| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9989071038251366, | |
| "eval_steps": 400, | |
| "global_step": 457, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01092896174863388, | |
| "grad_norm": 358.5620562461214, | |
| "learning_rate": 1.0869565217391303e-07, | |
| "logits/chosen": -1.0113575458526611, | |
| "logits/rejected": -1.0064939260482788, | |
| "logps/chosen": -0.2803983986377716, | |
| "logps/rejected": -0.2860395908355713, | |
| "loss": 5.2315, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -2.8039839267730713, | |
| "rewards/margins": 0.056411754339933395, | |
| "rewards/rejected": -2.860395669937134, | |
| "semantic_entropy": 0.7518940567970276, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.02185792349726776, | |
| "grad_norm": 233.11198507849488, | |
| "learning_rate": 2.1739130434782607e-07, | |
| "logits/chosen": -1.056563138961792, | |
| "logits/rejected": -1.0053507089614868, | |
| "logps/chosen": -0.2568749487400055, | |
| "logps/rejected": -0.27021342515945435, | |
| "loss": 5.206, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -2.568748950958252, | |
| "rewards/margins": 0.13338527083396912, | |
| "rewards/rejected": -2.702134370803833, | |
| "semantic_entropy": 0.7094504237174988, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03278688524590164, | |
| "grad_norm": 210.79380420075688, | |
| "learning_rate": 3.260869565217391e-07, | |
| "logits/chosen": -1.0091139078140259, | |
| "logits/rejected": -0.9631060361862183, | |
| "logps/chosen": -0.2674282491207123, | |
| "logps/rejected": -0.27336788177490234, | |
| "loss": 5.1278, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -2.6742825508117676, | |
| "rewards/margins": 0.05939612537622452, | |
| "rewards/rejected": -2.7336785793304443, | |
| "semantic_entropy": 0.7274739742279053, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.04371584699453552, | |
| "grad_norm": 375.0299222500816, | |
| "learning_rate": 4.3478260869565214e-07, | |
| "logits/chosen": -0.9487798810005188, | |
| "logits/rejected": -0.8998070955276489, | |
| "logps/chosen": -0.2723819613456726, | |
| "logps/rejected": -0.28497135639190674, | |
| "loss": 5.222, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -2.7238194942474365, | |
| "rewards/margins": 0.1258939504623413, | |
| "rewards/rejected": -2.8497138023376465, | |
| "semantic_entropy": 0.7452942132949829, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.0546448087431694, | |
| "grad_norm": 299.8105605992341, | |
| "learning_rate": 5.434782608695652e-07, | |
| "logits/chosen": -0.9505411386489868, | |
| "logits/rejected": -0.8759678602218628, | |
| "logps/chosen": -0.27557066082954407, | |
| "logps/rejected": -0.29382389783859253, | |
| "loss": 5.1223, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -2.755706548690796, | |
| "rewards/margins": 0.1825324147939682, | |
| "rewards/rejected": -2.9382388591766357, | |
| "semantic_entropy": 0.7546485662460327, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.06557377049180328, | |
| "grad_norm": 300.511189676041, | |
| "learning_rate": 6.521739130434782e-07, | |
| "logits/chosen": -1.0557540655136108, | |
| "logits/rejected": -0.990186333656311, | |
| "logps/chosen": -0.267598032951355, | |
| "logps/rejected": -0.28400346636772156, | |
| "loss": 5.1993, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -2.6759800910949707, | |
| "rewards/margins": 0.1640542596578598, | |
| "rewards/rejected": -2.8400347232818604, | |
| "semantic_entropy": 0.7248786091804504, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.07650273224043716, | |
| "grad_norm": 186.14457687250982, | |
| "learning_rate": 7.608695652173913e-07, | |
| "logits/chosen": -1.0076847076416016, | |
| "logits/rejected": -0.9405019879341125, | |
| "logps/chosen": -0.2580474615097046, | |
| "logps/rejected": -0.279694139957428, | |
| "loss": 5.0812, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -2.580474615097046, | |
| "rewards/margins": 0.2164669930934906, | |
| "rewards/rejected": -2.7969415187835693, | |
| "semantic_entropy": 0.7201561331748962, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.08743169398907104, | |
| "grad_norm": 223.67562662995005, | |
| "learning_rate": 8.695652173913043e-07, | |
| "logits/chosen": -0.9655061960220337, | |
| "logits/rejected": -0.9045132398605347, | |
| "logps/chosen": -0.2820321023464203, | |
