{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.9987368421052631, "eval_steps": 10000, "global_step": 593, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.003368421052631579, "grad_norm": 1021.0826362249041, "learning_rate": 8e-09, "logits/chosen": -6.098826885223389, "logits/rejected": -8.285457611083984, "logps/chosen": -1105.89208984375, "logps/rejected": -1840.68798828125, "loss": 0.6675, "nll_loss": 4.724418640136719, "rewards/accuracies": 0.5, "rewards/chosen": 0.03404197469353676, "rewards/margins": 0.10088615119457245, "rewards/rejected": -0.06684418022632599, "step": 2 }, { "epoch": 0.006736842105263158, "grad_norm": 775.2893008419502, "learning_rate": 1.6e-08, "logits/chosen": -6.404820919036865, "logits/rejected": -8.010771751403809, "logps/chosen": -1317.140625, "logps/rejected": -1826.8807373046875, "loss": 0.7069, "nll_loss": 4.841824054718018, "rewards/accuracies": 0.625, "rewards/chosen": 0.024968720972537994, "rewards/margins": 0.0651191771030426, "rewards/rejected": -0.04015045240521431, "step": 4 }, { "epoch": 0.010105263157894737, "grad_norm": 1152.6434834761353, "learning_rate": 2.3999999999999997e-08, "logits/chosen": -7.275630950927734, "logits/rejected": -7.599463939666748, "logps/chosen": -1705.133544921875, "logps/rejected": -1831.8782958984375, "loss": 0.7266, "nll_loss": 4.776337623596191, "rewards/accuracies": 0.5, "rewards/chosen": 0.00952148623764515, "rewards/margins": -0.00301208533346653, "rewards/rejected": 0.01253356970846653, "step": 6 }, { "epoch": 0.013473684210526317, "grad_norm": 1340.686158278663, "learning_rate": 3.2e-08, "logits/chosen": -7.0847063064575195, "logits/rejected": -7.663762092590332, "logps/chosen": -1612.9189453125, "logps/rejected": -1863.4921875, "loss": 0.6938, "nll_loss": 4.129978656768799, "rewards/accuracies": 0.375, "rewards/chosen": 0.05372562259435654, "rewards/margins": -0.047937583178281784, "rewards/rejected": 0.10166320949792862, "step": 8 }, { "epoch": 0.016842105263157894, "grad_norm": 830.1962870309648, "learning_rate": 4e-08, "logits/chosen": -7.328038692474365, "logits/rejected": -7.968114852905273, "logps/chosen": -1624.5255126953125, "logps/rejected": -1856.871826171875, "loss": 0.7296, "nll_loss": 4.658226490020752, "rewards/accuracies": 0.375, "rewards/chosen": 0.03902282938361168, "rewards/margins": -0.01064453274011612, "rewards/rejected": 0.0496673583984375, "step": 10 }, { "epoch": 0.020210526315789474, "grad_norm": 663.540527209906, "learning_rate": 4.799999999999999e-08, "logits/chosen": -7.396481513977051, "logits/rejected": -7.322075843811035, "logps/chosen": -1599.8494873046875, "logps/rejected": -1831.381103515625, "loss": 0.6782, "nll_loss": 4.17366361618042, "rewards/accuracies": 0.5, "rewards/chosen": -0.10295410454273224, "rewards/margins": 0.06537018716335297, "rewards/rejected": -0.168324276804924, "step": 12 }, { "epoch": 0.023578947368421053, "grad_norm": 1099.037021669425, "learning_rate": 5.6000000000000005e-08, "logits/chosen": -7.1363630294799805, "logits/rejected": -7.741977691650391, "logps/chosen": -1470.75537109375, "logps/rejected": -1858.2095947265625, "loss": 0.6512, "nll_loss": 4.299507141113281, "rewards/accuracies": 0.625, "rewards/chosen": -0.2573639154434204, "rewards/margins": 0.10828553140163422, "rewards/rejected": -0.36564940214157104, "step": 14 }, { "epoch": 0.026947368421052633, "grad_norm": 566.7332650815549, "learning_rate": 6.4e-08, "logits/chosen": -6.818307876586914, "logits/rejected": -7.9370036125183105, "logps/chosen": -1442.171875, "logps/rejected": -1850.56298828125, "loss": 0.6185, "nll_loss": 4.646788597106934, "rewards/accuracies": 0.875, "rewards/chosen": -0.3849685788154602, "rewards/margins": 0.29212191700935364, "rewards/rejected": -0.6770904660224915, "step": 16 }, { "epoch": 0.03031578947368421, "grad_norm": 743.3397608092816, "learning_rate": 7.2e-08, "logits/chosen": -7.3267436027526855, "logits/rejected": -7.727769374847412, "logps/chosen": -1639.974853515625, "logps/rejected": -1839.6697998046875, "loss": 0.5848, "nll_loss": 4.681390285491943, "rewards/accuracies": 0.5, "rewards/chosen": -0.806835949420929, "rewards/margins": 0.1156768724322319, "rewards/rejected": -0.9225128889083862, "step": 18 }, { "epoch": 0.03368421052631579, "grad_norm": 555.7381352759589, "learning_rate": 8e-08, "logits/chosen": -6.984310150146484, "logits/rejected": -7.973010540008545, "logps/chosen": -1658.0970458984375, "logps/rejected": -1865.5997314453125, "loss": 0.5718, "nll_loss": 4.629044532775879, "rewards/accuracies": 0.625, "rewards/chosen": -1.5665146112442017, "rewards/margins": 0.23055119812488556, "rewards/rejected": -1.7970658540725708, "step": 20 }, { "epoch": 0.03705263157894737, "grad_norm": 537.2764069824767, "learning_rate": 8.8e-08, "logits/chosen": -7.248239040374756, "logits/rejected": -7.833005428314209, "logps/chosen": -1698.72998046875, "logps/rejected": -1851.3116455078125, "loss": 0.5141, "nll_loss": 4.778887748718262, "rewards/accuracies": 0.75, "rewards/chosen": -2.040907382965088, "rewards/margins": 0.4460693299770355, "rewards/rejected": -2.4869766235351562, "step": 22 }, { "epoch": 0.04042105263157895, "grad_norm": 427.6716943228766, "learning_rate": 9.599999999999999e-08, "logits/chosen": -6.897974967956543, "logits/rejected": -8.095926284790039, "logps/chosen": -1618.291748046875, "logps/rejected": -1899.2071533203125, "loss": 0.5628, "nll_loss": 4.640469074249268, "rewards/accuracies": 0.625, "rewards/chosen": -2.707034111022949, "rewards/margins": 0.4983123540878296, "rewards/rejected": -3.2053468227386475, "step": 24 }, { "epoch": 0.043789473684210524, "grad_norm": 502.48124965017354, "learning_rate": 1.04e-07, "logits/chosen": -6.930876731872559, "logits/rejected": -7.992618083953857, "logps/chosen": -1484.425048828125, "logps/rejected": -1923.680908203125, "loss": 0.5462, "nll_loss": 4.688654899597168, "rewards/accuracies": 0.625, "rewards/chosen": -4.368743419647217, "rewards/margins": 1.1556202173233032, "rewards/rejected": -5.524363994598389, "step": 26 }, { "epoch": 0.04715789473684211, "grad_norm": 383.7692991889789, "learning_rate": 1.1200000000000001e-07, "logits/chosen": -7.199338912963867, "logits/rejected": -7.51033353805542, "logps/chosen": -1769.854248046875, "logps/rejected": -1886.64501953125, "loss": 0.4697, "nll_loss": 4.6844096183776855, "rewards/accuracies": 0.75, "rewards/chosen": -6.4538960456848145, "rewards/margins": 0.6539719104766846, "rewards/rejected": -7.107868194580078, "step": 28 }, { "epoch": 0.05052631578947368, "grad_norm": 481.75715160752935, "learning_rate": 1.2e-07, "logits/chosen": -6.897581577301025, "logits/rejected": -7.895030498504639, "logps/chosen": -1615.928466796875, "logps/rejected": -1956.94287109375, "loss": 0.4974, "nll_loss": 4.612817764282227, "rewards/accuracies": 0.625, "rewards/chosen": -6.480710029602051, "rewards/margins": 1.3521087169647217, "rewards/rejected": -7.832818984985352, "step": 30 }, { "epoch": 0.053894736842105266, "grad_norm": 433.1144839993232, "learning_rate": 1.28e-07, "logits/chosen": -7.140986442565918, "logits/rejected": -7.891180515289307, "logps/chosen": -1722.1290283203125, "logps/rejected": -1961.19140625, "loss": 0.3723, "nll_loss": 4.612584590911865, "rewards/accuracies": 0.75, "rewards/chosen": -7.3035502433776855, "rewards/margins": 1.1011128425598145, "rewards/rejected": -8.4046630859375, "step": 32 }, { "epoch": 0.05726315789473684, "grad_norm": 460.5421567294356, "learning_rate": 1.36e-07, "logits/chosen": -5.8924641609191895, "logits/rejected": -7.968005657196045, "logps/chosen": -1063.8079833984375, "logps/rejected": -1961.5302734375, "loss": 0.3717, "nll_loss": 4.887433052062988, "rewards/accuracies": 0.875, "rewards/chosen": -4.523468017578125, "rewards/margins": 4.804460525512695, "rewards/rejected": -9.327927589416504, "step": 34 }, { "epoch": 0.06063157894736842, "grad_norm": 382.3217926367787, "learning_rate": 1.44e-07, "logits/chosen": -7.277335166931152, "logits/rejected": -8.038162231445312, "logps/chosen": -1612.2705078125, "logps/rejected": -1970.5556640625, "loss": 0.4893, "nll_loss": 4.758667469024658, "rewards/accuracies": 0.75, "rewards/chosen": -8.101587295532227, "rewards/margins": 1.8269003629684448, "rewards/rejected": -9.928487777709961, "step": 36 }, { "epoch": 0.064, "grad_norm": 327.81609268765317, "learning_rate": 1.5199999999999998e-07, "logits/chosen": -6.232708930969238, "logits/rejected": -7.9314703941345215, "logps/chosen": -1311.58984375, "logps/rejected": -1898.37353515625, "loss": 0.4384, "nll_loss": 4.995038986206055, "rewards/accuracies": 0.875, "rewards/chosen": -7.171569347381592, "rewards/margins": 4.5418925285339355, "rewards/rejected": -11.713461875915527, "step": 38 }, { "epoch": 0.06736842105263158, "grad_norm": 411.60155946011673, "learning_rate": 1.6e-07, "logits/chosen": -7.439314365386963, "logits/rejected": -7.884708404541016, "logps/chosen": -1801.27978515625, "logps/rejected": -2006.94921875, "loss": 0.3972, "nll_loss": 4.832527160644531, "rewards/accuracies": 0.5, "rewards/chosen": -12.396756172180176, "rewards/margins": 1.0853164196014404, "rewards/rejected": -13.482072830200195, "step": 40 }, { "epoch": 0.07073684210526315, "grad_norm": 353.27396194384323, "learning_rate": 1.68e-07, "logits/chosen": -6.52924108505249, "logits/rejected": -7.981770038604736, "logps/chosen": -1366.396484375, "logps/rejected": -2025.9642333984375, "loss": 0.5679, "nll_loss": 4.827330112457275, "rewards/accuracies": 0.75, "rewards/chosen": -9.584479331970215, "rewards/margins": 5.323737621307373, "rewards/rejected": -14.90821647644043, "step": 42 }, { "epoch": 0.07410526315789474, "grad_norm": 259.81954374623166, "learning_rate": 1.76e-07, "logits/chosen": -7.801054000854492, "logits/rejected": -7.639019966125488, "logps/chosen": -1827.47314453125, "logps/rejected": -1996.3192138671875, "loss": 0.3678, "nll_loss": 4.5667243003845215, "rewards/accuracies": 1.0, "rewards/chosen": -14.310271263122559, "rewards/margins": 2.2248902320861816, "rewards/rejected": -16.5351619720459, "step": 44 }, { "epoch": 0.07747368421052632, "grad_norm": 406.5448161027269, "learning_rate": 1.84e-07, "logits/chosen": -8.297408103942871, "logits/rejected": -7.499330043792725, "logps/chosen": -1998.0411376953125, "logps/rejected": -1991.69921875, "loss": 0.3749, "nll_loss": 4.313129425048828, "rewards/accuracies": 0.75, "rewards/chosen": -16.31363296508789, "rewards/margins": 0.6383074522018433, "rewards/rejected": -16.951940536499023, "step": 46 }, { "epoch": 0.0808421052631579, "grad_norm": 462.82095282625227, "learning_rate": 1.9199999999999997e-07, "logits/chosen": -7.885348320007324, "logits/rejected": -8.057937622070312, "logps/chosen": -1869.8287353515625, "logps/rejected": -1985.5797119140625, "loss": 0.4191, "nll_loss": 4.887444019317627, "rewards/accuracies": 0.625, "rewards/chosen": -15.957789421081543, "rewards/margins": 1.4249342679977417, "rewards/rejected": -17.382722854614258, "step": 48 }, { "epoch": 0.08421052631578947, "grad_norm": 612.0204500259774, "learning_rate": 2e-07, "logits/chosen": -5.61109733581543, "logits/rejected": -7.90653657913208, "logps/chosen": -1070.6937255859375, "logps/rejected": -2036.2265625, "loss": 0.3329, "nll_loss": 5.1021575927734375, "rewards/accuracies": 1.0, "rewards/chosen": -8.67391586303711, "rewards/margins": 8.88609504699707, "rewards/rejected": -17.560009002685547, "step": 50 }, { "epoch": 0.08757894736842105, "grad_norm": 450.4075931864382, "learning_rate": 1.9999330539070613e-07, "logits/chosen": -7.206405162811279, "logits/rejected": -8.117046356201172, "logps/chosen": -1794.2418212890625, "logps/rejected": -2010.4686279296875, "loss": 0.3961, "nll_loss": 5.458052158355713, "rewards/accuracies": 0.75, "rewards/chosen": -15.731528282165527, "rewards/margins": 1.6168347597122192, "rewards/rejected": -17.348363876342773, "step": 52 }, { "epoch": 0.09094736842105264, "grad_norm": 521.1752044658302, "learning_rate": 1.9997322245918037e-07, "logits/chosen": -7.272748947143555, "logits/rejected": -8.118557929992676, "logps/chosen": -1745.1031494140625, "logps/rejected": -2038.525390625, "loss": 0.4714, "nll_loss": 5.170174598693848, "rewards/accuracies": 0.875, "rewards/chosen": -15.204635620117188, "rewards/margins": 2.791910409927368, "rewards/rejected": -17.996545791625977, "step": 54 }, { "epoch": 0.09431578947368421, "grad_norm": 217.38647360461505, "learning_rate": 1.9993975389437037e-07, "logits/chosen": -6.186915397644043, "logits/rejected": -8.047518730163574, "logps/chosen": -1388.68603515625, "logps/rejected": -2048.745361328125, "loss": 0.2578, "nll_loss": 5.35891056060791, "rewards/accuracies": 1.0, "rewards/chosen": -11.742466926574707, "rewards/margins": 7.281023025512695, "rewards/rejected": -19.023488998413086, "step": 56 }, { "epoch": 0.09768421052631579, "grad_norm": 405.6486896009896, "learning_rate": 1.9989290417745539e-07, "logits/chosen": -6.964948654174805, "logits/rejected": -8.299407005310059, "logps/chosen": -1703.4879150390625, "logps/rejected": -2057.886962890625, "loss": 0.4002, "nll_loss": 5.260710716247559, "rewards/accuracies": 1.0, "rewards/chosen": -14.706157684326172, "rewards/margins": 3.9658584594726562, "rewards/rejected": -18.672016143798828, "step": 58 }, { "epoch": 0.10105263157894737, "grad_norm": 475.2633809268288, "learning_rate": 1.9983267958124644e-07, "logits/chosen": -6.648141384124756, "logits/rejected": -7.772933006286621, "logps/chosen": -1497.2598876953125, "logps/rejected": -2023.4912109375, "loss": 0.474, "nll_loss": 5.289524555206299, "rewards/accuracies": 0.625, "rewards/chosen": -13.703365325927734, "rewards/margins": 5.52022647857666, "rewards/rejected": -19.22359275817871, "step": 60 }, { "epoch": 0.10442105263157894, "grad_norm": 295.526199134696, "learning_rate": 1.9975908816934638e-07, "logits/chosen": -7.280607223510742, "logits/rejected": -7.836946964263916, "logps/chosen": -1767.7076416015625, "logps/rejected": -2027.5374755859375, "loss": 0.3311, "nll_loss": 4.935500144958496, "rewards/accuracies": 0.625, "rewards/chosen": -16.52497100830078, "rewards/margins": 2.8025190830230713, "rewards/rejected": -19.327489852905273, "step": 62 }, { "epoch": 0.10778947368421053, "grad_norm": 451.01621367751295, "learning_rate": 1.9967213979507017e-07, "logits/chosen": -7.747291564941406, "logits/rejected": -7.578754901885986, "logps/chosen": -1903.654052734375, "logps/rejected": -1995.362060546875, "loss": 0.2961, "nll_loss": 4.874569416046143, "rewards/accuracies": 0.625, "rewards/chosen": -18.415132522583008, "rewards/margins": 1.7832107543945312, "rewards/rejected": -20.19834327697754, "step": 64 }, { "epoch": 0.11115789473684211, "grad_norm": 391.98836079659304, "learning_rate": 1.995718461001257e-07, "logits/chosen": -6.887873649597168, "logits/rejected": -7.892035484313965, "logps/chosen": -1613.168701171875, "logps/rejected": -2109.185546875, "loss": 0.2514, "nll_loss": 4.979562282562256, "rewards/accuracies": 0.75, "rewards/chosen": -15.48417854309082, "rewards/margins": 5.469597816467285, "rewards/rejected": -20.953777313232422, "step": 66 }, { "epoch": 0.11452631578947368, "grad_norm": 357.9128298203349, "learning_rate": 1.9945822051305505e-07, "logits/chosen": -6.562501430511475, "logits/rejected": -8.093667984008789, "logps/chosen": -1624.415771484375, "logps/rejected": -2102.240966796875, "loss": 0.2549, "nll_loss": 5.153851509094238, "rewards/accuracies": 1.0, "rewards/chosen": -15.476547241210938, "rewards/margins": 6.269686222076416, "rewards/rejected": -21.746234893798828, "step": 68 }, { "epoch": 0.11789473684210526, "grad_norm": 324.2814050905479, "learning_rate": 1.9933127824743643e-07, "logits/chosen": -7.621331214904785, "logits/rejected": -7.522034645080566, "logps/chosen": -1808.6651611328125, "logps/rejected": -2016.3822021484375, "loss": 0.1968, "nll_loss": 5.224420070648193, "rewards/accuracies": 0.875, "rewards/chosen": -17.95054817199707, "rewards/margins": 3.378469944000244, "rewards/rejected": -21.329015731811523, "step": 70 }, { "epoch": 0.12126315789473684, "grad_norm": 507.77270997957345, "learning_rate": 1.9919103629984725e-07, "logits/chosen": -7.971831321716309, "logits/rejected": -7.918793201446533, "logps/chosen": -1930.361083984375, "logps/rejected": -2080.251953125, "loss": 0.3995, "nll_loss": 4.725856781005859, "rewards/accuracies": 0.75, "rewards/chosen": -19.6121826171875, "rewards/margins": 1.6514708995819092, "rewards/rejected": -21.263654708862305, "step": 72 }, { "epoch": 0.12463157894736843, "grad_norm": 405.2199441352807, "learning_rate": 1.9903751344758845e-07, "logits/chosen": -7.292391300201416, "logits/rejected": -8.180887222290039, "logps/chosen": -1775.805908203125, "logps/rejected": -2013.7442626953125, "loss": 0.4428, "nll_loss": 5.191011905670166, "rewards/accuracies": 0.625, "rewards/chosen": -17.799779891967773, "rewards/margins": 3.329552173614502, "rewards/rejected": -21.129331588745117, "step": 74 }, { "epoch": 0.128, "grad_norm": 