,
+      "learning_rate": 9.926773451196301e-05,
+      "loss": 0.0395,
+      "step": 6110
+    },
+    {
+      "action_loss": 0.009163173846900463,
+      "epoch": 5.494604316546763,
+      "step": 6110
+    },
+    {
+      "epoch": 5.494604316546763,
+      "step": 6110,
+      "torque_loss": 0.08509886264801025
+    },
+    {
+      "epoch": 5.494604316546763,
+      "force_loss": 0.007788896095007658,
+      "step": 6110
+    },
+    {
+      "epoch": 5.503597122302159,
+      "grad_norm": 0.9088740944862366,
+      "learning_rate": 9.926302794349306e-05,
+      "loss": 0.0306,
+      "step": 6120
+    },
+    {
+      "action_loss": 0.030656510964035988,
+      "epoch": 5.503597122302159,
+      "step": 6120
+    },
+    {
+      "epoch": 5.503597122302159,
+      "step": 6120,
+      "torque_loss": 0.23709328472614288
+    },
+    {
+      "epoch": 5.503597122302159,
+      "force_loss": 0.030489442870020866,
+      "step": 6120
+    },
+    {
+      "epoch": 5.512589928057554,
+      "grad_norm": 1.0816494226455688,
+      "learning_rate": 9.925830641021594e-05,
+      "loss": 0.0398,
+      "step": 6130
+    },
+    {
+      "action_loss": 0.02450937032699585,
+      "epoch": 5.512589928057554,
+      "step": 6130
+    },
+    {
+      "epoch": 5.512589928057554,
+      "step": 6130,
+      "torque_loss": 0.16555556654930115
+    },
+    {
+      "epoch": 5.512589928057554,
+      "force_loss": 0.022962206974625587,
+      "step": 6130
+    },
+    {
+      "epoch": 5.5215827338129495,
+      "grad_norm": 0.9822729825973511,
+      "learning_rate": 9.925356991356593e-05,
+      "loss": 0.0367,
+      "step": 6140
+    },
+    {
+      "action_loss": 0.033981140702962875,
+      "epoch": 5.5215827338129495,
+      "step": 6140
+    },
+    {
+      "epoch": 5.5215827338129495,
+      "step": 6140,
+      "torque_loss": 0.16091638803482056
+    },
+    {
+      "epoch": 5.5215827338129495,
+      "force_loss": 0.01706506870687008,
+      "step": 6140
+    },
+    {
+      "epoch": 5.530575539568345,
+      "grad_norm": 0.41172078251838684,
+      "learning_rate": 9.924881845498184e-05,
+      "loss": 0.0297,
+      "step": 6150
+    },
+    {
+      "action_loss": 0.022704074159264565,
+      "epoch": 5.530575539568345,
+      "step": 6150
+    },
+    {
+      "epoch": 5.530575539568345,
+      "step": 6150,
+      "torque_loss": 0.1596505045890808
+    },
+    {
+      "epoch": 5.530575539568345,
+      "force_loss": 0.01549468468874693,
+      "step": 6150
+    },
+    {
+      "epoch": 5.539568345323741,
+      "grad_norm": 0.7721667885780334,
+      "learning_rate": 9.924405203590705e-05,
+      "loss": 0.0349,
+      "step": 6160
+    },
+    {
+      "action_loss": 0.0241011381149292,
+      "epoch": 5.539568345323741,
+      "step": 6160
+    },
+    {
+      "epoch": 5.539568345323741,
+      "step": 6160,
+      "torque_loss": 0.17437483370304108
+    },
+    {
+      "epoch": 5.539568345323741,
+      "force_loss": 0.026363296434283257,
+      "step": 6160
+    },
+    {
+      "epoch": 5.548561151079137,
+      "grad_norm": 1.2850946187973022,
+      "learning_rate": 9.923927065778946e-05,
+      "loss": 0.0345,
+      "step": 6170
+    },
+    {
+      "action_loss": 0.01621459610760212,
+      "epoch": 5.548561151079137,
+      "step": 6170
+    },
+    {
+      "epoch": 5.548561151079137,
+      "step": 6170,
+      "torque_loss": 0.1293100118637085
+    },
+    {
+      "epoch": 5.548561151079137,
+      "force_loss": 0.012890475802123547,
+      "step": 6170
+    },
+    {
+      "epoch": 5.557553956834532,
+      "grad_norm": 1.1787601709365845,
+      "learning_rate": 9.923447432208154e-05,
+      "loss": 0.0363,
+      "step": 6180
+    },
+    {
+      "action_loss": 0.010358665138483047,
+      "epoch": 5.557553956834532,
+      "step": 6180
+    },
+    {
+      "epoch": 5.557553956834532,
+      "step": 6180,
+      "torque_loss": 0.10722432285547256
+    },
+    {
+      "epoch": 5.557553956834532,
+      "force_loss": 0.007596289739012718,
+      "step": 6180
+    },
+    {
+      "epoch": 5.566546762589928,
+      "grad_norm": 0.7916187644004822,
+      "learning_rate": 9.922966303024027e-05,
+      "loss": 0.0317,
+      "step": 6190
+    },
+    {
+      "action_loss": 0.016377853229641914,
+      "epoch": 5.566546762589928,
+      "step": 6190
+    },
+    {
+      "epoch": 5.566546762589928,
+      "step": 6190,
+      "torque_loss": 0.18232186138629913
+    },
+    {
+      "epoch": 5.566546762589928,
+      "force_loss": 0.010570134967565536,
+      "step": 6190
+    },
+    {
+      "epoch": 5.575539568345324,
+      "grad_norm": 0.779882550239563,
+      "learning_rate": 9.922483678372721e-05,
+      "loss": 0.0423,
+      "step": 6200
+    },
+    {
+      "action_loss": 0.029877014458179474,
+      "epoch": 5.575539568345324,
+      "step": 6200
+    },
+    {
+      "epoch": 5.575539568345324,
+      "step": 6200,
+      "torque_loss": 0.14230668544769287
+    },
+    {
+      "epoch": 5.575539568345324,
+      "force_loss": 0.021593831479549408,
+      "step": 6200
+    },
+    {
+      "epoch": 5.58453237410072,
+      "grad_norm": 1.3249753713607788,
+      "learning_rate": 9.921999558400845e-05,
+      "loss": 0.0344,
+      "step": 6210
+    },
+    {
+      "action_loss": 0.016748622059822083,
+      "epoch": 5.58453237410072,
+      "step": 6210
+    },
+    {
+      "epoch": 5.58453237410072,
+      "step": 6210,
+      "torque_loss": 0.1596631556749344
+    },
+    {
+      "epoch": 5.58453237410072,
+      "force_loss": 0.01481617521494627,
+      "step": 6210
+    },
+    {
+      "epoch": 5.593525179856115,
+      "grad_norm": 0.6485111713409424,
+      "learning_rate": 9.92151394325546e-05,
+      "loss": 0.0407,
+      "step": 6220
+    },
+    {
+      "action_loss": 0.026288457214832306,
+      "epoch": 5.593525179856115,
+      "step": 6220
+    },
+    {
+      "epoch": 5.593525179856115,
+      "step": 6220,
+      "torque_loss": 0.13771633803844452
+    },
+    {
+      "epoch": 5.593525179856115,
+      "force_loss": 0.01655431091785431,
+      "step": 6220
+    },
+    {
+      "epoch": 5.602517985611511,
+      "grad_norm": 1.0226564407348633,
+      "learning_rate": 9.921026833084084e-05,
+      "loss": 0.0341,
+      "step": 6230
+    },
+    {
+      "action_loss": 0.02121158502995968,
+      "epoch": 5.602517985611511,
+      "step": 6230
+    },
+    {
+      "epoch": 5.602517985611511,
+      "step": 6230,
+      "torque_loss": 0.17241235077381134
+    },
+    {
+      "epoch": 5.602517985611511,
+      "force_loss": 0.016864560544490814,
+      "step": 6230
+    },
+    {
+      "epoch": 5.611510791366906,
+      "grad_norm": 0.9332301020622253,
+      "learning_rate": 9.920538228034689e-05,
+      "loss": 0.0308,
+      "step": 6240
+    },
+    {
+      "action_loss": 0.024339988827705383,
+      "epoch": 5.611510791366906,
+      "step": 6240
+    },
+    {
+      "epoch": 5.611510791366906,
+      "step": 6240,
+      "torque_loss": 0.14449548721313477
+    },
+    {
+      "epoch": 5.611510791366906,
+      "force_loss": 0.0294595118612051,
+      "step": 6240
+    },
+    {
+      "epoch": 5.620503597122302,
+      "grad_norm": 0.9044941663742065,
+      "learning_rate": 9.920048128255699e-05,
+      "loss": 0.0348,
+      "step": 6250
+    },
+    {
+      "action_loss": 0.011431700550019741,
+      "epoch": 5.620503597122302,
+      "step": 6250
+    },
+    {
+      "epoch": 5.620503597122302,
+      "step": 6250,
+      "torque_loss": 0.13255928456783295
+    },
+    {
+      "epoch": 5.620503597122302,
+      "force_loss": 0.009113145992159843,
+      "step": 6250
+    },
+    {
+      "epoch": 5.629496402877698,
+      "grad_norm": 1.7834445238113403,
+      "learning_rate": 9.919556533895995e-05,
+      "loss": 0.0395,
+      "step": 6260
+    },
+    {
+      "action_loss": 0.028955094516277313,
+      "epoch": 5.629496402877698,
+      "step": 6260
+    },
+    {
+      "epoch": 5.629496402877698,
+      "step": 6260,
+      "torque_loss": 0.1273440271615982
+    },
+    {
+      "epoch": 5.629496402877698,
+      "force_loss": 0.014864925295114517,
+      "step": 6260
+    },
+    {
+      "epoch": 5.638489208633094,
+      "grad_norm": 1.134705662727356,
+      "learning_rate": 9.919063445104907e-05,
+      "loss": 0.0371,
+      "step": 6270
+    },
+    {
+      "action_loss": 0.01801454648375511,
+      "epoch": 5.638489208633094,
+      "step": 6270
+    },
+    {
+      "epoch": 5.638489208633094,
+      "step": 6270,
+      "torque_loss": 0.1646595448255539
+    },
+    {
+      "epoch": 5.638489208633094,
+      "force_loss": 0.013859125785529613,
+      "step": 6270
+    },
+    {
+      "epoch": 5.647482014388489,
+      "grad_norm": 0.9329164624214172,
+      "learning_rate": 9.918568862032227e-05,
+      "loss": 0.0327,
+      "step": 6280
+    },
+    {
+      "action_loss": 0.02373000979423523,
+      "epoch": 5.647482014388489,
+      "step": 6280
+    },
+    {
+      "epoch": 5.647482014388489,
+      "step": 6280,
+      "torque_loss": 0.1502731889486313
+    },
+    {
+      "epoch": 5.647482014388489,
+      "force_loss": 0.01534126978367567,
+      "step": 6280
+    },
+    {
+      "epoch": 5.656474820143885,
+      "grad_norm": 0.8661342263221741,
+      "learning_rate": 9.918072784828194e-05,
+      "loss": 0.032,
+      "step": 6290
+    },
+    {
+      "action_loss": 0.016404708847403526,
+      "epoch": 5.656474820143885,
+      "step": 6290
+    },
+    {
+      "epoch": 5.656474820143885,
+      "step": 6290,
+      "torque_loss": 0.14487525820732117
+    },
+    {
+      "epoch": 5.656474820143885,
+      "force_loss": 0.015142071060836315,
+      "step": 6290
+    },
+    {
+      "epoch": 5.66546762589928,
+      "grad_norm": 0.9574587941169739,
+      "learning_rate": 9.917575213643501e-05,
+      "loss": 0.0378,
+      "step": 6300
+    },
+    {
+      "action_loss": 0.015606609173119068,
+      "epoch": 5.66546762589928,
+      "step": 6300
+    },
+    {
+      "epoch": 5.66546762589928,
+      "step": 6300,
+      "torque_loss": 0.14903031289577484
+    },
+    {
+      "epoch": 5.66546762589928,
+      "force_loss": 0.01487062405794859,
+      "step": 6300
+    },
+    {
+      "epoch": 5.674460431654676,
+      "grad_norm": 1.2378849983215332,
+      "learning_rate": 9.917076148629302e-05,
+      "loss": 0.032,
+      "step": 6310
+    },
+    {
+      "action_loss": 0.020867319777607918,
+      "epoch": 5.674460431654676,
+      "step": 6310
+    },
+    {
+      "epoch": 5.674460431654676,
+      "step": 6310,
+      "torque_loss": 0.09977815300226212
+    },
+    {
+      "epoch": 5.674460431654676,
+      "force_loss": 0.010460575111210346,
+      "step": 6310
+    },
+    {
+      "epoch": 5.683453237410072,
+      "grad_norm": 1.5370500087738037,
+      "learning_rate": 9.916575589937196e-05,
+      "loss": 0.0364,
+      "step": 6320
+    },
+    {
+      "action_loss": 0.01641274057328701,
+      "epoch": 5.683453237410072,
+      "step": 6320
+    },
+    {
+      "epoch": 5.683453237410072,
+      "step": 6320,
+      "torque_loss": 0.1378653198480606
+    },
+    {
+      "epoch": 5.683453237410072,
+      "force_loss": 0.020352089777588844,
+      "step": 6320
+    },
+    {
+      "epoch": 5.692446043165468,
+      "grad_norm": 0.7210316061973572,
+      "learning_rate": 9.916073537719239e-05,
+      "loss": 0.0326,
+      "step": 6330
+    },
+    {
+      "action_loss": 0.01788381300866604,
+      "epoch": 5.692446043165468,
+      "step": 6330
+    },
+    {
+      "epoch": 5.692446043165468,
+      "step": 6330,
+      "torque_loss": 0.1495245099067688
+    },
+    {
+      "epoch": 5.692446043165468,
+      "force_loss": 0.008984549902379513,
+      "step": 6330
+    },
+    {
+      "epoch": 5.701438848920863,
+      "grad_norm": 0.8334928154945374,
+      "learning_rate": 9.915569992127944e-05,
+      "loss": 0.0361,
+      "step": 6340
+    },
+    {
+      "action_loss": 0.022939607501029968,
+      "epoch": 5.701438848920863,
+      "step": 6340
+    },
+    {
+      "epoch": 5.701438848920863,
+      "step": 6340,
+      "torque_loss": 0.15945382416248322
+    },
+    {
+      "epoch": 5.701438848920863,
+      "force_loss": 0.018436981365084648,
+      "step": 6340
+    },
+    {
+      "epoch": 5.710431654676259,
+      "grad_norm": 0.4778817892074585,
+      "learning_rate": 9.915064953316273e-05,
+      "loss": 0.0297,
+      "step": 6350
+    },
+    {
+      "action_loss": 0.010276437737047672,
+      "epoch": 5.710431654676259,
+      "step": 6350
+    },
+    {
+      "epoch": 5.710431654676259,
+      "step": 6350,
+      "torque_loss": 0.15242727100849152
+    },
+    {
+      "epoch": 5.710431654676259,
+      "force_loss": 0.007281759288161993,
+      "step": 6350
+    },
+    {
+      "epoch": 5.719424460431655,
+      "grad_norm": 0.6238363981246948,
+      "learning_rate": 9.914558421437645e-05,
+      "loss": 0.0305,
+      "step": 6360
+    },
+    {
+      "action_loss": 0.012519605457782745,
+      "epoch": 5.719424460431655,
+      "step": 6360
+    },
+    {
+      "epoch": 5.719424460431655,
+      "step": 6360,
+      "torque_loss": 0.10555735230445862
+    },
+    {
+      "epoch": 5.719424460431655,
+      "force_loss": 0.012490061111748219,
+      "step": 6360
+    },
+    {
+      "epoch": 5.7284172661870505,
+      "grad_norm": 0.6641746759414673,
+      "learning_rate": 9.914050396645929e-05,
+      "loss": 0.0335,
+      "step": 6370
+    },
+    {
+      "action_loss": 0.022814318537712097,
+      "epoch": 5.7284172661870505,
+      "step": 6370
+    },
+    {
+      "epoch": 5.7284172661870505,
+      "step": 6370,
+      "torque_loss": 0.10337921231985092
+    },
+    {
+      "epoch": 5.7284172661870505,
+      "force_loss": 0.013225001282989979,
+      "step": 6370
+    },
+    {
+      "epoch": 5.737410071942446,
+      "grad_norm": 0.5852411389350891,
+      "learning_rate": 9.913540879095452e-05,
+      "loss": 0.0382,
+      "step": 6380
+    },
+    {
+      "action_loss": 0.02187674678862095,
+      "epoch": 5.737410071942446,
+      "step": 6380
+    },
+    {
+      "epoch": 5.737410071942446,
+      "step": 6380,
+      "torque_loss": 0.18658983707427979
+    },
+    {
+      "epoch": 5.737410071942446,
+      "force_loss": 0.018079740926623344,
+      "step": 6380
+    },
+    {
+      "epoch": 5.746402877697841,
+      "grad_norm": 0.8768635988235474,
+      "learning_rate": 9.913029868940987e-05,
+      "loss": 0.038,
+      "step": 6390
+    },
+    {
+      "action_loss": 0.05202604830265045,
+      "epoch": 5.746402877697841,
+      "step": 6390
+    },
+    {
+      "epoch": 5.746402877697841,
+      "step": 6390,
+      "torque_loss": 0.1624857634305954
+    },
+    {
+      "epoch": 5.746402877697841,
+      "force_loss": 0.03932690620422363,
+      "step": 6390
+    },
+    {
+      "epoch": 5.755395683453237,
+      "grad_norm": 1.5830304622650146,
+      "learning_rate": 9.912517366337772e-05,
+      "loss": 0.0371,
+      "step": 6400
+    },
+    {
+      "action_loss": 0.013611477799713612,
+      "epoch": 5.755395683453237,
+      "step": 6400
+    },
+    {
+      "epoch": 5.755395683453237,
+      "step": 6400,
+      "torque_loss": 0.1407223343849182
+    },
+    {
+      "epoch": 5.755395683453237,
+      "force_loss": 0.010399263352155685,
+      "step": 6400
+    },
+    {
+      "epoch": 5.764388489208633,
+      "grad_norm": 0.5996264219284058,
+      "learning_rate": 9.912003371441487e-05,
+      "loss": 0.0356,
+      "step": 6410
+    },
+    {
+      "action_loss": 0.02901286445558071,
+      "epoch": 5.764388489208633,
+      "step": 6410
+    },
+    {
+      "epoch": 5.764388489208633,
+      "step": 6410,
+      "torque_loss": 0.1307028979063034
+    },
+    {
+      "epoch": 5.764388489208633,
+      "force_loss": 0.017980530858039856,
+      "step": 6410
+    },
+    {
+      "epoch": 5.773381294964029,
+      "grad_norm": 0.5171142816543579,
+      "learning_rate": 9.911487884408271e-05,
+      "loss": 0.0331,
+      "step": 6420
+    },
+    {
+      "action_loss": 0.009266541339457035,
+      "epoch": 5.773381294964029,
+      "step": 6420
+    },
+    {
+      "epoch": 5.773381294964029,
+      "step": 6420,
+      "torque_loss": 0.11346965283155441
+    },
+    {
+      "epoch": 5.773381294964029,
+      "force_loss": 0.005412036087363958,
+      "step": 6420
+    },
+    {
+      "epoch": 5.782374100719425,
+      "grad_norm": 1.4605542421340942,
+      "learning_rate": 9.910970905394719e-05,
+      "loss": 0.0347,
+      "step": 6430
+    },
+    {
+      "action_loss": 0.012880965135991573,
+      "epoch": 5.782374100719425,
+      "step": 6430
+    },
+    {
+      "epoch": 5.782374100719425,
+      "step": 6430,
+      "torque_loss": 0.12281922250986099
+    },
+    {
+      "epoch": 5.782374100719425,
+      "force_loss": 0.008332737721502781,
+      "step": 6430
+    },
+    {
+      "epoch": 5.7913669064748206,
+      "grad_norm": 0.7657610774040222,
+      "learning_rate": 9.91045243455787e-05,
+      "loss": 0.0284,
+      "step": 6440
+    },
+    {
+      "action_loss": 0.01047398429363966,
+      "epoch": 5.7913669064748206,
+      "step": 6440
+    },
+    {
+      "epoch": 5.7913669064748206,
+      "step": 6440,
+      "torque_loss": 0.10461390018463135
+    },
+    {
+      "epoch": 5.7913669064748206,
+      "force_loss": 0.008071732707321644,
+      "step": 6440
+    },
+    {
+      "epoch": 5.8003597122302155,
+      "grad_norm": 0.6476050615310669,
+      "learning_rate": 9.909932472055225e-05,
+      "loss": 0.0364,
+      "step": 6450
+    },
+    {
+      "action_loss": 0.0158433485776186,
+      "epoch": 5.8003597122302155,
+      "step": 6450
+    },
+    {
+      "epoch": 5.8003597122302155,
+      "step": 6450,
+      "torque_loss": 0.13918693363666534
+    },
+    {
+      "epoch": 5.8003597122302155,
+      "force_loss": 0.009686020202934742,
+      "step": 6450
+    },
+    {
+      "epoch": 5.809352517985611,
+      "grad_norm": 0.5028572082519531,
+      "learning_rate": 9.909411018044734e-05,
+      "loss": 0.0299,
+      "step": 6460
+    },
+    {
+      "action_loss": 0.01332776714116335,
+      "epoch": 5.809352517985611,
+      "step": 6460
+    },
+    {
+      "epoch": 5.809352517985611,
+      "step": 6460,
+      "torque_loss": 0.12181492894887924
+    },
+    {
+      "epoch": 5.809352517985611,
+      "force_loss": 0.010035021230578423,
+      "step": 6460
+    },
+    {
+      "epoch": 5.818345323741007,
+      "grad_norm": 0.8717068433761597,
+      "learning_rate": 9.908888072684802e-05,
+      "loss": 0.0329,
+      "step": 6470
+    },
+    {
+      "action_loss": 0.01626739092171192,
+      "epoch": 5.818345323741007,
+      "step": 6470
+    },
+    {
+      "epoch": 5.818345323741007,
+      "step": 6470,
+      "torque_loss": 0.12440016865730286
+    },
+    {
+      "epoch": 5.818345323741007,
+      "force_loss": 0.013880801387131214,
+      "step": 6470
+    },
+    {
+      "epoch": 5.827338129496403,
+      "grad_norm": 1.1392223834991455,
+      "learning_rate": 9.908363636134285e-05,
+      "loss": 0.0282,
+      "step": 6480
+    },
+    {
+      "action_loss": 0.013950206339359283,
+      "epoch": 5.827338129496403,
+      "step": 6480
+    },
+    {
+      "epoch": 5.827338129496403,
+      "step": 6480,
+      "torque_loss": 0.1354544758796692
+    },
+    {
+      "epoch": 5.827338129496403,
+      "force_loss": 0.010613635182380676,
+      "step": 6480
+    },
+    {
+      "epoch": 5.836330935251799,
+      "grad_norm": 0.6433037519454956,
+      "learning_rate": 9.907837708552493e-05,
+      "loss": 0.0276,
+      "step": 6490
+    },
+    {
+      "action_loss": 0.0168133731931448,
+      "epoch": 5.836330935251799,
+      "step": 6490
+    },
+    {
+      "epoch": 5.836330935251799,
+      "step": 6490,
+      "torque_loss": 0.1305723488330841
+    },
+    {
+      "epoch": 5.836330935251799,
+      "force_loss": 0.024245254695415497,
+      "step": 6490
+    },
+    {
+      "epoch": 5.845323741007194,
+      "grad_norm": 1.059878945350647,
+      "learning_rate": 9.90731029009919e-05,
+      "loss": 0.0291,
+      "step": 6500
+    },
+    {
+      "action_loss": 0.027319571003317833,
+      "epoch": 5.845323741007194,
+      "step": 6500
+    },
+    {
+      "epoch": 5.845323741007194,
+      "step": 6500,
+      "torque_loss": 0.16758990287780762
+    },
+    {
+      "epoch": 5.845323741007194,
+      "force_loss": 0.025982022285461426,
+      "step": 6500
+    },
+    {
+      "epoch": 5.85431654676259,
+      "grad_norm": 1.482749104499817,
+      "learning_rate": 9.906781380934589e-05,
+      "loss": 0.0439,
+      "step": 6510
+    },
+    {
+      "action_loss": 0.02713160216808319,
+      "epoch": 5.85431654676259,
+      "step": 6510
+    },
+    {
+      "epoch": 5.85431654676259,
+      "step": 6510,
+      "torque_loss": 0.16857409477233887
+    },
+    {
+      "epoch": 5.85431654676259,
+      "force_loss": 0.03157292306423187,
+      "step": 6510
+    },
+    {
+      "epoch": 5.863309352517986,
+      "grad_norm": 0.8787999749183655,
+      "learning_rate": 9.906250981219362e-05,
+      "loss": 0.0347,
+      "step": 6520
+    },
+    {
+      "action_loss": 0.012791815213859081,
+      "epoch": 5.863309352517986,
+      "step": 6520
+    },
+    {
+      "epoch": 5.863309352517986,
+      "step": 6520,
+      "torque_loss": 0.19160287082195282
+    },
+    {
+      "epoch": 5.863309352517986,
+      "force_loss": 0.008362742140889168,
+      "step": 6520
+    },
+    {
+      "epoch": 5.872302158273381,
+      "grad_norm": 0.752284586429596,
+      "learning_rate": 9.905719091114628e-05,
+      "loss": 0.0296,
+      "step": 6530
+    },
+    {
+      "action_loss": 0.012131291441619396,
+      "epoch": 5.872302158273381,
+      "step": 6530
+    },
+    {
+      "epoch": 5.872302158273381,
+      "step": 6530,
+      "torque_loss": 0.12434646487236023
+    },
+    {
+      "epoch": 5.872302158273381,
+      "force_loss": 0.010255965404212475,
+      "step": 6530
+    },
+    {
+      "epoch": 5.881294964028777,
+      "grad_norm": 1.9539376497268677,
+      "learning_rate": 9.905185710781964e-05,
+      "loss": 0.039,
+      "step": 6540
+    },
+    {
+      "action_loss": 0.013075563125312328,
+      "epoch": 5.881294964028777,
+      "step": 6540
+    },
+    {
+      "epoch": 5.881294964028777,
+      "step": 6540,
+      "torque_loss": 0.14066021144390106
+    },
+    {
+      "epoch": 5.881294964028777,
+      "force_loss": 0.009071187116205692,
+      "step": 6540
+    },
+    {
+      "epoch": 5.890287769784173,
+      "grad_norm": 0.39630407094955444,
+      "learning_rate": 9.904650840383392e-05,
+      "loss": 0.035,
+      "step": 6550
+    },
+    {
+      "action_loss": 0.012409411370754242,
+      "epoch": 5.890287769784173,
+      "step": 6550
+    },
+    {
+      "epoch": 5.890287769784173,
+      "step": 6550,
+      "torque_loss": 0.10571638494729996
+    },
+    {
+      "epoch": 5.890287769784173,
+      "force_loss": 0.012267733924090862,
+      "step": 6550
+    },
+    {
+      "epoch": 5.899280575539568,
+      "grad_norm": 1.1717088222503662,
+      "learning_rate": 9.904114480081397e-05,
+      "loss": 0.027,
+      "step": 6560
+    },
+    {
+      "action_loss": 0.02597968839108944,
+      "epoch": 5.899280575539568,
+      "step": 6560
+    },
+    {
+      "epoch": 5.899280575539568,
+      "step": 6560,
+      "torque_loss": 0.1780753880739212
+    },
+    {
+      "epoch": 5.899280575539568,
+      "force_loss": 0.02346406877040863,
+      "step": 6560
+    },
+    {
+      "epoch": 5.908273381294964,
+      "grad_norm": 0.9946366548538208,
+      "learning_rate": 9.903576630038906e-05,
+      "loss": 0.0375,
+      "step": 6570
+    },
+    {
+      "action_loss": 0.015300107188522816,
+      "epoch": 5.908273381294964,
+      "step": 6570
+    },
+    {
+      "epoch": 5.908273381294964,
+      "step": 6570,
+      "torque_loss": 0.16440166532993317
+    },
+    {
+      "epoch": 5.908273381294964,
+      "force_loss": 0.016858940944075584,
+      "step": 6570
+    },
+    {
+      "epoch": 5.91726618705036,
+      "grad_norm": 0.5869100689888,
+      "learning_rate": 9.903037290419309e-05,
+      "loss": 0.0309,
+      "step": 6580
+    },
+    {
+      "action_loss": 0.014476188458502293,
+      "epoch": 5.91726618705036,
+      "step": 6580
+    },
+    {
+      "epoch": 5.91726618705036,
+      "step": 6580,
+      "torque_loss": 0.127194344997406
+    },
+    {
+      "epoch": 5.91726618705036,
+      "force_loss": 0.013970345258712769,
+      "step": 6580
+    },
+    {
+      "epoch": 5.926258992805756,
+      "grad_norm": 0.7035734057426453,
+      "learning_rate": 9.902496461386439e-05,
+      "loss": 0.0305,
+      "step": 6590
+    },
+    {
+      "action_loss": 0.010824356228113174,
+      "epoch": 5.926258992805756,
+      "step": 6590
+    },
+    {
+      "epoch": 5.926258992805756,
+      "step": 6590,
+      "torque_loss": 0.1059899553656578
+    },
+    {
+      "epoch": 5.926258992805756,
+      "force_loss": 0.01181915681809187,
+      "step": 6590
+    },
+    {
+      "epoch": 5.935251798561151,
+      "grad_norm": 1.3031914234161377,
+      "learning_rate": 9.901954143104588e-05,
+      "loss": 0.0314,
+      "step": 6600
+    },
+    {
+      "action_loss": 0.010905037634074688,
+      "epoch": 5.935251798561151,
+      "step": 6600
+    },
+    {
+      "epoch": 5.935251798561151,
+      "step": 6600,
+      "torque_loss": 0.07843795418739319
+    },
+    {
+      "epoch": 5.935251798561151,
+      "force_loss": 0.006029366049915552,
+      "step": 6600
+    },
+    {
+      "epoch": 5.944244604316546,
+      "grad_norm": 0.8794500231742859,
+      "learning_rate": 9.901410335738496e-05,
+      "loss": 0.025,
+      "step": 6610
+    },
+    {
+      "action_loss": 0.022479886189103127,
+      "epoch": 5.944244604316546,
+      "step": 6610
+    },
+    {
+      "epoch": 5.944244604316546,
+      "step": 6610,
+      "torque_loss": 0.14966750144958496
+    },
+    {
+      "epoch": 5.944244604316546,
+      "force_loss": 0.013774256221950054,
+      "step": 6610
+    },
+    {
+      "epoch": 5.953237410071942,
+      "grad_norm": 0.49869632720947266,
+      "learning_rate": 9.900865039453358e-05,
+      "loss": 0.0342,
+      "step": 6620
+    },
+    {
+      "action_loss": 0.00962988194078207,
+      "epoch": 5.953237410071942,
+      "step": 6620
+    },
+    {
+      "epoch": 5.953237410071942,
+      "step": 6620,
+      "torque_loss": 0.12475377321243286
+    },
+    {
+      "epoch": 5.953237410071942,
+      "force_loss": 0.008503995835781097,
+      "step": 6620
+    },
+    {
+      "epoch": 5.962230215827338,
+      "grad_norm": 1.0734554529190063,
+      "learning_rate": 9.900318254414821e-05,
+      "loss": 0.0315,
+      "step": 6630
+    },
+    {
+      "action_loss": 0.020362840965390205,
+      "epoch": 5.962230215827338,
+      "step": 6630
+    },
+    {
+      "epoch": 5.962230215827338,
+      "step": 6630,
+      "torque_loss": 0.22567391395568848
+    },
+    {
+      "epoch": 5.962230215827338,
+      "force_loss": 0.010840315371751785,
+      "step": 6630
+    },
+    {
+      "epoch": 5.971223021582734,
+      "grad_norm": 0.7541504502296448,
+      "learning_rate": 9.899769980788985e-05,
+      "loss": 0.0287,
+      "step": 6640
+    },
+    {
+      "action_loss": 0.009102792479097843,
+      "epoch": 5.971223021582734,
+      "step": 6640
+    },
+    {
+      "epoch": 5.971223021582734,
+      "step": 6640,
+      "torque_loss": 0.11625024676322937
+    },
+    {
+      "epoch": 5.971223021582734,
+      "force_loss": 0.0074315182864665985,
+      "step": 6640
+    },
+    {
+      "epoch": 5.98021582733813,
+      "grad_norm": 1.1257209777832031,
+      "learning_rate": 9.899220218742398e-05,
+      "loss": 0.0264,
+      "step": 6650
+    },
+    {
+      "action_loss": 0.008441298268735409,
+      "epoch": 5.98021582733813,
+      "step": 6650
+    },
+    {
+      "epoch": 5.98021582733813,
+      "step": 6650,
+      "torque_loss": 0.12980546057224274
+    },
+    {
+      "epoch": 5.98021582733813,
+      "force_loss": 0.008751760236918926,
+      "step": 6650
+    },
+    {
+      "epoch": 5.989208633093525,
+      "grad_norm": 1.6277531385421753,
+      "learning_rate": 9.898668968442066e-05,
+      "loss": 0.0373,
+      "step": 6660
+    },
+    {
+      "action_loss": 0.02496408484876156,
+      "epoch": 5.989208633093525,
+      "step": 6660
+    },
+    {
+      "epoch": 5.989208633093525,
+      "step": 6660,
+      "torque_loss": 0.16968567669391632
+    },
+    {
+      "epoch": 5.989208633093525,
+      "force_loss": 0.019812721759080887,
+      "step": 6660
+    },
+    {
+      "epoch": 5.998201438848921,
+      "grad_norm": 0.550177812576294,
+      "learning_rate": 9.898116230055443e-05,
+      "loss": 0.0327,
+      "step": 6670
+    },
+    {
+      "action_loss": 0.01624131202697754,
+      "epoch": 5.998201438848921,
+      "step": 6670
+    },
+    {
+      "epoch": 5.998201438848921,
+      "step": 6670,
+      "torque_loss": 0.1611861139535904
+    },
+    {
+      "epoch": 5.998201438848921,
+      "force_loss": 0.009481154382228851,
+      "step": 6670
+    },
+    {
+      "epoch": 6.0071942446043165,
+      "grad_norm": 0.3939690589904785,
+      "learning_rate": 9.897562003750437e-05,
+      "loss": 0.0286,
+      "step": 6680
+    },
+    {
+      "action_loss": 0.022057220339775085,
+      "epoch": 6.0071942446043165,
+      "step": 6680
+    },
+    {
+      "epoch": 6.0071942446043165,
+      "step": 6680,
+      "torque_loss": 0.18737749755382538
+    },
+    {
+      "epoch": 6.0071942446043165,
+      "force_loss": 0.02907741628587246,
+      "step": 6680
+    },
+    {
+      "epoch": 6.016187050359712,
+      "grad_norm": 0.7050179243087769,
+      "learning_rate": 9.897006289695407e-05,
+      "loss": 0.0336,
+      "step": 6690
+    },
+    {
+      "action_loss": 0.021385440602898598,
+      "epoch": 6.016187050359712,
+      "step": 6690
+    },
+    {
+      "epoch": 6.016187050359712,
+      "step": 6690,
+      "torque_loss": 0.1306765228509903
+    },
+    {
+      "epoch": 6.016187050359712,
+      "force_loss": 0.020024040713906288,
+      "step": 6690
+    },
+    {
+      "epoch": 6.025179856115108,
+      "grad_norm": 1.4871512651443481,
+      "learning_rate": 9.896449088059164e-05,
+      "loss": 0.0368,
+      "step": 6700
+    },
+    {
+      "action_loss": 0.01957033760845661,
+      "epoch": 6.025179856115108,
+      "step": 6700
+    },
+    {
+      "epoch": 6.025179856115108,
+      "step": 6700,
+      "torque_loss": 0.1466013640165329
+    },
+    {
+      "epoch": 6.025179856115108,
+      "force_loss": 0.017637547105550766,
+      "step": 6700
+    },
+    {
+      "epoch": 6.034172661870503,
+      "grad_norm": 0.9981018304824829,
+      "learning_rate": 9.89589039901097e-05,
+      "loss": 0.0329,
+      "step": 6710
+    },
+    {
+      "action_loss": 0.008246971294283867,
+      "epoch": 6.034172661870503,
+      "step": 6710
+    },
+    {
+      "epoch": 6.034172661870503,
+      "step": 6710,
+      "torque_loss": 0.16157709062099457
+    },
+    {
+      "epoch": 6.034172661870503,
+      "force_loss": 0.007431477308273315,
+      "step": 6710
+    },
+    {
+      "epoch": 6.043165467625899,
+      "grad_norm": 0.9391647577285767,
+      "learning_rate": 9.895330222720542e-05,
+      "loss": 0.0301,
+      "step": 6720
+    },
+    {
+      "action_loss": 0.013659723103046417,
+      "epoch": 6.043165467625899,
+      "step": 6720
+    },
+    {
+      "epoch": 6.043165467625899,
+      "step": 6720,
+      "torque_loss": 0.14471757411956787
+    },
+    {
+      "epoch": 6.043165467625899,
+      "force_loss": 0.012654148042201996,
+      "step": 6720
+    },
+    {
+      "epoch": 6.052158273381295,
+      "grad_norm": 0.7203466892242432,
+      "learning_rate": 9.894768559358047e-05,
+      "loss": 0.0442,
+      "step": 6730
+    },
+    {
+      "action_loss": 0.016545280814170837,
+      "epoch": 6.052158273381295,
+      "step": 6730
+    },
+    {
+      "epoch": 6.052158273381295,
+      "step": 6730,
+      "torque_loss": 0.18547111749649048
+    },
+    {
+      "epoch": 6.052158273381295,
+      "force_loss": 0.01861295849084854,
+      "step": 6730
+    },
+    {
+      "epoch": 6.061151079136691,
+      "grad_norm": 1.0598502159118652,
+      "learning_rate": 9.894205409094101e-05,
+      "loss": 0.0381,
+      "step": 6740
+    },
+    {
+      "action_loss": 0.03386823832988739,
+      "epoch": 6.061151079136691,
+      "step": 6740
+    },
+    {
+      "epoch": 6.061151079136691,
+      "step": 6740,
+      "torque_loss": 0.23352237045764923
+    },
+    {
+      "epoch": 6.061151079136691,
+      "force_loss": 0.025537485256791115,
+      "step": 6740
+    },
+    {
+      "epoch": 6.070143884892087,
+      "grad_norm": 0.6828194856643677,
+      "learning_rate": 9.893640772099777e-05,
+      "loss": 0.0313,
+      "step": 6750
+    },
+    {
+      "action_loss": 0.018237030133605003,
+      "epoch": 6.070143884892087,
+      "step": 6750
+    },
+    {
+      "epoch": 6.070143884892087,
+      "step": 6750,
+      "torque_loss": 0.13684232532978058
+    },
+    {
+      "epoch": 6.070143884892087,
+      "force_loss": 0.015216675586998463,
+      "step": 6750
+    },
+    {
+      "epoch": 6.079136690647482,
+      "grad_norm": 1.2732737064361572,
+      "learning_rate": 9.893074648546595e-05,
+      "loss": 0.0381,
+      "step": 6760
+    },
+    {
+      "action_loss": 0.010769792832434177,
+      "epoch": 6.079136690647482,
+      "step": 6760
+    },
+    {
+      "epoch": 6.079136690647482,
+      "step": 6760,
+      "torque_loss": 0.10087388008832932
+    },
+    {
+      "epoch": 6.079136690647482,
+      "force_loss": 0.009020409546792507,
+      "step": 6760
+    },
+    {
+      "epoch": 6.088129496402877,
+      "grad_norm": 0.6518918871879578,
+      "learning_rate": 9.892507038606528e-05,
+      "loss": 0.0377,
+      "step": 6770
+    },
+    {
+      "action_loss": 0.010382195934653282,
+      "epoch": 6.088129496402877,
+      "step": 6770
+    },
+    {
+      "epoch": 6.088129496402877,
+      "step": 6770,
+      "torque_loss": 0.1361764818429947
+    },
+    {
+      "epoch": 6.088129496402877,
+      "force_loss": 0.011860187165439129,
+      "step": 6770
+    },
+    {
+      "epoch": 6.097122302158273,
+      "grad_norm": 0.6612163186073303,
+      "learning_rate": 9.891937942452003e-05,
+      "loss": 0.0325,
+      "step": 6780
+    },
+    {
+      "action_loss": 0.02234676480293274,
+      "epoch": 6.097122302158273,
+      "step": 6780
+    },
+    {
+      "epoch": 6.097122302158273,
+      "step": 6780,
+      "torque_loss": 0.14713360369205475
+    },
+    {
+      "epoch": 6.097122302158273,
+      "force_loss": 0.013437267392873764,
+      "step": 6780
+    },
+    {
+      "epoch": 6.106115107913669,
+      "grad_norm": 0.785275936126709,
+      "learning_rate": 9.891367360255895e-05,
+      "loss": 0.0328,
+      "step": 6790
+    },
+    {
+      "action_loss": 0.022016048431396484,
+      "epoch": 6.106115107913669,
+      "step": 6790
+    },
+    {
+      "epoch": 6.106115107913669,
+      "step": 6790,
+      "torque_loss": 0.19155645370483398
+    },
+    {
+      "epoch": 6.106115107913669,
+      "force_loss": 0.026529178023338318,
+      "step": 6790
+    },
+    {
+      "epoch": 6.115107913669065,
+      "grad_norm": 1.501050591468811,
+      "learning_rate": 9.890795292191532e-05,
+      "loss": 0.0402,
+      "step": 6800
+    },
+    {
+      "action_loss": 0.02979736216366291,
+      "epoch": 6.115107913669065,
+      "step": 6800
+    },
+    {
+      "epoch": 6.115107913669065,
+      "step": 6800,
+      "torque_loss": 0.13777391612529755
+    },
+    {
+      "epoch": 6.115107913669065,
+      "force_loss": 0.021183064207434654,
+      "step": 6800
+    },
+    {
+      "epoch": 6.124100719424461,
+      "grad_norm": 0.622193455696106,
+      "learning_rate": 9.890221738432694e-05,
+      "loss": 0.0386,
+      "step": 6810
+    },
+    {
+      "action_loss": 0.00862081628292799,
+      "epoch": 6.124100719424461,
+      "step": 6810
+    },
+    {
+      "epoch": 6.124100719424461,
+      "step": 6810,
+      "torque_loss": 0.13776084780693054
+    },
+    {
+      "epoch": 6.124100719424461,
+      "force_loss": 0.013538769446313381,
+      "step": 6810
+    },
+    {
+      "epoch": 6.133093525179856,
+      "grad_norm": 0.969052791595459,
+      "learning_rate": 9.88964669915361e-05,
+      "loss": 0.0291,
+      "step": 6820
+    },
+    {
+      "action_loss": 0.009601211175322533,
+      "epoch": 6.133093525179856,
+      "step": 6820
+    },
+    {
+      "epoch": 6.133093525179856,
+      "step": 6820,
+      "torque_loss": 0.10650753229856491
+    },
+    {
+      "epoch": 6.133093525179856,
+      "force_loss": 0.011245968751609325,
+      "step": 6820
+    },
+    {
+      "epoch": 6.142086330935252,
+      "grad_norm": 0.682878851890564,
+      "learning_rate": 9.889070174528963e-05,
+      "loss": 0.0383,
+      "step": 6830
+    },
+    {
+      "action_loss": 0.010681535117328167,
+      "epoch": 6.142086330935252,
+      "step": 6830
+    },
+    {
+      "epoch": 6.142086330935252,
+      "step": 6830,
+      "torque_loss": 0.09342696517705917
+    },
+    {
+      "epoch": 6.142086330935252,
+      "force_loss": 0.010287155397236347,
+      "step": 6830
+    },
+    {
+      "epoch": 6.151079136690647,
+      "grad_norm": 0.9495276212692261,
+      "learning_rate": 9.888492164733883e-05,
+      "loss": 0.0307,
+      "step": 6840
+    },
+    {
+      "action_loss": 0.016709303483366966,
+      "epoch": 6.151079136690647,
+      "step": 6840
+    },
+    {
+      "epoch": 6.151079136690647,
+      "step": 6840,
+      "torque_loss": 0.12592746317386627
+    },
+    {
+      "epoch": 6.151079136690647,
+      "force_loss": 0.01575835421681404,
+      "step": 6840
+    },
+    {
+      "epoch": 6.160071942446043,
+      "grad_norm": 0.3741442859172821,
+      "learning_rate": 9.88791266994396e-05,
+      "loss": 0.0336,
+      "step": 6850
+    },
+    {
+      "action_loss": 0.010481658391654491,
+      "epoch": 6.160071942446043,
+      "step": 6850
+    },
+    {
+      "epoch": 6.160071942446043,
+      "step": 6850,
+      "torque_loss": 0.12324514240026474
+    },
+    {
+      "epoch": 6.160071942446043,
+      "force_loss": 0.008142626844346523,
+      "step": 6850
+    },
+    {
+      "epoch": 6.169064748201439,
+      "grad_norm": 0.5587148666381836,
+      "learning_rate": 9.887331690335223e-05,
+      "loss": 0.0297,
+      "step": 6860
+    },
+    {
+      "action_loss": 0.018462836742401123,
+      "epoch": 6.169064748201439,
+      "step": 6860
+    },
+    {
+      "epoch": 6.169064748201439,
+      "step": 6860,
+      "torque_loss": 0.1041853204369545
+    },
+    {
+      "epoch": 6.169064748201439,
+      "force_loss": 0.015111044049263,
+      "step": 6860
+    },
+    {
+      "epoch": 6.178057553956835,
+      "grad_norm": 0.927023708820343,
+      "learning_rate": 9.886749226084163e-05,
+      "loss": 0.0284,
+      "step": 6870
+    },
+    {
+      "action_loss": 0.014140564016997814,
+      "epoch": 6.178057553956835,
+      "step": 6870
+    },
+    {
+      "epoch": 6.178057553956835,
+      "step": 6870,
+      "torque_loss": 0.11716964840888977
+    },
+    {
+      "epoch": 6.178057553956835,
+      "force_loss": 0.01707938127219677,
+      "step": 6870
+    },
+    {
+      "epoch": 6.18705035971223,
+      "grad_norm": 1.0379085540771484,
+      "learning_rate": 9.886165277367714e-05,
+      "loss": 0.0374,
+      "step": 6880
+    },
+    {
+      "action_loss": 0.01943344622850418,
+      "epoch": 6.18705035971223,
+      "step": 6880
+    },
+    {
+      "epoch": 6.18705035971223,
+      "step": 6880,
+      "torque_loss": 0.1765347123146057
+    },
+    {
+      "epoch": 6.18705035971223,
+      "force_loss": 0.02587093412876129,
+      "step": 6880
+    },
+    {
+      "epoch": 6.196043165467626,
+      "grad_norm": 1.0784411430358887,
+      "learning_rate": 9.885579844363265e-05,
+      "loss": 0.0289,
+      "step": 6890
+    },
+    {
+      "action_loss": 0.014207050204277039,
+      "epoch": 6.196043165467626,
+      "step": 6890
+    },
+    {
+      "epoch": 6.196043165467626,
+      "step": 6890,
+      "torque_loss": 0.13490696251392365
+    },
+    {
+      "epoch": 6.196043165467626,
+      "force_loss": 0.015821872279047966,
+      "step": 6890
+    },
+    {
+      "epoch": 6.205035971223022,
+      "grad_norm": 1.1022746562957764,
+      "learning_rate": 9.884992927248656e-05,
+      "loss": 0.0299,
+      "step": 6900
+    },
+    {
+      "action_loss": 0.006934223230928183,
+      "epoch": 6.205035971223022,
+      "step": 6900
+    },
+    {
+      "epoch": 6.205035971223022,
+      "step": 6900,
+      "torque_loss": 0.1258738487958908
+    },
+    {
+      "epoch": 6.205035971223022,
+      "force_loss": 0.00647265138104558,
+      "step": 6900
+    },
+    {
+      "epoch": 6.2140287769784175,
+      "grad_norm": 0.9378265738487244,
+      "learning_rate": 9.884404526202178e-05,
+      "loss": 0.0356,
+      "step": 6910
+    },
+    {
+      "action_loss": 0.028104035183787346,
+      "epoch": 6.2140287769784175,
+      "step": 6910
+    },
+    {
+      "epoch": 6.2140287769784175,
+      "step": 6910,
+      "torque_loss": 0.15238653123378754
+    },
+    {
+      "epoch": 6.2140287769784175,
+      "force_loss": 0.033355068415403366,
+      "step": 6910
+    },
+    {
+      "epoch": 6.223021582733813,
+      "grad_norm": 1.419150948524475,
+      "learning_rate": 9.883814641402568e-05,
+      "loss": 0.0334,
+      "step": 6920
+    },
+    {
+      "action_loss": 0.024264290928840637,
+      "epoch": 6.223021582733813,
+      "step": 6920
+    },
+    {
+      "epoch": 6.223021582733813,
+      "step": 6920,
+      "torque_loss": 0.10440417379140854
+    },
+    {
+      "epoch": 6.223021582733813,
+      "force_loss": 0.019781582057476044,
+      "step": 6920
+    },
+    {
+      "epoch": 6.232014388489208,
+      "grad_norm": 1.142441987991333,
+      "learning_rate": 9.88322327302902e-05,
+      "loss": 0.0276,
+      "step": 6930
+    },
+    {
+      "action_loss": 0.014163498766720295,
+      "epoch": 6.232014388489208,
+      "step": 6930
+    },
+    {
+      "epoch": 6.232014388489208,
+      "step": 6930,
+      "torque_loss": 0.12220502644777298
+    },
+    {
+      "epoch": 6.232014388489208,
+      "force_loss": 0.0158188845962286,
+      "step": 6930
+    },
+    {
+      "epoch": 6.241007194244604,
+      "grad_norm": 0.591874361038208,
+      "learning_rate": 9.882630421261176e-05,
+      "loss": 0.0269,
+      "step": 6940
+    },
+    {
+      "action_loss": 0.008671222254633904,
+      "epoch": 6.241007194244604,
+      "step": 6940
+    },
+    {
+      "epoch": 6.241007194244604,
+      "step": 6940,
+      "torque_loss": 0.1492723971605301
+    },
+    {
+      "epoch": 6.241007194244604,
+      "force_loss": 0.007767186965793371,
+      "step": 6940
+    },
+    {
+      "epoch": 6.25,
+      "grad_norm": 0.8983883857727051,
+      "learning_rate": 9.88203608627913e-05,
+      "loss": 0.031,
+      "step": 6950
+    },
+    {
+      "action_loss": 0.026276089251041412,
+      "epoch": 6.25,
+      "step": 6950
+    },
+    {
+      "epoch": 6.25,
+      "step": 6950,
+      "torque_loss": 0.15878672897815704
+    },
+    {
+      "epoch": 6.25,
+      "force_loss": 0.022327935323119164,
+      "step": 6950
+    },
+    {
+      "epoch": 6.258992805755396,
+      "grad_norm": 1.1845142841339111,
+      "learning_rate": 9.881440268263422e-05,
+      "loss": 0.0329,
+      "step": 6960
+    },
+    {
+      "action_loss": 0.012182827107608318,
+      "epoch": 6.258992805755396,
+      "step": 6960
+    },
+    {
+      "epoch": 6.258992805755396,
+      "step": 6960,
+      "torque_loss": 0.11265624314546585
+    },
+    {
+      "epoch": 6.258992805755396,
+      "force_loss": 0.017472369596362114,
+      "step": 6960
+    },
+    {
+      "epoch": 6.267985611510792,
+      "grad_norm": 0.7261620759963989,
+      "learning_rate": 9.880842967395048e-05,
+      "loss": 0.0314,
+      "step": 6970
+    },
+    {
+      "action_loss": 0.016557425260543823,
+      "epoch": 6.267985611510792,
+      "step": 6970
+    },
+    {
+      "epoch": 6.267985611510792,
+      "step": 6970,
+      "torque_loss": 0.1438518911600113
+    },
+    {
+      "epoch": 6.267985611510792,
+      "force_loss": 0.010495592840015888,
+      "step": 6970
+    },
+    {
+      "epoch": 6.276978417266187,
+      "grad_norm": 1.1216435432434082,
+      "learning_rate": 9.880244183855452e-05,
+      "loss": 0.0338,
+      "step": 6980
+    },
+    {
+      "action_loss": 0.009683732874691486,
+      "epoch": 6.276978417266187,
+      "step": 6980
+    },
+    {
+      "epoch": 6.276978417266187,
+      "step": 6980,
+      "torque_loss": 0.14600324630737305
+    },
+    {
+      "epoch": 6.276978417266187,
+      "force_loss": 0.009540971368551254,
+      "step": 6980
+    },
+    {
+      "epoch": 6.2859712230215825,
+      "grad_norm": 0.6765707731246948,
+      "learning_rate": 9.879643917826527e-05,
+      "loss": 0.0425,
+      "step": 6990
+    },
+    {
+      "action_loss": 0.023353347554802895,
+      "epoch": 6.2859712230215825,
+      "step": 6990
+    },
+    {
+      "epoch": 6.2859712230215825,
+      "step": 6990,
+      "torque_loss": 0.14869339764118195
+    },
+    {
+      "epoch": 6.2859712230215825,
+      "force_loss": 0.01961287297308445,
+      "step": 6990
+    },
+    {
+      "epoch": 6.294964028776978,
+      "grad_norm": 1.4713687896728516,
+      "learning_rate": 9.87904216949062e-05,
+      "loss": 0.032,
+      "step": 7000
+    },
+    {
+      "action_loss": 0.011836816556751728,
+      "epoch": 6.294964028776978,
+      "step": 7000
+    },
+    {
+      "epoch": 6.294964028776978,
+      "step": 7000,
+      "torque_loss": 0.14368630945682526
+    },
+    {
+      "epoch": 6.294964028776978,
+      "force_loss": 0.008611331693828106,
+      "step": 7000
+    },
+    {
+      "epoch": 6.303956834532374,
+      "grad_norm": 1.2804861068725586,
+      "learning_rate": 9.878438939030526e-05,
+      "loss": 0.0347,
+      "step": 7010
+    },
+    {
+      "action_loss": 0.015133117325603962,
+      "epoch": 6.303956834532374,
+      "step": 7010
+    },
+    {
+      "epoch": 6.303956834532374,
+      "step": 7010,
+      "torque_loss": 0.27548733353614807
+    },
+    {
+      "epoch": 6.303956834532374,
+      "force_loss": 0.01170522253960371,
+      "step": 7010
+    },
+    {
+      "epoch": 6.31294964028777,
+      "grad_norm": 0.39458712935447693,
+      "learning_rate": 9.877834226629489e-05,
+      "loss": 0.0314,
+      "step": 7020
+    },
+    {
+      "action_loss": 0.006892053876072168,
+      "epoch": 6.31294964028777,
+      "step": 7020
+    },
+    {
+      "epoch": 6.31294964028777,
+      "step": 7020,
+      "torque_loss": 0.14708781242370605
+    },
+    {
+      "epoch": 6.31294964028777,
+      "force_loss": 0.0062276930548250675,
+      "step": 7020
+    },
+    {
+      "epoch": 6.321942446043165,
+      "grad_norm": 1.543850064277649,
+      "learning_rate": 9.877228032471206e-05,
+      "loss": 0.0303,
+      "step": 7030
+    },
+    {
+      "action_loss": 0.009968861006200314,
+      "epoch": 6.321942446043165,
+      "step": 7030
+    },
+    {
+      "epoch": 6.321942446043165,
+      "step": 7030,
+      "torque_loss": 0.1613566130399704
+    },
+    {
+      "epoch": 6.321942446043165,
+      "force_loss": 0.009990171529352665,
+      "step": 7030
+    },
+    {
+      "epoch": 6.330935251798561,
+      "grad_norm": 0.5653095841407776,
+      "learning_rate": 9.876620356739823e-05,
+      "loss": 0.0294,
+      "step": 7040
+    },
+    {
+      "action_loss": 0.013471047393977642,
+      "epoch": 6.330935251798561,
+      "step": 7040
+    },
+    {
+      "epoch": 6.330935251798561,
+      "step": 7040,
+      "torque_loss": 0.15058602392673492
+    },
+    {
+      "epoch": 6.330935251798561,
+      "force_loss": 0.013712618499994278,
+      "step": 7040
+    },
+    {
+      "epoch": 6.339928057553957,
+      "grad_norm": 0.4438684284687042,
+      "learning_rate": 9.876011199619935e-05,
+      "loss": 0.0289,
+      "step": 7050
+    },
+    {
+      "action_loss": 0.0074244216084480286,
+      "epoch": 6.339928057553957,
+      "step": 7050
+    },
+    {
+      "epoch": 6.339928057553957,
+      "step": 7050,
+      "torque_loss": 0.14606742560863495
+    },
+    {
+      "epoch": 6.339928057553957,
+      "force_loss": 0.008681533858180046,
+      "step": 7050
+    },
+    {
+      "epoch": 6.348920863309353,
+      "grad_norm": 0.8378824591636658,
+      "learning_rate": 9.875400561296589e-05,
+      "loss": 0.032,
+      "step": 7060
+    },
+    {
+      "action_loss": 0.021747564896941185,
+      "epoch": 6.348920863309353,
+      "step": 7060
+    },
+    {
+      "epoch": 6.348920863309353,
+      "step": 7060,
+      "torque_loss": 0.11303558200597763
+    },
+    {
+      "epoch": 6.348920863309353,
+      "force_loss": 0.010669874958693981,
+      "step": 7060
+    },
+    {
+      "epoch": 6.357913669064748,
+      "grad_norm": 1.1590083837509155,
+      "learning_rate": 9.874788441955278e-05,
+      "loss": 0.0364,
+      "step": 7070
+    },
+    {
+      "action_loss": 0.008278983645141125,
+      "epoch": 6.357913669064748,
+      "step": 7070
+    },
+    {
+      "epoch": 6.357913669064748,
+      "step": 7070,
+      "torque_loss": 0.09738612174987793
+    },
+    {
+      "epoch": 6.357913669064748,
+      "force_loss": 0.009098508395254612,
+      "step": 7070
+    },
+    {
+      "epoch": 6.366906474820144,
+      "grad_norm": 1.0467586517333984,
+      "learning_rate": 9.874174841781951e-05,
+      "loss": 0.038,
+      "step": 7080
+    },
+    {
+      "action_loss": 0.013057406060397625,
+      "epoch": 6.366906474820144,
+      "step": 7080
+    },
+    {
+      "epoch": 6.366906474820144,
+      "step": 7080,
+      "torque_loss": 0.1664576530456543
+    },
+    {
+      "epoch": 6.366906474820144,
+      "force_loss": 0.017414381727576256,
+      "step": 7080
+    },
+    {
+      "epoch": 6.375899280575539,
+      "grad_norm": 0.56081622838974,
+      "learning_rate": 9.873559760963003e-05,
+      "loss": 0.0313,
+      "step": 7090
+    },
+    {
+      "action_loss": 0.020285315811634064,
+      "epoch": 6.375899280575539,
+      "step": 7090
+    },
+    {
+      "epoch": 6.375899280575539,
+      "step": 7090,
+      "torque_loss": 0.1311166137456894
+    },
+    {
+      "epoch": 6.375899280575539,
+      "force_loss": 0.025589032098650932,
+      "step": 7090
+    },
+    {
+      "epoch": 6.384892086330935,
+      "grad_norm": 1.2313724756240845,
+      "learning_rate": 9.872943199685278e-05,
+      "loss": 0.0348,
+      "step": 7100
+    },
+    {
+      "action_loss": 0.01696467585861683,
+      "epoch": 6.384892086330935,
+      "step": 7100
+    },
+    {
+      "epoch": 6.384892086330935,
+      "step": 7100,
+      "torque_loss": 0.1781919002532959
+    },
+    {
+      "epoch": 6.384892086330935,
+      "force_loss": 0.011121484450995922,
+      "step": 7100
+    },
+    {
+      "epoch": 6.393884892086331,
+      "grad_norm": 0.9991695284843445,
+      "learning_rate": 9.872325158136071e-05,
+      "loss": 0.0334,
+      "step": 7110
+    },
+    {
+      "action_loss": 0.021009324118494987,
+      "epoch": 6.393884892086331,
+      "step": 7110
+    },
+    {
+      "epoch": 6.393884892086331,
+      "step": 7110,
+      "torque_loss": 0.17891235649585724
+    },
+    {
+      "epoch": 6.393884892086331,
+      "force_loss": 0.02270572818815708,
+      "step": 7110
+    },
+    {
+      "epoch": 6.402877697841727,
+      "grad_norm": 0.693954348564148,
+      "learning_rate": 9.871705636503128e-05,
+      "loss": 0.0334,
+      "step": 7120
+    },
+    {
+      "action_loss": 0.020892560482025146,
+      "epoch": 6.402877697841727,
+      "step": 7120
+    },
+    {
+      "epoch": 6.402877697841727,
+      "step": 7120,
+      "torque_loss": 0.13139615952968597
+    },
+    {
+      "epoch": 6.402877697841727,
+      "force_loss": 0.02533758245408535,
+      "step": 7120
+    },
+    {
+      "epoch": 6.411870503597123,
+      "grad_norm": 0.8198838233947754,
+      "learning_rate": 9.871084634974641e-05,
+      "loss": 0.0312,
+      "step": 7130
+    },
+    {
+      "action_loss": 0.030722124502062798,
+      "epoch": 6.411870503597123,
+      "step": 7130
+    },
+    {
+      "epoch": 6.411870503597123,
+      "step": 7130,
+      "torque_loss": 0.19522593915462494
+    },
+    {
+      "epoch": 6.411870503597123,
+      "force_loss": 0.019080134108662605,
+      "step": 7130
+    },
+    {
+      "epoch": 6.420863309352518,
+      "grad_norm": 0.9934543371200562,
+      "learning_rate": 9.870462153739257e-05,
+      "loss": 0.0381,
+      "step": 7140
+    },
+    {
+      "action_loss": 0.012233416549861431,
+      "epoch": 6.420863309352518,
+      "step": 7140
+    },
+    {
+      "epoch": 6.420863309352518,
+      "step": 7140,
+      "torque_loss": 0.13385216891765594
+    },
+    {
+      "epoch": 6.420863309352518,
+      "force_loss": 0.012103955261409283,
+      "step": 7140
+    },
+    {
+      "epoch": 6.429856115107913,
+      "grad_norm": 1.9219950437545776,
+      "learning_rate": 9.869838192986067e-05,
+      "loss": 0.0317,
+      "step": 7150
+    },
+    {
+      "action_loss": 0.010699831880629063,
+      "epoch": 6.429856115107913,
+      "step": 7150
+    },
+    {
+      "epoch": 6.429856115107913,
+      "step": 7150,
+      "torque_loss": 0.1361990123987198
+    },
+    {
+      "epoch": 6.429856115107913,
+      "force_loss": 0.011185181327164173,
+      "step": 7150
+    },
+    {
+      "epoch": 6.438848920863309,
+      "grad_norm": 0.7379173636436462,
+      "learning_rate": 9.869212752904616e-05,
+      "loss": 0.0404,
+      "step": 7160
+    },
+    {
+      "action_loss": 0.012237799353897572,
+      "epoch": 6.438848920863309,
+      "step": 7160
+    },
+    {
+      "epoch": 6.438848920863309,
+      "step": 7160,
+      "torque_loss": 0.16572776436805725
+    },
+    {
+      "epoch": 6.438848920863309,
+      "force_loss": 0.008981684222817421,
+      "step": 7160
+    },
+    {
+      "epoch": 6.447841726618705,
+      "grad_norm": 0.5905617475509644,
+      "learning_rate": 9.868585833684894e-05,
+      "loss": 0.0298,
+      "step": 7170
+    },
+    {
+      "action_loss": 0.01441280823200941,
+      "epoch": 6.447841726618705,
+      "step": 7170
+    },
+    {
+      "epoch": 6.447841726618705,
+      "step": 7170,
+      "torque_loss": 0.13176214694976807
+    },
+    {
+      "epoch": 6.447841726618705,
+      "force_loss": 0.014526325277984142,
+      "step": 7170
+    },
+    {
+      "epoch": 6.456834532374101,
+      "grad_norm": 0.7369416952133179,
+      "learning_rate": 9.867957435517342e-05,
+      "loss": 0.0331,
+      "step": 7180
+    },
+    {
+      "action_loss": 0.021134106442332268,
+      "epoch": 6.456834532374101,
+      "step": 7180
+    },
+    {
+      "epoch": 6.456834532374101,
+      "step": 7180,
+      "torque_loss": 0.18894974887371063
+    },
+    {
+      "epoch": 6.456834532374101,
+      "force_loss": 0.016715548932552338,
+      "step": 7180
+    },
+    {
+      "epoch": 6.465827338129497,
+      "grad_norm": 0.9571877717971802,
+      "learning_rate": 9.867327558592854e-05,
+      "loss": 0.0317,
+      "step": 7190
+    },
+    {
+      "action_loss": 0.012671258300542831,
+      "epoch": 6.465827338129497,
+      "step": 7190
+    },
+    {
+      "epoch": 6.465827338129497,
+      "step": 7190,
+      "torque_loss": 0.17883343994617462
+    },
+    {
+      "epoch": 6.465827338129497,
+      "force_loss": 0.00944325141608715,
+      "step": 7190
+    },
+    {
+      "epoch": 6.474820143884892,
+      "grad_norm": 1.0005165338516235,
+      "learning_rate": 9.866696203102766e-05,
+      "loss": 0.0391,
+      "step": 7200
+    },
+    {
+      "action_loss": 0.013150811195373535,
+      "epoch": 6.474820143884892,
+      "step": 7200
+    },
+    {
+      "epoch": 6.474820143884892,
+      "step": 7200,
+      "torque_loss": 0.11808394640684128
+    },
+    {
+      "epoch": 6.474820143884892,
+      "force_loss": 0.024321196600794792,
+      "step": 7200
+    },
+    {
+      "epoch": 6.483812949640288,
+      "grad_norm": 0.5357975363731384,
+      "learning_rate": 9.86606336923887e-05,
+      "loss": 0.0298,
+      "step": 7210
+    },
+    {
+      "action_loss": 0.00911055225878954,
+      "epoch": 6.483812949640288,
+      "step": 7210
+    },
+    {
+      "epoch": 6.483812949640288,
+      "step": 7210,
+      "torque_loss": 0.12453941255807877
+    },
+    {
+      "epoch": 6.483812949640288,
+      "force_loss": 0.006998790893703699,
+      "step": 7210
+    },
+    {
+      "epoch": 6.4928057553956835,
+      "grad_norm": 1.3047868013381958,
+      "learning_rate": 9.865429057193403e-05,
+      "loss": 0.0322,
+      "step": 7220
+    },
+    {
+      "action_loss": 0.021523533388972282,
+      "epoch": 6.4928057553956835,
+      "step": 7220
+    },
+    {
+      "epoch": 6.4928057553956835,
+      "step": 7220,
+      "torque_loss": 0.1434011161327362
+    },
+    {
+      "epoch": 6.4928057553956835,
+      "force_loss": 0.01496762316673994,
+      "step": 7220
+    },
+    {
+      "epoch": 6.501798561151079,
+      "grad_norm": 0.6869087815284729,
+      "learning_rate": 9.864793267159053e-05,
+      "loss": 0.0298,
+      "step": 7230
+    },
+    {
+      "action_loss": 0.012286696583032608,
+      "epoch": 6.501798561151079,
+      "step": 7230
+    },
+    {
+      "epoch": 6.501798561151079,
+      "step": 7230,
+      "torque_loss": 0.1482219696044922
+    },
+    {
+      "epoch": 6.501798561151079,
+      "force_loss": 0.009055552072823048,
+      "step": 7230
+    },
+    {
+      "epoch": 6.510791366906475,
+      "grad_norm": 0.9738610982894897,
+      "learning_rate": 9.864155999328957e-05,
+      "loss": 0.0307,
+      "step": 7240
+    },
+    {
+      "action_loss": 0.014220502227544785,
+      "epoch": 6.510791366906475,
+      "step": 7240
+    },
+    {
+      "epoch": 6.510791366906475,
+      "step": 7240,
+      "torque_loss": 0.12077126652002335
+    },
+    {
+      "epoch": 6.510791366906475,
+      "force_loss": 0.013996350578963757,
+      "step": 7240
+    },
+    {
+      "epoch": 6.51978417266187,
+      "grad_norm": 1.5982275009155273,
+      "learning_rate": 9.8635172538967e-05,
+      "loss": 0.036,
+      "step": 7250
+    },
+    {
+      "action_loss": 0.01579565368592739,
+      "epoch": 6.51978417266187,
+      "step": 7250
+    },
+    {
+      "epoch": 6.51978417266187,
+      "step": 7250,
+      "torque_loss": 0.12311286479234695
+    },
+    {
+      "epoch": 6.51978417266187,
+      "force_loss": 0.025801658630371094,
+      "step": 7250
+    },
+    {
+      "epoch": 6.528776978417266,
+      "grad_norm": 0.9958117008209229,
+      "learning_rate": 9.862877031056312e-05,
+      "loss": 0.0285,
+      "step": 7260
+    },
+    {
+      "action_loss": 0.009172475896775723,
+      "epoch": 6.528776978417266,
+      "step": 7260
+    },
+    {
+      "epoch": 6.528776978417266,
+      "step": 7260,
+      "torque_loss": 0.11217430979013443
+    },
+    {
+      "epoch": 6.528776978417266,
+      "force_loss": 0.007703617215156555,
+      "step": 7260
+    },
+    {
+      "epoch": 6.537769784172662,
+      "grad_norm": 0.6896548271179199,
+      "learning_rate": 9.862235331002279e-05,
+      "loss": 0.038,
+      "step": 7270
+    },
+    {
+      "action_loss": 0.004068915732204914,
+      "epoch": 6.537769784172662,
+      "step": 7270
+    },
+    {
+      "epoch": 6.537769784172662,
+      "step": 7270,
+      "torque_loss": 0.13171564042568207
+    },
+    {
+      "epoch": 6.537769784172662,
+      "force_loss": 0.008077028207480907,
+      "step": 7270
+    },
+    {
+      "epoch": 6.546762589928058,
+      "grad_norm": 0.6242184042930603,
+      "learning_rate": 9.861592153929533e-05,
+      "loss": 0.0284,
+      "step": 7280
+    },
+    {
+      "action_loss": 0.008289997465908527,
+      "epoch": 6.546762589928058,
+      "step": 7280
+    },
+    {
+      "epoch": 6.546762589928058,
+      "step": 7280,
+      "torque_loss": 0.11463453620672226
+    },
+    {
+      "epoch": 6.546762589928058,
+      "force_loss": 0.006958022713661194,
+      "step": 7280
+    },
+    {
+      "epoch": 6.555755395683454,
+      "grad_norm": 1.4783086776733398,
+      "learning_rate": 9.860947500033455e-05,
+      "loss": 0.0281,
+      "step": 7290
+    },
+    {
+      "action_loss": 0.05710189417004585,
+      "epoch": 6.555755395683454,
+      "step": 7290
+    },
+    {
+      "epoch": 6.555755395683454,
+      "step": 7290,
+      "torque_loss": 0.19084270298480988
+    },
+    {
+      "epoch": 6.555755395683454,
+      "force_loss": 0.05748428776860237,
+      "step": 7290
+    },
+    {
+      "epoch": 6.564748201438849,
+      "grad_norm": 0.4605726897716522,
+      "learning_rate": 9.86030136950987e-05,
+      "loss": 0.0379,
+      "step": 7300
+    },
+    {
+      "action_loss": 0.008308137767016888,
+      "epoch": 6.564748201438849,
+      "step": 7300
+    },
+    {
+      "epoch": 6.564748201438849,
+      "step": 7300,
+      "torque_loss": 0.09644180536270142
+    },
+    {
+      "epoch": 6.564748201438849,
+      "force_loss": 0.007345800753682852,
+      "step": 7300
+    },
+    {
+      "epoch": 6.573741007194244,
+      "grad_norm": 0.7289091944694519,
+      "learning_rate": 9.85965376255506e-05,
+      "loss": 0.0291,
+      "step": 7310
+    },
+    {
+      "action_loss": 0.010317725129425526,
+      "epoch": 6.573741007194244,
+      "step": 7310
+    },
+    {
+      "epoch": 6.573741007194244,
+      "step": 7310,
+      "torque_loss": 0.09413930028676987
+    },
+    {
+      "epoch": 6.573741007194244,
+      "force_loss": 0.009280539117753506,
+      "step": 7310
+    },
+    {
+      "epoch": 6.58273381294964,
+      "grad_norm": 1.0729559659957886,
+      "learning_rate": 9.859004679365747e-05,
+      "loss": 0.0273,
+      "step": 7320
+    },
+    {
+      "action_loss": 0.009764806367456913,
+      "epoch": 6.58273381294964,
+      "step": 7320
+    },
+    {
+      "epoch": 6.58273381294964,
+      "step": 7320,
+      "torque_loss": 0.10927480459213257
+    },
+    {
+      "epoch": 6.58273381294964,
+      "force_loss": 0.009876654483377934,
+      "step": 7320
+    },
+    {
+      "epoch": 6.591726618705036,
+      "grad_norm": 0.67609041929245,
+      "learning_rate": 9.858354120139108e-05,
+      "loss": 0.0291,
+      "step": 7330
+    },
+    {
+      "action_loss": 0.026859885081648827,
+      "epoch": 6.591726618705036,
+      "step": 7330
+    },
+    {
+      "epoch": 6.591726618705036,
+      "step": 7330,
+      "torque_loss": 0.22407126426696777
+    },
+    {
+      "epoch": 6.591726618705036,
+      "force_loss": 0.0318852998316288,
+      "step": 7330
+    },
+    {
+      "epoch": 6.600719424460432,
+      "grad_norm": 0.969810426235199,
+      "learning_rate": 9.857702085072764e-05,
+      "loss": 0.0318,
+      "step": 7340
+    },
+    {
+      "action_loss": 0.011868906207382679,
+      "epoch": 6.600719424460432,
+      "step": 7340
+    },
+    {
+      "epoch": 6.600719424460432,
+      "step": 7340,
+      "torque_loss": 0.11096364259719849
+    },
+    {
+      "epoch": 6.600719424460432,
+      "force_loss": 0.008112683892250061,
+      "step": 7340
+    },
+    {
+      "epoch": 6.609712230215827,
+      "grad_norm": 0.7616317272186279,
+      "learning_rate": 9.857048574364787e-05,
+      "loss": 0.034,
+      "step": 7350
+    },
+    {
+      "action_loss": 0.012131169438362122,
+      "epoch": 6.609712230215827,
+      "step": 7350
+    },
+    {
+      "epoch": 6.609712230215827,
+      "step": 7350,
+      "torque_loss": 0.14511050283908844
+    },
+    {
+      "epoch": 6.609712230215827,
+      "force_loss": 0.010916545987129211,
+      "step": 7350
+    },
+    {
+      "epoch": 6.618705035971223,
+      "grad_norm": 0.965458333492279,
+      "learning_rate": 9.856393588213698e-05,
+      "loss": 0.0345,
+      "step": 7360
+    },
+    {
+      "action_loss": 0.01754717156291008,
+      "epoch": 6.618705035971223,
+      "step": 7360
+    },
+    {
+      "epoch": 6.618705035971223,
+      "step": 7360,
+      "torque_loss": 0.13961432874202728
+    },
+    {
+      "epoch": 6.618705035971223,
+      "force_loss": 0.025284044444561005,
+      "step": 7360
+    },
+    {
+      "epoch": 6.627697841726619,
+      "grad_norm": 0.7501308917999268,
+      "learning_rate": 9.855737126818458e-05,
+      "loss": 0.0343,
+      "step": 7370
+    },
+    {
+      "action_loss": 0.008759301155805588,
+      "epoch": 6.627697841726619,
+      "step": 7370
+    },
+    {
+      "epoch": 6.627697841726619,
+      "step": 7370,
+      "torque_loss": 0.11119731515645981
+    },
+    {
+      "epoch": 6.627697841726619,
+      "force_loss": 0.008812778629362583,
+      "step": 7370
+    },
+    {
+      "epoch": 6.636690647482014,
+      "grad_norm": 0.5349105596542358,
+      "learning_rate": 9.855079190378491e-05,
+      "loss": 0.0356,
+      "step": 7380
+    },
+    {
+      "action_loss": 0.01553251687437296,
+      "epoch": 6.636690647482014,
+      "step": 7380
+    },
+    {
+      "epoch": 6.636690647482014,
+      "step": 7380,
+      "torque_loss": 0.18644464015960693
+    },
+    {
+      "epoch": 6.636690647482014,
+      "force_loss": 0.011027391068637371,
+      "step": 7380
+    },
+    {
+      "epoch": 6.64568345323741,
+      "grad_norm": 0.7108652591705322,
+      "learning_rate": 9.854419779093655e-05,
+      "loss": 0.0304,
+      "step": 7390
+    },
+    {
+      "action_loss": 0.015734700486063957,
+      "epoch": 6.64568345323741,
+      "step": 7390
+    },
+    {
+      "epoch": 6.64568345323741,
+      "step": 7390,
+      "torque_loss": 0.1646876484155655
+    },
+    {
+      "epoch": 6.64568345323741,
+      "force_loss": 0.009140515699982643,
+      "step": 7390
+    },
+    {
+      "epoch": 6.654676258992806,
+      "grad_norm": 0.7373231649398804,
+      "learning_rate": 9.853758893164264e-05,
+      "loss": 0.0306,
+      "step": 7400
+    },
+    {
+      "action_loss": 0.01355788391083479,
+      "epoch": 6.654676258992806,
+      "step": 7400
+    },
+    {
+      "epoch": 6.654676258992806,
+      "step": 7400,
+      "torque_loss": 0.136037677526474
+    },
+    {
+      "epoch": 6.654676258992806,
+      "force_loss": 0.008646302856504917,
+      "step": 7400
+    },
+    {
+      "epoch": 6.663669064748201,
+      "grad_norm": 0.5388456583023071,
+      "learning_rate": 9.853096532791078e-05,
+      "loss": 0.0296,
+      "step": 7410
+    },
+    {
+      "action_loss": 0.017288384959101677,
+      "epoch": 6.663669064748201,
+      "step": 7410
+    },
+    {
+      "epoch": 6.663669064748201,
+      "step": 7410,
+      "torque_loss": 0.13616986572742462
+    },
+    {
+      "epoch": 6.663669064748201,
+      "force_loss": 0.012311995029449463,
+      "step": 7410
+    },
+    {
+      "epoch": 6.672661870503597,
+      "grad_norm": 0.5630616545677185,
+      "learning_rate": 9.852432698175304e-05,
+      "loss": 0.0414,
+      "step": 7420
+    },
+    {
+      "action_loss": 0.00943202804774046,
+      "epoch": 6.672661870503597,
+      "step": 7420
+    },
+    {
+      "epoch": 6.672661870503597,
+      "step": 7420,
+      "torque_loss": 0.11193060874938965
+    },
+    {
+      "epoch": 6.672661870503597,
+      "force_loss": 0.013097594492137432,
+      "step": 7420
+    },
+    {
+      "epoch": 6.681654676258993,
+      "grad_norm": 1.0656129121780396,
+      "learning_rate": 9.851767389518597e-05,
+      "loss": 0.0397,
+      "step": 7430
+    },
+    {
+      "action_loss": 0.014476067386567593,
+      "epoch": 6.681654676258993,
+      "step": 7430
+    },
+    {
+      "epoch": 6.681654676258993,
+      "step": 7430,
+      "torque_loss": 0.15549862384796143
+    },
+    {
+      "epoch": 6.681654676258993,
+      "force_loss": 0.009197737090289593,
+      "step": 7430
+    },
+    {
+      "epoch": 6.690647482014389,
+      "grad_norm": 1.3346757888793945,
+      "learning_rate": 9.85110060702306e-05,
+      "loss": 0.0308,
+      "step": 7440
+    },
+    {
+      "action_loss": 0.012891768477857113,
+      "epoch": 6.690647482014389,
+      "step": 7440
+    },
+    {
+      "epoch": 6.690647482014389,
+      "step": 7440,
+      "torque_loss": 0.15279309451580048
+    },
+    {
+      "epoch": 6.690647482014389,
+      "force_loss": 0.0094189727678895,
+      "step": 7440
+    },
+    {
+      "epoch": 6.6996402877697845,
+      "grad_norm": 0.5730975866317749,
+      "learning_rate": 9.850432350891245e-05,
+      "loss": 0.0311,
+      "step": 7450
+    },
+    {
+      "action_loss": 0.017355715855956078,
+      "epoch": 6.6996402877697845,
+      "step": 7450
+    },
+    {
+      "epoch": 6.6996402877697845,
+      "step": 7450,
+      "torque_loss": 0.14554846286773682
+    },
+    {
+      "epoch": 6.6996402877697845,
+      "force_loss": 0.011299587786197662,
+      "step": 7450
+    },
+    {
+      "epoch": 6.7086330935251794,
+      "grad_norm": 1.064045786857605,
+      "learning_rate": 9.84976262132615e-05,
+      "loss": 0.0308,
+      "step": 7460
+    },
+    {
+      "action_loss": 0.0111715542152524,
+      "epoch": 6.7086330935251794,
+      "step": 7460
+    },
+    {
+      "epoch": 6.7086330935251794,
+      "step": 7460,
+      "torque_loss": 0.15227502584457397
+    },
+    {
+      "epoch": 6.7086330935251794,
+      "force_loss": 0.00802071113139391,
+      "step": 7460
+    },
+    {
+      "epoch": 6.717625899280575,
+      "grad_norm": 1.085392713546753,
+      "learning_rate": 9.849091418531222e-05,
+      "loss": 0.0344,
+      "step": 7470
+    },
+    {
+      "action_loss": 0.014177029021084309,
+      "epoch": 6.717625899280575,
+      "step": 7470
+    },
+    {
+      "epoch": 6.717625899280575,
+      "step": 7470,
+      "torque_loss": 0.1624159812927246
+    },
+    {
+      "epoch": 6.717625899280575,
+      "force_loss": 0.012642182409763336,
+      "step": 7470
+    },
+    {
+      "epoch": 6.726618705035971,
+      "grad_norm": 0.8970483541488647,
+      "learning_rate": 9.848418742710353e-05,
+      "loss": 0.0391,
+      "step": 7480
+    },
+    {
+      "action_loss": 0.007151718717068434,
+      "epoch": 6.726618705035971,
+      "step": 7480
+    },
+    {
+      "epoch": 6.726618705035971,
+      "step": 7480,
+      "torque_loss": 0.1363411396741867
+    },
+    {
+      "epoch": 6.726618705035971,
+      "force_loss": 0.004806325305253267,
+      "step": 7480
+    },
+    {
+      "epoch": 6.735611510791367,
+      "grad_norm": 1.2199770212173462,
+      "learning_rate": 9.847744594067885e-05,
+      "loss": 0.0341,
+      "step": 7490
+    },
+    {
+      "action_loss": 0.010929140262305737,
+      "epoch": 6.735611510791367,
+      "step": 7490
+    },
+    {
+      "epoch": 6.735611510791367,
+      "step": 7490,
+      "torque_loss": 0.11901833862066269
+    },
+    {
+      "epoch": 6.735611510791367,
+      "force_loss": 0.00839619804173708,
+      "step": 7490
+    },
+    {
+      "epoch": 6.744604316546763,
+      "grad_norm": 0.8313911557197571,
+      "learning_rate": 9.847068972808607e-05,
+      "loss": 0.033,
+      "step": 7500
+    },
+    {
+      "action_loss": 0.006041519343852997,
+      "epoch": 6.744604316546763,
+      "step": 7500
+    },
+    {
+      "epoch": 6.744604316546763,
+      "step": 7500,
+      "torque_loss": 0.11097083240747452
+    },
+    {
+      "epoch": 6.744604316546763,
+      "force_loss": 0.006382436957210302,
+      "step": 7500
+    },
+    {
+      "epoch": 6.753597122302159,
+      "grad_norm": 0.8322669267654419,
+      "learning_rate": 9.846391879137756e-05,
+      "loss": 0.0287,
+      "step": 7510
+    },
+    {
+      "action_loss": 0.011735928244888783,
+      "epoch": 6.753597122302159,
+      "step": 7510
+    },
+    {
+      "epoch": 6.753597122302159,
+      "step": 7510,
+      "torque_loss": 0.13664768636226654
+    },
+    {
+      "epoch": 6.753597122302159,
+      "force_loss": 0.014030583202838898,
+      "step": 7510
+    },
+    {
+      "epoch": 6.762589928057554,
+      "grad_norm": 0.850685715675354,
+      "learning_rate": 9.845713313261012e-05,
+      "loss": 0.0372,
+      "step": 7520
+    },
+    {
+      "action_loss": 0.005964975338429213,
+      "epoch": 6.762589928057554,
+      "step": 7520
+    },
+    {
+      "epoch": 6.762589928057554,
+      "step": 7520,
+      "torque_loss": 0.09368737787008286
+    },
+    {
+      "epoch": 6.762589928057554,
+      "force_loss": 0.007000280078500509,
+      "step": 7520
+    },
+    {
+      "epoch": 6.7715827338129495,
+      "grad_norm": 0.7753496170043945,
+      "learning_rate": 9.845033275384505e-05,
+      "loss": 0.0394,
+      "step": 7530
+    },
+    {
+      "action_loss": 0.00790408719331026,
+      "epoch": 6.7715827338129495,
+      "step": 7530
+    },
+    {
+      "epoch": 6.7715827338129495,
+      "step": 7530,
+      "torque_loss": 0.10145465284585953
+    },
+    {
+      "epoch": 6.7715827338129495,
+      "force_loss": 0.012381188571453094,
+      "step": 7530
+    },
+    {
+      "epoch": 6.780575539568345,
+      "grad_norm": 0.6639843583106995,
+      "learning_rate": 9.844351765714818e-05,
+      "loss": 0.0322,
+      "step": 7540
+    },
+    {
+      "action_loss": 0.02301245927810669,
+      "epoch": 6.780575539568345,
+      "step": 7540
+    },
+    {
+      "epoch": 6.780575539568345,
+      "step": 7540,
+      "torque_loss": 0.19160182774066925
+    },
+    {
+      "epoch": 6.780575539568345,
+      "force_loss": 0.015690037980675697,
+      "step": 7540
+    },
+    {
+      "epoch": 6.789568345323741,
+      "grad_norm": 0.9610219597816467,
+      "learning_rate": 9.843668784458971e-05,
+      "loss": 0.0351,
+      "step": 7550
+    },
+    {
+      "action_loss": 0.01086804736405611,
+      "epoch": 6.789568345323741,
+      "step": 7550
+    },
+    {
+      "epoch": 6.789568345323741,
+      "step": 7550,
+      "torque_loss": 0.13474760949611664
+    },
+    {
+      "epoch": 6.789568345323741,
+      "force_loss": 0.012556198053061962,
+      "step": 7550
+    },
+    {
+      "epoch": 6.798561151079137,
+      "grad_norm": 0.4313907027244568,
+      "learning_rate": 9.842984331824437e-05,
+      "loss": 0.0287,
+      "step": 7560
+    },
+    {
+      "action_loss": 0.017279664054512978,
+      "epoch": 6.798561151079137,
+      "step": 7560
+    },
+    {
+      "epoch": 6.798561151079137,
+      "step": 7560,
+      "torque_loss": 0.17883305251598358
+    },
+    {
+      "epoch": 6.798561151079137,
+      "force_loss": 0.01957053877413273,
+      "step": 7560
+    },
+    {
+      "epoch": 6.807553956834532,
+      "grad_norm": 0.7270711064338684,
+      "learning_rate": 9.842298408019133e-05,
+      "loss": 0.0241,
+      "step": 7570
+    },
+    {
+      "action_loss": 0.00688102887943387,
+      "epoch": 6.807553956834532,
+      "step": 7570
+    },
+    {
+      "epoch": 6.807553956834532,
+      "step": 7570,
+      "torque_loss": 0.15468500554561615
+    },
+    {
+      "epoch": 6.807553956834532,
+      "force_loss": 0.00769207626581192,
+      "step": 7570
+    },
+    {
+      "epoch": 6.816546762589928,
+      "grad_norm": 0.6725876331329346,
+      "learning_rate": 9.841611013251429e-05,
+      "loss": 0.0327,
+      "step": 7580
+    },
+    {
+      "action_loss": 0.013837548904120922,
+      "epoch": 6.816546762589928,
+      "step": 7580
+    },
+    {
+      "epoch": 6.816546762589928,
+      "step": 7580,
+      "torque_loss": 0.15249620378017426
+    },
+    {
+      "epoch": 6.816546762589928,
+      "force_loss": 0.010722585953772068,
+      "step": 7580
+    },
+    {
+      "epoch": 6.825539568345324,
+      "grad_norm": 1.0520435571670532,
+      "learning_rate": 9.840922147730133e-05,
+      "loss": 0.034,
+      "step": 7590
+    },
+    {
+      "action_loss": 0.008364955894649029,
+      "epoch": 6.825539568345324,
+      "step": 7590
+    },
+    {
+      "epoch": 6.825539568345324,
+      "step": 7590,
+      "torque_loss": 0.16778381168842316
+    },
+    {
+      "epoch": 6.825539568345324,
+      "force_loss": 0.00917450338602066,
+      "step": 7590
+    },
+    {
+      "epoch": 6.83453237410072,
+      "grad_norm": 1.3950741291046143,
+      "learning_rate": 9.840231811664506e-05,
+      "loss": 0.0352,
+      "step": 7600
+    },
+    {
+      "action_loss": 0.02204935997724533,
+      "epoch": 6.83453237410072,
+      "step": 7600
+    },
+    {
+      "epoch": 6.83453237410072,
+      "step": 7600,
+      "torque_loss": 0.14785687625408173
+    },
+    {
+      "epoch": 6.83453237410072,
+      "force_loss": 0.02144394814968109,
+      "step": 7600
+    },
+    {
+      "epoch": 6.843525179856115,
+      "grad_norm": 1.2162723541259766,
+      "learning_rate": 9.839540005264252e-05,
+      "loss": 0.0319,
+      "step": 7610
+    },
+    {
+      "action_loss": 0.015045590698719025,
+      "epoch": 6.843525179856115,
+      "step": 7610
+    },
+    {
+      "epoch": 6.843525179856115,
+      "step": 7610,
+      "torque_loss": 0.13497065007686615
+    },
+    {
+      "epoch": 6.843525179856115,
+      "force_loss": 0.010974928736686707,
+      "step": 7610
+    },
+    {
+      "epoch": 6.852517985611511,
+      "grad_norm": 0.5862897038459778,
+      "learning_rate": 9.838846728739527e-05,
+      "loss": 0.0255,
+      "step": 7620
+    },
+    {
+      "action_loss": 0.015867173671722412,
+      "epoch": 6.852517985611511,
+      "step": 7620
+    },
+    {
+      "epoch": 6.852517985611511,
+      "step": 7620,
+      "torque_loss": 0.15003100037574768
+    },
+    {
+      "epoch": 6.852517985611511,
+      "force_loss": 0.01031761709600687,
+      "step": 7620
+    },
+    {
+      "epoch": 6.861510791366906,
+      "grad_norm": 1.037456750869751,
+      "learning_rate": 9.838151982300927e-05,
+      "loss": 0.0394,
+      "step": 7630
+    },
+    {
+      "action_loss": 0.012813209556043148,
+      "epoch": 6.861510791366906,
+      "step": 7630
+    },
+    {
+      "epoch": 6.861510791366906,
+      "step": 7630,
+      "torque_loss": 0.1465476006269455
+    },
+    {
+      "epoch": 6.861510791366906,
+      "force_loss": 0.012145946733653545,
+      "step": 7630
+    },
+    {
+      "epoch": 6.870503597122302,
+      "grad_norm": 0.7836095094680786,
+      "learning_rate": 9.8374557661595e-05,
+      "loss": 0.0297,
+      "step": 7640
+    },
+    {
+      "action_loss": 0.028261477127671242,
+      "epoch": 6.870503597122302,
+      "step": 7640
+    },
+    {
+      "epoch": 6.870503597122302,
+      "step": 7640,
+      "torque_loss": 0.1583840399980545
+    },
+    {
+      "epoch": 6.870503597122302,
+      "force_loss": 0.018456261605024338,
+      "step": 7640
+    },
+    {
+      "epoch": 6.879496402877698,
+      "grad_norm": 0.702251672744751,
+      "learning_rate": 9.836758080526735e-05,
+      "loss": 0.0341,
+      "step": 7650
+    },
+    {
+      "action_loss": 0.01756826601922512,
+      "epoch": 6.879496402877698,
+      "step": 7650
+    },
+    {
+      "epoch": 6.879496402877698,
+      "step": 7650,
+      "torque_loss": 0.16502438485622406
+    },
+    {
+      "epoch": 6.879496402877698,
+      "force_loss": 0.013490281067788601,
+      "step": 7650
+    },
+    {
+      "epoch": 6.888489208633094,
+      "grad_norm": 0.8463592529296875,
+      "learning_rate": 9.836058925614575e-05,
+      "loss": 0.0313,
+      "step": 7660
+    },
+    {
+      "action_loss": 0.0048490907065570354,
+      "epoch": 6.888489208633094,
+      "step": 7660
+    },
+    {
+      "epoch": 6.888489208633094,
+      "step": 7660,
+      "torque_loss": 0.16393551230430603
+    },
+    {
+      "epoch": 6.888489208633094,
+      "force_loss": 0.006976804230362177,
+      "step": 7660
+    },
+    {
+      "epoch": 6.897482014388489,
+      "grad_norm": 0.7803255319595337,
+      "learning_rate": 9.8353583016354e-05,
+      "loss": 0.0287,
+      "step": 7670
+    },
+    {
+      "action_loss": 0.012085658498108387,
+      "epoch": 6.897482014388489,
+      "step": 7670
+    },
+    {
+      "epoch": 6.897482014388489,
+      "step": 7670,
+      "torque_loss": 0.17637775838375092
+    },
+    {
+      "epoch": 6.897482014388489,
+      "force_loss": 0.008509885519742966,
+      "step": 7670
+    },
+    {
+      "epoch": 6.906474820143885,
+      "grad_norm": 1.180240511894226,
+      "learning_rate": 9.834656208802044e-05,
+      "loss": 0.0323,
+      "step": 7680
+    },
+    {
+      "action_loss": 0.018915629014372826,
+      "epoch": 6.906474820143885,
+      "step": 7680
+    },
+    {
+      "epoch": 6.906474820143885,
+      "step": 7680,
+      "torque_loss": 0.21235208213329315
+    },
+    {
+      "epoch": 6.906474820143885,
+      "force_loss": 0.01931210793554783,
+      "step": 7680
+    },
+    {
+      "epoch": 6.91546762589928,
+      "grad_norm": 0.3223646283149719,
+      "learning_rate": 9.833952647327784e-05,
+      "loss": 0.0301,
+      "step": 7690
+    },
+    {
+      "action_loss": 0.009821752086281776,
+      "epoch": 6.91546762589928,
+      "step": 7690
+    },
+    {
+      "epoch": 6.91546762589928,
+      "step": 7690,
+      "torque_loss": 0.1348685473203659
+    },
+    {
+      "epoch": 6.91546762589928,
+      "force_loss": 0.004952156450599432,
+      "step": 7690
+    },
+    {
+      "epoch": 6.924460431654676,
+      "grad_norm": 0.7646491527557373,
+      "learning_rate": 9.833247617426342e-05,
+      "loss": 0.0284,
+      "step": 7700
+    },
+    {
+      "action_loss": 0.013955810107290745,
+      "epoch": 6.924460431654676,
+      "step": 7700
+    },
+    {
+      "epoch": 6.924460431654676,
+      "step": 7700,
+      "torque_loss": 0.17015063762664795
+    },
+    {
+      "epoch": 6.924460431654676,
+      "force_loss": 0.009839924052357674,
+      "step": 7700
+    },
+    {
+      "epoch": 6.933453237410072,
+      "grad_norm": 0.6083731055259705,
+      "learning_rate": 9.832541119311889e-05,
+      "loss": 0.0328,
+      "step": 7710
+    },
+    {
+      "action_loss": 0.024154866114258766,
+      "epoch": 6.933453237410072,
+      "step": 7710
+    },
+    {
+      "epoch": 6.933453237410072,
+      "step": 7710,
+      "torque_loss": 0.16915397346019745
+    },
+    {
+      "epoch": 6.933453237410072,
+      "force_loss": 0.019542505964636803,
+      "step": 7710
+    },
+    {
+      "epoch": 6.942446043165468,
+      "grad_norm": 1.2142714262008667,
+      "learning_rate": 9.83183315319904e-05,
+      "loss": 0.0402,
+      "step": 7720
+    },
+    {
+      "action_loss": 0.028392454609274864,
+      "epoch": 6.942446043165468,
+      "step": 7720
+    },
+    {
+      "epoch": 6.942446043165468,
+      "step": 7720,
+      "torque_loss": 0.19171525537967682
+    },
+    {
+      "epoch": 6.942446043165468,
+      "force_loss": 0.016506226733326912,
+      "step": 7720
+    },
+    {
+      "epoch": 6.951438848920863,
+      "grad_norm": 0.8586347699165344,
+      "learning_rate": 9.831123719302855e-05,
+      "loss": 0.0309,
+      "step": 7730
+    },
+    {
+      "action_loss": 0.022412672638893127,
+      "epoch": 6.951438848920863,
+      "step": 7730
+    },
+    {
+      "epoch": 6.951438848920863,
+      "step": 7730,
+      "torque_loss": 0.10771119594573975
+    },
+    {
+      "epoch": 6.951438848920863,
+      "force_loss": 0.026249969378113747,
+      "step": 7730
+    },
+    {
+      "epoch": 6.960431654676259,
+      "grad_norm": 1.0157043933868408,
+      "learning_rate": 9.830412817838842e-05,
+      "loss": 0.0306,
+      "step": 7740
+    },
+    {
+      "action_loss": 0.020980997011065483,
+      "epoch": 6.960431654676259,
+      "step": 7740
+    },
+    {
+      "epoch": 6.960431654676259,
+      "step": 7740,
+      "torque_loss": 0.169539213180542
+    },
+    {
+      "epoch": 6.960431654676259,
+      "force_loss": 0.020628642290830612,
+      "step": 7740
+    },
+    {
+      "epoch": 6.969424460431655,
+      "grad_norm": 0.5320214629173279,
+      "learning_rate": 9.829700449022956e-05,
+      "loss": 0.0347,
+      "step": 7750
+    },
+    {
+      "action_loss": 0.012410610914230347,
+      "epoch": 6.969424460431655,
+      "step": 7750
+    },
+    {
+      "epoch": 6.969424460431655,
+      "step": 7750,
+      "torque_loss": 0.15810233354568481
+    },
+    {
+      "epoch": 6.969424460431655,
+      "force_loss": 0.012388341128826141,
+      "step": 7750
+    },
+    {
+      "epoch": 6.9784172661870505,
+      "grad_norm": 1.3180557489395142,
+      "learning_rate": 9.828986613071593e-05,
+      "loss": 0.0281,
+      "step": 7760
+    },
+    {
+      "action_loss": 0.007189827039837837,
+      "epoch": 6.9784172661870505,
+      "step": 7760
+    },
+    {
+      "epoch": 6.9784172661870505,
+      "step": 7760,
+      "torque_loss": 0.09706615656614304
+    },
+    {
+      "epoch": 6.9784172661870505,
+      "force_loss": 0.007400956004858017,
+      "step": 7760
+    },
+    {
+      "epoch": 6.987410071942446,
+      "grad_norm": 0.7778557538986206,
+      "learning_rate": 9.828271310201601e-05,
+      "loss": 0.0277,
+      "step": 7770
+    },
+    {
+      "action_loss": 0.013597776181995869,
+      "epoch": 6.987410071942446,
+      "step": 7770
+    },
+    {
+      "epoch": 6.987410071942446,
+      "step": 7770,
+      "torque_loss": 0.1417521834373474
+    },
+    {
+      "epoch": 6.987410071942446,
+      "force_loss": 0.019632231444120407,
+      "step": 7770
+    },
+    {
+      "epoch": 6.996402877697841,
+      "grad_norm": 0.7626013159751892,
+      "learning_rate": 9.827554540630268e-05,
+      "loss": 0.0305,
+      "step": 7780
+    },
+    {
+      "action_loss": 0.014794292859733105,
+      "epoch": 6.996402877697841,
+      "step": 7780
+    },
+    {
+      "epoch": 6.996402877697841,
+      "step": 7780,
+      "torque_loss": 0.20905987918376923
+    },
+    {
+      "epoch": 6.996402877697841,
+      "force_loss": 0.009513375349342823,
+      "step": 7780
+    },
+    {
+      "epoch": 7.005395683453237,
+      "grad_norm": 0.3939371407032013,
+      "learning_rate": 9.826836304575329e-05,
+      "loss": 0.0303,
+      "step": 7790
+    },
+    {
+      "action_loss": 0.011105640791356564,
+      "epoch": 7.005395683453237,
+      "step": 7790
+    },
+    {
+      "epoch": 7.005395683453237,
+      "step": 7790,
+      "torque_loss": 0.1300303339958191
+    },
+    {
+      "epoch": 7.005395683453237,
+      "force_loss": 0.011433079838752747,
+      "step": 7790
+    },
+    {
+      "epoch": 7.014388489208633,
+      "grad_norm": 0.980540931224823,
+      "learning_rate": 9.826116602254966e-05,
+      "loss": 0.0315,
+      "step": 7800
+    },
+    {
+      "action_loss": 0.018765976652503014,
+      "epoch": 7.014388489208633,
+      "step": 7800
+    },
+    {
+      "epoch": 7.014388489208633,
+      "step": 7800,
+      "torque_loss": 0.1850420981645584
+    },
+    {
+      "epoch": 7.014388489208633,
+      "force_loss": 0.01585307903587818,
+      "step": 7800
+    },
+    {
+      "epoch": 7.023381294964029,
+      "grad_norm": 0.5051628947257996,
+      "learning_rate": 9.825395433887805e-05,
+      "loss": 0.0332,
+      "step": 7810
+    },
+    {
+      "action_loss": 0.029291892424225807,
+      "epoch": 7.023381294964029,
+      "step": 7810
+    },
+    {
+      "epoch": 7.023381294964029,
+      "step": 7810,
+      "torque_loss": 0.15354830026626587
+    },
+    {
+      "epoch": 7.023381294964029,
+      "force_loss": 0.033693213015794754,
+      "step": 7810
+    },
+    {
+      "epoch": 7.032374100719425,
+      "grad_norm": 1.2920396327972412,
+      "learning_rate": 9.824672799692917e-05,
+      "loss": 0.0326,
+      "step": 7820
+    },
+    {
+      "action_loss": 0.016909083351492882,
+      "epoch": 7.032374100719425,
+      "step": 7820
+    },
+    {
+      "epoch": 7.032374100719425,
+      "step": 7820,
+      "torque_loss": 0.1852744072675705
+    },
+    {
+      "epoch": 7.032374100719425,
+      "force_loss": 0.014743109233677387,
+      "step": 7820
+    },
+    {
+      "epoch": 7.0413669064748206,
+      "grad_norm": 1.7694145441055298,
+      "learning_rate": 9.823948699889823e-05,
+      "loss": 0.0334,
+      "step": 7830
+    },
+    {
+      "action_loss": 0.009295588359236717,
+      "epoch": 7.0413669064748206,
+      "step": 7830
+    },
+    {
+      "epoch": 7.0413669064748206,
+      "step": 7830,
+      "torque_loss": 0.11250940710306168
+    },
+    {
+      "epoch": 7.0413669064748206,
+      "force_loss": 0.00965843815356493,
+      "step": 7830
+    },
+    {
+      "epoch": 7.0503597122302155,
+      "grad_norm": 1.013838768005371,
+      "learning_rate": 9.823223134698483e-05,
+      "loss": 0.0299,
+      "step": 7840
+    },
+    {
+      "action_loss": 0.021403102204203606,
+      "epoch": 7.0503597122302155,
+      "step": 7840
+    },
+    {
+      "epoch": 7.0503597122302155,
+      "step": 7840,
+      "torque_loss": 0.1695012003183365
+    },
+    {
+      "epoch": 7.0503597122302155,
+      "force_loss": 0.015083979815244675,
+      "step": 7840
+    },
+    {
+      "epoch": 7.059352517985611,
+      "grad_norm": 0.733481764793396,
+      "learning_rate": 9.822496104339303e-05,
+      "loss": 0.0319,
+      "step": 7850
+    },
+    {
+      "action_loss": 0.008760634809732437,
+      "epoch": 7.059352517985611,
+      "step": 7850
+    },
+    {
+      "epoch": 7.059352517985611,
+      "step": 7850,
+      "torque_loss": 0.15025722980499268
+    },
+    {
+      "epoch": 7.059352517985611,
+      "force_loss": 0.007303763180971146,
+      "step": 7850
+    },
+    {
+      "epoch": 7.068345323741007,
+      "grad_norm": 0.5384030342102051,
+      "learning_rate": 9.821767609033138e-05,
+      "loss": 0.0306,
+      "step": 7860
+    },
+    {
+      "action_loss": 0.014348868280649185,
+      "epoch": 7.068345323741007,
+      "step": 7860
+    },
+    {
+      "epoch": 7.068345323741007,
+      "step": 7860,
+      "torque_loss": 0.14780882000923157
+    },
+    {
+      "epoch": 7.068345323741007,
+      "force_loss": 0.024701185524463654,
+      "step": 7860
+    },
+    {
+      "epoch": 7.077338129496403,
+      "grad_norm": 0.6079313158988953,
+      "learning_rate": 9.821037649001284e-05,
+      "loss": 0.0353,
+      "step": 7870
+    },
+    {
+      "action_loss": 0.011603743769228458,
+      "epoch": 7.077338129496403,
+      "step": 7870
+    },
+    {
+      "epoch": 7.077338129496403,
+      "step": 7870,
+      "torque_loss": 0.14331673085689545
+    },
+    {
+      "epoch": 7.077338129496403,
+      "force_loss": 0.012257256545126438,
+      "step": 7870
+    },
+    {
+      "epoch": 7.086330935251799,
+      "grad_norm": 0.8938829302787781,
+      "learning_rate": 9.820306224465486e-05,
+      "loss": 0.0299,
+      "step": 7880
+    },
+    {
+      "action_loss": 0.009041390381753445,
+      "epoch": 7.086330935251799,
+      "step": 7880
+    },
+    {
+      "epoch": 7.086330935251799,
+      "step": 7880,
+      "torque_loss": 0.14269673824310303
+    },
+    {
+      "epoch": 7.086330935251799,
+      "force_loss": 0.007723438087850809,
+      "step": 7880
+    },
+    {
+      "epoch": 7.095323741007194,
+      "grad_norm": 0.6975097060203552,
+      "learning_rate": 9.819573335647928e-05,
+      "loss": 0.034,
+      "step": 7890
+    },
+    {
+      "action_loss": 0.007311273366212845,
+      "epoch": 7.095323741007194,
+      "step": 7890
+    },
+    {
+      "epoch": 7.095323741007194,
+      "step": 7890,
+      "torque_loss": 0.11497911065816879
+    },
+    {
+      "epoch": 7.095323741007194,
+      "force_loss": 0.005881762597709894,
+      "step": 7890
+    },
+    {
+      "epoch": 7.10431654676259,
+      "grad_norm": 0.9963082075119019,
+      "learning_rate": 9.818838982771246e-05,
+      "loss": 0.0361,
+      "step": 7900
+    },
+    {
+      "action_loss": 0.013384412042796612,
+      "epoch": 7.10431654676259,
+      "step": 7900
+    },
+    {
+      "epoch": 7.10431654676259,
+      "step": 7900,
+      "torque_loss": 0.1763625591993332
+    },
+    {
+      "epoch": 7.10431654676259,
+      "force_loss": 0.01059798151254654,
+      "step": 7900
+    },
+    {
+      "epoch": 7.113309352517986,
+      "grad_norm": 1.1353193521499634,
+      "learning_rate": 9.818103166058514e-05,
+      "loss": 0.0316,
+      "step": 7910
+    },
+    {
+      "action_loss": 0.011034096591174603,
+      "epoch": 7.113309352517986,
+      "step": 7910
+    },
+    {
+      "epoch": 7.113309352517986,
+      "step": 7910,
+      "torque_loss": 0.17189419269561768
+    },
+    {
+      "epoch": 7.113309352517986,
+      "force_loss": 0.011203192174434662,
+      "step": 7910
+    },
+    {
+      "epoch": 7.122302158273381,
+      "grad_norm": 1.559002161026001,
+      "learning_rate": 9.817365885733254e-05,
+      "loss": 0.0315,
+      "step": 7920
+    },
+    {
+      "action_loss": 0.014221259392797947,
+      "epoch": 7.122302158273381,
+      "step": 7920
+    },
+    {
+      "epoch": 7.122302158273381,
+      "step": 7920,
+      "torque_loss": 0.12348828464746475
+    },
+    {
+      "epoch": 7.122302158273381,
+      "force_loss": 0.010941113345324993,
+      "step": 7920
+    },
+    {
+      "epoch": 7.131294964028777,
+      "grad_norm": 0.6356121897697449,
+      "learning_rate": 9.816627142019434e-05,
+      "loss": 0.0299,
+      "step": 7930
+    },
+    {
+      "action_loss": 0.019868925213813782,
+      "epoch": 7.131294964028777,
+      "step": 7930
+    },
+    {
+      "epoch": 7.131294964028777,
+      "step": 7930,
+      "torque_loss": 0.15779805183410645
+    },
+    {
+      "epoch": 7.131294964028777,
+      "force_loss": 0.01079738512635231,
+      "step": 7930
+    },
+    {
+      "epoch": 7.140287769784172,
+      "grad_norm": 1.2879135608673096,
+      "learning_rate": 9.815886935141463e-05,
+      "loss": 0.0366,
+      "step": 7940
+    },
+    {
+      "action_loss": 0.015258598141372204,
+      "epoch": 7.140287769784172,
+      "step": 7940
+    },
+    {
+      "epoch": 7.140287769784172,
+      "step": 7940,
+      "torque_loss": 0.11968880891799927
+    },
+    {
+      "epoch": 7.140287769784172,
+      "force_loss": 0.011036291718482971,
+      "step": 7940
+    },
+    {
+      "epoch": 7.149280575539568,
+      "grad_norm": 0.8261431455612183,
+      "learning_rate": 9.8151452653242e-05,
+      "loss": 0.0288,
+      "step": 7950
+    },
+    {
+      "action_loss": 0.015346703119575977,
+      "epoch": 7.149280575539568,
+      "step": 7950
+    },
+    {
+      "epoch": 7.149280575539568,
+      "step": 7950,
+      "torque_loss": 0.14154738187789917
+    },
+    {
+      "epoch": 7.149280575539568,
+      "force_loss": 0.02692827582359314,
+      "step": 7950
+    },
+    {
+      "epoch": 7.158273381294964,
+      "grad_norm": 0.9473282098770142,
+      "learning_rate": 9.814402132792939e-05,
+      "loss": 0.0299,
+      "step": 7960
+    },
+    {
+      "action_loss": 0.026272349059581757,
+      "epoch": 7.158273381294964,
+      "step": 7960
+    },
+    {
+      "epoch": 7.158273381294964,
+      "step": 7960,
+      "torque_loss": 0.1546822041273117
+    },
+    {
+      "epoch": 7.158273381294964,
+      "force_loss": 0.024519843980669975,
+      "step": 7960
+    },
+    {
+      "epoch": 7.16726618705036,
+      "grad_norm": 0.6530436873435974,
+      "learning_rate": 9.813657537773428e-05,
+      "loss": 0.0319,
+      "step": 7970
+    },
+    {
+      "action_loss": 0.01660039834678173,
+      "epoch": 7.16726618705036,
+      "step": 7970
+    },
+    {
+      "epoch": 7.16726618705036,
+      "step": 7970,
+      "torque_loss": 0.11716824769973755
+    },
+    {
+      "epoch": 7.16726618705036,
+      "force_loss": 0.009484336711466312,
+      "step": 7970
+    },
+    {
+      "epoch": 7.176258992805756,
+      "grad_norm": 0.6228723526000977,
+      "learning_rate": 9.812911480491854e-05,
+      "loss": 0.0284,
+      "step": 7980
+    },
+    {
+      "action_loss": 0.015811538323760033,
+      "epoch": 7.176258992805756,
+      "step": 7980
+    },
+    {
+      "epoch": 7.176258992805756,
+      "step": 7980,
+      "torque_loss": 0.14678019285202026
+    },
+    {
+      "epoch": 7.176258992805756,
+      "force_loss": 0.019527392461895943,
+      "step": 7980
+    },
+    {
+      "epoch": 7.1852517985611515,
+      "grad_norm": 0.6273057460784912,
+      "learning_rate": 9.81216396117485e-05,
+      "loss": 0.0301,
+      "step": 7990
+    },
+    {
+      "action_loss": 0.014880210161209106,
+      "epoch": 7.1852517985611515,
+      "step": 7990
+    },
+    {
+      "epoch": 7.1852517985611515,
+      "step": 7990,
+      "torque_loss": 0.1347213238477707
+    },
+    {
+      "epoch": 7.1852517985611515,
+      "force_loss": 0.01680292934179306,
+      "step": 7990
+    },
+    {
+      "epoch": 7.194244604316546,
+      "grad_norm": 0.8474001884460449,
+      "learning_rate": 9.811414980049491e-05,
+      "loss": 0.032,
+      "step": 8000
+    },
+    {
+      "action_loss": 0.02514682523906231,
+      "epoch": 7.194244604316546,
+      "step": 8000
+    },
+    {
+      "epoch": 7.194244604316546,
+      "step": 8000,
+      "torque_loss": 0.1738385111093521
+    },
+    {
+      "epoch": 7.194244604316546,
+      "force_loss": 0.016921082511544228,
+      "step": 8000
+    },
+    {
+      "epoch": 7.203237410071942,
+      "grad_norm": 0.5634039640426636,
+      "learning_rate": 9.810664537343301e-05,
+      "loss": 0.0296,
+      "step": 8010
+    },
+    {
+      "action_loss": 0.023893803358078003,
+      "epoch": 7.203237410071942,
+      "step": 8010
+    },
+    {
+      "epoch": 7.203237410071942,
+      "step": 8010,
+      "torque_loss": 0.14144329726696014
+    },
+    {
+      "epoch": 7.203237410071942,
+      "force_loss": 0.017246244475245476,
+      "step": 8010
+    },
+    {
+      "epoch": 7.212230215827338,
+      "grad_norm": 0.8652377128601074,
+      "learning_rate": 9.809912633284243e-05,
+      "loss": 0.0307,
+      "step": 8020
+    },
+    {
+      "action_loss": 0.01660771109163761,
+      "epoch": 7.212230215827338,
+      "step": 8020
+    },
+    {
+      "epoch": 7.212230215827338,
+      "step": 8020,
+      "torque_loss": 0.15797439217567444
+    },
+    {
+      "epoch": 7.212230215827338,
+      "force_loss": 0.012745119631290436,
+      "step": 8020
+    },
+    {
+      "epoch": 7.221223021582734,
+      "grad_norm": 0.8667154312133789,
+      "learning_rate": 9.809159268100725e-05,
+      "loss": 0.0296,
+      "step": 8030
+    },
+    {
+      "action_loss": 0.00831642746925354,
+      "epoch": 7.221223021582734,
+      "step": 8030
+    },
+    {
+      "epoch": 7.221223021582734,
+      "step": 8030,
+      "torque_loss": 0.14936727285385132
+    },
+    {
+      "epoch": 7.221223021582734,
+      "force_loss": 0.006427200045436621,
+      "step": 8030
+    },
+    {
+      "epoch": 7.23021582733813,
+      "grad_norm": 0.843334436416626,
+      "learning_rate": 9.808404442021599e-05,
+      "loss": 0.031,
+      "step": 8040
+    },
+    {
+      "action_loss": 0.008103086613118649,
+      "epoch": 7.23021582733813,
+      "step": 8040
+    },
+    {
+      "epoch": 7.23021582733813,
+      "step": 8040,
+      "torque_loss": 0.14466805756092072
+    },
+    {
+      "epoch": 7.23021582733813,
+      "force_loss": 0.009107830002903938,
+      "step": 8040
+    },
+    {
+      "epoch": 7.239208633093525,
+      "grad_norm": 1.0166850090026855,
+      "learning_rate": 9.807648155276163e-05,
+      "loss": 0.0313,
+      "step": 8050
+    },
+    {
+      "action_loss": 0.016420679166913033,
+      "epoch": 7.239208633093525,
+      "step": 8050
+    },
+    {
+      "epoch": 7.239208633093525,
+      "step": 8050,
+      "torque_loss": 0.13282187283039093
+    },
+    {
+      "epoch": 7.239208633093525,
+      "force_loss": 0.014203312806785107,
+      "step": 8050
+    },
+    {
+      "epoch": 7.248201438848921,
+      "grad_norm": 0.8945342898368835,
+      "learning_rate": 9.806890408094156e-05,
+      "loss": 0.0325,
+      "step": 8060
+    },
+    {
+      "action_loss": 0.021993523463606834,
+      "epoch": 7.248201438848921,
+      "step": 8060
+    },
+    {
+      "epoch": 7.248201438848921,
+      "step": 8060,
+      "torque_loss": 0.11349403858184814
+    },
+    {
+      "epoch": 7.248201438848921,
+      "force_loss": 0.03570099547505379,
+      "step": 8060
+    },
+    {
+      "epoch": 7.2571942446043165,
+      "grad_norm": 1.963100552558899,
+      "learning_rate": 9.806131200705761e-05,
+      "loss": 0.0326,
+      "step": 8070
+    },
+    {
+      "action_loss": 0.01009267196059227,
+      "epoch": 7.2571942446043165,
+      "step": 8070
+    },
+    {
+      "epoch": 7.2571942446043165,
+      "step": 8070,
+      "torque_loss": 0.13126297295093536
+    },
+    {
+      "epoch": 7.2571942446043165,
+      "force_loss": 0.015860306099057198,
+      "step": 8070
+    },
+    {
+      "epoch": 7.266187050359712,
+      "grad_norm": 0.19442281126976013,
+      "learning_rate": 9.805370533341605e-05,
+      "loss": 0.0334,
+      "step": 8080
+    },
+    {
+      "action_loss": 0.01956632174551487,
+      "epoch": 7.266187050359712,
+      "step": 8080
+    },
+    {
+      "epoch": 7.266187050359712,
+      "step": 8080,
+      "torque_loss": 0.1147812083363533
+    },
+    {
+      "epoch": 7.266187050359712,
+      "force_loss": 0.012806649319827557,
+      "step": 8080
+    },
+    {
+      "epoch": 7.275179856115108,
+      "grad_norm": 0.6823049187660217,
+      "learning_rate": 9.804608406232762e-05,
+      "loss": 0.0272,
+      "step": 8090
+    },
+    {
+      "action_loss": 0.007072980049997568,
+      "epoch": 7.275179856115108,
+      "step": 8090
+    },
+    {
+      "epoch": 7.275179856115108,
+      "step": 8090,
+      "torque_loss": 0.13415533304214478
+    },
+    {
+      "epoch": 7.275179856115108,
+      "force_loss": 0.007197019178420305,
+      "step": 8090
+    },
+    {
+      "epoch": 7.284172661870503,
+      "grad_norm": 0.589218020439148,
+      "learning_rate": 9.803844819610741e-05,
+      "loss": 0.0316,
+      "step": 8100
+    },
+    {
+      "action_loss": 0.015059317462146282,
+      "epoch": 7.284172661870503,
+      "step": 8100
+    },
+    {
+      "epoch": 7.284172661870503,
+      "step": 8100,
+      "torque_loss": 0.15744559466838837
+    },
+    {
+      "epoch": 7.284172661870503,
+      "force_loss": 0.015000968240201473,
+      "step": 8100
+    },
+    {
+      "epoch": 7.293165467625899,
+      "grad_norm": 0.48272275924682617,
+      "learning_rate": 9.803079773707504e-05,
+      "loss": 0.0284,
+      "step": 8110
+    },
+    {
+      "action_loss": 0.015221621841192245,
+      "epoch": 7.293165467625899,
+      "step": 8110
+    },
+    {
+      "epoch": 7.293165467625899,
+      "step": 8110,
+      "torque_loss": 0.18278925120830536
+    },
+    {
+      "epoch": 7.293165467625899,
+      "force_loss": 0.014225400984287262,
+      "step": 8110
+    },
+    {
+      "epoch": 7.302158273381295,
+      "grad_norm": 0.7312065362930298,
+      "learning_rate": 9.802313268755447e-05,
+      "loss": 0.0371,
+      "step": 8120
+    },
+    {
+      "action_loss": 0.027010740712285042,
+      "epoch": 7.302158273381295,
+      "step": 8120
+    },
+    {
+      "epoch": 7.302158273381295,
+      "step": 8120,
+      "torque_loss": 0.147797092795372
+    },
+    {
+      "epoch": 7.302158273381295,
+      "force_loss": 0.02799549140036106,
+      "step": 8120
+    },
+    {
+      "epoch": 7.311151079136691,
+      "grad_norm": 1.2887564897537231,
+      "learning_rate": 9.801545304987419e-05,
+      "loss": 0.0329,
+      "step": 8130
+    },
+    {
+      "action_loss": 0.015190199948847294,
+      "epoch": 7.311151079136691,
+      "step": 8130
+    },
+    {
+      "epoch": 7.311151079136691,
+      "step": 8130,
+      "torque_loss": 0.14864178001880646
+    },
+    {
+      "epoch": 7.311151079136691,
+      "force_loss": 0.01111955102533102,
+      "step": 8130
+    },
+    {
+      "epoch": 7.320143884892087,
+      "grad_norm": 0.3698447048664093,
+      "learning_rate": 9.800775882636704e-05,
+      "loss": 0.032,
+      "step": 8140
+    },
+    {
+      "action_loss": 0.02153765596449375,
+      "epoch": 7.320143884892087,
+      "step": 8140
+    },
+    {
+      "epoch": 7.320143884892087,
+      "step": 8140,
+      "torque_loss": 0.1726057082414627
+    },
+    {
+      "epoch": 7.320143884892087,
+      "force_loss": 0.031489450484514236,
+      "step": 8140
+    },
+    {
+      "epoch": 7.329136690647482,
+      "grad_norm": 0.7618156671524048,
+      "learning_rate": 9.800005001937034e-05,
+      "loss": 0.0267,
+      "step": 8150
+    },
+    {
+      "action_loss": 0.014157422818243504,
+      "epoch": 7.329136690647482,
+      "step": 8150
+    },
+    {
+      "epoch": 7.329136690647482,
+      "step": 8150,
+      "torque_loss": 0.13258889317512512
+    },
+    {
+      "epoch": 7.329136690647482,
+      "force_loss": 0.012225021608173847,
+      "step": 8150
+    },
+    {
+      "epoch": 7.338129496402877,
+      "grad_norm": 1.3190714120864868,
+      "learning_rate": 9.79923266312258e-05,
+      "loss": 0.0321,
+      "step": 8160
+    },
+    {
+      "action_loss": 0.01977492682635784,
+      "epoch": 7.338129496402877,
+      "step": 8160
+    },
+    {
+      "epoch": 7.338129496402877,
+      "step": 8160,
+      "torque_loss": 0.13922341167926788
+    },
+    {
+      "epoch": 7.338129496402877,
+      "force_loss": 0.01874031312763691,
+      "step": 8160
+    },
+    {
+      "epoch": 7.347122302158273,
+      "grad_norm": 0.620725691318512,
+      "learning_rate": 9.79845886642796e-05,
+      "loss": 0.029,
+      "step": 8170
+    },
+    {
+      "action_loss": 0.008612917736172676,
+      "epoch": 7.347122302158273,
+      "step": 8170
+    },
+    {
+      "epoch": 7.347122302158273,
+      "step": 8170,
+      "torque_loss": 0.12805069983005524
+    },
+    {
+      "epoch": 7.347122302158273,
+      "force_loss": 0.014394022524356842,
+      "step": 8170
+    },
+    {
+      "epoch": 7.356115107913669,
+      "grad_norm": 0.6814894080162048,
+      "learning_rate": 9.797683612088233e-05,
+      "loss": 0.0277,
+      "step": 8180
+    },
+    {
+      "action_loss": 0.020145997405052185,
+      "epoch": 7.356115107913669,
+      "step": 8180
+    },
+    {
+      "epoch": 7.356115107913669,
+      "step": 8180,
+      "torque_loss": 0.11983004957437515
+    },
+    {
+      "epoch": 7.356115107913669,
+      "force_loss": 0.020338408648967743,
+      "step": 8180
+    },
+    {
+      "epoch": 7.365107913669065,
+      "grad_norm": 0.8076351284980774,
+      "learning_rate": 9.796906900338898e-05,
+      "loss": 0.0321,
+      "step": 8190
+    },
+    {
+      "action_loss": 0.012436367571353912,
+      "epoch": 7.365107913669065,
+      "step": 8190
+    },
+    {
+      "epoch": 7.365107913669065,
+      "step": 8190,
+      "torque_loss": 0.15969730913639069
+    },
+    {
+      "epoch": 7.365107913669065,
+      "force_loss": 0.013429229147732258,
+      "step": 8190
+    },
+    {
+      "epoch": 7.374100719424461,
+      "grad_norm": 0.5265650749206543,
+      "learning_rate": 9.796128731415903e-05,
+      "loss": 0.032,
+      "step": 8200
+    },
+    {
+      "action_loss": 0.01803666166961193,
+      "epoch": 7.374100719424461,
+      "step": 8200
+    },
+    {
+      "epoch": 7.374100719424461,
+      "step": 8200,
+      "torque_loss": 0.15755203366279602
+    },
+    {
+      "epoch": 7.374100719424461,
+      "force_loss": 0.013762987218797207,
+      "step": 8200
+    },
+    {
+      "epoch": 7.383093525179856,
+      "grad_norm": 0.48601090908050537,
+      "learning_rate": 9.795349105555634e-05,
+      "loss": 0.0321,
+      "step": 8210
+    },
+    {
+      "action_loss": 0.012979594059288502,
+      "epoch": 7.383093525179856,
+      "step": 8210
+    },
+    {
+      "epoch": 7.383093525179856,
+      "step": 8210,
+      "torque_loss": 0.13602197170257568
+    },
+    {
+      "epoch": 7.383093525179856,
+      "force_loss": 0.016358325257897377,
+      "step": 8210
+    },
+    {
+      "epoch": 7.392086330935252,
+      "grad_norm": 0.28090977668762207,
+      "learning_rate": 9.794568022994922e-05,
+      "loss": 0.0309,
+      "step": 8220
+    },
+    {
+      "action_loss": 0.01126746367663145,
+      "epoch": 7.392086330935252,
+      "step": 8220
+    },
+    {
+      "epoch": 7.392086330935252,
+      "step": 8220,
+      "torque_loss": 0.1472492814064026
+    },
+    {
+      "epoch": 7.392086330935252,
+      "force_loss": 0.010812432505190372,
+      "step": 8220
+    },
+    {
+      "epoch": 7.401079136690647,
+      "grad_norm": 0.6091715693473816,
+      "learning_rate": 9.793785483971034e-05,
+      "loss": 0.0279,
+      "step": 8230
+    },
+    {
+      "action_loss": 0.009267676621675491,
+      "epoch": 7.401079136690647,
+      "step": 8230
+    },
+    {
+      "epoch": 7.401079136690647,
+      "step": 8230,
+      "torque_loss": 0.1545286327600479
+    },
+    {
+      "epoch": 7.401079136690647,
+      "force_loss": 0.007919063791632652,
+      "step": 8230
+    },
+    {
+      "epoch": 7.410071942446043,
+      "grad_norm": 1.1030200719833374,
+      "learning_rate": 9.793001488721691e-05,
+      "loss": 0.0326,
+      "step": 8240
+    },
+    {
+      "action_loss": 0.010712782852351665,
+      "epoch": 7.410071942446043,
+      "step": 8240
+    },
+    {
+      "epoch": 7.410071942446043,
+      "step": 8240,
+      "torque_loss": 0.10139413923025131
+    },
+    {
+      "epoch": 7.410071942446043,
+      "force_loss": 0.00794156827032566,
+      "step": 8240
+    },
+    {
+      "epoch": 7.419064748201439,
+      "grad_norm": 1.2643754482269287,
+      "learning_rate": 9.792216037485047e-05,
+      "loss": 0.0314,
+      "step": 8250
+    },
+    {
+      "action_loss": 0.013765938580036163,
+      "epoch": 7.419064748201439,
+      "step": 8250
+    },
+    {
+      "epoch": 7.419064748201439,
+      "step": 8250,
+      "torque_loss": 0.1910661906003952
+    },
+    {
+      "epoch": 7.419064748201439,
+      "force_loss": 0.013252117671072483,
+      "step": 8250
+    },
+    {
+      "epoch": 7.428057553956835,
+      "grad_norm": 0.582712709903717,
+      "learning_rate": 9.791429130499704e-05,
+      "loss": 0.0308,
+      "step": 8260
+    },
+    {
+      "action_loss": 0.018466627225279808,
+      "epoch": 7.428057553956835,
+      "step": 8260
+    },
+    {
+      "epoch": 7.428057553956835,
+      "step": 8260,
+      "torque_loss": 0.13622981309890747
+    },
+    {
+      "epoch": 7.428057553956835,
+      "force_loss": 0.016750799492001534,
+      "step": 8260
+    },
+    {
+      "epoch": 7.43705035971223,
+      "grad_norm": 0.5868211984634399,
+      "learning_rate": 9.790640768004698e-05,
+      "loss": 0.03,
+      "step": 8270
+    },
+    {
+      "action_loss": 0.01863769069314003,
+      "epoch": 7.43705035971223,
+      "step": 8270
+    },
+    {
+      "epoch": 7.43705035971223,
+      "step": 8270,
+      "torque_loss": 0.13692380487918854
+    },
+    {
+      "epoch": 7.43705035971223,
+      "force_loss": 0.021651098504662514,
+      "step": 8270
+    },
+    {
+      "epoch": 7.446043165467626,
+      "grad_norm": 1.2289541959762573,
+      "learning_rate": 9.789850950239518e-05,
+      "loss": 0.0351,
+      "step": 8280
+    },
+    {
+      "action_loss": 0.013784590177237988,
+      "epoch": 7.446043165467626,
+      "step": 8280
+    },
+    {
+      "epoch": 7.446043165467626,
+      "step": 8280,
+      "torque_loss": 0.1744844913482666
+    },
+    {
+      "epoch": 7.446043165467626,
+      "force_loss": 0.010908793658018112,
+      "step": 8280
+    },
+    {
+      "epoch": 7.455035971223022,
+      "grad_norm": 0.647811770439148,
+      "learning_rate": 9.789059677444089e-05,
+      "loss": 0.0295,
+      "step": 8290
+    },
+    {
+      "action_loss": 0.008474532514810562,
+      "epoch": 7.455035971223022,
+      "step": 8290
+    },
+    {
+      "epoch": 7.455035971223022,
+      "step": 8290,
+      "torque_loss": 0.1466941386461258
+    },
+    {
+      "epoch": 7.455035971223022,
+      "force_loss": 0.007238794583827257,
+      "step": 8290
+    },
+    {
+      "epoch": 7.4640287769784175,
+      "grad_norm": 0.5557348728179932,
+      "learning_rate": 9.788266949858776e-05,
+      "loss": 0.0249,
+      "step": 8300
+    },
+    {
+      "action_loss": 0.012704807333648205,
+      "epoch": 7.4640287769784175,
+      "step": 8300
+    },
+    {
+      "epoch": 7.4640287769784175,
+      "step": 8300,
+      "torque_loss": 0.13206680119037628
+    },
+    {
+      "epoch": 7.4640287769784175,
+      "force_loss": 0.010610085912048817,
+      "step": 8300
+    },
+    {
+      "epoch": 7.473021582733813,
+      "grad_norm": 0.5853516459465027,
+      "learning_rate": 9.787472767724392e-05,
+      "loss": 0.0245,
+      "step": 8310
+    },
+    {
+      "action_loss": 0.007080281153321266,
+      "epoch": 7.473021582733813,
+      "step": 8310
+    },
+    {
+      "epoch": 7.473021582733813,
+      "step": 8310,
+      "torque_loss": 0.11170990020036697
+    },
+    {
+      "epoch": 7.473021582733813,
+      "force_loss": 0.0078382333740592,
+      "step": 8310
+    },
+    {
+      "epoch": 7.482014388489208,
+      "grad_norm": 0.8179770708084106,
+      "learning_rate": 9.786677131282185e-05,
+      "loss": 0.0311,
+      "step": 8320
+    },
+    {
+      "action_loss": 0.007529696449637413,
+      "epoch": 7.482014388489208,
+      "step": 8320
+    },
+    {
+      "epoch": 7.482014388489208,
+      "step": 8320,
+      "torque_loss": 0.09825699776411057
+    },
+    {
+      "epoch": 7.482014388489208,
+      "force_loss": 0.009542413055896759,
+      "step": 8320
+    },
+    {
+      "epoch": 7.491007194244604,
+      "grad_norm": 0.38167914748191833,
+      "learning_rate": 9.785880040773853e-05,
+      "loss": 0.0278,
+      "step": 8330
+    },
+    {
+      "action_loss": 0.01627209782600403,
+      "epoch": 7.491007194244604,
+      "step": 8330
+    },
+    {
+      "epoch": 7.491007194244604,
+      "step": 8330,
+      "torque_loss": 0.1791231781244278
+    },
+    {
+      "epoch": 7.491007194244604,
+      "force_loss": 0.00902497861534357,
+      "step": 8330
+    },
+    {
+      "epoch": 7.5,
+      "grad_norm": 1.403141736984253,
+      "learning_rate": 9.785081496441527e-05,
+      "loss": 0.0368,
+      "step": 8340
+    },
+    {
+      "action_loss": 0.012243962846696377,
+      "epoch": 7.5,
+      "step": 8340
+    },
+    {
+      "epoch": 7.5,
+      "step": 8340,
+      "torque_loss": 0.12263405323028564
+    },
+    {
+      "epoch": 7.5,
+      "force_loss": 0.010465558618307114,
+      "step": 8340
+    },
+    {
+      "epoch": 7.508992805755396,
+      "grad_norm": 1.248809814453125,
+      "learning_rate": 9.784281498527785e-05,
+      "loss": 0.0315,
+      "step": 8350
+    },
+    {
+      "action_loss": 0.013629280030727386,
+      "epoch": 7.508992805755396,
+      "step": 8350
+    },
+    {
+      "epoch": 7.508992805755396,
+      "step": 8350,
+      "torque_loss": 0.13109223544597626
+    },
+    {
+      "epoch": 7.508992805755396,
+      "force_loss": 0.015668490901589394,
+      "step": 8350
+    },
+    {
+      "epoch": 7.517985611510792,
+      "grad_norm": 1.1658705472946167,
+      "learning_rate": 9.783480047275646e-05,
+      "loss": 0.0376,
+      "step": 8360
+    },
+    {
+      "action_loss": 0.01723673939704895,
+      "epoch": 7.517985611510792,
+      "step": 8360
+    },
+    {
+      "epoch": 7.517985611510792,
+      "step": 8360,
+      "torque_loss": 0.11456664651632309
+    },
+    {
+      "epoch": 7.517985611510792,
+      "force_loss": 0.01445191353559494,
+      "step": 8360
+    },
+    {
+      "epoch": 7.5269784172661875,
+      "grad_norm": 0.44663509726524353,
+      "learning_rate": 9.78267714292857e-05,
+      "loss": 0.0358,
+      "step": 8370
+    },
+    {
+      "action_loss": 0.006761031690984964,
+      "epoch": 7.5269784172661875,
+      "step": 8370
+    },
+    {
+      "epoch": 7.5269784172661875,
+      "step": 8370,
+      "torque_loss": 0.11758812516927719
+    },
+    {
+      "epoch": 7.5269784172661875,
+      "force_loss": 0.007492410484701395,
+      "step": 8370
+    },
+    {
+      "epoch": 7.5359712230215825,
+      "grad_norm": 0.5402593612670898,
+      "learning_rate": 9.781872785730454e-05,
+      "loss": 0.031,
+      "step": 8380
+    },
+    {
+      "action_loss": 0.010334520600736141,
+      "epoch": 7.5359712230215825,
+      "step": 8380
+    },
+    {
+      "epoch": 7.5359712230215825,
+      "step": 8380,
+      "torque_loss": 0.1553741842508316
+    },
+    {
+      "epoch": 7.5359712230215825,
+      "force_loss": 0.009717588312923908,
+      "step": 8380
+    },
+    {
+      "epoch": 7.544964028776978,
+      "grad_norm": 0.9734342098236084,
+      "learning_rate": 9.781066975925646e-05,
+      "loss": 0.0277,
+      "step": 8390
+    },
+    {
+      "action_loss": 0.018119921907782555,
+      "epoch": 7.544964028776978,
+      "step": 8390
+    },
+    {
+      "epoch": 7.544964028776978,
+      "step": 8390,
+      "torque_loss": 0.11200815439224243
+    },
+    {
+      "epoch": 7.544964028776978,
+      "force_loss": 0.01347032655030489,
+      "step": 8390
+    },
+    {
+      "epoch": 7.553956834532374,
+      "grad_norm": 0.5535122156143188,
+      "learning_rate": 9.780259713758928e-05,
+      "loss": 0.0293,
+      "step": 8400
+    },
+    {
+      "action_loss": 0.007479672785848379,
+      "epoch": 7.553956834532374,
+      "step": 8400
+    },
+    {
+      "epoch": 7.553956834532374,
+      "step": 8400,
+      "torque_loss": 0.12148300558328629
+    },
+    {
+      "epoch": 7.553956834532374,
+      "force_loss": 0.01105521246790886,
+      "step": 8400
+    },
+    {
+      "epoch": 7.56294964028777,
+      "grad_norm": 0.7123938798904419,
+      "learning_rate": 9.779450999475524e-05,
+      "loss": 0.0259,
+      "step": 8410
+    },
+    {
+      "action_loss": 0.010749242268502712,
+      "epoch": 7.56294964028777,
+      "step": 8410
+    },
+    {
+      "epoch": 7.56294964028777,
+      "step": 8410,
+      "torque_loss": 0.1403711587190628
+    },
+    {
+      "epoch": 7.56294964028777,
+      "force_loss": 0.021566079929471016,
+      "step": 8410
+    },
+    {
+      "epoch": 7.571942446043165,
+      "grad_norm": 0.8686343431472778,
+      "learning_rate": 9.7786408333211e-05,
+      "loss": 0.0289,
+      "step": 8420
+    },
+    {
+      "action_loss": 0.00947518553584814,
+      "epoch": 7.571942446043165,
+      "step": 8420
+    },
+    {
+      "epoch": 7.571942446043165,
+      "step": 8420,
+      "torque_loss": 0.16997653245925903
+    },
+    {
+      "epoch": 7.571942446043165,
+      "force_loss": 0.007192822638899088,
+      "step": 8420
+    },
+    {
+      "epoch": 7.580935251798561,
+      "grad_norm": 0.4283559322357178,
+      "learning_rate": 9.777829215541764e-05,
+      "loss": 0.0308,
+      "step": 8430
+    },
+    {
+      "action_loss": 0.011530887335538864,
+      "epoch": 7.580935251798561,
+      "step": 8430
+    },
+    {
+      "epoch": 7.580935251798561,
+      "step": 8430,
+      "torque_loss": 0.1305180937051773
+    },
+    {
+      "epoch": 7.580935251798561,
+      "force_loss": 0.010568656958639622,
+      "step": 8430
+    },
+    {
+      "epoch": 7.589928057553957,
+      "grad_norm": 0.48228919506073,
+      "learning_rate": 9.777016146384064e-05,
+      "loss": 0.0284,
+      "step": 8440
+    },
+    {
+      "action_loss": 0.027326839044690132,
+      "epoch": 7.589928057553957,
+      "step": 8440
+    },
+    {
+      "epoch": 7.589928057553957,
+      "step": 8440,
+      "torque_loss": 0.14576278626918793
+    },
+    {
+      "epoch": 7.589928057553957,
+      "force_loss": 0.024153592064976692,
+      "step": 8440
+    },
+    {
+      "epoch": 7.598920863309353,
+      "grad_norm": 0.927008330821991,
+      "learning_rate": 9.776201626094988e-05,
+      "loss": 0.0259,
+      "step": 8450
+    },
+    {
+      "action_loss": 0.01435435563325882,
+      "epoch": 7.598920863309353,
+      "step": 8450
+    },
+    {
+      "epoch": 7.598920863309353,
+      "step": 8450,
+      "torque_loss": 0.1355314701795578
+    },
+    {
+      "epoch": 7.598920863309353,
+      "force_loss": 0.007660705130547285,
+      "step": 8450
+    },
+    {
+      "epoch": 7.607913669064748,
+      "grad_norm": 0.43980029225349426,
+      "learning_rate": 9.775385654921965e-05,
+      "loss": 0.0268,
+      "step": 8460
+    },
+    {
+      "action_loss": 0.013210107572376728,
+      "epoch": 7.607913669064748,
+      "step": 8460
+    },
+    {
+      "epoch": 7.607913669064748,
+      "step": 8460,
+      "torque_loss": 0.16153599321842194
+    },
+    {
+      "epoch": 7.607913669064748,
+      "force_loss": 0.012342006899416447,
+      "step": 8460
+    },
+    {
+      "epoch": 7.616906474820144,
+      "grad_norm": 0.793596625328064,
+      "learning_rate": 9.774568233112868e-05,
+      "loss": 0.032,
+      "step": 8470
+    },
+    {
+      "action_loss": 0.009867441840469837,
+      "epoch": 7.616906474820144,
+      "step": 8470
+    },
+    {
+      "epoch": 7.616906474820144,
+      "step": 8470,
+      "torque_loss": 0.0989953801035881
+    },
+    {
+      "epoch": 7.616906474820144,
+      "force_loss": 0.007182652596384287,
+      "step": 8470
+    },
+    {
+      "epoch": 7.625899280575539,
+      "grad_norm": 0.7396289706230164,
+      "learning_rate": 9.773749360916007e-05,
+      "loss": 0.0321,
+      "step": 8480
+    },
+    {
+      "action_loss": 0.017461545765399933,
+      "epoch": 7.625899280575539,
+      "step": 8480
+    },
+    {
+      "epoch": 7.625899280575539,
+      "step": 8480,
+      "torque_loss": 0.12636858224868774
+    },
+    {
+      "epoch": 7.625899280575539,
+      "force_loss": 0.01598258875310421,
+      "step": 8480
+    },
+    {
+      "epoch": 7.634892086330935,
+      "grad_norm": 1.1434593200683594,
+      "learning_rate": 9.772929038580134e-05,
+      "loss": 0.0349,
+      "step": 8490
+    },
+    {
+      "action_loss": 0.012149865739047527,
+      "epoch": 7.634892086330935,
+      "step": 8490
+    },
+    {
+      "epoch": 7.634892086330935,
+      "step": 8490,
+      "torque_loss": 0.1339341253042221
+    },
+    {
+      "epoch": 7.634892086330935,
+      "force_loss": 0.01241689920425415,
+      "step": 8490
+    },
+    {
+      "epoch": 7.643884892086331,
+      "grad_norm": 1.122774362564087,
+      "learning_rate": 9.772107266354439e-05,
+      "loss": 0.0356,
+      "step": 8500
+    },
+    {
+      "action_loss": 0.0162548515945673,
+      "epoch": 7.643884892086331,
+      "step": 8500
+    },
+    {
+      "epoch": 7.643884892086331,
+      "step": 8500,
+      "torque_loss": 0.19368226826190948
+    },
+    {
+      "epoch": 7.643884892086331,
+      "force_loss": 0.010798695497214794,
+      "step": 8500
+    },
+    {
+      "epoch": 7.652877697841727,
+      "grad_norm": 0.6731473803520203,
+      "learning_rate": 9.77128404448856e-05,
+      "loss": 0.0378,
+      "step": 8510
+    },
+    {
+      "action_loss": 0.008567575365304947,
+      "epoch": 7.652877697841727,
+      "step": 8510
+    },
+    {
+      "epoch": 7.652877697841727,
+      "step": 8510,
+      "torque_loss": 0.14658944308757782
+    },
+    {
+      "epoch": 7.652877697841727,
+      "force_loss": 0.006835754960775375,
+      "step": 8510
+    },
+    {
+      "epoch": 7.661870503597123,
+      "grad_norm": 1.0794564485549927,
+      "learning_rate": 9.770459373232565e-05,
+      "loss": 0.0268,
+      "step": 8520
+    },
+    {
+      "action_loss": 0.004705252591520548,
+      "epoch": 7.661870503597123,
+      "step": 8520
+    },
+    {
+      "epoch": 7.661870503597123,
+      "step": 8520,
+      "torque_loss": 0.09792991727590561
+    },
+    {
+      "epoch": 7.661870503597123,
+      "force_loss": 0.006157430354505777,
+      "step": 8520
+    },
+    {
+      "epoch": 7.670863309352518,
+      "grad_norm": 0.55097895860672,
+      "learning_rate": 9.769633252836969e-05,
+      "loss": 0.0309,
+      "step": 8530
+    },
+    {
+      "action_loss": 0.010069550015032291,
+      "epoch": 7.670863309352518,
+      "step": 8530
+    },
+    {
+      "epoch": 7.670863309352518,
+      "step": 8530,
+      "torque_loss": 0.09578531980514526
+    },
+    {
+      "epoch": 7.670863309352518,
+      "force_loss": 0.007114263717085123,
+      "step": 8530
+    },
+    {
+      "epoch": 7.679856115107913,
+      "grad_norm": 1.0670244693756104,
+      "learning_rate": 9.768805683552724e-05,
+      "loss": 0.026,
+      "step": 8540
+    },
+    {
+      "action_loss": 0.01338367909193039,
+      "epoch": 7.679856115107913,
+      "step": 8540
+    },
+    {
+      "epoch": 7.679856115107913,
+      "step": 8540,
+      "torque_loss": 0.10631659626960754
+    },
+    {
+      "epoch": 7.679856115107913,
+      "force_loss": 0.00999636109918356,
+      "step": 8540
+    },
+    {
+      "epoch": 7.688848920863309,
+      "grad_norm": 0.9078983664512634,
+      "learning_rate": 9.767976665631228e-05,
+      "loss": 0.0281,
+      "step": 8550
+    },
+    {
+      "action_loss": 0.007202318403869867,
+      "epoch": 7.688848920863309,
+      "step": 8550
+    },
+    {
+      "epoch": 7.688848920863309,
+      "step": 8550,
+      "torque_loss": 0.12337324768304825
+    },
+    {
+      "epoch": 7.688848920863309,
+      "force_loss": 0.006715590599924326,
+      "step": 8550
+    },
+    {
+      "epoch": 7.697841726618705,
+      "grad_norm": 0.40653520822525024,
+      "learning_rate": 9.767146199324311e-05,
+      "loss": 0.0279,
+      "step": 8560
+    },
+    {
+      "action_loss": 0.01734321378171444,
+      "epoch": 7.697841726618705,
+      "step": 8560
+    },
+    {
+      "epoch": 7.697841726618705,
+      "step": 8560,
+      "torque_loss": 0.17517967522144318
+    },
+    {
+      "epoch": 7.697841726618705,
+      "force_loss": 0.021837225183844566,
+      "step": 8560
+    },
+    {
+      "epoch": 7.706834532374101,
+      "grad_norm": 0.5068568587303162,
+      "learning_rate": 9.766314284884249e-05,
+      "loss": 0.0259,
+      "step": 8570
+    },
+    {
+      "action_loss": 0.016721414402127266,
+      "epoch": 7.706834532374101,
+      "step": 8570
+    },
+    {
+      "epoch": 7.706834532374101,
+      "step": 8570,
+      "torque_loss": 0.13541068136692047
+    },
+    {
+      "epoch": 7.706834532374101,
+      "force_loss": 0.020000582560896873,
+      "step": 8570
+    },
+    {
+      "epoch": 7.715827338129497,
+      "grad_norm": 1.5598217248916626,
+      "learning_rate": 9.765480922563752e-05,
+      "loss": 0.0301,
+      "step": 8580
+    },
+    {
+      "action_loss": 0.020986055955290794,
+      "epoch": 7.715827338129497,
+      "step": 8580
+    },
+    {
+      "epoch": 7.715827338129497,
+      "step": 8580,
+      "torque_loss": 0.1576753705739975
+    },
+    {
+      "epoch": 7.715827338129497,
+      "force_loss": 0.015498235821723938,
+      "step": 8580
+    },
+    {
+      "epoch": 7.724820143884892,
+      "grad_norm": 0.7705401182174683,
+      "learning_rate": 9.764646112615978e-05,
+      "loss": 0.0301,
+      "step": 8590
+    },
+    {
+      "action_loss": 0.016865838319063187,
+      "epoch": 7.724820143884892,
+      "step": 8590
+    },
+    {
+      "epoch": 7.724820143884892,
+      "step": 8590,
+      "torque_loss": 0.17905347049236298
+    },
+    {
+      "epoch": 7.724820143884892,
+      "force_loss": 0.01965884305536747,
+      "step": 8590
+    },
+    {
+      "epoch": 7.733812949640288,
+      "grad_norm": 0.9407777786254883,
+      "learning_rate": 9.763809855294517e-05,
+      "loss": 0.0305,
+      "step": 8600
+    },
+    {
+      "action_loss": 0.020679781213402748,
+      "epoch": 7.733812949640288,
+      "step": 8600
+    },
+    {
+      "epoch": 7.733812949640288,
+      "step": 8600,
+      "torque_loss": 0.1808556765317917
+    },
+    {
+      "epoch": 7.733812949640288,
+      "force_loss": 0.03475014492869377,
+      "step": 8600
+    },
+    {
+      "epoch": 7.7428057553956835,
+      "grad_norm": 0.7985976338386536,
+      "learning_rate": 9.762972150853404e-05,
+      "loss": 0.0286,
+      "step": 8610
+    },
+    {
+      "action_loss": 0.008926980197429657,
+      "epoch": 7.7428057553956835,
+      "step": 8610
+    },
+    {
+      "epoch": 7.7428057553956835,
+      "step": 8610,
+      "torque_loss": 0.16814644634723663
+    },
+    {
+      "epoch": 7.7428057553956835,
+      "force_loss": 0.010465036146342754,
+      "step": 8610
+    },
+    {
+      "epoch": 7.751798561151079,
+      "grad_norm": 1.1951462030410767,
+      "learning_rate": 9.762132999547111e-05,
+      "loss": 0.0278,
+      "step": 8620
+    },
+    {
+      "action_loss": 0.06406699866056442,
+      "epoch": 7.751798561151079,
+      "step": 8620
+    },
+    {
+      "epoch": 7.751798561151079,
+      "step": 8620,
+      "torque_loss": 0.17853160202503204
+    },
+    {
+      "epoch": 7.751798561151079,
+      "force_loss": 0.06099233031272888,
+      "step": 8620
+    },
+    {
+      "epoch": 7.760791366906475,
+      "grad_norm": 1.1469022035598755,
+      "learning_rate": 9.761292401630549e-05,
+      "loss": 0.0335,
+      "step": 8630
+    },
+    {
+      "action_loss": 0.011628423817455769,
+      "epoch": 7.760791366906475,
+      "step": 8630
+    },
+    {
+      "epoch": 7.760791366906475,
+      "step": 8630,
+      "torque_loss": 0.14463847875595093
+    },
+    {
+      "epoch": 7.760791366906475,
+      "force_loss": 0.018112724646925926,
+      "step": 8630
+    },
+    {
+      "epoch": 7.76978417266187,
+      "grad_norm": 0.6961926221847534,
+      "learning_rate": 9.76045035735907e-05,
+      "loss": 0.028,
+      "step": 8640
+    },
+    {
+      "action_loss": 0.008572163991630077,
+      "epoch": 7.76978417266187,
+      "step": 8640
+    },
+    {
+      "epoch": 7.76978417266187,
+      "step": 8640,
+      "torque_loss": 0.18002033233642578
+    },
+    {
+      "epoch": 7.76978417266187,
+      "force_loss": 0.008644177578389645,
+      "step": 8640
+    },
+    {
+      "epoch": 7.778776978417266,
+      "grad_norm": 1.1041719913482666,
+      "learning_rate": 9.759606866988464e-05,
+      "loss": 0.0292,
+      "step": 8650
+    },
+    {
+      "action_loss": 0.016385432332754135,
+      "epoch": 7.778776978417266,
+      "step": 8650
+    },
+    {
+      "epoch": 7.778776978417266,
+      "step": 8650,
+      "torque_loss": 0.1338357925415039
+    },
+    {
+      "epoch": 7.778776978417266,
+      "force_loss": 0.014097783714532852,
+      "step": 8650
+    },
+    {
+      "epoch": 7.787769784172662,
+      "grad_norm": 1.7157633304595947,
+      "learning_rate": 9.758761930774963e-05,
+      "loss": 0.0254,
+      "step": 8660
+    },
+    {
+      "action_loss": 0.017985621467232704,
+      "epoch": 7.787769784172662,
+      "step": 8660
+    },
+    {
+      "epoch": 7.787769784172662,
+      "step": 8660,
+      "torque_loss": 0.1349223554134369
+    },
+    {
+      "epoch": 7.787769784172662,
+      "force_loss": 0.014778709970414639,
+      "step": 8660
+    },
+    {
+      "epoch": 7.796762589928058,
+      "grad_norm": 0.7497832179069519,
+      "learning_rate": 9.757915548975235e-05,
+      "loss": 0.0271,
+      "step": 8670
+    },
+    {
+      "action_loss": 0.019775187596678734,
+      "epoch": 7.796762589928058,
+      "step": 8670
+    },
+    {
+      "epoch": 7.796762589928058,
+      "step": 8670,
+      "torque_loss": 0.13471777737140656
+    },
+    {
+      "epoch": 7.796762589928058,
+      "force_loss": 0.026757625862956047,
+      "step": 8670
+    },
+    {
+      "epoch": 7.805755395683454,
+      "grad_norm": 1.151245355606079,
+      "learning_rate": 9.757067721846389e-05,
+      "loss": 0.0295,
+      "step": 8680
+    },
+    {
+      "action_loss": 0.01734592393040657,
+      "epoch": 7.805755395683454,
+      "step": 8680
+    },
+    {
+      "epoch": 7.805755395683454,
+      "step": 8680,
+      "torque_loss": 0.10999622195959091
+    },
+    {
+      "epoch": 7.805755395683454,
+      "force_loss": 0.01157617848366499,
+      "step": 8680
+    },
+    {
+      "epoch": 7.814748201438849,
+      "grad_norm": 0.6144757270812988,
+      "learning_rate": 9.756218449645971e-05,
+      "loss": 0.0297,
+      "step": 8690
+    },
+    {
+      "action_loss": 0.007410380989313126,
+      "epoch": 7.814748201438849,
+      "step": 8690
+    },
+    {
+      "epoch": 7.814748201438849,
+      "step": 8690,
+      "torque_loss": 0.17453980445861816
+    },
+    {
+      "epoch": 7.814748201438849,
+      "force_loss": 0.006194221321493387,
+      "step": 8690
+    },
+    {
+      "epoch": 7.823741007194244,
+      "grad_norm": 1.1518076658248901,
+      "learning_rate": 9.75536773263197e-05,
+      "loss": 0.0312,
+      "step": 8700
+    },
+    {
+      "action_loss": 0.013584806583821774,
+      "epoch": 7.823741007194244,
+      "step": 8700
+    },
+    {
+      "epoch": 7.823741007194244,
+      "step": 8700,
+      "torque_loss": 0.15442724525928497
+    },
+    {
+      "epoch": 7.823741007194244,
+      "force_loss": 0.019365733489394188,
+      "step": 8700
+    },
+    {
+      "epoch": 7.83273381294964,
+      "grad_norm": 1.2157481908798218,
+      "learning_rate": 9.75451557106281e-05,
+      "loss": 0.0273,
+      "step": 8710
+    },
+    {
+      "action_loss": 0.013075500726699829,
+      "epoch": 7.83273381294964,
+      "step": 8710
+    },
+    {
+      "epoch": 7.83273381294964,
+      "step": 8710,
+      "torque_loss": 0.1235220655798912
+    },
+    {
+      "epoch": 7.83273381294964,
+      "force_loss": 0.008428572677075863,
+      "step": 8710
+    },
+    {
+      "epoch": 7.841726618705036,
+      "grad_norm": 0.925574779510498,
+      "learning_rate": 9.753661965197354e-05,
+      "loss": 0.0304,
+      "step": 8720
+    },
+    {
+      "action_loss": 0.011055998504161835,
+      "epoch": 7.841726618705036,
+      "step": 8720
+    },
+    {
+      "epoch": 7.841726618705036,
+      "step": 8720,
+      "torque_loss": 0.16982437670230865
+    },
+    {
+      "epoch": 7.841726618705036,
+      "force_loss": 0.009769403375685215,
+      "step": 8720
+    },
+    {
+      "epoch": 7.850719424460432,
+      "grad_norm": 0.841577410697937,
+      "learning_rate": 9.752806915294908e-05,
+      "loss": 0.0304,
+      "step": 8730
+    },
+    {
+      "action_loss": 0.017190979793667793,
+      "epoch": 7.850719424460432,
+      "step": 8730
+    },
+    {
+      "epoch": 7.850719424460432,
+      "step": 8730,
+      "torque_loss": 0.15277527272701263
+    },
+    {
+      "epoch": 7.850719424460432,
+      "force_loss": 0.012404471635818481,
+      "step": 8730
+    },
+    {
+      "epoch": 7.859712230215827,
+      "grad_norm": 0.5871753096580505,
+      "learning_rate": 9.75195042161521e-05,
+      "loss": 0.0278,
+      "step": 8740
+    },
+    {
+      "action_loss": 0.009524501860141754,
+      "epoch": 7.859712230215827,
+      "step": 8740
+    },
+    {
+      "epoch": 7.859712230215827,
+      "step": 8740,
+      "torque_loss": 0.11257639527320862
+    },
+    {
+      "epoch": 7.859712230215827,
+      "force_loss": 0.007627982646226883,
+      "step": 8740
+    },
+    {
+      "epoch": 7.868705035971223,
+      "grad_norm": 0.8407716751098633,
+      "learning_rate": 9.751092484418442e-05,
+      "loss": 0.0253,
+      "step": 8750
+    },
+    {
+      "action_loss": 0.007104267831891775,
+      "epoch": 7.868705035971223,
+      "step": 8750
+    },
+    {
+      "epoch": 7.868705035971223,
+      "step": 8750,
+      "torque_loss": 0.17534182965755463
+    },
+    {
+      "epoch": 7.868705035971223,
+      "force_loss": 0.007171860430389643,
+      "step": 8750
+    },
+    {
+      "epoch": 7.877697841726619,
+      "grad_norm": 0.7447162866592407,
+      "learning_rate": 9.750233103965224e-05,
+      "loss": 0.0305,
+      "step": 8760
+    },
+    {
+      "action_loss": 0.026125816628336906,
+      "epoch": 7.877697841726619,
+      "step": 8760
+    },
+    {
+      "epoch": 7.877697841726619,
+      "step": 8760,
+      "torque_loss": 0.18450210988521576
+    },
+    {
+      "epoch": 7.877697841726619,
+      "force_loss": 0.022638479247689247,
+      "step": 8760
+    },
+    {
+      "epoch": 7.886690647482014,
+      "grad_norm": 0.7517133951187134,
+      "learning_rate": 9.749372280516611e-05,
+      "loss": 0.0309,
+      "step": 8770
+    },
+    {
+      "action_loss": 0.009107548743486404,
+      "epoch": 7.886690647482014,
+      "step": 8770
+    },
+    {
+      "epoch": 7.886690647482014,
+      "step": 8770,
+      "torque_loss": 0.1936049610376358
+    },
+    {
+      "epoch": 7.886690647482014,
+      "force_loss": 0.0101165184751153,
+      "step": 8770
+    },
+    {
+      "epoch": 7.89568345323741,
+      "grad_norm": 0.4790474474430084,
+      "learning_rate": 9.748510014334097e-05,
+      "loss": 0.035,
+      "step": 8780
+    },
+    {
+      "action_loss": 0.024564029648900032,
+      "epoch": 7.89568345323741,
+      "step": 8780
+    },
+    {
+      "epoch": 7.89568345323741,
+      "step": 8780,
+      "torque_loss": 0.169158473610878
+    },
+    {
+      "epoch": 7.89568345323741,
+      "force_loss": 0.02265465259552002,
+      "step": 8780
+    },
+    {
+      "epoch": 7.904676258992806,
+      "grad_norm": 0.5702713131904602,
+      "learning_rate": 9.747646305679621e-05,
+      "loss": 0.0331,
+      "step": 8790
+    },
+    {
+      "action_loss": 0.019576692953705788,
+      "epoch": 7.904676258992806,
+      "step": 8790
+    },
+    {
+      "epoch": 7.904676258992806,
+      "step": 8790,
+      "torque_loss": 0.15589185059070587
+    },
+    {
+      "epoch": 7.904676258992806,
+      "force_loss": 0.020213015377521515,
+      "step": 8790
+    },
+    {
+      "epoch": 7.913669064748201,
+      "grad_norm": 0.47509273886680603,
+      "learning_rate": 9.74678115481555e-05,
+      "loss": 0.0362,
+      "step": 8800
+    },
+    {
+      "action_loss": 0.020954512059688568,
+      "epoch": 7.913669064748201,
+      "step": 8800
+    },
+    {
+      "epoch": 7.913669064748201,
+      "step": 8800,
+      "torque_loss": 0.12004091590642929
+    },
+    {
+      "epoch": 7.913669064748201,
+      "force_loss": 0.02293638326227665,
+      "step": 8800
+    },
+    {
+      "epoch": 7.922661870503597,
+      "grad_norm": 0.9124072194099426,
+      "learning_rate": 9.745914562004696e-05,
+      "loss": 0.0324,
+      "step": 8810
+    },
+    {
+      "action_loss": 0.007807473186403513,
+      "epoch": 7.922661870503597,
+      "step": 8810
+    },
+    {
+      "epoch": 7.922661870503597,
+      "step": 8810,
+      "torque_loss": 0.16129188239574432
+    },
+    {
+      "epoch": 7.922661870503597,
+      "force_loss": 0.006094722542911768,
+      "step": 8810
+    },
+    {
+      "epoch": 7.931654676258993,
+      "grad_norm": 0.8787750601768494,
+      "learning_rate": 9.745046527510307e-05,
+      "loss": 0.0297,
+      "step": 8820
+    },
+    {
+      "action_loss": 0.014297188259661198,
+      "epoch": 7.931654676258993,
+      "step": 8820
+    },
+    {
+      "epoch": 7.931654676258993,
+      "step": 8820,
+      "torque_loss": 0.14133186638355255
+    },
+    {
+      "epoch": 7.931654676258993,
+      "force_loss": 0.015875620767474174,
+      "step": 8820
+    },
+    {
+      "epoch": 7.940647482014389,
+      "grad_norm": 0.8169597387313843,
+      "learning_rate": 9.744177051596068e-05,
+      "loss": 0.03,
+      "step": 8830
+    },
+    {
+      "action_loss": 0.009781970642507076,
+      "epoch": 7.940647482014389,
+      "step": 8830
+    },
+    {
+      "epoch": 7.940647482014389,
+      "step": 8830,
+      "torque_loss": 0.13550764322280884
+    },
+    {
+      "epoch": 7.940647482014389,
+      "force_loss": 0.007129937410354614,
+      "step": 8830
+    },
+    {
+      "epoch": 7.9496402877697845,
+      "grad_norm": 1.2415217161178589,
+      "learning_rate": 9.743306134526105e-05,
+      "loss": 0.0304,
+      "step": 8840
+    },
+    {
+      "action_loss": 0.028044523671269417,
+      "epoch": 7.9496402877697845,
+      "step": 8840
+    },
+    {
+      "epoch": 7.9496402877697845,
+      "step": 8840,
+      "torque_loss": 0.1481942981481552
+    },
+    {
+      "epoch": 7.9496402877697845,
+      "force_loss": 0.027820894494652748,
+      "step": 8840
+    },
+    {
+      "epoch": 7.9586330935251794,
+      "grad_norm": 0.8753522038459778,
+      "learning_rate": 9.742433776564977e-05,
+      "loss": 0.0301,
+      "step": 8850
+    },
+    {
+      "action_loss": 0.0076384288258850574,
+      "epoch": 7.9586330935251794,
+      "step": 8850
+    },
+    {
+      "epoch": 7.9586330935251794,
+      "step": 8850,
+      "torque_loss": 0.1496245265007019
+    },
+    {
+      "epoch": 7.9586330935251794,
+      "force_loss": 0.007920184172689915,
+      "step": 8850
+    },
+    {
+      "epoch": 7.967625899280575,
+      "grad_norm": 0.4720075726509094,
+      "learning_rate": 9.741559977977683e-05,
+      "loss": 0.0283,
+      "step": 8860
+    },
+    {
+      "action_loss": 0.012563151307404041,
+      "epoch": 7.967625899280575,
+      "step": 8860
+    },
+    {
+      "epoch": 7.967625899280575,
+      "step": 8860,
+      "torque_loss": 0.13388486206531525
+    },
+    {
+      "epoch": 7.967625899280575,
+      "force_loss": 0.011008444242179394,
+      "step": 8860
+    },
+    {
+      "epoch": 7.976618705035971,
+      "grad_norm": 0.6866222620010376,
+      "learning_rate": 9.740684739029661e-05,
+      "loss": 0.0332,
+      "step": 8870
+    },
+    {
+      "action_loss": 0.014407560229301453,
+      "epoch": 7.976618705035971,
+      "step": 8870
+    },
+    {
+      "epoch": 7.976618705035971,
+      "step": 8870,
+      "torque_loss": 0.17427630722522736
+    },
+    {
+      "epoch": 7.976618705035971,
+      "force_loss": 0.01010412722826004,
+      "step": 8870
+    },
+    {
+      "epoch": 7.985611510791367,
+      "grad_norm": 0.9170600175857544,
+      "learning_rate": 9.739808059986789e-05,
+      "loss": 0.0387,
+      "step": 8880
+    },
+    {
+      "action_loss": 0.009781739674508572,
+      "epoch": 7.985611510791367,
+      "step": 8880
+    },
+    {
+      "epoch": 7.985611510791367,
+      "step": 8880,
+      "torque_loss": 0.13599783182144165
+    },
+    {
+      "epoch": 7.985611510791367,
+      "force_loss": 0.009926320053637028,
+      "step": 8880
+    },
+    {
+      "epoch": 7.994604316546763,
+      "grad_norm": 0.8177639245986938,
+      "learning_rate": 9.738929941115373e-05,
+      "loss": 0.0318,
+      "step": 8890
+    },
+    {
+      "action_loss": 0.004782909527420998,
+      "epoch": 7.994604316546763,
+      "step": 8890
+    },
+    {
+      "epoch": 7.994604316546763,
+      "step": 8890,
+      "torque_loss": 0.12804840505123138
+    },
+    {
+      "epoch": 7.994604316546763,
+      "force_loss": 0.006489449646323919,
+      "step": 8890
+    },
+    {
+      "epoch": 8.003597122302159,
+      "grad_norm": 0.4400154948234558,
+      "learning_rate": 9.738050382682167e-05,
+      "loss": 0.0225,
+      "step": 8900
+    },
+    {
+      "action_loss": 0.007002175319939852,
+      "epoch": 8.003597122302159,
+      "step": 8900
+    },
+    {
+      "epoch": 8.003597122302159,
+      "step": 8900,
+      "torque_loss": 0.10859952121973038
+    },
+    {
+      "epoch": 8.003597122302159,
+      "force_loss": 0.007067775819450617,
+      "step": 8900
+    },
+    {
+      "epoch": 8.012589928057555,
+      "grad_norm": 1.134573221206665,
+      "learning_rate": 9.737169384954355e-05,
+      "loss": 0.0346,
+      "step": 8910
+    },
+    {
+      "action_loss": 0.009668984450399876,
+      "epoch": 8.012589928057555,
+      "step": 8910
+    },
+    {
+      "epoch": 8.012589928057555,
+      "step": 8910,
+      "torque_loss": 0.13476596772670746
+    },
+    {
+      "epoch": 8.012589928057555,
+      "force_loss": 0.011064928956329823,
+      "step": 8910
+    },
+    {
+      "epoch": 8.02158273381295,
+      "grad_norm": 1.1572602987289429,
+      "learning_rate": 9.736286948199562e-05,
+      "loss": 0.0292,
+      "step": 8920
+    },
+    {
+      "action_loss": 0.01806609332561493,
+      "epoch": 8.02158273381295,
+      "step": 8920
+    },
+    {
+      "epoch": 8.02158273381295,
+      "step": 8920,
+      "torque_loss": 0.1390966773033142
+    },
+    {
+      "epoch": 8.02158273381295,
+      "force_loss": 0.009013565257191658,
+      "step": 8920
+    },
+    {
+      "epoch": 8.030575539568344,
+      "grad_norm": 1.6132428646087646,
+      "learning_rate": 9.735403072685848e-05,
+      "loss": 0.0304,
+      "step": 8930
+    },
+    {
+      "action_loss": 0.012617386877536774,
+      "epoch": 8.030575539568344,
+      "step": 8930
+    },
+    {
+      "epoch": 8.030575539568344,
+      "step": 8930,
+      "torque_loss": 0.125461146235466
+    },
+    {
+      "epoch": 8.030575539568344,
+      "force_loss": 0.010436777025461197,
+      "step": 8930
+    },
+    {
+      "epoch": 8.03956834532374,
+      "grad_norm": 0.6063098907470703,
+      "learning_rate": 9.734517758681712e-05,
+      "loss": 0.0295,
+      "step": 8940
+    },
+    {
+      "action_loss": 0.012668609619140625,
+      "epoch": 8.03956834532374,
+      "step": 8940
+    },
+    {
+      "epoch": 8.03956834532374,
+      "step": 8940,
+      "torque_loss": 0.11099114269018173
+    },
+    {
+      "epoch": 8.03956834532374,
+      "force_loss": 0.015590664930641651,
+      "step": 8940
+    },
+    {
+      "epoch": 8.048561151079136,
+      "grad_norm": 1.1803076267242432,
+      "learning_rate": 9.733631006456088e-05,
+      "loss": 0.0288,
+      "step": 8950
+    },
+    {
+      "action_loss": 0.016963742673397064,
+      "epoch": 8.048561151079136,
+      "step": 8950
+    },
+    {
+      "epoch": 8.048561151079136,
+      "step": 8950,
+      "torque_loss": 0.17014066874980927
+    },
+    {
+      "epoch": 8.048561151079136,
+      "force_loss": 0.020709695294499397,
+      "step": 8950
+    },
+    {
+      "epoch": 8.057553956834532,
+      "grad_norm": 0.5683021545410156,
+      "learning_rate": 9.732742816278348e-05,
+      "loss": 0.0321,
+      "step": 8960
+    },
+    {
+      "action_loss": 0.01965462975203991,
+      "epoch": 8.057553956834532,
+      "step": 8960
+    },
+    {
+      "epoch": 8.057553956834532,
+      "step": 8960,
+      "torque_loss": 0.14173145592212677
+    },
+    {
+      "epoch": 8.057553956834532,
+      "force_loss": 0.026922591030597687,
+      "step": 8960
+    },
+    {
+      "epoch": 8.066546762589928,
+      "grad_norm": 0.870956540107727,
+      "learning_rate": 9.731853188418302e-05,
+      "loss": 0.0278,
+      "step": 8970
+    },
+    {
+      "action_loss": 0.04899698495864868,
+      "epoch": 8.066546762589928,
+      "step": 8970
+    },
+    {
+      "epoch": 8.066546762589928,
+      "step": 8970,
+      "torque_loss": 0.18799400329589844
+    },
+    {
+      "epoch": 8.066546762589928,
+      "force_loss": 0.05203847959637642,
+      "step": 8970
+    },
+    {
+      "epoch": 8.075539568345324,
+      "grad_norm": 0.8082453012466431,
+      "learning_rate": 9.730962123146194e-05,
+      "loss": 0.0345,
+      "step": 8980
+    },
+    {
+      "action_loss": 0.008148455992341042,
+      "epoch": 8.075539568345324,
+      "step": 8980
+    },
+    {
+      "epoch": 8.075539568345324,
+      "step": 8980,
+      "torque_loss": 0.13181421160697937
+    },
+    {
+      "epoch": 8.075539568345324,
+      "force_loss": 0.010034661740064621,
+      "step": 8980
+    },
+    {
+      "epoch": 8.08453237410072,
+      "grad_norm": 0.6588174700737,
+      "learning_rate": 9.730069620732709e-05,
+      "loss": 0.0326,
+      "step": 8990
+    },
+    {
+      "action_loss": 0.01672670803964138,
+      "epoch": 8.08453237410072,
+      "step": 8990
+    },
+    {
+      "epoch": 8.08453237410072,
+      "step": 8990,
+      "torque_loss": 0.12110227346420288
+    },
+    {
+      "epoch": 8.08453237410072,
+      "force_loss": 0.014961506240069866,
+      "step": 8990
+    },
+    {
+      "epoch": 8.093525179856115,
+      "grad_norm": 0.8046431541442871,
+      "learning_rate": 9.72917568144896e-05,
+      "loss": 0.0301,
+      "step": 9000
+    },
+    {
+      "action_loss": 0.016531439498066902,
+      "epoch": 8.093525179856115,
+      "step": 9000
+    },
+    {
+      "epoch": 8.093525179856115,
+      "step": 9000,
+      "torque_loss": 0.16104300320148468
+    },
+    {
+      "epoch": 8.093525179856115,
+      "force_loss": 0.011993743479251862,
+      "step": 9000
+    },
+    {
+      "epoch": 8.102517985611511,
+      "grad_norm": 1.2161035537719727,
+      "learning_rate": 9.728280305566509e-05,
+      "loss": 0.0358,
+      "step": 9010
+    },
+    {
+      "action_loss": 0.02981303632259369,
+      "epoch": 8.102517985611511,
+      "step": 9010
+    },
+    {
+      "epoch": 8.102517985611511,
+      "step": 9010,
+      "torque_loss": 0.1625359207391739
+    },
+    {
+      "epoch": 8.102517985611511,
+      "force_loss": 0.03526288643479347,
+      "step": 9010
+    },
+    {
+      "epoch": 8.111510791366907,
+      "grad_norm": 0.5446035861968994,
+      "learning_rate": 9.727383493357343e-05,
+      "loss": 0.0346,
+      "step": 9020
+    },
+    {
+      "action_loss": 0.018178652971982956,
+      "epoch": 8.111510791366907,
+      "step": 9020
+    },
+    {
+      "epoch": 8.111510791366907,
+      "step": 9020,
+      "torque_loss": 0.1328810602426529
+    },
+    {
+      "epoch": 8.111510791366907,
+      "force_loss": 0.016255365684628487,
+      "step": 9020
+    },
+    {
+      "epoch": 8.120503597122303,
+      "grad_norm": 0.7841731309890747,
+      "learning_rate": 9.726485245093891e-05,
+      "loss": 0.0343,
+      "step": 9030
+    },
+    {
+      "action_loss": 0.01261367741972208,
+      "epoch": 8.120503597122303,
+      "step": 9030
+    },
+    {
+      "epoch": 8.120503597122303,
+      "step": 9030,
+      "torque_loss": 0.1043577790260315
+    },
+    {
+      "epoch": 8.120503597122303,
+      "force_loss": 0.009312078356742859,
+      "step": 9030
+    },
+    {
+      "epoch": 8.129496402877697,
+      "grad_norm": 1.0097514390945435,
+      "learning_rate": 9.725585561049018e-05,
+      "loss": 0.0296,
+      "step": 9040
+    },
+    {
+      "action_loss": 0.032424528151750565,
+      "epoch": 8.129496402877697,
+      "step": 9040
+    },
+    {
+      "epoch": 8.129496402877697,
+      "step": 9040,
+      "torque_loss": 0.1894434541463852
+    },
+    {
+      "epoch": 8.129496402877697,
+      "force_loss": 0.03689010068774223,
+      "step": 9040
+    },
+    {
+      "epoch": 8.138489208633093,
+      "grad_norm": 0.6864043474197388,
+      "learning_rate": 9.724684441496022e-05,
+      "loss": 0.0326,
+      "step": 9050
+    },
+    {
+      "action_loss": 0.012702490203082561,
+      "epoch": 8.138489208633093,
+      "step": 9050
+    },
+    {
+      "epoch": 8.138489208633093,
+      "step": 9050,
+      "torque_loss": 0.08355443924665451
+    },
+    {
+      "epoch": 8.138489208633093,
+      "force_loss": 0.008524672128260136,
+      "step": 9050
+    },
+    {
+      "epoch": 8.147482014388489,
+      "grad_norm": 1.1375627517700195,
+      "learning_rate": 9.72378188670864e-05,
+      "loss": 0.0281,
+      "step": 9060
+    },
+    {
+      "action_loss": 0.007547351997345686,
+      "epoch": 8.147482014388489,
+      "step": 9060
+    },
+    {
+      "epoch": 8.147482014388489,
+      "step": 9060,
+      "torque_loss": 0.11102074384689331
+    },
+    {
+      "epoch": 8.147482014388489,
+      "force_loss": 0.006551103666424751,
+      "step": 9060
+    },
+    {
+      "epoch": 8.156474820143885,
+      "grad_norm": 0.4873935580253601,
+      "learning_rate": 9.722877896961047e-05,
+      "loss": 0.0254,
+      "step": 9070
+    },
+    {
+      "action_loss": 0.011491470970213413,
+      "epoch": 8.156474820143885,
+      "step": 9070
+    },
+    {
+      "epoch": 8.156474820143885,
+      "step": 9070,
+      "torque_loss": 0.1838628202676773
+    },
+    {
+      "epoch": 8.156474820143885,
+      "force_loss": 0.00972656812518835,
+      "step": 9070
+    },
+    {
+      "epoch": 8.16546762589928,
+      "grad_norm": 0.47592806816101074,
+      "learning_rate": 9.721972472527848e-05,
+      "loss": 0.0289,
+      "step": 9080
+    },
+    {
+      "action_loss": 0.010076670907437801,
+      "epoch": 8.16546762589928,
+      "step": 9080
+    },
+    {
+      "epoch": 8.16546762589928,
+      "step": 9080,
+      "torque_loss": 0.0921001061797142
+    },
+    {
+      "epoch": 8.16546762589928,
+      "force_loss": 0.01084989309310913,
+      "step": 9080
+    },
+    {
+      "epoch": 8.174460431654676,
+      "grad_norm": 0.6903480887413025,
+      "learning_rate": 9.721065613684089e-05,
+      "loss": 0.026,
+      "step": 9090
+    },
+    {
+      "action_loss": 0.009710323996841908,
+      "epoch": 8.174460431654676,
+      "step": 9090
+    },
+    {
+      "epoch": 8.174460431654676,
+      "step": 9090,
+      "torque_loss": 0.09566020965576172
+    },
+    {
+      "epoch": 8.174460431654676,
+      "force_loss": 0.013454924337565899,
+      "step": 9090
+    },
+    {
+      "epoch": 8.183453237410072,
+      "grad_norm": 0.47146815061569214,
+      "learning_rate": 9.72015732070525e-05,
+      "loss": 0.0252,
+      "step": 9100
+    },
+    {
+      "action_loss": 0.025996504351496696,
+      "epoch": 8.183453237410072,
+      "step": 9100
+    },
+    {
+      "epoch": 8.183453237410072,
+      "step": 9100,
+      "torque_loss": 0.1910381317138672
+    },
+    {
+      "epoch": 8.183453237410072,
+      "force_loss": 0.02081509865820408,
+      "step": 9100
+    },
+    {
+      "epoch": 8.192446043165468,
+      "grad_norm": 1.7140965461730957,
+      "learning_rate": 9.719247593867244e-05,
+      "loss": 0.0268,
+      "step": 9110
+    },
+    {
+      "action_loss": 0.018666816875338554,
+      "epoch": 8.192446043165468,
+      "step": 9110
+    },
+    {
+      "epoch": 8.192446043165468,
+      "step": 9110,
+      "torque_loss": 0.1385652869939804
+    },
+    {
+      "epoch": 8.192446043165468,
+      "force_loss": 0.016636190935969353,
+      "step": 9110
+    },
+    {
+      "epoch": 8.201438848920864,
+      "grad_norm": 0.647611141204834,
+      "learning_rate": 9.718336433446423e-05,
+      "loss": 0.032,
+      "step": 9120
+    },
+    {
+      "action_loss": 0.011723405681550503,
+      "epoch": 8.201438848920864,
+      "step": 9120
+    },
+    {
+      "epoch": 8.201438848920864,
+      "step": 9120,
+      "torque_loss": 0.15605856478214264
+    },
+    {
+      "epoch": 8.201438848920864,
+      "force_loss": 0.01640763320028782,
+      "step": 9120
+    },
+    {
+      "epoch": 8.21043165467626,
+      "grad_norm": 0.9738593101501465,
+      "learning_rate": 9.717423839719574e-05,
+      "loss": 0.028,
+      "step": 9130
+    },
+    {
+      "action_loss": 0.016471194103360176,
+      "epoch": 8.21043165467626,
+      "step": 9130
+    },
+    {
+      "epoch": 8.21043165467626,
+      "step": 9130,
+      "torque_loss": 0.16527016460895538
+    },
+    {
+      "epoch": 8.21043165467626,
+      "force_loss": 0.019555814564228058,
+      "step": 9130
+    },
+    {
+      "epoch": 8.219424460431656,
+      "grad_norm": 0.6147652864456177,
+      "learning_rate": 9.71650981296392e-05,
+      "loss": 0.0283,
+      "step": 9140
+    },
+    {
+      "action_loss": 0.017719784751534462,
+      "epoch": 8.219424460431656,
+      "step": 9140
+    },
+    {
+      "epoch": 8.219424460431656,
+      "step": 9140,
+      "torque_loss": 0.12981177866458893
+    },
+    {
+      "epoch": 8.219424460431656,
+      "force_loss": 0.017089011147618294,
+      "step": 9140
+    },
+    {
+      "epoch": 8.22841726618705,
+      "grad_norm": 0.4535752534866333,
+      "learning_rate": 9.715594353457118e-05,
+      "loss": 0.0266,
+      "step": 9150
+    },
+    {
+      "action_loss": 0.00867116916924715,
+      "epoch": 8.22841726618705,
+      "step": 9150
+    },
+    {
+      "epoch": 8.22841726618705,
+      "step": 9150,
+      "torque_loss": 0.09822756052017212
+    },
+    {
+      "epoch": 8.22841726618705,
+      "force_loss": 0.01133808959275484,
+      "step": 9150
+    },
+    {
+      "epoch": 8.237410071942445,
+      "grad_norm": 0.813336968421936,
+      "learning_rate": 9.714677461477257e-05,
+      "loss": 0.0299,
+      "step": 9160
+    },
+    {
+      "action_loss": 0.015352415852248669,
+      "epoch": 8.237410071942445,
+      "step": 9160
+    },
+    {
+      "epoch": 8.237410071942445,
+      "step": 9160,
+      "torque_loss": 0.14493830502033234
+    },
+    {
+      "epoch": 8.237410071942445,
+      "force_loss": 0.007876926101744175,
+      "step": 9160
+    },
+    {
+      "epoch": 8.246402877697841,
+      "grad_norm": 0.5332842469215393,
+      "learning_rate": 9.713759137302869e-05,
+      "loss": 0.0279,
+      "step": 9170
+    },
+    {
+      "action_loss": 0.020986506715416908,
+      "epoch": 8.246402877697841,
+      "step": 9170
+    },
+    {
+      "epoch": 8.246402877697841,
+      "step": 9170,
+      "torque_loss": 0.13961315155029297
+    },
+    {
+      "epoch": 8.246402877697841,
+      "force_loss": 0.016814392060041428,
+      "step": 9170
+    },
+    {
+      "epoch": 8.255395683453237,
+      "grad_norm": 1.2883217334747314,
+      "learning_rate": 9.712839381212914e-05,
+      "loss": 0.0294,
+      "step": 9180
+    },
+    {
+      "action_loss": 0.012940999120473862,
+      "epoch": 8.255395683453237,
+      "step": 9180
+    },
+    {
+      "epoch": 8.255395683453237,
+      "step": 9180,
+      "torque_loss": 0.1373448222875595
+    },
+    {
+      "epoch": 8.255395683453237,
+      "force_loss": 0.008718389086425304,
+      "step": 9180
+    },
+    {
+      "epoch": 8.264388489208633,
+      "grad_norm": 0.7358983159065247,
+      "learning_rate": 9.71191819348679e-05,
+      "loss": 0.0285,
+      "step": 9190
+    },
+    {
+      "action_loss": 0.00819359626621008,
+      "epoch": 8.264388489208633,
+      "step": 9190
+    },
+    {
+      "epoch": 8.264388489208633,
+      "step": 9190,
+      "torque_loss": 0.0880383625626564
+    },
+    {
+      "epoch": 8.264388489208633,
+      "force_loss": 0.006998899858444929,
+      "step": 9190
+    },
+    {
+      "epoch": 8.273381294964029,
+      "grad_norm": 0.4908333718776703,
+      "learning_rate": 9.710995574404331e-05,
+      "loss": 0.0295,
+      "step": 9200
+    },
+    {
+      "action_loss": 0.009041265584528446,
+      "epoch": 8.273381294964029,
+      "step": 9200
+    },
+    {
+      "epoch": 8.273381294964029,
+      "step": 9200,
+      "torque_loss": 0.12641453742980957
+    },
+    {
+      "epoch": 8.273381294964029,
+      "force_loss": 0.008404488675296307,
+      "step": 9200
+    },
+    {
+      "epoch": 8.282374100719425,
+      "grad_norm": 1.0569324493408203,
+      "learning_rate": 9.710071524245802e-05,
+      "loss": 0.0344,
+      "step": 9210
+    },
+    {
+      "action_loss": 0.023268500342965126,
+      "epoch": 8.282374100719425,
+      "step": 9210
+    },
+    {
+      "epoch": 8.282374100719425,
+      "step": 9210,
+      "torque_loss": 0.14671923220157623
+    },
+    {
+      "epoch": 8.282374100719425,
+      "force_loss": 0.02295270748436451,
+      "step": 9210
+    },
+    {
+      "epoch": 8.29136690647482,
+      "grad_norm": 0.43037769198417664,
+      "learning_rate": 9.709146043291906e-05,
+      "loss": 0.0258,
+      "step": 9220
+    },
+    {
+      "action_loss": 0.011890962719917297,
+      "epoch": 8.29136690647482,
+      "step": 9220
+    },
+    {
+      "epoch": 8.29136690647482,
+      "step": 9220,
+      "torque_loss": 0.1297449916601181
+    },
+    {
+      "epoch": 8.29136690647482,
+      "force_loss": 0.008125021122395992,
+      "step": 9220
+    },
+    {
+      "epoch": 8.300359712230216,
+      "grad_norm": 1.2475656270980835,
+      "learning_rate": 9.70821913182378e-05,
+      "loss": 0.0323,
+      "step": 9230
+    },
+    {
+      "action_loss": 0.01498037576675415,
+      "epoch": 8.300359712230216,
+      "step": 9230
+    },
+    {
+      "epoch": 8.300359712230216,
+      "step": 9230,
+      "torque_loss": 0.16476303339004517
+    },
+    {
+      "epoch": 8.300359712230216,
+      "force_loss": 0.01710757426917553,
+      "step": 9230
+    },
+    {
+      "epoch": 8.309352517985612,
+      "grad_norm": 0.6913616061210632,
+      "learning_rate": 9.707290790122995e-05,
+      "loss": 0.0324,
+      "step": 9240
+    },
+    {
+      "action_loss": 0.008063911460340023,
+      "epoch": 8.309352517985612,
+      "step": 9240
+    },
+    {
+      "epoch": 8.309352517985612,
+      "step": 9240,
+      "torque_loss": 0.09638575464487076
+    },
+    {
+      "epoch": 8.309352517985612,
+      "force_loss": 0.005585936829447746,
+      "step": 9240
+    },
+    {
+      "epoch": 8.318345323741006,
+      "grad_norm": 0.6526975035667419,
+      "learning_rate": 9.706361018471557e-05,
+      "loss": 0.0313,
+      "step": 9250
+    },
+    {
+      "action_loss": 0.022348126396536827,
+      "epoch": 8.318345323741006,
+      "step": 9250
+    },
+    {
+      "epoch": 8.318345323741006,
+      "step": 9250,
+      "torque_loss": 0.13248980045318604
+    },
+    {
+      "epoch": 8.318345323741006,
+      "force_loss": 0.02697359211742878,
+      "step": 9250
+    },
+    {
+      "epoch": 8.327338129496402,
+      "grad_norm": 0.947041928768158,
+      "learning_rate": 9.705429817151906e-05,
+      "loss": 0.0307,
+      "step": 9260
+    },
+    {
+      "action_loss": 0.01929389126598835,
+      "epoch": 8.327338129496402,
+      "step": 9260
+    },
+    {
+      "epoch": 8.327338129496402,
+      "step": 9260,
+      "torque_loss": 0.13584868609905243
+    },
+    {
+      "epoch": 8.327338129496402,
+      "force_loss": 0.028136268258094788,
+      "step": 9260
+    },
+    {
+      "epoch": 8.336330935251798,
+      "grad_norm": 0.9014036059379578,
+      "learning_rate": 9.704497186446917e-05,
+      "loss": 0.0276,
+      "step": 9270
+    },
+    {
+      "action_loss": 0.01147465780377388,
+      "epoch": 8.336330935251798,
+      "step": 9270
+    },
+    {
+      "epoch": 8.336330935251798,
+      "step": 9270,
+      "torque_loss": 0.12540923058986664
+    },
+    {
+      "epoch": 8.336330935251798,
+      "force_loss": 0.016301970928907394,
+      "step": 9270
+    },
+    {
+      "epoch": 8.345323741007194,
+      "grad_norm": 0.28941667079925537,
+      "learning_rate": 9.703563126639896e-05,
+      "loss": 0.033,
+      "step": 9280
+    },
+    {
+      "action_loss": 0.010872160084545612,
+      "epoch": 8.345323741007194,
+      "step": 9280
+    },
+    {
+      "epoch": 8.345323741007194,
+      "step": 9280,
+      "torque_loss": 0.13422954082489014
+    },
+    {
+      "epoch": 8.345323741007194,
+      "force_loss": 0.015045229345560074,
+      "step": 9280
+    },
+    {
+      "epoch": 8.35431654676259,
+      "grad_norm": 0.4963284730911255,
+      "learning_rate": 9.70262763801459e-05,
+      "loss": 0.0287,
+      "step": 9290
+    },
+    {
+      "action_loss": 0.020897654816508293,
+      "epoch": 8.35431654676259,
+      "step": 9290
+    },
+    {
+      "epoch": 8.35431654676259,
+      "step": 9290,
+      "torque_loss": 0.12592101097106934
+    },
+    {
+      "epoch": 8.35431654676259,
+      "force_loss": 0.014290876686573029,
+      "step": 9290
+    },
+    {
+      "epoch": 8.363309352517986,
+      "grad_norm": 0.9091506004333496,
+      "learning_rate": 9.701690720855171e-05,
+      "loss": 0.0335,
+      "step": 9300
+    },
+    {
+      "action_loss": 0.004191944841295481,
+      "epoch": 8.363309352517986,
+      "step": 9300
+    },
+    {
+      "epoch": 8.363309352517986,
+      "step": 9300,
+      "torque_loss": 0.06838779896497726
+    },
+    {
+      "epoch": 8.363309352517986,
+      "force_loss": 0.00599181791767478,
+      "step": 9300
+    },
+    {
+      "epoch": 8.372302158273381,
+      "grad_norm": 0.46863675117492676,
+      "learning_rate": 9.700752375446253e-05,
+      "loss": 0.0299,
+      "step": 9310
+    },
+    {
+      "action_loss": 0.0628635510802269,
+      "epoch": 8.372302158273381,
+      "step": 9310
+    },
+    {
+      "epoch": 8.372302158273381,
+      "step": 9310,
+      "torque_loss": 0.16405658423900604
+    },
+    {
+      "epoch": 8.372302158273381,
+      "force_loss": 0.05716918036341667,
+      "step": 9310
+    },
+    {
+      "epoch": 8.381294964028777,
+      "grad_norm": 1.1366662979125977,
+      "learning_rate": 9.69981260207288e-05,
+      "loss": 0.0414,
+      "step": 9320
+    },
+    {
+      "action_loss": 0.017923591658473015,
+      "epoch": 8.381294964028777,
+      "step": 9320
+    },
+    {
+      "epoch": 8.381294964028777,
+      "step": 9320,
+      "torque_loss": 0.22768175601959229
+    },
+    {
+      "epoch": 8.381294964028777,
+      "force_loss": 0.01648772321641445,
+      "step": 9320
+    },
+    {
+      "epoch": 8.390287769784173,
+      "grad_norm": 0.7111708521842957,
+      "learning_rate": 9.698871401020529e-05,
+      "loss": 0.0335,
+      "step": 9330
+    },
+    {
+      "action_loss": 0.007409380748867989,
+      "epoch": 8.390287769784173,
+      "step": 9330
+    },
+    {
+      "epoch": 8.390287769784173,
+      "step": 9330,
+      "torque_loss": 0.1321495622396469
+    },
+    {
+      "epoch": 8.390287769784173,
+      "force_loss": 0.007271577138453722,
+      "step": 9330
+    },
+    {
+      "epoch": 8.399280575539569,
+      "grad_norm": 0.9520056247711182,
+      "learning_rate": 9.697928772575112e-05,
+      "loss": 0.0309,
+      "step": 9340
+    },
+    {
+      "action_loss": 0.015844479203224182,
+      "epoch": 8.399280575539569,
+      "step": 9340
+    },
+    {
+      "epoch": 8.399280575539569,
+      "step": 9340,
+      "torque_loss": 0.15661227703094482
+    },
+    {
+      "epoch": 8.399280575539569,
+      "force_loss": 0.017089197412133217,
+      "step": 9340
+    },
+    {
+      "epoch": 8.408273381294965,
+      "grad_norm": 0.6098338961601257,
+      "learning_rate": 9.696984717022976e-05,
+      "loss": 0.0255,
+      "step": 9350
+    },
+    {
+      "action_loss": 0.005137607455253601,
+      "epoch": 8.408273381294965,
+      "step": 9350
+    },
+    {
+      "epoch": 8.408273381294965,
+      "step": 9350,
+      "torque_loss": 0.115314781665802
+    },
+    {
+      "epoch": 8.408273381294965,
+      "force_loss": 0.005152374505996704,
+      "step": 9350
+    },
+    {
+      "epoch": 8.417266187050359,
+      "grad_norm": 0.7887623906135559,
+      "learning_rate": 9.6960392346509e-05,
+      "loss": 0.026,
+      "step": 9360
+    },
+    {
+      "action_loss": 0.00956383440643549,
+      "epoch": 8.417266187050359,
+      "step": 9360
+    },
+    {
+      "epoch": 8.417266187050359,
+      "step": 9360,
+      "torque_loss": 0.11685600876808167
+    },
+    {
+      "epoch": 8.417266187050359,
+      "force_loss": 0.00842523854225874,
+      "step": 9360
+    },
+    {
+      "epoch": 8.426258992805755,
+      "grad_norm": 0.38678234815597534,
+      "learning_rate": 9.695092325746097e-05,
+      "loss": 0.0329,
+      "step": 9370
+    },
+    {
+      "action_loss": 0.005833383649587631,
+      "epoch": 8.426258992805755,
+      "step": 9370
+    },
+    {
+      "epoch": 8.426258992805755,
+      "step": 9370,
+      "torque_loss": 0.11251125484704971
+    },
+    {
+      "epoch": 8.426258992805755,
+      "force_loss": 0.006181512027978897,
+      "step": 9370
+    },
+    {
+      "epoch": 8.43525179856115,
+      "grad_norm": 0.8431825637817383,
+      "learning_rate": 9.694143990596211e-05,
+      "loss": 0.0259,
+      "step": 9380
+    },
+    {
+      "action_loss": 0.010571137070655823,
+      "epoch": 8.43525179856115,
+      "step": 9380
+    },
+    {
+      "epoch": 8.43525179856115,
+      "step": 9380,
+      "torque_loss": 0.1533479243516922
+    },
+    {
+      "epoch": 8.43525179856115,
+      "force_loss": 0.012814153917133808,
+      "step": 9380
+    },
+    {
+      "epoch": 8.444244604316546,
+      "grad_norm": 0.7134740948677063,
+      "learning_rate": 9.693194229489325e-05,
+      "loss": 0.0425,
+      "step": 9390
+    },
+    {
+      "action_loss": 0.011143607087433338,
+      "epoch": 8.444244604316546,
+      "step": 9390
+    },
+    {
+      "epoch": 8.444244604316546,
+      "step": 9390,
+      "torque_loss": 0.1590392142534256
+    },
+    {
+      "epoch": 8.444244604316546,
+      "force_loss": 0.017856024205684662,
+      "step": 9390
+    },
+    {
+      "epoch": 8.453237410071942,
+      "grad_norm": 1.0953168869018555,
+      "learning_rate": 9.692243042713944e-05,
+      "loss": 0.0289,
+      "step": 9400
+    },
+    {
+      "action_loss": 0.014227758161723614,
+      "epoch": 8.453237410071942,
+      "step": 9400
+    },
+    {
+      "epoch": 8.453237410071942,
+      "step": 9400,
+      "torque_loss": 0.12394676357507706
+    },
+    {
+      "epoch": 8.453237410071942,
+      "force_loss": 0.014927543699741364,
+      "step": 9400
+    },
+    {
+      "epoch": 8.462230215827338,
+      "grad_norm": 1.3834631443023682,
+      "learning_rate": 9.691290430559022e-05,
+      "loss": 0.028,
+      "step": 9410
+    },
+    {
+      "action_loss": 0.012397505342960358,
+      "epoch": 8.462230215827338,
+      "step": 9410
+    },
+    {
+      "epoch": 8.462230215827338,
+      "step": 9410,
+      "torque_loss": 0.18198783695697784
+    },
+    {
+      "epoch": 8.462230215827338,
+      "force_loss": 0.007384064141660929,
+      "step": 9410
+    },
+    {
+      "epoch": 8.471223021582734,
+      "grad_norm": 0.7003529667854309,
+      "learning_rate": 9.690336393313932e-05,
+      "loss": 0.0275,
+      "step": 9420
+    },
+    {
+      "action_loss": 0.009323895908892155,
+      "epoch": 8.471223021582734,
+      "step": 9420
+    },
+    {
+      "epoch": 8.471223021582734,
+      "step": 9420,
+      "torque_loss": 0.12896889448165894
+    },
+    {
+      "epoch": 8.471223021582734,
+      "force_loss": 0.0062034521251916885,
+      "step": 9420
+    },
+    {
+      "epoch": 8.48021582733813,
+      "grad_norm": 0.6547080874443054,
+      "learning_rate": 9.689380931268487e-05,
+      "loss": 0.031,
+      "step": 9430
+    },
+    {
+      "action_loss": 0.01772930659353733,
+      "epoch": 8.48021582733813,
+      "step": 9430
+    },
+    {
+      "epoch": 8.48021582733813,
+      "step": 9430,
+      "torque_loss": 0.1869361847639084
+    },
+    {
+      "epoch": 8.48021582733813,
+      "force_loss": 0.01642507128417492,
+      "step": 9430
+    },
+    {
+      "epoch": 8.489208633093526,
+      "grad_norm": 0.8937145471572876,
+      "learning_rate": 9.688424044712932e-05,
+      "loss": 0.0316,
+      "step": 9440
+    },
+    {
+      "action_loss": 0.009877237491309643,
+      "epoch": 8.489208633093526,
+      "step": 9440
+    },
+    {
+      "epoch": 8.489208633093526,
+      "step": 9440,
+      "torque_loss": 0.157605841755867
+    },
+    {
+      "epoch": 8.489208633093526,
+      "force_loss": 0.011405526660382748,
+      "step": 9440
+    },
+    {
+      "epoch": 8.498201438848922,
+      "grad_norm": 0.6158521771430969,
+      "learning_rate": 9.687465733937942e-05,
+      "loss": 0.0358,
+      "step": 9450
+    },
+    {
+      "action_loss": 0.0104767931625247,
+      "epoch": 8.498201438848922,
+      "step": 9450
+    },
+    {
+      "epoch": 8.498201438848922,
+      "step": 9450,
+      "torque_loss": 0.10644660145044327
+    },
+    {
+      "epoch": 8.498201438848922,
+      "force_loss": 0.009099212475121021,
+      "step": 9450
+    },
+    {
+      "epoch": 8.507194244604317,
+      "grad_norm": 1.7294045686721802,
+      "learning_rate": 9.686505999234627e-05,
+      "loss": 0.028,
+      "step": 9460
+    },
+    {
+      "action_loss": 0.007561964448541403,
+      "epoch": 8.507194244604317,
+      "step": 9460
+    },
+    {
+      "epoch": 8.507194244604317,
+      "step": 9460,
+      "torque_loss": 0.17132334411144257
+    },
+    {
+      "epoch": 8.507194244604317,
+      "force_loss": 0.009764394722878933,
+      "step": 9460
+    },
+    {
+      "epoch": 8.516187050359711,
+      "grad_norm": 0.8151748180389404,
+      "learning_rate": 9.685544840894529e-05,
+      "loss": 0.0272,
+      "step": 9470
+    },
+    {
+      "action_loss": 0.007332219276577234,
+      "epoch": 8.516187050359711,
+      "step": 9470
+    },
+    {
+      "epoch": 8.516187050359711,
+      "step": 9470,
+      "torque_loss": 0.14351345598697662
+    },
+    {
+      "epoch": 8.516187050359711,
+      "force_loss": 0.012972989119589329,
+      "step": 9470
+    },
+    {
+      "epoch": 8.525179856115107,
+      "grad_norm": 0.6972848773002625,
+      "learning_rate": 9.684582259209624e-05,
+      "loss": 0.0271,
+      "step": 9480
+    },
+    {
+      "action_loss": 0.009656401351094246,
+      "epoch": 8.525179856115107,
+      "step": 9480
+    },
+    {
+      "epoch": 8.525179856115107,
+      "step": 9480,
+      "torque_loss": 0.15328960120677948
+    },
+    {
+      "epoch": 8.525179856115107,
+      "force_loss": 0.012785851024091244,
+      "step": 9480
+    },
+    {
+      "epoch": 8.534172661870503,
+      "grad_norm": 0.5664857029914856,
+      "learning_rate": 9.683618254472317e-05,
+      "loss": 0.0277,
+      "step": 9490
+    },
+    {
+      "action_loss": 0.007961034774780273,
+      "epoch": 8.534172661870503,
+      "step": 9490
+    },
+    {
+      "epoch": 8.534172661870503,
+      "step": 9490,
+      "torque_loss": 0.15656043589115143
+    },
+    {
+      "epoch": 8.534172661870503,
+      "force_loss": 0.007319868076592684,
+      "step": 9490
+    },
+    {
+      "epoch": 8.543165467625899,
+      "grad_norm": 0.8491553068161011,
+      "learning_rate": 9.682652826975449e-05,
+      "loss": 0.0291,
+      "step": 9500
+    },
+    {
+      "action_loss": 0.019876837730407715,
+      "epoch": 8.543165467625899,
+      "step": 9500
+    },
+    {
+      "epoch": 8.543165467625899,
+      "step": 9500,
+      "torque_loss": 0.11075549572706223
+    },
+    {
+      "epoch": 8.543165467625899,
+      "force_loss": 0.02449529431760311,
+      "step": 9500
+    },
+    {
+      "epoch": 8.552158273381295,
+      "grad_norm": 0.7463903427124023,
+      "learning_rate": 9.681685977012291e-05,
+      "loss": 0.027,
+      "step": 9510
+    },
+    {
+      "action_loss": 0.008659898303449154,
+      "epoch": 8.552158273381295,
+      "step": 9510
+    },
+    {
+      "epoch": 8.552158273381295,
+      "step": 9510,
+      "torque_loss": 0.10944860428571701
+    },
+    {
+      "epoch": 8.552158273381295,
+      "force_loss": 0.008656457997858524,
+      "step": 9510
+    },
+    {
+      "epoch": 8.56115107913669,
+      "grad_norm": 0.7965874671936035,
+      "learning_rate": 9.680717704876546e-05,
+      "loss": 0.0302,
+      "step": 9520
+    },
+    {
+      "action_loss": 0.009310754016041756,
+      "epoch": 8.56115107913669,
+      "step": 9520
+    },
+    {
+      "epoch": 8.56115107913669,
+      "step": 9520,
+      "torque_loss": 0.1385524719953537
+    },
+    {
+      "epoch": 8.56115107913669,
+      "force_loss": 0.008410467766225338,
+      "step": 9520
+    },
+    {
+      "epoch": 8.570143884892087,
+      "grad_norm": 0.6976439952850342,
+      "learning_rate": 9.679748010862349e-05,
+      "loss": 0.0294,
+      "step": 9530
+    },
+    {
+      "action_loss": 0.013744421303272247,
+      "epoch": 8.570143884892087,
+      "step": 9530
+    },
+    {
+      "epoch": 8.570143884892087,
+      "step": 9530,
+      "torque_loss": 0.11806178838014603
+    },
+    {
+      "epoch": 8.570143884892087,
+      "force_loss": 0.01753784529864788,
+      "step": 9530
+    },
+    {
+      "epoch": 8.579136690647482,
+      "grad_norm": 0.6330084800720215,
+      "learning_rate": 9.678776895264267e-05,
+      "loss": 0.0325,
+      "step": 9540
+    },
+    {
+      "action_loss": 0.020705049857497215,
+      "epoch": 8.579136690647482,
+      "step": 9540
+    },
+    {
+      "epoch": 8.579136690647482,
+      "step": 9540,
+      "torque_loss": 0.16956502199172974
+    },
+    {
+      "epoch": 8.579136690647482,
+      "force_loss": 0.021446580067276955,
+      "step": 9540
+    },
+    {
+      "epoch": 8.588129496402878,
+      "grad_norm": 0.3010088801383972,
+      "learning_rate": 9.6778043583773e-05,
+      "loss": 0.0341,
+      "step": 9550
+    },
+    {
+      "action_loss": 0.011592283844947815,
+      "epoch": 8.588129496402878,
+      "step": 9550
+    },
+    {
+      "epoch": 8.588129496402878,
+      "step": 9550,
+      "torque_loss": 0.17431075870990753
+    },
+    {
+      "epoch": 8.588129496402878,
+      "force_loss": 0.019685033708810806,
+      "step": 9550
+    },
+    {
+      "epoch": 8.597122302158274,
+      "grad_norm": 0.3165673613548279,
+      "learning_rate": 9.67683040049688e-05,
+      "loss": 0.0259,
+      "step": 9560
+    },
+    {
+      "action_loss": 0.02787325717508793,
+      "epoch": 8.597122302158274,
+      "step": 9560
+    },
+    {
+      "epoch": 8.597122302158274,
+      "step": 9560,
+      "torque_loss": 0.16775758564472198
+    },
+    {
+      "epoch": 8.597122302158274,
+      "force_loss": 0.025492852553725243,
+      "step": 9560
+    },
+    {
+      "epoch": 8.60611510791367,
+      "grad_norm": 0.6373989582061768,
+      "learning_rate": 9.675855021918869e-05,
+      "loss": 0.031,
+      "step": 9570
+    },
+    {
+      "action_loss": 0.005025099031627178,
+      "epoch": 8.60611510791367,
+      "step": 9570
+    },
+    {
+      "epoch": 8.60611510791367,
+      "step": 9570,
+      "torque_loss": 0.14153970777988434
+    },
+    {
+      "epoch": 8.60611510791367,
+      "force_loss": 0.006119674537330866,
+      "step": 9570
+    },
+    {
+      "epoch": 8.615107913669064,
+      "grad_norm": 0.4211874008178711,
+      "learning_rate": 9.674878222939561e-05,
+      "loss": 0.0258,
+      "step": 9580
+    },
+    {
+      "action_loss": 0.0036755630280822515,
+      "epoch": 8.615107913669064,
+      "step": 9580
+    },
+    {
+      "epoch": 8.615107913669064,
+      "step": 9580,
+      "torque_loss": 0.07633873075246811
+    },
+    {
+      "epoch": 8.615107913669064,
+      "force_loss": 0.004712820518761873,
+      "step": 9580
+    },
+    {
+      "epoch": 8.62410071942446,
+      "grad_norm": 0.45486578345298767,
+      "learning_rate": 9.673900003855681e-05,
+      "loss": 0.0281,
+      "step": 9590
+    },
+    {
+      "action_loss": 0.016435489058494568,
+      "epoch": 8.62410071942446,
+      "step": 9590
+    },
+    {
+      "epoch": 8.62410071942446,
+      "step": 9590,
+      "torque_loss": 0.16094237565994263
+    },
+    {
+      "epoch": 8.62410071942446,
+      "force_loss": 0.016656536608934402,
+      "step": 9590
+    },
+    {
+      "epoch": 8.633093525179856,
+      "grad_norm": 0.5245792865753174,
+      "learning_rate": 9.672920364964389e-05,
+      "loss": 0.0359,
+      "step": 9600
+    },
+    {
+      "action_loss": 0.021780768409371376,
+      "epoch": 8.633093525179856,
+      "step": 9600
+    },
+    {
+      "epoch": 8.633093525179856,
+      "step": 9600,
+      "torque_loss": 0.18047575652599335
+    },
+    {
+      "epoch": 8.633093525179856,
+      "force_loss": 0.022455209866166115,
+      "step": 9600
+    },
+    {
+      "epoch": 8.642086330935252,
+      "grad_norm": 0.4266988933086395,
+      "learning_rate": 9.671939306563269e-05,
+      "loss": 0.025,
+      "step": 9610
+    },
+    {
+      "action_loss": 0.018251975998282433,
+      "epoch": 8.642086330935252,
+      "step": 9610
+    },
+    {
+      "epoch": 8.642086330935252,
+      "step": 9610,
+      "torque_loss": 0.1401875764131546
+    },
+    {
+      "epoch": 8.642086330935252,
+      "force_loss": 0.027092456817626953,
+      "step": 9610
+    },
+    {
+      "epoch": 8.651079136690647,
+      "grad_norm": 1.2824878692626953,
+      "learning_rate": 9.670956828950345e-05,
+      "loss": 0.0318,
+      "step": 9620
+    },
+    {
+      "action_loss": 0.011505433358252048,
+      "epoch": 8.651079136690647,
+      "step": 9620
+    },
+    {
+      "epoch": 8.651079136690647,
+      "step": 9620,
+      "torque_loss": 0.12270506471395493
+    },
+    {
+      "epoch": 8.651079136690647,
+      "force_loss": 0.013477717526257038,
+      "step": 9620
+    },
+    {
+      "epoch": 8.660071942446043,
+      "grad_norm": 0.8039194345474243,
+      "learning_rate": 9.669972932424065e-05,
+      "loss": 0.0299,
+      "step": 9630
+    },
+    {
+      "action_loss": 0.008318931795656681,
+      "epoch": 8.660071942446043,
+      "step": 9630
+    },
+    {
+      "epoch": 8.660071942446043,
+      "step": 9630,
+      "torque_loss": 0.13181714713573456
+    },
+    {
+      "epoch": 8.660071942446043,
+      "force_loss": 0.006616103928536177,
+      "step": 9630
+    },
+    {
+      "epoch": 8.66906474820144,
+      "grad_norm": 0.9702808856964111,
+      "learning_rate": 9.668987617283312e-05,
+      "loss": 0.0277,
+      "step": 9640
+    },
+    {
+      "action_loss": 0.009567559696733952,
+      "epoch": 8.66906474820144,
+      "step": 9640
+    },
+    {
+      "epoch": 8.66906474820144,
+      "step": 9640,
+      "torque_loss": 0.11925510317087173
+    },
+    {
+      "epoch": 8.66906474820144,
+      "force_loss": 0.006967818830162287,
+      "step": 9640
+    },
+    {
+      "epoch": 8.678057553956835,
+      "grad_norm": 0.4645603895187378,
+      "learning_rate": 9.668000883827397e-05,
+      "loss": 0.0305,
+      "step": 9650
+    },
+    {
+      "action_loss": 0.007450137287378311,
+      "epoch": 8.678057553956835,
+      "step": 9650
+    },
+    {
+      "epoch": 8.678057553956835,
+      "step": 9650,
+      "torque_loss": 0.12383129447698593
+    },
+    {
+      "epoch": 8.678057553956835,
+      "force_loss": 0.006323601584881544,
+      "step": 9650
+    },
+    {
+      "epoch": 8.68705035971223,
+      "grad_norm": 0.6332986354827881,
+      "learning_rate": 9.667012732356067e-05,
+      "loss": 0.0316,
+      "step": 9660
+    },
+    {
+      "action_loss": 0.0139608820900321,
+      "epoch": 8.68705035971223,
+      "step": 9660
+    },
+    {
+      "epoch": 8.68705035971223,
+      "step": 9660,
+      "torque_loss": 0.1393394023180008
+    },
+    {
+      "epoch": 8.68705035971223,
+      "force_loss": 0.014705373905599117,
+      "step": 9660
+    },
+    {
+      "epoch": 8.696043165467627,
+      "grad_norm": 1.1068260669708252,
+      "learning_rate": 9.666023163169493e-05,
+      "loss": 0.0479,
+      "step": 9670
+    },
+    {
+      "action_loss": 0.009434792213141918,
+      "epoch": 8.696043165467627,
+      "step": 9670
+    },
+    {
+      "epoch": 8.696043165467627,
+      "step": 9670,
+      "torque_loss": 0.17337889969348907
+    },
+    {
+      "epoch": 8.696043165467627,
+      "force_loss": 0.006190603133291006,
+      "step": 9670
+    },
+    {
+      "epoch": 8.70503597122302,
+      "grad_norm": 0.6589455604553223,
+      "learning_rate": 9.665032176568281e-05,
+      "loss": 0.0304,
+      "step": 9680
+    },
+    {
+      "action_loss": 0.01564081571996212,
+      "epoch": 8.70503597122302,
+      "step": 9680
+    },
+    {
+      "epoch": 8.70503597122302,
+      "step": 9680,
+      "torque_loss": 0.15311110019683838
+    },
+    {
+      "epoch": 8.70503597122302,
+      "force_loss": 0.009150044992566109,
+      "step": 9680
+    },
+    {
+      "epoch": 8.714028776978417,
+      "grad_norm": 1.3331222534179688,
+      "learning_rate": 9.664039772853469e-05,
+      "loss": 0.0324,
+      "step": 9690
+    },
+    {
+      "action_loss": 0.007561507169157267,
+      "epoch": 8.714028776978417,
+      "step": 9690
+    },
+    {
+      "epoch": 8.714028776978417,
+      "step": 9690,
+      "torque_loss": 0.05610576644539833
+    },
+    {
+      "epoch": 8.714028776978417,
+      "force_loss": 0.007908865809440613,
+      "step": 9690
+    },
+    {
+      "epoch": 8.723021582733812,
+      "grad_norm": 0.37813442945480347,
+      "learning_rate": 9.663045952326518e-05,
+      "loss": 0.027,
+      "step": 9700
+    },
+    {
+      "action_loss": 0.006428262684494257,
+      "epoch": 8.723021582733812,
+      "step": 9700
+    },
+    {
+      "epoch": 8.723021582733812,
+      "step": 9700,
+      "torque_loss": 0.1249067410826683
+    },
+    {
+      "epoch": 8.723021582733812,
+      "force_loss": 0.00746154272928834,
+      "step": 9700
+    },
+    {
+      "epoch": 8.732014388489208,
+      "grad_norm": 0.691757082939148,
+      "learning_rate": 9.662050715289328e-05,
+      "loss": 0.0266,
+      "step": 9710
+    },
+    {
+      "action_loss": 0.018051935359835625,
+      "epoch": 8.732014388489208,
+      "step": 9710
+    },
+    {
+      "epoch": 8.732014388489208,
+      "step": 9710,
+      "torque_loss": 0.17711806297302246
+    },
+    {
+      "epoch": 8.732014388489208,
+      "force_loss": 0.010310347191989422,
+      "step": 9710
+    },
+    {
+      "epoch": 8.741007194244604,
+      "grad_norm": 0.643276572227478,
+      "learning_rate": 9.661054062044226e-05,
+      "loss": 0.0345,
+      "step": 9720
+    },
+    {
+      "action_loss": 0.021279161795973778,
+      "epoch": 8.741007194244604,
+      "step": 9720
+    },
+    {
+      "epoch": 8.741007194244604,
+      "step": 9720,
+      "torque_loss": 0.14824149012565613
+    },
+    {
+      "epoch": 8.741007194244604,
+      "force_loss": 0.014712401665747166,
+      "step": 9720
+    },
+    {
+      "epoch": 8.75,
+      "grad_norm": 0.48943644762039185,
+      "learning_rate": 9.660055992893968e-05,
+      "loss": 0.0282,
+      "step": 9730
+    },
+    {
+      "action_loss": 0.019215544685721397,
+      "epoch": 8.75,
+      "step": 9730
+    },
+    {
+      "epoch": 8.75,
+      "step": 9730,
+      "torque_loss": 0.16288113594055176
+    },
+    {
+      "epoch": 8.75,
+      "force_loss": 0.01412635762244463,
+      "step": 9730
+    },
+    {
+      "epoch": 8.758992805755396,
+      "grad_norm": 0.46070775389671326,
+      "learning_rate": 9.659056508141739e-05,
+      "loss": 0.0286,
+      "step": 9740
+    },
+    {
+      "action_loss": 0.014236348681151867,
+      "epoch": 8.758992805755396,
+      "step": 9740
+    },
+    {
+      "epoch": 8.758992805755396,
+      "step": 9740,
+      "torque_loss": 0.12780122458934784
+    },
+    {
+      "epoch": 8.758992805755396,
+      "force_loss": 0.01821557618677616,
+      "step": 9740
+    },
+    {
+      "epoch": 8.767985611510792,
+      "grad_norm": 0.5840524435043335,
+      "learning_rate": 9.658055608091161e-05,
+      "loss": 0.0346,
+      "step": 9750
+    },
+    {
+      "action_loss": 0.011465773917734623,
+      "epoch": 8.767985611510792,
+      "step": 9750
+    },
+    {
+      "epoch": 8.767985611510792,
+      "step": 9750,
+      "torque_loss": 0.17164932191371918
+    },
+    {
+      "epoch": 8.767985611510792,
+      "force_loss": 0.008555716834962368,
+      "step": 9750
+    },
+    {
+      "epoch": 8.776978417266188,
+      "grad_norm": 1.5898200273513794,
+      "learning_rate": 9.657053293046276e-05,
+      "loss": 0.0311,
+      "step": 9760
+    },
+    {
+      "action_loss": 0.010616987943649292,
+      "epoch": 8.776978417266188,
+      "step": 9760
+    },
+    {
+      "epoch": 8.776978417266188,
+      "step": 9760,
+      "torque_loss": 0.12834854423999786
+    },
+    {
+      "epoch": 8.776978417266188,
+      "force_loss": 0.012776236049830914,
+      "step": 9760
+    },
+    {
+      "epoch": 8.785971223021583,
+      "grad_norm": 1.011033296585083,
+      "learning_rate": 9.656049563311564e-05,
+      "loss": 0.0272,
+      "step": 9770
+    },
+    {
+      "action_loss": 0.015393629670143127,
+      "epoch": 8.785971223021583,
+      "step": 9770
+    },
+    {
+      "epoch": 8.785971223021583,
+      "step": 9770,
+      "torque_loss": 0.14998044073581696
+    },
+    {
+      "epoch": 8.785971223021583,
+      "force_loss": 0.00780859449878335,
+      "step": 9770
+    },
+    {
+      "epoch": 8.79496402877698,
+      "grad_norm": 0.6138122081756592,
+      "learning_rate": 9.655044419191929e-05,
+      "loss": 0.0332,
+      "step": 9780
+    },
+    {
+      "action_loss": 0.009867786429822445,
+      "epoch": 8.79496402877698,
+      "step": 9780
+    },
+    {
+      "epoch": 8.79496402877698,
+      "step": 9780,
+      "torque_loss": 0.20817220211029053
+    },
+    {
+      "epoch": 8.79496402877698,
+      "force_loss": 0.005335627589374781,
+      "step": 9780
+    },
+    {
+      "epoch": 8.803956834532373,
+      "grad_norm": 0.5172603726387024,
+      "learning_rate": 9.654037860992711e-05,
+      "loss": 0.0313,
+      "step": 9790
+    },
+    {
+      "action_loss": 0.019808964803814888,
+      "epoch": 8.803956834532373,
+      "step": 9790
+    },
+    {
+      "epoch": 8.803956834532373,
+      "step": 9790,
+      "torque_loss": 0.1766410619020462
+    },
+    {
+      "epoch": 8.803956834532373,
+      "force_loss": 0.023585960268974304,
+      "step": 9790
+    },
+    {
+      "epoch": 8.81294964028777,
+      "grad_norm": 0.7335936427116394,
+      "learning_rate": 9.653029889019672e-05,
+      "loss": 0.031,
+      "step": 9800
+    },
+    {
+      "action_loss": 0.016281189396977425,
+      "epoch": 8.81294964028777,
+      "step": 9800
+    },
+    {
+      "epoch": 8.81294964028777,
+      "step": 9800,
+      "torque_loss": 0.14487840235233307
+    },
+    {
+      "epoch": 8.81294964028777,
+      "force_loss": 0.010766711086034775,
+      "step": 9800
+    },
+    {
+      "epoch": 8.821942446043165,
+      "grad_norm": 0.46491530537605286,
+      "learning_rate": 9.65202050357901e-05,
+      "loss": 0.0318,
+      "step": 9810
+    },
+    {
+      "action_loss": 0.042640089988708496,
+      "epoch": 8.821942446043165,
+      "step": 9810
+    },
+    {
+      "epoch": 8.821942446043165,
+      "step": 9810,
+      "torque_loss": 0.18616162240505219
+    },
+    {
+      "epoch": 8.821942446043165,
+      "force_loss": 0.03156326711177826,
+      "step": 9810
+    },
+    {
+      "epoch": 8.83093525179856,
+      "grad_norm": 0.29078832268714905,
+      "learning_rate": 9.651009704977347e-05,
+      "loss": 0.0277,
+      "step": 9820
+    },
+    {
+      "action_loss": 0.006332410965114832,
+      "epoch": 8.83093525179856,
+      "step": 9820
+    },
+    {
+      "epoch": 8.83093525179856,
+      "step": 9820,
+      "torque_loss": 0.1467493623495102
+    },
+    {
+      "epoch": 8.83093525179856,
+      "force_loss": 0.006051637232303619,
+      "step": 9820
+    },
+    {
+      "epoch": 8.839928057553957,
+      "grad_norm": 0.662349283695221,
+      "learning_rate": 9.649997493521738e-05,
+      "loss": 0.0307,
+      "step": 9830
+    },
+    {
+      "action_loss": 0.008931110613048077,
+      "epoch": 8.839928057553957,
+      "step": 9830
+    },
+    {
+      "epoch": 8.839928057553957,
+      "step": 9830,
+      "torque_loss": 0.13099968433380127
+    },
+    {
+      "epoch": 8.839928057553957,
+      "force_loss": 0.006029070820659399,
+      "step": 9830
+    },
+    {
+      "epoch": 8.848920863309353,
+      "grad_norm": 0.374646931886673,
+      "learning_rate": 9.64898386951967e-05,
+      "loss": 0.0282,
+      "step": 9840
+    },
+    {
+      "action_loss": 0.038531046360731125,
+      "epoch": 8.848920863309353,
+      "step": 9840
+    },
+    {
+      "epoch": 8.848920863309353,
+      "step": 9840,
+      "torque_loss": 0.2690763771533966
+    },
+    {
+      "epoch": 8.848920863309353,
+      "force_loss": 0.039697516709566116,
+      "step": 9840
+    },
+    {
+      "epoch": 8.857913669064748,
+      "grad_norm": 0.4819861054420471,
+      "learning_rate": 9.647968833279049e-05,
+      "loss": 0.032,
+      "step": 9850
+    },
+    {
+      "action_loss": 0.015623677521944046,
+      "epoch": 8.857913669064748,
+      "step": 9850
+    },
+    {
+      "epoch": 8.857913669064748,
+      "step": 9850,
+      "torque_loss": 0.19835130870342255
+    },
+    {
+      "epoch": 8.857913669064748,
+      "force_loss": 0.019565140828490257,
+      "step": 9850
+    },
+    {
+      "epoch": 8.866906474820144,
+      "grad_norm": 0.5853265523910522,
+      "learning_rate": 9.646952385108218e-05,
+      "loss": 0.0308,
+      "step": 9860
+    },
+    {
+      "action_loss": 0.011000948958098888,
+      "epoch": 8.866906474820144,
+      "step": 9860
+    },
+    {
+      "epoch": 8.866906474820144,
+      "step": 9860,
+      "torque_loss": 0.13960133492946625
+    },
+    {
+      "epoch": 8.866906474820144,
+      "force_loss": 0.0076547362841665745,
+      "step": 9860
+    },
+    {
+      "epoch": 8.87589928057554,
+      "grad_norm": 0.5242618322372437,
+      "learning_rate": 9.645934525315951e-05,
+      "loss": 0.026,
+      "step": 9870
+    },
+    {
+      "action_loss": 0.0023689023219048977,
+      "epoch": 8.87589928057554,
+      "step": 9870
+    },
+    {
+      "epoch": 8.87589928057554,
+      "step": 9870,
+      "torque_loss": 0.09525662660598755
+    },
+    {
+      "epoch": 8.87589928057554,
+      "force_loss": 0.004539574030786753,
+      "step": 9870
+    },
+    {
+      "epoch": 8.884892086330936,
+      "grad_norm": 0.7364248633384705,
+      "learning_rate": 9.644915254211442e-05,
+      "loss": 0.029,
+      "step": 9880
+    },
+    {
+      "action_loss": 0.007957479916512966,
+      "epoch": 8.884892086330936,
+      "step": 9880
+    },
+    {
+      "epoch": 8.884892086330936,
+      "step": 9880,
+      "torque_loss": 0.12385199218988419
+    },
+    {
+      "epoch": 8.884892086330936,
+      "force_loss": 0.010562228970229626,
+      "step": 9880
+    },
+    {
+      "epoch": 8.89388489208633,
+      "grad_norm": 0.6464004516601562,
+      "learning_rate": 9.643894572104321e-05,
+      "loss": 0.0281,
+      "step": 9890
+    },
+    {
+      "action_loss": 0.027421554550528526,
+      "epoch": 8.89388489208633,
+      "step": 9890
+    },
+    {
+      "epoch": 8.89388489208633,
+      "step": 9890,
+      "torque_loss": 0.19964134693145752
+    },
+    {
+      "epoch": 8.89388489208633,
+      "force_loss": 0.03211475536227226,
+      "step": 9890
+    },
+    {
+      "epoch": 8.902877697841726,
+      "grad_norm": 0.5398285984992981,
+      "learning_rate": 9.642872479304644e-05,
+      "loss": 0.033,
+      "step": 9900
+    },
+    {
+      "action_loss": 0.010402509942650795,
+      "epoch": 8.902877697841726,
+      "step": 9900
+    },
+    {
+      "epoch": 8.902877697841726,
+      "step": 9900,
+      "torque_loss": 0.16058401763439178
+    },
+    {
+      "epoch": 8.902877697841726,
+      "force_loss": 0.011009358800947666,
+      "step": 9900
+    },
+    {
+      "epoch": 8.911870503597122,
+      "grad_norm": 0.5709487795829773,
+      "learning_rate": 9.641848976122895e-05,
+      "loss": 0.0286,
+      "step": 9910
+    },
+    {
+      "action_loss": 0.011459200643002987,
+      "epoch": 8.911870503597122,
+      "step": 9910
+    },
+    {
+      "epoch": 8.911870503597122,
+      "step": 9910,
+      "torque_loss": 0.105069600045681
+    },
+    {
+      "epoch": 8.911870503597122,
+      "force_loss": 0.014452147297561169,
+      "step": 9910
+    },
+    {
+      "epoch": 8.920863309352518,
+      "grad_norm": 0.5692060589790344,
+      "learning_rate": 9.64082406286999e-05,
+      "loss": 0.0241,
+      "step": 9920
+    },
+    {
+      "action_loss": 0.01126785296946764,
+      "epoch": 8.920863309352518,
+      "step": 9920
+    },
+    {
+      "epoch": 8.920863309352518,
+      "step": 9920,
+      "torque_loss": 0.12175858020782471
+    },
+    {
+      "epoch": 8.920863309352518,
+      "force_loss": 0.010023976676166058,
+      "step": 9920
+    },
+    {
+      "epoch": 8.929856115107913,
+      "grad_norm": 0.9473484754562378,
+      "learning_rate": 9.639797739857269e-05,
+      "loss": 0.038,
+      "step": 9930
+    },
+    {
+      "action_loss": 0.006853196769952774,
+      "epoch": 8.929856115107913,
+      "step": 9930
+    },
+    {
+      "epoch": 8.929856115107913,
+      "step": 9930,
+      "torque_loss": 0.181155726313591
+    },
+    {
+      "epoch": 8.929856115107913,
+      "force_loss": 0.0068139140494167805,
+      "step": 9930
+    },
+    {
+      "epoch": 8.93884892086331,
+      "grad_norm": 0.3575853109359741,
+      "learning_rate": 9.638770007396498e-05,
+      "loss": 0.03,
+      "step": 9940
+    },
+    {
+      "action_loss": 0.00487502058967948,
+      "epoch": 8.93884892086331,
+      "step": 9940
+    },
+    {
+      "epoch": 8.93884892086331,
+      "step": 9940,
+      "torque_loss": 0.10531579703092575
+    },
+    {
+      "epoch": 8.93884892086331,
+      "force_loss": 0.006260173860937357,
+      "step": 9940
+    },
+    {
+      "epoch": 8.947841726618705,
+      "grad_norm": 0.5004148483276367,
+      "learning_rate": 9.63774086579988e-05,
+      "loss": 0.0399,
+      "step": 9950
+    },
+    {
+      "action_loss": 0.01306405384093523,
+      "epoch": 8.947841726618705,
+      "step": 9950
+    },
+    {
+      "epoch": 8.947841726618705,
+      "step": 9950,
+      "torque_loss": 0.1490817666053772
+    },
+    {
+      "epoch": 8.947841726618705,
+      "force_loss": 0.013406097888946533,
+      "step": 9950
+    },
+    {
+      "epoch": 8.956834532374101,
+      "grad_norm": 0.8100035786628723,
+      "learning_rate": 9.63671031538004e-05,
+      "loss": 0.0271,
+      "step": 9960
+    },
+    {
+      "action_loss": 0.007933643646538258,
+      "epoch": 8.956834532374101,
+      "step": 9960
+    },
+    {
+      "epoch": 8.956834532374101,
+      "step": 9960,
+      "torque_loss": 0.0904865637421608
+    },
+    {
+      "epoch": 8.956834532374101,
+      "force_loss": 0.0069241710007190704,
+      "step": 9960
+    },
+    {
+      "epoch": 8.965827338129497,
+      "grad_norm": 0.8812164664268494,
+      "learning_rate": 9.635678356450031e-05,
+      "loss": 0.029,
+      "step": 9970
+    },
+    {
+      "action_loss": 0.005592092871665955,
+      "epoch": 8.965827338129497,
+      "step": 9970
+    },
+    {
+      "epoch": 8.965827338129497,
+      "step": 9970,
+      "torque_loss": 0.11497309058904648
+    },
+    {
+      "epoch": 8.965827338129497,
+      "force_loss": 0.005749887321144342,
+      "step": 9970
+    },
+    {
+      "epoch": 8.974820143884893,
+      "grad_norm": 0.6939945816993713,
+      "learning_rate": 9.634644989323336e-05,
+      "loss": 0.0286,
+      "step": 9980
+    },
+    {
+      "action_loss": 0.02402697503566742,
+      "epoch": 8.974820143884893,
+      "step": 9980
+    },
+    {
+      "epoch": 8.974820143884893,
+      "step": 9980,
+      "torque_loss": 0.14686010777950287
+    },
+    {
+      "epoch": 8.974820143884893,
+      "force_loss": 0.020023677498102188,
+      "step": 9980
+    },
+    {
+      "epoch": 8.983812949640289,
+      "grad_norm": 0.6381686329841614,
+      "learning_rate": 9.633610214313861e-05,
+      "loss": 0.0303,
+      "step": 9990
+    },
+    {
+      "action_loss": 0.010235019028186798,
+      "epoch": 8.983812949640289,
+      "step": 9990
+    },
+    {
+      "epoch": 8.983812949640289,
+      "step": 9990,
+      "torque_loss": 0.1290939599275589
+    },
+    {
+      "epoch": 8.983812949640289,
+      "force_loss": 0.012070546858012676,
+      "step": 9990
+    },
+    {
+      "epoch": 8.992805755395683,
+      "grad_norm": 0.6441584825515747,
+      "learning_rate": 9.632574031735951e-05,
+      "loss": 0.0283,
+      "step": 10000
+    },
+    {
+      "action_loss": 0.01705154776573181,
+      "epoch": 8.992805755395683,
+      "step": 10000
+    },
+    {
+      "epoch": 8.992805755395683,
+      "step": 10000,
+      "torque_loss": 0.13199247419834137
+    },
+    {
+      "epoch": 8.992805755395683,
+      "force_loss": 0.012645289301872253,
+      "step": 10000
+    },
+    {
+      "epoch": 9.001798561151078,
+      "grad_norm": 0.5573319792747498,
+      "learning_rate": 9.631536441904364e-05,
+      "loss": 0.0271,
+      "step": 10010
+    },
+    {
+      "action_loss": 0.011835657991468906,
+      "epoch": 9.001798561151078,
+      "step": 10010
+    },
+    {
+      "epoch": 9.001798561151078,
+      "step": 10010,
+      "torque_loss": 0.12104404717683792
+    },
+    {
+      "epoch": 9.001798561151078,
+      "force_loss": 0.01840282790362835,
+      "step": 10010
+    },
+    {
+      "epoch": 9.010791366906474,
+      "grad_norm": 0.3133363723754883,
+      "learning_rate": 9.630497445134293e-05,
+      "loss": 0.0266,
+      "step": 10020
+    },
+    {
+      "action_loss": 0.01563299633562565,
+      "epoch": 9.010791366906474,
+      "step": 10020
+    },
+    {
+      "epoch": 9.010791366906474,
+      "step": 10020,
+      "torque_loss": 0.19062598049640656
+    },
+    {
+      "epoch": 9.010791366906474,
+      "force_loss": 0.018884213641285896,
+      "step": 10020
+    },
+    {
+      "epoch": 9.01978417266187,
+      "grad_norm": 0.4759218394756317,
+      "learning_rate": 9.62945704174136e-05,
+      "loss": 0.0289,
+      "step": 10030
+    },
+    {
+      "action_loss": 0.02218557894229889,
+      "epoch": 9.01978417266187,
+      "step": 10030
+    },
+    {
+      "epoch": 9.01978417266187,
+      "step": 10030,
+      "torque_loss": 0.11240116506814957
+    },
+    {
+      "epoch": 9.01978417266187,
+      "force_loss": 0.030858725309371948,
+      "step": 10030
+    },
+    {
+      "epoch": 9.028776978417266,
+      "grad_norm": 0.8062295317649841,
+      "learning_rate": 9.628415232041612e-05,
+      "loss": 0.0331,
+      "step": 10040
+    },
+    {
+      "action_loss": 0.02885535918176174,
+      "epoch": 9.028776978417266,
+      "step": 10040
+    },
+    {
+      "epoch": 9.028776978417266,
+      "step": 10040,
+      "torque_loss": 0.142742320895195
+    },
+    {
+      "epoch": 9.028776978417266,
+      "force_loss": 0.030764369294047356,
+      "step": 10040
+    },
+    {
+      "epoch": 9.037769784172662,
+      "grad_norm": 0.9538134932518005,
+      "learning_rate": 9.627372016351524e-05,
+      "loss": 0.0289,
+      "step": 10050
+    },
+    {
+      "action_loss": 0.014418564736843109,
+      "epoch": 9.037769784172662,
+      "step": 10050
+    },
+    {
+      "epoch": 9.037769784172662,
+      "step": 10050,
+      "torque_loss": 0.22218851745128632
+    },
+    {
+      "epoch": 9.037769784172662,
+      "force_loss": 0.01950434036552906,
+      "step": 10050
+    },
+    {
+      "epoch": 9.046762589928058,
+      "grad_norm": 0.8337439298629761,
+      "learning_rate": 9.626327394987995e-05,
+      "loss": 0.0309,
+      "step": 10060
+    },
+    {
+      "action_loss": 0.009677976369857788,
+      "epoch": 9.046762589928058,
+      "step": 10060
+    },
+    {
+      "epoch": 9.046762589928058,
+      "step": 10060,
+      "torque_loss": 0.15954314172267914
+    },
+    {
+      "epoch": 9.046762589928058,
+      "force_loss": 0.014350146986544132,
+      "step": 10060
+    },
+    {
+      "epoch": 9.055755395683454,
+      "grad_norm": 0.8795813322067261,
+      "learning_rate": 9.625281368268355e-05,
+      "loss": 0.0307,
+      "step": 10070
+    },
+    {
+      "action_loss": 0.007890405133366585,
+      "epoch": 9.055755395683454,
+      "step": 10070
+    },
+    {
+      "epoch": 9.055755395683454,
+      "step": 10070,
+      "torque_loss": 0.11823488026857376
+    },
+    {
+      "epoch": 9.055755395683454,
+      "force_loss": 0.004497329238802195,
+      "step": 10070
+    },
+    {
+      "epoch": 9.06474820143885,
+      "grad_norm": 0.4431263208389282,
+      "learning_rate": 9.624233936510357e-05,
+      "loss": 0.0268,
+      "step": 10080
+    },
+    {
+      "action_loss": 0.008813533931970596,
+      "epoch": 9.06474820143885,
+      "step": 10080
+    },
+    {
+      "epoch": 9.06474820143885,
+      "step": 10080,
+      "torque_loss": 0.1318175047636032
+    },
+    {
+      "epoch": 9.06474820143885,
+      "force_loss": 0.014058738946914673,
+      "step": 10080
+    },
+    {
+      "epoch": 9.073741007194245,
+      "grad_norm": 0.8673973083496094,
+      "learning_rate": 9.623185100032187e-05,
+      "loss": 0.031,
+      "step": 10090
+    },
+    {
+      "action_loss": 0.009516742080450058,
+      "epoch": 9.073741007194245,
+      "step": 10090
+    },
+    {
+      "epoch": 9.073741007194245,
+      "step": 10090,
+      "torque_loss": 0.1367594599723816
+    },
+    {
+      "epoch": 9.073741007194245,
+      "force_loss": 0.00948475394397974,
+      "step": 10090
+    },
+    {
+      "epoch": 9.082733812949641,
+      "grad_norm": 0.4420112669467926,
+      "learning_rate": 9.62213485915245e-05,
+      "loss": 0.0282,
+      "step": 10100
+    },
+    {
+      "action_loss": 0.012250508181750774,
+      "epoch": 9.082733812949641,
+      "step": 10100
+    },
+    {
+      "epoch": 9.082733812949641,
+      "step": 10100,
+      "torque_loss": 0.17002497613430023
+    },
+    {
+      "epoch": 9.082733812949641,
+      "force_loss": 0.008248533122241497,
+      "step": 10100
+    },
+    {
+      "epoch": 9.091726618705035,
+      "grad_norm": 0.3544306755065918,
+      "learning_rate": 9.621083214190186e-05,
+      "loss": 0.0263,
+      "step": 10110
+    },
+    {
+      "action_loss": 0.010369070805609226,
+      "epoch": 9.091726618705035,
+      "step": 10110
+    },
+    {
+      "epoch": 9.091726618705035,
+      "step": 10110,
+      "torque_loss": 0.11875941604375839
+    },
+    {
+      "epoch": 9.091726618705035,
+      "force_loss": 0.012134049087762833,
+      "step": 10110
+    },
+    {
+      "epoch": 9.100719424460431,
+      "grad_norm": 0.8716443777084351,
+      "learning_rate": 9.62003016546485e-05,
+      "loss": 0.0312,
+      "step": 10120
+    },
+    {
+      "action_loss": 0.017326300963759422,
+      "epoch": 9.100719424460431,
+      "step": 10120
+    },
+    {
+      "epoch": 9.100719424460431,
+      "step": 10120,
+      "torque_loss": 0.15255622565746307
+    },
+    {
+      "epoch": 9.100719424460431,
+      "force_loss": 0.018519021570682526,
+      "step": 10120
+    },
+    {
+      "epoch": 9.109712230215827,
+      "grad_norm": 0.32582125067710876,
+      "learning_rate": 9.618975713296339e-05,
+      "loss": 0.0247,
+      "step": 10130
+    },
+    {
+      "action_loss": 0.00469488138332963,
+      "epoch": 9.109712230215827,
+      "step": 10130
+    },
+    {
+      "epoch": 9.109712230215827,
+      "step": 10130,
+      "torque_loss": 0.15045224130153656
+    },
+    {
+      "epoch": 9.109712230215827,
+      "force_loss": 0.007093103602528572,
+      "step": 10130
+    },
+    {
+      "epoch": 9.118705035971223,
+      "grad_norm": 0.35480907559394836,
+      "learning_rate": 9.61791985800496e-05,
+      "loss": 0.0229,
+      "step": 10140
+    },
+    {
+      "action_loss": 0.022707441821694374,
+      "epoch": 9.118705035971223,
+      "step": 10140
+    },
+    {
+      "epoch": 9.118705035971223,
+      "step": 10140,
+      "torque_loss": 0.14760971069335938
+    },
+    {
+      "epoch": 9.118705035971223,
+      "force_loss": 0.022919008508324623,
+      "step": 10140
+    },
+    {
+      "epoch": 9.127697841726619,
+      "grad_norm": 0.7083514332771301,
+      "learning_rate": 9.616862599911458e-05,
+      "loss": 0.0303,
+      "step": 10150
+    },
+    {
+      "action_loss": 0.01765744388103485,
+      "epoch": 9.127697841726619,
+      "step": 10150
+    },
+    {
+      "epoch": 9.127697841726619,
+      "step": 10150,
+      "torque_loss": 0.1288621872663498
+    },
+    {
+      "epoch": 9.127697841726619,
+      "force_loss": 0.028317635878920555,
+      "step": 10150
+    },
+    {
+      "epoch": 9.136690647482014,
+      "grad_norm": 0.42028236389160156,
+      "learning_rate": 9.615803939337e-05,
+      "loss": 0.0287,
+      "step": 10160
+    },
+    {
+      "action_loss": 0.014068935066461563,
+      "epoch": 9.136690647482014,
+      "step": 10160
+    },
+    {
+      "epoch": 9.136690647482014,
+      "step": 10160,
+      "torque_loss": 0.12550655007362366
+    },
+    {
+      "epoch": 9.136690647482014,
+      "force_loss": 0.01162963267415762,
+      "step": 10160
+    },
+    {
+      "epoch": 9.14568345323741,
+      "grad_norm": 0.4233534038066864,
+      "learning_rate": 9.614743876603178e-05,
+      "loss": 0.0274,
+      "step": 10170
+    },
+    {
+      "action_loss": 0.017068881541490555,
+      "epoch": 9.14568345323741,
+      "step": 10170
+    },
+    {
+      "epoch": 9.14568345323741,
+      "step": 10170,
+      "torque_loss": 0.13255678117275238
+    },
+    {
+      "epoch": 9.14568345323741,
+      "force_loss": 0.013134836219251156,
+      "step": 10170
+    },
+    {
+      "epoch": 9.154676258992806,
+      "grad_norm": 0.6075559258460999,
+      "learning_rate": 9.613682412032013e-05,
+      "loss": 0.0286,
+      "step": 10180
+    },
+    {
+      "action_loss": 0.02802620828151703,
+      "epoch": 9.154676258992806,
+      "step": 10180
+    },
+    {
+      "epoch": 9.154676258992806,
+      "step": 10180,
+      "torque_loss": 0.15462185442447662
+    },
+    {
+      "epoch": 9.154676258992806,
+      "force_loss": 0.023218432441353798,
+      "step": 10180
+    },
+    {
+      "epoch": 9.163669064748202,
+      "grad_norm": 0.5244428515434265,
+      "learning_rate": 9.612619545945947e-05,
+      "loss": 0.0364,
+      "step": 10190
+    },
+    {
+      "action_loss": 0.005900113377720118,
+      "epoch": 9.163669064748202,
+      "step": 10190
+    },
+    {
+      "epoch": 9.163669064748202,
+      "step": 10190,
+      "torque_loss": 0.13469235599040985
+    },
+    {
+      "epoch": 9.163669064748202,
+      "force_loss": 0.010282638482749462,
+      "step": 10190
+    },
+    {
+      "epoch": 9.172661870503598,
+      "grad_norm": 0.6628869771957397,
+      "learning_rate": 9.611555278667852e-05,
+      "loss": 0.0266,
+      "step": 10200
+    },
+    {
+      "action_loss": 0.012148316018283367,
+      "epoch": 9.172661870503598,
+      "step": 10200
+    },
+    {
+      "epoch": 9.172661870503598,
+      "step": 10200,
+      "torque_loss": 0.15297381579875946
+    },
+    {
+      "epoch": 9.172661870503598,
+      "force_loss": 0.009634559042751789,
+      "step": 10200
+    },
+    {
+      "epoch": 9.181654676258994,
+      "grad_norm": 0.4624648690223694,
+      "learning_rate": 9.610489610521024e-05,
+      "loss": 0.0275,
+      "step": 10210
+    },
+    {
+      "action_loss": 0.008905833587050438,
+      "epoch": 9.181654676258994,
+      "step": 10210
+    },
+    {
+      "epoch": 9.181654676258994,
+      "step": 10210,
+      "torque_loss": 0.08853333443403244
+    },
+    {
+      "epoch": 9.181654676258994,
+      "force_loss": 0.010210342705249786,
+      "step": 10210
+    },
+    {
+      "epoch": 9.190647482014388,
+      "grad_norm": 0.6700626611709595,
+      "learning_rate": 9.609422541829187e-05,
+      "loss": 0.0255,
+      "step": 10220
+    },
+    {
+      "action_loss": 0.02126232720911503,
+      "epoch": 9.190647482014388,
+      "step": 10220
+    },
+    {
+      "epoch": 9.190647482014388,
+      "step": 10220,
+      "torque_loss": 0.1380685716867447
+    },
+    {
+      "epoch": 9.190647482014388,
+      "force_loss": 0.012280315160751343,
+      "step": 10220
+    },
+    {
+      "epoch": 9.199640287769784,
+      "grad_norm": 0.40555429458618164,
+      "learning_rate": 9.608354072916486e-05,
+      "loss": 0.0356,
+      "step": 10230
+    },
+    {
+      "action_loss": 0.006029499229043722,
+      "epoch": 9.199640287769784,
+      "step": 10230
+    },
+    {
+      "epoch": 9.199640287769784,
+      "step": 10230,
+      "torque_loss": 0.1045355424284935
+    },
+    {
+      "epoch": 9.199640287769784,
+      "force_loss": 0.007806442677974701,
+      "step": 10230
+    },
+    {
+      "epoch": 9.20863309352518,
+      "grad_norm": 0.9280624389648438,
+      "learning_rate": 9.607284204107493e-05,
+      "loss": 0.0314,
+      "step": 10240
+    },
+    {
+      "action_loss": 0.009026926942169666,
+      "epoch": 9.20863309352518,
+      "step": 10240
+    },
+    {
+      "epoch": 9.20863309352518,
+      "step": 10240,
+      "torque_loss": 0.16856233775615692
+    },
+    {
+      "epoch": 9.20863309352518,
+      "force_loss": 0.006835187319666147,
+      "step": 10240
+    },
+    {
+      "epoch": 9.217625899280575,
+      "grad_norm": 1.458818793296814,
+      "learning_rate": 9.606212935727208e-05,
+      "loss": 0.0257,
+      "step": 10250
+    },
+    {
+      "action_loss": 0.01133593637496233,
+      "epoch": 9.217625899280575,
+      "step": 10250
+    },
+    {
+      "epoch": 9.217625899280575,
+      "step": 10250,
+      "torque_loss": 0.1393679827451706
+    },
+    {
+      "epoch": 9.217625899280575,
+      "force_loss": 0.008330075070261955,
+      "step": 10250
+    },
+    {
+      "epoch": 9.226618705035971,
+      "grad_norm": 0.6630749702453613,
+      "learning_rate": 9.605140268101052e-05,
+      "loss": 0.0278,
+      "step": 10260
+    },
+    {
+      "action_loss": 0.00489208335056901,
+      "epoch": 9.226618705035971,
+      "step": 10260
+    },
+    {
+      "epoch": 9.226618705035971,
+      "step": 10260,
+      "torque_loss": 0.14676211774349213
+    },
+    {
+      "epoch": 9.226618705035971,
+      "force_loss": 0.006073640193790197,
+      "step": 10260
+    },
+    {
+      "epoch": 9.235611510791367,
+      "grad_norm": 0.6277463436126709,
+      "learning_rate": 9.604066201554875e-05,
+      "loss": 0.0254,
+      "step": 10270
+    },
+    {
+      "action_loss": 0.01280087698251009,
+      "epoch": 9.235611510791367,
+      "step": 10270
+    },
+    {
+      "epoch": 9.235611510791367,
+      "step": 10270,
+      "torque_loss": 0.13024820387363434
+    },
+    {
+      "epoch": 9.235611510791367,
+      "force_loss": 0.02367275021970272,
+      "step": 10270
+    },
+    {
+      "epoch": 9.244604316546763,
+      "grad_norm": 0.7234767079353333,
+      "learning_rate": 9.60299073641495e-05,
+      "loss": 0.029,
+      "step": 10280
+    },
+    {
+      "action_loss": 0.008430641144514084,
+      "epoch": 9.244604316546763,
+      "step": 10280
+    },
+    {
+      "epoch": 9.244604316546763,
+      "step": 10280,
+      "torque_loss": 0.1173025369644165
+    },
+    {
+      "epoch": 9.244604316546763,
+      "force_loss": 0.008532408624887466,
+      "step": 10280
+    },
+    {
+      "epoch": 9.253597122302159,
+      "grad_norm": 0.6215731501579285,
+      "learning_rate": 9.601913873007974e-05,
+      "loss": 0.0238,
+      "step": 10290
+    },
+    {
+      "action_loss": 0.012716807425022125,
+      "epoch": 9.253597122302159,
+      "step": 10290
+    },
+    {
+      "epoch": 9.253597122302159,
+      "step": 10290,
+      "torque_loss": 0.14327767491340637
+    },
+    {
+      "epoch": 9.253597122302159,
+      "force_loss": 0.021583078429102898,
+      "step": 10290
+    },
+    {
+      "epoch": 9.262589928057555,
+      "grad_norm": 0.6028773784637451,
+      "learning_rate": 9.60083561166107e-05,
+      "loss": 0.0317,
+      "step": 10300
+    },
+    {
+      "action_loss": 0.006474910769611597,
+      "epoch": 9.262589928057555,
+      "step": 10300
+    },
+    {
+      "epoch": 9.262589928057555,
+      "step": 10300,
+      "torque_loss": 0.11099831014871597
+    },
+    {
+      "epoch": 9.262589928057555,
+      "force_loss": 0.006636692676693201,
+      "step": 10300
+    },
+    {
+      "epoch": 9.27158273381295,
+      "grad_norm": 0.6199492812156677,
+      "learning_rate": 9.599755952701783e-05,
+      "loss": 0.0296,
+      "step": 10310
+    },
+    {
+      "action_loss": 0.014989577233791351,
+      "epoch": 9.27158273381295,
+      "step": 10310
+    },
+    {
+      "epoch": 9.27158273381295,
+      "step": 10310,
+      "torque_loss": 0.14685826003551483
+    },
+    {
+      "epoch": 9.27158273381295,
+      "force_loss": 0.010977485217154026,
+      "step": 10310
+    },
+    {
+      "epoch": 9.280575539568344,
+      "grad_norm": 0.4850398600101471,
+      "learning_rate": 9.598674896458089e-05,
+      "loss": 0.0312,
+      "step": 10320
+    },
+    {
+      "action_loss": 0.004792255815118551,
+      "epoch": 9.280575539568344,
+      "step": 10320
+    },
+    {
+      "epoch": 9.280575539568344,
+      "step": 10320,
+      "torque_loss": 0.09150233119726181
+    },
+    {
+      "epoch": 9.280575539568344,
+      "force_loss": 0.004847896751016378,
+      "step": 10320
+    },
+    {
+      "epoch": 9.28956834532374,
+      "grad_norm": 0.5110653042793274,
+      "learning_rate": 9.597592443258383e-05,
+      "loss": 0.0221,
+      "step": 10330
+    },
+    {
+      "action_loss": 0.00591149041429162,
+      "epoch": 9.28956834532374,
+      "step": 10330
+    },
+    {
+      "epoch": 9.28956834532374,
+      "step": 10330,
+      "torque_loss": 0.1200617179274559
+    },
+    {
+      "epoch": 9.28956834532374,
+      "force_loss": 0.005874169524759054,
+      "step": 10330
+    },
+    {
+      "epoch": 9.298561151079136,
+      "grad_norm": 0.6226730942726135,
+      "learning_rate": 9.596508593431483e-05,
+      "loss": 0.0232,
+      "step": 10340
+    },
+    {
+      "action_loss": 0.006625715643167496,
+      "epoch": 9.298561151079136,
+      "step": 10340
+    },
+    {
+      "epoch": 9.298561151079136,
+      "step": 10340,
+      "torque_loss": 0.09729776531457901
+    },
+    {
+      "epoch": 9.298561151079136,
+      "force_loss": 0.0058612339198589325,
+      "step": 10340
+    },
+    {
+      "epoch": 9.307553956834532,
+      "grad_norm": 0.5901241302490234,
+      "learning_rate": 9.59542334730664e-05,
+      "loss": 0.026,
+      "step": 10350
+    },
+    {
+      "action_loss": 0.01095897238701582,
+      "epoch": 9.307553956834532,
+      "step": 10350
+    },
+    {
+      "epoch": 9.307553956834532,
+      "step": 10350,
+      "torque_loss": 0.12940442562103271
+    },
+    {
+      "epoch": 9.307553956834532,
+      "force_loss": 0.009427826851606369,
+      "step": 10350
+    },
+    {
+      "epoch": 9.316546762589928,
+      "grad_norm": 0.5962520837783813,
+      "learning_rate": 9.594336705213516e-05,
+      "loss": 0.0283,
+      "step": 10360
+    },
+    {
+      "action_loss": 0.00791415199637413,
+      "epoch": 9.316546762589928,
+      "step": 10360
+    },
+    {
+      "epoch": 9.316546762589928,
+      "step": 10360,
+      "torque_loss": 0.12326723337173462
+    },
+    {
+      "epoch": 9.316546762589928,
+      "force_loss": 0.008934843353927135,
+      "step": 10360
+    },
+    {
+      "epoch": 9.325539568345324,
+      "grad_norm": 0.35113057494163513,
+      "learning_rate": 9.593248667482208e-05,
+      "loss": 0.0322,
+      "step": 10370
+    },
+    {
+      "action_loss": 0.03247269615530968,
+      "epoch": 9.325539568345324,
+      "step": 10370
+    },
+    {
+      "epoch": 9.325539568345324,
+      "step": 10370,
+      "torque_loss": 0.1403774619102478
+    },
+    {
+      "epoch": 9.325539568345324,
+      "force_loss": 0.04250963404774666,
+      "step": 10370
+    },
+    {
+      "epoch": 9.33453237410072,
+      "grad_norm": 0.6529799103736877,
+      "learning_rate": 9.592159234443233e-05,
+      "loss": 0.0339,
+      "step": 10380
+    },
+    {
+      "action_loss": 0.020657196640968323,
+      "epoch": 9.33453237410072,
+      "step": 10380
+    },
+    {
+      "epoch": 9.33453237410072,
+      "step": 10380,
+      "torque_loss": 0.15035028755664825
+    },
+    {
+      "epoch": 9.33453237410072,
+      "force_loss": 0.01743476837873459,
+      "step": 10380
+    },
+    {
+      "epoch": 9.343525179856115,
+      "grad_norm": 0.8577680587768555,
+      "learning_rate": 9.59106840642753e-05,
+      "loss": 0.0351,
+      "step": 10390
+    },
+    {
+      "action_loss": 0.016335487365722656,
+      "epoch": 9.343525179856115,
+      "step": 10390
+    },
+    {
+      "epoch": 9.343525179856115,
+      "step": 10390,
+      "torque_loss": 0.11230379343032837
+    },
+    {
+      "epoch": 9.343525179856115,
+      "force_loss": 0.024720648303627968,
+      "step": 10390
+    },
+    {
+      "epoch": 9.352517985611511,
+      "grad_norm": 0.7196080088615417,
+      "learning_rate": 9.589976183766467e-05,
+      "loss": 0.0299,
+      "step": 10400
+    },
+    {
+      "action_loss": 0.007247576955705881,
+      "epoch": 9.352517985611511,
+      "step": 10400
+    },
+    {
+      "epoch": 9.352517985611511,
+      "step": 10400,
+      "torque_loss": 0.09510833024978638
+    },
+    {
+      "epoch": 9.352517985611511,
+      "force_loss": 0.009220664389431477,
+      "step": 10400
+    },
+    {
+      "epoch": 9.361510791366907,
+      "grad_norm": 1.1985657215118408,
+      "learning_rate": 9.58888256679183e-05,
+      "loss": 0.0308,
+      "step": 10410
+    },
+    {
+      "action_loss": 0.008461021818220615,
+      "epoch": 9.361510791366907,
+      "step": 10410
+    },
+    {
+      "epoch": 9.361510791366907,
+      "step": 10410,
+      "torque_loss": 0.1299380362033844
+    },
+    {
+      "epoch": 9.361510791366907,
+      "force_loss": 0.007819849997758865,
+      "step": 10410
+    },
+    {
+      "epoch": 9.370503597122303,
+      "grad_norm": 1.1465376615524292,
+      "learning_rate": 9.587787555835832e-05,
+      "loss": 0.0279,
+      "step": 10420
+    },
+    {
+      "action_loss": 0.022557318210601807,
+      "epoch": 9.370503597122303,
+      "step": 10420
+    },
+    {
+      "epoch": 9.370503597122303,
+      "step": 10420,
+      "torque_loss": 0.16325058043003082
+    },
+    {
+      "epoch": 9.370503597122303,
+      "force_loss": 0.024700278416275978,
+      "step": 10420
+    },
+    {
+      "epoch": 9.379496402877697,
+      "grad_norm": 0.41409018635749817,
+      "learning_rate": 9.586691151231107e-05,
+      "loss": 0.0297,
+      "step": 10430
+    },
+    {
+      "action_loss": 0.007909419946372509,
+      "epoch": 9.379496402877697,
+      "step": 10430
+    },
+    {
+      "epoch": 9.379496402877697,
+      "step": 10430,
+      "torque_loss": 0.108908511698246
+    },
+    {
+      "epoch": 9.379496402877697,
+      "force_loss": 0.007891024462878704,
+      "step": 10430
+    },
+    {
+      "epoch": 9.388489208633093,
+      "grad_norm": 0.6971279978752136,
+      "learning_rate": 9.585593353310715e-05,
+      "loss": 0.0307,
+      "step": 10440
+    },
+    {
+      "action_loss": 0.02104235254228115,
+      "epoch": 9.388489208633093,
+      "step": 10440
+    },
+    {
+      "epoch": 9.388489208633093,
+      "step": 10440,
+      "torque_loss": 0.2005579024553299
+    },
+    {
+      "epoch": 9.388489208633093,
+      "force_loss": 0.018684981390833855,
+      "step": 10440
+    },
+    {
+      "epoch": 9.397482014388489,
+      "grad_norm": 0.49678513407707214,
+      "learning_rate": 9.58449416240814e-05,
+      "loss": 0.031,
+      "step": 10450
+    },
+    {
+      "action_loss": 0.006697792559862137,
+      "epoch": 9.397482014388489,
+      "step": 10450
+    },
+    {
+      "epoch": 9.397482014388489,
+      "step": 10450,
+      "torque_loss": 0.13596634566783905
+    },
+    {
+      "epoch": 9.397482014388489,
+      "force_loss": 0.00531019875779748,
+      "step": 10450
+    },
+    {
+      "epoch": 9.406474820143885,
+      "grad_norm": 0.3666587173938751,
+      "learning_rate": 9.583393578857283e-05,
+      "loss": 0.0275,
+      "step": 10460
+    },
+    {
+      "action_loss": 0.011480015702545643,
+      "epoch": 9.406474820143885,
+      "step": 10460
+    },
+    {
+      "epoch": 9.406474820143885,
+      "step": 10460,
+      "torque_loss": 0.11620926111936569
+    },
+    {
+      "epoch": 9.406474820143885,
+      "force_loss": 0.010091166943311691,
+      "step": 10460
+    },
+    {
+      "epoch": 9.41546762589928,
+      "grad_norm": 0.3673293888568878,
+      "learning_rate": 9.582291602992474e-05,
+      "loss": 0.0267,
+      "step": 10470
+    },
+    {
+      "action_loss": 0.013791979290544987,
+      "epoch": 9.41546762589928,
+      "step": 10470
+    },
+    {
+      "epoch": 9.41546762589928,
+      "step": 10470,
+      "torque_loss": 0.12247496098279953
+    },
+    {
+      "epoch": 9.41546762589928,
+      "force_loss": 0.013123749755322933,
+      "step": 10470
+    },
+    {
+      "epoch": 9.424460431654676,
+      "grad_norm": 0.4128532111644745,
+      "learning_rate": 9.581188235148466e-05,
+      "loss": 0.0287,
+      "step": 10480
+    },
+    {
+      "action_loss": 0.004348634276539087,
+      "epoch": 9.424460431654676,
+      "step": 10480
+    },
+    {
+      "epoch": 9.424460431654676,
+      "step": 10480,
+      "torque_loss": 0.13451822102069855
+    },
+    {
+      "epoch": 9.424460431654676,
+      "force_loss": 0.005837304051965475,
+      "step": 10480
+    },
+    {
+      "epoch": 9.433453237410072,
+      "grad_norm": 0.8465983271598816,
+      "learning_rate": 9.58008347566043e-05,
+      "loss": 0.0257,
+      "step": 10490
+    },
+    {
+      "action_loss": 0.01786666549742222,
+      "epoch": 9.433453237410072,
+      "step": 10490
+    },
+    {
+      "epoch": 9.433453237410072,
+      "step": 10490,
+      "torque_loss": 0.19296510517597198
+    },
+    {
+      "epoch": 9.433453237410072,
+      "force_loss": 0.016864048317074776,
+      "step": 10490
+    },
+    {
+      "epoch": 9.442446043165468,
+      "grad_norm": 0.47716739773750305,
+      "learning_rate": 9.578977324863965e-05,
+      "loss": 0.0277,
+      "step": 10500
+    },
+    {
+      "action_loss": 0.009205177426338196,
+      "epoch": 9.442446043165468,
+      "step": 10500
+    },
+    {
+      "epoch": 9.442446043165468,
+      "step": 10500,
+      "torque_loss": 0.17087721824645996
+    },
+    {
+      "epoch": 9.442446043165468,
+      "force_loss": 0.00732346111908555,
+      "step": 10500
+    },
+    {
+      "epoch": 9.451438848920864,
+      "grad_norm": 0.7659825086593628,
+      "learning_rate": 9.577869783095089e-05,
+      "loss": 0.0292,
+      "step": 10510
+    },
+    {
+      "action_loss": 0.016239508986473083,
+      "epoch": 9.451438848920864,
+      "step": 10510
+    },
+    {
+      "epoch": 9.451438848920864,
+      "step": 10510,
+      "torque_loss": 0.15334972739219666
+    },
+    {
+      "epoch": 9.451438848920864,
+      "force_loss": 0.008361724205315113,
+      "step": 10510
+    },
+    {
+      "epoch": 9.46043165467626,
+      "grad_norm": 0.8403934836387634,
+      "learning_rate": 9.576760850690245e-05,
+      "loss": 0.0302,
+      "step": 10520
+    },
+    {
+      "action_loss": 0.014726934023201466,
+      "epoch": 9.46043165467626,
+      "step": 10520
+    },
+    {
+      "epoch": 9.46043165467626,
+      "step": 10520,
+      "torque_loss": 0.19925636053085327
+    },
+    {
+      "epoch": 9.46043165467626,
+      "force_loss": 0.01486267801374197,
+      "step": 10520
+    },
+    {
+      "epoch": 9.469424460431656,
+      "grad_norm": 0.8984691500663757,
+      "learning_rate": 9.575650527986298e-05,
+      "loss": 0.0371,
+      "step": 10530
+    },
+    {
+      "action_loss": 0.006004458758980036,
+      "epoch": 9.469424460431656,
+      "step": 10530
+    },
+    {
+      "epoch": 9.469424460431656,
+      "step": 10530,
+      "torque_loss": 0.11854443699121475
+    },
+    {
+      "epoch": 9.469424460431656,
+      "force_loss": 0.010365949012339115,
+      "step": 10530
+    },
+    {
+      "epoch": 9.47841726618705,
+      "grad_norm": 0.6599291563034058,
+      "learning_rate": 9.574538815320531e-05,
+      "loss": 0.0247,
+      "step": 10540
+    },
+    {
+      "action_loss": 0.01251822430640459,
+      "epoch": 9.47841726618705,
+      "step": 10540
+    },
+    {
+      "epoch": 9.47841726618705,
+      "step": 10540,
+      "torque_loss": 0.15485882759094238
+    },
+    {
+      "epoch": 9.47841726618705,
+      "force_loss": 0.013469722121953964,
+      "step": 10540
+    },
+    {
+      "epoch": 9.487410071942445,
+      "grad_norm": 0.6501151323318481,
+      "learning_rate": 9.573425713030656e-05,
+      "loss": 0.0323,
+      "step": 10550
+    },
+    {
+      "action_loss": 0.006832594517618418,
+      "epoch": 9.487410071942445,
+      "step": 10550
+    },
+    {
+      "epoch": 9.487410071942445,
+      "step": 10550,
+      "torque_loss": 0.15817953646183014
+    },
+    {
+      "epoch": 9.487410071942445,
+      "force_loss": 0.0060196369886398315,
+      "step": 10550
+    },
+    {
+      "epoch": 9.496402877697841,
+      "grad_norm": 0.7456974983215332,
+      "learning_rate": 9.572311221454806e-05,
+      "loss": 0.0257,
+      "step": 10560
+    },
+    {
+      "action_loss": 0.024950919672846794,
+      "epoch": 9.496402877697841,
+      "step": 10560
+    },
+    {
+      "epoch": 9.496402877697841,
+      "step": 10560,
+      "torque_loss": 0.13057518005371094
+    },
+    {
+      "epoch": 9.496402877697841,
+      "force_loss": 0.019463906064629555,
+      "step": 10560
+    },
+    {
+      "epoch": 9.505395683453237,
+      "grad_norm": 0.9070531725883484,
+      "learning_rate": 9.57119534093153e-05,
+      "loss": 0.0315,
+      "step": 10570
+    },
+    {
+      "action_loss": 0.010373858734965324,
+      "epoch": 9.505395683453237,
+      "step": 10570
+    },
+    {
+      "epoch": 9.505395683453237,
+      "step": 10570,
+      "torque_loss": 0.14706474542617798
+    },
+    {
+      "epoch": 9.505395683453237,
+      "force_loss": 0.00816503819078207,
+      "step": 10570
+    },
+    {
+      "epoch": 9.514388489208633,
+      "grad_norm": 1.062094807624817,
+      "learning_rate": 9.570078071799806e-05,
+      "loss": 0.0296,
+      "step": 10580
+    },
+    {
+      "action_loss": 0.009817644022405148,
+      "epoch": 9.514388489208633,
+      "step": 10580
+    },
+    {
+      "epoch": 9.514388489208633,
+      "step": 10580,
+      "torque_loss": 0.12185729295015335
+    },
+    {
+      "epoch": 9.514388489208633,
+      "force_loss": 0.012036922387778759,
+      "step": 10580
+    },
+    {
+      "epoch": 9.523381294964029,
+      "grad_norm": 0.8121934533119202,
+      "learning_rate": 9.568959414399028e-05,
+      "loss": 0.0274,
+      "step": 10590
+    },
+    {
+      "action_loss": 0.025352761149406433,
+      "epoch": 9.523381294964029,
+      "step": 10590
+    },
+    {
+      "epoch": 9.523381294964029,
+      "step": 10590,
+      "torque_loss": 0.11938933283090591
+    },
+    {
+      "epoch": 9.523381294964029,
+      "force_loss": 0.025092795491218567,
+      "step": 10590
+    },
+    {
+      "epoch": 9.532374100719425,
+      "grad_norm": 0.4860723614692688,
+      "learning_rate": 9.567839369069018e-05,
+      "loss": 0.0289,
+      "step": 10600
+    },
+    {
+      "action_loss": 0.009876593947410583,
+      "epoch": 9.532374100719425,
+      "step": 10600
+    },
+    {
+      "epoch": 9.532374100719425,
+      "step": 10600,
+      "torque_loss": 0.14806832373142242
+    },
+    {
+      "epoch": 9.532374100719425,
+      "force_loss": 0.009916700422763824,
+      "step": 10600
+    },
+    {
+      "epoch": 9.54136690647482,
+      "grad_norm": 0.6804236769676208,
+      "learning_rate": 9.566717936150013e-05,
+      "loss": 0.029,
+      "step": 10610
+    },
+    {
+      "action_loss": 0.021389300003647804,
+      "epoch": 9.54136690647482,
+      "step": 10610
+    },
+    {
+      "epoch": 9.54136690647482,
+      "step": 10610,
+      "torque_loss": 0.17321930825710297
+    },
+    {
+      "epoch": 9.54136690647482,
+      "force_loss": 0.01682678610086441,
+      "step": 10610
+    },
+    {
+      "epoch": 9.550359712230216,
+      "grad_norm": 0.6666664481163025,
+      "learning_rate": 9.565595115982678e-05,
+      "loss": 0.0266,
+      "step": 10620
+    },
+    {
+      "action_loss": 0.009704948402941227,
+      "epoch": 9.550359712230216,
+      "step": 10620
+    },
+    {
+      "epoch": 9.550359712230216,
+      "step": 10620,
+      "torque_loss": 0.14480507373809814
+    },
+    {
+      "epoch": 9.550359712230216,
+      "force_loss": 0.016170455142855644,
+      "step": 10620
+    },
+    {
+      "epoch": 9.559352517985612,
+      "grad_norm": 0.4129558801651001,
+      "learning_rate": 9.564470908908094e-05,
+      "loss": 0.0303,
+      "step": 10630
+    },
+    {
+      "action_loss": 0.008663482964038849,
+      "epoch": 9.559352517985612,
+      "step": 10630
+    },
+    {
+      "epoch": 9.559352517985612,
+      "step": 10630,
+      "torque_loss": 0.10041690617799759
+    },
+    {
+      "epoch": 9.559352517985612,
+      "force_loss": 0.00800513755530119,
+      "step": 10630
+    },
+    {
+      "epoch": 9.568345323741006,
+      "grad_norm": 0.7748432755470276,
+      "learning_rate": 9.563345315267764e-05,
+      "loss": 0.0261,
+      "step": 10640
+    },
+    {
+      "action_loss": 0.02055913209915161,
+      "epoch": 9.568345323741006,
+      "step": 10640
+    },
+    {
+      "epoch": 9.568345323741006,
+      "step": 10640,
+      "torque_loss": 0.17869479954242706
+    },
+    {
+      "epoch": 9.568345323741006,
+      "force_loss": 0.015143473632633686,
+      "step": 10640
+    },
+    {
+      "epoch": 9.577338129496402,
+      "grad_norm": 0.44985678791999817,
+      "learning_rate": 9.562218335403616e-05,
+      "loss": 0.0283,
+      "step": 10650
+    },
+    {
+      "action_loss": 0.008624009788036346,
+      "epoch": 9.577338129496402,
+      "step": 10650
+    },
+    {
+      "epoch": 9.577338129496402,
+      "step": 10650,
+      "torque_loss": 0.17126911878585815
+    },
+    {
+      "epoch": 9.577338129496402,
+      "force_loss": 0.009126664139330387,
+      "step": 10650
+    },
+    {
+      "epoch": 9.586330935251798,
+      "grad_norm": 0.37473663687705994,
+      "learning_rate": 9.561089969657999e-05,
+      "loss": 0.0251,
+      "step": 10660
+    },
+    {
+      "action_loss": 0.02132773958146572,
+      "epoch": 9.586330935251798,
+      "step": 10660
+    },
+    {
+      "epoch": 9.586330935251798,
+      "step": 10660,
+      "torque_loss": 0.15396931767463684
+    },
+    {
+      "epoch": 9.586330935251798,
+      "force_loss": 0.02179611660540104,
+      "step": 10660
+    },
+    {
+      "epoch": 9.595323741007194,
+      "grad_norm": 0.4981321394443512,
+      "learning_rate": 9.559960218373673e-05,
+      "loss": 0.0288,
+      "step": 10670
+    },
+    {
+      "action_loss": 0.01068919152021408,
+      "epoch": 9.595323741007194,
+      "step": 10670
+    },
+    {
+      "epoch": 9.595323741007194,
+      "step": 10670,
+      "torque_loss": 0.11746841669082642
+    },
+    {
+      "epoch": 9.595323741007194,
+      "force_loss": 0.012537163682281971,
+      "step": 10670
+    },
+    {
+      "epoch": 9.60431654676259,
+      "grad_norm": 0.5899496674537659,
+      "learning_rate": 9.558829081893836e-05,
+      "loss": 0.0281,
+      "step": 10680
+    },
+    {
+      "action_loss": 0.014468804001808167,
+      "epoch": 9.60431654676259,
+      "step": 10680
+    },
+    {
+      "epoch": 9.60431654676259,
+      "step": 10680,
+      "torque_loss": 0.11580521613359451
+    },
+    {
+      "epoch": 9.60431654676259,
+      "force_loss": 0.010967371053993702,
+      "step": 10680
+    },
+    {
+      "epoch": 9.613309352517986,
+      "grad_norm": 0.6741436719894409,
+      "learning_rate": 9.55769656056209e-05,
+      "loss": 0.0267,
+      "step": 10690
+    },
+    {
+      "action_loss": 0.007974970154464245,
+      "epoch": 9.613309352517986,
+      "step": 10690
+    },
+    {
+      "epoch": 9.613309352517986,
+      "step": 10690,
+      "torque_loss": 0.11001706123352051
+    },
+    {
+      "epoch": 9.613309352517986,
+      "force_loss": 0.01084264274686575,
+      "step": 10690
+    },
+    {
+      "epoch": 9.622302158273381,
+      "grad_norm": 1.2521754503250122,
+      "learning_rate": 9.556562654722469e-05,
+      "loss": 0.0294,
+      "step": 10700
+    },
+    {
+      "action_loss": 0.010585453361272812,
+      "epoch": 9.622302158273381,
+      "step": 10700
+    },
+    {
+      "epoch": 9.622302158273381,
+      "step": 10700,
+      "torque_loss": 0.13352619111537933
+    },
+    {
+      "epoch": 9.622302158273381,
+      "force_loss": 0.009462296031415462,
+      "step": 10700
+    },
+    {
+      "epoch": 9.631294964028777,
+      "grad_norm": 0.8372606039047241,
+      "learning_rate": 9.555427364719422e-05,
+      "loss": 0.0312,
+      "step": 10710
+    },
+    {
+      "action_loss": 0.014742672443389893,
+      "epoch": 9.631294964028777,
+      "step": 10710
+    },
+    {
+      "epoch": 9.631294964028777,
+      "step": 10710,
+      "torque_loss": 0.11293081194162369
+    },
+    {
+      "epoch": 9.631294964028777,
+      "force_loss": 0.023100676015019417,
+      "step": 10710
+    },
+    {
+      "epoch": 9.640287769784173,
+      "grad_norm": 0.46177878975868225,
+      "learning_rate": 9.55429069089782e-05,
+      "loss": 0.0274,
+      "step": 10720
+    },
+    {
+      "action_loss": 0.019772058352828026,
+      "epoch": 9.640287769784173,
+      "step": 10720
+    },
+    {
+      "epoch": 9.640287769784173,
+      "step": 10720,
+      "torque_loss": 0.2091190367937088
+    },
+    {
+      "epoch": 9.640287769784173,
+      "force_loss": 0.019305391237139702,
+      "step": 10720
+    },
+    {
+      "epoch": 9.649280575539569,
+      "grad_norm": 0.7177428603172302,
+      "learning_rate": 9.553152633602956e-05,
+      "loss": 0.0328,
+      "step": 10730
+    },
+    {
+      "action_loss": 0.01486204657703638,
+      "epoch": 9.649280575539569,
+      "step": 10730
+    },
+    {
+      "epoch": 9.649280575539569,
+      "step": 10730,
+      "torque_loss": 0.14825069904327393
+    },
+    {
+      "epoch": 9.649280575539569,
+      "force_loss": 0.020056938752532005,
+      "step": 10730
+    },
+    {
+      "epoch": 9.658273381294965,
+      "grad_norm": 0.3259560465812683,
+      "learning_rate": 9.552013193180543e-05,
+      "loss": 0.0333,
+      "step": 10740
+    },
+    {
+      "action_loss": 0.008836471475660801,
+      "epoch": 9.658273381294965,
+      "step": 10740
+    },
+    {
+      "epoch": 9.658273381294965,
+      "step": 10740,
+      "torque_loss": 0.15030623972415924
+    },
+    {
+      "epoch": 9.658273381294965,
+      "force_loss": 0.009539115242660046,
+      "step": 10740
+    },
+    {
+      "epoch": 9.667266187050359,
+      "grad_norm": 0.9642545580863953,
+      "learning_rate": 9.550872369976707e-05,
+      "loss": 0.0286,
+      "step": 10750
+    },
+    {
+      "action_loss": 0.0097518190741539,
+      "epoch": 9.667266187050359,
+      "step": 10750
+    },
+    {
+      "epoch": 9.667266187050359,
+      "step": 10750,
+      "torque_loss": 0.14309723675251007
+    },
+    {
+      "epoch": 9.667266187050359,
+      "force_loss": 0.00840467307716608,
+      "step": 10750
+    },
+    {
+      "epoch": 9.676258992805755,
+      "grad_norm": 0.4511452317237854,
+      "learning_rate": 9.549730164338007e-05,
+      "loss": 0.0278,
+      "step": 10760
+    },
+    {
+      "action_loss": 0.01700589247047901,
+      "epoch": 9.676258992805755,
+      "step": 10760
+    },
+    {
+      "epoch": 9.676258992805755,
+      "step": 10760,
+      "torque_loss": 0.145081028342247
+    },
+    {
+      "epoch": 9.676258992805755,
+      "force_loss": 0.008743572980165482,
+      "step": 10760
+    },
+    {
+      "epoch": 9.68525179856115,
+      "grad_norm": 0.4731258451938629,
+      "learning_rate": 9.548586576611408e-05,
+      "loss": 0.028,
+      "step": 10770
+    },
+    {
+      "action_loss": 0.009498675353825092,
+      "epoch": 9.68525179856115,
+      "step": 10770
+    },
+    {
+      "epoch": 9.68525179856115,
+      "step": 10770,
+      "torque_loss": 0.11881691217422485
+    },
+    {
+      "epoch": 9.68525179856115,
+      "force_loss": 0.007521938998252153,
+      "step": 10770
+    },
+    {
+      "epoch": 9.694244604316546,
+      "grad_norm": 0.4556579887866974,
+      "learning_rate": 9.54744160714431e-05,
+      "loss": 0.026,
+      "step": 10780
+    },
+    {
+      "action_loss": 0.009431825019419193,
+      "epoch": 9.694244604316546,
+      "step": 10780
+    },
+    {
+      "epoch": 9.694244604316546,
+      "step": 10780,
+      "torque_loss": 0.1173371970653534
+    },
+    {
+      "epoch": 9.694244604316546,
+      "force_loss": 0.007316796109080315,
+      "step": 10780
+    },
+    {
+      "epoch": 9.703237410071942,
+      "grad_norm": 0.390954852104187,
+      "learning_rate": 9.546295256284516e-05,
+      "loss": 0.0265,
+      "step": 10790
+    },
+    {
+      "action_loss": 0.027017468586564064,
+      "epoch": 9.703237410071942,
+      "step": 10790
+    },
+    {
+      "epoch": 9.703237410071942,
+      "step": 10790,
+      "torque_loss": 0.16717219352722168
+    },
+    {
+      "epoch": 9.703237410071942,
+      "force_loss": 0.030752798542380333,
+      "step": 10790
+    },
+    {
+      "epoch": 9.712230215827338,
+      "grad_norm": 0.4961920976638794,
+      "learning_rate": 9.545147524380265e-05,
+      "loss": 0.0301,
+      "step": 10800
+    },
+    {
+      "action_loss": 0.015549294650554657,
+      "epoch": 9.712230215827338,
+      "step": 10800
+    },
+    {
+      "epoch": 9.712230215827338,
+      "step": 10800,
+      "torque_loss": 0.1894809752702713
+    },
+    {
+      "epoch": 9.712230215827338,
+      "force_loss": 0.022457608953118324,
+      "step": 10800
+    },
+    {
+      "epoch": 9.721223021582734,
+      "grad_norm": 0.6873412728309631,
+      "learning_rate": 9.543998411780201e-05,
+      "loss": 0.0324,
+      "step": 10810
+    },
+    {
+      "action_loss": 0.009921756573021412,
+      "epoch": 9.721223021582734,
+      "step": 10810
+    },
+    {
+      "epoch": 9.721223021582734,
+      "step": 10810,
+      "torque_loss": 0.148454949259758
+    },
+    {
+      "epoch": 9.721223021582734,
+      "force_loss": 0.008466281928122044,
+      "step": 10810
+    },
+    {
+      "epoch": 9.73021582733813,
+      "grad_norm": 0.6092079877853394,
+      "learning_rate": 9.542847918833397e-05,
+      "loss": 0.0273,
+      "step": 10820
+    },
+    {
+      "action_loss": 0.010918482206761837,
+      "epoch": 9.73021582733813,
+      "step": 10820
+    },
+    {
+      "epoch": 9.73021582733813,
+      "step": 10820,
+      "torque_loss": 0.1380562037229538
+    },
+    {
+      "epoch": 9.73021582733813,
+      "force_loss": 0.014418157748878002,
+      "step": 10820
+    },
+    {
+      "epoch": 9.739208633093526,
+      "grad_norm": 0.5506334900856018,
+      "learning_rate": 9.541696045889343e-05,
+      "loss": 0.028,
+      "step": 10830
+    },
+    {
+      "action_loss": 0.009138555265963078,
+      "epoch": 9.739208633093526,
+      "step": 10830
+    },
+    {
+      "epoch": 9.739208633093526,
+      "step": 10830,
+      "torque_loss": 0.12567301094532013
+    },
+    {
+      "epoch": 9.739208633093526,
+      "force_loss": 0.010566767305135727,
+      "step": 10830
+    },
+    {
+      "epoch": 9.748201438848922,
+      "grad_norm": 0.43438073992729187,
+      "learning_rate": 9.540542793297947e-05,
+      "loss": 0.0262,
+      "step": 10840
+    },
+    {
+      "action_loss": 0.011103163473308086,
+      "epoch": 9.748201438848922,
+      "step": 10840
+    },
+    {
+      "epoch": 9.748201438848922,
+      "step": 10840,
+      "torque_loss": 0.18787236511707306
+    },
+    {
+      "epoch": 9.748201438848922,
+      "force_loss": 0.009045661427080631,
+      "step": 10840
+    },
+    {
+      "epoch": 9.757194244604317,
+      "grad_norm": 0.7502126097679138,
+      "learning_rate": 9.539388161409537e-05,
+      "loss": 0.0283,
+      "step": 10850
+    },
+    {
+      "action_loss": 0.010768038220703602,
+      "epoch": 9.757194244604317,
+      "step": 10850
+    },
+    {
+      "epoch": 9.757194244604317,
+      "step": 10850,
+      "torque_loss": 0.1307252198457718
+    },
+    {
+      "epoch": 9.757194244604317,
+      "force_loss": 0.017645349726080894,
+      "step": 10850
+    },
+    {
+      "epoch": 9.766187050359711,
+      "grad_norm": 0.4489460289478302,
+      "learning_rate": 9.538232150574857e-05,
+      "loss": 0.0301,
+      "step": 10860
+    },
+    {
+      "action_loss": 0.008940274827182293,
+      "epoch": 9.766187050359711,
+      "step": 10860
+    },
+    {
+      "epoch": 9.766187050359711,
+      "step": 10860,
+      "torque_loss": 0.14543353021144867
+    },
+    {
+      "epoch": 9.766187050359711,
+      "force_loss": 0.009979513473808765,
+      "step": 10860
+    },
+    {
+      "epoch": 9.775179856115107,
+      "grad_norm": 0.8527157306671143,
+      "learning_rate": 9.537074761145076e-05,
+      "loss": 0.0252,
+      "step": 10870
+    },
+    {
+      "action_loss": 0.012462868355214596,
+      "epoch": 9.775179856115107,
+      "step": 10870
+    },
+    {
+      "epoch": 9.775179856115107,
+      "step": 10870,
+      "torque_loss": 0.1327194720506668
+    },
+    {
+      "epoch": 9.775179856115107,
+      "force_loss": 0.01534614060074091,
+      "step": 10870
+    },
+    {
+      "epoch": 9.784172661870503,
+      "grad_norm": 0.8630487322807312,
+      "learning_rate": 9.535915993471778e-05,
+      "loss": 0.0299,
+      "step": 10880
+    },
+    {
+      "action_loss": 0.01268719881772995,
+      "epoch": 9.784172661870503,
+      "step": 10880
+    },
+    {
+      "epoch": 9.784172661870503,
+      "step": 10880,
+      "torque_loss": 0.15173323452472687
+    },
+    {
+      "epoch": 9.784172661870503,
+      "force_loss": 0.015962107107043266,
+      "step": 10880
+    },
+    {
+      "epoch": 9.793165467625899,
+      "grad_norm": 0.5018957257270813,
+      "learning_rate": 9.534755847906964e-05,
+      "loss": 0.0373,
+      "step": 10890
+    },
+    {
+      "action_loss": 0.008178669027984142,
+      "epoch": 9.793165467625899,
+      "step": 10890
+    },
+    {
+      "epoch": 9.793165467625899,
+      "step": 10890,
+      "torque_loss": 0.1684926152229309
+    },
+    {
+      "epoch": 9.793165467625899,
+      "force_loss": 0.006757232826203108,
+      "step": 10890
+    },
+    {
+      "epoch": 9.802158273381295,
+      "grad_norm": 0.7773537039756775,
+      "learning_rate": 9.533594324803057e-05,
+      "loss": 0.0298,
+      "step": 10900
+    },
+    {
+      "action_loss": 0.012452635914087296,
+      "epoch": 9.802158273381295,
+      "step": 10900
+    },
+    {
+      "epoch": 9.802158273381295,
+      "step": 10900,
+      "torque_loss": 0.1251821219921112
+    },
+    {
+      "epoch": 9.802158273381295,
+      "force_loss": 0.010154018178582191,
+      "step": 10900
+    },
+    {
+      "epoch": 9.81115107913669,
+      "grad_norm": 0.5600789189338684,
+      "learning_rate": 9.532431424512895e-05,
+      "loss": 0.0243,
+      "step": 10910
+    },
+    {
+      "action_loss": 0.013393896631896496,
+      "epoch": 9.81115107913669,
+      "step": 10910
+    },
+    {
+      "epoch": 9.81115107913669,
+      "step": 10910,
+      "torque_loss": 0.15502311289310455
+    },
+    {
+      "epoch": 9.81115107913669,
+      "force_loss": 0.013520360924303532,
+      "step": 10910
+    },
+    {
+      "epoch": 9.820143884892087,
+      "grad_norm": 0.6210010647773743,
+      "learning_rate": 9.531267147389741e-05,
+      "loss": 0.0311,
+      "step": 10920
+    },
+    {
+      "action_loss": 0.017992960289120674,
+      "epoch": 9.820143884892087,
+      "step": 10920
+    },
+    {
+      "epoch": 9.820143884892087,
+      "step": 10920,
+      "torque_loss": 0.15930990874767303
+    },
+    {
+      "epoch": 9.820143884892087,
+      "force_loss": 0.012523858807981014,
+      "step": 10920
+    },
+    {
+      "epoch": 9.829136690647482,
+      "grad_norm": 0.6890802383422852,
+      "learning_rate": 9.530101493787266e-05,
+      "loss": 0.0301,
+      "step": 10930
+    },
+    {
+      "action_loss": 0.009423593990504742,
+      "epoch": 9.829136690647482,
+      "step": 10930
+    },
+    {
+      "epoch": 9.829136690647482,
+      "step": 10930,
+      "torque_loss": 0.11356351524591446
+    },
+    {
+      "epoch": 9.829136690647482,
+      "force_loss": 0.017987238243222237,
+      "step": 10930
+    },
+    {
+      "epoch": 9.838129496402878,
+      "grad_norm": 0.6505790948867798,
+      "learning_rate": 9.528934464059571e-05,
+      "loss": 0.025,
+      "step": 10940
+    },
+    {
+      "action_loss": 0.01121435035020113,
+      "epoch": 9.838129496402878,
+      "step": 10940
+    },
+    {
+      "epoch": 9.838129496402878,
+      "step": 10940,
+      "torque_loss": 0.1927977055311203
+    },
+    {
+      "epoch": 9.838129496402878,
+      "force_loss": 0.011948294937610626,
+      "step": 10940
+    },
+    {
+      "epoch": 9.847122302158274,
+      "grad_norm": 0.5397326350212097,
+      "learning_rate": 9.527766058561163e-05,
+      "loss": 0.0243,
+      "step": 10950
+    },
+    {
+      "action_loss": 0.014351271092891693,
+      "epoch": 9.847122302158274,
+      "step": 10950
+    },
+    {
+      "epoch": 9.847122302158274,
+      "step": 10950,
+      "torque_loss": 0.1684246063232422
+    },
+    {
+      "epoch": 9.847122302158274,
+      "force_loss": 0.009647504426538944,
+      "step": 10950
+    },
+    {
+      "epoch": 9.85611510791367,
+      "grad_norm": 0.9588680267333984,
+      "learning_rate": 9.526596277646976e-05,
+      "loss": 0.0278,
+      "step": 10960
+    },
+    {
+      "action_loss": 0.017118675634264946,
+      "epoch": 9.85611510791367,
+      "step": 10960
+    },
+    {
+      "epoch": 9.85611510791367,
+      "step": 10960,
+      "torque_loss": 0.13536207377910614
+    },
+    {
+      "epoch": 9.85611510791367,
+      "force_loss": 0.019327238202095032,
+      "step": 10960
+    },
+    {
+      "epoch": 9.865107913669064,
+      "grad_norm": 0.6891125440597534,
+      "learning_rate": 9.525425121672358e-05,
+      "loss": 0.0263,
+      "step": 10970
+    },
+    {
+      "action_loss": 0.0067167277447879314,
+      "epoch": 9.865107913669064,
+      "step": 10970
+    },
+    {
+      "epoch": 9.865107913669064,
+      "step": 10970,
+      "torque_loss": 0.16867922246456146
+    },
+    {
+      "epoch": 9.865107913669064,
+      "force_loss": 0.004338176921010017,
+      "step": 10970
+    },
+    {
+      "epoch": 9.87410071942446,
+      "grad_norm": 0.9081276059150696,
+      "learning_rate": 9.524252590993074e-05,
+      "loss": 0.0249,
+      "step": 10980
+    },
+    {
+      "action_loss": 0.008715505711734295,
+      "epoch": 9.87410071942446,
+      "step": 10980
+    },
+    {
+      "epoch": 9.87410071942446,
+      "step": 10980,
+      "torque_loss": 0.129295215010643
+    },
+    {
+      "epoch": 9.87410071942446,
+      "force_loss": 0.013092712499201298,
+      "step": 10980
+    },
+    {
+      "epoch": 9.883093525179856,
+      "grad_norm": 0.43884941935539246,
+      "learning_rate": 9.523078685965309e-05,
+      "loss": 0.0248,
+      "step": 10990
+    },
+    {
+      "action_loss": 0.00759974867105484,
+      "epoch": 9.883093525179856,
+      "step": 10990
+    },
+    {
+      "epoch": 9.883093525179856,
+      "step": 10990,
+      "torque_loss": 0.16198456287384033
+    },
+    {
+      "epoch": 9.883093525179856,
+      "force_loss": 0.006631703581660986,
+      "step": 10990
+    },
+    {
+      "epoch": 9.892086330935252,
+      "grad_norm": 0.5851729512214661,
+      "learning_rate": 9.521903406945664e-05,
+      "loss": 0.0267,
+      "step": 11000
+    },
+    {
+      "action_loss": 0.009170866571366787,
+      "epoch": 9.892086330935252,
+      "step": 11000
+    },
+    {
+      "epoch": 9.892086330935252,
+      "step": 11000,
+      "torque_loss": 0.08510661125183105
+    },
+    {
+      "epoch": 9.892086330935252,
+      "force_loss": 0.018143830820918083,
+      "step": 11000
+    },
+    {
+      "epoch": 9.901079136690647,
+      "grad_norm": 0.704824686050415,
+      "learning_rate": 9.520726754291158e-05,
+      "loss": 0.027,
+      "step": 11010
+    },
+    {
+      "action_loss": 0.017982706427574158,
+      "epoch": 9.901079136690647,
+      "step": 11010
+    },
+    {
+      "epoch": 9.901079136690647,
+      "step": 11010,
+      "torque_loss": 0.10604137182235718
+    },
+    {
+      "epoch": 9.901079136690647,
+      "force_loss": 0.019240552559494972,
+      "step": 11010
+    },
+    {
+      "epoch": 9.910071942446043,
+      "grad_norm": 0.6339232325553894,
+      "learning_rate": 9.519548728359227e-05,
+      "loss": 0.0279,
+      "step": 11020
+    },
+    {
+      "action_loss": 0.010949399322271347,
+      "epoch": 9.910071942446043,
+      "step": 11020
+    },
+    {
+      "epoch": 9.910071942446043,
+      "step": 11020,
+      "torque_loss": 0.14291070401668549
+    },
+    {
+      "epoch": 9.910071942446043,
+      "force_loss": 0.010003216564655304,
+      "step": 11020
+    },
+    {
+      "epoch": 9.91906474820144,
+      "grad_norm": 0.7300173044204712,
+      "learning_rate": 9.518369329507726e-05,
+      "loss": 0.0238,
+      "step": 11030
+    },
+    {
+      "action_loss": 0.011186365969479084,
+      "epoch": 9.91906474820144,
+      "step": 11030
+    },
+    {
+      "epoch": 9.91906474820144,
+      "step": 11030,
+      "torque_loss": 0.13138091564178467
+    },
+    {
+      "epoch": 9.91906474820144,
+      "force_loss": 0.01066580694168806,
+      "step": 11030
+    },
+    {
+      "epoch": 9.928057553956835,
+      "grad_norm": 0.7806556820869446,
+      "learning_rate": 9.51718855809492e-05,
+      "loss": 0.0249,
+      "step": 11040
+    },
+    {
+      "action_loss": 0.009963017888367176,
+      "epoch": 9.928057553956835,
+      "step": 11040
+    },
+    {
+      "epoch": 9.928057553956835,
+      "step": 11040,
+      "torque_loss": 0.16646207869052887
+    },
+    {
+      "epoch": 9.928057553956835,
+      "force_loss": 0.010010754689574242,
+      "step": 11040
+    },
+    {
+      "epoch": 9.93705035971223,
+      "grad_norm": 0.4272620975971222,
+      "learning_rate": 9.516006414479502e-05,
+      "loss": 0.0304,
+      "step": 11050
+    },
+    {
+      "action_loss": 0.010648422874510288,
+      "epoch": 9.93705035971223,
+      "step": 11050
+    },
+    {
+      "epoch": 9.93705035971223,
+      "step": 11050,
+      "torque_loss": 0.1669343262910843
+    },
+    {
+      "epoch": 9.93705035971223,
+      "force_loss": 0.011715386062860489,
+      "step": 11050
+    },
+    {
+      "epoch": 9.946043165467627,
+      "grad_norm": 0.7803710699081421,
+      "learning_rate": 9.514822899020572e-05,
+      "loss": 0.0385,
+      "step": 11060
+    },
+    {
+      "action_loss": 0.01156608760356903,
+      "epoch": 9.946043165467627,
+      "step": 11060
+    },
+    {
+      "epoch": 9.946043165467627,
+      "step": 11060,
+      "torque_loss": 0.11944753676652908
+    },
+    {
+      "epoch": 9.946043165467627,
+      "force_loss": 0.010257956571877003,
+      "step": 11060
+    },
+    {
+      "epoch": 9.95503597122302,
+      "grad_norm": 0.46194639801979065,
+      "learning_rate": 9.513638012077654e-05,
+      "loss": 0.0296,
+      "step": 11070
+    },
+    {
+      "action_loss": 0.01348278671503067,
+      "epoch": 9.95503597122302,
+      "step": 11070
+    },
+    {
+      "epoch": 9.95503597122302,
+      "step": 11070,
+      "torque_loss": 0.15218187868595123
+    },
+    {
+      "epoch": 9.95503597122302,
+      "force_loss": 0.010267366655170918,
+      "step": 11070
+    },
+    {
+      "epoch": 9.964028776978417,
+      "grad_norm": 0.436085045337677,
+      "learning_rate": 9.512451754010683e-05,
+      "loss": 0.0261,
+      "step": 11080
+    },
+    {
+      "action_loss": 0.012708738446235657,
+      "epoch": 9.964028776978417,
+      "step": 11080
+    },
+    {
+      "epoch": 9.964028776978417,
+      "step": 11080,
+      "torque_loss": 0.1555136889219284
+    },
+    {
+      "epoch": 9.964028776978417,
+      "force_loss": 0.009203293360769749,
+      "step": 11080
+    },
+    {
+      "epoch": 9.973021582733812,
+      "grad_norm": 0.6861566305160522,
+      "learning_rate": 9.511264125180013e-05,
+      "loss": 0.0249,
+      "step": 11090
+    },
+    {
+      "action_loss": 0.01185446698218584,
+      "epoch": 9.973021582733812,
+      "step": 11090
+    },
+    {
+      "epoch": 9.973021582733812,
+      "step": 11090,
+      "torque_loss": 0.16799135506153107
+    },
+    {
+      "epoch": 9.973021582733812,
+      "force_loss": 0.01190869975835085,
+      "step": 11090
+    },
+    {
+      "epoch": 9.982014388489208,
+      "grad_norm": 0.5024900436401367,
+      "learning_rate": 9.510075125946414e-05,
+      "loss": 0.0268,
+      "step": 11100
+    },
+    {
+      "action_loss": 0.015929432585835457,
+      "epoch": 9.982014388489208,
+      "step": 11100
+    },
+    {
+      "epoch": 9.982014388489208,
+      "step": 11100,
+      "torque_loss": 0.1766018122434616
+    },
+    {
+      "epoch": 9.982014388489208,
+      "force_loss": 0.013793278485536575,
+      "step": 11100
+    },
+    {
+      "epoch": 9.991007194244604,
+      "grad_norm": 0.5480630397796631,
+      "learning_rate": 9.508884756671075e-05,
+      "loss": 0.0291,
+      "step": 11110
+    },
+    {
+      "action_loss": 0.008085153996944427,
+      "epoch": 9.991007194244604,
+      "step": 11110
+    },
+    {
+      "epoch": 9.991007194244604,
+      "step": 11110,
+      "torque_loss": 0.15915803611278534
+    },
+    {
+      "epoch": 9.991007194244604,
+      "force_loss": 0.006082264240831137,
+      "step": 11110
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 0.7992289066314697,
+      "learning_rate": 9.507693017715596e-05,
+      "loss": 0.0286,
+      "step": 11120
+    },
+    {
+      "action_loss": 0.004015494138002396,
+      "epoch": 10.0,
+      "step": 11120
+    },
+    {
+      "epoch": 10.0,
+      "step": 11120,
+      "torque_loss": 0.08135860413312912
+    },
+    {
+      "epoch": 10.0,
+      "force_loss": 0.004399992059916258,
+      "step": 11120
+    },
+    {
+      "epoch": 10.008992805755396,
+      "grad_norm": 0.3683903217315674,
+      "learning_rate": 9.506499909441997e-05,
+      "loss": 0.0299,
+      "step": 11130
+    },
+    {
+      "action_loss": 0.00492004444822669,
+      "epoch": 10.008992805755396,
+      "step": 11130
+    },
+    {
+      "epoch": 10.008992805755396,
+      "step": 11130,
+      "torque_loss": 0.11255601048469543
+    },
+    {
+      "epoch": 10.008992805755396,
+      "force_loss": 0.004956409800797701,
+      "step": 11130
+    },
+    {
+      "epoch": 10.017985611510792,
+      "grad_norm": 0.7232905030250549,
+      "learning_rate": 9.505305432212713e-05,
+      "loss": 0.0248,
+      "step": 11140
+    },
+    {
+      "action_loss": 0.011769063770771027,
+      "epoch": 10.017985611510792,
+      "step": 11140
+    },
+    {
+      "epoch": 10.017985611510792,
+      "step": 11140,
+      "torque_loss": 0.11822396516799927
+    },
+    {
+      "epoch": 10.017985611510792,
+      "force_loss": 0.019234588369727135,
+      "step": 11140
+    },
+    {
+      "epoch": 10.026978417266188,
+      "grad_norm": 0.5520509481430054,
+      "learning_rate": 9.504109586390595e-05,
+      "loss": 0.026,
+      "step": 11150
+    },
+    {
+      "action_loss": 0.0076273479498922825,
+      "epoch": 10.026978417266188,
+      "step": 11150
+    },
+    {
+      "epoch": 10.026978417266188,
+      "step": 11150,
+      "torque_loss": 0.12028899788856506
+    },
+    {
+      "epoch": 10.026978417266188,
+      "force_loss": 0.0052971444092690945,
+      "step": 11150
+    },
+    {
+      "epoch": 10.035971223021583,
+      "grad_norm": 0.39104026556015015,
+      "learning_rate": 9.502912372338908e-05,
+      "loss": 0.0301,
+      "step": 11160
+    },
+    {
+      "action_loss": 0.007534921169281006,
+      "epoch": 10.035971223021583,
+      "step": 11160
+    },
+    {
+      "epoch": 10.035971223021583,
+      "step": 11160,
+      "torque_loss": 0.14494585990905762
+    },
+    {
+      "epoch": 10.035971223021583,
+      "force_loss": 0.008034944534301758,
+      "step": 11160
+    },
+    {
+      "epoch": 10.04496402877698,
+      "grad_norm": 0.744063675403595,
+      "learning_rate": 9.501713790421335e-05,
+      "loss": 0.0294,
+      "step": 11170
+    },
+    {
+      "action_loss": 0.006292261183261871,
+      "epoch": 10.04496402877698,
+      "step": 11170
+    },
+    {
+      "epoch": 10.04496402877698,
+      "step": 11170,
+      "torque_loss": 0.14006108045578003
+    },
+    {
+      "epoch": 10.04496402877698,
+      "force_loss": 0.005716968793421984,
+      "step": 11170
+    },
+    {
+      "epoch": 10.053956834532373,
+      "grad_norm": 1.0360183715820312,
+      "learning_rate": 9.500513841001974e-05,
+      "loss": 0.0277,
+      "step": 11180
+    },
+    {
+      "action_loss": 0.019556446000933647,
+      "epoch": 10.053956834532373,
+      "step": 11180
+    },
+    {
+      "epoch": 10.053956834532373,
+      "step": 11180,
+      "torque_loss": 0.17862176895141602
+    },
+    {
+      "epoch": 10.053956834532373,
+      "force_loss": 0.016817452386021614,
+      "step": 11180
+    },
+    {
+      "epoch": 10.06294964028777,
+      "grad_norm": 0.5944479703903198,
+      "learning_rate": 9.499312524445336e-05,
+      "loss": 0.0323,
+      "step": 11190
+    },
+    {
+      "action_loss": 0.005233692470937967,
+      "epoch": 10.06294964028777,
+      "step": 11190
+    },
+    {
+      "epoch": 10.06294964028777,
+      "step": 11190,
+      "torque_loss": 0.11514333635568619
+    },
+    {
+      "epoch": 10.06294964028777,
+      "force_loss": 0.005665199365466833,
+      "step": 11190
+    },
+    {
+      "epoch": 10.071942446043165,
+      "grad_norm": 0.32586511969566345,
+      "learning_rate": 9.498109841116351e-05,
+      "loss": 0.028,
+      "step": 11200
+    },
+    {
+      "action_loss": 0.016690149903297424,
+      "epoch": 10.071942446043165,
+      "step": 11200
+    },
+    {
+      "epoch": 10.071942446043165,
+      "step": 11200,
+      "torque_loss": 0.15541528165340424
+    },
+    {
+      "epoch": 10.071942446043165,
+      "force_loss": 0.017409134656190872,
+      "step": 11200
+    },
+    {
+      "epoch": 10.08093525179856,
+      "grad_norm": 0.5725913047790527,
+      "learning_rate": 9.496905791380363e-05,
+      "loss": 0.0366,
+      "step": 11210
+    },
+    {
+      "action_loss": 0.014832686632871628,
+      "epoch": 10.08093525179856,
+      "step": 11210
+    },
+    {
+      "epoch": 10.08093525179856,
+      "step": 11210,
+      "torque_loss": 0.13739167153835297
+    },
+    {
+      "epoch": 10.08093525179856,
+      "force_loss": 0.01938687637448311,
+      "step": 11210
+    },
+    {
+      "epoch": 10.089928057553957,
+      "grad_norm": 0.4827929437160492,
+      "learning_rate": 9.495700375603129e-05,
+      "loss": 0.0276,
+      "step": 11220
+    },
+    {
+      "action_loss": 0.014541879296302795,
+      "epoch": 10.089928057553957,
+      "step": 11220
+    },
+    {
+      "epoch": 10.089928057553957,
+      "step": 11220,
+      "torque_loss": 0.17996855080127716
+    },
+    {
+      "epoch": 10.089928057553957,
+      "force_loss": 0.014629076234996319,
+      "step": 11220
+    },
+    {
+      "epoch": 10.098920863309353,
+      "grad_norm": 0.6409385800361633,
+      "learning_rate": 9.494493594150822e-05,
+      "loss": 0.0308,
+      "step": 11230
+    },
+    {
+      "action_loss": 0.008698993362486362,
+      "epoch": 10.098920863309353,
+      "step": 11230
+    },
+    {
+      "epoch": 10.098920863309353,
+      "step": 11230,
+      "torque_loss": 0.09485579282045364
+    },
+    {
+      "epoch": 10.098920863309353,
+      "force_loss": 0.007293974980711937,
+      "step": 11230
+    },
+    {
+      "epoch": 10.107913669064748,
+      "grad_norm": 0.9839959740638733,
+      "learning_rate": 9.493285447390032e-05,
+      "loss": 0.0312,
+      "step": 11240
+    },
+    {
+      "action_loss": 0.013759724795818329,
+      "epoch": 10.107913669064748,
+      "step": 11240
+    },
+    {
+      "epoch": 10.107913669064748,
+      "step": 11240,
+      "torque_loss": 0.14273864030838013
+    },
+    {
+      "epoch": 10.107913669064748,
+      "force_loss": 0.01785869710147381,
+      "step": 11240
+    },
+    {
+      "epoch": 10.116906474820144,
+      "grad_norm": 0.3355298936367035,
+      "learning_rate": 9.492075935687761e-05,
+      "loss": 0.0274,
+      "step": 11250
+    },
+    {
+      "action_loss": 0.01849684678018093,
+      "epoch": 10.116906474820144,
+      "step": 11250
+    },
+    {
+      "epoch": 10.116906474820144,
+      "step": 11250,
+      "torque_loss": 0.10407856851816177
+    },
+    {
+      "epoch": 10.116906474820144,
+      "force_loss": 0.007260418031364679,
+      "step": 11250
+    },
+    {
+      "epoch": 10.12589928057554,
+      "grad_norm": 0.34183377027511597,
+      "learning_rate": 9.490865059411427e-05,
+      "loss": 0.0261,
+      "step": 11260
+    },
+    {
+      "action_loss": 0.06025155261158943,
+      "epoch": 10.12589928057554,
+      "step": 11260
+    },
+    {
+      "epoch": 10.12589928057554,
+      "step": 11260,
+      "torque_loss": 0.17157061398029327
+    },
+    {
+      "epoch": 10.12589928057554,
+      "force_loss": 0.06326596438884735,
+      "step": 11260
+    },
+    {
+      "epoch": 10.134892086330936,
+      "grad_norm": 0.33712413907051086,
+      "learning_rate": 9.489652818928863e-05,
+      "loss": 0.0277,
+      "step": 11270
+    },
+    {
+      "action_loss": 0.02074318192899227,
+      "epoch": 10.134892086330936,
+      "step": 11270
+    },
+    {
+      "epoch": 10.134892086330936,
+      "step": 11270,
+      "torque_loss": 0.1118684709072113
+    },
+    {
+      "epoch": 10.134892086330936,
+      "force_loss": 0.02923084795475006,
+      "step": 11270
+    },
+    {
+      "epoch": 10.14388489208633,
+      "grad_norm": 0.5187804102897644,
+      "learning_rate": 9.488439214608315e-05,
+      "loss": 0.0394,
+      "step": 11280
+    },
+    {
+      "action_loss": 0.015663141384720802,
+      "epoch": 10.14388489208633,
+      "step": 11280
+    },
+    {
+      "epoch": 10.14388489208633,
+      "step": 11280,
+      "torque_loss": 0.18274737894535065
+    },
+    {
+      "epoch": 10.14388489208633,
+      "force_loss": 0.014046695083379745,
+      "step": 11280
+    },
+    {
+      "epoch": 10.152877697841726,
+      "grad_norm": 0.49942636489868164,
+      "learning_rate": 9.487224246818444e-05,
+      "loss": 0.0254,
+      "step": 11290
+    },
+    {
+      "action_loss": 0.014512017369270325,
+      "epoch": 10.152877697841726,
+      "step": 11290
+    },
+    {
+      "epoch": 10.152877697841726,
+      "step": 11290,
+      "torque_loss": 0.16143356263637543
+    },
+    {
+      "epoch": 10.152877697841726,
+      "force_loss": 0.012113831005990505,
+      "step": 11290
+    },
+    {
+      "epoch": 10.161870503597122,
+      "grad_norm": 0.33942297101020813,
+      "learning_rate": 9.486007915928325e-05,
+      "loss": 0.0263,
+      "step": 11300
+    },
+    {
+      "action_loss": 0.008723310194909573,
+      "epoch": 10.161870503597122,
+      "step": 11300
+    },
+    {
+      "epoch": 10.161870503597122,
+      "step": 11300,
+      "torque_loss": 0.18283824622631073
+    },
+    {
+      "epoch": 10.161870503597122,
+      "force_loss": 0.006449391599744558,
+      "step": 11300
+    },
+    {
+      "epoch": 10.170863309352518,
+      "grad_norm": 0.6692873239517212,
+      "learning_rate": 9.484790222307448e-05,
+      "loss": 0.0277,
+      "step": 11310
+    },
+    {
+      "action_loss": 0.003828558139503002,
+      "epoch": 10.170863309352518,
+      "step": 11310
+    },
+    {
+      "epoch": 10.170863309352518,
+      "step": 11310,
+      "torque_loss": 0.1257026344537735
+    },
+    {
+      "epoch": 10.170863309352518,
+      "force_loss": 0.005450219381600618,
+      "step": 11310
+    },
+    {
+      "epoch": 10.179856115107913,
+      "grad_norm": 0.7476741075515747,
+      "learning_rate": 9.483571166325716e-05,
+      "loss": 0.0313,
+      "step": 11320
+    },
+    {
+      "action_loss": 0.010637554340064526,
+      "epoch": 10.179856115107913,
+      "step": 11320
+    },
+    {
+      "epoch": 10.179856115107913,
+      "step": 11320,
+      "torque_loss": 0.13662751019001007
+    },
+    {
+      "epoch": 10.179856115107913,
+      "force_loss": 0.010956441052258015,
+      "step": 11320
+    },
+    {
+      "epoch": 10.18884892086331,
+      "grad_norm": 0.7356482744216919,
+      "learning_rate": 9.482350748353444e-05,
+      "loss": 0.0281,
+      "step": 11330
+    },
+    {
+      "action_loss": 0.006456850096583366,
+      "epoch": 10.18884892086331,
+      "step": 11330
+    },
+    {
+      "epoch": 10.18884892086331,
+      "step": 11330,
+      "torque_loss": 0.11050192266702652
+    },
+    {
+      "epoch": 10.18884892086331,
+      "force_loss": 0.007361396681517363,
+      "step": 11330
+    },
+    {
+      "epoch": 10.197841726618705,
+      "grad_norm": 0.5127983689308167,
+      "learning_rate": 9.481128968761363e-05,
+      "loss": 0.0264,
+      "step": 11340
+    },
+    {
+      "action_loss": 0.010660975240170956,
+      "epoch": 10.197841726618705,
+      "step": 11340
+    },
+    {
+      "epoch": 10.197841726618705,
+      "step": 11340,
+      "torque_loss": 0.12635789811611176
+    },
+    {
+      "epoch": 10.197841726618705,
+      "force_loss": 0.0071951462887227535,
+      "step": 11340
+    },
+    {
+      "epoch": 10.206834532374101,
+      "grad_norm": 0.5391913056373596,
+      "learning_rate": 9.479905827920621e-05,
+      "loss": 0.0263,
+      "step": 11350
+    },
+    {
+      "action_loss": 0.016435103490948677,
+      "epoch": 10.206834532374101,
+      "step": 11350
+    },
+    {
+      "epoch": 10.206834532374101,
+      "step": 11350,
+      "torque_loss": 0.13317665457725525
+    },
+    {
+      "epoch": 10.206834532374101,
+      "force_loss": 0.01528336200863123,
+      "step": 11350
+    },
+    {
+      "epoch": 10.215827338129497,
+      "grad_norm": 0.8430057764053345,
+      "learning_rate": 9.478681326202773e-05,
+      "loss": 0.0296,
+      "step": 11360
+    },
+    {
+      "action_loss": 0.012869472615420818,
+      "epoch": 10.215827338129497,
+      "step": 11360
+    },
+    {
+      "epoch": 10.215827338129497,
+      "step": 11360,
+      "torque_loss": 0.15476185083389282
+    },
+    {
+      "epoch": 10.215827338129497,
+      "force_loss": 0.013207174837589264,
+      "step": 11360
+    },
+    {
+      "epoch": 10.224820143884893,
+      "grad_norm": 0.5041043162345886,
+      "learning_rate": 9.477455463979791e-05,
+      "loss": 0.0268,
+      "step": 11370
+    },
+    {
+      "action_loss": 0.007422102149575949,
+      "epoch": 10.224820143884893,
+      "step": 11370
+    },
+    {
+      "epoch": 10.224820143884893,
+      "step": 11370,
+      "torque_loss": 0.11074385792016983
+    },
+    {
+      "epoch": 10.224820143884893,
+      "force_loss": 0.004675613716244698,
+      "step": 11370
+    },
+    {
+      "epoch": 10.233812949640289,
+      "grad_norm": 0.9896944761276245,
+      "learning_rate": 9.476228241624059e-05,
+      "loss": 0.0357,
+      "step": 11380
+    },
+    {
+      "action_loss": 0.008223764598369598,
+      "epoch": 10.233812949640289,
+      "step": 11380
+    },
+    {
+      "epoch": 10.233812949640289,
+      "step": 11380,
+      "torque_loss": 0.12861405313014984
+    },
+    {
+      "epoch": 10.233812949640289,
+      "force_loss": 0.006207530852407217,
+      "step": 11380
+    },
+    {
+      "epoch": 10.242805755395683,
+      "grad_norm": 0.5693405866622925,
+      "learning_rate": 9.474999659508374e-05,
+      "loss": 0.0273,
+      "step": 11390
+    },
+    {
+      "action_loss": 0.005569235887378454,
+      "epoch": 10.242805755395683,
+      "step": 11390
+    },
+    {
+      "epoch": 10.242805755395683,
+      "step": 11390,
+      "torque_loss": 0.09525828808546066
+    },
+    {
+      "epoch": 10.242805755395683,
+      "force_loss": 0.0063684540800750256,
+      "step": 11390
+    },
+    {
+      "epoch": 10.251798561151078,
+      "grad_norm": 0.4121360182762146,
+      "learning_rate": 9.47376971800595e-05,
+      "loss": 0.0237,
+      "step": 11400
+    },
+    {
+      "action_loss": 0.006347741931676865,
+      "epoch": 10.251798561151078,
+      "step": 11400
+    },
+    {
+      "epoch": 10.251798561151078,
+      "step": 11400,
+      "torque_loss": 0.09575831890106201
+    },
+    {
+      "epoch": 10.251798561151078,
+      "force_loss": 0.004907703958451748,
+      "step": 11400
+    },
+    {
+      "epoch": 10.260791366906474,
+      "grad_norm": 0.5295513272285461,
+      "learning_rate": 9.472538417490409e-05,
+      "loss": 0.0323,
+      "step": 11410
+    },
+    {
+      "action_loss": 0.007348763290792704,
+      "epoch": 10.260791366906474,
+      "step": 11410
+    },
+    {
+      "epoch": 10.260791366906474,
+      "step": 11410,
+      "torque_loss": 0.10169955343008041
+    },
+    {
+      "epoch": 10.260791366906474,
+      "force_loss": 0.0060578309930861,
+      "step": 11410
+    },
+    {
+      "epoch": 10.26978417266187,
+      "grad_norm": 0.3044617772102356,
+      "learning_rate": 9.471305758335784e-05,
+      "loss": 0.0286,
+      "step": 11420
+    },
+    {
+      "action_loss": 0.005669146776199341,
+      "epoch": 10.26978417266187,
+      "step": 11420
+    },
+    {
+      "epoch": 10.26978417266187,
+      "step": 11420,
+      "torque_loss": 0.12416287511587143
+    },
+    {
+      "epoch": 10.26978417266187,
+      "force_loss": 0.00506193283945322,
+      "step": 11420
+    },
+    {
+      "epoch": 10.278776978417266,
+      "grad_norm": 0.4337250888347626,
+      "learning_rate": 9.47007174091653e-05,
+      "loss": 0.0345,
+      "step": 11430
+    },
+    {
+      "action_loss": 0.01646108366549015,
+      "epoch": 10.278776978417266,
+      "step": 11430
+    },
+    {
+      "epoch": 10.278776978417266,
+      "step": 11430,
+      "torque_loss": 0.1496351808309555
+    },
+    {
+      "epoch": 10.278776978417266,
+      "force_loss": 0.008551005274057388,
+      "step": 11430
+    },
+    {
+      "epoch": 10.287769784172662,
+      "grad_norm": 0.39356017112731934,
+      "learning_rate": 9.468836365607507e-05,
+      "loss": 0.032,
+      "step": 11440
+    },
+    {
+      "action_loss": 0.021052265539765358,
+      "epoch": 10.287769784172662,
+      "step": 11440
+    },
+    {
+      "epoch": 10.287769784172662,
+      "step": 11440,
+      "torque_loss": 0.22915327548980713
+    },
+    {
+      "epoch": 10.287769784172662,
+      "force_loss": 0.01646186225116253,
+      "step": 11440
+    },
+    {
+      "epoch": 10.296762589928058,
+      "grad_norm": 1.0541752576828003,
+      "learning_rate": 9.467599632783988e-05,
+      "loss": 0.0338,
+      "step": 11450
+    },
+    {
+      "action_loss": 0.009766120463609695,
+      "epoch": 10.296762589928058,
+      "step": 11450
+    },
+    {
+      "epoch": 10.296762589928058,
+      "step": 11450,
+      "torque_loss": 0.15463463962078094
+    },
+    {
+      "epoch": 10.296762589928058,
+      "force_loss": 0.016685975715517998,
+      "step": 11450
+    },
+    {
+      "epoch": 10.305755395683454,
+      "grad_norm": 0.7311800718307495,
+      "learning_rate": 9.466361542821662e-05,
+      "loss": 0.0293,
+      "step": 11460
+    },
+    {
+      "action_loss": 0.012333345599472523,
+      "epoch": 10.305755395683454,
+      "step": 11460
+    },
+    {
+      "epoch": 10.305755395683454,
+      "step": 11460,
+      "torque_loss": 0.1635270118713379
+    },
+    {
+      "epoch": 10.305755395683454,
+      "force_loss": 0.010200919583439827,
+      "step": 11460
+    },
+    {
+      "epoch": 10.31474820143885,
+      "grad_norm": 0.33149927854537964,
+      "learning_rate": 9.465122096096625e-05,
+      "loss": 0.0248,
+      "step": 11470
+    },
+    {
+      "action_loss": 0.012985029257833958,
+      "epoch": 10.31474820143885,
+      "step": 11470
+    },
+    {
+      "epoch": 10.31474820143885,
+      "step": 11470,
+      "torque_loss": 0.1617441475391388
+    },
+    {
+      "epoch": 10.31474820143885,
+      "force_loss": 0.01682000607252121,
+      "step": 11470
+    },
+    {
+      "epoch": 10.323741007194245,
+      "grad_norm": 0.7311398386955261,
+      "learning_rate": 9.463881292985391e-05,
+      "loss": 0.0304,
+      "step": 11480
+    },
+    {
+      "action_loss": 0.011703367345035076,
+      "epoch": 10.323741007194245,
+      "step": 11480
+    },
+    {
+      "epoch": 10.323741007194245,
+      "step": 11480,
+      "torque_loss": 0.1737755686044693
+    },
+    {
+      "epoch": 10.323741007194245,
+      "force_loss": 0.015321140177547932,
+      "step": 11480
+    },
+    {
+      "epoch": 10.332733812949641,
+      "grad_norm": 0.6987825632095337,
+      "learning_rate": 9.462639133864881e-05,
+      "loss": 0.026,
+      "step": 11490
+    },
+    {
+      "action_loss": 0.008947235532104969,
+      "epoch": 10.332733812949641,
+      "step": 11490
+    },
+    {
+      "epoch": 10.332733812949641,
+      "step": 11490,
+      "torque_loss": 0.18349778652191162
+    },
+    {
+      "epoch": 10.332733812949641,
+      "force_loss": 0.011550438590347767,
+      "step": 11490
+    },
+    {
+      "epoch": 10.341726618705035,
+      "grad_norm": 1.3858733177185059,
+      "learning_rate": 9.461395619112432e-05,
+      "loss": 0.0272,
+      "step": 11500
+    },
+    {
+      "action_loss": 0.008816898800432682,
+      "epoch": 10.341726618705035,
+      "step": 11500
+    },
+    {
+      "epoch": 10.341726618705035,
+      "step": 11500,
+      "torque_loss": 0.14297789335250854
+    },
+    {
+      "epoch": 10.341726618705035,
+      "force_loss": 0.007423428352922201,
+      "step": 11500
+    },
+    {
+      "epoch": 10.350719424460431,
+      "grad_norm": 0.8188197016716003,
+      "learning_rate": 9.460150749105791e-05,
+      "loss": 0.0233,
+      "step": 11510
+    },
+    {
+      "action_loss": 0.008732032030820847,
+      "epoch": 10.350719424460431,
+      "step": 11510
+    },
+    {
+      "epoch": 10.350719424460431,
+      "step": 11510,
+      "torque_loss": 0.17078770697116852
+    },
+    {
+      "epoch": 10.350719424460431,
+      "force_loss": 0.011270567774772644,
+      "step": 11510
+    },
+    {
+      "epoch": 10.359712230215827,
+      "grad_norm": 0.7546162605285645,
+      "learning_rate": 9.458904524223116e-05,
+      "loss": 0.0298,
+      "step": 11520
+    },
+    {
+      "action_loss": 0.021800050511956215,
+      "epoch": 10.359712230215827,
+      "step": 11520
+    },
+    {
+      "epoch": 10.359712230215827,
+      "step": 11520,
+      "torque_loss": 0.1783391833305359
+    },
+    {
+      "epoch": 10.359712230215827,
+      "force_loss": 0.012747381813824177,
+      "step": 11520
+    },
+    {
+      "epoch": 10.368705035971223,
+      "grad_norm": 1.0759632587432861,
+      "learning_rate": 9.457656944842976e-05,
+      "loss": 0.0319,
+      "step": 11530
+    },
+    {
+      "action_loss": 0.007475728634744883,
+      "epoch": 10.368705035971223,
+      "step": 11530
+    },
+    {
+      "epoch": 10.368705035971223,
+      "step": 11530,
+      "torque_loss": 0.16289037466049194
+    },
+    {
+      "epoch": 10.368705035971223,
+      "force_loss": 0.006704117637127638,
+      "step": 11530
+    },
+    {
+      "epoch": 10.377697841726619,
+      "grad_norm": 0.9225375056266785,
+      "learning_rate": 9.456408011344353e-05,
+      "loss": 0.0248,
+      "step": 11540
+    },
+    {
+      "action_loss": 0.012882369570434093,
+      "epoch": 10.377697841726619,
+      "step": 11540
+    },
+    {
+      "epoch": 10.377697841726619,
+      "step": 11540,
+      "torque_loss": 0.1400008648633957
+    },
+    {
+      "epoch": 10.377697841726619,
+      "force_loss": 0.00630096485838294,
+      "step": 11540
+    },
+    {
+      "epoch": 10.386690647482014,
+      "grad_norm": 0.4783977270126343,
+      "learning_rate": 9.455157724106643e-05,
+      "loss": 0.03,
+      "step": 11550
+    },
+    {
+      "action_loss": 0.0238427072763443,
+      "epoch": 10.386690647482014,
+      "step": 11550
+    },
+    {
+      "epoch": 10.386690647482014,
+      "step": 11550,
+      "torque_loss": 0.22121857106685638
+    },
+    {
+      "epoch": 10.386690647482014,
+      "force_loss": 0.034579694271087646,
+      "step": 11550
+    },
+    {
+      "epoch": 10.39568345323741,
+      "grad_norm": 0.8244315981864929,
+      "learning_rate": 9.453906083509647e-05,
+      "loss": 0.0318,
+      "step": 11560
+    },
+    {
+      "action_loss": 0.014974999241530895,
+      "epoch": 10.39568345323741,
+      "step": 11560
+    },
+    {
+      "epoch": 10.39568345323741,
+      "step": 11560,
+      "torque_loss": 0.11194352060556412
+    },
+    {
+      "epoch": 10.39568345323741,
+      "force_loss": 0.018318871036171913,
+      "step": 11560
+    },
+    {
+      "epoch": 10.404676258992806,
+      "grad_norm": 0.7440147995948792,
+      "learning_rate": 9.45265308993358e-05,
+      "loss": 0.0239,
+      "step": 11570
+    },
+    {
+      "action_loss": 0.037725578993558884,
+      "epoch": 10.404676258992806,
+      "step": 11570
+    },
+    {
+      "epoch": 10.404676258992806,
+      "step": 11570,
+      "torque_loss": 0.2239779233932495
+    },
+    {
+      "epoch": 10.404676258992806,
+      "force_loss": 0.04689591005444527,
+      "step": 11570
+    },
+    {
+      "epoch": 10.413669064748202,
+      "grad_norm": 0.34233060479164124,
+      "learning_rate": 9.451398743759071e-05,
+      "loss": 0.0294,
+      "step": 11580
+    },
+    {
+      "action_loss": 0.017021184787154198,
+      "epoch": 10.413669064748202,
+      "step": 11580
+    },
+    {
+      "epoch": 10.413669064748202,
+      "step": 11580,
+      "torque_loss": 0.13548994064331055
+    },
+    {
+      "epoch": 10.413669064748202,
+      "force_loss": 0.015162634663283825,
+      "step": 11580
+    },
+    {
+      "epoch": 10.422661870503598,
+      "grad_norm": 0.6761921644210815,
+      "learning_rate": 9.450143045367156e-05,
+      "loss": 0.0272,
+      "step": 11590
+    },
+    {
+      "action_loss": 0.011015105061233044,
+      "epoch": 10.422661870503598,
+      "step": 11590
+    },
+    {
+      "epoch": 10.422661870503598,
+      "step": 11590,
+      "torque_loss": 0.13699547946453094
+    },
+    {
+      "epoch": 10.422661870503598,
+      "force_loss": 0.010299679823219776,
+      "step": 11590
+    },
+    {
+      "epoch": 10.431654676258994,
+      "grad_norm": 0.6008148789405823,
+      "learning_rate": 9.448885995139283e-05,
+      "loss": 0.0234,
+      "step": 11600
+    },
+    {
+      "action_loss": 0.011528166942298412,
+      "epoch": 10.431654676258994,
+      "step": 11600
+    },
+    {
+      "epoch": 10.431654676258994,
+      "step": 11600,
+      "torque_loss": 0.11631127446889877
+    },
+    {
+      "epoch": 10.431654676258994,
+      "force_loss": 0.009262864477932453,
+      "step": 11600
+    },
+    {
+      "epoch": 10.440647482014388,
+      "grad_norm": 0.6367351412773132,
+      "learning_rate": 9.44762759345731e-05,
+      "loss": 0.0239,
+      "step": 11610
+    },
+    {
+      "action_loss": 0.007582036778330803,
+      "epoch": 10.440647482014388,
+      "step": 11610
+    },
+    {
+      "epoch": 10.440647482014388,
+      "step": 11610,
+      "torque_loss": 0.19286580383777618
+    },
+    {
+      "epoch": 10.440647482014388,
+      "force_loss": 0.0071846297942101955,
+      "step": 11610
+    },
+    {
+      "epoch": 10.449640287769784,
+      "grad_norm": 0.6339514255523682,
+      "learning_rate": 9.446367840703509e-05,
+      "loss": 0.026,
+      "step": 11620
+    },
+    {
+      "action_loss": 0.012275408953428268,
+      "epoch": 10.449640287769784,
+      "step": 11620
+    },
+    {
+      "epoch": 10.449640287769784,
+      "step": 11620,
+      "torque_loss": 0.13773991167545319
+    },
+    {
+      "epoch": 10.449640287769784,
+      "force_loss": 0.018510090187191963,
+      "step": 11620
+    },
+    {
+      "epoch": 10.45863309352518,
+      "grad_norm": 0.47833752632141113,
+      "learning_rate": 9.445106737260556e-05,
+      "loss": 0.022,
+      "step": 11630
+    },
+    {
+      "action_loss": 0.005452519282698631,
+      "epoch": 10.45863309352518,
+      "step": 11630
+    },
+    {
+      "epoch": 10.45863309352518,
+      "step": 11630,
+      "torque_loss": 0.14784525334835052
+    },
+    {
+      "epoch": 10.45863309352518,
+      "force_loss": 0.0059990473091602325,
+      "step": 11630
+    },
+    {
+      "epoch": 10.467625899280575,
+      "grad_norm": 0.6986682415008545,
+      "learning_rate": 9.443844283511543e-05,
+      "loss": 0.0283,
+      "step": 11640
+    },
+    {
+      "action_loss": 0.011630523018538952,
+      "epoch": 10.467625899280575,
+      "step": 11640
+    },
+    {
+      "epoch": 10.467625899280575,
+      "step": 11640,
+      "torque_loss": 0.1290273517370224
+    },
+    {
+      "epoch": 10.467625899280575,
+      "force_loss": 0.01168547198176384,
+      "step": 11640
+    },
+    {
+      "epoch": 10.476618705035971,
+      "grad_norm": 0.45972588658332825,
+      "learning_rate": 9.442580479839968e-05,
+      "loss": 0.0338,
+      "step": 11650
+    },
+    {
+      "action_loss": 0.015398656018078327,
+      "epoch": 10.476618705035971,
+      "step": 11650
+    },
+    {
+      "epoch": 10.476618705035971,
+      "step": 11650,
+      "torque_loss": 0.14129537343978882
+    },
+    {
+      "epoch": 10.476618705035971,