| "logps/rejected": -0.2990434765815735, | |
| "loss": 5.3116, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -2.820321559906006, | |
| "rewards/margins": 0.17011362314224243, | |
| "rewards/rejected": -2.9904348850250244, | |
| "semantic_entropy": 0.7597802877426147, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.09836065573770492, | |
| "grad_norm": 131.00541589477967, | |
| "learning_rate": 9.782608695652173e-07, | |
| "logits/chosen": -1.0179851055145264, | |
| "logits/rejected": -0.9359539151191711, | |
| "logps/chosen": -0.2855134606361389, | |
| "logps/rejected": -0.3078162968158722, | |
| "loss": 4.919, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -2.8551342487335205, | |
| "rewards/margins": 0.22302868962287903, | |
| "rewards/rejected": -3.078162908554077, | |
| "semantic_entropy": 0.7599790096282959, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.1092896174863388, | |
| "grad_norm": 360.5211336376954, | |
| "learning_rate": 9.997663088532014e-07, | |
| "logits/chosen": -0.9701619148254395, | |
| "logits/rejected": -0.887865424156189, | |
| "logps/chosen": -0.28138467669487, | |
| "logps/rejected": -0.28913217782974243, | |
| "loss": 5.1014, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -2.813847064971924, | |
| "rewards/margins": 0.07747481018304825, | |
| "rewards/rejected": -2.891321897506714, | |
| "semantic_entropy": 0.7510823607444763, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12021857923497267, | |
| "grad_norm": 162.42445549849919, | |
| "learning_rate": 9.98817312944725e-07, | |
| "logits/chosen": -1.0028311014175415, | |
| "logits/rejected": -0.8850187063217163, | |
| "logps/chosen": -0.27895885705947876, | |
| "logps/rejected": -0.3143305778503418, | |
| "loss": 4.8519, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": -2.7895889282226562, | |
| "rewards/margins": 0.3537173271179199, | |
| "rewards/rejected": -3.143306255340576, | |
| "semantic_entropy": 0.7608937621116638, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.13114754098360656, | |
| "grad_norm": 120.127890644715, | |
| "learning_rate": 9.971397915250336e-07, | |
| "logits/chosen": -1.0246554613113403, | |
| "logits/rejected": -0.9786936640739441, | |
| "logps/chosen": -0.26926669478416443, | |
| "logps/rejected": -0.3113505244255066, | |
| "loss": 4.6493, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -2.692667245864868, | |
| "rewards/margins": 0.42083826661109924, | |
| "rewards/rejected": -3.1135053634643555, | |
| "semantic_entropy": 0.7576217651367188, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14207650273224043, | |
| "grad_norm": 214.34275301709462, | |
| "learning_rate": 9.94736194623663e-07, | |
| "logits/chosen": -0.9859918355941772, | |
| "logits/rejected": -0.9181090593338013, | |
| "logps/chosen": -0.3084973096847534, | |
| "logps/rejected": -0.3399081826210022, | |
| "loss": 4.9942, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -3.084973096847534, | |
| "rewards/margins": 0.31410855054855347, | |
| "rewards/rejected": -3.3990814685821533, | |
| "semantic_entropy": 0.8060176968574524, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.15300546448087432, | |
| "grad_norm": 544.9393207205336, | |
| "learning_rate": 9.916100327075037e-07, | |
| "logits/chosen": -0.9521551132202148, | |
| "logits/rejected": -0.9334642291069031, | |
| "logps/chosen": -0.29637694358825684, | |
| "logps/rejected": -0.3240143656730652, | |
| "loss": 4.7857, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -2.9637694358825684, | |
| "rewards/margins": 0.27637460827827454, | |
| "rewards/rejected": -3.2401435375213623, | |
| "semantic_entropy": 0.7712014317512512, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16393442622950818, | |
| "grad_norm": 123.2584567983735, | |
| "learning_rate": 9.877658715537428e-07, | |
| "logits/chosen": -0.9457874298095703, | |
| "logits/rejected": -0.9289388656616211, | |
| "logps/chosen": -0.31344500184059143, | |
| "logps/rejected": -0.3473803997039795, | |
| "loss": 4.8544, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -3.1344501972198486, | |
| "rewards/margins": 0.33935409784317017, | |
| "rewards/rejected": -3.473804473876953, | |
| "semantic_entropy": 0.8020299077033997, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.17486338797814208, | |
| "grad_norm": 136.55956527115663, | |
| "learning_rate": 9.832093255815216e-07, | |