240.9685453409843, "learning_rate": 1.9887073024617028e-07, "logits/chosen": -6.470153331756592, "logits/rejected": -7.746395587921143, "logps/chosen": -1489.3653564453125, "logps/rejected": -2056.020751953125, "loss": 0.2493, "nll_loss": 5.174279689788818, "rewards/accuracies": 0.875, "rewards/chosen": -15.556097030639648, "rewards/margins": 6.754159927368164, "rewards/rejected": -22.310256958007812, "step": 76 }, { "epoch": 0.13136842105263158, "grad_norm": 493.99746962914696, "learning_rate": 1.9869070902656017e-07, "logits/chosen": -5.889953136444092, "logits/rejected": -8.0224027633667, "logps/chosen": -1382.2691650390625, "logps/rejected": -2080.0107421875, "loss": 0.485, "nll_loss": 5.714942932128906, "rewards/accuracies": 0.875, "rewards/chosen": -14.33112907409668, "rewards/margins": 7.287192344665527, "rewards/rejected": -21.61832046508789, "step": 78 }, { "epoch": 0.13473684210526315, "grad_norm": 137.24435752290512, "learning_rate": 1.984974738921927e-07, "logits/chosen": -6.03630256652832, "logits/rejected": -8.135429382324219, "logps/chosen": -1310.54833984375, "logps/rejected": -2067.27490234375, "loss": 0.2926, "nll_loss": 5.408031463623047, "rewards/accuracies": 1.0, "rewards/chosen": -13.000334739685059, "rewards/margins": 9.575756072998047, "rewards/rejected": -22.57608985900879, "step": 80 }, { "epoch": 0.13810526315789473, "grad_norm": 332.6447170969351, "learning_rate": 1.982910507157424e-07, "logits/chosen": -7.524165630340576, "logits/rejected": -7.871054649353027, "logps/chosen": -1923.2384033203125, "logps/rejected": -2104.744384765625, "loss": 0.4062, "nll_loss": 4.740773677825928, "rewards/accuracies": 0.875, "rewards/chosen": -20.172786712646484, "rewards/margins": 2.7083492279052734, "rewards/rejected": -22.881135940551758, "step": 82 }, { "epoch": 0.1414736842105263, "grad_norm": 434.7785715165825, "learning_rate": 1.9807146713565955e-07, "logits/chosen": -7.366245746612549, "logits/rejected": -7.99586296081543, "logps/chosen": -1834.388916015625, "logps/rejected": -2039.752685546875, "loss": 0.4183, "nll_loss": 5.197725296020508, "rewards/accuracies": 0.625, "rewards/chosen": -19.303865432739258, "rewards/margins": 3.0796072483062744, "rewards/rejected": -22.38347053527832, "step": 84 }, { "epoch": 0.14484210526315788, "grad_norm": 394.456469398082, "learning_rate": 1.9783875255246973e-07, "logits/chosen": -7.369488716125488, "logits/rejected": -7.899564743041992, "logps/chosen": -2014.7353515625, "logps/rejected": -2079.622802734375, "loss": 0.435, "nll_loss": 5.506885051727295, "rewards/accuracies": 0.5, "rewards/chosen": -22.415340423583984, "rewards/margins": 1.3391568660736084, "rewards/rejected": -23.75449562072754, "step": 86 }, { "epoch": 0.1482105263157895, "grad_norm": 474.52038286042045, "learning_rate": 1.9759293812483712e-07, "logits/chosen": -6.315229892730713, "logits/rejected": -7.918384075164795, "logps/chosen": -1505.7802734375, "logps/rejected": -2153.10986328125, "loss": 0.518, "nll_loss": 5.55679988861084, "rewards/accuracies": 1.0, "rewards/chosen": -16.11681365966797, "rewards/margins": 8.099884033203125, "rewards/rejected": -24.216697692871094, "step": 88 }, { "epoch": 0.15157894736842106, "grad_norm": 463.6020653534667, "learning_rate": 1.973340567653928e-07, "logits/chosen": -7.8903656005859375, "logits/rejected": -7.871662616729736, "logps/chosen": -2007.9490966796875, "logps/rejected": -2079.7353515625, "loss": 0.3285, "nll_loss": 4.673924922943115, "rewards/accuracies": 0.875, "rewards/chosen": -21.700756072998047, "rewards/margins": 2.0794034004211426, "rewards/rejected": -23.78015899658203, "step": 90 }, { "epoch": 0.15494736842105264, "grad_norm": 469.1854810466752, "learning_rate": 1.9706214313632782e-07, "logits/chosen": -7.816376686096191, "logits/rejected": -7.4226765632629395, "logps/chosen": -1935.4766845703125, "logps/rejected": -2112.067138671875, "loss": 0.4765, "nll_loss": 4.788717746734619, "rewards/accuracies": 0.875, "rewards/chosen": -21.18560791015625, "rewards/margins": 3.7070724964141846, "rewards/rejected": -24.892681121826172, "step": 92 }, { "epoch": 0.15831578947368422, "grad_norm": 203.68587948112653, "learning_rate": 1.9677723364475236e-07, "logits/chosen": -7.06157112121582, "logits/rejected": -8.203742027282715, "logps/chosen": -1748.9493408203125, "logps/rejected": -2103.68212890625, "loss": 0.2824, "nll_loss": 5.13286828994751, "rewards/accuracies": 0.875, "rewards/chosen": -18.6251220703125, "rewards/margins": 5.7565789222717285, "rewards/rejected": -24.381698608398438, "step": 94 }, { "epoch": 0.1616842105263158, "grad_norm": 454.0753654405841, "learning_rate": 1.9647936643782106e-07, "logits/chosen": -7.063906669616699, "logits/rejected": -7.930819511413574, "logps/chosen": -1851.276611328125, "logps/rejected": -2139.397216796875, "loss": 0.4842, "nll_loss": 4.915232181549072, "rewards/accuracies": 0.75, "rewards/chosen": -20.60812759399414, "rewards/margins": 3.6515331268310547, "rewards/rejected": -24.259662628173828, "step": 96 }, { "epoch": 0.16505263157894737, "grad_norm": 290.76223910181324, "learning_rate": 1.961685813976253e-07, "logits/chosen": -7.524170875549316, "logits/rejected": -8.006389617919922, "logps/chosen": -1945.13623046875, "logps/rejected": -2088.755126953125, "loss": 0.3862, "nll_loss": 4.894111633300781, "rewards/accuracies": 0.75, "rewards/chosen": -22.57034683227539, "rewards/margins": 2.1726748943328857, "rewards/rejected": -24.74302101135254, "step": 98 }, { "epoch": 0.16842105263157894, "grad_norm": 245.46799212168173, "learning_rate": 1.9584492013585354e-07, "logits/chosen": -7.0596513748168945, "logits/rejected": -7.93139123916626, "logps/chosen": -1557.81787109375, "logps/rejected": -2036.077880859375, "loss": 0.192, "nll_loss": 5.297558784484863, "rewards/accuracies": 0.875, "rewards/chosen": -16.414541244506836, "rewards/margins": 8.017242431640625, "rewards/rejected": -24.431781768798828, "step": 100 }, { "epoch": 0.17178947368421052, "grad_norm": 404.1873598654466, "learning_rate": 1.955084259882195e-07, "logits/chosen": -5.742030143737793, "logits/rejected": -7.751094818115234, "logps/chosen": -1235.095947265625, "logps/rejected": -2130.07080078125, "loss": 0.3401, "nll_loss": 5.310590744018555, "rewards/accuracies": 1.0, "rewards/chosen": -13.479273796081543, "rewards/margins": 12.036299705505371, "rewards/rejected": -25.515575408935547, "step": 102 }, { "epoch": 0.1751578947368421, "grad_norm": 116.3442830030767, "learning_rate": 1.9515914400866017e-07, "logits/chosen": -7.13689661026001, "logits/rejected": -7.9674553871154785, "logps/chosen": -1809.794677734375, "logps/rejected": -2110.8828125, "loss": 0.2132, "nll_loss": 5.315598487854004, "rewards/accuracies": 0.875, "rewards/chosen": -21.165102005004883, "rewards/margins": 4.648441791534424, "rewards/rejected": -25.81354522705078, "step": 104 }, { "epoch": 0.17852631578947367, "grad_norm": 290.204628775111, "learning_rate": 1.9479712096330334e-07, "logits/chosen": -7.378734111785889, "logits/rejected": -7.894670486450195, "logps/chosen": -1803.5947265625, "logps/rejected": -2066.1181640625, "loss": 0.3482, "nll_loss": 5.543490409851074, "rewards/accuracies": 0.75, "rewards/chosen": -20.527360916137695, "rewards/margins": 5.067646026611328, "rewards/rejected": -25.595006942749023, "step": 106 }, { "epoch": 0.18189473684210528, "grad_norm": 361.78273867060045, "learning_rate": 1.944224053242058e-07, "logits/chosen": -7.142321586608887, "logits/rejected": -7.959826946258545, "logps/chosen": -1579.7891845703125, "logps/rejected": -2068.02197265625, "loss": 0.2804, "nll_loss": 4.833759307861328, "rewards/accuracies": 0.875, "rewards/chosen": -18.271129608154297, "rewards/margins": 7.227394104003906, "rewards/rejected": -25.498523712158203, "step": 108 }, { "epoch": 0.18526315789473685, "grad_norm": 235.0435714261294, "learning_rate": 1.9403504726286367e-07, "logits/chosen": -5.593753337860107, "logits/rejected": -8.2781400680542, "logps/chosen": -1415.476806640625, "logps/rejected": -2148.1337890625, "loss": 0.259, "nll_loss": 4.7400031089782715, "rewards/accuracies": 0.75, "rewards/chosen": -16.4371337890625, "rewards/margins": 9.744308471679688, "rewards/rejected": -26.18144416809082, "step": 110 }, { "epoch": 0.18863157894736843, "grad_norm": 347.2912929912007, "learning_rate": 1.9363509864349436e-07, "logits/chosen": -6.396492004394531, "logits/rejected": -8.129210472106934, "logps/chosen": -1428.15185546875, "logps/rejected": -2139.29931640625, "loss": 0.1951, "nll_loss": 5.208731174468994, "rewards/accuracies": 1.0, "rewards/chosen": -17.481124877929688, "rewards/margins": 9.886039733886719, "rewards/rejected": -27.367164611816406, "step": 112 }, { "epoch": 0.192, "grad_norm": 443.2420860827636, "learning_rate": 1.9322261301609284e-07, "logits/chosen": -7.289253234863281, "logits/rejected": -8.456584930419922, "logps/chosen": -1952.02685546875, "logps/rejected": -2108.48876953125, "loss": 0.4061, "nll_loss": 5.445844650268555, "rewards/accuracies": 1.0, "rewards/chosen": -24.95429229736328, "rewards/margins": 2.7193362712860107, "rewards/rejected": -27.673629760742188, "step": 114 }, { "epoch": 0.19536842105263158, "grad_norm": 322.7074429727334, "learning_rate": 1.927976456092614e-07, "logits/chosen": -7.254305362701416, "logits/rejected": -8.170782089233398, "logps/chosen": -1910.8201904296875, "logps/rejected": -2137.612548828125, "loss": 0.3789, "nll_loss": 5.32393217086792, "rewards/accuracies": 0.875, "rewards/chosen": -23.50674819946289, "rewards/margins": 3.4448273181915283, "rewards/rejected": -26.951576232910156, "step": 116 }, { "epoch": 0.19873684210526316, "grad_norm": 230.28383471544416, "learning_rate": 1.9236025332281506e-07, "logits/chosen": -6.197360515594482, "logits/rejected": -8.328180313110352, "logps/chosen": -1564.211669921875, "logps/rejected": -2164.58056640625, "loss": 0.1884, "nll_loss": 5.403533458709717, "rewards/accuracies": 1.0, "rewards/chosen": -18.896282196044922, "rewards/margins": 8.365097999572754, "rewards/rejected": -27.261381149291992, "step": 118 }, { "epoch": 0.20210526315789473, "grad_norm": 375.89177298791736, "learning_rate": 1.9191049472016313e-07, "logits/chosen": -7.113523006439209, "logits/rejected": -7.964775562286377, "logps/chosen": -1672.96923828125, "logps/rejected": -2136.83642578125, "loss": 0.2439, "nll_loss": 4.739628791809082, "rewards/accuracies": 0.625, "rewards/chosen": -20.55191993713379, "rewards/margins": 5.649931907653809, "rewards/rejected": -26.20184898376465, "step": 120 }, { "epoch": 0.2054736842105263, "grad_norm": 323.0978953815701, "learning_rate": 1.9144843002046805e-07, "logits/chosen": -6.809453964233398, "logits/rejected": -8.00958251953125, "logps/chosen": -1658.2227783203125, "logps/rejected": -2084.864990234375, "loss": 0.5958, "nll_loss": 5.457843780517578, "rewards/accuracies": 0.75, "rewards/chosen": -20.677547454833984, "rewards/margins": 6.124557018280029, "rewards/rejected": -26.802101135253906, "step": 122 }, { "epoch": 0.20884210526315788, "grad_norm": 258.6436979707799, "learning_rate": 1.9097412109058243e-07, "logits/chosen": -7.692294597625732, "logits/rejected": -8.052206993103027, "logps/chosen": -1873.5733642578125, "logps/rejected": -2030.232177734375, "loss": 0.4711, "nll_loss": 5.26876163482666, "rewards/accuracies": 1.0, "rewards/chosen": -22.60447120666504, "rewards/margins": 3.8998751640319824, "rewards/rejected": -26.50434684753418, "step": 124 }, { "epoch": 0.21221052631578946, "grad_norm": 372.3923082583742, "learning_rate": 1.9048763143676575e-07, "logits/chosen": -7.407994747161865, "logits/rejected": -7.782034397125244, "logps/chosen": -1862.9501953125, "logps/rejected": -2138.9091796875, "loss": 0.1584, "nll_loss": 5.26196813583374, "rewards/accuracies": 1.0, "rewards/chosen": -21.54726791381836, "rewards/margins": 5.700539588928223, "rewards/rejected": -27.247806549072266, "step": 126 }, { "epoch": 0.21557894736842106, "grad_norm": 176.97835981766247, "learning_rate": 1.8998902619618113e-07, "logits/chosen": -6.254648208618164, "logits/rejected": -8.410988807678223, "logps/chosen": -1320.6219482421875, "logps/rejected": -2096.9287109375, "loss": 0.2115, "nll_loss": 5.321068286895752, "rewards/accuracies": 1.0, "rewards/chosen": -15.538348197937012, "rewards/margins": 11.237432479858398, "rewards/rejected": -26.775779724121094, "step": 128 }, { "epoch": 0.21894736842105264, "grad_norm": 547.3377613503384, "learning_rate": 1.8947837212817413e-07, "logits/chosen": -6.936631202697754, "logits/rejected": -8.201074600219727, "logps/chosen": -1785.872314453125, "logps/rejected": -2129.34375, "loss": 0.3615, "nll_loss": 5.191522598266602, "rewards/accuracies": 0.875, "rewards/chosen": -21.109575271606445, "rewards/margins": 5.265467643737793, "rewards/rejected": -26.37504005432129, "step": 130 }, { "epoch": 0.22231578947368422, "grad_norm": 353.0941136935024, "learning_rate": 1.8895573760533412e-07, "logits/chosen": -7.306629180908203, "logits/rejected": -7.669928073883057, "logps/chosen": -1730.935546875, "logps/rejected": -2167.73095703125, "loss": 0.293, "nll_loss": 4.846038818359375, "rewards/accuracies": 0.75, "rewards/chosen": -20.911489486694336, "rewards/margins": 6.074740409851074, "rewards/rejected": -26.986228942871094, "step": 132 }, { "epoch": 0.2256842105263158, "grad_norm": 212.13304713669675, "learning_rate": 1.884211926043398e-07, "logits/chosen": -6.941915035247803, "logits/rejected": -8.13010311126709, "logps/chosen": -1714.9908447265625, "logps/rejected": -2106.353515625, "loss": 0.1782, "nll_loss": 5.203182220458984, "rewards/accuracies": 0.875, "rewards/chosen": -20.565366744995117, "rewards/margins": 5.9251708984375, "rewards/rejected": -26.490537643432617, "step": 134 }, { "epoch": 0.22905263157894737, "grad_norm": 279.1889760986275, "learning_rate": 1.8787480869658978e-07, "logits/chosen": -7.706783294677734, "logits/rejected": -7.740918159484863, "logps/chosen": -2039.910888671875, "logps/rejected": -2072.723876953125, "loss": 0.2864, "nll_loss": 4.911942958831787, "rewards/accuracies": 0.875, "rewards/chosen": -23.88863754272461, "rewards/margins": 1.29465651512146, "rewards/rejected": -25.183292388916016, "step": 136 }, { "epoch": 0.23242105263157894, "grad_norm": 550.1062681428017, "learning_rate": 1.8731665903861985e-07, "logits/chosen": -6.753871917724609, "logits/rejected": -8.267376899719238, "logps/chosen": -1722.739990234375, "logps/rejected": -2134.793212890625, "loss": 0.2628, "nll_loss": 5.156876087188721, "rewards/accuracies": 1.0, "rewards/chosen": -19.4663143157959, "rewards/margins": 6.938652038574219, "rewards/rejected": -26.404966354370117, "step": 138 }, { "epoch": 0.23578947368421052, "grad_norm": 387.2010540404042, "learning_rate": 1.8674681836230768e-07, "logits/chosen": -7.421821117401123, "logits/rejected": -8.06303882598877, "logps/chosen": -1804.21875, "logps/rejected": -2085.41943359375, "loss": 0.4242, "nll_loss": 4.941443920135498, "rewards/accuracies": 0.875, "rewards/chosen": -21.495800018310547, "rewards/margins": 4.65245246887207, "rewards/rejected": -26.148252487182617, "step": 140 }, { "epoch": 0.2391578947368421, "grad_norm": 319.34935073129554, "learning_rate": 1.8616536296486708e-07, "logits/chosen": -6.959980487823486, "logits/rejected": -8.008804321289062, "logps/chosen": -1645.2440185546875, "logps/rejected": -2129.713623046875, "loss": 0.2486, "nll_loss": 5.249794960021973, "rewards/accuracies": 0.875, "rewards/chosen": -19.205081939697266, "rewards/margins": 6.898474216461182, "rewards/rejected": -26.103557586669922, "step": 142 }, { "epoch": 0.24252631578947367, "grad_norm": 179.51700063120762, "learning_rate": 1.855723706986322e-07, "logits/chosen": -6.863515377044678, "logits/rejected": -8.17676067352295, "logps/chosen": -1595.7620849609375, "logps/rejected": -2139.49951171875, "loss": 0.2893, "nll_loss": 5.239679336547852, "rewards/accuracies": 0.75, "rewards/chosen": -19.65725326538086, "rewards/margins": 7.443434238433838, "rewards/rejected": -27.10068702697754, "step": 144 }, { "epoch": 0.24589473684210525, "grad_norm": 275.3682193899652, "learning_rate": 1.8496792096063377e-07, "logits/chosen": -6.194632530212402, "logits/rejected": -8.473308563232422, "logps/chosen": -1469.7490234375, "logps/rejected": -2113.530029296875, "loss": 0.2238, "nll_loss": 5.825500965118408, "rewards/accuracies": 0.875, "rewards/chosen": -18.392675399780273, "rewards/margins": 9.299560546875, "rewards/rejected": -27.692235946655273, "step": 146 }, { "epoch": 0.24926315789473685, "grad_norm": 323.66578296523716, "learning_rate": 1.8435209468196847e-07, "logits/chosen": -7.083681106567383, "logits/rejected": -8.149633407592773, "logps/chosen": -1779.7303466796875, "logps/rejected": -2133.3251953125, "loss": 0.5144, "nll_loss": 4.9437971115112305, "rewards/accuracies": 0.75, "rewards/chosen": -22.35186767578125, "rewards/margins": 6.203129291534424, "rewards/rejected": -28.55499839782715, "step": 148 }, { "epoch": 0.25263157894736843, "grad_norm": 