| {"step": 0, "epoch": 0, "total_loss": 0.0380440428853035, "alignment_loss": 0.3175742030143738, "kl_loss": 0.00628662109375, "ert_budget": 0.75} |
| {"step": 1, "epoch": 0, "total_loss": 0.07489337772130966, "alignment_loss": 0.19015692174434662, "kl_loss": 0.055877685546875, "ert_budget": 0.75} |
| {"step": 2, "epoch": 0, "total_loss": 0.027266738936305046, "alignment_loss": 0.041954491287469864, "kl_loss": 0.0230712890625, "ert_budget": 0.75} |
| {"step": 3, "epoch": 0, "total_loss": 0.9451772570610046, "alignment_loss": 0.4039212167263031, "kl_loss": 0.90478515625, "ert_budget": 0.25} |
| {"step": 4, "epoch": 0, "total_loss": 0.2780088186264038, "alignment_loss": 0.07012724876403809, "kl_loss": 0.27099609375, "ert_budget": 0.25} |
| {"step": 5, "epoch": 0, "total_loss": 1.8360483646392822, "alignment_loss": 1.3878273963928223, "kl_loss": 1.697265625, "ert_budget": 0.25} |
| {"step": 6, "epoch": 0, "total_loss": 0.6480252742767334, "alignment_loss": 0.12771396338939667, "kl_loss": 0.63525390625, "ert_budget": 0.25} |
| {"step": 7, "epoch": 0, "total_loss": 2.323340892791748, "alignment_loss": 1.1045033931732178, "kl_loss": 2.212890625, "ert_budget": 0.25} |
| {"step": 8, "epoch": 0, "total_loss": 0.22986358404159546, "alignment_loss": 0.7068389654159546, "kl_loss": 0.1591796875, "ert_budget": 0.5} |
| {"step": 9, "epoch": 0, "total_loss": 1.649030327796936, "alignment_loss": 3.2969441413879395, "kl_loss": 1.3193359375, "ert_budget": 0.25} |
| {"step": 10, "epoch": 0, "total_loss": 0.11880166083574295, "alignment_loss": 0.8779579997062683, "kl_loss": 0.031005859375, "ert_budget": 0.75} |
| {"step": 11, "epoch": 0, "total_loss": 0.16462111473083496, "alignment_loss": 0.7691358327865601, "kl_loss": 0.08770751953125, "ert_budget": 0.75} |
| {"step": 12, "epoch": 0, "total_loss": 0.7627257108688354, "alignment_loss": 0.2981553077697754, "kl_loss": 0.73291015625, "ert_budget": 0.25} |
| {"step": 13, "epoch": 0, "total_loss": 0.06469056010246277, "alignment_loss": 0.16930550336837769, "kl_loss": 0.047760009765625, "ert_budget": 0.5} |
| {"step": 14, "epoch": 0, "total_loss": 0.32167500257492065, "alignment_loss": 0.013624962419271469, "kl_loss": 0.3203125, "ert_budget": 0.75} |
| {"step": 15, "epoch": 0, "total_loss": 0.03488689661026001, "alignment_loss": 0.2393871545791626, "kl_loss": 0.01094818115234375, "ert_budget": 0.75} |
| {"step": 16, "epoch": 0, "total_loss": 0.032873451709747314, "alignment_loss": 0.01913367584347725, "kl_loss": 0.0309600830078125, "ert_budget": 0.75} |
| {"step": 17, "epoch": 0, "total_loss": 0.10748714953660965, "alignment_loss": 0.3064388632774353, "kl_loss": 0.07684326171875, "ert_budget": 0.5} |
| {"step": 18, "epoch": 0, "total_loss": 1.29465913772583, "alignment_loss": 0.6126066446304321, "kl_loss": 1.2333984375, "ert_budget": 0.25} |
| {"step": 19, "epoch": 0, "total_loss": 1.4503875970840454, "alignment_loss": 0.6464542150497437, "kl_loss": 1.3857421875, "ert_budget": 0.5} |
| {"step": 20, "epoch": 0, "total_loss": 2.3323428630828857, "alignment_loss": 0.23749154806137085, "kl_loss": 2.30859375, "ert_budget": 0.25} |
| {"step": 21, "epoch": 0, "total_loss": 0.015629643574357033, "alignment_loss": 0.016449633985757828, "kl_loss": 0.01398468017578125, "ert_budget": 0.75} |