+      "force_loss": 0.017711279913783073,
+      "step": 11650
+    },
+    {
+      "epoch": 10.485611510791367,
+      "grad_norm": 0.7560446262359619,
+      "learning_rate": 9.441315326629745e-05,
+      "loss": 0.0328,
+      "step": 11660
+    },
+    {
+      "action_loss": 0.014941698871552944,
+      "epoch": 10.485611510791367,
+      "step": 11660
+    },
+    {
+      "epoch": 10.485611510791367,
+      "step": 11660,
+      "torque_loss": 0.12922406196594238
+    },
+    {
+      "epoch": 10.485611510791367,
+      "force_loss": 0.009879964403808117,
+      "step": 11660
+    },
+    {
+      "epoch": 10.494604316546763,
+      "grad_norm": 0.6328633427619934,
+      "learning_rate": 9.44004882426519e-05,
+      "loss": 0.0308,
+      "step": 11670
+    },
+    {
+      "action_loss": 0.009849830530583858,
+      "epoch": 10.494604316546763,
+      "step": 11670
+    },
+    {
+      "epoch": 10.494604316546763,
+      "step": 11670,
+      "torque_loss": 0.1196376383304596
+    },
+    {
+      "epoch": 10.494604316546763,
+      "force_loss": 0.00854521244764328,
+      "step": 11670
+    },
+    {
+      "epoch": 10.503597122302159,
+      "grad_norm": 0.8618049025535583,
+      "learning_rate": 9.438780973131037e-05,
+      "loss": 0.0296,
+      "step": 11680
+    },
+    {
+      "action_loss": 0.007375011686235666,
+      "epoch": 10.503597122302159,
+      "step": 11680
+    },
+    {
+      "epoch": 10.503597122302159,
+      "step": 11680,
+      "torque_loss": 0.1726502627134323
+    },
+    {
+      "epoch": 10.503597122302159,
+      "force_loss": 0.0059968214482069016,
+      "step": 11680
+    },
+    {
+      "epoch": 10.512589928057555,
+      "grad_norm": 0.8481230139732361,
+      "learning_rate": 9.437511773612423e-05,
+      "loss": 0.0311,
+      "step": 11690
+    },
+    {
+      "action_loss": 0.01055148709565401,
+      "epoch": 10.512589928057555,
+      "step": 11690
+    },
+    {
+      "epoch": 10.512589928057555,
+      "step": 11690,
+      "torque_loss": 0.10800374299287796
+    },
+    {
+      "epoch": 10.512589928057555,
+      "force_loss": 0.008011865429580212,
+      "step": 11690
+    },
+    {
+      "epoch": 10.52158273381295,
+      "grad_norm": 0.3983135521411896,
+      "learning_rate": 9.436241226094896e-05,
+      "loss": 0.0351,
+      "step": 11700
+    },
+    {
+      "action_loss": 0.044229984283447266,
+      "epoch": 10.52158273381295,
+      "step": 11700
+    },
+    {
+      "epoch": 10.52158273381295,
+      "step": 11700,
+      "torque_loss": 0.15636643767356873
+    },
+    {
+      "epoch": 10.52158273381295,
+      "force_loss": 0.0358545146882534,
+      "step": 11700
+    },
+    {
+      "epoch": 10.530575539568346,
+      "grad_norm": 0.5260031819343567,
+      "learning_rate": 9.434969330964418e-05,
+      "loss": 0.0282,
+      "step": 11710
+    },
+    {
+      "action_loss": 0.00560585455968976,
+      "epoch": 10.530575539568346,
+      "step": 11710
+    },
+    {
+      "epoch": 10.530575539568346,
+      "step": 11710,
+      "torque_loss": 0.15837110579013824
+    },
+    {
+      "epoch": 10.530575539568346,
+      "force_loss": 0.006068322341889143,
+      "step": 11710
+    },
+    {
+      "epoch": 10.53956834532374,
+      "grad_norm": 0.7153356671333313,
+      "learning_rate": 9.433696088607356e-05,
+      "loss": 0.0267,
+      "step": 11720
+    },
+    {
+      "action_loss": 0.011927996762096882,
+      "epoch": 10.53956834532374,
+      "step": 11720
+    },
+    {
+      "epoch": 10.53956834532374,
+      "step": 11720,
+      "torque_loss": 0.12231427431106567
+    },
+    {
+      "epoch": 10.53956834532374,
+      "force_loss": 0.02651011385023594,
+      "step": 11720
+    },
+    {
+      "epoch": 10.548561151079136,
+      "grad_norm": 0.5481546521186829,
+      "learning_rate": 9.432421499410486e-05,
+      "loss": 0.0256,
+      "step": 11730
+    },
+    {
+      "action_loss": 0.012892067432403564,
+      "epoch": 10.548561151079136,
+      "step": 11730
+    },
+    {
+      "epoch": 10.548561151079136,
+      "step": 11730,
+      "torque_loss": 0.15337607264518738
+    },
+    {
+      "epoch": 10.548561151079136,
+      "force_loss": 0.010688972659409046,
+      "step": 11730
+    },
+    {
+      "epoch": 10.557553956834532,
+      "grad_norm": 0.506474494934082,
+      "learning_rate": 9.431145563760998e-05,
+      "loss": 0.0275,
+      "step": 11740
+    },
+    {
+      "action_loss": 0.019991571083664894,
+      "epoch": 10.557553956834532,
+      "step": 11740
+    },
+    {
+      "epoch": 10.557553956834532,
+      "step": 11740,
+      "torque_loss": 0.1517646312713623
+    },
+    {
+      "epoch": 10.557553956834532,
+      "force_loss": 0.02033424936234951,
+      "step": 11740
+    },
+    {
+      "epoch": 10.566546762589928,
+      "grad_norm": 1.1932791471481323,
+      "learning_rate": 9.429868282046484e-05,
+      "loss": 0.034,
+      "step": 11750
+    },
+    {
+      "action_loss": 0.011276725679636002,
+      "epoch": 10.566546762589928,
+      "step": 11750
+    },
+    {
+      "epoch": 10.566546762589928,
+      "step": 11750,
+      "torque_loss": 0.14569266140460968
+    },
+    {
+      "epoch": 10.566546762589928,
+      "force_loss": 0.014366704039275646,
+      "step": 11750
+    },
+    {
+      "epoch": 10.575539568345324,
+      "grad_norm": 0.9943850636482239,
+      "learning_rate": 9.428589654654951e-05,
+      "loss": 0.0367,
+      "step": 11760
+    },
+    {
+      "action_loss": 0.011266603134572506,
+      "epoch": 10.575539568345324,
+      "step": 11760
+    },
+    {
+      "epoch": 10.575539568345324,
+      "step": 11760,
+      "torque_loss": 0.14086222648620605
+    },
+    {
+      "epoch": 10.575539568345324,
+      "force_loss": 0.009158913046121597,
+      "step": 11760
+    },
+    {
+      "epoch": 10.58453237410072,
+      "grad_norm": 0.6900865435600281,
+      "learning_rate": 9.42730968197481e-05,
+      "loss": 0.0272,
+      "step": 11770
+    },
+    {
+      "action_loss": 0.008175398223102093,
+      "epoch": 10.58453237410072,
+      "step": 11770
+    },
+    {
+      "epoch": 10.58453237410072,
+      "step": 11770,
+      "torque_loss": 0.16871309280395508
+    },
+    {
+      "epoch": 10.58453237410072,
+      "force_loss": 0.0069722323678433895,
+      "step": 11770
+    },
+    {
+      "epoch": 10.593525179856115,
+      "grad_norm": 0.4927932620048523,
+      "learning_rate": 9.426028364394883e-05,
+      "loss": 0.0271,
+      "step": 11780
+    },
+    {
+      "action_loss": 0.011634415946900845,
+      "epoch": 10.593525179856115,
+      "step": 11780
+    },
+    {
+      "epoch": 10.593525179856115,
+      "step": 11780,
+      "torque_loss": 0.10223275423049927
+    },
+    {
+      "epoch": 10.593525179856115,
+      "force_loss": 0.00927719846367836,
+      "step": 11780
+    },
+    {
+      "epoch": 10.602517985611511,
+      "grad_norm": 1.3977556228637695,
+      "learning_rate": 9.424745702304402e-05,
+      "loss": 0.0255,
+      "step": 11790
+    },
+    {
+      "action_loss": 0.009241175837814808,
+      "epoch": 10.602517985611511,
+      "step": 11790
+    },
+    {
+      "epoch": 10.602517985611511,
+      "step": 11790,
+      "torque_loss": 0.15933172404766083
+    },
+    {
+      "epoch": 10.602517985611511,
+      "force_loss": 0.010606445372104645,
+      "step": 11790
+    },
+    {
+      "epoch": 10.611510791366907,
+      "grad_norm": 0.8013253808021545,
+      "learning_rate": 9.423461696093006e-05,
+      "loss": 0.0316,
+      "step": 11800
+    },
+    {
+      "action_loss": 0.029430806636810303,
+      "epoch": 10.611510791366907,
+      "step": 11800
+    },
+    {
+      "epoch": 10.611510791366907,
+      "step": 11800,
+      "torque_loss": 0.24555903673171997
+    },
+    {
+      "epoch": 10.611510791366907,
+      "force_loss": 0.047616854310035706,
+      "step": 11800
+    },
+    {
+      "epoch": 10.620503597122303,
+      "grad_norm": 0.31562918424606323,
+      "learning_rate": 9.422176346150741e-05,
+      "loss": 0.0279,
+      "step": 11810
+    },
+    {
+      "action_loss": 0.00797791127115488,
+      "epoch": 10.620503597122303,
+      "step": 11810
+    },
+    {
+      "epoch": 10.620503597122303,
+      "step": 11810,
+      "torque_loss": 0.10208842158317566
+    },
+    {
+      "epoch": 10.620503597122303,
+      "force_loss": 0.010428149253129959,
+      "step": 11810
+    },
+    {
+      "epoch": 10.629496402877697,
+      "grad_norm": 0.9264644980430603,
+      "learning_rate": 9.420889652868063e-05,
+      "loss": 0.027,
+      "step": 11820
+    },
+    {
+      "action_loss": 0.005660125520080328,
+      "epoch": 10.629496402877697,
+      "step": 11820
+    },
+    {
+      "epoch": 10.629496402877697,
+      "step": 11820,
+      "torque_loss": 0.11972040683031082
+    },
+    {
+      "epoch": 10.629496402877697,
+      "force_loss": 0.011650193482637405,
+      "step": 11820
+    },
+    {
+      "epoch": 10.638489208633093,
+      "grad_norm": 0.3744155168533325,
+      "learning_rate": 9.419601616635836e-05,
+      "loss": 0.0269,
+      "step": 11830
+    },
+    {
+      "action_loss": 0.011947602033615112,
+      "epoch": 10.638489208633093,
+      "step": 11830
+    },
+    {
+      "epoch": 10.638489208633093,
+      "step": 11830,
+      "torque_loss": 0.14063523709774017
+    },
+    {
+      "epoch": 10.638489208633093,
+      "force_loss": 0.009198009967803955,
+      "step": 11830
+    },
+    {
+      "epoch": 10.647482014388489,
+      "grad_norm": 0.372519314289093,
+      "learning_rate": 9.418312237845331e-05,
+      "loss": 0.0302,
+      "step": 11840
+    },
+    {
+      "action_loss": 0.008128146640956402,
+      "epoch": 10.647482014388489,
+      "step": 11840
+    },
+    {
+      "epoch": 10.647482014388489,
+      "step": 11840,
+      "torque_loss": 0.1194186806678772
+    },
+    {
+      "epoch": 10.647482014388489,
+      "force_loss": 0.005439929664134979,
+      "step": 11840
+    },
+    {
+      "epoch": 10.656474820143885,
+      "grad_norm": 2.5389938354492188,
+      "learning_rate": 9.417021516888225e-05,
+      "loss": 0.0287,
+      "step": 11850
+    },
+    {
+      "action_loss": 0.010768379084765911,
+      "epoch": 10.656474820143885,
+      "step": 11850
+    },
+    {
+      "epoch": 10.656474820143885,
+      "step": 11850,
+      "torque_loss": 0.12309720367193222
+    },
+    {
+      "epoch": 10.656474820143885,
+      "force_loss": 0.01216029655188322,
+      "step": 11850
+    },
+    {
+      "epoch": 10.66546762589928,
+      "grad_norm": 0.6142351031303406,
+      "learning_rate": 9.415729454156608e-05,
+      "loss": 0.026,
+      "step": 11860
+    },
+    {
+      "action_loss": 0.024880558252334595,
+      "epoch": 10.66546762589928,
+      "step": 11860
+    },
+    {
+      "epoch": 10.66546762589928,
+      "step": 11860,
+      "torque_loss": 0.12127278000116348
+    },
+    {
+      "epoch": 10.66546762589928,
+      "force_loss": 0.02288181148469448,
+      "step": 11860
+    },
+    {
+      "epoch": 10.674460431654676,
+      "grad_norm": 0.22927068173885345,
+      "learning_rate": 9.414436050042973e-05,
+      "loss": 0.0285,
+      "step": 11870
+    },
+    {
+      "action_loss": 0.01672782376408577,
+      "epoch": 10.674460431654676,
+      "step": 11870
+    },
+    {
+      "epoch": 10.674460431654676,
+      "step": 11870,
+      "torque_loss": 0.10165462642908096
+    },
+    {
+      "epoch": 10.674460431654676,
+      "force_loss": 0.015788892284035683,
+      "step": 11870
+    },
+    {
+      "epoch": 10.683453237410072,
+      "grad_norm": 0.5687105655670166,
+      "learning_rate": 9.413141304940223e-05,
+      "loss": 0.0273,
+      "step": 11880
+    },
+    {
+      "action_loss": 0.011985464952886105,
+      "epoch": 10.683453237410072,
+      "step": 11880
+    },
+    {
+      "epoch": 10.683453237410072,
+      "step": 11880,
+      "torque_loss": 0.1488141566514969
+    },
+    {
+      "epoch": 10.683453237410072,
+      "force_loss": 0.016240663826465607,
+      "step": 11880
+    },
+    {
+      "epoch": 10.692446043165468,
+      "grad_norm": 0.7064986824989319,
+      "learning_rate": 9.411845219241666e-05,
+      "loss": 0.0263,
+      "step": 11890
+    },
+    {
+      "action_loss": 0.0039049230981618166,
+      "epoch": 10.692446043165468,
+      "step": 11890
+    },
+    {
+      "epoch": 10.692446043165468,
+      "step": 11890,
+      "torque_loss": 0.11934468895196915
+    },
+    {
+      "epoch": 10.692446043165468,
+      "force_loss": 0.004777942318469286,
+      "step": 11890
+    },
+    {
+      "epoch": 10.701438848920864,
+      "grad_norm": 0.3298688232898712,
+      "learning_rate": 9.410547793341021e-05,
+      "loss": 0.0271,
+      "step": 11900
+    },
+    {
+      "action_loss": 0.009596536867320538,
+      "epoch": 10.701438848920864,
+      "step": 11900
+    },
+    {
+      "epoch": 10.701438848920864,
+      "step": 11900,
+      "torque_loss": 0.12201857566833496
+    },
+    {
+      "epoch": 10.701438848920864,
+      "force_loss": 0.007582666352391243,
+      "step": 11900
+    },
+    {
+      "epoch": 10.71043165467626,
+      "grad_norm": 0.5734339952468872,
+      "learning_rate": 9.409249027632408e-05,
+      "loss": 0.0252,
+      "step": 11910
+    },
+    {
+      "action_loss": 0.004827613011002541,
+      "epoch": 10.71043165467626,
+      "step": 11910
+    },
+    {
+      "epoch": 10.71043165467626,
+      "step": 11910,
+      "torque_loss": 0.14751964807510376
+    },
+    {
+      "epoch": 10.71043165467626,
+      "force_loss": 0.005613662302494049,
+      "step": 11910
+    },
+    {
+      "epoch": 10.719424460431654,
+      "grad_norm": 0.3818264603614807,
+      "learning_rate": 9.407948922510362e-05,
+      "loss": 0.0255,
+      "step": 11920
+    },
+    {
+      "action_loss": 0.009109928272664547,
+      "epoch": 10.719424460431654,
+      "step": 11920
+    },
+    {
+      "epoch": 10.719424460431654,
+      "step": 11920,
+      "torque_loss": 0.08783774822950363
+    },
+    {
+      "epoch": 10.719424460431654,
+      "force_loss": 0.008770224638283253,
+      "step": 11920
+    },
+    {
+      "epoch": 10.72841726618705,
+      "grad_norm": 0.5351333618164062,
+      "learning_rate": 9.406647478369817e-05,
+      "loss": 0.0235,
+      "step": 11930
+    },
+    {
+      "action_loss": 0.017036335542798042,
+      "epoch": 10.72841726618705,
+      "step": 11930
+    },
+    {
+      "epoch": 10.72841726618705,
+      "step": 11930,
+      "torque_loss": 0.16139338910579681
+    },
+    {
+      "epoch": 10.72841726618705,
+      "force_loss": 0.02104955166578293,
+      "step": 11930
+    },
+    {
+      "epoch": 10.737410071942445,
+      "grad_norm": 0.46916478872299194,
+      "learning_rate": 9.405344695606118e-05,
+      "loss": 0.0263,
+      "step": 11940
+    },
+    {
+      "action_loss": 0.004225591663271189,
+      "epoch": 10.737410071942445,
+      "step": 11940
+    },
+    {
+      "epoch": 10.737410071942445,
+      "step": 11940,
+      "torque_loss": 0.12437965720891953
+    },
+    {
+      "epoch": 10.737410071942445,
+      "force_loss": 0.0045722294598817825,
+      "step": 11940
+    },
+    {
+      "epoch": 10.746402877697841,
+      "grad_norm": 0.7353121638298035,
+      "learning_rate": 9.404040574615018e-05,
+      "loss": 0.0257,
+      "step": 11950
+    },
+    {
+      "action_loss": 0.02080698497593403,
+      "epoch": 10.746402877697841,
+      "step": 11950
+    },
+    {
+      "epoch": 10.746402877697841,
+      "step": 11950,
+      "torque_loss": 0.16435514390468597
+    },
+    {
+      "epoch": 10.746402877697841,
+      "force_loss": 0.017621351405978203,
+      "step": 11950
+    },
+    {
+      "epoch": 10.755395683453237,
+      "grad_norm": 0.5872335433959961,
+      "learning_rate": 9.402735115792674e-05,
+      "loss": 0.0281,
+      "step": 11960
+    },
+    {
+      "action_loss": 0.004947586450725794,
+      "epoch": 10.755395683453237,
+      "step": 11960
+    },
+    {
+      "epoch": 10.755395683453237,
+      "step": 11960,
+      "torque_loss": 0.14728674292564392
+    },
+    {
+      "epoch": 10.755395683453237,
+      "force_loss": 0.004926704335957766,
+      "step": 11960
+    },
+    {
+      "epoch": 10.764388489208633,
+      "grad_norm": 0.4793684184551239,
+      "learning_rate": 9.401428319535649e-05,
+      "loss": 0.0232,
+      "step": 11970
+    },
+    {
+      "action_loss": 0.008999993093311787,
+      "epoch": 10.764388489208633,
+      "step": 11970
+    },
+    {
+      "epoch": 10.764388489208633,
+      "step": 11970,
+      "torque_loss": 0.1286764293909073
+    },
+    {
+      "epoch": 10.764388489208633,
+      "force_loss": 0.008132883347570896,
+      "step": 11970
+    },
+    {
+      "epoch": 10.773381294964029,
+      "grad_norm": 0.4197932779788971,
+      "learning_rate": 9.400120186240912e-05,
+      "loss": 0.027,
+      "step": 11980
+    },
+    {
+      "action_loss": 0.011099553667008877,
+      "epoch": 10.773381294964029,
+      "step": 11980
+    },
+    {
+      "epoch": 10.773381294964029,
+      "step": 11980,
+      "torque_loss": 0.11338604241609573
+    },
+    {
+      "epoch": 10.773381294964029,
+      "force_loss": 0.010286360047757626,
+      "step": 11980
+    },
+    {
+      "epoch": 10.782374100719425,
+      "grad_norm": 0.6464086771011353,
+      "learning_rate": 9.398810716305844e-05,
+      "loss": 0.026,
+      "step": 11990
+    },
+    {
+      "action_loss": 0.019942304119467735,
+      "epoch": 10.782374100719425,
+      "step": 11990
+    },
+    {
+      "epoch": 10.782374100719425,
+      "step": 11990,
+      "torque_loss": 0.1684349775314331
+    },
+    {
+      "epoch": 10.782374100719425,
+      "force_loss": 0.013247677125036716,
+      "step": 11990
+    },
+    {
+      "epoch": 10.79136690647482,
+      "grad_norm": 0.6473360657691956,
+      "learning_rate": 9.397499910128222e-05,
+      "loss": 0.0272,
+      "step": 12000
+    },
+    {
+      "action_loss": 0.006471951026469469,
+      "epoch": 10.79136690647482,
+      "step": 12000
+    },
+    {
+      "epoch": 10.79136690647482,
+      "step": 12000,
+      "torque_loss": 0.1268175095319748
+    },
+    {
+      "epoch": 10.79136690647482,
+      "force_loss": 0.005824244115501642,
+      "step": 12000
+    },
+    {
+      "epoch": 10.800359712230216,
+      "grad_norm": 0.26278069615364075,
+      "learning_rate": 9.396187768106237e-05,
+      "loss": 0.0232,
+      "step": 12010
+    },
+    {
+      "action_loss": 0.010234486311674118,
+      "epoch": 10.800359712230216,
+      "step": 12010
+    },
+    {
+      "epoch": 10.800359712230216,
+      "step": 12010,
+      "torque_loss": 0.13185055553913116
+    },
+    {
+      "epoch": 10.800359712230216,
+      "force_loss": 0.006136443465948105,
+      "step": 12010
+    },
+    {
+      "epoch": 10.809352517985612,
+      "grad_norm": 0.7460739612579346,
+      "learning_rate": 9.394874290638482e-05,
+      "loss": 0.0268,
+      "step": 12020
+    },
+    {
+      "action_loss": 0.016185598447918892,
+      "epoch": 10.809352517985612,
+      "step": 12020
+    },
+    {
+      "epoch": 10.809352517985612,
+      "step": 12020,
+      "torque_loss": 0.1253592073917389
+    },
+    {
+      "epoch": 10.809352517985612,
+      "force_loss": 0.012232418172061443,
+      "step": 12020
+    },
+    {
+      "epoch": 10.818345323741006,
+      "grad_norm": 0.5842387080192566,
+      "learning_rate": 9.393559478123959e-05,
+      "loss": 0.0291,
+      "step": 12030
+    },
+    {
+      "action_loss": 0.015043514780700207,
+      "epoch": 10.818345323741006,
+      "step": 12030
+    },
+    {
+      "epoch": 10.818345323741006,
+      "step": 12030,
+      "torque_loss": 0.16573457419872284
+    },
+    {
+      "epoch": 10.818345323741006,
+      "force_loss": 0.020701823756098747,
+      "step": 12030
+    },
+    {
+      "epoch": 10.827338129496402,
+      "grad_norm": 0.764706552028656,
+      "learning_rate": 9.39224333096207e-05,
+      "loss": 0.0262,
+      "step": 12040
+    },
+    {
+      "action_loss": 0.006273603066802025,
+      "epoch": 10.827338129496402,
+      "step": 12040
+    },
+    {
+      "epoch": 10.827338129496402,
+      "step": 12040,
+      "torque_loss": 0.09680929780006409
+    },
+    {
+      "epoch": 10.827338129496402,
+      "force_loss": 0.007812932133674622,
+      "step": 12040
+    },
+    {
+      "epoch": 10.836330935251798,
+      "grad_norm": 0.5986528992652893,
+      "learning_rate": 9.390925849552629e-05,
+      "loss": 0.0275,
+      "step": 12050
+    },
+    {
+      "action_loss": 0.010162926279008389,
+      "epoch": 10.836330935251798,
+      "step": 12050
+    },
+    {
+      "epoch": 10.836330935251798,
+      "step": 12050,
+      "torque_loss": 0.1518034189939499
+    },
+    {
+      "epoch": 10.836330935251798,
+      "force_loss": 0.0305942390114069,
+      "step": 12050
+    },
+    {
+      "epoch": 10.845323741007194,
+      "grad_norm": 0.5597752928733826,
+      "learning_rate": 9.389607034295849e-05,
+      "loss": 0.0242,
+      "step": 12060
+    },
+    {
+      "action_loss": 0.014853022992610931,
+      "epoch": 10.845323741007194,
+      "step": 12060
+    },
+    {
+      "epoch": 10.845323741007194,
+      "step": 12060,
+      "torque_loss": 0.14084966480731964
+    },
+    {
+      "epoch": 10.845323741007194,
+      "force_loss": 0.013705846853554249,
+      "step": 12060
+    },
+    {
+      "epoch": 10.85431654676259,
+      "grad_norm": 0.44602155685424805,
+      "learning_rate": 9.388286885592355e-05,
+      "loss": 0.0326,
+      "step": 12070
+    },
+    {
+      "action_loss": 0.007571047637611628,
+      "epoch": 10.85431654676259,
+      "step": 12070
+    },
+    {
+      "epoch": 10.85431654676259,
+      "step": 12070,
+      "torque_loss": 0.1311572641134262
+    },
+    {
+      "epoch": 10.85431654676259,
+      "force_loss": 0.015139713883399963,
+      "step": 12070
+    },
+    {
+      "epoch": 10.863309352517986,
+      "grad_norm": 0.3522129952907562,
+      "learning_rate": 9.386965403843168e-05,
+      "loss": 0.0255,
+      "step": 12080
+    },
+    {
+      "action_loss": 0.009334375150501728,
+      "epoch": 10.863309352517986,
+      "step": 12080
+    },
+    {
+      "epoch": 10.863309352517986,
+      "step": 12080,
+      "torque_loss": 0.1605527400970459
+    },
+    {
+      "epoch": 10.863309352517986,
+      "force_loss": 0.009085362777113914,
+      "step": 12080
+    },
+    {
+      "epoch": 10.872302158273381,
+      "grad_norm": 0.5320757031440735,
+      "learning_rate": 9.385642589449726e-05,
+      "loss": 0.0258,
+      "step": 12090
+    },
+    {
+      "action_loss": 0.0033140864688903093,
+      "epoch": 10.872302158273381,
+      "step": 12090
+    },
+    {
+      "epoch": 10.872302158273381,
+      "step": 12090,
+      "torque_loss": 0.10213208198547363
+    },
+    {
+      "epoch": 10.872302158273381,
+      "force_loss": 0.0058080703020095825,
+      "step": 12090
+    },
+    {
+      "epoch": 10.881294964028777,
+      "grad_norm": 0.45063185691833496,
+      "learning_rate": 9.38431844281386e-05,
+      "loss": 0.0227,
+      "step": 12100
+    },
+    {
+      "action_loss": 0.00553587032482028,
+      "epoch": 10.881294964028777,
+      "step": 12100
+    },
+    {
+      "epoch": 10.881294964028777,
+      "step": 12100,
+      "torque_loss": 0.1320485919713974
+    },
+    {
+      "epoch": 10.881294964028777,
+      "force_loss": 0.00692811468616128,
+      "step": 12100
+    },
+    {
+      "epoch": 10.890287769784173,
+      "grad_norm": 0.2876790761947632,
+      "learning_rate": 9.38299296433781e-05,
+      "loss": 0.0254,
+      "step": 12110
+    },
+    {
+      "action_loss": 0.01288487296551466,
+      "epoch": 10.890287769784173,
+      "step": 12110
+    },
+    {
+      "epoch": 10.890287769784173,
+      "step": 12110,
+      "torque_loss": 0.12228699773550034
+    },
+    {
+      "epoch": 10.890287769784173,
+      "force_loss": 0.009581823833286762,
+      "step": 12110
+    },
+    {
+      "epoch": 10.899280575539569,
+      "grad_norm": 0.6605455875396729,
+      "learning_rate": 9.381666154424226e-05,
+      "loss": 0.0266,
+      "step": 12120
+    },
+    {
+      "action_loss": 0.008676379919052124,
+      "epoch": 10.899280575539569,
+      "step": 12120
+    },
+    {
+      "epoch": 10.899280575539569,
+      "step": 12120,
+      "torque_loss": 0.11593180149793625
+    },
+    {
+      "epoch": 10.899280575539569,
+      "force_loss": 0.006400665733963251,
+      "step": 12120
+    },
+    {
+      "epoch": 10.908273381294965,
+      "grad_norm": 0.9221895337104797,
+      "learning_rate": 9.380338013476157e-05,
+      "loss": 0.024,
+      "step": 12130
+    },
+    {
+      "action_loss": 0.014356992207467556,
+      "epoch": 10.908273381294965,
+      "step": 12130
+    },
+    {
+      "epoch": 10.908273381294965,
+      "step": 12130,
+      "torque_loss": 0.14662493765354156
+    },
+    {
+      "epoch": 10.908273381294965,
+      "force_loss": 0.018285570666193962,
+      "step": 12130
+    },
+    {
+      "epoch": 10.917266187050359,
+      "grad_norm": 0.9484205842018127,
+      "learning_rate": 9.379008541897054e-05,
+      "loss": 0.0262,
+      "step": 12140
+    },
+    {
+      "action_loss": 0.0054193236865103245,
+      "epoch": 10.917266187050359,
+      "step": 12140
+    },
+    {
+      "epoch": 10.917266187050359,
+      "step": 12140,
+      "torque_loss": 0.12005052715539932
+    },
+    {
+      "epoch": 10.917266187050359,
+      "force_loss": 0.004963604733347893,
+      "step": 12140
+    },
+    {
+      "epoch": 10.926258992805755,
+      "grad_norm": 0.24158206582069397,
+      "learning_rate": 9.377677740090777e-05,
+      "loss": 0.0319,
+      "step": 12150
+    },
+    {
+      "action_loss": 0.009313946589827538,
+      "epoch": 10.926258992805755,
+      "step": 12150
+    },
+    {
+      "epoch": 10.926258992805755,
+      "step": 12150,
+      "torque_loss": 0.1337345987558365
+    },
+    {
+      "epoch": 10.926258992805755,
+      "force_loss": 0.012692399322986603,
+      "step": 12150
+    },
+    {
+      "epoch": 10.93525179856115,
+      "grad_norm": 0.5572620034217834,
+      "learning_rate": 9.376345608461588e-05,
+      "loss": 0.0266,
+      "step": 12160
+    },
+    {
+      "action_loss": 0.010906982235610485,
+      "epoch": 10.93525179856115,
+      "step": 12160
+    },
+    {
+      "epoch": 10.93525179856115,
+      "step": 12160,
+      "torque_loss": 0.1310039907693863
+    },
+    {
+      "epoch": 10.93525179856115,
+      "force_loss": 0.008720236830413342,
+      "step": 12160
+    },
+    {
+      "epoch": 10.944244604316546,
+      "grad_norm": 0.4573909044265747,
+      "learning_rate": 9.375012147414155e-05,
+      "loss": 0.0228,
+      "step": 12170
+    },
+    {
+      "action_loss": 0.011558172293007374,
+      "epoch": 10.944244604316546,
+      "step": 12170
+    },
+    {
+      "epoch": 10.944244604316546,
+      "step": 12170,
+      "torque_loss": 0.17224471271038055
+    },
+    {
+      "epoch": 10.944244604316546,
+      "force_loss": 0.01524873822927475,
+      "step": 12170
+    },
+    {
+      "epoch": 10.953237410071942,
+      "grad_norm": 0.40970471501350403,
+      "learning_rate": 9.373677357353545e-05,
+      "loss": 0.0263,
+      "step": 12180
+    },
+    {
+      "action_loss": 0.019229477271437645,
+      "epoch": 10.953237410071942,
+      "step": 12180
+    },
+    {
+      "epoch": 10.953237410071942,
+      "step": 12180,
+      "torque_loss": 0.1303432136774063
+    },
+    {
+      "epoch": 10.953237410071942,
+      "force_loss": 0.02540399134159088,
+      "step": 12180
+    },
+    {
+      "epoch": 10.962230215827338,
+      "grad_norm": 0.6801599860191345,
+      "learning_rate": 9.372341238685237e-05,
+      "loss": 0.0251,
+      "step": 12190
+    },
+    {
+      "action_loss": 0.005424709524959326,
+      "epoch": 10.962230215827338,
+      "step": 12190
+    },
+    {
+      "epoch": 10.962230215827338,
+      "step": 12190,
+      "torque_loss": 0.17813409864902496
+    },
+    {
+      "epoch": 10.962230215827338,
+      "force_loss": 0.004139252007007599,
+      "step": 12190
+    },
+    {
+      "epoch": 10.971223021582734,
+      "grad_norm": 0.6432246565818787,
+      "learning_rate": 9.371003791815102e-05,
+      "loss": 0.0228,
+      "step": 12200
+    },
+    {
+      "action_loss": 0.004331761505454779,
+      "epoch": 10.971223021582734,
+      "step": 12200
+    },
+    {
+      "epoch": 10.971223021582734,
+      "step": 12200,
+      "torque_loss": 0.10015038400888443
+    },
+    {
+      "epoch": 10.971223021582734,
+      "force_loss": 0.004639073740690947,
+      "step": 12200
+    },
+    {
+      "epoch": 10.98021582733813,
+      "grad_norm": 0.6928061246871948,
+      "learning_rate": 9.369665017149429e-05,
+      "loss": 0.0251,
+      "step": 12210
+    },
+    {
+      "action_loss": 0.01653742603957653,
+      "epoch": 10.98021582733813,
+      "step": 12210
+    },
+    {
+      "epoch": 10.98021582733813,
+      "step": 12210,
+      "torque_loss": 0.1767071932554245
+    },
+    {
+      "epoch": 10.98021582733813,
+      "force_loss": 0.008227099664509296,
+      "step": 12210
+    },
+    {
+      "epoch": 10.989208633093526,
+      "grad_norm": 0.3463161587715149,
+      "learning_rate": 9.368324915094895e-05,
+      "loss": 0.0228,
+      "step": 12220
+    },
+    {
+      "action_loss": 0.025904029607772827,
+      "epoch": 10.989208633093526,
+      "step": 12220
+    },
+    {
+      "epoch": 10.989208633093526,
+      "step": 12220,
+      "torque_loss": 0.1508151888847351
+    },
+    {
+      "epoch": 10.989208633093526,
+      "force_loss": 0.03892321512103081,
+      "step": 12220
+    },
+    {
+      "epoch": 10.998201438848922,
+      "grad_norm": 0.6688510775566101,
+      "learning_rate": 9.366983486058591e-05,
+      "loss": 0.0313,
+      "step": 12230
+    },
+    {
+      "action_loss": 0.004591356497257948,
+      "epoch": 10.998201438848922,
+      "step": 12230
+    },
+    {
+      "epoch": 10.998201438848922,
+      "step": 12230,
+      "torque_loss": 0.1349913328886032
+    },
+    {
+      "epoch": 10.998201438848922,
+      "force_loss": 0.004875443410128355,
+      "step": 12230
+    },
+    {
+      "epoch": 11.007194244604317,
+      "grad_norm": 0.6722777485847473,
+      "learning_rate": 9.365640730448009e-05,
+      "loss": 0.0343,
+      "step": 12240
+    },
+    {
+      "action_loss": 0.010542180389165878,
+      "epoch": 11.007194244604317,
+      "step": 12240
+    },
+    {
+      "epoch": 11.007194244604317,
+      "step": 12240,
+      "torque_loss": 0.1542787104845047
+    },
+    {
+      "epoch": 11.007194244604317,
+      "force_loss": 0.010751006193459034,
+      "step": 12240
+    },
+    {
+      "epoch": 11.016187050359711,
+      "grad_norm": 0.8943272829055786,
+      "learning_rate": 9.36429664867104e-05,
+      "loss": 0.0338,
+      "step": 12250
+    },
+    {
+      "action_loss": 0.007278565317392349,
+      "epoch": 11.016187050359711,
+      "step": 12250
+    },
+    {
+      "epoch": 11.016187050359711,
+      "step": 12250,
+      "torque_loss": 0.11434619873762131
+    },
+    {
+      "epoch": 11.016187050359711,
+      "force_loss": 0.007861269637942314,
+      "step": 12250
+    },
+    {
+      "epoch": 11.025179856115107,
+      "grad_norm": 0.31568488478660583,
+      "learning_rate": 9.362951241135982e-05,
+      "loss": 0.0297,
+      "step": 12260
+    },
+    {
+      "action_loss": 0.010769858956336975,
+      "epoch": 11.025179856115107,
+      "step": 12260
+    },
+    {
+      "epoch": 11.025179856115107,
+      "step": 12260,
+      "torque_loss": 0.13687936961650848
+    },
+    {
+      "epoch": 11.025179856115107,
+      "force_loss": 0.010664626955986023,
+      "step": 12260
+    },
+    {
+      "epoch": 11.034172661870503,
+      "grad_norm": 0.8092395663261414,
+      "learning_rate": 9.361604508251534e-05,
+      "loss": 0.026,
+      "step": 12270
+    },
+    {
+      "action_loss": 0.007826884277164936,
+      "epoch": 11.034172661870503,
+      "step": 12270
+    },
+    {
+      "epoch": 11.034172661870503,
+      "step": 12270,
+      "torque_loss": 0.1676703691482544
+    },
+    {
+      "epoch": 11.034172661870503,
+      "force_loss": 0.007892599329352379,
+      "step": 12270
+    },
+    {
+      "epoch": 11.043165467625899,
+      "grad_norm": 0.4204530417919159,
+      "learning_rate": 9.360256450426799e-05,
+      "loss": 0.0254,
+      "step": 12280
+    },
+    {
+      "action_loss": 0.01079513132572174,
+      "epoch": 11.043165467625899,
+      "step": 12280
+    },
+    {
+      "epoch": 11.043165467625899,
+      "step": 12280,
+      "torque_loss": 0.16246967017650604
+    },
+    {
+      "epoch": 11.043165467625899,
+      "force_loss": 0.00886585097759962,
+      "step": 12280
+    },
+    {
+      "epoch": 11.052158273381295,
+      "grad_norm": 0.5435256361961365,
+      "learning_rate": 9.358907068071279e-05,
+      "loss": 0.0254,
+      "step": 12290
+    },
+    {
+      "action_loss": 0.005063696298748255,
+      "epoch": 11.052158273381295,
+      "step": 12290
+    },
+    {
+      "epoch": 11.052158273381295,
+      "step": 12290,
+      "torque_loss": 0.13453452289104462
+    },
+    {
+      "epoch": 11.052158273381295,
+      "force_loss": 0.0044000535272061825,
+      "step": 12290
+    },
+    {
+      "epoch": 11.06115107913669,
+      "grad_norm": 0.328739196062088,
+      "learning_rate": 9.357556361594882e-05,
+      "loss": 0.0226,
+      "step": 12300
+    },
+    {
+      "action_loss": 0.004865929950028658,
+      "epoch": 11.06115107913669,
+      "step": 12300
+    },
+    {
+      "epoch": 11.06115107913669,
+      "step": 12300,
+      "torque_loss": 0.13972128927707672
+    },
+    {
+      "epoch": 11.06115107913669,
+      "force_loss": 0.0032887046691030264,
+      "step": 12300
+    },
+    {
+      "epoch": 11.070143884892087,
+      "grad_norm": 0.6979000568389893,
+      "learning_rate": 9.356204331407917e-05,
+      "loss": 0.0266,
+      "step": 12310
+    },
+    {
+      "action_loss": 0.037070319056510925,
+      "epoch": 11.070143884892087,
+      "step": 12310
+    },
+    {
+      "epoch": 11.070143884892087,
+      "step": 12310,
+      "torque_loss": 0.1636311560869217
+    },
+    {
+      "epoch": 11.070143884892087,
+      "force_loss": 0.04909073933959007,
+      "step": 12310
+    },
+    {
+      "epoch": 11.079136690647482,
+      "grad_norm": 0.599552571773529,
+      "learning_rate": 9.354850977921094e-05,
+      "loss": 0.0312,
+      "step": 12320
+    },
+    {
+      "action_loss": 0.008449557237327099,
+      "epoch": 11.079136690647482,
+      "step": 12320
+    },
+    {
+      "epoch": 11.079136690647482,
+      "step": 12320,
+      "torque_loss": 0.12021370977163315
+    },
+    {
+      "epoch": 11.079136690647482,
+      "force_loss": 0.00739591708406806,
+      "step": 12320
+    },
+    {
+      "epoch": 11.088129496402878,
+      "grad_norm": 0.3833772838115692,
+      "learning_rate": 9.353496301545529e-05,
+      "loss": 0.0263,
+      "step": 12330
+    },
+    {
+      "action_loss": 0.010584007017314434,
+      "epoch": 11.088129496402878,
+      "step": 12330
+    },
+    {
+      "epoch": 11.088129496402878,
+      "step": 12330,
+      "torque_loss": 0.13035576045513153
+    },
+    {
+      "epoch": 11.088129496402878,
+      "force_loss": 0.01557536143809557,
+      "step": 12330
+    },
+    {
+      "epoch": 11.097122302158274,
+      "grad_norm": 0.6187864542007446,
+      "learning_rate": 9.352140302692733e-05,
+      "loss": 0.0269,
+      "step": 12340
+    },
+    {
+      "action_loss": 0.018573325127363205,
+      "epoch": 11.097122302158274,
+      "step": 12340
+    },
+    {
+      "epoch": 11.097122302158274,
+      "step": 12340,
+      "torque_loss": 0.17013008892536163
+    },
+    {
+      "epoch": 11.097122302158274,
+      "force_loss": 0.024161985144019127,
+      "step": 12340
+    },
+    {
+      "epoch": 11.10611510791367,
+      "grad_norm": 0.9494650363922119,
+      "learning_rate": 9.350782981774627e-05,
+      "loss": 0.0285,
+      "step": 12350
+    },
+    {
+      "action_loss": 0.011366683058440685,
+      "epoch": 11.10611510791367,
+      "step": 12350
+    },
+    {
+      "epoch": 11.10611510791367,
+      "step": 12350,
+      "torque_loss": 0.15562209486961365
+    },
+    {
+      "epoch": 11.10611510791367,
+      "force_loss": 0.010943818837404251,
+      "step": 12350
+    },
+    {
+      "epoch": 11.115107913669064,
+      "grad_norm": 0.8779278993606567,
+      "learning_rate": 9.349424339203526e-05,
+      "loss": 0.0271,
+      "step": 12360
+    },
+    {
+      "action_loss": 0.014526404440402985,
+      "epoch": 11.115107913669064,
+      "step": 12360
+    },
+    {
+      "epoch": 11.115107913669064,
+      "step": 12360,
+      "torque_loss": 0.13063812255859375
+    },
+    {
+      "epoch": 11.115107913669064,
+      "force_loss": 0.015571859665215015,
+      "step": 12360
+    },
+    {
+      "epoch": 11.12410071942446,
+      "grad_norm": 0.806794285774231,
+      "learning_rate": 9.34806437539215e-05,
+      "loss": 0.0289,
+      "step": 12370
+    },
+    {
+      "action_loss": 0.01971241645514965,
+      "epoch": 11.12410071942446,
+      "step": 12370
+    },
+    {
+      "epoch": 11.12410071942446,
+      "step": 12370,
+      "torque_loss": 0.15413321554660797
+    },
+    {
+      "epoch": 11.12410071942446,
+      "force_loss": 0.025549190118908882,
+      "step": 12370
+    },
+    {
+      "epoch": 11.133093525179856,
+      "grad_norm": 0.6933779120445251,
+      "learning_rate": 9.346703090753622e-05,
+      "loss": 0.0299,
+      "step": 12380
+    },
+    {
+      "action_loss": 0.011384930461645126,
+      "epoch": 11.133093525179856,
+      "step": 12380
+    },
+    {
+      "epoch": 11.133093525179856,
+      "step": 12380,
+      "torque_loss": 0.17711804807186127
+    },
+    {
+      "epoch": 11.133093525179856,
+      "force_loss": 0.011022083461284637,
+      "step": 12380
+    },
+    {
+      "epoch": 11.142086330935252,
+      "grad_norm": 0.6616423726081848,
+      "learning_rate": 9.345340485701461e-05,
+      "loss": 0.0296,
+      "step": 12390
+    },
+    {
+      "action_loss": 0.010563376359641552,
+      "epoch": 11.142086330935252,
+      "step": 12390
+    },
+    {
+      "epoch": 11.142086330935252,
+      "step": 12390,
+      "torque_loss": 0.12168272584676743
+    },
+    {
+      "epoch": 11.142086330935252,
+      "force_loss": 0.006697139237076044,
+      "step": 12390
+    },
+    {
+      "epoch": 11.151079136690647,
+      "grad_norm": 0.5267893075942993,
+      "learning_rate": 9.343976560649595e-05,
+      "loss": 0.0235,
+      "step": 12400
+    },
+    {
+      "action_loss": 0.008783641271293163,
+      "epoch": 11.151079136690647,
+      "step": 12400
+    },
+    {
+      "epoch": 11.151079136690647,
+      "step": 12400,
+      "torque_loss": 0.15881018340587616
+    },
+    {
+      "epoch": 11.151079136690647,
+      "force_loss": 0.01086476445198059,
+      "step": 12400
+    },
+    {
+      "epoch": 11.160071942446043,
+      "grad_norm": 0.3190533518791199,
+      "learning_rate": 9.342611316012344e-05,
+      "loss": 0.0258,
+      "step": 12410
+    },
+    {
+      "action_loss": 0.014200364239513874,
+      "epoch": 11.160071942446043,
+      "step": 12410
+    },
+    {
+      "epoch": 11.160071942446043,
+      "step": 12410,
+      "torque_loss": 0.15340310335159302
+    },
+    {
+      "epoch": 11.160071942446043,
+      "force_loss": 0.014808292500674725,
+      "step": 12410
+    },
+    {
+      "epoch": 11.16906474820144,
+      "grad_norm": 0.8420072197914124,
+      "learning_rate": 9.341244752204437e-05,
+      "loss": 0.0282,
+      "step": 12420
+    },
+    {
+      "action_loss": 0.011746775358915329,
+      "epoch": 11.16906474820144,
+      "step": 12420
+    },
+    {
+      "epoch": 11.16906474820144,
+      "step": 12420,
+      "torque_loss": 0.14641153812408447
+    },
+    {
+      "epoch": 11.16906474820144,
+      "force_loss": 0.00971127301454544,
+      "step": 12420
+    },
+    {
+      "epoch": 11.178057553956835,
+      "grad_norm": 0.36430823802948,
+      "learning_rate": 9.339876869640995e-05,
+      "loss": 0.0216,
+      "step": 12430
+    },
+    {
+      "action_loss": 0.009064631536602974,
+      "epoch": 11.178057553956835,
+      "step": 12430
+    },
+    {
+      "epoch": 11.178057553956835,
+      "step": 12430,
+      "torque_loss": 0.13181424140930176
+    },
+    {
+      "epoch": 11.178057553956835,
+      "force_loss": 0.011014536023139954,
+      "step": 12430
+    },
+    {
+      "epoch": 11.18705035971223,
+      "grad_norm": 0.36544182896614075,
+      "learning_rate": 9.33850766873755e-05,
+      "loss": 0.0289,
+      "step": 12440
+    },
+    {
+      "action_loss": 0.02069510892033577,
+      "epoch": 11.18705035971223,
+      "step": 12440
+    },
+    {
+      "epoch": 11.18705035971223,
+      "step": 12440,
+      "torque_loss": 0.1885557919740677
+    },
+    {
+      "epoch": 11.18705035971223,
+      "force_loss": 0.02032717876136303,
+      "step": 12440
+    },
+    {
+      "epoch": 11.196043165467627,
+      "grad_norm": 0.8094887137413025,
+      "learning_rate": 9.337137149910028e-05,
+      "loss": 0.0298,
+      "step": 12450
+    },
+    {
+      "action_loss": 0.0070904274471104145,
+      "epoch": 11.196043165467627,
+      "step": 12450
+    },
+    {
+      "epoch": 11.196043165467627,
+      "step": 12450,
+      "torque_loss": 0.12666967511177063
+    },
+    {
+      "epoch": 11.196043165467627,
+      "force_loss": 0.006781961768865585,
+      "step": 12450
+    },
+    {
+      "epoch": 11.20503597122302,
+      "grad_norm": 0.6613590121269226,
+      "learning_rate": 9.335765313574753e-05,
+      "loss": 0.0242,
+      "step": 12460
+    },
+    {
+      "action_loss": 0.008711927570402622,
+      "epoch": 11.20503597122302,
+      "step": 12460
+    },
+    {
+      "epoch": 11.20503597122302,
+      "step": 12460,
+      "torque_loss": 0.139108344912529
+    },
+    {
+      "epoch": 11.20503597122302,
+      "force_loss": 0.005266461055725813,
+      "step": 12460
+    },
+    {
+      "epoch": 11.214028776978417,
+      "grad_norm": 0.5062190890312195,
+      "learning_rate": 9.334392160148457e-05,
+      "loss": 0.0262,
+      "step": 12470
+    },
+    {
+      "action_loss": 0.008867678232491016,
+      "epoch": 11.214028776978417,
+      "step": 12470
+    },
+    {
+      "epoch": 11.214028776978417,
+      "step": 12470,
+      "torque_loss": 0.12036507576704025
+    },
+    {
+      "epoch": 11.214028776978417,
+      "force_loss": 0.015314076095819473,
+      "step": 12470
+    },
+    {
+      "epoch": 11.223021582733812,
+      "grad_norm": 0.922947108745575,
+      "learning_rate": 9.333017690048264e-05,
+      "loss": 0.0279,
+      "step": 12480
+    },
+    {
+      "action_loss": 0.008514330722391605,
+      "epoch": 11.223021582733812,
+      "step": 12480
+    },
+    {
+      "epoch": 11.223021582733812,
+      "step": 12480,
+      "torque_loss": 0.11133581399917603
+    },
+    {
+      "epoch": 11.223021582733812,
+      "force_loss": 0.006068896502256393,
+      "step": 12480
+    },
+    {
+      "epoch": 11.232014388489208,
+      "grad_norm": 0.8144896626472473,
+      "learning_rate": 9.331641903691706e-05,
+      "loss": 0.0259,
+      "step": 12490
+    },
+    {
+      "action_loss": 0.014711529016494751,
+      "epoch": 11.232014388489208,
+      "step": 12490
+    },
+    {
+      "epoch": 11.232014388489208,
+      "step": 12490,
+      "torque_loss": 0.12662211060523987
+    },
+    {
+      "epoch": 11.232014388489208,
+      "force_loss": 0.012998889200389385,
+      "step": 12490
+    },
+    {
+      "epoch": 11.241007194244604,
+      "grad_norm": 0.495222270488739,
+      "learning_rate": 9.330264801496707e-05,
+      "loss": 0.0343,
+      "step": 12500
+    },
+    {
+      "action_loss": 0.021865472197532654,
+      "epoch": 11.241007194244604,
+      "step": 12500
+    },
+    {
+      "epoch": 11.241007194244604,
+      "step": 12500,
+      "torque_loss": 0.18252168595790863
+    },
+    {
+      "epoch": 11.241007194244604,
+      "force_loss": 0.028112413361668587,
+      "step": 12500
+    },
+    {
+      "epoch": 11.25,
+      "grad_norm": 0.42036470770835876,
+      "learning_rate": 9.328886383881594e-05,
+      "loss": 0.0324,
+      "step": 12510
+    },
+    {
+      "action_loss": 0.019588559865951538,
+      "epoch": 11.25,
+      "step": 12510
+    },
+    {
+      "epoch": 11.25,
+      "step": 12510,
+      "torque_loss": 0.21246810257434845
+    },
+    {
+      "epoch": 11.25,
+      "force_loss": 0.00969478115439415,
+      "step": 12510
+    },
+    {
+      "epoch": 11.258992805755396,
+      "grad_norm": 0.49514368176460266,
+      "learning_rate": 9.327506651265095e-05,
+      "loss": 0.0338,
+      "step": 12520
+    },
+    {
+      "action_loss": 0.01497744768857956,
+      "epoch": 11.258992805755396,
+      "step": 12520
+    },
+    {
+      "epoch": 11.258992805755396,
+      "step": 12520,
+      "torque_loss": 0.12542618811130524
+    },
+    {
+      "epoch": 11.258992805755396,
+      "force_loss": 0.009002608247101307,
+      "step": 12520
+    },
+    {
+      "epoch": 11.267985611510792,
+      "grad_norm": 0.5654804706573486,
+      "learning_rate": 9.326125604066338e-05,
+      "loss": 0.0287,
+      "step": 12530
+    },
+    {
+      "action_loss": 0.005585464183241129,
+      "epoch": 11.267985611510792,
+      "step": 12530
+    },
+    {
+      "epoch": 11.267985611510792,
+      "step": 12530,
+      "torque_loss": 0.10648307204246521
+    },
+    {
+      "epoch": 11.267985611510792,
+      "force_loss": 0.00592239061370492,
+      "step": 12530
+    },
+    {
+      "epoch": 11.276978417266188,
+      "grad_norm": 0.49654582142829895,
+      "learning_rate": 9.324743242704847e-05,
+      "loss": 0.0258,
+      "step": 12540
+    },
+    {
+      "action_loss": 0.01231973897665739,
+      "epoch": 11.276978417266188,
+      "step": 12540
+    },
+    {
+      "epoch": 11.276978417266188,
+      "step": 12540,
+      "torque_loss": 0.19205810129642487
+    },
+    {
+      "epoch": 11.276978417266188,
+      "force_loss": 0.011897188611328602,
+      "step": 12540
+    },
+    {
+      "epoch": 11.285971223021583,
+      "grad_norm": 0.4989960789680481,
+      "learning_rate": 9.323359567600546e-05,
+      "loss": 0.0309,
+      "step": 12550
+    },
+    {
+      "action_loss": 0.014130854047834873,
+      "epoch": 11.285971223021583,
+      "step": 12550
+    },
+    {
+      "epoch": 11.285971223021583,
+      "step": 12550,
+      "torque_loss": 0.13981622457504272
+    },
+    {
+      "epoch": 11.285971223021583,
+      "force_loss": 0.018393585458397865,
+      "step": 12550
+    },
+    {
+      "epoch": 11.29496402877698,
+      "grad_norm": 0.693494975566864,
+      "learning_rate": 9.321974579173761e-05,
+      "loss": 0.0252,
+      "step": 12560
+    },
+    {
+      "action_loss": 0.007720166817307472,
+      "epoch": 11.29496402877698,
+      "step": 12560
+    },
+    {
+      "epoch": 11.29496402877698,
+      "step": 12560,
+      "torque_loss": 0.1686241775751114
+    },
+    {
+      "epoch": 11.29496402877698,
+      "force_loss": 0.006033910438418388,
+      "step": 12560
+    },
+    {
+      "epoch": 11.303956834532373,
+      "grad_norm": 0.44988879561424255,
+      "learning_rate": 9.320588277845213e-05,
+      "loss": 0.0345,
+      "step": 12570
+    },
+    {
+      "action_loss": 0.013009771704673767,
+      "epoch": 11.303956834532373,
+      "step": 12570
+    },
+    {
+      "epoch": 11.303956834532373,
+      "step": 12570,
+      "torque_loss": 0.18443401157855988
+    },
+    {
+      "epoch": 11.303956834532373,
+      "force_loss": 0.014809790067374706,
+      "step": 12570
+    },
+    {
+      "epoch": 11.31294964028777,
+      "grad_norm": 0.3342968225479126,
+      "learning_rate": 9.319200664036026e-05,
+      "loss": 0.0257,
+      "step": 12580
+    },
+    {
+      "action_loss": 0.004076090175658464,
+      "epoch": 11.31294964028777,
+      "step": 12580
+    },
+    {
+      "epoch": 11.31294964028777,
+      "step": 12580,
+      "torque_loss": 0.12488287687301636
+    },
+    {
+      "epoch": 11.31294964028777,
+      "force_loss": 0.004519579000771046,
+      "step": 12580
+    },
+    {
+      "epoch": 11.321942446043165,
+      "grad_norm": 0.6651833057403564,
+      "learning_rate": 9.31781173816772e-05,
+      "loss": 0.0245,
+      "step": 12590
+    },
+    {
+      "action_loss": 0.011581075377762318,
+      "epoch": 11.321942446043165,
+      "step": 12590
+    },
+    {
+      "epoch": 11.321942446043165,
+      "step": 12590,
+      "torque_loss": 0.13377225399017334
+    },
+    {
+      "epoch": 11.321942446043165,
+      "force_loss": 0.014463149011135101,
+      "step": 12590
+    },
+    {
+      "epoch": 11.33093525179856,
+      "grad_norm": 0.33726048469543457,
+      "learning_rate": 9.316421500662212e-05,
+      "loss": 0.0249,
+      "step": 12600
+    },
+    {
+      "action_loss": 0.0035350958351045847,
+      "epoch": 11.33093525179856,
+      "step": 12600
+    },
+    {
+      "epoch": 11.33093525179856,
+      "step": 12600,
+      "torque_loss": 0.08888860791921616
+    },
+    {
+      "epoch": 11.33093525179856,
+      "force_loss": 0.004291305784136057,
+      "step": 12600
+    },
+    {
+      "epoch": 11.339928057553957,
+      "grad_norm": 0.34957993030548096,
+      "learning_rate": 9.31502995194182e-05,
+      "loss": 0.0263,
+      "step": 12610
+    },
+    {
+      "action_loss": 0.009594803676009178,
+      "epoch": 11.339928057553957,
+      "step": 12610
+    },
+    {
+      "epoch": 11.339928057553957,
+      "step": 12610,
+      "torque_loss": 0.17818152904510498
+    },
+    {
+      "epoch": 11.339928057553957,
+      "force_loss": 0.009814041666686535,
+      "step": 12610
+    },
+    {
+      "epoch": 11.348920863309353,
+      "grad_norm": 0.3459634482860565,
+      "learning_rate": 9.31363709242926e-05,
+      "loss": 0.0276,
+      "step": 12620
+    },
+    {
+      "action_loss": 0.012794236652553082,
+      "epoch": 11.348920863309353,
+      "step": 12620
+    },
+    {
+      "epoch": 11.348920863309353,
+      "step": 12620,
+      "torque_loss": 0.14862807095050812
+    },
+    {
+      "epoch": 11.348920863309353,
+      "force_loss": 0.008390388451516628,
+      "step": 12620
+    },
+    {
+      "epoch": 11.357913669064748,
+      "grad_norm": 0.31086641550064087,
+      "learning_rate": 9.312242922547647e-05,
+      "loss": 0.0209,
+      "step": 12630
+    },
+    {
+      "action_loss": 0.008075055666267872,
+      "epoch": 11.357913669064748,
+      "step": 12630
+    },
+    {
+      "epoch": 11.357913669064748,
+      "step": 12630,
+      "torque_loss": 0.1388295441865921
+    },
+    {
+      "epoch": 11.357913669064748,
+      "force_loss": 0.010876013897359371,
+      "step": 12630
+    },
+    {
+      "epoch": 11.366906474820144,
+      "grad_norm": 0.515978991985321,
+      "learning_rate": 9.310847442720492e-05,
+      "loss": 0.0281,
+      "step": 12640
+    },
+    {
+      "action_loss": 0.01441405713558197,
+      "epoch": 11.366906474820144,
+      "step": 12640
+    },
+    {
+      "epoch": 11.366906474820144,
+      "step": 12640,
+      "torque_loss": 0.1357070356607437
+    },
+    {
+      "epoch": 11.366906474820144,
+      "force_loss": 0.008753783069550991,
+      "step": 12640
+    },
+    {
+      "epoch": 11.37589928057554,
+      "grad_norm": 0.443697452545166,
+      "learning_rate": 9.309450653371706e-05,
+      "loss": 0.0264,
+      "step": 12650
+    },
+    {
+      "action_loss": 0.014335624873638153,
+      "epoch": 11.37589928057554,
+      "step": 12650
+    },
+    {
+      "epoch": 11.37589928057554,
+      "step": 12650,
+      "torque_loss": 0.13199064135551453
+    },
+    {
+      "epoch": 11.37589928057554,
+      "force_loss": 0.015366551466286182,
+      "step": 12650
+    },
+    {
+      "epoch": 11.384892086330936,
+      "grad_norm": 0.8964945673942566,
+      "learning_rate": 9.308052554925595e-05,
+      "loss": 0.0264,
+      "step": 12660
+    },
+    {
+      "action_loss": 0.02808235026896,
+      "epoch": 11.384892086330936,
+      "step": 12660
+    },
+    {
+      "epoch": 11.384892086330936,
+      "step": 12660,
+      "torque_loss": 0.14010505378246307
+    },
+    {
+      "epoch": 11.384892086330936,
+      "force_loss": 0.03326769545674324,
+      "step": 12660
+    },
+    {
+      "epoch": 11.39388489208633,
+      "grad_norm": 0.5438146591186523,
+      "learning_rate": 9.306653147806867e-05,
+      "loss": 0.0276,
+      "step": 12670
+    },
+    {
+      "action_loss": 0.0042573525570333,
+      "epoch": 11.39388489208633,
+      "step": 12670
+    },
+    {
+      "epoch": 11.39388489208633,
+      "step": 12670,
+      "torque_loss": 0.10584481805562973
+    },
+    {
+      "epoch": 11.39388489208633,
+      "force_loss": 0.004664700012654066,
+      "step": 12670
+    },
+    {
+      "epoch": 11.402877697841726,
+      "grad_norm": 0.3931425213813782,
+      "learning_rate": 9.305252432440622e-05,
+      "loss": 0.0258,
+      "step": 12680
+    },
+    {
+      "action_loss": 0.005934089422225952,
+      "epoch": 11.402877697841726,
+      "step": 12680
+    },
+    {
+      "epoch": 11.402877697841726,
+      "step": 12680,
+      "torque_loss": 0.10848554223775864
+    },
+    {
+      "epoch": 11.402877697841726,
+      "force_loss": 0.011862953193485737,
+      "step": 12680
+    },
+    {
+      "epoch": 11.411870503597122,
+      "grad_norm": 0.4016677439212799,
+      "learning_rate": 9.303850409252361e-05,
+      "loss": 0.0237,
+      "step": 12690
+    },
+    {
+      "action_loss": 0.008334841579198837,
+      "epoch": 11.411870503597122,
+      "step": 12690
+    },
+    {
+      "epoch": 11.411870503597122,
+      "step": 12690,
+      "torque_loss": 0.11518295854330063
+    },
+    {
+      "epoch": 11.411870503597122,
+      "force_loss": 0.0063664354383945465,
+      "step": 12690
+    },
+    {
+      "epoch": 11.420863309352518,
+      "grad_norm": 0.502417266368866,
+      "learning_rate": 9.302447078667985e-05,
+      "loss": 0.0236,
+      "step": 12700
+    },
+    {
+      "action_loss": 0.00721868546679616,
+      "epoch": 11.420863309352518,
+      "step": 12700
+    },
+    {
+      "epoch": 11.420863309352518,
+      "step": 12700,
+      "torque_loss": 0.11511284857988358
+    },
+    {
+      "epoch": 11.420863309352518,
+      "force_loss": 0.0068063754588365555,
+      "step": 12700
+    },
+    {
+      "epoch": 11.429856115107913,
+      "grad_norm": 0.7590480446815491,
+      "learning_rate": 9.301042441113783e-05,
+      "loss": 0.025,
+      "step": 12710
+    },
+    {
+      "action_loss": 0.008520825766026974,
+      "epoch": 11.429856115107913,
+      "step": 12710
+    },
+    {
+      "epoch": 11.429856115107913,
+      "step": 12710,
+      "torque_loss": 0.15272606909275055
+    },
+    {
+      "epoch": 11.429856115107913,
+      "force_loss": 0.007586202118545771,
+      "step": 12710
+    },
+    {
+      "epoch": 11.43884892086331,
+      "grad_norm": 0.3182271718978882,
+      "learning_rate": 9.299636497016451e-05,
+      "loss": 0.0304,
+      "step": 12720
+    },
+    {
+      "action_loss": 0.042789313942193985,
+      "epoch": 11.43884892086331,
+      "step": 12720
+    },
+    {
+      "epoch": 11.43884892086331,
+      "step": 12720,
+      "torque_loss": 0.20183898508548737
+    },
+    {
+      "epoch": 11.43884892086331,
+      "force_loss": 0.060008030384778976,
+      "step": 12720
+    },
+    {
+      "epoch": 11.447841726618705,
+      "grad_norm": 0.46768414974212646,
+      "learning_rate": 9.298229246803076e-05,
+      "loss": 0.0268,
+      "step": 12730
+    },
+    {
+      "action_loss": 0.008419915102422237,
+      "epoch": 11.447841726618705,
+      "step": 12730
+    },
+    {
+      "epoch": 11.447841726618705,
+      "step": 12730,
+      "torque_loss": 0.09191560000181198
+    },
+    {
+      "epoch": 11.447841726618705,
+      "force_loss": 0.005917018745094538,
+      "step": 12730
+    },
+    {
+      "epoch": 11.456834532374101,
+      "grad_norm": 0.7658388018608093,
+      "learning_rate": 9.296820690901144e-05,
+      "loss": 0.0294,
+      "step": 12740
+    },
+    {
+      "action_loss": 0.004596956539899111,
+      "epoch": 11.456834532374101,
+      "step": 12740
+    },
+    {
+      "epoch": 11.456834532374101,
+      "step": 12740,
+      "torque_loss": 0.11617550998926163
+    },
+    {
+      "epoch": 11.456834532374101,
+      "force_loss": 0.004909941460937262,
+      "step": 12740
+    },
+    {
+      "epoch": 11.465827338129497,
+      "grad_norm": 0.41740214824676514,
+      "learning_rate": 9.295410829738539e-05,
+      "loss": 0.0254,
+      "step": 12750
+    },
+    {
+      "action_loss": 0.009858358651399612,
+      "epoch": 11.465827338129497,
+      "step": 12750
+    },
+    {
+      "epoch": 11.465827338129497,
+      "step": 12750,
+      "torque_loss": 0.14883457124233246
+    },
+    {
+      "epoch": 11.465827338129497,
+      "force_loss": 0.011447202414274216,
+      "step": 12750
+    },
+    {
+      "epoch": 11.474820143884893,
+      "grad_norm": 0.5770699977874756,
+      "learning_rate": 9.293999663743535e-05,
+      "loss": 0.0262,
+      "step": 12760
+    },
+    {
+      "action_loss": 0.010206750594079494,
+      "epoch": 11.474820143884893,
+      "step": 12760
+    },
+    {
+      "epoch": 11.474820143884893,
+      "step": 12760,
+      "torque_loss": 0.185093954205513
+    },
+    {
+      "epoch": 11.474820143884893,
+      "force_loss": 0.018447376787662506,
+      "step": 12760
+    },
+    {
+      "epoch": 11.483812949640289,
+      "grad_norm": 0.5014813542366028,
+      "learning_rate": 9.292587193344813e-05,
+      "loss": 0.0303,
+      "step": 12770
+    },
+    {
+      "action_loss": 0.009192741475999355,
+      "epoch": 11.483812949640289,
+      "step": 12770
+    },
+    {
+      "epoch": 11.483812949640289,
+      "step": 12770,
+      "torque_loss": 0.14501263201236725
+    },
+    {
+      "epoch": 11.483812949640289,
+      "force_loss": 0.007580628152936697,
+      "step": 12770
+    },
+    {
+      "epoch": 11.492805755395683,
+      "grad_norm": 0.44411325454711914,
+      "learning_rate": 9.291173418971437e-05,
+      "loss": 0.0259,
+      "step": 12780
+    },
+    {
+      "action_loss": 0.0186526607722044,
+      "epoch": 11.492805755395683,
+      "step": 12780
+    },
+    {
+      "epoch": 11.492805755395683,
+      "step": 12780,
+      "torque_loss": 0.17377136647701263
+    },
+    {
+      "epoch": 11.492805755395683,
+      "force_loss": 0.014825411140918732,
+      "step": 12780
+    },
+    {
+      "epoch": 11.501798561151078,
+      "grad_norm": 0.4844873547554016,
+      "learning_rate": 9.28975834105288e-05,
+      "loss": 0.0278,
+      "step": 12790
+    },
+    {
+      "action_loss": 0.020406672731041908,
+      "epoch": 11.501798561151078,
+      "step": 12790
+    },
+    {
+      "epoch": 11.501798561151078,
+      "step": 12790,
+      "torque_loss": 0.1572578400373459
+    },
+    {
+      "epoch": 11.501798561151078,
+      "force_loss": 0.016273969784379005,
+      "step": 12790
+    },
+    {
+      "epoch": 11.510791366906474,
+      "grad_norm": 0.40174224972724915,
+      "learning_rate": 9.288341960019004e-05,
+      "loss": 0.0263,
+      "step": 12800
+    },
+    {
+      "action_loss": 0.004163198638707399,
+      "epoch": 11.510791366906474,
+      "step": 12800
+    },
+    {
+      "epoch": 11.510791366906474,
+      "step": 12800,
+      "torque_loss": 0.09697920083999634
+    },
+    {
+      "epoch": 11.510791366906474,
+      "force_loss": 0.0047917976044118404,
+      "step": 12800
+    },
+    {
+      "epoch": 11.51978417266187,
+      "grad_norm": 0.5060114860534668,
+      "learning_rate": 9.286924276300067e-05,
+      "loss": 0.0281,
+      "step": 12810
+    },
+    {
+      "action_loss": 0.00661865109577775,
+      "epoch": 11.51978417266187,
+      "step": 12810
+    },
+    {
+      "epoch": 11.51978417266187,
+      "step": 12810,
+      "torque_loss": 0.12898625433444977
+    },
+    {
+      "epoch": 11.51978417266187,
+      "force_loss": 0.010716870427131653,
+      "step": 12810
+    },
+    {
+      "epoch": 11.528776978417266,
+      "grad_norm": 0.5622050762176514,
+      "learning_rate": 9.285505290326726e-05,
+      "loss": 0.0256,
+      "step": 12820
+    },
+    {
+      "action_loss": 0.004694139119237661,
+      "epoch": 11.528776978417266,
+      "step": 12820
+    },
+    {
+      "epoch": 11.528776978417266,
+      "step": 12820,
+      "torque_loss": 0.10522476583719254
+    },
+    {
+      "epoch": 11.528776978417266,
+      "force_loss": 0.005791978910565376,
+      "step": 12820
+    },
+    {
+      "epoch": 11.537769784172662,
+      "grad_norm": 0.49937760829925537,
+      "learning_rate": 9.284085002530027e-05,
+      "loss": 0.0226,
+      "step": 12830
+    },
+    {
+      "action_loss": 0.0157787948846817,
+      "epoch": 11.537769784172662,
+      "step": 12830
+    },
+    {
+      "epoch": 11.537769784172662,
+      "step": 12830,
+      "torque_loss": 0.14168919622898102
+    },
+    {
+      "epoch": 11.537769784172662,
+      "force_loss": 0.013614925555884838,
+      "step": 12830
+    },
+    {
+      "epoch": 11.546762589928058,
+      "grad_norm": 0.48086875677108765,
+      "learning_rate": 9.282663413341422e-05,
+      "loss": 0.0314,
+      "step": 12840
+    },
+    {
+      "action_loss": 0.00636838236823678,
+      "epoch": 11.546762589928058,
+      "step": 12840
+    },
+    {
+      "epoch": 11.546762589928058,
+      "step": 12840,
+      "torque_loss": 0.15166963636875153
+    },
+    {
+      "epoch": 11.546762589928058,
+      "force_loss": 0.006735137198120356,
+      "step": 12840
+    },
+    {
+      "epoch": 11.555755395683454,
+      "grad_norm": 0.4735391139984131,
+      "learning_rate": 9.281240523192747e-05,
+      "loss": 0.0235,
+      "step": 12850
+    },
+    {
+      "action_loss": 0.013812369666993618,
+      "epoch": 11.555755395683454,
+      "step": 12850
+    },
+    {
+      "epoch": 11.555755395683454,
+      "step": 12850,
+      "torque_loss": 0.1245051696896553
+    },
+    {
+      "epoch": 11.555755395683454,
+      "force_loss": 0.016836203634738922,
+      "step": 12850
+    },
+    {
+      "epoch": 11.56474820143885,
+      "grad_norm": 0.49611353874206543,
+      "learning_rate": 9.279816332516242e-05,
+      "loss": 0.0232,
+      "step": 12860
+    },
+    {
+      "action_loss": 0.018690388649702072,
+      "epoch": 11.56474820143885,
+      "step": 12860
+    },
+    {
+      "epoch": 11.56474820143885,
+      "step": 12860,
+      "torque_loss": 0.14858348667621613
+    },
+    {
+      "epoch": 11.56474820143885,
+      "force_loss": 0.017054811120033264,
+      "step": 12860
+    },
+    {
+      "epoch": 11.573741007194245,
+      "grad_norm": 0.6356492638587952,
+      "learning_rate": 9.278390841744536e-05,
+      "loss": 0.0288,
+      "step": 12870
+    },
+    {
+      "action_loss": 0.005330819170922041,
+      "epoch": 11.573741007194245,
+      "step": 12870
+    },
+    {
+      "epoch": 11.573741007194245,
+      "step": 12870,
+      "torque_loss": 0.11260536313056946
+    },
+    {
+      "epoch": 11.573741007194245,
+      "force_loss": 0.004948218818753958,
+      "step": 12870
+    },
+    {
+      "epoch": 11.582733812949641,
+      "grad_norm": 0.7857191562652588,
+      "learning_rate": 9.276964051310658e-05,
+      "loss": 0.0259,
+      "step": 12880
+    },
+    {
+      "action_loss": 0.010961071588099003,
+      "epoch": 11.582733812949641,
+      "step": 12880
+    },
+    {
+      "epoch": 11.582733812949641,
+      "step": 12880,
+      "torque_loss": 0.1459680199623108
+    },
+    {
+      "epoch": 11.582733812949641,
+      "force_loss": 0.010086453519761562,
+      "step": 12880
+    },
+    {
+      "epoch": 11.591726618705035,
+      "grad_norm": 0.7839841842651367,
+      "learning_rate": 9.275535961648027e-05,
+      "loss": 0.0221,
+      "step": 12890
+    },
+    {
+      "action_loss": 0.005960522685199976,
+      "epoch": 11.591726618705035,
+      "step": 12890
+    },
+    {
+      "epoch": 11.591726618705035,
+      "step": 12890,
+      "torque_loss": 0.15388502180576324
+    },
+    {
+      "epoch": 11.591726618705035,
+      "force_loss": 0.0074750198982656,
+      "step": 12890
+    },
+    {
+      "epoch": 11.600719424460431,
+      "grad_norm": 0.4410642385482788,
+      "learning_rate": 9.274106573190459e-05,
+      "loss": 0.0227,
+      "step": 12900
+    },
+    {
+      "action_loss": 0.009813723154366016,
+      "epoch": 11.600719424460431,
+      "step": 12900
+    },
+    {
+      "epoch": 11.600719424460431,
+      "step": 12900,
+      "torque_loss": 0.1576903611421585
+    },
+    {
+      "epoch": 11.600719424460431,
+      "force_loss": 0.00767472991719842,
+      "step": 12900
+    },
+    {
+      "epoch": 11.609712230215827,
+      "grad_norm": 0.4177224934101105,
+      "learning_rate": 9.272675886372168e-05,
+      "loss": 0.0223,
+      "step": 12910
+    },
+    {
+      "action_loss": 0.013900388963520527,
+      "epoch": 11.609712230215827,
+      "step": 12910
+    },
+    {
+      "epoch": 11.609712230215827,
+      "step": 12910,
+      "torque_loss": 0.13374271988868713
+    },
+    {
+      "epoch": 11.609712230215827,
+      "force_loss": 0.006211420521140099,
+      "step": 12910
+    },
+    {
+      "epoch": 11.618705035971223,
+      "grad_norm": 0.42956387996673584,
+      "learning_rate": 9.271243901627754e-05,
+      "loss": 0.0261,
+      "step": 12920
+    },
+    {
+      "action_loss": 0.007312507834285498,
+      "epoch": 11.618705035971223,
+      "step": 12920
+    },
+    {
+      "epoch": 11.618705035971223,
+      "step": 12920,
+      "torque_loss": 0.12143468111753464
+    },
+    {
+      "epoch": 11.618705035971223,
+      "force_loss": 0.006631376687437296,
+      "step": 12920
+    },
+    {
+      "epoch": 11.627697841726619,
+      "grad_norm": 1.1085774898529053,
+      "learning_rate": 9.269810619392219e-05,
+      "loss": 0.0234,
+      "step": 12930
+    },
+    {
+      "action_loss": 0.0069969408214092255,
+      "epoch": 11.627697841726619,
+      "step": 12930
+    },
+    {
+      "epoch": 11.627697841726619,
+      "step": 12930,
+      "torque_loss": 0.07154836505651474
+    },
+    {
+      "epoch": 11.627697841726619,
+      "force_loss": 0.005450129508972168,
+      "step": 12930
+    },
+    {
+      "epoch": 11.636690647482014,
+      "grad_norm": 0.5631125569343567,
+      "learning_rate": 9.268376040100955e-05,
+      "loss": 0.0296,
+      "step": 12940
+    },
+    {
+      "action_loss": 0.014250400476157665,
+      "epoch": 11.636690647482014,
+      "step": 12940
+    },
+    {
+      "epoch": 11.636690647482014,
+      "step": 12940,
+      "torque_loss": 0.1898178607225418
+    },
+    {
+      "epoch": 11.636690647482014,
+      "force_loss": 0.01934933476150036,
+      "step": 12940
+    },
+    {
+      "epoch": 11.64568345323741,
+      "grad_norm": 0.6192155480384827,
+      "learning_rate": 9.266940164189752e-05,
+      "loss": 0.0225,
+      "step": 12950
+    },
+    {
+      "action_loss": 0.022436590865254402,
+      "epoch": 11.64568345323741,
+      "step": 12950
+    },
+    {
+      "epoch": 11.64568345323741,
+      "step": 12950,
+      "torque_loss": 0.11736515909433365
+    },
+    {
+      "epoch": 11.64568345323741,
+      "force_loss": 0.01960713043808937,
+      "step": 12950
+    },
+    {
+      "epoch": 11.654676258992806,
+      "grad_norm": 0.6199759244918823,
+      "learning_rate": 9.265502992094787e-05,
+      "loss": 0.0266,
+      "step": 12960
+    },
+    {
+      "action_loss": 0.0043710218742489815,
+      "epoch": 11.654676258992806,
+      "step": 12960
+    },
+    {
+      "epoch": 11.654676258992806,
+      "step": 12960,
+      "torque_loss": 0.10056054592132568
+    },
+    {
+      "epoch": 11.654676258992806,
+      "force_loss": 0.004458553157746792,
+      "step": 12960
+    },
+    {
+      "epoch": 11.663669064748202,
+      "grad_norm": 0.4860585033893585,
+      "learning_rate": 9.264064524252638e-05,
+      "loss": 0.0229,
+      "step": 12970
+    },
+    {
+      "action_loss": 0.003602929413318634,
+      "epoch": 11.663669064748202,
+      "step": 12970
+    },
+    {
+      "epoch": 11.663669064748202,
+      "step": 12970,
+      "torque_loss": 0.13326184451580048
+    },
+    {
+      "epoch": 11.663669064748202,
+      "force_loss": 0.004667218774557114,
+      "step": 12970
+    },
+    {
+      "epoch": 11.672661870503598,
+      "grad_norm": 0.3669048547744751,
+      "learning_rate": 9.262624761100271e-05,
+      "loss": 0.0267,
+      "step": 12980
+    },
+    {
+      "action_loss": 0.005022808909416199,
+      "epoch": 11.672661870503598,
+      "step": 12980
+    },
+    {
+      "epoch": 11.672661870503598,
+      "step": 12980,
+      "torque_loss": 0.10578330606222153
+    },
+    {
+      "epoch": 11.672661870503598,
+      "force_loss": 0.006402730941772461,
+      "step": 12980
+    },
+    {
+      "epoch": 11.681654676258994,
+      "grad_norm": 0.7687297463417053,
+      "learning_rate": 9.261183703075051e-05,
+      "loss": 0.0245,
+      "step": 12990
+    },
+    {
+      "action_loss": 0.015614025294780731,
+      "epoch": 11.681654676258994,
+      "step": 12990
+    },
+    {
+      "epoch": 11.681654676258994,
+      "step": 12990,
+      "torque_loss": 0.1742136925458908
+    },
+    {
+      "epoch": 11.681654676258994,
+      "force_loss": 0.012966055423021317,
+      "step": 12990
+    },
+    {
+      "epoch": 11.690647482014388,
+      "grad_norm": 1.1082276105880737,
+      "learning_rate": 9.259741350614733e-05,
+      "loss": 0.0243,
+      "step": 13000
+    },
+    {
+      "action_loss": 0.0065927873365581036,
+      "epoch": 11.690647482014388,
+      "step": 13000
+    },
+    {
+      "epoch": 11.690647482014388,
+      "step": 13000,
+      "torque_loss": 0.15699200332164764
+    },
+    {
+      "epoch": 11.690647482014388,
+      "force_loss": 0.0077374279499053955,
+      "step": 13000
+    },
+    {
+      "epoch": 11.699640287769784,
+      "grad_norm": 0.3770151734352112,
+      "learning_rate": 9.258297704157464e-05,
+      "loss": 0.0279,
+      "step": 13010
+    },
+    {
+      "action_loss": 0.005304222460836172,
+      "epoch": 11.699640287769784,
+      "step": 13010
+    },
+    {
+      "epoch": 11.699640287769784,
+      "step": 13010,
+      "torque_loss": 0.11752446740865707
+    },
+    {
+      "epoch": 11.699640287769784,
+      "force_loss": 0.0047838119789958,
+      "step": 13010
+    },
+    {
+      "epoch": 11.70863309352518,
+      "grad_norm": 0.43294763565063477,
+      "learning_rate": 9.256852764141786e-05,
+      "loss": 0.0265,
+      "step": 13020
+    },
+    {
+      "action_loss": 0.010032732039690018,
+      "epoch": 11.70863309352518,
+      "step": 13020
+    },
+    {
+      "epoch": 11.70863309352518,
+      "step": 13020,
+      "torque_loss": 0.1645011454820633
+    },
+    {
+      "epoch": 11.70863309352518,
+      "force_loss": 0.014812429435551167,
+      "step": 13020
+    },
+    {
+      "epoch": 11.717625899280575,
+      "grad_norm": 0.5424838066101074,
+      "learning_rate": 9.255406531006634e-05,
+      "loss": 0.0241,
+      "step": 13030
+    },
+    {
+      "action_loss": 0.008985103107988834,
+      "epoch": 11.717625899280575,
+      "step": 13030
+    },
+    {
+      "epoch": 11.717625899280575,
+      "step": 13030,
+      "torque_loss": 0.14888577163219452
+    },
+    {
+      "epoch": 11.717625899280575,
+      "force_loss": 0.011289309710264206,
+      "step": 13030
+    },
+    {
+      "epoch": 11.726618705035971,
+      "grad_norm": 0.34489864110946655,
+      "learning_rate": 9.253959005191335e-05,
+      "loss": 0.036,
+      "step": 13040
+    },
+    {
+      "action_loss": 0.010893319733440876,
+      "epoch": 11.726618705035971,
+      "step": 13040
+    },
+    {
+      "epoch": 11.726618705035971,
+      "step": 13040,
+      "torque_loss": 0.16150763630867004
+    },
+    {
+      "epoch": 11.726618705035971,
+      "force_loss": 0.00955308135598898,
+      "step": 13040
+    },
+    {
+      "epoch": 11.735611510791367,
+      "grad_norm": 0.5194714069366455,
+      "learning_rate": 9.25251018713561e-05,
+      "loss": 0.0299,
+      "step": 13050
+    },
+    {
+      "action_loss": 0.007421460468322039,
+      "epoch": 11.735611510791367,
+      "step": 13050
+    },
+    {
+      "epoch": 11.735611510791367,
+      "step": 13050,
+      "torque_loss": 0.14697431027889252
+    },
+    {
+      "epoch": 11.735611510791367,
+      "force_loss": 0.010507498867809772,
+      "step": 13050
+    },
+    {
+      "epoch": 11.744604316546763,
+      "grad_norm": 0.5162916779518127,
+      "learning_rate": 9.251060077279571e-05,
+      "loss": 0.0236,
+      "step": 13060
+    },
+    {
+      "action_loss": 0.02863156795501709,
+      "epoch": 11.744604316546763,
+      "step": 13060
+    },
+    {
+      "epoch": 11.744604316546763,
+      "step": 13060,
+      "torque_loss": 0.18968182802200317
+    },
+    {
+      "epoch": 11.744604316546763,
+      "force_loss": 0.05160766467452049,
+      "step": 13060
+    },
+    {
+      "epoch": 11.753597122302159,
+      "grad_norm": 0.43360474705696106,
+      "learning_rate": 9.249608676063724e-05,
+      "loss": 0.0313,
+      "step": 13070
+    },
+    {
+      "action_loss": 0.009612124413251877,
+      "epoch": 11.753597122302159,
+      "step": 13070
+    },
+    {
+      "epoch": 11.753597122302159,
+      "step": 13070,
+      "torque_loss": 0.13511134684085846
+    },
+    {
+      "epoch": 11.753597122302159,
+      "force_loss": 0.010957022197544575,
+      "step": 13070
+    },
+    {
+      "epoch": 11.762589928057555,
+      "grad_norm": 0.3442375361919403,
+      "learning_rate": 9.248155983928964e-05,
+      "loss": 0.0215,
+      "step": 13080
+    },
+    {
+      "action_loss": 0.008205555379390717,
+      "epoch": 11.762589928057555,
+      "step": 13080
+    },
+    {
+      "epoch": 11.762589928057555,
+      "step": 13080,
+      "torque_loss": 0.09010195732116699
+    },
+    {
+      "epoch": 11.762589928057555,
+      "force_loss": 0.009156804531812668,
+      "step": 13080
+    },
+    {
+      "epoch": 11.77158273381295,
+      "grad_norm": 0.521977961063385,
+      "learning_rate": 9.246702001316583e-05,
+      "loss": 0.0251,
+      "step": 13090
+    },
+    {
+      "action_loss": 0.01836331933736801,
+      "epoch": 11.77158273381295,
+      "step": 13090
+    },
+    {
+      "epoch": 11.77158273381295,
+      "step": 13090,
+      "torque_loss": 0.19094258546829224
+    },
+    {
+      "epoch": 11.77158273381295,
+      "force_loss": 0.01802162639796734,
+      "step": 13090
+    },
+    {
+      "epoch": 11.780575539568346,
+      "grad_norm": 0.7244899868965149,
+      "learning_rate": 9.245246728668262e-05,
+      "loss": 0.0265,
+      "step": 13100
+    },
+    {
+      "action_loss": 0.010238493792712688,
+      "epoch": 11.780575539568346,
+      "step": 13100
+    },
+    {
+      "epoch": 11.780575539568346,
+      "step": 13100,
+      "torque_loss": 0.13114742934703827
+    },
+    {
+      "epoch": 11.780575539568346,
+      "force_loss": 0.011190635152161121,
+      "step": 13100
+    },
+    {
+      "epoch": 11.78956834532374,
+      "grad_norm": 0.7381954789161682,
+      "learning_rate": 9.243790166426073e-05,
+      "loss": 0.0297,
+      "step": 13110
+    },
+    {
+      "action_loss": 0.028110459446907043,
+      "epoch": 11.78956834532374,
+      "step": 13110
+    },
+    {
+      "epoch": 11.78956834532374,
+      "step": 13110,
+      "torque_loss": 0.14432667195796967
+    },
+    {
+      "epoch": 11.78956834532374,
+      "force_loss": 0.024785900488495827,
+      "step": 13110
+    },
+    {
+      "epoch": 11.798561151079136,
+      "grad_norm": 0.5881648659706116,
+      "learning_rate": 9.242332315032484e-05,
+      "loss": 0.0267,
+      "step": 13120
+    },
+    {
+      "action_loss": 0.007070481777191162,
+      "epoch": 11.798561151079136,
+      "step": 13120
+    },
+    {
+      "epoch": 11.798561151079136,
+      "step": 13120,
+      "torque_loss": 0.11087200790643692
+    },
+    {
+      "epoch": 11.798561151079136,
+      "force_loss": 0.007086490746587515,
+      "step": 13120
+    },
+    {
+      "epoch": 11.807553956834532,
+      "grad_norm": 0.5413809418678284,
+      "learning_rate": 9.240873174930349e-05,
+      "loss": 0.0325,
+      "step": 13130
+    },
+    {
+      "action_loss": 0.004751246888190508,
+      "epoch": 11.807553956834532,
+      "step": 13130
+    },
+    {
+      "epoch": 11.807553956834532,
+      "step": 13130,
+      "torque_loss": 0.1028279960155487
+    },
+    {
+      "epoch": 11.807553956834532,
+      "force_loss": 0.005823200102895498,
+      "step": 13130
+    },
+    {
+      "epoch": 11.816546762589928,
+      "grad_norm": 0.634873628616333,
+      "learning_rate": 9.239412746562917e-05,
+      "loss": 0.031,
+      "step": 13140
+    },
+    {
+      "action_loss": 0.019993731752038002,
+      "epoch": 11.816546762589928,
+      "step": 13140
+    },
+    {
+      "epoch": 11.816546762589928,
+      "step": 13140,
+      "torque_loss": 0.1469378024339676
+    },
+    {
+      "epoch": 11.816546762589928,
+      "force_loss": 0.010980661027133465,
+      "step": 13140
+    },
+    {
+      "epoch": 11.825539568345324,
+      "grad_norm": 0.4987242817878723,
+      "learning_rate": 9.237951030373828e-05,
+      "loss": 0.0343,
+      "step": 13150
+    },
+    {
+      "action_loss": 0.016738135367631912,
+      "epoch": 11.825539568345324,
+      "step": 13150
+    },
+    {
+      "epoch": 11.825539568345324,
+      "step": 13150,
+      "torque_loss": 0.16462506353855133
+    },
+    {
+      "epoch": 11.825539568345324,
+      "force_loss": 0.012665067799389362,
+      "step": 13150
+    },
+    {
+      "epoch": 11.83453237410072,
+      "grad_norm": 0.6414262652397156,
+      "learning_rate": 9.236488026807113e-05,
+      "loss": 0.0328,
+      "step": 13160
+    },
+    {
+      "action_loss": 0.006865693721920252,
+      "epoch": 11.83453237410072,
+      "step": 13160
+    },
+    {
+      "epoch": 11.83453237410072,
+      "step": 13160,
+      "torque_loss": 0.12536965310573578
+    },
+    {
+      "epoch": 11.83453237410072,
+      "force_loss": 0.00795132014900446,
+      "step": 13160
+    },
+    {
+      "epoch": 11.843525179856115,
+      "grad_norm": 0.4860328435897827,
+      "learning_rate": 9.235023736307193e-05,
+      "loss": 0.0293,
+      "step": 13170
+    },
+    {
+      "action_loss": 0.006817122921347618,
+      "epoch": 11.843525179856115,
+      "step": 13170
+    },
+    {
+      "epoch": 11.843525179856115,
+      "step": 13170,
+      "torque_loss": 0.12449777126312256
+    },
+    {
+      "epoch": 11.843525179856115,
+      "force_loss": 0.0076742880046367645,
+      "step": 13170
+    },
+    {
+      "epoch": 11.852517985611511,
+      "grad_norm": 0.6348298788070679,
+      "learning_rate": 9.233558159318881e-05,
+      "loss": 0.0263,
+      "step": 13180
+    },
+    {
+      "action_loss": 0.0048419334925711155,
+      "epoch": 11.852517985611511,
+      "step": 13180
+    },
+    {
+      "epoch": 11.852517985611511,
+      "step": 13180,
+      "torque_loss": 0.15208874642848969
+    },
+    {
+      "epoch": 11.852517985611511,
+      "force_loss": 0.006664233282208443,
+      "step": 13180
+    },
+    {
+      "epoch": 11.861510791366907,
+      "grad_norm": 0.449191153049469,
+      "learning_rate": 9.232091296287382e-05,
+      "loss": 0.024,
+      "step": 13190
+    },
+    {
+      "action_loss": 0.0073325857520103455,
+      "epoch": 11.861510791366907,
+      "step": 13190
+    },
+    {
+      "epoch": 11.861510791366907,
+      "step": 13190,
+      "torque_loss": 0.13216394186019897
+    },
+    {
+      "epoch": 11.861510791366907,
+      "force_loss": 0.010304437950253487,
+      "step": 13190
+    },
+    {
+      "epoch": 11.870503597122303,
+      "grad_norm": 1.0398387908935547,
+      "learning_rate": 9.230623147658288e-05,
+      "loss": 0.0274,
+      "step": 13200
+    },
+    {
+      "action_loss": 0.010095992125570774,
+      "epoch": 11.870503597122303,
+      "step": 13200
+    },
+    {
+      "epoch": 11.870503597122303,
+      "step": 13200,
+      "torque_loss": 0.09582933038473129
+    },
+    {
+      "epoch": 11.870503597122303,
+      "force_loss": 0.007539154961705208,
+      "step": 13200
+    },
+    {
+      "epoch": 11.879496402877697,
+      "grad_norm": 0.30475932359695435,
+      "learning_rate": 9.229153713877586e-05,
+      "loss": 0.0249,
+      "step": 13210
+    },
+    {
+      "action_loss": 0.01730423979461193,
+      "epoch": 11.879496402877697,
+      "step": 13210
+    },
+    {
+      "epoch": 11.879496402877697,
+      "step": 13210,
+      "torque_loss": 0.1708080768585205
+    },
+    {
+      "epoch": 11.879496402877697,
+      "force_loss": 0.01646292582154274,
+      "step": 13210
+    },
+    {
+      "epoch": 11.888489208633093,
+      "grad_norm": 0.47347113490104675,
+      "learning_rate": 9.227682995391649e-05,
+      "loss": 0.0304,
+      "step": 13220
+    },
+    {
+      "action_loss": 0.026272406801581383,
+      "epoch": 11.888489208633093,
+      "step": 13220
+    },
+    {
+      "epoch": 11.888489208633093,
+      "step": 13220,
+      "torque_loss": 0.15797697007656097
+    },
+    {
+      "epoch": 11.888489208633093,
+      "force_loss": 0.019032509997487068,
+      "step": 13220
+    },
+    {
+      "epoch": 11.897482014388489,
+      "grad_norm": 0.420165479183197,
+      "learning_rate": 9.226210992647243e-05,
+      "loss": 0.0269,
+      "step": 13230
+    },
+    {
+      "action_loss": 0.008027888834476471,
+      "epoch": 11.897482014388489,
+      "step": 13230
+    },
+    {
+      "epoch": 11.897482014388489,
+      "step": 13230,
+      "torque_loss": 0.13189497590065002
+    },
+    {
+      "epoch": 11.897482014388489,
+      "force_loss": 0.005640607327222824,
+      "step": 13230
+    },
+    {
+      "epoch": 11.906474820143885,
+      "grad_norm": 0.5762290954589844,
+      "learning_rate": 9.224737706091525e-05,
+      "loss": 0.024,
+      "step": 13240
+    },
+    {
+      "action_loss": 0.017476534470915794,
+      "epoch": 11.906474820143885,
+      "step": 13240
+    },
+    {
+      "epoch": 11.906474820143885,
+      "step": 13240,
+      "torque_loss": 0.206381693482399
+    },
+    {
+      "epoch": 11.906474820143885,
+      "force_loss": 0.01907743513584137,
+      "step": 13240
+    },
+    {
+      "epoch": 11.91546762589928,
+      "grad_norm": 0.8608695268630981,
+      "learning_rate": 9.223263136172039e-05,
+      "loss": 0.0324,
+      "step": 13250
+    },
+    {
+      "action_loss": 0.014742610044777393,
+      "epoch": 11.91546762589928,
+      "step": 13250
+    },
+    {
+      "epoch": 11.91546762589928,
+      "step": 13250,
+      "torque_loss": 0.18341481685638428
+    },
+    {
+      "epoch": 11.91546762589928,
+      "force_loss": 0.015031460672616959,
+      "step": 13250
+    },
+    {
+      "epoch": 11.924460431654676,
+      "grad_norm": 0.5124015212059021,
+      "learning_rate": 9.22178728333672e-05,
+      "loss": 0.0269,
+      "step": 13260
+    },
+    {
+      "action_loss": 0.01137368381023407,
+      "epoch": 11.924460431654676,
+      "step": 13260
+    },
+    {
+      "epoch": 11.924460431654676,
+      "step": 13260,
+      "torque_loss": 0.11460211127996445
+    },
+    {
+      "epoch": 11.924460431654676,
+      "force_loss": 0.01490174699574709,
+      "step": 13260
+    },
+    {
+      "epoch": 11.933453237410072,
+      "grad_norm": 0.5288162231445312,
+      "learning_rate": 9.220310148033897e-05,
+      "loss": 0.0287,
+      "step": 13270
+    },
+    {
+      "action_loss": 0.0071401894092559814,
+      "epoch": 11.933453237410072,
+      "step": 13270
+    },
+    {
+      "epoch": 11.933453237410072,
+      "step": 13270,
+      "torque_loss": 0.10579943656921387
+    },
+    {
+      "epoch": 11.933453237410072,
+      "force_loss": 0.005128185730427504,
+      "step": 13270
+    },
+    {
+      "epoch": 11.942446043165468,
+      "grad_norm": 0.7475045323371887,
+      "learning_rate": 9.21883173071228e-05,
+      "loss": 0.0258,
+      "step": 13280
+    },
+    {
+      "action_loss": 0.005564589519053698,
+      "epoch": 11.942446043165468,
+      "step": 13280
+    },
+    {
+      "epoch": 11.942446043165468,
+      "step": 13280,
+      "torque_loss": 0.13392579555511475
+    },
+    {
+      "epoch": 11.942446043165468,
+      "force_loss": 0.005843350198119879,
+      "step": 13280
+    },
+    {
+      "epoch": 11.951438848920864,
+      "grad_norm": 0.3970196545124054,
+      "learning_rate": 9.217352031820976e-05,
+      "loss": 0.0216,
+      "step": 13290
+    },
+    {
+      "action_loss": 0.02049397863447666,
+      "epoch": 11.951438848920864,
+      "step": 13290
+    },
+    {
+      "epoch": 11.951438848920864,
+      "step": 13290,
+      "torque_loss": 0.12116687744855881
+    },
+    {
+      "epoch": 11.951438848920864,
+      "force_loss": 0.029718657955527306,
+      "step": 13290
+    },
+    {
+      "epoch": 11.96043165467626,
+      "grad_norm": 0.41319262981414795,
+      "learning_rate": 9.215871051809477e-05,
+      "loss": 0.0259,
+      "step": 13300
+    },
+    {
+      "action_loss": 0.004336673300713301,
+      "epoch": 11.96043165467626,
+      "step": 13300
+    },
+    {
+      "epoch": 11.96043165467626,
+      "step": 13300,
+      "torque_loss": 0.11368455737829208
+    },
+    {
+      "epoch": 11.96043165467626,
+      "force_loss": 0.004420553799718618,
+      "step": 13300
+    },
+    {
+      "epoch": 11.969424460431654,
+      "grad_norm": 0.7107324600219727,
+      "learning_rate": 9.214388791127666e-05,
+      "loss": 0.0257,
+      "step": 13310
+    },
+    {
+      "action_loss": 0.010378026403486729,
+      "epoch": 11.969424460431654,
+      "step": 13310
+    },
+    {
+      "epoch": 11.969424460431654,
+      "step": 13310,
+      "torque_loss": 0.15950961410999298
+    },
+    {
+      "epoch": 11.969424460431654,
+      "force_loss": 0.011539414525032043,
+      "step": 13310
+    },
+    {
+      "epoch": 11.97841726618705,
+      "grad_norm": 0.40262076258659363,
+      "learning_rate": 9.212905250225814e-05,
+      "loss": 0.0253,
+      "step": 13320
+    },
+    {
+      "action_loss": 0.0076032900251448154,
+      "epoch": 11.97841726618705,
+      "step": 13320
+    },
+    {
+      "epoch": 11.97841726618705,
+      "step": 13320,
+      "torque_loss": 0.18420910835266113
+    },
+    {
+      "epoch": 11.97841726618705,
+      "force_loss": 0.0048744529485702515,
+      "step": 13320
+    },
+    {
+      "epoch": 11.987410071942445,
+      "grad_norm": 0.660062313079834,
+      "learning_rate": 9.211420429554583e-05,
+      "loss": 0.0266,
+      "step": 13330
+    },
+    {
+      "action_loss": 0.004844684153795242,
+      "epoch": 11.987410071942445,
+      "step": 13330
+    },
+    {
+      "epoch": 11.987410071942445,
+      "step": 13330,
+      "torque_loss": 0.1788283735513687
+    },
+    {
+      "epoch": 11.987410071942445,
+      "force_loss": 0.006775469984859228,
+      "step": 13330
+    },
+    {
+      "epoch": 11.996402877697841,
+      "grad_norm": 1.07358717918396,
+      "learning_rate": 9.209934329565022e-05,
+      "loss": 0.0268,
+      "step": 13340
+    },
+    {
+      "action_loss": 0.0032872071024030447,
+      "epoch": 11.996402877697841,
+      "step": 13340
+    },
+    {
+      "epoch": 11.996402877697841,
+      "step": 13340,
+      "torque_loss": 0.133455291390419
+    },
+    {
+      "epoch": 11.996402877697841,
+      "force_loss": 0.004084920976310968,
+      "step": 13340
+    },
+    {
+      "epoch": 12.005395683453237,
+      "grad_norm": 0.6031844615936279,
+      "learning_rate": 9.208446950708568e-05,
+      "loss": 0.0223,
+      "step": 13350
+    },
+    {
+      "action_loss": 0.016608580946922302,
+      "epoch": 12.005395683453237,
+      "step": 13350
+    },
+    {
+      "epoch": 12.005395683453237,
+      "step": 13350,
+      "torque_loss": 0.19303618371486664
+    },
+    {
+      "epoch": 12.005395683453237,
+      "force_loss": 0.021011648699641228,
+      "step": 13350
+    },
+    {
+      "epoch": 12.014388489208633,
+      "grad_norm": 0.5078012943267822,
+      "learning_rate": 9.20695829343705e-05,
+      "loss": 0.0279,
+      "step": 13360
+    },
+    {
+      "action_loss": 0.004569401498883963,
+      "epoch": 12.014388489208633,
+      "step": 13360
+    },
+    {
+      "epoch": 12.014388489208633,
+      "step": 13360,
+      "torque_loss": 0.13210737705230713
+    },
+    {
+      "epoch": 12.014388489208633,
+      "force_loss": 0.0033373974729329348,
+      "step": 13360
+    },
+    {
+      "epoch": 12.023381294964029,
+      "grad_norm": 0.5666680335998535,
+      "learning_rate": 9.205468358202678e-05,
+      "loss": 0.0226,
+      "step": 13370
+    },
+    {
+      "action_loss": 0.008932110853493214,
+      "epoch": 12.023381294964029,
+      "step": 13370
+    },
+    {
+      "epoch": 12.023381294964029,
+      "step": 13370,
+      "torque_loss": 0.13694554567337036
+    },
+    {
+      "epoch": 12.023381294964029,
+      "force_loss": 0.00886202696710825,
+      "step": 13370
+    },
+    {
+      "epoch": 12.032374100719425,
+      "grad_norm": 0.5579889416694641,
+      "learning_rate": 9.203977145458059e-05,
+      "loss": 0.0286,
+      "step": 13380
+    },
+    {
+      "action_loss": 0.0067925299517810345,
+      "epoch": 12.032374100719425,
+      "step": 13380
+    },
+    {
+      "epoch": 12.032374100719425,
+      "step": 13380,
+      "torque_loss": 0.2041359692811966
+    },
+    {
+      "epoch": 12.032374100719425,
+      "force_loss": 0.008642517030239105,
+      "step": 13380
+    },
+    {
+      "epoch": 12.04136690647482,
+      "grad_norm": 0.5864238142967224,
+      "learning_rate": 9.202484655656182e-05,
+      "loss": 0.0259,
+      "step": 13390
+    },
+    {
+      "action_loss": 0.013879512436687946,
+      "epoch": 12.04136690647482,
+      "step": 13390
+    },
+    {
+      "epoch": 12.04136690647482,
+      "step": 13390,
+      "torque_loss": 0.16266094148159027
+    },
+    {
+      "epoch": 12.04136690647482,
+      "force_loss": 0.021635999903082848,
+      "step": 13390
+    },
+    {
+      "epoch": 12.050359712230216,
+      "grad_norm": 0.5106484889984131,
+      "learning_rate": 9.200990889250427e-05,
+      "loss": 0.0275,
+      "step": 13400
+    },
+    {
+      "action_loss": 0.005630366504192352,
+      "epoch": 12.050359712230216,
+      "step": 13400
+    },
+    {
+      "epoch": 12.050359712230216,
+      "step": 13400,
+      "torque_loss": 0.1035974845290184
+    },
+    {
+      "epoch": 12.050359712230216,
+      "force_loss": 0.008144255727529526,
+      "step": 13400
+    },
+    {
+      "epoch": 12.059352517985612,
+      "grad_norm": 0.3532458245754242,
+      "learning_rate": 9.19949584669456e-05,
+      "loss": 0.0226,
+      "step": 13410
+    },
+    {
+      "action_loss": 0.004766761790961027,
+      "epoch": 12.059352517985612,
+      "step": 13410
+    },
+    {
+      "epoch": 12.059352517985612,
+      "step": 13410,
+      "torque_loss": 0.152730792760849
+    },
+    {
+      "epoch": 12.059352517985612,
+      "force_loss": 0.0050053163431584835,
+      "step": 13410
+    },
+    {
+      "epoch": 12.068345323741006,
+      "grad_norm": 0.5839273929595947,
+      "learning_rate": 9.197999528442738e-05,
+      "loss": 0.0258,
+      "step": 13420
+    },
+    {
+      "action_loss": 0.006152821704745293,
+      "epoch": 12.068345323741006,
+      "step": 13420
+    },
+    {
+      "epoch": 12.068345323741006,
+      "step": 13420,
+      "torque_loss": 0.1193094328045845
+    },
+    {
+      "epoch": 12.068345323741006,
+      "force_loss": 0.005353698972612619,
+      "step": 13420
+    },
+    {
+      "epoch": 12.077338129496402,
+      "grad_norm": 0.278118759393692,
+      "learning_rate": 9.196501934949499e-05,
+      "loss": 0.0236,
+      "step": 13430
+    },
+    {
+      "action_loss": 0.0039503020234405994,
+      "epoch": 12.077338129496402,
+      "step": 13430
+    },
+    {
+      "epoch": 12.077338129496402,
+      "step": 13430,
+      "torque_loss": 0.0906924232840538
+    },
+    {
+      "epoch": 12.077338129496402,
+      "force_loss": 0.005831554066389799,
+      "step": 13430
+    },
+    {
+      "epoch": 12.086330935251798,
+      "grad_norm": 0.8984665870666504,
+      "learning_rate": 9.195003066669776e-05,
+      "loss": 0.0248,
+      "step": 13440
+    },
+    {
+      "action_loss": 0.017666278406977654,
+      "epoch": 12.086330935251798,
+      "step": 13440
+    },
+    {
+      "epoch": 12.086330935251798,
+      "step": 13440,
+      "torque_loss": 0.15089526772499084
+    },
+    {
+      "epoch": 12.086330935251798,
+      "force_loss": 0.01574471779167652,
+      "step": 13440
+    },
+    {
+      "epoch": 12.095323741007194,
+      "grad_norm": 0.6333910822868347,
+      "learning_rate": 9.193502924058884e-05,
+      "loss": 0.0237,
+      "step": 13450
+    },
+    {
+      "action_loss": 0.0071053714491426945,
+      "epoch": 12.095323741007194,
+      "step": 13450
+    },
+    {
+      "epoch": 12.095323741007194,
+      "step": 13450,
+      "torque_loss": 0.1242046132683754
+    },
+    {
+      "epoch": 12.095323741007194,
+      "force_loss": 0.008712325245141983,
+      "step": 13450
+    },
+    {
+      "epoch": 12.10431654676259,
+      "grad_norm": 0.6001847982406616,
+      "learning_rate": 9.192001507572526e-05,
+      "loss": 0.0267,
+      "step": 13460
+    },
+    {
+      "action_loss": 0.011919531039893627,
+      "epoch": 12.10431654676259,
+      "step": 13460
+    },
+    {
+      "epoch": 12.10431654676259,
+      "step": 13460,
+      "torque_loss": 0.1875322312116623
+    },
+    {
+      "epoch": 12.10431654676259,
+      "force_loss": 0.011915062554180622,
+      "step": 13460
+    },
+    {
+      "epoch": 12.113309352517986,
+      "grad_norm": 0.6838002800941467,
+      "learning_rate": 9.190498817666793e-05,
+      "loss": 0.0316,
+      "step": 13470
+    },
+    {
+      "action_loss": 0.00992722436785698,
+      "epoch": 12.113309352517986,
+      "step": 13470
+    },
+    {
+      "epoch": 12.113309352517986,
+      "step": 13470,
+      "torque_loss": 0.15418122708797455
+    },
+    {
+      "epoch": 12.113309352517986,
+      "force_loss": 0.012280096299946308,
+      "step": 13470
+    },
+    {
+      "epoch": 12.122302158273381,
+      "grad_norm": 0.5721536874771118,
+      "learning_rate": 9.188994854798163e-05,
+      "loss": 0.0245,
+      "step": 13480
+    },
+    {
+      "action_loss": 0.007268022745847702,
+      "epoch": 12.122302158273381,
+      "step": 13480
+    },
+    {
+      "epoch": 12.122302158273381,
+      "step": 13480,
+      "torque_loss": 0.11003028601408005
+    },
+    {
+      "epoch": 12.122302158273381,
+      "force_loss": 0.007329506333917379,
+      "step": 13480
+    },
+    {
+      "epoch": 12.131294964028777,
+      "grad_norm": 0.4797108769416809,
+      "learning_rate": 9.187489619423499e-05,
+      "loss": 0.0237,
+      "step": 13490
+    },
+    {
+      "action_loss": 0.004335633013397455,
+      "epoch": 12.131294964028777,
+      "step": 13490
+    },
+    {
+      "epoch": 12.131294964028777,
+      "step": 13490,
+      "torque_loss": 0.19956548511981964
+    },
+    {
+      "epoch": 12.131294964028777,
+      "force_loss": 0.0036118945572525263,
+      "step": 13490
+    },
+    {
+      "epoch": 12.140287769784173,
+      "grad_norm": 0.4975634813308716,
+      "learning_rate": 9.185983112000056e-05,
+      "loss": 0.0232,
+      "step": 13500
+    },
+    {
+      "action_loss": 0.006885308772325516,
+      "epoch": 12.140287769784173,
+      "step": 13500
+    },
+    {
+      "epoch": 12.140287769784173,
+      "step": 13500,
+      "torque_loss": 0.10912331193685532
+    },
+    {
+      "epoch": 12.140287769784173,
+      "force_loss": 0.0068674408830702305,
+      "step": 13500
+    },
+    {
+      "epoch": 12.149280575539569,
+      "grad_norm": 0.6214430928230286,
+      "learning_rate": 9.184475332985464e-05,
+      "loss": 0.0215,
+      "step": 13510
+    },
+    {
+      "action_loss": 0.004478698596358299,
+      "epoch": 12.149280575539569,
+      "step": 13510
+    },
+    {
+      "epoch": 12.149280575539569,
+      "step": 13510,
+      "torque_loss": 0.11821621656417847
+    },
+    {
+      "epoch": 12.149280575539569,
+      "force_loss": 0.004517280030995607,
+      "step": 13510
+    },
+    {
+      "epoch": 12.158273381294965,
+      "grad_norm": 0.22256101667881012,
+      "learning_rate": 9.182966282837754e-05,
+      "loss": 0.0248,
+      "step": 13520
+    },
+    {
+      "action_loss": 0.0061637721955776215,
+      "epoch": 12.158273381294965,
+      "step": 13520
+    },
+    {
+      "epoch": 12.158273381294965,
+      "step": 13520,
+      "torque_loss": 0.1224641278386116
+    },
+    {
+      "epoch": 12.158273381294965,
+      "force_loss": 0.006317870691418648,
+      "step": 13520
+    },
+    {
+      "epoch": 12.167266187050359,
+      "grad_norm": 0.23535418510437012,
+      "learning_rate": 9.18145596201533e-05,
+      "loss": 0.0206,
+      "step": 13530
+    },
+    {
+      "action_loss": 0.015792125836014748,
+      "epoch": 12.167266187050359,
+      "step": 13530
+    },
+    {
+      "epoch": 12.167266187050359,
+      "step": 13530,
+      "torque_loss": 0.12630845606327057
+    },
+    {
+      "epoch": 12.167266187050359,
+      "force_loss": 0.00916248094290495,
+      "step": 13530
+    },
+    {
+      "epoch": 12.176258992805755,
+      "grad_norm": 0.5914263725280762,
+      "learning_rate": 9.179944370976991e-05,
+      "loss": 0.0242,
+      "step": 13540
+    },
+    {
+      "action_loss": 0.010050378739833832,
+      "epoch": 12.176258992805755,
+      "step": 13540
+    },
+    {
+      "epoch": 12.176258992805755,
+      "step": 13540,
+      "torque_loss": 0.12712135910987854
+    },
+    {
+      "epoch": 12.176258992805755,
+      "force_loss": 0.006975242402404547,
+      "step": 13540
+    },
+    {
+      "epoch": 12.18525179856115,
+      "grad_norm": 0.5053604245185852,
+      "learning_rate": 9.178431510181918e-05,
+      "loss": 0.0208,
+      "step": 13550
+    },
+    {
+      "action_loss": 0.00847268383949995,
+      "epoch": 12.18525179856115,
+      "step": 13550
+    },
+    {
+      "epoch": 12.18525179856115,
+      "step": 13550,
+      "torque_loss": 0.06895231455564499
+    },
+    {
+      "epoch": 12.18525179856115,
+      "force_loss": 0.005697391927242279,
+      "step": 13550
+    },
+    {
+      "epoch": 12.194244604316546,
+      "grad_norm": 0.5936676859855652,
+      "learning_rate": 9.176917380089675e-05,
+      "loss": 0.0269,
+      "step": 13560
+    },
+    {
+      "action_loss": 0.007517187390476465,
+      "epoch": 12.194244604316546,
+      "step": 13560
+    },
+    {
+      "epoch": 12.194244604316546,
+      "step": 13560,
+      "torque_loss": 0.12213506549596786
+    },
+    {
+      "epoch": 12.194244604316546,
+      "force_loss": 0.0050654965452849865,
+      "step": 13560
+    },
+    {
+      "epoch": 12.203237410071942,
+      "grad_norm": 0.5953434109687805,
+      "learning_rate": 9.175401981160219e-05,
+      "loss": 0.0277,
+      "step": 13570
+    },
+    {
+      "action_loss": 0.006305044982582331,
+      "epoch": 12.203237410071942,
+      "step": 13570
+    },
+    {
+      "epoch": 12.203237410071942,
+      "step": 13570,
+      "torque_loss": 0.10977330058813095
+    },
+    {
+      "epoch": 12.203237410071942,
+      "force_loss": 0.005491024348884821,
+      "step": 13570
+    },
+    {
+      "epoch": 12.212230215827338,
+      "grad_norm": 0.40034228563308716,
+      "learning_rate": 9.173885313853885e-05,
+      "loss": 0.0241,
+      "step": 13580
+    },
+    {
+      "action_loss": 0.006344802677631378,
+      "epoch": 12.212230215827338,
+      "step": 13580
+    },
+    {
+      "epoch": 12.212230215827338,
+      "step": 13580,
+      "torque_loss": 0.149864062666893
+    },
+    {
+      "epoch": 12.212230215827338,
+      "force_loss": 0.005465306341648102,
+      "step": 13580
+    },
+    {
+      "epoch": 12.221223021582734,
+      "grad_norm": 0.60642409324646,
+      "learning_rate": 9.172367378631398e-05,
+      "loss": 0.0316,
+      "step": 13590
+    },
+    {
+      "action_loss": 0.018529774621129036,
+      "epoch": 12.221223021582734,
+      "step": 13590
+    },
+    {
+      "epoch": 12.221223021582734,
+      "step": 13590,
+      "torque_loss": 0.14816080033779144
+    },
+    {
+      "epoch": 12.221223021582734,
+      "force_loss": 0.014012637548148632,
+      "step": 13590
+    },
+    {
+      "epoch": 12.23021582733813,
+      "grad_norm": 0.6717132925987244,
+      "learning_rate": 9.170848175953866e-05,
+      "loss": 0.0335,
+      "step": 13600
+    },
+    {
+      "action_loss": 0.005674402695149183,
+      "epoch": 12.23021582733813,
+      "step": 13600
+    },
+    {
+      "epoch": 12.23021582733813,
+      "step": 13600,
+      "torque_loss": 0.10468759387731552
+    },
+    {
+      "epoch": 12.23021582733813,
+      "force_loss": 0.00724840210750699,
+      "step": 13600
+    },
+    {
+      "epoch": 12.239208633093526,
+      "grad_norm": 0.33241763710975647,
+      "learning_rate": 9.169327706282784e-05,
+      "loss": 0.0262,
+      "step": 13610
+    },
+    {
+      "action_loss": 0.0071988957934081554,
+      "epoch": 12.239208633093526,
+      "step": 13610
+    },
+    {
+      "epoch": 12.239208633093526,
+      "step": 13610,
+      "torque_loss": 0.11864650249481201
+    },
+    {
+      "epoch": 12.239208633093526,
+      "force_loss": 0.009233271703124046,
+      "step": 13610
+    },
+    {
+      "epoch": 12.248201438848922,
+      "grad_norm": 0.5763484239578247,
+      "learning_rate": 9.167805970080029e-05,
+      "loss": 0.0234,
+      "step": 13620
+    },
+    {
+      "action_loss": 0.01656474359333515,
+      "epoch": 12.248201438848922,
+      "step": 13620
+    },
+    {
+      "epoch": 12.248201438848922,
+      "step": 13620,
+      "torque_loss": 0.12728768587112427
+    },
+    {
+      "epoch": 12.248201438848922,
+      "force_loss": 0.020039258524775505,
+      "step": 13620
+    },
+    {
+      "epoch": 12.257194244604317,
+      "grad_norm": 0.38775160908699036,
+      "learning_rate": 9.166282967807864e-05,
+      "loss": 0.0251,
+      "step": 13630
+    },
+    {
+      "action_loss": 0.009864387102425098,
+      "epoch": 12.257194244604317,
+      "step": 13630
+    },
+    {
+      "epoch": 12.257194244604317,
+      "step": 13630,
+      "torque_loss": 0.09635195881128311
+    },
+    {
+      "epoch": 12.257194244604317,
+      "force_loss": 0.00766908610239625,
+      "step": 13630
+    },
+    {
+      "epoch": 12.266187050359711,
+      "grad_norm": 0.47758951783180237,
+      "learning_rate": 9.16475869992894e-05,
+      "loss": 0.0214,
+      "step": 13640
+    },
+    {
+      "action_loss": 0.009075481444597244,
+      "epoch": 12.266187050359711,
+      "step": 13640
+    },
+    {
+      "epoch": 12.266187050359711,
+      "step": 13640,
+      "torque_loss": 0.12635932862758636
+    },
+    {
+      "epoch": 12.266187050359711,
+      "force_loss": 0.006310591474175453,
+      "step": 13640
+    },
+    {
+      "epoch": 12.275179856115107,
+      "grad_norm": 0.5335314273834229,
+      "learning_rate": 9.163233166906284e-05,
+      "loss": 0.022,
+      "step": 13650
+    },
+    {
+      "action_loss": 0.00784197449684143,
+      "epoch": 12.275179856115107,
+      "step": 13650
+    },
+    {
+      "epoch": 12.275179856115107,
+      "step": 13650,
+      "torque_loss": 0.1303221732378006
+    },
+    {
+      "epoch": 12.275179856115107,
+      "force_loss": 0.005011549219489098,
+      "step": 13650
+    },
+    {
+      "epoch": 12.284172661870503,
+      "grad_norm": 0.5810086131095886,
+      "learning_rate": 9.161706369203317e-05,
+      "loss": 0.0253,
+      "step": 13660
+    },
+    {
+      "action_loss": 0.004765343386679888,
+      "epoch": 12.284172661870503,
+      "step": 13660
+    },
+    {
+      "epoch": 12.284172661870503,
+      "step": 13660,
+      "torque_loss": 0.0924394354224205
+    },
+    {
+      "epoch": 12.284172661870503,
+      "force_loss": 0.006362430285662413,
+      "step": 13660
+    },
+    {
+      "epoch": 12.293165467625899,
+      "grad_norm": 0.6272698640823364,
+      "learning_rate": 9.16017830728384e-05,
+      "loss": 0.022,
+      "step": 13670
+    },
+    {
+      "action_loss": 0.0043923030607402325,
+      "epoch": 12.293165467625899,
+      "step": 13670
+    },
+    {
+      "epoch": 12.293165467625899,
+      "step": 13670,
+      "torque_loss": 0.13075794279575348
+    },
+    {
+      "epoch": 12.293165467625899,
+      "force_loss": 0.003732785815373063,
+      "step": 13670
+    },
+    {
+      "epoch": 12.302158273381295,
+      "grad_norm": 0.7331401705741882,
+      "learning_rate": 9.158648981612035e-05,
+      "loss": 0.0252,
+      "step": 13680
+    },
+    {
+      "action_loss": 0.008111704140901566,
+      "epoch": 12.302158273381295,
+      "step": 13680
+    },
+    {
+      "epoch": 12.302158273381295,
+      "step": 13680,
+      "torque_loss": 0.12140810489654541
+    },
+    {
+      "epoch": 12.302158273381295,
+      "force_loss": 0.008820204995572567,
+      "step": 13680
+    },
+    {
+      "epoch": 12.31115107913669,
+      "grad_norm": 0.74886554479599,
+      "learning_rate": 9.157118392652472e-05,
+      "loss": 0.0241,
+      "step": 13690
+    },
+    {
+      "action_loss": 0.01673649065196514,
+      "epoch": 12.31115107913669,
+      "step": 13690
+    },
+    {
+      "epoch": 12.31115107913669,
+      "step": 13690,
+      "torque_loss": 0.14195752143859863
+    },
+    {
+      "epoch": 12.31115107913669,
+      "force_loss": 0.015183858573436737,
+      "step": 13690
+    },
+    {
+      "epoch": 12.320143884892087,
+      "grad_norm": 0.3879823088645935,
+      "learning_rate": 9.155586540870104e-05,
+      "loss": 0.0218,
+      "step": 13700
+    },
+    {
+      "action_loss": 0.009805773384869099,
+      "epoch": 12.320143884892087,
+      "step": 13700
+    },
+    {
+      "epoch": 12.320143884892087,
+      "step": 13700,
+      "torque_loss": 0.16158168017864227
+    },
+    {
+      "epoch": 12.320143884892087,
+      "force_loss": 0.007298277225345373,
+      "step": 13700
+    },
+    {
+      "epoch": 12.329136690647482,
+      "grad_norm": 0.4145404100418091,
+      "learning_rate": 9.154053426730267e-05,
+      "loss": 0.0254,
+      "step": 13710
+    },
+    {
+      "action_loss": 0.00787991564720869,
+      "epoch": 12.329136690647482,
+      "step": 13710
+    },
+    {
+      "epoch": 12.329136690647482,
+      "step": 13710,
+      "torque_loss": 0.12812145054340363
+    },
+    {
+      "epoch": 12.329136690647482,
+      "force_loss": 0.007166024763137102,
+      "step": 13710
+    },
+    {
+      "epoch": 12.338129496402878,
+      "grad_norm": 0.2867864966392517,
+      "learning_rate": 9.15251905069868e-05,
+      "loss": 0.0289,
+      "step": 13720
+    },
+    {
+      "action_loss": 0.007813703268766403,
+      "epoch": 12.338129496402878,
+      "step": 13720
+    },
+    {
+      "epoch": 12.338129496402878,
+      "step": 13720,
+      "torque_loss": 0.17746762931346893
+    },
+    {
+      "epoch": 12.338129496402878,
+      "force_loss": 0.011906628496944904,
+      "step": 13720
+    },
+    {
+      "epoch": 12.347122302158274,
+      "grad_norm": 0.33069443702697754,
+      "learning_rate": 9.150983413241446e-05,
+      "loss": 0.025,
+      "step": 13730
+    },
+    {
+      "action_loss": 0.011023983359336853,
+      "epoch": 12.347122302158274,
+      "step": 13730
+    },
+    {
+      "epoch": 12.347122302158274,
+      "step": 13730,
+      "torque_loss": 0.15857093036174774
+    },
+    {
+      "epoch": 12.347122302158274,
+      "force_loss": 0.009825992397964,
+      "step": 13730
+    },
+    {
+      "epoch": 12.35611510791367,
+      "grad_norm": 0.4829291105270386,
+      "learning_rate": 9.149446514825051e-05,
+      "loss": 0.0299,
+      "step": 13740
+    },
+    {
+      "action_loss": 0.005325306672602892,
+      "epoch": 12.35611510791367,
+      "step": 13740
+    },
+    {
+      "epoch": 12.35611510791367,
+      "step": 13740,
+      "torque_loss": 0.11463263630867004
+    },
+    {
+      "epoch": 12.35611510791367,
+      "force_loss": 0.005153453443199396,
+      "step": 13740
+    },
+    {
+      "epoch": 12.365107913669064,
+      "grad_norm": 0.6793624758720398,
+      "learning_rate": 9.147908355916365e-05,
+      "loss": 0.0278,
+      "step": 13750
+    },
+    {
+      "action_loss": 0.036259062588214874,
+      "epoch": 12.365107913669064,
+      "step": 13750
+    },
+    {
+      "epoch": 12.365107913669064,
+      "step": 13750,
+      "torque_loss": 0.20039959251880646
+    },
+    {
+      "epoch": 12.365107913669064,
+      "force_loss": 0.04130266234278679,
+      "step": 13750
+    },
+    {
+      "epoch": 12.37410071942446,
+      "grad_norm": 0.26231881976127625,
+      "learning_rate": 9.146368936982642e-05,
+      "loss": 0.0306,
+      "step": 13760
+    },
+    {
+      "action_loss": 0.013152661733329296,
+      "epoch": 12.37410071942446,
+      "step": 13760
+    },
+    {
+      "epoch": 12.37410071942446,
+      "step": 13760,
+      "torque_loss": 0.15253083407878876
+    },
+    {
+      "epoch": 12.37410071942446,
+      "force_loss": 0.013423360884189606,
+      "step": 13760
+    },
+    {
+      "epoch": 12.383093525179856,
+      "grad_norm": 0.3501366972923279,
+      "learning_rate": 9.144828258491511e-05,
+      "loss": 0.0277,
+      "step": 13770
+    },
+    {
+      "action_loss": 0.014741982333362103,
+      "epoch": 12.383093525179856,
+      "step": 13770
+    },
+    {
+      "epoch": 12.383093525179856,
+      "step": 13770,
+      "torque_loss": 0.14439605176448822
+    },
+    {
+      "epoch": 12.383093525179856,
+      "force_loss": 0.015146692283451557,
+      "step": 13770
+    },
+    {
+      "epoch": 12.392086330935252,
+      "grad_norm": 0.5469654202461243,
+      "learning_rate": 9.143286320910996e-05,
+      "loss": 0.0295,
+      "step": 13780
+    },
+    {
+      "action_loss": 0.011230617761611938,
+      "epoch": 12.392086330935252,
+      "step": 13780
+    },
+    {
+      "epoch": 12.392086330935252,
+      "step": 13780,
+      "torque_loss": 0.14415091276168823
+    },
+    {
+      "epoch": 12.392086330935252,
+      "force_loss": 0.007145017385482788,
+      "step": 13780
+    },
+    {
+      "epoch": 12.401079136690647,
+      "grad_norm": 0.5919626355171204,
+      "learning_rate": 9.141743124709491e-05,
+      "loss": 0.0258,
+      "step": 13790
+    },
+    {
+      "action_loss": 0.011638651601970196,
+      "epoch": 12.401079136690647,
+      "step": 13790
+    },
+    {
+      "epoch": 12.401079136690647,
+      "step": 13790,
+      "torque_loss": 0.17612363398075104
+    },
+    {
+      "epoch": 12.401079136690647,
+      "force_loss": 0.007354178931564093,
+      "step": 13790
+    },
+    {
+      "epoch": 12.410071942446043,
+      "grad_norm": 0.725797176361084,
+      "learning_rate": 9.140198670355784e-05,
+      "loss": 0.0252,
+      "step": 13800
+    },
+    {
+      "action_loss": 0.016116155311465263,
+      "epoch": 12.410071942446043,
+      "step": 13800
+    },
+    {
+      "epoch": 12.410071942446043,
+      "step": 13800,
+      "torque_loss": 0.1170927956700325
+    },
+    {
+      "epoch": 12.410071942446043,
+      "force_loss": 0.011015300638973713,
+      "step": 13800
+    },
+    {
+      "epoch": 12.41906474820144,
+      "grad_norm": 0.7040033936500549,
+      "learning_rate": 9.138652958319034e-05,
+      "loss": 0.0282,
+      "step": 13810
+    },
+    {
+      "action_loss": 0.0031529993284493685,
+      "epoch": 12.41906474820144,
+      "step": 13810
+    },
+    {
+      "epoch": 12.41906474820144,
+      "step": 13810,
+      "torque_loss": 0.1225736141204834
+    },
+    {
+      "epoch": 12.41906474820144,
+      "force_loss": 0.005469519644975662,
+      "step": 13810
+    },
+    {
+      "epoch": 12.428057553956835,
+      "grad_norm": 0.5092853307723999,
+      "learning_rate": 9.137105989068791e-05,
+      "loss": 0.0213,
+      "step": 13820
+    },
+    {
+      "action_loss": 0.01232543308287859,
+      "epoch": 12.428057553956835,
+      "step": 13820
+    },
+    {
+      "epoch": 12.428057553956835,
+      "step": 13820,
+      "torque_loss": 0.14313648641109467
+    },
+    {
+      "epoch": 12.428057553956835,
+      "force_loss": 0.02127975784242153,
+      "step": 13820
+    },
+    {
+      "epoch": 12.43705035971223,
+      "grad_norm": 0.4835895895957947,
+      "learning_rate": 9.135557763074983e-05,
+      "loss": 0.0327,
+      "step": 13830
+    },
+    {
+      "action_loss": 0.006219408009201288,
+      "epoch": 12.43705035971223,
+      "step": 13830
+    },
+    {
+      "epoch": 12.43705035971223,
+      "step": 13830,
+      "torque_loss": 0.1067761704325676
+    },
+    {
+      "epoch": 12.43705035971223,
+      "force_loss": 0.006293964106589556,
+      "step": 13830
+    },
+    {
+      "epoch": 12.446043165467627,
+      "grad_norm": 0.9791175127029419,
+      "learning_rate": 9.13400828080792e-05,
+      "loss": 0.0274,
+      "step": 13840
+    },
+    {
+      "action_loss": 0.0038767734076827765,
+      "epoch": 12.446043165467627,
+      "step": 13840
+    },
+    {
+      "epoch": 12.446043165467627,
+      "step": 13840,
+      "torque_loss": 0.09044162184000015
+    },
+    {
+      "epoch": 12.446043165467627,
+      "force_loss": 0.00449793878942728,
+      "step": 13840
+    },
+    {
+      "epoch": 12.45503597122302,
+      "grad_norm": 0.5134682655334473,
+      "learning_rate": 9.132457542738292e-05,
+      "loss": 0.0254,
+      "step": 13850
+    },
+    {
+      "action_loss": 0.005738839972764254,
+      "epoch": 12.45503597122302,
+      "step": 13850
+    },
+    {
+      "epoch": 12.45503597122302,
+      "step": 13850,
+      "torque_loss": 0.14481998980045319
+    },
+    {
+      "epoch": 12.45503597122302,
+      "force_loss": 0.005820055957883596,
+      "step": 13850
+    },
+    {
+      "epoch": 12.464028776978417,
+      "grad_norm": 0.29155710339546204,
+      "learning_rate": 9.130905549337174e-05,
+      "loss": 0.0216,
+      "step": 13860
+    },
+    {
+      "action_loss": 0.01345804426819086,
+      "epoch": 12.464028776978417,
+      "step": 13860
+    },
+    {
+      "epoch": 12.464028776978417,
+      "step": 13860,
+      "torque_loss": 0.14391465485095978
+    },
+    {
+      "epoch": 12.464028776978417,
+      "force_loss": 0.007132288068532944,
+      "step": 13860
+    },
+    {
+      "epoch": 12.473021582733812,
+      "grad_norm": 0.3183196783065796,
+      "learning_rate": 9.129352301076021e-05,
+      "loss": 0.0248,
+      "step": 13870
+    },
+    {
+      "action_loss": 0.009230182506144047,
+      "epoch": 12.473021582733812,
+      "step": 13870
+    },
+    {
+      "epoch": 12.473021582733812,
+      "step": 13870,
+      "torque_loss": 0.14582642912864685
+    },
+    {
+      "epoch": 12.473021582733812,
+      "force_loss": 0.008144509047269821,
+      "step": 13870
+    },
+    {
+      "epoch": 12.482014388489208,
+      "grad_norm": 0.6510902047157288,
+      "learning_rate": 9.127797798426668e-05,
+      "loss": 0.0247,
+      "step": 13880
+    },
+    {
+      "action_loss": 0.024639353156089783,
+      "epoch": 12.482014388489208,
+      "step": 13880
+    },
+    {
+      "epoch": 12.482014388489208,
+      "step": 13880,
+      "torque_loss": 0.16467545926570892
+    },
+    {
+      "epoch": 12.482014388489208,
+      "force_loss": 0.03665347024798393,
+      "step": 13880
+    },
+    {
+      "epoch": 12.491007194244604,
+      "grad_norm": 0.42777055501937866,
+      "learning_rate": 9.126242041861333e-05,
+      "loss": 0.0293,
+      "step": 13890
+    },
+    {
+      "action_loss": 0.01301325112581253,
+      "epoch": 12.491007194244604,
+      "step": 13890
+    },
+    {
+      "epoch": 12.491007194244604,
+      "step": 13890,
+      "torque_loss": 0.14655320346355438
+    },
+    {
+      "epoch": 12.491007194244604,
+      "force_loss": 0.008423476479947567,
+      "step": 13890
+    },
+    {
+      "epoch": 12.5,
+      "grad_norm": 0.42942848801612854,
+      "learning_rate": 9.124685031852611e-05,
+      "loss": 0.0279,
+      "step": 13900
+    },
+    {
+      "action_loss": 0.0063703334890306,
+      "epoch": 12.5,
+      "step": 13900
+    },
+    {
+      "epoch": 12.5,
+      "step": 13900,
+      "torque_loss": 0.1505795270204544
+    },
+    {
+      "epoch": 12.5,
+      "force_loss": 0.007232634350657463,
+      "step": 13900
+    },
+    {
+      "epoch": 12.508992805755396,
+      "grad_norm": 0.6118781566619873,
+      "learning_rate": 9.123126768873482e-05,
+      "loss": 0.0264,
+      "step": 13910
+    },
+    {
+      "action_loss": 0.009435432031750679,
+      "epoch": 12.508992805755396,
+      "step": 13910
+    },
+    {
+      "epoch": 12.508992805755396,
+      "step": 13910,
+      "torque_loss": 0.12627960741519928
+    },
+    {
+      "epoch": 12.508992805755396,
+      "force_loss": 0.00904346164315939,
+      "step": 13910
+    },
+    {
+      "epoch": 12.517985611510792,
+      "grad_norm": 0.5503523945808411,
+      "learning_rate": 9.121567253397308e-05,
+      "loss": 0.0279,
+      "step": 13920
+    },
+    {
+      "action_loss": 0.004087453242391348,
+      "epoch": 12.517985611510792,
+      "step": 13920
+    },
+    {
+      "epoch": 12.517985611510792,
+      "step": 13920,
+      "torque_loss": 0.10465057939291
+    },
+    {
+      "epoch": 12.517985611510792,
+      "force_loss": 0.007370705250650644,
+      "step": 13920
+    },
+    {
+      "epoch": 12.526978417266188,
+      "grad_norm": 0.5884429812431335,
+      "learning_rate": 9.120006485897824e-05,
+      "loss": 0.0223,
+      "step": 13930
+    },
+    {
+      "action_loss": 0.004481791984289885,
+      "epoch": 12.526978417266188,
+      "step": 13930
+    },
+    {
+      "epoch": 12.526978417266188,
+      "step": 13930,
+      "torque_loss": 0.10764376074075699
+    },
+    {
+      "epoch": 12.526978417266188,
+      "force_loss": 0.00729639595374465,
+      "step": 13930
+    },
+    {
+      "epoch": 12.535971223021583,
+      "grad_norm": 0.5687770247459412,
+      "learning_rate": 9.118444466849152e-05,
+      "loss": 0.0309,
+      "step": 13940
+    },
+    {
+      "action_loss": 0.01760326884686947,
+      "epoch": 12.535971223021583,
+      "step": 13940
+    },
+    {
+      "epoch": 12.535971223021583,
+      "step": 13940,
+      "torque_loss": 0.12070111185312271
+    },
+    {
+      "epoch": 12.535971223021583,
+      "force_loss": 0.021414687857031822,
+      "step": 13940
+    },
+    {
+      "epoch": 12.54496402877698,
+      "grad_norm": 0.40998855233192444,
+      "learning_rate": 9.116881196725793e-05,
+      "loss": 0.0249,
+      "step": 13950
+    },
+    {
+      "action_loss": 0.005727668758481741,
+      "epoch": 12.54496402877698,
+      "step": 13950
+    },
+    {
+      "epoch": 12.54496402877698,
+      "step": 13950,
+      "torque_loss": 0.10952623933553696
+    },
+    {
+      "epoch": 12.54496402877698,
+      "force_loss": 0.005893405992537737,
+      "step": 13950
+    },
+    {
+      "epoch": 12.553956834532373,
+      "grad_norm": 0.41071924567222595,
+      "learning_rate": 9.115316676002627e-05,
+      "loss": 0.0251,
+      "step": 13960
+    },
+    {
+      "action_loss": 0.015424457378685474,
+      "epoch": 12.553956834532373,
+      "step": 13960
+    },
+    {
+      "epoch": 12.553956834532373,
+      "step": 13960,
+      "torque_loss": 0.15265995264053345
+    },
+    {
+      "epoch": 12.553956834532373,
+      "force_loss": 0.012465615756809711,
+      "step": 13960
+    },
+    {
+      "epoch": 12.56294964028777,
+      "grad_norm": 0.40640953183174133,
+      "learning_rate": 9.113750905154911e-05,
+      "loss": 0.0262,
+      "step": 13970
+    },
+    {
+      "action_loss": 0.007475083228200674,
+      "epoch": 12.56294964028777,
+      "step": 13970
+    },
+    {
+      "epoch": 12.56294964028777,
+      "step": 13970,
+      "torque_loss": 0.14392529428005219
+    },
+    {
+      "epoch": 12.56294964028777,
+      "force_loss": 0.013792566955089569,
+      "step": 13970
+    },
+    {
+      "epoch": 12.571942446043165,
+      "grad_norm": 0.49539658427238464,
+      "learning_rate": 9.112183884658289e-05,
+      "loss": 0.0299,
+      "step": 13980
+    },
+    {
+      "action_loss": 0.007323987782001495,
+      "epoch": 12.571942446043165,
+      "step": 13980
+    },
+    {
+      "epoch": 12.571942446043165,
+      "step": 13980,
+      "torque_loss": 0.1938466578722
+    },
+    {
+      "epoch": 12.571942446043165,
+      "force_loss": 0.008375290781259537,
+      "step": 13980
+    },
+    {
+      "epoch": 12.58093525179856,
+      "grad_norm": 0.466087281703949,
+      "learning_rate": 9.11061561498878e-05,
+      "loss": 0.021,
+      "step": 13990
+    },
+    {
+      "action_loss": 0.02233838103711605,
+      "epoch": 12.58093525179856,
+      "step": 13990
+    },
+    {
+      "epoch": 12.58093525179856,
+      "step": 13990,
+      "torque_loss": 0.15092484652996063
+    },
+    {
+      "epoch": 12.58093525179856,
+      "force_loss": 0.018663665279746056,
+      "step": 13990
+    },
+    {
+      "epoch": 12.589928057553957,
+      "grad_norm": 0.7940137386322021,
+      "learning_rate": 9.109046096622779e-05,
+      "loss": 0.0282,
+      "step": 14000
+    },
+    {
+      "action_loss": 0.014001191593706608,
+      "epoch": 12.589928057553957,
+      "step": 14000
+    },
+    {
+      "epoch": 12.589928057553957,
+      "step": 14000,
+      "torque_loss": 0.1394633799791336
+    },
+    {
+      "epoch": 12.589928057553957,
+      "force_loss": 0.013311732560396194,
+      "step": 14000
+    },
+    {
+      "epoch": 12.598920863309353,
+      "grad_norm": 0.6612755656242371,
+      "learning_rate": 9.107475330037069e-05,
+      "loss": 0.0263,
+      "step": 14010
+    },
+    {
+      "action_loss": 0.01692165620625019,
+      "epoch": 12.598920863309353,
+      "step": 14010
+    },
+    {
+      "epoch": 12.598920863309353,
+      "step": 14010,
+      "torque_loss": 0.1343950480222702
+    },
+    {
+      "epoch": 12.598920863309353,
+      "force_loss": 0.010985836386680603,
+      "step": 14010
+    },
+    {
+      "epoch": 12.607913669064748,
+      "grad_norm": 0.38929954171180725,
+      "learning_rate": 9.105903315708806e-05,
+      "loss": 0.0267,
+      "step": 14020
+    },
+    {
+      "action_loss": 0.006525618489831686,
+      "epoch": 12.607913669064748,
+      "step": 14020
+    },
+    {
+      "epoch": 12.607913669064748,
+      "step": 14020,
+      "torque_loss": 0.1254248023033142
+    },
+    {
+      "epoch": 12.607913669064748,
+      "force_loss": 0.006911094766110182,
+      "step": 14020
+    },
+    {
+      "epoch": 12.616906474820144,
+      "grad_norm": 0.3836361765861511,
+      "learning_rate": 9.104330054115524e-05,
+      "loss": 0.0238,
+      "step": 14030
+    },
+    {
+      "action_loss": 0.006212487816810608,
+      "epoch": 12.616906474820144,
+      "step": 14030
+    },
+    {
+      "epoch": 12.616906474820144,
+      "step": 14030,
+      "torque_loss": 0.10357270389795303
+    },
+    {
+      "epoch": 12.616906474820144,
+      "force_loss": 0.005381166469305754,
+      "step": 14030
+    },
+    {
+      "epoch": 12.62589928057554,
+      "grad_norm": 0.547963559627533,
+      "learning_rate": 9.102755545735141e-05,
+      "loss": 0.0218,
+      "step": 14040
+    },
+    {
+      "action_loss": 0.006468549370765686,
+      "epoch": 12.62589928057554,
+      "step": 14040
+    },
+    {
+      "epoch": 12.62589928057554,
+      "step": 14040,
+      "torque_loss": 0.10992669314146042
+    },
+    {
+      "epoch": 12.62589928057554,
+      "force_loss": 0.0067933350801467896,
+      "step": 14040
+    },
+    {
+      "epoch": 12.634892086330936,
+      "grad_norm": 0.3229418992996216,
+      "learning_rate": 9.10117979104595e-05,
+      "loss": 0.0259,
+      "step": 14050
+    },
+    {
+      "action_loss": 0.011358804069459438,
+      "epoch": 12.634892086330936,
+      "step": 14050
+    },
+    {
+      "epoch": 12.634892086330936,
+      "step": 14050,
+      "torque_loss": 0.1362975686788559
+    },
+    {
+      "epoch": 12.634892086330936,
+      "force_loss": 0.013899271376430988,
+      "step": 14050
+    },
+    {
+      "epoch": 12.64388489208633,
+      "grad_norm": 0.4975723624229431,
+      "learning_rate": 9.099602790526624e-05,
+      "loss": 0.0274,
+      "step": 14060
+    },
+    {
+      "action_loss": 0.007108205463737249,
+      "epoch": 12.64388489208633,
+      "step": 14060
+    },
+    {
+      "epoch": 12.64388489208633,
+      "step": 14060,
+      "torque_loss": 0.13089169561862946
+    },
+    {
+      "epoch": 12.64388489208633,
+      "force_loss": 0.006654033903032541,
+      "step": 14060
+    },
+    {
+      "epoch": 12.652877697841726,
+      "grad_norm": 0.5616057515144348,
+      "learning_rate": 9.098024544656212e-05,
+      "loss": 0.0219,
+      "step": 14070
+    },
+    {
+      "action_loss": 0.005977492779493332,
+      "epoch": 12.652877697841726,
+      "step": 14070
+    },
+    {
+      "epoch": 12.652877697841726,
+      "step": 14070,
+      "torque_loss": 0.13386978209018707
+    },
+    {
+      "epoch": 12.652877697841726,
+      "force_loss": 0.009287810884416103,
+      "step": 14070
+    },
+    {
+      "epoch": 12.661870503597122,
+      "grad_norm": 0.6663073301315308,
+      "learning_rate": 9.096445053914148e-05,
+      "loss": 0.0336,
+      "step": 14080
+    },
+    {
+      "action_loss": 0.00646496145054698,
+      "epoch": 12.661870503597122,
+      "step": 14080
+    },
+    {
+      "epoch": 12.661870503597122,
+      "step": 14080,
+      "torque_loss": 0.12621334195137024
+    },
+    {
+      "epoch": 12.661870503597122,
+      "force_loss": 0.006149812135845423,
+      "step": 14080
+    },
+    {
+      "epoch": 12.670863309352518,
+      "grad_norm": 0.2959502637386322,
+      "learning_rate": 9.094864318780236e-05,
+      "loss": 0.0275,
+      "step": 14090
+    },
+    {
+      "action_loss": 0.04846276715397835,
+      "epoch": 12.670863309352518,
+      "step": 14090
+    },
+    {
+      "epoch": 12.670863309352518,
+      "step": 14090,
+      "torque_loss": 0.2021758109331131
+    },
+    {
+      "epoch": 12.670863309352518,
+      "force_loss": 0.0776110514998436,
+      "step": 14090
+    },
+    {
+      "epoch": 12.679856115107913,
+      "grad_norm": 0.4802953898906708,
+      "learning_rate": 9.093282339734663e-05,
+      "loss": 0.0366,
+      "step": 14100
+    },
+    {
+      "action_loss": 0.007498738821595907,
+      "epoch": 12.679856115107913,
+      "step": 14100
+    },
+    {
+      "epoch": 12.679856115107913,
+      "step": 14100,
+      "torque_loss": 0.12763415277004242
+    },
+    {
+      "epoch": 12.679856115107913,
+      "force_loss": 0.009072769433259964,
+      "step": 14100
+    },
+    {
+      "epoch": 12.68884892086331,
+      "grad_norm": 0.30895718932151794,
+      "learning_rate": 9.091699117257992e-05,
+      "loss": 0.0257,
+      "step": 14110
+    },
+    {
+      "action_loss": 0.005022708792239428,
+      "epoch": 12.68884892086331,
+      "step": 14110
+    },
+    {
+      "epoch": 12.68884892086331,
+      "step": 14110,
+      "torque_loss": 0.15535520017147064
+    },
+    {
+      "epoch": 12.68884892086331,
+      "force_loss": 0.005461633671075106,
+      "step": 14110
+    },
+    {
+      "epoch": 12.697841726618705,
+      "grad_norm": 0.46144697070121765,
+      "learning_rate": 9.090114651831163e-05,
+      "loss": 0.0247,
+      "step": 14120
+    },
+    {
+      "action_loss": 0.017287636175751686,
+      "epoch": 12.697841726618705,
+      "step": 14120
+    },
+    {
+      "epoch": 12.697841726618705,
+      "step": 14120,
+      "torque_loss": 0.1857975721359253
+    },
+    {
+      "epoch": 12.697841726618705,
+      "force_loss": 0.016247255727648735,
+      "step": 14120
+    },
+    {
+      "epoch": 12.706834532374101,
+      "grad_norm": 0.6360578536987305,
+      "learning_rate": 9.088528943935497e-05,
+      "loss": 0.0292,
+      "step": 14130
+    },
+    {
+      "action_loss": 0.016388682648539543,
+      "epoch": 12.706834532374101,
+      "step": 14130
+    },
+    {
+      "epoch": 12.706834532374101,
+      "step": 14130,
+      "torque_loss": 0.12045708298683167
+    },
+    {
+      "epoch": 12.706834532374101,
+      "force_loss": 0.012151663191616535,
+      "step": 14130
+    },
+    {
+      "epoch": 12.715827338129497,
+      "grad_norm": 0.7302191853523254,
+      "learning_rate": 9.086941994052689e-05,
+      "loss": 0.025,
+      "step": 14140
+    },
+    {
+      "action_loss": 0.008230630308389664,
+      "epoch": 12.715827338129497,
+      "step": 14140
+    },
+    {
+      "epoch": 12.715827338129497,
+      "step": 14140,
+      "torque_loss": 0.13785457611083984
+    },
+    {
+      "epoch": 12.715827338129497,
+      "force_loss": 0.007931435480713844,
+      "step": 14140
+    },
+    {
+      "epoch": 12.724820143884893,
+      "grad_norm": 0.36570969223976135,
+      "learning_rate": 9.085353802664813e-05,
+      "loss": 0.0233,
+      "step": 14150
+    },
+    {
+      "action_loss": 0.008267399854958057,
+      "epoch": 12.724820143884893,
+      "step": 14150
+    },
+    {
+      "epoch": 12.724820143884893,
+      "step": 14150,
+      "torque_loss": 0.1470588594675064
+    },
+    {
+      "epoch": 12.724820143884893,
+      "force_loss": 0.010002859868109226,
+      "step": 14150
+    },
+    {
+      "epoch": 12.733812949640289,
+      "grad_norm": 0.4242355525493622,
+      "learning_rate": 9.08376437025432e-05,
+      "loss": 0.0254,
+      "step": 14160
+    },
+    {
+      "action_loss": 0.006476314272731543,
+      "epoch": 12.733812949640289,
+      "step": 14160
+    },
+    {
+      "epoch": 12.733812949640289,
+      "step": 14160,
+      "torque_loss": 0.12720048427581787
+    },
+    {
+      "epoch": 12.733812949640289,
+      "force_loss": 0.011430182494223118,
+      "step": 14160
+    },
+    {
+      "epoch": 12.742805755395683,
+      "grad_norm": 0.4574201703071594,
+      "learning_rate": 9.082173697304035e-05,
+      "loss": 0.0263,
+      "step": 14170
+    },
+    {
+      "action_loss": 0.015163346193730831,
+      "epoch": 12.742805755395683,
+      "step": 14170
+    },
+    {
+      "epoch": 12.742805755395683,
+      "step": 14170,
+      "torque_loss": 0.15634046494960785
+    },
+    {
+      "epoch": 12.742805755395683,
+      "force_loss": 0.015708336606621742,
+      "step": 14170
+    },
+    {
+      "epoch": 12.751798561151078,
+      "grad_norm": 0.22592872381210327,
+      "learning_rate": 9.080581784297166e-05,
+      "loss": 0.0212,
+      "step": 14180
+    },
+    {
+      "action_loss": 0.016290761530399323,
+      "epoch": 12.751798561151078,
+      "step": 14180
+    },
+    {
+      "epoch": 12.751798561151078,
+      "step": 14180,
+      "torque_loss": 0.13783107697963715
+    },
+    {
+      "epoch": 12.751798561151078,
+      "force_loss": 0.0128251938149333,
+      "step": 14180
+    },
+    {
+      "epoch": 12.760791366906474,
+      "grad_norm": 0.6080386638641357,
+      "learning_rate": 9.078988631717291e-05,
+      "loss": 0.0247,
+      "step": 14190
+    },
+    {
+      "action_loss": 0.010328296571969986,
+      "epoch": 12.760791366906474,
+      "step": 14190
+    },
+    {
+      "epoch": 12.760791366906474,
+      "step": 14190,
+      "torque_loss": 0.12926451861858368
+    },
+    {
+      "epoch": 12.760791366906474,
+      "force_loss": 0.013141573406755924,
+      "step": 14190
+    },
+    {
+      "epoch": 12.76978417266187,
+      "grad_norm": 0.6275120377540588,
+      "learning_rate": 9.077394240048369e-05,
+      "loss": 0.0245,
+      "step": 14200
+    },
+    {
+      "action_loss": 0.004028057213872671,
+      "epoch": 12.76978417266187,
+      "step": 14200
+    },
+    {
+      "epoch": 12.76978417266187,
+      "step": 14200,
+      "torque_loss": 0.147388756275177
+    },
+    {
+      "epoch": 12.76978417266187,
+      "force_loss": 0.005528328474611044,
+      "step": 14200
+    },
+    {
+      "epoch": 12.778776978417266,
+      "grad_norm": 0.38772672414779663,
+      "learning_rate": 9.075798609774736e-05,
+      "loss": 0.0286,
+      "step": 14210
+    },
+    {
+      "action_loss": 0.004324526991695166,
+      "epoch": 12.778776978417266,
+      "step": 14210
+    },
+    {
+      "epoch": 12.778776978417266,
+      "step": 14210,
+      "torque_loss": 0.13769835233688354
+    },
+    {
+      "epoch": 12.778776978417266,
+      "force_loss": 0.0045209601521492004,
+      "step": 14210
+    },
+    {
+      "epoch": 12.787769784172662,
+      "grad_norm": 0.5099405646324158,
+      "learning_rate": 9.0742017413811e-05,
+      "loss": 0.0232,
+      "step": 14220
+    },
+    {
+      "action_loss": 0.008415845222771168,
+      "epoch": 12.787769784172662,
+      "step": 14220
+    },
+    {
+      "epoch": 12.787769784172662,
+      "step": 14220,
+      "torque_loss": 0.14204971492290497
+    },
+    {
+      "epoch": 12.787769784172662,
+      "force_loss": 0.010749172419309616,
+      "step": 14220
+    },
+    {
+      "epoch": 12.796762589928058,
+      "grad_norm": 0.6851628422737122,
+      "learning_rate": 9.072603635352548e-05,
+      "loss": 0.0258,
+      "step": 14230
+    },
+    {
+      "action_loss": 0.005719442386180162,
+      "epoch": 12.796762589928058,
+      "step": 14230
+    },
+    {
+      "epoch": 12.796762589928058,
+      "step": 14230,
+      "torque_loss": 0.15508398413658142
+    },
+    {
+      "epoch": 12.796762589928058,
+      "force_loss": 0.006898975465446711,
+      "step": 14230
+    },
+    {
+      "epoch": 12.805755395683454,
+      "grad_norm": 0.4878329932689667,
+      "learning_rate": 9.071004292174541e-05,
+      "loss": 0.0282,
+      "step": 14240
+    },
+    {
+      "action_loss": 0.0068457648158073425,
+      "epoch": 12.805755395683454,
+      "step": 14240
+    },
+    {
+      "epoch": 12.805755395683454,
+      "step": 14240,
+      "torque_loss": 0.08456694334745407
+    },
+    {
+      "epoch": 12.805755395683454,
+      "force_loss": 0.008580955676734447,
+      "step": 14240
+    },
+    {
+      "epoch": 12.81474820143885,
+      "grad_norm": 0.9357104301452637,
+      "learning_rate": 9.06940371233292e-05,
+      "loss": 0.024,
+      "step": 14250
+    },
+    {
+      "action_loss": 0.0034057635348290205,
+      "epoch": 12.81474820143885,
+      "step": 14250
+    },
+    {
+      "epoch": 12.81474820143885,
+      "step": 14250,
+      "torque_loss": 0.15382985770702362
+    },
+    {
+      "epoch": 12.81474820143885,
+      "force_loss": 0.003341132774949074,
+      "step": 14250
+    },
+    {
+      "epoch": 12.823741007194245,
+      "grad_norm": 0.34692761301994324,
+      "learning_rate": 9.067801896313898e-05,
+      "loss": 0.0279,
+      "step": 14260
+    },
+    {
+      "action_loss": 0.005263553466647863,
+      "epoch": 12.823741007194245,
+      "step": 14260
+    },
+    {
+      "epoch": 12.823741007194245,
+      "step": 14260,
+      "torque_loss": 0.10076197981834412
+    },
+    {
+      "epoch": 12.823741007194245,
+      "force_loss": 0.008891968987882137,
+      "step": 14260
+    },
+    {
+      "epoch": 12.832733812949641,
+      "grad_norm": 0.3128192126750946,
+      "learning_rate": 9.066198844604064e-05,
+      "loss": 0.0265,
+      "step": 14270
+    },
+    {
+      "action_loss": 0.007821241393685341,
+      "epoch": 12.832733812949641,
+      "step": 14270
+    },
+    {
+      "epoch": 12.832733812949641,
+      "step": 14270,
+      "torque_loss": 0.12226840853691101
+    },
+    {
+      "epoch": 12.832733812949641,
+      "force_loss": 0.005744075868278742,
+      "step": 14270
+    },
+    {
+      "epoch": 12.841726618705035,
+      "grad_norm": 0.42730021476745605,
+      "learning_rate": 9.06459455769038e-05,
+      "loss": 0.033,
+      "step": 14280
+    },
+    {
+      "action_loss": 0.0123290391638875,
+      "epoch": 12.841726618705035,
+      "step": 14280
+    },
+    {
+      "epoch": 12.841726618705035,
+      "step": 14280,
+      "torque_loss": 0.2124076634645462
+    },
+    {
+      "epoch": 12.841726618705035,
+      "force_loss": 0.012979678809642792,
+      "step": 14280
+    },
+    {
+      "epoch": 12.850719424460431,
+      "grad_norm": 0.7423769235610962,
+      "learning_rate": 9.062989036060193e-05,
+      "loss": 0.0338,
+      "step": 14290
+    },
+    {
+      "action_loss": 0.013494514860212803,
+      "epoch": 12.850719424460431,
+      "step": 14290
+    },
+    {
+      "epoch": 12.850719424460431,
+      "step": 14290,
+      "torque_loss": 0.14486967027187347
+    },
+    {
+      "epoch": 12.850719424460431,
+      "force_loss": 0.008547775447368622,
+      "step": 14290
+    },
+    {
+      "epoch": 12.859712230215827,
+      "grad_norm": 0.26264509558677673,
+      "learning_rate": 9.061382280201212e-05,
+      "loss": 0.023,
+      "step": 14300
+    },
+    {
+      "action_loss": 0.03842094540596008,
+      "epoch": 12.859712230215827,
+      "step": 14300
+    },
+    {
+      "epoch": 12.859712230215827,
+      "step": 14300,
+      "torque_loss": 0.19657373428344727
+    },
+    {
+      "epoch": 12.859712230215827,
+      "force_loss": 0.050675153732299805,
+      "step": 14300
+    },
+    {
+      "epoch": 12.868705035971223,
+      "grad_norm": 0.5091500878334045,
+      "learning_rate": 9.059774290601528e-05,
+      "loss": 0.0309,
+      "step": 14310
+    },
+    {
+      "action_loss": 0.005208694841712713,
+      "epoch": 12.868705035971223,
+      "step": 14310
+    },
+    {
+      "epoch": 12.868705035971223,
+      "step": 14310,
+      "torque_loss": 0.13491962850093842
+    },
+    {
+      "epoch": 12.868705035971223,
+      "force_loss": 0.006779221352189779,
+      "step": 14310
+    },
+    {
+      "epoch": 12.877697841726619,
+      "grad_norm": 0.3828144967556,
+      "learning_rate": 9.058165067749606e-05,
+      "loss": 0.0247,
+      "step": 14320
+    },
+    {
+      "action_loss": 0.012771953828632832,
+      "epoch": 12.877697841726619,
+      "step": 14320
+    },
+    {
+      "epoch": 12.877697841726619,
+      "step": 14320,
+      "torque_loss": 0.16346320509910583
+    },
+    {
+      "epoch": 12.877697841726619,
+      "force_loss": 0.011284224689006805,
+      "step": 14320
+    },
+    {
+      "epoch": 12.886690647482014,
+      "grad_norm": 0.5729163885116577,
+      "learning_rate": 9.056554612134288e-05,
+      "loss": 0.029,
+      "step": 14330
+    },
+    {
+      "action_loss": 0.009330385364592075,
+      "epoch": 12.886690647482014,
+      "step": 14330
+    },
+    {
+      "epoch": 12.886690647482014,
+      "step": 14330,
+      "torque_loss": 0.1821797639131546
+    },
+    {
+      "epoch": 12.886690647482014,
+      "force_loss": 0.006371926516294479,
+      "step": 14330
+    },
+    {
+      "epoch": 12.89568345323741,
+      "grad_norm": 0.4588429927825928,
+      "learning_rate": 9.054942924244785e-05,
+      "loss": 0.0264,
+      "step": 14340
+    },
+    {
+      "action_loss": 0.020296970382332802,
+      "epoch": 12.89568345323741,
+      "step": 14340
+    },
+    {
+      "epoch": 12.89568345323741,
+      "step": 14340,
+      "torque_loss": 0.14829859137535095
+    },
+    {
+      "epoch": 12.89568345323741,
+      "force_loss": 0.025701550766825676,
+      "step": 14340
+    },
+    {
+      "epoch": 12.904676258992806,
+      "grad_norm": 0.3945634067058563,
+      "learning_rate": 9.053330004570686e-05,
+      "loss": 0.0273,
+      "step": 14350
+    },
+    {
+      "action_loss": 0.006152736488729715,
+      "epoch": 12.904676258992806,
+      "step": 14350
+    },
+    {
+      "epoch": 12.904676258992806,
+      "step": 14350,
+      "torque_loss": 0.18100863695144653
+    },
+    {
+      "epoch": 12.904676258992806,
+      "force_loss": 0.007932188920676708,
+      "step": 14350
+    },
+    {
+      "epoch": 12.913669064748202,
+      "grad_norm": 0.5920173525810242,
+      "learning_rate": 9.051715853601955e-05,
+      "loss": 0.0251,
+      "step": 14360
+    },
+    {
+      "action_loss": 0.013649355620145798,
+      "epoch": 12.913669064748202,
+      "step": 14360
+    },
+    {
+      "epoch": 12.913669064748202,
+      "step": 14360,
+      "torque_loss": 0.14411768317222595
+    },
+    {
+      "epoch": 12.913669064748202,
+      "force_loss": 0.010054810903966427,
+      "step": 14360
+    },
+    {
+      "epoch": 12.922661870503598,
+      "grad_norm": 0.36797231435775757,
+      "learning_rate": 9.050100471828926e-05,
+      "loss": 0.0306,
+      "step": 14370
+    },
+    {
+      "action_loss": 0.009905297309160233,
+      "epoch": 12.922661870503598,
+      "step": 14370
+    },
+    {
+      "epoch": 12.922661870503598,
+      "step": 14370,
+      "torque_loss": 0.15208782255649567
+    },
+    {
+      "epoch": 12.922661870503598,
+      "force_loss": 0.015537414699792862,
+      "step": 14370
+    },
+    {
+      "epoch": 12.931654676258994,
+      "grad_norm": 0.33273354172706604,
+      "learning_rate": 9.048483859742311e-05,
+      "loss": 0.0293,
+      "step": 14380
+    },
+    {
+      "action_loss": 0.009125870652496815,
+      "epoch": 12.931654676258994,
+      "step": 14380
+    },
+    {
+      "epoch": 12.931654676258994,
+      "step": 14380,
+      "torque_loss": 0.1439599245786667
+    },
+    {
+      "epoch": 12.931654676258994,
+      "force_loss": 0.011200557462871075,
+      "step": 14380
+    },
+    {
+      "epoch": 12.940647482014388,
+      "grad_norm": 0.46101492643356323,
+      "learning_rate": 9.046866017833193e-05,
+      "loss": 0.0215,
+      "step": 14390
+    },
+    {
+      "action_loss": 0.011300328187644482,
+      "epoch": 12.940647482014388,
+      "step": 14390
+    },
+    {
+      "epoch": 12.940647482014388,
+      "step": 14390,
+      "torque_loss": 0.13503290712833405
+    },
+    {
+      "epoch": 12.940647482014388,
+      "force_loss": 0.009805439971387386,
+      "step": 14390
+    },
+    {
+      "epoch": 12.949640287769784,
+      "grad_norm": 0.3314456641674042,
+      "learning_rate": 9.045246946593029e-05,
+      "loss": 0.0254,
+      "step": 14400
+    },
+    {
+      "action_loss": 0.005221536383032799,
+      "epoch": 12.949640287769784,
+      "step": 14400
+    },
+    {
+      "epoch": 12.949640287769784,
+      "step": 14400,
+      "torque_loss": 0.14278076589107513
+    },
+    {
+      "epoch": 12.949640287769784,
+      "force_loss": 0.00457338010892272,
+      "step": 14400
+    },
+    {
+      "epoch": 12.95863309352518,
+      "grad_norm": 0.710391104221344,
+      "learning_rate": 9.043626646513652e-05,
+      "loss": 0.0233,
+      "step": 14410
+    },
+    {
+      "action_loss": 0.004302104935050011,
+      "epoch": 12.95863309352518,
+      "step": 14410
+    },
+    {
+      "epoch": 12.95863309352518,
+      "step": 14410,
+      "torque_loss": 0.14447474479675293
+    },
+    {
+      "epoch": 12.95863309352518,
+      "force_loss": 0.0045445640571415424,
+      "step": 14410
+    },
+    {
+      "epoch": 12.967625899280575,
+      "grad_norm": 0.6171999573707581,
+      "learning_rate": 9.042005118087267e-05,
+      "loss": 0.0225,
+      "step": 14420
+    },
+    {
+      "action_loss": 0.002514268271625042,
+      "epoch": 12.967625899280575,
+      "step": 14420
+    },
+    {
+      "epoch": 12.967625899280575,
+      "step": 14420,
+      "torque_loss": 0.09670746326446533
+    },
+    {
+      "epoch": 12.967625899280575,
+      "force_loss": 0.0036105215549468994,
+      "step": 14420
+    },
+    {
+      "epoch": 12.976618705035971,
+      "grad_norm": 0.441057950258255,
+      "learning_rate": 9.040382361806448e-05,
+      "loss": 0.0281,
+      "step": 14430
+    },
+    {
+      "action_loss": 0.004521968308836222,
+      "epoch": 12.976618705035971,
+      "step": 14430
+    },
+    {
+      "epoch": 12.976618705035971,
+      "step": 14430,
+      "torque_loss": 0.13695268332958221
+    },
+    {
+      "epoch": 12.976618705035971,
+      "force_loss": 0.005111285485327244,
+      "step": 14430
+    },
+    {
+      "epoch": 12.985611510791367,
+      "grad_norm": 0.36860525608062744,
+      "learning_rate": 9.038758378164148e-05,
+      "loss": 0.0317,
+      "step": 14440
+    },
+    {
+      "action_loss": 0.008770656771957874,
+      "epoch": 12.985611510791367,
+      "step": 14440
+    },
+    {
+      "epoch": 12.985611510791367,
+      "step": 14440,
+      "torque_loss": 0.13073667883872986
+    },
+    {
+      "epoch": 12.985611510791367,
+      "force_loss": 0.0072477455250918865,
+      "step": 14440
+    },
+    {
+      "epoch": 12.994604316546763,
+      "grad_norm": 0.42620325088500977,
+      "learning_rate": 9.037133167653691e-05,
+      "loss": 0.0245,
+      "step": 14450
+    },
+    {
+      "action_loss": 0.017914768308401108,
+      "epoch": 12.994604316546763,
+      "step": 14450
+    },
+    {
+      "epoch": 12.994604316546763,
+      "step": 14450,
+      "torque_loss": 0.15631122887134552
+    },
+    {
+      "epoch": 12.994604316546763,
+      "force_loss": 0.012198311276733875,
+      "step": 14450
+    },
+    {
+      "epoch": 13.003597122302159,
+      "grad_norm": 0.3327738344669342,
+      "learning_rate": 9.035506730768771e-05,
+      "loss": 0.0245,
+      "step": 14460
+    },
+    {
+      "action_loss": 0.0034663667902350426,
+      "epoch": 13.003597122302159,
+      "step": 14460
+    },
+    {
+      "epoch": 13.003597122302159,
+      "step": 14460,
+      "torque_loss": 0.10766801983118057
+    },
+    {
+      "epoch": 13.003597122302159,
+      "force_loss": 0.0042332750745117664,
+      "step": 14460
+    },
+    {
+      "epoch": 13.012589928057555,
+      "grad_norm": 0.4596000611782074,
+      "learning_rate": 9.033879068003458e-05,
+      "loss": 0.0242,
+      "step": 14470
+    },
+    {
+      "action_loss": 0.007420101668685675,
+      "epoch": 13.012589928057555,
+      "step": 14470
+    },
+    {
+      "epoch": 13.012589928057555,
+      "step": 14470,
+      "torque_loss": 0.12554533779621124
+    },
+    {
+      "epoch": 13.012589928057555,
+      "force_loss": 0.00997112225741148,
+      "step": 14470
+    },
+    {
+      "epoch": 13.02158273381295,
+      "grad_norm": 0.441500723361969,
+      "learning_rate": 9.032250179852193e-05,
+      "loss": 0.0255,
+      "step": 14480
+    },
+    {
+      "action_loss": 0.007700768765062094,
+      "epoch": 13.02158273381295,
+      "step": 14480
+    },
+    {
+      "epoch": 13.02158273381295,
+      "step": 14480,
+      "torque_loss": 0.12382280081510544
+    },
+    {
+      "epoch": 13.02158273381295,
+      "force_loss": 0.004308726638555527,
+      "step": 14480
+    },
+    {
+      "epoch": 13.030575539568344,
+      "grad_norm": 0.26572030782699585,
+      "learning_rate": 9.030620066809787e-05,
+      "loss": 0.0228,
+      "step": 14490
+    },
+    {
+      "action_loss": 0.006130896508693695,
+      "epoch": 13.030575539568344,
+      "step": 14490
+    },
+    {
+      "epoch": 13.030575539568344,
+      "step": 14490,
+      "torque_loss": 0.11724263429641724
+    },
+    {
+      "epoch": 13.030575539568344,
+      "force_loss": 0.006546203047037125,
+      "step": 14490
+    },
+    {
+      "epoch": 13.03956834532374,
+      "grad_norm": 0.32110661268234253,
+      "learning_rate": 9.028988729371428e-05,
+      "loss": 0.0233,
+      "step": 14500
+    },
+    {
+      "action_loss": 0.011450658552348614,
+      "epoch": 13.03956834532374,
+      "step": 14500
+    },
+    {
+      "epoch": 13.03956834532374,
+      "step": 14500,
+      "torque_loss": 0.15253585577011108
+    },
+    {
+      "epoch": 13.03956834532374,
+      "force_loss": 0.017569176852703094,
+      "step": 14500
+    },
+    {
+      "epoch": 13.048561151079136,
+      "grad_norm": 0.3928966522216797,
+      "learning_rate": 9.027356168032673e-05,
+      "loss": 0.0256,
+      "step": 14510
+    },
+    {
+      "action_loss": 0.004405443090945482,
+      "epoch": 13.048561151079136,
+      "step": 14510
+    },
+    {
+      "epoch": 13.048561151079136,
+      "step": 14510,
+      "torque_loss": 0.12579254806041718
+    },
+    {
+      "epoch": 13.048561151079136,
+      "force_loss": 0.005160333588719368,
+      "step": 14510
+    },
+    {
+      "epoch": 13.057553956834532,
+      "grad_norm": 0.677519679069519,
+      "learning_rate": 9.02572238328945e-05,
+      "loss": 0.0274,
+      "step": 14520
+    },
+    {
+      "action_loss": 0.011080168187618256,
+      "epoch": 13.057553956834532,
+      "step": 14520
+    },
+    {
+      "epoch": 13.057553956834532,
+      "step": 14520,
+      "torque_loss": 0.09257916361093521
+    },
+    {
+      "epoch": 13.057553956834532,
+      "force_loss": 0.009711046703159809,
+      "step": 14520
+    },
+    {
+      "epoch": 13.066546762589928,
+      "grad_norm": 0.4754875600337982,
+      "learning_rate": 9.02408737563806e-05,
+      "loss": 0.0264,
+      "step": 14530
+    },
+    {
+      "action_loss": 0.006562229245901108,
+      "epoch": 13.066546762589928,
+      "step": 14530
+    },
+    {
+      "epoch": 13.066546762589928,
+      "step": 14530,
+      "torque_loss": 0.15128637850284576
+    },
+    {
+      "epoch": 13.066546762589928,
+      "force_loss": 0.006391929928213358,
+      "step": 14530
+    },
+    {
+      "epoch": 13.075539568345324,
+      "grad_norm": 0.3993264138698578,
+      "learning_rate": 9.022451145575174e-05,
+      "loss": 0.0249,
+      "step": 14540
+    },
+    {
+      "action_loss": 0.003944835159927607,
+      "epoch": 13.075539568345324,
+      "step": 14540
+    },
+    {
+      "epoch": 13.075539568345324,
+      "step": 14540,
+      "torque_loss": 0.1241079792380333
+    },
+    {
+      "epoch": 13.075539568345324,
+      "force_loss": 0.004877950064837933,
+      "step": 14540
+    },
+    {
+      "epoch": 13.08453237410072,
+      "grad_norm": 0.8384110927581787,
+      "learning_rate": 9.02081369359784e-05,
+      "loss": 0.0233,
+      "step": 14550
+    },
+    {
+      "action_loss": 0.0061592659913003445,
+      "epoch": 13.08453237410072,
+      "step": 14550
+    },
+    {
+      "epoch": 13.08453237410072,
+      "step": 14550,
+      "torque_loss": 0.09389689564704895
+    },
+    {
+      "epoch": 13.08453237410072,
+      "force_loss": 0.007377520203590393,
+      "step": 14550
+    },
+    {
+      "epoch": 13.093525179856115,
+      "grad_norm": 0.5473024845123291,
+      "learning_rate": 9.019175020203465e-05,
+      "loss": 0.0214,
+      "step": 14560
+    },
+    {
+      "action_loss": 0.00808979943394661,
+      "epoch": 13.093525179856115,
+      "step": 14560
+    },
+    {
+      "epoch": 13.093525179856115,
+      "step": 14560,
+      "torque_loss": 0.18886379897594452
+    },
+    {
+      "epoch": 13.093525179856115,
+      "force_loss": 0.0075078862719237804,
+      "step": 14560
+    },
+    {
+      "epoch": 13.102517985611511,
+      "grad_norm": 0.8063585162162781,
+      "learning_rate": 9.017535125889842e-05,
+      "loss": 0.0212,
+      "step": 14570
+    },
+    {
+      "action_loss": 0.01819506473839283,
+      "epoch": 13.102517985611511,
+      "step": 14570
+    },
+    {
+      "epoch": 13.102517985611511,
+      "step": 14570,
+      "torque_loss": 0.15776431560516357
+    },
+    {
+      "epoch": 13.102517985611511,
+      "force_loss": 0.01641392707824707,
+      "step": 14570
+    },
+    {
+      "epoch": 13.111510791366907,
+      "grad_norm": 0.34871530532836914,
+      "learning_rate": 9.015894011155124e-05,
+      "loss": 0.0272,
+      "step": 14580
+    },
+    {
+      "action_loss": 0.018761910498142242,
+      "epoch": 13.111510791366907,
+      "step": 14580
+    },
+    {
+      "epoch": 13.111510791366907,
+      "step": 14580,
+      "torque_loss": 0.17186866700649261
+    },
+    {
+      "epoch": 13.111510791366907,
+      "force_loss": 0.01540849357843399,
+      "step": 14580
+    },
+    {
+      "epoch": 13.120503597122303,
+      "grad_norm": 0.5542654395103455,
+      "learning_rate": 9.014251676497838e-05,
+      "loss": 0.0297,
+      "step": 14590
+    },
+    {
+      "action_loss": 0.0072064208798110485,
+      "epoch": 13.120503597122303,
+      "step": 14590
+    },
+    {
+      "epoch": 13.120503597122303,
+      "step": 14590,
+      "torque_loss": 0.11462417989969254
+    },
+    {
+      "epoch": 13.120503597122303,
+      "force_loss": 0.007690413389354944,
+      "step": 14590
+    },
+    {
+      "epoch": 13.129496402877697,
+      "grad_norm": 0.3077492415904999,
+      "learning_rate": 9.012608122416884e-05,
+      "loss": 0.024,
+      "step": 14600
+    },
+    {
+      "action_loss": 0.006153901573270559,
+      "epoch": 13.129496402877697,
+      "step": 14600
+    },
+    {
+      "epoch": 13.129496402877697,
+      "step": 14600,
+      "torque_loss": 0.1357250064611435
+    },
+    {
+      "epoch": 13.129496402877697,
+      "force_loss": 0.006215449422597885,
+      "step": 14600
+    },
+    {
+      "epoch": 13.138489208633093,
+      "grad_norm": 0.3804683983325958,
+      "learning_rate": 9.010963349411529e-05,
+      "loss": 0.0277,
+      "step": 14610
+    },
+    {
+      "action_loss": 0.028855418786406517,
+      "epoch": 13.138489208633093,
+      "step": 14610
+    },
+    {
+      "epoch": 13.138489208633093,
+      "step": 14610,
+      "torque_loss": 0.14814309775829315
+    },
+    {
+      "epoch": 13.138489208633093,
+      "force_loss": 0.024809643626213074,
+      "step": 14610
+    },
+    {
+      "epoch": 13.147482014388489,
+      "grad_norm": 0.3638586699962616,
+      "learning_rate": 9.00931735798141e-05,
+      "loss": 0.0279,
+      "step": 14620
+    },
+    {
+      "action_loss": 0.007410066667944193,
+      "epoch": 13.147482014388489,
+      "step": 14620
+    },
+    {
+      "epoch": 13.147482014388489,
+      "step": 14620,
+      "torque_loss": 0.11786165088415146
+    },
+    {
+      "epoch": 13.147482014388489,
+      "force_loss": 0.004824485629796982,
+      "step": 14620
+    },
+    {
+      "epoch": 13.156474820143885,
+      "grad_norm": 0.40190863609313965,
+      "learning_rate": 9.00767014862654e-05,
+      "loss": 0.0227,
+      "step": 14630
+    },
+    {
+      "action_loss": 0.0038182036951184273,
+      "epoch": 13.156474820143885,
+      "step": 14630
+    },
+    {
+      "epoch": 13.156474820143885,
+      "step": 14630,
+      "torque_loss": 0.13455848395824432
+    },
+    {
+      "epoch": 13.156474820143885,
+      "force_loss": 0.003665941534563899,
+      "step": 14630
+    },
+    {
+      "epoch": 13.16546762589928,
+      "grad_norm": 0.4742845296859741,
+      "learning_rate": 9.006021721847295e-05,
+      "loss": 0.0229,
+      "step": 14640
+    },
+    {
+      "action_loss": 0.017086701467633247,
+      "epoch": 13.16546762589928,
+      "step": 14640
+    },
+    {
+      "epoch": 13.16546762589928,
+      "step": 14640,
+      "torque_loss": 0.16399197280406952
+    },
+    {
+      "epoch": 13.16546762589928,
+      "force_loss": 0.0141557976603508,
+      "step": 14640
+    },
+    {
+      "epoch": 13.174460431654676,
+      "grad_norm": 0.6930900812149048,
+      "learning_rate": 9.004372078144423e-05,
+      "loss": 0.0255,
+      "step": 14650
+    },
+    {
+      "action_loss": 0.017907066270709038,
+      "epoch": 13.174460431654676,
+      "step": 14650
+    },
+    {
+      "epoch": 13.174460431654676,
+      "step": 14650,
+      "torque_loss": 0.16293351352214813
+    },
+    {
+      "epoch": 13.174460431654676,
+      "force_loss": 0.024945497512817383,
+      "step": 14650
+    },
+    {
+      "epoch": 13.183453237410072,
+      "grad_norm": 0.5228962898254395,
+      "learning_rate": 9.002721218019043e-05,
+      "loss": 0.0265,
+      "step": 14660
+    },
+    {
+      "action_loss": 0.006672768387943506,
+      "epoch": 13.183453237410072,
+      "step": 14660
+    },
+    {
+      "epoch": 13.183453237410072,
+      "step": 14660,
+      "torque_loss": 0.11546932905912399
+    },
+    {
+      "epoch": 13.183453237410072,
+      "force_loss": 0.007314169779419899,
+      "step": 14660
+    },
+    {
+      "epoch": 13.192446043165468,
+      "grad_norm": 0.40475955605506897,
+      "learning_rate": 9.001069141972642e-05,
+      "loss": 0.0233,
+      "step": 14670
+    },
+    {
+      "action_loss": 0.012470702640712261,
+      "epoch": 13.192446043165468,
+      "step": 14670
+    },
+    {
+      "epoch": 13.192446043165468,
+      "step": 14670,
+      "torque_loss": 0.09179221838712692
+    },
+    {
+      "epoch": 13.192446043165468,
+      "force_loss": 0.024287141859531403,
+      "step": 14670
+    },
+    {
+      "epoch": 13.201438848920864,
+      "grad_norm": 0.40371808409690857,
+      "learning_rate": 8.99941585050708e-05,
+      "loss": 0.0274,
+      "step": 14680
+    },
+    {
+      "action_loss": 0.010434582829475403,
+      "epoch": 13.201438848920864,
+      "step": 14680
+    },
+    {
+      "epoch": 13.201438848920864,
+      "step": 14680,
+      "torque_loss": 0.16278213262557983
+    },
+    {
+      "epoch": 13.201438848920864,
+      "force_loss": 0.011291473172605038,
+      "step": 14680
+    },
+    {
+      "epoch": 13.21043165467626,
+      "grad_norm": 0.431395947933197,
+      "learning_rate": 8.997761344124578e-05,
+      "loss": 0.0262,
+      "step": 14690
+    },
+    {
+      "action_loss": 0.004833641927689314,
+      "epoch": 13.21043165467626,
+      "step": 14690
+    },
+    {
+      "epoch": 13.21043165467626,
+      "step": 14690,
+      "torque_loss": 0.16232754290103912
+    },
+    {
+      "epoch": 13.21043165467626,
+      "force_loss": 0.004413580056279898,
+      "step": 14690
+    },
+    {
+      "epoch": 13.219424460431656,
+      "grad_norm": 0.315735399723053,
+      "learning_rate": 8.996105623327737e-05,
+      "loss": 0.0224,
+      "step": 14700
+    },
+    {
+      "action_loss": 0.008078006096184254,
+      "epoch": 13.219424460431656,
+      "step": 14700
+    },
+    {
+      "epoch": 13.219424460431656,
+      "step": 14700,
+      "torque_loss": 0.13483105599880219
+    },
+    {
+      "epoch": 13.219424460431656,
+      "force_loss": 0.0045743403024971485,
+      "step": 14700
+    },
+    {
+      "epoch": 13.22841726618705,
+      "grad_norm": 0.3824552297592163,
+      "learning_rate": 8.994448688619517e-05,
+      "loss": 0.0235,
+      "step": 14710
+    },
+    {
+      "action_loss": 0.006124551873654127,
+      "epoch": 13.22841726618705,
+      "step": 14710
+    },
+    {
+      "epoch": 13.22841726618705,
+      "step": 14710,
+      "torque_loss": 0.11799757927656174
+    },
+    {
+      "epoch": 13.22841726618705,
+      "force_loss": 0.006544084753841162,
+      "step": 14710
+    },
+    {
+      "epoch": 13.237410071942445,
+      "grad_norm": 0.3821107745170593,
+      "learning_rate": 8.992790540503253e-05,
+      "loss": 0.0245,
+      "step": 14720
+    },
+    {
+      "action_loss": 0.007243847008794546,
+      "epoch": 13.237410071942445,
+      "step": 14720
+    },
+    {
+      "epoch": 13.237410071942445,
+      "step": 14720,
+      "torque_loss": 0.1174304261803627
+    },
+    {
+      "epoch": 13.237410071942445,
+      "force_loss": 0.004242307972162962,
+      "step": 14720
+    },
+    {
+      "epoch": 13.246402877697841,
+      "grad_norm": 0.2759091556072235,
+      "learning_rate": 8.991131179482648e-05,
+      "loss": 0.0219,
+      "step": 14730
+    },
+    {
+      "action_loss": 0.006294190883636475,
+      "epoch": 13.246402877697841,
+      "step": 14730
+    },
+    {
+      "epoch": 13.246402877697841,
+      "step": 14730,
+      "torque_loss": 0.13765421509742737
+    },
+    {
+      "epoch": 13.246402877697841,
+      "force_loss": 0.00725896842777729,
+      "step": 14730
+    },
+    {
+      "epoch": 13.255395683453237,
+      "grad_norm": 0.4401913583278656,
+      "learning_rate": 8.989470606061768e-05,
+      "loss": 0.0273,
+      "step": 14740
+    },
+    {
+      "action_loss": 0.009603387676179409,
+      "epoch": 13.255395683453237,
+      "step": 14740
+    },
+    {
+      "epoch": 13.255395683453237,
+      "step": 14740,
+      "torque_loss": 0.13843569159507751
+    },
+    {
+      "epoch": 13.255395683453237,
+      "force_loss": 0.007131265010684729,
+      "step": 14740
+    },
+    {
+      "epoch": 13.264388489208633,
+      "grad_norm": 0.4701067805290222,
+      "learning_rate": 8.987808820745056e-05,
+      "loss": 0.0267,
+      "step": 14750
+    },
+    {
+      "action_loss": 0.011600494384765625,
+      "epoch": 13.264388489208633,
+      "step": 14750
+    },
+    {
+      "epoch": 13.264388489208633,
+      "step": 14750,
+      "torque_loss": 0.1349010467529297
+    },
+    {
+      "epoch": 13.264388489208633,
+      "force_loss": 0.016512449830770493,
+      "step": 14750
+    },
+    {
+      "epoch": 13.273381294964029,
+      "grad_norm": 0.5634354948997498,
+      "learning_rate": 8.986145824037315e-05,
+      "loss": 0.0235,
+      "step": 14760
+    },
+    {
+      "action_loss": 0.018544087186455727,
+      "epoch": 13.273381294964029,
+      "step": 14760
+    },
+    {
+      "epoch": 13.273381294964029,
+      "step": 14760,
+      "torque_loss": 0.13523080945014954
+    },
+    {
+      "epoch": 13.273381294964029,
+      "force_loss": 0.02288237400352955,
+      "step": 14760
+    },
+    {
+      "epoch": 13.282374100719425,
+      "grad_norm": 0.4378294050693512,
+      "learning_rate": 8.984481616443721e-05,
+      "loss": 0.0336,
+      "step": 14770
+    },
+    {
+      "action_loss": 0.010487220250070095,
+      "epoch": 13.282374100719425,
+      "step": 14770
+    },
+    {
+      "epoch": 13.282374100719425,
+      "step": 14770,
+      "torque_loss": 0.15143419802188873
+    },
+    {
+      "epoch": 13.282374100719425,
+      "force_loss": 0.00698075070977211,
+      "step": 14770
+    },
+    {
+      "epoch": 13.29136690647482,
+      "grad_norm": 0.3537493944168091,
+      "learning_rate": 8.982816198469815e-05,
+      "loss": 0.023,
+      "step": 14780
+    },
+    {
+      "action_loss": 0.0047593191266059875,
+      "epoch": 13.29136690647482,
+      "step": 14780
+    },
+    {
+      "epoch": 13.29136690647482,
+      "step": 14780,
+      "torque_loss": 0.10496661812067032
+    },
+    {
+      "epoch": 13.29136690647482,
+      "force_loss": 0.003831305541098118,
+      "step": 14780
+    },
+    {
+      "epoch": 13.300359712230216,
+      "grad_norm": 0.21889355778694153,
+      "learning_rate": 8.98114957062151e-05,
+      "loss": 0.0269,
+      "step": 14790
+    },
+    {
+      "action_loss": 0.007127129938453436,
+      "epoch": 13.300359712230216,
+      "step": 14790
+    },
+    {
+      "epoch": 13.300359712230216,
+      "step": 14790,
+      "torque_loss": 0.14200273156166077
+    },
+    {
+      "epoch": 13.300359712230216,
+      "force_loss": 0.013411667197942734,
+      "step": 14790
+    },
+    {
+      "epoch": 13.309352517985612,
+      "grad_norm": 0.3099175691604614,
+      "learning_rate": 8.97948173340508e-05,
+      "loss": 0.0221,
+      "step": 14800
+    },
+    {
+      "action_loss": 0.006687544751912355,
+      "epoch": 13.309352517985612,
+      "step": 14800
+    },
+    {
+      "epoch": 13.309352517985612,
+      "step": 14800,
+      "torque_loss": 0.14193305373191833
+    },
+    {
+      "epoch": 13.309352517985612,
+      "force_loss": 0.0054398588836193085,
+      "step": 14800
+    },
+    {
+      "epoch": 13.318345323741006,
+      "grad_norm": 0.6632286906242371,
+      "learning_rate": 8.977812687327172e-05,
+      "loss": 0.0202,
+      "step": 14810
+    },
+    {
+      "action_loss": 0.03508426994085312,
+      "epoch": 13.318345323741006,
+      "step": 14810
+    },
+    {
+      "epoch": 13.318345323741006,
+      "step": 14810,
+      "torque_loss": 0.1740388721227646
+    },
+    {
+      "epoch": 13.318345323741006,
+      "force_loss": 0.02941388450562954,
+      "step": 14810
+    },
+    {
+      "epoch": 13.327338129496402,
+      "grad_norm": 0.5804286599159241,
+      "learning_rate": 8.976142432894798e-05,
+      "loss": 0.0261,
+      "step": 14820
+    },
+    {
+      "action_loss": 0.006049271672964096,
+      "epoch": 13.327338129496402,
+      "step": 14820
+    },
+    {
+      "epoch": 13.327338129496402,
+      "step": 14820,
+      "torque_loss": 0.10105837136507034
+    },
+    {
+      "epoch": 13.327338129496402,
+      "force_loss": 0.006374318152666092,
+      "step": 14820
+    },
+    {
+      "epoch": 13.336330935251798,
+      "grad_norm": 0.5489843487739563,
+      "learning_rate": 8.974470970615336e-05,
+      "loss": 0.0271,
+      "step": 14830
+    },
+    {
+      "action_loss": 0.019841017201542854,
+      "epoch": 13.336330935251798,
+      "step": 14830
+    },
+    {
+      "epoch": 13.336330935251798,
+      "step": 14830,
+      "torque_loss": 0.1333627551794052
+    },
+    {
+      "epoch": 13.336330935251798,
+      "force_loss": 0.018949931487441063,
+      "step": 14830
+    },
+    {
+      "epoch": 13.345323741007194,
+      "grad_norm": 0.5089920163154602,
+      "learning_rate": 8.972798300996534e-05,
+      "loss": 0.0284,
+      "step": 14840
+    },
+    {
+      "action_loss": 0.004613576456904411,
+      "epoch": 13.345323741007194,
+      "step": 14840
+    },
+    {
+      "epoch": 13.345323741007194,
+      "step": 14840,
+      "torque_loss": 0.12690582871437073
+    },
+    {
+      "epoch": 13.345323741007194,
+      "force_loss": 0.006062695290893316,
+      "step": 14840
+    },
+    {
+      "epoch": 13.35431654676259,
+      "grad_norm": 0.5182651281356812,
+      "learning_rate": 8.971124424546504e-05,
+      "loss": 0.023,
+      "step": 14850
+    },
+    {
+      "action_loss": 0.006306968629360199,
+      "epoch": 13.35431654676259,
+      "step": 14850
+    },
+    {
+      "epoch": 13.35431654676259,
+      "step": 14850,
+      "torque_loss": 0.1303333193063736
+    },
+    {
+      "epoch": 13.35431654676259,
+      "force_loss": 0.009413695894181728,
+      "step": 14850
+    },
+    {
+      "epoch": 13.363309352517986,
+      "grad_norm": 0.28933581709861755,
+      "learning_rate": 8.969449341773724e-05,
+      "loss": 0.0257,
+      "step": 14860
+    },
+    {
+      "action_loss": 0.009394045919179916,
+      "epoch": 13.363309352517986,
+      "step": 14860
+    },
+    {
+      "epoch": 13.363309352517986,
+      "step": 14860,
+      "torque_loss": 0.1231849193572998
+    },
+    {
+      "epoch": 13.363309352517986,
+      "force_loss": 0.010514688678085804,
+      "step": 14860
+    },
+    {
+      "epoch": 13.372302158273381,
+      "grad_norm": 0.5404540300369263,
+      "learning_rate": 8.967773053187042e-05,
+      "loss": 0.0298,
+      "step": 14870
+    },
+    {
+      "action_loss": 0.007335355039685965,
+      "epoch": 13.372302158273381,
+      "step": 14870
+    },
+    {
+      "epoch": 13.372302158273381,
+      "step": 14870,
+      "torque_loss": 0.09684888273477554
+    },
+    {
+      "epoch": 13.372302158273381,
+      "force_loss": 0.010095199570059776,
+      "step": 14870
+    },
+    {
+      "epoch": 13.381294964028777,
+      "grad_norm": 0.2733374834060669,
+      "learning_rate": 8.966095559295668e-05,
+      "loss": 0.0242,
+      "step": 14880
+    },
+    {
+      "action_loss": 0.004407978150993586,
+      "epoch": 13.381294964028777,
+      "step": 14880
+    },
+    {
+      "epoch": 13.381294964028777,
+      "step": 14880,
+      "torque_loss": 0.16121593117713928
+    },
+    {
+      "epoch": 13.381294964028777,
+      "force_loss": 0.00505354767665267,
+      "step": 14880
+    },
+    {
+      "epoch": 13.390287769784173,
+      "grad_norm": 0.3442193567752838,
+      "learning_rate": 8.964416860609184e-05,
+      "loss": 0.0268,
+      "step": 14890
+    },
+    {
+      "action_loss": 0.008913670666515827,
+      "epoch": 13.390287769784173,
+      "step": 14890
+    },
+    {
+      "epoch": 13.390287769784173,
+      "step": 14890,
+      "torque_loss": 0.10350557416677475
+    },
+    {
+      "epoch": 13.390287769784173,
+      "force_loss": 0.008378827013075352,
+      "step": 14890
+    },
+    {
+      "epoch": 13.399280575539569,
+      "grad_norm": 0.7161759734153748,
+      "learning_rate": 8.962736957637532e-05,
+      "loss": 0.0211,
+      "step": 14900
+    },
+    {
+      "action_loss": 0.002452212618663907,
+      "epoch": 13.399280575539569,
+      "step": 14900
+    },
+    {
+      "epoch": 13.399280575539569,
+      "step": 14900,
+      "torque_loss": 0.0714641883969307
+    },
+    {
+      "epoch": 13.399280575539569,
+      "force_loss": 0.0027752381283789873,
+      "step": 14900
+    },
+    {
+      "epoch": 13.408273381294965,
+      "grad_norm": 0.23730121552944183,
+      "learning_rate": 8.96105585089102e-05,
+      "loss": 0.0188,
+      "step": 14910
+    },
+    {
+      "action_loss": 0.03291613236069679,
+      "epoch": 13.408273381294965,
+      "step": 14910
+    },
+    {
+      "epoch": 13.408273381294965,
+      "step": 14910,
+      "torque_loss": 0.1959151029586792
+    },
+    {
+      "epoch": 13.408273381294965,
+      "force_loss": 0.024787189438939095,
+      "step": 14910
+    },
+    {
+      "epoch": 13.417266187050359,
+      "grad_norm": 0.32745203375816345,
+      "learning_rate": 8.959373540880329e-05,
+      "loss": 0.0307,
+      "step": 14920
+    },
+    {
+      "action_loss": 0.008861460722982883,
+      "epoch": 13.417266187050359,
+      "step": 14920
+    },
+    {
+      "epoch": 13.417266187050359,
+      "step": 14920,
+      "torque_loss": 0.1019737720489502
+    },
+    {
+      "epoch": 13.417266187050359,
+      "force_loss": 0.0068197958171367645,
+      "step": 14920
+    },
+    {
+      "epoch": 13.426258992805755,
+      "grad_norm": 0.3431726396083832,
+      "learning_rate": 8.957690028116495e-05,
+      "loss": 0.0268,
+      "step": 14930
+    },
+    {
+      "action_loss": 0.007505863904953003,
+      "epoch": 13.426258992805755,
+      "step": 14930
+    },
+    {
+      "epoch": 13.426258992805755,
+      "step": 14930,
+      "torque_loss": 0.12415751814842224
+    },
+    {
+      "epoch": 13.426258992805755,
+      "force_loss": 0.0055229924619197845,
+      "step": 14930
+    },
+    {
+      "epoch": 13.43525179856115,
+      "grad_norm": 0.2953338027000427,
+      "learning_rate": 8.956005313110928e-05,
+      "loss": 0.0234,
+      "step": 14940
+    },
+    {
+      "action_loss": 0.0041083176620304585,
+      "epoch": 13.43525179856115,
+      "step": 14940
+    },
+    {
+      "epoch": 13.43525179856115,
+      "step": 14940,
+      "torque_loss": 0.16240069270133972
+    },
+    {
+      "epoch": 13.43525179856115,
+      "force_loss": 0.005068450700491667,
+      "step": 14940
+    },
+    {
+      "epoch": 13.444244604316546,
+      "grad_norm": 0.3561532199382782,
+      "learning_rate": 8.9543193963754e-05,
+      "loss": 0.0265,
+      "step": 14950
+    },
+    {
+      "action_loss": 0.011054660193622112,
+      "epoch": 13.444244604316546,
+      "step": 14950
+    },
+    {
+      "epoch": 13.444244604316546,
+      "step": 14950,
+      "torque_loss": 0.11108561605215073
+    },
+    {
+      "epoch": 13.444244604316546,
+      "force_loss": 0.009395153261721134,
+      "step": 14950
+    },
+    {
+      "epoch": 13.453237410071942,
+      "grad_norm": 0.6054490208625793,
+      "learning_rate": 8.952632278422048e-05,
+      "loss": 0.0247,
+      "step": 14960
+    },
+    {
+      "action_loss": 0.010276337154209614,
+      "epoch": 13.453237410071942,
+      "step": 14960
+    },
+    {
+      "epoch": 13.453237410071942,
+      "step": 14960,
+      "torque_loss": 0.11921677738428116
+    },
+    {
+      "epoch": 13.453237410071942,
+      "force_loss": 0.008038091473281384,
+      "step": 14960
+    },
+    {
+      "epoch": 13.462230215827338,
+      "grad_norm": 0.42296022176742554,
+      "learning_rate": 8.95094395976337e-05,
+      "loss": 0.0289,
+      "step": 14970
+    },
+    {
+      "action_loss": 0.011814437806606293,
+      "epoch": 13.462230215827338,
+      "step": 14970
+    },
+    {
+      "epoch": 13.462230215827338,
+      "step": 14970,
+      "torque_loss": 0.15883596241474152
+    },
+    {
+      "epoch": 13.462230215827338,
+      "force_loss": 0.0067948270589113235,
+      "step": 14970
+    },
+    {
+      "epoch": 13.471223021582734,
+      "grad_norm": 0.36728227138519287,
+      "learning_rate": 8.949254440912239e-05,
+      "loss": 0.0271,
+      "step": 14980
+    },
+    {
+      "action_loss": 0.00897889118641615,
+      "epoch": 13.471223021582734,
+      "step": 14980
+    },
+    {
+      "epoch": 13.471223021582734,
+      "step": 14980,
+      "torque_loss": 0.14344914257526398
+    },
+    {
+      "epoch": 13.471223021582734,
+      "force_loss": 0.0068875751458108425,
+      "step": 14980
+    },
+    {
+      "epoch": 13.48021582733813,
+      "grad_norm": 0.5201298594474792,
+      "learning_rate": 8.94756372238188e-05,
+      "loss": 0.027,
+      "step": 14990
+    },
+    {
+      "action_loss": 0.012513409368693829,
+      "epoch": 13.48021582733813,
+      "step": 14990
+    },
+    {
+      "epoch": 13.48021582733813,
+      "step": 14990,
+      "torque_loss": 0.15048789978027344
+    },
+    {
+      "epoch": 13.48021582733813,
+      "force_loss": 0.010743759572505951,
+      "step": 14990
+    },
+    {
+      "epoch": 13.489208633093526,
+      "grad_norm": 0.4812636375427246,
+      "learning_rate": 8.945871804685892e-05,
+      "loss": 0.0249,
+      "step": 15000
+    },
+    {
+      "action_loss": 0.011180832982063293,
+      "epoch": 13.489208633093526,
+      "step": 15000
+    },
+    {
+      "epoch": 13.489208633093526,
+      "step": 15000,
+      "torque_loss": 0.18805503845214844
+    },
+    {
+      "epoch": 13.489208633093526,
+      "force_loss": 0.0050537544302642345,
+      "step": 15000
+    },
+    {
+      "epoch": 13.498201438848922,
+      "grad_norm": 0.2539704740047455,
+      "learning_rate": 8.944178688338236e-05,
+      "loss": 0.023,
+      "step": 15010
+    },
+    {
+      "action_loss": 0.018569903448224068,
+      "epoch": 13.498201438848922,
+      "step": 15010
+    },
+    {
+      "epoch": 13.498201438848922,
+      "step": 15010,
+      "torque_loss": 0.11188825964927673
+    },
+    {
+      "epoch": 13.498201438848922,
+      "force_loss": 0.017913958057761192,
+      "step": 15010
+    },
+    {
+      "epoch": 13.507194244604317,
+      "grad_norm": 0.5525025725364685,
+      "learning_rate": 8.942484373853233e-05,
+      "loss": 0.0206,
+      "step": 15020
+    },
+    {
+      "action_loss": 0.00969319324940443,
+      "epoch": 13.507194244604317,
+      "step": 15020
+    },
+    {
+      "epoch": 13.507194244604317,
+      "step": 15020,
+      "torque_loss": 0.13361899554729462
+    },
+    {
+      "epoch": 13.507194244604317,
+      "force_loss": 0.0065063959918916225,
+      "step": 15020
+    },
+    {
+      "epoch": 13.516187050359711,
+      "grad_norm": 0.6648332476615906,
+      "learning_rate": 8.940788861745572e-05,
+      "loss": 0.0295,
+      "step": 15030
+    },
+    {
+      "action_loss": 0.004434612579643726,
+      "epoch": 13.516187050359711,
+      "step": 15030
+    },
+    {
+      "epoch": 13.516187050359711,
+      "step": 15030,
+      "torque_loss": 0.12967358529567719
+    },
+    {
+      "epoch": 13.516187050359711,
+      "force_loss": 0.00821138545870781,
+      "step": 15030
+    },
+    {
+      "epoch": 13.525179856115107,
+      "grad_norm": 0.4235135614871979,
+      "learning_rate": 8.939092152530308e-05,
+      "loss": 0.0242,
+      "step": 15040
+    },
+    {
+      "action_loss": 0.007038809359073639,
+      "epoch": 13.525179856115107,
+      "step": 15040
+    },
+    {
+      "epoch": 13.525179856115107,
+      "step": 15040,
+      "torque_loss": 0.1582469940185547
+    },
+    {
+      "epoch": 13.525179856115107,
+      "force_loss": 0.007739563938230276,
+      "step": 15040
+    },
+    {
+      "epoch": 13.534172661870503,
+      "grad_norm": 0.6255474090576172,
+      "learning_rate": 8.937394246722853e-05,
+      "loss": 0.0233,
+      "step": 15050
+    },
+    {
+      "action_loss": 0.011083395220339298,
+      "epoch": 13.534172661870503,
+      "step": 15050
+    },
+    {
+      "epoch": 13.534172661870503,
+      "step": 15050,
+      "torque_loss": 0.15567262470722198
+    },
+    {
+      "epoch": 13.534172661870503,
+      "force_loss": 0.008961132727563381,
+      "step": 15050
+    },
+    {
+      "epoch": 13.543165467625899,
+      "grad_norm": 0.6343432664871216,
+      "learning_rate": 8.935695144838984e-05,
+      "loss": 0.03,
+      "step": 15060
+    },
+    {
+      "action_loss": 0.01805730350315571,
+      "epoch": 13.543165467625899,
+      "step": 15060
+    },
+    {
+      "epoch": 13.543165467625899,
+      "step": 15060,
+      "torque_loss": 0.13492028415203094
+    },
+    {
+      "epoch": 13.543165467625899,
+      "force_loss": 0.012581147253513336,
+      "step": 15060
+    },
+    {
+      "epoch": 13.552158273381295,
+      "grad_norm": 1.0415568351745605,
+      "learning_rate": 8.933994847394849e-05,
+      "loss": 0.0242,
+      "step": 15070
+    },
+    {
+      "action_loss": 0.007867749780416489,
+      "epoch": 13.552158273381295,
+      "step": 15070
+    },
+    {
+      "epoch": 13.552158273381295,
+      "step": 15070,
+      "torque_loss": 0.14760714769363403
+    },
+    {
+      "epoch": 13.552158273381295,
+      "force_loss": 0.006027327850461006,
+      "step": 15070
+    },
+    {
+      "epoch": 13.56115107913669,
+      "grad_norm": 0.2590728998184204,
+      "learning_rate": 8.932293354906949e-05,
+      "loss": 0.0243,
+      "step": 15080
+    },
+    {
+      "action_loss": 0.0067918128333985806,
+      "epoch": 13.56115107913669,
+      "step": 15080
+    },
+    {
+      "epoch": 13.56115107913669,
+      "step": 15080,
+      "torque_loss": 0.16374565660953522
+    },
+    {
+      "epoch": 13.56115107913669,
+      "force_loss": 0.0048575554974377155,
+      "step": 15080
+    },
+    {
+      "epoch": 13.570143884892087,
+      "grad_norm": 0.3504115641117096,
+      "learning_rate": 8.930590667892153e-05,
+      "loss": 0.0255,
+      "step": 15090
+    },
+    {
+      "action_loss": 0.009288015775382519,
+      "epoch": 13.570143884892087,
+      "step": 15090
+    },
+    {
+      "epoch": 13.570143884892087,
+      "step": 15090,
+      "torque_loss": 0.10502278804779053
+    },
+    {
+      "epoch": 13.570143884892087,
+      "force_loss": 0.008320000022649765,
+      "step": 15090
+    },
+    {
+      "epoch": 13.579136690647482,
+      "grad_norm": 0.4551716148853302,
+      "learning_rate": 8.928886786867696e-05,
+      "loss": 0.0231,
+      "step": 15100
+    },
+    {
+      "action_loss": 0.005145413335412741,
+      "epoch": 13.579136690647482,
+      "step": 15100
+    },
+    {
+      "epoch": 13.579136690647482,
+      "step": 15100,
+      "torque_loss": 0.1395842432975769
+    },
+    {
+      "epoch": 13.579136690647482,
+      "force_loss": 0.007424239069223404,
+      "step": 15100
+    },
+    {
+      "epoch": 13.588129496402878,
+      "grad_norm": 0.33234086632728577,
+      "learning_rate": 8.927181712351168e-05,
+      "loss": 0.0205,
+      "step": 15110
+    },
+    {
+      "action_loss": 0.007719407323747873,
+      "epoch": 13.588129496402878,
+      "step": 15110
+    },
+    {
+      "epoch": 13.588129496402878,
+      "step": 15110,
+      "torque_loss": 0.12911628186702728
+    },
+    {
+      "epoch": 13.588129496402878,
+      "force_loss": 0.00613104784861207,
+      "step": 15110
+    },
+    {
+      "epoch": 13.597122302158274,
+      "grad_norm": 0.6952317953109741,
+      "learning_rate": 8.925475444860527e-05,
+      "loss": 0.0243,
+      "step": 15120
+    },
+    {
+      "action_loss": 0.005236349534243345,
+      "epoch": 13.597122302158274,
+      "step": 15120
+    },
+    {
+      "epoch": 13.597122302158274,
+      "step": 15120,
+      "torque_loss": 0.09986937791109085
+    },
+    {
+      "epoch": 13.597122302158274,
+      "force_loss": 0.007467932999134064,
+      "step": 15120
+    },
+    {
+      "epoch": 13.60611510791367,
+      "grad_norm": 0.2264104187488556,
+      "learning_rate": 8.923767984914092e-05,
+      "loss": 0.0202,
+      "step": 15130
+    },
+    {
+      "action_loss": 0.005853123497217894,
+      "epoch": 13.60611510791367,
+      "step": 15130
+    },
+    {
+      "epoch": 13.60611510791367,
+      "step": 15130,
+      "torque_loss": 0.14523430168628693
+    },
+    {
+      "epoch": 13.60611510791367,
+      "force_loss": 0.004577724263072014,
+      "step": 15130
+    },
+    {
+      "epoch": 13.615107913669064,
+      "grad_norm": 0.7617133259773254,
+      "learning_rate": 8.922059333030545e-05,
+      "loss": 0.0324,
+      "step": 15140
+    },
+    {
+      "action_loss": 0.003778591752052307,
+      "epoch": 13.615107913669064,
+      "step": 15140
+    },
+    {
+      "epoch": 13.615107913669064,
+      "step": 15140,
+      "torque_loss": 0.09954670071601868
+    },
+    {
+      "epoch": 13.615107913669064,
+      "force_loss": 0.003781534731388092,
+      "step": 15140
+    },
+    {
+      "epoch": 13.62410071942446,
+      "grad_norm": 0.3433564603328705,
+      "learning_rate": 8.920349489728928e-05,
+      "loss": 0.0242,
+      "step": 15150
+    },
+    {
+      "action_loss": 0.0068816691637039185,
+      "epoch": 13.62410071942446,
+      "step": 15150
+    },
+    {
+      "epoch": 13.62410071942446,
+      "step": 15150,
+      "torque_loss": 0.16625277698040009
+    },
+    {
+      "epoch": 13.62410071942446,
+      "force_loss": 0.006247811019420624,
+      "step": 15150
+    },
+    {
+      "epoch": 13.633093525179856,
+      "grad_norm": 0.8180539011955261,
+      "learning_rate": 8.918638455528646e-05,
+      "loss": 0.0315,
+      "step": 15160
+    },
+    {
+      "action_loss": 0.006157733500003815,
+      "epoch": 13.633093525179856,
+      "step": 15160
+    },
+    {
+      "epoch": 13.633093525179856,
+      "step": 15160,
+      "torque_loss": 0.12670379877090454
+    },
+    {
+      "epoch": 13.633093525179856,
+      "force_loss": 0.01235104352235794,
+      "step": 15160
+    },
+    {
+      "epoch": 13.642086330935252,
+      "grad_norm": 0.24847963452339172,
+      "learning_rate": 8.916926230949468e-05,
+      "loss": 0.0262,
+      "step": 15170
+    },
+    {
+      "action_loss": 0.007512775715440512,
+      "epoch": 13.642086330935252,
+      "step": 15170
+    },
+    {
+      "epoch": 13.642086330935252,
+      "step": 15170,
+      "torque_loss": 0.13621143996715546
+    },
+    {
+      "epoch": 13.642086330935252,
+      "force_loss": 0.0039144582115113735,
+      "step": 15170
+    },
+    {
+      "epoch": 13.651079136690647,
+      "grad_norm": 0.6130836009979248,
+      "learning_rate": 8.915212816511522e-05,
+      "loss": 0.0238,
+      "step": 15180
+    },
+    {
+      "action_loss": 0.006106710527092218,
+      "epoch": 13.651079136690647,
+      "step": 15180
+    },
+    {
+      "epoch": 13.651079136690647,
+      "step": 15180,
+      "torque_loss": 0.1378335952758789
+    },
+    {
+      "epoch": 13.651079136690647,
+      "force_loss": 0.004762370139360428,
+      "step": 15180
+    },
+    {
+      "epoch": 13.660071942446043,
+      "grad_norm": 0.513279914855957,
+      "learning_rate": 8.913498212735296e-05,
+      "loss": 0.0266,
+      "step": 15190
+    },
+    {
+      "action_loss": 0.011448142118752003,
+      "epoch": 13.660071942446043,
+      "step": 15190
+    },
+    {
+      "epoch": 13.660071942446043,
+      "step": 15190,
+      "torque_loss": 0.17824681103229523
+    },
+    {
+      "epoch": 13.660071942446043,
+      "force_loss": 0.022146375849843025,
+      "step": 15190
+    },
+    {
+      "epoch": 13.66906474820144,
+      "grad_norm": 0.5906609296798706,
+      "learning_rate": 8.911782420141643e-05,
+      "loss": 0.0266,
+      "step": 15200
+    },
+    {
+      "action_loss": 0.005149825010448694,
+      "epoch": 13.66906474820144,
+      "step": 15200
+    },
+    {
+      "epoch": 13.66906474820144,
+      "step": 15200,
+      "torque_loss": 0.0891626700758934
+    },
+    {
+      "epoch": 13.66906474820144,
+      "force_loss": 0.004900491330772638,
+      "step": 15200
+    },
+    {
+      "epoch": 13.678057553956835,
+      "grad_norm": 0.48827168345451355,
+      "learning_rate": 8.910065439251775e-05,
+      "loss": 0.0212,
+      "step": 15210
+    },
+    {
+      "action_loss": 0.00458289822563529,
+      "epoch": 13.678057553956835,
+      "step": 15210
+    },
+    {
+      "epoch": 13.678057553956835,
+      "step": 15210,
+      "torque_loss": 0.10856308788061142
+    },
+    {
+      "epoch": 13.678057553956835,
+      "force_loss": 0.004713992588222027,
+      "step": 15210
+    },
+    {
+      "epoch": 13.68705035971223,
+      "grad_norm": 0.9654158353805542,
+      "learning_rate": 8.908347270587268e-05,
+      "loss": 0.0225,
+      "step": 15220
+    },
+    {
+      "action_loss": 0.005796946119517088,
+      "epoch": 13.68705035971223,
+      "step": 15220
+    },
+    {
+      "epoch": 13.68705035971223,
+      "step": 15220,
+      "torque_loss": 0.11669004708528519
+    },
+    {
+      "epoch": 13.68705035971223,
+      "force_loss": 0.011185649782419205,
+      "step": 15220
+    },
+    {
+      "epoch": 13.696043165467627,
+      "grad_norm": 0.463498592376709,
+      "learning_rate": 8.906627914670054e-05,
+      "loss": 0.0247,
+      "step": 15230
+    },
+    {
+      "action_loss": 0.004784261807799339,
+      "epoch": 13.696043165467627,
+      "step": 15230
+    },
+    {
+      "epoch": 13.696043165467627,
+      "step": 15230,
+      "torque_loss": 0.16224533319473267
+    },
+    {
+      "epoch": 13.696043165467627,
+      "force_loss": 0.004794577602297068,
+      "step": 15230
+    },
+    {
+      "epoch": 13.70503597122302,
+      "grad_norm": 0.2059027999639511,
+      "learning_rate": 8.904907372022427e-05,
+      "loss": 0.0229,
+      "step": 15240
+    },
+    {
+      "action_loss": 0.006131669040769339,
+      "epoch": 13.70503597122302,
+      "step": 15240
+    },
+    {
+      "epoch": 13.70503597122302,
+      "step": 15240,
+      "torque_loss": 0.12671232223510742
+    },
+    {
+      "epoch": 13.70503597122302,
+      "force_loss": 0.009363076649606228,
+      "step": 15240
+    },
+    {
+      "epoch": 13.714028776978417,
+      "grad_norm": 0.612990140914917,
+      "learning_rate": 8.903185643167042e-05,
+      "loss": 0.0274,
+      "step": 15250
+    },
+    {
+      "action_loss": 0.0032046486157923937,
+      "epoch": 13.714028776978417,
+      "step": 15250
+    },
+    {
+      "epoch": 13.714028776978417,
+      "step": 15250,
+      "torque_loss": 0.053358182311058044
+    },
+    {
+      "epoch": 13.714028776978417,
+      "force_loss": 0.005081131588667631,
+      "step": 15250
+    },
+    {
+      "epoch": 13.723021582733812,
+      "grad_norm": 0.304829478263855,
+      "learning_rate": 8.901462728626919e-05,
+      "loss": 0.0221,
+      "step": 15260
+    },
+    {
+      "action_loss": 0.018701009452342987,
+      "epoch": 13.723021582733812,
+      "step": 15260
+    },
+    {
+      "epoch": 13.723021582733812,
+      "step": 15260,
+      "torque_loss": 0.14688244462013245
+    },
+    {
+      "epoch": 13.723021582733812,
+      "force_loss": 0.027699897065758705,
+      "step": 15260
+    },
+    {
+      "epoch": 13.732014388489208,
+      "grad_norm": 0.704147458076477,
+      "learning_rate": 8.899738628925429e-05,
+      "loss": 0.0275,
+      "step": 15270
+    },
+    {
+      "action_loss": 0.005055636633187532,
+      "epoch": 13.732014388489208,
+      "step": 15270
+    },
+    {
+      "epoch": 13.732014388489208,
+      "step": 15270,
+      "torque_loss": 0.14347197115421295
+    },
+    {
+      "epoch": 13.732014388489208,
+      "force_loss": 0.005784823093563318,
+      "step": 15270
+    },
+    {
+      "epoch": 13.741007194244604,
+      "grad_norm": 0.33921340107917786,
+      "learning_rate": 8.898013344586312e-05,
+      "loss": 0.0214,
+      "step": 15280
+    },
+    {
+      "action_loss": 0.03477925434708595,
+      "epoch": 13.741007194244604,
+      "step": 15280
+    },
+    {
+      "epoch": 13.741007194244604,
+      "step": 15280,
+      "torque_loss": 0.15228581428527832
+    },
+    {
+      "epoch": 13.741007194244604,
+      "force_loss": 0.04001857712864876,
+      "step": 15280
+    },
+    {
+      "epoch": 13.75,
+      "grad_norm": 0.4462069272994995,
+      "learning_rate": 8.896286876133661e-05,
+      "loss": 0.0282,
+      "step": 15290
+    },
+    {
+      "action_loss": 0.004876232240349054,
+      "epoch": 13.75,
+      "step": 15290
+    },
+    {
+      "epoch": 13.75,
+      "step": 15290,
+      "torque_loss": 0.13058489561080933
+    },
+    {
+      "epoch": 13.75,
+      "force_loss": 0.006025476381182671,
+      "step": 15290
+    },
+    {
+      "epoch": 13.758992805755396,
+      "grad_norm": 0.7754341959953308,
+      "learning_rate": 8.894559224091933e-05,
+      "loss": 0.0238,
+      "step": 15300
+    },
+    {
+      "action_loss": 0.018312720581889153,
+      "epoch": 13.758992805755396,
+      "step": 15300
+    },
+    {
+      "epoch": 13.758992805755396,
+      "step": 15300,
+      "torque_loss": 0.15676982700824738
+    },
+    {
+      "epoch": 13.758992805755396,
+      "force_loss": 0.02750256098806858,
+      "step": 15300
+    },
+    {
+      "epoch": 13.767985611510792,
+      "grad_norm": 0.41330650448799133,
+      "learning_rate": 8.892830388985942e-05,
+      "loss": 0.0264,
+      "step": 15310
+    },
+    {
+      "action_loss": 0.00818963348865509,
+      "epoch": 13.767985611510792,
+      "step": 15310
+    },
+    {
+      "epoch": 13.767985611510792,
+      "step": 15310,
+      "torque_loss": 0.18316282331943512
+    },
+    {
+      "epoch": 13.767985611510792,
+      "force_loss": 0.009668399579823017,
+      "step": 15310
+    },
+    {
+      "epoch": 13.776978417266188,
+      "grad_norm": 0.4735620319843292,
+      "learning_rate": 8.891100371340864e-05,
+      "loss": 0.0235,
+      "step": 15320
+    },
+    {
+      "action_loss": 0.01410482544451952,
+      "epoch": 13.776978417266188,
+      "step": 15320
+    },
+    {
+      "epoch": 13.776978417266188,
+      "step": 15320,
+      "torque_loss": 0.11079797893762589
+    },
+    {
+      "epoch": 13.776978417266188,
+      "force_loss": 0.008774803020060062,
+      "step": 15320
+    },
+    {
+      "epoch": 13.785971223021583,
+      "grad_norm": 0.556950569152832,
+      "learning_rate": 8.889369171682231e-05,
+      "loss": 0.0279,
+      "step": 15330
+    },
+    {
+      "action_loss": 0.006459081545472145,
+      "epoch": 13.785971223021583,
+      "step": 15330
+    },
+    {
+      "epoch": 13.785971223021583,
+      "step": 15330,
+      "torque_loss": 0.18019777536392212
+    },
+    {
+      "epoch": 13.785971223021583,
+      "force_loss": 0.006488729268312454,
+      "step": 15330
+    },
+    {
+      "epoch": 13.79496402877698,
+      "grad_norm": 0.5908033847808838,
+      "learning_rate": 8.887636790535936e-05,
+      "loss": 0.0226,
+      "step": 15340
+    },
+    {
+      "action_loss": 0.005060199648141861,
+      "epoch": 13.79496402877698,
+      "step": 15340
+    },
+    {
+      "epoch": 13.79496402877698,
+      "step": 15340,
+      "torque_loss": 0.1453528255224228
+    },
+    {
+      "epoch": 13.79496402877698,
+      "force_loss": 0.004749844316393137,
+      "step": 15340
+    },
+    {
+      "epoch": 13.803956834532373,
+      "grad_norm": 0.2986871004104614,
+      "learning_rate": 8.885903228428231e-05,
+      "loss": 0.0227,
+      "step": 15350
+    },
+    {
+      "action_loss": 0.01119461003690958,
+      "epoch": 13.803956834532373,
+      "step": 15350
+    },
+    {
+      "epoch": 13.803956834532373,
+      "step": 15350,
+      "torque_loss": 0.15257616341114044
+    },
+    {
+      "epoch": 13.803956834532373,
+      "force_loss": 0.009376834146678448,
+      "step": 15350
+    },
+    {
+      "epoch": 13.81294964028777,
+      "grad_norm": 0.5822528004646301,
+      "learning_rate": 8.884168485885727e-05,
+      "loss": 0.0228,
+      "step": 15360
+    },
+    {
+      "action_loss": 0.005400268826633692,
+      "epoch": 13.81294964028777,
+      "step": 15360
+    },
+    {
+      "epoch": 13.81294964028777,
+      "step": 15360,
+      "torque_loss": 0.1534426063299179
+    },
+    {
+      "epoch": 13.81294964028777,
+      "force_loss": 0.0056499093770980835,
+      "step": 15360
+    },
+    {
+      "epoch": 13.821942446043165,
+      "grad_norm": 0.3519933521747589,
+      "learning_rate": 8.882432563435393e-05,
+      "loss": 0.0255,
+      "step": 15370
+    },
+    {
+      "action_loss": 0.008613456971943378,
+      "epoch": 13.821942446043165,
+      "step": 15370
+    },
+    {
+      "epoch": 13.821942446043165,
+      "step": 15370,
+      "torque_loss": 0.16823451220989227
+    },
+    {
+      "epoch": 13.821942446043165,
+      "force_loss": 0.009782267734408379,
+      "step": 15370
+    },
+    {
+      "epoch": 13.83093525179856,
+      "grad_norm": 0.5148043036460876,
+      "learning_rate": 8.880695461604556e-05,
+      "loss": 0.0223,
+      "step": 15380
+    },
+    {
+      "action_loss": 0.0024448649492114782,
+      "epoch": 13.83093525179856,
+      "step": 15380
+    },
+    {
+      "epoch": 13.83093525179856,
+      "step": 15380,
+      "torque_loss": 0.13449786603450775
+    },
+    {
+      "epoch": 13.83093525179856,
+      "force_loss": 0.004522038623690605,
+      "step": 15380
+    },
+    {
+      "epoch": 13.839928057553957,
+      "grad_norm": 0.6680033802986145,
+      "learning_rate": 8.878957180920901e-05,
+      "loss": 0.023,
+      "step": 15390
+    },
+    {
+      "action_loss": 0.004556481260806322,
+      "epoch": 13.839928057553957,
+      "step": 15390
+    },
+    {
+      "epoch": 13.839928057553957,
+      "step": 15390,
+      "torque_loss": 0.13199438154697418
+    },
+    {
+      "epoch": 13.839928057553957,
+      "force_loss": 0.005707425530999899,
+      "step": 15390
+    },
+    {
+      "epoch": 13.848920863309353,
+      "grad_norm": 0.41075780987739563,
+      "learning_rate": 8.877217721912473e-05,
+      "loss": 0.0234,
+      "step": 15400
+    },
+    {
+      "action_loss": 0.005182987544685602,
+      "epoch": 13.848920863309353,
+      "step": 15400
+    },
+    {
+      "epoch": 13.848920863309353,
+      "step": 15400,
+      "torque_loss": 0.11845013499259949
+    },
+    {
+      "epoch": 13.848920863309353,
+      "force_loss": 0.006118921097368002,
+      "step": 15400
+    },
+    {
+      "epoch": 13.857913669064748,
+      "grad_norm": 0.420786052942276,
+      "learning_rate": 8.875477085107673e-05,
+      "loss": 0.0238,
+      "step": 15410
+    },
+    {
+      "action_loss": 0.0063112773932516575,
+      "epoch": 13.857913669064748,
+      "step": 15410
+    },
+    {
+      "epoch": 13.857913669064748,
+      "step": 15410,
+      "torque_loss": 0.12901552021503448
+    },
+    {
+      "epoch": 13.857913669064748,
+      "force_loss": 0.0066695548593997955,
+      "step": 15410
+    },
+    {
+      "epoch": 13.866906474820144,
+      "grad_norm": 0.4857789874076843,
+      "learning_rate": 8.87373527103526e-05,
+      "loss": 0.0263,
+      "step": 15420
+    },
+    {
+      "action_loss": 0.0046065147034823895,
+      "epoch": 13.866906474820144,
+      "step": 15420
+    },
+    {
+      "epoch": 13.866906474820144,
+      "step": 15420,
+      "torque_loss": 0.10553174465894699
+    },
+    {
+      "epoch": 13.866906474820144,
+      "force_loss": 0.005900380667299032,
+      "step": 15420
+    },
+    {
+      "epoch": 13.87589928057554,
+      "grad_norm": 0.5139258503913879,
+      "learning_rate": 8.871992280224353e-05,
+      "loss": 0.0236,
+      "step": 15430
+    },
+    {
+      "action_loss": 0.006548629608005285,
+      "epoch": 13.87589928057554,
+      "step": 15430
+    },
+    {
+      "epoch": 13.87589928057554,
+      "step": 15430,
+      "torque_loss": 0.14469869434833527
+    },
+    {
+      "epoch": 13.87589928057554,
+      "force_loss": 0.008571171201765537,
+      "step": 15430
+    },
+    {
+      "epoch": 13.884892086330936,
+      "grad_norm": 0.4033961296081543,
+      "learning_rate": 8.870248113204422e-05,
+      "loss": 0.0229,
+      "step": 15440
+    },
+    {
+      "action_loss": 0.0070077404379844666,
+      "epoch": 13.884892086330936,
+      "step": 15440
+    },
+    {
+      "epoch": 13.884892086330936,
+      "step": 15440,
+      "torque_loss": 0.12000685185194016
+    },
+    {
+      "epoch": 13.884892086330936,
+      "force_loss": 0.007769885007292032,
+      "step": 15440
+    },
+    {
+      "epoch": 13.89388489208633,
+      "grad_norm": 0.9563060998916626,
+      "learning_rate": 8.868502770505306e-05,
+      "loss": 0.0259,
+      "step": 15450
+    },
+    {
+      "action_loss": 0.011506802402436733,
+      "epoch": 13.89388489208633,
+      "step": 15450
+    },
+    {
+      "epoch": 13.89388489208633,
+      "step": 15450,
+      "torque_loss": 0.13790783286094666
+    },
+    {
+      "epoch": 13.89388489208633,
+      "force_loss": 0.008600750006735325,
+      "step": 15450
+    },
+    {
+      "epoch": 13.902877697841726,
+      "grad_norm": 0.46110257506370544,
+      "learning_rate": 8.86675625265719e-05,
+      "loss": 0.029,
+      "step": 15460
+    },
+    {
+      "action_loss": 0.00926098134368658,
+      "epoch": 13.902877697841726,
+      "step": 15460
+    },
+    {
+      "epoch": 13.902877697841726,
+      "step": 15460,
+      "torque_loss": 0.17907463014125824
+    },
+    {
+      "epoch": 13.902877697841726,
+      "force_loss": 0.008339088410139084,
+      "step": 15460
+    },
+    {
+      "epoch": 13.911870503597122,
+      "grad_norm": 0.3648589551448822,
+      "learning_rate": 8.865008560190618e-05,
+      "loss": 0.0227,
+      "step": 15470
+    },
+    {
+      "action_loss": 0.03231241554021835,
+      "epoch": 13.911870503597122,
+      "step": 15470
+    },
+    {
+      "epoch": 13.911870503597122,
+      "step": 15470,
+      "torque_loss": 0.19119863212108612
+    },
+    {
+      "epoch": 13.911870503597122,
+      "force_loss": 0.03968670964241028,
+      "step": 15470
+    },
+    {
+      "epoch": 13.920863309352518,
+      "grad_norm": 0.29791513085365295,
+      "learning_rate": 8.863259693636496e-05,
+      "loss": 0.0261,
+      "step": 15480
+    },
+    {
+      "action_loss": 0.012710523791611195,
+      "epoch": 13.920863309352518,
+      "step": 15480
+    },
+    {
+      "epoch": 13.920863309352518,
+      "step": 15480,
+      "torque_loss": 0.13339431583881378
+    },
+    {
+      "epoch": 13.920863309352518,
+      "force_loss": 0.012615206651389599,
+      "step": 15480
+    },
+    {
+      "epoch": 13.929856115107913,
+      "grad_norm": 0.44263288378715515,
+      "learning_rate": 8.861509653526083e-05,
+      "loss": 0.0241,
+      "step": 15490
+    },
+    {
+      "action_loss": 0.01084295567125082,
+      "epoch": 13.929856115107913,
+      "step": 15490
+    },
+    {
+      "epoch": 13.929856115107913,
+      "step": 15490,
+      "torque_loss": 0.19021831452846527
+    },
+    {
+      "epoch": 13.929856115107913,
+      "force_loss": 0.013560633175075054,
+      "step": 15490
+    },
+    {
+      "epoch": 13.93884892086331,
+      "grad_norm": 0.20355476438999176,
+      "learning_rate": 8.859758440390993e-05,
+      "loss": 0.0248,
+      "step": 15500
+    },
+    {
+      "action_loss": 0.005911046639084816,
+      "epoch": 13.93884892086331,
+      "step": 15500
+    },
+    {
+      "epoch": 13.93884892086331,
+      "step": 15500,
+      "torque_loss": 0.13253700733184814
+    },
+    {
+      "epoch": 13.93884892086331,
+      "force_loss": 0.011428029276430607,
+      "step": 15500
+    },
+    {
+      "epoch": 13.947841726618705,
+      "grad_norm": 0.4832836091518402,
+      "learning_rate": 8.858006054763202e-05,
+      "loss": 0.0213,
+      "step": 15510
+    },
+    {
+      "action_loss": 0.007079319562762976,
+      "epoch": 13.947841726618705,
+      "step": 15510
+    },
+    {
+      "epoch": 13.947841726618705,
+      "step": 15510,
+      "torque_loss": 0.11854541301727295
+    },
+    {
+      "epoch": 13.947841726618705,
+      "force_loss": 0.009166515432298183,
+      "step": 15510
+    },
+    {
+      "epoch": 13.956834532374101,
+      "grad_norm": 0.36489251255989075,
+      "learning_rate": 8.856252497175035e-05,
+      "loss": 0.025,
+      "step": 15520
+    },
+    {
+      "action_loss": 0.005048893857747316,
+      "epoch": 13.956834532374101,
+      "step": 15520
+    },
+    {
+      "epoch": 13.956834532374101,
+      "step": 15520,
+      "torque_loss": 0.1286717802286148
+    },
+    {
+      "epoch": 13.956834532374101,
+      "force_loss": 0.004200175404548645,
+      "step": 15520
+    },
+    {
+      "epoch": 13.965827338129497,
+      "grad_norm": 0.2418166548013687,
+      "learning_rate": 8.854497768159178e-05,
+      "loss": 0.0196,
+      "step": 15530
+    },
+    {
+      "action_loss": 0.012202930636703968,
+      "epoch": 13.965827338129497,
+      "step": 15530
+    },
+    {
+      "epoch": 13.965827338129497,
+      "step": 15530,
+      "torque_loss": 0.1705620288848877
+    },
+    {
+      "epoch": 13.965827338129497,
+      "force_loss": 0.007922484539449215,
+      "step": 15530
+    },
+    {
+      "epoch": 13.974820143884893,
+      "grad_norm": 0.5926933288574219,
+      "learning_rate": 8.852741868248671e-05,
+      "loss": 0.0229,
+      "step": 15540
+    },
+    {
+      "action_loss": 0.010377899743616581,
+      "epoch": 13.974820143884893,
+      "step": 15540
+    },
+    {
+      "epoch": 13.974820143884893,
+      "step": 15540,
+      "torque_loss": 0.16820472478866577
+    },
+    {
+      "epoch": 13.974820143884893,
+      "force_loss": 0.009452663362026215,
+      "step": 15540
+    },
+    {
+      "epoch": 13.983812949640289,
+      "grad_norm": 0.9483239054679871,
+      "learning_rate": 8.85098479797691e-05,
+      "loss": 0.0263,
+      "step": 15550
+    },
+    {
+      "action_loss": 0.011345763690769672,
+      "epoch": 13.983812949640289,
+      "step": 15550
+    },
+    {
+      "epoch": 13.983812949640289,
+      "step": 15550,
+      "torque_loss": 0.17303144931793213
+    },
+    {
+      "epoch": 13.983812949640289,
+      "force_loss": 0.005850790534168482,
+      "step": 15550
+    },
+    {
+      "epoch": 13.992805755395683,
+      "grad_norm": 0.4478403329849243,
+      "learning_rate": 8.849226557877646e-05,
+      "loss": 0.0249,
+      "step": 15560
+    },
+    {
+      "action_loss": 0.007818459533154964,
+      "epoch": 13.992805755395683,
+      "step": 15560
+    },
+    {
+      "epoch": 13.992805755395683,
+      "step": 15560,
+      "torque_loss": 0.12763364613056183
+    },
+    {
+      "epoch": 13.992805755395683,
+      "force_loss": 0.007843210361897945,
+      "step": 15560
+    },
+    {
+      "epoch": 14.001798561151078,
+      "grad_norm": 0.8684555888175964,
+      "learning_rate": 8.84746714848499e-05,
+      "loss": 0.0252,
+      "step": 15570
+    },
+    {
+      "action_loss": 0.004722162615507841,
+      "epoch": 14.001798561151078,
+      "step": 15570
+    },
+    {
+      "epoch": 14.001798561151078,
+      "step": 15570,
+      "torque_loss": 0.13987940549850464
+    },
+    {
+      "epoch": 14.001798561151078,
+      "force_loss": 0.007056157570332289,
+      "step": 15570
+    },
+    {
+      "epoch": 14.010791366906474,
+      "grad_norm": 0.23410508036613464,
+      "learning_rate": 8.845706570333397e-05,
+      "loss": 0.0219,
+      "step": 15580
+    },
+    {
+      "action_loss": 0.007552033755928278,
+      "epoch": 14.010791366906474,
+      "step": 15580
+    },
+    {
+      "epoch": 14.010791366906474,
+      "step": 15580,
+      "torque_loss": 0.13435308635234833
+    },
+    {
+      "epoch": 14.010791366906474,
+      "force_loss": 0.010103273205459118,
+      "step": 15580
+    },
+    {
+      "epoch": 14.01978417266187,
+      "grad_norm": 0.2973777651786804,
+      "learning_rate": 8.84394482395769e-05,
+      "loss": 0.0289,
+      "step": 15590
+    },
+    {
+      "action_loss": 0.005697654094547033,
+      "epoch": 14.01978417266187,
+      "step": 15590
+    },
+    {
+      "epoch": 14.01978417266187,
+      "step": 15590,
+      "torque_loss": 0.13407254219055176
+    },
+    {
+      "epoch": 14.01978417266187,
+      "force_loss": 0.012538072653114796,
+      "step": 15590
+    },
+    {
+      "epoch": 14.028776978417266,
+      "grad_norm": 0.5000838041305542,
+      "learning_rate": 8.842181909893038e-05,
+      "loss": 0.0193,
+      "step": 15600
+    },
+    {
+      "action_loss": 0.009854399599134922,
+      "epoch": 14.028776978417266,
+      "step": 15600
+    },
+    {
+      "epoch": 14.028776978417266,
+      "step": 15600,
+      "torque_loss": 0.14689792692661285
+    },
+    {
+      "epoch": 14.028776978417266,
+      "force_loss": 0.011832110583782196,
+      "step": 15600
+    },
+    {
+      "epoch": 14.037769784172662,
+      "grad_norm": 0.21835018694400787,
+      "learning_rate": 8.840417828674969e-05,
+      "loss": 0.0239,
+      "step": 15610
+    },
+    {
+      "action_loss": 0.006035699043422937,
+      "epoch": 14.037769784172662,
+      "step": 15610
+    },
+    {
+      "epoch": 14.037769784172662,
+      "step": 15610,
+      "torque_loss": 0.12497264891862869
+    },
+    {
+      "epoch": 14.037769784172662,
+      "force_loss": 0.008065945468842983,
+      "step": 15610
+    },
+    {
+      "epoch": 14.046762589928058,
+      "grad_norm": 0.32977935671806335,
+      "learning_rate": 8.838652580839364e-05,
+      "loss": 0.0261,
+      "step": 15620
+    },
+    {
+      "action_loss": 0.008962932042777538,
+      "epoch": 14.046762589928058,
+      "step": 15620
+    },
+    {
+      "epoch": 14.046762589928058,
+      "step": 15620,
+      "torque_loss": 0.14133070409297943
+    },
+    {
+      "epoch": 14.046762589928058,
+      "force_loss": 0.008453644812107086,
+      "step": 15620
+    },
+    {
+      "epoch": 14.055755395683454,
+      "grad_norm": 0.5109555125236511,
+      "learning_rate": 8.836886166922458e-05,
+      "loss": 0.0322,
+      "step": 15630
+    },
+    {
+      "action_loss": 0.0034796353429555893,
+      "epoch": 14.055755395683454,
+      "step": 15630
+    },
+    {
+      "epoch": 14.055755395683454,
+      "step": 15630,
+      "torque_loss": 0.07112327218055725
+    },
+    {
+      "epoch": 14.055755395683454,
+      "force_loss": 0.0038140921387821436,
+      "step": 15630
+    },
+    {
+      "epoch": 14.06474820143885,
+      "grad_norm": 0.34478431940078735,
+      "learning_rate": 8.835118587460844e-05,
+      "loss": 0.0232,
+      "step": 15640
+    },
+    {
+      "action_loss": 0.009331970475614071,
+      "epoch": 14.06474820143885,
+      "step": 15640
+    },
+    {
+      "epoch": 14.06474820143885,
+      "step": 15640,
+      "torque_loss": 0.1184937134385109
+    },
+    {
+      "epoch": 14.06474820143885,
+      "force_loss": 0.014365400187671185,
+      "step": 15640
+    },
+    {
+      "epoch": 14.073741007194245,
+      "grad_norm": 0.7019870281219482,
+      "learning_rate": 8.83334984299146e-05,
+      "loss": 0.031,
+      "step": 15650
+    },
+    {
+      "action_loss": 0.009050595574080944,
+      "epoch": 14.073741007194245,
+      "step": 15650
+    },
+    {
+      "epoch": 14.073741007194245,
+      "step": 15650,
+      "torque_loss": 0.13808287680149078
+    },
+    {
+      "epoch": 14.073741007194245,
+      "force_loss": 0.008266493678092957,
+      "step": 15650
+    },
+    {
+      "epoch": 14.082733812949641,
+      "grad_norm": 0.691999077796936,
+      "learning_rate": 8.83157993405161e-05,
+      "loss": 0.0315,
+      "step": 15660
+    },
+    {
+      "action_loss": 0.004204514902085066,
+      "epoch": 14.082733812949641,
+      "step": 15660
+    },
+    {
+      "epoch": 14.082733812949641,
+      "step": 15660,
+      "torque_loss": 0.15880849957466125
+    },
+    {
+      "epoch": 14.082733812949641,
+      "force_loss": 0.004555519670248032,
+      "step": 15660
+    },
+    {
+      "epoch": 14.091726618705035,
+      "grad_norm": 0.6082274317741394,
+      "learning_rate": 8.829808861178943e-05,
+      "loss": 0.0248,
+      "step": 15670
+    },
+    {
+      "action_loss": 0.0051100715063512325,
+      "epoch": 14.091726618705035,
+      "step": 15670
+    },
+    {
+      "epoch": 14.091726618705035,
+      "step": 15670,
+      "torque_loss": 0.08284465223550797
+    },
+    {
+      "epoch": 14.091726618705035,
+      "force_loss": 0.011337089352309704,
+      "step": 15670
+    },
+    {
+      "epoch": 14.100719424460431,
+      "grad_norm": 0.4260847568511963,
+      "learning_rate": 8.828036624911464e-05,
+      "loss": 0.0271,
+      "step": 15680
+    },
+    {
+      "action_loss": 0.007330385968089104,
+      "epoch": 14.100719424460431,
+      "step": 15680
+    },
+    {
+      "epoch": 14.100719424460431,
+      "step": 15680,
+      "torque_loss": 0.1403549164533615
+    },
+    {
+      "epoch": 14.100719424460431,
+      "force_loss": 0.009846248663961887,
+      "step": 15680
+    },
+    {
+      "epoch": 14.109712230215827,
+      "grad_norm": 0.3712562918663025,
+      "learning_rate": 8.826263225787532e-05,
+      "loss": 0.0247,
+      "step": 15690
+    },
+    {
+      "action_loss": 0.004076914396136999,
+      "epoch": 14.109712230215827,
+      "step": 15690
+    },
+    {
+      "epoch": 14.109712230215827,
+      "step": 15690,
+      "torque_loss": 0.17920346558094025
+    },
+    {
+      "epoch": 14.109712230215827,
+      "force_loss": 0.004728124476969242,
+      "step": 15690
+    },
+    {
+      "epoch": 14.118705035971223,
+      "grad_norm": 0.487091988325119,
+      "learning_rate": 8.824488664345858e-05,
+      "loss": 0.022,
+      "step": 15700
+    },
+    {
+      "action_loss": 0.020175985991954803,
+      "epoch": 14.118705035971223,
+      "step": 15700
+    },
+    {
+      "epoch": 14.118705035971223,
+      "step": 15700,
+      "torque_loss": 0.1243928000330925
+    },
+    {
+      "epoch": 14.118705035971223,
+      "force_loss": 0.01809723861515522,
+      "step": 15700
+    },
+    {
+      "epoch": 14.127697841726619,
+      "grad_norm": 0.5377398729324341,
+      "learning_rate": 8.822712941125508e-05,
+      "loss": 0.0222,
+      "step": 15710
+    },
+    {
+      "action_loss": 0.00976415816694498,
+      "epoch": 14.127697841726619,
+      "step": 15710
+    },
+    {
+      "epoch": 14.127697841726619,
+      "step": 15710,
+      "torque_loss": 0.16083715856075287
+    },
+    {
+      "epoch": 14.127697841726619,
+      "force_loss": 0.014677882194519043,
+      "step": 15710
+    },
+    {
+      "epoch": 14.136690647482014,
+      "grad_norm": 0.7602187991142273,
+      "learning_rate": 8.820936056665898e-05,
+      "loss": 0.0252,
+      "step": 15720
+    },
+    {
+      "action_loss": 0.005407609045505524,
+      "epoch": 14.136690647482014,
+      "step": 15720
+    },
+    {
+      "epoch": 14.136690647482014,
+      "step": 15720,
+      "torque_loss": 0.1260979324579239
+    },
+    {
+      "epoch": 14.136690647482014,
+      "force_loss": 0.003310616360977292,
+      "step": 15720
+    },
+    {
+      "epoch": 14.14568345323741,
+      "grad_norm": 0.2339577078819275,
+      "learning_rate": 8.819158011506801e-05,
+      "loss": 0.0246,
+      "step": 15730
+    },
+    {
+      "action_loss": 0.004923509433865547,
+      "epoch": 14.14568345323741,
+      "step": 15730
+    },
+    {
+      "epoch": 14.14568345323741,
+      "step": 15730,
+      "torque_loss": 0.09556185454130173
+    },
+    {
+      "epoch": 14.14568345323741,
+      "force_loss": 0.00431869737803936,
+      "step": 15730
+    },
+    {
+      "epoch": 14.154676258992806,
+      "grad_norm": 0.2566169202327728,
+      "learning_rate": 8.81737880618834e-05,
+      "loss": 0.0229,
+      "step": 15740
+    },
+    {
+      "action_loss": 0.007789431139826775,
+      "epoch": 14.154676258992806,
+      "step": 15740
+    },
+    {
+      "epoch": 14.154676258992806,
+      "step": 15740,
+      "torque_loss": 0.15237168967723846
+    },
+    {
+      "epoch": 14.154676258992806,
+      "force_loss": 0.009500320069491863,
+      "step": 15740
+    },
+    {
+      "epoch": 14.163669064748202,
+      "grad_norm": 0.24414263665676117,
+      "learning_rate": 8.815598441250987e-05,
+      "loss": 0.0269,
+      "step": 15750
+    },
+    {
+      "action_loss": 0.0037577971816062927,
+      "epoch": 14.163669064748202,
+      "step": 15750
+    },
+    {
+      "epoch": 14.163669064748202,
+      "step": 15750,
+      "torque_loss": 0.11859043687582016
+    },
+    {
+      "epoch": 14.163669064748202,
+      "force_loss": 0.0034807927440851927,
+      "step": 15750
+    },
+    {
+      "epoch": 14.172661870503598,
+      "grad_norm": 0.3403555750846863,
+      "learning_rate": 8.813816917235576e-05,
+      "loss": 0.025,
+      "step": 15760
+    },
+    {
+      "action_loss": 0.014798753894865513,
+      "epoch": 14.172661870503598,
+      "step": 15760
+    },
+    {
+      "epoch": 14.172661870503598,
+      "step": 15760,
+      "torque_loss": 0.149000883102417
+    },
+    {
+      "epoch": 14.172661870503598,
+      "force_loss": 0.024646872654557228,
+      "step": 15760
+    },
+    {
+      "epoch": 14.181654676258994,
+      "grad_norm": 0.33511635661125183,
+      "learning_rate": 8.812034234683282e-05,
+      "loss": 0.0253,
+      "step": 15770
+    },
+    {
+      "action_loss": 0.009019267745316029,
+      "epoch": 14.181654676258994,
+      "step": 15770
+    },
+    {
+      "epoch": 14.181654676258994,
+      "step": 15770,
+      "torque_loss": 0.14723806083202362
+    },
+    {
+      "epoch": 14.181654676258994,
+      "force_loss": 0.01763060688972473,
+      "step": 15770
+    },
+    {
+      "epoch": 14.190647482014388,
+      "grad_norm": 0.5340002775192261,
+      "learning_rate": 8.810250394135637e-05,
+      "loss": 0.024,
+      "step": 15780
+    },
+    {
+      "action_loss": 0.01773165911436081,
+      "epoch": 14.190647482014388,
+      "step": 15780
+    },
+    {
+      "epoch": 14.190647482014388,
+      "step": 15780,
+      "torque_loss": 0.09388729184865952
+    },
+    {
+      "epoch": 14.190647482014388,
+      "force_loss": 0.022907018661499023,
+      "step": 15780
+    },
+    {
+      "epoch": 14.199640287769784,
+      "grad_norm": 0.5116387605667114,
+      "learning_rate": 8.808465396134529e-05,
+      "loss": 0.0266,
+      "step": 15790
+    },
+    {
+      "action_loss": 0.014380519278347492,
+      "epoch": 14.199640287769784,
+      "step": 15790
+    },
+    {
+      "epoch": 14.199640287769784,
+      "step": 15790,
+      "torque_loss": 0.11947671324014664
+    },
+    {
+      "epoch": 14.199640287769784,
+      "force_loss": 0.02432125248014927,
+      "step": 15790
+    },
+    {
+      "epoch": 14.20863309352518,
+      "grad_norm": 0.5784803032875061,
+      "learning_rate": 8.806679241222189e-05,
+      "loss": 0.0221,
+      "step": 15800
+    },
+    {
+      "action_loss": 0.007958601228892803,
+      "epoch": 14.20863309352518,
+      "step": 15800
+    },
+    {
+      "epoch": 14.20863309352518,
+      "step": 15800,
+      "torque_loss": 0.12594085931777954
+    },
+    {
+      "epoch": 14.20863309352518,
+      "force_loss": 0.00979639682918787,
+      "step": 15800
+    },
+    {
+      "epoch": 14.217625899280575,
+      "grad_norm": 0.6483083367347717,
+      "learning_rate": 8.804891929941203e-05,
+      "loss": 0.0288,
+      "step": 15810
+    },
+    {
+      "action_loss": 0.008797496557235718,
+      "epoch": 14.217625899280575,
+      "step": 15810
+    },
+    {
+      "epoch": 14.217625899280575,
+      "step": 15810,
+      "torque_loss": 0.1706986427307129
+    },
+    {
+      "epoch": 14.217625899280575,
+      "force_loss": 0.011068209074437618,
+      "step": 15810
+    },
+    {
+      "epoch": 14.226618705035971,
+      "grad_norm": 0.41875794529914856,
+      "learning_rate": 8.803103462834514e-05,
+      "loss": 0.0279,
+      "step": 15820
+    },
+    {
+      "action_loss": 0.00846771989017725,
+      "epoch": 14.226618705035971,
+      "step": 15820
+    },
+    {
+      "epoch": 14.226618705035971,
+      "step": 15820,
+      "torque_loss": 0.25680890679359436
+    },
+    {
+      "epoch": 14.226618705035971,
+      "force_loss": 0.004131515976041555,
+      "step": 15820
+    },
+    {
+      "epoch": 14.235611510791367,
+      "grad_norm": 0.4922666847705841,
+      "learning_rate": 8.801313840445408e-05,
+      "loss": 0.0255,
+      "step": 15830
+    },
+    {
+      "action_loss": 0.00815120991319418,
+      "epoch": 14.235611510791367,
+      "step": 15830
+    },
+    {
+      "epoch": 14.235611510791367,
+      "step": 15830,
+      "torque_loss": 0.10232289880514145
+    },
+    {
+      "epoch": 14.235611510791367,
+      "force_loss": 0.006058324594050646,
+      "step": 15830
+    },
+    {
+      "epoch": 14.244604316546763,
+      "grad_norm": 0.38295236229896545,
+      "learning_rate": 8.799523063317524e-05,
+      "loss": 0.0229,
+      "step": 15840
+    },
+    {
+      "action_loss": 0.00865989550948143,
+      "epoch": 14.244604316546763,
+      "step": 15840
+    },
+    {
+      "epoch": 14.244604316546763,
+      "step": 15840,
+      "torque_loss": 0.1718813180923462
+    },
+    {
+      "epoch": 14.244604316546763,
+      "force_loss": 0.006196597125381231,
+      "step": 15840
+    },
+    {
+      "epoch": 14.253597122302159,
+      "grad_norm": 0.29465678334236145,
+      "learning_rate": 8.797731131994854e-05,
+      "loss": 0.0227,
+      "step": 15850
+    },
+    {
+      "action_loss": 0.02060581184923649,
+      "epoch": 14.253597122302159,
+      "step": 15850
+    },
+    {
+      "epoch": 14.253597122302159,
+      "step": 15850,
+      "torque_loss": 0.15192629396915436
+    },
+    {
+      "epoch": 14.253597122302159,
+      "force_loss": 0.04282194375991821,
+      "step": 15850
+    },
+    {
+      "epoch": 14.262589928057555,
+      "grad_norm": 0.3826133906841278,
+      "learning_rate": 8.795938047021739e-05,
+      "loss": 0.0281,
+      "step": 15860
+    },
+    {
+      "action_loss": 0.006589151918888092,
+      "epoch": 14.262589928057555,
+      "step": 15860
+    },
+    {
+      "epoch": 14.262589928057555,
+      "step": 15860,
+      "torque_loss": 0.11835622787475586
+    },
+    {
+      "epoch": 14.262589928057555,
+      "force_loss": 0.0061923484317958355,
+      "step": 15860
+    },
+    {
+      "epoch": 14.27158273381295,
+      "grad_norm": 0.48784953355789185,
+      "learning_rate": 8.794143808942872e-05,
+      "loss": 0.0244,
+      "step": 15870
+    },
+    {
+      "action_loss": 0.010280195623636246,
+      "epoch": 14.27158273381295,
+      "step": 15870
+    },
+    {
+      "epoch": 14.27158273381295,
+      "step": 15870,
+      "torque_loss": 0.1272284835577011
+    },
+    {
+      "epoch": 14.27158273381295,
+      "force_loss": 0.0068427324295043945,
+      "step": 15870
+    },
+    {
+      "epoch": 14.280575539568344,
+      "grad_norm": 0.5823810696601868,
+      "learning_rate": 8.792348418303296e-05,
+      "loss": 0.0257,
+      "step": 15880
+    },
+    {
+      "action_loss": 0.028448795899748802,
+      "epoch": 14.280575539568344,
+      "step": 15880
+    },
+    {
+      "epoch": 14.280575539568344,
+      "step": 15880,
+      "torque_loss": 0.16948986053466797
+    },
+    {
+      "epoch": 14.280575539568344,
+      "force_loss": 0.02727888710796833,
+      "step": 15880
+    },
+    {
+      "epoch": 14.28956834532374,
+      "grad_norm": 0.7188531756401062,
+      "learning_rate": 8.790551875648398e-05,
+      "loss": 0.0299,
+      "step": 15890
+    },
+    {
+      "action_loss": 0.019151948392391205,
+      "epoch": 14.28956834532374,
+      "step": 15890
+    },
+    {
+      "epoch": 14.28956834532374,
+      "step": 15890,
+      "torque_loss": 0.12116354703903198
+    },
+    {
+      "epoch": 14.28956834532374,
+      "force_loss": 0.020395763218402863,
+      "step": 15890
+    },
+    {
+      "epoch": 14.298561151079136,
+      "grad_norm": 0.4301486015319824,
+      "learning_rate": 8.788754181523926e-05,
+      "loss": 0.0314,
+      "step": 15900
+    },
+    {
+      "action_loss": 0.03381485119462013,
+      "epoch": 14.298561151079136,
+      "step": 15900
+    },
+    {
+      "epoch": 14.298561151079136,
+      "step": 15900,
+      "torque_loss": 0.20251469314098358
+    },
+    {
+      "epoch": 14.298561151079136,
+      "force_loss": 0.049152523279190063,
+      "step": 15900
+    },
+    {
+      "epoch": 14.307553956834532,
+      "grad_norm": 0.4857807159423828,
+      "learning_rate": 8.78695533647597e-05,
+      "loss": 0.03,
+      "step": 15910
+    },
+    {
+      "action_loss": 0.01003540400415659,
+      "epoch": 14.307553956834532,
+      "step": 15910
+    },
+    {
+      "epoch": 14.307553956834532,
+      "step": 15910,
+      "torque_loss": 0.15890617668628693
+    },
+    {
+      "epoch": 14.307553956834532,
+      "force_loss": 0.008113347925245762,
+      "step": 15910
+    },
+    {
+      "epoch": 14.316546762589928,
+      "grad_norm": 0.5157390236854553,
+      "learning_rate": 8.785155341050972e-05,
+      "loss": 0.035,
+      "step": 15920
+    },
+    {
+      "action_loss": 0.012707020156085491,
+      "epoch": 14.316546762589928,
+      "step": 15920
+    },
+    {
+      "epoch": 14.316546762589928,
+      "step": 15920,
+      "torque_loss": 0.15442773699760437
+    },
+    {
+      "epoch": 14.316546762589928,
+      "force_loss": 0.009181453846395016,
+      "step": 15920
+    },
+    {
+      "epoch": 14.325539568345324,
+      "grad_norm": 0.2399837076663971,
+      "learning_rate": 8.783354195795721e-05,
+      "loss": 0.0263,
+      "step": 15930
+    },
+    {
+      "action_loss": 0.012498520314693451,
+      "epoch": 14.325539568345324,
+      "step": 15930
+    },
+    {
+      "epoch": 14.325539568345324,
+      "step": 15930,
+      "torque_loss": 0.11486140638589859
+    },
+    {
+      "epoch": 14.325539568345324,
+      "force_loss": 0.01330383587628603,
+      "step": 15930
+    },
+    {
+      "epoch": 14.33453237410072,
+      "grad_norm": 0.6676514744758606,
+      "learning_rate": 8.78155190125736e-05,
+      "loss": 0.0259,
+      "step": 15940
+    },
+    {
+      "action_loss": 0.00856395810842514,
+      "epoch": 14.33453237410072,
+      "step": 15940
+    },
+    {
+      "epoch": 14.33453237410072,
+      "step": 15940,
+      "torque_loss": 0.1485954225063324
+    },
+    {
+      "epoch": 14.33453237410072,
+      "force_loss": 0.010877604596316814,
+      "step": 15940
+    },
+    {
+      "epoch": 14.343525179856115,
+      "grad_norm": 0.45051026344299316,
+      "learning_rate": 8.779748457983378e-05,
+      "loss": 0.0308,
+      "step": 15950
+    },
+    {
+      "action_loss": 0.0058227889239788055,
+      "epoch": 14.343525179856115,
+      "step": 15950
+    },
+    {
+      "epoch": 14.343525179856115,
+      "step": 15950,
+      "torque_loss": 0.1357569545507431
+    },
+    {
+      "epoch": 14.343525179856115,
+      "force_loss": 0.005941580981016159,
+      "step": 15950
+    },
+    {
+      "epoch": 14.352517985611511,
+      "grad_norm": 0.3117270767688751,
+      "learning_rate": 8.777943866521612e-05,
+      "loss": 0.0343,
+      "step": 15960
+    },
+    {
+      "action_loss": 0.007253460586071014,
+      "epoch": 14.352517985611511,
+      "step": 15960
+    },
+    {
+      "epoch": 14.352517985611511,
+      "step": 15960,
+      "torque_loss": 0.15881727635860443
+    },
+    {
+      "epoch": 14.352517985611511,
+      "force_loss": 0.006982345134019852,
+      "step": 15960
+    },
+    {
+      "epoch": 14.361510791366907,
+      "grad_norm": 0.5149706602096558,
+      "learning_rate": 8.77613812742025e-05,
+      "loss": 0.0226,
+      "step": 15970
+    },
+    {
+      "action_loss": 0.010725936852395535,
+      "epoch": 14.361510791366907,
+      "step": 15970
+    },
+    {
+      "epoch": 14.361510791366907,
+      "step": 15970,
+      "torque_loss": 0.13262148201465607
+    },
+    {
+      "epoch": 14.361510791366907,
+      "force_loss": 0.009241418913006783,
+      "step": 15970
+    },
+    {
+      "epoch": 14.370503597122303,
+      "grad_norm": 0.7514430284500122,
+      "learning_rate": 8.774331241227829e-05,
+      "loss": 0.0296,
+      "step": 15980
+    },
+    {
+      "action_loss": 0.010539411567151546,
+      "epoch": 14.370503597122303,
+      "step": 15980
+    },
+    {
+      "epoch": 14.370503597122303,
+      "step": 15980,
+      "torque_loss": 0.12404807657003403
+    },
+    {
+      "epoch": 14.370503597122303,
+      "force_loss": 0.013775442726910114,
+      "step": 15980
+    },
+    {
+      "epoch": 14.379496402877697,
+      "grad_norm": 0.29295358061790466,
+      "learning_rate": 8.772523208493232e-05,
+      "loss": 0.0228,
+      "step": 15990
+    },
+    {
+      "action_loss": 0.014459223486483097,
+      "epoch": 14.379496402877697,
+      "step": 15990
+    },
+    {
+      "epoch": 14.379496402877697,
+      "step": 15990,
+      "torque_loss": 0.1487133502960205
+    },
+    {
+      "epoch": 14.379496402877697,
+      "force_loss": 0.009827100671827793,
+      "step": 15990
+    },
+    {
+      "epoch": 14.388489208633093,
+      "grad_norm": 0.6131527423858643,
+      "learning_rate": 8.770714029765692e-05,
+      "loss": 0.0311,
+      "step": 16000
+    },
+    {
+      "action_loss": 0.015181586146354675,
+      "epoch": 14.388489208633093,
+      "step": 16000
+    },
+    {
+      "epoch": 14.388489208633093,
+      "step": 16000,
+      "torque_loss": 0.14069007337093353
+    },
+    {
+      "epoch": 14.388489208633093,
+      "force_loss": 0.011419035494327545,
+      "step": 16000
+    },
+    {
+      "epoch": 14.397482014388489,
+      "grad_norm": 0.4668498933315277,
+      "learning_rate": 8.768903705594789e-05,
+      "loss": 0.0303,
+      "step": 16010
+    },
+    {
+      "action_loss": 0.01304984837770462,
+      "epoch": 14.397482014388489,
+      "step": 16010
+    },
+    {
+      "epoch": 14.397482014388489,
+      "step": 16010,
+      "torque_loss": 0.16862152516841888
+    },
+    {
+      "epoch": 14.397482014388489,
+      "force_loss": 0.011156275868415833,
+      "step": 16010
+    },
+    {
+      "epoch": 14.406474820143885,
+      "grad_norm": 0.44027066230773926,
+      "learning_rate": 8.767092236530453e-05,
+      "loss": 0.0277,
+      "step": 16020
+    },
+    {
+      "action_loss": 0.016446055844426155,
+      "epoch": 14.406474820143885,
+      "step": 16020
+    },
+    {
+      "epoch": 14.406474820143885,
+      "step": 16020,
+      "torque_loss": 0.14097067713737488
+    },
+    {
+      "epoch": 14.406474820143885,
+      "force_loss": 0.016288094222545624,
+      "step": 16020
+    },
+    {
+      "epoch": 14.41546762589928,
+      "grad_norm": 0.6568413376808167,
+      "learning_rate": 8.76527962312296e-05,
+      "loss": 0.031,
+      "step": 16030
+    },
+    {
+      "action_loss": 0.006981673184782267,
+      "epoch": 14.41546762589928,
+      "step": 16030
+    },
+    {
+      "epoch": 14.41546762589928,
+      "step": 16030,
+      "torque_loss": 0.12521032989025116
+    },
+    {
+      "epoch": 14.41546762589928,
+      "force_loss": 0.006815902888774872,
+      "step": 16030
+    },
+    {
+      "epoch": 14.424460431654676,
+      "grad_norm": 0.33631667494773865,
+      "learning_rate": 8.763465865922934e-05,
+      "loss": 0.024,
+      "step": 16040
+    },
+    {
+      "action_loss": 0.031969260424375534,
+      "epoch": 14.424460431654676,
+      "step": 16040
+    },
+    {
+      "epoch": 14.424460431654676,
+      "step": 16040,
+      "torque_loss": 0.1744951754808426
+    },
+    {
+      "epoch": 14.424460431654676,
+      "force_loss": 0.035487834364175797,
+      "step": 16040
+    },
+    {
+      "epoch": 14.433453237410072,
+      "grad_norm": 0.5534899234771729,
+      "learning_rate": 8.761650965481347e-05,
+      "loss": 0.0299,
+      "step": 16050
+    },
+    {
+      "action_loss": 0.011798910796642303,
+      "epoch": 14.433453237410072,
+      "step": 16050
+    },
+    {
+      "epoch": 14.433453237410072,
+      "step": 16050,
+      "torque_loss": 0.1327945739030838
+    },
+    {
+      "epoch": 14.433453237410072,
+      "force_loss": 0.007753260433673859,
+      "step": 16050
+    },
+    {
+      "epoch": 14.442446043165468,
+      "grad_norm": 0.4077386260032654,
+      "learning_rate": 8.759834922349516e-05,
+      "loss": 0.0248,
+      "step": 16060
+    },
+    {
+      "action_loss": 0.006675384473055601,
+      "epoch": 14.442446043165468,
+      "step": 16060
+    },
+    {
+      "epoch": 14.442446043165468,
+      "step": 16060,
+      "torque_loss": 0.16668426990509033
+    },
+    {
+      "epoch": 14.442446043165468,
+      "force_loss": 0.0039557721465826035,
+      "step": 16060
+    },
+    {
+      "epoch": 14.451438848920864,
+      "grad_norm": 0.3269275426864624,
+      "learning_rate": 8.758017737079108e-05,
+      "loss": 0.0261,
+      "step": 16070
+    },
+    {
+      "action_loss": 0.007361941039562225,
+      "epoch": 14.451438848920864,
+      "step": 16070
+    },
+    {
+      "epoch": 14.451438848920864,
+      "step": 16070,
+      "torque_loss": 0.11449036747217178
+    },
+    {
+      "epoch": 14.451438848920864,
+      "force_loss": 0.005952224135398865,
+      "step": 16070
+    },
+    {
+      "epoch": 14.46043165467626,
+      "grad_norm": 0.5325619578361511,
+      "learning_rate": 8.756199410222137e-05,
+      "loss": 0.0205,
+      "step": 16080
+    },
+    {
+      "action_loss": 0.0062446617521345615,
+      "epoch": 14.46043165467626,
+      "step": 16080
+    },
+    {
+      "epoch": 14.46043165467626,
+      "step": 16080,
+      "torque_loss": 0.14131304621696472
+    },
+    {
+      "epoch": 14.46043165467626,
+      "force_loss": 0.0054467604495584965,
+      "step": 16080
+    },
+    {
+      "epoch": 14.469424460431656,
+      "grad_norm": 0.5006517767906189,
+      "learning_rate": 8.754379942330963e-05,
+      "loss": 0.0241,
+      "step": 16090
+    },
+    {
+      "action_loss": 0.006374505814164877,
+      "epoch": 14.469424460431656,
+      "step": 16090
+    },
+    {
+      "epoch": 14.469424460431656,
+      "step": 16090,
+      "torque_loss": 0.14624978601932526
+    },
+    {
+      "epoch": 14.469424460431656,
+      "force_loss": 0.005361600313335657,
+      "step": 16090
+    },
+    {
+      "epoch": 14.47841726618705,
+      "grad_norm": 0.2686760127544403,
+      "learning_rate": 8.75255933395829e-05,
+      "loss": 0.0224,
+      "step": 16100
+    },
+    {
+      "action_loss": 0.01720951683819294,
+      "epoch": 14.47841726618705,
+      "step": 16100
+    },
+    {
+      "epoch": 14.47841726618705,
+      "step": 16100,
+      "torque_loss": 0.16719822585582733
+    },
+    {
+      "epoch": 14.47841726618705,
+      "force_loss": 0.014178097248077393,
+      "step": 16100
+    },
+    {
+      "epoch": 14.487410071942445,
+      "grad_norm": 0.4230107069015503,
+      "learning_rate": 8.750737585657171e-05,
+      "loss": 0.0265,
+      "step": 16110
+    },
+    {
+      "action_loss": 0.006234128028154373,
+      "epoch": 14.487410071942445,
+      "step": 16110
+    },
+    {
+      "epoch": 14.487410071942445,
+      "step": 16110,
+      "torque_loss": 0.10787993669509888
+    },
+    {
+      "epoch": 14.487410071942445,
+      "force_loss": 0.0037304898723959923,
+      "step": 16110
+    },
+    {
+      "epoch": 14.496402877697841,
+      "grad_norm": 0.6370156407356262,
+      "learning_rate": 8.748914697981008e-05,
+      "loss": 0.0247,
+      "step": 16120
+    },
+    {
+      "action_loss": 0.008517253212630749,
+      "epoch": 14.496402877697841,
+      "step": 16120
+    },
+    {
+      "epoch": 14.496402877697841,
+      "step": 16120,
+      "torque_loss": 0.08847641944885254
+    },
+    {
+      "epoch": 14.496402877697841,
+      "force_loss": 0.0049628303386271,
+      "step": 16120
+    },
+    {
+      "epoch": 14.505395683453237,
+      "grad_norm": 0.7003411650657654,
+      "learning_rate": 8.747090671483542e-05,
+      "loss": 0.0249,
+      "step": 16130
+    },
+    {
+      "action_loss": 0.002361086430028081,
+      "epoch": 14.505395683453237,
+      "step": 16130
+    },
+    {
+      "epoch": 14.505395683453237,
+      "step": 16130,
+      "torque_loss": 0.07902330905199051
+    },
+    {
+      "epoch": 14.505395683453237,
+      "force_loss": 0.004729801788926125,
+      "step": 16130
+    },
+    {
+      "epoch": 14.514388489208633,
+      "grad_norm": 0.34488117694854736,
+      "learning_rate": 8.745265506718869e-05,
+      "loss": 0.019,
+      "step": 16140
+    },
+    {
+      "action_loss": 0.004949633032083511,
+      "epoch": 14.514388489208633,
+      "step": 16140
+    },
+    {
+      "epoch": 14.514388489208633,
+      "step": 16140,
+      "torque_loss": 0.1301463097333908
+    },
+    {
+      "epoch": 14.514388489208633,
+      "force_loss": 0.006746729835867882,
+      "step": 16140
+    },
+    {
+      "epoch": 14.523381294964029,
+      "grad_norm": 0.8607816696166992,
+      "learning_rate": 8.74343920424142e-05,
+      "loss": 0.0245,
+      "step": 16150
+    },
+    {
+      "action_loss": 0.009195582009851933,
+      "epoch": 14.523381294964029,
+      "step": 16150
+    },
+    {
+      "epoch": 14.523381294964029,
+      "step": 16150,
+      "torque_loss": 0.12722842395305634
+    },
+    {
+      "epoch": 14.523381294964029,
+      "force_loss": 0.01190421637147665,
+      "step": 16150
+    },
+    {
+      "epoch": 14.532374100719425,
+      "grad_norm": 0.46893593668937683,
+      "learning_rate": 8.741611764605982e-05,
+      "loss": 0.0236,
+      "step": 16160
+    },
+    {
+      "action_loss": 0.005657311528921127,
+      "epoch": 14.532374100719425,
+      "step": 16160
+    },
+    {
+      "epoch": 14.532374100719425,
+      "step": 16160,
+      "torque_loss": 0.1434476524591446
+    },
+    {
+      "epoch": 14.532374100719425,
+      "force_loss": 0.005069083068519831,
+      "step": 16160
+    },
+    {
+      "epoch": 14.54136690647482,
+      "grad_norm": 0.28299471735954285,
+      "learning_rate": 8.739783188367682e-05,
+      "loss": 0.0245,
+      "step": 16170
+    },
+    {
+      "action_loss": 0.005101411137729883,
+      "epoch": 14.54136690647482,
+      "step": 16170
+    },
+    {
+      "epoch": 14.54136690647482,
+      "step": 16170,
+      "torque_loss": 0.11687228083610535
+    },
+    {
+      "epoch": 14.54136690647482,
+      "force_loss": 0.004526472184807062,
+      "step": 16170
+    },
+    {
+      "epoch": 14.550359712230216,
+      "grad_norm": 0.5864328145980835,
+      "learning_rate": 8.737953476081991e-05,
+      "loss": 0.0236,
+      "step": 16180
+    },
+    {
+      "action_loss": 0.00856483168900013,
+      "epoch": 14.550359712230216,
+      "step": 16180
+    },
+    {
+      "epoch": 14.550359712230216,
+      "step": 16180,
+      "torque_loss": 0.13133420050144196
+    },
+    {
+      "epoch": 14.550359712230216,
+      "force_loss": 0.009303194470703602,
+      "step": 16180
+    },
+    {
+      "epoch": 14.559352517985612,
+      "grad_norm": 0.36528876423835754,
+      "learning_rate": 8.73612262830473e-05,
+      "loss": 0.0233,
+      "step": 16190
+    },
+    {
+      "action_loss": 0.0048843990080058575,
+      "epoch": 14.559352517985612,
+      "step": 16190
+    },
+    {
+      "epoch": 14.559352517985612,
+      "step": 16190,
+      "torque_loss": 0.10311051458120346
+    },
+    {
+      "epoch": 14.559352517985612,
+      "force_loss": 0.004061442334204912,
+      "step": 16190
+    },
+    {
+      "epoch": 14.568345323741006,
+      "grad_norm": 0.362733393907547,
+      "learning_rate": 8.734290645592061e-05,
+      "loss": 0.0198,
+      "step": 16200
+    },
+    {
+      "action_loss": 0.00472608907148242,
+      "epoch": 14.568345323741006,
+      "step": 16200
+    },
+    {
+      "epoch": 14.568345323741006,
+      "step": 16200,
+      "torque_loss": 0.16307905316352844
+    },
+    {
+      "epoch": 14.568345323741006,
+      "force_loss": 0.0038498062640428543,
+      "step": 16200
+    },
+    {
+      "epoch": 14.577338129496402,
+      "grad_norm": 0.6847972273826599,
+      "learning_rate": 8.732457528500493e-05,
+      "loss": 0.0233,
+      "step": 16210
+    },
+    {
+      "action_loss": 0.006037202198058367,
+      "epoch": 14.577338129496402,
+      "step": 16210
+    },
+    {
+      "epoch": 14.577338129496402,
+      "step": 16210,
+      "torque_loss": 0.18439090251922607
+    },
+    {
+      "epoch": 14.577338129496402,
+      "force_loss": 0.0037185444962233305,
+      "step": 16210
+    },
+    {
+      "epoch": 14.586330935251798,
+      "grad_norm": 0.578105092048645,
+      "learning_rate": 8.730623277586875e-05,
+      "loss": 0.0245,
+      "step": 16220
+    },
+    {
+      "action_loss": 0.009703953750431538,
+      "epoch": 14.586330935251798,
+      "step": 16220
+    },
+    {
+      "epoch": 14.586330935251798,
+      "step": 16220,
+      "torque_loss": 0.11771911382675171
+    },
+    {
+      "epoch": 14.586330935251798,
+      "force_loss": 0.008038971573114395,
+      "step": 16220
+    },
+    {
+      "epoch": 14.595323741007194,
+      "grad_norm": 0.3750212490558624,
+      "learning_rate": 8.72878789340841e-05,
+      "loss": 0.0217,
+      "step": 16230
+    },
+    {
+      "action_loss": 0.00477495975792408,
+      "epoch": 14.595323741007194,
+      "step": 16230
+    },
+    {
+      "epoch": 14.595323741007194,
+      "step": 16230,
+      "torque_loss": 0.11629226803779602
+    },
+    {
+      "epoch": 14.595323741007194,
+      "force_loss": 0.0035047137644141912,
+      "step": 16230
+    },
+    {
+      "epoch": 14.60431654676259,
+      "grad_norm": 0.31527113914489746,
+      "learning_rate": 8.726951376522635e-05,
+      "loss": 0.0256,
+      "step": 16240
+    },
+    {
+      "action_loss": 0.007542117964476347,
+      "epoch": 14.60431654676259,
+      "step": 16240
+    },
+    {
+      "epoch": 14.60431654676259,
+      "step": 16240,
+      "torque_loss": 0.12231564521789551
+    },
+    {
+      "epoch": 14.60431654676259,
+      "force_loss": 0.005131407640874386,
+      "step": 16240
+    },
+    {
+      "epoch": 14.613309352517986,
+      "grad_norm": 0.4643120765686035,
+      "learning_rate": 8.725113727487435e-05,
+      "loss": 0.0223,
+      "step": 16250
+    },
+    {
+      "action_loss": 0.01127786934375763,
+      "epoch": 14.613309352517986,
+      "step": 16250
+    },
+    {
+      "epoch": 14.613309352517986,
+      "step": 16250,
+      "torque_loss": 0.13514098525047302
+    },
+    {
+      "epoch": 14.613309352517986,
+      "force_loss": 0.007229645270854235,
+      "step": 16250
+    },
+    {
+      "epoch": 14.622302158273381,
+      "grad_norm": 0.33437561988830566,
+      "learning_rate": 8.723274946861042e-05,
+      "loss": 0.0215,
+      "step": 16260
+    },
+    {
+      "action_loss": 0.004695317707955837,
+      "epoch": 14.622302158273381,
+      "step": 16260
+    },
+    {
+      "epoch": 14.622302158273381,
+      "step": 16260,
+      "torque_loss": 0.11406657844781876
+    },
+    {
+      "epoch": 14.622302158273381,
+      "force_loss": 0.007661603856831789,
+      "step": 16260
+    },
+    {
+      "epoch": 14.631294964028777,
+      "grad_norm": 0.7634636163711548,
+      "learning_rate": 8.721435035202026e-05,
+      "loss": 0.0232,
+      "step": 16270
+    },
+    {
+      "action_loss": 0.015687547624111176,
+      "epoch": 14.631294964028777,
+      "step": 16270
+    },
+    {
+      "epoch": 14.631294964028777,
+      "step": 16270,
+      "torque_loss": 0.16938884556293488
+    },
+    {
+      "epoch": 14.631294964028777,
+      "force_loss": 0.02431638538837433,
+      "step": 16270
+    },
+    {
+      "epoch": 14.640287769784173,
+      "grad_norm": 0.38326090574264526,
+      "learning_rate": 8.719593993069306e-05,
+      "loss": 0.0304,
+      "step": 16280
+    },
+    {
+      "action_loss": 0.00968262366950512,
+      "epoch": 14.640287769784173,
+      "step": 16280
+    },
+    {
+      "epoch": 14.640287769784173,
+      "step": 16280,
+      "torque_loss": 0.10673090070486069
+    },
+    {
+      "epoch": 14.640287769784173,
+      "force_loss": 0.01634608395397663,
+      "step": 16280
+    },
+    {
+      "epoch": 14.649280575539569,
+      "grad_norm": 0.7706455588340759,
+      "learning_rate": 8.717751821022139e-05,
+      "loss": 0.0256,
+      "step": 16290
+    },
+    {
+      "action_loss": 0.008624917827546597,
+      "epoch": 14.649280575539569,
+      "step": 16290
+    },
+    {
+      "epoch": 14.649280575539569,
+      "step": 16290,
+      "torque_loss": 0.1214180588722229
+    },
+    {
+      "epoch": 14.649280575539569,
+      "force_loss": 0.011329042725265026,
+      "step": 16290
+    },
+    {
+      "epoch": 14.658273381294965,
+      "grad_norm": 0.42331382632255554,
+      "learning_rate": 8.715908519620134e-05,
+      "loss": 0.0255,
+      "step": 16300
+    },
+    {
+      "action_loss": 0.023016376420855522,
+      "epoch": 14.658273381294965,
+      "step": 16300
+    },
+    {
+      "epoch": 14.658273381294965,
+      "step": 16300,
+      "torque_loss": 0.15137328207492828
+    },
+    {
+      "epoch": 14.658273381294965,
+      "force_loss": 0.021671846508979797,
+      "step": 16300
+    },
+    {
+      "epoch": 14.667266187050359,
+      "grad_norm": 0.33794134855270386,
+      "learning_rate": 8.71406408942323e-05,
+      "loss": 0.0262,
+      "step": 16310
+    },
+    {
+      "action_loss": 0.010126424953341484,
+      "epoch": 14.667266187050359,
+      "step": 16310
+    },
+    {
+      "epoch": 14.667266187050359,
+      "step": 16310,
+      "torque_loss": 0.10819221287965775
+    },
+    {
+      "epoch": 14.667266187050359,
+      "force_loss": 0.007030328270047903,
+      "step": 16310
+    },
+    {
+      "epoch": 14.676258992805755,
+      "grad_norm": 0.2426094263792038,
+      "learning_rate": 8.712218530991723e-05,
+      "loss": 0.0211,
+      "step": 16320
+    },
+    {
+      "action_loss": 0.011374286375939846,
+      "epoch": 14.676258992805755,
+      "step": 16320
+    },
+    {
+      "epoch": 14.676258992805755,
+      "step": 16320,
+      "torque_loss": 0.1683548241853714
+    },
+    {
+      "epoch": 14.676258992805755,
+      "force_loss": 0.013959120959043503,
+      "step": 16320
+    },
+    {
+      "epoch": 14.68525179856115,
+      "grad_norm": 0.7456856966018677,
+      "learning_rate": 8.710371844886241e-05,
+      "loss": 0.0226,
+      "step": 16330
+    },
+    {
+      "action_loss": 0.014729071408510208,
+      "epoch": 14.68525179856115,
+      "step": 16330
+    },
+    {
+      "epoch": 14.68525179856115,
+      "step": 16330,
+      "torque_loss": 0.1768568903207779
+    },
+    {
+      "epoch": 14.68525179856115,
+      "force_loss": 0.01182624977082014,
+      "step": 16330
+    },
+    {
+      "epoch": 14.694244604316546,
+      "grad_norm": 0.4690254032611847,
+      "learning_rate": 8.708524031667758e-05,
+      "loss": 0.0238,
+      "step": 16340
+    },
+    {
+      "action_loss": 0.015546525828540325,
+      "epoch": 14.694244604316546,
+      "step": 16340
+    },
+    {
+      "epoch": 14.694244604316546,
+      "step": 16340,
+      "torque_loss": 0.14213943481445312
+    },
+    {
+      "epoch": 14.694244604316546,
+      "force_loss": 0.01066828053444624,
+      "step": 16340
+    },
+    {
+      "epoch": 14.703237410071942,
+      "grad_norm": 0.4124860465526581,
+      "learning_rate": 8.706675091897592e-05,
+      "loss": 0.0273,
+      "step": 16350
+    },
+    {
+      "action_loss": 0.008112015202641487,
+      "epoch": 14.703237410071942,
+      "step": 16350
+    },
+    {
+      "epoch": 14.703237410071942,
+      "step": 16350,
+      "torque_loss": 0.1799161434173584
+    },
+    {
+      "epoch": 14.703237410071942,
+      "force_loss": 0.004709279630333185,
+      "step": 16350
+    },
+    {
+      "epoch": 14.712230215827338,
+      "grad_norm": 0.2928280532360077,
+      "learning_rate": 8.704825026137404e-05,
+      "loss": 0.0258,
+      "step": 16360
+    },
+    {
+      "action_loss": 0.011181623674929142,
+      "epoch": 14.712230215827338,
+      "step": 16360
+    },
+    {
+      "epoch": 14.712230215827338,
+      "step": 16360,
+      "torque_loss": 0.15154887735843658
+    },
+    {
+      "epoch": 14.712230215827338,
+      "force_loss": 0.011326193809509277,
+      "step": 16360
+    },
+    {
+      "epoch": 14.721223021582734,
+      "grad_norm": 0.2906883955001831,
+      "learning_rate": 8.702973834949192e-05,
+      "loss": 0.025,
+      "step": 16370
+    },
+    {
+      "action_loss": 0.010558349080383778,
+      "epoch": 14.721223021582734,
+      "step": 16370
+    },
+    {
+      "epoch": 14.721223021582734,
+      "step": 16370,
+      "torque_loss": 0.13773809373378754
+    },
+    {
+      "epoch": 14.721223021582734,
+      "force_loss": 0.011783388443291187,
+      "step": 16370
+    },
+    {
+      "epoch": 14.73021582733813,
+      "grad_norm": 0.7011801600456238,
+      "learning_rate": 8.701121518895301e-05,
+      "loss": 0.0252,
+      "step": 16380
+    },
+    {
+      "action_loss": 0.011409520171582699,
+      "epoch": 14.73021582733813,
+      "step": 16380
+    },
+    {
+      "epoch": 14.73021582733813,
+      "step": 16380,
+      "torque_loss": 0.15905313193798065
+    },
+    {
+      "epoch": 14.73021582733813,
+      "force_loss": 0.017366239801049232,
+      "step": 16380
+    },
+    {
+      "epoch": 14.739208633093526,
+      "grad_norm": 0.4976460635662079,
+      "learning_rate": 8.699268078538414e-05,
+      "loss": 0.0277,
+      "step": 16390
+    },
+    {
+      "action_loss": 0.00357674527913332,
+      "epoch": 14.739208633093526,
+      "step": 16390
+    },
+    {
+      "epoch": 14.739208633093526,
+      "step": 16390,
+      "torque_loss": 0.1501704305410385
+    },
+    {
+      "epoch": 14.739208633093526,
+      "force_loss": 0.0039963372983038425,
+      "step": 16390
+    },
+    {
+      "epoch": 14.748201438848922,
+      "grad_norm": 0.4516439437866211,
+      "learning_rate": 8.69741351444156e-05,
+      "loss": 0.0223,
+      "step": 16400
+    },
+    {
+      "action_loss": 0.011151435784995556,
+      "epoch": 14.748201438848922,
+      "step": 16400
+    },
+    {
+      "epoch": 14.748201438848922,
+      "step": 16400,
+      "torque_loss": 0.1500273495912552
+    },
+    {
+      "epoch": 14.748201438848922,
+      "force_loss": 0.010211797431111336,
+      "step": 16400
+    },
+    {
+      "epoch": 14.757194244604317,
+      "grad_norm": 0.42409786581993103,
+      "learning_rate": 8.695557827168101e-05,
+      "loss": 0.022,
+      "step": 16410
+    },
+    {
+      "action_loss": 0.007657613139599562,
+      "epoch": 14.757194244604317,
+      "step": 16410
+    },
+    {
+      "epoch": 14.757194244604317,
+      "step": 16410,
+      "torque_loss": 0.1123104989528656
+    },
+    {
+      "epoch": 14.757194244604317,
+      "force_loss": 0.00562768941745162,
+      "step": 16410
+    },
+    {
+      "epoch": 14.766187050359711,
+      "grad_norm": 0.3427654504776001,
+      "learning_rate": 8.693701017281753e-05,
+      "loss": 0.0254,
+      "step": 16420
+    },
+    {
+      "action_loss": 0.013873311690986156,
+      "epoch": 14.766187050359711,
+      "step": 16420
+    },
+    {
+      "epoch": 14.766187050359711,
+      "step": 16420,
+      "torque_loss": 0.17615483701229095
+    },
+    {
+      "epoch": 14.766187050359711,
+      "force_loss": 0.024800434708595276,
+      "step": 16420
+    },
+    {
+      "epoch": 14.775179856115107,
+      "grad_norm": 1.050300121307373,
+      "learning_rate": 8.691843085346563e-05,
+      "loss": 0.0283,
+      "step": 16430
+    },
+    {
+      "action_loss": 0.006628282833844423,
+      "epoch": 14.775179856115107,
+      "step": 16430
+    },
+    {
+      "epoch": 14.775179856115107,
+      "step": 16430,
+      "torque_loss": 0.09419536590576172
+    },
+    {
+      "epoch": 14.775179856115107,
+      "force_loss": 0.006657319143414497,
+      "step": 16430
+    },
+    {
+      "epoch": 14.784172661870503,
+      "grad_norm": 0.5347671508789062,
+      "learning_rate": 8.689984031926919e-05,
+      "loss": 0.0226,
+      "step": 16440
+    },
+    {
+      "action_loss": 0.008300053887069225,
+      "epoch": 14.784172661870503,
+      "step": 16440
+    },
+    {
+      "epoch": 14.784172661870503,
+      "step": 16440,
+      "torque_loss": 0.14323776960372925
+    },
+    {
+      "epoch": 14.784172661870503,
+      "force_loss": 0.007797125726938248,
+      "step": 16440
+    },
+    {
+      "epoch": 14.793165467625899,
+      "grad_norm": 0.32349851727485657,
+      "learning_rate": 8.688123857587555e-05,
+      "loss": 0.0242,
+      "step": 16450
+    },
+    {
+      "action_loss": 0.0037973488215357065,
+      "epoch": 14.793165467625899,
+      "step": 16450
+    },
+    {
+      "epoch": 14.793165467625899,
+      "step": 16450,
+      "torque_loss": 0.12730592489242554
+    },
+    {
+      "epoch": 14.793165467625899,
+      "force_loss": 0.0064923143945634365,
+      "step": 16450
+    },
+    {
+      "epoch": 14.802158273381295,
+      "grad_norm": 0.6720380783081055,
+      "learning_rate": 8.686262562893544e-05,
+      "loss": 0.0255,
+      "step": 16460
+    },
+    {
+      "action_loss": 0.008057990111410618,
+      "epoch": 14.802158273381295,
+      "step": 16460
+    },
+    {
+      "epoch": 14.802158273381295,
+      "step": 16460,
+      "torque_loss": 0.120577372610569
+    },
+    {
+      "epoch": 14.802158273381295,
+      "force_loss": 0.007687706965953112,
+      "step": 16460
+    },
+    {
+      "epoch": 14.81115107913669,
+      "grad_norm": 0.25501903891563416,
+      "learning_rate": 8.684400148410294e-05,
+      "loss": 0.0212,
+      "step": 16470
+    },
+    {
+      "action_loss": 0.0035678327549248934,
+      "epoch": 14.81115107913669,
+      "step": 16470
+    },
+    {
+      "epoch": 14.81115107913669,
+      "step": 16470,
+      "torque_loss": 0.12409751862287521
+    },
+    {
+      "epoch": 14.81115107913669,
+      "force_loss": 0.006432787049561739,
+      "step": 16470
+    },
+    {
+      "epoch": 14.820143884892087,
+      "grad_norm": 0.22578421235084534,
+      "learning_rate": 8.682536614703562e-05,
+      "loss": 0.0227,
+      "step": 16480
+    },
+    {
+      "action_loss": 0.008843987248837948,
+      "epoch": 14.820143884892087,
+      "step": 16480
+    },
+    {
+      "epoch": 14.820143884892087,
+      "step": 16480,
+      "torque_loss": 0.1574493944644928
+    },
+    {
+      "epoch": 14.820143884892087,
+      "force_loss": 0.011142074130475521,
+      "step": 16480
+    },
+    {
+      "epoch": 14.829136690647482,
+      "grad_norm": 0.5097362399101257,
+      "learning_rate": 8.680671962339437e-05,
+      "loss": 0.0224,
+      "step": 16490
+    },
+    {
+      "action_loss": 0.008273611776530743,
+      "epoch": 14.829136690647482,
+      "step": 16490
+    },
+    {
+      "epoch": 14.829136690647482,
+      "step": 16490,
+      "torque_loss": 0.1288774013519287
+    },
+    {
+      "epoch": 14.829136690647482,
+      "force_loss": 0.010392571799457073,
+      "step": 16490
+    },
+    {
+      "epoch": 14.838129496402878,
+      "grad_norm": 0.45725390315055847,
+      "learning_rate": 8.678806191884352e-05,
+      "loss": 0.0253,
+      "step": 16500
+    },
+    {
+      "action_loss": 0.008720788173377514,
+      "epoch": 14.838129496402878,
+      "step": 16500
+    },
+    {
+      "epoch": 14.838129496402878,
+      "step": 16500,
+      "torque_loss": 0.16255836188793182
+    },
+    {
+      "epoch": 14.838129496402878,
+      "force_loss": 0.008494697511196136,
+      "step": 16500
+    },
+    {
+      "epoch": 14.847122302158274,
+      "grad_norm": 0.21858061850070953,
+      "learning_rate": 8.67693930390508e-05,
+      "loss": 0.0217,
+      "step": 16510
+    },
+    {
+      "action_loss": 0.028219209983944893,
+      "epoch": 14.847122302158274,
+      "step": 16510
+    },
+    {
+      "epoch": 14.847122302158274,
+      "step": 16510,
+      "torque_loss": 0.11343064159154892
+    },
+    {
+      "epoch": 14.847122302158274,
+      "force_loss": 0.021377915516495705,
+      "step": 16510
+    },
+    {
+      "epoch": 14.85611510791367,
+      "grad_norm": 0.32938069105148315,
+      "learning_rate": 8.67507129896873e-05,
+      "loss": 0.0216,
+      "step": 16520
+    },
+    {
+      "action_loss": 0.013413523323833942,
+      "epoch": 14.85611510791367,
+      "step": 16520
+    },
+    {
+      "epoch": 14.85611510791367,
+      "step": 16520,
+      "torque_loss": 0.1496639847755432
+    },
+    {
+      "epoch": 14.85611510791367,
+      "force_loss": 0.009586935862898827,
+      "step": 16520
+    },
+    {
+      "epoch": 14.865107913669064,
+      "grad_norm": 0.5014426708221436,
+      "learning_rate": 8.673202177642757e-05,
+      "loss": 0.0303,
+      "step": 16530
+    },
+    {
+      "action_loss": 0.014454229734838009,
+      "epoch": 14.865107913669064,
+      "step": 16530
+    },
+    {
+      "epoch": 14.865107913669064,
+      "step": 16530,
+      "torque_loss": 0.19893985986709595
+    },
+    {
+      "epoch": 14.865107913669064,
+      "force_loss": 0.01378584373742342,
+      "step": 16530
+    },
+    {
+      "epoch": 14.87410071942446,
+      "grad_norm": 0.38330456614494324,
+      "learning_rate": 8.671331940494945e-05,
+      "loss": 0.0331,
+      "step": 16540
+    },
+    {
+      "action_loss": 0.013897606171667576,
+      "epoch": 14.87410071942446,
+      "step": 16540
+    },
+    {
+      "epoch": 14.87410071942446,
+      "step": 16540,
+      "torque_loss": 0.1659325212240219
+    },
+    {
+      "epoch": 14.87410071942446,
+      "force_loss": 0.013821497559547424,
+      "step": 16540
+    },
+    {
+      "epoch": 14.883093525179856,
+      "grad_norm": 0.34845295548439026,
+      "learning_rate": 8.669460588093427e-05,
+      "loss": 0.0273,
+      "step": 16550
+    },
+    {
+      "action_loss": 0.006512706633657217,
+      "epoch": 14.883093525179856,
+      "step": 16550
+    },
+    {
+      "epoch": 14.883093525179856,
+      "step": 16550,
+      "torque_loss": 0.17432083189487457
+    },
+    {
+      "epoch": 14.883093525179856,
+      "force_loss": 0.007318615447729826,
+      "step": 16550
+    },
+    {
+      "epoch": 14.892086330935252,
+      "grad_norm": 0.277150958776474,
+      "learning_rate": 8.667588121006667e-05,
+      "loss": 0.0231,
+      "step": 16560
+    },
+    {
+      "action_loss": 0.030604219064116478,
+      "epoch": 14.892086330935252,
+      "step": 16560
+    },
+    {
+      "epoch": 14.892086330935252,
+      "step": 16560,
+      "torque_loss": 0.12815971672534943
+    },
+    {
+      "epoch": 14.892086330935252,
+      "force_loss": 0.040705639868974686,
+      "step": 16560
+    },
+    {
+      "epoch": 14.901079136690647,
+      "grad_norm": 0.4285851716995239,
+      "learning_rate": 8.665714539803475e-05,
+      "loss": 0.0293,
+      "step": 16570
+    },
+    {
+      "action_loss": 0.006151033565402031,
+      "epoch": 14.901079136690647,
+      "step": 16570
+    },
+    {
+      "epoch": 14.901079136690647,
+      "step": 16570,
+      "torque_loss": 0.09936177730560303
+    },
+    {
+      "epoch": 14.901079136690647,
+      "force_loss": 0.0033768664579838514,
+      "step": 16570
+    },
+    {
+      "epoch": 14.910071942446043,
+      "grad_norm": 0.7274307012557983,
+      "learning_rate": 8.663839845052993e-05,
+      "loss": 0.0211,
+      "step": 16580
+    },
+    {
+      "action_loss": 0.009478853084146976,
+      "epoch": 14.910071942446043,
+      "step": 16580
+    },
+    {
+      "epoch": 14.910071942446043,
+      "step": 16580,
+      "torque_loss": 0.149641215801239
+    },
+    {
+      "epoch": 14.910071942446043,
+      "force_loss": 0.0051117464900016785,
+      "step": 16580
+    },
+    {
+      "epoch": 14.91906474820144,
+      "grad_norm": 0.8926911354064941,
+      "learning_rate": 8.661964037324703e-05,
+      "loss": 0.0264,
+      "step": 16590
+    },
+    {
+      "action_loss": 0.018504738807678223,
+      "epoch": 14.91906474820144,
+      "step": 16590
+    },
+    {
+      "epoch": 14.91906474820144,
+      "step": 16590,
+      "torque_loss": 0.12704424560070038
+    },
+    {
+      "epoch": 14.91906474820144,
+      "force_loss": 0.021473711356520653,
+      "step": 16590
+    },
+    {
+      "epoch": 14.928057553956835,
+      "grad_norm": 0.3026418089866638,
+      "learning_rate": 8.660087117188427e-05,
+      "loss": 0.0206,
+      "step": 16600
+    },
+    {
+      "action_loss": 0.011007758788764477,
+      "epoch": 14.928057553956835,
+      "step": 16600
+    },
+    {
+      "epoch": 14.928057553956835,
+      "step": 16600,
+      "torque_loss": 0.14828522503376007
+    },
+    {
+      "epoch": 14.928057553956835,
+      "force_loss": 0.008979643695056438,
+      "step": 16600
+    },
+    {
+      "epoch": 14.93705035971223,
+      "grad_norm": 0.34866297245025635,
+      "learning_rate": 8.658209085214325e-05,
+      "loss": 0.0235,
+      "step": 16610
+    },
+    {
+      "action_loss": 0.011805638670921326,
+      "epoch": 14.93705035971223,
+      "step": 16610
+    },
+    {
+      "epoch": 14.93705035971223,
+      "step": 16610,
+      "torque_loss": 0.16651158034801483
+    },
+    {
+      "epoch": 14.93705035971223,
+      "force_loss": 0.010302189737558365,
+      "step": 16610
+    },
+    {
+      "epoch": 14.946043165467627,
+      "grad_norm": 0.43696945905685425,
+      "learning_rate": 8.656329941972891e-05,
+      "loss": 0.0325,
+      "step": 16620
+    },
+    {
+      "action_loss": 0.006582924630492926,
+      "epoch": 14.946043165467627,
+      "step": 16620
+    },
+    {
+      "epoch": 14.946043165467627,
+      "step": 16620,
+      "torque_loss": 0.09430796653032303
+    },
+    {
+      "epoch": 14.946043165467627,
+      "force_loss": 0.006717576179653406,
+      "step": 16620
+    },
+    {
+      "epoch": 14.95503597122302,
+      "grad_norm": 0.28286707401275635,
+      "learning_rate": 8.654449688034963e-05,
+      "loss": 0.026,
+      "step": 16630
+    },
+    {
+      "action_loss": 0.013563995249569416,
+      "epoch": 14.95503597122302,
+      "step": 16630
+    },
+    {
+      "epoch": 14.95503597122302,
+      "step": 16630,
+      "torque_loss": 0.1204179897904396
+    },
+    {
+      "epoch": 14.95503597122302,
+      "force_loss": 0.01628933474421501,
+      "step": 16630
+    },
+    {
+      "epoch": 14.964028776978417,
+      "grad_norm": 0.27962011098861694,
+      "learning_rate": 8.652568323971706e-05,
+      "loss": 0.0259,
+      "step": 16640
+    },
+    {
+      "action_loss": 0.004835537634789944,
+      "epoch": 14.964028776978417,
+      "step": 16640
+    },
+    {
+      "epoch": 14.964028776978417,
+      "step": 16640,
+      "torque_loss": 0.12375181913375854
+    },
+    {
+      "epoch": 14.964028776978417,
+      "force_loss": 0.007731992285698652,
+      "step": 16640
+    },
+    {
+      "epoch": 14.973021582733812,
+      "grad_norm": 0.2986242175102234,
+      "learning_rate": 8.650685850354636e-05,
+      "loss": 0.0229,
+      "step": 16650
+    },
+    {
+      "action_loss": 0.01162691693753004,
+      "epoch": 14.973021582733812,
+      "step": 16650
+    },
+    {
+      "epoch": 14.973021582733812,
+      "step": 16650,
+      "torque_loss": 0.1142582893371582
+    },
+    {
+      "epoch": 14.973021582733812,
+      "force_loss": 0.01357144396752119,
+      "step": 16650
+    },
+    {
+      "epoch": 14.982014388489208,
+      "grad_norm": 0.40612348914146423,
+      "learning_rate": 8.648802267755593e-05,
+      "loss": 0.0271,
+      "step": 16660
+    },
+    {
+      "action_loss": 0.0030443326104432344,
+      "epoch": 14.982014388489208,
+      "step": 16660
+    },
+    {
+      "epoch": 14.982014388489208,
+      "step": 16660,
+      "torque_loss": 0.10475953668355942
+    },
+    {
+      "epoch": 14.982014388489208,
+      "force_loss": 0.0034353325609117746,
+      "step": 16660
+    },
+    {
+      "epoch": 14.991007194244604,
+      "grad_norm": 0.3785460889339447,
+      "learning_rate": 8.646917576746764e-05,
+      "loss": 0.0258,
+      "step": 16670
+    },
+    {
+      "action_loss": 0.006419685203582048,
+      "epoch": 14.991007194244604,
+      "step": 16670
+    },
+    {
+      "epoch": 14.991007194244604,
+      "step": 16670,
+      "torque_loss": 0.13704830408096313
+    },
+    {
+      "epoch": 14.991007194244604,
+      "force_loss": 0.00422352971509099,
+      "step": 16670
+    },
+    {
+      "epoch": 15.0,
+      "grad_norm": 0.3477333188056946,
+      "learning_rate": 8.645031777900666e-05,
+      "loss": 0.0215,
+      "step": 16680
+    },
+    {
+      "action_loss": 0.009861718863248825,
+      "epoch": 15.0,
+      "step": 16680
+    },
+    {
+      "epoch": 15.0,
+      "step": 16680,
+      "torque_loss": 0.11473336815834045
+    },
+    {
+      "epoch": 15.0,
+      "force_loss": 0.008207684382796288,
+      "step": 16680
+    },
+    {
+      "epoch": 15.008992805755396,
+      "grad_norm": 0.33410927653312683,
+      "learning_rate": 8.643144871790154e-05,
+      "loss": 0.0233,
+      "step": 16690
+    },
+    {
+      "action_loss": 0.005458649713546038,
+      "epoch": 15.008992805755396,
+      "step": 16690
+    },
+    {
+      "epoch": 15.008992805755396,
+      "step": 16690,
+      "torque_loss": 0.1323322057723999
+    },
+    {
+      "epoch": 15.008992805755396,
+      "force_loss": 0.004970806650817394,
+      "step": 16690
+    },
+    {
+      "epoch": 15.017985611510792,
+      "grad_norm": 0.26669541001319885,
+      "learning_rate": 8.641256858988424e-05,
+      "loss": 0.0212,
+      "step": 16700
+    },
+    {
+      "action_loss": 0.005379742477089167,
+      "epoch": 15.017985611510792,
+      "step": 16700
+    },
+    {
+      "epoch": 15.017985611510792,
+      "step": 16700,
+      "torque_loss": 0.096858911216259
+    },
+    {
+      "epoch": 15.017985611510792,
+      "force_loss": 0.010163123719394207,
+      "step": 16700
+    },
+    {
+      "epoch": 15.026978417266188,
+      "grad_norm": 0.47695598006248474,
+      "learning_rate": 8.639367740069e-05,
+      "loss": 0.0228,
+      "step": 16710
+    },
+    {
+      "action_loss": 0.00768280727788806,
+      "epoch": 15.026978417266188,
+      "step": 16710
+    },
+    {
+      "epoch": 15.026978417266188,
+      "step": 16710,
+      "torque_loss": 0.12936653196811676
+    },
+    {
+      "epoch": 15.026978417266188,
+      "force_loss": 0.006900383625179529,
+      "step": 16710
+    },
+    {
+      "epoch": 15.035971223021583,
+      "grad_norm": 0.4819808900356293,
+      "learning_rate": 8.63747751560575e-05,
+      "loss": 0.0229,
+      "step": 16720
+    },
+    {
+      "action_loss": 0.005033209454268217,
+      "epoch": 15.035971223021583,
+      "step": 16720
+    },
+    {
+      "epoch": 15.035971223021583,
+      "step": 16720,
+      "torque_loss": 0.11651767045259476
+    },
+    {
+      "epoch": 15.035971223021583,
+      "force_loss": 0.004356353543698788,
+      "step": 16720
+    },
+    {
+      "epoch": 15.04496402877698,
+      "grad_norm": 0.3256521224975586,
+      "learning_rate": 8.635586186172871e-05,
+      "loss": 0.0209,
+      "step": 16730
+    },
+    {
+      "action_loss": 0.004675617441534996,
+      "epoch": 15.04496402877698,
+      "step": 16730
+    },
+    {
+      "epoch": 15.04496402877698,
+      "step": 16730,
+      "torque_loss": 0.14101998507976532
+    },
+    {
+      "epoch": 15.04496402877698,
+      "force_loss": 0.005581811536103487,
+      "step": 16730
+    },
+    {
+      "epoch": 15.053956834532373,
+      "grad_norm": 0.4358675479888916,
+      "learning_rate": 8.633693752344902e-05,
+      "loss": 0.0209,
+      "step": 16740
+    },
+    {
+      "action_loss": 0.004803153220564127,
+      "epoch": 15.053956834532373,
+      "step": 16740
+    },
+    {
+      "epoch": 15.053956834532373,
+      "step": 16740,
+      "torque_loss": 0.12264394015073776
+    },
+    {
+      "epoch": 15.053956834532373,
+      "force_loss": 0.005168326664716005,
+      "step": 16740
+    },
+    {
+      "epoch": 15.06294964028777,
+      "grad_norm": 0.6398131847381592,
+      "learning_rate": 8.631800214696713e-05,
+      "loss": 0.0214,
+      "step": 16750
+    },
+    {
+      "action_loss": 0.0083686001598835,
+      "epoch": 15.06294964028777,
+      "step": 16750
+    },
+    {
+      "epoch": 15.06294964028777,
+      "step": 16750,
+      "torque_loss": 0.1393813043832779
+    },
+    {
+      "epoch": 15.06294964028777,
+      "force_loss": 0.0058719017542898655,
+      "step": 16750
+    },
+    {
+      "epoch": 15.071942446043165,
+      "grad_norm": 0.22567465901374817,
+      "learning_rate": 8.629905573803511e-05,
+      "loss": 0.0231,
+      "step": 16760
+    },
+    {
+      "action_loss": 0.0068164789117872715,
+      "epoch": 15.071942446043165,
+      "step": 16760
+    },
+    {
+      "epoch": 15.071942446043165,
+      "step": 16760,
+      "torque_loss": 0.11044230312108994
+    },
+    {
+      "epoch": 15.071942446043165,
+      "force_loss": 0.005865610670298338,
+      "step": 16760
+    },
+    {
+      "epoch": 15.08093525179856,
+      "grad_norm": 0.33949726819992065,
+      "learning_rate": 8.628009830240839e-05,
+      "loss": 0.0235,
+      "step": 16770
+    },
+    {
+      "action_loss": 0.0038899772334843874,
+      "epoch": 15.08093525179856,
+      "step": 16770
+    },
+    {
+      "epoch": 15.08093525179856,
+      "step": 16770,
+      "torque_loss": 0.0786958560347557
+    },
+    {
+      "epoch": 15.08093525179856,
+      "force_loss": 0.004531762097030878,
+      "step": 16770
+    },
+    {
+      "epoch": 15.089928057553957,
+      "grad_norm": 0.2313470095396042,
+      "learning_rate": 8.626112984584571e-05,
+      "loss": 0.0236,
+      "step": 16780
+    },
+    {
+      "action_loss": 0.004389610607177019,
+      "epoch": 15.089928057553957,
+      "step": 16780
+    },
+    {
+      "epoch": 15.089928057553957,
+      "step": 16780,
+      "torque_loss": 0.09413108974695206
+    },
+    {
+      "epoch": 15.089928057553957,
+      "force_loss": 0.006620475556701422,
+      "step": 16780
+    },
+    {
+      "epoch": 15.098920863309353,
+      "grad_norm": 0.28301331400871277,
+      "learning_rate": 8.62421503741092e-05,
+      "loss": 0.0233,
+      "step": 16790
+    },
+    {
+      "action_loss": 0.004337919410318136,
+      "epoch": 15.098920863309353,
+      "step": 16790
+    },
+    {
+      "epoch": 15.098920863309353,
+      "step": 16790,
+      "torque_loss": 0.1020972952246666
+    },
+    {
+      "epoch": 15.098920863309353,
+      "force_loss": 0.005252324044704437,
+      "step": 16790
+    },
+    {
+      "epoch": 15.107913669064748,
+      "grad_norm": 0.19795122742652893,
+      "learning_rate": 8.622315989296432e-05,
+      "loss": 0.0183,
+      "step": 16800
+    },
+    {
+      "action_loss": 0.0032609812915325165,
+      "epoch": 15.107913669064748,
+      "step": 16800
+    },
+    {
+      "epoch": 15.107913669064748,
+      "step": 16800,
+      "torque_loss": 0.11267898231744766
+    },
+    {
+      "epoch": 15.107913669064748,
+      "force_loss": 0.003354280022904277,
+      "step": 16800
+    },
+    {
+      "epoch": 15.116906474820144,
+      "grad_norm": 0.24201449751853943,
+      "learning_rate": 8.62041584081799e-05,
+      "loss": 0.0228,
+      "step": 16810
+    },
+    {
+      "action_loss": 0.007624324411153793,
+      "epoch": 15.116906474820144,
+      "step": 16810
+    },
+    {
+      "epoch": 15.116906474820144,
+      "step": 16810,
+      "torque_loss": 0.09145904332399368
+    },
+    {
+      "epoch": 15.116906474820144,
+      "force_loss": 0.011103170923888683,
+      "step": 16810
+    },
+    {
+      "epoch": 15.12589928057554,
+      "grad_norm": 0.4952791631221771,
+      "learning_rate": 8.618514592552807e-05,
+      "loss": 0.0233,
+      "step": 16820
+    },
+    {
+      "action_loss": 0.005998767446726561,
+      "epoch": 15.12589928057554,
+      "step": 16820
+    },
+    {
+      "epoch": 15.12589928057554,
+      "step": 16820,
+      "torque_loss": 0.10756029933691025
+    },
+    {
+      "epoch": 15.12589928057554,
+      "force_loss": 0.006025247275829315,
+      "step": 16820
+    },
+    {
+      "epoch": 15.134892086330936,
+      "grad_norm": 0.3274683654308319,
+      "learning_rate": 8.616612245078431e-05,
+      "loss": 0.023,
+      "step": 16830
+    },
+    {
+      "action_loss": 0.014972631819546223,
+      "epoch": 15.134892086330936,
+      "step": 16830
+    },
+    {
+      "epoch": 15.134892086330936,
+      "step": 16830,
+      "torque_loss": 0.12423815578222275
+    },
+    {
+      "epoch": 15.134892086330936,
+      "force_loss": 0.012586652301251888,
+      "step": 16830
+    },
+    {
+      "epoch": 15.14388489208633,
+      "grad_norm": 0.2552795708179474,
+      "learning_rate": 8.614708798972746e-05,
+      "loss": 0.0223,
+      "step": 16840
+    },
+    {
+      "action_loss": 0.009253703989088535,
+      "epoch": 15.14388489208633,
+      "step": 16840
+    },
+    {
+      "epoch": 15.14388489208633,
+      "step": 16840,
+      "torque_loss": 0.1225009635090828
+    },
+    {
+      "epoch": 15.14388489208633,
+      "force_loss": 0.006945467088371515,
+      "step": 16840
+    },
+    {
+      "epoch": 15.152877697841726,
+      "grad_norm": 0.402065247297287,
+      "learning_rate": 8.61280425481397e-05,
+      "loss": 0.0208,
+      "step": 16850
+    },
+    {
+      "action_loss": 0.013394013047218323,
+      "epoch": 15.152877697841726,
+      "step": 16850
+    },
+    {
+      "epoch": 15.152877697841726,
+      "step": 16850,
+      "torque_loss": 0.14408591389656067
+    },
+    {
+      "epoch": 15.152877697841726,
+      "force_loss": 0.016764340922236443,
+      "step": 16850
+    },
+    {
+      "epoch": 15.161870503597122,
+      "grad_norm": 0.5309786200523376,
+      "learning_rate": 8.61089861318065e-05,
+      "loss": 0.0232,
+      "step": 16860
+    },
+    {
+      "action_loss": 0.006734972819685936,
+      "epoch": 15.161870503597122,
+      "step": 16860
+    },
+    {
+      "epoch": 15.161870503597122,
+      "step": 16860,
+      "torque_loss": 0.11887051910161972
+    },
+    {
+      "epoch": 15.161870503597122,
+      "force_loss": 0.004631218034774065,
+      "step": 16860
+    },
+    {
+      "epoch": 15.170863309352518,
+      "grad_norm": 0.5137143731117249,
+      "learning_rate": 8.608991874651673e-05,
+      "loss": 0.019,
+      "step": 16870
+    },
+    {
+      "action_loss": 0.018816007301211357,
+      "epoch": 15.170863309352518,
+      "step": 16870
+    },
+    {
+      "epoch": 15.170863309352518,
+      "step": 16870,
+      "torque_loss": 0.16703300178050995
+    },
+    {
+      "epoch": 15.170863309352518,
+      "force_loss": 0.010347138158977032,
+      "step": 16870
+    },
+    {
+      "epoch": 15.179856115107913,
+      "grad_norm": 0.2642015814781189,
+      "learning_rate": 8.607084039806255e-05,
+      "loss": 0.0236,
+      "step": 16880
+    },
+    {
+      "action_loss": 0.014908102340996265,
+      "epoch": 15.179856115107913,
+      "step": 16880
+    },
+    {
+      "epoch": 15.179856115107913,
+      "step": 16880,
+      "torque_loss": 0.1228601410984993
+    },
+    {
+      "epoch": 15.179856115107913,
+      "force_loss": 0.01598764955997467,
+      "step": 16880
+    },
+    {
+      "epoch": 15.18884892086331,
+      "grad_norm": 0.6216331124305725,
+      "learning_rate": 8.605175109223944e-05,
+      "loss": 0.0218,
+      "step": 16890
+    },
+    {
+      "action_loss": 0.012915591709315777,
+      "epoch": 15.18884892086331,
+      "step": 16890
+    },
+    {
+      "epoch": 15.18884892086331,
+      "step": 16890,
+      "torque_loss": 0.17278863489627838
+    },
+    {
+      "epoch": 15.18884892086331,
+      "force_loss": 0.01144468691200018,
+      "step": 16890
+    },
+    {
+      "epoch": 15.197841726618705,
+      "grad_norm": 0.5609726905822754,
+      "learning_rate": 8.603265083484624e-05,
+      "loss": 0.022,
+      "step": 16900
+    },
+    {
+      "action_loss": 0.004875303711742163,
+      "epoch": 15.197841726618705,
+      "step": 16900
+    },
+    {
+      "epoch": 15.197841726618705,
+      "step": 16900,
+      "torque_loss": 0.14107267558574677
+    },
+    {
+      "epoch": 15.197841726618705,
+      "force_loss": 0.005135141778737307,
+      "step": 16900
+    },
+    {
+      "epoch": 15.206834532374101,
+      "grad_norm": 0.18485307693481445,
+      "learning_rate": 8.60135396316851e-05,
+      "loss": 0.0204,
+      "step": 16910
+    },
+    {
+      "action_loss": 0.006360320840030909,
+      "epoch": 15.206834532374101,
+      "step": 16910
+    },
+    {
+      "epoch": 15.206834532374101,
+      "step": 16910,
+      "torque_loss": 0.11764228343963623
+    },
+    {
+      "epoch": 15.206834532374101,
+      "force_loss": 0.017634594812989235,
+      "step": 16910
+    },
+    {
+      "epoch": 15.215827338129497,
+      "grad_norm": 0.4924125075340271,
+      "learning_rate": 8.599441748856152e-05,
+      "loss": 0.02,
+      "step": 16920
+    },
+    {
+      "action_loss": 0.011829804629087448,
+      "epoch": 15.215827338129497,
+      "step": 16920
+    },
+    {
+      "epoch": 15.215827338129497,
+      "step": 16920,
+      "torque_loss": 0.12550397217273712
+    },
+    {
+      "epoch": 15.215827338129497,
+      "force_loss": 0.014131136238574982,
+      "step": 16920
+    },
+    {
+      "epoch": 15.224820143884893,
+      "grad_norm": 0.31032586097717285,
+      "learning_rate": 8.597528441128427e-05,
+      "loss": 0.0235,
+      "step": 16930
+    },
+    {
+      "action_loss": 0.004942980129271746,
+      "epoch": 15.224820143884893,
+      "step": 16930
+    },
+    {
+      "epoch": 15.224820143884893,
+      "step": 16930,
+      "torque_loss": 0.09666839987039566
+    },
+    {
+      "epoch": 15.224820143884893,
+      "force_loss": 0.003027847735211253,
+      "step": 16930
+    },
+    {
+      "epoch": 15.233812949640289,
+      "grad_norm": 0.2096644639968872,
+      "learning_rate": 8.595614040566549e-05,
+      "loss": 0.021,
+      "step": 16940
+    },
+    {
+      "action_loss": 0.002870767144486308,
+      "epoch": 15.233812949640289,
+      "step": 16940
+    },
+    {
+      "epoch": 15.233812949640289,
+      "step": 16940,
+      "torque_loss": 0.1073073148727417
+    },
+    {
+      "epoch": 15.233812949640289,
+      "force_loss": 0.003411786863580346,
+      "step": 16940
+    },
+    {
+      "epoch": 15.242805755395683,
+      "grad_norm": 0.4573443830013275,
+      "learning_rate": 8.593698547752063e-05,
+      "loss": 0.0198,
+      "step": 16950
+    },
+    {
+      "action_loss": 0.004562361165881157,
+      "epoch": 15.242805755395683,
+      "step": 16950
+    },
+    {
+      "epoch": 15.242805755395683,
+      "step": 16950,
+      "torque_loss": 0.11304136365652084
+    },
+    {
+      "epoch": 15.242805755395683,
+      "force_loss": 0.006221102550625801,
+      "step": 16950
+    },
+    {
+      "epoch": 15.251798561151078,
+      "grad_norm": 0.461597740650177,
+      "learning_rate": 8.591781963266843e-05,
+      "loss": 0.0349,
+      "step": 16960
+    },
+    {
+      "action_loss": 0.003422211855649948,
+      "epoch": 15.251798561151078,
+      "step": 16960
+    },
+    {
+      "epoch": 15.251798561151078,
+      "step": 16960,
+      "torque_loss": 0.12643198668956757
+    },
+    {
+      "epoch": 15.251798561151078,
+      "force_loss": 0.0030089672654867172,
+      "step": 16960
+    },
+    {
+      "epoch": 15.260791366906474,
+      "grad_norm": 0.422500878572464,
+      "learning_rate": 8.5898642876931e-05,
+      "loss": 0.0208,
+      "step": 16970
+    },
+    {
+      "action_loss": 0.025540227070450783,
+      "epoch": 15.260791366906474,
+      "step": 16970
+    },
+    {
+      "epoch": 15.260791366906474,
+      "step": 16970,
+      "torque_loss": 0.12221430987119675
+    },
+    {
+      "epoch": 15.260791366906474,
+      "force_loss": 0.026424996554851532,
+      "step": 16970
+    },
+    {
+      "epoch": 15.26978417266187,
+      "grad_norm": 0.4246219992637634,
+      "learning_rate": 8.587945521613369e-05,
+      "loss": 0.0317,
+      "step": 16980
+    },
+    {
+      "action_loss": 0.00594263756647706,
+      "epoch": 15.26978417266187,
+      "step": 16980
+    },
+    {
+      "epoch": 15.26978417266187,
+      "step": 16980,
+      "torque_loss": 0.10616742819547653
+    },
+    {
+      "epoch": 15.26978417266187,
+      "force_loss": 0.003391991602256894,
+      "step": 16980
+    },
+    {
+      "epoch": 15.278776978417266,
+      "grad_norm": 0.23231947422027588,
+      "learning_rate": 8.586025665610524e-05,
+      "loss": 0.0203,
+      "step": 16990
+    },
+    {
+      "action_loss": 0.011393420398235321,
+      "epoch": 15.278776978417266,
+      "step": 16990
+    },
+    {
+      "epoch": 15.278776978417266,
+      "step": 16990,
+      "torque_loss": 0.12851528823375702
+    },
+    {
+      "epoch": 15.278776978417266,
+      "force_loss": 0.009426434524357319,
+      "step": 16990
+    },
+    {
+      "epoch": 15.287769784172662,
+      "grad_norm": 0.4100874066352844,
+      "learning_rate": 8.584104720267765e-05,
+      "loss": 0.0228,
+      "step": 17000
+    },
+    {
+      "action_loss": 0.010372314602136612,
+      "epoch": 15.287769784172662,
+      "step": 17000
+    },
+    {
+      "epoch": 15.287769784172662,
+      "step": 17000,
+      "torque_loss": 0.10932165384292603
+    },
+    {
+      "epoch": 15.287769784172662,
+      "force_loss": 0.013693836517632008,
+      "step": 17000
+    },
+    {
+      "epoch": 15.296762589928058,
+      "grad_norm": 0.256041020154953,
+      "learning_rate": 8.582182686168625e-05,
+      "loss": 0.0242,
+      "step": 17010
+    },
+    {
+      "action_loss": 0.00900339987128973,
+      "epoch": 15.296762589928058,
+      "step": 17010
+    },
+    {
+      "epoch": 15.296762589928058,
+      "step": 17010,
+      "torque_loss": 0.12491494417190552
+    },
+    {
+      "epoch": 15.296762589928058,
+      "force_loss": 0.009845726191997528,
+      "step": 17010
+    },
+    {
+      "epoch": 15.305755395683454,
+      "grad_norm": 0.2653682827949524,
+      "learning_rate": 8.580259563896967e-05,
+      "loss": 0.0248,
+      "step": 17020
+    },
+    {
+      "action_loss": 0.008330421522259712,
+      "epoch": 15.305755395683454,
+      "step": 17020
+    },
+    {
+      "epoch": 15.305755395683454,
+      "step": 17020,
+      "torque_loss": 0.17204131186008453
+    },
+    {
+      "epoch": 15.305755395683454,
+      "force_loss": 0.006588909775018692,
+      "step": 17020
+    },
+    {
+      "epoch": 15.31474820143885,
+      "grad_norm": 0.4202973246574402,
+      "learning_rate": 8.578335354036983e-05,
+      "loss": 0.0238,
+      "step": 17030
+    },
+    {
+      "action_loss": 0.014478645287454128,
+      "epoch": 15.31474820143885,
+      "step": 17030
+    },
+    {
+      "epoch": 15.31474820143885,
+      "step": 17030,
+      "torque_loss": 0.10385163873434067
+    },
+    {
+      "epoch": 15.31474820143885,
+      "force_loss": 0.008455420844256878,
+      "step": 17030
+    },
+    {
+      "epoch": 15.323741007194245,
+      "grad_norm": 0.27224811911582947,
+      "learning_rate": 8.576410057173201e-05,
+      "loss": 0.0231,
+      "step": 17040
+    },
+    {
+      "action_loss": 0.004931575618684292,
+      "epoch": 15.323741007194245,
+      "step": 17040
+    },
+    {
+      "epoch": 15.323741007194245,
+      "step": 17040,
+      "torque_loss": 0.18013717234134674
+    },
+    {
+      "epoch": 15.323741007194245,
+      "force_loss": 0.006074054166674614,
+      "step": 17040
+    },
+    {
+      "epoch": 15.332733812949641,
+      "grad_norm": 0.31736278533935547,
+      "learning_rate": 8.574483673890474e-05,
+      "loss": 0.0209,
+      "step": 17050
+    },
+    {
+      "action_loss": 0.003609471023082733,
+      "epoch": 15.332733812949641,
+      "step": 17050
+    },
+    {
+      "epoch": 15.332733812949641,
+      "step": 17050,
+      "torque_loss": 0.1309669464826584
+    },
+    {
+      "epoch": 15.332733812949641,
+      "force_loss": 0.005024902056902647,
+      "step": 17050
+    },
+    {
+      "epoch": 15.341726618705035,
+      "grad_norm": 0.38658303022384644,
+      "learning_rate": 8.572556204773983e-05,
+      "loss": 0.0206,
+      "step": 17060
+    },
+    {
+      "action_loss": 0.0042090336792171,
+      "epoch": 15.341726618705035,
+      "step": 17060
+    },
+    {
+      "epoch": 15.341726618705035,
+      "step": 17060,
+      "torque_loss": 0.14667350053787231
+    },
+    {
+      "epoch": 15.341726618705035,
+      "force_loss": 0.00515960855409503,
+      "step": 17060
+    },
+    {
+      "epoch": 15.350719424460431,
+      "grad_norm": 0.5411636233329773,
+      "learning_rate": 8.570627650409246e-05,
+      "loss": 0.0201,
+      "step": 17070
+    },
+    {
+      "action_loss": 0.002961714519187808,
+      "epoch": 15.350719424460431,
+      "step": 17070
+    },
+    {
+      "epoch": 15.350719424460431,
+      "step": 17070,
+      "torque_loss": 0.12429994344711304
+    },
+    {
+      "epoch": 15.350719424460431,
+      "force_loss": 0.003751396434381604,
+      "step": 17070
+    },
+    {
+      "epoch": 15.359712230215827,
+      "grad_norm": 0.20276440680027008,
+      "learning_rate": 8.568698011382107e-05,
+      "loss": 0.0199,
+      "step": 17080
+    },
+    {
+      "action_loss": 0.022901957854628563,
+      "epoch": 15.359712230215827,
+      "step": 17080
+    },
+    {
+      "epoch": 15.359712230215827,
+      "step": 17080,
+      "torque_loss": 0.16564421355724335
+    },
+    {
+      "epoch": 15.359712230215827,
+      "force_loss": 0.02848103828728199,
+      "step": 17080
+    },
+    {
+      "epoch": 15.368705035971223,
+      "grad_norm": 0.4422018826007843,
+      "learning_rate": 8.566767288278738e-05,
+      "loss": 0.0265,
+      "step": 17090
+    },
+    {
+      "action_loss": 0.0053179445676505566,
+      "epoch": 15.368705035971223,
+      "step": 17090
+    },
+    {
+      "epoch": 15.368705035971223,
+      "step": 17090,
+      "torque_loss": 0.14988398551940918
+    },
+    {
+      "epoch": 15.368705035971223,
+      "force_loss": 0.0034148860722780228,
+      "step": 17090
+    },
+    {
+      "epoch": 15.377697841726619,
+      "grad_norm": 0.29468706250190735,
+      "learning_rate": 8.56483548168564e-05,
+      "loss": 0.0216,
+      "step": 17100
+    },
+    {
+      "action_loss": 0.006086260080337524,
+      "epoch": 15.377697841726619,
+      "step": 17100
+    },
+    {
+      "epoch": 15.377697841726619,
+      "step": 17100,
+      "torque_loss": 0.08516178280115128
+    },
+    {
+      "epoch": 15.377697841726619,
+      "force_loss": 0.00398709112778306,
+      "step": 17100
+    },
+    {
+      "epoch": 15.386690647482014,
+      "grad_norm": 0.3179362416267395,
+      "learning_rate": 8.562902592189648e-05,
+      "loss": 0.0212,
+      "step": 17110
+    },
+    {
+      "action_loss": 0.0031117647886276245,
+      "epoch": 15.386690647482014,
+      "step": 17110
+    },
+    {
+      "epoch": 15.386690647482014,
+      "step": 17110,
+      "torque_loss": 0.1303016096353531
+    },
+    {
+      "epoch": 15.386690647482014,
+      "force_loss": 0.004306287970393896,
+      "step": 17110
+    },
+    {
+      "epoch": 15.39568345323741,
+      "grad_norm": 0.18887697160243988,
+      "learning_rate": 8.560968620377921e-05,
+      "loss": 0.021,
+      "step": 17120
+    },
+    {
+      "action_loss": 0.0083672646433115,
+      "epoch": 15.39568345323741,
+      "step": 17120
+    },
+    {
+      "epoch": 15.39568345323741,
+      "step": 17120,
+      "torque_loss": 0.15263231098651886
+    },
+    {
+      "epoch": 15.39568345323741,
+      "force_loss": 0.006824602838605642,
+      "step": 17120
+    },
+    {
+      "epoch": 15.404676258992806,
+      "grad_norm": 0.5132378935813904,
+      "learning_rate": 8.559033566837951e-05,
+      "loss": 0.0238,
+      "step": 17130
+    },
+    {
+      "action_loss": 0.0025514133740216494,
+      "epoch": 15.404676258992806,
+      "step": 17130
+    },
+    {
+      "epoch": 15.404676258992806,
+      "step": 17130,