461.451311082762, "learning_rate": 1.8372497431696285e-07, "logits/chosen": -7.620912551879883, "logits/rejected": -8.02045726776123, "logps/chosen": -1918.67724609375, "logps/rejected": -2131.48193359375, "loss": 0.4635, "nll_loss": 4.790881633758545, "rewards/accuracies": 0.75, "rewards/chosen": -24.607433319091797, "rewards/margins": 3.646448850631714, "rewards/rejected": -28.25388526916504, "step": 150 }, { "epoch": 0.256, "grad_norm": 405.566848682187, "learning_rate": 1.830866438321334e-07, "logits/chosen": -7.111250877380371, "logits/rejected": -8.096766471862793, "logps/chosen": -1748.1270751953125, "logps/rejected": -2097.0595703125, "loss": 0.4518, "nll_loss": 5.408980846405029, "rewards/accuracies": 0.75, "rewards/chosen": -23.680131912231445, "rewards/margins": 5.77828311920166, "rewards/rejected": -29.458415985107422, "step": 152 }, { "epoch": 0.2593684210526316, "grad_norm": 298.617818561619, "learning_rate": 1.8243718869494405e-07, "logits/chosen": -7.182644844055176, "logits/rejected": -8.200958251953125, "logps/chosen": -1616.265625, "logps/rejected": -2086.645751953125, "loss": 0.2766, "nll_loss": 5.574889183044434, "rewards/accuracies": 0.875, "rewards/chosen": -21.693281173706055, "rewards/margins": 7.690347671508789, "rewards/rejected": -29.38362693786621, "step": 154 }, { "epoch": 0.26273684210526316, "grad_norm": 297.68208713323685, "learning_rate": 1.8177669586236274e-07, "logits/chosen": -7.177743434906006, "logits/rejected": -8.266185760498047, "logps/chosen": -1737.14892578125, "logps/rejected": -2121.933837890625, "loss": 0.1476, "nll_loss": 5.227602481842041, "rewards/accuracies": 1.0, "rewards/chosen": -23.120718002319336, "rewards/margins": 7.246493816375732, "rewards/rejected": -30.367212295532227, "step": 156 }, { "epoch": 0.26610526315789473, "grad_norm": 321.45477422959516, "learning_rate": 1.811052537692186e-07, "logits/chosen": -7.373076915740967, "logits/rejected": -8.062271118164062, "logps/chosen": -1772.658447265625, "logps/rejected": -2124.108642578125, "loss": 0.1706, "nll_loss": 5.48671817779541, "rewards/accuracies": 0.875, "rewards/chosen": -22.967531204223633, "rewards/margins": 7.354568004608154, "rewards/rejected": -30.322099685668945, "step": 158 }, { "epoch": 0.2694736842105263, "grad_norm": 265.3611042051943, "learning_rate": 1.8042295231636113e-07, "logits/chosen": -6.719966888427734, "logits/rejected": -8.064167976379395, "logps/chosen": -1674.589111328125, "logps/rejected": -2176.885009765625, "loss": 0.3049, "nll_loss": 5.396821022033691, "rewards/accuracies": 1.0, "rewards/chosen": -22.00328254699707, "rewards/margins": 8.764394760131836, "rewards/rejected": -30.767677307128906, "step": 160 }, { "epoch": 0.2728421052631579, "grad_norm": 442.39075275710724, "learning_rate": 1.7972988285862333e-07, "logits/chosen": -7.613976001739502, "logits/rejected": -8.213547706604004, "logps/chosen": -2117.5400390625, "logps/rejected": -2208.39599609375, "loss": 0.2877, "nll_loss": 5.046686172485352, "rewards/accuracies": 1.0, "rewards/chosen": -28.775413513183594, "rewards/margins": 1.5463881492614746, "rewards/rejected": -30.32179832458496, "step": 162 }, { "epoch": 0.27621052631578946, "grad_norm": 232.11714950375813, "learning_rate": 1.7902613819258983e-07, "logits/chosen": -6.078405857086182, "logits/rejected": -8.073712348937988, "logps/chosen": -1325.989990234375, "logps/rejected": -2109.148681640625, "loss": 0.1362, "nll_loss": 5.590985298156738, "rewards/accuracies": 1.0, "rewards/chosen": -17.27791976928711, "rewards/margins": 13.148313522338867, "rewards/rejected": -30.426233291625977, "step": 164 }, { "epoch": 0.27957894736842104, "grad_norm": 178.08792377292178, "learning_rate": 1.7831181254417226e-07, "logits/chosen": -7.393200397491455, "logits/rejected": -7.933506965637207, "logps/chosen": -1770.2550048828125, "logps/rejected": -2183.81884765625, "loss": 0.1839, "nll_loss": 4.8053669929504395, "rewards/accuracies": 1.0, "rewards/chosen": -22.983407974243164, "rewards/margins": 7.7809247970581055, "rewards/rejected": -30.764333724975586, "step": 166 }, { "epoch": 0.2829473684210526, "grad_norm": 422.36016546899873, "learning_rate": 1.7758700155599316e-07, "logits/chosen": -6.5516037940979, "logits/rejected": -7.964224815368652, "logps/chosen": -1520.2445068359375, "logps/rejected": -2135.012451171875, "loss": 0.1957, "nll_loss": 5.540028095245361, "rewards/accuracies": 0.875, "rewards/chosen": -20.019882202148438, "rewards/margins": 11.298989295959473, "rewards/rejected": -31.318870544433594, "step": 168 }, { "epoch": 0.2863157894736842, "grad_norm": 496.76355668466664, "learning_rate": 1.7685180227458002e-07, "logits/chosen": -7.281660556793213, "logits/rejected": -8.243330955505371, "logps/chosen": -1699.278564453125, "logps/rejected": -2149.191650390625, "loss": 0.4263, "nll_loss": 5.4164533615112305, "rewards/accuracies": 0.75, "rewards/chosen": -24.19361686706543, "rewards/margins": 6.679977893829346, "rewards/rejected": -30.873594284057617, "step": 170 }, { "epoch": 0.28968421052631577, "grad_norm": 368.2965007979305, "learning_rate": 1.7610631313737172e-07, "logits/chosen": -7.080427646636963, "logits/rejected": -8.119704246520996, "logps/chosen": -1611.391357421875, "logps/rejected": -2172.8203125, "loss": 0.3694, "nll_loss": 5.258100509643555, "rewards/accuracies": 0.875, "rewards/chosen": -22.043684005737305, "rewards/margins": 9.61347770690918, "rewards/rejected": -31.657163619995117, "step": 172 }, { "epoch": 0.29305263157894734, "grad_norm": 670.6700241176729, "learning_rate": 1.753506339595384e-07, "logits/chosen": -7.517750263214111, "logits/rejected": -7.757349967956543, "logps/chosen": -1787.9560546875, "logps/rejected": -2129.67822265625, "loss": 0.298, "nll_loss": 5.433893203735352, "rewards/accuracies": 0.625, "rewards/chosen": -25.235340118408203, "rewards/margins": 6.193837642669678, "rewards/rejected": -31.42917823791504, "step": 174 }, { "epoch": 0.296421052631579, "grad_norm": 410.62113772917246, "learning_rate": 1.7458486592061701e-07, "logits/chosen": -6.625953674316406, "logits/rejected": -8.243332862854004, "logps/chosen": -1637.386474609375, "logps/rejected": -2249.2197265625, "loss": 0.3263, "nll_loss": 5.135756969451904, "rewards/accuracies": 1.0, "rewards/chosen": -22.93773651123047, "rewards/margins": 10.093299865722656, "rewards/rejected": -33.03103256225586, "step": 176 }, { "epoch": 0.29978947368421055, "grad_norm": 518.14733064062, "learning_rate": 1.7380911155096407e-07, "logits/chosen": -6.312992572784424, "logits/rejected": -7.9758687019348145, "logps/chosen": -1488.8402099609375, "logps/rejected": -2148.6962890625, "loss": 0.591, "nll_loss": 5.555506706237793, "rewards/accuracies": 0.75, "rewards/chosen": -20.87957763671875, "rewards/margins": 10.946573257446289, "rewards/rejected": -31.82615089416504, "step": 178 }, { "epoch": 0.3031578947368421, "grad_norm": 230.12282725753218, "learning_rate": 1.7302347471802795e-07, "logits/chosen": -6.5913238525390625, "logits/rejected": -8.363908767700195, "logps/chosen": -1680.2467041015625, "logps/rejected": -2223.9873046875, "loss": 0.3578, "nll_loss": 5.693438529968262, "rewards/accuracies": 1.0, "rewards/chosen": -23.647768020629883, "rewards/margins": 8.622623443603516, "rewards/rejected": -32.27039337158203, "step": 180 }, { "epoch": 0.3065263157894737, "grad_norm": 290.90448926580694, "learning_rate": 1.7222806061244146e-07, "logits/chosen": -6.316551208496094, "logits/rejected": -8.457919120788574, "logps/chosen": -1597.56103515625, "logps/rejected": -2268.66259765625, "loss": 0.1266, "nll_loss": 5.54366397857666, "rewards/accuracies": 1.0, "rewards/chosen": -22.09076499938965, "rewards/margins": 9.565703392028809, "rewards/rejected": -31.656469345092773, "step": 182 }, { "epoch": 0.3098947368421053, "grad_norm": 219.15996405308613, "learning_rate": 1.7142297573393788e-07, "logits/chosen": -7.972108840942383, "logits/rejected": -8.042094230651855, "logps/chosen": -1998.596923828125, "logps/rejected": -2172.12255859375, "loss": 0.3072, "nll_loss": 5.00222110748291, "rewards/accuracies": 0.875, "rewards/chosen": -28.76752281188965, "rewards/margins": 3.3450047969818115, "rewards/rejected": -32.112525939941406, "step": 184 }, { "epoch": 0.31326315789473685, "grad_norm": 322.85198814627114, "learning_rate": 1.7060832787709138e-07, "logits/chosen": -6.03144645690918, "logits/rejected": -8.47148323059082, "logps/chosen": -1274.5035400390625, "logps/rejected": -2136.2109375, "loss": 0.3631, "nll_loss": 5.840561866760254, "rewards/accuracies": 0.875, "rewards/chosen": -17.418729782104492, "rewards/margins": 15.012275695800781, "rewards/rejected": -32.43100357055664, "step": 186 }, { "epoch": 0.31663157894736843, "grad_norm": 513.7476885592771, "learning_rate": 1.697842261168843e-07, "logits/chosen": -7.334516525268555, "logits/rejected": -8.225534439086914, "logps/chosen": -1927.53515625, "logps/rejected": -2148.278564453125, "loss": 0.3361, "nll_loss": 5.305893898010254, "rewards/accuracies": 0.875, "rewards/chosen": -26.597455978393555, "rewards/margins": 5.804078102111816, "rewards/rejected": -32.40153503417969, "step": 188 }, { "epoch": 0.32, "grad_norm": 163.682762913509, "learning_rate": 1.6895078079410266e-07, "logits/chosen": -5.538216590881348, "logits/rejected": -8.494758605957031, "logps/chosen": -1233.1563720703125, "logps/rejected": -2203.232177734375, "loss": 0.179, "nll_loss": 5.9964189529418945, "rewards/accuracies": 0.875, "rewards/chosen": -17.165884017944336, "rewards/margins": 15.676169395446777, "rewards/rejected": -32.84205627441406, "step": 190 }, { "epoch": 0.3233684210526316, "grad_norm": 308.2898266039604, "learning_rate": 1.6810810350056258e-07, "logits/chosen": -6.971872329711914, "logits/rejected": -8.188929557800293, "logps/chosen": -1851.5859375, "logps/rejected": -2194.031982421875, "loss": 0.4115, "nll_loss": 5.3788743019104, "rewards/accuracies": 0.75, "rewards/chosen": -26.422870635986328, "rewards/margins": 6.5016984939575195, "rewards/rejected": -32.92456817626953, "step": 192 }, { "epoch": 0.32673684210526316, "grad_norm": 412.5565491541735, "learning_rate": 1.672563070641688e-07, "logits/chosen": -7.640994548797607, "logits/rejected": -8.178326606750488, "logps/chosen": -2078.18115234375, "logps/rejected": -2178.9091796875, "loss": 0.2875, "nll_loss": 5.410080432891846, "rewards/accuracies": 0.625, "rewards/chosen": -30.432947158813477, "rewards/margins": 2.4267215728759766, "rewards/rejected": -32.85966873168945, "step": 194 }, { "epoch": 0.33010526315789473, "grad_norm": 469.2498668087184, "learning_rate": 1.6639550553380816e-07, "logits/chosen": -7.088956356048584, "logits/rejected": -8.258764266967773, "logps/chosen": -1782.86376953125, "logps/rejected": -2211.70703125, "loss": 0.369, "nll_loss": 5.0121073722839355, "rewards/accuracies": 0.875, "rewards/chosen": -26.950332641601562, "rewards/margins": 7.496822357177734, "rewards/rejected": -34.44715118408203, "step": 196 }, { "epoch": 0.3334736842105263, "grad_norm": 236.59088336990078, "learning_rate": 1.6552581416407916e-07, "logits/chosen": -8.196027755737305, "logits/rejected": -7.7416181564331055, "logps/chosen": -2154.635498046875, "logps/rejected": -2205.575927734375, "loss": 0.3022, "nll_loss": 4.687681198120117, "rewards/accuracies": 0.75, "rewards/chosen": -32.5702018737793, "rewards/margins": 2.931349515914917, "rewards/rejected": -35.501556396484375, "step": 198 }, { "epoch": 0.3368421052631579, "grad_norm": 413.82963735092727, "learning_rate": 1.6464734939986035e-07, "logits/chosen": -7.169932842254639, "logits/rejected": -8.24282455444336, "logps/chosen": -1997.083984375, "logps/rejected": -2223.7470703125, "loss": 0.3572, "nll_loss": 5.281017303466797, "rewards/accuracies": 1.0, "rewards/chosen": -30.46786117553711, "rewards/margins": 5.10554313659668, "rewards/rejected": -35.573402404785156, "step": 200 }, { "epoch": 0.34021052631578946, "grad_norm": 415.7728496426865, "learning_rate": 1.637602288607192e-07, "logits/chosen": -6.219871520996094, "logits/rejected": -8.299065589904785, "logps/chosen": -1463.1075439453125, "logps/rejected": -2223.69775390625, "loss": 0.2563, "nll_loss": 5.272424697875977, "rewards/accuracies": 0.875, "rewards/chosen": -20.643104553222656, "rewards/margins": 13.95057487487793, "rewards/rejected": -34.59368133544922, "step": 202 }, { "epoch": 0.34357894736842104, "grad_norm": 422.85323621204554, "learning_rate": 1.6286457132516383e-07, "logits/chosen": -6.83880615234375, "logits/rejected": -8.385416030883789, "logps/chosen": -1577.858154296875, "logps/rejected": -2194.265869140625, "loss": 0.3599, "nll_loss": 5.622811317443848, "rewards/accuracies": 0.875, "rewards/chosen": -22.948638916015625, "rewards/margins": 11.672065734863281, "rewards/rejected": -34.620704650878906, "step": 204 }, { "epoch": 0.3469473684210526, "grad_norm": 490.11145097314346, "learning_rate": 1.6196049671473953e-07, "logits/chosen": -7.446739196777344, "logits/rejected": -8.425362586975098, "logps/chosen": -1768.897705078125, "logps/rejected": -2202.29443359375, "loss": 0.2982, "nll_loss": 5.6991496086120605, "rewards/accuracies": 0.75, "rewards/chosen": -28.04994773864746, "rewards/margins": 6.368475914001465, "rewards/rejected": -34.41842269897461, "step": 206 }, { "epoch": 0.3503157894736842, "grad_norm": 530.9413220874789, "learning_rate": 1.61048126077972e-07, "logits/chosen": -7.343864440917969, "logits/rejected": -8.190047264099121, "logps/chosen": -1754.0421142578125, "logps/rejected": -2148.49072265625, "loss": 0.2674, "nll_loss": 5.666429042816162, "rewards/accuracies": 1.0, "rewards/chosen": -25.846054077148438, "rewards/margins": 8.071301460266113, "rewards/rejected": -33.917354583740234, "step": 208 }, { "epoch": 0.35368421052631577, "grad_norm": 166.06683989436303, "learning_rate": 1.6012758157416018e-07, "logits/chosen": -7.329439640045166, "logits/rejected": -7.938960075378418, "logps/chosen": -1708.5814208984375, "logps/rejected": -2173.199462890625, "loss": 0.1976, "nll_loss": 5.0503830909729, "rewards/accuracies": 0.875, "rewards/chosen": -24.269054412841797, "rewards/margins": 9.374040603637695, "rewards/rejected": -33.643096923828125, "step": 210 }, { "epoch": 0.35705263157894734, "grad_norm": 525.5297922097282, "learning_rate": 1.5919898645701988e-07, "logits/chosen": -6.739994049072266, "logits/rejected": -8.293057441711426, "logps/chosen": -1717.248046875, "logps/rejected": -2190.51171875, "loss": 0.4755, "nll_loss": 4.879263877868652, "rewards/accuracies": 0.875, "rewards/chosen": -24.138900756835938, "rewards/margins": 8.57326602935791, "rewards/rejected": -32.71216583251953, "step": 212 }, { "epoch": 0.3604210526315789, "grad_norm": 548.5346673183018, "learning_rate": 1.5826246505818112e-07, "logits/chosen": -6.99255895614624, "logits/rejected": -8.398011207580566, "logps/chosen": -1611.18701171875, "logps/rejected": -2188.08642578125, "loss": 0.432, "nll_loss": 5.333250999450684, "rewards/accuracies": 0.875, "rewards/chosen": -23.24384880065918, "rewards/margins": 10.111581802368164, "rewards/rejected": -33.355430603027344, "step": 214 }, { "epoch": 0.36378947368421055, "grad_norm": 185.93773109229338, "learning_rate": 1.573181427705411e-07, "logits/chosen": -7.210558891296387, "logits/rejected": -8.003637313842773, "logps/chosen": -1898.2894287109375, "logps/rejected": -2259.249755859375, "loss": 0.1156, "nll_loss": 5.2748212814331055, "rewards/accuracies": 1.0, "rewards/chosen": -27.62005043029785, "rewards/margins": 6.235610485076904, "rewards/rejected": -33.85565948486328, "step": 216 }, { "epoch": 0.3671578947368421, "grad_norm": 298.31090901524203, "learning_rate": 1.5636614603147512e-07, "logits/chosen": -7.3358025550842285, "logits/rejected": -8.385098457336426, "logps/chosen": -1998.4923095703125, "logps/rejected": -2187.198486328125, "loss": 0.2525, "nll_loss": 5.315785884857178, "rewards/accuracies": 0.625, "rewards/chosen": -29.09065818786621, "rewards/margins": 3.6531078815460205, "rewards/rejected": -32.74376678466797, "step": 218 }, { "epoch": 0.3705263157894737, "grad_norm": 377.3450841073725, "learning_rate": 1.5540660230590748e-07, "logits/chosen": -6.510415077209473, "logits/rejected": -8.296943664550781, "logps/chosen": -1665.1639404296875, "logps/rejected": -2214.30078125, "loss": 0.438, "nll_loss": 5.418496131896973, "rewards/accuracies": 1.0, "rewards/chosen": -23.495925903320312, "rewards/margins": 10.108709335327148, "rewards/rejected": -33.604637145996094, "step": 220 }, { "epoch": 0.3738947368421053, "grad_norm": 443.30578117705613, "learning_rate": 1.5443964006924507e-07, "logits/chosen": -6.681422710418701, "logits/rejected": -8.464709281921387, "logps/chosen": -1588.643310546875, "logps/rejected": -2177.72705078125, "loss": 0.2768, "nll_loss": 6.007861137390137, "rewards/accuracies": 0.75, "rewards/chosen": -23.003562927246094, "rewards/margins": 10.166757583618164, "rewards/rejected": -33.17032241821289, "step": 222 }, { "epoch": 