| {"step": 22, "epoch": 0, "total_loss": 1.7519681453704834, "alignment_loss": 0.10757209360599518, "kl_loss": 1.7412109375, "ert_budget": 0.75} |
| {"step": 23, "epoch": 0, "total_loss": 0.05736752972006798, "alignment_loss": 0.3136655390262604, "kl_loss": 0.0260009765625, "ert_budget": 0.5} |
| {"step": 24, "epoch": 0, "total_loss": 0.035560354590415955, "alignment_loss": 0.3048299252986908, "kl_loss": 0.005077362060546875, "ert_budget": 0.75} |
| {"step": 25, "epoch": 0, "total_loss": 0.03080439381301403, "alignment_loss": 0.007445787079632282, "kl_loss": 0.030059814453125, "ert_budget": 0.5} |
| {"step": 26, "epoch": 0, "total_loss": 0.02591961808502674, "alignment_loss": 0.019175425171852112, "kl_loss": 0.0240020751953125, "ert_budget": 0.75} |
| {"step": 27, "epoch": 0, "total_loss": 0.013377207331359386, "alignment_loss": 0.049009498208761215, "kl_loss": 0.00847625732421875, "ert_budget": 0.5} |
| {"step": 28, "epoch": 0, "total_loss": 0.06042125076055527, "alignment_loss": 0.041468385607004166, "kl_loss": 0.0562744140625, "ert_budget": 0.75} |
| {"step": 29, "epoch": 0, "total_loss": 0.015709349885582924, "alignment_loss": 0.003513783449307084, "kl_loss": 0.01535797119140625, "ert_budget": 0.5} |
| {"step": 30, "epoch": 0, "total_loss": 0.041080087423324585, "alignment_loss": 0.00888438057154417, "kl_loss": 0.040191650390625, "ert_budget": 0.5} |
| {"step": 31, "epoch": 0, "total_loss": 0.13435135781764984, "alignment_loss": 0.614143431186676, "kl_loss": 0.07293701171875, "ert_budget": 0.75} |
| {"step": 32, "epoch": 0, "total_loss": 0.5123229622840881, "alignment_loss": 1.588073492050171, "kl_loss": 0.353515625, "ert_budget": 0.5} |
| {"step": 33, "epoch": 0, "total_loss": 3.8968636989593506, "alignment_loss": 0.003792651928961277, "kl_loss": 3.896484375, "ert_budget": 0.25} |
| {"step": 34, "epoch": 0, "total_loss": 0.2857065796852112, "alignment_loss": 1.8591409921646118, "kl_loss": 0.09979248046875, "ert_budget": 0.5} |
| {"step": 35, "epoch": 0, "total_loss": 2.0257506370544434, "alignment_loss": 2.9528181552886963, "kl_loss": 1.73046875, "ert_budget": 0.25} |
| {"step": 36, "epoch": 0, "total_loss": 3.6764674186706543, "alignment_loss": 2.897487163543701, "kl_loss": 3.38671875, "ert_budget": 0.25} |
| {"step": 37, "epoch": 0, "total_loss": 1.7791311740875244, "alignment_loss": 0.8088901042938232, "kl_loss": 1.6982421875, "ert_budget": 0.25} |
| {"step": 38, "epoch": 0, "total_loss": 0.056623365730047226, "alignment_loss": 0.020274190232157707, "kl_loss": 0.054595947265625, "ert_budget": 0.75} |
| {"step": 39, "epoch": 0, "total_loss": 0.03924161195755005, "alignment_loss": 0.2775174081325531, "kl_loss": 0.0114898681640625, "ert_budget": 0.5} |
| {"step": 40, "epoch": 0, "total_loss": 1.8858648538589478, "alignment_loss": 1.1828668117523193, "kl_loss": 1.767578125, "ert_budget": 0.25} |
| {"step": 41, "epoch": 0, "total_loss": 0.08013869822025299, "alignment_loss": 0.7038832902908325, "kl_loss": 0.0097503662109375, "ert_budget": 0.75} |
| {"step": 42, "epoch": 0, "total_loss": 2.6273603439331055, "alignment_loss": 1.1368842124938965, "kl_loss": 2.513671875, "ert_budget": 0.25} |