+      "torque_loss": 0.1241193637251854
+    },
+    {
+      "epoch": 15.404676258992806,
+      "force_loss": 0.002342240186408162,
+      "step": 17130
+    },
+    {
+      "epoch": 15.413669064748202,
+      "grad_norm": 0.5962299704551697,
+      "learning_rate": 8.557097432157551e-05,
+      "loss": 0.0249,
+      "step": 17140
+    },
+    {
+      "action_loss": 0.007867868058383465,
+      "epoch": 15.413669064748202,
+      "step": 17140
+    },
+    {
+      "epoch": 15.413669064748202,
+      "step": 17140,
+      "torque_loss": 0.12188458442687988
+    },
+    {
+      "epoch": 15.413669064748202,
+      "force_loss": 0.008350138552486897,
+      "step": 17140
+    },
+    {
+      "epoch": 15.422661870503598,
+      "grad_norm": 0.3489937484264374,
+      "learning_rate": 8.555160216924872e-05,
+      "loss": 0.0204,
+      "step": 17150
+    },
+    {
+      "action_loss": 0.013648436404764652,
+      "epoch": 15.422661870503598,
+      "step": 17150
+    },
+    {
+      "epoch": 15.422661870503598,
+      "step": 17150,
+      "torque_loss": 0.11687064170837402
+    },
+    {
+      "epoch": 15.422661870503598,
+      "force_loss": 0.011117999441921711,
+      "step": 17150
+    },
+    {
+      "epoch": 15.431654676258994,
+      "grad_norm": 0.6030886769294739,
+      "learning_rate": 8.55322192172839e-05,
+      "loss": 0.0228,
+      "step": 17160
+    },
+    {
+      "action_loss": 0.004624569788575172,
+      "epoch": 15.431654676258994,
+      "step": 17160
+    },
+    {
+      "epoch": 15.431654676258994,
+      "step": 17160,
+      "torque_loss": 0.1281377524137497
+    },
+    {
+      "epoch": 15.431654676258994,
+      "force_loss": 0.0034989865962415934,
+      "step": 17160
+    },
+    {
+      "epoch": 15.440647482014388,
+      "grad_norm": 0.2881278693675995,
+      "learning_rate": 8.551282547156902e-05,
+      "loss": 0.02,
+      "step": 17170
+    },
+    {
+      "action_loss": 0.010662811808288097,
+      "epoch": 15.440647482014388,
+      "step": 17170
+    },
+    {
+      "epoch": 15.440647482014388,
+      "step": 17170,
+      "torque_loss": 0.16588138043880463
+    },
+    {
+      "epoch": 15.440647482014388,
+      "force_loss": 0.007802148815244436,
+      "step": 17170
+    },
+    {
+      "epoch": 15.449640287769784,
+      "grad_norm": 0.8668406009674072,
+      "learning_rate": 8.549342093799544e-05,
+      "loss": 0.0236,
+      "step": 17180
+    },
+    {
+      "action_loss": 0.009070406667888165,
+      "epoch": 15.449640287769784,
+      "step": 17180
+    },
+    {
+      "epoch": 15.449640287769784,
+      "step": 17180,
+      "torque_loss": 0.15621691942214966
+    },
+    {
+      "epoch": 15.449640287769784,
+      "force_loss": 0.017987458035349846,
+      "step": 17180
+    },
+    {
+      "epoch": 15.45863309352518,
+      "grad_norm": 0.33376264572143555,
+      "learning_rate": 8.547400562245773e-05,
+      "loss": 0.0252,
+      "step": 17190
+    },
+    {
+      "action_loss": 0.019972102716565132,
+      "epoch": 15.45863309352518,
+      "step": 17190
+    },
+    {
+      "epoch": 15.45863309352518,
+      "step": 17190,
+      "torque_loss": 0.10239126533269882
+    },
+    {
+      "epoch": 15.45863309352518,
+      "force_loss": 0.014145704917609692,
+      "step": 17190
+    },
+    {
+      "epoch": 15.467625899280575,
+      "grad_norm": 0.35148489475250244,
+      "learning_rate": 8.545457953085374e-05,
+      "loss": 0.0266,
+      "step": 17200
+    },
+    {
+      "action_loss": 0.005947902332991362,
+      "epoch": 15.467625899280575,
+      "step": 17200
+    },
+    {
+      "epoch": 15.467625899280575,
+      "step": 17200,
+      "torque_loss": 0.11391753703355789
+    },
+    {
+      "epoch": 15.467625899280575,
+      "force_loss": 0.005423885304480791,
+      "step": 17200
+    },
+    {
+      "epoch": 15.476618705035971,
+      "grad_norm": 0.25836482644081116,
+      "learning_rate": 8.543514266908463e-05,
+      "loss": 0.0243,
+      "step": 17210
+    },
+    {
+      "action_loss": 0.00973950419574976,
+      "epoch": 15.476618705035971,
+      "step": 17210
+    },
+    {
+      "epoch": 15.476618705035971,
+      "step": 17210,
+      "torque_loss": 0.12382733821868896
+    },
+    {
+      "epoch": 15.476618705035971,
+      "force_loss": 0.004502378404140472,
+      "step": 17210
+    },
+    {
+      "epoch": 15.485611510791367,
+      "grad_norm": 0.41935157775878906,
+      "learning_rate": 8.541569504305478e-05,
+      "loss": 0.0257,
+      "step": 17220
+    },
+    {
+      "action_loss": 0.004793783184140921,
+      "epoch": 15.485611510791367,
+      "step": 17220
+    },
+    {
+      "epoch": 15.485611510791367,
+      "step": 17220,
+      "torque_loss": 0.11694937944412231
+    },
+    {
+      "epoch": 15.485611510791367,
+      "force_loss": 0.008017126470804214,
+      "step": 17220
+    },
+    {
+      "epoch": 15.494604316546763,
+      "grad_norm": 0.3241329491138458,
+      "learning_rate": 8.539623665867187e-05,
+      "loss": 0.0237,
+      "step": 17230
+    },
+    {
+      "action_loss": 0.03657834231853485,
+      "epoch": 15.494604316546763,
+      "step": 17230
+    },
+    {
+      "epoch": 15.494604316546763,
+      "step": 17230,
+      "torque_loss": 0.13412193953990936
+    },
+    {
+      "epoch": 15.494604316546763,
+      "force_loss": 0.02088378369808197,
+      "step": 17230
+    },
+    {
+      "epoch": 15.503597122302159,
+      "grad_norm": 0.3684811592102051,
+      "learning_rate": 8.537676752184685e-05,
+      "loss": 0.0256,
+      "step": 17240
+    },
+    {
+      "action_loss": 0.006749139633029699,
+      "epoch": 15.503597122302159,
+      "step": 17240
+    },
+    {
+      "epoch": 15.503597122302159,
+      "step": 17240,
+      "torque_loss": 0.19073623418807983
+    },
+    {
+      "epoch": 15.503597122302159,
+      "force_loss": 0.006471940781921148,
+      "step": 17240
+    },
+    {
+      "epoch": 15.512589928057555,
+      "grad_norm": 0.46356040239334106,
+      "learning_rate": 8.53572876384939e-05,
+      "loss": 0.0223,
+      "step": 17250
+    },
+    {
+      "action_loss": 0.005972277373075485,
+      "epoch": 15.512589928057555,
+      "step": 17250
+    },
+    {
+      "epoch": 15.512589928057555,
+      "step": 17250,
+      "torque_loss": 0.11475479602813721
+    },
+    {
+      "epoch": 15.512589928057555,
+      "force_loss": 0.006718062330037355,
+      "step": 17250
+    },
+    {
+      "epoch": 15.52158273381295,
+      "grad_norm": 0.5765788555145264,
+      "learning_rate": 8.533779701453056e-05,
+      "loss": 0.0233,
+      "step": 17260
+    },
+    {
+      "action_loss": 0.026722656562924385,
+      "epoch": 15.52158273381295,
+      "step": 17260
+    },
+    {
+      "epoch": 15.52158273381295,
+      "step": 17260,
+      "torque_loss": 0.159148171544075
+    },
+    {
+      "epoch": 15.52158273381295,
+      "force_loss": 0.02406766451895237,
+      "step": 17260
+    },
+    {
+      "epoch": 15.530575539568346,
+      "grad_norm": 0.4047594368457794,
+      "learning_rate": 8.53182956558775e-05,
+      "loss": 0.0241,
+      "step": 17270
+    },
+    {
+      "action_loss": 0.003972392529249191,
+      "epoch": 15.530575539568346,
+      "step": 17270
+    },
+    {
+      "epoch": 15.530575539568346,
+      "step": 17270,
+      "torque_loss": 0.1654411107301712
+    },
+    {
+      "epoch": 15.530575539568346,
+      "force_loss": 0.004558879882097244,
+      "step": 17270
+    },
+    {
+      "epoch": 15.53956834532374,
+      "grad_norm": 0.2250213921070099,
+      "learning_rate": 8.529878356845877e-05,
+      "loss": 0.021,
+      "step": 17280
+    },
+    {
+      "action_loss": 0.006601112429052591,
+      "epoch": 15.53956834532374,
+      "step": 17280
+    },
+    {
+      "epoch": 15.53956834532374,
+      "step": 17280,
+      "torque_loss": 0.13410568237304688
+    },
+    {
+      "epoch": 15.53956834532374,
+      "force_loss": 0.01160566508769989,
+      "step": 17280
+    },
+    {
+      "epoch": 15.548561151079136,
+      "grad_norm": 0.6944681406021118,
+      "learning_rate": 8.527926075820158e-05,
+      "loss": 0.0254,
+      "step": 17290
+    },
+    {
+      "action_loss": 0.0055807214230299,
+      "epoch": 15.548561151079136,
+      "step": 17290
+    },
+    {
+      "epoch": 15.548561151079136,
+      "step": 17290,
+      "torque_loss": 0.08707667142152786
+    },
+    {
+      "epoch": 15.548561151079136,
+      "force_loss": 0.003927208017557859,
+      "step": 17290
+    },
+    {
+      "epoch": 15.557553956834532,
+      "grad_norm": 0.6500819325447083,
+      "learning_rate": 8.525972723103648e-05,
+      "loss": 0.0251,
+      "step": 17300
+    },
+    {
+      "action_loss": 0.004456877708435059,
+      "epoch": 15.557553956834532,
+      "step": 17300
+    },
+    {
+      "epoch": 15.557553956834532,
+      "step": 17300,
+      "torque_loss": 0.1163586750626564
+    },
+    {
+      "epoch": 15.557553956834532,
+      "force_loss": 0.003028958337381482,
+      "step": 17300
+    },
+    {
+      "epoch": 15.566546762589928,
+      "grad_norm": 0.3894810378551483,
+      "learning_rate": 8.524018299289722e-05,
+      "loss": 0.0207,
+      "step": 17310
+    },
+    {
+      "action_loss": 0.011233396828174591,
+      "epoch": 15.566546762589928,
+      "step": 17310
+    },
+    {
+      "epoch": 15.566546762589928,
+      "step": 17310,
+      "torque_loss": 0.14090825617313385
+    },
+    {
+      "epoch": 15.566546762589928,
+      "force_loss": 0.006974514573812485,
+      "step": 17310
+    },
+    {
+      "epoch": 15.575539568345324,
+      "grad_norm": 0.4652196764945984,
+      "learning_rate": 8.522062804972083e-05,
+      "loss": 0.0241,
+      "step": 17320
+    },
+    {
+      "action_loss": 0.007066812366247177,
+      "epoch": 15.575539568345324,
+      "step": 17320
+    },
+    {
+      "epoch": 15.575539568345324,
+      "step": 17320,
+      "torque_loss": 0.10062506794929504
+    },
+    {
+      "epoch": 15.575539568345324,
+      "force_loss": 0.006861022207885981,
+      "step": 17320
+    },
+    {
+      "epoch": 15.58453237410072,
+      "grad_norm": 0.35327690839767456,
+      "learning_rate": 8.520106240744759e-05,
+      "loss": 0.022,
+      "step": 17330
+    },
+    {
+      "action_loss": 0.005429770797491074,
+      "epoch": 15.58453237410072,
+      "step": 17330
+    },
+    {
+      "epoch": 15.58453237410072,
+      "step": 17330,
+      "torque_loss": 0.12060769647359848
+    },
+    {
+      "epoch": 15.58453237410072,
+      "force_loss": 0.0049434867687523365,
+      "step": 17330
+    },
+    {
+      "epoch": 15.593525179856115,
+      "grad_norm": 0.4524448812007904,
+      "learning_rate": 8.518148607202102e-05,
+      "loss": 0.0259,
+      "step": 17340
+    },
+    {
+      "action_loss": 0.0075156292878091335,
+      "epoch": 15.593525179856115,
+      "step": 17340
+    },
+    {
+      "epoch": 15.593525179856115,
+      "step": 17340,
+      "torque_loss": 0.1567445546388626
+    },
+    {
+      "epoch": 15.593525179856115,
+      "force_loss": 0.004765639081597328,
+      "step": 17340
+    },
+    {
+      "epoch": 15.602517985611511,
+      "grad_norm": 0.45095622539520264,
+      "learning_rate": 8.51618990493879e-05,
+      "loss": 0.0254,
+      "step": 17350
+    },
+    {
+      "action_loss": 0.012920449487864971,
+      "epoch": 15.602517985611511,
+      "step": 17350
+    },
+    {
+      "epoch": 15.602517985611511,
+      "step": 17350,
+      "torque_loss": 0.17410211265087128
+    },
+    {
+      "epoch": 15.602517985611511,
+      "force_loss": 0.006048997398465872,
+      "step": 17350
+    },
+    {
+      "epoch": 15.611510791366907,
+      "grad_norm": 0.42172640562057495,
+      "learning_rate": 8.514230134549823e-05,
+      "loss": 0.0279,
+      "step": 17360
+    },
+    {
+      "action_loss": 0.013524171896278858,
+      "epoch": 15.611510791366907,
+      "step": 17360
+    },
+    {
+      "epoch": 15.611510791366907,
+      "step": 17360,
+      "torque_loss": 0.159952774643898
+    },
+    {
+      "epoch": 15.611510791366907,
+      "force_loss": 0.013305045664310455,
+      "step": 17360
+    },
+    {
+      "epoch": 15.620503597122303,
+      "grad_norm": 0.4205578863620758,
+      "learning_rate": 8.51226929663053e-05,
+      "loss": 0.0236,
+      "step": 17370
+    },
+    {
+      "action_loss": 0.012213137932121754,
+      "epoch": 15.620503597122303,
+      "step": 17370
+    },
+    {
+      "epoch": 15.620503597122303,
+      "step": 17370,
+      "torque_loss": 0.12412973493337631
+    },
+    {
+      "epoch": 15.620503597122303,
+      "force_loss": 0.020799381658434868,
+      "step": 17370
+    },
+    {
+      "epoch": 15.629496402877697,
+      "grad_norm": 0.7652576565742493,
+      "learning_rate": 8.51030739177656e-05,
+      "loss": 0.0225,
+      "step": 17380
+    },
+    {
+      "action_loss": 0.0036580178420990705,
+      "epoch": 15.629496402877697,
+      "step": 17380
+    },
+    {
+      "epoch": 15.629496402877697,
+      "step": 17380,
+      "torque_loss": 0.08535952121019363
+    },
+    {
+      "epoch": 15.629496402877697,
+      "force_loss": 0.004635924007743597,
+      "step": 17380
+    },
+    {
+      "epoch": 15.638489208633093,
+      "grad_norm": 0.26303476095199585,
+      "learning_rate": 8.508344420583889e-05,
+      "loss": 0.0247,
+      "step": 17390
+    },
+    {
+      "action_loss": 0.009836475364863873,
+      "epoch": 15.638489208633093,
+      "step": 17390
+    },
+    {
+      "epoch": 15.638489208633093,
+      "step": 17390,
+      "torque_loss": 0.12285927683115005
+    },
+    {
+      "epoch": 15.638489208633093,
+      "force_loss": 0.014383415691554546,
+      "step": 17390
+    },
+    {
+      "epoch": 15.647482014388489,
+      "grad_norm": 0.26400992274284363,
+      "learning_rate": 8.506380383648816e-05,
+      "loss": 0.0221,
+      "step": 17400
+    },
+    {
+      "action_loss": 0.006597287952899933,
+      "epoch": 15.647482014388489,
+      "step": 17400
+    },
+    {
+      "epoch": 15.647482014388489,
+      "step": 17400,
+      "torque_loss": 0.13825170695781708
+    },
+    {
+      "epoch": 15.647482014388489,
+      "force_loss": 0.004263477399945259,
+      "step": 17400
+    },
+    {
+      "epoch": 15.656474820143885,
+      "grad_norm": 0.3148339092731476,
+      "learning_rate": 8.504415281567963e-05,
+      "loss": 0.0214,
+      "step": 17410
+    },
+    {
+      "action_loss": 0.004615604877471924,
+      "epoch": 15.656474820143885,
+      "step": 17410
+    },
+    {
+      "epoch": 15.656474820143885,
+      "step": 17410,
+      "torque_loss": 0.10837849229574203
+    },
+    {
+      "epoch": 15.656474820143885,
+      "force_loss": 0.0037073504645377398,
+      "step": 17410
+    },
+    {
+      "epoch": 15.66546762589928,
+      "grad_norm": 0.3267367482185364,
+      "learning_rate": 8.502449114938275e-05,
+      "loss": 0.0212,
+      "step": 17420
+    },
+    {
+      "action_loss": 0.0038669221103191376,
+      "epoch": 15.66546762589928,
+      "step": 17420
+    },
+    {
+      "epoch": 15.66546762589928,
+      "step": 17420,
+      "torque_loss": 0.11064451187849045
+    },
+    {
+      "epoch": 15.66546762589928,
+      "force_loss": 0.0056570447050035,
+      "step": 17420
+    },
+    {
+      "epoch": 15.674460431654676,
+      "grad_norm": 0.7484758496284485,
+      "learning_rate": 8.500481884357025e-05,
+      "loss": 0.0248,
+      "step": 17430
+    },
+    {
+      "action_loss": 0.008798401802778244,
+      "epoch": 15.674460431654676,
+      "step": 17430
+    },
+    {
+      "epoch": 15.674460431654676,
+      "step": 17430,
+      "torque_loss": 0.1286756694316864
+    },
+    {
+      "epoch": 15.674460431654676,
+      "force_loss": 0.00599707430228591,
+      "step": 17430
+    },
+    {
+      "epoch": 15.683453237410072,
+      "grad_norm": 0.7126136422157288,
+      "learning_rate": 8.498513590421801e-05,
+      "loss": 0.025,
+      "step": 17440
+    },
+    {
+      "action_loss": 0.009946263395249844,
+      "epoch": 15.683453237410072,
+      "step": 17440
+    },
+    {
+      "epoch": 15.683453237410072,
+      "step": 17440,
+      "torque_loss": 0.1469493955373764
+    },
+    {
+      "epoch": 15.683453237410072,
+      "force_loss": 0.022330351173877716,
+      "step": 17440
+    },
+    {
+      "epoch": 15.692446043165468,
+      "grad_norm": 0.4793100357055664,
+      "learning_rate": 8.496544233730522e-05,
+      "loss": 0.0247,
+      "step": 17450
+    },
+    {
+      "action_loss": 0.005858680699020624,
+      "epoch": 15.692446043165468,
+      "step": 17450
+    },
+    {
+      "epoch": 15.692446043165468,
+      "step": 17450,
+      "torque_loss": 0.12280535697937012
+    },
+    {
+      "epoch": 15.692446043165468,
+      "force_loss": 0.0036531102377921343,
+      "step": 17450
+    },
+    {
+      "epoch": 15.701438848920864,
+      "grad_norm": 0.31725069880485535,
+      "learning_rate": 8.494573814881426e-05,
+      "loss": 0.0234,
+      "step": 17460
+    },
+    {
+      "action_loss": 0.005597864743322134,
+      "epoch": 15.701438848920864,
+      "step": 17460
+    },
+    {
+      "epoch": 15.701438848920864,
+      "step": 17460,
+      "torque_loss": 0.14102140069007874
+    },
+    {
+      "epoch": 15.701438848920864,
+      "force_loss": 0.004687194246798754,
+      "step": 17460
+    },
+    {
+      "epoch": 15.71043165467626,
+      "grad_norm": 0.35858169198036194,
+      "learning_rate": 8.492602334473074e-05,
+      "loss": 0.0203,
+      "step": 17470
+    },
+    {
+      "action_loss": 0.02361586131155491,
+      "epoch": 15.71043165467626,
+      "step": 17470
+    },
+    {
+      "epoch": 15.71043165467626,
+      "step": 17470,
+      "torque_loss": 0.2368074655532837
+    },
+    {
+      "epoch": 15.71043165467626,
+      "force_loss": 0.020131416618824005,
+      "step": 17470
+    },
+    {
+      "epoch": 15.719424460431654,
+      "grad_norm": 0.43791642785072327,
+      "learning_rate": 8.49062979310435e-05,
+      "loss": 0.0259,
+      "step": 17480
+    },
+    {
+      "action_loss": 0.015914402902126312,
+      "epoch": 15.719424460431654,
+      "step": 17480
+    },
+    {
+      "epoch": 15.719424460431654,
+      "step": 17480,
+      "torque_loss": 0.1638699322938919
+    },
+    {
+      "epoch": 15.719424460431654,
+      "force_loss": 0.009551900438964367,
+      "step": 17480
+    },
+    {
+      "epoch": 15.72841726618705,
+      "grad_norm": 0.17880485951900482,
+      "learning_rate": 8.488656191374458e-05,
+      "loss": 0.0255,
+      "step": 17490
+    },
+    {
+      "action_loss": 0.005156724713742733,
+      "epoch": 15.72841726618705,
+      "step": 17490
+    },
+    {
+      "epoch": 15.72841726618705,
+      "step": 17490,
+      "torque_loss": 0.09796721488237381
+    },
+    {
+      "epoch": 15.72841726618705,
+      "force_loss": 0.00782252661883831,
+      "step": 17490
+    },
+    {
+      "epoch": 15.737410071942445,
+      "grad_norm": 0.3123587965965271,
+      "learning_rate": 8.48668152988293e-05,
+      "loss": 0.0228,
+      "step": 17500
+    },
+    {
+      "action_loss": 0.005358940456062555,
+      "epoch": 15.737410071942445,
+      "step": 17500
+    },
+    {
+      "epoch": 15.737410071942445,
+      "step": 17500,
+      "torque_loss": 0.15764956176280975
+    },
+    {
+      "epoch": 15.737410071942445,
+      "force_loss": 0.006400526966899633,
+      "step": 17500
+    },
+    {
+      "epoch": 15.746402877697841,
+      "grad_norm": 0.2304433435201645,
+      "learning_rate": 8.484705809229612e-05,
+      "loss": 0.0226,
+      "step": 17510
+    },
+    {
+      "action_loss": 0.015522944740951061,
+      "epoch": 15.746402877697841,
+      "step": 17510
+    },
+    {
+      "epoch": 15.746402877697841,
+      "step": 17510,
+      "torque_loss": 0.17891037464141846
+    },
+    {
+      "epoch": 15.746402877697841,
+      "force_loss": 0.02278672158718109,
+      "step": 17510
+    },
+    {
+      "epoch": 15.755395683453237,
+      "grad_norm": 0.9582626223564148,
+      "learning_rate": 8.482729030014677e-05,
+      "loss": 0.0263,
+      "step": 17520
+    },
+    {
+      "action_loss": 0.0069336616434156895,
+      "epoch": 15.755395683453237,
+      "step": 17520
+    },
+    {
+      "epoch": 15.755395683453237,
+      "step": 17520,
+      "torque_loss": 0.13514092564582825
+    },
+    {
+      "epoch": 15.755395683453237,
+      "force_loss": 0.004314844962209463,
+      "step": 17520
+    },
+    {
+      "epoch": 15.764388489208633,
+      "grad_norm": 0.3698276877403259,
+      "learning_rate": 8.48075119283862e-05,
+      "loss": 0.0224,
+      "step": 17530
+    },
+    {
+      "action_loss": 0.013405833393335342,
+      "epoch": 15.764388489208633,
+      "step": 17530
+    },
+    {
+      "epoch": 15.764388489208633,
+      "step": 17530,
+      "torque_loss": 0.13743506371974945
+    },
+    {
+      "epoch": 15.764388489208633,
+      "force_loss": 0.02497536689043045,
+      "step": 17530
+    },
+    {
+      "epoch": 15.773381294964029,
+      "grad_norm": 0.5154282450675964,
+      "learning_rate": 8.478772298302254e-05,
+      "loss": 0.0237,
+      "step": 17540
+    },
+    {
+      "action_loss": 0.0063224430195987225,
+      "epoch": 15.773381294964029,
+      "step": 17540
+    },
+    {
+      "epoch": 15.773381294964029,
+      "step": 17540,
+      "torque_loss": 0.13486050069332123
+    },
+    {
+      "epoch": 15.773381294964029,
+      "force_loss": 0.01241572480648756,
+      "step": 17540
+    },
+    {
+      "epoch": 15.782374100719425,
+      "grad_norm": 0.41657909750938416,
+      "learning_rate": 8.476792347006716e-05,
+      "loss": 0.0278,
+      "step": 17550
+    },
+    {
+      "action_loss": 0.004343951120972633,
+      "epoch": 15.782374100719425,
+      "step": 17550
+    },
+    {
+      "epoch": 15.782374100719425,
+      "step": 17550,
+      "torque_loss": 0.09667249768972397
+    },
+    {
+      "epoch": 15.782374100719425,
+      "force_loss": 0.004796222317963839,
+      "step": 17550
+    },
+    {
+      "epoch": 15.79136690647482,
+      "grad_norm": 0.5779966711997986,
+      "learning_rate": 8.474811339553462e-05,
+      "loss": 0.0245,
+      "step": 17560
+    },
+    {
+      "action_loss": 0.007796111050993204,
+      "epoch": 15.79136690647482,
+      "step": 17560
+    },
+    {
+      "epoch": 15.79136690647482,
+      "step": 17560,
+      "torque_loss": 0.12458092719316483
+    },
+    {
+      "epoch": 15.79136690647482,
+      "force_loss": 0.010667835362255573,
+      "step": 17560
+    },
+    {
+      "epoch": 15.800359712230216,
+      "grad_norm": 0.8348226547241211,
+      "learning_rate": 8.47282927654427e-05,
+      "loss": 0.0327,
+      "step": 17570
+    },
+    {
+      "action_loss": 0.004865535534918308,
+      "epoch": 15.800359712230216,
+      "step": 17570
+    },
+    {
+      "epoch": 15.800359712230216,
+      "step": 17570,
+      "torque_loss": 0.09535191208124161
+    },
+    {
+      "epoch": 15.800359712230216,
+      "force_loss": 0.005711412522941828,
+      "step": 17570
+    },
+    {
+      "epoch": 15.809352517985612,
+      "grad_norm": 0.4629124104976654,
+      "learning_rate": 8.470846158581238e-05,
+      "loss": 0.0225,
+      "step": 17580
+    },
+    {
+      "action_loss": 0.010356389917433262,
+      "epoch": 15.809352517985612,
+      "step": 17580
+    },
+    {
+      "epoch": 15.809352517985612,
+      "step": 17580,
+      "torque_loss": 0.13352568447589874
+    },
+    {
+      "epoch": 15.809352517985612,
+      "force_loss": 0.0201144777238369,
+      "step": 17580
+    },
+    {
+      "epoch": 15.818345323741006,
+      "grad_norm": 0.45410624146461487,
+      "learning_rate": 8.468861986266787e-05,
+      "loss": 0.0246,
+      "step": 17590
+    },
+    {
+      "action_loss": 0.020586155354976654,
+      "epoch": 15.818345323741006,
+      "step": 17590
+    },
+    {
+      "epoch": 15.818345323741006,
+      "step": 17590,
+      "torque_loss": 0.15210437774658203
+    },
+    {
+      "epoch": 15.818345323741006,
+      "force_loss": 0.02451533079147339,
+      "step": 17590
+    },
+    {
+      "epoch": 15.827338129496402,
+      "grad_norm": 0.2987782657146454,
+      "learning_rate": 8.466876760203654e-05,
+      "loss": 0.0255,
+      "step": 17600
+    },
+    {
+      "action_loss": 0.052640676498413086,
+      "epoch": 15.827338129496402,
+      "step": 17600
+    },
+    {
+      "epoch": 15.827338129496402,
+      "step": 17600,
+      "torque_loss": 0.21721559762954712
+    },
+    {
+      "epoch": 15.827338129496402,
+      "force_loss": 0.05710431933403015,
+      "step": 17600
+    },
+    {
+      "epoch": 15.836330935251798,
+      "grad_norm": 0.35147571563720703,
+      "learning_rate": 8.464890480994898e-05,
+      "loss": 0.0307,
+      "step": 17610
+    },
+    {
+      "action_loss": 0.006250986363738775,
+      "epoch": 15.836330935251798,
+      "step": 17610
+    },
+    {
+      "epoch": 15.836330935251798,
+      "step": 17610,
+      "torque_loss": 0.13218986988067627
+    },
+    {
+      "epoch": 15.836330935251798,
+      "force_loss": 0.011742700822651386,
+      "step": 17610
+    },
+    {
+      "epoch": 15.845323741007194,
+      "grad_norm": 0.27377206087112427,
+      "learning_rate": 8.462903149243899e-05,
+      "loss": 0.0217,
+      "step": 17620
+    },
+    {
+      "action_loss": 0.005419587716460228,
+      "epoch": 15.845323741007194,
+      "step": 17620
+    },
+    {
+      "epoch": 15.845323741007194,
+      "step": 17620,
+      "torque_loss": 0.10276583582162857
+    },
+    {
+      "epoch": 15.845323741007194,
+      "force_loss": 0.00598743325099349,
+      "step": 17620
+    },
+    {
+      "epoch": 15.85431654676259,
+      "grad_norm": 0.27440932393074036,
+      "learning_rate": 8.460914765554357e-05,
+      "loss": 0.0215,
+      "step": 17630
+    },
+    {
+      "action_loss": 0.006008013617247343,
+      "epoch": 15.85431654676259,
+      "step": 17630
+    },
+    {
+      "epoch": 15.85431654676259,
+      "step": 17630,
+      "torque_loss": 0.135933056473732
+    },
+    {
+      "epoch": 15.85431654676259,
+      "force_loss": 0.004844170529395342,
+      "step": 17630
+    },
+    {
+      "epoch": 15.863309352517986,
+      "grad_norm": 0.4325622022151947,
+      "learning_rate": 8.458925330530288e-05,
+      "loss": 0.03,
+      "step": 17640
+    },
+    {
+      "action_loss": 0.004168189596384764,
+      "epoch": 15.863309352517986,
+      "step": 17640
+    },
+    {
+      "epoch": 15.863309352517986,
+      "step": 17640,
+      "torque_loss": 0.15087778866291046
+    },
+    {
+      "epoch": 15.863309352517986,
+      "force_loss": 0.0033332640305161476,
+      "step": 17640
+    },
+    {
+      "epoch": 15.872302158273381,
+      "grad_norm": 0.440936416387558,
+      "learning_rate": 8.456934844776032e-05,
+      "loss": 0.0234,
+      "step": 17650
+    },
+    {
+      "action_loss": 0.006525044795125723,
+      "epoch": 15.872302158273381,
+      "step": 17650
+    },
+    {
+      "epoch": 15.872302158273381,
+      "step": 17650,
+      "torque_loss": 0.1332537978887558
+    },
+    {
+      "epoch": 15.872302158273381,
+      "force_loss": 0.011580362915992737,
+      "step": 17650
+    },
+    {
+      "epoch": 15.881294964028777,
+      "grad_norm": 1.0107548236846924,
+      "learning_rate": 8.454943308896246e-05,
+      "loss": 0.0271,
+      "step": 17660
+    },
+    {
+      "action_loss": 0.01941235363483429,
+      "epoch": 15.881294964028777,
+      "step": 17660
+    },
+    {
+      "epoch": 15.881294964028777,
+      "step": 17660,
+      "torque_loss": 0.17526470124721527
+    },
+    {
+      "epoch": 15.881294964028777,
+      "force_loss": 0.016152264550328255,
+      "step": 17660
+    },
+    {
+      "epoch": 15.890287769784173,
+      "grad_norm": 0.3381480872631073,
+      "learning_rate": 8.452950723495905e-05,
+      "loss": 0.0232,
+      "step": 17670
+    },
+    {
+      "action_loss": 0.008447933942079544,
+      "epoch": 15.890287769784173,
+      "step": 17670
+    },
+    {
+      "epoch": 15.890287769784173,
+      "step": 17670,
+      "torque_loss": 0.13460573554039001
+    },
+    {
+      "epoch": 15.890287769784173,
+      "force_loss": 0.006527024786919355,
+      "step": 17670
+    },
+    {
+      "epoch": 15.899280575539569,
+      "grad_norm": 0.626726508140564,
+      "learning_rate": 8.450957089180303e-05,
+      "loss": 0.0245,
+      "step": 17680
+    },
+    {
+      "action_loss": 0.0064415582455694675,
+      "epoch": 15.899280575539569,
+      "step": 17680
+    },
+    {
+      "epoch": 15.899280575539569,
+      "step": 17680,
+      "torque_loss": 0.1289413571357727
+    },
+    {
+      "epoch": 15.899280575539569,
+      "force_loss": 0.004623195622116327,
+      "step": 17680
+    },
+    {
+      "epoch": 15.908273381294965,
+      "grad_norm": 0.735294759273529,
+      "learning_rate": 8.448962406555055e-05,
+      "loss": 0.0235,
+      "step": 17690
+    },
+    {
+      "action_loss": 0.024541936814785004,
+      "epoch": 15.908273381294965,
+      "step": 17690
+    },
+    {
+      "epoch": 15.908273381294965,
+      "step": 17690,
+      "torque_loss": 0.14886687695980072
+    },
+    {
+      "epoch": 15.908273381294965,
+      "force_loss": 0.024909773841500282,
+      "step": 17690
+    },
+    {
+      "epoch": 15.917266187050359,
+      "grad_norm": 0.33747783303260803,
+      "learning_rate": 8.446966676226093e-05,
+      "loss": 0.0274,
+      "step": 17700
+    },
+    {
+      "action_loss": 0.004488543141633272,
+      "epoch": 15.917266187050359,
+      "step": 17700
+    },
+    {
+      "epoch": 15.917266187050359,
+      "step": 17700,
+      "torque_loss": 0.11510869115591049
+    },
+    {
+      "epoch": 15.917266187050359,
+      "force_loss": 0.0038391754496842623,
+      "step": 17700
+    },
+    {
+      "epoch": 15.926258992805755,
+      "grad_norm": 0.5885621309280396,
+      "learning_rate": 8.444969898799667e-05,
+      "loss": 0.0273,
+      "step": 17710
+    },
+    {
+      "action_loss": 0.01990235038101673,
+      "epoch": 15.926258992805755,
+      "step": 17710
+    },
+    {
+      "epoch": 15.926258992805755,
+      "step": 17710,
+      "torque_loss": 0.13860540091991425
+    },
+    {
+      "epoch": 15.926258992805755,
+      "force_loss": 0.018544703722000122,
+      "step": 17710
+    },
+    {
+      "epoch": 15.93525179856115,
+      "grad_norm": 0.7729743719100952,
+      "learning_rate": 8.442972074882343e-05,
+      "loss": 0.0253,
+      "step": 17720
+    },
+    {
+      "action_loss": 0.01067747175693512,
+      "epoch": 15.93525179856115,
+      "step": 17720
+    },
+    {
+      "epoch": 15.93525179856115,
+      "step": 17720,
+      "torque_loss": 0.128219872713089
+    },
+    {
+      "epoch": 15.93525179856115,
+      "force_loss": 0.009551775641739368,
+      "step": 17720
+    },
+    {
+      "epoch": 15.944244604316546,
+      "grad_norm": 0.6176245212554932,
+      "learning_rate": 8.44097320508101e-05,
+      "loss": 0.0241,
+      "step": 17730
+    },
+    {
+      "action_loss": 0.005875390022993088,
+      "epoch": 15.944244604316546,
+      "step": 17730
+    },
+    {
+      "epoch": 15.944244604316546,
+      "step": 17730,
+      "torque_loss": 0.1448507159948349
+    },
+    {
+      "epoch": 15.944244604316546,
+      "force_loss": 0.005019858945161104,
+      "step": 17730
+    },
+    {
+      "epoch": 15.953237410071942,
+      "grad_norm": 0.23681005835533142,
+      "learning_rate": 8.43897329000287e-05,
+      "loss": 0.0194,
+      "step": 17740
+    },
+    {
+      "action_loss": 0.010999500751495361,
+      "epoch": 15.953237410071942,
+      "step": 17740
+    },
+    {
+      "epoch": 15.953237410071942,
+      "step": 17740,
+      "torque_loss": 0.12151765078306198
+    },
+    {
+      "epoch": 15.953237410071942,
+      "force_loss": 0.007289711385965347,
+      "step": 17740
+    },
+    {
+      "epoch": 15.962230215827338,
+      "grad_norm": 0.29228949546813965,
+      "learning_rate": 8.436972330255448e-05,
+      "loss": 0.0223,
+      "step": 17750
+    },
+    {
+      "action_loss": 0.005345213692635298,
+      "epoch": 15.962230215827338,
+      "step": 17750
+    },
+    {
+      "epoch": 15.962230215827338,
+      "step": 17750,
+      "torque_loss": 0.11086401343345642
+    },
+    {
+      "epoch": 15.962230215827338,
+      "force_loss": 0.0031974802259355783,
+      "step": 17750
+    },
+    {
+      "epoch": 15.971223021582734,
+      "grad_norm": 0.4384174346923828,
+      "learning_rate": 8.434970326446579e-05,
+      "loss": 0.025,
+      "step": 17760
+    },
+    {
+      "action_loss": 0.007547078654170036,
+      "epoch": 15.971223021582734,
+      "step": 17760
+    },
+    {
+      "epoch": 15.971223021582734,
+      "step": 17760,
+      "torque_loss": 0.08681625127792358
+    },
+    {
+      "epoch": 15.971223021582734,
+      "force_loss": 0.00496362941339612,
+      "step": 17760
+    },
+    {
+      "epoch": 15.98021582733813,
+      "grad_norm": 0.23871272802352905,
+      "learning_rate": 8.432967279184418e-05,
+      "loss": 0.0214,
+      "step": 17770
+    },
+    {
+      "action_loss": 0.009915996342897415,
+      "epoch": 15.98021582733813,
+      "step": 17770
+    },
+    {
+      "epoch": 15.98021582733813,
+      "step": 17770,
+      "torque_loss": 0.10791996866464615
+    },
+    {
+      "epoch": 15.98021582733813,
+      "force_loss": 0.0038088730070739985,
+      "step": 17770
+    },
+    {
+      "epoch": 15.989208633093526,
+      "grad_norm": 0.5143328309059143,
+      "learning_rate": 8.430963189077441e-05,
+      "loss": 0.0199,
+      "step": 17780
+    },
+    {
+      "action_loss": 0.004039436113089323,
+      "epoch": 15.989208633093526,
+      "step": 17780
+    },
+    {
+      "epoch": 15.989208633093526,
+      "step": 17780,
+      "torque_loss": 0.15265363454818726
+    },
+    {
+      "epoch": 15.989208633093526,
+      "force_loss": 0.0063683390617370605,
+      "step": 17780
+    },
+    {
+      "epoch": 15.998201438848922,
+      "grad_norm": 0.5695040822029114,
+      "learning_rate": 8.428958056734437e-05,
+      "loss": 0.0207,
+      "step": 17790
+    },
+    {
+      "action_loss": 0.007465250324457884,
+      "epoch": 15.998201438848922,
+      "step": 17790
+    },
+    {
+      "epoch": 15.998201438848922,
+      "step": 17790,
+      "torque_loss": 0.13948342204093933
+    },
+    {
+      "epoch": 15.998201438848922,
+      "force_loss": 0.005301920231431723,
+      "step": 17790
+    },
+    {
+      "epoch": 16.007194244604317,
+      "grad_norm": 0.3909229040145874,
+      "learning_rate": 8.426951882764513e-05,
+      "loss": 0.0314,
+      "step": 17800
+    },
+    {
+      "action_loss": 0.013685323297977448,
+      "epoch": 16.007194244604317,
+      "step": 17800
+    },
+    {
+      "epoch": 16.007194244604317,
+      "step": 17800,
+      "torque_loss": 0.16999734938144684
+    },
+    {
+      "epoch": 16.007194244604317,
+      "force_loss": 0.019315673038363457,
+      "step": 17800
+    },
+    {
+      "epoch": 16.01618705035971,
+      "grad_norm": 0.33626413345336914,
+      "learning_rate": 8.424944667777089e-05,
+      "loss": 0.0214,
+      "step": 17810
+    },
+    {
+      "action_loss": 0.00577952666208148,
+      "epoch": 16.01618705035971,
+      "step": 17810
+    },
+    {
+      "epoch": 16.01618705035971,
+      "step": 17810,
+      "torque_loss": 0.1201549544930458
+    },
+    {
+      "epoch": 16.01618705035971,
+      "force_loss": 0.007053063716739416,
+      "step": 17810
+    },
+    {
+      "epoch": 16.02517985611511,
+      "grad_norm": 0.48775607347488403,
+      "learning_rate": 8.422936412381905e-05,
+      "loss": 0.0216,
+      "step": 17820
+    },
+    {
+      "action_loss": 0.0107901431620121,
+      "epoch": 16.02517985611511,
+      "step": 17820
+    },
+    {
+      "epoch": 16.02517985611511,
+      "step": 17820,
+      "torque_loss": 0.15228381752967834
+    },
+    {
+      "epoch": 16.02517985611511,
+      "force_loss": 0.0058626290410757065,
+      "step": 17820
+    },
+    {
+      "epoch": 16.034172661870503,
+      "grad_norm": 0.3989713191986084,
+      "learning_rate": 8.420927117189017e-05,
+      "loss": 0.0222,
+      "step": 17830
+    },
+    {
+      "action_loss": 0.01462673768401146,
+      "epoch": 16.034172661870503,
+      "step": 17830
+    },
+    {
+      "epoch": 16.034172661870503,
+      "step": 17830,
+      "torque_loss": 0.18573270738124847
+    },
+    {
+      "epoch": 16.034172661870503,
+      "force_loss": 0.010324918664991856,
+      "step": 17830
+    },
+    {
+      "epoch": 16.0431654676259,
+      "grad_norm": 0.3277093768119812,
+      "learning_rate": 8.418916782808795e-05,
+      "loss": 0.0241,
+      "step": 17840
+    },
+    {
+      "action_loss": 0.006221791263669729,
+      "epoch": 16.0431654676259,
+      "step": 17840
+    },
+    {
+      "epoch": 16.0431654676259,
+      "step": 17840,
+      "torque_loss": 0.10587122291326523
+    },
+    {
+      "epoch": 16.0431654676259,
+      "force_loss": 0.005029916763305664,
+      "step": 17840
+    },
+    {
+      "epoch": 16.052158273381295,
+      "grad_norm": 0.2585214376449585,
+      "learning_rate": 8.416905409851926e-05,
+      "loss": 0.0214,
+      "step": 17850
+    },
+    {
+      "action_loss": 0.0056740292347967625,
+      "epoch": 16.052158273381295,
+      "step": 17850
+    },
+    {
+      "epoch": 16.052158273381295,
+      "step": 17850,
+      "torque_loss": 0.16839216649532318
+    },
+    {
+      "epoch": 16.052158273381295,
+      "force_loss": 0.005034435074776411,
+      "step": 17850
+    },
+    {
+      "epoch": 16.06115107913669,
+      "grad_norm": 0.24692974984645844,
+      "learning_rate": 8.41489299892941e-05,
+      "loss": 0.02,
+      "step": 17860
+    },
+    {
+      "action_loss": 0.017583243548870087,
+      "epoch": 16.06115107913669,
+      "step": 17860
+    },
+    {
+      "epoch": 16.06115107913669,
+      "step": 17860,
+      "torque_loss": 0.19254548847675323
+    },
+    {
+      "epoch": 16.06115107913669,
+      "force_loss": 0.013399135321378708,
+      "step": 17860
+    },
+    {
+      "epoch": 16.070143884892087,
+      "grad_norm": 0.2261732518672943,
+      "learning_rate": 8.412879550652566e-05,
+      "loss": 0.0223,
+      "step": 17870
+    },
+    {
+      "action_loss": 0.006178155075758696,
+      "epoch": 16.070143884892087,
+      "step": 17870
+    },
+    {
+      "epoch": 16.070143884892087,
+      "step": 17870,
+      "torque_loss": 0.13173258304595947
+    },
+    {
+      "epoch": 16.070143884892087,
+      "force_loss": 0.006183135788887739,
+      "step": 17870
+    },
+    {
+      "epoch": 16.07913669064748,
+      "grad_norm": 0.4054189622402191,
+      "learning_rate": 8.410865065633029e-05,
+      "loss": 0.0224,
+      "step": 17880
+    },
+    {
+      "action_loss": 0.005918122828006744,
+      "epoch": 16.07913669064748,
+      "step": 17880
+    },
+    {
+      "epoch": 16.07913669064748,
+      "step": 17880,
+      "torque_loss": 0.084866464138031
+    },
+    {
+      "epoch": 16.07913669064748,
+      "force_loss": 0.0033826958388090134,
+      "step": 17880
+    },
+    {
+      "epoch": 16.08812949640288,
+      "grad_norm": 0.6187392473220825,
+      "learning_rate": 8.408849544482742e-05,
+      "loss": 0.0248,
+      "step": 17890
+    },
+    {
+      "action_loss": 0.007516726851463318,
+      "epoch": 16.08812949640288,
+      "step": 17890
+    },
+    {
+      "epoch": 16.08812949640288,
+      "step": 17890,
+      "torque_loss": 0.1508045345544815
+    },
+    {
+      "epoch": 16.08812949640288,
+      "force_loss": 0.011879724450409412,
+      "step": 17890
+    },
+    {
+      "epoch": 16.097122302158272,
+      "grad_norm": 0.6458420753479004,
+      "learning_rate": 8.406832987813968e-05,
+      "loss": 0.023,
+      "step": 17900
+    },
+    {
+      "action_loss": 0.012775298207998276,
+      "epoch": 16.097122302158272,
+      "step": 17900
+    },
+    {
+      "epoch": 16.097122302158272,
+      "step": 17900,
+      "torque_loss": 0.1499585658311844
+    },
+    {
+      "epoch": 16.097122302158272,
+      "force_loss": 0.02126883901655674,
+      "step": 17900
+    },
+    {
+      "epoch": 16.10611510791367,
+      "grad_norm": 0.8367826342582703,
+      "learning_rate": 8.404815396239286e-05,
+      "loss": 0.024,
+      "step": 17910
+    },
+    {
+      "action_loss": 0.0034704997669905424,
+      "epoch": 16.10611510791367,
+      "step": 17910
+    },
+    {
+      "epoch": 16.10611510791367,
+      "step": 17910,
+      "torque_loss": 0.15362930297851562
+    },
+    {
+      "epoch": 16.10611510791367,
+      "force_loss": 0.0028502587229013443,
+      "step": 17910
+    },
+    {
+      "epoch": 16.115107913669064,
+      "grad_norm": 0.6870766282081604,
+      "learning_rate": 8.402796770371587e-05,
+      "loss": 0.0258,
+      "step": 17920
+    },
+    {
+      "action_loss": 0.007104405667632818,
+      "epoch": 16.115107913669064,
+      "step": 17920
+    },
+    {
+      "epoch": 16.115107913669064,
+      "step": 17920,
+      "torque_loss": 0.10150986909866333
+    },
+    {
+      "epoch": 16.115107913669064,
+      "force_loss": 0.012554663233458996,
+      "step": 17920
+    },
+    {
+      "epoch": 16.12410071942446,
+      "grad_norm": 0.2720673382282257,
+      "learning_rate": 8.400777110824071e-05,
+      "loss": 0.0218,
+      "step": 17930
+    },
+    {
+      "action_loss": 0.00911865197122097,
+      "epoch": 16.12410071942446,
+      "step": 17930
+    },
+    {
+      "epoch": 16.12410071942446,
+      "step": 17930,
+      "torque_loss": 0.11493996530771255
+    },
+    {
+      "epoch": 16.12410071942446,
+      "force_loss": 0.008109722286462784,
+      "step": 17930
+    },
+    {
+      "epoch": 16.133093525179856,
+      "grad_norm": 0.7344787120819092,
+      "learning_rate": 8.398756418210263e-05,
+      "loss": 0.0202,
+      "step": 17940
+    },
+    {
+      "action_loss": 0.0041253212839365005,
+      "epoch": 16.133093525179856,
+      "step": 17940
+    },
+    {
+      "epoch": 16.133093525179856,
+      "step": 17940,
+      "torque_loss": 0.1408514380455017
+    },
+    {
+      "epoch": 16.133093525179856,
+      "force_loss": 0.002820750931277871,
+      "step": 17940
+    },
+    {
+      "epoch": 16.142086330935253,
+      "grad_norm": 0.2454034835100174,
+      "learning_rate": 8.396734693143993e-05,
+      "loss": 0.021,
+      "step": 17950
+    },
+    {
+      "action_loss": 0.008053547702729702,
+      "epoch": 16.142086330935253,
+      "step": 17950
+    },
+    {
+      "epoch": 16.142086330935253,
+      "step": 17950,
+      "torque_loss": 0.1375073790550232
+    },
+    {
+      "epoch": 16.142086330935253,
+      "force_loss": 0.006553851068019867,
+      "step": 17950
+    },
+    {
+      "epoch": 16.151079136690647,
+      "grad_norm": 0.37337568402290344,
+      "learning_rate": 8.39471193623941e-05,
+      "loss": 0.0297,
+      "step": 17960
+    },
+    {
+      "action_loss": 0.004111962392926216,
+      "epoch": 16.151079136690647,
+      "step": 17960
+    },
+    {
+      "epoch": 16.151079136690647,
+      "step": 17960,
+      "torque_loss": 0.11576279252767563
+    },
+    {
+      "epoch": 16.151079136690647,
+      "force_loss": 0.004006903152912855,
+      "step": 17960
+    },
+    {
+      "epoch": 16.16007194244604,
+      "grad_norm": 0.3513309955596924,
+      "learning_rate": 8.392688148110974e-05,
+      "loss": 0.0252,
+      "step": 17970
+    },
+    {
+      "action_loss": 0.003856303170323372,
+      "epoch": 16.16007194244604,
+      "step": 17970
+    },
+    {
+      "epoch": 16.16007194244604,
+      "step": 17970,
+      "torque_loss": 0.13178487122058868
+    },
+    {
+      "epoch": 16.16007194244604,
+      "force_loss": 0.0045032622292637825,
+      "step": 17970
+    },
+    {
+      "epoch": 16.16906474820144,
+      "grad_norm": 0.5031517744064331,
+      "learning_rate": 8.390663329373456e-05,
+      "loss": 0.0226,
+      "step": 17980
+    },
+    {
+      "action_loss": 0.007378889247775078,
+      "epoch": 16.16906474820144,
+      "step": 17980
+    },
+    {
+      "epoch": 16.16906474820144,
+      "step": 17980,
+      "torque_loss": 0.13647036254405975
+    },
+    {
+      "epoch": 16.16906474820144,
+      "force_loss": 0.006954573094844818,
+      "step": 17980
+    },
+    {
+      "epoch": 16.178057553956833,
+      "grad_norm": 0.6193506717681885,
+      "learning_rate": 8.388637480641944e-05,
+      "loss": 0.0297,
+      "step": 17990
+    },
+    {
+      "action_loss": 0.015983037650585175,
+      "epoch": 16.178057553956833,
+      "step": 17990
+    },
+    {
+      "epoch": 16.178057553956833,
+      "step": 17990,
+      "torque_loss": 0.1552460938692093
+    },
+    {
+      "epoch": 16.178057553956833,
+      "force_loss": 0.011983555741608143,
+      "step": 17990
+    },
+    {
+      "epoch": 16.18705035971223,
+      "grad_norm": 0.4312192499637604,
+      "learning_rate": 8.386610602531837e-05,
+      "loss": 0.0257,
+      "step": 18000
+    },
+    {
+      "action_loss": 0.007568379398435354,
+      "epoch": 16.18705035971223,
+      "step": 18000
+    },
+    {
+      "epoch": 16.18705035971223,
+      "step": 18000,
+      "torque_loss": 0.11109446734189987
+    },
+    {
+      "epoch": 16.18705035971223,
+      "force_loss": 0.003926304634660482,
+      "step": 18000
+    },
+    {
+      "epoch": 16.196043165467625,
+      "grad_norm": 0.4163177013397217,
+      "learning_rate": 8.384582695658847e-05,
+      "loss": 0.0235,
+      "step": 18010
+    },
+    {
+      "action_loss": 0.015307377092540264,
+      "epoch": 16.196043165467625,
+      "step": 18010
+    },
+    {
+      "epoch": 16.196043165467625,
+      "step": 18010,
+      "torque_loss": 0.101185142993927
+    },
+    {
+      "epoch": 16.196043165467625,
+      "force_loss": 0.02133381925523281,
+      "step": 18010
+    },
+    {
+      "epoch": 16.205035971223023,
+      "grad_norm": 0.44741207361221313,
+      "learning_rate": 8.382553760638999e-05,
+      "loss": 0.0237,
+      "step": 18020
+    },
+    {
+      "action_loss": 0.009381234645843506,
+      "epoch": 16.205035971223023,
+      "step": 18020
+    },
+    {
+      "epoch": 16.205035971223023,
+      "step": 18020,
+      "torque_loss": 0.12503407895565033
+    },
+    {
+      "epoch": 16.205035971223023,
+      "force_loss": 0.00686328811571002,
+      "step": 18020
+    },
+    {
+      "epoch": 16.214028776978417,
+      "grad_norm": 0.36904048919677734,
+      "learning_rate": 8.380523798088631e-05,
+      "loss": 0.0206,
+      "step": 18030
+    },
+    {
+      "action_loss": 0.005536861252039671,
+      "epoch": 16.214028776978417,
+      "step": 18030
+    },
+    {
+      "epoch": 16.214028776978417,
+      "step": 18030,
+      "torque_loss": 0.11668294668197632
+    },
+    {
+      "epoch": 16.214028776978417,
+      "force_loss": 0.005136905703693628,
+      "step": 18030
+    },
+    {
+      "epoch": 16.223021582733814,
+      "grad_norm": 0.3489348888397217,
+      "learning_rate": 8.378492808624389e-05,
+      "loss": 0.0213,
+      "step": 18040
+    },
+    {
+      "action_loss": 0.004436939489096403,
+      "epoch": 16.223021582733814,
+      "step": 18040
+    },
+    {
+      "epoch": 16.223021582733814,
+      "step": 18040,
+      "torque_loss": 0.10620620846748352
+    },
+    {
+      "epoch": 16.223021582733814,
+      "force_loss": 0.004801291972398758,
+      "step": 18040
+    },
+    {
+      "epoch": 16.23201438848921,
+      "grad_norm": 0.5666268467903137,
+      "learning_rate": 8.376460792863237e-05,
+      "loss": 0.0258,
+      "step": 18050
+    },
+    {
+      "action_loss": 0.009859741665422916,
+      "epoch": 16.23201438848921,
+      "step": 18050
+    },
+    {
+      "epoch": 16.23201438848921,
+      "step": 18050,
+      "torque_loss": 0.13223198056221008
+    },
+    {
+      "epoch": 16.23201438848921,
+      "force_loss": 0.005797772202640772,
+      "step": 18050
+    },
+    {
+      "epoch": 16.241007194244606,
+      "grad_norm": 0.3594134449958801,
+      "learning_rate": 8.374427751422444e-05,
+      "loss": 0.0183,
+      "step": 18060
+    },
+    {
+      "action_loss": 0.005189632531255484,
+      "epoch": 16.241007194244606,
+      "step": 18060
+    },
+    {
+      "epoch": 16.241007194244606,
+      "step": 18060,
+      "torque_loss": 0.10445895791053772
+    },
+    {
+      "epoch": 16.241007194244606,
+      "force_loss": 0.0031314559746533632,
+      "step": 18060
+    },
+    {
+      "epoch": 16.25,
+      "grad_norm": 0.2947518825531006,
+      "learning_rate": 8.3723936849196e-05,
+      "loss": 0.0199,
+      "step": 18070
+    },
+    {
+      "action_loss": 0.004197425674647093,
+      "epoch": 16.25,
+      "step": 18070
+    },
+    {
+      "epoch": 16.25,
+      "step": 18070,
+      "torque_loss": 0.15771900117397308
+    },
+    {
+      "epoch": 16.25,
+      "force_loss": 0.0050789485685527325,
+      "step": 18070
+    },
+    {
+      "epoch": 16.258992805755394,
+      "grad_norm": 0.34019678831100464,
+      "learning_rate": 8.370358593972595e-05,
+      "loss": 0.0208,
+      "step": 18080
+    },
+    {
+      "action_loss": 0.004253249149769545,
+      "epoch": 16.258992805755394,
+      "step": 18080
+    },
+    {
+      "epoch": 16.258992805755394,
+      "step": 18080,
+      "torque_loss": 0.1208881139755249
+    },
+    {
+      "epoch": 16.258992805755394,
+      "force_loss": 0.007525982800871134,
+      "step": 18080
+    },
+    {
+      "epoch": 16.26798561151079,
+      "grad_norm": 0.4985514283180237,
+      "learning_rate": 8.36832247919964e-05,
+      "loss": 0.0195,
+      "step": 18090
+    },
+    {
+      "action_loss": 0.015206550247967243,
+      "epoch": 16.26798561151079,
+      "step": 18090
+    },
+    {
+      "epoch": 16.26798561151079,
+      "step": 18090,
+      "torque_loss": 0.11427616328001022
+    },
+    {
+      "epoch": 16.26798561151079,
+      "force_loss": 0.0120158726349473,
+      "step": 18090
+    },
+    {
+      "epoch": 16.276978417266186,
+      "grad_norm": 0.4713435769081116,
+      "learning_rate": 8.36628534121925e-05,
+      "loss": 0.0244,
+      "step": 18100
+    },
+    {
+      "action_loss": 0.004326095338910818,
+      "epoch": 16.276978417266186,
+      "step": 18100
+    },
+    {
+      "epoch": 16.276978417266186,
+      "step": 18100,
+      "torque_loss": 0.12609809637069702
+    },
+    {
+      "epoch": 16.276978417266186,
+      "force_loss": 0.0025467786472290754,
+      "step": 18100
+    },
+    {
+      "epoch": 16.285971223021583,
+      "grad_norm": 0.9442657232284546,
+      "learning_rate": 8.364247180650254e-05,
+      "loss": 0.0262,
+      "step": 18110
+    },
+    {
+      "action_loss": 0.011486049741506577,
+      "epoch": 16.285971223021583,
+      "step": 18110
+    },
+    {
+      "epoch": 16.285971223021583,
+      "step": 18110,
+      "torque_loss": 0.1206124797463417
+    },
+    {
+      "epoch": 16.285971223021583,
+      "force_loss": 0.011293943971395493,
+      "step": 18110
+    },
+    {
+      "epoch": 16.294964028776977,
+      "grad_norm": 0.25877389311790466,
+      "learning_rate": 8.362207998111794e-05,
+      "loss": 0.0191,
+      "step": 18120
+    },
+    {
+      "action_loss": 0.00729757035151124,
+      "epoch": 16.294964028776977,
+      "step": 18120
+    },
+    {
+      "epoch": 16.294964028776977,
+      "step": 18120,
+      "torque_loss": 0.12804141640663147
+    },
+    {
+      "epoch": 16.294964028776977,
+      "force_loss": 0.0034155212342739105,
+      "step": 18120
+    },
+    {
+      "epoch": 16.303956834532375,
+      "grad_norm": 0.2360243946313858,
+      "learning_rate": 8.360167794223318e-05,
+      "loss": 0.0209,
+      "step": 18130
+    },
+    {
+      "action_loss": 0.009071596898138523,
+      "epoch": 16.303956834532375,
+      "step": 18130
+    },
+    {
+      "epoch": 16.303956834532375,
+      "step": 18130,
+      "torque_loss": 0.17007555067539215
+    },
+    {
+      "epoch": 16.303956834532375,
+      "force_loss": 0.010613895952701569,
+      "step": 18130
+    },
+    {
+      "epoch": 16.31294964028777,
+      "grad_norm": 0.6451969146728516,
+      "learning_rate": 8.358126569604586e-05,
+      "loss": 0.0234,
+      "step": 18140
+    },
+    {
+      "action_loss": 0.004848698619753122,
+      "epoch": 16.31294964028777,
+      "step": 18140
+    },
+    {
+      "epoch": 16.31294964028777,
+      "step": 18140,
+      "torque_loss": 0.1646655946969986
+    },
+    {
+      "epoch": 16.31294964028777,
+      "force_loss": 0.005273627582937479,
+      "step": 18140
+    },
+    {
+      "epoch": 16.321942446043167,
+      "grad_norm": 0.25522100925445557,
+      "learning_rate": 8.356084324875668e-05,
+      "loss": 0.0257,
+      "step": 18150
+    },
+    {
+      "action_loss": 0.0047170682810246944,
+      "epoch": 16.321942446043167,
+      "step": 18150
+    },
+    {
+      "epoch": 16.321942446043167,
+      "step": 18150,
+      "torque_loss": 0.11787024885416031
+    },
+    {
+      "epoch": 16.321942446043167,
+      "force_loss": 0.006361795123666525,
+      "step": 18150
+    },
+    {
+      "epoch": 16.33093525179856,
+      "grad_norm": 0.3561152219772339,
+      "learning_rate": 8.354041060656945e-05,
+      "loss": 0.0198,
+      "step": 18160
+    },
+    {
+      "action_loss": 0.0031788635533303022,
+      "epoch": 16.33093525179856,
+      "step": 18160
+    },
+    {
+      "epoch": 16.33093525179856,
+      "step": 18160,
+      "torque_loss": 0.09880802035331726
+    },
+    {
+      "epoch": 16.33093525179856,
+      "force_loss": 0.002557651838287711,
+      "step": 18160
+    },
+    {
+      "epoch": 16.33992805755396,
+      "grad_norm": 0.26321348547935486,
+      "learning_rate": 8.351996777569106e-05,
+      "loss": 0.0204,
+      "step": 18170
+    },
+    {
+      "action_loss": 0.006902188062667847,
+      "epoch": 16.33992805755396,
+      "step": 18170
+    },
+    {
+      "epoch": 16.33992805755396,
+      "step": 18170,
+      "torque_loss": 0.13893133401870728
+    },
+    {
+      "epoch": 16.33992805755396,
+      "force_loss": 0.006122795399278402,
+      "step": 18170
+    },
+    {
+      "epoch": 16.348920863309353,
+      "grad_norm": 0.38124632835388184,
+      "learning_rate": 8.349951476233148e-05,
+      "loss": 0.0226,
+      "step": 18180
+    },
+    {
+      "action_loss": 0.003286626422777772,
+      "epoch": 16.348920863309353,
+      "step": 18180
+    },
+    {
+      "epoch": 16.348920863309353,
+      "step": 18180,
+      "torque_loss": 0.10841871052980423
+    },
+    {
+      "epoch": 16.348920863309353,
+      "force_loss": 0.003879651427268982,
+      "step": 18180
+    },
+    {
+      "epoch": 16.357913669064747,
+      "grad_norm": 0.22667431831359863,
+      "learning_rate": 8.347905157270386e-05,
+      "loss": 0.0215,
+      "step": 18190
+    },
+    {
+      "action_loss": 0.0055561005137860775,
+      "epoch": 16.357913669064747,
+      "step": 18190
+    },
+    {
+      "epoch": 16.357913669064747,
+      "step": 18190,
+      "torque_loss": 0.15022213757038116
+    },
+    {
+      "epoch": 16.357913669064747,
+      "force_loss": 0.004825991578400135,
+      "step": 18190
+    },
+    {
+      "epoch": 16.366906474820144,
+      "grad_norm": 0.2512531280517578,
+      "learning_rate": 8.345857821302432e-05,
+      "loss": 0.0239,
+      "step": 18200
+    },
+    {
+      "action_loss": 0.008582408539950848,
+      "epoch": 16.366906474820144,
+      "step": 18200
+    },
+    {
+      "epoch": 16.366906474820144,
+      "step": 18200,
+      "torque_loss": 0.16552527248859406
+    },
+    {
+      "epoch": 16.366906474820144,
+      "force_loss": 0.01397332176566124,
+      "step": 18200
+    },
+    {
+      "epoch": 16.37589928057554,
+      "grad_norm": 0.2697646915912628,
+      "learning_rate": 8.343809468951213e-05,
+      "loss": 0.0214,
+      "step": 18210
+    },
+    {
+      "action_loss": 0.006092951167374849,
+      "epoch": 16.37589928057554,
+      "step": 18210
+    },
+    {
+      "epoch": 16.37589928057554,
+      "step": 18210,
+      "torque_loss": 0.1761222928762436
+    },
+    {
+      "epoch": 16.37589928057554,
+      "force_loss": 0.007903401739895344,
+      "step": 18210
+    },
+    {
+      "epoch": 16.384892086330936,
+      "grad_norm": 0.5763723254203796,
+      "learning_rate": 8.341760100838965e-05,
+      "loss": 0.0278,
+      "step": 18220
+    },
+    {
+      "action_loss": 0.009349461644887924,
+      "epoch": 16.384892086330936,
+      "step": 18220
+    },
+    {
+      "epoch": 16.384892086330936,
+      "step": 18220,
+      "torque_loss": 0.13287697732448578
+    },
+    {
+      "epoch": 16.384892086330936,
+      "force_loss": 0.007232888136059046,
+      "step": 18220
+    },
+    {
+      "epoch": 16.39388489208633,
+      "grad_norm": 0.30847808718681335,
+      "learning_rate": 8.339709717588233e-05,
+      "loss": 0.0267,
+      "step": 18230
+    },
+    {
+      "action_loss": 0.007372267544269562,
+      "epoch": 16.39388489208633,
+      "step": 18230
+    },
+    {
+      "epoch": 16.39388489208633,
+      "step": 18230,
+      "torque_loss": 0.1381007581949234
+    },
+    {
+      "epoch": 16.39388489208633,
+      "force_loss": 0.005219774786382914,
+      "step": 18230
+    },
+    {
+      "epoch": 16.402877697841728,
+      "grad_norm": 0.5407735109329224,
+      "learning_rate": 8.33765831982187e-05,
+      "loss": 0.022,
+      "step": 18240
+    },
+    {
+      "action_loss": 0.005897350609302521,
+      "epoch": 16.402877697841728,
+      "step": 18240
+    },
+    {
+      "epoch": 16.402877697841728,
+      "step": 18240,
+      "torque_loss": 0.08377639204263687
+    },
+    {
+      "epoch": 16.402877697841728,
+      "force_loss": 0.004883553367108107,
+      "step": 18240
+    },
+    {
+      "epoch": 16.41187050359712,
+      "grad_norm": 0.22648492455482483,
+      "learning_rate": 8.335605908163035e-05,
+      "loss": 0.0179,
+      "step": 18250
+    },
+    {
+      "action_loss": 0.008108261041343212,
+      "epoch": 16.41187050359712,
+      "step": 18250
+    },
+    {
+      "epoch": 16.41187050359712,
+      "step": 18250,
+      "torque_loss": 0.1093992218375206
+    },
+    {
+      "epoch": 16.41187050359712,
+      "force_loss": 0.004886269569396973,
+      "step": 18250
+    },
+    {
+      "epoch": 16.42086330935252,
+      "grad_norm": 0.39348939061164856,
+      "learning_rate": 8.333552483235196e-05,
+      "loss": 0.021,
+      "step": 18260
+    },
+    {
+      "action_loss": 0.0043386840261518955,
+      "epoch": 16.42086330935252,
+      "step": 18260
+    },
+    {
+      "epoch": 16.42086330935252,
+      "step": 18260,
+      "torque_loss": 0.12040645629167557
+    },
+    {
+      "epoch": 16.42086330935252,
+      "force_loss": 0.004096080083400011,
+      "step": 18260
+    },
+    {
+      "epoch": 16.429856115107913,
+      "grad_norm": 0.47013697028160095,
+      "learning_rate": 8.33149804566213e-05,
+      "loss": 0.0226,
+      "step": 18270
+    },
+    {
+      "action_loss": 0.010142925195395947,
+      "epoch": 16.429856115107913,
+      "step": 18270
+    },
+    {
+      "epoch": 16.429856115107913,
+      "step": 18270,
+      "torque_loss": 0.14123956859111786
+    },
+    {
+      "epoch": 16.429856115107913,
+      "force_loss": 0.006758174393326044,
+      "step": 18270
+    },
+    {
+      "epoch": 16.43884892086331,
+      "grad_norm": 0.2910516560077667,
+      "learning_rate": 8.329442596067921e-05,
+      "loss": 0.0245,
+      "step": 18280
+    },
+    {
+      "action_loss": 0.003754260018467903,
+      "epoch": 16.43884892086331,
+      "step": 18280
+    },
+    {
+      "epoch": 16.43884892086331,
+      "step": 18280,
+      "torque_loss": 0.11502655595541
+    },
+    {
+      "epoch": 16.43884892086331,
+      "force_loss": 0.0023056436330080032,
+      "step": 18280
+    },
+    {
+      "epoch": 16.447841726618705,
+      "grad_norm": 0.3777787387371063,
+      "learning_rate": 8.32738613507696e-05,
+      "loss": 0.0251,
+      "step": 18290
+    },
+    {
+      "action_loss": 0.024292265996336937,
+      "epoch": 16.447841726618705,
+      "step": 18290
+    },
+    {
+      "epoch": 16.447841726618705,
+      "step": 18290,
+      "torque_loss": 0.18905021250247955
+    },
+    {
+      "epoch": 16.447841726618705,
+      "force_loss": 0.032001394778490067,
+      "step": 18290
+    },
+    {
+      "epoch": 16.4568345323741,
+      "grad_norm": 0.7325795888900757,
+      "learning_rate": 8.325328663313946e-05,
+      "loss": 0.0283,
+      "step": 18300
+    },
+    {
+      "action_loss": 0.012889673002064228,
+      "epoch": 16.4568345323741,
+      "step": 18300
+    },
+    {
+      "epoch": 16.4568345323741,
+      "step": 18300,
+      "torque_loss": 0.17282839119434357
+    },
+    {
+      "epoch": 16.4568345323741,
+      "force_loss": 0.023282423615455627,
+      "step": 18300
+    },
+    {
+      "epoch": 16.465827338129497,
+      "grad_norm": 0.3477124571800232,
+      "learning_rate": 8.323270181403884e-05,
+      "loss": 0.0249,
+      "step": 18310
+    },
+    {
+      "action_loss": 0.01892814226448536,
+      "epoch": 16.465827338129497,
+      "step": 18310
+    },
+    {
+      "epoch": 16.465827338129497,
+      "step": 18310,
+      "torque_loss": 0.13234031200408936
+    },
+    {
+      "epoch": 16.465827338129497,
+      "force_loss": 0.017151379957795143,
+      "step": 18310
+    },
+    {
+      "epoch": 16.47482014388489,
+      "grad_norm": 0.2253669798374176,
+      "learning_rate": 8.321210689972086e-05,
+      "loss": 0.0256,
+      "step": 18320
+    },
+    {
+      "action_loss": 0.005118671339005232,
+      "epoch": 16.47482014388489,
+      "step": 18320
+    },
+    {
+      "epoch": 16.47482014388489,
+      "step": 18320,
+      "torque_loss": 0.1254182606935501
+    },
+    {
+      "epoch": 16.47482014388489,
+      "force_loss": 0.00862070545554161,
+      "step": 18320
+    },
+    {
+      "epoch": 16.48381294964029,
+      "grad_norm": 0.3347220718860626,
+      "learning_rate": 8.319150189644174e-05,
+      "loss": 0.0239,
+      "step": 18330
+    },
+    {
+      "action_loss": 0.00308927777223289,
+      "epoch": 16.48381294964029,
+      "step": 18330
+    },
+    {
+      "epoch": 16.48381294964029,
+      "step": 18330,
+      "torque_loss": 0.10159683227539062
+    },
+    {
+      "epoch": 16.48381294964029,
+      "force_loss": 0.004008258227258921,
+      "step": 18330
+    },
+    {
+      "epoch": 16.492805755395683,
+      "grad_norm": 0.6985332369804382,
+      "learning_rate": 8.31708868104607e-05,
+      "loss": 0.0299,
+      "step": 18340
+    },
+    {
+      "action_loss": 0.010480965487658978,
+      "epoch": 16.492805755395683,
+      "step": 18340
+    },
+    {
+      "epoch": 16.492805755395683,
+      "step": 18340,
+      "torque_loss": 0.1386459618806839
+    },
+    {
+      "epoch": 16.492805755395683,
+      "force_loss": 0.007932504639029503,
+      "step": 18340
+    },
+    {
+      "epoch": 16.50179856115108,
+      "grad_norm": 0.37076595425605774,
+      "learning_rate": 8.315026164804007e-05,
+      "loss": 0.0238,
+      "step": 18350
+    },
+    {
+      "action_loss": 0.006796905305236578,
+      "epoch": 16.50179856115108,
+      "step": 18350
+    },
+    {
+      "epoch": 16.50179856115108,
+      "step": 18350,
+      "torque_loss": 0.13150005042552948
+    },
+    {
+      "epoch": 16.50179856115108,
+      "force_loss": 0.0044395411387085915,
+      "step": 18350
+    },
+    {
+      "epoch": 16.510791366906474,
+      "grad_norm": 1.0180201530456543,
+      "learning_rate": 8.312962641544524e-05,
+      "loss": 0.025,
+      "step": 18360
+    },
+    {
+      "action_loss": 0.007877036929130554,
+      "epoch": 16.510791366906474,
+      "step": 18360
+    },
+    {
+      "epoch": 16.510791366906474,
+      "step": 18360,
+      "torque_loss": 0.1178581491112709
+    },
+    {
+      "epoch": 16.510791366906474,
+      "force_loss": 0.009159534238278866,
+      "step": 18360
+    },
+    {
+      "epoch": 16.519784172661872,
+      "grad_norm": 0.39785414934158325,
+      "learning_rate": 8.310898111894465e-05,
+      "loss": 0.0216,
+      "step": 18370
+    },
+    {
+      "action_loss": 0.008548031561076641,
+      "epoch": 16.519784172661872,
+      "step": 18370
+    },
+    {
+      "epoch": 16.519784172661872,
+      "step": 18370,
+      "torque_loss": 0.13693289458751678
+    },
+    {
+      "epoch": 16.519784172661872,
+      "force_loss": 0.016244065016508102,
+      "step": 18370
+    },
+    {
+      "epoch": 16.528776978417266,
+      "grad_norm": 0.4085237979888916,
+      "learning_rate": 8.308832576480977e-05,
+      "loss": 0.0214,
+      "step": 18380
+    },
+    {
+      "action_loss": 0.004474898334592581,
+      "epoch": 16.528776978417266,
+      "step": 18380
+    },
+    {
+      "epoch": 16.528776978417266,
+      "step": 18380,
+      "torque_loss": 0.11087968200445175
+    },
+    {
+      "epoch": 16.528776978417266,
+      "force_loss": 0.004281978588551283,
+      "step": 18380
+    },
+    {
+      "epoch": 16.53776978417266,
+      "grad_norm": 0.24151866137981415,
+      "learning_rate": 8.306766035931519e-05,
+      "loss": 0.0262,
+      "step": 18390
+    },
+    {
+      "action_loss": 0.006306593772023916,
+      "epoch": 16.53776978417266,
+      "step": 18390
+    },
+    {
+      "epoch": 16.53776978417266,
+      "step": 18390,
+      "torque_loss": 0.2008480280637741
+    },
+    {
+      "epoch": 16.53776978417266,
+      "force_loss": 0.011665627360343933,
+      "step": 18390
+    },
+    {
+      "epoch": 16.546762589928058,
+      "grad_norm": 0.26691973209381104,
+      "learning_rate": 8.304698490873847e-05,
+      "loss": 0.0229,
+      "step": 18400
+    },
+    {
+      "action_loss": 0.005176689941436052,
+      "epoch": 16.546762589928058,
+      "step": 18400
+    },
+    {
+      "epoch": 16.546762589928058,
+      "step": 18400,
+      "torque_loss": 0.13689620792865753
+    },
+    {
+      "epoch": 16.546762589928058,
+      "force_loss": 0.0031080376356840134,
+      "step": 18400
+    },
+    {
+      "epoch": 16.555755395683452,
+      "grad_norm": 0.3118777275085449,
+      "learning_rate": 8.30262994193603e-05,
+      "loss": 0.0243,
+      "step": 18410
+    },
+    {
+      "action_loss": 0.0033146876376122236,
+      "epoch": 16.555755395683452,
+      "step": 18410
+    },
+    {
+      "epoch": 16.555755395683452,
+      "step": 18410,
+      "torque_loss": 0.10385149717330933
+    },
+    {
+      "epoch": 16.555755395683452,
+      "force_loss": 0.003780081868171692,
+      "step": 18410
+    },
+    {
+      "epoch": 16.56474820143885,
+      "grad_norm": 0.4881945252418518,
+      "learning_rate": 8.300560389746438e-05,
+      "loss": 0.0222,
+      "step": 18420
+    },
+    {
+      "action_loss": 0.007430974394083023,
+      "epoch": 16.56474820143885,
+      "step": 18420
+    },
+    {
+      "epoch": 16.56474820143885,
+      "step": 18420,
+      "torque_loss": 0.11570519953966141
+    },
+    {
+      "epoch": 16.56474820143885,
+      "force_loss": 0.008030847646296024,
+      "step": 18420
+    },
+    {
+      "epoch": 16.573741007194243,
+      "grad_norm": 0.7185755968093872,
+      "learning_rate": 8.298489834933745e-05,
+      "loss": 0.0263,
+      "step": 18430
+    },
+    {
+      "action_loss": 0.003142673522233963,
+      "epoch": 16.573741007194243,
+      "step": 18430
+    },
+    {
+      "epoch": 16.573741007194243,
+      "step": 18430,
+      "torque_loss": 0.12953804433345795
+    },
+    {
+      "epoch": 16.573741007194243,
+      "force_loss": 0.0025068691466003656,
+      "step": 18430
+    },
+    {
+      "epoch": 16.58273381294964,
+      "grad_norm": 0.25372314453125,
+      "learning_rate": 8.296418278126934e-05,
+      "loss": 0.0283,
+      "step": 18440
+    },
+    {
+      "action_loss": 0.005031744949519634,
+      "epoch": 16.58273381294964,
+      "step": 18440
+    },
+    {
+      "epoch": 16.58273381294964,
+      "step": 18440,
+      "torque_loss": 0.09780289977788925
+    },
+    {
+      "epoch": 16.58273381294964,
+      "force_loss": 0.00647571450099349,
+      "step": 18440
+    },
+    {
+      "epoch": 16.591726618705035,
+      "grad_norm": 0.2521689832210541,
+      "learning_rate": 8.294345719955284e-05,
+      "loss": 0.0223,
+      "step": 18450
+    },
+    {
+      "action_loss": 0.004054040182381868,
+      "epoch": 16.591726618705035,
+      "step": 18450
+    },
+    {
+      "epoch": 16.591726618705035,
+      "step": 18450,
+      "torque_loss": 0.21708768606185913
+    },
+    {
+      "epoch": 16.591726618705035,
+      "force_loss": 0.00399880250915885,
+      "step": 18450
+    },
+    {
+      "epoch": 16.600719424460433,
+      "grad_norm": 0.4184238314628601,
+      "learning_rate": 8.29227216104839e-05,
+      "loss": 0.0228,
+      "step": 18460
+    },
+    {
+      "action_loss": 0.006403022911399603,
+      "epoch": 16.600719424460433,
+      "step": 18460
+    },
+    {
+      "epoch": 16.600719424460433,
+      "step": 18460,
+      "torque_loss": 0.12503254413604736
+    },
+    {
+      "epoch": 16.600719424460433,
+      "force_loss": 0.005357746034860611,
+      "step": 18460
+    },
+    {
+      "epoch": 16.609712230215827,
+      "grad_norm": 0.34864628314971924,
+      "learning_rate": 8.290197602036137e-05,
+      "loss": 0.0236,
+      "step": 18470
+    },
+    {
+      "action_loss": 0.005995148792862892,
+      "epoch": 16.609712230215827,
+      "step": 18470
+    },
+    {
+      "epoch": 16.609712230215827,
+      "step": 18470,
+      "torque_loss": 0.1445033699274063
+    },
+    {
+      "epoch": 16.609712230215827,
+      "force_loss": 0.003798463148996234,
+      "step": 18470
+    },
+    {
+      "epoch": 16.618705035971225,
+      "grad_norm": 0.4197278618812561,
+      "learning_rate": 8.288122043548725e-05,
+      "loss": 0.0203,
+      "step": 18480
+    },
+    {
+      "action_loss": 0.008849951438605785,
+      "epoch": 16.618705035971225,
+      "step": 18480
+    },
+    {
+      "epoch": 16.618705035971225,
+      "step": 18480,
+      "torque_loss": 0.09843218326568604
+    },
+    {
+      "epoch": 16.618705035971225,
+      "force_loss": 0.009366246871650219,
+      "step": 18480
+    },
+    {
+      "epoch": 16.62769784172662,
+      "grad_norm": 0.3057209849357605,
+      "learning_rate": 8.286045486216657e-05,
+      "loss": 0.0254,
+      "step": 18490
+    },
+    {
+      "action_loss": 0.006491031032055616,
+      "epoch": 16.62769784172662,
+      "step": 18490
+    },
+    {
+      "epoch": 16.62769784172662,
+      "step": 18490,
+      "torque_loss": 0.1364649534225464
+    },
+    {
+      "epoch": 16.62769784172662,
+      "force_loss": 0.00570320850238204,
+      "step": 18490
+    },
+    {
+      "epoch": 16.636690647482013,
+      "grad_norm": 0.49683183431625366,
+      "learning_rate": 8.283967930670733e-05,
+      "loss": 0.0232,
+      "step": 18500
+    },
+    {
+      "action_loss": 0.008977101184427738,
+      "epoch": 16.636690647482013,
+      "step": 18500
+    },
+    {
+      "epoch": 16.636690647482013,
+      "step": 18500,
+      "torque_loss": 0.17831246554851532
+    },
+    {
+      "epoch": 16.636690647482013,
+      "force_loss": 0.005677130538970232,
+      "step": 18500
+    },
+    {
+      "epoch": 16.64568345323741,
+      "grad_norm": 0.25140267610549927,
+      "learning_rate": 8.281889377542058e-05,
+      "loss": 0.0248,
+      "step": 18510
+    },
+    {
+      "action_loss": 0.011087329126894474,
+      "epoch": 16.64568345323741,
+      "step": 18510
+    },
+    {
+      "epoch": 16.64568345323741,
+      "step": 18510,
+      "torque_loss": 0.12683016061782837
+    },
+    {
+      "epoch": 16.64568345323741,
+      "force_loss": 0.005519574508070946,
+      "step": 18510
+    },
+    {
+      "epoch": 16.654676258992804,
+      "grad_norm": 0.7398377656936646,
+      "learning_rate": 8.279809827462045e-05,
+      "loss": 0.0213,
+      "step": 18520
+    },
+    {
+      "action_loss": 0.01726895570755005,
+      "epoch": 16.654676258992804,
+      "step": 18520
+    },
+    {
+      "epoch": 16.654676258992804,
+      "step": 18520,
+      "torque_loss": 0.19766275584697723
+    },
+    {
+      "epoch": 16.654676258992804,
+      "force_loss": 0.012134368531405926,
+      "step": 18520
+    },
+    {
+      "epoch": 16.663669064748202,
+      "grad_norm": 0.5961434841156006,
+      "learning_rate": 8.277729281062402e-05,
+      "loss": 0.0267,
+      "step": 18530
+    },
+    {
+      "action_loss": 0.005861076060682535,
+      "epoch": 16.663669064748202,
+      "step": 18530
+    },
+    {
+      "epoch": 16.663669064748202,
+      "step": 18530,
+      "torque_loss": 0.13658632338047028
+    },
+    {
+      "epoch": 16.663669064748202,
+      "force_loss": 0.0032646150793880224,
+      "step": 18530
+    },
+    {
+      "epoch": 16.672661870503596,
+      "grad_norm": 0.4042477011680603,
+      "learning_rate": 8.27564773897515e-05,
+      "loss": 0.0244,
+      "step": 18540
+    },
+    {
+      "action_loss": 0.0073905400931835175,
+      "epoch": 16.672661870503596,
+      "step": 18540
+    },
+    {
+      "epoch": 16.672661870503596,
+      "step": 18540,
+      "torque_loss": 0.11590138077735901
+    },
+    {
+      "epoch": 16.672661870503596,
+      "force_loss": 0.004519374575465918,
+      "step": 18540
+    },
+    {
+      "epoch": 16.681654676258994,
+      "grad_norm": 0.23871681094169617,
+      "learning_rate": 8.273565201832602e-05,
+      "loss": 0.0214,
+      "step": 18550
+    },
+    {
+      "action_loss": 0.01045029517263174,
+      "epoch": 16.681654676258994,
+      "step": 18550
+    },
+    {
+      "epoch": 16.681654676258994,
+      "step": 18550,
+      "torque_loss": 0.15892839431762695
+    },
+    {
+      "epoch": 16.681654676258994,
+      "force_loss": 0.009049165062606335,
+      "step": 18550
+    },
+    {
+      "epoch": 16.690647482014388,
+      "grad_norm": 0.27389827370643616,
+      "learning_rate": 8.27148167026738e-05,
+      "loss": 0.0219,
+      "step": 18560
+    },
+    {
+      "action_loss": 0.007756886538118124,
+      "epoch": 16.690647482014388,
+      "step": 18560
+    },
+    {
+      "epoch": 16.690647482014388,
+      "step": 18560,
+      "torque_loss": 0.1800699681043625
+    },
+    {
+      "epoch": 16.690647482014388,
+      "force_loss": 0.004350780043751001,
+      "step": 18560
+    },
+    {
+      "epoch": 16.699640287769785,
+      "grad_norm": 0.3494758903980255,
+      "learning_rate": 8.269397144912405e-05,
+      "loss": 0.0201,
+      "step": 18570
+    },
+    {
+      "action_loss": 0.005796968936920166,
+      "epoch": 16.699640287769785,
+      "step": 18570
+    },
+    {
+      "epoch": 16.699640287769785,
+      "step": 18570,
+      "torque_loss": 0.13563187420368195
+    },
+    {
+      "epoch": 16.699640287769785,
+      "force_loss": 0.0042900522239506245,
+      "step": 18570
+    },
+    {
+      "epoch": 16.70863309352518,
+      "grad_norm": 0.1897176206111908,
+      "learning_rate": 8.267311626400899e-05,
+      "loss": 0.0203,
+      "step": 18580
+    },
+    {
+      "action_loss": 0.0036982849705964327,
+      "epoch": 16.70863309352518,
+      "step": 18580
+    },
+    {
+      "epoch": 16.70863309352518,
+      "step": 18580,
+      "torque_loss": 0.13515764474868774
+    },
+    {
+      "epoch": 16.70863309352518,
+      "force_loss": 0.004321611020714045,
+      "step": 18580
+    },
+    {
+      "epoch": 16.717625899280577,
+      "grad_norm": 0.648822546005249,
+      "learning_rate": 8.26522511536639e-05,
+      "loss": 0.0266,
+      "step": 18590
+    },
+    {
+      "action_loss": 0.005962255876511335,
+      "epoch": 16.717625899280577,
+      "step": 18590
+    },
+    {
+      "epoch": 16.717625899280577,
+      "step": 18590,
+      "torque_loss": 0.1289035528898239
+    },
+    {
+      "epoch": 16.717625899280577,
+      "force_loss": 0.007715058047324419,
+      "step": 18590
+    },
+    {
+      "epoch": 16.72661870503597,
+      "grad_norm": 0.6050834059715271,
+      "learning_rate": 8.263137612442706e-05,
+      "loss": 0.0228,
+      "step": 18600
+    },
+    {
+      "action_loss": 0.00279523734934628,
+      "epoch": 16.72661870503597,
+      "step": 18600
+    },
+    {
+      "epoch": 16.72661870503597,
+      "step": 18600,
+      "torque_loss": 0.1604495793581009
+    },
+    {
+      "epoch": 16.72661870503597,
+      "force_loss": 0.003315134672448039,
+      "step": 18600
+    },
+    {
+      "epoch": 16.735611510791365,
+      "grad_norm": 0.5683088898658752,
+      "learning_rate": 8.261049118263971e-05,
+      "loss": 0.0236,
+      "step": 18610
+    },
+    {
+      "action_loss": 0.00667849974706769,
+      "epoch": 16.735611510791365,
+      "step": 18610
+    },
+    {
+      "epoch": 16.735611510791365,
+      "step": 18610,
+      "torque_loss": 0.10446900129318237
+    },
+    {
+      "epoch": 16.735611510791365,
+      "force_loss": 0.005121113266795874,
+      "step": 18610
+    },
+    {
+      "epoch": 16.744604316546763,
+      "grad_norm": 0.312840074300766,
+      "learning_rate": 8.258959633464619e-05,
+      "loss": 0.0256,
+      "step": 18620
+    },
+    {
+      "action_loss": 0.011705388315021992,
+      "epoch": 16.744604316546763,
+      "step": 18620
+    },
+    {
+      "epoch": 16.744604316546763,
+      "step": 18620,
+      "torque_loss": 0.139545276761055
+    },
+    {
+      "epoch": 16.744604316546763,
+      "force_loss": 0.01052594929933548,
+      "step": 18620
+    },
+    {
+      "epoch": 16.753597122302157,
+      "grad_norm": 0.23400932550430298,
+      "learning_rate": 8.256869158679377e-05,
+      "loss": 0.0215,
+      "step": 18630
+    },
+    {
+      "action_loss": 0.013369572348892689,
+      "epoch": 16.753597122302157,
+      "step": 18630
+    },
+    {
+      "epoch": 16.753597122302157,
+      "step": 18630,
+      "torque_loss": 0.14347916841506958
+    },
+    {
+      "epoch": 16.753597122302157,
+      "force_loss": 0.02529866062104702,
+      "step": 18630
+    },
+    {
+      "epoch": 16.762589928057555,
+      "grad_norm": 0.2445773333311081,
+      "learning_rate": 8.254777694543278e-05,
+      "loss": 0.0244,
+      "step": 18640
+    },
+    {
+      "action_loss": 0.003455832600593567,
+      "epoch": 16.762589928057555,
+      "step": 18640
+    },
+    {
+      "epoch": 16.762589928057555,
+      "step": 18640,
+      "torque_loss": 0.10403459519147873
+    },
+    {
+      "epoch": 16.762589928057555,
+      "force_loss": 0.002834821818396449,
+      "step": 18640
+    },
+    {
+      "epoch": 16.77158273381295,
+      "grad_norm": 0.170840322971344,
+      "learning_rate": 8.252685241691651e-05,
+      "loss": 0.0223,
+      "step": 18650
+    },
+    {
+      "action_loss": 0.014742475003004074,
+      "epoch": 16.77158273381295,
+      "step": 18650
+    },
+    {
+      "epoch": 16.77158273381295,
+      "step": 18650,
+      "torque_loss": 0.1419486254453659
+    },
+    {
+      "epoch": 16.77158273381295,
+      "force_loss": 0.013319268822669983,
+      "step": 18650
+    },
+    {
+      "epoch": 16.780575539568346,
+      "grad_norm": 0.25448960065841675,
+      "learning_rate": 8.250591800760133e-05,
+      "loss": 0.0259,
+      "step": 18660
+    },
+    {
+      "action_loss": 0.0067754946649074554,
+      "epoch": 16.780575539568346,
+      "step": 18660
+    },
+    {
+      "epoch": 16.780575539568346,
+      "step": 18660,
+      "torque_loss": 0.12209559231996536
+    },
+    {
+      "epoch": 16.780575539568346,
+      "force_loss": 0.006957089062780142,
+      "step": 18660
+    },
+    {
+      "epoch": 16.78956834532374,
+      "grad_norm": 0.46032240986824036,
+      "learning_rate": 8.248497372384649e-05,
+      "loss": 0.0233,
+      "step": 18670
+    },
+    {
+      "action_loss": 0.005185444373637438,
+      "epoch": 16.78956834532374,
+      "step": 18670
+    },
+    {
+      "epoch": 16.78956834532374,
+      "step": 18670,
+      "torque_loss": 0.13454100489616394
+    },
+    {
+      "epoch": 16.78956834532374,
+      "force_loss": 0.004870919976383448,
+      "step": 18670
+    },
+    {
+      "epoch": 16.798561151079138,
+      "grad_norm": 0.3743554949760437,
+      "learning_rate": 8.246401957201437e-05,
+      "loss": 0.0226,
+      "step": 18680
+    },
+    {
+      "action_loss": 0.01026403158903122,
+      "epoch": 16.798561151079138,
+      "step": 18680
+    },
+    {
+      "epoch": 16.798561151079138,
+      "step": 18680,
+      "torque_loss": 0.09702879190444946
+    },
+    {
+      "epoch": 16.798561151079138,
+      "force_loss": 0.00985734909772873,
+      "step": 18680
+    },
+    {
+      "epoch": 16.807553956834532,
+      "grad_norm": 0.5881900787353516,
+      "learning_rate": 8.244305555847027e-05,
+      "loss": 0.0223,
+      "step": 18690
+    },
+    {
+      "action_loss": 0.006126718129962683,
+      "epoch": 16.807553956834532,
+      "step": 18690
+    },
+    {
+      "epoch": 16.807553956834532,
+      "step": 18690,
+      "torque_loss": 0.14007656276226044
+    },
+    {
+      "epoch": 16.807553956834532,
+      "force_loss": 0.005367644131183624,
+      "step": 18690
+    },
+    {
+      "epoch": 16.81654676258993,
+      "grad_norm": 0.1845822036266327,
+      "learning_rate": 8.24220816895825e-05,
+      "loss": 0.0208,
+      "step": 18700
+    },
+    {
+      "action_loss": 0.0038799226749688387,
+      "epoch": 16.81654676258993,
+      "step": 18700
+    },
+    {
+      "epoch": 16.81654676258993,
+      "step": 18700,
+      "torque_loss": 0.14139671623706818
+    },
+    {
+      "epoch": 16.81654676258993,
+      "force_loss": 0.004548170138150454,
+      "step": 18700
+    },
+    {
+      "epoch": 16.825539568345324,
+      "grad_norm": 0.33281707763671875,
+      "learning_rate": 8.240109797172237e-05,
+      "loss": 0.0238,
+      "step": 18710
+    },
+    {
+      "action_loss": 0.007475856691598892,
+      "epoch": 16.825539568345324,
+      "step": 18710
+    },
+    {
+      "epoch": 16.825539568345324,
+      "step": 18710,
+      "torque_loss": 0.12454714626073837
+    },
+    {
+      "epoch": 16.825539568345324,
+      "force_loss": 0.005322887096554041,
+      "step": 18710
+    },
+    {
+      "epoch": 16.834532374100718,
+      "grad_norm": 0.45379719138145447,
+      "learning_rate": 8.238010441126416e-05,
+      "loss": 0.0307,
+      "step": 18720
+    },
+    {
+      "action_loss": 0.00832696445286274,
+      "epoch": 16.834532374100718,
+      "step": 18720
+    },
+    {
+      "epoch": 16.834532374100718,
+      "step": 18720,
+      "torque_loss": 0.14609622955322266
+    },
+    {
+      "epoch": 16.834532374100718,
+      "force_loss": 0.01168263703584671,
+      "step": 18720
+    },
+    {
+      "epoch": 16.843525179856115,
+      "grad_norm": 0.2713542878627777,
+      "learning_rate": 8.23591010145852e-05,
+      "loss": 0.0274,
+      "step": 18730
+    },
+    {
+      "action_loss": 0.023935288190841675,
+      "epoch": 16.843525179856115,
+      "step": 18730
+    },
+    {
+      "epoch": 16.843525179856115,
+      "step": 18730,
+      "torque_loss": 0.16119518876075745
+    },
+    {
+      "epoch": 16.843525179856115,
+      "force_loss": 0.03651067987084389,
+      "step": 18730
+    },
+    {
+      "epoch": 16.85251798561151,
+      "grad_norm": 0.3632880747318268,
+      "learning_rate": 8.233808778806571e-05,
+      "loss": 0.0273,
+      "step": 18740
+    },
+    {
+      "action_loss": 0.007144462782889605,
+      "epoch": 16.85251798561151,
+      "step": 18740
+    },
+    {
+      "epoch": 16.85251798561151,
+      "step": 18740,
+      "torque_loss": 0.18217028677463531
+    },
+    {
+      "epoch": 16.85251798561151,
+      "force_loss": 0.0072443666867911816,
+      "step": 18740
+    },
+    {
+      "epoch": 16.861510791366907,
+      "grad_norm": 0.4320412874221802,
+      "learning_rate": 8.231706473808903e-05,
+      "loss": 0.024,
+      "step": 18750
+    },
+    {
+      "action_loss": 0.010637740604579449,
+      "epoch": 16.861510791366907,
+      "step": 18750
+    },
+    {
+      "epoch": 16.861510791366907,
+      "step": 18750,
+      "torque_loss": 0.15085817873477936
+    },
+    {
+      "epoch": 16.861510791366907,
+      "force_loss": 0.01596071571111679,
+      "step": 18750
+    },
+    {
+      "epoch": 16.8705035971223,
+      "grad_norm": 0.4479663074016571,
+      "learning_rate": 8.229603187104133e-05,
+      "loss": 0.0294,
+      "step": 18760
+    },
+    {
+      "action_loss": 0.004178466275334358,
+      "epoch": 16.8705035971223,
+      "step": 18760
+    },
+    {
+      "epoch": 16.8705035971223,
+      "step": 18760,
+      "torque_loss": 0.08190203458070755
+    },
+    {
+      "epoch": 16.8705035971223,
+      "force_loss": 0.0028483474161475897,
+      "step": 18760
+    },
+    {
+      "epoch": 16.8794964028777,
+      "grad_norm": 0.780994176864624,
+      "learning_rate": 8.22749891933119e-05,
+      "loss": 0.0244,
+      "step": 18770
+    },
+    {
+      "action_loss": 0.023183442652225494,
+      "epoch": 16.8794964028777,
+      "step": 18770
+    },
+    {
+      "epoch": 16.8794964028777,
+      "step": 18770,
+      "torque_loss": 0.18326187133789062
+    },
+    {
+      "epoch": 16.8794964028777,
+      "force_loss": 0.014718101359903812,
+      "step": 18770
+    },
+    {
+      "epoch": 16.888489208633093,
+      "grad_norm": 0.62701815366745,
+      "learning_rate": 8.225393671129291e-05,
+      "loss": 0.029,
+      "step": 18780
+    },
+    {
+      "action_loss": 0.008167502470314503,
+      "epoch": 16.888489208633093,
+      "step": 18780
+    },
+    {
+      "epoch": 16.888489208633093,
+      "step": 18780,
+      "torque_loss": 0.1430051028728485
+    },
+    {
+      "epoch": 16.888489208633093,
+      "force_loss": 0.0054850452579557896,
+      "step": 18780
+    },
+    {
+      "epoch": 16.89748201438849,
+      "grad_norm": 0.26363375782966614,
+      "learning_rate": 8.223287443137957e-05,
+      "loss": 0.0211,
+      "step": 18790
+    },
+    {
+      "action_loss": 0.0046821762807667255,
+      "epoch": 16.89748201438849,
+      "step": 18790
+    },
+    {
+      "epoch": 16.89748201438849,
+      "step": 18790,
+      "torque_loss": 0.12745413184165955
+    },
+    {
+      "epoch": 16.89748201438849,
+      "force_loss": 0.003840351477265358,
+      "step": 18790
+    },
+    {
+      "epoch": 16.906474820143885,
+      "grad_norm": 0.70247882604599,
+      "learning_rate": 8.221180235997004e-05,
+      "loss": 0.0231,
+      "step": 18800
+    },
+    {
+      "action_loss": 0.003294161520898342,
+      "epoch": 16.906474820143885,
+      "step": 18800
+    },
+    {
+      "epoch": 16.906474820143885,
+      "step": 18800,
+      "torque_loss": 0.13857722282409668
+    },
+    {
+      "epoch": 16.906474820143885,
+      "force_loss": 0.00698641873896122,
+      "step": 18800
+    },
+    {
+      "epoch": 16.915467625899282,
+      "grad_norm": 0.18306425213813782,
+      "learning_rate": 8.219072050346544e-05,
+      "loss": 0.0219,
+      "step": 18810
+    },
+    {
+      "action_loss": 0.006904096808284521,
+      "epoch": 16.915467625899282,
+      "step": 18810
+    },
+    {
+      "epoch": 16.915467625899282,
+      "step": 18810,
+      "torque_loss": 0.13770753145217896
+    },
+    {
+      "epoch": 16.915467625899282,
+      "force_loss": 0.009075612761080265,
+      "step": 18810
+    },
+    {
+      "epoch": 16.924460431654676,
+      "grad_norm": 0.3748214542865753,
+      "learning_rate": 8.216962886826992e-05,
+      "loss": 0.0222,
+      "step": 18820
+    },
+    {
+      "action_loss": 0.0046668448485434055,
+      "epoch": 16.924460431654676,
+      "step": 18820
+    },
+    {
+      "epoch": 16.924460431654676,
+      "step": 18820,
+      "torque_loss": 0.17053575813770294
+    },
+    {
+      "epoch": 16.924460431654676,
+      "force_loss": 0.004429507534950972,
+      "step": 18820
+    },
+    {
+      "epoch": 16.93345323741007,
+      "grad_norm": 0.7287952303886414,
+      "learning_rate": 8.214852746079054e-05,
+      "loss": 0.0264,
+      "step": 18830
+    },
+    {
+      "action_loss": 0.006570905912667513,
+      "epoch": 16.93345323741007,
+      "step": 18830
+    },
+    {
+      "epoch": 16.93345323741007,
+      "step": 18830,
+      "torque_loss": 0.16397243738174438
+    },
+    {
+      "epoch": 16.93345323741007,
+      "force_loss": 0.002944026840850711,
+      "step": 18830
+    },
+    {
+      "epoch": 16.942446043165468,
+      "grad_norm": 0.40586352348327637,
+      "learning_rate": 8.212741628743732e-05,
+      "loss": 0.0206,
+      "step": 18840
+    },
+    {
+      "action_loss": 0.0093872444704175,
+      "epoch": 16.942446043165468,
+      "step": 18840
+    },
+    {
+      "epoch": 16.942446043165468,
+      "step": 18840,
+      "torque_loss": 0.1676502227783203
+    },
+    {
+      "epoch": 16.942446043165468,
+      "force_loss": 0.005877931136637926,
+      "step": 18840
+    },
+    {
+      "epoch": 16.951438848920862,
+      "grad_norm": 0.3222416937351227,
+      "learning_rate": 8.210629535462333e-05,
+      "loss": 0.0239,
+      "step": 18850
+    },
+    {
+      "action_loss": 0.006561322137713432,
+      "epoch": 16.951438848920862,
+      "step": 18850
+    },
+    {
+      "epoch": 16.951438848920862,
+      "step": 18850,
+      "torque_loss": 0.08255260437726974
+    },
+    {
+      "epoch": 16.951438848920862,
+      "force_loss": 0.007167476695030928,
+      "step": 18850
+    },
+    {
+      "epoch": 16.96043165467626,
+      "grad_norm": 0.3633570671081543,
+      "learning_rate": 8.208516466876453e-05,
+      "loss": 0.0221,
+      "step": 18860
+    },
+    {
+      "action_loss": 0.009111897088587284,
+      "epoch": 16.96043165467626,
+      "step": 18860
+    },
+    {
+      "epoch": 16.96043165467626,
+      "step": 18860,
+      "torque_loss": 0.1447877287864685
+    },
+    {
+      "epoch": 16.96043165467626,
+      "force_loss": 0.011469370685517788,
+      "step": 18860
+    },
+    {
+      "epoch": 16.969424460431654,
+      "grad_norm": 0.457709401845932,
+      "learning_rate": 8.206402423627986e-05,
+      "loss": 0.0199,
+      "step": 18870
+    },
+    {
+      "action_loss": 0.00664312532171607,
+      "epoch": 16.969424460431654,
+      "step": 18870
+    },
+    {
+      "epoch": 16.969424460431654,
+      "step": 18870,
+      "torque_loss": 0.11739363521337509
+    },
+    {
+      "epoch": 16.969424460431654,
+      "force_loss": 0.011712410487234592,
+      "step": 18870
+    },
+    {
+      "epoch": 16.97841726618705,
+      "grad_norm": 0.3963375389575958,
+      "learning_rate": 8.204287406359124e-05,
+      "loss": 0.0311,
+      "step": 18880
+    },
+    {
+      "action_loss": 0.005666142795234919,
+      "epoch": 16.97841726618705,
+      "step": 18880
+    },
+    {
+      "epoch": 16.97841726618705,
+      "step": 18880,
+      "torque_loss": 0.13782495260238647
+    },
+    {
+      "epoch": 16.97841726618705,
+      "force_loss": 0.007117301691323519,
+      "step": 18880
+    },
+    {
+      "epoch": 16.987410071942445,
+      "grad_norm": 0.42669326066970825,
+      "learning_rate": 8.20217141571235e-05,
+      "loss": 0.0233,
+      "step": 18890
+    },
+    {
+      "action_loss": 0.0026051367167383432,
+      "epoch": 16.987410071942445,
+      "step": 18890
+    },
+    {
+      "epoch": 16.987410071942445,
+      "step": 18890,
+      "torque_loss": 0.10621228069067001
+    },
+    {
+      "epoch": 16.987410071942445,
+      "force_loss": 0.0038599625695496798,
+      "step": 18890
+    },
+    {
+      "epoch": 16.996402877697843,
+      "grad_norm": 0.2006705105304718,
+      "learning_rate": 8.200054452330449e-05,
+      "loss": 0.0187,
+      "step": 18900
+    },
+    {
+      "action_loss": 0.0020557413809001446,
+      "epoch": 16.996402877697843,
+      "step": 18900
+    },
+    {
+      "epoch": 16.996402877697843,
+      "step": 18900,
+      "torque_loss": 0.12934191524982452
+    },
+    {
+      "epoch": 16.996402877697843,
+      "force_loss": 0.0024156435392796993,
+      "step": 18900
+    },
+    {
+      "epoch": 17.005395683453237,
+      "grad_norm": 0.36919236183166504,
+      "learning_rate": 8.197936516856499e-05,
+      "loss": 0.0226,
+      "step": 18910
+    },
+    {
+      "action_loss": 0.007828912697732449,
+      "epoch": 17.005395683453237,
+      "step": 18910
+    },
+    {
+      "epoch": 17.005395683453237,
+      "step": 18910,
+      "torque_loss": 0.13782049715518951
+    },
+    {
+      "epoch": 17.005395683453237,
+      "force_loss": 0.008543119765818119,
+      "step": 18910
+    },
+    {
+      "epoch": 17.014388489208635,
+      "grad_norm": 0.32674938440322876,
+      "learning_rate": 8.195817609933871e-05,
+      "loss": 0.0314,
+      "step": 18920
+    },
+    {
+      "action_loss": 0.005228050053119659,
+      "epoch": 17.014388489208635,
+      "step": 18920
+    },
+    {
+      "epoch": 17.014388489208635,
+      "step": 18920,
+      "torque_loss": 0.11597766727209091
+    },
+    {
+      "epoch": 17.014388489208635,
+      "force_loss": 0.0029003219678997993,
+      "step": 18920
+    },
+    {
+      "epoch": 17.02338129496403,
+      "grad_norm": 0.4104541838169098,
+      "learning_rate": 8.193697732206233e-05,
+      "loss": 0.0209,
+      "step": 18930
+    },
+    {
+      "action_loss": 0.007867767475545406,
+      "epoch": 17.02338129496403,
+      "step": 18930
+    },
+    {
+      "epoch": 17.02338129496403,
+      "step": 18930,
+      "torque_loss": 0.12521807849407196
+    },
+    {
+      "epoch": 17.02338129496403,
+      "force_loss": 0.005270702298730612,
+      "step": 18930
+    },
+    {
+      "epoch": 17.032374100719423,
+      "grad_norm": 0.4457932710647583,
+      "learning_rate": 8.19157688431755e-05,
+      "loss": 0.0259,
+      "step": 18940
+    },
+    {
+      "action_loss": 0.007062478456646204,
+      "epoch": 17.032374100719423,
+      "step": 18940
+    },
+    {
+      "epoch": 17.032374100719423,
+      "step": 18940,
+      "torque_loss": 0.15664128959178925
+    },
+    {
+      "epoch": 17.032374100719423,
+      "force_loss": 0.004922620952129364,
+      "step": 18940
+    },
+    {
+      "epoch": 17.04136690647482,
+      "grad_norm": 0.8380558490753174,
+      "learning_rate": 8.189455066912077e-05,
+      "loss": 0.0229,
+      "step": 18950
+    },
+    {
+      "action_loss": 0.010789395309984684,
+      "epoch": 17.04136690647482,
+      "step": 18950
+    },
+    {
+      "epoch": 17.04136690647482,
+      "step": 18950,
+      "torque_loss": 0.14502167701721191
+    },
+    {
+      "epoch": 17.04136690647482,
+      "force_loss": 0.009628080762922764,
+      "step": 18950
+    },
+    {
+      "epoch": 17.050359712230215,
+      "grad_norm": 0.4603548049926758,
+      "learning_rate": 8.187332280634369e-05,
+      "loss": 0.0238,
+      "step": 18960
+    },
+    {
+      "action_loss": 0.008165935985744,
+      "epoch": 17.050359712230215,
+      "step": 18960
+    },
+    {
+      "epoch": 17.050359712230215,
+      "step": 18960,
+      "torque_loss": 0.15990819036960602
+    },
+    {
+      "epoch": 17.050359712230215,
+      "force_loss": 0.01032984908670187,
+      "step": 18960
+    },
+    {
+      "epoch": 17.059352517985612,
+      "grad_norm": 0.33527544140815735,
+      "learning_rate": 8.18520852612927e-05,
+      "loss": 0.0238,
+      "step": 18970
+    },
+    {
+      "action_loss": 0.006526540964841843,
+      "epoch": 17.059352517985612,
+      "step": 18970
+    },
+    {
+      "epoch": 17.059352517985612,
+      "step": 18970,
+      "torque_loss": 0.17004112899303436
+    },
+    {
+      "epoch": 17.059352517985612,
+      "force_loss": 0.0052053299732506275,
+      "step": 18970
+    },
+    {
+      "epoch": 17.068345323741006,
+      "grad_norm": 0.46330389380455017,
+      "learning_rate": 8.183083804041921e-05,
+      "loss": 0.0269,
+      "step": 18980
+    },
+    {
+      "action_loss": 0.010260792449116707,
+      "epoch": 17.068345323741006,
+      "step": 18980
+    },
+    {
+      "epoch": 17.068345323741006,
+      "step": 18980,
+      "torque_loss": 0.14924046397209167
+    },
+    {
+      "epoch": 17.068345323741006,
+      "force_loss": 0.007903140969574451,
+      "step": 18980
+    },
+    {
+      "epoch": 17.077338129496404,
+      "grad_norm": 0.3768214285373688,
+      "learning_rate": 8.180958115017757e-05,
+      "loss": 0.0211,
+      "step": 18990
+    },
+    {
+      "action_loss": 0.003264196217060089,
+      "epoch": 17.077338129496404,
+      "step": 18990
+    },
+    {
+      "epoch": 17.077338129496404,
+      "step": 18990,
+      "torque_loss": 0.12123546749353409
+    },
+    {
+      "epoch": 17.077338129496404,
+      "force_loss": 0.0032893174793571234,
+      "step": 18990
+    },
+    {
+      "epoch": 17.086330935251798,
+      "grad_norm": 0.6482235193252563,
+      "learning_rate": 8.178831459702505e-05,
+      "loss": 0.0306,
+      "step": 19000
+    },
+    {
+      "action_loss": 0.0057056681253015995,
+      "epoch": 17.086330935251798,
+      "step": 19000
+    },
+    {
+      "epoch": 17.086330935251798,
+      "step": 19000,
+      "torque_loss": 0.14882366359233856
+    },
+    {
+      "epoch": 17.086330935251798,
+      "force_loss": 0.005644096527248621,
+      "step": 19000
+    },
+    {
+      "epoch": 17.095323741007196,
+      "grad_norm": 0.423765629529953,
+      "learning_rate": 8.17670383874219e-05,
+      "loss": 0.0254,
+      "step": 19010
+    },
+    {
+      "action_loss": 0.02130795083940029,
+      "epoch": 17.095323741007196,
+      "step": 19010
+    },
+    {
+      "epoch": 17.095323741007196,
+      "step": 19010,
+      "torque_loss": 0.12126599997282028
+    },
+    {
+      "epoch": 17.095323741007196,
+      "force_loss": 0.014218921773135662,
+      "step": 19010
+    },
+    {
+      "epoch": 17.10431654676259,
+      "grad_norm": 0.22702963650226593,
+      "learning_rate": 8.174575252783124e-05,
+      "loss": 0.0208,
+      "step": 19020
+    },
+    {
+      "action_loss": 0.02861572802066803,
+      "epoch": 17.10431654676259,
+      "step": 19020
+    },
+    {
+      "epoch": 17.10431654676259,
+      "step": 19020,
+      "torque_loss": 0.27991050481796265
+    },
+    {
+      "epoch": 17.10431654676259,
+      "force_loss": 0.05038939043879509,
+      "step": 19020
+    },
+    {
+      "epoch": 17.113309352517987,
+      "grad_norm": 0.30597561597824097,
+      "learning_rate": 8.172445702471914e-05,
+      "loss": 0.0301,
+      "step": 19030
+    },
+    {
+      "action_loss": 0.013531462289392948,
+      "epoch": 17.113309352517987,
+      "step": 19030
+    },
+    {
+      "epoch": 17.113309352517987,
+      "step": 19030,
+      "torque_loss": 0.16070210933685303
+    },
+    {
+      "epoch": 17.113309352517987,
+      "force_loss": 0.005723575595766306,
+      "step": 19030
+    },
+    {
+      "epoch": 17.12230215827338,
+      "grad_norm": 0.5818365812301636,
+      "learning_rate": 8.170315188455466e-05,
+      "loss": 0.0259,
+      "step": 19040
+    },
+    {
+      "action_loss": 0.01446501910686493,
+      "epoch": 17.12230215827338,
+      "step": 19040
+    },
+    {
+      "epoch": 17.12230215827338,
+      "step": 19040,
+      "torque_loss": 0.10388970375061035
+    },
+    {
+      "epoch": 17.12230215827338,
+      "force_loss": 0.012621755711734295,
+      "step": 19040
+    },
+    {
+      "epoch": 17.131294964028775,
+      "grad_norm": 1.3864644765853882,
+      "learning_rate": 8.168183711380969e-05,
+      "loss": 0.0246,
+      "step": 19050
+    },
+    {
+      "action_loss": 0.005052044056355953,
+      "epoch": 17.131294964028775,
+      "step": 19050
+    },
+    {
+      "epoch": 17.131294964028775,
+      "step": 19050,
+      "torque_loss": 0.18764309585094452
+    },
+    {
+      "epoch": 17.131294964028775,
+      "force_loss": 0.004084060899913311,
+      "step": 19050
+    },
+    {
+      "epoch": 17.140287769784173,
+      "grad_norm": 0.3341905176639557,
+      "learning_rate": 8.166051271895913e-05,
+      "loss": 0.0263,
+      "step": 19060
+    },
+    {
+      "action_loss": 0.009814356453716755,
+      "epoch": 17.140287769784173,
+      "step": 19060
+    },
+    {
+      "epoch": 17.140287769784173,
+      "step": 19060,
+      "torque_loss": 0.12652598321437836
+    },
+    {
+      "epoch": 17.140287769784173,
+      "force_loss": 0.005368433892726898,
+      "step": 19060
+    },
+    {
+      "epoch": 17.149280575539567,
+      "grad_norm": 0.6932147741317749,
+      "learning_rate": 8.163917870648075e-05,
+      "loss": 0.0352,
+      "step": 19070
+    },
+    {
+      "action_loss": 0.014251135289669037,
+      "epoch": 17.149280575539567,
+      "step": 19070
+    },
+    {
+      "epoch": 17.149280575539567,
+      "step": 19070,
+      "torque_loss": 0.13219048082828522
+    },
+    {
+      "epoch": 17.149280575539567,
+      "force_loss": 0.0063622742891311646,
+      "step": 19070
+    },
+    {
+      "epoch": 17.158273381294965,
+      "grad_norm": 0.46337074041366577,
+      "learning_rate": 8.161783508285526e-05,
+      "loss": 0.0316,
+      "step": 19080
+    },
+    {
+      "action_loss": 0.05375169590115547,
+      "epoch": 17.158273381294965,
+      "step": 19080
+    },
+    {
+      "epoch": 17.158273381294965,
+      "step": 19080,
+      "torque_loss": 0.16900013387203217
+    },
+    {
+      "epoch": 17.158273381294965,
+      "force_loss": 0.028509723022580147,
+      "step": 19080
+    },
+    {
+      "epoch": 17.16726618705036,
+      "grad_norm": 0.3895530104637146,
+      "learning_rate": 8.159648185456628e-05,
+      "loss": 0.038,
+      "step": 19090
+    },
+    {
+      "action_loss": 0.008286750875413418,
+      "epoch": 17.16726618705036,
+      "step": 19090
+    },
+    {
+      "epoch": 17.16726618705036,
+      "step": 19090,
+      "torque_loss": 0.09403803199529648
+    },
+    {
+      "epoch": 17.16726618705036,
+      "force_loss": 0.0056198700331151485,
+      "step": 19090
+    },
+    {
+      "epoch": 17.176258992805757,
+      "grad_norm": 0.5421538949012756,
+      "learning_rate": 8.157511902810038e-05,
+      "loss": 0.0365,
+      "step": 19100
+    },
+    {
+      "action_loss": 0.012842250056564808,
+      "epoch": 17.176258992805757,
+      "step": 19100
+    },
+    {
+      "epoch": 17.176258992805757,
+      "step": 19100,
+      "torque_loss": 0.13496725261211395
+    },
+    {
+      "epoch": 17.176258992805757,
+      "force_loss": 0.013069468550384045,
+      "step": 19100
+    },
+    {
+      "epoch": 17.18525179856115,
+      "grad_norm": 0.8404548168182373,
+      "learning_rate": 8.155374660994701e-05,
+      "loss": 0.0267,
+      "step": 19110
+    },
+    {
+      "action_loss": 0.007729492615908384,
+      "epoch": 17.18525179856115,
+      "step": 19110
+    },
+    {
+      "epoch": 17.18525179856115,
+      "step": 19110,
+      "torque_loss": 0.11826196312904358
+    },
+    {
+      "epoch": 17.18525179856115,
+      "force_loss": 0.004309921991080046,
+      "step": 19110
+    },
+    {
+      "epoch": 17.194244604316548,
+      "grad_norm": 0.38744139671325684,
+      "learning_rate": 8.153236460659857e-05,
+      "loss": 0.0333,
+      "step": 19120
+    },
+    {
+      "action_loss": 0.011896542273461819,
+      "epoch": 17.194244604316548,
+      "step": 19120
+    },
+    {
+      "epoch": 17.194244604316548,
+      "step": 19120,
+      "torque_loss": 0.1816953867673874
+    },
+    {
+      "epoch": 17.194244604316548,
+      "force_loss": 0.008728374727070332,
+      "step": 19120
+    },
+    {
+      "epoch": 17.203237410071942,
+      "grad_norm": 0.5814850330352783,
+      "learning_rate": 8.151097302455031e-05,
+      "loss": 0.0246,
+      "step": 19130
+    },
+    {
+      "action_loss": 0.008752156980335712,
+      "epoch": 17.203237410071942,
+      "step": 19130
+    },
+    {
+      "epoch": 17.203237410071942,
+      "step": 19130,
+      "torque_loss": 0.14554281532764435
+    },
+    {
+      "epoch": 17.203237410071942,
+      "force_loss": 0.005143011454492807,
+      "step": 19130
+    },
+    {
+      "epoch": 17.21223021582734,
+      "grad_norm": 0.6042574644088745,
+      "learning_rate": 8.148957187030044e-05,
+      "loss": 0.03,
+      "step": 19140
+    },
+    {
+      "action_loss": 0.0060353465378284454,
+      "epoch": 17.21223021582734,
+      "step": 19140
+    },
+    {
+      "epoch": 17.21223021582734,
+      "step": 19140,
+      "torque_loss": 0.17903532087802887
+    },
+    {
+      "epoch": 17.21223021582734,
+      "force_loss": 0.0028662399854511023,
+      "step": 19140
+    },
+    {
+      "epoch": 17.221223021582734,
+      "grad_norm": 0.42693963646888733,
+      "learning_rate": 8.146816115035006e-05,
+      "loss": 0.0263,
+      "step": 19150
+    },
+    {
+      "action_loss": 0.0046659205108881,
+      "epoch": 17.221223021582734,
+      "step": 19150
+    },
+    {
+      "epoch": 17.221223021582734,
+      "step": 19150,
+      "torque_loss": 0.13491903245449066
+    },
+    {
+      "epoch": 17.221223021582734,
+      "force_loss": 0.0027668692637234926,
+      "step": 19150
+    },
+    {
+      "epoch": 17.230215827338128,
+      "grad_norm": 0.36125826835632324,
+      "learning_rate": 8.14467408712032e-05,
+      "loss": 0.0304,
+      "step": 19160
+    },
+    {
+      "action_loss": 0.0030811659526079893,
+      "epoch": 17.230215827338128,
+      "step": 19160
+    },
+    {
+      "epoch": 17.230215827338128,
+      "step": 19160,
+      "torque_loss": 0.11700669676065445
+    },
+    {
+      "epoch": 17.230215827338128,
+      "force_loss": 0.0032936956267803907,
+      "step": 19160
+    },
+    {
+      "epoch": 17.239208633093526,
+      "grad_norm": 0.3138788938522339,
+      "learning_rate": 8.142531103936678e-05,
+      "loss": 0.025,
+      "step": 19170
+    },
+    {
+      "action_loss": 0.006638602819293737,
+      "epoch": 17.239208633093526,
+      "step": 19170
+    },
+    {
+      "epoch": 17.239208633093526,
+      "step": 19170,
+      "torque_loss": 0.09379879385232925
+    },
+    {
+      "epoch": 17.239208633093526,
+      "force_loss": 0.0046833292581140995,
+      "step": 19170
+    },
+    {
+      "epoch": 17.24820143884892,
+      "grad_norm": 0.1880631446838379,
+      "learning_rate": 8.14038716613506e-05,
+      "loss": 0.0186,
+      "step": 19180
+    },
+    {
+      "action_loss": 0.01769532449543476,
+      "epoch": 17.24820143884892,
+      "step": 19180
+    },
+    {
+      "epoch": 17.24820143884892,
+      "step": 19180,
+      "torque_loss": 0.10830444097518921
+    },
+    {
+      "epoch": 17.24820143884892,
+      "force_loss": 0.01403549313545227,
+      "step": 19180
+    },
+    {
+      "epoch": 17.257194244604317,
+      "grad_norm": 0.4972555637359619,
+      "learning_rate": 8.138242274366736e-05,
+      "loss": 0.0234,
+      "step": 19190
+    },
+    {
+      "action_loss": 0.004576841834932566,
+      "epoch": 17.257194244604317,
+      "step": 19190
+    },
+    {
+      "epoch": 17.257194244604317,
+      "step": 19190,
+      "torque_loss": 0.11655453592538834
+    },
+    {
+      "epoch": 17.257194244604317,
+      "force_loss": 0.004046024288982153,
+      "step": 19190
+    },
+    {
+      "epoch": 17.26618705035971,
+      "grad_norm": 0.4342607855796814,
+      "learning_rate": 8.136096429283271e-05,
+      "loss": 0.0263,
+      "step": 19200
+    },
+    {
+      "action_loss": 0.006233396008610725,
+      "epoch": 17.26618705035971,
+      "step": 19200
+    },
+    {
+      "epoch": 17.26618705035971,
+      "step": 19200,
+      "torque_loss": 0.07891140133142471
+    },
+    {
+      "epoch": 17.26618705035971,
+      "force_loss": 0.005459946114569902,
+      "step": 19200
+    },
+    {
+      "epoch": 17.27517985611511,
+      "grad_norm": 0.6128469705581665,
+      "learning_rate": 8.133949631536515e-05,
+      "loss": 0.028,
+      "step": 19210
+    },
+    {
+      "action_loss": 0.012323073111474514,
+      "epoch": 17.27517985611511,
+      "step": 19210
+    },
+    {
+      "epoch": 17.27517985611511,
+      "step": 19210,
+      "torque_loss": 0.13757109642028809
+    },
+    {
+      "epoch": 17.27517985611511,
+      "force_loss": 0.00769123574718833,
+      "step": 19210
+    },
+    {
+      "epoch": 17.284172661870503,
+      "grad_norm": 0.5576527714729309,
+      "learning_rate": 8.131801881778607e-05,
+      "loss": 0.0293,
+      "step": 19220
+    },
+    {
+      "action_loss": 0.007438432425260544,
+      "epoch": 17.284172661870503,
+      "step": 19220
+    },
+    {
+      "epoch": 17.284172661870503,
+      "step": 19220,
+      "torque_loss": 0.0926814153790474
+    },
+    {
+      "epoch": 17.284172661870503,
+      "force_loss": 0.010401430539786816,
+      "step": 19220
+    },
+    {
+      "epoch": 17.2931654676259,
+      "grad_norm": 0.42516401410102844,
+      "learning_rate": 8.129653180661978e-05,
+      "loss": 0.0269,
+      "step": 19230
+    },
+    {
+      "action_loss": 0.006882320623844862,
+      "epoch": 17.2931654676259,
+      "step": 19230
+    },
+    {
+      "epoch": 17.2931654676259,
+      "step": 19230,
+      "torque_loss": 0.1604544073343277
+    },
+    {
+      "epoch": 17.2931654676259,
+      "force_loss": 0.0053317672573029995,
+      "step": 19230
+    },
+    {
+      "epoch": 17.302158273381295,
+      "grad_norm": 0.2955699563026428,
+      "learning_rate": 8.127503528839346e-05,
+      "loss": 0.0261,
+      "step": 19240
+    },
+    {
+      "action_loss": 0.01876833103597164,
+      "epoch": 17.302158273381295,
+      "step": 19240
+    },
+    {
+      "epoch": 17.302158273381295,
+      "step": 19240,
+      "torque_loss": 0.17759306728839874
+    },
+    {
+      "epoch": 17.302158273381295,
+      "force_loss": 0.015544223599135876,
+      "step": 19240
+    },
+    {
+      "epoch": 17.31115107913669,
+      "grad_norm": 0.5286865234375,
+      "learning_rate": 8.125352926963721e-05,
+      "loss": 0.0337,
+      "step": 19250
+    },
+    {
+      "action_loss": 0.004853442776948214,
+      "epoch": 17.31115107913669,
+      "step": 19250
+    },
+    {
+      "epoch": 17.31115107913669,
+      "step": 19250,
+      "torque_loss": 0.08871179074048996
+    },
+    {
+      "epoch": 17.31115107913669,
+      "force_loss": 0.003862435230985284,
+      "step": 19250
+    },
+    {
+      "epoch": 17.320143884892087,
+      "grad_norm": 0.4135963022708893,
+      "learning_rate": 8.123201375688395e-05,
+      "loss": 0.0239,
+      "step": 19260
+    },
+    {
+      "action_loss": 0.0029513679910451174,
+      "epoch": 17.320143884892087,
+      "step": 19260
+    },
+    {
+      "epoch": 17.320143884892087,
+      "step": 19260,
+      "torque_loss": 0.12203123420476913
+    },
+    {
+      "epoch": 17.320143884892087,
+      "force_loss": 0.0028603628743439913,
+      "step": 19260
+    },
+    {
+      "epoch": 17.32913669064748,
+      "grad_norm": 0.3187497556209564,
+      "learning_rate": 8.121048875666954e-05,
+      "loss": 0.0248,
+      "step": 19270
+    },
+    {
+      "action_loss": 0.00624515675008297,
+      "epoch": 17.32913669064748,
+      "step": 19270
+    },
+    {
+      "epoch": 17.32913669064748,
+      "step": 19270,
+      "torque_loss": 0.13843059539794922
+    },
+    {
+      "epoch": 17.32913669064748,
+      "force_loss": 0.0032186105381697416,
+      "step": 19270
+    },
+    {
+      "epoch": 17.33812949640288,
+      "grad_norm": 0.2930435836315155,
+      "learning_rate": 8.118895427553274e-05,
+      "loss": 0.0201,
+      "step": 19280
+    },
+    {
+      "action_loss": 0.00903305597603321,
+      "epoch": 17.33812949640288,
+      "step": 19280
+    },
+    {
+      "epoch": 17.33812949640288,
+      "step": 19280,
+      "torque_loss": 0.1215401291847229
+    },
+    {
+      "epoch": 17.33812949640288,
+      "force_loss": 0.00682861777022481,
+      "step": 19280
+    },
+    {
+      "epoch": 17.347122302158272,
+      "grad_norm": 0.37510836124420166,
+      "learning_rate": 8.116741032001511e-05,
+      "loss": 0.0248,
+      "step": 19290
+    },
+    {
+      "action_loss": 0.0066780210472643375,
+      "epoch": 17.347122302158272,
+      "step": 19290
+    },
+    {
+      "epoch": 17.347122302158272,
+      "step": 19290,
+      "torque_loss": 0.12911011278629303
+    },
+    {
+      "epoch": 17.347122302158272,
+      "force_loss": 0.005036401562392712,
+      "step": 19290
+    },
+    {
+      "epoch": 17.35611510791367,
+      "grad_norm": 0.5693814754486084,
+      "learning_rate": 8.114585689666114e-05,
+      "loss": 0.0251,
+      "step": 19300
+    },
+    {
+      "action_loss": 0.004626793321222067,
+      "epoch": 17.35611510791367,
+      "step": 19300
+    },
+    {
+      "epoch": 17.35611510791367,
+      "step": 19300,
+      "torque_loss": 0.11726341396570206
+    },
+    {
+      "epoch": 17.35611510791367,
+      "force_loss": 0.0034203932154923677,
+      "step": 19300
+    },
+    {
+      "epoch": 17.365107913669064,
+      "grad_norm": 0.35918116569519043,
+      "learning_rate": 8.112429401201821e-05,
+      "loss": 0.0208,
+      "step": 19310
+    },
+    {
+      "action_loss": 0.03519358113408089,
+      "epoch": 17.365107913669064,
+      "step": 19310
+    },
+    {
+      "epoch": 17.365107913669064,
+      "step": 19310,
+      "torque_loss": 0.21112732589244843
+    },
+    {
+      "epoch": 17.365107913669064,
+      "force_loss": 0.02771923877298832,
+      "step": 19310
+    },
+    {
+      "epoch": 17.37410071942446,
+      "grad_norm": 0.6362349987030029,
+      "learning_rate": 8.110272167263656e-05,
+      "loss": 0.0253,
+      "step": 19320
+    },
+    {
+      "action_loss": 0.011003575287759304,
+      "epoch": 17.37410071942446,
+      "step": 19320
+    },
+    {
+      "epoch": 17.37410071942446,
+      "step": 19320,
+      "torque_loss": 0.1277102380990982
+    },
+    {
+      "epoch": 17.37410071942446,
+      "force_loss": 0.007746491581201553,
+      "step": 19320
+    },
+    {
+      "epoch": 17.383093525179856,
+      "grad_norm": 0.27018672227859497,
+      "learning_rate": 8.108113988506929e-05,
+      "loss": 0.0234,
+      "step": 19330
+    },
+    {
+      "action_loss": 0.011912592686712742,
+      "epoch": 17.383093525179856,
+      "step": 19330
+    },
+    {
+      "epoch": 17.383093525179856,
+      "step": 19330,
+      "torque_loss": 0.19017989933490753
+    },
+    {
+      "epoch": 17.383093525179856,
+      "force_loss": 0.015125956386327744,
+      "step": 19330
+    },
+    {
+      "epoch": 17.392086330935253,
+      "grad_norm": 0.3717755675315857,
+      "learning_rate": 8.105954865587235e-05,
+      "loss": 0.0242,
+      "step": 19340
+    },
+    {
+      "action_loss": 0.006878519430756569,
+      "epoch": 17.392086330935253,
+      "step": 19340
+    },
+    {
+      "epoch": 17.392086330935253,
+      "step": 19340,
+      "torque_loss": 0.10891749709844589
+    },
+    {
+      "epoch": 17.392086330935253,
+      "force_loss": 0.004361771512776613,
+      "step": 19340
+    },
+    {
+      "epoch": 17.401079136690647,
+      "grad_norm": 0.25477027893066406,
+      "learning_rate": 8.103794799160463e-05,
+      "loss": 0.0221,
+      "step": 19350
+    },
+    {
+      "action_loss": 0.037407178431749344,
+      "epoch": 17.401079136690647,
+      "step": 19350
+    },
+    {
+      "epoch": 17.401079136690647,
+      "step": 19350,
+      "torque_loss": 0.20230621099472046
+    },
+    {
+      "epoch": 17.401079136690647,
+      "force_loss": 0.04858658090233803,
+      "step": 19350
+    },
+    {
+      "epoch": 17.41007194244604,
+      "grad_norm": 0.6735423803329468,
+      "learning_rate": 8.101633789882781e-05,
+      "loss": 0.026,
+      "step": 19360
+    },
+    {
+      "action_loss": 0.004760397132486105,
+      "epoch": 17.41007194244604,
+      "step": 19360
+    },
+    {
+      "epoch": 17.41007194244604,
+      "step": 19360,
+      "torque_loss": 0.1070166826248169
+    },
+    {
+      "epoch": 17.41007194244604,
+      "force_loss": 0.003806935390457511,
+      "step": 19360
+    },
+    {
+      "epoch": 17.41906474820144,
+      "grad_norm": 0.4031885862350464,
+      "learning_rate": 8.099471838410648e-05,
+      "loss": 0.0193,
+      "step": 19370
+    },
+    {
+      "action_loss": 0.010101772844791412,
+      "epoch": 17.41906474820144,
+      "step": 19370
+    },
+    {
+      "epoch": 17.41906474820144,
+      "step": 19370,
+      "torque_loss": 0.15674875676631927
+    },
+    {
+      "epoch": 17.41906474820144,
+      "force_loss": 0.008931778371334076,
+      "step": 19370
+    },
+    {
+      "epoch": 17.428057553956833,
+      "grad_norm": 0.4902113378047943,
+      "learning_rate": 8.097308945400806e-05,
+      "loss": 0.022,
+      "step": 19380
+    },
+    {
+      "action_loss": 0.00328111182898283,
+      "epoch": 17.428057553956833,
+      "step": 19380
+    },
+    {
+      "epoch": 17.428057553956833,
+      "step": 19380,
+      "torque_loss": 0.08215201646089554
+    },
+    {
+      "epoch": 17.428057553956833,
+      "force_loss": 0.002679927973076701,
+      "step": 19380
+    },
+    {
+      "epoch": 17.43705035971223,
+      "grad_norm": 0.6395204663276672,
+      "learning_rate": 8.095145111510288e-05,
+      "loss": 0.0227,
+      "step": 19390
+    },
+    {
+      "action_loss": 0.005782816559076309,
+      "epoch": 17.43705035971223,
+      "step": 19390
+    },
+    {
+      "epoch": 17.43705035971223,
+      "step": 19390,
+      "torque_loss": 0.10056646913290024
+    },
+    {
+      "epoch": 17.43705035971223,
+      "force_loss": 0.010992948897182941,
+      "step": 19390
+    },
+    {
+      "epoch": 17.446043165467625,
+      "grad_norm": 0.3419896364212036,
+      "learning_rate": 8.092980337396406e-05,
+      "loss": 0.0201,
+      "step": 19400
+    },
+    {
+      "action_loss": 0.010971141047775745,
+      "epoch": 17.446043165467625,
+      "step": 19400
+    },
+    {
+      "epoch": 17.446043165467625,
+      "step": 19400,
+      "torque_loss": 0.15733571350574493
+    },
+    {
+      "epoch": 17.446043165467625,
+      "force_loss": 0.017646750435233116,
+      "step": 19400
+    },
+    {
+      "epoch": 17.455035971223023,
+      "grad_norm": 0.40668272972106934,
+      "learning_rate": 8.090814623716763e-05,
+      "loss": 0.0246,
+      "step": 19410
+    },
+    {
+      "action_loss": 0.00327402469702065,
+      "epoch": 17.455035971223023,
+      "step": 19410
+    },
+    {
+      "epoch": 17.455035971223023,
+      "step": 19410,
+      "torque_loss": 0.11614451557397842
+    },
+    {
+      "epoch": 17.455035971223023,
+      "force_loss": 0.005468995776027441,
+      "step": 19410
+    },
+    {
+      "epoch": 17.464028776978417,
+      "grad_norm": 0.23284949362277985,
+      "learning_rate": 8.088647971129246e-05,
+      "loss": 0.0262,
+      "step": 19420
+    },
+    {
+      "action_loss": 0.025572746992111206,
+      "epoch": 17.464028776978417,
+      "step": 19420
+    },
+    {
+      "epoch": 17.464028776978417,
+      "step": 19420,
+      "torque_loss": 0.1209123507142067
+    },
+    {
+      "epoch": 17.464028776978417,
+      "force_loss": 0.026003820821642876,
+      "step": 19420
+    },
+    {
+      "epoch": 17.473021582733814,
+      "grad_norm": 0.20005621016025543,
+      "learning_rate": 8.086480380292026e-05,
+      "loss": 0.0265,
+      "step": 19430
+    },
+    {
+      "action_loss": 0.0027618110179901123,
+      "epoch": 17.473021582733814,
+      "step": 19430
+    },
+    {
+      "epoch": 17.473021582733814,
+      "step": 19430,
+      "torque_loss": 0.0853128433227539
+    },
+    {
+      "epoch": 17.473021582733814,
+      "force_loss": 0.002930806949734688,
+      "step": 19430
+    },
+    {
+      "epoch": 17.48201438848921,
+      "grad_norm": 0.4512649178504944,
+      "learning_rate": 8.084311851863562e-05,
+      "loss": 0.0228,
+      "step": 19440
+    },
+    {
+      "action_loss": 0.0056221564300358295,
+      "epoch": 17.48201438848921,
+      "step": 19440
+    },
+    {
+      "epoch": 17.48201438848921,
+      "step": 19440,
+      "torque_loss": 0.10249606519937515
+    },
+    {
+      "epoch": 17.48201438848921,
+      "force_loss": 0.003846965730190277,
+      "step": 19440
+    },
+    {
+      "epoch": 17.491007194244606,
+      "grad_norm": 0.3204338550567627,
+      "learning_rate": 8.082142386502591e-05,
+      "loss": 0.022,
+      "step": 19450
+    },
+    {
+      "action_loss": 0.012489952147006989,
+      "epoch": 17.491007194244606,
+      "step": 19450
+    },
+    {
+      "epoch": 17.491007194244606,
+      "step": 19450,
+      "torque_loss": 0.11867877840995789
+    },
+    {
+      "epoch": 17.491007194244606,
+      "force_loss": 0.014584623277187347,
+      "step": 19450
+    },
+    {
+      "epoch": 17.5,
+      "grad_norm": 0.8133110404014587,
+      "learning_rate": 8.079971984868145e-05,
+      "loss": 0.0252,
+      "step": 19460
+    },
+    {
+      "action_loss": 0.005288057029247284,
+      "epoch": 17.5,
+      "step": 19460
+    },
+    {
+      "epoch": 17.5,
+      "step": 19460,
+      "torque_loss": 0.1301623433828354
+    },
+    {
+      "epoch": 17.5,
+      "force_loss": 0.006689710076898336,
+      "step": 19460
+    },
+    {
+      "epoch": 17.508992805755394,
+      "grad_norm": 0.34816107153892517,
+      "learning_rate": 8.077800647619532e-05,
+      "loss": 0.0241,
+      "step": 19470
+    },
+    {
+      "action_loss": 0.004750456660985947,
+      "epoch": 17.508992805755394,
+      "step": 19470
+    },
+    {
+      "epoch": 17.508992805755394,
+      "step": 19470,
+      "torque_loss": 0.1748230904340744
+    },
+    {
+      "epoch": 17.508992805755394,
+      "force_loss": 0.0048001110553741455,
+      "step": 19470
+    },
+    {
+      "epoch": 17.51798561151079,
+      "grad_norm": 0.2986980676651001,
+      "learning_rate": 8.075628375416345e-05,
+      "loss": 0.022,
+      "step": 19480
+    },
+    {
+      "action_loss": 0.004213675856590271,
+      "epoch": 17.51798561151079,
+      "step": 19480
+    },
+    {
+      "epoch": 17.51798561151079,
+      "step": 19480,
+      "torque_loss": 0.11202915757894516
+    },
+    {
+      "epoch": 17.51798561151079,
+      "force_loss": 0.005337308626621962,
+      "step": 19480
+    },
+    {
+      "epoch": 17.526978417266186,
+      "grad_norm": 0.4014330804347992,
+      "learning_rate": 8.073455168918464e-05,
+      "loss": 0.0245,
+      "step": 19490
+    },
+    {
+      "action_loss": 0.006691730115562677,
+      "epoch": 17.526978417266186,
+      "step": 19490
+    },
+    {
+      "epoch": 17.526978417266186,
+      "step": 19490,
+      "torque_loss": 0.16882272064685822
+    },
+    {
+      "epoch": 17.526978417266186,
+      "force_loss": 0.0055504050105810165,
+      "step": 19490
+    },
+    {
+      "epoch": 17.535971223021583,
+      "grad_norm": 0.2635452449321747,
+      "learning_rate": 8.071281028786055e-05,
+      "loss": 0.0208,
+      "step": 19500
+    },
+    {
+      "action_loss": 0.005574541632086039,
+      "epoch": 17.535971223021583,
+      "step": 19500
+    },
+    {
+      "epoch": 17.535971223021583,
+      "step": 19500,
+      "torque_loss": 0.15276120603084564
+    },
+    {
+      "epoch": 17.535971223021583,
+      "force_loss": 0.0036612292751669884,
+      "step": 19500
+    },
+    {
+      "epoch": 17.544964028776977,
+      "grad_norm": 0.2833903729915619,
+      "learning_rate": 8.069105955679562e-05,
+      "loss": 0.0222,
+      "step": 19510
+    },
+    {
+      "action_loss": 0.009073040448129177,
+      "epoch": 17.544964028776977,
+      "step": 19510
+    },
+    {
+      "epoch": 17.544964028776977,
+      "step": 19510,
+      "torque_loss": 0.16100911796092987
+    },
+    {
+      "epoch": 17.544964028776977,
+      "force_loss": 0.005548717454075813,
+      "step": 19510
+    },
+    {
+      "epoch": 17.553956834532375,
+      "grad_norm": 0.31475210189819336,
+      "learning_rate": 8.066929950259713e-05,
+      "loss": 0.0231,
+      "step": 19520
+    },
+    {
+      "action_loss": 0.007994409650564194,
+      "epoch": 17.553956834532375,
+      "step": 19520
+    },
+    {
+      "epoch": 17.553956834532375,
+      "step": 19520,
+      "torque_loss": 0.14138777554035187
+    },
+    {
+      "epoch": 17.553956834532375,
+      "force_loss": 0.010852988809347153,
+      "step": 19520
+    },
+    {
+      "epoch": 17.56294964028777,
+      "grad_norm": 0.3428254723548889,
+      "learning_rate": 8.064753013187522e-05,
+      "loss": 0.0227,
+      "step": 19530
+    },
+    {
+      "action_loss": 0.020793737843632698,
+      "epoch": 17.56294964028777,
+      "step": 19530
+    },
+    {
+      "epoch": 17.56294964028777,
+      "step": 19530,
+      "torque_loss": 0.15864413976669312
+    },
+    {
+      "epoch": 17.56294964028777,
+      "force_loss": 0.018360594287514687,
+      "step": 19530
+    },
+    {
+      "epoch": 17.571942446043167,
+      "grad_norm": 0.3125179409980774,
+      "learning_rate": 8.062575145124289e-05,
+      "loss": 0.0232,
+      "step": 19540
+    },
+    {
+      "action_loss": 0.008672011084854603,
+      "epoch": 17.571942446043167,
+      "step": 19540
+    },
+    {
+      "epoch": 17.571942446043167,
+      "step": 19540,
+      "torque_loss": 0.1733013391494751
+    },
+    {
+      "epoch": 17.571942446043167,
+      "force_loss": 0.005243752151727676,
+      "step": 19540
+    },
+    {
+      "epoch": 17.58093525179856,
+      "grad_norm": 0.39535972476005554,
+      "learning_rate": 8.060396346731587e-05,
+      "loss": 0.0231,
+      "step": 19550
+    },
+    {
+      "action_loss": 0.0033565405756235123,
+      "epoch": 17.58093525179856,
+      "step": 19550
+    },
+    {
+      "epoch": 17.58093525179856,
+      "step": 19550,
+      "torque_loss": 0.11748424917459488
+    },
+    {
+      "epoch": 17.58093525179856,
+      "force_loss": 0.0031257914379239082,
+      "step": 19550
+    },
+    {
+      "epoch": 17.58992805755396,
+      "grad_norm": 0.5443530678749084,
+      "learning_rate": 8.058216618671281e-05,
+      "loss": 0.0196,
+      "step": 19560
+    },
+    {
+      "action_loss": 0.012148463167250156,
+      "epoch": 17.58992805755396,
+      "step": 19560
+    },
+    {
+      "epoch": 17.58992805755396,
+      "step": 19560,
+      "torque_loss": 0.10654151439666748
+    },
+    {
+      "epoch": 17.58992805755396,
+      "force_loss": 0.008956816047430038,
+      "step": 19560
+    },
+    {
+      "epoch": 17.598920863309353,
+      "grad_norm": 0.30746304988861084,
+      "learning_rate": 8.056035961605514e-05,
+      "loss": 0.0192,
+      "step": 19570
+    },
+    {
+      "action_loss": 0.0071060024201869965,
+      "epoch": 17.598920863309353,
+      "step": 19570
+    },
+    {
+      "epoch": 17.598920863309353,
+      "step": 19570,
+      "torque_loss": 0.12469115108251572
+    },
+    {
+      "epoch": 17.598920863309353,
+      "force_loss": 0.004781091120094061,
+      "step": 19570
+    },
+    {
+      "epoch": 17.607913669064747,
+      "grad_norm": 0.4439001679420471,
+      "learning_rate": 8.05385437619671e-05,
+      "loss": 0.028,
+      "step": 19580
+    },
+    {
+      "action_loss": 0.003755138022825122,
+      "epoch": 17.607913669064747,
+      "step": 19580
+    },
+    {
+      "epoch": 17.607913669064747,
+      "step": 19580,
+      "torque_loss": 0.14100150763988495
+    },
+    {
+      "epoch": 17.607913669064747,
+      "force_loss": 0.006742970552295446,
+      "step": 19580
+    },
+    {
+      "epoch": 17.616906474820144,
+      "grad_norm": 0.30653971433639526,
+      "learning_rate": 8.05167186310758e-05,
+      "loss": 0.0201,
+      "step": 19590
+    },
+    {
+      "action_loss": 0.008167531341314316,
+      "epoch": 17.616906474820144,
+      "step": 19590
+    },
+    {
+      "epoch": 17.616906474820144,
+      "step": 19590,
+      "torque_loss": 0.1529424637556076
+    },
+    {
+      "epoch": 17.616906474820144,
+      "force_loss": 0.005340991076081991,
+      "step": 19590
+    },
+    {
+      "epoch": 17.62589928057554,
+      "grad_norm": 0.3988235294818878,
+      "learning_rate": 8.049488423001113e-05,
+      "loss": 0.0228,
+      "step": 19600
+    },
+    {
+      "action_loss": 0.01556698139756918,
+      "epoch": 17.62589928057554,
+      "step": 19600
+    },
+    {
+      "epoch": 17.62589928057554,
+      "step": 19600,
+      "torque_loss": 0.1938236951828003
+    },
+    {
+      "epoch": 17.62589928057554,
+      "force_loss": 0.008753178641200066,
+      "step": 19600
+    },
+    {
+      "epoch": 17.634892086330936,
+      "grad_norm": 0.4130746126174927,
+      "learning_rate": 8.047304056540581e-05,
+      "loss": 0.0251,
+      "step": 19610
+    },
+    {
+      "action_loss": 0.005363933742046356,
+      "epoch": 17.634892086330936,
+      "step": 19610
+    },
+    {
+      "epoch": 17.634892086330936,
+      "step": 19610,
+      "torque_loss": 0.137730672955513
+    },
+    {
+      "epoch": 17.634892086330936,
+      "force_loss": 0.005873801652342081,
+      "step": 19610
+    },
+    {
+      "epoch": 17.64388489208633,
+      "grad_norm": 0.40095967054367065,
+      "learning_rate": 8.045118764389534e-05,
+      "loss": 0.0261,
+      "step": 19620
+    },
+    {
+      "action_loss": 0.002637271536514163,
+      "epoch": 17.64388489208633,
+      "step": 19620
+    },
+    {
+      "epoch": 17.64388489208633,
+      "step": 19620,
+      "torque_loss": 0.12004473805427551
+    },
+    {
+      "epoch": 17.64388489208633,
+      "force_loss": 0.0027399174869060516,
+      "step": 19620
+    },
+    {
+      "epoch": 17.652877697841728,
+      "grad_norm": 0.23908305168151855,
+      "learning_rate": 8.042932547211809e-05,
+      "loss": 0.0182,
+      "step": 19630
+    },
+    {
+      "action_loss": 0.006685630884021521,
+      "epoch": 17.652877697841728,
+      "step": 19630
+    },
+    {
+      "epoch": 17.652877697841728,
+      "step": 19630,
+      "torque_loss": 0.14698709547519684
+    },
+    {
+      "epoch": 17.652877697841728,
+      "force_loss": 0.004281893372535706,
+      "step": 19630
+    },
+    {
+      "epoch": 17.66187050359712,
+      "grad_norm": 0.5244224071502686,
+      "learning_rate": 8.04074540567152e-05,
+      "loss": 0.0229,
+      "step": 19640
+    },
+    {
+      "action_loss": 0.0032203607261180878,
+      "epoch": 17.66187050359712,
+      "step": 19640
+    },
+    {
+      "epoch": 17.66187050359712,
+      "step": 19640,
+      "torque_loss": 0.14612585306167603
+    },
+    {
+      "epoch": 17.66187050359712,
+      "force_loss": 0.0025413024704903364,
+      "step": 19640
+    },
+    {
+      "epoch": 17.67086330935252,
+      "grad_norm": 0.43039417266845703,
+      "learning_rate": 8.038557340433063e-05,
+      "loss": 0.0301,
+      "step": 19650
+    },
+    {
+      "action_loss": 0.009502124972641468,
+      "epoch": 17.67086330935252,
+      "step": 19650
+    },
+    {
+      "epoch": 17.67086330935252,
+      "step": 19650,
+      "torque_loss": 0.1313255876302719
+    },
+    {
+      "epoch": 17.67086330935252,
+      "force_loss": 0.008214607834815979,
+      "step": 19650
+    },
+    {
+      "epoch": 17.679856115107913,
+      "grad_norm": 0.21753033995628357,
+      "learning_rate": 8.036368352161115e-05,
+      "loss": 0.0208,
+      "step": 19660
+    },
+    {
+      "action_loss": 0.01917554996907711,
+      "epoch": 17.679856115107913,
+      "step": 19660
+    },
+    {
+      "epoch": 17.679856115107913,
+      "step": 19660,
+      "torque_loss": 0.13388876616954803
+    },
+    {
+      "epoch": 17.679856115107913,
+      "force_loss": 0.023362377658486366,
+      "step": 19660
+    },
+    {
+      "epoch": 17.68884892086331,
+      "grad_norm": 0.170255646109581,
+      "learning_rate": 8.034178441520633e-05,
+      "loss": 0.0204,
+      "step": 19670
+    },
+    {
+      "action_loss": 0.005621526390314102,
+      "epoch": 17.68884892086331,
+      "step": 19670
+    },
+    {
+      "epoch": 17.68884892086331,
+      "step": 19670,
+      "torque_loss": 0.11105605214834213
+    },
+    {
+      "epoch": 17.68884892086331,
+      "force_loss": 0.00489856768399477,
+      "step": 19670
+    },
+    {
+      "epoch": 17.697841726618705,
+      "grad_norm": 0.4363051652908325,
+      "learning_rate": 8.031987609176852e-05,
+      "loss": 0.0262,
+      "step": 19680
+    },
+    {
+      "action_loss": 0.0037665802519768476,
+      "epoch": 17.697841726618705,
+      "step": 19680
+    },
+    {
+      "epoch": 17.697841726618705,
+      "step": 19680,
+      "torque_loss": 0.10609819740056992
+    },
+    {
+      "epoch": 17.697841726618705,
+      "force_loss": 0.003781653940677643,
+      "step": 19680
+    },
+    {
+      "epoch": 17.7068345323741,
+      "grad_norm": 0.21985115110874176,
+      "learning_rate": 8.02979585579529e-05,
+      "loss": 0.0231,
+      "step": 19690
+    },
+    {
+      "action_loss": 0.007497600745409727,
+      "epoch": 17.7068345323741,
+      "step": 19690
+    },
+    {
+      "epoch": 17.7068345323741,
+      "step": 19690,
+      "torque_loss": 0.10320410877466202
+    },
+    {
+      "epoch": 17.7068345323741,
+      "force_loss": 0.010489988140761852,
+      "step": 19690
+    },
+    {
+      "epoch": 17.715827338129497,
+      "grad_norm": 0.6059436798095703,
+      "learning_rate": 8.027603182041745e-05,
+      "loss": 0.0291,
+      "step": 19700
+    },
+    {
+      "action_loss": 0.007499633356928825,
+      "epoch": 17.715827338129497,
+      "step": 19700
+    },
+    {
+      "epoch": 17.715827338129497,
+      "step": 19700,
+      "torque_loss": 0.12829194962978363
+    },
+    {
+      "epoch": 17.715827338129497,
+      "force_loss": 0.005964389070868492,
+      "step": 19700
+    },
+    {
+      "epoch": 17.72482014388489,
+      "grad_norm": 0.3505343794822693,
+      "learning_rate": 8.025409588582292e-05,
+      "loss": 0.0259,
+      "step": 19710
+    },
+    {
+      "action_loss": 0.006138558033853769,
+      "epoch": 17.72482014388489,
+      "step": 19710
+    },
+    {
+      "epoch": 17.72482014388489,
+      "step": 19710,
+      "torque_loss": 0.14246180653572083
+    },
+    {
+      "epoch": 17.72482014388489,
+      "force_loss": 0.0058523621410131454,
+      "step": 19710
+    },
+    {
+      "epoch": 17.73381294964029,
+      "grad_norm": 0.7307343482971191,
+      "learning_rate": 8.023215076083288e-05,
+      "loss": 0.0234,
+      "step": 19720
+    },
+    {
+      "action_loss": 0.0041977656073868275,
+      "epoch": 17.73381294964029,
+      "step": 19720
+    },
+    {
+      "epoch": 17.73381294964029,
+      "step": 19720,
+      "torque_loss": 0.12356024980545044
+    },
+    {
+      "epoch": 17.73381294964029,
+      "force_loss": 0.010036000050604343,
+      "step": 19720
+    },
+    {
+      "epoch": 17.742805755395683,
+      "grad_norm": 0.36447811126708984,
+      "learning_rate": 8.021019645211367e-05,
+      "loss": 0.0201,
+      "step": 19730
+    },
+    {
+      "action_loss": 0.002569265430793166,
+      "epoch": 17.742805755395683,
+      "step": 19730
+    },
+    {
+      "epoch": 17.742805755395683,
+      "step": 19730,
+      "torque_loss": 0.1548398733139038
+    },
+    {
+      "epoch": 17.742805755395683,
+      "force_loss": 0.003124109236523509,
+      "step": 19730
+    },
+    {
+      "epoch": 17.75179856115108,
+      "grad_norm": 0.30815255641937256,
+      "learning_rate": 8.018823296633441e-05,
+      "loss": 0.0207,
+      "step": 19740
+    },
+    {
+      "action_loss": 0.010665304027497768,
+      "epoch": 17.75179856115108,
+      "step": 19740
+    },
+    {
+      "epoch": 17.75179856115108,
+      "step": 19740,
+      "torque_loss": 0.1426772028207779
+    },
+    {
+      "epoch": 17.75179856115108,
+      "force_loss": 0.013071350753307343,
+      "step": 19740
+    },
+    {
+      "epoch": 17.760791366906474,
+      "grad_norm": 0.31556180119514465,
+      "learning_rate": 8.016626031016708e-05,
+      "loss": 0.0264,
+      "step": 19750
+    },
+    {
+      "action_loss": 0.008767900057137012,
+      "epoch": 17.760791366906474,
+      "step": 19750
+    },
+    {
+      "epoch": 17.760791366906474,
+      "step": 19750,
+      "torque_loss": 0.13452835381031036
+    },
+    {
+      "epoch": 17.760791366906474,
+      "force_loss": 0.011867806315422058,
+      "step": 19750
+    },
+    {
+      "epoch": 17.769784172661872,
+      "grad_norm": 0.4469786286354065,
+      "learning_rate": 8.014427849028636e-05,
+      "loss": 0.0215,
+      "step": 19760
+    },
+    {
+      "action_loss": 0.01100847590714693,
+      "epoch": 17.769784172661872,
+      "step": 19760
+    },
+    {
+      "epoch": 17.769784172661872,
+      "step": 19760,
+      "torque_loss": 0.14512237906455994
+    },
+    {
+      "epoch": 17.769784172661872,
+      "force_loss": 0.011651960201561451,
+      "step": 19760
+    },
+    {
+      "epoch": 17.778776978417266,
+      "grad_norm": 0.588995099067688,
+      "learning_rate": 8.012228751336974e-05,
+      "loss": 0.0243,
+      "step": 19770
+    },
+    {
+      "action_loss": 0.008913128636777401,
+      "epoch": 17.778776978417266,
+      "step": 19770
+    },
+    {
+      "epoch": 17.778776978417266,
+      "step": 19770,
+      "torque_loss": 0.19584007561206818
+    },
+    {
+      "epoch": 17.778776978417266,
+      "force_loss": 0.008015934377908707,
+      "step": 19770
+    },
+    {
+      "epoch": 17.78776978417266,
+      "grad_norm": 0.4769243597984314,
+      "learning_rate": 8.01002873860975e-05,
+      "loss": 0.0317,
+      "step": 19780
+    },
+    {
+      "action_loss": 0.0070003787986934185,
+      "epoch": 17.78776978417266,
+      "step": 19780
+    },
+    {
+      "epoch": 17.78776978417266,
+      "step": 19780,
+      "torque_loss": 0.1274857521057129
+    },
+    {
+      "epoch": 17.78776978417266,
+      "force_loss": 0.006243351846933365,
+      "step": 19780
+    },
+    {
+      "epoch": 17.796762589928058,
+      "grad_norm": 0.35653167963027954,
+      "learning_rate": 8.00782781151527e-05,
+      "loss": 0.0212,
+      "step": 19790
+    },
+    {
+      "action_loss": 0.02327529340982437,
+      "epoch": 17.796762589928058,
+      "step": 19790
+    },
+    {
+      "epoch": 17.796762589928058,
+      "step": 19790,
+      "torque_loss": 0.16975641250610352
+    },
+    {
+      "epoch": 17.796762589928058,
+      "force_loss": 0.01342775672674179,
+      "step": 19790
+    },
+    {
+      "epoch": 17.805755395683452,
+      "grad_norm": 0.335950642824173,
+      "learning_rate": 8.005625970722119e-05,
+      "loss": 0.0263,
+      "step": 19800
+    },
+    {
+      "action_loss": 0.004299199674278498,
+      "epoch": 17.805755395683452,
+      "step": 19800
+    },
+    {
+      "epoch": 17.805755395683452,
+      "step": 19800,
+      "torque_loss": 0.08940321952104568
+    },
+    {
+      "epoch": 17.805755395683452,
+      "force_loss": 0.00388413411565125,
+      "step": 19800
+    },
+    {
+      "epoch": 17.81474820143885,
+      "grad_norm": 0.441150039434433,
+      "learning_rate": 8.003423216899158e-05,
+      "loss": 0.0201,
+      "step": 19810
+    },
+    {
+      "action_loss": 0.0038878542836755514,
+      "epoch": 17.81474820143885,
+      "step": 19810
+    },
+    {
+      "epoch": 17.81474820143885,
+      "step": 19810,
+      "torque_loss": 0.19903464615345
+    },
+    {
+      "epoch": 17.81474820143885,
+      "force_loss": 0.002141745062544942,
+      "step": 19810
+    },
+    {
+      "epoch": 17.823741007194243,
+      "grad_norm": 0.24629060924053192,
+      "learning_rate": 8.001219550715522e-05,
+      "loss": 0.0285,
+      "step": 19820
+    },
+    {
+      "action_loss": 0.008333519101142883,
+      "epoch": 17.823741007194243,
+      "step": 19820
+    },
+    {
+      "epoch": 17.823741007194243,
+      "step": 19820,
+      "torque_loss": 0.10464733093976974
+    },
+    {
+      "epoch": 17.823741007194243,
+      "force_loss": 0.015250998549163342,
+      "step": 19820
+    },
+    {
+      "epoch": 17.83273381294964,
+      "grad_norm": 0.28278669714927673,
+      "learning_rate": 7.999014972840632e-05,
+      "loss": 0.0266,
+      "step": 19830
+    },
+    {
+      "action_loss": 0.0053902133367955685,
+      "epoch": 17.83273381294964,
+      "step": 19830
+    },
+    {
+      "epoch": 17.83273381294964,
+      "step": 19830,
+      "torque_loss": 0.12154543399810791
+    },
+    {
+      "epoch": 17.83273381294964,
+      "force_loss": 0.005527317523956299,
+      "step": 19830
+    },
+    {
+      "epoch": 17.841726618705035,
+      "grad_norm": 0.25460949540138245,
+      "learning_rate": 7.996809483944174e-05,
+      "loss": 0.02,
+      "step": 19840
+    },
+    {
+      "action_loss": 0.020251328125596046,
+      "epoch": 17.841726618705035,
+      "step": 19840
+    },
+    {
+      "epoch": 17.841726618705035,
+      "step": 19840,
+      "torque_loss": 0.16169770061969757
+    },
+    {
+      "epoch": 17.841726618705035,
+      "force_loss": 0.013819930143654346,
+      "step": 19840
+    },
+    {
+      "epoch": 17.850719424460433,
+      "grad_norm": 0.39666181802749634,
+      "learning_rate": 7.994603084696124e-05,
+      "loss": 0.0197,
+      "step": 19850
+    },
+    {
+      "action_loss": 0.014225137419998646,
+      "epoch": 17.850719424460433,
+      "step": 19850
+    },
+    {
+      "epoch": 17.850719424460433,
+      "step": 19850,
+      "torque_loss": 0.1906416416168213
+    },
+    {
+      "epoch": 17.850719424460433,
+      "force_loss": 0.011307370848953724,
+      "step": 19850
+    },
+    {
+      "epoch": 17.859712230215827,
+      "grad_norm": 0.19151118397712708,
+      "learning_rate": 7.992395775766724e-05,
+      "loss": 0.0192,
+      "step": 19860
+    },
+    {
+      "action_loss": 0.005128639284521341,
+      "epoch": 17.859712230215827,
+      "step": 19860
+    },
+    {
+      "epoch": 17.859712230215827,
+      "step": 19860,
+      "torque_loss": 0.11203759163618088
+    },
+    {
+      "epoch": 17.859712230215827,
+      "force_loss": 0.004943978972733021,
+      "step": 19860
+    },
+    {
+      "epoch": 17.868705035971225,
+      "grad_norm": 0.17765939235687256,
+      "learning_rate": 7.990187557826497e-05,
+      "loss": 0.0267,
+      "step": 19870
+    },
+    {
+      "action_loss": 0.011818304657936096,
+      "epoch": 17.868705035971225,
+      "step": 19870
+    },
+    {
+      "epoch": 17.868705035971225,
+      "step": 19870,
+      "torque_loss": 0.1765238642692566
+    },
+    {
+      "epoch": 17.868705035971225,
+      "force_loss": 0.014577222056686878,
+      "step": 19870
+    },
+    {
+      "epoch": 17.87769784172662,
+      "grad_norm": 0.2867027521133423,
+      "learning_rate": 7.987978431546242e-05,
+      "loss": 0.0259,
+      "step": 19880
+    },
+    {
+      "action_loss": 0.008290189318358898,
+      "epoch": 17.87769784172662,
+      "step": 19880
+    },
+    {
+      "epoch": 17.87769784172662,
+      "step": 19880,
+      "torque_loss": 0.1209149956703186
+    },
+    {
+      "epoch": 17.87769784172662,
+      "force_loss": 0.007892039604485035,
+      "step": 19880
+    },
+    {
+      "epoch": 17.886690647482013,
+      "grad_norm": 0.6111899018287659,
+      "learning_rate": 7.985768397597031e-05,
+      "loss": 0.0238,
+      "step": 19890
+    },
+    {
+      "action_loss": 0.008832314051687717,
+      "epoch": 17.886690647482013,
+      "step": 19890
+    },
+    {
+      "epoch": 17.886690647482013,
+      "step": 19890,
+      "torque_loss": 0.13851268589496613
+    },
+    {
+      "epoch": 17.886690647482013,
+      "force_loss": 0.008938915096223354,
+      "step": 19890
+    },
+    {
+      "epoch": 17.89568345323741,
+      "grad_norm": 0.6564391851425171,
+      "learning_rate": 7.983557456650216e-05,
+      "loss": 0.0263,
+      "step": 19900
+    },
+    {
+      "action_loss": 0.022660313174128532,
+      "epoch": 17.89568345323741,
+      "step": 19900
+    },
+    {
+      "epoch": 17.89568345323741,
+      "step": 19900,
+      "torque_loss": 0.16038145124912262
+    },
+    {
+      "epoch": 17.89568345323741,
+      "force_loss": 0.029534464702010155,
+      "step": 19900
+    },
+    {
+      "epoch": 17.904676258992804,
+      "grad_norm": 0.4128296375274658,
+      "learning_rate": 7.981345609377422e-05,
+      "loss": 0.0267,
+      "step": 19910
+    },
+    {
+      "action_loss": 0.018345626071095467,
+      "epoch": 17.904676258992804,
+      "step": 19910
+    },
+    {
+      "epoch": 17.904676258992804,
+      "step": 19910,
+      "torque_loss": 0.1732242852449417
+    },
+    {
+      "epoch": 17.904676258992804,
+      "force_loss": 0.013446040451526642,
+      "step": 19910
+    },
+    {
+      "epoch": 17.913669064748202,
+      "grad_norm": 0.3570900559425354,
+      "learning_rate": 7.97913285645055e-05,
+      "loss": 0.0225,
+      "step": 19920
+    },
+    {
+      "action_loss": 0.011159618385136127,
+      "epoch": 17.913669064748202,
+      "step": 19920
+    },
+    {
+      "epoch": 17.913669064748202,
+      "step": 19920,
+      "torque_loss": 0.1497364044189453
+    },
+    {
+      "epoch": 17.913669064748202,
+      "force_loss": 0.011165213771164417,
+      "step": 19920
+    },
+    {
+      "epoch": 17.922661870503596,
+      "grad_norm": 0.3876411020755768,
+      "learning_rate": 7.976919198541776e-05,
+      "loss": 0.0212,
+      "step": 19930
+    },
+    {
+      "action_loss": 0.007042673882097006,
+      "epoch": 17.922661870503596,
+      "step": 19930
+    },
+    {
+      "epoch": 17.922661870503596,
+      "step": 19930,
+      "torque_loss": 0.1251247674226761
+    },
+    {
+      "epoch": 17.922661870503596,
+      "force_loss": 0.00474808132275939,
+      "step": 19930
+    },
+    {
+      "epoch": 17.931654676258994,
+      "grad_norm": 0.45121076703071594,
+      "learning_rate": 7.974704636323548e-05,
+      "loss": 0.024,
+      "step": 19940
+    },
+    {
+      "action_loss": 0.007809826638549566,
+      "epoch": 17.931654676258994,
+      "step": 19940
+    },
+    {
+      "epoch": 17.931654676258994,
+      "step": 19940,
+      "torque_loss": 0.10121830552816391
+    },
+    {
+      "epoch": 17.931654676258994,
+      "force_loss": 0.0060829236172139645,
+      "step": 19940
+    },
+    {
+      "epoch": 17.940647482014388,
+      "grad_norm": 0.3448992371559143,
+      "learning_rate": 7.972489170468597e-05,
+      "loss": 0.0198,
+      "step": 19950
+    },
+    {
+      "action_loss": 0.004157028626650572,
+      "epoch": 17.940647482014388,
+      "step": 19950
+    },
+    {
+      "epoch": 17.940647482014388,
+      "step": 19950,
+      "torque_loss": 0.12187572568655014
+    },
+    {
+      "epoch": 17.940647482014388,
+      "force_loss": 0.004202096723020077,
+      "step": 19950
+    },
+    {
+      "epoch": 17.949640287769785,
+      "grad_norm": 0.41944149136543274,
+      "learning_rate": 7.970272801649918e-05,
+      "loss": 0.0247,
+      "step": 19960
+    },
+    {
+      "action_loss": 0.006942410487681627,
+      "epoch": 17.949640287769785,
+      "step": 19960
+    },
+    {
+      "epoch": 17.949640287769785,
+      "step": 19960,
+      "torque_loss": 0.14761404693126678
+    },
+    {
+      "epoch": 17.949640287769785,
+      "force_loss": 0.006611744407564402,
+      "step": 19960
+    },
+    {
+      "epoch": 17.95863309352518,
+      "grad_norm": 0.2848796546459198,
+      "learning_rate": 7.96805553054079e-05,
+      "loss": 0.0252,
+      "step": 19970
+    },
+    {
+      "action_loss": 0.007320711854845285,
+      "epoch": 17.95863309352518,
+      "step": 19970
+    },
+    {
+      "epoch": 17.95863309352518,
+      "step": 19970,
+      "torque_loss": 0.1549905389547348
+    },
+    {
+      "epoch": 17.95863309352518,
+      "force_loss": 0.0038721465971320868,
+      "step": 19970
+    },
+    {
+      "epoch": 17.967625899280577,
+      "grad_norm": 0.5131918787956238,
+      "learning_rate": 7.965837357814756e-05,
+      "loss": 0.0233,
+      "step": 19980
+    },
+    {
+      "action_loss": 0.007778391242027283,
+      "epoch": 17.967625899280577,
+      "step": 19980
+    },
+    {
+      "epoch": 17.967625899280577,
+      "step": 19980,
+      "torque_loss": 0.12595798075199127
+    },
+    {
+      "epoch": 17.967625899280577,
+      "force_loss": 0.004015080630779266,
+      "step": 19980
+    },
+    {
+      "epoch": 17.97661870503597,
+      "grad_norm": 0.23807628452777863,
+      "learning_rate": 7.963618284145643e-05,
+      "loss": 0.0248,
+      "step": 19990
+    },
+    {
+      "action_loss": 0.012418706901371479,
+      "epoch": 17.97661870503597,
+      "step": 19990
+    },
+    {
+      "epoch": 17.97661870503597,
+      "step": 19990,
+      "torque_loss": 0.19964194297790527
+    },
+    {
+      "epoch": 17.97661870503597,
+      "force_loss": 0.00856059230864048,
+      "step": 19990
+    },
+    {
+      "epoch": 17.985611510791365,
+      "grad_norm": 0.3788957893848419,
+      "learning_rate": 7.961398310207544e-05,
+      "loss": 0.0256,
+      "step": 20000
+    },
+    {
+      "action_loss": 0.0044211591593921185,
+      "epoch": 17.985611510791365,
+      "step": 20000
+    },
+    {
+      "epoch": 17.985611510791365,
+      "step": 20000,
+      "torque_loss": 0.11479606479406357
+    },
+    {
+      "epoch": 17.985611510791365,
+      "force_loss": 0.0035031491424888372,
+      "step": 20000
+    },
+    {
+      "epoch": 17.994604316546763,
+      "grad_norm": 0.3506962060928345,
+      "learning_rate": 7.95917743667483e-05,
+      "loss": 0.0224,
+      "step": 20010
+    },
+    {
+      "action_loss": 0.004047689493745565,
+      "epoch": 17.994604316546763,
+      "step": 20010
+    },
+    {
+      "epoch": 17.994604316546763,
+      "step": 20010,
+      "torque_loss": 0.14169400930404663
+    },
+    {
+      "epoch": 17.994604316546763,
+      "force_loss": 0.0023170565254986286,
+      "step": 20010
+    },
+    {
+      "epoch": 18.003597122302157,
+      "grad_norm": 0.5437688827514648,
+      "learning_rate": 7.956955664222144e-05,
+      "loss": 0.0234,
+      "step": 20020
+    },
+    {
+      "action_loss": 0.010446433909237385,
+      "epoch": 18.003597122302157,
+      "step": 20020
+    },
+    {
+      "epoch": 18.003597122302157,
+      "step": 20020,
+      "torque_loss": 0.14005039632320404
+    },
+    {
+      "epoch": 18.003597122302157,
+      "force_loss": 0.014068144373595715,
+      "step": 20020
+    },
+    {
+      "epoch": 18.012589928057555,
+      "grad_norm": 0.19264504313468933,
+      "learning_rate": 7.954732993524399e-05,
+      "loss": 0.0237,
+      "step": 20030
+    },
+    {
+      "action_loss": 0.01298863347619772,
+      "epoch": 18.012589928057555,
+      "step": 20030
+    },
+    {
+      "epoch": 18.012589928057555,
+      "step": 20030,
+      "torque_loss": 0.11979428678750992
+    },
+    {
+      "epoch": 18.012589928057555,
+      "force_loss": 0.0061620366759598255,
+      "step": 20030
+    },
+    {
+      "epoch": 18.02158273381295,
+      "grad_norm": 0.26759570837020874,
+      "learning_rate": 7.952509425256786e-05,
+      "loss": 0.024,
+      "step": 20040
+    },
+    {
+      "action_loss": 0.0036432023625820875,
+      "epoch": 18.02158273381295,
+      "step": 20040
+    },
+    {
+      "epoch": 18.02158273381295,
+      "step": 20040,
+      "torque_loss": 0.11801549792289734
+    },
+    {
+      "epoch": 18.02158273381295,
+      "force_loss": 0.003185616573318839,
+      "step": 20040
+    },
+    {
+      "epoch": 18.030575539568346,
+      "grad_norm": 0.3666815757751465,
+      "learning_rate": 7.950284960094767e-05,
+      "loss": 0.0238,
+      "step": 20050
+    },
+    {
+      "action_loss": 0.0030627876985818148,
+      "epoch": 18.030575539568346,
+      "step": 20050
+    },
+    {
+      "epoch": 18.030575539568346,
+      "step": 20050,
+      "torque_loss": 0.12197041511535645
+    },
+    {
+      "epoch": 18.030575539568346,
+      "force_loss": 0.002873636782169342,
+      "step": 20050
+    },
+    {
+      "epoch": 18.03956834532374,
+      "grad_norm": 0.41907984018325806,
+      "learning_rate": 7.948059598714076e-05,
+      "loss": 0.021,
+      "step": 20060
+    },
+    {
+      "action_loss": 0.003818368539214134,
+      "epoch": 18.03956834532374,
+      "step": 20060
+    },
+    {
+      "epoch": 18.03956834532374,
+      "step": 20060,
+      "torque_loss": 0.08319190889596939
+    },
+    {
+      "epoch": 18.03956834532374,
+      "force_loss": 0.006896860897541046,
+      "step": 20060
+    },
+    {
+      "epoch": 18.048561151079138,
+      "grad_norm": 0.425952285528183,
+      "learning_rate": 7.945833341790717e-05,
+      "loss": 0.0259,
+      "step": 20070
+    },
+    {
+      "action_loss": 0.010059972293674946,
+      "epoch": 18.048561151079138,
+      "step": 20070
+    },
+    {
+      "epoch": 18.048561151079138,
+      "step": 20070,
+      "torque_loss": 0.20924358069896698
+    },
+    {
+      "epoch": 18.048561151079138,
+      "force_loss": 0.008788040839135647,
+      "step": 20070
+    },
+    {
+      "epoch": 18.057553956834532,
+      "grad_norm": 0.6132962703704834,
+      "learning_rate": 7.94360619000097e-05,
+      "loss": 0.0278,
+      "step": 20080
+    },
+    {
+      "action_loss": 0.003754944773390889,
+      "epoch": 18.057553956834532,
+      "step": 20080
+    },
+    {
+      "epoch": 18.057553956834532,
+      "step": 20080,
+      "torque_loss": 0.10349229723215103
+    },
+    {
+      "epoch": 18.057553956834532,
+      "force_loss": 0.005687051918357611,
+      "step": 20080
+    },
+    {
+      "epoch": 18.06654676258993,
+      "grad_norm": 0.3191094398498535,
+      "learning_rate": 7.941378144021381e-05,
+      "loss": 0.0218,
+      "step": 20090
+    },
+    {
+      "action_loss": 0.03181866556406021,
+      "epoch": 18.06654676258993,
+      "step": 20090
+    },
+    {
+      "epoch": 18.06654676258993,
+      "step": 20090,
+      "torque_loss": 0.15198683738708496
+    },
+    {
+      "epoch": 18.06654676258993,
+      "force_loss": 0.04333909973502159,
+      "step": 20090
+    },
+    {
+      "epoch": 18.075539568345324,
+      "grad_norm": 0.2699708044528961,
+      "learning_rate": 7.939149204528777e-05,
+      "loss": 0.022,
+      "step": 20100
+    },
+    {
+      "action_loss": 0.011830595321953297,
+      "epoch": 18.075539568345324,
+      "step": 20100
+    },
+    {
+      "epoch": 18.075539568345324,
+      "step": 20100,
+      "torque_loss": 0.18377113342285156
+    },
+    {
+      "epoch": 18.075539568345324,
+      "force_loss": 0.016688520088791847,
+      "step": 20100
+    },
+    {
+      "epoch": 18.084532374100718,
+      "grad_norm": 0.5913984179496765,
+      "learning_rate": 7.936919372200246e-05,
+      "loss": 0.0264,
+      "step": 20110
+    },
+    {
+      "action_loss": 0.006229867693036795,
+      "epoch": 18.084532374100718,
+      "step": 20110
+    },
+    {
+      "epoch": 18.084532374100718,
+      "step": 20110,
+      "torque_loss": 0.1246085986495018
+    },
+    {
+      "epoch": 18.084532374100718,
+      "force_loss": 0.008133932948112488,
+      "step": 20110
+    },
+    {
+      "epoch": 18.093525179856115,
+      "grad_norm": 0.26789072155952454,
+      "learning_rate": 7.934688647713158e-05,
+      "loss": 0.0239,
+      "step": 20120
+    },
+    {
+      "action_loss": 0.003590468317270279,
+      "epoch": 18.093525179856115,
+      "step": 20120
+    },
+    {
+      "epoch": 18.093525179856115,
+      "step": 20120,
+      "torque_loss": 0.14061181247234344
+    },
+    {
+      "epoch": 18.093525179856115,
+      "force_loss": 0.003195495344698429,
+      "step": 20120
+    },
+    {
+      "epoch": 18.10251798561151,
+      "grad_norm": 0.29276755452156067,
+      "learning_rate": 7.932457031745143e-05,
+      "loss": 0.0216,
+      "step": 20130
+    },
+    {
+      "action_loss": 0.06483045220375061,
+      "epoch": 18.10251798561151,
+      "step": 20130
+    },
+    {
+      "epoch": 18.10251798561151,
+      "step": 20130,
+      "torque_loss": 0.1954166144132614
+    },
+    {
+      "epoch": 18.10251798561151,
+      "force_loss": 0.06355533748865128,
+      "step": 20130
+    },
+    {
+      "epoch": 18.111510791366907,
+      "grad_norm": 0.43592166900634766,
+      "learning_rate": 7.930224524974108e-05,
+      "loss": 0.0245,
+      "step": 20140
+    },
+    {
+      "action_loss": 0.014609490521252155,
+      "epoch": 18.111510791366907,
+      "step": 20140
+    },
+    {
+      "epoch": 18.111510791366907,
+      "step": 20140,
+      "torque_loss": 0.17122183740139008
+    },
+    {
+      "epoch": 18.111510791366907,
+      "force_loss": 0.013039768673479557,
+      "step": 20140
+    },
+    {
+      "epoch": 18.1205035971223,
+      "grad_norm": 0.39575302600860596,
+      "learning_rate": 7.927991128078232e-05,
+      "loss": 0.0229,
+      "step": 20150
+    },
+    {
+      "action_loss": 0.006208075676113367,
+      "epoch": 18.1205035971223,
+      "step": 20150
+    },
+    {
+      "epoch": 18.1205035971223,
+      "step": 20150,
+      "torque_loss": 0.15075568854808807
+    },
+    {
+      "epoch": 18.1205035971223,
+      "force_loss": 0.006185729056596756,
+      "step": 20150
+    },
+    {
+      "epoch": 18.1294964028777,
+      "grad_norm": 0.16428524255752563,
+      "learning_rate": 7.925756841735958e-05,
+      "loss": 0.0217,
+      "step": 20160
+    },
+    {
+      "action_loss": 0.012076959013938904,
+      "epoch": 18.1294964028777,
+      "step": 20160
+    },
+    {
+      "epoch": 18.1294964028777,
+      "step": 20160,
+      "torque_loss": 0.13658148050308228
+    },
+    {
+      "epoch": 18.1294964028777,
+      "force_loss": 0.008835576474666595,
+      "step": 20160
+    },
+    {
+      "epoch": 18.138489208633093,
+      "grad_norm": 0.4757242202758789,
+      "learning_rate": 7.923521666626008e-05,
+      "loss": 0.02,
+      "step": 20170
+    },
+    {
+      "action_loss": 0.015908218920230865,
+      "epoch": 18.138489208633093,
+      "step": 20170
+    },
+    {
+      "epoch": 18.138489208633093,
+      "step": 20170,
+      "torque_loss": 0.13368353247642517
+    },
+    {
+      "epoch": 18.138489208633093,
+      "force_loss": 0.022351935505867004,
+      "step": 20170
+    },
+    {
+      "epoch": 18.14748201438849,
+      "grad_norm": 0.3321283459663391,
+      "learning_rate": 7.921285603427366e-05,
+      "loss": 0.0211,
+      "step": 20180
+    },
+    {
+      "action_loss": 0.00522614223882556,
+      "epoch": 18.14748201438849,
+      "step": 20180
+    },
+    {
+      "epoch": 18.14748201438849,
+      "step": 20180,
+      "torque_loss": 0.10176622122526169
+    },
+    {
+      "epoch": 18.14748201438849,
+      "force_loss": 0.0024137133732438087,
+      "step": 20180
+    },
+    {
+      "epoch": 18.156474820143885,
+      "grad_norm": 0.3831334710121155,
+      "learning_rate": 7.91904865281929e-05,
+      "loss": 0.022,
+      "step": 20190
+    },
+    {
+      "action_loss": 0.003217557677999139,
+      "epoch": 18.156474820143885,
+      "step": 20190
+    },
+    {
+      "epoch": 18.156474820143885,
+      "step": 20190,
+      "torque_loss": 0.1516803503036499
+    },
+    {
+      "epoch": 18.156474820143885,
+      "force_loss": 0.0023491624742746353,
+      "step": 20190
+    },
+    {
+      "epoch": 18.165467625899282,
+      "grad_norm": 0.6456699967384338,
+      "learning_rate": 7.916810815481307e-05,
+      "loss": 0.0217,
+      "step": 20200
+    },
+    {
+      "action_loss": 0.004174186848104,
+      "epoch": 18.165467625899282,
+      "step": 20200
+    },
+    {
+      "epoch": 18.165467625899282,
+      "step": 20200,
+      "torque_loss": 0.11544787883758545
+    },
+    {
+      "epoch": 18.165467625899282,
+      "force_loss": 0.00204020575620234,
+      "step": 20200
+    },
+    {
+      "epoch": 18.174460431654676,
+      "grad_norm": 0.3879193961620331,
+      "learning_rate": 7.914572092093211e-05,
+      "loss": 0.0233,
+      "step": 20210
+    },
+    {
+      "action_loss": 0.01638071797788143,
+      "epoch": 18.174460431654676,
+      "step": 20210
+    },
+    {
+      "epoch": 18.174460431654676,
+      "step": 20210,
+      "torque_loss": 0.1639884114265442
+    },
+    {
+      "epoch": 18.174460431654676,
+      "force_loss": 0.025213507935404778,
+      "step": 20210
+    },
+    {
+      "epoch": 18.18345323741007,
+      "grad_norm": 0.3586960732936859,
+      "learning_rate": 7.912332483335068e-05,
+      "loss": 0.0231,
+      "step": 20220
+    },
+    {
+      "action_loss": 0.0165130365639925,
+      "epoch": 18.18345323741007,
+      "step": 20220
+    },
+    {
+      "epoch": 18.18345323741007,
+      "step": 20220,
+      "torque_loss": 0.12713371217250824
+    },
+    {
+      "epoch": 18.18345323741007,
+      "force_loss": 0.019182147458195686,
+      "step": 20220
+    },
+    {
+      "epoch": 18.192446043165468,
+      "grad_norm": 0.3122924268245697,
+      "learning_rate": 7.910091989887213e-05,
+      "loss": 0.0214,
+      "step": 20230
+    },
+    {
+      "action_loss": 0.008303462527692318,
+      "epoch": 18.192446043165468,
+      "step": 20230
+    },
+    {
+      "epoch": 18.192446043165468,
+      "step": 20230,
+      "torque_loss": 0.16290760040283203
+    },
+    {
+      "epoch": 18.192446043165468,
+      "force_loss": 0.008843469433486462,
+      "step": 20230
+    },
+    {
+      "epoch": 18.201438848920862,
+      "grad_norm": 0.5016494989395142,
+      "learning_rate": 7.907850612430248e-05,
+      "loss": 0.0246,
+      "step": 20240
+    },
+    {
+      "action_loss": 0.007878461852669716,
+      "epoch": 18.201438848920862,
+      "step": 20240
+    },
+    {
+      "epoch": 18.201438848920862,
+      "step": 20240,
+      "torque_loss": 0.16675801575183868
+    },
+    {
+      "epoch": 18.201438848920862,
+      "force_loss": 0.007841571234166622,
+      "step": 20240
+    },
+    {
+      "epoch": 18.21043165467626,
+      "grad_norm": 0.6306307911872864,
+      "learning_rate": 7.905608351645044e-05,
+      "loss": 0.0216,
+      "step": 20250
+    },
+    {
+      "action_loss": 0.009518210776150227,
+      "epoch": 18.21043165467626,
+      "step": 20250
+    },
+    {
+      "epoch": 18.21043165467626,
+      "step": 20250,
+      "torque_loss": 0.17719995975494385
+    },
+    {
+      "epoch": 18.21043165467626,
+      "force_loss": 0.010739825665950775,
+      "step": 20250
+    },
+    {
+      "epoch": 18.219424460431654,
+      "grad_norm": 0.4184080958366394,
+      "learning_rate": 7.90336520821274e-05,
+      "loss": 0.023,
+      "step": 20260
+    },
+    {
+      "action_loss": 0.009486421011388302,
+      "epoch": 18.219424460431654,
+      "step": 20260
+    },
+    {
+      "epoch": 18.219424460431654,
+      "step": 20260,
+      "torque_loss": 0.16221953928470612
+    },
+    {
+      "epoch": 18.219424460431654,
+      "force_loss": 0.017822276800870895,
+      "step": 20260
+    },
+    {
+      "epoch": 18.22841726618705,
+      "grad_norm": 0.19861479103565216,
+      "learning_rate": 7.901121182814746e-05,
+      "loss": 0.0203,
+      "step": 20270
+    },
+    {
+      "action_loss": 0.0024029635824263096,
+      "epoch": 18.22841726618705,
+      "step": 20270
+    },
+    {
+      "epoch": 18.22841726618705,
+      "step": 20270,
+      "torque_loss": 0.12884463369846344
+    },
+    {
+      "epoch": 18.22841726618705,
+      "force_loss": 0.0030409630853682756,
+      "step": 20270
+    },
+    {
+      "epoch": 18.237410071942445,
+      "grad_norm": 0.16220778226852417,
+      "learning_rate": 7.898876276132736e-05,
+      "loss": 0.0214,
+      "step": 20280
+    },
+    {
+      "action_loss": 0.004468801897019148,
+      "epoch": 18.237410071942445,
+      "step": 20280
+    },
+    {
+      "epoch": 18.237410071942445,
+      "step": 20280,
+      "torque_loss": 0.11765783280134201
+    },
+    {
+      "epoch": 18.237410071942445,
+      "force_loss": 0.0026512593030929565,
+      "step": 20280
+    },
+    {
+      "epoch": 18.246402877697843,
+      "grad_norm": 0.2431236356496811,
+      "learning_rate": 7.896630488848654e-05,
+      "loss": 0.0192,
+      "step": 20290
+    },
+    {
+      "action_loss": 0.009883382357656956,
+      "epoch": 18.246402877697843,
+      "step": 20290
+    },
+    {
+      "epoch": 18.246402877697843,
+      "step": 20290,
+      "torque_loss": 0.16536231338977814
+    },
+    {
+      "epoch": 18.246402877697843,
+      "force_loss": 0.01941848360002041,
+      "step": 20290
+    },
+    {
+      "epoch": 18.255395683453237,
+      "grad_norm": 0.29355722665786743,
+      "learning_rate": 7.89438382164471e-05,
+      "loss": 0.0228,
+      "step": 20300
+    },
+    {
+      "action_loss": 0.011039885692298412,
+      "epoch": 18.255395683453237,
+      "step": 20300
+    },
+    {
+      "epoch": 18.255395683453237,
+      "step": 20300,
+      "torque_loss": 0.17612044513225555
+    },
+    {
+      "epoch": 18.255395683453237,
+      "force_loss": 0.007762711029499769,
+      "step": 20300
+    },
+    {
+      "epoch": 18.264388489208635,
+      "grad_norm": 0.40812596678733826,
+      "learning_rate": 7.892136275203383e-05,
+      "loss": 0.0216,
+      "step": 20310
+    },
+    {
+      "action_loss": 0.0024956001434475183,
+      "epoch": 18.264388489208635,
+      "step": 20310
+    },
+    {
+      "epoch": 18.264388489208635,
+      "step": 20310,
+      "torque_loss": 0.07250147312879562
+    },
+    {
+      "epoch": 18.264388489208635,
+      "force_loss": 0.004585580434650183,
+      "step": 20310
+    },
+    {
+      "epoch": 18.27338129496403,
+      "grad_norm": 0.5498508810997009,
+      "learning_rate": 7.889887850207418e-05,
+      "loss": 0.0208,
+      "step": 20320
+    },
+    {
+      "action_loss": 0.012911583296954632,
+      "epoch": 18.27338129496403,
+      "step": 20320
+    },
+    {
+      "epoch": 18.27338129496403,
+      "step": 20320,
+      "torque_loss": 0.12616920471191406
+    },
+    {
+      "epoch": 18.27338129496403,
+      "force_loss": 0.013621024787425995,
+      "step": 20320
+    },
+    {
+      "epoch": 18.282374100719423,
+      "grad_norm": 0.4653564691543579,
+      "learning_rate": 7.887638547339827e-05,
+      "loss": 0.0245,
+      "step": 20330
+    },
+    {
+      "action_loss": 0.006385440472513437,
+      "epoch": 18.282374100719423,
+      "step": 20330
+    },
+    {
+      "epoch": 18.282374100719423,
+      "step": 20330,
+      "torque_loss": 0.13101039826869965
+    },
+    {
+      "epoch": 18.282374100719423,
+      "force_loss": 0.005413890350610018,
+      "step": 20330
+    },
+    {
+      "epoch": 18.29136690647482,
+      "grad_norm": 0.48891064524650574,
+      "learning_rate": 7.885388367283891e-05,
+      "loss": 0.0215,
+      "step": 20340
+    },
+    {
+      "action_loss": 0.008322554640471935,
+      "epoch": 18.29136690647482,
+      "step": 20340
+    },
+    {
+      "epoch": 18.29136690647482,
+      "step": 20340,
+      "torque_loss": 0.11535265296697617
+    },
+    {
+      "epoch": 18.29136690647482,
+      "force_loss": 0.011928697116672993,
+      "step": 20340
+    },
+    {
+      "epoch": 18.300359712230215,
+      "grad_norm": 0.33788347244262695,
+      "learning_rate": 7.88313731072315e-05,
+      "loss": 0.0223,
+      "step": 20350
+    },
+    {
+      "action_loss": 0.007125012576580048,
+      "epoch": 18.300359712230215,
+      "step": 20350
+    },
+    {
+      "epoch": 18.300359712230215,
+      "step": 20350,
+      "torque_loss": 0.16455142199993134
+    },
+    {
+      "epoch": 18.300359712230215,
+      "force_loss": 0.009632878936827183,
+      "step": 20350
+    },
+    {
+      "epoch": 18.309352517985612,
+      "grad_norm": 0.4722294211387634,
+      "learning_rate": 7.88088537834142e-05,
+      "loss": 0.0219,
+      "step": 20360
+    },
+    {
+      "action_loss": 0.022035181522369385,
+      "epoch": 18.309352517985612,
+      "step": 20360
+    },
+    {
+      "epoch": 18.309352517985612,
+      "step": 20360,
+      "torque_loss": 0.1472807675600052
+    },
+    {
+      "epoch": 18.309352517985612,
+      "force_loss": 0.029020635411143303,
+      "step": 20360
+    },
+    {
+      "epoch": 18.318345323741006,
+      "grad_norm": 0.2702522873878479,
+      "learning_rate": 7.878632570822778e-05,
+      "loss": 0.0225,
+      "step": 20370
+    },
+    {
+      "action_loss": 0.010567483492195606,
+      "epoch": 18.318345323741006,
+      "step": 20370
+    },
+    {
+      "epoch": 18.318345323741006,
+      "step": 20370,
+      "torque_loss": 0.12350978702306747
+    },
+    {
+      "epoch": 18.318345323741006,
+      "force_loss": 0.015000718645751476,
+      "step": 20370
+    },
+    {
+      "epoch": 18.327338129496404,
+      "grad_norm": 0.5101580619812012,
+      "learning_rate": 7.876378888851567e-05,
+      "loss": 0.022,
+      "step": 20380
+    },
+    {
+      "action_loss": 0.011632178910076618,
+      "epoch": 18.327338129496404,
+      "step": 20380
+    },
+    {
+      "epoch": 18.327338129496404,
+      "step": 20380,
+      "torque_loss": 0.11865854263305664
+    },
+    {
+      "epoch": 18.327338129496404,
+      "force_loss": 0.015955500304698944,
+      "step": 20380
+    },
+    {
+      "epoch": 18.336330935251798,
+      "grad_norm": 0.6490873694419861,
+      "learning_rate": 7.874124333112396e-05,
+      "loss": 0.023,
+      "step": 20390
+    },
+    {
+      "action_loss": 0.01874011754989624,
+      "epoch": 18.336330935251798,
+      "step": 20390
+    },
+    {
+      "epoch": 18.336330935251798,
+      "step": 20390,
+      "torque_loss": 0.07617810368537903
+    },
+    {
+      "epoch": 18.336330935251798,
+      "force_loss": 0.026293203234672546,
+      "step": 20390
+    },
+    {
+      "epoch": 18.345323741007196,
+      "grad_norm": 0.2603389024734497,
+      "learning_rate": 7.871868904290138e-05,
+      "loss": 0.0218,
+      "step": 20400
+    },
+    {
+      "action_loss": 0.018921151757240295,
+      "epoch": 18.345323741007196,
+      "step": 20400
+    },
+    {
+      "epoch": 18.345323741007196,
+      "step": 20400,
+      "torque_loss": 0.15553446114063263
+    },
+    {
+      "epoch": 18.345323741007196,
+      "force_loss": 0.024872565641999245,
+      "step": 20400
+    },
+    {
+      "epoch": 18.35431654676259,
+      "grad_norm": 0.3065906763076782,
+      "learning_rate": 7.869612603069935e-05,
+      "loss": 0.0243,
+      "step": 20410
+    },
+    {
+      "action_loss": 0.005943920463323593,
+      "epoch": 18.35431654676259,
+      "step": 20410
+    },
+    {
+      "epoch": 18.35431654676259,
+      "step": 20410,
+      "torque_loss": 0.13668511807918549
+    },
+    {
+      "epoch": 18.35431654676259,
+      "force_loss": 0.005031559616327286,
+      "step": 20410
+    },
+    {
+      "epoch": 18.363309352517987,
+      "grad_norm": 0.20989049971103668,
+      "learning_rate": 7.867355430137192e-05,
+      "loss": 0.0223,
+      "step": 20420
+    },
+    {
+      "action_loss": 0.003977446351200342,
+      "epoch": 18.363309352517987,
+      "step": 20420
+    },
+    {
+      "epoch": 18.363309352517987,
+      "step": 20420,
+      "torque_loss": 0.08900357037782669
+    },
+    {
+      "epoch": 18.363309352517987,
+      "force_loss": 0.00794149935245514,
+      "step": 20420
+    },
+    {
+      "epoch": 18.37230215827338,
+      "grad_norm": 0.2944192588329315,
+      "learning_rate": 7.865097386177577e-05,
+      "loss": 0.0182,
+      "step": 20430
+    },
+    {
+      "action_loss": 0.0303004402667284,
+      "epoch": 18.37230215827338,
+      "step": 20430
+    },
+    {
+      "epoch": 18.37230215827338,
+      "step": 20430,
+      "torque_loss": 0.16158780455589294
+    },
+    {
+      "epoch": 18.37230215827338,
+      "force_loss": 0.02947285957634449,
+      "step": 20430
+    },
+    {
+      "epoch": 18.381294964028775,
+      "grad_norm": 0.32577598094940186,
+      "learning_rate": 7.862838471877023e-05,
+      "loss": 0.0255,
+      "step": 20440
+    },
+    {
+      "action_loss": 0.017475208267569542,
+      "epoch": 18.381294964028775,
+      "step": 20440
+    },
+    {
+      "epoch": 18.381294964028775,
+      "step": 20440,
+      "torque_loss": 0.16580072045326233
+    },
+    {
+      "epoch": 18.381294964028775,
+      "force_loss": 0.014205612242221832,
+      "step": 20440
+    },
+    {
+      "epoch": 18.390287769784173,
+      "grad_norm": 0.3017246127128601,
+      "learning_rate": 7.860578687921731e-05,
+      "loss": 0.0245,
+      "step": 20450
+    },
+    {
+      "action_loss": 0.010980118997395039,
+      "epoch": 18.390287769784173,
+      "step": 20450
+    },
+    {
+      "epoch": 18.390287769784173,
+      "step": 20450,
+      "torque_loss": 0.12659159302711487
+    },
+    {
+      "epoch": 18.390287769784173,
+      "force_loss": 0.019411476328969002,
+      "step": 20450
+    },
+    {
+      "epoch": 18.399280575539567,
+      "grad_norm": 0.3701028823852539,
+      "learning_rate": 7.858318034998164e-05,
+      "loss": 0.0219,
+      "step": 20460
+    },
+    {
+      "action_loss": 0.0061258901841938496,
+      "epoch": 18.399280575539567,
+      "step": 20460
+    },
+    {
+      "epoch": 18.399280575539567,
+      "step": 20460,
+      "torque_loss": 0.13757547736167908
+    },
+    {
+      "epoch": 18.399280575539567,
+      "force_loss": 0.005352713633328676,
+      "step": 20460
+    },
+    {
+      "epoch": 18.408273381294965,
+      "grad_norm": 0.2128572314977646,
+      "learning_rate": 7.856056513793046e-05,
+      "loss": 0.0196,
+      "step": 20470
+    },
+    {
+      "action_loss": 0.009321019984781742,
+      "epoch": 18.408273381294965,
+      "step": 20470
+    },
+    {
+      "epoch": 18.408273381294965,
+      "step": 20470,
+      "torque_loss": 0.13114409148693085
+    },
+    {
+      "epoch": 18.408273381294965,
+      "force_loss": 0.004438444506376982,
+      "step": 20470
+    },
+    {
+      "epoch": 18.41726618705036,
+      "grad_norm": 0.31634169816970825,
+      "learning_rate": 7.85379412499337e-05,
+      "loss": 0.0221,
+      "step": 20480
+    },
+    {
+      "action_loss": 0.006537764798849821,
+      "epoch": 18.41726618705036,
+      "step": 20480
+    },
+    {
+      "epoch": 18.41726618705036,
+      "step": 20480,
+      "torque_loss": 0.1347230076789856
+    },
+    {
+      "epoch": 18.41726618705036,
+      "force_loss": 0.005477122962474823,
+      "step": 20480
+    },
+    {
+      "epoch": 18.426258992805757,
+      "grad_norm": 0.23366904258728027,
+      "learning_rate": 7.851530869286389e-05,
+      "loss": 0.0219,
+      "step": 20490
+    },
+    {
+      "action_loss": 0.00228289351798594,
+      "epoch": 18.426258992805757,
+      "step": 20490
+    },
+    {
+      "epoch": 18.426258992805757,
+      "step": 20490,
+      "torque_loss": 0.125401109457016
+    },
+    {
+      "epoch": 18.426258992805757,
+      "force_loss": 0.0026794367004185915,
+      "step": 20490
+    },
+    {
+      "epoch": 18.43525179856115,
+      "grad_norm": 0.597470760345459,
+      "learning_rate": 7.849266747359619e-05,
+      "loss": 0.021,
+      "step": 20500
+    },
+    {
+      "action_loss": 0.002604519948363304,
+      "epoch": 18.43525179856115,
+      "step": 20500
+    },
+    {
+      "epoch": 18.43525179856115,
+      "step": 20500,
+      "torque_loss": 0.12323381751775742
+    },
+    {
+      "epoch": 18.43525179856115,
+      "force_loss": 0.002440603682771325,
+      "step": 20500
+    },
+    {
+      "epoch": 18.444244604316548,
+      "grad_norm": 0.4298843741416931,
+      "learning_rate": 7.847001759900843e-05,
+      "loss": 0.0174,
+      "step": 20510
+    },
+    {
+      "action_loss": 0.003394864732399583,
+      "epoch": 18.444244604316548,
+      "step": 20510
+    },
+    {
+      "epoch": 18.444244604316548,
+      "step": 20510,
+      "torque_loss": 0.11813098192214966
+    },
+    {
+      "epoch": 18.444244604316548,
+      "force_loss": 0.004520654212683439,
+      "step": 20510
+    },
+    {
+      "epoch": 18.453237410071942,
+      "grad_norm": 0.4122294485569,
+      "learning_rate": 7.844735907598102e-05,
+      "loss": 0.0272,
+      "step": 20520
+    },
+    {
+      "action_loss": 0.02835668809711933,
+      "epoch": 18.453237410071942,
+      "step": 20520
+    },
+    {
+      "epoch": 18.453237410071942,
+      "step": 20520,
+      "torque_loss": 0.12698209285736084
+    },
+    {
+      "epoch": 18.453237410071942,
+      "force_loss": 0.03437325730919838,
+      "step": 20520
+    },
+    {
+      "epoch": 18.46223021582734,
+      "grad_norm": 0.23517845571041107,
+      "learning_rate": 7.842469191139703e-05,
+      "loss": 0.0218,
+      "step": 20530
+    },
+    {
+      "action_loss": 0.00439832080155611,
+      "epoch": 18.46223021582734,
+      "step": 20530
+    },
+    {
+      "epoch": 18.46223021582734,
+      "step": 20530,
+      "torque_loss": 0.15723256766796112
+    },
+    {
+      "epoch": 18.46223021582734,
+      "force_loss": 0.0030727973207831383,
+      "step": 20530
+    },
+    {
+      "epoch": 18.471223021582734,
+      "grad_norm": 0.26438984274864197,
+      "learning_rate": 7.840201611214215e-05,
+      "loss": 0.0215,
+      "step": 20540
+    },
+    {
+      "action_loss": 0.01083553209900856,
+      "epoch": 18.471223021582734,
+      "step": 20540
+    },
+    {
+      "epoch": 18.471223021582734,
+      "step": 20540,
+      "torque_loss": 0.17683565616607666
+    },
+    {
+      "epoch": 18.471223021582734,
+      "force_loss": 0.026332667097449303,
+      "step": 20540
+    },
+    {
+      "epoch": 18.480215827338128,
+      "grad_norm": 0.3710186779499054,
+      "learning_rate": 7.837933168510469e-05,
+      "loss": 0.0298,
+      "step": 20550
+    },
+    {
+      "action_loss": 0.00446340162307024,
+      "epoch": 18.480215827338128,
+      "step": 20550
+    },
+    {
+      "epoch": 18.480215827338128,
+      "step": 20550,
+      "torque_loss": 0.1782977133989334
+    },
+    {
+      "epoch": 18.480215827338128,
+      "force_loss": 0.003301275661215186,
+      "step": 20550
+    },
+    {
+      "epoch": 18.489208633093526,
+      "grad_norm": 0.438558429479599,
+      "learning_rate": 7.835663863717559e-05,
+      "loss": 0.0251,
+      "step": 20560
+    },
+    {
+      "action_loss": 0.012829662300646305,
+      "epoch": 18.489208633093526,
+      "step": 20560
+    },
+    {
+      "epoch": 18.489208633093526,
+      "step": 20560,
+      "torque_loss": 0.12239912897348404
+    },
+    {
+      "epoch": 18.489208633093526,
+      "force_loss": 0.012078560888767242,
+      "step": 20560
+    },
+    {
+      "epoch": 18.49820143884892,
+      "grad_norm": 0.33304500579833984,
+      "learning_rate": 7.833393697524838e-05,
+      "loss": 0.0264,
+      "step": 20570
+    },
+    {
+      "action_loss": 0.003589646890759468,
+      "epoch": 18.49820143884892,
+      "step": 20570
+    },
+    {
+      "epoch": 18.49820143884892,
+      "step": 20570,
+      "torque_loss": 0.08568349480628967
+    },
+    {
+      "epoch": 18.49820143884892,
+      "force_loss": 0.0027285218238830566,
+      "step": 20570
+    },
+    {
+      "epoch": 18.507194244604317,
+      "grad_norm": 0.3924681842327118,
+      "learning_rate": 7.831122670621922e-05,
+      "loss": 0.0226,
+      "step": 20580
+    },
+    {
+      "action_loss": 0.0037728771567344666,
+      "epoch": 18.507194244604317,
+      "step": 20580
+    },
+    {
+      "epoch": 18.507194244604317,
+      "step": 20580,
+      "torque_loss": 0.11202045530080795
+    },
+    {
+      "epoch": 18.507194244604317,
+      "force_loss": 0.004348220303654671,
+      "step": 20580
+    },
+    {
+      "epoch": 18.51618705035971,
+      "grad_norm": 0.23032504320144653,
+      "learning_rate": 7.82885078369869e-05,
+      "loss": 0.0217,
+      "step": 20590
+    },
+    {
+      "action_loss": 0.0051721553318202496,
+      "epoch": 18.51618705035971,
+      "step": 20590
+    },
+    {
+      "epoch": 18.51618705035971,
+      "step": 20590,
+      "torque_loss": 0.1278543621301651
+    },
+    {
+      "epoch": 18.51618705035971,
+      "force_loss": 0.005688309669494629,
+      "step": 20590
+    },
+    {
+      "epoch": 18.52517985611511,
+      "grad_norm": 0.4717794358730316,
+      "learning_rate": 7.826578037445283e-05,
+      "loss": 0.0205,
+      "step": 20600
+    },
+    {
+      "action_loss": 0.0063773407600820065,
+      "epoch": 18.52517985611511,
+      "step": 20600
+    },
+    {
+      "epoch": 18.52517985611511,
+      "step": 20600,
+      "torque_loss": 0.18620699644088745
+    },
+    {
+      "epoch": 18.52517985611511,
+      "force_loss": 0.008187116123735905,
+      "step": 20600
+    },
+    {
+      "epoch": 18.534172661870503,
+      "grad_norm": 0.29804527759552,
+      "learning_rate": 7.824304432552097e-05,
+      "loss": 0.019,
+      "step": 20610
+    },
+    {
+      "action_loss": 0.013948939740657806,
+      "epoch": 18.534172661870503,
+      "step": 20610
+    },
+    {
+      "epoch": 18.534172661870503,
+      "step": 20610,
+      "torque_loss": 0.19339405000209808
+    },
+    {
+      "epoch": 18.534172661870503,
+      "force_loss": 0.010422687977552414,
+      "step": 20610
+    },
+    {
+      "epoch": 18.5431654676259,
+      "grad_norm": 0.39517560601234436,
+      "learning_rate": 7.822029969709798e-05,
+      "loss": 0.0249,
+      "step": 20620
+    },
+    {
+      "action_loss": 0.00812717154622078,
+      "epoch": 18.5431654676259,
+      "step": 20620
+    },
+    {
+      "epoch": 18.5431654676259,
+      "step": 20620,
+      "torque_loss": 0.07682988047599792
+    },
+    {
+      "epoch": 18.5431654676259,
+      "force_loss": 0.006491567939519882,
+      "step": 20620
+    },
+    {
+      "epoch": 18.552158273381295,
+      "grad_norm": 0.27267733216285706,
+      "learning_rate": 7.819754649609306e-05,
+      "loss": 0.0261,
+      "step": 20630
+    },
+    {
+      "action_loss": 0.004313321318477392,
+      "epoch": 18.552158273381295,
+      "step": 20630
+    },
+    {
+      "epoch": 18.552158273381295,
+      "step": 20630,
+      "torque_loss": 0.11302149295806885
+    },
+    {
+      "epoch": 18.552158273381295,
+      "force_loss": 0.0033273298759013414,
+      "step": 20630
+    },
+    {
+      "epoch": 18.56115107913669,
+      "grad_norm": 0.22700157761573792,
+      "learning_rate": 7.817478472941802e-05,
+      "loss": 0.0232,
+      "step": 20640
+    },
+    {
+      "action_loss": 0.005358824972063303,
+      "epoch": 18.56115107913669,
+      "step": 20640
+    },
+    {
+      "epoch": 18.56115107913669,
+      "step": 20640,
+      "torque_loss": 0.11660012602806091
+    },
+    {
+      "epoch": 18.56115107913669,
+      "force_loss": 0.0028624318074434996,
+      "step": 20640
+    },
+    {
+      "epoch": 18.570143884892087,
+      "grad_norm": 0.506298303604126,
+      "learning_rate": 7.815201440398727e-05,
+      "loss": 0.0183,
+      "step": 20650
+    },
+    {
+      "action_loss": 0.01943872682750225,
+      "epoch": 18.570143884892087,
+      "step": 20650
+    },
+    {
+      "epoch": 18.570143884892087,
+      "step": 20650,
+      "torque_loss": 0.19819898903369904
+    },
+    {
+      "epoch": 18.570143884892087,
+      "force_loss": 0.027110660448670387,
+      "step": 20650
+    },
+    {
+      "epoch": 18.57913669064748,
+      "grad_norm": 0.26202455163002014,
+      "learning_rate": 7.812923552671789e-05,
+      "loss": 0.0242,
+      "step": 20660
+    },
+    {
+      "action_loss": 0.005351244937628508,
+      "epoch": 18.57913669064748,
+      "step": 20660
+    },
+    {
+      "epoch": 18.57913669064748,
+      "step": 20660,
+      "torque_loss": 0.1428813487291336
+    },
+    {
+      "epoch": 18.57913669064748,
+      "force_loss": 0.0048134420067071915,
+      "step": 20660
+    },
+    {
+      "epoch": 18.58812949640288,
+      "grad_norm": 0.2282099574804306,
+      "learning_rate": 7.810644810452945e-05,
+      "loss": 0.0195,
+      "step": 20670
+    },
+    {
+      "action_loss": 0.005451902747154236,
+      "epoch": 18.58812949640288,
+      "step": 20670
+    },
+    {
+      "epoch": 18.58812949640288,
+      "step": 20670,
+      "torque_loss": 0.1219620481133461
+    },
+    {
+      "epoch": 18.58812949640288,
+      "force_loss": 0.004392911214381456,
+      "step": 20670
+    },
+    {
+      "epoch": 18.597122302158272,
+      "grad_norm": 0.45334017276763916,
+      "learning_rate": 7.808365214434417e-05,
+      "loss": 0.0209,
+      "step": 20680
+    },
+    {
+      "action_loss": 0.0045912074856460094,
+      "epoch": 18.597122302158272,
+      "step": 20680
+    },
+    {
+      "epoch": 18.597122302158272,
+      "step": 20680,
+      "torque_loss": 0.13276012241840363
+    },
+    {
+      "epoch": 18.597122302158272,
+      "force_loss": 0.0067670163698494434,
+      "step": 20680
+    },
+    {
+      "epoch": 18.60611510791367,
+      "grad_norm": 0.2921150326728821,
+      "learning_rate": 7.80608476530869e-05,
+      "loss": 0.0219,
+      "step": 20690
+    },
+    {
+      "action_loss": 0.002647256478667259,
+      "epoch": 18.60611510791367,
+      "step": 20690
+    },
+    {
+      "epoch": 18.60611510791367,
+      "step": 20690,
+      "torque_loss": 0.128000870347023
+    },
+    {
+      "epoch": 18.60611510791367,
+      "force_loss": 0.002313948003575206,
+      "step": 20690
+    },
+    {
+      "epoch": 18.615107913669064,
+      "grad_norm": 0.2665504813194275,
+      "learning_rate": 7.8038034637685e-05,
+      "loss": 0.0197,
+      "step": 20700
+    },
+    {
+      "action_loss": 0.01289258524775505,
+      "epoch": 18.615107913669064,
+      "step": 20700
+    },
+    {
+      "epoch": 18.615107913669064,
+      "step": 20700,
+      "torque_loss": 0.12954920530319214
+    },
+    {
+      "epoch": 18.615107913669064,
+      "force_loss": 0.011452610604465008,
+      "step": 20700
+    },
+    {
+      "epoch": 18.62410071942446,
+      "grad_norm": 0.37744781374931335,
+      "learning_rate": 7.801521310506848e-05,
+      "loss": 0.0193,
+      "step": 20710
+    },
+    {
+      "action_loss": 0.008950660936534405,
+      "epoch": 18.62410071942446,
+      "step": 20710
+    },
+    {
+      "epoch": 18.62410071942446,
+      "step": 20710,
+      "torque_loss": 0.11152449995279312
+    },
+    {
+      "epoch": 18.62410071942446,
+      "force_loss": 0.005761043634265661,
+      "step": 20710
+    },
+    {
+      "epoch": 18.633093525179856,
+      "grad_norm": 0.4123731851577759,
+      "learning_rate": 7.799238306216994e-05,
+      "loss": 0.0188,
+      "step": 20720
+    },
+    {
+      "action_loss": 0.010444770567119122,
+      "epoch": 18.633093525179856,
+      "step": 20720
+    },
+    {
+      "epoch": 18.633093525179856,
+      "step": 20720,
+      "torque_loss": 0.1856756955385208
+    },
+    {
+      "epoch": 18.633093525179856,
+      "force_loss": 0.009819758124649525,
+      "step": 20720
+    },
+    {
+      "epoch": 18.642086330935253,
+      "grad_norm": 0.2715679407119751,
+      "learning_rate": 7.796954451592448e-05,
+      "loss": 0.0251,
+      "step": 20730
+    },
+    {
+      "action_loss": 0.00551832839846611,
+      "epoch": 18.642086330935253,
+      "step": 20730
+    },
+    {
+      "epoch": 18.642086330935253,
+      "step": 20730,
+      "torque_loss": 0.12976853549480438
+    },
+    {
+      "epoch": 18.642086330935253,
+      "force_loss": 0.0028945046942681074,
+      "step": 20730
+    },
+    {
+      "epoch": 18.651079136690647,
+      "grad_norm": 0.38369905948638916,
+      "learning_rate": 7.794669747326992e-05,
+      "loss": 0.0241,
+      "step": 20740
+    },
+    {
+      "action_loss": 0.005488332360982895,
+      "epoch": 18.651079136690647,
+      "step": 20740
+    },
+    {
+      "epoch": 18.651079136690647,
+      "step": 20740,
+      "torque_loss": 0.12438354641199112
+    },
+    {
+      "epoch": 18.651079136690647,
+      "force_loss": 0.007082134485244751,
+      "step": 20740
+    },
+    {
+      "epoch": 18.66007194244604,
+      "grad_norm": 0.27323734760284424,
+      "learning_rate": 7.792384194114654e-05,
+      "loss": 0.0205,
+      "step": 20750
+    },
+    {
+      "action_loss": 0.008285093121230602,
+      "epoch": 18.66007194244604,
+      "step": 20750
+    },
+    {
+      "epoch": 18.66007194244604,
+      "step": 20750,
+      "torque_loss": 0.13652460277080536
+    },
+    {
+      "epoch": 18.66007194244604,
+      "force_loss": 0.010674399323761463,
+      "step": 20750
+    },
+    {
+      "epoch": 18.66906474820144,
+      "grad_norm": 1.0483168363571167,
+      "learning_rate": 7.790097792649729e-05,
+      "loss": 0.0228,
+      "step": 20760
+    },
+    {
+      "action_loss": 0.0167890265583992,
+      "epoch": 18.66906474820144,
+      "step": 20760
+    },
+    {
+      "epoch": 18.66906474820144,
+      "step": 20760,
+      "torque_loss": 0.16304481029510498
+    },
+    {
+      "epoch": 18.66906474820144,
+      "force_loss": 0.018465260043740273,
+      "step": 20760
+    },
+    {
+      "epoch": 18.678057553956833,
+      "grad_norm": 0.28037571907043457,
+      "learning_rate": 7.787810543626762e-05,
+      "loss": 0.0209,
+      "step": 20770
+    },
+    {
+      "action_loss": 0.008035396225750446,
+      "epoch": 18.678057553956833,
+      "step": 20770
+    },
+    {
+      "epoch": 18.678057553956833,
+      "step": 20770,
+      "torque_loss": 0.1455891877412796
+    },
+    {
+      "epoch": 18.678057553956833,
+      "force_loss": 0.006655184086412191,
+      "step": 20770
+    },
+    {
+      "epoch": 18.68705035971223,
+      "grad_norm": 0.34441569447517395,
+      "learning_rate": 7.785522447740558e-05,
+      "loss": 0.0194,
+      "step": 20780
+    },
+    {
+      "action_loss": 0.004798682872205973,
+      "epoch": 18.68705035971223,
+      "step": 20780
+    },
+    {
+      "epoch": 18.68705035971223,
+      "step": 20780,
+      "torque_loss": 0.11174693703651428
+    },
+    {
+      "epoch": 18.68705035971223,
+      "force_loss": 0.0041641718707978725,
+      "step": 20780
+    },
+    {
+      "epoch": 18.696043165467625,
+      "grad_norm": 0.34278228878974915,
+      "learning_rate": 7.783233505686182e-05,
+      "loss": 0.0274,
+      "step": 20790
+    },
+    {
+      "action_loss": 0.0032451439183205366,
+      "epoch": 18.696043165467625,
+      "step": 20790
+    },
+    {
+      "epoch": 18.696043165467625,
+      "step": 20790,
+      "torque_loss": 0.138041689991951
+    },
+    {
+      "epoch": 18.696043165467625,
+      "force_loss": 0.0033103812020272017,
+      "step": 20790
+    },
+    {
+      "epoch": 18.705035971223023,
+      "grad_norm": 0.2710281014442444,
+      "learning_rate": 7.780943718158955e-05,
+      "loss": 0.0189,
+      "step": 20800
+    },
+    {
+      "action_loss": 0.004741458687931299,
+      "epoch": 18.705035971223023,
+      "step": 20800
+    },
+    {
+      "epoch": 18.705035971223023,
+      "step": 20800,
+      "torque_loss": 0.14033029973506927
+    },
+    {
+      "epoch": 18.705035971223023,
+      "force_loss": 0.008525731973350048,
+      "step": 20800
+    },
+    {
+      "epoch": 18.714028776978417,
+      "grad_norm": 0.4619952142238617,
+      "learning_rate": 7.778653085854453e-05,
+      "loss": 0.0226,
+      "step": 20810
+    },
+    {
+      "action_loss": 0.0035277584102004766,
+      "epoch": 18.714028776978417,
+      "step": 20810
+    },
+    {
+      "epoch": 18.714028776978417,
+      "step": 20810,
+      "torque_loss": 0.06143399700522423
+    },
+    {
+      "epoch": 18.714028776978417,
+      "force_loss": 0.005184220615774393,
+      "step": 20810
+    },
+    {
+      "epoch": 18.723021582733814,
+      "grad_norm": 0.35060372948646545,
+      "learning_rate": 7.77636160946851e-05,
+      "loss": 0.0195,
+      "step": 20820
+    },
+    {
+      "action_loss": 0.0045479764230549335,
+      "epoch": 18.723021582733814,
+      "step": 20820
+    },
+    {
+      "epoch": 18.723021582733814,
+      "step": 20820,
+      "torque_loss": 0.12226519733667374
+    },
+    {
+      "epoch": 18.723021582733814,
+      "force_loss": 0.009367171674966812,
+      "step": 20820
+    },
+    {
+      "epoch": 18.73201438848921,
+      "grad_norm": 0.45519742369651794,
+      "learning_rate": 7.774069289697215e-05,
+      "loss": 0.0229,
+      "step": 20830
+    },
+    {
+      "action_loss": 0.005259813740849495,
+      "epoch": 18.73201438848921,
+      "step": 20830
+    },
+    {
+      "epoch": 18.73201438848921,
+      "step": 20830,
+      "torque_loss": 0.13091827929019928
+    },
+    {
+      "epoch": 18.73201438848921,
+      "force_loss": 0.0031828416977077723,
+      "step": 20830
+    },
+    {
+      "epoch": 18.741007194244606,
+      "grad_norm": 0.2008289098739624,
+      "learning_rate": 7.771776127236913e-05,
+      "loss": 0.0208,
+      "step": 20840
+    },
+    {
+      "action_loss": 0.006448762025684118,
+      "epoch": 18.741007194244606,
+      "step": 20840
+    },
+    {
+      "epoch": 18.741007194244606,
+      "step": 20840,
+      "torque_loss": 0.14278720319271088
+    },
+    {
+      "epoch": 18.741007194244606,
+      "force_loss": 0.004897276405245066,
+      "step": 20840
+    },
+    {
+      "epoch": 18.75,
+      "grad_norm": 0.22035515308380127,
+      "learning_rate": 7.769482122784212e-05,
+      "loss": 0.02,
+      "step": 20850
+    },
+    {
+      "action_loss": 0.012750628404319286,
+      "epoch": 18.75,
+      "step": 20850
+    },
+    {
+      "epoch": 18.75,
+      "step": 20850,
+      "torque_loss": 0.17558342218399048
+    },
+    {
+      "epoch": 18.75,
+      "force_loss": 0.01920117251574993,
+      "step": 20850
+    },
+    {
+      "epoch": 18.758992805755394,
+      "grad_norm": 0.5520920157432556,
+      "learning_rate": 7.767187277035963e-05,
+      "loss": 0.0206,
+      "step": 20860
+    },
+    {
+      "action_loss": 0.013377487659454346,
+      "epoch": 18.758992805755394,
+      "step": 20860
+    },
+    {
+      "epoch": 18.758992805755394,
+      "step": 20860,
+      "torque_loss": 0.13168296217918396
+    },
+    {
+      "epoch": 18.758992805755394,
+      "force_loss": 0.01273645181208849,
+      "step": 20860
+    },
+    {
+      "epoch": 18.76798561151079,
+      "grad_norm": 0.3895902633666992,
+      "learning_rate": 7.764891590689285e-05,
+      "loss": 0.0245,
+      "step": 20870
+    },
+    {
+      "action_loss": 0.007989990524947643,
+      "epoch": 18.76798561151079,
+      "step": 20870
+    },
+    {
+      "epoch": 18.76798561151079,
+      "step": 20870,
+      "torque_loss": 0.20341791212558746
+    },
+    {
+      "epoch": 18.76798561151079,
+      "force_loss": 0.006274865940213203,
+      "step": 20870
+    },
+    {
+      "epoch": 18.776978417266186,
+      "grad_norm": 0.3786901533603668,
+      "learning_rate": 7.762595064441542e-05,
+      "loss": 0.0227,
+      "step": 20880
+    },
+    {
+      "action_loss": 0.020759470760822296,
+      "epoch": 18.776978417266186,
+      "step": 20880
+    },
+    {
+      "epoch": 18.776978417266186,
+      "step": 20880,
+      "torque_loss": 0.15657936036586761
+    },
+    {
+      "epoch": 18.776978417266186,
+      "force_loss": 0.022274455055594444,
+      "step": 20880
+    },
+    {
+      "epoch": 18.785971223021583,
+      "grad_norm": 0.3006162941455841,
+      "learning_rate": 7.760297698990362e-05,
+      "loss": 0.0211,
+      "step": 20890
+    },
+    {
+      "action_loss": 0.005678398068994284,
+      "epoch": 18.785971223021583,
+      "step": 20890
+    },
+    {
+      "epoch": 18.785971223021583,
+      "step": 20890,
+      "torque_loss": 0.08947417140007019
+    },
+    {
+      "epoch": 18.785971223021583,
+      "force_loss": 0.00516447564586997,
+      "step": 20890
+    },
+    {
+      "epoch": 18.794964028776977,
+      "grad_norm": 0.5261251926422119,
+      "learning_rate": 7.757999495033623e-05,
+      "loss": 0.0231,
+      "step": 20900
+    },
+    {
+      "action_loss": 0.00572761707007885,
+      "epoch": 18.794964028776977,
+      "step": 20900
+    },
+    {
+      "epoch": 18.794964028776977,
+      "step": 20900,
+      "torque_loss": 0.16962863504886627
+    },
+    {
+      "epoch": 18.794964028776977,
+      "force_loss": 0.004866601899266243,
+      "step": 20900
+    },
+    {
+      "epoch": 18.803956834532375,
+      "grad_norm": 0.20297500491142273,
+      "learning_rate": 7.755700453269456e-05,
+      "loss": 0.019,
+      "step": 20910
+    },
+    {
+      "action_loss": 0.004213298205286264,
+      "epoch": 18.803956834532375,
+      "step": 20910
+    },
+    {
+      "epoch": 18.803956834532375,
+      "step": 20910,
+      "torque_loss": 0.21491968631744385
+    },
+    {
+      "epoch": 18.803956834532375,
+      "force_loss": 0.002340235747396946,
+      "step": 20910
+    },
+    {
+      "epoch": 18.81294964028777,
+      "grad_norm": 0.6163339614868164,
+      "learning_rate": 7.753400574396254e-05,
+      "loss": 0.0237,
+      "step": 20920
+    },
+    {
+      "action_loss": 0.0036244343500584364,
+      "epoch": 18.81294964028777,
+      "step": 20920
+    },
+    {
+      "epoch": 18.81294964028777,
+      "step": 20920,
+      "torque_loss": 0.13420718908309937
+    },
+    {
+      "epoch": 18.81294964028777,
+      "force_loss": 0.002731218934059143,
+      "step": 20920
+    },
+    {
+      "epoch": 18.821942446043167,
+      "grad_norm": 0.5458612442016602,
+      "learning_rate": 7.751099859112655e-05,
+      "loss": 0.025,
+      "step": 20930
+    },
+    {
+      "action_loss": 0.006525054574012756,
+      "epoch": 18.821942446043167,
+      "step": 20930
+    },
+    {
+      "epoch": 18.821942446043167,
+      "step": 20930,
+      "torque_loss": 0.12040508538484573
+    },
+    {
+      "epoch": 18.821942446043167,
+      "force_loss": 0.014152037911117077,
+      "step": 20930
+    },
+    {
+      "epoch": 18.83093525179856,
+      "grad_norm": 0.2515645921230316,
+      "learning_rate": 7.748798308117557e-05,
+      "loss": 0.0196,
+      "step": 20940
+    },
+    {
+      "action_loss": 0.005239488091319799,
+      "epoch": 18.83093525179856,
+      "step": 20940
+    },
+    {
+      "epoch": 18.83093525179856,
+      "step": 20940,
+      "torque_loss": 0.11633515357971191
+    },
+    {
+      "epoch": 18.83093525179856,
+      "force_loss": 0.005713296588510275,
+      "step": 20940
+    },
+    {
+      "epoch": 18.83992805755396,
+      "grad_norm": 0.46672412753105164,
+      "learning_rate": 7.746495922110112e-05,
+      "loss": 0.0223,
+      "step": 20950
+    },
+    {
+      "action_loss": 0.008477890864014626,
+      "epoch": 18.83992805755396,
+      "step": 20950
+    },
+    {
+      "epoch": 18.83992805755396,
+      "step": 20950,
+      "torque_loss": 0.17128349840641022
+    },
+    {
+      "epoch": 18.83992805755396,
+      "force_loss": 0.00535630201920867,
+      "step": 20950
+    },
+    {
+      "epoch": 18.848920863309353,
+      "grad_norm": 0.14538341760635376,
+      "learning_rate": 7.744192701789723e-05,
+      "loss": 0.0201,
+      "step": 20960
+    },
+    {
+      "action_loss": 0.0038794800639152527,
+      "epoch": 18.848920863309353,
+      "step": 20960
+    },
+    {
+      "epoch": 18.848920863309353,
+      "step": 20960,
+      "torque_loss": 0.1321813017129898
+    },
+    {
+      "epoch": 18.848920863309353,
+      "force_loss": 0.002145926933735609,
+      "step": 20960
+    },
+    {
+      "epoch": 18.857913669064747,
+      "grad_norm": 0.30010467767715454,
+      "learning_rate": 7.741888647856046e-05,
+      "loss": 0.0201,
+      "step": 20970
+    },
+    {
+      "action_loss": 0.00981205701828003,
+      "epoch": 18.857913669064747,
+      "step": 20970
+    },
+    {
+      "epoch": 18.857913669064747,
+      "step": 20970,
+      "torque_loss": 0.1619131714105606
+    },
+    {
+      "epoch": 18.857913669064747,
+      "force_loss": 0.00961390696465969,
+      "step": 20970
+    },
+    {
+      "epoch": 18.866906474820144,
+      "grad_norm": 0.2195420116186142,
+      "learning_rate": 7.739583761008994e-05,
+      "loss": 0.022,
+      "step": 20980
+    },
+    {
+      "action_loss": 0.01965527981519699,
+      "epoch": 18.866906474820144,
+      "step": 20980
+    },
+    {
+      "epoch": 18.866906474820144,
+      "step": 20980,
+      "torque_loss": 0.13485606014728546
+    },
+    {
+      "epoch": 18.866906474820144,
+      "force_loss": 0.010004352778196335,
+      "step": 20980
+    },
+    {
+      "epoch": 18.87589928057554,
+      "grad_norm": 0.580980658531189,
+      "learning_rate": 7.73727804194873e-05,
+      "loss": 0.0208,
+      "step": 20990
+    },
+    {
+      "action_loss": 0.005588372703641653,
+      "epoch": 18.87589928057554,
+      "step": 20990
+    },
+    {
+      "epoch": 18.87589928057554,
+      "step": 20990,
+      "torque_loss": 0.18242399394512177
+    },
+    {
+      "epoch": 18.87589928057554,
+      "force_loss": 0.00622784486040473,
+      "step": 20990
+    },
+    {
+      "epoch": 18.884892086330936,
+      "grad_norm": 0.3247342109680176,
+      "learning_rate": 7.734971491375671e-05,
+      "loss": 0.0209,
+      "step": 21000
+    },
+    {
+      "action_loss": 0.002195393666625023,
+      "epoch": 18.884892086330936,
+      "step": 21000
+    },
+    {
+      "epoch": 18.884892086330936,
+      "step": 21000,
+      "torque_loss": 0.08245361596345901
+    },
+    {
+      "epoch": 18.884892086330936,
+      "force_loss": 0.0029244597535580397,
+      "step": 21000
+    },
+    {
+      "epoch": 18.89388489208633,
+      "grad_norm": 0.2926085293292999,
+      "learning_rate": 7.732664109990485e-05,
+      "loss": 0.0208,
+      "step": 21010
+    },
+    {
+      "action_loss": 0.006941869854927063,
+      "epoch": 18.89388489208633,
+      "step": 21010
+    },
+    {
+      "epoch": 18.89388489208633,
+      "step": 21010,
+      "torque_loss": 0.1329679787158966
+    },
+    {
+      "epoch": 18.89388489208633,
+      "force_loss": 0.0030126480851322412,
+      "step": 21010
+    },
+    {
+      "epoch": 18.902877697841728,
+      "grad_norm": 0.2950783967971802,
+      "learning_rate": 7.730355898494095e-05,
+      "loss": 0.0304,
+      "step": 21020
+    },
+    {
+      "action_loss": 0.011579719372093678,
+      "epoch": 18.902877697841728,
+      "step": 21020
+    },
+    {
+      "epoch": 18.902877697841728,
+      "step": 21020,
+      "torque_loss": 0.17243927717208862
+    },
+    {
+      "epoch": 18.902877697841728,
+      "force_loss": 0.009910565800964832,
+      "step": 21020
+    },
+    {
+      "epoch": 18.91187050359712,
+      "grad_norm": 0.3713012635707855,
+      "learning_rate": 7.728046857587673e-05,
+      "loss": 0.0269,
+      "step": 21030
+    },
+    {
+      "action_loss": 0.00507223280146718,
+      "epoch": 18.91187050359712,
+      "step": 21030
+    },
+    {
+      "epoch": 18.91187050359712,
+      "step": 21030,
+      "torque_loss": 0.12365678697824478
+    },
+    {
+      "epoch": 18.91187050359712,
+      "force_loss": 0.008763331919908524,
+      "step": 21030
+    },
+    {
+      "epoch": 18.92086330935252,
+      "grad_norm": 0.7931478023529053,
+      "learning_rate": 7.725736987972647e-05,
+      "loss": 0.0248,
+      "step": 21040
+    },
+    {
+      "action_loss": 0.009880068711936474,
+      "epoch": 18.92086330935252,
+      "step": 21040
+    },
+    {
+      "epoch": 18.92086330935252,
+      "step": 21040,
+      "torque_loss": 0.12279900908470154
+    },
+    {
+      "epoch": 18.92086330935252,
+      "force_loss": 0.009871678426861763,
+      "step": 21040
+    },
+    {
+      "epoch": 18.929856115107913,
+      "grad_norm": 0.4834195673465729,
+      "learning_rate": 7.723426290350691e-05,
+      "loss": 0.0245,
+      "step": 21050
+    },
+    {
+      "action_loss": 0.011661668308079243,
+      "epoch": 18.929856115107913,
+      "step": 21050
+    },
+    {
+      "epoch": 18.929856115107913,
+      "step": 21050,
+      "torque_loss": 0.14733369648456573
+    },
+    {
+      "epoch": 18.929856115107913,
+      "force_loss": 0.012337010353803635,
+      "step": 21050
+    },
+    {
+      "epoch": 18.93884892086331,
+      "grad_norm": 0.3356728255748749,
+      "learning_rate": 7.721114765423736e-05,
+      "loss": 0.0249,
+      "step": 21060
+    },
+    {
+      "action_loss": 0.008570664562284946,
+      "epoch": 18.93884892086331,
+      "step": 21060
+    },
+    {
+      "epoch": 18.93884892086331,
+      "step": 21060,
+      "torque_loss": 0.10762292146682739
+    },
+    {
+      "epoch": 18.93884892086331,
+      "force_loss": 0.009123653173446655,
+      "step": 21060
+    },
+    {
+      "epoch": 18.947841726618705,
+      "grad_norm": 0.27851220965385437,
+      "learning_rate": 7.718802413893963e-05,
+      "loss": 0.0215,
+      "step": 21070
+    },
+    {
+      "action_loss": 0.007585987448692322,
+      "epoch": 18.947841726618705,
+      "step": 21070
+    },
+    {
+      "epoch": 18.947841726618705,
+      "step": 21070,
+      "torque_loss": 0.11948993057012558
+    },
+    {
+      "epoch": 18.947841726618705,
+      "force_loss": 0.0066976421512663364,
+      "step": 21070
+    },
+    {
+      "epoch": 18.9568345323741,
+      "grad_norm": 0.35844120383262634,
+      "learning_rate": 7.716489236463802e-05,
+      "loss": 0.0208,
+      "step": 21080
+    },
+    {
+      "action_loss": 0.007779898587614298,
+      "epoch": 18.9568345323741,
+      "step": 21080
+    },
+    {
+      "epoch": 18.9568345323741,
+      "step": 21080,
+      "torque_loss": 0.12000761181116104
+    },
+    {
+      "epoch": 18.9568345323741,
+      "force_loss": 0.015624753199517727,
+      "step": 21080
+    },
+    {
+      "epoch": 18.965827338129497,
+      "grad_norm": 0.16129791736602783,
+      "learning_rate": 7.714175233835936e-05,
+      "loss": 0.0197,
+      "step": 21090
+    },
+    {
+      "action_loss": 0.0030143007170408964,
+      "epoch": 18.965827338129497,
+      "step": 21090
+    },
+    {
+      "epoch": 18.965827338129497,
+      "step": 21090,
+      "torque_loss": 0.12497544288635254
+    },
+    {
+      "epoch": 18.965827338129497,
+      "force_loss": 0.0027070678770542145,
+      "step": 21090
+    },
+    {
+      "epoch": 18.97482014388489,
+      "grad_norm": 0.3578173816204071,
+      "learning_rate": 7.711860406713299e-05,
+      "loss": 0.0202,
+      "step": 21100
+    },
+    {
+      "action_loss": 0.005643816199153662,
+      "epoch": 18.97482014388489,
+      "step": 21100
+    },
+    {
+      "epoch": 18.97482014388489,
+      "step": 21100,
+      "torque_loss": 0.12769551575183868
+    },
+    {
+      "epoch": 18.97482014388489,
+      "force_loss": 0.0054486957378685474,
+      "step": 21100
+    },
+    {
+      "epoch": 18.98381294964029,
+      "grad_norm": 0.20371337234973907,
+      "learning_rate": 7.70954475579907e-05,
+      "loss": 0.0206,
+      "step": 21110
+    },
+    {
+      "action_loss": 0.004199328366667032,
+      "epoch": 18.98381294964029,
+      "step": 21110
+    },
+    {
+      "epoch": 18.98381294964029,
+      "step": 21110,
+      "torque_loss": 0.13115596771240234
+    },
+    {
+      "epoch": 18.98381294964029,
+      "force_loss": 0.003864569589495659,
+      "step": 21110
+    },
+    {
+      "epoch": 18.992805755395683,
+      "grad_norm": 0.4181339740753174,
+      "learning_rate": 7.707228281796688e-05,
+      "loss": 0.0211,
+      "step": 21120
+    },
+    {
+      "action_loss": 0.004545002710074186,
+      "epoch": 18.992805755395683,
+      "step": 21120
+    },
+    {
+      "epoch": 18.992805755395683,
+      "step": 21120,
+      "torque_loss": 0.14200468361377716
+    },
+    {
+      "epoch": 18.992805755395683,
+      "force_loss": 0.002078678458929062,
+      "step": 21120
+    },
+    {
+      "epoch": 19.00179856115108,
+      "grad_norm": 0.45750585198402405,
+      "learning_rate": 7.704910985409833e-05,
+      "loss": 0.0242,
+      "step": 21130
+    },
+    {
+      "action_loss": 0.01091001182794571,
+      "epoch": 19.00179856115108,
+      "step": 21130
+    },
+    {
+      "epoch": 19.00179856115108,
+      "step": 21130,
+      "torque_loss": 0.15371786057949066
+    },
+    {
+      "epoch": 19.00179856115108,
+      "force_loss": 0.014564561657607555,
+      "step": 21130
+    },
+    {
+      "epoch": 19.010791366906474,
+      "grad_norm": 0.37320059537887573,
+      "learning_rate": 7.702592867342439e-05,
+      "loss": 0.0264,
+      "step": 21140
+    },
+    {
+      "action_loss": 0.009730189107358456,
+      "epoch": 19.010791366906474,
+      "step": 21140
+    },
+    {
+      "epoch": 19.010791366906474,
+      "step": 21140,
+      "torque_loss": 0.18975459039211273
+    },
+    {
+      "epoch": 19.010791366906474,
+      "force_loss": 0.006682038307189941,
+      "step": 21140
+    },
+    {
+      "epoch": 19.019784172661872,
+      "grad_norm": 0.2717139422893524,
+      "learning_rate": 7.700273928298691e-05,
+      "loss": 0.0201,
+      "step": 21150
+    },
+    {
+      "action_loss": 0.006085825618356466,
+      "epoch": 19.019784172661872,
+      "step": 21150
+    },
+    {
+      "epoch": 19.019784172661872,
+      "step": 21150,
+      "torque_loss": 0.12803782522678375
+    },
+    {
+      "epoch": 19.019784172661872,
+      "force_loss": 0.008105587214231491,
+      "step": 21150
+    },
+    {
+      "epoch": 19.028776978417266,
+      "grad_norm": 0.2965158224105835,
+      "learning_rate": 7.697954168983021e-05,
+      "loss": 0.0223,
+      "step": 21160
+    },
+    {
+      "action_loss": 0.005119136068969965,
+      "epoch": 19.028776978417266,
+      "step": 21160
+    },
+    {
+      "epoch": 19.028776978417266,
+      "step": 21160,
+      "torque_loss": 0.0973953828215599
+    },
+    {
+      "epoch": 19.028776978417266,
+      "force_loss": 0.0038145689759403467,
+      "step": 21160
+    },
+    {
+      "epoch": 19.037769784172664,
+      "grad_norm": 0.30059152841567993,
+      "learning_rate": 7.695633590100109e-05,
+      "loss": 0.0232,
+      "step": 21170
+    },
+    {
+      "action_loss": 0.00632014824077487,
+      "epoch": 19.037769784172664,
+      "step": 21170
+    },
+    {
+      "epoch": 19.037769784172664,
+      "step": 21170,
+      "torque_loss": 0.14109350740909576
+    },
+    {
+      "epoch": 19.037769784172664,
+      "force_loss": 0.009804566390812397,
+      "step": 21170
+    },
+    {
+      "epoch": 19.046762589928058,
+      "grad_norm": 0.41945117712020874,
+      "learning_rate": 7.693312192354886e-05,
+      "loss": 0.0224,
+      "step": 21180
+    },
+    {
+      "action_loss": 0.0068938713520765305,
+      "epoch": 19.046762589928058,
+      "step": 21180
+    },
+    {
+      "epoch": 19.046762589928058,
+      "step": 21180,
+      "torque_loss": 0.13942526280879974
+    },
+    {
+      "epoch": 19.046762589928058,
+      "force_loss": 0.01081293448805809,
+      "step": 21180
+    },
+    {
+      "epoch": 19.055755395683452,
+      "grad_norm": 0.2478008270263672,
+      "learning_rate": 7.690989976452532e-05,
+      "loss": 0.0207,
+      "step": 21190
+    },
+    {
+      "action_loss": 0.006615053862333298,
+      "epoch": 19.055755395683452,
+      "step": 21190
+    },
+    {
+      "epoch": 19.055755395683452,
+      "step": 21190,
+      "torque_loss": 0.1412479132413864
+    },
+    {
+      "epoch": 19.055755395683452,
+      "force_loss": 0.005072843749076128,
+      "step": 21190
+    },
+    {
+      "epoch": 19.06474820143885,
+      "grad_norm": 0.41966503858566284,
+      "learning_rate": 7.688666943098475e-05,
+      "loss": 0.0219,
+      "step": 21200
+    },
+    {
+      "action_loss": 0.00664830207824707,
+      "epoch": 19.06474820143885,
+      "step": 21200
+    },
+    {
+      "epoch": 19.06474820143885,
+      "step": 21200,
+      "torque_loss": 0.15645386278629303
+    },
+    {
+      "epoch": 19.06474820143885,
+      "force_loss": 0.00666293129324913,
+      "step": 21200
+    },
+    {
+      "epoch": 19.073741007194243,
+      "grad_norm": 0.21348387002944946,
+      "learning_rate": 7.686343092998389e-05,
+      "loss": 0.019,
+      "step": 21210
+    },
+    {
+      "action_loss": 0.004536744207143784,
+      "epoch": 19.073741007194243,
+      "step": 21210
+    },
+    {
+      "epoch": 19.073741007194243,
+      "step": 21210,
+      "torque_loss": 0.14316974580287933
+    },
+    {
+      "epoch": 19.073741007194243,
+      "force_loss": 0.004935022909194231,
+      "step": 21210
+    },
+    {
+      "epoch": 19.08273381294964,
+      "grad_norm": 0.31518983840942383,
+      "learning_rate": 7.684018426858202e-05,
+      "loss": 0.0254,
+      "step": 21220
+    },
+    {
+      "action_loss": 0.003023432567715645,
+      "epoch": 19.08273381294964,
+      "step": 21220
+    },
+    {
+      "epoch": 19.08273381294964,
+      "step": 21220,
+      "torque_loss": 0.1416092962026596
+    },
+    {
+      "epoch": 19.08273381294964,
+      "force_loss": 0.0018521612510085106,
+      "step": 21220
+    },
+    {
+      "epoch": 19.091726618705035,
+      "grad_norm": 0.42931514978408813,
+      "learning_rate": 7.681692945384084e-05,
+      "loss": 0.0217,
+      "step": 21230
+    },
+    {
+      "action_loss": 0.004434697795659304,
+      "epoch": 19.091726618705035,
+      "step": 21230
+    },
+    {
+      "epoch": 19.091726618705035,
+      "step": 21230,
+      "torque_loss": 0.10643320530653
+    },
+    {
+      "epoch": 19.091726618705035,
+      "force_loss": 0.006954862270504236,
+      "step": 21230
+    },
+    {
+      "epoch": 19.100719424460433,
+      "grad_norm": 0.18268223106861115,
+      "learning_rate": 7.679366649282456e-05,
+      "loss": 0.0185,
+      "step": 21240
+    },
+    {
+      "action_loss": 0.0451851524412632,
+      "epoch": 19.100719424460433,
+      "step": 21240
+    },
+    {
+      "epoch": 19.100719424460433,
+      "step": 21240,
+      "torque_loss": 0.17587248980998993
+    },
+    {
+      "epoch": 19.100719424460433,
+      "force_loss": 0.0445452444255352,
+      "step": 21240
+    },
+    {
+      "epoch": 19.109712230215827,
+      "grad_norm": 0.23430714011192322,
+      "learning_rate": 7.677039539259983e-05,
+      "loss": 0.0237,
+      "step": 21250
+    },
+    {
+      "action_loss": 0.019339272752404213,
+      "epoch": 19.109712230215827,
+      "step": 21250
+    },
+    {
+      "epoch": 19.109712230215827,
+      "step": 21250,
+      "torque_loss": 0.17584173381328583
+    },
+    {
+      "epoch": 19.109712230215827,
+      "force_loss": 0.021408798173069954,
+      "step": 21250
+    },
+    {
+      "epoch": 19.118705035971225,
+      "grad_norm": 0.36534684896469116,
+      "learning_rate": 7.674711616023581e-05,
+      "loss": 0.0243,
+      "step": 21260
+    },
+    {
+      "action_loss": 0.005523163825273514,
+      "epoch": 19.118705035971225,
+      "step": 21260
+    },
+    {
+      "epoch": 19.118705035971225,
+      "step": 21260,
+      "torque_loss": 0.15490221977233887
+    },
+    {
+      "epoch": 19.118705035971225,
+      "force_loss": 0.007282101083546877,
+      "step": 21260
+    },
+    {
+      "epoch": 19.12769784172662,
+      "grad_norm": 0.44240134954452515,
+      "learning_rate": 7.672382880280413e-05,
+      "loss": 0.0203,
+      "step": 21270
+    },
+    {
+      "action_loss": 0.009951199404895306,
+      "epoch": 19.12769784172662,
+      "step": 21270
+    },
+    {
+      "epoch": 19.12769784172662,
+      "step": 21270,
+      "torque_loss": 0.13297080993652344
+    },
+    {
+      "epoch": 19.12769784172662,
+      "force_loss": 0.008881856687366962,
+      "step": 21270
+    },
+    {
+      "epoch": 19.136690647482013,
+      "grad_norm": 0.36340591311454773,
+      "learning_rate": 7.670053332737885e-05,
+      "loss": 0.0217,
+      "step": 21280
+    },
+    {
+      "action_loss": 0.002128706779330969,
+      "epoch": 19.136690647482013,
+      "step": 21280
+    },
+    {
+      "epoch": 19.136690647482013,
+      "step": 21280,
+      "torque_loss": 0.07681339234113693
+    },
+    {
+      "epoch": 19.136690647482013,
+      "force_loss": 0.001706802868284285,
+      "step": 21280
+    },
+    {
+      "epoch": 19.14568345323741,
+      "grad_norm": 0.4394150674343109,
+      "learning_rate": 7.667722974103654e-05,
+      "loss": 0.0172,
+      "step": 21290
+    },
+    {
+      "action_loss": 0.0016647170996293426,
+      "epoch": 19.14568345323741,
+      "step": 21290
+    },
+    {
+      "epoch": 19.14568345323741,
+      "step": 21290,
+      "torque_loss": 0.06562650948762894
+    },
+    {
+      "epoch": 19.14568345323741,
+      "force_loss": 0.0011401548981666565,
+      "step": 21290
+    },
+    {
+      "epoch": 19.154676258992804,
+      "grad_norm": 1.3382195234298706,
+      "learning_rate": 7.66539180508562e-05,
+      "loss": 0.0199,
+      "step": 21300
+    },
+    {
+      "action_loss": 0.009306578896939754,
+      "epoch": 19.154676258992804,
+      "step": 21300
+    },
+    {
+      "epoch": 19.154676258992804,
+      "step": 21300,
+      "torque_loss": 0.17482580244541168
+    },
+    {
+      "epoch": 19.154676258992804,
+      "force_loss": 0.013695008121430874,
+      "step": 21300
+    },
+    {
+      "epoch": 19.163669064748202,
+      "grad_norm": 0.5320211052894592,
+      "learning_rate": 7.663059826391932e-05,
+      "loss": 0.0246,
+      "step": 21310
+    },
+    {
+      "action_loss": 0.004609776195138693,
+      "epoch": 19.163669064748202,
+      "step": 21310
+    },
+    {
+      "epoch": 19.163669064748202,
+      "step": 21310,
+      "torque_loss": 0.16104266047477722
+    },
+    {
+      "epoch": 19.163669064748202,
+      "force_loss": 0.006079719867557287,
+      "step": 21310
+    },
+    {
+      "epoch": 19.172661870503596,
+      "grad_norm": 0.2695898413658142,
+      "learning_rate": 7.660727038730981e-05,
+      "loss": 0.0219,
+      "step": 21320
+    },
+    {
+      "action_loss": 0.022043129429221153,
+      "epoch": 19.172661870503596,
+      "step": 21320
+    },
+    {
+      "epoch": 19.172661870503596,
+      "step": 21320,
+      "torque_loss": 0.15773947536945343
+    },
+    {
+      "epoch": 19.172661870503596,
+      "force_loss": 0.0237929355353117,
+      "step": 21320
+    },
+    {
+      "epoch": 19.181654676258994,
+      "grad_norm": 0.3807615041732788,
+      "learning_rate": 7.65839344281141e-05,
+      "loss": 0.0237,
+      "step": 21330
+    },
+    {
+      "action_loss": 0.005357636604458094,
+      "epoch": 19.181654676258994,
+      "step": 21330
+    },
+    {
+      "epoch": 19.181654676258994,
+      "step": 21330,
+      "torque_loss": 0.13021297752857208
+    },
+    {
+      "epoch": 19.181654676258994,
+      "force_loss": 0.008785915561020374,
+      "step": 21330
+    },
+    {
+      "epoch": 19.190647482014388,
+      "grad_norm": 0.28463417291641235,
+      "learning_rate": 7.656059039342101e-05,
+      "loss": 0.0189,
+      "step": 21340
+    },
+    {
+      "action_loss": 0.009215909987688065,
+      "epoch": 19.190647482014388,
+      "step": 21340
+    },
+    {
+      "epoch": 19.190647482014388,
+      "step": 21340,
+      "torque_loss": 0.11108382791280746
+    },
+    {
+      "epoch": 19.190647482014388,
+      "force_loss": 0.011105313897132874,
+      "step": 21340
+    },
+    {
+      "epoch": 19.199640287769785,
+      "grad_norm": 0.33906736969947815,
+      "learning_rate": 7.653723829032187e-05,
+      "loss": 0.0216,
+      "step": 21350
+    },
+    {
+      "action_loss": 0.00422265287488699,
+      "epoch": 19.199640287769785,
+      "step": 21350
+    },
+    {
+      "epoch": 19.199640287769785,
+      "step": 21350,
+      "torque_loss": 0.09489398449659348
+    },
+    {
+      "epoch": 19.199640287769785,
+      "force_loss": 0.00896951463073492,
+      "step": 21350
+    },
+    {
+      "epoch": 19.20863309352518,
+      "grad_norm": 0.3440510630607605,
+      "learning_rate": 7.65138781259104e-05,
+      "loss": 0.0258,
+      "step": 21360
+    },
+    {
+      "action_loss": 0.003073754021897912,
+      "epoch": 19.20863309352518,
+      "step": 21360
+    },
+    {
+      "epoch": 19.20863309352518,
+      "step": 21360,
+      "torque_loss": 0.1115085706114769
+    },
+    {
+      "epoch": 19.20863309352518,
+      "force_loss": 0.003067217767238617,
+      "step": 21360
+    },
+    {
+      "epoch": 19.217625899280577,
+      "grad_norm": 0.3840801417827606,
+      "learning_rate": 7.649050990728279e-05,
+      "loss": 0.0183,
+      "step": 21370
+    },
+    {
+      "action_loss": 0.009511918760836124,
+      "epoch": 19.217625899280577,
+      "step": 21370
+    },
+    {
+      "epoch": 19.217625899280577,
+      "step": 21370,
+      "torque_loss": 0.1170787587761879
+    },
+    {
+      "epoch": 19.217625899280577,
+      "force_loss": 0.007953859865665436,
+      "step": 21370
+    },
+    {
+      "epoch": 19.22661870503597,
+      "grad_norm": 0.4349346458911896,
+      "learning_rate": 7.646713364153774e-05,
+      "loss": 0.0224,
+      "step": 21380
+    },
+    {
+      "action_loss": 0.00421341834589839,
+      "epoch": 19.22661870503597,
+      "step": 21380
+    },
+    {
+      "epoch": 19.22661870503597,
+      "step": 21380,
+      "torque_loss": 0.1454629749059677
+    },
+    {
+      "epoch": 19.22661870503597,
+      "force_loss": 0.0028723319992423058,
+      "step": 21380
+    },
+    {
+      "epoch": 19.235611510791365,
+      "grad_norm": 0.29153531789779663,
+      "learning_rate": 7.64437493357763e-05,
+      "loss": 0.0216,
+      "step": 21390
+    },
+    {
+      "action_loss": 0.005913767963647842,
+      "epoch": 19.235611510791365,
+      "step": 21390
+    },
+    {
+      "epoch": 19.235611510791365,
+      "step": 21390,
+      "torque_loss": 0.10143619775772095
+    },
+    {
+      "epoch": 19.235611510791365,
+      "force_loss": 0.006469981744885445,
+      "step": 21390
+    },
+    {
+      "epoch": 19.244604316546763,
+      "grad_norm": 0.21742607653141022,
+      "learning_rate": 7.642035699710202e-05,
+      "loss": 0.0218,
+      "step": 21400
+    },
+    {
+      "action_loss": 0.00501607172191143,
+      "epoch": 19.244604316546763,
+      "step": 21400
+    },
+    {
+      "epoch": 19.244604316546763,
+      "step": 21400,
+      "torque_loss": 0.14102411270141602
+    },
+    {
+      "epoch": 19.244604316546763,
+      "force_loss": 0.0037264544516801834,
+      "step": 21400
+    },
+    {
+      "epoch": 19.253597122302157,
+      "grad_norm": 0.39896100759506226,
+      "learning_rate": 7.639695663262089e-05,
+      "loss": 0.0197,
+      "step": 21410
+    },
+    {
+      "action_loss": 0.012978407554328442,
+      "epoch": 19.253597122302157,
+      "step": 21410
+    },
+    {
+      "epoch": 19.253597122302157,
+      "step": 21410,
+      "torque_loss": 0.14753873646259308
+    },
+    {
+      "epoch": 19.253597122302157,
+      "force_loss": 0.01547118928283453,
+      "step": 21410
+    },
+    {
+      "epoch": 19.262589928057555,
+      "grad_norm": 0.442873477935791,
+      "learning_rate": 7.637354824944128e-05,
+      "loss": 0.0293,
+      "step": 21420
+    },
+    {
+      "action_loss": 0.002996989758685231,
+      "epoch": 19.262589928057555,
+      "step": 21420
+    },
+    {
+      "epoch": 19.262589928057555,
+      "step": 21420,
+      "torque_loss": 0.12058514356613159
+    },
+    {
+      "epoch": 19.262589928057555,
+      "force_loss": 0.002383910119533539,
+      "step": 21420
+    },
+    {
+      "epoch": 19.27158273381295,
+      "grad_norm": 0.4413894712924957,
+      "learning_rate": 7.635013185467408e-05,
+      "loss": 0.018,
+      "step": 21430
+    },
+    {
+      "action_loss": 0.005001666955649853,
+      "epoch": 19.27158273381295,
+      "step": 21430
+    },
+    {
+      "epoch": 19.27158273381295,
+      "step": 21430,
+      "torque_loss": 0.12552034854888916
+    },
+    {
+      "epoch": 19.27158273381295,
+      "force_loss": 0.003579702228307724,
+      "step": 21430
+    },
+    {
+      "epoch": 19.280575539568346,
+      "grad_norm": 0.19401530921459198,
+      "learning_rate": 7.632670745543256e-05,
+      "loss": 0.0202,
+      "step": 21440
+    },
+    {
+      "action_loss": 0.0051520224660634995,
+      "epoch": 19.280575539568346,
+      "step": 21440
+    },
+    {
+      "epoch": 19.280575539568346,
+      "step": 21440,
+      "torque_loss": 0.10939887166023254
+    },
+    {
+      "epoch": 19.280575539568346,
+      "force_loss": 0.0027365863788872957,
+      "step": 21440
+    },
+    {
+      "epoch": 19.28956834532374,
+      "grad_norm": 0.3055570125579834,
+      "learning_rate": 7.630327505883242e-05,
+      "loss": 0.0233,
+      "step": 21450
+    },
+    {
+      "action_loss": 0.0033240392804145813,
+      "epoch": 19.28956834532374,
+      "step": 21450
+    },
+    {
+      "epoch": 19.28956834532374,
+      "step": 21450,
+      "torque_loss": 0.13942258059978485
+    },
+    {
+      "epoch": 19.28956834532374,
+      "force_loss": 0.002692107344046235,
+      "step": 21450
+    },
+    {
+      "epoch": 19.298561151079138,
+      "grad_norm": 0.2008047252893448,
+      "learning_rate": 7.627983467199182e-05,
+      "loss": 0.019,
+      "step": 21460
+    },
+    {
+      "action_loss": 0.006855081766843796,
+      "epoch": 19.298561151079138,
+      "step": 21460
+    },
+    {
+      "epoch": 19.298561151079138,
+      "step": 21460,
+      "torque_loss": 0.21656499803066254
+    },
+    {
+      "epoch": 19.298561151079138,
+      "force_loss": 0.007732841651886702,
+      "step": 21460
+    },
+    {
+      "epoch": 19.307553956834532,
+      "grad_norm": 1.2336971759796143,
+      "learning_rate": 7.625638630203132e-05,
+      "loss": 0.0261,
+      "step": 21470
+    },
+    {
+      "action_loss": 0.005216373596340418,
+      "epoch": 19.307553956834532,
+      "step": 21470
+    },
+    {
+      "epoch": 19.307553956834532,
+      "step": 21470,
+      "torque_loss": 0.1514403074979782
+    },
+    {
+      "epoch": 19.307553956834532,
+      "force_loss": 0.0029113886412233114,
+      "step": 21470
+    },
+    {
+      "epoch": 19.31654676258993,
+      "grad_norm": 0.30147963762283325,
+      "learning_rate": 7.623292995607394e-05,
+      "loss": 0.021,
+      "step": 21480
+    },
+    {
+      "action_loss": 0.0085029611364007,
+      "epoch": 19.31654676258993,
+      "step": 21480
+    },
+    {
+      "epoch": 19.31654676258993,
+      "step": 21480,
+      "torque_loss": 0.17489902675151825
+    },
+    {
+      "epoch": 19.31654676258993,
+      "force_loss": 0.006882034242153168,
+      "step": 21480
+    },
+    {
+      "epoch": 19.325539568345324,
+      "grad_norm": 0.4513334035873413,
+      "learning_rate": 7.620946564124507e-05,
+      "loss": 0.0217,
+      "step": 21490
+    },
+    {
+      "action_loss": 0.012447998858988285,
+      "epoch": 19.325539568345324,
+      "step": 21490
+    },
+    {
+      "epoch": 19.325539568345324,
+      "step": 21490,
+      "torque_loss": 0.14656271040439606
+    },
+    {
+      "epoch": 19.325539568345324,
+      "force_loss": 0.013763678260147572,
+      "step": 21490
+    },
+    {
+      "epoch": 19.334532374100718,
+      "grad_norm": 0.3689354956150055,
+      "learning_rate": 7.618599336467256e-05,
+      "loss": 0.0218,
+      "step": 21500
+    },
+    {
+      "action_loss": 0.005984982941299677,
+      "epoch": 19.334532374100718,
+      "step": 21500
+    },
+    {
+      "epoch": 19.334532374100718,
+      "step": 21500,
+      "torque_loss": 0.11968144029378891
+    },
+    {
+      "epoch": 19.334532374100718,
+      "force_loss": 0.007563177030533552,
+      "step": 21500
+    },
+    {
+      "epoch": 19.343525179856115,
+      "grad_norm": 0.33535292744636536,
+      "learning_rate": 7.616251313348666e-05,
+      "loss": 0.0194,
+      "step": 21510
+    },
+    {
+      "action_loss": 0.004599094856530428,
+      "epoch": 19.343525179856115,
+      "step": 21510
+    },
+    {
+      "epoch": 19.343525179856115,
+      "step": 21510,
+      "torque_loss": 0.14619536697864532
+    },
+    {
+      "epoch": 19.343525179856115,
+      "force_loss": 0.009815558791160583,
+      "step": 21510
+    },
+    {
+      "epoch": 19.35251798561151,
+      "grad_norm": 0.7966991066932678,
+      "learning_rate": 7.613902495482005e-05,
+      "loss": 0.025,
+      "step": 21520
+    },
+    {
+      "action_loss": 0.00572822755202651,
+      "epoch": 19.35251798561151,
+      "step": 21520
+    },
+    {
+      "epoch": 19.35251798561151,
+      "step": 21520,
+      "torque_loss": 0.12078839540481567
+    },
+    {
+      "epoch": 19.35251798561151,
+      "force_loss": 0.0027878370601683855,
+      "step": 21520
+    },
+    {
+      "epoch": 19.361510791366907,
+      "grad_norm": 0.34956786036491394,
+      "learning_rate": 7.611552883580784e-05,
+      "loss": 0.0193,
+      "step": 21530
+    },
+    {
+      "action_loss": 0.009181947447359562,
+      "epoch": 19.361510791366907,
+      "step": 21530
+    },
+    {
+      "epoch": 19.361510791366907,
+      "step": 21530,
+      "torque_loss": 0.15982826054096222
+    },
+    {
+      "epoch": 19.361510791366907,
+      "force_loss": 0.00671034911647439,
+      "step": 21530
+    },
+    {
+      "epoch": 19.3705035971223,
+      "grad_norm": 0.2988636791706085,
+      "learning_rate": 7.609202478358748e-05,
+      "loss": 0.022,
+      "step": 21540
+    },
+    {
+      "action_loss": 0.009565108455717564,
+      "epoch": 19.3705035971223,
+      "step": 21540
+    },
+    {
+      "epoch": 19.3705035971223,
+      "step": 21540,
+      "torque_loss": 0.16315078735351562
+    },
+    {
+      "epoch": 19.3705035971223,
+      "force_loss": 0.008048557676374912,
+      "step": 21540
+    },
+    {
+      "epoch": 19.3794964028777,
+      "grad_norm": 0.22613811492919922,
+      "learning_rate": 7.606851280529895e-05,
+      "loss": 0.0221,
+      "step": 21550
+    },
+    {
+      "action_loss": 0.006762331817299128,
+      "epoch": 19.3794964028777,
+      "step": 21550
+    },
+    {
+      "epoch": 19.3794964028777,
+      "step": 21550,
+      "torque_loss": 0.13161323964595795
+    },
+    {
+      "epoch": 19.3794964028777,
+      "force_loss": 0.006115918513387442,
+      "step": 21550
+    },
+    {
+      "epoch": 19.388489208633093,
+      "grad_norm": 0.24901631474494934,
+      "learning_rate": 7.604499290808449e-05,
+      "loss": 0.0217,
+      "step": 21560
+    },
+    {
+      "action_loss": 0.011465758085250854,
+      "epoch": 19.388489208633093,
+      "step": 21560
+    },
+    {
+      "epoch": 19.388489208633093,
+      "step": 21560,
+      "torque_loss": 0.10021913796663284
+    },
+    {
+      "epoch": 19.388489208633093,
+      "force_loss": 0.005063768941909075,
+      "step": 21560
+    },
+    {
+      "epoch": 19.39748201438849,
+      "grad_norm": 0.32080399990081787,
+      "learning_rate": 7.602146509908888e-05,
+      "loss": 0.0195,
+      "step": 21570
+    },
+    {
+      "action_loss": 0.0021028409246355295,
+      "epoch": 19.39748201438849,
+      "step": 21570
+    },
+    {
+      "epoch": 19.39748201438849,
+      "step": 21570,
+      "torque_loss": 0.08726945519447327
+    },
+    {
+      "epoch": 19.39748201438849,
+      "force_loss": 0.001828385517001152,
+      "step": 21570
+    },
+    {
+      "epoch": 19.406474820143885,
+      "grad_norm": 0.5371273159980774,
+      "learning_rate": 7.599792938545921e-05,
+      "loss": 0.027,
+      "step": 21580
+    },
+    {
+      "action_loss": 0.007085882127285004,
+      "epoch": 19.406474820143885,
+      "step": 21580
+    },
+    {
+      "epoch": 19.406474820143885,
+      "step": 21580,
+      "torque_loss": 0.09476127475500107
+    },
+    {
+      "epoch": 19.406474820143885,
+      "force_loss": 0.004616372287273407,
+      "step": 21580
+    },
+    {
+      "epoch": 19.415467625899282,
+      "grad_norm": 0.33448728919029236,
+      "learning_rate": 7.597438577434506e-05,
+      "loss": 0.0226,
+      "step": 21590
+    },
+    {
+      "action_loss": 0.0068463473580777645,
+      "epoch": 19.415467625899282,
+      "step": 21590
+    },
+    {
+      "epoch": 19.415467625899282,
+      "step": 21590,
+      "torque_loss": 0.15716011822223663
+    },
+    {
+      "epoch": 19.415467625899282,
+      "force_loss": 0.008225476369261742,
+      "step": 21590
+    },
+    {
+      "epoch": 19.424460431654676,
+      "grad_norm": 0.36186304688453674,
+      "learning_rate": 7.595083427289831e-05,
+      "loss": 0.0212,
+      "step": 21600
+    },
+    {
+      "action_loss": 0.0033750608563423157,
+      "epoch": 19.424460431654676,
+      "step": 21600
+    },
+    {
+      "epoch": 19.424460431654676,
+      "step": 21600,
+      "torque_loss": 0.1819905787706375
+    },
+    {
+      "epoch": 19.424460431654676,
+      "force_loss": 0.004477418027818203,
+      "step": 21600
+    },
+    {
+      "epoch": 19.43345323741007,
+      "grad_norm": 0.19142687320709229,
+      "learning_rate": 7.59272748882733e-05,
+      "loss": 0.0191,
+      "step": 21610
+    },
+    {
+      "action_loss": 0.022279134020209312,
+      "epoch": 19.43345323741007,
+      "step": 21610
+    },
+    {
+      "epoch": 19.43345323741007,
+      "step": 21610,
+      "torque_loss": 0.18531103432178497
+    },
+    {
+      "epoch": 19.43345323741007,
+      "force_loss": 0.025092462077736855,
+      "step": 21610
+    },
+    {
+      "epoch": 19.442446043165468,
+      "grad_norm": 0.20963843166828156,
+      "learning_rate": 7.590370762762675e-05,
+      "loss": 0.0218,
+      "step": 21620
+    },
+    {
+      "action_loss": 0.003140711458399892,
+      "epoch": 19.442446043165468,
+      "step": 21620
+    },
+    {
+      "epoch": 19.442446043165468,
+      "step": 21620,
+      "torque_loss": 0.1628819704055786
+    },
+    {
+      "epoch": 19.442446043165468,
+      "force_loss": 0.002686604857444763,
+      "step": 21620
+    },
+    {
+      "epoch": 19.451438848920862,
+      "grad_norm": 0.5249615907669067,
+      "learning_rate": 7.588013249811777e-05,
+      "loss": 0.0273,
+      "step": 21630
+    },
+    {
+      "action_loss": 0.006182766053825617,
+      "epoch": 19.451438848920862,
+      "step": 21630
+    },
+    {
+      "epoch": 19.451438848920862,
+      "step": 21630,
+      "torque_loss": 0.13689465820789337
+    },
+    {
+      "epoch": 19.451438848920862,
+      "force_loss": 0.004190466832369566,
+      "step": 21630
+    },
+    {
+      "epoch": 19.46043165467626,
+      "grad_norm": 0.2777141034603119,
+      "learning_rate": 7.585654950690786e-05,
+      "loss": 0.0217,
+      "step": 21640
+    },
+    {
+      "action_loss": 0.006458205636590719,
+      "epoch": 19.46043165467626,
+      "step": 21640
+    },
+    {
+      "epoch": 19.46043165467626,
+      "step": 21640,
+      "torque_loss": 0.13719987869262695
+    },
+    {
+      "epoch": 19.46043165467626,
+      "force_loss": 0.00828318577259779,
+      "step": 21640
+    },
+    {
+      "epoch": 19.469424460431654,
+      "grad_norm": 0.45955348014831543,
+      "learning_rate": 7.583295866116091e-05,
+      "loss": 0.0236,
+      "step": 21650
+    },
+    {
+      "action_loss": 0.007151912897825241,
+      "epoch": 19.469424460431654,
+      "step": 21650
+    },
+    {
+      "epoch": 19.469424460431654,
+      "step": 21650,
+      "torque_loss": 0.1097402349114418
+    },
+    {
+      "epoch": 19.469424460431654,
+      "force_loss": 0.004806819837540388,
+      "step": 21650
+    },
+    {
+      "epoch": 19.47841726618705,
+      "grad_norm": 0.2692463994026184,
+      "learning_rate": 7.580935996804321e-05,
+      "loss": 0.0233,
+      "step": 21660
+    },
+    {
+      "action_loss": 0.018180709332227707,
+      "epoch": 19.47841726618705,
+      "step": 21660
+    },
+    {
+      "epoch": 19.47841726618705,
+      "step": 21660,
+      "torque_loss": 0.14763160049915314
+    },
+    {
+      "epoch": 19.47841726618705,
+      "force_loss": 0.014515109360218048,
+      "step": 21660
+    },
+    {
+      "epoch": 19.487410071942445,
+      "grad_norm": 0.22884929180145264,
+      "learning_rate": 7.57857534347234e-05,
+      "loss": 0.0232,
+      "step": 21670
+    },
+    {
+      "action_loss": 0.004157568793743849,
+      "epoch": 19.487410071942445,
+      "step": 21670
+    },
+    {
+      "epoch": 19.487410071942445,
+      "step": 21670,
+      "torque_loss": 0.12420075386762619
+    },
+    {
+      "epoch": 19.487410071942445,
+      "force_loss": 0.0033161770552396774,
+      "step": 21670
+    },
+    {
+      "epoch": 19.496402877697843,
+      "grad_norm": 0.4245760142803192,
+      "learning_rate": 7.576213906837254e-05,
+      "loss": 0.0211,
+      "step": 21680
+    },
+    {
+      "action_loss": 0.020072678104043007,
+      "epoch": 19.496402877697843,
+      "step": 21680
+    },
+    {
+      "epoch": 19.496402877697843,
+      "step": 21680,
+      "torque_loss": 0.18354733288288116
+    },
+    {
+      "epoch": 19.496402877697843,
+      "force_loss": 0.019310100004076958,
+      "step": 21680
+    },
+    {
+      "epoch": 19.505395683453237,
+      "grad_norm": 0.2859075665473938,
+      "learning_rate": 7.573851687616403e-05,
+      "loss": 0.0231,
+      "step": 21690
+    },
+    {
+      "action_loss": 0.008961736224591732,
+      "epoch": 19.505395683453237,
+      "step": 21690
+    },
+    {
+      "epoch": 19.505395683453237,
+      "step": 21690,
+      "torque_loss": 0.14467954635620117
+    },
+    {
+      "epoch": 19.505395683453237,
+      "force_loss": 0.014233012683689594,
+      "step": 21690
+    },
+    {
+      "epoch": 19.514388489208635,
+      "grad_norm": 0.22410668432712555,
+      "learning_rate": 7.571488686527368e-05,
+      "loss": 0.019,
+      "step": 21700
+    },
+    {
+      "action_loss": 0.005172547418624163,
+      "epoch": 19.514388489208635,
+      "step": 21700
+    },
+    {
+      "epoch": 19.514388489208635,
+      "step": 21700,
+      "torque_loss": 0.13156788051128387
+    },
+    {
+      "epoch": 19.514388489208635,
+      "force_loss": 0.009319930337369442,
+      "step": 21700
+    },
+    {
+      "epoch": 19.52338129496403,
+      "grad_norm": 0.37554094195365906,
+      "learning_rate": 7.569124904287968e-05,
+      "loss": 0.0195,
+      "step": 21710
+    },
+    {
+      "action_loss": 0.009185795672237873,
+      "epoch": 19.52338129496403,
+      "step": 21710
+    },
+    {
+      "epoch": 19.52338129496403,
+      "step": 21710,
+      "torque_loss": 0.15854012966156006
+    },
+    {
+      "epoch": 19.52338129496403,
+      "force_loss": 0.009237014688551426,
+      "step": 21710
+    },
+    {
+      "epoch": 19.532374100719423,
+      "grad_norm": 0.4086345136165619,
+      "learning_rate": 7.566760341616254e-05,
+      "loss": 0.0225,
+      "step": 21720
+    },
+    {
+      "action_loss": 0.0032329161185771227,
+      "epoch": 19.532374100719423,
+      "step": 21720
+    },
+    {
+      "epoch": 19.532374100719423,
+      "step": 21720,
+      "torque_loss": 0.12092572450637817
+    },
+    {
+      "epoch": 19.532374100719423,
+      "force_loss": 0.002039354993030429,
+      "step": 21720
+    },
+    {
+      "epoch": 19.54136690647482,
+      "grad_norm": 0.3729719817638397,
+      "learning_rate": 7.564394999230519e-05,
+      "loss": 0.0246,
+      "step": 21730
+    },
+    {
+      "action_loss": 0.0032205404713749886,
+      "epoch": 19.54136690647482,
+      "step": 21730
+    },
+    {
+      "epoch": 19.54136690647482,
+      "step": 21730,
+      "torque_loss": 0.11641553789377213
+    },
+    {
+      "epoch": 19.54136690647482,
+      "force_loss": 0.002129241591319442,
+      "step": 21730
+    },
+    {
+      "epoch": 19.550359712230215,
+      "grad_norm": 0.2299564927816391,
+      "learning_rate": 7.562028877849294e-05,
+      "loss": 0.0183,
+      "step": 21740
+    },
+    {
+      "action_loss": 0.013398644514381886,
+      "epoch": 19.550359712230215,
+      "step": 21740
+    },
+    {
+      "epoch": 19.550359712230215,
+      "step": 21740,
+      "torque_loss": 0.1278691589832306
+    },
+    {
+      "epoch": 19.550359712230215,
+      "force_loss": 0.011579960584640503,
+      "step": 21740
+    },
+    {
+      "epoch": 19.559352517985612,
+      "grad_norm": 0.3825627267360687,
+      "learning_rate": 7.559661978191341e-05,
+      "loss": 0.0215,
+      "step": 21750
+    },
+    {
+      "action_loss": 0.011062725447118282,
+      "epoch": 19.559352517985612,
+      "step": 21750
+    },
+    {
+      "epoch": 19.559352517985612,
+      "step": 21750,
+      "torque_loss": 0.09550317376852036
+    },
+    {
+      "epoch": 19.559352517985612,
+      "force_loss": 0.007611427456140518,
+      "step": 21750
+    },
+    {
+      "epoch": 19.568345323741006,
+      "grad_norm": 0.3207350969314575,
+      "learning_rate": 7.557294300975664e-05,
+      "loss": 0.0206,
+      "step": 21760
+    },
+    {
+      "action_loss": 0.003980598878115416,
+      "epoch": 19.568345323741006,
+      "step": 21760
+    },
+    {
+      "epoch": 19.568345323741006,
+      "step": 21760,
+      "torque_loss": 0.11463906615972519
+    },
+    {
+      "epoch": 19.568345323741006,
+      "force_loss": 0.0025576932821422815,
+      "step": 21760
+    },
+    {
+      "epoch": 19.577338129496404,
+      "grad_norm": 0.4431993067264557,
+      "learning_rate": 7.554925846921499e-05,
+      "loss": 0.0205,
+      "step": 21770
+    },
+    {
+      "action_loss": 0.0061083766631782055,
+      "epoch": 19.577338129496404,
+      "step": 21770
+    },
+    {
+      "epoch": 19.577338129496404,
+      "step": 21770,
+      "torque_loss": 0.12348593026399612
+    },
+    {
+      "epoch": 19.577338129496404,
+      "force_loss": 0.003276041941717267,
+      "step": 21770
+    },
+    {
+      "epoch": 19.586330935251798,
+      "grad_norm": 0.4156457781791687,
+      "learning_rate": 7.552556616748321e-05,
+      "loss": 0.0208,
+      "step": 21780
+    },
+    {
+      "action_loss": 0.008796994574368,
+      "epoch": 19.586330935251798,
+      "step": 21780
+    },
+    {
+      "epoch": 19.586330935251798,
+      "step": 21780,
+      "torque_loss": 0.13290178775787354
+    },
+    {
+      "epoch": 19.586330935251798,
+      "force_loss": 0.012139275670051575,
+      "step": 21780
+    },
+    {
+      "epoch": 19.595323741007196,
+      "grad_norm": 0.28027454018592834,
+      "learning_rate": 7.550186611175838e-05,
+      "loss": 0.0236,
+      "step": 21790
+    },
+    {
+      "action_loss": 0.003140410641208291,
+      "epoch": 19.595323741007196,
+      "step": 21790
+    },
+    {
+      "epoch": 19.595323741007196,
+      "step": 21790,
+      "torque_loss": 0.12866981327533722
+    },
+    {
+      "epoch": 19.595323741007196,
+      "force_loss": 0.0019691220950335264,
+      "step": 21790
+    },
+    {
+      "epoch": 19.60431654676259,
+      "grad_norm": 0.3398789167404175,
+      "learning_rate": 7.547815830923998e-05,
+      "loss": 0.0187,
+      "step": 21800
+    },
+    {
+      "action_loss": 0.008160281926393509,
+      "epoch": 19.60431654676259,
+      "step": 21800
+    },
+    {
+      "epoch": 19.60431654676259,
+      "step": 21800,
+      "torque_loss": 0.13390092551708221
+    },
+    {
+      "epoch": 19.60431654676259,
+      "force_loss": 0.004792905878275633,
+      "step": 21800
+    },
+    {
+      "epoch": 19.613309352517987,
+      "grad_norm": 0.43377819657325745,
+      "learning_rate": 7.54544427671298e-05,
+      "loss": 0.0229,
+      "step": 21810
+    },
+    {
+      "action_loss": 0.011805598624050617,
+      "epoch": 19.613309352517987,
+      "step": 21810
+    },
+    {
+      "epoch": 19.613309352517987,
+      "step": 21810,
+      "torque_loss": 0.14656482636928558
+    },
+    {
+      "epoch": 19.613309352517987,
+      "force_loss": 0.013147692196071148,
+      "step": 21810
+    },
+    {
+      "epoch": 19.62230215827338,
+      "grad_norm": 0.4121027886867523,
+      "learning_rate": 7.543071949263198e-05,
+      "loss": 0.02,
+      "step": 21820
+    },
+    {
+      "action_loss": 0.0031295297667384148,
+      "epoch": 19.62230215827338,
+      "step": 21820
+    },
+    {
+      "epoch": 19.62230215827338,
+      "step": 21820,
+      "torque_loss": 0.11127686500549316
+    },
+    {
+      "epoch": 19.62230215827338,
+      "force_loss": 0.002735495800152421,
+      "step": 21820
+    },
+    {
+      "epoch": 19.631294964028775,
+      "grad_norm": 0.331376850605011,
+      "learning_rate": 7.540698849295305e-05,
+      "loss": 0.0178,
+      "step": 21830
+    },
+    {
+      "action_loss": 0.006875521969050169,
+      "epoch": 19.631294964028775,
+      "step": 21830
+    },
+    {
+      "epoch": 19.631294964028775,
+      "step": 21830,
+      "torque_loss": 0.1267344206571579
+    },
+    {
+      "epoch": 19.631294964028775,
+      "force_loss": 0.007150923367589712,
+      "step": 21830
+    },
+    {
+      "epoch": 19.640287769784173,
+      "grad_norm": 0.20037895441055298,
+      "learning_rate": 7.538324977530183e-05,
+      "loss": 0.0203,
+      "step": 21840
+    },
+    {
+      "action_loss": 0.021925918757915497,
+      "epoch": 19.640287769784173,
+      "step": 21840
+    },
+    {
+      "epoch": 19.640287769784173,
+      "step": 21840,
+      "torque_loss": 0.20452392101287842
+    },
+    {
+      "epoch": 19.640287769784173,
+      "force_loss": 0.03243473544716835,
+      "step": 21840
+    },
+    {
+      "epoch": 19.649280575539567,
+      "grad_norm": 0.3448210656642914,
+      "learning_rate": 7.535950334688955e-05,
+      "loss": 0.0243,
+      "step": 21850
+    },
+    {
+      "action_loss": 0.00136956840287894,
+      "epoch": 19.649280575539567,
+      "step": 21850
+    },
+    {
+      "epoch": 19.649280575539567,
+      "step": 21850,
+      "torque_loss": 0.07476025819778442
+    },
+    {
+      "epoch": 19.649280575539567,
+      "force_loss": 0.001339944894425571,
+      "step": 21850
+    },
+    {
+      "epoch": 19.658273381294965,
+      "grad_norm": 0.18949294090270996,
+      "learning_rate": 7.533574921492972e-05,
+      "loss": 0.0214,
+      "step": 21860
+    },
+    {
+      "action_loss": 0.0027416015509516,
+      "epoch": 19.658273381294965,
+      "step": 21860
+    },
+    {
+      "epoch": 19.658273381294965,
+      "step": 21860,
+      "torque_loss": 0.11827247589826584
+    },
+    {
+      "epoch": 19.658273381294965,
+      "force_loss": 0.002873457269743085,
+      "step": 21860
+    },
+    {
+      "epoch": 19.66726618705036,
+      "grad_norm": 0.20143093168735504,
+      "learning_rate": 7.531198738663824e-05,
+      "loss": 0.0191,
+      "step": 21870
+    },
+    {
+      "action_loss": 0.005855960305780172,
+      "epoch": 19.66726618705036,
+      "step": 21870
+    },
+    {
+      "epoch": 19.66726618705036,
+      "step": 21870,
+      "torque_loss": 0.17841953039169312
+    },
+    {
+      "epoch": 19.66726618705036,
+      "force_loss": 0.006000491324812174,
+      "step": 21870
+    },
+    {
+      "epoch": 19.676258992805757,
+      "grad_norm": 0.20106497406959534,
+      "learning_rate": 7.528821786923333e-05,
+      "loss": 0.0204,
+      "step": 21880
+    },
+    {
+      "action_loss": 0.008288371376693249,
+      "epoch": 19.676258992805757,
+      "step": 21880
+    },
+    {
+      "epoch": 19.676258992805757,
+      "step": 21880,
+      "torque_loss": 0.13893398642539978
+    },
+    {
+      "epoch": 19.676258992805757,
+      "force_loss": 0.004750822205096483,
+      "step": 21880
+    },
+    {
+      "epoch": 19.68525179856115,
+      "grad_norm": 0.3063576817512512,
+      "learning_rate": 7.52644406699355e-05,
+      "loss": 0.0227,
+      "step": 21890
+    },
+    {
+      "action_loss": 0.005023397039622068,
+      "epoch": 19.68525179856115,
+      "step": 21890
+    },
+    {
+      "epoch": 19.68525179856115,
+      "step": 21890,
+      "torque_loss": 0.15836508572101593
+    },
+    {
+      "epoch": 19.68525179856115,
+      "force_loss": 0.0050356886349618435,
+      "step": 21890
+    },
+    {
+      "epoch": 19.694244604316548,
+      "grad_norm": 0.20005039870738983,
+      "learning_rate": 7.524065579596766e-05,
+      "loss": 0.0205,
+      "step": 21900
+    },
+    {
+      "action_loss": 0.003932314459234476,
+      "epoch": 19.694244604316548,
+      "step": 21900
+    },
+    {
+      "epoch": 19.694244604316548,
+      "step": 21900,
+      "torque_loss": 0.12649227678775787
+    },
+    {
+      "epoch": 19.694244604316548,
+      "force_loss": 0.0038639113772660494,
+      "step": 21900
+    },
+    {
+      "epoch": 19.703237410071942,
+      "grad_norm": 0.20406463742256165,
+      "learning_rate": 7.521686325455506e-05,
+      "loss": 0.0191,
+      "step": 21910
+    },
+    {
+      "action_loss": 0.007570803165435791,
+      "epoch": 19.703237410071942,
+      "step": 21910
+    },
+    {
+      "epoch": 19.703237410071942,
+      "step": 21910,
+      "torque_loss": 0.16327345371246338
+    },
+    {
+      "epoch": 19.703237410071942,
+      "force_loss": 0.004229263868182898,
+      "step": 21910
+    },
+    {
+      "epoch": 19.71223021582734,
+      "grad_norm": 0.5840118527412415,
+      "learning_rate": 7.51930630529252e-05,
+      "loss": 0.0217,
+      "step": 21920
+    },
+    {
+      "action_loss": 0.005566589534282684,
+      "epoch": 19.71223021582734,
+      "step": 21920
+    },
+    {
+      "epoch": 19.71223021582734,
+      "step": 21920,
+      "torque_loss": 0.15318508446216583
+    },
+    {
+      "epoch": 19.71223021582734,
+      "force_loss": 0.008140143938362598,
+      "step": 21920
+    },
+    {
+      "epoch": 19.721223021582734,
+      "grad_norm": 0.30791956186294556,
+      "learning_rate": 7.516925519830797e-05,
+      "loss": 0.0196,
+      "step": 21930
+    },
+    {
+      "action_loss": 0.010097134858369827,
+      "epoch": 19.721223021582734,
+      "step": 21930
+    },
+    {
+      "epoch": 19.721223021582734,
+      "step": 21930,
+      "torque_loss": 0.16301332414150238
+    },
+    {
+      "epoch": 19.721223021582734,
+      "force_loss": 0.011774941347539425,
+      "step": 21930
+    },
+    {
+      "epoch": 19.730215827338128,
+      "grad_norm": 0.29583054780960083,
+      "learning_rate": 7.514543969793557e-05,
+      "loss": 0.0213,
+      "step": 21940
+    },
+    {
+      "action_loss": 0.0032733362168073654,
+      "epoch": 19.730215827338128,
+      "step": 21940
+    },
+    {
+      "epoch": 19.730215827338128,
+      "step": 21940,
+      "torque_loss": 0.11166312545537949
+    },
+    {
+      "epoch": 19.730215827338128,
+      "force_loss": 0.0047770277597010136,
+      "step": 21940
+    },
+    {
+      "epoch": 19.739208633093526,
+      "grad_norm": 0.6144985556602478,
+      "learning_rate": 7.512161655904251e-05,
+      "loss": 0.0221,
+      "step": 21950
+    },
+    {
+      "action_loss": 0.005275297444313765,
+      "epoch": 19.739208633093526,
+      "step": 21950
+    },
+    {
+      "epoch": 19.739208633093526,
+      "step": 21950,
+      "torque_loss": 0.13560757040977478
+    },
+    {
+      "epoch": 19.739208633093526,
+      "force_loss": 0.004693909082561731,
+      "step": 21950
+    },
+    {
+      "epoch": 19.74820143884892,
+      "grad_norm": 0.3211604356765747,
+      "learning_rate": 7.509778578886563e-05,
+      "loss": 0.0218,
+      "step": 21960
+    },
+    {
+      "action_loss": 0.003604740835726261,
+      "epoch": 19.74820143884892,
+      "step": 21960
+    },
+    {
+      "epoch": 19.74820143884892,
+      "step": 21960,
+      "torque_loss": 0.11745858192443848
+    },
+    {
+      "epoch": 19.74820143884892,
+      "force_loss": 0.006334047764539719,
+      "step": 21960
+    },
+    {
+      "epoch": 19.757194244604317,
+      "grad_norm": 0.71286940574646,
+      "learning_rate": 7.507394739464412e-05,
+      "loss": 0.0221,
+      "step": 21970
+    },
+    {
+      "action_loss": 0.004232177510857582,
+      "epoch": 19.757194244604317,
+      "step": 21970
+    },
+    {
+      "epoch": 19.757194244604317,
+      "step": 21970,
+      "torque_loss": 0.09373342990875244
+    },
+    {
+      "epoch": 19.757194244604317,
+      "force_loss": 0.006249606609344482,
+      "step": 21970
+    },
+    {
+      "epoch": 19.76618705035971,
+      "grad_norm": 0.7615202069282532,
+      "learning_rate": 7.50501013836194e-05,
+      "loss": 0.0242,
+      "step": 21980
+    },
+    {
+      "action_loss": 0.004300723783671856,
+      "epoch": 19.76618705035971,
+      "step": 21980
+    },
+    {
+      "epoch": 19.76618705035971,
+      "step": 21980,
+      "torque_loss": 0.1120564416050911
+    },
+    {
+      "epoch": 19.76618705035971,
+      "force_loss": 0.004500760231167078,
+      "step": 21980
+    },
+    {
+      "epoch": 19.77517985611511,
+      "grad_norm": 0.5227411389350891,
+      "learning_rate": 7.50262477630353e-05,
+      "loss": 0.025,
+      "step": 21990
+    },
+    {
+      "action_loss": 0.01617811806499958,
+      "epoch": 19.77517985611511,
+      "step": 21990
+    },
+    {
+      "epoch": 19.77517985611511,
+      "step": 21990,
+      "torque_loss": 0.11352749913930893
+    },
+    {
+      "epoch": 19.77517985611511,
+      "force_loss": 0.009329590015113354,
+      "step": 21990
+    },
+    {
+      "epoch": 19.784172661870503,
+      "grad_norm": 0.35520029067993164,
+      "learning_rate": 7.500238654013794e-05,
+      "loss": 0.0248,
+      "step": 22000
+    },
+    {
+      "action_loss": 0.0038139354437589645,
+      "epoch": 19.784172661870503,
+      "step": 22000
+    },
+    {
+      "epoch": 19.784172661870503,
+      "step": 22000,
+      "torque_loss": 0.12948231399059296
+    },
+    {
+      "epoch": 19.784172661870503,
+      "force_loss": 0.0033348749857395887,
+      "step": 22000
+    },
+    {
+      "epoch": 19.7931654676259,
+      "grad_norm": 0.1954745501279831,
+      "learning_rate": 7.497851772217566e-05,
+      "loss": 0.0203,
+      "step": 22010
+    },
+    {
+      "action_loss": 0.005455614533275366,
+      "epoch": 19.7931654676259,
+      "step": 22010
+    },
+    {
+      "epoch": 19.7931654676259,
+      "step": 22010,
+      "torque_loss": 0.15897557139396667
+    },
+    {
+      "epoch": 19.7931654676259,
+      "force_loss": 0.003561886725947261,
+      "step": 22010
+    },
+    {
+      "epoch": 19.802158273381295,
+      "grad_norm": 0.2572297751903534,
+      "learning_rate": 7.495464131639924e-05,
+      "loss": 0.0195,
+      "step": 22020
+    },
+    {
+      "action_loss": 0.007881480269134045,
+      "epoch": 19.802158273381295,
+      "step": 22020
+    },
+    {
+      "epoch": 19.802158273381295,
+      "step": 22020,
+      "torque_loss": 0.0848357081413269
+    },
+    {
+      "epoch": 19.802158273381295,
+      "force_loss": 0.005228835623711348,
+      "step": 22020
+    },
+    {
+      "epoch": 19.81115107913669,
+      "grad_norm": 0.5305858850479126,
+      "learning_rate": 7.493075733006166e-05,
+      "loss": 0.0214,
+      "step": 22030
+    },
+    {
+      "action_loss": 0.0029701683670282364,
+      "epoch": 19.81115107913669,
+      "step": 22030
+    },
+    {
+      "epoch": 19.81115107913669,
+      "step": 22030,
+      "torque_loss": 0.1005556508898735
+    },
+    {
+      "epoch": 19.81115107913669,
+      "force_loss": 0.006433894857764244,
+      "step": 22030
+    },
+    {
+      "epoch": 19.820143884892087,
+      "grad_norm": 0.33326831459999084,
+      "learning_rate": 7.490686577041828e-05,
+      "loss": 0.0205,
+      "step": 22040
+    },
+    {
+      "action_loss": 0.005682650487869978,
+      "epoch": 19.820143884892087,
+      "step": 22040
+    },
+    {
+      "epoch": 19.820143884892087,
+      "step": 22040,
+      "torque_loss": 0.17791688442230225
+    },
+    {
+      "epoch": 19.820143884892087,
+      "force_loss": 0.003731469390913844,
+      "step": 22040
+    },
+    {
+      "epoch": 19.82913669064748,
+      "grad_norm": 0.33978524804115295,
+      "learning_rate": 7.488296664472668e-05,
+      "loss": 0.0282,
+      "step": 22050
+    },
+    {
+      "action_loss": 0.007012306246906519,
+      "epoch": 19.82913669064748,
+      "step": 22050
+    },
+    {
+      "epoch": 19.82913669064748,
+      "step": 22050,
+      "torque_loss": 0.09265895932912827
+    },
+    {
+      "epoch": 19.82913669064748,
+      "force_loss": 0.00769916782155633,
+      "step": 22050
+    },
+    {
+      "epoch": 19.83812949640288,
+      "grad_norm": 0.5150094032287598,
+      "learning_rate": 7.485905996024682e-05,
+      "loss": 0.0228,
+      "step": 22060
+    },
+    {
+      "action_loss": 0.006248244550079107,
+      "epoch": 19.83812949640288,
+      "step": 22060
+    },
+    {
+      "epoch": 19.83812949640288,
+      "step": 22060,
+      "torque_loss": 0.19567523896694183
+    },
+    {
+      "epoch": 19.83812949640288,
+      "force_loss": 0.004354776348918676,
+      "step": 22060
+    },
+    {
+      "epoch": 19.847122302158272,
+      "grad_norm": 0.5206131339073181,
+      "learning_rate": 7.483514572424093e-05,
+      "loss": 0.0199,
+      "step": 22070
+    },
+    {
+      "action_loss": 0.010819584131240845,
+      "epoch": 19.847122302158272,
+      "step": 22070
+    },
+    {
+      "epoch": 19.847122302158272,
+      "step": 22070,
+      "torque_loss": 0.08632185310125351
+    },
+    {
+      "epoch": 19.847122302158272,
+      "force_loss": 0.013305111788213253,
+      "step": 22070
+    },
+    {
+      "epoch": 19.85611510791367,
+      "grad_norm": 0.27759236097335815,
+      "learning_rate": 7.481122394397349e-05,
+      "loss": 0.0197,
+      "step": 22080
+    },
+    {
+      "action_loss": 0.004525866359472275,
+      "epoch": 19.85611510791367,
+      "step": 22080
+    },
+    {
+      "epoch": 19.85611510791367,
+      "step": 22080,
+      "torque_loss": 0.14517642557621002
+    },
+    {
+      "epoch": 19.85611510791367,
+      "force_loss": 0.0021787364967167377,
+      "step": 22080
+    },
+    {
+      "epoch": 19.865107913669064,
+      "grad_norm": 0.1399390548467636,
+      "learning_rate": 7.478729462671131e-05,
+      "loss": 0.0193,
+      "step": 22090
+    },
+    {
+      "action_loss": 0.010387607850134373,
+      "epoch": 19.865107913669064,
+      "step": 22090
+    },
+    {
+      "epoch": 19.865107913669064,
+      "step": 22090,
+      "torque_loss": 0.16290795803070068
+    },
+    {
+      "epoch": 19.865107913669064,
+      "force_loss": 0.011659813113510609,
+      "step": 22090
+    },
+    {
+      "epoch": 19.87410071942446,
+      "grad_norm": 0.36216187477111816,
+      "learning_rate": 7.47633577797235e-05,
+      "loss": 0.0264,
+      "step": 22100
+    },
+    {
+      "action_loss": 0.003574792528524995,
+      "epoch": 19.87410071942446,
+      "step": 22100
+    },
+    {
+      "epoch": 19.87410071942446,
+      "step": 22100,
+      "torque_loss": 0.12696026265621185
+    },
+    {
+      "epoch": 19.87410071942446,
+      "force_loss": 0.003652961691841483,
+      "step": 22100
+    },
+    {
+      "epoch": 19.883093525179856,
+      "grad_norm": 0.46038782596588135,
+      "learning_rate": 7.473941341028144e-05,
+      "loss": 0.0198,
+      "step": 22110
+    },
+    {
+      "action_loss": 0.0029956044163554907,
+      "epoch": 19.883093525179856,
+      "step": 22110
+    },
+    {
+      "epoch": 19.883093525179856,
+      "step": 22110,
+      "torque_loss": 0.14681954681873322
+    },
+    {
+      "epoch": 19.883093525179856,
+      "force_loss": 0.0023136690724641085,
+      "step": 22110
+    },
+    {
+      "epoch": 19.892086330935253,
+      "grad_norm": 0.23139101266860962,
+      "learning_rate": 7.471546152565879e-05,
+      "loss": 0.0189,
+      "step": 22120
+    },
+    {
+      "action_loss": 0.020397908985614777,
+      "epoch": 19.892086330935253,
+      "step": 22120
+    },
+    {
+      "epoch": 19.892086330935253,
+      "step": 22120,
+      "torque_loss": 0.1635989099740982
+    },
+    {
+      "epoch": 19.892086330935253,
+      "force_loss": 0.03053855150938034,
+      "step": 22120
+    },
+    {
+      "epoch": 19.901079136690647,
+      "grad_norm": 0.22430942952632904,
+      "learning_rate": 7.46915021331315e-05,
+      "loss": 0.0258,
+      "step": 22130
+    },
+    {
+      "action_loss": 0.0238180011510849,
+      "epoch": 19.901079136690647,
+      "step": 22130
+    },
+    {
+      "epoch": 19.901079136690647,
+      "step": 22130,
+      "torque_loss": 0.13356070220470428
+    },
+    {
+      "epoch": 19.901079136690647,
+      "force_loss": 0.04048687219619751,
+      "step": 22130
+    },
+    {
+      "epoch": 19.91007194244604,
+      "grad_norm": 0.2783410847187042,
+      "learning_rate": 7.466753523997778e-05,
+      "loss": 0.0242,
+      "step": 22140
+    },
+    {
+      "action_loss": 0.008494014851748943,
+      "epoch": 19.91007194244604,
+      "step": 22140
+    },
+    {
+      "epoch": 19.91007194244604,
+      "step": 22140,
+      "torque_loss": 0.13085371255874634
+    },
+    {
+      "epoch": 19.91007194244604,
+      "force_loss": 0.0026112135965377092,
+      "step": 22140
+    },
+    {
+      "epoch": 19.91906474820144,
+      "grad_norm": 0.4878132939338684,
+      "learning_rate": 7.464356085347819e-05,
+      "loss": 0.0251,
+      "step": 22150
+    },
+    {
+      "action_loss": 0.004474810790270567,
+      "epoch": 19.91906474820144,
+      "step": 22150
+    },
+    {
+      "epoch": 19.91906474820144,
+      "step": 22150,
+      "torque_loss": 0.13291238248348236
+    },
+    {
+      "epoch": 19.91906474820144,
+      "force_loss": 0.005239900201559067,
+      "step": 22150
+    },
+    {
+      "epoch": 19.928057553956833,
+      "grad_norm": 0.3961816430091858,
+      "learning_rate": 7.461957898091548e-05,
+      "loss": 0.0224,
+      "step": 22160
+    },
+    {
+      "action_loss": 0.0038818896282464266,
+      "epoch": 19.928057553956833,
+      "step": 22160
+    },
+    {
+      "epoch": 19.928057553956833,
+      "step": 22160,
+      "torque_loss": 0.12886212766170502
+    },
+    {
+      "epoch": 19.928057553956833,
+      "force_loss": 0.0030095858965069056,
+      "step": 22160
+    },
+    {
+      "epoch": 19.93705035971223,
+      "grad_norm": 0.21624428033828735,
+      "learning_rate": 7.459558962957473e-05,
+      "loss": 0.0223,
+      "step": 22170
+    },
+    {
+      "action_loss": 0.0031913446728140116,
+      "epoch": 19.93705035971223,
+      "step": 22170
+    },
+    {
+      "epoch": 19.93705035971223,
+      "step": 22170,
+      "torque_loss": 0.1169213280081749
+    },
+    {
+      "epoch": 19.93705035971223,
+      "force_loss": 0.0029216452967375517,
+      "step": 22170
+    },
+    {
+      "epoch": 19.946043165467625,
+      "grad_norm": 0.4567829668521881,
+      "learning_rate": 7.457159280674326e-05,
+      "loss": 0.0214,
+      "step": 22180
+    },
+    {
+      "action_loss": 0.006738933268934488,
+      "epoch": 19.946043165467625,
+      "step": 22180
+    },
+    {
+      "epoch": 19.946043165467625,
+      "step": 22180,
+      "torque_loss": 0.0989663302898407
+    },
+    {
+      "epoch": 19.946043165467625,
+      "force_loss": 0.007388899568468332,
+      "step": 22180
+    },
+    {
+      "epoch": 19.955035971223023,
+      "grad_norm": 0.4367087781429291,
+      "learning_rate": 7.454758851971066e-05,
+      "loss": 0.0237,
+      "step": 22190
+    },
+    {
+      "action_loss": 0.004280680324882269,
+      "epoch": 19.955035971223023,
+      "step": 22190
+    },
+    {
+      "epoch": 19.955035971223023,
+      "step": 22190,
+      "torque_loss": 0.1810615062713623
+    },
+    {
+      "epoch": 19.955035971223023,
+      "force_loss": 0.004259243607521057,
+      "step": 22190
+    },
+    {
+      "epoch": 19.964028776978417,
+      "grad_norm": 0.21038062870502472,
+      "learning_rate": 7.45235767757688e-05,
+      "loss": 0.0225,
+      "step": 22200
+    },
+    {
+      "action_loss": 0.0524432547390461,
+      "epoch": 19.964028776978417,
+      "step": 22200
+    },
+    {
+      "epoch": 19.964028776978417,
+      "step": 22200,
+      "torque_loss": 0.20673634111881256
+    },
+    {
+      "epoch": 19.964028776978417,
+      "force_loss": 0.051325708627700806,
+      "step": 22200
+    },
+    {
+      "epoch": 19.973021582733814,
+      "grad_norm": 0.6872680187225342,
+      "learning_rate": 7.449955758221183e-05,
+      "loss": 0.0296,
+      "step": 22210
+    },
+    {
+      "action_loss": 0.010176043026149273,
+      "epoch": 19.973021582733814,
+      "step": 22210
+    },
+    {
+      "epoch": 19.973021582733814,
+      "step": 22210,
+      "torque_loss": 0.16881322860717773
+    },
+    {
+      "epoch": 19.973021582733814,
+      "force_loss": 0.011119500733911991,
+      "step": 22210
+    },
+    {
+      "epoch": 19.98201438848921,
+      "grad_norm": 0.2892797291278839,
+      "learning_rate": 7.447553094633615e-05,
+      "loss": 0.0207,
+      "step": 22220
+    },
+    {
+      "action_loss": 0.010465911589562893,
+      "epoch": 19.98201438848921,
+      "step": 22220
+    },
+    {
+      "epoch": 19.98201438848921,
+      "step": 22220,
+      "torque_loss": 0.11822440475225449
+    },
+    {
+      "epoch": 19.98201438848921,
+      "force_loss": 0.010354467667639256,
+      "step": 22220
+    },
+    {
+      "epoch": 19.991007194244606,
+      "grad_norm": 0.5477696061134338,
+      "learning_rate": 7.445149687544039e-05,
+      "loss": 0.0216,
+      "step": 22230
+    },
+    {
+      "action_loss": 0.009246482513844967,
+      "epoch": 19.991007194244606,
+      "step": 22230
+    },
+    {
+      "epoch": 19.991007194244606,
+      "step": 22230,
+      "torque_loss": 0.1304655373096466
+    },
+    {
+      "epoch": 19.991007194244606,
+      "force_loss": 0.006503938231617212,
+      "step": 22230
+    },
+    {
+      "epoch": 20.0,
+      "grad_norm": 0.24211806058883667,
+      "learning_rate": 7.44274553768255e-05,
+      "loss": 0.0224,
+      "step": 22240
+    },
+    {
+      "action_loss": 0.015477842651307583,
+      "epoch": 20.0,
+      "step": 22240
+    },
+    {
+      "epoch": 20.0,
+      "step": 22240,
+      "torque_loss": 0.13669709861278534
+    },
+    {
+      "epoch": 20.0,
+      "force_loss": 0.019504958763718605,
+      "step": 22240
+    },
+    {
+      "epoch": 20.008992805755394,
+      "grad_norm": 0.35836857557296753,
+      "learning_rate": 7.440340645779464e-05,
+      "loss": 0.0225,
+      "step": 22250
+    },
+    {
+      "action_loss": 0.008046984672546387,
+      "epoch": 20.008992805755394,
+      "step": 22250
+    },
+    {
+      "epoch": 20.008992805755394,
+      "step": 22250,
+      "torque_loss": 0.12169583886861801
+    },
+    {
+      "epoch": 20.008992805755394,
+      "force_loss": 0.004948172718286514,
+      "step": 22250
+    },
+    {
+      "epoch": 20.01798561151079,
+      "grad_norm": 0.23541094362735748,
+      "learning_rate": 7.437935012565322e-05,
+      "loss": 0.0219,
+      "step": 22260
+    },
+    {
+      "action_loss": 0.003909599035978317,
+      "epoch": 20.01798561151079,
+      "step": 22260
+    },
+    {
+      "epoch": 20.01798561151079,
+      "step": 22260,
+      "torque_loss": 0.08229057490825653
+    },
+    {
+      "epoch": 20.01798561151079,
+      "force_loss": 0.005453682038933039,
+      "step": 22260
+    },
+    {
+      "epoch": 20.026978417266186,
+      "grad_norm": 0.23424990475177765,
+      "learning_rate": 7.435528638770893e-05,
+      "loss": 0.0234,
+      "step": 22270
+    },
+    {
+      "action_loss": 0.007413262501358986,
+      "epoch": 20.026978417266186,
+      "step": 22270
+    },
+    {
+      "epoch": 20.026978417266186,
+      "step": 22270,
+      "torque_loss": 0.1534394770860672
+    },
+    {
+      "epoch": 20.026978417266186,
+      "force_loss": 0.003463149070739746,
+      "step": 22270
+    },
+    {
+      "epoch": 20.035971223021583,
+      "grad_norm": 0.3123205602169037,
+      "learning_rate": 7.433121525127171e-05,
+      "loss": 0.0284,
+      "step": 22280
+    },
+    {
+      "action_loss": 0.011486530303955078,
+      "epoch": 20.035971223021583,
+      "step": 22280
+    },
+    {
+      "epoch": 20.035971223021583,
+      "step": 22280,
+      "torque_loss": 0.14047278463840485
+    },
+    {
+      "epoch": 20.035971223021583,
+      "force_loss": 0.00781856756657362,
+      "step": 22280
+    },
+    {
+      "epoch": 20.044964028776977,
+      "grad_norm": 0.29927730560302734,
+      "learning_rate": 7.430713672365371e-05,
+      "loss": 0.0341,
+      "step": 22290
+    },
+    {
+      "action_loss": 0.0031803464516997337,
+      "epoch": 20.044964028776977,
+      "step": 22290
+    },
+    {
+      "epoch": 20.044964028776977,
+      "step": 22290,
+      "torque_loss": 0.12818396091461182
+    },
+    {
+      "epoch": 20.044964028776977,
+      "force_loss": 0.003122147871181369,
+      "step": 22290
+    },
+    {
+      "epoch": 20.053956834532375,
+      "grad_norm": 0.2114167958498001,
+      "learning_rate": 7.428305081216938e-05,
+      "loss": 0.023,
+      "step": 22300
+    },
+    {
+      "action_loss": 0.00613408163189888,
+      "epoch": 20.053956834532375,
+      "step": 22300
+    },
+    {
+      "epoch": 20.053956834532375,
+      "step": 22300,
+      "torque_loss": 0.14041538536548615
+    },
+    {
+      "epoch": 20.053956834532375,
+      "force_loss": 0.005299183074384928,
+      "step": 22300
+    },
+    {
+      "epoch": 20.06294964028777,
+      "grad_norm": 0.38932931423187256,
+      "learning_rate": 7.425895752413536e-05,
+      "loss": 0.0236,
+      "step": 22310
+    },
+    {
+      "action_loss": 0.017816336825489998,
+      "epoch": 20.06294964028777,
+      "step": 22310
+    },
+    {
+      "epoch": 20.06294964028777,
+      "step": 22310,
+      "torque_loss": 0.10648417472839355
+    },
+    {
+      "epoch": 20.06294964028777,
+      "force_loss": 0.03322720527648926,
+      "step": 22310
+    },
+    {
+      "epoch": 20.071942446043167,
+      "grad_norm": 0.2532728910446167,
+      "learning_rate": 7.423485686687057e-05,
+      "loss": 0.0274,
+      "step": 22320
+    },
+    {
+      "action_loss": 0.007788743823766708,
+      "epoch": 20.071942446043167,
+      "step": 22320
+    },
+    {
+      "epoch": 20.071942446043167,
+      "step": 22320,
+      "torque_loss": 0.13962529599666595
+    },
+    {
+      "epoch": 20.071942446043167,
+      "force_loss": 0.009766372852027416,
+      "step": 22320
+    },
+    {
+      "epoch": 20.08093525179856,
+      "grad_norm": 0.3897517919540405,
+      "learning_rate": 7.421074884769616e-05,
+      "loss": 0.0235,
+      "step": 22330
+    },
+    {
+      "action_loss": 0.014374022372066975,
+      "epoch": 20.08093525179856,
+      "step": 22330
+    },
+    {
+      "epoch": 20.08093525179856,
+      "step": 22330,
+      "torque_loss": 0.13326890766620636
+    },
+    {
+      "epoch": 20.08093525179856,
+      "force_loss": 0.014441430568695068,
+      "step": 22330
+    },
+    {
+      "epoch": 20.08992805755396,
+      "grad_norm": 0.3673367202281952,
+      "learning_rate": 7.418663347393548e-05,
+      "loss": 0.0238,
+      "step": 22340
+    },
+    {
+      "action_loss": 0.006223693490028381,
+      "epoch": 20.08992805755396,
+      "step": 22340
+    },
+    {
+      "epoch": 20.08992805755396,
+      "step": 22340,
+      "torque_loss": 0.14779575169086456
+    },
+    {
+      "epoch": 20.08992805755396,
+      "force_loss": 0.010368873365223408,
+      "step": 22340
+    },
+    {
+      "epoch": 20.098920863309353,
+      "grad_norm": 0.29293516278266907,
+      "learning_rate": 7.416251075291418e-05,
+      "loss": 0.0222,
+      "step": 22350
+    },
+    {
+      "action_loss": 0.004675194155424833,
+      "epoch": 20.098920863309353,
+      "step": 22350
+    },
+    {
+      "epoch": 20.098920863309353,
+      "step": 22350,
+      "torque_loss": 0.11690375953912735
+    },
+    {
+      "epoch": 20.098920863309353,
+      "force_loss": 0.0035475061740726233,
+      "step": 22350
+    },
+    {
+      "epoch": 20.107913669064747,
+      "grad_norm": 0.16194823384284973,
+      "learning_rate": 7.413838069196007e-05,
+      "loss": 0.0172,
+      "step": 22360
+    },
+    {
+      "action_loss": 0.004984290339052677,
+      "epoch": 20.107913669064747,
+      "step": 22360
+    },
+    {
+      "epoch": 20.107913669064747,
+      "step": 22360,
+      "torque_loss": 0.10476479679346085
+    },
+    {
+      "epoch": 20.107913669064747,
+      "force_loss": 0.006919895764440298,
+      "step": 22360
+    },
+    {
+      "epoch": 20.116906474820144,
+      "grad_norm": 0.18410949409008026,
+      "learning_rate": 7.411424329840324e-05,
+      "loss": 0.0232,
+      "step": 22370
+    },
+    {
+      "action_loss": 0.006483596283942461,
+      "epoch": 20.116906474820144,
+      "step": 22370
+    },
+    {
+      "epoch": 20.116906474820144,
+      "step": 22370,
+      "torque_loss": 0.12645767629146576
+    },
+    {
+      "epoch": 20.116906474820144,
+      "force_loss": 0.004071300849318504,
+      "step": 22370
+    },
+    {
+      "epoch": 20.12589928057554,
+      "grad_norm": 0.2255152463912964,
+      "learning_rate": 7.409009857957601e-05,
+      "loss": 0.0286,
+      "step": 22380
+    },
+    {
+      "action_loss": 0.01818394847214222,
+      "epoch": 20.12589928057554,
+      "step": 22380
+    },
+    {
+      "epoch": 20.12589928057554,
+      "step": 22380,
+      "torque_loss": 0.1535508930683136
+    },
+    {
+      "epoch": 20.12589928057554,
+      "force_loss": 0.01527877151966095,
+      "step": 22380
+    },
+    {
+      "epoch": 20.134892086330936,
+      "grad_norm": 0.23035389184951782,
+      "learning_rate": 7.40659465428129e-05,
+      "loss": 0.0273,
+      "step": 22390
+    },
+    {
+      "action_loss": 0.05318012833595276,
+      "epoch": 20.134892086330936,
+      "step": 22390
+    },
+    {
+      "epoch": 20.134892086330936,
+      "step": 22390,
+      "torque_loss": 0.18275542557239532
+    },
+    {
+      "epoch": 20.134892086330936,
+      "force_loss": 0.04396986588835716,
+      "step": 22390
+    },
+    {
+      "epoch": 20.14388489208633,
+      "grad_norm": 0.243258997797966,
+      "learning_rate": 7.404178719545063e-05,
+      "loss": 0.025,
+      "step": 22400
+    },
+    {
+      "action_loss": 0.010891060344874859,
+      "epoch": 20.14388489208633,
+      "step": 22400
+    },
+    {
+      "epoch": 20.14388489208633,
+      "step": 22400,
+      "torque_loss": 0.1615249067544937
+    },
+    {
+      "epoch": 20.14388489208633,
+      "force_loss": 0.007259754464030266,
+      "step": 22400
+    },
+    {
+      "epoch": 20.152877697841728,
+      "grad_norm": 0.33244723081588745,
+      "learning_rate": 7.401762054482822e-05,
+      "loss": 0.0276,
+      "step": 22410
+    },
+    {
+      "action_loss": 0.00869003590196371,
+      "epoch": 20.152877697841728,
+      "step": 22410
+    },
+    {
+      "epoch": 20.152877697841728,
+      "step": 22410,
+      "torque_loss": 0.11507581919431686
+    },
+    {
+      "epoch": 20.152877697841728,
+      "force_loss": 0.008545215241611004,
+      "step": 22410
+    },
+    {
+      "epoch": 20.16187050359712,
+      "grad_norm": 1.2588152885437012,
+      "learning_rate": 7.39934465982868e-05,
+      "loss": 0.0235,
+      "step": 22420
+    },
+    {
+      "action_loss": 0.005899036768823862,
+      "epoch": 20.16187050359712,
+      "step": 22420
+    },
+    {
+      "epoch": 20.16187050359712,
+      "step": 22420,
+      "torque_loss": 0.16736231744289398
+    },
+    {
+      "epoch": 20.16187050359712,
+      "force_loss": 0.004097914323210716,
+      "step": 22420
+    },
+    {
+      "epoch": 20.17086330935252,
+      "grad_norm": 0.44440752267837524,
+      "learning_rate": 7.396926536316984e-05,
+      "loss": 0.0212,
+      "step": 22430
+    },
+    {
+      "action_loss": 0.006229629274457693,
+      "epoch": 20.17086330935252,
+      "step": 22430
+    },
+    {
+      "epoch": 20.17086330935252,
+      "step": 22430,
+      "torque_loss": 0.13538970053195953
+    },
+    {
+      "epoch": 20.17086330935252,
+      "force_loss": 0.006261768285185099,
+      "step": 22430
+    },
+    {
+      "epoch": 20.179856115107913,
+      "grad_norm": 0.46360284090042114,
+      "learning_rate": 7.394507684682293e-05,
+      "loss": 0.0249,
+      "step": 22440
+    },
+    {
+      "action_loss": 0.0048889233730733395,
+      "epoch": 20.179856115107913,
+      "step": 22440
+    },
+    {
+      "epoch": 20.179856115107913,
+      "step": 22440,
+      "torque_loss": 0.07896500080823898
+    },
+    {
+      "epoch": 20.179856115107913,
+      "force_loss": 0.003936445340514183,
+      "step": 22440
+    },
+    {
+      "epoch": 20.18884892086331,
+      "grad_norm": 0.36999088525772095,
+      "learning_rate": 7.392088105659393e-05,
+      "loss": 0.022,
+      "step": 22450
+    },
+    {
+      "action_loss": 0.006878270301967859,
+      "epoch": 20.18884892086331,
+      "step": 22450
+    },
+    {
+      "epoch": 20.18884892086331,
+      "step": 22450,
+      "torque_loss": 0.11815337091684341
+    },
+    {
+      "epoch": 20.18884892086331,
+      "force_loss": 0.0078524649143219,
+      "step": 22450
+    },
+    {
+      "epoch": 20.197841726618705,
+      "grad_norm": 0.2953590154647827,
+      "learning_rate": 7.389667799983284e-05,
+      "loss": 0.0265,
+      "step": 22460
+    },
+    {
+      "action_loss": 0.004856071434915066,
+      "epoch": 20.197841726618705,
+      "step": 22460
+    },
+    {
+      "epoch": 20.197841726618705,
+      "step": 22460,
+      "torque_loss": 0.10968545824289322
+    },
+    {
+      "epoch": 20.197841726618705,
+      "force_loss": 0.0033237107563763857,
+      "step": 22460
+    },
+    {
+      "epoch": 20.2068345323741,
+      "grad_norm": 0.3318917453289032,
+      "learning_rate": 7.387246768389193e-05,
+      "loss": 0.0208,
+      "step": 22470
+    },
+    {
+      "action_loss": 0.017688913270831108,
+      "epoch": 20.2068345323741,
+      "step": 22470
+    },
+    {
+      "epoch": 20.2068345323741,
+      "step": 22470,
+      "torque_loss": 0.1457376927137375
+    },
+    {
+      "epoch": 20.2068345323741,
+      "force_loss": 0.020751332864165306,
+      "step": 22470
+    },
+    {
+      "epoch": 20.215827338129497,
+      "grad_norm": 0.25988060235977173,
+      "learning_rate": 7.384825011612563e-05,
+      "loss": 0.0201,
+      "step": 22480
+    },
+    {
+      "action_loss": 0.006920827552676201,
+      "epoch": 20.215827338129497,
+      "step": 22480
+    },
+    {
+      "epoch": 20.215827338129497,
+      "step": 22480,
+      "torque_loss": 0.1720820814371109
+    },
+    {
+      "epoch": 20.215827338129497,
+      "force_loss": 0.006652161478996277,
+      "step": 22480
+    },
+    {
+      "epoch": 20.22482014388489,
+      "grad_norm": 0.5164228081703186,
+      "learning_rate": 7.382402530389066e-05,
+      "loss": 0.0288,
+      "step": 22490
+    },
+    {
+      "action_loss": 0.004785399418324232,
+      "epoch": 20.22482014388489,
+      "step": 22490
+    },
+    {
+      "epoch": 20.22482014388489,
+      "step": 22490,
+      "torque_loss": 0.13154055178165436
+    },
+    {
+      "epoch": 20.22482014388489,
+      "force_loss": 0.0029237607959657907,
+      "step": 22490
+    },
+    {
+      "epoch": 20.23381294964029,
+      "grad_norm": 0.4319625198841095,
+      "learning_rate": 7.379979325454582e-05,
+      "loss": 0.0244,
+      "step": 22500
+    },
+    {
+      "action_loss": 0.0043952311389148235,
+      "epoch": 20.23381294964029,
+      "step": 22500
+    },
+    {
+      "epoch": 20.23381294964029,
+      "step": 22500,
+      "torque_loss": 0.12781020998954773
+    },
+    {
+      "epoch": 20.23381294964029,
+      "force_loss": 0.0027401261031627655,
+      "step": 22500
+    },
+    {
+      "epoch": 20.242805755395683,
+      "grad_norm": 0.3425253927707672,
+      "learning_rate": 7.37755539754522e-05,
+      "loss": 0.0228,
+      "step": 22510
+    },
+    {
+      "action_loss": 0.009167276322841644,
+      "epoch": 20.242805755395683,
+      "step": 22510
+    },
+    {
+      "epoch": 20.242805755395683,
+      "step": 22510,
+      "torque_loss": 0.1435106247663498
+    },
+    {
+      "epoch": 20.242805755395683,
+      "force_loss": 0.00964704342186451,
+      "step": 22510
+    },
+    {
+      "epoch": 20.25179856115108,
+      "grad_norm": 0.24721738696098328,
+      "learning_rate": 7.375130747397302e-05,
+      "loss": 0.02,
+      "step": 22520
+    },
+    {
+      "action_loss": 0.0033436038065701723,
+      "epoch": 20.25179856115108,
+      "step": 22520
+    },
+    {
+      "epoch": 20.25179856115108,
+      "step": 22520,
+      "torque_loss": 0.11359447240829468
+    },
+    {
+      "epoch": 20.25179856115108,
+      "force_loss": 0.0030099814757704735,
+      "step": 22520
+    },
+    {
+      "epoch": 20.260791366906474,
+      "grad_norm": 0.4744258224964142,
+      "learning_rate": 7.372705375747377e-05,
+      "loss": 0.0209,
+      "step": 22530
+    },
+    {
+      "action_loss": 0.006460311356931925,
+      "epoch": 20.260791366906474,
+      "step": 22530
+    },
+    {
+      "epoch": 20.260791366906474,
+      "step": 22530,
+      "torque_loss": 0.1257258802652359
+    },
+    {
+      "epoch": 20.260791366906474,
+      "force_loss": 0.008514966815710068,
+      "step": 22530
+    },
+    {
+      "epoch": 20.269784172661872,
+      "grad_norm": 0.5593452453613281,
+      "learning_rate": 7.370279283332205e-05,
+      "loss": 0.0201,
+      "step": 22540
+    },
+    {
+      "action_loss": 0.01024629082530737,
+      "epoch": 20.269784172661872,
+      "step": 22540
+    },
+    {
+      "epoch": 20.269784172661872,
+      "step": 22540,
+      "torque_loss": 0.1480618119239807
+    },
+    {
+      "epoch": 20.269784172661872,
+      "force_loss": 0.012057830579578876,
+      "step": 22540
+    },
+    {
+      "epoch": 20.278776978417266,
+      "grad_norm": 0.35618850588798523,
+      "learning_rate": 7.36785247088877e-05,
+      "loss": 0.0203,
+      "step": 22550
+    },
+    {
+      "action_loss": 0.007292577531188726,
+      "epoch": 20.278776978417266,
+      "step": 22550
+    },
+    {
+      "epoch": 20.278776978417266,
+      "step": 22550,
+      "torque_loss": 0.1357903629541397
+    },
+    {
+      "epoch": 20.278776978417266,
+      "force_loss": 0.004702517297118902,
+      "step": 22550
+    },
+    {
+      "epoch": 20.28776978417266,
+      "grad_norm": 0.21865610778331757,
+      "learning_rate": 7.365424939154275e-05,
+      "loss": 0.0226,
+      "step": 22560
+    },
+    {
+      "action_loss": 0.013641615398228168,
+      "epoch": 20.28776978417266,
+      "step": 22560
+    },
+    {
+      "epoch": 20.28776978417266,
+      "step": 22560,
+      "torque_loss": 0.12795692682266235
+    },
+    {
+      "epoch": 20.28776978417266,
+      "force_loss": 0.013013812713325024,
+      "step": 22560
+    },
+    {
+      "epoch": 20.296762589928058,
+      "grad_norm": 0.22523823380470276,
+      "learning_rate": 7.362996688866138e-05,
+      "loss": 0.0214,
+      "step": 22570
+    },
+    {
+      "action_loss": 0.0032046933192759752,
+      "epoch": 20.296762589928058,
+      "step": 22570
+    },
+    {
+      "epoch": 20.296762589928058,
+      "step": 22570,
+      "torque_loss": 0.11089793592691422
+    },
+    {
+      "epoch": 20.296762589928058,
+      "force_loss": 0.002070546383038163,
+      "step": 22570
+    },
+    {
+      "epoch": 20.305755395683452,
+      "grad_norm": 0.5252678394317627,
+      "learning_rate": 7.360567720761999e-05,
+      "loss": 0.0239,
+      "step": 22580
+    },
+    {
+      "action_loss": 0.008866540156304836,
+      "epoch": 20.305755395683452,
+      "step": 22580
+    },
+    {
+      "epoch": 20.305755395683452,
+      "step": 22580,
+      "torque_loss": 0.1712295413017273
+    },
+    {
+      "epoch": 20.305755395683452,
+      "force_loss": 0.009010610170662403,
+      "step": 22580
+    },
+    {
+      "epoch": 20.31474820143885,
+      "grad_norm": 0.22528356313705444,
+      "learning_rate": 7.358138035579711e-05,
+      "loss": 0.0225,
+      "step": 22590
+    },
+    {
+      "action_loss": 0.003371112747117877,
+      "epoch": 20.31474820143885,
+      "step": 22590
+    },
+    {
+      "epoch": 20.31474820143885,
+      "step": 22590,
+      "torque_loss": 0.1017804741859436
+    },
+    {
+      "epoch": 20.31474820143885,
+      "force_loss": 0.0027128432411700487,
+      "step": 22590
+    },
+    {
+      "epoch": 20.323741007194243,
+      "grad_norm": 0.21611282229423523,
+      "learning_rate": 7.355707634057354e-05,
+      "loss": 0.0214,
+      "step": 22600
+    },
+    {
+      "action_loss": 0.005914825946092606,
+      "epoch": 20.323741007194243,
+      "step": 22600
+    },
+    {
+      "epoch": 20.323741007194243,
+      "step": 22600,
+      "torque_loss": 0.14993257820606232
+    },
+    {
+      "epoch": 20.323741007194243,
+      "force_loss": 0.004669476766139269,
+      "step": 22600
+    },
+    {
+      "epoch": 20.33273381294964,
+      "grad_norm": 0.287702739238739,
+      "learning_rate": 7.353276516933215e-05,
+      "loss": 0.0172,
+      "step": 22610
+    },
+    {
+      "action_loss": 0.014122466556727886,
+      "epoch": 20.33273381294964,
+      "step": 22610
+    },
+    {
+      "epoch": 20.33273381294964,
+      "step": 22610,
+      "torque_loss": 0.15321184694766998
+    },
+    {
+      "epoch": 20.33273381294964,
+      "force_loss": 0.02719024010002613,
+      "step": 22610