0.37726315789473686, "grad_norm": 163.54164941613496, "learning_rate": 1.534653887901754e-07, "logits/chosen": -5.698173999786377, "logits/rejected": -8.45270824432373, "logps/chosen": -1357.1298828125, "logps/rejected": -2201.384765625, "loss": 0.1406, "nll_loss": 5.912946701049805, "rewards/accuracies": 1.0, "rewards/chosen": -17.897022247314453, "rewards/margins": 14.996172904968262, "rewards/rejected": -32.8931999206543, "step": 224 }, { "epoch": 0.38063157894736843, "grad_norm": 319.17016425411697, "learning_rate": 1.5248397891333183e-07, "logits/chosen": -6.813119888305664, "logits/rejected": -8.12168025970459, "logps/chosen": -1691.3763427734375, "logps/rejected": -2194.62646484375, "loss": 0.2993, "nll_loss": 5.037168025970459, "rewards/accuracies": 0.625, "rewards/chosen": -24.049442291259766, "rewards/margins": 8.878725051879883, "rewards/rejected": -32.928165435791016, "step": 226 }, { "epoch": 0.384, "grad_norm": 284.69578061502773, "learning_rate": 1.51495541841828e-07, "logits/chosen": -6.420520305633545, "logits/rejected": -8.212552070617676, "logps/chosen": -1653.12890625, "logps/rejected": -2208.427734375, "loss": 0.2744, "nll_loss": 5.3444905281066895, "rewards/accuracies": 0.875, "rewards/chosen": -23.33785629272461, "rewards/margins": 9.556280136108398, "rewards/rejected": -32.894134521484375, "step": 228 }, { "epoch": 0.3873684210526316, "grad_norm": 284.69191132696415, "learning_rate": 1.5050020991966403e-07, "logits/chosen": -6.635988712310791, "logits/rejected": -7.83236837387085, "logps/chosen": -1533.177978515625, "logps/rejected": -2170.71044921875, "loss": 0.1905, "nll_loss": 5.389209270477295, "rewards/accuracies": 1.0, "rewards/chosen": -21.042217254638672, "rewards/margins": 11.956886291503906, "rewards/rejected": -32.99910354614258, "step": 230 }, { "epoch": 0.39073684210526316, "grad_norm": 733.6966760862421, "learning_rate": 1.4949811641400668e-07, "logits/chosen": -7.570094108581543, "logits/rejected": -7.502762794494629, "logps/chosen": -1841.1671142578125, "logps/rejected": -2136.026611328125, "loss": 0.5221, "nll_loss": 4.983242988586426, "rewards/accuracies": 0.75, "rewards/chosen": -25.91825294494629, "rewards/margins": 6.6433634757995605, "rewards/rejected": -32.561614990234375, "step": 232 }, { "epoch": 0.39410526315789474, "grad_norm": 188.5369451526676, "learning_rate": 1.484893954973458e-07, "logits/chosen": -6.149730205535889, "logits/rejected": -8.597711563110352, "logps/chosen": -1442.3626708984375, "logps/rejected": -2134.160888671875, "loss": 0.129, "nll_loss": 5.606182098388672, "rewards/accuracies": 0.875, "rewards/chosen": -18.86325454711914, "rewards/margins": 13.016372680664062, "rewards/rejected": -31.879629135131836, "step": 234 }, { "epoch": 0.3974736842105263, "grad_norm": 305.89732243842866, "learning_rate": 1.4747418222952995e-07, "logits/chosen": -6.757414817810059, "logits/rejected": -8.115114212036133, "logps/chosen": -1697.4263916015625, "logps/rejected": -2184.919921875, "loss": 0.2413, "nll_loss": 5.372623920440674, "rewards/accuracies": 0.875, "rewards/chosen": -22.677663803100586, "rewards/margins": 8.713125228881836, "rewards/rejected": -31.39078712463379, "step": 236 }, { "epoch": 0.4008421052631579, "grad_norm": 341.4679249212142, "learning_rate": 1.4645261253968259e-07, "logits/chosen": -6.988116264343262, "logits/rejected": -8.155961036682129, "logps/chosen": -1654.0662841796875, "logps/rejected": -2215.248291015625, "loss": 0.2575, "nll_loss": 5.236739158630371, "rewards/accuracies": 1.0, "rewards/chosen": -21.976306915283203, "rewards/margins": 10.24085807800293, "rewards/rejected": -32.2171630859375, "step": 238 }, { "epoch": 0.40421052631578946, "grad_norm": 132.8280211518827, "learning_rate": 1.454248232080026e-07, "logits/chosen": -7.433531761169434, "logits/rejected": -8.119049072265625, "logps/chosen": -1924.980712890625, "logps/rejected": -2134.167724609375, "loss": 0.2044, "nll_loss": 5.498749256134033, "rewards/accuracies": 0.875, "rewards/chosen": -26.4937801361084, "rewards/margins": 5.2426652908325195, "rewards/rejected": -31.736446380615234, "step": 240 }, { "epoch": 0.40757894736842104, "grad_norm": 340.2818569766834, "learning_rate": 1.4439095184745024e-07, "logits/chosen": -6.053187370300293, "logits/rejected": -8.49085807800293, "logps/chosen": -1365.32568359375, "logps/rejected": -2177.580078125, "loss": 0.271, "nll_loss": 5.570743083953857, "rewards/accuracies": 0.875, "rewards/chosen": -18.466915130615234, "rewards/margins": 12.624629974365234, "rewards/rejected": -31.09154510498047, "step": 242 }, { "epoch": 0.4109473684210526, "grad_norm": 214.41734275706338, "learning_rate": 1.4335113688532182e-07, "logits/chosen": -6.56601619720459, "logits/rejected": -8.04720401763916, "logps/chosen": -1554.8018798828125, "logps/rejected": -2144.083251953125, "loss": 0.138, "nll_loss": 5.4142374992370605, "rewards/accuracies": 1.0, "rewards/chosen": -20.857574462890625, "rewards/margins": 11.797863006591797, "rewards/rejected": -32.655433654785156, "step": 244 }, { "epoch": 0.4143157894736842, "grad_norm": 337.35948463767846, "learning_rate": 1.423055175447155e-07, "logits/chosen": -6.994697570800781, "logits/rejected": -7.979827880859375, "logps/chosen": -1652.507568359375, "logps/rejected": -2181.640625, "loss": 0.329, "nll_loss": 5.518115043640137, "rewards/accuracies": 0.875, "rewards/chosen": -23.70856475830078, "rewards/margins": 8.838756561279297, "rewards/rejected": -32.54732131958008, "step": 246 }, { "epoch": 0.41768421052631577, "grad_norm": 274.454131527474, "learning_rate": 1.4125423382589048e-07, "logits/chosen": -6.993724346160889, "logits/rejected": -8.257303237915039, "logps/chosen": -1779.2265625, "logps/rejected": -2149.052490234375, "loss": 0.2965, "nll_loss": 5.484703063964844, "rewards/accuracies": 1.0, "rewards/chosen": -24.876686096191406, "rewards/margins": 7.389163970947266, "rewards/rejected": -32.26585006713867, "step": 248 }, { "epoch": 0.42105263157894735, "grad_norm": 352.35818290838654, "learning_rate": 1.401974264875218e-07, "logits/chosen": -7.193760871887207, "logits/rejected": -8.0586519241333, "logps/chosen": -1831.6397705078125, "logps/rejected": -2227.03857421875, "loss": 0.2662, "nll_loss": 5.117783546447754, "rewards/accuracies": 0.875, "rewards/chosen": -25.96319007873535, "rewards/margins": 6.454355239868164, "rewards/rejected": -32.41754150390625, "step": 250 }, { "epoch": 0.4244210526315789, "grad_norm": 255.0171256504159, "learning_rate": 1.391352370278541e-07, "logits/chosen": -6.700958728790283, "logits/rejected": -8.45335865020752, "logps/chosen": -1634.654541015625, "logps/rejected": -2221.049072265625, "loss": 0.2647, "nll_loss": 6.0273308753967285, "rewards/accuracies": 0.875, "rewards/chosen": -23.497007369995117, "rewards/margins": 10.51921558380127, "rewards/rejected": -34.0162239074707, "step": 252 }, { "epoch": 0.42778947368421055, "grad_norm": 481.3772072138439, "learning_rate": 1.3806780766575588e-07, "logits/chosen": -7.795599460601807, "logits/rejected": -8.443253517150879, "logps/chosen": -1971.610107421875, "logps/rejected": -2192.519287109375, "loss": 0.203, "nll_loss": 5.141743183135986, "rewards/accuracies": 0.875, "rewards/chosen": -28.521339416503906, "rewards/margins": 5.977480888366699, "rewards/rejected": -34.49882125854492, "step": 254 }, { "epoch": 0.43115789473684213, "grad_norm": 451.03866242263535, "learning_rate": 1.3699528132167776e-07, "logits/chosen": -7.792696475982666, "logits/rejected": -7.79255485534668, "logps/chosen": -1888.9193115234375, "logps/rejected": -2135.4482421875, "loss": 0.3684, "nll_loss": 4.749884128570557, "rewards/accuracies": 0.875, "rewards/chosen": -26.909269332885742, "rewards/margins": 6.3855719566345215, "rewards/rejected": -33.29484176635742, "step": 256 }, { "epoch": 0.4345263157894737, "grad_norm": 242.47130951413612, "learning_rate": 1.3591780159851627e-07, "logits/chosen": -6.771208763122559, "logits/rejected": -8.258780479431152, "logps/chosen": -1740.8980712890625, "logps/rejected": -2198.968994140625, "loss": 0.1598, "nll_loss": 5.290096759796143, "rewards/accuracies": 1.0, "rewards/chosen": -24.707876205444336, "rewards/margins": 9.988024711608887, "rewards/rejected": -34.695899963378906, "step": 258 }, { "epoch": 0.4378947368421053, "grad_norm": 499.1497771210561, "learning_rate": 1.3483551276238686e-07, "logits/chosen": -6.881796836853027, "logits/rejected": -8.311073303222656, "logps/chosen": -1801.54443359375, "logps/rejected": -2261.04150390625, "loss": 0.5297, "nll_loss": 4.906632900238037, "rewards/accuracies": 1.0, "rewards/chosen": -25.078868865966797, "rewards/margins": 9.138972282409668, "rewards/rejected": -34.217838287353516, "step": 260 }, { "epoch": 0.44126315789473686, "grad_norm": 340.10309787989286, "learning_rate": 1.3374855972330756e-07, "logits/chosen": -7.035615921020508, "logits/rejected": -8.248941421508789, "logps/chosen": -1872.8837890625, "logps/rejected": -2217.878173828125, "loss": 0.306, "nll_loss": 5.385906219482422, "rewards/accuracies": 1.0, "rewards/chosen": -26.4267520904541, "rewards/margins": 7.073578357696533, "rewards/rejected": -33.500328063964844, "step": 262 }, { "epoch": 0.44463157894736843, "grad_norm": 225.67805816583055, "learning_rate": 1.3265708801579666e-07, "logits/chosen": -6.937837600708008, "logits/rejected": -8.376426696777344, "logps/chosen": -1765.403076171875, "logps/rejected": -2198.61767578125, "loss": 0.1877, "nll_loss": 5.633481502532959, "rewards/accuracies": 0.875, "rewards/chosen": -25.389394760131836, "rewards/margins": 7.419838905334473, "rewards/rejected": -32.80923080444336, "step": 264 }, { "epoch": 0.448, "grad_norm": 91.96492729058454, "learning_rate": 1.3156124377938697e-07, "logits/chosen": -7.8583831787109375, "logits/rejected": -8.153332710266113, "logps/chosen": -2142.6328125, "logps/rejected": -2167.61376953125, "loss": 0.2784, "nll_loss": 4.982174396514893, "rewards/accuracies": 0.875, "rewards/chosen": -30.6492919921875, "rewards/margins": 2.7910072803497314, "rewards/rejected": -33.4402961730957, "step": 266 }, { "epoch": 0.4513684210526316, "grad_norm": 199.4223395992786, "learning_rate": 1.3046117373905865e-07, "logits/chosen": -6.933570861816406, "logits/rejected": -8.099403381347656, "logps/chosen": -1728.806396484375, "logps/rejected": -2209.177734375, "loss": 0.0685, "nll_loss": 5.4580488204956055, "rewards/accuracies": 1.0, "rewards/chosen": -23.95172882080078, "rewards/margins": 10.292177200317383, "rewards/rejected": -34.2439079284668, "step": 268 }, { "epoch": 0.45473684210526316, "grad_norm": 216.84232462782722, "learning_rate": 1.2935702518559397e-07, "logits/chosen": -7.183603763580322, "logits/rejected": -8.098438262939453, "logps/chosen": -1903.0699462890625, "logps/rejected": -2204.51904296875, "loss": 0.2503, "nll_loss": 5.397114276885986, "rewards/accuracies": 0.625, "rewards/chosen": -27.064674377441406, "rewards/margins": 5.608869552612305, "rewards/rejected": -32.673545837402344, "step": 270 }, { "epoch": 0.45810526315789474, "grad_norm": 380.9095200963084, "learning_rate": 1.2824894595585636e-07, "logits/chosen": -6.462231636047363, "logits/rejected": -8.369811058044434, "logps/chosen": -1741.0838623046875, "logps/rejected": -2199.78466796875, "loss": 0.2068, "nll_loss": 6.009423732757568, "rewards/accuracies": 0.75, "rewards/chosen": -26.056018829345703, "rewards/margins": 7.744605541229248, "rewards/rejected": -33.80062484741211, "step": 272 }, { "epoch": 0.4614736842105263, "grad_norm": 137.60298699973777, "learning_rate": 1.27137084412996e-07, "logits/chosen": -6.126290321350098, "logits/rejected": -8.357515335083008, "logps/chosen": -1254.0537109375, "logps/rejected": -2182.109130859375, "loss": 0.1443, "nll_loss": 5.386259078979492, "rewards/accuracies": 1.0, "rewards/chosen": -17.421083450317383, "rewards/margins": 17.08145523071289, "rewards/rejected": -34.50253677368164, "step": 274 }, { "epoch": 0.4648421052631579, "grad_norm": 238.72881941389062, "learning_rate": 1.260215894265852e-07, "logits/chosen": -6.577013969421387, "logits/rejected": -8.490443229675293, "logps/chosen": -1655.7412109375, "logps/rejected": -2204.900390625, "loss": 0.2314, "nll_loss": 5.632746696472168, "rewards/accuracies": 1.0, "rewards/chosen": -23.928146362304688, "rewards/margins": 11.161284446716309, "rewards/rejected": -35.08943176269531, "step": 276 }, { "epoch": 0.46821052631578947, "grad_norm": 205.0849481007809, "learning_rate": 1.2490261035268613e-07, "logits/chosen": -7.492171287536621, "logits/rejected": -8.131207466125488, "logps/chosen": -1958.902099609375, "logps/rejected": -2155.704345703125, "loss": 0.1155, "nll_loss": 5.272936820983887, "rewards/accuracies": 1.0, "rewards/chosen": -30.05508804321289, "rewards/margins": 5.22835111618042, "rewards/rejected": -35.2834358215332, "step": 278 }, { "epoch": 0.47157894736842104, "grad_norm": 226.262655032818, "learning_rate": 1.2378029701385286e-07, "logits/chosen": -6.374780178070068, "logits/rejected": -8.473326683044434, "logps/chosen": -1493.491455078125, "logps/rejected": -2233.69140625, "loss": 0.1602, "nll_loss": 5.51650333404541, "rewards/accuracies": 0.875, "rewards/chosen": -22.54718017578125, "rewards/margins": 13.27789306640625, "rewards/rejected": -35.8250732421875, "step": 280 }, { "epoch": 0.4749473684210526, "grad_norm": 304.14594820990135, "learning_rate": 1.2265479967907158e-07, "logits/chosen": -7.050374507904053, "logits/rejected": -8.2347412109375, "logps/chosen": -1850.6962890625, "logps/rejected": -2199.18798828125, "loss": 0.3842, "nll_loss": 5.64528226852417, "rewards/accuracies": 0.875, "rewards/chosen": -28.668458938598633, "rewards/margins": 7.283104419708252, "rewards/rejected": -35.95156478881836, "step": 282 }, { "epoch": 0.4783157894736842, "grad_norm": 250.05279911200753, "learning_rate": 1.2152626904364064e-07, "logits/chosen": -6.778600215911865, "logits/rejected": -8.34770679473877, "logps/chosen": -1604.2724609375, "logps/rejected": -2214.350830078125, "loss": 0.1214, "nll_loss": 5.6927690505981445, "rewards/accuracies": 1.0, "rewards/chosen": -24.976463317871094, "rewards/margins": 11.123945236206055, "rewards/rejected": -36.10041046142578, "step": 284 }, { "epoch": 0.48168421052631577, "grad_norm": 49.88924745592731, "learning_rate": 1.2039485620899367e-07, "logits/chosen": -7.990779876708984, "logits/rejected": -7.396949291229248, "logps/chosen": -2097.546142578125, "logps/rejected": -2177.14501953125, "loss": 0.092, "nll_loss": 5.315004348754883, "rewards/accuracies": 0.875, "rewards/chosen": -33.4441032409668, "rewards/margins": 4.195204734802246, "rewards/rejected": -37.63930892944336, "step": 286 }, { "epoch": 0.48505263157894735, "grad_norm": 298.7045075195552, "learning_rate": 1.1926071266246824e-07, "logits/chosen": -7.262898921966553, "logits/rejected": -7.93293571472168, "logps/chosen": -1817.0556640625, "logps/rejected": -2241.826904296875, "loss": 0.378, "nll_loss": 5.347080230712891, "rewards/accuracies": 0.875, "rewards/chosen": -29.16703224182129, "rewards/margins": 7.644922733306885, "rewards/rejected": -36.81195831298828, "step": 288 }, { "epoch": 0.4884210526315789, "grad_norm": 305.83167858317097, "learning_rate": 1.1812399025702289e-07, "logits/chosen": -7.386429309844971, "logits/rejected": -7.993374347686768, "logps/chosen": -1947.408935546875, "logps/rejected": -2193.298583984375, "loss": 0.2834, "nll_loss": 5.327097415924072, "rewards/accuracies": 0.75, "rewards/chosen": -30.927370071411133, "rewards/margins": 5.908793926239014, "rewards/rejected": -36.83616256713867, "step": 290 }, { "epoch": 0.4917894736842105, "grad_norm": 457.2760039079243, "learning_rate": 1.1698484119090518e-07, "logits/chosen": -7.0947346687316895, "logits/rejected": -8.351741790771484, "logps/chosen": -1785.537353515625, "logps/rejected": -2246.34912109375, "loss": 0.2948, "nll_loss": 5.337271213531494, "rewards/accuracies": 0.875, "rewards/chosen": -28.525087356567383, "rewards/margins": 9.390640258789062, "rewards/rejected": -37.91572952270508, "step": 292 }, { "epoch": 0.49515789473684213, "grad_norm": 296.42125851564697, "learning_rate": 1.1584341798727364e-07, "logits/chosen": -6.732482433319092, "logits/rejected": -8.335723876953125, "logps/chosen": -1633.8388671875, "logps/rejected": -2277.510498046875, "loss": 0.2649, "nll_loss": 5.798310279846191, "rewards/accuracies": 0.875, "rewards/chosen": -25.833433151245117, "rewards/margins": 11.791748046875, "rewards/rejected": -37.62518310546875, "step": 294 }, { "epoch": 0.4985263157894737, "grad_norm": 439.27146717605046, "learning_rate": 1.1469987347377601e-07, "logits/chosen": -7.6690354347229, "logits/rejected": -7.856517791748047, "logps/chosen": -2050.155029296875, "logps/rejected": -2208.57666015625, "loss": 0.3627, "nll_loss": 5.227299213409424, "rewards/accuracies": 0.875, "rewards/chosen": -33.6103515625, "rewards/margins": 2.915879249572754, "rewards/rejected": -36.52622985839844, "step": 296 }, { "epoch": 0.5018947368421053, "grad_norm": 