| "logits/chosen": -0.9409273266792297, | |
| "logits/rejected": -0.8797443509101868, | |
| "logps/chosen": -0.3109641373157501, | |
| "logps/rejected": -0.3346250057220459, | |
| "loss": 4.7853, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -3.1096413135528564, | |
| "rewards/margins": 0.23660895228385925, | |
| "rewards/rejected": -3.346250534057617, | |
| "semantic_entropy": 0.7786868810653687, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18579234972677597, | |
| "grad_norm": 202.99615488786125, | |
| "learning_rate": 9.779470496520441e-07, | |
| "logits/chosen": -0.9390329122543335, | |
| "logits/rejected": -0.89063560962677, | |
| "logps/chosen": -0.31133827567100525, | |
| "logps/rejected": -0.3710102438926697, | |
| "loss": 4.6394, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -3.1133828163146973, | |
| "rewards/margins": 0.5967196226119995, | |
| "rewards/rejected": -3.7101027965545654, | |
| "semantic_entropy": 0.8020746111869812, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.19672131147540983, | |
| "grad_norm": 147.8274369377715, | |
| "learning_rate": 9.719867293491144e-07, | |
| "logits/chosen": -0.9997242093086243, | |
| "logits/rejected": -0.9218411445617676, | |
| "logps/chosen": -0.33917468786239624, | |
| "logps/rejected": -0.3771332800388336, | |
| "loss": 4.7466, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -3.391746997833252, | |
| "rewards/margins": 0.37958595156669617, | |
| "rewards/rejected": -3.7713329792022705, | |
| "semantic_entropy": 0.8518031239509583, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20765027322404372, | |
| "grad_norm": 150.5965714109422, | |
| "learning_rate": 9.653370697542987e-07, | |
| "logits/chosen": -0.949033260345459, | |
| "logits/rejected": -0.9504354596138, | |
| "logps/chosen": -0.34006524085998535, | |
| "logps/rejected": -0.3670746684074402, | |
| "loss": 4.5023, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -3.4006524085998535, | |
| "rewards/margins": 0.27009397745132446, | |
| "rewards/rejected": -3.6707465648651123, | |
| "semantic_entropy": 0.8378095626831055, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.2185792349726776, | |
| "grad_norm": 205.55603229224332, | |
| "learning_rate": 9.580077827331037e-07, | |
| "logits/chosen": -0.9659525752067566, | |
| "logits/rejected": -0.9214147329330444, | |
| "logps/chosen": -0.37815189361572266, | |
| "logps/rejected": -0.4408392012119293, | |
| "loss": 4.4615, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -3.7815189361572266, | |
| "rewards/margins": 0.6268730163574219, | |
| "rewards/rejected": -4.408391952514648, | |
| "semantic_entropy": 0.8833344578742981, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.22950819672131148, | |
| "grad_norm": 168.46156689250674, | |
| "learning_rate": 9.500095727507419e-07, | |
| "logits/chosen": -1.0187479257583618, | |
| "logits/rejected": -0.9876689910888672, | |
| "logps/chosen": -0.3568420708179474, | |
| "logps/rejected": -0.4102093577384949, | |
| "loss": 4.4882, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -3.568420886993408, | |
| "rewards/margins": 0.5336726903915405, | |
| "rewards/rejected": -4.102093696594238, | |
| "semantic_entropy": 0.8596333265304565, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.24043715846994534, | |
| "grad_norm": 163.05795657074705, | |
| "learning_rate": 9.413541212382004e-07, | |
| "logits/chosen": -1.013091802597046, | |
| "logits/rejected": -0.9950464367866516, | |
| "logps/chosen": -0.37098902463912964, | |
| "logps/rejected": -0.45781344175338745, | |
| "loss": 4.3876, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -3.709890842437744, | |
| "rewards/margins": 0.8682438135147095, | |
| "rewards/rejected": -4.578134059906006, | |
| "semantic_entropy": 0.9028980135917664, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25136612021857924, | |
| "grad_norm": 159.92608074418774, | |
| "learning_rate": 9.320540695314438e-07, | |
| "logits/chosen": -1.0206435918807983, | |
| "logits/rejected": -0.9817934036254883, | |
| "logps/chosen": -0.37152332067489624, | |
| "logps/rejected": -0.4835759103298187, | |
| "loss": 4.1862, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -3.715233325958252, | |
| "rewards/margins": 1.1205257177352905, | |
| "rewards/rejected": -4.835759162902832, | |