+    },
+    {
+      "epoch": 20.341726618705035,
+      "grad_norm": 0.1807129979133606,
+      "learning_rate": 7.350844684945806e-05,
+      "loss": 0.0228,
+      "step": 22620
+    },
+    {
+      "action_loss": 0.008383541367948055,
+      "epoch": 20.341726618705035,
+      "step": 22620
+    },
+    {
+      "epoch": 20.341726618705035,
+      "step": 22620,
+      "torque_loss": 0.17359912395477295
+    },
+    {
+      "epoch": 20.341726618705035,
+      "force_loss": 0.014319824986159801,
+      "step": 22620
+    },
+    {
+      "epoch": 20.350719424460433,
+      "grad_norm": 0.26065948605537415,
+      "learning_rate": 7.348412138833851e-05,
+      "loss": 0.0194,
+      "step": 22630
+    },
+    {
+      "action_loss": 0.005191102158278227,
+      "epoch": 20.350719424460433,
+      "step": 22630
+    },
+    {
+      "epoch": 20.350719424460433,
+      "step": 22630,
+      "torque_loss": 0.10668083280324936
+    },
+    {
+      "epoch": 20.350719424460433,
+      "force_loss": 0.005676930770277977,
+      "step": 22630
+    },
+    {
+      "epoch": 20.359712230215827,
+      "grad_norm": 0.29040876030921936,
+      "learning_rate": 7.345978879336295e-05,
+      "loss": 0.0194,
+      "step": 22640
+    },
+    {
+      "action_loss": 0.004312486387789249,
+      "epoch": 20.359712230215827,
+      "step": 22640
+    },
+    {
+      "epoch": 20.359712230215827,
+      "step": 22640,
+      "torque_loss": 0.11470244079828262
+    },
+    {
+      "epoch": 20.359712230215827,
+      "force_loss": 0.006611905992031097,
+      "step": 22640
+    },
+    {
+      "epoch": 20.368705035971225,
+      "grad_norm": 0.25477704405784607,
+      "learning_rate": 7.343544907192296e-05,
+      "loss": 0.0251,
+      "step": 22650
+    },
+    {
+      "action_loss": 0.007612638175487518,
+      "epoch": 20.368705035971225,
+      "step": 22650
+    },
+    {
+      "epoch": 20.368705035971225,
+      "step": 22650,
+      "torque_loss": 0.19843995571136475
+    },
+    {
+      "epoch": 20.368705035971225,
+      "force_loss": 0.005092856008559465,
+      "step": 22650
+    },
+    {
+      "epoch": 20.37769784172662,
+      "grad_norm": 0.263653427362442,
+      "learning_rate": 7.341110223141235e-05,
+      "loss": 0.0223,
+      "step": 22660
+    },
+    {
+      "action_loss": 0.004654020071029663,
+      "epoch": 20.37769784172662,
+      "step": 22660
+    },
+    {
+      "epoch": 20.37769784172662,
+      "step": 22660,
+      "torque_loss": 0.1299055814743042
+    },
+    {
+      "epoch": 20.37769784172662,
+      "force_loss": 0.0033027261961251497,
+      "step": 22660
+    },
+    {
+      "epoch": 20.386690647482013,
+      "grad_norm": 0.6859336495399475,
+      "learning_rate": 7.3386748279227e-05,
+      "loss": 0.0252,
+      "step": 22670
+    },
+    {
+      "action_loss": 0.004958470351994038,
+      "epoch": 20.386690647482013,
+      "step": 22670
+    },
+    {
+      "epoch": 20.386690647482013,
+      "step": 22670,
+      "torque_loss": 0.12926582992076874
+    },
+    {
+      "epoch": 20.386690647482013,
+      "force_loss": 0.0032022378873080015,
+      "step": 22670
+    },
+    {
+      "epoch": 20.39568345323741,
+      "grad_norm": 0.8659705519676208,
+      "learning_rate": 7.336238722276501e-05,
+      "loss": 0.0267,
+      "step": 22680
+    },
+    {
+      "action_loss": 0.011049383319914341,
+      "epoch": 20.39568345323741,
+      "step": 22680
+    },
+    {
+      "epoch": 20.39568345323741,
+      "step": 22680,
+      "torque_loss": 0.12452337890863419
+    },
+    {
+      "epoch": 20.39568345323741,
+      "force_loss": 0.005410314071923494,
+      "step": 22680
+    },
+    {
+      "epoch": 20.404676258992804,
+      "grad_norm": 0.4141238331794739,
+      "learning_rate": 7.333801906942663e-05,
+      "loss": 0.0203,
+      "step": 22690
+    },
+    {
+      "action_loss": 0.010135426186025143,
+      "epoch": 20.404676258992804,
+      "step": 22690
+    },
+    {
+      "epoch": 20.404676258992804,
+      "step": 22690,
+      "torque_loss": 0.12890994548797607
+    },
+    {
+      "epoch": 20.404676258992804,
+      "force_loss": 0.01397536601871252,
+      "step": 22690
+    },
+    {
+      "epoch": 20.413669064748202,
+      "grad_norm": 0.21891717612743378,
+      "learning_rate": 7.331364382661428e-05,
+      "loss": 0.0201,
+      "step": 22700
+    },
+    {
+      "action_loss": 0.004239857662469149,
+      "epoch": 20.413669064748202,
+      "step": 22700
+    },
+    {
+      "epoch": 20.413669064748202,
+      "step": 22700,
+      "torque_loss": 0.14295683801174164
+    },
+    {
+      "epoch": 20.413669064748202,
+      "force_loss": 0.006196148693561554,
+      "step": 22700
+    },
+    {
+      "epoch": 20.422661870503596,
+      "grad_norm": 0.2961306869983673,
+      "learning_rate": 7.328926150173248e-05,
+      "loss": 0.0211,
+      "step": 22710
+    },
+    {
+      "action_loss": 0.005986800882965326,
+      "epoch": 20.422661870503596,
+      "step": 22710
+    },
+    {
+      "epoch": 20.422661870503596,
+      "step": 22710,
+      "torque_loss": 0.09392008930444717
+    },
+    {
+      "epoch": 20.422661870503596,
+      "force_loss": 0.0027146136853843927,
+      "step": 22710
+    },
+    {
+      "epoch": 20.431654676258994,
+      "grad_norm": 0.4203747510910034,
+      "learning_rate": 7.326487210218795e-05,
+      "loss": 0.02,
+      "step": 22720
+    },
+    {
+      "action_loss": 0.008271696977317333,
+      "epoch": 20.431654676258994,
+      "step": 22720
+    },
+    {
+      "epoch": 20.431654676258994,
+      "step": 22720,
+      "torque_loss": 0.12983034551143646
+    },
+    {
+      "epoch": 20.431654676258994,
+      "force_loss": 0.004994562827050686,
+      "step": 22720
+    },
+    {
+      "epoch": 20.440647482014388,
+      "grad_norm": 0.31495434045791626,
+      "learning_rate": 7.324047563538955e-05,
+      "loss": 0.026,
+      "step": 22730
+    },
+    {
+      "action_loss": 0.00360013241879642,
+      "epoch": 20.440647482014388,
+      "step": 22730
+    },
+    {
+      "epoch": 20.440647482014388,
+      "step": 22730,
+      "torque_loss": 0.15073582530021667
+    },
+    {
+      "epoch": 20.440647482014388,
+      "force_loss": 0.0022492518182843924,
+      "step": 22730
+    },
+    {
+      "epoch": 20.449640287769785,
+      "grad_norm": 0.2842729091644287,
+      "learning_rate": 7.321607210874828e-05,
+      "loss": 0.0221,
+      "step": 22740
+    },
+    {
+      "action_loss": 0.011417836882174015,
+      "epoch": 20.449640287769785,
+      "step": 22740
+    },
+    {
+      "epoch": 20.449640287769785,
+      "step": 22740,
+      "torque_loss": 0.15755224227905273
+    },
+    {
+      "epoch": 20.449640287769785,
+      "force_loss": 0.019844690337777138,
+      "step": 22740
+    },
+    {
+      "epoch": 20.45863309352518,
+      "grad_norm": 0.2664801776409149,
+      "learning_rate": 7.31916615296773e-05,
+      "loss": 0.0208,
+      "step": 22750
+    },
+    {
+      "action_loss": 0.008020293898880482,
+      "epoch": 20.45863309352518,
+      "step": 22750
+    },
+    {
+      "epoch": 20.45863309352518,
+      "step": 22750,
+      "torque_loss": 0.14187000691890717
+    },
+    {
+      "epoch": 20.45863309352518,
+      "force_loss": 0.005518303718417883,
+      "step": 22750
+    },
+    {
+      "epoch": 20.467625899280577,
+      "grad_norm": 0.20545458793640137,
+      "learning_rate": 7.316724390559188e-05,
+      "loss": 0.0193,
+      "step": 22760
+    },
+    {
+      "action_loss": 0.0034446020144969225,
+      "epoch": 20.467625899280577,
+      "step": 22760
+    },
+    {
+      "epoch": 20.467625899280577,
+      "step": 22760,
+      "torque_loss": 0.10390549153089523
+    },
+    {
+      "epoch": 20.467625899280577,
+      "force_loss": 0.0020792176946997643,
+      "step": 22760
+    },
+    {
+      "epoch": 20.47661870503597,
+      "grad_norm": 0.2803266942501068,
+      "learning_rate": 7.314281924390946e-05,
+      "loss": 0.0187,
+      "step": 22770
+    },
+    {
+      "action_loss": 0.025641782209277153,
+      "epoch": 20.47661870503597,
+      "step": 22770
+    },
+    {
+      "epoch": 20.47661870503597,
+      "step": 22770,
+      "torque_loss": 0.14045365154743195
+    },
+    {
+      "epoch": 20.47661870503597,
+      "force_loss": 0.024263186380267143,
+      "step": 22770
+    },
+    {
+      "epoch": 20.485611510791365,
+      "grad_norm": 0.3588990271091461,
+      "learning_rate": 7.311838755204959e-05,
+      "loss": 0.0224,
+      "step": 22780
+    },
+    {
+      "action_loss": 0.004766867961734533,
+      "epoch": 20.485611510791365,
+      "step": 22780
+    },
+    {
+      "epoch": 20.485611510791365,
+      "step": 22780,
+      "torque_loss": 0.10751763731241226
+    },
+    {
+      "epoch": 20.485611510791365,
+      "force_loss": 0.0041995481587946415,
+      "step": 22780
+    },
+    {
+      "epoch": 20.494604316546763,
+      "grad_norm": 0.2514796853065491,
+      "learning_rate": 7.3093948837434e-05,
+      "loss": 0.0189,
+      "step": 22790
+    },
+    {
+      "action_loss": 0.0024767823051661253,
+      "epoch": 20.494604316546763,
+      "step": 22790
+    },
+    {
+      "epoch": 20.494604316546763,
+      "step": 22790,
+      "torque_loss": 0.08609968423843384
+    },
+    {
+      "epoch": 20.494604316546763,
+      "force_loss": 0.00434294156730175,
+      "step": 22790
+    },
+    {
+      "epoch": 20.503597122302157,
+      "grad_norm": 0.3953184187412262,
+      "learning_rate": 7.306950310748651e-05,
+      "loss": 0.0181,
+      "step": 22800
+    },
+    {
+      "action_loss": 0.00486801890656352,
+      "epoch": 20.503597122302157,
+      "step": 22800
+    },
+    {
+      "epoch": 20.503597122302157,
+      "step": 22800,
+      "torque_loss": 0.12692274153232574
+    },
+    {
+      "epoch": 20.503597122302157,
+      "force_loss": 0.003190352814272046,
+      "step": 22800
+    },
+    {
+      "epoch": 20.512589928057555,
+      "grad_norm": 0.9914472699165344,
+      "learning_rate": 7.304505036963311e-05,
+      "loss": 0.026,
+      "step": 22810
+    },
+    {
+      "action_loss": 0.008738495409488678,
+      "epoch": 20.512589928057555,
+      "step": 22810
+    },
+    {
+      "epoch": 20.512589928057555,
+      "step": 22810,
+      "torque_loss": 0.12337524443864822
+    },
+    {
+      "epoch": 20.512589928057555,
+      "force_loss": 0.015178129076957703,
+      "step": 22810
+    },
+    {
+      "epoch": 20.52158273381295,
+      "grad_norm": 0.38276538252830505,
+      "learning_rate": 7.302059063130186e-05,
+      "loss": 0.0215,
+      "step": 22820
+    },
+    {
+      "action_loss": 0.005425394978374243,
+      "epoch": 20.52158273381295,
+      "step": 22820
+    },
+    {
+      "epoch": 20.52158273381295,
+      "step": 22820,
+      "torque_loss": 0.13263416290283203
+    },
+    {
+      "epoch": 20.52158273381295,
+      "force_loss": 0.007371599320322275,
+      "step": 22820
+    },
+    {
+      "epoch": 20.530575539568346,
+      "grad_norm": 0.40876755118370056,
+      "learning_rate": 7.2996123899923e-05,
+      "loss": 0.0204,
+      "step": 22830
+    },
+    {
+      "action_loss": 0.0035579076502472162,
+      "epoch": 20.530575539568346,
+      "step": 22830
+    },
+    {
+      "epoch": 20.530575539568346,
+      "step": 22830,
+      "torque_loss": 0.16174590587615967
+    },
+    {
+      "epoch": 20.530575539568346,
+      "force_loss": 0.0021231574937701225,
+      "step": 22830
+    },
+    {
+      "epoch": 20.53956834532374,
+      "grad_norm": 0.3351193070411682,
+      "learning_rate": 7.297165018292886e-05,
+      "loss": 0.0237,
+      "step": 22840
+    },
+    {
+      "action_loss": 0.004874266218394041,
+      "epoch": 20.53956834532374,
+      "step": 22840
+    },
+    {
+      "epoch": 20.53956834532374,
+      "step": 22840,
+      "torque_loss": 0.1806125044822693
+    },
+    {
+      "epoch": 20.53956834532374,
+      "force_loss": 0.011274074204266071,
+      "step": 22840
+    },
+    {
+      "epoch": 20.548561151079138,
+      "grad_norm": 0.28446999192237854,
+      "learning_rate": 7.294716948775396e-05,
+      "loss": 0.0205,
+      "step": 22850
+    },
+    {
+      "action_loss": 0.009164727292954922,
+      "epoch": 20.548561151079138,
+      "step": 22850
+    },
+    {
+      "epoch": 20.548561151079138,
+      "step": 22850,
+      "torque_loss": 0.16843517124652863
+    },
+    {
+      "epoch": 20.548561151079138,
+      "force_loss": 0.008355320431292057,
+      "step": 22850
+    },
+    {
+      "epoch": 20.557553956834532,
+      "grad_norm": 0.26358136534690857,
+      "learning_rate": 7.292268182183484e-05,
+      "loss": 0.0206,
+      "step": 22860
+    },
+    {
+      "action_loss": 0.003183215158060193,
+      "epoch": 20.557553956834532,
+      "step": 22860
+    },
+    {
+      "epoch": 20.557553956834532,
+      "step": 22860,
+      "torque_loss": 0.1107417568564415
+    },
+    {
+      "epoch": 20.557553956834532,
+      "force_loss": 0.003787040011957288,
+      "step": 22860
+    },
+    {
+      "epoch": 20.56654676258993,
+      "grad_norm": 0.39613863825798035,
+      "learning_rate": 7.28981871926102e-05,
+      "loss": 0.0196,
+      "step": 22870
+    },
+    {
+      "action_loss": 0.009816057980060577,
+      "epoch": 20.56654676258993,
+      "step": 22870
+    },
+    {
+      "epoch": 20.56654676258993,
+      "step": 22870,
+      "torque_loss": 0.13265351951122284
+    },
+    {
+      "epoch": 20.56654676258993,
+      "force_loss": 0.005459487903863192,
+      "step": 22870
+    },
+    {
+      "epoch": 20.575539568345324,
+      "grad_norm": 0.4228397309780121,
+      "learning_rate": 7.28736856075209e-05,
+      "loss": 0.023,
+      "step": 22880
+    },
+    {
+      "action_loss": 0.004934034775942564,
+      "epoch": 20.575539568345324,
+      "step": 22880
+    },
+    {
+      "epoch": 20.575539568345324,
+      "step": 22880,
+      "torque_loss": 0.10597292333841324
+    },
+    {
+      "epoch": 20.575539568345324,
+      "force_loss": 0.00351532269269228,
+      "step": 22880
+    },
+    {
+      "epoch": 20.584532374100718,
+      "grad_norm": 0.10540630668401718,
+      "learning_rate": 7.284917707400985e-05,
+      "loss": 0.0198,
+      "step": 22890
+    },
+    {
+      "action_loss": 0.025004873052239418,
+      "epoch": 20.584532374100718,
+      "step": 22890
+    },
+    {
+      "epoch": 20.584532374100718,
+      "step": 22890,
+      "torque_loss": 0.18900613486766815
+    },
+    {
+      "epoch": 20.584532374100718,
+      "force_loss": 0.02924586832523346,
+      "step": 22890
+    },
+    {
+      "epoch": 20.593525179856115,
+      "grad_norm": 0.298160195350647,
+      "learning_rate": 7.282466159952212e-05,
+      "loss": 0.0244,
+      "step": 22900
+    },
+    {
+      "action_loss": 0.006574613507837057,
+      "epoch": 20.593525179856115,
+      "step": 22900
+    },
+    {
+      "epoch": 20.593525179856115,
+      "step": 22900,
+      "torque_loss": 0.1357024759054184
+    },
+    {
+      "epoch": 20.593525179856115,
+      "force_loss": 0.004133729729801416,
+      "step": 22900
+    },
+    {
+      "epoch": 20.60251798561151,
+      "grad_norm": 0.1741696447134018,
+      "learning_rate": 7.280013919150483e-05,
+      "loss": 0.0224,
+      "step": 22910
+    },
+    {
+      "action_loss": 0.028599942103028297,
+      "epoch": 20.60251798561151,
+      "step": 22910
+    },
+    {
+      "epoch": 20.60251798561151,
+      "step": 22910,
+      "torque_loss": 0.24114887416362762
+    },
+    {
+      "epoch": 20.60251798561151,
+      "force_loss": 0.036634355783462524,
+      "step": 22910
+    },
+    {
+      "epoch": 20.611510791366907,
+      "grad_norm": 0.37376120686531067,
+      "learning_rate": 7.277560985740728e-05,
+      "loss": 0.0316,
+      "step": 22920
+    },
+    {
+      "action_loss": 0.010220159776508808,
+      "epoch": 20.611510791366907,
+      "step": 22920
+    },
+    {
+      "epoch": 20.611510791366907,
+      "step": 22920,
+      "torque_loss": 0.13908237218856812
+    },
+    {
+      "epoch": 20.611510791366907,
+      "force_loss": 0.024738037958741188,
+      "step": 22920
+    },
+    {
+      "epoch": 20.6205035971223,
+      "grad_norm": 0.24346213042736053,
+      "learning_rate": 7.275107360468079e-05,
+      "loss": 0.0242,
+      "step": 22930
+    },
+    {
+      "action_loss": 0.006254466716200113,
+      "epoch": 20.6205035971223,
+      "step": 22930
+    },
+    {
+      "epoch": 20.6205035971223,
+      "step": 22930,
+      "torque_loss": 0.1192498728632927
+    },
+    {
+      "epoch": 20.6205035971223,
+      "force_loss": 0.006305792834609747,
+      "step": 22930
+    },
+    {
+      "epoch": 20.6294964028777,
+      "grad_norm": 0.25157687067985535,
+      "learning_rate": 7.272653044077885e-05,
+      "loss": 0.0218,
+      "step": 22940
+    },
+    {
+      "action_loss": 0.02662225253880024,
+      "epoch": 20.6294964028777,
+      "step": 22940
+    },
+    {
+      "epoch": 20.6294964028777,
+      "step": 22940,
+      "torque_loss": 0.2028772085905075
+    },
+    {
+      "epoch": 20.6294964028777,
+      "force_loss": 0.028705840930342674,
+      "step": 22940
+    },
+    {
+      "epoch": 20.638489208633093,
+      "grad_norm": 0.4135701060295105,
+      "learning_rate": 7.270198037315703e-05,
+      "loss": 0.0277,
+      "step": 22950
+    },
+    {
+      "action_loss": 0.006027355790138245,
+      "epoch": 20.638489208633093,
+      "step": 22950
+    },
+    {
+      "epoch": 20.638489208633093,
+      "step": 22950,
+      "torque_loss": 0.14523166418075562
+    },
+    {
+      "epoch": 20.638489208633093,
+      "force_loss": 0.003223589388653636,
+      "step": 22950
+    },
+    {
+      "epoch": 20.64748201438849,
+      "grad_norm": 0.1827203780412674,
+      "learning_rate": 7.267742340927297e-05,
+      "loss": 0.0191,
+      "step": 22960
+    },
+    {
+      "action_loss": 0.008993715979158878,
+      "epoch": 20.64748201438849,
+      "step": 22960
+    },
+    {
+      "epoch": 20.64748201438849,
+      "step": 22960,
+      "torque_loss": 0.1408950835466385
+    },
+    {
+      "epoch": 20.64748201438849,
+      "force_loss": 0.007974785752594471,
+      "step": 22960
+    },
+    {
+      "epoch": 20.656474820143885,
+      "grad_norm": 0.1610017716884613,
+      "learning_rate": 7.265285955658645e-05,
+      "loss": 0.0186,
+      "step": 22970
+    },
+    {
+      "action_loss": 0.012898233719170094,
+      "epoch": 20.656474820143885,
+      "step": 22970
+    },
+    {
+      "epoch": 20.656474820143885,
+      "step": 22970,
+      "torque_loss": 0.14384199678897858
+    },
+    {
+      "epoch": 20.656474820143885,
+      "force_loss": 0.02432405948638916,
+      "step": 22970
+    },
+    {
+      "epoch": 20.665467625899282,
+      "grad_norm": 0.2561543583869934,
+      "learning_rate": 7.26282888225593e-05,
+      "loss": 0.021,
+      "step": 22980
+    },
+    {
+      "action_loss": 0.005474910605698824,
+      "epoch": 20.665467625899282,
+      "step": 22980
+    },
+    {
+      "epoch": 20.665467625899282,
+      "step": 22980,
+      "torque_loss": 0.10223334282636642
+    },
+    {
+      "epoch": 20.665467625899282,
+      "force_loss": 0.009249657392501831,
+      "step": 22980
+    },
+    {
+      "epoch": 20.674460431654676,
+      "grad_norm": 0.2637881636619568,
+      "learning_rate": 7.260371121465548e-05,
+      "loss": 0.0205,
+      "step": 22990
+    },
+    {
+      "action_loss": 0.004197610542178154,
+      "epoch": 20.674460431654676,
+      "step": 22990
+    },
+    {
+      "epoch": 20.674460431654676,
+      "step": 22990,
+      "torque_loss": 0.09877309948205948
+    },
+    {
+      "epoch": 20.674460431654676,
+      "force_loss": 0.0035306450445204973,
+      "step": 22990
+    },
+    {
+      "epoch": 20.68345323741007,
+      "grad_norm": 0.21508187055587769,
+      "learning_rate": 7.2579126740341e-05,
+      "loss": 0.024,
+      "step": 23000
+    },
+    {
+      "action_loss": 0.015079661272466183,
+      "epoch": 20.68345323741007,
+      "step": 23000
+    },
+    {
+      "epoch": 20.68345323741007,
+      "step": 23000,
+      "torque_loss": 0.15902179479599
+    },
+    {
+      "epoch": 20.68345323741007,
+      "force_loss": 0.01978220045566559,
+      "step": 23000
+    },
+    {
+      "epoch": 20.692446043165468,
+      "grad_norm": 0.15836142003536224,
+      "learning_rate": 7.2554535407084e-05,
+      "loss": 0.0189,
+      "step": 23010
+    },
+    {
+      "action_loss": 0.002242977963760495,
+      "epoch": 20.692446043165468,
+      "step": 23010
+    },
+    {
+      "epoch": 20.692446043165468,
+      "step": 23010,
+      "torque_loss": 0.1343315690755844
+    },
+    {
+      "epoch": 20.692446043165468,
+      "force_loss": 0.00287189451046288,
+      "step": 23010
+    },
+    {
+      "epoch": 20.701438848920862,
+      "grad_norm": 0.14945478737354279,
+      "learning_rate": 7.252993722235464e-05,
+      "loss": 0.0216,
+      "step": 23020
+    },
+    {
+      "action_loss": 0.00448503065854311,
+      "epoch": 20.701438848920862,
+      "step": 23020
+    },
+    {
+      "epoch": 20.701438848920862,
+      "step": 23020,
+      "torque_loss": 0.10860016196966171
+    },
+    {
+      "epoch": 20.701438848920862,
+      "force_loss": 0.004214582499116659,
+      "step": 23020
+    },
+    {
+      "epoch": 20.71043165467626,
+      "grad_norm": 0.24460574984550476,
+      "learning_rate": 7.250533219362523e-05,
+      "loss": 0.0188,
+      "step": 23030
+    },
+    {
+      "action_loss": 0.007515007629990578,
+      "epoch": 20.71043165467626,
+      "step": 23030
+    },
+    {
+      "epoch": 20.71043165467626,
+      "step": 23030,
+      "torque_loss": 0.15320223569869995
+    },
+    {
+      "epoch": 20.71043165467626,
+      "force_loss": 0.008584956638514996,
+      "step": 23030
+    },
+    {
+      "epoch": 20.719424460431654,
+      "grad_norm": 0.33232659101486206,
+      "learning_rate": 7.248072032837012e-05,
+      "loss": 0.024,
+      "step": 23040
+    },
+    {
+      "action_loss": 0.018211496993899345,
+      "epoch": 20.719424460431654,
+      "step": 23040
+    },
+    {
+      "epoch": 20.719424460431654,
+      "step": 23040,
+      "torque_loss": 0.12002400308847427
+    },
+    {
+      "epoch": 20.719424460431654,
+      "force_loss": 0.012630761601030827,
+      "step": 23040
+    },
+    {
+      "epoch": 20.72841726618705,
+      "grad_norm": 0.4287187159061432,
+      "learning_rate": 7.245610163406575e-05,
+      "loss": 0.0221,
+      "step": 23050
+    },
+    {
+      "action_loss": 0.009795115329325199,
+      "epoch": 20.72841726618705,
+      "step": 23050
+    },
+    {
+      "epoch": 20.72841726618705,
+      "step": 23050,
+      "torque_loss": 0.10163215547800064
+    },
+    {
+      "epoch": 20.72841726618705,
+      "force_loss": 0.009978638030588627,
+      "step": 23050
+    },
+    {
+      "epoch": 20.737410071942445,
+      "grad_norm": 0.3301442861557007,
+      "learning_rate": 7.243147611819061e-05,
+      "loss": 0.0226,
+      "step": 23060
+    },
+    {
+      "action_loss": 0.0025839603040367365,
+      "epoch": 20.737410071942445,
+      "step": 23060
+    },
+    {
+      "epoch": 20.737410071942445,
+      "step": 23060,
+      "torque_loss": 0.13534292578697205
+    },
+    {
+      "epoch": 20.737410071942445,
+      "force_loss": 0.002022884087637067,
+      "step": 23060
+    },
+    {
+      "epoch": 20.746402877697843,
+      "grad_norm": 0.3564178943634033,
+      "learning_rate": 7.240684378822531e-05,
+      "loss": 0.0201,
+      "step": 23070
+    },
+    {
+      "action_loss": 0.0025712875649333,
+      "epoch": 20.746402877697843,
+      "step": 23070
+    },
+    {
+      "epoch": 20.746402877697843,
+      "step": 23070,
+      "torque_loss": 0.13158206641674042
+    },
+    {
+      "epoch": 20.746402877697843,
+      "force_loss": 0.0022691027261316776,
+      "step": 23070
+    },
+    {
+      "epoch": 20.755395683453237,
+      "grad_norm": 0.4516822099685669,
+      "learning_rate": 7.238220465165248e-05,
+      "loss": 0.0194,
+      "step": 23080
+    },
+    {
+      "action_loss": 0.004054612014442682,
+      "epoch": 20.755395683453237,
+      "step": 23080
+    },
+    {
+      "epoch": 20.755395683453237,
+      "step": 23080,
+      "torque_loss": 0.14073596894741058
+    },
+    {
+      "epoch": 20.755395683453237,
+      "force_loss": 0.0026126548182219267,
+      "step": 23080
+    },
+    {
+      "epoch": 20.764388489208635,
+      "grad_norm": 0.4884900748729706,
+      "learning_rate": 7.235755871595684e-05,
+      "loss": 0.0248,
+      "step": 23090
+    },
+    {
+      "action_loss": 0.005712583661079407,
+      "epoch": 20.764388489208635,
+      "step": 23090
+    },
+    {
+      "epoch": 20.764388489208635,
+      "step": 23090,
+      "torque_loss": 0.09353253990411758
+    },
+    {
+      "epoch": 20.764388489208635,
+      "force_loss": 0.006905198562890291,
+      "step": 23090
+    },
+    {
+      "epoch": 20.77338129496403,
+      "grad_norm": 0.24786247313022614,
+      "learning_rate": 7.233290598862517e-05,
+      "loss": 0.0216,
+      "step": 23100
+    },
+    {
+      "action_loss": 0.00905452948063612,
+      "epoch": 20.77338129496403,
+      "step": 23100
+    },
+    {
+      "epoch": 20.77338129496403,
+      "step": 23100,
+      "torque_loss": 0.13807465136051178
+    },
+    {
+      "epoch": 20.77338129496403,
+      "force_loss": 0.005232261493802071,
+      "step": 23100
+    },
+    {
+      "epoch": 20.782374100719423,
+      "grad_norm": 0.39007019996643066,
+      "learning_rate": 7.230824647714635e-05,
+      "loss": 0.0211,
+      "step": 23110
+    },
+    {
+      "action_loss": 0.00991443544626236,
+      "epoch": 20.782374100719423,
+      "step": 23110
+    },
+    {
+      "epoch": 20.782374100719423,
+      "step": 23110,
+      "torque_loss": 0.13694830238819122
+    },
+    {
+      "epoch": 20.782374100719423,
+      "force_loss": 0.007330181542783976,
+      "step": 23110
+    },
+    {
+      "epoch": 20.79136690647482,
+      "grad_norm": 0.3686298131942749,
+      "learning_rate": 7.228358018901124e-05,
+      "loss": 0.0213,
+      "step": 23120
+    },
+    {
+      "action_loss": 0.005644665565341711,
+      "epoch": 20.79136690647482,
+      "step": 23120
+    },
+    {
+      "epoch": 20.79136690647482,
+      "step": 23120,
+      "torque_loss": 0.12125851958990097
+    },
+    {
+      "epoch": 20.79136690647482,
+      "force_loss": 0.0046434178948402405,
+      "step": 23120
+    },
+    {
+      "epoch": 20.800359712230215,
+      "grad_norm": 0.4194779396057129,
+      "learning_rate": 7.225890713171286e-05,
+      "loss": 0.0233,
+      "step": 23130
+    },
+    {
+      "action_loss": 0.010150275193154812,
+      "epoch": 20.800359712230215,
+      "step": 23130
+    },
+    {
+      "epoch": 20.800359712230215,
+      "step": 23130,
+      "torque_loss": 0.12016569823026657
+    },
+    {
+      "epoch": 20.800359712230215,
+      "force_loss": 0.008466050960123539,
+      "step": 23130
+    },
+    {
+      "epoch": 20.809352517985612,
+      "grad_norm": 0.25126224756240845,
+      "learning_rate": 7.223422731274618e-05,
+      "loss": 0.0202,
+      "step": 23140
+    },
+    {
+      "action_loss": 0.004344898741692305,
+      "epoch": 20.809352517985612,
+      "step": 23140
+    },
+    {
+      "epoch": 20.809352517985612,
+      "step": 23140,
+      "torque_loss": 0.11202933639287949
+    },
+    {
+      "epoch": 20.809352517985612,
+      "force_loss": 0.0033856763038784266,
+      "step": 23140
+    },
+    {
+      "epoch": 20.818345323741006,
+      "grad_norm": 0.5618451833724976,
+      "learning_rate": 7.220954073960832e-05,
+      "loss": 0.0193,
+      "step": 23150
+    },
+    {
+      "action_loss": 0.00615106662735343,
+      "epoch": 20.818345323741006,
+      "step": 23150
+    },
+    {
+      "epoch": 20.818345323741006,
+      "step": 23150,
+      "torque_loss": 0.12374911457300186
+    },
+    {
+      "epoch": 20.818345323741006,
+      "force_loss": 0.012422936968505383,
+      "step": 23150
+    },
+    {
+      "epoch": 20.827338129496404,
+      "grad_norm": 0.3195774555206299,
+      "learning_rate": 7.218484741979838e-05,
+      "loss": 0.0202,
+      "step": 23160
+    },
+    {
+      "action_loss": 0.004305599257349968,
+      "epoch": 20.827338129496404,
+      "step": 23160
+    },
+    {
+      "epoch": 20.827338129496404,
+      "step": 23160,
+      "torque_loss": 0.12096554040908813
+    },
+    {
+      "epoch": 20.827338129496404,
+      "force_loss": 0.004232310224324465,
+      "step": 23160
+    },
+    {
+      "epoch": 20.836330935251798,
+      "grad_norm": 0.36266034841537476,
+      "learning_rate": 7.216014736081756e-05,
+      "loss": 0.0246,
+      "step": 23170
+    },
+    {
+      "action_loss": 0.014980564825236797,
+      "epoch": 20.836330935251798,
+      "step": 23170
+    },
+    {
+      "epoch": 20.836330935251798,
+      "step": 23170,
+      "torque_loss": 0.16716547310352325
+    },
+    {
+      "epoch": 20.836330935251798,
+      "force_loss": 0.023115748539566994,
+      "step": 23170
+    },
+    {
+      "epoch": 20.845323741007196,
+      "grad_norm": 0.22026123106479645,
+      "learning_rate": 7.213544057016906e-05,
+      "loss": 0.0225,
+      "step": 23180
+    },
+    {
+      "action_loss": 0.0049982825294137,
+      "epoch": 20.845323741007196,
+      "step": 23180
+    },
+    {
+      "epoch": 20.845323741007196,
+      "step": 23180,
+      "torque_loss": 0.13764846324920654
+    },
+    {
+      "epoch": 20.845323741007196,
+      "force_loss": 0.0037465591449290514,
+      "step": 23180
+    },
+    {
+      "epoch": 20.85431654676259,
+      "grad_norm": 0.24484923481941223,
+      "learning_rate": 7.211072705535819e-05,
+      "loss": 0.0214,
+      "step": 23190
+    },
+    {
+      "action_loss": 0.006374166812747717,
+      "epoch": 20.85431654676259,
+      "step": 23190
+    },
+    {
+      "epoch": 20.85431654676259,
+      "step": 23190,
+      "torque_loss": 0.1526750773191452
+    },
+    {
+      "epoch": 20.85431654676259,
+      "force_loss": 0.004419608507305384,
+      "step": 23190
+    },
+    {
+      "epoch": 20.863309352517987,
+      "grad_norm": 0.2755548655986786,
+      "learning_rate": 7.208600682389224e-05,
+      "loss": 0.0211,
+      "step": 23200
+    },
+    {
+      "action_loss": 0.002987870015203953,
+      "epoch": 20.863309352517987,
+      "step": 23200
+    },
+    {
+      "epoch": 20.863309352517987,
+      "step": 23200,
+      "torque_loss": 0.1289345771074295
+    },
+    {
+      "epoch": 20.863309352517987,
+      "force_loss": 0.0045778024941682816,
+      "step": 23200
+    },
+    {
+      "epoch": 20.87230215827338,
+      "grad_norm": 0.6204583048820496,
+      "learning_rate": 7.206127988328055e-05,
+      "loss": 0.0225,
+      "step": 23210
+    },
+    {
+      "action_loss": 0.01421859860420227,
+      "epoch": 20.87230215827338,
+      "step": 23210
+    },
+    {
+      "epoch": 20.87230215827338,
+      "step": 23210,
+      "torque_loss": 0.124875508248806
+    },
+    {
+      "epoch": 20.87230215827338,
+      "force_loss": 0.02199861966073513,
+      "step": 23210
+    },
+    {
+      "epoch": 20.881294964028775,
+      "grad_norm": 0.24115519225597382,
+      "learning_rate": 7.203654624103453e-05,
+      "loss": 0.0212,
+      "step": 23220
+    },
+    {
+      "action_loss": 0.007234225515276194,
+      "epoch": 20.881294964028775,
+      "step": 23220
+    },
+    {
+      "epoch": 20.881294964028775,
+      "step": 23220,
+      "torque_loss": 0.15499652922153473
+    },
+    {
+      "epoch": 20.881294964028775,
+      "force_loss": 0.008450089953839779,
+      "step": 23220
+    },
+    {
+      "epoch": 20.890287769784173,
+      "grad_norm": 0.2984203100204468,
+      "learning_rate": 7.201180590466761e-05,
+      "loss": 0.0219,
+      "step": 23230
+    },
+    {
+      "action_loss": 0.003522929735481739,
+      "epoch": 20.890287769784173,
+      "step": 23230
+    },
+    {
+      "epoch": 20.890287769784173,
+      "step": 23230,
+      "torque_loss": 0.10293682664632797
+    },
+    {
+      "epoch": 20.890287769784173,
+      "force_loss": 0.004324816633015871,
+      "step": 23230
+    },
+    {
+      "epoch": 20.899280575539567,
+      "grad_norm": 0.18120406568050385,
+      "learning_rate": 7.198705888169523e-05,
+      "loss": 0.0254,
+      "step": 23240
+    },
+    {
+      "action_loss": 0.012180055491626263,
+      "epoch": 20.899280575539567,
+      "step": 23240
+    },
+    {
+      "epoch": 20.899280575539567,
+      "step": 23240,
+      "torque_loss": 0.16695387661457062
+    },
+    {
+      "epoch": 20.899280575539567,
+      "force_loss": 0.014077156782150269,
+      "step": 23240
+    },
+    {
+      "epoch": 20.908273381294965,
+      "grad_norm": 0.4418695569038391,
+      "learning_rate": 7.196230517963491e-05,
+      "loss": 0.0204,
+      "step": 23250
+    },
+    {
+      "action_loss": 0.031053727492690086,
+      "epoch": 20.908273381294965,
+      "step": 23250
+    },
+    {
+      "epoch": 20.908273381294965,
+      "step": 23250,
+      "torque_loss": 0.17252886295318604
+    },
+    {
+      "epoch": 20.908273381294965,
+      "force_loss": 0.021286115050315857,
+      "step": 23250
+    },
+    {
+      "epoch": 20.91726618705036,
+      "grad_norm": 0.2474447786808014,
+      "learning_rate": 7.193754480600615e-05,
+      "loss": 0.0203,
+      "step": 23260
+    },
+    {
+      "action_loss": 0.0039863549172878265,
+      "epoch": 20.91726618705036,
+      "step": 23260
+    },
+    {
+      "epoch": 20.91726618705036,
+      "step": 23260,
+      "torque_loss": 0.11533962935209274
+    },
+    {
+      "epoch": 20.91726618705036,
+      "force_loss": 0.0024454023223370314,
+      "step": 23260
+    },
+    {
+      "epoch": 20.926258992805757,
+      "grad_norm": 0.31433430314064026,
+      "learning_rate": 7.19127777683305e-05,
+      "loss": 0.0303,
+      "step": 23270
+    },
+    {
+      "action_loss": 0.002052891766652465,
+      "epoch": 20.926258992805757,
+      "step": 23270
+    },
+    {
+      "epoch": 20.926258992805757,
+      "step": 23270,
+      "torque_loss": 0.06688390672206879
+    },
+    {
+      "epoch": 20.926258992805757,
+      "force_loss": 0.0032323088962584734,
+      "step": 23270
+    },
+    {
+      "epoch": 20.93525179856115,
+      "grad_norm": 0.16507881879806519,
+      "learning_rate": 7.188800407413156e-05,
+      "loss": 0.0169,
+      "step": 23280
+    },
+    {
+      "action_loss": 0.002124940976500511,
+      "epoch": 20.93525179856115,
+      "step": 23280
+    },
+    {
+      "epoch": 20.93525179856115,
+      "step": 23280,
+      "torque_loss": 0.11510109901428223
+    },
+    {
+      "epoch": 20.93525179856115,
+      "force_loss": 0.0015078651485964656,
+      "step": 23280
+    },
+    {
+      "epoch": 20.944244604316548,
+      "grad_norm": 0.1809292882680893,
+      "learning_rate": 7.186322373093489e-05,
+      "loss": 0.0214,
+      "step": 23290
+    },
+    {
+      "action_loss": 0.002750733867287636,
+      "epoch": 20.944244604316548,
+      "step": 23290
+    },
+    {
+      "epoch": 20.944244604316548,
+      "step": 23290,
+      "torque_loss": 0.1321621686220169
+    },
+    {
+      "epoch": 20.944244604316548,
+      "force_loss": 0.003203033236786723,
+      "step": 23290
+    },
+    {
+      "epoch": 20.953237410071942,
+      "grad_norm": 0.420011967420578,
+      "learning_rate": 7.18384367462681e-05,
+      "loss": 0.0183,
+      "step": 23300
+    },
+    {
+      "action_loss": 0.011628490872681141,
+      "epoch": 20.953237410071942,
+      "step": 23300
+    },
+    {
+      "epoch": 20.953237410071942,
+      "step": 23300,
+      "torque_loss": 0.1385202556848526
+    },
+    {
+      "epoch": 20.953237410071942,
+      "force_loss": 0.009227244183421135,
+      "step": 23300
+    },
+    {
+      "epoch": 20.96223021582734,
+      "grad_norm": 0.31038522720336914,
+      "learning_rate": 7.181364312766085e-05,
+      "loss": 0.0212,
+      "step": 23310
+    },
+    {
+      "action_loss": 0.006483044475317001,
+      "epoch": 20.96223021582734,
+      "step": 23310
+    },
+    {
+      "epoch": 20.96223021582734,
+      "step": 23310,
+      "torque_loss": 0.20367340743541718
+    },
+    {
+      "epoch": 20.96223021582734,
+      "force_loss": 0.0034793762024492025,
+      "step": 23310
+    },
+    {
+      "epoch": 20.971223021582734,
+      "grad_norm": 0.2559487521648407,
+      "learning_rate": 7.178884288264477e-05,
+      "loss": 0.0183,
+      "step": 23320
+    },
+    {
+      "action_loss": 0.011040893383324146,
+      "epoch": 20.971223021582734,
+      "step": 23320
+    },
+    {
+      "epoch": 20.971223021582734,
+      "step": 23320,
+      "torque_loss": 0.156141459941864
+    },
+    {
+      "epoch": 20.971223021582734,
+      "force_loss": 0.011167776770889759,
+      "step": 23320
+    },
+    {
+      "epoch": 20.980215827338128,
+      "grad_norm": 0.4533368945121765,
+      "learning_rate": 7.176403601875353e-05,
+      "loss": 0.0226,
+      "step": 23330
+    },
+    {
+      "action_loss": 0.003421274945139885,
+      "epoch": 20.980215827338128,
+      "step": 23330
+    },
+    {
+      "epoch": 20.980215827338128,
+      "step": 23330,
+      "torque_loss": 0.1442335546016693
+    },
+    {
+      "epoch": 20.980215827338128,
+      "force_loss": 0.0024091105442494154,
+      "step": 23330
+    },
+    {
+      "epoch": 20.989208633093526,
+      "grad_norm": 0.3069468140602112,
+      "learning_rate": 7.173922254352279e-05,
+      "loss": 0.0251,
+      "step": 23340
+    },
+    {
+      "action_loss": 0.00260400609113276,
+      "epoch": 20.989208633093526,
+      "step": 23340
+    },
+    {
+      "epoch": 20.989208633093526,
+      "step": 23340,
+      "torque_loss": 0.13419514894485474
+    },
+    {
+      "epoch": 20.989208633093526,
+      "force_loss": 0.0029714603442698717,
+      "step": 23340
+    },
+    {
+      "epoch": 20.99820143884892,
+      "grad_norm": 0.2399994432926178,
+      "learning_rate": 7.171440246449024e-05,
+      "loss": 0.0196,
+      "step": 23350
+    },
+    {
+      "action_loss": 0.005962350871413946,
+      "epoch": 20.99820143884892,
+      "step": 23350
+    },
+    {
+      "epoch": 20.99820143884892,
+      "step": 23350,
+      "torque_loss": 0.13464462757110596
+    },
+    {
+      "epoch": 20.99820143884892,
+      "force_loss": 0.0050678919069468975,
+      "step": 23350
+    },
+    {
+      "epoch": 21.007194244604317,
+      "grad_norm": 0.19311892986297607,
+      "learning_rate": 7.168957578919555e-05,
+      "loss": 0.0176,
+      "step": 23360
+    },
+    {
+      "action_loss": 0.003959774039685726,
+      "epoch": 21.007194244604317,
+      "step": 23360
+    },
+    {
+      "epoch": 21.007194244604317,
+      "step": 23360,
+      "torque_loss": 0.1621791124343872
+    },
+    {
+      "epoch": 21.007194244604317,
+      "force_loss": 0.004346047528088093,
+      "step": 23360
+    },
+    {
+      "epoch": 21.01618705035971,
+      "grad_norm": 0.24874347448349,
+      "learning_rate": 7.16647425251804e-05,
+      "loss": 0.0214,
+      "step": 23370
+    },
+    {
+      "action_loss": 0.00828962866216898,
+      "epoch": 21.01618705035971,
+      "step": 23370
+    },
+    {
+      "epoch": 21.01618705035971,
+      "step": 23370,
+      "torque_loss": 0.12503741681575775
+    },
+    {
+      "epoch": 21.01618705035971,
+      "force_loss": 0.0070351199246943,
+      "step": 23370
+    },
+    {
+      "epoch": 21.02517985611511,
+      "grad_norm": 0.35728171467781067,
+      "learning_rate": 7.163990267998852e-05,
+      "loss": 0.023,
+      "step": 23380
+    },
+    {
+      "action_loss": 0.00807653833180666,
+      "epoch": 21.02517985611511,
+      "step": 23380
+    },
+    {
+      "epoch": 21.02517985611511,
+      "step": 23380,
+      "torque_loss": 0.1355939358472824
+    },
+    {
+      "epoch": 21.02517985611511,
+      "force_loss": 0.004546436946839094,
+      "step": 23380
+    },
+    {
+      "epoch": 21.034172661870503,
+      "grad_norm": 0.19132237136363983,
+      "learning_rate": 7.161505626116556e-05,
+      "loss": 0.0229,
+      "step": 23390
+    },
+    {
+      "action_loss": 0.003996682818979025,
+      "epoch": 21.034172661870503,
+      "step": 23390
+    },
+    {
+      "epoch": 21.034172661870503,
+      "step": 23390,
+      "torque_loss": 0.13249383866786957
+    },
+    {
+      "epoch": 21.034172661870503,
+      "force_loss": 0.002313007367774844,
+      "step": 23390
+    },
+    {
+      "epoch": 21.0431654676259,
+      "grad_norm": 0.63074791431427,
+      "learning_rate": 7.159020327625923e-05,
+      "loss": 0.0217,
+      "step": 23400
+    },
+    {
+      "action_loss": 0.013399988412857056,
+      "epoch": 21.0431654676259,
+      "step": 23400
+    },
+    {
+      "epoch": 21.0431654676259,
+      "step": 23400,
+      "torque_loss": 0.1996813416481018
+    },
+    {
+      "epoch": 21.0431654676259,
+      "force_loss": 0.017626948654651642,
+      "step": 23400
+    },
+    {
+      "epoch": 21.052158273381295,
+      "grad_norm": 0.4147021472454071,
+      "learning_rate": 7.15653437328192e-05,
+      "loss": 0.0298,
+      "step": 23410
+    },
+    {
+      "action_loss": 0.004978652112185955,
+      "epoch": 21.052158273381295,
+      "step": 23410
+    },
+    {
+      "epoch": 21.052158273381295,
+      "step": 23410,
+      "torque_loss": 0.13333041965961456
+    },
+    {
+      "epoch": 21.052158273381295,
+      "force_loss": 0.003872217610478401,
+      "step": 23410
+    },
+    {
+      "epoch": 21.06115107913669,
+      "grad_norm": 0.2319970726966858,
+      "learning_rate": 7.154047763839713e-05,
+      "loss": 0.0217,
+      "step": 23420
+    },
+    {
+      "action_loss": 0.002744711237028241,
+      "epoch": 21.06115107913669,
+      "step": 23420
+    },
+    {
+      "epoch": 21.06115107913669,
+      "step": 23420,
+      "torque_loss": 0.10988708585500717
+    },
+    {
+      "epoch": 21.06115107913669,
+      "force_loss": 0.0029605224262923002,
+      "step": 23420
+    },
+    {
+      "epoch": 21.070143884892087,
+      "grad_norm": 0.26481521129608154,
+      "learning_rate": 7.15156050005467e-05,
+      "loss": 0.0183,
+      "step": 23430
+    },
+    {
+      "action_loss": 0.009172564372420311,
+      "epoch": 21.070143884892087,
+      "step": 23430
+    },
+    {
+      "epoch": 21.070143884892087,
+      "step": 23430,
+      "torque_loss": 0.11302395910024643
+    },
+    {
+      "epoch": 21.070143884892087,
+      "force_loss": 0.015311741270124912,
+      "step": 23430
+    },
+    {
+      "epoch": 21.07913669064748,
+      "grad_norm": 0.360005646944046,
+      "learning_rate": 7.149072582682357e-05,
+      "loss": 0.0201,
+      "step": 23440
+    },
+    {
+      "action_loss": 0.013230375945568085,
+      "epoch": 21.07913669064748,
+      "step": 23440
+    },
+    {
+      "epoch": 21.07913669064748,
+      "step": 23440,
+      "torque_loss": 0.09906294196844101
+    },
+    {
+      "epoch": 21.07913669064748,
+      "force_loss": 0.013757321052253246,
+      "step": 23440
+    },
+    {
+      "epoch": 21.08812949640288,
+      "grad_norm": 0.1836540400981903,
+      "learning_rate": 7.146584012478535e-05,
+      "loss": 0.0181,
+      "step": 23450
+    },
+    {
+      "action_loss": 0.01820623129606247,
+      "epoch": 21.08812949640288,
+      "step": 23450
+    },
+    {
+      "epoch": 21.08812949640288,
+      "step": 23450,
+      "torque_loss": 0.10784520953893661
+    },
+    {
+      "epoch": 21.08812949640288,
+      "force_loss": 0.025196099653840065,
+      "step": 23450
+    },
+    {
+      "epoch": 21.097122302158272,
+      "grad_norm": 0.2669019401073456,
+      "learning_rate": 7.144094790199169e-05,
+      "loss": 0.021,
+      "step": 23460
+    },
+    {
+      "action_loss": 0.007781012449413538,
+      "epoch": 21.097122302158272,
+      "step": 23460
+    },
+    {
+      "epoch": 21.097122302158272,
+      "step": 23460,
+      "torque_loss": 0.11886411905288696
+    },
+    {
+      "epoch": 21.097122302158272,
+      "force_loss": 0.012984029948711395,
+      "step": 23460
+    },
+    {
+      "epoch": 21.10611510791367,
+      "grad_norm": 0.2791656255722046,
+      "learning_rate": 7.141604916600415e-05,
+      "loss": 0.02,
+      "step": 23470
+    },
+    {
+      "action_loss": 0.008043314330279827,
+      "epoch": 21.10611510791367,
+      "step": 23470
+    },
+    {
+      "epoch": 21.10611510791367,
+      "step": 23470,
+      "torque_loss": 0.20208413898944855
+    },
+    {
+      "epoch": 21.10611510791367,
+      "force_loss": 0.007641212549060583,
+      "step": 23470
+    },
+    {
+      "epoch": 21.115107913669064,
+      "grad_norm": 0.27927857637405396,
+      "learning_rate": 7.139114392438635e-05,
+      "loss": 0.0193,
+      "step": 23480
+    },
+    {
+      "action_loss": 0.026578575372695923,
+      "epoch": 21.115107913669064,
+      "step": 23480
+    },
+    {
+      "epoch": 21.115107913669064,
+      "step": 23480,
+      "torque_loss": 0.11552394181489944
+    },
+    {
+      "epoch": 21.115107913669064,
+      "force_loss": 0.02971476875245571,
+      "step": 23480
+    },
+    {
+      "epoch": 21.12410071942446,
+      "grad_norm": 0.23018686473369598,
+      "learning_rate": 7.136623218470382e-05,
+      "loss": 0.0218,
+      "step": 23490
+    },
+    {
+      "action_loss": 0.003139466280117631,
+      "epoch": 21.12410071942446,
+      "step": 23490
+    },
+    {
+      "epoch": 21.12410071942446,
+      "step": 23490,
+      "torque_loss": 0.14302611351013184
+    },
+    {
+      "epoch": 21.12410071942446,
+      "force_loss": 0.0020197771955281496,
+      "step": 23490
+    },
+    {
+      "epoch": 21.133093525179856,
+      "grad_norm": 0.23918592929840088,
+      "learning_rate": 7.13413139545241e-05,
+      "loss": 0.0201,
+      "step": 23500
+    },
+    {
+      "action_loss": 0.013414248824119568,
+      "epoch": 21.133093525179856,
+      "step": 23500
+    },
+    {
+      "epoch": 21.133093525179856,
+      "step": 23500,
+      "torque_loss": 0.15465418994426727
+    },
+    {
+      "epoch": 21.133093525179856,
+      "force_loss": 0.007328353356570005,
+      "step": 23500
+    },
+    {
+      "epoch": 21.142086330935253,
+      "grad_norm": 0.19082936644554138,
+      "learning_rate": 7.131638924141668e-05,
+      "loss": 0.024,
+      "step": 23510
+    },
+    {
+      "action_loss": 0.0034979190677404404,
+      "epoch": 21.142086330935253,
+      "step": 23510
+    },
+    {
+      "epoch": 21.142086330935253,
+      "step": 23510,
+      "torque_loss": 0.07275998592376709
+    },
+    {
+      "epoch": 21.142086330935253,
+      "force_loss": 0.006297590211033821,
+      "step": 23510
+    },
+    {
+      "epoch": 21.151079136690647,
+      "grad_norm": 0.5941745638847351,
+      "learning_rate": 7.129145805295304e-05,
+      "loss": 0.0221,
+      "step": 23520
+    },
+    {
+      "action_loss": 0.002326295478269458,
+      "epoch": 21.151079136690647,
+      "step": 23520
+    },
+    {
+      "epoch": 21.151079136690647,
+      "step": 23520,
+      "torque_loss": 0.10932417958974838
+    },
+    {
+      "epoch": 21.151079136690647,
+      "force_loss": 0.0025577836204320192,
+      "step": 23520
+    },
+    {
+      "epoch": 21.16007194244604,
+      "grad_norm": 0.10179499536752701,
+      "learning_rate": 7.126652039670661e-05,
+      "loss": 0.0155,
+      "step": 23530
+    },
+    {
+      "action_loss": 0.004577077459543943,
+      "epoch": 21.16007194244604,
+      "step": 23530
+    },
+    {
+      "epoch": 21.16007194244604,
+      "step": 23530,
+      "torque_loss": 0.12532149255275726
+    },
+    {
+      "epoch": 21.16007194244604,
+      "force_loss": 0.0070156678557395935,
+      "step": 23530
+    },
+    {
+      "epoch": 21.16906474820144,
+      "grad_norm": 0.256354957818985,
+      "learning_rate": 7.124157628025278e-05,
+      "loss": 0.02,
+      "step": 23540
+    },
+    {
+      "action_loss": 0.006174854468554258,
+      "epoch": 21.16906474820144,
+      "step": 23540
+    },
+    {
+      "epoch": 21.16906474820144,
+      "step": 23540,
+      "torque_loss": 0.11213638633489609
+    },
+    {
+      "epoch": 21.16906474820144,
+      "force_loss": 0.008580025285482407,
+      "step": 23540
+    },
+    {
+      "epoch": 21.178057553956833,
+      "grad_norm": 0.3622494041919708,
+      "learning_rate": 7.121662571116894e-05,
+      "loss": 0.0206,
+      "step": 23550
+    },
+    {
+      "action_loss": 0.00479300320148468,
+      "epoch": 21.178057553956833,
+      "step": 23550
+    },
+    {
+      "epoch": 21.178057553956833,
+      "step": 23550,
+      "torque_loss": 0.11562427133321762
+    },
+    {
+      "epoch": 21.178057553956833,
+      "force_loss": 0.002547664800658822,
+      "step": 23550
+    },
+    {
+      "epoch": 21.18705035971223,
+      "grad_norm": 0.44891542196273804,
+      "learning_rate": 7.119166869703441e-05,
+      "loss": 0.0166,
+      "step": 23560
+    },
+    {
+      "action_loss": 0.0022924719378352165,
+      "epoch": 21.18705035971223,
+      "step": 23560
+    },
+    {
+      "epoch": 21.18705035971223,
+      "step": 23560,
+      "torque_loss": 0.10936244577169418
+    },
+    {
+      "epoch": 21.18705035971223,
+      "force_loss": 0.004271793644875288,
+      "step": 23560
+    },
+    {
+      "epoch": 21.196043165467625,
+      "grad_norm": 0.26346316933631897,
+      "learning_rate": 7.116670524543044e-05,
+      "loss": 0.0229,
+      "step": 23570
+    },
+    {
+      "action_loss": 0.006651930045336485,
+      "epoch": 21.196043165467625,
+      "step": 23570
+    },
+    {
+      "epoch": 21.196043165467625,
+      "step": 23570,
+      "torque_loss": 0.1137520894408226
+    },
+    {
+      "epoch": 21.196043165467625,
+      "force_loss": 0.008779712952673435,
+      "step": 23570
+    },
+    {
+      "epoch": 21.205035971223023,
+      "grad_norm": 0.5915934443473816,
+      "learning_rate": 7.114173536394032e-05,
+      "loss": 0.0218,
+      "step": 23580
+    },
+    {
+      "action_loss": 0.005308545660227537,
+      "epoch": 21.205035971223023,
+      "step": 23580
+    },
+    {
+      "epoch": 21.205035971223023,
+      "step": 23580,
+      "torque_loss": 0.11905542761087418
+    },
+    {
+      "epoch": 21.205035971223023,
+      "force_loss": 0.00404251366853714,
+      "step": 23580
+    },
+    {
+      "epoch": 21.214028776978417,
+      "grad_norm": 0.23304638266563416,
+      "learning_rate": 7.111675906014917e-05,
+      "loss": 0.0228,
+      "step": 23590
+    },
+    {
+      "action_loss": 0.010965664871037006,
+      "epoch": 21.214028776978417,
+      "step": 23590
+    },
+    {
+      "epoch": 21.214028776978417,
+      "step": 23590,
+      "torque_loss": 0.1191180869936943
+    },
+    {
+      "epoch": 21.214028776978417,
+      "force_loss": 0.012466054409742355,
+      "step": 23590
+    },
+    {
+      "epoch": 21.223021582733814,
+      "grad_norm": 0.23632240295410156,
+      "learning_rate": 7.109177634164421e-05,
+      "loss": 0.0191,
+      "step": 23600
+    },
+    {
+      "action_loss": 0.002272161887958646,
+      "epoch": 21.223021582733814,
+      "step": 23600
+    },
+    {
+      "epoch": 21.223021582733814,
+      "step": 23600,
+      "torque_loss": 0.08585027605295181
+    },
+    {
+      "epoch": 21.223021582733814,
+      "force_loss": 0.0030193133279681206,
+      "step": 23600
+    },
+    {
+      "epoch": 21.23201438848921,
+      "grad_norm": 0.2457566112279892,
+      "learning_rate": 7.106678721601449e-05,
+      "loss": 0.0184,
+      "step": 23610
+    },
+    {
+      "action_loss": 0.009018865413963795,
+      "epoch": 21.23201438848921,
+      "step": 23610
+    },
+    {
+      "epoch": 21.23201438848921,
+      "step": 23610,
+      "torque_loss": 0.11562643200159073
+    },
+    {
+      "epoch": 21.23201438848921,
+      "force_loss": 0.009257269091904163,
+      "step": 23610
+    },
+    {
+      "epoch": 21.241007194244606,
+      "grad_norm": 0.34829962253570557,
+      "learning_rate": 7.104179169085103e-05,
+      "loss": 0.0218,
+      "step": 23620
+    },
+    {
+      "action_loss": 0.00873131025582552,
+      "epoch": 21.241007194244606,
+      "step": 23620
+    },
+    {
+      "epoch": 21.241007194244606,
+      "step": 23620,
+      "torque_loss": 0.11321505904197693
+    },
+    {
+      "epoch": 21.241007194244606,
+      "force_loss": 0.010424568317830563,
+      "step": 23620
+    },
+    {
+      "epoch": 21.25,
+      "grad_norm": 0.3373963534832001,
+      "learning_rate": 7.101678977374683e-05,
+      "loss": 0.0265,
+      "step": 23630
+    },
+    {
+      "action_loss": 0.003254042938351631,
+      "epoch": 21.25,
+      "step": 23630
+    },
+    {
+      "epoch": 21.25,
+      "step": 23630,
+      "torque_loss": 0.126604363322258
+    },
+    {
+      "epoch": 21.25,
+      "force_loss": 0.0015665879473090172,
+      "step": 23630
+    },
+    {
+      "epoch": 21.258992805755394,
+      "grad_norm": 0.31189942359924316,
+      "learning_rate": 7.099178147229685e-05,
+      "loss": 0.0205,
+      "step": 23640
+    },
+    {
+      "action_loss": 0.011472395621240139,
+      "epoch": 21.258992805755394,
+      "step": 23640
+    },
+    {
+      "epoch": 21.258992805755394,
+      "step": 23640,
+      "torque_loss": 0.13908208906650543
+    },
+    {
+      "epoch": 21.258992805755394,
+      "force_loss": 0.008833193220198154,
+      "step": 23640
+    },
+    {
+      "epoch": 21.26798561151079,
+      "grad_norm": 0.1859862357378006,
+      "learning_rate": 7.096676679409789e-05,
+      "loss": 0.022,
+      "step": 23650
+    },
+    {
+      "action_loss": 0.015845075249671936,
+      "epoch": 21.26798561151079,
+      "step": 23650
+    },
+    {
+      "epoch": 21.26798561151079,
+      "step": 23650,
+      "torque_loss": 0.13396631181240082
+    },
+    {
+      "epoch": 21.26798561151079,
+      "force_loss": 0.021496951580047607,
+      "step": 23650
+    },
+    {
+      "epoch": 21.276978417266186,
+      "grad_norm": 0.36624208092689514,
+      "learning_rate": 7.094174574674877e-05,
+      "loss": 0.0214,
+      "step": 23660
+    },
+    {
+      "action_loss": 0.005504108965396881,
+      "epoch": 21.276978417266186,
+      "step": 23660
+    },
+    {
+      "epoch": 21.276978417266186,
+      "step": 23660,
+      "torque_loss": 0.14271610975265503
+    },
+    {
+      "epoch": 21.276978417266186,
+      "force_loss": 0.004780036862939596,
+      "step": 23660
+    },
+    {
+      "epoch": 21.285971223021583,
+      "grad_norm": 0.44943949580192566,
+      "learning_rate": 7.091671833785025e-05,
+      "loss": 0.018,
+      "step": 23670
+    },
+    {
+      "action_loss": 0.0040962230414152145,
+      "epoch": 21.285971223021583,
+      "step": 23670
+    },
+    {
+      "epoch": 21.285971223021583,
+      "step": 23670,
+      "torque_loss": 0.11015243083238602
+    },
+    {
+      "epoch": 21.285971223021583,
+      "force_loss": 0.004503512289375067,
+      "step": 23670
+    },
+    {
+      "epoch": 21.294964028776977,
+      "grad_norm": 0.4552011489868164,
+      "learning_rate": 7.089168457500493e-05,
+      "loss": 0.0211,
+      "step": 23680
+    },
+    {
+      "action_loss": 0.006273546721786261,
+      "epoch": 21.294964028776977,
+      "step": 23680
+    },
+    {
+      "epoch": 21.294964028776977,
+      "step": 23680,
+      "torque_loss": 0.12740683555603027
+    },
+    {
+      "epoch": 21.294964028776977,
+      "force_loss": 0.00599891459569335,
+      "step": 23680
+    },
+    {
+      "epoch": 21.303956834532375,
+      "grad_norm": 0.15325917303562164,
+      "learning_rate": 7.086664446581747e-05,
+      "loss": 0.0256,
+      "step": 23690
+    },
+    {
+      "action_loss": 0.004658068064600229,
+      "epoch": 21.303956834532375,
+      "step": 23690
+    },
+    {
+      "epoch": 21.303956834532375,
+      "step": 23690,
+      "torque_loss": 0.18447761237621307
+    },
+    {
+      "epoch": 21.303956834532375,
+      "force_loss": 0.008792830631136894,
+      "step": 23690
+    },
+    {
+      "epoch": 21.31294964028777,
+      "grad_norm": 0.21590489149093628,
+      "learning_rate": 7.084159801789438e-05,
+      "loss": 0.0201,
+      "step": 23700
+    },
+    {
+      "action_loss": 0.0029024232644587755,
+      "epoch": 21.31294964028777,
+      "step": 23700
+    },
+    {
+      "epoch": 21.31294964028777,
+      "step": 23700,
+      "torque_loss": 0.11336324363946915
+    },
+    {
+      "epoch": 21.31294964028777,
+      "force_loss": 0.001504159881733358,
+      "step": 23700
+    },
+    {
+      "epoch": 21.321942446043167,
+      "grad_norm": 0.22023874521255493,
+      "learning_rate": 7.081654523884411e-05,
+      "loss": 0.0222,
+      "step": 23710
+    },
+    {
+      "action_loss": 0.003117077983915806,
+      "epoch": 21.321942446043167,
+      "step": 23710
+    },
+    {
+      "epoch": 21.321942446043167,
+      "step": 23710,
+      "torque_loss": 0.11545255780220032
+    },
+    {
+      "epoch": 21.321942446043167,
+      "force_loss": 0.0031235376372933388,
+      "step": 23710
+    },
+    {
+      "epoch": 21.33093525179856,
+      "grad_norm": 0.7098838686943054,
+      "learning_rate": 7.0791486136277e-05,
+      "loss": 0.0235,
+      "step": 23720
+    },
+    {
+      "action_loss": 0.004025460220873356,
+      "epoch": 21.33093525179856,
+      "step": 23720
+    },
+    {
+      "epoch": 21.33093525179856,
+      "step": 23720,
+      "torque_loss": 0.1383565217256546
+    },
+    {
+      "epoch": 21.33093525179856,
+      "force_loss": 0.003235556883737445,
+      "step": 23720
+    },
+    {
+      "epoch": 21.33992805755396,
+      "grad_norm": 0.29444053769111633,
+      "learning_rate": 7.07664207178054e-05,
+      "loss": 0.0237,
+      "step": 23730
+    },
+    {
+      "action_loss": 0.003561244113370776,
+      "epoch": 21.33992805755396,
+      "step": 23730
+    },
+    {
+      "epoch": 21.33992805755396,
+      "step": 23730,
+      "torque_loss": 0.14339353144168854
+    },
+    {
+      "epoch": 21.33992805755396,
+      "force_loss": 0.005450025200843811,
+      "step": 23730
+    },
+    {
+      "epoch": 21.348920863309353,
+      "grad_norm": 0.42307940125465393,
+      "learning_rate": 7.074134899104345e-05,
+      "loss": 0.0229,
+      "step": 23740
+    },
+    {
+      "action_loss": 0.003302047261968255,
+      "epoch": 21.348920863309353,
+      "step": 23740
+    },
+    {
+      "epoch": 21.348920863309353,
+      "step": 23740,
+      "torque_loss": 0.15255039930343628
+    },
+    {
+      "epoch": 21.348920863309353,
+      "force_loss": 0.0042688664980232716,
+      "step": 23740
+    },
+    {
+      "epoch": 21.357913669064747,
+      "grad_norm": 0.19251471757888794,
+      "learning_rate": 7.071627096360735e-05,
+      "loss": 0.019,
+      "step": 23750
+    },
+    {
+      "action_loss": 0.003922722768038511,
+      "epoch": 21.357913669064747,
+      "step": 23750
+    },
+    {
+      "epoch": 21.357913669064747,
+      "step": 23750,
+      "torque_loss": 0.17961324751377106
+    },
+    {
+      "epoch": 21.357913669064747,
+      "force_loss": 0.008094881661236286,
+      "step": 23750
+    },
+    {
+      "epoch": 21.366906474820144,
+      "grad_norm": 0.1816745102405548,
+      "learning_rate": 7.069118664311511e-05,
+      "loss": 0.0202,
+      "step": 23760
+    },
+    {
+      "action_loss": 0.006597679108381271,
+      "epoch": 21.366906474820144,
+      "step": 23760
+    },
+    {
+      "epoch": 21.366906474820144,
+      "step": 23760,
+      "torque_loss": 0.09441333264112473
+    },
+    {
+      "epoch": 21.366906474820144,
+      "force_loss": 0.002418153453618288,
+      "step": 23760
+    },
+    {
+      "epoch": 21.37589928057554,
+      "grad_norm": 0.29194214940071106,
+      "learning_rate": 7.06660960371867e-05,
+      "loss": 0.0191,
+      "step": 23770
+    },
+    {
+      "action_loss": 0.008289594203233719,
+      "epoch": 21.37589928057554,
+      "step": 23770
+    },
+    {
+      "epoch": 21.37589928057554,
+      "step": 23770,
+      "torque_loss": 0.13637441396713257
+    },
+    {
+      "epoch": 21.37589928057554,
+      "force_loss": 0.008279447443783283,
+      "step": 23770
+    },
+    {
+      "epoch": 21.384892086330936,
+      "grad_norm": 0.4021297097206116,
+      "learning_rate": 7.064099915344396e-05,
+      "loss": 0.025,
+      "step": 23780
+    },
+    {
+      "action_loss": 0.006939854472875595,
+      "epoch": 21.384892086330936,
+      "step": 23780
+    },
+    {
+      "epoch": 21.384892086330936,
+      "step": 23780,
+      "torque_loss": 0.12058931589126587
+    },
+    {
+      "epoch": 21.384892086330936,
+      "force_loss": 0.003337736474350095,
+      "step": 23780
+    },
+    {
+      "epoch": 21.39388489208633,
+      "grad_norm": 0.415184885263443,
+      "learning_rate": 7.061589599951066e-05,
+      "loss": 0.0225,
+      "step": 23790
+    },
+    {
+      "action_loss": 0.003426339477300644,
+      "epoch": 21.39388489208633,
+      "step": 23790
+    },
+    {
+      "epoch": 21.39388489208633,
+      "step": 23790,
+      "torque_loss": 0.1108291745185852
+    },
+    {
+      "epoch": 21.39388489208633,
+      "force_loss": 0.004369981121271849,
+      "step": 23790
+    },
+    {
+      "epoch": 21.402877697841728,
+      "grad_norm": 0.2257554829120636,
+      "learning_rate": 7.05907865830125e-05,
+      "loss": 0.0271,
+      "step": 23800
+    },
+    {
+      "action_loss": 0.00312385824508965,
+      "epoch": 21.402877697841728,
+      "step": 23800
+    },
+    {
+      "epoch": 21.402877697841728,
+      "step": 23800,
+      "torque_loss": 0.1279628425836563
+    },
+    {
+      "epoch": 21.402877697841728,
+      "force_loss": 0.004388665780425072,
+      "step": 23800
+    },
+    {
+      "epoch": 21.41187050359712,
+      "grad_norm": 0.389212429523468,
+      "learning_rate": 7.056567091157703e-05,
+      "loss": 0.0192,
+      "step": 23810
+    },
+    {
+      "action_loss": 0.010184607468545437,
+      "epoch": 21.41187050359712,
+      "step": 23810
+    },
+    {
+      "epoch": 21.41187050359712,
+      "step": 23810,
+      "torque_loss": 0.12354090064764023
+    },
+    {
+      "epoch": 21.41187050359712,
+      "force_loss": 0.0047220769338309765,
+      "step": 23810
+    },
+    {
+      "epoch": 21.42086330935252,
+      "grad_norm": 0.23627988994121552,
+      "learning_rate": 7.054054899283375e-05,
+      "loss": 0.0226,
+      "step": 23820
+    },
+    {
+      "action_loss": 0.0036707476247102022,
+      "epoch": 21.42086330935252,
+      "step": 23820
+    },
+    {
+      "epoch": 21.42086330935252,
+      "step": 23820,
+      "torque_loss": 0.12478425353765488
+    },
+    {
+      "epoch": 21.42086330935252,
+      "force_loss": 0.004485785495489836,
+      "step": 23820
+    },
+    {
+      "epoch": 21.429856115107913,
+      "grad_norm": 0.16367127001285553,
+      "learning_rate": 7.051542083441403e-05,
+      "loss": 0.0197,
+      "step": 23830
+    },
+    {
+      "action_loss": 0.0046365465968847275,
+      "epoch": 21.429856115107913,
+      "step": 23830
+    },
+    {
+      "epoch": 21.429856115107913,
+      "step": 23830,
+      "torque_loss": 0.11964243650436401
+    },
+    {
+      "epoch": 21.429856115107913,
+      "force_loss": 0.003397740190848708,
+      "step": 23830
+    },
+    {
+      "epoch": 21.43884892086331,
+      "grad_norm": 0.4269733428955078,
+      "learning_rate": 7.049028644395113e-05,
+      "loss": 0.0211,
+      "step": 23840
+    },
+    {
+      "action_loss": 0.002511439146474004,
+      "epoch": 21.43884892086331,
+      "step": 23840
+    },
+    {
+      "epoch": 21.43884892086331,
+      "step": 23840,
+      "torque_loss": 0.12440451234579086
+    },
+    {
+      "epoch": 21.43884892086331,
+      "force_loss": 0.005141694564372301,
+      "step": 23840
+    },
+    {
+      "epoch": 21.447841726618705,
+      "grad_norm": 0.18345722556114197,
+      "learning_rate": 7.046514582908024e-05,
+      "loss": 0.0174,
+      "step": 23850
+    },
+    {
+      "action_loss": 0.004589290823787451,
+      "epoch": 21.447841726618705,
+      "step": 23850
+    },
+    {
+      "epoch": 21.447841726618705,
+      "step": 23850,
+      "torque_loss": 0.12272452563047409
+    },
+    {
+      "epoch": 21.447841726618705,
+      "force_loss": 0.004211971070617437,
+      "step": 23850
+    },
+    {
+      "epoch": 21.4568345323741,
+      "grad_norm": 0.3202456533908844,
+      "learning_rate": 7.043999899743838e-05,
+      "loss": 0.0198,
+      "step": 23860
+    },
+    {
+      "action_loss": 0.010315182618796825,
+      "epoch": 21.4568345323741,
+      "step": 23860
+    },
+    {
+      "epoch": 21.4568345323741,
+      "step": 23860,
+      "torque_loss": 0.14328861236572266
+    },
+    {
+      "epoch": 21.4568345323741,
+      "force_loss": 0.01128691527992487,
+      "step": 23860
+    },
+    {
+      "epoch": 21.465827338129497,
+      "grad_norm": 0.18338774144649506,
+      "learning_rate": 7.041484595666451e-05,
+      "loss": 0.0243,
+      "step": 23870
+    },
+    {
+      "action_loss": 0.005853708367794752,
+      "epoch": 21.465827338129497,
+      "step": 23870
+    },
+    {
+      "epoch": 21.465827338129497,
+      "step": 23870,
+      "torque_loss": 0.14396049082279205
+    },
+    {
+      "epoch": 21.465827338129497,
+      "force_loss": 0.00822159182280302,
+      "step": 23870
+    },
+    {
+      "epoch": 21.47482014388489,
+      "grad_norm": 0.31921061873435974,
+      "learning_rate": 7.038968671439948e-05,
+      "loss": 0.0222,
+      "step": 23880
+    },
+    {
+      "action_loss": 0.008314737118780613,
+      "epoch": 21.47482014388489,
+      "step": 23880
+    },
+    {
+      "epoch": 21.47482014388489,
+      "step": 23880,
+      "torque_loss": 0.12615780532360077
+    },
+    {
+      "epoch": 21.47482014388489,
+      "force_loss": 0.010880435816943645,
+      "step": 23880
+    },
+    {
+      "epoch": 21.48381294964029,
+      "grad_norm": 0.18777592480182648,
+      "learning_rate": 7.036452127828596e-05,
+      "loss": 0.0225,
+      "step": 23890
+    },
+    {
+      "action_loss": 0.0036884669680148363,
+      "epoch": 21.48381294964029,
+      "step": 23890
+    },
+    {
+      "epoch": 21.48381294964029,
+      "step": 23890,
+      "torque_loss": 0.12386516481637955
+    },
+    {
+      "epoch": 21.48381294964029,
+      "force_loss": 0.0038495806511491537,
+      "step": 23890
+    },
+    {
+      "epoch": 21.492805755395683,
+      "grad_norm": 0.3598543703556061,
+      "learning_rate": 7.033934965596859e-05,
+      "loss": 0.0177,
+      "step": 23900
+    },
+    {
+      "action_loss": 0.010872217826545238,
+      "epoch": 21.492805755395683,
+      "step": 23900
+    },
+    {
+      "epoch": 21.492805755395683,
+      "step": 23900,
+      "torque_loss": 0.1956467181444168
+    },
+    {
+      "epoch": 21.492805755395683,
+      "force_loss": 0.010237191803753376,
+      "step": 23900
+    },
+    {
+      "epoch": 21.50179856115108,
+      "grad_norm": 0.19021499156951904,
+      "learning_rate": 7.031417185509381e-05,
+      "loss": 0.0211,
+      "step": 23910
+    },
+    {
+      "action_loss": 0.006348499562591314,
+      "epoch": 21.50179856115108,
+      "step": 23910
+    },
+    {
+      "epoch": 21.50179856115108,
+      "step": 23910,
+      "torque_loss": 0.12852859497070312
+    },
+    {
+      "epoch": 21.50179856115108,
+      "force_loss": 0.0033777987118810415,
+      "step": 23910
+    },
+    {
+      "epoch": 21.510791366906474,
+      "grad_norm": 0.22508279979228973,
+      "learning_rate": 7.028898788331e-05,
+      "loss": 0.0199,
+      "step": 23920
+    },
+    {
+      "action_loss": 0.05363455042243004,
+      "epoch": 21.510791366906474,
+      "step": 23920
+    },
+    {
+      "epoch": 21.510791366906474,
+      "step": 23920,
+      "torque_loss": 0.18643148243427277
+    },
+    {
+      "epoch": 21.510791366906474,
+      "force_loss": 0.06643527001142502,
+      "step": 23920
+    },
+    {
+      "epoch": 21.519784172661872,
+      "grad_norm": 0.13674558699131012,
+      "learning_rate": 7.026379774826736e-05,
+      "loss": 0.0272,
+      "step": 23930
+    },
+    {
+      "action_loss": 0.007402696181088686,
+      "epoch": 21.519784172661872,
+      "step": 23930
+    },
+    {
+      "epoch": 21.519784172661872,
+      "step": 23930,
+      "torque_loss": 0.16087555885314941
+    },
+    {
+      "epoch": 21.519784172661872,
+      "force_loss": 0.016275329515337944,
+      "step": 23930
+    },
+    {
+      "epoch": 21.528776978417266,
+      "grad_norm": 0.3391278088092804,
+      "learning_rate": 7.0238601457618e-05,
+      "loss": 0.0177,
+      "step": 23940
+    },
+    {
+      "action_loss": 0.014212819747626781,
+      "epoch": 21.528776978417266,
+      "step": 23940
+    },
+    {
+      "epoch": 21.528776978417266,
+      "step": 23940,
+      "torque_loss": 0.1407519429922104
+    },
+    {
+      "epoch": 21.528776978417266,
+      "force_loss": 0.008294091559946537,
+      "step": 23940
+    },
+    {
+      "epoch": 21.53776978417266,
+      "grad_norm": 0.24297048151493073,
+      "learning_rate": 7.02133990190159e-05,
+      "loss": 0.0201,
+      "step": 23950
+    },
+    {
+      "action_loss": 0.0034887150395661592,
+      "epoch": 21.53776978417266,
+      "step": 23950
+    },
+    {
+      "epoch": 21.53776978417266,
+      "step": 23950,
+      "torque_loss": 0.13968603312969208
+    },
+    {
+      "epoch": 21.53776978417266,
+      "force_loss": 0.002215509070083499,
+      "step": 23950
+    },
+    {
+      "epoch": 21.546762589928058,
+      "grad_norm": 0.3339236378669739,
+      "learning_rate": 7.018819044011687e-05,
+      "loss": 0.021,
+      "step": 23960
+    },
+    {
+      "action_loss": 0.005108047276735306,
+      "epoch": 21.546762589928058,
+      "step": 23960
+    },
+    {
+      "epoch": 21.546762589928058,
+      "step": 23960,
+      "torque_loss": 0.1285800188779831
+    },
+    {
+      "epoch": 21.546762589928058,
+      "force_loss": 0.00446784682571888,
+      "step": 23960
+    },
+    {
+      "epoch": 21.555755395683452,
+      "grad_norm": 0.3731296956539154,
+      "learning_rate": 7.016297572857863e-05,
+      "loss": 0.0188,
+      "step": 23970
+    },
+    {
+      "action_loss": 0.0029835563618689775,
+      "epoch": 21.555755395683452,
+      "step": 23970
+    },
+    {
+      "epoch": 21.555755395683452,
+      "step": 23970,
+      "torque_loss": 0.11705994606018066
+    },
+    {
+      "epoch": 21.555755395683452,
+      "force_loss": 0.0021975317504256964,
+      "step": 23970
+    },
+    {
+      "epoch": 21.56474820143885,
+      "grad_norm": 0.212856262922287,
+      "learning_rate": 7.013775489206072e-05,
+      "loss": 0.0208,
+      "step": 23980
+    },
+    {
+      "action_loss": 0.010328707285225391,
+      "epoch": 21.56474820143885,
+      "step": 23980
+    },
+    {
+      "epoch": 21.56474820143885,
+      "step": 23980,
+      "torque_loss": 0.11490169167518616
+    },
+    {
+      "epoch": 21.56474820143885,
+      "force_loss": 0.012079688720405102,
+      "step": 23980
+    },
+    {
+      "epoch": 21.573741007194243,
+      "grad_norm": 0.29255232214927673,
+      "learning_rate": 7.01125279382246e-05,
+      "loss": 0.0175,
+      "step": 23990
+    },
+    {
+      "action_loss": 0.013218914158642292,
+      "epoch": 21.573741007194243,
+      "step": 23990
+    },
+    {
+      "epoch": 21.573741007194243,
+      "step": 23990,
+      "torque_loss": 0.16297321021556854
+    },
+    {
+      "epoch": 21.573741007194243,
+      "force_loss": 0.013694614171981812,
+      "step": 23990
+    },
+    {
+      "epoch": 21.58273381294964,
+      "grad_norm": 0.39558789134025574,
+      "learning_rate": 7.008729487473351e-05,
+      "loss": 0.0211,
+      "step": 24000
+    },
+    {
+      "action_loss": 0.005202788393944502,
+      "epoch": 21.58273381294964,
+      "step": 24000
+    },
+    {
+      "epoch": 21.58273381294964,
+      "step": 24000,
+      "torque_loss": 0.12038284540176392
+    },
+    {
+      "epoch": 21.58273381294964,
+      "force_loss": 0.008069656789302826,
+      "step": 24000
+    },
+    {
+      "epoch": 21.591726618705035,
+      "grad_norm": 0.3813171684741974,
+      "learning_rate": 7.006205570925263e-05,
+      "loss": 0.0233,
+      "step": 24010
+    },
+    {
+      "action_loss": 0.00404118187725544,
+      "epoch": 21.591726618705035,
+      "step": 24010
+    },
+    {
+      "epoch": 21.591726618705035,
+      "step": 24010,
+      "torque_loss": 0.15226443111896515
+    },
+    {
+      "epoch": 21.591726618705035,
+      "force_loss": 0.0024941328447312117,
+      "step": 24010
+    },
+    {
+      "epoch": 21.600719424460433,
+      "grad_norm": 0.2981605529785156,
+      "learning_rate": 7.003681044944892e-05,
+      "loss": 0.0205,
+      "step": 24020
+    },
+    {
+      "action_loss": 0.0040172552689909935,
+      "epoch": 21.600719424460433,
+      "step": 24020
+    },
+    {
+      "epoch": 21.600719424460433,
+      "step": 24020,
+      "torque_loss": 0.11741326004266739
+    },
+    {
+      "epoch": 21.600719424460433,
+      "force_loss": 0.0046316650696098804,
+      "step": 24020
+    },
+    {
+      "epoch": 21.609712230215827,
+      "grad_norm": 0.4226076304912567,
+      "learning_rate": 7.001155910299126e-05,
+      "loss": 0.0216,
+      "step": 24030
+    },
+    {
+      "action_loss": 0.017753953114151955,
+      "epoch": 21.609712230215827,
+      "step": 24030
+    },
+    {
+      "epoch": 21.609712230215827,
+      "step": 24030,
+      "torque_loss": 0.1803961545228958
+    },
+    {
+      "epoch": 21.609712230215827,
+      "force_loss": 0.02056589350104332,
+      "step": 24030
+    },
+    {
+      "epoch": 21.618705035971225,
+      "grad_norm": 0.2300824373960495,
+      "learning_rate": 6.99863016775503e-05,
+      "loss": 0.0261,
+      "step": 24040
+    },
+    {
+      "action_loss": 0.006227040663361549,
+      "epoch": 21.618705035971225,
+      "step": 24040
+    },
+    {
+      "epoch": 21.618705035971225,
+      "step": 24040,
+      "torque_loss": 0.11970778554677963
+    },
+    {
+      "epoch": 21.618705035971225,
+      "force_loss": 0.009214621037244797,
+      "step": 24040
+    },
+    {
+      "epoch": 21.62769784172662,
+      "grad_norm": 0.40676695108413696,
+      "learning_rate": 6.996103818079859e-05,
+      "loss": 0.0196,
+      "step": 24050
+    },
+    {
+      "action_loss": 0.008792642503976822,
+      "epoch": 21.62769784172662,
+      "step": 24050
+    },
+    {
+      "epoch": 21.62769784172662,
+      "step": 24050,
+      "torque_loss": 0.10972053557634354
+    },
+    {
+      "epoch": 21.62769784172662,
+      "force_loss": 0.005656917113810778,
+      "step": 24050
+    },
+    {
+      "epoch": 21.636690647482013,
+      "grad_norm": 0.506324291229248,
+      "learning_rate": 6.993576862041054e-05,
+      "loss": 0.0218,
+      "step": 24060
+    },
+    {
+      "action_loss": 0.0061348178423941135,
+      "epoch": 21.636690647482013,
+      "step": 24060
+    },
+    {
+      "epoch": 21.636690647482013,
+      "step": 24060,
+      "torque_loss": 0.13718056678771973
+    },
+    {
+      "epoch": 21.636690647482013,
+      "force_loss": 0.00756865506991744,
+      "step": 24060
+    },
+    {
+      "epoch": 21.64568345323741,
+      "grad_norm": 0.4522571563720703,
+      "learning_rate": 6.991049300406235e-05,
+      "loss": 0.0179,
+      "step": 24070
+    },
+    {
+      "action_loss": 0.005923125892877579,
+      "epoch": 21.64568345323741,
+      "step": 24070
+    },
+    {
+      "epoch": 21.64568345323741,
+      "step": 24070,
+      "torque_loss": 0.14005731046199799
+    },
+    {
+      "epoch": 21.64568345323741,
+      "force_loss": 0.0036965280305594206,
+      "step": 24070
+    },
+    {
+      "epoch": 21.654676258992804,
+      "grad_norm": 0.21875688433647156,
+      "learning_rate": 6.988521133943209e-05,
+      "loss": 0.0202,
+      "step": 24080
+    },
+    {
+      "action_loss": 0.017261147499084473,
+      "epoch": 21.654676258992804,
+      "step": 24080
+    },
+    {
+      "epoch": 21.654676258992804,
+      "step": 24080,
+      "torque_loss": 0.1414811760187149
+    },
+    {
+      "epoch": 21.654676258992804,
+      "force_loss": 0.01806695945560932,
+      "step": 24080
+    },
+    {
+      "epoch": 21.663669064748202,
+      "grad_norm": 0.5208433866500854,
+      "learning_rate": 6.985992363419966e-05,
+      "loss": 0.0245,
+      "step": 24090
+    },
+    {
+      "action_loss": 0.0029069986194372177,
+      "epoch": 21.663669064748202,
+      "step": 24090
+    },
+    {
+      "epoch": 21.663669064748202,
+      "step": 24090,
+      "torque_loss": 0.09179780632257462
+    },
+    {
+      "epoch": 21.663669064748202,
+      "force_loss": 0.001528251334093511,
+      "step": 24090
+    },
+    {
+      "epoch": 21.672661870503596,
+      "grad_norm": 0.3127578794956207,
+      "learning_rate": 6.983462989604682e-05,
+      "loss": 0.021,
+      "step": 24100
+    },
+    {
+      "action_loss": 0.0032168340403586626,
+      "epoch": 21.672661870503596,
+      "step": 24100
+    },
+    {
+      "epoch": 21.672661870503596,
+      "step": 24100,
+      "torque_loss": 0.12450555711984634
+    },
+    {
+      "epoch": 21.672661870503596,
+      "force_loss": 0.002190846251323819,
+      "step": 24100
+    },
+    {
+      "epoch": 21.681654676258994,
+      "grad_norm": 0.3800390660762787,
+      "learning_rate": 6.980933013265709e-05,
+      "loss": 0.0177,
+      "step": 24110
+    },
+    {
+      "action_loss": 0.006540060043334961,
+      "epoch": 21.681654676258994,
+      "step": 24110
+    },
+    {
+      "epoch": 21.681654676258994,
+      "step": 24110,
+      "torque_loss": 0.13025234639644623
+    },
+    {
+      "epoch": 21.681654676258994,
+      "force_loss": 0.009451337158679962,
+      "step": 24110
+    },
+    {
+      "epoch": 21.690647482014388,
+      "grad_norm": 0.4404729902744293,
+      "learning_rate": 6.978402435171592e-05,
+      "loss": 0.0303,
+      "step": 24120
+    },
+    {
+      "action_loss": 0.007622348610311747,
+      "epoch": 21.690647482014388,
+      "step": 24120
+    },
+    {
+      "epoch": 21.690647482014388,
+      "step": 24120,
+      "torque_loss": 0.19697785377502441
+    },
+    {
+      "epoch": 21.690647482014388,
+      "force_loss": 0.01301037147641182,
+      "step": 24120
+    },
+    {
+      "epoch": 21.699640287769785,
+      "grad_norm": 0.25689589977264404,
+      "learning_rate": 6.975871256091052e-05,
+      "loss": 0.0232,
+      "step": 24130
+    },
+    {
+      "action_loss": 0.0025068081449717283,
+      "epoch": 21.699640287769785,
+      "step": 24130
+    },
+    {
+      "epoch": 21.699640287769785,
+      "step": 24130,
+      "torque_loss": 0.12582151591777802
+    },
+    {
+      "epoch": 21.699640287769785,
+      "force_loss": 0.002489696955308318,
+      "step": 24130
+    },
+    {
+      "epoch": 21.70863309352518,
+      "grad_norm": 0.24892547726631165,
+      "learning_rate": 6.973339476792995e-05,
+      "loss": 0.0201,
+      "step": 24140
+    },
+    {
+      "action_loss": 0.005375228356570005,
+      "epoch": 21.70863309352518,
+      "step": 24140
+    },
+    {
+      "epoch": 21.70863309352518,
+      "step": 24140,
+      "torque_loss": 0.18812493979930878
+    },
+    {
+      "epoch": 21.70863309352518,
+      "force_loss": 0.004086604341864586,
+      "step": 24140
+    },
+    {
+      "epoch": 21.717625899280577,
+      "grad_norm": 0.28377944231033325,
+      "learning_rate": 6.970807098046505e-05,
+      "loss": 0.0191,
+      "step": 24150
+    },
+    {
+      "action_loss": 0.0025983082596212626,
+      "epoch": 21.717625899280577,
+      "step": 24150
+    },
+    {
+      "epoch": 21.717625899280577,
+      "step": 24150,
+      "torque_loss": 0.12086061388254166
+    },
+    {
+      "epoch": 21.717625899280577,
+      "force_loss": 0.0032244424801319838,
+      "step": 24150
+    },
+    {
+      "epoch": 21.72661870503597,
+      "grad_norm": 0.4414898157119751,
+      "learning_rate": 6.968274120620858e-05,
+      "loss": 0.0214,
+      "step": 24160
+    },
+    {
+      "action_loss": 0.002598340855911374,
+      "epoch": 21.72661870503597,
+      "step": 24160
+    },
+    {
+      "epoch": 21.72661870503597,
+      "step": 24160,
+      "torque_loss": 0.15716147422790527
+    },
+    {
+      "epoch": 21.72661870503597,
+      "force_loss": 0.0021320506930351257,
+      "step": 24160
+    },
+    {
+      "epoch": 21.735611510791365,
+      "grad_norm": 0.23562584817409515,
+      "learning_rate": 6.965740545285499e-05,
+      "loss": 0.0184,
+      "step": 24170
+    },
+    {
+      "action_loss": 0.0031706802546977997,
+      "epoch": 21.735611510791365,
+      "step": 24170
+    },
+    {
+      "epoch": 21.735611510791365,
+      "step": 24170,
+      "torque_loss": 0.1128600463271141
+    },
+    {
+      "epoch": 21.735611510791365,
+      "force_loss": 0.004506824538111687,
+      "step": 24170
+    },
+    {
+      "epoch": 21.744604316546763,
+      "grad_norm": 0.16314929723739624,
+      "learning_rate": 6.963206372810068e-05,
+      "loss": 0.0187,
+      "step": 24180
+    },
+    {
+      "action_loss": 0.008436471223831177,
+      "epoch": 21.744604316546763,
+      "step": 24180
+    },
+    {
+      "epoch": 21.744604316546763,
+      "step": 24180,
+      "torque_loss": 0.11788768321275711
+    },
+    {
+      "epoch": 21.744604316546763,
+      "force_loss": 0.004664561245590448,
+      "step": 24180
+    },
+    {
+      "epoch": 21.753597122302157,
+      "grad_norm": 0.2003936767578125,
+      "learning_rate": 6.960671603964375e-05,
+      "loss": 0.023,
+      "step": 24190
+    },
+    {
+      "action_loss": 0.03077048249542713,
+      "epoch": 21.753597122302157,
+      "step": 24190
+    },
+    {
+      "epoch": 21.753597122302157,
+      "step": 24190,
+      "torque_loss": 0.13748450577259064
+    },
+    {
+      "epoch": 21.753597122302157,
+      "force_loss": 0.04616479203104973,
+      "step": 24190
+    },
+    {
+      "epoch": 21.762589928057555,
+      "grad_norm": 0.257639080286026,
+      "learning_rate": 6.958136239518418e-05,
+      "loss": 0.0212,
+      "step": 24200
+    },
+    {
+      "action_loss": 0.0019250352634117007,
+      "epoch": 21.762589928057555,
+      "step": 24200
+    },
+    {
+      "epoch": 21.762589928057555,
+      "step": 24200,
+      "torque_loss": 0.09396030753850937
+    },
+    {
+      "epoch": 21.762589928057555,
+      "force_loss": 0.002093286719173193,
+      "step": 24200
+    },
+    {
+      "epoch": 21.77158273381295,
+      "grad_norm": 0.23291504383087158,
+      "learning_rate": 6.955600280242371e-05,
+      "loss": 0.0197,
+      "step": 24210
+    },
+    {
+      "action_loss": 0.005376171320676804,
+      "epoch": 21.77158273381295,
+      "step": 24210
+    },
+    {
+      "epoch": 21.77158273381295,
+      "step": 24210,
+      "torque_loss": 0.1337880641222
+    },
+    {
+      "epoch": 21.77158273381295,
+      "force_loss": 0.00854805763810873,
+      "step": 24210
+    },
+    {
+      "epoch": 21.780575539568346,
+      "grad_norm": 0.3987133204936981,
+      "learning_rate": 6.953063726906596e-05,
+      "loss": 0.0204,
+      "step": 24220
+    },
+    {
+      "action_loss": 0.00652548810467124,
+      "epoch": 21.780575539568346,
+      "step": 24220
+    },
+    {
+      "epoch": 21.780575539568346,
+      "step": 24220,
+      "torque_loss": 0.15214596688747406
+    },
+    {
+      "epoch": 21.780575539568346,
+      "force_loss": 0.011250283569097519,
+      "step": 24220
+    },
+    {
+      "epoch": 21.78956834532374,
+      "grad_norm": 0.31795546412467957,
+      "learning_rate": 6.950526580281626e-05,
+      "loss": 0.0202,
+      "step": 24230
+    },
+    {
+      "action_loss": 0.0022861079778522253,
+      "epoch": 21.78956834532374,
+      "step": 24230
+    },
+    {
+      "epoch": 21.78956834532374,
+      "step": 24230,
+      "torque_loss": 0.1178806722164154
+    },
+    {
+      "epoch": 21.78956834532374,
+      "force_loss": 0.003101978451013565,
+      "step": 24230
+    },
+    {
+      "epoch": 21.798561151079138,
+      "grad_norm": 0.2745036780834198,
+      "learning_rate": 6.947988841138184e-05,
+      "loss": 0.022,
+      "step": 24240
+    },
+    {
+      "action_loss": 0.0025741548743098974,
+      "epoch": 21.798561151079138,
+      "step": 24240
+    },
+    {
+      "epoch": 21.798561151079138,
+      "step": 24240,
+      "torque_loss": 0.10255517810583115
+    },
+    {
+      "epoch": 21.798561151079138,
+      "force_loss": 0.002620341954752803,
+      "step": 24240
+    },
+    {
+      "epoch": 21.807553956834532,
+      "grad_norm": 0.3730088174343109,
+      "learning_rate": 6.945450510247165e-05,
+      "loss": 0.0207,
+      "step": 24250
+    },
+    {
+      "action_loss": 0.0075866542756557465,
+      "epoch": 21.807553956834532,
+      "step": 24250
+    },
+    {
+      "epoch": 21.807553956834532,
+      "step": 24250,
+      "torque_loss": 0.1519446223974228
+    },
+    {
+      "epoch": 21.807553956834532,
+      "force_loss": 0.004284977447241545,
+      "step": 24250
+    },
+    {
+      "epoch": 21.81654676258993,
+      "grad_norm": 0.3975403606891632,
+      "learning_rate": 6.942911588379647e-05,
+      "loss": 0.0207,
+      "step": 24260
+    },
+    {
+      "action_loss": 0.004833179991692305,
+      "epoch": 21.81654676258993,
+      "step": 24260
+    },
+    {
+      "epoch": 21.81654676258993,
+      "step": 24260,
+      "torque_loss": 0.12231109291315079
+    },
+    {
+      "epoch": 21.81654676258993,
+      "force_loss": 0.007674244698137045,
+      "step": 24260
+    },
+    {
+      "epoch": 21.825539568345324,
+      "grad_norm": 0.2681105434894562,
+      "learning_rate": 6.940372076306888e-05,
+      "loss": 0.0201,
+      "step": 24270
+    },
+    {
+      "action_loss": 0.005627890583127737,
+      "epoch": 21.825539568345324,
+      "step": 24270
+    },
+    {
+      "epoch": 21.825539568345324,
+      "step": 24270,
+      "torque_loss": 0.18219177424907684
+    },
+    {
+      "epoch": 21.825539568345324,
+      "force_loss": 0.0068586901761591434,
+      "step": 24270
+    },
+    {
+      "epoch": 21.834532374100718,
+      "grad_norm": 0.14424137771129608,
+      "learning_rate": 6.937831974800326e-05,
+      "loss": 0.0202,
+      "step": 24280
+    },
+    {
+      "action_loss": 0.004893868695944548,
+      "epoch": 21.834532374100718,
+      "step": 24280
+    },
+    {
+      "epoch": 21.834532374100718,
+      "step": 24280,
+      "torque_loss": 0.10660874843597412
+    },
+    {
+      "epoch": 21.834532374100718,
+      "force_loss": 0.004262237343937159,
+      "step": 24280
+    },
+    {
+      "epoch": 21.843525179856115,
+      "grad_norm": 0.7608571648597717,
+      "learning_rate": 6.935291284631574e-05,
+      "loss": 0.0238,
+      "step": 24290
+    },
+    {
+      "action_loss": 0.015205886214971542,
+      "epoch": 21.843525179856115,
+      "step": 24290
+    },
+    {
+      "epoch": 21.843525179856115,
+      "step": 24290,
+      "torque_loss": 0.11672701686620712
+    },
+    {
+      "epoch": 21.843525179856115,
+      "force_loss": 0.017855575308203697,
+      "step": 24290
+    },
+    {
+      "epoch": 21.85251798561151,
+      "grad_norm": 0.6793885231018066,
+      "learning_rate": 6.932750006572428e-05,
+      "loss": 0.019,
+      "step": 24300
+    },
+    {
+      "action_loss": 0.005090448074042797,
+      "epoch": 21.85251798561151,
+      "step": 24300
+    },
+    {
+      "epoch": 21.85251798561151,
+      "step": 24300,
+      "torque_loss": 0.1350153088569641
+    },
+    {
+      "epoch": 21.85251798561151,
+      "force_loss": 0.010580133646726608,
+      "step": 24300
+    },
+    {
+      "epoch": 21.861510791366907,
+      "grad_norm": 0.25309425592422485,
+      "learning_rate": 6.930208141394863e-05,
+      "loss": 0.0367,
+      "step": 24310
+    },
+    {
+      "action_loss": 0.007332662586122751,
+      "epoch": 21.861510791366907,
+      "step": 24310
+    },
+    {
+      "epoch": 21.861510791366907,
+      "step": 24310,
+      "torque_loss": 0.15832073986530304
+    },
+    {
+      "epoch": 21.861510791366907,
+      "force_loss": 0.013494604267179966,
+      "step": 24310
+    },
+    {
+      "epoch": 21.8705035971223,
+      "grad_norm": 0.42558208107948303,
+      "learning_rate": 6.927665689871026e-05,
+      "loss": 0.0202,
+      "step": 24320
+    },
+    {
+      "action_loss": 0.012143700383603573,
+      "epoch": 21.8705035971223,
+      "step": 24320
+    },
+    {
+      "epoch": 21.8705035971223,
+      "step": 24320,
+      "torque_loss": 0.14808912575244904
+    },
+    {
+      "epoch": 21.8705035971223,
+      "force_loss": 0.008264814503490925,
+      "step": 24320
+    },
+    {
+      "epoch": 21.8794964028777,
+      "grad_norm": 0.3452666103839874,
+      "learning_rate": 6.925122652773253e-05,
+      "loss": 0.0216,
+      "step": 24330
+    },
+    {
+      "action_loss": 0.0038850188720971346,
+      "epoch": 21.8794964028777,
+      "step": 24330
+    },
+    {
+      "epoch": 21.8794964028777,
+      "step": 24330,
+      "torque_loss": 0.14941705763339996
+    },
+    {
+      "epoch": 21.8794964028777,
+      "force_loss": 0.004443696234375238,
+      "step": 24330
+    },
+    {
+      "epoch": 21.888489208633093,
+      "grad_norm": 0.20415908098220825,
+      "learning_rate": 6.922579030874046e-05,
+      "loss": 0.0212,
+      "step": 24340
+    },
+    {
+      "action_loss": 0.011489741504192352,
+      "epoch": 21.888489208633093,
+      "step": 24340
+    },
+    {
+      "epoch": 21.888489208633093,
+      "step": 24340,
+      "torque_loss": 0.19305045902729034
+    },
+    {
+      "epoch": 21.888489208633093,
+      "force_loss": 0.02003069408237934,
+      "step": 24340
+    },
+    {
+      "epoch": 21.89748201438849,
+      "grad_norm": 0.29833388328552246,
+      "learning_rate": 6.920034824946093e-05,
+      "loss": 0.025,
+      "step": 24350
+    },
+    {
+      "action_loss": 0.005486502777785063,
+      "epoch": 21.89748201438849,
+      "step": 24350
+    },
+    {
+      "epoch": 21.89748201438849,
+      "step": 24350,
+      "torque_loss": 0.16189566254615784
+    },
+    {
+      "epoch": 21.89748201438849,
+      "force_loss": 0.003585736034438014,
+      "step": 24350
+    },
+    {
+      "epoch": 21.906474820143885,
+      "grad_norm": 0.23319552838802338,
+      "learning_rate": 6.917490035762255e-05,
+      "loss": 0.023,
+      "step": 24360
+    },
+    {
+      "action_loss": 0.024321740493178368,
+      "epoch": 21.906474820143885,
+      "step": 24360
+    },
+    {
+      "epoch": 21.906474820143885,
+      "step": 24360,
+      "torque_loss": 0.20250184834003448
+    },
+    {
+      "epoch": 21.906474820143885,
+      "force_loss": 0.04470481351017952,
+      "step": 24360
+    },
+    {
+      "epoch": 21.915467625899282,
+      "grad_norm": 0.2829186022281647,
+      "learning_rate": 6.914944664095573e-05,
+      "loss": 0.0281,
+      "step": 24370
+    },
+    {
+      "action_loss": 0.004660788923501968,
+      "epoch": 21.915467625899282,
+      "step": 24370
+    },
+    {
+      "epoch": 21.915467625899282,
+      "step": 24370,
+      "torque_loss": 0.14200960099697113
+    },
+    {
+      "epoch": 21.915467625899282,
+      "force_loss": 0.004201762843877077,
+      "step": 24370
+    },
+    {
+      "epoch": 21.924460431654676,
+      "grad_norm": 0.5323184132575989,
+      "learning_rate": 6.912398710719264e-05,
+      "loss": 0.0226,
+      "step": 24380
+    },
+    {
+      "action_loss": 0.0049565755762159824,
+      "epoch": 21.924460431654676,
+      "step": 24380
+    },
+    {
+      "epoch": 21.924460431654676,
+      "step": 24380,
+      "torque_loss": 0.15860824286937714
+    },
+    {
+      "epoch": 21.924460431654676,
+      "force_loss": 0.006551635917276144,
+      "step": 24380
+    },
+    {
+      "epoch": 21.93345323741007,
+      "grad_norm": 0.26677796244621277,
+      "learning_rate": 6.90985217640672e-05,
+      "loss": 0.0189,
+      "step": 24390
+    },
+    {
+      "action_loss": 0.004493407439440489,
+      "epoch": 21.93345323741007,
+      "step": 24390
+    },
+    {
+      "epoch": 21.93345323741007,
+      "step": 24390,
+      "torque_loss": 0.10732502490282059
+    },
+    {
+      "epoch": 21.93345323741007,
+      "force_loss": 0.002163970610126853,
+      "step": 24390
+    },
+    {
+      "epoch": 21.942446043165468,
+      "grad_norm": 0.6263008117675781,
+      "learning_rate": 6.90730506193151e-05,
+      "loss": 0.0201,
+      "step": 24400
+    },
+    {
+      "action_loss": 0.005693149287253618,
+      "epoch": 21.942446043165468,
+      "step": 24400
+    },
+    {
+      "epoch": 21.942446043165468,
+      "step": 24400,
+      "torque_loss": 0.12757854163646698
+    },
+    {
+      "epoch": 21.942446043165468,
+      "force_loss": 0.0025026879739016294,
+      "step": 24400
+    },
+    {
+      "epoch": 21.951438848920862,
+      "grad_norm": 0.3909143805503845,
+      "learning_rate": 6.904757368067384e-05,
+      "loss": 0.023,
+      "step": 24410
+    },
+    {
+      "action_loss": 0.0076774596236646175,
+      "epoch": 21.951438848920862,
+      "step": 24410
+    },
+    {
+      "epoch": 21.951438848920862,
+      "step": 24410,
+      "torque_loss": 0.09612607955932617
+    },
+    {
+      "epoch": 21.951438848920862,
+      "force_loss": 0.01226059254258871,
+      "step": 24410
+    },
+    {
+      "epoch": 21.96043165467626,
+      "grad_norm": 0.26202166080474854,
+      "learning_rate": 6.90220909558826e-05,
+      "loss": 0.0251,
+      "step": 24420
+    },
+    {
+      "action_loss": 0.010712314397096634,
+      "epoch": 21.96043165467626,
+      "step": 24420
+    },
+    {
+      "epoch": 21.96043165467626,
+      "step": 24420,
+      "torque_loss": 0.12895852327346802
+    },
+    {
+      "epoch": 21.96043165467626,
+      "force_loss": 0.00807793065905571,
+      "step": 24420
+    },
+    {
+      "epoch": 21.969424460431654,
+      "grad_norm": 0.3353864252567291,
+      "learning_rate": 6.899660245268237e-05,
+      "loss": 0.0213,
+      "step": 24430
+    },
+    {
+      "action_loss": 0.0037603459786623716,
+      "epoch": 21.969424460431654,
+      "step": 24430
+    },
+    {
+      "epoch": 21.969424460431654,
+      "step": 24430,
+      "torque_loss": 0.1198287382721901
+    },
+    {
+      "epoch": 21.969424460431654,
+      "force_loss": 0.004271873738616705,
+      "step": 24430
+    },
+    {
+      "epoch": 21.97841726618705,
+      "grad_norm": 0.11434774845838547,
+      "learning_rate": 6.897110817881592e-05,
+      "loss": 0.0214,
+      "step": 24440
+    },
+    {
+      "action_loss": 0.006766705308109522,
+      "epoch": 21.97841726618705,
+      "step": 24440
+    },
+    {
+      "epoch": 21.97841726618705,
+      "step": 24440,
+      "torque_loss": 0.1438988894224167
+    },
+    {
+      "epoch": 21.97841726618705,
+      "force_loss": 0.006048687268048525,
+      "step": 24440
+    },
+    {
+      "epoch": 21.987410071942445,
+      "grad_norm": 0.6172817945480347,
+      "learning_rate": 6.894560814202769e-05,
+      "loss": 0.0229,
+      "step": 24450
+    },
+    {
+      "action_loss": 0.0024692120496183634,
+      "epoch": 21.987410071942445,
+      "step": 24450
+    },
+    {
+      "epoch": 21.987410071942445,
+      "step": 24450,
+      "torque_loss": 0.11306580156087875
+    },
+    {
+      "epoch": 21.987410071942445,
+      "force_loss": 0.0031304650474339724,
+      "step": 24450
+    },
+    {
+      "epoch": 21.996402877697843,
+      "grad_norm": 0.4427724778652191,
+      "learning_rate": 6.892010235006394e-05,
+      "loss": 0.0268,
+      "step": 24460
+    },
+    {
+      "action_loss": 0.003611071268096566,
+      "epoch": 21.996402877697843,
+      "step": 24460
+    },
+    {
+      "epoch": 21.996402877697843,
+      "step": 24460,
+      "torque_loss": 0.1440209299325943
+    },
+    {
+      "epoch": 21.996402877697843,
+      "force_loss": 0.003615929977968335,
+      "step": 24460
+    },
+    {
+      "epoch": 22.005395683453237,
+      "grad_norm": 0.21930928528308868,
+      "learning_rate": 6.889459081067264e-05,
+      "loss": 0.0204,
+      "step": 24470
+    },
+    {
+      "action_loss": 0.0037955709267407656,
+      "epoch": 22.005395683453237,
+      "step": 24470
+    },
+    {
+      "epoch": 22.005395683453237,
+      "step": 24470,
+      "torque_loss": 0.14295008778572083
+    },
+    {
+      "epoch": 22.005395683453237,
+      "force_loss": 0.004967763088643551,
+      "step": 24470
+    },
+    {
+      "epoch": 22.014388489208635,
+      "grad_norm": 0.43003565073013306,
+      "learning_rate": 6.886907353160356e-05,
+      "loss": 0.0202,
+      "step": 24480
+    },
+    {
+      "action_loss": 0.007930673658847809,
+      "epoch": 22.014388489208635,
+      "step": 24480
+    },
+    {
+      "epoch": 22.014388489208635,
+      "step": 24480,
+      "torque_loss": 0.14471480250358582
+    },
+    {
+      "epoch": 22.014388489208635,
+      "force_loss": 0.00403394503518939,
+      "step": 24480
+    },
+    {
+      "epoch": 22.02338129496403,
+      "grad_norm": 0.31343305110931396,
+      "learning_rate": 6.884355052060814e-05,
+      "loss": 0.0187,
+      "step": 24490
+    },
+    {
+      "action_loss": 0.00326588936150074,
+      "epoch": 22.02338129496403,
+      "step": 24490
+    },
+    {
+      "epoch": 22.02338129496403,
+      "step": 24490,
+      "torque_loss": 0.09171837568283081
+    },
+    {
+      "epoch": 22.02338129496403,
+      "force_loss": 0.005965802818536758,
+      "step": 24490
+    },
+    {
+      "epoch": 22.032374100719423,
+      "grad_norm": 0.2788020074367523,
+      "learning_rate": 6.88180217854396e-05,
+      "loss": 0.0197,
+      "step": 24500
+    },
+    {
+      "action_loss": 0.0050150505267083645,
+      "epoch": 22.032374100719423,
+      "step": 24500
+    },
+    {
+      "epoch": 22.032374100719423,
+      "step": 24500,
+      "torque_loss": 0.14749211072921753
+    },
+    {
+      "epoch": 22.032374100719423,
+      "force_loss": 0.002626863308250904,
+      "step": 24500
+    },
+    {
+      "epoch": 22.04136690647482,
+      "grad_norm": 0.32771310210227966,
+      "learning_rate": 6.87924873338529e-05,
+      "loss": 0.0231,
+      "step": 24510
+    },
+    {
+      "action_loss": 0.006207337602972984,
+      "epoch": 22.04136690647482,
+      "step": 24510
+    },
+    {
+      "epoch": 22.04136690647482,
+      "step": 24510,
+      "torque_loss": 0.11077922582626343
+    },
+    {
+      "epoch": 22.04136690647482,
+      "force_loss": 0.0047624483704566956,
+      "step": 24510
+    },
+    {
+      "epoch": 22.050359712230215,
+      "grad_norm": 0.4554392099380493,
+      "learning_rate": 6.876694717360475e-05,
+      "loss": 0.0217,
+      "step": 24520
+    },
+    {
+      "action_loss": 0.02835608273744583,
+      "epoch": 22.050359712230215,
+      "step": 24520
+    },
+    {
+      "epoch": 22.050359712230215,
+      "step": 24520,
+      "torque_loss": 0.12670385837554932
+    },
+    {
+      "epoch": 22.050359712230215,
+      "force_loss": 0.015614225529134274,
+      "step": 24520
+    },
+    {
+      "epoch": 22.059352517985612,
+      "grad_norm": 0.5012367963790894,
+      "learning_rate": 6.874140131245355e-05,
+      "loss": 0.0224,
+      "step": 24530
+    },
+    {
+      "action_loss": 0.012275311164557934,
+      "epoch": 22.059352517985612,
+      "step": 24530
+    },
+    {
+      "epoch": 22.059352517985612,
+      "step": 24530,
+      "torque_loss": 0.15115399658679962
+    },
+    {
+      "epoch": 22.059352517985612,
+      "force_loss": 0.01023236196488142,
+      "step": 24530
+    },
+    {
+      "epoch": 22.068345323741006,
+      "grad_norm": 0.2496970146894455,
+      "learning_rate": 6.871584975815948e-05,
+      "loss": 0.0217,
+      "step": 24540
+    },
+    {
+      "action_loss": 0.0030946796759963036,
+      "epoch": 22.068345323741006,
+      "step": 24540
+    },
+    {
+      "epoch": 22.068345323741006,
+      "step": 24540,
+      "torque_loss": 0.10134420543909073
+    },
+    {
+      "epoch": 22.068345323741006,
+      "force_loss": 0.0022605503909289837,
+      "step": 24540
+    },
+    {
+      "epoch": 22.077338129496404,
+      "grad_norm": 0.14811429381370544,
+      "learning_rate": 6.86902925184844e-05,
+      "loss": 0.0193,
+      "step": 24550
+    },
+    {
+      "action_loss": 0.01476606447249651,
+      "epoch": 22.077338129496404,
+      "step": 24550
+    },
+    {
+      "epoch": 22.077338129496404,
+      "step": 24550,
+      "torque_loss": 0.1842147558927536
+    },
+    {
+      "epoch": 22.077338129496404,
+      "force_loss": 0.013570576906204224,
+      "step": 24550
+    },
+    {
+      "epoch": 22.086330935251798,
+      "grad_norm": 0.3305526375770569,
+      "learning_rate": 6.866472960119195e-05,
+      "loss": 0.0234,
+      "step": 24560
+    },
+    {
+      "action_loss": 0.005050330422818661,
+      "epoch": 22.086330935251798,
+      "step": 24560
+    },
+    {
+      "epoch": 22.086330935251798,
+      "step": 24560,
+      "torque_loss": 0.14061351120471954
+    },
+    {
+      "epoch": 22.086330935251798,
+      "force_loss": 0.00714228255674243,
+      "step": 24560
+    },
+    {
+      "epoch": 22.095323741007196,
+      "grad_norm": 0.1732715666294098,
+      "learning_rate": 6.863916101404748e-05,
+      "loss": 0.0195,
+      "step": 24570
+    },
+    {
+      "action_loss": 0.004619399551302195,
+      "epoch": 22.095323741007196,
+      "step": 24570
+    },
+    {
+      "epoch": 22.095323741007196,
+      "step": 24570,
+      "torque_loss": 0.09905781596899033
+    },
+    {
+      "epoch": 22.095323741007196,
+      "force_loss": 0.004102995153516531,
+      "step": 24570
+    },
+    {
+      "epoch": 22.10431654676259,
+      "grad_norm": 0.28794586658477783,
+      "learning_rate": 6.8613586764818e-05,
+      "loss": 0.0178,
+      "step": 24580
+    },
+    {
+      "action_loss": 0.003019990399479866,
+      "epoch": 22.10431654676259,
+      "step": 24580
+    },
+    {
+      "epoch": 22.10431654676259,
+      "step": 24580,
+      "torque_loss": 0.16483174264431
+    },
+    {
+      "epoch": 22.10431654676259,
+      "force_loss": 0.0052350894547998905,
+      "step": 24580
+    },
+    {
+      "epoch": 22.113309352517987,
+      "grad_norm": 0.3921899199485779,
+      "learning_rate": 6.858800686127233e-05,
+      "loss": 0.0198,
+      "step": 24590
+    },
+    {
+      "action_loss": 0.0059818304143846035,
+      "epoch": 22.113309352517987,
+      "step": 24590
+    },
+    {
+      "epoch": 22.113309352517987,
+      "step": 24590,
+      "torque_loss": 0.16413426399230957
+    },
+    {
+      "epoch": 22.113309352517987,
+      "force_loss": 0.006394246127456427,
+      "step": 24590
+    },
+    {
+      "epoch": 22.12230215827338,
+      "grad_norm": 0.3071562349796295,
+      "learning_rate": 6.856242131118097e-05,
+      "loss": 0.0193,
+      "step": 24600
+    },
+    {
+      "action_loss": 0.007390322629362345,
+      "epoch": 22.12230215827338,
+      "step": 24600
+    },
+    {
+      "epoch": 22.12230215827338,
+      "step": 24600,
+      "torque_loss": 0.08253031224012375
+    },
+    {
+      "epoch": 22.12230215827338,
+      "force_loss": 0.008102931082248688,
+      "step": 24600
+    },
+    {
+      "epoch": 22.131294964028775,
+      "grad_norm": 0.432756245136261,
+      "learning_rate": 6.853683012231614e-05,
+      "loss": 0.0186,
+      "step": 24610
+    },
+    {
+      "action_loss": 0.0032958232332021,
+      "epoch": 22.131294964028775,
+      "step": 24610
+    },
+    {
+      "epoch": 22.131294964028775,
+      "step": 24610,
+      "torque_loss": 0.19886575639247894
+    },
+    {
+      "epoch": 22.131294964028775,
+      "force_loss": 0.0029224285390228033,
+      "step": 24610
+    },
+    {
+      "epoch": 22.140287769784173,
+      "grad_norm": 0.48232972621917725,
+      "learning_rate": 6.851123330245173e-05,
+      "loss": 0.0213,
+      "step": 24620
+    },
+    {
+      "action_loss": 0.011466942727565765,
+      "epoch": 22.140287769784173,
+      "step": 24620
+    },
+    {
+      "epoch": 22.140287769784173,
+      "step": 24620,
+      "torque_loss": 0.1438477784395218
+    },
+    {
+      "epoch": 22.140287769784173,
+      "force_loss": 0.018884124234318733,
+      "step": 24620
+    },
+    {
+      "epoch": 22.149280575539567,
+      "grad_norm": 0.39638736844062805,
+      "learning_rate": 6.848563085936343e-05,
+      "loss": 0.0181,
+      "step": 24630
+    },
+    {
+      "action_loss": 0.0053353155963122845,
+      "epoch": 22.149280575539567,
+      "step": 24630
+    },
+    {
+      "epoch": 22.149280575539567,
+      "step": 24630,
+      "torque_loss": 0.1025819256901741
+    },
+    {
+      "epoch": 22.149280575539567,
+      "force_loss": 0.0041658105328679085,
+      "step": 24630
+    },
+    {
+      "epoch": 22.158273381294965,
+      "grad_norm": 0.319553941488266,
+      "learning_rate": 6.846002280082853e-05,
+      "loss": 0.0195,
+      "step": 24640
+    },
+    {
+      "action_loss": 0.006258409004658461,
+      "epoch": 22.158273381294965,
+      "step": 24640
+    },
+    {
+      "epoch": 22.158273381294965,
+      "step": 24640,
+      "torque_loss": 0.10401511192321777
+    },
+    {
+      "epoch": 22.158273381294965,
+      "force_loss": 0.003696775995194912,
+      "step": 24640
+    },
+    {
+      "epoch": 22.16726618705036,
+      "grad_norm": 0.23440581560134888,
+      "learning_rate": 6.843440913462614e-05,
+      "loss": 0.0181,
+      "step": 24650
+    },
+    {
+      "action_loss": 0.0025956497993320227,
+      "epoch": 22.16726618705036,
+      "step": 24650
+    },
+    {
+      "epoch": 22.16726618705036,
+      "step": 24650,
+      "torque_loss": 0.0990554690361023
+    },
+    {
+      "epoch": 22.16726618705036,
+      "force_loss": 0.003344378201290965,
+      "step": 24650
+    },
+    {
+      "epoch": 22.176258992805757,
+      "grad_norm": 0.28230926394462585,
+      "learning_rate": 6.840878986853698e-05,
+      "loss": 0.02,
+      "step": 24660
+    },
+    {
+      "action_loss": 0.0030577369034290314,
+      "epoch": 22.176258992805757,
+      "step": 24660
+    },
+    {
+      "epoch": 22.176258992805757,
+      "step": 24660,
+      "torque_loss": 0.13943682610988617
+    },
+    {
+      "epoch": 22.176258992805757,
+      "force_loss": 0.003802162827923894,
+      "step": 24660
+    },
+    {
+      "epoch": 22.18525179856115,
+      "grad_norm": 0.2650807499885559,
+      "learning_rate": 6.838316501034352e-05,
+      "loss": 0.0199,
+      "step": 24670
+    },
+    {
+      "action_loss": 0.0055350023321807384,
+      "epoch": 22.18525179856115,
+      "step": 24670
+    },
+    {
+      "epoch": 22.18525179856115,
+      "step": 24670,
+      "torque_loss": 0.09273897856473923
+    },
+    {
+      "epoch": 22.18525179856115,
+      "force_loss": 0.00897206086665392,
+      "step": 24670
+    },
+    {
+      "epoch": 22.194244604316548,
+      "grad_norm": 0.15893565118312836,
+      "learning_rate": 6.83575345678299e-05,
+      "loss": 0.0187,
+      "step": 24680
+    },
+    {
+      "action_loss": 0.011666147969663143,
+      "epoch": 22.194244604316548,
+      "step": 24680
+    },
+    {
+      "epoch": 22.194244604316548,
+      "step": 24680,
+      "torque_loss": 0.1414196938276291
+    },
+    {
+      "epoch": 22.194244604316548,
+      "force_loss": 0.014527139253914356,
+      "step": 24680
+    },
+    {
+      "epoch": 22.203237410071942,
+      "grad_norm": 0.3142419159412384,
+      "learning_rate": 6.833189854878196e-05,
+      "loss": 0.0215,
+      "step": 24690
+    },
+    {
+      "action_loss": 0.0065027461387217045,
+      "epoch": 22.203237410071942,
+      "step": 24690
+    },
+    {
+      "epoch": 22.203237410071942,
+      "step": 24690,
+      "torque_loss": 0.11355862766504288
+    },
+    {
+      "epoch": 22.203237410071942,
+      "force_loss": 0.009570063091814518,
+      "step": 24690
+    },
+    {
+      "epoch": 22.21223021582734,
+      "grad_norm": 0.1881752759218216,
+      "learning_rate": 6.83062569609873e-05,
+      "loss": 0.0206,
+      "step": 24700
+    },
+    {
+      "action_loss": 0.006756490562111139,
+      "epoch": 22.21223021582734,
+      "step": 24700
+    },
+    {
+      "epoch": 22.21223021582734,
+      "step": 24700,
+      "torque_loss": 0.14468832314014435
+    },
+    {
+      "epoch": 22.21223021582734,
+      "force_loss": 0.003405606606975198,
+      "step": 24700
+    },
+    {
+      "epoch": 22.221223021582734,
+      "grad_norm": 0.2073754072189331,
+      "learning_rate": 6.828060981223512e-05,
+      "loss": 0.0195,
+      "step": 24710
+    },
+    {
+      "action_loss": 0.004220070783048868,
+      "epoch": 22.221223021582734,
+      "step": 24710
+    },
+    {
+      "epoch": 22.221223021582734,
+      "step": 24710,
+      "torque_loss": 0.1627507209777832
+    },
+    {
+      "epoch": 22.221223021582734,
+      "force_loss": 0.008783784694969654,
+      "step": 24710
+    },
+    {
+      "epoch": 22.230215827338128,
+      "grad_norm": 0.28720879554748535,
+      "learning_rate": 6.825495711031634e-05,
+      "loss": 0.0197,
+      "step": 24720
+    },
+    {
+      "action_loss": 0.005291081964969635,
+      "epoch": 22.230215827338128,
+      "step": 24720
+    },
+    {
+      "epoch": 22.230215827338128,
+      "step": 24720,
+      "torque_loss": 0.1927880048751831
+    },
+    {
+      "epoch": 22.230215827338128,
+      "force_loss": 0.006160981487482786,
+      "step": 24720
+    },
+    {
+      "epoch": 22.239208633093526,
+      "grad_norm": 0.4726666510105133,
+      "learning_rate": 6.822929886302359e-05,
+      "loss": 0.0174,
+      "step": 24730
+    },
+    {
+      "action_loss": 0.005073511507362127,
+      "epoch": 22.239208633093526,
+      "step": 24730
+    },
+    {
+      "epoch": 22.239208633093526,
+      "step": 24730,
+      "torque_loss": 0.07566569000482559
+    },
+    {
+      "epoch": 22.239208633093526,
+      "force_loss": 0.002500078408047557,
+      "step": 24730
+    },
+    {
+      "epoch": 22.24820143884892,
+      "grad_norm": 0.2170042246580124,
+      "learning_rate": 6.820363507815116e-05,
+      "loss": 0.0183,
+      "step": 24740
+    },
+    {
+      "action_loss": 0.011065724305808544,
+      "epoch": 22.24820143884892,
+      "step": 24740
+    },
+    {
+      "epoch": 22.24820143884892,
+      "step": 24740,
+      "torque_loss": 0.10500673204660416
+    },
+    {
+      "epoch": 22.24820143884892,
+      "force_loss": 0.014746363274753094,
+      "step": 24740
+    },
+    {
+      "epoch": 22.257194244604317,
+      "grad_norm": 0.29405534267425537,
+      "learning_rate": 6.817796576349501e-05,
+      "loss": 0.0217,
+      "step": 24750
+    },
+    {
+      "action_loss": 0.005858331453055143,
+      "epoch": 22.257194244604317,
+      "step": 24750
+    },
+    {
+      "epoch": 22.257194244604317,
+      "step": 24750,
+      "torque_loss": 0.15028120577335358
+    },
+    {
+      "epoch": 22.257194244604317,
+      "force_loss": 0.0070252916775643826,
+      "step": 24750
+    },
+    {
+      "epoch": 22.26618705035971,
+      "grad_norm": 0.3350282609462738,
+      "learning_rate": 6.815229092685285e-05,
+      "loss": 0.0235,
+      "step": 24760
+    },
+    {
+      "action_loss": 0.005081116687506437,
+      "epoch": 22.26618705035971,
+      "step": 24760
+    },
+    {
+      "epoch": 22.26618705035971,
+      "step": 24760,
+      "torque_loss": 0.0663282573223114
+    },
+    {
+      "epoch": 22.26618705035971,
+      "force_loss": 0.004096261691302061,
+      "step": 24760
+    },
+    {
+      "epoch": 22.27517985611511,
+      "grad_norm": 0.2524561285972595,
+      "learning_rate": 6.812661057602399e-05,
+      "loss": 0.0188,
+      "step": 24770
+    },
+    {
+      "action_loss": 0.003354328451678157,
+      "epoch": 22.27517985611511,
+      "step": 24770
+    },
+    {
+      "epoch": 22.27517985611511,
+      "step": 24770,
+      "torque_loss": 0.08592530339956284
+    },
+    {
+      "epoch": 22.27517985611511,
+      "force_loss": 0.002191887004300952,
+      "step": 24770
+    },
+    {
+      "epoch": 22.284172661870503,
+      "grad_norm": 0.2644980251789093,
+      "learning_rate": 6.810092471880943e-05,
+      "loss": 0.0217,
+      "step": 24780
+    },
+    {
+      "action_loss": 0.0054429336450994015,
+      "epoch": 22.284172661870503,
+      "step": 24780
+    },
+    {
+      "epoch": 22.284172661870503,
+      "step": 24780,
+      "torque_loss": 0.1347578763961792
+    },
+    {
+      "epoch": 22.284172661870503,
+      "force_loss": 0.006393214222043753,
+      "step": 24780
+    },
+    {
+      "epoch": 22.2931654676259,
+      "grad_norm": 0.7584131360054016,
+      "learning_rate": 6.807523336301187e-05,
+      "loss": 0.023,
+      "step": 24790
+    },
+    {
+      "action_loss": 0.015225163660943508,
+      "epoch": 22.2931654676259,
+      "step": 24790
+    },
+    {
+      "epoch": 22.2931654676259,
+      "step": 24790,
+      "torque_loss": 0.17529775202274323
+    },
+    {
+      "epoch": 22.2931654676259,
+      "force_loss": 0.022977866232395172,
+      "step": 24790
+    },
+    {
+      "epoch": 22.302158273381295,
+      "grad_norm": 0.49336618185043335,
+      "learning_rate": 6.804953651643566e-05,
+      "loss": 0.0253,
+      "step": 24800
+    },
+    {
+      "action_loss": 0.008526027202606201,
+      "epoch": 22.302158273381295,
+      "step": 24800
+    },
+    {
+      "epoch": 22.302158273381295,
+      "step": 24800,
+      "torque_loss": 0.11459165811538696
+    },
+    {
+      "epoch": 22.302158273381295,
+      "force_loss": 0.0073268115520477295,
+      "step": 24800
+    },
+    {
+      "epoch": 22.31115107913669,
+      "grad_norm": 0.2615133821964264,
+      "learning_rate": 6.802383418688685e-05,
+      "loss": 0.023,
+      "step": 24810
+    },
+    {
+      "action_loss": 0.0180727019906044,
+      "epoch": 22.31115107913669,
+      "step": 24810
+    },
+    {
+      "epoch": 22.31115107913669,
+      "step": 24810,
+      "torque_loss": 0.1488514542579651
+    },
+    {
+      "epoch": 22.31115107913669,
+      "force_loss": 0.02162788063287735,
+      "step": 24810
+    },
+    {
+      "epoch": 22.320143884892087,
+      "grad_norm": 0.3022231459617615,
+      "learning_rate": 6.799812638217309e-05,
+      "loss": 0.0246,
+      "step": 24820
+    },
+    {
+      "action_loss": 0.014785663224756718,
+      "epoch": 22.320143884892087,
+      "step": 24820
+    },
+    {
+      "epoch": 22.320143884892087,
+      "step": 24820,
+      "torque_loss": 0.16015289723873138
+    },
+    {
+      "epoch": 22.320143884892087,
+      "force_loss": 0.014687980525195599,
+      "step": 24820
+    },
+    {
+      "epoch": 22.32913669064748,
+      "grad_norm": 0.43572723865509033,
+      "learning_rate": 6.797241311010373e-05,
+      "loss": 0.023,
+      "step": 24830
+    },
+    {
+      "action_loss": 0.008995863609015942,
+      "epoch": 22.32913669064748,
+      "step": 24830
+    },
+    {
+      "epoch": 22.32913669064748,
+      "step": 24830,
+      "torque_loss": 0.13701839745044708
+    },
+    {
+      "epoch": 22.32913669064748,
+      "force_loss": 0.005755970720201731,
+      "step": 24830
+    },
+    {
+      "epoch": 22.33812949640288,
+      "grad_norm": 0.8672119975090027,
+      "learning_rate": 6.794669437848982e-05,
+      "loss": 0.0231,
+      "step": 24840
+    },
+    {
+      "action_loss": 0.004539123270660639,
+      "epoch": 22.33812949640288,
+      "step": 24840
+    },
+    {
+      "epoch": 22.33812949640288,
+      "step": 24840,
+      "torque_loss": 0.13663187623023987
+    },
+    {
+      "epoch": 22.33812949640288,
+      "force_loss": 0.004633542615920305,
+      "step": 24840
+    },
+    {
+      "epoch": 22.347122302158272,
+      "grad_norm": 0.3586977422237396,
+      "learning_rate": 6.792097019514402e-05,
+      "loss": 0.0213,
+      "step": 24850
+    },
+    {
+      "action_loss": 0.002659488469362259,
+      "epoch": 22.347122302158272,
+      "step": 24850
+    },
+    {
+      "epoch": 22.347122302158272,
+      "step": 24850,
+      "torque_loss": 0.09668034315109253
+    },
+    {
+      "epoch": 22.347122302158272,
+      "force_loss": 0.0020048124715685844,
+      "step": 24850
+    },
+    {
+      "epoch": 22.35611510791367,
+      "grad_norm": 0.4001941978931427,
+      "learning_rate": 6.789524056788064e-05,
+      "loss": 0.0224,
+      "step": 24860
+    },
+    {
+      "action_loss": 0.003226390341296792,
+      "epoch": 22.35611510791367,
+      "step": 24860
+    },
+    {
+      "epoch": 22.35611510791367,
+      "step": 24860,
+      "torque_loss": 0.10833647847175598
+    },
+    {
+      "epoch": 22.35611510791367,
+      "force_loss": 0.0017195852706208825,
+      "step": 24860
+    },
+    {
+      "epoch": 22.365107913669064,
+      "grad_norm": 0.39897388219833374,
+      "learning_rate": 6.786950550451567e-05,
+      "loss": 0.0181,
+      "step": 24870
+    },
+    {
+      "action_loss": 0.014251780696213245,
+      "epoch": 22.365107913669064,
+      "step": 24870
+    },
+    {
+      "epoch": 22.365107913669064,
+      "step": 24870,
+      "torque_loss": 0.17052705585956573
+    },
+    {
+      "epoch": 22.365107913669064,
+      "force_loss": 0.01718895509839058,
+      "step": 24870
+    },
+    {
+      "epoch": 22.37410071942446,
+      "grad_norm": 0.21309426426887512,
+      "learning_rate": 6.784376501286676e-05,
+      "loss": 0.0216,
+      "step": 24880
+    },
+    {
+      "action_loss": 0.006995067000389099,
+      "epoch": 22.37410071942446,
+      "step": 24880
+    },
+    {
+      "epoch": 22.37410071942446,
+      "step": 24880,
+      "torque_loss": 0.14151662588119507
+    },
+    {
+      "epoch": 22.37410071942446,
+      "force_loss": 0.004014255944639444,
+      "step": 24880
+    },
+    {
+      "epoch": 22.383093525179856,
+      "grad_norm": 0.14126957952976227,
+      "learning_rate": 6.781801910075316e-05,
+      "loss": 0.0192,
+      "step": 24890
+    },
+    {
+      "action_loss": 0.007174244616180658,
+      "epoch": 22.383093525179856,
+      "step": 24890
+    },
+    {
+      "epoch": 22.383093525179856,
+      "step": 24890,
+      "torque_loss": 0.19454972445964813
+    },
+    {
+      "epoch": 22.383093525179856,
+      "force_loss": 0.008282207883894444,
+      "step": 24890
+    },
+    {
+      "epoch": 22.392086330935253,
+      "grad_norm": 0.24397610127925873,
+      "learning_rate": 6.779226777599581e-05,
+      "loss": 0.0192,
+      "step": 24900
+    },
+    {
+      "action_loss": 0.013853867538273335,
+      "epoch": 22.392086330935253,
+      "step": 24900
+    },
+    {
+      "epoch": 22.392086330935253,
+      "step": 24900,
+      "torque_loss": 0.14847691357135773
+    },
+    {
+      "epoch": 22.392086330935253,
+      "force_loss": 0.012605573982000351,
+      "step": 24900
+    },
+    {
+      "epoch": 22.401079136690647,
+      "grad_norm": 0.2652047574520111,
+      "learning_rate": 6.776651104641729e-05,
+      "loss": 0.0218,
+      "step": 24910
+    },
+    {
+      "action_loss": 0.009297415614128113,
+      "epoch": 22.401079136690647,
+      "step": 24910
+    },
+    {
+      "epoch": 22.401079136690647,
+      "step": 24910,
+      "torque_loss": 0.1373676210641861
+    },
+    {
+      "epoch": 22.401079136690647,
+      "force_loss": 0.0061741359531879425,
+      "step": 24910
+    },
+    {
+      "epoch": 22.41007194244604,
+      "grad_norm": 0.2956402599811554,
+      "learning_rate": 6.774074891984183e-05,
+      "loss": 0.021,
+      "step": 24920
+    },
+    {
+      "action_loss": 0.010844677686691284,
+      "epoch": 22.41007194244604,
+      "step": 24920
+    },
+    {
+      "epoch": 22.41007194244604,
+      "step": 24920,
+      "torque_loss": 0.08772342652082443
+    },
+    {
+      "epoch": 22.41007194244604,
+      "force_loss": 0.005964852403849363,
+      "step": 24920
+    },
+    {
+      "epoch": 22.41906474820144,
+      "grad_norm": 0.31046637892723083,
+      "learning_rate": 6.771498140409526e-05,
+      "loss": 0.0197,
+      "step": 24930
+    },
+    {
+      "action_loss": 0.004734450485557318,
+      "epoch": 22.41906474820144,
+      "step": 24930
+    },
+    {
+      "epoch": 22.41906474820144,
+      "step": 24930,
+      "torque_loss": 0.20352749526500702
+    },
+    {
+      "epoch": 22.41906474820144,
+      "force_loss": 0.0023477189242839813,
+      "step": 24930
+    },
+    {
+      "epoch": 22.428057553956833,
+      "grad_norm": 0.43902724981307983,
+      "learning_rate": 6.768920850700506e-05,
+      "loss": 0.02,
+      "step": 24940
+    },
+    {
+      "action_loss": 0.03939766809344292,
+      "epoch": 22.428057553956833,
+      "step": 24940
+    },
+    {
+      "epoch": 22.428057553956833,
+      "step": 24940,
+      "torque_loss": 0.18034398555755615
+    },
+    {
+      "epoch": 22.428057553956833,
+      "force_loss": 0.06515728682279587,
+      "step": 24940
+    },
+    {
+      "epoch": 22.43705035971223,
+      "grad_norm": 0.20046652853488922,
+      "learning_rate": 6.766343023640039e-05,
+      "loss": 0.0259,
+      "step": 24950
+    },
+    {
+      "action_loss": 0.0051938616670668125,
+      "epoch": 22.43705035971223,
+      "step": 24950
+    },
+    {
+      "epoch": 22.43705035971223,
+      "step": 24950,
+      "torque_loss": 0.10448554158210754
+    },
+    {
+      "epoch": 22.43705035971223,
+      "force_loss": 0.008075357414782047,
+      "step": 24950
+    },
+    {
+      "epoch": 22.446043165467625,
+      "grad_norm": 0.2738473117351532,
+      "learning_rate": 6.763764660011198e-05,
+      "loss": 0.0191,
+      "step": 24960
+    },
+    {
+      "action_loss": 0.003480554325506091,
+      "epoch": 22.446043165467625,
+      "step": 24960
+    },
+    {
+      "epoch": 22.446043165467625,
+      "step": 24960,
+      "torque_loss": 0.20879770815372467
+    },
+    {
+      "epoch": 22.446043165467625,
+      "force_loss": 0.003547949017956853,
+      "step": 24960
+    },
+    {
+      "epoch": 22.455035971223023,
+      "grad_norm": 0.234994575381279,
+      "learning_rate": 6.761185760597223e-05,
+      "loss": 0.0188,
+      "step": 24970
+    },
+    {
+      "action_loss": 0.01314420998096466,
+      "epoch": 22.455035971223023,
+      "step": 24970
+    },
+    {
+      "epoch": 22.455035971223023,
+      "step": 24970,
+      "torque_loss": 0.16362716257572174
+    },
+    {
+      "epoch": 22.455035971223023,
+      "force_loss": 0.015762537717819214,
+      "step": 24970
+    },
+    {
+      "epoch": 22.464028776978417,
+      "grad_norm": 0.2934909164905548,
+      "learning_rate": 6.758606326181515e-05,
+      "loss": 0.0222,
+      "step": 24980
+    },
+    {
+      "action_loss": 0.005077384412288666,
+      "epoch": 22.464028776978417,
+      "step": 24980
+    },
+    {
+      "epoch": 22.464028776978417,
+      "step": 24980,
+      "torque_loss": 0.14764131605625153
+    },
+    {
+      "epoch": 22.464028776978417,
+      "force_loss": 0.003645877121016383,
+      "step": 24980
+    },
+    {
+      "epoch": 22.473021582733814,
+      "grad_norm": 0.26081541180610657,
+      "learning_rate": 6.75602635754764e-05,
+      "loss": 0.0202,
+      "step": 24990
+    },
+    {
+      "action_loss": 0.008632371202111244,
+      "epoch": 22.473021582733814,
+      "step": 24990
+    },
+    {
+      "epoch": 22.473021582733814,
+      "step": 24990,
+      "torque_loss": 0.12616242468357086
+    },
+    {
+      "epoch": 22.473021582733814,
+      "force_loss": 0.018579445779323578,
+      "step": 24990
+    },
+    {
+      "epoch": 22.48201438848921,
+      "grad_norm": 0.7656939029693604,
+      "learning_rate": 6.75344585547932e-05,
+      "loss": 0.0232,
+      "step": 25000
+    },
+    {
+      "action_loss": 0.0028978276532143354,
+      "epoch": 22.48201438848921,
+      "step": 25000
+    },
+    {
+      "epoch": 22.48201438848921,
+      "step": 25000,
+      "torque_loss": 0.09270501881837845
+    },
+    {
+      "epoch": 22.48201438848921,
+      "force_loss": 0.0019726885948330164,
+      "step": 25000
+    },
+    {
+      "epoch": 22.491007194244606,
+      "grad_norm": 0.44807904958724976,
+      "learning_rate": 6.750864820760449e-05,
+      "loss": 0.0182,
+      "step": 25010
+    },
+    {
+      "action_loss": 0.0037822090089321136,
+      "epoch": 22.491007194244606,
+      "step": 25010
+    },
+    {
+      "epoch": 22.491007194244606,
+      "step": 25010,
+      "torque_loss": 0.11277223378419876
+    },
+    {
+      "epoch": 22.491007194244606,
+      "force_loss": 0.0042450325563549995,
+      "step": 25010
+    },
+    {
+      "epoch": 22.5,
+      "grad_norm": 0.6256704330444336,
+      "learning_rate": 6.748283254175072e-05,
+      "loss": 0.021,
+      "step": 25020
+    },
+    {
+      "action_loss": 0.0028590282890945673,
+      "epoch": 22.5,
+      "step": 25020
+    },
+    {
+      "epoch": 22.5,
+      "step": 25020,
+      "torque_loss": 0.13656139373779297
+    },
+    {
+      "epoch": 22.5,
+      "force_loss": 0.00206926092505455,
+      "step": 25020
+    },
+    {
+      "epoch": 22.508992805755394,
+      "grad_norm": 0.21707117557525635,
+      "learning_rate": 6.745701156507404e-05,
+      "loss": 0.0185,
+      "step": 25030
+    },
+    {
+      "action_loss": 0.0020222412422299385,
+      "epoch": 22.508992805755394,
+      "step": 25030
+    },
+    {
+      "epoch": 22.508992805755394,
+      "step": 25030,
+      "torque_loss": 0.09703171253204346
+    },
+    {
+      "epoch": 22.508992805755394,
+      "force_loss": 0.0013680089032277465,
+      "step": 25030
+    },
+    {
+      "epoch": 22.51798561151079,
+      "grad_norm": 0.16367389261722565,
+      "learning_rate": 6.743118528541818e-05,
+      "loss": 0.0191,
+      "step": 25040
+    },
+    {
+      "action_loss": 0.0050452034920454025,
+      "epoch": 22.51798561151079,
+      "step": 25040
+    },
+    {
+      "epoch": 22.51798561151079,
+      "step": 25040,
+      "torque_loss": 0.13396082818508148
+    },
+    {
+      "epoch": 22.51798561151079,
+      "force_loss": 0.006993439514189959,
+      "step": 25040
+    },
+    {
+      "epoch": 22.526978417266186,
+      "grad_norm": 0.40440526604652405,
+      "learning_rate": 6.740535371062846e-05,
+      "loss": 0.0197,
+      "step": 25050
+    },
+    {
+      "action_loss": 0.008025317452847958,
+      "epoch": 22.526978417266186,
+      "step": 25050
+    },
+    {
+      "epoch": 22.526978417266186,
+      "step": 25050,
+      "torque_loss": 0.13185955584049225
+    },
+    {
+      "epoch": 22.526978417266186,
+      "force_loss": 0.010429500602185726,
+      "step": 25050
+    },
+    {
+      "epoch": 22.535971223021583,
+      "grad_norm": 0.14783775806427002,
+      "learning_rate": 6.737951684855185e-05,
+      "loss": 0.0188,
+      "step": 25060
+    },
+    {
+      "action_loss": 0.011450252495706081,
+      "epoch": 22.535971223021583,
+      "step": 25060
+    },
+    {
+      "epoch": 22.535971223021583,
+      "step": 25060,
+      "torque_loss": 0.15720783174037933
+    },
+    {
+      "epoch": 22.535971223021583,
+      "force_loss": 0.00878047477453947,
+      "step": 25060
+    },
+    {
+      "epoch": 22.544964028776977,
+      "grad_norm": 0.3013652265071869,
+      "learning_rate": 6.735367470703691e-05,
+      "loss": 0.02,
+      "step": 25070
+    },
+    {
+      "action_loss": 0.0033889685291796923,
+      "epoch": 22.544964028776977,
+      "step": 25070
+    },
+    {
+      "epoch": 22.544964028776977,
+      "step": 25070,
+      "torque_loss": 0.0825701579451561
+    },
+    {
+      "epoch": 22.544964028776977,
+      "force_loss": 0.0018574000569060445,
+      "step": 25070
+    },
+    {
+      "epoch": 22.553956834532375,
+      "grad_norm": 0.3465050756931305,
+      "learning_rate": 6.732782729393379e-05,
+      "loss": 0.0173,
+      "step": 25080
+    },
+    {
+      "action_loss": 0.00617151940241456,
+      "epoch": 22.553956834532375,
+      "step": 25080
+    },
+    {
+      "epoch": 22.553956834532375,
+      "step": 25080,
+      "torque_loss": 0.15709702670574188
+    },
+    {
+      "epoch": 22.553956834532375,
+      "force_loss": 0.010902347974479198,
+      "step": 25080
+    },
+    {
+      "epoch": 22.56294964028777,
+      "grad_norm": 0.3559282720088959,
+      "learning_rate": 6.730197461709425e-05,
+      "loss": 0.0188,
+      "step": 25090
+    },
+    {
+      "action_loss": 0.006413564085960388,
+      "epoch": 22.56294964028777,
+      "step": 25090
+    },
+    {
+      "epoch": 22.56294964028777,
+      "step": 25090,
+      "torque_loss": 0.12822692096233368
+    },
+    {
+      "epoch": 22.56294964028777,
+      "force_loss": 0.004938144702464342,
+      "step": 25090
+    },
+    {
+      "epoch": 22.571942446043167,
+      "grad_norm": 0.18942776322364807,
+      "learning_rate": 6.727611668437164e-05,
+      "loss": 0.0244,
+      "step": 25100
+    },
+    {
+      "action_loss": 0.008657535538077354,
+      "epoch": 22.571942446043167,
+      "step": 25100
+    },
+    {
+      "epoch": 22.571942446043167,
+      "step": 25100,
+      "torque_loss": 0.19749164581298828
+    },
+    {
+      "epoch": 22.571942446043167,
+      "force_loss": 0.006594389211386442,
+      "step": 25100
+    },
+    {
+      "epoch": 22.58093525179856,
+      "grad_norm": 0.348819762468338,
+      "learning_rate": 6.725025350362094e-05,
+      "loss": 0.0184,
+      "step": 25110
+    },
+    {
+      "action_loss": 0.0034841757733374834,
+      "epoch": 22.58093525179856,
+      "step": 25110
+    },
+    {
+      "epoch": 22.58093525179856,
+      "step": 25110,
+      "torque_loss": 0.13012182712554932
+    },
+    {
+      "epoch": 22.58093525179856,
+      "force_loss": 0.003398830071091652,
+      "step": 25110
+    },
+    {
+      "epoch": 22.58992805755396,
+      "grad_norm": 0.26889199018478394,
+      "learning_rate": 6.72243850826987e-05,
+      "loss": 0.0171,
+      "step": 25120
+    },
+    {
+      "action_loss": 0.003253486705943942,
+      "epoch": 22.58992805755396,
+      "step": 25120
+    },
+    {
+      "epoch": 22.58992805755396,
+      "step": 25120,
+      "torque_loss": 0.10711225122213364
+    },
+    {
+      "epoch": 22.58992805755396,
+      "force_loss": 0.004229988437145948,
+      "step": 25120
+    },
+    {
+      "epoch": 22.598920863309353,
+      "grad_norm": 0.5290204882621765,
+      "learning_rate": 6.719851142946305e-05,
+      "loss": 0.0176,
+      "step": 25130
+    },
+    {
+      "action_loss": 0.002730436623096466,
+      "epoch": 22.598920863309353,
+      "step": 25130
+    },
+    {
+      "epoch": 22.598920863309353,
+      "step": 25130,
+      "torque_loss": 0.09934938699007034
+    },
+    {
+      "epoch": 22.598920863309353,
+      "force_loss": 0.0018018983537331223,
+      "step": 25130
+    },
+    {
+      "epoch": 22.607913669064747,
+      "grad_norm": 0.27351611852645874,
+      "learning_rate": 6.717263255177372e-05,
+      "loss": 0.0252,
+      "step": 25140
+    },
+    {
+      "action_loss": 0.005032676737755537,
+      "epoch": 22.607913669064747,
+      "step": 25140
+    },
+    {
+      "epoch": 22.607913669064747,
+      "step": 25140,
+      "torque_loss": 0.17156310379505157
+    },
+    {
+      "epoch": 22.607913669064747,
+      "force_loss": 0.004266476724296808,
+      "step": 25140
+    },
+    {
+      "epoch": 22.616906474820144,
+      "grad_norm": 0.12947946786880493,
+      "learning_rate": 6.714674845749205e-05,
+      "loss": 0.0205,
+      "step": 25150
+    },
+    {
+      "action_loss": 0.008485876955091953,
+      "epoch": 22.616906474820144,
+      "step": 25150
+    },
+    {
+      "epoch": 22.616906474820144,
+      "step": 25150,
+      "torque_loss": 0.13253921270370483
+    },
+    {
+      "epoch": 22.616906474820144,
+      "force_loss": 0.011147302575409412,
+      "step": 25150
+    },
+    {
+      "epoch": 22.62589928057554,
+      "grad_norm": 0.17825235426425934,
+      "learning_rate": 6.712085915448092e-05,
+      "loss": 0.0208,
+      "step": 25160
+    },
+    {
+      "action_loss": 0.004266685806214809,
+      "epoch": 22.62589928057554,
+      "step": 25160
+    },
+    {
+      "epoch": 22.62589928057554,
+      "step": 25160,
+      "torque_loss": 0.11097616702318192
+    },
+    {
+      "epoch": 22.62589928057554,
+      "force_loss": 0.008491885848343372,
+      "step": 25160
+    },
+    {
+      "epoch": 22.634892086330936,
+      "grad_norm": 0.3936159908771515,
+      "learning_rate": 6.709496465060486e-05,
+      "loss": 0.0214,
+      "step": 25170
+    },
+    {
+      "action_loss": 0.00317780370824039,
+      "epoch": 22.634892086330936,
+      "step": 25170
+    },
+    {
+      "epoch": 22.634892086330936,
+      "step": 25170,
+      "torque_loss": 0.1275627166032791
+    },
+    {
+      "epoch": 22.634892086330936,
+      "force_loss": 0.005806193221360445,
+      "step": 25170
+    },
+    {
+      "epoch": 22.64388489208633,
+      "grad_norm": 0.29340144991874695,
+      "learning_rate": 6.706906495372987e-05,
+      "loss": 0.0187,
+      "step": 25180
+    },
+    {
+      "action_loss": 0.006735945586115122,
+      "epoch": 22.64388489208633,
+      "step": 25180
+    },
+    {
+      "epoch": 22.64388489208633,
+      "step": 25180,
+      "torque_loss": 0.16585038602352142
+    },
+    {
+      "epoch": 22.64388489208633,
+      "force_loss": 0.006408578250557184,
+      "step": 25180
+    },
+    {
+      "epoch": 22.652877697841728,
+      "grad_norm": 0.5214519500732422,
+      "learning_rate": 6.704316007172365e-05,
+      "loss": 0.0201,
+      "step": 25190
+    },
+    {
+      "action_loss": 0.005629087332636118,
+      "epoch": 22.652877697841728,
+      "step": 25190
+    },
+    {
+      "epoch": 22.652877697841728,
+      "step": 25190,
+      "torque_loss": 0.17105501890182495
+    },
+    {
+      "epoch": 22.652877697841728,
+      "force_loss": 0.0044324579648673534,
+      "step": 25190
+    },
+    {
+      "epoch": 22.66187050359712,
+      "grad_norm": 0.327261745929718,
+      "learning_rate": 6.701725001245539e-05,
+      "loss": 0.026,
+      "step": 25200
+    },
+    {
+      "action_loss": 0.005990903824567795,
+      "epoch": 22.66187050359712,
+      "step": 25200
+    },
+    {
+      "epoch": 22.66187050359712,
+      "step": 25200,
+      "torque_loss": 0.10485409945249557
+    },
+    {
+      "epoch": 22.66187050359712,
+      "force_loss": 0.005060315132141113,
+      "step": 25200
+    },
+    {
+      "epoch": 22.67086330935252,
+      "grad_norm": 0.3446037173271179,
+      "learning_rate": 6.699133478379588e-05,
+      "loss": 0.0179,
+      "step": 25210
+    },
+    {
+      "action_loss": 0.0031948660034686327,
+      "epoch": 22.67086330935252,
+      "step": 25210
+    },
+    {
+      "epoch": 22.67086330935252,
+      "step": 25210,
+      "torque_loss": 0.08439884334802628
+    },
+    {
+      "epoch": 22.67086330935252,
+      "force_loss": 0.002503957599401474,
+      "step": 25210
+    },
+    {
+      "epoch": 22.679856115107913,
+      "grad_norm": 0.2616611123085022,
+      "learning_rate": 6.69654143936175e-05,
+      "loss": 0.0187,
+      "step": 25220
+    },
+    {
+      "action_loss": 0.003216393291950226,
+      "epoch": 22.679856115107913,
+      "step": 25220
+    },
+    {
+      "epoch": 22.679856115107913,
+      "step": 25220,
+      "torque_loss": 0.12741748988628387
+    },
+    {
+      "epoch": 22.679856115107913,
+      "force_loss": 0.004136684816330671,
+      "step": 25220
+    },
+    {
+      "epoch": 22.68884892086331,
+      "grad_norm": 0.20833612978458405,
+      "learning_rate": 6.693948884979419e-05,
+      "loss": 0.0215,
+      "step": 25230
+    },
+    {
+      "action_loss": 0.0038545706775039434,
+      "epoch": 22.68884892086331,
+      "step": 25230
+    },
+    {
+      "epoch": 22.68884892086331,
+      "step": 25230,
+      "torque_loss": 0.14403019845485687
+    },
+    {
+      "epoch": 22.68884892086331,
+      "force_loss": 0.002605097135528922,
+      "step": 25230
+    },
+    {
+      "epoch": 22.697841726618705,
+      "grad_norm": 0.2755392789840698,
+      "learning_rate": 6.691355816020142e-05,
+      "loss": 0.0192,
+      "step": 25240
+    },
+    {
+      "action_loss": 0.004064594861119986,
+      "epoch": 22.697841726618705,
+      "step": 25240
+    },
+    {
+      "epoch": 22.697841726618705,
+      "step": 25240,
+      "torque_loss": 0.13240979611873627
+    },
+    {
+      "epoch": 22.697841726618705,
+      "force_loss": 0.005109542980790138,
+      "step": 25240
+    },
+    {
+      "epoch": 22.7068345323741,
+      "grad_norm": 0.37402114272117615,
+      "learning_rate": 6.688762233271624e-05,
+      "loss": 0.0198,
+      "step": 25250
+    },
+    {
+      "action_loss": 0.0055160499177873135,
+      "epoch": 22.7068345323741,
+      "step": 25250
+    },
+    {
+      "epoch": 22.7068345323741,
+      "step": 25250,
+      "torque_loss": 0.12401332706212997
+    },
+    {
+      "epoch": 22.7068345323741,
+      "force_loss": 0.008280246518552303,
+      "step": 25250
+    },
+    {
+      "epoch": 22.715827338129497,
+      "grad_norm": 0.12323077023029327,
+      "learning_rate": 6.68616813752173e-05,
+      "loss": 0.0225,
+      "step": 25260
+    },
+    {
+      "action_loss": 0.004778133239597082,
+      "epoch": 22.715827338129497,
+      "step": 25260
+    },
+    {
+      "epoch": 22.715827338129497,
+      "step": 25260,
+      "torque_loss": 0.11605324596166611
+    },
+    {
+      "epoch": 22.715827338129497,
+      "force_loss": 0.004602286498993635,
+      "step": 25260
+    },
+    {
+      "epoch": 22.72482014388489,
+      "grad_norm": 0.18610790371894836,
+      "learning_rate": 6.683573529558477e-05,
+      "loss": 0.021,
+      "step": 25270
+    },
+    {
+      "action_loss": 0.0023338731843978167,
+      "epoch": 22.72482014388489,
+      "step": 25270
+    },
+    {
+      "epoch": 22.72482014388489,
+      "step": 25270,
+      "torque_loss": 0.10729467868804932
+    },
+    {
+      "epoch": 22.72482014388489,
+      "force_loss": 0.002586699789389968,
+      "step": 25270
+    },
+    {
+      "epoch": 22.73381294964029,
+      "grad_norm": 0.20692098140716553,
+      "learning_rate": 6.680978410170037e-05,
+      "loss": 0.0198,
+      "step": 25280
+    },
+    {
+      "action_loss": 0.0026255929842591286,
+      "epoch": 22.73381294964029,
+      "step": 25280
+    },
+    {
+      "epoch": 22.73381294964029,
+      "step": 25280,
+      "torque_loss": 0.15501603484153748
+    },
+    {
+      "epoch": 22.73381294964029,
+      "force_loss": 0.00565456785261631,
+      "step": 25280
+    },
+    {
+      "epoch": 22.742805755395683,
+      "grad_norm": 0.41962599754333496,
+      "learning_rate": 6.678382780144741e-05,
+      "loss": 0.0222,
+      "step": 25290
+    },
+    {
+      "action_loss": 0.009218104183673859,
+      "epoch": 22.742805755395683,
+      "step": 25290
+    },
+    {
+      "epoch": 22.742805755395683,
+      "step": 25290,
+      "torque_loss": 0.1374831348657608
+    },
+    {
+      "epoch": 22.742805755395683,
+      "force_loss": 0.007170087192207575,
+      "step": 25290
+    },
+    {
+      "epoch": 22.75179856115108,
+      "grad_norm": 0.15407513082027435,
+      "learning_rate": 6.675786640271071e-05,
+      "loss": 0.0186,
+      "step": 25300
+    },
+    {
+      "action_loss": 0.002584812231361866,
+      "epoch": 22.75179856115108,
+      "step": 25300
+    },
+    {
+      "epoch": 22.75179856115108,
+      "step": 25300,
+      "torque_loss": 0.09250873327255249
+    },
+    {
+      "epoch": 22.75179856115108,
+      "force_loss": 0.003979390021413565,
+      "step": 25300
+    },
+    {
+      "epoch": 22.760791366906474,
+      "grad_norm": 0.16511009633541107,
+      "learning_rate": 6.673189991337665e-05,
+      "loss": 0.0193,
+      "step": 25310
+    },
+    {
+      "action_loss": 0.005039873067289591,
+      "epoch": 22.760791366906474,
+      "step": 25310
+    },
+    {
+      "epoch": 22.760791366906474,
+      "step": 25310,
+      "torque_loss": 0.17084509134292603
+    },
+    {
+      "epoch": 22.760791366906474,
+      "force_loss": 0.005266187712550163,
+      "step": 25310
+    },
+    {
+      "epoch": 22.769784172661872,
+      "grad_norm": 0.18709422647953033,
+      "learning_rate": 6.670592834133317e-05,
+      "loss": 0.0241,
+      "step": 25320
+    },
+    {
+      "action_loss": 0.00991649180650711,
+      "epoch": 22.769784172661872,
+      "step": 25320
+    },
+    {
+      "epoch": 22.769784172661872,
+      "step": 25320,
+      "torque_loss": 0.16399723291397095
+    },
+    {
+      "epoch": 22.769784172661872,
+      "force_loss": 0.011603073216974735,
+      "step": 25320
+    },
+    {
+      "epoch": 22.778776978417266,
+      "grad_norm": 0.5876155495643616,
+      "learning_rate": 6.667995169446979e-05,
+      "loss": 0.0211,
+      "step": 25330
+    },
+    {
+      "action_loss": 0.00250917742960155,
+      "epoch": 22.778776978417266,
+      "step": 25330
+    },
+    {
+      "epoch": 22.778776978417266,
+      "step": 25330,
+      "torque_loss": 0.11301350593566895
+    },
+    {
+      "epoch": 22.778776978417266,
+      "force_loss": 0.0021972907707095146,
+      "step": 25330
+    },
+    {
+      "epoch": 22.78776978417266,
+      "grad_norm": 0.20074112713336945,
+      "learning_rate": 6.665396998067747e-05,
+      "loss": 0.0239,
+      "step": 25340
+    },
+    {
+      "action_loss": 0.006846604403108358,
+      "epoch": 22.78776978417266,
+      "step": 25340
+    },
+    {
+      "epoch": 22.78776978417266,
+      "step": 25340,
+      "torque_loss": 0.1528770476579666
+    },
+    {
+      "epoch": 22.78776978417266,
+      "force_loss": 0.006580863147974014,
+      "step": 25340
+    },
+    {
+      "epoch": 22.796762589928058,
+      "grad_norm": 0.2842159867286682,
+      "learning_rate": 6.66279832078488e-05,
+      "loss": 0.0195,
+      "step": 25350
+    },
+    {
+      "action_loss": 0.004146437626332045,
+      "epoch": 22.796762589928058,
+      "step": 25350
+    },
+    {
+      "epoch": 22.796762589928058,
+      "step": 25350,
+      "torque_loss": 0.08629711717367172
+    },
+    {
+      "epoch": 22.796762589928058,
+      "force_loss": 0.0037164564710110426,
+      "step": 25350
+    },
+    {
+      "epoch": 22.805755395683452,
+      "grad_norm": 0.2562117874622345,
+      "learning_rate": 6.660199138387786e-05,
+      "loss": 0.0198,
+      "step": 25360
+    },
+    {
+      "action_loss": 0.009775112383067608,
+      "epoch": 22.805755395683452,
+      "step": 25360
+    },
+    {
+      "epoch": 22.805755395683452,
+      "step": 25360,
+      "torque_loss": 0.11194447427988052
+    },
+    {
+      "epoch": 22.805755395683452,
+      "force_loss": 0.010112606920301914,
+      "step": 25360
+    },
+    {
+      "epoch": 22.81474820143885,
+      "grad_norm": 0.2811155915260315,
+      "learning_rate": 6.65759945166603e-05,
+      "loss": 0.0224,
+      "step": 25370
+    },
+    {
+      "action_loss": 0.004484861623495817,
+      "epoch": 22.81474820143885,
+      "step": 25370
+    },
+    {
+      "epoch": 22.81474820143885,
+      "step": 25370,
+      "torque_loss": 0.1898781657218933
+    },
+    {
+      "epoch": 22.81474820143885,
+      "force_loss": 0.0037356652319431305,
+      "step": 25370
+    },
+    {
+      "epoch": 22.823741007194243,
+      "grad_norm": 0.44249197840690613,
+      "learning_rate": 6.654999261409326e-05,
+      "loss": 0.0282,
+      "step": 25380
+    },
+    {
+      "action_loss": 0.01969350315630436,
+      "epoch": 22.823741007194243,
+      "step": 25380
+    },
+    {
+      "epoch": 22.823741007194243,
+      "step": 25380,
+      "torque_loss": 0.1598086953163147
+    },
+    {
+      "epoch": 22.823741007194243,
+      "force_loss": 0.023778853937983513,
+      "step": 25380
+    },
+    {
+      "epoch": 22.83273381294964,
+      "grad_norm": 0.6640731692314148,
+      "learning_rate": 6.652398568407544e-05,
+      "loss": 0.0256,
+      "step": 25390
+    },
+    {
+      "action_loss": 0.0049319821409881115,
+      "epoch": 22.83273381294964,
+      "step": 25390
+    },
+    {
+      "epoch": 22.83273381294964,
+      "step": 25390,
+      "torque_loss": 0.10348508507013321
+    },
+    {
+      "epoch": 22.83273381294964,
+      "force_loss": 0.0038687579799443483,
+      "step": 25390
+    },
+    {
+      "epoch": 22.841726618705035,
+      "grad_norm": 0.29850515723228455,
+      "learning_rate": 6.649797373450707e-05,
+      "loss": 0.023,
+      "step": 25400
+    },
+    {
+      "action_loss": 0.007974536158144474,
+      "epoch": 22.841726618705035,
+      "step": 25400
+    },
+    {
+      "epoch": 22.841726618705035,
+      "step": 25400,
+      "torque_loss": 0.12296003848314285
+    },
+    {
+      "epoch": 22.841726618705035,
+      "force_loss": 0.005034271627664566,
+      "step": 25400
+    },
+    {
+      "epoch": 22.850719424460433,
+      "grad_norm": 0.19731557369232178,
+      "learning_rate": 6.647195677328988e-05,
+      "loss": 0.0181,
+      "step": 25410
+    },
+    {
+      "action_loss": 0.009311657398939133,
+      "epoch": 22.850719424460433,
+      "step": 25410
+    },
+    {
+      "epoch": 22.850719424460433,
+      "step": 25410,
+      "torque_loss": 0.13293598592281342
+    },
+    {
+      "epoch": 22.850719424460433,
+      "force_loss": 0.011000546626746655,
+      "step": 25410
+    },
+    {
+      "epoch": 22.859712230215827,
+      "grad_norm": 0.1860690414905548,
+      "learning_rate": 6.644593480832712e-05,
+      "loss": 0.018,
+      "step": 25420
+    },
+    {
+      "action_loss": 0.005338681396096945,
+      "epoch": 22.859712230215827,
+      "step": 25420
+    },
+    {
+      "epoch": 22.859712230215827,
+      "step": 25420,
+      "torque_loss": 0.1254522204399109
+    },
+    {
+      "epoch": 22.859712230215827,
+      "force_loss": 0.004750473890453577,
+      "step": 25420
+    },
+    {
+      "epoch": 22.868705035971225,
+      "grad_norm": 0.2702329754829407,
+      "learning_rate": 6.641990784752363e-05,
+      "loss": 0.0206,
+      "step": 25430
+    },
+    {
+      "action_loss": 0.003232441609725356,
+      "epoch": 22.868705035971225,
+      "step": 25430
+    },
+    {
+      "epoch": 22.868705035971225,
+      "step": 25430,
+      "torque_loss": 0.09547365456819534
+    },
+    {
+      "epoch": 22.868705035971225,
+      "force_loss": 0.004664943087846041,
+      "step": 25430
+    },
+    {
+      "epoch": 22.87769784172662,
+      "grad_norm": 0.7561193108558655,
+      "learning_rate": 6.639387589878566e-05,
+      "loss": 0.0222,
+      "step": 25440
+    },
+    {
+      "action_loss": 0.003841092111542821,
+      "epoch": 22.87769784172662,
+      "step": 25440
+    },
+    {
+      "epoch": 22.87769784172662,
+      "step": 25440,
+      "torque_loss": 0.13480597734451294
+    },
+    {
+      "epoch": 22.87769784172662,
+      "force_loss": 0.002297803992405534,
+      "step": 25440
+    },
+    {
+      "epoch": 22.886690647482013,
+      "grad_norm": 0.3329511880874634,
+      "learning_rate": 6.636783897002103e-05,
+      "loss": 0.0208,
+      "step": 25450
+    },
+    {
+      "action_loss": 0.0023310650140047073,
+      "epoch": 22.886690647482013,
+      "step": 25450
+    },
+    {
+      "epoch": 22.886690647482013,
+      "step": 25450,
+      "torque_loss": 0.15812350809574127
+    },
+    {
+      "epoch": 22.886690647482013,
+      "force_loss": 0.0025320190470665693,
+      "step": 25450
+    },
+    {
+      "epoch": 22.89568345323741,
+      "grad_norm": 0.19684022665023804,
+      "learning_rate": 6.63417970691391e-05,
+      "loss": 0.0213,
+      "step": 25460
+    },
+    {
+      "action_loss": 0.0045658499002456665,
+      "epoch": 22.89568345323741,
+      "step": 25460
+    },
+    {
+      "epoch": 22.89568345323741,
+      "step": 25460,
+      "torque_loss": 0.13276435434818268
+    },
+    {
+      "epoch": 22.89568345323741,
+      "force_loss": 0.006904436741024256,
+      "step": 25460
+    },
+    {
+      "epoch": 22.904676258992804,
+      "grad_norm": 0.4892916977405548,
+      "learning_rate": 6.63157502040507e-05,
+      "loss": 0.0182,
+      "step": 25470
+    },
+    {
+      "action_loss": 0.02397470735013485,
+      "epoch": 22.904676258992804,
+      "step": 25470
+    },
+    {
+      "epoch": 22.904676258992804,
+      "step": 25470,
+      "torque_loss": 0.17074398696422577
+    },
+    {
+      "epoch": 22.904676258992804,
+      "force_loss": 0.02185717783868313,
+      "step": 25470
+    },
+    {
+      "epoch": 22.913669064748202,
+      "grad_norm": 0.2766764760017395,
+      "learning_rate": 6.628969838266819e-05,
+      "loss": 0.023,
+      "step": 25480
+    },
+    {
+      "action_loss": 0.008708766661584377,
+      "epoch": 22.913669064748202,
+      "step": 25480
+    },
+    {
+      "epoch": 22.913669064748202,
+      "step": 25480,
+      "torque_loss": 0.12777329981327057
+    },
+    {
+      "epoch": 22.913669064748202,
+      "force_loss": 0.007226257119327784,
+      "step": 25480
+    },
+    {
+      "epoch": 22.922661870503596,
+      "grad_norm": 0.229297935962677,
+      "learning_rate": 6.626364161290541e-05,
+      "loss": 0.0197,
+      "step": 25490
+    },
+    {
+      "action_loss": 0.0076741487719118595,
+      "epoch": 22.922661870503596,
+      "step": 25490
+    },
+    {
+      "epoch": 22.922661870503596,
+      "step": 25490,
+      "torque_loss": 0.13525649905204773
+    },
+    {
+      "epoch": 22.922661870503596,
+      "force_loss": 0.0064540766179561615,
+      "step": 25490
+    },
+    {
+      "epoch": 22.931654676258994,
+      "grad_norm": 0.40219736099243164,
+      "learning_rate": 6.623757990267774e-05,
+      "loss": 0.0221,
+      "step": 25500
+    },
+    {
+      "action_loss": 0.0067048631608486176,
+      "epoch": 22.931654676258994,
+      "step": 25500
+    },
+    {
+      "epoch": 22.931654676258994,
+      "step": 25500,
+      "torque_loss": 0.13198016583919525
+    },
+    {
+      "epoch": 22.931654676258994,
+      "force_loss": 0.004208454862236977,
+      "step": 25500
+    },
+    {
+      "epoch": 22.940647482014388,
+      "grad_norm": 0.4207010567188263,
+      "learning_rate": 6.621151325990201e-05,
+      "loss": 0.0203,
+      "step": 25510
+    },
+    {
+      "action_loss": 0.014719352126121521,
+      "epoch": 22.940647482014388,
+      "step": 25510
+    },
+    {
+      "epoch": 22.940647482014388,
+      "step": 25510,
+      "torque_loss": 0.13498015701770782
+    },
+    {
+      "epoch": 22.940647482014388,
+      "force_loss": 0.008890347555279732,
+      "step": 25510
+    },
+    {
+      "epoch": 22.949640287769785,
+      "grad_norm": 0.18830212950706482,
+      "learning_rate": 6.618544169249657e-05,
+      "loss": 0.0213,
+      "step": 25520
+    },
+    {
+      "action_loss": 0.0021579600870609283,
+      "epoch": 22.949640287769785,
+      "step": 25520
+    },
+    {
+      "epoch": 22.949640287769785,
+      "step": 25520,
+      "torque_loss": 0.10413229465484619
+    },
+    {
+      "epoch": 22.949640287769785,
+      "force_loss": 0.0014954224461689591,
+      "step": 25520
+    },
+    {
+      "epoch": 22.95863309352518,
+      "grad_norm": 0.330130010843277,
+      "learning_rate": 6.615936520838133e-05,
+      "loss": 0.0219,
+      "step": 25530
+    },
+    {
+      "action_loss": 0.0031004927586764097,
+      "epoch": 22.95863309352518,
+      "step": 25530
+    },
+    {
+      "epoch": 22.95863309352518,
+      "step": 25530,
+      "torque_loss": 0.12198827415704727
+    },
+    {
+      "epoch": 22.95863309352518,
+      "force_loss": 0.002654009498655796,
+      "step": 25530
+    },
+    {
+      "epoch": 22.967625899280577,
+      "grad_norm": 0.5394603610038757,
+      "learning_rate": 6.613328381547759e-05,
+      "loss": 0.0179,
+      "step": 25540
+    },
+    {
+      "action_loss": 0.0065713063813745975,
+      "epoch": 22.967625899280577,
+      "step": 25540
+    },
+    {
+      "epoch": 22.967625899280577,
+      "step": 25540,
+      "torque_loss": 0.18813152611255646
+    },
+    {
+      "epoch": 22.967625899280577,
+      "force_loss": 0.00476634269580245,
+      "step": 25540
+    },
+    {
+      "epoch": 22.97661870503597,
+      "grad_norm": 0.27255162596702576,
+      "learning_rate": 6.610719752170821e-05,
+      "loss": 0.0313,
+      "step": 25550
+    },
+    {
+      "action_loss": 0.002748814644291997,
+      "epoch": 22.97661870503597,
+      "step": 25550
+    },
+    {
+      "epoch": 22.97661870503597,
+      "step": 25550,
+      "torque_loss": 0.13115152716636658
+    },
+    {
+      "epoch": 22.97661870503597,
+      "force_loss": 0.0018982872134074569,
+      "step": 25550
+    },
+    {
+      "epoch": 22.985611510791365,
+      "grad_norm": 0.3695186376571655,
+      "learning_rate": 6.60811063349975e-05,
+      "loss": 0.02,
+      "step": 25560
+    },
+    {
+      "action_loss": 0.0027263539377599955,
+      "epoch": 22.985611510791365,
+      "step": 25560
+    },
+    {
+      "epoch": 22.985611510791365,
+      "step": 25560,
+      "torque_loss": 0.08767839521169662
+    },
+    {
+      "epoch": 22.985611510791365,
+      "force_loss": 0.0022533198352903128,
+      "step": 25560
+    },
+    {
+      "epoch": 22.994604316546763,
+      "grad_norm": 0.24433258175849915,
+      "learning_rate": 6.605501026327127e-05,
+      "loss": 0.0206,
+      "step": 25570
+    },
+    {
+      "action_loss": 0.00752577418461442,
+      "epoch": 22.994604316546763,
+      "step": 25570
+    },
+    {
+      "epoch": 22.994604316546763,
+      "step": 25570,
+      "torque_loss": 0.11854282766580582
+    },
+    {
+      "epoch": 22.994604316546763,
+      "force_loss": 0.003949462901800871,
+      "step": 25570
+    },
+    {
+      "epoch": 23.003597122302157,
+      "grad_norm": 0.588006854057312,
+      "learning_rate": 6.602890931445685e-05,
+      "loss": 0.0257,
+      "step": 25580
+    },
+    {
+      "action_loss": 0.008192433975636959,
+      "epoch": 23.003597122302157,
+      "step": 25580
+    },
+    {
+      "epoch": 23.003597122302157,
+      "step": 25580,
+      "torque_loss": 0.14182865619659424
+    },
+    {
+      "epoch": 23.003597122302157,
+      "force_loss": 0.008779712952673435,
+      "step": 25580
+    },
+    {
+      "epoch": 23.012589928057555,
+      "grad_norm": 0.3446831703186035,
+      "learning_rate": 6.6002803496483e-05,
+      "loss": 0.0223,
+      "step": 25590
+    },
+    {
+      "action_loss": 0.0035741832107305527,
+      "epoch": 23.012589928057555,
+      "step": 25590
+    },
+    {
+      "epoch": 23.012589928057555,
+      "step": 25590,
+      "torque_loss": 0.13330493867397308
+    },
+    {
+      "epoch": 23.012589928057555,
+      "force_loss": 0.004079489968717098,
+      "step": 25590
+    },
+    {
+      "epoch": 23.02158273381295,
+      "grad_norm": 0.19388410449028015,
+      "learning_rate": 6.597669281727997e-05,
+      "loss": 0.0178,
+      "step": 25600
+    },
+    {
+      "action_loss": 0.009083190932869911,
+      "epoch": 23.02158273381295,
+      "step": 25600
+    },
+    {
+      "epoch": 23.02158273381295,
+      "step": 25600,
+      "torque_loss": 0.10139637440443039
+    },
+    {
+      "epoch": 23.02158273381295,
+      "force_loss": 0.00901173148304224,
+      "step": 25600
+    },
+    {
+      "epoch": 23.030575539568346,
+      "grad_norm": 0.22701629996299744,
+      "learning_rate": 6.595057728477949e-05,
+      "loss": 0.0201,
+      "step": 25610
+    },
+    {
+      "action_loss": 0.0063840169459581375,
+      "epoch": 23.030575539568346,
+      "step": 25610
+    },
+    {
+      "epoch": 23.030575539568346,
+      "step": 25610,
+      "torque_loss": 0.11656978726387024
+    },
+    {
+      "epoch": 23.030575539568346,
+      "force_loss": 0.011731003411114216,
+      "step": 25610
+    },
+    {
+      "epoch": 23.03956834532374,
+      "grad_norm": 0.33431360125541687,
+      "learning_rate": 6.59244569069148e-05,
+      "loss": 0.0234,
+      "step": 25620
+    },
+    {
+      "action_loss": 0.0028157334309071302,
+      "epoch": 23.03956834532374,
+      "step": 25620
+    },
+    {
+      "epoch": 23.03956834532374,
+      "step": 25620,
+      "torque_loss": 0.14123941957950592
+    },
+    {
+      "epoch": 23.03956834532374,
+      "force_loss": 0.0038219671696424484,
+      "step": 25620
+    },
+    {
+      "epoch": 23.048561151079138,
+      "grad_norm": 0.3215562105178833,
+      "learning_rate": 6.589833169162054e-05,
+      "loss": 0.0186,
+      "step": 25630
+    },
+    {
+      "action_loss": 0.005002176854759455,
+      "epoch": 23.048561151079138,
+      "step": 25630
+    },
+    {
+      "epoch": 23.048561151079138,
+      "step": 25630,
+      "torque_loss": 0.15622670948505402
+    },
+    {
+      "epoch": 23.048561151079138,
+      "force_loss": 0.006401598919183016,
+      "step": 25630
+    },
+    {
+      "epoch": 23.057553956834532,
+      "grad_norm": 0.5653586983680725,
+      "learning_rate": 6.587220164683291e-05,
+      "loss": 0.0231,
+      "step": 25640
+    },
+    {
+      "action_loss": 0.0023761941120028496,
+      "epoch": 23.057553956834532,
+      "step": 25640
+    },
+    {
+      "epoch": 23.057553956834532,
+      "step": 25640,
+      "torque_loss": 0.1033112183213234
+    },
+    {
+      "epoch": 23.057553956834532,
+      "force_loss": 0.0028601244557648897,
+      "step": 25640
+    },
+    {
+      "epoch": 23.06654676258993,
+      "grad_norm": 0.5096305012702942,
+      "learning_rate": 6.58460667804895e-05,
+      "loss": 0.0201,
+      "step": 25650
+    },
+    {
+      "action_loss": 0.007477613165974617,
+      "epoch": 23.06654676258993,
+      "step": 25650
+    },
+    {
+      "epoch": 23.06654676258993,
+      "step": 25650,
+      "torque_loss": 0.09020773321390152
+    },
+    {
+      "epoch": 23.06654676258993,
+      "force_loss": 0.005367389414459467,
+      "step": 25650
+    },
+    {
+      "epoch": 23.075539568345324,
+      "grad_norm": 0.30646347999572754,
+      "learning_rate": 6.581992710052938e-05,
+      "loss": 0.0236,
+      "step": 25660
+    },
+    {
+      "action_loss": 0.005354841705411673,
+      "epoch": 23.075539568345324,
+      "step": 25660
+    },
+    {
+      "epoch": 23.075539568345324,
+      "step": 25660,
+      "torque_loss": 0.14324022829532623
+    },
+    {
+      "epoch": 23.075539568345324,
+      "force_loss": 0.005249425303190947,
+      "step": 25660
+    },
+    {
+      "epoch": 23.084532374100718,
+      "grad_norm": 0.6165043115615845,
+      "learning_rate": 6.579378261489311e-05,
+      "loss": 0.0238,
+      "step": 25670
+    },
+    {
+      "action_loss": 0.020665811374783516,
+      "epoch": 23.084532374100718,
+      "step": 25670
+    },
+    {
+      "epoch": 23.084532374100718,
+      "step": 25670,
+      "torque_loss": 0.11552337557077408
+    },
+    {
+      "epoch": 23.084532374100718,
+      "force_loss": 0.024774039164185524,
+      "step": 25670
+    },
+    {
+      "epoch": 23.093525179856115,
+      "grad_norm": 0.23046307265758514,
+      "learning_rate": 6.576763333152268e-05,
+      "loss": 0.0213,
+      "step": 25680
+    },
+    {
+      "action_loss": 0.004617140628397465,
+      "epoch": 23.093525179856115,
+      "step": 25680
+    },
+    {
+      "epoch": 23.093525179856115,
+      "step": 25680,
+      "torque_loss": 0.19429004192352295
+    },
+    {
+      "epoch": 23.093525179856115,
+      "force_loss": 0.003150787902995944,
+      "step": 25680
+    },
+    {
+      "epoch": 23.10251798561151,
+      "grad_norm": 0.29696786403656006,
+      "learning_rate": 6.574147925836159e-05,
+      "loss": 0.0184,
+      "step": 25690
+    },
+    {
+      "action_loss": 0.01016028132289648,
+      "epoch": 23.10251798561151,
+      "step": 25690
+    },
+    {
+      "epoch": 23.10251798561151,
+      "step": 25690,
+      "torque_loss": 0.1103476956486702
+    },
+    {
+      "epoch": 23.10251798561151,
+      "force_loss": 0.006418560165911913,
+      "step": 25690
+    },
+    {
+      "epoch": 23.111510791366907,
+      "grad_norm": 0.22070881724357605,
+      "learning_rate": 6.571532040335472e-05,
+      "loss": 0.0206,
+      "step": 25700
+    },
+    {
+      "action_loss": 0.002738197334110737,
+      "epoch": 23.111510791366907,
+      "step": 25700
+    },
+    {
+      "epoch": 23.111510791366907,
+      "step": 25700,
+      "torque_loss": 0.1095813736319542
+    },
+    {
+      "epoch": 23.111510791366907,
+      "force_loss": 0.0031854708213359118,
+      "step": 25700
+    },
+    {
+      "epoch": 23.1205035971223,
+      "grad_norm": 0.494768351316452,
+      "learning_rate": 6.568915677444845e-05,
+      "loss": 0.0202,
+      "step": 25710
+    },
+    {
+      "action_loss": 0.011660532094538212,
+      "epoch": 23.1205035971223,
+      "step": 25710
+    },
+    {
+      "epoch": 23.1205035971223,
+      "step": 25710,
+      "torque_loss": 0.12994426488876343
+    },
+    {
+      "epoch": 23.1205035971223,
+      "force_loss": 0.023325903341174126,
+      "step": 25710
+    },
+    {
+      "epoch": 23.1294964028777,
+      "grad_norm": 0.32806041836738586,
+      "learning_rate": 6.56629883795906e-05,
+      "loss": 0.0211,
+      "step": 25720
+    },
+    {
+      "action_loss": 0.00436545442789793,
+      "epoch": 23.1294964028777,
+      "step": 25720
+    },
+    {
+      "epoch": 23.1294964028777,
+      "step": 25720,
+      "torque_loss": 0.1534643918275833
+    },
+    {
+      "epoch": 23.1294964028777,
+      "force_loss": 0.002581397071480751,
+      "step": 25720
+    },
+    {
+      "epoch": 23.138489208633093,
+      "grad_norm": 0.26697176694869995,
+      "learning_rate": 6.563681522673043e-05,
+      "loss": 0.0215,
+      "step": 25730
+    },
+    {
+      "action_loss": 0.004084462765604258,
+      "epoch": 23.138489208633093,
+      "step": 25730
+    },
+    {
+      "epoch": 23.138489208633093,
+      "step": 25730,
+      "torque_loss": 0.1061796173453331
+    },
+    {
+      "epoch": 23.138489208633093,
+      "force_loss": 0.008010889403522015,
+      "step": 25730
+    },
+    {
+      "epoch": 23.14748201438849,
+      "grad_norm": 0.32571542263031006,
+      "learning_rate": 6.561063732381867e-05,
+      "loss": 0.0236,
+      "step": 25740
+    },
+    {
+      "action_loss": 0.007778527680784464,
+      "epoch": 23.14748201438849,
+      "step": 25740
+    },
+    {
+      "epoch": 23.14748201438849,
+      "step": 25740,
+      "torque_loss": 0.14481233060359955
+    },
+    {
+      "epoch": 23.14748201438849,
+      "force_loss": 0.004333335906267166,
+      "step": 25740
+    },
+    {
+      "epoch": 23.156474820143885,
+      "grad_norm": 0.18872971832752228,
+      "learning_rate": 6.558445467880745e-05,
+      "loss": 0.02,
+      "step": 25750
+    },
+    {
+      "action_loss": 0.0035236794501543045,
+      "epoch": 23.156474820143885,
+      "step": 25750
+    },
+    {
+      "epoch": 23.156474820143885,
+      "step": 25750,
+      "torque_loss": 0.12785522639751434
+    },
+    {
+      "epoch": 23.156474820143885,
+      "force_loss": 0.0025902960915118456,
+      "step": 25750
+    },
+    {
+      "epoch": 23.165467625899282,
+      "grad_norm": 0.18403694033622742,
+      "learning_rate": 6.55582672996504e-05,
+      "loss": 0.0212,
+      "step": 25760
+    },
+    {
+      "action_loss": 0.00574125349521637,
+      "epoch": 23.165467625899282,
+      "step": 25760
+    },
+    {
+      "epoch": 23.165467625899282,
+      "step": 25760,
+      "torque_loss": 0.12348950654268265
+    },
+    {
+      "epoch": 23.165467625899282,
+      "force_loss": 0.005121069028973579,
+      "step": 25760
+    },
+    {
+      "epoch": 23.174460431654676,
+      "grad_norm": 0.35492053627967834,
+      "learning_rate": 6.553207519430253e-05,
+      "loss": 0.0182,
+      "step": 25770
+    },
+    {
+      "action_loss": 0.009262331761419773,
+      "epoch": 23.174460431654676,
+      "step": 25770
+    },
+    {
+      "epoch": 23.174460431654676,
+      "step": 25770,
+      "torque_loss": 0.1523372381925583
+    },
+    {
+      "epoch": 23.174460431654676,
+      "force_loss": 0.012263941578567028,
+      "step": 25770
+    },
+    {
+      "epoch": 23.18345323741007,
+      "grad_norm": 0.16774387657642365,
+      "learning_rate": 6.550587837072032e-05,
+      "loss": 0.0207,
+      "step": 25780
+    },
+    {
+      "action_loss": 0.01723303459584713,
+      "epoch": 23.18345323741007,
+      "step": 25780
+    },
+    {
+      "epoch": 23.18345323741007,
+      "step": 25780,
+      "torque_loss": 0.12551383674144745
+    },
+    {
+      "epoch": 23.18345323741007,
+      "force_loss": 0.01829545572400093,
+      "step": 25780
+    },
+    {
+      "epoch": 23.192446043165468,
+      "grad_norm": 0.19067732989788055,
+      "learning_rate": 6.547967683686166e-05,
+      "loss": 0.0228,
+      "step": 25790
+    },
+    {
+      "action_loss": 0.006243444979190826,
+      "epoch": 23.192446043165468,
+      "step": 25790
+    },
+    {
+      "epoch": 23.192446043165468,
+      "step": 25790,
+      "torque_loss": 0.14092488586902618
+    },
+    {
+      "epoch": 23.192446043165468,
+      "force_loss": 0.007283204700797796,
+      "step": 25790
+    },
+    {
+      "epoch": 23.201438848920862,
+      "grad_norm": 0.1651083081960678,
+      "learning_rate": 6.545347060068591e-05,
+      "loss": 0.0181,
+      "step": 25800
+    },
+    {
+      "action_loss": 0.0152888810262084,
+      "epoch": 23.201438848920862,
+      "step": 25800
+    },
+    {
+      "epoch": 23.201438848920862,
+      "step": 25800,
+      "torque_loss": 0.1812572032213211
+    },
+    {
+      "epoch": 23.201438848920862,
+      "force_loss": 0.024338068440556526,
+      "step": 25800
+    },
+    {
+      "epoch": 23.21043165467626,
+      "grad_norm": 0.42186781764030457,
+      "learning_rate": 6.542725967015382e-05,
+      "loss": 0.0212,
+      "step": 25810
+    },
+    {
+      "action_loss": 0.006021048873662949,
+      "epoch": 23.21043165467626,
+      "step": 25810
+    },
+    {
+      "epoch": 23.21043165467626,
+      "step": 25810,
+      "torque_loss": 0.14184804260730743
+    },
+    {
+      "epoch": 23.21043165467626,
+      "force_loss": 0.0029388871043920517,
+      "step": 25810
+    },
+    {
+      "epoch": 23.219424460431654,
+      "grad_norm": 0.3582594692707062,
+      "learning_rate": 6.540104405322757e-05,
+      "loss": 0.0199,
+      "step": 25820
+    },
+    {
+      "action_loss": 0.002735581947490573,
+      "epoch": 23.219424460431654,
+      "step": 25820
+    },
+    {
+      "epoch": 23.219424460431654,
+      "step": 25820,
+      "torque_loss": 0.12251565605401993
+    },
+    {
+      "epoch": 23.219424460431654,
+      "force_loss": 0.0027741764206439257,
+      "step": 25820
+    },
+    {
+      "epoch": 23.22841726618705,
+      "grad_norm": 0.3141603469848633,
+      "learning_rate": 6.537482375787077e-05,
+      "loss": 0.0228,
+      "step": 25830
+    },
+    {
+      "action_loss": 0.003992381971329451,
+      "epoch": 23.22841726618705,
+      "step": 25830
+    },
+    {
+      "epoch": 23.22841726618705,
+      "step": 25830,
+      "torque_loss": 0.0965239629149437
+    },
+    {
+      "epoch": 23.22841726618705,
+      "force_loss": 0.0037076727021485567,
+      "step": 25830
+    },
+    {
+      "epoch": 23.237410071942445,
+      "grad_norm": 0.4681563377380371,
+      "learning_rate": 6.534859879204845e-05,
+      "loss": 0.0226,
+      "step": 25840
+    },
+    {
+      "action_loss": 0.015236998908221722,
+      "epoch": 23.237410071942445,
+      "step": 25840
+    },
+    {
+      "epoch": 23.237410071942445,
+      "step": 25840,
+      "torque_loss": 0.14281809329986572
+    },
+    {
+      "epoch": 23.237410071942445,
+      "force_loss": 0.013179329223930836,
+      "step": 25840
+    },
+    {
+      "epoch": 23.246402877697843,
+      "grad_norm": 0.22874422371387482,
+      "learning_rate": 6.532236916372709e-05,
+      "loss": 0.0202,
+      "step": 25850
+    },
+    {
+      "action_loss": 0.017200618982315063,
+      "epoch": 23.246402877697843,
+      "step": 25850
+    },
+    {
+      "epoch": 23.246402877697843,
+      "step": 25850,
+      "torque_loss": 0.12735037505626678
+    },
+    {
+      "epoch": 23.246402877697843,
+      "force_loss": 0.01724884659051895,
+      "step": 25850
+    },
+    {
+      "epoch": 23.255395683453237,
+      "grad_norm": 0.46542343497276306,
+      "learning_rate": 6.529613488087454e-05,
+      "loss": 0.0218,
+      "step": 25860
+    },
+    {
+      "action_loss": 0.0030846295412629843,
+      "epoch": 23.255395683453237,
+      "step": 25860
+    },
+    {
+      "epoch": 23.255395683453237,
+      "step": 25860,
+      "torque_loss": 0.10783765465021133
+    },
+    {
+      "epoch": 23.255395683453237,
+      "force_loss": 0.0020216742996126413,
+      "step": 25860
+    },
+    {
+      "epoch": 23.264388489208635,
+      "grad_norm": 0.3403064012527466,
+      "learning_rate": 6.526989595146009e-05,
+      "loss": 0.0179,
+      "step": 25870
+    },
+    {
+      "action_loss": 0.010160609148442745,
+      "epoch": 23.264388489208635,
+      "step": 25870
+    },
+    {
+      "epoch": 23.264388489208635,
+      "step": 25870,
+      "torque_loss": 0.08791982382535934
+    },
+    {
+      "epoch": 23.264388489208635,
+      "force_loss": 0.004336020443588495,
+      "step": 25870
+    },
+    {
+      "epoch": 23.27338129496403,
+      "grad_norm": 0.31536662578582764,
+      "learning_rate": 6.524365238345441e-05,
+      "loss": 0.0188,
+      "step": 25880
+    },
+    {
+      "action_loss": 0.007521332707256079,
+      "epoch": 23.27338129496403,
+      "step": 25880
+    },
+    {
+      "epoch": 23.27338129496403,
+      "step": 25880,
+      "torque_loss": 0.1705770492553711
+    },
+    {
+      "epoch": 23.27338129496403,
+      "force_loss": 0.008855112828314304,
+      "step": 25880
+    },
+    {
+      "epoch": 23.282374100719423,
+      "grad_norm": 0.4632514417171478,
+      "learning_rate": 6.521740418482964e-05,
+      "loss": 0.0221,
+      "step": 25890
+    },
+    {
+      "action_loss": 0.0031656965147703886,
+      "epoch": 23.282374100719423,
+      "step": 25890
+    },
+    {
+      "epoch": 23.282374100719423,
+      "step": 25890,
+      "torque_loss": 0.13370315730571747
+    },
+    {
+      "epoch": 23.282374100719423,
+      "force_loss": 0.0024058662820607424,
+      "step": 25890
+    },
+    {
+      "epoch": 23.29136690647482,
+      "grad_norm": 0.1726468801498413,
+      "learning_rate": 6.519115136355925e-05,
+      "loss": 0.0197,
+      "step": 25900
+    },
+    {
+      "action_loss": 0.009261681698262691,
+      "epoch": 23.29136690647482,
+      "step": 25900
+    },
+    {
+      "epoch": 23.29136690647482,
+      "step": 25900,
+      "torque_loss": 0.114000104367733
+    },
+    {
+      "epoch": 23.29136690647482,
+      "force_loss": 0.009563510306179523,
+      "step": 25900
+    },
+    {
+      "epoch": 23.300359712230215,
+      "grad_norm": 0.49877873063087463,
+      "learning_rate": 6.51648939276182e-05,
+      "loss": 0.0213,
+      "step": 25910
+    },
+    {
+      "action_loss": 0.007213458884507418,
+      "epoch": 23.300359712230215,
+      "step": 25910
+    },
+    {
+      "epoch": 23.300359712230215,
+      "step": 25910,
+      "torque_loss": 0.1358034461736679
+    },
+    {
+      "epoch": 23.300359712230215,
+      "force_loss": 0.011465176939964294,
+      "step": 25910
+    },
+    {
+      "epoch": 23.309352517985612,
+      "grad_norm": 0.20892885327339172,
+      "learning_rate": 6.513863188498277e-05,
+      "loss": 0.0211,
+      "step": 25920
+    },
+    {
+      "action_loss": 0.004518096800893545,
+      "epoch": 23.309352517985612,
+      "step": 25920
+    },
+    {
+      "epoch": 23.309352517985612,
+      "step": 25920,
+      "torque_loss": 0.15882180631160736
+    },
+    {
+      "epoch": 23.309352517985612,
+      "force_loss": 0.002603568835183978,
+      "step": 25920
+    },
+    {
+      "epoch": 23.318345323741006,
+      "grad_norm": 0.19917403161525726,
+      "learning_rate": 6.511236524363068e-05,
+      "loss": 0.0212,
+      "step": 25930
+    },
+    {
+      "action_loss": 0.006028077099472284,
+      "epoch": 23.318345323741006,
+      "step": 25930
+    },
+    {
+      "epoch": 23.318345323741006,
+      "step": 25930,
+      "torque_loss": 0.09007533639669418
+    },
+    {
+      "epoch": 23.318345323741006,
+      "force_loss": 0.007944750599563122,
+      "step": 25930
+    },
+    {
+      "epoch": 23.327338129496404,
+      "grad_norm": 0.24674886465072632,
+      "learning_rate": 6.508609401154104e-05,
+      "loss": 0.0243,
+      "step": 25940
+    },
+    {
+      "action_loss": 0.002575724618509412,
+      "epoch": 23.327338129496404,
+      "step": 25940
+    },
+    {
+      "epoch": 23.327338129496404,
+      "step": 25940,
+      "torque_loss": 0.07920623570680618
+    },
+    {
+      "epoch": 23.327338129496404,
+      "force_loss": 0.001899032387882471,
+      "step": 25940
+    },
+    {
+      "epoch": 23.336330935251798,
+      "grad_norm": 0.23422738909721375,
+      "learning_rate": 6.505981819669439e-05,
+      "loss": 0.0228,
+      "step": 25950
+    },
+    {
+      "action_loss": 0.004276049789041281,
+      "epoch": 23.336330935251798,
+      "step": 25950
+    },
+    {
+      "epoch": 23.336330935251798,
+      "step": 25950,
+      "torque_loss": 0.07456272840499878
+    },
+    {
+      "epoch": 23.336330935251798,
+      "force_loss": 0.002772013656795025,
+      "step": 25950
+    },
+    {
+      "epoch": 23.345323741007196,
+      "grad_norm": 0.3791617751121521,
+      "learning_rate": 6.503353780707258e-05,
+      "loss": 0.02,
+      "step": 25960
+    },
+    {
+      "action_loss": 0.004444082733243704,
+      "epoch": 23.345323741007196,
+      "step": 25960
+    },
+    {
+      "epoch": 23.345323741007196,
+      "step": 25960,
+      "torque_loss": 0.09586221724748611
+    },
+    {
+      "epoch": 23.345323741007196,
+      "force_loss": 0.004566045943647623,
+      "step": 25960
+    },
+    {
+      "epoch": 23.35431654676259,
+      "grad_norm": 0.16332946717739105,
+      "learning_rate": 6.500725285065895e-05,
+      "loss": 0.0204,
+      "step": 25970
+    },
+    {
+      "action_loss": 0.00948419887572527,
+      "epoch": 23.35431654676259,
+      "step": 25970
+    },
+    {
+      "epoch": 23.35431654676259,
+      "step": 25970,
+      "torque_loss": 0.14972086250782013
+    },
+    {
+      "epoch": 23.35431654676259,
+      "force_loss": 0.007919912226498127,
+      "step": 25970
+    },
+    {
+      "epoch": 23.363309352517987,
+      "grad_norm": 0.3652385473251343,
+      "learning_rate": 6.498096333543813e-05,
+      "loss": 0.0217,
+      "step": 25980
+    },
+    {
+      "action_loss": 0.007189897820353508,
+      "epoch": 23.363309352517987,
+      "step": 25980
+    },
+    {
+      "epoch": 23.363309352517987,
+      "step": 25980,
+      "torque_loss": 0.10546346753835678
+    },
+    {
+      "epoch": 23.363309352517987,
+      "force_loss": 0.009700865484774113,
+      "step": 25980
+    },
+    {
+      "epoch": 23.37230215827338,
+      "grad_norm": 0.331389844417572,
+      "learning_rate": 6.49546692693962e-05,
+      "loss": 0.0187,
+      "step": 25990
+    },
+    {
+      "action_loss": 0.011981263756752014,
+      "epoch": 23.37230215827338,
+      "step": 25990
+    },
+    {
+      "epoch": 23.37230215827338,
+      "step": 25990,
+      "torque_loss": 0.09773851186037064
+    },
+    {
+      "epoch": 23.37230215827338,
+      "force_loss": 0.011358651332557201,
+      "step": 25990
+    },
+    {
+      "epoch": 23.381294964028775,
+      "grad_norm": 0.42301568388938904,
+      "learning_rate": 6.492837066052059e-05,
+      "loss": 0.0236,
+      "step": 26000
+    },
+    {
+      "action_loss": 0.004145585000514984,
+      "epoch": 23.381294964028775,
+      "step": 26000
+    },
+    {
+      "epoch": 23.381294964028775,
+      "step": 26000,
+      "torque_loss": 0.14076244831085205
+    },
+    {
+      "epoch": 23.381294964028775,
+      "force_loss": 0.00511195370927453,
+      "step": 26000
+    },
+    {
+      "epoch": 23.390287769784173,
+      "grad_norm": 0.33904626965522766,
+      "learning_rate": 6.490206751680014e-05,
+      "loss": 0.0233,
+      "step": 26010
+    },
+    {
+      "action_loss": 0.0027019595727324486,
+      "epoch": 23.390287769784173,
+      "step": 26010
+    },
+    {
+      "epoch": 23.390287769784173,
+      "step": 26010,
+      "torque_loss": 0.11031558364629745
+    },
+    {
+      "epoch": 23.390287769784173,
+      "force_loss": 0.002878331346437335,
+      "step": 26010
+    },
+    {
+      "epoch": 23.399280575539567,
+      "grad_norm": 0.21259473264217377,
+      "learning_rate": 6.487575984622505e-05,
+      "loss": 0.0223,
+      "step": 26020
+    },
+    {
+      "action_loss": 0.008811467327177525,
+      "epoch": 23.399280575539567,
+      "step": 26020
+    },
+    {
+      "epoch": 23.399280575539567,
+      "step": 26020,
+      "torque_loss": 0.12438362091779709
+    },
+    {
+      "epoch": 23.399280575539567,
+      "force_loss": 0.013224628753960133,
+      "step": 26020
+    },
+    {
+      "epoch": 23.408273381294965,
+      "grad_norm": 0.17997276782989502,
+      "learning_rate": 6.484944765678689e-05,
+      "loss": 0.0205,
+      "step": 26030
+    },
+    {
+      "action_loss": 0.003967522643506527,
+      "epoch": 23.408273381294965,
+      "step": 26030
+    },
+    {
+      "epoch": 23.408273381294965,
+      "step": 26030,
+      "torque_loss": 0.11782846599817276
+    },
+    {
+      "epoch": 23.408273381294965,
+      "force_loss": 0.0046944585628807545,
+      "step": 26030
+    },
+    {
+      "epoch": 23.41726618705036,
+      "grad_norm": 0.38092041015625,
+      "learning_rate": 6.482313095647861e-05,
+      "loss": 0.0201,
+      "step": 26040
+    },
+    {
+      "action_loss": 0.0027728062123060226,
+      "epoch": 23.41726618705036,
+      "step": 26040
+    },
+    {
+      "epoch": 23.41726618705036,
+      "step": 26040,
+      "torque_loss": 0.09379864484071732
+    },
+    {
+      "epoch": 23.41726618705036,
+      "force_loss": 0.0022747565526515245,
+      "step": 26040
+    },
+    {
+      "epoch": 23.426258992805757,
+      "grad_norm": 0.2724270522594452,
+      "learning_rate": 6.479680975329451e-05,
+      "loss": 0.0223,
+      "step": 26050
+    },
+    {
+      "action_loss": 0.005345206707715988,
+      "epoch": 23.426258992805757,
+      "step": 26050
+    },
+    {
+      "epoch": 23.426258992805757,
+      "step": 26050,
+      "torque_loss": 0.16053888201713562
+    },
+    {
+      "epoch": 23.426258992805757,
+      "force_loss": 0.003982897847890854,
+      "step": 26050
+    },
+    {
+      "epoch": 23.43525179856115,
+      "grad_norm": 0.17012503743171692,
+      "learning_rate": 6.477048405523031e-05,
+      "loss": 0.0194,
+      "step": 26060
+    },
+    {
+      "action_loss": 0.0031624275725334883,
+      "epoch": 23.43525179856115,
+      "step": 26060
+    },
+    {
+      "epoch": 23.43525179856115,
+      "step": 26060,
+      "torque_loss": 0.1522766798734665
+    },
+    {
+      "epoch": 23.43525179856115,
+      "force_loss": 0.0062133558094501495,
+      "step": 26060
+    },
+    {
+      "epoch": 23.444244604316548,
+      "grad_norm": 0.28893211483955383,
+      "learning_rate": 6.474415387028304e-05,
+      "loss": 0.023,
+      "step": 26070
+    },
+    {
+      "action_loss": 0.0028869889210909605,
+      "epoch": 23.444244604316548,
+      "step": 26070
+    },
+    {
+      "epoch": 23.444244604316548,
+      "step": 26070,
+      "torque_loss": 0.12058921903371811
+    },
+    {
+      "epoch": 23.444244604316548,
+      "force_loss": 0.004029951523989439,
+      "step": 26070
+    },
+    {
+      "epoch": 23.453237410071942,
+      "grad_norm": 0.18101710081100464,
+      "learning_rate": 6.471781920645114e-05,
+      "loss": 0.0183,
+      "step": 26080
+    },
+    {
+      "action_loss": 0.008361860178411007,
+      "epoch": 23.453237410071942,
+      "step": 26080
+    },
+    {
+      "epoch": 23.453237410071942,
+      "step": 26080,
+      "torque_loss": 0.08174052089452744
+    },
+    {
+      "epoch": 23.453237410071942,
+      "force_loss": 0.0051781609654426575,
+      "step": 26080
+    },
+    {
+      "epoch": 23.46223021582734,
+      "grad_norm": 0.6979445815086365,
+      "learning_rate": 6.469148007173434e-05,
+      "loss": 0.0198,
+      "step": 26090
+    },
+    {
+      "action_loss": 0.004804322961717844,
+      "epoch": 23.46223021582734,
+      "step": 26090
+    },
+    {
+      "epoch": 23.46223021582734,
+      "step": 26090,
+      "torque_loss": 0.2135559767484665
+    },
+    {
+      "epoch": 23.46223021582734,
+      "force_loss": 0.0033818751107901335,
+      "step": 26090
+    },
+    {
+      "epoch": 23.471223021582734,
+      "grad_norm": 0.32751765847206116,
+      "learning_rate": 6.466513647413381e-05,
+      "loss": 0.0221,
+      "step": 26100
+    },
+    {
+      "action_loss": 0.003584170015528798,
+      "epoch": 23.471223021582734,
+      "step": 26100
+    },
+    {
+      "epoch": 23.471223021582734,
+      "step": 26100,
+      "torque_loss": 0.13833053410053253
+    },
+    {
+      "epoch": 23.471223021582734,
+      "force_loss": 0.005440997425466776,
+      "step": 26100
+    },
+    {
+      "epoch": 23.480215827338128,
+      "grad_norm": 0.3299177587032318,
+      "learning_rate": 6.463878842165203e-05,
+      "loss": 0.0196,
+      "step": 26110
+    },
+    {
+      "action_loss": 0.0022978943306952715,
+      "epoch": 23.480215827338128,
+      "step": 26110
+    },
+    {
+      "epoch": 23.480215827338128,
+      "step": 26110,
+      "torque_loss": 0.1417805403470993
+    },
+    {
+      "epoch": 23.480215827338128,
+      "force_loss": 0.001809374545700848,
+      "step": 26110
+    },
+    {
+      "epoch": 23.489208633093526,
+      "grad_norm": 0.23897022008895874,
+      "learning_rate": 6.461243592229286e-05,
+      "loss": 0.0218,
+      "step": 26120
+    },
+    {
+      "action_loss": 0.0033195570576936007,
+      "epoch": 23.489208633093526,
+      "step": 26120
+    },
+    {
+      "epoch": 23.489208633093526,
+      "step": 26120,
+      "torque_loss": 0.11360173672437668
+    },
+    {
+      "epoch": 23.489208633093526,
+      "force_loss": 0.0024100469890981913,
+      "step": 26120
+    },
+    {
+      "epoch": 23.49820143884892,
+      "grad_norm": 0.23758238554000854,
+      "learning_rate": 6.458607898406146e-05,
+      "loss": 0.0162,
+      "step": 26130
+    },
+    {
+      "action_loss": 0.0023425498511642218,
+      "epoch": 23.49820143884892,
+      "step": 26130
+    },
+    {
+      "epoch": 23.49820143884892,
+      "step": 26130,
+      "torque_loss": 0.15933813154697418
+    },
+    {
+      "epoch": 23.49820143884892,
+      "force_loss": 0.002432143082842231,
+      "step": 26130
+    },
+    {
+      "epoch": 23.507194244604317,
+      "grad_norm": 0.1860436201095581,
+      "learning_rate": 6.455971761496439e-05,
+      "loss": 0.0162,
+      "step": 26140
+    },
+    {
+      "action_loss": 0.008270727470517159,
+      "epoch": 23.507194244604317,
+      "step": 26140
+    },
+    {
+      "epoch": 23.507194244604317,
+      "step": 26140,
+      "torque_loss": 0.13363142311573029
+    },
+    {
+      "epoch": 23.507194244604317,
+      "force_loss": 0.01450568437576294,
+      "step": 26140
+    },
+    {
+      "epoch": 23.51618705035971,
+      "grad_norm": 0.19848312437534332,
+      "learning_rate": 6.453335182300953e-05,
+      "loss": 0.019,
+      "step": 26150
+    },
+    {
+      "action_loss": 0.00577367702499032,
+      "epoch": 23.51618705035971,
+      "step": 26150
+    },
+    {
+      "epoch": 23.51618705035971,
+      "step": 26150,
+      "torque_loss": 0.14000141620635986
+    },
+    {
+      "epoch": 23.51618705035971,
+      "force_loss": 0.009334307163953781,
+      "step": 26150
+    },
+    {
+      "epoch": 23.52517985611511,
+      "grad_norm": 0.41124746203422546,
+      "learning_rate": 6.450698161620612e-05,
+      "loss": 0.0205,
+      "step": 26160
+    },
+    {
+      "action_loss": 0.0034524749498814344,
+      "epoch": 23.52517985611511,
+      "step": 26160
+    },
+    {
+      "epoch": 23.52517985611511,
+      "step": 26160,
+      "torque_loss": 0.1307174116373062
+    },
+    {
+      "epoch": 23.52517985611511,
+      "force_loss": 0.0021633952856063843,
+      "step": 26160
+    },
+    {
+      "epoch": 23.534172661870503,
+      "grad_norm": 0.40090763568878174,
+      "learning_rate": 6.448060700256473e-05,
+      "loss": 0.0236,
+      "step": 26170
+    },
+    {
+      "action_loss": 0.00533563457429409,
+      "epoch": 23.534172661870503,
+      "step": 26170
+    },
+    {
+      "epoch": 23.534172661870503,
+      "step": 26170,
+      "torque_loss": 0.20002329349517822
+    },
+    {
+      "epoch": 23.534172661870503,
+      "force_loss": 0.004222980234771967,
+      "step": 26170
+    },
+    {
+      "epoch": 23.5431654676259,
+      "grad_norm": 0.2789079248905182,
+      "learning_rate": 6.445422799009726e-05,
+      "loss": 0.0186,
+      "step": 26180
+    },
+    {
+      "action_loss": 0.004568005446344614,
+      "epoch": 23.5431654676259,
+      "step": 26180
+    },
+    {
+      "epoch": 23.5431654676259,
+      "step": 26180,
+      "torque_loss": 0.10053543001413345
+    },
+    {
+      "epoch": 23.5431654676259,
+      "force_loss": 0.0037038931623101234,
+      "step": 26180
+    },
+    {
+      "epoch": 23.552158273381295,
+      "grad_norm": 0.585283637046814,
+      "learning_rate": 6.442784458681699e-05,
+      "loss": 0.0211,
+      "step": 26190
+    },
+    {
+      "action_loss": 0.005224842578172684,
+      "epoch": 23.552158273381295,
+      "step": 26190
+    },
+    {
+      "epoch": 23.552158273381295,
+      "step": 26190,
+      "torque_loss": 0.11199595779180527
+    },
+    {
+      "epoch": 23.552158273381295,
+      "force_loss": 0.0018089317018166184,
+      "step": 26190
+    },
+    {
+      "epoch": 23.56115107913669,
+      "grad_norm": 0.44279634952545166,
+      "learning_rate": 6.440145680073847e-05,
+      "loss": 0.0206,
+      "step": 26200
+    },
+    {
+      "action_loss": 0.003384426236152649,
+      "epoch": 23.56115107913669,
+      "step": 26200
+    },
+    {
+      "epoch": 23.56115107913669,
+      "step": 26200,
+      "torque_loss": 0.12964686751365662
+    },
+    {
+      "epoch": 23.56115107913669,
+      "force_loss": 0.00263400562107563,
+      "step": 26200
+    },
+    {
+      "epoch": 23.570143884892087,
+      "grad_norm": 0.41551002860069275,
+      "learning_rate": 6.437506463987762e-05,
+      "loss": 0.0226,
+      "step": 26210
+    },
+    {
+      "action_loss": 0.006423190236091614,
+      "epoch": 23.570143884892087,
+      "step": 26210
+    },
+    {
+      "epoch": 23.570143884892087,
+      "step": 26210,
+      "torque_loss": 0.14234891533851624
+    },
+    {
+      "epoch": 23.570143884892087,
+      "force_loss": 0.0073415786027908325,
+      "step": 26210
+    },
+    {
+      "epoch": 23.57913669064748,
+      "grad_norm": 0.15321233868598938,
+      "learning_rate": 6.434866811225168e-05,
+      "loss": 0.0176,
+      "step": 26220
+    },
+    {
+      "action_loss": 0.00684645539149642,
+      "epoch": 23.57913669064748,
+      "step": 26220
+    },
+    {
+      "epoch": 23.57913669064748,
+      "step": 26220,
+      "torque_loss": 0.07743039727210999
+    },
+    {
+      "epoch": 23.57913669064748,
+      "force_loss": 0.009907073341310024,
+      "step": 26220
+    },
+    {
+      "epoch": 23.58812949640288,
+      "grad_norm": 0.1923304796218872,
+      "learning_rate": 6.432226722587923e-05,
+      "loss": 0.0227,
+      "step": 26230
+    },
+    {
+      "action_loss": 0.005282173398882151,
+      "epoch": 23.58812949640288,
+      "step": 26230
+    },
+    {
+      "epoch": 23.58812949640288,
+      "step": 26230,
+      "torque_loss": 0.14174269139766693
+    },
+    {
+      "epoch": 23.58812949640288,
+      "force_loss": 0.00299402535893023,
+      "step": 26230
+    },
+    {
+      "epoch": 23.597122302158272,
+      "grad_norm": 0.29661592841148376,
+      "learning_rate": 6.429586198878015e-05,
+      "loss": 0.0201,
+      "step": 26240
+    },
+    {
+      "action_loss": 0.06906166672706604,
+      "epoch": 23.597122302158272,
+      "step": 26240
+    },
+    {
+      "epoch": 23.597122302158272,
+      "step": 26240,
+      "torque_loss": 0.18621863424777985
+    },
+    {
+      "epoch": 23.597122302158272,
+      "force_loss": 0.08686783164739609,
+      "step": 26240
+    },
+    {
+      "epoch": 23.60611510791367,
+      "grad_norm": 0.1875166893005371,
+      "learning_rate": 6.426945240897566e-05,
+      "loss": 0.0264,
+      "step": 26250
+    },
+    {
+      "action_loss": 0.007652702275663614,
+      "epoch": 23.60611510791367,
+      "step": 26250
+    },
+    {
+      "epoch": 23.60611510791367,
+      "step": 26250,
+      "torque_loss": 0.19763286411762238
+    },
+    {
+      "epoch": 23.60611510791367,
+      "force_loss": 0.006358183454722166,
+      "step": 26250
+    },
+    {
+      "epoch": 23.615107913669064,
+      "grad_norm": 0.1927485167980194,
+      "learning_rate": 6.424303849448829e-05,
+      "loss": 0.0203,
+      "step": 26260
+    },
+    {
+      "action_loss": 0.001317686983384192,
+      "epoch": 23.615107913669064,
+      "step": 26260
+    },
+    {
+      "epoch": 23.615107913669064,
+      "step": 26260,
+      "torque_loss": 0.1169653907418251
+    },
+    {
+      "epoch": 23.615107913669064,
+      "force_loss": 0.001565055805258453,
+      "step": 26260
+    },
+    {
+      "epoch": 23.62410071942446,
+      "grad_norm": 0.34047549962997437,
+      "learning_rate": 6.42166202533419e-05,
+      "loss": 0.0194,
+      "step": 26270
+    },
+    {
+      "action_loss": 0.004502443131059408,
+      "epoch": 23.62410071942446,
+      "step": 26270
+    },
+    {
+      "epoch": 23.62410071942446,
+      "step": 26270,
+      "torque_loss": 0.157049298286438
+    },
+    {
+      "epoch": 23.62410071942446,
+      "force_loss": 0.0043264105916023254,
+      "step": 26270
+    },
+    {
+      "epoch": 23.633093525179856,
+      "grad_norm": 0.23518070578575134,
+      "learning_rate": 6.419019769356164e-05,
+      "loss": 0.0192,
+      "step": 26280
+    },
+    {
+      "action_loss": 0.005145172122865915,
+      "epoch": 23.633093525179856,
+      "step": 26280
+    },
+    {
+      "epoch": 23.633093525179856,
+      "step": 26280,
+      "torque_loss": 0.14803844690322876
+    },
+    {
+      "epoch": 23.633093525179856,
+      "force_loss": 0.006488631013780832,
+      "step": 26280
+    },
+    {
+      "epoch": 23.642086330935253,
+      "grad_norm": 0.4682842493057251,
+      "learning_rate": 6.416377082317398e-05,
+      "loss": 0.0196,
+      "step": 26290
+    },
+    {
+      "action_loss": 0.0036925256717950106,
+      "epoch": 23.642086330935253,
+      "step": 26290
+    },
+    {
+      "epoch": 23.642086330935253,
+      "step": 26290,
+      "torque_loss": 0.12495657056570053
+    },
+    {
+      "epoch": 23.642086330935253,
+      "force_loss": 0.0037626430857926607,
+      "step": 26290
+    },
+    {
+      "epoch": 23.651079136690647,
+      "grad_norm": 0.31905946135520935,
+      "learning_rate": 6.413733965020674e-05,
+      "loss": 0.017,
+      "step": 26300
+    },
+    {
+      "action_loss": 0.008717426098883152,
+      "epoch": 23.651079136690647,
+      "step": 26300
+    },
+    {
+      "epoch": 23.651079136690647,
+      "step": 26300,
+      "torque_loss": 0.09252175688743591
+    },
+    {
+      "epoch": 23.651079136690647,
+      "force_loss": 0.006333283614367247,
+      "step": 26300
+    },
+    {
+      "epoch": 23.66007194244604,
+      "grad_norm": 0.29095759987831116,
+      "learning_rate": 6.411090418268896e-05,
+      "loss": 0.0179,
+      "step": 26310
+    },
+    {
+      "action_loss": 0.005678208079189062,
+      "epoch": 23.66007194244604,
+      "step": 26310
+    },
+    {
+      "epoch": 23.66007194244604,
+      "step": 26310,
+      "torque_loss": 0.12558908760547638
+    },
+    {
+      "epoch": 23.66007194244604,
+      "force_loss": 0.013269971124827862,
+      "step": 26310
+    },
+    {
+      "epoch": 23.66906474820144,
+      "grad_norm": 0.1718796044588089,
+      "learning_rate": 6.408446442865109e-05,
+      "loss": 0.0169,
+      "step": 26320
+    },
+    {
+      "action_loss": 0.0051812659949064255,
+      "epoch": 23.66906474820144,
+      "step": 26320
+    },
+    {
+      "epoch": 23.66906474820144,
+      "step": 26320,
+      "torque_loss": 0.08716332167387009
+    },
+    {
+      "epoch": 23.66906474820144,
+      "force_loss": 0.004113268107175827,
+      "step": 26320
+    },
+    {
+      "epoch": 23.678057553956833,
+      "grad_norm": 0.24299782514572144,
+      "learning_rate": 6.405802039612479e-05,
+      "loss": 0.0231,
+      "step": 26330
+    },
+    {
+      "action_loss": 0.003167175455018878,
+      "epoch": 23.678057553956833,
+      "step": 26330
+    },
+    {
+      "epoch": 23.678057553956833,
+      "step": 26330,
+      "torque_loss": 0.098402239382267
+    },
+    {
+      "epoch": 23.678057553956833,
+      "force_loss": 0.001812900067307055,
+      "step": 26330
+    },
+    {
+      "epoch": 23.68705035971223,
+      "grad_norm": 0.7098739743232727,
+      "learning_rate": 6.403157209314308e-05,
+      "loss": 0.0216,
+      "step": 26340
+    },
+    {
+      "action_loss": 0.00874798372387886,
+      "epoch": 23.68705035971223,
+      "step": 26340
+    },
+    {
+      "epoch": 23.68705035971223,
+      "step": 26340,
+      "torque_loss": 0.10877111554145813
+    },
+    {
+      "epoch": 23.68705035971223,
+      "force_loss": 0.005317417439073324,
+      "step": 26340
+    },
+    {
+      "epoch": 23.696043165467625,
+      "grad_norm": 0.5119842290878296,
+      "learning_rate": 6.400511952774024e-05,
+      "loss": 0.0212,
+      "step": 26350
+    },
+    {
+      "action_loss": 0.003754744306206703,
+      "epoch": 23.696043165467625,
+      "step": 26350
+    },
+    {
+      "epoch": 23.696043165467625,
+      "step": 26350,
+      "torque_loss": 0.14863841235637665
+    },
+    {
+      "epoch": 23.696043165467625,
+      "force_loss": 0.0018192650750279427,
+      "step": 26350
+    },
+    {
+      "epoch": 23.705035971223023,
+      "grad_norm": 0.3002479076385498,
+      "learning_rate": 6.397866270795187e-05,
+      "loss": 0.0191,
+      "step": 26360
+    },
+    {
+      "action_loss": 0.0065006990917027,
+      "epoch": 23.705035971223023,
+      "step": 26360
+    },
+    {
+      "epoch": 23.705035971223023,
+      "step": 26360,
+      "torque_loss": 0.12278381735086441
+    },
+    {
+      "epoch": 23.705035971223023,
+      "force_loss": 0.009819511324167252,
+      "step": 26360
+    },
+    {
+      "epoch": 23.714028776978417,
+      "grad_norm": 0.33609211444854736,
+      "learning_rate": 6.395220164181489e-05,
+      "loss": 0.0188,
+      "step": 26370
+    },
+    {
+      "action_loss": 0.009771792218089104,
+      "epoch": 23.714028776978417,
+      "step": 26370
+    },
+    {
+      "epoch": 23.714028776978417,
+      "step": 26370,
+      "torque_loss": 0.11551538109779358
+    },
+    {
+      "epoch": 23.714028776978417,
+      "force_loss": 0.01165012177079916,
+      "step": 26370
+    },
+    {
+      "epoch": 23.723021582733814,
+      "grad_norm": 0.27792075276374817,
+      "learning_rate": 6.39257363373674e-05,
+      "loss": 0.0177,
+      "step": 26380
+    },
+    {
+      "action_loss": 0.009556450881063938,
+      "epoch": 23.723021582733814,
+      "step": 26380
+    },
+    {
+      "epoch": 23.723021582733814,
+      "step": 26380,
+      "torque_loss": 0.13581602275371552
+    },
+    {
+      "epoch": 23.723021582733814,
+      "force_loss": 0.020313812419772148,
+      "step": 26380
+    },
+    {
+      "epoch": 23.73201438848921,
+      "grad_norm": 0.22145555913448334,
+      "learning_rate": 6.389926680264892e-05,
+      "loss": 0.022,
+      "step": 26390
+    },
+    {
+      "action_loss": 0.0036732947919517756,
+      "epoch": 23.73201438848921,
+      "step": 26390
+    },
+    {
+      "epoch": 23.73201438848921,
+      "step": 26390,
+      "torque_loss": 0.1463232785463333
+    },
+    {
+      "epoch": 23.73201438848921,
+      "force_loss": 0.003843870246782899,
+      "step": 26390
+    },
+    {
+      "epoch": 23.741007194244606,
+      "grad_norm": 0.43495306372642517,
+      "learning_rate": 6.387279304570017e-05,
+      "loss": 0.0264,
+      "step": 26400
+    },
+    {
+      "action_loss": 0.0023827191907912493,
+      "epoch": 23.741007194244606,
+      "step": 26400
+    },
+    {
+      "epoch": 23.741007194244606,
+      "step": 26400,
+      "torque_loss": 0.11775105446577072
+    },
+    {
+      "epoch": 23.741007194244606,
+      "force_loss": 0.0022756941616535187,
+      "step": 26400
+    },
+    {
+      "epoch": 23.75,
+      "grad_norm": 0.17927949130535126,
+      "learning_rate": 6.384631507456319e-05,
+      "loss": 0.0204,
+      "step": 26410
+    },
+    {
+      "action_loss": 0.006513234693557024,
+      "epoch": 23.75,
+      "step": 26410
+    },
+    {
+      "epoch": 23.75,
+      "step": 26410,
+      "torque_loss": 0.14344631135463715
+    },
+    {
+      "epoch": 23.75,
+      "force_loss": 0.005243296269327402,
+      "step": 26410
+    },
+    {
+      "epoch": 23.758992805755394,
+      "grad_norm": 0.35449185967445374,
+      "learning_rate": 6.381983289728126e-05,
+      "loss": 0.0199,
+      "step": 26420
+    },
+    {
+      "action_loss": 0.0040943678468465805,
+      "epoch": 23.758992805755394,
+      "step": 26420
+    },
+    {
+      "epoch": 23.758992805755394,
+      "step": 26420,
+      "torque_loss": 0.13349874317646027
+    },
+    {
+      "epoch": 23.758992805755394,
+      "force_loss": 0.00539036700502038,
+      "step": 26420
+    },
+    {
+      "epoch": 23.76798561151079,
+      "grad_norm": 0.28454479575157166,
+      "learning_rate": 6.3793346521899e-05,
+      "loss": 0.0213,
+      "step": 26430
+    },
+    {
+      "action_loss": 0.010007210075855255,
+      "epoch": 23.76798561151079,
+      "step": 26430
+    },
+    {
+      "epoch": 23.76798561151079,
+      "step": 26430,
+      "torque_loss": 0.19674928486347198
+    },
+    {
+      "epoch": 23.76798561151079,
+      "force_loss": 0.006232852581888437,
+      "step": 26430
+    },
+    {
+      "epoch": 23.776978417266186,
+      "grad_norm": 0.21880799531936646,
+      "learning_rate": 6.376685595646226e-05,
+      "loss": 0.0218,
+      "step": 26440
+    },
+    {
+      "action_loss": 0.008988688699901104,
+      "epoch": 23.776978417266186,
+      "step": 26440
+    },
+    {
+      "epoch": 23.776978417266186,
+      "step": 26440,
+      "torque_loss": 0.16790910065174103
+    },
+    {
+      "epoch": 23.776978417266186,
+      "force_loss": 0.013626438565552235,
+      "step": 26440
+    },
+    {
+      "epoch": 23.785971223021583,
+      "grad_norm": 0.20643463730812073,
+      "learning_rate": 6.374036120901816e-05,
+      "loss": 0.0216,
+      "step": 26450
+    },
+    {
+      "action_loss": 0.007177682593464851,
+      "epoch": 23.785971223021583,
+      "step": 26450
+    },
+    {
+      "epoch": 23.785971223021583,
+      "step": 26450,
+      "torque_loss": 0.1461370885372162
+    },
+    {
+      "epoch": 23.785971223021583,
+      "force_loss": 0.010485065169632435,
+      "step": 26450
+    },
+    {
+      "epoch": 23.794964028776977,
+      "grad_norm": 0.27559787034988403,
+      "learning_rate": 6.371386228761514e-05,
+      "loss": 0.0195,
+      "step": 26460
+    },
+    {
+      "action_loss": 0.008658151142299175,
+      "epoch": 23.794964028776977,
+      "step": 26460
+    },
+    {
+      "epoch": 23.794964028776977,
+      "step": 26460,
+      "torque_loss": 0.1518176943063736
+    },
+    {
+      "epoch": 23.794964028776977,
+      "force_loss": 0.002371664857491851,
+      "step": 26460
+    },
+    {
+      "epoch": 23.803956834532375,
+      "grad_norm": 0.23754039406776428,
+      "learning_rate": 6.368735920030283e-05,
+      "loss": 0.0263,
+      "step": 26470
+    },
+    {
+      "action_loss": 0.0035001190844923258,
+      "epoch": 23.803956834532375,
+      "step": 26470
+    },
+    {
+      "epoch": 23.803956834532375,
+      "step": 26470,
+      "torque_loss": 0.15118958055973053
+    },
+    {
+      "epoch": 23.803956834532375,
+      "force_loss": 0.005252012982964516,
+      "step": 26470
+    },
+    {
+      "epoch": 23.81294964028777,
+      "grad_norm": 0.17233656346797943,
+      "learning_rate": 6.366085195513218e-05,
+      "loss": 0.0206,
+      "step": 26480
+    },
+    {
+      "action_loss": 0.0035643859300762415,
+      "epoch": 23.81294964028777,
+      "step": 26480
+    },
+    {
+      "epoch": 23.81294964028777,
+      "step": 26480,
+      "torque_loss": 0.18411237001419067
+    },
+    {
+      "epoch": 23.81294964028777,
+      "force_loss": 0.0026406904216855764,
+      "step": 26480
+    },
+    {
+      "epoch": 23.821942446043167,
+      "grad_norm": 0.23081789910793304,
+      "learning_rate": 6.363434056015543e-05,
+      "loss": 0.0252,
+      "step": 26490
+    },
+    {
+      "action_loss": 0.007509656250476837,
+      "epoch": 23.821942446043167,
+      "step": 26490
+    },
+    {
+      "epoch": 23.821942446043167,
+      "step": 26490,
+      "torque_loss": 0.13051988184452057
+    },
+    {
+      "epoch": 23.821942446043167,
+      "force_loss": 0.007185162510722876,
+      "step": 26490
+    },
+    {
+      "epoch": 23.83093525179856,
+      "grad_norm": 0.29633012413978577,
+      "learning_rate": 6.360782502342599e-05,
+      "loss": 0.0192,
+      "step": 26500
+    },
+    {
+      "action_loss": 0.004706589970737696,
+      "epoch": 23.83093525179856,
+      "step": 26500
+    },
+    {
+      "epoch": 23.83093525179856,
+      "step": 26500,
+      "torque_loss": 0.11669912189245224
+    },
+    {
+      "epoch": 23.83093525179856,
+      "force_loss": 0.004882893059402704,
+      "step": 26500
+    },
+    {
+      "epoch": 23.83992805755396,
+      "grad_norm": 0.26792123913764954,
+      "learning_rate": 6.358130535299862e-05,
+      "loss": 0.0249,
+      "step": 26510
+    },
+    {
+      "action_loss": 0.002868209732696414,
+      "epoch": 23.83992805755396,
+      "step": 26510
+    },
+    {
+      "epoch": 23.83992805755396,
+      "step": 26510,
+      "torque_loss": 0.10993697494268417
+    },
+    {
+      "epoch": 23.83992805755396,
+      "force_loss": 0.0021622648928314447,
+      "step": 26510
+    },
+    {
+      "epoch": 23.848920863309353,
+      "grad_norm": 0.3968540132045746,
+      "learning_rate": 6.355478155692926e-05,
+      "loss": 0.0227,
+      "step": 26520
+    },
+    {
+      "action_loss": 0.007956896908581257,
+      "epoch": 23.848920863309353,
+      "step": 26520
+    },
+    {
+      "epoch": 23.848920863309353,
+      "step": 26520,
+      "torque_loss": 0.12510977685451508
+    },
+    {
+      "epoch": 23.848920863309353,
+      "force_loss": 0.0060476139187812805,
+      "step": 26520
+    },
+    {
+      "epoch": 23.857913669064747,
+      "grad_norm": 0.45184972882270813,
+      "learning_rate": 6.352825364327517e-05,
+      "loss": 0.0209,
+      "step": 26530
+    },
+    {
+      "action_loss": 0.015073730610311031,
+      "epoch": 23.857913669064747,
+      "step": 26530
+    },
+    {
+      "epoch": 23.857913669064747,
+      "step": 26530,
+      "torque_loss": 0.12922672927379608
+    },
+    {
+      "epoch": 23.857913669064747,
+      "force_loss": 0.01225163321942091,
+      "step": 26530
+    },
+    {
+      "epoch": 23.866906474820144,
+      "grad_norm": 0.41845831274986267,
+      "learning_rate": 6.350172162009482e-05,
+      "loss": 0.0268,
+      "step": 26540
+    },
+    {
+      "action_loss": 0.014782925136387348,
+      "epoch": 23.866906474820144,
+      "step": 26540
+    },
+    {
+      "epoch": 23.866906474820144,
+      "step": 26540,
+      "torque_loss": 0.13937172293663025
+    },
+    {
+      "epoch": 23.866906474820144,
+      "force_loss": 0.02035324089229107,
+      "step": 26540
+    },
+    {
+      "epoch": 23.87589928057554,
+      "grad_norm": 0.334474116563797,
+      "learning_rate": 6.347518549544793e-05,
+      "loss": 0.0221,
+      "step": 26550
+    },
+    {
+      "action_loss": 0.003326263278722763,
+      "epoch": 23.87589928057554,
+      "step": 26550
+    },
+    {
+      "epoch": 23.87589928057554,
+      "step": 26550,
+      "torque_loss": 0.1415240317583084
+    },
+    {
+      "epoch": 23.87589928057554,
+      "force_loss": 0.0048951092176139355,
+      "step": 26550
+    },
+    {
+      "epoch": 23.884892086330936,
+      "grad_norm": 0.22813700139522552,
+      "learning_rate": 6.344864527739547e-05,
+      "loss": 0.0199,
+      "step": 26560
+    },
+    {
+      "action_loss": 0.004978131968528032,
+      "epoch": 23.884892086330936,
+      "step": 26560
+    },
+    {
+      "epoch": 23.884892086330936,
+      "step": 26560,
+      "torque_loss": 0.11734786629676819
+    },
+    {
+      "epoch": 23.884892086330936,
+      "force_loss": 0.005578598473221064,
+      "step": 26560
+    },
+    {
+      "epoch": 23.89388489208633,
+      "grad_norm": 0.19286710023880005,
+      "learning_rate": 6.342210097399966e-05,
+      "loss": 0.0204,
+      "step": 26570
+    },
+    {
+      "action_loss": 0.008035891689360142,
+      "epoch": 23.89388489208633,
+      "step": 26570
+    },
+    {
+      "epoch": 23.89388489208633,
+      "step": 26570,
+      "torque_loss": 0.17052531242370605
+    },
+    {
+      "epoch": 23.89388489208633,
+      "force_loss": 0.004830732475966215,
+      "step": 26570
+    },
+    {
+      "epoch": 23.902877697841728,
+      "grad_norm": 0.2038847953081131,
+      "learning_rate": 6.339555259332398e-05,
+      "loss": 0.0274,
+      "step": 26580
+    },
+    {
+      "action_loss": 0.005652909632772207,
+      "epoch": 23.902877697841728,
+      "step": 26580
+    },
+    {
+      "epoch": 23.902877697841728,
+      "step": 26580,
+      "torque_loss": 0.1375824362039566
+    },
+    {
+      "epoch": 23.902877697841728,
+      "force_loss": 0.005405073519796133,
+      "step": 26580
+    },
+    {
+      "epoch": 23.91187050359712,
+      "grad_norm": 0.4427371621131897,
+      "learning_rate": 6.33690001434331e-05,
+      "loss": 0.0246,
+      "step": 26590
+    },
+    {
+      "action_loss": 0.018171582370996475,
+      "epoch": 23.91187050359712,
+      "step": 26590
+    },
+    {
+      "epoch": 23.91187050359712,
+      "step": 26590,
+      "torque_loss": 0.16622121632099152
+    },
+    {
+      "epoch": 23.91187050359712,
+      "force_loss": 0.029015885666012764,
+      "step": 26590
+    },
+    {
+      "epoch": 23.92086330935252,
+      "grad_norm": 0.358405202627182,
+      "learning_rate": 6.334244363239296e-05,
+      "loss": 0.023,
+      "step": 26600
+    },
+    {
+      "action_loss": 0.003429584437981248,
+      "epoch": 23.92086330935252,
+      "step": 26600
+    },
+    {
+      "epoch": 23.92086330935252,
+      "step": 26600,
+      "torque_loss": 0.08238402009010315
+    },
+    {
+      "epoch": 23.92086330935252,
+      "force_loss": 0.002532379003241658,
+      "step": 26600
+    },
+    {
+      "epoch": 23.929856115107913,
+      "grad_norm": 0.13266392052173615,
+      "learning_rate": 6.331588306827073e-05,
+      "loss": 0.018,
+      "step": 26610
+    },
+    {
+      "action_loss": 0.0064096166752278805,
+      "epoch": 23.929856115107913,
+      "step": 26610
+    },
+    {
+      "epoch": 23.929856115107913,
+      "step": 26610,
+      "torque_loss": 0.15423421561717987
+    },
+    {
+      "epoch": 23.929856115107913,
+      "force_loss": 0.009012925438582897,
+      "step": 26610
+    },
+    {
+      "epoch": 23.93884892086331,
+      "grad_norm": 0.1737116128206253,
+      "learning_rate": 6.328931845913483e-05,
+      "loss": 0.0171,
+      "step": 26620
+    },
+    {
+      "action_loss": 0.003073088126257062,
+      "epoch": 23.93884892086331,
+      "step": 26620
+    },
+    {
+      "epoch": 23.93884892086331,
+      "step": 26620,
+      "torque_loss": 0.14280617237091064
+    },
+    {
+      "epoch": 23.93884892086331,
+      "force_loss": 0.0023988857865333557,
+      "step": 26620
+    },
+    {
+      "epoch": 23.947841726618705,
+      "grad_norm": 0.2824406921863556,
+      "learning_rate": 6.326274981305484e-05,
+      "loss": 0.0189,
+      "step": 26630
+    },
+    {
+      "action_loss": 0.007629834581166506,
+      "epoch": 23.947841726618705,
+      "step": 26630
+    },
+    {
+      "epoch": 23.947841726618705,
+      "step": 26630,
+      "torque_loss": 0.12465822696685791
+    },
+    {
+      "epoch": 23.947841726618705,
+      "force_loss": 0.010988034307956696,
+      "step": 26630
+    },
+    {
+      "epoch": 23.9568345323741,
+      "grad_norm": 0.18834997713565826,
+      "learning_rate": 6.323617713810166e-05,
+      "loss": 0.0214,
+      "step": 26640
+    },
+    {
+      "action_loss": 0.003116500796750188,
+      "epoch": 23.9568345323741,
+      "step": 26640
+    },
+    {
+      "epoch": 23.9568345323741,
+      "step": 26640,
+      "torque_loss": 0.07612895220518112
+    },
+    {
+      "epoch": 23.9568345323741,
+      "force_loss": 0.003354916349053383,
+      "step": 26640
+    },
+    {
+      "epoch": 23.965827338129497,
+      "grad_norm": 0.11095096915960312,
+      "learning_rate": 6.320960044234734e-05,
+      "loss": 0.0198,
+      "step": 26650
+    },
+    {
+      "action_loss": 0.010923963040113449,
+      "epoch": 23.965827338129497,
+      "step": 26650
+    },
+    {
+      "epoch": 23.965827338129497,
+      "step": 26650,
+      "torque_loss": 0.1912527084350586
+    },
+    {
+      "epoch": 23.965827338129497,
+      "force_loss": 0.01209546159952879,
+      "step": 26650
+    },
+    {
+      "epoch": 23.97482014388489,
+      "grad_norm": 0.27918171882629395,
+      "learning_rate": 6.318301973386518e-05,
+      "loss": 0.0276,
+      "step": 26660
+    },
+    {
+      "action_loss": 0.006705548148602247,
+      "epoch": 23.97482014388489,
+      "step": 26660
+    },
+    {
+      "epoch": 23.97482014388489,
+      "step": 26660,
+      "torque_loss": 0.16599969565868378
+    },
+    {
+      "epoch": 23.97482014388489,
+      "force_loss": 0.005112719256430864,
+      "step": 26660
+    },
+    {
+      "epoch": 23.98381294964029,
+      "grad_norm": 0.41221168637275696,
+      "learning_rate": 6.315643502072971e-05,
+      "loss": 0.021,
+      "step": 26670
+    },
+    {
+      "action_loss": 0.0027291413862258196,
+      "epoch": 23.98381294964029,
+      "step": 26670
+    },
+    {
+      "epoch": 23.98381294964029,
+      "step": 26670,
+      "torque_loss": 0.1066742017865181
+    },
+    {
+      "epoch": 23.98381294964029,
+      "force_loss": 0.0014603679301217198,
+      "step": 26670
+    },
+    {
+      "epoch": 23.992805755395683,
+      "grad_norm": 0.2750396132469177,
+      "learning_rate": 6.312984631101667e-05,
+      "loss": 0.018,
+      "step": 26680
+    },
+    {
+      "action_loss": 0.004309473093599081,
+      "epoch": 23.992805755395683,
+      "step": 26680
+    },
+    {
+      "epoch": 23.992805755395683,
+      "step": 26680,
+      "torque_loss": 0.12925082445144653
+    },
+    {
+      "epoch": 23.992805755395683,
+      "force_loss": 0.0049682422541081905,
+      "step": 26680
+    },
+    {
+      "epoch": 24.00179856115108,
+      "grad_norm": 0.14965416491031647,
+      "learning_rate": 6.310325361280297e-05,
+      "loss": 0.0195,
+      "step": 26690
+    },
+    {
+      "action_loss": 0.0023170060012489557,
+      "epoch": 24.00179856115108,
+      "step": 26690
+    },
+    {
+      "epoch": 24.00179856115108,
+      "step": 26690,
+      "torque_loss": 0.14102952182292938
+    },
+    {
+      "epoch": 24.00179856115108,
+      "force_loss": 0.002769235521554947,
+      "step": 26690
+    },
+    {
+      "epoch": 24.010791366906474,
+      "grad_norm": 0.5179844498634338,
+      "learning_rate": 6.30766569341668e-05,
+      "loss": 0.0172,
+      "step": 26700
+    },
+    {
+      "action_loss": 0.004210879560559988,
+      "epoch": 24.010791366906474,
+      "step": 26700
+    },
+    {
+      "epoch": 24.010791366906474,
+      "step": 26700,
+      "torque_loss": 0.13975654542446136
+    },
+    {
+      "epoch": 24.010791366906474,
+      "force_loss": 0.013249740935862064,
+      "step": 26700
+    },
+    {
+      "epoch": 24.019784172661872,
+      "grad_norm": 0.27187013626098633,
+      "learning_rate": 6.305005628318753e-05,
+      "loss": 0.0224,
+      "step": 26710
+    },
+    {
+      "action_loss": 0.0035000566858798265,
+      "epoch": 24.019784172661872,
+      "step": 26710
+    },
+    {
+      "epoch": 24.019784172661872,
+      "step": 26710,
+      "torque_loss": 0.12807540595531464
+    },
+    {
+      "epoch": 24.019784172661872,
+      "force_loss": 0.013426161371171474,
+      "step": 26710
+    },
+    {
+      "epoch": 24.028776978417266,
+      "grad_norm": 0.21601654589176178,
+      "learning_rate": 6.302345166794572e-05,
+      "loss": 0.0189,
+      "step": 26720
+    },
+    {
+      "action_loss": 0.013133309781551361,
+      "epoch": 24.028776978417266,
+      "step": 26720
+    },
+    {
+      "epoch": 24.028776978417266,
+      "step": 26720,
+      "torque_loss": 0.09281337261199951
+    },
+    {
+      "epoch": 24.028776978417266,
+      "force_loss": 0.015573568642139435,
+      "step": 26720
+    },
+    {
+      "epoch": 24.037769784172664,
+      "grad_norm": 0.34611037373542786,
+      "learning_rate": 6.299684309652316e-05,
+      "loss": 0.021,
+      "step": 26730
+    },
+    {
+      "action_loss": 0.0021687422413378954,
+      "epoch": 24.037769784172664,
+      "step": 26730
+    },
+    {
+      "epoch": 24.037769784172664,
+      "step": 26730,
+      "torque_loss": 0.08817531913518906
+    },
+    {
+      "epoch": 24.037769784172664,
+      "force_loss": 0.002704254351556301,
+      "step": 26730
+    },
+    {
+      "epoch": 24.046762589928058,
+      "grad_norm": 0.5585370659828186,
+      "learning_rate": 6.297023057700283e-05,
+      "loss": 0.019,
+      "step": 26740
+    },
+    {
+      "action_loss": 0.010020464658737183,
+      "epoch": 24.046762589928058,
+      "step": 26740
+    },
+    {
+      "epoch": 24.046762589928058,
+      "step": 26740,
+      "torque_loss": 0.10950187593698502
+    },
+    {
+      "epoch": 24.046762589928058,
+      "force_loss": 0.010822725482285023,
+      "step": 26740
+    },
+    {
+      "epoch": 24.055755395683452,
+      "grad_norm": 0.3086368143558502,
+      "learning_rate": 6.294361411746891e-05,
+      "loss": 0.0206,
+      "step": 26750
+    },
+    {
+      "action_loss": 0.0038403042126446962,
+      "epoch": 24.055755395683452,
+      "step": 26750
+    },
+    {
+      "epoch": 24.055755395683452,
+      "step": 26750,
+      "torque_loss": 0.14038939774036407
+    },
+    {
+      "epoch": 24.055755395683452,
+      "force_loss": 0.0019277144456282258,
+      "step": 26750
+    },
+    {
+      "epoch": 24.06474820143885,
+      "grad_norm": 0.11394848674535751,
+      "learning_rate": 6.291699372600677e-05,
+      "loss": 0.019,
+      "step": 26760
+    },
+    {
+      "action_loss": 0.005254644900560379,
+      "epoch": 24.06474820143885,
+      "step": 26760
+    },
+    {
+      "epoch": 24.06474820143885,
+      "step": 26760,
+      "torque_loss": 0.07805917412042618
+    },
+    {
+      "epoch": 24.06474820143885,
+      "force_loss": 0.01563364453613758,
+      "step": 26760
+    },
+    {
+      "epoch": 24.073741007194243,
+      "grad_norm": 0.39980843663215637,
+      "learning_rate": 6.2890369410703e-05,
+      "loss": 0.0239,
+      "step": 26770
+    },
+    {
+      "action_loss": 0.0024700958747416735,
+      "epoch": 24.073741007194243,
+      "step": 26770
+    },
+    {
+      "epoch": 24.073741007194243,
+      "step": 26770,
+      "torque_loss": 0.1352996975183487
+    },
+    {
+      "epoch": 24.073741007194243,
+      "force_loss": 0.0021215060260146856,
+      "step": 26770
+    },
+    {
+      "epoch": 24.08273381294964,
+      "grad_norm": 0.1333743929862976,
+      "learning_rate": 6.286374117964534e-05,
+      "loss": 0.0206,
+      "step": 26780
+    },
+    {
+      "action_loss": 0.002684928709641099,
+      "epoch": 24.08273381294964,
+      "step": 26780
+    },
+    {
+      "epoch": 24.08273381294964,
+      "step": 26780,
+      "torque_loss": 0.1584099978208542
+    },
+    {
+      "epoch": 24.08273381294964,
+      "force_loss": 0.0018392599886283278,
+      "step": 26780
+    },
+    {
+      "epoch": 24.091726618705035,
+      "grad_norm": 0.148902028799057,
+      "learning_rate": 6.283710904092277e-05,
+      "loss": 0.0171,
+      "step": 26790
+    },
+    {
+      "action_loss": 0.005456050392240286,
+      "epoch": 24.091726618705035,
+      "step": 26790
+    },
+    {
+      "epoch": 24.091726618705035,
+      "step": 26790,
+      "torque_loss": 0.14530979096889496
+    },
+    {
+      "epoch": 24.091726618705035,
+      "force_loss": 0.006891926284879446,
+      "step": 26790
+    },
+    {
+      "epoch": 24.100719424460433,
+      "grad_norm": 0.34021157026290894,
+      "learning_rate": 6.281047300262542e-05,
+      "loss": 0.0183,
+      "step": 26800
+    },
+    {
+      "action_loss": 0.009326084516942501,
+      "epoch": 24.100719424460433,
+      "step": 26800
+    },
+    {
+      "epoch": 24.100719424460433,
+      "step": 26800,
+      "torque_loss": 0.09304801374673843
+    },
+    {
+      "epoch": 24.100719424460433,
+      "force_loss": 0.00395691953599453,
+      "step": 26800
+    },
+    {
+      "epoch": 24.109712230215827,
+      "grad_norm": 0.4762473404407501,
+      "learning_rate": 6.278383307284461e-05,
+      "loss": 0.0205,
+      "step": 26810
+    },
+    {
+      "action_loss": 0.0042495871894061565,
+      "epoch": 24.109712230215827,
+      "step": 26810
+    },
+    {
+      "epoch": 24.109712230215827,
+      "step": 26810,
+      "torque_loss": 0.14461617171764374
+    },
+    {
+      "epoch": 24.109712230215827,
+      "force_loss": 0.00489812670275569,
+      "step": 26810
+    },
+    {
+      "epoch": 24.118705035971225,
+      "grad_norm": 0.32843729853630066,
+      "learning_rate": 6.275718925967284e-05,
+      "loss": 0.0212,
+      "step": 26820
+    },
+    {
+      "action_loss": 0.007712294813245535,
+      "epoch": 24.118705035971225,
+      "step": 26820
+    },
+    {
+      "epoch": 24.118705035971225,
+      "step": 26820,
+      "torque_loss": 0.14035049080848694
+    },
+    {
+      "epoch": 24.118705035971225,
+      "force_loss": 0.017633654177188873,
+      "step": 26820
+    },
+    {
+      "epoch": 24.12769784172662,
+      "grad_norm": 0.4606230556964874,
+      "learning_rate": 6.273054157120382e-05,
+      "loss": 0.0213,
+      "step": 26830
+    },
+    {
+      "action_loss": 0.008606442250311375,
+      "epoch": 24.12769784172662,
+      "step": 26830
+    },
+    {
+      "epoch": 24.12769784172662,
+      "step": 26830,
+      "torque_loss": 0.14684230089187622
+    },
+    {
+      "epoch": 24.12769784172662,
+      "force_loss": 0.01178403664380312,
+      "step": 26830
+    },
+    {
+      "epoch": 24.136690647482013,
+      "grad_norm": 0.23865234851837158,
+      "learning_rate": 6.270389001553238e-05,
+      "loss": 0.0223,
+      "step": 26840
+    },
+    {
+      "action_loss": 0.005808299873024225,
+      "epoch": 24.136690647482013,
+      "step": 26840
+    },
+    {
+      "epoch": 24.136690647482013,
+      "step": 26840,
+      "torque_loss": 0.10328838229179382
+    },
+    {
+      "epoch": 24.136690647482013,
+      "force_loss": 0.005072253290563822,
+      "step": 26840
+    },
+    {
+      "epoch": 24.14568345323741,
+      "grad_norm": 0.3093968629837036,
+      "learning_rate": 6.26772346007546e-05,
+      "loss": 0.0238,
+      "step": 26850
+    },
+    {
+      "action_loss": 0.004079446662217379,
+      "epoch": 24.14568345323741,
+      "step": 26850
+    },
+    {
+      "epoch": 24.14568345323741,
+      "step": 26850,
+      "torque_loss": 0.10266079753637314
+    },
+    {
+      "epoch": 24.14568345323741,
+      "force_loss": 0.0023946769069880247,
+      "step": 26850
+    },
+    {
+      "epoch": 24.154676258992804,
+      "grad_norm": 0.48012447357177734,
+      "learning_rate": 6.265057533496767e-05,
+      "loss": 0.0206,
+      "step": 26860
+    },
+    {
+      "action_loss": 0.0037369942292571068,
+      "epoch": 24.154676258992804,
+      "step": 26860
+    },
+    {
+      "epoch": 24.154676258992804,
+      "step": 26860,
+      "torque_loss": 0.16007035970687866
+    },
+    {
+      "epoch": 24.154676258992804,
+      "force_loss": 0.0044674742966890335,
+      "step": 26860
+    },
+    {
+      "epoch": 24.163669064748202,
+      "grad_norm": 0.37928563356399536,
+      "learning_rate": 6.262391222626997e-05,
+      "loss": 0.0233,
+      "step": 26870
+    },
+    {
+      "action_loss": 0.002601821906864643,
+      "epoch": 24.163669064748202,
+      "step": 26870
+    },
+    {
+      "epoch": 24.163669064748202,
+      "step": 26870,
+      "torque_loss": 0.15268342196941376
+    },
+    {
+      "epoch": 24.163669064748202,
+      "force_loss": 0.0025890101678669453,
+      "step": 26870
+    },
+    {
+      "epoch": 24.172661870503596,
+      "grad_norm": 0.23401878774166107,
+      "learning_rate": 6.259724528276106e-05,
+      "loss": 0.0253,
+      "step": 26880
+    },
+    {
+      "action_loss": 0.002314871409907937,
+      "epoch": 24.172661870503596,
+      "step": 26880
+    },
+    {
+      "epoch": 24.172661870503596,
+      "step": 26880,
+      "torque_loss": 0.11141131073236465
+    },
+    {
+      "epoch": 24.172661870503596,
+      "force_loss": 0.0026015767361968756,
+      "step": 26880
+    },
+    {
+      "epoch": 24.181654676258994,
+      "grad_norm": 0.6529308557510376,
+      "learning_rate": 6.257057451254162e-05,
+      "loss": 0.0188,
+      "step": 26890
+    },
+    {
+      "action_loss": 0.003148353425785899,
+      "epoch": 24.181654676258994,
+      "step": 26890
+    },
+    {
+      "epoch": 24.181654676258994,
+      "step": 26890,
+      "torque_loss": 0.1261998862028122
+    },
+    {
+      "epoch": 24.181654676258994,
+      "force_loss": 0.010966651141643524,
+      "step": 26890
+    },
+    {
+      "epoch": 24.190647482014388,
+      "grad_norm": 0.21692568063735962,
+      "learning_rate": 6.254389992371357e-05,
+      "loss": 0.0201,
+      "step": 26900
+    },
+    {
+      "action_loss": 0.011829155497252941,
+      "epoch": 24.190647482014388,
+      "step": 26900
+    },
+    {
+      "epoch": 24.190647482014388,
+      "step": 26900,
+      "torque_loss": 0.11395353823900223
+    },
+    {
+      "epoch": 24.190647482014388,
+      "force_loss": 0.012464761734008789,
+      "step": 26900
+    },
+    {
+      "epoch": 24.199640287769785,
+      "grad_norm": 0.5146127939224243,
+      "learning_rate": 6.25172215243799e-05,
+      "loss": 0.0215,
+      "step": 26910
+    },
+    {
+      "action_loss": 0.003134808735921979,
+      "epoch": 24.199640287769785,
+      "step": 26910
+    },
+    {
+      "epoch": 24.199640287769785,
+      "step": 26910,
+      "torque_loss": 0.10611215233802795
+    },
+    {
+      "epoch": 24.199640287769785,
+      "force_loss": 0.008303600363433361,
+      "step": 26910
+    },
+    {
+      "epoch": 24.20863309352518,
+      "grad_norm": 0.16701440513134003,
+      "learning_rate": 6.249053932264486e-05,
+      "loss": 0.02,
+      "step": 26920
+    },
+    {
+      "action_loss": 0.022442683577537537,
+      "epoch": 24.20863309352518,
+      "step": 26920
+    },
+    {
+      "epoch": 24.20863309352518,
+      "step": 26920,
+      "torque_loss": 0.1562359780073166
+    },
+    {
+      "epoch": 24.20863309352518,
+      "force_loss": 0.031283047050237656,
+      "step": 26920
+    },
+    {
+      "epoch": 24.217625899280577,
+      "grad_norm": 0.28590187430381775,
+      "learning_rate": 6.246385332661376e-05,
+      "loss": 0.0215,
+      "step": 26930
+    },
+    {
+      "action_loss": 0.006464321166276932,
+      "epoch": 24.217625899280577,
+      "step": 26930
+    },
+    {
+      "epoch": 24.217625899280577,
+      "step": 26930,
+      "torque_loss": 0.14696358144283295
+    },
+    {
+      "epoch": 24.217625899280577,
+      "force_loss": 0.008115951903164387,
+      "step": 26930
+    },
+    {
+      "epoch": 24.22661870503597,
+      "grad_norm": 0.3090987801551819,
+      "learning_rate": 6.24371635443931e-05,
+      "loss": 0.0202,
+      "step": 26940
+    },
+    {
+      "action_loss": 0.0022081492934376,
+      "epoch": 24.22661870503597,
+      "step": 26940
+    },
+    {
+      "epoch": 24.22661870503597,
+      "step": 26940,
+      "torque_loss": 0.17350207269191742
+    },
+    {
+      "epoch": 24.22661870503597,
+      "force_loss": 0.0024397282395511866,
+      "step": 26940
+    },
+    {
+      "epoch": 24.235611510791365,
+      "grad_norm": 0.21155443787574768,
+      "learning_rate": 6.241046998409054e-05,
+      "loss": 0.0156,
+      "step": 26950
+    },
+    {
+      "action_loss": 0.0036125865299254656,
+      "epoch": 24.235611510791365,
+      "step": 26950
+    },
+    {
+      "epoch": 24.235611510791365,
+      "step": 26950,
+      "torque_loss": 0.07176991552114487
+    },
+    {
+      "epoch": 24.235611510791365,
+      "force_loss": 0.002854272723197937,
+      "step": 26950
+    },
+    {
+      "epoch": 24.244604316546763,
+      "grad_norm": 0.43226367235183716,
+      "learning_rate": 6.238377265381489e-05,
+      "loss": 0.0207,
+      "step": 26960
+    },
+    {
+      "action_loss": 0.005308577325195074,
+      "epoch": 24.244604316546763,
+      "step": 26960
+    },
+    {
+      "epoch": 24.244604316546763,
+      "step": 26960,
+      "torque_loss": 0.13910140097141266
+    },
+    {
+      "epoch": 24.244604316546763,
+      "force_loss": 0.009794993326067924,
+      "step": 26960
+    },
+    {
+      "epoch": 24.253597122302157,
+      "grad_norm": 0.3633691370487213,
+      "learning_rate": 6.235707156167607e-05,
+      "loss": 0.0263,
+      "step": 26970
+    },
+    {
+      "action_loss": 0.003366742981597781,
+      "epoch": 24.253597122302157,
+      "step": 26970
+    },
+    {
+      "epoch": 24.253597122302157,
+      "step": 26970,
+      "torque_loss": 0.11560416221618652
+    },
+    {
+      "epoch": 24.253597122302157,
+      "force_loss": 0.0039924634620547295,
+      "step": 26970
+    },
+    {
+      "epoch": 24.262589928057555,
+      "grad_norm": 0.2185497283935547,
+      "learning_rate": 6.233036671578519e-05,
+      "loss": 0.0265,
+      "step": 26980
+    },
+    {
+      "action_loss": 0.027073144912719727,
+      "epoch": 24.262589928057555,
+      "step": 26980
+    },
+    {
+      "epoch": 24.262589928057555,
+      "step": 26980,
+      "torque_loss": 0.11673716455698013
+    },
+    {
+      "epoch": 24.262589928057555,
+      "force_loss": 0.026011614128947258,
+      "step": 26980
+    },
+    {
+      "epoch": 24.27158273381295,
+      "grad_norm": 0.5467181205749512,
+      "learning_rate": 6.230365812425445e-05,
+      "loss": 0.0227,
+      "step": 26990
+    },
+    {
+      "action_loss": 0.003463304601609707,
+      "epoch": 24.27158273381295,
+      "step": 26990
+    },
+    {
+      "epoch": 24.27158273381295,
+      "step": 26990,
+      "torque_loss": 0.13536104559898376
+    },
+    {
+      "epoch": 24.27158273381295,
+      "force_loss": 0.0035492137540131807,
+      "step": 26990
+    },
+    {
+      "epoch": 24.280575539568346,
+      "grad_norm": 0.30241721868515015,
+      "learning_rate": 6.227694579519724e-05,
+      "loss": 0.0227,
+      "step": 27000
+    },
+    {
+      "action_loss": 0.013331294059753418,
+      "epoch": 24.280575539568346,
+      "step": 27000
+    },
+    {
+      "epoch": 24.280575539568346,
+      "step": 27000,
+      "torque_loss": 0.1690327674150467
+    },
+    {
+      "epoch": 24.280575539568346,
+      "force_loss": 0.021961594000458717,
+      "step": 27000
+    },
+    {
+      "epoch": 24.28956834532374,
+      "grad_norm": 0.19231468439102173,
+      "learning_rate": 6.225022973672805e-05,
+      "loss": 0.0204,
+      "step": 27010
+    },
+    {
+      "action_loss": 0.021512677893042564,
+      "epoch": 24.28956834532374,
+      "step": 27010
+    },
+    {
+      "epoch": 24.28956834532374,
+      "step": 27010,
+      "torque_loss": 0.16950565576553345
+    },
+    {
+      "epoch": 24.28956834532374,
+      "force_loss": 0.026604605838656425,
+      "step": 27010
+    },
+    {
+      "epoch": 24.298561151079138,
+      "grad_norm": 0.2754470407962799,
+      "learning_rate": 6.222350995696253e-05,
+      "loss": 0.0257,
+      "step": 27020
+    },
+    {
+      "action_loss": 0.010160579346120358,
+      "epoch": 24.298561151079138,
+      "step": 27020
+    },
+    {
+      "epoch": 24.298561151079138,
+      "step": 27020,
+      "torque_loss": 0.11716454476118088
+    },
+    {
+      "epoch": 24.298561151079138,
+      "force_loss": 0.017418472096323967,
+      "step": 27020
+    },
+    {
+      "epoch": 24.307553956834532,
+      "grad_norm": 0.45914801955223083,
+      "learning_rate": 6.21967864640174e-05,
+      "loss": 0.0199,
+      "step": 27030
+    },
+    {
+      "action_loss": 0.0032418230548501015,
+      "epoch": 24.307553956834532,
+      "step": 27030
+    },
+    {
+      "epoch": 24.307553956834532,
+      "step": 27030,
+      "torque_loss": 0.12885306775569916
+    },
+    {
+      "epoch": 24.307553956834532,
+      "force_loss": 0.0039643216878175735,
+      "step": 27030
+    },
+    {
+      "epoch": 24.31654676258993,
+      "grad_norm": 0.3468298017978668,
+      "learning_rate": 6.217005926601059e-05,
+      "loss": 0.0174,
+      "step": 27040
+    },
+    {
+      "action_loss": 0.013095776550471783,
+      "epoch": 24.31654676258993,
+      "step": 27040
+    },
+    {
+      "epoch": 24.31654676258993,
+      "step": 27040,
+      "torque_loss": 0.12410322576761246
+    },
+    {
+      "epoch": 24.31654676258993,
+      "force_loss": 0.013539652340114117,
+      "step": 27040
+    },
+    {
+      "epoch": 24.325539568345324,
+      "grad_norm": 0.578674852848053,
+      "learning_rate": 6.214332837106111e-05,
+      "loss": 0.02,
+      "step": 27050
+    },
+    {
+      "action_loss": 0.0065324814058840275,
+      "epoch": 24.325539568345324,
+      "step": 27050
+    },
+    {
+      "epoch": 24.325539568345324,
+      "step": 27050,
+      "torque_loss": 0.09553492814302444
+    },
+    {
+      "epoch": 24.325539568345324,
+      "force_loss": 0.009036104194819927,
+      "step": 27050
+    },
+    {
+      "epoch": 24.334532374100718,
+      "grad_norm": 0.4983639121055603,
+      "learning_rate": 6.21165937872891e-05,
+      "loss": 0.0197,
+      "step": 27060
+    },
+    {
+      "action_loss": 0.030352450907230377,
+      "epoch": 24.334532374100718,
+      "step": 27060
+    },
+    {
+      "epoch": 24.334532374100718,
+      "step": 27060,
+      "torque_loss": 0.1310928910970688
+    },
+    {
+      "epoch": 24.334532374100718,
+      "force_loss": 0.02637692727148533,
+      "step": 27060
+    },
+    {
+      "epoch": 24.343525179856115,
+      "grad_norm": 0.5432457327842712,
+      "learning_rate": 6.208985552281582e-05,
+      "loss": 0.0218,
+      "step": 27070
+    },
+    {
+      "action_loss": 0.026522735133767128,
+      "epoch": 24.343525179856115,
+      "step": 27070
+    },
+    {
+      "epoch": 24.343525179856115,
+      "step": 27070,
+      "torque_loss": 0.1456541270017624
+    },
+    {
+      "epoch": 24.343525179856115,
+      "force_loss": 0.031462885439395905,
+      "step": 27070
+    },
+    {
+      "epoch": 24.35251798561151,
+      "grad_norm": 0.3561514616012573,
+      "learning_rate": 6.206311358576364e-05,
+      "loss": 0.0252,
+      "step": 27080
+    },
+    {
+      "action_loss": 0.006840948015451431,
+      "epoch": 24.35251798561151,
+      "step": 27080
+    },
+    {
+      "epoch": 24.35251798561151,
+      "step": 27080,
+      "torque_loss": 0.18298260867595673
+    },
+    {
+      "epoch": 24.35251798561151,
+      "force_loss": 0.004947637673467398,
+      "step": 27080
+    },
+    {
+      "epoch": 24.361510791366907,
+      "grad_norm": 0.27993276715278625,
+      "learning_rate": 6.203636798425608e-05,
+      "loss": 0.0199,
+      "step": 27090
+    },
+    {
+      "action_loss": 0.018681637942790985,
+      "epoch": 24.361510791366907,
+      "step": 27090
+    },
+    {
+      "epoch": 24.361510791366907,
+      "step": 27090,
+      "torque_loss": 0.18137353658676147
+    },
+    {
+      "epoch": 24.361510791366907,
+      "force_loss": 0.011561229825019836,
+      "step": 27090
+    },
+    {
+      "epoch": 24.3705035971223,
+      "grad_norm": 0.2753047049045563,
+      "learning_rate": 6.20096187264177e-05,
+      "loss": 0.0229,
+      "step": 27100
+    },
+    {
+      "action_loss": 0.009925643913447857,
+      "epoch": 24.3705035971223,
+      "step": 27100
+    },
+    {
+      "epoch": 24.3705035971223,
+      "step": 27100,
+      "torque_loss": 0.11520782858133316
+    },
+    {
+      "epoch": 24.3705035971223,
+      "force_loss": 0.011404546909034252,
+      "step": 27100
+    },
+    {
+      "epoch": 24.3794964028777,
+      "grad_norm": 0.33238694071769714,
+      "learning_rate": 6.198286582037425e-05,
+      "loss": 0.022,
+      "step": 27110
+    },
+    {
+      "action_loss": 0.00781126506626606,
+      "epoch": 24.3794964028777,
+      "step": 27110
+    },
+    {
+      "epoch": 24.3794964028777,
+      "step": 27110,
+      "torque_loss": 0.13570380210876465
+    },
+    {
+      "epoch": 24.3794964028777,
+      "force_loss": 0.004932140465825796,
+      "step": 27110
+    },
+    {
+      "epoch": 24.388489208633093,
+      "grad_norm": 0.17936363816261292,
+      "learning_rate": 6.195610927425256e-05,
+      "loss": 0.0229,
+      "step": 27120
+    },
+    {
+      "action_loss": 0.0032112086191773415,
+      "epoch": 24.388489208633093,
+      "step": 27120
+    },
+    {
+      "epoch": 24.388489208633093,
+      "step": 27120,
+      "torque_loss": 0.1424873024225235
+    },
+    {
+      "epoch": 24.388489208633093,
+      "force_loss": 0.0028595321346074343,
+      "step": 27120
+    },
+    {
+      "epoch": 24.39748201438849,
+      "grad_norm": 0.12826332449913025,
+      "learning_rate": 6.192934909618056e-05,
+      "loss": 0.0164,
+      "step": 27130
+    },
+    {
+      "action_loss": 0.0033648523967713118,
+      "epoch": 24.39748201438849,
+      "step": 27130
+    },
+    {
+      "epoch": 24.39748201438849,
+      "step": 27130,
+      "torque_loss": 0.12207072973251343
+    },
+    {
+      "epoch": 24.39748201438849,
+      "force_loss": 0.002059903694316745,
+      "step": 27130
+    },
+    {
+      "epoch": 24.406474820143885,
+      "grad_norm": 0.12039807438850403,
+      "learning_rate": 6.190258529428728e-05,
+      "loss": 0.0175,
+      "step": 27140
+    },
+    {
+      "action_loss": 0.06430386751890182,
+      "epoch": 24.406474820143885,
+      "step": 27140
+    },
+    {
+      "epoch": 24.406474820143885,
+      "step": 27140,
+      "torque_loss": 0.1655750274658203
+    },
+    {
+      "epoch": 24.406474820143885,
+      "force_loss": 0.07303141802549362,
+      "step": 27140
+    },
+    {
+      "epoch": 24.415467625899282,
+      "grad_norm": 0.2760085165500641,
+      "learning_rate": 6.187581787670285e-05,
+      "loss": 0.0263,
+      "step": 27150
+    },
+    {
+      "action_loss": 0.0025449099484831095,
+      "epoch": 24.415467625899282,
+      "step": 27150
+    },
+    {
+      "epoch": 24.415467625899282,
+      "step": 27150,
+      "torque_loss": 0.12741833925247192
+    },
+    {
+      "epoch": 24.415467625899282,
+      "force_loss": 0.001235464820638299,
+      "step": 27150
+    },
+    {
+      "epoch": 24.424460431654676,
+      "grad_norm": 0.23651844263076782,
+      "learning_rate": 6.184904685155852e-05,
+      "loss": 0.0204,
+      "step": 27160
+    },
+    {
+      "action_loss": 0.023961732164025307,
+      "epoch": 24.424460431654676,
+      "step": 27160
+    },
+    {
+      "epoch": 24.424460431654676,
+      "step": 27160,
+      "torque_loss": 0.17916782200336456
+    },
+    {
+      "epoch": 24.424460431654676,
+      "force_loss": 0.023182248696684837,
+      "step": 27160
+    },
+    {
+      "epoch": 24.43345323741007,
+      "grad_norm": 0.4052882492542267,
+      "learning_rate": 6.18222722269866e-05,
+      "loss": 0.0231,
+      "step": 27170
+    },
+    {
+      "action_loss": 0.003978236578404903,
+      "epoch": 24.43345323741007,
+      "step": 27170
+    },
+    {
+      "epoch": 24.43345323741007,
+      "step": 27170,
+      "torque_loss": 0.1772550791501999
+    },
+    {
+      "epoch": 24.43345323741007,
+      "force_loss": 0.003809179412201047,
+      "step": 27170
+    },
+    {
+      "epoch": 24.442446043165468,
+      "grad_norm": 0.16913823783397675,
+      "learning_rate": 6.179549401112053e-05,
+      "loss": 0.0187,
+      "step": 27180
+    },
+    {
+      "action_loss": 0.005313266068696976,
+      "epoch": 24.442446043165468,
+      "step": 27180
+    },
+    {
+      "epoch": 24.442446043165468,
+      "step": 27180,
+      "torque_loss": 0.15760749578475952
+    },
+    {
+      "epoch": 24.442446043165468,
+      "force_loss": 0.005538178142160177,
+      "step": 27180
+    },
+    {
+      "epoch": 24.451438848920862,
+      "grad_norm": 0.2454654723405838,
+      "learning_rate": 6.176871221209482e-05,
+      "loss": 0.0221,
+      "step": 27190
+    },
+    {
+      "action_loss": 0.00831401627510786,
+      "epoch": 24.451438848920862,
+      "step": 27190
+    },
+    {
+      "epoch": 24.451438848920862,
+      "step": 27190,
+      "torque_loss": 0.13839416205883026
+    },
+    {
+      "epoch": 24.451438848920862,
+      "force_loss": 0.00549171632155776,
+      "step": 27190
+    },
+    {
+      "epoch": 24.46043165467626,
+      "grad_norm": 0.19690600037574768,
+      "learning_rate": 6.174192683804508e-05,
+      "loss": 0.0196,
+      "step": 27200
+    },
+    {
+      "action_loss": 0.0033114359248429537,
+      "epoch": 24.46043165467626,
+      "step": 27200
+    },
+    {
+      "epoch": 24.46043165467626,
+      "step": 27200,
+      "torque_loss": 0.22195708751678467
+    },
+    {
+      "epoch": 24.46043165467626,
+      "force_loss": 0.002759550465270877,
+      "step": 27200
+    },
+    {
+      "epoch": 24.469424460431654,
+      "grad_norm": 0.17955347895622253,
+      "learning_rate": 6.1715137897108e-05,
+      "loss": 0.0193,
+      "step": 27210
+    },
+    {
+      "action_loss": 0.0019358963472768664,
+      "epoch": 24.469424460431654,
+      "step": 27210
+    },
+    {
+      "epoch": 24.469424460431654,
+      "step": 27210,
+      "torque_loss": 0.10629644244909286
+    },
+    {
+      "epoch": 24.469424460431654,
+      "force_loss": 0.0025763895828276873,
+      "step": 27210
+    },
+    {
+      "epoch": 24.47841726618705,
+      "grad_norm": 0.17782427370548248,
+      "learning_rate": 6.168834539742134e-05,
+      "loss": 0.0212,
+      "step": 27220
+    },
+    {
+      "action_loss": 0.005482169333845377,
+      "epoch": 24.47841726618705,
+      "step": 27220
+    },
+    {
+      "epoch": 24.47841726618705,
+      "step": 27220,
+      "torque_loss": 0.2010514736175537
+    },
+    {
+      "epoch": 24.47841726618705,
+      "force_loss": 0.005316395312547684,
+      "step": 27220
+    },
+    {
+      "epoch": 24.487410071942445,
+      "grad_norm": 0.32597222924232483,
+      "learning_rate": 6.166154934712397e-05,
+      "loss": 0.0213,
+      "step": 27230
+    },
+    {
+      "action_loss": 0.0035998879466205835,
+      "epoch": 24.487410071942445,
+      "step": 27230
+    },
+    {
+      "epoch": 24.487410071942445,
+      "step": 27230,
+      "torque_loss": 0.1276254504919052
+    },
+    {
+      "epoch": 24.487410071942445,
+      "force_loss": 0.0035867150872945786,
+      "step": 27230
+    },
+    {
+      "epoch": 24.496402877697843,
+      "grad_norm": 0.3753611743450165,
+      "learning_rate": 6.163474975435581e-05,
+      "loss": 0.0172,
+      "step": 27240
+    },
+    {
+      "action_loss": 0.011537007987499237,
+      "epoch": 24.496402877697843,
+      "step": 27240
+    },
+    {
+      "epoch": 24.496402877697843,
+      "step": 27240,
+      "torque_loss": 0.1387958526611328
+    },
+    {
+      "epoch": 24.496402877697843,
+      "force_loss": 0.008628747425973415,
+      "step": 27240
+    },
+    {
+      "epoch": 24.505395683453237,
+      "grad_norm": 0.295878142118454,
+      "learning_rate": 6.160794662725787e-05,
+      "loss": 0.0234,
+      "step": 27250
+    },
+    {
+      "action_loss": 0.002969893626868725,
+      "epoch": 24.505395683453237,
+      "step": 27250
+    },
+    {
+      "epoch": 24.505395683453237,
+      "step": 27250,
+      "torque_loss": 0.13085995614528656
+    },
+    {
+      "epoch": 24.505395683453237,
+      "force_loss": 0.0024577591102570295,
+      "step": 27250
+    },
+    {
+      "epoch": 24.514388489208635,
+      "grad_norm": 0.36321496963500977,
+      "learning_rate": 6.158113997397222e-05,
+      "loss": 0.0205,
+      "step": 27260
+    },
+    {
+      "action_loss": 0.003001876873895526,
+      "epoch": 24.514388489208635,
+      "step": 27260
+    },
+    {
+      "epoch": 24.514388489208635,
+      "step": 27260,
+      "torque_loss": 0.13007371127605438
+    },
+    {
+      "epoch": 24.514388489208635,
+      "force_loss": 0.005574739072471857,
+      "step": 27260
+    },
+    {
+      "epoch": 24.52338129496403,
+      "grad_norm": 0.2102063000202179,
+      "learning_rate": 6.155432980264205e-05,
+      "loss": 0.0212,
+      "step": 27270
+    },
+    {
+      "action_loss": 0.004698061849921942,
+      "epoch": 24.52338129496403,
+      "step": 27270
+    },
+    {
+      "epoch": 24.52338129496403,
+      "step": 27270,
+      "torque_loss": 0.1412065476179123
+    },
+    {
+      "epoch": 24.52338129496403,
+      "force_loss": 0.002502820221707225,
+      "step": 27270
+    },
+    {
+      "epoch": 24.532374100719423,
+      "grad_norm": 0.161998450756073,
+      "learning_rate": 6.152751612141156e-05,
+      "loss": 0.0234,
+      "step": 27280
+    },
+    {
+      "action_loss": 0.0026472285389900208,
+      "epoch": 24.532374100719423,
+      "step": 27280
+    },
+    {
+      "epoch": 24.532374100719423,
+      "step": 27280,
+      "torque_loss": 0.10982010513544083
+    },
+    {
+      "epoch": 24.532374100719423,
+      "force_loss": 0.00370503100566566,
+      "step": 27280
+    },
+    {
+      "epoch": 24.54136690647482,
+      "grad_norm": 0.20840491354465485,
+      "learning_rate": 6.150069893842602e-05,
+      "loss": 0.0206,
+      "step": 27290
+    },
+    {
+      "action_loss": 0.0054594953544437885,
+      "epoch": 24.54136690647482,
+      "step": 27290
+    },
+    {
+      "epoch": 24.54136690647482,
+      "step": 27290,
+      "torque_loss": 0.10830133408308029
+    },
+    {
+      "epoch": 24.54136690647482,
+      "force_loss": 0.003318127943202853,
+      "step": 27290
+    },
+    {
+      "epoch": 24.550359712230215,
+      "grad_norm": 0.5064330697059631,
+      "learning_rate": 6.147387826183182e-05,
+      "loss": 0.0199,
+      "step": 27300
+    },
+    {
+      "action_loss": 0.011984252370893955,
+      "epoch": 24.550359712230215,
+      "step": 27300
+    },
+    {
+      "epoch": 24.550359712230215,
+      "step": 27300,
+      "torque_loss": 0.14927031099796295
+    },
+    {
+      "epoch": 24.550359712230215,
+      "force_loss": 0.013501625508069992,
+      "step": 27300
+    },
+    {
+      "epoch": 24.559352517985612,
+      "grad_norm": 0.4219507575035095,
+      "learning_rate": 6.144705409977635e-05,
+      "loss": 0.0242,
+      "step": 27310
+    },
+    {
+      "action_loss": 0.004247523378580809,
+      "epoch": 24.559352517985612,
+      "step": 27310
+    },
+    {
+      "epoch": 24.559352517985612,
+      "step": 27310,
+      "torque_loss": 0.11038357019424438
+    },
+    {
+      "epoch": 24.559352517985612,
+      "force_loss": 0.004107773769646883,
+      "step": 27310
+    },
+    {
+      "epoch": 24.568345323741006,
+      "grad_norm": 0.2655084729194641,
+      "learning_rate": 6.142022646040808e-05,
+      "loss": 0.0176,
+      "step": 27320
+    },
+    {
+      "action_loss": 0.00814251322299242,
+      "epoch": 24.568345323741006,
+      "step": 27320
+    },
+    {
+      "epoch": 24.568345323741006,
+      "step": 27320,
+      "torque_loss": 0.15498368442058563
+    },
+    {
+      "epoch": 24.568345323741006,
+      "force_loss": 0.007861743681132793,
+      "step": 27320
+    },
+    {
+      "epoch": 24.577338129496404,
+      "grad_norm": 0.14792513847351074,
+      "learning_rate": 6.139339535187653e-05,
+      "loss": 0.0192,
+      "step": 27330
+    },
+    {
+      "action_loss": 0.00445583974942565,
+      "epoch": 24.577338129496404,
+      "step": 27330
+    },
+    {
+      "epoch": 24.577338129496404,
+      "step": 27330,
+      "torque_loss": 0.15024316310882568
+    },
+    {
+      "epoch": 24.577338129496404,
+      "force_loss": 0.007579391356557608,
+      "step": 27330
+    },
+    {
+      "epoch": 24.586330935251798,
+      "grad_norm": 0.270242840051651,
+      "learning_rate": 6.136656078233232e-05,
+      "loss": 0.017,
+      "step": 27340
+    },
+    {
+      "action_loss": 0.0024240335915237665,
+      "epoch": 24.586330935251798,
+      "step": 27340
+    },
+    {
+      "epoch": 24.586330935251798,
+      "step": 27340,
+      "torque_loss": 0.13511238992214203
+    },
+    {
+      "epoch": 24.586330935251798,
+      "force_loss": 0.004678442608565092,
+      "step": 27340
+    },
+    {
+      "epoch": 24.595323741007196,
+      "grad_norm": 0.15835519134998322,
+      "learning_rate": 6.133972275992707e-05,
+      "loss": 0.0221,
+      "step": 27350
+    },
+    {
+      "action_loss": 0.0024552172981202602,
+      "epoch": 24.595323741007196,
+      "step": 27350
+    },
+    {
+      "epoch": 24.595323741007196,
+      "step": 27350,
+      "torque_loss": 0.12071740627288818
+    },
+    {
+      "epoch": 24.595323741007196,
+      "force_loss": 0.0018101140158250928,
+      "step": 27350
+    },
+    {
+      "epoch": 24.60431654676259,
+      "grad_norm": 0.2091076374053955,
+      "learning_rate": 6.131288129281342e-05,
+      "loss": 0.0217,
+      "step": 27360
+    },
+    {
+      "action_loss": 0.010779518634080887,
+      "epoch": 24.60431654676259,
+      "step": 27360
+    },
+    {
+      "epoch": 24.60431654676259,
+      "step": 27360,
+      "torque_loss": 0.1460753232240677
+    },
+    {
+      "epoch": 24.60431654676259,
+      "force_loss": 0.01828053593635559,
+      "step": 27360
+    },
+    {
+      "epoch": 24.613309352517987,
+      "grad_norm": 0.36564478278160095,
+      "learning_rate": 6.128603638914516e-05,
+      "loss": 0.0209,
+      "step": 27370
+    },
+    {
+      "action_loss": 0.006637075450271368,
+      "epoch": 24.613309352517987,
+      "step": 27370
+    },
+    {
+      "epoch": 24.613309352517987,
+      "step": 27370,
+      "torque_loss": 0.11818021535873413
+    },
+    {
+      "epoch": 24.613309352517987,
+      "force_loss": 0.005635035689920187,
+      "step": 27370
+    },
+    {
+      "epoch": 24.62230215827338,
+      "grad_norm": 0.7099848985671997,
+      "learning_rate": 6.125918805707704e-05,
+      "loss": 0.0184,
+      "step": 27380
+    },
+    {
+      "action_loss": 0.014135800302028656,
+      "epoch": 24.62230215827338,
+      "step": 27380
+    },
+    {
+      "epoch": 24.62230215827338,
+      "step": 27380,
+      "torque_loss": 0.14051954448223114
+    },
+    {
+      "epoch": 24.62230215827338,
+      "force_loss": 0.018400678411126137,
+      "step": 27380
+    },
+    {
+      "epoch": 24.631294964028775,
+      "grad_norm": 0.37354403734207153,
+      "learning_rate": 6.123233630476485e-05,
+      "loss": 0.0227,
+      "step": 27390
+    },
+    {
+      "action_loss": 0.0023275529965758324,
+      "epoch": 24.631294964028775,
+      "step": 27390
+    },
+    {
+      "epoch": 24.631294964028775,
+      "step": 27390,
+      "torque_loss": 0.08032269030809402
+    },
+    {
+      "epoch": 24.631294964028775,
+      "force_loss": 0.004027464892715216,
+      "step": 27390
+    },
+    {
+      "epoch": 24.640287769784173,
+      "grad_norm": 0.19231006503105164,
+      "learning_rate": 6.120548114036547e-05,
+      "loss": 0.0193,
+      "step": 27400
+    },
+    {
+      "action_loss": 0.010985583066940308,
+      "epoch": 24.640287769784173,
+      "step": 27400
+    },
+    {
+      "epoch": 24.640287769784173,
+      "step": 27400,
+      "torque_loss": 0.12024138122797012
+    },
+    {
+      "epoch": 24.640287769784173,
+      "force_loss": 0.008251228369772434,
+      "step": 27400
+    },
+    {
+      "epoch": 24.649280575539567,
+      "grad_norm": 0.4444057047367096,
+      "learning_rate": 6.117862257203679e-05,
+      "loss": 0.0203,
+      "step": 27410
+    },
+    {
+      "action_loss": 0.0030823620036244392,
+      "epoch": 24.649280575539567,
+      "step": 27410
+    },
+    {
+      "epoch": 24.649280575539567,
+      "step": 27410,
+      "torque_loss": 0.0979972556233406
+    },
+    {
+      "epoch": 24.649280575539567,
+      "force_loss": 0.0031780097633600235,
+      "step": 27410
+    },
+    {
+      "epoch": 24.658273381294965,
+      "grad_norm": 0.617821216583252,
+      "learning_rate": 6.115176060793771e-05,
+      "loss": 0.0225,
+      "step": 27420
+    },
+    {
+      "action_loss": 0.0027088781353086233,
+      "epoch": 24.658273381294965,
+      "step": 27420
+    },
+    {
+      "epoch": 24.658273381294965,
+      "step": 27420,
+      "torque_loss": 0.17940177023410797
+    },
+    {
+      "epoch": 24.658273381294965,
+      "force_loss": 0.002800578251481056,
+      "step": 27420
+    },
+    {
+      "epoch": 24.66726618705036,
+      "grad_norm": 0.2714875638484955,
+      "learning_rate": 6.112489525622822e-05,
+      "loss": 0.017,
+      "step": 27430
+    },
+    {
+      "action_loss": 0.003958713263273239,
+      "epoch": 24.66726618705036,
+      "step": 27430
+    },
+    {
+      "epoch": 24.66726618705036,
+      "step": 27430,
+      "torque_loss": 0.10730186849832535
+    },
+    {
+      "epoch": 24.66726618705036,
+      "force_loss": 0.004312272649258375,
+      "step": 27430
+    },
+    {
+      "epoch": 24.676258992805757,
+      "grad_norm": 0.4587785005569458,
+      "learning_rate": 6.109802652506928e-05,
+      "loss": 0.0188,
+      "step": 27440
+    },
+    {
+      "action_loss": 0.005403958726674318,
+      "epoch": 24.676258992805757,
+      "step": 27440
+    },
+    {
+      "epoch": 24.676258992805757,
+      "step": 27440,
+      "torque_loss": 0.11038515716791153
+    },
+    {
+      "epoch": 24.676258992805757,
+      "force_loss": 0.006019880995154381,
+      "step": 27440
+    },
+    {
+      "epoch": 24.68525179856115,
+      "grad_norm": 0.5030703544616699,
+      "learning_rate": 6.107115442262291e-05,
+      "loss": 0.0195,
+      "step": 27450
+    },
+    {
+      "action_loss": 0.00563558004796505,
+      "epoch": 24.68525179856115,
+      "step": 27450
+    },
+    {
+      "epoch": 24.68525179856115,
+      "step": 27450,
+      "torque_loss": 0.16837310791015625
+    },
+    {
+      "epoch": 24.68525179856115,
+      "force_loss": 0.008612028323113918,
+      "step": 27450
+    },
+    {
+      "epoch": 24.694244604316548,
+      "grad_norm": 0.3643386960029602,
+      "learning_rate": 6.104427895705214e-05,
+      "loss": 0.0206,
+      "step": 27460
+    },
+    {
+      "action_loss": 0.0025047778617590666,
+      "epoch": 24.694244604316548,
+      "step": 27460
+    },
+    {
+      "epoch": 24.694244604316548,
+      "step": 27460,
+      "torque_loss": 0.10358033329248428
+    },
+    {
+      "epoch": 24.694244604316548,
+      "force_loss": 0.005856382194906473,
+      "step": 27460
+    },
+    {
+      "epoch": 24.703237410071942,
+      "grad_norm": 0.2279098778963089,
+      "learning_rate": 6.101740013652103e-05,
+      "loss": 0.0176,
+      "step": 27470
+    },
+    {
+      "action_loss": 0.006774865090847015,
+      "epoch": 24.703237410071942,
+      "step": 27470
+    },
+    {
+      "epoch": 24.703237410071942,
+      "step": 27470,
+      "torque_loss": 0.15938641130924225
+    },
+    {
+      "epoch": 24.703237410071942,
+      "force_loss": 0.005501022096723318,
+      "step": 27470
+    },
+    {
+      "epoch": 24.71223021582734,
+      "grad_norm": 0.18981735408306122,
+      "learning_rate": 6.099051796919465e-05,
+      "loss": 0.0224,
+      "step": 27480
+    },
+    {
+      "action_loss": 0.06200120225548744,
+      "epoch": 24.71223021582734,
+      "step": 27480
+    },
+    {
+      "epoch": 24.71223021582734,
+      "step": 27480,
+      "torque_loss": 0.22226791083812714
+    },
+    {
+      "epoch": 24.71223021582734,
+      "force_loss": 0.0631442740559578,
+      "step": 27480
+    },
+    {
+      "epoch": 24.721223021582734,
+      "grad_norm": 0.28622543811798096,
+      "learning_rate": 6.096363246323911e-05,
+      "loss": 0.0246,
+      "step": 27490
+    },
+    {
+      "action_loss": 0.0039014138747006655,
+      "epoch": 24.721223021582734,
+      "step": 27490
+    },
+    {
+      "epoch": 24.721223021582734,
+      "step": 27490,
+      "torque_loss": 0.16382472217082977
+    },
+    {
+      "epoch": 24.721223021582734,
+      "force_loss": 0.0027531490195542574,
+      "step": 27490
+    },
+    {
+      "epoch": 24.730215827338128,
+      "grad_norm": 0.28025224804878235,
+      "learning_rate": 6.0936743626821504e-05,
+      "loss": 0.024,
+      "step": 27500
+    },
+    {
+      "action_loss": 0.003201183630153537,
+      "epoch": 24.730215827338128,
+      "step": 27500
+    },
+    {
+      "epoch": 24.730215827338128,
+      "step": 27500,
+      "torque_loss": 0.1074504628777504
+    },
+    {
+      "epoch": 24.730215827338128,
+      "force_loss": 0.004344044253230095,
+      "step": 27500
+    },
+    {
+      "epoch": 24.739208633093526,
+      "grad_norm": 0.49105799198150635,
+      "learning_rate": 6.090985146810996e-05,
+      "loss": 0.0193,
+      "step": 27510
+    },
+    {
+      "action_loss": 0.00403236597776413,
+      "epoch": 24.739208633093526,
+      "step": 27510
+    },
+    {
+      "epoch": 24.739208633093526,
+      "step": 27510,
+      "torque_loss": 0.12196534872055054
+    },
+    {
+      "epoch": 24.739208633093526,
+      "force_loss": 0.006179175805300474,
+      "step": 27510
+    },
+    {
+      "epoch": 24.74820143884892,
+      "grad_norm": 0.4832867681980133,
+      "learning_rate": 6.088295599527357e-05,
+      "loss": 0.0201,
+      "step": 27520
+    },
+    {
+      "action_loss": 0.00697707012295723,
+      "epoch": 24.74820143884892,
+      "step": 27520
+    },
+    {
+      "epoch": 24.74820143884892,
+      "step": 27520,
+      "torque_loss": 0.1521146595478058
+    },
+    {
+      "epoch": 24.74820143884892,
+      "force_loss": 0.01482886541634798,
+      "step": 27520
+    },
+    {
+      "epoch": 24.757194244604317,
+      "grad_norm": 0.20004992187023163,
+      "learning_rate": 6.085605721648252e-05,
+      "loss": 0.0186,
+      "step": 27530
+    },
+    {
+      "action_loss": 0.009972660802304745,
+      "epoch": 24.757194244604317,
+      "step": 27530
+    },
+    {
+      "epoch": 24.757194244604317,
+      "step": 27530,
+      "torque_loss": 0.1113722026348114
+    },
+    {
+      "epoch": 24.757194244604317,
+      "force_loss": 0.016840727999806404,
+      "step": 27530
+    },
+    {
+      "epoch": 24.76618705035971,
+      "grad_norm": 0.6739809513092041,
+      "learning_rate": 6.082915513990792e-05,
+      "loss": 0.0205,
+      "step": 27540
+    },
+    {
+      "action_loss": 0.004202758893370628,
+      "epoch": 24.76618705035971,
+      "step": 27540
+    },
+    {
+      "epoch": 24.76618705035971,
+      "step": 27540,
+      "torque_loss": 0.11291039735078812
+    },
+    {
+      "epoch": 24.76618705035971,
+      "force_loss": 0.00447657098993659,
+      "step": 27540
+    },
+    {
+      "epoch": 24.77517985611511,
+      "grad_norm": 0.20661500096321106,
+      "learning_rate": 6.080224977372192e-05,
+      "loss": 0.0196,
+      "step": 27550
+    },
+    {
+      "action_loss": 0.015018738806247711,
+      "epoch": 24.77517985611511,
+      "step": 27550
+    },
+    {
+      "epoch": 24.77517985611511,
+      "step": 27550,
+      "torque_loss": 0.12736612558364868
+    },
+    {
+      "epoch": 24.77517985611511,
+      "force_loss": 0.008811368606984615,
+      "step": 27550
+    },
+    {
+      "epoch": 24.784172661870503,
+      "grad_norm": 0.30827438831329346,
+      "learning_rate": 6.0775341126097666e-05,
+      "loss": 0.0184,
+      "step": 27560
+    },
+    {
+      "action_loss": 0.004452539142221212,
+      "epoch": 24.784172661870503,
+      "step": 27560
+    },
+    {
+      "epoch": 24.784172661870503,
+      "step": 27560,
+      "torque_loss": 0.13114501535892487
+    },
+    {
+      "epoch": 24.784172661870503,
+      "force_loss": 0.005534987431019545,
+      "step": 27560
+    },
+    {
+      "epoch": 24.7931654676259,
+      "grad_norm": 0.21373234689235687,
+      "learning_rate": 6.074842920520926e-05,
+      "loss": 0.0215,
+      "step": 27570
+    },
+    {
+      "action_loss": 0.003416118212044239,
+      "epoch": 24.7931654676259,
+      "step": 27570
+    },
+    {
+      "epoch": 24.7931654676259,
+      "step": 27570,
+      "torque_loss": 0.13711723685264587
+    },
+    {
+      "epoch": 24.7931654676259,
+      "force_loss": 0.003428607014939189,
+      "step": 27570
+    },
+    {
+      "epoch": 24.802158273381295,
+      "grad_norm": 0.23638705909252167,
+      "learning_rate": 6.072151401923186e-05,
+      "loss": 0.0183,
+      "step": 27580
+    },
+    {
+      "action_loss": 0.005607984960079193,
+      "epoch": 24.802158273381295,
+      "step": 27580
+    },
+    {
+      "epoch": 24.802158273381295,
+      "step": 27580,
+      "torque_loss": 0.1508052498102188
+    },
+    {
+      "epoch": 24.802158273381295,
+      "force_loss": 0.005074174143373966,
+      "step": 27580
+    },
+    {
+      "epoch": 24.81115107913669,
+      "grad_norm": 0.21366402506828308,
+      "learning_rate": 6.069459557634159e-05,
+      "loss": 0.0222,
+      "step": 27590
+    },
+    {
+      "action_loss": 0.0031138500198721886,
+      "epoch": 24.81115107913669,
+      "step": 27590
+    },
+    {
+      "epoch": 24.81115107913669,
+      "step": 27590,
+      "torque_loss": 0.12468931823968887
+    },
+    {
+      "epoch": 24.81115107913669,
+      "force_loss": 0.007567358668893576,
+      "step": 27590
+    },
+    {
+      "epoch": 24.820143884892087,
+      "grad_norm": 0.4918735921382904,
+      "learning_rate": 6.066767388471557e-05,
+      "loss": 0.0226,
+      "step": 27600
+    },
+    {
+      "action_loss": 0.012953785248100758,
+      "epoch": 24.820143884892087,
+      "step": 27600
+    },
+    {
+      "epoch": 24.820143884892087,
+      "step": 27600,
+      "torque_loss": 0.16156236827373505
+    },
+    {
+      "epoch": 24.820143884892087,
+      "force_loss": 0.018003307282924652,
+      "step": 27600
+    },
+    {
+      "epoch": 24.82913669064748,
+      "grad_norm": 0.2150358259677887,
+      "learning_rate": 6.064074895253188e-05,
+      "loss": 0.0219,
+      "step": 27610
+    },
+    {
+      "action_loss": 0.007434766739606857,
+      "epoch": 24.82913669064748,
+      "step": 27610
+    },
+    {
+      "epoch": 24.82913669064748,
+      "step": 27610,
+      "torque_loss": 0.12523068487644196
+    },
+    {
+      "epoch": 24.82913669064748,
+      "force_loss": 0.011812425218522549,
+      "step": 27610
+    },
+    {
+      "epoch": 24.83812949640288,
+      "grad_norm": 0.22323743999004364,
+      "learning_rate": 6.061382078796961e-05,
+      "loss": 0.0252,
+      "step": 27620
+    },
+    {
+      "action_loss": 0.01067497581243515,
+      "epoch": 24.83812949640288,
+      "step": 27620
+    },
+    {
+      "epoch": 24.83812949640288,
+      "step": 27620,
+      "torque_loss": 0.17736369371414185
+    },
+    {
+      "epoch": 24.83812949640288,
+      "force_loss": 0.01045769453048706,
+      "step": 27620
+    },
+    {
+      "epoch": 24.847122302158272,
+      "grad_norm": 0.2883903682231903,
+      "learning_rate": 6.0586889399208814e-05,
+      "loss": 0.0194,
+      "step": 27630
+    },
+    {
+      "action_loss": 0.004756091628223658,
+      "epoch": 24.847122302158272,
+      "step": 27630
+    },
+    {
+      "epoch": 24.847122302158272,
+      "step": 27630,
+      "torque_loss": 0.13929414749145508
+    },
+    {
+      "epoch": 24.847122302158272,
+      "force_loss": 0.005363310221582651,
+      "step": 27630
+    },
+    {
+      "epoch": 24.85611510791367,
+      "grad_norm": 0.6270283460617065,
+      "learning_rate": 6.0559954794430565e-05,
+      "loss": 0.0241,
+      "step": 27640
+    },
+    {
+      "action_loss": 0.023654451593756676,
+      "epoch": 24.85611510791367,
+      "step": 27640
+    },
+    {
+      "epoch": 24.85611510791367,
+      "step": 27640,
+      "torque_loss": 0.15856419503688812
+    },
+    {
+      "epoch": 24.85611510791367,
+      "force_loss": 0.030192309990525246,
+      "step": 27640
+    },
+    {
+      "epoch": 24.865107913669064,
+      "grad_norm": 0.2929843068122864,
+      "learning_rate": 6.053301698181687e-05,
+      "loss": 0.0207,
+      "step": 27650
+    },
+    {
+      "action_loss": 0.002177464310079813,
+      "epoch": 24.865107913669064,
+      "step": 27650
+    },
+    {
+      "epoch": 24.865107913669064,
+      "step": 27650,
+      "torque_loss": 0.17063532769680023
+    },
+    {
+      "epoch": 24.865107913669064,
+      "force_loss": 0.0013247322058305144,
+      "step": 27650
+    },
+    {
+      "epoch": 24.87410071942446,
+      "grad_norm": 0.1990029364824295,
+      "learning_rate": 6.0506075969550725e-05,
+      "loss": 0.0195,
+      "step": 27660
+    },
+    {
+      "action_loss": 0.003913574386388063,
+      "epoch": 24.87410071942446,
+      "step": 27660
+    },
+    {
+      "epoch": 24.87410071942446,
+      "step": 27660,
+      "torque_loss": 0.16323544085025787
+    },
+    {
+      "epoch": 24.87410071942446,
+      "force_loss": 0.0020988702308386564,
+      "step": 27660
+    },
+    {
+      "epoch": 24.883093525179856,
+      "grad_norm": 0.1366661936044693,
+      "learning_rate": 6.047913176581609e-05,
+      "loss": 0.0202,
+      "step": 27670
+    },
+    {
+      "action_loss": 0.0026836830656975508,
+      "epoch": 24.883093525179856,
+      "step": 27670
+    },
+    {
+      "epoch": 24.883093525179856,
+      "step": 27670,
+      "torque_loss": 0.1376323252916336
+    },
+    {
+      "epoch": 24.883093525179856,
+      "force_loss": 0.0028415534179657698,
+      "step": 27670
+    },
+    {
+      "epoch": 24.892086330935253,
+      "grad_norm": 0.7001839280128479,
+      "learning_rate": 6.0452184378797904e-05,
+      "loss": 0.0192,
+      "step": 27680
+    },
+    {
+      "action_loss": 0.006577400956302881,
+      "epoch": 24.892086330935253,
+      "step": 27680
+    },
+    {
+      "epoch": 24.892086330935253,
+      "step": 27680,
+      "torque_loss": 0.10539466142654419
+    },
+    {
+      "epoch": 24.892086330935253,
+      "force_loss": 0.013131454586982727,
+      "step": 27680
+    },
+    {
+      "epoch": 24.901079136690647,
+      "grad_norm": 0.1965656727552414,
+      "learning_rate": 6.042523381668209e-05,
+      "loss": 0.0192,
+      "step": 27690
+    },
+    {
+      "action_loss": 0.007105072494596243,
+      "epoch": 24.901079136690647,
+      "step": 27690
+    },
+    {
+      "epoch": 24.901079136690647,
+      "step": 27690,
+      "torque_loss": 0.08462647348642349
+    },
+    {
+      "epoch": 24.901079136690647,
+      "force_loss": 0.003805331187322736,
+      "step": 27690
+    },
+    {
+      "epoch": 24.91007194244604,
+      "grad_norm": 0.30609291791915894,
+      "learning_rate": 6.03982800876555e-05,
+      "loss": 0.0189,
+      "step": 27700
+    },
+    {
+      "action_loss": 0.03476320207118988,
+      "epoch": 24.91007194244604,
+      "step": 27700
+    },
+    {
+      "epoch": 24.91007194244604,
+      "step": 27700,
+      "torque_loss": 0.16271285712718964
+    },
+    {
+      "epoch": 24.91007194244604,
+      "force_loss": 0.019819073379039764,
+      "step": 27700
+    },
+    {
+      "epoch": 24.91906474820144,
+      "grad_norm": 0.28838419914245605,
+      "learning_rate": 6.0371323199905975e-05,
+      "loss": 0.024,
+      "step": 27710
+    },
+    {
+      "action_loss": 0.0034765799064189196,
+      "epoch": 24.91906474820144,
+      "step": 27710
+    },
+    {
+      "epoch": 24.91906474820144,
+      "step": 27710,
+      "torque_loss": 0.13389629125595093
+    },
+    {
+      "epoch": 24.91906474820144,
+      "force_loss": 0.004095780197530985,
+      "step": 27710
+    },
+    {
+      "epoch": 24.928057553956833,
+      "grad_norm": 0.2344946414232254,
+      "learning_rate": 6.03443631616223e-05,
+      "loss": 0.0207,
+      "step": 27720
+    },
+    {
+      "action_loss": 0.001939621870405972,
+      "epoch": 24.928057553956833,
+      "step": 27720
+    },
+    {
+      "epoch": 24.928057553956833,
+      "step": 27720,
+      "torque_loss": 0.10942911356687546
+    },
+    {
+      "epoch": 24.928057553956833,
+      "force_loss": 0.0010275017702952027,
+      "step": 27720
+    },
+    {
+      "epoch": 24.93705035971223,
+      "grad_norm": 0.470210999250412,
+      "learning_rate": 6.031739998099421e-05,
+      "loss": 0.0183,
+      "step": 27730
+    },
+    {
+      "action_loss": 0.0030129754450172186,
+      "epoch": 24.93705035971223,
+      "step": 27730
+    },
+    {
+      "epoch": 24.93705035971223,
+      "step": 27730,
+      "torque_loss": 0.13927406072616577
+    },
+    {
+      "epoch": 24.93705035971223,
+      "force_loss": 0.003387420205399394,
+      "step": 27730
+    },
+    {
+      "epoch": 24.946043165467625,
+      "grad_norm": 0.27437713742256165,
+      "learning_rate": 6.029043366621243e-05,
+      "loss": 0.0165,
+      "step": 27740
+    },
+    {
+      "action_loss": 0.0032256992999464273,
+      "epoch": 24.946043165467625,
+      "step": 27740
+    },
+    {
+      "epoch": 24.946043165467625,
+      "step": 27740,
+      "torque_loss": 0.1392812728881836
+    },
+    {
+      "epoch": 24.946043165467625,
+      "force_loss": 0.002643261104822159,
+      "step": 27740
+    },
+    {
+      "epoch": 24.955035971223023,
+      "grad_norm": 0.2755166292190552,
+      "learning_rate": 6.0263464225468615e-05,
+      "loss": 0.0226,
+      "step": 27750
+    },
+    {
+      "action_loss": 0.0022924989461898804,
+      "epoch": 24.955035971223023,
+      "step": 27750
+    },
+    {
+      "epoch": 24.955035971223023,
+      "step": 27750,
+      "torque_loss": 0.1275307983160019
+    },
+    {
+      "epoch": 24.955035971223023,
+      "force_loss": 0.003888562321662903,
+      "step": 27750
+    },
+    {
+      "epoch": 24.964028776978417,
+      "grad_norm": 0.2888088524341583,
+      "learning_rate": 6.023649166695534e-05,
+      "loss": 0.0183,
+      "step": 27760
+    },
+    {
+      "action_loss": 0.004289816599339247,
+      "epoch": 24.964028776978417,
+      "step": 27760
+    },
+    {
+      "epoch": 24.964028776978417,
+      "step": 27760,
+      "torque_loss": 0.14794236421585083
+    },
+    {
+      "epoch": 24.964028776978417,
+      "force_loss": 0.008980989456176758,
+      "step": 27760
+    },
+    {
+      "epoch": 24.973021582733814,
+      "grad_norm": 0.21840260922908783,
+      "learning_rate": 6.0209515998866186e-05,
+      "loss": 0.0196,
+      "step": 27770
+    },
+    {
+      "action_loss": 0.0021711259614676237,
+      "epoch": 24.973021582733814,
+      "step": 27770
+    },
+    {
+      "epoch": 24.973021582733814,
+      "step": 27770,
+      "torque_loss": 0.12228868156671524
+    },
+    {
+      "epoch": 24.973021582733814,
+      "force_loss": 0.003480076789855957,
+      "step": 27770
+    },
+    {
+      "epoch": 24.98201438848921,
+      "grad_norm": 0.12162240594625473,
+      "learning_rate": 6.018253722939563e-05,
+      "loss": 0.0203,
+      "step": 27780
+    },
+    {
+      "action_loss": 0.009355143643915653,
+      "epoch": 24.98201438848921,
+      "step": 27780
+    },
+    {
+      "epoch": 24.98201438848921,
+      "step": 27780,
+      "torque_loss": 0.09987518936395645
+    },
+    {
+      "epoch": 24.98201438848921,
+      "force_loss": 0.004401336889714003,
+      "step": 27780
+    },
+    {
+      "epoch": 24.991007194244606,
+      "grad_norm": 0.512770414352417,
+      "learning_rate": 6.015555536673914e-05,
+      "loss": 0.0214,
+      "step": 27790
+    },
+    {
+      "action_loss": 0.003254072042182088,
+      "epoch": 24.991007194244606,
+      "step": 27790
+    },
+    {
+      "epoch": 24.991007194244606,
+      "step": 27790,
+      "torque_loss": 0.14153435826301575
+    },
+    {
+      "epoch": 24.991007194244606,
+      "force_loss": 0.001677804277278483,
+      "step": 27790
+    },
+    {
+      "epoch": 25.0,
+      "grad_norm": 0.2779077887535095,
+      "learning_rate": 6.0128570419093054e-05,
+      "loss": 0.0235,
+      "step": 27800
+    },
+    {
+      "action_loss": 0.009039516560733318,
+      "epoch": 25.0,
+      "step": 27800
+    },
+    {
+      "epoch": 25.0,
+      "step": 27800,
+      "torque_loss": 0.09193138033151627
+    },
+    {
+      "epoch": 25.0,
+      "force_loss": 0.0035536736249923706,
+      "step": 27800
+    },
+    {
+      "epoch": 25.008992805755394,
+      "grad_norm": 0.13003653287887573,
+      "learning_rate": 6.010158239465471e-05,
+      "loss": 0.0196,
+      "step": 27810
+    },
+    {
+      "action_loss": 0.009136388078331947,
+      "epoch": 25.008992805755394,
+      "step": 27810
+    },
+    {
+      "epoch": 25.008992805755394,
+      "step": 27810,
+      "torque_loss": 0.13190436363220215
+    },
+    {
+      "epoch": 25.008992805755394,
+      "force_loss": 0.008251729421317577,
+      "step": 27810
+    },
+    {
+      "epoch": 25.01798561151079,
+      "grad_norm": 0.4156569540500641,
+      "learning_rate": 6.007459130162235e-05,
+      "loss": 0.0207,
+      "step": 27820
+    },
+    {
+      "action_loss": 0.003521420294418931,
+      "epoch": 25.01798561151079,
+      "step": 27820
+    },
+    {
+      "epoch": 25.01798561151079,
+      "step": 27820,
+      "torque_loss": 0.115324966609478
+    },
+    {
+      "epoch": 25.01798561151079,
+      "force_loss": 0.009384633041918278,
+      "step": 27820
+    },
+    {
+      "epoch": 25.026978417266186,
+      "grad_norm": 0.24368155002593994,
+      "learning_rate": 6.004759714819516e-05,
+      "loss": 0.0211,
+      "step": 27830
+    },
+    {
+      "action_loss": 0.001910856575705111,
+      "epoch": 25.026978417266186,
+      "step": 27830
+    },
+    {
+      "epoch": 25.026978417266186,
+      "step": 27830,
+      "torque_loss": 0.10204290598630905
+    },
+    {
+      "epoch": 25.026978417266186,
+      "force_loss": 0.0018200107151642442,
+      "step": 27830
+    },
+    {
+      "epoch": 25.035971223021583,
+      "grad_norm": 0.23502390086650848,
+      "learning_rate": 6.002059994257323e-05,
+      "loss": 0.0236,
+      "step": 27840
+    },
+    {
+      "action_loss": 0.006032480392605066,
+      "epoch": 25.035971223021583,
+      "step": 27840
+    },
+    {
+      "epoch": 25.035971223021583,
+      "step": 27840,
+      "torque_loss": 0.10046689957380295
+    },
+    {
+      "epoch": 25.035971223021583,
+      "force_loss": 0.004050326067954302,
+      "step": 27840
+    },
+    {
+      "epoch": 25.044964028776977,
+      "grad_norm": 0.37774401903152466,
+      "learning_rate": 5.999359969295764e-05,
+      "loss": 0.0204,
+      "step": 27850
+    },
+    {
+      "action_loss": 0.0031546875834465027,
+      "epoch": 25.044964028776977,
+      "step": 27850
+    },
+    {
+      "epoch": 25.044964028776977,
+      "step": 27850,
+      "torque_loss": 0.14888004958629608
+    },
+    {
+      "epoch": 25.044964028776977,
+      "force_loss": 0.00252740946598351,
+      "step": 27850
+    },
+    {
+      "epoch": 25.053956834532375,
+      "grad_norm": 0.2149512618780136,
+      "learning_rate": 5.9966596407550314e-05,
+      "loss": 0.0173,
+      "step": 27860
+    },
+    {
+      "action_loss": 0.011078891344368458,
+      "epoch": 25.053956834532375,
+      "step": 27860
+    },
+    {
+      "epoch": 25.053956834532375,
+      "step": 27860,
+      "torque_loss": 0.13848333060741425
+    },
+    {
+      "epoch": 25.053956834532375,
+      "force_loss": 0.011729366146028042,
+      "step": 27860
+    },
+    {
+      "epoch": 25.06294964028777,
+      "grad_norm": 0.2878969609737396,
+      "learning_rate": 5.993959009455416e-05,
+      "loss": 0.0229,
+      "step": 27870
+    },
+    {
+      "action_loss": 0.007101371884346008,
+      "epoch": 25.06294964028777,
+      "step": 27870
+    },
+    {
+      "epoch": 25.06294964028777,
+      "step": 27870,
+      "torque_loss": 0.1358853280544281
+    },
+    {
+      "epoch": 25.06294964028777,
+      "force_loss": 0.009204776957631111,
+      "step": 27870
+    },
+    {
+      "epoch": 25.071942446043167,
+      "grad_norm": 0.19891761243343353,
+      "learning_rate": 5.991258076217298e-05,
+      "loss": 0.0171,
+      "step": 27880
+    },
+    {
+      "action_loss": 0.013965328224003315,
+      "epoch": 25.071942446043167,
+      "step": 27880
+    },
+    {
+      "epoch": 25.071942446043167,
+      "step": 27880,
+      "torque_loss": 0.15057577192783356
+    },
+    {
+      "epoch": 25.071942446043167,
+      "force_loss": 0.014601573348045349,
+      "step": 27880
+    },
+    {
+      "epoch": 25.08093525179856,
+      "grad_norm": 0.31790658831596375,
+      "learning_rate": 5.988556841861147e-05,
+      "loss": 0.022,
+      "step": 27890
+    },
+    {
+      "action_loss": 0.006994236260652542,
+      "epoch": 25.08093525179856,
+      "step": 27890
+    },
+    {
+      "epoch": 25.08093525179856,
+      "step": 27890,
+      "torque_loss": 0.13063372671604156
+    },
+    {
+      "epoch": 25.08093525179856,
+      "force_loss": 0.0073255919851362705,
+      "step": 27890
+    },
+    {
+      "epoch": 25.08992805755396,
+      "grad_norm": 0.5693853497505188,
+      "learning_rate": 5.985855307207531e-05,
+      "loss": 0.0238,
+      "step": 27900
+    },
+    {
+      "action_loss": 0.012989816255867481,
+      "epoch": 25.08992805755396,
+      "step": 27900
+    },
+    {
+      "epoch": 25.08992805755396,
+      "step": 27900,
+      "torque_loss": 0.11889564990997314
+    },
+    {
+      "epoch": 25.08992805755396,
+      "force_loss": 0.008619645610451698,
+      "step": 27900
+    },
+    {
+      "epoch": 25.098920863309353,
+      "grad_norm": 0.5171234011650085,
+      "learning_rate": 5.9831534730771e-05,
+      "loss": 0.0183,
+      "step": 27910
+    },
+    {
+      "action_loss": 0.005285435821861029,
+      "epoch": 25.098920863309353,
+      "step": 27910
+    },
+    {
+      "epoch": 25.098920863309353,
+      "step": 27910,
+      "torque_loss": 0.113880455493927
+    },
+    {
+      "epoch": 25.098920863309353,
+      "force_loss": 0.004809692036360502,
+      "step": 27910
+    },
+    {
+      "epoch": 25.107913669064747,
+      "grad_norm": 0.21858294308185577,
+      "learning_rate": 5.980451340290605e-05,
+      "loss": 0.0193,
+      "step": 27920
+    },
+    {
+      "action_loss": 0.0023250191006809473,
+      "epoch": 25.107913669064747,
+      "step": 27920
+    },
+    {
+      "epoch": 25.107913669064747,
+      "step": 27920,
+      "torque_loss": 0.10454988479614258
+    },
+    {
+      "epoch": 25.107913669064747,
+      "force_loss": 0.0014954075450077653,
+      "step": 27920
+    },
+    {
+      "epoch": 25.116906474820144,
+      "grad_norm": 0.19935208559036255,
+      "learning_rate": 5.97774890966888e-05,
+      "loss": 0.0173,
+      "step": 27930
+    },
+    {
+      "action_loss": 0.005066411104053259,
+      "epoch": 25.116906474820144,
+      "step": 27930
+    },
+    {
+      "epoch": 25.116906474820144,
+      "step": 27930,
+      "torque_loss": 0.11119184643030167
+    },
+    {
+      "epoch": 25.116906474820144,
+      "force_loss": 0.002239025430753827,
+      "step": 27930
+    },
+    {
+      "epoch": 25.12589928057554,
+      "grad_norm": 0.4909704923629761,
+      "learning_rate": 5.975046182032851e-05,
+      "loss": 0.0201,
+      "step": 27940
+    },
+    {
+      "action_loss": 0.021295903250575066,
+      "epoch": 25.12589928057554,
+      "step": 27940
+    },
+    {
+      "epoch": 25.12589928057554,
+      "step": 27940,
+      "torque_loss": 0.10453003644943237
+    },
+    {
+      "epoch": 25.12589928057554,
+      "force_loss": 0.02358761429786682,
+      "step": 27940
+    },
+    {
+      "epoch": 25.134892086330936,
+      "grad_norm": 0.2959257960319519,
+      "learning_rate": 5.972343158203537e-05,
+      "loss": 0.0202,
+      "step": 27950
+    },
+    {
+      "action_loss": 0.010482084937393665,
+      "epoch": 25.134892086330936,
+      "step": 27950
+    },
+    {
+      "epoch": 25.134892086330936,
+      "step": 27950,
+      "torque_loss": 0.1191159263253212
+    },
+    {
+      "epoch": 25.134892086330936,
+      "force_loss": 0.013482670299708843,
+      "step": 27950
+    },
+    {
+      "epoch": 25.14388489208633,
+      "grad_norm": 0.3135308027267456,
+      "learning_rate": 5.969639839002045e-05,
+      "loss": 0.0183,
+      "step": 27960
+    },
+    {
+      "action_loss": 0.0136274928227067,
+      "epoch": 25.14388489208633,
+      "step": 27960
+    },
+    {
+      "epoch": 25.14388489208633,
+      "step": 27960,
+      "torque_loss": 0.1246398463845253
+    },
+    {
+      "epoch": 25.14388489208633,
+      "force_loss": 0.01063587237149477,
+      "step": 27960
+    },
+    {
+      "epoch": 25.152877697841728,
+      "grad_norm": 0.19451689720153809,
+      "learning_rate": 5.966936225249572e-05,
+      "loss": 0.0198,
+      "step": 27970
+    },
+    {
+      "action_loss": 0.008372939191758633,
+      "epoch": 25.152877697841728,
+      "step": 27970
+    },
+    {
+      "epoch": 25.152877697841728,
+      "step": 27970,
+      "torque_loss": 0.09813759475946426
+    },
+    {
+      "epoch": 25.152877697841728,
+      "force_loss": 0.008229688741266727,
+      "step": 27970
+    },
+    {
+      "epoch": 25.16187050359712,
+      "grad_norm": 0.24217577278614044,
+      "learning_rate": 5.9642323177674044e-05,
+      "loss": 0.0178,
+      "step": 27980
+    },
+    {
+      "action_loss": 0.0019485506927594543,
+      "epoch": 25.16187050359712,
+      "step": 27980
+    },
+    {
+      "epoch": 25.16187050359712,
+      "step": 27980,
+      "torque_loss": 0.12314663082361221
+    },
+    {
+      "epoch": 25.16187050359712,
+      "force_loss": 0.0016669895267114043,
+      "step": 27980
+    },
+    {
+      "epoch": 25.17086330935252,
+      "grad_norm": 0.32251855731010437,
+      "learning_rate": 5.9615281173769154e-05,
+      "loss": 0.0187,
+      "step": 27990
+    },
+    {
+      "action_loss": 0.02054518461227417,
+      "epoch": 25.17086330935252,
+      "step": 27990
+    },
+    {
+      "epoch": 25.17086330935252,
+      "step": 27990,
+      "torque_loss": 0.17750942707061768
+    },
+    {
+      "epoch": 25.17086330935252,
+      "force_loss": 0.012470967136323452,
+      "step": 27990
+    },
+    {
+      "epoch": 25.179856115107913,
+      "grad_norm": 0.1701742261648178,
+      "learning_rate": 5.958823624899574e-05,
+      "loss": 0.0176,
+      "step": 28000
+    },
+    {
+      "action_loss": 0.006592078600078821,
+      "epoch": 25.179856115107913,
+      "step": 28000
+    },
+    {
+      "epoch": 25.179856115107913,
+      "step": 28000,
+      "torque_loss": 0.1097002848982811
+    },
+    {
+      "epoch": 25.179856115107913,
+      "force_loss": 0.002601605374366045,
+      "step": 28000
+    },
+    {
+      "epoch": 25.18884892086331,
+      "grad_norm": 0.3739883303642273,
+      "learning_rate": 5.956118841156933e-05,
+      "loss": 0.0189,
+      "step": 28010
+    },
+    {
+      "action_loss": 0.00389285571873188,
+      "epoch": 25.18884892086331,
+      "step": 28010
+    },
+    {
+      "epoch": 25.18884892086331,
+      "step": 28010,
+      "torque_loss": 0.15238027274608612
+    },
+    {
+      "epoch": 25.18884892086331,
+      "force_loss": 0.006241050083190203,
+      "step": 28010
+    },
+    {
+      "epoch": 25.197841726618705,
+      "grad_norm": 0.14512798190116882,
+      "learning_rate": 5.953413766970631e-05,
+      "loss": 0.0209,
+      "step": 28020
+    },
+    {
+      "action_loss": 0.004007064737379551,
+      "epoch": 25.197841726618705,
+      "step": 28020
+    },
+    {
+      "epoch": 25.197841726618705,
+      "step": 28020,
+      "torque_loss": 0.15665273368358612
+    },
+    {
+      "epoch": 25.197841726618705,
+      "force_loss": 0.002668733475729823,
+      "step": 28020
+    },
+    {
+      "epoch": 25.2068345323741,
+      "grad_norm": 0.24719534814357758,
+      "learning_rate": 5.9507084031624e-05,
+      "loss": 0.0213,
+      "step": 28030
+    },
+    {
+      "action_loss": 0.0036468645557761192,
+      "epoch": 25.2068345323741,
+      "step": 28030
+    },
+    {
+      "epoch": 25.2068345323741,
+      "step": 28030,
+      "torque_loss": 0.11742961406707764
+    },
+    {
+      "epoch": 25.2068345323741,
+      "force_loss": 0.0037753216456621885,
+      "step": 28030
+    },
+    {
+      "epoch": 25.215827338129497,
+      "grad_norm": 0.4668901860713959,
+      "learning_rate": 5.948002750554058e-05,
+      "loss": 0.0225,
+      "step": 28040
+    },
+    {
+      "action_loss": 0.006935764104127884,
+      "epoch": 25.215827338129497,
+      "step": 28040
+    },
+    {
+      "epoch": 25.215827338129497,
+      "step": 28040,
+      "torque_loss": 0.15631915628910065
+    },
+    {
+      "epoch": 25.215827338129497,
+      "force_loss": 0.018520066514611244,
+      "step": 28040
+    },
+    {
+      "epoch": 25.22482014388489,
+      "grad_norm": 0.34241315722465515,
+      "learning_rate": 5.9452968099675124e-05,
+      "loss": 0.0187,
+      "step": 28050
+    },
+    {
+      "action_loss": 0.00364665687084198,
+      "epoch": 25.22482014388489,
+      "step": 28050
+    },
+    {
+      "epoch": 25.22482014388489,
+      "step": 28050,
+      "torque_loss": 0.12001895904541016
+    },
+    {
+      "epoch": 25.22482014388489,
+      "force_loss": 0.002398020587861538,
+      "step": 28050
+    },
+    {
+      "epoch": 25.23381294964029,
+      "grad_norm": 0.3288611173629761,
+      "learning_rate": 5.9425905822247527e-05,
+      "loss": 0.0208,
+      "step": 28060
+    },
+    {
+      "action_loss": 0.008102469146251678,
+      "epoch": 25.23381294964029,
+      "step": 28060
+    },
+    {
+      "epoch": 25.23381294964029,
+      "step": 28060,
+      "torque_loss": 0.1445932537317276
+    },
+    {
+      "epoch": 25.23381294964029,
+      "force_loss": 0.003417715197429061,
+      "step": 28060
+    },
+    {
+      "epoch": 25.242805755395683,
+      "grad_norm": 0.8010976910591125,
+      "learning_rate": 5.939884068147864e-05,
+      "loss": 0.0198,
+      "step": 28070
+    },
+    {
+      "action_loss": 0.0019338391721248627,
+      "epoch": 25.242805755395683,
+      "step": 28070
+    },
+    {
+      "epoch": 25.242805755395683,
+      "step": 28070,
+      "torque_loss": 0.10716181993484497
+    },
+    {
+      "epoch": 25.242805755395683,
+      "force_loss": 0.0027629092801362276,
+      "step": 28070
+    },
+    {
+      "epoch": 25.25179856115108,
+      "grad_norm": 0.2820752263069153,
+      "learning_rate": 5.937177268559011e-05,
+      "loss": 0.0183,
+      "step": 28080
+    },
+    {
+      "action_loss": 0.004666687920689583,
+      "epoch": 25.25179856115108,
+      "step": 28080
+    },
+    {
+      "epoch": 25.25179856115108,
+      "step": 28080,
+      "torque_loss": 0.11430569738149643
+    },
+    {
+      "epoch": 25.25179856115108,
+      "force_loss": 0.0045473817735910416,
+      "step": 28080
+    },
+    {
+      "epoch": 25.260791366906474,
+      "grad_norm": 0.2296970933675766,
+      "learning_rate": 5.934470184280448e-05,
+      "loss": 0.0161,
+      "step": 28090
+    },
+    {
+      "action_loss": 0.0031554270535707474,
+      "epoch": 25.260791366906474,
+      "step": 28090
+    },
+    {
+      "epoch": 25.260791366906474,
+      "step": 28090,
+      "torque_loss": 0.1344473659992218
+    },
+    {
+      "epoch": 25.260791366906474,
+      "force_loss": 0.0032503902912139893,