318.9522275161282, "learning_rate": 1.1355436076208687e-07, "logits/chosen": -7.616082668304443, "logits/rejected": -8.118494033813477, "logps/chosen": -1956.5233154296875, "logps/rejected": -2223.263916015625, "loss": 0.2973, "nll_loss": 5.363955497741699, "rewards/accuracies": 0.75, "rewards/chosen": -32.195335388183594, "rewards/margins": 5.070850849151611, "rewards/rejected": -37.26618576049805, "step": 298 }, { "epoch": 0.5052631578947369, "grad_norm": 141.04978897095168, "learning_rate": 1.124070332274071e-07, "logits/chosen": -7.996880054473877, "logits/rejected": -8.085912704467773, "logps/chosen": -2062.050537109375, "logps/rejected": -2186.66162109375, "loss": 0.0746, "nll_loss": 5.05142879486084, "rewards/accuracies": 0.875, "rewards/chosen": -33.29553985595703, "rewards/margins": 4.88584041595459, "rewards/rejected": -38.18138122558594, "step": 300 }, { "epoch": 0.5086315789473684, "grad_norm": 373.17587592849566, "learning_rate": 1.112580444879283e-07, "logits/chosen": -7.303493022918701, "logits/rejected": -7.817137718200684, "logps/chosen": -1930.81689453125, "logps/rejected": -2200.581298828125, "loss": 0.4021, "nll_loss": 5.62571907043457, "rewards/accuracies": 0.75, "rewards/chosen": -31.353313446044922, "rewards/margins": 5.793161869049072, "rewards/rejected": -37.14647674560547, "step": 302 }, { "epoch": 0.512, "grad_norm": 282.00796475897624, "learning_rate": 1.1010754838426426e-07, "logits/chosen": -7.355342388153076, "logits/rejected": -8.244458198547363, "logps/chosen": -1883.871337890625, "logps/rejected": -2235.58447265625, "loss": 0.2066, "nll_loss": 5.333528995513916, "rewards/accuracies": 1.0, "rewards/chosen": -29.595571517944336, "rewards/margins": 7.5849127769470215, "rewards/rejected": -37.180484771728516, "step": 304 }, { "epoch": 0.5153684210526316, "grad_norm": 243.79920977620088, "learning_rate": 1.089556989588532e-07, "logits/chosen": -7.213363170623779, "logits/rejected": -8.331066131591797, "logps/chosen": -1723.94287109375, "logps/rejected": -2178.452880859375, "loss": 0.1619, "nll_loss": 5.337440490722656, "rewards/accuracies": 0.625, "rewards/chosen": -27.8447265625, "rewards/margins": 9.870349884033203, "rewards/rejected": -37.71508026123047, "step": 306 }, { "epoch": 0.5187368421052632, "grad_norm": 199.39976335241724, "learning_rate": 1.078026504353325e-07, "logits/chosen": -7.63919734954834, "logits/rejected": -8.216974258422852, "logps/chosen": -2052.788330078125, "logps/rejected": -2217.202880859375, "loss": 0.2817, "nll_loss": 5.197584629058838, "rewards/accuracies": 1.0, "rewards/chosen": -31.69765853881836, "rewards/margins": 5.5009331703186035, "rewards/rejected": -37.19859313964844, "step": 308 }, { "epoch": 0.5221052631578947, "grad_norm": 620.1262472271271, "learning_rate": 1.0664855719788934e-07, "logits/chosen": -7.262815952301025, "logits/rejected": -8.217065811157227, "logps/chosen": -2147.777587890625, "logps/rejected": -2262.103271484375, "loss": 0.2195, "nll_loss": 5.310882568359375, "rewards/accuracies": 0.875, "rewards/chosen": -35.0344123840332, "rewards/margins": 2.332733631134033, "rewards/rejected": -37.367149353027344, "step": 310 }, { "epoch": 0.5254736842105263, "grad_norm": 395.93023375337225, "learning_rate": 1.0549357377059005e-07, "logits/chosen": -7.084537506103516, "logits/rejected": -8.080052375793457, "logps/chosen": -1756.028076171875, "logps/rejected": -2257.95703125, "loss": 0.479, "nll_loss": 5.514889717102051, "rewards/accuracies": 0.875, "rewards/chosen": -28.614755630493164, "rewards/margins": 9.187152862548828, "rewards/rejected": -37.801910400390625, "step": 312 }, { "epoch": 0.5288421052631579, "grad_norm": 200.0389317576706, "learning_rate": 1.0433785479669037e-07, "logits/chosen": -7.246962547302246, "logits/rejected": -8.294339179992676, "logps/chosen": -1767.36376953125, "logps/rejected": -2233.66064453125, "loss": 0.2677, "nll_loss": 5.436453342437744, "rewards/accuracies": 1.0, "rewards/chosen": -28.277070999145508, "rewards/margins": 9.897624969482422, "rewards/rejected": -38.17469787597656, "step": 314 }, { "epoch": 0.5322105263157895, "grad_norm": 232.98730015013695, "learning_rate": 1.0318155501792987e-07, "logits/chosen": -7.37540340423584, "logits/rejected": -8.02101993560791, "logps/chosen": -1990.3394775390625, "logps/rejected": -2271.31201171875, "loss": 0.0943, "nll_loss": 5.673530578613281, "rewards/accuracies": 0.875, "rewards/chosen": -32.0615348815918, "rewards/margins": 6.438616752624512, "rewards/rejected": -38.500152587890625, "step": 316 }, { "epoch": 0.535578947368421, "grad_norm": 226.92107830972864, "learning_rate": 1.0202482925381357e-07, "logits/chosen": -6.763341426849365, "logits/rejected": -8.087518692016602, "logps/chosen": -1665.9189453125, "logps/rejected": -2236.61376953125, "loss": 0.2185, "nll_loss": 5.170358657836914, "rewards/accuracies": 0.875, "rewards/chosen": -26.559932708740234, "rewards/margins": 10.590608596801758, "rewards/rejected": -37.15054702758789, "step": 318 }, { "epoch": 0.5389473684210526, "grad_norm": 393.3203409753937, "learning_rate": 1.0086783238088243e-07, "logits/chosen": -7.136436939239502, "logits/rejected": -8.437828063964844, "logps/chosen": -1846.78369140625, "logps/rejected": -2221.16650390625, "loss": 0.2647, "nll_loss": 6.048235893249512, "rewards/accuracies": 0.875, "rewards/chosen": -30.3648681640625, "rewards/margins": 7.082037925720215, "rewards/rejected": -37.44690704345703, "step": 320 }, { "epoch": 0.5423157894736842, "grad_norm": 395.9473213067617, "learning_rate": 9.971071931197684e-08, "logits/chosen": -7.869472503662109, "logits/rejected": -8.018000602722168, "logps/chosen": -2073.983642578125, "logps/rejected": -2261.07470703125, "loss": 0.1895, "nll_loss": 5.238686561584473, "rewards/accuracies": 0.875, "rewards/chosen": -34.49671936035156, "rewards/margins": 3.941061019897461, "rewards/rejected": -38.437782287597656, "step": 322 }, { "epoch": 0.5456842105263158, "grad_norm": 407.67917839889736, "learning_rate": 9.855364497549495e-08, "logits/chosen": -7.1182451248168945, "logits/rejected": -8.209175109863281, "logps/chosen": -2025.939697265625, "logps/rejected": -2232.8544921875, "loss": 0.3687, "nll_loss": 5.539498329162598, "rewards/accuracies": 0.875, "rewards/chosen": -32.55401611328125, "rewards/margins": 4.3816399574279785, "rewards/rejected": -36.9356575012207, "step": 324 }, { "epoch": 0.5490526315789473, "grad_norm": 82.29409674682019, "learning_rate": 9.73967642946488e-08, "logits/chosen": -6.609898567199707, "logits/rejected": -8.180124282836914, "logps/chosen": -1846.933837890625, "logps/rejected": -2254.260009765625, "loss": 0.1058, "nll_loss": 6.1715264320373535, "rewards/accuracies": 1.0, "rewards/chosen": -31.290800094604492, "rewards/margins": 6.582897186279297, "rewards/rejected": -37.87369918823242, "step": 326 }, { "epoch": 0.5524210526315789, "grad_norm": 263.7142452068147, "learning_rate": 9.62402321667216e-08, "logits/chosen": -7.458375453948975, "logits/rejected": -8.259882926940918, "logps/chosen": -2037.841064453125, "logps/rejected": -2244.8076171875, "loss": 0.1896, "nll_loss": 5.539731502532959, "rewards/accuracies": 1.0, "rewards/chosen": -33.32776641845703, "rewards/margins": 5.299670219421387, "rewards/rejected": -38.62743377685547, "step": 328 }, { "epoch": 0.5557894736842105, "grad_norm": 326.97700368298075, "learning_rate": 9.508420344232799e-08, "logits/chosen": -6.740803241729736, "logits/rejected": -8.295230865478516, "logps/chosen": -1766.736083984375, "logps/rejected": -2225.2880859375, "loss": 0.3129, "nll_loss": 5.566542625427246, "rewards/accuracies": 0.75, "rewards/chosen": -29.658353805541992, "rewards/margins": 8.34937858581543, "rewards/rejected": -38.00773239135742, "step": 330 }, { "epoch": 0.5591578947368421, "grad_norm": 208.334466663584, "learning_rate": 9.392883290468082e-08, "logits/chosen": -6.8002543449401855, "logits/rejected": -8.274842262268066, "logps/chosen": -1685.3468017578125, "logps/rejected": -2213.642578125, "loss": 0.265, "nll_loss": 5.414514541625977, "rewards/accuracies": 0.875, "rewards/chosen": -27.97294044494629, "rewards/margins": 10.61333179473877, "rewards/rejected": -38.586273193359375, "step": 332 }, { "epoch": 0.5625263157894737, "grad_norm": 63.08646332985191, "learning_rate": 9.277427524886687e-08, "logits/chosen": -7.172906398773193, "logits/rejected": -8.222369194030762, "logps/chosen": -1881.0384521484375, "logps/rejected": -2289.237548828125, "loss": 0.1563, "nll_loss": 5.601876258850098, "rewards/accuracies": 0.875, "rewards/chosen": -30.56769561767578, "rewards/margins": 7.970741271972656, "rewards/rejected": -38.53843688964844, "step": 334 }, { "epoch": 0.5658947368421052, "grad_norm": 500.5421904045652, "learning_rate": 9.16206850611344e-08, "logits/chosen": -6.209731578826904, "logits/rejected": -8.183979034423828, "logps/chosen": -1466.2998046875, "logps/rejected": -2199.416015625, "loss": 0.2086, "nll_loss": 5.533670425415039, "rewards/accuracies": 0.875, "rewards/chosen": -23.00306510925293, "rewards/margins": 15.483078002929688, "rewards/rejected": -38.48614501953125, "step": 336 }, { "epoch": 0.5692631578947368, "grad_norm": 135.0392014842357, "learning_rate": 9.046821679819527e-08, "logits/chosen": -7.247140884399414, "logits/rejected": -8.511941909790039, "logps/chosen": -1807.933349609375, "logps/rejected": -2225.65185546875, "loss": 0.2196, "nll_loss": 5.2833781242370605, "rewards/accuracies": 0.875, "rewards/chosen": -30.05931282043457, "rewards/margins": 9.361750602722168, "rewards/rejected": -39.42106246948242, "step": 338 }, { "epoch": 0.5726315789473684, "grad_norm": 359.17470505312593, "learning_rate": 8.93170247665443e-08, "logits/chosen": -7.181203365325928, "logits/rejected": -7.774561882019043, "logps/chosen": -1932.3189697265625, "logps/rejected": -2207.670166015625, "loss": 0.1509, "nll_loss": 5.3413214683532715, "rewards/accuracies": 0.875, "rewards/chosen": -31.930519104003906, "rewards/margins": 7.081175804138184, "rewards/rejected": -39.011695861816406, "step": 340 }, { "epoch": 0.576, "grad_norm": 227.67264225472942, "learning_rate": 8.816726310179903e-08, "logits/chosen": -6.475002765655518, "logits/rejected": -8.516534805297852, "logps/chosen": -1540.990478515625, "logps/rejected": -2179.462158203125, "loss": 0.2961, "nll_loss": 5.707268714904785, "rewards/accuracies": 0.875, "rewards/chosen": -24.699859619140625, "rewards/margins": 14.47529411315918, "rewards/rejected": -39.17515563964844, "step": 342 }, { "epoch": 0.5793684210526315, "grad_norm": 97.40013916506999, "learning_rate": 8.701908574806198e-08, "logits/chosen": -7.5892767906188965, "logits/rejected": -8.08768081665039, "logps/chosen": -2175.58203125, "logps/rejected": -2273.545166015625, "loss": 0.2472, "nll_loss": 5.250769138336182, "rewards/accuracies": 0.875, "rewards/chosen": -35.854244232177734, "rewards/margins": 3.6291019916534424, "rewards/rejected": -39.48334884643555, "step": 344 }, { "epoch": 0.5827368421052631, "grad_norm": 248.47568888877237, "learning_rate": 8.587264643730875e-08, "logits/chosen": -7.8160400390625, "logits/rejected": -7.814428329467773, "logps/chosen": -2092.376708984375, "logps/rejected": -2212.744873046875, "loss": 0.2602, "nll_loss": 5.270852565765381, "rewards/accuracies": 1.0, "rewards/chosen": -35.40900802612305, "rewards/margins": 4.026213645935059, "rewards/rejected": -39.43522262573242, "step": 346 }, { "epoch": 0.5861052631578947, "grad_norm": 216.54285625194382, "learning_rate": 8.472809866880475e-08, "logits/chosen": -6.873068332672119, "logits/rejected": -8.404152870178223, "logps/chosen": -1594.337158203125, "logps/rejected": -2241.162109375, "loss": 0.167, "nll_loss": 5.596911907196045, "rewards/accuracies": 1.0, "rewards/chosen": -25.399681091308594, "rewards/margins": 13.986969947814941, "rewards/rejected": -39.38665771484375, "step": 348 }, { "epoch": 0.5894736842105263, "grad_norm": 394.3825078524655, "learning_rate": 8.358559568855248e-08, "logits/chosen": -7.783346652984619, "logits/rejected": -7.84064245223999, "logps/chosen": -2028.794189453125, "logps/rejected": -2198.99365234375, "loss": 0.2492, "nll_loss": 5.23142671585083, "rewards/accuracies": 0.875, "rewards/chosen": -33.149845123291016, "rewards/margins": 5.621738910675049, "rewards/rejected": -38.77158737182617, "step": 350 }, { "epoch": 0.592842105263158, "grad_norm": 476.30094871834973, "learning_rate": 8.244529046877334e-08, "logits/chosen": -6.29586935043335, "logits/rejected": -8.341531753540039, "logps/chosen": -1501.43212890625, "logps/rejected": -2209.18115234375, "loss": 0.336, "nll_loss": 5.435293197631836, "rewards/accuracies": 0.875, "rewards/chosen": -24.25966453552246, "rewards/margins": 15.036050796508789, "rewards/rejected": -39.29571533203125, "step": 352 }, { "epoch": 0.5962105263157895, "grad_norm": 478.94423279229494, "learning_rate": 8.130733568742579e-08, "logits/chosen": -8.36856460571289, "logits/rejected": -8.123823165893555, "logps/chosen": -2097.725830078125, "logps/rejected": -2244.956298828125, "loss": 0.2053, "nll_loss": 5.296268939971924, "rewards/accuracies": 0.875, "rewards/chosen": -33.903053283691406, "rewards/margins": 4.607615947723389, "rewards/rejected": -38.51067352294922, "step": 354 }, { "epoch": 0.5995789473684211, "grad_norm": 456.66250188186024, "learning_rate": 8.017188370776291e-08, "logits/chosen": -6.48959493637085, "logits/rejected": -7.693470478057861, "logps/chosen": -1521.6650390625, "logps/rejected": -2250.853271484375, "loss": 0.2331, "nll_loss": 5.437237739562988, "rewards/accuracies": 1.0, "rewards/chosen": -23.872224807739258, "rewards/margins": 15.529407501220703, "rewards/rejected": -39.401634216308594, "step": 356 }, { "epoch": 0.6029473684210527, "grad_norm": 220.8432613727437, "learning_rate": 7.903908655793223e-08, "logits/chosen": -6.895531177520752, "logits/rejected": -8.068331718444824, "logps/chosen": -1690.776611328125, "logps/rejected": -2218.949462890625, "loss": 0.1547, "nll_loss": 5.907761096954346, "rewards/accuracies": 0.875, "rewards/chosen": -28.69055938720703, "rewards/margins": 10.130281448364258, "rewards/rejected": -38.820838928222656, "step": 358 }, { "epoch": 0.6063157894736843, "grad_norm": 360.83010848251007, "learning_rate": 7.790909591062032e-08, "logits/chosen": -7.703709602355957, "logits/rejected": -8.051048278808594, "logps/chosen": -2004.1722412109375, "logps/rejected": -2261.4755859375, "loss": 0.2567, "nll_loss": 5.125810146331787, "rewards/accuracies": 0.875, "rewards/chosen": -32.635921478271484, "rewards/margins": 6.34180212020874, "rewards/rejected": -38.977725982666016, "step": 360 }, { "epoch": 0.6096842105263158, "grad_norm": 173.19431712494207, "learning_rate": 7.678206306274494e-08, "logits/chosen": -7.508633613586426, "logits/rejected": -8.247699737548828, "logps/chosen": -1943.2529296875, "logps/rejected": -2260.33447265625, "loss": 0.1663, "nll_loss": 5.228693008422852, "rewards/accuracies": 0.875, "rewards/chosen": -30.90469741821289, "rewards/margins": 7.72345495223999, "rewards/rejected": -38.628150939941406, "step": 362 }, { "epoch": 0.6130526315789474, "grad_norm": 387.0681464548588, "learning_rate": 7.565813891519764e-08, "logits/chosen": -7.763317108154297, "logits/rejected": -8.09384822845459, "logps/chosen": -1969.88134765625, "logps/rejected": -2286.67333984375, "loss": 0.2615, "nll_loss": 5.108150005340576, "rewards/accuracies": 0.875, "rewards/chosen": -31.575706481933594, "rewards/margins": 7.403677463531494, "rewards/rejected": -38.97938537597656, "step": 364 }, { "epoch": 0.616421052631579, "grad_norm": 267.106705620619, "learning_rate": 7.45374739526393e-08, "logits/chosen": -7.11317253112793, "logits/rejected": -8.084968566894531, "logps/chosen": -1716.8660888671875, "logps/rejected": -2228.20751953125, "loss": 0.311, "nll_loss": 4.8913702964782715, "rewards/accuracies": 1.0, "rewards/chosen": -27.476198196411133, "rewards/margins": 11.33434772491455, "rewards/rejected": -38.810546875, "step": 366 }, { "epoch": 0.6197894736842106, "grad_norm": 295.702658932922, "learning_rate": 7.342021822335143e-08, "logits/chosen": -6.858173847198486, "logits/rejected": -8.155499458312988, "logps/chosen": -1734.608642578125, "logps/rejected": -2249.30029296875, "loss": 0.2118, "nll_loss": 5.381309986114502, "rewards/accuracies": 1.0, "rewards/chosen": -27.76462745666504, "rewards/margins": 11.01102066040039, "rewards/rejected": -38.7756462097168, "step": 368 }, { "epoch": 0.6231578947368421, "grad_norm": 253.96226497503523, "learning_rate": 7.230652131914573e-08, "logits/chosen": -6.647583484649658, "logits/rejected": -8.132351875305176, "logps/chosen": -1633.9908447265625, "logps/rejected": -2257.0341796875, "loss": 0.1478, "nll_loss": 5.517997741699219, "rewards/accuracies": 0.875, "rewards/chosen": -26.471336364746094, "rewards/margins": 11.363603591918945, "rewards/rejected": -37.83494186401367, "step": 370 }, { "epoch": 0.6265263157894737, "grad_norm": 195.7616335942771, "learning_rate": 