| {"step": 43, "epoch": 0, "total_loss": 4.572003364562988, "alignment_loss": 0.7200352549552917, "kl_loss": 4.5, "ert_budget": 0.25} |
| {"step": 44, "epoch": 0, "total_loss": 0.07037802040576935, "alignment_loss": 0.0537557527422905, "kl_loss": 0.06500244140625, "ert_budget": 0.5} |
| {"step": 45, "epoch": 0, "total_loss": 0.02003193460404873, "alignment_loss": 0.026521746069192886, "kl_loss": 0.0173797607421875, "ert_budget": 0.75} |
| {"step": 46, "epoch": 0, "total_loss": 0.21546965837478638, "alignment_loss": 0.027011100202798843, "kl_loss": 0.2127685546875, "ert_budget": 0.75} |
| {"step": 47, "epoch": 0, "total_loss": 0.8648312091827393, "alignment_loss": 0.1815153956413269, "kl_loss": 0.8466796875, "ert_budget": 0.25} |
| {"step": 48, "epoch": 0, "total_loss": 0.05332450568675995, "alignment_loss": 0.2656058669090271, "kl_loss": 0.026763916015625, "ert_budget": 0.5} |
| {"step": 49, "epoch": 0, "total_loss": 0.07062545418739319, "alignment_loss": 0.4378523826599121, "kl_loss": 0.0268402099609375, "ert_budget": 0.5} |
| {"step": 50, "epoch": 0, "total_loss": 0.13119438290596008, "alignment_loss": 0.12419970333576202, "kl_loss": 0.1187744140625, "ert_budget": 0.5} |
| {"step": 51, "epoch": 0, "total_loss": 0.7289183139801025, "alignment_loss": 0.06750358641147614, "kl_loss": 0.72216796875, "ert_budget": 0.25} |
| {"step": 52, "epoch": 0, "total_loss": 2.3378841876983643, "alignment_loss": 0.01946737989783287, "kl_loss": 2.3359375, "ert_budget": 0.25} |
| {"step": 53, "epoch": 0, "total_loss": 0.05274183303117752, "alignment_loss": 0.013502303510904312, "kl_loss": 0.0513916015625, "ert_budget": 0.5} |
| {"step": 54, "epoch": 0, "total_loss": 0.0396343357861042, "alignment_loss": 0.06950011104345322, "kl_loss": 0.032684326171875, "ert_budget": 0.5} |
| {"step": 55, "epoch": 0, "total_loss": 1.018385410308838, "alignment_loss": 0.008072306402027607, "kl_loss": 1.017578125, "ert_budget": 0.25} |
| {"step": 56, "epoch": 0, "total_loss": 0.24863402545452118, "alignment_loss": 0.0034300731495022774, "kl_loss": 0.248291015625, "ert_budget": 0.5} |
| {"step": 57, "epoch": 0, "total_loss": 0.11496273428201675, "alignment_loss": 0.6070247888565063, "kl_loss": 0.05426025390625, "ert_budget": 0.5} |
| {"step": 58, "epoch": 0, "total_loss": 0.05089804530143738, "alignment_loss": 0.16535250842571259, "kl_loss": 0.03436279296875, "ert_budget": 0.75} |
| {"step": 59, "epoch": 0, "total_loss": 0.05798006057739258, "alignment_loss": 0.10250568389892578, "kl_loss": 0.0477294921875, "ert_budget": 0.5} |
| {"step": 60, "epoch": 0, "total_loss": 0.1406940072774887, "alignment_loss": 0.053180284798145294, "kl_loss": 0.1353759765625, "ert_budget": 0.75} |
| {"step": 61, "epoch": 0, "total_loss": 0.7856789827346802, "alignment_loss": 0.20542249083518982, "kl_loss": 0.76513671875, "ert_budget": 0.25} |
| {"step": 62, "epoch": 0, "total_loss": 0.04840497672557831, "alignment_loss": 0.015910102054476738, "kl_loss": 0.04681396484375, "ert_budget": 0.75} |
| {"step": 63, "epoch": 0, "total_loss": 0.06701909005641937, "alignment_loss": 0.08455856889486313, "kl_loss": 0.058563232421875, "ert_budget": 0.75} |