+      "step": 28090
+    },
+    {
+      "epoch": 25.269784172661872,
+      "grad_norm": 0.3302646577358246,
+      "learning_rate": 5.931762816134516e-05,
+      "loss": 0.0189,
+      "step": 28100
+    },
+    {
+      "action_loss": 0.00474116625264287,
+      "epoch": 25.269784172661872,
+      "step": 28100
+    },
+    {
+      "epoch": 25.269784172661872,
+      "step": 28100,
+      "torque_loss": 0.17696833610534668
+    },
+    {
+      "epoch": 25.269784172661872,
+      "force_loss": 0.0024352208711206913,
+      "step": 28100
+    },
+    {
+      "epoch": 25.278776978417266,
+      "grad_norm": 0.2758291959762573,
+      "learning_rate": 5.9290551649436434e-05,
+      "loss": 0.0204,
+      "step": 28110
+    },
+    {
+      "action_loss": 0.004586617927998304,
+      "epoch": 25.278776978417266,
+      "step": 28110
+    },
+    {
+      "epoch": 25.278776978417266,
+      "step": 28110,
+      "torque_loss": 0.13379128277301788
+    },
+    {
+      "epoch": 25.278776978417266,
+      "force_loss": 0.003668044460937381,
+      "step": 28110
+    },
+    {
+      "epoch": 25.28776978417266,
+      "grad_norm": 0.47487664222717285,
+      "learning_rate": 5.9263472315303416e-05,
+      "loss": 0.0208,
+      "step": 28120
+    },
+    {
+      "action_loss": 0.0049309092573821545,
+      "epoch": 25.28776978417266,
+      "step": 28120
+    },
+    {
+      "epoch": 25.28776978417266,
+      "step": 28120,
+      "torque_loss": 0.13639463484287262
+    },
+    {
+      "epoch": 25.28776978417266,
+      "force_loss": 0.007970382459461689,
+      "step": 28120
+    },
+    {
+      "epoch": 25.296762589928058,
+      "grad_norm": 0.5768781304359436,
+      "learning_rate": 5.9236390167172096e-05,
+      "loss": 0.0196,
+      "step": 28130
+    },
+    {
+      "action_loss": 0.00518127903342247,
+      "epoch": 25.296762589928058,
+      "step": 28130
+    },
+    {
+      "epoch": 25.296762589928058,
+      "step": 28130,
+      "torque_loss": 0.1471085548400879
+    },
+    {
+      "epoch": 25.296762589928058,
+      "force_loss": 0.0037640798836946487,
+      "step": 28130
+    },
+    {
+      "epoch": 25.305755395683452,
+      "grad_norm": 0.3404545485973358,
+      "learning_rate": 5.920930521326932e-05,
+      "loss": 0.0204,
+      "step": 28140
+    },
+    {
+      "action_loss": 0.02042549103498459,
+      "epoch": 25.305755395683452,
+      "step": 28140
+    },
+    {
+      "epoch": 25.305755395683452,
+      "step": 28140,
+      "torque_loss": 0.2146511673927307
+    },
+    {
+      "epoch": 25.305755395683452,
+      "force_loss": 0.03472762927412987,
+      "step": 28140
+    },
+    {
+      "epoch": 25.31474820143885,
+      "grad_norm": 0.1535160094499588,
+      "learning_rate": 5.918221746182276e-05,
+      "loss": 0.0211,
+      "step": 28150
+    },
+    {
+      "action_loss": 0.009396887384355068,
+      "epoch": 25.31474820143885,
+      "step": 28150
+    },
+    {
+      "epoch": 25.31474820143885,
+      "step": 28150,
+      "torque_loss": 0.12705738842487335
+    },
+    {
+      "epoch": 25.31474820143885,
+      "force_loss": 0.0069696661084890366,
+      "step": 28150
+    },
+    {
+      "epoch": 25.323741007194243,
+      "grad_norm": 0.16612853109836578,
+      "learning_rate": 5.9155126921061e-05,
+      "loss": 0.0181,
+      "step": 28160
+    },
+    {
+      "action_loss": 0.0062751155346632,
+      "epoch": 25.323741007194243,
+      "step": 28160
+    },
+    {
+      "epoch": 25.323741007194243,
+      "step": 28160,
+      "torque_loss": 0.13922913372516632
+    },
+    {
+      "epoch": 25.323741007194243,
+      "force_loss": 0.010980424471199512,
+      "step": 28160
+    },
+    {
+      "epoch": 25.33273381294964,
+      "grad_norm": 0.19602370262145996,
+      "learning_rate": 5.91280335992134e-05,
+      "loss": 0.0165,
+      "step": 28170
+    },
+    {
+      "action_loss": 0.0038515112828463316,
+      "epoch": 25.33273381294964,
+      "step": 28170
+    },
+    {
+      "epoch": 25.33273381294964,
+      "step": 28170,
+      "torque_loss": 0.1181967630982399
+    },
+    {
+      "epoch": 25.33273381294964,
+      "force_loss": 0.003768881084397435,
+      "step": 28170
+    },
+    {
+      "epoch": 25.341726618705035,
+      "grad_norm": 0.2185714840888977,
+      "learning_rate": 5.91009375045102e-05,
+      "loss": 0.0195,
+      "step": 28180
+    },
+    {
+      "action_loss": 0.002243555150926113,
+      "epoch": 25.341726618705035,
+      "step": 28180
+    },
+    {
+      "epoch": 25.341726618705035,
+      "step": 28180,
+      "torque_loss": 0.11303848028182983
+    },
+    {
+      "epoch": 25.341726618705035,
+      "force_loss": 0.003053064690902829,
+      "step": 28180
+    },
+    {
+      "epoch": 25.350719424460433,
+      "grad_norm": 0.21856456995010376,
+      "learning_rate": 5.9073838645182476e-05,
+      "loss": 0.0177,
+      "step": 28190
+    },
+    {
+      "action_loss": 0.001767792389728129,
+      "epoch": 25.350719424460433,
+      "step": 28190
+    },
+    {
+      "epoch": 25.350719424460433,
+      "step": 28190,
+      "torque_loss": 0.1569480150938034
+    },
+    {
+      "epoch": 25.350719424460433,
+      "force_loss": 0.0018111895769834518,
+      "step": 28190
+    },
+    {
+      "epoch": 25.359712230215827,
+      "grad_norm": 0.34505677223205566,
+      "learning_rate": 5.904673702946217e-05,
+      "loss": 0.0214,
+      "step": 28200
+    },
+    {
+      "action_loss": 0.008137241937220097,
+      "epoch": 25.359712230215827,
+      "step": 28200
+    },
+    {
+      "epoch": 25.359712230215827,
+      "step": 28200,
+      "torque_loss": 0.11258846521377563
+    },
+    {
+      "epoch": 25.359712230215827,
+      "force_loss": 0.0054153744131326675,
+      "step": 28200
+    },
+    {
+      "epoch": 25.368705035971225,
+      "grad_norm": 0.21682202816009521,
+      "learning_rate": 5.9019632665582004e-05,
+      "loss": 0.0177,
+      "step": 28210
+    },
+    {
+      "action_loss": 0.0020642413292080164,
+      "epoch": 25.368705035971225,
+      "step": 28210
+    },
+    {
+      "epoch": 25.368705035971225,
+      "step": 28210,
+      "torque_loss": 0.1540975421667099
+    },
+    {
+      "epoch": 25.368705035971225,
+      "force_loss": 0.002290738048031926,
+      "step": 28210
+    },
+    {
+      "epoch": 25.37769784172662,
+      "grad_norm": 0.5267677307128906,
+      "learning_rate": 5.899252556177559e-05,
+      "loss": 0.0181,
+      "step": 28220
+    },
+    {
+      "action_loss": 0.009411846287548542,
+      "epoch": 25.37769784172662,
+      "step": 28220
+    },
+    {
+      "epoch": 25.37769784172662,
+      "step": 28220,
+      "torque_loss": 0.09601055830717087
+    },
+    {
+      "epoch": 25.37769784172662,
+      "force_loss": 0.006265006959438324,
+      "step": 28220
+    },
+    {
+      "epoch": 25.386690647482013,
+      "grad_norm": 0.1786697804927826,
+      "learning_rate": 5.896541572627735e-05,
+      "loss": 0.0193,
+      "step": 28230
+    },
+    {
+      "action_loss": 0.002352232113480568,
+      "epoch": 25.386690647482013,
+      "step": 28230
+    },
+    {
+      "epoch": 25.386690647482013,
+      "step": 28230,
+      "torque_loss": 0.1546822041273117
+    },
+    {
+      "epoch": 25.386690647482013,
+      "force_loss": 0.005397506523877382,
+      "step": 28230
+    },
+    {
+      "epoch": 25.39568345323741,
+      "grad_norm": 0.23854495584964752,
+      "learning_rate": 5.893830316732253e-05,
+      "loss": 0.018,
+      "step": 28240
+    },
+    {
+      "action_loss": 0.002737324684858322,
+      "epoch": 25.39568345323741,
+      "step": 28240
+    },
+    {
+      "epoch": 25.39568345323741,
+      "step": 28240,
+      "torque_loss": 0.10519962757825851
+    },
+    {
+      "epoch": 25.39568345323741,
+      "force_loss": 0.004168535117059946,
+      "step": 28240
+    },
+    {
+      "epoch": 25.404676258992804,
+      "grad_norm": 0.4477432370185852,
+      "learning_rate": 5.8911187893147214e-05,
+      "loss": 0.0201,
+      "step": 28250
+    },
+    {
+      "action_loss": 0.0025989983696490526,
+      "epoch": 25.404676258992804,
+      "step": 28250
+    },
+    {
+      "epoch": 25.404676258992804,
+      "step": 28250,
+      "torque_loss": 0.1423255205154419
+    },
+    {
+      "epoch": 25.404676258992804,
+      "force_loss": 0.0019936368335038424,
+      "step": 28250
+    },
+    {
+      "epoch": 25.413669064748202,
+      "grad_norm": 0.3411311209201813,
+      "learning_rate": 5.888406991198828e-05,
+      "loss": 0.0195,
+      "step": 28260
+    },
+    {
+      "action_loss": 0.0035081792157143354,
+      "epoch": 25.413669064748202,
+      "step": 28260
+    },
+    {
+      "epoch": 25.413669064748202,
+      "step": 28260,
+      "torque_loss": 0.11614874005317688
+    },
+    {
+      "epoch": 25.413669064748202,
+      "force_loss": 0.003963755909353495,
+      "step": 28260
+    },
+    {
+      "epoch": 25.422661870503596,
+      "grad_norm": 0.30606362223625183,
+      "learning_rate": 5.885694923208349e-05,
+      "loss": 0.0232,
+      "step": 28270
+    },
+    {
+      "action_loss": 0.009681626223027706,
+      "epoch": 25.422661870503596,
+      "step": 28270
+    },
+    {
+      "epoch": 25.422661870503596,
+      "step": 28270,
+      "torque_loss": 0.12211836129426956
+    },
+    {
+      "epoch": 25.422661870503596,
+      "force_loss": 0.006401500198990107,
+      "step": 28270
+    },
+    {
+      "epoch": 25.431654676258994,
+      "grad_norm": 0.18935605883598328,
+      "learning_rate": 5.882982586167138e-05,
+      "loss": 0.0204,
+      "step": 28280
+    },
+    {
+      "action_loss": 0.0030593425035476685,
+      "epoch": 25.431654676258994,
+      "step": 28280
+    },
+    {
+      "epoch": 25.431654676258994,
+      "step": 28280,
+      "torque_loss": 0.11507879942655563
+    },
+    {
+      "epoch": 25.431654676258994,
+      "force_loss": 0.003175983903929591,
+      "step": 28280
+    },
+    {
+      "epoch": 25.440647482014388,
+      "grad_norm": 0.16558599472045898,
+      "learning_rate": 5.880269980899131e-05,
+      "loss": 0.0191,
+      "step": 28290
+    },
+    {
+      "action_loss": 0.00271633081138134,
+      "epoch": 25.440647482014388,
+      "step": 28290
+    },
+    {
+      "epoch": 25.440647482014388,
+      "step": 28290,
+      "torque_loss": 0.12530149519443512
+    },
+    {
+      "epoch": 25.440647482014388,
+      "force_loss": 0.0022884663194417953,
+      "step": 28290
+    },
+    {
+      "epoch": 25.449640287769785,
+      "grad_norm": 0.19729189574718475,
+      "learning_rate": 5.8775571082283465e-05,
+      "loss": 0.0178,
+      "step": 28300
+    },
+    {
+      "action_loss": 0.004341150168329477,
+      "epoch": 25.449640287769785,
+      "step": 28300
+    },
+    {
+      "epoch": 25.449640287769785,
+      "step": 28300,
+      "torque_loss": 0.11109278351068497
+    },
+    {
+      "epoch": 25.449640287769785,
+      "force_loss": 0.014054469764232635,
+      "step": 28300
+    },
+    {
+      "epoch": 25.45863309352518,
+      "grad_norm": 0.11685407161712646,
+      "learning_rate": 5.8748439689788824e-05,
+      "loss": 0.0211,
+      "step": 28310
+    },
+    {
+      "action_loss": 0.0035375170409679413,
+      "epoch": 25.45863309352518,
+      "step": 28310
+    },
+    {
+      "epoch": 25.45863309352518,
+      "step": 28310,
+      "torque_loss": 0.1559673398733139
+    },
+    {
+      "epoch": 25.45863309352518,
+      "force_loss": 0.004950309172272682,
+      "step": 28310
+    },
+    {
+      "epoch": 25.467625899280577,
+      "grad_norm": 0.22283989191055298,
+      "learning_rate": 5.87213056397492e-05,
+      "loss": 0.0184,
+      "step": 28320
+    },
+    {
+      "action_loss": 0.013143475167453289,
+      "epoch": 25.467625899280577,
+      "step": 28320
+    },
+    {
+      "epoch": 25.467625899280577,
+      "step": 28320,
+      "torque_loss": 0.1311284452676773
+    },
+    {
+      "epoch": 25.467625899280577,
+      "force_loss": 0.013893249444663525,
+      "step": 28320
+    },
+    {
+      "epoch": 25.47661870503597,
+      "grad_norm": 0.16183613240718842,
+      "learning_rate": 5.869416894040719e-05,
+      "loss": 0.0249,
+      "step": 28330
+    },
+    {
+      "action_loss": 0.00408804090693593,
+      "epoch": 25.47661870503597,
+      "step": 28330
+    },
+    {
+      "epoch": 25.47661870503597,
+      "step": 28330,
+      "torque_loss": 0.12245913594961166
+    },
+    {
+      "epoch": 25.47661870503597,
+      "force_loss": 0.003849039552733302,
+      "step": 28330
+    },
+    {
+      "epoch": 25.485611510791365,
+      "grad_norm": 0.10859288275241852,
+      "learning_rate": 5.866702960000621e-05,
+      "loss": 0.0191,
+      "step": 28340
+    },
+    {
+      "action_loss": 0.0031465962529182434,
+      "epoch": 25.485611510791365,
+      "step": 28340
+    },
+    {
+      "epoch": 25.485611510791365,
+      "step": 28340,
+      "torque_loss": 0.1350121945142746
+    },
+    {
+      "epoch": 25.485611510791365,
+      "force_loss": 0.006646886467933655,
+      "step": 28340
+    },
+    {
+      "epoch": 25.494604316546763,
+      "grad_norm": 0.3632952868938446,
+      "learning_rate": 5.863988762679048e-05,
+      "loss": 0.0212,
+      "step": 28350
+    },
+    {
+      "action_loss": 0.0031209969893097878,
+      "epoch": 25.494604316546763,
+      "step": 28350
+    },
+    {
+      "epoch": 25.494604316546763,
+      "step": 28350,
+      "torque_loss": 0.12694773077964783
+    },
+    {
+      "epoch": 25.494604316546763,
+      "force_loss": 0.0029710112139582634,
+      "step": 28350
+    },
+    {
+      "epoch": 25.503597122302157,
+      "grad_norm": 0.33126339316368103,
+      "learning_rate": 5.8612743029005e-05,
+      "loss": 0.0176,
+      "step": 28360
+    },
+    {
+      "action_loss": 0.001578190946020186,
+      "epoch": 25.503597122302157,
+      "step": 28360
+    },
+    {
+      "epoch": 25.503597122302157,
+      "step": 28360,
+      "torque_loss": 0.1360173374414444
+    },
+    {
+      "epoch": 25.503597122302157,
+      "force_loss": 0.0012496812269091606,
+      "step": 28360
+    },
+    {
+      "epoch": 25.512589928057555,
+      "grad_norm": 0.26428744196891785,
+      "learning_rate": 5.858559581489561e-05,
+      "loss": 0.0189,
+      "step": 28370
+    },
+    {
+      "action_loss": 0.017098193988204002,
+      "epoch": 25.512589928057555,
+      "step": 28370
+    },
+    {
+      "epoch": 25.512589928057555,
+      "step": 28370,
+      "torque_loss": 0.17933158576488495
+    },
+    {
+      "epoch": 25.512589928057555,
+      "force_loss": 0.010257538408041,
+      "step": 28370
+    },
+    {
+      "epoch": 25.52158273381295,
+      "grad_norm": 0.17129705846309662,
+      "learning_rate": 5.85584459927089e-05,
+      "loss": 0.0204,
+      "step": 28380
+    },
+    {
+      "action_loss": 0.011494812555611134,
+      "epoch": 25.52158273381295,
+      "step": 28380
+    },
+    {
+      "epoch": 25.52158273381295,
+      "step": 28380,
+      "torque_loss": 0.13039420545101166
+    },
+    {
+      "epoch": 25.52158273381295,
+      "force_loss": 0.016698889434337616,
+      "step": 28380
+    },
+    {
+      "epoch": 25.530575539568346,
+      "grad_norm": 0.18305902183055878,
+      "learning_rate": 5.853129357069227e-05,
+      "loss": 0.0187,
+      "step": 28390
+    },
+    {
+      "action_loss": 0.0047340672463178635,
+      "epoch": 25.530575539568346,
+      "step": 28390
+    },
+    {
+      "epoch": 25.530575539568346,
+      "step": 28390,
+      "torque_loss": 0.1560722142457962
+    },
+    {
+      "epoch": 25.530575539568346,
+      "force_loss": 0.004592828452587128,
+      "step": 28390
+    },
+    {
+      "epoch": 25.53956834532374,
+      "grad_norm": 0.34422537684440613,
+      "learning_rate": 5.8504138557093913e-05,
+      "loss": 0.0203,
+      "step": 28400
+    },
+    {
+      "action_loss": 0.015056215226650238,
+      "epoch": 25.53956834532374,
+      "step": 28400
+    },
+    {
+      "epoch": 25.53956834532374,
+      "step": 28400,
+      "torque_loss": 0.15267789363861084
+    },
+    {
+      "epoch": 25.53956834532374,
+      "force_loss": 0.0188570749014616,
+      "step": 28400
+    },
+    {
+      "epoch": 25.548561151079138,
+      "grad_norm": 0.23806895315647125,
+      "learning_rate": 5.8476980960162784e-05,
+      "loss": 0.0198,
+      "step": 28410
+    },
+    {
+      "action_loss": 0.004191949497908354,
+      "epoch": 25.548561151079138,
+      "step": 28410
+    },
+    {
+      "epoch": 25.548561151079138,
+      "step": 28410,
+      "torque_loss": 0.1481306254863739
+    },
+    {
+      "epoch": 25.548561151079138,
+      "force_loss": 0.003596404567360878,
+      "step": 28410
+    },
+    {
+      "epoch": 25.557553956834532,
+      "grad_norm": 0.21229428052902222,
+      "learning_rate": 5.844982078814868e-05,
+      "loss": 0.0206,
+      "step": 28420
+    },
+    {
+      "action_loss": 0.004433061927556992,
+      "epoch": 25.557553956834532,
+      "step": 28420
+    },
+    {
+      "epoch": 25.557553956834532,
+      "step": 28420,
+      "torque_loss": 0.1579483300447464
+    },
+    {
+      "epoch": 25.557553956834532,
+      "force_loss": 0.004343610722571611,
+      "step": 28420
+    },
+    {
+      "epoch": 25.56654676258993,
+      "grad_norm": 0.4190731346607208,
+      "learning_rate": 5.842265804930211e-05,
+      "loss": 0.0238,
+      "step": 28430
+    },
+    {
+      "action_loss": 0.0030550810042768717,
+      "epoch": 25.56654676258993,
+      "step": 28430
+    },
+    {
+      "epoch": 25.56654676258993,
+      "step": 28430,
+      "torque_loss": 0.12878479063510895
+    },
+    {
+      "epoch": 25.56654676258993,
+      "force_loss": 0.0020467147696763277,
+      "step": 28430
+    },
+    {
+      "epoch": 25.575539568345324,
+      "grad_norm": 0.2838549315929413,
+      "learning_rate": 5.839549275187444e-05,
+      "loss": 0.018,
+      "step": 28440
+    },
+    {
+      "action_loss": 0.012333524413406849,
+      "epoch": 25.575539568345324,
+      "step": 28440
+    },
+    {
+      "epoch": 25.575539568345324,
+      "step": 28440,
+      "torque_loss": 0.15713512897491455
+    },
+    {
+      "epoch": 25.575539568345324,
+      "force_loss": 0.01004349160939455,
+      "step": 28440
+    },
+    {
+      "epoch": 25.584532374100718,
+      "grad_norm": 0.5280764102935791,
+      "learning_rate": 5.836832490411771e-05,
+      "loss": 0.0199,
+      "step": 28450
+    },
+    {
+      "action_loss": 0.002958972007036209,
+      "epoch": 25.584532374100718,
+      "step": 28450
+    },
+    {
+      "epoch": 25.584532374100718,
+      "step": 28450,
+      "torque_loss": 0.14007197320461273
+    },
+    {
+      "epoch": 25.584532374100718,
+      "force_loss": 0.002835891442373395,
+      "step": 28450
+    },
+    {
+      "epoch": 25.593525179856115,
+      "grad_norm": 0.2873149812221527,
+      "learning_rate": 5.834115451428485e-05,
+      "loss": 0.0199,
+      "step": 28460
+    },
+    {
+      "action_loss": 0.007463912013918161,
+      "epoch": 25.593525179856115,
+      "step": 28460
+    },
+    {
+      "epoch": 25.593525179856115,
+      "step": 28460,
+      "torque_loss": 0.09955960512161255
+    },
+    {
+      "epoch": 25.593525179856115,
+      "force_loss": 0.006850830744951963,
+      "step": 28460
+    },
+    {
+      "epoch": 25.60251798561151,
+      "grad_norm": 0.16323289275169373,
+      "learning_rate": 5.831398159062946e-05,
+      "loss": 0.0179,
+      "step": 28470
+    },
+    {
+      "action_loss": 0.002226385287940502,
+      "epoch": 25.60251798561151,
+      "step": 28470
+    },
+    {
+      "epoch": 25.60251798561151,
+      "step": 28470,
+      "torque_loss": 0.17174698412418365
+    },
+    {
+      "epoch": 25.60251798561151,
+      "force_loss": 0.0024590285029262304,
+      "step": 28470
+    },
+    {
+      "epoch": 25.611510791366907,
+      "grad_norm": 0.27848368883132935,
+      "learning_rate": 5.828680614140599e-05,
+      "loss": 0.0184,
+      "step": 28480
+    },
+    {
+      "action_loss": 0.01565668359398842,
+      "epoch": 25.611510791366907,
+      "step": 28480
+    },
+    {
+      "epoch": 25.611510791366907,
+      "step": 28480,
+      "torque_loss": 0.1555856466293335
+    },
+    {
+      "epoch": 25.611510791366907,
+      "force_loss": 0.023662244901061058,
+      "step": 28480
+    },
+    {
+      "epoch": 25.6205035971223,
+      "grad_norm": 0.5817075371742249,
+      "learning_rate": 5.825962817486962e-05,
+      "loss": 0.0215,
+      "step": 28490
+    },
+    {
+      "action_loss": 0.004613556433469057,
+      "epoch": 25.6205035971223,
+      "step": 28490
+    },
+    {
+      "epoch": 25.6205035971223,
+      "step": 28490,
+      "torque_loss": 0.10848802328109741
+    },
+    {
+      "epoch": 25.6205035971223,
+      "force_loss": 0.0048154438845813274,
+      "step": 28490
+    },
+    {
+      "epoch": 25.6294964028777,
+      "grad_norm": 0.15831421315670013,
+      "learning_rate": 5.823244769927629e-05,
+      "loss": 0.0195,
+      "step": 28500
+    },
+    {
+      "action_loss": 0.005210065748542547,
+      "epoch": 25.6294964028777,
+      "step": 28500
+    },
+    {
+      "epoch": 25.6294964028777,
+      "step": 28500,
+      "torque_loss": 0.12404090911149979
+    },
+    {
+      "epoch": 25.6294964028777,
+      "force_loss": 0.011453814804553986,
+      "step": 28500
+    },
+    {
+      "epoch": 25.638489208633093,
+      "grad_norm": 0.2887755036354065,
+      "learning_rate": 5.8205264722882716e-05,
+      "loss": 0.0266,
+      "step": 28510
+    },
+    {
+      "action_loss": 0.002105491003021598,
+      "epoch": 25.638489208633093,
+      "step": 28510
+    },
+    {
+      "epoch": 25.638489208633093,
+      "step": 28510,
+      "torque_loss": 0.0879632830619812
+    },
+    {
+      "epoch": 25.638489208633093,
+      "force_loss": 0.002024345565587282,
+      "step": 28510
+    },
+    {
+      "epoch": 25.64748201438849,
+      "grad_norm": 0.46345382928848267,
+      "learning_rate": 5.817807925394636e-05,
+      "loss": 0.0182,
+      "step": 28520
+    },
+    {
+      "action_loss": 0.0032017540652304888,
+      "epoch": 25.64748201438849,
+      "step": 28520
+    },
+    {
+      "epoch": 25.64748201438849,
+      "step": 28520,
+      "torque_loss": 0.16883087158203125
+    },
+    {
+      "epoch": 25.64748201438849,
+      "force_loss": 0.0020397247280925512,
+      "step": 28520
+    },
+    {
+      "epoch": 25.656474820143885,
+      "grad_norm": 0.2489829808473587,
+      "learning_rate": 5.815089130072546e-05,
+      "loss": 0.0188,
+      "step": 28530
+    },
+    {
+      "action_loss": 0.004855359438806772,
+      "epoch": 25.656474820143885,
+      "step": 28530
+    },
+    {
+      "epoch": 25.656474820143885,
+      "step": 28530,
+      "torque_loss": 0.08536690473556519
+    },
+    {
+      "epoch": 25.656474820143885,
+      "force_loss": 0.008929177187383175,
+      "step": 28530
+    },
+    {
+      "epoch": 25.665467625899282,
+      "grad_norm": 0.18484389781951904,
+      "learning_rate": 5.8123700871479e-05,
+      "loss": 0.0161,
+      "step": 28540
+    },
+    {
+      "action_loss": 0.0024327461142092943,
+      "epoch": 25.665467625899282,
+      "step": 28540
+    },
+    {
+      "epoch": 25.665467625899282,
+      "step": 28540,
+      "torque_loss": 0.1123182475566864
+    },
+    {
+      "epoch": 25.665467625899282,
+      "force_loss": 0.0032714849803596735,
+      "step": 28540
+    },
+    {
+      "epoch": 25.674460431654676,
+      "grad_norm": 0.10311053693294525,
+      "learning_rate": 5.809650797446671e-05,
+      "loss": 0.018,
+      "step": 28550
+    },
+    {
+      "action_loss": 0.007736308965831995,
+      "epoch": 25.674460431654676,
+      "step": 28550
+    },
+    {
+      "epoch": 25.674460431654676,
+      "step": 28550,
+      "torque_loss": 0.1223251223564148
+    },
+    {
+      "epoch": 25.674460431654676,
+      "force_loss": 0.0098481560125947,
+      "step": 28550
+    },
+    {
+      "epoch": 25.68345323741007,
+      "grad_norm": 0.195352703332901,
+      "learning_rate": 5.806931261794907e-05,
+      "loss": 0.0194,
+      "step": 28560
+    },
+    {
+      "action_loss": 0.001999068772420287,
+      "epoch": 25.68345323741007,
+      "step": 28560
+    },
+    {
+      "epoch": 25.68345323741007,
+      "step": 28560,
+      "torque_loss": 0.12419679760932922
+    },
+    {
+      "epoch": 25.68345323741007,
+      "force_loss": 0.003961937036365271,
+      "step": 28560
+    },
+    {
+      "epoch": 25.692446043165468,
+      "grad_norm": 0.13347913324832916,
+      "learning_rate": 5.804211481018731e-05,
+      "loss": 0.0153,
+      "step": 28570
+    },
+    {
+      "action_loss": 0.01377819012850523,
+      "epoch": 25.692446043165468,
+      "step": 28570
+    },
+    {
+      "epoch": 25.692446043165468,
+      "step": 28570,
+      "torque_loss": 0.16717176139354706
+    },
+    {
+      "epoch": 25.692446043165468,
+      "force_loss": 0.009943501092493534,
+      "step": 28570
+    },
+    {
+      "epoch": 25.701438848920862,
+      "grad_norm": 0.1758003681898117,
+      "learning_rate": 5.801491455944341e-05,
+      "loss": 0.0181,
+      "step": 28580
+    },
+    {
+      "action_loss": 0.003849254921078682,
+      "epoch": 25.701438848920862,
+      "step": 28580
+    },
+    {
+      "epoch": 25.701438848920862,
+      "step": 28580,
+      "torque_loss": 0.11864777654409409
+    },
+    {
+      "epoch": 25.701438848920862,
+      "force_loss": 0.002902369014918804,
+      "step": 28580
+    },
+    {
+      "epoch": 25.71043165467626,
+      "grad_norm": 0.2609722912311554,
+      "learning_rate": 5.79877118739801e-05,
+      "loss": 0.0185,
+      "step": 28590
+    },
+    {
+      "action_loss": 0.007190841715782881,
+      "epoch": 25.71043165467626,
+      "step": 28590
+    },
+    {
+      "epoch": 25.71043165467626,
+      "step": 28590,
+      "torque_loss": 0.14964285492897034
+    },
+    {
+      "epoch": 25.71043165467626,
+      "force_loss": 0.009538150392472744,
+      "step": 28590
+    },
+    {
+      "epoch": 25.719424460431654,
+      "grad_norm": 0.3223738968372345,
+      "learning_rate": 5.7960506762060816e-05,
+      "loss": 0.0196,
+      "step": 28600
+    },
+    {
+      "action_loss": 0.005924099590629339,
+      "epoch": 25.719424460431654,
+      "step": 28600
+    },
+    {
+      "epoch": 25.719424460431654,
+      "step": 28600,
+      "torque_loss": 0.10150185972452164
+    },
+    {
+      "epoch": 25.719424460431654,
+      "force_loss": 0.009191926568746567,
+      "step": 28600
+    },
+    {
+      "epoch": 25.72841726618705,
+      "grad_norm": 0.4019058644771576,
+      "learning_rate": 5.793329923194977e-05,
+      "loss": 0.0167,
+      "step": 28610
+    },
+    {
+      "action_loss": 0.003359417198225856,
+      "epoch": 25.72841726618705,
+      "step": 28610
+    },
+    {
+      "epoch": 25.72841726618705,
+      "step": 28610,
+      "torque_loss": 0.12953747808933258
+    },
+    {
+      "epoch": 25.72841726618705,
+      "force_loss": 0.004174624104052782,
+      "step": 28610
+    },
+    {
+      "epoch": 25.737410071942445,
+      "grad_norm": 0.4349093735218048,
+      "learning_rate": 5.790608929191187e-05,
+      "loss": 0.0191,
+      "step": 28620
+    },
+    {
+      "action_loss": 0.003896950511261821,
+      "epoch": 25.737410071942445,
+      "step": 28620
+    },
+    {
+      "epoch": 25.737410071942445,
+      "step": 28620,
+      "torque_loss": 0.13709606230258942
+    },
+    {
+      "epoch": 25.737410071942445,
+      "force_loss": 0.0024140216410160065,
+      "step": 28620
+    },
+    {
+      "epoch": 25.746402877697843,
+      "grad_norm": 0.22858905792236328,
+      "learning_rate": 5.78788769502128e-05,
+      "loss": 0.0231,
+      "step": 28630
+    },
+    {
+      "action_loss": 0.008807025849819183,
+      "epoch": 25.746402877697843,
+      "step": 28630
+    },
+    {
+      "epoch": 25.746402877697843,
+      "step": 28630,
+      "torque_loss": 0.15493394434452057
+    },
+    {
+      "epoch": 25.746402877697843,
+      "force_loss": 0.011327589862048626,
+      "step": 28630
+    },
+    {
+      "epoch": 25.755395683453237,
+      "grad_norm": 0.32832226157188416,
+      "learning_rate": 5.785166221511894e-05,
+      "loss": 0.0208,
+      "step": 28640
+    },
+    {
+      "action_loss": 0.003166560083627701,
+      "epoch": 25.755395683453237,
+      "step": 28640
+    },
+    {
+      "epoch": 25.755395683453237,
+      "step": 28640,
+      "torque_loss": 0.110824815928936
+    },
+    {
+      "epoch": 25.755395683453237,
+      "force_loss": 0.0033940270077437162,
+      "step": 28640
+    },
+    {
+      "epoch": 25.764388489208635,
+      "grad_norm": 0.47827088832855225,
+      "learning_rate": 5.7824445094897415e-05,
+      "loss": 0.0229,
+      "step": 28650
+    },
+    {
+      "action_loss": 0.005630446132272482,
+      "epoch": 25.764388489208635,
+      "step": 28650
+    },
+    {
+      "epoch": 25.764388489208635,
+      "step": 28650,
+      "torque_loss": 0.13710464537143707
+    },
+    {
+      "epoch": 25.764388489208635,
+      "force_loss": 0.0062186564318835735,
+      "step": 28650
+    },
+    {
+      "epoch": 25.77338129496403,
+      "grad_norm": 0.4051119089126587,
+      "learning_rate": 5.7797225597816065e-05,
+      "loss": 0.0217,
+      "step": 28660
+    },
+    {
+      "action_loss": 0.005888625513762236,
+      "epoch": 25.77338129496403,
+      "step": 28660
+    },
+    {
+      "epoch": 25.77338129496403,
+      "step": 28660,
+      "torque_loss": 0.10672199726104736
+    },
+    {
+      "epoch": 25.77338129496403,
+      "force_loss": 0.005437417421489954,
+      "step": 28660
+    },
+    {
+      "epoch": 25.782374100719423,
+      "grad_norm": 0.27474355697631836,
+      "learning_rate": 5.777000373214345e-05,
+      "loss": 0.0185,
+      "step": 28670
+    },
+    {
+      "action_loss": 0.003132052021101117,
+      "epoch": 25.782374100719423,
+      "step": 28670
+    },
+    {
+      "epoch": 25.782374100719423,
+      "step": 28670,
+      "torque_loss": 0.10297843813896179
+    },
+    {
+      "epoch": 25.782374100719423,
+      "force_loss": 0.0027979493606835604,
+      "step": 28670
+    },
+    {
+      "epoch": 25.79136690647482,
+      "grad_norm": 0.30740973353385925,
+      "learning_rate": 5.774277950614885e-05,
+      "loss": 0.024,
+      "step": 28680
+    },
+    {
+      "action_loss": 0.008867667056620121,
+      "epoch": 25.79136690647482,
+      "step": 28680
+    },
+    {
+      "epoch": 25.79136690647482,
+      "step": 28680,
+      "torque_loss": 0.12430638074874878
+    },
+    {
+      "epoch": 25.79136690647482,
+      "force_loss": 0.007247579749673605,
+      "step": 28680
+    },
+    {
+      "epoch": 25.800359712230215,
+      "grad_norm": 0.20495644211769104,
+      "learning_rate": 5.771555292810227e-05,
+      "loss": 0.0191,
+      "step": 28690
+    },
+    {
+      "action_loss": 0.0036730438005179167,
+      "epoch": 25.800359712230215,
+      "step": 28690
+    },
+    {
+      "epoch": 25.800359712230215,
+      "step": 28690,
+      "torque_loss": 0.12670987844467163
+    },
+    {
+      "epoch": 25.800359712230215,
+      "force_loss": 0.003855082904919982,
+      "step": 28690
+    },
+    {
+      "epoch": 25.809352517985612,
+      "grad_norm": 0.4939091205596924,
+      "learning_rate": 5.768832400627444e-05,
+      "loss": 0.0182,
+      "step": 28700
+    },
+    {
+      "action_loss": 0.005129032302647829,
+      "epoch": 25.809352517985612,
+      "step": 28700
+    },
+    {
+      "epoch": 25.809352517985612,
+      "step": 28700,
+      "torque_loss": 0.1300128698348999
+    },
+    {
+      "epoch": 25.809352517985612,
+      "force_loss": 0.012985593639314175,
+      "step": 28700
+    },
+    {
+      "epoch": 25.818345323741006,
+      "grad_norm": 0.5835957527160645,
+      "learning_rate": 5.7661092748936775e-05,
+      "loss": 0.0219,
+      "step": 28710
+    },
+    {
+      "action_loss": 0.0049516293220222,
+      "epoch": 25.818345323741006,
+      "step": 28710
+    },
+    {
+      "epoch": 25.818345323741006,
+      "step": 28710,
+      "torque_loss": 0.09968063980340958
+    },
+    {
+      "epoch": 25.818345323741006,
+      "force_loss": 0.0068278550170362,
+      "step": 28710
+    },
+    {
+      "epoch": 25.827338129496404,
+      "grad_norm": 0.21389810740947723,
+      "learning_rate": 5.76338591643614e-05,
+      "loss": 0.025,
+      "step": 28720
+    },
+    {
+      "action_loss": 0.005397242959588766,
+      "epoch": 25.827338129496404,
+      "step": 28720
+    },
+    {
+      "epoch": 25.827338129496404,
+      "step": 28720,
+      "torque_loss": 0.10741696506738663
+    },
+    {
+      "epoch": 25.827338129496404,
+      "force_loss": 0.004168794024735689,
+      "step": 28720
+    },
+    {
+      "epoch": 25.836330935251798,
+      "grad_norm": 0.16188180446624756,
+      "learning_rate": 5.760662326082118e-05,
+      "loss": 0.0238,
+      "step": 28730
+    },
+    {
+      "action_loss": 0.003408056451007724,
+      "epoch": 25.836330935251798,
+      "step": 28730
+    },
+    {
+      "epoch": 25.836330935251798,
+      "step": 28730,
+      "torque_loss": 0.12680047750473022
+    },
+    {
+      "epoch": 25.836330935251798,
+      "force_loss": 0.006992408540099859,
+      "step": 28730
+    },
+    {
+      "epoch": 25.845323741007196,
+      "grad_norm": 0.36538028717041016,
+      "learning_rate": 5.757938504658965e-05,
+      "loss": 0.0204,
+      "step": 28740
+    },
+    {
+      "action_loss": 0.009054531343281269,
+      "epoch": 25.845323741007196,
+      "step": 28740
+    },
+    {
+      "epoch": 25.845323741007196,
+      "step": 28740,
+      "torque_loss": 0.11967268586158752
+    },
+    {
+      "epoch": 25.845323741007196,
+      "force_loss": 0.004422473255544901,
+      "step": 28740
+    },
+    {
+      "epoch": 25.85431654676259,
+      "grad_norm": 0.20116569101810455,
+      "learning_rate": 5.755214452994107e-05,
+      "loss": 0.0199,
+      "step": 28750
+    },
+    {
+      "action_loss": 0.005058668088167906,
+      "epoch": 25.85431654676259,
+      "step": 28750
+    },
+    {
+      "epoch": 25.85431654676259,
+      "step": 28750,
+      "torque_loss": 0.14605657756328583
+    },
+    {
+      "epoch": 25.85431654676259,
+      "force_loss": 0.0036439981777220964,
+      "step": 28750
+    },
+    {
+      "epoch": 25.863309352517987,
+      "grad_norm": 0.17597800493240356,
+      "learning_rate": 5.752490171915039e-05,
+      "loss": 0.0222,
+      "step": 28760
+    },
+    {
+      "action_loss": 0.002458389615640044,
+      "epoch": 25.863309352517987,
+      "step": 28760
+    },
+    {
+      "epoch": 25.863309352517987,
+      "step": 28760,
+      "torque_loss": 0.1379869431257248
+    },
+    {
+      "epoch": 25.863309352517987,
+      "force_loss": 0.00291249412111938,
+      "step": 28760
+    },
+    {
+      "epoch": 25.87230215827338,
+      "grad_norm": 0.3704226315021515,
+      "learning_rate": 5.749765662249324e-05,
+      "loss": 0.0239,
+      "step": 28770
+    },
+    {
+      "action_loss": 0.005187605042010546,
+      "epoch": 25.87230215827338,
+      "step": 28770
+    },
+    {
+      "epoch": 25.87230215827338,
+      "step": 28770,
+      "torque_loss": 0.13833649456501007
+    },
+    {
+      "epoch": 25.87230215827338,
+      "force_loss": 0.007807048503309488,
+      "step": 28770
+    },
+    {
+      "epoch": 25.881294964028775,
+      "grad_norm": 0.18758834898471832,
+      "learning_rate": 5.747040924824596e-05,
+      "loss": 0.0161,
+      "step": 28780
+    },
+    {
+      "action_loss": 0.0019194980850443244,
+      "epoch": 25.881294964028775,
+      "step": 28780
+    },
+    {
+      "epoch": 25.881294964028775,
+      "step": 28780,
+      "torque_loss": 0.09164443612098694
+    },
+    {
+      "epoch": 25.881294964028775,
+      "force_loss": 0.0008873723563738167,
+      "step": 28780
+    },
+    {
+      "epoch": 25.890287769784173,
+      "grad_norm": 0.29400554299354553,
+      "learning_rate": 5.7443159604685613e-05,
+      "loss": 0.0198,
+      "step": 28790
+    },
+    {
+      "action_loss": 0.010663188993930817,
+      "epoch": 25.890287769784173,
+      "step": 28790
+    },
+    {
+      "epoch": 25.890287769784173,
+      "step": 28790,
+      "torque_loss": 0.14566320180892944
+    },
+    {
+      "epoch": 25.890287769784173,
+      "force_loss": 0.013719928450882435,
+      "step": 28790
+    },
+    {
+      "epoch": 25.899280575539567,
+      "grad_norm": 0.30214011669158936,
+      "learning_rate": 5.74159077000899e-05,
+      "loss": 0.0217,
+      "step": 28800
+    },
+    {
+      "action_loss": 0.00703857047483325,
+      "epoch": 25.899280575539567,
+      "step": 28800
+    },
+    {
+      "epoch": 25.899280575539567,
+      "step": 28800,
+      "torque_loss": 0.12661753594875336
+    },
+    {
+      "epoch": 25.899280575539567,
+      "force_loss": 0.007603875827044249,
+      "step": 28800
+    },
+    {
+      "epoch": 25.908273381294965,
+      "grad_norm": 0.5199934840202332,
+      "learning_rate": 5.7388653542737235e-05,
+      "loss": 0.0203,
+      "step": 28810
+    },
+    {
+      "action_loss": 0.0063810646533966064,
+      "epoch": 25.908273381294965,
+      "step": 28810
+    },
+    {
+      "epoch": 25.908273381294965,
+      "step": 28810,
+      "torque_loss": 0.14510506391525269
+    },
+    {
+      "epoch": 25.908273381294965,
+      "force_loss": 0.010953851044178009,
+      "step": 28810
+    },
+    {
+      "epoch": 25.91726618705036,
+      "grad_norm": 0.2643183469772339,
+      "learning_rate": 5.736139714090672e-05,
+      "loss": 0.0205,
+      "step": 28820
+    },
+    {
+      "action_loss": 0.03102993406355381,
+      "epoch": 25.91726618705036,
+      "step": 28820
+    },
+    {
+      "epoch": 25.91726618705036,
+      "step": 28820,
+      "torque_loss": 0.1717318296432495
+    },
+    {
+      "epoch": 25.91726618705036,
+      "force_loss": 0.02739032171666622,
+      "step": 28820
+    },
+    {
+      "epoch": 25.926258992805757,
+      "grad_norm": 0.34062471985816956,
+      "learning_rate": 5.73341385028781e-05,
+      "loss": 0.022,
+      "step": 28830
+    },
+    {
+      "action_loss": 0.0028952148277312517,
+      "epoch": 25.926258992805757,
+      "step": 28830
+    },
+    {
+      "epoch": 25.926258992805757,
+      "step": 28830,
+      "torque_loss": 0.0797184482216835
+    },
+    {
+      "epoch": 25.926258992805757,
+      "force_loss": 0.002789615886285901,
+      "step": 28830
+    },
+    {
+      "epoch": 25.93525179856115,
+      "grad_norm": 0.28866004943847656,
+      "learning_rate": 5.7306877636931855e-05,
+      "loss": 0.0164,
+      "step": 28840
+    },
+    {
+      "action_loss": 0.003368654288351536,
+      "epoch": 25.93525179856115,
+      "step": 28840
+    },
+    {
+      "epoch": 25.93525179856115,
+      "step": 28840,
+      "torque_loss": 0.10057816654443741
+    },
+    {
+      "epoch": 25.93525179856115,
+      "force_loss": 0.0037018340080976486,
+      "step": 28840
+    },
+    {
+      "epoch": 25.944244604316548,
+      "grad_norm": 0.1584034115076065,
+      "learning_rate": 5.7279614551349125e-05,
+      "loss": 0.0173,
+      "step": 28850
+    },
+    {
+      "action_loss": 0.0019302628934383392,
+      "epoch": 25.944244604316548,
+      "step": 28850
+    },
+    {
+      "epoch": 25.944244604316548,
+      "step": 28850,
+      "torque_loss": 0.15853749215602875
+    },
+    {
+      "epoch": 25.944244604316548,
+      "force_loss": 0.0033526085317134857,
+      "step": 28850
+    },
+    {
+      "epoch": 25.953237410071942,
+      "grad_norm": 0.1823311746120453,
+      "learning_rate": 5.725234925441169e-05,
+      "loss": 0.0164,
+      "step": 28860
+    },
+    {
+      "action_loss": 0.0013245727168396115,
+      "epoch": 25.953237410071942,
+      "step": 28860
+    },
+    {
+      "epoch": 25.953237410071942,
+      "step": 28860,
+      "torque_loss": 0.09203815460205078
+    },
+    {
+      "epoch": 25.953237410071942,
+      "force_loss": 0.0012865253956988454,
+      "step": 28860
+    },
+    {
+      "epoch": 25.96223021582734,
+      "grad_norm": 0.6473246216773987,
+      "learning_rate": 5.7225081754402044e-05,
+      "loss": 0.0171,
+      "step": 28870
+    },
+    {
+      "action_loss": 0.00738756125792861,
+      "epoch": 25.96223021582734,
+      "step": 28870
+    },
+    {
+      "epoch": 25.96223021582734,
+      "step": 28870,
+      "torque_loss": 0.12493404000997543
+    },
+    {
+      "epoch": 25.96223021582734,
+      "force_loss": 0.005382902454584837,
+      "step": 28870
+    },
+    {
+      "epoch": 25.971223021582734,
+      "grad_norm": 0.21924792230129242,
+      "learning_rate": 5.7197812059603326e-05,
+      "loss": 0.0185,
+      "step": 28880
+    },
+    {
+      "action_loss": 0.008321907371282578,
+      "epoch": 25.971223021582734,
+      "step": 28880
+    },
+    {
+      "epoch": 25.971223021582734,
+      "step": 28880,
+      "torque_loss": 0.13138620555400848
+    },
+    {
+      "epoch": 25.971223021582734,
+      "force_loss": 0.005462937522679567,
+      "step": 28880
+    },
+    {
+      "epoch": 25.980215827338128,
+      "grad_norm": 0.3269522786140442,
+      "learning_rate": 5.717054017829934e-05,
+      "loss": 0.0213,
+      "step": 28890
+    },
+    {
+      "action_loss": 0.002165122888982296,
+      "epoch": 25.980215827338128,
+      "step": 28890
+    },
+    {
+      "epoch": 25.980215827338128,
+      "step": 28890,
+      "torque_loss": 0.10722139477729797
+    },
+    {
+      "epoch": 25.980215827338128,
+      "force_loss": 0.0011654613772407174,
+      "step": 28890
+    },
+    {
+      "epoch": 25.989208633093526,
+      "grad_norm": 0.3549667298793793,
+      "learning_rate": 5.7143266118774584e-05,
+      "loss": 0.0287,
+      "step": 28900
+    },
+    {
+      "action_loss": 0.0026221785228699446,
+      "epoch": 25.989208633093526,
+      "step": 28900
+    },
+    {
+      "epoch": 25.989208633093526,
+      "step": 28900,
+      "torque_loss": 0.12324017286300659
+    },
+    {
+      "epoch": 25.989208633093526,
+      "force_loss": 0.0051282597705721855,
+      "step": 28900
+    },
+    {
+      "epoch": 25.99820143884892,
+      "grad_norm": 0.28808143734931946,
+      "learning_rate": 5.711598988931418e-05,
+      "loss": 0.0188,
+      "step": 28910
+    },
+    {
+      "action_loss": 0.003999910783022642,
+      "epoch": 25.99820143884892,
+      "step": 28910
+    },
+    {
+      "epoch": 25.99820143884892,
+      "step": 28910,
+      "torque_loss": 0.1733754426240921
+    },
+    {
+      "epoch": 25.99820143884892,
+      "force_loss": 0.002792246639728546,
+      "step": 28910
+    },
+    {
+      "epoch": 26.007194244604317,
+      "grad_norm": 0.32484543323516846,
+      "learning_rate": 5.7088711498203954e-05,
+      "loss": 0.0195,
+      "step": 28920
+    },
+    {
+      "action_loss": 0.00449376692995429,
+      "epoch": 26.007194244604317,
+      "step": 28920
+    },
+    {
+      "epoch": 26.007194244604317,
+      "step": 28920,
+      "torque_loss": 0.1329164057970047
+    },
+    {
+      "epoch": 26.007194244604317,
+      "force_loss": 0.005821609403938055,
+      "step": 28920
+    },
+    {
+      "epoch": 26.01618705035971,
+      "grad_norm": 0.14284273982048035,
+      "learning_rate": 5.706143095373033e-05,
+      "loss": 0.021,
+      "step": 28930
+    },
+    {
+      "action_loss": 0.0031853385735303164,
+      "epoch": 26.01618705035971,
+      "step": 28930
+    },
+    {
+      "epoch": 26.01618705035971,
+      "step": 28930,
+      "torque_loss": 0.1327640861272812
+    },
+    {
+      "epoch": 26.01618705035971,
+      "force_loss": 0.005952553357928991,
+      "step": 28930
+    },
+    {
+      "epoch": 26.02517985611511,
+      "grad_norm": 0.5342223048210144,
+      "learning_rate": 5.703414826418042e-05,
+      "loss": 0.0194,
+      "step": 28940
+    },
+    {
+      "action_loss": 0.009315653704106808,
+      "epoch": 26.02517985611511,
+      "step": 28940
+    },
+    {
+      "epoch": 26.02517985611511,
+      "step": 28940,
+      "torque_loss": 0.1404716521501541
+    },
+    {
+      "epoch": 26.02517985611511,
+      "force_loss": 0.006468483712524176,
+      "step": 28940
+    },
+    {
+      "epoch": 26.034172661870503,
+      "grad_norm": 0.2200510948896408,
+      "learning_rate": 5.7006863437842007e-05,
+      "loss": 0.0199,
+      "step": 28950
+    },
+    {
+      "action_loss": 0.005508519243448973,
+      "epoch": 26.034172661870503,
+      "step": 28950
+    },
+    {
+      "epoch": 26.034172661870503,
+      "step": 28950,
+      "torque_loss": 0.14688707888126373
+    },
+    {
+      "epoch": 26.034172661870503,
+      "force_loss": 0.003007271094247699,
+      "step": 28950
+    },
+    {
+      "epoch": 26.0431654676259,
+      "grad_norm": 0.3171609044075012,
+      "learning_rate": 5.697957648300348e-05,
+      "loss": 0.0274,
+      "step": 28960
+    },
+    {
+      "action_loss": 0.0037698314990848303,
+      "epoch": 26.0431654676259,
+      "step": 28960
+    },
+    {
+      "epoch": 26.0431654676259,
+      "step": 28960,
+      "torque_loss": 0.12635333836078644
+    },
+    {
+      "epoch": 26.0431654676259,
+      "force_loss": 0.0025174252223223448,
+      "step": 28960
+    },
+    {
+      "epoch": 26.052158273381295,
+      "grad_norm": 0.20609548687934875,
+      "learning_rate": 5.695228740795391e-05,
+      "loss": 0.0216,
+      "step": 28970
+    },
+    {
+      "action_loss": 0.0028065908700227737,
+      "epoch": 26.052158273381295,
+      "step": 28970
+    },
+    {
+      "epoch": 26.052158273381295,
+      "step": 28970,
+      "torque_loss": 0.13230383396148682
+    },
+    {
+      "epoch": 26.052158273381295,
+      "force_loss": 0.003663417650386691,
+      "step": 28970
+    },
+    {
+      "epoch": 26.06115107913669,
+      "grad_norm": 0.42974433302879333,
+      "learning_rate": 5.6924996220982985e-05,
+      "loss": 0.0196,
+      "step": 28980
+    },
+    {
+      "action_loss": 0.002372222952544689,
+      "epoch": 26.06115107913669,
+      "step": 28980
+    },
+    {
+      "epoch": 26.06115107913669,
+      "step": 28980,
+      "torque_loss": 0.1182672381401062
+    },
+    {
+      "epoch": 26.06115107913669,
+      "force_loss": 0.0022993513848632574,
+      "step": 28980
+    },
+    {
+      "epoch": 26.070143884892087,
+      "grad_norm": 0.16391532123088837,
+      "learning_rate": 5.6897702930381045e-05,
+      "loss": 0.0167,
+      "step": 28990
+    },
+    {
+      "action_loss": 0.0022096759639680386,
+      "epoch": 26.070143884892087,
+      "step": 28990
+    },
+    {
+      "epoch": 26.070143884892087,
+      "step": 28990,
+      "torque_loss": 0.08953986316919327
+    },
+    {
+      "epoch": 26.070143884892087,
+      "force_loss": 0.003215731354430318,
+      "step": 28990
+    },
+    {
+      "epoch": 26.07913669064748,
+      "grad_norm": 0.12691636383533478,
+      "learning_rate": 5.687040754443908e-05,
+      "loss": 0.0164,
+      "step": 29000
+    },
+    {
+      "action_loss": 0.002988505410030484,
+      "epoch": 26.07913669064748,
+      "step": 29000
+    },
+    {
+      "epoch": 26.07913669064748,
+      "step": 29000,
+      "torque_loss": 0.12361162900924683
+    },
+    {
+      "epoch": 26.07913669064748,
+      "force_loss": 0.004084745887666941,
+      "step": 29000
+    },
+    {
+      "epoch": 26.08812949640288,
+      "grad_norm": 0.20146793127059937,
+      "learning_rate": 5.6843110071448725e-05,
+      "loss": 0.0192,
+      "step": 29010
+    },
+    {
+      "action_loss": 0.003533797338604927,
+      "epoch": 26.08812949640288,
+      "step": 29010
+    },
+    {
+      "epoch": 26.08812949640288,
+      "step": 29010,
+      "torque_loss": 0.11133870482444763
+    },
+    {
+      "epoch": 26.08812949640288,
+      "force_loss": 0.004257683642208576,
+      "step": 29010
+    },
+    {
+      "epoch": 26.097122302158272,
+      "grad_norm": 0.308725506067276,
+      "learning_rate": 5.6815810519702194e-05,
+      "loss": 0.0195,
+      "step": 29020
+    },
+    {
+      "action_loss": 0.0036037983372807503,
+      "epoch": 26.097122302158272,
+      "step": 29020
+    },
+    {
+      "epoch": 26.097122302158272,
+      "step": 29020,
+      "torque_loss": 0.08814121037721634
+    },
+    {
+      "epoch": 26.097122302158272,
+      "force_loss": 0.0018510861555114388,
+      "step": 29020
+    },
+    {
+      "epoch": 26.10611510791367,
+      "grad_norm": 0.3805563449859619,
+      "learning_rate": 5.6788508897492396e-05,
+      "loss": 0.0168,
+      "step": 29030
+    },
+    {
+      "action_loss": 0.002368503948673606,
+      "epoch": 26.10611510791367,
+      "step": 29030
+    },
+    {
+      "epoch": 26.10611510791367,
+      "step": 29030,
+      "torque_loss": 0.11723294109106064
+    },
+    {
+      "epoch": 26.10611510791367,
+      "force_loss": 0.0014365953393280506,
+      "step": 29030
+    },
+    {
+      "epoch": 26.115107913669064,
+      "grad_norm": 0.3679788112640381,
+      "learning_rate": 5.676120521311282e-05,
+      "loss": 0.0184,
+      "step": 29040
+    },
+    {
+      "action_loss": 0.007685080636292696,
+      "epoch": 26.115107913669064,
+      "step": 29040
+    },
+    {
+      "epoch": 26.115107913669064,
+      "step": 29040,
+      "torque_loss": 0.1312701404094696
+    },
+    {
+      "epoch": 26.115107913669064,
+      "force_loss": 0.013920996338129044,
+      "step": 29040
+    },
+    {
+      "epoch": 26.12410071942446,
+      "grad_norm": 0.13891716301441193,
+      "learning_rate": 5.6733899474857634e-05,
+      "loss": 0.0196,
+      "step": 29050
+    },
+    {
+      "action_loss": 0.0030181354377418756,
+      "epoch": 26.12410071942446,
+      "step": 29050
+    },
+    {
+      "epoch": 26.12410071942446,
+      "step": 29050,
+      "torque_loss": 0.12050659209489822
+    },
+    {
+      "epoch": 26.12410071942446,
+      "force_loss": 0.003781926818192005,
+      "step": 29050
+    },
+    {
+      "epoch": 26.133093525179856,
+      "grad_norm": 0.18489331007003784,
+      "learning_rate": 5.670659169102157e-05,
+      "loss": 0.0169,
+      "step": 29060
+    },
+    {
+      "action_loss": 0.02172222174704075,
+      "epoch": 26.133093525179856,
+      "step": 29060
+    },
+    {
+      "epoch": 26.133093525179856,
+      "step": 29060,
+      "torque_loss": 0.20113885402679443
+    },
+    {
+      "epoch": 26.133093525179856,
+      "force_loss": 0.019537802785634995,
+      "step": 29060
+    },
+    {
+      "epoch": 26.142086330935253,
+      "grad_norm": 0.43244943022727966,
+      "learning_rate": 5.6679281869900044e-05,
+      "loss": 0.0251,
+      "step": 29070
+    },
+    {
+      "action_loss": 0.0025063930079340935,
+      "epoch": 26.142086330935253,
+      "step": 29070
+    },
+    {
+      "epoch": 26.142086330935253,
+      "step": 29070,
+      "torque_loss": 0.11865457147359848
+    },
+    {
+      "epoch": 26.142086330935253,
+      "force_loss": 0.0032049445435404778,
+      "step": 29070
+    },
+    {
+      "epoch": 26.151079136690647,
+      "grad_norm": 0.6014410853385925,
+      "learning_rate": 5.6651970019789045e-05,
+      "loss": 0.0191,
+      "step": 29080
+    },
+    {
+      "action_loss": 0.006351388990879059,
+      "epoch": 26.151079136690647,
+      "step": 29080
+    },
+    {
+      "epoch": 26.151079136690647,
+      "step": 29080,
+      "torque_loss": 0.1085766926407814
+    },
+    {
+      "epoch": 26.151079136690647,
+      "force_loss": 0.011536809615790844,
+      "step": 29080
+    },
+    {
+      "epoch": 26.16007194244604,
+      "grad_norm": 0.4379007816314697,
+      "learning_rate": 5.662465614898519e-05,
+      "loss": 0.0168,
+      "step": 29090
+    },
+    {
+      "action_loss": 0.002459225943312049,
+      "epoch": 26.16007194244604,
+      "step": 29090
+    },
+    {
+      "epoch": 26.16007194244604,
+      "step": 29090,
+      "torque_loss": 0.11237046867609024
+    },
+    {
+      "epoch": 26.16007194244604,
+      "force_loss": 0.003944239113479853,
+      "step": 29090
+    },
+    {
+      "epoch": 26.16906474820144,
+      "grad_norm": 0.2697606384754181,
+      "learning_rate": 5.6597340265785695e-05,
+      "loss": 0.024,
+      "step": 29100
+    },
+    {
+      "action_loss": 0.0027110965456813574,
+      "epoch": 26.16906474820144,
+      "step": 29100
+    },
+    {
+      "epoch": 26.16906474820144,
+      "step": 29100,
+      "torque_loss": 0.1028432548046112
+    },
+    {
+      "epoch": 26.16906474820144,
+      "force_loss": 0.002248074160888791,
+      "step": 29100
+    },
+    {
+      "epoch": 26.178057553956833,
+      "grad_norm": 0.13784776628017426,
+      "learning_rate": 5.657002237848843e-05,
+      "loss": 0.0193,
+      "step": 29110
+    },
+    {
+      "action_loss": 0.0029081034008413553,
+      "epoch": 26.178057553956833,
+      "step": 29110
+    },
+    {
+      "epoch": 26.178057553956833,
+      "step": 29110,
+      "torque_loss": 0.12708336114883423
+    },
+    {
+      "epoch": 26.178057553956833,
+      "force_loss": 0.002756238216534257,
+      "step": 29110
+    },
+    {
+      "epoch": 26.18705035971223,
+      "grad_norm": 0.27970412373542786,
+      "learning_rate": 5.654270249539183e-05,
+      "loss": 0.0177,
+      "step": 29120
+    },
+    {
+      "action_loss": 0.012793789617717266,
+      "epoch": 26.18705035971223,
+      "step": 29120
+    },
+    {
+      "epoch": 26.18705035971223,
+      "step": 29120,
+      "torque_loss": 0.1278197318315506
+    },
+    {
+      "epoch": 26.18705035971223,
+      "force_loss": 0.01568256877362728,
+      "step": 29120
+    },
+    {
+      "epoch": 26.196043165467625,
+      "grad_norm": 0.2465725541114807,
+      "learning_rate": 5.651538062479498e-05,
+      "loss": 0.0214,
+      "step": 29130
+    },
+    {
+      "action_loss": 0.01291924063116312,
+      "epoch": 26.196043165467625,
+      "step": 29130
+    },
+    {
+      "epoch": 26.196043165467625,
+      "step": 29130,
+      "torque_loss": 0.13238336145877838
+    },
+    {
+      "epoch": 26.196043165467625,
+      "force_loss": 0.01366290170699358,
+      "step": 29130
+    },
+    {
+      "epoch": 26.205035971223023,
+      "grad_norm": 0.28811484575271606,
+      "learning_rate": 5.648805677499751e-05,
+      "loss": 0.0196,
+      "step": 29140
+    },
+    {
+      "action_loss": 0.0030822723638266325,
+      "epoch": 26.205035971223023,
+      "step": 29140
+    },
+    {
+      "epoch": 26.205035971223023,
+      "step": 29140,
+      "torque_loss": 0.17272700369358063
+    },
+    {
+      "epoch": 26.205035971223023,
+      "force_loss": 0.002299186307936907,
+      "step": 29140
+    },
+    {
+      "epoch": 26.214028776978417,
+      "grad_norm": 0.18055102229118347,
+      "learning_rate": 5.646073095429969e-05,
+      "loss": 0.0221,
+      "step": 29150
+    },
+    {
+      "action_loss": 0.007661381736397743,
+      "epoch": 26.214028776978417,
+      "step": 29150
+    },
+    {
+      "epoch": 26.214028776978417,
+      "step": 29150,
+      "torque_loss": 0.13292427361011505
+    },
+    {
+      "epoch": 26.214028776978417,
+      "force_loss": 0.00930861011147499,
+      "step": 29150
+    },
+    {
+      "epoch": 26.223021582733814,
+      "grad_norm": 0.29005342721939087,
+      "learning_rate": 5.643340317100241e-05,
+      "loss": 0.0192,
+      "step": 29160
+    },
+    {
+      "action_loss": 0.0030203352216631174,
+      "epoch": 26.223021582733814,
+      "step": 29160
+    },
+    {
+      "epoch": 26.223021582733814,
+      "step": 29160,
+      "torque_loss": 0.12590394914150238
+    },
+    {
+      "epoch": 26.223021582733814,
+      "force_loss": 0.004403008613735437,
+      "step": 29160
+    },
+    {
+      "epoch": 26.23201438848921,
+      "grad_norm": 0.8712238073348999,
+      "learning_rate": 5.64060734334071e-05,
+      "loss": 0.02,
+      "step": 29170
+    },
+    {
+      "action_loss": 0.0030623984057456255,
+      "epoch": 26.23201438848921,
+      "step": 29170
+    },
+    {
+      "epoch": 26.23201438848921,
+      "step": 29170,
+      "torque_loss": 0.11342933028936386
+    },
+    {
+      "epoch": 26.23201438848921,
+      "force_loss": 0.003106334013864398,
+      "step": 29170
+    },
+    {
+      "epoch": 26.241007194244606,
+      "grad_norm": 0.22020100057125092,
+      "learning_rate": 5.637874174981583e-05,
+      "loss": 0.0179,
+      "step": 29180
+    },
+    {
+      "action_loss": 0.004982968792319298,
+      "epoch": 26.241007194244606,
+      "step": 29180
+    },
+    {
+      "epoch": 26.241007194244606,
+      "step": 29180,
+      "torque_loss": 0.1274523288011551
+    },
+    {
+      "epoch": 26.241007194244606,
+      "force_loss": 0.003581090597435832,
+      "step": 29180
+    },
+    {
+      "epoch": 26.25,
+      "grad_norm": 0.31618985533714294,
+      "learning_rate": 5.635140812853124e-05,
+      "loss": 0.0191,
+      "step": 29190
+    },
+    {
+      "action_loss": 0.0024433142971247435,
+      "epoch": 26.25,
+      "step": 29190
+    },
+    {
+      "epoch": 26.25,
+      "step": 29190,
+      "torque_loss": 0.10915540903806686
+    },
+    {
+      "epoch": 26.25,
+      "force_loss": 0.0013166745193302631,
+      "step": 29190
+    },
+    {
+      "epoch": 26.258992805755394,
+      "grad_norm": 0.26701149344444275,
+      "learning_rate": 5.6324072577856544e-05,
+      "loss": 0.0186,
+      "step": 29200
+    },
+    {
+      "action_loss": 0.01693028025329113,
+      "epoch": 26.258992805755394,
+      "step": 29200
+    },
+    {
+      "epoch": 26.258992805755394,
+      "step": 29200,
+      "torque_loss": 0.165596142411232
+    },
+    {
+      "epoch": 26.258992805755394,
+      "force_loss": 0.020471083000302315,
+      "step": 29200
+    },
+    {
+      "epoch": 26.26798561151079,
+      "grad_norm": 0.29948124289512634,
+      "learning_rate": 5.629673510609559e-05,
+      "loss": 0.0191,
+      "step": 29210
+    },
+    {
+      "action_loss": 0.002859042724594474,
+      "epoch": 26.26798561151079,
+      "step": 29210
+    },
+    {
+      "epoch": 26.26798561151079,
+      "step": 29210,
+      "torque_loss": 0.12832000851631165
+    },
+    {
+      "epoch": 26.26798561151079,
+      "force_loss": 0.006577649619430304,
+      "step": 29210
+    },
+    {
+      "epoch": 26.276978417266186,
+      "grad_norm": 0.28308770060539246,
+      "learning_rate": 5.626939572155276e-05,
+      "loss": 0.0204,
+      "step": 29220
+    },
+    {
+      "action_loss": 0.0046681291423738,
+      "epoch": 26.276978417266186,
+      "step": 29220
+    },
+    {
+      "epoch": 26.276978417266186,
+      "step": 29220,
+      "torque_loss": 0.0951072946190834
+    },
+    {
+      "epoch": 26.276978417266186,
+      "force_loss": 0.007679952774196863,
+      "step": 29220
+    },
+    {
+      "epoch": 26.285971223021583,
+      "grad_norm": 0.21266824007034302,
+      "learning_rate": 5.6242054432533054e-05,
+      "loss": 0.0191,
+      "step": 29230
+    },
+    {
+      "action_loss": 0.004444051533937454,
+      "epoch": 26.285971223021583,
+      "step": 29230
+    },
+    {
+      "epoch": 26.285971223021583,
+      "step": 29230,
+      "torque_loss": 0.16044634580612183
+    },
+    {
+      "epoch": 26.285971223021583,
+      "force_loss": 0.006212200969457626,
+      "step": 29230
+    },
+    {
+      "epoch": 26.294964028776977,
+      "grad_norm": 0.5010757446289062,
+      "learning_rate": 5.621471124734201e-05,
+      "loss": 0.019,
+      "step": 29240
+    },
+    {
+      "action_loss": 0.004960187710821629,
+      "epoch": 26.294964028776977,
+      "step": 29240
+    },
+    {
+      "epoch": 26.294964028776977,
+      "step": 29240,
+      "torque_loss": 0.13931436836719513
+    },
+    {
+      "epoch": 26.294964028776977,
+      "force_loss": 0.0017164964228868484,
+      "step": 29240
+    },
+    {
+      "epoch": 26.303956834532375,
+      "grad_norm": 0.5563440918922424,
+      "learning_rate": 5.6187366174285794e-05,
+      "loss": 0.0248,
+      "step": 29250
+    },
+    {
+      "action_loss": 0.003914941567927599,
+      "epoch": 26.303956834532375,
+      "step": 29250
+    },
+    {
+      "epoch": 26.303956834532375,
+      "step": 29250,
+      "torque_loss": 0.19958682358264923
+    },
+    {
+      "epoch": 26.303956834532375,
+      "force_loss": 0.003421303117647767,
+      "step": 29250
+    },
+    {
+      "epoch": 26.31294964028777,
+      "grad_norm": 0.3078116476535797,
+      "learning_rate": 5.616001922167109e-05,
+      "loss": 0.0205,
+      "step": 29260
+    },
+    {
+      "action_loss": 0.0027273904997855425,
+      "epoch": 26.31294964028777,
+      "step": 29260
+    },
+    {
+      "epoch": 26.31294964028777,
+      "step": 29260,
+      "torque_loss": 0.12430980056524277
+    },
+    {
+      "epoch": 26.31294964028777,
+      "force_loss": 0.0021364546846598387,
+      "step": 29260
+    },
+    {
+      "epoch": 26.321942446043167,
+      "grad_norm": 0.5256577730178833,
+      "learning_rate": 5.61326703978052e-05,
+      "loss": 0.0209,
+      "step": 29270
+    },
+    {
+      "action_loss": 0.001910290215164423,
+      "epoch": 26.321942446043167,
+      "step": 29270
+    },
+    {
+      "epoch": 26.321942446043167,
+      "step": 29270,
+      "torque_loss": 0.12559562921524048
+    },
+    {
+      "epoch": 26.321942446043167,
+      "force_loss": 0.0015483573079109192,
+      "step": 29270
+    },
+    {
+      "epoch": 26.33093525179856,
+      "grad_norm": 0.42534586787223816,
+      "learning_rate": 5.6105319710995964e-05,
+      "loss": 0.0189,
+      "step": 29280
+    },
+    {
+      "action_loss": 0.0027956177946180105,
+      "epoch": 26.33093525179856,
+      "step": 29280
+    },
+    {
+      "epoch": 26.33093525179856,
+      "step": 29280,
+      "torque_loss": 0.13496915996074677
+    },
+    {
+      "epoch": 26.33093525179856,
+      "force_loss": 0.0017784927040338516,
+      "step": 29280
+    },
+    {
+      "epoch": 26.33992805755396,
+      "grad_norm": 0.22000323235988617,
+      "learning_rate": 5.60779671695518e-05,
+      "loss": 0.0192,
+      "step": 29290
+    },
+    {
+      "action_loss": 0.003970836754888296,
+      "epoch": 26.33992805755396,
+      "step": 29290
+    },
+    {
+      "epoch": 26.33992805755396,
+      "step": 29290,
+      "torque_loss": 0.1045316755771637
+    },
+    {
+      "epoch": 26.33992805755396,
+      "force_loss": 0.0021295619662851095,
+      "step": 29290
+    },
+    {
+      "epoch": 26.348920863309353,
+      "grad_norm": 0.2960430383682251,
+      "learning_rate": 5.6050612781781684e-05,
+      "loss": 0.0191,
+      "step": 29300
+    },
+    {
+      "action_loss": 0.0014416227350011468,
+      "epoch": 26.348920863309353,
+      "step": 29300
+    },
+    {
+      "epoch": 26.348920863309353,
+      "step": 29300,
+      "torque_loss": 0.12320087105035782
+    },
+    {
+      "epoch": 26.348920863309353,
+      "force_loss": 0.001710645854473114,
+      "step": 29300
+    },
+    {
+      "epoch": 26.357913669064747,
+      "grad_norm": 0.2702453136444092,
+      "learning_rate": 5.602325655599516e-05,
+      "loss": 0.0182,
+      "step": 29310
+    },
+    {
+      "action_loss": 0.012626200914382935,
+      "epoch": 26.357913669064747,
+      "step": 29310
+    },
+    {
+      "epoch": 26.357913669064747,
+      "step": 29310,
+      "torque_loss": 0.13518483936786652
+    },
+    {
+      "epoch": 26.357913669064747,
+      "force_loss": 0.011082902550697327,
+      "step": 29310
+    },
+    {
+      "epoch": 26.366906474820144,
+      "grad_norm": 0.2371370941400528,
+      "learning_rate": 5.599589850050234e-05,
+      "loss": 0.0187,
+      "step": 29320
+    },
+    {
+      "action_loss": 0.002585978014394641,
+      "epoch": 26.366906474820144,
+      "step": 29320
+    },
+    {
+      "epoch": 26.366906474820144,
+      "step": 29320,
+      "torque_loss": 0.14565639197826385
+    },
+    {
+      "epoch": 26.366906474820144,
+      "force_loss": 0.001961111556738615,
+      "step": 29320
+    },
+    {
+      "epoch": 26.37589928057554,
+      "grad_norm": 0.1642235964536667,
+      "learning_rate": 5.5968538623613874e-05,
+      "loss": 0.0189,
+      "step": 29330
+    },
+    {
+      "action_loss": 0.0021901302970945835,
+      "epoch": 26.37589928057554,
+      "step": 29330
+    },
+    {
+      "epoch": 26.37589928057554,
+      "step": 29330,
+      "torque_loss": 0.1027688980102539
+    },
+    {
+      "epoch": 26.37589928057554,
+      "force_loss": 0.0055335513316094875,
+      "step": 29330
+    },
+    {
+      "epoch": 26.384892086330936,
+      "grad_norm": 0.302939772605896,
+      "learning_rate": 5.594117693364095e-05,
+      "loss": 0.0178,
+      "step": 29340
+    },
+    {
+      "action_loss": 0.011924631893634796,
+      "epoch": 26.384892086330936,
+      "step": 29340
+    },
+    {
+      "epoch": 26.384892086330936,
+      "step": 29340,
+      "torque_loss": 0.12345170974731445
+    },
+    {
+      "epoch": 26.384892086330936,
+      "force_loss": 0.004974150564521551,
+      "step": 29340
+    },
+    {
+      "epoch": 26.39388489208633,
+      "grad_norm": 0.2409353405237198,
+      "learning_rate": 5.591381343889535e-05,
+      "loss": 0.0194,
+      "step": 29350
+    },
+    {
+      "action_loss": 0.008909042924642563,
+      "epoch": 26.39388489208633,
+      "step": 29350
+    },
+    {
+      "epoch": 26.39388489208633,
+      "step": 29350,
+      "torque_loss": 0.1389877200126648
+    },
+    {
+      "epoch": 26.39388489208633,
+      "force_loss": 0.010384388267993927,
+      "step": 29350
+    },
+    {
+      "epoch": 26.402877697841728,
+      "grad_norm": 0.14128169417381287,
+      "learning_rate": 5.5886448147689355e-05,
+      "loss": 0.02,
+      "step": 29360
+    },
+    {
+      "action_loss": 0.004463754128664732,
+      "epoch": 26.402877697841728,
+      "step": 29360
+    },
+    {
+      "epoch": 26.402877697841728,
+      "step": 29360,
+      "torque_loss": 0.13355658948421478
+    },
+    {
+      "epoch": 26.402877697841728,
+      "force_loss": 0.008837021887302399,
+      "step": 29360
+    },
+    {
+      "epoch": 26.41187050359712,
+      "grad_norm": 0.3179236650466919,
+      "learning_rate": 5.585908106833585e-05,
+      "loss": 0.0159,
+      "step": 29370
+    },
+    {
+      "action_loss": 0.0023016862105578184,
+      "epoch": 26.41187050359712,
+      "step": 29370
+    },
+    {
+      "epoch": 26.41187050359712,
+      "step": 29370,
+      "torque_loss": 0.1341308206319809
+    },
+    {
+      "epoch": 26.41187050359712,
+      "force_loss": 0.0015278387581929564,
+      "step": 29370
+    },
+    {
+      "epoch": 26.42086330935252,
+      "grad_norm": 0.21255384385585785,
+      "learning_rate": 5.5831712209148226e-05,
+      "loss": 0.0197,
+      "step": 29380
+    },
+    {
+      "action_loss": 0.014123695902526379,
+      "epoch": 26.42086330935252,
+      "step": 29380
+    },
+    {
+      "epoch": 26.42086330935252,
+      "step": 29380,
+      "torque_loss": 0.18206168711185455
+    },
+    {
+      "epoch": 26.42086330935252,
+      "force_loss": 0.023765431717038155,
+      "step": 29380
+    },
+    {
+      "epoch": 26.429856115107913,
+      "grad_norm": 0.23572397232055664,
+      "learning_rate": 5.58043415784404e-05,
+      "loss": 0.0207,
+      "step": 29390
+    },
+    {
+      "action_loss": 0.005147491116076708,
+      "epoch": 26.429856115107913,
+      "step": 29390
+    },
+    {
+      "epoch": 26.429856115107913,
+      "step": 29390,
+      "torque_loss": 0.12641461193561554
+    },
+    {
+      "epoch": 26.429856115107913,
+      "force_loss": 0.005691490601748228,
+      "step": 29390
+    },
+    {
+      "epoch": 26.43884892086331,
+      "grad_norm": 0.32689836621284485,
+      "learning_rate": 5.577696918452686e-05,
+      "loss": 0.0186,
+      "step": 29400
+    },
+    {
+      "action_loss": 0.0063631427474319935,
+      "epoch": 26.43884892086331,
+      "step": 29400
+    },
+    {
+      "epoch": 26.43884892086331,
+      "step": 29400,
+      "torque_loss": 0.2093459963798523
+    },
+    {
+      "epoch": 26.43884892086331,
+      "force_loss": 0.004051268566399813,
+      "step": 29400
+    },
+    {
+      "epoch": 26.447841726618705,
+      "grad_norm": 0.1445700079202652,
+      "learning_rate": 5.5749595035722604e-05,
+      "loss": 0.0345,
+      "step": 29410
+    },
+    {
+      "action_loss": 0.005245273467153311,
+      "epoch": 26.447841726618705,
+      "step": 29410
+    },
+    {
+      "epoch": 26.447841726618705,
+      "step": 29410,
+      "torque_loss": 0.10218346863985062
+    },
+    {
+      "epoch": 26.447841726618705,
+      "force_loss": 0.0031239886302500963,
+      "step": 29410
+    },
+    {
+      "epoch": 26.4568345323741,
+      "grad_norm": 0.1571255773305893,
+      "learning_rate": 5.5722219140343193e-05,
+      "loss": 0.0241,
+      "step": 29420
+    },
+    {
+      "action_loss": 0.003722322406247258,
+      "epoch": 26.4568345323741,
+      "step": 29420
+    },
+    {
+      "epoch": 26.4568345323741,
+      "step": 29420,
+      "torque_loss": 0.10764622688293457
+    },
+    {
+      "epoch": 26.4568345323741,
+      "force_loss": 0.0037944603245705366,
+      "step": 29420
+    },
+    {
+      "epoch": 26.465827338129497,
+      "grad_norm": 0.4807431399822235,
+      "learning_rate": 5.56948415067047e-05,
+      "loss": 0.0206,
+      "step": 29430
+    },
+    {
+      "action_loss": 0.018917372450232506,
+      "epoch": 26.465827338129497,
+      "step": 29430
+    },
+    {
+      "epoch": 26.465827338129497,
+      "step": 29430,
+      "torque_loss": 0.18301136791706085
+    },
+    {
+      "epoch": 26.465827338129497,
+      "force_loss": 0.020734211429953575,
+      "step": 29430
+    },
+    {
+      "epoch": 26.47482014388489,
+      "grad_norm": 0.232110857963562,
+      "learning_rate": 5.5667462143123704e-05,
+      "loss": 0.0245,
+      "step": 29440
+    },
+    {
+      "action_loss": 0.002458919771015644,
+      "epoch": 26.47482014388489,
+      "step": 29440
+    },
+    {
+      "epoch": 26.47482014388489,
+      "step": 29440,
+      "torque_loss": 0.15756447613239288
+    },
+    {
+      "epoch": 26.47482014388489,
+      "force_loss": 0.001853745779953897,
+      "step": 29440
+    },
+    {
+      "epoch": 26.48381294964029,
+      "grad_norm": 0.1807197630405426,
+      "learning_rate": 5.564008105791737e-05,
+      "loss": 0.0171,
+      "step": 29450
+    },
+    {
+      "action_loss": 0.004783970769494772,
+      "epoch": 26.48381294964029,
+      "step": 29450
+    },
+    {
+      "epoch": 26.48381294964029,
+      "step": 29450,
+      "torque_loss": 0.10162629932165146
+    },
+    {
+      "epoch": 26.48381294964029,
+      "force_loss": 0.0039339023642241955,
+      "step": 29450
+    },
+    {
+      "epoch": 26.492805755395683,
+      "grad_norm": 0.35580533742904663,
+      "learning_rate": 5.5612698259403316e-05,
+      "loss": 0.023,
+      "step": 29460
+    },
+    {
+      "action_loss": 0.0036192673724144697,
+      "epoch": 26.492805755395683,
+      "step": 29460
+    },
+    {
+      "epoch": 26.492805755395683,
+      "step": 29460,
+      "torque_loss": 0.10597450286149979
+    },
+    {
+      "epoch": 26.492805755395683,
+      "force_loss": 0.004650396760553122,
+      "step": 29460
+    },
+    {
+      "epoch": 26.50179856115108,
+      "grad_norm": 0.24788348376750946,
+      "learning_rate": 5.5585313755899724e-05,
+      "loss": 0.0224,
+      "step": 29470
+    },
+    {
+      "action_loss": 0.010820182971656322,
+      "epoch": 26.50179856115108,
+      "step": 29470
+    },
+    {
+      "epoch": 26.50179856115108,
+      "step": 29470,
+      "torque_loss": 0.11108478903770447
+    },
+    {
+      "epoch": 26.50179856115108,
+      "force_loss": 0.005138583481311798,
+      "step": 29470
+    },
+    {
+      "epoch": 26.510791366906474,
+      "grad_norm": 0.6962348222732544,
+      "learning_rate": 5.5557927555725285e-05,
+      "loss": 0.0243,
+      "step": 29480
+    },
+    {
+      "action_loss": 0.003334566717967391,
+      "epoch": 26.510791366906474,
+      "step": 29480
+    },
+    {
+      "epoch": 26.510791366906474,
+      "step": 29480,
+      "torque_loss": 0.1051768884062767
+    },
+    {
+      "epoch": 26.510791366906474,
+      "force_loss": 0.004038270562887192,
+      "step": 29480
+    },
+    {
+      "epoch": 26.519784172661872,
+      "grad_norm": 0.2491716593503952,
+      "learning_rate": 5.55305396671992e-05,
+      "loss": 0.0177,
+      "step": 29490
+    },
+    {
+      "action_loss": 0.004448357503861189,
+      "epoch": 26.519784172661872,
+      "step": 29490
+    },
+    {
+      "epoch": 26.519784172661872,
+      "step": 29490,
+      "torque_loss": 0.09822774678468704
+    },
+    {
+      "epoch": 26.519784172661872,
+      "force_loss": 0.00592459924519062,
+      "step": 29490
+    },
+    {
+      "epoch": 26.528776978417266,
+      "grad_norm": 0.4447100758552551,
+      "learning_rate": 5.55031500986412e-05,
+      "loss": 0.0238,
+      "step": 29500
+    },
+    {
+      "action_loss": 0.0025491274427622557,
+      "epoch": 26.528776978417266,
+      "step": 29500
+    },
+    {
+      "epoch": 26.528776978417266,
+      "step": 29500,
+      "torque_loss": 0.10773605853319168
+    },
+    {
+      "epoch": 26.528776978417266,
+      "force_loss": 0.0015268739080056548,
+      "step": 29500
+    },
+    {
+      "epoch": 26.53776978417266,
+      "grad_norm": 0.5370615720748901,
+      "learning_rate": 5.547575885837149e-05,
+      "loss": 0.021,
+      "step": 29510
+    },
+    {
+      "action_loss": 0.004978204611688852,
+      "epoch": 26.53776978417266,
+      "step": 29510
+    },
+    {
+      "epoch": 26.53776978417266,
+      "step": 29510,
+      "torque_loss": 0.16674725711345673
+    },
+    {
+      "epoch": 26.53776978417266,
+      "force_loss": 0.00587942311540246,
+      "step": 29510
+    },
+    {
+      "epoch": 26.546762589928058,
+      "grad_norm": 0.20173229277133942,
+      "learning_rate": 5.5448365954710825e-05,
+      "loss": 0.022,
+      "step": 29520
+    },
+    {
+      "action_loss": 0.018634973093867302,
+      "epoch": 26.546762589928058,
+      "step": 29520
+    },
+    {
+      "epoch": 26.546762589928058,
+      "step": 29520,
+      "torque_loss": 0.17407572269439697
+    },
+    {
+      "epoch": 26.546762589928058,
+      "force_loss": 0.0200891625136137,
+      "step": 29520
+    },
+    {
+      "epoch": 26.555755395683452,
+      "grad_norm": 0.4266655743122101,
+      "learning_rate": 5.5420971395980446e-05,
+      "loss": 0.0229,
+      "step": 29530
+    },
+    {
+      "action_loss": 0.011320815421640873,
+      "epoch": 26.555755395683452,
+      "step": 29530
+    },
+    {
+      "epoch": 26.555755395683452,
+      "step": 29530,
+      "torque_loss": 0.12391838431358337
+    },
+    {
+      "epoch": 26.555755395683452,
+      "force_loss": 0.011937855742871761,
+      "step": 29530
+    },
+    {
+      "epoch": 26.56474820143885,
+      "grad_norm": 0.3762349486351013,
+      "learning_rate": 5.539357519050209e-05,
+      "loss": 0.022,
+      "step": 29540
+    },
+    {
+      "action_loss": 0.004747358616441488,
+      "epoch": 26.56474820143885,
+      "step": 29540
+    },
+    {
+      "epoch": 26.56474820143885,
+      "step": 29540,
+      "torque_loss": 0.09421396255493164
+    },
+    {
+      "epoch": 26.56474820143885,
+      "force_loss": 0.004423636011779308,
+      "step": 29540
+    },
+    {
+      "epoch": 26.573741007194243,
+      "grad_norm": 0.2552447021007538,
+      "learning_rate": 5.536617734659799e-05,
+      "loss": 0.0224,
+      "step": 29550
+    },
+    {
+      "action_loss": 0.0025371320080012083,
+      "epoch": 26.573741007194243,
+      "step": 29550
+    },
+    {
+      "epoch": 26.573741007194243,
+      "step": 29550,
+      "torque_loss": 0.09377231448888779
+    },
+    {
+      "epoch": 26.573741007194243,
+      "force_loss": 0.0017519989050924778,
+      "step": 29550
+    },
+    {
+      "epoch": 26.58273381294964,
+      "grad_norm": 0.46805620193481445,
+      "learning_rate": 5.533877787259091e-05,
+      "loss": 0.0177,
+      "step": 29560
+    },
+    {
+      "action_loss": 0.004050894174724817,
+      "epoch": 26.58273381294964,
+      "step": 29560
+    },
+    {
+      "epoch": 26.58273381294964,
+      "step": 29560,
+      "torque_loss": 0.12956112623214722
+    },
+    {
+      "epoch": 26.58273381294964,
+      "force_loss": 0.0105274124071002,
+      "step": 29560
+    },
+    {
+      "epoch": 26.591726618705035,
+      "grad_norm": 0.22202341258525848,
+      "learning_rate": 5.5311376776804044e-05,
+      "loss": 0.0202,
+      "step": 29570
+    },
+    {
+      "action_loss": 0.0020610897336155176,
+      "epoch": 26.591726618705035,
+      "step": 29570
+    },
+    {
+      "epoch": 26.591726618705035,
+      "step": 29570,
+      "torque_loss": 0.1485256403684616
+    },
+    {
+      "epoch": 26.591726618705035,
+      "force_loss": 0.0010620758403092623,
+      "step": 29570
+    },
+    {
+      "epoch": 26.600719424460433,
+      "grad_norm": 0.22027987241744995,
+      "learning_rate": 5.528397406756118e-05,
+      "loss": 0.0213,
+      "step": 29580
+    },
+    {
+      "action_loss": 0.007421335205435753,
+      "epoch": 26.600719424460433,
+      "step": 29580
+    },
+    {
+      "epoch": 26.600719424460433,
+      "step": 29580,
+      "torque_loss": 0.15243853628635406
+    },
+    {
+      "epoch": 26.600719424460433,
+      "force_loss": 0.008105410262942314,
+      "step": 29580
+    },
+    {
+      "epoch": 26.609712230215827,
+      "grad_norm": 0.18835632503032684,
+      "learning_rate": 5.525656975318652e-05,
+      "loss": 0.0174,
+      "step": 29590
+    },
+    {
+      "action_loss": 0.0021108174696564674,
+      "epoch": 26.609712230215827,
+      "step": 29590
+    },
+    {
+      "epoch": 26.609712230215827,
+      "step": 29590,
+      "torque_loss": 0.15360350906848907
+    },
+    {
+      "epoch": 26.609712230215827,
+      "force_loss": 0.0019917127210646868,
+      "step": 29590
+    },
+    {
+      "epoch": 26.618705035971225,
+      "grad_norm": 0.1570572853088379,
+      "learning_rate": 5.522916384200474e-05,
+      "loss": 0.0158,
+      "step": 29600
+    },
+    {
+      "action_loss": 0.003603878431022167,
+      "epoch": 26.618705035971225,
+      "step": 29600
+    },
+    {
+      "epoch": 26.618705035971225,
+      "step": 29600,
+      "torque_loss": 0.1303596943616867
+    },
+    {
+      "epoch": 26.618705035971225,
+      "force_loss": 0.004055473022162914,
+      "step": 29600
+    },
+    {
+      "epoch": 26.62769784172662,
+      "grad_norm": 0.215669184923172,
+      "learning_rate": 5.520175634234106e-05,
+      "loss": 0.0182,
+      "step": 29610
+    },
+    {
+      "action_loss": 0.0026220858562737703,
+      "epoch": 26.62769784172662,
+      "step": 29610
+    },
+    {
+      "epoch": 26.62769784172662,
+      "step": 29610,
+      "torque_loss": 0.12466021627187729
+    },
+    {
+      "epoch": 26.62769784172662,
+      "force_loss": 0.002202659845352173,
+      "step": 29610
+    },
+    {
+      "epoch": 26.636690647482013,
+      "grad_norm": 0.45026540756225586,
+      "learning_rate": 5.517434726252113e-05,
+      "loss": 0.0201,
+      "step": 29620
+    },
+    {
+      "action_loss": 0.004916761536151171,
+      "epoch": 26.636690647482013,
+      "step": 29620
+    },
+    {
+      "epoch": 26.636690647482013,
+      "step": 29620,
+      "torque_loss": 0.13714423775672913
+    },
+    {
+      "epoch": 26.636690647482013,
+      "force_loss": 0.0044364831410348415,
+      "step": 29620
+    },
+    {
+      "epoch": 26.64568345323741,
+      "grad_norm": 0.2543215751647949,
+      "learning_rate": 5.514693661087113e-05,
+      "loss": 0.0187,
+      "step": 29630
+    },
+    {
+      "action_loss": 0.0063333637081086636,
+      "epoch": 26.64568345323741,
+      "step": 29630
+    },
+    {
+      "epoch": 26.64568345323741,
+      "step": 29630,
+      "torque_loss": 0.13830113410949707
+    },
+    {
+      "epoch": 26.64568345323741,
+      "force_loss": 0.004002048168331385,
+      "step": 29630
+    },
+    {
+      "epoch": 26.654676258992804,
+      "grad_norm": 0.30882659554481506,
+      "learning_rate": 5.511952439571769e-05,
+      "loss": 0.0176,
+      "step": 29640
+    },
+    {
+      "action_loss": 0.013768695294857025,
+      "epoch": 26.654676258992804,
+      "step": 29640
+    },
+    {
+      "epoch": 26.654676258992804,
+      "step": 29640,
+      "torque_loss": 0.13508917391300201
+    },
+    {
+      "epoch": 26.654676258992804,
+      "force_loss": 0.02008710615336895,
+      "step": 29640
+    },
+    {
+      "epoch": 26.663669064748202,
+      "grad_norm": 0.19365213811397552,
+      "learning_rate": 5.509211062538791e-05,
+      "loss": 0.0187,
+      "step": 29650
+    },
+    {
+      "action_loss": 0.013977513648569584,
+      "epoch": 26.663669064748202,
+      "step": 29650
+    },
+    {
+      "epoch": 26.663669064748202,
+      "step": 29650,
+      "torque_loss": 0.11426761001348495
+    },
+    {
+      "epoch": 26.663669064748202,
+      "force_loss": 0.006032731384038925,
+      "step": 29650
+    },
+    {
+      "epoch": 26.672661870503596,
+      "grad_norm": 0.3829696774482727,
+      "learning_rate": 5.506469530820939e-05,
+      "loss": 0.0214,
+      "step": 29660
+    },
+    {
+      "action_loss": 0.005374563857913017,
+      "epoch": 26.672661870503596,
+      "step": 29660
+    },
+    {
+      "epoch": 26.672661870503596,
+      "step": 29660,
+      "torque_loss": 0.09295514971017838
+    },
+    {
+      "epoch": 26.672661870503596,
+      "force_loss": 0.0066895452328026295,
+      "step": 29660
+    },
+    {
+      "epoch": 26.681654676258994,
+      "grad_norm": 0.2453082650899887,
+      "learning_rate": 5.503727845251014e-05,
+      "loss": 0.0187,
+      "step": 29670
+    },
+    {
+      "action_loss": 0.0045484015718102455,
+      "epoch": 26.681654676258994,
+      "step": 29670
+    },
+    {
+      "epoch": 26.681654676258994,
+      "step": 29670,
+      "torque_loss": 0.1226678192615509
+    },
+    {
+      "epoch": 26.681654676258994,
+      "force_loss": 0.0032400668133050203,
+      "step": 29670
+    },
+    {
+      "epoch": 26.690647482014388,
+      "grad_norm": 0.18619810044765472,
+      "learning_rate": 5.50098600666187e-05,
+      "loss": 0.021,
+      "step": 29680
+    },
+    {
+      "action_loss": 0.0037671327590942383,
+      "epoch": 26.690647482014388,
+      "step": 29680
+    },
+    {
+      "epoch": 26.690647482014388,
+      "step": 29680,
+      "torque_loss": 0.13072626292705536
+    },
+    {
+      "epoch": 26.690647482014388,
+      "force_loss": 0.005459554959088564,
+      "step": 29680
+    },
+    {
+      "epoch": 26.699640287769785,
+      "grad_norm": 0.3322921395301819,
+      "learning_rate": 5.498244015886406e-05,
+      "loss": 0.0188,
+      "step": 29690
+    },
+    {
+      "action_loss": 0.003932202234864235,
+      "epoch": 26.699640287769785,
+      "step": 29690
+    },
+    {
+      "epoch": 26.699640287769785,
+      "step": 29690,
+      "torque_loss": 0.15101401507854462
+    },
+    {
+      "epoch": 26.699640287769785,
+      "force_loss": 0.00469462713226676,
+      "step": 29690
+    },
+    {
+      "epoch": 26.70863309352518,
+      "grad_norm": 0.2403191328048706,
+      "learning_rate": 5.495501873757565e-05,
+      "loss": 0.0192,
+      "step": 29700
+    },
+    {
+      "action_loss": 0.005818183068186045,
+      "epoch": 26.70863309352518,
+      "step": 29700
+    },
+    {
+      "epoch": 26.70863309352518,
+      "step": 29700,
+      "torque_loss": 0.15367823839187622
+    },
+    {
+      "epoch": 26.70863309352518,
+      "force_loss": 0.007292923983186483,
+      "step": 29700
+    },
+    {
+      "epoch": 26.717625899280577,
+      "grad_norm": 0.21703964471817017,
+      "learning_rate": 5.492759581108336e-05,
+      "loss": 0.0188,
+      "step": 29710
+    },
+    {
+      "action_loss": 0.006756692659109831,
+      "epoch": 26.717625899280577,
+      "step": 29710
+    },
+    {
+      "epoch": 26.717625899280577,
+      "step": 29710,
+      "torque_loss": 0.13324739038944244
+    },
+    {
+      "epoch": 26.717625899280577,
+      "force_loss": 0.004057153128087521,
+      "step": 29710
+    },
+    {
+      "epoch": 26.72661870503597,
+      "grad_norm": 0.3107157349586487,
+      "learning_rate": 5.490017138771759e-05,
+      "loss": 0.021,
+      "step": 29720
+    },
+    {
+      "action_loss": 0.010310097597539425,
+      "epoch": 26.72661870503597,
+      "step": 29720
+    },
+    {
+      "epoch": 26.72661870503597,
+      "step": 29720,
+      "torque_loss": 0.15190069377422333
+    },
+    {
+      "epoch": 26.72661870503597,
+      "force_loss": 0.009750566445291042,
+      "step": 29720
+    },
+    {
+      "epoch": 26.735611510791365,
+      "grad_norm": 0.21111157536506653,
+      "learning_rate": 5.487274547580912e-05,
+      "loss": 0.0196,
+      "step": 29730
+    },
+    {
+      "action_loss": 0.013572419993579388,
+      "epoch": 26.735611510791365,
+      "step": 29730
+    },
+    {
+      "epoch": 26.735611510791365,
+      "step": 29730,
+      "torque_loss": 0.14122606813907623
+    },
+    {
+      "epoch": 26.735611510791365,
+      "force_loss": 0.013210773468017578,
+      "step": 29730
+    },
+    {
+      "epoch": 26.744604316546763,
+      "grad_norm": 0.30866900086402893,
+      "learning_rate": 5.484531808368923e-05,
+      "loss": 0.0198,
+      "step": 29740
+    },
+    {
+      "action_loss": 0.002811520593240857,
+      "epoch": 26.744604316546763,
+      "step": 29740
+    },
+    {
+      "epoch": 26.744604316546763,
+      "step": 29740,
+      "torque_loss": 0.13623277842998505
+    },
+    {
+      "epoch": 26.744604316546763,
+      "force_loss": 0.0024728986900299788,
+      "step": 29740
+    },
+    {
+      "epoch": 26.753597122302157,
+      "grad_norm": 0.2113165259361267,
+      "learning_rate": 5.4817889219689656e-05,
+      "loss": 0.0185,
+      "step": 29750
+    },
+    {
+      "action_loss": 0.00368649885058403,
+      "epoch": 26.753597122302157,
+      "step": 29750
+    },
+    {
+      "epoch": 26.753597122302157,
+      "step": 29750,
+      "torque_loss": 0.09717867523431778
+    },
+    {
+      "epoch": 26.753597122302157,
+      "force_loss": 0.0054848454892635345,
+      "step": 29750
+    },
+    {
+      "epoch": 26.762589928057555,
+      "grad_norm": 0.37648633122444153,
+      "learning_rate": 5.4790458892142536e-05,
+      "loss": 0.0188,
+      "step": 29760
+    },
+    {
+      "action_loss": 0.0013744235038757324,
+      "epoch": 26.762589928057555,
+      "step": 29760
+    },
+    {
+      "epoch": 26.762589928057555,
+      "step": 29760,
+      "torque_loss": 0.09067843109369278
+    },
+    {
+      "epoch": 26.762589928057555,
+      "force_loss": 0.0010968871647492051,
+      "step": 29760
+    },
+    {
+      "epoch": 26.77158273381295,
+      "grad_norm": 0.3552525043487549,
+      "learning_rate": 5.476302710938048e-05,
+      "loss": 0.0177,
+      "step": 29770
+    },
+    {
+      "action_loss": 0.005411986727267504,
+      "epoch": 26.77158273381295,
+      "step": 29770
+    },
+    {
+      "epoch": 26.77158273381295,
+      "step": 29770,
+      "torque_loss": 0.10153061151504517
+    },
+    {
+      "epoch": 26.77158273381295,
+      "force_loss": 0.008366730064153671,
+      "step": 29770
+    },
+    {
+      "epoch": 26.780575539568346,
+      "grad_norm": 0.13714280724525452,
+      "learning_rate": 5.473559387973657e-05,
+      "loss": 0.0202,
+      "step": 29780
+    },
+    {
+      "action_loss": 0.008373484946787357,
+      "epoch": 26.780575539568346,
+      "step": 29780
+    },
+    {
+      "epoch": 26.780575539568346,
+      "step": 29780,
+      "torque_loss": 0.13095788657665253
+    },
+    {
+      "epoch": 26.780575539568346,
+      "force_loss": 0.007003776729106903,
+      "step": 29780
+    },
+    {
+      "epoch": 26.78956834532374,
+      "grad_norm": 0.3039674162864685,
+      "learning_rate": 5.470815921154425e-05,
+      "loss": 0.0196,
+      "step": 29790
+    },
+    {
+      "action_loss": 0.0020388788543641567,
+      "epoch": 26.78956834532374,
+      "step": 29790
+    },
+    {
+      "epoch": 26.78956834532374,
+      "step": 29790,
+      "torque_loss": 0.09394717216491699
+    },
+    {
+      "epoch": 26.78956834532374,
+      "force_loss": 0.001609127502888441,
+      "step": 29790
+    },
+    {
+      "epoch": 26.798561151079138,
+      "grad_norm": 0.11895343661308289,
+      "learning_rate": 5.468072311313749e-05,
+      "loss": 0.0155,
+      "step": 29800
+    },
+    {
+      "action_loss": 0.003226721426472068,
+      "epoch": 26.798561151079138,
+      "step": 29800
+    },
+    {
+      "epoch": 26.798561151079138,
+      "step": 29800,
+      "torque_loss": 0.13606567680835724
+    },
+    {
+      "epoch": 26.798561151079138,
+      "force_loss": 0.005955897271633148,
+      "step": 29800
+    },
+    {
+      "epoch": 26.807553956834532,
+      "grad_norm": 0.21088936924934387,
+      "learning_rate": 5.465328559285063e-05,
+      "loss": 0.0186,
+      "step": 29810
+    },
+    {
+      "action_loss": 0.0025117734912782907,
+      "epoch": 26.807553956834532,
+      "step": 29810
+    },
+    {
+      "epoch": 26.807553956834532,
+      "step": 29810,
+      "torque_loss": 0.1175898090004921
+    },
+    {
+      "epoch": 26.807553956834532,
+      "force_loss": 0.0012616555904969573,
+      "step": 29810
+    },
+    {
+      "epoch": 26.81654676258993,
+      "grad_norm": 0.24970100820064545,
+      "learning_rate": 5.462584665901849e-05,
+      "loss": 0.0174,
+      "step": 29820
+    },
+    {
+      "action_loss": 0.0017017270438373089,
+      "epoch": 26.81654676258993,
+      "step": 29820
+    },
+    {
+      "epoch": 26.81654676258993,
+      "step": 29820,
+      "torque_loss": 0.12260963767766953
+    },
+    {
+      "epoch": 26.81654676258993,
+      "force_loss": 0.0018574738642200828,
+      "step": 29820
+    },
+    {
+      "epoch": 26.825539568345324,
+      "grad_norm": 0.5178494453430176,
+      "learning_rate": 5.4598406319976235e-05,
+      "loss": 0.0215,
+      "step": 29830
+    },
+    {
+      "action_loss": 0.040359657257795334,
+      "epoch": 26.825539568345324,
+      "step": 29830
+    },
+    {
+      "epoch": 26.825539568345324,
+      "step": 29830,
+      "torque_loss": 0.1942255049943924
+    },
+    {
+      "epoch": 26.825539568345324,
+      "force_loss": 0.04720534011721611,
+      "step": 29830
+    },
+    {
+      "epoch": 26.834532374100718,
+      "grad_norm": 0.4567413628101349,
+      "learning_rate": 5.457096458405958e-05,
+      "loss": 0.0269,
+      "step": 29840
+    },
+    {
+      "action_loss": 0.0025946127716451883,
+      "epoch": 26.834532374100718,
+      "step": 29840
+    },
+    {
+      "epoch": 26.834532374100718,
+      "step": 29840,
+      "torque_loss": 0.10776170343160629
+    },
+    {
+      "epoch": 26.834532374100718,
+      "force_loss": 0.003184608416631818,
+      "step": 29840
+    },
+    {
+      "epoch": 26.843525179856115,
+      "grad_norm": 0.3079434931278229,
+      "learning_rate": 5.454352145960457e-05,
+      "loss": 0.0194,
+      "step": 29850
+    },
+    {
+      "action_loss": 0.002686734078451991,
+      "epoch": 26.843525179856115,
+      "step": 29850
+    },
+    {
+      "epoch": 26.843525179856115,
+      "step": 29850,
+      "torque_loss": 0.14496742188930511
+    },
+    {
+      "epoch": 26.843525179856115,
+      "force_loss": 0.002164140110835433,
+      "step": 29850
+    },
+    {
+      "epoch": 26.85251798561151,
+      "grad_norm": 0.18310968577861786,
+      "learning_rate": 5.4516076954947715e-05,
+      "loss": 0.0181,
+      "step": 29860
+    },
+    {
+      "action_loss": 0.0035923875402659178,
+      "epoch": 26.85251798561151,
+      "step": 29860
+    },
+    {
+      "epoch": 26.85251798561151,
+      "step": 29860,
+      "torque_loss": 0.13412249088287354
+    },
+    {
+      "epoch": 26.85251798561151,
+      "force_loss": 0.002886866917833686,
+      "step": 29860
+    },
+    {
+      "epoch": 26.861510791366907,
+      "grad_norm": 0.5331716537475586,
+      "learning_rate": 5.448863107842591e-05,
+      "loss": 0.0201,
+      "step": 29870
+    },
+    {
+      "action_loss": 0.00453979754820466,
+      "epoch": 26.861510791366907,
+      "step": 29870
+    },
+    {
+      "epoch": 26.861510791366907,
+      "step": 29870,
+      "torque_loss": 0.13979299366474152
+    },
+    {
+      "epoch": 26.861510791366907,
+      "force_loss": 0.005563817452639341,
+      "step": 29870
+    },
+    {
+      "epoch": 26.8705035971223,
+      "grad_norm": 0.35070788860321045,
+      "learning_rate": 5.446118383837651e-05,
+      "loss": 0.0183,
+      "step": 29880
+    },
+    {
+      "action_loss": 0.019334664568305016,
+      "epoch": 26.8705035971223,
+      "step": 29880
+    },
+    {
+      "epoch": 26.8705035971223,
+      "step": 29880,
+      "torque_loss": 0.16253288090229034
+    },
+    {
+      "epoch": 26.8705035971223,
+      "force_loss": 0.03553791344165802,
+      "step": 29880
+    },
+    {
+      "epoch": 26.8794964028777,
+      "grad_norm": 0.3043481707572937,
+      "learning_rate": 5.443373524313722e-05,
+      "loss": 0.0217,
+      "step": 29890
+    },
+    {
+      "action_loss": 0.0036943492013961077,
+      "epoch": 26.8794964028777,
+      "step": 29890
+    },
+    {
+      "epoch": 26.8794964028777,
+      "step": 29890,
+      "torque_loss": 0.10507988184690475
+    },
+    {
+      "epoch": 26.8794964028777,
+      "force_loss": 0.002917099744081497,
+      "step": 29890
+    },
+    {
+      "epoch": 26.888489208633093,
+      "grad_norm": 0.21925991773605347,
+      "learning_rate": 5.440628530104626e-05,
+      "loss": 0.0176,
+      "step": 29900
+    },
+    {
+      "action_loss": 0.0013846675865352154,
+      "epoch": 26.888489208633093,
+      "step": 29900
+    },
+    {
+      "epoch": 26.888489208633093,
+      "step": 29900,
+      "torque_loss": 0.13571317493915558
+    },
+    {
+      "epoch": 26.888489208633093,
+      "force_loss": 0.001583204255439341,
+      "step": 29900
+    },
+    {
+      "epoch": 26.89748201438849,
+      "grad_norm": 0.11622114479541779,
+      "learning_rate": 5.4378834020442146e-05,
+      "loss": 0.0178,
+      "step": 29910
+    },
+    {
+      "action_loss": 0.005388239398598671,
+      "epoch": 26.89748201438849,
+      "step": 29910
+    },
+    {
+      "epoch": 26.89748201438849,
+      "step": 29910,
+      "torque_loss": 0.1476847529411316
+    },
+    {
+      "epoch": 26.89748201438849,
+      "force_loss": 0.0067005157470703125,
+      "step": 29910
+    },
+    {
+      "epoch": 26.906474820143885,
+      "grad_norm": 0.3066387474536896,
+      "learning_rate": 5.4351381409663884e-05,
+      "loss": 0.0168,
+      "step": 29920
+    },
+    {
+      "action_loss": 0.0020175655372440815,
+      "epoch": 26.906474820143885,
+      "step": 29920
+    },
+    {
+      "epoch": 26.906474820143885,
+      "step": 29920,
+      "torque_loss": 0.1331719160079956
+    },
+    {
+      "epoch": 26.906474820143885,
+      "force_loss": 0.002690813271328807,
+      "step": 29920
+    },
+    {
+      "epoch": 26.915467625899282,
+      "grad_norm": 0.27886462211608887,
+      "learning_rate": 5.432392747705084e-05,
+      "loss": 0.0173,
+      "step": 29930
+    },
+    {
+      "action_loss": 0.004540421534329653,
+      "epoch": 26.915467625899282,
+      "step": 29930
+    },
+    {
+      "epoch": 26.915467625899282,
+      "step": 29930,
+      "torque_loss": 0.16133111715316772
+    },
+    {
+      "epoch": 26.915467625899282,
+      "force_loss": 0.0035180847626179457,
+      "step": 29930
+    },
+    {
+      "epoch": 26.924460431654676,
+      "grad_norm": 0.13023129105567932,
+      "learning_rate": 5.429647223094278e-05,
+      "loss": 0.017,
+      "step": 29940
+    },
+    {
+      "action_loss": 0.007440451066941023,
+      "epoch": 26.924460431654676,
+      "step": 29940
+    },
+    {
+      "epoch": 26.924460431654676,
+      "step": 29940,
+      "torque_loss": 0.13692821562290192
+    },
+    {
+      "epoch": 26.924460431654676,
+      "force_loss": 0.007159040775150061,
+      "step": 29940
+    },
+    {
+      "epoch": 26.93345323741007,
+      "grad_norm": 0.124339759349823,
+      "learning_rate": 5.4269015679679924e-05,
+      "loss": 0.019,
+      "step": 29950
+    },
+    {
+      "action_loss": 0.0022405320778489113,
+      "epoch": 26.93345323741007,
+      "step": 29950
+    },
+    {
+      "epoch": 26.93345323741007,
+      "step": 29950,
+      "torque_loss": 0.114387147128582
+    },
+    {
+      "epoch": 26.93345323741007,
+      "force_loss": 0.0014298348687589169,
+      "step": 29950
+    },
+    {
+      "epoch": 26.942446043165468,
+      "grad_norm": 0.37802475690841675,
+      "learning_rate": 5.424155783160281e-05,
+      "loss": 0.0153,
+      "step": 29960
+    },
+    {
+      "action_loss": 0.009256296791136265,
+      "epoch": 26.942446043165468,
+      "step": 29960
+    },
+    {
+      "epoch": 26.942446043165468,
+      "step": 29960,
+      "torque_loss": 0.1303991824388504
+    },
+    {
+      "epoch": 26.942446043165468,
+      "force_loss": 0.003956787753850222,
+      "step": 29960
+    },
+    {
+      "epoch": 26.951438848920862,
+      "grad_norm": 0.27281561493873596,
+      "learning_rate": 5.4214098695052415e-05,
+      "loss": 0.0213,
+      "step": 29970
+    },
+    {
+      "action_loss": 0.019115492701530457,
+      "epoch": 26.951438848920862,
+      "step": 29970
+    },
+    {
+      "epoch": 26.951438848920862,
+      "step": 29970,
+      "torque_loss": 0.1212594285607338
+    },
+    {
+      "epoch": 26.951438848920862,
+      "force_loss": 0.017049849033355713,
+      "step": 29970
+    },
+    {
+      "epoch": 26.96043165467626,
+      "grad_norm": 0.28068768978118896,
+      "learning_rate": 5.418663827837012e-05,
+      "loss": 0.0226,
+      "step": 29980
+    },
+    {
+      "action_loss": 0.0035388183314353228,
+      "epoch": 26.96043165467626,
+      "step": 29980
+    },
+    {
+      "epoch": 26.96043165467626,
+      "step": 29980,
+      "torque_loss": 0.12805412709712982
+    },
+    {
+      "epoch": 26.96043165467626,
+      "force_loss": 0.0059658922255039215,
+      "step": 29980
+    },
+    {
+      "epoch": 26.969424460431654,
+      "grad_norm": 0.18451905250549316,
+      "learning_rate": 5.415917658989763e-05,
+      "loss": 0.0173,
+      "step": 29990
+    },
+    {
+      "action_loss": 0.0013209698954597116,
+      "epoch": 26.969424460431654,
+      "step": 29990
+    },
+    {
+      "epoch": 26.969424460431654,
+      "step": 29990,
+      "torque_loss": 0.09117806702852249
+    },
+    {
+      "epoch": 26.969424460431654,
+      "force_loss": 0.0021667832043021917,
+      "step": 29990
+    },
+    {
+      "epoch": 26.97841726618705,
+      "grad_norm": 0.26481354236602783,
+      "learning_rate": 5.413171363797713e-05,
+      "loss": 0.0207,
+      "step": 30000
+    },
+    {
+      "action_loss": 0.00438087759539485,
+      "epoch": 26.97841726618705,
+      "step": 30000
+    },
+    {
+      "epoch": 26.97841726618705,
+      "step": 30000,
+      "torque_loss": 0.17940616607666016
+    },
+    {
+      "epoch": 26.97841726618705,
+      "force_loss": 0.0053337570279836655,
+      "step": 30000
+    },
+    {
+      "epoch": 26.987410071942445,
+      "grad_norm": 0.2747499346733093,
+      "learning_rate": 5.4104249430951116e-05,
+      "loss": 0.0176,
+      "step": 30010
+    },
+    {
+      "action_loss": 0.0031281306874006987,
+      "epoch": 26.987410071942445,
+      "step": 30010
+    },
+    {
+      "epoch": 26.987410071942445,
+      "step": 30010,
+      "torque_loss": 0.15624935925006866
+    },
+    {
+      "epoch": 26.987410071942445,
+      "force_loss": 0.006158431526273489,
+      "step": 30010
+    },
+    {
+      "epoch": 26.996402877697843,
+      "grad_norm": 0.525160014629364,
+      "learning_rate": 5.4076783977162494e-05,
+      "loss": 0.019,
+      "step": 30020
+    },
+    {
+      "action_loss": 0.005360439419746399,
+      "epoch": 26.996402877697843,
+      "step": 30020
+    },
+    {
+      "epoch": 26.996402877697843,
+      "step": 30020,
+      "torque_loss": 0.12258223444223404
+    },
+    {
+      "epoch": 26.996402877697843,
+      "force_loss": 0.009998547844588757,
+      "step": 30020
+    },
+    {
+      "epoch": 27.005395683453237,
+      "grad_norm": 0.27929526567459106,
+      "learning_rate": 5.4049317284954525e-05,
+      "loss": 0.0178,
+      "step": 30030
+    },
+    {
+      "action_loss": 0.003984289709478617,
+      "epoch": 27.005395683453237,
+      "step": 30030
+    },
+    {
+      "epoch": 27.005395683453237,
+      "step": 30030,
+      "torque_loss": 0.1588083952665329
+    },
+    {
+      "epoch": 27.005395683453237,
+      "force_loss": 0.015379909425973892,
+      "step": 30030
+    },
+    {
+      "epoch": 27.014388489208635,
+      "grad_norm": 0.17650508880615234,
+      "learning_rate": 5.4021849362670884e-05,
+      "loss": 0.0235,
+      "step": 30040
+    },
+    {
+      "action_loss": 0.0028035659343004227,
+      "epoch": 27.014388489208635,
+      "step": 30040
+    },
+    {
+      "epoch": 27.014388489208635,
+      "step": 30040,
+      "torque_loss": 0.08240079879760742
+    },
+    {
+      "epoch": 27.014388489208635,
+      "force_loss": 0.0018409885233268142,
+      "step": 30040
+    },
+    {
+      "epoch": 27.02338129496403,
+      "grad_norm": 0.2881733775138855,
+      "learning_rate": 5.3994380218655604e-05,
+      "loss": 0.0209,
+      "step": 30050
+    },
+    {
+      "action_loss": 0.005864724982529879,
+      "epoch": 27.02338129496403,
+      "step": 30050
+    },
+    {
+      "epoch": 27.02338129496403,
+      "step": 30050,
+      "torque_loss": 0.11897751688957214
+    },
+    {
+      "epoch": 27.02338129496403,
+      "force_loss": 0.007947350852191448,
+      "step": 30050
+    },
+    {
+      "epoch": 27.032374100719423,
+      "grad_norm": 0.28892263770103455,
+      "learning_rate": 5.396690986125309e-05,
+      "loss": 0.0198,
+      "step": 30060
+    },
+    {
+      "action_loss": 0.0025683999992907047,
+      "epoch": 27.032374100719423,
+      "step": 30060
+    },
+    {
+      "epoch": 27.032374100719423,
+      "step": 30060,
+      "torque_loss": 0.13800550997257233
+    },
+    {
+      "epoch": 27.032374100719423,
+      "force_loss": 0.0018803342245519161,
+      "step": 30060
+    },
+    {
+      "epoch": 27.04136690647482,
+      "grad_norm": 0.3276672959327698,
+      "learning_rate": 5.3939438298808075e-05,
+      "loss": 0.0204,
+      "step": 30070
+    },
+    {
+      "action_loss": 0.0027023537550121546,
+      "epoch": 27.04136690647482,
+      "step": 30070
+    },
+    {
+      "epoch": 27.04136690647482,
+      "step": 30070,
+      "torque_loss": 0.13884642720222473
+    },
+    {
+      "epoch": 27.04136690647482,
+      "force_loss": 0.003128311363980174,
+      "step": 30070
+    },
+    {
+      "epoch": 27.050359712230215,
+      "grad_norm": 0.450895756483078,
+      "learning_rate": 5.3911965539665744e-05,
+      "loss": 0.0216,
+      "step": 30080
+    },
+    {
+      "action_loss": 0.007366721052676439,
+      "epoch": 27.050359712230215,
+      "step": 30080
+    },
+    {
+      "epoch": 27.050359712230215,
+      "step": 30080,
+      "torque_loss": 0.11305730789899826
+    },
+    {
+      "epoch": 27.050359712230215,
+      "force_loss": 0.0037769523914903402,
+      "step": 30080
+    },
+    {
+      "epoch": 27.059352517985612,
+      "grad_norm": 0.21291057765483856,
+      "learning_rate": 5.388449159217156e-05,
+      "loss": 0.0203,
+      "step": 30090
+    },
+    {
+      "action_loss": 0.004496064502745867,
+      "epoch": 27.059352517985612,
+      "step": 30090
+    },
+    {
+      "epoch": 27.059352517985612,
+      "step": 30090,
+      "torque_loss": 0.1814671754837036
+    },
+    {
+      "epoch": 27.059352517985612,
+      "force_loss": 0.004324872046709061,
+      "step": 30090
+    },
+    {
+      "epoch": 27.068345323741006,
+      "grad_norm": 0.15907500684261322,
+      "learning_rate": 5.3857016464671385e-05,
+      "loss": 0.0182,
+      "step": 30100
+    },
+    {
+      "action_loss": 0.0033876586239784956,
+      "epoch": 27.068345323741006,
+      "step": 30100
+    },
+    {
+      "epoch": 27.068345323741006,
+      "step": 30100,
+      "torque_loss": 0.12786339223384857
+    },
+    {
+      "epoch": 27.068345323741006,
+      "force_loss": 0.0033877205569297075,
+      "step": 30100
+    },
+    {
+      "epoch": 27.077338129496404,
+      "grad_norm": 0.1309467852115631,
+      "learning_rate": 5.382954016551146e-05,
+      "loss": 0.0171,
+      "step": 30110
+    },
+    {
+      "action_loss": 0.002663847291842103,
+      "epoch": 27.077338129496404,
+      "step": 30110
+    },
+    {
+      "epoch": 27.077338129496404,
+      "step": 30110,
+      "torque_loss": 0.11073434352874756
+    },
+    {
+      "epoch": 27.077338129496404,
+      "force_loss": 0.004452723544090986,
+      "step": 30110
+    },
+    {
+      "epoch": 27.086330935251798,
+      "grad_norm": 0.46172890067100525,
+      "learning_rate": 5.380206270303835e-05,
+      "loss": 0.0196,
+      "step": 30120
+    },
+    {
+      "action_loss": 0.004135601222515106,
+      "epoch": 27.086330935251798,
+      "step": 30120
+    },
+    {
+      "epoch": 27.086330935251798,
+      "step": 30120,
+      "torque_loss": 0.138643279671669
+    },
+    {
+      "epoch": 27.086330935251798,
+      "force_loss": 0.00646270951256156,
+      "step": 30120
+    },
+    {
+      "epoch": 27.095323741007196,
+      "grad_norm": 0.17744766175746918,
+      "learning_rate": 5.377458408559897e-05,
+      "loss": 0.019,
+      "step": 30130
+    },
+    {
+      "action_loss": 0.005453914403915405,
+      "epoch": 27.095323741007196,
+      "step": 30130
+    },
+    {
+      "epoch": 27.095323741007196,
+      "step": 30130,
+      "torque_loss": 0.1135575994849205
+    },
+    {
+      "epoch": 27.095323741007196,
+      "force_loss": 0.003186618909239769,
+      "step": 30130
+    },
+    {
+      "epoch": 27.10431654676259,
+      "grad_norm": 0.2936372756958008,
+      "learning_rate": 5.374710432154061e-05,
+      "loss": 0.0193,
+      "step": 30140
+    },
+    {
+      "action_loss": 0.006100647617131472,
+      "epoch": 27.10431654676259,
+      "step": 30140
+    },
+    {
+      "epoch": 27.10431654676259,
+      "step": 30140,
+      "torque_loss": 0.17243479192256927
+    },
+    {
+      "epoch": 27.10431654676259,
+      "force_loss": 0.004507630597800016,
+      "step": 30140
+    },
+    {
+      "epoch": 27.113309352517987,
+      "grad_norm": 0.20095570385456085,
+      "learning_rate": 5.3719623419210886e-05,
+      "loss": 0.0181,
+      "step": 30150
+    },
+    {
+      "action_loss": 0.004813058767467737,
+      "epoch": 27.113309352517987,
+      "step": 30150
+    },
+    {
+      "epoch": 27.113309352517987,
+      "step": 30150,
+      "torque_loss": 0.17041879892349243
+    },
+    {
+      "epoch": 27.113309352517987,
+      "force_loss": 0.004747232887893915,
+      "step": 30150
+    },
+    {
+      "epoch": 27.12230215827338,
+      "grad_norm": 0.3171005845069885,
+      "learning_rate": 5.3692141386957786e-05,
+      "loss": 0.0172,
+      "step": 30160
+    },
+    {
+      "action_loss": 0.004677954595535994,
+      "epoch": 27.12230215827338,
+      "step": 30160
+    },
+    {
+      "epoch": 27.12230215827338,
+      "step": 30160,
+      "torque_loss": 0.0752430260181427
+    },
+    {
+      "epoch": 27.12230215827338,
+      "force_loss": 0.007557352539151907,
+      "step": 30160
+    },
+    {
+      "epoch": 27.131294964028775,
+      "grad_norm": 0.15077587962150574,
+      "learning_rate": 5.3664658233129616e-05,
+      "loss": 0.021,
+      "step": 30170
+    },
+    {
+      "action_loss": 0.005904344841837883,
+      "epoch": 27.131294964028775,
+      "step": 30170
+    },
+    {
+      "epoch": 27.131294964028775,
+      "step": 30170,
+      "torque_loss": 0.1867804378271103
+    },
+    {
+      "epoch": 27.131294964028775,
+      "force_loss": 0.0032666174229234457,
+      "step": 30170
+    },
+    {
+      "epoch": 27.140287769784173,
+      "grad_norm": 0.27926909923553467,
+      "learning_rate": 5.363717396607504e-05,
+      "loss": 0.0193,
+      "step": 30180
+    },
+    {
+      "action_loss": 0.0019679819233715534,
+      "epoch": 27.140287769784173,
+      "step": 30180
+    },
+    {
+      "epoch": 27.140287769784173,
+      "step": 30180,
+      "torque_loss": 0.09325297921895981
+    },
+    {
+      "epoch": 27.140287769784173,
+      "force_loss": 0.0022979744244366884,
+      "step": 30180
+    },
+    {
+      "epoch": 27.149280575539567,
+      "grad_norm": 0.21495363116264343,
+      "learning_rate": 5.360968859414305e-05,
+      "loss": 0.0173,
+      "step": 30190
+    },
+    {
+      "action_loss": 0.005858736112713814,
+      "epoch": 27.149280575539567,
+      "step": 30190
+    },
+    {
+      "epoch": 27.149280575539567,
+      "step": 30190,
+      "torque_loss": 0.13998264074325562
+    },
+    {
+      "epoch": 27.149280575539567,
+      "force_loss": 0.008681301027536392,
+      "step": 30190
+    },
+    {
+      "epoch": 27.158273381294965,
+      "grad_norm": 0.18916158378124237,
+      "learning_rate": 5.358220212568295e-05,
+      "loss": 0.0167,
+      "step": 30200
+    },
+    {
+      "action_loss": 0.022345414385199547,
+      "epoch": 27.158273381294965,
+      "step": 30200
+    },
+    {
+      "epoch": 27.158273381294965,
+      "step": 30200,
+      "torque_loss": 0.17376093566417694
+    },
+    {
+      "epoch": 27.158273381294965,
+      "force_loss": 0.023212112486362457,
+      "step": 30200
+    },
+    {
+      "epoch": 27.16726618705036,
+      "grad_norm": 0.17001235485076904,
+      "learning_rate": 5.355471456904444e-05,
+      "loss": 0.0179,
+      "step": 30210
+    },
+    {
+      "action_loss": 0.03106633387506008,
+      "epoch": 27.16726618705036,
+      "step": 30210
+    },
+    {
+      "epoch": 27.16726618705036,
+      "step": 30210,
+      "torque_loss": 0.19194336235523224
+    },
+    {
+      "epoch": 27.16726618705036,
+      "force_loss": 0.028198689222335815,
+      "step": 30210
+    },
+    {
+      "epoch": 27.176258992805757,
+      "grad_norm": 0.5119553804397583,
+      "learning_rate": 5.3527225932577495e-05,
+      "loss": 0.0241,
+      "step": 30220
+    },
+    {
+      "action_loss": 0.028138378635048866,
+      "epoch": 27.176258992805757,
+      "step": 30220
+    },
+    {
+      "epoch": 27.176258992805757,
+      "step": 30220,
+      "torque_loss": 0.22290174663066864
+    },
+    {
+      "epoch": 27.176258992805757,
+      "force_loss": 0.04579739645123482,
+      "step": 30220
+    },
+    {
+      "epoch": 27.18525179856115,
+      "grad_norm": 0.32575497031211853,
+      "learning_rate": 5.349973622463246e-05,
+      "loss": 0.0238,
+      "step": 30230
+    },
+    {
+      "action_loss": 0.008045557886362076,
+      "epoch": 27.18525179856115,
+      "step": 30230
+    },
+    {
+      "epoch": 27.18525179856115,
+      "step": 30230,
+      "torque_loss": 0.1500275880098343
+    },
+    {
+      "epoch": 27.18525179856115,
+      "force_loss": 0.023438720032572746,
+      "step": 30230
+    },
+    {
+      "epoch": 27.194244604316548,
+      "grad_norm": 0.20883388817310333,
+      "learning_rate": 5.3472245453559956e-05,
+      "loss": 0.0179,
+      "step": 30240
+    },
+    {
+      "action_loss": 0.004487939178943634,
+      "epoch": 27.194244604316548,
+      "step": 30240
+    },
+    {
+      "epoch": 27.194244604316548,
+      "step": 30240,
+      "torque_loss": 0.14595277607440948
+    },
+    {
+      "epoch": 27.194244604316548,
+      "force_loss": 0.009683411568403244,
+      "step": 30240
+    },
+    {
+      "epoch": 27.203237410071942,
+      "grad_norm": 0.23641663789749146,
+      "learning_rate": 5.3444753627710955e-05,
+      "loss": 0.0207,
+      "step": 30250
+    },
+    {
+      "action_loss": 0.005362626165151596,
+      "epoch": 27.203237410071942,
+      "step": 30250
+    },
+    {
+      "epoch": 27.203237410071942,
+      "step": 30250,
+      "torque_loss": 0.12094292789697647
+    },
+    {
+      "epoch": 27.203237410071942,
+      "force_loss": 0.004349933937191963,
+      "step": 30250
+    },
+    {
+      "epoch": 27.21223021582734,
+      "grad_norm": 0.3433016240596771,
+      "learning_rate": 5.341726075543676e-05,
+      "loss": 0.0185,
+      "step": 30260
+    },
+    {
+      "action_loss": 0.0029825549572706223,
+      "epoch": 27.21223021582734,
+      "step": 30260
+    },
+    {
+      "epoch": 27.21223021582734,
+      "step": 30260,
+      "torque_loss": 0.15964572131633759
+    },
+    {
+      "epoch": 27.21223021582734,
+      "force_loss": 0.0013917317846789956,
+      "step": 30260
+    },
+    {
+      "epoch": 27.221223021582734,
+      "grad_norm": 0.7519928216934204,
+      "learning_rate": 5.338976684508898e-05,
+      "loss": 0.019,
+      "step": 30270
+    },
+    {
+      "action_loss": 0.021886220201849937,
+      "epoch": 27.221223021582734,
+      "step": 30270
+    },
+    {
+      "epoch": 27.221223021582734,
+      "step": 30270,
+      "torque_loss": 0.15924613177776337
+    },
+    {
+      "epoch": 27.221223021582734,
+      "force_loss": 0.016282370314002037,
+      "step": 30270
+    },
+    {
+      "epoch": 27.230215827338128,
+      "grad_norm": 0.1829599291086197,
+      "learning_rate": 5.336227190501953e-05,
+      "loss": 0.0182,
+      "step": 30280
+    },
+    {
+      "action_loss": 0.0018750490853562951,
+      "epoch": 27.230215827338128,
+      "step": 30280
+    },
+    {
+      "epoch": 27.230215827338128,
+      "step": 30280,
+      "torque_loss": 0.12079592794179916
+    },
+    {
+      "epoch": 27.230215827338128,
+      "force_loss": 0.0016502775251865387,
+      "step": 30280
+    },
+    {
+      "epoch": 27.239208633093526,
+      "grad_norm": 0.08746183663606644,
+      "learning_rate": 5.3334775943580664e-05,
+      "loss": 0.0193,
+      "step": 30290
+    },
+    {
+      "action_loss": 0.0016757560661062598,
+      "epoch": 27.239208633093526,
+      "step": 30290
+    },
+    {
+      "epoch": 27.239208633093526,
+      "step": 30290,
+      "torque_loss": 0.08195025473833084
+    },
+    {
+      "epoch": 27.239208633093526,
+      "force_loss": 0.0013245666632428765,
+      "step": 30290
+    },
+    {
+      "epoch": 27.24820143884892,
+      "grad_norm": 0.18650297820568085,
+      "learning_rate": 5.330727896912491e-05,
+      "loss": 0.0151,
+      "step": 30300
+    },
+    {
+      "action_loss": 0.003503574291244149,
+      "epoch": 27.24820143884892,
+      "step": 30300
+    },
+    {
+      "epoch": 27.24820143884892,
+      "step": 30300,
+      "torque_loss": 0.08101185411214828
+    },
+    {
+      "epoch": 27.24820143884892,
+      "force_loss": 0.004829870071262121,
+      "step": 30300
+    },
+    {
+      "epoch": 27.257194244604317,
+      "grad_norm": 0.3734641671180725,
+      "learning_rate": 5.327978099000511e-05,
+      "loss": 0.0167,
+      "step": 30310
+    },
+    {
+      "action_loss": 0.0027369551826268435,
+      "epoch": 27.257194244604317,
+      "step": 30310
+    },
+    {
+      "epoch": 27.257194244604317,
+      "step": 30310,
+      "torque_loss": 0.11790794134140015
+    },
+    {
+      "epoch": 27.257194244604317,
+      "force_loss": 0.00968188513070345,
+      "step": 30310
+    },
+    {
+      "epoch": 27.26618705035971,
+      "grad_norm": 0.20427779853343964,
+      "learning_rate": 5.3252282014574465e-05,
+      "loss": 0.018,
+      "step": 30320
+    },
+    {
+      "action_loss": 0.005135164130479097,
+      "epoch": 27.26618705035971,
+      "step": 30320
+    },
+    {
+      "epoch": 27.26618705035971,
+      "step": 30320,
+      "torque_loss": 0.11646892875432968
+    },
+    {
+      "epoch": 27.26618705035971,
+      "force_loss": 0.00904246885329485,
+      "step": 30320
+    },
+    {
+      "epoch": 27.27517985611511,
+      "grad_norm": 0.09920324385166168,
+      "learning_rate": 5.322478205118641e-05,
+      "loss": 0.0157,
+      "step": 30330
+    },
+    {
+      "action_loss": 0.00667195999994874,
+      "epoch": 27.27517985611511,
+      "step": 30330
+    },
+    {
+      "epoch": 27.27517985611511,
+      "step": 30330,
+      "torque_loss": 0.1300421804189682
+    },
+    {
+      "epoch": 27.27517985611511,
+      "force_loss": 0.0047349329106509686,
+      "step": 30330
+    },
+    {
+      "epoch": 27.284172661870503,
+      "grad_norm": 0.300405889749527,
+      "learning_rate": 5.3197281108194704e-05,
+      "loss": 0.0222,
+      "step": 30340
+    },
+    {
+      "action_loss": 0.004093907307833433,
+      "epoch": 27.284172661870503,
+      "step": 30340
+    },
+    {
+      "epoch": 27.284172661870503,
+      "step": 30340,
+      "torque_loss": 0.0945718064904213
+    },
+    {
+      "epoch": 27.284172661870503,
+      "force_loss": 0.009305437095463276,
+      "step": 30340
+    },
+    {
+      "epoch": 27.2931654676259,
+      "grad_norm": 0.27889901399612427,
+      "learning_rate": 5.316977919395342e-05,
+      "loss": 0.0174,
+      "step": 30350
+    },
+    {
+      "action_loss": 0.006806804332882166,
+      "epoch": 27.2931654676259,
+      "step": 30350
+    },
+    {
+      "epoch": 27.2931654676259,
+      "step": 30350,
+      "torque_loss": 0.15492509305477142
+    },
+    {
+      "epoch": 27.2931654676259,
+      "force_loss": 0.007464412599802017,
+      "step": 30350
+    },
+    {
+      "epoch": 27.302158273381295,
+      "grad_norm": 0.35391148924827576,
+      "learning_rate": 5.314227631681691e-05,
+      "loss": 0.0191,
+      "step": 30360
+    },
+    {
+      "action_loss": 0.006740959826856852,
+      "epoch": 27.302158273381295,
+      "step": 30360
+    },
+    {
+      "epoch": 27.302158273381295,
+      "step": 30360,
+      "torque_loss": 0.1226671114563942
+    },
+    {
+      "epoch": 27.302158273381295,
+      "force_loss": 0.004497290588915348,
+      "step": 30360
+    },
+    {
+      "epoch": 27.31115107913669,
+      "grad_norm": 0.17393139004707336,
+      "learning_rate": 5.311477248513982e-05,
+      "loss": 0.0171,
+      "step": 30370
+    },
+    {
+      "action_loss": 0.0036024879664182663,
+      "epoch": 27.31115107913669,
+      "step": 30370
+    },
+    {
+      "epoch": 27.31115107913669,
+      "step": 30370,
+      "torque_loss": 0.11064424365758896
+    },
+    {
+      "epoch": 27.31115107913669,
+      "force_loss": 0.004538074601441622,
+      "step": 30370
+    },
+    {
+      "epoch": 27.320143884892087,
+      "grad_norm": 0.2588514983654022,
+      "learning_rate": 5.30872677072771e-05,
+      "loss": 0.0219,
+      "step": 30380
+    },
+    {
+      "action_loss": 0.002123519079759717,
+      "epoch": 27.320143884892087,
+      "step": 30380
+    },
+    {
+      "epoch": 27.320143884892087,
+      "step": 30380,
+      "torque_loss": 0.13704858720302582
+    },
+    {
+      "epoch": 27.320143884892087,
+      "force_loss": 0.001857923693023622,
+      "step": 30380
+    },
+    {
+      "epoch": 27.32913669064748,
+      "grad_norm": 0.11446554958820343,
+      "learning_rate": 5.3059761991583954e-05,
+      "loss": 0.0211,
+      "step": 30390
+    },
+    {
+      "action_loss": 0.017162993550300598,
+      "epoch": 27.32913669064748,
+      "step": 30390
+    },
+    {
+      "epoch": 27.32913669064748,
+      "step": 30390,
+      "torque_loss": 0.15154393017292023
+    },
+    {
+      "epoch": 27.32913669064748,
+      "force_loss": 0.015158439986407757,
+      "step": 30390
+    },
+    {
+      "epoch": 27.33812949640288,
+      "grad_norm": 0.14296375215053558,
+      "learning_rate": 5.303225534641592e-05,
+      "loss": 0.0185,
+      "step": 30400
+    },
+    {
+      "action_loss": 0.004359942860901356,
+      "epoch": 27.33812949640288,
+      "step": 30400
+    },
+    {
+      "epoch": 27.33812949640288,
+      "step": 30400,
+      "torque_loss": 0.12148231267929077
+    },
+    {
+      "epoch": 27.33812949640288,
+      "force_loss": 0.007306875195354223,
+      "step": 30400
+    },
+    {
+      "epoch": 27.347122302158272,
+      "grad_norm": 0.2763224244117737,
+      "learning_rate": 5.300474778012875e-05,
+      "loss": 0.0183,
+      "step": 30410
+    },
+    {
+      "action_loss": 0.0025603987742215395,
+      "epoch": 27.347122302158272,
+      "step": 30410
+    },
+    {
+      "epoch": 27.347122302158272,
+      "step": 30410,
+      "torque_loss": 0.13627688586711884
+    },
+    {
+      "epoch": 27.347122302158272,
+      "force_loss": 0.005227160174399614,
+      "step": 30410
+    },
+    {
+      "epoch": 27.35611510791367,
+      "grad_norm": 0.18527504801750183,
+      "learning_rate": 5.297723930107855e-05,
+      "loss": 0.0181,
+      "step": 30420
+    },
+    {
+      "action_loss": 0.008046342991292477,
+      "epoch": 27.35611510791367,
+      "step": 30420
+    },
+    {
+      "epoch": 27.35611510791367,
+      "step": 30420,
+      "torque_loss": 0.11432340741157532
+    },
+    {
+      "epoch": 27.35611510791367,
+      "force_loss": 0.01383527647703886,
+      "step": 30420
+    },
+    {
+      "epoch": 27.365107913669064,
+      "grad_norm": 0.17165902256965637,
+      "learning_rate": 5.294972991762167e-05,
+      "loss": 0.0186,
+      "step": 30430
+    },
+    {
+      "action_loss": 0.00251086731441319,
+      "epoch": 27.365107913669064,
+      "step": 30430
+    },
+    {
+      "epoch": 27.365107913669064,
+      "step": 30430,
+      "torque_loss": 0.13733571767807007
+    },
+    {
+      "epoch": 27.365107913669064,
+      "force_loss": 0.0028186349663883448,
+      "step": 30430
+    },
+    {
+      "epoch": 27.37410071942446,
+      "grad_norm": 0.1820177286863327,
+      "learning_rate": 5.292221963811472e-05,
+      "loss": 0.0161,
+      "step": 30440
+    },
+    {
+      "action_loss": 0.0028685119468718767,
+      "epoch": 27.37410071942446,
+      "step": 30440
+    },
+    {
+      "epoch": 27.37410071942446,
+      "step": 30440,
+      "torque_loss": 0.11441189050674438
+    },
+    {
+      "epoch": 27.37410071942446,
+      "force_loss": 0.00430305814370513,
+      "step": 30440
+    },
+    {
+      "epoch": 27.383093525179856,
+      "grad_norm": 0.2745693624019623,
+      "learning_rate": 5.28947084709146e-05,
+      "loss": 0.0183,
+      "step": 30450
+    },
+    {
+      "action_loss": 0.0054676649160683155,
+      "epoch": 27.383093525179856,
+      "step": 30450
+    },
+    {
+      "epoch": 27.383093525179856,
+      "step": 30450,
+      "torque_loss": 0.13324573636054993
+    },
+    {
+      "epoch": 27.383093525179856,
+      "force_loss": 0.003394828410819173,
+      "step": 30450
+    },
+    {
+      "epoch": 27.392086330935253,
+      "grad_norm": 0.41122880578041077,
+      "learning_rate": 5.2867196424378465e-05,
+      "loss": 0.0198,
+      "step": 30460
+    },
+    {
+      "action_loss": 0.0030081484001129866,
+      "epoch": 27.392086330935253,
+      "step": 30460
+    },
+    {
+      "epoch": 27.392086330935253,
+      "step": 30460,
+      "torque_loss": 0.0999179258942604
+    },
+    {
+      "epoch": 27.392086330935253,
+      "force_loss": 0.004558890126645565,
+      "step": 30460
+    },
+    {
+      "epoch": 27.401079136690647,
+      "grad_norm": 0.6369983553886414,
+      "learning_rate": 5.2839683506863765e-05,
+      "loss": 0.0183,
+      "step": 30470
+    },
+    {
+      "action_loss": 0.0035806421656161547,
+      "epoch": 27.401079136690647,
+      "step": 30470
+    },
+    {
+      "epoch": 27.401079136690647,
+      "step": 30470,
+      "torque_loss": 0.17002622783184052
+    },
+    {
+      "epoch": 27.401079136690647,
+      "force_loss": 0.002485469914972782,
+      "step": 30470
+    },
+    {
+      "epoch": 27.41007194244604,
+      "grad_norm": 0.2785317301750183,
+      "learning_rate": 5.281216972672821e-05,
+      "loss": 0.0186,
+      "step": 30480
+    },
+    {
+      "action_loss": 0.007724123075604439,
+      "epoch": 27.41007194244604,
+      "step": 30480
+    },
+    {
+      "epoch": 27.41007194244604,
+      "step": 30480,
+      "torque_loss": 0.08777821063995361
+    },
+    {
+      "epoch": 27.41007194244604,
+      "force_loss": 0.006125069689005613,
+      "step": 30480
+    },
+    {
+      "epoch": 27.41906474820144,
+      "grad_norm": 0.6122022271156311,
+      "learning_rate": 5.278465509232973e-05,
+      "loss": 0.0241,
+      "step": 30490
+    },
+    {
+      "action_loss": 0.0026271052192896605,
+      "epoch": 27.41906474820144,
+      "step": 30490
+    },
+    {
+      "epoch": 27.41906474820144,
+      "step": 30490,
+      "torque_loss": 0.14399898052215576
+    },
+    {
+      "epoch": 27.41906474820144,
+      "force_loss": 0.0026529596652835608,
+      "step": 30490
+    },
+    {
+      "epoch": 27.428057553956833,
+      "grad_norm": 0.2580064833164215,
+      "learning_rate": 5.275713961202655e-05,
+      "loss": 0.0191,
+      "step": 30500
+    },
+    {
+      "action_loss": 0.006645339075475931,
+      "epoch": 27.428057553956833,
+      "step": 30500
+    },
+    {
+      "epoch": 27.428057553956833,
+      "step": 30500,
+      "torque_loss": 0.13289107382297516
+    },
+    {
+      "epoch": 27.428057553956833,
+      "force_loss": 0.006983119994401932,
+      "step": 30500
+    },
+    {
+      "epoch": 27.43705035971223,
+      "grad_norm": 0.21055203676223755,
+      "learning_rate": 5.2729623294177165e-05,
+      "loss": 0.0169,
+      "step": 30510
+    },
+    {
+      "action_loss": 0.0030639495234936476,
+      "epoch": 27.43705035971223,
+      "step": 30510
+    },
+    {
+      "epoch": 27.43705035971223,
+      "step": 30510,
+      "torque_loss": 0.095901720225811
+    },
+    {
+      "epoch": 27.43705035971223,
+      "force_loss": 0.00638936460018158,
+      "step": 30510
+    },
+    {
+      "epoch": 27.446043165467625,
+      "grad_norm": 0.2256162315607071,
+      "learning_rate": 5.270210614714028e-05,
+      "loss": 0.0186,
+      "step": 30520
+    },
+    {
+      "action_loss": 0.009638610295951366,
+      "epoch": 27.446043165467625,
+      "step": 30520
+    },
+    {
+      "epoch": 27.446043165467625,
+      "step": 30520,
+      "torque_loss": 0.14846175909042358
+    },
+    {
+      "epoch": 27.446043165467625,
+      "force_loss": 0.007739686872810125,
+      "step": 30520
+    },
+    {
+      "epoch": 27.455035971223023,
+      "grad_norm": 0.10142328590154648,
+      "learning_rate": 5.267458817927491e-05,
+      "loss": 0.0201,
+      "step": 30530
+    },
+    {
+      "action_loss": 0.021939754486083984,
+      "epoch": 27.455035971223023,
+      "step": 30530
+    },
+    {
+      "epoch": 27.455035971223023,
+      "step": 30530,
+      "torque_loss": 0.14862318336963654
+    },
+    {
+      "epoch": 27.455035971223023,
+      "force_loss": 0.03013339452445507,
+      "step": 30530
+    },
+    {
+      "epoch": 27.464028776978417,
+      "grad_norm": 0.20716537535190582,
+      "learning_rate": 5.264706939894026e-05,
+      "loss": 0.0214,
+      "step": 30540
+    },
+    {
+      "action_loss": 0.00564810074865818,
+      "epoch": 27.464028776978417,
+      "step": 30540
+    },
+    {
+      "epoch": 27.464028776978417,
+      "step": 30540,
+      "torque_loss": 0.11731040477752686
+    },
+    {
+      "epoch": 27.464028776978417,
+      "force_loss": 0.004977627191692591,
+      "step": 30540
+    },
+    {
+      "epoch": 27.473021582733814,
+      "grad_norm": 0.1534290909767151,
+      "learning_rate": 5.261954981449584e-05,
+      "loss": 0.0188,
+      "step": 30550
+    },
+    {
+      "action_loss": 0.002763771452009678,
+      "epoch": 27.473021582733814,
+      "step": 30550
+    },
+    {
+      "epoch": 27.473021582733814,
+      "step": 30550,
+      "torque_loss": 0.09688911586999893
+    },
+    {
+      "epoch": 27.473021582733814,
+      "force_loss": 0.0027837317902594805,
+      "step": 30550
+    },
+    {
+      "epoch": 27.48201438848921,
+      "grad_norm": 0.3309992849826813,
+      "learning_rate": 5.2592029434301324e-05,
+      "loss": 0.0183,
+      "step": 30560
+    },
+    {
+      "action_loss": 0.012967114336788654,
+      "epoch": 27.48201438848921,
+      "step": 30560
+    },
+    {
+      "epoch": 27.48201438848921,
+      "step": 30560,
+      "torque_loss": 0.1345806121826172
+    },
+    {
+      "epoch": 27.48201438848921,
+      "force_loss": 0.00974154844880104,
+      "step": 30560
+    },
+    {
+      "epoch": 27.491007194244606,
+      "grad_norm": 0.329872190952301,
+      "learning_rate": 5.256450826671672e-05,
+      "loss": 0.0241,
+      "step": 30570
+    },
+    {
+      "action_loss": 0.006892070639878511,
+      "epoch": 27.491007194244606,
+      "step": 30570
+    },
+    {
+      "epoch": 27.491007194244606,
+      "step": 30570,
+      "torque_loss": 0.12427874654531479
+    },
+    {
+      "epoch": 27.491007194244606,
+      "force_loss": 0.00313993520103395,
+      "step": 30570
+    },
+    {
+      "epoch": 27.5,
+      "grad_norm": 0.2719530165195465,
+      "learning_rate": 5.253698632010221e-05,
+      "loss": 0.0201,
+      "step": 30580
+    },
+    {
+      "action_loss": 0.0027273735031485558,
+      "epoch": 27.5,
+      "step": 30580
+    },
+    {
+      "epoch": 27.5,
+      "step": 30580,
+      "torque_loss": 0.15049169957637787
+    },
+    {
+      "epoch": 27.5,
+      "force_loss": 0.001852459623478353,
+      "step": 30580
+    },
+    {
+      "epoch": 27.508992805755394,
+      "grad_norm": 0.2190684825181961,
+      "learning_rate": 5.2509463602818246e-05,
+      "loss": 0.0198,
+      "step": 30590
+    },
+    {
+      "action_loss": 0.006629563868045807,
+      "epoch": 27.508992805755394,
+      "step": 30590
+    },
+    {
+      "epoch": 27.508992805755394,
+      "step": 30590,
+      "torque_loss": 0.14348743855953217
+    },
+    {
+      "epoch": 27.508992805755394,
+      "force_loss": 0.010787595994770527,
+      "step": 30590
+    },
+    {
+      "epoch": 27.51798561151079,
+      "grad_norm": 0.38350728154182434,
+      "learning_rate": 5.248194012322549e-05,
+      "loss": 0.0178,
+      "step": 30600
+    },
+    {
+      "action_loss": 0.01122139673680067,
+      "epoch": 27.51798561151079,
+      "step": 30600
+    },
+    {
+      "epoch": 27.51798561151079,
+      "step": 30600,
+      "torque_loss": 0.10846803337335587
+    },
+    {
+      "epoch": 27.51798561151079,
+      "force_loss": 0.011705628596246243,
+      "step": 30600
+    },
+    {
+      "epoch": 27.526978417266186,
+      "grad_norm": 0.3710283935070038,
+      "learning_rate": 5.245441588968486e-05,
+      "loss": 0.019,
+      "step": 30610
+    },
+    {
+      "action_loss": 0.00363993551582098,
+      "epoch": 27.526978417266186,
+      "step": 30610
+    },
+    {
+      "epoch": 27.526978417266186,
+      "step": 30610,
+      "torque_loss": 0.11770457029342651
+    },
+    {
+      "epoch": 27.526978417266186,
+      "force_loss": 0.006347086746245623,
+      "step": 30610
+    },
+    {
+      "epoch": 27.535971223021583,
+      "grad_norm": 0.4064185321331024,
+      "learning_rate": 5.242689091055748e-05,
+      "loss": 0.0212,
+      "step": 30620
+    },
+    {
+      "action_loss": 0.0039041517302393913,
+      "epoch": 27.535971223021583,
+      "step": 30620
+    },
+    {
+      "epoch": 27.535971223021583,
+      "step": 30620,
+      "torque_loss": 0.18442851305007935
+    },
+    {
+      "epoch": 27.535971223021583,
+      "force_loss": 0.00466783344745636,
+      "step": 30620
+    },
+    {
+      "epoch": 27.544964028776977,
+      "grad_norm": 0.38118013739585876,
+      "learning_rate": 5.239936519420473e-05,
+      "loss": 0.0255,
+      "step": 30630
+    },
+    {
+      "action_loss": 0.0021402251441031694,
+      "epoch": 27.544964028776977,
+      "step": 30630
+    },
+    {
+      "epoch": 27.544964028776977,
+      "step": 30630,
+      "torque_loss": 0.11720553785562515
+    },
+    {
+      "epoch": 27.544964028776977,
+      "force_loss": 0.002472731051966548,
+      "step": 30630
+    },
+    {
+      "epoch": 27.553956834532375,
+      "grad_norm": 0.34569528698921204,
+      "learning_rate": 5.2371838748988175e-05,
+      "loss": 0.0221,
+      "step": 30640
+    },
+    {
+      "action_loss": 0.005891752894967794,
+      "epoch": 27.553956834532375,
+      "step": 30640
+    },
+    {
+      "epoch": 27.553956834532375,
+      "step": 30640,
+      "torque_loss": 0.11999966949224472
+    },
+    {
+      "epoch": 27.553956834532375,
+      "force_loss": 0.007090532686561346,
+      "step": 30640
+    },
+    {
+      "epoch": 27.56294964028777,
+      "grad_norm": 0.26249584555625916,
+      "learning_rate": 5.234431158326965e-05,
+      "loss": 0.0224,
+      "step": 30650
+    },
+    {
+      "action_loss": 0.00848119705915451,
+      "epoch": 27.56294964028777,
+      "step": 30650
+    },
+    {
+      "epoch": 27.56294964028777,
+      "step": 30650,
+      "torque_loss": 0.15211278200149536
+    },
+    {
+      "epoch": 27.56294964028777,
+      "force_loss": 0.011708579957485199,
+      "step": 30650
+    },
+    {
+      "epoch": 27.571942446043167,
+      "grad_norm": 0.233008474111557,
+      "learning_rate": 5.231678370541115e-05,
+      "loss": 0.0201,
+      "step": 30660
+    },
+    {
+      "action_loss": 0.01889168657362461,
+      "epoch": 27.571942446043167,
+      "step": 30660
+    },
+    {
+      "epoch": 27.571942446043167,
+      "step": 30660,
+      "torque_loss": 0.1622740626335144
+    },
+    {
+      "epoch": 27.571942446043167,
+      "force_loss": 0.031231537461280823,
+      "step": 30660
+    },
+    {
+      "epoch": 27.58093525179856,
+      "grad_norm": 0.1595122516155243,
+      "learning_rate": 5.228925512377495e-05,
+      "loss": 0.0198,
+      "step": 30670
+    },
+    {
+      "action_loss": 0.00274129887111485,
+      "epoch": 27.58093525179856,
+      "step": 30670
+    },
+    {
+      "epoch": 27.58093525179856,
+      "step": 30670,
+      "torque_loss": 0.17599289119243622
+    },
+    {
+      "epoch": 27.58093525179856,
+      "force_loss": 0.001764418208040297,
+      "step": 30670
+    },
+    {
+      "epoch": 27.58992805755396,
+      "grad_norm": 0.19278109073638916,
+      "learning_rate": 5.2261725846723465e-05,
+      "loss": 0.0185,
+      "step": 30680
+    },
+    {
+      "action_loss": 0.008990000002086163,
+      "epoch": 27.58992805755396,
+      "step": 30680
+    },
+    {
+      "epoch": 27.58992805755396,
+      "step": 30680,
+      "torque_loss": 0.1137491837143898
+    },
+    {
+      "epoch": 27.58992805755396,
+      "force_loss": 0.014333575963973999,
+      "step": 30680
+    },
+    {
+      "epoch": 27.598920863309353,
+      "grad_norm": 0.16543430089950562,
+      "learning_rate": 5.22341958826194e-05,
+      "loss": 0.0192,
+      "step": 30690
+    },
+    {
+      "action_loss": 0.007123140152543783,
+      "epoch": 27.598920863309353,
+      "step": 30690
+    },
+    {
+      "epoch": 27.598920863309353,
+      "step": 30690,
+      "torque_loss": 0.11266070604324341
+    },
+    {
+      "epoch": 27.598920863309353,
+      "force_loss": 0.004885811358690262,
+      "step": 30690
+    },
+    {
+      "epoch": 27.607913669064747,
+      "grad_norm": 0.16478745639324188,
+      "learning_rate": 5.22066652398256e-05,
+      "loss": 0.0175,
+      "step": 30700
+    },
+    {
+      "action_loss": 0.003914508502930403,
+      "epoch": 27.607913669064747,
+      "step": 30700
+    },
+    {
+      "epoch": 27.607913669064747,
+      "step": 30700,
+      "torque_loss": 0.10943523794412613
+    },
+    {
+      "epoch": 27.607913669064747,
+      "force_loss": 0.005109761841595173,
+      "step": 30700
+    },
+    {
+      "epoch": 27.616906474820144,
+      "grad_norm": 0.26371338963508606,
+      "learning_rate": 5.2179133926705185e-05,
+      "loss": 0.0169,
+      "step": 30710
+    },
+    {
+      "action_loss": 0.006401343736797571,
+      "epoch": 27.616906474820144,
+      "step": 30710
+    },
+    {
+      "epoch": 27.616906474820144,
+      "step": 30710,
+      "torque_loss": 0.1256575733423233
+    },
+    {
+      "epoch": 27.616906474820144,
+      "force_loss": 0.003221782622858882,
+      "step": 30710
+    },
+    {
+      "epoch": 27.62589928057554,
+      "grad_norm": 0.13252998888492584,
+      "learning_rate": 5.215160195162141e-05,
+      "loss": 0.0186,
+      "step": 30720
+    },
+    {
+      "action_loss": 0.0028024381026625633,
+      "epoch": 27.62589928057554,
+      "step": 30720
+    },
+    {
+      "epoch": 27.62589928057554,
+      "step": 30720,
+      "torque_loss": 0.10049884766340256
+    },
+    {
+      "epoch": 27.62589928057554,
+      "force_loss": 0.004064956679940224,
+      "step": 30720
+    },
+    {
+      "epoch": 27.634892086330936,
+      "grad_norm": 0.18607401847839355,
+      "learning_rate": 5.212406932293776e-05,
+      "loss": 0.0165,
+      "step": 30730
+    },
+    {
+      "action_loss": 0.003135338658466935,
+      "epoch": 27.634892086330936,
+      "step": 30730
+    },
+    {
+      "epoch": 27.634892086330936,
+      "step": 30730,
+      "torque_loss": 0.12078125029802322
+    },
+    {
+      "epoch": 27.634892086330936,
+      "force_loss": 0.002940795151516795,
+      "step": 30730
+    },
+    {
+      "epoch": 27.64388489208633,
+      "grad_norm": 0.13245603442192078,
+      "learning_rate": 5.209653604901795e-05,
+      "loss": 0.0178,
+      "step": 30740
+    },
+    {
+      "action_loss": 0.005995666608214378,
+      "epoch": 27.64388489208633,
+      "step": 30740
+    },
+    {
+      "epoch": 27.64388489208633,
+      "step": 30740,
+      "torque_loss": 0.24334824085235596
+    },
+    {
+      "epoch": 27.64388489208633,
+      "force_loss": 0.00946358498185873,
+      "step": 30740
+    },
+    {
+      "epoch": 27.652877697841728,
+      "grad_norm": 0.28509190678596497,
+      "learning_rate": 5.206900213822584e-05,
+      "loss": 0.0184,
+      "step": 30750
+    },
+    {
+      "action_loss": 0.009454579092562199,
+      "epoch": 27.652877697841728,
+      "step": 30750
+    },
+    {
+      "epoch": 27.652877697841728,
+      "step": 30750,
+      "torque_loss": 0.1258021593093872
+    },
+    {
+      "epoch": 27.652877697841728,
+      "force_loss": 0.0027389039751142263,
+      "step": 30750
+    },
+    {
+      "epoch": 27.66187050359712,
+      "grad_norm": 0.152934268116951,
+      "learning_rate": 5.204146759892551e-05,
+      "loss": 0.0201,
+      "step": 30760
+    },
+    {
+      "action_loss": 0.001829686458222568,
+      "epoch": 27.66187050359712,
+      "step": 30760
+    },
+    {
+      "epoch": 27.66187050359712,
+      "step": 30760,
+      "torque_loss": 0.10581382364034653
+    },
+    {
+      "epoch": 27.66187050359712,
+      "force_loss": 0.001624145545065403,
+      "step": 30760
+    },
+    {
+      "epoch": 27.67086330935252,
+      "grad_norm": 0.2874409854412079,
+      "learning_rate": 5.2013932439481216e-05,
+      "loss": 0.0186,
+      "step": 30770
+    },
+    {
+      "action_loss": 0.01221607904881239,
+      "epoch": 27.67086330935252,
+      "step": 30770
+    },
+    {
+      "epoch": 27.67086330935252,
+      "step": 30770,
+      "torque_loss": 0.18106383085250854
+    },
+    {
+      "epoch": 27.67086330935252,
+      "force_loss": 0.01274307444691658,
+      "step": 30770
+    },
+    {
+      "epoch": 27.679856115107913,
+      "grad_norm": 0.241510808467865,
+      "learning_rate": 5.198639666825743e-05,
+      "loss": 0.0267,
+      "step": 30780
+    },
+    {
+      "action_loss": 0.006198238581418991,
+      "epoch": 27.679856115107913,
+      "step": 30780
+    },
+    {
+      "epoch": 27.679856115107913,
+      "step": 30780,
+      "torque_loss": 0.10350052267313004
+    },
+    {
+      "epoch": 27.679856115107913,
+      "force_loss": 0.0062993853352963924,
+      "step": 30780
+    },
+    {
+      "epoch": 27.68884892086331,
+      "grad_norm": 0.16487713158130646,
+      "learning_rate": 5.195886029361877e-05,
+      "loss": 0.019,
+      "step": 30790
+    },
+    {
+      "action_loss": 0.009002833627164364,
+      "epoch": 27.68884892086331,
+      "step": 30790
+    },
+    {
+      "epoch": 27.68884892086331,
+      "step": 30790,
+      "torque_loss": 0.15934205055236816
+    },
+    {
+      "epoch": 27.68884892086331,
+      "force_loss": 0.006100728642195463,
+      "step": 30790
+    },
+    {
+      "epoch": 27.697841726618705,
+      "grad_norm": 0.5091961622238159,
+      "learning_rate": 5.193132332393009e-05,
+      "loss": 0.0231,
+      "step": 30800
+    },
+    {
+      "action_loss": 0.004027887247502804,
+      "epoch": 27.697841726618705,
+      "step": 30800
+    },
+    {
+      "epoch": 27.697841726618705,
+      "step": 30800,
+      "torque_loss": 0.17949558794498444
+    },
+    {
+      "epoch": 27.697841726618705,
+      "force_loss": 0.0030570297967642546,
+      "step": 30800
+    },
+    {
+      "epoch": 27.7068345323741,
+      "grad_norm": 0.3422536551952362,
+      "learning_rate": 5.1903785767556376e-05,
+      "loss": 0.0201,
+      "step": 30810
+    },
+    {
+      "action_loss": 0.004884196911007166,
+      "epoch": 27.7068345323741,
+      "step": 30810
+    },
+    {
+      "epoch": 27.7068345323741,
+      "step": 30810,
+      "torque_loss": 0.07200178503990173
+    },
+    {
+      "epoch": 27.7068345323741,
+      "force_loss": 0.006248340476304293,
+      "step": 30810
+    },
+    {
+      "epoch": 27.715827338129497,
+      "grad_norm": 0.1799556463956833,
+      "learning_rate": 5.187624763286282e-05,
+      "loss": 0.0164,
+      "step": 30820
+    },
+    {
+      "action_loss": 0.008699198253452778,
+      "epoch": 27.715827338129497,
+      "step": 30820
+    },
+    {
+      "epoch": 27.715827338129497,
+      "step": 30820,
+      "torque_loss": 0.10013705492019653
+    },
+    {
+      "epoch": 27.715827338129497,
+      "force_loss": 0.007091708946973085,
+      "step": 30820
+    },
+    {
+      "epoch": 27.72482014388489,
+      "grad_norm": 0.6433326005935669,
+      "learning_rate": 5.184870892821475e-05,
+      "loss": 0.0243,
+      "step": 30830
+    },
+    {
+      "action_loss": 0.007349391933530569,
+      "epoch": 27.72482014388489,
+      "step": 30830
+    },
+    {
+      "epoch": 27.72482014388489,
+      "step": 30830,
+      "torque_loss": 0.1412038803100586
+    },
+    {
+      "epoch": 27.72482014388489,
+      "force_loss": 0.005403654184192419,
+      "step": 30830
+    },
+    {
+      "epoch": 27.73381294964029,
+      "grad_norm": 0.13341479003429413,
+      "learning_rate": 5.182116966197773e-05,
+      "loss": 0.0191,
+      "step": 30840
+    },
+    {
+      "action_loss": 0.005824791267514229,
+      "epoch": 27.73381294964029,
+      "step": 30840
+    },
+    {
+      "epoch": 27.73381294964029,
+      "step": 30840,
+      "torque_loss": 0.1492571383714676
+    },
+    {
+      "epoch": 27.73381294964029,
+      "force_loss": 0.006928595248609781,
+      "step": 30840
+    },
+    {
+      "epoch": 27.742805755395683,
+      "grad_norm": 0.1712639182806015,
+      "learning_rate": 5.1793629842517466e-05,
+      "loss": 0.0184,
+      "step": 30850
+    },
+    {
+      "action_loss": 0.004385871347039938,
+      "epoch": 27.742805755395683,
+      "step": 30850
+    },
+    {
+      "epoch": 27.742805755395683,
+      "step": 30850,
+      "torque_loss": 0.17530196905136108
+    },
+    {
+      "epoch": 27.742805755395683,
+      "force_loss": 0.002255291910842061,
+      "step": 30850
+    },
+    {
+      "epoch": 27.75179856115108,
+      "grad_norm": 0.15210753679275513,
+      "learning_rate": 5.17660894781998e-05,
+      "loss": 0.0184,
+      "step": 30860
+    },
+    {
+      "action_loss": 0.0028193567413836718,
+      "epoch": 27.75179856115108,
+      "step": 30860
+    },
+    {
+      "epoch": 27.75179856115108,
+      "step": 30860,
+      "torque_loss": 0.11524689197540283
+    },
+    {
+      "epoch": 27.75179856115108,
+      "force_loss": 0.0037801319267600775,
+      "step": 30860
+    },
+    {
+      "epoch": 27.760791366906474,
+      "grad_norm": 0.41138550639152527,
+      "learning_rate": 5.173854857739079e-05,
+      "loss": 0.0189,
+      "step": 30870
+    },
+    {
+      "action_loss": 0.004349606577306986,
+      "epoch": 27.760791366906474,
+      "step": 30870
+    },
+    {
+      "epoch": 27.760791366906474,
+      "step": 30870,
+      "torque_loss": 0.1613604575395584
+    },
+    {
+      "epoch": 27.760791366906474,
+      "force_loss": 0.005117138847708702,
+      "step": 30870
+    },
+    {
+      "epoch": 27.769784172661872,
+      "grad_norm": 0.35171759128570557,
+      "learning_rate": 5.171100714845661e-05,
+      "loss": 0.0179,
+      "step": 30880
+    },
+    {
+      "action_loss": 0.010870051570236683,
+      "epoch": 27.769784172661872,
+      "step": 30880
+    },
+    {
+      "epoch": 27.769784172661872,
+      "step": 30880,
+      "torque_loss": 0.1550101935863495
+    },
+    {
+      "epoch": 27.769784172661872,
+      "force_loss": 0.013455321080982685,
+      "step": 30880
+    },
+    {
+      "epoch": 27.778776978417266,
+      "grad_norm": 0.3980942964553833,
+      "learning_rate": 5.1683465199763646e-05,
+      "loss": 0.0211,
+      "step": 30890
+    },
+    {
+      "action_loss": 0.002072709845378995,
+      "epoch": 27.778776978417266,
+      "step": 30890
+    },
+    {
+      "epoch": 27.778776978417266,
+      "step": 30890,
+      "torque_loss": 0.1543998122215271
+    },
+    {
+      "epoch": 27.778776978417266,
+      "force_loss": 0.0012014691019430757,
+      "step": 30890
+    },
+    {
+      "epoch": 27.78776978417266,
+      "grad_norm": 0.21990813314914703,
+      "learning_rate": 5.16559227396784e-05,
+      "loss": 0.0171,
+      "step": 30900
+    },
+    {
+      "action_loss": 0.005485077854245901,
+      "epoch": 27.78776978417266,
+      "step": 30900
+    },
+    {
+      "epoch": 27.78776978417266,
+      "step": 30900,
+      "torque_loss": 0.13829806447029114
+    },
+    {
+      "epoch": 27.78776978417266,
+      "force_loss": 0.00719534931704402,
+      "step": 30900
+    },
+    {
+      "epoch": 27.796762589928058,
+      "grad_norm": 0.21184466779232025,
+      "learning_rate": 5.1628379776567556e-05,
+      "loss": 0.0196,
+      "step": 30910
+    },
+    {
+      "action_loss": 0.0020861951634287834,
+      "epoch": 27.796762589928058,
+      "step": 30910
+    },
+    {
+      "epoch": 27.796762589928058,
+      "step": 30910,
+      "torque_loss": 0.08997714519500732
+    },
+    {
+      "epoch": 27.796762589928058,
+      "force_loss": 0.0049504549242556095,
+      "step": 30910
+    },
+    {
+      "epoch": 27.805755395683452,
+      "grad_norm": 0.3364858627319336,
+      "learning_rate": 5.160083631879792e-05,
+      "loss": 0.0186,
+      "step": 30920
+    },
+    {
+      "action_loss": 0.004167437087744474,
+      "epoch": 27.805755395683452,
+      "step": 30920
+    },
+    {
+      "epoch": 27.805755395683452,
+      "step": 30920,
+      "torque_loss": 0.0762174054980278
+    },
+    {
+      "epoch": 27.805755395683452,
+      "force_loss": 0.010606507770717144,
+      "step": 30920
+    },
+    {
+      "epoch": 27.81474820143885,
+      "grad_norm": 0.28972136974334717,
+      "learning_rate": 5.1573292374736484e-05,
+      "loss": 0.0178,
+      "step": 30930
+    },
+    {
+      "action_loss": 0.0051741148345172405,
+      "epoch": 27.81474820143885,
+      "step": 30930
+    },
+    {
+      "epoch": 27.81474820143885,
+      "step": 30930,
+      "torque_loss": 0.16094911098480225
+    },
+    {
+      "epoch": 27.81474820143885,
+      "force_loss": 0.002703087404370308,
+      "step": 30930
+    },
+    {
+      "epoch": 27.823741007194243,
+      "grad_norm": 0.32346779108047485,
+      "learning_rate": 5.1545747952750356e-05,
+      "loss": 0.0203,
+      "step": 30940
+    },
+    {
+      "action_loss": 0.0021596148144453764,
+      "epoch": 27.823741007194243,
+      "step": 30940
+    },
+    {
+      "epoch": 27.823741007194243,
+      "step": 30940,
+      "torque_loss": 0.10066640377044678
+    },
+    {
+      "epoch": 27.823741007194243,
+      "force_loss": 0.006780712399631739,
+      "step": 30940
+    },
+    {
+      "epoch": 27.83273381294964,
+      "grad_norm": 0.22288697957992554,
+      "learning_rate": 5.151820306120682e-05,
+      "loss": 0.0164,
+      "step": 30950
+    },
+    {
+      "action_loss": 0.002462455304339528,
+      "epoch": 27.83273381294964,
+      "step": 30950
+    },
+    {
+      "epoch": 27.83273381294964,
+      "step": 30950,
+      "torque_loss": 0.11628085374832153
+    },
+    {
+      "epoch": 27.83273381294964,
+      "force_loss": 0.0019165457924827933,
+      "step": 30950
+    },
+    {
+      "epoch": 27.841726618705035,
+      "grad_norm": 0.20578458905220032,
+      "learning_rate": 5.149065770847328e-05,
+      "loss": 0.0176,
+      "step": 30960
+    },
+    {
+      "action_loss": 0.00621893210336566,
+      "epoch": 27.841726618705035,
+      "step": 30960
+    },
+    {
+      "epoch": 27.841726618705035,
+      "step": 30960,
+      "torque_loss": 0.1364258974790573
+    },
+    {
+      "epoch": 27.841726618705035,
+      "force_loss": 0.00447434326633811,
+      "step": 30960
+    },
+    {
+      "epoch": 27.850719424460433,
+      "grad_norm": 0.3081282079219818,
+      "learning_rate": 5.1463111902917297e-05,
+      "loss": 0.021,
+      "step": 30970
+    },
+    {
+      "action_loss": 0.004593221936374903,
+      "epoch": 27.850719424460433,
+      "step": 30970
+    },
+    {
+      "epoch": 27.850719424460433,
+      "step": 30970,
+      "torque_loss": 0.16942958533763885
+    },
+    {
+      "epoch": 27.850719424460433,
+      "force_loss": 0.004925244953483343,
+      "step": 30970
+    },
+    {
+      "epoch": 27.859712230215827,
+      "grad_norm": 0.2719041705131531,
+      "learning_rate": 5.143556565290654e-05,
+      "loss": 0.0194,
+      "step": 30980
+    },
+    {
+      "action_loss": 0.007191583979874849,
+      "epoch": 27.859712230215827,
+      "step": 30980
+    },
+    {
+      "epoch": 27.859712230215827,
+      "step": 30980,
+      "torque_loss": 0.15952707827091217
+    },
+    {
+      "epoch": 27.859712230215827,
+      "force_loss": 0.007490354124456644,
+      "step": 30980
+    },
+    {
+      "epoch": 27.868705035971225,
+      "grad_norm": 0.4152081608772278,
+      "learning_rate": 5.140801896680882e-05,
+      "loss": 0.0184,
+      "step": 30990
+    },
+    {
+      "action_loss": 0.007137230131775141,
+      "epoch": 27.868705035971225,
+      "step": 30990
+    },
+    {
+      "epoch": 27.868705035971225,
+      "step": 30990,
+      "torque_loss": 0.14315825700759888
+    },
+    {
+      "epoch": 27.868705035971225,
+      "force_loss": 0.007594504859298468,
+      "step": 30990
+    },
+    {
+      "epoch": 27.87769784172662,
+      "grad_norm": 0.15846173465251923,
+      "learning_rate": 5.1380471852992144e-05,
+      "loss": 0.0186,
+      "step": 31000
+    },
+    {
+      "action_loss": 0.004153303802013397,
+      "epoch": 27.87769784172662,
+      "step": 31000
+    },
+    {
+      "epoch": 27.87769784172662,
+      "step": 31000,
+      "torque_loss": 0.1316547840833664
+    },
+    {
+      "epoch": 27.87769784172662,
+      "force_loss": 0.00436300178989768,
+      "step": 31000
+    },
+    {
+      "epoch": 27.886690647482013,
+      "grad_norm": 0.22018012404441833,
+      "learning_rate": 5.135292431982457e-05,
+      "loss": 0.0185,
+      "step": 31010
+    },
+    {
+      "action_loss": 0.002149861538782716,
+      "epoch": 27.886690647482013,
+      "step": 31010
+    },
+    {
+      "epoch": 27.886690647482013,
+      "step": 31010,
+      "torque_loss": 0.12831994891166687
+    },
+    {
+      "epoch": 27.886690647482013,
+      "force_loss": 0.004885531030595303,
+      "step": 31010
+    },
+    {
+      "epoch": 27.89568345323741,
+      "grad_norm": 0.2307564616203308,
+      "learning_rate": 5.1325376375674294e-05,
+      "loss": 0.0175,
+      "step": 31020
+    },
+    {
+      "action_loss": 0.010385863482952118,
+      "epoch": 27.89568345323741,
+      "step": 31020
+    },
+    {
+      "epoch": 27.89568345323741,
+      "step": 31020,
+      "torque_loss": 0.16394855082035065
+    },
+    {
+      "epoch": 27.89568345323741,
+      "force_loss": 0.017414994537830353,
+      "step": 31020
+    },
+    {
+      "epoch": 27.904676258992804,
+      "grad_norm": 0.161410853266716,
+      "learning_rate": 5.129782802890968e-05,
+      "loss": 0.0199,
+      "step": 31030
+    },
+    {
+      "action_loss": 0.004073642659932375,
+      "epoch": 27.904676258992804,
+      "step": 31030
+    },
+    {
+      "epoch": 27.904676258992804,
+      "step": 31030,
+      "torque_loss": 0.15164317190647125
+    },
+    {
+      "epoch": 27.904676258992804,
+      "force_loss": 0.0034698061645030975,
+      "step": 31030
+    },
+    {
+      "epoch": 27.913669064748202,
+      "grad_norm": 0.2141055315732956,
+      "learning_rate": 5.127027928789916e-05,
+      "loss": 0.0181,
+      "step": 31040
+    },
+    {
+      "action_loss": 0.0021478633861988783,
+      "epoch": 27.913669064748202,
+      "step": 31040
+    },
+    {
+      "epoch": 27.913669064748202,
+      "step": 31040,
+      "torque_loss": 0.11276334524154663
+    },
+    {
+      "epoch": 27.913669064748202,
+      "force_loss": 0.0028508733958005905,
+      "step": 31040
+    },
+    {
+      "epoch": 27.922661870503596,
+      "grad_norm": 0.1585572212934494,
+      "learning_rate": 5.124273016101135e-05,
+      "loss": 0.028,
+      "step": 31050
+    },
+    {
+      "action_loss": 0.005104747600853443,
+      "epoch": 27.922661870503596,
+      "step": 31050
+    },
+    {
+      "epoch": 27.922661870503596,
+      "step": 31050,
+      "torque_loss": 0.15689681470394135
+    },
+    {
+      "epoch": 27.922661870503596,
+      "force_loss": 0.003291598754003644,
+      "step": 31050
+    },
+    {
+      "epoch": 27.931654676258994,
+      "grad_norm": 0.20900021493434906,
+      "learning_rate": 5.121518065661492e-05,
+      "loss": 0.0198,
+      "step": 31060
+    },
+    {
+      "action_loss": 0.0021802065894007683,
+      "epoch": 27.931654676258994,
+      "step": 31060
+    },
+    {
+      "epoch": 27.931654676258994,
+      "step": 31060,
+      "torque_loss": 0.1086922362446785
+    },
+    {
+      "epoch": 27.931654676258994,
+      "force_loss": 0.0023348331451416016,
+      "step": 31060
+    },
+    {
+      "epoch": 27.940647482014388,
+      "grad_norm": 0.25486835837364197,
+      "learning_rate": 5.11876307830787e-05,
+      "loss": 0.0176,
+      "step": 31070
+    },
+    {
+      "action_loss": 0.010240408591926098,
+      "epoch": 27.940647482014388,
+      "step": 31070
+    },
+    {
+      "epoch": 27.940647482014388,
+      "step": 31070,
+      "torque_loss": 0.12701404094696045
+    },
+    {
+      "epoch": 27.940647482014388,
+      "force_loss": 0.006227945443242788,
+      "step": 31070
+    },
+    {
+      "epoch": 27.949640287769785,
+      "grad_norm": 0.21160483360290527,
+      "learning_rate": 5.1160080548771596e-05,
+      "loss": 0.022,
+      "step": 31080
+    },
+    {
+      "action_loss": 0.0023566593881696463,
+      "epoch": 27.949640287769785,
+      "step": 31080
+    },
+    {
+      "epoch": 27.949640287769785,
+      "step": 31080,
+      "torque_loss": 0.14099116623401642
+    },
+    {
+      "epoch": 27.949640287769785,
+      "force_loss": 0.0016366178169846535,
+      "step": 31080
+    },
+    {
+      "epoch": 27.95863309352518,
+      "grad_norm": 0.5432224869728088,
+      "learning_rate": 5.1132529962062656e-05,
+      "loss": 0.019,
+      "step": 31090
+    },
+    {
+      "action_loss": 0.002328429138287902,
+      "epoch": 27.95863309352518,
+      "step": 31090
+    },
+    {
+      "epoch": 27.95863309352518,
+      "step": 31090,
+      "torque_loss": 0.15630583465099335
+    },
+    {
+      "epoch": 27.95863309352518,
+      "force_loss": 0.001649947720579803,
+      "step": 31090
+    },
+    {
+      "epoch": 27.967625899280577,
+      "grad_norm": 0.2914307415485382,
+      "learning_rate": 5.110497903132101e-05,
+      "loss": 0.0198,
+      "step": 31100
+    },
+    {
+      "action_loss": 0.0015938254073262215,
+      "epoch": 27.967625899280577,
+      "step": 31100
+    },
+    {
+      "epoch": 27.967625899280577,
+      "step": 31100,
+      "torque_loss": 0.0902237519621849
+    },
+    {
+      "epoch": 27.967625899280577,
+      "force_loss": 0.0016521824290975928,
+      "step": 31100
+    },
+    {
+      "epoch": 27.97661870503597,
+      "grad_norm": 0.2809683680534363,
+      "learning_rate": 5.107742776491592e-05,
+      "loss": 0.0191,
+      "step": 31110
+    },
+    {
+      "action_loss": 0.0029244229663163424,
+      "epoch": 27.97661870503597,
+      "step": 31110
+    },
+    {
+      "epoch": 27.97661870503597,
+      "step": 31110,
+      "torque_loss": 0.13668620586395264
+    },
+    {
+      "epoch": 27.97661870503597,
+      "force_loss": 0.002639237092807889,
+      "step": 31110
+    },
+    {
+      "epoch": 27.985611510791365,
+      "grad_norm": 0.16187016665935516,
+      "learning_rate": 5.104987617121673e-05,
+      "loss": 0.0189,
+      "step": 31120
+    },
+    {
+      "action_loss": 0.005097375717014074,
+      "epoch": 27.985611510791365,
+      "step": 31120
+    },
+    {
+      "epoch": 27.985611510791365,
+      "step": 31120,
+      "torque_loss": 0.0963716208934784
+    },
+    {
+      "epoch": 27.985611510791365,
+      "force_loss": 0.008771772496402264,
+      "step": 31120
+    },
+    {
+      "epoch": 27.994604316546763,
+      "grad_norm": 0.19888824224472046,
+      "learning_rate": 5.102232425859287e-05,
+      "loss": 0.0183,
+      "step": 31130
+    },
+    {
+      "action_loss": 0.001765283290296793,
+      "epoch": 27.994604316546763,
+      "step": 31130
+    },
+    {
+      "epoch": 27.994604316546763,
+      "step": 31130,
+      "torque_loss": 0.1185249462723732
+    },
+    {
+      "epoch": 27.994604316546763,
+      "force_loss": 0.001336389104835689,
+      "step": 31130
+    },
+    {
+      "epoch": 28.003597122302157,
+      "grad_norm": 0.23520608246326447,
+      "learning_rate": 5.09947720354139e-05,
+      "loss": 0.0175,
+      "step": 31140
+    },
+    {
+      "action_loss": 0.003011924447491765,
+      "epoch": 28.003597122302157,
+      "step": 31140
+    },
+    {
+      "epoch": 28.003597122302157,
+      "step": 31140,
+      "torque_loss": 0.08293508738279343
+    },
+    {
+      "epoch": 28.003597122302157,
+      "force_loss": 0.0032558289822191,
+      "step": 31140
+    },
+    {
+      "epoch": 28.012589928057555,
+      "grad_norm": 0.3765396177768707,
+      "learning_rate": 5.096721951004942e-05,
+      "loss": 0.0187,
+      "step": 31150
+    },
+    {
+      "action_loss": 0.01665198989212513,
+      "epoch": 28.012589928057555,
+      "step": 31150
+    },
+    {
+      "epoch": 28.012589928057555,
+      "step": 31150,
+      "torque_loss": 0.10503220558166504
+    },
+    {
+      "epoch": 28.012589928057555,
+      "force_loss": 0.027542075142264366,
+      "step": 31150
+    },
+    {
+      "epoch": 28.02158273381295,
+      "grad_norm": 0.26802268624305725,
+      "learning_rate": 5.0939666690869227e-05,
+      "loss": 0.0198,
+      "step": 31160
+    },
+    {
+      "action_loss": 0.00382077693939209,
+      "epoch": 28.02158273381295,
+      "step": 31160
+    },
+    {
+      "epoch": 28.02158273381295,
+      "step": 31160,
+      "torque_loss": 0.12123435735702515
+    },
+    {
+      "epoch": 28.02158273381295,
+      "force_loss": 0.0025932935532182455,
+      "step": 31160
+    },
+    {
+      "epoch": 28.030575539568346,
+      "grad_norm": 0.3540857434272766,
+      "learning_rate": 5.0912113586243096e-05,
+      "loss": 0.0189,
+      "step": 31170
+    },
+    {
+      "action_loss": 0.0028544620145112276,
+      "epoch": 28.030575539568346,
+      "step": 31170
+    },
+    {
+      "epoch": 28.030575539568346,
+      "step": 31170,
+      "torque_loss": 0.12507446110248566
+    },
+    {
+      "epoch": 28.030575539568346,
+      "force_loss": 0.004324841778725386,
+      "step": 31170
+    },
+    {
+      "epoch": 28.03956834532374,
+      "grad_norm": 0.616055428981781,
+      "learning_rate": 5.0884560204540935e-05,
+      "loss": 0.0203,
+      "step": 31180
+    },
+    {
+      "action_loss": 0.011461793445050716,
+      "epoch": 28.03956834532374,
+      "step": 31180
+    },
+    {
+      "epoch": 28.03956834532374,
+      "step": 31180,
+      "torque_loss": 0.08699599653482437
+    },
+    {
+      "epoch": 28.03956834532374,
+      "force_loss": 0.014050401747226715,
+      "step": 31180
+    },
+    {
+      "epoch": 28.048561151079138,
+      "grad_norm": 0.1850016564130783,
+      "learning_rate": 5.0857006554132736e-05,
+      "loss": 0.0189,
+      "step": 31190
+    },
+    {
+      "action_loss": 0.008330575190484524,
+      "epoch": 28.048561151079138,
+      "step": 31190
+    },
+    {
+      "epoch": 28.048561151079138,
+      "step": 31190,
+      "torque_loss": 0.22060860693454742
+    },
+    {
+      "epoch": 28.048561151079138,
+      "force_loss": 0.009200750850141048,
+      "step": 31190
+    },
+    {
+      "epoch": 28.057553956834532,
+      "grad_norm": 0.1515791118144989,
+      "learning_rate": 5.0829452643388575e-05,
+      "loss": 0.0172,
+      "step": 31200
+    },
+    {
+      "action_loss": 0.003502137726172805,
+      "epoch": 28.057553956834532,
+      "step": 31200
+    },
+    {
+      "epoch": 28.057553956834532,
+      "step": 31200,
+      "torque_loss": 0.12400698661804199
+    },
+    {
+      "epoch": 28.057553956834532,
+      "force_loss": 0.004505326505750418,
+      "step": 31200
+    },
+    {
+      "epoch": 28.06654676258993,
+      "grad_norm": 0.2560032904148102,
+      "learning_rate": 5.08018984806786e-05,
+      "loss": 0.0167,
+      "step": 31210
+    },
+    {
+      "action_loss": 0.0030723672825843096,
+      "epoch": 28.06654676258993,
+      "step": 31210
+    },
+    {
+      "epoch": 28.06654676258993,
+      "step": 31210,
+      "torque_loss": 0.12460774928331375
+    },
+    {
+      "epoch": 28.06654676258993,
+      "force_loss": 0.0016551404260098934,
+      "step": 31210
+    },
+    {
+      "epoch": 28.075539568345324,
+      "grad_norm": 0.3447650372982025,
+      "learning_rate": 5.0774344074373036e-05,
+      "loss": 0.0199,
+      "step": 31220
+    },
+    {
+      "action_loss": 0.002020900370553136,
+      "epoch": 28.075539568345324,
+      "step": 31220
+    },
+    {
+      "epoch": 28.075539568345324,
+      "step": 31220,
+      "torque_loss": 0.14845220744609833
+    },
+    {
+      "epoch": 28.075539568345324,
+      "force_loss": 0.0034141484647989273,
+      "step": 31220
+    },
+    {
+      "epoch": 28.084532374100718,
+      "grad_norm": 0.2701382040977478,
+      "learning_rate": 5.07467894328422e-05,
+      "loss": 0.0191,
+      "step": 31230
+    },
+    {
+      "action_loss": 0.0048328242264688015,
+      "epoch": 28.084532374100718,
+      "step": 31230
+    },
+    {
+      "epoch": 28.084532374100718,
+      "step": 31230,
+      "torque_loss": 0.10102403163909912
+    },
+    {
+      "epoch": 28.084532374100718,
+      "force_loss": 0.0024929549545049667,
+      "step": 31230
+    },
+    {
+      "epoch": 28.093525179856115,
+      "grad_norm": 0.4575992226600647,
+      "learning_rate": 5.0719234564456454e-05,
+      "loss": 0.018,
+      "step": 31240
+    },
+    {
+      "action_loss": 0.0035630997736006975,
+      "epoch": 28.093525179856115,
+      "step": 31240
+    },
+    {
+      "epoch": 28.093525179856115,
+      "step": 31240,
+      "torque_loss": 0.14600805938243866
+    },
+    {
+      "epoch": 28.093525179856115,
+      "force_loss": 0.003792768344283104,
+      "step": 31240
+    },
+    {
+      "epoch": 28.10251798561151,
+      "grad_norm": 0.14927834272384644,
+      "learning_rate": 5.0691679477586216e-05,
+      "loss": 0.0147,
+      "step": 31250
+    },
+    {
+      "action_loss": 0.0029871861916035414,
+      "epoch": 28.10251798561151,
+      "step": 31250
+    },
+    {
+      "epoch": 28.10251798561151,
+      "step": 31250,
+      "torque_loss": 0.13111555576324463
+    },
+    {
+      "epoch": 28.10251798561151,
+      "force_loss": 0.005156335886567831,
+      "step": 31250
+    },
+    {
+      "epoch": 28.111510791366907,
+      "grad_norm": 0.21014977991580963,
+      "learning_rate": 5.0664124180602035e-05,
+      "loss": 0.017,
+      "step": 31260
+    },
+    {
+      "action_loss": 0.0018609591061249375,
+      "epoch": 28.111510791366907,
+      "step": 31260
+    },
+    {
+      "epoch": 28.111510791366907,
+      "step": 31260,
+      "torque_loss": 0.08364728838205338
+    },
+    {
+      "epoch": 28.111510791366907,
+      "force_loss": 0.003630842315033078,
+      "step": 31260
+    },
+    {
+      "epoch": 28.1205035971223,
+      "grad_norm": 0.2883286476135254,
+      "learning_rate": 5.063656868187447e-05,
+      "loss": 0.0174,
+      "step": 31270
+    },
+    {
+      "action_loss": 0.004610724281519651,
+      "epoch": 28.1205035971223,
+      "step": 31270
+    },
+    {
+      "epoch": 28.1205035971223,
+      "step": 31270,
+      "torque_loss": 0.11254767328500748
+    },
+    {
+      "epoch": 28.1205035971223,
+      "force_loss": 0.008744189515709877,
+      "step": 31270
+    },
+    {
+      "epoch": 28.1294964028777,
+      "grad_norm": 0.26366499066352844,
+      "learning_rate": 5.060901298977413e-05,
+      "loss": 0.0185,
+      "step": 31280
+    },
+    {
+      "action_loss": 0.011458448134362698,
+      "epoch": 28.1294964028777,
+      "step": 31280
+    },
+    {
+      "epoch": 28.1294964028777,
+      "step": 31280,
+      "torque_loss": 0.13499057292938232
+    },
+    {
+      "epoch": 28.1294964028777,
+      "force_loss": 0.010758512653410435,
+      "step": 31280
+    },
+    {
+      "epoch": 28.138489208633093,
+      "grad_norm": 0.42329901456832886,
+      "learning_rate": 5.0581457112671725e-05,
+      "loss": 0.028,
+      "step": 31290
+    },
+    {
+      "action_loss": 0.006546385586261749,
+      "epoch": 28.138489208633093,
+      "step": 31290
+    },
+    {
+      "epoch": 28.138489208633093,
+      "step": 31290,
+      "torque_loss": 0.10845958441495895
+    },
+    {
+      "epoch": 28.138489208633093,
+      "force_loss": 0.010698576457798481,
+      "step": 31290
+    },
+    {
+      "epoch": 28.14748201438849,
+      "grad_norm": 0.28220894932746887,
+      "learning_rate": 5.0553901058938016e-05,
+      "loss": 0.0177,
+      "step": 31300
+    },
+    {
+      "action_loss": 0.00202188896946609,
+      "epoch": 28.14748201438849,
+      "step": 31300
+    },
+    {
+      "epoch": 28.14748201438849,
+      "step": 31300,
+      "torque_loss": 0.09039351344108582
+    },
+    {
+      "epoch": 28.14748201438849,
+      "force_loss": 0.0016027702949941158,
+      "step": 31300
+    },
+    {
+      "epoch": 28.156474820143885,
+      "grad_norm": 0.1555168777704239,
+      "learning_rate": 5.052634483694377e-05,
+      "loss": 0.0178,
+      "step": 31310
+    },
+    {
+      "action_loss": 0.003646890399977565,
+      "epoch": 28.156474820143885,
+      "step": 31310
+    },
+    {
+      "epoch": 28.156474820143885,
+      "step": 31310,
+      "torque_loss": 0.1680331975221634
+    },
+    {
+      "epoch": 28.156474820143885,
+      "force_loss": 0.002329886658117175,
+      "step": 31310
+    },
+    {
+      "epoch": 28.165467625899282,
+      "grad_norm": 0.3172502815723419,
+      "learning_rate": 5.049878845505988e-05,
+      "loss": 0.0232,
+      "step": 31320
+    },
+    {
+      "action_loss": 0.0020603330340236425,
+      "epoch": 28.165467625899282,
+      "step": 31320
+    },
+    {
+      "epoch": 28.165467625899282,
+      "step": 31320,
+      "torque_loss": 0.12851010262966156
+    },
+    {
+      "epoch": 28.165467625899282,
+      "force_loss": 0.001722722896374762,
+      "step": 31320
+    },
+    {
+      "epoch": 28.174460431654676,
+      "grad_norm": 0.1463720053434372,
+      "learning_rate": 5.047123192165721e-05,
+      "loss": 0.0195,
+      "step": 31330
+    },
+    {
+      "action_loss": 0.003479136386886239,
+      "epoch": 28.174460431654676,
+      "step": 31330
+    },
+    {
+      "epoch": 28.174460431654676,
+      "step": 31330,
+      "torque_loss": 0.11305113881826401
+    },
+    {
+      "epoch": 28.174460431654676,
+      "force_loss": 0.003512629074975848,
+      "step": 31330
+    },
+    {
+      "epoch": 28.18345323741007,
+      "grad_norm": 0.18156354129314423,
+      "learning_rate": 5.0443675245106735e-05,
+      "loss": 0.0184,
+      "step": 31340
+    },
+    {
+      "action_loss": 0.0027709228452295065,
+      "epoch": 28.18345323741007,
+      "step": 31340
+    },
+    {
+      "epoch": 28.18345323741007,
+      "step": 31340,
+      "torque_loss": 0.12847156822681427
+    },
+    {
+      "epoch": 28.18345323741007,
+      "force_loss": 0.002399100223556161,
+      "step": 31340
+    },
+    {
+      "epoch": 28.192446043165468,
+      "grad_norm": 0.24686631560325623,
+      "learning_rate": 5.0416118433779426e-05,
+      "loss": 0.0165,
+      "step": 31350
+    },
+    {
+      "action_loss": 0.007418935652822256,
+      "epoch": 28.192446043165468,
+      "step": 31350
+    },
+    {
+      "epoch": 28.192446043165468,
+      "step": 31350,
+      "torque_loss": 0.1823035627603531
+    },
+    {
+      "epoch": 28.192446043165468,
+      "force_loss": 0.00943506695330143,
+      "step": 31350
+    },
+    {
+      "epoch": 28.201438848920862,
+      "grad_norm": 0.21008306741714478,
+      "learning_rate": 5.038856149604633e-05,
+      "loss": 0.0178,
+      "step": 31360
+    },
+    {
+      "action_loss": 0.009561304934322834,
+      "epoch": 28.201438848920862,
+      "step": 31360
+    },
+    {
+      "epoch": 28.201438848920862,
+      "step": 31360,
+      "torque_loss": 0.17719674110412598
+    },
+    {
+      "epoch": 28.201438848920862,
+      "force_loss": 0.009404500015079975,
+      "step": 31360
+    },
+    {
+      "epoch": 28.21043165467626,
+      "grad_norm": 0.28640440106391907,
+      "learning_rate": 5.03610044402785e-05,
+      "loss": 0.0197,
+      "step": 31370
+    },
+    {
+      "action_loss": 0.0022651939652860165,
+      "epoch": 28.21043165467626,
+      "step": 31370
+    },
+    {
+      "epoch": 28.21043165467626,
+      "step": 31370,
+      "torque_loss": 0.12357058376073837
+    },
+    {
+      "epoch": 28.21043165467626,
+      "force_loss": 0.0031259600073099136,
+      "step": 31370
+    },
+    {
+      "epoch": 28.219424460431654,
+      "grad_norm": 0.1845446527004242,
+      "learning_rate": 5.033344727484707e-05,
+      "loss": 0.0214,
+      "step": 31380
+    },
+    {
+      "action_loss": 0.00495353527367115,
+      "epoch": 28.219424460431654,
+      "step": 31380
+    },
+    {
+      "epoch": 28.219424460431654,
+      "step": 31380,
+      "torque_loss": 0.16547518968582153
+    },
+    {
+      "epoch": 28.219424460431654,
+      "force_loss": 0.005297411698848009,
+      "step": 31380
+    },
+    {
+      "epoch": 28.22841726618705,
+      "grad_norm": 0.15085051953792572,
+      "learning_rate": 5.030589000812315e-05,
+      "loss": 0.0173,
+      "step": 31390
+    },
+    {
+      "action_loss": 0.003035797504708171,
+      "epoch": 28.22841726618705,
+      "step": 31390
+    },
+    {
+      "epoch": 28.22841726618705,
+      "step": 31390,
+      "torque_loss": 0.12055474519729614
+    },
+    {
+      "epoch": 28.22841726618705,
+      "force_loss": 0.006290100514888763,
+      "step": 31390
+    },
+    {
+      "epoch": 28.237410071942445,
+      "grad_norm": 0.36627402901649475,
+      "learning_rate": 5.027833264847793e-05,
+      "loss": 0.0178,
+      "step": 31400
+    },
+    {
+      "action_loss": 0.005598003510385752,
+      "epoch": 28.237410071942445,
+      "step": 31400
+    },
+    {
+      "epoch": 28.237410071942445,
+      "step": 31400,
+      "torque_loss": 0.1297248750925064
+    },
+    {
+      "epoch": 28.237410071942445,
+      "force_loss": 0.0020954047795385122,
+      "step": 31400
+    },
+    {
+      "epoch": 28.246402877697843,
+      "grad_norm": 0.22061501443386078,
+      "learning_rate": 5.025077520428258e-05,
+      "loss": 0.0213,
+      "step": 31410
+    },
+    {
+      "action_loss": 0.0025652453768998384,
+      "epoch": 28.246402877697843,
+      "step": 31410
+    },
+    {
+      "epoch": 28.246402877697843,
+      "step": 31410,
+      "torque_loss": 0.12894245982170105
+    },
+    {
+      "epoch": 28.246402877697843,
+      "force_loss": 0.003746572183445096,
+      "step": 31410
+    },
+    {
+      "epoch": 28.255395683453237,
+      "grad_norm": 0.15218910574913025,
+      "learning_rate": 5.022321768390837e-05,
+      "loss": 0.0166,
+      "step": 31420
+    },
+    {
+      "action_loss": 0.014568574726581573,
+      "epoch": 28.255395683453237,
+      "step": 31420
+    },
+    {
+      "epoch": 28.255395683453237,
+      "step": 31420,
+      "torque_loss": 0.12638191878795624
+    },
+    {
+      "epoch": 28.255395683453237,
+      "force_loss": 0.01743035577237606,
+      "step": 31420
+    },
+    {
+      "epoch": 28.264388489208635,
+      "grad_norm": 0.25253164768218994,
+      "learning_rate": 5.0195660095726516e-05,
+      "loss": 0.0199,
+      "step": 31430
+    },
+    {
+      "action_loss": 0.0025993958115577698,
+      "epoch": 28.264388489208635,
+      "step": 31430
+    },
+    {
+      "epoch": 28.264388489208635,
+      "step": 31430,
+      "torque_loss": 0.11008652299642563
+    },
+    {
+      "epoch": 28.264388489208635,
+      "force_loss": 0.0024869602639228106,
+      "step": 31430
+    },
+    {
+      "epoch": 28.27338129496403,
+      "grad_norm": 0.147151917219162,
+      "learning_rate": 5.016810244810829e-05,
+      "loss": 0.0192,
+      "step": 31440
+    },
+    {
+      "action_loss": 0.00280805304646492,
+      "epoch": 28.27338129496403,
+      "step": 31440
+    },
+    {
+      "epoch": 28.27338129496403,
+      "step": 31440,
+      "torque_loss": 0.1111692562699318
+    },
+    {
+      "epoch": 28.27338129496403,
+      "force_loss": 0.0022786271292716265,
+      "step": 31440
+    },
+    {
+      "epoch": 28.282374100719423,
+      "grad_norm": 0.56572425365448,
+      "learning_rate": 5.0140544749424976e-05,
+      "loss": 0.017,
+      "step": 31450
+    },
+    {
+      "action_loss": 0.00293734110891819,
+      "epoch": 28.282374100719423,
+      "step": 31450
+    },
+    {
+      "epoch": 28.282374100719423,
+      "step": 31450,
+      "torque_loss": 0.16455556452274323
+    },
+    {
+      "epoch": 28.282374100719423,
+      "force_loss": 0.002204336691647768,
+      "step": 31450
+    },
+    {
+      "epoch": 28.29136690647482,
+      "grad_norm": 0.27453145384788513,
+      "learning_rate": 5.0112987008047874e-05,
+      "loss": 0.0163,
+      "step": 31460
+    },
+    {
+      "action_loss": 0.019957998767495155,
+      "epoch": 28.29136690647482,
+      "step": 31460
+    },
+    {
+      "epoch": 28.29136690647482,
+      "step": 31460,
+      "torque_loss": 0.15143389999866486
+    },
+    {
+      "epoch": 28.29136690647482,
+      "force_loss": 0.025196874514222145,
+      "step": 31460
+    },
+    {
+      "epoch": 28.300359712230215,
+      "grad_norm": 0.1736888736486435,
+      "learning_rate": 5.008542923234831e-05,
+      "loss": 0.0197,
+      "step": 31470
+    },
+    {
+      "action_loss": 0.00510973297059536,
+      "epoch": 28.300359712230215,
+      "step": 31470
+    },
+    {
+      "epoch": 28.300359712230215,
+      "step": 31470,
+      "torque_loss": 0.11906745284795761
+    },
+    {
+      "epoch": 28.300359712230215,
+      "force_loss": 0.009846851229667664,
+      "step": 31470
+    },
+    {
+      "epoch": 28.309352517985612,
+      "grad_norm": 0.1319504976272583,
+      "learning_rate": 5.00578714306976e-05,
+      "loss": 0.0198,
+      "step": 31480
+    },
+    {
+      "action_loss": 0.00496891001239419,
+      "epoch": 28.309352517985612,
+      "step": 31480
+    },
+    {
+      "epoch": 28.309352517985612,
+      "step": 31480,
+      "torque_loss": 0.13380688428878784
+    },
+    {
+      "epoch": 28.309352517985612,
+      "force_loss": 0.004541265778243542,
+      "step": 31480
+    },
+    {
+      "epoch": 28.318345323741006,
+      "grad_norm": 0.2568466067314148,
+      "learning_rate": 5.0030313611467084e-05,
+      "loss": 0.0192,
+      "step": 31490
+    },
+    {
+      "action_loss": 0.0056123011745512486,
+      "epoch": 28.318345323741006,
+      "step": 31490
+    },
+    {
+      "epoch": 28.318345323741006,
+      "step": 31490,
+      "torque_loss": 0.12392618507146835
+    },
+    {
+      "epoch": 28.318345323741006,
+      "force_loss": 0.00703589478507638,
+      "step": 31490
+    },
+    {
+      "epoch": 28.327338129496404,
+      "grad_norm": 0.33176225423812866,
+      "learning_rate": 5.0002755783028074e-05,
+      "loss": 0.0177,
+      "step": 31500
+    },
+    {
+      "action_loss": 0.012295238673686981,
+      "epoch": 28.327338129496404,
+      "step": 31500
+    },
+    {
+      "epoch": 28.327338129496404,
+      "step": 31500,
+      "torque_loss": 0.12550227344036102
+    },
+    {
+      "epoch": 28.327338129496404,
+      "force_loss": 0.013909243047237396,
+      "step": 31500
+    },
+    {
+      "epoch": 28.336330935251798,
+      "grad_norm": 0.6499210000038147,
+      "learning_rate": 4.997519795375194e-05,
+      "loss": 0.0191,
+      "step": 31510
+    },
+    {
+      "action_loss": 0.0024094032123684883,
+      "epoch": 28.336330935251798,
+      "step": 31510
+    },
+    {
+      "epoch": 28.336330935251798,
+      "step": 31510,
+      "torque_loss": 0.11365920305252075
+    },
+    {
+      "epoch": 28.336330935251798,
+      "force_loss": 0.0022959013003855944,
+      "step": 31510
+    },
+    {
+      "epoch": 28.345323741007196,
+      "grad_norm": 0.14598187804222107,
+      "learning_rate": 4.9947640132010016e-05,
+      "loss": 0.0191,
+      "step": 31520
+    },
+    {
+      "action_loss": 0.010777813382446766,
+      "epoch": 28.345323741007196,
+      "step": 31520
+    },
+    {
+      "epoch": 28.345323741007196,
+      "step": 31520,
+      "torque_loss": 0.13041585683822632
+    },
+    {
+      "epoch": 28.345323741007196,
+      "force_loss": 0.01159286592155695,
+      "step": 31520
+    },
+    {
+      "epoch": 28.35431654676259,
+      "grad_norm": 0.28974649310112,
+      "learning_rate": 4.9920082326173625e-05,
+      "loss": 0.0189,
+      "step": 31530
+    },
+    {
+      "action_loss": 0.007011393550783396,
+      "epoch": 28.35431654676259,
+      "step": 31530
+    },
+    {
+      "epoch": 28.35431654676259,
+      "step": 31530,
+      "torque_loss": 0.12972255051136017
+    },
+    {
+      "epoch": 28.35431654676259,
+      "force_loss": 0.003591722808778286,
+      "step": 31530
+    },
+    {
+      "epoch": 28.363309352517987,
+      "grad_norm": 0.2784913182258606,
+      "learning_rate": 4.9892524544614114e-05,
+      "loss": 0.0204,
+      "step": 31540
+    },
+    {
+      "action_loss": 0.00864789541810751,
+      "epoch": 28.363309352517987,
+      "step": 31540
+    },
+    {
+      "epoch": 28.363309352517987,
+      "step": 31540,
+      "torque_loss": 0.12829583883285522
+    },
+    {
+      "epoch": 28.363309352517987,
+      "force_loss": 0.009747441858053207,
+      "step": 31540
+    },
+    {
+      "epoch": 28.37230215827338,
+      "grad_norm": 0.2597323954105377,
+      "learning_rate": 4.986496679570283e-05,
+      "loss": 0.0197,
+      "step": 31550
+    },
+    {
+      "action_loss": 0.009286990389227867,
+      "epoch": 28.37230215827338,
+      "step": 31550
+    },
+    {
+      "epoch": 28.37230215827338,
+      "step": 31550,
+      "torque_loss": 0.11492288112640381
+    },
+    {
+      "epoch": 28.37230215827338,
+      "force_loss": 0.006870301906019449,
+      "step": 31550
+    },
+    {
+      "epoch": 28.381294964028775,
+      "grad_norm": 0.6100925207138062,
+      "learning_rate": 4.983740908781105e-05,
+      "loss": 0.0217,
+      "step": 31560
+    },
+    {
+      "action_loss": 0.002930929185822606,
+      "epoch": 28.381294964028775,
+      "step": 31560
+    },
+    {
+      "epoch": 28.381294964028775,
+      "step": 31560,
+      "torque_loss": 0.15469162166118622
+    },
+    {
+      "epoch": 28.381294964028775,
+      "force_loss": 0.003778379410505295,
+      "step": 31560
+    },
+    {
+      "epoch": 28.390287769784173,
+      "grad_norm": 0.3829963505268097,
+      "learning_rate": 4.9809851429310116e-05,
+      "loss": 0.0175,
+      "step": 31570
+    },
+    {
+      "action_loss": 0.003712309757247567,
+      "epoch": 28.390287769784173,
+      "step": 31570
+    },
+    {
+      "epoch": 28.390287769784173,
+      "step": 31570,
+      "torque_loss": 0.12879160046577454
+    },
+    {
+      "epoch": 28.390287769784173,
+      "force_loss": 0.0027420343831181526,
+      "step": 31570
+    },
+    {
+      "epoch": 28.399280575539567,
+      "grad_norm": 0.23992829024791718,
+      "learning_rate": 4.9782293828571275e-05,
+      "loss": 0.022,
+      "step": 31580
+    },
+    {
+      "action_loss": 0.004066178109496832,
+      "epoch": 28.399280575539567,
+      "step": 31580
+    },
+    {
+      "epoch": 28.399280575539567,
+      "step": 31580,
+      "torque_loss": 0.08748013526201248
+    },
+    {
+      "epoch": 28.399280575539567,
+      "force_loss": 0.002831103513017297,
+      "step": 31580
+    },
+    {
+      "epoch": 28.408273381294965,
+      "grad_norm": 0.2187284380197525,
+      "learning_rate": 4.9754736293965846e-05,
+      "loss": 0.0167,
+      "step": 31590
+    },
+    {
+      "action_loss": 0.0063866134732961655,
+      "epoch": 28.408273381294965,
+      "step": 31590
+    },
+    {
+      "epoch": 28.408273381294965,
+      "step": 31590,
+      "torque_loss": 0.15923090279102325
+    },
+    {
+      "epoch": 28.408273381294965,
+      "force_loss": 0.0038600738625973463,
+      "step": 31590
+    },
+    {
+      "epoch": 28.41726618705036,
+      "grad_norm": 0.5560687780380249,
+      "learning_rate": 4.972717883386502e-05,
+      "loss": 0.0219,
+      "step": 31600
+    },
+    {
+      "action_loss": 0.0023293609265238047,
+      "epoch": 28.41726618705036,
+      "step": 31600
+    },
+    {
+      "epoch": 28.41726618705036,
+      "step": 31600,
+      "torque_loss": 0.1120397076010704
+    },
+    {
+      "epoch": 28.41726618705036,
+      "force_loss": 0.001863547950051725,
+      "step": 31600
+    },
+    {
+      "epoch": 28.426258992805757,
+      "grad_norm": 0.2414456009864807,
+      "learning_rate": 4.9699621456640075e-05,
+      "loss": 0.0192,
+      "step": 31610
+    },
+    {
+      "action_loss": 0.0022713118232786655,
+      "epoch": 28.426258992805757,
+      "step": 31610
+    },
+    {
+      "epoch": 28.426258992805757,
+      "step": 31610,
+      "torque_loss": 0.09746291488409042
+    },
+    {
+      "epoch": 28.426258992805757,
+      "force_loss": 0.0019517071777954698,
+      "step": 31610
+    },
+    {
+      "epoch": 28.43525179856115,
+      "grad_norm": 0.44207125902175903,
+      "learning_rate": 4.9672064170662214e-05,
+      "loss": 0.0166,
+      "step": 31620
+    },
+    {
+      "action_loss": 0.0016194864874705672,
+      "epoch": 28.43525179856115,
+      "step": 31620
+    },
+    {
+      "epoch": 28.43525179856115,
+      "step": 31620,
+      "torque_loss": 0.11972985416650772
+    },
+    {
+      "epoch": 28.43525179856115,
+      "force_loss": 0.0015085259219631553,
+      "step": 31620
+    },
+    {
+      "epoch": 28.444244604316548,
+      "grad_norm": 0.6432380080223083,
+      "learning_rate": 4.9644506984302583e-05,
+      "loss": 0.0229,
+      "step": 31630
+    },
+    {
+      "action_loss": 0.0028622543904930353,
+      "epoch": 28.444244604316548,
+      "step": 31630
+    },
+    {
+      "epoch": 28.444244604316548,
+      "step": 31630,
+      "torque_loss": 0.11647132784128189
+    },
+    {
+      "epoch": 28.444244604316548,
+      "force_loss": 0.0022364670876413584,
+      "step": 31630
+    },
+    {
+      "epoch": 28.453237410071942,
+      "grad_norm": 0.2759113013744354,
+      "learning_rate": 4.9616949905932356e-05,
+      "loss": 0.0184,
+      "step": 31640
+    },
+    {
+      "action_loss": 0.0022402871400117874,
+      "epoch": 28.453237410071942,
+      "step": 31640
+    },
+    {
+      "epoch": 28.453237410071942,
+      "step": 31640,
+      "torque_loss": 0.07226861268281937
+    },
+    {
+      "epoch": 28.453237410071942,
+      "force_loss": 0.002943779109045863,
+      "step": 31640
+    },
+    {
+      "epoch": 28.46223021582734,
+      "grad_norm": 0.26935675740242004,
+      "learning_rate": 4.9589392943922615e-05,
+      "loss": 0.02,
+      "step": 31650
+    },
+    {
+      "action_loss": 0.003252230817452073,
+      "epoch": 28.46223021582734,
+      "step": 31650
+    },
+    {
+      "epoch": 28.46223021582734,
+      "step": 31650,
+      "torque_loss": 0.13012158870697021
+    },
+    {
+      "epoch": 28.46223021582734,
+      "force_loss": 0.0021820305846631527,
+      "step": 31650
+    },
+    {
+      "epoch": 28.471223021582734,
+      "grad_norm": 0.3346239924430847,
+      "learning_rate": 4.956183610664447e-05,
+      "loss": 0.0207,
+      "step": 31660
+    },
+    {
+      "action_loss": 0.004996235948055983,
+      "epoch": 28.471223021582734,
+      "step": 31660
+    },
+    {
+      "epoch": 28.471223021582734,
+      "step": 31660,
+      "torque_loss": 0.1323365569114685
+    },
+    {
+      "epoch": 28.471223021582734,
+      "force_loss": 0.006417748052626848,
+      "step": 31660
+    },
+    {
+      "epoch": 28.480215827338128,
+      "grad_norm": 0.2346462607383728,
+      "learning_rate": 4.9534279402468945e-05,
+      "loss": 0.0172,
+      "step": 31670
+    },
+    {
+      "action_loss": 0.015941092744469643,
+      "epoch": 28.480215827338128,
+      "step": 31670
+    },
+    {
+      "epoch": 28.480215827338128,
+      "step": 31670,
+      "torque_loss": 0.1559155136346817
+    },
+    {
+      "epoch": 28.480215827338128,
+      "force_loss": 0.019919589161872864,
+      "step": 31670
+    },
+    {
+      "epoch": 28.489208633093526,
+      "grad_norm": 0.11588010936975479,
+      "learning_rate": 4.9506722839767036e-05,
+      "loss": 0.0182,
+      "step": 31680
+    },
+    {
+      "action_loss": 0.008000590838491917,
+      "epoch": 28.489208633093526,
+      "step": 31680
+    },
+    {
+      "epoch": 28.489208633093526,
+      "step": 31680,
+      "torque_loss": 0.12118353694677353
+    },
+    {
+      "epoch": 28.489208633093526,
+      "force_loss": 0.0066853477619588375,
+      "step": 31680
+    },
+    {
+      "epoch": 28.49820143884892,
+      "grad_norm": 0.18816101551055908,
+      "learning_rate": 4.947916642690972e-05,
+      "loss": 0.026,
+      "step": 31690
+    },
+    {
+      "action_loss": 0.0025853884872049093,
+      "epoch": 28.49820143884892,
+      "step": 31690
+    },
+    {
+      "epoch": 28.49820143884892,
+      "step": 31690,
+      "torque_loss": 0.12101741880178452
+    },
+    {
+      "epoch": 28.49820143884892,
+      "force_loss": 0.0020138428080826998,
+      "step": 31690
+    },
+    {
+      "epoch": 28.507194244604317,
+      "grad_norm": 0.16970165073871613,
+      "learning_rate": 4.9451610172267874e-05,
+      "loss": 0.0155,
+      "step": 31700
+    },
+    {
+      "action_loss": 0.00584329292178154,
+      "epoch": 28.507194244604317,
+      "step": 31700
+    },
+    {
+      "epoch": 28.507194244604317,
+      "step": 31700,
+      "torque_loss": 0.13173158466815948
+    },
+    {
+      "epoch": 28.507194244604317,
+      "force_loss": 0.006607357412576675,
+      "step": 31700
+    },
+    {
+      "epoch": 28.51618705035971,
+      "grad_norm": 0.5986936688423157,
+      "learning_rate": 4.9424054084212376e-05,
+      "loss": 0.0179,
+      "step": 31710
+    },
+    {
+      "action_loss": 0.002945384941995144,
+      "epoch": 28.51618705035971,
+      "step": 31710
+    },
+    {
+      "epoch": 28.51618705035971,
+      "step": 31710,
+      "torque_loss": 0.1225002184510231
+    },
+    {
+      "epoch": 28.51618705035971,
+      "force_loss": 0.00539809325709939,
+      "step": 31710
+    },
+    {
+      "epoch": 28.52517985611511,
+      "grad_norm": 0.25696679949760437,
+      "learning_rate": 4.939649817111407e-05,
+      "loss": 0.0181,
+      "step": 31720
+    },
+    {
+      "action_loss": 0.0038010275457054377,
+      "epoch": 28.52517985611511,
+      "step": 31720
+    },
+    {
+      "epoch": 28.52517985611511,
+      "step": 31720,
+      "torque_loss": 0.14548729360103607
+    },
+    {
+      "epoch": 28.52517985611511,
+      "force_loss": 0.0033391022589057684,
+      "step": 31720
+    },
+    {
+      "epoch": 28.534172661870503,
+      "grad_norm": 0.23569805920124054,
+      "learning_rate": 4.936894244134365e-05,
+      "loss": 0.0171,
+      "step": 31730
+    },
+    {
+      "action_loss": 0.0066145905293524265,
+      "epoch": 28.534172661870503,
+      "step": 31730
+    },
+    {
+      "epoch": 28.534172661870503,
+      "step": 31730,
+      "torque_loss": 0.17376501858234406
+    },
+    {
+      "epoch": 28.534172661870503,
+      "force_loss": 0.005036663729697466,
+      "step": 31730
+    },
+    {
+      "epoch": 28.5431654676259,
+      "grad_norm": 0.27967146039009094,
+      "learning_rate": 4.9341386903271886e-05,
+      "loss": 0.0185,
+      "step": 31740
+    },
+    {
+      "action_loss": 0.014002300798892975,
+      "epoch": 28.5431654676259,
+      "step": 31740
+    },
+    {
+      "epoch": 28.5431654676259,
+      "step": 31740,
+      "torque_loss": 0.1286267787218094
+    },
+    {
+      "epoch": 28.5431654676259,
+      "force_loss": 0.014078046195209026,
+      "step": 31740
+    },
+    {
+      "epoch": 28.552158273381295,
+      "grad_norm": 0.2546086013317108,
+      "learning_rate": 4.931383156526936e-05,
+      "loss": 0.0229,
+      "step": 31750
+    },
+    {
+      "action_loss": 0.015389028005301952,
+      "epoch": 28.552158273381295,
+      "step": 31750
+    },
+    {
+      "epoch": 28.552158273381295,
+      "step": 31750,
+      "torque_loss": 0.10529843717813492
+    },
+    {
+      "epoch": 28.552158273381295,
+      "force_loss": 0.013241861946880817,
+      "step": 31750
+    },
+    {
+      "epoch": 28.56115107913669,
+      "grad_norm": 0.19817949831485748,
+      "learning_rate": 4.92862764357067e-05,
+      "loss": 0.0211,
+      "step": 31760
+    },
+    {
+      "action_loss": 0.0013725543394684792,
+      "epoch": 28.56115107913669,
+      "step": 31760
+    },
+    {
+      "epoch": 28.56115107913669,
+      "step": 31760,
+      "torque_loss": 0.10182621330022812
+    },
+    {
+      "epoch": 28.56115107913669,
+      "force_loss": 0.0017656233394518495,
+      "step": 31760
+    },
+    {
+      "epoch": 28.570143884892087,
+      "grad_norm": 0.13827678561210632,
+      "learning_rate": 4.925872152295443e-05,
+      "loss": 0.0171,
+      "step": 31770
+    },
+    {
+      "action_loss": 0.0016914677107706666,
+      "epoch": 28.570143884892087,
+      "step": 31770
+    },
+    {
+      "epoch": 28.570143884892087,
+      "step": 31770,
+      "torque_loss": 0.07993989437818527
+    },
+    {
+      "epoch": 28.570143884892087,
+      "force_loss": 0.0034969402477145195,
+      "step": 31770
+    },
+    {
+      "epoch": 28.57913669064748,
+      "grad_norm": 0.2877127528190613,
+      "learning_rate": 4.923116683538296e-05,
+      "loss": 0.0188,
+      "step": 31780
+    },
+    {
+      "action_loss": 0.010105691850185394,
+      "epoch": 28.57913669064748,
+      "step": 31780
+    },
+    {
+      "epoch": 28.57913669064748,
+      "step": 31780,
+      "torque_loss": 0.11479947715997696
+    },
+    {
+      "epoch": 28.57913669064748,
+      "force_loss": 0.011219676584005356,
+      "step": 31780
+    },
+    {
+      "epoch": 28.58812949640288,
+      "grad_norm": 0.11982277780771255,
+      "learning_rate": 4.920361238136273e-05,
+      "loss": 0.0191,
+      "step": 31790
+    },
+    {
+      "action_loss": 0.007315091788768768,
+      "epoch": 28.58812949640288,
+      "step": 31790
+    },
+    {
+      "epoch": 28.58812949640288,
+      "step": 31790,
+      "torque_loss": 0.1151256263256073
+    },
+    {
+      "epoch": 28.58812949640288,
+      "force_loss": 0.010056677274405956,
+      "step": 31790
+    },
+    {
+      "epoch": 28.597122302158272,
+      "grad_norm": 0.09665153920650482,
+      "learning_rate": 4.9176058169264014e-05,
+      "loss": 0.0193,
+      "step": 31800
+    },
+    {
+      "action_loss": 0.00902259536087513,
+      "epoch": 28.597122302158272,
+      "step": 31800
+    },
+    {
+      "epoch": 28.597122302158272,
+      "step": 31800,
+      "torque_loss": 0.10882830619812012
+    },
+    {
+      "epoch": 28.597122302158272,
+      "force_loss": 0.022170864045619965,
+      "step": 31800
+    },
+    {
+      "epoch": 28.60611510791367,
+      "grad_norm": 0.16901050508022308,
+      "learning_rate": 4.9148504207457074e-05,
+      "loss": 0.0212,
+      "step": 31810
+    },
+    {
+      "action_loss": 0.002538155997171998,
+      "epoch": 28.60611510791367,
+      "step": 31810
+    },
+    {
+      "epoch": 28.60611510791367,
+      "step": 31810,
+      "torque_loss": 0.15164512395858765
+    },
+    {
+      "epoch": 28.60611510791367,
+      "force_loss": 0.002965029561892152,
+      "step": 31810
+    },
+    {
+      "epoch": 28.615107913669064,
+      "grad_norm": 0.2416938692331314,
+      "learning_rate": 4.912095050431208e-05,
+      "loss": 0.0217,
+      "step": 31820
+    },
+    {
+      "action_loss": 0.004177636466920376,
+      "epoch": 28.615107913669064,
+      "step": 31820
+    },
+    {
+      "epoch": 28.615107913669064,
+      "step": 31820,
+      "torque_loss": 0.09113707393407822
+    },
+    {
+      "epoch": 28.615107913669064,
+      "force_loss": 0.008358192630112171,
+      "step": 31820
+    },
+    {
+      "epoch": 28.62410071942446,
+      "grad_norm": 0.1306232362985611,
+      "learning_rate": 4.909339706819911e-05,
+      "loss": 0.0194,
+      "step": 31830
+    },
+    {
+      "action_loss": 0.00744277611374855,
+      "epoch": 28.62410071942446,
+      "step": 31830
+    },
+    {
+      "epoch": 28.62410071942446,
+      "step": 31830,
+      "torque_loss": 0.15952813625335693
+    },
+    {
+      "epoch": 28.62410071942446,
+      "force_loss": 0.004684206563979387,
+      "step": 31830
+    },
+    {
+      "epoch": 28.633093525179856,
+      "grad_norm": 0.11685939133167267,
+      "learning_rate": 4.906584390748819e-05,
+      "loss": 0.0181,
+      "step": 31840
+    },
+    {
+      "action_loss": 0.010659020394086838,
+      "epoch": 28.633093525179856,
+      "step": 31840
+    },
+    {
+      "epoch": 28.633093525179856,
+      "step": 31840,
+      "torque_loss": 0.12336523085832596
+    },
+    {
+      "epoch": 28.633093525179856,
+      "force_loss": 0.023605575785040855,
+      "step": 31840
+    },
+    {
+      "epoch": 28.642086330935253,
+      "grad_norm": 0.3025333881378174,
+      "learning_rate": 4.9038291030549195e-05,
+      "loss": 0.0285,
+      "step": 31850
+    },
+    {
+      "action_loss": 0.0026493777986615896,
+      "epoch": 28.642086330935253,
+      "step": 31850
+    },
+    {
+      "epoch": 28.642086330935253,
+      "step": 31850,
+      "torque_loss": 0.1358613222837448
+    },
+    {
+      "epoch": 28.642086330935253,
+      "force_loss": 0.003288527950644493,
+      "step": 31850
+    },
+    {
+      "epoch": 28.651079136690647,
+      "grad_norm": 0.1424875408411026,
+      "learning_rate": 4.9010738445751995e-05,
+      "loss": 0.0187,
+      "step": 31860
+    },
+    {
+      "action_loss": 0.006690997630357742,
+      "epoch": 28.651079136690647,
+      "step": 31860
+    },
+    {
+      "epoch": 28.651079136690647,
+      "step": 31860,
+      "torque_loss": 0.1267218440771103
+    },
+    {
+      "epoch": 28.651079136690647,
+      "force_loss": 0.010846694000065327,
+      "step": 31860
+    },
+    {
+      "epoch": 28.66007194244604,
+      "grad_norm": 0.10953143239021301,
+      "learning_rate": 4.8983186161466364e-05,
+      "loss": 0.0172,
+      "step": 31870
+    },
+    {
+      "action_loss": 0.0069437469355762005,
+      "epoch": 28.66007194244604,
+      "step": 31870
+    },
+    {
+      "epoch": 28.66007194244604,
+      "step": 31870,
+      "torque_loss": 0.13778366148471832
+    },
+    {
+      "epoch": 28.66007194244604,
+      "force_loss": 0.0042550694197416306,
+      "step": 31870
+    },
+    {
+      "epoch": 28.66906474820144,
+      "grad_norm": 0.22482098639011383,
+      "learning_rate": 4.89556341860619e-05,
+      "loss": 0.0185,
+      "step": 31880
+    },
+    {
+      "action_loss": 0.00789960753172636,
+      "epoch": 28.66906474820144,
+      "step": 31880
+    },
+    {
+      "epoch": 28.66906474820144,
+      "step": 31880,
+      "torque_loss": 0.10606253147125244
+    },
+    {
+      "epoch": 28.66906474820144,
+      "force_loss": 0.01030336506664753,
+      "step": 31880
+    },
+    {
+      "epoch": 28.678057553956833,
+      "grad_norm": 0.25729864835739136,
+      "learning_rate": 4.892808252790822e-05,
+      "loss": 0.0199,
+      "step": 31890
+    },
+    {
+      "action_loss": 0.0035116756334900856,
+      "epoch": 28.678057553956833,
+      "step": 31890
+    },
+    {
+      "epoch": 28.678057553956833,
+      "step": 31890,
+      "torque_loss": 0.13032568991184235
+    },
+    {
+      "epoch": 28.678057553956833,
+      "force_loss": 0.002061855746433139,
+      "step": 31890
+    },
+    {
+      "epoch": 28.68705035971223,
+      "grad_norm": 0.21324051916599274,
+      "learning_rate": 4.890053119537475e-05,
+      "loss": 0.0184,
+      "step": 31900
+    },
+    {
+      "action_loss": 0.009989923797547817,
+      "epoch": 28.68705035971223,
+      "step": 31900
+    },
+    {
+      "epoch": 28.68705035971223,
+      "step": 31900,
+      "torque_loss": 0.11143625527620316
+    },
+    {
+      "epoch": 28.68705035971223,
+      "force_loss": 0.014830250293016434,
+      "step": 31900
+    },
+    {
+      "epoch": 28.696043165467625,
+      "grad_norm": 0.19445998966693878,
+      "learning_rate": 4.887298019683087e-05,
+      "loss": 0.0224,
+      "step": 31910
+    },
+    {
+      "action_loss": 0.004206265322864056,
+      "epoch": 28.696043165467625,
+      "step": 31910
+    },
+    {
+      "epoch": 28.696043165467625,
+      "step": 31910,
+      "torque_loss": 0.10461030155420303
+    },
+    {
+      "epoch": 28.696043165467625,
+      "force_loss": 0.002050551353022456,
+      "step": 31910
+    },
+    {
+      "epoch": 28.705035971223023,
+      "grad_norm": 0.11839746683835983,
+      "learning_rate": 4.884542954064587e-05,
+      "loss": 0.0231,
+      "step": 31920
+    },
+    {
+      "action_loss": 0.00904050562530756,
+      "epoch": 28.705035971223023,
+      "step": 31920
+    },
+    {
+      "epoch": 28.705035971223023,
+      "step": 31920,
+      "torque_loss": 0.16358019411563873
+    },
+    {
+      "epoch": 28.705035971223023,
+      "force_loss": 0.007803190499544144,
+      "step": 31920
+    },
+    {
+      "epoch": 28.714028776978417,
+      "grad_norm": 0.1934649497270584,
+      "learning_rate": 4.881787923518887e-05,
+      "loss": 0.0199,
+      "step": 31930
+    },
+    {
+      "action_loss": 0.004410514608025551,
+      "epoch": 28.714028776978417,
+      "step": 31930
+    },
+    {
+      "epoch": 28.714028776978417,
+      "step": 31930,
+      "torque_loss": 0.0705103725194931
+    },
+    {
+      "epoch": 28.714028776978417,
+      "force_loss": 0.007975791580975056,
+      "step": 31930
+    },
+    {
+      "epoch": 28.723021582733814,
+      "grad_norm": 0.44208401441574097,
+      "learning_rate": 4.879032928882896e-05,
+      "loss": 0.0177,
+      "step": 31940
+    },
+    {
+      "action_loss": 0.0029375366866588593,
+      "epoch": 28.723021582733814,
+      "step": 31940
+    },
+    {
+      "epoch": 28.723021582733814,
+      "step": 31940,
+      "torque_loss": 0.1289682239294052
+    },
+    {
+      "epoch": 28.723021582733814,
+      "force_loss": 0.0037312351632863283,
+      "step": 31940
+    },
+    {
+      "epoch": 28.73201438848921,
+      "grad_norm": 0.14763537049293518,
+      "learning_rate": 4.876277970993505e-05,
+      "loss": 0.0171,
+      "step": 31950
+    },
+    {
+      "action_loss": 0.007328465580940247,
+      "epoch": 28.73201438848921,
+      "step": 31950
+    },
+    {
+      "epoch": 28.73201438848921,
+      "step": 31950,
+      "torque_loss": 0.19901573657989502
+    },
+    {
+      "epoch": 28.73201438848921,
+      "force_loss": 0.008131143637001514,
+      "step": 31950
+    },
+    {
+      "epoch": 28.741007194244606,
+      "grad_norm": 0.1703796088695526,
+      "learning_rate": 4.873523050687602e-05,
+      "loss": 0.0209,
+      "step": 31960
+    },
+    {
+      "action_loss": 0.003658779663965106,
+      "epoch": 28.741007194244606,
+      "step": 31960
+    },
+    {
+      "epoch": 28.741007194244606,
+      "step": 31960,
+      "torque_loss": 0.11144045740365982
+    },
+    {
+      "epoch": 28.741007194244606,
+      "force_loss": 0.002541125286370516,
+      "step": 31960
+    },
+    {
+      "epoch": 28.75,
+      "grad_norm": 0.12297157943248749,
+      "learning_rate": 4.870768168802056e-05,
+      "loss": 0.0169,
+      "step": 31970
+    },
+    {
+      "action_loss": 0.0037621979136019945,
+      "epoch": 28.75,
+      "step": 31970
+    },
+    {
+      "epoch": 28.75,
+      "step": 31970,
+      "torque_loss": 0.12326759099960327
+    },
+    {
+      "epoch": 28.75,
+      "force_loss": 0.0085930610075593,
+      "step": 31970
+    },
+    {
+      "epoch": 28.758992805755394,
+      "grad_norm": 0.192536398768425,
+      "learning_rate": 4.868013326173728e-05,
+      "loss": 0.0177,
+      "step": 31980
+    },
+    {
+      "action_loss": 0.003589194267988205,
+      "epoch": 28.758992805755394,
+      "step": 31980
+    },
+    {
+      "epoch": 28.758992805755394,
+      "step": 31980,
+      "torque_loss": 0.11283278465270996
+    },
+    {
+      "epoch": 28.758992805755394,
+      "force_loss": 0.010618864558637142,
+      "step": 31980
+    },
+    {
+      "epoch": 28.76798561151079,
+      "grad_norm": 0.3223935663700104,
+      "learning_rate": 4.865258523639468e-05,
+      "loss": 0.0172,
+      "step": 31990
+    },
+    {
+      "action_loss": 0.0027538712602108717,
+      "epoch": 28.76798561151079,
+      "step": 31990
+    },
+    {
+      "epoch": 28.76798561151079,
+      "step": 31990,
+      "torque_loss": 0.15496884286403656
+    },
+    {
+      "epoch": 28.76798561151079,
+      "force_loss": 0.003215826814994216,
+      "step": 31990
+    },
+    {
+      "epoch": 28.776978417266186,
+      "grad_norm": 0.18327292799949646,
+      "learning_rate": 4.862503762036109e-05,
+      "loss": 0.0189,
+      "step": 32000
+    },
+    {
+      "action_loss": 0.012443861924111843,
+      "epoch": 28.776978417266186,
+      "step": 32000
+    },
+    {
+      "epoch": 28.776978417266186,
+      "step": 32000,
+      "torque_loss": 0.13682183623313904
+    },
+    {
+      "epoch": 28.776978417266186,
+      "force_loss": 0.010915939696133137,
+      "step": 32000
+    },
+    {
+      "epoch": 28.785971223021583,
+      "grad_norm": 0.13749296963214874,
+      "learning_rate": 4.859749042200478e-05,
+      "loss": 0.0184,
+      "step": 32010
+    },
+    {
+      "action_loss": 0.0048415870405733585,
+      "epoch": 28.785971223021583,
+      "step": 32010
+    },
+    {
+      "epoch": 28.785971223021583,
+      "step": 32010,
+      "torque_loss": 0.1794721633195877
+    },
+    {
+      "epoch": 28.785971223021583,
+      "force_loss": 0.007957744412124157,
+      "step": 32010
+    },
+    {
+      "epoch": 28.794964028776977,
+      "grad_norm": 0.16890855133533478,
+      "learning_rate": 4.856994364969384e-05,
+      "loss": 0.0186,
+      "step": 32020
+    },
+    {
+      "action_loss": 0.0025698344688862562,
+      "epoch": 28.794964028776977,
+      "step": 32020
+    },
+    {
+      "epoch": 28.794964028776977,
+      "step": 32020,
+      "torque_loss": 0.11772950738668442
+    },
+    {
+      "epoch": 28.794964028776977,
+      "force_loss": 0.0036689986009150743,
+      "step": 32020
+    },
+    {
+      "epoch": 28.803956834532375,
+      "grad_norm": 0.11281709372997284,
+      "learning_rate": 4.854239731179625e-05,
+      "loss": 0.0153,
+      "step": 32030
+    },
+    {
+      "action_loss": 0.0019129790598526597,
+      "epoch": 28.803956834532375,
+      "step": 32030
+    },
+    {
+      "epoch": 28.803956834532375,
+      "step": 32030,
+      "torque_loss": 0.1618196964263916
+    },
+    {
+      "epoch": 28.803956834532375,
+      "force_loss": 0.0015649019042029977,
+      "step": 32030
+    },
+    {
+      "epoch": 28.81294964028777,
+      "grad_norm": 0.2970485985279083,
+      "learning_rate": 4.85148514166799e-05,
+      "loss": 0.0177,
+      "step": 32040
+    },
+    {
+      "action_loss": 0.0009857051772996783,
+      "epoch": 28.81294964028777,
+      "step": 32040
+    },
+    {
+      "epoch": 28.81294964028777,
+      "step": 32040,
+      "torque_loss": 0.10178893059492111
+    },
+    {
+      "epoch": 28.81294964028777,
+      "force_loss": 0.0007166683790273964,
+      "step": 32040
+    },
+    {
+      "epoch": 28.821942446043167,
+      "grad_norm": 0.17108264565467834,
+      "learning_rate": 4.8487305972712456e-05,
+      "loss": 0.0178,
+      "step": 32050
+    },
+    {
+      "action_loss": 0.007683031260967255,
+      "epoch": 28.821942446043167,
+      "step": 32050
+    },
+    {
+      "epoch": 28.821942446043167,
+      "step": 32050,
+      "torque_loss": 0.1575421839952469
+    },
+    {
+      "epoch": 28.821942446043167,
+      "force_loss": 0.007862170226871967,
+      "step": 32050
+    },
+    {
+      "epoch": 28.83093525179856,
+      "grad_norm": 0.24892657995224,
+      "learning_rate": 4.8459760988261526e-05,
+      "loss": 0.0165,
+      "step": 32060
+    },
+    {
+      "action_loss": 0.0024876922834664583,
+      "epoch": 28.83093525179856,
+      "step": 32060
+    },
+    {
+      "epoch": 28.83093525179856,
+      "step": 32060,
+      "torque_loss": 0.14223970472812653
+    },
+    {
+      "epoch": 28.83093525179856,
+      "force_loss": 0.001781096332706511,
+      "step": 32060
+    },
+    {
+      "epoch": 28.83992805755396,
+      "grad_norm": 0.1431387960910797,
+      "learning_rate": 4.843221647169453e-05,
+      "loss": 0.0207,
+      "step": 32070
+    },
+    {
+      "action_loss": 0.005526240915060043,
+      "epoch": 28.83992805755396,
+      "step": 32070
+    },
+    {
+      "epoch": 28.83992805755396,
+      "step": 32070,
+      "torque_loss": 0.10630917549133301
+    },
+    {
+      "epoch": 28.83992805755396,
+      "force_loss": 0.007075069472193718,
+      "step": 32070
+    },
+    {
+      "epoch": 28.848920863309353,
+      "grad_norm": 0.17426855862140656,
+      "learning_rate": 4.840467243137878e-05,
+      "loss": 0.0206,
+      "step": 32080
+    },
+    {
+      "action_loss": 0.002417096635326743,
+      "epoch": 28.848920863309353,
+      "step": 32080
+    },
+    {
+      "epoch": 28.848920863309353,
+      "step": 32080,
+      "torque_loss": 0.13800916075706482
+    },
+    {
+      "epoch": 28.848920863309353,
+      "force_loss": 0.0017435640329495072,
+      "step": 32080
+    },
+    {
+      "epoch": 28.857913669064747,
+      "grad_norm": 0.3956768214702606,
+      "learning_rate": 4.837712887568143e-05,
+      "loss": 0.0209,
+      "step": 32090
+    },
+    {
+      "action_loss": 0.004531675949692726,
+      "epoch": 28.857913669064747,
+      "step": 32090
+    },
+    {
+      "epoch": 28.857913669064747,
+      "step": 32090,
+      "torque_loss": 0.13119183480739594
+    },
+    {
+      "epoch": 28.857913669064747,
+      "force_loss": 0.006331259850412607,
+      "step": 32090
+    },
+    {
+      "epoch": 28.866906474820144,
+      "grad_norm": 0.24706391990184784,
+      "learning_rate": 4.8349585812969464e-05,
+      "loss": 0.0191,
+      "step": 32100
+    },
+    {
+      "action_loss": 0.006737230345606804,
+      "epoch": 28.866906474820144,
+      "step": 32100
+    },
+    {
+      "epoch": 28.866906474820144,
+      "step": 32100,
+      "torque_loss": 0.08923038095235825
+    },
+    {
+      "epoch": 28.866906474820144,
+      "force_loss": 0.006674828007817268,
+      "step": 32100
+    },
+    {
+      "epoch": 28.87589928057554,
+      "grad_norm": 0.19916436076164246,
+      "learning_rate": 4.8322043251609775e-05,
+      "loss": 0.0185,
+      "step": 32110
+    },
+    {
+      "action_loss": 0.001197325182147324,
+      "epoch": 28.87589928057554,
+      "step": 32110
+    },
+    {
+      "epoch": 28.87589928057554,
+      "step": 32110,
+      "torque_loss": 0.13352930545806885
+    },
+    {
+      "epoch": 28.87589928057554,
+      "force_loss": 0.0032023098319768906,
+      "step": 32110
+    },
+    {
+      "epoch": 28.884892086330936,
+      "grad_norm": 0.293727308511734,
+      "learning_rate": 4.8294501199969015e-05,
+      "loss": 0.0178,
+      "step": 32120
+    },
+    {
+      "action_loss": 0.0013039332116022706,
+      "epoch": 28.884892086330936,
+      "step": 32120
+    },
+    {
+      "epoch": 28.884892086330936,
+      "step": 32120,
+      "torque_loss": 0.09936841577291489
+    },
+    {
+      "epoch": 28.884892086330936,
+      "force_loss": 0.002565234201028943,
+      "step": 32120
+    },
+    {
+      "epoch": 28.89388489208633,
+      "grad_norm": 0.1831299513578415,
+      "learning_rate": 4.826695966641376e-05,
+      "loss": 0.0174,
+      "step": 32130
+    },
+    {
+      "action_loss": 0.0028378423303365707,
+      "epoch": 28.89388489208633,
+      "step": 32130
+    },
+    {
+      "epoch": 28.89388489208633,
+      "step": 32130,
+      "torque_loss": 0.1480691134929657
+    },
+    {
+      "epoch": 28.89388489208633,
+      "force_loss": 0.0015282105887308717,
+      "step": 32130
+    },
+    {
+      "epoch": 28.902877697841728,
+      "grad_norm": 0.27493205666542053,
+      "learning_rate": 4.823941865931043e-05,
+      "loss": 0.0192,
+      "step": 32140
+    },
+    {
+      "action_loss": 0.0119972825050354,
+      "epoch": 28.902877697841728,
+      "step": 32140
+    },
+    {
+      "epoch": 28.902877697841728,
+      "step": 32140,
+      "torque_loss": 0.16558034718036652
+    },
+    {
+      "epoch": 28.902877697841728,
+      "force_loss": 0.013252168893814087,
+      "step": 32140
+    },
+    {
+      "epoch": 28.91187050359712,
+      "grad_norm": 0.27799856662750244,
+      "learning_rate": 4.82118781870252e-05,
+      "loss": 0.0218,
+      "step": 32150
+    },
+    {
+      "action_loss": 0.0039338041096925735,
+      "epoch": 28.91187050359712,
+      "step": 32150
+    },
+    {
+      "epoch": 28.91187050359712,
+      "step": 32150,
+      "torque_loss": 0.1132541075348854
+    },
+    {
+      "epoch": 28.91187050359712,
+      "force_loss": 0.011024055071175098,
+      "step": 32150
+    },
+    {
+      "epoch": 28.92086330935252,
+      "grad_norm": 0.4246889650821686,
+      "learning_rate": 4.8184338257924185e-05,
+      "loss": 0.0188,
+      "step": 32160
+    },
+    {
+      "action_loss": 0.01281951367855072,
+      "epoch": 28.92086330935252,
+      "step": 32160
+    },
+    {
+      "epoch": 28.92086330935252,
+      "step": 32160,
+      "torque_loss": 0.14115017652511597
+    },
+    {
+      "epoch": 28.92086330935252,
+      "force_loss": 0.010982475243508816,
+      "step": 32160
+    },
+    {
+      "epoch": 28.929856115107913,
+      "grad_norm": 0.3644884526729584,
+      "learning_rate": 4.815679888037324e-05,
+      "loss": 0.0215,
+      "step": 32170
+    },
+    {
+      "action_loss": 0.001956092892214656,
+      "epoch": 28.929856115107913,
+      "step": 32170
+    },
+    {
+      "epoch": 28.929856115107913,
+      "step": 32170,
+      "torque_loss": 0.11333179473876953
+    },
+    {
+      "epoch": 28.929856115107913,
+      "force_loss": 0.0022038861643522978,
+      "step": 32170
+    },
+    {
+      "epoch": 28.93884892086331,
+      "grad_norm": 0.2975490093231201,
+      "learning_rate": 4.8129260062738135e-05,
+      "loss": 0.0165,
+      "step": 32180
+    },
+    {
+      "action_loss": 0.0011761150090023875,
+      "epoch": 28.93884892086331,
+      "step": 32180
+    },
+    {
+      "epoch": 28.93884892086331,
+      "step": 32180,
+      "torque_loss": 0.0744553878903389
+    },
+    {
+      "epoch": 28.93884892086331,
+      "force_loss": 0.0010741964215412736,
+      "step": 32180
+    },
+    {
+      "epoch": 28.947841726618705,
+      "grad_norm": 0.2999323606491089,
+      "learning_rate": 4.810172181338445e-05,
+      "loss": 0.0182,
+      "step": 32190
+    },
+    {
+      "action_loss": 0.023830803111195564,
+      "epoch": 28.947841726618705,
+      "step": 32190
+    },
+    {
+      "epoch": 28.947841726618705,
+      "step": 32190,
+      "torque_loss": 0.10264798998832703
+    },
+    {
+      "epoch": 28.947841726618705,
+      "force_loss": 0.03188973665237427,
+      "step": 32190
+    },
+    {
+      "epoch": 28.9568345323741,
+      "grad_norm": 0.1338382512331009,
+      "learning_rate": 4.807418414067753e-05,
+      "loss": 0.0183,
+      "step": 32200
+    },
+    {
+      "action_loss": 0.0027587839867919683,
+      "epoch": 28.9568345323741,
+      "step": 32200
+    },
+    {
+      "epoch": 28.9568345323741,
+      "step": 32200,
+      "torque_loss": 0.10562457889318466
+    },
+    {
+      "epoch": 28.9568345323741,
+      "force_loss": 0.002024925546720624,
+      "step": 32200
+    },
+    {
+      "epoch": 28.965827338129497,
+      "grad_norm": 0.13481660187244415,
+      "learning_rate": 4.804664705298264e-05,
+      "loss": 0.0214,
+      "step": 32210
+    },
+    {
+      "action_loss": 0.0025620502419769764,
+      "epoch": 28.965827338129497,
+      "step": 32210
+    },
+    {
+      "epoch": 28.965827338129497,
+      "step": 32210,
+      "torque_loss": 0.12761642038822174
+    },
+    {
+      "epoch": 28.965827338129497,
+      "force_loss": 0.004373255651444197,
+      "step": 32210
+    },
+    {
+      "epoch": 28.97482014388489,
+      "grad_norm": 0.2162230908870697,
+      "learning_rate": 4.80191105586648e-05,
+      "loss": 0.0172,
+      "step": 32220
+    },
+    {
+      "action_loss": 0.002540873596444726,
+      "epoch": 28.97482014388489,
+      "step": 32220
+    },
+    {
+      "epoch": 28.97482014388489,
+      "step": 32220,
+      "torque_loss": 0.16925716400146484
+    },
+    {
+      "epoch": 28.97482014388489,
+      "force_loss": 0.002540864981710911,
+      "step": 32220
+    },
+    {
+      "epoch": 28.98381294964029,
+      "grad_norm": 0.4287576973438263,
+      "learning_rate": 4.799157466608886e-05,
+      "loss": 0.0177,
+      "step": 32230
+    },
+    {
+      "action_loss": 0.005182547960430384,
+      "epoch": 28.98381294964029,
+      "step": 32230
+    },
+    {
+      "epoch": 28.98381294964029,
+      "step": 32230,
+      "torque_loss": 0.15841057896614075
+    },
+    {
+      "epoch": 28.98381294964029,
+      "force_loss": 0.002919477177783847,
+      "step": 32230
+    },
+    {
+      "epoch": 28.992805755395683,
+      "grad_norm": 0.10567975044250488,
+      "learning_rate": 4.796403938361951e-05,
+      "loss": 0.0198,
+      "step": 32240
+    },
+    {
+      "action_loss": 0.0012668374693021178,
+      "epoch": 28.992805755395683,
+      "step": 32240
+    },
+    {
+      "epoch": 28.992805755395683,
+      "step": 32240,
+      "torque_loss": 0.11107317358255386
+    },
+    {
+      "epoch": 28.992805755395683,
+      "force_loss": 0.001408244133926928,
+      "step": 32240
+    },
+    {
+      "epoch": 29.00179856115108,
+      "grad_norm": 0.41919589042663574,
+      "learning_rate": 4.793650471962123e-05,
+      "loss": 0.0184,
+      "step": 32250
+    },
+    {
+      "action_loss": 0.0016538627678528428,
+      "epoch": 29.00179856115108,
+      "step": 32250
+    },
+    {
+      "epoch": 29.00179856115108,
+      "step": 32250,
+      "torque_loss": 0.09937354922294617
+    },
+    {
+      "epoch": 29.00179856115108,
+      "force_loss": 0.0019216047367081046,
+      "step": 32250
+    },
+    {
+      "epoch": 29.010791366906474,
+      "grad_norm": 0.27208903431892395,
+      "learning_rate": 4.790897068245835e-05,
+      "loss": 0.0183,
+      "step": 32260
+    },
+    {
+      "action_loss": 0.007077125366777182,
+      "epoch": 29.010791366906474,
+      "step": 32260
+    },
+    {
+      "epoch": 29.010791366906474,
+      "step": 32260,
+      "torque_loss": 0.09648188948631287
+    },
+    {
+      "epoch": 29.010791366906474,
+      "force_loss": 0.00590849993750453,
+      "step": 32260
+    },
+    {
+      "epoch": 29.019784172661872,
+      "grad_norm": 0.14177268743515015,
+      "learning_rate": 4.7881437280494954e-05,
+      "loss": 0.0174,
+      "step": 32270
+    },
+    {
+      "action_loss": 0.0015265486435964704,
+      "epoch": 29.019784172661872,
+      "step": 32270
+    },
+    {
+      "epoch": 29.019784172661872,
+      "step": 32270,
+      "torque_loss": 0.10215165466070175
+    },
+    {
+      "epoch": 29.019784172661872,
+      "force_loss": 0.0038197848480194807,
+      "step": 32270
+    },
+    {
+      "epoch": 29.028776978417266,
+      "grad_norm": 0.23310257494449615,
+      "learning_rate": 4.7853904522094965e-05,
+      "loss": 0.0214,
+      "step": 32280
+    },
+    {
+      "action_loss": 0.0017850763397291303,
+      "epoch": 29.028776978417266,
+      "step": 32280
+    },
+    {
+      "epoch": 29.028776978417266,
+      "step": 32280,
+      "torque_loss": 0.10656794160604477
+    },
+    {
+      "epoch": 29.028776978417266,
+      "force_loss": 0.001972500467672944,
+      "step": 32280
+    },
+    {
+      "epoch": 29.037769784172664,
+      "grad_norm": 0.21488600969314575,
+      "learning_rate": 4.782637241562215e-05,
+      "loss": 0.0199,
+      "step": 32290
+    },
+    {
+      "action_loss": 0.003017843933776021,
+      "epoch": 29.037769784172664,
+      "step": 32290
+    },
+    {
+      "epoch": 29.037769784172664,
+      "step": 32290,
+      "torque_loss": 0.11039341241121292
+    },
+    {
+      "epoch": 29.037769784172664,
+      "force_loss": 0.008591358549892902,
+      "step": 32290
+    },
+    {
+      "epoch": 29.046762589928058,
+      "grad_norm": 0.18550518155097961,
+      "learning_rate": 4.779884096943997e-05,
+      "loss": 0.0172,
+      "step": 32300
+    },
+    {
+      "action_loss": 0.0059019471518695354,
+      "epoch": 29.046762589928058,
+      "step": 32300
+    },
+    {
+      "epoch": 29.046762589928058,
+      "step": 32300,
+      "torque_loss": 0.1212267205119133
+    },
+    {
+      "epoch": 29.046762589928058,
+      "force_loss": 0.01663847826421261,
+      "step": 32300
+    },
+    {
+      "epoch": 29.055755395683452,
+      "grad_norm": 0.45165735483169556,
+      "learning_rate": 4.777131019191182e-05,
+      "loss": 0.0277,
+      "step": 32310
+    },
+    {
+      "action_loss": 0.0054965815506875515,
+      "epoch": 29.055755395683452,
+      "step": 32310
+    },
+    {
+      "epoch": 29.055755395683452,
+      "step": 32310,
+      "torque_loss": 0.10573756694793701
+    },
+    {
+      "epoch": 29.055755395683452,
+      "force_loss": 0.0016368599608540535,
+      "step": 32310
+    },
+    {
+      "epoch": 29.06474820143885,
+      "grad_norm": 0.2798215448856354,
+      "learning_rate": 4.774378009140076e-05,
+      "loss": 0.0189,
+      "step": 32320
+    },
+    {
+      "action_loss": 0.01228462252765894,
+      "epoch": 29.06474820143885,
+      "step": 32320
+    },
+    {
+      "epoch": 29.06474820143885,
+      "step": 32320,
+      "torque_loss": 0.10527908802032471
+    },
+    {
+      "epoch": 29.06474820143885,
+      "force_loss": 0.010610915720462799,
+      "step": 32320
+    },
+    {
+      "epoch": 29.073741007194243,
+      "grad_norm": 0.4051395654678345,
+      "learning_rate": 4.7716250676269735e-05,
+      "loss": 0.018,
+      "step": 32330
+    },
+    {
+      "action_loss": 0.004625875037163496,
+      "epoch": 29.073741007194243,
+      "step": 32330
+    },
+    {
+      "epoch": 29.073741007194243,
+      "step": 32330,
+      "torque_loss": 0.12205108255147934
+    },
+    {
+      "epoch": 29.073741007194243,
+      "force_loss": 0.004206816665828228,
+      "step": 32330
+    },
+    {
+      "epoch": 29.08273381294964,
+      "grad_norm": 0.1723063588142395,
+      "learning_rate": 4.7688721954881485e-05,
+      "loss": 0.0186,
+      "step": 32340
+    },
+    {
+      "action_loss": 0.0039085340686142445,
+      "epoch": 29.08273381294964,
+      "step": 32340
+    },
+    {
+      "epoch": 29.08273381294964,
+      "step": 32340,
+      "torque_loss": 0.18239738047122955
+    },
+    {
+      "epoch": 29.08273381294964,
+      "force_loss": 0.002797734225168824,
+      "step": 32340
+    },
+    {
+      "epoch": 29.091726618705035,
+      "grad_norm": 0.2537984848022461,
+      "learning_rate": 4.7661193935598446e-05,
+      "loss": 0.0277,
+      "step": 32350
+    },
+    {
+      "action_loss": 0.004381152801215649,
+      "epoch": 29.091726618705035,
+      "step": 32350
+    },
+    {
+      "epoch": 29.091726618705035,
+      "step": 32350,
+      "torque_loss": 0.095862478017807
+    },
+    {
+      "epoch": 29.091726618705035,
+      "force_loss": 0.006126149091869593,
+      "step": 32350
+    },
+    {
+      "epoch": 29.100719424460433,
+      "grad_norm": 0.2501719295978546,
+      "learning_rate": 4.763366662678296e-05,
+      "loss": 0.019,
+      "step": 32360
+    },
+    {
+      "action_loss": 0.004788230173289776,
+      "epoch": 29.100719424460433,
+      "step": 32360
+    },
+    {
+      "epoch": 29.100719424460433,
+      "step": 32360,
+      "torque_loss": 0.11477348953485489
+    },
+    {
+      "epoch": 29.100719424460433,
+      "force_loss": 0.003308061510324478,
+      "step": 32360
+    },
+    {
+      "epoch": 29.109712230215827,
+      "grad_norm": 0.30460652709007263,
+      "learning_rate": 4.7606140036797064e-05,
+      "loss": 0.0206,
+      "step": 32370
+    },
+    {
+      "action_loss": 0.0032112309709191322,
+      "epoch": 29.109712230215827,
+      "step": 32370
+    },
+    {
+      "epoch": 29.109712230215827,
+      "step": 32370,
+      "torque_loss": 0.15254874527454376
+    },
+    {
+      "epoch": 29.109712230215827,
+      "force_loss": 0.008447595871984959,
+      "step": 32370
+    },
+    {
+      "epoch": 29.118705035971225,
+      "grad_norm": 0.23224332928657532,
+      "learning_rate": 4.7578614174002614e-05,
+      "loss": 0.0178,
+      "step": 32380
+    },
+    {
+      "action_loss": 0.003697851672768593,
+      "epoch": 29.118705035971225,
+      "step": 32380
+    },
+    {
+      "epoch": 29.118705035971225,
+      "step": 32380,
+      "torque_loss": 0.15930971503257751
+    },
+    {
+      "epoch": 29.118705035971225,
+      "force_loss": 0.006726616993546486,
+      "step": 32380
+    },
+    {
+      "epoch": 29.12769784172662,
+      "grad_norm": 0.2128390222787857,
+      "learning_rate": 4.755108904676125e-05,
+      "loss": 0.0173,
+      "step": 32390
+    },
+    {
+      "action_loss": 0.007075525354593992,
+      "epoch": 29.12769784172662,
+      "step": 32390
+    },
+    {
+      "epoch": 29.12769784172662,
+      "step": 32390,
+      "torque_loss": 0.14321278035640717
+    },
+    {
+      "epoch": 29.12769784172662,
+      "force_loss": 0.012733793817460537,
+      "step": 32390
+    },
+    {
+      "epoch": 29.136690647482013,
+      "grad_norm": 0.17491494119167328,
+      "learning_rate": 4.752356466343436e-05,
+      "loss": 0.0243,
+      "step": 32400
+    },
+    {
+      "action_loss": 0.0029850397258996964,
+      "epoch": 29.136690647482013,
+      "step": 32400
+    },
+    {
+      "epoch": 29.136690647482013,
+      "step": 32400,
+      "torque_loss": 0.11526184529066086
+    },
+    {
+      "epoch": 29.136690647482013,
+      "force_loss": 0.005373217165470123,
+      "step": 32400
+    },
+    {
+      "epoch": 29.14568345323741,
+      "grad_norm": 0.140650674700737,
+      "learning_rate": 4.7496041032383174e-05,
+      "loss": 0.0174,
+      "step": 32410
+    },
+    {
+      "action_loss": 0.0027762267272919416,
+      "epoch": 29.14568345323741,
+      "step": 32410
+    },
+    {
+      "epoch": 29.14568345323741,
+      "step": 32410,
+      "torque_loss": 0.08548658341169357
+    },
+    {
+      "epoch": 29.14568345323741,
+      "force_loss": 0.0017282020999118686,
+      "step": 32410
+    },
+    {
+      "epoch": 29.154676258992804,
+      "grad_norm": 0.521045982837677,
+      "learning_rate": 4.746851816196858e-05,
+      "loss": 0.0161,
+      "step": 32420
+    },
+    {
+      "action_loss": 0.00275931297801435,
+      "epoch": 29.154676258992804,
+      "step": 32420
+    },
+    {
+      "epoch": 29.154676258992804,
+      "step": 32420,
+      "torque_loss": 0.11880142241716385
+    },
+    {
+      "epoch": 29.154676258992804,
+      "force_loss": 0.0073845237493515015,
+      "step": 32420
+    },
+    {
+      "epoch": 29.163669064748202,
+      "grad_norm": 0.2052893489599228,
+      "learning_rate": 4.744099606055135e-05,
+      "loss": 0.0176,
+      "step": 32430
+    },
+    {
+      "action_loss": 0.015046007931232452,
+      "epoch": 29.163669064748202,
+      "step": 32430
+    },
+    {
+      "epoch": 29.163669064748202,
+      "step": 32430,
+      "torque_loss": 0.1811208575963974
+    },
+    {
+      "epoch": 29.163669064748202,
+      "force_loss": 0.014213413000106812,
+      "step": 32430
+    },
+    {
+      "epoch": 29.172661870503596,
+      "grad_norm": 0.27192258834838867,
+      "learning_rate": 4.741347473649193e-05,
+      "loss": 0.0193,
+      "step": 32440
+    },
+    {
+      "action_loss": 0.005661696195602417,
+      "epoch": 29.172661870503596,
+      "step": 32440
+    },
+    {
+      "epoch": 29.172661870503596,
+      "step": 32440,
+      "torque_loss": 0.0931566059589386
+    },
+    {
+      "epoch": 29.172661870503596,
+      "force_loss": 0.0030042286962270737,
+      "step": 32440
+    },
+    {
+      "epoch": 29.181654676258994,
+      "grad_norm": 0.10476517677307129,
+      "learning_rate": 4.738595419815058e-05,
+      "loss": 0.0175,
+      "step": 32450
+    },
+    {
+      "action_loss": 0.005921449512243271,
+      "epoch": 29.181654676258994,
+      "step": 32450
+    },
+    {
+      "epoch": 29.181654676258994,
+      "step": 32450,
+      "torque_loss": 0.10327693819999695
+    },
+    {
+      "epoch": 29.181654676258994,
+      "force_loss": 0.009495220147073269,
+      "step": 32450
+    },
+    {
+      "epoch": 29.190647482014388,
+      "grad_norm": 0.22602613270282745,
+      "learning_rate": 4.7358434453887365e-05,
+      "loss": 0.0197,
+      "step": 32460
+    },
+    {
+      "action_loss": 0.008681467734277248,
+      "epoch": 29.190647482014388,
+      "step": 32460
+    },
+    {
+      "epoch": 29.190647482014388,
+      "step": 32460,
+      "torque_loss": 0.07998576760292053
+    },
+    {
+      "epoch": 29.190647482014388,
+      "force_loss": 0.0065096099860966206,
+      "step": 32460
+    },
+    {
+      "epoch": 29.199640287769785,
+      "grad_norm": 0.2433185875415802,
+      "learning_rate": 4.7330915512061976e-05,
+      "loss": 0.0184,
+      "step": 32470
+    },
+    {
+      "action_loss": 0.003693939186632633,
+      "epoch": 29.199640287769785,
+      "step": 32470
+    },
+    {
+      "epoch": 29.199640287769785,
+      "step": 32470,
+      "torque_loss": 0.10094495862722397
+    },
+    {
+      "epoch": 29.199640287769785,
+      "force_loss": 0.007911822758615017,
+      "step": 32470
+    },
+    {
+      "epoch": 29.20863309352518,
+      "grad_norm": 0.2534589469432831,
+      "learning_rate": 4.730339738103402e-05,
+      "loss": 0.0224,
+      "step": 32480
+    },
+    {
+      "action_loss": 0.004707122687250376,
+      "epoch": 29.20863309352518,
+      "step": 32480
+    },
+    {
+      "epoch": 29.20863309352518,
+      "step": 32480,
+      "torque_loss": 0.132128044962883
+    },
+    {
+      "epoch": 29.20863309352518,
+      "force_loss": 0.00502332067117095,
+      "step": 32480
+    },
+    {
+      "epoch": 29.217625899280577,
+      "grad_norm": 0.2297942191362381,
+      "learning_rate": 4.727588006916271e-05,
+      "loss": 0.0214,
+      "step": 32490
+    },
+    {
+      "action_loss": 0.002568019786849618,
+      "epoch": 29.217625899280577,
+      "step": 32490
+    },
+    {
+      "epoch": 29.217625899280577,
+      "step": 32490,
+      "torque_loss": 0.11688452214002609
+    },
+    {
+      "epoch": 29.217625899280577,
+      "force_loss": 0.002252336358651519,
+      "step": 32490
+    },
+    {
+      "epoch": 29.22661870503597,
+      "grad_norm": 0.2703711688518524,
+      "learning_rate": 4.724836358480711e-05,
+      "loss": 0.0169,
+      "step": 32500
+    },
+    {
+      "action_loss": 0.01019375305622816,
+      "epoch": 29.22661870503597,
+      "step": 32500
+    },
+    {
+      "epoch": 29.22661870503597,
+      "step": 32500,
+      "torque_loss": 0.1894327849149704
+    },
+    {
+      "epoch": 29.22661870503597,
+      "force_loss": 0.02001328580081463,
+      "step": 32500
+    },
+    {
+      "epoch": 29.235611510791365,
+      "grad_norm": 0.2101975381374359,
+      "learning_rate": 4.722084793632601e-05,
+      "loss": 0.0195,
+      "step": 32510
+    },
+    {
+      "action_loss": 0.006329949479550123,
+      "epoch": 29.235611510791365,
+      "step": 32510
+    },
+    {
+      "epoch": 29.235611510791365,
+      "step": 32510,
+      "torque_loss": 0.14494441449642181
+    },
+    {
+      "epoch": 29.235611510791365,
+      "force_loss": 0.00831457506865263,
+      "step": 32510
+    },
+    {
+      "epoch": 29.244604316546763,
+      "grad_norm": 0.2567432224750519,
+      "learning_rate": 4.719333313207792e-05,
+      "loss": 0.0197,
+      "step": 32520
+    },
+    {
+      "action_loss": 0.0057547432370483875,
+      "epoch": 29.244604316546763,
+      "step": 32520
+    },
+    {
+      "epoch": 29.244604316546763,
+      "step": 32520,
+      "torque_loss": 0.14106887578964233
+    },
+    {
+      "epoch": 29.244604316546763,
+      "force_loss": 0.004578470718115568,
+      "step": 32520
+    },
+    {
+      "epoch": 29.253597122302157,
+      "grad_norm": 0.27158123254776,
+      "learning_rate": 4.716581918042114e-05,
+      "loss": 0.0194,
+      "step": 32530
+    },
+    {
+      "action_loss": 0.008752552792429924,
+      "epoch": 29.253597122302157,
+      "step": 32530
+    },
+    {
+      "epoch": 29.253597122302157,
+      "step": 32530,
+      "torque_loss": 0.1290217787027359
+    },
+    {
+      "epoch": 29.253597122302157,
+      "force_loss": 0.014507760293781757,
+      "step": 32530
+    },
+    {
+      "epoch": 29.262589928057555,
+      "grad_norm": 0.17042985558509827,
+      "learning_rate": 4.7138306089713636e-05,
+      "loss": 0.0172,
+      "step": 32540
+    },
+    {
+      "action_loss": 0.0013127584243193269,
+      "epoch": 29.262589928057555,
+      "step": 32540
+    },
+    {
+      "epoch": 29.262589928057555,
+      "step": 32540,
+      "torque_loss": 0.07442139834165573
+    },
+    {
+      "epoch": 29.262589928057555,
+      "force_loss": 0.001402603113092482,
+      "step": 32540
+    },
+    {
+      "epoch": 29.27158273381295,
+      "grad_norm": 0.14285852015018463,
+      "learning_rate": 4.7110793868313183e-05,
+      "loss": 0.0153,
+      "step": 32550
+    },
+    {
+      "action_loss": 0.00603299168869853,
+      "epoch": 29.27158273381295,
+      "step": 32550
+    },
+    {
+      "epoch": 29.27158273381295,
+      "step": 32550,
+      "torque_loss": 0.09963829070329666
+    },
+    {
+      "epoch": 29.27158273381295,
+      "force_loss": 0.0076375603675842285,
+      "step": 32550
+    },
+    {
+      "epoch": 29.280575539568346,
+      "grad_norm": 0.513832688331604,
+      "learning_rate": 4.708328252457729e-05,
+      "loss": 0.0185,
+      "step": 32560
+    },
+    {
+      "action_loss": 0.011000323109328747,
+      "epoch": 29.280575539568346,
+      "step": 32560
+    },
+    {
+      "epoch": 29.280575539568346,
+      "step": 32560,
+      "torque_loss": 0.13765794038772583
+    },
+    {
+      "epoch": 29.280575539568346,
+      "force_loss": 0.015668898820877075,
+      "step": 32560
+    },
+    {
+      "epoch": 29.28956834532374,
+      "grad_norm": 0.18672128021717072,
+      "learning_rate": 4.7055772066863135e-05,
+      "loss": 0.0236,
+      "step": 32570
+    },
+    {
+      "action_loss": 0.00420399708673358,
+      "epoch": 29.28956834532374,
+      "step": 32570
+    },
+    {
+      "epoch": 29.28956834532374,
+      "step": 32570,
+      "torque_loss": 0.1491629034280777
+    },
+    {
+      "epoch": 29.28956834532374,
+      "force_loss": 0.005232276860624552,
+      "step": 32570
+    },
+    {
+      "epoch": 29.298561151079138,
+      "grad_norm": 0.1859084963798523,
+      "learning_rate": 4.702826250352771e-05,
+      "loss": 0.0161,
+      "step": 32580
+    },
+    {
+      "action_loss": 0.005574436392635107,
+      "epoch": 29.298561151079138,
+      "step": 32580
+    },
+    {
+      "epoch": 29.298561151079138,
+      "step": 32580,
+      "torque_loss": 0.1271819919347763
+    },
+    {
+      "epoch": 29.298561151079138,
+      "force_loss": 0.009282768703997135,
+      "step": 32580
+    },
+    {
+      "epoch": 29.307553956834532,
+      "grad_norm": 0.2615342140197754,
+      "learning_rate": 4.7000753842927653e-05,
+      "loss": 0.0181,
+      "step": 32590
+    },
+    {
+      "action_loss": 0.0013217715313658118,
+      "epoch": 29.307553956834532,
+      "step": 32590
+    },
+    {
+      "epoch": 29.307553956834532,
+      "step": 32590,
+      "torque_loss": 0.09290742874145508
+    },
+    {
+      "epoch": 29.307553956834532,
+      "force_loss": 0.0027795464266091585,
+      "step": 32590
+    },
+    {
+      "epoch": 29.31654676258993,
+      "grad_norm": 0.35575351119041443,
+      "learning_rate": 4.6973246093419384e-05,
+      "loss": 0.0183,
+      "step": 32600
+    },
+    {
+      "action_loss": 0.0021544245537370443,
+      "epoch": 29.31654676258993,
+      "step": 32600
+    },
+    {
+      "epoch": 29.31654676258993,
+      "step": 32600,
+      "torque_loss": 0.08024931699037552
+    },
+    {
+      "epoch": 29.31654676258993,
+      "force_loss": 0.0024534789845347404,
+      "step": 32600
+    },
+    {
+      "epoch": 29.325539568345324,
+      "grad_norm": 0.31143516302108765,
+      "learning_rate": 4.694573926335906e-05,
+      "loss": 0.021,
+      "step": 32610
+    },
+    {
+      "action_loss": 0.0023949090391397476,
+      "epoch": 29.325539568345324,
+      "step": 32610
+    },
+    {
+      "epoch": 29.325539568345324,
+      "step": 32610,
+      "torque_loss": 0.0861029401421547
+    },
+    {
+      "epoch": 29.325539568345324,
+      "force_loss": 0.0039823721162974834,
+      "step": 32610
+    },
+    {
+      "epoch": 29.334532374100718,
+      "grad_norm": 0.3357939124107361,
+      "learning_rate": 4.6918233361102476e-05,
+      "loss": 0.0172,
+      "step": 32620
+    },
+    {
+      "action_loss": 0.0055252485908567905,
+      "epoch": 29.334532374100718,
+      "step": 32620
+    },
+    {
+      "epoch": 29.334532374100718,
+      "step": 32620,
+      "torque_loss": 0.11591412872076035
+    },
+    {
+      "epoch": 29.334532374100718,
+      "force_loss": 0.008989318273961544,
+      "step": 32620
+    },
+    {
+      "epoch": 29.343525179856115,
+      "grad_norm": 0.5258576273918152,
+      "learning_rate": 4.689072839500525e-05,
+      "loss": 0.0187,
+      "step": 32630
+    },
+    {
+      "action_loss": 0.006293279584497213,
+      "epoch": 29.343525179856115,
+      "step": 32630
+    },
+    {
+      "epoch": 29.343525179856115,
+      "step": 32630,
+      "torque_loss": 0.11771956831216812
+    },
+    {
+      "epoch": 29.343525179856115,
+      "force_loss": 0.0033822599798440933,
+      "step": 32630
+    },
+    {
+      "epoch": 29.35251798561151,
+      "grad_norm": 0.5485567450523376,
+      "learning_rate": 4.6863224373422635e-05,
+      "loss": 0.0181,
+      "step": 32640
+    },
+    {
+      "action_loss": 0.004397273063659668,
+      "epoch": 29.35251798561151,
+      "step": 32640
+    },
+    {
+      "epoch": 29.35251798561151,
+      "step": 32640,
+      "torque_loss": 0.10283096879720688
+    },
+    {
+      "epoch": 29.35251798561151,
+      "force_loss": 0.0037207824643701315,
+      "step": 32640
+    },
+    {
+      "epoch": 29.361510791366907,
+      "grad_norm": 0.09677112102508545,
+      "learning_rate": 4.683572130470962e-05,
+      "loss": 0.0177,
+      "step": 32650
+    },
+    {
+      "action_loss": 0.001985314302146435,
+      "epoch": 29.361510791366907,
+      "step": 32650
+    },
+    {
+      "epoch": 29.361510791366907,
+      "step": 32650,
+      "torque_loss": 0.10480914264917374
+    },
+    {
+      "epoch": 29.361510791366907,
+      "force_loss": 0.0034442029427736998,
+      "step": 32650
+    },
+    {
+      "epoch": 29.3705035971223,
+      "grad_norm": 0.1992332637310028,
+      "learning_rate": 4.680821919722094e-05,
+      "loss": 0.0206,
+      "step": 32660
+    },
+    {
+      "action_loss": 0.011067663319408894,
+      "epoch": 29.3705035971223,
+      "step": 32660
+    },
+    {
+      "epoch": 29.3705035971223,
+      "step": 32660,
+      "torque_loss": 0.12920941412448883
+    },
+    {
+      "epoch": 29.3705035971223,
+      "force_loss": 0.008987017907202244,
+      "step": 32660
+    },
+    {
+      "epoch": 29.3794964028777,
+      "grad_norm": 0.3118537962436676,
+      "learning_rate": 4.6780718059310975e-05,
+      "loss": 0.0217,
+      "step": 32670
+    },
+    {
+      "action_loss": 0.004925935994833708,
+      "epoch": 29.3794964028777,
+      "step": 32670
+    },
+    {
+      "epoch": 29.3794964028777,
+      "step": 32670,
+      "torque_loss": 0.1036626324057579
+    },
+    {
+      "epoch": 29.3794964028777,
+      "force_loss": 0.0030708431731909513,
+      "step": 32670
+    },
+    {
+      "epoch": 29.388489208633093,
+      "grad_norm": 0.16304568946361542,
+      "learning_rate": 4.675321789933389e-05,
+      "loss": 0.0171,
+      "step": 32680
+    },
+    {
+      "action_loss": 0.0061505562625825405,
+      "epoch": 29.388489208633093,
+      "step": 32680
+    },
+    {
+      "epoch": 29.388489208633093,
+      "step": 32680,
+      "torque_loss": 0.10227325558662415
+    },
+    {
+      "epoch": 29.388489208633093,
+      "force_loss": 0.001908316626213491,
+      "step": 32680
+    },
+    {
+      "epoch": 29.39748201438849,
+      "grad_norm": 0.20562444627285004,
+      "learning_rate": 4.6725718725643464e-05,
+      "loss": 0.0151,
+      "step": 32690
+    },
+    {
+      "action_loss": 0.02423258125782013,
+      "epoch": 29.39748201438849,
+      "step": 32690
+    },
+    {
+      "epoch": 29.39748201438849,
+      "step": 32690,
+      "torque_loss": 0.16639037430286407
+    },
+    {
+      "epoch": 29.39748201438849,
+      "force_loss": 0.038983047008514404,
+      "step": 32690
+    },
+    {
+      "epoch": 29.406474820143885,
+      "grad_norm": 0.15895070135593414,
+      "learning_rate": 4.669822054659323e-05,
+      "loss": 0.0209,
+      "step": 32700
+    },
+    {
+      "action_loss": 0.01568661816418171,
+      "epoch": 29.406474820143885,
+      "step": 32700
+    },
+    {
+      "epoch": 29.406474820143885,
+      "step": 32700,
+      "torque_loss": 0.09653004258871078
+    },
+    {
+      "epoch": 29.406474820143885,
+      "force_loss": 0.0159252118319273,
+      "step": 32700
+    },
+    {
+      "epoch": 29.415467625899282,
+      "grad_norm": 0.3581959307193756,
+      "learning_rate": 4.667072337053644e-05,
+      "loss": 0.0184,
+      "step": 32710
+    },
+    {
+      "action_loss": 0.010245249606668949,
+      "epoch": 29.415467625899282,
+      "step": 32710
+    },
+    {
+      "epoch": 29.415467625899282,
+      "step": 32710,
+      "torque_loss": 0.11866069585084915
+    },
+    {
+      "epoch": 29.415467625899282,
+      "force_loss": 0.006856624037027359,
+      "step": 32710
+    },
+    {
+      "epoch": 29.424460431654676,
+      "grad_norm": 0.17700177431106567,
+      "learning_rate": 4.6643227205825965e-05,
+      "loss": 0.0205,
+      "step": 32720
+    },
+    {
+      "action_loss": 0.0032879996579140425,
+      "epoch": 29.424460431654676,
+      "step": 32720
+    },
+    {
+      "epoch": 29.424460431654676,
+      "step": 32720,
+      "torque_loss": 0.11360689997673035
+    },
+    {
+      "epoch": 29.424460431654676,
+      "force_loss": 0.004228583537042141,
+      "step": 32720
+    },
+    {
+      "epoch": 29.43345323741007,
+      "grad_norm": 0.2889029085636139,
+      "learning_rate": 4.6615732060814454e-05,
+      "loss": 0.018,
+      "step": 32730
+    },
+    {
+      "action_loss": 0.003413207596167922,
+      "epoch": 29.43345323741007,
+      "step": 32730
+    },
+    {
+      "epoch": 29.43345323741007,
+      "step": 32730,
+      "torque_loss": 0.15451616048812866
+    },
+    {
+      "epoch": 29.43345323741007,
+      "force_loss": 0.004634483251720667,
+      "step": 32730
+    },
+    {
+      "epoch": 29.442446043165468,
+      "grad_norm": 0.1654779464006424,
+      "learning_rate": 4.658823794385417e-05,
+      "loss": 0.019,
+      "step": 32740
+    },
+    {
+      "action_loss": 0.005410236772149801,
+      "epoch": 29.442446043165468,
+      "step": 32740
+    },
+    {
+      "epoch": 29.442446043165468,
+      "step": 32740,
+      "torque_loss": 0.09852194786071777
+    },
+    {
+      "epoch": 29.442446043165468,
+      "force_loss": 0.0046052876859903336,
+      "step": 32740
+    },
+    {
+      "epoch": 29.451438848920862,
+      "grad_norm": 0.4740029573440552,
+      "learning_rate": 4.6560744863297115e-05,
+      "loss": 0.0191,
+      "step": 32750
+    },
+    {
+      "action_loss": 0.0026829184498637915,
+      "epoch": 29.451438848920862,
+      "step": 32750
+    },
+    {
+      "epoch": 29.451438848920862,
+      "step": 32750,
+      "torque_loss": 0.10991368442773819
+    },
+    {
+      "epoch": 29.451438848920862,
+      "force_loss": 0.002388973021879792,
+      "step": 32750
+    },
+    {
+      "epoch": 29.46043165467626,
+      "grad_norm": 0.39620524644851685,
+      "learning_rate": 4.653325282749498e-05,
+      "loss": 0.0205,
+      "step": 32760
+    },
+    {
+      "action_loss": 0.0012941869208589196,
+      "epoch": 29.46043165467626,
+      "step": 32760
+    },
+    {
+      "epoch": 29.46043165467626,
+      "step": 32760,
+      "torque_loss": 0.13842102885246277
+    },
+    {
+      "epoch": 29.46043165467626,
+      "force_loss": 0.0016389338998124003,
+      "step": 32760
+    },
+    {
+      "epoch": 29.469424460431654,
+      "grad_norm": 0.27129366993904114,
+      "learning_rate": 4.6505761844799075e-05,
+      "loss": 0.019,
+      "step": 32770
+    },
+    {
+      "action_loss": 0.004043588414788246,
+      "epoch": 29.469424460431654,
+      "step": 32770
+    },
+    {
+      "epoch": 29.469424460431654,
+      "step": 32770,
+      "torque_loss": 0.15159566700458527
+    },
+    {
+      "epoch": 29.469424460431654,
+      "force_loss": 0.009783412329852581,
+      "step": 32770
+    },
+    {
+      "epoch": 29.47841726618705,
+      "grad_norm": 0.3575635850429535,
+      "learning_rate": 4.647827192356048e-05,
+      "loss": 0.0191,
+      "step": 32780
+    },
+    {
+      "action_loss": 0.004523263778537512,
+      "epoch": 29.47841726618705,
+      "step": 32780
+    },
+    {
+      "epoch": 29.47841726618705,
+      "step": 32780,
+      "torque_loss": 0.12116900831460953
+    },
+    {
+      "epoch": 29.47841726618705,
+      "force_loss": 0.008808523416519165,
+      "step": 32780
+    },
+    {
+      "epoch": 29.487410071942445,
+      "grad_norm": 0.29399749636650085,
+      "learning_rate": 4.645078307212989e-05,
+      "loss": 0.0202,
+      "step": 32790
+    },
+    {
+      "action_loss": 0.004080316983163357,
+      "epoch": 29.487410071942445,
+      "step": 32790
+    },
+    {
+      "epoch": 29.487410071942445,
+      "step": 32790,
+      "torque_loss": 0.10931116342544556
+    },
+    {
+      "epoch": 29.487410071942445,
+      "force_loss": 0.003771195886656642,
+      "step": 32790
+    },
+    {
+      "epoch": 29.496402877697843,
+      "grad_norm": 0.23084160685539246,
+      "learning_rate": 4.642329529885768e-05,
+      "loss": 0.0168,
+      "step": 32800
+    },
+    {
+      "action_loss": 0.003108937991783023,
+      "epoch": 29.496402877697843,
+      "step": 32800
+    },
+    {
+      "epoch": 29.496402877697843,
+      "step": 32800,
+      "torque_loss": 0.06809798628091812
+    },
+    {
+      "epoch": 29.496402877697843,
+      "force_loss": 0.005005677696317434,
+      "step": 32800
+    },
+    {
+      "epoch": 29.505395683453237,
+      "grad_norm": 0.5844730138778687,
+      "learning_rate": 4.639580861209393e-05,
+      "loss": 0.0175,
+      "step": 32810
+    },
+    {
+      "action_loss": 0.00632280670106411,
+      "epoch": 29.505395683453237,
+      "step": 32810
+    },
+    {
+      "epoch": 29.505395683453237,
+      "step": 32810,
+      "torque_loss": 0.15455208718776703
+    },
+    {
+      "epoch": 29.505395683453237,
+      "force_loss": 0.006788300815969706,
+      "step": 32810
+    },
+    {
+      "epoch": 29.514388489208635,
+      "grad_norm": 0.2245039939880371,
+      "learning_rate": 4.636832302018835e-05,
+      "loss": 0.0208,
+      "step": 32820
+    },
+    {
+      "action_loss": 0.0030229396652430296,
+      "epoch": 29.514388489208635,
+      "step": 32820
+    },
+    {
+      "epoch": 29.514388489208635,
+      "step": 32820,
+      "torque_loss": 0.13324254751205444
+    },
+    {
+      "epoch": 29.514388489208635,
+      "force_loss": 0.004435569513589144,
+      "step": 32820
+    },
+    {
+      "epoch": 29.52338129496403,
+      "grad_norm": 0.1462528258562088,
+      "learning_rate": 4.6340838531490365e-05,
+      "loss": 0.0166,
+      "step": 32830
+    },
+    {
+      "action_loss": 0.00136242201551795,
+      "epoch": 29.52338129496403,
+      "step": 32830
+    },
+    {
+      "epoch": 29.52338129496403,
+      "step": 32830,
+      "torque_loss": 0.0685676857829094
+    },
+    {
+      "epoch": 29.52338129496403,
+      "force_loss": 0.0013977460330352187,
+      "step": 32830
+    },
+    {
+      "epoch": 29.532374100719423,
+      "grad_norm": 0.22453230619430542,
+      "learning_rate": 4.6313355154349e-05,
+      "loss": 0.0176,
+      "step": 32840
+    },
+    {
+      "action_loss": 0.005980899091809988,
+      "epoch": 29.532374100719423,
+      "step": 32840
+    },
+    {
+      "epoch": 29.532374100719423,
+      "step": 32840,
+      "torque_loss": 0.10074014216661453
+    },
+    {
+      "epoch": 29.532374100719423,
+      "force_loss": 0.004690586123615503,
+      "step": 32840
+    },
+    {
+      "epoch": 29.54136690647482,
+      "grad_norm": 0.1655723750591278,
+      "learning_rate": 4.6285872897113025e-05,
+      "loss": 0.017,
+      "step": 32850
+    },
+    {
+      "action_loss": 0.004144648090004921,
+      "epoch": 29.54136690647482,
+      "step": 32850
+    },
+    {
+      "epoch": 29.54136690647482,
+      "step": 32850,
+      "torque_loss": 0.1202431321144104
+    },
+    {
+      "epoch": 29.54136690647482,
+      "force_loss": 0.0019840833265334368,
+      "step": 32850
+    },
+    {
+      "epoch": 29.550359712230215,
+      "grad_norm": 0.1184563934803009,
+      "learning_rate": 4.625839176813077e-05,
+      "loss": 0.0163,
+      "step": 32860
+    },
+    {
+      "action_loss": 0.0019806192722171545,
+      "epoch": 29.550359712230215,
+      "step": 32860
+    },
+    {
+      "epoch": 29.550359712230215,
+      "step": 32860,
+      "torque_loss": 0.10810966044664383
+    },
+    {
+      "epoch": 29.550359712230215,
+      "force_loss": 0.004407880827784538,
+      "step": 32860
+    },
+    {
+      "epoch": 29.559352517985612,
+      "grad_norm": 0.12117161601781845,
+      "learning_rate": 4.623091177575031e-05,
+      "loss": 0.0179,
+      "step": 32870
+    },
+    {
+      "action_loss": 0.003812926122918725,
+      "epoch": 29.559352517985612,
+      "step": 32870
+    },
+    {
+      "epoch": 29.559352517985612,
+      "step": 32870,
+      "torque_loss": 0.08688423782587051
+    },
+    {
+      "epoch": 29.559352517985612,
+      "force_loss": 0.005498215556144714,
+      "step": 32870
+    },
+    {
+      "epoch": 29.568345323741006,
+      "grad_norm": 0.16760703921318054,
+      "learning_rate": 4.620343292831936e-05,
+      "loss": 0.018,
+      "step": 32880
+    },
+    {
+      "action_loss": 0.003354416461661458,
+      "epoch": 29.568345323741006,
+      "step": 32880
+    },
+    {
+      "epoch": 29.568345323741006,
+      "step": 32880,
+      "torque_loss": 0.09763610363006592
+    },
+    {
+      "epoch": 29.568345323741006,
+      "force_loss": 0.0025031135883182287,
+      "step": 32880
+    },
+    {
+      "epoch": 29.577338129496404,
+      "grad_norm": 0.2660999894142151,
+      "learning_rate": 4.6175955234185206e-05,
+      "loss": 0.0181,
+      "step": 32890
+    },
+    {
+      "action_loss": 0.009728002361953259,
+      "epoch": 29.577338129496404,
+      "step": 32890
+    },
+    {
+      "epoch": 29.577338129496404,
+      "step": 32890,
+      "torque_loss": 0.1275230199098587
+    },
+    {
+      "epoch": 29.577338129496404,
+      "force_loss": 0.010925649665296078,
+      "step": 32890
+    },
+    {
+      "epoch": 29.586330935251798,
+      "grad_norm": 0.3399328887462616,
+      "learning_rate": 4.614847870169492e-05,
+      "loss": 0.0168,
+      "step": 32900
+    },
+    {
+      "action_loss": 0.004511337727308273,
+      "epoch": 29.586330935251798,
+      "step": 32900
+    },
+    {
+      "epoch": 29.586330935251798,
+      "step": 32900,
+      "torque_loss": 0.1351112574338913
+    },
+    {
+      "epoch": 29.586330935251798,
+      "force_loss": 0.006338247563689947,
+      "step": 32900
+    },
+    {
+      "epoch": 29.595323741007196,
+      "grad_norm": 0.16598787903785706,
+      "learning_rate": 4.612100333919509e-05,
+      "loss": 0.0164,
+      "step": 32910
+    },
+    {
+      "action_loss": 0.0022442378103733063,
+      "epoch": 29.595323741007196,
+      "step": 32910
+    },
+    {
+      "epoch": 29.595323741007196,
+      "step": 32910,
+      "torque_loss": 0.1249852403998375
+    },
+    {
+      "epoch": 29.595323741007196,
+      "force_loss": 0.0027145585045218468,
+      "step": 32910
+    },
+    {
+      "epoch": 29.60431654676259,
+      "grad_norm": 0.8419445157051086,
+      "learning_rate": 4.609352915503202e-05,
+      "loss": 0.0193,
+      "step": 32920
+    },
+    {
+      "action_loss": 0.0026799149345606565,
+      "epoch": 29.60431654676259,
+      "step": 32920
+    },
+    {
+      "epoch": 29.60431654676259,
+      "step": 32920,
+      "torque_loss": 0.10197170823812485
+    },
+    {
+      "epoch": 29.60431654676259,
+      "force_loss": 0.004200031515210867,
+      "step": 32920
+    },
+    {
+      "epoch": 29.613309352517987,
+      "grad_norm": 0.17716027796268463,
+      "learning_rate": 4.606605615755166e-05,
+      "loss": 0.0169,
+      "step": 32930
+    },
+    {
+      "action_loss": 0.003672495484352112,
+      "epoch": 29.613309352517987,
+      "step": 32930
+    },
+    {
+      "epoch": 29.613309352517987,
+      "step": 32930,
+      "torque_loss": 0.09633766859769821
+    },
+    {
+      "epoch": 29.613309352517987,
+      "force_loss": 0.002823582151904702,
+      "step": 32930
+    },
+    {
+      "epoch": 29.62230215827338,
+      "grad_norm": 0.1838403046131134,
+      "learning_rate": 4.6038584355099576e-05,
+      "loss": 0.0195,
+      "step": 32940
+    },
+    {
+      "action_loss": 0.0030016282107681036,
+      "epoch": 29.62230215827338,
+      "step": 32940
+    },
+    {
+      "epoch": 29.62230215827338,
+      "step": 32940,
+      "torque_loss": 0.10197654366493225
+    },
+    {
+      "epoch": 29.62230215827338,
+      "force_loss": 0.0039540971629321575,
+      "step": 32940
+    },
+    {
+      "epoch": 29.631294964028775,
+      "grad_norm": 0.1331952065229416,
+      "learning_rate": 4.6011113756020964e-05,
+      "loss": 0.0195,
+      "step": 32950
+    },
+    {
+      "action_loss": 0.00984635204076767,
+      "epoch": 29.631294964028775,
+      "step": 32950
+    },
+    {
+      "epoch": 29.631294964028775,
+      "step": 32950,
+      "torque_loss": 0.13478004932403564
+    },
+    {
+      "epoch": 29.631294964028775,
+      "force_loss": 0.020974332466721535,
+      "step": 32950
+    },
+    {
+      "epoch": 29.640287769784173,
+      "grad_norm": 0.14265209436416626,
+      "learning_rate": 4.598364436866066e-05,
+      "loss": 0.0202,
+      "step": 32960
+    },
+    {
+      "action_loss": 0.003692147321999073,
+      "epoch": 29.640287769784173,
+      "step": 32960
+    },
+    {
+      "epoch": 29.640287769784173,
+      "step": 32960,
+      "torque_loss": 0.11055556684732437
+    },
+    {
+      "epoch": 29.640287769784173,
+      "force_loss": 0.008668371476233006,
+      "step": 32960
+    },
+    {
+      "epoch": 29.649280575539567,
+      "grad_norm": 0.303299218416214,
+      "learning_rate": 4.595617620136316e-05,
+      "loss": 0.02,
+      "step": 32970
+    },
+    {
+      "action_loss": 0.007363330572843552,
+      "epoch": 29.649280575539567,
+      "step": 32970
+    },
+    {
+      "epoch": 29.649280575539567,
+      "step": 32970,
+      "torque_loss": 0.13789105415344238
+    },
+    {
+      "epoch": 29.649280575539567,
+      "force_loss": 0.010800435207784176,
+      "step": 32970
+    },
+    {
+      "epoch": 29.658273381294965,
+      "grad_norm": 0.20006980001926422,
+      "learning_rate": 4.592870926247257e-05,
+      "loss": 0.0207,
+      "step": 32980
+    },
+    {
+      "action_loss": 0.0036609049420803785,
+      "epoch": 29.658273381294965,
+      "step": 32980
+    },
+    {
+      "epoch": 29.658273381294965,
+      "step": 32980,
+      "torque_loss": 0.1461556851863861
+    },
+    {
+      "epoch": 29.658273381294965,
+      "force_loss": 0.00689267972484231,
+      "step": 32980
+    },
+    {
+      "epoch": 29.66726618705036,
+      "grad_norm": 0.19028274714946747,
+      "learning_rate": 4.5901243560332594e-05,
+      "loss": 0.0186,
+      "step": 32990
+    },
+    {
+      "action_loss": 0.003864791477099061,
+      "epoch": 29.66726618705036,
+      "step": 32990
+    },
+    {
+      "epoch": 29.66726618705036,
+      "step": 32990,
+      "torque_loss": 0.11155102401971817
+    },
+    {
+      "epoch": 29.66726618705036,
+      "force_loss": 0.004080370534211397,
+      "step": 32990
+    },
+    {
+      "epoch": 29.676258992805757,
+      "grad_norm": 0.22748075425624847,
+      "learning_rate": 4.587377910328662e-05,
+      "loss": 0.0164,
+      "step": 33000
+    },
+    {
+      "action_loss": 0.0033679369371384382,
+      "epoch": 29.676258992805757,
+      "step": 33000
+    },
+    {
+      "epoch": 29.676258992805757,
+      "step": 33000,
+      "torque_loss": 0.12033047527074814
+    },
+    {
+      "epoch": 29.676258992805757,
+      "force_loss": 0.002878366969525814,
+      "step": 33000
+    },
+    {
+      "epoch": 29.68525179856115,
+      "grad_norm": 0.11324349790811539,
+      "learning_rate": 4.5846315899677586e-05,
+      "loss": 0.0165,
+      "step": 33010
+    },
+    {
+      "action_loss": 0.002293139696121216,
+      "epoch": 29.68525179856115,
+      "step": 33010
+    },
+    {
+      "epoch": 29.68525179856115,
+      "step": 33010,
+      "torque_loss": 0.11542467027902603
+    },
+    {
+      "epoch": 29.68525179856115,
+      "force_loss": 0.004007941577583551,
+      "step": 33010
+    },
+    {
+      "epoch": 29.694244604316548,
+      "grad_norm": 0.23012064397335052,
+      "learning_rate": 4.5818853957848114e-05,
+      "loss": 0.0175,
+      "step": 33020
+    },
+    {
+      "action_loss": 0.007706517819315195,
+      "epoch": 29.694244604316548,
+      "step": 33020
+    },
+    {
+      "epoch": 29.694244604316548,
+      "step": 33020,
+      "torque_loss": 0.1225409284234047
+    },
+    {
+      "epoch": 29.694244604316548,
+      "force_loss": 0.007782285567373037,
+      "step": 33020
+    },
+    {
+      "epoch": 29.703237410071942,
+      "grad_norm": 0.2857019305229187,
+      "learning_rate": 4.579139328614043e-05,
+      "loss": 0.0183,
+      "step": 33030
+    },
+    {
+      "action_loss": 0.0113174794241786,
+      "epoch": 29.703237410071942,
+      "step": 33030
+    },
+    {
+      "epoch": 29.703237410071942,
+      "step": 33030,
+      "torque_loss": 0.14130271971225739
+    },
+    {
+      "epoch": 29.703237410071942,
+      "force_loss": 0.008874858729541302,
+      "step": 33030
+    },
+    {
+      "epoch": 29.71223021582734,
+      "grad_norm": 0.21743011474609375,
+      "learning_rate": 4.576393389289633e-05,
+      "loss": 0.0187,
+      "step": 33040
+    },
+    {
+      "action_loss": 0.004120629280805588,
+      "epoch": 29.71223021582734,
+      "step": 33040
+    },
+    {
+      "epoch": 29.71223021582734,
+      "step": 33040,
+      "torque_loss": 0.1165727749466896
+    },
+    {
+      "epoch": 29.71223021582734,
+      "force_loss": 0.00596576975658536,
+      "step": 33040
+    },
+    {
+      "epoch": 29.721223021582734,
+      "grad_norm": 0.14444954693317413,
+      "learning_rate": 4.573647578645728e-05,
+      "loss": 0.02,
+      "step": 33050
+    },
+    {
+      "action_loss": 0.002949989167973399,
+      "epoch": 29.721223021582734,
+      "step": 33050
+    },
+    {
+      "epoch": 29.721223021582734,
+      "step": 33050,
+      "torque_loss": 0.13620251417160034
+    },
+    {
+      "epoch": 29.721223021582734,
+      "force_loss": 0.0032610802445560694,
+      "step": 33050
+    },
+    {
+      "epoch": 29.730215827338128,
+      "grad_norm": 0.11691472679376602,
+      "learning_rate": 4.57090189751643e-05,
+      "loss": 0.0153,
+      "step": 33060
+    },
+    {
+      "action_loss": 0.0019554237369447947,
+      "epoch": 29.730215827338128,
+      "step": 33060
+    },
+    {
+      "epoch": 29.730215827338128,
+      "step": 33060,
+      "torque_loss": 0.12198758125305176
+    },
+    {
+      "epoch": 29.730215827338128,
+      "force_loss": 0.0035621251445263624,
+      "step": 33060
+    },
+    {
+      "epoch": 29.739208633093526,
+      "grad_norm": 0.3457503020763397,
+      "learning_rate": 4.568156346735806e-05,
+      "loss": 0.0186,
+      "step": 33070
+    },
+    {
+      "action_loss": 0.002127479063346982,
+      "epoch": 29.739208633093526,
+      "step": 33070
+    },
+    {
+      "epoch": 29.739208633093526,
+      "step": 33070,
+      "torque_loss": 0.11081872135400772
+    },
+    {
+      "epoch": 29.739208633093526,
+      "force_loss": 0.003237922443076968,
+      "step": 33070
+    },
+    {
+      "epoch": 29.74820143884892,
+      "grad_norm": 0.14352914690971375,
+      "learning_rate": 4.565410927137882e-05,
+      "loss": 0.0172,
+      "step": 33080
+    },
+    {
+      "action_loss": 0.0018772659823298454,
+      "epoch": 29.74820143884892,
+      "step": 33080
+    },
+    {
+      "epoch": 29.74820143884892,
+      "step": 33080,
+      "torque_loss": 0.11937590688467026
+    },
+    {
+      "epoch": 29.74820143884892,
+      "force_loss": 0.003528344677761197,
+      "step": 33080
+    },
+    {
+      "epoch": 29.757194244604317,
+      "grad_norm": 0.2091064751148224,
+      "learning_rate": 4.562665639556644e-05,
+      "loss": 0.0208,
+      "step": 33090
+    },
+    {
+      "action_loss": 0.0020316801965236664,
+      "epoch": 29.757194244604317,
+      "step": 33090
+    },
+    {
+      "epoch": 29.757194244604317,
+      "step": 33090,
+      "torque_loss": 0.10107707977294922
+    },
+    {
+      "epoch": 29.757194244604317,
+      "force_loss": 0.003617035923525691,
+      "step": 33090
+    },
+    {
+      "epoch": 29.76618705035971,
+      "grad_norm": 0.18773119151592255,
+      "learning_rate": 4.559920484826037e-05,
+      "loss": 0.0176,
+      "step": 33100
+    },
+    {
+      "action_loss": 0.0025949233677238226,
+      "epoch": 29.76618705035971,
+      "step": 33100
+    },
+    {
+      "epoch": 29.76618705035971,
+      "step": 33100,
+      "torque_loss": 0.0989958867430687
+    },
+    {
+      "epoch": 29.76618705035971,
+      "force_loss": 0.0015219500055536628,
+      "step": 33100
+    },
+    {
+      "epoch": 29.77517985611511,
+      "grad_norm": 0.3331056535243988,
+      "learning_rate": 4.5571754637799665e-05,
+      "loss": 0.0184,
+      "step": 33110
+    },
+    {
+      "action_loss": 0.009025699459016323,
+      "epoch": 29.77517985611511,
+      "step": 33110
+    },
+    {
+      "epoch": 29.77517985611511,
+      "step": 33110,
+      "torque_loss": 0.10117089003324509
+    },
+    {
+      "epoch": 29.77517985611511,
+      "force_loss": 0.006383243948221207,
+      "step": 33110
+    },
+    {
+      "epoch": 29.784172661870503,
+      "grad_norm": 0.23136396706104279,
+      "learning_rate": 4.554430577252298e-05,
+      "loss": 0.0166,
+      "step": 33120
+    },
+    {
+      "action_loss": 0.008841276168823242,
+      "epoch": 29.784172661870503,
+      "step": 33120
+    },
+    {
+      "epoch": 29.784172661870503,
+      "step": 33120,
+      "torque_loss": 0.15196190774440765
+    },
+    {
+      "epoch": 29.784172661870503,
+      "force_loss": 0.005283138249069452,
+      "step": 33120
+    },
+    {
+      "epoch": 29.7931654676259,
+      "grad_norm": 0.17802201211452484,
+      "learning_rate": 4.551685826076858e-05,
+      "loss": 0.0186,
+      "step": 33130
+    },
+    {
+      "action_loss": 0.0073266527615487576,
+      "epoch": 29.7931654676259,
+      "step": 33130
+    },
+    {
+      "epoch": 29.7931654676259,
+      "step": 33130,
+      "torque_loss": 0.14417017996311188
+    },
+    {
+      "epoch": 29.7931654676259,
+      "force_loss": 0.0027394357603043318,
+      "step": 33130
+    },
+    {
+      "epoch": 29.802158273381295,
+      "grad_norm": 0.10232173651456833,
+      "learning_rate": 4.5489412110874246e-05,
+      "loss": 0.0186,
+      "step": 33140
+    },
+    {
+      "action_loss": 0.0026322982739657164,
+      "epoch": 29.802158273381295,
+      "step": 33140
+    },
+    {
+      "epoch": 29.802158273381295,
+      "step": 33140,
+      "torque_loss": 0.09275990724563599
+    },
+    {
+      "epoch": 29.802158273381295,
+      "force_loss": 0.0025260073598474264,
+      "step": 33140
+    },
+    {
+      "epoch": 29.81115107913669,
+      "grad_norm": 0.2526633143424988,
+      "learning_rate": 4.5461967331177444e-05,
+      "loss": 0.0191,
+      "step": 33150
+    },
+    {
+      "action_loss": 0.0031592827290296555,
+      "epoch": 29.81115107913669,
+      "step": 33150
+    },
+    {
+      "epoch": 29.81115107913669,
+      "step": 33150,
+      "torque_loss": 0.08308272808790207
+    },
+    {
+      "epoch": 29.81115107913669,
+      "force_loss": 0.0032640835270285606,
+      "step": 33150
+    },
+    {
+      "epoch": 29.820143884892087,
+      "grad_norm": 0.23503777384757996,
+      "learning_rate": 4.5434523930015115e-05,
+      "loss": 0.0185,
+      "step": 33160
+    },
+    {
+      "action_loss": 0.014097201637923717,
+      "epoch": 29.820143884892087,
+      "step": 33160
+    },
+    {
+      "epoch": 29.820143884892087,
+      "step": 33160,
+      "torque_loss": 0.15288548171520233
+    },
+    {
+      "epoch": 29.820143884892087,
+      "force_loss": 0.0197067242115736,
+      "step": 33160
+    },
+    {
+      "epoch": 29.82913669064748,
+      "grad_norm": 0.5467575788497925,
+      "learning_rate": 4.540708191572388e-05,
+      "loss": 0.0205,
+      "step": 33170
+    },
+    {
+      "action_loss": 0.0020516656804829836,
+      "epoch": 29.82913669064748,
+      "step": 33170
+    },
+    {
+      "epoch": 29.82913669064748,
+      "step": 33170,
+      "torque_loss": 0.07463323324918747
+    },
+    {
+      "epoch": 29.82913669064748,
+      "force_loss": 0.0032436437904834747,
+      "step": 33170
+    },
+    {
+      "epoch": 29.83812949640288,
+      "grad_norm": 0.31882551312446594,
+      "learning_rate": 4.537964129663991e-05,
+      "loss": 0.0173,
+      "step": 33180
+    },
+    {
+      "action_loss": 0.002108158776536584,
+      "epoch": 29.83812949640288,
+      "step": 33180
+    },
+    {
+      "epoch": 29.83812949640288,
+      "step": 33180,
+      "torque_loss": 0.15018528699874878
+    },
+    {
+      "epoch": 29.83812949640288,
+      "force_loss": 0.002047862159088254,
+      "step": 33180
+    },
+    {
+      "epoch": 29.847122302158272,
+      "grad_norm": 0.3126389980316162,
+      "learning_rate": 4.535220208109889e-05,
+      "loss": 0.0201,
+      "step": 33190
+    },
+    {
+      "action_loss": 0.0034148755948990583,
+      "epoch": 29.847122302158272,
+      "step": 33190
+    },
+    {
+      "epoch": 29.847122302158272,
+      "step": 33190,
+      "torque_loss": 0.14748704433441162
+    },
+    {
+      "epoch": 29.847122302158272,
+      "force_loss": 0.0017256079008802772,
+      "step": 33190
+    },
+    {
+      "epoch": 29.85611510791367,
+      "grad_norm": 0.15697187185287476,
+      "learning_rate": 4.5324764277436194e-05,
+      "loss": 0.0164,
+      "step": 33200
+    },
+    {
+      "action_loss": 0.0034842982422560453,
+      "epoch": 29.85611510791367,
+      "step": 33200
+    },
+    {
+      "epoch": 29.85611510791367,
+      "step": 33200,
+      "torque_loss": 0.13564208149909973
+    },
+    {
+      "epoch": 29.85611510791367,
+      "force_loss": 0.001287777558900416,
+      "step": 33200
+    },
+    {
+      "epoch": 29.865107913669064,
+      "grad_norm": 0.24154828488826752,
+      "learning_rate": 4.529732789398664e-05,
+      "loss": 0.0201,
+      "step": 33210
+    },
+    {
+      "action_loss": 0.003454518737271428,
+      "epoch": 29.865107913669064,
+      "step": 33210
+    },
+    {
+      "epoch": 29.865107913669064,
+      "step": 33210,
+      "torque_loss": 0.17831595242023468
+    },
+    {
+      "epoch": 29.865107913669064,
+      "force_loss": 0.001054757391102612,
+      "step": 33210
+    },
+    {
+      "epoch": 29.87410071942446,
+      "grad_norm": 0.1497463583946228,
+      "learning_rate": 4.526989293908472e-05,
+      "loss": 0.0208,
+      "step": 33220
+    },
+    {
+      "action_loss": 0.002602895488962531,
+      "epoch": 29.87410071942446,
+      "step": 33220
+    },
+    {
+      "epoch": 29.87410071942446,
+      "step": 33220,
+      "torque_loss": 0.15655624866485596
+    },
+    {
+      "epoch": 29.87410071942446,
+      "force_loss": 0.0034274160861968994,
+      "step": 33220
+    },
+    {
+      "epoch": 29.883093525179856,
+      "grad_norm": 0.2563171982765198,
+      "learning_rate": 4.524245942106442e-05,
+      "loss": 0.0184,
+      "step": 33230
+    },
+    {
+      "action_loss": 0.007358677219599485,
+      "epoch": 29.883093525179856,
+      "step": 33230
+    },
+    {
+      "epoch": 29.883093525179856,
+      "step": 33230,
+      "torque_loss": 0.16274546086788177
+    },
+    {
+      "epoch": 29.883093525179856,
+      "force_loss": 0.007875451818108559,
+      "step": 33230
+    },
+    {
+      "epoch": 29.892086330935253,
+      "grad_norm": 0.2812725007534027,
+      "learning_rate": 4.5215027348259345e-05,
+      "loss": 0.0209,
+      "step": 33240
+    },
+    {
+      "action_loss": 0.004986295476555824,
+      "epoch": 29.892086330935253,
+      "step": 33240
+    },
+    {
+      "epoch": 29.892086330935253,
+      "step": 33240,
+      "torque_loss": 0.11599931865930557
+    },
+    {
+      "epoch": 29.892086330935253,
+      "force_loss": 0.019289052113890648,
+      "step": 33240
+    },
+    {
+      "epoch": 29.901079136690647,
+      "grad_norm": 0.18221893906593323,
+      "learning_rate": 4.5187596729002616e-05,
+      "loss": 0.0196,
+      "step": 33250
+    },
+    {
+      "action_loss": 0.012487287633121014,
+      "epoch": 29.901079136690647,
+      "step": 33250
+    },
+    {
+      "epoch": 29.901079136690647,
+      "step": 33250,
+      "torque_loss": 0.07998945564031601
+    },
+    {
+      "epoch": 29.901079136690647,
+      "force_loss": 0.025370754301548004,
+      "step": 33250
+    },
+    {
+      "epoch": 29.91007194244604,
+      "grad_norm": 0.2595881521701813,
+      "learning_rate": 4.516016757162693e-05,
+      "loss": 0.0167,
+      "step": 33260
+    },
+    {
+      "action_loss": 0.019895339384675026,
+      "epoch": 29.91007194244604,
+      "step": 33260
+    },
+    {
+      "epoch": 29.91007194244604,
+      "step": 33260,
+      "torque_loss": 0.16868579387664795
+    },
+    {
+      "epoch": 29.91007194244604,
+      "force_loss": 0.008317751809954643,
+      "step": 33260
+    },
+    {
+      "epoch": 29.91906474820144,
+      "grad_norm": 0.16835442185401917,
+      "learning_rate": 4.513273988446457e-05,
+      "loss": 0.0187,
+      "step": 33270
+    },
+    {
+      "action_loss": 0.0020997589454054832,
+      "epoch": 29.91906474820144,
+      "step": 33270
+    },
+    {
+      "epoch": 29.91906474820144,
+      "step": 33270,
+      "torque_loss": 0.1342884749174118
+    },
+    {
+      "epoch": 29.91906474820144,
+      "force_loss": 0.005571839865297079,
+      "step": 33270
+    },
+    {
+      "epoch": 29.928057553956833,
+      "grad_norm": 0.22526520490646362,
+      "learning_rate": 4.5105313675847296e-05,
+      "loss": 0.0164,
+      "step": 33280
+    },
+    {
+      "action_loss": 0.004929483402520418,
+      "epoch": 29.928057553956833,
+      "step": 33280
+    },
+    {
+      "epoch": 29.928057553956833,
+      "step": 33280,
+      "torque_loss": 0.16145117580890656
+    },
+    {
+      "epoch": 29.928057553956833,
+      "force_loss": 0.005234928801655769,
+      "step": 33280
+    },
+    {
+      "epoch": 29.93705035971223,
+      "grad_norm": 0.20873883366584778,
+      "learning_rate": 4.5077888954106495e-05,
+      "loss": 0.0178,
+      "step": 33290
+    },
+    {
+      "action_loss": 0.0016989628784358501,
+      "epoch": 29.93705035971223,
+      "step": 33290
+    },
+    {
+      "epoch": 29.93705035971223,
+      "step": 33290,
+      "torque_loss": 0.09238481521606445
+    },
+    {
+      "epoch": 29.93705035971223,
+      "force_loss": 0.0024258980993181467,
+      "step": 33290
+    },
+    {
+      "epoch": 29.946043165467625,
+      "grad_norm": 0.1721733659505844,
+      "learning_rate": 4.505046572757309e-05,
+      "loss": 0.0202,
+      "step": 33300
+    },
+    {
+      "action_loss": 0.014291095547378063,
+      "epoch": 29.946043165467625,
+      "step": 33300
+    },
+    {
+      "epoch": 29.946043165467625,
+      "step": 33300,
+      "torque_loss": 0.13486896455287933
+    },
+    {
+      "epoch": 29.946043165467625,
+      "force_loss": 0.01131768524646759,
+      "step": 33300
+    },
+    {
+      "epoch": 29.955035971223023,
+      "grad_norm": 0.26423338055610657,
+      "learning_rate": 4.502304400457749e-05,
+      "loss": 0.0205,
+      "step": 33310
+    },
+    {
+      "action_loss": 0.002390447538346052,
+      "epoch": 29.955035971223023,
+      "step": 33310
+    },
+    {
+      "epoch": 29.955035971223023,
+      "step": 33310,
+      "torque_loss": 0.09509583562612534
+    },
+    {
+      "epoch": 29.955035971223023,
+      "force_loss": 0.0037280935794115067,
+      "step": 33310
+    },
+    {
+      "epoch": 29.964028776978417,
+      "grad_norm": 0.39813458919525146,
+      "learning_rate": 4.499562379344973e-05,
+      "loss": 0.0168,
+      "step": 33320
+    },
+    {
+      "action_loss": 0.01173321157693863,
+      "epoch": 29.964028776978417,
+      "step": 33320
+    },
+    {
+      "epoch": 29.964028776978417,
+      "step": 33320,
+      "torque_loss": 0.11689639091491699
+    },
+    {
+      "epoch": 29.964028776978417,
+      "force_loss": 0.011147972196340561,
+      "step": 33320
+    },
+    {
+      "epoch": 29.973021582733814,
+      "grad_norm": 0.09179895371198654,
+      "learning_rate": 4.4968205102519306e-05,
+      "loss": 0.0179,
+      "step": 33330
+    },
+    {
+      "action_loss": 0.0064166150987148285,
+      "epoch": 29.973021582733814,
+      "step": 33330
+    },
+    {
+      "epoch": 29.973021582733814,
+      "step": 33330,
+      "torque_loss": 0.1534060388803482
+    },
+    {
+      "epoch": 29.973021582733814,
+      "force_loss": 0.009151996113359928,
+      "step": 33330
+    },
+    {
+      "epoch": 29.98201438848921,
+      "grad_norm": 0.32343992590904236,
+      "learning_rate": 4.494078794011532e-05,
+      "loss": 0.0185,
+      "step": 33340
+    },
+    {
+      "action_loss": 0.0036133381072431803,
+      "epoch": 29.98201438848921,
+      "step": 33340
+    },
+    {
+      "epoch": 29.98201438848921,
+      "step": 33340,
+      "torque_loss": 0.11830791085958481
+    },
+    {
+      "epoch": 29.98201438848921,
+      "force_loss": 0.0028760216664522886,
+      "step": 33340
+    },
+    {
+      "epoch": 29.991007194244606,
+      "grad_norm": 0.4685978889465332,
+      "learning_rate": 4.491337231456639e-05,
+      "loss": 0.018,
+      "step": 33350
+    },
+    {
+      "action_loss": 0.004563137423247099,
+      "epoch": 29.991007194244606,
+      "step": 33350
+    },
+    {
+      "epoch": 29.991007194244606,
+      "step": 33350,
+      "torque_loss": 0.112477146089077
+    },
+    {
+      "epoch": 29.991007194244606,
+      "force_loss": 0.002000430366024375,
+      "step": 33350
+    },
+    {
+      "epoch": 30.0,
+      "grad_norm": 0.22582240402698517,
+      "learning_rate": 4.4885958234200634e-05,
+      "loss": 0.0164,
+      "step": 33360
+    },
+    {
+      "action_loss": 0.05387647822499275,
+      "epoch": 30.0,
+      "step": 33360
+    },
+    {
+      "epoch": 30.0,
+      "step": 33360,
+      "torque_loss": 0.15187954902648926
+    },
+    {
+      "epoch": 30.0,
+      "force_loss": 0.052053917199373245,
+      "step": 33360
+    },
+    {
+      "epoch": 30.008992805755394,
+      "grad_norm": 0.1293298304080963,
+      "learning_rate": 4.485854570734575e-05,
+      "loss": 0.0212,
+      "step": 33370
+    },
+    {
+      "action_loss": 0.02416323870420456,
+      "epoch": 30.008992805755394,
+      "step": 33370
+    },
+    {
+      "epoch": 30.008992805755394,
+      "step": 33370,
+      "torque_loss": 0.16026407480239868
+    },
+    {
+      "epoch": 30.008992805755394,
+      "force_loss": 0.03959124907851219,
+      "step": 33370
+    },
+    {
+      "epoch": 30.01798561151079,
+      "grad_norm": 0.4653620719909668,
+      "learning_rate": 4.483113474232891e-05,
+      "loss": 0.0203,
+      "step": 33380
+    },
+    {
+      "action_loss": 0.006471555680036545,
+      "epoch": 30.01798561151079,
+      "step": 33380
+    },
+    {
+      "epoch": 30.01798561151079,
+      "step": 33380,
+      "torque_loss": 0.10242074728012085
+    },
+    {
+      "epoch": 30.01798561151079,
+      "force_loss": 0.007715189829468727,
+      "step": 33380
+    },
+    {
+      "epoch": 30.026978417266186,
+      "grad_norm": 0.1794714629650116,
+      "learning_rate": 4.480372534747688e-05,
+      "loss": 0.0195,
+      "step": 33390
+    },
+    {
+      "action_loss": 0.002253317041322589,
+      "epoch": 30.026978417266186,
+      "step": 33390
+    },
+    {
+      "epoch": 30.026978417266186,
+      "step": 33390,
+      "torque_loss": 0.1080147996544838
+    },
+    {
+      "epoch": 30.026978417266186,
+      "force_loss": 0.0016623451374471188,
+      "step": 33390
+    },
+    {
+      "epoch": 30.035971223021583,
+      "grad_norm": 0.2829987704753876,
+      "learning_rate": 4.477631753111588e-05,
+      "loss": 0.022,
+      "step": 33400
+    },
+    {
+      "action_loss": 0.012816439382731915,
+      "epoch": 30.035971223021583,
+      "step": 33400
+    },
+    {
+      "epoch": 30.035971223021583,
+      "step": 33400,
+      "torque_loss": 0.12620994448661804
+    },
+    {
+      "epoch": 30.035971223021583,
+      "force_loss": 0.005510048475116491,
+      "step": 33400
+    },
+    {
+      "epoch": 30.044964028776977,
+      "grad_norm": 0.15960019826889038,
+      "learning_rate": 4.4748911301571686e-05,
+      "loss": 0.0188,
+      "step": 33410
+    },
+    {
+      "action_loss": 0.0019453917630016804,
+      "epoch": 30.044964028776977,
+      "step": 33410
+    },
+    {
+      "epoch": 30.044964028776977,
+      "step": 33410,
+      "torque_loss": 0.15751896798610687
+    },
+    {
+      "epoch": 30.044964028776977,
+      "force_loss": 0.0013027280801907182,
+      "step": 33410
+    },
+    {
+      "epoch": 30.053956834532375,
+      "grad_norm": 0.16101831197738647,
+      "learning_rate": 4.472150666716961e-05,
+      "loss": 0.0162,
+      "step": 33420
+    },
+    {
+      "action_loss": 0.0054649473167955875,
+      "epoch": 30.053956834532375,
+      "step": 33420
+    },
+    {
+      "epoch": 30.053956834532375,
+      "step": 33420,
+      "torque_loss": 0.1371350735425949
+    },
+    {
+      "epoch": 30.053956834532375,
+      "force_loss": 0.007342675235122442,
+      "step": 33420
+    },
+    {
+      "epoch": 30.06294964028777,
+      "grad_norm": 0.21752667427062988,
+      "learning_rate": 4.469410363623442e-05,
+      "loss": 0.0148,
+      "step": 33430
+    },
+    {
+      "action_loss": 0.0023852363228797913,
+      "epoch": 30.06294964028777,
+      "step": 33430
+    },
+    {
+      "epoch": 30.06294964028777,
+      "step": 33430,
+      "torque_loss": 0.11105093359947205
+    },
+    {
+      "epoch": 30.06294964028777,
+      "force_loss": 0.007129593286663294,
+      "step": 33430
+    },
+    {
+      "epoch": 30.071942446043167,
+      "grad_norm": 0.31025755405426025,
+      "learning_rate": 4.466670221709044e-05,
+      "loss": 0.0272,
+      "step": 33440
+    },
+    {
+      "action_loss": 0.0015622731298208237,
+      "epoch": 30.071942446043167,
+      "step": 33440
+    },
+    {
+      "epoch": 30.071942446043167,
+      "step": 33440,
+      "torque_loss": 0.07495087385177612
+    },
+    {
+      "epoch": 30.071942446043167,
+      "force_loss": 0.0027264754753559828,
+      "step": 33440
+    },
+    {
+      "epoch": 30.08093525179856,
+      "grad_norm": 0.16972821950912476,
+      "learning_rate": 4.463930241806154e-05,
+      "loss": 0.0186,
+      "step": 33450
+    },
+    {
+      "action_loss": 0.0013034323928877711,
+      "epoch": 30.08093525179856,
+      "step": 33450
+    },
+    {
+      "epoch": 30.08093525179856,
+      "step": 33450,
+      "torque_loss": 0.08380503207445145
+    },
+    {
+      "epoch": 30.08093525179856,
+      "force_loss": 0.0033943590242415667,
+      "step": 33450
+    },
+    {
+      "epoch": 30.08992805755396,
+      "grad_norm": 0.17786522209644318,
+      "learning_rate": 4.4611904247471006e-05,
+      "loss": 0.0155,
+      "step": 33460
+    },
+    {
+      "action_loss": 0.004154016263782978,
+      "epoch": 30.08992805755396,
+      "step": 33460
+    },
+    {
+      "epoch": 30.08992805755396,
+      "step": 33460,
+      "torque_loss": 0.12680579721927643
+    },
+    {
+      "epoch": 30.08992805755396,
+      "force_loss": 0.005134863313287497,
+      "step": 33460
+    },
+    {
+      "epoch": 30.098920863309353,
+      "grad_norm": 0.24578897655010223,
+      "learning_rate": 4.458450771364171e-05,
+      "loss": 0.0159,
+      "step": 33470
+    },
+    {
+      "action_loss": 0.0011644229525700212,
+      "epoch": 30.098920863309353,
+      "step": 33470
+    },
+    {
+      "epoch": 30.098920863309353,
+      "step": 33470,
+      "torque_loss": 0.08433590084314346
+    },
+    {
+      "epoch": 30.098920863309353,
+      "force_loss": 0.001824203529395163,
+      "step": 33470
+    },
+    {
+      "epoch": 30.107913669064747,
+      "grad_norm": 0.23850153386592865,
+      "learning_rate": 4.4557112824895965e-05,
+      "loss": 0.0155,
+      "step": 33480
+    },
+    {
+      "action_loss": 0.004469658713787794,
+      "epoch": 30.107913669064747,
+      "step": 33480
+    },
+    {
+      "epoch": 30.107913669064747,
+      "step": 33480,
+      "torque_loss": 0.13697348535060883
+    },
+    {
+      "epoch": 30.107913669064747,
+      "force_loss": 0.012753158807754517,
+      "step": 33480
+    },
+    {
+      "epoch": 30.116906474820144,
+      "grad_norm": 0.23066876828670502,
+      "learning_rate": 4.452971958955563e-05,
+      "loss": 0.0171,
+      "step": 33490
+    },
+    {
+      "action_loss": 0.004219897091388702,
+      "epoch": 30.116906474820144,
+      "step": 33490
+    },
+    {
+      "epoch": 30.116906474820144,
+      "step": 33490,
+      "torque_loss": 0.10240279883146286
+    },
+    {
+      "epoch": 30.116906474820144,
+      "force_loss": 0.0023729100357741117,
+      "step": 33490
+    },
+    {
+      "epoch": 30.12589928057554,
+      "grad_norm": 0.183693066239357,
+      "learning_rate": 4.450232801594208e-05,
+      "loss": 0.0166,
+      "step": 33500
+    },
+    {
+      "action_loss": 0.008158431388437748,
+      "epoch": 30.12589928057554,
+      "step": 33500
+    },
+    {
+      "epoch": 30.12589928057554,
+      "step": 33500,
+      "torque_loss": 0.11823907494544983
+    },
+    {
+      "epoch": 30.12589928057554,
+      "force_loss": 0.009602522477507591,
+      "step": 33500
+    },
+    {
+      "epoch": 30.134892086330936,
+      "grad_norm": 0.23512250185012817,
+      "learning_rate": 4.447493811237609e-05,
+      "loss": 0.0178,
+      "step": 33510
+    },
+    {
+      "action_loss": 0.012874695472419262,
+      "epoch": 30.134892086330936,
+      "step": 33510
+    },
+    {
+      "epoch": 30.134892086330936,
+      "step": 33510,
+      "torque_loss": 0.1644458770751953
+    },
+    {
+      "epoch": 30.134892086330936,
+      "force_loss": 0.018665602430701256,
+      "step": 33510
+    },
+    {
+      "epoch": 30.14388489208633,
+      "grad_norm": 0.2832042872905731,
+      "learning_rate": 4.444754988717804e-05,
+      "loss": 0.0197,
+      "step": 33520
+    },
+    {
+      "action_loss": 0.0023626608308404684,
+      "epoch": 30.14388489208633,
+      "step": 33520
+    },
+    {
+      "epoch": 30.14388489208633,
+      "step": 33520,
+      "torque_loss": 0.13426519930362701
+    },
+    {
+      "epoch": 30.14388489208633,
+      "force_loss": 0.0022349469363689423,
+      "step": 33520
+    },
+    {
+      "epoch": 30.152877697841728,
+      "grad_norm": 0.3160881996154785,
+      "learning_rate": 4.442016334866771e-05,
+      "loss": 0.0234,
+      "step": 33530
+    },
+    {
+      "action_loss": 0.007441936060786247,
+      "epoch": 30.152877697841728,
+      "step": 33530
+    },
+    {
+      "epoch": 30.152877697841728,
+      "step": 33530,
+      "torque_loss": 0.10917917639017105
+    },
+    {
+      "epoch": 30.152877697841728,
+      "force_loss": 0.0067154522985219955,
+      "step": 33530
+    },
+    {
+      "epoch": 30.16187050359712,
+      "grad_norm": 0.17749209702014923,
+      "learning_rate": 4.4392778505164445e-05,
+      "loss": 0.0173,
+      "step": 33540
+    },
+    {
+      "action_loss": 0.0013483191141858697,
+      "epoch": 30.16187050359712,
+      "step": 33540
+    },
+    {
+      "epoch": 30.16187050359712,
+      "step": 33540,
+      "torque_loss": 0.13838058710098267
+    },
+    {
+      "epoch": 30.16187050359712,
+      "force_loss": 0.0011774327140301466,
+      "step": 33540
+    },
+    {
+      "epoch": 30.17086330935252,
+      "grad_norm": 0.3133264482021332,
+      "learning_rate": 4.436539536498702e-05,
+      "loss": 0.0165,
+      "step": 33550
+    },
+    {
+      "action_loss": 0.00280605535954237,
+      "epoch": 30.17086330935252,
+      "step": 33550
+    },
+    {
+      "epoch": 30.17086330935252,
+      "step": 33550,
+      "torque_loss": 0.13557972013950348
+    },
+    {
+      "epoch": 30.17086330935252,
+      "force_loss": 0.0042430199682712555,
+      "step": 33550
+    },
+    {
+      "epoch": 30.179856115107913,
+      "grad_norm": 0.2542215883731842,
+      "learning_rate": 4.433801393645369e-05,
+      "loss": 0.0175,
+      "step": 33560
+    },
+    {
+      "action_loss": 0.004041679203510284,
+      "epoch": 30.179856115107913,
+      "step": 33560
+    },
+    {
+      "epoch": 30.179856115107913,
+      "step": 33560,
+      "torque_loss": 0.13326963782310486
+    },
+    {
+      "epoch": 30.179856115107913,
+      "force_loss": 0.0023123843129724264,
+      "step": 33560
+    },
+    {
+      "epoch": 30.18884892086331,
+      "grad_norm": 0.6925213932991028,
+      "learning_rate": 4.431063422788226e-05,
+      "loss": 0.0181,