7.119653235533519e-08, "logits/chosen": -6.492335319519043, "logits/rejected": -8.236223220825195, "logps/chosen": -1685.419677734375, "logps/rejected": -2270.370361328125, "loss": 0.2562, "nll_loss": 5.618862628936768, "rewards/accuracies": 1.0, "rewards/chosen": -26.345735549926758, "rewards/margins": 10.770453453063965, "rewards/rejected": -37.116188049316406, "step": 372 }, { "epoch": 0.6298947368421053, "grad_norm": 157.13258856973306, "learning_rate": 7.009039995076844e-08, "logits/chosen": -7.430450916290283, "logits/rejected": -8.231632232666016, "logps/chosen": -2042.21044921875, "logps/rejected": -2247.43798828125, "loss": 0.177, "nll_loss": 5.24857234954834, "rewards/accuracies": 0.875, "rewards/chosen": -32.96132278442383, "rewards/margins": 5.063416004180908, "rewards/rejected": -38.02473831176758, "step": 374 }, { "epoch": 0.6332631578947369, "grad_norm": 306.3698872482412, "learning_rate": 6.898827220793101e-08, "logits/chosen": -7.346684455871582, "logits/rejected": -8.37730884552002, "logps/chosen": -1900.4794921875, "logps/rejected": -2195.65380859375, "loss": 0.2348, "nll_loss": 5.606869697570801, "rewards/accuracies": 0.75, "rewards/chosen": -30.2860107421875, "rewards/margins": 6.685296535491943, "rewards/rejected": -36.97130584716797, "step": 376 }, { "epoch": 0.6366315789473684, "grad_norm": 404.1225835065395, "learning_rate": 6.78902966931155e-08, "logits/chosen": -6.475454807281494, "logits/rejected": -8.423490524291992, "logps/chosen": -1562.2786865234375, "logps/rejected": -2233.23779296875, "loss": 0.2088, "nll_loss": 5.756392478942871, "rewards/accuracies": 1.0, "rewards/chosen": -24.448637008666992, "rewards/margins": 13.247220993041992, "rewards/rejected": -37.69586181640625, "step": 378 }, { "epoch": 0.64, "grad_norm": 154.48168757776762, "learning_rate": 6.67966204166636e-08, "logits/chosen": -7.25782585144043, "logits/rejected": -8.18947982788086, "logps/chosen": -2031.32568359375, "logps/rejected": -2208.31298828125, "loss": 0.0898, "nll_loss": 5.3062825202941895, "rewards/accuracies": 1.0, "rewards/chosen": -30.2602596282959, "rewards/margins": 6.670921325683594, "rewards/rejected": -36.93117904663086, "step": 380 }, { "epoch": 0.6433684210526316, "grad_norm": 118.13213437450032, "learning_rate": 6.570738981328266e-08, "logits/chosen": -6.853590488433838, "logits/rejected": -8.224015235900879, "logps/chosen": -1710.7021484375, "logps/rejected": -2247.219482421875, "loss": 0.092, "nll_loss": 5.132368564605713, "rewards/accuracies": 1.0, "rewards/chosen": -26.538768768310547, "rewards/margins": 10.044065475463867, "rewards/rejected": -36.58283615112305, "step": 382 }, { "epoch": 0.6467368421052632, "grad_norm": 173.64933572252733, "learning_rate": 6.462275072243906e-08, "logits/chosen": -6.30927848815918, "logits/rejected": -8.526763916015625, "logps/chosen": -1525.7376708984375, "logps/rejected": -2231.30859375, "loss": 0.1534, "nll_loss": 5.520993709564209, "rewards/accuracies": 0.875, "rewards/chosen": -23.9798641204834, "rewards/margins": 13.774813652038574, "rewards/rejected": -37.75468063354492, "step": 384 }, { "epoch": 0.6501052631578947, "grad_norm": 210.67565229221435, "learning_rate": 6.354284836883156e-08, "logits/chosen": -7.883388042449951, "logits/rejected": -7.760946273803711, "logps/chosen": -2060.9970703125, "logps/rejected": -2203.468505859375, "loss": 0.1437, "nll_loss": 4.833434104919434, "rewards/accuracies": 0.875, "rewards/chosen": -32.70746612548828, "rewards/margins": 4.838762283325195, "rewards/rejected": -37.54623031616211, "step": 386 }, { "epoch": 0.6534736842105263, "grad_norm": 194.1019283617502, "learning_rate": 6.246782734294683e-08, "logits/chosen": -6.361388206481934, "logits/rejected": -8.105430603027344, "logps/chosen": -1549.34228515625, "logps/rejected": -2235.9560546875, "loss": 0.0999, "nll_loss": 5.782830238342285, "rewards/accuracies": 1.0, "rewards/chosen": -23.41505241394043, "rewards/margins": 15.236028671264648, "rewards/rejected": -38.651084899902344, "step": 388 }, { "epoch": 0.6568421052631579, "grad_norm": 151.3635743520382, "learning_rate": 6.139783158169984e-08, "logits/chosen": -7.068288803100586, "logits/rejected": -7.906414985656738, "logps/chosen": -1757.916259765625, "logps/rejected": -2189.158203125, "loss": 0.2537, "nll_loss": 5.233246326446533, "rewards/accuracies": 1.0, "rewards/chosen": -26.988143920898438, "rewards/margins": 10.64902114868164, "rewards/rejected": -37.63716506958008, "step": 390 }, { "epoch": 0.6602105263157895, "grad_norm": 202.61630533473854, "learning_rate": 6.033300434916202e-08, "logits/chosen": -6.130324363708496, "logits/rejected": -8.429763793945312, "logps/chosen": -1517.0257568359375, "logps/rejected": -2255.234619140625, "loss": 0.0935, "nll_loss": 5.869035243988037, "rewards/accuracies": 1.0, "rewards/chosen": -24.74703598022461, "rewards/margins": 14.127795219421387, "rewards/rejected": -38.87482833862305, "step": 392 }, { "epoch": 0.663578947368421, "grad_norm": 51.935332949676464, "learning_rate": 5.9273488217379056e-08, "logits/chosen": -6.12472677230835, "logits/rejected": -8.31787109375, "logps/chosen": -1496.01025390625, "logps/rejected": -2250.830078125, "loss": 0.0405, "nll_loss": 5.127452373504639, "rewards/accuracies": 1.0, "rewards/chosen": -23.69045066833496, "rewards/margins": 14.94088363647461, "rewards/rejected": -38.6313362121582, "step": 394 }, { "epoch": 0.6669473684210526, "grad_norm": 157.62065071319176, "learning_rate": 5.821942504728182e-08, "logits/chosen": -5.9207844734191895, "logits/rejected": -8.412075996398926, "logps/chosen": -1410.460205078125, "logps/rejected": -2238.791259765625, "loss": 0.2002, "nll_loss": 5.9735589027404785, "rewards/accuracies": 1.0, "rewards/chosen": -21.796749114990234, "rewards/margins": 17.19828224182129, "rewards/rejected": -38.99502944946289, "step": 396 }, { "epoch": 0.6703157894736842, "grad_norm": 409.7869110891659, "learning_rate": 5.7170955969692257e-08, "logits/chosen": -7.128916263580322, "logits/rejected": -8.091632843017578, "logps/chosen": -1968.1507568359375, "logps/rejected": -2279.8564453125, "loss": 0.2463, "nll_loss": 5.298381328582764, "rewards/accuracies": 0.875, "rewards/chosen": -31.842985153198242, "rewards/margins": 7.260722637176514, "rewards/rejected": -39.10371017456055, "step": 398 }, { "epoch": 0.6736842105263158, "grad_norm": 196.72465489549091, "learning_rate": 5.612822136642696e-08, "logits/chosen": -7.994609832763672, "logits/rejected": -8.255779266357422, "logps/chosen": -2162.84375, "logps/rejected": -2256.470458984375, "loss": 0.1349, "nll_loss": 5.209514617919922, "rewards/accuracies": 0.875, "rewards/chosen": -35.79097366333008, "rewards/margins": 3.0046000480651855, "rewards/rejected": -38.795570373535156, "step": 400 }, { "epoch": 0.6770526315789474, "grad_norm": 276.3858546036682, "learning_rate": 5.5091360851501214e-08, "logits/chosen": -7.058955192565918, "logits/rejected": -8.353060722351074, "logps/chosen": -1752.0609130859375, "logps/rejected": -2249.98486328125, "loss": 0.1673, "nll_loss": 5.86051607131958, "rewards/accuracies": 0.75, "rewards/chosen": -29.15858268737793, "rewards/margins": 8.361825942993164, "rewards/rejected": -37.520408630371094, "step": 402 }, { "epoch": 0.6804210526315789, "grad_norm": 283.88427439715645, "learning_rate": 5.406051325243586e-08, "logits/chosen": -6.989212989807129, "logits/rejected": -7.9651103019714355, "logps/chosen": -1695.2578125, "logps/rejected": -2204.557373046875, "loss": 0.173, "nll_loss": 5.183251857757568, "rewards/accuracies": 0.875, "rewards/chosen": -27.098966598510742, "rewards/margins": 11.072956085205078, "rewards/rejected": -38.17192077636719, "step": 404 }, { "epoch": 0.6837894736842105, "grad_norm": 262.0694683666909, "learning_rate": 5.30358165916692e-08, "logits/chosen": -7.981943130493164, "logits/rejected": -8.023259162902832, "logps/chosen": -2143.8271484375, "logps/rejected": -2240.49609375, "loss": 0.1934, "nll_loss": 5.243900775909424, "rewards/accuracies": 1.0, "rewards/chosen": -34.75602722167969, "rewards/margins": 2.599457263946533, "rewards/rejected": -37.3554801940918, "step": 406 }, { "epoch": 0.6871578947368421, "grad_norm": 375.4614868731545, "learning_rate": 5.201740806807706e-08, "logits/chosen": -6.7205047607421875, "logits/rejected": -8.098003387451172, "logps/chosen": -1731.788818359375, "logps/rejected": -2216.237060546875, "loss": 0.2756, "nll_loss": 5.406951904296875, "rewards/accuracies": 0.75, "rewards/chosen": -27.372705459594727, "rewards/margins": 11.000872611999512, "rewards/rejected": -38.37357711791992, "step": 408 }, { "epoch": 0.6905263157894737, "grad_norm": 210.62564910879206, "learning_rate": 5.100542403860272e-08, "logits/chosen": -7.43436861038208, "logits/rejected": -7.938074111938477, "logps/chosen": -2066.572265625, "logps/rejected": -2238.23876953125, "loss": 0.2334, "nll_loss": 5.189877033233643, "rewards/accuracies": 0.875, "rewards/chosen": -34.191410064697266, "rewards/margins": 4.024079322814941, "rewards/rejected": -38.215492248535156, "step": 410 }, { "epoch": 0.6938947368421052, "grad_norm": 136.5153093521016, "learning_rate": 5.000000000000002e-08, "logits/chosen": -5.983953475952148, "logits/rejected": -8.266366004943848, "logps/chosen": -1458.96875, "logps/rejected": -2244.586181640625, "loss": 0.1569, "nll_loss": 6.0212860107421875, "rewards/accuracies": 1.0, "rewards/chosen": -23.258224487304688, "rewards/margins": 14.572385787963867, "rewards/rejected": -37.83060836791992, "step": 412 }, { "epoch": 0.6972631578947368, "grad_norm": 403.7847701746563, "learning_rate": 4.900127057069116e-08, "logits/chosen": -6.388440132141113, "logits/rejected": -8.440248489379883, "logps/chosen": -1643.813720703125, "logps/rejected": -2231.506591796875, "loss": 0.2028, "nll_loss": 5.681787490844727, "rewards/accuracies": 0.875, "rewards/chosen": -26.718189239501953, "rewards/margins": 11.000421524047852, "rewards/rejected": -37.71861267089844, "step": 414 }, { "epoch": 0.7006315789473684, "grad_norm": 384.7121082705872, "learning_rate": 4.8009369472742546e-08, "logits/chosen": -6.855067729949951, "logits/rejected": -8.357769966125488, "logps/chosen": -1806.01318359375, "logps/rejected": -2285.265869140625, "loss": 0.2982, "nll_loss": 5.6631622314453125, "rewards/accuracies": 1.0, "rewards/chosen": -29.00935935974121, "rewards/margins": 9.206945419311523, "rewards/rejected": -38.216304779052734, "step": 416 }, { "epoch": 0.704, "grad_norm": 205.84300753243963, "learning_rate": 4.7024429513960414e-08, "logits/chosen": -6.546984672546387, "logits/rejected": -8.477337837219238, "logps/chosen": -1774.5728759765625, "logps/rejected": -2264.49169921875, "loss": 0.1813, "nll_loss": 5.92774772644043, "rewards/accuracies": 1.0, "rewards/chosen": -28.638093948364258, "rewards/margins": 9.986715316772461, "rewards/rejected": -38.62480926513672, "step": 418 }, { "epoch": 0.7073684210526315, "grad_norm": 143.96793262737364, "learning_rate": 4.604658257010874e-08, "logits/chosen": -6.683577060699463, "logits/rejected": -8.190496444702148, "logps/chosen": -1780.205322265625, "logps/rejected": -2335.033203125, "loss": 0.0523, "nll_loss": 5.4027886390686035, "rewards/accuracies": 1.0, "rewards/chosen": -26.897085189819336, "rewards/margins": 11.124114036560059, "rewards/rejected": -38.02119827270508, "step": 420 }, { "epoch": 0.7107368421052631, "grad_norm": 168.66393059683804, "learning_rate": 4.507595956725233e-08, "logits/chosen": -6.165244102478027, "logits/rejected": -8.504766464233398, "logps/chosen": -1527.099609375, "logps/rejected": -2197.563720703125, "loss": 0.2034, "nll_loss": 6.075287818908691, "rewards/accuracies": 1.0, "rewards/chosen": -24.494510650634766, "rewards/margins": 13.510110855102539, "rewards/rejected": -38.00461959838867, "step": 422 }, { "epoch": 0.7141052631578947, "grad_norm": 412.195910642015, "learning_rate": 4.4112690464226697e-08, "logits/chosen": -6.30023717880249, "logits/rejected": -8.322918891906738, "logps/chosen": -1680.518310546875, "logps/rejected": -2284.201416015625, "loss": 0.274, "nll_loss": 5.803287982940674, "rewards/accuracies": 0.75, "rewards/chosen": -26.578989028930664, "rewards/margins": 11.078222274780273, "rewards/rejected": -37.65721130371094, "step": 424 }, { "epoch": 0.7174736842105263, "grad_norm": 292.02246007370735, "learning_rate": 4.315690423523757e-08, "logits/chosen": -6.849490165710449, "logits/rejected": -8.255549430847168, "logps/chosen": -1616.78857421875, "logps/rejected": -2214.650390625, "loss": 0.2553, "nll_loss": 5.239808559417725, "rewards/accuracies": 0.75, "rewards/chosen": -24.917198181152344, "rewards/margins": 12.881205558776855, "rewards/rejected": -37.79840087890625, "step": 426 }, { "epoch": 0.7208421052631578, "grad_norm": 256.5445563152725, "learning_rate": 4.2208728852592465e-08, "logits/chosen": -6.686156272888184, "logits/rejected": -8.05382251739502, "logps/chosen": -1576.5635986328125, "logps/rejected": -2256.39697265625, "loss": 0.3019, "nll_loss": 5.413787841796875, "rewards/accuracies": 0.875, "rewards/chosen": -25.76823616027832, "rewards/margins": 12.707958221435547, "rewards/rejected": -38.476192474365234, "step": 428 }, { "epoch": 0.7242105263157895, "grad_norm": 244.67144743001788, "learning_rate": 4.126829126956588e-08, "logits/chosen": -6.939605236053467, "logits/rejected": -8.691452026367188, "logps/chosen": -1630.356689453125, "logps/rejected": -2211.13671875, "loss": 0.252, "nll_loss": 5.6548848152160645, "rewards/accuracies": 0.875, "rewards/chosen": -25.576709747314453, "rewards/margins": 13.35787582397461, "rewards/rejected": -38.93458938598633, "step": 430 }, { "epoch": 0.7275789473684211, "grad_norm": 141.2458931925102, "learning_rate": 4.033571740340157e-08, "logits/chosen": -6.9448723793029785, "logits/rejected": -8.11502456665039, "logps/chosen": -1771.171875, "logps/rejected": -2253.62353515625, "loss": 0.0496, "nll_loss": 5.7086663246154785, "rewards/accuracies": 1.0, "rewards/chosen": -27.938629150390625, "rewards/margins": 10.429763793945312, "rewards/rejected": -38.36838912963867, "step": 432 }, { "epoch": 0.7309473684210527, "grad_norm": 70.65354096973839, "learning_rate": 3.9411132118452896e-08, "logits/chosen": -7.186574935913086, "logits/rejected": -8.068167686462402, "logps/chosen": -1631.0499267578125, "logps/rejected": -2222.7802734375, "loss": 0.0387, "nll_loss": 4.864493370056152, "rewards/accuracies": 1.0, "rewards/chosen": -25.45480728149414, "rewards/margins": 13.78203010559082, "rewards/rejected": -39.236839294433594, "step": 434 }, { "epoch": 0.7343157894736843, "grad_norm": 206.48565136287314, "learning_rate": 3.849465920946475e-08, "logits/chosen": -6.413349151611328, "logits/rejected": -8.138294219970703, "logps/chosen": -1650.3955078125, "logps/rejected": -2234.84912109375, "loss": 0.1177, "nll_loss": 5.74521541595459, "rewards/accuracies": 0.875, "rewards/chosen": -26.769577026367188, "rewards/margins": 10.97209358215332, "rewards/rejected": -37.74167251586914, "step": 436 }, { "epoch": 0.7376842105263158, "grad_norm": 204.9937975433647, "learning_rate": 3.758642138499819e-08, "logits/chosen": -6.9706597328186035, "logits/rejected": -8.29416275024414, "logps/chosen": -1775.1536865234375, "logps/rejected": -2233.42138671875, "loss": 0.1547, "nll_loss": 5.631414890289307, "rewards/accuracies": 0.75, "rewards/chosen": -28.840255737304688, "rewards/margins": 9.066543579101562, "rewards/rejected": -37.90679931640625, "step": 438 }, { "epoch": 0.7410526315789474, "grad_norm": 277.05848886110675, "learning_rate": 3.668654025100075e-08, "logits/chosen": -7.38844108581543, "logits/rejected": -8.166891098022461, "logps/chosen": -1834.601318359375, "logps/rejected": -2221.83984375, "loss": 0.151, "nll_loss": 5.479992389678955, "rewards/accuracies": 1.0, "rewards/chosen": -28.740129470825195, "rewards/margins": 10.051774978637695, "rewards/rejected": -38.791900634765625, "step": 440 }, { "epoch": 0.744421052631579, "grad_norm": 67.30116112714921, "learning_rate": 3.579513629452464e-08, "logits/chosen": -6.464494228363037, "logits/rejected": -8.139457702636719, "logps/chosen": -1527.421630859375, "logps/rejected": -2210.33740234375, "loss": 0.0763, "nll_loss": 5.879497528076172, "rewards/accuracies": 1.0, "rewards/chosen": -24.09062385559082, "rewards/margins": 14.153335571289062, "rewards/rejected": -38.243961334228516, "step": 442 }, { "epoch": 0.7477894736842106, "grad_norm": 198.27913956717686, "learning_rate": 3.491232886759398e-08, "logits/chosen": -7.406864166259766, "logits/rejected": -8.350690841674805, "logps/chosen": -1810.20166015625, "logps/rejected": -2223.15478515625, "loss": 0.1604, "nll_loss": 5.455080032348633, "rewards/accuracies": 0.875, "rewards/chosen": -30.108121871948242, "rewards/margins": 7.683882236480713, "rewards/rejected": -37.7920036315918, "step": 444 }, { "epoch": 0.7511578947368421, "grad_norm": 57.27184908194972, "learning_rate": 