| "semantic_entropy": 0.8893026113510132, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.26229508196721313, | |
| "grad_norm": 126.53869344949423, | |
| "learning_rate": 9.221230004086721e-07, | |
| "logits/chosen": -1.0430892705917358, | |
| "logits/rejected": -0.9731811285018921, | |
| "logps/chosen": -0.38018742203712463, | |
| "logps/rejected": -0.44267600774765015, | |
| "loss": 4.3042, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -3.8018736839294434, | |
| "rewards/margins": 0.6248863935470581, | |
| "rewards/rejected": -4.426759719848633, | |
| "semantic_entropy": 0.9051868319511414, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.273224043715847, | |
| "grad_norm": 330.21792860177044, | |
| "learning_rate": 9.11575418252596e-07, | |
| "logits/chosen": -0.9472485780715942, | |
| "logits/rejected": -0.9132539629936218, | |
| "logps/chosen": -0.396454393863678, | |
| "logps/rejected": -0.4733741283416748, | |
| "loss": 4.0912, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -3.9645442962646484, | |
| "rewards/margins": 0.7691973447799683, | |
| "rewards/rejected": -4.733741283416748, | |
| "semantic_entropy": 0.9098461866378784, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.28415300546448086, | |
| "grad_norm": 136.7218535491149, | |
| "learning_rate": 9.004267278667031e-07, | |
| "logits/chosen": -0.9847833514213562, | |
| "logits/rejected": -0.9780336618423462, | |
| "logps/chosen": -0.4162030816078186, | |
| "logps/rejected": -0.5500742793083191, | |
| "loss": 4.1012, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -4.162030220031738, | |
| "rewards/margins": 1.3387129306793213, | |
| "rewards/rejected": -5.500743389129639, | |
| "semantic_entropy": 0.9023580551147461, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29508196721311475, | |
| "grad_norm": 114.59260330847327, | |
| "learning_rate": 8.886932119764565e-07, | |
| "logits/chosen": -1.018243670463562, | |
| "logits/rejected": -0.934001624584198, | |
| "logps/chosen": -0.40604203939437866, | |
| "logps/rejected": -0.521537184715271, | |
| "loss": 3.933, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -4.060420036315918, | |
| "rewards/margins": 1.154951810836792, | |
| "rewards/rejected": -5.215372562408447, | |
| "semantic_entropy": 0.9238536953926086, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.30601092896174864, | |
| "grad_norm": 151.22961611030988, | |
| "learning_rate": 8.763920074482809e-07, | |
| "logits/chosen": -1.02057683467865, | |
| "logits/rejected": -0.9660438299179077, | |
| "logps/chosen": -0.43255481123924255, | |
| "logps/rejected": -0.5811036229133606, | |
| "loss": 3.5202, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -4.32554817199707, | |
| "rewards/margins": 1.485487699508667, | |
| "rewards/rejected": -5.811036109924316, | |
| "semantic_entropy": 0.9533751606941223, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31693989071038253, | |
| "grad_norm": 171.9485188405326, | |
| "learning_rate": 8.635410802610723e-07, | |
| "logits/chosen": -1.0028008222579956, | |
| "logits/rejected": -0.9838630557060242, | |
| "logps/chosen": -0.4235480725765228, | |
| "logps/rejected": -0.49707216024398804, | |
| "loss": 3.7809, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -4.235480785369873, | |
| "rewards/margins": 0.7352409958839417, | |
| "rewards/rejected": -4.970722198486328, | |
| "semantic_entropy": 0.9417489767074585, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.32786885245901637, | |
| "grad_norm": 135.1994551251087, | |
| "learning_rate": 8.501591992667849e-07, | |
| "logits/chosen": -1.0660603046417236, | |
| "logits/rejected": -1.0330188274383545, | |
| "logps/chosen": -0.458204448223114, | |
| "logps/rejected": -0.6369711756706238, | |
| "loss": 3.6343, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -4.58204460144043, | |
| "rewards/margins": 1.7876676321029663, | |
| "rewards/rejected": -6.369711875915527, | |
| "semantic_entropy": 0.9522278904914856, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33879781420765026, | |
| "grad_norm": 123.13149743090504, | |
| "learning_rate": 8.362659087784152e-07, | |
| "logits/chosen": -0.9992470741271973, | |
| "logits/rejected": -0.946983814239502, | |
| "logps/chosen": -0.45955339074134827, | |
| "logps/rejected": -0.5732384324073792, | |
| "loss": 3.6764, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -4.5955328941345215, | |
| "rewards/margins": 1.1368510723114014, | |