+      "step": 33570
+    },
+    {
+      "action_loss": 0.003081059781834483,
+      "epoch": 30.18884892086331,
+      "step": 33570
+    },
+    {
+      "epoch": 30.18884892086331,
+      "step": 33570,
+      "torque_loss": 0.1330539584159851
+    },
+    {
+      "epoch": 30.18884892086331,
+      "force_loss": 0.007840510457754135,
+      "step": 33570
+    },
+    {
+      "epoch": 30.197841726618705,
+      "grad_norm": 0.3583317697048187,
+      "learning_rate": 4.428325624758991e-05,
+      "loss": 0.0177,
+      "step": 33580
+    },
+    {
+      "action_loss": 0.003772567957639694,
+      "epoch": 30.197841726618705,
+      "step": 33580
+    },
+    {
+      "epoch": 30.197841726618705,
+      "step": 33580,
+      "torque_loss": 0.10897797346115112
+    },
+    {
+      "epoch": 30.197841726618705,
+      "force_loss": 0.003163060173392296,
+      "step": 33580
+    },
+    {
+      "epoch": 30.2068345323741,
+      "grad_norm": 0.4390246868133545,
+      "learning_rate": 4.4255880003893366e-05,
+      "loss": 0.0217,
+      "step": 33590
+    },
+    {
+      "action_loss": 0.002452633110806346,
+      "epoch": 30.2068345323741,
+      "step": 33590
+    },
+    {
+      "epoch": 30.2068345323741,
+      "step": 33590,
+      "torque_loss": 0.13239872455596924
+    },
+    {
+      "epoch": 30.2068345323741,
+      "force_loss": 0.003955683205276728,
+      "step": 33590
+    },
+    {
+      "epoch": 30.215827338129497,
+      "grad_norm": 0.23491589725017548,
+      "learning_rate": 4.422850550510884e-05,
+      "loss": 0.0166,
+      "step": 33600
+    },
+    {
+      "action_loss": 0.007710794452577829,
+      "epoch": 30.215827338129497,
+      "step": 33600
+    },
+    {
+      "epoch": 30.215827338129497,
+      "step": 33600,
+      "torque_loss": 0.21398377418518066
+    },
+    {
+      "epoch": 30.215827338129497,
+      "force_loss": 0.012334768660366535,
+      "step": 33600
+    },
+    {
+      "epoch": 30.22482014388489,
+      "grad_norm": 0.29465538263320923,
+      "learning_rate": 4.4201132759551934e-05,
+      "loss": 0.0205,
+      "step": 33610
+    },
+    {
+      "action_loss": 0.004370903596282005,
+      "epoch": 30.22482014388489,
+      "step": 33610
+    },
+    {
+      "epoch": 30.22482014388489,
+      "step": 33610,
+      "torque_loss": 0.1381704956293106
+    },
+    {
+      "epoch": 30.22482014388489,
+      "force_loss": 0.0032118279486894608,
+      "step": 33610
+    },
+    {
+      "epoch": 30.23381294964029,
+      "grad_norm": 0.21574260294437408,
+      "learning_rate": 4.4173761775537804e-05,
+      "loss": 0.0207,
+      "step": 33620
+    },
+    {
+      "action_loss": 0.0028051387052983046,
+      "epoch": 30.23381294964029,
+      "step": 33620
+    },
+    {
+      "epoch": 30.23381294964029,
+      "step": 33620,
+      "torque_loss": 0.10356395691633224
+    },
+    {
+      "epoch": 30.23381294964029,
+      "force_loss": 0.0020543793216347694,
+      "step": 33620
+    },
+    {
+      "epoch": 30.242805755395683,
+      "grad_norm": 0.320326566696167,
+      "learning_rate": 4.414639256138099e-05,
+      "loss": 0.0185,
+      "step": 33630
+    },
+    {
+      "action_loss": 0.00179259292781353,
+      "epoch": 30.242805755395683,
+      "step": 33630
+    },
+    {
+      "epoch": 30.242805755395683,
+      "step": 33630,
+      "torque_loss": 0.1169067919254303
+    },
+    {
+      "epoch": 30.242805755395683,
+      "force_loss": 0.003782205516472459,
+      "step": 33630
+    },
+    {
+      "epoch": 30.25179856115108,
+      "grad_norm": 0.24306286871433258,
+      "learning_rate": 4.411902512539557e-05,
+      "loss": 0.0186,
+      "step": 33640
+    },
+    {
+      "action_loss": 0.002906916895881295,
+      "epoch": 30.25179856115108,
+      "step": 33640
+    },
+    {
+      "epoch": 30.25179856115108,
+      "step": 33640,
+      "torque_loss": 0.1894126981496811
+    },
+    {
+      "epoch": 30.25179856115108,
+      "force_loss": 0.0016277733957394958,
+      "step": 33640
+    },
+    {
+      "epoch": 30.260791366906474,
+      "grad_norm": 0.36424165964126587,
+      "learning_rate": 4.4091659475895044e-05,
+      "loss": 0.0173,
+      "step": 33650
+    },
+    {
+      "action_loss": 0.0023266819771379232,
+      "epoch": 30.260791366906474,
+      "step": 33650
+    },
+    {
+      "epoch": 30.260791366906474,
+      "step": 33650,
+      "torque_loss": 0.12251365184783936
+    },
+    {
+      "epoch": 30.260791366906474,
+      "force_loss": 0.005497979465872049,
+      "step": 33650
+    },
+    {
+      "epoch": 30.269784172661872,
+      "grad_norm": 0.23687084019184113,
+      "learning_rate": 4.406429562119235e-05,
+      "loss": 0.0184,
+      "step": 33660
+    },
+    {
+      "action_loss": 0.008143831975758076,
+      "epoch": 30.269784172661872,
+      "step": 33660
+    },
+    {
+      "epoch": 30.269784172661872,
+      "step": 33660,
+      "torque_loss": 0.13751937448978424
+    },
+    {
+      "epoch": 30.269784172661872,
+      "force_loss": 0.00900186225771904,
+      "step": 33660
+    },
+    {
+      "epoch": 30.278776978417266,
+      "grad_norm": 0.37069347500801086,
+      "learning_rate": 4.4036933569599945e-05,
+      "loss": 0.0192,
+      "step": 33670
+    },
+    {
+      "action_loss": 0.003267800435423851,
+      "epoch": 30.278776978417266,
+      "step": 33670
+    },
+    {
+      "epoch": 30.278776978417266,
+      "step": 33670,
+      "torque_loss": 0.14806006848812103
+    },
+    {
+      "epoch": 30.278776978417266,
+      "force_loss": 0.003686965210363269,
+      "step": 33670
+    },
+    {
+      "epoch": 30.28776978417266,
+      "grad_norm": 0.22585856914520264,
+      "learning_rate": 4.400957332942965e-05,
+      "loss": 0.0192,
+      "step": 33680
+    },
+    {
+      "action_loss": 0.004052708391100168,
+      "epoch": 30.28776978417266,
+      "step": 33680
+    },
+    {
+      "epoch": 30.28776978417266,
+      "step": 33680,
+      "torque_loss": 0.109751857817173
+    },
+    {
+      "epoch": 30.28776978417266,
+      "force_loss": 0.005819523707032204,
+      "step": 33680
+    },
+    {
+      "epoch": 30.296762589928058,
+      "grad_norm": 0.33284908533096313,
+      "learning_rate": 4.3982214908992844e-05,
+      "loss": 0.0173,
+      "step": 33690
+    },
+    {
+      "action_loss": 0.00896837655454874,
+      "epoch": 30.296762589928058,
+      "step": 33690
+    },
+    {
+      "epoch": 30.296762589928058,
+      "step": 33690,
+      "torque_loss": 0.12630636990070343
+    },
+    {
+      "epoch": 30.296762589928058,
+      "force_loss": 0.009608641266822815,
+      "step": 33690
+    },
+    {
+      "epoch": 30.305755395683452,
+      "grad_norm": 0.3137732148170471,
+      "learning_rate": 4.3954858316600235e-05,
+      "loss": 0.0176,
+      "step": 33700
+    },
+    {
+      "action_loss": 0.0027881485875695944,
+      "epoch": 30.305755395683452,
+      "step": 33700
+    },
+    {
+      "epoch": 30.305755395683452,
+      "step": 33700,
+      "torque_loss": 0.14884357154369354
+    },
+    {
+      "epoch": 30.305755395683452,
+      "force_loss": 0.0032020965591073036,
+      "step": 33700
+    },
+    {
+      "epoch": 30.31474820143885,
+      "grad_norm": 0.21497157216072083,
+      "learning_rate": 4.392750356056205e-05,
+      "loss": 0.0173,
+      "step": 33710
+    },
+    {
+      "action_loss": 0.0028638120274990797,
+      "epoch": 30.31474820143885,
+      "step": 33710
+    },
+    {
+      "epoch": 30.31474820143885,
+      "step": 33710,
+      "torque_loss": 0.07155754417181015
+    },
+    {
+      "epoch": 30.31474820143885,
+      "force_loss": 0.00342006329447031,
+      "step": 33710
+    },
+    {
+      "epoch": 30.323741007194243,
+      "grad_norm": 0.13051877915859222,
+      "learning_rate": 4.390015064918798e-05,
+      "loss": 0.0161,
+      "step": 33720
+    },
+    {
+      "action_loss": 0.0015397461829707026,
+      "epoch": 30.323741007194243,
+      "step": 33720
+    },
+    {
+      "epoch": 30.323741007194243,
+      "step": 33720,
+      "torque_loss": 0.17357449233531952
+    },
+    {
+      "epoch": 30.323741007194243,
+      "force_loss": 0.0026060156524181366,
+      "step": 33720
+    },
+    {
+      "epoch": 30.33273381294964,
+      "grad_norm": 0.13377784192562103,
+      "learning_rate": 4.387279959078705e-05,
+      "loss": 0.0165,
+      "step": 33730
+    },
+    {
+      "action_loss": 0.008763168938457966,
+      "epoch": 30.33273381294964,
+      "step": 33730
+    },
+    {
+      "epoch": 30.33273381294964,
+      "step": 33730,
+      "torque_loss": 0.14452257752418518
+    },
+    {
+      "epoch": 30.33273381294964,
+      "force_loss": 0.010376120917499065,
+      "step": 33730
+    },
+    {
+      "epoch": 30.341726618705035,
+      "grad_norm": 0.14394210278987885,
+      "learning_rate": 4.384545039366786e-05,
+      "loss": 0.0176,
+      "step": 33740
+    },
+    {
+      "action_loss": 0.005538098514080048,
+      "epoch": 30.341726618705035,
+      "step": 33740
+    },
+    {
+      "epoch": 30.341726618705035,
+      "step": 33740,
+      "torque_loss": 0.22733789682388306
+    },
+    {
+      "epoch": 30.341726618705035,
+      "force_loss": 0.004319025669246912,
+      "step": 33740
+    },
+    {
+      "epoch": 30.350719424460433,
+      "grad_norm": 0.3404535949230194,
+      "learning_rate": 4.381810306613831e-05,
+      "loss": 0.0216,
+      "step": 33750
+    },
+    {
+      "action_loss": 0.006918886210769415,
+      "epoch": 30.350719424460433,
+      "step": 33750
+    },
+    {
+      "epoch": 30.350719424460433,
+      "step": 33750,
+      "torque_loss": 0.1701337844133377
+    },
+    {
+      "epoch": 30.350719424460433,
+      "force_loss": 0.011591176502406597,
+      "step": 33750
+    },
+    {
+      "epoch": 30.359712230215827,
+      "grad_norm": 0.13359081745147705,
+      "learning_rate": 4.3790757616505826e-05,
+      "loss": 0.0171,
+      "step": 33760
+    },
+    {
+      "action_loss": 0.013710635714232922,
+      "epoch": 30.359712230215827,
+      "step": 33760
+    },
+    {
+      "epoch": 30.359712230215827,
+      "step": 33760,
+      "torque_loss": 0.10910716652870178
+    },
+    {
+      "epoch": 30.359712230215827,
+      "force_loss": 0.01313174981623888,
+      "step": 33760
+    },
+    {
+      "epoch": 30.368705035971225,
+      "grad_norm": 0.33932897448539734,
+      "learning_rate": 4.376341405307725e-05,
+      "loss": 0.0221,
+      "step": 33770
+    },
+    {
+      "action_loss": 0.0016877828165888786,
+      "epoch": 30.368705035971225,
+      "step": 33770
+    },
+    {
+      "epoch": 30.368705035971225,
+      "step": 33770,
+      "torque_loss": 0.12250272184610367
+    },
+    {
+      "epoch": 30.368705035971225,
+      "force_loss": 0.001253430382348597,
+      "step": 33770
+    },
+    {
+      "epoch": 30.37769784172662,
+      "grad_norm": 0.15971879661083221,
+      "learning_rate": 4.37360723841588e-05,
+      "loss": 0.0208,
+      "step": 33780
+    },
+    {
+      "action_loss": 0.006034780293703079,
+      "epoch": 30.37769784172662,
+      "step": 33780
+    },
+    {
+      "epoch": 30.37769784172662,
+      "step": 33780,
+      "torque_loss": 0.16161827743053436
+    },
+    {
+      "epoch": 30.37769784172662,
+      "force_loss": 0.004980245605111122,
+      "step": 33780
+    },
+    {
+      "epoch": 30.386690647482013,
+      "grad_norm": 0.3918313980102539,
+      "learning_rate": 4.370873261805619e-05,
+      "loss": 0.0245,
+      "step": 33790
+    },
+    {
+      "action_loss": 0.01284277718514204,
+      "epoch": 30.386690647482013,
+      "step": 33790
+    },
+    {
+      "epoch": 30.386690647482013,
+      "step": 33790,
+      "torque_loss": 0.16608965396881104
+    },
+    {
+      "epoch": 30.386690647482013,
+      "force_loss": 0.00825351383537054,
+      "step": 33790
+    },
+    {
+      "epoch": 30.39568345323741,
+      "grad_norm": 0.21983924508094788,
+      "learning_rate": 4.368139476307449e-05,
+      "loss": 0.0221,
+      "step": 33800
+    },
+    {
+      "action_loss": 0.007103261072188616,
+      "epoch": 30.39568345323741,
+      "step": 33800
+    },
+    {
+      "epoch": 30.39568345323741,
+      "step": 33800,
+      "torque_loss": 0.11381250619888306
+    },
+    {
+      "epoch": 30.39568345323741,
+      "force_loss": 0.007707206532359123,
+      "step": 33800
+    },
+    {
+      "epoch": 30.404676258992804,
+      "grad_norm": 0.32419559359550476,
+      "learning_rate": 4.365405882751822e-05,
+      "loss": 0.019,
+      "step": 33810
+    },
+    {
+      "action_loss": 0.0071796514093875885,
+      "epoch": 30.404676258992804,
+      "step": 33810
+    },
+    {
+      "epoch": 30.404676258992804,
+      "step": 33810,
+      "torque_loss": 0.16883157193660736
+    },
+    {
+      "epoch": 30.404676258992804,
+      "force_loss": 0.008002888411283493,
+      "step": 33810
+    },
+    {
+      "epoch": 30.413669064748202,
+      "grad_norm": 0.30898603796958923,
+      "learning_rate": 4.3626724819691326e-05,
+      "loss": 0.0164,
+      "step": 33820
+    },
+    {
+      "action_loss": 0.003020596457645297,
+      "epoch": 30.413669064748202,
+      "step": 33820
+    },
+    {
+      "epoch": 30.413669064748202,
+      "step": 33820,
+      "torque_loss": 0.08189874142408371
+    },
+    {
+      "epoch": 30.413669064748202,
+      "force_loss": 0.0019158223876729608,
+      "step": 33820
+    },
+    {
+      "epoch": 30.422661870503596,
+      "grad_norm": 0.18998531997203827,
+      "learning_rate": 4.359939274789715e-05,
+      "loss": 0.0165,
+      "step": 33830
+    },
+    {
+      "action_loss": 0.003167690709233284,
+      "epoch": 30.422661870503596,
+      "step": 33830
+    },
+    {
+      "epoch": 30.422661870503596,
+      "step": 33830,
+      "torque_loss": 0.11638975143432617
+    },
+    {
+      "epoch": 30.422661870503596,
+      "force_loss": 0.001651280210353434,
+      "step": 33830
+    },
+    {
+      "epoch": 30.431654676258994,
+      "grad_norm": 0.128376767039299,
+      "learning_rate": 4.357206262043848e-05,
+      "loss": 0.0191,
+      "step": 33840
+    },
+    {
+      "action_loss": 0.0014494335046038032,
+      "epoch": 30.431654676258994,
+      "step": 33840
+    },
+    {
+      "epoch": 30.431654676258994,
+      "step": 33840,
+      "torque_loss": 0.10283250361680984
+    },
+    {
+      "epoch": 30.431654676258994,
+      "force_loss": 0.0010659225517883897,
+      "step": 33840
+    },
+    {
+      "epoch": 30.440647482014388,
+      "grad_norm": 0.14935314655303955,
+      "learning_rate": 4.354473444561745e-05,
+      "loss": 0.0154,
+      "step": 33850
+    },
+    {
+      "action_loss": 0.0070334733463823795,
+      "epoch": 30.440647482014388,
+      "step": 33850
+    },
+    {
+      "epoch": 30.440647482014388,
+      "step": 33850,
+      "torque_loss": 0.13468003273010254
+    },
+    {
+      "epoch": 30.440647482014388,
+      "force_loss": 0.01708083599805832,
+      "step": 33850
+    },
+    {
+      "epoch": 30.449640287769785,
+      "grad_norm": 0.3647903501987457,
+      "learning_rate": 4.3517408231735644e-05,
+      "loss": 0.0179,
+      "step": 33860
+    },
+    {
+      "action_loss": 0.004732321482151747,
+      "epoch": 30.449640287769785,
+      "step": 33860
+    },
+    {
+      "epoch": 30.449640287769785,
+      "step": 33860,
+      "torque_loss": 0.12329337000846863
+    },
+    {
+      "epoch": 30.449640287769785,
+      "force_loss": 0.018216587603092194,
+      "step": 33860
+    },
+    {
+      "epoch": 30.45863309352518,
+      "grad_norm": 0.1677381694316864,
+      "learning_rate": 4.3490083987094086e-05,
+      "loss": 0.0162,
+      "step": 33870
+    },
+    {
+      "action_loss": 0.0013670871267095208,
+      "epoch": 30.45863309352518,
+      "step": 33870
+    },
+    {
+      "epoch": 30.45863309352518,
+      "step": 33870,
+      "torque_loss": 0.10352057218551636
+    },
+    {
+      "epoch": 30.45863309352518,
+      "force_loss": 0.0013810965465381742,
+      "step": 33870
+    },
+    {
+      "epoch": 30.467625899280577,
+      "grad_norm": 0.131908118724823,
+      "learning_rate": 4.34627617199931e-05,
+      "loss": 0.0206,
+      "step": 33880
+    },
+    {
+      "action_loss": 0.0023977644741535187,
+      "epoch": 30.467625899280577,
+      "step": 33880
+    },
+    {
+      "epoch": 30.467625899280577,
+      "step": 33880,
+      "torque_loss": 0.11434202641248703
+    },
+    {
+      "epoch": 30.467625899280577,
+      "force_loss": 0.002391970017924905,
+      "step": 33880
+    },
+    {
+      "epoch": 30.47661870503597,
+      "grad_norm": 0.3285118341445923,
+      "learning_rate": 4.3435441438732526e-05,
+      "loss": 0.0168,
+      "step": 33890
+    },
+    {
+      "action_loss": 0.003221740946173668,
+      "epoch": 30.47661870503597,
+      "step": 33890
+    },
+    {
+      "epoch": 30.47661870503597,
+      "step": 33890,
+      "torque_loss": 0.12017184495925903
+    },
+    {
+      "epoch": 30.47661870503597,
+      "force_loss": 0.002145914128050208,
+      "step": 33890
+    },
+    {
+      "epoch": 30.485611510791365,
+      "grad_norm": 0.11932337284088135,
+      "learning_rate": 4.340812315161149e-05,
+      "loss": 0.0183,
+      "step": 33900
+    },
+    {
+      "action_loss": 0.00820623617619276,
+      "epoch": 30.485611510791365,
+      "step": 33900
+    },
+    {
+      "epoch": 30.485611510791365,
+      "step": 33900,
+      "torque_loss": 0.13924716413021088
+    },
+    {
+      "epoch": 30.485611510791365,
+      "force_loss": 0.010649804025888443,
+      "step": 33900
+    },
+    {
+      "epoch": 30.494604316546763,
+      "grad_norm": 0.9350066184997559,
+      "learning_rate": 4.338080686692859e-05,
+      "loss": 0.0217,
+      "step": 33910
+    },
+    {
+      "action_loss": 0.004893421661108732,
+      "epoch": 30.494604316546763,
+      "step": 33910
+    },
+    {
+      "epoch": 30.494604316546763,
+      "step": 33910,
+      "torque_loss": 0.10739513486623764
+    },
+    {
+      "epoch": 30.494604316546763,
+      "force_loss": 0.009484733454883099,
+      "step": 33910
+    },
+    {
+      "epoch": 30.503597122302157,
+      "grad_norm": 0.2702327072620392,
+      "learning_rate": 4.3353492592981816e-05,
+      "loss": 0.0164,
+      "step": 33920
+    },
+    {
+      "action_loss": 0.0021730372682213783,
+      "epoch": 30.503597122302157,
+      "step": 33920
+    },
+    {
+      "epoch": 30.503597122302157,
+      "step": 33920,
+      "torque_loss": 0.14096295833587646
+    },
+    {
+      "epoch": 30.503597122302157,
+      "force_loss": 0.003891879925504327,
+      "step": 33920
+    },
+    {
+      "epoch": 30.512589928057555,
+      "grad_norm": 0.29781439900398254,
+      "learning_rate": 4.3326180338068485e-05,
+      "loss": 0.0224,
+      "step": 33930
+    },
+    {
+      "action_loss": 0.0019507581600919366,
+      "epoch": 30.512589928057555,
+      "step": 33930
+    },
+    {
+      "epoch": 30.512589928057555,
+      "step": 33930,
+      "torque_loss": 0.10935936123132706
+    },
+    {
+      "epoch": 30.512589928057555,
+      "force_loss": 0.004337312187999487,
+      "step": 33930
+    },
+    {
+      "epoch": 30.52158273381295,
+      "grad_norm": 0.18422941863536835,
+      "learning_rate": 4.3298870110485356e-05,
+      "loss": 0.0189,
+      "step": 33940
+    },
+    {
+      "action_loss": 0.003793962299823761,
+      "epoch": 30.52158273381295,
+      "step": 33940
+    },
+    {
+      "epoch": 30.52158273381295,
+      "step": 33940,
+      "torque_loss": 0.09891750663518906
+    },
+    {
+      "epoch": 30.52158273381295,
+      "force_loss": 0.003323591547086835,
+      "step": 33940
+    },
+    {
+      "epoch": 30.530575539568346,
+      "grad_norm": 0.13422386348247528,
+      "learning_rate": 4.3271561918528567e-05,
+      "loss": 0.0159,
+      "step": 33950
+    },
+    {
+      "action_loss": 0.0013124371180310845,
+      "epoch": 30.530575539568346,
+      "step": 33950
+    },
+    {
+      "epoch": 30.530575539568346,
+      "step": 33950,
+      "torque_loss": 0.1449468731880188
+    },
+    {
+      "epoch": 30.530575539568346,
+      "force_loss": 0.0011732857674360275,
+      "step": 33950
+    },
+    {
+      "epoch": 30.53956834532374,
+      "grad_norm": 0.12928429245948792,
+      "learning_rate": 4.324425577049359e-05,
+      "loss": 0.0188,
+      "step": 33960
+    },
+    {
+      "action_loss": 0.004642493557184935,
+      "epoch": 30.53956834532374,
+      "step": 33960
+    },
+    {
+      "epoch": 30.53956834532374,
+      "step": 33960,
+      "torque_loss": 0.1675863415002823
+    },
+    {
+      "epoch": 30.53956834532374,
+      "force_loss": 0.01138545572757721,
+      "step": 33960
+    },
+    {
+      "epoch": 30.548561151079138,
+      "grad_norm": 0.2427515685558319,
+      "learning_rate": 4.321695167467535e-05,
+      "loss": 0.0179,
+      "step": 33970
+    },
+    {
+      "action_loss": 0.0022519289050251245,
+      "epoch": 30.548561151079138,
+      "step": 33970
+    },
+    {
+      "epoch": 30.548561151079138,
+      "step": 33970,
+      "torque_loss": 0.09331681579351425
+    },
+    {
+      "epoch": 30.548561151079138,
+      "force_loss": 0.004880804102867842,
+      "step": 33970
+    },
+    {
+      "epoch": 30.557553956834532,
+      "grad_norm": 0.24680624902248383,
+      "learning_rate": 4.3189649639368093e-05,
+      "loss": 0.0223,
+      "step": 33980
+    },
+    {
+      "action_loss": 0.00781483668833971,
+      "epoch": 30.557553956834532,
+      "step": 33980
+    },
+    {
+      "epoch": 30.557553956834532,
+      "step": 33980,
+      "torque_loss": 0.10990267992019653
+    },
+    {
+      "epoch": 30.557553956834532,
+      "force_loss": 0.004496725741773844,
+      "step": 33980
+    },
+    {
+      "epoch": 30.56654676258993,
+      "grad_norm": 0.27627503871917725,
+      "learning_rate": 4.316234967286547e-05,
+      "loss": 0.018,
+      "step": 33990
+    },
+    {
+      "action_loss": 0.002367333509027958,
+      "epoch": 30.56654676258993,
+      "step": 33990
+    },
+    {
+      "epoch": 30.56654676258993,
+      "step": 33990,
+      "torque_loss": 0.17009912431240082
+    },
+    {
+      "epoch": 30.56654676258993,
+      "force_loss": 0.002819410292431712,
+      "step": 33990
+    },
+    {
+      "epoch": 30.575539568345324,
+      "grad_norm": 0.42873600125312805,
+      "learning_rate": 4.313505178346046e-05,
+      "loss": 0.0234,
+      "step": 34000
+    },
+    {
+      "action_loss": 0.0024660711642354727,
+      "epoch": 30.575539568345324,
+      "step": 34000
+    },
+    {
+      "epoch": 30.575539568345324,
+      "step": 34000,
+      "torque_loss": 0.0902329683303833
+    },
+    {
+      "epoch": 30.575539568345324,
+      "force_loss": 0.0014743782812729478,
+      "step": 34000
+    },
+    {
+      "epoch": 30.584532374100718,
+      "grad_norm": 0.245400533080101,
+      "learning_rate": 4.3107755979445465e-05,
+      "loss": 0.0189,
+      "step": 34010
+    },
+    {
+      "action_loss": 0.00279278215020895,
+      "epoch": 30.584532374100718,
+      "step": 34010
+    },
+    {
+      "epoch": 30.584532374100718,
+      "step": 34010,
+      "torque_loss": 0.18927471339702606
+    },
+    {
+      "epoch": 30.584532374100718,
+      "force_loss": 0.0023267928045243025,
+      "step": 34010
+    },
+    {
+      "epoch": 30.593525179856115,
+      "grad_norm": 0.14921408891677856,
+      "learning_rate": 4.308046226911224e-05,
+      "loss": 0.017,
+      "step": 34020
+    },
+    {
+      "action_loss": 0.0031119815539568663,
+      "epoch": 30.593525179856115,
+      "step": 34020
+    },
+    {
+      "epoch": 30.593525179856115,
+      "step": 34020,
+      "torque_loss": 0.11340288072824478
+    },
+    {
+      "epoch": 30.593525179856115,
+      "force_loss": 0.0018722632667049766,
+      "step": 34020
+    },
+    {
+      "epoch": 30.60251798561151,
+      "grad_norm": 0.1204448863863945,
+      "learning_rate": 4.305317066075185e-05,
+      "loss": 0.019,
+      "step": 34030
+    },
+    {
+      "action_loss": 0.00167891348246485,
+      "epoch": 30.60251798561151,
+      "step": 34030
+    },
+    {
+      "epoch": 30.60251798561151,
+      "step": 34030,
+      "torque_loss": 0.17503099143505096
+    },
+    {
+      "epoch": 30.60251798561151,
+      "force_loss": 0.0012211333960294724,
+      "step": 34030
+    },
+    {
+      "epoch": 30.611510791366907,
+      "grad_norm": 0.21496781706809998,
+      "learning_rate": 4.302588116265482e-05,
+      "loss": 0.0179,
+      "step": 34040
+    },
+    {
+      "action_loss": 0.006426831241697073,
+      "epoch": 30.611510791366907,
+      "step": 34040
+    },
+    {
+      "epoch": 30.611510791366907,
+      "step": 34040,
+      "torque_loss": 0.1717856377363205
+    },
+    {
+      "epoch": 30.611510791366907,
+      "force_loss": 0.009729974903166294,
+      "step": 34040
+    },
+    {
+      "epoch": 30.6205035971223,
+      "grad_norm": 0.2579571008682251,
+      "learning_rate": 4.299859378311094e-05,
+      "loss": 0.0224,
+      "step": 34050
+    },
+    {
+      "action_loss": 0.00375187024474144,
+      "epoch": 30.6205035971223,
+      "step": 34050
+    },
+    {
+      "epoch": 30.6205035971223,
+      "step": 34050,
+      "torque_loss": 0.10367164760828018
+    },
+    {
+      "epoch": 30.6205035971223,
+      "force_loss": 0.009184802882373333,
+      "step": 34050
+    },
+    {
+      "epoch": 30.6294964028777,
+      "grad_norm": 0.31832313537597656,
+      "learning_rate": 4.2971308530409424e-05,
+      "loss": 0.0164,
+      "step": 34060
+    },
+    {
+      "action_loss": 0.01362632866948843,
+      "epoch": 30.6294964028777,
+      "step": 34060
+    },
+    {
+      "epoch": 30.6294964028777,
+      "step": 34060,
+      "torque_loss": 0.13499194383621216
+    },
+    {
+      "epoch": 30.6294964028777,
+      "force_loss": 0.017313940450549126,
+      "step": 34060
+    },
+    {
+      "epoch": 30.638489208633093,
+      "grad_norm": 0.29899221658706665,
+      "learning_rate": 4.2944025412838765e-05,
+      "loss": 0.0167,
+      "step": 34070
+    },
+    {
+      "action_loss": 0.004860586486756802,
+      "epoch": 30.638489208633093,
+      "step": 34070
+    },
+    {
+      "epoch": 30.638489208633093,
+      "step": 34070,
+      "torque_loss": 0.16460739076137543
+    },
+    {
+      "epoch": 30.638489208633093,
+      "force_loss": 0.003158821724355221,
+      "step": 34070
+    },
+    {
+      "epoch": 30.64748201438849,
+      "grad_norm": 0.2701317071914673,
+      "learning_rate": 4.291674443868689e-05,
+      "loss": 0.0177,
+      "step": 34080
+    },
+    {
+      "action_loss": 0.004506182856857777,
+      "epoch": 30.64748201438849,
+      "step": 34080
+    },
+    {
+      "epoch": 30.64748201438849,
+      "step": 34080,
+      "torque_loss": 0.14024712145328522
+    },
+    {
+      "epoch": 30.64748201438849,
+      "force_loss": 0.0014873739564791322,
+      "step": 34080
+    },
+    {
+      "epoch": 30.656474820143885,
+      "grad_norm": 0.1805032640695572,
+      "learning_rate": 4.288946561624104e-05,
+      "loss": 0.0168,
+      "step": 34090
+    },
+    {
+      "action_loss": 0.002074779011309147,
+      "epoch": 30.656474820143885,
+      "step": 34090
+    },
+    {
+      "epoch": 30.656474820143885,
+      "step": 34090,
+      "torque_loss": 0.09850070625543594
+    },
+    {
+      "epoch": 30.656474820143885,
+      "force_loss": 0.0031660879030823708,
+      "step": 34090
+    },
+    {
+      "epoch": 30.665467625899282,
+      "grad_norm": 0.10007679462432861,
+      "learning_rate": 4.2862188953787794e-05,
+      "loss": 0.0182,
+      "step": 34100
+    },
+    {
+      "action_loss": 0.005797685589641333,
+      "epoch": 30.665467625899282,
+      "step": 34100
+    },
+    {
+      "epoch": 30.665467625899282,
+      "step": 34100,
+      "torque_loss": 0.12656909227371216
+    },
+    {
+      "epoch": 30.665467625899282,
+      "force_loss": 0.010923708789050579,
+      "step": 34100
+    },
+    {
+      "epoch": 30.674460431654676,
+      "grad_norm": 0.295224130153656,
+      "learning_rate": 4.283491445961308e-05,
+      "loss": 0.0178,
+      "step": 34110
+    },
+    {
+      "action_loss": 0.004737206734716892,
+      "epoch": 30.674460431654676,
+      "step": 34110
+    },
+    {
+      "epoch": 30.674460431654676,
+      "step": 34110,
+      "torque_loss": 0.12477246671915054
+    },
+    {
+      "epoch": 30.674460431654676,
+      "force_loss": 0.003697453299537301,
+      "step": 34110
+    },
+    {
+      "epoch": 30.68345323741007,
+      "grad_norm": 0.2363261878490448,
+      "learning_rate": 4.2807642142002155e-05,
+      "loss": 0.0179,
+      "step": 34120
+    },
+    {
+      "action_loss": 0.0015916851116344333,
+      "epoch": 30.68345323741007,
+      "step": 34120
+    },
+    {
+      "epoch": 30.68345323741007,
+      "step": 34120,
+      "torque_loss": 0.15076014399528503
+    },
+    {
+      "epoch": 30.68345323741007,
+      "force_loss": 0.0037832444068044424,
+      "step": 34120
+    },
+    {
+      "epoch": 30.692446043165468,
+      "grad_norm": 0.2545408606529236,
+      "learning_rate": 4.278037200923966e-05,
+      "loss": 0.0162,
+      "step": 34130
+    },
+    {
+      "action_loss": 0.0030171528924256563,
+      "epoch": 30.692446043165468,
+      "step": 34130
+    },
+    {
+      "epoch": 30.692446043165468,
+      "step": 34130,
+      "torque_loss": 0.12975893914699554
+    },
+    {
+      "epoch": 30.692446043165468,
+      "force_loss": 0.003769857110455632,
+      "step": 34130
+    },
+    {
+      "epoch": 30.701438848920862,
+      "grad_norm": 0.19470427930355072,
+      "learning_rate": 4.275310406960953e-05,
+      "loss": 0.0225,
+      "step": 34140
+    },
+    {
+      "action_loss": 0.02639020048081875,
+      "epoch": 30.701438848920862,
+      "step": 34140
+    },
+    {
+      "epoch": 30.701438848920862,
+      "step": 34140,
+      "torque_loss": 0.2032298594713211
+    },
+    {
+      "epoch": 30.701438848920862,
+      "force_loss": 0.044616106897592545,
+      "step": 34140
+    },
+    {
+      "epoch": 30.71043165467626,
+      "grad_norm": 0.19291940331459045,
+      "learning_rate": 4.272583833139502e-05,
+      "loss": 0.0237,
+      "step": 34150
+    },
+    {
+      "action_loss": 0.0019570773001760244,
+      "epoch": 30.71043165467626,
+      "step": 34150
+    },
+    {
+      "epoch": 30.71043165467626,
+      "step": 34150,
+      "torque_loss": 0.15377110242843628
+    },
+    {
+      "epoch": 30.71043165467626,
+      "force_loss": 0.0014714262215420604,
+      "step": 34150
+    },
+    {
+      "epoch": 30.719424460431654,
+      "grad_norm": 0.35958021879196167,
+      "learning_rate": 4.2698574802878794e-05,
+      "loss": 0.0187,
+      "step": 34160
+    },
+    {
+      "action_loss": 0.008195770904421806,
+      "epoch": 30.719424460431654,
+      "step": 34160
+    },
+    {
+      "epoch": 30.719424460431654,
+      "step": 34160,
+      "torque_loss": 0.10825852304697037
+    },
+    {
+      "epoch": 30.719424460431654,
+      "force_loss": 0.0059619080275297165,
+      "step": 34160
+    },
+    {
+      "epoch": 30.72841726618705,
+      "grad_norm": 0.25675928592681885,
+      "learning_rate": 4.2671313492342734e-05,
+      "loss": 0.0164,
+      "step": 34170
+    },
+    {
+      "action_loss": 0.004299210850149393,
+      "epoch": 30.72841726618705,
+      "step": 34170
+    },
+    {
+      "epoch": 30.72841726618705,
+      "step": 34170,
+      "torque_loss": 0.08962175995111465
+    },
+    {
+      "epoch": 30.72841726618705,
+      "force_loss": 0.009974678978323936,
+      "step": 34170
+    },
+    {
+      "epoch": 30.737410071942445,
+      "grad_norm": 0.337780624628067,
+      "learning_rate": 4.264405440806813e-05,
+      "loss": 0.0172,
+      "step": 34180
+    },
+    {
+      "action_loss": 0.0011797873303294182,
+      "epoch": 30.737410071942445,
+      "step": 34180
+    },
+    {
+      "epoch": 30.737410071942445,
+      "step": 34180,
+      "torque_loss": 0.12248945981264114
+    },
+    {
+      "epoch": 30.737410071942445,
+      "force_loss": 0.001072687329724431,
+      "step": 34180
+    },
+    {
+      "epoch": 30.746402877697843,
+      "grad_norm": 0.2946050465106964,
+      "learning_rate": 4.26167975583356e-05,
+      "loss": 0.0227,
+      "step": 34190
+    },
+    {
+      "action_loss": 0.0018826182931661606,
+      "epoch": 30.746402877697843,
+      "step": 34190
+    },
+    {
+      "epoch": 30.746402877697843,
+      "step": 34190,
+      "torque_loss": 0.1174575462937355
+    },
+    {
+      "epoch": 30.746402877697843,
+      "force_loss": 0.0018945131450891495,
+      "step": 34190
+    },
+    {
+      "epoch": 30.755395683453237,
+      "grad_norm": 0.23761264979839325,
+      "learning_rate": 4.2589542951425e-05,
+      "loss": 0.0165,
+      "step": 34200
+    },
+    {
+      "action_loss": 0.003781116334721446,
+      "epoch": 30.755395683453237,
+      "step": 34200
+    },
+    {
+      "epoch": 30.755395683453237,
+      "step": 34200,
+      "torque_loss": 0.15604843199253082
+    },
+    {
+      "epoch": 30.755395683453237,
+      "force_loss": 0.003606157610192895,
+      "step": 34200
+    },
+    {
+      "epoch": 30.764388489208635,
+      "grad_norm": 0.2848336696624756,
+      "learning_rate": 4.2562290595615615e-05,
+      "loss": 0.0181,
+      "step": 34210
+    },
+    {
+      "action_loss": 0.003935912624001503,
+      "epoch": 30.764388489208635,
+      "step": 34210
+    },
+    {
+      "epoch": 30.764388489208635,
+      "step": 34210,
+      "torque_loss": 0.11677604913711548
+    },
+    {
+      "epoch": 30.764388489208635,
+      "force_loss": 0.005298606585711241,
+      "step": 34210
+    },
+    {
+      "epoch": 30.77338129496403,
+      "grad_norm": 0.18371392786502838,
+      "learning_rate": 4.2535040499185946e-05,
+      "loss": 0.0206,
+      "step": 34220
+    },
+    {
+      "action_loss": 0.003213894786313176,
+      "epoch": 30.77338129496403,
+      "step": 34220
+    },
+    {
+      "epoch": 30.77338129496403,
+      "step": 34220,
+      "torque_loss": 0.12822218239307404
+    },
+    {
+      "epoch": 30.77338129496403,
+      "force_loss": 0.0016867819940671325,
+      "step": 34220
+    },
+    {
+      "epoch": 30.782374100719423,
+      "grad_norm": 0.4384494423866272,
+      "learning_rate": 4.250779267041387e-05,
+      "loss": 0.0197,
+      "step": 34230
+    },
+    {
+      "action_loss": 0.0026134636718779802,
+      "epoch": 30.782374100719423,
+      "step": 34230
+    },
+    {
+      "epoch": 30.782374100719423,
+      "step": 34230,
+      "torque_loss": 0.09193725138902664
+    },
+    {
+      "epoch": 30.782374100719423,
+      "force_loss": 0.002919525606557727,
+      "step": 34230
+    },
+    {
+      "epoch": 30.79136690647482,
+      "grad_norm": 0.3756982088088989,
+      "learning_rate": 4.248054711757657e-05,
+      "loss": 0.0195,
+      "step": 34240
+    },
+    {
+      "action_loss": 0.0046348064206540585,
+      "epoch": 30.79136690647482,
+      "step": 34240
+    },
+    {
+      "epoch": 30.79136690647482,
+      "step": 34240,
+      "torque_loss": 0.10517802834510803
+    },
+    {
+      "epoch": 30.79136690647482,
+      "force_loss": 0.0036902560386806726,
+      "step": 34240
+    },
+    {
+      "epoch": 30.800359712230215,
+      "grad_norm": 0.18024572730064392,
+      "learning_rate": 4.245330384895052e-05,
+      "loss": 0.0199,
+      "step": 34250
+    },
+    {
+      "action_loss": 0.016274841502308846,
+      "epoch": 30.800359712230215,
+      "step": 34250
+    },
+    {
+      "epoch": 30.800359712230215,
+      "step": 34250,
+      "torque_loss": 0.18422596156597137
+    },
+    {
+      "epoch": 30.800359712230215,
+      "force_loss": 0.01038005668669939,
+      "step": 34250
+    },
+    {
+      "epoch": 30.809352517985612,
+      "grad_norm": 0.2398492991924286,
+      "learning_rate": 4.242606287281151e-05,
+      "loss": 0.0241,
+      "step": 34260
+    },
+    {
+      "action_loss": 0.012687087059020996,
+      "epoch": 30.809352517985612,
+      "step": 34260
+    },
+    {
+      "epoch": 30.809352517985612,
+      "step": 34260,
+      "torque_loss": 0.14182384312152863
+    },
+    {
+      "epoch": 30.809352517985612,
+      "force_loss": 0.02039421536028385,
+      "step": 34260
+    },
+    {
+      "epoch": 30.818345323741006,
+      "grad_norm": 0.2453010231256485,
+      "learning_rate": 4.2398824197434595e-05,
+      "loss": 0.0252,
+      "step": 34270
+    },
+    {
+      "action_loss": 0.0049733868800103664,
+      "epoch": 30.818345323741006,
+      "step": 34270
+    },
+    {
+      "epoch": 30.818345323741006,
+      "step": 34270,
+      "torque_loss": 0.16707904636859894
+    },
+    {
+      "epoch": 30.818345323741006,
+      "force_loss": 0.010182465426623821,
+      "step": 34270
+    },
+    {
+      "epoch": 30.827338129496404,
+      "grad_norm": 0.11318911612033844,
+      "learning_rate": 4.23715878310942e-05,
+      "loss": 0.019,
+      "step": 34280
+    },
+    {
+      "action_loss": 0.0023354210425168276,
+      "epoch": 30.827338129496404,
+      "step": 34280
+    },
+    {
+      "epoch": 30.827338129496404,
+      "step": 34280,
+      "torque_loss": 0.08740288019180298
+    },
+    {
+      "epoch": 30.827338129496404,
+      "force_loss": 0.0018805539002642035,
+      "step": 34280
+    },
+    {
+      "epoch": 30.836330935251798,
+      "grad_norm": 0.27674686908721924,
+      "learning_rate": 4.234435378206402e-05,
+      "loss": 0.0171,
+      "step": 34290
+    },
+    {
+      "action_loss": 0.005275351461023092,
+      "epoch": 30.836330935251798,
+      "step": 34290
+    },
+    {
+      "epoch": 30.836330935251798,
+      "step": 34290,
+      "torque_loss": 0.1801472306251526
+    },
+    {
+      "epoch": 30.836330935251798,
+      "force_loss": 0.008779492229223251,
+      "step": 34290
+    },
+    {
+      "epoch": 30.845323741007196,
+      "grad_norm": 0.22926506400108337,
+      "learning_rate": 4.2317122058617006e-05,
+      "loss": 0.0189,
+      "step": 34300
+    },
+    {
+      "action_loss": 0.0030152134131640196,
+      "epoch": 30.845323741007196,
+      "step": 34300
+    },
+    {
+      "epoch": 30.845323741007196,
+      "step": 34300,
+      "torque_loss": 0.11216884106397629
+    },
+    {
+      "epoch": 30.845323741007196,
+      "force_loss": 0.0015955861890688539,
+      "step": 34300
+    },
+    {
+      "epoch": 30.85431654676259,
+      "grad_norm": 0.22880510985851288,
+      "learning_rate": 4.2289892669025485e-05,
+      "loss": 0.0185,
+      "step": 34310
+    },
+    {
+      "action_loss": 0.0020585735328495502,
+      "epoch": 30.85431654676259,
+      "step": 34310
+    },
+    {
+      "epoch": 30.85431654676259,
+      "step": 34310,
+      "torque_loss": 0.12581248581409454
+    },
+    {
+      "epoch": 30.85431654676259,
+      "force_loss": 0.0010630955221131444,
+      "step": 34310
+    },
+    {
+      "epoch": 30.863309352517987,
+      "grad_norm": 0.3811962902545929,
+      "learning_rate": 4.226266562156097e-05,
+      "loss": 0.0194,
+      "step": 34320
+    },
+    {
+      "action_loss": 0.0009385032462887466,
+      "epoch": 30.863309352517987,
+      "step": 34320
+    },
+    {
+      "epoch": 30.863309352517987,
+      "step": 34320,
+      "torque_loss": 0.10718772560358047
+    },
+    {
+      "epoch": 30.863309352517987,
+      "force_loss": 0.0010268172482028604,
+      "step": 34320
+    },
+    {
+      "epoch": 30.87230215827338,
+      "grad_norm": 0.2305983603000641,
+      "learning_rate": 4.223544092449435e-05,
+      "loss": 0.0182,
+      "step": 34330
+    },
+    {
+      "action_loss": 0.002569180214777589,
+      "epoch": 30.87230215827338,
+      "step": 34330
+    },
+    {
+      "epoch": 30.87230215827338,
+      "step": 34330,
+      "torque_loss": 0.1357460469007492
+    },
+    {
+      "epoch": 30.87230215827338,
+      "force_loss": 0.0052687483839690685,
+      "step": 34330
+    },
+    {
+      "epoch": 30.881294964028775,
+      "grad_norm": 0.1490369290113449,
+      "learning_rate": 4.2208218586095784e-05,
+      "loss": 0.0168,
+      "step": 34340
+    },
+    {
+      "action_loss": 0.002988257212564349,
+      "epoch": 30.881294964028775,
+      "step": 34340
+    },
+    {
+      "epoch": 30.881294964028775,
+      "step": 34340,
+      "torque_loss": 0.10673753172159195
+    },
+    {
+      "epoch": 30.881294964028775,
+      "force_loss": 0.0021501716692000628,
+      "step": 34340
+    },
+    {
+      "epoch": 30.890287769784173,
+      "grad_norm": 0.31900548934936523,
+      "learning_rate": 4.218099861463466e-05,
+      "loss": 0.0183,
+      "step": 34350
+    },
+    {
+      "action_loss": 0.0057869236916303635,
+      "epoch": 30.890287769784173,
+      "step": 34350
+    },
+    {
+      "epoch": 30.890287769784173,
+      "step": 34350,
+      "torque_loss": 0.13345248997211456
+    },
+    {
+      "epoch": 30.890287769784173,
+      "force_loss": 0.006804469972848892,
+      "step": 34350
+    },
+    {
+      "epoch": 30.899280575539567,
+      "grad_norm": 0.19831882417201996,
+      "learning_rate": 4.215378101837972e-05,
+      "loss": 0.0212,
+      "step": 34360
+    },
+    {
+      "action_loss": 0.004619716200977564,
+      "epoch": 30.899280575539567,
+      "step": 34360
+    },
+    {
+      "epoch": 30.899280575539567,
+      "step": 34360,
+      "torque_loss": 0.16431497037410736
+    },
+    {
+      "epoch": 30.899280575539567,
+      "force_loss": 0.003166493261232972,
+      "step": 34360
+    },
+    {
+      "epoch": 30.908273381294965,
+      "grad_norm": 0.11223962157964706,
+      "learning_rate": 4.2126565805598937e-05,
+      "loss": 0.0183,
+      "step": 34370
+    },
+    {
+      "action_loss": 0.0017998417606577277,
+      "epoch": 30.908273381294965,
+      "step": 34370
+    },
+    {
+      "epoch": 30.908273381294965,
+      "step": 34370,
+      "torque_loss": 0.09303891658782959
+    },
+    {
+      "epoch": 30.908273381294965,
+      "force_loss": 0.0026637546252459288,
+      "step": 34370
+    },
+    {
+      "epoch": 30.91726618705036,
+      "grad_norm": 0.29248949885368347,
+      "learning_rate": 4.209935298455957e-05,
+      "loss": 0.0165,
+      "step": 34380
+    },
+    {
+      "action_loss": 0.0026354186702519655,
+      "epoch": 30.91726618705036,
+      "step": 34380
+    },
+    {
+      "epoch": 30.91726618705036,
+      "step": 34380,
+      "torque_loss": 0.10319829732179642
+    },
+    {
+      "epoch": 30.91726618705036,
+      "force_loss": 0.0020492158364504576,
+      "step": 34380
+    },
+    {
+      "epoch": 30.926258992805757,
+      "grad_norm": 0.17651765048503876,
+      "learning_rate": 4.207214256352817e-05,
+      "loss": 0.0168,
+      "step": 34390
+    },
+    {
+      "action_loss": 0.0015649498673155904,
+      "epoch": 30.926258992805757,
+      "step": 34390
+    },
+    {
+      "epoch": 30.926258992805757,
+      "step": 34390,
+      "torque_loss": 0.0876346305012703
+    },
+    {
+      "epoch": 30.926258992805757,
+      "force_loss": 0.002223416231572628,
+      "step": 34390
+    },
+    {
+      "epoch": 30.93525179856115,
+      "grad_norm": 0.23415303230285645,
+      "learning_rate": 4.2044934550770524e-05,
+      "loss": 0.0206,
+      "step": 34400
+    },
+    {
+      "action_loss": 0.048367906361818314,
+      "epoch": 30.93525179856115,
+      "step": 34400
+    },
+    {
+      "epoch": 30.93525179856115,
+      "step": 34400,
+      "torque_loss": 0.17771399021148682
+    },
+    {
+      "epoch": 30.93525179856115,
+      "force_loss": 0.07640401273965836,
+      "step": 34400
+    },
+    {
+      "epoch": 30.944244604316548,
+      "grad_norm": 0.1210138276219368,
+      "learning_rate": 4.201772895455174e-05,
+      "loss": 0.0219,
+      "step": 34410
+    },
+    {
+      "action_loss": 0.002077364129945636,
+      "epoch": 30.944244604316548,
+      "step": 34410
+    },
+    {
+      "epoch": 30.944244604316548,
+      "step": 34410,
+      "torque_loss": 0.1436670571565628
+    },
+    {
+      "epoch": 30.944244604316548,
+      "force_loss": 0.0031349908094853163,
+      "step": 34410
+    },
+    {
+      "epoch": 30.953237410071942,
+      "grad_norm": 0.23080870509147644,
+      "learning_rate": 4.199052578313613e-05,
+      "loss": 0.0202,
+      "step": 34420
+    },
+    {
+      "action_loss": 0.0038766537327319384,
+      "epoch": 30.953237410071942,
+      "step": 34420
+    },
+    {
+      "epoch": 30.953237410071942,
+      "step": 34420,
+      "torque_loss": 0.12370244413614273
+    },
+    {
+      "epoch": 30.953237410071942,
+      "force_loss": 0.00461923610419035,
+      "step": 34420
+    },
+    {
+      "epoch": 30.96223021582734,
+      "grad_norm": 0.3497636020183563,
+      "learning_rate": 4.1963325044787294e-05,
+      "loss": 0.0196,
+      "step": 34430
+    },
+    {
+      "action_loss": 0.0013608367880806327,
+      "epoch": 30.96223021582734,
+      "step": 34430
+    },
+    {
+      "epoch": 30.96223021582734,
+      "step": 34430,
+      "torque_loss": 0.11652910709381104
+    },
+    {
+      "epoch": 30.96223021582734,
+      "force_loss": 0.00146779406350106,
+      "step": 34430
+    },
+    {
+      "epoch": 30.971223021582734,
+      "grad_norm": 0.18539556860923767,
+      "learning_rate": 4.193612674776814e-05,
+      "loss": 0.0167,
+      "step": 34440
+    },
+    {
+      "action_loss": 0.005846960935741663,
+      "epoch": 30.971223021582734,
+      "step": 34440
+    },
+    {
+      "epoch": 30.971223021582734,
+      "step": 34440,
+      "torque_loss": 0.14601987600326538
+    },
+    {
+      "epoch": 30.971223021582734,
+      "force_loss": 0.007143509108573198,
+      "step": 34440
+    },
+    {
+      "epoch": 30.980215827338128,
+      "grad_norm": 0.1306850165128708,
+      "learning_rate": 4.1908930900340745e-05,
+      "loss": 0.0184,
+      "step": 34450
+    },
+    {
+      "action_loss": 0.005658076610416174,
+      "epoch": 30.980215827338128,
+      "step": 34450
+    },
+    {
+      "epoch": 30.980215827338128,
+      "step": 34450,
+      "torque_loss": 0.13187600672245026
+    },
+    {
+      "epoch": 30.980215827338128,
+      "force_loss": 0.0058209956623613834,
+      "step": 34450
+    },
+    {
+      "epoch": 30.989208633093526,
+      "grad_norm": 0.21161605417728424,
+      "learning_rate": 4.1881737510766536e-05,
+      "loss": 0.0168,
+      "step": 34460
+    },
+    {
+      "action_loss": 0.002968036336824298,
+      "epoch": 30.989208633093526,
+      "step": 34460
+    },
+    {
+      "epoch": 30.989208633093526,
+      "step": 34460,
+      "torque_loss": 0.1066054105758667
+    },
+    {
+      "epoch": 30.989208633093526,
+      "force_loss": 0.0036938395351171494,
+      "step": 34460
+    },
+    {
+      "epoch": 30.99820143884892,
+      "grad_norm": 0.08903868496417999,
+      "learning_rate": 4.185454658730609e-05,
+      "loss": 0.016,
+      "step": 34470
+    },
+    {
+      "action_loss": 0.003074484644457698,
+      "epoch": 30.99820143884892,
+      "step": 34470
+    },
+    {
+      "epoch": 30.99820143884892,
+      "step": 34470,
+      "torque_loss": 0.17268311977386475
+    },
+    {
+      "epoch": 30.99820143884892,
+      "force_loss": 0.0022521994542330503,
+      "step": 34470
+    },
+    {
+      "epoch": 31.007194244604317,
+      "grad_norm": 0.28496021032333374,
+      "learning_rate": 4.1827358138219355e-05,
+      "loss": 0.0176,
+      "step": 34480
+    },
+    {
+      "action_loss": 0.0025270762853324413,
+      "epoch": 31.007194244604317,
+      "step": 34480
+    },
+    {
+      "epoch": 31.007194244604317,
+      "step": 34480,
+      "torque_loss": 0.11888081580400467
+    },
+    {
+      "epoch": 31.007194244604317,
+      "force_loss": 0.006741913501173258,
+      "step": 34480
+    },
+    {
+      "epoch": 31.01618705035971,
+      "grad_norm": 0.29523223638534546,
+      "learning_rate": 4.1800172171765404e-05,
+      "loss": 0.0273,
+      "step": 34490
+    },
+    {
+      "action_loss": 0.005727287847548723,
+      "epoch": 31.01618705035971,
+      "step": 34490
+    },
+    {
+      "epoch": 31.01618705035971,
+      "step": 34490,
+      "torque_loss": 0.08258908987045288
+    },
+    {
+      "epoch": 31.01618705035971,
+      "force_loss": 0.005919950548559427,
+      "step": 34490
+    },
+    {
+      "epoch": 31.02517985611511,
+      "grad_norm": 0.18150265514850616,
+      "learning_rate": 4.177298869620264e-05,
+      "loss": 0.0154,
+      "step": 34500
+    },
+    {
+      "action_loss": 0.0028090246487408876,
+      "epoch": 31.02517985611511,
+      "step": 34500
+    },
+    {
+      "epoch": 31.02517985611511,
+      "step": 34500,
+      "torque_loss": 0.14135436713695526
+    },
+    {
+      "epoch": 31.02517985611511,
+      "force_loss": 0.0021016087848693132,
+      "step": 34500
+    },
+    {
+      "epoch": 31.034172661870503,
+      "grad_norm": 0.13717269897460938,
+      "learning_rate": 4.1745807719788705e-05,
+      "loss": 0.019,
+      "step": 34510
+    },
+    {
+      "action_loss": 0.0021398004610091448,
+      "epoch": 31.034172661870503,
+      "step": 34510
+    },
+    {
+      "epoch": 31.034172661870503,
+      "step": 34510,
+      "torque_loss": 0.10547112673521042
+    },
+    {
+      "epoch": 31.034172661870503,
+      "force_loss": 0.0011371915461495519,
+      "step": 34510
+    },
+    {
+      "epoch": 31.0431654676259,
+      "grad_norm": 0.1943628489971161,
+      "learning_rate": 4.1718629250780445e-05,
+      "loss": 0.0178,
+      "step": 34520
+    },
+    {
+      "action_loss": 0.0027561886236071587,
+      "epoch": 31.0431654676259,
+      "step": 34520
+    },
+    {
+      "epoch": 31.0431654676259,
+      "step": 34520,
+      "torque_loss": 0.15017829835414886
+    },
+    {
+      "epoch": 31.0431654676259,
+      "force_loss": 0.009032641537487507,
+      "step": 34520
+    },
+    {
+      "epoch": 31.052158273381295,
+      "grad_norm": 0.11110750585794449,
+      "learning_rate": 4.1691453297433956e-05,
+      "loss": 0.0179,
+      "step": 34530
+    },
+    {
+      "action_loss": 0.005441917572170496,
+      "epoch": 31.052158273381295,
+      "step": 34530
+    },
+    {
+      "epoch": 31.052158273381295,
+      "step": 34530,
+      "torque_loss": 0.14995302259922028
+    },
+    {
+      "epoch": 31.052158273381295,
+      "force_loss": 0.002294342964887619,
+      "step": 34530
+    },
+    {
+      "epoch": 31.06115107913669,
+      "grad_norm": 0.5229678750038147,
+      "learning_rate": 4.166427986800457e-05,
+      "loss": 0.0196,
+      "step": 34540
+    },
+    {
+      "action_loss": 0.004152171779423952,
+      "epoch": 31.06115107913669,
+      "step": 34540
+    },
+    {
+      "epoch": 31.06115107913669,
+      "step": 34540,
+      "torque_loss": 0.12068601697683334
+    },
+    {
+      "epoch": 31.06115107913669,
+      "force_loss": 0.0015859386185184121,
+      "step": 34540
+    },
+    {
+      "epoch": 31.070143884892087,
+      "grad_norm": 0.19499029219150543,
+      "learning_rate": 4.163710897074688e-05,
+      "loss": 0.0172,
+      "step": 34550
+    },
+    {
+      "action_loss": 0.005862883757799864,
+      "epoch": 31.070143884892087,
+      "step": 34550
+    },
+    {
+      "epoch": 31.070143884892087,
+      "step": 34550,
+      "torque_loss": 0.09602329134941101
+    },
+    {
+      "epoch": 31.070143884892087,
+      "force_loss": 0.0055970698595047,
+      "step": 34550
+    },
+    {
+      "epoch": 31.07913669064748,
+      "grad_norm": 0.47074493765830994,
+      "learning_rate": 4.1609940613914686e-05,
+      "loss": 0.0237,
+      "step": 34560
+    },
+    {
+      "action_loss": 0.006736569106578827,
+      "epoch": 31.07913669064748,
+      "step": 34560
+    },
+    {
+      "epoch": 31.07913669064748,
+      "step": 34560,
+      "torque_loss": 0.1278313547372818
+    },
+    {
+      "epoch": 31.07913669064748,
+      "force_loss": 0.004935703705996275,
+      "step": 34560
+    },
+    {
+      "epoch": 31.08812949640288,
+      "grad_norm": 0.2099589854478836,
+      "learning_rate": 4.1582774805760996e-05,
+      "loss": 0.0176,
+      "step": 34570
+    },
+    {
+      "action_loss": 0.0035369135439395905,
+      "epoch": 31.08812949640288,
+      "step": 34570
+    },
+    {
+      "epoch": 31.08812949640288,
+      "step": 34570,
+      "torque_loss": 0.1384759545326233
+    },
+    {
+      "epoch": 31.08812949640288,
+      "force_loss": 0.004823770374059677,
+      "step": 34570
+    },
+    {
+      "epoch": 31.097122302158272,
+      "grad_norm": 0.35145261883735657,
+      "learning_rate": 4.155561155453809e-05,
+      "loss": 0.0179,
+      "step": 34580
+    },
+    {
+      "action_loss": 0.0020043589174747467,
+      "epoch": 31.097122302158272,
+      "step": 34580
+    },
+    {
+      "epoch": 31.097122302158272,
+      "step": 34580,
+      "torque_loss": 0.13524475693702698
+    },
+    {
+      "epoch": 31.097122302158272,
+      "force_loss": 0.0017389642307534814,
+      "step": 34580
+    },
+    {
+      "epoch": 31.10611510791367,
+      "grad_norm": 0.2711109519004822,
+      "learning_rate": 4.15284508684974e-05,
+      "loss": 0.0207,
+      "step": 34590
+    },
+    {
+      "action_loss": 0.001917586661875248,
+      "epoch": 31.10611510791367,
+      "step": 34590
+    },
+    {
+      "epoch": 31.10611510791367,
+      "step": 34590,
+      "torque_loss": 0.14557045698165894
+    },
+    {
+      "epoch": 31.10611510791367,
+      "force_loss": 0.002514700870960951,
+      "step": 34590
+    },
+    {
+      "epoch": 31.115107913669064,
+      "grad_norm": 0.2572793662548065,
+      "learning_rate": 4.1501292755889675e-05,
+      "loss": 0.0195,
+      "step": 34600
+    },
+    {
+      "action_loss": 0.00617902772501111,
+      "epoch": 31.115107913669064,
+      "step": 34600
+    },
+    {
+      "epoch": 31.115107913669064,
+      "step": 34600,
+      "torque_loss": 0.10713726282119751
+    },
+    {
+      "epoch": 31.115107913669064,
+      "force_loss": 0.009111873805522919,
+      "step": 34600
+    },
+    {
+      "epoch": 31.12410071942446,
+      "grad_norm": 0.3115302622318268,
+      "learning_rate": 4.1474137224964833e-05,
+      "loss": 0.0208,
+      "step": 34610
+    },
+    {
+      "action_loss": 0.004423326347023249,
+      "epoch": 31.12410071942446,
+      "step": 34610
+    },
+    {
+      "epoch": 31.12410071942446,
+      "step": 34610,
+      "torque_loss": 0.14599765837192535
+    },
+    {
+      "epoch": 31.12410071942446,
+      "force_loss": 0.016374656930565834,
+      "step": 34610
+    },
+    {
+      "epoch": 31.133093525179856,
+      "grad_norm": 0.338373601436615,
+      "learning_rate": 4.144698428397197e-05,
+      "loss": 0.0211,
+      "step": 34620
+    },
+    {
+      "action_loss": 0.003914201632142067,
+      "epoch": 31.133093525179856,
+      "step": 34620
+    },
+    {
+      "epoch": 31.133093525179856,
+      "step": 34620,
+      "torque_loss": 0.1427883356809616
+    },
+    {
+      "epoch": 31.133093525179856,
+      "force_loss": 0.002898996928706765,
+      "step": 34620
+    },
+    {
+      "epoch": 31.142086330935253,
+      "grad_norm": 0.09954176098108292,
+      "learning_rate": 4.1419833941159466e-05,
+      "loss": 0.0165,
+      "step": 34630
+    },
+    {
+      "action_loss": 0.0015698192873969674,
+      "epoch": 31.142086330935253,
+      "step": 34630
+    },
+    {
+      "epoch": 31.142086330935253,
+      "step": 34630,
+      "torque_loss": 0.11284714937210083
+    },
+    {
+      "epoch": 31.142086330935253,
+      "force_loss": 0.003355605760589242,
+      "step": 34630
+    },
+    {
+      "epoch": 31.151079136690647,
+      "grad_norm": 0.24442985653877258,
+      "learning_rate": 4.1392686204774846e-05,
+      "loss": 0.0167,
+      "step": 34640
+    },
+    {
+      "action_loss": 0.001921254675835371,
+      "epoch": 31.151079136690647,
+      "step": 34640
+    },
+    {
+      "epoch": 31.151079136690647,
+      "step": 34640,
+      "torque_loss": 0.13938771188259125
+    },
+    {
+      "epoch": 31.151079136690647,
+      "force_loss": 0.0030783768743276596,
+      "step": 34640
+    },
+    {
+      "epoch": 31.16007194244604,
+      "grad_norm": 0.1379452645778656,
+      "learning_rate": 4.13655410830649e-05,
+      "loss": 0.0155,
+      "step": 34650
+    },
+    {
+      "action_loss": 0.0036887743044644594,
+      "epoch": 31.16007194244604,
+      "step": 34650
+    },
+    {
+      "epoch": 31.16007194244604,
+      "step": 34650,
+      "torque_loss": 0.16108696162700653
+    },
+    {
+      "epoch": 31.16007194244604,
+      "force_loss": 0.005511805415153503,
+      "step": 34650
+    },
+    {
+      "epoch": 31.16906474820144,
+      "grad_norm": 0.3105632960796356,
+      "learning_rate": 4.1338398584275594e-05,
+      "loss": 0.02,
+      "step": 34660
+    },
+    {
+      "action_loss": 0.016464168205857277,
+      "epoch": 31.16906474820144,
+      "step": 34660
+    },
+    {
+      "epoch": 31.16906474820144,
+      "step": 34660,
+      "torque_loss": 0.10444417595863342
+    },
+    {
+      "epoch": 31.16906474820144,
+      "force_loss": 0.03631044924259186,
+      "step": 34660
+    },
+    {
+      "epoch": 31.178057553956833,
+      "grad_norm": 0.12203385680913925,
+      "learning_rate": 4.1311258716652104e-05,
+      "loss": 0.019,
+      "step": 34670
+    },
+    {
+      "action_loss": 0.00409175269305706,
+      "epoch": 31.178057553956833,
+      "step": 34670
+    },
+    {
+      "epoch": 31.178057553956833,
+      "step": 34670,
+      "torque_loss": 0.12638358771800995
+    },
+    {
+      "epoch": 31.178057553956833,
+      "force_loss": 0.01283965166658163,
+      "step": 34670
+    },
+    {
+      "epoch": 31.18705035971223,
+      "grad_norm": 0.17293165624141693,
+      "learning_rate": 4.128412148843881e-05,
+      "loss": 0.0178,
+      "step": 34680
+    },
+    {
+      "action_loss": 0.005139355082064867,
+      "epoch": 31.18705035971223,
+      "step": 34680
+    },
+    {
+      "epoch": 31.18705035971223,
+      "step": 34680,
+      "torque_loss": 0.13874301314353943
+    },
+    {
+      "epoch": 31.18705035971223,
+      "force_loss": 0.011036957614123821,
+      "step": 34680
+    },
+    {
+      "epoch": 31.196043165467625,
+      "grad_norm": 0.2990374267101288,
+      "learning_rate": 4.125698690787926e-05,
+      "loss": 0.0202,
+      "step": 34690
+    },
+    {
+      "action_loss": 0.005877554416656494,
+      "epoch": 31.196043165467625,
+      "step": 34690
+    },
+    {
+      "epoch": 31.196043165467625,
+      "step": 34690,
+      "torque_loss": 0.13707762956619263
+    },
+    {
+      "epoch": 31.196043165467625,
+      "force_loss": 0.005564381834119558,
+      "step": 34690
+    },
+    {
+      "epoch": 31.205035971223023,
+      "grad_norm": 0.2713855504989624,
+      "learning_rate": 4.1229854983216245e-05,
+      "loss": 0.0192,
+      "step": 34700
+    },
+    {
+      "action_loss": 0.0044330330565571785,
+      "epoch": 31.205035971223023,
+      "step": 34700
+    },
+    {
+      "epoch": 31.205035971223023,
+      "step": 34700,
+      "torque_loss": 0.15669254958629608
+    },
+    {
+      "epoch": 31.205035971223023,
+      "force_loss": 0.004603779409080744,
+      "step": 34700
+    },
+    {
+      "epoch": 31.214028776978417,
+      "grad_norm": 0.2822035551071167,
+      "learning_rate": 4.120272572269175e-05,
+      "loss": 0.0195,
+      "step": 34710
+    },
+    {
+      "action_loss": 0.00781963299959898,
+      "epoch": 31.214028776978417,
+      "step": 34710
+    },
+    {
+      "epoch": 31.214028776978417,
+      "step": 34710,
+      "torque_loss": 0.1621348261833191
+    },
+    {
+      "epoch": 31.214028776978417,
+      "force_loss": 0.013986830599606037,
+      "step": 34710
+    },
+    {
+      "epoch": 31.223021582733814,
+      "grad_norm": 0.33111459016799927,
+      "learning_rate": 4.117559913454687e-05,
+      "loss": 0.0193,
+      "step": 34720
+    },
+    {
+      "action_loss": 0.0026872241869568825,
+      "epoch": 31.223021582733814,
+      "step": 34720
+    },
+    {
+      "epoch": 31.223021582733814,
+      "step": 34720,
+      "torque_loss": 0.14999866485595703
+    },
+    {
+      "epoch": 31.223021582733814,
+      "force_loss": 0.00711226649582386,
+      "step": 34720
+    },
+    {
+      "epoch": 31.23201438848921,
+      "grad_norm": 0.20210996270179749,
+      "learning_rate": 4.114847522702201e-05,
+      "loss": 0.0193,
+      "step": 34730
+    },
+    {
+      "action_loss": 0.004023254383355379,
+      "epoch": 31.23201438848921,
+      "step": 34730
+    },
+    {
+      "epoch": 31.23201438848921,
+      "step": 34730,
+      "torque_loss": 0.13857853412628174
+    },
+    {
+      "epoch": 31.23201438848921,
+      "force_loss": 0.00549306022003293,
+      "step": 34730
+    },
+    {
+      "epoch": 31.241007194244606,
+      "grad_norm": 0.1407058835029602,
+      "learning_rate": 4.112135400835664e-05,
+      "loss": 0.0169,
+      "step": 34740
+    },
+    {
+      "action_loss": 0.005574887152761221,
+      "epoch": 31.241007194244606,
+      "step": 34740
+    },
+    {
+      "epoch": 31.241007194244606,
+      "step": 34740,
+      "torque_loss": 0.10836869478225708
+    },
+    {
+      "epoch": 31.241007194244606,
+      "force_loss": 0.0074792117811739445,
+      "step": 34740
+    },
+    {
+      "epoch": 31.25,
+      "grad_norm": 0.19229404628276825,
+      "learning_rate": 4.109423548678949e-05,
+      "loss": 0.0151,
+      "step": 34750
+    },
+    {
+      "action_loss": 0.0059226867742836475,
+      "epoch": 31.25,
+      "step": 34750
+    },
+    {
+      "epoch": 31.25,
+      "step": 34750,
+      "torque_loss": 0.12453003972768784
+    },
+    {
+      "epoch": 31.25,
+      "force_loss": 0.003832560032606125,
+      "step": 34750
+    },
+    {
+      "epoch": 31.258992805755394,
+      "grad_norm": 0.35395365953445435,
+      "learning_rate": 4.106711967055848e-05,
+      "loss": 0.0178,
+      "step": 34760
+    },
+    {
+      "action_loss": 0.003903917735442519,
+      "epoch": 31.258992805755394,
+      "step": 34760
+    },
+    {
+      "epoch": 31.258992805755394,
+      "step": 34760,
+      "torque_loss": 0.09834903478622437
+    },
+    {
+      "epoch": 31.258992805755394,
+      "force_loss": 0.007758116815239191,
+      "step": 34760
+    },
+    {
+      "epoch": 31.26798561151079,
+      "grad_norm": 0.2640562355518341,
+      "learning_rate": 4.1040006567900636e-05,
+      "loss": 0.0168,
+      "step": 34770
+    },
+    {
+      "action_loss": 0.0020288738887757063,
+      "epoch": 31.26798561151079,
+      "step": 34770
+    },
+    {
+      "epoch": 31.26798561151079,
+      "step": 34770,
+      "torque_loss": 0.11705183237791061
+    },
+    {
+      "epoch": 31.26798561151079,
+      "force_loss": 0.00440516322851181,
+      "step": 34770
+    },
+    {
+      "epoch": 31.276978417266186,
+      "grad_norm": 0.4126938581466675,
+      "learning_rate": 4.101289618705224e-05,
+      "loss": 0.0158,
+      "step": 34780
+    },
+    {
+      "action_loss": 0.0020967319142073393,
+      "epoch": 31.276978417266186,
+      "step": 34780
+    },
+    {
+      "epoch": 31.276978417266186,
+      "step": 34780,
+      "torque_loss": 0.12304145097732544
+    },
+    {
+      "epoch": 31.276978417266186,
+      "force_loss": 0.001527313725091517,
+      "step": 34780
+    },
+    {
+      "epoch": 31.285971223021583,
+      "grad_norm": 0.35957178473472595,
+      "learning_rate": 4.0985788536248675e-05,
+      "loss": 0.0175,
+      "step": 34790
+    },
+    {
+      "action_loss": 0.0024529846850782633,
+      "epoch": 31.285971223021583,
+      "step": 34790
+    },
+    {
+      "epoch": 31.285971223021583,
+      "step": 34790,
+      "torque_loss": 0.12623198330402374
+    },
+    {
+      "epoch": 31.285971223021583,
+      "force_loss": 0.0035021493677049875,
+      "step": 34790
+    },
+    {
+      "epoch": 31.294964028776977,
+      "grad_norm": 0.4234926104545593,
+      "learning_rate": 4.095868362372454e-05,
+      "loss": 0.017,
+      "step": 34800
+    },
+    {
+      "action_loss": 0.0035946767311543226,
+      "epoch": 31.294964028776977,
+      "step": 34800
+    },
+    {
+      "epoch": 31.294964028776977,
+      "step": 34800,
+      "torque_loss": 0.13089805841445923
+    },
+    {
+      "epoch": 31.294964028776977,
+      "force_loss": 0.001584723242558539,
+      "step": 34800
+    },
+    {
+      "epoch": 31.303956834532375,
+      "grad_norm": 0.33862611651420593,
+      "learning_rate": 4.0931581457713614e-05,
+      "loss": 0.0178,
+      "step": 34810
+    },
+    {
+      "action_loss": 0.007343566045165062,
+      "epoch": 31.303956834532375,
+      "step": 34810
+    },
+    {
+      "epoch": 31.303956834532375,
+      "step": 34810,
+      "torque_loss": 0.14513994753360748
+    },
+    {
+      "epoch": 31.303956834532375,
+      "force_loss": 0.008060489781200886,
+      "step": 34810
+    },
+    {
+      "epoch": 31.31294964028777,
+      "grad_norm": 0.15803027153015137,
+      "learning_rate": 4.09044820464488e-05,
+      "loss": 0.0188,
+      "step": 34820
+    },
+    {
+      "action_loss": 0.0023536905646324158,
+      "epoch": 31.31294964028777,
+      "step": 34820
+    },
+    {
+      "epoch": 31.31294964028777,
+      "step": 34820,
+      "torque_loss": 0.13476301729679108
+    },
+    {
+      "epoch": 31.31294964028777,
+      "force_loss": 0.0015555074205622077,
+      "step": 34820
+    },
+    {
+      "epoch": 31.321942446043167,
+      "grad_norm": 0.18544569611549377,
+      "learning_rate": 4.087738539816219e-05,
+      "loss": 0.0176,
+      "step": 34830
+    },
+    {
+      "action_loss": 0.01503182202577591,
+      "epoch": 31.321942446043167,
+      "step": 34830
+    },
+    {
+      "epoch": 31.321942446043167,
+      "step": 34830,
+      "torque_loss": 0.14875905215740204
+    },
+    {
+      "epoch": 31.321942446043167,
+      "force_loss": 0.017478706315159798,
+      "step": 34830
+    },
+    {
+      "epoch": 31.33093525179856,
+      "grad_norm": 1.007502555847168,
+      "learning_rate": 4.085029152108501e-05,
+      "loss": 0.022,
+      "step": 34840
+    },
+    {
+      "action_loss": 0.0013019072357565165,
+      "epoch": 31.33093525179856,
+      "step": 34840
+    },
+    {
+      "epoch": 31.33093525179856,
+      "step": 34840,
+      "torque_loss": 0.12316927313804626
+    },
+    {
+      "epoch": 31.33093525179856,
+      "force_loss": 0.0011967908358201385,
+      "step": 34840
+    },
+    {
+      "epoch": 31.33992805755396,
+      "grad_norm": 0.41491687297821045,
+      "learning_rate": 4.0823200423447714e-05,
+      "loss": 0.0196,
+      "step": 34850
+    },
+    {
+      "action_loss": 0.0028079384937882423,
+      "epoch": 31.33992805755396,
+      "step": 34850
+    },
+    {
+      "epoch": 31.33992805755396,
+      "step": 34850,
+      "torque_loss": 0.13180182874202728
+    },
+    {
+      "epoch": 31.33992805755396,
+      "force_loss": 0.004840850364416838,
+      "step": 34850
+    },
+    {
+      "epoch": 31.348920863309353,
+      "grad_norm": 0.1371115744113922,
+      "learning_rate": 4.079611211347981e-05,
+      "loss": 0.0241,
+      "step": 34860
+    },
+    {
+      "action_loss": 0.007915817201137543,
+      "epoch": 31.348920863309353,
+      "step": 34860
+    },
+    {
+      "epoch": 31.348920863309353,
+      "step": 34860,
+      "torque_loss": 0.15225368738174438
+    },
+    {
+      "epoch": 31.348920863309353,
+      "force_loss": 0.006319840904325247,
+      "step": 34860
+    },
+    {
+      "epoch": 31.357913669064747,
+      "grad_norm": 0.35097813606262207,
+      "learning_rate": 4.076902659941002e-05,
+      "loss": 0.0167,
+      "step": 34870
+    },
+    {
+      "action_loss": 0.002877927152439952,
+      "epoch": 31.357913669064747,
+      "step": 34870
+    },
+    {
+      "epoch": 31.357913669064747,
+      "step": 34870,
+      "torque_loss": 0.12672726809978485
+    },
+    {
+      "epoch": 31.357913669064747,
+      "force_loss": 0.0036767490673810244,
+      "step": 34870
+    },
+    {
+      "epoch": 31.366906474820144,
+      "grad_norm": 0.3157791495323181,
+      "learning_rate": 4.074194388946624e-05,
+      "loss": 0.0157,
+      "step": 34880
+    },
+    {
+      "action_loss": 0.0014395791804417968,
+      "epoch": 31.366906474820144,
+      "step": 34880
+    },
+    {
+      "epoch": 31.366906474820144,
+      "step": 34880,
+      "torque_loss": 0.12489987164735794
+    },
+    {
+      "epoch": 31.366906474820144,
+      "force_loss": 0.001917771645821631,
+      "step": 34880
+    },
+    {
+      "epoch": 31.37589928057554,
+      "grad_norm": 0.38226935267448425,
+      "learning_rate": 4.071486399187545e-05,
+      "loss": 0.0164,
+      "step": 34890
+    },
+    {
+      "action_loss": 0.008211738429963589,
+      "epoch": 31.37589928057554,
+      "step": 34890
+    },
+    {
+      "epoch": 31.37589928057554,
+      "step": 34890,
+      "torque_loss": 0.12687841057777405
+    },
+    {
+      "epoch": 31.37589928057554,
+      "force_loss": 0.010581615380942822,
+      "step": 34890
+    },
+    {
+      "epoch": 31.384892086330936,
+      "grad_norm": 0.2458968311548233,
+      "learning_rate": 4.0687786914863836e-05,
+      "loss": 0.0185,
+      "step": 34900
+    },
+    {
+      "action_loss": 0.0017622933955863118,
+      "epoch": 31.384892086330936,
+      "step": 34900
+    },
+    {
+      "epoch": 31.384892086330936,
+      "step": 34900,
+      "torque_loss": 0.10126352310180664
+    },
+    {
+      "epoch": 31.384892086330936,
+      "force_loss": 0.0025832951068878174,
+      "step": 34900
+    },
+    {
+      "epoch": 31.39388489208633,
+      "grad_norm": 0.272097110748291,
+      "learning_rate": 4.0660712666656666e-05,
+      "loss": 0.0157,
+      "step": 34910
+    },
+    {
+      "action_loss": 0.005375581327825785,
+      "epoch": 31.39388489208633,
+      "step": 34910
+    },
+    {
+      "epoch": 31.39388489208633,
+      "step": 34910,
+      "torque_loss": 0.14981211721897125
+    },
+    {
+      "epoch": 31.39388489208633,
+      "force_loss": 0.00587962381541729,
+      "step": 34910
+    },
+    {
+      "epoch": 31.402877697841728,
+      "grad_norm": 0.18457461893558502,
+      "learning_rate": 4.0633641255478394e-05,
+      "loss": 0.0207,
+      "step": 34920
+    },
+    {
+      "action_loss": 0.0011194973485544324,
+      "epoch": 31.402877697841728,
+      "step": 34920
+    },
+    {
+      "epoch": 31.402877697841728,
+      "step": 34920,
+      "torque_loss": 0.09443652629852295
+    },
+    {
+      "epoch": 31.402877697841728,
+      "force_loss": 0.0017360160127282143,
+      "step": 34920
+    },
+    {
+      "epoch": 31.41187050359712,
+      "grad_norm": 0.1195453479886055,
+      "learning_rate": 4.0606572689552624e-05,
+      "loss": 0.0154,
+      "step": 34930
+    },
+    {
+      "action_loss": 0.0014406010741367936,
+      "epoch": 31.41187050359712,
+      "step": 34930
+    },
+    {
+      "epoch": 31.41187050359712,
+      "step": 34930,
+      "torque_loss": 0.13830390572547913
+    },
+    {
+      "epoch": 31.41187050359712,
+      "force_loss": 0.0007887175306677818,
+      "step": 34930
+    },
+    {
+      "epoch": 31.42086330935252,
+      "grad_norm": 0.270561158657074,
+      "learning_rate": 4.0579506977102036e-05,
+      "loss": 0.0187,
+      "step": 34940
+    },
+    {
+      "action_loss": 0.0025075108278542757,
+      "epoch": 31.42086330935252,
+      "step": 34940
+    },
+    {
+      "epoch": 31.42086330935252,
+      "step": 34940,
+      "torque_loss": 0.15471245348453522
+    },
+    {
+      "epoch": 31.42086330935252,
+      "force_loss": 0.004849970806390047,
+      "step": 34940
+    },
+    {
+      "epoch": 31.429856115107913,
+      "grad_norm": 0.6065011620521545,
+      "learning_rate": 4.055244412634849e-05,
+      "loss": 0.0186,
+      "step": 34950
+    },
+    {
+      "action_loss": 0.003092721104621887,
+      "epoch": 31.429856115107913,
+      "step": 34950
+    },
+    {
+      "epoch": 31.429856115107913,
+      "step": 34950,
+      "torque_loss": 0.1049649640917778
+    },
+    {
+      "epoch": 31.429856115107913,
+      "force_loss": 0.0021366889122873545,
+      "step": 34950
+    },
+    {
+      "epoch": 31.43884892086331,
+      "grad_norm": 0.14073814451694489,
+      "learning_rate": 4.052538414551298e-05,
+      "loss": 0.0189,
+      "step": 34960
+    },
+    {
+      "action_loss": 0.012841415591537952,
+      "epoch": 31.43884892086331,
+      "step": 34960
+    },
+    {
+      "epoch": 31.43884892086331,
+      "step": 34960,
+      "torque_loss": 0.15012070536613464
+    },
+    {
+      "epoch": 31.43884892086331,
+      "force_loss": 0.0143703892827034,
+      "step": 34960
+    },
+    {
+      "epoch": 31.447841726618705,
+      "grad_norm": 0.11514277011156082,
+      "learning_rate": 4.0498327042815596e-05,
+      "loss": 0.0169,
+      "step": 34970
+    },
+    {
+      "action_loss": 0.0035906240809708834,
+      "epoch": 31.447841726618705,
+      "step": 34970
+    },
+    {
+      "epoch": 31.447841726618705,
+      "step": 34970,
+      "torque_loss": 0.14721103012561798
+    },
+    {
+      "epoch": 31.447841726618705,
+      "force_loss": 0.005858649965375662,
+      "step": 34970
+    },
+    {
+      "epoch": 31.4568345323741,
+      "grad_norm": 0.20259389281272888,
+      "learning_rate": 4.047127282647559e-05,
+      "loss": 0.0179,
+      "step": 34980
+    },
+    {
+      "action_loss": 0.003659149631857872,
+      "epoch": 31.4568345323741,
+      "step": 34980
+    },
+    {
+      "epoch": 31.4568345323741,
+      "step": 34980,
+      "torque_loss": 0.09909810870885849
+    },
+    {
+      "epoch": 31.4568345323741,
+      "force_loss": 0.0038765103090554476,
+      "step": 34980
+    },
+    {
+      "epoch": 31.465827338129497,
+      "grad_norm": 0.12645934522151947,
+      "learning_rate": 4.04442215047113e-05,
+      "loss": 0.0167,
+      "step": 34990
+    },
+    {
+      "action_loss": 0.0025366703048348427,
+      "epoch": 31.465827338129497,
+      "step": 34990
+    },
+    {
+      "epoch": 31.465827338129497,
+      "step": 34990,
+      "torque_loss": 0.12167666107416153
+    },
+    {
+      "epoch": 31.465827338129497,
+      "force_loss": 0.005248036235570908,
+      "step": 34990
+    },
+    {
+      "epoch": 31.47482014388489,
+      "grad_norm": 0.15403643250465393,
+      "learning_rate": 4.041717308574023e-05,
+      "loss": 0.0153,
+      "step": 35000
+    },
+    {
+      "action_loss": 0.004632141441106796,
+      "epoch": 31.47482014388489,
+      "step": 35000
+    },
+    {
+      "epoch": 31.47482014388489,
+      "step": 35000,
+      "torque_loss": 0.1133936420083046
+    },
+    {
+      "epoch": 31.47482014388489,
+      "force_loss": 0.005289702210575342,
+      "step": 35000
+    },
+    {
+      "epoch": 31.48381294964029,
+      "grad_norm": 0.1703784316778183,
+      "learning_rate": 4.039012757777893e-05,
+      "loss": 0.0177,
+      "step": 35010
+    },
+    {
+      "action_loss": 0.0033064030576497316,
+      "epoch": 31.48381294964029,
+      "step": 35010
+    },
+    {
+      "epoch": 31.48381294964029,
+      "step": 35010,
+      "torque_loss": 0.10487169027328491
+    },
+    {
+      "epoch": 31.48381294964029,
+      "force_loss": 0.0018594572320580482,
+      "step": 35010
+    },
+    {
+      "epoch": 31.492805755395683,
+      "grad_norm": 0.11044184863567352,
+      "learning_rate": 4.036308498904314e-05,
+      "loss": 0.0149,
+      "step": 35020
+    },
+    {
+      "action_loss": 0.0055665522813797,
+      "epoch": 31.492805755395683,
+      "step": 35020
+    },
+    {
+      "epoch": 31.492805755395683,
+      "step": 35020,
+      "torque_loss": 0.1490526795387268
+    },
+    {
+      "epoch": 31.492805755395683,
+      "force_loss": 0.005140155553817749,
+      "step": 35020
+    },
+    {
+      "epoch": 31.50179856115108,
+      "grad_norm": 0.31837034225463867,
+      "learning_rate": 4.033604532774771e-05,
+      "loss": 0.0186,
+      "step": 35030
+    },
+    {
+      "action_loss": 0.003204680746421218,
+      "epoch": 31.50179856115108,
+      "step": 35030
+    },
+    {
+      "epoch": 31.50179856115108,
+      "step": 35030,
+      "torque_loss": 0.15786059200763702
+    },
+    {
+      "epoch": 31.50179856115108,
+      "force_loss": 0.0030774895567446947,
+      "step": 35030
+    },
+    {
+      "epoch": 31.510791366906474,
+      "grad_norm": 0.20702697336673737,
+      "learning_rate": 4.030900860210652e-05,
+      "loss": 0.0161,
+      "step": 35040
+    },
+    {
+      "action_loss": 0.015199624001979828,
+      "epoch": 31.510791366906474,
+      "step": 35040
+    },
+    {
+      "epoch": 31.510791366906474,
+      "step": 35040,
+      "torque_loss": 0.13615868985652924
+    },
+    {
+      "epoch": 31.510791366906474,
+      "force_loss": 0.013585063628852367,
+      "step": 35040
+    },
+    {
+      "epoch": 31.519784172661872,
+      "grad_norm": 0.1472933441400528,
+      "learning_rate": 4.028197482033266e-05,
+      "loss": 0.018,
+      "step": 35050
+    },
+    {
+      "action_loss": 0.002620900049805641,
+      "epoch": 31.519784172661872,
+      "step": 35050
+    },
+    {
+      "epoch": 31.519784172661872,
+      "step": 35050,
+      "torque_loss": 0.14953002333641052
+    },
+    {
+      "epoch": 31.519784172661872,
+      "force_loss": 0.01394414808601141,
+      "step": 35050
+    },
+    {
+      "epoch": 31.528776978417266,
+      "grad_norm": 0.3911173343658447,
+      "learning_rate": 4.0254943990638246e-05,
+      "loss": 0.0189,
+      "step": 35060
+    },
+    {
+      "action_loss": 0.0028132274746894836,
+      "epoch": 31.528776978417266,
+      "step": 35060
+    },
+    {
+      "epoch": 31.528776978417266,
+      "step": 35060,
+      "torque_loss": 0.11279907077550888
+    },
+    {
+      "epoch": 31.528776978417266,
+      "force_loss": 0.003741770749911666,
+      "step": 35060
+    },
+    {
+      "epoch": 31.53776978417266,
+      "grad_norm": 0.21598149836063385,
+      "learning_rate": 4.022791612123454e-05,
+      "loss": 0.0252,
+      "step": 35070
+    },
+    {
+      "action_loss": 0.0020898503717035055,
+      "epoch": 31.53776978417266,
+      "step": 35070
+    },
+    {
+      "epoch": 31.53776978417266,
+      "step": 35070,
+      "torque_loss": 0.13717205822467804
+    },
+    {
+      "epoch": 31.53776978417266,
+      "force_loss": 0.002998681738972664,
+      "step": 35070
+    },
+    {
+      "epoch": 31.546762589928058,
+      "grad_norm": 0.14628909528255463,
+      "learning_rate": 4.020089122033192e-05,
+      "loss": 0.0182,
+      "step": 35080
+    },
+    {
+      "action_loss": 0.007679073140025139,
+      "epoch": 31.546762589928058,
+      "step": 35080
+    },
+    {
+      "epoch": 31.546762589928058,
+      "step": 35080,
+      "torque_loss": 0.09587617963552475
+    },
+    {
+      "epoch": 31.546762589928058,
+      "force_loss": 0.002372180810198188,
+      "step": 35080
+    },
+    {
+      "epoch": 31.555755395683452,
+      "grad_norm": 0.20146098732948303,
+      "learning_rate": 4.01738692961398e-05,
+      "loss": 0.0245,
+      "step": 35090
+    },
+    {
+      "action_loss": 0.00114073243457824,
+      "epoch": 31.555755395683452,
+      "step": 35090
+    },
+    {
+      "epoch": 31.555755395683452,
+      "step": 35090,
+      "torque_loss": 0.0986347496509552
+    },
+    {
+      "epoch": 31.555755395683452,
+      "force_loss": 0.0011868586298078299,
+      "step": 35090
+    },
+    {
+      "epoch": 31.56474820143885,
+      "grad_norm": 0.42345449328422546,
+      "learning_rate": 4.014685035686675e-05,
+      "loss": 0.016,
+      "step": 35100
+    },
+    {
+      "action_loss": 0.004441409837454557,
+      "epoch": 31.56474820143885,
+      "step": 35100
+    },
+    {
+      "epoch": 31.56474820143885,
+      "step": 35100,
+      "torque_loss": 0.10088644176721573
+    },
+    {
+      "epoch": 31.56474820143885,
+      "force_loss": 0.007341319229453802,
+      "step": 35100
+    },
+    {
+      "epoch": 31.573741007194243,
+      "grad_norm": 0.14468078315258026,
+      "learning_rate": 4.011983441072039e-05,
+      "loss": 0.0191,
+      "step": 35110
+    },
+    {
+      "action_loss": 0.004453874658793211,
+      "epoch": 31.573741007194243,
+      "step": 35110
+    },
+    {
+      "epoch": 31.573741007194243,
+      "step": 35110,
+      "torque_loss": 0.12231036275625229
+    },
+    {
+      "epoch": 31.573741007194243,
+      "force_loss": 0.004258755128830671,
+      "step": 35110
+    },
+    {
+      "epoch": 31.58273381294964,
+      "grad_norm": 0.3526107370853424,
+      "learning_rate": 4.0092821465907485e-05,
+      "loss": 0.0212,
+      "step": 35120
+    },
+    {
+      "action_loss": 0.0022445085924118757,
+      "epoch": 31.58273381294964,
+      "step": 35120
+    },
+    {
+      "epoch": 31.58273381294964,
+      "step": 35120,
+      "torque_loss": 0.10221400111913681
+    },
+    {
+      "epoch": 31.58273381294964,
+      "force_loss": 0.006604932714253664,
+      "step": 35120
+    },
+    {
+      "epoch": 31.591726618705035,
+      "grad_norm": 0.32770583033561707,
+      "learning_rate": 4.006581153063383e-05,
+      "loss": 0.0163,
+      "step": 35130
+    },
+    {
+      "action_loss": 0.0018231518333777785,
+      "epoch": 31.591726618705035,
+      "step": 35130
+    },
+    {
+      "epoch": 31.591726618705035,
+      "step": 35130,
+      "torque_loss": 0.1334356814622879
+    },
+    {
+      "epoch": 31.591726618705035,
+      "force_loss": 0.0011773643782362342,
+      "step": 35130
+    },
+    {
+      "epoch": 31.600719424460433,
+      "grad_norm": 0.3421180844306946,
+      "learning_rate": 4.003880461310432e-05,
+      "loss": 0.0196,
+      "step": 35140
+    },
+    {
+      "action_loss": 0.0029285720083862543,
+      "epoch": 31.600719424460433,
+      "step": 35140
+    },
+    {
+      "epoch": 31.600719424460433,
+      "step": 35140,
+      "torque_loss": 0.14738458395004272
+    },
+    {
+      "epoch": 31.600719424460433,
+      "force_loss": 0.0024381878320127726,
+      "step": 35140
+    },
+    {
+      "epoch": 31.609712230215827,
+      "grad_norm": 0.6190105080604553,
+      "learning_rate": 4.001180072152298e-05,
+      "loss": 0.0196,
+      "step": 35150
+    },
+    {
+      "action_loss": 0.0028955843299627304,
+      "epoch": 31.609712230215827,
+      "step": 35150
+    },
+    {
+      "epoch": 31.609712230215827,
+      "step": 35150,
+      "torque_loss": 0.16135770082473755
+    },
+    {
+      "epoch": 31.609712230215827,
+      "force_loss": 0.002871215110644698,
+      "step": 35150
+    },
+    {
+      "epoch": 31.618705035971225,
+      "grad_norm": 0.2656359076499939,
+      "learning_rate": 3.998479986409285e-05,
+      "loss": 0.0192,
+      "step": 35160
+    },
+    {
+      "action_loss": 0.002504608826711774,
+      "epoch": 31.618705035971225,
+      "step": 35160
+    },
+    {
+      "epoch": 31.618705035971225,
+      "step": 35160,
+      "torque_loss": 0.11583981662988663
+    },
+    {
+      "epoch": 31.618705035971225,
+      "force_loss": 0.009636103175580502,
+      "step": 35160
+    },
+    {
+      "epoch": 31.62769784172662,
+      "grad_norm": 0.10041473060846329,
+      "learning_rate": 3.995780204901607e-05,
+      "loss": 0.0152,
+      "step": 35170
+    },
+    {
+      "action_loss": 0.0036317389458417892,
+      "epoch": 31.62769784172662,
+      "step": 35170
+    },
+    {
+      "epoch": 31.62769784172662,
+      "step": 35170,
+      "torque_loss": 0.12671585381031036
+    },
+    {
+      "epoch": 31.62769784172662,
+      "force_loss": 0.006318715866655111,
+      "step": 35170
+    },
+    {
+      "epoch": 31.636690647482013,
+      "grad_norm": 0.1791400909423828,
+      "learning_rate": 3.993080728449391e-05,
+      "loss": 0.0169,
+      "step": 35180
+    },
+    {
+      "action_loss": 0.0016373790567740798,
+      "epoch": 31.636690647482013,
+      "step": 35180
+    },
+    {
+      "epoch": 31.636690647482013,
+      "step": 35180,
+      "torque_loss": 0.1573609709739685
+    },
+    {
+      "epoch": 31.636690647482013,
+      "force_loss": 0.002217449015006423,
+      "step": 35180
+    },
+    {
+      "epoch": 31.64568345323741,
+      "grad_norm": 0.17255602777004242,
+      "learning_rate": 3.990381557872661e-05,
+      "loss": 0.0155,
+      "step": 35190
+    },
+    {
+      "action_loss": 0.006632767617702484,
+      "epoch": 31.64568345323741,
+      "step": 35190
+    },
+    {
+      "epoch": 31.64568345323741,
+      "step": 35190,
+      "torque_loss": 0.1648382991552353
+    },
+    {
+      "epoch": 31.64568345323741,
+      "force_loss": 0.006704200059175491,
+      "step": 35190
+    },
+    {
+      "epoch": 31.654676258992804,
+      "grad_norm": 0.12096426635980606,
+      "learning_rate": 3.987682693991359e-05,
+      "loss": 0.017,
+      "step": 35200
+    },
+    {
+      "action_loss": 0.004801938775926828,
+      "epoch": 31.654676258992804,
+      "step": 35200
+    },
+    {
+      "epoch": 31.654676258992804,
+      "step": 35200,
+      "torque_loss": 0.1270293891429901
+    },
+    {
+      "epoch": 31.654676258992804,
+      "force_loss": 0.005421049892902374,
+      "step": 35200
+    },
+    {
+      "epoch": 31.663669064748202,
+      "grad_norm": 0.10935506969690323,
+      "learning_rate": 3.9849841376253226e-05,
+      "loss": 0.015,
+      "step": 35210
+    },
+    {
+      "action_loss": 0.0019251549383625388,
+      "epoch": 31.663669064748202,
+      "step": 35210
+    },
+    {
+      "epoch": 31.663669064748202,
+      "step": 35210,
+      "torque_loss": 0.14209115505218506
+    },
+    {
+      "epoch": 31.663669064748202,
+      "force_loss": 0.0015115119749680161,
+      "step": 35210
+    },
+    {
+      "epoch": 31.672661870503596,
+      "grad_norm": 0.15032987296581268,
+      "learning_rate": 3.982285889594306e-05,
+      "loss": 0.0191,
+      "step": 35220
+    },
+    {
+      "action_loss": 0.0037692366167902946,
+      "epoch": 31.672661870503596,
+      "step": 35220
+    },
+    {
+      "epoch": 31.672661870503596,
+      "step": 35220,
+      "torque_loss": 0.11902087926864624
+    },
+    {
+      "epoch": 31.672661870503596,
+      "force_loss": 0.0031990210991352797,
+      "step": 35220
+    },
+    {
+      "epoch": 31.681654676258994,
+      "grad_norm": 0.5698276162147522,
+      "learning_rate": 3.9795879507179665e-05,
+      "loss": 0.0174,
+      "step": 35230
+    },
+    {
+      "action_loss": 0.008475713431835175,
+      "epoch": 31.681654676258994,
+      "step": 35230
+    },
+    {
+      "epoch": 31.681654676258994,
+      "step": 35230,
+      "torque_loss": 0.11896976828575134
+    },
+    {
+      "epoch": 31.681654676258994,
+      "force_loss": 0.011556281708180904,
+      "step": 35230
+    },
+    {
+      "epoch": 31.690647482014388,
+      "grad_norm": 0.07939211279153824,
+      "learning_rate": 3.9768903218158634e-05,
+      "loss": 0.0167,
+      "step": 35240
+    },
+    {
+      "action_loss": 0.003014965681359172,
+      "epoch": 31.690647482014388,
+      "step": 35240
+    },
+    {
+      "epoch": 31.690647482014388,
+      "step": 35240,
+      "torque_loss": 0.13165317475795746
+    },
+    {
+      "epoch": 31.690647482014388,
+      "force_loss": 0.0021797276567667723,
+      "step": 35240
+    },
+    {
+      "epoch": 31.699640287769785,
+      "grad_norm": 0.17850062251091003,
+      "learning_rate": 3.974193003707468e-05,
+      "loss": 0.0178,
+      "step": 35250
+    },
+    {
+      "action_loss": 0.0028750200290232897,
+      "epoch": 31.699640287769785,
+      "step": 35250
+    },
+    {
+      "epoch": 31.699640287769785,
+      "step": 35250,
+      "torque_loss": 0.12155556678771973
+    },
+    {
+      "epoch": 31.699640287769785,
+      "force_loss": 0.004737237934023142,
+      "step": 35250
+    },
+    {
+      "epoch": 31.70863309352518,
+      "grad_norm": 0.17754119634628296,
+      "learning_rate": 3.971495997212152e-05,
+      "loss": 0.016,
+      "step": 35260
+    },
+    {
+      "action_loss": 0.0023565383162349463,
+      "epoch": 31.70863309352518,
+      "step": 35260
+    },
+    {
+      "epoch": 31.70863309352518,
+      "step": 35260,
+      "torque_loss": 0.1414964497089386
+    },
+    {
+      "epoch": 31.70863309352518,
+      "force_loss": 0.0033865266013890505,
+      "step": 35260
+    },
+    {
+      "epoch": 31.717625899280577,
+      "grad_norm": 0.6097290515899658,
+      "learning_rate": 3.9687993031491985e-05,
+      "loss": 0.018,
+      "step": 35270
+    },
+    {
+      "action_loss": 0.005043547600507736,
+      "epoch": 31.717625899280577,
+      "step": 35270
+    },
+    {
+      "epoch": 31.717625899280577,
+      "step": 35270,
+      "torque_loss": 0.17120970785617828
+    },
+    {
+      "epoch": 31.717625899280577,
+      "force_loss": 0.007256325799971819,
+      "step": 35270
+    },
+    {
+      "epoch": 31.72661870503597,
+      "grad_norm": 0.3341626226902008,
+      "learning_rate": 3.966102922337787e-05,
+      "loss": 0.0196,
+      "step": 35280
+    },
+    {
+      "action_loss": 0.0037151221185922623,
+      "epoch": 31.72661870503597,
+      "step": 35280
+    },
+    {
+      "epoch": 31.72661870503597,
+      "step": 35280,
+      "torque_loss": 0.1417021006345749
+    },
+    {
+      "epoch": 31.72661870503597,
+      "force_loss": 0.003333625616505742,
+      "step": 35280
+    },
+    {
+      "epoch": 31.735611510791365,
+      "grad_norm": 0.1533995270729065,
+      "learning_rate": 3.963406855597009e-05,
+      "loss": 0.0174,
+      "step": 35290
+    },
+    {
+      "action_loss": 0.00558444345369935,
+      "epoch": 31.735611510791365,
+      "step": 35290
+    },
+    {
+      "epoch": 31.735611510791365,
+      "step": 35290,
+      "torque_loss": 0.16073264181613922
+    },
+    {
+      "epoch": 31.735611510791365,
+      "force_loss": 0.005496786441653967,
+      "step": 35290
+    },
+    {
+      "epoch": 31.744604316546763,
+      "grad_norm": 0.20038332045078278,
+      "learning_rate": 3.960711103745861e-05,
+      "loss": 0.0183,
+      "step": 35300
+    },
+    {
+      "action_loss": 0.005106428172439337,
+      "epoch": 31.744604316546763,
+      "step": 35300
+    },
+    {
+      "epoch": 31.744604316546763,
+      "step": 35300,
+      "torque_loss": 0.12848752737045288
+    },
+    {
+      "epoch": 31.744604316546763,
+      "force_loss": 0.007534404750913382,
+      "step": 35300
+    },
+    {
+      "epoch": 31.753597122302157,
+      "grad_norm": 0.3179410398006439,
+      "learning_rate": 3.958015667603237e-05,
+      "loss": 0.019,
+      "step": 35310
+    },
+    {
+      "action_loss": 0.007186559494584799,
+      "epoch": 31.753597122302157,
+      "step": 35310
+    },
+    {
+      "epoch": 31.753597122302157,
+      "step": 35310,
+      "torque_loss": 0.13391290605068207
+    },
+    {
+      "epoch": 31.753597122302157,
+      "force_loss": 0.013319111429154873,
+      "step": 35310
+    },
+    {
+      "epoch": 31.762589928057555,
+      "grad_norm": 0.29443496465682983,
+      "learning_rate": 3.955320547987943e-05,
+      "loss": 0.018,
+      "step": 35320
+    },
+    {
+      "action_loss": 0.0034029714297503233,
+      "epoch": 31.762589928057555,
+      "step": 35320
+    },
+    {
+      "epoch": 31.762589928057555,
+      "step": 35320,
+      "torque_loss": 0.08257120102643967
+    },
+    {
+      "epoch": 31.762589928057555,
+      "force_loss": 0.005713058169931173,
+      "step": 35320
+    },
+    {
+      "epoch": 31.77158273381295,
+      "grad_norm": 0.9021161198616028,
+      "learning_rate": 3.952625745718681e-05,
+      "loss": 0.025,
+      "step": 35330
+    },
+    {
+      "action_loss": 0.0028725943993777037,
+      "epoch": 31.77158273381295,
+      "step": 35330
+    },
+    {
+      "epoch": 31.77158273381295,
+      "step": 35330,
+      "torque_loss": 0.08668620139360428
+    },
+    {
+      "epoch": 31.77158273381295,
+      "force_loss": 0.00623175548389554,
+      "step": 35330
+    },
+    {
+      "epoch": 31.780575539568346,
+      "grad_norm": 0.35724562406539917,
+      "learning_rate": 3.949931261614064e-05,
+      "loss": 0.0197,
+      "step": 35340
+    },
+    {
+      "action_loss": 0.0032996945083141327,
+      "epoch": 31.780575539568346,
+      "step": 35340
+    },
+    {
+      "epoch": 31.780575539568346,
+      "step": 35340,
+      "torque_loss": 0.10863769799470901
+    },
+    {
+      "epoch": 31.780575539568346,
+      "force_loss": 0.004071838688105345,
+      "step": 35340
+    },
+    {
+      "epoch": 31.78956834532374,
+      "grad_norm": 0.6367750763893127,
+      "learning_rate": 3.947237096492605e-05,
+      "loss": 0.0161,
+      "step": 35350
+    },
+    {
+      "action_loss": 0.008309929631650448,
+      "epoch": 31.78956834532374,
+      "step": 35350
+    },
+    {
+      "epoch": 31.78956834532374,
+      "step": 35350,
+      "torque_loss": 0.10108502954244614
+    },
+    {
+      "epoch": 31.78956834532374,
+      "force_loss": 0.0099371038377285,
+      "step": 35350
+    },
+    {
+      "epoch": 31.798561151079138,
+      "grad_norm": 0.27515581250190735,
+      "learning_rate": 3.944543251172719e-05,
+      "loss": 0.0182,
+      "step": 35360
+    },
+    {
+      "action_loss": 0.0035193238873034716,
+      "epoch": 31.798561151079138,
+      "step": 35360
+    },
+    {
+      "epoch": 31.798561151079138,
+      "step": 35360,
+      "torque_loss": 0.12900541722774506
+    },
+    {
+      "epoch": 31.798561151079138,
+      "force_loss": 0.0037006158381700516,
+      "step": 35360
+    },
+    {
+      "epoch": 31.807553956834532,
+      "grad_norm": 0.19117578864097595,
+      "learning_rate": 3.941849726472725e-05,
+      "loss": 0.0166,
+      "step": 35370
+    },
+    {
+      "action_loss": 0.010736651718616486,
+      "epoch": 31.807553956834532,
+      "step": 35370
+    },
+    {
+      "epoch": 31.807553956834532,
+      "step": 35370,
+      "torque_loss": 0.17377400398254395
+    },
+    {
+      "epoch": 31.807553956834532,
+      "force_loss": 0.007627354934811592,
+      "step": 35370
+    },
+    {
+      "epoch": 31.81654676258993,
+      "grad_norm": 0.1995638906955719,
+      "learning_rate": 3.939156523210846e-05,
+      "loss": 0.0159,
+      "step": 35380
+    },
+    {
+      "action_loss": 0.014721316285431385,
+      "epoch": 31.81654676258993,
+      "step": 35380
+    },
+    {
+      "epoch": 31.81654676258993,
+      "step": 35380,
+      "torque_loss": 0.1905713826417923
+    },
+    {
+      "epoch": 31.81654676258993,
+      "force_loss": 0.015707792714238167,
+      "step": 35380
+    },
+    {
+      "epoch": 31.825539568345324,
+      "grad_norm": 0.23033098876476288,
+      "learning_rate": 3.9364636422052046e-05,
+      "loss": 0.0225,
+      "step": 35390
+    },
+    {
+      "action_loss": 0.0027905143797397614,
+      "epoch": 31.825539568345324,
+      "step": 35390
+    },
+    {
+      "epoch": 31.825539568345324,
+      "step": 35390,
+      "torque_loss": 0.1308002918958664
+    },
+    {
+      "epoch": 31.825539568345324,
+      "force_loss": 0.005041670519858599,
+      "step": 35390
+    },
+    {
+      "epoch": 31.834532374100718,
+      "grad_norm": 0.1351657509803772,
+      "learning_rate": 3.933771084273828e-05,
+      "loss": 0.0233,
+      "step": 35400
+    },
+    {
+      "action_loss": 0.008600800298154354,
+      "epoch": 31.834532374100718,
+      "step": 35400
+    },
+    {
+      "epoch": 31.834532374100718,
+      "step": 35400,
+      "torque_loss": 0.12727564573287964
+    },
+    {
+      "epoch": 31.834532374100718,
+      "force_loss": 0.004926920402795076,
+      "step": 35400
+    },
+    {
+      "epoch": 31.843525179856115,
+      "grad_norm": 0.5700240135192871,
+      "learning_rate": 3.931078850234643e-05,
+      "loss": 0.0196,
+      "step": 35410
+    },
+    {
+      "action_loss": 0.0034690566826611757,
+      "epoch": 31.843525179856115,
+      "step": 35410
+    },
+    {
+      "epoch": 31.843525179856115,
+      "step": 35410,
+      "torque_loss": 0.10629952698945999
+    },
+    {
+      "epoch": 31.843525179856115,
+      "force_loss": 0.0024806184228509665,
+      "step": 35410
+    },
+    {
+      "epoch": 31.85251798561151,
+      "grad_norm": 0.2264889031648636,
+      "learning_rate": 3.928386940905483e-05,
+      "loss": 0.0156,
+      "step": 35420
+    },
+    {
+      "action_loss": 0.002343257423490286,
+      "epoch": 31.85251798561151,
+      "step": 35420
+    },
+    {
+      "epoch": 31.85251798561151,
+      "step": 35420,
+      "torque_loss": 0.17719678580760956
+    },
+    {
+      "epoch": 31.85251798561151,
+      "force_loss": 0.003980493173003197,
+      "step": 35420
+    },
+    {
+      "epoch": 31.861510791366907,
+      "grad_norm": 0.09385884553194046,
+      "learning_rate": 3.925695357104073e-05,
+      "loss": 0.0164,
+      "step": 35430
+    },
+    {
+      "action_loss": 0.00855818297713995,
+      "epoch": 31.861510791366907,
+      "step": 35430
+    },
+    {
+      "epoch": 31.861510791366907,
+      "step": 35430,
+      "torque_loss": 0.13382574915885925
+    },
+    {
+      "epoch": 31.861510791366907,
+      "force_loss": 0.01465965062379837,
+      "step": 35430
+    },
+    {
+      "epoch": 31.8705035971223,
+      "grad_norm": 0.2396390438079834,
+      "learning_rate": 3.923004099648049e-05,
+      "loss": 0.0196,
+      "step": 35440
+    },
+    {
+      "action_loss": 0.0022025706712156534,
+      "epoch": 31.8705035971223,
+      "step": 35440
+    },
+    {
+      "epoch": 31.8705035971223,
+      "step": 35440,
+      "torque_loss": 0.12100708484649658
+    },
+    {
+      "epoch": 31.8705035971223,
+      "force_loss": 0.0025210340972989798,
+      "step": 35440
+    },
+    {
+      "epoch": 31.8794964028777,
+      "grad_norm": 0.26506224274635315,
+      "learning_rate": 3.920313169354944e-05,
+      "loss": 0.0168,
+      "step": 35450
+    },
+    {
+      "action_loss": 0.0011666034115478396,
+      "epoch": 31.8794964028777,
+      "step": 35450
+    },
+    {
+      "epoch": 31.8794964028777,
+      "step": 35450,
+      "torque_loss": 0.10555519908666611
+    },
+    {
+      "epoch": 31.8794964028777,
+      "force_loss": 0.002271487610414624,
+      "step": 35450
+    },
+    {
+      "epoch": 31.888489208633093,
+      "grad_norm": 0.19065934419631958,
+      "learning_rate": 3.9176225670421897e-05,
+      "loss": 0.023,
+      "step": 35460
+    },
+    {
+      "action_loss": 0.002745841396972537,
+      "epoch": 31.888489208633093,
+      "step": 35460
+    },
+    {
+      "epoch": 31.888489208633093,
+      "step": 35460,
+      "torque_loss": 0.13876836001873016
+    },
+    {
+      "epoch": 31.888489208633093,
+      "force_loss": 0.005210071336477995,
+      "step": 35460
+    },
+    {
+      "epoch": 31.89748201438849,
+      "grad_norm": 0.17335911095142365,
+      "learning_rate": 3.9149322935271224e-05,
+      "loss": 0.0176,
+      "step": 35470
+    },
+    {
+      "action_loss": 0.009939249604940414,
+      "epoch": 31.89748201438849,
+      "step": 35470
+    },
+    {
+      "epoch": 31.89748201438849,
+      "step": 35470,
+      "torque_loss": 0.1244107261300087
+    },
+    {
+      "epoch": 31.89748201438849,
+      "force_loss": 0.0119258351624012,
+      "step": 35470
+    },
+    {
+      "epoch": 31.906474820143885,
+      "grad_norm": 0.3328714072704315,
+      "learning_rate": 3.9122423496269725e-05,
+      "loss": 0.0212,
+      "step": 35480
+    },
+    {
+      "action_loss": 0.001553779817186296,
+      "epoch": 31.906474820143885,
+      "step": 35480
+    },
+    {
+      "epoch": 31.906474820143885,
+      "step": 35480,
+      "torque_loss": 0.1323273777961731
+    },
+    {
+      "epoch": 31.906474820143885,
+      "force_loss": 0.0013025113148614764,
+      "step": 35480
+    },
+    {
+      "epoch": 31.915467625899282,
+      "grad_norm": 0.21877729892730713,
+      "learning_rate": 3.909552736158877e-05,
+      "loss": 0.0174,
+      "step": 35490
+    },
+    {
+      "action_loss": 0.003774484619498253,
+      "epoch": 31.915467625899282,
+      "step": 35490
+    },
+    {
+      "epoch": 31.915467625899282,
+      "step": 35490,
+      "torque_loss": 0.18078093230724335
+    },
+    {
+      "epoch": 31.915467625899282,
+      "force_loss": 0.0039477222599089146,
+      "step": 35490
+    },
+    {
+      "epoch": 31.924460431654676,
+      "grad_norm": 0.10050065070390701,
+      "learning_rate": 3.90686345393987e-05,
+      "loss": 0.017,
+      "step": 35500
+    },
+    {
+      "action_loss": 0.0015745697310194373,
+      "epoch": 31.924460431654676,
+      "step": 35500
+    },
+    {
+      "epoch": 31.924460431654676,
+      "step": 35500,
+      "torque_loss": 0.13969404995441437
+    },
+    {
+      "epoch": 31.924460431654676,
+      "force_loss": 0.0036092661321163177,
+      "step": 35500
+    },
+    {
+      "epoch": 31.93345323741007,
+      "grad_norm": 0.19695894420146942,
+      "learning_rate": 3.9041745037868816e-05,
+      "loss": 0.016,
+      "step": 35510
+    },
+    {
+      "action_loss": 0.0019424165366217494,
+      "epoch": 31.93345323741007,
+      "step": 35510
+    },
+    {
+      "epoch": 31.93345323741007,
+      "step": 35510,
+      "torque_loss": 0.1555541306734085
+    },
+    {
+      "epoch": 31.93345323741007,
+      "force_loss": 0.0014627467608079314,
+      "step": 35510
+    },
+    {
+      "epoch": 31.942446043165468,
+      "grad_norm": 0.3071577250957489,
+      "learning_rate": 3.9014858865167465e-05,
+      "loss": 0.0203,
+      "step": 35520
+    },
+    {
+      "action_loss": 0.004073005635291338,
+      "epoch": 31.942446043165468,
+      "step": 35520
+    },
+    {
+      "epoch": 31.942446043165468,
+      "step": 35520,
+      "torque_loss": 0.14763735234737396
+    },
+    {
+      "epoch": 31.942446043165468,
+      "force_loss": 0.002649690257385373,
+      "step": 35520
+    },
+    {
+      "epoch": 31.951438848920862,
+      "grad_norm": 0.14206725358963013,
+      "learning_rate": 3.8987976029461935e-05,
+      "loss": 0.0231,
+      "step": 35530
+    },
+    {
+      "action_loss": 0.002284081419929862,
+      "epoch": 31.951438848920862,
+      "step": 35530
+    },
+    {
+      "epoch": 31.951438848920862,
+      "step": 35530,
+      "torque_loss": 0.12091898173093796
+    },
+    {
+      "epoch": 31.951438848920862,
+      "force_loss": 0.005671159829944372,
+      "step": 35530
+    },
+    {
+      "epoch": 31.96043165467626,
+      "grad_norm": 0.18670997023582458,
+      "learning_rate": 3.896109653891853e-05,
+      "loss": 0.0172,
+      "step": 35540
+    },
+    {
+      "action_loss": 0.0024247097317129374,
+      "epoch": 31.96043165467626,
+      "step": 35540
+    },
+    {
+      "epoch": 31.96043165467626,
+      "step": 35540,
+      "torque_loss": 0.10384371876716614
+    },
+    {
+      "epoch": 31.96043165467626,
+      "force_loss": 0.003963085822761059,
+      "step": 35540
+    },
+    {
+      "epoch": 31.969424460431654,
+      "grad_norm": 0.21620497107505798,
+      "learning_rate": 3.893422040170254e-05,
+      "loss": 0.0159,
+      "step": 35550
+    },
+    {
+      "action_loss": 0.006690721958875656,
+      "epoch": 31.969424460431654,
+      "step": 35550
+    },
+    {
+      "epoch": 31.969424460431654,
+      "step": 35550,
+      "torque_loss": 0.10844894498586655
+    },
+    {
+      "epoch": 31.969424460431654,
+      "force_loss": 0.016566084697842598,
+      "step": 35550
+    },
+    {
+      "epoch": 31.97841726618705,
+      "grad_norm": 0.16152913868427277,
+      "learning_rate": 3.8907347625978207e-05,
+      "loss": 0.0213,
+      "step": 35560
+    },
+    {
+      "action_loss": 0.01430537085980177,
+      "epoch": 31.97841726618705,
+      "step": 35560
+    },
+    {
+      "epoch": 31.97841726618705,
+      "step": 35560,
+      "torque_loss": 0.21666081249713898
+    },
+    {
+      "epoch": 31.97841726618705,
+      "force_loss": 0.018692273646593094,
+      "step": 35560
+    },
+    {
+      "epoch": 31.987410071942445,
+      "grad_norm": 0.23225396871566772,
+      "learning_rate": 3.88804782199088e-05,
+      "loss": 0.0181,
+      "step": 35570
+    },
+    {
+      "action_loss": 0.017697514966130257,
+      "epoch": 31.987410071942445,
+      "step": 35570
+    },
+    {
+      "epoch": 31.987410071942445,
+      "step": 35570,
+      "torque_loss": 0.16339315474033356
+    },
+    {
+      "epoch": 31.987410071942445,
+      "force_loss": 0.026077555492520332,
+      "step": 35570
+    },
+    {
+      "epoch": 31.996402877697843,
+      "grad_norm": 0.6275593042373657,
+      "learning_rate": 3.8853612191656495e-05,
+      "loss": 0.0173,
+      "step": 35580
+    },
+    {
+      "action_loss": 0.002076135715469718,
+      "epoch": 31.996402877697843,
+      "step": 35580
+    },
+    {
+      "epoch": 31.996402877697843,
+      "step": 35580,
+      "torque_loss": 0.15213413536548615
+    },
+    {
+      "epoch": 31.996402877697843,
+      "force_loss": 0.0013605583226308227,
+      "step": 35580
+    },
+    {
+      "epoch": 32.00539568345324,
+      "grad_norm": 0.31636619567871094,
+      "learning_rate": 3.88267495493825e-05,
+      "loss": 0.0179,
+      "step": 35590
+    },
+    {
+      "action_loss": 0.0017146820900961757,
+      "epoch": 32.00539568345324,
+      "step": 35590
+    },
+    {
+      "epoch": 32.00539568345324,
+      "step": 35590,
+      "torque_loss": 0.0939028263092041
+    },
+    {
+      "epoch": 32.00539568345324,
+      "force_loss": 0.0031150151044130325,
+      "step": 35590
+    },
+    {
+      "epoch": 32.014388489208635,
+      "grad_norm": 0.24996823072433472,
+      "learning_rate": 3.8799890301247004e-05,
+      "loss": 0.0211,
+      "step": 35600
+    },
+    {
+      "action_loss": 0.0191951896995306,
+      "epoch": 32.014388489208635,
+      "step": 35600
+    },
+    {
+      "epoch": 32.014388489208635,
+      "step": 35600,
+      "torque_loss": 0.14893077313899994
+    },
+    {
+      "epoch": 32.014388489208635,
+      "force_loss": 0.03253011777997017,
+      "step": 35600
+    },
+    {
+      "epoch": 32.023381294964025,
+      "grad_norm": 0.21590636670589447,
+      "learning_rate": 3.8773034455409096e-05,
+      "loss": 0.0195,
+      "step": 35610
+    },
+    {
+      "action_loss": 0.007882803678512573,
+      "epoch": 32.023381294964025,
+      "step": 35610
+    },
+    {
+      "epoch": 32.023381294964025,
+      "step": 35610,
+      "torque_loss": 0.15608543157577515
+    },
+    {
+      "epoch": 32.023381294964025,
+      "force_loss": 0.007862423546612263,
+      "step": 35610
+    },
+    {
+      "epoch": 32.03237410071942,
+      "grad_norm": 0.228277787566185,
+      "learning_rate": 3.8746182020026904e-05,
+      "loss": 0.0193,
+      "step": 35620
+    },
+    {
+      "action_loss": 0.003483865410089493,
+      "epoch": 32.03237410071942,
+      "step": 35620
+    },
+    {
+      "epoch": 32.03237410071942,
+      "step": 35620,
+      "torque_loss": 0.1525992751121521
+    },
+    {
+      "epoch": 32.03237410071942,
+      "force_loss": 0.002559076063334942,
+      "step": 35620
+    },
+    {
+      "epoch": 32.04136690647482,
+      "grad_norm": 0.25834381580352783,
+      "learning_rate": 3.871933300325745e-05,
+      "loss": 0.0167,
+      "step": 35630
+    },
+    {
+      "action_loss": 0.0043231225572526455,
+      "epoch": 32.04136690647482,
+      "step": 35630
+    },
+    {
+      "epoch": 32.04136690647482,
+      "step": 35630,
+      "torque_loss": 0.13323475420475006
+    },
+    {
+      "epoch": 32.04136690647482,
+      "force_loss": 0.009907946921885014,
+      "step": 35630
+    },
+    {
+      "epoch": 32.05035971223022,
+      "grad_norm": 0.3000139594078064,
+      "learning_rate": 3.869248741325679e-05,
+      "loss": 0.02,
+      "step": 35640
+    },
+    {
+      "action_loss": 0.004427507985383272,
+      "epoch": 32.05035971223022,
+      "step": 35640
+    },
+    {
+      "epoch": 32.05035971223022,
+      "step": 35640,
+      "torque_loss": 0.14015614986419678
+    },
+    {
+      "epoch": 32.05035971223022,
+      "force_loss": 0.005604162812232971,
+      "step": 35640
+    },
+    {
+      "epoch": 32.05935251798561,
+      "grad_norm": 0.17594636976718903,
+      "learning_rate": 3.866564525817992e-05,
+      "loss": 0.0167,
+      "step": 35650
+    },
+    {
+      "action_loss": 0.004938953090459108,
+      "epoch": 32.05935251798561,
+      "step": 35650
+    },
+    {
+      "epoch": 32.05935251798561,
+      "step": 35650,
+      "torque_loss": 0.1780819296836853
+    },
+    {
+      "epoch": 32.05935251798561,
+      "force_loss": 0.002592293079942465,
+      "step": 35650
+    },
+    {
+      "epoch": 32.068345323741006,
+      "grad_norm": 0.29027655720710754,
+      "learning_rate": 3.8638806546180725e-05,
+      "loss": 0.0175,
+      "step": 35660
+    },
+    {
+      "action_loss": 0.0020971298217773438,
+      "epoch": 32.068345323741006,
+      "step": 35660
+    },
+    {
+      "epoch": 32.068345323741006,
+      "step": 35660,
+      "torque_loss": 0.08877771347761154
+    },
+    {
+      "epoch": 32.068345323741006,
+      "force_loss": 0.002040656516328454,
+      "step": 35660
+    },
+    {
+      "epoch": 32.077338129496404,
+      "grad_norm": 0.18215881288051605,
+      "learning_rate": 3.861197128541213e-05,
+      "loss": 0.02,
+      "step": 35670
+    },
+    {
+      "action_loss": 0.00523406034335494,
+      "epoch": 32.077338129496404,
+      "step": 35670
+    },
+    {
+      "epoch": 32.077338129496404,
+      "step": 35670,
+      "torque_loss": 0.13208834826946259
+    },
+    {
+      "epoch": 32.077338129496404,
+      "force_loss": 0.0069520375691354275,
+      "step": 35670
+    },
+    {
+      "epoch": 32.0863309352518,
+      "grad_norm": 0.11743181198835373,
+      "learning_rate": 3.858513948402599e-05,
+      "loss": 0.0191,
+      "step": 35680
+    },
+    {
+      "action_loss": 0.005054831970483065,
+      "epoch": 32.0863309352518,
+      "step": 35680
+    },
+    {
+      "epoch": 32.0863309352518,
+      "step": 35680,
+      "torque_loss": 0.14573615789413452
+    },
+    {
+      "epoch": 32.0863309352518,
+      "force_loss": 0.006392117589712143,
+      "step": 35680
+    },
+    {
+      "epoch": 32.09532374100719,
+      "grad_norm": 0.32260438799858093,
+      "learning_rate": 3.8558311150173077e-05,
+      "loss": 0.0328,
+      "step": 35690
+    },
+    {
+      "action_loss": 0.005089659709483385,
+      "epoch": 32.09532374100719,
+      "step": 35690
+    },
+    {
+      "epoch": 32.09532374100719,
+      "step": 35690,
+      "torque_loss": 0.19206349551677704
+    },
+    {
+      "epoch": 32.09532374100719,
+      "force_loss": 0.0048776851035654545,
+      "step": 35690
+    },
+    {
+      "epoch": 32.10431654676259,
+      "grad_norm": 0.19433145225048065,
+      "learning_rate": 3.853148629200312e-05,
+      "loss": 0.0207,
+      "step": 35700
+    },
+    {
+      "action_loss": 0.0017868741415441036,
+      "epoch": 32.10431654676259,
+      "step": 35700
+    },
+    {
+      "epoch": 32.10431654676259,
+      "step": 35700,
+      "torque_loss": 0.1520131379365921
+    },
+    {
+      "epoch": 32.10431654676259,
+      "force_loss": 0.0017587350448593497,
+      "step": 35700
+    },
+    {
+      "epoch": 32.11330935251799,
+      "grad_norm": 0.1347096711397171,
+      "learning_rate": 3.850466491766482e-05,
+      "loss": 0.0185,
+      "step": 35710
+    },
+    {
+      "action_loss": 0.0031681815162301064,
+      "epoch": 32.11330935251799,
+      "step": 35710
+    },
+    {
+      "epoch": 32.11330935251799,
+      "step": 35710,
+      "torque_loss": 0.16818366944789886
+    },
+    {
+      "epoch": 32.11330935251799,
+      "force_loss": 0.009126334451138973,
+      "step": 35710
+    },
+    {
+      "epoch": 32.12230215827338,
+      "grad_norm": 0.22764194011688232,
+      "learning_rate": 3.847784703530583e-05,
+      "loss": 0.0165,
+      "step": 35720
+    },
+    {
+      "action_loss": 0.003924195189028978,
+      "epoch": 32.12230215827338,
+      "step": 35720
+    },
+    {
+      "epoch": 32.12230215827338,
+      "step": 35720,
+      "torque_loss": 0.09507862478494644
+    },
+    {
+      "epoch": 32.12230215827338,
+      "force_loss": 0.007018100470304489,
+      "step": 35720
+    },
+    {
+      "epoch": 32.131294964028775,
+      "grad_norm": 0.11621231585741043,
+      "learning_rate": 3.845103265307266e-05,
+      "loss": 0.0173,
+      "step": 35730
+    },
+    {
+      "action_loss": 0.002146992366760969,
+      "epoch": 32.131294964028775,
+      "step": 35730
+    },
+    {
+      "epoch": 32.131294964028775,
+      "step": 35730,
+      "torque_loss": 0.146606907248497
+    },
+    {
+      "epoch": 32.131294964028775,
+      "force_loss": 0.0018671402940526605,
+      "step": 35730
+    },
+    {
+      "epoch": 32.14028776978417,
+      "grad_norm": 0.10835649073123932,
+      "learning_rate": 3.842422177911086e-05,
+      "loss": 0.0174,
+      "step": 35740
+    },
+    {
+      "action_loss": 0.008047571405768394,
+      "epoch": 32.14028776978417,
+      "step": 35740
+    },
+    {
+      "epoch": 32.14028776978417,
+      "step": 35740,
+      "torque_loss": 0.11531472206115723
+    },
+    {
+      "epoch": 32.14028776978417,
+      "force_loss": 0.014328345656394958,
+      "step": 35740
+    },
+    {
+      "epoch": 32.14928057553957,
+      "grad_norm": 0.1983528435230255,
+      "learning_rate": 3.8397414421564826e-05,
+      "loss": 0.0191,
+      "step": 35750
+    },
+    {
+      "action_loss": 0.004859645385295153,
+      "epoch": 32.14928057553957,
+      "step": 35750
+    },
+    {
+      "epoch": 32.14928057553957,
+      "step": 35750,
+      "torque_loss": 0.16065071523189545
+    },
+    {
+      "epoch": 32.14928057553957,
+      "force_loss": 0.0056686424650251865,
+      "step": 35750
+    },
+    {
+      "epoch": 32.15827338129496,
+      "grad_norm": 0.37957534193992615,
+      "learning_rate": 3.8370610588577935e-05,
+      "loss": 0.0199,
+      "step": 35760
+    },
+    {
+      "action_loss": 0.004888215567916632,
+      "epoch": 32.15827338129496,
+      "step": 35760
+    },
+    {
+      "epoch": 32.15827338129496,
+      "step": 35760,
+      "torque_loss": 0.15222908556461334
+    },
+    {
+      "epoch": 32.15827338129496,
+      "force_loss": 0.01704634726047516,
+      "step": 35760
+    },
+    {
+      "epoch": 32.16726618705036,
+      "grad_norm": 0.17114831507205963,
+      "learning_rate": 3.834381028829251e-05,
+      "loss": 0.0174,
+      "step": 35770
+    },
+    {
+      "action_loss": 0.003907196223735809,
+      "epoch": 32.16726618705036,
+      "step": 35770
+    },
+    {
+      "epoch": 32.16726618705036,
+      "step": 35770,
+      "torque_loss": 0.15016154944896698
+    },
+    {
+      "epoch": 32.16726618705036,
+      "force_loss": 0.002589662792161107,
+      "step": 35770
+    },
+    {
+      "epoch": 32.17625899280576,
+      "grad_norm": 0.19417725503444672,
+      "learning_rate": 3.8317013528849745e-05,
+      "loss": 0.0158,
+      "step": 35780
+    },
+    {
+      "action_loss": 0.0029246974736452103,
+      "epoch": 32.17625899280576,
+      "step": 35780
+    },
+    {
+      "epoch": 32.17625899280576,
+      "step": 35780,
+      "torque_loss": 0.12482305616140366
+    },
+    {
+      "epoch": 32.17625899280576,
+      "force_loss": 0.004284083843231201,
+      "step": 35780
+    },
+    {
+      "epoch": 32.185251798561154,
+      "grad_norm": 0.1793966442346573,
+      "learning_rate": 3.8290220318389815e-05,
+      "loss": 0.0213,
+      "step": 35790
+    },
+    {
+      "action_loss": 0.0022399509325623512,
+      "epoch": 32.185251798561154,
+      "step": 35790
+    },
+    {
+      "epoch": 32.185251798561154,
+      "step": 35790,
+      "torque_loss": 0.09926352649927139
+    },
+    {
+      "epoch": 32.185251798561154,
+      "force_loss": 0.0037671260070055723,
+      "step": 35790
+    },
+    {
+      "epoch": 32.194244604316545,
+      "grad_norm": 0.1799623966217041,
+      "learning_rate": 3.8263430665051746e-05,
+      "loss": 0.0155,
+      "step": 35800
+    },
+    {
+      "action_loss": 0.0030329227447509766,
+      "epoch": 32.194244604316545,
+      "step": 35800
+    },
+    {
+      "epoch": 32.194244604316545,
+      "step": 35800,
+      "torque_loss": 0.14444126188755035
+    },
+    {
+      "epoch": 32.194244604316545,
+      "force_loss": 0.005087635014206171,
+      "step": 35800
+    },
+    {
+      "epoch": 32.20323741007194,
+      "grad_norm": 0.24660754203796387,
+      "learning_rate": 3.8236644576973554e-05,
+      "loss": 0.0165,
+      "step": 35810
+    },
+    {
+      "action_loss": 0.004632997792214155,
+      "epoch": 32.20323741007194,
+      "step": 35810
+    },
+    {
+      "epoch": 32.20323741007194,
+      "step": 35810,
+      "torque_loss": 0.1333283632993698
+    },
+    {
+      "epoch": 32.20323741007194,
+      "force_loss": 0.006720694247633219,
+      "step": 35810
+    },
+    {
+      "epoch": 32.21223021582734,
+      "grad_norm": 0.07578231394290924,
+      "learning_rate": 3.820986206229217e-05,
+      "loss": 0.0161,
+      "step": 35820
+    },
+    {
+      "action_loss": 0.00815785676240921,
+      "epoch": 32.21223021582734,
+      "step": 35820
+    },
+    {
+      "epoch": 32.21223021582734,
+      "step": 35820,
+      "torque_loss": 0.15545953810214996
+    },
+    {
+      "epoch": 32.21223021582734,
+      "force_loss": 0.0039089578203856945,
+      "step": 35820
+    },
+    {
+      "epoch": 32.22122302158273,
+      "grad_norm": 0.49912646412849426,
+      "learning_rate": 3.8183083129143384e-05,
+      "loss": 0.0172,
+      "step": 35830
+    },
+    {
+      "action_loss": 0.06644818186759949,
+      "epoch": 32.22122302158273,
+      "step": 35830
+    },
+    {
+      "epoch": 32.22122302158273,
+      "step": 35830,
+      "torque_loss": 0.19193215668201447
+    },
+    {
+      "epoch": 32.22122302158273,
+      "force_loss": 0.06332853436470032,
+      "step": 35830
+    },
+    {
+      "epoch": 32.23021582733813,
+      "grad_norm": 0.22522485256195068,
+      "learning_rate": 3.815630778566193e-05,
+      "loss": 0.0273,
+      "step": 35840
+    },
+    {
+      "action_loss": 0.006677614990621805,
+      "epoch": 32.23021582733813,
+      "step": 35840
+    },
+    {
+      "epoch": 32.23021582733813,
+      "step": 35840,
+      "torque_loss": 0.13887529075145721
+    },
+    {
+      "epoch": 32.23021582733813,
+      "force_loss": 0.004647038411349058,
+      "step": 35840
+    },
+    {
+      "epoch": 32.239208633093526,
+      "grad_norm": 0.15656091272830963,
+      "learning_rate": 3.812953603998145e-05,
+      "loss": 0.0175,
+      "step": 35850
+    },
+    {
+      "action_loss": 0.0009108425583690405,
+      "epoch": 32.239208633093526,
+      "step": 35850
+    },
+    {
+      "epoch": 32.239208633093526,
+      "step": 35850,
+      "torque_loss": 0.07493776828050613
+    },
+    {
+      "epoch": 32.239208633093526,
+      "force_loss": 0.0010153655894100666,
+      "step": 35850
+    },
+    {
+      "epoch": 32.24820143884892,
+      "grad_norm": 0.28867635130882263,
+      "learning_rate": 3.8102767900234504e-05,
+      "loss": 0.0169,
+      "step": 35860
+    },
+    {
+      "action_loss": 0.001914641004987061,
+      "epoch": 32.24820143884892,
+      "step": 35860
+    },
+    {
+      "epoch": 32.24820143884892,
+      "step": 35860,
+      "torque_loss": 0.10751917958259583
+    },
+    {
+      "epoch": 32.24820143884892,
+      "force_loss": 0.0030521564185619354,
+      "step": 35860
+    },
+    {
+      "epoch": 32.257194244604314,
+      "grad_norm": 0.23864106833934784,
+      "learning_rate": 3.807600337455256e-05,
+      "loss": 0.0193,
+      "step": 35870
+    },
+    {
+      "action_loss": 0.0036270718555897474,
+      "epoch": 32.257194244604314,
+      "step": 35870
+    },
+    {
+      "epoch": 32.257194244604314,
+      "step": 35870,
+      "torque_loss": 0.09673986583948135
+    },
+    {
+      "epoch": 32.257194244604314,
+      "force_loss": 0.003295250004157424,
+      "step": 35870
+    },
+    {
+      "epoch": 32.26618705035971,
+      "grad_norm": 0.2653827965259552,
+      "learning_rate": 3.804924247106593e-05,
+      "loss": 0.019,
+      "step": 35880
+    },
+    {
+      "action_loss": 0.0020589062478393316,
+      "epoch": 32.26618705035971,
+      "step": 35880
+    },
+    {
+      "epoch": 32.26618705035971,
+      "step": 35880,
+      "torque_loss": 0.06579836457967758
+    },
+    {
+      "epoch": 32.26618705035971,
+      "force_loss": 0.0023911793250590563,
+      "step": 35880
+    },
+    {
+      "epoch": 32.27517985611511,
+      "grad_norm": 0.18982920050621033,
+      "learning_rate": 3.8022485197903925e-05,
+      "loss": 0.0175,
+      "step": 35890
+    },
+    {
+      "action_loss": 0.0015938496217131615,
+      "epoch": 32.27517985611511,
+      "step": 35890
+    },
+    {
+      "epoch": 32.27517985611511,
+      "step": 35890,
+      "torque_loss": 0.10908424109220505
+    },
+    {
+      "epoch": 32.27517985611511,
+      "force_loss": 0.00077058095484972,
+      "step": 35890
+    },
+    {
+      "epoch": 32.28417266187051,
+      "grad_norm": 0.13052676618099213,
+      "learning_rate": 3.799573156319464e-05,
+      "loss": 0.0172,
+      "step": 35900
+    },
+    {
+      "action_loss": 0.014239001087844372,
+      "epoch": 32.28417266187051,
+      "step": 35900
+    },
+    {
+      "epoch": 32.28417266187051,
+      "step": 35900,
+      "torque_loss": 0.10351020842790604
+    },
+    {
+      "epoch": 32.28417266187051,
+      "force_loss": 0.010057463310658932,
+      "step": 35900
+    },
+    {
+      "epoch": 32.2931654676259,
+      "grad_norm": 0.2521089017391205,
+      "learning_rate": 3.796898157506515e-05,
+      "loss": 0.0211,
+      "step": 35910
+    },
+    {
+      "action_loss": 0.002401305129751563,
+      "epoch": 32.2931654676259,
+      "step": 35910
+    },
+    {
+      "epoch": 32.2931654676259,
+      "step": 35910,
+      "torque_loss": 0.133083775639534
+    },
+    {
+      "epoch": 32.2931654676259,
+      "force_loss": 0.00283030909486115,
+      "step": 35910
+    },
+    {
+      "epoch": 32.302158273381295,
+      "grad_norm": 0.13918083906173706,
+      "learning_rate": 3.794223524164143e-05,
+      "loss": 0.0178,
+      "step": 35920
+    },
+    {
+      "action_loss": 0.0015532016986981034,
+      "epoch": 32.302158273381295,
+      "step": 35920
+    },
+    {
+      "epoch": 32.302158273381295,
+      "step": 35920,
+      "torque_loss": 0.08964937925338745
+    },
+    {
+      "epoch": 32.302158273381295,
+      "force_loss": 0.0019377600401639938,
+      "step": 35920
+    },
+    {
+      "epoch": 32.31115107913669,
+      "grad_norm": 0.1956116110086441,
+      "learning_rate": 3.7915492571048245e-05,
+      "loss": 0.0205,
+      "step": 35930
+    },
+    {
+      "action_loss": 0.002844024682417512,
+      "epoch": 32.31115107913669,
+      "step": 35930
+    },
+    {
+      "epoch": 32.31115107913669,
+      "step": 35930,
+      "torque_loss": 0.10727709531784058
+    },
+    {
+      "epoch": 32.31115107913669,
+      "force_loss": 0.004802302923053503,
+      "step": 35930
+    },
+    {
+      "epoch": 32.32014388489208,
+      "grad_norm": 0.2805778980255127,
+      "learning_rate": 3.788875357140937e-05,
+      "loss": 0.0238,
+      "step": 35940
+    },
+    {
+      "action_loss": 0.004543399903923273,
+      "epoch": 32.32014388489208,
+      "step": 35940
+    },
+    {
+      "epoch": 32.32014388489208,
+      "step": 35940,
+      "torque_loss": 0.12029657512903214
+    },
+    {
+      "epoch": 32.32014388489208,
+      "force_loss": 0.00520290806889534,
+      "step": 35940
+    },
+    {
+      "epoch": 32.32913669064748,
+      "grad_norm": 0.24458202719688416,
+      "learning_rate": 3.786201825084736e-05,
+      "loss": 0.0163,
+      "step": 35950
+    },
+    {
+      "action_loss": 0.002751231426373124,
+      "epoch": 32.32913669064748,
+      "step": 35950
+    },
+    {
+      "epoch": 32.32913669064748,
+      "step": 35950,
+      "torque_loss": 0.11642998456954956
+    },
+    {
+      "epoch": 32.32913669064748,
+      "force_loss": 0.0029301319736987352,
+      "step": 35950
+    },
+    {
+      "epoch": 32.33812949640288,
+      "grad_norm": 0.2637021243572235,
+      "learning_rate": 3.783528661748372e-05,
+      "loss": 0.016,
+      "step": 35960
+    },
+    {
+      "action_loss": 0.0029750268440693617,
+      "epoch": 32.33812949640288,
+      "step": 35960
+    },
+    {
+      "epoch": 32.33812949640288,
+      "step": 35960,
+      "torque_loss": 0.11280087381601334
+    },
+    {
+      "epoch": 32.33812949640288,
+      "force_loss": 0.0060738143511116505,
+      "step": 35960
+    },
+    {
+      "epoch": 32.347122302158276,
+      "grad_norm": 0.3076011538505554,
+      "learning_rate": 3.780855867943882e-05,
+      "loss": 0.022,
+      "step": 35970
+    },
+    {
+      "action_loss": 0.001550456159748137,
+      "epoch": 32.347122302158276,
+      "step": 35970
+    },
+    {
+      "epoch": 32.347122302158276,
+      "step": 35970,
+      "torque_loss": 0.10353758931159973
+    },
+    {
+      "epoch": 32.347122302158276,
+      "force_loss": 0.0014127379981800914,
+      "step": 35970
+    },
+    {
+      "epoch": 32.356115107913666,
+      "grad_norm": 0.2607734501361847,
+      "learning_rate": 3.778183444483189e-05,
+      "loss": 0.0177,
+      "step": 35980
+    },
+    {
+      "action_loss": 0.005373851861804724,
+      "epoch": 32.356115107913666,
+      "step": 35980
+    },
+    {
+      "epoch": 32.356115107913666,
+      "step": 35980,
+      "torque_loss": 0.10691497474908829
+    },
+    {
+      "epoch": 32.356115107913666,
+      "force_loss": 0.004640842322260141,
+      "step": 35980
+    },
+    {
+      "epoch": 32.365107913669064,
+      "grad_norm": 0.2998850345611572,
+      "learning_rate": 3.775511392178108e-05,
+      "loss": 0.0181,
+      "step": 35990
+    },
+    {
+      "action_loss": 0.0013369083171710372,
+      "epoch": 32.365107913669064,
+      "step": 35990
+    },
+    {
+      "epoch": 32.365107913669064,
+      "step": 35990,
+      "torque_loss": 0.12341249734163284
+    },
+    {
+      "epoch": 32.365107913669064,
+      "force_loss": 0.0023934622295200825,
+      "step": 35990
+    },
+    {
+      "epoch": 32.37410071942446,
+      "grad_norm": 0.32738152146339417,
+      "learning_rate": 3.772839711840332e-05,
+      "loss": 0.0202,
+      "step": 36000
+    },
+    {
+      "action_loss": 0.007295499090105295,
+      "epoch": 32.37410071942446,
+      "step": 36000
+    },
+    {
+      "epoch": 32.37410071942446,
+      "step": 36000,
+      "torque_loss": 0.15303842723369598
+    },
+    {
+      "epoch": 32.37410071942446,
+      "force_loss": 0.005727540235966444,
+      "step": 36000
+    },
+    {
+      "epoch": 32.38309352517986,
+      "grad_norm": 0.39588892459869385,
+      "learning_rate": 3.7701684042814515e-05,
+      "loss": 0.0211,
+      "step": 36010
+    },
+    {
+      "action_loss": 0.0037621904630213976,
+      "epoch": 32.38309352517986,
+      "step": 36010
+    },
+    {
+      "epoch": 32.38309352517986,
+      "step": 36010,
+      "torque_loss": 0.12148018926382065
+    },
+    {
+      "epoch": 32.38309352517986,
+      "force_loss": 0.004511638078838587,
+      "step": 36010
+    },
+    {
+      "epoch": 32.39208633093525,
+      "grad_norm": 0.29814431071281433,
+      "learning_rate": 3.76749747031294e-05,
+      "loss": 0.0164,
+      "step": 36020
+    },
+    {
+      "action_loss": 0.024477383121848106,
+      "epoch": 32.39208633093525,
+      "step": 36020
+    },
+    {
+      "epoch": 32.39208633093525,
+      "step": 36020,
+      "torque_loss": 0.19140727818012238
+    },
+    {
+      "epoch": 32.39208633093525,
+      "force_loss": 0.040710482746362686,
+      "step": 36020
+    },
+    {
+      "epoch": 32.40107913669065,
+      "grad_norm": 0.25167617201805115,
+      "learning_rate": 3.764826910746152e-05,
+      "loss": 0.0245,
+      "step": 36030
+    },
+    {
+      "action_loss": 0.002894095377996564,
+      "epoch": 32.40107913669065,
+      "step": 36030
+    },
+    {
+      "epoch": 32.40107913669065,
+      "step": 36030,
+      "torque_loss": 0.168232262134552
+    },
+    {
+      "epoch": 32.40107913669065,
+      "force_loss": 0.0017592711374163628,
+      "step": 36030
+    },
+    {
+      "epoch": 32.410071942446045,
+      "grad_norm": 0.17676064372062683,
+      "learning_rate": 3.762156726392338e-05,
+      "loss": 0.0163,
+      "step": 36040
+    },
+    {
+      "action_loss": 0.0027098748832941055,
+      "epoch": 32.410071942446045,
+      "step": 36040
+    },
+    {
+      "epoch": 32.410071942446045,
+      "step": 36040,
+      "torque_loss": 0.09334651380777359
+    },
+    {
+      "epoch": 32.410071942446045,
+      "force_loss": 0.0035172405187040567,
+      "step": 36040
+    },
+    {
+      "epoch": 32.419064748201436,
+      "grad_norm": 0.19449324905872345,
+      "learning_rate": 3.759486918062625e-05,
+      "loss": 0.0222,
+      "step": 36050
+    },
+    {
+      "action_loss": 0.002983634127303958,
+      "epoch": 32.419064748201436,
+      "step": 36050
+    },
+    {
+      "epoch": 32.419064748201436,
+      "step": 36050,
+      "torque_loss": 0.16478627920150757
+    },
+    {
+      "epoch": 32.419064748201436,
+      "force_loss": 0.005507923662662506,
+      "step": 36050
+    },
+    {
+      "epoch": 32.42805755395683,
+      "grad_norm": 0.2805757522583008,
+      "learning_rate": 3.756817486568033e-05,
+      "loss": 0.0197,
+      "step": 36060
+    },
+    {
+      "action_loss": 0.007810443639755249,
+      "epoch": 32.42805755395683,
+      "step": 36060
+    },
+    {
+      "epoch": 32.42805755395683,
+      "step": 36060,
+      "torque_loss": 0.11835313588380814
+    },
+    {
+      "epoch": 32.42805755395683,
+      "force_loss": 0.012032014317810535,
+      "step": 36060
+    },
+    {
+      "epoch": 32.43705035971223,
+      "grad_norm": 0.29620882868766785,
+      "learning_rate": 3.7541484327194654e-05,
+      "loss": 0.0176,
+      "step": 36070
+    },
+    {
+      "action_loss": 0.002350898226723075,
+      "epoch": 32.43705035971223,
+      "step": 36070
+    },
+    {
+      "epoch": 32.43705035971223,
+      "step": 36070,
+      "torque_loss": 0.07381812483072281
+    },
+    {
+      "epoch": 32.43705035971223,
+      "force_loss": 0.0035089829470962286,
+      "step": 36070
+    },
+    {
+      "epoch": 32.44604316546763,
+      "grad_norm": 0.1146189272403717,
+      "learning_rate": 3.751479757327707e-05,
+      "loss": 0.0202,
+      "step": 36080
+    },
+    {
+      "action_loss": 0.003432293189689517,
+      "epoch": 32.44604316546763,
+      "step": 36080
+    },
+    {
+      "epoch": 32.44604316546763,
+      "step": 36080,
+      "torque_loss": 0.10028309375047684
+    },
+    {
+      "epoch": 32.44604316546763,
+      "force_loss": 0.002584822243079543,
+      "step": 36080
+    },
+    {
+      "epoch": 32.45503597122302,
+      "grad_norm": 0.28914424777030945,
+      "learning_rate": 3.7488114612034345e-05,
+      "loss": 0.0184,
+      "step": 36090
+    },
+    {
+      "action_loss": 0.000974094495177269,
+      "epoch": 32.45503597122302,
+      "step": 36090
+    },
+    {
+      "epoch": 32.45503597122302,
+      "step": 36090,
+      "torque_loss": 0.16874247789382935
+    },
+    {
+      "epoch": 32.45503597122302,
+      "force_loss": 0.00147256872151047,
+      "step": 36090
+    },
+    {
+      "epoch": 32.46402877697842,
+      "grad_norm": 0.25618690252304077,
+      "learning_rate": 3.7461435451572044e-05,
+      "loss": 0.0201,
+      "step": 36100
+    },
+    {
+      "action_loss": 0.010891355574131012,
+      "epoch": 32.46402877697842,
+      "step": 36100
+    },
+    {
+      "epoch": 32.46402877697842,
+      "step": 36100,
+      "torque_loss": 0.12587206065654755
+    },
+    {
+      "epoch": 32.46402877697842,
+      "force_loss": 0.008505494333803654,
+      "step": 36100
+    },
+    {
+      "epoch": 32.473021582733814,
+      "grad_norm": 0.09253151714801788,
+      "learning_rate": 3.743476009999459e-05,
+      "loss": 0.0215,
+      "step": 36110
+    },
+    {
+      "action_loss": 0.002712973626330495,
+      "epoch": 32.473021582733814,
+      "step": 36110
+    },
+    {
+      "epoch": 32.473021582733814,
+      "step": 36110,
+      "torque_loss": 0.16960680484771729
+    },
+    {
+      "epoch": 32.473021582733814,
+      "force_loss": 0.0028641382232308388,
+      "step": 36110
+    },
+    {
+      "epoch": 32.48201438848921,
+      "grad_norm": 0.17254550755023956,
+      "learning_rate": 3.7408088565405245e-05,
+      "loss": 0.02,
+      "step": 36120
+    },
+    {
+      "action_loss": 0.0022241489496082067,
+      "epoch": 32.48201438848921,
+      "step": 36120
+    },
+    {
+      "epoch": 32.48201438848921,
+      "step": 36120,
+      "torque_loss": 0.15643279254436493
+    },
+    {
+      "epoch": 32.48201438848921,
+      "force_loss": 0.0014043691335245967,
+      "step": 36120
+    },
+    {
+      "epoch": 32.4910071942446,
+      "grad_norm": 0.1474131941795349,
+      "learning_rate": 3.738142085590612e-05,
+      "loss": 0.0184,
+      "step": 36130
+    },
+    {
+      "action_loss": 0.005370520055294037,
+      "epoch": 32.4910071942446,
+      "step": 36130
+    },
+    {
+      "epoch": 32.4910071942446,
+      "step": 36130,
+      "torque_loss": 0.12520332634449005
+    },
+    {
+      "epoch": 32.4910071942446,
+      "force_loss": 0.004635232966393232,
+      "step": 36130
+    },
+    {
+      "epoch": 32.5,
+      "grad_norm": 0.2439391314983368,
+      "learning_rate": 3.7354756979598194e-05,
+      "loss": 0.0172,
+      "step": 36140
+    },
+    {
+      "action_loss": 0.004163598176091909,
+      "epoch": 32.5,
+      "step": 36140
+    },
+    {
+      "epoch": 32.5,
+      "step": 36140,
+      "torque_loss": 0.13880373537540436
+    },
+    {
+      "epoch": 32.5,
+      "force_loss": 0.011748421005904675,
+      "step": 36140
+    },
+    {
+      "epoch": 32.5089928057554,
+      "grad_norm": 0.2441454827785492,
+      "learning_rate": 3.7328096944581187e-05,
+      "loss": 0.0162,
+      "step": 36150
+    },
+    {
+      "action_loss": 0.0010668698232620955,
+      "epoch": 32.5089928057554,
+      "step": 36150
+    },
+    {
+      "epoch": 32.5089928057554,
+      "step": 36150,
+      "torque_loss": 0.10977080464363098
+    },
+    {
+      "epoch": 32.5089928057554,
+      "force_loss": 0.0009263657848350704,
+      "step": 36150
+    },
+    {
+      "epoch": 32.51798561151079,
+      "grad_norm": 0.2893800437450409,
+      "learning_rate": 3.730144075895377e-05,
+      "loss": 0.0192,
+      "step": 36160
+    },
+    {
+      "action_loss": 0.014249789528548717,
+      "epoch": 32.51798561151079,
+      "step": 36160
+    },
+    {
+      "epoch": 32.51798561151079,
+      "step": 36160,
+      "torque_loss": 0.11918381601572037
+    },
+    {
+      "epoch": 32.51798561151079,
+      "force_loss": 0.02451418526470661,
+      "step": 36160
+    },
+    {
+      "epoch": 32.526978417266186,
+      "grad_norm": 0.1465277224779129,
+      "learning_rate": 3.727478843081335e-05,
+      "loss": 0.0165,
+      "step": 36170
+    },
+    {
+      "action_loss": 0.002034307224676013,
+      "epoch": 32.526978417266186,
+      "step": 36170
+    },
+    {
+      "epoch": 32.526978417266186,
+      "step": 36170,
+      "torque_loss": 0.11345982551574707
+    },
+    {
+      "epoch": 32.526978417266186,
+      "force_loss": 0.00411411514505744,
+      "step": 36170
+    },
+    {
+      "epoch": 32.53597122302158,
+      "grad_norm": 0.7144786715507507,
+      "learning_rate": 3.72481399682562e-05,
+      "loss": 0.0174,
+      "step": 36180
+    },
+    {
+      "action_loss": 0.03424415737390518,
+      "epoch": 32.53597122302158,
+      "step": 36180
+    },
+    {
+      "epoch": 32.53597122302158,
+      "step": 36180,
+      "torque_loss": 0.14526519179344177
+    },
+    {
+      "epoch": 32.53597122302158,
+      "force_loss": 0.0341915637254715,
+      "step": 36180
+    },
+    {
+      "epoch": 32.54496402877698,
+      "grad_norm": 0.24556899070739746,
+      "learning_rate": 3.722149537937747e-05,
+      "loss": 0.0195,
+      "step": 36190
+    },
+    {
+      "action_loss": 0.0016800950979813933,
+      "epoch": 32.54496402877698,
+      "step": 36190
+    },
+    {
+      "epoch": 32.54496402877698,
+      "step": 36190,
+      "torque_loss": 0.07097528129816055
+    },
+    {
+      "epoch": 32.54496402877698,
+      "force_loss": 0.0025095471646636724,
+      "step": 36190
+    },
+    {
+      "epoch": 32.55395683453237,
+      "grad_norm": 0.12260957062244415,
+      "learning_rate": 3.7194854672271015e-05,
+      "loss": 0.0142,
+      "step": 36200
+    },
+    {
+      "action_loss": 0.0065669403411448,
+      "epoch": 32.55395683453237,
+      "step": 36200
+    },
+    {
+      "epoch": 32.55395683453237,
+      "step": 36200,
+      "torque_loss": 0.16885767877101898
+    },
+    {
+      "epoch": 32.55395683453237,
+      "force_loss": 0.010066266171634197,
+      "step": 36200
+    },
+    {
+      "epoch": 32.56294964028777,
+      "grad_norm": 0.23737619817256927,
+      "learning_rate": 3.7168217855029644e-05,
+      "loss": 0.0179,
+      "step": 36210
+    },
+    {
+      "action_loss": 0.0021387652959674597,
+      "epoch": 32.56294964028777,
+      "step": 36210
+    },
+    {
+      "epoch": 32.56294964028777,
+      "step": 36210,
+      "torque_loss": 0.13318897783756256
+    },
+    {
+      "epoch": 32.56294964028777,
+      "force_loss": 0.0055229864083230495,
+      "step": 36210
+    },
+    {
+      "epoch": 32.57194244604317,
+      "grad_norm": 0.16022609174251556,
+      "learning_rate": 3.7141584935744856e-05,
+      "loss": 0.0176,
+      "step": 36220
+    },
+    {
+      "action_loss": 0.0031720492988824844,
+      "epoch": 32.57194244604317,
+      "step": 36220
+    },
+    {
+      "epoch": 32.57194244604317,
+      "step": 36220,
+      "torque_loss": 0.1342240422964096
+    },
+    {
+      "epoch": 32.57194244604317,
+      "force_loss": 0.0026714131236076355,
+      "step": 36220
+    },
+    {
+      "epoch": 32.580935251798564,
+      "grad_norm": 0.16397269070148468,
+      "learning_rate": 3.7114955922507055e-05,
+      "loss": 0.0162,
+      "step": 36230
+    },
+    {
+      "action_loss": 0.0010003969073295593,
+      "epoch": 32.580935251798564,
+      "step": 36230
+    },
+    {
+      "epoch": 32.580935251798564,
+      "step": 36230,
+      "torque_loss": 0.08848855644464493
+    },
+    {
+      "epoch": 32.580935251798564,
+      "force_loss": 0.0011519277468323708,
+      "step": 36230
+    },
+    {
+      "epoch": 32.589928057553955,
+      "grad_norm": 0.17343898117542267,
+      "learning_rate": 3.708833082340545e-05,
+      "loss": 0.0192,
+      "step": 36240
+    },
+    {
+      "action_loss": 0.001305138343013823,
+      "epoch": 32.589928057553955,
+      "step": 36240
+    },
+    {
+      "epoch": 32.589928057553955,
+      "step": 36240,
+      "torque_loss": 0.06593629717826843
+    },
+    {
+      "epoch": 32.589928057553955,
+      "force_loss": 0.0015580812469124794,
+      "step": 36240
+    },
+    {
+      "epoch": 32.59892086330935,
+      "grad_norm": 0.1266532987356186,
+      "learning_rate": 3.7061709646528034e-05,
+      "loss": 0.0141,
+      "step": 36250
+    },
+    {
+      "action_loss": 0.007717927452176809,
+      "epoch": 32.59892086330935,
+      "step": 36250
+    },
+    {
+      "epoch": 32.59892086330935,
+      "step": 36250,
+      "torque_loss": 0.12802881002426147
+    },
+    {
+      "epoch": 32.59892086330935,
+      "force_loss": 0.01575583405792713,
+      "step": 36250
+    },
+    {
+      "epoch": 32.60791366906475,
+      "grad_norm": 0.1754145324230194,
+      "learning_rate": 3.7035092399961604e-05,
+      "loss": 0.0166,
+      "step": 36260
+    },
+    {
+      "action_loss": 0.001969613367691636,
+      "epoch": 32.60791366906475,
+      "step": 36260
+    },
+    {
+      "epoch": 32.60791366906475,
+      "step": 36260,
+      "torque_loss": 0.1301606446504593
+    },
+    {
+      "epoch": 32.60791366906475,
+      "force_loss": 0.004181955009698868,
+      "step": 36260
+    },
+    {
+      "epoch": 32.61690647482014,
+      "grad_norm": 0.7388187646865845,
+      "learning_rate": 3.700847909179177e-05,
+      "loss": 0.0194,
+      "step": 36270
+    },
+    {
+      "action_loss": 0.0048209298402071,
+      "epoch": 32.61690647482014,
+      "step": 36270
+    },
+    {
+      "epoch": 32.61690647482014,
+      "step": 36270,
+      "torque_loss": 0.15288764238357544
+    },
+    {
+      "epoch": 32.61690647482014,
+      "force_loss": 0.003654737723991275,
+      "step": 36270
+    },
+    {
+      "epoch": 32.62589928057554,
+      "grad_norm": 0.3836144208908081,
+      "learning_rate": 3.698186973010297e-05,
+      "loss": 0.0232,
+      "step": 36280
+    },
+    {
+      "action_loss": 0.005664277821779251,
+      "epoch": 32.62589928057554,
+      "step": 36280
+    },
+    {
+      "epoch": 32.62589928057554,
+      "step": 36280,
+      "torque_loss": 0.13004820048809052
+    },
+    {
+      "epoch": 32.62589928057554,
+      "force_loss": 0.013694548048079014,
+      "step": 36280
+    },
+    {
+      "epoch": 32.634892086330936,
+      "grad_norm": 0.3084618151187897,
+      "learning_rate": 3.695526432297844e-05,
+      "loss": 0.0161,
+      "step": 36290
+    },
+    {
+      "action_loss": 0.006037670653313398,
+      "epoch": 32.634892086330936,
+      "step": 36290
+    },
+    {
+      "epoch": 32.634892086330936,
+      "step": 36290,
+      "torque_loss": 0.1703004091978073
+    },
+    {
+      "epoch": 32.634892086330936,
+      "force_loss": 0.007248884532600641,
+      "step": 36290
+    },
+    {
+      "epoch": 32.643884892086334,
+      "grad_norm": 0.5387348532676697,
+      "learning_rate": 3.692866287850017e-05,
+      "loss": 0.0163,
+      "step": 36300
+    },
+    {
+      "action_loss": 0.0015967251965776086,
+      "epoch": 32.643884892086334,
+      "step": 36300
+    },
+    {
+      "epoch": 32.643884892086334,
+      "step": 36300,
+      "torque_loss": 0.12389391660690308
+    },
+    {
+      "epoch": 32.643884892086334,
+      "force_loss": 0.005228487309068441,
+      "step": 36300
+    },
+    {
+      "epoch": 32.652877697841724,
+      "grad_norm": 0.2087627500295639,
+      "learning_rate": 3.6902065404749006e-05,
+      "loss": 0.0171,
+      "step": 36310
+    },
+    {
+      "action_loss": 0.011323176324367523,
+      "epoch": 32.652877697841724,
+      "step": 36310
+    },
+    {
+      "epoch": 32.652877697841724,
+      "step": 36310,
+      "torque_loss": 0.14483921229839325
+    },
+    {
+      "epoch": 32.652877697841724,
+      "force_loss": 0.005349026992917061,
+      "step": 36310
+    },
+    {
+      "epoch": 32.66187050359712,
+      "grad_norm": 0.41825759410858154,
+      "learning_rate": 3.6875471909804516e-05,
+      "loss": 0.0228,
+      "step": 36320
+    },
+    {
+      "action_loss": 0.000972981855738908,
+      "epoch": 32.66187050359712,
+      "step": 36320
+    },
+    {
+      "epoch": 32.66187050359712,
+      "step": 36320,
+      "torque_loss": 0.09398885816335678
+    },
+    {
+      "epoch": 32.66187050359712,
+      "force_loss": 0.000758222013246268,
+      "step": 36320
+    },
+    {
+      "epoch": 32.67086330935252,
+      "grad_norm": 0.23640744388103485,
+      "learning_rate": 3.6848882401745135e-05,
+      "loss": 0.0139,
+      "step": 36330
+    },
+    {
+      "action_loss": 0.002796156331896782,
+      "epoch": 32.67086330935252,
+      "step": 36330
+    },
+    {
+      "epoch": 32.67086330935252,
+      "step": 36330,
+      "torque_loss": 0.127136692404747
+    },
+    {
+      "epoch": 32.67086330935252,
+      "force_loss": 0.004370329901576042,
+      "step": 36330
+    },
+    {
+      "epoch": 32.67985611510792,
+      "grad_norm": 0.279753714799881,
+      "learning_rate": 3.682229688864806e-05,
+      "loss": 0.0169,
+      "step": 36340
+    },
+    {
+      "action_loss": 0.0012949536321684718,
+      "epoch": 32.67985611510792,
+      "step": 36340
+    },
+    {
+      "epoch": 32.67985611510792,
+      "step": 36340,
+      "torque_loss": 0.1081714853644371
+    },
+    {
+      "epoch": 32.67985611510792,
+      "force_loss": 0.0017533120699226856,
+      "step": 36340
+    },
+    {
+      "epoch": 32.68884892086331,
+      "grad_norm": 0.10611731559038162,
+      "learning_rate": 3.6795715378589235e-05,
+      "loss": 0.0159,
+      "step": 36350
+    },
+    {
+      "action_loss": 0.001235096831806004,
+      "epoch": 32.68884892086331,
+      "step": 36350
+    },
+    {
+      "epoch": 32.68884892086331,
+      "step": 36350,
+      "torque_loss": 0.11640948057174683
+    },
+    {
+      "epoch": 32.68884892086331,
+      "force_loss": 0.001461779116652906,
+      "step": 36350
+    },
+    {
+      "epoch": 32.697841726618705,
+      "grad_norm": 0.34805500507354736,
+      "learning_rate": 3.676913787964345e-05,
+      "loss": 0.0217,
+      "step": 36360
+    },
+    {
+      "action_loss": 0.002664159284904599,
+      "epoch": 32.697841726618705,
+      "step": 36360
+    },
+    {
+      "epoch": 32.697841726618705,
+      "step": 36360,
+      "torque_loss": 0.10745004564523697
+    },
+    {
+      "epoch": 32.697841726618705,
+      "force_loss": 0.0024546687491238117,
+      "step": 36360
+    },
+    {
+      "epoch": 32.7068345323741,
+      "grad_norm": 0.2172185331583023,
+      "learning_rate": 3.674256439988423e-05,
+      "loss": 0.0198,
+      "step": 36370
+    },
+    {
+      "action_loss": 0.002869424642995,
+      "epoch": 32.7068345323741,
+      "step": 36370
+    },
+    {
+      "epoch": 32.7068345323741,
+      "step": 36370,
+      "torque_loss": 0.09453406184911728
+    },
+    {
+      "epoch": 32.7068345323741,
+      "force_loss": 0.00511117884889245,
+      "step": 36370
+    },
+    {
+      "epoch": 32.71582733812949,
+      "grad_norm": 0.4886157512664795,
+      "learning_rate": 3.6715994947383904e-05,
+      "loss": 0.0168,
+      "step": 36380
+    },
+    {
+      "action_loss": 0.0028554575983434916,
+      "epoch": 32.71582733812949,
+      "step": 36380
+    },
+    {
+      "epoch": 32.71582733812949,
+      "step": 36380,
+      "torque_loss": 0.08057756721973419
+    },
+    {
+      "epoch": 32.71582733812949,
+      "force_loss": 0.0010780870215967298,
+      "step": 36380
+    },
+    {
+      "epoch": 32.72482014388489,
+      "grad_norm": 0.1216602623462677,
+      "learning_rate": 3.668942953021357e-05,
+      "loss": 0.0159,
+      "step": 36390
+    },
+    {
+      "action_loss": 0.008342002518475056,
+      "epoch": 32.72482014388489,
+      "step": 36390
+    },
+    {
+      "epoch": 32.72482014388489,
+      "step": 36390,
+      "torque_loss": 0.14404691755771637
+    },
+    {
+      "epoch": 32.72482014388489,
+      "force_loss": 0.015245121903717518,
+      "step": 36390
+    },
+    {
+      "epoch": 32.73381294964029,
+      "grad_norm": 0.11196604371070862,
+      "learning_rate": 3.66628681564431e-05,
+      "loss": 0.0156,
+      "step": 36400
+    },
+    {
+      "action_loss": 0.0013187493896111846,
+      "epoch": 32.73381294964029,
+      "step": 36400
+    },
+    {
+      "epoch": 32.73381294964029,
+      "step": 36400,
+      "torque_loss": 0.14906540513038635
+    },
+    {
+      "epoch": 32.73381294964029,
+      "force_loss": 0.004711283836513758,
+      "step": 36400
+    },
+    {
+      "epoch": 32.742805755395686,
+      "grad_norm": 0.09018264710903168,
+      "learning_rate": 3.663631083414114e-05,
+      "loss": 0.0149,
+      "step": 36410
+    },
+    {
+      "action_loss": 0.0025115821044892073,
+      "epoch": 32.742805755395686,
+      "step": 36410
+    },
+    {
+      "epoch": 32.742805755395686,
+      "step": 36410,
+      "torque_loss": 0.1465323567390442
+    },
+    {
+      "epoch": 32.742805755395686,
+      "force_loss": 0.003507686546072364,
+      "step": 36410
+    },
+    {
+      "epoch": 32.75179856115108,
+      "grad_norm": 0.3198237121105194,
+      "learning_rate": 3.660975757137509e-05,
+      "loss": 0.0165,
+      "step": 36420
+    },
+    {
+      "action_loss": 0.0016209982568398118,
+      "epoch": 32.75179856115108,
+      "step": 36420
+    },
+    {
+      "epoch": 32.75179856115108,
+      "step": 36420,
+      "torque_loss": 0.09486324340105057
+    },
+    {
+      "epoch": 32.75179856115108,
+      "force_loss": 0.004021722823381424,
+      "step": 36420
+    },
+    {
+      "epoch": 32.760791366906474,
+      "grad_norm": 0.27158528566360474,
+      "learning_rate": 3.658320837621114e-05,
+      "loss": 0.016,
+      "step": 36430
+    },
+    {
+      "action_loss": 0.0017069907626137137,
+      "epoch": 32.760791366906474,
+      "step": 36430
+    },
+    {
+      "epoch": 32.760791366906474,
+      "step": 36430,
+      "torque_loss": 0.11200741678476334
+    },
+    {
+      "epoch": 32.760791366906474,
+      "force_loss": 0.002729924162849784,
+      "step": 36430
+    },
+    {
+      "epoch": 32.76978417266187,
+      "grad_norm": 0.19988465309143066,
+      "learning_rate": 3.655666325671426e-05,
+      "loss": 0.0182,
+      "step": 36440
+    },
+    {
+      "action_loss": 0.0011555716628208756,
+      "epoch": 32.76978417266187,
+      "step": 36440
+    },
+    {
+      "epoch": 32.76978417266187,
+      "step": 36440,
+      "torque_loss": 0.11958108097314835
+    },
+    {
+      "epoch": 32.76978417266187,
+      "force_loss": 0.0012702671810984612,
+      "step": 36440
+    },
+    {
+      "epoch": 32.77877697841727,
+      "grad_norm": 0.16070693731307983,
+      "learning_rate": 3.65301222209481e-05,
+      "loss": 0.0158,
+      "step": 36450
+    },
+    {
+      "action_loss": 0.00433853454887867,
+      "epoch": 32.77877697841727,
+      "step": 36450
+    },
+    {
+      "epoch": 32.77877697841727,
+      "step": 36450,
+      "torque_loss": 0.12327710539102554
+    },
+    {
+      "epoch": 32.77877697841727,
+      "force_loss": 0.0031252000480890274,
+      "step": 36450
+    },
+    {
+      "epoch": 32.78776978417266,
+      "grad_norm": 0.36357322335243225,
+      "learning_rate": 3.650358527697519e-05,
+      "loss": 0.022,
+      "step": 36460
+    },
+    {
+      "action_loss": 0.004157778341323137,
+      "epoch": 32.78776978417266,
+      "step": 36460
+    },
+    {
+      "epoch": 32.78776978417266,
+      "step": 36460,
+      "torque_loss": 0.10977357625961304
+    },
+    {
+      "epoch": 32.78776978417266,
+      "force_loss": 0.01005170401185751,
+      "step": 36460
+    },
+    {
+      "epoch": 32.79676258992806,
+      "grad_norm": 0.17972682416439056,
+      "learning_rate": 3.64770524328567e-05,
+      "loss": 0.0184,
+      "step": 36470
+    },
+    {
+      "action_loss": 0.00646154535934329,
+      "epoch": 32.79676258992806,
+      "step": 36470
+    },
+    {
+      "epoch": 32.79676258992806,
+      "step": 36470,
+      "torque_loss": 0.09342347830533981
+    },
+    {
+      "epoch": 32.79676258992806,
+      "force_loss": 0.006371751893311739,
+      "step": 36470
+    },
+    {
+      "epoch": 32.805755395683455,
+      "grad_norm": 0.2528167963027954,
+      "learning_rate": 3.645052369665265e-05,
+      "loss": 0.0183,
+      "step": 36480
+    },
+    {
+      "action_loss": 0.005299969110637903,
+      "epoch": 32.805755395683455,
+      "step": 36480
+    },
+    {
+      "epoch": 32.805755395683455,
+      "step": 36480,
+      "torque_loss": 0.12035242468118668
+    },
+    {
+      "epoch": 32.805755395683455,
+      "force_loss": 0.012455065734684467,
+      "step": 36480
+    },
+    {
+      "epoch": 32.814748201438846,
+      "grad_norm": 0.295187383890152,
+      "learning_rate": 3.6423999076421724e-05,
+      "loss": 0.0191,
+      "step": 36490
+    },
+    {
+      "action_loss": 0.002935246331617236,
+      "epoch": 32.814748201438846,
+      "step": 36490
+    },
+    {
+      "epoch": 32.814748201438846,
+      "step": 36490,
+      "torque_loss": 0.17187924683094025
+    },
+    {
+      "epoch": 32.814748201438846,
+      "force_loss": 0.0024404069408774376,
+      "step": 36490
+    },
+    {
+      "epoch": 32.82374100719424,
+      "grad_norm": 0.3179472088813782,
+      "learning_rate": 3.639747858022142e-05,
+      "loss": 0.0209,
+      "step": 36500
+    },
+    {
+      "action_loss": 0.005375315900892019,
+      "epoch": 32.82374100719424,
+      "step": 36500
+    },
+    {
+      "epoch": 32.82374100719424,
+      "step": 36500,
+      "torque_loss": 0.09595170617103577
+    },
+    {
+      "epoch": 32.82374100719424,
+      "force_loss": 0.015310156159102917,
+      "step": 36500
+    },
+    {
+      "epoch": 32.83273381294964,
+      "grad_norm": 0.12300152331590652,
+      "learning_rate": 3.637096221610799e-05,
+      "loss": 0.0179,
+      "step": 36510
+    },
+    {
+      "action_loss": 0.0014070924371480942,
+      "epoch": 32.83273381294964,
+      "step": 36510
+    },
+    {
+      "epoch": 32.83273381294964,
+      "step": 36510,
+      "torque_loss": 0.11820673942565918
+    },
+    {
+      "epoch": 32.83273381294964,
+      "force_loss": 0.0017465772107243538,
+      "step": 36510
+    },
+    {
+      "epoch": 32.84172661870504,
+      "grad_norm": 0.13506054878234863,
+      "learning_rate": 3.634444999213638e-05,
+      "loss": 0.0159,
+      "step": 36520
+    },
+    {
+      "action_loss": 0.0023763671051710844,
+      "epoch": 32.84172661870504,
+      "step": 36520
+    },
+    {
+      "epoch": 32.84172661870504,
+      "step": 36520,
+      "torque_loss": 0.17207233607769012
+    },
+    {
+      "epoch": 32.84172661870504,
+      "force_loss": 0.0017365027451887727,
+      "step": 36520
+    },
+    {
+      "epoch": 32.85071942446043,
+      "grad_norm": 0.25424519181251526,
+      "learning_rate": 3.6317941916360296e-05,
+      "loss": 0.019,
+      "step": 36530
+    },
+    {
+      "action_loss": 0.0013623054837808013,
+      "epoch": 32.85071942446043,
+      "step": 36530
+    },
+    {
+      "epoch": 32.85071942446043,
+      "step": 36530,
+      "torque_loss": 0.12819351255893707
+    },
+    {
+      "epoch": 32.85071942446043,
+      "force_loss": 0.00106936099473387,
+      "step": 36530
+    },
+    {
+      "epoch": 32.85971223021583,
+      "grad_norm": 0.15003550052642822,
+      "learning_rate": 3.629143799683221e-05,
+      "loss": 0.0166,
+      "step": 36540
+    },
+    {
+      "action_loss": 0.0020373777952045202,
+      "epoch": 32.85971223021583,
+      "step": 36540
+    },
+    {
+      "epoch": 32.85971223021583,
+      "step": 36540,
+      "torque_loss": 0.14321736991405487
+    },
+    {
+      "epoch": 32.85971223021583,
+      "force_loss": 0.002855272265151143,
+      "step": 36540
+    },
+    {
+      "epoch": 32.868705035971225,
+      "grad_norm": 0.24814960360527039,
+      "learning_rate": 3.626493824160331e-05,
+      "loss": 0.0184,
+      "step": 36550
+    },
+    {
+      "action_loss": 0.0011365163372829556,
+      "epoch": 32.868705035971225,
+      "step": 36550
+    },
+    {
+      "epoch": 32.868705035971225,
+      "step": 36550,
+      "torque_loss": 0.12206246703863144
+    },
+    {
+      "epoch": 32.868705035971225,
+      "force_loss": 0.002020656829699874,
+      "step": 36550
+    },
+    {
+      "epoch": 32.87769784172662,
+      "grad_norm": 0.17621426284313202,
+      "learning_rate": 3.623844265872352e-05,
+      "loss": 0.0135,
+      "step": 36560
+    },
+    {
+      "action_loss": 0.0018490878865122795,
+      "epoch": 32.87769784172662,
+      "step": 36560
+    },
+    {
+      "epoch": 32.87769784172662,
+      "step": 36560,
+      "torque_loss": 0.11712668091058731
+    },
+    {
+      "epoch": 32.87769784172662,
+      "force_loss": 0.0028901107143610716,
+      "step": 36560
+    },
+    {
+      "epoch": 32.88669064748201,
+      "grad_norm": 0.2913665175437927,
+      "learning_rate": 3.621195125624149e-05,
+      "loss": 0.0191,
+      "step": 36570
+    },
+    {
+      "action_loss": 0.006535492837429047,
+      "epoch": 32.88669064748201,
+      "step": 36570
+    },
+    {
+      "epoch": 32.88669064748201,
+      "step": 36570,
+      "torque_loss": 0.15344056487083435
+    },
+    {
+      "epoch": 32.88669064748201,
+      "force_loss": 0.0064294240437448025,
+      "step": 36570
+    },
+    {
+      "epoch": 32.89568345323741,
+      "grad_norm": 0.22605034708976746,
+      "learning_rate": 3.618546404220463e-05,
+      "loss": 0.0212,
+      "step": 36580
+    },
+    {
+      "action_loss": 0.0054314048029482365,
+      "epoch": 32.89568345323741,
+      "step": 36580
+    },
+    {
+      "epoch": 32.89568345323741,
+      "step": 36580,
+      "torque_loss": 0.12828782200813293
+    },
+    {
+      "epoch": 32.89568345323741,
+      "force_loss": 0.006194684188812971,
+      "step": 36580
+    },
+    {
+      "epoch": 32.90467625899281,
+      "grad_norm": 0.14173366129398346,
+      "learning_rate": 3.615898102465903e-05,
+      "loss": 0.0189,
+      "step": 36590
+    },
+    {
+      "action_loss": 0.012773125432431698,
+      "epoch": 32.90467625899281,
+      "step": 36590
+    },
+    {
+      "epoch": 32.90467625899281,
+      "step": 36590,
+      "torque_loss": 0.19824115931987762
+    },
+    {
+      "epoch": 32.90467625899281,
+      "force_loss": 0.007875685580074787,
+      "step": 36590
+    },
+    {
+      "epoch": 32.9136690647482,
+      "grad_norm": 0.19740355014801025,
+      "learning_rate": 3.6132502211649544e-05,
+      "loss": 0.0203,
+      "step": 36600
+    },
+    {
+      "action_loss": 0.0032737914007157087,
+      "epoch": 32.9136690647482,
+      "step": 36600
+    },
+    {
+      "epoch": 32.9136690647482,
+      "step": 36600,
+      "torque_loss": 0.10628330707550049
+    },
+    {
+      "epoch": 32.9136690647482,
+      "force_loss": 0.0048421830870211124,
+      "step": 36600
+    },
+    {
+      "epoch": 32.922661870503596,
+      "grad_norm": 0.320834219455719,
+      "learning_rate": 3.610602761121975e-05,
+      "loss": 0.0165,
+      "step": 36610
+    },
+    {
+      "action_loss": 0.002689277520403266,
+      "epoch": 32.922661870503596,
+      "step": 36610
+    },
+    {
+      "epoch": 32.922661870503596,
+      "step": 36610,
+      "torque_loss": 0.15329383313655853
+    },
+    {
+      "epoch": 32.922661870503596,
+      "force_loss": 0.003482213243842125,
+      "step": 36610
+    },
+    {
+      "epoch": 32.931654676258994,
+      "grad_norm": 0.16296826303005219,
+      "learning_rate": 3.6079557231411897e-05,
+      "loss": 0.0182,
+      "step": 36620
+    },
+    {
+      "action_loss": 0.002387323882430792,
+      "epoch": 32.931654676258994,
+      "step": 36620
+    },
+    {
+      "epoch": 32.931654676258994,
+      "step": 36620,
+      "torque_loss": 0.1422547698020935
+    },
+    {
+      "epoch": 32.931654676258994,
+      "force_loss": 0.004744958132505417,
+      "step": 36620
+    },
+    {
+      "epoch": 32.94064748201439,
+      "grad_norm": 0.09802334755659103,
+      "learning_rate": 3.6053091080267035e-05,
+      "loss": 0.0156,
+      "step": 36630
+    },
+    {
+      "action_loss": 0.0044053904712200165,
+      "epoch": 32.94064748201439,
+      "step": 36630
+    },
+    {
+      "epoch": 32.94064748201439,
+      "step": 36630,
+      "torque_loss": 0.11917876452207565
+    },
+    {
+      "epoch": 32.94064748201439,
+      "force_loss": 0.003263379679992795,
+      "step": 36630
+    },
+    {
+      "epoch": 32.94964028776978,
+      "grad_norm": 0.09336693584918976,
+      "learning_rate": 3.602662916582483e-05,
+      "loss": 0.0156,
+      "step": 36640
+    },
+    {
+      "action_loss": 0.003043554723262787,
+      "epoch": 32.94964028776978,
+      "step": 36640
+    },
+    {
+      "epoch": 32.94964028776978,
+      "step": 36640,
+      "torque_loss": 0.12353932857513428
+    },
+    {
+      "epoch": 32.94964028776978,
+      "force_loss": 0.0018917560810223222,
+      "step": 36640
+    },
+    {
+      "epoch": 32.95863309352518,
+      "grad_norm": 0.14380191266536713,
+      "learning_rate": 3.600017149612375e-05,
+      "loss": 0.0169,
+      "step": 36650
+    },
+    {
+      "action_loss": 0.0009219881030730903,
+      "epoch": 32.95863309352518,
+      "step": 36650
+    },
+    {
+      "epoch": 32.95863309352518,
+      "step": 36650,
+      "torque_loss": 0.1801210641860962
+    },
+    {
+      "epoch": 32.95863309352518,
+      "force_loss": 0.0009728195145726204,
+      "step": 36650
+    },
+    {
+      "epoch": 32.96762589928058,
+      "grad_norm": 0.21092064678668976,
+      "learning_rate": 3.5973718079200935e-05,
+      "loss": 0.0199,
+      "step": 36660
+    },
+    {
+      "action_loss": 0.0006425574538297951,
+      "epoch": 32.96762589928058,
+      "step": 36660
+    },
+    {
+      "epoch": 32.96762589928058,
+      "step": 36660,
+      "torque_loss": 0.09892163425683975
+    },
+    {
+      "epoch": 32.96762589928058,
+      "force_loss": 0.0011274392018094659,
+      "step": 36660
+    },
+    {
+      "epoch": 32.976618705035975,
+      "grad_norm": 0.0832480788230896,
+      "learning_rate": 3.5947268923092216e-05,
+      "loss": 0.0157,
+      "step": 36670
+    },
+    {
+      "action_loss": 0.00119773019105196,
+      "epoch": 32.976618705035975,
+      "step": 36670
+    },
+    {
+      "epoch": 32.976618705035975,
+      "step": 36670,
+      "torque_loss": 0.1590159684419632
+    },
+    {
+      "epoch": 32.976618705035975,
+      "force_loss": 0.0012387711321935058,
+      "step": 36670
+    },
+    {
+      "epoch": 32.985611510791365,
+      "grad_norm": 0.13251672685146332,
+      "learning_rate": 3.592082403583216e-05,
+      "loss": 0.0166,
+      "step": 36680
+    },
+    {
+      "action_loss": 0.0008663971093483269,
+      "epoch": 32.985611510791365,
+      "step": 36680
+    },
+    {
+      "epoch": 32.985611510791365,
+      "step": 36680,
+      "torque_loss": 0.10590314865112305
+    },
+    {
+      "epoch": 32.985611510791365,
+      "force_loss": 0.0011033309856429696,
+      "step": 36680
+    },
+    {
+      "epoch": 32.99460431654676,
+      "grad_norm": 0.1287786364555359,
+      "learning_rate": 3.5894383425454004e-05,
+      "loss": 0.0166,
+      "step": 36690
+    },
+    {
+      "action_loss": 0.0021171323023736477,
+      "epoch": 32.99460431654676,
+      "step": 36690
+    },
+    {
+      "epoch": 32.99460431654676,
+      "step": 36690,
+      "torque_loss": 0.13737957179546356
+    },
+    {
+      "epoch": 32.99460431654676,
+      "force_loss": 0.0046308995224535465,
+      "step": 36690
+    },
+    {
+      "epoch": 33.00359712230216,
+      "grad_norm": 0.3398348391056061,
+      "learning_rate": 3.586794709998975e-05,
+      "loss": 0.0161,
+      "step": 36700
+    },
+    {
+      "action_loss": 0.006835767533630133,
+      "epoch": 33.00359712230216,
+      "step": 36700
+    },
+    {
+      "epoch": 33.00359712230216,
+      "step": 36700,
+      "torque_loss": 0.10759087651968002
+    },
+    {
+      "epoch": 33.00359712230216,
+      "force_loss": 0.008363459259271622,
+      "step": 36700
+    },
+    {
+      "epoch": 33.01258992805755,
+      "grad_norm": 0.19697205722332,
+      "learning_rate": 3.584151506747002e-05,
+      "loss": 0.0177,
+      "step": 36710
+    },
+    {
+      "action_loss": 0.0015691443113610148,
+      "epoch": 33.01258992805755,
+      "step": 36710
+    },
+    {
+      "epoch": 33.01258992805755,
+      "step": 36710,
+      "torque_loss": 0.11815029382705688
+    },
+    {
+      "epoch": 33.01258992805755,
+      "force_loss": 0.002755112946033478,
+      "step": 36710
+    },
+    {
+      "epoch": 33.02158273381295,
+      "grad_norm": 0.3054356575012207,
+      "learning_rate": 3.581508733592418e-05,
+      "loss": 0.0188,
+      "step": 36720
+    },
+    {
+      "action_loss": 0.010482266545295715,
+      "epoch": 33.02158273381295,
+      "step": 36720
+    },
+    {
+      "epoch": 33.02158273381295,
+      "step": 36720,
+      "torque_loss": 0.12076478451490402
+    },
+    {
+      "epoch": 33.02158273381295,
+      "force_loss": 0.009484922513365746,
+      "step": 36720
+    },
+    {
+      "epoch": 33.030575539568346,
+      "grad_norm": 0.14727911353111267,
+      "learning_rate": 3.5788663913380297e-05,
+      "loss": 0.0177,
+      "step": 36730
+    },
+    {
+      "action_loss": 0.005812238901853561,
+      "epoch": 33.030575539568346,
+      "step": 36730
+    },
+    {
+      "epoch": 33.030575539568346,
+      "step": 36730,
+      "torque_loss": 0.1283504217863083
+    },
+    {
+      "epoch": 33.030575539568346,
+      "force_loss": 0.006615789141505957,
+      "step": 36730
+    },
+    {
+      "epoch": 33.039568345323744,
+      "grad_norm": 0.13999532163143158,
+      "learning_rate": 3.576224480786506e-05,
+      "loss": 0.0233,
+      "step": 36740
+    },
+    {
+      "action_loss": 0.0031441580504179,
+      "epoch": 33.039568345323744,
+      "step": 36740
+    },
+    {
+      "epoch": 33.039568345323744,
+      "step": 36740,
+      "torque_loss": 0.09089095145463943
+    },
+    {
+      "epoch": 33.039568345323744,
+      "force_loss": 0.006952656898647547,
+      "step": 36740
+    },
+    {
+      "epoch": 33.048561151079134,
+      "grad_norm": 0.1500289887189865,
+      "learning_rate": 3.573583002740393e-05,
+      "loss": 0.0181,
+      "step": 36750
+    },
+    {
+      "action_loss": 0.009716532193124294,
+      "epoch": 33.048561151079134,
+      "step": 36750
+    },
+    {
+      "epoch": 33.048561151079134,
+      "step": 36750,
+      "torque_loss": 0.170123890042305
+    },
+    {
+      "epoch": 33.048561151079134,
+      "force_loss": 0.010109642520546913,
+      "step": 36750
+    },
+    {
+      "epoch": 33.05755395683453,
+      "grad_norm": 0.30803382396698,
+      "learning_rate": 3.570941958002103e-05,
+      "loss": 0.0187,
+      "step": 36760
+    },
+    {
+      "action_loss": 0.001115148887038231,
+      "epoch": 33.05755395683453,
+      "step": 36760
+    },
+    {
+      "epoch": 33.05755395683453,
+      "step": 36760,
+      "torque_loss": 0.11222013831138611
+    },
+    {
+      "epoch": 33.05755395683453,
+      "force_loss": 0.002676828531548381,
+      "step": 36760
+    },
+    {
+      "epoch": 33.06654676258993,
+      "grad_norm": 0.1865822970867157,
+      "learning_rate": 3.568301347373912e-05,
+      "loss": 0.0157,
+      "step": 36770
+    },
+    {
+      "action_loss": 0.0011608978966251016,
+      "epoch": 33.06654676258993,
+      "step": 36770
+    },
+    {
+      "epoch": 33.06654676258993,
+      "step": 36770,
+      "torque_loss": 0.1131066158413887
+    },
+    {
+      "epoch": 33.06654676258993,
+      "force_loss": 0.0014423249522224069,
+      "step": 36770
+    },
+    {
+      "epoch": 33.07553956834533,
+      "grad_norm": 0.11792761087417603,
+      "learning_rate": 3.5656611716579726e-05,
+      "loss": 0.017,
+      "step": 36780
+    },
+    {
+      "action_loss": 0.0028036274015903473,
+      "epoch": 33.07553956834533,
+      "step": 36780
+    },
+    {
+      "epoch": 33.07553956834533,
+      "step": 36780,
+      "torque_loss": 0.13723760843276978
+    },
+    {
+      "epoch": 33.07553956834533,
+      "force_loss": 0.004048797767609358,
+      "step": 36780
+    },
+    {
+      "epoch": 33.08453237410072,
+      "grad_norm": 0.1791292130947113,
+      "learning_rate": 3.5630214316562946e-05,
+      "loss": 0.0218,
+      "step": 36790
+    },
+    {
+      "action_loss": 0.001995891099795699,
+      "epoch": 33.08453237410072,
+      "step": 36790
+    },
+    {
+      "epoch": 33.08453237410072,
+      "step": 36790,
+      "torque_loss": 0.11438407748937607
+    },
+    {
+      "epoch": 33.08453237410072,
+      "force_loss": 0.003643640084192157,
+      "step": 36790
+    },
+    {
+      "epoch": 33.093525179856115,
+      "grad_norm": 0.1332855075597763,
+      "learning_rate": 3.560382128170766e-05,
+      "loss": 0.016,
+      "step": 36800
+    },
+    {
+      "action_loss": 0.06602529436349869,
+      "epoch": 33.093525179856115,
+      "step": 36800
+    },
+    {
+      "epoch": 33.093525179856115,
+      "step": 36800,
+      "torque_loss": 0.26105237007141113
+    },
+    {
+      "epoch": 33.093525179856115,
+      "force_loss": 0.05167825147509575,
+      "step": 36800
+    },
+    {
+      "epoch": 33.10251798561151,
+      "grad_norm": 0.07148727774620056,
+      "learning_rate": 3.5577432620031374e-05,
+      "loss": 0.0223,
+      "step": 36810
+    },
+    {
+      "action_loss": 0.002395481104031205,
+      "epoch": 33.10251798561151,
+      "step": 36810
+    },
+    {
+      "epoch": 33.10251798561151,
+      "step": 36810,
+      "torque_loss": 0.10932862758636475
+    },
+    {
+      "epoch": 33.10251798561151,
+      "force_loss": 0.003633117536082864,
+      "step": 36810
+    },
+    {
+      "epoch": 33.111510791366904,
+      "grad_norm": 0.1899031400680542,
+      "learning_rate": 3.5551048339550216e-05,
+      "loss": 0.0174,
+      "step": 36820
+    },
+    {
+      "action_loss": 0.008279651403427124,
+      "epoch": 33.111510791366904,
+      "step": 36820
+    },
+    {
+      "epoch": 33.111510791366904,
+      "step": 36820,
+      "torque_loss": 0.14984314143657684
+    },
+    {
+      "epoch": 33.111510791366904,
+      "force_loss": 0.0064571392722427845,
+      "step": 36820
+    },
+    {
+      "epoch": 33.1205035971223,
+      "grad_norm": 0.247861847281456,
+      "learning_rate": 3.55246684482791e-05,
+      "loss": 0.0195,
+      "step": 36830
+    },
+    {
+      "action_loss": 0.005978280212730169,
+      "epoch": 33.1205035971223,
+      "step": 36830
+    },
+    {
+      "epoch": 33.1205035971223,
+      "step": 36830,
+      "torque_loss": 0.09209439903497696
+    },
+    {
+      "epoch": 33.1205035971223,
+      "force_loss": 0.0068796114064753056,
+      "step": 36830
+    },
+    {
+      "epoch": 33.1294964028777,
+      "grad_norm": 0.16852621734142303,
+      "learning_rate": 3.5498292954231496e-05,
+      "loss": 0.0168,
+      "step": 36840
+    },
+    {
+      "action_loss": 0.0018792202463373542,
+      "epoch": 33.1294964028777,
+      "step": 36840
+    },
+    {
+      "epoch": 33.1294964028777,
+      "step": 36840,
+      "torque_loss": 0.15244042873382568
+    },
+    {
+      "epoch": 33.1294964028777,
+      "force_loss": 0.00158885947894305,
+      "step": 36840
+    },
+    {
+      "epoch": 33.138489208633096,
+      "grad_norm": 0.357898086309433,
+      "learning_rate": 3.54719218654196e-05,
+      "loss": 0.0196,
+      "step": 36850
+    },
+    {
+      "action_loss": 0.016745124012231827,
+      "epoch": 33.138489208633096,
+      "step": 36850
+    },
+    {
+      "epoch": 33.138489208633096,
+      "step": 36850,
+      "torque_loss": 0.09298679977655411
+    },
+    {
+      "epoch": 33.138489208633096,
+      "force_loss": 0.011909134685993195,
+      "step": 36850
+    },
+    {
+      "epoch": 33.14748201438849,
+      "grad_norm": 0.28340256214141846,
+      "learning_rate": 3.544555518985425e-05,
+      "loss": 0.0174,
+      "step": 36860
+    },
+    {
+      "action_loss": 0.002974806586280465,
+      "epoch": 33.14748201438849,
+      "step": 36860
+    },
+    {
+      "epoch": 33.14748201438849,
+      "step": 36860,
+      "torque_loss": 0.13511574268341064
+    },
+    {
+      "epoch": 33.14748201438849,
+      "force_loss": 0.0018620028859004378,
+      "step": 36860
+    },
+    {
+      "epoch": 33.156474820143885,
+      "grad_norm": 0.2921615540981293,
+      "learning_rate": 3.541919293554494e-05,
+      "loss": 0.0165,
+      "step": 36870
+    },
+    {
+      "action_loss": 0.0021552962716668844,
+      "epoch": 33.156474820143885,
+      "step": 36870
+    },
+    {
+      "epoch": 33.156474820143885,
+      "step": 36870,
+      "torque_loss": 0.13561786711215973
+    },
+    {
+      "epoch": 33.156474820143885,
+      "force_loss": 0.001641663839109242,
+      "step": 36870
+    },
+    {
+      "epoch": 33.16546762589928,
+      "grad_norm": 0.18524503707885742,
+      "learning_rate": 3.539283511049985e-05,
+      "loss": 0.0193,
+      "step": 36880
+    },
+    {
+      "action_loss": 0.00771550415083766,
+      "epoch": 33.16546762589928,
+      "step": 36880
+    },
+    {
+      "epoch": 33.16546762589928,
+      "step": 36880,
+      "torque_loss": 0.20462913811206818
+    },
+    {
+      "epoch": 33.16546762589928,
+      "force_loss": 0.006421716418117285,
+      "step": 36880
+    },
+    {
+      "epoch": 33.17446043165467,
+      "grad_norm": 0.24745787680149078,
+      "learning_rate": 3.5366481722725755e-05,
+      "loss": 0.0173,
+      "step": 36890
+    },
+    {
+      "action_loss": 0.01282097864896059,
+      "epoch": 33.17446043165467,
+      "step": 36890
+    },
+    {
+      "epoch": 33.17446043165467,
+      "step": 36890,
+      "torque_loss": 0.10666092485189438
+    },
+    {
+      "epoch": 33.17446043165467,
+      "force_loss": 0.022650517523288727,
+      "step": 36890
+    },
+    {
+      "epoch": 33.18345323741007,
+      "grad_norm": 0.21280813217163086,
+      "learning_rate": 3.534013278022816e-05,
+      "loss": 0.0163,
+      "step": 36900
+    },
+    {
+      "action_loss": 0.01911107450723648,
+      "epoch": 33.18345323741007,
+      "step": 36900
+    },
+    {
+      "epoch": 33.18345323741007,
+      "step": 36900,
+      "torque_loss": 0.16768568754196167
+    },
+    {
+      "epoch": 33.18345323741007,
+      "force_loss": 0.023493608459830284,
+      "step": 36900
+    },
+    {
+      "epoch": 33.19244604316547,
+      "grad_norm": 0.40644174814224243,
+      "learning_rate": 3.531378829101113e-05,
+      "loss": 0.0196,
+      "step": 36910
+    },
+    {
+      "action_loss": 0.004462155979126692,
+      "epoch": 33.19244604316547,
+      "step": 36910
+    },
+    {
+      "epoch": 33.19244604316547,
+      "step": 36910,
+      "torque_loss": 0.10983768850564957
+    },
+    {
+      "epoch": 33.19244604316547,
+      "force_loss": 0.005608848761767149,
+      "step": 36910
+    },
+    {
+      "epoch": 33.201438848920866,
+      "grad_norm": 0.2414204627275467,
+      "learning_rate": 3.528744826307746e-05,
+      "loss": 0.019,
+      "step": 36920
+    },
+    {
+      "action_loss": 0.002865339396521449,
+      "epoch": 33.201438848920866,
+      "step": 36920
+    },
+    {
+      "epoch": 33.201438848920866,
+      "step": 36920,
+      "torque_loss": 0.1108912006020546
+    },
+    {
+      "epoch": 33.201438848920866,
+      "force_loss": 0.005613286048173904,
+      "step": 36920
+    },
+    {
+      "epoch": 33.210431654676256,
+      "grad_norm": 0.2705686688423157,
+      "learning_rate": 3.5261112704428554e-05,
+      "loss": 0.018,
+      "step": 36930
+    },
+    {
+      "action_loss": 0.0017391294240951538,
+      "epoch": 33.210431654676256,
+      "step": 36930
+    },
+    {
+      "epoch": 33.210431654676256,
+      "step": 36930,
+      "torque_loss": 0.15713131427764893
+    },
+    {
+      "epoch": 33.210431654676256,
+      "force_loss": 0.0019209147430956364,
+      "step": 36930
+    },
+    {
+      "epoch": 33.219424460431654,
+      "grad_norm": 0.2855124771595001,
+      "learning_rate": 3.523478162306443e-05,
+      "loss": 0.0169,
+      "step": 36940
+    },
+    {
+      "action_loss": 0.0016624769195914268,
+      "epoch": 33.219424460431654,
+      "step": 36940
+    },
+    {
+      "epoch": 33.219424460431654,
+      "step": 36940,
+      "torque_loss": 0.1696600317955017
+    },
+    {
+      "epoch": 33.219424460431654,
+      "force_loss": 0.0020895209163427353,
+      "step": 36940
+    },
+    {
+      "epoch": 33.22841726618705,
+      "grad_norm": 0.136759951710701,
+      "learning_rate": 3.520845502698381e-05,
+      "loss": 0.0161,
+      "step": 36950
+    },
+    {
+      "action_loss": 0.001486244611442089,
+      "epoch": 33.22841726618705,
+      "step": 36950
+    },
+    {
+      "epoch": 33.22841726618705,
+      "step": 36950,
+      "torque_loss": 0.09340252727270126
+    },
+    {
+      "epoch": 33.22841726618705,
+      "force_loss": 0.0017930116737261415,
+      "step": 36950
+    },
+    {
+      "epoch": 33.23741007194245,
+      "grad_norm": 0.3832060694694519,
+      "learning_rate": 3.5182132924184005e-05,
+      "loss": 0.0169,
+      "step": 36960
+    },
+    {
+      "action_loss": 0.0057553499937057495,
+      "epoch": 33.23741007194245,
+      "step": 36960
+    },
+    {
+      "epoch": 33.23741007194245,
+      "step": 36960,
+      "torque_loss": 0.10780555009841919
+    },
+    {
+      "epoch": 33.23741007194245,
+      "force_loss": 0.007818729616701603,
+      "step": 36960
+    },
+    {
+      "epoch": 33.24640287769784,
+      "grad_norm": 0.458911269903183,
+      "learning_rate": 3.5155815322660966e-05,
+      "loss": 0.0195,
+      "step": 36970
+    },
+    {
+      "action_loss": 0.010337863117456436,
+      "epoch": 33.24640287769784,
+      "step": 36970
+    },
+    {
+      "epoch": 33.24640287769784,
+      "step": 36970,
+      "torque_loss": 0.1338314265012741
+    },
+    {
+      "epoch": 33.24640287769784,
+      "force_loss": 0.0112711600959301,
+      "step": 36970
+    },
+    {
+      "epoch": 33.25539568345324,
+      "grad_norm": 0.5410187840461731,
+      "learning_rate": 3.512950223040931e-05,
+      "loss": 0.0171,
+      "step": 36980
+    },
+    {
+      "action_loss": 0.005152761470526457,
+      "epoch": 33.25539568345324,
+      "step": 36980
+    },
+    {
+      "epoch": 33.25539568345324,
+      "step": 36980,
+      "torque_loss": 0.09719046205282211
+    },
+    {
+      "epoch": 33.25539568345324,
+      "force_loss": 0.004004388581961393,
+      "step": 36980
+    },
+    {
+      "epoch": 33.264388489208635,
+      "grad_norm": 0.3527100086212158,
+      "learning_rate": 3.5103193655422216e-05,
+      "loss": 0.0181,
+      "step": 36990
+    },
+    {
+      "action_loss": 0.0041131447069346905,
+      "epoch": 33.264388489208635,
+      "step": 36990
+    },
+    {
+      "epoch": 33.264388489208635,
+      "step": 36990,
+      "torque_loss": 0.10060495883226395
+    },
+    {
+      "epoch": 33.264388489208635,
+      "force_loss": 0.004139088559895754,
+      "step": 36990
+    },
+    {
+      "epoch": 33.273381294964025,
+      "grad_norm": 0.16072429716587067,
+      "learning_rate": 3.5076889605691596e-05,
+      "loss": 0.0176,
+      "step": 37000
+    },
+    {
+      "action_loss": 0.0035816740710288286,
+      "epoch": 33.273381294964025,
+      "step": 37000
+    },
+    {
+      "epoch": 33.273381294964025,
+      "step": 37000,
+      "torque_loss": 0.10252014547586441
+    },
+    {
+      "epoch": 33.273381294964025,
+      "force_loss": 0.002722412347793579,
+      "step": 37000
+    },
+    {
+      "epoch": 33.28237410071942,
+      "grad_norm": 0.17735706269741058,
+      "learning_rate": 3.505059008920787e-05,
+      "loss": 0.019,
+      "step": 37010
+    },
+    {
+      "action_loss": 0.00379945314489305,
+      "epoch": 33.28237410071942,
+      "step": 37010
+    },
+    {
+      "epoch": 33.28237410071942,
+      "step": 37010,
+      "torque_loss": 0.11861858516931534
+    },
+    {
+      "epoch": 33.28237410071942,
+      "force_loss": 0.0024994960986077785,
+      "step": 37010
+    },
+    {
+      "epoch": 33.29136690647482,
+      "grad_norm": 0.22336888313293457,
+      "learning_rate": 3.502429511396016e-05,
+      "loss": 0.0233,
+      "step": 37020
+    },
+    {
+      "action_loss": 0.0022373658139258623,
+      "epoch": 33.29136690647482,
+      "step": 37020
+    },
+    {
+      "epoch": 33.29136690647482,
+      "step": 37020,
+      "torque_loss": 0.0988149642944336
+    },
+    {
+      "epoch": 33.29136690647482,
+      "force_loss": 0.0013564390828832984,
+      "step": 37020
+    },
+    {
+      "epoch": 33.30035971223022,
+      "grad_norm": 0.20216836035251617,
+      "learning_rate": 3.4998004687936196e-05,
+      "loss": 0.0181,
+      "step": 37030
+    },
+    {
+      "action_loss": 0.0014282582560554147,
+      "epoch": 33.30035971223022,
+      "step": 37030
+    },
+    {
+      "epoch": 33.30035971223022,
+      "step": 37030,
+      "torque_loss": 0.10763565450906754
+    },
+    {
+      "epoch": 33.30035971223022,
+      "force_loss": 0.0018201001221314073,
+      "step": 37030
+    },
+    {
+      "epoch": 33.30935251798561,
+      "grad_norm": 0.22850529849529266,
+      "learning_rate": 3.497171881912229e-05,
+      "loss": 0.017,
+      "step": 37040
+    },
+    {
+      "action_loss": 0.003305848455056548,
+      "epoch": 33.30935251798561,
+      "step": 37040
+    },
+    {
+      "epoch": 33.30935251798561,
+      "step": 37040,
+      "torque_loss": 0.14173872768878937
+    },
+    {
+      "epoch": 33.30935251798561,
+      "force_loss": 0.0017273053526878357,
+      "step": 37040
+    },
+    {
+      "epoch": 33.318345323741006,
+      "grad_norm": 0.12240440398454666,
+      "learning_rate": 3.494543751550342e-05,
+      "loss": 0.0191,
+      "step": 37050
+    },
+    {
+      "action_loss": 0.0035281090531498194,
+      "epoch": 33.318345323741006,
+      "step": 37050
+    },
+    {
+      "epoch": 33.318345323741006,
+      "step": 37050,
+      "torque_loss": 0.10068196058273315
+    },
+    {
+      "epoch": 33.318345323741006,
+      "force_loss": 0.0063293371349573135,
+      "step": 37050
+    },
+    {
+      "epoch": 33.327338129496404,
+      "grad_norm": 0.3412223160266876,
+      "learning_rate": 3.491916078506313e-05,
+      "loss": 0.0195,
+      "step": 37060
+    },
+    {
+      "action_loss": 0.00252143875695765,
+      "epoch": 33.327338129496404,
+      "step": 37060
+    },
+    {
+      "epoch": 33.327338129496404,
+      "step": 37060,
+      "torque_loss": 0.08746298402547836
+    },
+    {
+      "epoch": 33.327338129496404,
+      "force_loss": 0.002379532204940915,
+      "step": 37060
+    },
+    {
+      "epoch": 33.3363309352518,
+      "grad_norm": 0.2112552970647812,
+      "learning_rate": 3.489288863578361e-05,
+      "loss": 0.0219,
+      "step": 37070
+    },
+    {
+      "action_loss": 0.007381019648164511,
+      "epoch": 33.3363309352518,
+      "step": 37070
+    },
+    {
+      "epoch": 33.3363309352518,
+      "step": 37070,
+      "torque_loss": 0.09913697838783264
+    },
+    {
+      "epoch": 33.3363309352518,
+      "force_loss": 0.008412917144596577,
+      "step": 37070
+    },
+    {
+      "epoch": 33.34532374100719,
+      "grad_norm": 0.20136091113090515,
+      "learning_rate": 3.4866621075645646e-05,
+      "loss": 0.0196,
+      "step": 37080
+    },
+    {
+      "action_loss": 0.0018523618346080184,
+      "epoch": 33.34532374100719,
+      "step": 37080
+    },
+    {
+      "epoch": 33.34532374100719,
+      "step": 37080,
+      "torque_loss": 0.12294883280992508
+    },
+    {
+      "epoch": 33.34532374100719,
+      "force_loss": 0.003359299385920167,
+      "step": 37080
+    },
+    {
+      "epoch": 33.35431654676259,
+      "grad_norm": 0.2694815695285797,
+      "learning_rate": 3.4840358112628614e-05,
+      "loss": 0.0189,
+      "step": 37090
+    },
+    {
+      "action_loss": 0.007186661008745432,
+      "epoch": 33.35431654676259,
+      "step": 37090
+    },
+    {
+      "epoch": 33.35431654676259,
+      "step": 37090,
+      "torque_loss": 0.12499719858169556
+    },
+    {
+      "epoch": 33.35431654676259,
+      "force_loss": 0.01146344467997551,
+      "step": 37090
+    },
+    {
+      "epoch": 33.36330935251799,
+      "grad_norm": 0.17808374762535095,
+      "learning_rate": 3.481409975471053e-05,
+      "loss": 0.0151,
+      "step": 37100
+    },
+    {
+      "action_loss": 0.0011240991298109293,
+      "epoch": 33.36330935251799,
+      "step": 37100
+    },
+    {
+      "epoch": 33.36330935251799,
+      "step": 37100,
+      "torque_loss": 0.08603618294000626
+    },
+    {
+      "epoch": 33.36330935251799,
+      "force_loss": 0.0016839904710650444,
+      "step": 37100
+    },
+    {
+      "epoch": 33.37230215827338,
+      "grad_norm": 0.18893858790397644,
+      "learning_rate": 3.4787846009867986e-05,
+      "loss": 0.0204,
+      "step": 37110
+    },
+    {
+      "action_loss": 0.005135738756507635,
+      "epoch": 33.37230215827338,
+      "step": 37110
+    },
+    {
+      "epoch": 33.37230215827338,
+      "step": 37110,
+      "torque_loss": 0.09564564377069473
+    },
+    {
+      "epoch": 33.37230215827338,
+      "force_loss": 0.01340966671705246,
+      "step": 37110
+    },
+    {
+      "epoch": 33.381294964028775,
+      "grad_norm": 0.38284650444984436,
+      "learning_rate": 3.476159688607615e-05,
+      "loss": 0.0196,
+      "step": 37120
+    },
+    {
+      "action_loss": 0.001931976992636919,
+      "epoch": 33.381294964028775,
+      "step": 37120
+    },
+    {
+      "epoch": 33.381294964028775,
+      "step": 37120,
+      "torque_loss": 0.0948227047920227
+    },
+    {
+      "epoch": 33.381294964028775,
+      "force_loss": 0.00308020762167871,
+      "step": 37120
+    },
+    {
+      "epoch": 33.39028776978417,
+      "grad_norm": 0.2643629014492035,
+      "learning_rate": 3.4735352391308854e-05,
+      "loss": 0.0182,
+      "step": 37130
+    },
+    {
+      "action_loss": 0.0031101114582270384,
+      "epoch": 33.39028776978417,
+      "step": 37130
+    },
+    {
+      "epoch": 33.39028776978417,
+      "step": 37130,
+      "torque_loss": 0.10445722192525864
+    },
+    {
+      "epoch": 33.39028776978417,
+      "force_loss": 0.004710142035037279,
+      "step": 37130
+    },
+    {
+      "epoch": 33.39928057553957,
+      "grad_norm": 0.273506224155426,
+      "learning_rate": 3.4709112533538446e-05,
+      "loss": 0.0182,
+      "step": 37140
+    },
+    {
+      "action_loss": 0.001914509921334684,
+      "epoch": 33.39928057553957,
+      "step": 37140
+    },
+    {
+      "epoch": 33.39928057553957,
+      "step": 37140,
+      "torque_loss": 0.11653489619493484
+    },
+    {
+      "epoch": 33.39928057553957,
+      "force_loss": 0.002778188558295369,
+      "step": 37140
+    },
+    {
+      "epoch": 33.40827338129496,
+      "grad_norm": 0.20317287743091583,
+      "learning_rate": 3.4682877320735934e-05,
+      "loss": 0.013,
+      "step": 37150
+    },
+    {
+      "action_loss": 0.0028175953775644302,
+      "epoch": 33.40827338129496,
+      "step": 37150
+    },
+    {
+      "epoch": 33.40827338129496,
+      "step": 37150,
+      "torque_loss": 0.13847455382347107
+    },
+    {
+      "epoch": 33.40827338129496,
+      "force_loss": 0.0015900194412097335,
+      "step": 37150
+    },
+    {
+      "epoch": 33.41726618705036,
+      "grad_norm": 0.16540081799030304,
+      "learning_rate": 3.465664676087085e-05,
+      "loss": 0.0185,
+      "step": 37160
+    },
+    {
+      "action_loss": 0.006076851859688759,
+      "epoch": 33.41726618705036,
+      "step": 37160
+    },
+    {
+      "epoch": 33.41726618705036,
+      "step": 37160,
+      "torque_loss": 0.1168292760848999
+    },
+    {
+      "epoch": 33.41726618705036,
+      "force_loss": 0.004960244055837393,
+      "step": 37160
+    },
+    {
+      "epoch": 33.42625899280576,
+      "grad_norm": 0.12980490922927856,
+      "learning_rate": 3.463042086191136e-05,
+      "loss": 0.0184,
+      "step": 37170
+    },
+    {
+      "action_loss": 0.00509148882701993,
+      "epoch": 33.42625899280576,
+      "step": 37170
+    },
+    {
+      "epoch": 33.42625899280576,
+      "step": 37170,
+      "torque_loss": 0.14884944260120392
+    },
+    {
+      "epoch": 33.42625899280576,
+      "force_loss": 0.006237113382667303,
+      "step": 37170
+    },
+    {
+      "epoch": 33.435251798561154,
+      "grad_norm": 0.2110956907272339,
+      "learning_rate": 3.460419963182423e-05,
+      "loss": 0.0182,
+      "step": 37180
+    },
+    {
+      "action_loss": 0.003542988793924451,
+      "epoch": 33.435251798561154,
+      "step": 37180
+    },
+    {
+      "epoch": 33.435251798561154,
+      "step": 37180,
+      "torque_loss": 0.1327694058418274
+    },
+    {
+      "epoch": 33.435251798561154,
+      "force_loss": 0.007422368507832289,
+      "step": 37180
+    },
+    {
+      "epoch": 33.444244604316545,
+      "grad_norm": 0.12511028349399567,
+      "learning_rate": 3.457798307857473e-05,
+      "loss": 0.017,
+      "step": 37190
+    },
+    {
+      "action_loss": 0.0014243010664358735,
+      "epoch": 33.444244604316545,
+      "step": 37190
+    },
+    {
+      "epoch": 33.444244604316545,
+      "step": 37190,
+      "torque_loss": 0.07623125612735748
+    },
+    {
+      "epoch": 33.444244604316545,
+      "force_loss": 0.001706250011920929,
+      "step": 37190
+    },
+    {
+      "epoch": 33.45323741007194,
+      "grad_norm": 0.32506227493286133,
+      "learning_rate": 3.455177121012678e-05,
+      "loss": 0.0182,
+      "step": 37200
+    },
+    {
+      "action_loss": 0.0025057748425751925,
+      "epoch": 33.45323741007194,
+      "step": 37200
+    },
+    {
+      "epoch": 33.45323741007194,
+      "step": 37200,
+      "torque_loss": 0.0937427505850792
+    },
+    {
+      "epoch": 33.45323741007194,
+      "force_loss": 0.002118638949468732,
+      "step": 37200
+    },
+    {
+      "epoch": 33.46223021582734,
+      "grad_norm": 0.2868201434612274,
+      "learning_rate": 3.452556403444285e-05,
+      "loss": 0.0186,
+      "step": 37210
+    },
+    {
+      "action_loss": 0.0016576891066506505,
+      "epoch": 33.46223021582734,
+      "step": 37210
+    },
+    {
+      "epoch": 33.46223021582734,
+      "step": 37210,
+      "torque_loss": 0.13382239639759064
+    },
+    {
+      "epoch": 33.46223021582734,
+      "force_loss": 0.0017278121085837483,
+      "step": 37210
+    },
+    {
+      "epoch": 33.47122302158273,
+      "grad_norm": 0.45674651861190796,
+      "learning_rate": 3.4499361559483975e-05,
+      "loss": 0.0177,
+      "step": 37220
+    },
+    {
+      "action_loss": 0.0016105164540931582,
+      "epoch": 33.47122302158273,
+      "step": 37220
+    },
+    {
+      "epoch": 33.47122302158273,
+      "step": 37220,
+      "torque_loss": 0.09298282116651535
+    },
+    {
+      "epoch": 33.47122302158273,
+      "force_loss": 0.00311430380679667,
+      "step": 37220
+    },
+    {
+      "epoch": 33.48021582733813,
+      "grad_norm": 0.11264558881521225,
+      "learning_rate": 3.44731637932098e-05,
+      "loss": 0.0154,
+      "step": 37230
+    },
+    {
+      "action_loss": 0.002587829949334264,
+      "epoch": 33.48021582733813,
+      "step": 37230
+    },
+    {
+      "epoch": 33.48021582733813,
+      "step": 37230,
+      "torque_loss": 0.17612791061401367
+    },
+    {
+      "epoch": 33.48021582733813,
+      "force_loss": 0.0030575525015592575,
+      "step": 37230
+    },
+    {
+      "epoch": 33.489208633093526,
+      "grad_norm": 0.13175669312477112,
+      "learning_rate": 3.44469707435785e-05,
+      "loss": 0.0159,
+      "step": 37240
+    },
+    {
+      "action_loss": 0.0034480781760066748,
+      "epoch": 33.489208633093526,
+      "step": 37240
+    },
+    {
+      "epoch": 33.489208633093526,
+      "step": 37240,
+      "torque_loss": 0.12548349797725677
+    },
+    {
+      "epoch": 33.489208633093526,
+      "force_loss": 0.004469641018658876,
+      "step": 37240
+    },
+    {
+      "epoch": 33.49820143884892,
+      "grad_norm": 0.20179106295108795,
+      "learning_rate": 3.4420782418546835e-05,
+      "loss": 0.0169,
+      "step": 37250
+    },
+    {
+      "action_loss": 0.004843270871788263,
+      "epoch": 33.49820143884892,
+      "step": 37250
+    },
+    {
+      "epoch": 33.49820143884892,
+      "step": 37250,
+      "torque_loss": 0.12337341904640198
+    },
+    {
+      "epoch": 33.49820143884892,
+      "force_loss": 0.004069761838763952,
+      "step": 37250
+    },
+    {
+      "epoch": 33.507194244604314,
+      "grad_norm": 0.15321849286556244,
+      "learning_rate": 3.439459882607012e-05,
+      "loss": 0.0159,
+      "step": 37260
+    },
+    {
+      "action_loss": 0.0017042580293491483,
+      "epoch": 33.507194244604314,
+      "step": 37260
+    },
+    {
+      "epoch": 33.507194244604314,
+      "step": 37260,
+      "torque_loss": 0.15580761432647705
+    },
+    {
+      "epoch": 33.507194244604314,
+      "force_loss": 0.0032556771766394377,
+      "step": 37260
+    },
+    {
+      "epoch": 33.51618705035971,
+      "grad_norm": 0.596222460269928,
+      "learning_rate": 3.436841997410225e-05,
+      "loss": 0.0166,
+      "step": 37270
+    },
+    {
+      "action_loss": 0.0015397807583212852,
+      "epoch": 33.51618705035971,
+      "step": 37270
+    },
+    {
+      "epoch": 33.51618705035971,
+      "step": 37270,
+      "torque_loss": 0.08465570211410522
+    },
+    {
+      "epoch": 33.51618705035971,
+      "force_loss": 0.005037181545048952,
+      "step": 37270
+    },
+    {
+      "epoch": 33.52517985611511,
+      "grad_norm": 0.13629741966724396,
+      "learning_rate": 3.434224587059567e-05,
+      "loss": 0.0163,
+      "step": 37280
+    },
+    {
+      "action_loss": 0.00683363014832139,
+      "epoch": 33.52517985611511,
+      "step": 37280
+    },
+    {
+      "epoch": 33.52517985611511,
+      "step": 37280,
+      "torque_loss": 0.1719367504119873
+    },
+    {
+      "epoch": 33.52517985611511,
+      "force_loss": 0.0065572988241910934,
+      "step": 37280
+    },
+    {
+      "epoch": 33.53417266187051,
+      "grad_norm": 0.18977464735507965,
+      "learning_rate": 3.431607652350136e-05,
+      "loss": 0.0165,
+      "step": 37290
+    },
+    {
+      "action_loss": 0.004459542687982321,
+      "epoch": 33.53417266187051,
+      "step": 37290
+    },
+    {
+      "epoch": 33.53417266187051,
+      "step": 37290,
+      "torque_loss": 0.15703324973583221
+    },
+    {
+      "epoch": 33.53417266187051,
+      "force_loss": 0.005290301516652107,
+      "step": 37290
+    },
+    {
+      "epoch": 33.5431654676259,
+      "grad_norm": 0.16229213774204254,
+      "learning_rate": 3.428991194076891e-05,
+      "loss": 0.0167,
+      "step": 37300
+    },
+    {
+      "action_loss": 0.0021986865904182196,
+      "epoch": 33.5431654676259,
+      "step": 37300
+    },
+    {
+      "epoch": 33.5431654676259,
+      "step": 37300,
+      "torque_loss": 0.06772855669260025
+    },
+    {
+      "epoch": 33.5431654676259,
+      "force_loss": 0.0009641043725423515,
+      "step": 37300
+    },
+    {
+      "epoch": 33.552158273381295,
+      "grad_norm": 0.5432208776473999,
+      "learning_rate": 3.4263752130346394e-05,
+      "loss": 0.0154,
+      "step": 37310
+    },
+    {
+      "action_loss": 0.0035273574758321047,
+      "epoch": 33.552158273381295,
+      "step": 37310
+    },
+    {
+      "epoch": 33.552158273381295,
+      "step": 37310,
+      "torque_loss": 0.0995974913239479
+    },
+    {
+      "epoch": 33.552158273381295,
+      "force_loss": 0.002668227069079876,
+      "step": 37310
+    },
+    {
+      "epoch": 33.56115107913669,
+      "grad_norm": 0.2932492196559906,
+      "learning_rate": 3.4237597100180515e-05,
+      "loss": 0.0173,
+      "step": 37320
+    },
+    {
+      "action_loss": 0.0009626168757677078,
+      "epoch": 33.56115107913669,
+      "step": 37320
+    },
+    {
+      "epoch": 33.56115107913669,
+      "step": 37320,
+      "torque_loss": 0.09957525879144669
+    },
+    {
+      "epoch": 33.56115107913669,
+      "force_loss": 0.0013694021617993712,
+      "step": 37320
+    },
+    {
+      "epoch": 33.57014388489208,
+      "grad_norm": 0.39091816544532776,
+      "learning_rate": 3.4211446858216427e-05,
+      "loss": 0.0207,
+      "step": 37330
+    },
+    {
+      "action_loss": 0.00854301918298006,
+      "epoch": 33.57014388489208,
+      "step": 37330
+    },
+    {
+      "epoch": 33.57014388489208,
+      "step": 37330,
+      "torque_loss": 0.09822949022054672
+    },
+    {
+      "epoch": 33.57014388489208,
+      "force_loss": 0.010932791978120804,
+      "step": 37330
+    },
+    {
+      "epoch": 33.57913669064748,
+      "grad_norm": 0.3789336085319519,
+      "learning_rate": 3.4185301412397915e-05,
+      "loss": 0.0281,
+      "step": 37340
+    },
+    {
+      "action_loss": 0.003750154748558998,
+      "epoch": 33.57913669064748,
+      "step": 37340
+    },
+    {
+      "epoch": 33.57913669064748,
+      "step": 37340,
+      "torque_loss": 0.11640580743551254
+    },
+    {
+      "epoch": 33.57913669064748,
+      "force_loss": 0.006120224948972464,
+      "step": 37340
+    },
+    {
+      "epoch": 33.58812949640288,
+      "grad_norm": 0.125992089509964,
+      "learning_rate": 3.415916077066729e-05,
+      "loss": 0.0211,
+      "step": 37350
+    },
+    {
+      "action_loss": 0.0021435495000332594,
+      "epoch": 33.58812949640288,
+      "step": 37350
+    },
+    {
+      "epoch": 33.58812949640288,
+      "step": 37350,
+      "torque_loss": 0.11021417379379272
+    },
+    {
+      "epoch": 33.58812949640288,
+      "force_loss": 0.002189154038205743,
+      "step": 37350
+    },
+    {
+      "epoch": 33.597122302158276,
+      "grad_norm": 0.2969815135002136,
+      "learning_rate": 3.413302494096535e-05,
+      "loss": 0.015,
+      "step": 37360
+    },
+    {
+      "action_loss": 0.005637716967612505,
+      "epoch": 33.597122302158276,
+      "step": 37360
+    },
+    {
+      "epoch": 33.597122302158276,
+      "step": 37360,
+      "torque_loss": 0.15141957998275757
+    },
+    {
+      "epoch": 33.597122302158276,
+      "force_loss": 0.009561696089804173,
+      "step": 37360
+    },
+    {
+      "epoch": 33.606115107913666,
+      "grad_norm": 0.1678556352853775,
+      "learning_rate": 3.410689393123151e-05,
+      "loss": 0.0179,
+      "step": 37370
+    },
+    {
+      "action_loss": 0.003360767848789692,
+      "epoch": 33.606115107913666,
+      "step": 37370
+    },
+    {
+      "epoch": 33.606115107913666,
+      "step": 37370,
+      "torque_loss": 0.1640554666519165
+    },
+    {
+      "epoch": 33.606115107913666,
+      "force_loss": 0.002035049954429269,
+      "step": 37370
+    },
+    {
+      "epoch": 33.615107913669064,
+      "grad_norm": 0.25882089138031006,
+      "learning_rate": 3.408076774940364e-05,
+      "loss": 0.0188,
+      "step": 37380
+    },
+    {
+      "action_loss": 0.004847631324082613,
+      "epoch": 33.615107913669064,
+      "step": 37380
+    },
+    {
+      "epoch": 33.615107913669064,
+      "step": 37380,
+      "torque_loss": 0.09756847470998764
+    },
+    {
+      "epoch": 33.615107913669064,
+      "force_loss": 0.004225565120577812,
+      "step": 37380
+    },
+    {
+      "epoch": 33.62410071942446,
+      "grad_norm": 0.3377942740917206,
+      "learning_rate": 3.40546464034182e-05,
+      "loss": 0.0178,
+      "step": 37390
+    },
+    {
+      "action_loss": 0.0038517692591995,
+      "epoch": 33.62410071942446,
+      "step": 37390
+    },
+    {
+      "epoch": 33.62410071942446,
+      "step": 37390,
+      "torque_loss": 0.12156283110380173
+    },
+    {
+      "epoch": 33.62410071942446,
+      "force_loss": 0.003442132845520973,
+      "step": 37390
+    },
+    {
+      "epoch": 33.63309352517986,
+      "grad_norm": 0.08597448468208313,
+      "learning_rate": 3.4028529901210185e-05,
+      "loss": 0.0227,
+      "step": 37400
+    },
+    {
+      "action_loss": 0.008343946188688278,
+      "epoch": 33.63309352517986,
+      "step": 37400
+    },
+    {
+      "epoch": 33.63309352517986,
+      "step": 37400,
+      "torque_loss": 0.13320598006248474
+    },
+    {
+      "epoch": 33.63309352517986,
+      "force_loss": 0.0168051365762949,
+      "step": 37400
+    },
+    {
+      "epoch": 33.64208633093525,
+      "grad_norm": 0.17610159516334534,
+      "learning_rate": 3.4002418250713086e-05,
+      "loss": 0.0213,
+      "step": 37410
+    },
+    {
+      "action_loss": 0.003728461218997836,
+      "epoch": 33.64208633093525,
+      "step": 37410
+    },
+    {
+      "epoch": 33.64208633093525,
+      "step": 37410,
+      "torque_loss": 0.12109140306711197
+    },
+    {
+      "epoch": 33.64208633093525,
+      "force_loss": 0.00312162097543478,
+      "step": 37410
+    },
+    {
+      "epoch": 33.65107913669065,
+      "grad_norm": 0.2590765655040741,
+      "learning_rate": 3.3976311459858936e-05,
+      "loss": 0.0155,
+      "step": 37420
+    },
+    {
+      "action_loss": 0.0029412468429654837,
+      "epoch": 33.65107913669065,
+      "step": 37420
+    },
+    {
+      "epoch": 33.65107913669065,
+      "step": 37420,
+      "torque_loss": 0.17498338222503662
+    },
+    {
+      "epoch": 33.65107913669065,
+      "force_loss": 0.0039038872346282005,
+      "step": 37420
+    },
+    {
+      "epoch": 33.660071942446045,
+      "grad_norm": 0.33347541093826294,
+      "learning_rate": 3.395020953657826e-05,
+      "loss": 0.0179,
+      "step": 37430
+    },
+    {
+      "action_loss": 0.0020078737288713455,
+      "epoch": 33.660071942446045,
+      "step": 37430
+    },
+    {
+      "epoch": 33.660071942446045,
+      "step": 37430,
+      "torque_loss": 0.14252640306949615
+    },
+    {
+      "epoch": 33.660071942446045,
+      "force_loss": 0.007143103983253241,
+      "step": 37430
+    },
+    {
+      "epoch": 33.669064748201436,
+      "grad_norm": 0.15101240575313568,
+      "learning_rate": 3.3924112488800165e-05,
+      "loss": 0.0161,
+      "step": 37440
+    },
+    {
+      "action_loss": 0.0037033017724752426,
+      "epoch": 33.669064748201436,
+      "step": 37440
+    },
+    {
+      "epoch": 33.669064748201436,
+      "step": 37440,
+      "torque_loss": 0.12488061189651489
+    },
+    {
+      "epoch": 33.669064748201436,
+      "force_loss": 0.002581848530098796,
+      "step": 37440
+    },
+    {
+      "epoch": 33.67805755395683,
+      "grad_norm": 0.12471426278352737,
+      "learning_rate": 3.389802032445225e-05,
+      "loss": 0.0162,
+      "step": 37450
+    },
+    {
+      "action_loss": 0.0028155117761343718,
+      "epoch": 33.67805755395683,
+      "step": 37450
+    },
+    {
+      "epoch": 33.67805755395683,
+      "step": 37450,
+      "torque_loss": 0.09916329383850098
+    },
+    {
+      "epoch": 33.67805755395683,
+      "force_loss": 0.003710748627781868,
+      "step": 37450
+    },
+    {
+      "epoch": 33.68705035971223,
+      "grad_norm": 0.20393705368041992,
+      "learning_rate": 3.38719330514606e-05,
+      "loss": 0.0158,
+      "step": 37460
+    },
+    {
+      "action_loss": 0.0023817417677491903,
+      "epoch": 33.68705035971223,
+      "step": 37460
+    },
+    {
+      "epoch": 33.68705035971223,
+      "step": 37460,
+      "torque_loss": 0.11883729696273804
+    },
+    {
+      "epoch": 33.68705035971223,
+      "force_loss": 0.0021876280661672354,
+      "step": 37460
+    },
+    {
+      "epoch": 33.69604316546763,
+      "grad_norm": 0.3987010717391968,
+      "learning_rate": 3.3845850677749866e-05,
+      "loss": 0.0183,
+      "step": 37470
+    },
+    {
+      "action_loss": 0.010615058243274689,
+      "epoch": 33.69604316546763,
+      "step": 37470
+    },
+    {
+      "epoch": 33.69604316546763,
+      "step": 37470,
+      "torque_loss": 0.15040643513202667
+    },
+    {
+      "epoch": 33.69604316546763,
+      "force_loss": 0.009057716466486454,
+      "step": 37470
+    },
+    {
+      "epoch": 33.70503597122302,
+      "grad_norm": 0.3651612102985382,
+      "learning_rate": 3.3819773211243157e-05,
+      "loss": 0.0172,
+      "step": 37480
+    },
+    {
+      "action_loss": 0.002133977832272649,
+      "epoch": 33.70503597122302,
+      "step": 37480
+    },
+    {
+      "epoch": 33.70503597122302,
+      "step": 37480,
+      "torque_loss": 0.1405663937330246
+    },
+    {
+      "epoch": 33.70503597122302,
+      "force_loss": 0.007784413173794746,
+      "step": 37480
+    },
+    {
+      "epoch": 33.71402877697842,
+      "grad_norm": 0.10498624294996262,
+      "learning_rate": 3.379370065986213e-05,
+      "loss": 0.0167,
+      "step": 37490
+    },
+    {
+      "action_loss": 0.0021034483797848225,
+      "epoch": 33.71402877697842,
+      "step": 37490
+    },
+    {
+      "epoch": 33.71402877697842,
+      "step": 37490,
+      "torque_loss": 0.0771850124001503
+    },
+    {
+      "epoch": 33.71402877697842,
+      "force_loss": 0.004364440683275461,
+      "step": 37490
+    },
+    {
+      "epoch": 33.723021582733814,
+      "grad_norm": 0.14413706958293915,
+      "learning_rate": 3.3767633031526955e-05,
+      "loss": 0.0165,
+      "step": 37500
+    },
+    {
+      "action_loss": 0.00519385002553463,
+      "epoch": 33.723021582733814,
+      "step": 37500
+    },
+    {
+      "epoch": 33.723021582733814,
+      "step": 37500,
+      "torque_loss": 0.1302100270986557
+    },
+    {
+      "epoch": 33.723021582733814,
+      "force_loss": 0.014179304242134094,
+      "step": 37500
+    },
+    {
+      "epoch": 33.73201438848921,
+      "grad_norm": 0.13210685551166534,
+      "learning_rate": 3.374157033415626e-05,
+      "loss": 0.0179,
+      "step": 37510
+    },
+    {
+      "action_loss": 0.007926639169454575,
+      "epoch": 33.73201438848921,
+      "step": 37510
+    },
+    {
+      "epoch": 33.73201438848921,
+      "step": 37510,
+      "torque_loss": 0.1363096535205841
+    },
+    {
+      "epoch": 33.73201438848921,
+      "force_loss": 0.016951395198702812,
+      "step": 37510
+    },
+    {
+      "epoch": 33.7410071942446,
+      "grad_norm": 0.3256091773509979,
+      "learning_rate": 3.371551257566723e-05,
+      "loss": 0.0167,
+      "step": 37520
+    },
+    {
+      "action_loss": 0.007042469456791878,
+      "epoch": 33.7410071942446,
+      "step": 37520
+    },
+    {
+      "epoch": 33.7410071942446,
+      "step": 37520,
+      "torque_loss": 0.14798277616500854
+    },
+    {
+      "epoch": 33.7410071942446,
+      "force_loss": 0.007989879697561264,
+      "step": 37520
+    },
+    {
+      "epoch": 33.75,
+      "grad_norm": 0.19257792830467224,
+      "learning_rate": 3.36894597639755e-05,
+      "loss": 0.0168,
+      "step": 37530
+    },
+    {
+      "action_loss": 0.007201740052551031,
+      "epoch": 33.75,
+      "step": 37530
+    },
+    {
+      "epoch": 33.75,
+      "step": 37530,
+      "torque_loss": 0.16540561616420746
+    },
+    {
+      "epoch": 33.75,
+      "force_loss": 0.012398049235343933,
+      "step": 37530
+    },
+    {
+      "epoch": 33.7589928057554,
+      "grad_norm": 0.2116713672876358,
+      "learning_rate": 3.366341190699523e-05,
+      "loss": 0.0169,
+      "step": 37540
+    },
+    {
+      "action_loss": 0.0036671289708465338,
+      "epoch": 33.7589928057554,
+      "step": 37540
+    },
+    {
+      "epoch": 33.7589928057554,
+      "step": 37540,
+      "torque_loss": 0.14058664441108704
+    },
+    {
+      "epoch": 33.7589928057554,
+      "force_loss": 0.010138331912457943,
+      "step": 37540
+    },
+    {
+      "epoch": 33.76798561151079,
+      "grad_norm": 0.32120534777641296,
+      "learning_rate": 3.36373690126391e-05,
+      "loss": 0.0168,
+      "step": 37550
+    },
+    {
+      "action_loss": 0.0008958268444985151,
+      "epoch": 33.76798561151079,
+      "step": 37550
+    },
+    {
+      "epoch": 33.76798561151079,
+      "step": 37550,
+      "torque_loss": 0.1516445428133011
+    },
+    {
+      "epoch": 33.76798561151079,
+      "force_loss": 0.0012197787873446941,
+      "step": 37550
+    },
+    {
+      "epoch": 33.776978417266186,
+      "grad_norm": 0.3115787208080292,
+      "learning_rate": 3.3611331088818234e-05,
+      "loss": 0.0181,
+      "step": 37560
+    },
+    {
+      "action_loss": 0.002802715403959155,
+      "epoch": 33.776978417266186,
+      "step": 37560
+    },
+    {
+      "epoch": 33.776978417266186,
+      "step": 37560,
+      "torque_loss": 0.11129637807607651
+    },
+    {
+      "epoch": 33.776978417266186,
+      "force_loss": 0.004637276753783226,
+      "step": 37560
+    },
+    {
+      "epoch": 33.78597122302158,
+      "grad_norm": 0.09075756371021271,
+      "learning_rate": 3.3585298143442265e-05,
+      "loss": 0.0178,
+      "step": 37570
+    },
+    {
+      "action_loss": 0.0008286790107376873,
+      "epoch": 33.78597122302158,
+      "step": 37570
+    },
+    {
+      "epoch": 33.78597122302158,
+      "step": 37570,
+      "torque_loss": 0.14661164581775665
+    },
+    {
+      "epoch": 33.78597122302158,
+      "force_loss": 0.0009299348457716405,
+      "step": 37570
+    },
+    {
+      "epoch": 33.79496402877698,
+      "grad_norm": 0.23820370435714722,
+      "learning_rate": 3.35592701844193e-05,
+      "loss": 0.0186,
+      "step": 37580
+    },
+    {
+      "action_loss": 0.0022285699378699064,
+      "epoch": 33.79496402877698,
+      "step": 37580
+    },
+    {
+      "epoch": 33.79496402877698,
+      "step": 37580,
+      "torque_loss": 0.1465781182050705
+    },
+    {
+      "epoch": 33.79496402877698,
+      "force_loss": 0.0012841963907703757,
+      "step": 37580
+    },
+    {
+      "epoch": 33.80395683453237,
+      "grad_norm": 0.2742636501789093,
+      "learning_rate": 3.353324721965596e-05,
+      "loss": 0.0208,
+      "step": 37590
+    },
+    {
+      "action_loss": 0.0014164196327328682,
+      "epoch": 33.80395683453237,
+      "step": 37590
+    },
+    {
+      "epoch": 33.80395683453237,
+      "step": 37590,
+      "torque_loss": 0.15090857446193695
+    },
+    {
+      "epoch": 33.80395683453237,
+      "force_loss": 0.0011573833180591464,
+      "step": 37590
+    },
+    {
+      "epoch": 33.81294964028777,
+      "grad_norm": 0.39198556542396545,
+      "learning_rate": 3.350722925705736e-05,
+      "loss": 0.0151,
+      "step": 37600
+    },
+    {
+      "action_loss": 0.008726746775209904,
+      "epoch": 33.81294964028777,
+      "step": 37600
+    },
+    {
+      "epoch": 33.81294964028777,
+      "step": 37600,
+      "torque_loss": 0.14815856516361237
+    },
+    {
+      "epoch": 33.81294964028777,
+      "force_loss": 0.007682301104068756,
+      "step": 37600
+    },
+    {
+      "epoch": 33.82194244604317,
+      "grad_norm": 0.15155120193958282,
+      "learning_rate": 3.348121630452703e-05,
+      "loss": 0.0201,
+      "step": 37610
+    },
+    {
+      "action_loss": 0.005731402430683374,
+      "epoch": 33.82194244604317,
+      "step": 37610
+    },
+    {
+      "epoch": 33.82194244604317,
+      "step": 37610,
+      "torque_loss": 0.16444124281406403
+    },
+    {
+      "epoch": 33.82194244604317,
+      "force_loss": 0.007580975536257029,
+      "step": 37610
+    },
+    {
+      "epoch": 33.830935251798564,
+      "grad_norm": 0.1278384029865265,
+      "learning_rate": 3.3455208369967044e-05,
+      "loss": 0.0169,
+      "step": 37620
+    },
+    {
+      "action_loss": 0.0018593333661556244,
+      "epoch": 33.830935251798564,
+      "step": 37620
+    },
+    {
+      "epoch": 33.830935251798564,
+      "step": 37620,
+      "torque_loss": 0.19654791057109833
+    },
+    {
+      "epoch": 33.830935251798564,
+      "force_loss": 0.0014461595565080643,
+      "step": 37620
+    },
+    {
+      "epoch": 33.839928057553955,
+      "grad_norm": 0.22377993166446686,
+      "learning_rate": 3.34292054612779e-05,
+      "loss": 0.0176,
+      "step": 37630
+    },
+    {
+      "action_loss": 0.00334533560089767,
+      "epoch": 33.839928057553955,
+      "step": 37630
+    },
+    {
+      "epoch": 33.839928057553955,
+      "step": 37630,
+      "torque_loss": 0.1333996206521988
+    },
+    {
+      "epoch": 33.839928057553955,
+      "force_loss": 0.0028983645606786013,
+      "step": 37630
+    },
+    {
+      "epoch": 33.84892086330935,
+      "grad_norm": 0.20579461753368378,
+      "learning_rate": 3.340320758635861e-05,
+      "loss": 0.0173,
+      "step": 37640
+    },
+    {
+      "action_loss": 0.0016261772252619267,
+      "epoch": 33.84892086330935,
+      "step": 37640
+    },
+    {
+      "epoch": 33.84892086330935,
+      "step": 37640,
+      "torque_loss": 0.10516867786645889
+    },
+    {
+      "epoch": 33.84892086330935,
+      "force_loss": 0.001366125768981874,
+      "step": 37640
+    },
+    {
+      "epoch": 33.85791366906475,
+      "grad_norm": 0.3685949146747589,
+      "learning_rate": 3.337721475310666e-05,
+      "loss": 0.0168,
+      "step": 37650
+    },
+    {
+      "action_loss": 0.002672509290277958,
+      "epoch": 33.85791366906475,
+      "step": 37650
+    },
+    {
+      "epoch": 33.85791366906475,
+      "step": 37650,
+      "torque_loss": 0.12032388895750046
+    },
+    {
+      "epoch": 33.85791366906475,
+      "force_loss": 0.0033669155091047287,
+      "step": 37650
+    },
+    {
+      "epoch": 33.86690647482014,
+      "grad_norm": 0.15373122692108154,
+      "learning_rate": 3.335122696941795e-05,
+      "loss": 0.0156,
+      "step": 37660
+    },
+    {
+      "action_loss": 0.002857193350791931,
+      "epoch": 33.86690647482014,
+      "step": 37660
+    },
+    {
+      "epoch": 33.86690647482014,
+      "step": 37660,
+      "torque_loss": 0.11864816397428513
+    },
+    {
+      "epoch": 33.86690647482014,
+      "force_loss": 0.0032869998831301928,
+      "step": 37660
+    },
+    {
+      "epoch": 33.87589928057554,
+      "grad_norm": 0.6034438014030457,
+      "learning_rate": 3.332524424318692e-05,
+      "loss": 0.017,
+      "step": 37670
+    },
+    {
+      "action_loss": 0.000992569955997169,
+      "epoch": 33.87589928057554,
+      "step": 37670
+    },
+    {
+      "epoch": 33.87589928057554,
+      "step": 37670,
+      "torque_loss": 0.12354931980371475
+    },
+    {
+      "epoch": 33.87589928057554,
+      "force_loss": 0.002155066467821598,
+      "step": 37670
+    },
+    {
+      "epoch": 33.884892086330936,
+      "grad_norm": 0.18322548270225525,
+      "learning_rate": 3.32992665823064e-05,
+      "loss": 0.016,
+      "step": 37680
+    },
+    {
+      "action_loss": 0.006615945603698492,
+      "epoch": 33.884892086330936,
+      "step": 37680
+    },
+    {
+      "epoch": 33.884892086330936,
+      "step": 37680,
+      "torque_loss": 0.08322913199663162
+    },
+    {
+      "epoch": 33.884892086330936,
+      "force_loss": 0.01546612661331892,
+      "step": 37680
+    },
+    {
+      "epoch": 33.893884892086334,
+      "grad_norm": 0.2370956689119339,
+      "learning_rate": 3.327329399466774e-05,
+      "loss": 0.0187,
+      "step": 37690
+    },
+    {
+      "action_loss": 0.0012947594514116645,
+      "epoch": 33.893884892086334,
+      "step": 37690
+    },
+    {
+      "epoch": 33.893884892086334,
+      "step": 37690,
+      "torque_loss": 0.11286928504705429
+    },
+    {
+      "epoch": 33.893884892086334,
+      "force_loss": 0.001266099396161735,
+      "step": 37690
+    },
+    {
+      "epoch": 33.902877697841724,
+      "grad_norm": 0.3126492500305176,
+      "learning_rate": 3.324732648816072e-05,
+      "loss": 0.017,
+      "step": 37700
+    },
+    {
+      "action_loss": 0.002898208796977997,
+      "epoch": 33.902877697841724,
+      "step": 37700
+    },
+    {
+      "epoch": 33.902877697841724,
+      "step": 37700,
+      "torque_loss": 0.1359972506761551
+    },
+    {
+      "epoch": 33.902877697841724,
+      "force_loss": 0.0015772896585986018,
+      "step": 37700
+    },
+    {
+      "epoch": 33.91187050359712,
+      "grad_norm": 0.29794228076934814,
+      "learning_rate": 3.322136407067358e-05,
+      "loss": 0.0151,
+      "step": 37710
+    },
+    {
+      "action_loss": 0.0062476820312440395,
+      "epoch": 33.91187050359712,
+      "step": 37710
+    },
+    {
+      "epoch": 33.91187050359712,
+      "step": 37710,
+      "torque_loss": 0.13420605659484863
+    },
+    {
+      "epoch": 33.91187050359712,
+      "force_loss": 0.02077285759150982,
+      "step": 37710
+    },
+    {
+      "epoch": 33.92086330935252,
+      "grad_norm": 0.2918078303337097,
+      "learning_rate": 3.3195406750093036e-05,
+      "loss": 0.0168,
+      "step": 37720
+    },
+    {
+      "action_loss": 0.0014831653097644448,
+      "epoch": 33.92086330935252,
+      "step": 37720
+    },
+    {
+      "epoch": 33.92086330935252,
+      "step": 37720,
+      "torque_loss": 0.09238388389348984
+    },
+    {
+      "epoch": 33.92086330935252,
+      "force_loss": 0.0015495987609028816,
+      "step": 37720
+    },
+    {
+      "epoch": 33.92985611510792,
+      "grad_norm": 0.28916600346565247,
+      "learning_rate": 3.3169454534304205e-05,
+      "loss": 0.0161,
+      "step": 37730
+    },
+    {
+      "action_loss": 0.0015696684131398797,
+      "epoch": 33.92985611510792,
+      "step": 37730
+    },
+    {
+      "epoch": 33.92985611510792,
+      "step": 37730,
+      "torque_loss": 0.12014540284872055
+    },
+    {
+      "epoch": 33.92985611510792,
+      "force_loss": 0.0015735877677798271,
+      "step": 37730
+    },
+    {
+      "epoch": 33.93884892086331,
+      "grad_norm": 0.21858280897140503,
+      "learning_rate": 3.3143507431190725e-05,
+      "loss": 0.0168,
+      "step": 37740
+    },
+    {
+      "action_loss": 0.004588219802826643,
+      "epoch": 33.93884892086331,
+      "step": 37740
+    },
+    {
+      "epoch": 33.93884892086331,
+      "step": 37740,
+      "torque_loss": 0.11913526803255081
+    },
+    {
+      "epoch": 33.93884892086331,
+      "force_loss": 0.004644671455025673,
+      "step": 37740
+    },
+    {
+      "epoch": 33.947841726618705,
+      "grad_norm": 0.2118322253227234,
+      "learning_rate": 3.311756544863459e-05,
+      "loss": 0.0164,
+      "step": 37750
+    },
+    {
+      "action_loss": 0.0020320925395935774,
+      "epoch": 33.947841726618705,
+      "step": 37750
+    },
+    {
+      "epoch": 33.947841726618705,
+      "step": 37750,
+      "torque_loss": 0.14661842584609985
+    },
+    {
+      "epoch": 33.947841726618705,
+      "force_loss": 0.003945678006857634,
+      "step": 37750
+    },
+    {
+      "epoch": 33.9568345323741,
+      "grad_norm": 0.1361873596906662,
+      "learning_rate": 3.309162859451633e-05,
+      "loss": 0.0196,
+      "step": 37760
+    },
+    {
+      "action_loss": 0.0033473484218120575,
+      "epoch": 33.9568345323741,
+      "step": 37760
+    },
+    {
+      "epoch": 33.9568345323741,
+      "step": 37760,
+      "torque_loss": 0.06191965565085411
+    },
+    {
+      "epoch": 33.9568345323741,
+      "force_loss": 0.002128748456016183,
+      "step": 37760
+    },
+    {
+      "epoch": 33.96582733812949,
+      "grad_norm": 0.08888915926218033,
+      "learning_rate": 3.306569687671487e-05,
+      "loss": 0.0162,
+      "step": 37770
+    },
+    {
+      "action_loss": 0.00863057654350996,
+      "epoch": 33.96582733812949,
+      "step": 37770
+    },
+    {
+      "epoch": 33.96582733812949,
+      "step": 37770,
+      "torque_loss": 0.15201717615127563
+    },
+    {
+      "epoch": 33.96582733812949,
+      "force_loss": 0.010936747305095196,
+      "step": 37770
+    },
+    {
+      "epoch": 33.97482014388489,
+      "grad_norm": 0.27153369784355164,
+      "learning_rate": 3.303977030310756e-05,
+      "loss": 0.017,
+      "step": 37780
+    },
+    {
+      "action_loss": 0.0035764712374657393,
+      "epoch": 33.97482014388489,
+      "step": 37780
+    },
+    {
+      "epoch": 33.97482014388489,
+      "step": 37780,
+      "torque_loss": 0.12162601947784424
+    },
+    {
+      "epoch": 33.97482014388489,
+      "force_loss": 0.002379960147663951,
+      "step": 37780
+    },
+    {
+      "epoch": 33.98381294964029,
+      "grad_norm": 0.12568314373493195,
+      "learning_rate": 3.3013848881570245e-05,
+      "loss": 0.0155,
+      "step": 37790
+    },
+    {
+      "action_loss": 0.0024474619422107935,
+      "epoch": 33.98381294964029,
+      "step": 37790
+    },
+    {
+      "epoch": 33.98381294964029,
+      "step": 37790,
+      "torque_loss": 0.09655970335006714
+    },
+    {
+      "epoch": 33.98381294964029,
+      "force_loss": 0.0016897293971851468,
+      "step": 37790
+    },
+    {
+      "epoch": 33.992805755395686,
+      "grad_norm": 0.17658117413520813,
+      "learning_rate": 3.298793261997712e-05,
+      "loss": 0.0202,
+      "step": 37800
+    },
+    {
+      "action_loss": 0.0010149498702958226,
+      "epoch": 33.992805755395686,
+      "step": 37800
+    },
+    {
+      "epoch": 33.992805755395686,
+      "step": 37800,
+      "torque_loss": 0.11661273241043091
+    },
+    {
+      "epoch": 33.992805755395686,
+      "force_loss": 0.0007563618128187954,
+      "step": 37800
+    },
+    {
+      "epoch": 34.00179856115108,
+      "grad_norm": 0.12374042719602585,
+      "learning_rate": 3.2962021526200893e-05,
+      "loss": 0.0155,
+      "step": 37810
+    },
+    {
+      "action_loss": 0.00854000449180603,
+      "epoch": 34.00179856115108,
+      "step": 37810
+    },
+    {
+      "epoch": 34.00179856115108,
+      "step": 37810,
+      "torque_loss": 0.16836285591125488
+    },
+    {
+      "epoch": 34.00179856115108,
+      "force_loss": 0.017972860485315323,
+      "step": 37810
+    },
+    {
+      "epoch": 34.010791366906474,
+      "grad_norm": 0.2548501491546631,
+      "learning_rate": 3.293611560811268e-05,
+      "loss": 0.019,
+      "step": 37820
+    },
+    {
+      "action_loss": 0.003282478079199791,
+      "epoch": 34.010791366906474,
+      "step": 37820
+    },
+    {
+      "epoch": 34.010791366906474,
+      "step": 37820,
+      "torque_loss": 0.09992154687643051
+    },
+    {
+      "epoch": 34.010791366906474,
+      "force_loss": 0.003299129195511341,
+      "step": 37820
+    },
+    {
+      "epoch": 34.01978417266187,
+      "grad_norm": 0.2718057930469513,
+      "learning_rate": 3.291021487358199e-05,
+      "loss": 0.0197,
+      "step": 37830
+    },
+    {
+      "action_loss": 0.014899407513439655,
+      "epoch": 34.01978417266187,
+      "step": 37830
+    },
+    {
+      "epoch": 34.01978417266187,
+      "step": 37830,
+      "torque_loss": 0.11702007055282593
+    },
+    {
+      "epoch": 34.01978417266187,
+      "force_loss": 0.03293338045477867,
+      "step": 37830
+    },
+    {
+      "epoch": 34.02877697841727,
+      "grad_norm": 0.3501644730567932,
+      "learning_rate": 3.28843193304768e-05,
+      "loss": 0.0209,
+      "step": 37840
+    },
+    {
+      "action_loss": 0.010363319888710976,
+      "epoch": 34.02877697841727,
+      "step": 37840
+    },
+    {
+      "epoch": 34.02877697841727,
+      "step": 37840,
+      "torque_loss": 0.07841602712869644
+    },
+    {
+      "epoch": 34.02877697841727,
+      "force_loss": 0.009520623832941055,
+      "step": 37840
+    },
+    {
+      "epoch": 34.03776978417266,
+      "grad_norm": 0.17710961401462555,
+      "learning_rate": 3.2858428986663456e-05,
+      "loss": 0.0197,
+      "step": 37850
+    },
+    {
+      "action_loss": 0.003229064168408513,
+      "epoch": 34.03776978417266,
+      "step": 37850
+    },
+    {
+      "epoch": 34.03776978417266,
+      "step": 37850,
+      "torque_loss": 0.08983471244573593
+    },
+    {
+      "epoch": 34.03776978417266,
+      "force_loss": 0.008613090962171555,
+      "step": 37850
+    },
+    {
+      "epoch": 34.04676258992806,
+      "grad_norm": 0.13705997169017792,
+      "learning_rate": 3.283254385000681e-05,
+      "loss": 0.0152,
+      "step": 37860
+    },
+    {
+      "action_loss": 0.008400142192840576,
+      "epoch": 34.04676258992806,
+      "step": 37860
+    },
+    {
+      "epoch": 34.04676258992806,
+      "step": 37860,
+      "torque_loss": 0.14736247062683105
+    },
+    {
+      "epoch": 34.04676258992806,
+      "force_loss": 0.01469422597438097,
+      "step": 37860
+    },
+    {
+      "epoch": 34.055755395683455,
+      "grad_norm": 0.17904780805110931,
+      "learning_rate": 3.2806663928370076e-05,
+      "loss": 0.0179,
+      "step": 37870
+    },
+    {
+      "action_loss": 0.00401114160194993,
+      "epoch": 34.055755395683455,
+      "step": 37870
+    },
+    {
+      "epoch": 34.055755395683455,
+      "step": 37870,
+      "torque_loss": 0.14634637534618378
+    },
+    {
+      "epoch": 34.055755395683455,
+      "force_loss": 0.002393824979662895,
+      "step": 37870
+    },
+    {
+      "epoch": 34.064748201438846,
+      "grad_norm": 0.1617443561553955,
+      "learning_rate": 3.278078922961485e-05,
+      "loss": 0.0163,
+      "step": 37880
+    },
+    {
+      "action_loss": 0.0014823386445641518,
+      "epoch": 34.064748201438846,
+      "step": 37880
+    },
+    {
+      "epoch": 34.064748201438846,
+      "step": 37880,
+      "torque_loss": 0.11770256608724594
+    },
+    {
+      "epoch": 34.064748201438846,
+      "force_loss": 0.003905919613316655,
+      "step": 37880
+    },
+    {
+      "epoch": 34.07374100719424,
+      "grad_norm": 0.09863841533660889,
+      "learning_rate": 3.275491976160123e-05,
+      "loss": 0.0155,
+      "step": 37890
+    },
+    {
+      "action_loss": 0.005528856068849564,
+      "epoch": 34.07374100719424,
+      "step": 37890
+    },
+    {
+      "epoch": 34.07374100719424,
+      "step": 37890,
+      "torque_loss": 0.11923859268426895
+    },
+    {
+      "epoch": 34.07374100719424,
+      "force_loss": 0.00419391505420208,
+      "step": 37890
+    },
+    {
+      "epoch": 34.08273381294964,
+      "grad_norm": 0.18858498334884644,
+      "learning_rate": 3.2729055532187645e-05,
+      "loss": 0.0158,
+      "step": 37900
+    },
+    {
+      "action_loss": 0.005414047744125128,
+      "epoch": 34.08273381294964,
+      "step": 37900
+    },
+    {
+      "epoch": 34.08273381294964,
+      "step": 37900,
+      "torque_loss": 0.21118660271167755
+    },
+    {
+      "epoch": 34.08273381294964,
+      "force_loss": 0.0015715138288214803,
+      "step": 37900
+    },
+    {
+      "epoch": 34.09172661870504,
+      "grad_norm": 0.4234582781791687,
+      "learning_rate": 3.270319654923097e-05,
+      "loss": 0.0199,
+      "step": 37910
+    },
+    {
+      "action_loss": 0.015277204103767872,
+      "epoch": 34.09172661870504,
+      "step": 37910
+    },
+    {
+      "epoch": 34.09172661870504,
+      "step": 37910,
+      "torque_loss": 0.12573844194412231
+    },
+    {
+      "epoch": 34.09172661870504,
+      "force_loss": 0.02044801227748394,
+      "step": 37910
+    },
+    {
+      "epoch": 34.10071942446043,
+      "grad_norm": 0.2833522856235504,
+      "learning_rate": 3.2677342820586506e-05,
+      "loss": 0.0177,
+      "step": 37920
+    },
+    {
+      "action_loss": 0.0018413658253848553,
+      "epoch": 34.10071942446043,
+      "step": 37920
+    },
+    {
+      "epoch": 34.10071942446043,
+      "step": 37920,
+      "torque_loss": 0.10017908364534378
+    },
+    {
+      "epoch": 34.10071942446043,
+      "force_loss": 0.0016164866974577308,
+      "step": 37920
+    },
+    {
+      "epoch": 34.10971223021583,
+      "grad_norm": 0.2937682569026947,
+      "learning_rate": 3.2651494354107905e-05,
+      "loss": 0.0178,
+      "step": 37930
+    },
+    {
+      "action_loss": 0.001016736845485866,
+      "epoch": 34.10971223021583,
+      "step": 37930
+    },
+    {
+      "epoch": 34.10971223021583,
+      "step": 37930,
+      "torque_loss": 0.11007937043905258
+    },
+    {
+      "epoch": 34.10971223021583,
+      "force_loss": 0.0016252034110948443,
+      "step": 37930
+    },
+    {
+      "epoch": 34.118705035971225,
+      "grad_norm": 0.08290261030197144,
+      "learning_rate": 3.2625651157647266e-05,
+      "loss": 0.0141,
+      "step": 37940
+    },
+    {
+      "action_loss": 0.0014338326873257756,
+      "epoch": 34.118705035971225,
+      "step": 37940
+    },
+    {
+      "epoch": 34.118705035971225,
+      "step": 37940,
+      "torque_loss": 0.1151399090886116
+    },
+    {
+      "epoch": 34.118705035971225,
+      "force_loss": 0.0014972379431128502,
+      "step": 37940
+    },
+    {
+      "epoch": 34.12769784172662,
+      "grad_norm": 0.19066950678825378,
+      "learning_rate": 3.259981323905505e-05,
+      "loss": 0.0163,
+      "step": 37950
+    },
+    {
+      "action_loss": 0.0012854525120928884,
+      "epoch": 34.12769784172662,
+      "step": 37950
+    },
+    {
+      "epoch": 34.12769784172662,
+      "step": 37950,
+      "torque_loss": 0.10462678223848343
+    },
+    {
+      "epoch": 34.12769784172662,
+      "force_loss": 0.00450166454538703,
+      "step": 37950
+    },
+    {
+      "epoch": 34.13669064748201,
+      "grad_norm": 0.1492910236120224,
+      "learning_rate": 3.257398060618014e-05,
+      "loss": 0.0161,
+      "step": 37960
+    },
+    {
+      "action_loss": 0.0031411610543727875,
+      "epoch": 34.13669064748201,
+      "step": 37960
+    },
+    {
+      "epoch": 34.13669064748201,
+      "step": 37960,
+      "torque_loss": 0.11208828538656235
+    },
+    {
+      "epoch": 34.13669064748201,
+      "force_loss": 0.002766592660918832,
+      "step": 37960
+    },
+    {
+      "epoch": 34.14568345323741,
+      "grad_norm": 0.1473616361618042,
+      "learning_rate": 3.254815326686983e-05,
+      "loss": 0.0147,
+      "step": 37970
+    },
+    {
+      "action_loss": 0.0011684807250276208,
+      "epoch": 34.14568345323741,
+      "step": 37970
+    },
+    {
+      "epoch": 34.14568345323741,
+      "step": 37970,
+      "torque_loss": 0.0926390290260315
+    },
+    {
+      "epoch": 34.14568345323741,
+      "force_loss": 0.0008684794302098453,
+      "step": 37970
+    },
+    {
+      "epoch": 34.15467625899281,
+      "grad_norm": 0.14073944091796875,
+      "learning_rate": 3.2522331228969774e-05,
+      "loss": 0.0138,
+      "step": 37980
+    },
+    {
+      "action_loss": 0.0021744014229625463,
+      "epoch": 34.15467625899281,
+      "step": 37980
+    },
+    {
+      "epoch": 34.15467625899281,
+      "step": 37980,
+      "torque_loss": 0.12354063987731934
+    },
+    {
+      "epoch": 34.15467625899281,
+      "force_loss": 0.007990377955138683,
+      "step": 37980
+    },
+    {
+      "epoch": 34.1636690647482,
+      "grad_norm": 0.1969049721956253,
+      "learning_rate": 3.2496514500324006e-05,
+      "loss": 0.016,
+      "step": 37990
+    },
+    {
+      "action_loss": 0.0008678183075971901,
+      "epoch": 34.1636690647482,
+      "step": 37990
+    },
+    {
+      "epoch": 34.1636690647482,
+      "step": 37990,
+      "torque_loss": 0.10964993387460709
+    },
+    {
+      "epoch": 34.1636690647482,
+      "force_loss": 0.0005905279540456831,
+      "step": 37990
+    },
+    {
+      "epoch": 34.172661870503596,
+      "grad_norm": 0.09668637067079544,
+      "learning_rate": 3.247070308877498e-05,
+      "loss": 0.0138,
+      "step": 38000
+    },
+    {
+      "action_loss": 0.009854384697973728,
+      "epoch": 34.172661870503596,
+      "step": 38000
+    },
+    {
+      "epoch": 34.172661870503596,
+      "step": 38000,
+      "torque_loss": 0.1738036870956421
+    },
+    {
+      "epoch": 34.172661870503596,
+      "force_loss": 0.024181565269827843,
+      "step": 38000
+    },
+    {
+      "epoch": 34.181654676258994,
+      "grad_norm": 0.1287219077348709,
+      "learning_rate": 3.2444897002163515e-05,
+      "loss": 0.0164,
+      "step": 38010
+    },
+    {
+      "action_loss": 0.0009754174388945103,
+      "epoch": 34.181654676258994,
+      "step": 38010
+    },
+    {
+      "epoch": 34.181654676258994,
+      "step": 38010,
+      "torque_loss": 0.07954701781272888
+    },
+    {
+      "epoch": 34.181654676258994,
+      "force_loss": 0.006283354014158249,
+      "step": 38010
+    },
+    {
+      "epoch": 34.19064748201439,
+      "grad_norm": 0.21000033617019653,
+      "learning_rate": 3.241909624832885e-05,
+      "loss": 0.0149,
+      "step": 38020
+    },
+    {
+      "action_loss": 0.002194909146055579,
+      "epoch": 34.19064748201439,
+      "step": 38020
+    },
+    {
+      "epoch": 34.19064748201439,
+      "step": 38020,
+      "torque_loss": 0.08859354257583618
+    },
+    {
+      "epoch": 34.19064748201439,
+      "force_loss": 0.0030823368579149246,
+      "step": 38020
+    },
+    {
+      "epoch": 34.19964028776978,
+      "grad_norm": 0.34158578515052795,
+      "learning_rate": 3.239330083510852e-05,
+      "loss": 0.0185,
+      "step": 38030
+    },
+    {
+      "action_loss": 0.010723414830863476,
+      "epoch": 34.19964028776978,
+      "step": 38030
+    },
+    {
+      "epoch": 34.19964028776978,
+      "step": 38030,
+      "torque_loss": 0.10285502672195435
+    },
+    {
+      "epoch": 34.19964028776978,
+      "force_loss": 0.021538028493523598,
+      "step": 38030
+    },
+    {
+      "epoch": 34.20863309352518,
+      "grad_norm": 0.20425674319267273,
+      "learning_rate": 3.236751077033855e-05,
+      "loss": 0.0187,
+      "step": 38040
+    },
+    {
+      "action_loss": 0.0016714368248358369,
+      "epoch": 34.20863309352518,
+      "step": 38040
+    },
+    {
+      "epoch": 34.20863309352518,
+      "step": 38040,
+      "torque_loss": 0.1403820663690567
+    },
+    {
+      "epoch": 34.20863309352518,
+      "force_loss": 0.0013357937568798661,
+      "step": 38040
+    },
+    {
+      "epoch": 34.21762589928058,
+      "grad_norm": 0.6160386800765991,
+      "learning_rate": 3.234172606185322e-05,
+      "loss": 0.0207,
+      "step": 38050
+    },
+    {
+      "action_loss": 0.0022034465800970793,
+      "epoch": 34.21762589928058,
+      "step": 38050
+    },
+    {
+      "epoch": 34.21762589928058,
+      "step": 38050,
+      "torque_loss": 0.16391442716121674
+    },
+    {
+      "epoch": 34.21762589928058,
+      "force_loss": 0.003691088641062379,
+      "step": 38050
+    },
+    {
+      "epoch": 34.226618705035975,
+      "grad_norm": 0.42136967182159424,
+      "learning_rate": 3.231594671748528e-05,
+      "loss": 0.0183,
+      "step": 38060
+    },
+    {
+      "action_loss": 0.0058631435967981815,
+      "epoch": 34.226618705035975,
+      "step": 38060
+    },
+    {
+      "epoch": 34.226618705035975,
+      "step": 38060,
+      "torque_loss": 0.17233215272426605
+    },
+    {
+      "epoch": 34.226618705035975,
+      "force_loss": 0.0033209084067493677,
+      "step": 38060
+    },
+    {
+      "epoch": 34.235611510791365,
+      "grad_norm": 0.1460287868976593,
+      "learning_rate": 3.2290172745065815e-05,
+      "loss": 0.0162,
+      "step": 38070
+    },
+    {
+      "action_loss": 0.00554612884297967,
+      "epoch": 34.235611510791365,
+      "step": 38070
+    },
+    {
+      "epoch": 34.235611510791365,
+      "step": 38070,
+      "torque_loss": 0.11947265267372131
+    },
+    {
+      "epoch": 34.235611510791365,
+      "force_loss": 0.012024116702377796,
+      "step": 38070
+    },
+    {
+      "epoch": 34.24460431654676,
+      "grad_norm": 0.31487590074539185,
+      "learning_rate": 3.226440415242426e-05,
+      "loss": 0.0162,
+      "step": 38080
+    },
+    {
+      "action_loss": 0.0022396717686206102,
+      "epoch": 34.24460431654676,
+      "step": 38080
+    },
+    {
+      "epoch": 34.24460431654676,
+      "step": 38080,
+      "torque_loss": 0.15118329226970673
+    },
+    {
+      "epoch": 34.24460431654676,
+      "force_loss": 0.0028562508523464203,
+      "step": 38080
+    },
+    {
+      "epoch": 34.25359712230216,
+      "grad_norm": 0.896686851978302,
+      "learning_rate": 3.223864094738846e-05,
+      "loss": 0.0175,
+      "step": 38090
+    },
+    {
+      "action_loss": 0.0029583973810076714,
+      "epoch": 34.25359712230216,
+      "step": 38090
+    },
+    {
+      "epoch": 34.25359712230216,
+      "step": 38090,
+      "torque_loss": 0.16267897188663483
+    },
+    {
+      "epoch": 34.25359712230216,
+      "force_loss": 0.0029362214263528585,
+      "step": 38090
+    },
+    {
+      "epoch": 34.26258992805755,
+      "grad_norm": 0.13580796122550964,
+      "learning_rate": 3.221288313778456e-05,
+      "loss": 0.0151,
+      "step": 38100
+    },
+    {
+      "action_loss": 0.0009770820615813136,
+      "epoch": 34.26258992805755,
+      "step": 38100
+    },
+    {
+      "epoch": 34.26258992805755,
+      "step": 38100,
+      "torque_loss": 0.08829469233751297
+    },
+    {
+      "epoch": 34.26258992805755,
+      "force_loss": 0.0013094767928123474,
+      "step": 38100
+    },
+    {
+      "epoch": 34.27158273381295,
+      "grad_norm": 0.09662114083766937,
+      "learning_rate": 3.2187130731437125e-05,
+      "loss": 0.0142,
+      "step": 38110
+    },
+    {
+      "action_loss": 0.008527331985533237,
+      "epoch": 34.27158273381295,
+      "step": 38110
+    },
+    {
+      "epoch": 34.27158273381295,
+      "step": 38110,
+      "torque_loss": 0.11693272739648819
+    },
+    {
+      "epoch": 34.27158273381295,
+      "force_loss": 0.010731442831456661,
+      "step": 38110
+    },
+    {
+      "epoch": 34.280575539568346,
+      "grad_norm": 0.11308939754962921,
+      "learning_rate": 3.216138373616905e-05,
+      "loss": 0.0172,
+      "step": 38120
+    },
+    {
+      "action_loss": 0.0016659159446135163,
+      "epoch": 34.280575539568346,
+      "step": 38120
+    },
+    {
+      "epoch": 34.280575539568346,
+      "step": 38120,
+      "torque_loss": 0.1553521752357483
+    },
+    {
+      "epoch": 34.280575539568346,
+      "force_loss": 0.005824326071888208,
+      "step": 38120
+    },
+    {
+      "epoch": 34.289568345323744,
+      "grad_norm": 0.10985405743122101,
+      "learning_rate": 3.21356421598016e-05,
+      "loss": 0.0169,
+      "step": 38130
+    },
+    {
+      "action_loss": 0.002218943787738681,
+      "epoch": 34.289568345323744,
+      "step": 38130
+    },
+    {
+      "epoch": 34.289568345323744,
+      "step": 38130,
+      "torque_loss": 0.11694708466529846
+    },
+    {
+      "epoch": 34.289568345323744,
+      "force_loss": 0.003227776614949107,
+      "step": 38130
+    },
+    {
+      "epoch": 34.298561151079134,
+      "grad_norm": 0.309419184923172,
+      "learning_rate": 3.210990601015438e-05,
+      "loss": 0.0167,
+      "step": 38140
+    },
+    {
+      "action_loss": 0.001612002495676279,
+      "epoch": 34.298561151079134,
+      "step": 38140
+    },
+    {
+      "epoch": 34.298561151079134,
+      "step": 38140,
+      "torque_loss": 0.09101120382547379
+    },
+    {
+      "epoch": 34.298561151079134,
+      "force_loss": 0.005857498850673437,
+      "step": 38140
+    },
+    {
+      "epoch": 34.30755395683453,
+      "grad_norm": 0.08458050340414047,
+      "learning_rate": 3.208417529504535e-05,
+      "loss": 0.016,
+      "step": 38150
+    },
+    {
+      "action_loss": 0.004689517430961132,
+      "epoch": 34.30755395683453,
+      "step": 38150
+    },
+    {
+      "epoch": 34.30755395683453,
+      "step": 38150,
+      "torque_loss": 0.14398176968097687
+    },
+    {
+      "epoch": 34.30755395683453,
+      "force_loss": 0.009167391806840897,
+      "step": 38150
+    },
+    {
+      "epoch": 34.31654676258993,
+      "grad_norm": 0.1201728954911232,
+      "learning_rate": 3.205845002229084e-05,
+      "loss": 0.0179,
+      "step": 38160
+    },
+    {
+      "action_loss": 0.002028467832133174,
+      "epoch": 34.31654676258993,
+      "step": 38160
+    },
+    {
+      "epoch": 34.31654676258993,
+      "step": 38160,
+      "torque_loss": 0.12843093276023865
+    },
+    {
+      "epoch": 34.31654676258993,
+      "force_loss": 0.003033468732610345,
+      "step": 38160
+    },
+    {
+      "epoch": 34.32553956834532,
+      "grad_norm": 0.23771627247333527,
+      "learning_rate": 3.203273019970547e-05,
+      "loss": 0.0158,
+      "step": 38170
+    },
+    {
+      "action_loss": 0.01165491621941328,
+      "epoch": 34.32553956834532,
+      "step": 38170
+    },
+    {
+      "epoch": 34.32553956834532,
+      "step": 38170,
+      "torque_loss": 0.10926521569490433
+    },
+    {
+      "epoch": 34.32553956834532,
+      "force_loss": 0.01667793281376362,
+      "step": 38170
+    },
+    {
+      "epoch": 34.33453237410072,
+      "grad_norm": 0.38787347078323364,
+      "learning_rate": 3.200701583510227e-05,
+      "loss": 0.0171,
+      "step": 38180
+    },
+    {
+      "action_loss": 0.0010124507825821638,
+      "epoch": 34.33453237410072,
+      "step": 38180
+    },
+    {
+      "epoch": 34.33453237410072,
+      "step": 38180,
+      "torque_loss": 0.1000855341553688
+    },
+    {
+      "epoch": 34.33453237410072,
+      "force_loss": 0.003066321136429906,
+      "step": 38180
+    },
+    {
+      "epoch": 34.343525179856115,
+      "grad_norm": 0.15007776021957397,
+      "learning_rate": 3.198130693629261e-05,
+      "loss": 0.0169,
+      "step": 38190
+    },
+    {
+      "action_loss": 0.0010897573083639145,
+      "epoch": 34.343525179856115,
+      "step": 38190
+    },
+    {
+      "epoch": 34.343525179856115,
+      "step": 38190,
+      "torque_loss": 0.12579868733882904
+    },
+    {
+      "epoch": 34.343525179856115,
+      "force_loss": 0.001298569142818451,
+      "step": 38190
+    },
+    {
+      "epoch": 34.35251798561151,
+      "grad_norm": 0.33359494805336,
+      "learning_rate": 3.195560351108612e-05,
+      "loss": 0.0175,
+      "step": 38200
+    },
+    {
+      "action_loss": 0.0017828574636951089,
+      "epoch": 34.35251798561151,
+      "step": 38200
+    },
+    {
+      "epoch": 34.35251798561151,
+      "step": 38200,
+      "torque_loss": 0.13059596717357635
+    },
+    {
+      "epoch": 34.35251798561151,
+      "force_loss": 0.0023853592574596405,
+      "step": 38200
+    },
+    {
+      "epoch": 34.361510791366904,
+      "grad_norm": 0.125379741191864,
+      "learning_rate": 3.1929905567290865e-05,
+      "loss": 0.0157,
+      "step": 38210
+    },
+    {
+      "action_loss": 0.009774972684681416,
+      "epoch": 34.361510791366904,
+      "step": 38210
+    },
+    {
+      "epoch": 34.361510791366904,
+      "step": 38210,
+      "torque_loss": 0.11848794668912888
+    },
+    {
+      "epoch": 34.361510791366904,
+      "force_loss": 0.016782671213150024,
+      "step": 38210
+    },
+    {
+      "epoch": 34.3705035971223,
+      "grad_norm": 0.22163516283035278,
+      "learning_rate": 3.1904213112713164e-05,
+      "loss": 0.0178,
+      "step": 38220
+    },
+    {
+      "action_loss": 0.0012675829930230975,
+      "epoch": 34.3705035971223,
+      "step": 38220
+    },
+    {
+      "epoch": 34.3705035971223,
+      "step": 38220,
+      "torque_loss": 0.11763477325439453
+    },
+    {
+      "epoch": 34.3705035971223,
+      "force_loss": 0.002317859558388591,
+      "step": 38220
+    },
+    {
+      "epoch": 34.3794964028777,
+      "grad_norm": 0.11981499195098877,
+      "learning_rate": 3.187852615515774e-05,
+      "loss": 0.018,
+      "step": 38230
+    },
+    {
+      "action_loss": 0.0013185711577534676,
+      "epoch": 34.3794964028777,
+      "step": 38230
+    },
+    {
+      "epoch": 34.3794964028777,
+      "step": 38230,
+      "torque_loss": 0.10611226409673691
+    },
+    {
+      "epoch": 34.3794964028777,
+      "force_loss": 0.0019390900852158666,
+      "step": 38230
+    },
+    {
+      "epoch": 34.388489208633096,
+      "grad_norm": 0.1892155259847641,
+      "learning_rate": 3.1852844702427606e-05,
+      "loss": 0.0157,
+      "step": 38240
+    },
+    {
+      "action_loss": 0.001037661568261683,
+      "epoch": 34.388489208633096,
+      "step": 38240
+    },
+    {
+      "epoch": 34.388489208633096,
+      "step": 38240,
+      "torque_loss": 0.10556551069021225
+    },
+    {
+      "epoch": 34.388489208633096,
+      "force_loss": 0.0007493208977393806,
+      "step": 38240
+    },
+    {
+      "epoch": 34.39748201438849,
+      "grad_norm": 0.1841997504234314,
+      "learning_rate": 3.18271687623241e-05,
+      "loss": 0.0175,
+      "step": 38250
+    },
+    {
+      "action_loss": 0.0012132179690524936,
+      "epoch": 34.39748201438849,
+      "step": 38250
+    },
+    {
+      "epoch": 34.39748201438849,
+      "step": 38250,
+      "torque_loss": 0.12356892973184586
+    },
+    {
+      "epoch": 34.39748201438849,
+      "force_loss": 0.0009360490366816521,
+      "step": 38250
+    },
+    {
+      "epoch": 34.406474820143885,
+      "grad_norm": 0.24160510301589966,
+      "learning_rate": 3.1801498342646896e-05,
+      "loss": 0.0173,
+      "step": 38260
+    },
+    {
+      "action_loss": 0.001780887134373188,
+      "epoch": 34.406474820143885,
+      "step": 38260
+    },
+    {
+      "epoch": 34.406474820143885,
+      "step": 38260,
+      "torque_loss": 0.10922298580408096
+    },
+    {
+      "epoch": 34.406474820143885,
+      "force_loss": 0.0032669713255017996,
+      "step": 38260
+    },
+    {
+      "epoch": 34.41546762589928,
+      "grad_norm": 0.3818244934082031,
+      "learning_rate": 3.177583345119398e-05,
+      "loss": 0.0179,
+      "step": 38270
+    },
+    {
+      "action_loss": 0.0012223607627674937,
+      "epoch": 34.41546762589928,
+      "step": 38270
+    },
+    {
+      "epoch": 34.41546762589928,
+      "step": 38270,
+      "torque_loss": 0.1331145316362381
+    },
+    {
+      "epoch": 34.41546762589928,
+      "force_loss": 0.0014526421437039971,
+      "step": 38270
+    },
+    {
+      "epoch": 34.42446043165468,
+      "grad_norm": 0.2520965039730072,
+      "learning_rate": 3.17501740957617e-05,
+      "loss": 0.0154,
+      "step": 38280
+    },
+    {
+      "action_loss": 0.002908057300373912,
+      "epoch": 34.42446043165468,
+      "step": 38280
+    },
+    {
+      "epoch": 34.42446043165468,
+      "step": 38280,
+      "torque_loss": 0.1257518231868744
+    },
+    {
+      "epoch": 34.42446043165468,
+      "force_loss": 0.003900927258655429,
+      "step": 38280
+    },
+    {
+      "epoch": 34.43345323741007,
+      "grad_norm": 0.14366383850574493,
+      "learning_rate": 3.172452028414467e-05,
+      "loss": 0.0173,
+      "step": 38290
+    },
+    {
+      "action_loss": 0.002626248402521014,
+      "epoch": 34.43345323741007,
+      "step": 38290
+    },
+    {
+      "epoch": 34.43345323741007,
+      "step": 38290,
+      "torque_loss": 0.1364048719406128
+    },
+    {
+      "epoch": 34.43345323741007,
+      "force_loss": 0.004100541118532419,
+      "step": 38290
+    },
+    {
+      "epoch": 34.44244604316547,
+      "grad_norm": 0.20616415143013,
+      "learning_rate": 3.169887202413583e-05,
+      "loss": 0.0181,
+      "step": 38300
+    },
+    {
+      "action_loss": 0.011716465465724468,
+      "epoch": 34.44244604316547,
+      "step": 38300
+    },
+    {
+      "epoch": 34.44244604316547,
+      "step": 38300,
+      "torque_loss": 0.12543442845344543
+    },
+    {
+      "epoch": 34.44244604316547,
+      "force_loss": 0.016384921967983246,
+      "step": 38300
+    },
+    {
+      "epoch": 34.451438848920866,
+      "grad_norm": 0.2401425689458847,
+      "learning_rate": 3.167322932352646e-05,
+      "loss": 0.0178,
+      "step": 38310
+    },
+    {
+      "action_loss": 0.0014225519262254238,
+      "epoch": 34.451438848920866,
+      "step": 38310
+    },
+    {
+      "epoch": 34.451438848920866,
+      "step": 38310,
+      "torque_loss": 0.09976992756128311
+    },
+    {
+      "epoch": 34.451438848920866,
+      "force_loss": 0.0016255304217338562,
+      "step": 38310
+    },
+    {
+      "epoch": 34.460431654676256,
+      "grad_norm": 0.11276914179325104,
+      "learning_rate": 3.164759219010613e-05,
+      "loss": 0.0183,
+      "step": 38320
+    },
+    {
+      "action_loss": 0.002597688464447856,
+      "epoch": 34.460431654676256,
+      "step": 38320
+    },
+    {
+      "epoch": 34.460431654676256,
+      "step": 38320,
+      "torque_loss": 0.12056461721658707
+    },
+    {
+      "epoch": 34.460431654676256,
+      "force_loss": 0.002494994318112731,
+      "step": 38320
+    },
+    {
+      "epoch": 34.469424460431654,
+      "grad_norm": 0.2172938734292984,
+      "learning_rate": 3.1621960631662725e-05,
+      "loss": 0.0168,
+      "step": 38330
+    },
+    {
+      "action_loss": 0.0031617535278201103,
+      "epoch": 34.469424460431654,
+      "step": 38330
+    },
+    {
+      "epoch": 34.469424460431654,
+      "step": 38330,
+      "torque_loss": 0.13513052463531494
+    },
+    {
+      "epoch": 34.469424460431654,
+      "force_loss": 0.004643291234970093,
+      "step": 38330
+    },
+    {
+      "epoch": 34.47841726618705,
+      "grad_norm": 0.29820922017097473,
+      "learning_rate": 3.159633465598245e-05,
+      "loss": 0.0228,
+      "step": 38340
+    },
+    {
+      "action_loss": 0.0019786886405199766,
+      "epoch": 34.47841726618705,
+      "step": 38340
+    },
+    {
+      "epoch": 34.47841726618705,
+      "step": 38340,
+      "torque_loss": 0.16076822578907013
+    },
+    {
+      "epoch": 34.47841726618705,
+      "force_loss": 0.0020407189149409533,
+      "step": 38340
+    },
+    {
+      "epoch": 34.48741007194245,
+      "grad_norm": 0.2596455216407776,
+      "learning_rate": 3.1570714270849767e-05,
+      "loss": 0.0159,
+      "step": 38350
+    },
+    {
+      "action_loss": 0.0019989933352917433,
+      "epoch": 34.48741007194245,
+      "step": 38350
+    },
+    {
+      "epoch": 34.48741007194245,
+      "step": 38350,
+      "torque_loss": 0.16867582499980927
+    },
+    {
+      "epoch": 34.48741007194245,
+      "force_loss": 0.0039395857602357864,
+      "step": 38350
+    },
+    {
+      "epoch": 34.49640287769784,
+      "grad_norm": 0.20967614650726318,
+      "learning_rate": 3.1545099484047516e-05,
+      "loss": 0.0163,
+      "step": 38360
+    },
+    {
+      "action_loss": 0.003609751584008336,
+      "epoch": 34.49640287769784,
+      "step": 38360
+    },
+    {
+      "epoch": 34.49640287769784,
+      "step": 38360,
+      "torque_loss": 0.09196087718009949
+    },
+    {
+      "epoch": 34.49640287769784,
+      "force_loss": 0.00652661407366395,
+      "step": 38360
+    },
+    {
+      "epoch": 34.50539568345324,
+      "grad_norm": 0.30926424264907837,
+      "learning_rate": 3.151949030335674e-05,
+      "loss": 0.015,
+      "step": 38370
+    },
+    {
+      "action_loss": 0.0023019700311124325,
+      "epoch": 34.50539568345324,
+      "step": 38370
+    },
+    {
+      "epoch": 34.50539568345324,
+      "step": 38370,
+      "torque_loss": 0.1672786921262741
+    },
+    {
+      "epoch": 34.50539568345324,
+      "force_loss": 0.003482109634205699,
+      "step": 38370
+    },
+    {
+      "epoch": 34.514388489208635,
+      "grad_norm": 0.3725399076938629,
+      "learning_rate": 3.149388673655687e-05,
+      "loss": 0.018,
+      "step": 38380
+    },
+    {
+      "action_loss": 0.0032252792734652758,
+      "epoch": 34.514388489208635,
+      "step": 38380
+    },
+    {
+      "epoch": 34.514388489208635,
+      "step": 38380,
+      "torque_loss": 0.18382048606872559
+    },
+    {
+      "epoch": 34.514388489208635,
+      "force_loss": 0.009460299275815487,
+      "step": 38380
+    },
+    {
+      "epoch": 34.523381294964025,
+      "grad_norm": 0.22372767329216003,
+      "learning_rate": 3.146828879142559e-05,
+      "loss": 0.0188,
+      "step": 38390
+    },
+    {
+      "action_loss": 0.0013847621157765388,
+      "epoch": 34.523381294964025,
+      "step": 38390
+    },
+    {
+      "epoch": 34.523381294964025,
+      "step": 38390,
+      "torque_loss": 0.11002520471811295
+    },
+    {
+      "epoch": 34.523381294964025,
+      "force_loss": 0.0018376881489530206,
+      "step": 38390
+    },
+    {
+      "epoch": 34.53237410071942,
+      "grad_norm": 0.07798569649457932,
+      "learning_rate": 3.1442696475738866e-05,
+      "loss": 0.0146,
+      "step": 38400
+    },
+    {
+      "action_loss": 0.001473208307288587,
+      "epoch": 34.53237410071942,
+      "step": 38400
+    },
+    {
+      "epoch": 34.53237410071942,
+      "step": 38400,
+      "torque_loss": 0.11487217992544174
+    },
+    {
+      "epoch": 34.53237410071942,
+      "force_loss": 0.000947553024161607,
+      "step": 38400
+    },
+    {
+      "epoch": 34.54136690647482,
+      "grad_norm": 0.18024161458015442,
+      "learning_rate": 3.141710979727098e-05,
+      "loss": 0.0151,
+      "step": 38410
+    },
+    {
+      "action_loss": 0.0012842240976169705,
+      "epoch": 34.54136690647482,
+      "step": 38410
+    },
+    {
+      "epoch": 34.54136690647482,
+      "step": 38410,
+      "torque_loss": 0.14827051758766174
+    },
+    {
+      "epoch": 34.54136690647482,
+      "force_loss": 0.0008141950820572674,
+      "step": 38410
+    },
+    {
+      "epoch": 34.55035971223022,
+      "grad_norm": 0.14035546779632568,
+      "learning_rate": 3.139152876379447e-05,
+      "loss": 0.0164,
+      "step": 38420
+    },
+    {
+      "action_loss": 0.0026300896424800158,
+      "epoch": 34.55035971223022,
+      "step": 38420
+    },
+    {
+      "epoch": 34.55035971223022,
+      "step": 38420,
+      "torque_loss": 0.10661149024963379
+    },
+    {
+      "epoch": 34.55035971223022,
+      "force_loss": 0.0066487849690020084,
+      "step": 38420
+    },
+    {
+      "epoch": 34.55935251798561,
+      "grad_norm": 0.2546481490135193,
+      "learning_rate": 3.1365953383080214e-05,
+      "loss": 0.0182,
+      "step": 38430
+    },
+    {
+      "action_loss": 0.0016450881958007812,
+      "epoch": 34.55935251798561,
+      "step": 38430
+    },
+    {
+      "epoch": 34.55935251798561,
+      "step": 38430,
+      "torque_loss": 0.08318792283535004
+    },
+    {
+      "epoch": 34.55935251798561,
+      "force_loss": 0.0018876157701015472,
+      "step": 38430
+    },
+    {
+      "epoch": 34.568345323741006,
+      "grad_norm": 0.1375221461057663,
+      "learning_rate": 3.134038366289731e-05,
+      "loss": 0.0151,
+      "step": 38440
+    },
+    {
+      "action_loss": 0.0013049558037891984,
+      "epoch": 34.568345323741006,
+      "step": 38440
+    },
+    {
+      "epoch": 34.568345323741006,
+      "step": 38440,
+      "torque_loss": 0.0981418564915657
+    },
+    {
+      "epoch": 34.568345323741006,
+      "force_loss": 0.0027469974011182785,
+      "step": 38440
+    },
+    {
+      "epoch": 34.577338129496404,
+      "grad_norm": 0.24912779033184052,
+      "learning_rate": 3.131481961101317e-05,
+      "loss": 0.019,
+      "step": 38450
+    },
+    {
+      "action_loss": 0.0014133708318695426,
+      "epoch": 34.577338129496404,
+      "step": 38450
+    },
+    {
+      "epoch": 34.577338129496404,
+      "step": 38450,
+      "torque_loss": 0.16744482517242432
+    },
+    {
+      "epoch": 34.577338129496404,
+      "force_loss": 0.002211084822192788,
+      "step": 38450
+    },
+    {
+      "epoch": 34.5863309352518,
+      "grad_norm": 0.15376850962638855,
+      "learning_rate": 3.128926123519349e-05,
+      "loss": 0.0182,
+      "step": 38460
+    },
+    {
+      "action_loss": 0.0012096124701201916,
+      "epoch": 34.5863309352518,
+      "step": 38460
+    },
+    {
+      "epoch": 34.5863309352518,
+      "step": 38460,
+      "torque_loss": 0.13167576491832733
+    },
+    {
+      "epoch": 34.5863309352518,
+      "force_loss": 0.002058493671938777,
+      "step": 38460
+    },
+    {
+      "epoch": 34.59532374100719,
+      "grad_norm": 0.10142739862203598,
+      "learning_rate": 3.1263708543202194e-05,
+      "loss": 0.0135,
+      "step": 38470
+    },
+    {
+      "action_loss": 0.006791311781853437,
+      "epoch": 34.59532374100719,
+      "step": 38470
+    },
+    {
+      "epoch": 34.59532374100719,
+      "step": 38470,
+      "torque_loss": 0.12551186978816986
+    },
+    {
+      "epoch": 34.59532374100719,
+      "force_loss": 0.0074044340290129185,
+      "step": 38470
+    },
+    {
+      "epoch": 34.60431654676259,
+      "grad_norm": 0.17957811057567596,
+      "learning_rate": 3.123816154280155e-05,
+      "loss": 0.0202,
+      "step": 38480
+    },
+    {
+      "action_loss": 0.0025393737014383078,
+      "epoch": 34.60431654676259,
+      "step": 38480
+    },
+    {
+      "epoch": 34.60431654676259,
+      "step": 38480,
+      "torque_loss": 0.14315222203731537
+    },
+    {
+      "epoch": 34.60431654676259,
+      "force_loss": 0.002506350865587592,
+      "step": 38480
+    },
+    {
+      "epoch": 34.61330935251799,
+      "grad_norm": 0.17599982023239136,
+      "learning_rate": 3.121262024175207e-05,
+      "loss": 0.0177,
+      "step": 38490
+    },
+    {
+      "action_loss": 0.005657270550727844,
+      "epoch": 34.61330935251799,
+      "step": 38490
+    },
+    {
+      "epoch": 34.61330935251799,
+      "step": 38490,
+      "torque_loss": 0.10008245706558228
+    },
+    {
+      "epoch": 34.61330935251799,
+      "force_loss": 0.006468428764492273,
+      "step": 38490
+    },
+    {
+      "epoch": 34.62230215827338,
+      "grad_norm": 0.27707439661026,
+      "learning_rate": 3.118708464781248e-05,
+      "loss": 0.0166,
+      "step": 38500
+    },
+    {
+      "action_loss": 0.0019723237492144108,
+      "epoch": 34.62230215827338,
+      "step": 38500
+    },
+    {
+      "epoch": 34.62230215827338,
+      "step": 38500,
+      "torque_loss": 0.12207069993019104
+    },
+    {
+      "epoch": 34.62230215827338,
+      "force_loss": 0.0031738094985485077,
+      "step": 38500
+    },
+    {
+      "epoch": 34.631294964028775,
+      "grad_norm": 0.3856644034385681,
+      "learning_rate": 3.116155476873987e-05,
+      "loss": 0.0174,
+      "step": 38510
+    },
+    {
+      "action_loss": 0.0012843600707128644,
+      "epoch": 34.631294964028775,
+      "step": 38510
+    },
+    {
+      "epoch": 34.631294964028775,
+      "step": 38510,
+      "torque_loss": 0.10053249448537827
+    },
+    {
+      "epoch": 34.631294964028775,
+      "force_loss": 0.0038113368209451437,
+      "step": 38510
+    },
+    {
+      "epoch": 34.64028776978417,
+      "grad_norm": 0.10629621893167496,
+      "learning_rate": 3.11360306122895e-05,
+      "loss": 0.0167,
+      "step": 38520
+    },
+    {
+      "action_loss": 0.0037077355664223433,
+      "epoch": 34.64028776978417,
+      "step": 38520
+    },
+    {
+      "epoch": 34.64028776978417,
+      "step": 38520,
+      "torque_loss": 0.14237217605113983
+    },
+    {
+      "epoch": 34.64028776978417,
+      "force_loss": 0.007574303541332483,
+      "step": 38520
+    },
+    {
+      "epoch": 34.64928057553957,
+      "grad_norm": 0.21006788313388824,
+      "learning_rate": 3.1110512186214975e-05,
+      "loss": 0.017,
+      "step": 38530
+    },
+    {
+      "action_loss": 0.0016152107855305076,
+      "epoch": 34.64928057553957,
+      "step": 38530
+    },
+    {
+      "epoch": 34.64928057553957,
+      "step": 38530,
+      "torque_loss": 0.1222589835524559
+    },
+    {
+      "epoch": 34.64928057553957,
+      "force_loss": 0.0008252934203483164,
+      "step": 38530
+    },
+    {
+      "epoch": 34.65827338129496,
+      "grad_norm": 0.24139325320720673,
+      "learning_rate": 3.1084999498268095e-05,
+      "loss": 0.0183,
+      "step": 38540
+    },
+    {
+      "action_loss": 0.0011581552680581808,
+      "epoch": 34.65827338129496,
+      "step": 38540
+    },
+    {
+      "epoch": 34.65827338129496,
+      "step": 38540,
+      "torque_loss": 0.1371845006942749
+    },
+    {
+      "epoch": 34.65827338129496,
+      "force_loss": 0.0009283709223382175,
+      "step": 38540
+    },
+    {
+      "epoch": 34.66726618705036,
+      "grad_norm": 0.2640725076198578,
+      "learning_rate": 3.1059492556198934e-05,
+      "loss": 0.0168,
+      "step": 38550
+    },
+    {
+      "action_loss": 0.018619609996676445,
+      "epoch": 34.66726618705036,
+      "step": 38550
+    },
+    {
+      "epoch": 34.66726618705036,
+      "step": 38550,
+      "torque_loss": 0.1586941033601761
+    },
+    {
+      "epoch": 34.66726618705036,
+      "force_loss": 0.019095636904239655,
+      "step": 38550
+    },
+    {
+      "epoch": 34.67625899280576,
+      "grad_norm": 0.39664411544799805,
+      "learning_rate": 3.103399136775586e-05,
+      "loss": 0.0183,
+      "step": 38560
+    },
+    {
+      "action_loss": 0.007044801488518715,
+      "epoch": 34.67625899280576,
+      "step": 38560
+    },
+    {
+      "epoch": 34.67625899280576,
+      "step": 38560,
+      "torque_loss": 0.14300976693630219
+    },
+    {
+      "epoch": 34.67625899280576,
+      "force_loss": 0.0095159150660038,
+      "step": 38560
+    },
+    {
+      "epoch": 34.685251798561154,
+      "grad_norm": 0.4605175256729126,
+      "learning_rate": 3.100849594068541e-05,
+      "loss": 0.0194,
+      "step": 38570
+    },
+    {
+      "action_loss": 0.003579928306862712,
+      "epoch": 34.685251798561154,
+      "step": 38570
+    },
+    {
+      "epoch": 34.685251798561154,
+      "step": 38570,
+      "torque_loss": 0.11204127222299576
+    },
+    {
+      "epoch": 34.685251798561154,
+      "force_loss": 0.004976348951458931,
+      "step": 38570
+    },
+    {
+      "epoch": 34.694244604316545,
+      "grad_norm": 0.2618870437145233,
+      "learning_rate": 3.0983006282732484e-05,
+      "loss": 0.0172,
+      "step": 38580
+    },
+    {
+      "action_loss": 0.005463303532451391,
+      "epoch": 34.694244604316545,
+      "step": 38580
+    },
+    {
+      "epoch": 34.694244604316545,
+      "step": 38580,
+      "torque_loss": 0.1450069695711136
+    },
+    {
+      "epoch": 34.694244604316545,
+      "force_loss": 0.004828445613384247,
+      "step": 38580
+    },
+    {
+      "epoch": 34.70323741007194,
+      "grad_norm": 0.200755313038826,
+      "learning_rate": 3.0957522401640116e-05,
+      "loss": 0.0179,
+      "step": 38590
+    },
+    {
+      "action_loss": 0.0009448258206248283,
+      "epoch": 34.70323741007194,
+      "step": 38590
+    },
+    {
+      "epoch": 34.70323741007194,
+      "step": 38590,
+      "torque_loss": 0.11735858768224716
+    },
+    {
+      "epoch": 34.70323741007194,
+      "force_loss": 0.0005795633187517524,
+      "step": 38590
+    },
+    {
+      "epoch": 34.71223021582734,
+      "grad_norm": 0.10308893769979477,
+      "learning_rate": 3.0932044305149645e-05,
+      "loss": 0.0156,
+      "step": 38600
+    },
+    {
+      "action_loss": 0.00312026753090322,
+      "epoch": 34.71223021582734,
+      "step": 38600
+    },
+    {
+      "epoch": 34.71223021582734,
+      "step": 38600,
+      "torque_loss": 0.1819179505109787
+    },
+    {
+      "epoch": 34.71223021582734,
+      "force_loss": 0.0033644726499915123,
+      "step": 38600
+    },
+    {
+      "epoch": 34.72122302158273,
+      "grad_norm": 0.28257471323013306,
+      "learning_rate": 3.090657200100068e-05,
+      "loss": 0.0161,
+      "step": 38610
+    },
+    {
+      "action_loss": 0.005166333634406328,
+      "epoch": 34.72122302158273,
+      "step": 38610
+    },
+    {
+      "epoch": 34.72122302158273,
+      "step": 38610,
+      "torque_loss": 0.15472187101840973
+    },
+    {
+      "epoch": 34.72122302158273,
+      "force_loss": 0.0045619686134159565,
+      "step": 38610
+    },
+    {
+      "epoch": 34.73021582733813,
+      "grad_norm": 0.10393625497817993,
+      "learning_rate": 3.088110549693099e-05,
+      "loss": 0.0166,
+      "step": 38620
+    },
+    {
+      "action_loss": 0.003556174226105213,
+      "epoch": 34.73021582733813,
+      "step": 38620
+    },
+    {
+      "epoch": 34.73021582733813,
+      "step": 38620,
+      "torque_loss": 0.14934581518173218
+    },
+    {
+      "epoch": 34.73021582733813,
+      "force_loss": 0.004354837816208601,
+      "step": 38620
+    },
+    {
+      "epoch": 34.739208633093526,
+      "grad_norm": 0.17032960057258606,
+      "learning_rate": 3.085564480067667e-05,
+      "loss": 0.0163,
+      "step": 38630
+    },
+    {
+      "action_loss": 0.0020976655650883913,
+      "epoch": 34.739208633093526,
+      "step": 38630
+    },
+    {
+      "epoch": 34.739208633093526,
+      "step": 38630,
+      "torque_loss": 0.13000796735286713
+    },
+    {
+      "epoch": 34.739208633093526,
+      "force_loss": 0.0012645056704059243,
+      "step": 38630
+    },
+    {
+      "epoch": 34.74820143884892,
+      "grad_norm": 0.0839029997587204,
+      "learning_rate": 3.0830189919971955e-05,
+      "loss": 0.0147,
+      "step": 38640
+    },
+    {
+      "action_loss": 0.0012977635487914085,
+      "epoch": 34.74820143884892,
+      "step": 38640
+    },
+    {
+      "epoch": 34.74820143884892,
+      "step": 38640,
+      "torque_loss": 0.12684597074985504
+    },
+    {
+      "epoch": 34.74820143884892,
+      "force_loss": 0.004534391686320305,
+      "step": 38640
+    },
+    {
+      "epoch": 34.757194244604314,
+      "grad_norm": 0.4317879378795624,
+      "learning_rate": 3.080474086254939e-05,
+      "loss": 0.0168,
+      "step": 38650
+    },
+    {
+      "action_loss": 0.0022805833723396063,
+      "epoch": 34.757194244604314,
+      "step": 38650
+    },
+    {
+      "epoch": 34.757194244604314,
+      "step": 38650,
+      "torque_loss": 0.10338791459798813
+    },
+    {
+      "epoch": 34.757194244604314,
+      "force_loss": 0.0035688721109181643,
+      "step": 38650
+    },
+    {
+      "epoch": 34.76618705035971,
+      "grad_norm": 0.11154210567474365,
+      "learning_rate": 3.077929763613975e-05,
+      "loss": 0.0162,
+      "step": 38660
+    },
+    {
+      "action_loss": 0.003112196223810315,
+      "epoch": 34.76618705035971,
+      "step": 38660
+    },
+    {
+      "epoch": 34.76618705035971,
+      "step": 38660,
+      "torque_loss": 0.13109798729419708
+    },
+    {
+      "epoch": 34.76618705035971,
+      "force_loss": 0.004263064358383417,
+      "step": 38660
+    },
+    {
+      "epoch": 34.77517985611511,
+      "grad_norm": 0.10955115407705307,
+      "learning_rate": 3.075386024847198e-05,
+      "loss": 0.0188,
+      "step": 38670
+    },
+    {
+      "action_loss": 0.00954432599246502,
+      "epoch": 34.77517985611511,
+      "step": 38670
+    },
+    {
+      "epoch": 34.77517985611511,
+      "step": 38670,
+      "torque_loss": 0.12226337194442749
+    },
+    {
+      "epoch": 34.77517985611511,
+      "force_loss": 0.013184207491576672,
+      "step": 38670
+    },
+    {
+      "epoch": 34.78417266187051,
+      "grad_norm": 0.12276234477758408,
+      "learning_rate": 3.072842870727331e-05,
+      "loss": 0.0166,
+      "step": 38680
+    },
+    {
+      "action_loss": 0.0012173851719126105,
+      "epoch": 34.78417266187051,
+      "step": 38680
+    },
+    {
+      "epoch": 34.78417266187051,
+      "step": 38680,
+      "torque_loss": 0.1090126633644104
+    },
+    {
+      "epoch": 34.78417266187051,
+      "force_loss": 0.001330745406448841,
+      "step": 38680
+    },
+    {
+      "epoch": 34.7931654676259,
+      "grad_norm": 0.2330036610364914,
+      "learning_rate": 3.070300302026916e-05,
+      "loss": 0.015,
+      "step": 38690
+    },
+    {
+      "action_loss": 0.00153315847273916,
+      "epoch": 34.7931654676259,
+      "step": 38690
+    },
+    {
+      "epoch": 34.7931654676259,
+      "step": 38690,
+      "torque_loss": 0.11496663093566895
+    },
+    {
+      "epoch": 34.7931654676259,
+      "force_loss": 0.0012268581194803119,
+      "step": 38690
+    },
+    {
+      "epoch": 34.802158273381295,
+      "grad_norm": 0.3656274974346161,
+      "learning_rate": 3.067758319518318e-05,
+      "loss": 0.0203,
+      "step": 38700
+    },
+    {
+      "action_loss": 0.002117444993928075,
+      "epoch": 34.802158273381295,
+      "step": 38700
+    },
+    {
+      "epoch": 34.802158273381295,
+      "step": 38700,
+      "torque_loss": 0.07872550934553146
+    },
+    {
+      "epoch": 34.802158273381295,
+      "force_loss": 0.0019782257732003927,
+      "step": 38700
+    },
+    {
+      "epoch": 34.81115107913669,
+      "grad_norm": 0.17873837053775787,
+      "learning_rate": 3.065216923973725e-05,
+      "loss": 0.0146,
+      "step": 38710
+    },
+    {
+      "action_loss": 0.0021198915783315897,
+      "epoch": 34.81115107913669,
+      "step": 38710
+    },
+    {
+      "epoch": 34.81115107913669,
+      "step": 38710,
+      "torque_loss": 0.14273861050605774
+    },
+    {
+      "epoch": 34.81115107913669,
+      "force_loss": 0.007720018271356821,
+      "step": 38710
+    },
+    {
+      "epoch": 34.82014388489208,
+      "grad_norm": 0.19084160029888153,
+      "learning_rate": 3.062676116165145e-05,
+      "loss": 0.0161,
+      "step": 38720
+    },
+    {
+      "action_loss": 0.003642566502094269,
+      "epoch": 34.82014388489208,
+      "step": 38720
+    },
+    {
+      "epoch": 34.82014388489208,
+      "step": 38720,
+      "torque_loss": 0.1195937916636467
+    },
+    {
+      "epoch": 34.82014388489208,
+      "force_loss": 0.008990968577563763,
+      "step": 38720
+    },
+    {
+      "epoch": 34.82913669064748,
+      "grad_norm": 0.3843153715133667,
+      "learning_rate": 3.06013589686441e-05,
+      "loss": 0.0165,
+      "step": 38730
+    },
+    {
+      "action_loss": 0.0009784154826775193,
+      "epoch": 34.82913669064748,
+      "step": 38730
+    },
+    {
+      "epoch": 34.82913669064748,
+      "step": 38730,
+      "torque_loss": 0.08457992225885391
+    },
+    {
+      "epoch": 34.82913669064748,
+      "force_loss": 0.001979834632948041,
+      "step": 38730
+    },
+    {
+      "epoch": 34.83812949640288,
+      "grad_norm": 0.39605802297592163,
+      "learning_rate": 3.05759626684317e-05,
+      "loss": 0.0179,
+      "step": 38740
+    },
+    {
+      "action_loss": 0.0043804761953651905,
+      "epoch": 34.83812949640288,
+      "step": 38740
+    },
+    {
+      "epoch": 34.83812949640288,
+      "step": 38740,
+      "torque_loss": 0.14209119975566864
+    },
+    {
+      "epoch": 34.83812949640288,
+      "force_loss": 0.005905414465814829,
+      "step": 38740
+    },
+    {
+      "epoch": 34.847122302158276,
+      "grad_norm": 0.284212201833725,
+      "learning_rate": 3.055057226872896e-05,
+      "loss": 0.0164,
+      "step": 38750
+    },
+    {
+      "action_loss": 0.004344114568084478,
+      "epoch": 34.847122302158276,
+      "step": 38750
+    },
+    {
+      "epoch": 34.847122302158276,
+      "step": 38750,
+      "torque_loss": 0.06664899736642838
+    },
+    {
+      "epoch": 34.847122302158276,
+      "force_loss": 0.003713706275448203,
+      "step": 38750
+    },
+    {
+      "epoch": 34.856115107913666,
+      "grad_norm": 0.1520061492919922,
+      "learning_rate": 3.052518777724887e-05,
+      "loss": 0.0161,
+      "step": 38760
+    },
+    {
+      "action_loss": 0.004284176509827375,
+      "epoch": 34.856115107913666,
+      "step": 38760
+    },
+    {
+      "epoch": 34.856115107913666,
+      "step": 38760,
+      "torque_loss": 0.13820405304431915
+    },
+    {
+      "epoch": 34.856115107913666,
+      "force_loss": 0.001813484006561339,
+      "step": 38760
+    },
+    {
+      "epoch": 34.865107913669064,
+      "grad_norm": 0.1246003657579422,
+      "learning_rate": 3.04998092017025e-05,
+      "loss": 0.016,
+      "step": 38770
+    },
+    {
+      "action_loss": 0.009232689626514912,
+      "epoch": 34.865107913669064,
+      "step": 38770
+    },
+    {
+      "epoch": 34.865107913669064,
+      "step": 38770,
+      "torque_loss": 0.18753080070018768
+    },
+    {
+      "epoch": 34.865107913669064,
+      "force_loss": 0.00506631750613451,
+      "step": 38770
+    },
+    {
+      "epoch": 34.87410071942446,
+      "grad_norm": 0.4100170433521271,
+      "learning_rate": 3.0474436549799246e-05,
+      "loss": 0.02,
+      "step": 38780
+    },
+    {
+      "action_loss": 0.0022725574672222137,
+      "epoch": 34.87410071942446,
+      "step": 38780
+    },
+    {
+      "epoch": 34.87410071942446,
+      "step": 38780,
+      "torque_loss": 0.18606919050216675
+    },
+    {
+      "epoch": 34.87410071942446,
+      "force_loss": 0.002096575917676091,
+      "step": 38780
+    },
+    {
+      "epoch": 34.88309352517986,
+      "grad_norm": 0.3019394278526306,
+      "learning_rate": 3.044906982924661e-05,
+      "loss": 0.0177,
+      "step": 38790
+    },
+    {
+      "action_loss": 0.0009394208900630474,
+      "epoch": 34.88309352517986,
+      "step": 38790
+    },
+    {
+      "epoch": 34.88309352517986,
+      "step": 38790,
+      "torque_loss": 0.10499713569879532
+    },
+    {
+      "epoch": 34.88309352517986,
+      "force_loss": 0.0010895855957642198,
+      "step": 38790
+    },
+    {
+      "epoch": 34.89208633093525,
+      "grad_norm": 0.23774023354053497,
+      "learning_rate": 3.0423709047750337e-05,
+      "loss": 0.0155,
+      "step": 38800
+    },
+    {
+      "action_loss": 0.003401062684133649,
+      "epoch": 34.89208633093525,
+      "step": 38800
+    },
+    {
+      "epoch": 34.89208633093525,
+      "step": 38800,
+      "torque_loss": 0.08398023992776871
+    },
+    {
+      "epoch": 34.89208633093525,
+      "force_loss": 0.005365829914808273,
+      "step": 38800
+    },
+    {
+      "epoch": 34.90107913669065,
+      "grad_norm": 0.14938440918922424,
+      "learning_rate": 3.03983542130144e-05,
+      "loss": 0.0168,
+      "step": 38810
+    },
+    {
+      "action_loss": 0.0028705305885523558,
+      "epoch": 34.90107913669065,
+      "step": 38810
+    },
+    {
+      "epoch": 34.90107913669065,
+      "step": 38810,
+      "torque_loss": 0.08145985007286072
+    },
+    {
+      "epoch": 34.90107913669065,
+      "force_loss": 0.003623089985921979,
+      "step": 38810
+    },
+    {
+      "epoch": 34.910071942446045,
+      "grad_norm": 0.3361145853996277,
+      "learning_rate": 3.0373005332740877e-05,
+      "loss": 0.0207,
+      "step": 38820
+    },
+    {
+      "action_loss": 0.006614338140934706,
+      "epoch": 34.910071942446045,
+      "step": 38820
+    },
+    {
+      "epoch": 34.910071942446045,
+      "step": 38820,
+      "torque_loss": 0.114862360060215
+    },
+    {
+      "epoch": 34.910071942446045,
+      "force_loss": 0.00877420324832201,
+      "step": 38820
+    },
+    {
+      "epoch": 34.919064748201436,
+      "grad_norm": 0.19874995946884155,
+      "learning_rate": 3.034766241463013e-05,
+      "loss": 0.0165,
+      "step": 38830
+    },
+    {
+      "action_loss": 0.001063765143044293,
+      "epoch": 34.919064748201436,
+      "step": 38830
+    },
+    {
+      "epoch": 34.919064748201436,
+      "step": 38830,
+      "torque_loss": 0.11082597821950912
+    },
+    {
+      "epoch": 34.919064748201436,
+      "force_loss": 0.0019464752404019237,
+      "step": 38830
+    },
+    {
+      "epoch": 34.92805755395683,
+      "grad_norm": 0.35252317786216736,
+      "learning_rate": 3.032232546638064e-05,
+      "loss": 0.0163,
+      "step": 38840
+    },
+    {
+      "action_loss": 0.004800503142178059,
+      "epoch": 34.92805755395683,
+      "step": 38840
+    },
+    {
+      "epoch": 34.92805755395683,
+      "step": 38840,
+      "torque_loss": 0.11431223899126053
+    },
+    {
+      "epoch": 34.92805755395683,
+      "force_loss": 0.004633750766515732,
+      "step": 38840
+    },
+    {
+      "epoch": 34.93705035971223,
+      "grad_norm": 0.36122679710388184,
+      "learning_rate": 3.0296994495689114e-05,
+      "loss": 0.0163,
+      "step": 38850
+    },
+    {
+      "action_loss": 0.007631280925124884,
+      "epoch": 34.93705035971223,
+      "step": 38850
+    },
+    {
+      "epoch": 34.93705035971223,
+      "step": 38850,
+      "torque_loss": 0.13991303741931915
+    },
+    {
+      "epoch": 34.93705035971223,
+      "force_loss": 0.003343901364132762,
+      "step": 38850
+    },
+    {
+      "epoch": 34.94604316546763,
+      "grad_norm": 0.13534283638000488,
+      "learning_rate": 3.0271669510250444e-05,
+      "loss": 0.0166,
+      "step": 38860
+    },
+    {
+      "action_loss": 0.002593817189335823,
+      "epoch": 34.94604316546763,
+      "step": 38860
+    },
+    {
+      "epoch": 34.94604316546763,
+      "step": 38860,
+      "torque_loss": 0.0812850072979927
+    },
+    {
+      "epoch": 34.94604316546763,
+      "force_loss": 0.0023991798516362906,
+      "step": 38860
+    },
+    {
+      "epoch": 34.95503597122302,
+      "grad_norm": 0.19206443428993225,
+      "learning_rate": 3.024635051775766e-05,
+      "loss": 0.0167,
+      "step": 38870
+    },
+    {
+      "action_loss": 0.0023118644021451473,
+      "epoch": 34.95503597122302,
+      "step": 38870
+    },
+    {
+      "epoch": 34.95503597122302,
+      "step": 38870,
+      "torque_loss": 0.1302252858877182
+    },
+    {
+      "epoch": 34.95503597122302,
+      "force_loss": 0.0034537147730588913,
+      "step": 38870
+    },
+    {
+      "epoch": 34.96402877697842,
+      "grad_norm": 0.10967405885457993,
+      "learning_rate": 3.022103752590205e-05,
+      "loss": 0.0196,
+      "step": 38880
+    },
+    {
+      "action_loss": 0.002898571314290166,
+      "epoch": 34.96402877697842,
+      "step": 38880
+    },
+    {
+      "epoch": 34.96402877697842,
+      "step": 38880,
+      "torque_loss": 0.1515694558620453
+    },
+    {
+      "epoch": 34.96402877697842,
+      "force_loss": 0.00570349395275116,
+      "step": 38880
+    },
+    {
+      "epoch": 34.973021582733814,
+      "grad_norm": 0.3706991672515869,
+      "learning_rate": 3.0195730542372992e-05,
+      "loss": 0.0189,
+      "step": 38890
+    },
+    {
+      "action_loss": 0.00310687068849802,
+      "epoch": 34.973021582733814,
+      "step": 38890
+    },
+    {
+      "epoch": 34.973021582733814,
+      "step": 38890,
+      "torque_loss": 0.21740974485874176
+    },
+    {
+      "epoch": 34.973021582733814,
+      "force_loss": 0.004394970368593931,
+      "step": 38890
+    },
+    {
+      "epoch": 34.98201438848921,
+      "grad_norm": 0.32958418130874634,
+      "learning_rate": 3.0170429574858084e-05,
+      "loss": 0.0168,
+      "step": 38900
+    },
+    {
+      "action_loss": 0.002011655131354928,
+      "epoch": 34.98201438848921,
+      "step": 38900
+    },
+    {
+      "epoch": 34.98201438848921,
+      "step": 38900,
+      "torque_loss": 0.09887155145406723
+    },
+    {
+      "epoch": 34.98201438848921,
+      "force_loss": 0.006040406879037619,
+      "step": 38900
+    },
+    {
+      "epoch": 34.9910071942446,
+      "grad_norm": 0.09772446751594543,
+      "learning_rate": 3.0145134631043127e-05,
+      "loss": 0.0149,
+      "step": 38910
+    },
+    {
+      "action_loss": 0.0020546845626085997,
+      "epoch": 34.9910071942446,
+      "step": 38910
+    },
+    {
+      "epoch": 34.9910071942446,
+      "step": 38910,
+      "torque_loss": 0.14445792138576508
+    },
+    {
+      "epoch": 34.9910071942446,
+      "force_loss": 0.0013244961155578494,
+      "step": 38910
+    },
+    {
+      "epoch": 35.0,
+      "grad_norm": 0.11525759845972061,
+      "learning_rate": 3.0119845718612018e-05,
+      "loss": 0.016,
+      "step": 38920
+    },
+    {
+      "action_loss": 0.002357466844841838,
+      "epoch": 35.0,
+      "step": 38920
+    },
+    {
+      "epoch": 35.0,
+      "step": 38920,
+      "torque_loss": 0.15787339210510254
+    },
+    {
+      "epoch": 35.0,
+      "force_loss": 0.002151248510926962,
+      "step": 38920
+    },
+    {
+      "epoch": 35.0089928057554,
+      "grad_norm": 0.5724800825119019,
+      "learning_rate": 3.009456284524688e-05,
+      "loss": 0.0204,
+      "step": 38930
+    },
+    {
+      "action_loss": 0.008969360962510109,
+      "epoch": 35.0089928057554,
+      "step": 38930
+    },
+    {
+      "epoch": 35.0089928057554,
+      "step": 38930,
+      "torque_loss": 0.11319664865732193
+    },
+    {
+      "epoch": 35.0089928057554,
+      "force_loss": 0.012574061751365662,
+      "step": 38930
+    },
+    {
+      "epoch": 35.01798561151079,
+      "grad_norm": 0.11487343907356262,
+      "learning_rate": 3.0069286018627967e-05,
+      "loss": 0.0181,
+      "step": 38940
+    },
+    {
+      "action_loss": 0.0036686931271106005,
+      "epoch": 35.01798561151079,
+      "step": 38940
+    },
+    {
+      "epoch": 35.01798561151079,
+      "step": 38940,
+      "torque_loss": 0.13395781815052032
+    },
+    {
+      "epoch": 35.01798561151079,
+      "force_loss": 0.013992962427437305,
+      "step": 38940
+    },
+    {
+      "epoch": 35.026978417266186,
+      "grad_norm": 0.3809404671192169,
+      "learning_rate": 3.0044015246433743e-05,
+      "loss": 0.021,
+      "step": 38950
+    },
+    {
+      "action_loss": 0.013758336193859577,
+      "epoch": 35.026978417266186,
+      "step": 38950
+    },
+    {
+      "epoch": 35.026978417266186,
+      "step": 38950,
+      "torque_loss": 0.14249445497989655
+    },
+    {
+      "epoch": 35.026978417266186,
+      "force_loss": 0.01452005747705698,
+      "step": 38950
+    },
+    {
+      "epoch": 35.03597122302158,
+      "grad_norm": 0.12666234374046326,
+      "learning_rate": 3.0018750536340755e-05,
+      "loss": 0.0201,
+      "step": 38960
+    },
+    {
+      "action_loss": 0.03742603957653046,
+      "epoch": 35.03597122302158,
+      "step": 38960
+    },
+    {
+      "epoch": 35.03597122302158,
+      "step": 38960,
+      "torque_loss": 0.22465144097805023
+    },
+    {
+      "epoch": 35.03597122302158,
+      "force_loss": 0.02060875855386257,
+      "step": 38960
+    },
+    {
+      "epoch": 35.04496402877698,
+      "grad_norm": 0.16837553679943085,
+      "learning_rate": 2.999349189602378e-05,
+      "loss": 0.0251,
+      "step": 38970
+    },
+    {
+      "action_loss": 0.0012964889174327254,
+      "epoch": 35.04496402877698,
+      "step": 38970
+    },
+    {
+      "epoch": 35.04496402877698,
+      "step": 38970,
+      "torque_loss": 0.10754083842039108
+    },
+    {
+      "epoch": 35.04496402877698,
+      "force_loss": 0.0011754450388252735,
+      "step": 38970
+    },
+    {
+      "epoch": 35.05395683453237,
+      "grad_norm": 0.24246415495872498,
+      "learning_rate": 2.9968239333155733e-05,
+      "loss": 0.0189,
+      "step": 38980
+    },
+    {
+      "action_loss": 0.002997683361172676,
+      "epoch": 35.05395683453237,
+      "step": 38980
+    },
+    {
+      "epoch": 35.05395683453237,
+      "step": 38980,
+      "torque_loss": 0.14538516104221344
+    },
+    {
+      "epoch": 35.05395683453237,
+      "force_loss": 0.006573410239070654,
+      "step": 38980
+    },
+    {
+      "epoch": 35.06294964028777,
+      "grad_norm": 0.11037357151508331,
+      "learning_rate": 2.994299285540767e-05,
+      "loss": 0.0151,
+      "step": 38990
+    },
+    {
+      "action_loss": 0.0011510731419548392,
+      "epoch": 35.06294964028777,
+      "step": 38990
+    },
+    {
+      "epoch": 35.06294964028777,
+      "step": 38990,
+      "torque_loss": 0.10932561010122299
+    },
+    {
+      "epoch": 35.06294964028777,
+      "force_loss": 0.0018023927696049213,
+      "step": 38990
+    },
+    {
+      "epoch": 35.07194244604317,
+      "grad_norm": 0.2527114748954773,
+      "learning_rate": 2.9917752470448813e-05,
+      "loss": 0.017,
+      "step": 39000
+    },
+    {
+      "action_loss": 0.0013888214016333222,
+      "epoch": 35.07194244604317,
+      "step": 39000
+    },
+    {
+      "epoch": 35.07194244604317,
+      "step": 39000,
+      "torque_loss": 0.09995990991592407
+    },
+    {
+      "epoch": 35.07194244604317,
+      "force_loss": 0.0023165512830018997,
+      "step": 39000
+    },
+    {
+      "epoch": 35.080935251798564,
+      "grad_norm": 0.13498377799987793,
+      "learning_rate": 2.9892518185946495e-05,
+      "loss": 0.0172,
+      "step": 39010
+    },
+    {
+      "action_loss": 0.004496482200920582,
+      "epoch": 35.080935251798564,
+      "step": 39010
+    },
+    {
+      "epoch": 35.080935251798564,
+      "step": 39010,
+      "torque_loss": 0.12354707717895508
+    },
+    {
+      "epoch": 35.080935251798564,
+      "force_loss": 0.0062436000443995,
+      "step": 39010
+    },
+    {
+      "epoch": 35.089928057553955,
+      "grad_norm": 0.10693199932575226,
+      "learning_rate": 2.986729000956624e-05,
+      "loss": 0.0215,
+      "step": 39020
+    },
+    {
+      "action_loss": 0.0035669002681970596,
+      "epoch": 35.089928057553955,
+      "step": 39020
+    },
+    {
+      "epoch": 35.089928057553955,
+      "step": 39020,
+      "torque_loss": 0.14371724426746368
+    },
+    {
+      "epoch": 35.089928057553955,
+      "force_loss": 0.010771863162517548,
+      "step": 39020
+    },
+    {
+      "epoch": 35.09892086330935,
+      "grad_norm": 0.2423935979604721,
+      "learning_rate": 2.9842067948971736e-05,
+      "loss": 0.0174,
+      "step": 39030
+    },
+    {
+      "action_loss": 0.0022365062031894922,
+      "epoch": 35.09892086330935,
+      "step": 39030
+    },
+    {
+      "epoch": 35.09892086330935,
+      "step": 39030,
+      "torque_loss": 0.10635358095169067
+    },
+    {
+      "epoch": 35.09892086330935,
+      "force_loss": 0.0015481151640415192,
+      "step": 39030
+    },
+    {
+      "epoch": 35.10791366906475,
+      "grad_norm": 0.1286407858133316,
+      "learning_rate": 2.9816852011824727e-05,
+      "loss": 0.0147,
+      "step": 39040
+    },
+    {
+      "action_loss": 0.002132033696398139,
+      "epoch": 35.10791366906475,
+      "step": 39040
+    },
+    {
+      "epoch": 35.10791366906475,
+      "step": 39040,
+      "torque_loss": 0.09622766822576523
+    },
+    {
+      "epoch": 35.10791366906475,
+      "force_loss": 0.001295006019063294,
+      "step": 39040
+    },
+    {
+      "epoch": 35.11690647482014,
+      "grad_norm": 0.08749210089445114,
+      "learning_rate": 2.979164220578519e-05,
+      "loss": 0.0165,