3.4038236171224943e-08, "logits/chosen": -6.585896968841553, "logits/rejected": -8.207939147949219, "logps/chosen": -1619.34423828125, "logps/rejected": -2219.935546875, "loss": 0.1412, "nll_loss": 5.543335914611816, "rewards/accuracies": 1.0, "rewards/chosen": -24.96392250061035, "rewards/margins": 12.792692184448242, "rewards/rejected": -37.756614685058594, "step": 446 }, { "epoch": 0.7545263157894737, "grad_norm": 162.97400818458752, "learning_rate": 3.317297523959927e-08, "logits/chosen": -7.09434175491333, "logits/rejected": -8.217880249023438, "logps/chosen": -1831.752685546875, "logps/rejected": -2260.440185546875, "loss": 0.1862, "nll_loss": 5.406676292419434, "rewards/accuracies": 1.0, "rewards/chosen": -30.033018112182617, "rewards/margins": 9.028882026672363, "rewards/rejected": -39.0619010925293, "step": 448 }, { "epoch": 0.7578947368421053, "grad_norm": 154.08278897844173, "learning_rate": 3.231666192439442e-08, "logits/chosen": -7.439211368560791, "logits/rejected": -8.138461112976074, "logps/chosen": -1908.0830078125, "logps/rejected": -2208.218505859375, "loss": 0.1308, "nll_loss": 6.031886100769043, "rewards/accuracies": 0.875, "rewards/chosen": -31.519283294677734, "rewards/margins": 8.400613784790039, "rewards/rejected": -39.919898986816406, "step": 450 }, { "epoch": 0.7612631578947369, "grad_norm": 73.20681520848144, "learning_rate": 3.146941087927203e-08, "logits/chosen": -7.356697082519531, "logits/rejected": -7.933921813964844, "logps/chosen": -1768.478271484375, "logps/rejected": -2268.8671875, "loss": 0.0731, "nll_loss": 5.609357833862305, "rewards/accuracies": 1.0, "rewards/chosen": -27.863664627075195, "rewards/margins": 11.745465278625488, "rewards/rejected": -39.609130859375, "step": 452 }, { "epoch": 0.7646315789473684, "grad_norm": 304.7554603694575, "learning_rate": 3.063133554452645e-08, "logits/chosen": -7.254971027374268, "logits/rejected": -8.151304244995117, "logps/chosen": -1963.06396484375, "logps/rejected": -2218.695068359375, "loss": 0.2756, "nll_loss": 5.318238258361816, "rewards/accuracies": 0.875, "rewards/chosen": -32.04559326171875, "rewards/margins": 6.272039890289307, "rewards/rejected": -38.31763458251953, "step": 454 }, { "epoch": 0.768, "grad_norm": 196.76512807151576, "learning_rate": 2.980254813189623e-08, "logits/chosen": -7.754366397857666, "logits/rejected": -8.369587898254395, "logps/chosen": -1897.6082763671875, "logps/rejected": -2181.4443359375, "loss": 0.0918, "nll_loss": 5.094344139099121, "rewards/accuracies": 0.75, "rewards/chosen": -30.451175689697266, "rewards/margins": 7.756158351898193, "rewards/rejected": -38.207332611083984, "step": 456 }, { "epoch": 0.7713684210526316, "grad_norm": 232.4816020946604, "learning_rate": 2.8983159609539633e-08, "logits/chosen": -8.158332824707031, "logits/rejected": -7.675507545471191, "logps/chosen": -2173.97265625, "logps/rejected": -2187.316650390625, "loss": 0.1702, "nll_loss": 4.910658359527588, "rewards/accuracies": 1.0, "rewards/chosen": -35.214569091796875, "rewards/margins": 4.455814838409424, "rewards/rejected": -39.67039108276367, "step": 458 }, { "epoch": 0.7747368421052632, "grad_norm": 185.23537602534242, "learning_rate": 2.8173279687177054e-08, "logits/chosen": -6.689702987670898, "logits/rejected": -8.015731811523438, "logps/chosen": -1809.507568359375, "logps/rejected": -2290.198486328125, "loss": 0.079, "nll_loss": 5.7478179931640625, "rewards/accuracies": 0.875, "rewards/chosen": -29.168781280517578, "rewards/margins": 10.092117309570312, "rewards/rejected": -39.26089859008789, "step": 460 }, { "epoch": 0.7781052631578947, "grad_norm": 228.5238209000322, "learning_rate": 2.7373016801401573e-08, "logits/chosen": -7.5323591232299805, "logits/rejected": -8.179119110107422, "logps/chosen": -2089.93896484375, "logps/rejected": -2171.359130859375, "loss": 0.2045, "nll_loss": 5.747359275817871, "rewards/accuracies": 0.75, "rewards/chosen": -34.591739654541016, "rewards/margins": 3.1891961097717285, "rewards/rejected": -37.78093719482422, "step": 462 }, { "epoch": 0.7814736842105263, "grad_norm": 176.4295959972419, "learning_rate": 2.6582478101160165e-08, "logits/chosen": -6.586089134216309, "logits/rejected": -8.126249313354492, "logps/chosen": -1404.34130859375, "logps/rejected": -2220.29345703125, "loss": 0.0916, "nll_loss": 5.6755828857421875, "rewards/accuracies": 1.0, "rewards/chosen": -21.904067993164062, "rewards/margins": 16.568695068359375, "rewards/rejected": -38.4727668762207, "step": 464 }, { "epoch": 0.7848421052631579, "grad_norm": 109.08978437364105, "learning_rate": 2.580176943340756e-08, "logits/chosen": -6.541893005371094, "logits/rejected": -8.444816589355469, "logps/chosen": -1707.611328125, "logps/rejected": -2249.0712890625, "loss": 0.1225, "nll_loss": 5.4323201179504395, "rewards/accuracies": 1.0, "rewards/chosen": -27.17432403564453, "rewards/margins": 12.210725784301758, "rewards/rejected": -39.38505172729492, "step": 466 }, { "epoch": 0.7882105263157895, "grad_norm": 266.4174356156143, "learning_rate": 2.5030995328933723e-08, "logits/chosen": -6.337734222412109, "logits/rejected": -8.216695785522461, "logps/chosen": -1433.7144775390625, "logps/rejected": -2253.561279296875, "loss": 0.0827, "nll_loss": 5.440419673919678, "rewards/accuracies": 1.0, "rewards/chosen": -21.820663452148438, "rewards/margins": 16.79531478881836, "rewards/rejected": -38.61597442626953, "step": 468 }, { "epoch": 0.791578947368421, "grad_norm": 444.435808075812, "learning_rate": 2.4270258988368376e-08, "logits/chosen": -6.913492202758789, "logits/rejected": -8.389839172363281, "logps/chosen": -1789.693603515625, "logps/rejected": -2235.32421875, "loss": 0.4713, "nll_loss": 5.414417743682861, "rewards/accuracies": 0.875, "rewards/chosen": -29.466032028198242, "rewards/margins": 8.25886344909668, "rewards/rejected": -37.72489547729492, "step": 470 }, { "epoch": 0.7949473684210526, "grad_norm": 227.88051110998938, "learning_rate": 2.3519662268363006e-08, "logits/chosen": -7.451695442199707, "logits/rejected": -8.134040832519531, "logps/chosen": -1984.7669677734375, "logps/rejected": -2170.09326171875, "loss": 0.3332, "nll_loss": 5.461165904998779, "rewards/accuracies": 0.75, "rewards/chosen": -32.05513381958008, "rewards/margins": 4.828782081604004, "rewards/rejected": -36.883914947509766, "step": 472 }, { "epoch": 0.7983157894736842, "grad_norm": 161.35000532615425, "learning_rate": 2.2779305667953152e-08, "logits/chosen": -7.434182167053223, "logits/rejected": -7.826815605163574, "logps/chosen": -1815.1357421875, "logps/rejected": -2225.9404296875, "loss": 0.1478, "nll_loss": 4.895688533782959, "rewards/accuracies": 1.0, "rewards/chosen": -28.85767936706543, "rewards/margins": 8.826580047607422, "rewards/rejected": -37.684261322021484, "step": 474 }, { "epoch": 0.8016842105263158, "grad_norm": 252.39419967830085, "learning_rate": 2.204928831510241e-08, "logits/chosen": -7.645818710327148, "logits/rejected": -8.229476928710938, "logps/chosen": -1907.0130615234375, "logps/rejected": -2305.749755859375, "loss": 0.1393, "nll_loss": 5.33302116394043, "rewards/accuracies": 1.0, "rewards/chosen": -30.992538452148438, "rewards/margins": 9.087993621826172, "rewards/rejected": -40.080535888671875, "step": 476 }, { "epoch": 0.8050526315789474, "grad_norm": 519.421493105551, "learning_rate": 2.132970795342982e-08, "logits/chosen": -7.341223239898682, "logits/rejected": -8.555485725402832, "logps/chosen": -1928.36865234375, "logps/rejected": -2207.854736328125, "loss": 0.3261, "nll_loss": 5.729315280914307, "rewards/accuracies": 0.625, "rewards/chosen": -32.422027587890625, "rewards/margins": 6.0826640129089355, "rewards/rejected": -38.504695892333984, "step": 478 }, { "epoch": 0.8084210526315789, "grad_norm": 291.74584937064657, "learning_rate": 2.0620660929123e-08, "logits/chosen": -6.5453057289123535, "logits/rejected": -7.966734886169434, "logps/chosen": -1665.8740234375, "logps/rejected": -2183.277099609375, "loss": 0.1242, "nll_loss": 5.9209136962890625, "rewards/accuracies": 0.875, "rewards/chosen": -27.65888786315918, "rewards/margins": 10.57414436340332, "rewards/rejected": -38.2330322265625, "step": 480 }, { "epoch": 0.8117894736842105, "grad_norm": 324.06037246167574, "learning_rate": 1.992224217803786e-08, "logits/chosen": -6.169129371643066, "logits/rejected": -7.988075256347656, "logps/chosen": -1408.3675537109375, "logps/rejected": -2238.338134765625, "loss": 0.1118, "nll_loss": 5.187215805053711, "rewards/accuracies": 1.0, "rewards/chosen": -22.458417892456055, "rewards/margins": 16.477596282958984, "rewards/rejected": -38.93601608276367, "step": 482 }, { "epoch": 0.8151578947368421, "grad_norm": 394.80006807243586, "learning_rate": 1.9234545212987686e-08, "logits/chosen": -7.038122653961182, "logits/rejected": -8.259831428527832, "logps/chosen": -1790.6435546875, "logps/rejected": -2252.951416015625, "loss": 0.237, "nll_loss": 5.162839412689209, "rewards/accuracies": 0.875, "rewards/chosen": -29.044601440429688, "rewards/margins": 9.581314086914062, "rewards/rejected": -38.62591552734375, "step": 484 }, { "epoch": 0.8185263157894737, "grad_norm": 245.72106899897696, "learning_rate": 1.855766211122234e-08, "logits/chosen": -6.452915191650391, "logits/rejected": -8.34599494934082, "logps/chosen": -1615.44775390625, "logps/rejected": -2228.80810546875, "loss": 0.1125, "nll_loss": 5.462100028991699, "rewards/accuracies": 1.0, "rewards/chosen": -25.96318244934082, "rewards/margins": 12.65347671508789, "rewards/rejected": -38.61665725708008, "step": 486 }, { "epoch": 0.8218947368421052, "grad_norm": 265.58626188114476, "learning_rate": 1.789168350209983e-08, "logits/chosen": -7.410139083862305, "logits/rejected": -8.122781753540039, "logps/chosen": -2041.22216796875, "logps/rejected": -2253.333740234375, "loss": 0.1616, "nll_loss": 5.257528781890869, "rewards/accuracies": 1.0, "rewards/chosen": -32.81037521362305, "rewards/margins": 5.845396518707275, "rewards/rejected": -38.65576934814453, "step": 488 }, { "epoch": 0.8252631578947368, "grad_norm": 195.8964969058916, "learning_rate": 1.723669855495199e-08, "logits/chosen": -7.151376247406006, "logits/rejected": -7.946229934692383, "logps/chosen": -1715.78955078125, "logps/rejected": -2244.12841796875, "loss": 0.1499, "nll_loss": 5.343441963195801, "rewards/accuracies": 0.875, "rewards/chosen": -27.61037826538086, "rewards/margins": 11.252402305603027, "rewards/rejected": -38.8627815246582, "step": 490 }, { "epoch": 0.8286315789473684, "grad_norm": 186.530933259071, "learning_rate": 1.659279496714503e-08, "logits/chosen": -6.534136772155762, "logits/rejected": -8.039037704467773, "logps/chosen": -1630.31298828125, "logps/rejected": -2252.484130859375, "loss": 0.097, "nll_loss": 5.664884567260742, "rewards/accuracies": 1.0, "rewards/chosen": -26.512128829956055, "rewards/margins": 13.004049301147461, "rewards/rejected": -39.51618194580078, "step": 492 }, { "epoch": 0.832, "grad_norm": 335.92762194779397, "learning_rate": 1.5960058952337884e-08, "logits/chosen": -6.526432514190674, "logits/rejected": -8.54305648803711, "logps/chosen": -1744.421875, "logps/rejected": -2280.98486328125, "loss": 0.2653, "nll_loss": 5.502665996551514, "rewards/accuracies": 0.875, "rewards/chosen": -29.049074172973633, "rewards/margins": 9.935380935668945, "rewards/rejected": -38.98445129394531, "step": 494 }, { "epoch": 0.8353684210526315, "grad_norm": 218.92976542147028, "learning_rate": 1.5338575228938612e-08, "logits/chosen": -6.907125949859619, "logits/rejected": -8.139141082763672, "logps/chosen": -1731.1072998046875, "logps/rejected": -2325.4013671875, "loss": 0.2407, "nll_loss": 4.939925670623779, "rewards/accuracies": 0.875, "rewards/chosen": -27.770357131958008, "rewards/margins": 11.083059310913086, "rewards/rejected": -38.85342025756836, "step": 496 }, { "epoch": 0.8387368421052631, "grad_norm": 159.02247163333553, "learning_rate": 1.4728427008761401e-08, "logits/chosen": -7.341606140136719, "logits/rejected": -8.316031455993652, "logps/chosen": -1968.184326171875, "logps/rejected": -2243.793701171875, "loss": 0.1247, "nll_loss": 5.373488426208496, "rewards/accuracies": 1.0, "rewards/chosen": -32.148460388183594, "rewards/margins": 6.647695541381836, "rewards/rejected": -38.7961540222168, "step": 498 }, { "epoch": 0.8421052631578947, "grad_norm": 324.2851031499905, "learning_rate": 1.4129695985885225e-08, "logits/chosen": -7.782922267913818, "logits/rejected": -7.933620452880859, "logps/chosen": -2051.697265625, "logps/rejected": -2286.102783203125, "loss": 0.2646, "nll_loss": 5.484002590179443, "rewards/accuracies": 0.875, "rewards/chosen": -32.9913215637207, "rewards/margins": 6.650774955749512, "rewards/rejected": -39.642093658447266, "step": 500 }, { "epoch": 0.8454736842105263, "grad_norm": 383.21308122896846, "learning_rate": 1.3542462325715442e-08, "logits/chosen": -7.545998573303223, "logits/rejected": -8.588085174560547, "logps/chosen": -1862.868896484375, "logps/rejected": -2210.8505859375, "loss": 0.2632, "nll_loss": 5.365788459777832, "rewards/accuracies": 0.75, "rewards/chosen": -31.22475814819336, "rewards/margins": 7.746358871459961, "rewards/rejected": -38.97111511230469, "step": 502 }, { "epoch": 0.8488421052631578, "grad_norm": 252.10082928265032, "learning_rate": 1.2966804654250462e-08, "logits/chosen": -6.983362674713135, "logits/rejected": -8.178686141967773, "logps/chosen": -1801.875244140625, "logps/rejected": -2263.5986328125, "loss": 0.1475, "nll_loss": 6.0845537185668945, "rewards/accuracies": 1.0, "rewards/chosen": -30.309919357299805, "rewards/margins": 8.438333511352539, "rewards/rejected": -38.748252868652344, "step": 504 }, { "epoch": 0.8522105263157894, "grad_norm": 100.48131239501701, "learning_rate": 1.2402800047554207e-08, "logits/chosen": -6.94020938873291, "logits/rejected": -8.300201416015625, "logps/chosen": -1838.616943359375, "logps/rejected": -2285.8544921875, "loss": 0.2172, "nll_loss": 5.775844573974609, "rewards/accuracies": 1.0, "rewards/chosen": -30.875246047973633, "rewards/margins": 8.517755508422852, "rewards/rejected": -39.393001556396484, "step": 506 }, { "epoch": 0.8555789473684211, "grad_norm": 158.9115319918602, "learning_rate": 1.1850524021436336e-08, "logits/chosen": -6.291950225830078, "logits/rejected": -8.398715019226074, "logps/chosen": -1423.4140625, "logps/rejected": -2280.9658203125, "loss": 0.1451, "nll_loss": 5.541550636291504, "rewards/accuracies": 1.0, "rewards/chosen": -22.83027458190918, "rewards/margins": 16.543231964111328, "rewards/rejected": -39.373504638671875, "step": 508 }, { "epoch": 0.8589473684210527, "grad_norm": 139.85116251304456, "learning_rate": 1.1310050521341197e-08, "logits/chosen": -7.1967291831970215, "logits/rejected": -8.252025604248047, "logps/chosen": -1780.57080078125, "logps/rejected": -2163.275390625, "loss": 0.2197, "nll_loss": 5.365113258361816, "rewards/accuracies": 1.0, "rewards/chosen": -29.839202880859375, "rewards/margins": 9.690923690795898, "rewards/rejected": -39.53012466430664, "step": 510 }, { "epoch": 0.8623157894736843, "grad_norm": 320.90667451768593, "learning_rate": 1.0781451912447059e-08, "logits/chosen": -6.554062366485596, "logits/rejected": -8.399422645568848, "logps/chosen": -1562.239013671875, "logps/rejected": -2252.832763671875, "loss": 0.2041, "nll_loss": 5.663952827453613, "rewards/accuracies": 1.0, "rewards/chosen": -25.284164428710938, "rewards/margins": 13.453445434570312, "rewards/rejected": -38.73760986328125, "step": 512 }, { "epoch": 0.8656842105263158, "grad_norm": 466.8053997652555, "learning_rate": 1.0264798969977228e-08, "logits/chosen": -7.735861301422119, "logits/rejected": -8.004199981689453, "logps/chosen": -1992.1900634765625, "logps/rejected": -2198.92041015625, "loss": 0.3714, "nll_loss": 5.44398307800293, "rewards/accuracies": 0.875, "rewards/chosen": -33.00054168701172, "rewards/margins": 4.714691162109375, "rewards/rejected": -37.715232849121094, "step": 514 }, { "epoch": 0.8690526315789474, "grad_norm": 265.87276679833155, "learning_rate": 9.760160869723455e-09, "logits/chosen": -8.163463592529297, "logits/rejected": -7.93167781829834, "logps/chosen": -2085.497314453125, "logps/rejected": -2193.436279296875, "loss": 0.2273, "nll_loss": 5.032775402069092, "rewards/accuracies": 0.875, "rewards/chosen": -33.65901565551758, "rewards/margins": 5.517548561096191, "rewards/rejected": -39.17656326293945, "step": 516 }, { "epoch": 0.872421052631579, "grad_norm": 150.83909835897987, "learning_rate": 9.267605178784032e-09, "logits/chosen": -7.286246299743652, "logits/rejected": -8.414872169494629, "logps/chosen": -1961.04541015625, "logps/rejected": -2198.59912109375, "loss": 0.1727, "nll_loss": 6.030449867248535, "rewards/accuracies": 1.0, "rewards/chosen": -32.829463958740234, "rewards/margins": 5.171086311340332, "rewards/rejected": -38.000553131103516, "step": 518 }, { "epoch": 0.8757894736842106, "grad_norm": 251.81035926119085, "learning_rate": 8.787197846517147e-09, "logits/chosen": -5.83320426940918, "logits/rejected": -8.448880195617676, "logps/chosen": -1317.8983154296875, "logps/rejected": -2223.787109375, "loss": 