| "rewards/rejected": -5.732384204864502, | |
| "semantic_entropy": 0.9510468244552612, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.34972677595628415, | |
| "grad_norm": 245.75242739092204, | |
| "learning_rate": 8.218815000254231e-07, | |
| "logits/chosen": -1.0509494543075562, | |
| "logits/rejected": -0.9942834973335266, | |
| "logps/chosen": -0.5233258008956909, | |
| "logps/rejected": -0.6250703930854797, | |
| "loss": 3.7281, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -5.233258247375488, | |
| "rewards/margins": 1.0174460411071777, | |
| "rewards/rejected": -6.250703811645508, | |
| "semantic_entropy": 0.9671458005905151, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36065573770491804, | |
| "grad_norm": 158.28325663577223, | |
| "learning_rate": 8.07026981518276e-07, | |
| "logits/chosen": -1.0312683582305908, | |
| "logits/rejected": -0.9780334234237671, | |
| "logps/chosen": -0.5141728520393372, | |
| "logps/rejected": -0.6252259016036987, | |
| "loss": 3.6091, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -5.141728401184082, | |
| "rewards/margins": 1.1105303764343262, | |
| "rewards/rejected": -6.252258777618408, | |
| "semantic_entropy": 0.9860242605209351, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.37158469945355194, | |
| "grad_norm": 136.70967385020978, | |
| "learning_rate": 7.917240483654e-07, | |
| "logits/chosen": -1.0271486043930054, | |
| "logits/rejected": -0.9650506973266602, | |
| "logps/chosen": -0.5105966329574585, | |
| "logps/rejected": -0.6147508025169373, | |
| "loss": 3.7493, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -5.105965614318848, | |
| "rewards/margins": 1.0415422916412354, | |
| "rewards/rejected": -6.147508144378662, | |
| "semantic_entropy": 0.9853399991989136, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.3825136612021858, | |
| "grad_norm": 192.89943041726585, | |
| "learning_rate": 7.759950505873521e-07, | |
| "logits/chosen": -1.0819661617279053, | |
| "logits/rejected": -1.050621747970581, | |
| "logps/chosen": -0.5413715839385986, | |
| "logps/rejected": -0.6315657496452332, | |
| "loss": 3.5161, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -5.413715362548828, | |
| "rewards/margins": 0.9019424319267273, | |
| "rewards/rejected": -6.315657615661621, | |
| "semantic_entropy": 0.9645811319351196, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.39344262295081966, | |
| "grad_norm": 152.89832261081025, | |
| "learning_rate": 7.598629604744872e-07, | |
| "logits/chosen": -1.0869873762130737, | |
| "logits/rejected": -1.0785080194473267, | |
| "logps/chosen": -0.5242325067520142, | |
| "logps/rejected": -0.7012667655944824, | |
| "loss": 3.3521, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -5.242325305938721, | |
| "rewards/margins": 1.7703425884246826, | |
| "rewards/rejected": -7.012667655944824, | |
| "semantic_entropy": 1.0011693239212036, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.40437158469945356, | |
| "grad_norm": 150.16456326171243, | |
| "learning_rate": 7.433513390357989e-07, | |
| "logits/chosen": -1.1106479167938232, | |
| "logits/rejected": -1.1203057765960693, | |
| "logps/chosen": -0.5613775253295898, | |
| "logps/rejected": -0.7369329333305359, | |
| "loss": 3.3325, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -5.61377477645874, | |
| "rewards/margins": 1.755553960800171, | |
| "rewards/rejected": -7.36932897567749, | |
| "semantic_entropy": 1.0098798274993896, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 0.41530054644808745, | |
| "grad_norm": 177.03981164325452, | |
| "learning_rate": 7.264843015879321e-07, | |
| "logits/chosen": -1.118817925453186, | |
| "logits/rejected": -1.075656771659851, | |
| "logps/chosen": -0.5620681047439575, | |
| "logps/rejected": -0.7573049664497375, | |
| "loss": 3.3889, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -5.620680332183838, | |
| "rewards/margins": 1.9523694515228271, | |
| "rewards/rejected": -7.573049068450928, | |
| "semantic_entropy": 0.998263955116272, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4262295081967213, | |
| "grad_norm": 126.76170022708793, | |
| "learning_rate": 7.092864825346266e-07, | |
| "logits/chosen": -1.1385692358016968, | |
| "logits/rejected": -1.1158863306045532, | |
| "logps/chosen": -0.6738488078117371, | |
| "logps/rejected": -0.8919061422348022, | |