| {"step": 64, "epoch": 0, "total_loss": 0.06489812582731247, "alignment_loss": 0.3894292414188385, "kl_loss": 0.0259552001953125, "ert_budget": 0.75} |
| {"step": 65, "epoch": 0, "total_loss": 2.7165751457214355, "alignment_loss": 0.09543977677822113, "kl_loss": 2.70703125, "ert_budget": 0.25} |
| {"step": 66, "epoch": 0, "total_loss": 0.030559323728084564, "alignment_loss": 0.06023190915584564, "kl_loss": 0.0245361328125, "ert_budget": 0.75} |
| {"step": 67, "epoch": 0, "total_loss": 0.04364381358027458, "alignment_loss": 0.06107192486524582, "kl_loss": 0.03753662109375, "ert_budget": 0.5} |
| {"step": 68, "epoch": 0, "total_loss": 0.05471475422382355, "alignment_loss": 0.07107331603765488, "kl_loss": 0.047607421875, "ert_budget": 0.5} |
| {"step": 69, "epoch": 0, "total_loss": 0.8071276545524597, "alignment_loss": 0.014635691419243813, "kl_loss": 0.8056640625, "ert_budget": 0.25} |
| {"step": 70, "epoch": 0, "total_loss": 0.031328022480010986, "alignment_loss": 0.09477437287569046, "kl_loss": 0.0218505859375, "ert_budget": 0.5} |
| {"step": 71, "epoch": 0, "total_loss": 0.07454878836870193, "alignment_loss": 0.03809039667248726, "kl_loss": 0.07073974609375, "ert_budget": 0.5} |
| {"step": 72, "epoch": 0, "total_loss": 0.10160369426012039, "alignment_loss": 0.05717463046312332, "kl_loss": 0.09588623046875, "ert_budget": 0.75} |
| {"step": 73, "epoch": 0, "total_loss": 0.6217407584190369, "alignment_loss": 0.10412657260894775, "kl_loss": 0.611328125, "ert_budget": 0.25} |
| {"step": 74, "epoch": 0, "total_loss": 0.24316415190696716, "alignment_loss": 0.017090732231736183, "kl_loss": 0.241455078125, "ert_budget": 0.5} |
| {"step": 75, "epoch": 0, "total_loss": 0.7731280326843262, "alignment_loss": 0.1189754530787468, "kl_loss": 0.76123046875, "ert_budget": 0.25} |
| {"step": 76, "epoch": 0, "total_loss": 0.07487978041172028, "alignment_loss": 0.06398334354162216, "kl_loss": 0.0684814453125, "ert_budget": 0.75} |
| {"step": 77, "epoch": 0, "total_loss": 4.350356101989746, "alignment_loss": 0.026997651904821396, "kl_loss": 4.34765625, "ert_budget": 0.25} |
| {"step": 78, "epoch": 0, "total_loss": 0.03243019059300423, "alignment_loss": 0.14546890556812286, "kl_loss": 0.01788330078125, "ert_budget": 0.75} |
| {"step": 79, "epoch": 0, "total_loss": 0.2034475952386856, "alignment_loss": 1.2440706491470337, "kl_loss": 0.07904052734375, "ert_budget": 0.5} |
| {"step": 80, "epoch": 0, "total_loss": 0.824140727519989, "alignment_loss": 0.4337899684906006, "kl_loss": 0.78076171875, "ert_budget": 0.25} |
| {"step": 81, "epoch": 0, "total_loss": 1.5864249467849731, "alignment_loss": 0.0048742881044745445, "kl_loss": 1.5859375, "ert_budget": 0.25} |
| {"step": 82, "epoch": 0, "total_loss": 0.6960474252700806, "alignment_loss": 0.08059155941009521, "kl_loss": 0.68798828125, "ert_budget": 0.25} |
| {"step": 83, "epoch": 0, "total_loss": 0.04383426159620285, "alignment_loss": 0.0547366589307785, "kl_loss": 0.038360595703125, "ert_budget": 0.5} |
| {"step": 84, "epoch": 0, "total_loss": 0.04118781536817551, "alignment_loss": 0.24128486216068268, "kl_loss": 0.017059326171875, "ert_budget": 0.75} |
| {"step": 85, "epoch": 0, "total_loss": 0.8350921273231506, "alignment_loss": 0.07455426454544067, "kl_loss": 0.82763671875, "ert_budget": 0.25} |