0.1032, "nll_loss": 5.769347190856934, "rewards/accuracies": 1.0, "rewards/chosen": -20.78626823425293, "rewards/margins": 16.798099517822266, "rewards/rejected": -37.58436965942383, "step": 520 }, { "epoch": 0.8791578947368421, "grad_norm": 363.5013970055686, "learning_rate": 8.319003195710571e-09, "logits/chosen": -7.96688175201416, "logits/rejected": -8.128907203674316, "logps/chosen": -2162.495361328125, "logps/rejected": -2259.739013671875, "loss": 0.2421, "nll_loss": 5.122936725616455, "rewards/accuracies": 1.0, "rewards/chosen": -35.563453674316406, "rewards/margins": 4.058735370635986, "rewards/rejected": -39.6221923828125, "step": 522 }, { "epoch": 0.8825263157894737, "grad_norm": 196.5174829213676, "learning_rate": 7.86308391396956e-09, "logits/chosen": -7.527979850769043, "logits/rejected": -8.271844863891602, "logps/chosen": -2047.7938232421875, "logps/rejected": -2268.5478515625, "loss": 0.2425, "nll_loss": 5.69685697555542, "rewards/accuracies": 0.875, "rewards/chosen": -34.301979064941406, "rewards/margins": 4.180940628051758, "rewards/rejected": -38.48291778564453, "step": 524 }, { "epoch": 0.8858947368421053, "grad_norm": 268.6123339266088, "learning_rate": 7.4195010453232956e-09, "logits/chosen": -7.325321197509766, "logits/rejected": -8.17194938659668, "logps/chosen": -1752.6014404296875, "logps/rejected": -2256.585205078125, "loss": 0.1815, "nll_loss": 5.504175186157227, "rewards/accuracies": 0.75, "rewards/chosen": -28.291757583618164, "rewards/margins": 9.903141021728516, "rewards/rejected": -38.19489669799805, "step": 526 }, { "epoch": 0.8892631578947369, "grad_norm": 246.9449900787623, "learning_rate": 6.988313982051719e-09, "logits/chosen": -7.29826021194458, "logits/rejected": -7.989704608917236, "logps/chosen": -1862.6717529296875, "logps/rejected": -2166.964599609375, "loss": 0.1301, "nll_loss": 5.650459289550781, "rewards/accuracies": 0.75, "rewards/chosen": -30.91683578491211, "rewards/margins": 7.5222368240356445, "rewards/rejected": -38.4390754699707, "step": 528 }, { "epoch": 0.8926315789473684, "grad_norm": 378.94244285907064, "learning_rate": 6.5695804567332036e-09, "logits/chosen": -6.940848350524902, "logits/rejected": -8.17110824584961, "logps/chosen": -1906.43212890625, "logps/rejected": -2278.51025390625, "loss": 0.2445, "nll_loss": 5.465893745422363, "rewards/accuracies": 0.75, "rewards/chosen": -31.02324867248535, "rewards/margins": 8.205381393432617, "rewards/rejected": -39.228633880615234, "step": 530 }, { "epoch": 0.896, "grad_norm": 490.0475742211158, "learning_rate": 6.163356534514807e-09, "logits/chosen": -6.172750949859619, "logits/rejected": -8.489022254943848, "logps/chosen": -1533.9468994140625, "logps/rejected": -2212.72802734375, "loss": 0.1805, "nll_loss": 5.867823123931885, "rewards/accuracies": 1.0, "rewards/chosen": -24.704307556152344, "rewards/margins": 14.34754753112793, "rewards/rejected": -39.05185317993164, "step": 532 }, { "epoch": 0.8993684210526316, "grad_norm": 255.92774706692123, "learning_rate": 5.7696966056053785e-09, "logits/chosen": -6.2565388679504395, "logits/rejected": -8.29428482055664, "logps/chosen": -1412.822998046875, "logps/rejected": -2203.464599609375, "loss": 0.2004, "nll_loss": 5.558704853057861, "rewards/accuracies": 1.0, "rewards/chosen": -20.468172073364258, "rewards/margins": 18.679214477539062, "rewards/rejected": -39.14738845825195, "step": 534 }, { "epoch": 0.9027368421052632, "grad_norm": 114.13752241676485, "learning_rate": 5.388653377993324e-09, "logits/chosen": -6.097024917602539, "logits/rejected": -8.490983963012695, "logps/chosen": -1212.191162109375, "logps/rejected": -2282.863037109375, "loss": 0.1083, "nll_loss": 5.277416229248047, "rewards/accuracies": 0.875, "rewards/chosen": -18.971294403076172, "rewards/margins": 20.884796142578125, "rewards/rejected": -39.8560905456543, "step": 536 }, { "epoch": 0.9061052631578947, "grad_norm": 280.6301781259457, "learning_rate": 5.020277870389311e-09, "logits/chosen": -7.719882965087891, "logits/rejected": -8.041402816772461, "logps/chosen": -2189.172119140625, "logps/rejected": -2295.2314453125, "loss": 0.2457, "nll_loss": 5.236344337463379, "rewards/accuracies": 0.875, "rewards/chosen": -36.655120849609375, "rewards/margins": 2.5539703369140625, "rewards/rejected": -39.20909118652344, "step": 538 }, { "epoch": 0.9094736842105263, "grad_norm": 79.48759546011723, "learning_rate": 4.664619405395265e-09, "logits/chosen": -6.635951995849609, "logits/rejected": -8.384746551513672, "logps/chosen": -1718.5721435546875, "logps/rejected": -2246.451904296875, "loss": 0.1078, "nll_loss": 5.581665992736816, "rewards/accuracies": 1.0, "rewards/chosen": -27.542470932006836, "rewards/margins": 12.282429695129395, "rewards/rejected": -39.82490158081055, "step": 540 }, { "epoch": 0.9128421052631579, "grad_norm": 229.40394912686224, "learning_rate": 4.321725602900472e-09, "logits/chosen": -7.116404056549072, "logits/rejected": -8.396138191223145, "logps/chosen": -1757.71484375, "logps/rejected": -2282.0869140625, "loss": 0.1307, "nll_loss": 4.875128269195557, "rewards/accuracies": 0.875, "rewards/chosen": -27.32893180847168, "rewards/margins": 12.180450439453125, "rewards/rejected": -39.50938415527344, "step": 542 }, { "epoch": 0.9162105263157895, "grad_norm": 289.16292174568304, "learning_rate": 3.991642373705695e-09, "logits/chosen": -6.898987770080566, "logits/rejected": -8.000998497009277, "logps/chosen": -1806.2984619140625, "logps/rejected": -2228.53076171875, "loss": 0.1709, "nll_loss": 5.912762641906738, "rewards/accuracies": 0.875, "rewards/chosen": -30.397703170776367, "rewards/margins": 9.081558227539062, "rewards/rejected": -39.4792594909668, "step": 544 }, { "epoch": 0.919578947368421, "grad_norm": 239.60349733987607, "learning_rate": 3.6744139133759956e-09, "logits/chosen": -7.233107089996338, "logits/rejected": -8.430794715881348, "logps/chosen": -1778.7083740234375, "logps/rejected": -2200.23876953125, "loss": 0.171, "nll_loss": 5.608141899108887, "rewards/accuracies": 1.0, "rewards/chosen": -28.902019500732422, "rewards/margins": 10.116137504577637, "rewards/rejected": -39.018157958984375, "step": 546 }, { "epoch": 0.9229473684210526, "grad_norm": 203.79868091759081, "learning_rate": 3.370082696323373e-09, "logits/chosen": -7.195704460144043, "logits/rejected": -8.166756629943848, "logps/chosen": -1911.45361328125, "logps/rejected": -2244.798095703125, "loss": 0.151, "nll_loss": 5.114628314971924, "rewards/accuracies": 1.0, "rewards/chosen": -30.203014373779297, "rewards/margins": 8.864540100097656, "rewards/rejected": -39.06755447387695, "step": 548 }, { "epoch": 0.9263157894736842, "grad_norm": 120.3056248035602, "learning_rate": 3.0786894701196775e-09, "logits/chosen": -7.326858997344971, "logits/rejected": -8.087577819824219, "logps/chosen": -2035.1636962890625, "logps/rejected": -2246.181884765625, "loss": 0.0989, "nll_loss": 5.451916694641113, "rewards/accuracies": 1.0, "rewards/chosen": -33.032073974609375, "rewards/margins": 5.780745506286621, "rewards/rejected": -38.81281661987305, "step": 550 }, { "epoch": 0.9296842105263158, "grad_norm": 275.770327333272, "learning_rate": 2.800273250040952e-09, "logits/chosen": -7.597753524780273, "logits/rejected": -7.824631690979004, "logps/chosen": -2004.7808837890625, "logps/rejected": -2244.8212890625, "loss": 0.1247, "nll_loss": 4.721861839294434, "rewards/accuracies": 1.0, "rewards/chosen": -31.433671951293945, "rewards/margins": 7.14993143081665, "rewards/rejected": -38.58360290527344, "step": 552 }, { "epoch": 0.9330526315789474, "grad_norm": 422.0746379492613, "learning_rate": 2.5348713138434565e-09, "logits/chosen": -7.181792259216309, "logits/rejected": -8.049543380737305, "logps/chosen": -1707.939208984375, "logps/rejected": -2208.77734375, "loss": 0.2438, "nll_loss": 5.61822509765625, "rewards/accuracies": 1.0, "rewards/chosen": -27.62373161315918, "rewards/margins": 11.296683311462402, "rewards/rejected": -38.92041015625, "step": 554 }, { "epoch": 0.9364210526315789, "grad_norm": 37.31015421332207, "learning_rate": 2.28251919677257e-09, "logits/chosen": -6.2887396812438965, "logits/rejected": -8.343204498291016, "logps/chosen": -1522.6220703125, "logps/rejected": -2214.747802734375, "loss": 0.024, "nll_loss": 5.970060348510742, "rewards/accuracies": 1.0, "rewards/chosen": -23.73328971862793, "rewards/margins": 15.055632591247559, "rewards/rejected": -38.78892135620117, "step": 556 }, { "epoch": 0.9397894736842105, "grad_norm": 310.37839808852675, "learning_rate": 2.043250686804865e-09, "logits/chosen": -7.500852108001709, "logits/rejected": -7.99564266204834, "logps/chosen": -1894.286376953125, "logps/rejected": -2235.044921875, "loss": 0.2862, "nll_loss": 5.685905933380127, "rewards/accuracies": 0.875, "rewards/chosen": -31.6041202545166, "rewards/margins": 7.6095404624938965, "rewards/rejected": -39.21365737915039, "step": 558 }, { "epoch": 0.9431578947368421, "grad_norm": 294.5349485842045, "learning_rate": 1.817097820124147e-09, "logits/chosen": -6.809659957885742, "logits/rejected": -8.295966148376465, "logps/chosen": -1886.675537109375, "logps/rejected": -2271.10986328125, "loss": 0.2189, "nll_loss": 5.538888931274414, "rewards/accuracies": 0.875, "rewards/chosen": -30.958091735839844, "rewards/margins": 7.502249240875244, "rewards/rejected": -38.4603385925293, "step": 560 }, { "epoch": 0.9465263157894737, "grad_norm": 390.96399290727527, "learning_rate": 1.604090876832087e-09, "logits/chosen": -7.394793510437012, "logits/rejected": -7.84267520904541, "logps/chosen": -1881.2449951171875, "logps/rejected": -2232.389404296875, "loss": 0.2396, "nll_loss": 5.126293659210205, "rewards/accuracies": 0.875, "rewards/chosen": -31.569351196289062, "rewards/margins": 7.6439924240112305, "rewards/rejected": -39.213340759277344, "step": 562 }, { "epoch": 0.9498947368421052, "grad_norm": 310.3158547149984, "learning_rate": 1.4042583768939299e-09, "logits/chosen": -6.972746849060059, "logits/rejected": -8.371156692504883, "logps/chosen": -1714.3399658203125, "logps/rejected": -2206.8310546875, "loss": 0.1199, "nll_loss": 5.605197906494141, "rewards/accuracies": 0.875, "rewards/chosen": -27.692014694213867, "rewards/margins": 11.318819046020508, "rewards/rejected": -39.010833740234375, "step": 564 }, { "epoch": 0.9532631578947368, "grad_norm": 184.23275233044205, "learning_rate": 1.2176270763198826e-09, "logits/chosen": -7.448263168334961, "logits/rejected": -8.297872543334961, "logps/chosen": -1999.669189453125, "logps/rejected": -2279.085693359375, "loss": 0.1391, "nll_loss": 5.474849224090576, "rewards/accuracies": 1.0, "rewards/chosen": -33.233360290527344, "rewards/margins": 6.506428241729736, "rewards/rejected": -39.73978805541992, "step": 566 }, { "epoch": 0.9566315789473684, "grad_norm": 212.45164891099338, "learning_rate": 1.0442219635827587e-09, "logits/chosen": -6.379326820373535, "logits/rejected": -8.392101287841797, "logps/chosen": -1411.836669921875, "logps/rejected": -2236.45556640625, "loss": 0.1531, "nll_loss": 5.742588520050049, "rewards/accuracies": 1.0, "rewards/chosen": -22.78182601928711, "rewards/margins": 15.909248352050781, "rewards/rejected": -38.69107437133789, "step": 568 }, { "epoch": 0.96, "grad_norm": 176.85132902561122, "learning_rate": 8.840662562721313e-10, "logits/chosen": -7.5018157958984375, "logits/rejected": -8.164826393127441, "logps/chosen": -2023.5302734375, "logps/rejected": -2211.3251953125, "loss": 0.1309, "nll_loss": 5.339632034301758, "rewards/accuracies": 1.0, "rewards/chosen": -33.29636764526367, "rewards/margins": 5.157810211181641, "rewards/rejected": -38.45417404174805, "step": 570 }, { "epoch": 0.9633684210526315, "grad_norm": 175.63839472941274, "learning_rate": 7.371813979857311e-10, "logits/chosen": -7.2445173263549805, "logits/rejected": -8.058394432067871, "logps/chosen": -1971.2177734375, "logps/rejected": -2261.788818359375, "loss": 0.1188, "nll_loss": 4.997542381286621, "rewards/accuracies": 0.875, "rewards/chosen": -31.539989471435547, "rewards/margins": 7.252119064331055, "rewards/rejected": -38.792110443115234, "step": 572 }, { "epoch": 0.9667368421052631, "grad_norm": 190.37162464177797, "learning_rate": 6.035870554582989e-10, "logits/chosen": -6.96453332901001, "logits/rejected": -8.222492218017578, "logps/chosen": -1799.5657958984375, "logps/rejected": -2261.03857421875, "loss": 0.1363, "nll_loss": 5.406635284423828, "rewards/accuracies": 0.875, "rewards/chosen": -29.174453735351562, "rewards/margins": 9.996255874633789, "rewards/rejected": -39.170711517333984, "step": 574 }, { "epoch": 0.9701052631578947, "grad_norm": 179.27962403393755, "learning_rate": 4.833011159284028e-10, "logits/chosen": -7.209228515625, "logits/rejected": -7.93266487121582, "logps/chosen": -2057.040771484375, "logps/rejected": -2237.497314453125, "loss": 0.0949, "nll_loss": 5.6815595626831055, "rewards/accuracies": 0.875, "rewards/chosen": -33.02680969238281, "rewards/margins": 6.363800048828125, "rewards/rejected": -39.3906135559082, "step": 576 }, { "epoch": 0.9734736842105263, "grad_norm": 182.8402649855879, "learning_rate": 3.763396847433875e-10, "logits/chosen": -7.02055549621582, "logits/rejected": -8.420388221740723, "logps/chosen": -1753.1495361328125, "logps/rejected": -2262.48388671875, "loss": 0.2172, "nll_loss": 5.674689292907715, "rewards/accuracies": 1.0, "rewards/chosen": -28.21030044555664, "rewards/margins": 11.214584350585938, "rewards/rejected": -39.42488479614258, "step": 578 }, { "epoch": 0.9768421052631578, "grad_norm": 59.175625376263454, "learning_rate": 2.8271708320309893e-10, "logits/chosen": -7.082157135009766, "logits/rejected": -7.969208240509033, "logps/chosen": -1872.389404296875, "logps/rejected": -2212.38916015625, "loss": 0.0631, "nll_loss": 5.4750566482543945, "rewards/accuracies": 1.0, "rewards/chosen": -30.0280704498291, "rewards/margins": 8.51849365234375, "rewards/rejected": -38.54656219482422, "step": 580 }, { "epoch": 0.9802105263157894, "grad_norm": 31.130838237276375, "learning_rate": 2.0244584664229624e-10, "logits/chosen": -6.935577392578125, "logits/rejected": -8.142889022827148, "logps/chosen": -1751.22998046875, "logps/rejected": -2215.97900390625, "loss": 0.1858, "nll_loss": 5.484932899475098, "rewards/accuracies": 1.0, "rewards/chosen": -27.302139282226562, "rewards/margins": 10.821523666381836, "rewards/rejected": -38.123661041259766, "step": 582 }, { "epoch": 0.983578947368421, "grad_norm": 300.9076244256699, "learning_rate": 1.3553672275230522e-10, "logits/chosen": -6.667942047119141, "logits/rejected": -8.240449905395508, "logps/chosen": -1669.30908203125, "logps/rejected": -2250.7109375, "loss": 0.2112, "nll_loss": 5.46246337890625, "rewards/accuracies": 0.875, "rewards/chosen": -26.676097869873047, "rewards/margins": 11.850287437438965, "rewards/rejected": -38.526390075683594, "step": 584 }, { "epoch": 0.9869473684210527, "grad_norm": 205.37779593349424, "learning_rate": 8.199867014198059e-11, "logits/chosen": -7.167195796966553, "logits/rejected": -8.354471206665039, "logps/chosen": -1912.6531982421875, "logps/rejected": -2276.97509765625, "loss": 0.1158, "nll_loss": 5.5597991943359375, "rewards/accuracies": 1.0, "rewards/chosen": -30.361347198486328, "rewards/margins": 7.688362121582031, "rewards/rejected": -38.049713134765625, "step": 586 }, { "epoch": 0.9903157894736843, "grad_norm": 276.43754945737743, "learning_rate": 4.1838857138221105e-11, "logits/chosen": -6.897299766540527, "logits/rejected": -8.224220275878906, "logps/chosen": -1726.572265625, "logps/rejected": -2215.0009765625, "loss": 0.3376, "nll_loss": 5.524778842926025, "rewards/accuracies": 0.875, "rewards/chosen": -27.695968627929688, "rewards/margins": 10.981616973876953, "rewards/rejected": -38.677589416503906, "step": 588 }, { "epoch": 0.9936842105263158, "grad_norm": 101.87640432314295, "learning_rate": 1.506266082615948e-11, "logits/chosen": -6.690047264099121, "logits/rejected": -8.358760833740234, "logps/chosen": -1721.9786376953125, "logps/rejected": -2250.2998046875, "loss": 0.1519, "nll_loss": 5.525182247161865, "rewards/accuracies": 0.875, "rewards/chosen": -28.625076293945312, "rewards/margins": 10.398566246032715, "rewards/rejected": -39.02363967895508, "step": 590 }, { "epoch": 0.9970526315789474, "grad_norm": 109.18438747710223, "learning_rate": 1.6736663292604703e-12, "logits/chosen": -7.776650428771973, "logits/rejected": -8.136452674865723, "logps/chosen": -2102.580810546875, "logps/rejected": -2248.21337890625, "loss": 0.1902, "nll_loss": 5.207364559173584, "rewards/accuracies": 0.875, "rewards/chosen": -34.941654205322266, "rewards/margins": 4.452713489532471, "rewards/rejected": -39.39436721801758, "step": 592 } ], "logging_steps": 2, "max_steps": 593, "num_input_tokens_seen": 0, "num_train_epochs": 1, "save_steps": 50, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 1, "trial_name": null, "trial_params": null }