| "loss": 3.431, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -6.73848819732666, | |
| "rewards/margins": 2.180572986602783, | |
| "rewards/rejected": -8.919061660766602, | |
| "semantic_entropy": 0.9914792776107788, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.4371584699453552, | |
| "grad_norm": 146.12776815167663, | |
| "learning_rate": 6.917829993880302e-07, | |
| "logits/chosen": -1.1107165813446045, | |
| "logits/rejected": -1.0282893180847168, | |
| "logps/chosen": -0.6398170590400696, | |
| "logps/rejected": -0.8206952810287476, | |
| "loss": 3.2848, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -6.398170471191406, | |
| "rewards/margins": 1.8087825775146484, | |
| "rewards/rejected": -8.206953048706055, | |
| "semantic_entropy": 1.0069334506988525, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44808743169398907, | |
| "grad_norm": 114.70543290599623, | |
| "learning_rate": 6.739994160844309e-07, | |
| "logits/chosen": -1.0968120098114014, | |
| "logits/rejected": -1.1100647449493408, | |
| "logps/chosen": -0.6191304326057434, | |
| "logps/rejected": -0.8075233697891235, | |
| "loss": 3.113, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -6.1913042068481445, | |
| "rewards/margins": 1.8839294910430908, | |
| "rewards/rejected": -8.075233459472656, | |
| "semantic_entropy": 1.0141515731811523, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 0.45901639344262296, | |
| "grad_norm": 207.23320066468338, | |
| "learning_rate": 6.559617056479827e-07, | |
| "logits/chosen": -1.1229215860366821, | |
| "logits/rejected": -1.1227028369903564, | |
| "logps/chosen": -0.6795850396156311, | |
| "logps/rejected": -0.9186260104179382, | |
| "loss": 3.1442, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -6.7958502769470215, | |
| "rewards/margins": 2.3904099464416504, | |
| "rewards/rejected": -9.186259269714355, | |
| "semantic_entropy": 0.9720403552055359, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46994535519125685, | |
| "grad_norm": 154.50502979258272, | |
| "learning_rate": 6.376962122569567e-07, | |
| "logits/chosen": -1.1284302473068237, | |
| "logits/rejected": -1.0769437551498413, | |
| "logps/chosen": -0.6974600553512573, | |
| "logps/rejected": -0.9393995404243469, | |
| "loss": 3.4054, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -6.974600315093994, | |
| "rewards/margins": 2.4193947315216064, | |
| "rewards/rejected": -9.39399528503418, | |
| "semantic_entropy": 1.0159441232681274, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 0.4808743169398907, | |
| "grad_norm": 138.6943703552703, | |
| "learning_rate": 6.192296127679192e-07, | |
| "logits/chosen": -1.1995110511779785, | |
| "logits/rejected": -1.1735769510269165, | |
| "logps/chosen": -0.7303147912025452, | |
| "logps/rejected": -0.9362783432006836, | |
| "loss": 3.1146, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -7.303147792816162, | |
| "rewards/margins": 2.0596346855163574, | |
| "rewards/rejected": -9.362783432006836, | |
| "semantic_entropy": 0.9887905120849609, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.4918032786885246, | |
| "grad_norm": 120.75185480754473, | |
| "learning_rate": 6.005888777540319e-07, | |
| "logits/chosen": -1.2185839414596558, | |
| "logits/rejected": -1.1783679723739624, | |
| "logps/chosen": -0.7471610307693481, | |
| "logps/rejected": -0.9852075576782227, | |
| "loss": 3.1221, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -7.471610069274902, | |
| "rewards/margins": 2.380465030670166, | |
| "rewards/rejected": -9.852075576782227, | |
| "semantic_entropy": 0.9979068040847778, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.5027322404371585, | |
| "grad_norm": 122.09354415954165, | |
| "learning_rate": 5.818012321143773e-07, | |
| "logits/chosen": -1.1167972087860107, | |
| "logits/rejected": -1.1158543825149536, | |
| "logps/chosen": -0.7506524920463562, | |
| "logps/rejected": -1.0082188844680786, | |
| "loss": 3.1675, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -7.50652551651001, | |
| "rewards/margins": 2.5756633281707764, | |
| "rewards/rejected": -10.082188606262207, | |
| "semantic_entropy": 0.993320643901825, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5136612021857924, | |
| "grad_norm": 147.5670346142303, | |
| "learning_rate": 5.628941153118388e-07, | |
| "logits/chosen": -1.1291579008102417, | |