+      "step": 39050
+    },
+    {
+      "action_loss": 0.01790345273911953,
+      "epoch": 35.11690647482014,
+      "step": 39050
+    },
+    {
+      "epoch": 35.11690647482014,
+      "step": 39050,
+      "torque_loss": 0.11699700355529785
+    },
+    {
+      "epoch": 35.11690647482014,
+      "force_loss": 0.02455134503543377,
+      "step": 39050
+    },
+    {
+      "epoch": 35.12589928057554,
+      "grad_norm": 0.19494834542274475,
+      "learning_rate": 2.9766438538511165e-05,
+      "loss": 0.017,
+      "step": 39060
+    },
+    {
+      "action_loss": 0.002067138673737645,
+      "epoch": 35.12589928057554,
+      "step": 39060
+    },
+    {
+      "epoch": 35.12589928057554,
+      "step": 39060,
+      "torque_loss": 0.10579248517751694
+    },
+    {
+      "epoch": 35.12589928057554,
+      "force_loss": 0.004045093897730112,
+      "step": 39060
+    },
+    {
+      "epoch": 35.134892086330936,
+      "grad_norm": 0.31693965196609497,
+      "learning_rate": 2.9741241017658873e-05,
+      "loss": 0.0162,
+      "step": 39070
+    },
+    {
+      "action_loss": 0.003981070127338171,
+      "epoch": 35.134892086330936,
+      "step": 39070
+    },
+    {
+      "epoch": 35.134892086330936,
+      "step": 39070,
+      "torque_loss": 0.1272837519645691
+    },
+    {
+      "epoch": 35.134892086330936,
+      "force_loss": 0.00482361251488328,
+      "step": 39070
+    },
+    {
+      "epoch": 35.143884892086334,
+      "grad_norm": 0.3081154525279999,
+      "learning_rate": 2.971604965088267e-05,
+      "loss": 0.0163,
+      "step": 39080
+    },
+    {
+      "action_loss": 0.0040407381020486355,
+      "epoch": 35.143884892086334,
+      "step": 39080
+    },
+    {
+      "epoch": 35.143884892086334,
+      "step": 39080,
+      "torque_loss": 0.12722699344158173
+    },
+    {
+      "epoch": 35.143884892086334,
+      "force_loss": 0.00501455832272768,
+      "step": 39080
+    },
+    {
+      "epoch": 35.152877697841724,
+      "grad_norm": 0.17428624629974365,
+      "learning_rate": 2.9690864445835008e-05,
+      "loss": 0.0167,
+      "step": 39090
+    },
+    {
+      "action_loss": 0.004586245398968458,
+      "epoch": 35.152877697841724,
+      "step": 39090
+    },
+    {
+      "epoch": 35.152877697841724,
+      "step": 39090,
+      "torque_loss": 0.1103062704205513
+    },
+    {
+      "epoch": 35.152877697841724,
+      "force_loss": 0.010136381722986698,
+      "step": 39090
+    },
+    {
+      "epoch": 35.16187050359712,
+      "grad_norm": 0.2923608422279358,
+      "learning_rate": 2.966568541016651e-05,
+      "loss": 0.0255,
+      "step": 39100
+    },
+    {
+      "action_loss": 0.0013203794369474053,
+      "epoch": 35.16187050359712,
+      "step": 39100
+    },
+    {
+      "epoch": 35.16187050359712,
+      "step": 39100,
+      "torque_loss": 0.14165626466274261
+    },
+    {
+      "epoch": 35.16187050359712,
+      "force_loss": 0.0013694696826860309,
+      "step": 39100
+    },
+    {
+      "epoch": 35.17086330935252,
+      "grad_norm": 0.25528061389923096,
+      "learning_rate": 2.9640512551525867e-05,
+      "loss": 0.0201,
+      "step": 39110
+    },
+    {
+      "action_loss": 0.0019578454084694386,
+      "epoch": 35.17086330935252,
+      "step": 39110
+    },
+    {
+      "epoch": 35.17086330935252,
+      "step": 39110,
+      "torque_loss": 0.17274141311645508
+    },
+    {
+      "epoch": 35.17086330935252,
+      "force_loss": 0.0021876245737075806,
+      "step": 39110
+    },
+    {
+      "epoch": 35.17985611510792,
+      "grad_norm": 0.24375079572200775,
+      "learning_rate": 2.961534587755995e-05,
+      "loss": 0.0163,
+      "step": 39120
+    },
+    {
+      "action_loss": 0.0041252062655985355,
+      "epoch": 35.17985611510792,
+      "step": 39120
+    },
+    {
+      "epoch": 35.17985611510792,
+      "step": 39120,
+      "torque_loss": 0.09842798113822937
+    },
+    {
+      "epoch": 35.17985611510792,
+      "force_loss": 0.0029616865795105696,
+      "step": 39120
+    },
+    {
+      "epoch": 35.18884892086331,
+      "grad_norm": 0.2040484994649887,
+      "learning_rate": 2.959018539591375e-05,
+      "loss": 0.0188,
+      "step": 39130
+    },
+    {
+      "action_loss": 0.00480501726269722,
+      "epoch": 35.18884892086331,
+      "step": 39130
+    },
+    {
+      "epoch": 35.18884892086331,
+      "step": 39130,
+      "torque_loss": 0.1627064347267151
+    },
+    {
+      "epoch": 35.18884892086331,
+      "force_loss": 0.013547913171350956,
+      "step": 39130
+    },
+    {
+      "epoch": 35.197841726618705,
+      "grad_norm": 0.18873806297779083,
+      "learning_rate": 2.9565031114230325e-05,
+      "loss": 0.0199,
+      "step": 39140
+    },
+    {
+      "action_loss": 0.0026393691077828407,
+      "epoch": 35.197841726618705,
+      "step": 39140
+    },
+    {
+      "epoch": 35.197841726618705,
+      "step": 39140,
+      "torque_loss": 0.13031871616840363
+    },
+    {
+      "epoch": 35.197841726618705,
+      "force_loss": 0.0018990668468177319,
+      "step": 39140
+    },
+    {
+      "epoch": 35.2068345323741,
+      "grad_norm": 0.3732083737850189,
+      "learning_rate": 2.9539883040150895e-05,
+      "loss": 0.0178,
+      "step": 39150
+    },
+    {
+      "action_loss": 0.0017531081102788448,
+      "epoch": 35.2068345323741,
+      "step": 39150
+    },
+    {
+      "epoch": 35.2068345323741,
+      "step": 39150,
+      "torque_loss": 0.1356305032968521
+    },
+    {
+      "epoch": 35.2068345323741,
+      "force_loss": 0.0031208631116896868,
+      "step": 39150
+    },
+    {
+      "epoch": 35.21582733812949,
+      "grad_norm": 0.10988090187311172,
+      "learning_rate": 2.9514741181314774e-05,
+      "loss": 0.0181,
+      "step": 39160
+    },
+    {
+      "action_loss": 0.0022965031675994396,
+      "epoch": 35.21582733812949,
+      "step": 39160
+    },
+    {
+      "epoch": 35.21582733812949,
+      "step": 39160,
+      "torque_loss": 0.172601580619812
+    },
+    {
+      "epoch": 35.21582733812949,
+      "force_loss": 0.008354817517101765,
+      "step": 39160
+    },
+    {
+      "epoch": 35.22482014388489,
+      "grad_norm": 0.26232537627220154,
+      "learning_rate": 2.94896055453594e-05,
+      "loss": 0.015,
+      "step": 39170
+    },
+    {
+      "action_loss": 0.0013383496552705765,
+      "epoch": 35.22482014388489,
+      "step": 39170
+    },
+    {
+      "epoch": 35.22482014388489,
+      "step": 39170,
+      "torque_loss": 0.12118307501077652
+    },
+    {
+      "epoch": 35.22482014388489,
+      "force_loss": 0.000979963573627174,
+      "step": 39170
+    },
+    {
+      "epoch": 35.23381294964029,
+      "grad_norm": 0.20352475345134735,
+      "learning_rate": 2.9464476139920332e-05,
+      "loss": 0.0173,
+      "step": 39180
+    },
+    {
+      "action_loss": 0.002695555565878749,
+      "epoch": 35.23381294964029,
+      "step": 39180
+    },
+    {
+      "epoch": 35.23381294964029,
+      "step": 39180,
+      "torque_loss": 0.11937006562948227
+    },
+    {
+      "epoch": 35.23381294964029,
+      "force_loss": 0.002426528139039874,
+      "step": 39180
+    },
+    {
+      "epoch": 35.242805755395686,
+      "grad_norm": 0.17224377393722534,
+      "learning_rate": 2.9439352972631186e-05,
+      "loss": 0.0131,
+      "step": 39190
+    },
+    {
+      "action_loss": 0.0031537513714283705,
+      "epoch": 35.242805755395686,
+      "step": 39190
+    },
+    {
+      "epoch": 35.242805755395686,
+      "step": 39190,
+      "torque_loss": 0.15531350672245026
+    },
+    {
+      "epoch": 35.242805755395686,
+      "force_loss": 0.006157737225294113,
+      "step": 39190
+    },
+    {
+      "epoch": 35.25179856115108,
+      "grad_norm": 0.12478018552064896,
+      "learning_rate": 2.9414236051123757e-05,
+      "loss": 0.0173,
+      "step": 39200
+    },
+    {
+      "action_loss": 0.0021453325171023607,
+      "epoch": 35.25179856115108,
+      "step": 39200
+    },
+    {
+      "epoch": 35.25179856115108,
+      "step": 39200,
+      "torque_loss": 0.1252780556678772
+    },
+    {
+      "epoch": 35.25179856115108,
+      "force_loss": 0.0020287532825022936,
+      "step": 39200
+    },
+    {
+      "epoch": 35.260791366906474,
+      "grad_norm": 0.09150382876396179,
+      "learning_rate": 2.938912538302785e-05,
+      "loss": 0.0155,
+      "step": 39210
+    },
+    {
+      "action_loss": 0.0031896408181637526,
+      "epoch": 35.260791366906474,
+      "step": 39210
+    },
+    {
+      "epoch": 35.260791366906474,
+      "step": 39210,
+      "torque_loss": 0.14173874258995056
+    },
+    {
+      "epoch": 35.260791366906474,
+      "force_loss": 0.0050001805648207664,
+      "step": 39210
+    },
+    {
+      "epoch": 35.26978417266187,
+      "grad_norm": 0.23232616484165192,
+      "learning_rate": 2.9364020975971464e-05,
+      "loss": 0.0138,
+      "step": 39220
+    },
+    {
+      "action_loss": 0.0010937395272776484,
+      "epoch": 35.26978417266187,
+      "step": 39220
+    },
+    {
+      "epoch": 35.26978417266187,
+      "step": 39220,
+      "torque_loss": 0.10851342231035233
+    },
+    {
+      "epoch": 35.26978417266187,
+      "force_loss": 0.0009604405495338142,
+      "step": 39220
+    },
+    {
+      "epoch": 35.27877697841727,
+      "grad_norm": 0.08099891245365143,
+      "learning_rate": 2.9338922837580657e-05,
+      "loss": 0.016,
+      "step": 39230
+    },
+    {
+      "action_loss": 0.004228802863508463,
+      "epoch": 35.27877697841727,
+      "step": 39230
+    },
+    {
+      "epoch": 35.27877697841727,
+      "step": 39230,
+      "torque_loss": 0.11079681664705276
+    },
+    {
+      "epoch": 35.27877697841727,
+      "force_loss": 0.006905920337885618,
+      "step": 39230
+    },
+    {
+      "epoch": 35.28776978417266,
+      "grad_norm": 0.22082647681236267,
+      "learning_rate": 2.931383097547955e-05,
+      "loss": 0.0186,
+      "step": 39240
+    },
+    {
+      "action_loss": 0.006835064385086298,
+      "epoch": 35.28776978417266,
+      "step": 39240
+    },
+    {
+      "epoch": 35.28776978417266,
+      "step": 39240,
+      "torque_loss": 0.11697503924369812
+    },
+    {
+      "epoch": 35.28776978417266,
+      "force_loss": 0.01195115689188242,
+      "step": 39240
+    },
+    {
+      "epoch": 35.29676258992806,
+      "grad_norm": 0.298631489276886,
+      "learning_rate": 2.928874539729043e-05,
+      "loss": 0.0169,
+      "step": 39250
+    },
+    {
+      "action_loss": 0.0017788062104955316,
+      "epoch": 35.29676258992806,
+      "step": 39250
+    },
+    {
+      "epoch": 35.29676258992806,
+      "step": 39250,
+      "torque_loss": 0.0846819281578064
+    },
+    {
+      "epoch": 35.29676258992806,
+      "force_loss": 0.001323180622421205,
+      "step": 39250
+    },
+    {
+      "epoch": 35.305755395683455,
+      "grad_norm": 0.4005330204963684,
+      "learning_rate": 2.926366611063358e-05,
+      "loss": 0.0157,
+      "step": 39260
+    },
+    {
+      "action_loss": 0.002382399281486869,
+      "epoch": 35.305755395683455,
+      "step": 39260
+    },
+    {
+      "epoch": 35.305755395683455,
+      "step": 39260,
+      "torque_loss": 0.19659322500228882
+    },
+    {
+      "epoch": 35.305755395683455,
+      "force_loss": 0.0021546969655901194,
+      "step": 39260
+    },
+    {
+      "epoch": 35.314748201438846,
+      "grad_norm": 0.11796467006206512,
+      "learning_rate": 2.9238593123127463e-05,
+      "loss": 0.0168,
+      "step": 39270
+    },
+    {
+      "action_loss": 0.0013092047302052379,
+      "epoch": 35.314748201438846,
+      "step": 39270
+    },
+    {
+      "epoch": 35.314748201438846,
+      "step": 39270,
+      "torque_loss": 0.07257264107465744
+    },
+    {
+      "epoch": 35.314748201438846,
+      "force_loss": 0.003968419507145882,
+      "step": 39270
+    },
+    {
+      "epoch": 35.32374100719424,
+      "grad_norm": 0.08360259234905243,
+      "learning_rate": 2.9213526442388583e-05,
+      "loss": 0.0143,
+      "step": 39280
+    },
+    {
+      "action_loss": 0.0014142096042633057,
+      "epoch": 35.32374100719424,
+      "step": 39280
+    },
+    {
+      "epoch": 35.32374100719424,
+      "step": 39280,
+      "torque_loss": 0.18021027743816376
+    },
+    {
+      "epoch": 35.32374100719424,
+      "force_loss": 0.0019014537101611495,
+      "step": 39280
+    },
+    {
+      "epoch": 35.33273381294964,
+      "grad_norm": 0.08111292868852615,
+      "learning_rate": 2.9188466076031545e-05,
+      "loss": 0.0141,
+      "step": 39290
+    },
+    {
+      "action_loss": 0.0012782718986272812,
+      "epoch": 35.33273381294964,
+      "step": 39290
+    },
+    {
+      "epoch": 35.33273381294964,
+      "step": 39290,
+      "torque_loss": 0.14475487172603607
+    },
+    {
+      "epoch": 35.33273381294964,
+      "force_loss": 0.0027069367934018373,
+      "step": 39290
+    },
+    {
+      "epoch": 35.34172661870504,
+      "grad_norm": 0.19811029732227325,
+      "learning_rate": 2.9163412031669012e-05,
+      "loss": 0.0158,
+      "step": 39300
+    },
+    {
+      "action_loss": 0.0013705870369449258,
+      "epoch": 35.34172661870504,
+      "step": 39300
+    },
+    {
+      "epoch": 35.34172661870504,
+      "step": 39300,
+      "torque_loss": 0.11132822185754776
+    },
+    {
+      "epoch": 35.34172661870504,
+      "force_loss": 0.0021050074137747288,
+      "step": 39300
+    },
+    {
+      "epoch": 35.35071942446043,
+      "grad_norm": 0.08124122768640518,
+      "learning_rate": 2.913836431691175e-05,
+      "loss": 0.0181,
+      "step": 39310
+    },
+    {
+      "action_loss": 0.0010235168738290668,
+      "epoch": 35.35071942446043,
+      "step": 39310
+    },
+    {
+      "epoch": 35.35071942446043,
+      "step": 39310,
+      "torque_loss": 0.1073768138885498
+    },
+    {
+      "epoch": 35.35071942446043,
+      "force_loss": 0.0006691438029520214,
+      "step": 39310
+    },
+    {
+      "epoch": 35.35971223021583,
+      "grad_norm": 0.09625593572854996,
+      "learning_rate": 2.9113322939368583e-05,
+      "loss": 0.0179,
+      "step": 39320
+    },
+    {
+      "action_loss": 0.001894653425551951,
+      "epoch": 35.35971223021583,
+      "step": 39320
+    },
+    {
+      "epoch": 35.35971223021583,
+      "step": 39320,
+      "torque_loss": 0.1327088177204132
+    },
+    {
+      "epoch": 35.35971223021583,
+      "force_loss": 0.0035764549393206835,
+      "step": 39320
+    },
+    {
+      "epoch": 35.368705035971225,
+      "grad_norm": 0.13951651751995087,
+      "learning_rate": 2.9088287906646427e-05,
+      "loss": 0.0185,
+      "step": 39330
+    },
+    {
+      "action_loss": 0.005528716370463371,
+      "epoch": 35.368705035971225,
+      "step": 39330
+    },
+    {
+      "epoch": 35.368705035971225,
+      "step": 39330,
+      "torque_loss": 0.12613056600093842
+    },
+    {
+      "epoch": 35.368705035971225,
+      "force_loss": 0.006761878728866577,
+      "step": 39330
+    },
+    {
+      "epoch": 35.37769784172662,
+      "grad_norm": 0.2915889322757721,
+      "learning_rate": 2.906325922635024e-05,
+      "loss": 0.0169,
+      "step": 39340
+    },
+    {
+      "action_loss": 0.0013506049290299416,
+      "epoch": 35.37769784172662,
+      "step": 39340
+    },
+    {
+      "epoch": 35.37769784172662,
+      "step": 39340,
+      "torque_loss": 0.0970362201333046
+    },
+    {
+      "epoch": 35.37769784172662,
+      "force_loss": 0.0008510045590810478,
+      "step": 39340
+    },
+    {
+      "epoch": 35.38669064748201,
+      "grad_norm": 0.2827681601047516,
+      "learning_rate": 2.903823690608313e-05,
+      "loss": 0.0153,
+      "step": 39350
+    },
+    {
+      "action_loss": 0.004065061453729868,
+      "epoch": 35.38669064748201,
+      "step": 39350
+    },
+    {
+      "epoch": 35.38669064748201,
+      "step": 39350,
+      "torque_loss": 0.1831979751586914
+    },
+    {
+      "epoch": 35.38669064748201,
+      "force_loss": 0.005742497276514769,
+      "step": 39350
+    },
+    {
+      "epoch": 35.39568345323741,
+      "grad_norm": 0.127257838845253,
+      "learning_rate": 2.9013220953446174e-05,
+      "loss": 0.0171,
+      "step": 39360
+    },
+    {
+      "action_loss": 0.0010706715984269977,
+      "epoch": 35.39568345323741,
+      "step": 39360
+    },
+    {
+      "epoch": 35.39568345323741,
+      "step": 39360,
+      "torque_loss": 0.12138820439577103
+    },
+    {
+      "epoch": 35.39568345323741,
+      "force_loss": 0.0009163037757389247,
+      "step": 39360
+    },
+    {
+      "epoch": 35.40467625899281,
+      "grad_norm": 0.144105464220047,
+      "learning_rate": 2.8988211376038564e-05,
+      "loss": 0.0147,
+      "step": 39370
+    },
+    {
+      "action_loss": 0.0012387748574838042,
+      "epoch": 35.40467625899281,
+      "step": 39370
+    },
+    {
+      "epoch": 35.40467625899281,
+      "step": 39370,
+      "torque_loss": 0.12555968761444092
+    },
+    {
+      "epoch": 35.40467625899281,
+      "force_loss": 0.000849484873469919,
+      "step": 39370
+    },
+    {
+      "epoch": 35.4136690647482,
+      "grad_norm": 0.31135427951812744,
+      "learning_rate": 2.8963208181457564e-05,
+      "loss": 0.0155,
+      "step": 39380
+    },
+    {
+      "action_loss": 0.014255337417125702,
+      "epoch": 35.4136690647482,
+      "step": 39380
+    },
+    {
+      "epoch": 35.4136690647482,
+      "step": 39380,
+      "torque_loss": 0.15915845334529877
+    },
+    {
+      "epoch": 35.4136690647482,
+      "force_loss": 0.017163896933197975,
+      "step": 39380
+    },
+    {
+      "epoch": 35.422661870503596,
+      "grad_norm": 0.14365169405937195,
+      "learning_rate": 2.8938211377298453e-05,
+      "loss": 0.023,
+      "step": 39390
+    },
+    {
+      "action_loss": 0.004171127453446388,
+      "epoch": 35.422661870503596,
+      "step": 39390
+    },
+    {
+      "epoch": 35.422661870503596,
+      "step": 39390,
+      "torque_loss": 0.09644972532987595
+    },
+    {
+      "epoch": 35.422661870503596,
+      "force_loss": 0.0029431155417114496,
+      "step": 39390
+    },
+    {
+      "epoch": 35.431654676258994,
+      "grad_norm": 0.1570999026298523,
+      "learning_rate": 2.8913220971154652e-05,
+      "loss": 0.0232,
+      "step": 39400
+    },
+    {
+      "action_loss": 0.004454169888049364,
+      "epoch": 35.431654676258994,
+      "step": 39400
+    },
+    {
+      "epoch": 35.431654676258994,
+      "step": 39400,
+      "torque_loss": 0.12635324895381927
+    },
+    {
+      "epoch": 35.431654676258994,
+      "force_loss": 0.002220805501565337,
+      "step": 39400
+    },
+    {
+      "epoch": 35.44064748201439,
+      "grad_norm": 0.11610624939203262,
+      "learning_rate": 2.888823697061753e-05,
+      "loss": 0.0159,
+      "step": 39410
+    },
+    {
+      "action_loss": 0.0035125084687024355,
+      "epoch": 35.44064748201439,
+      "step": 39410
+    },
+    {
+      "epoch": 35.44064748201439,
+      "step": 39410,
+      "torque_loss": 0.1324479579925537
+    },
+    {
+      "epoch": 35.44064748201439,
+      "force_loss": 0.005658428650349379,
+      "step": 39410
+    },
+    {
+      "epoch": 35.44964028776978,
+      "grad_norm": 0.19316492974758148,
+      "learning_rate": 2.8863259383276618e-05,
+      "loss": 0.0178,
+      "step": 39420
+    },
+    {
+      "action_loss": 0.00290547008626163,
+      "epoch": 35.44964028776978,
+      "step": 39420
+    },
+    {
+      "epoch": 35.44964028776978,
+      "step": 39420,
+      "torque_loss": 0.12777863442897797
+    },
+    {
+      "epoch": 35.44964028776978,
+      "force_loss": 0.017759613692760468,
+      "step": 39420
+    },
+    {
+      "epoch": 35.45863309352518,
+      "grad_norm": 0.27461129426956177,
+      "learning_rate": 2.8838288216719395e-05,
+      "loss": 0.0166,
+      "step": 39430
+    },
+    {
+      "action_loss": 0.003124079667031765,
+      "epoch": 35.45863309352518,
+      "step": 39430
+    },
+    {
+      "epoch": 35.45863309352518,
+      "step": 39430,
+      "torque_loss": 0.15821687877178192
+    },
+    {
+      "epoch": 35.45863309352518,
+      "force_loss": 0.0033759765792638063,
+      "step": 39430
+    },
+    {
+      "epoch": 35.46762589928058,
+      "grad_norm": 0.17730678617954254,
+      "learning_rate": 2.8813323478531484e-05,
+      "loss": 0.0158,
+      "step": 39440
+    },
+    {
+      "action_loss": 0.0015664267120882869,
+      "epoch": 35.46762589928058,
+      "step": 39440
+    },
+    {
+      "epoch": 35.46762589928058,
+      "step": 39440,
+      "torque_loss": 0.1286115199327469
+    },
+    {
+      "epoch": 35.46762589928058,
+      "force_loss": 0.0024287817068398,
+      "step": 39440
+    },
+    {
+      "epoch": 35.476618705035975,
+      "grad_norm": 0.49196091294288635,
+      "learning_rate": 2.8788365176296496e-05,
+      "loss": 0.0231,
+      "step": 39450
+    },
+    {
+      "action_loss": 0.00135201180819422,
+      "epoch": 35.476618705035975,
+      "step": 39450
+    },
+    {
+      "epoch": 35.476618705035975,
+      "step": 39450,
+      "torque_loss": 0.12456002086400986
+    },
+    {
+      "epoch": 35.476618705035975,
+      "force_loss": 0.001781140803359449,
+      "step": 39450
+    },
+    {
+      "epoch": 35.485611510791365,
+      "grad_norm": 0.13427764177322388,
+      "learning_rate": 2.876341331759611e-05,
+      "loss": 0.0143,
+      "step": 39460
+    },
+    {
+      "action_loss": 0.0027898463886231184,
+      "epoch": 35.485611510791365,
+      "step": 39460
+    },
+    {
+      "epoch": 35.485611510791365,
+      "step": 39460,
+      "torque_loss": 0.13061822950839996
+    },
+    {
+      "epoch": 35.485611510791365,
+      "force_loss": 0.0048857200890779495,
+      "step": 39460
+    },
+    {
+      "epoch": 35.49460431654676,
+      "grad_norm": 0.07692616432905197,
+      "learning_rate": 2.8738467910010036e-05,
+      "loss": 0.0194,
+      "step": 39470
+    },
+    {
+      "action_loss": 0.0025288949254900217,
+      "epoch": 35.49460431654676,
+      "step": 39470
+    },
+    {
+      "epoch": 35.49460431654676,
+      "step": 39470,
+      "torque_loss": 0.10948216915130615
+    },
+    {
+      "epoch": 35.49460431654676,
+      "force_loss": 0.004075657110661268,
+      "step": 39470
+    },
+    {
+      "epoch": 35.50359712230216,
+      "grad_norm": 0.3425906300544739,
+      "learning_rate": 2.8713528961116032e-05,
+      "loss": 0.0158,
+      "step": 39480
+    },
+    {
+      "action_loss": 0.014046314172446728,
+      "epoch": 35.50359712230216,
+      "step": 39480
+    },
+    {
+      "epoch": 35.50359712230216,
+      "step": 39480,
+      "torque_loss": 0.15999191999435425
+    },
+    {
+      "epoch": 35.50359712230216,
+      "force_loss": 0.012685430236160755,
+      "step": 39480
+    },
+    {
+      "epoch": 35.51258992805755,
+      "grad_norm": 0.1947794407606125,
+      "learning_rate": 2.8688596478489875e-05,
+      "loss": 0.0192,
+      "step": 39490
+    },
+    {
+      "action_loss": 0.004564117640256882,
+      "epoch": 35.51258992805755,
+      "step": 39490
+    },
+    {
+      "epoch": 35.51258992805755,
+      "step": 39490,
+      "torque_loss": 0.1416541188955307
+    },
+    {
+      "epoch": 35.51258992805755,
+      "force_loss": 0.007333502173423767,
+      "step": 39490
+    },
+    {
+      "epoch": 35.52158273381295,
+      "grad_norm": 0.2513328492641449,
+      "learning_rate": 2.8663670469705434e-05,
+      "loss": 0.0172,
+      "step": 39500
+    },
+    {
+      "action_loss": 0.0026504509150981903,
+      "epoch": 35.52158273381295,
+      "step": 39500
+    },
+    {
+      "epoch": 35.52158273381295,
+      "step": 39500,
+      "torque_loss": 0.10963200777769089
+    },
+    {
+      "epoch": 35.52158273381295,
+      "force_loss": 0.004098037723451853,
+      "step": 39500
+    },
+    {
+      "epoch": 35.530575539568346,
+      "grad_norm": 0.11168992519378662,
+      "learning_rate": 2.8638750942334546e-05,
+      "loss": 0.0187,
+      "step": 39510
+    },
+    {
+      "action_loss": 0.002348459092900157,
+      "epoch": 35.530575539568346,
+      "step": 39510
+    },
+    {
+      "epoch": 35.530575539568346,
+      "step": 39510,
+      "torque_loss": 0.13284127414226532
+    },
+    {
+      "epoch": 35.530575539568346,
+      "force_loss": 0.002375292591750622,
+      "step": 39510
+    },
+    {
+      "epoch": 35.539568345323744,
+      "grad_norm": 0.282320499420166,
+      "learning_rate": 2.8613837903947115e-05,
+      "loss": 0.0151,
+      "step": 39520
+    },
+    {
+      "action_loss": 0.01882605440914631,
+      "epoch": 35.539568345323744,
+      "step": 39520
+    },
+    {
+      "epoch": 35.539568345323744,
+      "step": 39520,
+      "torque_loss": 0.14823205769062042
+    },
+    {
+      "epoch": 35.539568345323744,
+      "force_loss": 0.032957300543785095,
+      "step": 39520
+    },
+    {
+      "epoch": 35.548561151079134,
+      "grad_norm": 0.23333409428596497,
+      "learning_rate": 2.858893136211106e-05,
+      "loss": 0.0182,
+      "step": 39530
+    },
+    {
+      "action_loss": 0.007244527339935303,
+      "epoch": 35.548561151079134,
+      "step": 39530
+    },
+    {
+      "epoch": 35.548561151079134,
+      "step": 39530,
+      "torque_loss": 0.1048051193356514
+    },
+    {
+      "epoch": 35.548561151079134,
+      "force_loss": 0.006691901013255119,
+      "step": 39530
+    },
+    {
+      "epoch": 35.55755395683453,
+      "grad_norm": 0.1614960879087448,
+      "learning_rate": 2.8564031324392315e-05,
+      "loss": 0.0162,
+      "step": 39540
+    },
+    {
+      "action_loss": 0.0025056616868823767,
+      "epoch": 35.55755395683453,
+      "step": 39540
+    },
+    {
+      "epoch": 35.55755395683453,
+      "step": 39540,
+      "torque_loss": 0.12144249677658081
+    },
+    {
+      "epoch": 35.55755395683453,
+      "force_loss": 0.003249605419114232,
+      "step": 39540
+    },
+    {
+      "epoch": 35.56654676258993,
+      "grad_norm": 0.3132919669151306,
+      "learning_rate": 2.85391377983549e-05,
+      "loss": 0.0176,
+      "step": 39550
+    },
+    {
+      "action_loss": 0.00451565720140934,
+      "epoch": 35.56654676258993,
+      "step": 39550
+    },
+    {
+      "epoch": 35.56654676258993,
+      "step": 39550,
+      "torque_loss": 0.14093726873397827
+    },
+    {
+      "epoch": 35.56654676258993,
+      "force_loss": 0.008190018124878407,
+      "step": 39550
+    },
+    {
+      "epoch": 35.57553956834532,
+      "grad_norm": 0.3187926709651947,
+      "learning_rate": 2.851425079156075e-05,
+      "loss": 0.0211,
+      "step": 39560
+    },
+    {
+      "action_loss": 0.0034284184221178293,
+      "epoch": 35.57553956834532,
+      "step": 39560
+    },
+    {
+      "epoch": 35.57553956834532,
+      "step": 39560,
+      "torque_loss": 0.11422654241323471
+    },
+    {
+      "epoch": 35.57553956834532,
+      "force_loss": 0.004590218886733055,
+      "step": 39560
+    },
+    {
+      "epoch": 35.58453237410072,
+      "grad_norm": 0.42136937379837036,
+      "learning_rate": 2.848937031156994e-05,
+      "loss": 0.0156,
+      "step": 39570
+    },
+    {
+      "action_loss": 0.001175474957562983,
+      "epoch": 35.58453237410072,
+      "step": 39570
+    },
+    {
+      "epoch": 35.58453237410072,
+      "step": 39570,
+      "torque_loss": 0.14840276539325714
+    },
+    {
+      "epoch": 35.58453237410072,
+      "force_loss": 0.0011653221445158124,
+      "step": 39570
+    },
+    {
+      "epoch": 35.593525179856115,
+      "grad_norm": 0.22646832466125488,
+      "learning_rate": 2.846449636594044e-05,
+      "loss": 0.0217,
+      "step": 39580
+    },
+    {
+      "action_loss": 0.008363465778529644,
+      "epoch": 35.593525179856115,
+      "step": 39580
+    },
+    {
+      "epoch": 35.593525179856115,
+      "step": 39580,
+      "torque_loss": 0.1058172956109047
+    },
+    {
+      "epoch": 35.593525179856115,
+      "force_loss": 0.014881676994264126,
+      "step": 39580
+    },
+    {
+      "epoch": 35.60251798561151,
+      "grad_norm": 0.1466566026210785,
+      "learning_rate": 2.843962896222836e-05,
+      "loss": 0.0165,
+      "step": 39590
+    },
+    {
+      "action_loss": 0.008684664964675903,
+      "epoch": 35.60251798561151,
+      "step": 39590
+    },
+    {
+      "epoch": 35.60251798561151,
+      "step": 39590,
+      "torque_loss": 0.14061520993709564
+    },
+    {
+      "epoch": 35.60251798561151,
+      "force_loss": 0.016985170543193817,
+      "step": 39590
+    },
+    {
+      "epoch": 35.611510791366904,
+      "grad_norm": 0.17727145552635193,
+      "learning_rate": 2.8414768107987722e-05,
+      "loss": 0.0179,
+      "step": 39600
+    },
+    {
+      "action_loss": 0.0039934213273227215,
+      "epoch": 35.611510791366904,
+      "step": 39600
+    },
+    {
+      "epoch": 35.611510791366904,
+      "step": 39600,
+      "torque_loss": 0.11750852316617966
+    },
+    {
+      "epoch": 35.611510791366904,
+      "force_loss": 0.004186486825346947,
+      "step": 39600
+    },
+    {
+      "epoch": 35.6205035971223,
+      "grad_norm": 0.2128559947013855,
+      "learning_rate": 2.838991381077061e-05,
+      "loss": 0.0185,
+      "step": 39610
+    },
+    {
+      "action_loss": 0.003089507343247533,
+      "epoch": 35.6205035971223,
+      "step": 39610
+    },
+    {
+      "epoch": 35.6205035971223,
+      "step": 39610,
+      "torque_loss": 0.1126134991645813
+    },
+    {
+      "epoch": 35.6205035971223,
+      "force_loss": 0.006907335016876459,
+      "step": 39610
+    },
+    {
+      "epoch": 35.6294964028777,
+      "grad_norm": 0.15802417695522308,
+      "learning_rate": 2.83650660781271e-05,
+      "loss": 0.0154,
+      "step": 39620
+    },
+    {
+      "action_loss": 0.001239847973920405,
+      "epoch": 35.6294964028777,
+      "step": 39620
+    },
+    {
+      "epoch": 35.6294964028777,
+      "step": 39620,
+      "torque_loss": 0.09498241543769836
+    },
+    {
+      "epoch": 35.6294964028777,
+      "force_loss": 0.001859107636846602,
+      "step": 39620
+    },
+    {
+      "epoch": 35.638489208633096,
+      "grad_norm": 0.1919955164194107,
+      "learning_rate": 2.8340224917605285e-05,
+      "loss": 0.0145,
+      "step": 39630
+    },
+    {
+      "action_loss": 0.003541032550856471,
+      "epoch": 35.638489208633096,
+      "step": 39630
+    },
+    {
+      "epoch": 35.638489208633096,
+      "step": 39630,
+      "torque_loss": 0.13486574590206146
+    },
+    {
+      "epoch": 35.638489208633096,
+      "force_loss": 0.003753454191610217,
+      "step": 39630
+    },
+    {
+      "epoch": 35.64748201438849,
+      "grad_norm": 0.20450246334075928,
+      "learning_rate": 2.831539033675122e-05,
+      "loss": 0.0187,
+      "step": 39640
+    },
+    {
+      "action_loss": 0.003315453650429845,
+      "epoch": 35.64748201438849,
+      "step": 39640
+    },
+    {
+      "epoch": 35.64748201438849,
+      "step": 39640,
+      "torque_loss": 0.09184668213129044
+    },
+    {
+      "epoch": 35.64748201438849,
+      "force_loss": 0.001953155966475606,
+      "step": 39640
+    },
+    {
+      "epoch": 35.656474820143885,
+      "grad_norm": 0.1740974634885788,
+      "learning_rate": 2.8290562343109038e-05,
+      "loss": 0.0209,
+      "step": 39650
+    },
+    {
+      "action_loss": 0.004921962972730398,
+      "epoch": 35.656474820143885,
+      "step": 39650
+    },
+    {
+      "epoch": 35.656474820143885,
+      "step": 39650,
+      "torque_loss": 0.08533027768135071
+    },
+    {
+      "epoch": 35.656474820143885,
+      "force_loss": 0.006142523139715195,
+      "step": 39650
+    },
+    {
+      "epoch": 35.66546762589928,
+      "grad_norm": 0.11232845485210419,
+      "learning_rate": 2.826574094422082e-05,
+      "loss": 0.0154,
+      "step": 39660
+    },
+    {
+      "action_loss": 0.00197235937230289,
+      "epoch": 35.66546762589928,
+      "step": 39660
+    },
+    {
+      "epoch": 35.66546762589928,
+      "step": 39660,
+      "torque_loss": 0.11520601063966751
+    },
+    {
+      "epoch": 35.66546762589928,
+      "force_loss": 0.004631725139915943,
+      "step": 39660
+    },
+    {
+      "epoch": 35.67446043165468,
+      "grad_norm": 0.20036110281944275,
+      "learning_rate": 2.8240926147626645e-05,
+      "loss": 0.0193,
+      "step": 39670
+    },
+    {
+      "action_loss": 0.0008053396013565361,
+      "epoch": 35.67446043165468,
+      "step": 39670
+    },
+    {
+      "epoch": 35.67446043165468,
+      "step": 39670,
+      "torque_loss": 0.09567391872406006
+    },
+    {
+      "epoch": 35.67446043165468,
+      "force_loss": 0.0011568082263693213,
+      "step": 39670
+    },
+    {
+      "epoch": 35.68345323741007,
+      "grad_norm": 0.3572350740432739,
+      "learning_rate": 2.8216117960864586e-05,
+      "loss": 0.0192,
+      "step": 39680
+    },
+    {
+      "action_loss": 0.002147337654605508,
+      "epoch": 35.68345323741007,
+      "step": 39680
+    },
+    {
+      "epoch": 35.68345323741007,
+      "step": 39680,
+      "torque_loss": 0.1557895392179489
+    },
+    {
+      "epoch": 35.68345323741007,
+      "force_loss": 0.013895433396100998,
+      "step": 39680
+    },
+    {
+      "epoch": 35.69244604316547,
+      "grad_norm": 0.18563728034496307,
+      "learning_rate": 2.8191316391470703e-05,
+      "loss": 0.0164,
+      "step": 39690
+    },
+    {
+      "action_loss": 0.00875378493219614,
+      "epoch": 35.69244604316547,
+      "step": 39690
+    },
+    {
+      "epoch": 35.69244604316547,
+      "step": 39690,
+      "torque_loss": 0.13940010964870453
+    },
+    {
+      "epoch": 35.69244604316547,
+      "force_loss": 0.011029067449271679,
+      "step": 39690
+    },
+    {
+      "epoch": 35.701438848920866,
+      "grad_norm": 0.10199469327926636,
+      "learning_rate": 2.816652144697911e-05,
+      "loss": 0.0211,
+      "step": 39700
+    },
+    {
+      "action_loss": 0.0029193172231316566,
+      "epoch": 35.701438848920866,
+      "step": 39700
+    },
+    {
+      "epoch": 35.701438848920866,
+      "step": 39700,
+      "torque_loss": 0.1119891107082367
+    },
+    {
+      "epoch": 35.701438848920866,
+      "force_loss": 0.0015033232048153877,
+      "step": 39700
+    },
+    {
+      "epoch": 35.710431654676256,
+      "grad_norm": 0.1253720074892044,
+      "learning_rate": 2.8141733134921783e-05,
+      "loss": 0.0204,
+      "step": 39710
+    },
+    {
+      "action_loss": 0.0032067298889160156,
+      "epoch": 35.710431654676256,
+      "step": 39710
+    },
+    {
+      "epoch": 35.710431654676256,
+      "step": 39710,
+      "torque_loss": 0.11965510994195938
+    },
+    {
+      "epoch": 35.710431654676256,
+      "force_loss": 0.0026536390651017427,
+      "step": 39710
+    },
+    {
+      "epoch": 35.719424460431654,
+      "grad_norm": 0.41262295842170715,
+      "learning_rate": 2.811695146282884e-05,
+      "loss": 0.0179,
+      "step": 39720
+    },
+    {
+      "action_loss": 0.0026037513744086027,
+      "epoch": 35.719424460431654,
+      "step": 39720
+    },
+    {
+      "epoch": 35.719424460431654,
+      "step": 39720,
+      "torque_loss": 0.10272317379713058
+    },
+    {
+      "epoch": 35.719424460431654,
+      "force_loss": 0.0028136586770415306,
+      "step": 39720
+    },
+    {
+      "epoch": 35.72841726618705,
+      "grad_norm": 0.20123039186000824,
+      "learning_rate": 2.8092176438228212e-05,
+      "loss": 0.0161,
+      "step": 39730
+    },
+    {
+      "action_loss": 0.0025930143892765045,
+      "epoch": 35.72841726618705,
+      "step": 39730
+    },
+    {
+      "epoch": 35.72841726618705,
+      "step": 39730,
+      "torque_loss": 0.11339656263589859
+    },
+    {
+      "epoch": 35.72841726618705,
+      "force_loss": 0.005493659991770983,
+      "step": 39730
+    },
+    {
+      "epoch": 35.73741007194245,
+      "grad_norm": 0.31548067927360535,
+      "learning_rate": 2.806740806864598e-05,
+      "loss": 0.0194,
+      "step": 39740
+    },
+    {
+      "action_loss": 0.0027863343711942434,
+      "epoch": 35.73741007194245,
+      "step": 39740
+    },
+    {
+      "epoch": 35.73741007194245,
+      "step": 39740,
+      "torque_loss": 0.14005906879901886
+    },
+    {
+      "epoch": 35.73741007194245,
+      "force_loss": 0.0059461998753249645,
+      "step": 39740
+    },
+    {
+      "epoch": 35.74640287769784,
+      "grad_norm": 0.291875958442688,
+      "learning_rate": 2.804264636160604e-05,
+      "loss": 0.0161,
+      "step": 39750
+    },
+    {
+      "action_loss": 0.008138484321534634,
+      "epoch": 35.74640287769784,
+      "step": 39750
+    },
+    {
+      "epoch": 35.74640287769784,
+      "step": 39750,
+      "torque_loss": 0.1472003012895584
+    },
+    {
+      "epoch": 35.74640287769784,
+      "force_loss": 0.004367370158433914,
+      "step": 39750
+    },
+    {
+      "epoch": 35.75539568345324,
+      "grad_norm": 0.10295046120882034,
+      "learning_rate": 2.8017891324630402e-05,
+      "loss": 0.0152,
+      "step": 39760
+    },
+    {
+      "action_loss": 0.003908615559339523,
+      "epoch": 35.75539568345324,
+      "step": 39760
+    },
+    {
+      "epoch": 35.75539568345324,
+      "step": 39760,
+      "torque_loss": 0.16097819805145264
+    },
+    {
+      "epoch": 35.75539568345324,
+      "force_loss": 0.0038034298922866583,
+      "step": 39760
+    },
+    {
+      "epoch": 35.764388489208635,
+      "grad_norm": 0.13721863925457,
+      "learning_rate": 2.7993142965238976e-05,
+      "loss": 0.0193,
+      "step": 39770
+    },
+    {
+      "action_loss": 0.00265670008957386,
+      "epoch": 35.764388489208635,
+      "step": 39770
+    },
+    {
+      "epoch": 35.764388489208635,
+      "step": 39770,
+      "torque_loss": 0.10878171771764755
+    },
+    {
+      "epoch": 35.764388489208635,
+      "force_loss": 0.0066906786523759365,
+      "step": 39770
+    },
+    {
+      "epoch": 35.773381294964025,
+      "grad_norm": 0.12275765836238861,
+      "learning_rate": 2.7968401290949665e-05,
+      "loss": 0.0167,
+      "step": 39780
+    },
+    {
+      "action_loss": 0.004790095612406731,
+      "epoch": 35.773381294964025,
+      "step": 39780
+    },
+    {
+      "epoch": 35.773381294964025,
+      "step": 39780,
+      "torque_loss": 0.15864305198192596
+    },
+    {
+      "epoch": 35.773381294964025,
+      "force_loss": 0.006495725363492966,
+      "step": 39780
+    },
+    {
+      "epoch": 35.78237410071942,
+      "grad_norm": 0.11805963516235352,
+      "learning_rate": 2.7943666309278328e-05,
+      "loss": 0.0164,
+      "step": 39790
+    },
+    {
+      "action_loss": 0.0025398877914994955,
+      "epoch": 35.78237410071942,
+      "step": 39790
+    },
+    {
+      "epoch": 35.78237410071942,
+      "step": 39790,
+      "torque_loss": 0.08587618917226791
+    },
+    {
+      "epoch": 35.78237410071942,
+      "force_loss": 0.004442976322025061,
+      "step": 39790
+    },
+    {
+      "epoch": 35.79136690647482,
+      "grad_norm": 0.1223180741071701,
+      "learning_rate": 2.7918938027738783e-05,
+      "loss": 0.017,
+      "step": 39800
+    },
+    {
+      "action_loss": 0.0015568555099889636,
+      "epoch": 35.79136690647482,
+      "step": 39800
+    },
+    {
+      "epoch": 35.79136690647482,
+      "step": 39800,
+      "torque_loss": 0.11426212638616562
+    },
+    {
+      "epoch": 35.79136690647482,
+      "force_loss": 0.0010734944371506572,
+      "step": 39800
+    },
+    {
+      "epoch": 35.80035971223022,
+      "grad_norm": 0.12311927229166031,
+      "learning_rate": 2.789421645384287e-05,
+      "loss": 0.0144,
+      "step": 39810
+    },
+    {
+      "action_loss": 0.024767877534031868,
+      "epoch": 35.80035971223022,
+      "step": 39810
+    },
+    {
+      "epoch": 35.80035971223022,
+      "step": 39810,
+      "torque_loss": 0.12275147438049316
+    },
+    {
+      "epoch": 35.80035971223022,
+      "force_loss": 0.028525101020932198,
+      "step": 39810
+    },
+    {
+      "epoch": 35.80935251798561,
+      "grad_norm": 0.17425116896629333,
+      "learning_rate": 2.786950159510032e-05,
+      "loss": 0.0185,
+      "step": 39820
+    },
+    {
+      "action_loss": 0.0011173957027494907,
+      "epoch": 35.80935251798561,
+      "step": 39820
+    },
+    {
+      "epoch": 35.80935251798561,
+      "step": 39820,
+      "torque_loss": 0.1357005089521408
+    },
+    {
+      "epoch": 35.80935251798561,
+      "force_loss": 0.004809817299246788,
+      "step": 39820
+    },
+    {
+      "epoch": 35.818345323741006,
+      "grad_norm": 0.3158116936683655,
+      "learning_rate": 2.7844793459018876e-05,
+      "loss": 0.0171,
+      "step": 39830
+    },
+    {
+      "action_loss": 0.0019987565465271473,
+      "epoch": 35.818345323741006,
+      "step": 39830
+    },
+    {
+      "epoch": 35.818345323741006,
+      "step": 39830,
+      "torque_loss": 0.10832720994949341
+    },
+    {
+      "epoch": 35.818345323741006,
+      "force_loss": 0.0037875166162848473,
+      "step": 39830
+    },
+    {
+      "epoch": 35.827338129496404,
+      "grad_norm": 0.2635257840156555,
+      "learning_rate": 2.7820092053104195e-05,
+      "loss": 0.0158,
+      "step": 39840
+    },
+    {
+      "action_loss": 0.0046249027363955975,
+      "epoch": 35.827338129496404,
+      "step": 39840
+    },
+    {
+      "epoch": 35.827338129496404,
+      "step": 39840,
+      "torque_loss": 0.14352811872959137
+    },
+    {
+      "epoch": 35.827338129496404,
+      "force_loss": 0.005319828633219004,
+      "step": 39840
+    },
+    {
+      "epoch": 35.8363309352518,
+      "grad_norm": 0.0906476303935051,
+      "learning_rate": 2.7795397384859933e-05,
+      "loss": 0.0148,
+      "step": 39850
+    },
+    {
+      "action_loss": 0.007267903536558151,
+      "epoch": 35.8363309352518,
+      "step": 39850
+    },
+    {
+      "epoch": 35.8363309352518,
+      "step": 39850,
+      "torque_loss": 0.11004098504781723
+    },
+    {
+      "epoch": 35.8363309352518,
+      "force_loss": 0.014775082468986511,
+      "step": 39850
+    },
+    {
+      "epoch": 35.84532374100719,
+      "grad_norm": 0.1165788546204567,
+      "learning_rate": 2.7770709461787638e-05,
+      "loss": 0.0159,
+      "step": 39860
+    },
+    {
+      "action_loss": 0.0020856165792793036,
+      "epoch": 35.84532374100719,
+      "step": 39860
+    },
+    {
+      "epoch": 35.84532374100719,
+      "step": 39860,
+      "torque_loss": 0.1354133039712906
+    },
+    {
+      "epoch": 35.84532374100719,
+      "force_loss": 0.001596722169779241,
+      "step": 39860
+    },
+    {
+      "epoch": 35.85431654676259,
+      "grad_norm": 0.13126079738140106,
+      "learning_rate": 2.7746028291386915e-05,
+      "loss": 0.0153,
+      "step": 39870
+    },
+    {
+      "action_loss": 0.002711870474740863,
+      "epoch": 35.85431654676259,
+      "step": 39870
+    },
+    {
+      "epoch": 35.85431654676259,
+      "step": 39870,
+      "torque_loss": 0.13184605538845062
+    },
+    {
+      "epoch": 35.85431654676259,
+      "force_loss": 0.002248078351840377,
+      "step": 39870
+    },
+    {
+      "epoch": 35.86330935251799,
+      "grad_norm": 0.5663595795631409,
+      "learning_rate": 2.772135388115519e-05,
+      "loss": 0.0175,
+      "step": 39880
+    },
+    {
+      "action_loss": 0.002428376814350486,
+      "epoch": 35.86330935251799,
+      "step": 39880
+    },
+    {
+      "epoch": 35.86330935251799,
+      "step": 39880,
+      "torque_loss": 0.16598446667194366
+    },
+    {
+      "epoch": 35.86330935251799,
+      "force_loss": 0.0022639110684394836,
+      "step": 39880
+    },
+    {
+      "epoch": 35.87230215827338,
+      "grad_norm": 0.29061752557754517,
+      "learning_rate": 2.7696686238587945e-05,
+      "loss": 0.017,
+      "step": 39890
+    },
+    {
+      "action_loss": 0.005739705171436071,
+      "epoch": 35.87230215827338,
+      "step": 39890
+    },
+    {
+      "epoch": 35.87230215827338,
+      "step": 39890,
+      "torque_loss": 0.1431030035018921
+    },
+    {
+      "epoch": 35.87230215827338,
+      "force_loss": 0.009874660521745682,
+      "step": 39890
+    },
+    {
+      "epoch": 35.881294964028775,
+      "grad_norm": 0.11623810976743698,
+      "learning_rate": 2.7672025371178505e-05,
+      "loss": 0.0201,
+      "step": 39900
+    },
+    {
+      "action_loss": 0.0011906855506822467,
+      "epoch": 35.881294964028775,
+      "step": 39900
+    },
+    {
+      "epoch": 35.881294964028775,
+      "step": 39900,
+      "torque_loss": 0.0976327657699585
+    },
+    {
+      "epoch": 35.881294964028775,
+      "force_loss": 0.0008170453947968781,
+      "step": 39900
+    },
+    {
+      "epoch": 35.89028776978417,
+      "grad_norm": 0.5750964283943176,
+      "learning_rate": 2.7647371286418238e-05,
+      "loss": 0.0166,
+      "step": 39910
+    },
+    {
+      "action_loss": 0.0016528418054804206,
+      "epoch": 35.89028776978417,
+      "step": 39910
+    },
+    {
+      "epoch": 35.89028776978417,
+      "step": 39910,
+      "torque_loss": 0.09385880082845688
+    },
+    {
+      "epoch": 35.89028776978417,
+      "force_loss": 0.0024199041072279215,
+      "step": 39910
+    },
+    {
+      "epoch": 35.89928057553957,
+      "grad_norm": 0.2214822769165039,
+      "learning_rate": 2.762272399179639e-05,
+      "loss": 0.0154,
+      "step": 39920
+    },
+    {
+      "action_loss": 0.006354435812681913,
+      "epoch": 35.89928057553957,
+      "step": 39920
+    },
+    {
+      "epoch": 35.89928057553957,
+      "step": 39920,
+      "torque_loss": 0.14036713540554047
+    },
+    {
+      "epoch": 35.89928057553957,
+      "force_loss": 0.007456177845597267,
+      "step": 39920
+    },
+    {
+      "epoch": 35.90827338129496,
+      "grad_norm": 0.2445472776889801,
+      "learning_rate": 2.7598083494800154e-05,
+      "loss": 0.0155,
+      "step": 39930
+    },
+    {
+      "action_loss": 0.007851445116102695,
+      "epoch": 35.90827338129496,
+      "step": 39930
+    },
+    {
+      "epoch": 35.90827338129496,
+      "step": 39930,
+      "torque_loss": 0.19601039588451385
+    },
+    {
+      "epoch": 35.90827338129496,
+      "force_loss": 0.0154436519369483,
+      "step": 39930
+    },
+    {
+      "epoch": 35.91726618705036,
+      "grad_norm": 0.5414363741874695,
+      "learning_rate": 2.7573449802914664e-05,
+      "loss": 0.0186,
+      "step": 39940
+    },
+    {
+      "action_loss": 0.0016132229939103127,
+      "epoch": 35.91726618705036,
+      "step": 39940
+    },
+    {
+      "epoch": 35.91726618705036,
+      "step": 39940,
+      "torque_loss": 0.10824790596961975
+    },
+    {
+      "epoch": 35.91726618705036,
+      "force_loss": 0.001795265474356711,
+      "step": 39940
+    },
+    {
+      "epoch": 35.92625899280576,
+      "grad_norm": 0.31710657477378845,
+      "learning_rate": 2.7548822923622964e-05,
+      "loss": 0.0155,
+      "step": 39950
+    },
+    {
+      "action_loss": 0.0017026454443112016,
+      "epoch": 35.92625899280576,
+      "step": 39950
+    },
+    {
+      "epoch": 35.92625899280576,
+      "step": 39950,
+      "torque_loss": 0.0875890925526619
+    },
+    {
+      "epoch": 35.92625899280576,
+      "force_loss": 0.006363121792674065,
+      "step": 39950
+    },
+    {
+      "epoch": 35.935251798561154,
+      "grad_norm": 0.09589733928442001,
+      "learning_rate": 2.752420286440609e-05,
+      "loss": 0.0157,
+      "step": 39960
+    },
+    {
+      "action_loss": 0.0011475059436634183,
+      "epoch": 35.935251798561154,
+      "step": 39960
+    },
+    {
+      "epoch": 35.935251798561154,
+      "step": 39960,
+      "torque_loss": 0.1054466962814331
+    },
+    {
+      "epoch": 35.935251798561154,
+      "force_loss": 0.0019890100229531527,
+      "step": 39960
+    },
+    {
+      "epoch": 35.944244604316545,
+      "grad_norm": 0.09964091330766678,
+      "learning_rate": 2.749958963274295e-05,
+      "loss": 0.0169,
+      "step": 39970
+    },
+    {
+      "action_loss": 0.00537843769416213,
+      "epoch": 35.944244604316545,
+      "step": 39970
+    },
+    {
+      "epoch": 35.944244604316545,
+      "step": 39970,
+      "torque_loss": 0.1167064681649208
+    },
+    {
+      "epoch": 35.944244604316545,
+      "force_loss": 0.012418483383953571,
+      "step": 39970
+    },
+    {
+      "epoch": 35.95323741007194,
+      "grad_norm": 0.2967495024204254,
+      "learning_rate": 2.747498323611039e-05,
+      "loss": 0.0178,
+      "step": 39980
+    },
+    {
+      "action_loss": 0.0007548278081230819,
+      "epoch": 35.95323741007194,
+      "step": 39980
+    },
+    {
+      "epoch": 35.95323741007194,
+      "step": 39980,
+      "torque_loss": 0.09171827882528305
+    },
+    {
+      "epoch": 35.95323741007194,
+      "force_loss": 0.0007295326213352382,
+      "step": 39980
+    },
+    {
+      "epoch": 35.96223021582734,
+      "grad_norm": 0.14533080160617828,
+      "learning_rate": 2.7450383681983184e-05,
+      "loss": 0.0152,
+      "step": 39990
+    },
+    {
+      "action_loss": 0.0024219213519245386,
+      "epoch": 35.96223021582734,
+      "step": 39990
+    },
+    {
+      "epoch": 35.96223021582734,
+      "step": 39990,
+      "torque_loss": 0.17060492932796478
+    },
+    {
+      "epoch": 35.96223021582734,
+      "force_loss": 0.0016523669473826885,
+      "step": 39990
+    },
+    {
+      "epoch": 35.97122302158273,
+      "grad_norm": 0.31310293078422546,
+      "learning_rate": 2.742579097783403e-05,
+      "loss": 0.0163,
+      "step": 40000
+    },
+    {
+      "action_loss": 0.0009725075215101242,
+      "epoch": 35.97122302158273,
+      "step": 40000
+    },
+    {
+      "epoch": 35.97122302158273,
+      "step": 40000,
+      "torque_loss": 0.09078284353017807
+    },
+    {
+      "epoch": 35.97122302158273,
+      "force_loss": 0.0020989449694752693,
+      "step": 40000
+    },
+    {
+      "epoch": 35.98021582733813,
+      "grad_norm": 0.13676171004772186,
+      "learning_rate": 2.7401205131133512e-05,
+      "loss": 0.0155,
+      "step": 40010
+    },
+    {
+      "action_loss": 0.001372501370497048,
+      "epoch": 35.98021582733813,
+      "step": 40010
+    },
+    {
+      "epoch": 35.98021582733813,
+      "step": 40010,
+      "torque_loss": 0.11363661289215088
+    },
+    {
+      "epoch": 35.98021582733813,
+      "force_loss": 0.00143437460064888,
+      "step": 40010
+    },
+    {
+      "epoch": 35.989208633093526,
+      "grad_norm": 0.29066896438598633,
+      "learning_rate": 2.7376626149350238e-05,
+      "loss": 0.0176,
+      "step": 40020
+    },
+    {
+      "action_loss": 0.003074265317991376,
+      "epoch": 35.989208633093526,
+      "step": 40020
+    },
+    {
+      "epoch": 35.989208633093526,
+      "step": 40020,
+      "torque_loss": 0.1350867748260498
+    },
+    {
+      "epoch": 35.989208633093526,
+      "force_loss": 0.00463194539770484,
+      "step": 40020
+    },
+    {
+      "epoch": 35.99820143884892,
+      "grad_norm": 0.1654406487941742,
+      "learning_rate": 2.735205403995056e-05,
+      "loss": 0.0168,
+      "step": 40030
+    },
+    {
+      "action_loss": 0.0019327859627082944,
+      "epoch": 35.99820143884892,
+      "step": 40030
+    },
+    {
+      "epoch": 35.99820143884892,
+      "step": 40030,
+      "torque_loss": 0.1391221284866333
+    },
+    {
+      "epoch": 35.99820143884892,
+      "force_loss": 0.0014563221484422684,
+      "step": 40030
+    },
+    {
+      "epoch": 36.007194244604314,
+      "grad_norm": 0.156697615981102,
+      "learning_rate": 2.7327488810398917e-05,
+      "loss": 0.0148,
+      "step": 40040
+    },
+    {
+      "action_loss": 0.0022141123190522194,
+      "epoch": 36.007194244604314,
+      "step": 40040
+    },
+    {
+      "epoch": 36.007194244604314,
+      "step": 40040,
+      "torque_loss": 0.1411222219467163
+    },
+    {
+      "epoch": 36.007194244604314,
+      "force_loss": 0.011481490917503834,
+      "step": 40040
+    },
+    {
+      "epoch": 36.01618705035971,
+      "grad_norm": 0.168813094496727,
+      "learning_rate": 2.7302930468157507e-05,
+      "loss": 0.0168,
+      "step": 40050
+    },
+    {
+      "action_loss": 0.004035875666886568,
+      "epoch": 36.01618705035971,
+      "step": 40050
+    },
+    {
+      "epoch": 36.01618705035971,
+      "step": 40050,
+      "torque_loss": 0.09257396310567856
+    },
+    {
+      "epoch": 36.01618705035971,
+      "force_loss": 0.009550203569233418,
+      "step": 40050
+    },
+    {
+      "epoch": 36.02517985611511,
+      "grad_norm": 0.41819918155670166,
+      "learning_rate": 2.727837902068655e-05,
+      "loss": 0.017,
+      "step": 40060
+    },
+    {
+      "action_loss": 0.0009752670302987099,
+      "epoch": 36.02517985611511,
+      "step": 40060
+    },
+    {
+      "epoch": 36.02517985611511,
+      "step": 40060,
+      "torque_loss": 0.11066470295190811
+    },
+    {
+      "epoch": 36.02517985611511,
+      "force_loss": 0.0011010640300810337,
+      "step": 40060
+    },
+    {
+      "epoch": 36.03417266187051,
+      "grad_norm": 0.15850377082824707,
+      "learning_rate": 2.7253834475444123e-05,
+      "loss": 0.016,
+      "step": 40070
+    },
+    {
+      "action_loss": 0.0011355552123859525,
+      "epoch": 36.03417266187051,
+      "step": 40070
+    },
+    {
+      "epoch": 36.03417266187051,
+      "step": 40070,
+      "torque_loss": 0.14427512884140015
+    },
+    {
+      "epoch": 36.03417266187051,
+      "force_loss": 0.0008650435484014452,
+      "step": 40070
+    },
+    {
+      "epoch": 36.0431654676259,
+      "grad_norm": 0.23034508526325226,
+      "learning_rate": 2.7229296839886204e-05,
+      "loss": 0.0189,
+      "step": 40080
+    },
+    {
+      "action_loss": 0.001957319676876068,
+      "epoch": 36.0431654676259,
+      "step": 40080
+    },
+    {
+      "epoch": 36.0431654676259,
+      "step": 40080,
+      "torque_loss": 0.12914668023586273
+    },
+    {
+      "epoch": 36.0431654676259,
+      "force_loss": 0.0027458462864160538,
+      "step": 40080
+    },
+    {
+      "epoch": 36.052158273381295,
+      "grad_norm": 0.14659057557582855,
+      "learning_rate": 2.720476612146668e-05,
+      "loss": 0.0149,
+      "step": 40090
+    },
+    {
+      "action_loss": 0.0009491958771832287,
+      "epoch": 36.052158273381295,
+      "step": 40090
+    },
+    {
+      "epoch": 36.052158273381295,
+      "step": 40090,
+      "torque_loss": 0.11084743589162827
+    },
+    {
+      "epoch": 36.052158273381295,
+      "force_loss": 0.0010933068115264177,
+      "step": 40090
+    },
+    {
+      "epoch": 36.06115107913669,
+      "grad_norm": 0.49214720726013184,
+      "learning_rate": 2.7180242327637317e-05,
+      "loss": 0.0167,
+      "step": 40100
+    },
+    {
+      "action_loss": 0.0015685424441471696,
+      "epoch": 36.06115107913669,
+      "step": 40100
+    },
+    {
+      "epoch": 36.06115107913669,
+      "step": 40100,
+      "torque_loss": 0.13770383596420288
+    },
+    {
+      "epoch": 36.06115107913669,
+      "force_loss": 0.001006570179015398,
+      "step": 40100
+    },
+    {
+      "epoch": 36.07014388489208,
+      "grad_norm": 0.18703696131706238,
+      "learning_rate": 2.7155725465847826e-05,
+      "loss": 0.0204,
+      "step": 40110
+    },
+    {
+      "action_loss": 0.0051618232391774654,
+      "epoch": 36.07014388489208,
+      "step": 40110
+    },
+    {
+      "epoch": 36.07014388489208,
+      "step": 40110,
+      "torque_loss": 0.13218457996845245
+    },
+    {
+      "epoch": 36.07014388489208,
+      "force_loss": 0.004874333273619413,
+      "step": 40110
+    },
+    {
+      "epoch": 36.07913669064748,
+      "grad_norm": 0.14812196791172028,
+      "learning_rate": 2.713121554354578e-05,
+      "loss": 0.0157,
+      "step": 40120
+    },
+    {
+      "action_loss": 0.0013802592875435948,
+      "epoch": 36.07913669064748,
+      "step": 40120
+    },
+    {
+      "epoch": 36.07913669064748,
+      "step": 40120,
+      "torque_loss": 0.07893958687782288
+    },
+    {
+      "epoch": 36.07913669064748,
+      "force_loss": 0.0012137488229200244,
+      "step": 40120
+    },
+    {
+      "epoch": 36.08812949640288,
+      "grad_norm": 0.1326134204864502,
+      "learning_rate": 2.7106712568176628e-05,
+      "loss": 0.0173,
+      "step": 40130
+    },
+    {
+      "action_loss": 0.002022692235186696,
+      "epoch": 36.08812949640288,
+      "step": 40130
+    },
+    {
+      "epoch": 36.08812949640288,
+      "step": 40130,
+      "torque_loss": 0.11246875673532486
+    },
+    {
+      "epoch": 36.08812949640288,
+      "force_loss": 0.006079592276364565,
+      "step": 40130
+    },
+    {
+      "epoch": 36.097122302158276,
+      "grad_norm": 0.3330748975276947,
+      "learning_rate": 2.708221654718374e-05,
+      "loss": 0.0169,
+      "step": 40140
+    },
+    {
+      "action_loss": 0.007756896782666445,
+      "epoch": 36.097122302158276,
+      "step": 40140
+    },
+    {
+      "epoch": 36.097122302158276,
+      "step": 40140,
+      "torque_loss": 0.12113672494888306
+    },
+    {
+      "epoch": 36.097122302158276,
+      "force_loss": 0.008284587413072586,
+      "step": 40140
+    },
+    {
+      "epoch": 36.106115107913666,
+      "grad_norm": 0.1462162584066391,
+      "learning_rate": 2.7057727488008357e-05,
+      "loss": 0.0166,
+      "step": 40150
+    },
+    {
+      "action_loss": 0.0011988849146291614,
+      "epoch": 36.106115107913666,
+      "step": 40150
+    },
+    {
+      "epoch": 36.106115107913666,
+      "step": 40150,
+      "torque_loss": 0.1106100007891655
+    },
+    {
+      "epoch": 36.106115107913666,
+      "force_loss": 0.00159476010594517,
+      "step": 40150
+    },
+    {
+      "epoch": 36.115107913669064,
+      "grad_norm": 0.10286112874746323,
+      "learning_rate": 2.703324539808961e-05,
+      "loss": 0.0152,
+      "step": 40160
+    },
+    {
+      "action_loss": 0.0011922478443011642,
+      "epoch": 36.115107913669064,
+      "step": 40160
+    },
+    {
+      "epoch": 36.115107913669064,
+      "step": 40160,
+      "torque_loss": 0.09055280685424805
+    },
+    {
+      "epoch": 36.115107913669064,
+      "force_loss": 0.0021607305388897657,
+      "step": 40160
+    },
+    {
+      "epoch": 36.12410071942446,
+      "grad_norm": 0.14139248430728912,
+      "learning_rate": 2.7008770284864505e-05,
+      "loss": 0.0149,
+      "step": 40170
+    },
+    {
+      "action_loss": 0.0053743403404951096,
+      "epoch": 36.12410071942446,
+      "step": 40170
+    },
+    {
+      "epoch": 36.12410071942446,
+      "step": 40170,
+      "torque_loss": 0.16901345551013947
+    },
+    {
+      "epoch": 36.12410071942446,
+      "force_loss": 0.005718277767300606,
+      "step": 40170
+    },
+    {
+      "epoch": 36.13309352517986,
+      "grad_norm": 0.19509971141815186,
+      "learning_rate": 2.6984302155767916e-05,
+      "loss": 0.0171,
+      "step": 40180
+    },
+    {
+      "action_loss": 0.0008323395741172135,
+      "epoch": 36.13309352517986,
+      "step": 40180
+    },
+    {
+      "epoch": 36.13309352517986,
+      "step": 40180,
+      "torque_loss": 0.12096711248159409
+    },
+    {
+      "epoch": 36.13309352517986,
+      "force_loss": 0.0006870224024169147,
+      "step": 40180
+    },
+    {
+      "epoch": 36.14208633093525,
+      "grad_norm": 0.26619499921798706,
+      "learning_rate": 2.6959841018232683e-05,
+      "loss": 0.0163,
+      "step": 40190
+    },
+    {
+      "action_loss": 0.0011973791988566518,
+      "epoch": 36.14208633093525,
+      "step": 40190
+    },
+    {
+      "epoch": 36.14208633093525,
+      "step": 40190,
+      "torque_loss": 0.08380764722824097
+    },
+    {
+      "epoch": 36.14208633093525,
+      "force_loss": 0.003565125400200486,
+      "step": 40190
+    },
+    {
+      "epoch": 36.15107913669065,
+      "grad_norm": 0.35528653860092163,
+      "learning_rate": 2.693538687968937e-05,
+      "loss": 0.0159,
+      "step": 40200
+    },
+    {
+      "action_loss": 0.002721647033467889,
+      "epoch": 36.15107913669065,
+      "step": 40200
+    },
+    {
+      "epoch": 36.15107913669065,
+      "step": 40200,
+      "torque_loss": 0.13448883593082428
+    },
+    {
+      "epoch": 36.15107913669065,
+      "force_loss": 0.0056694429367780685,
+      "step": 40200
+    },
+    {
+      "epoch": 36.160071942446045,
+      "grad_norm": 0.16108831763267517,
+      "learning_rate": 2.6910939747566556e-05,
+      "loss": 0.0172,
+      "step": 40210
+    },
+    {
+      "action_loss": 0.0010758339194580913,
+      "epoch": 36.160071942446045,
+      "step": 40210
+    },
+    {
+      "epoch": 36.160071942446045,
+      "step": 40210,
+      "torque_loss": 0.12400656938552856
+    },
+    {
+      "epoch": 36.160071942446045,
+      "force_loss": 0.0024678672198206186,
+      "step": 40210
+    },
+    {
+      "epoch": 36.169064748201436,
+      "grad_norm": 0.078067846596241,
+      "learning_rate": 2.6886499629290607e-05,
+      "loss": 0.0129,
+      "step": 40220
+    },
+    {
+      "action_loss": 0.0013312241062521935,
+      "epoch": 36.169064748201436,
+      "step": 40220
+    },
+    {
+      "epoch": 36.169064748201436,
+      "step": 40220,
+      "torque_loss": 0.11408732086420059
+    },
+    {
+      "epoch": 36.169064748201436,
+      "force_loss": 0.002038830192759633,
+      "step": 40220
+    },
+    {
+      "epoch": 36.17805755395683,
+      "grad_norm": 0.061363305896520615,
+      "learning_rate": 2.6862066532285802e-05,
+      "loss": 0.0151,
+      "step": 40230
+    },
+    {
+      "action_loss": 0.0034331511706113815,
+      "epoch": 36.17805755395683,
+      "step": 40230
+    },
+    {
+      "epoch": 36.17805755395683,
+      "step": 40230,
+      "torque_loss": 0.14008234441280365
+    },
+    {
+      "epoch": 36.17805755395683,
+      "force_loss": 0.005968259181827307,
+      "step": 40230
+    },
+    {
+      "epoch": 36.18705035971223,
+      "grad_norm": 0.3427886664867401,
+      "learning_rate": 2.6837640463974262e-05,
+      "loss": 0.0157,
+      "step": 40240
+    },
+    {
+      "action_loss": 0.001401070854626596,
+      "epoch": 36.18705035971223,
+      "step": 40240
+    },
+    {
+      "epoch": 36.18705035971223,
+      "step": 40240,
+      "torque_loss": 0.13309124112129211
+    },
+    {
+      "epoch": 36.18705035971223,
+      "force_loss": 0.004553420003503561,
+      "step": 40240
+    },
+    {
+      "epoch": 36.19604316546763,
+      "grad_norm": 0.26022714376449585,
+      "learning_rate": 2.681322143177596e-05,
+      "loss": 0.0151,
+      "step": 40250
+    },
+    {
+      "action_loss": 0.0013003213098272681,
+      "epoch": 36.19604316546763,
+      "step": 40250
+    },
+    {
+      "epoch": 36.19604316546763,
+      "step": 40250,
+      "torque_loss": 0.09454932063817978
+    },
+    {
+      "epoch": 36.19604316546763,
+      "force_loss": 0.0035277120769023895,
+      "step": 40250
+    },
+    {
+      "epoch": 36.20503597122302,
+      "grad_norm": 0.20755928754806519,
+      "learning_rate": 2.678880944310882e-05,
+      "loss": 0.0166,
+      "step": 40260
+    },
+    {
+      "action_loss": 0.0049186162650585175,
+      "epoch": 36.20503597122302,
+      "step": 40260
+    },
+    {
+      "epoch": 36.20503597122302,
+      "step": 40260,
+      "torque_loss": 0.1464814394712448
+    },
+    {
+      "epoch": 36.20503597122302,
+      "force_loss": 0.0049349102191627026,
+      "step": 40260
+    },
+    {
+      "epoch": 36.21402877697842,
+      "grad_norm": 0.47558560967445374,
+      "learning_rate": 2.6764404505388474e-05,
+      "loss": 0.0188,
+      "step": 40270
+    },
+    {
+      "action_loss": 0.0011392353335395455,
+      "epoch": 36.21402877697842,
+      "step": 40270
+    },
+    {
+      "epoch": 36.21402877697842,
+      "step": 40270,
+      "torque_loss": 0.12501515448093414
+    },
+    {
+      "epoch": 36.21402877697842,
+      "force_loss": 0.003926285542547703,
+      "step": 40270
+    },
+    {
+      "epoch": 36.223021582733814,
+      "grad_norm": 0.27003738284111023,
+      "learning_rate": 2.6740006626028558e-05,
+      "loss": 0.0177,
+      "step": 40280
+    },
+    {
+      "action_loss": 0.0029524012934416533,
+      "epoch": 36.223021582733814,
+      "step": 40280
+    },
+    {
+      "epoch": 36.223021582733814,
+      "step": 40280,
+      "torque_loss": 0.13837145268917084
+    },
+    {
+      "epoch": 36.223021582733814,
+      "force_loss": 0.003736709477379918,
+      "step": 40280
+    },
+    {
+      "epoch": 36.23201438848921,
+      "grad_norm": 0.22201573848724365,
+      "learning_rate": 2.671561581244048e-05,
+      "loss": 0.0185,
+      "step": 40290
+    },
+    {
+      "action_loss": 0.0021585607901215553,
+      "epoch": 36.23201438848921,
+      "step": 40290
+    },
+    {
+      "epoch": 36.23201438848921,
+      "step": 40290,
+      "torque_loss": 0.0984804555773735
+    },
+    {
+      "epoch": 36.23201438848921,
+      "force_loss": 0.0020127766765654087,
+      "step": 40290
+    },
+    {
+      "epoch": 36.2410071942446,
+      "grad_norm": 0.2025788426399231,
+      "learning_rate": 2.6691232072033536e-05,
+      "loss": 0.0164,
+      "step": 40300
+    },
+    {
+      "action_loss": 0.0029947690200060606,
+      "epoch": 36.2410071942446,
+      "step": 40300
+    },
+    {
+      "epoch": 36.2410071942446,
+      "step": 40300,
+      "torque_loss": 0.12133028358221054
+    },
+    {
+      "epoch": 36.2410071942446,
+      "force_loss": 0.0017820410430431366,
+      "step": 40300
+    },
+    {
+      "epoch": 36.25,
+      "grad_norm": 0.22762535512447357,
+      "learning_rate": 2.6666855412214852e-05,
+      "loss": 0.0193,
+      "step": 40310
+    },
+    {
+      "action_loss": 0.004297464620321989,
+      "epoch": 36.25,
+      "step": 40310
+    },
+    {
+      "epoch": 36.25,
+      "step": 40310,
+      "torque_loss": 0.15519928932189941
+    },
+    {
+      "epoch": 36.25,
+      "force_loss": 0.003369540674611926,
+      "step": 40310
+    },
+    {
+      "epoch": 36.2589928057554,
+      "grad_norm": 0.11644995957612991,
+      "learning_rate": 2.664248584038942e-05,
+      "loss": 0.0183,
+      "step": 40320
+    },
+    {
+      "action_loss": 0.002315806457772851,
+      "epoch": 36.2589928057554,
+      "step": 40320
+    },
+    {
+      "epoch": 36.2589928057554,
+      "step": 40320,
+      "torque_loss": 0.12423001974821091
+    },
+    {
+      "epoch": 36.2589928057554,
+      "force_loss": 0.003705536248162389,
+      "step": 40320
+    },
+    {
+      "epoch": 36.26798561151079,
+      "grad_norm": 0.5651867985725403,
+      "learning_rate": 2.6618123363960047e-05,
+      "loss": 0.0149,
+      "step": 40330
+    },
+    {
+      "action_loss": 0.004449928645044565,
+      "epoch": 36.26798561151079,
+      "step": 40330
+    },
+    {
+      "epoch": 36.26798561151079,
+      "step": 40330,
+      "torque_loss": 0.12324311584234238
+    },
+    {
+      "epoch": 36.26798561151079,
+      "force_loss": 0.005421905312687159,
+      "step": 40330
+    },
+    {
+      "epoch": 36.276978417266186,
+      "grad_norm": 0.49224138259887695,
+      "learning_rate": 2.659376799032748e-05,
+      "loss": 0.0164,
+      "step": 40340
+    },
+    {
+      "action_loss": 0.0017632081871852279,
+      "epoch": 36.276978417266186,
+      "step": 40340
+    },
+    {
+      "epoch": 36.276978417266186,
+      "step": 40340,
+      "torque_loss": 0.11182568222284317
+    },
+    {
+      "epoch": 36.276978417266186,
+      "force_loss": 0.001757893362082541,
+      "step": 40340
+    },
+    {
+      "epoch": 36.28597122302158,
+      "grad_norm": 0.2579120099544525,
+      "learning_rate": 2.6569419726890145e-05,
+      "loss": 0.0204,
+      "step": 40350
+    },
+    {
+      "action_loss": 0.0043951706029474735,
+      "epoch": 36.28597122302158,
+      "step": 40350
+    },
+    {
+      "epoch": 36.28597122302158,
+      "step": 40350,
+      "torque_loss": 0.12350539118051529
+    },
+    {
+      "epoch": 36.28597122302158,
+      "force_loss": 0.003934036009013653,
+      "step": 40350
+    },
+    {
+      "epoch": 36.29496402877698,
+      "grad_norm": 0.2595391571521759,
+      "learning_rate": 2.654507858104447e-05,
+      "loss": 0.0191,
+      "step": 40360
+    },
+    {
+      "action_loss": 0.003790018381550908,
+      "epoch": 36.29496402877698,
+      "step": 40360
+    },
+    {
+      "epoch": 36.29496402877698,
+      "step": 40360,
+      "torque_loss": 0.13111385703086853
+    },
+    {
+      "epoch": 36.29496402877698,
+      "force_loss": 0.00280355685390532,
+      "step": 40360
+    },
+    {
+      "epoch": 36.30395683453237,
+      "grad_norm": 0.07278679311275482,
+      "learning_rate": 2.652074456018463e-05,
+      "loss": 0.0155,
+      "step": 40370
+    },
+    {
+      "action_loss": 0.0023485682904720306,
+      "epoch": 36.30395683453237,
+      "step": 40370
+    },
+    {
+      "epoch": 36.30395683453237,
+      "step": 40370,
+      "torque_loss": 0.22213876247406006
+    },
+    {
+      "epoch": 36.30395683453237,
+      "force_loss": 0.0021585547365248203,
+      "step": 40370
+    },
+    {
+      "epoch": 36.31294964028777,
+      "grad_norm": 0.27411389350891113,
+      "learning_rate": 2.6496417671702646e-05,
+      "loss": 0.017,
+      "step": 40380
+    },
+    {
+      "action_loss": 0.0031113314907997847,
+      "epoch": 36.31294964028777,
+      "step": 40380
+    },
+    {
+      "epoch": 36.31294964028777,
+      "step": 40380,
+      "torque_loss": 0.11473411321640015
+    },
+    {
+      "epoch": 36.31294964028777,
+      "force_loss": 0.003830675734207034,
+      "step": 40380
+    },
+    {
+      "epoch": 36.32194244604317,
+      "grad_norm": 0.1259770393371582,
+      "learning_rate": 2.6472097922988427e-05,
+      "loss": 0.0152,
+      "step": 40390
+    },
+    {
+      "action_loss": 0.014899145811796188,
+      "epoch": 36.32194244604317,
+      "step": 40390
+    },
+    {
+      "epoch": 36.32194244604317,
+      "step": 40390,
+      "torque_loss": 0.14831231534481049
+    },
+    {
+      "epoch": 36.32194244604317,
+      "force_loss": 0.028943797573447227,
+      "step": 40390
+    },
+    {
+      "epoch": 36.330935251798564,
+      "grad_norm": 0.10508344322443008,
+      "learning_rate": 2.6447785321429607e-05,
+      "loss": 0.0306,
+      "step": 40400
+    },
+    {
+      "action_loss": 0.0018155038123950362,
+      "epoch": 36.330935251798564,
+      "step": 40400
+    },
+    {
+      "epoch": 36.330935251798564,
+      "step": 40400,
+      "torque_loss": 0.1340511590242386
+    },
+    {
+      "epoch": 36.330935251798564,
+      "force_loss": 0.0014119663974270225,
+      "step": 40400
+    },
+    {
+      "epoch": 36.339928057553955,
+      "grad_norm": 0.14072927832603455,
+      "learning_rate": 2.6423479874411784e-05,
+      "loss": 0.0186,
+      "step": 40410
+    },
+    {
+      "action_loss": 0.002754220739006996,
+      "epoch": 36.339928057553955,
+      "step": 40410
+    },
+    {
+      "epoch": 36.339928057553955,
+      "step": 40410,
+      "torque_loss": 0.13490955531597137
+    },
+    {
+      "epoch": 36.339928057553955,
+      "force_loss": 0.007847055792808533,
+      "step": 40410
+    },
+    {
+      "epoch": 36.34892086330935,
+      "grad_norm": 0.27140411734580994,
+      "learning_rate": 2.6399181589318234e-05,
+      "loss": 0.0149,
+      "step": 40420
+    },
+    {
+      "action_loss": 0.0016590593149885535,
+      "epoch": 36.34892086330935,
+      "step": 40420
+    },
+    {
+      "epoch": 36.34892086330935,
+      "step": 40420,
+      "torque_loss": 0.1119220182299614
+    },
+    {
+      "epoch": 36.34892086330935,
+      "force_loss": 0.00261204713024199,
+      "step": 40420
+    },
+    {
+      "epoch": 36.35791366906475,
+      "grad_norm": 0.1714809089899063,
+      "learning_rate": 2.6374890473530188e-05,
+      "loss": 0.015,
+      "step": 40430
+    },
+    {
+      "action_loss": 0.002275634091347456,
+      "epoch": 36.35791366906475,
+      "step": 40430
+    },
+    {
+      "epoch": 36.35791366906475,
+      "step": 40430,
+      "torque_loss": 0.1370101273059845
+    },
+    {
+      "epoch": 36.35791366906475,
+      "force_loss": 0.007402322720736265,
+      "step": 40430
+    },
+    {
+      "epoch": 36.36690647482014,
+      "grad_norm": 0.07405562698841095,
+      "learning_rate": 2.635060653442664e-05,
+      "loss": 0.0183,
+      "step": 40440
+    },
+    {
+      "action_loss": 0.0008310635457746685,
+      "epoch": 36.36690647482014,
+      "step": 40440
+    },
+    {
+      "epoch": 36.36690647482014,
+      "step": 40440,
+      "torque_loss": 0.11837742477655411
+    },
+    {
+      "epoch": 36.36690647482014,
+      "force_loss": 0.0009524838533252478,
+      "step": 40440
+    },
+    {
+      "epoch": 36.37589928057554,
+      "grad_norm": 0.1030859500169754,
+      "learning_rate": 2.6326329779384395e-05,
+      "loss": 0.0165,
+      "step": 40450
+    },
+    {
+      "action_loss": 0.004299999680370092,
+      "epoch": 36.37589928057554,
+      "step": 40450
+    },
+    {
+      "epoch": 36.37589928057554,
+      "step": 40450,
+      "torque_loss": 0.11847180873155594
+    },
+    {
+      "epoch": 36.37589928057554,
+      "force_loss": 0.00947362557053566,
+      "step": 40450
+    },
+    {
+      "epoch": 36.384892086330936,
+      "grad_norm": 0.15667304396629333,
+      "learning_rate": 2.63020602157781e-05,
+      "loss": 0.0152,
+      "step": 40460
+    },
+    {
+      "action_loss": 0.0012257518246769905,
+      "epoch": 36.384892086330936,
+      "step": 40460
+    },
+    {
+      "epoch": 36.384892086330936,
+      "step": 40460,
+      "torque_loss": 0.1009514257311821
+    },
+    {
+      "epoch": 36.384892086330936,
+      "force_loss": 0.0029828825499862432,
+      "step": 40460
+    },
+    {
+      "epoch": 36.393884892086334,
+      "grad_norm": 0.11764544248580933,
+      "learning_rate": 2.62777978509802e-05,
+      "loss": 0.0169,
+      "step": 40470
+    },
+    {
+      "action_loss": 0.0022348554339259863,
+      "epoch": 36.393884892086334,
+      "step": 40470
+    },
+    {
+      "epoch": 36.393884892086334,
+      "step": 40470,
+      "torque_loss": 0.11265590041875839
+    },
+    {
+      "epoch": 36.393884892086334,
+      "force_loss": 0.003278246382251382,
+      "step": 40470
+    },
+    {
+      "epoch": 36.402877697841724,
+      "grad_norm": 0.10085401684045792,
+      "learning_rate": 2.6253542692360954e-05,
+      "loss": 0.0178,
+      "step": 40480
+    },
+    {
+      "action_loss": 0.0021885985042899847,
+      "epoch": 36.402877697841724,
+      "step": 40480
+    },
+    {
+      "epoch": 36.402877697841724,
+      "step": 40480,
+      "torque_loss": 0.09116091579198837
+    },
+    {
+      "epoch": 36.402877697841724,
+      "force_loss": 0.003922689240425825,
+      "step": 40480
+    },
+    {
+      "epoch": 36.41187050359712,
+      "grad_norm": 0.3292659521102905,
+      "learning_rate": 2.6229294747288458e-05,
+      "loss": 0.0139,
+      "step": 40490
+    },
+    {
+      "action_loss": 0.0016092568403109908,
+      "epoch": 36.41187050359712,
+      "step": 40490
+    },
+    {
+      "epoch": 36.41187050359712,
+      "step": 40490,
+      "torque_loss": 0.10183060169219971
+    },
+    {
+      "epoch": 36.41187050359712,
+      "force_loss": 0.0010820940369740129,
+      "step": 40490
+    },
+    {
+      "epoch": 36.42086330935252,
+      "grad_norm": 0.06739940494298935,
+      "learning_rate": 2.6205054023128596e-05,
+      "loss": 0.0138,
+      "step": 40500
+    },
+    {
+      "action_loss": 0.0036087885964661837,
+      "epoch": 36.42086330935252,
+      "step": 40500
+    },
+    {
+      "epoch": 36.42086330935252,
+      "step": 40500,
+      "torque_loss": 0.10821110755205154
+    },
+    {
+      "epoch": 36.42086330935252,
+      "force_loss": 0.004196793772280216,
+      "step": 40500
+    },
+    {
+      "epoch": 36.42985611510792,
+      "grad_norm": 0.25225770473480225,
+      "learning_rate": 2.6180820527245043e-05,
+      "loss": 0.0191,
+      "step": 40510
+    },
+    {
+      "action_loss": 0.0015113676199689507,
+      "epoch": 36.42985611510792,
+      "step": 40510
+    },
+    {
+      "epoch": 36.42985611510792,
+      "step": 40510,
+      "torque_loss": 0.1375342458486557
+    },
+    {
+      "epoch": 36.42985611510792,
+      "force_loss": 0.0015819409163668752,
+      "step": 40510
+    },
+    {
+      "epoch": 36.43884892086331,
+      "grad_norm": 0.23361550271511078,
+      "learning_rate": 2.6156594266999313e-05,
+      "loss": 0.0168,
+      "step": 40520
+    },
+    {
+      "action_loss": 0.004073174670338631,
+      "epoch": 36.43884892086331,
+      "step": 40520
+    },
+    {
+      "epoch": 36.43884892086331,
+      "step": 40520,
+      "torque_loss": 0.10116708278656006
+    },
+    {
+      "epoch": 36.43884892086331,
+      "force_loss": 0.006961511448025703,
+      "step": 40520
+    },
+    {
+      "epoch": 36.447841726618705,
+      "grad_norm": 0.2864074110984802,
+      "learning_rate": 2.6132375249750672e-05,
+      "loss": 0.0172,
+      "step": 40530
+    },
+    {
+      "action_loss": 0.005145722068846226,
+      "epoch": 36.447841726618705,
+      "step": 40530
+    },
+    {
+      "epoch": 36.447841726618705,
+      "step": 40530,
+      "torque_loss": 0.07767868041992188
+    },
+    {
+      "epoch": 36.447841726618705,
+      "force_loss": 0.005683968309313059,
+      "step": 40530
+    },
+    {
+      "epoch": 36.4568345323741,
+      "grad_norm": 0.11944551765918732,
+      "learning_rate": 2.6108163482856286e-05,
+      "loss": 0.0148,
+      "step": 40540
+    },
+    {
+      "action_loss": 0.00349957006983459,
+      "epoch": 36.4568345323741,
+      "step": 40540
+    },
+    {
+      "epoch": 36.4568345323741,
+      "step": 40540,
+      "torque_loss": 0.082356758415699
+    },
+    {
+      "epoch": 36.4568345323741,
+      "force_loss": 0.003870716318488121,
+      "step": 40540
+    },
+    {
+      "epoch": 36.46582733812949,
+      "grad_norm": 0.12395649403333664,
+      "learning_rate": 2.6083958973670964e-05,
+      "loss": 0.0158,
+      "step": 40550
+    },
+    {
+      "action_loss": 0.0013828134397044778,
+      "epoch": 36.46582733812949,
+      "step": 40550
+    },
+    {
+      "epoch": 36.46582733812949,
+      "step": 40550,
+      "torque_loss": 0.09643980115652084
+    },
+    {
+      "epoch": 36.46582733812949,
+      "force_loss": 0.003292340785264969,
+      "step": 40550
+    },
+    {
+      "epoch": 36.47482014388489,
+      "grad_norm": 0.14932620525360107,
+      "learning_rate": 2.6059761729547483e-05,
+      "loss": 0.0174,
+      "step": 40560
+    },
+    {
+      "action_loss": 0.0006636045873165131,
+      "epoch": 36.47482014388489,
+      "step": 40560
+    },
+    {
+      "epoch": 36.47482014388489,
+      "step": 40560,
+      "torque_loss": 0.11179139465093613
+    },
+    {
+      "epoch": 36.47482014388489,
+      "force_loss": 0.0016114813042804599,
+      "step": 40560
+    },
+    {
+      "epoch": 36.48381294964029,
+      "grad_norm": 0.09086962789297104,
+      "learning_rate": 2.603557175783624e-05,
+      "loss": 0.0147,
+      "step": 40570
+    },
+    {
+      "action_loss": 0.0006282440153881907,
+      "epoch": 36.48381294964029,
+      "step": 40570
+    },
+    {
+      "epoch": 36.48381294964029,
+      "step": 40570,
+      "torque_loss": 0.10010053962469101
+    },
+    {
+      "epoch": 36.48381294964029,
+      "force_loss": 0.001065872493200004,
+      "step": 40570
+    },
+    {
+      "epoch": 36.492805755395686,
+      "grad_norm": 0.06441736221313477,
+      "learning_rate": 2.601138906588559e-05,
+      "loss": 0.0153,
+      "step": 40580
+    },
+    {
+      "action_loss": 0.0014755738666281104,
+      "epoch": 36.492805755395686,
+      "step": 40580
+    },
+    {
+      "epoch": 36.492805755395686,
+      "step": 40580,
+      "torque_loss": 0.130422905087471
+    },
+    {
+      "epoch": 36.492805755395686,
+      "force_loss": 0.0017960546538233757,
+      "step": 40580
+    },
+    {
+      "epoch": 36.50179856115108,
+      "grad_norm": 0.3446483612060547,
+      "learning_rate": 2.598721366104152e-05,
+      "loss": 0.0163,
+      "step": 40590
+    },
+    {
+      "action_loss": 0.001830587163567543,
+      "epoch": 36.50179856115108,
+      "step": 40590
+    },
+    {
+      "epoch": 36.50179856115108,
+      "step": 40590,
+      "torque_loss": 0.13736669719219208
+    },
+    {
+      "epoch": 36.50179856115108,
+      "force_loss": 0.0017593413358554244,
+      "step": 40590
+    },
+    {
+      "epoch": 36.510791366906474,
+      "grad_norm": 0.09236304461956024,
+      "learning_rate": 2.5963045550647945e-05,
+      "loss": 0.0134,
+      "step": 40600
+    },
+    {
+      "action_loss": 0.0018125405767932534,
+      "epoch": 36.510791366906474,
+      "step": 40600
+    },
+    {
+      "epoch": 36.510791366906474,
+      "step": 40600,
+      "torque_loss": 0.14467743039131165
+    },
+    {
+      "epoch": 36.510791366906474,
+      "force_loss": 0.0026651702355593443,
+      "step": 40600
+    },
+    {
+      "epoch": 36.51978417266187,
+      "grad_norm": 0.27811768651008606,
+      "learning_rate": 2.5938884742046466e-05,
+      "loss": 0.015,
+      "step": 40610
+    },
+    {
+      "action_loss": 0.005231405142694712,
+      "epoch": 36.51978417266187,
+      "step": 40610
+    },
+    {
+      "epoch": 36.51978417266187,
+      "step": 40610,
+      "torque_loss": 0.1340021938085556
+    },
+    {
+      "epoch": 36.51978417266187,
+      "force_loss": 0.006940803956240416,
+      "step": 40610
+    },
+    {
+      "epoch": 36.52877697841727,
+      "grad_norm": 0.16698452830314636,
+      "learning_rate": 2.5914731242576507e-05,
+      "loss": 0.0168,
+      "step": 40620
+    },
+    {
+      "action_loss": 0.00139683170709759,
+      "epoch": 36.52877697841727,
+      "step": 40620
+    },
+    {
+      "epoch": 36.52877697841727,
+      "step": 40620,
+      "torque_loss": 0.0994524359703064
+    },
+    {
+      "epoch": 36.52877697841727,
+      "force_loss": 0.002018711995333433,
+      "step": 40620
+    },
+    {
+      "epoch": 36.53776978417266,
+      "grad_norm": 0.1333102583885193,
+      "learning_rate": 2.5890585059575268e-05,
+      "loss": 0.0164,
+      "step": 40630
+    },
+    {
+      "action_loss": 0.0010444204090163112,
+      "epoch": 36.53776978417266,
+      "step": 40630
+    },
+    {
+      "epoch": 36.53776978417266,
+      "step": 40630,
+      "torque_loss": 0.15608525276184082
+    },
+    {
+      "epoch": 36.53776978417266,
+      "force_loss": 0.0033776445779949427,
+      "step": 40630
+    },
+    {
+      "epoch": 36.54676258992806,
+      "grad_norm": 0.13561423122882843,
+      "learning_rate": 2.5866446200377688e-05,
+      "loss": 0.0159,
+      "step": 40640
+    },
+    {
+      "action_loss": 0.0017424138495698571,
+      "epoch": 36.54676258992806,
+      "step": 40640
+    },
+    {
+      "epoch": 36.54676258992806,
+      "step": 40640,
+      "torque_loss": 0.1100485697388649
+    },
+    {
+      "epoch": 36.54676258992806,
+      "force_loss": 0.0015539852902293205,
+      "step": 40640
+    },
+    {
+      "epoch": 36.555755395683455,
+      "grad_norm": 0.33544790744781494,
+      "learning_rate": 2.5842314672316566e-05,
+      "loss": 0.017,
+      "step": 40650
+    },
+    {
+      "action_loss": 0.0023810788989067078,
+      "epoch": 36.555755395683455,
+      "step": 40650
+    },
+    {
+      "epoch": 36.555755395683455,
+      "step": 40650,
+      "torque_loss": 0.10753890126943588
+    },
+    {
+      "epoch": 36.555755395683455,
+      "force_loss": 0.005487387999892235,
+      "step": 40650
+    },
+    {
+      "epoch": 36.564748201438846,
+      "grad_norm": 0.20060555636882782,
+      "learning_rate": 2.581819048272239e-05,
+      "loss": 0.0161,
+      "step": 40660
+    },
+    {
+      "action_loss": 0.013823983259499073,
+      "epoch": 36.564748201438846,
+      "step": 40660
+    },
+    {
+      "epoch": 36.564748201438846,
+      "step": 40660,
+      "torque_loss": 0.1418100744485855
+    },
+    {
+      "epoch": 36.564748201438846,
+      "force_loss": 0.013399017043411732,
+      "step": 40660
+    },
+    {
+      "epoch": 36.57374100719424,
+      "grad_norm": 0.21801358461380005,
+      "learning_rate": 2.5794073638923478e-05,
+      "loss": 0.0199,
+      "step": 40670
+    },
+    {
+      "action_loss": 0.005062831100076437,
+      "epoch": 36.57374100719424,
+      "step": 40670
+    },
+    {
+      "epoch": 36.57374100719424,
+      "step": 40670,
+      "torque_loss": 0.13114379346370697
+    },
+    {
+      "epoch": 36.57374100719424,
+      "force_loss": 0.003218039171770215,
+      "step": 40670
+    },
+    {
+      "epoch": 36.58273381294964,
+      "grad_norm": 0.4264189898967743,
+      "learning_rate": 2.576996414824586e-05,
+      "loss": 0.0228,
+      "step": 40680
+    },
+    {
+      "action_loss": 0.0015105524798855186,
+      "epoch": 36.58273381294964,
+      "step": 40680
+    },
+    {
+      "epoch": 36.58273381294964,
+      "step": 40680,
+      "torque_loss": 0.10605400800704956
+    },
+    {
+      "epoch": 36.58273381294964,
+      "force_loss": 0.002417312702164054,
+      "step": 40680
+    },
+    {
+      "epoch": 36.59172661870504,
+      "grad_norm": 0.20488204061985016,
+      "learning_rate": 2.574586201801339e-05,
+      "loss": 0.0163,
+      "step": 40690
+    },
+    {
+      "action_loss": 0.0009720135130919516,
+      "epoch": 36.59172661870504,
+      "step": 40690
+    },
+    {
+      "epoch": 36.59172661870504,
+      "step": 40690,
+      "torque_loss": 0.1611345410346985
+    },
+    {
+      "epoch": 36.59172661870504,
+      "force_loss": 0.0007408817764371634,
+      "step": 40690
+    },
+    {
+      "epoch": 36.60071942446043,
+      "grad_norm": 0.20045778155326843,
+      "learning_rate": 2.572176725554762e-05,
+      "loss": 0.0153,
+      "step": 40700
+    },
+    {
+      "action_loss": 0.007397270295768976,
+      "epoch": 36.60071942446043,
+      "step": 40700
+    },
+    {
+      "epoch": 36.60071942446043,
+      "step": 40700,
+      "torque_loss": 0.11989260464906693
+    },
+    {
+      "epoch": 36.60071942446043,
+      "force_loss": 0.012663266621530056,
+      "step": 40700
+    },
+    {
+      "epoch": 36.60971223021583,
+      "grad_norm": 0.5078071355819702,
+      "learning_rate": 2.5697679868167966e-05,
+      "loss": 0.0179,
+      "step": 40710
+    },
+    {
+      "action_loss": 0.001644335687160492,
+      "epoch": 36.60971223021583,
+      "step": 40710
+    },
+    {
+      "epoch": 36.60971223021583,
+      "step": 40710,
+      "torque_loss": 0.13805948197841644
+    },
+    {
+      "epoch": 36.60971223021583,
+      "force_loss": 0.0016886420780792832,
+      "step": 40710
+    },
+    {
+      "epoch": 36.618705035971225,
+      "grad_norm": 0.13483642041683197,
+      "learning_rate": 2.5673599863191468e-05,
+      "loss": 0.0166,
+      "step": 40720
+    },
+    {
+      "action_loss": 0.005367635283619165,
+      "epoch": 36.618705035971225,
+      "step": 40720
+    },
+    {
+      "epoch": 36.618705035971225,
+      "step": 40720,
+      "torque_loss": 0.09979171305894852
+    },
+    {
+      "epoch": 36.618705035971225,
+      "force_loss": 0.006961230654269457,
+      "step": 40720
+    },
+    {
+      "epoch": 36.62769784172662,
+      "grad_norm": 0.22946494817733765,
+      "learning_rate": 2.564952724793306e-05,
+      "loss": 0.0143,
+      "step": 40730
+    },
+    {
+      "action_loss": 0.0017493563937023282,
+      "epoch": 36.62769784172662,
+      "step": 40730
+    },
+    {
+      "epoch": 36.62769784172662,
+      "step": 40730,
+      "torque_loss": 0.140501007437706
+    },
+    {
+      "epoch": 36.62769784172662,
+      "force_loss": 0.0019741205032914877,
+      "step": 40730
+    },
+    {
+      "epoch": 36.63669064748201,
+      "grad_norm": 0.14705729484558105,
+      "learning_rate": 2.5625462029705306e-05,
+      "loss": 0.017,
+      "step": 40740
+    },
+    {
+      "action_loss": 0.00583826145157218,
+      "epoch": 36.63669064748201,
+      "step": 40740
+    },
+    {
+      "epoch": 36.63669064748201,
+      "step": 40740,
+      "torque_loss": 0.11922410130500793
+    },
+    {
+      "epoch": 36.63669064748201,
+      "force_loss": 0.004548977129161358,
+      "step": 40740
+    },
+    {
+      "epoch": 36.64568345323741,
+      "grad_norm": 0.10758621990680695,
+      "learning_rate": 2.5601404215818624e-05,
+      "loss": 0.0157,
+      "step": 40750
+    },
+    {
+      "action_loss": 0.0021942760795354843,
+      "epoch": 36.64568345323741,
+      "step": 40750
+    },
+    {
+      "epoch": 36.64568345323741,
+      "step": 40750,
+      "torque_loss": 0.09720811992883682
+    },
+    {
+      "epoch": 36.64568345323741,
+      "force_loss": 0.0034751601051539183,
+      "step": 40750
+    },
+    {
+      "epoch": 36.65467625899281,
+      "grad_norm": 0.17540010809898376,
+      "learning_rate": 2.5577353813581144e-05,
+      "loss": 0.0158,
+      "step": 40760
+    },
+    {
+      "action_loss": 0.003847033716738224,
+      "epoch": 36.65467625899281,
+      "step": 40760
+    },
+    {
+      "epoch": 36.65467625899281,
+      "step": 40760,
+      "torque_loss": 0.12497133016586304
+    },
+    {
+      "epoch": 36.65467625899281,
+      "force_loss": 0.005627034697681665,
+      "step": 40760
+    },
+    {
+      "epoch": 36.6636690647482,
+      "grad_norm": 0.3054220378398895,
+      "learning_rate": 2.5553310830298733e-05,
+      "loss": 0.0198,
+      "step": 40770
+    },
+    {
+      "action_loss": 0.012451942078769207,
+      "epoch": 36.6636690647482,
+      "step": 40770
+    },
+    {
+      "epoch": 36.6636690647482,
+      "step": 40770,
+      "torque_loss": 0.1235266700387001
+    },
+    {
+      "epoch": 36.6636690647482,
+      "force_loss": 0.016896799206733704,
+      "step": 40770
+    },
+    {
+      "epoch": 36.672661870503596,
+      "grad_norm": 0.16211777925491333,
+      "learning_rate": 2.5529275273275012e-05,
+      "loss": 0.015,
+      "step": 40780
+    },
+    {
+      "action_loss": 0.009844637475907803,
+      "epoch": 36.672661870503596,
+      "step": 40780
+    },
+    {
+      "epoch": 36.672661870503596,
+      "step": 40780,
+      "torque_loss": 0.10241729021072388
+    },
+    {
+      "epoch": 36.672661870503596,
+      "force_loss": 0.007657181471586227,
+      "step": 40780
+    },
+    {
+      "epoch": 36.681654676258994,
+      "grad_norm": 0.2554575800895691,
+      "learning_rate": 2.550524714981133e-05,
+      "loss": 0.0145,
+      "step": 40790
+    },
+    {
+      "action_loss": 0.0008564687450416386,
+      "epoch": 36.681654676258994,
+      "step": 40790
+    },
+    {
+      "epoch": 36.681654676258994,
+      "step": 40790,
+      "torque_loss": 0.12012094259262085
+    },
+    {
+      "epoch": 36.681654676258994,
+      "force_loss": 0.001241430756635964,
+      "step": 40790
+    },
+    {
+      "epoch": 36.69064748201439,
+      "grad_norm": 0.07543785870075226,
+      "learning_rate": 2.5481226467206837e-05,
+      "loss": 0.0155,
+      "step": 40800
+    },
+    {
+      "action_loss": 0.0014604610623791814,
+      "epoch": 36.69064748201439,
+      "step": 40800
+    },
+    {
+      "epoch": 36.69064748201439,
+      "step": 40800,
+      "torque_loss": 0.14680027961730957
+    },
+    {
+      "epoch": 36.69064748201439,
+      "force_loss": 0.0021571898832917213,
+      "step": 40800
+    },
+    {
+      "epoch": 36.69964028776978,
+      "grad_norm": 0.09962771087884903,
+      "learning_rate": 2.5457213232758365e-05,
+      "loss": 0.0146,
+      "step": 40810
+    },
+    {
+      "action_loss": 0.004125829786062241,
+      "epoch": 36.69964028776978,
+      "step": 40810
+    },
+    {
+      "epoch": 36.69964028776978,
+      "step": 40810,
+      "torque_loss": 0.1400660276412964
+    },
+    {
+      "epoch": 36.69964028776978,
+      "force_loss": 0.009941665455698967,
+      "step": 40810
+    },
+    {
+      "epoch": 36.70863309352518,
+      "grad_norm": 0.12375088781118393,
+      "learning_rate": 2.5433207453760498e-05,
+      "loss": 0.0206,
+      "step": 40820
+    },
+    {
+      "action_loss": 0.0012120419414713979,
+      "epoch": 36.70863309352518,
+      "step": 40820
+    },
+    {
+      "epoch": 36.70863309352518,
+      "step": 40820,
+      "torque_loss": 0.1398182511329651
+    },
+    {
+      "epoch": 36.70863309352518,
+      "force_loss": 0.0014342637732625008,
+      "step": 40820
+    },
+    {
+      "epoch": 36.71762589928058,
+      "grad_norm": 0.22068969905376434,
+      "learning_rate": 2.5409209137505552e-05,
+      "loss": 0.0167,
+      "step": 40830
+    },
+    {
+      "action_loss": 0.0047895521856844425,
+      "epoch": 36.71762589928058,
+      "step": 40830
+    },
+    {
+      "epoch": 36.71762589928058,
+      "step": 40830,
+      "torque_loss": 0.1456289291381836
+    },
+    {
+      "epoch": 36.71762589928058,
+      "force_loss": 0.008335483260452747,
+      "step": 40830
+    },
+    {
+      "epoch": 36.726618705035975,
+      "grad_norm": 0.19960246980190277,
+      "learning_rate": 2.5385218291283597e-05,
+      "loss": 0.015,
+      "step": 40840
+    },
+    {
+      "action_loss": 0.00313937128521502,
+      "epoch": 36.726618705035975,
+      "step": 40840
+    },
+    {
+      "epoch": 36.726618705035975,
+      "step": 40840,
+      "torque_loss": 0.19083677232265472
+    },
+    {
+      "epoch": 36.726618705035975,
+      "force_loss": 0.002062875544652343,
+      "step": 40840
+    },
+    {
+      "epoch": 36.735611510791365,
+      "grad_norm": 0.11217894405126572,
+      "learning_rate": 2.5361234922382383e-05,
+      "loss": 0.0184,
+      "step": 40850
+    },
+    {
+      "action_loss": 0.0013308579800650477,
+      "epoch": 36.735611510791365,
+      "step": 40850
+    },
+    {
+      "epoch": 36.735611510791365,
+      "step": 40850,
+      "torque_loss": 0.11408674716949463
+    },
+    {
+      "epoch": 36.735611510791365,
+      "force_loss": 0.002995135262608528,
+      "step": 40850
+    },
+    {
+      "epoch": 36.74460431654676,
+      "grad_norm": 0.15857085585594177,
+      "learning_rate": 2.533725903808749e-05,
+      "loss": 0.0162,
+      "step": 40860
+    },
+    {
+      "action_loss": 0.0016767798224464059,
+      "epoch": 36.74460431654676,
+      "step": 40860
+    },
+    {
+      "epoch": 36.74460431654676,
+      "step": 40860,
+      "torque_loss": 0.1243806704878807
+    },
+    {
+      "epoch": 36.74460431654676,
+      "force_loss": 0.0010130471782758832,
+      "step": 40860
+    },
+    {
+      "epoch": 36.75359712230216,
+      "grad_norm": 0.17636162042617798,
+      "learning_rate": 2.5313290645682085e-05,
+      "loss": 0.0178,
+      "step": 40870
+    },
+    {
+      "action_loss": 0.006714515388011932,
+      "epoch": 36.75359712230216,
+      "step": 40870
+    },
+    {
+      "epoch": 36.75359712230216,
+      "step": 40870,
+      "torque_loss": 0.13070610165596008
+    },
+    {
+      "epoch": 36.75359712230216,
+      "force_loss": 0.011135063134133816,
+      "step": 40870
+    },
+    {
+      "epoch": 36.76258992805755,
+      "grad_norm": 0.16088257730007172,
+      "learning_rate": 2.52893297524472e-05,
+      "loss": 0.0183,
+      "step": 40880
+    },
+    {
+      "action_loss": 0.006263503339141607,
+      "epoch": 36.76258992805755,
+      "step": 40880
+    },
+    {
+      "epoch": 36.76258992805755,
+      "step": 40880,
+      "torque_loss": 0.07483164221048355
+    },
+    {
+      "epoch": 36.76258992805755,
+      "force_loss": 0.007079185452312231,
+      "step": 40880
+    },
+    {
+      "epoch": 36.77158273381295,
+      "grad_norm": 0.31436511874198914,
+      "learning_rate": 2.526537636566145e-05,
+      "loss": 0.0137,
+      "step": 40890
+    },
+    {
+      "action_loss": 0.0019651756156235933,
+      "epoch": 36.77158273381295,
+      "step": 40890
+    },
+    {
+      "epoch": 36.77158273381295,
+      "step": 40890,
+      "torque_loss": 0.1034235954284668
+    },
+    {
+      "epoch": 36.77158273381295,
+      "force_loss": 0.005370249506086111,
+      "step": 40890
+    },
+    {
+      "epoch": 36.780575539568346,
+      "grad_norm": 0.13597896695137024,
+      "learning_rate": 2.5241430492601305e-05,
+      "loss": 0.0169,
+      "step": 40900
+    },
+    {
+      "action_loss": 0.0029434890020638704,
+      "epoch": 36.780575539568346,
+      "step": 40900
+    },
+    {
+      "epoch": 36.780575539568346,
+      "step": 40900,
+      "torque_loss": 0.12355612963438034
+    },
+    {
+      "epoch": 36.780575539568346,
+      "force_loss": 0.0021419788245111704,
+      "step": 40900
+    },
+    {
+      "epoch": 36.789568345323744,
+      "grad_norm": 0.11905980110168457,
+      "learning_rate": 2.5217492140540867e-05,
+      "loss": 0.0163,
+      "step": 40910
+    },
+    {
+      "action_loss": 0.0007639930699951947,
+      "epoch": 36.789568345323744,
+      "step": 40910
+    },
+    {
+      "epoch": 36.789568345323744,
+      "step": 40910,
+      "torque_loss": 0.11611571162939072
+    },
+    {
+      "epoch": 36.789568345323744,
+      "force_loss": 0.000702633464243263,
+      "step": 40910
+    },
+    {
+      "epoch": 36.798561151079134,
+      "grad_norm": 0.06944193691015244,
+      "learning_rate": 2.5193561316751967e-05,
+      "loss": 0.0164,
+      "step": 40920
+    },
+    {
+      "action_loss": 0.0014789592241868377,
+      "epoch": 36.798561151079134,
+      "step": 40920
+    },
+    {
+      "epoch": 36.798561151079134,
+      "step": 40920,
+      "torque_loss": 0.13382190465927124
+    },
+    {
+      "epoch": 36.798561151079134,
+      "force_loss": 0.00314895436167717,
+      "step": 40920
+    },
+    {
+      "epoch": 36.80755395683453,
+      "grad_norm": 0.09905727207660675,
+      "learning_rate": 2.516963802850416e-05,
+      "loss": 0.0181,
+      "step": 40930
+    },
+    {
+      "action_loss": 0.0007742529851384461,
+      "epoch": 36.80755395683453,
+      "step": 40930
+    },
+    {
+      "epoch": 36.80755395683453,
+      "step": 40930,
+      "torque_loss": 0.08615850657224655
+    },
+    {
+      "epoch": 36.80755395683453,
+      "force_loss": 0.000837352592498064,
+      "step": 40930
+    },
+    {
+      "epoch": 36.81654676258993,
+      "grad_norm": 0.1636144071817398,
+      "learning_rate": 2.5145722283064698e-05,
+      "loss": 0.0175,
+      "step": 40940
+    },
+    {
+      "action_loss": 0.0014748625690117478,
+      "epoch": 36.81654676258993,
+      "step": 40940
+    },
+    {
+      "epoch": 36.81654676258993,
+      "step": 40940,
+      "torque_loss": 0.12413579225540161
+    },
+    {
+      "epoch": 36.81654676258993,
+      "force_loss": 0.0024982497561722994,
+      "step": 40940
+    },
+    {
+      "epoch": 36.82553956834532,
+      "grad_norm": 0.24281755089759827,
+      "learning_rate": 2.5121814087698602e-05,
+      "loss": 0.0156,
+      "step": 40950
+    },
+    {
+      "action_loss": 0.0018437281250953674,
+      "epoch": 36.82553956834532,
+      "step": 40950
+    },
+    {
+      "epoch": 36.82553956834532,
+      "step": 40950,
+      "torque_loss": 0.13107097148895264
+    },
+    {
+      "epoch": 36.82553956834532,
+      "force_loss": 0.003890153719112277,
+      "step": 40950
+    },
+    {
+      "epoch": 36.83453237410072,
+      "grad_norm": 0.1569495052099228,
+      "learning_rate": 2.509791344966848e-05,
+      "loss": 0.0153,
+      "step": 40960
+    },
+    {
+      "action_loss": 0.00205861940048635,
+      "epoch": 36.83453237410072,
+      "step": 40960
+    },
+    {
+      "epoch": 36.83453237410072,
+      "step": 40960,
+      "torque_loss": 0.14420723915100098
+    },
+    {
+      "epoch": 36.83453237410072,
+      "force_loss": 0.005641359835863113,
+      "step": 40960
+    },
+    {
+      "epoch": 36.843525179856115,
+      "grad_norm": 0.09009571373462677,
+      "learning_rate": 2.5074020376234768e-05,
+      "loss": 0.0167,
+      "step": 40970
+    },
+    {
+      "action_loss": 0.0010952440788969398,
+      "epoch": 36.843525179856115,
+      "step": 40970
+    },
+    {
+      "epoch": 36.843525179856115,
+      "step": 40970,
+      "torque_loss": 0.12368500232696533
+    },
+    {
+      "epoch": 36.843525179856115,
+      "force_loss": 0.0014682303881272674,
+      "step": 40970
+    },
+    {
+      "epoch": 36.85251798561151,
+      "grad_norm": 0.12156790494918823,
+      "learning_rate": 2.5050134874655534e-05,
+      "loss": 0.0169,
+      "step": 40980
+    },
+    {
+      "action_loss": 0.0015372349880635738,
+      "epoch": 36.85251798561151,
+      "step": 40980
+    },
+    {
+      "epoch": 36.85251798561151,
+      "step": 40980,
+      "torque_loss": 0.16442887485027313
+    },
+    {
+      "epoch": 36.85251798561151,
+      "force_loss": 0.005887440871447325,
+      "step": 40980
+    },
+    {
+      "epoch": 36.861510791366904,
+      "grad_norm": 0.2620079517364502,
+      "learning_rate": 2.5026256952186566e-05,
+      "loss": 0.0135,
+      "step": 40990
+    },
+    {
+      "action_loss": 0.0025781861040741205,
+      "epoch": 36.861510791366904,
+      "step": 40990
+    },
+    {
+      "epoch": 36.861510791366904,
+      "step": 40990,
+      "torque_loss": 0.16907234489917755
+    },
+    {
+      "epoch": 36.861510791366904,
+      "force_loss": 0.006407263223081827,
+      "step": 40990
+    },
+    {
+      "epoch": 36.8705035971223,
+      "grad_norm": 0.1035318449139595,
+      "learning_rate": 2.5002386616081335e-05,
+      "loss": 0.0163,
+      "step": 41000
+    },
+    {
+      "action_loss": 0.0011775917373597622,
+      "epoch": 36.8705035971223,
+      "step": 41000
+    },
+    {
+      "epoch": 36.8705035971223,
+      "step": 41000,
+      "torque_loss": 0.10542478412389755
+    },
+    {
+      "epoch": 36.8705035971223,
+      "force_loss": 0.0017100091790780425,
+      "step": 41000
+    },
+    {
+      "epoch": 36.8794964028777,
+      "grad_norm": 0.09911273419857025,
+      "learning_rate": 2.497852387359103e-05,
+      "loss": 0.026,
+      "step": 41010
+    },
+    {
+      "action_loss": 0.0013963145902380347,
+      "epoch": 36.8794964028777,
+      "step": 41010
+    },
+    {
+      "epoch": 36.8794964028777,
+      "step": 41010,
+      "torque_loss": 0.10430901497602463
+    },
+    {
+      "epoch": 36.8794964028777,
+      "force_loss": 0.0026781449560076,
+      "step": 41010
+    },
+    {
+      "epoch": 36.888489208633096,
+      "grad_norm": 0.11811426281929016,
+      "learning_rate": 2.4954668731964496e-05,
+      "loss": 0.0163,
+      "step": 41020
+    },
+    {
+      "action_loss": 0.005028447601944208,
+      "epoch": 36.888489208633096,
+      "step": 41020
+    },
+    {
+      "epoch": 36.888489208633096,
+      "step": 41020,
+      "torque_loss": 0.12395540624856949
+    },
+    {
+      "epoch": 36.888489208633096,
+      "force_loss": 0.005967164412140846,
+      "step": 41020
+    },
+    {
+      "epoch": 36.89748201438849,
+      "grad_norm": 0.07472461462020874,
+      "learning_rate": 2.4930821198448364e-05,
+      "loss": 0.0175,
+      "step": 41030
+    },
+    {
+      "action_loss": 0.009155981242656708,
+      "epoch": 36.89748201438849,
+      "step": 41030
+    },
+    {
+      "epoch": 36.89748201438849,
+      "step": 41030,
+      "torque_loss": 0.16561651229858398
+    },
+    {
+      "epoch": 36.89748201438849,
+      "force_loss": 0.019507234916090965,
+      "step": 41030
+    },
+    {
+      "epoch": 36.906474820143885,
+      "grad_norm": 0.1970684975385666,
+      "learning_rate": 2.4906981280286796e-05,
+      "loss": 0.0172,
+      "step": 41040
+    },
+    {
+      "action_loss": 0.003002721117809415,
+      "epoch": 36.906474820143885,
+      "step": 41040
+    },
+    {
+      "epoch": 36.906474820143885,
+      "step": 41040,
+      "torque_loss": 0.19472616910934448
+    },
+    {
+      "epoch": 36.906474820143885,
+      "force_loss": 0.002908949973061681,
+      "step": 41040
+    },
+    {
+      "epoch": 36.91546762589928,
+      "grad_norm": 0.16730277240276337,
+      "learning_rate": 2.488314898472179e-05,
+      "loss": 0.017,
+      "step": 41050
+    },
+    {
+      "action_loss": 0.0017800036584958434,
+      "epoch": 36.91546762589928,
+      "step": 41050
+    },
+    {
+      "epoch": 36.91546762589928,
+      "step": 41050,
+      "torque_loss": 0.1606191098690033
+    },
+    {
+      "epoch": 36.91546762589928,
+      "force_loss": 0.0013810022501274943,
+      "step": 41050
+    },
+    {
+      "epoch": 36.92446043165468,
+      "grad_norm": 0.2171097993850708,
+      "learning_rate": 2.485932431899295e-05,
+      "loss": 0.0155,
+      "step": 41060
+    },
+    {
+      "action_loss": 0.0010319054126739502,
+      "epoch": 36.92446043165468,
+      "step": 41060
+    },
+    {
+      "epoch": 36.92446043165468,
+      "step": 41060,
+      "torque_loss": 0.14204587042331696
+    },
+    {
+      "epoch": 36.92446043165468,
+      "force_loss": 0.0026959876995533705,
+      "step": 41060
+    },
+    {
+      "epoch": 36.93345323741007,
+      "grad_norm": 0.13802260160446167,
+      "learning_rate": 2.4835507290337584e-05,
+      "loss": 0.0205,
+      "step": 41070
+    },
+    {
+      "action_loss": 0.004782230127602816,
+      "epoch": 36.93345323741007,
+      "step": 41070
+    },
+    {
+      "epoch": 36.93345323741007,
+      "step": 41070,
+      "torque_loss": 0.13187959790229797
+    },
+    {
+      "epoch": 36.93345323741007,
+      "force_loss": 0.002835092833265662,
+      "step": 41070
+    },
+    {
+      "epoch": 36.94244604316547,
+      "grad_norm": 0.24833816289901733,
+      "learning_rate": 2.4811697905990672e-05,
+      "loss": 0.015,
+      "step": 41080
+    },
+    {
+      "action_loss": 0.007749114651232958,
+      "epoch": 36.94244604316547,
+      "step": 41080
+    },
+    {
+      "epoch": 36.94244604316547,
+      "step": 41080,
+      "torque_loss": 0.16751426458358765
+    },
+    {
+      "epoch": 36.94244604316547,
+      "force_loss": 0.011407208628952503,
+      "step": 41080
+    },
+    {
+      "epoch": 36.951438848920866,
+      "grad_norm": 0.055707551538944244,
+      "learning_rate": 2.4787896173184854e-05,
+      "loss": 0.0156,
+      "step": 41090
+    },
+    {
+      "action_loss": 0.004178546369075775,
+      "epoch": 36.951438848920866,
+      "step": 41090
+    },
+    {
+      "epoch": 36.951438848920866,
+      "step": 41090,
+      "torque_loss": 0.11315649747848511
+    },
+    {
+      "epoch": 36.951438848920866,
+      "force_loss": 0.005426496267318726,
+      "step": 41090
+    },
+    {
+      "epoch": 36.960431654676256,
+      "grad_norm": 0.11946913599967957,
+      "learning_rate": 2.4764102099150534e-05,
+      "loss": 0.0199,
+      "step": 41100
+    },
+    {
+      "action_loss": 0.002004031091928482,
+      "epoch": 36.960431654676256,
+      "step": 41100
+    },
+    {
+      "epoch": 36.960431654676256,
+      "step": 41100,
+      "torque_loss": 0.08728880435228348
+    },
+    {
+      "epoch": 36.960431654676256,
+      "force_loss": 0.0021931633818894625,
+      "step": 41100
+    },
+    {
+      "epoch": 36.969424460431654,
+      "grad_norm": 0.11880341172218323,
+      "learning_rate": 2.4740315691115644e-05,
+      "loss": 0.0177,
+      "step": 41110
+    },
+    {
+      "action_loss": 0.01130113285034895,
+      "epoch": 36.969424460431654,
+      "step": 41110
+    },
+    {
+      "epoch": 36.969424460431654,
+      "step": 41110,
+      "torque_loss": 0.12452418357133865
+    },
+    {
+      "epoch": 36.969424460431654,
+      "force_loss": 0.021689504384994507,
+      "step": 41110
+    },
+    {
+      "epoch": 36.97841726618705,
+      "grad_norm": 0.12031841278076172,
+      "learning_rate": 2.4716536956305918e-05,
+      "loss": 0.0196,
+      "step": 41120
+    },
+    {
+      "action_loss": 0.010829283855855465,
+      "epoch": 36.97841726618705,
+      "step": 41120
+    },
+    {
+      "epoch": 36.97841726618705,
+      "step": 41120,
+      "torque_loss": 0.1601242870092392
+    },
+    {
+      "epoch": 36.97841726618705,
+      "force_loss": 0.005428771022707224,
+      "step": 41120
+    },
+    {
+      "epoch": 36.98741007194245,
+      "grad_norm": 0.1415182650089264,
+      "learning_rate": 2.4692765901944697e-05,
+      "loss": 0.0162,
+      "step": 41130
+    },
+    {
+      "action_loss": 0.001498539000749588,
+      "epoch": 36.98741007194245,
+      "step": 41130
+    },
+    {
+      "epoch": 36.98741007194245,
+      "step": 41130,
+      "torque_loss": 0.11062075942754745
+    },
+    {
+      "epoch": 36.98741007194245,
+      "force_loss": 0.0020529981702566147,
+      "step": 41130
+    },
+    {
+      "epoch": 36.99640287769784,
+      "grad_norm": 0.19059258699417114,
+      "learning_rate": 2.4669002535253e-05,
+      "loss": 0.0207,
+      "step": 41140
+    },
+    {
+      "action_loss": 0.0025275133084505796,
+      "epoch": 36.99640287769784,
+      "step": 41140
+    },
+    {
+      "epoch": 36.99640287769784,
+      "step": 41140,
+      "torque_loss": 0.167414590716362
+    },
+    {
+      "epoch": 36.99640287769784,
+      "force_loss": 0.0019015618599951267,
+      "step": 41140
+    },
+    {
+      "epoch": 37.00539568345324,
+      "grad_norm": 0.15231658518314362,
+      "learning_rate": 2.46452468634495e-05,
+      "loss": 0.017,
+      "step": 41150
+    },
+    {
+      "action_loss": 0.0024368467275053263,
+      "epoch": 37.00539568345324,
+      "step": 41150
+    },
+    {
+      "epoch": 37.00539568345324,
+      "step": 41150,
+      "torque_loss": 0.17986814677715302
+    },
+    {
+      "epoch": 37.00539568345324,
+      "force_loss": 0.004945660009980202,
+      "step": 41150
+    },
+    {
+      "epoch": 37.014388489208635,
+      "grad_norm": 0.07546962797641754,
+      "learning_rate": 2.462149889375055e-05,
+      "loss": 0.0193,
+      "step": 41160
+    },
+    {
+      "action_loss": 0.002899347571656108,
+      "epoch": 37.014388489208635,
+      "step": 41160
+    },
+    {
+      "epoch": 37.014388489208635,
+      "step": 41160,
+      "torque_loss": 0.14161519706249237
+    },
+    {
+      "epoch": 37.014388489208635,
+      "force_loss": 0.005148437339812517,
+      "step": 41160
+    },
+    {
+      "epoch": 37.023381294964025,
+      "grad_norm": 0.1663338989019394,
+      "learning_rate": 2.459775863337014e-05,
+      "loss": 0.0167,
+      "step": 41170
+    },
+    {
+      "action_loss": 0.0018307944992557168,
+      "epoch": 37.023381294964025,
+      "step": 41170
+    },
+    {
+      "epoch": 37.023381294964025,
+      "step": 41170,
+      "torque_loss": 0.12471288442611694
+    },
+    {
+      "epoch": 37.023381294964025,
+      "force_loss": 0.00840574037283659,
+      "step": 41170
+    },
+    {
+      "epoch": 37.03237410071942,
+      "grad_norm": 0.18716450035572052,
+      "learning_rate": 2.4574026089519985e-05,
+      "loss": 0.0157,
+      "step": 41180
+    },
+    {
+      "action_loss": 0.0017526658484712243,
+      "epoch": 37.03237410071942,
+      "step": 41180
+    },
+    {
+      "epoch": 37.03237410071942,
+      "step": 41180,
+      "torque_loss": 0.10260826349258423
+    },
+    {
+      "epoch": 37.03237410071942,
+      "force_loss": 0.0017096939263865352,
+      "step": 41180
+    },
+    {
+      "epoch": 37.04136690647482,
+      "grad_norm": 0.24679633975028992,
+      "learning_rate": 2.4550301269409333e-05,
+      "loss": 0.0174,
+      "step": 41190
+    },
+    {
+      "action_loss": 0.007488483563065529,
+      "epoch": 37.04136690647482,
+      "step": 41190
+    },
+    {
+      "epoch": 37.04136690647482,
+      "step": 41190,
+      "torque_loss": 0.10798462480306625
+    },
+    {
+      "epoch": 37.04136690647482,
+      "force_loss": 0.004979473073035479,
+      "step": 41190
+    },
+    {
+      "epoch": 37.05035971223022,
+      "grad_norm": 0.17777372896671295,
+      "learning_rate": 2.4526584180245216e-05,
+      "loss": 0.0181,
+      "step": 41200
+    },
+    {
+      "action_loss": 0.0021048991475254297,
+      "epoch": 37.05035971223022,
+      "step": 41200
+    },
+    {
+      "epoch": 37.05035971223022,
+      "step": 41200,
+      "torque_loss": 0.12938641011714935
+    },
+    {
+      "epoch": 37.05035971223022,
+      "force_loss": 0.001535317045636475,
+      "step": 41200
+    },
+    {
+      "epoch": 37.05935251798561,
+      "grad_norm": 0.17514784634113312,
+      "learning_rate": 2.4502874829232236e-05,
+      "loss": 0.0197,
+      "step": 41210
+    },
+    {
+      "action_loss": 0.001937543973326683,
+      "epoch": 37.05935251798561,
+      "step": 41210
+    },
+    {
+      "epoch": 37.05935251798561,
+      "step": 41210,
+      "torque_loss": 0.17843842506408691
+    },
+    {
+      "epoch": 37.05935251798561,
+      "force_loss": 0.003016873961314559,
+      "step": 41210
+    },
+    {
+      "epoch": 37.068345323741006,
+      "grad_norm": 0.3838598430156708,
+      "learning_rate": 2.447917322357267e-05,
+      "loss": 0.0163,
+      "step": 41220
+    },
+    {
+      "action_loss": 0.0018005785532295704,
+      "epoch": 37.068345323741006,
+      "step": 41220
+    },
+    {
+      "epoch": 37.068345323741006,
+      "step": 41220,
+      "torque_loss": 0.1490955501794815
+    },
+    {
+      "epoch": 37.068345323741006,
+      "force_loss": 0.004971009213477373,
+      "step": 41220
+    },
+    {
+      "epoch": 37.077338129496404,
+      "grad_norm": 0.14843182265758514,
+      "learning_rate": 2.4455479370466443e-05,
+      "loss": 0.017,
+      "step": 41230
+    },
+    {
+      "action_loss": 0.0018570529064163566,
+      "epoch": 37.077338129496404,
+      "step": 41230
+    },
+    {
+      "epoch": 37.077338129496404,
+      "step": 41230,
+      "torque_loss": 0.12969480454921722
+    },
+    {
+      "epoch": 37.077338129496404,
+      "force_loss": 0.003479142440482974,
+      "step": 41230
+    },
+    {
+      "epoch": 37.0863309352518,
+      "grad_norm": 0.3520604968070984,
+      "learning_rate": 2.4431793277111097e-05,
+      "loss": 0.0154,
+      "step": 41240
+    },
+    {
+      "action_loss": 0.0029263596516102552,
+      "epoch": 37.0863309352518,
+      "step": 41240
+    },
+    {
+      "epoch": 37.0863309352518,
+      "step": 41240,
+      "torque_loss": 0.1147025004029274
+    },
+    {
+      "epoch": 37.0863309352518,
+      "force_loss": 0.003374862251803279,
+      "step": 41240
+    },
+    {
+      "epoch": 37.09532374100719,
+      "grad_norm": 0.16331171989440918,
+      "learning_rate": 2.4408114950701905e-05,
+      "loss": 0.0145,
+      "step": 41250
+    },
+    {
+      "action_loss": 0.001004036865197122,
+      "epoch": 37.09532374100719,
+      "step": 41250
+    },
+    {
+      "epoch": 37.09532374100719,
+      "step": 41250,
+      "torque_loss": 0.10240895301103592
+    },
+    {
+      "epoch": 37.09532374100719,
+      "force_loss": 0.002176305279135704,
+      "step": 41250
+    },
+    {
+      "epoch": 37.10431654676259,
+      "grad_norm": 0.1545088291168213,
+      "learning_rate": 2.4384444398431634e-05,
+      "loss": 0.0147,
+      "step": 41260
+    },
+    {
+      "action_loss": 0.0008223604527302086,
+      "epoch": 37.10431654676259,
+      "step": 41260
+    },
+    {
+      "epoch": 37.10431654676259,
+      "step": 41260,
+      "torque_loss": 0.14022932946681976
+    },
+    {
+      "epoch": 37.10431654676259,
+      "force_loss": 0.0010246924357488751,
+      "step": 41260
+    },
+    {
+      "epoch": 37.11330935251799,
+      "grad_norm": 0.19908730685710907,
+      "learning_rate": 2.4360781627490837e-05,
+      "loss": 0.016,
+      "step": 41270
+    },
+    {
+      "action_loss": 0.001409512129612267,
+      "epoch": 37.11330935251799,
+      "step": 41270
+    },
+    {
+      "epoch": 37.11330935251799,
+      "step": 41270,
+      "torque_loss": 0.10714276880025864
+    },
+    {
+      "epoch": 37.11330935251799,
+      "force_loss": 0.0017570602940395474,
+      "step": 41270
+    },
+    {
+      "epoch": 37.12230215827338,
+      "grad_norm": 0.13742773234844208,
+      "learning_rate": 2.433712664506762e-05,
+      "loss": 0.015,
+      "step": 41280
+    },
+    {
+      "action_loss": 0.0016693522920832038,
+      "epoch": 37.12230215827338,
+      "step": 41280
+    },
+    {
+      "epoch": 37.12230215827338,
+      "step": 41280,
+      "torque_loss": 0.10153579711914062
+    },
+    {
+      "epoch": 37.12230215827338,
+      "force_loss": 0.005407848861068487,
+      "step": 41280
+    },
+    {
+      "epoch": 37.131294964028775,
+      "grad_norm": 0.341092973947525,
+      "learning_rate": 2.431347945834774e-05,
+      "loss": 0.0151,
+      "step": 41290
+    },
+    {
+      "action_loss": 0.0009617149480618536,
+      "epoch": 37.131294964028775,
+      "step": 41290
+    },
+    {
+      "epoch": 37.131294964028775,
+      "step": 41290,
+      "torque_loss": 0.15237867832183838
+    },
+    {
+      "epoch": 37.131294964028775,
+      "force_loss": 0.0007011820562183857,
+      "step": 41290
+    },
+    {
+      "epoch": 37.14028776978417,
+      "grad_norm": 0.10262230038642883,
+      "learning_rate": 2.428984007451458e-05,
+      "loss": 0.018,
+      "step": 41300
+    },
+    {
+      "action_loss": 0.0047562080435454845,
+      "epoch": 37.14028776978417,
+      "step": 41300
+    },
+    {
+      "epoch": 37.14028776978417,
+      "step": 41300,
+      "torque_loss": 0.12761260569095612
+    },
+    {
+      "epoch": 37.14028776978417,
+      "force_loss": 0.006924936082214117,
+      "step": 41300
+    },
+    {
+      "epoch": 37.14928057553957,
+      "grad_norm": 0.2014538049697876,
+      "learning_rate": 2.426620850074917e-05,
+      "loss": 0.0183,
+      "step": 41310
+    },
+    {
+      "action_loss": 0.002989858388900757,
+      "epoch": 37.14928057553957,
+      "step": 41310
+    },
+    {
+      "epoch": 37.14928057553957,
+      "step": 41310,
+      "torque_loss": 0.10058669000864029
+    },
+    {
+      "epoch": 37.14928057553957,
+      "force_loss": 0.0022126834373921156,
+      "step": 41310
+    },
+    {
+      "epoch": 37.15827338129496,
+      "grad_norm": 0.23876667022705078,
+      "learning_rate": 2.424258474423014e-05,
+      "loss": 0.0207,
+      "step": 41320
+    },
+    {
+      "action_loss": 0.0018437696853652596,
+      "epoch": 37.15827338129496,
+      "step": 41320
+    },
+    {
+      "epoch": 37.15827338129496,
+      "step": 41320,
+      "torque_loss": 0.10081741958856583
+    },
+    {
+      "epoch": 37.15827338129496,
+      "force_loss": 0.0060493722558021545,
+      "step": 41320
+    },
+    {
+      "epoch": 37.16726618705036,
+      "grad_norm": 0.20577901601791382,
+      "learning_rate": 2.421896881213382e-05,
+      "loss": 0.0167,
+      "step": 41330
+    },
+    {
+      "action_loss": 0.004117974545806646,
+      "epoch": 37.16726618705036,
+      "step": 41330
+    },
+    {
+      "epoch": 37.16726618705036,
+      "step": 41330,
+      "torque_loss": 0.14380398392677307
+    },
+    {
+      "epoch": 37.16726618705036,
+      "force_loss": 0.003966405522078276,
+      "step": 41330
+    },
+    {
+      "epoch": 37.17625899280576,
+      "grad_norm": 0.23002475500106812,
+      "learning_rate": 2.419536071163402e-05,
+      "loss": 0.0163,
+      "step": 41340
+    },
+    {
+      "action_loss": 0.0014744462678208947,
+      "epoch": 37.17625899280576,
+      "step": 41340
+    },
+    {
+      "epoch": 37.17625899280576,
+      "step": 41340,
+      "torque_loss": 0.09775781631469727
+    },
+    {
+      "epoch": 37.17625899280576,
+      "force_loss": 0.0056115384213626385,
+      "step": 41340
+    },
+    {
+      "epoch": 37.185251798561154,
+      "grad_norm": 0.2878395915031433,
+      "learning_rate": 2.417176044990233e-05,
+      "loss": 0.0152,
+      "step": 41350
+    },
+    {
+      "action_loss": 0.0076812454499304295,
+      "epoch": 37.185251798561154,
+      "step": 41350
+    },
+    {
+      "epoch": 37.185251798561154,
+      "step": 41350,
+      "torque_loss": 0.09224791079759598
+    },
+    {
+      "epoch": 37.185251798561154,
+      "force_loss": 0.017508333548903465,
+      "step": 41350
+    },
+    {
+      "epoch": 37.194244604316545,
+      "grad_norm": 0.2531919777393341,
+      "learning_rate": 2.4148168034107855e-05,
+      "loss": 0.0175,
+      "step": 41360
+    },
+    {
+      "action_loss": 0.0008937583188526332,
+      "epoch": 37.194244604316545,
+      "step": 41360
+    },
+    {
+      "epoch": 37.194244604316545,
+      "step": 41360,
+      "torque_loss": 0.15019066631793976
+    },
+    {
+      "epoch": 37.194244604316545,
+      "force_loss": 0.0014468939043581486,
+      "step": 41360
+    },
+    {
+      "epoch": 37.20323741007194,
+      "grad_norm": 0.19757236540317535,
+      "learning_rate": 2.4124583471417355e-05,
+      "loss": 0.0157,
+      "step": 41370
+    },
+    {
+      "action_loss": 0.0034667998552322388,
+      "epoch": 37.20323741007194,
+      "step": 41370
+    },
+    {
+      "epoch": 37.20323741007194,
+      "step": 41370,
+      "torque_loss": 0.1455114185810089
+    },
+    {
+      "epoch": 37.20323741007194,
+      "force_loss": 0.002801511436700821,
+      "step": 41370
+    },
+    {
+      "epoch": 37.21223021582734,
+      "grad_norm": 0.09754537791013718,
+      "learning_rate": 2.41010067689952e-05,
+      "loss": 0.0174,
+      "step": 41380
+    },
+    {
+      "action_loss": 0.006356456782668829,
+      "epoch": 37.21223021582734,
+      "step": 41380
+    },
+    {
+      "epoch": 37.21223021582734,
+      "step": 41380,
+      "torque_loss": 0.1255526840686798
+    },
+    {
+      "epoch": 37.21223021582734,
+      "force_loss": 0.0030926025938242674,
+      "step": 41380
+    },
+    {
+      "epoch": 37.22122302158273,
+      "grad_norm": 0.2362031191587448,
+      "learning_rate": 2.4077437934003338e-05,
+      "loss": 0.0168,
+      "step": 41390
+    },
+    {
+      "action_loss": 0.0017143426230177283,
+      "epoch": 37.22122302158273,
+      "step": 41390
+    },
+    {
+      "epoch": 37.22122302158273,
+      "step": 41390,
+      "torque_loss": 0.13515441119670868
+    },
+    {
+      "epoch": 37.22122302158273,
+      "force_loss": 0.0025035792496055365,
+      "step": 41390
+    },
+    {
+      "epoch": 37.23021582733813,
+      "grad_norm": 0.15564242005348206,
+      "learning_rate": 2.405387697360143e-05,
+      "loss": 0.0179,
+      "step": 41400
+    },
+    {
+      "action_loss": 0.002296850783750415,
+      "epoch": 37.23021582733813,
+      "step": 41400
+    },
+    {
+      "epoch": 37.23021582733813,
+      "step": 41400,
+      "torque_loss": 0.14443393051624298
+    },
+    {
+      "epoch": 37.23021582733813,
+      "force_loss": 0.005019243340939283,
+      "step": 41400
+    },
+    {
+      "epoch": 37.239208633093526,
+      "grad_norm": 0.17784880101680756,
+      "learning_rate": 2.4030323894946595e-05,
+      "loss": 0.015,
+      "step": 41410
+    },
+    {
+      "action_loss": 0.003928730264306068,
+      "epoch": 37.239208633093526,
+      "step": 41410
+    },
+    {
+      "epoch": 37.239208633093526,
+      "step": 41410,
+      "torque_loss": 0.13540036976337433
+    },
+    {
+      "epoch": 37.239208633093526,
+      "force_loss": 0.003075113520026207,
+      "step": 41410
+    },
+    {
+      "epoch": 37.24820143884892,
+      "grad_norm": 0.2360670268535614,
+      "learning_rate": 2.40067787051937e-05,
+      "loss": 0.0174,
+      "step": 41420
+    },
+    {
+      "action_loss": 0.0012221961515024304,
+      "epoch": 37.24820143884892,
+      "step": 41420
+    },
+    {
+      "epoch": 37.24820143884892,
+      "step": 41420,
+      "torque_loss": 0.13722175359725952
+    },
+    {
+      "epoch": 37.24820143884892,
+      "force_loss": 0.004681184887886047,
+      "step": 41420
+    },
+    {
+      "epoch": 37.257194244604314,
+      "grad_norm": 0.1011488288640976,
+      "learning_rate": 2.3983241411495087e-05,
+      "loss": 0.0168,
+      "step": 41430
+    },
+    {
+      "action_loss": 0.0016275119269266725,
+      "epoch": 37.257194244604314,
+      "step": 41430
+    },
+    {
+      "epoch": 37.257194244604314,
+      "step": 41430,
+      "torque_loss": 0.1198020949959755
+    },
+    {
+      "epoch": 37.257194244604314,
+      "force_loss": 0.0024211483541876078,
+      "step": 41430
+    },
+    {
+      "epoch": 37.26618705035971,
+      "grad_norm": 0.2071520984172821,
+      "learning_rate": 2.3959712021000823e-05,
+      "loss": 0.0169,
+      "step": 41440
+    },
+    {
+      "action_loss": 0.0016327359480783343,
+      "epoch": 37.26618705035971,
+      "step": 41440
+    },
+    {
+      "epoch": 37.26618705035971,
+      "step": 41440,
+      "torque_loss": 0.07270107418298721
+    },
+    {
+      "epoch": 37.26618705035971,
+      "force_loss": 0.004087215755134821,
+      "step": 41440
+    },
+    {
+      "epoch": 37.27517985611511,
+      "grad_norm": 0.4247923195362091,
+      "learning_rate": 2.3936190540858495e-05,
+      "loss": 0.016,
+      "step": 41450
+    },
+    {
+      "action_loss": 0.008242207579314709,
+      "epoch": 37.27517985611511,
+      "step": 41450
+    },
+    {
+      "epoch": 37.27517985611511,
+      "step": 41450,
+      "torque_loss": 0.1607363075017929
+    },
+    {
+      "epoch": 37.27517985611511,
+      "force_loss": 0.005622658878564835,
+      "step": 41450
+    },
+    {
+      "epoch": 37.28417266187051,
+      "grad_norm": 0.14170235395431519,
+      "learning_rate": 2.39126769782133e-05,
+      "loss": 0.018,
+      "step": 41460
+    },
+    {
+      "action_loss": 0.0017281798645853996,
+      "epoch": 37.28417266187051,
+      "step": 41460
+    },
+    {
+      "epoch": 37.28417266187051,
+      "step": 41460,
+      "torque_loss": 0.09212587028741837
+    },
+    {
+      "epoch": 37.28417266187051,
+      "force_loss": 0.004253492224961519,
+      "step": 41460
+    },
+    {
+      "epoch": 37.2931654676259,
+      "grad_norm": 0.06294012814760208,
+      "learning_rate": 2.388917134020805e-05,
+      "loss": 0.0155,
+      "step": 41470
+    },
+    {
+      "action_loss": 0.005167144350707531,
+      "epoch": 37.2931654676259,
+      "step": 41470
+    },
+    {
+      "epoch": 37.2931654676259,
+      "step": 41470,
+      "torque_loss": 0.12978367507457733
+    },
+    {
+      "epoch": 37.2931654676259,
+      "force_loss": 0.006586526986211538,
+      "step": 41470
+    },
+    {
+      "epoch": 37.302158273381295,
+      "grad_norm": 0.10713452845811844,
+      "learning_rate": 2.3865673633983128e-05,
+      "loss": 0.0171,
+      "step": 41480
+    },
+    {
+      "action_loss": 0.001562502351589501,
+      "epoch": 37.302158273381295,
+      "step": 41480
+    },
+    {
+      "epoch": 37.302158273381295,
+      "step": 41480,
+      "torque_loss": 0.10428479313850403
+    },
+    {
+      "epoch": 37.302158273381295,
+      "force_loss": 0.0011037611402571201,
+      "step": 41480
+    },
+    {
+      "epoch": 37.31115107913669,
+      "grad_norm": 0.16244597733020782,
+      "learning_rate": 2.3842183866676492e-05,
+      "loss": 0.0158,
+      "step": 41490
+    },
+    {
+      "action_loss": 0.00414640735834837,
+      "epoch": 37.31115107913669,
+      "step": 41490
+    },
+    {
+      "epoch": 37.31115107913669,
+      "step": 41490,
+      "torque_loss": 0.09372931718826294
+    },
+    {
+      "epoch": 37.31115107913669,
+      "force_loss": 0.013103918172419071,
+      "step": 41490
+    },
+    {
+      "epoch": 37.32014388489208,
+      "grad_norm": 0.10922183841466904,
+      "learning_rate": 2.381870204542377e-05,
+      "loss": 0.0155,
+      "step": 41500
+    },
+    {
+      "action_loss": 0.0007616126094944775,
+      "epoch": 37.32014388489208,
+      "step": 41500
+    },
+    {
+      "epoch": 37.32014388489208,
+      "step": 41500,
+      "torque_loss": 0.12131894379854202
+    },
+    {
+      "epoch": 37.32014388489208,
+      "force_loss": 0.0007598371594212949,
+      "step": 41500
+    },
+    {
+      "epoch": 37.32913669064748,
+      "grad_norm": 0.12311369180679321,
+      "learning_rate": 2.379522817735808e-05,
+      "loss": 0.0143,
+      "step": 41510
+    },
+    {
+      "action_loss": 0.0022795626427978277,
+      "epoch": 37.32913669064748,
+      "step": 41510
+    },
+    {
+      "epoch": 37.32913669064748,
+      "step": 41510,
+      "torque_loss": 0.16554562747478485
+    },
+    {
+      "epoch": 37.32913669064748,
+      "force_loss": 0.008074014447629452,
+      "step": 41510
+    },
+    {
+      "epoch": 37.33812949640288,
+      "grad_norm": 0.1777009814977646,
+      "learning_rate": 2.377176226961018e-05,
+      "loss": 0.0161,
+      "step": 41520
+    },
+    {
+      "action_loss": 0.003735143691301346,
+      "epoch": 37.33812949640288,
+      "step": 41520
+    },
+    {
+      "epoch": 37.33812949640288,
+      "step": 41520,
+      "torque_loss": 0.10782458633184433
+    },
+    {
+      "epoch": 37.33812949640288,
+      "force_loss": 0.01307226438075304,
+      "step": 41520
+    },
+    {
+      "epoch": 37.347122302158276,
+      "grad_norm": 0.1115955039858818,
+      "learning_rate": 2.3748304329308384e-05,
+      "loss": 0.0159,
+      "step": 41530
+    },
+    {
+      "action_loss": 0.0013150578597560525,
+      "epoch": 37.347122302158276,
+      "step": 41530
+    },
+    {
+      "epoch": 37.347122302158276,
+      "step": 41530,
+      "torque_loss": 0.10661900043487549
+    },
+    {
+      "epoch": 37.347122302158276,
+      "force_loss": 0.0032155411317944527,
+      "step": 41530
+    },
+    {
+      "epoch": 37.356115107913666,
+      "grad_norm": 0.46261322498321533,
+      "learning_rate": 2.372485436357858e-05,
+      "loss": 0.0168,
+      "step": 41540
+    },
+    {
+      "action_loss": 0.0012545754434540868,
+      "epoch": 37.356115107913666,
+      "step": 41540
+    },
+    {
+      "epoch": 37.356115107913666,
+      "step": 41540,
+      "torque_loss": 0.11633291095495224
+    },
+    {
+      "epoch": 37.356115107913666,
+      "force_loss": 0.0012813982320949435,
+      "step": 41540
+    },
+    {
+      "epoch": 37.365107913669064,
+      "grad_norm": 0.36423543095588684,
+      "learning_rate": 2.3701412379544296e-05,
+      "loss": 0.0187,
+      "step": 41550
+    },
+    {
+      "action_loss": 0.005754852667450905,
+      "epoch": 37.365107913669064,
+      "step": 41550
+    },
+    {
+      "epoch": 37.365107913669064,
+      "step": 41550,
+      "torque_loss": 0.14603643119335175
+    },
+    {
+      "epoch": 37.365107913669064,
+      "force_loss": 0.007697058375924826,
+      "step": 41550
+    },
+    {
+      "epoch": 37.37410071942446,
+      "grad_norm": 0.09964298456907272,
+      "learning_rate": 2.367797838432653e-05,
+      "loss": 0.0182,
+      "step": 41560
+    },
+    {
+      "action_loss": 0.01978965662419796,
+      "epoch": 37.37410071942446,
+      "step": 41560
+    },
+    {
+      "epoch": 37.37410071942446,
+      "step": 41560,
+      "torque_loss": 0.1627604216337204
+    },
+    {
+      "epoch": 37.37410071942446,
+      "force_loss": 0.019091719761490822,
+      "step": 41560
+    },
+    {
+      "epoch": 37.38309352517986,
+      "grad_norm": 0.1869172304868698,
+      "learning_rate": 2.3654552385043967e-05,
+      "loss": 0.0194,
+      "step": 41570
+    },
+    {
+      "action_loss": 0.0024029752239584923,
+      "epoch": 37.38309352517986,
+      "step": 41570
+    },
+    {
+      "epoch": 37.38309352517986,
+      "step": 41570,
+      "torque_loss": 0.15013819932937622
+    },
+    {
+      "epoch": 37.38309352517986,
+      "force_loss": 0.003192966105416417,
+      "step": 41570
+    },
+    {
+      "epoch": 37.39208633093525,
+      "grad_norm": 0.23554450273513794,
+      "learning_rate": 2.3631134388812742e-05,
+      "loss": 0.0194,
+      "step": 41580
+    },
+    {
+      "action_loss": 0.010989122092723846,
+      "epoch": 37.39208633093525,
+      "step": 41580
+    },
+    {
+      "epoch": 37.39208633093525,
+      "step": 41580,
+      "torque_loss": 0.13955427706241608
+    },
+    {
+      "epoch": 37.39208633093525,
+      "force_loss": 0.024112893268465996,
+      "step": 41580
+    },
+    {
+      "epoch": 37.40107913669065,
+      "grad_norm": 0.41500040888786316,
+      "learning_rate": 2.3607724402746684e-05,
+      "loss": 0.0187,
+      "step": 41590
+    },
+    {
+      "action_loss": 0.0011135967215523124,
+      "epoch": 37.40107913669065,
+      "step": 41590
+    },
+    {
+      "epoch": 37.40107913669065,
+      "step": 41590,
+      "torque_loss": 0.1296481043100357
+    },
+    {
+      "epoch": 37.40107913669065,
+      "force_loss": 0.000798785884398967,
+      "step": 41590
+    },
+    {
+      "epoch": 37.410071942446045,
+      "grad_norm": 0.08141325414180756,
+      "learning_rate": 2.35843224339571e-05,
+      "loss": 0.0148,
+      "step": 41600
+    },
+    {
+      "action_loss": 0.0013470901176333427,
+      "epoch": 37.410071942446045,
+      "step": 41600
+    },
+    {
+      "epoch": 37.410071942446045,
+      "step": 41600,
+      "torque_loss": 0.09661503881216049
+    },
+    {
+      "epoch": 37.410071942446045,
+      "force_loss": 0.0021243628580123186,
+      "step": 41600
+    },
+    {
+      "epoch": 37.419064748201436,
+      "grad_norm": 0.12963275611400604,
+      "learning_rate": 2.3560928489552897e-05,
+      "loss": 0.0133,
+      "step": 41610
+    },
+    {
+      "action_loss": 0.004545900505036116,
+      "epoch": 37.419064748201436,
+      "step": 41610
+    },
+    {
+      "epoch": 37.419064748201436,
+      "step": 41610,
+      "torque_loss": 0.1700814813375473
+    },
+    {
+      "epoch": 37.419064748201436,
+      "force_loss": 0.0026790162082761526,
+      "step": 41610
+    },
+    {
+      "epoch": 37.42805755395683,
+      "grad_norm": 0.1503637582063675,
+      "learning_rate": 2.353754257664053e-05,
+      "loss": 0.0183,
+      "step": 41620
+    },
+    {
+      "action_loss": 0.007140960078686476,
+      "epoch": 37.42805755395683,
+      "step": 41620
+    },
+    {
+      "epoch": 37.42805755395683,
+      "step": 41620,
+      "torque_loss": 0.13710756599903107
+    },
+    {
+      "epoch": 37.42805755395683,
+      "force_loss": 0.0037944645155221224,
+      "step": 41620
+    },
+    {
+      "epoch": 37.43705035971223,
+      "grad_norm": 0.08360598981380463,
+      "learning_rate": 2.3514164702324037e-05,
+      "loss": 0.0166,
+      "step": 41630
+    },
+    {
+      "action_loss": 0.004188325721770525,
+      "epoch": 37.43705035971223,
+      "step": 41630
+    },
+    {
+      "epoch": 37.43705035971223,
+      "step": 41630,
+      "torque_loss": 0.15347853302955627
+    },
+    {
+      "epoch": 37.43705035971223,
+      "force_loss": 0.009718392044305801,
+      "step": 41630
+    },
+    {
+      "epoch": 37.44604316546763,
+      "grad_norm": 0.14037662744522095,
+      "learning_rate": 2.3490794873704963e-05,
+      "loss": 0.0161,
+      "step": 41640
+    },
+    {
+      "action_loss": 0.0008292978745885193,
+      "epoch": 37.44604316546763,
+      "step": 41640
+    },
+    {
+      "epoch": 37.44604316546763,
+      "step": 41640,
+      "torque_loss": 0.1254401057958603
+    },
+    {
+      "epoch": 37.44604316546763,
+      "force_loss": 0.0013599743833765388,
+      "step": 41640
+    },
+    {
+      "epoch": 37.45503597122302,
+      "grad_norm": 0.25255194306373596,
+      "learning_rate": 2.3467433097882496e-05,
+      "loss": 0.0151,
+      "step": 41650
+    },
+    {
+      "action_loss": 0.0012644171947613358,
+      "epoch": 37.45503597122302,
+      "step": 41650
+    },
+    {
+      "epoch": 37.45503597122302,
+      "step": 41650,
+      "torque_loss": 0.11599936336278915
+    },
+    {
+      "epoch": 37.45503597122302,
+      "force_loss": 0.0030962014570832253,
+      "step": 41650
+    },
+    {
+      "epoch": 37.46402877697842,
+      "grad_norm": 0.1710033416748047,
+      "learning_rate": 2.34440793819533e-05,
+      "loss": 0.0165,
+      "step": 41660
+    },
+    {
+      "action_loss": 0.0017680843593552709,
+      "epoch": 37.46402877697842,
+      "step": 41660
+    },
+    {
+      "epoch": 37.46402877697842,
+      "step": 41660,
+      "torque_loss": 0.10299190133810043
+    },
+    {
+      "epoch": 37.46402877697842,
+      "force_loss": 0.001538500771857798,
+      "step": 41660
+    },
+    {
+      "epoch": 37.473021582733814,
+      "grad_norm": 0.15772037208080292,
+      "learning_rate": 2.3420733733011617e-05,
+      "loss": 0.0157,
+      "step": 41670
+    },
+    {
+      "action_loss": 0.005712547805160284,
+      "epoch": 37.473021582733814,
+      "step": 41670
+    },
+    {
+      "epoch": 37.473021582733814,
+      "step": 41670,
+      "torque_loss": 0.10994597524404526
+    },
+    {
+      "epoch": 37.473021582733814,
+      "force_loss": 0.007984291762113571,
+      "step": 41670
+    },
+    {
+      "epoch": 37.48201438848921,
+      "grad_norm": 0.29070183634757996,
+      "learning_rate": 2.3397396158149243e-05,
+      "loss": 0.0184,
+      "step": 41680
+    },
+    {
+      "action_loss": 0.005757478531450033,
+      "epoch": 37.48201438848921,
+      "step": 41680
+    },
+    {
+      "epoch": 37.48201438848921,
+      "step": 41680,
+      "torque_loss": 0.10696472972631454
+    },
+    {
+      "epoch": 37.48201438848921,
+      "force_loss": 0.009240470826625824,
+      "step": 41680
+    },
+    {
+      "epoch": 37.4910071942446,
+      "grad_norm": 0.13802073895931244,
+      "learning_rate": 2.3374066664455498e-05,
+      "loss": 0.0172,
+      "step": 41690
+    },
+    {
+      "action_loss": 0.0016419509192928672,
+      "epoch": 37.4910071942446,
+      "step": 41690
+    },
+    {
+      "epoch": 37.4910071942446,
+      "step": 41690,
+      "torque_loss": 0.11248960345983505
+    },
+    {
+      "epoch": 37.4910071942446,
+      "force_loss": 0.002011026255786419,
+      "step": 41690
+    },
+    {
+      "epoch": 37.5,
+      "grad_norm": 0.1483779400587082,
+      "learning_rate": 2.3350745259017315e-05,
+      "loss": 0.0146,
+      "step": 41700
+    },
+    {
+      "action_loss": 0.00665512727573514,
+      "epoch": 37.5,
+      "step": 41700
+    },
+    {
+      "epoch": 37.5,
+      "step": 41700,
+      "torque_loss": 0.11320441961288452
+    },
+    {
+      "epoch": 37.5,
+      "force_loss": 0.005524365231394768,
+      "step": 41700
+    },
+    {
+      "epoch": 37.5089928057554,
+      "grad_norm": 0.5056332349777222,
+      "learning_rate": 2.332743194891906e-05,
+      "loss": 0.0166,
+      "step": 41710
+    },
+    {
+      "action_loss": 0.0007786910864524543,
+      "epoch": 37.5089928057554,
+      "step": 41710
+    },
+    {
+      "epoch": 37.5089928057554,
+      "step": 41710,
+      "torque_loss": 0.09663679450750351
+    },
+    {
+      "epoch": 37.5089928057554,
+      "force_loss": 0.0010359091684222221,
+      "step": 41710
+    },
+    {
+      "epoch": 37.51798561151079,
+      "grad_norm": 0.09813045710325241,
+      "learning_rate": 2.330412674124276e-05,
+      "loss": 0.0138,
+      "step": 41720
+    },
+    {
+      "action_loss": 0.0011931395856663585,
+      "epoch": 37.51798561151079,
+      "step": 41720
+    },
+    {
+      "epoch": 37.51798561151079,
+      "step": 41720,
+      "torque_loss": 0.11519008129835129
+    },
+    {
+      "epoch": 37.51798561151079,
+      "force_loss": 0.004237674176692963,
+      "step": 41720
+    },
+    {
+      "epoch": 37.526978417266186,
+      "grad_norm": 0.19659537076950073,
+      "learning_rate": 2.328082964306786e-05,
+      "loss": 0.0148,
+      "step": 41730
+    },
+    {
+      "action_loss": 0.0006921563763171434,
+      "epoch": 37.526978417266186,
+      "step": 41730
+    },
+    {
+      "epoch": 37.526978417266186,
+      "step": 41730,
+      "torque_loss": 0.09513994306325912
+    },
+    {
+      "epoch": 37.526978417266186,
+      "force_loss": 0.002250265097245574,
+      "step": 41730
+    },
+    {
+      "epoch": 37.53597122302158,
+      "grad_norm": 0.08896561712026596,
+      "learning_rate": 2.325754066147145e-05,
+      "loss": 0.0145,
+      "step": 41740
+    },
+    {
+      "action_loss": 0.0012034709798172116,
+      "epoch": 37.53597122302158,
+      "step": 41740
+    },
+    {
+      "epoch": 37.53597122302158,
+      "step": 41740,
+      "torque_loss": 0.1256457418203354
+    },
+    {
+      "epoch": 37.53597122302158,
+      "force_loss": 0.0028368141502141953,
+      "step": 41740
+    },
+    {
+      "epoch": 37.54496402877698,
+      "grad_norm": 0.11042187362909317,
+      "learning_rate": 2.32342598035281e-05,
+      "loss": 0.0157,
+      "step": 41750
+    },
+    {
+      "action_loss": 0.002094205701723695,
+      "epoch": 37.54496402877698,
+      "step": 41750
+    },
+    {
+      "epoch": 37.54496402877698,
+      "step": 41750,
+      "torque_loss": 0.13694559037685394
+    },
+    {
+      "epoch": 37.54496402877698,
+      "force_loss": 0.008897403255105019,
+      "step": 41750
+    },
+    {
+      "epoch": 37.55395683453237,
+      "grad_norm": 0.22857968509197235,
+      "learning_rate": 2.321098707630991e-05,
+      "loss": 0.0163,
+      "step": 41760
+    },
+    {
+      "action_loss": 0.0010560353985056281,
+      "epoch": 37.55395683453237,
+      "step": 41760
+    },
+    {
+      "epoch": 37.55395683453237,
+      "step": 41760,
+      "torque_loss": 0.1850542575120926
+    },
+    {
+      "epoch": 37.55395683453237,
+      "force_loss": 0.0032666318584233522,
+      "step": 41760
+    },
+    {
+      "epoch": 37.56294964028777,
+      "grad_norm": 0.16816456615924835,
+      "learning_rate": 2.318772248688652e-05,
+      "loss": 0.0153,
+      "step": 41770
+    },
+    {
+      "action_loss": 0.0026116278022527695,
+      "epoch": 37.56294964028777,
+      "step": 41770
+    },
+    {
+      "epoch": 37.56294964028777,
+      "step": 41770,
+      "torque_loss": 0.16919898986816406
+    },
+    {
+      "epoch": 37.56294964028777,
+      "force_loss": 0.0057323165237903595,
+      "step": 41770
+    },
+    {
+      "epoch": 37.57194244604317,
+      "grad_norm": 0.49228811264038086,
+      "learning_rate": 2.3164466042325107e-05,
+      "loss": 0.0167,
+      "step": 41780
+    },
+    {
+      "action_loss": 0.0015910168876871467,
+      "epoch": 37.57194244604317,
+      "step": 41780
+    },
+    {
+      "epoch": 37.57194244604317,
+      "step": 41780,
+      "torque_loss": 0.1640763133764267
+    },
+    {
+      "epoch": 37.57194244604317,
+      "force_loss": 0.002082190243527293,
+      "step": 41780
+    },
+    {
+      "epoch": 37.580935251798564,
+      "grad_norm": 0.17482158541679382,
+      "learning_rate": 2.3141217749690353e-05,
+      "loss": 0.0174,
+      "step": 41790
+    },
+    {
+      "action_loss": 0.006881105247884989,
+      "epoch": 37.580935251798564,
+      "step": 41790
+    },
+    {
+      "epoch": 37.580935251798564,
+      "step": 41790,
+      "torque_loss": 0.15449021756649017
+    },
+    {
+      "epoch": 37.580935251798564,
+      "force_loss": 0.011124883778393269,
+      "step": 41790
+    },
+    {
+      "epoch": 37.589928057553955,
+      "grad_norm": 0.11784825474023819,
+      "learning_rate": 2.3117977616044466e-05,
+      "loss": 0.0153,
+      "step": 41800
+    },
+    {
+      "action_loss": 0.002115255920216441,
+      "epoch": 37.589928057553955,
+      "step": 41800
+    },
+    {
+      "epoch": 37.589928057553955,
+      "step": 41800,
+      "torque_loss": 0.0910719633102417
+    },
+    {
+      "epoch": 37.589928057553955,
+      "force_loss": 0.0017877034842967987,
+      "step": 41800
+    },
+    {
+      "epoch": 37.59892086330935,
+      "grad_norm": 0.5862287282943726,
+      "learning_rate": 2.309474564844722e-05,
+      "loss": 0.0177,
+      "step": 41810
+    },
+    {
+      "action_loss": 0.0015137457521632314,
+      "epoch": 37.59892086330935,
+      "step": 41810
+    },
+    {
+      "epoch": 37.59892086330935,
+      "step": 41810,
+      "torque_loss": 0.12903255224227905
+    },
+    {
+      "epoch": 37.59892086330935,
+      "force_loss": 0.002219780581071973,
+      "step": 41810
+    },
+    {
+      "epoch": 37.60791366906475,
+      "grad_norm": 0.0833015888929367,
+      "learning_rate": 2.307152185395585e-05,
+      "loss": 0.0167,
+      "step": 41820
+    },
+    {
+      "action_loss": 0.0014900037785992026,
+      "epoch": 37.60791366906475,
+      "step": 41820
+    },
+    {
+      "epoch": 37.60791366906475,
+      "step": 41820,
+      "torque_loss": 0.13498614728450775
+    },
+    {
+      "epoch": 37.60791366906475,
+      "force_loss": 0.0035768242087215185,
+      "step": 41820
+    },
+    {
+      "epoch": 37.61690647482014,
+      "grad_norm": 0.45619750022888184,
+      "learning_rate": 2.3048306239625144e-05,
+      "loss": 0.016,
+      "step": 41830
+    },
+    {
+      "action_loss": 0.002615357981994748,
+      "epoch": 37.61690647482014,
+      "step": 41830
+    },
+    {
+      "epoch": 37.61690647482014,
+      "step": 41830,
+      "torque_loss": 0.1370583325624466
+    },
+    {
+      "epoch": 37.61690647482014,
+      "force_loss": 0.004041640553623438,
+      "step": 41830
+    },
+    {
+      "epoch": 37.62589928057554,
+      "grad_norm": 0.2658224105834961,
+      "learning_rate": 2.3025098812507378e-05,
+      "loss": 0.0206,
+      "step": 41840
+    },
+    {
+      "action_loss": 0.0019492382416501641,
+      "epoch": 37.62589928057554,
+      "step": 41840
+    },
+    {
+      "epoch": 37.62589928057554,
+      "step": 41840,
+      "torque_loss": 0.1327981799840927
+    },
+    {
+      "epoch": 37.62589928057554,
+      "force_loss": 0.00436962116509676,
+      "step": 41840
+    },
+    {
+      "epoch": 37.634892086330936,
+      "grad_norm": 0.273560494184494,
+      "learning_rate": 2.3001899579652366e-05,
+      "loss": 0.015,
+      "step": 41850
+    },
+    {
+      "action_loss": 0.0014214725233614445,
+      "epoch": 37.634892086330936,
+      "step": 41850
+    },
+    {
+      "epoch": 37.634892086330936,
+      "step": 41850,
+      "torque_loss": 0.15702326595783234
+    },
+    {
+      "epoch": 37.634892086330936,
+      "force_loss": 0.0036117613781243563,
+      "step": 41850
+    },
+    {
+      "epoch": 37.643884892086334,
+      "grad_norm": 0.08339393138885498,
+      "learning_rate": 2.2978708548107393e-05,
+      "loss": 0.0156,
+      "step": 41860
+    },
+    {
+      "action_loss": 0.001197000965476036,
+      "epoch": 37.643884892086334,
+      "step": 41860
+    },
+    {
+      "epoch": 37.643884892086334,
+      "step": 41860,
+      "torque_loss": 0.16042494773864746
+    },
+    {
+      "epoch": 37.643884892086334,
+      "force_loss": 0.002590571530163288,
+      "step": 41860
+    },
+    {
+      "epoch": 37.652877697841724,
+      "grad_norm": 0.14148195087909698,
+      "learning_rate": 2.2955525724917348e-05,
+      "loss": 0.0156,
+      "step": 41870
+    },
+    {
+      "action_loss": 0.001695584156550467,
+      "epoch": 37.652877697841724,
+      "step": 41870
+    },
+    {
+      "epoch": 37.652877697841724,
+      "step": 41870,
+      "torque_loss": 0.1457749903202057
+    },
+    {
+      "epoch": 37.652877697841724,
+      "force_loss": 0.0023889620788395405,
+      "step": 41870
+    },
+    {
+      "epoch": 37.66187050359712,
+      "grad_norm": 0.31962788105010986,
+      "learning_rate": 2.2932351117124477e-05,
+      "loss": 0.0189,
+      "step": 41880
+    },
+    {
+      "action_loss": 0.002825948176905513,
+      "epoch": 37.66187050359712,
+      "step": 41880
+    },
+    {
+      "epoch": 37.66187050359712,
+      "step": 41880,
+      "torque_loss": 0.12647639214992523
+    },
+    {
+      "epoch": 37.66187050359712,
+      "force_loss": 0.005641849245876074,
+      "step": 41880
+    },
+    {
+      "epoch": 37.67086330935252,
+      "grad_norm": 0.09548363834619522,
+      "learning_rate": 2.29091847317687e-05,
+      "loss": 0.0155,
+      "step": 41890
+    },
+    {
+      "action_loss": 0.002940859878435731,
+      "epoch": 37.67086330935252,
+      "step": 41890
+    },
+    {
+      "epoch": 37.67086330935252,
+      "step": 41890,
+      "torque_loss": 0.11217710375785828
+    },
+    {
+      "epoch": 37.67086330935252,
+      "force_loss": 0.0021758966613560915,
+      "step": 41890
+    },
+    {
+      "epoch": 37.67985611510792,
+      "grad_norm": 0.20715530216693878,
+      "learning_rate": 2.2886026575887277e-05,
+      "loss": 0.0169,
+      "step": 41900
+    },
+    {
+      "action_loss": 0.0019089201232418418,
+      "epoch": 37.67985611510792,
+      "step": 41900
+    },
+    {
+      "epoch": 37.67985611510792,
+      "step": 41900,
+      "torque_loss": 0.1353653371334076
+    },
+    {
+      "epoch": 37.67985611510792,
+      "force_loss": 0.002920748433098197,
+      "step": 41900
+    },
+    {
+      "epoch": 37.68884892086331,
+      "grad_norm": 0.2596253752708435,
+      "learning_rate": 2.2862876656515094e-05,
+      "loss": 0.0166,
+      "step": 41910
+    },
+    {
+      "action_loss": 0.001412943354807794,
+      "epoch": 37.68884892086331,
+      "step": 41910
+    },
+    {
+      "epoch": 37.68884892086331,
+      "step": 41910,
+      "torque_loss": 0.1325875222682953
+    },
+    {
+      "epoch": 37.68884892086331,
+      "force_loss": 0.0016223323764279485,
+      "step": 41910
+    },
+    {
+      "epoch": 37.697841726618705,
+      "grad_norm": 0.2145116627216339,
+      "learning_rate": 2.2839734980684464e-05,
+      "loss": 0.0161,
+      "step": 41920
+    },
+    {
+      "action_loss": 0.007583460304886103,
+      "epoch": 37.697841726618705,
+      "step": 41920
+    },
+    {
+      "epoch": 37.697841726618705,
+      "step": 41920,
+      "torque_loss": 0.13984991610050201
+    },
+    {
+      "epoch": 37.697841726618705,
+      "force_loss": 0.012093673460185528,
+      "step": 41920
+    },
+    {
+      "epoch": 37.7068345323741,
+      "grad_norm": 0.1525728553533554,
+      "learning_rate": 2.281660155542522e-05,
+      "loss": 0.0186,
+      "step": 41930
+    },
+    {
+      "action_loss": 0.025422945618629456,
+      "epoch": 37.7068345323741,
+      "step": 41930
+    },
+    {
+      "epoch": 37.7068345323741,
+      "step": 41930,
+      "torque_loss": 0.10879912972450256
+    },
+    {
+      "epoch": 37.7068345323741,
+      "force_loss": 0.030627774074673653,
+      "step": 41930
+    },
+    {
+      "epoch": 37.71582733812949,
+      "grad_norm": 0.1916947066783905,
+      "learning_rate": 2.279347638776469e-05,
+      "loss": 0.0169,
+      "step": 41940
+    },
+    {
+      "action_loss": 0.004034934099763632,
+      "epoch": 37.71582733812949,
+      "step": 41940
+    },
+    {
+      "epoch": 37.71582733812949,
+      "step": 41940,
+      "torque_loss": 0.10615388303995132
+    },
+    {
+      "epoch": 37.71582733812949,
+      "force_loss": 0.004340244457125664,
+      "step": 41940
+    },
+    {
+      "epoch": 37.72482014388489,
+      "grad_norm": 0.4816558361053467,
+      "learning_rate": 2.2770359484727665e-05,
+      "loss": 0.0167,
+      "step": 41950
+    },
+    {
+      "action_loss": 0.0015975426649674773,
+      "epoch": 37.72482014388489,
+      "step": 41950
+    },
+    {
+      "epoch": 37.72482014388489,
+      "step": 41950,
+      "torque_loss": 0.09869460016489029
+    },
+    {
+      "epoch": 37.72482014388489,
+      "force_loss": 0.0016976045444607735,
+      "step": 41950
+    },
+    {
+      "epoch": 37.73381294964029,
+      "grad_norm": 0.11220408231019974,
+      "learning_rate": 2.27472508533365e-05,
+      "loss": 0.0169,
+      "step": 41960
+    },
+    {
+      "action_loss": 0.002514922758564353,
+      "epoch": 37.73381294964029,
+      "step": 41960
+    },
+    {
+      "epoch": 37.73381294964029,
+      "step": 41960,
+      "torque_loss": 0.1607474833726883
+    },
+    {
+      "epoch": 37.73381294964029,
+      "force_loss": 0.014331807382404804,
+      "step": 41960
+    },
+    {
+      "epoch": 37.742805755395686,
+      "grad_norm": 0.21857528388500214,
+      "learning_rate": 2.2724150500610948e-05,
+      "loss": 0.0202,
+      "step": 41970
+    },
+    {
+      "action_loss": 0.001497595920227468,
+      "epoch": 37.742805755395686,
+      "step": 41970
+    },
+    {
+      "epoch": 37.742805755395686,
+      "step": 41970,
+      "torque_loss": 0.12348630279302597
+    },
+    {
+      "epoch": 37.742805755395686,
+      "force_loss": 0.0011046187719330192,
+      "step": 41970
+    },
+    {
+      "epoch": 37.75179856115108,
+      "grad_norm": 0.3066596984863281,
+      "learning_rate": 2.2701058433568302e-05,
+      "loss": 0.0188,
+      "step": 41980
+    },
+    {
+      "action_loss": 0.0010938197374343872,
+      "epoch": 37.75179856115108,
+      "step": 41980
+    },
+    {
+      "epoch": 37.75179856115108,
+      "step": 41980,
+      "torque_loss": 0.09929013252258301
+    },
+    {
+      "epoch": 37.75179856115108,
+      "force_loss": 0.0019019526662304997,
+      "step": 41980
+    },
+    {
+      "epoch": 37.760791366906474,
+      "grad_norm": 0.13061729073524475,
+      "learning_rate": 2.2677974659223318e-05,
+      "loss": 0.0213,
+      "step": 41990
+    },
+    {
+      "action_loss": 0.007037907838821411,
+      "epoch": 37.760791366906474,
+      "step": 41990
+    },
+    {
+      "epoch": 37.760791366906474,
+      "step": 41990,
+      "torque_loss": 0.13844385743141174
+    },
+    {
+      "epoch": 37.760791366906474,
+      "force_loss": 0.00899954978376627,
+      "step": 41990
+    },
+    {
+      "epoch": 37.76978417266187,
+      "grad_norm": 0.29756227135658264,
+      "learning_rate": 2.2654899184588235e-05,
+      "loss": 0.0175,
+      "step": 42000
+    },
+    {
+      "action_loss": 0.0016529156127944589,
+      "epoch": 37.76978417266187,
+      "step": 42000
+    },
+    {
+      "epoch": 37.76978417266187,
+      "step": 42000,
+      "torque_loss": 0.17188672721385956
+    },
+    {
+      "epoch": 37.76978417266187,
+      "force_loss": 0.0018125292845070362,
+      "step": 42000
+    },
+    {
+      "epoch": 37.77877697841727,
+      "grad_norm": 0.1905078887939453,
+      "learning_rate": 2.2631832016672756e-05,
+      "loss": 0.0155,
+      "step": 42010
+    },
+    {
+      "action_loss": 0.0019595816265791655,
+      "epoch": 37.77877697841727,
+      "step": 42010
+    },
+    {
+      "epoch": 37.77877697841727,
+      "step": 42010,
+      "torque_loss": 0.14393800497055054
+    },
+    {
+      "epoch": 37.77877697841727,
+      "force_loss": 0.0009099827148020267,
+      "step": 42010
+    },
+    {
+      "epoch": 37.78776978417266,
+      "grad_norm": 0.3305841088294983,
+      "learning_rate": 2.2608773162484127e-05,
+      "loss": 0.0158,
+      "step": 42020
+    },
+    {
+      "action_loss": 0.005438648164272308,
+      "epoch": 37.78776978417266,
+      "step": 42020
+    },
+    {
+      "epoch": 37.78776978417266,
+      "step": 42020,
+      "torque_loss": 0.12238804250955582
+    },
+    {
+      "epoch": 37.78776978417266,
+      "force_loss": 0.009795799851417542,
+      "step": 42020
+    },
+    {
+      "epoch": 37.79676258992806,
+      "grad_norm": 0.30900031328201294,
+      "learning_rate": 2.2585722629026958e-05,
+      "loss": 0.0164,
+      "step": 42030
+    },
+    {
+      "action_loss": 0.0017492446349933743,
+      "epoch": 37.79676258992806,
+      "step": 42030
+    },
+    {
+      "epoch": 37.79676258992806,
+      "step": 42030,
+      "torque_loss": 0.1076723113656044
+    },
+    {
+      "epoch": 37.79676258992806,
+      "force_loss": 0.0033653213176876307,
+      "step": 42030
+    },
+    {
+      "epoch": 37.805755395683455,
+      "grad_norm": 0.07750316709280014,
+      "learning_rate": 2.2562680423303457e-05,
+      "loss": 0.0156,
+      "step": 42040
+    },
+    {
+      "action_loss": 0.031596552580595016,
+      "epoch": 37.805755395683455,
+      "step": 42040
+    },
+    {
+      "epoch": 37.805755395683455,
+      "step": 42040,
+      "torque_loss": 0.13096915185451508
+    },
+    {
+      "epoch": 37.805755395683455,
+      "force_loss": 0.045666519552469254,
+      "step": 42040
+    },
+    {
+      "epoch": 37.814748201438846,
+      "grad_norm": 0.08575436472892761,
+      "learning_rate": 2.2539646552313165e-05,
+      "loss": 0.0201,
+      "step": 42050
+    },
+    {
+      "action_loss": 0.01117153000086546,
+      "epoch": 37.814748201438846,
+      "step": 42050
+    },
+    {
+      "epoch": 37.814748201438846,
+      "step": 42050,
+      "torque_loss": 0.18546605110168457
+    },
+    {
+      "epoch": 37.814748201438846,
+      "force_loss": 0.015107501298189163,
+      "step": 42050
+    },
+    {
+      "epoch": 37.82374100719424,
+      "grad_norm": 0.1526290327310562,
+      "learning_rate": 2.251662102305322e-05,
+      "loss": 0.0163,
+      "step": 42060
+    },
+    {
+      "action_loss": 0.0016643124399706721,
+      "epoch": 37.82374100719424,
+      "step": 42060
+    },
+    {
+      "epoch": 37.82374100719424,
+      "step": 42060,
+      "torque_loss": 0.1225610077381134
+    },
+    {
+      "epoch": 37.82374100719424,
+      "force_loss": 0.005740236956626177,
+      "step": 42060
+    },
+    {
+      "epoch": 37.83273381294964,
+      "grad_norm": 0.10917676240205765,
+      "learning_rate": 2.2493603842518152e-05,
+      "loss": 0.0143,
+      "step": 42070
+    },
+    {
+      "action_loss": 0.003165190340951085,
+      "epoch": 37.83273381294964,
+      "step": 42070
+    },
+    {
+      "epoch": 37.83273381294964,
+      "step": 42070,
+      "torque_loss": 0.1288745254278183
+    },
+    {
+      "epoch": 37.83273381294964,
+      "force_loss": 0.0036003708373755217,
+      "step": 42070
+    },
+    {
+      "epoch": 37.84172661870504,
+      "grad_norm": 0.18314294517040253,
+      "learning_rate": 2.2470595017699974e-05,
+      "loss": 0.0155,
+      "step": 42080
+    },
+    {
+      "action_loss": 0.0015404910081997514,
+      "epoch": 37.84172661870504,
+      "step": 42080
+    },
+    {
+      "epoch": 37.84172661870504,
+      "step": 42080,
+      "torque_loss": 0.13327224552631378
+    },
+    {
+      "epoch": 37.84172661870504,
+      "force_loss": 0.0014931490877643228,
+      "step": 42080
+    },
+    {
+      "epoch": 37.85071942446043,
+      "grad_norm": 0.08783494681119919,
+      "learning_rate": 2.244759455558816e-05,
+      "loss": 0.0167,
+      "step": 42090
+    },
+    {
+      "action_loss": 0.0007188048330135643,
+      "epoch": 37.85071942446043,
+      "step": 42090
+    },
+    {
+      "epoch": 37.85071942446043,
+      "step": 42090,
+      "torque_loss": 0.11901036649942398
+    },
+    {
+      "epoch": 37.85071942446043,
+      "force_loss": 0.0007325056940317154,
+      "step": 42090
+    },
+    {
+      "epoch": 37.85971223021583,
+      "grad_norm": 0.11911371350288391,
+      "learning_rate": 2.2424602463169614e-05,
+      "loss": 0.0162,
+      "step": 42100
+    },
+    {
+      "action_loss": 0.0037666999269276857,
+      "epoch": 37.85971223021583,
+      "step": 42100
+    },
+    {
+      "epoch": 37.85971223021583,
+      "step": 42100,
+      "torque_loss": 0.09223818778991699
+    },
+    {
+      "epoch": 37.85971223021583,
+      "force_loss": 0.001488087815232575,
+      "step": 42100
+    },
+    {
+      "epoch": 37.868705035971225,
+      "grad_norm": 0.10478778183460236,
+      "learning_rate": 2.2401618747428776e-05,
+      "loss": 0.0159,
+      "step": 42110
+    },
+    {
+      "action_loss": 0.001334990723989904,
+      "epoch": 37.868705035971225,
+      "step": 42110
+    },
+    {
+      "epoch": 37.868705035971225,
+      "step": 42110,
+      "torque_loss": 0.1115121841430664
+    },
+    {
+      "epoch": 37.868705035971225,
+      "force_loss": 0.002197865629568696,
+      "step": 42110
+    },
+    {
+      "epoch": 37.87769784172662,
+      "grad_norm": 0.052730903029441833,
+      "learning_rate": 2.237864341534747e-05,
+      "loss": 0.0171,
+      "step": 42120
+    },
+    {
+      "action_loss": 0.0028694781940430403,
+      "epoch": 37.87769784172662,
+      "step": 42120
+    },
+    {
+      "epoch": 37.87769784172662,
+      "step": 42120,
+      "torque_loss": 0.1752682477235794
+    },
+    {
+      "epoch": 37.87769784172662,
+      "force_loss": 0.009043095633387566,
+      "step": 42120
+    },
+    {
+      "epoch": 37.88669064748201,
+      "grad_norm": 0.05406825989484787,
+      "learning_rate": 2.2355676473904998e-05,
+      "loss": 0.0148,
+      "step": 42130
+    },
+    {
+      "action_loss": 0.0006184595404192805,
+      "epoch": 37.88669064748201,
+      "step": 42130
+    },
+    {
+      "epoch": 37.88669064748201,
+      "step": 42130,
+      "torque_loss": 0.12273558974266052
+    },
+    {
+      "epoch": 37.88669064748201,
+      "force_loss": 0.0016919871559366584,
+      "step": 42130
+    },
+    {
+      "epoch": 37.89568345323741,
+      "grad_norm": 0.15935902297496796,
+      "learning_rate": 2.2332717930078108e-05,
+      "loss": 0.0171,
+      "step": 42140
+    },
+    {
+      "action_loss": 0.0024945142213255167,
+      "epoch": 37.89568345323741,
+      "step": 42140
+    },
+    {
+      "epoch": 37.89568345323741,
+      "step": 42140,
+      "torque_loss": 0.10992572456598282
+    },
+    {
+      "epoch": 37.89568345323741,
+      "force_loss": 0.007920256815850735,
+      "step": 42140
+    },
+    {
+      "epoch": 37.90467625899281,
+      "grad_norm": 0.19391345977783203,
+      "learning_rate": 2.2309767790840992e-05,
+      "loss": 0.0212,
+      "step": 42150
+    },
+    {
+      "action_loss": 0.016867203637957573,
+      "epoch": 37.90467625899281,
+      "step": 42150
+    },
+    {
+      "epoch": 37.90467625899281,
+      "step": 42150,
+      "torque_loss": 0.1651969999074936
+    },
+    {
+      "epoch": 37.90467625899281,
+      "force_loss": 0.02321917749941349,
+      "step": 42150
+    },
+    {
+      "epoch": 37.9136690647482,
+      "grad_norm": 0.08190567046403885,
+      "learning_rate": 2.228682606316529e-05,
+      "loss": 0.019,
+      "step": 42160
+    },
+    {
+      "action_loss": 0.0016315663233399391,
+      "epoch": 37.9136690647482,
+      "step": 42160
+    },
+    {
+      "epoch": 37.9136690647482,
+      "step": 42160,
+      "torque_loss": 0.12805941700935364
+    },
+    {
+      "epoch": 37.9136690647482,
+      "force_loss": 0.0020304303616285324,
+      "step": 42160
+    },
+    {
+      "epoch": 37.922661870503596,
+      "grad_norm": 0.2823522686958313,
+      "learning_rate": 2.2263892754020138e-05,
+      "loss": 0.0168,
+      "step": 42170
+    },
+    {
+      "action_loss": 0.0018545013153925538,
+      "epoch": 37.922661870503596,
+      "step": 42170
+    },
+    {
+      "epoch": 37.922661870503596,
+      "step": 42170,
+      "torque_loss": 0.11755432933568954
+    },
+    {
+      "epoch": 37.922661870503596,
+      "force_loss": 0.0014761259080842137,
+      "step": 42170
+    },
+    {
+      "epoch": 37.931654676258994,
+      "grad_norm": 0.11751024425029755,
+      "learning_rate": 2.2240967870372004e-05,
+      "loss": 0.0161,
+      "step": 42180
+    },
+    {
+      "action_loss": 0.01340011041611433,
+      "epoch": 37.931654676258994,
+      "step": 42180
+    },
+    {
+      "epoch": 37.931654676258994,
+      "step": 42180,
+      "torque_loss": 0.11241209506988525
+    },
+    {
+      "epoch": 37.931654676258994,
+      "force_loss": 0.013013034127652645,
+      "step": 42180
+    },
+    {
+      "epoch": 37.94064748201439,
+      "grad_norm": 0.22758176922798157,
+      "learning_rate": 2.2218051419184933e-05,
+      "loss": 0.0175,
+      "step": 42190
+    },
+    {
+      "action_loss": 0.0016600164817646146,
+      "epoch": 37.94064748201439,
+      "step": 42190
+    },
+    {
+      "epoch": 37.94064748201439,
+      "step": 42190,
+      "torque_loss": 0.11962270736694336
+    },
+    {
+      "epoch": 37.94064748201439,
+      "force_loss": 0.0010911537101492286,
+      "step": 42190
+    },
+    {
+      "epoch": 37.94964028776978,
+      "grad_norm": 0.283041775226593,
+      "learning_rate": 2.219514340742026e-05,
+      "loss": 0.0188,
+      "step": 42200
+    },
+    {
+      "action_loss": 0.004071592353284359,
+      "epoch": 37.94964028776978,
+      "step": 42200
+    },
+    {
+      "epoch": 37.94964028776978,
+      "step": 42200,
+      "torque_loss": 0.17521530389785767
+    },
+    {
+      "epoch": 37.94964028776978,
+      "force_loss": 0.004676694516092539,
+      "step": 42200
+    },
+    {
+      "epoch": 37.95863309352518,
+      "grad_norm": 0.3204297423362732,
+      "learning_rate": 2.2172243842036898e-05,
+      "loss": 0.0159,
+      "step": 42210
+    },
+    {
+      "action_loss": 0.0008400164660997689,
+      "epoch": 37.95863309352518,
+      "step": 42210
+    },
+    {
+      "epoch": 37.95863309352518,
+      "step": 42210,
+      "torque_loss": 0.14027972519397736
+    },
+    {
+      "epoch": 37.95863309352518,
+      "force_loss": 0.0009440624271519482,
+      "step": 42210
+    },
+    {
+      "epoch": 37.96762589928058,
+      "grad_norm": 0.2375713735818863,
+      "learning_rate": 2.2149352729991107e-05,
+      "loss": 0.0157,
+      "step": 42220
+    },
+    {
+      "action_loss": 0.003954483196139336,
+      "epoch": 37.96762589928058,
+      "step": 42220
+    },
+    {
+      "epoch": 37.96762589928058,
+      "step": 42220,
+      "torque_loss": 0.08916950225830078
+    },
+    {
+      "epoch": 37.96762589928058,
+      "force_loss": 0.0017460359958931804,
+      "step": 42220
+    },
+    {
+      "epoch": 37.976618705035975,
+      "grad_norm": 0.08957651257514954,
+      "learning_rate": 2.2126470078236605e-05,
+      "loss": 0.0166,
+      "step": 42230
+    },
+    {
+      "action_loss": 0.0007148857112042606,
+      "epoch": 37.976618705035975,
+      "step": 42230
+    },
+    {
+      "epoch": 37.976618705035975,
+      "step": 42230,
+      "torque_loss": 0.1047196015715599
+    },
+    {
+      "epoch": 37.976618705035975,
+      "force_loss": 0.0006302558467723429,
+      "step": 42230
+    },
+    {
+      "epoch": 37.985611510791365,
+      "grad_norm": 0.13261087238788605,
+      "learning_rate": 2.2103595893724533e-05,
+      "loss": 0.0154,
+      "step": 42240
+    },
+    {
+      "action_loss": 0.0010683628497645259,
+      "epoch": 37.985611510791365,
+      "step": 42240
+    },
+    {
+      "epoch": 37.985611510791365,
+      "step": 42240,
+      "torque_loss": 0.1169266328215599
+    },
+    {
+      "epoch": 37.985611510791365,
+      "force_loss": 0.0010190032189711928,
+      "step": 42240
+    },
+    {
+      "epoch": 37.99460431654676,
+      "grad_norm": 0.1813839077949524,
+      "learning_rate": 2.208073018340345e-05,
+      "loss": 0.0156,
+      "step": 42250
+    },
+    {
+      "action_loss": 0.0008676950819790363,
+      "epoch": 37.99460431654676,
+      "step": 42250
+    },
+    {
+      "epoch": 37.99460431654676,
+      "step": 42250,
+      "torque_loss": 0.13542599976062775
+    },
+    {
+      "epoch": 37.99460431654676,
+      "force_loss": 0.0010492483852431178,
+      "step": 42250
+    },
+    {
+      "epoch": 38.00359712230216,
+      "grad_norm": 0.26701563596725464,
+      "learning_rate": 2.2057872954219405e-05,
+      "loss": 0.0186,
+      "step": 42260
+    },
+    {
+      "action_loss": 0.0012737991055473685,
+      "epoch": 38.00359712230216,
+      "step": 42260
+    },
+    {
+      "epoch": 38.00359712230216,
+      "step": 42260,
+      "torque_loss": 0.11293216794729233
+    },
+    {
+      "epoch": 38.00359712230216,
+      "force_loss": 0.0013094739988446236,
+      "step": 42260
+    },
+    {
+      "epoch": 38.01258992805755,
+      "grad_norm": 0.3566124439239502,
+      "learning_rate": 2.203502421311575e-05,
+      "loss": 0.018,
+      "step": 42270
+    },
+    {
+      "action_loss": 0.0052148811519145966,
+      "epoch": 38.01258992805755,
+      "step": 42270
+    },
+    {
+      "epoch": 38.01258992805755,
+      "step": 42270,
+      "torque_loss": 0.13218361139297485
+    },
+    {
+      "epoch": 38.01258992805755,
+      "force_loss": 0.005130449775606394,
+      "step": 42270
+    },
+    {
+      "epoch": 38.02158273381295,
+      "grad_norm": 0.28244003653526306,
+      "learning_rate": 2.2012183967033388e-05,
+      "loss": 0.015,
+      "step": 42280
+    },
+    {
+      "action_loss": 0.0017035113414749503,
+      "epoch": 38.02158273381295,
+      "step": 42280
+    },
+    {
+      "epoch": 38.02158273381295,
+      "step": 42280,
+      "torque_loss": 0.12922503054141998
+    },
+    {
+      "epoch": 38.02158273381295,
+      "force_loss": 0.0009031973895616829,
+      "step": 42280
+    },
+    {
+      "epoch": 38.030575539568346,
+      "grad_norm": 0.08576733618974686,
+      "learning_rate": 2.198935222291056e-05,
+      "loss": 0.0177,
+      "step": 42290
+    },
+    {
+      "action_loss": 0.0012579961912706494,
+      "epoch": 38.030575539568346,
+      "step": 42290
+    },
+    {
+      "epoch": 38.030575539568346,
+      "step": 42290,
+      "torque_loss": 0.12923459708690643
+    },
+    {
+      "epoch": 38.030575539568346,
+      "force_loss": 0.002322385786101222,
+      "step": 42290
+    },
+    {
+      "epoch": 38.039568345323744,
+      "grad_norm": 0.15510354936122894,
+      "learning_rate": 2.1966528987682948e-05,
+      "loss": 0.0163,
+      "step": 42300
+    },
+    {
+      "action_loss": 0.0021190308034420013,
+      "epoch": 38.039568345323744,
+      "step": 42300
+    },
+    {
+      "epoch": 38.039568345323744,
+      "step": 42300,
+      "torque_loss": 0.1178090050816536
+    },
+    {
+      "epoch": 38.039568345323744,
+      "force_loss": 0.00510313780978322,
+      "step": 42300
+    },
+    {
+      "epoch": 38.048561151079134,
+      "grad_norm": 0.37736085057258606,
+      "learning_rate": 2.194371426828365e-05,
+      "loss": 0.0169,
+      "step": 42310
+    },
+    {
+      "action_loss": 0.008922112174332142,
+      "epoch": 38.048561151079134,
+      "step": 42310
+    },
+    {
+      "epoch": 38.048561151079134,
+      "step": 42310,
+      "torque_loss": 0.1116734966635704
+    },
+    {
+      "epoch": 38.048561151079134,
+      "force_loss": 0.00866603385657072,
+      "step": 42310
+    },
+    {
+      "epoch": 38.05755395683453,
+      "grad_norm": 0.2952435314655304,
+      "learning_rate": 2.192090807164317e-05,
+      "loss": 0.0146,
+      "step": 42320
+    },
+    {
+      "action_loss": 0.006244678515940905,
+      "epoch": 38.05755395683453,
+      "step": 42320
+    },
+    {
+      "epoch": 38.05755395683453,
+      "step": 42320,
+      "torque_loss": 0.20723174512386322
+    },
+    {
+      "epoch": 38.05755395683453,
+      "force_loss": 0.0156722255051136,
+      "step": 42320
+    },
+    {
+      "epoch": 38.06654676258993,
+      "grad_norm": 0.16199855506420135,
+      "learning_rate": 2.1898110404689422e-05,
+      "loss": 0.0165,
+      "step": 42330
+    },
+    {
+      "action_loss": 0.012500171549618244,
+      "epoch": 38.06654676258993,
+      "step": 42330
+    },
+    {
+      "epoch": 38.06654676258993,
+      "step": 42330,
+      "torque_loss": 0.14990507066249847
+    },
+    {
+      "epoch": 38.06654676258993,
+      "force_loss": 0.010013852268457413,
+      "step": 42330
+    },
+    {
+      "epoch": 38.07553956834533,
+      "grad_norm": 0.2425270676612854,
+      "learning_rate": 2.1875321274347776e-05,
+      "loss": 0.0168,
+      "step": 42340
+    },
+    {
+      "action_loss": 0.0013319241115823388,
+      "epoch": 38.07553956834533,
+      "step": 42340
+    },
+    {
+      "epoch": 38.07553956834533,
+      "step": 42340,
+      "torque_loss": 0.15097080171108246
+    },
+    {
+      "epoch": 38.07553956834533,
+      "force_loss": 0.003812998766079545,
+      "step": 42340
+    },
+    {
+      "epoch": 38.08453237410072,
+      "grad_norm": 0.09177496284246445,
+      "learning_rate": 2.18525406875409e-05,
+      "loss": 0.0172,
+      "step": 42350
+    },
+    {
+      "action_loss": 0.001440504565834999,
+      "epoch": 38.08453237410072,
+      "step": 42350
+    },
+    {
+      "epoch": 38.08453237410072,
+      "step": 42350,
+      "torque_loss": 0.11144503206014633
+    },
+    {
+      "epoch": 38.08453237410072,
+      "force_loss": 0.002569103380665183,
+      "step": 42350
+    },
+    {
+      "epoch": 38.093525179856115,
+      "grad_norm": 0.4311312735080719,
+      "learning_rate": 2.1829768651188997e-05,
+      "loss": 0.014,
+      "step": 42360
+    },
+    {
+      "action_loss": 0.0008007041178643703,
+      "epoch": 38.093525179856115,
+      "step": 42360
+    },
+    {
+      "epoch": 38.093525179856115,
+      "step": 42360,
+      "torque_loss": 0.1781158447265625
+    },
+    {
+      "epoch": 38.093525179856115,
+      "force_loss": 0.0013596760109066963,
+      "step": 42360
+    },
+    {
+      "epoch": 38.10251798561151,
+      "grad_norm": 0.3303205668926239,
+      "learning_rate": 2.180700517220958e-05,
+      "loss": 0.0128,
+      "step": 42370
+    },
+    {
+      "action_loss": 0.002289427211508155,
+      "epoch": 38.10251798561151,
+      "step": 42370
+    },
+    {
+      "epoch": 38.10251798561151,
+      "step": 42370,
+      "torque_loss": 0.10773161798715591
+    },
+    {
+      "epoch": 38.10251798561151,
+      "force_loss": 0.005581540521234274,
+      "step": 42370
+    },
+    {
+      "epoch": 38.111510791366904,
+      "grad_norm": 0.15102706849575043,
+      "learning_rate": 2.1784250257517603e-05,
+      "loss": 0.0162,
+      "step": 42380
+    },
+    {
+      "action_loss": 0.003619803348556161,
+      "epoch": 38.111510791366904,
+      "step": 42380
+    },
+    {
+      "epoch": 38.111510791366904,
+      "step": 42380,
+      "torque_loss": 0.1511116772890091
+    },
+    {
+      "epoch": 38.111510791366904,
+      "force_loss": 0.0023690329398959875,
+      "step": 42380
+    },
+    {
+      "epoch": 38.1205035971223,
+      "grad_norm": 0.366584450006485,
+      "learning_rate": 2.1761503914025406e-05,
+      "loss": 0.0163,
+      "step": 42390
+    },
+    {
+      "action_loss": 0.002153784967958927,
+      "epoch": 38.1205035971223,
+      "step": 42390
+    },
+    {
+      "epoch": 38.1205035971223,
+      "step": 42390,
+      "torque_loss": 0.11411242932081223
+    },
+    {
+      "epoch": 38.1205035971223,
+      "force_loss": 0.002840594155713916,
+      "step": 42390
+    },
+    {
+      "epoch": 38.1294964028777,
+      "grad_norm": 0.14787189662456512,
+      "learning_rate": 2.1738766148642705e-05,
+      "loss": 0.0144,
+      "step": 42400
+    },
+    {
+      "action_loss": 0.004105289001017809,
+      "epoch": 38.1294964028777,
+      "step": 42400
+    },
+    {
+      "epoch": 38.1294964028777,
+      "step": 42400,
+      "torque_loss": 0.14026354253292084
+    },
+    {
+      "epoch": 38.1294964028777,
+      "force_loss": 0.00389136653393507,
+      "step": 42400
+    },
+    {
+      "epoch": 38.138489208633096,
+      "grad_norm": 0.15205298364162445,
+      "learning_rate": 2.1716036968276683e-05,
+      "loss": 0.017,
+      "step": 42410
+    },
+    {
+      "action_loss": 0.0013198625529184937,
+      "epoch": 38.138489208633096,
+      "step": 42410
+    },
+    {
+      "epoch": 38.138489208633096,
+      "step": 42410,
+      "torque_loss": 0.07007986307144165
+    },
+    {
+      "epoch": 38.138489208633096,
+      "force_loss": 0.005232428666204214,
+      "step": 42410
+    },
+    {
+      "epoch": 38.14748201438849,
+      "grad_norm": 0.12271703034639359,
+      "learning_rate": 2.1693316379831808e-05,
+      "loss": 0.0137,
+      "step": 42420
+    },
+    {
+      "action_loss": 0.0018609925173223019,
+      "epoch": 38.14748201438849,
+      "step": 42420
+    },
+    {
+      "epoch": 38.14748201438849,
+      "step": 42420,
+      "torque_loss": 0.16388960182666779
+    },
+    {
+      "epoch": 38.14748201438849,
+      "force_loss": 0.0015514730475842953,
+      "step": 42420
+    },
+    {
+      "epoch": 38.156474820143885,
+      "grad_norm": 0.14147064089775085,
+      "learning_rate": 2.1670604390210037e-05,
+      "loss": 0.0169,
+      "step": 42430
+    },
+    {
+      "action_loss": 0.0010573114268481731,
+      "epoch": 38.156474820143885,
+      "step": 42430
+    },
+    {
+      "epoch": 38.156474820143885,
+      "step": 42430,
+      "torque_loss": 0.15909452736377716
+    },
+    {
+      "epoch": 38.156474820143885,
+      "force_loss": 0.0011442030081525445,
+      "step": 42430
+    },
+    {
+      "epoch": 38.16546762589928,
+      "grad_norm": 0.3083512783050537,
+      "learning_rate": 2.1647901006310656e-05,
+      "loss": 0.0158,
+      "step": 42440
+    },
+    {
+      "action_loss": 0.0033844977151602507,
+      "epoch": 38.16546762589928,
+      "step": 42440
+    },
+    {
+      "epoch": 38.16546762589928,
+      "step": 42440,
+      "torque_loss": 0.09742780774831772
+    },
+    {
+      "epoch": 38.16546762589928,
+      "force_loss": 0.010735255666077137,
+      "step": 42440
+    },
+    {
+      "epoch": 38.17446043165467,
+      "grad_norm": 0.21156315505504608,
+      "learning_rate": 2.1625206235030353e-05,
+      "loss": 0.0175,
+      "step": 42450
+    },
+    {
+      "action_loss": 0.014417861588299274,
+      "epoch": 38.17446043165467,
+      "step": 42450
+    },
+    {
+      "epoch": 38.17446043165467,
+      "step": 42450,
+      "torque_loss": 0.12278424948453903
+    },
+    {
+      "epoch": 38.17446043165467,
+      "force_loss": 0.02201397716999054,
+      "step": 42450
+    },
+    {
+      "epoch": 38.18345323741007,
+      "grad_norm": 0.2523946464061737,
+      "learning_rate": 2.160252008326321e-05,
+      "loss": 0.0196,
+      "step": 42460
+    },
+    {
+      "action_loss": 0.005058374721556902,
+      "epoch": 38.18345323741007,
+      "step": 42460
+    },
+    {
+      "epoch": 38.18345323741007,
+      "step": 42460,
+      "torque_loss": 0.10087931156158447
+    },
+    {
+      "epoch": 38.18345323741007,
+      "force_loss": 0.0036174114793539047,
+      "step": 42460
+    },
+    {
+      "epoch": 38.19244604316547,
+      "grad_norm": 0.14239241182804108,
+      "learning_rate": 2.157984255790067e-05,
+      "loss": 0.0168,
+      "step": 42470
+    },
+    {
+      "action_loss": 0.0024516042321920395,
+      "epoch": 38.19244604316547,
+      "step": 42470
+    },
+    {
+      "epoch": 38.19244604316547,
+      "step": 42470,
+      "torque_loss": 0.13193939626216888
+    },
+    {
+      "epoch": 38.19244604316547,
+      "force_loss": 0.0028035396244376898,
+      "step": 42470
+    },
+    {
+      "epoch": 38.201438848920866,
+      "grad_norm": 0.22383785247802734,
+      "learning_rate": 2.1557173665831553e-05,
+      "loss": 0.0151,
+      "step": 42480
+    },
+    {
+      "action_loss": 0.0018066968768835068,
+      "epoch": 38.201438848920866,
+      "step": 42480
+    },
+    {
+      "epoch": 38.201438848920866,
+      "step": 42480,
+      "torque_loss": 0.16234518587589264
+    },
+    {
+      "epoch": 38.201438848920866,
+      "force_loss": 0.006594979669898748,
+      "step": 42480
+    },
+    {
+      "epoch": 38.210431654676256,
+      "grad_norm": 0.23351898789405823,
+      "learning_rate": 2.153451341394212e-05,
+      "loss": 0.0194,
+      "step": 42490
+    },
+    {
+      "action_loss": 0.001507127657532692,
+      "epoch": 38.210431654676256,
+      "step": 42490
+    },
+    {
+      "epoch": 38.210431654676256,
+      "step": 42490,
+      "torque_loss": 0.20244701206684113
+    },
+    {
+      "epoch": 38.210431654676256,
+      "force_loss": 0.0035984080750495195,
+      "step": 42490
+    },
+    {
+      "epoch": 38.219424460431654,
+      "grad_norm": 0.08941373229026794,
+      "learning_rate": 2.151186180911589e-05,
+      "loss": 0.0155,
+      "step": 42500
+    },
+    {
+      "action_loss": 0.0013980502262711525,
+      "epoch": 38.219424460431654,
+      "step": 42500
+    },
+    {
+      "epoch": 38.219424460431654,
+      "step": 42500,
+      "torque_loss": 0.13525958359241486
+    },
+    {
+      "epoch": 38.219424460431654,
+      "force_loss": 0.0053792609833180904,
+      "step": 42500
+    },
+    {
+      "epoch": 38.22841726618705,
+      "grad_norm": 0.1562460958957672,
+      "learning_rate": 2.1489218858233877e-05,
+      "loss": 0.0174,
+      "step": 42510
+    },
+    {
+      "action_loss": 0.001466710469685495,
+      "epoch": 38.22841726618705,
+      "step": 42510
+    },
+    {
+      "epoch": 38.22841726618705,
+      "step": 42510,
+      "torque_loss": 0.10757160931825638
+    },
+    {
+      "epoch": 38.22841726618705,
+      "force_loss": 0.0019725810270756483,
+      "step": 42510
+    },
+    {
+      "epoch": 38.23741007194245,
+      "grad_norm": 0.1342063993215561,
+      "learning_rate": 2.1466584568174392e-05,
+      "loss": 0.0195,
+      "step": 42520
+    },
+    {
+      "action_loss": 0.007308927830308676,
+      "epoch": 38.23741007194245,
+      "step": 42520
+    },
+    {
+      "epoch": 38.23741007194245,
+      "step": 42520,
+      "torque_loss": 0.10933581739664078
+    },
+    {
+      "epoch": 38.23741007194245,
+      "force_loss": 0.0039587477222085,
+      "step": 42520
+    },
+    {
+      "epoch": 38.24640287769784,
+      "grad_norm": 0.399356871843338,
+      "learning_rate": 2.1443958945813132e-05,
+      "loss": 0.0183,
+      "step": 42530
+    },
+    {
+      "action_loss": 0.0010282915318384767,
+      "epoch": 38.24640287769784,
+      "step": 42530
+    },
+    {
+      "epoch": 38.24640287769784,
+      "step": 42530,
+      "torque_loss": 0.13052575290203094
+    },
+    {
+      "epoch": 38.24640287769784,
+      "force_loss": 0.0025030726101249456,
+      "step": 42530
+    },
+    {
+      "epoch": 38.25539568345324,
+      "grad_norm": 0.1908128410577774,
+      "learning_rate": 2.1421341998023163e-05,
+      "loss": 0.0194,
+      "step": 42540
+    },
+    {
+      "action_loss": 0.021642101928591728,
+      "epoch": 38.25539568345324,
+      "step": 42540
+    },
+    {
+      "epoch": 38.25539568345324,
+      "step": 42540,
+      "torque_loss": 0.10879906266927719
+    },
+    {
+      "epoch": 38.25539568345324,
+      "force_loss": 0.02443387545645237,
+      "step": 42540
+    },
+    {
+      "epoch": 38.264388489208635,
+      "grad_norm": 0.3333670198917389,
+      "learning_rate": 2.139873373167491e-05,
+      "loss": 0.0178,
+      "step": 42550
+    },
+    {
+      "action_loss": 0.011166143231093884,
+      "epoch": 38.264388489208635,
+      "step": 42550
+    },
+    {
+      "epoch": 38.264388489208635,
+      "step": 42550,
+      "torque_loss": 0.09375491738319397
+    },
+    {
+      "epoch": 38.264388489208635,
+      "force_loss": 0.01609027199447155,
+      "step": 42550
+    },
+    {
+      "epoch": 38.273381294964025,
+      "grad_norm": 0.1628560870885849,
+      "learning_rate": 2.13761341536362e-05,
+      "loss": 0.0163,
+      "step": 42560
+    },
+    {
+      "action_loss": 0.0015759443631395698,
+      "epoch": 38.273381294964025,
+      "step": 42560
+    },
+    {
+      "epoch": 38.273381294964025,
+      "step": 42560,
+      "torque_loss": 0.14005199074745178
+    },
+    {
+      "epoch": 38.273381294964025,
+      "force_loss": 0.0020526631269603968,
+      "step": 42560
+    },
+    {
+      "epoch": 38.28237410071942,
+      "grad_norm": 0.16305041313171387,
+      "learning_rate": 2.1353543270772136e-05,
+      "loss": 0.0153,
+      "step": 42570
+    },
+    {
+      "action_loss": 0.0032874986063688993,
+      "epoch": 38.28237410071942,
+      "step": 42570
+    },
+    {
+      "epoch": 38.28237410071942,
+      "step": 42570,
+      "torque_loss": 0.19163213670253754
+    },
+    {
+      "epoch": 38.28237410071942,
+      "force_loss": 0.00328506319783628,
+      "step": 42570
+    },
+    {
+      "epoch": 38.29136690647482,
+      "grad_norm": 0.08315874636173248,
+      "learning_rate": 2.1330961089945297e-05,
+      "loss": 0.0174,
+      "step": 42580
+    },
+    {
+      "action_loss": 0.0031259332317858934,
+      "epoch": 38.29136690647482,
+      "step": 42580
+    },
+    {
+      "epoch": 38.29136690647482,
+      "step": 42580,
+      "torque_loss": 0.11001918464899063
+    },
+    {
+      "epoch": 38.29136690647482,
+      "force_loss": 0.003332960419356823,
+      "step": 42580
+    },
+    {
+      "epoch": 38.30035971223022,
+      "grad_norm": 0.20289459824562073,
+      "learning_rate": 2.130838761801548e-05,
+      "loss": 0.0198,
+      "step": 42590
+    },
+    {
+      "action_loss": 0.0016095106257125735,
+      "epoch": 38.30035971223022,
+      "step": 42590
+    },
+    {
+      "epoch": 38.30035971223022,
+      "step": 42590,
+      "torque_loss": 0.13450567424297333
+    },
+    {
+      "epoch": 38.30035971223022,
+      "force_loss": 0.0033097604755312204,
+      "step": 42590
+    },
+    {
+      "epoch": 38.30935251798561,
+      "grad_norm": 0.11926200985908508,
+      "learning_rate": 2.1285822861839966e-05,
+      "loss": 0.0165,
+      "step": 42600
+    },
+    {
+      "action_loss": 0.0022899971809238195,
+      "epoch": 38.30935251798561,
+      "step": 42600
+    },
+    {
+      "epoch": 38.30935251798561,
+      "step": 42600,
+      "torque_loss": 0.15427549183368683
+    },
+    {
+      "epoch": 38.30935251798561,
+      "force_loss": 0.0017726510995998979,
+      "step": 42600
+    },
+    {
+      "epoch": 38.318345323741006,
+      "grad_norm": 0.0767514556646347,
+      "learning_rate": 2.126326682827331e-05,
+      "loss": 0.0142,
+      "step": 42610
+    },
+    {
+      "action_loss": 0.0012972141848877072,
+      "epoch": 38.318345323741006,
+      "step": 42610
+    },
+    {
+      "epoch": 38.318345323741006,
+      "step": 42610,
+      "torque_loss": 0.12797673046588898
+    },
+    {
+      "epoch": 38.318345323741006,
+      "force_loss": 0.003716723993420601,
+      "step": 42610
+    },
+    {
+      "epoch": 38.327338129496404,
+      "grad_norm": 0.09448380023241043,
+      "learning_rate": 2.124071952416744e-05,
+      "loss": 0.0149,
+      "step": 42620
+    },
+    {
+      "action_loss": 0.0014604199677705765,
+      "epoch": 38.327338129496404,
+      "step": 42620
+    },
+    {
+      "epoch": 38.327338129496404,
+      "step": 42620,
+      "torque_loss": 0.12701942026615143
+    },
+    {
+      "epoch": 38.327338129496404,
+      "force_loss": 0.002675536321476102,
+      "step": 42620
+    },
+    {
+      "epoch": 38.3363309352518,
+      "grad_norm": 0.11596908420324326,
+      "learning_rate": 2.1218180956371634e-05,
+      "loss": 0.0136,
+      "step": 42630
+    },
+    {
+      "action_loss": 0.005087161436676979,
+      "epoch": 38.3363309352518,
+      "step": 42630
+    },
+    {
+      "epoch": 38.3363309352518,
+      "step": 42630,
+      "torque_loss": 0.07203080505132675
+    },
+    {
+      "epoch": 38.3363309352518,
+      "force_loss": 0.005233215633779764,
+      "step": 42630
+    },
+    {
+      "epoch": 38.34532374100719,
+      "grad_norm": 0.15473967790603638,
+      "learning_rate": 2.119565113173252e-05,
+      "loss": 0.0173,
+      "step": 42640
+    },
+    {
+      "action_loss": 0.012108213268220425,
+      "epoch": 38.34532374100719,
+      "step": 42640
+    },
+    {
+      "epoch": 38.34532374100719,
+      "step": 42640,
+      "torque_loss": 0.08756273239850998
+    },
+    {
+      "epoch": 38.34532374100719,
+      "force_loss": 0.018123289570212364,
+      "step": 42640
+    },
+    {
+      "epoch": 38.35431654676259,
+      "grad_norm": 0.1036149337887764,
+      "learning_rate": 2.1173130057094033e-05,
+      "loss": 0.0155,
+      "step": 42650
+    },
+    {
+      "action_loss": 0.0009599531185813248,
+      "epoch": 38.35431654676259,
+      "step": 42650
+    },
+    {
+      "epoch": 38.35431654676259,
+      "step": 42650,
+      "torque_loss": 0.11805254966020584
+    },
+    {
+      "epoch": 38.35431654676259,
+      "force_loss": 0.0018283972749486566,
+      "step": 42650
+    },
+    {
+      "epoch": 38.36330935251799,
+      "grad_norm": 0.1567688286304474,
+      "learning_rate": 2.115061773929753e-05,
+      "loss": 0.0147,
+      "step": 42660
+    },
+    {
+      "action_loss": 0.0014010765589773655,
+      "epoch": 38.36330935251799,
+      "step": 42660
+    },
+    {
+      "epoch": 38.36330935251799,
+      "step": 42660,
+      "torque_loss": 0.11579396575689316
+    },
+    {
+      "epoch": 38.36330935251799,
+      "force_loss": 0.0018234135350212455,
+      "step": 42660
+    },
+    {
+      "epoch": 38.37230215827338,
+      "grad_norm": 0.2620141804218292,
+      "learning_rate": 2.1128114185181623e-05,
+      "loss": 0.0172,
+      "step": 42670
+    },
+    {
+      "action_loss": 0.003817712888121605,
+      "epoch": 38.37230215827338,
+      "step": 42670
+    },
+    {
+      "epoch": 38.37230215827338,
+      "step": 42670,
+      "torque_loss": 0.10588070750236511
+    },
+    {
+      "epoch": 38.37230215827338,
+      "force_loss": 0.007507646456360817,
+      "step": 42670
+    },
+    {
+      "epoch": 38.381294964028775,
+      "grad_norm": 0.22709479928016663,
+      "learning_rate": 2.1105619401582317e-05,
+      "loss": 0.016,
+      "step": 42680
+    },
+    {
+      "action_loss": 0.0008732354617677629,
+      "epoch": 38.381294964028775,
+      "step": 42680
+    },
+    {
+      "epoch": 38.381294964028775,
+      "step": 42680,
+      "torque_loss": 0.08710754662752151
+    },
+    {
+      "epoch": 38.381294964028775,
+      "force_loss": 0.0022411842364817858,
+      "step": 42680
+    },
+    {
+      "epoch": 38.39028776978417,
+      "grad_norm": 0.21565333008766174,
+      "learning_rate": 2.1083133395332928e-05,
+      "loss": 0.0162,
+      "step": 42690
+    },
+    {
+      "action_loss": 0.003400401910766959,
+      "epoch": 38.39028776978417,
+      "step": 42690
+    },
+    {
+      "epoch": 38.39028776978417,
+      "step": 42690,
+      "torque_loss": 0.1361086517572403
+    },
+    {
+      "epoch": 38.39028776978417,
+      "force_loss": 0.0055230301804840565,
+      "step": 42690
+    },
+    {
+      "epoch": 38.39928057553957,
+      "grad_norm": 0.1529228836297989,
+      "learning_rate": 2.1060656173264082e-05,
+      "loss": 0.0151,
+      "step": 42700
+    },
+    {
+      "action_loss": 0.005311049986630678,
+      "epoch": 38.39928057553957,
+      "step": 42700
+    },
+    {
+      "epoch": 38.39928057553957,
+      "step": 42700,
+      "torque_loss": 0.14421053230762482
+    },
+    {
+      "epoch": 38.39928057553957,
+      "force_loss": 0.006098801735788584,
+      "step": 42700
+    },
+    {
+      "epoch": 38.40827338129496,
+      "grad_norm": 0.275186151266098,
+      "learning_rate": 2.103818774220383e-05,
+      "loss": 0.0226,
+      "step": 42710
+    },
+    {
+      "action_loss": 0.0014118897961452603,
+      "epoch": 38.40827338129496,
+      "step": 42710
+    },
+    {
+      "epoch": 38.40827338129496,
+      "step": 42710,
+      "torque_loss": 0.09845157712697983
+    },
+    {
+      "epoch": 38.40827338129496,
+      "force_loss": 0.0006217022310011089,
+      "step": 42710
+    },
+    {
+      "epoch": 38.41726618705036,
+      "grad_norm": 0.19079558551311493,
+      "learning_rate": 2.1015728108977412e-05,
+      "loss": 0.016,
+      "step": 42720
+    },
+    {
+      "action_loss": 0.017542317509651184,
+      "epoch": 38.41726618705036,
+      "step": 42720
+    },
+    {
+      "epoch": 38.41726618705036,
+      "step": 42720,
+      "torque_loss": 0.12345749139785767
+    },
+    {
+      "epoch": 38.41726618705036,
+      "force_loss": 0.017251184210181236,
+      "step": 42720
+    },
+    {
+      "epoch": 38.42625899280576,
+      "grad_norm": 0.23277685046195984,
+      "learning_rate": 2.0993277280407548e-05,
+      "loss": 0.0183,
+      "step": 42730
+    },
+    {
+      "action_loss": 0.007395673543214798,
+      "epoch": 38.42625899280576,
+      "step": 42730
+    },
+    {
+      "epoch": 38.42625899280576,
+      "step": 42730,
+      "torque_loss": 0.11387022584676743
+    },
+    {
+      "epoch": 38.42625899280576,
+      "force_loss": 0.011070760898292065,
+      "step": 42730
+    },
+    {
+      "epoch": 38.435251798561154,
+      "grad_norm": 0.06706535816192627,
+      "learning_rate": 2.0970835263314132e-05,
+      "loss": 0.0163,
+      "step": 42740
+    },
+    {
+      "action_loss": 0.0020951805636286736,
+      "epoch": 38.435251798561154,
+      "step": 42740
+    },
+    {
+      "epoch": 38.435251798561154,
+      "step": 42740,
+      "torque_loss": 0.1587638109922409
+    },
+    {
+      "epoch": 38.435251798561154,
+      "force_loss": 0.004596231505274773,
+      "step": 42740
+    },
+    {
+      "epoch": 38.444244604316545,
+      "grad_norm": 0.1208859458565712,
+      "learning_rate": 2.094840206451451e-05,
+      "loss": 0.0156,
+      "step": 42750
+    },
+    {
+      "action_loss": 0.0021530913654714823,
+      "epoch": 38.444244604316545,
+      "step": 42750
+    },
+    {
+      "epoch": 38.444244604316545,
+      "step": 42750,
+      "torque_loss": 0.14189133048057556
+    },
+    {
+      "epoch": 38.444244604316545,
+      "force_loss": 0.0074575818143785,
+      "step": 42750
+    },
+    {
+      "epoch": 38.45323741007194,
+      "grad_norm": 0.15104053914546967,
+      "learning_rate": 2.0925977690823273e-05,
+      "loss": 0.0159,
+      "step": 42760
+    },
+    {
+      "action_loss": 0.010436457581818104,
+      "epoch": 38.45323741007194,
+      "step": 42760
+    },
+    {
+      "epoch": 38.45323741007194,
+      "step": 42760,
+      "torque_loss": 0.11887097358703613
+    },
+    {
+      "epoch": 38.45323741007194,
+      "force_loss": 0.015443484298884869,
+      "step": 42760
+    },
+    {
+      "epoch": 38.46223021582734,
+      "grad_norm": 0.12667535245418549,
+      "learning_rate": 2.0903562149052364e-05,
+      "loss": 0.0147,
+      "step": 42770
+    },
+    {
+      "action_loss": 0.0013711719075217843,
+      "epoch": 38.46223021582734,
+      "step": 42770
+    },
+    {
+      "epoch": 38.46223021582734,
+      "step": 42770,
+      "torque_loss": 0.13282687962055206
+    },
+    {
+      "epoch": 38.46223021582734,
+      "force_loss": 0.0016530733555555344,
+      "step": 42770
+    },
+    {
+      "epoch": 38.47122302158273,
+      "grad_norm": 0.2508116662502289,
+      "learning_rate": 2.0881155446011025e-05,
+      "loss": 0.0161,
+      "step": 42780
+    },
+    {
+      "action_loss": 0.002459365176036954,
+      "epoch": 38.47122302158273,
+      "step": 42780
+    },
+    {
+      "epoch": 38.47122302158273,
+      "step": 42780,
+      "torque_loss": 0.11887223273515701
+    },
+    {
+      "epoch": 38.47122302158273,
+      "force_loss": 0.010309889912605286,
+      "step": 42780
+    },
+    {
+      "epoch": 38.48021582733813,
+      "grad_norm": 0.14695261418819427,
+      "learning_rate": 2.0858757588505823e-05,
+      "loss": 0.0183,
+      "step": 42790
+    },
+    {
+      "action_loss": 0.003857190487906337,
+      "epoch": 38.48021582733813,
+      "step": 42790
+    },
+    {
+      "epoch": 38.48021582733813,
+      "step": 42790,
+      "torque_loss": 0.16643297672271729
+    },
+    {
+      "epoch": 38.48021582733813,
+      "force_loss": 0.005898170173168182,
+      "step": 42790
+    },
+    {
+      "epoch": 38.489208633093526,
+      "grad_norm": 0.09625194221735,
+      "learning_rate": 2.0836368583340622e-05,
+      "loss": 0.0144,
+      "step": 42800
+    },
+    {
+      "action_loss": 0.003638644702732563,
+      "epoch": 38.489208633093526,
+      "step": 42800
+    },
+    {
+      "epoch": 38.489208633093526,
+      "step": 42800,
+      "torque_loss": 0.12514102458953857
+    },
+    {
+      "epoch": 38.489208633093526,
+      "force_loss": 0.0020719580352306366,
+      "step": 42800
+    },
+    {
+      "epoch": 38.49820143884892,
+      "grad_norm": 0.15096129477024078,
+      "learning_rate": 2.081398843731664e-05,
+      "loss": 0.0155,
+      "step": 42810
+    },
+    {
+      "action_loss": 0.001300158561207354,
+      "epoch": 38.49820143884892,
+      "step": 42810
+    },
+    {
+      "epoch": 38.49820143884892,
+      "step": 42810,
+      "torque_loss": 0.11797508597373962
+    },
+    {
+      "epoch": 38.49820143884892,
+      "force_loss": 0.00276996660977602,
+      "step": 42810
+    },
+    {
+      "epoch": 38.507194244604314,
+      "grad_norm": 0.15623484551906586,
+      "learning_rate": 2.0791617157232357e-05,
+      "loss": 0.0153,
+      "step": 42820
+    },
+    {
+      "action_loss": 0.0023197324480861425,
+      "epoch": 38.507194244604314,
+      "step": 42820
+    },
+    {
+      "epoch": 38.507194244604314,
+      "step": 42820,
+      "torque_loss": 0.1320878118276596
+    },
+    {
+      "epoch": 38.507194244604314,
+      "force_loss": 0.005294546484947205,
+      "step": 42820
+    },
+    {
+      "epoch": 38.51618705035971,
+      "grad_norm": 0.23853468894958496,
+      "learning_rate": 2.0769254749883576e-05,
+      "loss": 0.0168,
+      "step": 42830
+    },
+    {
+      "action_loss": 0.0017935411306098104,
+      "epoch": 38.51618705035971,
+      "step": 42830
+    },
+    {
+      "epoch": 38.51618705035971,
+      "step": 42830,
+      "torque_loss": 0.14692510664463043
+    },
+    {
+      "epoch": 38.51618705035971,
+      "force_loss": 0.012311834841966629,
+      "step": 42830
+    },
+    {
+      "epoch": 38.52517985611511,
+      "grad_norm": 0.1728084683418274,
+      "learning_rate": 2.0746901222063415e-05,
+      "loss": 0.0157,
+      "step": 42840
+    },
+    {
+      "action_loss": 0.0020139499101787806,
+      "epoch": 38.52517985611511,
+      "step": 42840
+    },
+    {
+      "epoch": 38.52517985611511,
+      "step": 42840,
+      "torque_loss": 0.13630957901477814
+    },
+    {
+      "epoch": 38.52517985611511,
+      "force_loss": 0.005008505191653967,
+      "step": 42840
+    },
+    {
+      "epoch": 38.53417266187051,
+      "grad_norm": 0.08252532035112381,
+      "learning_rate": 2.072455658056226e-05,
+      "loss": 0.0162,
+      "step": 42850
+    },
+    {
+      "action_loss": 0.0016188727458938956,
+      "epoch": 38.53417266187051,
+      "step": 42850
+    },
+    {
+      "epoch": 38.53417266187051,
+      "step": 42850,
+      "torque_loss": 0.17332512140274048
+    },
+    {
+      "epoch": 38.53417266187051,
+      "force_loss": 0.0020309491083025932,
+      "step": 42850
+    },
+    {
+      "epoch": 38.5431654676259,
+      "grad_norm": 0.20099502801895142,
+      "learning_rate": 2.0702220832167873e-05,
+      "loss": 0.0177,
+      "step": 42860
+    },
+    {
+      "action_loss": 0.0013149403966963291,
+      "epoch": 38.5431654676259,
+      "step": 42860
+    },
+    {
+      "epoch": 38.5431654676259,
+      "step": 42860,
+      "torque_loss": 0.07513218373060226
+    },
+    {
+      "epoch": 38.5431654676259,
+      "force_loss": 0.0021678388584405184,
+      "step": 42860
+    },
+    {
+      "epoch": 38.552158273381295,
+      "grad_norm": 0.145574688911438,
+      "learning_rate": 2.0679893983665205e-05,
+      "loss": 0.0156,
+      "step": 42870
+    },
+    {
+      "action_loss": 0.0014957472449168563,
+      "epoch": 38.552158273381295,
+      "step": 42870
+    },
+    {
+      "epoch": 38.552158273381295,
+      "step": 42870,
+      "torque_loss": 0.1247599795460701
+    },
+    {
+      "epoch": 38.552158273381295,
+      "force_loss": 0.0020375109743326902,
+      "step": 42870
+    },
+    {
+      "epoch": 38.56115107913669,
+      "grad_norm": 0.2791518270969391,
+      "learning_rate": 2.0657576041836622e-05,
+      "loss": 0.0208,
+      "step": 42880
+    },
+    {
+      "action_loss": 0.006603449583053589,
+      "epoch": 38.56115107913669,
+      "step": 42880
+    },
+    {
+      "epoch": 38.56115107913669,
+      "step": 42880,
+      "torque_loss": 0.14387445151805878
+    },
+    {
+      "epoch": 38.56115107913669,
+      "force_loss": 0.0029947077855467796,
+      "step": 42880
+    },
+    {
+      "epoch": 38.57014388489208,
+      "grad_norm": 0.12982387840747833,
+      "learning_rate": 2.0635267013461666e-05,
+      "loss": 0.0217,
+      "step": 42890
+    },
+    {
+      "action_loss": 0.0017544155707582831,
+      "epoch": 38.57014388489208,
+      "step": 42890
+    },
+    {
+      "epoch": 38.57014388489208,
+      "step": 42890,
+      "torque_loss": 0.10584614425897598
+    },
+    {
+      "epoch": 38.57014388489208,
+      "force_loss": 0.0037342163268476725,
+      "step": 42890
+    },
+    {
+      "epoch": 38.57913669064748,
+      "grad_norm": 0.12017540633678436,
+      "learning_rate": 2.061296690531728e-05,
+      "loss": 0.0166,
+      "step": 42900
+    },
+    {
+      "action_loss": 0.0021294003818184137,
+      "epoch": 38.57913669064748,
+      "step": 42900
+    },
+    {
+      "epoch": 38.57913669064748,
+      "step": 42900,
+      "torque_loss": 0.09758982807397842
+    },
+    {
+      "epoch": 38.57913669064748,
+      "force_loss": 0.004263147246092558,
+      "step": 42900
+    },
+    {
+      "epoch": 38.58812949640288,
+      "grad_norm": 0.1537683755159378,
+      "learning_rate": 2.0590675724177622e-05,
+      "loss": 0.015,
+      "step": 42910
+    },
+    {
+      "action_loss": 0.00441939989104867,
+      "epoch": 38.58812949640288,
+      "step": 42910
+    },
+    {
+      "epoch": 38.58812949640288,
+      "step": 42910,
+      "torque_loss": 0.15077464282512665
+    },
+    {
+      "epoch": 38.58812949640288,
+      "force_loss": 0.006337263155728579,
+      "step": 42910
+    },
+    {
+      "epoch": 38.597122302158276,
+      "grad_norm": 0.12387017905712128,
+      "learning_rate": 2.0568393476814167e-05,
+      "loss": 0.0153,
+      "step": 42920
+    },
+    {
+      "action_loss": 0.005837309639900923,
+      "epoch": 38.597122302158276,
+      "step": 42920
+    },
+    {
+      "epoch": 38.597122302158276,
+      "step": 42920,
+      "torque_loss": 0.1131918802857399
+    },
+    {
+      "epoch": 38.597122302158276,
+      "force_loss": 0.005272162612527609,
+      "step": 42920
+    },
+    {
+      "epoch": 38.606115107913666,
+      "grad_norm": 0.33255335688591003,
+      "learning_rate": 2.0546120169995685e-05,
+      "loss": 0.0175,
+      "step": 42930
+    },
+    {
+      "action_loss": 0.0009794255020096898,
+      "epoch": 38.606115107913666,
+      "step": 42930
+    },
+    {
+      "epoch": 38.606115107913666,
+      "step": 42930,
+      "torque_loss": 0.16273508965969086
+    },
+    {
+      "epoch": 38.606115107913666,
+      "force_loss": 0.0017201714217662811,
+      "step": 42930
+    },
+    {
+      "epoch": 38.615107913669064,
+      "grad_norm": 0.24979008734226227,
+      "learning_rate": 2.0523855810488214e-05,
+      "loss": 0.0173,
+      "step": 42940
+    },
+    {
+      "action_loss": 0.0019623811822384596,
+      "epoch": 38.615107913669064,
+      "step": 42940
+    },
+    {
+      "epoch": 38.615107913669064,
+      "step": 42940,
+      "torque_loss": 0.08859298378229141
+    },
+    {
+      "epoch": 38.615107913669064,
+      "force_loss": 0.004715843591839075,
+      "step": 42940
+    },
+    {
+      "epoch": 38.62410071942446,
+      "grad_norm": 0.18740585446357727,
+      "learning_rate": 2.050160040505505e-05,
+      "loss": 0.016,
+      "step": 42950
+    },
+    {
+      "action_loss": 0.003015686059370637,
+      "epoch": 38.62410071942446,
+      "step": 42950
+    },
+    {
+      "epoch": 38.62410071942446,
+      "step": 42950,
+      "torque_loss": 0.11927451938390732
+    },
+    {
+      "epoch": 38.62410071942446,
+      "force_loss": 0.008128714747726917,
+      "step": 42950
+    },
+    {
+      "epoch": 38.63309352517986,
+      "grad_norm": 0.210119366645813,
+      "learning_rate": 2.0479353960456843e-05,
+      "loss": 0.0167,
+      "step": 42960
+    },
+    {
+      "action_loss": 0.004030993673950434,
+      "epoch": 38.63309352517986,
+      "step": 42960
+    },
+    {
+      "epoch": 38.63309352517986,
+      "step": 42960,
+      "torque_loss": 0.12204185873270035
+    },
+    {
+      "epoch": 38.63309352517986,
+      "force_loss": 0.004106556531041861,
+      "step": 42960
+    },
+    {
+      "epoch": 38.64208633093525,
+      "grad_norm": 0.15033361315727234,
+      "learning_rate": 2.0457116483451456e-05,
+      "loss": 0.0165,
+      "step": 42970
+    },
+    {
+      "action_loss": 0.0024442195426672697,
+      "epoch": 38.64208633093525,
+      "step": 42970
+    },
+    {
+      "epoch": 38.64208633093525,
+      "step": 42970,
+      "torque_loss": 0.09865373373031616
+    },
+    {
+      "epoch": 38.64208633093525,
+      "force_loss": 0.001946457545273006,
+      "step": 42970
+    },
+    {
+      "epoch": 38.65107913669065,
+      "grad_norm": 0.21015511453151703,
+      "learning_rate": 2.0434887980794043e-05,
+      "loss": 0.0155,
+      "step": 42980
+    },
+    {
+      "action_loss": 0.0006661896477453411,
+      "epoch": 38.65107913669065,
+      "step": 42980
+    },
+    {
+      "epoch": 38.65107913669065,
+      "step": 42980,
+      "torque_loss": 0.11808957904577255
+    },
+    {
+      "epoch": 38.65107913669065,
+      "force_loss": 0.0008906096336431801,
+      "step": 42980
+    },
+    {
+      "epoch": 38.660071942446045,
+      "grad_norm": 0.3225348889827728,
+      "learning_rate": 2.0412668459237043e-05,
+      "loss": 0.0189,
+      "step": 42990
+    },
+    {
+      "action_loss": 0.0025322488509118557,
+      "epoch": 38.660071942446045,
+      "step": 42990
+    },
+    {
+      "epoch": 38.660071942446045,
+      "step": 42990,
+      "torque_loss": 0.17096595466136932
+    },
+    {
+      "epoch": 38.660071942446045,
+      "force_loss": 0.012234543450176716,
+      "step": 42990
+    },
+    {
+      "epoch": 38.669064748201436,
+      "grad_norm": 0.1543445885181427,
+      "learning_rate": 2.039045792553016e-05,
+      "loss": 0.0153,
+      "step": 43000
+    },
+    {
+      "action_loss": 0.0023059502709656954,
+      "epoch": 38.669064748201436,
+      "step": 43000
+    },
+    {
+      "epoch": 38.669064748201436,
+      "step": 43000,
+      "torque_loss": 0.10220726579427719
+    },
+    {
+      "epoch": 38.669064748201436,
+      "force_loss": 0.0031755883246660233,
+      "step": 43000
+    },
+    {
+      "epoch": 38.67805755395683,
+      "grad_norm": 0.1288253217935562,
+      "learning_rate": 2.036825638642036e-05,
+      "loss": 0.0173,
+      "step": 43010
+    },
+    {
+      "action_loss": 0.004959840327501297,
+      "epoch": 38.67805755395683,
+      "step": 43010
+    },
+    {
+      "epoch": 38.67805755395683,
+      "step": 43010,
+      "torque_loss": 0.12836463749408722
+    },
+    {
+      "epoch": 38.67805755395683,
+      "force_loss": 0.009414522908627987,
+      "step": 43010
+    },
+    {
+      "epoch": 38.68705035971223,
+      "grad_norm": 0.32297179102897644,
+      "learning_rate": 2.0346063848651868e-05,
+      "loss": 0.0146,
+      "step": 43020
+    },
+    {
+      "action_loss": 0.0065900362096726894,
+      "epoch": 38.68705035971223,
+      "step": 43020
+    },
+    {
+      "epoch": 38.68705035971223,
+      "step": 43020,
+      "torque_loss": 0.13217408955097198
+    },
+    {
+      "epoch": 38.68705035971223,
+      "force_loss": 0.006156821269541979,
+      "step": 43020
+    },
+    {
+      "epoch": 38.69604316546763,
+      "grad_norm": 0.19051147997379303,
+      "learning_rate": 2.0323880318966254e-05,
+      "loss": 0.0172,
+      "step": 43030
+    },
+    {
+      "action_loss": 0.0032602306455373764,
+      "epoch": 38.69604316546763,
+      "step": 43030
+    },
+    {
+      "epoch": 38.69604316546763,
+      "step": 43030,
+      "torque_loss": 0.1282208412885666
+    },
+    {
+      "epoch": 38.69604316546763,
+      "force_loss": 0.0020867667626589537,
+      "step": 43030
+    },
+    {
+      "epoch": 38.70503597122302,
+      "grad_norm": 0.13722969591617584,
+      "learning_rate": 2.030170580410221e-05,
+      "loss": 0.0182,
+      "step": 43040
+    },
+    {
+      "action_loss": 0.0012555635767057538,
+      "epoch": 38.70503597122302,
+      "step": 43040
+    },
+    {
+      "epoch": 38.70503597122302,
+      "step": 43040,
+      "torque_loss": 0.12337436527013779
+    },
+    {
+      "epoch": 38.70503597122302,
+      "force_loss": 0.0032531581819057465,
+      "step": 43040
+    },
+    {
+      "epoch": 38.71402877697842,
+      "grad_norm": 0.25402286648750305,
+      "learning_rate": 2.0279540310795837e-05,
+      "loss": 0.0145,
+      "step": 43050
+    },
+    {
+      "action_loss": 0.004971138201653957,
+      "epoch": 38.71402877697842,
+      "step": 43050
+    },
+    {
+      "epoch": 38.71402877697842,
+      "step": 43050,
+      "torque_loss": 0.08613301068544388
+    },
+    {
+      "epoch": 38.71402877697842,
+      "force_loss": 0.006066437810659409,
+      "step": 43050
+    },
+    {
+      "epoch": 38.723021582733814,
+      "grad_norm": 0.07191108167171478,
+      "learning_rate": 2.0257383845780365e-05,
+      "loss": 0.0152,
+      "step": 43060
+    },
+    {
+      "action_loss": 0.007087436970323324,
+      "epoch": 38.723021582733814,
+      "step": 43060
+    },
+    {
+      "epoch": 38.723021582733814,
+      "step": 43060,
+      "torque_loss": 0.11101161688566208
+    },
+    {
+      "epoch": 38.723021582733814,
+      "force_loss": 0.017721662297844887,
+      "step": 43060
+    },
+    {
+      "epoch": 38.73201438848921,
+      "grad_norm": 0.1300228387117386,
+      "learning_rate": 2.0235236415786384e-05,
+      "loss": 0.0187,
+      "step": 43070
+    },
+    {
+      "action_loss": 0.001134268706664443,
+      "epoch": 38.73201438848921,
+      "step": 43070
+    },
+    {
+      "epoch": 38.73201438848921,
+      "step": 43070,
+      "torque_loss": 0.13750235736370087
+    },
+    {
+      "epoch": 38.73201438848921,
+      "force_loss": 0.0009366504964418709,
+      "step": 43070
+    },
+    {
+      "epoch": 38.7410071942446,
+      "grad_norm": 0.08991255611181259,
+      "learning_rate": 2.021309802754169e-05,
+      "loss": 0.0154,
+      "step": 43080
+    },
+    {
+      "action_loss": 0.01060511264950037,
+      "epoch": 38.7410071942446,
+      "step": 43080
+    },
+    {
+      "epoch": 38.7410071942446,
+      "step": 43080,
+      "torque_loss": 0.11946281790733337
+    },
+    {
+      "epoch": 38.7410071942446,
+      "force_loss": 0.010707574896514416,
+      "step": 43080
+    },
+    {
+      "epoch": 38.75,
+      "grad_norm": 0.08264023810625076,
+      "learning_rate": 2.0190968687771332e-05,
+      "loss": 0.0133,
+      "step": 43090
+    },
+    {
+      "action_loss": 0.0007571717142127454,
+      "epoch": 38.75,
+      "step": 43090
+    },
+    {
+      "epoch": 38.75,
+      "step": 43090,
+      "torque_loss": 0.11499225348234177
+    },
+    {
+      "epoch": 38.75,
+      "force_loss": 0.0019671835470944643,
+      "step": 43090
+    },
+    {
+      "epoch": 38.7589928057554,
+      "grad_norm": 0.2653161287307739,
+      "learning_rate": 2.016884840319763e-05,
+      "loss": 0.0147,
+      "step": 43100
+    },
+    {
+      "action_loss": 0.0028366928454488516,
+      "epoch": 38.7589928057554,
+      "step": 43100
+    },
+    {
+      "epoch": 38.7589928057554,
+      "step": 43100,
+      "torque_loss": 0.1616523563861847
+    },
+    {
+      "epoch": 38.7589928057554,
+      "force_loss": 0.00493226433172822,
+      "step": 43100
+    },
+    {
+      "epoch": 38.76798561151079,
+      "grad_norm": 0.3793729841709137,
+      "learning_rate": 2.0146737180540122e-05,
+      "loss": 0.0162,
+      "step": 43110
+    },
+    {
+      "action_loss": 0.0036913391668349504,
+      "epoch": 38.76798561151079,
+      "step": 43110
+    },
+    {
+      "epoch": 38.76798561151079,
+      "step": 43110,
+      "torque_loss": 0.154439315199852
+    },
+    {
+      "epoch": 38.76798561151079,
+      "force_loss": 0.007478731218725443,
+      "step": 43110
+    },
+    {
+      "epoch": 38.776978417266186,
+      "grad_norm": 0.1540844589471817,
+      "learning_rate": 2.012463502651564e-05,
+      "loss": 0.0142,
+      "step": 43120
+    },
+    {
+      "action_loss": 0.007521331310272217,
+      "epoch": 38.776978417266186,
+      "step": 43120
+    },
+    {
+      "epoch": 38.776978417266186,
+      "step": 43120,
+      "torque_loss": 0.12096823006868362
+    },
+    {
+      "epoch": 38.776978417266186,
+      "force_loss": 0.018969228491187096,
+      "step": 43120
+    },
+    {
+      "epoch": 38.78597122302158,
+      "grad_norm": 0.1410590261220932,
+      "learning_rate": 2.0102541947838228e-05,
+      "loss": 0.017,
+      "step": 43130
+    },
+    {
+      "action_loss": 0.000987159670330584,
+      "epoch": 38.78597122302158,
+      "step": 43130
+    },
+    {
+      "epoch": 38.78597122302158,
+      "step": 43130,
+      "torque_loss": 0.13239984214305878
+    },
+    {
+      "epoch": 38.78597122302158,
+      "force_loss": 0.0014093982754275203,
+      "step": 43130
+    },
+    {
+      "epoch": 38.79496402877698,
+      "grad_norm": 0.12232816219329834,
+      "learning_rate": 2.0080457951219173e-05,
+      "loss": 0.0147,
+      "step": 43140
+    },
+    {
+      "action_loss": 0.0008844216354191303,
+      "epoch": 38.79496402877698,
+      "step": 43140
+    },
+    {
+      "epoch": 38.79496402877698,
+      "step": 43140,
+      "torque_loss": 0.14423374831676483
+    },
+    {
+      "epoch": 38.79496402877698,
+      "force_loss": 0.0007429769611917436,
+      "step": 43140
+    },
+    {
+      "epoch": 38.80395683453237,
+      "grad_norm": 0.15267446637153625,
+      "learning_rate": 2.0058383043367017e-05,
+      "loss": 0.0199,
+      "step": 43150
+    },
+    {
+      "action_loss": 0.0012336460640653968,
+      "epoch": 38.80395683453237,
+      "step": 43150
+    },
+    {
+      "epoch": 38.80395683453237,
+      "step": 43150,
+      "torque_loss": 0.1527603715658188
+    },
+    {
+      "epoch": 38.80395683453237,
+      "force_loss": 0.002973536029458046,
+      "step": 43150
+    },
+    {
+      "epoch": 38.81294964028777,
+      "grad_norm": 0.1822865754365921,
+      "learning_rate": 2.0036317230987528e-05,
+      "loss": 0.0133,
+      "step": 43160
+    },
+    {
+      "action_loss": 0.0008694579009898007,
+      "epoch": 38.81294964028777,
+      "step": 43160
+    },
+    {
+      "epoch": 38.81294964028777,
+      "step": 43160,
+      "torque_loss": 0.11263538151979446
+    },
+    {
+      "epoch": 38.81294964028777,
+      "force_loss": 0.0010804197518154979,
+      "step": 43160
+    },
+    {
+      "epoch": 38.82194244604317,
+      "grad_norm": 0.1791848987340927,
+      "learning_rate": 2.0014260520783696e-05,
+      "loss": 0.0196,
+      "step": 43170
+    },
+    {
+      "action_loss": 0.004213029518723488,
+      "epoch": 38.82194244604317,
+      "step": 43170
+    },
+    {
+      "epoch": 38.82194244604317,
+      "step": 43170,
+      "torque_loss": 0.12668021023273468
+    },
+    {
+      "epoch": 38.82194244604317,
+      "force_loss": 0.008832802064716816,
+      "step": 43170
+    },
+    {
+      "epoch": 38.830935251798564,
+      "grad_norm": 0.5062022805213928,
+      "learning_rate": 1.9992212919455834e-05,
+      "loss": 0.0184,
+      "step": 43180
+    },
+    {
+      "action_loss": 0.011418166570365429,
+      "epoch": 38.830935251798564,
+      "step": 43180
+    },
+    {
+      "epoch": 38.830935251798564,
+      "step": 43180,
+      "torque_loss": 0.1398744434118271
+    },
+    {
+      "epoch": 38.830935251798564,
+      "force_loss": 0.02036232315003872,
+      "step": 43180
+    },
+    {
+      "epoch": 38.839928057553955,
+      "grad_norm": 0.27486929297447205,
+      "learning_rate": 1.9970174433701333e-05,
+      "loss": 0.017,
+      "step": 43190
+    },
+    {
+      "action_loss": 0.0007710832287557423,
+      "epoch": 38.839928057553955,
+      "step": 43190
+    },
+    {
+      "epoch": 38.839928057553955,
+      "step": 43190,
+      "torque_loss": 0.09834077209234238
+    },
+    {
+      "epoch": 38.839928057553955,
+      "force_loss": 0.0009948507649824023,
+      "step": 43190
+    },
+    {
+      "epoch": 38.84892086330935,
+      "grad_norm": 0.33841633796691895,
+      "learning_rate": 1.9948145070214992e-05,
+      "loss": 0.0143,
+      "step": 43200
+    },
+    {
+      "action_loss": 0.0026311948895454407,
+      "epoch": 38.84892086330935,
+      "step": 43200
+    },
+    {
+      "epoch": 38.84892086330935,
+      "step": 43200,
+      "torque_loss": 0.1537993997335434
+    },
+    {
+      "epoch": 38.84892086330935,
+      "force_loss": 0.002414965769276023,
+      "step": 43200
+    },
+    {
+      "epoch": 38.85791366906475,
+      "grad_norm": 0.089418426156044,
+      "learning_rate": 1.9926124835688663e-05,
+      "loss": 0.0181,
+      "step": 43210
+    },
+    {
+      "action_loss": 0.004569074604660273,
+      "epoch": 38.85791366906475,
+      "step": 43210
+    },
+    {
+      "epoch": 38.85791366906475,
+      "step": 43210,
+      "torque_loss": 0.1313643455505371
+    },
+    {
+      "epoch": 38.85791366906475,
+      "force_loss": 0.0030327008571475744,
+      "step": 43210
+    },
+    {
+      "epoch": 38.86690647482014,
+      "grad_norm": 0.36725449562072754,
+      "learning_rate": 1.9904113736811576e-05,
+      "loss": 0.0175,
+      "step": 43220
+    },
+    {
+      "action_loss": 0.01263834536075592,
+      "epoch": 38.86690647482014,
+      "step": 43220
+    },
+    {
+      "epoch": 38.86690647482014,
+      "step": 43220,
+      "torque_loss": 0.10542973130941391
+    },
+    {
+      "epoch": 38.86690647482014,
+      "force_loss": 0.006002683658152819,
+      "step": 43220
+    },
+    {
+      "epoch": 38.87589928057554,
+      "grad_norm": 0.23696266114711761,
+      "learning_rate": 1.9882111780270096e-05,
+      "loss": 0.0177,
+      "step": 43230
+    },
+    {
+      "action_loss": 0.000753029016777873,
+      "epoch": 38.87589928057554,
+      "step": 43230
+    },
+    {
+      "epoch": 38.87589928057554,
+      "step": 43230,
+      "torque_loss": 0.11790001392364502
+    },
+    {
+      "epoch": 38.87589928057554,
+      "force_loss": 0.0018389280885457993,
+      "step": 43230
+    },
+    {
+      "epoch": 38.884892086330936,
+      "grad_norm": 0.1683730036020279,
+      "learning_rate": 1.986011897274784e-05,
+      "loss": 0.0152,
+      "step": 43240
+    },
+    {
+      "action_loss": 0.0014560552081093192,
+      "epoch": 38.884892086330936,
+      "step": 43240
+    },
+    {
+      "epoch": 38.884892086330936,
+      "step": 43240,
+      "torque_loss": 0.09332650899887085
+    },
+    {
+      "epoch": 38.884892086330936,
+      "force_loss": 0.005439750384539366,
+      "step": 43240
+    },
+    {
+      "epoch": 38.893884892086334,
+      "grad_norm": 0.20827648043632507,
+      "learning_rate": 1.983813532092565e-05,
+      "loss": 0.021,
+      "step": 43250
+    },
+    {
+      "action_loss": 0.002569853560999036,
+      "epoch": 38.893884892086334,
+      "step": 43250
+    },
+    {
+      "epoch": 38.893884892086334,
+      "step": 43250,
+      "torque_loss": 0.11992853879928589
+    },
+    {
+      "epoch": 38.893884892086334,
+      "force_loss": 0.001781979575753212,
+      "step": 43250
+    },
+    {
+      "epoch": 38.902877697841724,
+      "grad_norm": 0.2119198888540268,
+      "learning_rate": 1.981616083148155e-05,
+      "loss": 0.0157,
+      "step": 43260
+    },
+    {
+      "action_loss": 0.0006633300217799842,
+      "epoch": 38.902877697841724,
+      "step": 43260
+    },
+    {
+      "epoch": 38.902877697841724,
+      "step": 43260,
+      "torque_loss": 0.12181542068719864
+    },
+    {
+      "epoch": 38.902877697841724,
+      "force_loss": 0.0007153118494898081,
+      "step": 43260
+    },
+    {
+      "epoch": 38.91187050359712,
+      "grad_norm": 0.1701694279909134,
+      "learning_rate": 1.9794195511090845e-05,
+      "loss": 0.0153,
+      "step": 43270
+    },
+    {
+      "action_loss": 0.0020223597530275583,
+      "epoch": 38.91187050359712,
+      "step": 43270
+    },
+    {
+      "epoch": 38.91187050359712,
+      "step": 43270,
+      "torque_loss": 0.09244181960821152
+    },
+    {
+      "epoch": 38.91187050359712,
+      "force_loss": 0.010265180841088295,
+      "step": 43270
+    },
+    {
+      "epoch": 38.92086330935252,
+      "grad_norm": 0.40680307149887085,
+      "learning_rate": 1.977223936642601e-05,
+      "loss": 0.0156,
+      "step": 43280
+    },
+    {
+      "action_loss": 0.0019339105347171426,
+      "epoch": 38.92086330935252,
+      "step": 43280
+    },
+    {
+      "epoch": 38.92086330935252,
+      "step": 43280,
+      "torque_loss": 0.13848191499710083
+    },
+    {
+      "epoch": 38.92086330935252,
+      "force_loss": 0.007765419315546751,
+      "step": 43280
+    },
+    {
+      "epoch": 38.92985611510792,
+      "grad_norm": 0.10724852234125137,
+      "learning_rate": 1.975029240415674e-05,
+      "loss": 0.0151,
+      "step": 43290
+    },
+    {
+      "action_loss": 0.004592752084136009,
+      "epoch": 38.92985611510792,
+      "step": 43290
+    },
+    {
+      "epoch": 38.92985611510792,
+      "step": 43290,
+      "torque_loss": 0.1376928687095642
+    },
+    {
+      "epoch": 38.92985611510792,
+      "force_loss": 0.006820193957537413,
+      "step": 43290
+    },
+    {
+      "epoch": 38.93884892086331,
+      "grad_norm": 0.09468401223421097,
+      "learning_rate": 1.9728354630949936e-05,
+      "loss": 0.0163,
+      "step": 43300
+    },
+    {
+      "action_loss": 0.002079724334180355,
+      "epoch": 38.93884892086331,
+      "step": 43300
+    },
+    {
+      "epoch": 38.93884892086331,
+      "step": 43300,
+      "torque_loss": 0.11005637794733047
+    },
+    {
+      "epoch": 38.93884892086331,
+      "force_loss": 0.0040276930667459965,
+      "step": 43300
+    },
+    {
+      "epoch": 38.947841726618705,
+      "grad_norm": 0.10339222103357315,
+      "learning_rate": 1.9706426053469716e-05,
+      "loss": 0.0151,
+      "step": 43310
+    },
+    {
+      "action_loss": 0.008520000614225864,
+      "epoch": 38.947841726618705,
+      "step": 43310
+    },
+    {
+      "epoch": 38.947841726618705,
+      "step": 43310,
+      "torque_loss": 0.11181195825338364
+    },
+    {
+      "epoch": 38.947841726618705,
+      "force_loss": 0.015329674817621708,
+      "step": 43310
+    },
+    {
+      "epoch": 38.9568345323741,
+      "grad_norm": 0.07079189270734787,
+      "learning_rate": 1.9684506678377396e-05,
+      "loss": 0.0158,
+      "step": 43320
+    },
+    {
+      "action_loss": 0.000937213480938226,
+      "epoch": 38.9568345323741,
+      "step": 43320
+    },
+    {
+      "epoch": 38.9568345323741,
+      "step": 43320,
+      "torque_loss": 0.08863457292318344
+    },
+    {
+      "epoch": 38.9568345323741,
+      "force_loss": 0.0016422374173998833,
+      "step": 43320
+    },
+    {
+      "epoch": 38.96582733812949,
+      "grad_norm": 0.5183575749397278,
+      "learning_rate": 1.9662596512331544e-05,
+      "loss": 0.0157,
+      "step": 43330
+    },
+    {
+      "action_loss": 0.0006735995411872864,
+      "epoch": 38.96582733812949,
+      "step": 43330
+    },
+    {
+      "epoch": 38.96582733812949,
+      "step": 43330,
+      "torque_loss": 0.10106196254491806
+    },
+    {
+      "epoch": 38.96582733812949,
+      "force_loss": 0.0006834732485003769,
+      "step": 43330
+    },
+    {
+      "epoch": 38.97482014388489,
+      "grad_norm": 0.19415691494941711,
+      "learning_rate": 1.964069556198782e-05,
+      "loss": 0.0182,
+      "step": 43340
+    },
+    {
+      "action_loss": 0.002119858283549547,
+      "epoch": 38.97482014388489,
+      "step": 43340
+    },
+    {
+      "epoch": 38.97482014388489,
+      "step": 43340,
+      "torque_loss": 0.14544929563999176
+    },
+    {
+      "epoch": 38.97482014388489,
+      "force_loss": 0.004749286454170942,
+      "step": 43340
+    },
+    {
+      "epoch": 38.98381294964029,
+      "grad_norm": 0.12878499925136566,
+      "learning_rate": 1.9618803833999232e-05,
+      "loss": 0.0165,
+      "step": 43350
+    },
+    {
+      "action_loss": 0.0027808435261249542,
+      "epoch": 38.98381294964029,
+      "step": 43350
+    },
+    {
+      "epoch": 38.98381294964029,
+      "step": 43350,
+      "torque_loss": 0.09459487348794937
+    },
+    {
+      "epoch": 38.98381294964029,
+      "force_loss": 0.0014990046620368958,
+      "step": 43350
+    },
+    {
+      "epoch": 38.992805755395686,
+      "grad_norm": 0.18176470696926117,
+      "learning_rate": 1.9596921335015838e-05,
+      "loss": 0.016,
+      "step": 43360
+    },
+    {
+      "action_loss": 0.0027199166361242533,
+      "epoch": 38.992805755395686,
+      "step": 43360
+    },
+    {
+      "epoch": 38.992805755395686,
+      "step": 43360,
+      "torque_loss": 0.1447519063949585
+    },
+    {
+      "epoch": 38.992805755395686,
+      "force_loss": 0.0030312202870845795,
+      "step": 43360
+    },
+    {
+      "epoch": 39.00179856115108,
+      "grad_norm": 0.07352231442928314,
+      "learning_rate": 1.957504807168501e-05,
+      "loss": 0.0167,
+      "step": 43370
+    },
+    {
+      "action_loss": 0.0007470918935723603,
+      "epoch": 39.00179856115108,
+      "step": 43370
+    },
+    {
+      "epoch": 39.00179856115108,
+      "step": 43370,
+      "torque_loss": 0.15074753761291504
+    },
+    {
+      "epoch": 39.00179856115108,
+      "force_loss": 0.0018721274100244045,
+      "step": 43370
+    },
+    {
+      "epoch": 39.010791366906474,
+      "grad_norm": 0.270116925239563,
+      "learning_rate": 1.9553184050651253e-05,
+      "loss": 0.0194,
+      "step": 43380
+    },
+    {
+      "action_loss": 0.001262398436665535,
+      "epoch": 39.010791366906474,
+      "step": 43380
+    },
+    {
+      "epoch": 39.010791366906474,
+      "step": 43380,
+      "torque_loss": 0.12718075513839722
+    },
+    {
+      "epoch": 39.010791366906474,
+      "force_loss": 0.0016740221763029695,
+      "step": 43380
+    },
+    {
+      "epoch": 39.01978417266187,
+      "grad_norm": 0.09478301554918289,
+      "learning_rate": 1.953132927855628e-05,
+      "loss": 0.0193,
+      "step": 43390
+    },
+    {
+      "action_loss": 0.0025347634218633175,
+      "epoch": 39.01978417266187,
+      "step": 43390
+    },
+    {
+      "epoch": 39.01978417266187,
+      "step": 43390,
+      "torque_loss": 0.07287508249282837
+    },
+    {
+      "epoch": 39.01978417266187,
+      "force_loss": 0.010764396749436855,
+      "step": 43390
+    },
+    {
+      "epoch": 39.02877697841727,
+      "grad_norm": 0.2662072777748108,
+      "learning_rate": 1.9509483762038995e-05,
+      "loss": 0.0168,
+      "step": 43400
+    },
+    {
+      "action_loss": 0.002846834482625127,
+      "epoch": 39.02877697841727,
+      "step": 43400
+    },
+    {
+      "epoch": 39.02877697841727,
+      "step": 43400,
+      "torque_loss": 0.11030423641204834
+    },
+    {
+      "epoch": 39.02877697841727,
+      "force_loss": 0.006497098598629236,
+      "step": 43400
+    },
+    {
+      "epoch": 39.03776978417266,
+      "grad_norm": 0.24567677080631256,
+      "learning_rate": 1.9487647507735467e-05,
+      "loss": 0.015,
+      "step": 43410
+    },
+    {
+      "action_loss": 0.004797985777258873,
+      "epoch": 39.03776978417266,
+      "step": 43410
+    },
+    {
+      "epoch": 39.03776978417266,
+      "step": 43410,
+      "torque_loss": 0.11644471436738968
+    },
+    {
+      "epoch": 39.03776978417266,
+      "force_loss": 0.01078809890896082,
+      "step": 43410
+    },
+    {
+      "epoch": 39.04676258992806,
+      "grad_norm": 0.15378287434577942,
+      "learning_rate": 1.9465820522279032e-05,
+      "loss": 0.0152,
+      "step": 43420
+    },
+    {
+      "action_loss": 0.004549337085336447,
+      "epoch": 39.04676258992806,
+      "step": 43420
+    },
+    {
+      "epoch": 39.04676258992806,
+      "step": 43420,
+      "torque_loss": 0.1682567596435547
+    },
+    {
+      "epoch": 39.04676258992806,
+      "force_loss": 0.014509479515254498,
+      "step": 43420
+    },
+    {
+      "epoch": 39.055755395683455,
+      "grad_norm": 0.29289907217025757,
+      "learning_rate": 1.9444002812300078e-05,
+      "loss": 0.0178,
+      "step": 43430
+    },
+    {
+      "action_loss": 0.0018622586503624916,
+      "epoch": 39.055755395683455,
+      "step": 43430
+    },
+    {
+      "epoch": 39.055755395683455,
+      "step": 43430,
+      "torque_loss": 0.08266764134168625
+    },
+    {
+      "epoch": 39.055755395683455,
+      "force_loss": 0.0013394304551184177,
+      "step": 43430
+    },
+    {
+      "epoch": 39.064748201438846,
+      "grad_norm": 0.34031981229782104,
+      "learning_rate": 1.94221943844263e-05,
+      "loss": 0.014,
+      "step": 43440
+    },
+    {
+      "action_loss": 0.0012294509215280414,
+      "epoch": 39.064748201438846,
+      "step": 43440
+    },
+    {
+      "epoch": 39.064748201438846,
+      "step": 43440,
+      "torque_loss": 0.12060848623514175
+    },
+    {
+      "epoch": 39.064748201438846,
+      "force_loss": 0.00680111488327384,
+      "step": 43440
+    },
+    {
+      "epoch": 39.07374100719424,
+      "grad_norm": 0.25598829984664917,
+      "learning_rate": 1.9400395245282515e-05,
+      "loss": 0.0163,
+      "step": 43450
+    },
+    {
+      "action_loss": 0.0032631580252200365,
+      "epoch": 39.07374100719424,
+      "step": 43450
+    },
+    {
+      "epoch": 39.07374100719424,
+      "step": 43450,
+      "torque_loss": 0.14373677968978882
+    },
+    {
+      "epoch": 39.07374100719424,
+      "force_loss": 0.00581319211050868,
+      "step": 43450
+    },
+    {
+      "epoch": 39.08273381294964,
+      "grad_norm": 0.3662920892238617,
+      "learning_rate": 1.937860540149071e-05,
+      "loss": 0.0178,
+      "step": 43460
+    },
+    {
+      "action_loss": 0.0008543722215108573,
+      "epoch": 39.08273381294964,
+      "step": 43460
+    },
+    {
+      "epoch": 39.08273381294964,
+      "step": 43460,
+      "torque_loss": 0.1530778557062149
+    },
+    {
+      "epoch": 39.08273381294964,
+      "force_loss": 0.0006587388925254345,
+      "step": 43460
+    },
+    {
+      "epoch": 39.09172661870504,
+      "grad_norm": 0.13331736624240875,
+      "learning_rate": 1.9356824859670082e-05,
+      "loss": 0.0162,
+      "step": 43470
+    },
+    {
+      "action_loss": 0.001381900510750711,
+      "epoch": 39.09172661870504,
+      "step": 43470
+    },
+    {
+      "epoch": 39.09172661870504,
+      "step": 43470,
+      "torque_loss": 0.13231825828552246
+    },
+    {
+      "epoch": 39.09172661870504,
+      "force_loss": 0.005654952023178339,
+      "step": 43470
+    },
+    {
+      "epoch": 39.10071942446043,
+      "grad_norm": 0.09411843866109848,
+      "learning_rate": 1.9335053626436967e-05,
+      "loss": 0.0141,
+      "step": 43480
+    },
+    {
+      "action_loss": 0.0007644129800610244,
+      "epoch": 39.10071942446043,
+      "step": 43480
+    },
+    {
+      "epoch": 39.10071942446043,
+      "step": 43480,
+      "torque_loss": 0.10469192266464233
+    },
+    {
+      "epoch": 39.10071942446043,
+      "force_loss": 0.0006866563926450908,
+      "step": 43480
+    },
+    {
+      "epoch": 39.10971223021583,
+      "grad_norm": 0.20438140630722046,
+      "learning_rate": 1.9313291708404885e-05,
+      "loss": 0.0171,
+      "step": 43490
+    },
+    {
+      "action_loss": 0.0017541605047881603,
+      "epoch": 39.10971223021583,
+      "step": 43490
+    },
+    {
+      "epoch": 39.10971223021583,
+      "step": 43490,
+      "torque_loss": 0.19136440753936768
+    },
+    {
+      "epoch": 39.10971223021583,
+      "force_loss": 0.0028786181937903166,
+      "step": 43490
+    },
+    {
+      "epoch": 39.118705035971225,
+      "grad_norm": 0.06751357018947601,
+      "learning_rate": 1.9291539112184587e-05,
+      "loss": 0.0173,
+      "step": 43500
+    },
+    {
+      "action_loss": 0.019240109249949455,
+      "epoch": 39.118705035971225,
+      "step": 43500
+    },
+    {
+      "epoch": 39.118705035971225,
+      "step": 43500,
+      "torque_loss": 0.13838006556034088
+    },
+    {
+      "epoch": 39.118705035971225,
+      "force_loss": 0.035848468542099,
+      "step": 43500
+    },
+    {
+      "epoch": 39.12769784172662,
+      "grad_norm": 0.14523009955883026,
+      "learning_rate": 1.9269795844383854e-05,
+      "loss": 0.0192,
+      "step": 43510
+    },
+    {
+      "action_loss": 0.0014279716415330768,
+      "epoch": 39.12769784172662,
+      "step": 43510
+    },
+    {
+      "epoch": 39.12769784172662,
+      "step": 43510,
+      "torque_loss": 0.13327476382255554
+    },
+    {
+      "epoch": 39.12769784172662,
+      "force_loss": 0.004329290706664324,
+      "step": 43510
+    },
+    {
+      "epoch": 39.13669064748201,
+      "grad_norm": 0.2870192527770996,
+      "learning_rate": 1.9248061911607777e-05,
+      "loss": 0.0153,
+      "step": 43520
+    },
+    {
+      "action_loss": 0.000583467714022845,
+      "epoch": 39.13669064748201,
+      "step": 43520
+    },
+    {
+      "epoch": 39.13669064748201,
+      "step": 43520,
+      "torque_loss": 0.08029218763113022
+    },
+    {
+      "epoch": 39.13669064748201,
+      "force_loss": 0.0007255204836837947,
+      "step": 43520
+    },
+    {
+      "epoch": 39.14568345323741,
+      "grad_norm": 0.29372504353523254,
+      "learning_rate": 1.9226337320458538e-05,
+      "loss": 0.0161,
+      "step": 43530
+    },
+    {
+      "action_loss": 0.0008068454917520285,
+      "epoch": 39.14568345323741,
+      "step": 43530
+    },
+    {
+      "epoch": 39.14568345323741,
+      "step": 43530,
+      "torque_loss": 0.07926338911056519
+    },
+    {
+      "epoch": 39.14568345323741,
+      "force_loss": 0.000790182501077652,
+      "step": 43530
+    },
+    {
+      "epoch": 39.15467625899281,
+      "grad_norm": 0.06133270636200905,
+      "learning_rate": 1.9204622077535488e-05,
+      "loss": 0.0128,
+      "step": 43540
+    },
+    {
+      "action_loss": 0.0011543224100023508,
+      "epoch": 39.15467625899281,
+      "step": 43540
+    },
+    {
+      "epoch": 39.15467625899281,
+      "step": 43540,
+      "torque_loss": 0.15765444934368134
+    },
+    {
+      "epoch": 39.15467625899281,
+      "force_loss": 0.0037975467275828123,
+      "step": 43540
+    },
+    {
+      "epoch": 39.1636690647482,
+      "grad_norm": 0.2015220820903778,
+      "learning_rate": 1.9182916189435147e-05,
+      "loss": 0.017,
+      "step": 43550
+    },
+    {
+      "action_loss": 0.00102571165189147,
+      "epoch": 39.1636690647482,
+      "step": 43550
+    },
+    {
+      "epoch": 39.1636690647482,
+      "step": 43550,
+      "torque_loss": 0.10125064104795456
+    },
+    {
+      "epoch": 39.1636690647482,
+      "force_loss": 0.0007454375736415386,
+      "step": 43550
+    },
+    {
+      "epoch": 39.172661870503596,
+      "grad_norm": 0.6480226516723633,
+      "learning_rate": 1.916121966275117e-05,
+      "loss": 0.0173,
+      "step": 43560
+    },
+    {
+      "action_loss": 0.0016835493734106421,
+      "epoch": 39.172661870503596,
+      "step": 43560
+    },
+    {
+      "epoch": 39.172661870503596,
+      "step": 43560,
+      "torque_loss": 0.12694333493709564
+    },
+    {
+      "epoch": 39.172661870503596,
+      "force_loss": 0.001822114922106266,
+      "step": 43560
+    },
+    {
+      "epoch": 39.181654676258994,
+      "grad_norm": 0.060405977070331573,
+      "learning_rate": 1.9139532504074443e-05,
+      "loss": 0.0144,
+      "step": 43570
+    },
+    {
+      "action_loss": 0.011442703194916248,
+      "epoch": 39.181654676258994,
+      "step": 43570
+    },
+    {
+      "epoch": 39.181654676258994,
+      "step": 43570,
+      "torque_loss": 0.10561565309762955
+    },
+    {
+      "epoch": 39.181654676258994,
+      "force_loss": 0.01838916353881359,
+      "step": 43570
+    },
+    {
+      "epoch": 39.19064748201439,
+      "grad_norm": 0.3246228098869324,
+      "learning_rate": 1.9117854719992885e-05,
+      "loss": 0.0156,
+      "step": 43580
+    },
+    {
+      "action_loss": 0.0021977215074002743,
+      "epoch": 39.19064748201439,
+      "step": 43580
+    },
+    {
+      "epoch": 39.19064748201439,
+      "step": 43580,
+      "torque_loss": 0.08217433840036392
+    },
+    {
+      "epoch": 39.19064748201439,
+      "force_loss": 0.0026522884145379066,
+      "step": 43580
+    },
+    {
+      "epoch": 39.19964028776978,
+      "grad_norm": 0.26682373881340027,
+      "learning_rate": 1.9096186317091687e-05,
+      "loss": 0.014,
+      "step": 43590
+    },
+    {
+      "action_loss": 0.0030627388041466475,
+      "epoch": 39.19964028776978,
+      "step": 43590
+    },
+    {
+      "epoch": 39.19964028776978,
+      "step": 43590,
+      "torque_loss": 0.11929595470428467
+    },
+    {
+      "epoch": 39.19964028776978,
+      "force_loss": 0.006410196889191866,
+      "step": 43590
+    },
+    {
+      "epoch": 39.20863309352518,
+      "grad_norm": 0.08988889306783676,
+      "learning_rate": 1.9074527301953116e-05,
+      "loss": 0.0164,
+      "step": 43600
+    },
+    {
+      "action_loss": 0.0009041757439263165,
+      "epoch": 39.20863309352518,
+      "step": 43600
+    },
+    {
+      "epoch": 39.20863309352518,
+      "step": 43600,
+      "torque_loss": 0.1162225529551506
+    },
+    {
+      "epoch": 39.20863309352518,
+      "force_loss": 0.0009028881904669106,
+      "step": 43600
+    },
+    {
+      "epoch": 39.21762589928058,
+      "grad_norm": 0.5964419841766357,
+      "learning_rate": 1.9052877681156607e-05,
+      "loss": 0.0165,
+      "step": 43610
+    },
+    {
+      "action_loss": 0.0005472433404065669,
+      "epoch": 39.21762589928058,
+      "step": 43610
+    },
+    {
+      "epoch": 39.21762589928058,
+      "step": 43610,
+      "torque_loss": 0.09671727567911148
+    },
+    {
+      "epoch": 39.21762589928058,
+      "force_loss": 0.0014344975352287292,
+      "step": 43610
+    },
+    {
+      "epoch": 39.226618705035975,
+      "grad_norm": 0.21963979303836823,
+      "learning_rate": 1.903123746127875e-05,
+      "loss": 0.021,
+      "step": 43620
+    },
+    {
+      "action_loss": 0.0010812049731612206,
+      "epoch": 39.226618705035975,
+      "step": 43620
+    },
+    {
+      "epoch": 39.226618705035975,
+      "step": 43620,
+      "torque_loss": 0.1577730029821396
+    },
+    {
+      "epoch": 39.226618705035975,
+      "force_loss": 0.0008934948709793389,
+      "step": 43620
+    },
+    {
+      "epoch": 39.235611510791365,
+      "grad_norm": 0.2075311690568924,
+      "learning_rate": 1.900960664889327e-05,
+      "loss": 0.0147,
+      "step": 43630
+    },
+    {
+      "action_loss": 0.0016371005913242698,
+      "epoch": 39.235611510791365,
+      "step": 43630
+    },
+    {
+      "epoch": 39.235611510791365,
+      "step": 43630,
+      "torque_loss": 0.09523005038499832
+    },
+    {
+      "epoch": 39.235611510791365,
+      "force_loss": 0.0019344780594110489,
+      "step": 43630
+    },
+    {
+      "epoch": 39.24460431654676,
+      "grad_norm": 0.06286769360303879,
+      "learning_rate": 1.8987985250571015e-05,
+      "loss": 0.0131,
+      "step": 43640
+    },
+    {
+      "action_loss": 0.0026696836575865746,
+      "epoch": 39.24460431654676,
+      "step": 43640
+    },
+    {
+      "epoch": 39.24460431654676,
+      "step": 43640,
+      "torque_loss": 0.09047457575798035
+    },
+    {
+      "epoch": 39.24460431654676,
+      "force_loss": 0.002093771705403924,
+      "step": 43640
+    },
+    {
+      "epoch": 39.25359712230216,
+      "grad_norm": 0.4066935181617737,
+      "learning_rate": 1.8966373272880054e-05,
+      "loss": 0.0155,
+      "step": 43650
+    },
+    {
+      "action_loss": 0.003066253149881959,
+      "epoch": 39.25359712230216,
+      "step": 43650
+    },
+    {
+      "epoch": 39.25359712230216,
+      "step": 43650,
+      "torque_loss": 0.13323789834976196
+    },
+    {
+      "epoch": 39.25359712230216,
+      "force_loss": 0.0063616386614739895,
+      "step": 43650
+    },
+    {
+      "epoch": 39.26258992805755,
+      "grad_norm": 0.4084271192550659,
+      "learning_rate": 1.8944770722385462e-05,
+      "loss": 0.0164,
+      "step": 43660
+    },
+    {
+      "action_loss": 0.001409812830388546,
+      "epoch": 39.26258992805755,
+      "step": 43660
+    },
+    {
+      "epoch": 39.26258992805755,
+      "step": 43660,
+      "torque_loss": 0.1265808790922165
+    },
+    {
+      "epoch": 39.26258992805755,
+      "force_loss": 0.0019094987073913217,
+      "step": 43660
+    },
+    {
+      "epoch": 39.27158273381295,
+      "grad_norm": 0.18222594261169434,
+      "learning_rate": 1.8923177605649576e-05,
+      "loss": 0.0145,
+      "step": 43670
+    },
+    {
+      "action_loss": 0.0012905653566122055,
+      "epoch": 39.27158273381295,
+      "step": 43670
+    },
+    {
+      "epoch": 39.27158273381295,
+      "step": 43670,
+      "torque_loss": 0.1176639273762703
+    },
+    {
+      "epoch": 39.27158273381295,
+      "force_loss": 0.0024132535327225924,
+      "step": 43670
+    },
+    {
+      "epoch": 39.280575539568346,
+      "grad_norm": 0.22960710525512695,
+      "learning_rate": 1.8901593929231802e-05,
+      "loss": 0.0157,
+      "step": 43680
+    },
+    {
+      "action_loss": 0.0017899327212944627,
+      "epoch": 39.280575539568346,
+      "step": 43680
+    },
+    {
+      "epoch": 39.280575539568346,
+      "step": 43680,
+      "torque_loss": 0.13487543165683746
+    },
+    {
+      "epoch": 39.280575539568346,
+      "force_loss": 0.004524525720626116,
+      "step": 43680
+    },
+    {
+      "epoch": 39.289568345323744,
+      "grad_norm": 0.11069312691688538,
+      "learning_rate": 1.8880019699688684e-05,
+      "loss": 0.016,
+      "step": 43690
+    },
+    {
+      "action_loss": 0.0025324011221528053,
+      "epoch": 39.289568345323744,
+      "step": 43690
+    },
+    {
+      "epoch": 39.289568345323744,
+      "step": 43690,
+      "torque_loss": 0.13795801997184753
+    },
+    {
+      "epoch": 39.289568345323744,
+      "force_loss": 0.0021935102995485067,
+      "step": 43690
+    },
+    {
+      "epoch": 39.298561151079134,
+      "grad_norm": 0.29894110560417175,
+      "learning_rate": 1.8858454923573904e-05,
+      "loss": 0.0179,
+      "step": 43700
+    },
+    {
+      "action_loss": 0.000977706047706306,
+      "epoch": 39.298561151079134,
+      "step": 43700
+    },
+    {
+      "epoch": 39.298561151079134,
+      "step": 43700,
+      "torque_loss": 0.10749268531799316
+    },
+    {
+      "epoch": 39.298561151079134,
+      "force_loss": 0.0025428796652704477,
+      "step": 43700
+    },
+    {
+      "epoch": 39.30755395683453,
+      "grad_norm": 0.1514466106891632,
+      "learning_rate": 1.8836899607438253e-05,
+      "loss": 0.0157,
+      "step": 43710
+    },
+    {
+      "action_loss": 0.003464254317805171,
+      "epoch": 39.30755395683453,
+      "step": 43710
+    },
+    {
+      "epoch": 39.30755395683453,
+      "step": 43710,
+      "torque_loss": 0.1474972516298294
+    },
+    {
+      "epoch": 39.30755395683453,
+      "force_loss": 0.007144351955503225,
+      "step": 43710
+    },
+    {
+      "epoch": 39.31654676258993,
+      "grad_norm": 0.13160359859466553,
+      "learning_rate": 1.8815353757829723e-05,
+      "loss": 0.0173,
+      "step": 43720
+    },
+    {
+      "action_loss": 0.002079211175441742,
+      "epoch": 39.31654676258993,
+      "step": 43720
+    },
+    {
+      "epoch": 39.31654676258993,
+      "step": 43720,
+      "torque_loss": 0.14407895505428314
+    },
+    {
+      "epoch": 39.31654676258993,
+      "force_loss": 0.006182938348501921,
+      "step": 43720
+    },
+    {
+      "epoch": 39.32553956834532,
+      "grad_norm": 0.2642349302768707,
+      "learning_rate": 1.879381738129331e-05,
+      "loss": 0.016,
+      "step": 43730
+    },
+    {
+      "action_loss": 0.0025451451074332,
+      "epoch": 39.32553956834532,
+      "step": 43730
+    },
+    {
+      "epoch": 39.32553956834532,
+      "step": 43730,
+      "torque_loss": 0.1272275298833847
+    },
+    {
+      "epoch": 39.32553956834532,
+      "force_loss": 0.01034451648592949,
+      "step": 43730
+    },
+    {
+      "epoch": 39.33453237410072,
+      "grad_norm": 0.2915138304233551,
+      "learning_rate": 1.8772290484371236e-05,
+      "loss": 0.0169,
+      "step": 43740
+    },
+    {
+      "action_loss": 0.001227371278218925,
+      "epoch": 39.33453237410072,
+      "step": 43740
+    },
+    {
+      "epoch": 39.33453237410072,
+      "step": 43740,
+      "torque_loss": 0.12451880425214767
+    },
+    {
+      "epoch": 39.33453237410072,
+      "force_loss": 0.0034549012780189514,
+      "step": 43740
+    },
+    {
+      "epoch": 39.343525179856115,
+      "grad_norm": 0.28453344106674194,
+      "learning_rate": 1.8750773073602795e-05,
+      "loss": 0.0155,
+      "step": 43750
+    },
+    {
+      "action_loss": 0.0029809344559907913,
+      "epoch": 39.343525179856115,
+      "step": 43750
+    },
+    {
+      "epoch": 39.343525179856115,
+      "step": 43750,
+      "torque_loss": 0.11854053288698196
+    },
+    {
+      "epoch": 39.343525179856115,
+      "force_loss": 0.0020232771057635546,
+      "step": 43750
+    },
+    {
+      "epoch": 39.35251798561151,
+      "grad_norm": 0.1431475281715393,
+      "learning_rate": 1.8729265155524405e-05,
+      "loss": 0.0169,
+      "step": 43760
+    },
+    {
+      "action_loss": 0.0012319684028625488,
+      "epoch": 39.35251798561151,
+      "step": 43760
+    },
+    {
+      "epoch": 39.35251798561151,
+      "step": 43760,
+      "torque_loss": 0.12439505010843277
+    },
+    {
+      "epoch": 39.35251798561151,
+      "force_loss": 0.0032792461570352316,
+      "step": 43760
+    },
+    {
+      "epoch": 39.361510791366904,
+      "grad_norm": 0.14642590284347534,
+      "learning_rate": 1.8707766736669607e-05,
+      "loss": 0.0154,
+      "step": 43770
+    },
+    {
+      "action_loss": 0.002598799765110016,
+      "epoch": 39.361510791366904,
+      "step": 43770
+    },
+    {
+      "epoch": 39.361510791366904,
+      "step": 43770,
+      "torque_loss": 0.12145161628723145
+    },
+    {
+      "epoch": 39.361510791366904,
+      "force_loss": 0.0018740269588306546,
+      "step": 43770
+    },
+    {
+      "epoch": 39.3705035971223,
+      "grad_norm": 0.08802445977926254,
+      "learning_rate": 1.8686277823569055e-05,
+      "loss": 0.0152,
+      "step": 43780
+    },
+    {
+      "action_loss": 0.0019907497335225344,
+      "epoch": 39.3705035971223,
+      "step": 43780
+    },
+    {
+      "epoch": 39.3705035971223,
+      "step": 43780,
+      "torque_loss": 0.14405936002731323
+    },
+    {
+      "epoch": 39.3705035971223,
+      "force_loss": 0.0013877536403015256,
+      "step": 43780
+    },
+    {
+      "epoch": 39.3794964028777,
+      "grad_norm": 0.12408871948719025,
+      "learning_rate": 1.8664798422750484e-05,
+      "loss": 0.0153,
+      "step": 43790
+    },
+    {
+      "action_loss": 0.02651992253959179,
+      "epoch": 39.3794964028777,
+      "step": 43790
+    },
+    {
+      "epoch": 39.3794964028777,
+      "step": 43790,
+      "torque_loss": 0.14238248765468597
+    },
+    {
+      "epoch": 39.3794964028777,
+      "force_loss": 0.02361910045146942,
+      "step": 43790
+    },
+    {
+      "epoch": 39.388489208633096,
+      "grad_norm": 0.2672511339187622,
+      "learning_rate": 1.8643328540738832e-05,
+      "loss": 0.0173,
+      "step": 43800
+    },
+    {
+      "action_loss": 0.0005486967856995761,
+      "epoch": 39.388489208633096,
+      "step": 43800
+    },
+    {
+      "epoch": 39.388489208633096,
+      "step": 43800,
+      "torque_loss": 0.10640814155340195
+    },
+    {
+      "epoch": 39.388489208633096,
+      "force_loss": 0.0005647577345371246,
+      "step": 43800
+    },
+    {
+      "epoch": 39.39748201438849,
+      "grad_norm": 0.3469555974006653,
+      "learning_rate": 1.862186818405601e-05,
+      "loss": 0.018,
+      "step": 43810
+    },
+    {
+      "action_loss": 0.0018105361377820373,
+      "epoch": 39.39748201438849,
+      "step": 43810
+    },
+    {
+      "epoch": 39.39748201438849,
+      "step": 43810,
+      "torque_loss": 0.11959423869848251
+    },
+    {
+      "epoch": 39.39748201438849,
+      "force_loss": 0.0017830407014116645,
+      "step": 43810
+    },
+    {
+      "epoch": 39.406474820143885,
+      "grad_norm": 0.09855153411626816,
+      "learning_rate": 1.8600417359221156e-05,
+      "loss": 0.0153,
+      "step": 43820
+    },
+    {
+      "action_loss": 0.0022541515063494444,
+      "epoch": 39.406474820143885,
+      "step": 43820
+    },
+    {
+      "epoch": 39.406474820143885,
+      "step": 43820,
+      "torque_loss": 0.0888887345790863
+    },
+    {
+      "epoch": 39.406474820143885,
+      "force_loss": 0.0025707988534122705,
+      "step": 43820
+    },
+    {
+      "epoch": 39.41546762589928,
+      "grad_norm": 0.11229920387268066,
+      "learning_rate": 1.8578976072750454e-05,
+      "loss": 0.0141,
+      "step": 43830
+    },
+    {
+      "action_loss": 0.0006761560216546059,
+      "epoch": 39.41546762589928,
+      "step": 43830
+    },
+    {
+      "epoch": 39.41546762589928,
+      "step": 43830,
+      "torque_loss": 0.10884235054254532
+    },
+    {
+      "epoch": 39.41546762589928,
+      "force_loss": 0.0007971436716616154,
+      "step": 43830
+    },
+    {
+      "epoch": 39.42446043165468,
+      "grad_norm": 0.13020965456962585,
+      "learning_rate": 1.8557544331157194e-05,
+      "loss": 0.0145,
+      "step": 43840
+    },
+    {
+      "action_loss": 0.0022146382834762335,
+      "epoch": 39.42446043165468,
+      "step": 43840
+    },
+    {
+      "epoch": 39.42446043165468,
+      "step": 43840,
+      "torque_loss": 0.10193448513746262
+    },
+    {
+      "epoch": 39.42446043165468,
+      "force_loss": 0.0014279544120654464,
+      "step": 43840
+    },
+    {
+      "epoch": 39.43345323741007,
+      "grad_norm": 0.33270177245140076,
+      "learning_rate": 1.8536122140951785e-05,
+      "loss": 0.0192,
+      "step": 43850
+    },
+    {
+      "action_loss": 0.009690295904874802,
+      "epoch": 39.43345323741007,
+      "step": 43850
+    },
+    {
+      "epoch": 39.43345323741007,
+      "step": 43850,
+      "torque_loss": 0.12405824661254883
+    },
+    {
+      "epoch": 39.43345323741007,
+      "force_loss": 0.012639988213777542,
+      "step": 43850
+    },
+    {
+      "epoch": 39.44244604316547,
+      "grad_norm": 0.243993878364563,
+      "learning_rate": 1.8514709508641688e-05,
+      "loss": 0.0186,
+      "step": 43860
+    },
+    {
+      "action_loss": 0.015644175931811333,
+      "epoch": 39.44244604316547,
+      "step": 43860
+    },
+    {
+      "epoch": 39.44244604316547,
+      "step": 43860,
+      "torque_loss": 0.16225554049015045
+    },
+    {
+      "epoch": 39.44244604316547,
+      "force_loss": 0.02187289111316204,
+      "step": 43860
+    },
+    {
+      "epoch": 39.451438848920866,
+      "grad_norm": 0.22017475962638855,
+      "learning_rate": 1.8493306440731555e-05,
+      "loss": 0.016,
+      "step": 43870
+    },
+    {
+      "action_loss": 0.00458850571885705,
+      "epoch": 39.451438848920866,
+      "step": 43870
+    },
+    {
+      "epoch": 39.451438848920866,
+      "step": 43870,
+      "torque_loss": 0.11322059482336044
+    },
+    {
+      "epoch": 39.451438848920866,
+      "force_loss": 0.0096434922888875,
+      "step": 43870
+    },
+    {
+      "epoch": 39.460431654676256,
+      "grad_norm": 0.4645187556743622,
+      "learning_rate": 1.8471912943723013e-05,
+      "loss": 0.0165,
+      "step": 43880
+    },
+    {
+      "action_loss": 0.0016340486472472548,
+      "epoch": 39.460431654676256,
+      "step": 43880
+    },
+    {
+      "epoch": 39.460431654676256,
+      "step": 43880,
+      "torque_loss": 0.1482481211423874
+    },
+    {
+      "epoch": 39.460431654676256,
+      "force_loss": 0.0021107306238263845,
+      "step": 43880
+    },
+    {
+      "epoch": 39.469424460431654,
+      "grad_norm": 0.17871703207492828,
+      "learning_rate": 1.8450529024114894e-05,
+      "loss": 0.0152,
+      "step": 43890
+    },
+    {
+      "action_loss": 0.009491377510130405,
+      "epoch": 39.469424460431654,
+      "step": 43890
+    },
+    {
+      "epoch": 39.469424460431654,
+      "step": 43890,
+      "torque_loss": 0.1356562227010727
+    },
+    {
+      "epoch": 39.469424460431654,
+      "force_loss": 0.01624637469649315,
+      "step": 43890
+    },
+    {
+      "epoch": 39.47841726618705,
+      "grad_norm": 0.12953442335128784,
+      "learning_rate": 1.842915468840301e-05,
+      "loss": 0.0153,
+      "step": 43900
+    },
+    {
+      "action_loss": 0.009950876235961914,
+      "epoch": 39.47841726618705,
+      "step": 43900
+    },
+    {
+      "epoch": 39.47841726618705,
+      "step": 43900,
+      "torque_loss": 0.14285694062709808
+    },
+    {
+      "epoch": 39.47841726618705,
+      "force_loss": 0.00954663660377264,
+      "step": 43900
+    },
+    {
+      "epoch": 39.48741007194245,
+      "grad_norm": 0.24169917404651642,
+      "learning_rate": 1.840778994308037e-05,
+      "loss": 0.0167,
+      "step": 43910
+    },
+    {
+      "action_loss": 0.0016071697464212775,
+      "epoch": 39.48741007194245,
+      "step": 43910
+    },
+    {
+      "epoch": 39.48741007194245,
+      "step": 43910,
+      "torque_loss": 0.1528947800397873
+    },
+    {
+      "epoch": 39.48741007194245,
+      "force_loss": 0.001843883073888719,
+      "step": 43910
+    },
+    {
+      "epoch": 39.49640287769784,
+      "grad_norm": 0.09055488556623459,
+      "learning_rate": 1.8386434794637004e-05,
+      "loss": 0.0132,
+      "step": 43920
+    },
+    {
+      "action_loss": 0.0011045786086469889,
+      "epoch": 39.49640287769784,
+      "step": 43920
+    },
+    {
+      "epoch": 39.49640287769784,
+      "step": 43920,
+      "torque_loss": 0.07543081790208817
+    },
+    {
+      "epoch": 39.49640287769784,
+      "force_loss": 0.002623551758006215,
+      "step": 43920
+    },
+    {
+      "epoch": 39.50539568345324,
+      "grad_norm": 0.17763882875442505,
+      "learning_rate": 1.8365089249560034e-05,
+      "loss": 0.0135,
+      "step": 43930
+    },
+    {
+      "action_loss": 0.0021788382437080145,
+      "epoch": 39.50539568345324,
+      "step": 43930
+    },
+    {
+      "epoch": 39.50539568345324,
+      "step": 43930,
+      "torque_loss": 0.15388917922973633
+    },
+    {
+      "epoch": 39.50539568345324,
+      "force_loss": 0.011935091577470303,
+      "step": 43930
+    },
+    {
+      "epoch": 39.514388489208635,
+      "grad_norm": 0.3831996023654938,
+      "learning_rate": 1.8343753314333683e-05,
+      "loss": 0.0153,
+      "step": 43940
+    },
+    {
+      "action_loss": 0.0022421476896852255,
+      "epoch": 39.514388489208635,
+      "step": 43940
+    },
+    {
+      "epoch": 39.514388489208635,
+      "step": 43940,
+      "torque_loss": 0.1466163545846939
+    },
+    {
+      "epoch": 39.514388489208635,
+      "force_loss": 0.004129001870751381,
+      "step": 43940
+    },
+    {
+      "epoch": 39.523381294964025,
+      "grad_norm": 0.42577558755874634,
+      "learning_rate": 1.8322426995439236e-05,
+      "loss": 0.0167,
+      "step": 43950
+    },
+    {
+      "action_loss": 0.005300629884004593,
+      "epoch": 39.523381294964025,
+      "step": 43950
+    },
+    {
+      "epoch": 39.523381294964025,
+      "step": 43950,
+      "torque_loss": 0.11630934476852417
+    },
+    {
+      "epoch": 39.523381294964025,
+      "force_loss": 0.010095425881445408,
+      "step": 43950
+    },
+    {
+      "epoch": 39.53237410071942,
+      "grad_norm": 0.14609003067016602,
+      "learning_rate": 1.8301110299355058e-05,
+      "loss": 0.0242,
+      "step": 43960
+    },
+    {
+      "action_loss": 0.001404804759658873,
+      "epoch": 39.53237410071942,
+      "step": 43960
+    },
+    {
+      "epoch": 39.53237410071942,
+      "step": 43960,
+      "torque_loss": 0.10890919715166092
+    },
+    {
+      "epoch": 39.53237410071942,
+      "force_loss": 0.0014651529490947723,
+      "step": 43960
+    },
+    {
+      "epoch": 39.54136690647482,
+      "grad_norm": 0.11471282690763474,
+      "learning_rate": 1.8279803232556625e-05,
+      "loss": 0.0164,
+      "step": 43970
+    },
+    {
+      "action_loss": 0.002044142922386527,
+      "epoch": 39.54136690647482,
+      "step": 43970
+    },
+    {
+      "epoch": 39.54136690647482,
+      "step": 43970,
+      "torque_loss": 0.18028700351715088
+    },
+    {
+      "epoch": 39.54136690647482,
+      "force_loss": 0.002922143554314971,
+      "step": 43970
+    },
+    {
+      "epoch": 39.55035971223022,
+      "grad_norm": 0.11105646193027496,
+      "learning_rate": 1.8258505801516444e-05,
+      "loss": 0.015,
+      "step": 43980
+    },
+    {
+      "action_loss": 0.0015438087284564972,
+      "epoch": 39.55035971223022,
+      "step": 43980
+    },
+    {
+      "epoch": 39.55035971223022,
+      "step": 43980,
+      "torque_loss": 0.13766025006771088
+    },
+    {
+      "epoch": 39.55035971223022,
+      "force_loss": 0.007871488109230995,
+      "step": 43980
+    },
+    {
+      "epoch": 39.55935251798561,
+      "grad_norm": 0.2366029918193817,
+      "learning_rate": 1.8237218012704117e-05,
+      "loss": 0.0162,
+      "step": 43990
+    },
+    {
+      "action_loss": 0.009585704654455185,
+      "epoch": 39.55935251798561,
+      "step": 43990
+    },
+    {
+      "epoch": 39.55935251798561,
+      "step": 43990,
+      "torque_loss": 0.09305298328399658
+    },
+    {
+      "epoch": 39.55935251798561,
+      "force_loss": 0.011574490927159786,
+      "step": 43990
+    },
+    {
+      "epoch": 39.568345323741006,
+      "grad_norm": 0.10388772934675217,
+      "learning_rate": 1.821593987258631e-05,
+      "loss": 0.0158,
+      "step": 44000
+    },
+    {
+      "action_loss": 0.0023453370667994022,
+      "epoch": 39.568345323741006,
+      "step": 44000
+    },
+    {
+      "epoch": 39.568345323741006,
+      "step": 44000,
+      "torque_loss": 0.09254888445138931
+    },
+    {
+      "epoch": 39.568345323741006,
+      "force_loss": 0.00380254746414721,
+      "step": 44000
+    },
+    {
+      "epoch": 39.577338129496404,
+      "grad_norm": 0.11751534789800644,
+      "learning_rate": 1.8194671387626744e-05,
+      "loss": 0.0158,
+      "step": 44010
+    },
+    {
+      "action_loss": 0.002771947532892227,
+      "epoch": 39.577338129496404,
+      "step": 44010
+    },
+    {
+      "epoch": 39.577338129496404,
+      "step": 44010,
+      "torque_loss": 0.10505024343729019
+    },
+    {
+      "epoch": 39.577338129496404,
+      "force_loss": 0.003331434680148959,
+      "step": 44010
+    },
+    {
+      "epoch": 39.5863309352518,
+      "grad_norm": 0.09511953592300415,
+      "learning_rate": 1.8173412564286276e-05,
+      "loss": 0.0167,
+      "step": 44020
+    },
+    {
+      "action_loss": 0.002071511233225465,
+      "epoch": 39.5863309352518,
+      "step": 44020
+    },
+    {
+      "epoch": 39.5863309352518,
+      "step": 44020,
+      "torque_loss": 0.12547074258327484
+    },
+    {
+      "epoch": 39.5863309352518,
+      "force_loss": 0.0028882029000669718,
+      "step": 44020
+    },
+    {
+      "epoch": 39.59532374100719,
+      "grad_norm": 0.25395458936691284,
+      "learning_rate": 1.8152163409022697e-05,
+      "loss": 0.0163,
+      "step": 44030
+    },
+    {
+      "action_loss": 0.0027981549501419067,
+      "epoch": 39.59532374100719,
+      "step": 44030
+    },
+    {
+      "epoch": 39.59532374100719,
+      "step": 44030,
+      "torque_loss": 0.11936943978071213
+    },
+    {
+      "epoch": 39.59532374100719,
+      "force_loss": 0.002433473477140069,
+      "step": 44030
+    },
+    {
+      "epoch": 39.60431654676259,
+      "grad_norm": 0.12484948337078094,
+      "learning_rate": 1.8130923928291023e-05,
+      "loss": 0.0152,
+      "step": 44040
+    },
+    {
+      "action_loss": 0.004602337721735239,
+      "epoch": 39.60431654676259,
+      "step": 44040
+    },
+    {
+      "epoch": 39.60431654676259,
+      "step": 44040,
+      "torque_loss": 0.1243242621421814
+    },
+    {
+      "epoch": 39.60431654676259,
+      "force_loss": 0.004839560482650995,
+      "step": 44040
+    },
+    {
+      "epoch": 39.61330935251799,
+      "grad_norm": 0.31259891390800476,
+      "learning_rate": 1.8109694128543163e-05,
+      "loss": 0.0165,
+      "step": 44050
+    },
+    {
+      "action_loss": 0.0018528815126046538,
+      "epoch": 39.61330935251799,
+      "step": 44050
+    },
+    {
+      "epoch": 39.61330935251799,
+      "step": 44050,
+      "torque_loss": 0.0797339603304863
+    },
+    {
+      "epoch": 39.61330935251799,
+      "force_loss": 0.0032678258139640093,
+      "step": 44050
+    },
+    {
+      "epoch": 39.62230215827338,
+      "grad_norm": 0.08590597659349442,
+      "learning_rate": 1.8088474016228237e-05,
+      "loss": 0.0146,
+      "step": 44060
+    },
+    {
+      "action_loss": 0.0007265713065862656,
+      "epoch": 39.62230215827338,
+      "step": 44060
+    },
+    {
+      "epoch": 39.62230215827338,
+      "step": 44060,
+      "torque_loss": 0.11183180660009384
+    },
+    {
+      "epoch": 39.62230215827338,
+      "force_loss": 0.0012429512571543455,
+      "step": 44060
+    },
+    {
+      "epoch": 39.631294964028775,
+      "grad_norm": 0.340399831533432,
+      "learning_rate": 1.8067263597792328e-05,
+      "loss": 0.0156,
+      "step": 44070
+    },
+    {
+      "action_loss": 0.0024141687899827957,
+      "epoch": 39.631294964028775,
+      "step": 44070
+    },
+    {
+      "epoch": 39.631294964028775,
+      "step": 44070,
+      "torque_loss": 0.14910492300987244
+    },
+    {
+      "epoch": 39.631294964028775,
+      "force_loss": 0.015826279297471046,
+      "step": 44070
+    },
+    {
+      "epoch": 39.64028776978417,
+      "grad_norm": 0.14503931999206543,
+      "learning_rate": 1.80460628796786e-05,
+      "loss": 0.0172,
+      "step": 44080
+    },
+    {
+      "action_loss": 0.0008308072574436665,
+      "epoch": 39.64028776978417,
+      "step": 44080
+    },
+    {
+      "epoch": 39.64028776978417,
+      "step": 44080,
+      "torque_loss": 0.12460996955633163
+    },
+    {
+      "epoch": 39.64028776978417,
+      "force_loss": 0.0015784470597282052,
+      "step": 44080
+    },
+    {
+      "epoch": 39.64928057553957,
+      "grad_norm": 0.2599188983440399,
+      "learning_rate": 1.8024871868327276e-05,
+      "loss": 0.015,
+      "step": 44090
+    },
+    {
+      "action_loss": 0.0031471364200115204,
+      "epoch": 39.64928057553957,
+      "step": 44090
+    },
+    {
+      "epoch": 39.64928057553957,
+      "step": 44090,
+      "torque_loss": 0.10941216349601746
+    },
+    {
+      "epoch": 39.64928057553957,
+      "force_loss": 0.005341757088899612,
+      "step": 44090
+    },
+    {
+      "epoch": 39.65827338129496,
+      "grad_norm": 0.20297805964946747,
+      "learning_rate": 1.8003690570175608e-05,
+      "loss": 0.0165,
+      "step": 44100
+    },
+    {
+      "action_loss": 0.0034830167423933744,
+      "epoch": 39.65827338129496,
+      "step": 44100
+    },
+    {
+      "epoch": 39.65827338129496,
+      "step": 44100,
+      "torque_loss": 0.15224498510360718
+    },
+    {
+      "epoch": 39.65827338129496,
+      "force_loss": 0.0027494810055941343,
+      "step": 44100
+    },
+    {
+      "epoch": 39.66726618705036,
+      "grad_norm": 0.10213612765073776,
+      "learning_rate": 1.7982518991657943e-05,
+      "loss": 0.0165,
+      "step": 44110
+    },
+    {
+      "action_loss": 0.0008230170351453125,
+      "epoch": 39.66726618705036,
+      "step": 44110
+    },
+    {
+      "epoch": 39.66726618705036,
+      "step": 44110,
+      "torque_loss": 0.1021357998251915
+    },
+    {
+      "epoch": 39.66726618705036,
+      "force_loss": 0.003565584309399128,
+      "step": 44110
+    },
+    {
+      "epoch": 39.67625899280576,
+      "grad_norm": 0.11593269556760788,
+      "learning_rate": 1.7961357139205643e-05,
+      "loss": 0.0123,
+      "step": 44120
+    },
+    {
+      "action_loss": 0.0008481573313474655,
+      "epoch": 39.67625899280576,
+      "step": 44120
+    },
+    {
+      "epoch": 39.67625899280576,
+      "step": 44120,
+      "torque_loss": 0.12810324132442474
+    },
+    {
+      "epoch": 39.67625899280576,
+      "force_loss": 0.0015052525559440255,
+      "step": 44120
+    },
+    {
+      "epoch": 39.685251798561154,
+      "grad_norm": 0.1353565901517868,
+      "learning_rate": 1.7940205019247108e-05,
+      "loss": 0.0135,
+      "step": 44130
+    },
+    {
+      "action_loss": 0.0014148285845294595,
+      "epoch": 39.685251798561154,
+      "step": 44130
+    },
+    {
+      "epoch": 39.685251798561154,
+      "step": 44130,
+      "torque_loss": 0.13040098547935486
+    },
+    {
+      "epoch": 39.685251798561154,
+      "force_loss": 0.0022107947152107954,
+      "step": 44130
+    },
+    {
+      "epoch": 39.694244604316545,
+      "grad_norm": 0.28801798820495605,
+      "learning_rate": 1.79190626382078e-05,
+      "loss": 0.0164,
+      "step": 44140
+    },
+    {
+      "action_loss": 0.000787557743024081,
+      "epoch": 39.694244604316545,
+      "step": 44140
+    },
+    {
+      "epoch": 39.694244604316545,
+      "step": 44140,
+      "torque_loss": 0.14245951175689697
+    },
+    {
+      "epoch": 39.694244604316545,
+      "force_loss": 0.0037407006602734327,
+      "step": 44140
+    },
+    {
+      "epoch": 39.70323741007194,
+      "grad_norm": 0.11380057781934738,
+      "learning_rate": 1.7897930002510215e-05,
+      "loss": 0.0174,
+      "step": 44150
+    },
+    {
+      "action_loss": 0.0007630392792634666,
+      "epoch": 39.70323741007194,
+      "step": 44150
+    },
+    {
+      "epoch": 39.70323741007194,
+      "step": 44150,
+      "torque_loss": 0.11410563439130783
+    },
+    {
+      "epoch": 39.70323741007194,
+      "force_loss": 0.0006533423438668251,
+      "step": 44150
+    },
+    {
+      "epoch": 39.71223021582734,
+      "grad_norm": 0.20559054613113403,
+      "learning_rate": 1.787680711857387e-05,
+      "loss": 0.0149,
+      "step": 44160
+    },
+    {
+      "action_loss": 0.00552561366930604,
+      "epoch": 39.71223021582734,
+      "step": 44160
+    },
+    {
+      "epoch": 39.71223021582734,
+      "step": 44160,
+      "torque_loss": 0.14884339272975922
+    },
+    {
+      "epoch": 39.71223021582734,
+      "force_loss": 0.008594215847551823,
+      "step": 44160
+    },
+    {
+      "epoch": 39.72122302158273,
+      "grad_norm": 0.12301896512508392,
+      "learning_rate": 1.7855693992815398e-05,
+      "loss": 0.0165,
+      "step": 44170
+    },
+    {
+      "action_loss": 0.001032329979352653,
+      "epoch": 39.72122302158273,
+      "step": 44170
+    },
+    {
+      "epoch": 39.72122302158273,
+      "step": 44170,
+      "torque_loss": 0.1236443892121315
+    },
+    {
+      "epoch": 39.72122302158273,
+      "force_loss": 0.0016958294436335564,
+      "step": 44170
+    },
+    {
+      "epoch": 39.73021582733813,
+      "grad_norm": 0.24747337400913239,
+      "learning_rate": 1.7834590631648328e-05,
+      "loss": 0.0154,
+      "step": 44180
+    },
+    {
+      "action_loss": 0.001045675715431571,
+      "epoch": 39.73021582733813,
+      "step": 44180
+    },
+    {
+      "epoch": 39.73021582733813,
+      "step": 44180,
+      "torque_loss": 0.11847928911447525
+    },
+    {
+      "epoch": 39.73021582733813,
+      "force_loss": 0.0025990523863583803,
+      "step": 44180
+    },
+    {
+      "epoch": 39.739208633093526,
+      "grad_norm": 0.19082608819007874,
+      "learning_rate": 1.7813497041483384e-05,
+      "loss": 0.017,
+      "step": 44190
+    },
+    {
+      "action_loss": 0.004814065992832184,
+      "epoch": 39.739208633093526,
+      "step": 44190
+    },
+    {
+      "epoch": 39.739208633093526,
+      "step": 44190,
+      "torque_loss": 0.13301880657672882
+    },
+    {
+      "epoch": 39.739208633093526,
+      "force_loss": 0.008390623144805431,
+      "step": 44190
+    },
+    {
+      "epoch": 39.74820143884892,
+      "grad_norm": 0.2500182092189789,
+      "learning_rate": 1.779241322872817e-05,
+      "loss": 0.0164,
+      "step": 44200
+    },
+    {
+      "action_loss": 0.005697909742593765,
+      "epoch": 39.74820143884892,
+      "step": 44200
+    },
+    {
+      "epoch": 39.74820143884892,
+      "step": 44200,
+      "torque_loss": 0.1487625241279602
+    },
+    {
+      "epoch": 39.74820143884892,
+      "force_loss": 0.013683046214282513,
+      "step": 44200
+    },
+    {
+      "epoch": 39.757194244604314,
+      "grad_norm": 0.18140637874603271,
+      "learning_rate": 1.777133919978744e-05,
+      "loss": 0.0155,
+      "step": 44210
+    },
+    {
+      "action_loss": 0.008365617133677006,
+      "epoch": 39.757194244604314,
+      "step": 44210
+    },
+    {
+      "epoch": 39.757194244604314,
+      "step": 44210,
+      "torque_loss": 0.19642280042171478
+    },
+    {
+      "epoch": 39.757194244604314,
+      "force_loss": 0.01971357874572277,
+      "step": 44210
+    },
+    {
+      "epoch": 39.76618705035971,
+      "grad_norm": 0.09545369446277618,
+      "learning_rate": 1.7750274961062912e-05,
+      "loss": 0.0201,
+      "step": 44220
+    },
+    {
+      "action_loss": 0.003569573163986206,
+      "epoch": 39.76618705035971,
+      "step": 44220
+    },
+    {
+      "epoch": 39.76618705035971,
+      "step": 44220,
+      "torque_loss": 0.11544469743967056
+    },
+    {
+      "epoch": 39.76618705035971,
+      "force_loss": 0.01004467811435461,
+      "step": 44220
+    },
+    {
+      "epoch": 39.77517985611511,
+      "grad_norm": 0.1329287886619568,
+      "learning_rate": 1.772922051895335e-05,
+      "loss": 0.0164,
+      "step": 44230
+    },
+    {
+      "action_loss": 0.002476147608831525,
+      "epoch": 39.77517985611511,
+      "step": 44230
+    },
+    {
+      "epoch": 39.77517985611511,
+      "step": 44230,
+      "torque_loss": 0.08876601606607437
+    },
+    {
+      "epoch": 39.77517985611511,
+      "force_loss": 0.008319896645843983,
+      "step": 44230
+    },
+    {
+      "epoch": 39.78417266187051,
+      "grad_norm": 0.11444862186908722,
+      "learning_rate": 1.770817587985453e-05,
+      "loss": 0.0153,
+      "step": 44240
+    },
+    {
+      "action_loss": 0.0015230766730383039,
+      "epoch": 39.78417266187051,
+      "step": 44240
+    },
+    {
+      "epoch": 39.78417266187051,
+      "step": 44240,
+      "torque_loss": 0.10689255595207214
+    },
+    {
+      "epoch": 39.78417266187051,
+      "force_loss": 0.0018003984587267041,
+      "step": 44240
+    },
+    {
+      "epoch": 39.7931654676259,
+      "grad_norm": 0.10953350365161896,
+      "learning_rate": 1.7687141050159246e-05,
+      "loss": 0.015,
+      "step": 44250
+    },
+    {
+      "action_loss": 0.009383227676153183,
+      "epoch": 39.7931654676259,
+      "step": 44250
+    },
+    {
+      "epoch": 39.7931654676259,
+      "step": 44250,
+      "torque_loss": 0.18183942139148712
+    },
+    {
+      "epoch": 39.7931654676259,
+      "force_loss": 0.00998720433562994,
+      "step": 44250
+    },
+    {
+      "epoch": 39.802158273381295,
+      "grad_norm": 0.1178608164191246,
+      "learning_rate": 1.7666116036257375e-05,
+      "loss": 0.0156,
+      "step": 44260
+    },
+    {
+      "action_loss": 0.0023383963853120804,
+      "epoch": 39.802158273381295,
+      "step": 44260
+    },
+    {
+      "epoch": 39.802158273381295,
+      "step": 44260,
+      "torque_loss": 0.14484000205993652
+    },
+    {
+      "epoch": 39.802158273381295,
+      "force_loss": 0.007460723165422678,
+      "step": 44260
+    },
+    {
+      "epoch": 39.81115107913669,
+      "grad_norm": 0.06822218000888824,
+      "learning_rate": 1.764510084453569e-05,
+      "loss": 0.0138,
+      "step": 44270
+    },
+    {
+      "action_loss": 0.006252707447856665,
+      "epoch": 39.81115107913669,
+      "step": 44270
+    },
+    {
+      "epoch": 39.81115107913669,
+      "step": 44270,
+      "torque_loss": 0.104990653693676
+    },
+    {
+      "epoch": 39.81115107913669,
+      "force_loss": 0.018709400668740273,
+      "step": 44270
+    },
+    {
+      "epoch": 39.82014388489208,
+      "grad_norm": 0.1215263083577156,
+      "learning_rate": 1.76240954813781e-05,
+      "loss": 0.0161,
+      "step": 44280
+    },
+    {
+      "action_loss": 0.0024274829775094986,
+      "epoch": 39.82014388489208,
+      "step": 44280
+    },
+    {
+      "epoch": 39.82014388489208,
+      "step": 44280,
+      "torque_loss": 0.15360279381275177
+    },
+    {
+      "epoch": 39.82014388489208,
+      "force_loss": 0.0024420542176812887,
+      "step": 44280
+    },
+    {
+      "epoch": 39.82913669064748,
+      "grad_norm": 0.15674108266830444,
+      "learning_rate": 1.7603099953165476e-05,
+      "loss": 0.0154,
+      "step": 44290
+    },
+    {
+      "action_loss": 0.0028388623613864183,
+      "epoch": 39.82913669064748,
+      "step": 44290
+    },
+    {
+      "epoch": 39.82913669064748,
+      "step": 44290,
+      "torque_loss": 0.1412268579006195
+    },
+    {
+      "epoch": 39.82913669064748,
+      "force_loss": 0.003971495199948549,
+      "step": 44290
+    },
+    {
+      "epoch": 39.83812949640288,
+      "grad_norm": 0.18095889687538147,
+      "learning_rate": 1.7582114266275683e-05,
+      "loss": 0.0192,
+      "step": 44300
+    },
+    {
+      "action_loss": 0.007399098481982946,
+      "epoch": 39.83812949640288,
+      "step": 44300
+    },
+    {
+      "epoch": 39.83812949640288,
+      "step": 44300,
+      "torque_loss": 0.1313752382993698
+    },
+    {
+      "epoch": 39.83812949640288,
+      "force_loss": 0.006780050694942474,
+      "step": 44300
+    },
+    {
+      "epoch": 39.847122302158276,
+      "grad_norm": 0.06766969710588455,
+      "learning_rate": 1.756113842708364e-05,
+      "loss": 0.015,
+      "step": 44310
+    },
+    {
+      "action_loss": 0.005207750480622053,
+      "epoch": 39.847122302158276,
+      "step": 44310
+    },
+    {
+      "epoch": 39.847122302158276,
+      "step": 44310,
+      "torque_loss": 0.09442100673913956
+    },
+    {
+      "epoch": 39.847122302158276,
+      "force_loss": 0.0039870780892670155,
+      "step": 44310
+    },
+    {
+      "epoch": 39.856115107913666,
+      "grad_norm": 0.11593057960271835,
+      "learning_rate": 1.7540172441961245e-05,
+      "loss": 0.017,
+      "step": 44320
+    },
+    {
+      "action_loss": 0.0027600687462836504,
+      "epoch": 39.856115107913666,
+      "step": 44320
+    },
+    {
+      "epoch": 39.856115107913666,
+      "step": 44320,
+      "torque_loss": 0.11569730192422867
+    },
+    {
+      "epoch": 39.856115107913666,
+      "force_loss": 0.0012520946329459548,
+      "step": 44320
+    },
+    {
+      "epoch": 39.865107913669064,
+      "grad_norm": 0.09822013974189758,
+      "learning_rate": 1.7519216317277387e-05,
+      "loss": 0.0132,
+      "step": 44330
+    },
+    {
+      "action_loss": 0.0033128943759948015,
+      "epoch": 39.865107913669064,
+      "step": 44330
+    },
+    {
+      "epoch": 39.865107913669064,
+      "step": 44330,
+      "torque_loss": 0.21037475764751434
+    },
+    {
+      "epoch": 39.865107913669064,
+      "force_loss": 0.002765208249911666,
+      "step": 44330
+    },
+    {
+      "epoch": 39.87410071942446,
+      "grad_norm": 0.10911911725997925,
+      "learning_rate": 1.7498270059398046e-05,
+      "loss": 0.0151,
+      "step": 44340
+    },
+    {
+      "action_loss": 0.001463107648305595,
+      "epoch": 39.87410071942446,
+      "step": 44340
+    },
+    {
+      "epoch": 39.87410071942446,
+      "step": 44340,
+      "torque_loss": 0.13337288796901703
+    },
+    {
+      "epoch": 39.87410071942446,
+      "force_loss": 0.0008917516097426414,
+      "step": 44340
+    },
+    {
+      "epoch": 39.88309352517986,
+      "grad_norm": 0.34264805912971497,
+      "learning_rate": 1.7477333674686062e-05,
+      "loss": 0.0161,
+      "step": 44350
+    },
+    {
+      "action_loss": 0.0015004229499027133,
+      "epoch": 39.88309352517986,
+      "step": 44350
+    },
+    {
+      "epoch": 39.88309352517986,
+      "step": 44350,
+      "torque_loss": 0.1908785104751587
+    },
+    {
+      "epoch": 39.88309352517986,
+      "force_loss": 0.0020156477112323046,
+      "step": 44350
+    },
+    {
+      "epoch": 39.89208633093525,
+      "grad_norm": 0.2536611557006836,
+      "learning_rate": 1.745640716950142e-05,
+      "loss": 0.0185,
+      "step": 44360
+    },
+    {
+      "action_loss": 0.011019092053174973,
+      "epoch": 39.89208633093525,
+      "step": 44360
+    },
+    {
+      "epoch": 39.89208633093525,
+      "step": 44360,
+      "torque_loss": 0.14018167555332184
+    },
+    {
+      "epoch": 39.89208633093525,
+      "force_loss": 0.014119830913841724,
+      "step": 44360
+    },
+    {
+      "epoch": 39.90107913669065,
+      "grad_norm": 0.4099871814250946,
+      "learning_rate": 1.7435490550201017e-05,
+      "loss": 0.0192,
+      "step": 44370
+    },
+    {
+      "action_loss": 0.0010389521485194564,
+      "epoch": 39.90107913669065,
+      "step": 44370
+    },
+    {
+      "epoch": 39.90107913669065,
+      "step": 44370,
+      "torque_loss": 0.09374342113733292
+    },
+    {
+      "epoch": 39.90107913669065,
+      "force_loss": 0.0029633017256855965,
+      "step": 44370
+    },
+    {
+      "epoch": 39.910071942446045,
+      "grad_norm": 0.11975488066673279,
+      "learning_rate": 1.7414583823138762e-05,
+      "loss": 0.014,
+      "step": 44380
+    },
+    {
+      "action_loss": 0.0014100942062214017,
+      "epoch": 39.910071942446045,
+      "step": 44380
+    },
+    {
+      "epoch": 39.910071942446045,
+      "step": 44380,
+      "torque_loss": 0.12192344665527344
+    },
+    {
+      "epoch": 39.910071942446045,
+      "force_loss": 0.0009730260935612023,
+      "step": 44380
+    },
+    {
+      "epoch": 39.919064748201436,
+      "grad_norm": 0.0840345025062561,
+      "learning_rate": 1.739368699466558e-05,
+      "loss": 0.0162,
+      "step": 44390
+    },
+    {
+      "action_loss": 0.0023275448475033045,
+      "epoch": 39.919064748201436,
+      "step": 44390
+    },
+    {
+      "epoch": 39.919064748201436,
+      "step": 44390,
+      "torque_loss": 0.12090662121772766
+    },
+    {
+      "epoch": 39.919064748201436,
+      "force_loss": 0.004105397965759039,
+      "step": 44390
+    },
+    {
+      "epoch": 39.92805755395683,
+      "grad_norm": 0.09263237565755844,
+      "learning_rate": 1.737280007112935e-05,
+      "loss": 0.0165,
+      "step": 44400
+    },
+    {
+      "action_loss": 0.0005101033020764589,
+      "epoch": 39.92805755395683,
+      "step": 44400
+    },
+    {
+      "epoch": 39.92805755395683,
+      "step": 44400,
+      "torque_loss": 0.0995657816529274
+    },
+    {
+      "epoch": 39.92805755395683,
+      "force_loss": 0.0005334276938810945,
+      "step": 44400
+    },
+    {
+      "epoch": 39.93705035971223,
+      "grad_norm": 0.11976148188114166,
+      "learning_rate": 1.735192305887502e-05,
+      "loss": 0.0139,
+      "step": 44410
+    },
+    {
+      "action_loss": 0.0017665340565145016,
+      "epoch": 39.93705035971223,
+      "step": 44410
+    },
+    {
+      "epoch": 39.93705035971223,
+      "step": 44410,
+      "torque_loss": 0.10761773586273193
+    },
+    {
+      "epoch": 39.93705035971223,
+      "force_loss": 0.0019337544217705727,
+      "step": 44410
+    },
+    {
+      "epoch": 39.94604316546763,
+      "grad_norm": 0.1317591369152069,
+      "learning_rate": 1.733105596424441e-05,
+      "loss": 0.0163,
+      "step": 44420
+    },
+    {
+      "action_loss": 0.003290182910859585,
+      "epoch": 39.94604316546763,
+      "step": 44420
+    },
+    {
+      "epoch": 39.94604316546763,
+      "step": 44420,
+      "torque_loss": 0.1056455448269844
+    },
+    {
+      "epoch": 39.94604316546763,
+      "force_loss": 0.007925410754978657,
+      "step": 44420
+    },
+    {
+      "epoch": 39.95503597122302,
+      "grad_norm": 0.09897452592849731,
+      "learning_rate": 1.7310198793576437e-05,
+      "loss": 0.0162,
+      "step": 44430
+    },
+    {
+      "action_loss": 0.0020759671460837126,
+      "epoch": 39.95503597122302,
+      "step": 44430
+    },
+    {
+      "epoch": 39.95503597122302,
+      "step": 44430,
+      "torque_loss": 0.1621129810810089
+    },
+    {
+      "epoch": 39.95503597122302,
+      "force_loss": 0.011458312161266804,
+      "step": 44430
+    },
+    {
+      "epoch": 39.96402877697842,
+      "grad_norm": 0.14866717159748077,
+      "learning_rate": 1.7289351553206952e-05,
+      "loss": 0.0173,
+      "step": 44440
+    },
+    {
+      "action_loss": 0.001105463714338839,
+      "epoch": 39.96402877697842,
+      "step": 44440
+    },
+    {
+      "epoch": 39.96402877697842,
+      "step": 44440,
+      "torque_loss": 0.15111349523067474
+    },
+    {
+      "epoch": 39.96402877697842,
+      "force_loss": 0.0043312846682965755,
+      "step": 44440
+    },
+    {
+      "epoch": 39.973021582733814,
+      "grad_norm": 0.3920046091079712,
+      "learning_rate": 1.7268514249468788e-05,
+      "loss": 0.0172,
+      "step": 44450
+    },
+    {
+      "action_loss": 0.0065963841043412685,
+      "epoch": 39.973021582733814,
+      "step": 44450
+    },
+    {
+      "epoch": 39.973021582733814,
+      "step": 44450,
+      "torque_loss": 0.11663947254419327
+    },
+    {
+      "epoch": 39.973021582733814,
+      "force_loss": 0.020411519333720207,
+      "step": 44450
+    },
+    {
+      "epoch": 39.98201438848921,
+      "grad_norm": 0.23445050418376923,
+      "learning_rate": 1.7247686888691765e-05,
+      "loss": 0.0155,
+      "step": 44460
+    },
+    {
+      "action_loss": 0.0011924013961106539,
+      "epoch": 39.98201438848921,
+      "step": 44460
+    },
+    {
+      "epoch": 39.98201438848921,
+      "step": 44460,
+      "torque_loss": 0.13264767825603485
+    },
+    {
+      "epoch": 39.98201438848921,
+      "force_loss": 0.0013270940398797393,
+      "step": 44460
+    },
+    {
+      "epoch": 39.9910071942446,
+      "grad_norm": 0.14190790057182312,
+      "learning_rate": 1.7226869477202694e-05,
+      "loss": 0.014,
+      "step": 44470
+    },
+    {
+      "action_loss": 0.004054693039506674,
+      "epoch": 39.9910071942446,
+      "step": 44470
+    },
+    {
+      "epoch": 39.9910071942446,
+      "step": 44470,
+      "torque_loss": 0.1341247707605362
+    },
+    {
+      "epoch": 39.9910071942446,
+      "force_loss": 0.003079628571867943,
+      "step": 44470
+    },
+    {
+      "epoch": 40.0,
+      "grad_norm": 0.2483171969652176,
+      "learning_rate": 1.7206062021325336e-05,
+      "loss": 0.0185,
+      "step": 44480
+    },
+    {
+      "action_loss": 0.0015627824468538165,
+      "epoch": 40.0,
+      "step": 44480
+    },
+    {
+      "epoch": 40.0,
+      "step": 44480,
+      "torque_loss": 0.11845695972442627
+    },
+    {
+      "epoch": 40.0,
+      "force_loss": 0.0026303771883249283,
+      "step": 44480
+    },
+    {
+      "epoch": 40.0089928057554,
+      "grad_norm": 0.1087113544344902,
+      "learning_rate": 1.7185264527380502e-05,
+      "loss": 0.0166,
+      "step": 44490
+    },
+    {
+      "action_loss": 0.0012507130159065127,
+      "epoch": 40.0089928057554,
+      "step": 44490
+    },
+    {
+      "epoch": 40.0089928057554,
+      "step": 44490,
+      "torque_loss": 0.08470701426267624
+    },
+    {
+      "epoch": 40.0089928057554,
+      "force_loss": 0.0010447535896673799,
+      "step": 44490
+    },
+    {
+      "epoch": 40.01798561151079,
+      "grad_norm": 0.27818405628204346,
+      "learning_rate": 1.716447700168584e-05,
+      "loss": 0.017,
+      "step": 44500
+    },
+    {
+      "action_loss": 0.0033613082487136126,
+      "epoch": 40.01798561151079,
+      "step": 44500
+    },
+    {
+      "epoch": 40.01798561151079,
+      "step": 44500,
+      "torque_loss": 0.09948086738586426
+    },
+    {
+      "epoch": 40.01798561151079,
+      "force_loss": 0.005918712820857763,
+      "step": 44500
+    },
+    {
+      "epoch": 40.026978417266186,
+      "grad_norm": 0.6830267906188965,
+      "learning_rate": 1.714369945055611e-05,
+      "loss": 0.0208,
+      "step": 44510
+    },
+    {
+      "action_loss": 0.0006649745628237724,
+      "epoch": 40.026978417266186,
+      "step": 44510
+    },
+    {
+      "epoch": 40.026978417266186,
+      "step": 44510,
+      "torque_loss": 0.10315630584955215
+    },
+    {
+      "epoch": 40.026978417266186,
+      "force_loss": 0.0006237350753508508,
+      "step": 44510
+    },
+    {
+      "epoch": 40.03597122302158,
+      "grad_norm": 0.07630914449691772,
+      "learning_rate": 1.7122931880302968e-05,
+      "loss": 0.0141,
+      "step": 44520
+    },
+    {
+      "action_loss": 0.001907266560010612,
+      "epoch": 40.03597122302158,
+      "step": 44520
+    },
+    {
+      "epoch": 40.03597122302158,
+      "step": 44520,
+      "torque_loss": 0.12051069736480713
+    },
+    {
+      "epoch": 40.03597122302158,
+      "force_loss": 0.0050184838473796844,
+      "step": 44520
+    },
+    {
+      "epoch": 40.04496402877698,
+      "grad_norm": 0.40941163897514343,