| {"step": 86, "epoch": 0, "total_loss": 0.06150740012526512, "alignment_loss": 0.07308182120323181, "kl_loss": 0.05419921875, "ert_budget": 0.5} |
| {"step": 87, "epoch": 0, "total_loss": 0.01202632300555706, "alignment_loss": 0.024972088634967804, "kl_loss": 0.00952911376953125, "ert_budget": 0.75} |
| {"step": 88, "epoch": 0, "total_loss": 1.3873120546340942, "alignment_loss": 0.669995903968811, "kl_loss": 1.3203125, "ert_budget": 0.25} |
| {"step": 89, "epoch": 0, "total_loss": 0.1096559390425682, "alignment_loss": 0.7758196592330933, "kl_loss": 0.032073974609375, "ert_budget": 0.5} |
| {"step": 90, "epoch": 0, "total_loss": 2.883305549621582, "alignment_loss": 0.395554780960083, "kl_loss": 2.84375, "ert_budget": 0.25} |
| {"step": 91, "epoch": 0, "total_loss": 0.04580866917967796, "alignment_loss": 0.36134594678878784, "kl_loss": 0.009674072265625, "ert_budget": 0.75} |
| {"step": 92, "epoch": 0, "total_loss": 0.024629507213830948, "alignment_loss": 0.06990346312522888, "kl_loss": 0.01763916015625, "ert_budget": 0.5} |
| {"step": 93, "epoch": 0, "total_loss": 0.05926142632961273, "alignment_loss": 0.12539014220237732, "kl_loss": 0.046722412109375, "ert_budget": 0.75} |
| {"step": 94, "epoch": 0, "total_loss": 0.08774193376302719, "alignment_loss": 0.4065330922603607, "kl_loss": 0.047088623046875, "ert_budget": 0.5} |
| {"step": 95, "epoch": 0, "total_loss": 0.13421964645385742, "alignment_loss": 0.8752775192260742, "kl_loss": 0.04669189453125, "ert_budget": 0.75} |
| {"step": 96, "epoch": 0, "total_loss": 0.07609280943870544, "alignment_loss": 0.5573757886886597, "kl_loss": 0.020355224609375, "ert_budget": 0.5} |
| {"step": 97, "epoch": 0, "total_loss": 0.10295647382736206, "alignment_loss": 0.2300041764974594, "kl_loss": 0.0799560546875, "ert_budget": 0.5} |
| {"step": 98, "epoch": 0, "total_loss": 2.475581407546997, "alignment_loss": 0.2440946102142334, "kl_loss": 2.451171875, "ert_budget": 0.25} |
| {"step": 99, "epoch": 0, "total_loss": 0.01738891936838627, "alignment_loss": 0.08958438038825989, "kl_loss": 0.00843048095703125, "ert_budget": 0.5} |
| {"step": 100, "epoch": 0, "total_loss": 0.27233266830444336, "alignment_loss": 0.7335806488990784, "kl_loss": 0.198974609375, "ert_budget": 0.75} |
| {"step": 101, "epoch": 0, "total_loss": 0.15076567232608795, "alignment_loss": 1.2322355508804321, "kl_loss": 0.0275421142578125, "ert_budget": 0.5} |
| {"step": 102, "epoch": 0, "total_loss": 0.07945167273283005, "alignment_loss": 0.6766425967216492, "kl_loss": 0.01178741455078125, "ert_budget": 0.5} |
| {"step": 103, "epoch": 0, "total_loss": 0.05655734986066818, "alignment_loss": 0.267874538898468, "kl_loss": 0.0297698974609375, "ert_budget": 0.75} |
| {"step": 104, "epoch": 0, "total_loss": 5.088040351867676, "alignment_loss": 0.060090072453022, "kl_loss": 5.08203125, "ert_budget": 0.25} |
| {"step": 105, "epoch": 0, "total_loss": 0.056513115763664246, "alignment_loss": 0.46312615275382996, "kl_loss": 0.01020050048828125, "ert_budget": 0.75} |
| {"step": 106, "epoch": 0, "total_loss": 0.8719894289970398, "alignment_loss": 0.550949215888977, "kl_loss": 0.81689453125, "ert_budget": 0.5} |