| "logits/rejected": -1.0918009281158447, | |
| "logps/chosen": -0.775924563407898, | |
| "logps/rejected": -1.0093990564346313, | |
| "loss": 2.9612, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -7.759246826171875, | |
| "rewards/margins": 2.3347439765930176, | |
| "rewards/rejected": -10.093989372253418, | |
| "semantic_entropy": 0.981053352355957, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 0.5245901639344263, | |
| "grad_norm": 133.14046441252583, | |
| "learning_rate": 5.438951412976098e-07, | |
| "logits/chosen": -1.1797640323638916, | |
| "logits/rejected": -1.1900447607040405, | |
| "logps/chosen": -0.7702494859695435, | |
| "logps/rejected": -1.056715488433838, | |
| "loss": 2.7098, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -7.702495574951172, | |
| "rewards/margins": 2.864659547805786, | |
| "rewards/rejected": -10.567155838012695, | |
| "semantic_entropy": 0.991602897644043, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5355191256830601, | |
| "grad_norm": 133.0444183672, | |
| "learning_rate": 5.248320581808619e-07, | |
| "logits/chosen": -1.1087061166763306, | |
| "logits/rejected": -1.0653207302093506, | |
| "logps/chosen": -0.792129397392273, | |
| "logps/rejected": -1.075703501701355, | |
| "loss": 2.9413, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -7.921294212341309, | |
| "rewards/margins": 2.835740566253662, | |
| "rewards/rejected": -10.757036209106445, | |
| "semantic_entropy": 0.9732475280761719, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 0.546448087431694, | |
| "grad_norm": 122.1994564086186, | |
| "learning_rate": 5.057327077024744e-07, | |
| "logits/chosen": -1.1738090515136719, | |
| "logits/rejected": -1.1409178972244263, | |
| "logps/chosen": -0.806613564491272, | |
| "logps/rejected": -1.0335161685943604, | |
| "loss": 3.0864, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -8.06613540649414, | |
| "rewards/margins": 2.2690269947052, | |
| "rewards/rejected": -10.335162162780762, | |
| "semantic_entropy": 0.9622586965560913, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5573770491803278, | |
| "grad_norm": 232.10642636329976, | |
| "learning_rate": 4.866249845720132e-07, | |
| "logits/chosen": -1.1730918884277344, | |
| "logits/rejected": -1.1448460817337036, | |
| "logps/chosen": -0.8790150880813599, | |
| "logps/rejected": -1.1891063451766968, | |
| "loss": 2.8134, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -8.790151596069336, | |
| "rewards/margins": 3.1009116172790527, | |
| "rewards/rejected": -11.891061782836914, | |
| "semantic_entropy": 0.9595390558242798, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 0.5683060109289617, | |
| "grad_norm": 131.2755348822229, | |
| "learning_rate": 4.675367957273505e-07, | |
| "logits/chosen": -1.141157865524292, | |
| "logits/rejected": -1.1335737705230713, | |
| "logps/chosen": -0.8586977124214172, | |
| "logps/rejected": -1.1454424858093262, | |
| "loss": 2.8892, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -8.586977005004883, | |
| "rewards/margins": 2.8674476146698, | |
| "rewards/rejected": -11.454424858093262, | |
| "semantic_entropy": 0.9425627589225769, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5792349726775956, | |
| "grad_norm": 140.64129624293926, | |
| "learning_rate": 4.4849601957642285e-07, | |
| "logits/chosen": -1.1671048402786255, | |
| "logits/rejected": -1.1342850923538208, | |
| "logps/chosen": -0.884295642375946, | |
| "logps/rejected": -1.1824935674667358, | |
| "loss": 2.8754, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -8.84295654296875, | |
| "rewards/margins": 2.9819793701171875, | |
| "rewards/rejected": -11.824935913085938, | |
| "semantic_entropy": 0.9442464709281921, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 0.5901639344262295, | |
| "grad_norm": 147.4714622434341, | |
| "learning_rate": 4.295304652806592e-07, | |
| "logits/chosen": -1.161628246307373, | |
| "logits/rejected": -1.1403006315231323, | |
| "logps/chosen": -0.9028175473213196, | |
| "logps/rejected": -1.2475910186767578, | |
| "loss": 2.6617, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -9.028175354003906, | |
| "rewards/margins": 3.447734832763672, | |
| "rewards/rejected": -12.475909233093262, | |
| "semantic_entropy": 0.9366561770439148, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.6010928961748634, | |
| "grad_norm": 103.487924065158, | |