+      "learning_rate": 1.710217429723505e-05,
+      "loss": 0.0156,
+      "step": 44530
+    },
+    {
+      "action_loss": 0.0012308628065511584,
+      "epoch": 40.04496402877698,
+      "step": 44530
+    },
+    {
+      "epoch": 40.04496402877698,
+      "step": 44530,
+      "torque_loss": 0.14783541858196259
+    },
+    {
+      "epoch": 40.04496402877698,
+      "force_loss": 0.0012665995163843036,
+      "step": 44530
+    },
+    {
+      "epoch": 40.05395683453237,
+      "grad_norm": 0.11552932113409042,
+      "learning_rate": 1.7081426707657972e-05,
+      "loss": 0.0156,
+      "step": 44540
+    },
+    {
+      "action_loss": 0.003938527312129736,
+      "epoch": 40.05395683453237,
+      "step": 44540
+    },
+    {
+      "epoch": 40.05395683453237,
+      "step": 44540,
+      "torque_loss": 0.18767376244068146
+    },
+    {
+      "epoch": 40.05395683453237,
+      "force_loss": 0.011259282939136028,
+      "step": 44540
+    },
+    {
+      "epoch": 40.06294964028777,
+      "grad_norm": 0.16364285349845886,
+      "learning_rate": 1.7060689117874275e-05,
+      "loss": 0.017,
+      "step": 44550
+    },
+    {
+      "action_loss": 0.0019341300940141082,
+      "epoch": 40.06294964028777,
+      "step": 44550
+    },
+    {
+      "epoch": 40.06294964028777,
+      "step": 44550,
+      "torque_loss": 0.16263054311275482
+    },
+    {
+      "epoch": 40.06294964028777,
+      "force_loss": 0.0060359700582921505,
+      "step": 44550
+    },
+    {
+      "epoch": 40.07194244604317,
+      "grad_norm": 0.12794502079486847,
+      "learning_rate": 1.703996153418354e-05,
+      "loss": 0.0141,
+      "step": 44560
+    },
+    {
+      "action_loss": 0.0013586924178525805,
+      "epoch": 40.07194244604317,
+      "step": 44560
+    },
+    {
+      "epoch": 40.07194244604317,
+      "step": 44560,
+      "torque_loss": 0.10127987712621689
+    },
+    {
+      "epoch": 40.07194244604317,
+      "force_loss": 0.0018603671342134476,
+      "step": 44560
+    },
+    {
+      "epoch": 40.080935251798564,
+      "grad_norm": 0.2764982283115387,
+      "learning_rate": 1.7019243962882205e-05,
+      "loss": 0.0161,
+      "step": 44570
+    },
+    {
+      "action_loss": 0.00107524951454252,
+      "epoch": 40.080935251798564,
+      "step": 44570
+    },
+    {
+      "epoch": 40.080935251798564,
+      "step": 44570,
+      "torque_loss": 0.08953773230314255
+    },
+    {
+      "epoch": 40.080935251798564,
+      "force_loss": 0.003173167584463954,
+      "step": 44570
+    },
+    {
+      "epoch": 40.089928057553955,
+      "grad_norm": 0.08040502667427063,
+      "learning_rate": 1.6998536410263754e-05,
+      "loss": 0.0138,
+      "step": 44580
+    },
+    {
+      "action_loss": 0.0048240735195577145,
+      "epoch": 40.089928057553955,
+      "step": 44580
+    },
+    {
+      "epoch": 40.089928057553955,
+      "step": 44580,
+      "torque_loss": 0.14370672404766083
+    },
+    {
+      "epoch": 40.089928057553955,
+      "force_loss": 0.012612252496182919,
+      "step": 44580
+    },
+    {
+      "epoch": 40.09892086330935,
+      "grad_norm": 0.11724365502595901,
+      "learning_rate": 1.6977838882618596e-05,
+      "loss": 0.0143,
+      "step": 44590
+    },
+    {
+      "action_loss": 0.0014210754306986928,
+      "epoch": 40.09892086330935,
+      "step": 44590
+    },
+    {
+      "epoch": 40.09892086330935,
+      "step": 44590,
+      "torque_loss": 0.10057639330625534
+    },
+    {
+      "epoch": 40.09892086330935,
+      "force_loss": 0.0023515161592513323,
+      "step": 44590
+    },
+    {
+      "epoch": 40.10791366906475,
+      "grad_norm": 0.38110387325286865,
+      "learning_rate": 1.6957151386234088e-05,
+      "loss": 0.0143,
+      "step": 44600
+    },
+    {
+      "action_loss": 0.0024461140856146812,
+      "epoch": 40.10791366906475,
+      "step": 44600
+    },
+    {
+      "epoch": 40.10791366906475,
+      "step": 44600,
+      "torque_loss": 0.15897291898727417
+    },
+    {
+      "epoch": 40.10791366906475,
+      "force_loss": 0.007345343474298716,
+      "step": 44600
+    },
+    {
+      "epoch": 40.11690647482014,
+      "grad_norm": 0.0928143784403801,
+      "learning_rate": 1.6936473927394536e-05,
+      "loss": 0.0156,
+      "step": 44610
+    },
+    {
+      "action_loss": 0.003792443545535207,
+      "epoch": 40.11690647482014,
+      "step": 44610
+    },
+    {
+      "epoch": 40.11690647482014,
+      "step": 44610,
+      "torque_loss": 0.11347117274999619
+    },
+    {
+      "epoch": 40.11690647482014,
+      "force_loss": 0.0037435460835695267,
+      "step": 44610
+    },
+    {
+      "epoch": 40.12589928057554,
+      "grad_norm": 0.0897802785038948,
+      "learning_rate": 1.6915806512381222e-05,
+      "loss": 0.0157,
+      "step": 44620
+    },
+    {
+      "action_loss": 0.0036807898432016373,
+      "epoch": 40.12589928057554,
+      "step": 44620
+    },
+    {
+      "epoch": 40.12589928057554,
+      "step": 44620,
+      "torque_loss": 0.08641115576028824
+    },
+    {
+      "epoch": 40.12589928057554,
+      "force_loss": 0.0037523380015045404,
+      "step": 44620
+    },
+    {
+      "epoch": 40.134892086330936,
+      "grad_norm": 0.17474913597106934,
+      "learning_rate": 1.6895149147472344e-05,
+      "loss": 0.0166,
+      "step": 44630
+    },
+    {
+      "action_loss": 0.003175650490447879,
+      "epoch": 40.134892086330936,
+      "step": 44630
+    },
+    {
+      "epoch": 40.134892086330936,
+      "step": 44630,
+      "torque_loss": 0.09962426871061325
+    },
+    {
+      "epoch": 40.134892086330936,
+      "force_loss": 0.006837495137006044,
+      "step": 44630
+    },
+    {
+      "epoch": 40.143884892086334,
+      "grad_norm": 0.36691057682037354,
+      "learning_rate": 1.6874501838943073e-05,
+      "loss": 0.0174,
+      "step": 44640
+    },
+    {
+      "action_loss": 0.0024612240958958864,
+      "epoch": 40.143884892086334,
+      "step": 44640
+    },
+    {
+      "epoch": 40.143884892086334,
+      "step": 44640,
+      "torque_loss": 0.17060263454914093
+    },
+    {
+      "epoch": 40.143884892086334,
+      "force_loss": 0.011553235352039337,
+      "step": 44640
+    },
+    {
+      "epoch": 40.152877697841724,
+      "grad_norm": 0.4615657925605774,
+      "learning_rate": 1.6853864593065506e-05,
+      "loss": 0.016,
+      "step": 44650
+    },
+    {
+      "action_loss": 0.001044761505909264,
+      "epoch": 40.152877697841724,
+      "step": 44650
+    },
+    {
+      "epoch": 40.152877697841724,
+      "step": 44650,
+      "torque_loss": 0.11981693655252457
+    },
+    {
+      "epoch": 40.152877697841724,
+      "force_loss": 0.0020934399217367172,
+      "step": 44650
+    },
+    {
+      "epoch": 40.16187050359712,
+      "grad_norm": 0.2949974834918976,
+      "learning_rate": 1.683323741610871e-05,
+      "loss": 0.0157,
+      "step": 44660
+    },
+    {
+      "action_loss": 0.0012690444709733129,
+      "epoch": 40.16187050359712,
+      "step": 44660
+    },
+    {
+      "epoch": 40.16187050359712,
+      "step": 44660,
+      "torque_loss": 0.15798909962177277
+    },
+    {
+      "epoch": 40.16187050359712,
+      "force_loss": 0.0011633285321295261,
+      "step": 44660
+    },
+    {
+      "epoch": 40.17086330935252,
+      "grad_norm": 0.38125699758529663,
+      "learning_rate": 1.6812620314338674e-05,
+      "loss": 0.0166,
+      "step": 44670
+    },
+    {
+      "action_loss": 0.0026361290365457535,
+      "epoch": 40.17086330935252,
+      "step": 44670
+    },
+    {
+      "epoch": 40.17086330935252,
+      "step": 44670,
+      "torque_loss": 0.16972856223583221
+    },
+    {
+      "epoch": 40.17086330935252,
+      "force_loss": 0.002562169684097171,
+      "step": 44670
+    },
+    {
+      "epoch": 40.17985611510792,
+      "grad_norm": 0.1883958876132965,
+      "learning_rate": 1.6792013294018326e-05,
+      "loss": 0.0147,
+      "step": 44680
+    },
+    {
+      "action_loss": 0.004774322733283043,
+      "epoch": 40.17985611510792,
+      "step": 44680
+    },
+    {
+      "epoch": 40.17985611510792,
+      "step": 44680,
+      "torque_loss": 0.11046300083398819
+    },
+    {
+      "epoch": 40.17985611510792,
+      "force_loss": 0.002918919548392296,
+      "step": 44680
+    },
+    {
+      "epoch": 40.18884892086331,
+      "grad_norm": 0.13169023394584656,
+      "learning_rate": 1.6771416361407526e-05,
+      "loss": 0.0153,
+      "step": 44690
+    },
+    {
+      "action_loss": 0.000996866961941123,
+      "epoch": 40.18884892086331,
+      "step": 44690
+    },
+    {
+      "epoch": 40.18884892086331,
+      "step": 44690,
+      "torque_loss": 0.09149760007858276
+    },
+    {
+      "epoch": 40.18884892086331,
+      "force_loss": 0.0015735543565824628,
+      "step": 44690
+    },
+    {
+      "epoch": 40.197841726618705,
+      "grad_norm": 0.23048055171966553,
+      "learning_rate": 1.675082952276308e-05,
+      "loss": 0.0137,
+      "step": 44700
+    },
+    {
+      "action_loss": 0.0007311130757443607,
+      "epoch": 40.197841726618705,
+      "step": 44700
+    },
+    {
+      "epoch": 40.197841726618705,
+      "step": 44700,
+      "torque_loss": 0.09804457426071167
+    },
+    {
+      "epoch": 40.197841726618705,
+      "force_loss": 0.0007812807452864945,
+      "step": 44700
+    },
+    {
+      "epoch": 40.2068345323741,
+      "grad_norm": 0.22695928812026978,
+      "learning_rate": 1.6730252784338757e-05,
+      "loss": 0.0144,
+      "step": 44710
+    },
+    {
+      "action_loss": 0.0024429606273770332,
+      "epoch": 40.2068345323741,
+      "step": 44710
+    },
+    {
+      "epoch": 40.2068345323741,
+      "step": 44710,
+      "torque_loss": 0.13664865493774414
+    },
+    {
+      "epoch": 40.2068345323741,
+      "force_loss": 0.006511110812425613,
+      "step": 44710
+    },
+    {
+      "epoch": 40.21582733812949,
+      "grad_norm": 0.08444716781377792,
+      "learning_rate": 1.6709686152385166e-05,
+      "loss": 0.0148,
+      "step": 44720
+    },
+    {
+      "action_loss": 0.002623581560328603,
+      "epoch": 40.21582733812949,
+      "step": 44720
+    },
+    {
+      "epoch": 40.21582733812949,
+      "step": 44720,
+      "torque_loss": 0.12171286344528198
+    },
+    {
+      "epoch": 40.21582733812949,
+      "force_loss": 0.011893153190612793,
+      "step": 44720
+    },
+    {
+      "epoch": 40.22482014388489,
+      "grad_norm": 0.2250167280435562,
+      "learning_rate": 1.668912963314998e-05,
+      "loss": 0.0171,
+      "step": 44730
+    },
+    {
+      "action_loss": 0.008974972181022167,
+      "epoch": 40.22482014388489,
+      "step": 44730
+    },
+    {
+      "epoch": 40.22482014388489,
+      "step": 44730,
+      "torque_loss": 0.11451403051614761
+    },
+    {
+      "epoch": 40.22482014388489,
+      "force_loss": 0.020065506920218468,
+      "step": 44730
+    },
+    {
+      "epoch": 40.23381294964029,
+      "grad_norm": 0.1760580986738205,
+      "learning_rate": 1.6668583232877653e-05,
+      "loss": 0.0171,
+      "step": 44740
+    },
+    {
+      "action_loss": 0.0054635354317724705,
+      "epoch": 40.23381294964029,
+      "step": 44740
+    },
+    {
+      "epoch": 40.23381294964029,
+      "step": 44740,
+      "torque_loss": 0.09759384393692017
+    },
+    {
+      "epoch": 40.23381294964029,
+      "force_loss": 0.0041343956254422665,
+      "step": 44740
+    },
+    {
+      "epoch": 40.242805755395686,
+      "grad_norm": 0.12554466724395752,
+      "learning_rate": 1.6648046957809698e-05,
+      "loss": 0.0187,
+      "step": 44750
+    },
+    {
+      "action_loss": 0.008528563193976879,
+      "epoch": 40.242805755395686,
+      "step": 44750
+    },
+    {
+      "epoch": 40.242805755395686,
+      "step": 44750,
+      "torque_loss": 0.12795187532901764
+    },
+    {
+      "epoch": 40.242805755395686,
+      "force_loss": 0.013514988124370575,
+      "step": 44750
+    },
+    {
+      "epoch": 40.25179856115108,
+      "grad_norm": 0.12210524827241898,
+      "learning_rate": 1.6627520814184462e-05,
+      "loss": 0.0157,
+      "step": 44760
+    },
+    {
+      "action_loss": 0.006425800267606974,
+      "epoch": 40.25179856115108,
+      "step": 44760
+    },
+    {
+      "epoch": 40.25179856115108,
+      "step": 44760,
+      "torque_loss": 0.13641668856143951
+    },
+    {
+      "epoch": 40.25179856115108,
+      "force_loss": 0.01587783545255661,
+      "step": 44760
+    },
+    {
+      "epoch": 40.260791366906474,
+      "grad_norm": 0.13937558233737946,
+      "learning_rate": 1.660700480823726e-05,
+      "loss": 0.0162,
+      "step": 44770
+    },
+    {
+      "action_loss": 0.002961226971819997,
+      "epoch": 40.260791366906474,
+      "step": 44770
+    },
+    {
+      "epoch": 40.260791366906474,
+      "step": 44770,
+      "torque_loss": 0.1057615652680397
+    },
+    {
+      "epoch": 40.260791366906474,
+      "force_loss": 0.005273377988487482,
+      "step": 44770
+    },
+    {
+      "epoch": 40.26978417266187,
+      "grad_norm": 0.07900244742631912,
+      "learning_rate": 1.65864989462003e-05,
+      "loss": 0.0166,
+      "step": 44780
+    },
+    {
+      "action_loss": 0.001128497184254229,
+      "epoch": 40.26978417266187,
+      "step": 44780
+    },
+    {
+      "epoch": 40.26978417266187,
+      "step": 44780,
+      "torque_loss": 0.1287650316953659
+    },
+    {
+      "epoch": 40.26978417266187,
+      "force_loss": 0.0007052418659441173,
+      "step": 44780
+    },
+    {
+      "epoch": 40.27877697841727,
+      "grad_norm": 0.08304659277200699,
+      "learning_rate": 1.656600323430273e-05,
+      "loss": 0.019,
+      "step": 44790
+    },
+    {
+      "action_loss": 0.0005852977628819644,
+      "epoch": 40.27877697841727,
+      "step": 44790
+    },
+    {
+      "epoch": 40.27877697841727,
+      "step": 44790,
+      "torque_loss": 0.12477948516607285
+    },
+    {
+      "epoch": 40.27877697841727,
+      "force_loss": 0.0008666057256050408,
+      "step": 44790
+    },
+    {
+      "epoch": 40.28776978417266,
+      "grad_norm": 0.1904379278421402,
+      "learning_rate": 1.654551767877059e-05,
+      "loss": 0.0161,
+      "step": 44800
+    },
+    {
+      "action_loss": 0.000839108310174197,
+      "epoch": 40.28776978417266,
+      "step": 44800
+    },
+    {
+      "epoch": 40.28776978417266,
+      "step": 44800,
+      "torque_loss": 0.0952586904168129
+    },
+    {
+      "epoch": 40.28776978417266,
+      "force_loss": 0.0018974501872435212,
+      "step": 44800
+    },
+    {
+      "epoch": 40.29676258992806,
+      "grad_norm": 0.156271293759346,
+      "learning_rate": 1.6525042285826874e-05,
+      "loss": 0.0183,
+      "step": 44810
+    },
+    {
+      "action_loss": 0.002238394459709525,
+      "epoch": 40.29676258992806,
+      "step": 44810
+    },
+    {
+      "epoch": 40.29676258992806,
+      "step": 44810,
+      "torque_loss": 0.1386720985174179
+    },
+    {
+      "epoch": 40.29676258992806,
+      "force_loss": 0.0026462338864803314,
+      "step": 44810
+    },
+    {
+      "epoch": 40.305755395683455,
+      "grad_norm": 0.4390171766281128,
+      "learning_rate": 1.6504577061691468e-05,
+      "loss": 0.0183,
+      "step": 44820
+    },
+    {
+      "action_loss": 0.0010000247275456786,
+      "epoch": 40.305755395683455,
+      "step": 44820
+    },
+    {
+      "epoch": 40.305755395683455,
+      "step": 44820,
+      "torque_loss": 0.13093285262584686
+    },
+    {
+      "epoch": 40.305755395683455,
+      "force_loss": 0.0015082870377227664,
+      "step": 44820
+    },
+    {
+      "epoch": 40.314748201438846,
+      "grad_norm": 0.5617092847824097,
+      "learning_rate": 1.6484122012581143e-05,
+      "loss": 0.016,
+      "step": 44830
+    },
+    {
+      "action_loss": 0.0021910963114351034,
+      "epoch": 40.314748201438846,
+      "step": 44830
+    },
+    {
+      "epoch": 40.314748201438846,
+      "step": 44830,
+      "torque_loss": 0.1063302755355835
+    },
+    {
+      "epoch": 40.314748201438846,
+      "force_loss": 0.0035671256482601166,
+      "step": 44830
+    },
+    {
+      "epoch": 40.32374100719424,
+      "grad_norm": 0.22056697309017181,
+      "learning_rate": 1.6463677144709623e-05,
+      "loss": 0.0152,
+      "step": 44840
+    },
+    {
+      "action_loss": 0.00223253364674747,
+      "epoch": 40.32374100719424,
+      "step": 44840
+    },
+    {
+      "epoch": 40.32374100719424,
+      "step": 44840,
+      "torque_loss": 0.14081676304340363
+    },
+    {
+      "epoch": 40.32374100719424,
+      "force_loss": 0.004668839275836945,
+      "step": 44840
+    },
+    {
+      "epoch": 40.33273381294964,
+      "grad_norm": 0.07791223376989365,
+      "learning_rate": 1.6443242464287493e-05,
+      "loss": 0.0147,
+      "step": 44850
+    },
+    {
+      "action_loss": 0.0032717313151806593,
+      "epoch": 40.33273381294964,
+      "step": 44850
+    },
+    {
+      "epoch": 40.33273381294964,
+      "step": 44850,
+      "torque_loss": 0.11260297149419785
+    },
+    {
+      "epoch": 40.33273381294964,
+      "force_loss": 0.003491375595331192,
+      "step": 44850
+    },
+    {
+      "epoch": 40.34172661870504,
+      "grad_norm": 0.135382741689682,
+      "learning_rate": 1.642281797752232e-05,
+      "loss": 0.0159,
+      "step": 44860
+    },
+    {
+      "action_loss": 0.0015873191878199577,
+      "epoch": 40.34172661870504,
+      "step": 44860
+    },
+    {
+      "epoch": 40.34172661870504,
+      "step": 44860,
+      "torque_loss": 0.15371070802211761
+    },
+    {
+      "epoch": 40.34172661870504,
+      "force_loss": 0.0024278899654746056,
+      "step": 44860
+    },
+    {
+      "epoch": 40.35071942446043,
+      "grad_norm": 0.5832369923591614,
+      "learning_rate": 1.6402403690618456e-05,
+      "loss": 0.016,
+      "step": 44870
+    },
+    {
+      "action_loss": 0.001395586528815329,
+      "epoch": 40.35071942446043,
+      "step": 44870
+    },
+    {
+      "epoch": 40.35071942446043,
+      "step": 44870,
+      "torque_loss": 0.13407404720783234
+    },
+    {
+      "epoch": 40.35071942446043,
+      "force_loss": 0.002249868120998144,
+      "step": 44870
+    },
+    {
+      "epoch": 40.35971223021583,
+      "grad_norm": 0.09243930876255035,
+      "learning_rate": 1.6381999609777295e-05,
+      "loss": 0.0132,
+      "step": 44880
+    },
+    {
+      "action_loss": 0.0059046498499810696,
+      "epoch": 40.35971223021583,
+      "step": 44880
+    },
+    {
+      "epoch": 40.35971223021583,
+      "step": 44880,
+      "torque_loss": 0.11999446153640747
+    },
+    {
+      "epoch": 40.35971223021583,
+      "force_loss": 0.003569086315110326,
+      "step": 44880
+    },
+    {
+      "epoch": 40.368705035971225,
+      "grad_norm": 0.17036353051662445,
+      "learning_rate": 1.6361605741196983e-05,
+      "loss": 0.0251,
+      "step": 44890
+    },
+    {
+      "action_loss": 0.003418646054342389,
+      "epoch": 40.368705035971225,
+      "step": 44890
+    },
+    {
+      "epoch": 40.368705035971225,
+      "step": 44890,
+      "torque_loss": 0.14726051688194275
+    },
+    {
+      "epoch": 40.368705035971225,
+      "force_loss": 0.004301040433347225,
+      "step": 44890
+    },
+    {
+      "epoch": 40.37769784172662,
+      "grad_norm": 0.23966199159622192,
+      "learning_rate": 1.63412220910727e-05,
+      "loss": 0.0202,
+      "step": 44900
+    },
+    {
+      "action_loss": 0.009670862928032875,
+      "epoch": 40.37769784172662,
+      "step": 44900
+    },
+    {
+      "epoch": 40.37769784172662,
+      "step": 44900,
+      "torque_loss": 0.10886164754629135
+    },
+    {
+      "epoch": 40.37769784172662,
+      "force_loss": 0.0035433850716799498,
+      "step": 44900
+    },
+    {
+      "epoch": 40.38669064748201,
+      "grad_norm": 0.09231775999069214,
+      "learning_rate": 1.6320848665596433e-05,
+      "loss": 0.0178,
+      "step": 44910
+    },
+    {
+      "action_loss": 0.0011204906040802598,
+      "epoch": 40.38669064748201,
+      "step": 44910
+    },
+    {
+      "epoch": 40.38669064748201,
+      "step": 44910,
+      "torque_loss": 0.12653951346874237
+    },
+    {
+      "epoch": 40.38669064748201,
+      "force_loss": 0.002186434343457222,
+      "step": 44910
+    },
+    {
+      "epoch": 40.39568345323741,
+      "grad_norm": 0.21757644414901733,
+      "learning_rate": 1.6300485470957095e-05,
+      "loss": 0.0162,
+      "step": 44920
+    },
+    {
+      "action_loss": 0.002309808973222971,
+      "epoch": 40.39568345323741,
+      "step": 44920
+    },
+    {
+      "epoch": 40.39568345323741,
+      "step": 44920,
+      "torque_loss": 0.14443175494670868
+    },
+    {
+      "epoch": 40.39568345323741,
+      "force_loss": 0.004321340471506119,
+      "step": 44920
+    },
+    {
+      "epoch": 40.40467625899281,
+      "grad_norm": 0.1127144917845726,
+      "learning_rate": 1.6280132513340483e-05,
+      "loss": 0.0143,
+      "step": 44930
+    },
+    {
+      "action_loss": 0.0016665965085849166,
+      "epoch": 40.40467625899281,
+      "step": 44930
+    },
+    {
+      "epoch": 40.40467625899281,
+      "step": 44930,
+      "torque_loss": 0.12928690016269684
+    },
+    {
+      "epoch": 40.40467625899281,
+      "force_loss": 0.0019769768696278334,
+      "step": 44930
+    },
+    {
+      "epoch": 40.4136690647482,
+      "grad_norm": 0.11336503177881241,
+      "learning_rate": 1.62597897989293e-05,
+      "loss": 0.0149,
+      "step": 44940
+    },
+    {
+      "action_loss": 0.0010945176472887397,
+      "epoch": 40.4136690647482,
+      "step": 44940
+    },
+    {
+      "epoch": 40.4136690647482,
+      "step": 44940,
+      "torque_loss": 0.1361013948917389
+    },
+    {
+      "epoch": 40.4136690647482,
+      "force_loss": 0.0025673715863376856,
+      "step": 44940
+    },
+    {
+      "epoch": 40.422661870503596,
+      "grad_norm": 0.10752066224813461,
+      "learning_rate": 1.623945733390309e-05,
+      "loss": 0.0157,
+      "step": 44950
+    },
+    {
+      "action_loss": 0.002123811049386859,
+      "epoch": 40.422661870503596,
+      "step": 44950
+    },
+    {
+      "epoch": 40.422661870503596,
+      "step": 44950,
+      "torque_loss": 0.12423846125602722
+    },
+    {
+      "epoch": 40.422661870503596,
+      "force_loss": 0.0037513121496886015,
+      "step": 44950
+    },
+    {
+      "epoch": 40.431654676258994,
+      "grad_norm": 0.1287233978509903,
+      "learning_rate": 1.6219135124438374e-05,
+      "loss": 0.0141,
+      "step": 44960
+    },
+    {
+      "action_loss": 0.0007463112124241889,
+      "epoch": 40.431654676258994,
+      "step": 44960
+    },
+    {
+      "epoch": 40.431654676258994,
+      "step": 44960,
+      "torque_loss": 0.12744678556919098
+    },
+    {
+      "epoch": 40.431654676258994,
+      "force_loss": 0.0012125533539801836,
+      "step": 44960
+    },
+    {
+      "epoch": 40.44064748201439,
+      "grad_norm": 0.14318744838237762,
+      "learning_rate": 1.6198823176708465e-05,
+      "loss": 0.0151,
+      "step": 44970
+    },
+    {
+      "action_loss": 0.0009360754047520459,
+      "epoch": 40.44064748201439,
+      "step": 44970
+    },
+    {
+      "epoch": 40.44064748201439,
+      "step": 44970,
+      "torque_loss": 0.1614072471857071
+    },
+    {
+      "epoch": 40.44064748201439,
+      "force_loss": 0.0014803907833993435,
+      "step": 44970
+    },
+    {
+      "epoch": 40.44964028776978,
+      "grad_norm": 0.3382089138031006,
+      "learning_rate": 1.6178521496883613e-05,
+      "loss": 0.0196,
+      "step": 44980
+    },
+    {
+      "action_loss": 0.004354895558208227,
+      "epoch": 40.44964028776978,
+      "step": 44980
+    },
+    {
+      "epoch": 40.44964028776978,
+      "step": 44980,
+      "torque_loss": 0.12859804928302765
+    },
+    {
+      "epoch": 40.44964028776978,
+      "force_loss": 0.020465349778532982,
+      "step": 44980
+    },
+    {
+      "epoch": 40.45863309352518,
+      "grad_norm": 0.31221604347229004,
+      "learning_rate": 1.6158230091130926e-05,
+      "loss": 0.0143,
+      "step": 44990
+    },
+    {
+      "action_loss": 0.0008308025426231325,
+      "epoch": 40.45863309352518,
+      "step": 44990
+    },
+    {
+      "epoch": 40.45863309352518,
+      "step": 44990,
+      "torque_loss": 0.12496646493673325
+    },
+    {
+      "epoch": 40.45863309352518,
+      "force_loss": 0.001841373392380774,
+      "step": 44990
+    },
+    {
+      "epoch": 40.46762589928058,
+      "grad_norm": 0.14317816495895386,
+      "learning_rate": 1.613794896561438e-05,
+      "loss": 0.0152,
+      "step": 45000
+    },
+    {
+      "action_loss": 0.0012701962841674685,
+      "epoch": 40.46762589928058,
+      "step": 45000
+    },
+    {
+      "epoch": 40.46762589928058,
+      "step": 45000,
+      "torque_loss": 0.07607150822877884
+    },
+    {
+      "epoch": 40.46762589928058,
+      "force_loss": 0.0012807067250832915,
+      "step": 45000
+    },
+    {
+      "epoch": 40.476618705035975,
+      "grad_norm": 0.27331069111824036,
+      "learning_rate": 1.6117678126494894e-05,
+      "loss": 0.0148,
+      "step": 45010
+    },
+    {
+      "action_loss": 0.0020423864480108023,
+      "epoch": 40.476618705035975,
+      "step": 45010
+    },
+    {
+      "epoch": 40.476618705035975,
+      "step": 45010,
+      "torque_loss": 0.11822780221700668
+    },
+    {
+      "epoch": 40.476618705035975,
+      "force_loss": 0.005563536658883095,
+      "step": 45010
+    },
+    {
+      "epoch": 40.485611510791365,
+      "grad_norm": 0.20098598301410675,
+      "learning_rate": 1.6097417579930153e-05,
+      "loss": 0.0158,
+      "step": 45020
+    },
+    {
+      "action_loss": 0.0012331913458183408,
+      "epoch": 40.485611510791365,
+      "step": 45020
+    },
+    {
+      "epoch": 40.485611510791365,
+      "step": 45020,
+      "torque_loss": 0.11882391571998596
+    },
+    {
+      "epoch": 40.485611510791365,
+      "force_loss": 0.003346501151099801,
+      "step": 45020
+    },
+    {
+      "epoch": 40.49460431654676,
+      "grad_norm": 0.19315105676651,
+      "learning_rate": 1.6077167332074834e-05,
+      "loss": 0.0155,
+      "step": 45030
+    },
+    {
+      "action_loss": 0.0031557867769151926,
+      "epoch": 40.49460431654676,
+      "step": 45030
+    },
+    {
+      "epoch": 40.49460431654676,
+      "step": 45030,
+      "torque_loss": 0.09228048473596573
+    },
+    {
+      "epoch": 40.49460431654676,
+      "force_loss": 0.005141712725162506,
+      "step": 45030
+    },
+    {
+      "epoch": 40.50359712230216,
+      "grad_norm": 0.33281370997428894,
+      "learning_rate": 1.605692738908037e-05,
+      "loss": 0.0136,
+      "step": 45040
+    },
+    {
+      "action_loss": 0.0030528989154845476,
+      "epoch": 40.50359712230216,
+      "step": 45040
+    },
+    {
+      "epoch": 40.50359712230216,
+      "step": 45040,
+      "torque_loss": 0.17031632363796234
+    },
+    {
+      "epoch": 40.50359712230216,
+      "force_loss": 0.0069494531489908695,
+      "step": 45040
+    },
+    {
+      "epoch": 40.51258992805755,
+      "grad_norm": 0.27201807498931885,
+      "learning_rate": 1.6036697757095176e-05,
+      "loss": 0.0169,
+      "step": 45050
+    },
+    {
+      "action_loss": 0.0018646420212462544,
+      "epoch": 40.51258992805755,
+      "step": 45050
+    },
+    {
+      "epoch": 40.51258992805755,
+      "step": 45050,
+      "torque_loss": 0.09695889800786972
+    },
+    {
+      "epoch": 40.51258992805755,
+      "force_loss": 0.0022622295655310154,
+      "step": 45050
+    },
+    {
+      "epoch": 40.52158273381295,
+      "grad_norm": 0.15830875933170319,
+      "learning_rate": 1.6016478442264428e-05,
+      "loss": 0.0143,
+      "step": 45060
+    },
+    {
+      "action_loss": 0.0020414425525814295,
+      "epoch": 40.52158273381295,
+      "step": 45060
+    },
+    {
+      "epoch": 40.52158273381295,
+      "step": 45060,
+      "torque_loss": 0.10891855508089066
+    },
+    {
+      "epoch": 40.52158273381295,
+      "force_loss": 0.003519168123602867,
+      "step": 45060
+    },
+    {
+      "epoch": 40.530575539568346,
+      "grad_norm": 0.10168365389108658,
+      "learning_rate": 1.599626945073026e-05,
+      "loss": 0.0142,
+      "step": 45070
+    },
+    {
+      "action_loss": 0.0014845033874735236,
+      "epoch": 40.530575539568346,
+      "step": 45070
+    },
+    {
+      "epoch": 40.530575539568346,
+      "step": 45070,
+      "torque_loss": 0.11872855573892593
+    },
+    {
+      "epoch": 40.530575539568346,
+      "force_loss": 0.001696488237939775,
+      "step": 45070
+    },
+    {
+      "epoch": 40.539568345323744,
+      "grad_norm": 0.4593876302242279,
+      "learning_rate": 1.597607078863162e-05,
+      "loss": 0.0153,
+      "step": 45080
+    },
+    {
+      "action_loss": 0.001522569335065782,
+      "epoch": 40.539568345323744,
+      "step": 45080
+    },
+    {
+      "epoch": 40.539568345323744,
+      "step": 45080,
+      "torque_loss": 0.12280359119176865
+    },
+    {
+      "epoch": 40.539568345323744,
+      "force_loss": 0.005877872463315725,
+      "step": 45080
+    },
+    {
+      "epoch": 40.548561151079134,
+      "grad_norm": 0.08953931927680969,
+      "learning_rate": 1.595588246210432e-05,
+      "loss": 0.0155,
+      "step": 45090
+    },
+    {
+      "action_loss": 0.001429850235581398,
+      "epoch": 40.548561151079134,
+      "step": 45090
+    },
+    {
+      "epoch": 40.548561151079134,
+      "step": 45090,
+      "torque_loss": 0.11547132581472397
+    },
+    {
+      "epoch": 40.548561151079134,
+      "force_loss": 0.002205177675932646,
+      "step": 45090
+    },
+    {
+      "epoch": 40.55755395683453,
+      "grad_norm": 0.1410577893257141,
+      "learning_rate": 1.5935704477281048e-05,
+      "loss": 0.0147,
+      "step": 45100
+    },
+    {
+      "action_loss": 0.0018070373916998506,
+      "epoch": 40.55755395683453,
+      "step": 45100
+    },
+    {
+      "epoch": 40.55755395683453,
+      "step": 45100,
+      "torque_loss": 0.11407071352005005
+    },
+    {
+      "epoch": 40.55755395683453,
+      "force_loss": 0.0007607995066791773,
+      "step": 45100
+    },
+    {
+      "epoch": 40.56654676258993,
+      "grad_norm": 0.33400487899780273,
+      "learning_rate": 1.5915536840291323e-05,
+      "loss": 0.0145,
+      "step": 45110
+    },
+    {
+      "action_loss": 0.0018557592993602157,
+      "epoch": 40.56654676258993,
+      "step": 45110
+    },
+    {
+      "epoch": 40.56654676258993,
+      "step": 45110,
+      "torque_loss": 0.1627974957227707
+    },
+    {
+      "epoch": 40.56654676258993,
+      "force_loss": 0.0016722064465284348,
+      "step": 45110
+    },
+    {
+      "epoch": 40.57553956834532,
+      "grad_norm": 0.18177655339241028,
+      "learning_rate": 1.5895379557261576e-05,
+      "loss": 0.0153,
+      "step": 45120
+    },
+    {
+      "action_loss": 0.0011204140027984977,
+      "epoch": 40.57553956834532,
+      "step": 45120
+    },
+    {
+      "epoch": 40.57553956834532,
+      "step": 45120,
+      "torque_loss": 0.07101935893297195
+    },
+    {
+      "epoch": 40.57553956834532,
+      "force_loss": 0.0013590813614428043,
+      "step": 45120
+    },
+    {
+      "epoch": 40.58453237410072,
+      "grad_norm": 0.5464057922363281,
+      "learning_rate": 1.5875232634315033e-05,
+      "loss": 0.0146,
+      "step": 45130
+    },
+    {
+      "action_loss": 0.0014324383810162544,
+      "epoch": 40.58453237410072,
+      "step": 45130
+    },
+    {
+      "epoch": 40.58453237410072,
+      "step": 45130,
+      "torque_loss": 0.13846252858638763
+    },
+    {
+      "epoch": 40.58453237410072,
+      "force_loss": 0.0020315914880484343,
+      "step": 45130
+    },
+    {
+      "epoch": 40.593525179856115,
+      "grad_norm": 0.3913037180900574,
+      "learning_rate": 1.5855096077571812e-05,
+      "loss": 0.0168,
+      "step": 45140
+    },
+    {
+      "action_loss": 0.0036267004907131195,
+      "epoch": 40.593525179856115,
+      "step": 45140
+    },
+    {
+      "epoch": 40.593525179856115,
+      "step": 45140,
+      "torque_loss": 0.10154348611831665
+    },
+    {
+      "epoch": 40.593525179856115,
+      "force_loss": 0.003939494024962187,
+      "step": 45140
+    },
+    {
+      "epoch": 40.60251798561151,
+      "grad_norm": 0.19053564965724945,
+      "learning_rate": 1.5834969893148855e-05,
+      "loss": 0.0141,
+      "step": 45150
+    },
+    {
+      "action_loss": 0.0006660239887423813,
+      "epoch": 40.60251798561151,
+      "step": 45150
+    },
+    {
+      "epoch": 40.60251798561151,
+      "step": 45150,
+      "torque_loss": 0.13420797884464264
+    },
+    {
+      "epoch": 40.60251798561151,
+      "force_loss": 0.0005756028695032,
+      "step": 45150
+    },
+    {
+      "epoch": 40.611510791366904,
+      "grad_norm": 0.10361992567777634,
+      "learning_rate": 1.581485408715997e-05,
+      "loss": 0.0148,
+      "step": 45160
+    },
+    {
+      "action_loss": 0.0024870671331882477,
+      "epoch": 40.611510791366904,
+      "step": 45160
+    },
+    {
+      "epoch": 40.611510791366904,
+      "step": 45160,
+      "torque_loss": 0.1262078732252121
+    },
+    {
+      "epoch": 40.611510791366904,
+      "force_loss": 0.005188516806811094,
+      "step": 45160
+    },
+    {
+      "epoch": 40.6205035971223,
+      "grad_norm": 0.10443980991840363,
+      "learning_rate": 1.5794748665715785e-05,
+      "loss": 0.0149,
+      "step": 45170
+    },
+    {
+      "action_loss": 0.0007142227259464562,
+      "epoch": 40.6205035971223,
+      "step": 45170
+    },
+    {
+      "epoch": 40.6205035971223,
+      "step": 45170,
+      "torque_loss": 0.08687439560890198
+    },
+    {
+      "epoch": 40.6205035971223,
+      "force_loss": 0.0015860926359891891,
+      "step": 45170
+    },
+    {
+      "epoch": 40.6294964028777,
+      "grad_norm": 0.21977931261062622,
+      "learning_rate": 1.5774653634923857e-05,
+      "loss": 0.0137,
+      "step": 45180
+    },
+    {
+      "action_loss": 0.0007581103127449751,
+      "epoch": 40.6294964028777,
+      "step": 45180
+    },
+    {
+      "epoch": 40.6294964028777,
+      "step": 45180,
+      "torque_loss": 0.08276765793561935
+    },
+    {
+      "epoch": 40.6294964028777,
+      "force_loss": 0.0022683802526444197,
+      "step": 45180
+    },
+    {
+      "epoch": 40.638489208633096,
+      "grad_norm": 0.12322276085615158,
+      "learning_rate": 1.575456900088845e-05,
+      "loss": 0.0141,
+      "step": 45190
+    },
+    {
+      "action_loss": 0.016932303085923195,
+      "epoch": 40.638489208633096,
+      "step": 45190
+    },
+    {
+      "epoch": 40.638489208633096,
+      "step": 45190,
+      "torque_loss": 0.13495950400829315
+    },
+    {
+      "epoch": 40.638489208633096,
+      "force_loss": 0.012971772812306881,
+      "step": 45190
+    },
+    {
+      "epoch": 40.64748201438849,
+      "grad_norm": 0.07109040766954422,
+      "learning_rate": 1.5734494769710816e-05,
+      "loss": 0.0169,
+      "step": 45200
+    },
+    {
+      "action_loss": 0.001673296676017344,
+      "epoch": 40.64748201438849,
+      "step": 45200
+    },
+    {
+      "epoch": 40.64748201438849,
+      "step": 45200,
+      "torque_loss": 0.11823960393667221
+    },
+    {
+      "epoch": 40.64748201438849,
+      "force_loss": 0.0017986007733270526,
+      "step": 45200
+    },
+    {
+      "epoch": 40.656474820143885,
+      "grad_norm": 0.37495097517967224,
+      "learning_rate": 1.5714430947488912e-05,
+      "loss": 0.0145,
+      "step": 45210
+    },
+    {
+      "action_loss": 0.0019356566481292248,
+      "epoch": 40.656474820143885,
+      "step": 45210
+    },
+    {
+      "epoch": 40.656474820143885,
+      "step": 45210,
+      "torque_loss": 0.1061038076877594
+    },
+    {
+      "epoch": 40.656474820143885,
+      "force_loss": 0.006631011608988047,
+      "step": 45210
+    },
+    {
+      "epoch": 40.66546762589928,
+      "grad_norm": 0.3506937026977539,
+      "learning_rate": 1.5694377540317645e-05,
+      "loss": 0.0176,
+      "step": 45220
+    },
+    {
+      "action_loss": 0.0008853801409713924,
+      "epoch": 40.66546762589928,
+      "step": 45220
+    },
+    {
+      "epoch": 40.66546762589928,
+      "step": 45220,
+      "torque_loss": 0.10339949280023575
+    },
+    {
+      "epoch": 40.66546762589928,
+      "force_loss": 0.005537648219615221,
+      "step": 45220
+    },
+    {
+      "epoch": 40.67446043165468,
+      "grad_norm": 0.28692197799682617,
+      "learning_rate": 1.5674334554288694e-05,
+      "loss": 0.0159,
+      "step": 45230
+    },
+    {
+      "action_loss": 0.0023041001986712217,
+      "epoch": 40.67446043165468,
+      "step": 45230
+    },
+    {
+      "epoch": 40.67446043165468,
+      "step": 45230,
+      "torque_loss": 0.08491676300764084
+    },
+    {
+      "epoch": 40.67446043165468,
+      "force_loss": 0.0036495423410087824,
+      "step": 45230
+    },
+    {
+      "epoch": 40.68345323741007,
+      "grad_norm": 0.11309140920639038,
+      "learning_rate": 1.5654301995490582e-05,
+      "loss": 0.0151,
+      "step": 45240
+    },
+    {
+      "action_loss": 0.003722668858245015,
+      "epoch": 40.68345323741007,
+      "step": 45240
+    },
+    {
+      "epoch": 40.68345323741007,
+      "step": 45240,
+      "torque_loss": 0.16350679099559784
+    },
+    {
+      "epoch": 40.68345323741007,
+      "force_loss": 0.006182520184665918,
+      "step": 45240
+    },
+    {
+      "epoch": 40.69244604316547,
+      "grad_norm": 0.26787275075912476,
+      "learning_rate": 1.5634279870008685e-05,
+      "loss": 0.0144,
+      "step": 45250
+    },
+    {
+      "action_loss": 0.001883464865386486,
+      "epoch": 40.69244604316547,
+      "step": 45250
+    },
+    {
+      "epoch": 40.69244604316547,
+      "step": 45250,
+      "torque_loss": 0.1572028249502182
+    },
+    {
+      "epoch": 40.69244604316547,
+      "force_loss": 0.0016495849704369903,
+      "step": 45250
+    },
+    {
+      "epoch": 40.701438848920866,
+      "grad_norm": 0.0759255588054657,
+      "learning_rate": 1.5614268183925174e-05,
+      "loss": 0.014,
+      "step": 45260
+    },
+    {
+      "action_loss": 0.005943686235696077,
+      "epoch": 40.701438848920866,
+      "step": 45260
+    },
+    {
+      "epoch": 40.701438848920866,
+      "step": 45260,
+      "torque_loss": 0.082745261490345
+    },
+    {
+      "epoch": 40.701438848920866,
+      "force_loss": 0.006494285073131323,
+      "step": 45260
+    },
+    {
+      "epoch": 40.710431654676256,
+      "grad_norm": 0.12950587272644043,
+      "learning_rate": 1.5594266943319097e-05,
+      "loss": 0.0168,
+      "step": 45270
+    },
+    {
+      "action_loss": 0.002122250385582447,
+      "epoch": 40.710431654676256,
+      "step": 45270
+    },
+    {
+      "epoch": 40.710431654676256,
+      "step": 45270,
+      "torque_loss": 0.17228136956691742
+    },
+    {
+      "epoch": 40.710431654676256,
+      "force_loss": 0.0015712982276454568,
+      "step": 45270
+    },
+    {
+      "epoch": 40.719424460431654,
+      "grad_norm": 0.07140425592660904,
+      "learning_rate": 1.5574276154266294e-05,
+      "loss": 0.0147,
+      "step": 45280
+    },
+    {
+      "action_loss": 0.001982799032703042,
+      "epoch": 40.719424460431654,
+      "step": 45280
+    },
+    {
+      "epoch": 40.719424460431654,
+      "step": 45280,
+      "torque_loss": 0.11044812202453613
+    },
+    {
+      "epoch": 40.719424460431654,
+      "force_loss": 0.0032283940818160772,
+      "step": 45280
+    },
+    {
+      "epoch": 40.72841726618705,
+      "grad_norm": 0.07352296262979507,
+      "learning_rate": 1.5554295822839437e-05,
+      "loss": 0.0175,
+      "step": 45290
+    },
+    {
+      "action_loss": 0.0007347942446358502,
+      "epoch": 40.72841726618705,
+      "step": 45290
+    },
+    {
+      "epoch": 40.72841726618705,
+      "step": 45290,
+      "torque_loss": 0.08273150026798248
+    },
+    {
+      "epoch": 40.72841726618705,
+      "force_loss": 0.00250991084612906,
+      "step": 45290
+    },
+    {
+      "epoch": 40.73741007194245,
+      "grad_norm": 0.2720288336277008,
+      "learning_rate": 1.5534325955108025e-05,
+      "loss": 0.0149,
+      "step": 45300
+    },
+    {
+      "action_loss": 0.0036175574641674757,
+      "epoch": 40.73741007194245,
+      "step": 45300
+    },
+    {
+      "epoch": 40.73741007194245,
+      "step": 45300,
+      "torque_loss": 0.147342249751091
+    },
+    {
+      "epoch": 40.73741007194245,
+      "force_loss": 0.0033506855834275484,
+      "step": 45300
+    },
+    {
+      "epoch": 40.74640287769784,
+      "grad_norm": 0.10177946835756302,
+      "learning_rate": 1.5514366557138373e-05,
+      "loss": 0.0158,
+      "step": 45310
+    },
+    {
+      "action_loss": 0.0006860622670501471,
+      "epoch": 40.74640287769784,
+      "step": 45310
+    },
+    {
+      "epoch": 40.74640287769784,
+      "step": 45310,
+      "torque_loss": 0.13801836967468262
+    },
+    {
+      "epoch": 40.74640287769784,
+      "force_loss": 0.0010156398639082909,
+      "step": 45310
+    },
+    {
+      "epoch": 40.75539568345324,
+      "grad_norm": 0.11018849164247513,
+      "learning_rate": 1.5494417634993602e-05,
+      "loss": 0.0148,
+      "step": 45320
+    },
+    {
+      "action_loss": 0.0032422132790088654,
+      "epoch": 40.75539568345324,
+      "step": 45320
+    },
+    {
+      "epoch": 40.75539568345324,
+      "step": 45320,
+      "torque_loss": 0.1261252611875534
+    },
+    {
+      "epoch": 40.75539568345324,
+      "force_loss": 0.002237763023003936,
+      "step": 45320
+    },
+    {
+      "epoch": 40.764388489208635,
+      "grad_norm": 0.6076178550720215,
+      "learning_rate": 1.547447919473372e-05,
+      "loss": 0.0187,
+      "step": 45330
+    },
+    {
+      "action_loss": 0.0018011104548349977,
+      "epoch": 40.764388489208635,
+      "step": 45330
+    },
+    {
+      "epoch": 40.764388489208635,
+      "step": 45330,
+      "torque_loss": 0.10535530000925064
+    },
+    {
+      "epoch": 40.764388489208635,
+      "force_loss": 0.00834402535110712,
+      "step": 45330
+    },
+    {
+      "epoch": 40.773381294964025,
+      "grad_norm": 0.189901664853096,
+      "learning_rate": 1.5454551242415434e-05,
+      "loss": 0.0139,
+      "step": 45340
+    },
+    {
+      "action_loss": 0.0012888670898973942,
+      "epoch": 40.773381294964025,
+      "step": 45340
+    },
+    {
+      "epoch": 40.773381294964025,
+      "step": 45340,
+      "torque_loss": 0.19431431591510773
+    },
+    {
+      "epoch": 40.773381294964025,
+      "force_loss": 0.0008001910173334181,
+      "step": 45340
+    },
+    {
+      "epoch": 40.78237410071942,
+      "grad_norm": 0.1786688268184662,
+      "learning_rate": 1.543463378409239e-05,
+      "loss": 0.0141,
+      "step": 45350
+    },
+    {
+      "action_loss": 0.001953274942934513,
+      "epoch": 40.78237410071942,
+      "step": 45350
+    },
+    {
+      "epoch": 40.78237410071942,
+      "step": 45350,
+      "torque_loss": 0.10562432557344437
+    },
+    {
+      "epoch": 40.78237410071942,
+      "force_loss": 0.003258664859458804,
+      "step": 45350
+    },
+    {
+      "epoch": 40.79136690647482,
+      "grad_norm": 0.1174292042851448,
+      "learning_rate": 1.541472682581493e-05,
+      "loss": 0.0203,
+      "step": 45360
+    },
+    {
+      "action_loss": 0.0025638267397880554,
+      "epoch": 40.79136690647482,
+      "step": 45360
+    },
+    {
+      "epoch": 40.79136690647482,
+      "step": 45360,
+      "torque_loss": 0.1209598183631897
+    },
+    {
+      "epoch": 40.79136690647482,
+      "force_loss": 0.0020155783277004957,
+      "step": 45360
+    },
+    {
+      "epoch": 40.80035971223022,
+      "grad_norm": 0.15606212615966797,
+      "learning_rate": 1.5394830373630298e-05,
+      "loss": 0.0163,
+      "step": 45370
+    },
+    {
+      "action_loss": 0.004919996950775385,
+      "epoch": 40.80035971223022,
+      "step": 45370
+    },
+    {
+      "epoch": 40.80035971223022,
+      "step": 45370,
+      "torque_loss": 0.11587062478065491
+    },
+    {
+      "epoch": 40.80035971223022,
+      "force_loss": 0.004136399365961552,
+      "step": 45370
+    },
+    {
+      "epoch": 40.80935251798561,
+      "grad_norm": 0.27999019622802734,
+      "learning_rate": 1.5374944433582506e-05,
+      "loss": 0.0169,
+      "step": 45380
+    },
+    {
+      "action_loss": 0.0020243690814822912,
+      "epoch": 40.80935251798561,
+      "step": 45380
+    },
+    {
+      "epoch": 40.80935251798561,
+      "step": 45380,
+      "torque_loss": 0.1055983304977417
+    },
+    {
+      "epoch": 40.80935251798561,
+      "force_loss": 0.004501225892454386,
+      "step": 45380
+    },
+    {
+      "epoch": 40.818345323741006,
+      "grad_norm": 0.14808295667171478,
+      "learning_rate": 1.5355069011712375e-05,
+      "loss": 0.015,
+      "step": 45390
+    },
+    {
+      "action_loss": 0.006646020337939262,
+      "epoch": 40.818345323741006,
+      "step": 45390
+    },
+    {
+      "epoch": 40.818345323741006,
+      "step": 45390,
+      "torque_loss": 0.10003986209630966
+    },
+    {
+      "epoch": 40.818345323741006,
+      "force_loss": 0.017654258757829666,
+      "step": 45390
+    },
+    {
+      "epoch": 40.827338129496404,
+      "grad_norm": 0.3043540120124817,
+      "learning_rate": 1.5335204114057526e-05,
+      "loss": 0.0211,
+      "step": 45400
+    },
+    {
+      "action_loss": 0.001958661014214158,
+      "epoch": 40.827338129496404,
+      "step": 45400
+    },
+    {
+      "epoch": 40.827338129496404,
+      "step": 45400,
+      "torque_loss": 0.10480699688196182
+    },
+    {
+      "epoch": 40.827338129496404,
+      "force_loss": 0.0018539298325777054,
+      "step": 45400
+    },
+    {
+      "epoch": 40.8363309352518,
+      "grad_norm": 0.13785338401794434,
+      "learning_rate": 1.5315349746652387e-05,
+      "loss": 0.0154,
+      "step": 45410
+    },
+    {
+      "action_loss": 0.00223384122364223,
+      "epoch": 40.8363309352518,
+      "step": 45410
+    },
+    {
+      "epoch": 40.8363309352518,
+      "step": 45410,
+      "torque_loss": 0.13858433067798615
+    },
+    {
+      "epoch": 40.8363309352518,
+      "force_loss": 0.004815421532839537,
+      "step": 45410
+    },
+    {
+      "epoch": 40.84532374100719,
+      "grad_norm": 0.10613209009170532,
+      "learning_rate": 1.5295505915528212e-05,
+      "loss": 0.0159,
+      "step": 45420
+    },
+    {
+      "action_loss": 0.0032362292986363173,
+      "epoch": 40.84532374100719,
+      "step": 45420
+    },
+    {
+      "epoch": 40.84532374100719,
+      "step": 45420,
+      "torque_loss": 0.08778562396764755
+    },
+    {
+      "epoch": 40.84532374100719,
+      "force_loss": 0.008067467249929905,
+      "step": 45420
+    },
+    {
+      "epoch": 40.85431654676259,
+      "grad_norm": 0.10648372024297714,
+      "learning_rate": 1.5275672626713024e-05,
+      "loss": 0.0156,
+      "step": 45430
+    },
+    {
+      "action_loss": 0.0023236980196088552,
+      "epoch": 40.85431654676259,
+      "step": 45430
+    },
+    {
+      "epoch": 40.85431654676259,
+      "step": 45430,
+      "torque_loss": 0.13655267655849457
+    },
+    {
+      "epoch": 40.85431654676259,
+      "force_loss": 0.003679906716570258,
+      "step": 45430
+    },
+    {
+      "epoch": 40.86330935251799,
+      "grad_norm": 0.11320368200540543,
+      "learning_rate": 1.5255849886231643e-05,
+      "loss": 0.0166,
+      "step": 45440
+    },
+    {
+      "action_loss": 0.0014845350524410605,
+      "epoch": 40.86330935251799,
+      "step": 45440
+    },
+    {
+      "epoch": 40.86330935251799,
+      "step": 45440,
+      "torque_loss": 0.14364412426948547
+    },
+    {
+      "epoch": 40.86330935251799,
+      "force_loss": 0.004346174653619528,
+      "step": 45440
+    },
+    {
+      "epoch": 40.87230215827338,
+      "grad_norm": 0.15395492315292358,
+      "learning_rate": 1.523603770010571e-05,
+      "loss": 0.0183,
+      "step": 45450
+    },
+    {
+      "action_loss": 0.0007283451850526035,
+      "epoch": 40.87230215827338,
+      "step": 45450
+    },
+    {
+      "epoch": 40.87230215827338,
+      "step": 45450,
+      "torque_loss": 0.12713085114955902
+    },
+    {
+      "epoch": 40.87230215827338,
+      "force_loss": 0.0021612446289509535,
+      "step": 45450
+    },
+    {
+      "epoch": 40.881294964028775,
+      "grad_norm": 0.22847002744674683,
+      "learning_rate": 1.521623607435363e-05,
+      "loss": 0.0157,
+      "step": 45460
+    },
+    {
+      "action_loss": 0.008981908671557903,
+      "epoch": 40.881294964028775,
+      "step": 45460
+    },
+    {
+      "epoch": 40.881294964028775,
+      "step": 45460,
+      "torque_loss": 0.16054268181324005
+    },
+    {
+      "epoch": 40.881294964028775,
+      "force_loss": 0.013512018136680126,
+      "step": 45460
+    },
+    {
+      "epoch": 40.89028776978417,
+      "grad_norm": 0.21227699518203735,
+      "learning_rate": 1.5196445014990612e-05,
+      "loss": 0.0195,
+      "step": 45470
+    },
+    {
+      "action_loss": 0.008794314227998257,
+      "epoch": 40.89028776978417,
+      "step": 45470
+    },
+    {
+      "epoch": 40.89028776978417,
+      "step": 45470,
+      "torque_loss": 0.103215791285038
+    },
+    {
+      "epoch": 40.89028776978417,
+      "force_loss": 0.011172245256602764,
+      "step": 45470
+    },
+    {
+      "epoch": 40.89928057553957,
+      "grad_norm": 0.07943612337112427,
+      "learning_rate": 1.5176664528028672e-05,
+      "loss": 0.0148,
+      "step": 45480
+    },
+    {
+      "action_loss": 0.001446147565729916,
+      "epoch": 40.89928057553957,
+      "step": 45480
+    },
+    {
+      "epoch": 40.89928057553957,
+      "step": 45480,
+      "torque_loss": 0.14647291600704193
+    },
+    {
+      "epoch": 40.89928057553957,
+      "force_loss": 0.0014994280645623803,
+      "step": 45480
+    },
+    {
+      "epoch": 40.90827338129496,
+      "grad_norm": 0.08619990199804306,
+      "learning_rate": 1.5156894619476574e-05,
+      "loss": 0.0159,
+      "step": 45490
+    },
+    {
+      "action_loss": 0.003168029710650444,
+      "epoch": 40.90827338129496,
+      "step": 45490
+    },
+    {
+      "epoch": 40.90827338129496,
+      "step": 45490,
+      "torque_loss": 0.15680629014968872
+    },
+    {
+      "epoch": 40.90827338129496,
+      "force_loss": 0.01152495015412569,
+      "step": 45490
+    },
+    {
+      "epoch": 40.91726618705036,
+      "grad_norm": 0.18403546512126923,
+      "learning_rate": 1.5137135295339938e-05,
+      "loss": 0.0172,
+      "step": 45500
+    },
+    {
+      "action_loss": 0.0038366764783859253,
+      "epoch": 40.91726618705036,
+      "step": 45500
+    },
+    {
+      "epoch": 40.91726618705036,
+      "step": 45500,
+      "torque_loss": 0.12737993896007538
+    },
+    {
+      "epoch": 40.91726618705036,
+      "force_loss": 0.003249027533456683,
+      "step": 45500
+    },
+    {
+      "epoch": 40.92625899280576,
+      "grad_norm": 0.13323071599006653,
+      "learning_rate": 1.5117386561621073e-05,
+      "loss": 0.0159,
+      "step": 45510
+    },
+    {
+      "action_loss": 0.002135268412530422,
+      "epoch": 40.92625899280576,
+      "step": 45510
+    },
+    {
+      "epoch": 40.92625899280576,
+      "step": 45510,
+      "torque_loss": 0.0728834941983223
+    },
+    {
+      "epoch": 40.92625899280576,
+      "force_loss": 0.003202298656105995,
+      "step": 45510
+    },
+    {
+      "epoch": 40.935251798561154,
+      "grad_norm": 0.2703399658203125,
+      "learning_rate": 1.5097648424319167e-05,
+      "loss": 0.0159,
+      "step": 45520
+    },
+    {
+      "action_loss": 0.004016268998384476,
+      "epoch": 40.935251798561154,
+      "step": 45520
+    },
+    {
+      "epoch": 40.935251798561154,
+      "step": 45520,
+      "torque_loss": 0.0884188786149025
+    },
+    {
+      "epoch": 40.935251798561154,
+      "force_loss": 0.012089270167052746,
+      "step": 45520
+    },
+    {
+      "epoch": 40.944244604316545,
+      "grad_norm": 0.05581743270158768,
+      "learning_rate": 1.5077920889430119e-05,
+      "loss": 0.0152,
+      "step": 45530
+    },
+    {
+      "action_loss": 0.0008043153211474419,
+      "epoch": 40.944244604316545,
+      "step": 45530
+    },
+    {
+      "epoch": 40.944244604316545,
+      "step": 45530,
+      "torque_loss": 0.1398676186800003
+    },
+    {
+      "epoch": 40.944244604316545,
+      "force_loss": 0.0019333800300955772,
+      "step": 45530
+    },
+    {
+      "epoch": 40.95323741007194,
+      "grad_norm": 0.11849424242973328,
+      "learning_rate": 1.5058203962946644e-05,
+      "loss": 0.0176,
+      "step": 45540
+    },
+    {
+      "action_loss": 0.0020483946427702904,
+      "epoch": 40.95323741007194,
+      "step": 45540
+    },
+    {
+      "epoch": 40.95323741007194,
+      "step": 45540,
+      "torque_loss": 0.13761721551418304
+    },
+    {
+      "epoch": 40.95323741007194,
+      "force_loss": 0.0018013231456279755,
+      "step": 45540
+    },
+    {
+      "epoch": 40.96223021582734,
+      "grad_norm": 0.16001176834106445,
+      "learning_rate": 1.503849765085822e-05,
+      "loss": 0.0163,
+      "step": 45550
+    },
+    {
+      "action_loss": 0.0009472144884057343,
+      "epoch": 40.96223021582734,
+      "step": 45550
+    },
+    {
+      "epoch": 40.96223021582734,
+      "step": 45550,
+      "torque_loss": 0.13335828483104706
+    },
+    {
+      "epoch": 40.96223021582734,
+      "force_loss": 0.0007911059074103832,
+      "step": 45550
+    },
+    {
+      "epoch": 40.97122302158273,
+      "grad_norm": 0.1334599256515503,
+      "learning_rate": 1.501880195915109e-05,
+      "loss": 0.0166,
+      "step": 45560
+    },
+    {
+      "action_loss": 0.012951535172760487,
+      "epoch": 40.97122302158273,
+      "step": 45560
+    },
+    {
+      "epoch": 40.97122302158273,
+      "step": 45560,
+      "torque_loss": 0.12569712102413177
+    },
+    {
+      "epoch": 40.97122302158273,
+      "force_loss": 0.01916349306702614,
+      "step": 45560
+    },
+    {
+      "epoch": 40.98021582733813,
+      "grad_norm": 0.22745516896247864,
+      "learning_rate": 1.499911689380833e-05,
+      "loss": 0.0212,
+      "step": 45570
+    },
+    {
+      "action_loss": 0.003298509633168578,
+      "epoch": 40.98021582733813,
+      "step": 45570
+    },
+    {
+      "epoch": 40.98021582733813,
+      "step": 45570,
+      "torque_loss": 0.12431459873914719
+    },
+    {
+      "epoch": 40.98021582733813,
+      "force_loss": 0.0026190143544226885,
+      "step": 45570
+    },
+    {
+      "epoch": 40.989208633093526,
+      "grad_norm": 0.13201363384723663,
+      "learning_rate": 1.4979442460809683e-05,
+      "loss": 0.0147,
+      "step": 45580
+    },
+    {
+      "action_loss": 0.0020275565329939127,
+      "epoch": 40.989208633093526,
+      "step": 45580
+    },
+    {
+      "epoch": 40.989208633093526,
+      "step": 45580,
+      "torque_loss": 0.10227114707231522
+    },
+    {
+      "epoch": 40.989208633093526,
+      "force_loss": 0.0012670475989580154,
+      "step": 45580
+    },
+    {
+      "epoch": 40.99820143884892,
+      "grad_norm": 0.07656856626272202,
+      "learning_rate": 1.4959778666131763e-05,
+      "loss": 0.0169,
+      "step": 45590
+    },
+    {
+      "action_loss": 0.004446713253855705,
+      "epoch": 40.99820143884892,
+      "step": 45590
+    },
+    {
+      "epoch": 40.99820143884892,
+      "step": 45590,
+      "torque_loss": 0.17414820194244385
+    },
+    {
+      "epoch": 40.99820143884892,
+      "force_loss": 0.0033637837041169405,
+      "step": 45590
+    },
+    {
+      "epoch": 41.007194244604314,
+      "grad_norm": 0.05595703423023224,
+      "learning_rate": 1.4940125515747905e-05,
+      "loss": 0.0143,
+      "step": 45600
+    },
+    {
+      "action_loss": 0.0009173061698675156,
+      "epoch": 41.007194244604314,
+      "step": 45600
+    },
+    {
+      "epoch": 41.007194244604314,
+      "step": 45600,
+      "torque_loss": 0.13735492527484894
+    },
+    {
+      "epoch": 41.007194244604314,
+      "force_loss": 0.0036803781986236572,
+      "step": 45600
+    },
+    {
+      "epoch": 41.01618705035971,
+      "grad_norm": 0.07746853679418564,
+      "learning_rate": 1.4920483015628211e-05,
+      "loss": 0.0195,
+      "step": 45610
+    },
+    {
+      "action_loss": 0.0012599130859598517,
+      "epoch": 41.01618705035971,
+      "step": 45610
+    },
+    {
+      "epoch": 41.01618705035971,
+      "step": 45610,
+      "torque_loss": 0.11418485641479492
+    },
+    {
+      "epoch": 41.01618705035971,
+      "force_loss": 0.002129557076841593,
+      "step": 45610
+    },
+    {
+      "epoch": 41.02517985611511,
+      "grad_norm": 0.138280987739563,
+      "learning_rate": 1.490085117173956e-05,
+      "loss": 0.0172,
+      "step": 45620
+    },
+    {
+      "action_loss": 0.0014922342961654067,
+      "epoch": 41.02517985611511,
+      "step": 45620
+    },
+    {
+      "epoch": 41.02517985611511,
+      "step": 45620,
+      "torque_loss": 0.12018273025751114
+    },
+    {
+      "epoch": 41.02517985611511,
+      "force_loss": 0.0010604195995256305,
+      "step": 45620
+    },
+    {
+      "epoch": 41.03417266187051,
+      "grad_norm": 0.15245187282562256,
+      "learning_rate": 1.488122999004558e-05,
+      "loss": 0.0158,
+      "step": 45630
+    },
+    {
+      "action_loss": 0.0030966622289270163,
+      "epoch": 41.03417266187051,
+      "step": 45630
+    },
+    {
+      "epoch": 41.03417266187051,
+      "step": 45630,
+      "torque_loss": 0.15085382759571075
+    },
+    {
+      "epoch": 41.03417266187051,
+      "force_loss": 0.0018634916050359607,
+      "step": 45630
+    },
+    {
+      "epoch": 41.0431654676259,
+      "grad_norm": 0.15233322978019714,
+      "learning_rate": 1.486161947650666e-05,
+      "loss": 0.017,
+      "step": 45640
+    },
+    {
+      "action_loss": 0.0010055876336991787,
+      "epoch": 41.0431654676259,
+      "step": 45640
+    },
+    {
+      "epoch": 41.0431654676259,
+      "step": 45640,
+      "torque_loss": 0.11032122373580933
+    },
+    {
+      "epoch": 41.0431654676259,
+      "force_loss": 0.0014110886259004474,
+      "step": 45640
+    },
+    {
+      "epoch": 41.052158273381295,
+      "grad_norm": 0.21772274374961853,
+      "learning_rate": 1.4842019637079995e-05,
+      "loss": 0.0145,
+      "step": 45650
+    },
+    {
+      "action_loss": 0.0009151121485047042,
+      "epoch": 41.052158273381295,
+      "step": 45650
+    },
+    {
+      "epoch": 41.052158273381295,
+      "step": 45650,
+      "torque_loss": 0.11252357810735703
+    },
+    {
+      "epoch": 41.052158273381295,
+      "force_loss": 0.0015598125755786896,
+      "step": 45650
+    },
+    {
+      "epoch": 41.06115107913669,
+      "grad_norm": 0.3965555727481842,
+      "learning_rate": 1.482243047771944e-05,
+      "loss": 0.0156,
+      "step": 45660
+    },
+    {
+      "action_loss": 0.0006716644275002182,
+      "epoch": 41.06115107913669,
+      "step": 45660
+    },
+    {
+      "epoch": 41.06115107913669,
+      "step": 45660,
+      "torque_loss": 0.12471433728933334
+    },
+    {
+      "epoch": 41.06115107913669,
+      "force_loss": 0.0005602568271569908,
+      "step": 45660
+    },
+    {
+      "epoch": 41.07014388489208,
+      "grad_norm": 0.119111567735672,
+      "learning_rate": 1.4802852004375712e-05,
+      "loss": 0.0161,
+      "step": 45670
+    },
+    {
+      "action_loss": 0.0006047275965102017,
+      "epoch": 41.07014388489208,
+      "step": 45670
+    },
+    {
+      "epoch": 41.07014388489208,
+      "step": 45670,
+      "torque_loss": 0.08508279919624329
+    },
+    {
+      "epoch": 41.07014388489208,
+      "force_loss": 0.002354872180148959,
+      "step": 45670
+    },
+    {
+      "epoch": 41.07913669064748,
+      "grad_norm": 0.07040691375732422,
+      "learning_rate": 1.4783284222996218e-05,
+      "loss": 0.0164,
+      "step": 45680
+    },
+    {
+      "action_loss": 0.0008050825563259423,
+      "epoch": 41.07913669064748,
+      "step": 45680
+    },
+    {
+      "epoch": 41.07913669064748,
+      "step": 45680,
+      "torque_loss": 0.10329881310462952
+    },
+    {
+      "epoch": 41.07913669064748,
+      "force_loss": 0.001120261731557548,
+      "step": 45680
+    },
+    {
+      "epoch": 41.08812949640288,
+      "grad_norm": 0.052260711789131165,
+      "learning_rate": 1.4763727139525135e-05,
+      "loss": 0.0139,
+      "step": 45690
+    },
+    {
+      "action_loss": 0.0048795039765536785,
+      "epoch": 41.08812949640288,
+      "step": 45690
+    },
+    {
+      "epoch": 41.08812949640288,
+      "step": 45690,
+      "torque_loss": 0.12040800601243973
+    },
+    {
+      "epoch": 41.08812949640288,
+      "force_loss": 0.01828545518219471,
+      "step": 45690
+    },
+    {
+      "epoch": 41.097122302158276,
+      "grad_norm": 0.0828484520316124,
+      "learning_rate": 1.4744180759903392e-05,
+      "loss": 0.0154,
+      "step": 45700
+    },
+    {
+      "action_loss": 0.0006536262226291001,
+      "epoch": 41.097122302158276,
+      "step": 45700
+    },
+    {
+      "epoch": 41.097122302158276,
+      "step": 45700,
+      "torque_loss": 0.10166070610284805
+    },
+    {
+      "epoch": 41.097122302158276,
+      "force_loss": 0.0007730152574367821,
+      "step": 45700
+    },
+    {
+      "epoch": 41.106115107913666,
+      "grad_norm": 0.49192941188812256,
+      "learning_rate": 1.4724645090068635e-05,
+      "loss": 0.0172,
+      "step": 45710
+    },
+    {
+      "action_loss": 0.0007073748856782913,
+      "epoch": 41.106115107913666,
+      "step": 45710
+    },
+    {
+      "epoch": 41.106115107913666,
+      "step": 45710,
+      "torque_loss": 0.12907542288303375
+    },
+    {
+      "epoch": 41.106115107913666,
+      "force_loss": 0.0007066354155540466,
+      "step": 45710
+    },
+    {
+      "epoch": 41.115107913669064,
+      "grad_norm": 0.0751081332564354,
+      "learning_rate": 1.4705120135955341e-05,
+      "loss": 0.0148,
+      "step": 45720
+    },
+    {
+      "action_loss": 0.0018327077850699425,
+      "epoch": 41.115107913669064,
+      "step": 45720
+    },
+    {
+      "epoch": 41.115107913669064,
+      "step": 45720,
+      "torque_loss": 0.1009114608168602
+    },
+    {
+      "epoch": 41.115107913669064,
+      "force_loss": 0.007172440644353628,
+      "step": 45720
+    },
+    {
+      "epoch": 41.12410071942446,
+      "grad_norm": 0.05827561765909195,
+      "learning_rate": 1.4685605903494614e-05,
+      "loss": 0.014,
+      "step": 45730
+    },
+    {
+      "action_loss": 0.004469592124223709,
+      "epoch": 41.12410071942446,
+      "step": 45730
+    },
+    {
+      "epoch": 41.12410071942446,
+      "step": 45730,
+      "torque_loss": 0.11639633029699326
+    },
+    {
+      "epoch": 41.12410071942446,
+      "force_loss": 0.004198806826025248,
+      "step": 45730
+    },
+    {
+      "epoch": 41.13309352517986,
+      "grad_norm": 0.4126000702381134,
+      "learning_rate": 1.46661023986144e-05,
+      "loss": 0.0142,
+      "step": 45740
+    },
+    {
+      "action_loss": 0.003094307379797101,
+      "epoch": 41.13309352517986,
+      "step": 45740
+    },
+    {
+      "epoch": 41.13309352517986,
+      "step": 45740,
+      "torque_loss": 0.13494877517223358
+    },
+    {
+      "epoch": 41.13309352517986,
+      "force_loss": 0.0034474774729460478,
+      "step": 45740
+    },
+    {
+      "epoch": 41.14208633093525,
+      "grad_norm": 0.07729417085647583,
+      "learning_rate": 1.4646609627239344e-05,
+      "loss": 0.0151,
+      "step": 45750
+    },
+    {
+      "action_loss": 0.005464589688926935,
+      "epoch": 41.14208633093525,
+      "step": 45750
+    },
+    {
+      "epoch": 41.14208633093525,
+      "step": 45750,
+      "torque_loss": 0.11652740091085434
+    },
+    {
+      "epoch": 41.14208633093525,
+      "force_loss": 0.0057957409881055355,
+      "step": 45750
+    },
+    {
+      "epoch": 41.15107913669065,
+      "grad_norm": 0.0866890475153923,
+      "learning_rate": 1.4627127595290835e-05,
+      "loss": 0.0144,
+      "step": 45760
+    },
+    {
+      "action_loss": 0.0029875747859477997,
+      "epoch": 41.15107913669065,
+      "step": 45760
+    },
+    {
+      "epoch": 41.15107913669065,
+      "step": 45760,
+      "torque_loss": 0.12687520682811737
+    },
+    {
+      "epoch": 41.15107913669065,
+      "force_loss": 0.008059307001531124,
+      "step": 45760
+    },
+    {
+      "epoch": 41.160071942446045,
+      "grad_norm": 0.2610211968421936,
+      "learning_rate": 1.460765630868699e-05,
+      "loss": 0.0151,
+      "step": 45770
+    },
+    {
+      "action_loss": 0.0010249275946989655,
+      "epoch": 41.160071942446045,
+      "step": 45770
+    },
+    {
+      "epoch": 41.160071942446045,
+      "step": 45770,
+      "torque_loss": 0.15003173053264618
+    },
+    {
+      "epoch": 41.160071942446045,
+      "force_loss": 0.0017228905344381928,
+      "step": 45770
+    },
+    {
+      "epoch": 41.169064748201436,
+      "grad_norm": 0.13967429101467133,
+      "learning_rate": 1.4588195773342678e-05,
+      "loss": 0.0196,
+      "step": 45780
+    },
+    {
+      "action_loss": 0.0013811002718284726,
+      "epoch": 41.169064748201436,
+      "step": 45780
+    },
+    {
+      "epoch": 41.169064748201436,
+      "step": 45780,
+      "torque_loss": 0.1370999664068222
+    },
+    {
+      "epoch": 41.169064748201436,
+      "force_loss": 0.003218189114704728,
+      "step": 45780
+    },
+    {
+      "epoch": 41.17805755395683,
+      "grad_norm": 0.33587485551834106,
+      "learning_rate": 1.4568745995169485e-05,
+      "loss": 0.0145,
+      "step": 45790
+    },
+    {
+      "action_loss": 0.0018223609076812863,
+      "epoch": 41.17805755395683,
+      "step": 45790
+    },
+    {
+      "epoch": 41.17805755395683,
+      "step": 45790,
+      "torque_loss": 0.08675705641508102
+    },
+    {
+      "epoch": 41.17805755395683,
+      "force_loss": 0.005780382547527552,
+      "step": 45790
+    },
+    {
+      "epoch": 41.18705035971223,
+      "grad_norm": 0.14802373945713043,
+      "learning_rate": 1.4549306980075778e-05,
+      "loss": 0.0144,
+      "step": 45800
+    },
+    {
+      "action_loss": 0.0024017065297812223,
+      "epoch": 41.18705035971223,
+      "step": 45800
+    },
+    {
+      "epoch": 41.18705035971223,
+      "step": 45800,
+      "torque_loss": 0.14377476274967194
+    },
+    {
+      "epoch": 41.18705035971223,
+      "force_loss": 0.004742899909615517,
+      "step": 45800
+    },
+    {
+      "epoch": 41.19604316546763,
+      "grad_norm": 0.11063234508037567,
+      "learning_rate": 1.4529878733966557e-05,
+      "loss": 0.0156,
+      "step": 45810
+    },
+    {
+      "action_loss": 0.002027997048571706,
+      "epoch": 41.19604316546763,
+      "step": 45810
+    },
+    {
+      "epoch": 41.19604316546763,
+      "step": 45810,
+      "torque_loss": 0.08225687593221664
+    },
+    {
+      "epoch": 41.19604316546763,
+      "force_loss": 0.0023245718330144882,
+      "step": 45810
+    },
+    {
+      "epoch": 41.20503597122302,
+      "grad_norm": 0.11345665156841278,
+      "learning_rate": 1.4510461262743658e-05,
+      "loss": 0.0182,
+      "step": 45820
+    },
+    {
+      "action_loss": 0.002794221742078662,
+      "epoch": 41.20503597122302,
+      "step": 45820
+    },
+    {
+      "epoch": 41.20503597122302,
+      "step": 45820,
+      "torque_loss": 0.11140748113393784
+    },
+    {
+      "epoch": 41.20503597122302,
+      "force_loss": 0.0028088155668228865,
+      "step": 45820
+    },
+    {
+      "epoch": 41.21402877697842,
+      "grad_norm": 0.5280176997184753,
+      "learning_rate": 1.4491054572305585e-05,
+      "loss": 0.0177,
+      "step": 45830
+    },
+    {
+      "action_loss": 0.002045967150479555,
+      "epoch": 41.21402877697842,
+      "step": 45830
+    },
+    {
+      "epoch": 41.21402877697842,
+      "step": 45830,
+      "torque_loss": 0.17886781692504883
+    },
+    {
+      "epoch": 41.21402877697842,
+      "force_loss": 0.008194251917302608,
+      "step": 45830
+    },
+    {
+      "epoch": 41.223021582733814,
+      "grad_norm": 0.28480222821235657,
+      "learning_rate": 1.4471658668547566e-05,
+      "loss": 0.016,
+      "step": 45840
+    },
+    {
+      "action_loss": 0.003977763932198286,
+      "epoch": 41.223021582733814,
+      "step": 45840
+    },
+    {
+      "epoch": 41.223021582733814,
+      "step": 45840,
+      "torque_loss": 0.11935213953256607
+    },
+    {
+      "epoch": 41.223021582733814,
+      "force_loss": 0.009474127553403378,
+      "step": 45840
+    },
+    {
+      "epoch": 41.23201438848921,
+      "grad_norm": 0.4027993977069855,
+      "learning_rate": 1.4452273557361579e-05,
+      "loss": 0.016,
+      "step": 45850
+    },
+    {
+      "action_loss": 0.0012138989986851811,
+      "epoch": 41.23201438848921,
+      "step": 45850
+    },
+    {
+      "epoch": 41.23201438848921,
+      "step": 45850,
+      "torque_loss": 0.13976942002773285
+    },
+    {
+      "epoch": 41.23201438848921,
+      "force_loss": 0.0014497647061944008,
+      "step": 45850
+    },
+    {
+      "epoch": 41.2410071942446,
+      "grad_norm": 0.37084975838661194,
+      "learning_rate": 1.4432899244636282e-05,
+      "loss": 0.0147,
+      "step": 45860
+    },
+    {
+      "action_loss": 0.0019848018418997526,
+      "epoch": 41.2410071942446,
+      "step": 45860
+    },
+    {
+      "epoch": 41.2410071942446,
+      "step": 45860,
+      "torque_loss": 0.10346949100494385
+    },
+    {
+      "epoch": 41.2410071942446,
+      "force_loss": 0.006480767857283354,
+      "step": 45860
+    },
+    {
+      "epoch": 41.25,
+      "grad_norm": 0.07505764812231064,
+      "learning_rate": 1.4413535736257134e-05,
+      "loss": 0.0141,
+      "step": 45870
+    },
+    {
+      "action_loss": 0.0006394576630555093,
+      "epoch": 41.25,
+      "step": 45870
+    },
+    {
+      "epoch": 41.25,
+      "step": 45870,
+      "torque_loss": 0.1338426172733307
+    },
+    {
+      "epoch": 41.25,
+      "force_loss": 0.0005767183029092848,
+      "step": 45870
+    },
+    {
+      "epoch": 41.2589928057554,
+      "grad_norm": 0.14940722286701202,
+      "learning_rate": 1.439418303810619e-05,
+      "loss": 0.0132,
+      "step": 45880
+    },
+    {
+      "action_loss": 0.0008269302197732031,
+      "epoch": 41.2589928057554,
+      "step": 45880
+    },
+    {
+      "epoch": 41.2589928057554,
+      "step": 45880,
+      "torque_loss": 0.09742355346679688
+    },
+    {
+      "epoch": 41.2589928057554,
+      "force_loss": 0.0026347816456109285,
+      "step": 45880
+    },
+    {
+      "epoch": 41.26798561151079,
+      "grad_norm": 0.35172533988952637,
+      "learning_rate": 1.4374841156062352e-05,
+      "loss": 0.0156,
+      "step": 45890
+    },
+    {
+      "action_loss": 0.0020411249715834856,
+      "epoch": 41.26798561151079,
+      "step": 45890
+    },
+    {
+      "epoch": 41.26798561151079,
+      "step": 45890,
+      "torque_loss": 0.12388905882835388
+    },
+    {
+      "epoch": 41.26798561151079,
+      "force_loss": 0.004184046760201454,
+      "step": 45890
+    },
+    {
+      "epoch": 41.276978417266186,
+      "grad_norm": 0.19972166419029236,
+      "learning_rate": 1.4355510096001112e-05,
+      "loss": 0.0141,
+      "step": 45900
+    },
+    {
+      "action_loss": 0.0008028179290704429,
+      "epoch": 41.276978417266186,
+      "step": 45900
+    },
+    {
+      "epoch": 41.276978417266186,
+      "step": 45900,
+      "torque_loss": 0.11227947473526001
+    },
+    {
+      "epoch": 41.276978417266186,
+      "force_loss": 0.000998382456600666,
+      "step": 45900
+    },
+    {
+      "epoch": 41.28597122302158,
+      "grad_norm": 0.13529066741466522,
+      "learning_rate": 1.4336189863794786e-05,
+      "loss": 0.0151,
+      "step": 45910
+    },
+    {
+      "action_loss": 0.005704980343580246,
+      "epoch": 41.28597122302158,
+      "step": 45910
+    },
+    {
+      "epoch": 41.28597122302158,
+      "step": 45910,
+      "torque_loss": 0.15707606077194214
+    },
+    {
+      "epoch": 41.28597122302158,
+      "force_loss": 0.010159828700125217,
+      "step": 45910
+    },
+    {
+      "epoch": 41.29496402877698,
+      "grad_norm": 0.06918837130069733,
+      "learning_rate": 1.4316880465312327e-05,
+      "loss": 0.0152,
+      "step": 45920
+    },
+    {
+      "action_loss": 0.0011797536863014102,
+      "epoch": 41.29496402877698,
+      "step": 45920
+    },
+    {
+      "epoch": 41.29496402877698,
+      "step": 45920,
+      "torque_loss": 0.12955911457538605
+    },
+    {
+      "epoch": 41.29496402877698,
+      "force_loss": 0.0008417136850766838,
+      "step": 45920
+    },
+    {
+      "epoch": 41.30395683453237,
+      "grad_norm": 0.19777178764343262,
+      "learning_rate": 1.4297581906419426e-05,
+      "loss": 0.0138,
+      "step": 45930
+    },
+    {
+      "action_loss": 0.001145684625953436,
+      "epoch": 41.30395683453237,
+      "step": 45930
+    },
+    {
+      "epoch": 41.30395683453237,
+      "step": 45930,
+      "torque_loss": 0.20803320407867432
+    },
+    {
+      "epoch": 41.30395683453237,
+      "force_loss": 0.004739269148558378,
+      "step": 45930
+    },
+    {
+      "epoch": 41.31294964028777,
+      "grad_norm": 0.08228175342082977,
+      "learning_rate": 1.4278294192978475e-05,
+      "loss": 0.0145,
+      "step": 45940
+    },
+    {
+      "action_loss": 0.00227418658323586,
+      "epoch": 41.31294964028777,
+      "step": 45940
+    },
+    {
+      "epoch": 41.31294964028777,
+      "step": 45940,
+      "torque_loss": 0.12369390577077866
+    },
+    {
+      "epoch": 41.31294964028777,
+      "force_loss": 0.00150086369831115,
+      "step": 45940
+    },
+    {
+      "epoch": 41.32194244604317,
+      "grad_norm": 0.43653446435928345,
+      "learning_rate": 1.4259017330848574e-05,
+      "loss": 0.019,
+      "step": 45950
+    },
+    {
+      "action_loss": 0.0018179056933149695,
+      "epoch": 41.32194244604317,
+      "step": 45950
+    },
+    {
+      "epoch": 41.32194244604317,
+      "step": 45950,
+      "torque_loss": 0.16293329000473022
+    },
+    {
+      "epoch": 41.32194244604317,
+      "force_loss": 0.004935581702739,
+      "step": 45950
+    },
+    {
+      "epoch": 41.330935251798564,
+      "grad_norm": 0.09548771381378174,
+      "learning_rate": 1.4239751325885498e-05,
+      "loss": 0.0163,
+      "step": 45960
+    },
+    {
+      "action_loss": 0.0013453805586323142,
+      "epoch": 41.330935251798564,
+      "step": 45960
+    },
+    {
+      "epoch": 41.330935251798564,
+      "step": 45960,
+      "torque_loss": 0.10989055037498474
+    },
+    {
+      "epoch": 41.330935251798564,
+      "force_loss": 0.001306669320911169,
+      "step": 45960
+    },
+    {
+      "epoch": 41.339928057553955,
+      "grad_norm": 0.07037942856550217,
+      "learning_rate": 1.4220496183941795e-05,
+      "loss": 0.0145,
+      "step": 45970
+    },
+    {
+      "action_loss": 0.006145510822534561,
+      "epoch": 41.339928057553955,
+      "step": 45970
+    },
+    {
+      "epoch": 41.339928057553955,
+      "step": 45970,
+      "torque_loss": 0.11592301726341248
+    },
+    {
+      "epoch": 41.339928057553955,
+      "force_loss": 0.01623745821416378,
+      "step": 45970
+    },
+    {
+      "epoch": 41.34892086330935,
+      "grad_norm": 0.16258494555950165,
+      "learning_rate": 1.4201251910866648e-05,
+      "loss": 0.0152,
+      "step": 45980
+    },
+    {
+      "action_loss": 0.001714748446829617,
+      "epoch": 41.34892086330935,
+      "step": 45980
+    },
+    {
+      "epoch": 41.34892086330935,
+      "step": 45980,
+      "torque_loss": 0.10614802688360214
+    },
+    {
+      "epoch": 41.34892086330935,
+      "force_loss": 0.007494939956814051,
+      "step": 45980
+    },
+    {
+      "epoch": 41.35791366906475,
+      "grad_norm": 0.13360010087490082,
+      "learning_rate": 1.4182018512505957e-05,
+      "loss": 0.0141,
+      "step": 45990
+    },
+    {
+      "action_loss": 0.0036005883011966944,
+      "epoch": 41.35791366906475,
+      "step": 45990
+    },
+    {
+      "epoch": 41.35791366906475,
+      "step": 45990,
+      "torque_loss": 0.13439002633094788
+    },
+    {
+      "epoch": 41.35791366906475,
+      "force_loss": 0.01338754128664732,
+      "step": 45990
+    },
+    {
+      "epoch": 41.36690647482014,
+      "grad_norm": 0.44523322582244873,
+      "learning_rate": 1.4162795994702327e-05,
+      "loss": 0.0158,
+      "step": 46000
+    },
+    {
+      "action_loss": 0.0009819652186706662,
+      "epoch": 41.36690647482014,
+      "step": 46000
+    },
+    {
+      "epoch": 41.36690647482014,
+      "step": 46000,
+      "torque_loss": 0.12439358234405518
+    },
+    {
+      "epoch": 41.36690647482014,
+      "force_loss": 0.0018455515382811427,
+      "step": 46000
+    },
+    {
+      "epoch": 41.37589928057554,
+      "grad_norm": 0.36756545305252075,
+      "learning_rate": 1.4143584363295032e-05,
+      "loss": 0.0166,
+      "step": 46010
+    },
+    {
+      "action_loss": 0.0021100991871207952,
+      "epoch": 41.37589928057554,
+      "step": 46010
+    },
+    {
+      "epoch": 41.37589928057554,
+      "step": 46010,
+      "torque_loss": 0.11846810579299927
+    },
+    {
+      "epoch": 41.37589928057554,
+      "force_loss": 0.004856557119637728,
+      "step": 46010
+    },
+    {
+      "epoch": 41.384892086330936,
+      "grad_norm": 0.27665412425994873,
+      "learning_rate": 1.4124383624120101e-05,
+      "loss": 0.0161,
+      "step": 46020
+    },
+    {
+      "action_loss": 0.0018661081558093429,
+      "epoch": 41.384892086330936,
+      "step": 46020
+    },
+    {
+      "epoch": 41.384892086330936,
+      "step": 46020,
+      "torque_loss": 0.10356643050909042
+    },
+    {
+      "epoch": 41.384892086330936,
+      "force_loss": 0.002635879209265113,
+      "step": 46020
+    },
+    {
+      "epoch": 41.393884892086334,
+      "grad_norm": 0.12330401688814163,
+      "learning_rate": 1.4105193783010151e-05,
+      "loss": 0.016,
+      "step": 46030
+    },
+    {
+      "action_loss": 0.0014474744675680995,
+      "epoch": 41.393884892086334,
+      "step": 46030
+    },
+    {
+      "epoch": 41.393884892086334,
+      "step": 46030,
+      "torque_loss": 0.12947173416614532
+    },
+    {
+      "epoch": 41.393884892086334,
+      "force_loss": 0.001322103664278984,
+      "step": 46030
+    },
+    {
+      "epoch": 41.402877697841724,
+      "grad_norm": 0.34685513377189636,
+      "learning_rate": 1.4086014845794621e-05,
+      "loss": 0.0156,
+      "step": 46040
+    },
+    {
+      "action_loss": 0.009591239504516125,
+      "epoch": 41.402877697841724,
+      "step": 46040
+    },
+    {
+      "epoch": 41.402877697841724,
+      "step": 46040,
+      "torque_loss": 0.11023377627134323
+    },
+    {
+      "epoch": 41.402877697841724,
+      "force_loss": 0.012572459876537323,
+      "step": 46040
+    },
+    {
+      "epoch": 41.41187050359712,
+      "grad_norm": 0.11846359819173813,
+      "learning_rate": 1.4066846818299489e-05,
+      "loss": 0.0166,
+      "step": 46050
+    },
+    {
+      "action_loss": 0.001323712058365345,
+      "epoch": 41.41187050359712,
+      "step": 46050
+    },
+    {
+      "epoch": 41.41187050359712,
+      "step": 46050,
+      "torque_loss": 0.14367739856243134
+    },
+    {
+      "epoch": 41.41187050359712,
+      "force_loss": 0.0006256178603507578,
+      "step": 46050
+    },
+    {
+      "epoch": 41.42086330935252,
+      "grad_norm": 0.13408155739307404,
+      "learning_rate": 1.4047689706347555e-05,
+      "loss": 0.0152,
+      "step": 46060
+    },
+    {
+      "action_loss": 0.0007833559648133814,
+      "epoch": 41.42086330935252,
+      "step": 46060
+    },
+    {
+      "epoch": 41.42086330935252,
+      "step": 46060,
+      "torque_loss": 0.10750800371170044
+    },
+    {
+      "epoch": 41.42086330935252,
+      "force_loss": 0.0025953929871320724,
+      "step": 46060
+    },
+    {
+      "epoch": 41.42985611510792,
+      "grad_norm": 0.1011774018406868,
+      "learning_rate": 1.402854351575822e-05,
+      "loss": 0.0159,
+      "step": 46070
+    },
+    {
+      "action_loss": 0.005844214931130409,
+      "epoch": 41.42985611510792,
+      "step": 46070
+    },
+    {
+      "epoch": 41.42985611510792,
+      "step": 46070,
+      "torque_loss": 0.11894115805625916
+    },
+    {
+      "epoch": 41.42985611510792,
+      "force_loss": 0.00927754770964384,
+      "step": 46070
+    },
+    {
+      "epoch": 41.43884892086331,
+      "grad_norm": 0.0896984338760376,
+      "learning_rate": 1.4009408252347588e-05,
+      "loss": 0.0168,
+      "step": 46080
+    },
+    {
+      "action_loss": 0.00584026426076889,
+      "epoch": 41.43884892086331,
+      "step": 46080
+    },
+    {
+      "epoch": 41.43884892086331,
+      "step": 46080,
+      "torque_loss": 0.1551179140806198
+    },
+    {
+      "epoch": 41.43884892086331,
+      "force_loss": 0.006083325948566198,
+      "step": 46080
+    },
+    {
+      "epoch": 41.447841726618705,
+      "grad_norm": 0.18654625117778778,
+      "learning_rate": 1.399028392192846e-05,
+      "loss": 0.0182,
+      "step": 46090
+    },
+    {
+      "action_loss": 0.0019407831132411957,
+      "epoch": 41.447841726618705,
+      "step": 46090
+    },
+    {
+      "epoch": 41.447841726618705,
+      "step": 46090,
+      "torque_loss": 0.11483168601989746
+    },
+    {
+      "epoch": 41.447841726618705,
+      "force_loss": 0.003445149166509509,
+      "step": 46090
+    },
+    {
+      "epoch": 41.4568345323741,
+      "grad_norm": 0.09483020752668381,
+      "learning_rate": 1.397117053031029e-05,
+      "loss": 0.0222,
+      "step": 46100
+    },
+    {
+      "action_loss": 0.0031765957828611135,
+      "epoch": 41.4568345323741,
+      "step": 46100
+    },
+    {
+      "epoch": 41.4568345323741,
+      "step": 46100,
+      "torque_loss": 0.12828777730464935
+    },
+    {
+      "epoch": 41.4568345323741,
+      "force_loss": 0.005128649529069662,
+      "step": 46100
+    },
+    {
+      "epoch": 41.46582733812949,
+      "grad_norm": 0.25609779357910156,
+      "learning_rate": 1.3952068083299213e-05,
+      "loss": 0.0145,
+      "step": 46110
+    },
+    {
+      "action_loss": 0.0025511858984827995,
+      "epoch": 41.46582733812949,
+      "step": 46110
+    },
+    {
+      "epoch": 41.46582733812949,
+      "step": 46110,
+      "torque_loss": 0.1489318460226059
+    },
+    {
+      "epoch": 41.46582733812949,
+      "force_loss": 0.004632181953638792,
+      "step": 46110
+    },
+    {
+      "epoch": 41.47482014388489,
+      "grad_norm": 0.23925311863422394,
+      "learning_rate": 1.3932976586698082e-05,
+      "loss": 0.0182,
+      "step": 46120
+    },
+    {
+      "action_loss": 0.0010949471034109592,
+      "epoch": 41.47482014388489,
+      "step": 46120
+    },
+    {
+      "epoch": 41.47482014388489,
+      "step": 46120,
+      "torque_loss": 0.10768363624811172
+    },
+    {
+      "epoch": 41.47482014388489,
+      "force_loss": 0.002231505000963807,
+      "step": 46120
+    },
+    {
+      "epoch": 41.48381294964029,
+      "grad_norm": 0.30661842226982117,
+      "learning_rate": 1.3913896046306363e-05,
+      "loss": 0.0163,
+      "step": 46130
+    },
+    {
+      "action_loss": 0.0018126320792362094,
+      "epoch": 41.48381294964029,
+      "step": 46130
+    },
+    {
+      "epoch": 41.48381294964029,
+      "step": 46130,
+      "torque_loss": 0.13091595470905304
+    },
+    {
+      "epoch": 41.48381294964029,
+      "force_loss": 0.0016024975338950753,
+      "step": 46130
+    },
+    {
+      "epoch": 41.492805755395686,
+      "grad_norm": 0.23943215608596802,
+      "learning_rate": 1.389482646792023e-05,
+      "loss": 0.0219,
+      "step": 46140
+    },
+    {
+      "action_loss": 0.00118375348392874,
+      "epoch": 41.492805755395686,
+      "step": 46140
+    },
+    {
+      "epoch": 41.492805755395686,
+      "step": 46140,
+      "torque_loss": 0.10795984417200089
+    },
+    {
+      "epoch": 41.492805755395686,
+      "force_loss": 0.0016770964721217752,
+      "step": 46140
+    },
+    {
+      "epoch": 41.50179856115108,
+      "grad_norm": 0.23393382132053375,
+      "learning_rate": 1.387576785733251e-05,
+      "loss": 0.0155,
+      "step": 46150
+    },
+    {
+      "action_loss": 0.018763892352581024,
+      "epoch": 41.50179856115108,
+      "step": 46150
+    },
+    {
+      "epoch": 41.50179856115108,
+      "step": 46150,
+      "torque_loss": 0.13682280480861664
+    },
+    {
+      "epoch": 41.50179856115108,
+      "force_loss": 0.020032361149787903,
+      "step": 46150
+    },
+    {
+      "epoch": 41.510791366906474,
+      "grad_norm": 0.4050309658050537,
+      "learning_rate": 1.3856720220332703e-05,
+      "loss": 0.0185,
+      "step": 46160
+    },
+    {
+      "action_loss": 0.0017664622282609344,
+      "epoch": 41.510791366906474,
+      "step": 46160
+    },
+    {
+      "epoch": 41.510791366906474,
+      "step": 46160,
+      "torque_loss": 0.1190435066819191
+    },
+    {
+      "epoch": 41.510791366906474,
+      "force_loss": 0.0017898627556860447,
+      "step": 46160
+    },
+    {
+      "epoch": 41.51978417266187,
+      "grad_norm": 0.14559349417686462,
+      "learning_rate": 1.383768356270701e-05,
+      "loss": 0.0143,
+      "step": 46170
+    },
+    {
+      "action_loss": 0.0022800681181252003,
+      "epoch": 41.51978417266187,
+      "step": 46170
+    },
+    {
+      "epoch": 41.51978417266187,
+      "step": 46170,
+      "torque_loss": 0.15389318764209747
+    },
+    {
+      "epoch": 41.51978417266187,
+      "force_loss": 0.011671776883304119,
+      "step": 46170
+    },
+    {
+      "epoch": 41.52877697841727,
+      "grad_norm": 0.2197759747505188,
+      "learning_rate": 1.3818657890238207e-05,
+      "loss": 0.0174,
+      "step": 46180
+    },
+    {
+      "action_loss": 0.001531957183033228,
+      "epoch": 41.52877697841727,
+      "step": 46180
+    },
+    {
+      "epoch": 41.52877697841727,
+      "step": 46180,
+      "torque_loss": 0.10026818513870239
+    },
+    {
+      "epoch": 41.52877697841727,
+      "force_loss": 0.0018418096005916595,
+      "step": 46180
+    },
+    {
+      "epoch": 41.53776978417266,
+      "grad_norm": 0.16019059717655182,
+      "learning_rate": 1.3799643208705859e-05,
+      "loss": 0.0149,
+      "step": 46190
+    },
+    {
+      "action_loss": 0.011442805640399456,
+      "epoch": 41.53776978417266,
+      "step": 46190
+    },
+    {
+      "epoch": 41.53776978417266,
+      "step": 46190,
+      "torque_loss": 0.1324142962694168
+    },
+    {
+      "epoch": 41.53776978417266,
+      "force_loss": 0.013543106615543365,
+      "step": 46190
+    },
+    {
+      "epoch": 41.54676258992806,
+      "grad_norm": 0.26096782088279724,
+      "learning_rate": 1.3780639523886058e-05,
+      "loss": 0.0198,
+      "step": 46200
+    },
+    {
+      "action_loss": 0.0016514029121026397,
+      "epoch": 41.54676258992806,
+      "step": 46200
+    },
+    {
+      "epoch": 41.54676258992806,
+      "step": 46200,
+      "torque_loss": 0.12119301408529282
+    },
+    {
+      "epoch": 41.54676258992806,
+      "force_loss": 0.0034138669725507498,
+      "step": 46200
+    },
+    {
+      "epoch": 41.555755395683455,
+      "grad_norm": 0.0929505005478859,
+      "learning_rate": 1.3761646841551668e-05,
+      "loss": 0.0151,
+      "step": 46210
+    },
+    {
+      "action_loss": 0.0011883891420438886,
+      "epoch": 41.555755395683455,
+      "step": 46210
+    },
+    {
+      "epoch": 41.555755395683455,
+      "step": 46210,
+      "torque_loss": 0.10892126709222794
+    },
+    {
+      "epoch": 41.555755395683455,
+      "force_loss": 0.0010458383476361632,
+      "step": 46210
+    },
+    {
+      "epoch": 41.564748201438846,
+      "grad_norm": 0.057342126965522766,
+      "learning_rate": 1.3742665167472146e-05,
+      "loss": 0.0147,
+      "step": 46220
+    },
+    {
+      "action_loss": 0.006611010525375605,
+      "epoch": 41.564748201438846,
+      "step": 46220
+    },
+    {
+      "epoch": 41.564748201438846,
+      "step": 46220,
+      "torque_loss": 0.1379125565290451
+    },
+    {
+      "epoch": 41.564748201438846,
+      "force_loss": 0.011980275623500347,
+      "step": 46220
+    },
+    {
+      "epoch": 41.57374100719424,
+      "grad_norm": 0.3273448944091797,
+      "learning_rate": 1.372369450741363e-05,
+      "loss": 0.0175,
+      "step": 46230
+    },
+    {
+      "action_loss": 0.0010159645462408662,
+      "epoch": 41.57374100719424,
+      "step": 46230
+    },
+    {
+      "epoch": 41.57374100719424,
+      "step": 46230,
+      "torque_loss": 0.1348085254430771
+    },
+    {
+      "epoch": 41.57374100719424,
+      "force_loss": 0.0008509322069585323,
+      "step": 46230
+    },
+    {
+      "epoch": 41.58273381294964,
+      "grad_norm": 0.0944715365767479,
+      "learning_rate": 1.3704734867138901e-05,
+      "loss": 0.0167,
+      "step": 46240
+    },
+    {
+      "action_loss": 0.001384379924274981,
+      "epoch": 41.58273381294964,
+      "step": 46240
+    },
+    {
+      "epoch": 41.58273381294964,
+      "step": 46240,
+      "torque_loss": 0.10782928019762039
+    },
+    {
+      "epoch": 41.58273381294964,
+      "force_loss": 0.0019347543129697442,
+      "step": 46240
+    },
+    {
+      "epoch": 41.59172661870504,
+      "grad_norm": 0.1085263267159462,
+      "learning_rate": 1.36857862524074e-05,
+      "loss": 0.0153,
+      "step": 46250
+    },
+    {
+      "action_loss": 0.001557510462589562,
+      "epoch": 41.59172661870504,
+      "step": 46250
+    },
+    {
+      "epoch": 41.59172661870504,
+      "step": 46250,
+      "torque_loss": 0.17917267978191376
+    },
+    {
+      "epoch": 41.59172661870504,
+      "force_loss": 0.0009153726859949529,
+      "step": 46250
+    },
+    {
+      "epoch": 41.60071942446043,
+      "grad_norm": 0.10347095131874084,
+      "learning_rate": 1.3666848668975213e-05,
+      "loss": 0.0153,
+      "step": 46260
+    },
+    {
+      "action_loss": 0.0010037199826911092,
+      "epoch": 41.60071942446043,
+      "step": 46260
+    },
+    {
+      "epoch": 41.60071942446043,
+      "step": 46260,
+      "torque_loss": 0.12637753784656525
+    },
+    {
+      "epoch": 41.60071942446043,
+      "force_loss": 0.002662701765075326,
+      "step": 46260
+    },
+    {
+      "epoch": 41.60971223021583,
+      "grad_norm": 0.1711738407611847,
+      "learning_rate": 1.3647922122595063e-05,
+      "loss": 0.0147,
+      "step": 46270
+    },
+    {
+      "action_loss": 0.0010016815504059196,
+      "epoch": 41.60971223021583,
+      "step": 46270
+    },
+    {
+      "epoch": 41.60971223021583,
+      "step": 46270,
+      "torque_loss": 0.16279369592666626
+    },
+    {
+      "epoch": 41.60971223021583,
+      "force_loss": 0.0016701758140698075,
+      "step": 46270
+    },
+    {
+      "epoch": 41.618705035971225,
+      "grad_norm": 0.16988788545131683,
+      "learning_rate": 1.3629006619016366e-05,
+      "loss": 0.0184,
+      "step": 46280
+    },
+    {
+      "action_loss": 0.0006814058870077133,
+      "epoch": 41.618705035971225,
+      "step": 46280
+    },
+    {
+      "epoch": 41.618705035971225,
+      "step": 46280,
+      "torque_loss": 0.13192714750766754
+    },
+    {
+      "epoch": 41.618705035971225,
+      "force_loss": 0.004833744373172522,
+      "step": 46280
+    },
+    {
+      "epoch": 41.62769784172662,
+      "grad_norm": 0.06417033821344376,
+      "learning_rate": 1.3610102163985139e-05,
+      "loss": 0.0201,
+      "step": 46290
+    },
+    {
+      "action_loss": 0.002310168696567416,
+      "epoch": 41.62769784172662,
+      "step": 46290
+    },
+    {
+      "epoch": 41.62769784172662,
+      "step": 46290,
+      "torque_loss": 0.14548902213573456
+    },
+    {
+      "epoch": 41.62769784172662,
+      "force_loss": 0.004910944029688835,
+      "step": 46290
+    },
+    {
+      "epoch": 41.63669064748201,
+      "grad_norm": 0.09510938078165054,
+      "learning_rate": 1.3591208763244057e-05,
+      "loss": 0.0164,
+      "step": 46300
+    },
+    {
+      "action_loss": 0.0007438471657224,
+      "epoch": 41.63669064748201,
+      "step": 46300
+    },
+    {
+      "epoch": 41.63669064748201,
+      "step": 46300,
+      "torque_loss": 0.1430986076593399
+    },
+    {
+      "epoch": 41.63669064748201,
+      "force_loss": 0.0009638678166083992,
+      "step": 46300
+    },
+    {
+      "epoch": 41.64568345323741,
+      "grad_norm": 0.0993509292602539,
+      "learning_rate": 1.3572326422532428e-05,
+      "loss": 0.0199,
+      "step": 46310
+    },
+    {
+      "action_loss": 0.00236869347281754,
+      "epoch": 41.64568345323741,
+      "step": 46310
+    },
+    {
+      "epoch": 41.64568345323741,
+      "step": 46310,
+      "torque_loss": 0.13813556730747223
+    },
+    {
+      "epoch": 41.64568345323741,
+      "force_loss": 0.0011360583594068885,
+      "step": 46310
+    },
+    {
+      "epoch": 41.65467625899281,
+      "grad_norm": 0.30295559763908386,
+      "learning_rate": 1.355345514758622e-05,
+      "loss": 0.0155,
+      "step": 46320
+    },
+    {
+      "action_loss": 0.010639210231602192,
+      "epoch": 41.65467625899281,
+      "step": 46320
+    },
+    {
+      "epoch": 41.65467625899281,
+      "step": 46320,
+      "torque_loss": 0.1164890006184578
+    },
+    {
+      "epoch": 41.65467625899281,
+      "force_loss": 0.01661677099764347,
+      "step": 46320
+    },
+    {
+      "epoch": 41.6636690647482,
+      "grad_norm": 0.2708222568035126,
+      "learning_rate": 1.3534594944138007e-05,
+      "loss": 0.0165,
+      "step": 46330
+    },
+    {
+      "action_loss": 0.0009878772543743253,
+      "epoch": 41.6636690647482,
+      "step": 46330
+    },
+    {
+      "epoch": 41.6636690647482,
+      "step": 46330,
+      "torque_loss": 0.12461438030004501
+    },
+    {
+      "epoch": 41.6636690647482,
+      "force_loss": 0.0009639711934141815,
+      "step": 46330
+    },
+    {
+      "epoch": 41.672661870503596,
+      "grad_norm": 0.4043882489204407,
+      "learning_rate": 1.3515745817917069e-05,
+      "loss": 0.0146,
+      "step": 46340
+    },
+    {
+      "action_loss": 0.002103827428072691,
+      "epoch": 41.672661870503596,
+      "step": 46340
+    },
+    {
+      "epoch": 41.672661870503596,
+      "step": 46340,
+      "torque_loss": 0.1126183271408081
+    },
+    {
+      "epoch": 41.672661870503596,
+      "force_loss": 0.0028829958755522966,
+      "step": 46340
+    },
+    {
+      "epoch": 41.681654676258994,
+      "grad_norm": 0.5196036100387573,
+      "learning_rate": 1.3496907774649208e-05,
+      "loss": 0.0151,
+      "step": 46350
+    },
+    {
+      "action_loss": 0.000996283837594092,
+      "epoch": 41.681654676258994,
+      "step": 46350
+    },
+    {
+      "epoch": 41.681654676258994,
+      "step": 46350,
+      "torque_loss": 0.12013643980026245
+    },
+    {
+      "epoch": 41.681654676258994,
+      "force_loss": 0.0012420298298820853,
+      "step": 46350
+    },
+    {
+      "epoch": 41.69064748201439,
+      "grad_norm": 0.2126244753599167,
+      "learning_rate": 1.3478080820056987e-05,
+      "loss": 0.0207,
+      "step": 46360
+    },
+    {
+      "action_loss": 0.0007505409885197878,
+      "epoch": 41.69064748201439,
+      "step": 46360
+    },
+    {
+      "epoch": 41.69064748201439,
+      "step": 46360,
+      "torque_loss": 0.16304346919059753
+    },
+    {
+      "epoch": 41.69064748201439,
+      "force_loss": 0.0014648130163550377,
+      "step": 46360
+    },
+    {
+      "epoch": 41.69964028776978,
+      "grad_norm": 0.11799972504377365,
+      "learning_rate": 1.3459264959859474e-05,
+      "loss": 0.0163,
+      "step": 46370
+    },
+    {
+      "action_loss": 0.0010451226262375712,
+      "epoch": 41.69964028776978,
+      "step": 46370
+    },
+    {
+      "epoch": 41.69964028776978,
+      "step": 46370,
+      "torque_loss": 0.19468139111995697
+    },
+    {
+      "epoch": 41.69964028776978,
+      "force_loss": 0.0022757528349757195,
+      "step": 46370
+    },
+    {
+      "epoch": 41.70863309352518,
+      "grad_norm": 0.32826197147369385,
+      "learning_rate": 1.3440460199772487e-05,
+      "loss": 0.0131,
+      "step": 46380
+    },
+    {
+      "action_loss": 0.0007544537074863911,
+      "epoch": 41.70863309352518,
+      "step": 46380
+    },
+    {
+      "epoch": 41.70863309352518,
+      "step": 46380,
+      "torque_loss": 0.21511048078536987
+    },
+    {
+      "epoch": 41.70863309352518,
+      "force_loss": 0.0018087811768054962,
+      "step": 46380
+    },
+    {
+      "epoch": 41.71762589928058,
+      "grad_norm": 0.3819451630115509,
+      "learning_rate": 1.3421666545508382e-05,
+      "loss": 0.0167,
+      "step": 46390
+    },
+    {
+      "action_loss": 0.0023083712439984083,
+      "epoch": 41.71762589928058,
+      "step": 46390
+    },
+    {
+      "epoch": 41.71762589928058,
+      "step": 46390,
+      "torque_loss": 0.15789858996868134
+    },
+    {
+      "epoch": 41.71762589928058,
+      "force_loss": 0.007165252696722746,
+      "step": 46390
+    },
+    {
+      "epoch": 41.726618705035975,
+      "grad_norm": 0.15164703130722046,
+      "learning_rate": 1.3402884002776194e-05,
+      "loss": 0.0149,
+      "step": 46400
+    },
+    {
+      "action_loss": 0.0008625375921837986,
+      "epoch": 41.726618705035975,
+      "step": 46400
+    },
+    {
+      "epoch": 41.726618705035975,
+      "step": 46400,
+      "torque_loss": 0.18257851898670197
+    },
+    {
+      "epoch": 41.726618705035975,
+      "force_loss": 0.0007447314565069973,
+      "step": 46400
+    },
+    {
+      "epoch": 41.735611510791365,
+      "grad_norm": 0.14613275229930878,
+      "learning_rate": 1.3384112577281555e-05,
+      "loss": 0.0195,
+      "step": 46410
+    },
+    {
+      "action_loss": 0.005302126053720713,
+      "epoch": 41.735611510791365,
+      "step": 46410
+    },
+    {
+      "epoch": 41.735611510791365,
+      "step": 46410,
+      "torque_loss": 0.10265422612428665
+    },
+    {
+      "epoch": 41.735611510791365,
+      "force_loss": 0.009346667677164078,
+      "step": 46410
+    },
+    {
+      "epoch": 41.74460431654676,
+      "grad_norm": 0.17907406389713287,
+      "learning_rate": 1.3365352274726711e-05,
+      "loss": 0.0172,
+      "step": 46420
+    },
+    {
+      "action_loss": 0.0017045149579644203,
+      "epoch": 41.74460431654676,
+      "step": 46420
+    },
+    {
+      "epoch": 41.74460431654676,
+      "step": 46420,
+      "torque_loss": 0.13871800899505615
+    },
+    {
+      "epoch": 41.74460431654676,
+      "force_loss": 0.0034153710585087538,
+      "step": 46420
+    },
+    {
+      "epoch": 41.75359712230216,
+      "grad_norm": 0.1436508744955063,
+      "learning_rate": 1.3346603100810578e-05,
+      "loss": 0.0166,
+      "step": 46430
+    },
+    {
+      "action_loss": 0.0007342484313994646,
+      "epoch": 41.75359712230216,
+      "step": 46430
+    },
+    {
+      "epoch": 41.75359712230216,
+      "step": 46430,
+      "torque_loss": 0.08791147917509079
+    },
+    {
+      "epoch": 41.75359712230216,
+      "force_loss": 0.002309478120878339,
+      "step": 46430
+    },
+    {
+      "epoch": 41.76258992805755,
+      "grad_norm": 0.1547205150127411,
+      "learning_rate": 1.3327865061228645e-05,
+      "loss": 0.017,
+      "step": 46440
+    },
+    {
+      "action_loss": 0.003924580756574869,
+      "epoch": 41.76258992805755,
+      "step": 46440
+    },
+    {
+      "epoch": 41.76258992805755,
+      "step": 46440,
+      "torque_loss": 0.12193212658166885
+    },
+    {
+      "epoch": 41.76258992805755,
+      "force_loss": 0.0033480441197752953,
+      "step": 46440
+    },
+    {
+      "epoch": 41.77158273381295,
+      "grad_norm": 0.5368626713752747,
+      "learning_rate": 1.330913816167304e-05,
+      "loss": 0.0212,
+      "step": 46450
+    },
+    {
+      "action_loss": 0.0010073069715872407,
+      "epoch": 41.77158273381295,
+      "step": 46450
+    },
+    {
+      "epoch": 41.77158273381295,
+      "step": 46450,
+      "torque_loss": 0.09068405628204346
+    },
+    {
+      "epoch": 41.77158273381295,
+      "force_loss": 0.004405824933201075,
+      "step": 46450
+    },
+    {
+      "epoch": 41.780575539568346,
+      "grad_norm": 0.2417355477809906,
+      "learning_rate": 1.3290422407832492e-05,
+      "loss": 0.0178,
+      "step": 46460
+    },
+    {
+      "action_loss": 0.012405410408973694,
+      "epoch": 41.780575539568346,
+      "step": 46460
+    },
+    {
+      "epoch": 41.780575539568346,
+      "step": 46460,
+      "torque_loss": 0.2033160924911499
+    },
+    {
+      "epoch": 41.780575539568346,
+      "force_loss": 0.024507006630301476,
+      "step": 46460
+    },
+    {
+      "epoch": 41.789568345323744,
+      "grad_norm": 0.08595062047243118,
+      "learning_rate": 1.3271717805392354e-05,
+      "loss": 0.0177,
+      "step": 46470
+    },
+    {
+      "action_loss": 0.002071862341836095,
+      "epoch": 41.789568345323744,
+      "step": 46470
+    },
+    {
+      "epoch": 41.789568345323744,
+      "step": 46470,
+      "torque_loss": 0.1411842703819275
+    },
+    {
+      "epoch": 41.789568345323744,
+      "force_loss": 0.0019701633136719465,
+      "step": 46470
+    },
+    {
+      "epoch": 41.798561151079134,
+      "grad_norm": 0.10007981956005096,
+      "learning_rate": 1.3253024360034582e-05,
+      "loss": 0.0164,
+      "step": 46480
+    },
+    {
+      "action_loss": 0.0008854269399307668,
+      "epoch": 41.798561151079134,
+      "step": 46480
+    },
+    {
+      "epoch": 41.798561151079134,
+      "step": 46480,
+      "torque_loss": 0.1057625338435173
+    },
+    {
+      "epoch": 41.798561151079134,
+      "force_loss": 0.004036023747175932,
+      "step": 46480
+    },
+    {
+      "epoch": 41.80755395683453,
+      "grad_norm": 0.05173208937048912,
+      "learning_rate": 1.323434207743779e-05,
+      "loss": 0.0186,
+      "step": 46490
+    },
+    {
+      "action_loss": 0.002810916630551219,
+      "epoch": 41.80755395683453,
+      "step": 46490
+    },
+    {
+      "epoch": 41.80755395683453,
+      "step": 46490,
+      "torque_loss": 0.11307287216186523
+    },
+    {
+      "epoch": 41.80755395683453,
+      "force_loss": 0.001997811021283269,
+      "step": 46490
+    },
+    {
+      "epoch": 41.81654676258993,
+      "grad_norm": 0.114821657538414,
+      "learning_rate": 1.3215670963277105e-05,
+      "loss": 0.0188,
+      "step": 46500
+    },
+    {
+      "action_loss": 0.0010481419740244746,
+      "epoch": 41.81654676258993,
+      "step": 46500
+    },
+    {
+      "epoch": 41.81654676258993,
+      "step": 46500,
+      "torque_loss": 0.1736762970685959
+    },
+    {
+      "epoch": 41.81654676258993,
+      "force_loss": 0.0021222021896392107,
+      "step": 46500
+    },
+    {
+      "epoch": 41.82553956834532,
+      "grad_norm": 0.1164456307888031,
+      "learning_rate": 1.3197011023224376e-05,
+      "loss": 0.0146,
+      "step": 46510
+    },
+    {
+      "action_loss": 0.0010417585726827383,
+      "epoch": 41.82553956834532,
+      "step": 46510
+    },
+    {
+      "epoch": 41.82553956834532,
+      "step": 46510,
+      "torque_loss": 0.11614730209112167
+    },
+    {
+      "epoch": 41.82553956834532,
+      "force_loss": 0.002403555205091834,
+      "step": 46510
+    },
+    {
+      "epoch": 41.83453237410072,
+      "grad_norm": 0.29869675636291504,
+      "learning_rate": 1.3178362262947941e-05,
+      "loss": 0.0209,
+      "step": 46520
+    },
+    {
+      "action_loss": 0.004037730861455202,
+      "epoch": 41.83453237410072,
+      "step": 46520
+    },
+    {
+      "epoch": 41.83453237410072,
+      "step": 46520,
+      "torque_loss": 0.12442287057638168
+    },
+    {
+      "epoch": 41.83453237410072,
+      "force_loss": 0.008197152987122536,
+      "step": 46520
+    },
+    {
+      "epoch": 41.843525179856115,
+      "grad_norm": 0.11285030841827393,
+      "learning_rate": 1.3159724688112845e-05,
+      "loss": 0.0149,
+      "step": 46530
+    },
+    {
+      "action_loss": 0.0009287123684771359,
+      "epoch": 41.843525179856115,
+      "step": 46530
+    },
+    {
+      "epoch": 41.843525179856115,
+      "step": 46530,
+      "torque_loss": 0.14639179408550262
+    },
+    {
+      "epoch": 41.843525179856115,
+      "force_loss": 0.00194704404566437,
+      "step": 46530
+    },
+    {
+      "epoch": 41.85251798561151,
+      "grad_norm": 0.12745630741119385,
+      "learning_rate": 1.3141098304380683e-05,
+      "loss": 0.0159,
+      "step": 46540
+    },
+    {
+      "action_loss": 0.0009762852569110692,
+      "epoch": 41.85251798561151,
+      "step": 46540
+    },
+    {
+      "epoch": 41.85251798561151,
+      "step": 46540,
+      "torque_loss": 0.16845448315143585
+    },
+    {
+      "epoch": 41.85251798561151,
+      "force_loss": 0.0028479357715696096,
+      "step": 46540
+    },
+    {
+      "epoch": 41.861510791366904,
+      "grad_norm": 0.07736373692750931,
+      "learning_rate": 1.3122483117409651e-05,
+      "loss": 0.0161,
+      "step": 46550
+    },
+    {
+      "action_loss": 0.0008849836303852499,
+      "epoch": 41.861510791366904,
+      "step": 46550
+    },
+    {
+      "epoch": 41.861510791366904,
+      "step": 46550,
+      "torque_loss": 0.13746891915798187
+    },
+    {
+      "epoch": 41.861510791366904,
+      "force_loss": 0.0025917694438248873,
+      "step": 46550
+    },
+    {
+      "epoch": 41.8705035971223,
+      "grad_norm": 0.40781673789024353,
+      "learning_rate": 1.3103879132854552e-05,
+      "loss": 0.0173,
+      "step": 46560
+    },
+    {
+      "action_loss": 0.023091599345207214,
+      "epoch": 41.8705035971223,
+      "step": 46560
+    },
+    {
+      "epoch": 41.8705035971223,
+      "step": 46560,
+      "torque_loss": 0.12224598973989487
+    },
+    {
+      "epoch": 41.8705035971223,
+      "force_loss": 0.02782335877418518,
+      "step": 46560
+    },
+    {
+      "epoch": 41.8794964028777,
+      "grad_norm": 0.10636132955551147,
+      "learning_rate": 1.3085286356366771e-05,
+      "loss": 0.0174,
+      "step": 46570
+    },
+    {
+      "action_loss": 0.0009309242013841867,
+      "epoch": 41.8794964028777,
+      "step": 46570
+    },
+    {
+      "epoch": 41.8794964028777,
+      "step": 46570,
+      "torque_loss": 0.0996650829911232
+    },
+    {
+      "epoch": 41.8794964028777,
+      "force_loss": 0.0025415238924324512,
+      "step": 46570
+    },
+    {
+      "epoch": 41.888489208633096,
+      "grad_norm": 0.08773689717054367,
+      "learning_rate": 1.3066704793594337e-05,
+      "loss": 0.0164,
+      "step": 46580
+    },
+    {
+      "action_loss": 0.002911119954660535,
+      "epoch": 41.888489208633096,
+      "step": 46580
+    },
+    {
+      "epoch": 41.888489208633096,
+      "step": 46580,
+      "torque_loss": 0.1087740957736969
+    },
+    {
+      "epoch": 41.888489208633096,
+      "force_loss": 0.0035202335566282272,
+      "step": 46580
+    },
+    {
+      "epoch": 41.89748201438849,
+      "grad_norm": 0.32328110933303833,
+      "learning_rate": 1.3048134450181816e-05,
+      "loss": 0.0155,
+      "step": 46590
+    },
+    {
+      "action_loss": 0.00818340852856636,
+      "epoch": 41.89748201438849,
+      "step": 46590
+    },
+    {
+      "epoch": 41.89748201438849,
+      "step": 46590,
+      "torque_loss": 0.15708810091018677
+    },
+    {
+      "epoch": 41.89748201438849,
+      "force_loss": 0.012910461984574795,
+      "step": 46590
+    },
+    {
+      "epoch": 41.906474820143885,
+      "grad_norm": 0.11418038606643677,
+      "learning_rate": 1.3029575331770394e-05,
+      "loss": 0.0153,
+      "step": 46600
+    },
+    {
+      "action_loss": 0.0007504620589315891,
+      "epoch": 41.906474820143885,
+      "step": 46600
+    },
+    {
+      "epoch": 41.906474820143885,
+      "step": 46600,
+      "torque_loss": 0.12935860455036163
+    },
+    {
+      "epoch": 41.906474820143885,
+      "force_loss": 0.0008481654222123325,
+      "step": 46600
+    },
+    {
+      "epoch": 41.91546762589928,
+      "grad_norm": 0.3308919072151184,
+      "learning_rate": 1.3011027443997837e-05,
+      "loss": 0.0146,
+      "step": 46610
+    },
+    {
+      "action_loss": 0.0008278512395918369,
+      "epoch": 41.91546762589928,
+      "step": 46610
+    },
+    {
+      "epoch": 41.91546762589928,
+      "step": 46610,
+      "torque_loss": 0.13288386166095734
+    },
+    {
+      "epoch": 41.91546762589928,
+      "force_loss": 0.0007833181880414486,
+      "step": 46610
+    },
+    {
+      "epoch": 41.92446043165468,
+      "grad_norm": 0.21655842661857605,
+      "learning_rate": 1.2992490792498507e-05,
+      "loss": 0.0144,
+      "step": 46620
+    },
+    {
+      "action_loss": 0.003982121590524912,
+      "epoch": 41.92446043165468,
+      "step": 46620
+    },
+    {
+      "epoch": 41.92446043165468,
+      "step": 46620,
+      "torque_loss": 0.1748986840248108
+    },
+    {
+      "epoch": 41.92446043165468,
+      "force_loss": 0.012981380335986614,
+      "step": 46620
+    },
+    {
+      "epoch": 41.93345323741007,
+      "grad_norm": 0.18265517055988312,
+      "learning_rate": 1.297396538290333e-05,
+      "loss": 0.0171,
+      "step": 46630
+    },
+    {
+      "action_loss": 0.0009083410259336233,
+      "epoch": 41.93345323741007,
+      "step": 46630
+    },
+    {
+      "epoch": 41.93345323741007,
+      "step": 46630,
+      "torque_loss": 0.10882826894521713
+    },
+    {
+      "epoch": 41.93345323741007,
+      "force_loss": 0.0006618359475396574,
+      "step": 46630
+    },
+    {
+      "epoch": 41.94244604316547,
+      "grad_norm": 0.2547270357608795,
+      "learning_rate": 1.2955451220839888e-05,
+      "loss": 0.013,
+      "step": 46640
+    },
+    {
+      "action_loss": 0.0031260100658982992,
+      "epoch": 41.94244604316547,
+      "step": 46640
+    },
+    {
+      "epoch": 41.94244604316547,
+      "step": 46640,
+      "torque_loss": 0.14381995797157288
+    },
+    {
+      "epoch": 41.94244604316547,
+      "force_loss": 0.0017336081946268678,
+      "step": 46640
+    },
+    {
+      "epoch": 41.951438848920866,
+      "grad_norm": 0.6436915397644043,
+      "learning_rate": 1.2936948311932223e-05,
+      "loss": 0.0174,
+      "step": 46650
+    },
+    {
+      "action_loss": 0.0027956615667790174,
+      "epoch": 41.951438848920866,
+      "step": 46650
+    },
+    {
+      "epoch": 41.951438848920866,
+      "step": 46650,
+      "torque_loss": 0.11021675914525986
+    },
+    {
+      "epoch": 41.951438848920866,
+      "force_loss": 0.0063346861861646175,
+      "step": 46650
+    },
+    {
+      "epoch": 41.960431654676256,
+      "grad_norm": 0.09409542381763458,
+      "learning_rate": 1.2918456661801104e-05,
+      "loss": 0.0152,
+      "step": 46660
+    },
+    {
+      "action_loss": 0.000733214255888015,
+      "epoch": 41.960431654676256,
+      "step": 46660
+    },
+    {
+      "epoch": 41.960431654676256,
+      "step": 46660,
+      "torque_loss": 0.1312297135591507
+    },
+    {
+      "epoch": 41.960431654676256,
+      "force_loss": 0.0016367450589314103,
+      "step": 46660
+    },
+    {
+      "epoch": 41.969424460431654,
+      "grad_norm": 0.06914551556110382,
+      "learning_rate": 1.2899976276063736e-05,
+      "loss": 0.0149,
+      "step": 46670
+    },
+    {
+      "action_loss": 0.004222599323838949,
+      "epoch": 41.969424460431654,
+      "step": 46670
+    },
+    {
+      "epoch": 41.969424460431654,
+      "step": 46670,
+      "torque_loss": 0.10489914566278458
+    },
+    {
+      "epoch": 41.969424460431654,
+      "force_loss": 0.0031899672467261553,
+      "step": 46670
+    },
+    {
+      "epoch": 41.97841726618705,
+      "grad_norm": 0.18672610819339752,
+      "learning_rate": 1.2881507160334022e-05,
+      "loss": 0.0162,
+      "step": 46680
+    },
+    {
+      "action_loss": 0.0008945378358475864,
+      "epoch": 41.97841726618705,
+      "step": 46680
+    },
+    {
+      "epoch": 41.97841726618705,
+      "step": 46680,
+      "torque_loss": 0.11672995239496231
+    },
+    {
+      "epoch": 41.97841726618705,
+      "force_loss": 0.000731871637981385,
+      "step": 46680
+    },
+    {
+      "epoch": 41.98741007194245,
+      "grad_norm": 0.0604969821870327,
+      "learning_rate": 1.286304932022238e-05,
+      "loss": 0.0141,
+      "step": 46690
+    },
+    {
+      "action_loss": 0.001619009650312364,
+      "epoch": 41.98741007194245,
+      "step": 46690
+    },
+    {
+      "epoch": 41.98741007194245,
+      "step": 46690,
+      "torque_loss": 0.15916641056537628
+    },
+    {
+      "epoch": 41.98741007194245,
+      "force_loss": 0.0056321038864552975,
+      "step": 46690
+    },
+    {
+      "epoch": 41.99640287769784,
+      "grad_norm": 0.32196900248527527,
+      "learning_rate": 1.2844602761335806e-05,
+      "loss": 0.0195,
+      "step": 46700
+    },
+    {
+      "action_loss": 0.001419402309693396,
+      "epoch": 41.99640287769784,
+      "step": 46700
+    },
+    {
+      "epoch": 41.99640287769784,
+      "step": 46700,
+      "torque_loss": 0.15044492483139038
+    },
+    {
+      "epoch": 41.99640287769784,
+      "force_loss": 0.001633634907193482,
+      "step": 46700
+    },
+    {
+      "epoch": 42.00539568345324,
+      "grad_norm": 0.11786222457885742,
+      "learning_rate": 1.2826167489277885e-05,
+      "loss": 0.0142,
+      "step": 46710
+    },
+    {
+      "action_loss": 0.0036024730652570724,
+      "epoch": 42.00539568345324,
+      "step": 46710
+    },
+    {
+      "epoch": 42.00539568345324,
+      "step": 46710,
+      "torque_loss": 0.13186015188694
+    },
+    {
+      "epoch": 42.00539568345324,
+      "force_loss": 0.014024369418621063,
+      "step": 46710
+    },
+    {
+      "epoch": 42.014388489208635,
+      "grad_norm": 0.18105298280715942,
+      "learning_rate": 1.2807743509648745e-05,
+      "loss": 0.0139,
+      "step": 46720
+    },
+    {
+      "action_loss": 0.0012587569653987885,
+      "epoch": 42.014388489208635,
+      "step": 46720
+    },
+    {
+      "epoch": 42.014388489208635,
+      "step": 46720,
+      "torque_loss": 0.07623998075723648
+    },
+    {
+      "epoch": 42.014388489208635,
+      "force_loss": 0.000937749573495239,
+      "step": 46720
+    },
+    {
+      "epoch": 42.023381294964025,
+      "grad_norm": 0.20153272151947021,
+      "learning_rate": 1.2789330828045149e-05,
+      "loss": 0.0148,
+      "step": 46730
+    },
+    {
+      "action_loss": 0.0010120555525645614,
+      "epoch": 42.023381294964025,
+      "step": 46730
+    },
+    {
+      "epoch": 42.023381294964025,
+      "step": 46730,
+      "torque_loss": 0.12439671158790588
+    },
+    {
+      "epoch": 42.023381294964025,
+      "force_loss": 0.0018724799156188965,
+      "step": 46730
+    },
+    {
+      "epoch": 42.03237410071942,
+      "grad_norm": 0.14741671085357666,
+      "learning_rate": 1.2770929450060332e-05,
+      "loss": 0.0161,
+      "step": 46740
+    },
+    {
+      "action_loss": 0.0007091127336025238,
+      "epoch": 42.03237410071942,
+      "step": 46740
+    },
+    {
+      "epoch": 42.03237410071942,
+      "step": 46740,
+      "torque_loss": 0.1275375634431839
+    },
+    {
+      "epoch": 42.03237410071942,
+      "force_loss": 0.0010350263910368085,
+      "step": 46740
+    },
+    {
+      "epoch": 42.04136690647482,
+      "grad_norm": 0.30644872784614563,
+      "learning_rate": 1.2752539381284184e-05,
+      "loss": 0.0158,
+      "step": 46750
+    },
+    {
+      "action_loss": 0.0012461119331419468,
+      "epoch": 42.04136690647482,
+      "step": 46750
+    },
+    {
+      "epoch": 42.04136690647482,
+      "step": 46750,
+      "torque_loss": 0.10962500423192978
+    },
+    {
+      "epoch": 42.04136690647482,
+      "force_loss": 0.0029679357539862394,
+      "step": 46750
+    },
+    {
+      "epoch": 42.05035971223022,
+      "grad_norm": 0.05391109362244606,
+      "learning_rate": 1.273416062730311e-05,
+      "loss": 0.0151,
+      "step": 46760
+    },
+    {
+      "action_loss": 0.00622273376211524,
+      "epoch": 42.05035971223022,
+      "step": 46760
+    },
+    {
+      "epoch": 42.05035971223022,
+      "step": 46760,
+      "torque_loss": 0.142961323261261
+    },
+    {
+      "epoch": 42.05035971223022,
+      "force_loss": 0.008253638632595539,
+      "step": 46760
+    },
+    {
+      "epoch": 42.05935251798561,
+      "grad_norm": 0.06334792822599411,
+      "learning_rate": 1.2715793193700088e-05,
+      "loss": 0.0163,
+      "step": 46770
+    },
+    {
+      "action_loss": 0.0005098909023217857,
+      "epoch": 42.05935251798561,
+      "step": 46770
+    },
+    {
+      "epoch": 42.05935251798561,
+      "step": 46770,
+      "torque_loss": 0.11684909462928772
+    },
+    {
+      "epoch": 42.05935251798561,
+      "force_loss": 0.0008358575869351625,
+      "step": 46770
+    },
+    {
+      "epoch": 42.068345323741006,
+      "grad_norm": 0.12084853649139404,
+      "learning_rate": 1.2697437086054664e-05,
+      "loss": 0.0172,
+      "step": 46780
+    },
+    {
+      "action_loss": 0.010842480696737766,
+      "epoch": 42.068345323741006,
+      "step": 46780
+    },
+    {
+      "epoch": 42.068345323741006,
+      "step": 46780,
+      "torque_loss": 0.1762898713350296
+    },
+    {
+      "epoch": 42.068345323741006,
+      "force_loss": 0.012682903558015823,
+      "step": 46780
+    },
+    {
+      "epoch": 42.077338129496404,
+      "grad_norm": 0.17878368496894836,
+      "learning_rate": 1.2679092309942937e-05,
+      "loss": 0.0163,
+      "step": 46790
+    },
+    {
+      "action_loss": 0.0048118163831532,
+      "epoch": 42.077338129496404,
+      "step": 46790
+    },
+    {
+      "epoch": 42.077338129496404,
+      "step": 46790,
+      "torque_loss": 0.13234467804431915
+    },
+    {
+      "epoch": 42.077338129496404,
+      "force_loss": 0.009777412749826908,
+      "step": 46790
+    },
+    {
+      "epoch": 42.0863309352518,
+      "grad_norm": 0.2512463927268982,
+      "learning_rate": 1.266075887093755e-05,
+      "loss": 0.0147,
+      "step": 46800
+    },
+    {
+      "action_loss": 0.0015251734293997288,
+      "epoch": 42.0863309352518,
+      "step": 46800
+    },
+    {
+      "epoch": 42.0863309352518,
+      "step": 46800,
+      "torque_loss": 0.1434912234544754
+    },
+    {
+      "epoch": 42.0863309352518,
+      "force_loss": 0.004402507562190294,
+      "step": 46800
+    },
+    {
+      "epoch": 42.09532374100719,
+      "grad_norm": 0.30996543169021606,
+      "learning_rate": 1.2642436774607757e-05,
+      "loss": 0.0176,
+      "step": 46810
+    },
+    {
+      "action_loss": 0.0016457010060548782,
+      "epoch": 42.09532374100719,
+      "step": 46810
+    },
+    {
+      "epoch": 42.09532374100719,
+      "step": 46810,
+      "torque_loss": 0.1044391393661499
+    },
+    {
+      "epoch": 42.09532374100719,
+      "force_loss": 0.0029262385796755552,
+      "step": 46810
+    },
+    {
+      "epoch": 42.10431654676259,
+      "grad_norm": 0.11659257858991623,
+      "learning_rate": 1.2624126026519278e-05,
+      "loss": 0.0133,
+      "step": 46820
+    },
+    {
+      "action_loss": 0.009499938227236271,
+      "epoch": 42.10431654676259,
+      "step": 46820
+    },
+    {
+      "epoch": 42.10431654676259,
+      "step": 46820,
+      "torque_loss": 0.19349241256713867
+    },
+    {
+      "epoch": 42.10431654676259,
+      "force_loss": 0.030090009793639183,
+      "step": 46820
+    },
+    {
+      "epoch": 42.11330935251799,
+      "grad_norm": 0.16274572908878326,
+      "learning_rate": 1.2605826632234474e-05,
+      "loss": 0.0174,
+      "step": 46830
+    },
+    {
+      "action_loss": 0.002080908278003335,
+      "epoch": 42.11330935251799,
+      "step": 46830
+    },
+    {
+      "epoch": 42.11330935251799,
+      "step": 46830,
+      "torque_loss": 0.130034938454628
+    },
+    {
+      "epoch": 42.11330935251799,
+      "force_loss": 0.003555377246811986,
+      "step": 46830
+    },
+    {
+      "epoch": 42.12230215827338,
+      "grad_norm": 0.14463350176811218,
+      "learning_rate": 1.2587538597312198e-05,
+      "loss": 0.0147,
+      "step": 46840
+    },
+    {
+      "action_loss": 0.02536866068840027,
+      "epoch": 42.12230215827338,
+      "step": 46840
+    },
+    {
+      "epoch": 42.12230215827338,
+      "step": 46840,
+      "torque_loss": 0.11947962641716003
+    },
+    {
+      "epoch": 42.12230215827338,
+      "force_loss": 0.0279528945684433,
+      "step": 46840
+    },
+    {
+      "epoch": 42.131294964028775,
+      "grad_norm": 0.10108552128076553,
+      "learning_rate": 1.2569261927307884e-05,
+      "loss": 0.0177,
+      "step": 46850
+    },
+    {
+      "action_loss": 0.0006828121840953827,
+      "epoch": 42.131294964028775,
+      "step": 46850
+    },
+    {
+      "epoch": 42.131294964028775,
+      "step": 46850,
+      "torque_loss": 0.17187131941318512
+    },
+    {
+      "epoch": 42.131294964028775,
+      "force_loss": 0.0005717813619412482,
+      "step": 46850
+    },
+    {
+      "epoch": 42.14028776978417,
+      "grad_norm": 0.07092025130987167,
+      "learning_rate": 1.2550996627773493e-05,
+      "loss": 0.015,
+      "step": 46860
+    },
+    {
+      "action_loss": 0.0008692220435477793,
+      "epoch": 42.14028776978417,
+      "step": 46860
+    },
+    {
+      "epoch": 42.14028776978417,
+      "step": 46860,
+      "torque_loss": 0.14385050535202026
+    },
+    {
+      "epoch": 42.14028776978417,
+      "force_loss": 0.0017538117244839668,
+      "step": 46860
+    },
+    {
+      "epoch": 42.14928057553957,
+      "grad_norm": 0.35319799184799194,
+      "learning_rate": 1.2532742704257527e-05,
+      "loss": 0.0168,
+      "step": 46870
+    },
+    {
+      "action_loss": 0.003098190063610673,
+      "epoch": 42.14928057553957,
+      "step": 46870
+    },
+    {
+      "epoch": 42.14928057553957,
+      "step": 46870,
+      "torque_loss": 0.12692494690418243
+    },
+    {
+      "epoch": 42.14928057553957,
+      "force_loss": 0.006886093411594629,
+      "step": 46870
+    },
+    {
+      "epoch": 42.15827338129496,
+      "grad_norm": 0.11866920441389084,
+      "learning_rate": 1.2514500162305087e-05,
+      "loss": 0.0145,
+      "step": 46880
+    },
+    {
+      "action_loss": 0.0020604028832167387,
+      "epoch": 42.15827338129496,
+      "step": 46880
+    },
+    {
+      "epoch": 42.15827338129496,
+      "step": 46880,
+      "torque_loss": 0.10279615968465805
+    },
+    {
+      "epoch": 42.15827338129496,
+      "force_loss": 0.00376189430244267,
+      "step": 46880
+    },
+    {
+      "epoch": 42.16726618705036,
+      "grad_norm": 0.07028356939554214,
+      "learning_rate": 1.2496269007457728e-05,
+      "loss": 0.0155,
+      "step": 46890
+    },
+    {
+      "action_loss": 0.002010742900893092,
+      "epoch": 42.16726618705036,
+      "step": 46890
+    },
+    {
+      "epoch": 42.16726618705036,
+      "step": 46890,
+      "torque_loss": 0.1491326540708542
+    },
+    {
+      "epoch": 42.16726618705036,
+      "force_loss": 0.002597041428089142,
+      "step": 46890
+    },
+    {
+      "epoch": 42.17625899280576,
+      "grad_norm": 0.11444475501775742,
+      "learning_rate": 1.2478049245253625e-05,
+      "loss": 0.0146,
+      "step": 46900
+    },
+    {
+      "action_loss": 0.0010552414460107684,
+      "epoch": 42.17625899280576,
+      "step": 46900
+    },
+    {
+      "epoch": 42.17625899280576,
+      "step": 46900,
+      "torque_loss": 0.09426501393318176
+    },
+    {
+      "epoch": 42.17625899280576,
+      "force_loss": 0.002343730768188834,
+      "step": 46900
+    },
+    {
+      "epoch": 42.185251798561154,
+      "grad_norm": 0.33438387513160706,
+      "learning_rate": 1.2459840881227459e-05,
+      "loss": 0.017,
+      "step": 46910
+    },
+    {
+      "action_loss": 0.0014436669880524278,
+      "epoch": 42.185251798561154,
+      "step": 46910
+    },
+    {
+      "epoch": 42.185251798561154,
+      "step": 46910,
+      "torque_loss": 0.10698408633470535
+    },
+    {
+      "epoch": 42.185251798561154,
+      "force_loss": 0.002083592815324664,
+      "step": 46910
+    },
+    {
+      "epoch": 42.194244604316545,
+      "grad_norm": 0.0719999298453331,
+      "learning_rate": 1.2441643920910435e-05,
+      "loss": 0.0162,
+      "step": 46920
+    },
+    {
+      "action_loss": 0.0013842815533280373,
+      "epoch": 42.194244604316545,
+      "step": 46920
+    },
+    {
+      "epoch": 42.194244604316545,
+      "step": 46920,
+      "torque_loss": 0.12207833677530289
+    },
+    {
+      "epoch": 42.194244604316545,
+      "force_loss": 0.0017060752725228667,
+      "step": 46920
+    },
+    {
+      "epoch": 42.20323741007194,
+      "grad_norm": 0.10483632236719131,
+      "learning_rate": 1.2423458369830322e-05,
+      "loss": 0.0138,
+      "step": 46930
+    },
+    {
+      "action_loss": 0.004464512690901756,
+      "epoch": 42.20323741007194,
+      "step": 46930
+    },
+    {
+      "epoch": 42.20323741007194,
+      "step": 46930,
+      "torque_loss": 0.1396876573562622
+    },
+    {
+      "epoch": 42.20323741007194,
+      "force_loss": 0.007298026233911514,
+      "step": 46930
+    },
+    {
+      "epoch": 42.21223021582734,
+      "grad_norm": 0.10816202312707901,
+      "learning_rate": 1.2405284233511406e-05,
+      "loss": 0.0172,
+      "step": 46940
+    },
+    {
+      "action_loss": 0.010293533094227314,
+      "epoch": 42.21223021582734,
+      "step": 46940
+    },
+    {
+      "epoch": 42.21223021582734,
+      "step": 46940,
+      "torque_loss": 0.1770031452178955
+    },
+    {
+      "epoch": 42.21223021582734,
+      "force_loss": 0.01605510711669922,
+      "step": 46940
+    },
+    {
+      "epoch": 42.22122302158273,
+      "grad_norm": 0.272856742143631,
+      "learning_rate": 1.2387121517474487e-05,
+      "loss": 0.0166,
+      "step": 46950
+    },
+    {
+      "action_loss": 0.001577790710143745,
+      "epoch": 42.22122302158273,
+      "step": 46950
+    },
+    {
+      "epoch": 42.22122302158273,
+      "step": 46950,
+      "torque_loss": 0.12595312297344208
+    },
+    {
+      "epoch": 42.22122302158273,
+      "force_loss": 0.004614649806171656,
+      "step": 46950
+    },
+    {
+      "epoch": 42.23021582733813,
+      "grad_norm": 0.17100203037261963,
+      "learning_rate": 1.2368970227236975e-05,
+      "loss": 0.0176,
+      "step": 46960
+    },
+    {
+      "action_loss": 0.0012894395040348172,
+      "epoch": 42.23021582733813,
+      "step": 46960
+    },
+    {
+      "epoch": 42.23021582733813,
+      "step": 46960,
+      "torque_loss": 0.11722856760025024
+    },
+    {
+      "epoch": 42.23021582733813,
+      "force_loss": 0.0024467355106025934,
+      "step": 46960
+    },
+    {
+      "epoch": 42.239208633093526,
+      "grad_norm": 0.14707109332084656,
+      "learning_rate": 1.2350830368312688e-05,
+      "loss": 0.0158,
+      "step": 46970
+    },
+    {
+      "action_loss": 0.0006690708105452359,
+      "epoch": 42.239208633093526,
+      "step": 46970
+    },
+    {
+      "epoch": 42.239208633093526,
+      "step": 46970,
+      "torque_loss": 0.07282324135303497
+    },
+    {
+      "epoch": 42.239208633093526,
+      "force_loss": 0.000768433790653944,
+      "step": 46970
+    },
+    {
+      "epoch": 42.24820143884892,
+      "grad_norm": 0.28848797082901,
+      "learning_rate": 1.2332701946212083e-05,
+      "loss": 0.0153,
+      "step": 46980
+    },
+    {
+      "action_loss": 0.0026538728270679712,
+      "epoch": 42.24820143884892,
+      "step": 46980
+    },
+    {
+      "epoch": 42.24820143884892,
+      "step": 46980,
+      "torque_loss": 0.08095107227563858
+    },
+    {
+      "epoch": 42.24820143884892,
+      "force_loss": 0.006578672677278519,
+      "step": 46980
+    },
+    {
+      "epoch": 42.257194244604314,
+      "grad_norm": 0.36609283089637756,
+      "learning_rate": 1.2314584966442077e-05,
+      "loss": 0.0149,
+      "step": 46990
+    },
+    {
+      "action_loss": 0.004711670335382223,
+      "epoch": 42.257194244604314,
+      "step": 46990
+    },
+    {
+      "epoch": 42.257194244604314,
+      "step": 46990,
+      "torque_loss": 0.09482771158218384
+    },
+    {
+      "epoch": 42.257194244604314,
+      "force_loss": 0.013062643818557262,
+      "step": 46990
+    },
+    {
+      "epoch": 42.26618705035971,
+      "grad_norm": 0.07432954013347626,
+      "learning_rate": 1.2296479434506136e-05,
+      "loss": 0.0163,
+      "step": 47000
+    },
+    {
+      "action_loss": 0.0012593042338266969,
+      "epoch": 42.26618705035971,
+      "step": 47000
+    },
+    {
+      "epoch": 42.26618705035971,
+      "step": 47000,
+      "torque_loss": 0.08349712938070297
+    },
+    {
+      "epoch": 42.26618705035971,
+      "force_loss": 0.004426718223839998,
+      "step": 47000
+    },
+    {
+      "epoch": 42.27517985611511,
+      "grad_norm": 0.1788962036371231,
+      "learning_rate": 1.2278385355904232e-05,
+      "loss": 0.0149,
+      "step": 47010
+    },
+    {
+      "action_loss": 0.0008733691647648811,
+      "epoch": 42.27517985611511,
+      "step": 47010
+    },
+    {
+      "epoch": 42.27517985611511,
+      "step": 47010,
+      "torque_loss": 0.12246324867010117
+    },
+    {
+      "epoch": 42.27517985611511,
+      "force_loss": 0.0005289769032970071,
+      "step": 47010
+    },
+    {
+      "epoch": 42.28417266187051,
+      "grad_norm": 0.43475180864334106,
+      "learning_rate": 1.2260302736132867e-05,
+      "loss": 0.0151,
+      "step": 47020
+    },
+    {
+      "action_loss": 0.0011639531003311276,
+      "epoch": 42.28417266187051,
+      "step": 47020
+    },
+    {
+      "epoch": 42.28417266187051,
+      "step": 47020,
+      "torque_loss": 0.10735426098108292
+    },
+    {
+      "epoch": 42.28417266187051,
+      "force_loss": 0.0036427301820367575,
+      "step": 47020
+    },
+    {
+      "epoch": 42.2931654676259,
+      "grad_norm": 0.09938303381204605,
+      "learning_rate": 1.2242231580685098e-05,
+      "loss": 0.0175,
+      "step": 47030
+    },
+    {
+      "action_loss": 0.003048837184906006,
+      "epoch": 42.2931654676259,
+      "step": 47030
+    },
+    {
+      "epoch": 42.2931654676259,
+      "step": 47030,
+      "torque_loss": 0.14806193113327026
+    },
+    {
+      "epoch": 42.2931654676259,
+      "force_loss": 0.0018813660135492682,
+      "step": 47030
+    },
+    {
+      "epoch": 42.302158273381295,
+      "grad_norm": 0.09749345481395721,
+      "learning_rate": 1.2224171895050413e-05,
+      "loss": 0.0153,
+      "step": 47040
+    },
+    {
+      "action_loss": 0.006674807518720627,
+      "epoch": 42.302158273381295,
+      "step": 47040
+    },
+    {
+      "epoch": 42.302158273381295,
+      "step": 47040,
+      "torque_loss": 0.09648054838180542
+    },
+    {
+      "epoch": 42.302158273381295,
+      "force_loss": 0.010812963359057903,
+      "step": 47040
+    },
+    {
+      "epoch": 42.31115107913669,
+      "grad_norm": 0.3661826252937317,
+      "learning_rate": 1.2206123684714903e-05,
+      "loss": 0.0154,
+      "step": 47050
+    },
+    {
+      "action_loss": 0.0029531989712268114,
+      "epoch": 42.31115107913669,
+      "step": 47050
+    },
+    {
+      "epoch": 42.31115107913669,
+      "step": 47050,
+      "torque_loss": 0.17020951211452484
+    },
+    {
+      "epoch": 42.31115107913669,
+      "force_loss": 0.018691852688789368,
+      "step": 47050
+    },
+    {
+      "epoch": 42.32014388489208,
+      "grad_norm": 0.2752850353717804,
+      "learning_rate": 1.2188086955161132e-05,
+      "loss": 0.0145,
+      "step": 47060
+    },
+    {
+      "action_loss": 0.0015264186076819897,
+      "epoch": 42.32014388489208,
+      "step": 47060
+    },
+    {
+      "epoch": 42.32014388489208,
+      "step": 47060,
+      "torque_loss": 0.10866586118936539
+    },
+    {
+      "epoch": 42.32014388489208,
+      "force_loss": 0.0016908334800973535,
+      "step": 47060
+    },
+    {
+      "epoch": 42.32913669064748,
+      "grad_norm": 0.10309295356273651,
+      "learning_rate": 1.2170061711868175e-05,
+      "loss": 0.0157,
+      "step": 47070
+    },
+    {
+      "action_loss": 0.0029808704275637865,
+      "epoch": 42.32913669064748,
+      "step": 47070
+    },
+    {
+      "epoch": 42.32913669064748,
+      "step": 47070,
+      "torque_loss": 0.10860654711723328
+    },
+    {
+      "epoch": 42.32913669064748,
+      "force_loss": 0.010925118811428547,
+      "step": 47070
+    },
+    {
+      "epoch": 42.33812949640288,
+      "grad_norm": 0.20670431852340698,
+      "learning_rate": 1.215204796031163e-05,
+      "loss": 0.0145,
+      "step": 47080
+    },
+    {
+      "action_loss": 0.0036128342617303133,
+      "epoch": 42.33812949640288,
+      "step": 47080
+    },
+    {
+      "epoch": 42.33812949640288,
+      "step": 47080,
+      "torque_loss": 0.1813892275094986
+    },
+    {
+      "epoch": 42.33812949640288,
+      "force_loss": 0.018829522654414177,
+      "step": 47080
+    },
+    {
+      "epoch": 42.347122302158276,
+      "grad_norm": 0.05608697608113289,
+      "learning_rate": 1.2134045705963599e-05,
+      "loss": 0.0142,
+      "step": 47090
+    },
+    {
+      "action_loss": 0.0007133810431696475,
+      "epoch": 42.347122302158276,
+      "step": 47090
+    },
+    {
+      "epoch": 42.347122302158276,
+      "step": 47090,
+      "torque_loss": 0.0937839224934578
+    },
+    {
+      "epoch": 42.347122302158276,
+      "force_loss": 0.0009587321546860039,
+      "step": 47090
+    },
+    {
+      "epoch": 42.356115107913666,
+      "grad_norm": 0.10110324621200562,
+      "learning_rate": 1.2116054954292689e-05,
+      "loss": 0.0128,
+      "step": 47100
+    },
+    {
+      "action_loss": 0.001186329871416092,
+      "epoch": 42.356115107913666,
+      "step": 47100
+    },
+    {
+      "epoch": 42.356115107913666,
+      "step": 47100,
+      "torque_loss": 0.12169986963272095
+    },
+    {
+      "epoch": 42.356115107913666,
+      "force_loss": 0.0014533005887642503,
+      "step": 47100
+    },
+    {
+      "epoch": 42.365107913669064,
+      "grad_norm": 0.2030871957540512,
+      "learning_rate": 1.2098075710764011e-05,
+      "loss": 0.0157,
+      "step": 47110
+    },
+    {
+      "action_loss": 0.004844674840569496,
+      "epoch": 42.365107913669064,
+      "step": 47110
+    },
+    {
+      "epoch": 42.365107913669064,
+      "step": 47110,
+      "torque_loss": 0.1532885581254959
+    },
+    {
+      "epoch": 42.365107913669064,
+      "force_loss": 0.008660007268190384,
+      "step": 47110
+    },
+    {
+      "epoch": 42.37410071942446,
+      "grad_norm": 0.09478098899126053,
+      "learning_rate": 1.2080107980839183e-05,
+      "loss": 0.0157,
+      "step": 47120
+    },
+    {
+      "action_loss": 0.0007692959625273943,
+      "epoch": 42.37410071942446,
+      "step": 47120
+    },
+    {
+      "epoch": 42.37410071942446,
+      "step": 47120,
+      "torque_loss": 0.09850936383008957
+    },
+    {
+      "epoch": 42.37410071942446,
+      "force_loss": 0.0012820689007639885,
+      "step": 47120
+    },
+    {
+      "epoch": 42.38309352517986,
+      "grad_norm": 0.14071500301361084,
+      "learning_rate": 1.2062151769976343e-05,
+      "loss": 0.0174,
+      "step": 47130
+    },
+    {
+      "action_loss": 0.0007247214671224356,
+      "epoch": 42.38309352517986,
+      "step": 47130
+    },
+    {
+      "epoch": 42.38309352517986,
+      "step": 47130,
+      "torque_loss": 0.12936072051525116
+    },
+    {
+      "epoch": 42.38309352517986,
+      "force_loss": 0.0009547374211251736,
+      "step": 47130
+    },
+    {
+      "epoch": 42.39208633093525,
+      "grad_norm": 0.07789991050958633,
+      "learning_rate": 1.204420708363011e-05,
+      "loss": 0.0153,
+      "step": 47140
+    },
+    {
+      "action_loss": 0.0015309160808101296,
+      "epoch": 42.39208633093525,
+      "step": 47140
+    },
+    {
+      "epoch": 42.39208633093525,
+      "step": 47140,
+      "torque_loss": 0.08339641243219376
+    },
+    {
+      "epoch": 42.39208633093525,
+      "force_loss": 0.0027038194239139557,
+      "step": 47140
+    },
+    {
+      "epoch": 42.40107913669065,
+      "grad_norm": 0.07954797893762589,
+      "learning_rate": 1.2026273927251597e-05,
+      "loss": 0.0155,
+      "step": 47150
+    },
+    {
+      "action_loss": 0.0012581819901242852,
+      "epoch": 42.40107913669065,
+      "step": 47150
+    },
+    {
+      "epoch": 42.40107913669065,
+      "step": 47150,
+      "torque_loss": 0.1861734837293625
+    },
+    {
+      "epoch": 42.40107913669065,
+      "force_loss": 0.001045328681357205,
+      "step": 47150
+    },
+    {
+      "epoch": 42.410071942446045,
+      "grad_norm": 0.32643038034439087,
+      "learning_rate": 1.2008352306288424e-05,
+      "loss": 0.0158,
+      "step": 47160
+    },
+    {
+      "action_loss": 0.0018418746767565608,
+      "epoch": 42.410071942446045,
+      "step": 47160
+    },
+    {
+      "epoch": 42.410071942446045,
+      "step": 47160,
+      "torque_loss": 0.08608373999595642
+    },
+    {
+      "epoch": 42.410071942446045,
+      "force_loss": 0.0071082995273172855,
+      "step": 47160
+    },
+    {
+      "epoch": 42.419064748201436,
+      "grad_norm": 0.26478394865989685,
+      "learning_rate": 1.1990442226184695e-05,
+      "loss": 0.0132,
+      "step": 47170
+    },
+    {
+      "action_loss": 0.0007814674754627049,
+      "epoch": 42.419064748201436,
+      "step": 47170
+    },
+    {
+      "epoch": 42.419064748201436,
+      "step": 47170,
+      "torque_loss": 0.14269594848155975
+    },
+    {
+      "epoch": 42.419064748201436,
+      "force_loss": 0.0013496694155037403,
+      "step": 47170
+    },
+    {
+      "epoch": 42.42805755395683,
+      "grad_norm": 0.36180445551872253,
+      "learning_rate": 1.1972543692381066e-05,
+      "loss": 0.0165,
+      "step": 47180
+    },
+    {
+      "action_loss": 0.005655744578689337,
+      "epoch": 42.42805755395683,
+      "step": 47180
+    },
+    {
+      "epoch": 42.42805755395683,
+      "step": 47180,
+      "torque_loss": 0.14706192910671234
+    },
+    {
+      "epoch": 42.42805755395683,
+      "force_loss": 0.0059926752001047134,
+      "step": 47180
+    },
+    {
+      "epoch": 42.43705035971223,
+      "grad_norm": 0.28917908668518066,
+      "learning_rate": 1.1954656710314576e-05,
+      "loss": 0.0168,
+      "step": 47190
+    },
+    {
+      "action_loss": 0.000995397916994989,
+      "epoch": 42.43705035971223,
+      "step": 47190
+    },
+    {
+      "epoch": 42.43705035971223,
+      "step": 47190,
+      "torque_loss": 0.06930605322122574
+    },
+    {
+      "epoch": 42.43705035971223,
+      "force_loss": 0.007246797438710928,
+      "step": 47190
+    },
+    {
+      "epoch": 42.44604316546763,
+      "grad_norm": 0.1822378784418106,
+      "learning_rate": 1.1936781285418875e-05,
+      "loss": 0.0144,
+      "step": 47200
+    },
+    {
+      "action_loss": 0.0010932345176115632,
+      "epoch": 42.44604316546763,
+      "step": 47200
+    },
+    {
+      "epoch": 42.44604316546763,
+      "step": 47200,
+      "torque_loss": 0.14512300491333008
+    },
+    {
+      "epoch": 42.44604316546763,
+      "force_loss": 0.0010726271430030465,
+      "step": 47200
+    },
+    {
+      "epoch": 42.45503597122302,
+      "grad_norm": 0.07754116505384445,
+      "learning_rate": 1.1918917423123993e-05,
+      "loss": 0.0143,
+      "step": 47210
+    },
+    {
+      "action_loss": 0.00236272974871099,
+      "epoch": 42.45503597122302,
+      "step": 47210
+    },
+    {
+      "epoch": 42.45503597122302,
+      "step": 47210,
+      "torque_loss": 0.12109533697366714
+    },
+    {
+      "epoch": 42.45503597122302,
+      "force_loss": 0.010092931799590588,
+      "step": 47210
+    },
+    {
+      "epoch": 42.46402877697842,
+      "grad_norm": 0.36079341173171997,
+      "learning_rate": 1.1901065128856537e-05,
+      "loss": 0.0175,
+      "step": 47220
+    },
+    {
+      "action_loss": 0.0034934638533741236,
+      "epoch": 42.46402877697842,
+      "step": 47220
+    },
+    {
+      "epoch": 42.46402877697842,
+      "step": 47220,
+      "torque_loss": 0.08428329974412918
+    },
+    {
+      "epoch": 42.46402877697842,
+      "force_loss": 0.004261286463588476,
+      "step": 47220
+    },
+    {
+      "epoch": 42.473021582733814,
+      "grad_norm": 0.07354458421468735,
+      "learning_rate": 1.1883224408039551e-05,
+      "loss": 0.0149,
+      "step": 47230
+    },
+    {
+      "action_loss": 0.0019940249621868134,
+      "epoch": 42.473021582733814,
+      "step": 47230
+    },
+    {
+      "epoch": 42.473021582733814,
+      "step": 47230,
+      "torque_loss": 0.12210458517074585
+    },
+    {
+      "epoch": 42.473021582733814,
+      "force_loss": 0.002284869784489274,
+      "step": 47230
+    },
+    {
+      "epoch": 42.48201438848921,
+      "grad_norm": 0.3708144426345825,
+      "learning_rate": 1.1865395266092578e-05,
+      "loss": 0.0183,
+      "step": 47240
+    },
+    {
+      "action_loss": 0.0005899166571907699,
+      "epoch": 42.48201438848921,
+      "step": 47240
+    },
+    {
+      "epoch": 42.48201438848921,
+      "step": 47240,
+      "torque_loss": 0.07279258221387863
+    },
+    {
+      "epoch": 42.48201438848921,
+      "force_loss": 0.0008377400226891041,
+      "step": 47240
+    },
+    {
+      "epoch": 42.4910071942446,
+      "grad_norm": 0.12171059846878052,
+      "learning_rate": 1.1847577708431633e-05,
+      "loss": 0.0164,
+      "step": 47250
+    },
+    {
+      "action_loss": 0.001660170848481357,
+      "epoch": 42.4910071942446,
+      "step": 47250
+    },
+    {
+      "epoch": 42.4910071942446,
+      "step": 47250,
+      "torque_loss": 0.12421798706054688
+    },
+    {
+      "epoch": 42.4910071942446,
+      "force_loss": 0.004743021447211504,
+      "step": 47250
+    },
+    {
+      "epoch": 42.5,
+      "grad_norm": 0.1717991977930069,
+      "learning_rate": 1.1829771740469225e-05,
+      "loss": 0.0139,
+      "step": 47260
+    },
+    {
+      "action_loss": 0.0016320167342200875,
+      "epoch": 42.5,
+      "step": 47260
+    },
+    {
+      "epoch": 42.5,
+      "step": 47260,
+      "torque_loss": 0.1975756287574768
+    },
+    {
+      "epoch": 42.5,
+      "force_loss": 0.0021180794574320316,
+      "step": 47260
+    },
+    {
+      "epoch": 42.5089928057554,
+      "grad_norm": 0.08667226135730743,
+      "learning_rate": 1.1811977367614324e-05,
+      "loss": 0.0151,
+      "step": 47270
+    },
+    {
+      "action_loss": 0.0016210117610171437,
+      "epoch": 42.5089928057554,
+      "step": 47270
+    },
+    {
+      "epoch": 42.5089928057554,
+      "step": 47270,
+      "torque_loss": 0.11727314442396164
+    },
+    {
+      "epoch": 42.5089928057554,
+      "force_loss": 0.003396612824872136,
+      "step": 47270
+    },
+    {
+      "epoch": 42.51798561151079,
+      "grad_norm": 0.07232655584812164,
+      "learning_rate": 1.1794194595272412e-05,
+      "loss": 0.0235,
+      "step": 47280
+    },
+    {
+      "action_loss": 0.00337769091129303,
+      "epoch": 42.51798561151079,
+      "step": 47280
+    },
+    {
+      "epoch": 42.51798561151079,
+      "step": 47280,
+      "torque_loss": 0.13087229430675507
+    },
+    {
+      "epoch": 42.51798561151079,
+      "force_loss": 0.009339473210275173,
+      "step": 47280
+    },
+    {
+      "epoch": 42.526978417266186,
+      "grad_norm": 0.30817991495132446,
+      "learning_rate": 1.1776423428845423e-05,
+      "loss": 0.0137,
+      "step": 47290
+    },
+    {
+      "action_loss": 0.013060093857347965,
+      "epoch": 42.526978417266186,
+      "step": 47290
+    },
+    {
+      "epoch": 42.526978417266186,
+      "step": 47290,
+      "torque_loss": 0.13149075210094452
+    },
+    {
+      "epoch": 42.526978417266186,
+      "force_loss": 0.033541012555360794,
+      "step": 47290
+    },
+    {
+      "epoch": 42.53597122302158,
+      "grad_norm": 0.07581840455532074,
+      "learning_rate": 1.1758663873731756e-05,
+      "loss": 0.018,
+      "step": 47300
+    },
+    {
+      "action_loss": 0.0008062750566750765,
+      "epoch": 42.53597122302158,
+      "step": 47300
+    },
+    {
+      "epoch": 42.53597122302158,
+      "step": 47300,
+      "torque_loss": 0.09492490440607071
+    },
+    {
+      "epoch": 42.53597122302158,
+      "force_loss": 0.001890915329568088,
+      "step": 47300
+    },
+    {
+      "epoch": 42.54496402877698,
+      "grad_norm": 0.16609393060207367,
+      "learning_rate": 1.1740915935326302e-05,
+      "loss": 0.0209,
+      "step": 47310
+    },
+    {
+      "action_loss": 0.0017578349215909839,
+      "epoch": 42.54496402877698,
+      "step": 47310
+    },
+    {
+      "epoch": 42.54496402877698,
+      "step": 47310,
+      "torque_loss": 0.10964155942201614
+    },
+    {
+      "epoch": 42.54496402877698,
+      "force_loss": 0.0014121043495833874,
+      "step": 47310
+    },
+    {
+      "epoch": 42.55395683453237,
+      "grad_norm": 0.25634074211120605,
+      "learning_rate": 1.1723179619020396e-05,
+      "loss": 0.0139,
+      "step": 47320
+    },
+    {
+      "action_loss": 0.0015162843046709895,
+      "epoch": 42.55395683453237,
+      "step": 47320
+    },
+    {
+      "epoch": 42.55395683453237,
+      "step": 47320,
+      "torque_loss": 0.1422034651041031
+    },
+    {
+      "epoch": 42.55395683453237,
+      "force_loss": 0.00577051704749465,
+      "step": 47320
+    },
+    {
+      "epoch": 42.56294964028777,
+      "grad_norm": 0.2171502560377121,
+      "learning_rate": 1.1705454930201914e-05,
+      "loss": 0.0157,
+      "step": 47330
+    },
+    {
+      "action_loss": 0.011758401058614254,
+      "epoch": 42.56294964028777,
+      "step": 47330
+    },
+    {
+      "epoch": 42.56294964028777,
+      "step": 47330,
+      "torque_loss": 0.1420688033103943
+    },
+    {
+      "epoch": 42.56294964028777,
+      "force_loss": 0.019480887800455093,
+      "step": 47330
+    },
+    {
+      "epoch": 42.57194244604317,
+      "grad_norm": 0.09375839680433273,
+      "learning_rate": 1.1687741874255087e-05,
+      "loss": 0.0175,
+      "step": 47340
+    },
+    {
+      "action_loss": 0.0005921342526562512,
+      "epoch": 42.57194244604317,
+      "step": 47340
+    },
+    {
+      "epoch": 42.57194244604317,
+      "step": 47340,
+      "torque_loss": 0.11464810371398926
+    },
+    {
+      "epoch": 42.57194244604317,
+      "force_loss": 0.0009963855845853686,
+      "step": 47340
+    },
+    {
+      "epoch": 42.580935251798564,
+      "grad_norm": 0.10439424216747284,
+      "learning_rate": 1.1670040456560728e-05,
+      "loss": 0.0137,
+      "step": 47350
+    },
+    {
+      "action_loss": 0.0026397760957479477,
+      "epoch": 42.580935251798564,
+      "step": 47350
+    },
+    {
+      "epoch": 42.580935251798564,
+      "step": 47350,
+      "torque_loss": 0.1713700294494629
+    },
+    {
+      "epoch": 42.580935251798564,
+      "force_loss": 0.008073628880083561,
+      "step": 47350
+    },
+    {
+      "epoch": 42.589928057553955,
+      "grad_norm": 0.4963211119174957,
+      "learning_rate": 1.1652350682496005e-05,
+      "loss": 0.0183,
+      "step": 47360
+    },
+    {
+      "action_loss": 0.0018448602641001344,
+      "epoch": 42.589928057553955,
+      "step": 47360
+    },
+    {
+      "epoch": 42.589928057553955,
+      "step": 47360,
+      "torque_loss": 0.10342096537351608
+    },
+    {
+      "epoch": 42.589928057553955,
+      "force_loss": 0.0017703211633488536,
+      "step": 47360
+    },
+    {
+      "epoch": 42.59892086330935,
+      "grad_norm": 0.07963874936103821,
+      "learning_rate": 1.163467255743465e-05,
+      "loss": 0.014,
+      "step": 47370
+    },
+    {
+      "action_loss": 0.002963275881484151,
+      "epoch": 42.59892086330935,
+      "step": 47370
+    },
+    {
+      "epoch": 42.59892086330935,
+      "step": 47370,
+      "torque_loss": 0.09060266613960266
+    },
+    {
+      "epoch": 42.59892086330935,
+      "force_loss": 0.0015868988120928407,
+      "step": 47370
+    },
+    {
+      "epoch": 42.60791366906475,
+      "grad_norm": 0.07791832834482193,
+      "learning_rate": 1.1617006086746796e-05,
+      "loss": 0.0151,
+      "step": 47380
+    },
+    {
+      "action_loss": 0.013082985766232014,
+      "epoch": 42.60791366906475,
+      "step": 47380
+    },
+    {
+      "epoch": 42.60791366906475,
+      "step": 47380,
+      "torque_loss": 0.1579044908285141
+    },
+    {
+      "epoch": 42.60791366906475,
+      "force_loss": 0.012325276620686054,
+      "step": 47380
+    },
+    {
+      "epoch": 42.61690647482014,
+      "grad_norm": 0.1890263557434082,
+      "learning_rate": 1.1599351275799047e-05,
+      "loss": 0.0171,
+      "step": 47390
+    },
+    {
+      "action_loss": 0.0037095455918461084,
+      "epoch": 42.61690647482014,
+      "step": 47390
+    },
+    {
+      "epoch": 42.61690647482014,
+      "step": 47390,
+      "torque_loss": 0.09953423589468002
+    },
+    {
+      "epoch": 42.61690647482014,
+      "force_loss": 0.014167613349854946,
+      "step": 47390
+    },
+    {
+      "epoch": 42.62589928057554,
+      "grad_norm": 0.05807991698384285,
+      "learning_rate": 1.1581708129954466e-05,
+      "loss": 0.013,
+      "step": 47400
+    },
+    {
+      "action_loss": 0.0013249530456960201,
+      "epoch": 42.62589928057554,
+      "step": 47400
+    },
+    {
+      "epoch": 42.62589928057554,
+      "step": 47400,
+      "torque_loss": 0.08267392963171005
+    },
+    {
+      "epoch": 42.62589928057554,
+      "force_loss": 0.003339221002534032,
+      "step": 47400
+    },
+    {
+      "epoch": 42.634892086330936,
+      "grad_norm": 0.08891882002353668,
+      "learning_rate": 1.1564076654572587e-05,
+      "loss": 0.0137,
+      "step": 47410
+    },
+    {
+      "action_loss": 0.0017964104190468788,
+      "epoch": 42.634892086330936,
+      "step": 47410
+    },
+    {
+      "epoch": 42.634892086330936,
+      "step": 47410,
+      "torque_loss": 0.1630369871854782
+    },
+    {
+      "epoch": 42.634892086330936,
+      "force_loss": 0.0049766856245696545,
+      "step": 47410
+    },
+    {
+      "epoch": 42.643884892086334,
+      "grad_norm": 0.09315340965986252,
+      "learning_rate": 1.1546456855009358e-05,
+      "loss": 0.0153,
+      "step": 47420
+    },
+    {
+      "action_loss": 0.0008421282400377095,
+      "epoch": 42.643884892086334,
+      "step": 47420
+    },
+    {
+      "epoch": 42.643884892086334,
+      "step": 47420,
+      "torque_loss": 0.1481899619102478
+    },
+    {
+      "epoch": 42.643884892086334,
+      "force_loss": 0.001376375905238092,
+      "step": 47420
+    },
+    {
+      "epoch": 42.652877697841724,
+      "grad_norm": 0.1282106190919876,
+      "learning_rate": 1.1528848736617248e-05,
+      "loss": 0.0165,
+      "step": 47430
+    },
+    {
+      "action_loss": 0.0007081867661327124,
+      "epoch": 42.652877697841724,
+      "step": 47430
+    },
+    {
+      "epoch": 42.652877697841724,
+      "step": 47430,
+      "torque_loss": 0.08617168664932251
+    },
+    {
+      "epoch": 42.652877697841724,
+      "force_loss": 0.0006583284121006727,
+      "step": 47430
+    },
+    {
+      "epoch": 42.66187050359712,
+      "grad_norm": 0.14328855276107788,
+      "learning_rate": 1.1511252304745112e-05,
+      "loss": 0.017,
+      "step": 47440
+    },
+    {
+      "action_loss": 0.036643579602241516,
+      "epoch": 42.66187050359712,
+      "step": 47440
+    },
+    {
+      "epoch": 42.66187050359712,
+      "step": 47440,
+      "torque_loss": 0.15435637533664703
+    },
+    {
+      "epoch": 42.66187050359712,
+      "force_loss": 0.05480148270726204,
+      "step": 47440
+    },
+    {
+      "epoch": 42.67086330935252,
+      "grad_norm": 0.1529199630022049,
+      "learning_rate": 1.1493667564738297e-05,
+      "loss": 0.0265,
+      "step": 47450
+    },
+    {
+      "action_loss": 0.006078500300645828,
+      "epoch": 42.67086330935252,
+      "step": 47450
+    },
+    {
+      "epoch": 42.67086330935252,
+      "step": 47450,
+      "torque_loss": 0.11079076677560806
+    },
+    {
+      "epoch": 42.67086330935252,
+      "force_loss": 0.01721310243010521,
+      "step": 47450
+    },
+    {
+      "epoch": 42.67985611510792,
+      "grad_norm": 0.06626751273870468,
+      "learning_rate": 1.1476094521938574e-05,
+      "loss": 0.0138,
+      "step": 47460
+    },
+    {
+      "action_loss": 0.000707072380464524,
+      "epoch": 42.67985611510792,
+      "step": 47460
+    },
+    {
+      "epoch": 42.67985611510792,
+      "step": 47460,
+      "torque_loss": 0.08428502082824707
+    },
+    {
+      "epoch": 42.67985611510792,
+      "force_loss": 0.0025898588355630636,
+      "step": 47460
+    },
+    {
+      "epoch": 42.68884892086331,
+      "grad_norm": 0.12890978157520294,
+      "learning_rate": 1.1458533181684167e-05,
+      "loss": 0.0148,
+      "step": 47470
+    },
+    {
+      "action_loss": 0.0006818519905209541,
+      "epoch": 42.68884892086331,
+      "step": 47470
+    },
+    {
+      "epoch": 42.68884892086331,
+      "step": 47470,
+      "torque_loss": 0.14568443596363068
+    },
+    {
+      "epoch": 42.68884892086331,
+      "force_loss": 0.0012362306006252766,
+      "step": 47470
+    },
+    {
+      "epoch": 42.697841726618705,
+      "grad_norm": 0.0929703414440155,
+      "learning_rate": 1.1440983549309753e-05,
+      "loss": 0.0152,
+      "step": 47480
+    },
+    {
+      "action_loss": 0.0008459368837065995,
+      "epoch": 42.697841726618705,
+      "step": 47480
+    },
+    {
+      "epoch": 42.697841726618705,
+      "step": 47480,
+      "torque_loss": 0.12080162763595581
+    },
+    {
+      "epoch": 42.697841726618705,
+      "force_loss": 0.001601852010935545,
+      "step": 47480
+    },
+    {
+      "epoch": 42.7068345323741,
+      "grad_norm": 0.15074065327644348,
+      "learning_rate": 1.1423445630146434e-05,
+      "loss": 0.0142,
+      "step": 47490
+    },
+    {
+      "action_loss": 0.002071307273581624,
+      "epoch": 42.7068345323741,
+      "step": 47490
+    },
+    {
+      "epoch": 42.7068345323741,
+      "step": 47490,
+      "torque_loss": 0.11533572524785995
+    },
+    {
+      "epoch": 42.7068345323741,
+      "force_loss": 0.0038288019131869078,
+      "step": 47490
+    },
+    {
+      "epoch": 42.71582733812949,
+      "grad_norm": 0.15697233378887177,
+      "learning_rate": 1.1405919429521799e-05,
+      "loss": 0.0195,
+      "step": 47500
+    },
+    {
+      "action_loss": 0.0005225121858529747,
+      "epoch": 42.71582733812949,
+      "step": 47500
+    },
+    {
+      "epoch": 42.71582733812949,
+      "step": 47500,
+      "torque_loss": 0.08119718730449677
+    },
+    {
+      "epoch": 42.71582733812949,
+      "force_loss": 0.00042487154132686555,
+      "step": 47500
+    },
+    {
+      "epoch": 42.72482014388489,
+      "grad_norm": 0.09051039069890976,
+      "learning_rate": 1.1388404952759802e-05,
+      "loss": 0.0158,
+      "step": 47510
+    },
+    {
+      "action_loss": 0.0012061764718964696,
+      "epoch": 42.72482014388489,
+      "step": 47510
+    },
+    {
+      "epoch": 42.72482014388489,
+      "step": 47510,
+      "torque_loss": 0.11353711038827896
+    },
+    {
+      "epoch": 42.72482014388489,
+      "force_loss": 0.001552483648993075,
+      "step": 47510
+    },
+    {
+      "epoch": 42.73381294964029,
+      "grad_norm": 0.06365683674812317,
+      "learning_rate": 1.1370902205180923e-05,
+      "loss": 0.015,
+      "step": 47520
+    },
+    {
+      "action_loss": 0.0011349001433700323,
+      "epoch": 42.73381294964029,
+      "step": 47520
+    },
+    {
+      "epoch": 42.73381294964029,
+      "step": 47520,
+      "torque_loss": 0.12694470584392548
+    },
+    {
+      "epoch": 42.73381294964029,
+      "force_loss": 0.004099330399185419,
+      "step": 47520
+    },
+    {
+      "epoch": 42.742805755395686,
+      "grad_norm": 0.08204715698957443,
+      "learning_rate": 1.1353411192101987e-05,
+      "loss": 0.0172,
+      "step": 47530
+    },
+    {
+      "action_loss": 0.0005749166011810303,
+      "epoch": 42.742805755395686,
+      "step": 47530
+    },
+    {
+      "epoch": 42.742805755395686,
+      "step": 47530,
+      "torque_loss": 0.11609262228012085
+    },
+    {
+      "epoch": 42.742805755395686,
+      "force_loss": 0.0006579051841981709,
+      "step": 47530
+    },
+    {
+      "epoch": 42.75179856115108,
+      "grad_norm": 0.08540261536836624,
+      "learning_rate": 1.133593191883634e-05,
+      "loss": 0.0135,
+      "step": 47540
+    },
+    {
+      "action_loss": 0.0014629355864599347,
+      "epoch": 42.75179856115108,
+      "step": 47540
+    },
+    {
+      "epoch": 42.75179856115108,
+      "step": 47540,
+      "torque_loss": 0.12457611411809921
+    },
+    {
+      "epoch": 42.75179856115108,
+      "force_loss": 0.0027056329417973757,
+      "step": 47540
+    },
+    {
+      "epoch": 42.760791366906474,
+      "grad_norm": 0.11708243936300278,
+      "learning_rate": 1.1318464390693711e-05,
+      "loss": 0.015,
+      "step": 47550
+    },
+    {
+      "action_loss": 0.0018090964294970036,
+      "epoch": 42.760791366906474,
+      "step": 47550
+    },
+    {
+      "epoch": 42.760791366906474,
+      "step": 47550,
+      "torque_loss": 0.14880378544330597
+    },
+    {
+      "epoch": 42.760791366906474,
+      "force_loss": 0.0025596863124519587,
+      "step": 47550
+    },
+    {
+      "epoch": 42.76978417266187,
+      "grad_norm": 0.10936582833528519,
+      "learning_rate": 1.1301008612980257e-05,
+      "loss": 0.0155,
+      "step": 47560
+    },
+    {
+      "action_loss": 0.001809805049560964,
+      "epoch": 42.76978417266187,
+      "step": 47560
+    },
+    {
+      "epoch": 42.76978417266187,
+      "step": 47560,
+      "torque_loss": 0.11501473933458328
+    },
+    {
+      "epoch": 42.76978417266187,
+      "force_loss": 0.0019889865070581436,
+      "step": 47560
+    },
+    {
+      "epoch": 42.77877697841727,
+      "grad_norm": 0.24609656631946564,
+      "learning_rate": 1.128356459099863e-05,
+      "loss": 0.0177,
+      "step": 47570
+    },
+    {
+      "action_loss": 0.0010163206607103348,
+      "epoch": 42.77877697841727,
+      "step": 47570
+    },
+    {
+      "epoch": 42.77877697841727,
+      "step": 47570,
+      "torque_loss": 0.09729542583227158
+    },
+    {
+      "epoch": 42.77877697841727,
+      "force_loss": 0.0007158178486861289,
+      "step": 47570
+    },
+    {
+      "epoch": 42.78776978417266,
+      "grad_norm": 0.5613279342651367,
+      "learning_rate": 1.1266132330047802e-05,
+      "loss": 0.0162,
+      "step": 47580
+    },
+    {
+      "action_loss": 0.0008137661498039961,
+      "epoch": 42.78776978417266,
+      "step": 47580
+    },
+    {
+      "epoch": 42.78776978417266,
+      "step": 47580,
+      "torque_loss": 0.0968179702758789
+    },
+    {
+      "epoch": 42.78776978417266,
+      "force_loss": 0.0014220471493899822,
+      "step": 47580
+    },
+    {
+      "epoch": 42.79676258992806,
+      "grad_norm": 0.05134407803416252,
+      "learning_rate": 1.1248711835423281e-05,
+      "loss": 0.0165,
+      "step": 47590
+    },
+    {
+      "action_loss": 0.0011653086403384805,
+      "epoch": 42.79676258992806,
+      "step": 47590
+    },
+    {
+      "epoch": 42.79676258992806,
+      "step": 47590,
+      "torque_loss": 0.12019113451242447
+    },
+    {
+      "epoch": 42.79676258992806,
+      "force_loss": 0.003048662096261978,
+      "step": 47590
+    },
+    {
+      "epoch": 42.805755395683455,
+      "grad_norm": 0.05334530770778656,
+      "learning_rate": 1.123130311241693e-05,
+      "loss": 0.0138,
+      "step": 47600
+    },
+    {
+      "action_loss": 0.007098613306879997,
+      "epoch": 42.805755395683455,
+      "step": 47600
+    },
+    {
+      "epoch": 42.805755395683455,
+      "step": 47600,
+      "torque_loss": 0.08616843074560165
+    },
+    {
+      "epoch": 42.805755395683455,
+      "force_loss": 0.01013248786330223,
+      "step": 47600
+    },
+    {
+      "epoch": 42.814748201438846,
+      "grad_norm": 0.06116926297545433,
+      "learning_rate": 1.1213906166317068e-05,
+      "loss": 0.0149,
+      "step": 47610
+    },
+    {
+      "action_loss": 0.0006389851332642138,
+      "epoch": 42.814748201438846,
+      "step": 47610
+    },
+    {
+      "epoch": 42.814748201438846,
+      "step": 47610,
+      "torque_loss": 0.14744198322296143
+    },
+    {
+      "epoch": 42.814748201438846,
+      "force_loss": 0.0006112952250987291,
+      "step": 47610
+    },
+    {
+      "epoch": 42.82374100719424,
+      "grad_norm": 0.360599547624588,
+      "learning_rate": 1.1196521002408427e-05,
+      "loss": 0.0181,
+      "step": 47620
+    },
+    {
+      "action_loss": 0.0008904607966542244,
+      "epoch": 42.82374100719424,
+      "step": 47620
+    },
+    {
+      "epoch": 42.82374100719424,
+      "step": 47620,
+      "torque_loss": 0.12338205426931381
+    },
+    {
+      "epoch": 42.82374100719424,
+      "force_loss": 0.0040774052031338215,
+      "step": 47620
+    },
+    {
+      "epoch": 42.83273381294964,
+      "grad_norm": 0.08569305390119553,
+      "learning_rate": 1.1179147625972159e-05,
+      "loss": 0.0158,
+      "step": 47630
+    },
+    {
+      "action_loss": 0.0024999361485242844,
+      "epoch": 42.83273381294964,
+      "step": 47630
+    },
+    {
+      "epoch": 42.83273381294964,
+      "step": 47630,
+      "torque_loss": 0.10827334970235825
+    },
+    {
+      "epoch": 42.83273381294964,
+      "force_loss": 0.002221408998593688,
+      "step": 47630
+    },
+    {
+      "epoch": 42.84172661870504,
+      "grad_norm": 0.11933228373527527,
+      "learning_rate": 1.1161786042285822e-05,
+      "loss": 0.0153,
+      "step": 47640
+    },
+    {
+      "action_loss": 0.005867915228009224,
+      "epoch": 42.84172661870504,
+      "step": 47640
+    },
+    {
+      "epoch": 42.84172661870504,
+      "step": 47640,
+      "torque_loss": 0.11276962608098984
+    },
+    {
+      "epoch": 42.84172661870504,
+      "force_loss": 0.0065126265399158,
+      "step": 47640
+    },
+    {
+      "epoch": 42.85071942446043,
+      "grad_norm": 0.19709131121635437,
+      "learning_rate": 1.1144436256623447e-05,
+      "loss": 0.0162,
+      "step": 47650
+    },
+    {
+      "action_loss": 0.006689954549074173,
+      "epoch": 42.85071942446043,
+      "step": 47650
+    },
+    {
+      "epoch": 42.85071942446043,
+      "step": 47650,
+      "torque_loss": 0.15377478301525116
+    },
+    {
+      "epoch": 42.85071942446043,
+      "force_loss": 0.0033050470519810915,
+      "step": 47650
+    },
+    {
+      "epoch": 42.85971223021583,
+      "grad_norm": 0.10624817758798599,
+      "learning_rate": 1.1127098274255392e-05,
+      "loss": 0.018,
+      "step": 47660
+    },
+    {
+      "action_loss": 0.0006072015967220068,
+      "epoch": 42.85971223021583,
+      "step": 47660
+    },
+    {
+      "epoch": 42.85971223021583,
+      "step": 47660,
+      "torque_loss": 0.13391609489917755
+    },
+    {
+      "epoch": 42.85971223021583,
+      "force_loss": 0.0014309532707557082,
+      "step": 47660
+    },
+    {
+      "epoch": 42.868705035971225,
+      "grad_norm": 0.07417885959148407,
+      "learning_rate": 1.1109772100448512e-05,
+      "loss": 0.0152,
+      "step": 47670
+    },
+    {
+      "action_loss": 0.0029009487479925156,
+      "epoch": 42.868705035971225,
+      "step": 47670
+    },
+    {
+      "epoch": 42.868705035971225,
+      "step": 47670,
+      "torque_loss": 0.10411717742681503
+    },
+    {
+      "epoch": 42.868705035971225,
+      "force_loss": 0.004915538243949413,
+      "step": 47670
+    },
+    {
+      "epoch": 42.87769784172662,
+      "grad_norm": 0.2228514403104782,
+      "learning_rate": 1.1092457740466033e-05,
+      "loss": 0.0149,
+      "step": 47680
+    },
+    {
+      "action_loss": 0.002773801563307643,
+      "epoch": 42.87769784172662,
+      "step": 47680
+    },
+    {
+      "epoch": 42.87769784172662,
+      "step": 47680,
+      "torque_loss": 0.13222777843475342
+    },
+    {
+      "epoch": 42.87769784172662,
+      "force_loss": 0.0041300226002931595,
+      "step": 47680
+    },
+    {
+      "epoch": 42.88669064748201,
+      "grad_norm": 0.21478824317455292,
+      "learning_rate": 1.10751551995676e-05,
+      "loss": 0.0202,
+      "step": 47690
+    },
+    {
+      "action_loss": 0.003015379188582301,
+      "epoch": 42.88669064748201,
+      "step": 47690
+    },
+    {
+      "epoch": 42.88669064748201,
+      "step": 47690,
+      "torque_loss": 0.16523443162441254
+    },
+    {
+      "epoch": 42.88669064748201,
+      "force_loss": 0.005697846412658691,
+      "step": 47690
+    },
+    {
+      "epoch": 42.89568345323741,
+      "grad_norm": 0.1537024825811386,
+      "learning_rate": 1.1057864483009262e-05,
+      "loss": 0.0173,
+      "step": 47700
+    },
+    {
+      "action_loss": 0.004201934672892094,
+      "epoch": 42.89568345323741,
+      "step": 47700
+    },
+    {
+      "epoch": 42.89568345323741,
+      "step": 47700,
+      "torque_loss": 0.1378016471862793
+    },
+    {
+      "epoch": 42.89568345323741,
+      "force_loss": 0.004908794071525335,
+      "step": 47700
+    },
+    {
+      "epoch": 42.90467625899281,
+      "grad_norm": 0.19818437099456787,
+      "learning_rate": 1.1040585596043473e-05,
+      "loss": 0.0155,
+      "step": 47710
+    },
+    {
+      "action_loss": 0.001223340048454702,
+      "epoch": 42.90467625899281,
+      "step": 47710
+    },
+    {
+      "epoch": 42.90467625899281,
+      "step": 47710,
+      "torque_loss": 0.12380459159612656
+    },
+    {
+      "epoch": 42.90467625899281,
+      "force_loss": 0.0011330292327329516,
+      "step": 47710
+    },
+    {
+      "epoch": 42.9136690647482,
+      "grad_norm": 0.06923489272594452,
+      "learning_rate": 1.1023318543919148e-05,
+      "loss": 0.0158,
+      "step": 47720
+    },
+    {
+      "action_loss": 0.0033123933244496584,
+      "epoch": 42.9136690647482,
+      "step": 47720
+    },
+    {
+      "epoch": 42.9136690647482,
+      "step": 47720,
+      "torque_loss": 0.11249303817749023
+    },
+    {
+      "epoch": 42.9136690647482,
+      "force_loss": 0.00871638860553503,
+      "step": 47720
+    },
+    {
+      "epoch": 42.922661870503596,
+      "grad_norm": 0.15130597352981567,
+      "learning_rate": 1.10060633318815e-05,
+      "loss": 0.0153,
+      "step": 47730
+    },
+    {
+      "action_loss": 0.0005271590780466795,
+      "epoch": 42.922661870503596,
+      "step": 47730
+    },
+    {
+      "epoch": 42.922661870503596,
+      "step": 47730,
+      "torque_loss": 0.13004140555858612
+    },
+    {
+      "epoch": 42.922661870503596,
+      "force_loss": 0.000815410225186497,
+      "step": 47730
+    },
+    {
+      "epoch": 42.931654676258994,
+      "grad_norm": 0.08738364279270172,
+      "learning_rate": 1.0988819965172248e-05,
+      "loss": 0.0187,
+      "step": 47740
+    },
+    {
+      "action_loss": 0.00766528956592083,
+      "epoch": 42.931654676258994,
+      "step": 47740
+    },
+    {
+      "epoch": 42.931654676258994,
+      "step": 47740,
+      "torque_loss": 0.12799929082393646
+    },
+    {
+      "epoch": 42.931654676258994,
+      "force_loss": 0.012195862829685211,
+      "step": 47740
+    },
+    {
+      "epoch": 42.94064748201439,
+      "grad_norm": 0.17410610616207123,
+      "learning_rate": 1.0971588449029462e-05,
+      "loss": 0.0182,
+      "step": 47750
+    },
+    {
+      "action_loss": 0.004894527141004801,
+      "epoch": 42.94064748201439,
+      "step": 47750
+    },
+    {
+      "epoch": 42.94064748201439,
+      "step": 47750,
+      "torque_loss": 0.14845876395702362
+    },
+    {
+      "epoch": 42.94064748201439,
+      "force_loss": 0.0026279091835021973,
+      "step": 47750
+    },
+    {
+      "epoch": 42.94964028776978,
+      "grad_norm": 0.24348744750022888,
+      "learning_rate": 1.095436878868762e-05,
+      "loss": 0.0185,
+      "step": 47760
+    },
+    {
+      "action_loss": 0.0014869635924696922,
+      "epoch": 42.94964028776978,
+      "step": 47760
+    },
+    {
+      "epoch": 42.94964028776978,
+      "step": 47760,
+      "torque_loss": 0.1049477681517601
+    },
+    {
+      "epoch": 42.94964028776978,
+      "force_loss": 0.00132732221391052,
+      "step": 47760
+    },
+    {
+      "epoch": 42.95863309352518,
+      "grad_norm": 0.21606042981147766,
+      "learning_rate": 1.0937160989377598e-05,
+      "loss": 0.0178,
+      "step": 47770
+    },
+    {
+      "action_loss": 0.0005397333297878504,
+      "epoch": 42.95863309352518,
+      "step": 47770
+    },
+    {
+      "epoch": 42.95863309352518,
+      "step": 47770,
+      "torque_loss": 0.16108444333076477
+    },
+    {
+      "epoch": 42.95863309352518,
+      "force_loss": 0.0009688996360637248,
+      "step": 47770
+    },
+    {
+      "epoch": 42.96762589928058,
+      "grad_norm": 0.058848198503255844,
+      "learning_rate": 1.0919965056326676e-05,
+      "loss": 0.0159,
+      "step": 47780
+    },
+    {
+      "action_loss": 0.0009252442978322506,
+      "epoch": 42.96762589928058,
+      "step": 47780
+    },
+    {
+      "epoch": 42.96762589928058,
+      "step": 47780,
+      "torque_loss": 0.1137903556227684
+    },
+    {
+      "epoch": 42.96762589928058,
+      "force_loss": 0.0012039474677294493,
+      "step": 47780
+    },
+    {
+      "epoch": 42.976618705035975,
+      "grad_norm": 0.12039360404014587,
+      "learning_rate": 1.0902780994758504e-05,
+      "loss": 0.0149,
+      "step": 47790
+    },
+    {
+      "action_loss": 0.0012971563264727592,
+      "epoch": 42.976618705035975,
+      "step": 47790
+    },
+    {
+      "epoch": 42.976618705035975,
+      "step": 47790,
+      "torque_loss": 0.12286225706338882
+    },
+    {
+      "epoch": 42.976618705035975,
+      "force_loss": 0.001622358220629394,
+      "step": 47790
+    },
+    {
+      "epoch": 42.985611510791365,
+      "grad_norm": 0.24436727166175842,
+      "learning_rate": 1.0885608809893193e-05,
+      "loss": 0.0138,
+      "step": 47800
+    },
+    {
+      "action_loss": 0.0010201864643022418,
+      "epoch": 42.985611510791365,
+      "step": 47800
+    },
+    {
+      "epoch": 42.985611510791365,
+      "step": 47800,
+      "torque_loss": 0.10960216075181961
+    },
+    {
+      "epoch": 42.985611510791365,
+      "force_loss": 0.001070774276740849,
+      "step": 47800
+    },
+    {
+      "epoch": 42.99460431654676,
+      "grad_norm": 0.32260990142822266,
+      "learning_rate": 1.0868448506947142e-05,
+      "loss": 0.016,
+      "step": 47810
+    },
+    {
+      "action_loss": 0.0006577176391147077,
+      "epoch": 42.99460431654676,
+      "step": 47810
+    },
+    {
+      "epoch": 42.99460431654676,
+      "step": 47810,
+      "torque_loss": 0.13745005428791046
+    },
+    {
+      "epoch": 42.99460431654676,
+      "force_loss": 0.0011286836815997958,
+      "step": 47810
+    },
+    {
+      "epoch": 43.00359712230216,
+      "grad_norm": 0.06442060321569443,
+      "learning_rate": 1.0851300091133243e-05,
+      "loss": 0.0128,
+      "step": 47820
+    },
+    {
+      "action_loss": 0.002379534998908639,
+      "epoch": 43.00359712230216,
+      "step": 47820
+    },
+    {
+      "epoch": 43.00359712230216,
+      "step": 47820,
+      "torque_loss": 0.11185771226882935
+    },
+    {
+      "epoch": 43.00359712230216,
+      "force_loss": 0.0060440548695623875,
+      "step": 47820
+    },
+    {
+      "epoch": 43.01258992805755,
+      "grad_norm": 0.1262531578540802,
+      "learning_rate": 1.083416356766071e-05,
+      "loss": 0.0162,
+      "step": 47830
+    },
+    {
+      "action_loss": 0.012524661608040333,
+      "epoch": 43.01258992805755,
+      "step": 47830
+    },
+    {
+      "epoch": 43.01258992805755,
+      "step": 47830,
+      "torque_loss": 0.1212630495429039
+    },
+    {
+      "epoch": 43.01258992805755,
+      "force_loss": 0.007837825454771519,
+      "step": 47830
+    },
+    {
+      "epoch": 43.02158273381295,
+      "grad_norm": 0.1975613683462143,
+      "learning_rate": 1.0817038941735175e-05,
+      "loss": 0.0216,
+      "step": 47840
+    },
+    {
+      "action_loss": 0.0007798531441949308,
+      "epoch": 43.02158273381295,
+      "step": 47840
+    },
+    {
+      "epoch": 43.02158273381295,
+      "step": 47840,
+      "torque_loss": 0.10824859142303467
+    },
+    {
+      "epoch": 43.02158273381295,
+      "force_loss": 0.0011800049105659127,
+      "step": 47840
+    },
+    {
+      "epoch": 43.030575539568346,
+      "grad_norm": 0.381454199552536,
+      "learning_rate": 1.0799926218558642e-05,
+      "loss": 0.018,
+      "step": 47850
+    },
+    {
+      "action_loss": 0.002221165457740426,
+      "epoch": 43.030575539568346,
+      "step": 47850
+    },
+    {
+      "epoch": 43.030575539568346,
+      "step": 47850,
+      "torque_loss": 0.08142309635877609
+    },
+    {
+      "epoch": 43.030575539568346,
+      "force_loss": 0.00194599200040102,
+      "step": 47850
+    },
+    {
+      "epoch": 43.039568345323744,
+      "grad_norm": 0.37045004963874817,
+      "learning_rate": 1.0782825403329488e-05,
+      "loss": 0.0159,
+      "step": 47860
+    },
+    {
+      "action_loss": 0.011807483620941639,
+      "epoch": 43.039568345323744,
+      "step": 47860
+    },
+    {
+      "epoch": 43.039568345323744,
+      "step": 47860,
+      "torque_loss": 0.14993666112422943
+    },
+    {
+      "epoch": 43.039568345323744,
+      "force_loss": 0.03662684187293053,
+      "step": 47860
+    },
+    {
+      "epoch": 43.048561151079134,
+      "grad_norm": 0.28101295232772827,
+      "learning_rate": 1.076573650124254e-05,
+      "loss": 0.0153,
+      "step": 47870
+    },
+    {
+      "action_loss": 0.003324523800984025,
+      "epoch": 43.048561151079134,
+      "step": 47870
+    },
+    {
+      "epoch": 43.048561151079134,
+      "step": 47870,
+      "torque_loss": 0.13396786153316498
+    },
+    {
+      "epoch": 43.048561151079134,
+      "force_loss": 0.002152825240045786,
+      "step": 47870
+    },
+    {
+      "epoch": 43.05755395683453,
+      "grad_norm": 0.17905646562576294,
+      "learning_rate": 1.0748659517488891e-05,
+      "loss": 0.0164,
+      "step": 47880
+    },
+    {
+      "action_loss": 0.0008058263338170946,
+      "epoch": 43.05755395683453,
+      "step": 47880
+    },
+    {
+      "epoch": 43.05755395683453,
+      "step": 47880,
+      "torque_loss": 0.07485131919384003
+    },
+    {
+      "epoch": 43.05755395683453,
+      "force_loss": 0.0027047526091337204,
+      "step": 47880
+    },
+    {
+      "epoch": 43.06654676258993,
+      "grad_norm": 0.08739248663187027,
+      "learning_rate": 1.0731594457256138e-05,
+      "loss": 0.0155,
+      "step": 47890
+    },
+    {
+      "action_loss": 0.001667486852966249,
+      "epoch": 43.06654676258993,
+      "step": 47890
+    },
+    {
+      "epoch": 43.06654676258993,
+      "step": 47890,
+      "torque_loss": 0.13758164644241333
+    },
+    {
+      "epoch": 43.06654676258993,
+      "force_loss": 0.0022158322390168905,
+      "step": 47890
+    },
+    {
+      "epoch": 43.07553956834533,
+      "grad_norm": 0.05744181200861931,
+      "learning_rate": 1.0714541325728139e-05,
+      "loss": 0.0157,
+      "step": 47900
+    },
+    {
+      "action_loss": 0.0023959812242537737,
+      "epoch": 43.07553956834533,
+      "step": 47900
+    },
+    {
+      "epoch": 43.07553956834533,
+      "step": 47900,
+      "torque_loss": 0.13124150037765503
+    },
+    {
+      "epoch": 43.07553956834533,
+      "force_loss": 0.0031061277259141207,
+      "step": 47900
+    },
+    {
+      "epoch": 43.08453237410072,
+      "grad_norm": 0.14698629081249237,
+      "learning_rate": 1.0697500128085231e-05,
+      "loss": 0.0146,
+      "step": 47910
+    },
+    {
+      "action_loss": 0.00103837123606354,
+      "epoch": 43.08453237410072,
+      "step": 47910
+    },
+    {
+      "epoch": 43.08453237410072,
+      "step": 47910,
+      "torque_loss": 0.0659821629524231
+    },
+    {
+      "epoch": 43.08453237410072,
+      "force_loss": 0.0024087040219455957,
+      "step": 47910
+    },
+    {
+      "epoch": 43.093525179856115,
+      "grad_norm": 0.2218926101922989,
+      "learning_rate": 1.0680470869504055e-05,
+      "loss": 0.0164,
+      "step": 47920
+    },
+    {
+      "action_loss": 0.0012790969340130687,
+      "epoch": 43.093525179856115,
+      "step": 47920
+    },
+    {
+      "epoch": 43.093525179856115,
+      "step": 47920,
+      "torque_loss": 0.17348580062389374
+    },
+    {
+      "epoch": 43.093525179856115,
+      "force_loss": 0.0015070317313075066,
+      "step": 47920
+    },
+    {
+      "epoch": 43.10251798561151,
+      "grad_norm": 0.09568403661251068,
+      "learning_rate": 1.066345355515766e-05,
+      "loss": 0.0151,
+      "step": 47930
+    },
+    {
+      "action_loss": 0.0010845664655789733,
+      "epoch": 43.10251798561151,
+      "step": 47930
+    },
+    {
+      "epoch": 43.10251798561151,
+      "step": 47930,
+      "torque_loss": 0.11754051595926285
+    },
+    {
+      "epoch": 43.10251798561151,
+      "force_loss": 0.0034554668236523867,
+      "step": 47930
+    },
+    {
+      "epoch": 43.111510791366904,
+      "grad_norm": 0.12043570727109909,
+      "learning_rate": 1.0646448190215453e-05,
+      "loss": 0.0144,
+      "step": 47940
+    },
+    {
+      "action_loss": 0.0005038560484535992,
+      "epoch": 43.111510791366904,
+      "step": 47940
+    },
+    {
+      "epoch": 43.111510791366904,
+      "step": 47940,
+      "torque_loss": 0.09740033000707626
+    },
+    {
+      "epoch": 43.111510791366904,
+      "force_loss": 0.0004790698003489524,
+      "step": 47940
+    },
+    {
+      "epoch": 43.1205035971223,
+      "grad_norm": 0.0963497906923294,
+      "learning_rate": 1.0629454779843217e-05,
+      "loss": 0.0147,
+      "step": 47950
+    },
+    {
+      "action_loss": 0.005476843565702438,
+      "epoch": 43.1205035971223,
+      "step": 47950
+    },
+    {
+      "epoch": 43.1205035971223,
+      "step": 47950,
+      "torque_loss": 0.11254861950874329
+    },
+    {
+      "epoch": 43.1205035971223,
+      "force_loss": 0.011349198408424854,
+      "step": 47950
+    },
+    {
+      "epoch": 43.1294964028777,
+      "grad_norm": 0.05257950723171234,
+      "learning_rate": 1.0612473329203082e-05,
+      "loss": 0.0149,
+      "step": 47960
+    },
+    {
+      "action_loss": 0.0011553984368219972,
+      "epoch": 43.1294964028777,
+      "step": 47960
+    },
+    {
+      "epoch": 43.1294964028777,
+      "step": 47960,
+      "torque_loss": 0.14551182091236115
+    },
+    {
+      "epoch": 43.1294964028777,
+      "force_loss": 0.002435431582853198,
+      "step": 47960
+    },
+    {
+      "epoch": 43.138489208633096,
+      "grad_norm": 0.15176130831241608,
+      "learning_rate": 1.0595503843453596e-05,
+      "loss": 0.0173,
+      "step": 47970
+    },
+    {
+      "action_loss": 0.0014907823642715812,
+      "epoch": 43.138489208633096,
+      "step": 47970
+    },
+    {
+      "epoch": 43.138489208633096,
+      "step": 47970,
+      "torque_loss": 0.12223824113607407
+    },
+    {
+      "epoch": 43.138489208633096,
+      "force_loss": 0.002127317478880286,
+      "step": 47970
+    },
+    {
+      "epoch": 43.14748201438849,
+      "grad_norm": 0.4841651916503906,
+      "learning_rate": 1.0578546327749634e-05,
+      "loss": 0.0163,
+      "step": 47980
+    },
+    {
+      "action_loss": 0.000627431261818856,
+      "epoch": 43.14748201438849,
+      "step": 47980
+    },
+    {
+      "epoch": 43.14748201438849,
+      "step": 47980,
+      "torque_loss": 0.103617824614048
+    },
+    {
+      "epoch": 43.14748201438849,
+      "force_loss": 0.0006132099078968167,
+      "step": 47980
+    },
+    {
+      "epoch": 43.156474820143885,
+      "grad_norm": 0.12395449727773666,
+      "learning_rate": 1.0561600787242425e-05,
+      "loss": 0.0153,
+      "step": 47990
+    },
+    {
+      "action_loss": 0.0010367375798523426,
+      "epoch": 43.156474820143885,
+      "step": 47990
+    },
+    {
+      "epoch": 43.156474820143885,
+      "step": 47990,
+      "torque_loss": 0.13103224337100983
+    },
+    {
+      "epoch": 43.156474820143885,
+      "force_loss": 0.0014029039302840829,
+      "step": 47990
+    },
+    {
+      "epoch": 43.16546762589928,
+      "grad_norm": 0.06919924914836884,
+      "learning_rate": 1.0544667227079591e-05,
+      "loss": 0.0162,
+      "step": 48000
+    },
+    {
+      "action_loss": 0.006488892715424299,
+      "epoch": 43.16546762589928,
+      "step": 48000
+    },
+    {
+      "epoch": 43.16546762589928,
+      "step": 48000,
+      "torque_loss": 0.13330407440662384
+    },
+    {
+      "epoch": 43.16546762589928,
+      "force_loss": 0.009447026066482067,
+      "step": 48000
+    },
+    {
+      "epoch": 43.17446043165467,
+      "grad_norm": 0.0908312201499939,
+      "learning_rate": 1.0527745652405085e-05,
+      "loss": 0.0168,
+      "step": 48010
+    },
+    {
+      "action_loss": 0.0035829085391014814,
+      "epoch": 43.17446043165467,
+      "step": 48010
+    },
+    {
+      "epoch": 43.17446043165467,
+      "step": 48010,
+      "torque_loss": 0.13573874533176422
+    },
+    {
+      "epoch": 43.17446043165467,
+      "force_loss": 0.011387116275727749,
+      "step": 48010
+    },
+    {
+      "epoch": 43.18345323741007,
+      "grad_norm": 0.08729702234268188,
+      "learning_rate": 1.051083606835927e-05,
+      "loss": 0.0148,
+      "step": 48020
+    },
+    {
+      "action_loss": 0.0021713587921112776,
+      "epoch": 43.18345323741007,
+      "step": 48020
+    },
+    {
+      "epoch": 43.18345323741007,
+      "step": 48020,
+      "torque_loss": 0.11696640402078629
+    },
+    {
+      "epoch": 43.18345323741007,
+      "force_loss": 0.006436000112444162,
+      "step": 48020
+    },
+    {
+      "epoch": 43.19244604316547,
+      "grad_norm": 0.16368085145950317,
+      "learning_rate": 1.049393848007878e-05,
+      "loss": 0.0138,
+      "step": 48030
+    },
+    {
+      "action_loss": 0.0011179042048752308,
+      "epoch": 43.19244604316547,
+      "step": 48030
+    },
+    {
+      "epoch": 43.19244604316547,
+      "step": 48030,
+      "torque_loss": 0.08802339434623718
+    },
+    {
+      "epoch": 43.19244604316547,
+      "force_loss": 0.00294036534614861,
+      "step": 48030
+    },
+    {
+      "epoch": 43.201438848920866,
+      "grad_norm": 0.17621979117393494,
+      "learning_rate": 1.0477052892696709e-05,
+      "loss": 0.02,
+      "step": 48040
+    },
+    {
+      "action_loss": 0.000769799284171313,
+      "epoch": 43.201438848920866,
+      "step": 48040
+    },
+    {
+      "epoch": 43.201438848920866,
+      "step": 48040,
+      "torque_loss": 0.11906150728464127
+    },
+    {
+      "epoch": 43.201438848920866,
+      "force_loss": 0.01241263747215271,
+      "step": 48040
+    },
+    {
+      "epoch": 43.210431654676256,
+      "grad_norm": 0.0709763914346695,
+      "learning_rate": 1.0460179311342394e-05,
+      "loss": 0.0205,
+      "step": 48050
+    },
+    {
+      "action_loss": 0.001575362286530435,
+      "epoch": 43.210431654676256,
+      "step": 48050
+    },
+    {
+      "epoch": 43.210431654676256,
+      "step": 48050,
+      "torque_loss": 0.16489823162555695
+    },
+    {
+      "epoch": 43.210431654676256,
+      "force_loss": 0.005189526826143265,
+      "step": 48050
+    },
+    {
+      "epoch": 43.219424460431654,
+      "grad_norm": 0.06481238454580307,
+      "learning_rate": 1.0443317741141634e-05,
+      "loss": 0.0128,
+      "step": 48060
+    },
+    {
+      "action_loss": 0.003785705892369151,
+      "epoch": 43.219424460431654,
+      "step": 48060
+    },
+    {
+      "epoch": 43.219424460431654,
+      "step": 48060,
+      "torque_loss": 0.09890396147966385
+    },
+    {
+      "epoch": 43.219424460431654,
+      "force_loss": 0.002923009218648076,
+      "step": 48060
+    },
+    {
+      "epoch": 43.22841726618705,
+      "grad_norm": 0.12806440889835358,
+      "learning_rate": 1.0426468187216514e-05,
+      "loss": 0.017,
+      "step": 48070
+    },
+    {
+      "action_loss": 0.0007881092024035752,
+      "epoch": 43.22841726618705,
+      "step": 48070
+    },
+    {
+      "epoch": 43.22841726618705,
+      "step": 48070,
+      "torque_loss": 0.0918109342455864
+    },
+    {
+      "epoch": 43.22841726618705,
+      "force_loss": 0.0022175053600221872,
+      "step": 48070
+    },
+    {
+      "epoch": 43.23741007194245,
+      "grad_norm": 0.07798373699188232,
+      "learning_rate": 1.0409630654685477e-05,
+      "loss": 0.0132,
+      "step": 48080
+    },
+    {
+      "action_loss": 0.0009924840414896607,
+      "epoch": 43.23741007194245,
+      "step": 48080
+    },
+    {
+      "epoch": 43.23741007194245,
+      "step": 48080,
+      "torque_loss": 0.1431632786989212
+    },
+    {
+      "epoch": 43.23741007194245,
+      "force_loss": 0.00153402762953192,
+      "step": 48080
+    },
+    {
+      "epoch": 43.24640287769784,
+      "grad_norm": 0.04101170226931572,
+      "learning_rate": 1.039280514866332e-05,
+      "loss": 0.0149,
+      "step": 48090
+    },
+    {
+      "action_loss": 0.004973534028977156,
+      "epoch": 43.24640287769784,
+      "step": 48090
+    },
+    {
+      "epoch": 43.24640287769784,
+      "step": 48090,
+      "torque_loss": 0.09770975261926651
+    },
+    {
+      "epoch": 43.24640287769784,
+      "force_loss": 0.016257310286164284,
+      "step": 48090
+    },
+    {
+      "epoch": 43.25539568345324,
+      "grad_norm": 0.07817025482654572,
+      "learning_rate": 1.0375991674261198e-05,
+      "loss": 0.0149,
+      "step": 48100
+    },
+    {
+      "action_loss": 0.002991507528349757,
+      "epoch": 43.25539568345324,
+      "step": 48100
+    },
+    {
+      "epoch": 43.25539568345324,
+      "step": 48100,
+      "torque_loss": 0.11104949563741684
+    },
+    {
+      "epoch": 43.25539568345324,
+      "force_loss": 0.0017960690893232822,
+      "step": 48100
+    },
+    {
+      "epoch": 43.264388489208635,
+      "grad_norm": 0.05605383217334747,
+      "learning_rate": 1.0359190236586575e-05,
+      "loss": 0.016,
+      "step": 48110
+    },
+    {
+      "action_loss": 0.00559874577447772,
+      "epoch": 43.264388489208635,
+      "step": 48110
+    },
+    {
+      "epoch": 43.264388489208635,
+      "step": 48110,
+      "torque_loss": 0.07325108349323273
+    },
+    {
+      "epoch": 43.264388489208635,
+      "force_loss": 0.00607203459367156,
+      "step": 48110
+    },
+    {
+      "epoch": 43.273381294964025,
+      "grad_norm": 0.06400890648365021,
+      "learning_rate": 1.0342400840743322e-05,
+      "loss": 0.0132,
+      "step": 48120
+    },
+    {
+      "action_loss": 0.002850184217095375,
+      "epoch": 43.273381294964025,
+      "step": 48120
+    },
+    {
+      "epoch": 43.273381294964025,
+      "step": 48120,
+      "torque_loss": 0.15429259836673737
+    },
+    {
+      "epoch": 43.273381294964025,
+      "force_loss": 0.002299923449754715,
+      "step": 48120
+    },
+    {
+      "epoch": 43.28237410071942,
+      "grad_norm": 0.11021051555871964,
+      "learning_rate": 1.0325623491831593e-05,
+      "loss": 0.0139,
+      "step": 48130
+    },
+    {
+      "action_loss": 0.0017068613087758422,
+      "epoch": 43.28237410071942,
+      "step": 48130
+    },
+    {
+      "epoch": 43.28237410071942,
+      "step": 48130,
+      "torque_loss": 0.13638852536678314
+    },
+    {
+      "epoch": 43.28237410071942,
+      "force_loss": 0.004259928595274687,
+      "step": 48130
+    },
+    {
+      "epoch": 43.29136690647482,
+      "grad_norm": 0.05624956265091896,
+      "learning_rate": 1.0308858194947906e-05,
+      "loss": 0.0166,
+      "step": 48140
+    },
+    {
+      "action_loss": 0.002726856619119644,
+      "epoch": 43.29136690647482,
+      "step": 48140
+    },
+    {
+      "epoch": 43.29136690647482,
+      "step": 48140,
+      "torque_loss": 0.12415780872106552
+    },
+    {
+      "epoch": 43.29136690647482,
+      "force_loss": 0.0031326161697506905,
+      "step": 48140
+    },
+    {
+      "epoch": 43.30035971223022,
+      "grad_norm": 0.07504697889089584,
+      "learning_rate": 1.0292104955185111e-05,
+      "loss": 0.0151,
+      "step": 48150
+    },
+    {
+      "action_loss": 0.0004920857609249651,
+      "epoch": 43.30035971223022,
+      "step": 48150
+    },
+    {
+      "epoch": 43.30035971223022,
+      "step": 48150,
+      "torque_loss": 0.12791381776332855
+    },
+    {
+      "epoch": 43.30035971223022,
+      "force_loss": 0.0024852887727320194,
+      "step": 48150
+    },
+    {
+      "epoch": 43.30935251798561,
+      "grad_norm": 0.16796861588954926,
+      "learning_rate": 1.0275363777632396e-05,
+      "loss": 0.0139,
+      "step": 48160
+    },
+    {
+      "action_loss": 0.0007870792760513723,
+      "epoch": 43.30935251798561,
+      "step": 48160
+    },
+    {
+      "epoch": 43.30935251798561,
+      "step": 48160,
+      "torque_loss": 0.11213616281747818
+    },
+    {
+      "epoch": 43.30935251798561,
+      "force_loss": 0.0005961991264484823,
+      "step": 48160
+    },
+    {
+      "epoch": 43.318345323741006,
+      "grad_norm": 0.09855154156684875,
+      "learning_rate": 1.0258634667375321e-05,
+      "loss": 0.02,
+      "step": 48170
+    },
+    {
+      "action_loss": 0.002681231126189232,
+      "epoch": 43.318345323741006,
+      "step": 48170
+    },
+    {
+      "epoch": 43.318345323741006,
+      "step": 48170,
+      "torque_loss": 0.09299081563949585
+    },
+    {
+      "epoch": 43.318345323741006,
+      "force_loss": 0.003628367558121681,
+      "step": 48170
+    },
+    {
+      "epoch": 43.327338129496404,
+      "grad_norm": 0.18675276637077332,
+      "learning_rate": 1.02419176294957e-05,
+      "loss": 0.0165,
+      "step": 48180
+    },
+    {
+      "action_loss": 0.0019436705624684691,
+      "epoch": 43.327338129496404,
+      "step": 48180
+    },
+    {
+      "epoch": 43.327338129496404,
+      "step": 48180,
+      "torque_loss": 0.09857650846242905
+    },
+    {
+      "epoch": 43.327338129496404,
+      "force_loss": 0.012498490512371063,
+      "step": 48180
+    },
+    {
+      "epoch": 43.3363309352518,
+      "grad_norm": 0.06614652276039124,
+      "learning_rate": 1.0225212669071782e-05,
+      "loss": 0.0139,
+      "step": 48190
+    },
+    {
+      "action_loss": 0.0006972583942115307,
+      "epoch": 43.3363309352518,
+      "step": 48190
+    },
+    {
+      "epoch": 43.3363309352518,
+      "step": 48190,
+      "torque_loss": 0.08957979083061218
+    },
+    {
+      "epoch": 43.3363309352518,
+      "force_loss": 0.0018935035914182663,
+      "step": 48190
+    },
+    {
+      "epoch": 43.34532374100719,
+      "grad_norm": 0.1687958687543869,
+      "learning_rate": 1.0208519791178029e-05,
+      "loss": 0.0141,
+      "step": 48200
+    },
+    {
+      "action_loss": 0.0013340270379558206,
+      "epoch": 43.34532374100719,
+      "step": 48200
+    },
+    {
+      "epoch": 43.34532374100719,
+      "step": 48200,
+      "torque_loss": 0.12439454346895218
+    },
+    {
+      "epoch": 43.34532374100719,
+      "force_loss": 0.0025843174662441015,
+      "step": 48200
+    },
+    {
+      "epoch": 43.35431654676259,
+      "grad_norm": 0.2555897831916809,
+      "learning_rate": 1.019183900088535e-05,
+      "loss": 0.0159,
+      "step": 48210
+    },
+    {
+      "action_loss": 0.0019390127854421735,
+      "epoch": 43.35431654676259,
+      "step": 48210
+    },
+    {
+      "epoch": 43.35431654676259,
+      "step": 48210,
+      "torque_loss": 0.16420213878154755
+    },
+    {
+      "epoch": 43.35431654676259,
+      "force_loss": 0.003517964854836464,
+      "step": 48210
+    },
+    {
+      "epoch": 43.36330935251799,
+      "grad_norm": 0.08727017045021057,
+      "learning_rate": 1.0175170303260906e-05,
+      "loss": 0.0155,
+      "step": 48220
+    },
+    {
+      "action_loss": 0.0011818286729976535,
+      "epoch": 43.36330935251799,
+      "step": 48220
+    },
+    {
+      "epoch": 43.36330935251799,
+      "step": 48220,
+      "torque_loss": 0.12860801815986633
+    },
+    {
+      "epoch": 43.36330935251799,
+      "force_loss": 0.003914635628461838,
+      "step": 48220
+    },
+    {
+      "epoch": 43.37230215827338,
+      "grad_norm": 0.17293666303157806,
+      "learning_rate": 1.0158513703368206e-05,
+      "loss": 0.0183,
+      "step": 48230
+    },
+    {
+      "action_loss": 0.0008485823054797947,
+      "epoch": 43.37230215827338,
+      "step": 48230
+    },
+    {
+      "epoch": 43.37230215827338,
+      "step": 48230,
+      "torque_loss": 0.08316108584403992
+    },
+    {
+      "epoch": 43.37230215827338,
+      "force_loss": 0.0029302656184881926,
+      "step": 48230
+    },
+    {
+      "epoch": 43.381294964028775,
+      "grad_norm": 0.40532225370407104,
+      "learning_rate": 1.0141869206267095e-05,
+      "loss": 0.0152,
+      "step": 48240
+    },
+    {
+      "action_loss": 0.0006334370118565857,
+      "epoch": 43.381294964028775,
+      "step": 48240
+    },
+    {
+      "epoch": 43.381294964028775,
+      "step": 48240,
+      "torque_loss": 0.13413994014263153
+    },
+    {
+      "epoch": 43.381294964028775,
+      "force_loss": 0.0036111585795879364,
+      "step": 48240
+    },
+    {
+      "epoch": 43.39028776978417,
+      "grad_norm": 0.051989614963531494,
+      "learning_rate": 1.0125236817013723e-05,
+      "loss": 0.0147,
+      "step": 48250
+    },
+    {
+      "action_loss": 0.0008893844787962735,
+      "epoch": 43.39028776978417,
+      "step": 48250
+    },
+    {
+      "epoch": 43.39028776978417,
+      "step": 48250,
+      "torque_loss": 0.13448502123355865
+    },
+    {
+      "epoch": 43.39028776978417,
+      "force_loss": 0.0013115155743435025,
+      "step": 48250
+    },
+    {
+      "epoch": 43.39928057553957,
+      "grad_norm": 0.2579008936882019,
+      "learning_rate": 1.010861654066056e-05,
+      "loss": 0.0143,
+      "step": 48260
+    },
+    {
+      "action_loss": 0.0024942161981016397,
+      "epoch": 43.39928057553957,
+      "step": 48260
+    },
+    {
+      "epoch": 43.39928057553957,
+      "step": 48260,
+      "torque_loss": 0.1423993855714798
+    },
+    {
+      "epoch": 43.39928057553957,
+      "force_loss": 0.005319037940353155,
+      "step": 48260
+    },
+    {
+      "epoch": 43.40827338129496,
+      "grad_norm": 0.11613548547029495,
+      "learning_rate": 1.0092008382256434e-05,
+      "loss": 0.0147,
+      "step": 48270
+    },
+    {
+      "action_loss": 0.0005665691569447517,
+      "epoch": 43.40827338129496,
+      "step": 48270
+    },
+    {
+      "epoch": 43.40827338129496,
+      "step": 48270,
+      "torque_loss": 0.09662297368049622
+    },
+    {
+      "epoch": 43.40827338129496,
+      "force_loss": 0.0004258766712155193,
+      "step": 48270
+    },
+    {
+      "epoch": 43.41726618705036,
+      "grad_norm": 0.11186227947473526,
+      "learning_rate": 1.0075412346846458e-05,
+      "loss": 0.0139,
+      "step": 48280
+    },
+    {
+      "action_loss": 0.004276219289749861,
+      "epoch": 43.41726618705036,
+      "step": 48280
+    },
+    {
+      "epoch": 43.41726618705036,
+      "step": 48280,
+      "torque_loss": 0.09548377990722656
+    },
+    {
+      "epoch": 43.41726618705036,
+      "force_loss": 0.0027953078970313072,
+      "step": 48280
+    },
+    {
+      "epoch": 43.42625899280576,
+      "grad_norm": 0.08906149864196777,
+      "learning_rate": 1.0058828439472056e-05,
+      "loss": 0.0142,
+      "step": 48290
+    },
+    {
+      "action_loss": 0.002606047550216317,
+      "epoch": 43.42625899280576,
+      "step": 48290
+    },
+    {
+      "epoch": 43.42625899280576,
+      "step": 48290,
+      "torque_loss": 0.1044873595237732
+    },
+    {
+      "epoch": 43.42625899280576,
+      "force_loss": 0.0037624333053827286,
+      "step": 48290
+    },
+    {
+      "epoch": 43.435251798561154,
+      "grad_norm": 0.36315038800239563,
+      "learning_rate": 1.0042256665170996e-05,
+      "loss": 0.0173,
+      "step": 48300
+    },
+    {
+      "action_loss": 0.0008198173600248992,
+      "epoch": 43.435251798561154,
+      "step": 48300
+    },
+    {
+      "epoch": 43.435251798561154,
+      "step": 48300,
+      "torque_loss": 0.1483038067817688
+    },
+    {
+      "epoch": 43.435251798561154,
+      "force_loss": 0.001358009991236031,
+      "step": 48300
+    },
+    {
+      "epoch": 43.444244604316545,
+      "grad_norm": 0.4151192903518677,
+      "learning_rate": 1.0025697028977332e-05,
+      "loss": 0.0159,
+      "step": 48310
+    },
+    {
+      "action_loss": 0.00045225091162137687,
+      "epoch": 43.444244604316545,
+      "step": 48310
+    },
+    {
+      "epoch": 43.444244604316545,
+      "step": 48310,
+      "torque_loss": 0.07546117156744003
+    },
+    {
+      "epoch": 43.444244604316545,
+      "force_loss": 0.0013339482247829437,
+      "step": 48310
+    },
+    {
+      "epoch": 43.45323741007194,
+      "grad_norm": 0.20928062498569489,
+      "learning_rate": 1.0009149535921454e-05,
+      "loss": 0.013,
+      "step": 48320
+    },
+    {
+      "action_loss": 0.0013231117045506835,
+      "epoch": 43.45323741007194,
+      "step": 48320
+    },
+    {
+      "epoch": 43.45323741007194,
+      "step": 48320,
+      "torque_loss": 0.09207256883382797
+    },
+    {
+      "epoch": 43.45323741007194,
+      "force_loss": 0.003994135186076164,
+      "step": 48320
+    },
+    {
+      "epoch": 43.46223021582734,
+      "grad_norm": 0.11802376061677933,
+      "learning_rate": 9.992614191030031e-06,
+      "loss": 0.0141,
+      "step": 48330
+    },
+    {
+      "action_loss": 0.0004930900759063661,
+      "epoch": 43.46223021582734,
+      "step": 48330
+    },
+    {
+      "epoch": 43.46223021582734,
+      "step": 48330,
+      "torque_loss": 0.12215912342071533
+    },
+    {
+      "epoch": 43.46223021582734,
+      "force_loss": 0.0006290259771049023,
+      "step": 48330
+    },
+    {
+      "epoch": 43.47122302158273,
+      "grad_norm": 0.1930931806564331,
+      "learning_rate": 9.976090999326115e-06,
+      "loss": 0.0158,
+      "step": 48340
+    },
+    {
+      "action_loss": 0.001698052161373198,
+      "epoch": 43.47122302158273,
+      "step": 48340
+    },
+    {
+      "epoch": 43.47122302158273,
+      "step": 48340,
+      "torque_loss": 0.11253729462623596
+    },
+    {
+      "epoch": 43.47122302158273,
+      "force_loss": 0.006606638431549072,
+      "step": 48340
+    },
+    {
+      "epoch": 43.48021582733813,
+      "grad_norm": 0.19494445621967316,
+      "learning_rate": 9.959579965828952e-06,
+      "loss": 0.0154,
+      "step": 48350
+    },
+    {
+      "action_loss": 0.00363553105853498,
+      "epoch": 43.48021582733813,
+      "step": 48350
+    },
+    {
+      "epoch": 43.48021582733813,
+      "step": 48350,
+      "torque_loss": 0.11971917003393173
+    },
+    {
+      "epoch": 43.48021582733813,
+      "force_loss": 0.004944715183228254,
+      "step": 48350
+    },
+    {
+      "epoch": 43.489208633093526,
+      "grad_norm": 0.10635515302419662,
+      "learning_rate": 9.943081095554218e-06,
+      "loss": 0.0143,
+      "step": 48360
+    },
+    {
+      "action_loss": 0.0006834636442363262,
+      "epoch": 43.489208633093526,
+      "step": 48360
+    },
+    {
+      "epoch": 43.489208633093526,
+      "step": 48360,
+      "torque_loss": 0.09464162588119507
+    },
+    {
+      "epoch": 43.489208633093526,
+      "force_loss": 0.003731540171429515,
+      "step": 48360
+    },
+    {
+      "epoch": 43.49820143884892,
+      "grad_norm": 0.17328238487243652,
+      "learning_rate": 9.926594393513783e-06,
+      "loss": 0.0152,
+      "step": 48370
+    },
+    {
+      "action_loss": 0.002312740543857217,
+      "epoch": 43.49820143884892,
+      "step": 48370
+    },
+    {
+      "epoch": 43.49820143884892,
+      "step": 48370,
+      "torque_loss": 0.129380464553833
+    },
+    {
+      "epoch": 43.49820143884892,
+      "force_loss": 0.0029780417680740356,
+      "step": 48370
+    },
+    {
+      "epoch": 43.507194244604314,
+      "grad_norm": 0.09742707014083862,
+      "learning_rate": 9.910119864715906e-06,
+      "loss": 0.0165,
+      "step": 48380
+    },
+    {
+      "action_loss": 0.002389805158600211,
+      "epoch": 43.507194244604314,
+      "step": 48380
+    },
+    {
+      "epoch": 43.507194244604314,
+      "step": 48380,
+      "torque_loss": 0.1433708518743515
+    },
+    {
+      "epoch": 43.507194244604314,
+      "force_loss": 0.007369231432676315,
+      "step": 48380
+    },
+    {
+      "epoch": 43.51618705035971,
+      "grad_norm": 0.07832226157188416,
+      "learning_rate": 9.8936575141651e-06,
+      "loss": 0.0143,
+      "step": 48390
+    },
+    {
+      "action_loss": 0.0012067954521626234,
+      "epoch": 43.51618705035971,
+      "step": 48390
+    },
+    {
+      "epoch": 43.51618705035971,
+      "step": 48390,
+      "torque_loss": 0.11249709129333496
+    },
+    {
+      "epoch": 43.51618705035971,
+      "force_loss": 0.004242629278451204,
+      "step": 48390
+    },
+    {
+      "epoch": 43.52517985611511,
+      "grad_norm": 0.39420026540756226,
+      "learning_rate": 9.877207346862194e-06,
+      "loss": 0.0153,
+      "step": 48400
+    },
+    {
+      "action_loss": 0.001663988921791315,
+      "epoch": 43.52517985611511,
+      "step": 48400
+    },
+    {
+      "epoch": 43.52517985611511,
+      "step": 48400,
+      "torque_loss": 0.15344999730587006
+    },
+    {
+      "epoch": 43.52517985611511,
+      "force_loss": 0.0034708238672465086,
+      "step": 48400
+    },
+    {
+      "epoch": 43.53417266187051,
+      "grad_norm": 0.1646430641412735,
+      "learning_rate": 9.860769367804312e-06,
+      "loss": 0.0147,
+      "step": 48410
+    },
+    {
+      "action_loss": 0.0021479518618434668,
+      "epoch": 43.53417266187051,
+      "step": 48410
+    },
+    {
+      "epoch": 43.53417266187051,
+      "step": 48410,
+      "torque_loss": 0.1471019834280014
+    },
+    {
+      "epoch": 43.53417266187051,
+      "force_loss": 0.003607754362747073,
+      "step": 48410
+    },
+    {
+      "epoch": 43.5431654676259,
+      "grad_norm": 0.11888998001813889,
+      "learning_rate": 9.844343581984877e-06,
+      "loss": 0.016,
+      "step": 48420
+    },
+    {
+      "action_loss": 0.0011510709300637245,
+      "epoch": 43.5431654676259,
+      "step": 48420
+    },
+    {
+      "epoch": 43.5431654676259,
+      "step": 48420,
+      "torque_loss": 0.08036055415868759
+    },
+    {
+      "epoch": 43.5431654676259,
+      "force_loss": 0.0010939823696389794,
+      "step": 48420
+    },
+    {
+      "epoch": 43.552158273381295,
+      "grad_norm": 0.37382176518440247,
+      "learning_rate": 9.82792999439362e-06,
+      "loss": 0.015,
+      "step": 48430
+    },
+    {
+      "action_loss": 0.0010367751820012927,
+      "epoch": 43.552158273381295,
+      "step": 48430
+    },
+    {
+      "epoch": 43.552158273381295,
+      "step": 48430,
+      "torque_loss": 0.15117304027080536
+    },
+    {
+      "epoch": 43.552158273381295,
+      "force_loss": 0.0016779651632532477,
+      "step": 48430
+    },
+    {
+      "epoch": 43.56115107913669,
+      "grad_norm": 0.23981720209121704,
+      "learning_rate": 9.811528610016546e-06,
+      "loss": 0.0146,
+      "step": 48440
+    },
+    {
+      "action_loss": 0.0008652581018395722,
+      "epoch": 43.56115107913669,
+      "step": 48440
+    },
+    {
+      "epoch": 43.56115107913669,
+      "step": 48440,
+      "torque_loss": 0.09773878008127213
+    },
+    {
+      "epoch": 43.56115107913669,
+      "force_loss": 0.0007889159023761749,
+      "step": 48440
+    },
+    {
+      "epoch": 43.57014388489208,
+      "grad_norm": 0.1496325582265854,
+      "learning_rate": 9.79513943383597e-06,
+      "loss": 0.0133,
+      "step": 48450
+    },
+    {
+      "action_loss": 0.002451093401759863,
+      "epoch": 43.57014388489208,
+      "step": 48450
+    },
+    {
+      "epoch": 43.57014388489208,
+      "step": 48450,
+      "torque_loss": 0.08924409002065659
+    },
+    {
+      "epoch": 43.57014388489208,
+      "force_loss": 0.010365149937570095,
+      "step": 48450
+    },
+    {
+      "epoch": 43.57913669064748,
+      "grad_norm": 0.1631256490945816,
+      "learning_rate": 9.778762470830489e-06,
+      "loss": 0.0145,
+      "step": 48460
+    },
+    {
+      "action_loss": 0.002930335933342576,
+      "epoch": 43.57913669064748,
+      "step": 48460
+    },
+    {
+      "epoch": 43.57913669064748,
+      "step": 48460,
+      "torque_loss": 0.13075071573257446
+    },
+    {
+      "epoch": 43.57913669064748,
+      "force_loss": 0.008413630537688732,
+      "step": 48460
+    },
+    {
+      "epoch": 43.58812949640288,
+      "grad_norm": 0.07262802869081497,
+      "learning_rate": 9.762397725974982e-06,
+      "loss": 0.0148,
+      "step": 48470
+    },
+    {
+      "action_loss": 0.0033307112753391266,
+      "epoch": 43.58812949640288,
+      "step": 48470
+    },
+    {
+      "epoch": 43.58812949640288,
+      "step": 48470,
+      "torque_loss": 0.12713240087032318
+    },
+    {
+      "epoch": 43.58812949640288,
+      "force_loss": 0.0052453563548624516,
+      "step": 48470
+    },
+    {
+      "epoch": 43.597122302158276,
+      "grad_norm": 0.1431194543838501,
+      "learning_rate": 9.746045204240622e-06,
+      "loss": 0.0167,
+      "step": 48480
+    },
+    {
+      "action_loss": 0.0030724897515028715,
+      "epoch": 43.597122302158276,
+      "step": 48480
+    },
+    {
+      "epoch": 43.597122302158276,
+      "step": 48480,
+      "torque_loss": 0.11863982677459717
+    },
+    {
+      "epoch": 43.597122302158276,
+      "force_loss": 0.016462991014122963,
+      "step": 48480
+    },
+    {
+      "epoch": 43.606115107913666,
+      "grad_norm": 0.10779836028814316,
+      "learning_rate": 9.729704910594917e-06,
+      "loss": 0.0149,
+      "step": 48490
+    },
+    {
+      "action_loss": 0.0005644892808049917,
+      "epoch": 43.606115107913666,
+      "step": 48490
+    },
+    {
+      "epoch": 43.606115107913666,
+      "step": 48490,
+      "torque_loss": 0.13799667358398438
+    },
+    {
+      "epoch": 43.606115107913666,
+      "force_loss": 0.00060946837766096,
+      "step": 48490
+    },
+    {
+      "epoch": 43.615107913669064,
+      "grad_norm": 0.15771760046482086,
+      "learning_rate": 9.713376850001554e-06,
+      "loss": 0.0146,
+      "step": 48500
+    },
+    {
+      "action_loss": 0.0004909606068395078,
+      "epoch": 43.615107913669064,
+      "step": 48500
+    },
+    {
+      "epoch": 43.615107913669064,
+      "step": 48500,
+      "torque_loss": 0.11307934671640396
+    },
+    {
+      "epoch": 43.615107913669064,
+      "force_loss": 0.0008854271727614105,
+      "step": 48500
+    },
+    {
+      "epoch": 43.62410071942446,
+      "grad_norm": 0.10096035897731781,
+      "learning_rate": 9.697061027420622e-06,
+      "loss": 0.0145,
+      "step": 48510
+    },
+    {
+      "action_loss": 0.0008008410222828388,
+      "epoch": 43.62410071942446,
+      "step": 48510
+    },
+    {
+      "epoch": 43.62410071942446,
+      "step": 48510,
+      "torque_loss": 0.1088382825255394
+    },
+    {
+      "epoch": 43.62410071942446,
+      "force_loss": 0.0022940707858651876,
+      "step": 48510
+    },
+    {
+      "epoch": 43.63309352517986,
+      "grad_norm": 0.11563926190137863,
+      "learning_rate": 9.680757447808385e-06,
+      "loss": 0.0177,
+      "step": 48520
+    },
+    {
+      "action_loss": 0.0006132869748398662,
+      "epoch": 43.63309352517986,
+      "step": 48520
+    },
+    {
+      "epoch": 43.63309352517986,
+      "step": 48520,
+      "torque_loss": 0.1222463920712471
+    },
+    {
+      "epoch": 43.63309352517986,
+      "force_loss": 0.001937453169375658,
+      "step": 48520
+    },
+    {
+      "epoch": 43.64208633093525,
+      "grad_norm": 0.2531619966030121,
+      "learning_rate": 9.664466116117488e-06,
+      "loss": 0.0131,
+      "step": 48530
+    },
+    {
+      "action_loss": 0.000509663310367614,
+      "epoch": 43.64208633093525,
+      "step": 48530
+    },
+    {
+      "epoch": 43.64208633093525,
+      "step": 48530,
+      "torque_loss": 0.08936712145805359
+    },
+    {
+      "epoch": 43.64208633093525,
+      "force_loss": 0.0009677493944764137,
+      "step": 48530
+    },
+    {
+      "epoch": 43.65107913669065,
+      "grad_norm": 0.058159422129392624,
+      "learning_rate": 9.64818703729678e-06,
+      "loss": 0.014,
+      "step": 48540
+    },
+    {
+      "action_loss": 0.0007652916829101741,
+      "epoch": 43.65107913669065,
+      "step": 48540
+    },
+    {
+      "epoch": 43.65107913669065,
+      "step": 48540,
+      "torque_loss": 0.1308279037475586
+    },
+    {
+      "epoch": 43.65107913669065,
+      "force_loss": 0.0008879646775312722,
+      "step": 48540
+    },
+    {
+      "epoch": 43.660071942446045,
+      "grad_norm": 0.18933160603046417,
+      "learning_rate": 9.631920216291423e-06,
+      "loss": 0.0157,
+      "step": 48550
+    },
+    {
+      "action_loss": 0.009560401551425457,
+      "epoch": 43.660071942446045,
+      "step": 48550
+    },
+    {
+      "epoch": 43.660071942446045,
+      "step": 48550,
+      "torque_loss": 0.18331509828567505
+    },
+    {
+      "epoch": 43.660071942446045,
+      "force_loss": 0.012608930468559265,
+      "step": 48550
+    },
+    {
+      "epoch": 43.669064748201436,
+      "grad_norm": 0.14010848104953766,
+      "learning_rate": 9.615665658042849e-06,
+      "loss": 0.0165,
+      "step": 48560
+    },
+    {
+      "action_loss": 0.0013685687445104122,
+      "epoch": 43.669064748201436,
+      "step": 48560
+    },
+    {
+      "epoch": 43.669064748201436,
+      "step": 48560,
+      "torque_loss": 0.1093185618519783
+    },
+    {
+      "epoch": 43.669064748201436,
+      "force_loss": 0.0030134720727801323,
+      "step": 48560
+    },
+    {
+      "epoch": 43.67805755395683,
+      "grad_norm": 0.10423647612333298,
+      "learning_rate": 9.599423367488747e-06,
+      "loss": 0.015,
+      "step": 48570
+    },
+    {
+      "action_loss": 0.010759256780147552,
+      "epoch": 43.67805755395683,
+      "step": 48570
+    },
+    {
+      "epoch": 43.67805755395683,
+      "step": 48570,
+      "torque_loss": 0.09712091833353043
+    },
+    {
+      "epoch": 43.67805755395683,
+      "force_loss": 0.0132240429520607,
+      "step": 48570
+    },
+    {
+      "epoch": 43.68705035971223,
+      "grad_norm": 0.29065975546836853,
+      "learning_rate": 9.583193349563124e-06,
+      "loss": 0.0145,
+      "step": 48580
+    },
+    {
+      "action_loss": 0.0010066931135952473,
+      "epoch": 43.68705035971223,
+      "step": 48580
+    },
+    {
+      "epoch": 43.68705035971223,
+      "step": 48580,
+      "torque_loss": 0.1086641177535057
+    },
+    {
+      "epoch": 43.68705035971223,
+      "force_loss": 0.0013468527467921376,
+      "step": 48580
+    },
+    {
+      "epoch": 43.69604316546763,
+      "grad_norm": 0.17229269444942474,
+      "learning_rate": 9.566975609196216e-06,
+      "loss": 0.0148,
+      "step": 48590
+    },
+    {
+      "action_loss": 0.0010770100634545088,
+      "epoch": 43.69604316546763,
+      "step": 48590
+    },
+    {
+      "epoch": 43.69604316546763,
+      "step": 48590,
+      "torque_loss": 0.12984324991703033
+    },
+    {
+      "epoch": 43.69604316546763,
+      "force_loss": 0.002969875233247876,
+      "step": 48590
+    },
+    {
+      "epoch": 43.70503597122302,
+      "grad_norm": 0.11129924654960632,
+      "learning_rate": 9.550770151314548e-06,
+      "loss": 0.0145,
+      "step": 48600
+    },
+    {
+      "action_loss": 0.0027359535451978445,
+      "epoch": 43.70503597122302,
+      "step": 48600
+    },
+    {
+      "epoch": 43.70503597122302,
+      "step": 48600,
+      "torque_loss": 0.12264364957809448
+    },
+    {
+      "epoch": 43.70503597122302,
+      "force_loss": 0.009422987699508667,
+      "step": 48600
+    },
+    {
+      "epoch": 43.71402877697842,
+      "grad_norm": 0.09316616505384445,
+      "learning_rate": 9.53457698084091e-06,
+      "loss": 0.0139,
+      "step": 48610
+    },
+    {
+      "action_loss": 0.005461699794977903,
+      "epoch": 43.71402877697842,
+      "step": 48610
+    },
+    {
+      "epoch": 43.71402877697842,
+      "step": 48610,
+      "torque_loss": 0.10269784927368164
+    },
+    {
+      "epoch": 43.71402877697842,
+      "force_loss": 0.022500261664390564,
+      "step": 48610
+    },
+    {
+      "epoch": 43.723021582733814,
+      "grad_norm": 0.07619958370923996,
+      "learning_rate": 9.518396102694355e-06,
+      "loss": 0.0144,
+      "step": 48620
+    },
+    {
+      "action_loss": 0.01489965245127678,
+      "epoch": 43.723021582733814,
+      "step": 48620
+    },
+    {
+      "epoch": 43.723021582733814,
+      "step": 48620,
+      "torque_loss": 0.15996527671813965
+    },
+    {
+      "epoch": 43.723021582733814,
+      "force_loss": 0.016346463933587074,
+      "step": 48620
+    },
+    {
+      "epoch": 43.73201438848921,
+      "grad_norm": 0.08920487761497498,
+      "learning_rate": 9.502227521790198e-06,
+      "loss": 0.0164,
+      "step": 48630
+    },
+    {
+      "action_loss": 0.005055455956608057,
+      "epoch": 43.73201438848921,
+      "step": 48630
+    },
+    {
+      "epoch": 43.73201438848921,
+      "step": 48630,
+      "torque_loss": 0.140116885304451
+    },
+    {
+      "epoch": 43.73201438848921,
+      "force_loss": 0.003923205658793449,
+      "step": 48630
+    },
+    {
+      "epoch": 43.7410071942446,
+      "grad_norm": 0.2509458661079407,
+      "learning_rate": 9.486071243040063e-06,
+      "loss": 0.0184,
+      "step": 48640
+    },
+    {
+      "action_loss": 0.0010319643188267946,
+      "epoch": 43.7410071942446,
+      "step": 48640
+    },
+    {
+      "epoch": 43.7410071942446,
+      "step": 48640,
+      "torque_loss": 0.14142358303070068
+    },
+    {
+      "epoch": 43.7410071942446,
+      "force_loss": 0.0011054282076656818,
+      "step": 48640
+    },
+    {
+      "epoch": 43.75,
+      "grad_norm": 0.21261194348335266,
+      "learning_rate": 9.469927271351747e-06,
+      "loss": 0.0159,
+      "step": 48650
+    },
+    {
+      "action_loss": 0.0008157800766639411,
+      "epoch": 43.75,
+      "step": 48650
+    },
+    {
+      "epoch": 43.75,
+      "step": 48650,
+      "torque_loss": 0.16082684695720673
+    },
+    {
+      "epoch": 43.75,
+      "force_loss": 0.0017475209897384048,
+      "step": 48650
+    },
+    {
+      "epoch": 43.7589928057554,
+      "grad_norm": 0.23251934349536896,
+      "learning_rate": 9.453795611629419e-06,
+      "loss": 0.0142,
+      "step": 48660
+    },
+    {
+      "action_loss": 0.001602761447429657,
+      "epoch": 43.7589928057554,
+      "step": 48660
+    },
+    {
+      "epoch": 43.7589928057554,
+      "step": 48660,
+      "torque_loss": 0.12902593612670898
+    },
+    {
+      "epoch": 43.7589928057554,
+      "force_loss": 0.008653654716908932,
+      "step": 48660
+    },
+    {
+      "epoch": 43.76798561151079,
+      "grad_norm": 0.21277140080928802,
+      "learning_rate": 9.437676268773399e-06,
+      "loss": 0.0153,
+      "step": 48670
+    },
+    {
+      "action_loss": 0.0008176176925189793,
+      "epoch": 43.76798561151079,
+      "step": 48670
+    },
+    {
+      "epoch": 43.76798561151079,
+      "step": 48670,
+      "torque_loss": 0.1603170782327652
+    },
+    {
+      "epoch": 43.76798561151079,
+      "force_loss": 0.0009851822396740317,
+      "step": 48670
+    },
+    {
+      "epoch": 43.776978417266186,
+      "grad_norm": 0.07089299708604813,
+      "learning_rate": 9.421569247680357e-06,
+      "loss": 0.0147,
+      "step": 48680
+    },
+    {
+      "action_loss": 0.0010423255153000355,
+      "epoch": 43.776978417266186,
+      "step": 48680
+    },
+    {
+      "epoch": 43.776978417266186,
+      "step": 48680,
+      "torque_loss": 0.1136939525604248
+    },
+    {
+      "epoch": 43.776978417266186,
+      "force_loss": 0.0017612312221899629,
+      "step": 48680
+    },
+    {
+      "epoch": 43.78597122302158,
+      "grad_norm": 0.1810387820005417,
+      "learning_rate": 9.40547455324316e-06,
+      "loss": 0.0167,
+      "step": 48690
+    },
+    {
+      "action_loss": 0.005367903504520655,
+      "epoch": 43.78597122302158,
+      "step": 48690
+    },
+    {
+      "epoch": 43.78597122302158,
+      "step": 48690,
+      "torque_loss": 0.15231801569461823
+    },
+    {
+      "epoch": 43.78597122302158,
+      "force_loss": 0.00645092548802495,
+      "step": 48690
+    },
+    {
+      "epoch": 43.79496402877698,
+      "grad_norm": 0.2494833618402481,
+      "learning_rate": 9.389392190350965e-06,
+      "loss": 0.0158,
+      "step": 48700
+    },
+    {
+      "action_loss": 0.003078509122133255,
+      "epoch": 43.79496402877698,
+      "step": 48700
+    },
+    {
+      "epoch": 43.79496402877698,
+      "step": 48700,
+      "torque_loss": 0.15402019023895264
+    },
+    {
+      "epoch": 43.79496402877698,
+      "force_loss": 0.0024214035365730524,
+      "step": 48700
+    },
+    {
+      "epoch": 43.80395683453237,
+      "grad_norm": 0.17299841344356537,
+      "learning_rate": 9.373322163889153e-06,
+      "loss": 0.016,
+      "step": 48710
+    },
+    {
+      "action_loss": 0.0007417557644657791,
+      "epoch": 43.80395683453237,
+      "step": 48710
+    },
+    {
+      "epoch": 43.80395683453237,
+      "step": 48710,
+      "torque_loss": 0.1902492791414261
+    },
+    {
+      "epoch": 43.80395683453237,
+      "force_loss": 0.0018473620293661952,
+      "step": 48710
+    },
+    {
+      "epoch": 43.81294964028777,
+      "grad_norm": 0.40739917755126953,
+      "learning_rate": 9.357264478739375e-06,
+      "loss": 0.0161,
+      "step": 48720
+    },
+    {
+      "action_loss": 0.0008976270328275859,
+      "epoch": 43.81294964028777,
+      "step": 48720
+    },
+    {
+      "epoch": 43.81294964028777,
+      "step": 48720,
+      "torque_loss": 0.13780051469802856
+    },
+    {
+      "epoch": 43.81294964028777,
+      "force_loss": 0.0007627615705132484,
+      "step": 48720
+    },
+    {
+      "epoch": 43.82194244604317,
+      "grad_norm": 0.05911646783351898,
+      "learning_rate": 9.341219139779567e-06,
+      "loss": 0.0154,
+      "step": 48730
+    },
+    {
+      "action_loss": 0.0008540501003153622,
+      "epoch": 43.82194244604317,
+      "step": 48730
+    },
+    {
+      "epoch": 43.82194244604317,
+      "step": 48730,
+      "torque_loss": 0.14919531345367432
+    },
+    {
+      "epoch": 43.82194244604317,
+      "force_loss": 0.0017305637011304498,
+      "step": 48730
+    },
+    {
+      "epoch": 43.830935251798564,
+      "grad_norm": 0.4672837555408478,
+      "learning_rate": 9.325186151883824e-06,
+      "loss": 0.0153,
+      "step": 48740
+    },
+    {
+      "action_loss": 0.0017807031981647015,
+      "epoch": 43.830935251798564,
+      "step": 48740
+    },
+    {
+      "epoch": 43.830935251798564,
+      "step": 48740,
+      "torque_loss": 0.11897901445627213
+    },
+    {
+      "epoch": 43.830935251798564,
+      "force_loss": 0.0018606815719977021,
+      "step": 48740
+    },
+    {
+      "epoch": 43.839928057553955,
+      "grad_norm": 0.05451711267232895,
+      "learning_rate": 9.30916551992258e-06,
+      "loss": 0.0129,
+      "step": 48750
+    },
+    {
+      "action_loss": 0.003994535189121962,
+      "epoch": 43.839928057553955,
+      "step": 48750
+    },
+    {
+      "epoch": 43.839928057553955,
+      "step": 48750,
+      "torque_loss": 0.1472126841545105
+    },
+    {
+      "epoch": 43.839928057553955,
+      "force_loss": 0.0032157546374946833,
+      "step": 48750
+    },
+    {
+      "epoch": 43.84892086330935,
+      "grad_norm": 0.14822275936603546,
+      "learning_rate": 9.293157248762479e-06,
+      "loss": 0.0158,
+      "step": 48760
+    },
+    {
+      "action_loss": 0.002055221237242222,
+      "epoch": 43.84892086330935,
+      "step": 48760
+    },
+    {
+      "epoch": 43.84892086330935,
+      "step": 48760,
+      "torque_loss": 0.17756320536136627
+    },
+    {
+      "epoch": 43.84892086330935,
+      "force_loss": 0.0023576233070343733,
+      "step": 48760
+    },
+    {
+      "epoch": 43.85791366906475,
+      "grad_norm": 0.06743174046278,
+      "learning_rate": 9.2771613432664e-06,
+      "loss": 0.0166,
+      "step": 48770
+    },
+    {
+      "action_loss": 0.04644298925995827,
+      "epoch": 43.85791366906475,
+      "step": 48770
+    },
+    {
+      "epoch": 43.85791366906475,
+      "step": 48770,
+      "torque_loss": 0.16290383040905
+    },
+    {
+      "epoch": 43.85791366906475,
+      "force_loss": 0.05463673546910286,
+      "step": 48770
+    },
+    {
+      "epoch": 43.86690647482014,
+      "grad_norm": 0.0918324664235115,
+      "learning_rate": 9.261177808293481e-06,
+      "loss": 0.0227,
+      "step": 48780
+    },
+    {
+      "action_loss": 0.0009478906285949051,
+      "epoch": 43.86690647482014,
+      "step": 48780
+    },
+    {
+      "epoch": 43.86690647482014,
+      "step": 48780,
+      "torque_loss": 0.11124614626169205
+    },
+    {
+      "epoch": 43.86690647482014,
+      "force_loss": 0.0012306583812460303,
+      "step": 48780
+    },
+    {
+      "epoch": 43.87589928057554,
+      "grad_norm": 0.13011440634727478,
+      "learning_rate": 9.245206648699096e-06,
+      "loss": 0.0147,
+      "step": 48790
+    },
+    {
+      "action_loss": 0.0006437050178647041,
+      "epoch": 43.87589928057554,
+      "step": 48790
+    },
+    {
+      "epoch": 43.87589928057554,
+      "step": 48790,
+      "torque_loss": 0.16849011182785034
+    },
+    {
+      "epoch": 43.87589928057554,
+      "force_loss": 0.0013001452898606658,
+      "step": 48790
+    },
+    {
+      "epoch": 43.884892086330936,
+      "grad_norm": 0.43789413571357727,
+      "learning_rate": 9.22924786933485e-06,
+      "loss": 0.0158,
+      "step": 48800
+    },
+    {
+      "action_loss": 0.001970255747437477,
+      "epoch": 43.884892086330936,
+      "step": 48800
+    },
+    {
+      "epoch": 43.884892086330936,
+      "step": 48800,
+      "torque_loss": 0.16100837290287018
+    },
+    {
+      "epoch": 43.884892086330936,
+      "force_loss": 0.004228240344673395,
+      "step": 48800
+    },
+    {
+      "epoch": 43.893884892086334,
+      "grad_norm": 0.22035063803195953,
+      "learning_rate": 9.213301475048642e-06,
+      "loss": 0.015,
+      "step": 48810
+    },
+    {
+      "action_loss": 0.0014453298645094037,
+      "epoch": 43.893884892086334,
+      "step": 48810
+    },
+    {
+      "epoch": 43.893884892086334,
+      "step": 48810,
+      "torque_loss": 0.11721111088991165
+    },
+    {
+      "epoch": 43.893884892086334,
+      "force_loss": 0.0015289318980649114,
+      "step": 48810
+    },
+    {
+      "epoch": 43.902877697841724,
+      "grad_norm": 0.16457585990428925,
+      "learning_rate": 9.197367470684504e-06,
+      "loss": 0.0156,
+      "step": 48820
+    },
+    {
+      "action_loss": 0.0010235834633931518,
+      "epoch": 43.902877697841724,
+      "step": 48820
+    },
+    {
+      "epoch": 43.902877697841724,
+      "step": 48820,
+      "torque_loss": 0.17007851600646973
+    },
+    {
+      "epoch": 43.902877697841724,
+      "force_loss": 0.0010748450877144933,
+      "step": 48820
+    },
+    {
+      "epoch": 43.91187050359712,
+      "grad_norm": 0.13499639928340912,
+      "learning_rate": 9.181445861082816e-06,
+      "loss": 0.0169,
+      "step": 48830
+    },
+    {
+      "action_loss": 0.003606931073591113,
+      "epoch": 43.91187050359712,
+      "step": 48830
+    },
+    {
+      "epoch": 43.91187050359712,
+      "step": 48830,
+      "torque_loss": 0.11121616512537003
+    },
+    {
+      "epoch": 43.91187050359712,
+      "force_loss": 0.01151010300964117,
+      "step": 48830
+    },
+    {
+      "epoch": 43.92086330935252,
+      "grad_norm": 0.4295700788497925,
+      "learning_rate": 9.16553665108012e-06,
+      "loss": 0.02,
+      "step": 48840
+    },
+    {
+      "action_loss": 0.000658109609503299,
+      "epoch": 43.92086330935252,
+      "step": 48840
+    },
+    {
+      "epoch": 43.92086330935252,
+      "step": 48840,
+      "torque_loss": 0.11263301968574524
+    },
+    {
+      "epoch": 43.92086330935252,
+      "force_loss": 0.0013278181431815028,
+      "step": 48840
+    },
+    {
+      "epoch": 43.92985611510792,
+      "grad_norm": 0.1152895987033844,
+      "learning_rate": 9.149639845509223e-06,
+      "loss": 0.0154,
+      "step": 48850
+    },
+    {
+      "action_loss": 0.0008247655932791531,
+      "epoch": 43.92985611510792,
+      "step": 48850
+    },
+    {
+      "epoch": 43.92985611510792,
+      "step": 48850,
+      "torque_loss": 0.13623180985450745
+    },
+    {
+      "epoch": 43.92985611510792,
+      "force_loss": 0.0009287621360272169,
+      "step": 48850
+    },
+    {
+      "epoch": 43.93884892086331,
+      "grad_norm": 0.23939968645572662,
+      "learning_rate": 9.133755449199144e-06,
+      "loss": 0.0139,
+      "step": 48860
+    },
+    {
+      "action_loss": 0.0019896244630217552,
+      "epoch": 43.93884892086331,
+      "step": 48860
+    },
+    {
+      "epoch": 43.93884892086331,
+      "step": 48860,
+      "torque_loss": 0.14498431980609894
+    },
+    {
+      "epoch": 43.93884892086331,
+      "force_loss": 0.0045209359377622604,
+      "step": 48860
+    },
+    {
+      "epoch": 43.947841726618705,
+      "grad_norm": 0.0391739197075367,
+      "learning_rate": 9.117883466975135e-06,
+      "loss": 0.0141,
+      "step": 48870
+    },
+    {
+      "action_loss": 0.0010534119792282581,
+      "epoch": 43.947841726618705,
+      "step": 48870
+    },
+    {
+      "epoch": 43.947841726618705,
+      "step": 48870,
+      "torque_loss": 0.12999878823757172
+    },
+    {
+      "epoch": 43.947841726618705,
+      "force_loss": 0.002682490274310112,
+      "step": 48870
+    },
+    {
+      "epoch": 43.9568345323741,
+      "grad_norm": 0.0908980444073677,
+      "learning_rate": 9.10202390365873e-06,
+      "loss": 0.0157,
+      "step": 48880
+    },
+    {
+      "action_loss": 0.0020297958981245756,
+      "epoch": 43.9568345323741,
+      "step": 48880
+    },
+    {
+      "epoch": 43.9568345323741,
+      "step": 48880,
+      "torque_loss": 0.06899247318506241
+    },
+    {
+      "epoch": 43.9568345323741,
+      "force_loss": 0.0021482689771801233,
+      "step": 48880
+    },
+    {
+      "epoch": 43.96582733812949,
+      "grad_norm": 0.22997556626796722,
+      "learning_rate": 9.086176764067583e-06,
+      "loss": 0.0166,
+      "step": 48890
+    },
+    {
+      "action_loss": 0.0038438818883150816,
+      "epoch": 43.96582733812949,
+      "step": 48890
+    },
+    {
+      "epoch": 43.96582733812949,
+      "step": 48890,
+      "torque_loss": 0.13119101524353027
+    },
+    {
+      "epoch": 43.96582733812949,
+      "force_loss": 0.003268355503678322,
+      "step": 48890
+    },
+    {
+      "epoch": 43.97482014388489,
+      "grad_norm": 0.4135984182357788,
+      "learning_rate": 9.070342053015684e-06,
+      "loss": 0.0164,
+      "step": 48900
+    },
+    {
+      "action_loss": 0.009184107184410095,
+      "epoch": 43.97482014388489,
+      "step": 48900
+    },
+    {
+      "epoch": 43.97482014388489,
+      "step": 48900,
+      "torque_loss": 0.15513502061367035
+    },
+    {
+      "epoch": 43.97482014388489,
+      "force_loss": 0.011585988104343414,
+      "step": 48900
+    },
+    {
+      "epoch": 43.98381294964029,
+      "grad_norm": 0.24386382102966309,
+      "learning_rate": 9.054519775313187e-06,
+      "loss": 0.0162,
+      "step": 48910
+    },
+    {
+      "action_loss": 0.012753057293593884,
+      "epoch": 43.98381294964029,
+      "step": 48910
+    },
+    {
+      "epoch": 43.98381294964029,
+      "step": 48910,
+      "torque_loss": 0.16046780347824097
+    },
+    {
+      "epoch": 43.98381294964029,
+      "force_loss": 0.026838863268494606,
+      "step": 48910
+    },
+    {
+      "epoch": 43.992805755395686,
+      "grad_norm": 0.0535489022731781,
+      "learning_rate": 9.038709935766476e-06,
+      "loss": 0.0146,
+      "step": 48920
+    },
+    {
+      "action_loss": 0.0015181965427473187,
+      "epoch": 43.992805755395686,
+      "step": 48920
+    },
+    {
+      "epoch": 43.992805755395686,
+      "step": 48920,
+      "torque_loss": 0.1305958330631256
+    },
+    {
+      "epoch": 43.992805755395686,
+      "force_loss": 0.0020104984287172556,
+      "step": 48920
+    },
+    {
+      "epoch": 44.00179856115108,
+      "grad_norm": 0.4193684756755829,
+      "learning_rate": 9.02291253917817e-06,
+      "loss": 0.0175,
+      "step": 48930
+    },
+    {
+      "action_loss": 0.0006691753515042365,
+      "epoch": 44.00179856115108,
+      "step": 48930
+    },
+    {
+      "epoch": 44.00179856115108,
+      "step": 48930,
+      "torque_loss": 0.13290785253047943
+    },
+    {
+      "epoch": 44.00179856115108,
+      "force_loss": 0.0035969307646155357,
+      "step": 48930
+    },
+    {
+      "epoch": 44.010791366906474,
+      "grad_norm": 0.16773025691509247,
+      "learning_rate": 9.007127590347091e-06,
+      "loss": 0.0151,
+      "step": 48940
+    },
+    {
+      "action_loss": 0.0009007633198052645,
+      "epoch": 44.010791366906474,
+      "step": 48940
+    },
+    {
+      "epoch": 44.010791366906474,
+      "step": 48940,
+      "torque_loss": 0.1115138903260231
+    },
+    {
+      "epoch": 44.010791366906474,
+      "force_loss": 0.0032146081794053316,
+      "step": 48940
+    },
+    {
+      "epoch": 44.01978417266187,
+      "grad_norm": 0.3608867824077606,
+      "learning_rate": 8.991355094068288e-06,
+      "loss": 0.0168,
+      "step": 48950
+    },
+    {
+      "action_loss": 0.0011368446284905076,
+      "epoch": 44.01978417266187,
+      "step": 48950
+    },
+    {
+      "epoch": 44.01978417266187,
+      "step": 48950,
+      "torque_loss": 0.09599777311086655
+    },
+    {
+      "epoch": 44.01978417266187,
+      "force_loss": 0.0060363151133060455,
+      "step": 48950
+    },
+    {
+      "epoch": 44.02877697841727,
+      "grad_norm": 0.0976967141032219,
+      "learning_rate": 8.975595055133062e-06,
+      "loss": 0.0143,
+      "step": 48960
+    },
+    {
+      "action_loss": 0.0013883948558941483,
+      "epoch": 44.02877697841727,
+      "step": 48960
+    },
+    {
+      "epoch": 44.02877697841727,
+      "step": 48960,
+      "torque_loss": 0.12109386920928955
+    },
+    {
+      "epoch": 44.02877697841727,
+      "force_loss": 0.0020069105084985495,
+      "step": 48960
+    },
+    {
+      "epoch": 44.03776978417266,
+      "grad_norm": 0.08910827338695526,
+      "learning_rate": 8.959847478328848e-06,
+      "loss": 0.0161,
+      "step": 48970
+    },
+    {
+      "action_loss": 0.0008580006542615592,
+      "epoch": 44.03776978417266,
+      "step": 48970
+    },
+    {
+      "epoch": 44.03776978417266,
+      "step": 48970,
+      "torque_loss": 0.11592822521924973
+    },
+    {
+      "epoch": 44.03776978417266,
+      "force_loss": 0.00447820546105504,
+      "step": 48970
+    },
+    {
+      "epoch": 44.04676258992806,
+      "grad_norm": 0.29000189900398254,
+      "learning_rate": 8.944112368439378e-06,
+      "loss": 0.0149,
+      "step": 48980
+    },
+    {
+      "action_loss": 0.0020134069491177797,
+      "epoch": 44.04676258992806,
+      "step": 48980
+    },
+    {
+      "epoch": 44.04676258992806,
+      "step": 48980,
+      "torque_loss": 0.1304437518119812
+    },
+    {
+      "epoch": 44.04676258992806,
+      "force_loss": 0.010307746939361095,
+      "step": 48980
+    },
+    {
+      "epoch": 44.055755395683455,
+      "grad_norm": 0.04845907539129257,
+      "learning_rate": 8.928389730244552e-06,
+      "loss": 0.0142,
+      "step": 48990
+    },
+    {
+      "action_loss": 0.0007547863642685115,
+      "epoch": 44.055755395683455,
+      "step": 48990
+    },
+    {
+      "epoch": 44.055755395683455,
+      "step": 48990,
+      "torque_loss": 0.11735448241233826
+    },
+    {
+      "epoch": 44.055755395683455,
+      "force_loss": 0.0007481517386622727,
+      "step": 48990
+    },
+    {
+      "epoch": 44.064748201438846,
+      "grad_norm": 0.14411114156246185,
+      "learning_rate": 8.912679568520494e-06,
+      "loss": 0.0145,
+      "step": 49000
+    },
+    {
+      "action_loss": 0.0008495559450238943,
+      "epoch": 44.064748201438846,
+      "step": 49000
+    },
+    {
+      "epoch": 44.064748201438846,
+      "step": 49000,
+      "torque_loss": 0.1195276752114296
+    },
+    {
+      "epoch": 44.064748201438846,
+      "force_loss": 0.0035455410834401846,
+      "step": 49000
+    },
+    {
+      "epoch": 44.07374100719424,
+      "grad_norm": 0.16706620156764984,
+      "learning_rate": 8.896981888039534e-06,
+      "loss": 0.0136,
+      "step": 49010
+    },
+    {
+      "action_loss": 0.0011316556483507156,
+      "epoch": 44.07374100719424,
+      "step": 49010
+    },
+    {
+      "epoch": 44.07374100719424,
+      "step": 49010,
+      "torque_loss": 0.1310105174779892
+    },
+    {
+      "epoch": 44.07374100719424,
+      "force_loss": 0.0017588408663868904,
+      "step": 49010
+    },
+    {
+      "epoch": 44.08273381294964,
+      "grad_norm": 0.09481914341449738,
+      "learning_rate": 8.881296693570201e-06,
+      "loss": 0.0147,
+      "step": 49020
+    },
+    {
+      "action_loss": 0.008379969745874405,
+      "epoch": 44.08273381294964,
+      "step": 49020
+    },
+    {
+      "epoch": 44.08273381294964,
+      "step": 49020,
+      "torque_loss": 0.17608459293842316
+    },
+    {
+      "epoch": 44.08273381294964,
+      "force_loss": 0.01149552408605814,
+      "step": 49020
+    },
+    {
+      "epoch": 44.09172661870504,
+      "grad_norm": 0.15379805862903595,
+      "learning_rate": 8.865623989877281e-06,
+      "loss": 0.0165,
+      "step": 49030
+    },
+    {
+      "action_loss": 0.0005789054557681084,
+      "epoch": 44.09172661870504,
+      "step": 49030
+    },
+    {
+      "epoch": 44.09172661870504,
+      "step": 49030,
+      "torque_loss": 0.07873835414648056
+    },
+    {
+      "epoch": 44.09172661870504,
+      "force_loss": 0.003443564986810088,
+      "step": 49030
+    },
+    {
+      "epoch": 44.10071942446043,
+      "grad_norm": 0.11708827316761017,
+      "learning_rate": 8.849963781721681e-06,
+      "loss": 0.0149,
+      "step": 49040
+    },
+    {
+      "action_loss": 0.0032002823427319527,
+      "epoch": 44.10071942446043,
+      "step": 49040
+    },
+    {
+      "epoch": 44.10071942446043,
+      "step": 49040,
+      "torque_loss": 0.1290290653705597
+    },
+    {
+      "epoch": 44.10071942446043,
+      "force_loss": 0.005330126266926527,
+      "step": 49040
+    },
+    {
+      "epoch": 44.10971223021583,
+      "grad_norm": 0.0901990532875061,
+      "learning_rate": 8.834316073860588e-06,
+      "loss": 0.016,
+      "step": 49050
+    },
+    {
+      "action_loss": 0.014266385696828365,
+      "epoch": 44.10971223021583,
+      "step": 49050
+    },
+    {
+      "epoch": 44.10971223021583,
+      "step": 49050,
+      "torque_loss": 0.13990400731563568
+    },
+    {
+      "epoch": 44.10971223021583,
+      "force_loss": 0.01892738975584507,
+      "step": 49050
+    },
+    {
+      "epoch": 44.118705035971225,
+      "grad_norm": 0.15352337062358856,
+      "learning_rate": 8.818680871047357e-06,
+      "loss": 0.019,
+      "step": 49060
+    },
+    {
+      "action_loss": 0.0013577440986409783,
+      "epoch": 44.118705035971225,
+      "step": 49060
+    },
+    {
+      "epoch": 44.118705035971225,
+      "step": 49060,
+      "torque_loss": 0.09515384584665298
+    },
+    {
+      "epoch": 44.118705035971225,
+      "force_loss": 0.0034006417263299227,
+      "step": 49060
+    },
+    {
+      "epoch": 44.12769784172662,
+      "grad_norm": 0.10146159678697586,
+      "learning_rate": 8.803058178031549e-06,
+      "loss": 0.0148,
+      "step": 49070
+    },
+    {
+      "action_loss": 0.0005237226141616702,
+      "epoch": 44.12769784172662,
+      "step": 49070
+    },
+    {
+      "epoch": 44.12769784172662,
+      "step": 49070,
+      "torque_loss": 0.10528019815683365
+    },
+    {
+      "epoch": 44.12769784172662,
+      "force_loss": 0.0022101302165538073,
+      "step": 49070
+    },
+    {
+      "epoch": 44.13669064748201,
+      "grad_norm": 0.06642207503318787,
+      "learning_rate": 8.787447999558922e-06,
+      "loss": 0.0167,
+      "step": 49080
+    },
+    {
+      "action_loss": 0.005163056310266256,
+      "epoch": 44.13669064748201,
+      "step": 49080
+    },
+    {
+      "epoch": 44.13669064748201,
+      "step": 49080,
+      "torque_loss": 0.15639592707157135
+    },
+    {
+      "epoch": 44.13669064748201,
+      "force_loss": 0.009870877489447594,
+      "step": 49080
+    },
+    {
+      "epoch": 44.14568345323741,
+      "grad_norm": 0.12552420794963837,
+      "learning_rate": 8.77185034037144e-06,
+      "loss": 0.0159,
+      "step": 49090
+    },
+    {
+      "action_loss": 0.0008313118014484644,
+      "epoch": 44.14568345323741,
+      "step": 49090
+    },
+    {
+      "epoch": 44.14568345323741,
+      "step": 49090,
+      "torque_loss": 0.08804670721292496
+    },
+    {
+      "epoch": 44.14568345323741,
+      "force_loss": 0.0007858526078052819,
+      "step": 49090
+    },
+    {
+      "epoch": 44.15467625899281,
+      "grad_norm": 0.13472674787044525,
+      "learning_rate": 8.756265205207259e-06,
+      "loss": 0.0135,
+      "step": 49100
+    },
+    {
+      "action_loss": 0.0012137899175286293,
+      "epoch": 44.15467625899281,
+      "step": 49100
+    },
+    {
+      "epoch": 44.15467625899281,
+      "step": 49100,
+      "torque_loss": 0.13230256736278534
+    },
+    {
+      "epoch": 44.15467625899281,
+      "force_loss": 0.002622600644826889,
+      "step": 49100
+    },
+    {
+      "epoch": 44.1636690647482,
+      "grad_norm": 0.06753526628017426,
+      "learning_rate": 8.740692598800732e-06,
+      "loss": 0.0128,
+      "step": 49110
+    },
+    {
+      "action_loss": 0.00447540869936347,
+      "epoch": 44.1636690647482,
+      "step": 49110
+    },
+    {
+      "epoch": 44.1636690647482,
+      "step": 49110,
+      "torque_loss": 0.11595549434423447
+    },
+    {
+      "epoch": 44.1636690647482,
+      "force_loss": 0.0022706470917910337,
+      "step": 49110
+    },
+    {
+      "epoch": 44.172661870503596,
+      "grad_norm": 0.11529926210641861,
+      "learning_rate": 8.72513252588239e-06,
+      "loss": 0.0142,
+      "step": 49120
+    },
+    {
+      "action_loss": 0.0013867107918486,
+      "epoch": 44.172661870503596,
+      "step": 49120
+    },
+    {
+      "epoch": 44.172661870503596,
+      "step": 49120,
+      "torque_loss": 0.10046526044607162
+    },
+    {
+      "epoch": 44.172661870503596,
+      "force_loss": 0.0031199390068650246,
+      "step": 49120
+    },
+    {
+      "epoch": 44.181654676258994,
+      "grad_norm": 0.06633419543504715,
+      "learning_rate": 8.709584991178998e-06,
+      "loss": 0.0133,
+      "step": 49130
+    },
+    {
+      "action_loss": 0.0021849339827895164,
+      "epoch": 44.181654676258994,
+      "step": 49130
+    },
+    {
+      "epoch": 44.181654676258994,
+      "step": 49130,
+      "torque_loss": 0.09284258633852005
+    },
+    {
+      "epoch": 44.181654676258994,
+      "force_loss": 0.008053903467953205,
+      "step": 49130
+    },
+    {
+      "epoch": 44.19064748201439,
+      "grad_norm": 0.07379432767629623,
+      "learning_rate": 8.694049999413479e-06,
+      "loss": 0.0127,
+      "step": 49140
+    },
+    {
+      "action_loss": 0.0017946548759937286,
+      "epoch": 44.19064748201439,
+      "step": 49140
+    },
+    {
+      "epoch": 44.19064748201439,
+      "step": 49140,
+      "torque_loss": 0.04979671910405159
+    },
+    {
+      "epoch": 44.19064748201439,
+      "force_loss": 0.002636880613863468,
+      "step": 49140
+    },
+    {
+      "epoch": 44.19964028776978,
+      "grad_norm": 0.13538052141666412,
+      "learning_rate": 8.678527555304945e-06,
+      "loss": 0.0137,
+      "step": 49150
+    },
+    {
+      "action_loss": 0.0010748477652668953,
+      "epoch": 44.19964028776978,
+      "step": 49150
+    },
+    {
+      "epoch": 44.19964028776978,
+      "step": 49150,
+      "torque_loss": 0.16835923492908478
+    },
+    {
+      "epoch": 44.19964028776978,
+      "force_loss": 0.006916664075106382,
+      "step": 49150
+    },
+    {
+      "epoch": 44.20863309352518,
+      "grad_norm": 0.10933997482061386,
+      "learning_rate": 8.663017663568712e-06,
+      "loss": 0.0139,
+      "step": 49160
+    },
+    {
+      "action_loss": 0.005764565896242857,
+      "epoch": 44.20863309352518,
+      "step": 49160
+    },
+    {
+      "epoch": 44.20863309352518,
+      "step": 49160,
+      "torque_loss": 0.16226160526275635
+    },
+    {
+      "epoch": 44.20863309352518,
+      "force_loss": 0.0039595505222678185,
+      "step": 49160
+    },
+    {
+      "epoch": 44.21762589928058,
+      "grad_norm": 0.04308479279279709,
+      "learning_rate": 8.647520328916259e-06,
+      "loss": 0.0163,
+      "step": 49170
+    },
+    {
+      "action_loss": 0.001063380972482264,
+      "epoch": 44.21762589928058,
+      "step": 49170
+    },
+    {
+      "epoch": 44.21762589928058,
+      "step": 49170,
+      "torque_loss": 0.102976955473423
+    },
+    {
+      "epoch": 44.21762589928058,
+      "force_loss": 0.0026875126641243696,
+      "step": 49170
+    },
+    {
+      "epoch": 44.226618705035975,
+      "grad_norm": 0.2102208137512207,
+      "learning_rate": 8.632035556055307e-06,
+      "loss": 0.0169,
+      "step": 49180
+    },
+    {
+      "action_loss": 0.002664957894012332,
+      "epoch": 44.226618705035975,
+      "step": 49180
+    },
+    {
+      "epoch": 44.226618705035975,
+      "step": 49180,
+      "torque_loss": 0.11847998946905136
+    },
+    {
+      "epoch": 44.226618705035975,
+      "force_loss": 0.00101189233828336,
+      "step": 49180
+    },
+    {
+      "epoch": 44.235611510791365,
+      "grad_norm": 0.16477319598197937,
+      "learning_rate": 8.616563349689672e-06,
+      "loss": 0.0168,
+      "step": 49190
+    },
+    {
+      "action_loss": 0.0013223630376160145,
+      "epoch": 44.235611510791365,
+      "step": 49190
+    },
+    {
+      "epoch": 44.235611510791365,
+      "step": 49190,
+      "torque_loss": 0.10624772310256958
+    },
+    {
+      "epoch": 44.235611510791365,
+      "force_loss": 0.002198045840486884,
+      "step": 49190
+    },
+    {
+      "epoch": 44.24460431654676,
+      "grad_norm": 0.3269161283969879,
+      "learning_rate": 8.601103714519448e-06,
+      "loss": 0.0152,
+      "step": 49200
+    },
+    {
+      "action_loss": 0.0033661779016256332,
+      "epoch": 44.24460431654676,
+      "step": 49200
+    },
+    {
+      "epoch": 44.24460431654676,
+      "step": 49200,
+      "torque_loss": 0.1871439814567566
+    },
+    {
+      "epoch": 44.24460431654676,
+      "force_loss": 0.005143625661730766,
+      "step": 49200
+    },
+    {
+      "epoch": 44.25359712230216,
+      "grad_norm": 0.5780916213989258,
+      "learning_rate": 8.58565665524082e-06,
+      "loss": 0.0158,
+      "step": 49210
+    },
+    {
+      "action_loss": 0.001128057367168367,
+      "epoch": 44.25359712230216,
+      "step": 49210
+    },
+    {
+      "epoch": 44.25359712230216,
+      "step": 49210,
+      "torque_loss": 0.10066789388656616
+    },
+    {
+      "epoch": 44.25359712230216,
+      "force_loss": 0.005184227600693703,
+      "step": 49210
+    },
+    {
+      "epoch": 44.26258992805755,
+      "grad_norm": 0.06680929660797119,
+      "learning_rate": 8.570222176546222e-06,
+      "loss": 0.0171,
+      "step": 49220
+    },
+    {
+      "action_loss": 0.0006675591575913131,
+      "epoch": 44.26258992805755,
+      "step": 49220
+    },
+    {
+      "epoch": 44.26258992805755,
+      "step": 49220,
+      "torque_loss": 0.08851144462823868
+    },
+    {
+      "epoch": 44.26258992805755,
+      "force_loss": 0.0009518570150248706,
+      "step": 49220
+    },
+    {
+      "epoch": 44.27158273381295,
+      "grad_norm": 0.6446833610534668,
+      "learning_rate": 8.554800283124242e-06,
+      "loss": 0.0159,
+      "step": 49230
+    },
+    {
+      "action_loss": 0.006528045982122421,
+      "epoch": 44.27158273381295,
+      "step": 49230
+    },
+    {
+      "epoch": 44.27158273381295,
+      "step": 49230,
+      "torque_loss": 0.11312786489725113
+    },
+    {
+      "epoch": 44.27158273381295,
+      "force_loss": 0.008985977619886398,
+      "step": 49230
+    },
+    {
+      "epoch": 44.280575539568346,
+      "grad_norm": 0.06628572940826416,
+      "learning_rate": 8.539390979659639e-06,
+      "loss": 0.0165,
+      "step": 49240
+    },
+    {
+      "action_loss": 0.0010382866021245718,
+      "epoch": 44.280575539568346,
+      "step": 49240
+    },
+    {
+      "epoch": 44.280575539568346,
+      "step": 49240,
+      "torque_loss": 0.11408758908510208
+    },
+    {
+      "epoch": 44.280575539568346,
+      "force_loss": 0.0024445292074233294,
+      "step": 49240
+    },
+    {
+      "epoch": 44.289568345323744,
+      "grad_norm": 0.14445219933986664,
+      "learning_rate": 8.523994270833352e-06,
+      "loss": 0.0152,
+      "step": 49250
+    },
+    {
+      "action_loss": 0.001168295624665916,
+      "epoch": 44.289568345323744,
+      "step": 49250
+    },
+    {
+      "epoch": 44.289568345323744,
+      "step": 49250,
+      "torque_loss": 0.10078869014978409
+    },
+    {
+      "epoch": 44.289568345323744,
+      "force_loss": 0.001015942427329719,
+      "step": 49250
+    },
+    {
+      "epoch": 44.298561151079134,
+      "grad_norm": 0.07885625213384628,
+      "learning_rate": 8.5086101613225e-06,
+      "loss": 0.0173,
+      "step": 49260
+    },
+    {
+      "action_loss": 0.0004780030867550522,
+      "epoch": 44.298561151079134,
+      "step": 49260
+    },
+    {
+      "epoch": 44.298561151079134,
+      "step": 49260,
+      "torque_loss": 0.09015974402427673
+    },
+    {
+      "epoch": 44.298561151079134,
+      "force_loss": 0.0013808520743623376,
+      "step": 49260
+    },
+    {
+      "epoch": 44.30755395683453,
+      "grad_norm": 0.18947839736938477,
+      "learning_rate": 8.493238655800346e-06,
+      "loss": 0.0131,
+      "step": 49270
+    },
+    {
+      "action_loss": 0.004167734179645777,
+      "epoch": 44.30755395683453,
+      "step": 49270
+    },
+    {
+      "epoch": 44.30755395683453,
+      "step": 49270,
+      "torque_loss": 0.12240535020828247
+    },
+    {
+      "epoch": 44.30755395683453,
+      "force_loss": 0.010304260067641735,
+      "step": 49270
+    },
+    {
+      "epoch": 44.31654676258993,
+      "grad_norm": 0.0866895318031311,
+      "learning_rate": 8.47787975893638e-06,
+      "loss": 0.015,
+      "step": 49280
+    },
+    {
+      "action_loss": 0.006357399746775627,
+      "epoch": 44.31654676258993,
+      "step": 49280
+    },
+    {
+      "epoch": 44.31654676258993,
+      "step": 49280,
+      "torque_loss": 0.15336821973323822
+    },
+    {
+      "epoch": 44.31654676258993,
+      "force_loss": 0.011244997382164001,
+      "step": 49280
+    },
+    {
+      "epoch": 44.32553956834532,
+      "grad_norm": 0.1137361228466034,
+      "learning_rate": 8.462533475396211e-06,
+      "loss": 0.0154,
+      "step": 49290
+    },
+    {
+      "action_loss": 0.001686560339294374,
+      "epoch": 44.32553956834532,
+      "step": 49290
+    },
+    {
+      "epoch": 44.32553956834532,
+      "step": 49290,
+      "torque_loss": 0.10393055528402328
+    },
+    {
+      "epoch": 44.32553956834532,
+      "force_loss": 0.006603599991649389,
+      "step": 49290
+    },
+    {
+      "epoch": 44.33453237410072,
+      "grad_norm": 0.4816022217273712,
+      "learning_rate": 8.447199809841643e-06,
+      "loss": 0.0134,
+      "step": 49300
+    },
+    {
+      "action_loss": 0.0007192753255367279,
+      "epoch": 44.33453237410072,
+      "step": 49300
+    },
+    {
+      "epoch": 44.33453237410072,
+      "step": 49300,
+      "torque_loss": 0.12039162963628769
+    },
+    {
+      "epoch": 44.33453237410072,
+      "force_loss": 0.0037354787345975637,
+      "step": 49300
+    },
+    {
+      "epoch": 44.343525179856115,
+      "grad_norm": 0.06910062581300735,
+      "learning_rate": 8.431878766930635e-06,
+      "loss": 0.0129,
+      "step": 49310
+    },
+    {
+      "action_loss": 0.0005455666687339544,
+      "epoch": 44.343525179856115,
+      "step": 49310
+    },
+    {
+      "epoch": 44.343525179856115,
+      "step": 49310,
+      "torque_loss": 0.09510147571563721
+    },
+    {
+      "epoch": 44.343525179856115,
+      "force_loss": 0.0009796253871172667,
+      "step": 49310
+    },
+    {
+      "epoch": 44.35251798561151,
+      "grad_norm": 0.15499146282672882,
+      "learning_rate": 8.416570351317304e-06,
+      "loss": 0.015,
+      "step": 49320
+    },
+    {
+      "action_loss": 0.003919545561075211,
+      "epoch": 44.35251798561151,
+      "step": 49320
+    },
+    {
+      "epoch": 44.35251798561151,
+      "step": 49320,
+      "torque_loss": 0.13794682919979095
+    },
+    {
+      "epoch": 44.35251798561151,
+      "force_loss": 0.010099179111421108,
+      "step": 49320
+    },
+    {
+      "epoch": 44.361510791366904,
+      "grad_norm": 0.11637290567159653,
+      "learning_rate": 8.401274567651973e-06,
+      "loss": 0.0148,
+      "step": 49330
+    },
+    {
+      "action_loss": 0.0019072741270065308,
+      "epoch": 44.361510791366904,
+      "step": 49330
+    },
+    {
+      "epoch": 44.361510791366904,
+      "step": 49330,
+      "torque_loss": 0.14849670231342316
+    },
+    {
+      "epoch": 44.361510791366904,
+      "force_loss": 0.0016698756953701377,
+      "step": 49330
+    },
+    {
+      "epoch": 44.3705035971223,
+      "grad_norm": 0.35222741961479187,
+      "learning_rate": 8.385991420581058e-06,
+      "loss": 0.0171,
+      "step": 49340
+    },
+    {
+      "action_loss": 0.0018626055680215359,
+      "epoch": 44.3705035971223,
+      "step": 49340
+    },
+    {
+      "epoch": 44.3705035971223,
+      "step": 49340,
+      "torque_loss": 0.1551978439092636
+    },
+    {
+      "epoch": 44.3705035971223,
+      "force_loss": 0.0018640585476532578,
+      "step": 49340
+    },
+    {
+      "epoch": 44.3794964028777,
+      "grad_norm": 0.04401099681854248,
+      "learning_rate": 8.370720914747215e-06,
+      "loss": 0.0157,
+      "step": 49350
+    },
+    {
+      "action_loss": 0.0011097675887867808,
+      "epoch": 44.3794964028777,
+      "step": 49350
+    },
+    {
+      "epoch": 44.3794964028777,
+      "step": 49350,
+      "torque_loss": 0.060345251113176346
+    },
+    {
+      "epoch": 44.3794964028777,
+      "force_loss": 0.0016489513218402863,
+      "step": 49350
+    },
+    {
+      "epoch": 44.388489208633096,
+      "grad_norm": 0.34394213557243347,
+      "learning_rate": 8.355463054789181e-06,
+      "loss": 0.0155,
+      "step": 49360
+    },
+    {
+      "action_loss": 0.003258590353652835,
+      "epoch": 44.388489208633096,
+      "step": 49360
+    },
+    {
+      "epoch": 44.388489208633096,
+      "step": 49360,
+      "torque_loss": 0.12820221483707428
+    },
+    {
+      "epoch": 44.388489208633096,
+      "force_loss": 0.0032490657176822424,
+      "step": 49360
+    },
+    {
+      "epoch": 44.39748201438849,
+      "grad_norm": 0.5119810104370117,
+      "learning_rate": 8.340217845341919e-06,
+      "loss": 0.0183,
+      "step": 49370
+    },
+    {
+      "action_loss": 0.001605277881026268,
+      "epoch": 44.39748201438849,
+      "step": 49370
+    },
+    {
+      "epoch": 44.39748201438849,
+      "step": 49370,
+      "torque_loss": 0.13937672972679138
+    },
+    {
+      "epoch": 44.39748201438849,
+      "force_loss": 0.002603571629151702,
+      "step": 49370
+    },
+    {
+      "epoch": 44.406474820143885,
+      "grad_norm": 0.6002481579780579,
+      "learning_rate": 8.324985291036514e-06,
+      "loss": 0.0176,
+      "step": 49380
+    },
+    {
+      "action_loss": 0.0011088335886597633,
+      "epoch": 44.406474820143885,
+      "step": 49380
+    },
+    {
+      "epoch": 44.406474820143885,
+      "step": 49380,
+      "torque_loss": 0.12105637043714523
+    },
+    {
+      "epoch": 44.406474820143885,
+      "force_loss": 0.003954824525862932,
+      "step": 49380
+    },
+    {
+      "epoch": 44.41546762589928,
+      "grad_norm": 0.2561337649822235,
+      "learning_rate": 8.309765396500213e-06,
+      "loss": 0.0145,
+      "step": 49390
+    },
+    {
+      "action_loss": 0.002047070302069187,
+      "epoch": 44.41546762589928,
+      "step": 49390
+    },
+    {
+      "epoch": 44.41546762589928,
+      "step": 49390,
+      "torque_loss": 0.1376371830701828
+    },
+    {
+      "epoch": 44.41546762589928,
+      "force_loss": 0.00330272875726223,
+      "step": 49390
+    },
+    {
+      "epoch": 44.42446043165468,
+      "grad_norm": 0.18997599184513092,
+      "learning_rate": 8.294558166356419e-06,
+      "loss": 0.0144,
+      "step": 49400
+    },
+    {
+      "action_loss": 0.004301346838474274,
+      "epoch": 44.42446043165468,
+      "step": 49400
+    },
+    {
+      "epoch": 44.42446043165468,
+      "step": 49400,
+      "torque_loss": 0.11567624658346176
+    },
+    {
+      "epoch": 44.42446043165468,
+      "force_loss": 0.003047328442335129,
+      "step": 49400
+    },
+    {
+      "epoch": 44.43345323741007,
+      "grad_norm": 0.09938976168632507,
+      "learning_rate": 8.279363605224683e-06,
+      "loss": 0.0133,
+      "step": 49410
+    },
+    {
+      "action_loss": 0.003768334863707423,
+      "epoch": 44.43345323741007,
+      "step": 49410
+    },
+    {
+      "epoch": 44.43345323741007,
+      "step": 49410,
+      "torque_loss": 0.13831429183483124
+    },
+    {
+      "epoch": 44.43345323741007,
+      "force_loss": 0.0023436248302459717,
+      "step": 49410
+    },
+    {
+      "epoch": 44.44244604316547,
+      "grad_norm": 0.07708617299795151,
+      "learning_rate": 8.264181717720704e-06,
+      "loss": 0.014,
+      "step": 49420
+    },
+    {
+      "action_loss": 0.0058227344416081905,
+      "epoch": 44.44244604316547,
+      "step": 49420
+    },
+    {
+      "epoch": 44.44244604316547,
+      "step": 49420,
+      "torque_loss": 0.14846259355545044
+    },
+    {
+      "epoch": 44.44244604316547,
+      "force_loss": 0.013669963926076889,
+      "step": 49420
+    },
+    {
+      "epoch": 44.451438848920866,
+      "grad_norm": 0.11816935241222382,
+      "learning_rate": 8.249012508456361e-06,
+      "loss": 0.0159,
+      "step": 49430
+    },
+    {
+      "action_loss": 0.0023793091531842947,
+      "epoch": 44.451438848920866,
+      "step": 49430
+    },
+    {
+      "epoch": 44.451438848920866,
+      "step": 49430,
+      "torque_loss": 0.1441594362258911
+    },
+    {
+      "epoch": 44.451438848920866,
+      "force_loss": 0.003016282804310322,
+      "step": 49430
+    },
+    {
+      "epoch": 44.460431654676256,
+      "grad_norm": 0.08664724230766296,
+      "learning_rate": 8.233855982039646e-06,
+      "loss": 0.014,
+      "step": 49440
+    },
+    {
+      "action_loss": 0.001746080699376762,
+      "epoch": 44.460431654676256,
+      "step": 49440
+    },
+    {
+      "epoch": 44.460431654676256,
+      "step": 49440,
+      "torque_loss": 0.1471930742263794
+    },
+    {
+      "epoch": 44.460431654676256,
+      "force_loss": 0.0032019547652453184,
+      "step": 49440
+    },
+    {
+      "epoch": 44.469424460431654,
+      "grad_norm": 0.09860538691282272,
+      "learning_rate": 8.218712143074708e-06,
+      "loss": 0.0171,
+      "step": 49450
+    },
+    {
+      "action_loss": 0.0014260044554248452,
+      "epoch": 44.469424460431654,
+      "step": 49450
+    },
+    {
+      "epoch": 44.469424460431654,
+      "step": 49450,
+      "torque_loss": 0.12285075336694717
+    },
+    {
+      "epoch": 44.469424460431654,
+      "force_loss": 0.0033392657060176134,
+      "step": 49450
+    },
+    {
+      "epoch": 44.47841726618705,
+      "grad_norm": 0.06341202557086945,
+      "learning_rate": 8.203580996161858e-06,
+      "loss": 0.0135,
+      "step": 49460
+    },
+    {
+      "action_loss": 0.0006078608566895127,
+      "epoch": 44.47841726618705,
+      "step": 49460
+    },
+    {
+      "epoch": 44.47841726618705,
+      "step": 49460,
+      "torque_loss": 0.13634148240089417
+    },
+    {
+      "epoch": 44.47841726618705,
+      "force_loss": 0.001436496269889176,
+      "step": 49460
+    },
+    {
+      "epoch": 44.48741007194245,
+      "grad_norm": 0.2088707834482193,
+      "learning_rate": 8.188462545897512e-06,
+      "loss": 0.014,
+      "step": 49470
+    },
+    {
+      "action_loss": 0.002264623064547777,
+      "epoch": 44.48741007194245,
+      "step": 49470
+    },
+    {
+      "epoch": 44.48741007194245,
+      "step": 49470,
+      "torque_loss": 0.09197241067886353
+    },
+    {
+      "epoch": 44.48741007194245,
+      "force_loss": 0.0027277201879769564,
+      "step": 49470
+    },
+    {
+      "epoch": 44.49640287769784,
+      "grad_norm": 0.09228070825338364,
+      "learning_rate": 8.173356796874304e-06,
+      "loss": 0.0126,
+      "step": 49480
+    },
+    {
+      "action_loss": 0.002317664911970496,
+      "epoch": 44.49640287769784,
+      "step": 49480
+    },
+    {
+      "epoch": 44.49640287769784,
+      "step": 49480,
+      "torque_loss": 0.0934019684791565
+    },
+    {
+      "epoch": 44.49640287769784,
+      "force_loss": 0.009517152793705463,
+      "step": 49480
+    },
+    {
+      "epoch": 44.50539568345324,
+      "grad_norm": 0.06284049898386002,
+      "learning_rate": 8.158263753680906e-06,
+      "loss": 0.0172,
+      "step": 49490
+    },
+    {
+      "action_loss": 0.002216239692643285,
+      "epoch": 44.50539568345324,
+      "step": 49490
+    },
+    {
+      "epoch": 44.50539568345324,
+      "step": 49490,
+      "torque_loss": 0.14668191969394684
+    },
+    {
+      "epoch": 44.50539568345324,
+      "force_loss": 0.0034530481789261103,
+      "step": 49490
+    },
+    {
+      "epoch": 44.514388489208635,
+      "grad_norm": 0.09228930622339249,
+      "learning_rate": 8.143183420902239e-06,
+      "loss": 0.0144,
+      "step": 49500
+    },
+    {
+      "action_loss": 0.0014816907932981849,
+      "epoch": 44.514388489208635,
+      "step": 49500
+    },
+    {
+      "epoch": 44.514388489208635,
+      "step": 49500,
+      "torque_loss": 0.10810700803995132
+    },
+    {
+      "epoch": 44.514388489208635,
+      "force_loss": 0.009996780194342136,
+      "step": 49500
+    },
+    {
+      "epoch": 44.523381294964025,
+      "grad_norm": 0.2832128405570984,
+      "learning_rate": 8.128115803119258e-06,
+      "loss": 0.0135,
+      "step": 49510
+    },
+    {
+      "action_loss": 0.0010376755381003022,
+      "epoch": 44.523381294964025,
+      "step": 49510
+    },
+    {
+      "epoch": 44.523381294964025,
+      "step": 49510,
+      "torque_loss": 0.12448078393936157
+    },
+    {
+      "epoch": 44.523381294964025,
+      "force_loss": 0.0013605793938040733,
+      "step": 49510
+    },
+    {
+      "epoch": 44.53237410071942,
+      "grad_norm": 0.12625567615032196,
+      "learning_rate": 8.11306090490916e-06,
+      "loss": 0.0134,
+      "step": 49520
+    },
+    {
+      "action_loss": 0.0011574003146961331,
+      "epoch": 44.53237410071942,
+      "step": 49520
+    },
+    {
+      "epoch": 44.53237410071942,
+      "step": 49520,
+      "torque_loss": 0.13235025107860565
+    },
+    {
+      "epoch": 44.53237410071942,
+      "force_loss": 0.0017351250862702727,
+      "step": 49520
+    },
+    {
+      "epoch": 44.54136690647482,
+      "grad_norm": 0.05436331033706665,
+      "learning_rate": 8.098018730845169e-06,
+      "loss": 0.0139,
+      "step": 49530
+    },
+    {
+      "action_loss": 0.0009508302318863571,
+      "epoch": 44.54136690647482,
+      "step": 49530
+    },
+    {
+      "epoch": 44.54136690647482,
+      "step": 49530,
+      "torque_loss": 0.14779029786586761
+    },
+    {
+      "epoch": 44.54136690647482,
+      "force_loss": 0.0007588908192701638,
+      "step": 49530
+    },
+    {
+      "epoch": 44.55035971223022,
+      "grad_norm": 0.11662055552005768,
+      "learning_rate": 8.082989285496745e-06,
+      "loss": 0.014,
+      "step": 49540
+    },
+    {
+      "action_loss": 0.0014115944504737854,
+      "epoch": 44.55035971223022,
+      "step": 49540
+    },
+    {
+      "epoch": 44.55035971223022,
+      "step": 49540,
+      "torque_loss": 0.07627858966588974
+    },
+    {
+      "epoch": 44.55035971223022,
+      "force_loss": 0.003981321584433317,
+      "step": 49540
+    },
+    {
+      "epoch": 44.55935251798561,
+      "grad_norm": 0.047712501138448715,
+      "learning_rate": 8.067972573429416e-06,
+      "loss": 0.0149,
+      "step": 49550
+    },
+    {
+      "action_loss": 0.0008498501847498119,
+      "epoch": 44.55935251798561,
+      "step": 49550
+    },
+    {
+      "epoch": 44.55935251798561,
+      "step": 49550,
+      "torque_loss": 0.08005679398775101
+    },
+    {
+      "epoch": 44.55935251798561,
+      "force_loss": 0.0014413176104426384,
+      "step": 49550
+    },
+    {
+      "epoch": 44.568345323741006,
+      "grad_norm": 0.48789462447166443,
+      "learning_rate": 8.052968599204874e-06,
+      "loss": 0.0165,
+      "step": 49560
+    },
+    {
+      "action_loss": 0.004135263618081808,
+      "epoch": 44.568345323741006,
+      "step": 49560
+    },
+    {
+      "epoch": 44.568345323741006,
+      "step": 49560,
+      "torque_loss": 0.10088077932596207
+    },
+    {
+      "epoch": 44.568345323741006,
+      "force_loss": 0.006796931382268667,
+      "step": 49560
+    },
+    {
+      "epoch": 44.577338129496404,
+      "grad_norm": 0.10422906279563904,
+      "learning_rate": 8.037977367380922e-06,
+      "loss": 0.0141,
+      "step": 49570
+    },
+    {
+      "action_loss": 0.0007956252084113657,
+      "epoch": 44.577338129496404,
+      "step": 49570
+    },
+    {
+      "epoch": 44.577338129496404,
+      "step": 49570,
+      "torque_loss": 0.16722308099269867
+    },
+    {
+      "epoch": 44.577338129496404,
+      "force_loss": 0.001192752388305962,
+      "step": 49570
+    },
+    {
+      "epoch": 44.5863309352518,
+      "grad_norm": 0.29155808687210083,
+      "learning_rate": 8.022998882511495e-06,
+      "loss": 0.0169,
+      "step": 49580
+    },
+    {
+      "action_loss": 0.0014681977918371558,
+      "epoch": 44.5863309352518,
+      "step": 49580
+    },
+    {
+      "epoch": 44.5863309352518,
+      "step": 49580,
+      "torque_loss": 0.14347128570079803
+    },
+    {
+      "epoch": 44.5863309352518,
+      "force_loss": 0.0034073113929480314,
+      "step": 49580
+    },
+    {
+      "epoch": 44.59532374100719,
+      "grad_norm": 0.2645726501941681,
+      "learning_rate": 8.008033149146677e-06,
+      "loss": 0.0147,
+      "step": 49590
+    },
+    {
+      "action_loss": 0.005384637042880058,
+      "epoch": 44.59532374100719,
+      "step": 49590
+    },
+    {
+      "epoch": 44.59532374100719,
+      "step": 49590,
+      "torque_loss": 0.10157445818185806
+    },
+    {
+      "epoch": 44.59532374100719,
+      "force_loss": 0.010809145867824554,
+      "step": 49590
+    },
+    {
+      "epoch": 44.60431654676259,
+      "grad_norm": 0.21352891623973846,
+      "learning_rate": 7.993080171832656e-06,
+      "loss": 0.0144,
+      "step": 49600
+    },
+    {
+      "action_loss": 0.0006976434378884733,
+      "epoch": 44.60431654676259,
+      "step": 49600
+    },
+    {
+      "epoch": 44.60431654676259,
+      "step": 49600,
+      "torque_loss": 0.12141027301549911
+    },
+    {
+      "epoch": 44.60431654676259,
+      "force_loss": 0.0011753566795960069,
+      "step": 49600
+    },
+    {
+      "epoch": 44.61330935251799,
+      "grad_norm": 0.06730925291776657,
+      "learning_rate": 7.978139955111752e-06,
+      "loss": 0.0161,
+      "step": 49610
+    },
+    {
+      "action_loss": 0.003273202106356621,
+      "epoch": 44.61330935251799,
+      "step": 49610
+    },
+    {
+      "epoch": 44.61330935251799,
+      "step": 49610,
+      "torque_loss": 0.11216571182012558
+    },
+    {
+      "epoch": 44.61330935251799,
+      "force_loss": 0.010554053820669651,
+      "step": 49610
+    },
+    {
+      "epoch": 44.62230215827338,
+      "grad_norm": 0.04671144112944603,
+      "learning_rate": 7.9632125035224e-06,
+      "loss": 0.0157,
+      "step": 49620
+    },
+    {
+      "action_loss": 0.002263119211420417,
+      "epoch": 44.62230215827338,
+      "step": 49620
+    },
+    {
+      "epoch": 44.62230215827338,
+      "step": 49620,
+      "torque_loss": 0.1412576138973236
+    },
+    {
+      "epoch": 44.62230215827338,
+      "force_loss": 0.0049514793790876865,
+      "step": 49620
+    },
+    {
+      "epoch": 44.631294964028775,
+      "grad_norm": 0.31017956137657166,
+      "learning_rate": 7.948297821599177e-06,
+      "loss": 0.0138,
+      "step": 49630
+    },
+    {
+      "action_loss": 0.0010347046190872788,
+      "epoch": 44.631294964028775,
+      "step": 49630
+    },
+    {
+      "epoch": 44.631294964028775,
+      "step": 49630,
+      "torque_loss": 0.10887827724218369
+    },
+    {
+      "epoch": 44.631294964028775,
+      "force_loss": 0.0031630315352231264,
+      "step": 49630
+    },
+    {
+      "epoch": 44.64028776978417,
+      "grad_norm": 0.15130667388439178,
+      "learning_rate": 7.933395913872755e-06,
+      "loss": 0.0142,
+      "step": 49640
+    },
+    {
+      "action_loss": 0.008818390779197216,
+      "epoch": 44.64028776978417,
+      "step": 49640
+    },
+    {
+      "epoch": 44.64028776978417,
+      "step": 49640,
+      "torque_loss": 0.15408599376678467
+    },
+    {
+      "epoch": 44.64028776978417,
+      "force_loss": 0.011250992305576801,
+      "step": 49640
+    },
+    {
+      "epoch": 44.64928057553957,
+      "grad_norm": 0.2945329546928406,
+      "learning_rate": 7.918506784869972e-06,
+      "loss": 0.0152,
+      "step": 49650
+    },
+    {
+      "action_loss": 0.0011330684646964073,
+      "epoch": 44.64928057553957,
+      "step": 49650
+    },
+    {
+      "epoch": 44.64928057553957,
+      "step": 49650,
+      "torque_loss": 0.12201186269521713
+    },
+    {
+      "epoch": 44.64928057553957,
+      "force_loss": 0.0032301582396030426,
+      "step": 49650
+    },
+    {
+      "epoch": 44.65827338129496,
+      "grad_norm": 0.05091472342610359,
+      "learning_rate": 7.903630439113707e-06,
+      "loss": 0.0158,
+      "step": 49660
+    },
+    {
+      "action_loss": 0.0014601141447201371,
+      "epoch": 44.65827338129496,
+      "step": 49660
+    },
+    {
+      "epoch": 44.65827338129496,
+      "step": 49660,
+      "torque_loss": 0.1302904486656189
+    },
+    {
+      "epoch": 44.65827338129496,
+      "force_loss": 0.00391970993950963,
+      "step": 49660
+    },
+    {
+      "epoch": 44.66726618705036,
+      "grad_norm": 0.14738763868808746,
+      "learning_rate": 7.888766881123044e-06,
+      "loss": 0.0159,
+      "step": 49670
+    },
+    {
+      "action_loss": 0.0007106612320058048,
+      "epoch": 44.66726618705036,
+      "step": 49670
+    },
+    {
+      "epoch": 44.66726618705036,
+      "step": 49670,
+      "torque_loss": 0.10596217960119247
+    },
+    {
+      "epoch": 44.66726618705036,
+      "force_loss": 0.0027635181322693825,
+      "step": 49670
+    },
+    {
+      "epoch": 44.67625899280576,
+      "grad_norm": 0.18586307764053345,
+      "learning_rate": 7.873916115413099e-06,
+      "loss": 0.0141,
+      "step": 49680
+    },
+    {
+      "action_loss": 0.0012354138307273388,
+      "epoch": 44.67625899280576,
+      "step": 49680
+    },
+    {
+      "epoch": 44.67625899280576,
+      "step": 49680,
+      "torque_loss": 0.12358730286359787
+    },
+    {
+      "epoch": 44.67625899280576,
+      "force_loss": 0.0026613890659064054,
+      "step": 49680
+    },
+    {
+      "epoch": 44.685251798561154,
+      "grad_norm": 0.06968514621257782,
+      "learning_rate": 7.85907814649518e-06,
+      "loss": 0.0187,
+      "step": 49690
+    },
+    {
+      "action_loss": 0.0010834741406142712,
+      "epoch": 44.685251798561154,
+      "step": 49690
+    },
+    {
+      "epoch": 44.685251798561154,
+      "step": 49690,
+      "torque_loss": 0.12469297647476196
+    },
+    {
+      "epoch": 44.685251798561154,
+      "force_loss": 0.00414871284738183,
+      "step": 49690
+    },
+    {
+      "epoch": 44.694244604316545,
+      "grad_norm": 0.04685192182660103,
+      "learning_rate": 7.844252978876649e-06,
+      "loss": 0.016,
+      "step": 49700
+    },
+    {
+      "action_loss": 0.0021343789994716644,
+      "epoch": 44.694244604316545,
+      "step": 49700
+    },
+    {
+      "epoch": 44.694244604316545,
+      "step": 49700,
+      "torque_loss": 0.12869907915592194
+    },
+    {
+      "epoch": 44.694244604316545,
+      "force_loss": 0.010410395450890064,
+      "step": 49700
+    },
+    {
+      "epoch": 44.70323741007194,
+      "grad_norm": 0.5662211179733276,
+      "learning_rate": 7.829440617061001e-06,
+      "loss": 0.0169,
+      "step": 49710
+    },
+    {
+      "action_loss": 0.003900849027559161,
+      "epoch": 44.70323741007194,
+      "step": 49710
+    },
+    {
+      "epoch": 44.70323741007194,
+      "step": 49710,
+      "torque_loss": 0.11429709196090698
+    },
+    {
+      "epoch": 44.70323741007194,
+      "force_loss": 0.004489378537982702,
+      "step": 49710
+    },
+    {
+      "epoch": 44.71223021582734,
+      "grad_norm": 0.0439014658331871,
+      "learning_rate": 7.814641065547851e-06,
+      "loss": 0.0168,
+      "step": 49720
+    },
+    {
+      "action_loss": 0.0008110936614684761,
+      "epoch": 44.71223021582734,
+      "step": 49720
+    },
+    {
+      "epoch": 44.71223021582734,
+      "step": 49720,
+      "torque_loss": 0.15931938588619232
+    },
+    {
+      "epoch": 44.71223021582734,
+      "force_loss": 0.0017350729322060943,
+      "step": 49720
+    },
+    {
+      "epoch": 44.72122302158273,
+      "grad_norm": 0.05959755927324295,
+      "learning_rate": 7.79985432883289e-06,
+      "loss": 0.013,
+      "step": 49730
+    },
+    {
+      "action_loss": 0.0011860455852001905,
+      "epoch": 44.72122302158273,
+      "step": 49730
+    },
+    {
+      "epoch": 44.72122302158273,
+      "step": 49730,
+      "torque_loss": 0.12883572280406952
+    },
+    {
+      "epoch": 44.72122302158273,
+      "force_loss": 0.0029928397852927446,
+      "step": 49730
+    },
+    {
+      "epoch": 44.73021582733813,
+      "grad_norm": 0.08484892547130585,
+      "learning_rate": 7.78508041140797e-06,
+      "loss": 0.0143,
+      "step": 49740
+    },
+    {
+      "action_loss": 0.0005358399939723313,
+      "epoch": 44.73021582733813,
+      "step": 49740
+    },
+    {
+      "epoch": 44.73021582733813,
+      "step": 49740,
+      "torque_loss": 0.09793832153081894
+    },
+    {
+      "epoch": 44.73021582733813,
+      "force_loss": 0.0011657123686745763,
+      "step": 49740
+    },
+    {
+      "epoch": 44.739208633093526,
+      "grad_norm": 0.3519071042537689,
+      "learning_rate": 7.770319317760993e-06,
+      "loss": 0.0155,
+      "step": 49750
+    },
+    {
+      "action_loss": 0.000978046446107328,
+      "epoch": 44.739208633093526,
+      "step": 49750
+    },
+    {
+      "epoch": 44.739208633093526,
+      "step": 49750,
+      "torque_loss": 0.11245503276586533
+    },
+    {
+      "epoch": 44.739208633093526,
+      "force_loss": 0.003304282436147332,
+      "step": 49750
+    },
+    {
+      "epoch": 44.74820143884892,
+      "grad_norm": 0.06182730942964554,
+      "learning_rate": 7.755571052376004e-06,
+      "loss": 0.0146,
+      "step": 49760
+    },
+    {
+      "action_loss": 0.0021554799750447273,
+      "epoch": 44.74820143884892,
+      "step": 49760
+    },
+    {
+      "epoch": 44.74820143884892,
+      "step": 49760,
+      "torque_loss": 0.15364956855773926
+    },
+    {
+      "epoch": 44.74820143884892,
+      "force_loss": 0.004587630741298199,
+      "step": 49760
+    },
+    {
+      "epoch": 44.757194244604314,
+      "grad_norm": 0.07076491415500641,
+      "learning_rate": 7.740835619733128e-06,
+      "loss": 0.0135,
+      "step": 49770
+    },
+    {
+      "action_loss": 0.003983645234256983,
+      "epoch": 44.757194244604314,
+      "step": 49770
+    },
+    {
+      "epoch": 44.757194244604314,
+      "step": 49770,
+      "torque_loss": 0.1310625970363617
+    },
+    {
+      "epoch": 44.757194244604314,
+      "force_loss": 0.006321063730865717,
+      "step": 49770
+    },
+    {
+      "epoch": 44.76618705035971,
+      "grad_norm": 0.10592120885848999,
+      "learning_rate": 7.726113024308601e-06,
+      "loss": 0.0159,
+      "step": 49780
+    },
+    {
+      "action_loss": 0.0009449923527427018,
+      "epoch": 44.76618705035971,
+      "step": 49780
+    },
+    {
+      "epoch": 44.76618705035971,
+      "step": 49780,
+      "torque_loss": 0.16312932968139648
+    },
+    {
+      "epoch": 44.76618705035971,
+      "force_loss": 0.0033867955207824707,
+      "step": 49780
+    },
+    {
+      "epoch": 44.77517985611511,
+      "grad_norm": 0.15752017498016357,
+      "learning_rate": 7.711403270574746e-06,
+      "loss": 0.0167,
+      "step": 49790
+    },
+    {
+      "action_loss": 0.002825170988216996,
+      "epoch": 44.77517985611511,
+      "step": 49790
+    },
+    {
+      "epoch": 44.77517985611511,
+      "step": 49790,
+      "torque_loss": 0.1041218712925911
+    },
+    {
+      "epoch": 44.77517985611511,
+      "force_loss": 0.007644938305020332,
+      "step": 49790
+    },
+    {
+      "epoch": 44.78417266187051,
+      "grad_norm": 0.0501573346555233,
+      "learning_rate": 7.696706363000039e-06,
+      "loss": 0.0141,
+      "step": 49800
+    },
+    {
+      "action_loss": 0.0017316960729658604,
+      "epoch": 44.78417266187051,
+      "step": 49800
+    },
+    {
+      "epoch": 44.78417266187051,
+      "step": 49800,
+      "torque_loss": 0.12151408195495605
+    },
+    {
+      "epoch": 44.78417266187051,
+      "force_loss": 0.002438431140035391,
+      "step": 49800
+    },
+    {
+      "epoch": 44.7931654676259,
+      "grad_norm": 0.16803665459156036,
+      "learning_rate": 7.682022306048959e-06,
+      "loss": 0.0137,
+      "step": 49810
+    },
+    {
+      "action_loss": 0.00035620061680674553,
+      "epoch": 44.7931654676259,
+      "step": 49810
+    },
+    {
+      "epoch": 44.7931654676259,
+      "step": 49810,
+      "torque_loss": 0.14025768637657166
+    },
+    {
+      "epoch": 44.7931654676259,
+      "force_loss": 0.0005227602086961269,
+      "step": 49810
+    },
+    {
+      "epoch": 44.802158273381295,
+      "grad_norm": 0.20629088580608368,
+      "learning_rate": 7.667351104182186e-06,
+      "loss": 0.0149,
+      "step": 49820
+    },
+    {
+      "action_loss": 0.0008661316242069006,
+      "epoch": 44.802158273381295,
+      "step": 49820
+    },
+    {
+      "epoch": 44.802158273381295,
+      "step": 49820,
+      "torque_loss": 0.13550396263599396
+    },
+    {
+      "epoch": 44.802158273381295,
+      "force_loss": 0.002790125785395503,
+      "step": 49820
+    },
+    {
+      "epoch": 44.81115107913669,
+      "grad_norm": 0.051408298313617706,
+      "learning_rate": 7.652692761856395e-06,
+      "loss": 0.0134,
+      "step": 49830
+    },
+    {
+      "action_loss": 0.0007608411251567304,
+      "epoch": 44.81115107913669,
+      "step": 49830
+    },
+    {
+      "epoch": 44.81115107913669,
+      "step": 49830,
+      "torque_loss": 0.11166875809431076
+    },
+    {
+      "epoch": 44.81115107913669,
+      "force_loss": 0.0031106378883123398,
+      "step": 49830
+    },
+    {
+      "epoch": 44.82014388489208,
+      "grad_norm": 0.09923563152551651,
+      "learning_rate": 7.63804728352444e-06,
+      "loss": 0.0142,
+      "step": 49840
+    },
+    {
+      "action_loss": 0.0020305428188294172,
+      "epoch": 44.82014388489208,
+      "step": 49840
+    },
+    {
+      "epoch": 44.82014388489208,
+      "step": 49840,
+      "torque_loss": 0.1581265926361084
+    },
+    {
+      "epoch": 44.82014388489208,
+      "force_loss": 0.007345882710069418,
+      "step": 49840
+    },
+    {
+      "epoch": 44.82913669064748,
+      "grad_norm": 0.3739672303199768,
+      "learning_rate": 7.623414673635215e-06,
+      "loss": 0.0146,
+      "step": 49850
+    },
+    {
+      "action_loss": 0.0009307024884037673,
+      "epoch": 44.82913669064748,
+      "step": 49850
+    },
+    {
+      "epoch": 44.82913669064748,
+      "step": 49850,
+      "torque_loss": 0.1286327838897705
+    },
+    {
+      "epoch": 44.82913669064748,
+      "force_loss": 0.0067672282457351685,
+      "step": 49850
+    },
+    {
+      "epoch": 44.83812949640288,
+      "grad_norm": 0.28007903695106506,
+      "learning_rate": 7.608794936633723e-06,
+      "loss": 0.0205,
+      "step": 49860
+    },
+    {
+      "action_loss": 0.0015352214686572552,
+      "epoch": 44.83812949640288,
+      "step": 49860
+    },
+    {
+      "epoch": 44.83812949640288,
+      "step": 49860,
+      "torque_loss": 0.11712837219238281
+    },
+    {
+      "epoch": 44.83812949640288,
+      "force_loss": 0.0037732094060629606,
+      "step": 49860
+    },
+    {
+      "epoch": 44.847122302158276,
+      "grad_norm": 0.06705943495035172,
+      "learning_rate": 7.594188076961056e-06,
+      "loss": 0.0168,
+      "step": 49870
+    },
+    {
+      "action_loss": 0.0010187792358919978,
+      "epoch": 44.847122302158276,
+      "step": 49870
+    },
+    {
+      "epoch": 44.847122302158276,
+      "step": 49870,
+      "torque_loss": 0.110748790204525
+    },
+    {
+      "epoch": 44.847122302158276,
+      "force_loss": 0.0007314844406209886,
+      "step": 49870
+    },
+    {
+      "epoch": 44.856115107913666,
+      "grad_norm": 0.24257680773735046,
+      "learning_rate": 7.579594099054382e-06,
+      "loss": 0.0161,
+      "step": 49880
+    },
+    {
+      "action_loss": 0.001507192268036306,
+      "epoch": 44.856115107913666,
+      "step": 49880
+    },
+    {
+      "epoch": 44.856115107913666,
+      "step": 49880,
+      "torque_loss": 0.11246427893638611
+    },
+    {
+      "epoch": 44.856115107913666,
+      "force_loss": 0.006308602634817362,
+      "step": 49880
+    },
+    {
+      "epoch": 44.865107913669064,
+      "grad_norm": 0.07981456071138382,
+      "learning_rate": 7.565013007346983e-06,
+      "loss": 0.014,
+      "step": 49890
+    },
+    {
+      "action_loss": 0.005797725170850754,
+      "epoch": 44.865107913669064,
+      "step": 49890
+    },
+    {
+      "epoch": 44.865107913669064,
+      "step": 49890,
+      "torque_loss": 0.2362171858549118
+    },
+    {
+      "epoch": 44.865107913669064,
+      "force_loss": 0.0074445023201406,
+      "step": 49890
+    },
+    {
+      "epoch": 44.87410071942446,
+      "grad_norm": 0.11106700450181961,
+      "learning_rate": 7.5504448062682035e-06,
+      "loss": 0.0176,
+      "step": 49900
+    },
+    {
+      "action_loss": 0.0008172406814992428,
+      "epoch": 44.87410071942446,
+      "step": 49900
+    },
+    {
+      "epoch": 44.87410071942446,
+      "step": 49900,
+      "torque_loss": 0.13382510840892792
+    },
+    {
+      "epoch": 44.87410071942446,
+      "force_loss": 0.0021993634290993214,
+      "step": 49900
+    },
+    {
+      "epoch": 44.88309352517986,
+      "grad_norm": 0.30454111099243164,
+      "learning_rate": 7.53588950024347e-06,
+      "loss": 0.0156,
+      "step": 49910
+    },
+    {
+      "action_loss": 0.001323816948570311,
+      "epoch": 44.88309352517986,
+      "step": 49910
+    },
+    {
+      "epoch": 44.88309352517986,
+      "step": 49910,
+      "torque_loss": 0.10558053851127625
+    },
+    {
+      "epoch": 44.88309352517986,
+      "force_loss": 0.0026909213047474623,
+      "step": 49910
+    },
+    {
+      "epoch": 44.89208633093525,
+      "grad_norm": 0.11809254437685013,
+      "learning_rate": 7.5213470936943145e-06,
+      "loss": 0.0147,
+      "step": 49920
+    },
+    {
+      "action_loss": 0.004369642119854689,
+      "epoch": 44.89208633093525,
+      "step": 49920
+    },
+    {
+      "epoch": 44.89208633093525,
+      "step": 49920,
+      "torque_loss": 0.11722224950790405
+    },
+    {
+      "epoch": 44.89208633093525,
+      "force_loss": 0.010220506228506565,
+      "step": 49920
+    },
+    {
+      "epoch": 44.90107913669065,
+      "grad_norm": 0.07059824466705322,
+      "learning_rate": 7.506817591038323e-06,
+      "loss": 0.0172,
+      "step": 49930
+    },
+    {
+      "action_loss": 0.0011751716956496239,
+      "epoch": 44.90107913669065,
+      "step": 49930
+    },
+    {
+      "epoch": 44.90107913669065,
+      "step": 49930,
+      "torque_loss": 0.07866986840963364
+    },
+    {
+      "epoch": 44.90107913669065,
+      "force_loss": 0.0024234424345195293,
+      "step": 49930
+    },
+    {
+      "epoch": 44.910071942446045,
+      "grad_norm": 0.12487969547510147,
+      "learning_rate": 7.492300996689183e-06,
+      "loss": 0.0245,
+      "step": 49940
+    },
+    {
+      "action_loss": 0.0012136775767430663,
+      "epoch": 44.910071942446045,
+      "step": 49940
+    },
+    {
+      "epoch": 44.910071942446045,
+      "step": 49940,
+      "torque_loss": 0.08027943968772888
+    },
+    {
+      "epoch": 44.910071942446045,
+      "force_loss": 0.0012687615817412734,
+      "step": 49940
+    },
+    {
+      "epoch": 44.919064748201436,
+      "grad_norm": 0.36301109194755554,
+      "learning_rate": 7.477797315056645e-06,
+      "loss": 0.0155,
+      "step": 49950
+    },
+    {
+      "action_loss": 0.004123583901673555,
+      "epoch": 44.919064748201436,
+      "step": 49950
+    },
+    {
+      "epoch": 44.919064748201436,
+      "step": 49950,
+      "torque_loss": 0.15322358906269073
+    },
+    {
+      "epoch": 44.919064748201436,
+      "force_loss": 0.010679654777050018,
+      "step": 49950
+    },
+    {
+      "epoch": 44.92805755395683,
+      "grad_norm": 0.30525267124176025,
+      "learning_rate": 7.463306550546539e-06,
+      "loss": 0.0144,
+      "step": 49960
+    },
+    {
+      "action_loss": 0.003829879453405738,
+      "epoch": 44.92805755395683,
+      "step": 49960
+    },
+    {
+      "epoch": 44.92805755395683,
+      "step": 49960,
+      "torque_loss": 0.13717468082904816
+    },
+    {
+      "epoch": 44.92805755395683,
+      "force_loss": 0.0047232164070010185,
+      "step": 49960
+    },
+    {
+      "epoch": 44.93705035971223,
+      "grad_norm": 0.10683489590883255,
+      "learning_rate": 7.448828707560812e-06,
+      "loss": 0.0136,
+      "step": 49970
+    },
+    {
+      "action_loss": 0.006063038017600775,
+      "epoch": 44.93705035971223,
+      "step": 49970
+    },
+    {
+      "epoch": 44.93705035971223,
+      "step": 49970,
+      "torque_loss": 0.1413004845380783
+    },
+    {
+      "epoch": 44.93705035971223,
+      "force_loss": 0.009604431688785553,
+      "step": 49970
+    },
+    {
+      "epoch": 44.94604316546763,
+      "grad_norm": 0.07475993037223816,
+      "learning_rate": 7.4343637904974e-06,
+      "loss": 0.0151,
+      "step": 49980
+    },
+    {
+      "action_loss": 0.003994505386799574,
+      "epoch": 44.94604316546763,
+      "step": 49980
+    },
+    {
+      "epoch": 44.94604316546763,
+      "step": 49980,
+      "torque_loss": 0.10486545413732529
+    },
+    {
+      "epoch": 44.94604316546763,
+      "force_loss": 0.008612277917563915,
+      "step": 49980
+    },
+    {
+      "epoch": 44.95503597122302,
+      "grad_norm": 0.08893385529518127,
+      "learning_rate": 7.419911803750401e-06,
+      "loss": 0.015,
+      "step": 49990
+    },
+    {
+      "action_loss": 0.0025657371152192354,
+      "epoch": 44.95503597122302,
+      "step": 49990
+    },
+    {
+      "epoch": 44.95503597122302,
+      "step": 49990,
+      "torque_loss": 0.13204152882099152
+    },
+    {
+      "epoch": 44.95503597122302,
+      "force_loss": 0.006318904925137758,
+      "step": 49990
+    },
+    {
+      "epoch": 44.96402877697842,
+      "grad_norm": 0.21625585854053497,
+      "learning_rate": 7.405472751709935e-06,
+      "loss": 0.0191,
+      "step": 50000
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 60000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 54,
+  "save_steps": 10000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}