| {"step": 107, "epoch": 0, "total_loss": 0.19406349956989288, "alignment_loss": 0.9781105518341064, "kl_loss": 0.09625244140625, "ert_budget": 0.5} |
| {"step": 108, "epoch": 0, "total_loss": 1.6667838096618652, "alignment_loss": 0.9256504774093628, "kl_loss": 1.57421875, "ert_budget": 0.25} |
| {"step": 109, "epoch": 0, "total_loss": 0.49055740237236023, "alignment_loss": 0.6917067766189575, "kl_loss": 0.42138671875, "ert_budget": 0.25} |
| {"step": 110, "epoch": 0, "total_loss": 0.18326665461063385, "alignment_loss": 0.2713872790336609, "kl_loss": 0.1561279296875, "ert_budget": 0.75} |
| {"step": 111, "epoch": 0, "total_loss": 0.039371293038129807, "alignment_loss": 0.11600297689437866, "kl_loss": 0.02777099609375, "ert_budget": 0.75} |
| {"step": 112, "epoch": 0, "total_loss": 1.464139461517334, "alignment_loss": 0.11014411598443985, "kl_loss": 1.453125, "ert_budget": 0.25} |
| {"step": 113, "epoch": 0, "total_loss": 0.024687914177775383, "alignment_loss": 0.17569687962532043, "kl_loss": 0.00711822509765625, "ert_budget": 0.75} |
| {"step": 114, "epoch": 0, "total_loss": 0.009070439264178276, "alignment_loss": 0.01395268365740776, "kl_loss": 0.0076751708984375, "ert_budget": 0.75} |
| {"step": 115, "epoch": 0, "total_loss": 0.018684379756450653, "alignment_loss": 0.049514688551425934, "kl_loss": 0.01373291015625, "ert_budget": 0.75} |
| {"step": 116, "epoch": 0, "total_loss": 0.021968401968479156, "alignment_loss": 0.018878359347581863, "kl_loss": 0.02008056640625, "ert_budget": 0.5} |
| {"step": 117, "epoch": 0, "total_loss": 0.042297884821891785, "alignment_loss": 0.09430452436208725, "kl_loss": 0.032867431640625, "ert_budget": 0.75} |
| {"step": 118, "epoch": 0, "total_loss": 1.5464301109313965, "alignment_loss": 0.11273840069770813, "kl_loss": 1.53515625, "ert_budget": 0.25} |
| {"step": 119, "epoch": 0, "total_loss": 0.04959718510508537, "alignment_loss": 0.009216492995619774, "kl_loss": 0.048675537109375, "ert_budget": 0.5} |
| {"step": 120, "epoch": 0, "total_loss": 0.11182201653718948, "alignment_loss": 0.06048091500997543, "kl_loss": 0.10577392578125, "ert_budget": 0.5} |
| {"step": 121, "epoch": 0, "total_loss": 0.06526785343885422, "alignment_loss": 0.020964650437235832, "kl_loss": 0.06317138671875, "ert_budget": 0.5} |
| {"step": 122, "epoch": 0, "total_loss": 0.03498348966240883, "alignment_loss": 0.041302166879177094, "kl_loss": 0.030853271484375, "ert_budget": 0.5} |
| {"step": 123, "epoch": 0, "total_loss": 0.9095113277435303, "alignment_loss": 0.21816040575504303, "kl_loss": 0.8876953125, "ert_budget": 0.25} |
| {"step": 124, "epoch": 0, "total_loss": 0.045556724071502686, "alignment_loss": 0.12902913987636566, "kl_loss": 0.03265380859375, "ert_budget": 0.75} |
| {"step": 125, "epoch": 0, "total_loss": 0.130308598279953, "alignment_loss": 0.06041016802191734, "kl_loss": 0.124267578125, "ert_budget": 0.5} |
| {"step": 126, "epoch": 0, "total_loss": 0.057378802448511124, "alignment_loss": 0.24938616156578064, "kl_loss": 0.032440185546875, "ert_budget": 0.5} |
| {"step": 127, "epoch": 0, "total_loss": 1.8353954553604126, "alignment_loss": 0.014110494405031204, "kl_loss": 1.833984375, "ert_budget": 0.25} |
|
|