| "learning_rate": 4.106678321395433e-07, | |
| "logits/chosen": -1.149213194847107, | |
| "logits/rejected": -1.091584324836731, | |
| "logps/chosen": -0.944200873374939, | |
| "logps/rejected": -1.1400898694992065, | |
| "loss": 2.9059, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -9.442008972167969, | |
| "rewards/margins": 1.9588892459869385, | |
| "rewards/rejected": -11.400897979736328, | |
| "semantic_entropy": 0.9353906512260437, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.6120218579234973, | |
| "grad_norm": 134.66475933979152, | |
| "learning_rate": 3.9193566913562915e-07, | |
| "logits/chosen": -1.112555742263794, | |
| "logits/rejected": -1.1183173656463623, | |
| "logps/chosen": -0.9673782587051392, | |
| "logps/rejected": -1.330810308456421, | |
| "loss": 2.8642, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -9.673782348632812, | |
| "rewards/margins": 3.6343204975128174, | |
| "rewards/rejected": -13.30810260772705, | |
| "semantic_entropy": 0.8982528448104858, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6229508196721312, | |
| "grad_norm": 122.90153644042246, | |
| "learning_rate": 3.7336133469909623e-07, | |
| "logits/chosen": -1.2511584758758545, | |
| "logits/rejected": -1.2238370180130005, | |
| "logps/chosen": -0.9596298933029175, | |
| "logps/rejected": -1.3390361070632935, | |
| "loss": 2.5195, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -9.59630012512207, | |
| "rewards/margins": 3.794060468673706, | |
| "rewards/rejected": -13.390359878540039, | |
| "semantic_entropy": 0.9061362147331238, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 0.6338797814207651, | |
| "grad_norm": 155.0095981855771, | |
| "learning_rate": 3.549719567506076e-07, | |
| "logits/chosen": -1.190308690071106, | |
| "logits/rejected": -1.1568708419799805, | |
| "logps/chosen": -1.0390634536743164, | |
| "logps/rejected": -1.3806968927383423, | |
| "loss": 2.8348, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -10.390633583068848, | |
| "rewards/margins": 3.4163336753845215, | |
| "rewards/rejected": -13.806968688964844, | |
| "semantic_entropy": 0.8884953260421753, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.644808743169399, | |
| "grad_norm": 128.49285257186182, | |
| "learning_rate": 3.3679439308082774e-07, | |
| "logits/chosen": -1.1692029237747192, | |
| "logits/rejected": -1.172719120979309, | |
| "logps/chosen": -1.0511713027954102, | |
| "logps/rejected": -1.4550367593765259, | |
| "loss": 2.2826, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -10.511713027954102, | |
| "rewards/margins": 4.0386552810668945, | |
| "rewards/rejected": -14.55036735534668, | |
| "semantic_entropy": 0.878921389579773, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 0.6557377049180327, | |
| "grad_norm": 113.97800775884723, | |
| "learning_rate": 3.1885519212446716e-07, | |
| "logits/chosen": -1.2200191020965576, | |
| "logits/rejected": -1.209221601486206, | |
| "logps/chosen": -1.0968992710113525, | |
| "logps/rejected": -1.4734210968017578, | |
| "loss": 2.5534, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -10.968993186950684, | |
| "rewards/margins": 3.7652194499969482, | |
| "rewards/rejected": -14.734212875366211, | |
| "semantic_entropy": 0.8491713404655457, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6666666666666666, | |
| "grad_norm": 153.40901841678263, | |
| "learning_rate": 3.0118055418614295e-07, | |
| "logits/chosen": -1.210296392440796, | |
| "logits/rejected": -1.166100263595581, | |
| "logps/chosen": -1.0915873050689697, | |
| "logps/rejected": -1.5003201961517334, | |
| "loss": 2.7077, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -10.915873527526855, | |
| "rewards/margins": 4.087328910827637, | |
| "rewards/rejected": -15.003199577331543, | |
| "semantic_entropy": 0.851865291595459, | |
| "step": 305 | |
| }, | |
| { | |
| "epoch": 0.6775956284153005, | |
| "grad_norm": 188.92294277142176, | |
| "learning_rate": 2.83796293174686e-07, | |
| "logits/chosen": -1.144047498703003, | |
| "logits/rejected": -1.1552207469940186, | |
| "logps/chosen": -1.1141725778579712, | |
| "logps/rejected": -1.5596559047698975, | |
| "loss": 2.9239, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -11.141725540161133, | |
| "rewards/margins": 4.454832077026367, | |
| "rewards/rejected": -15.5965576171875, | |
| "semantic_entropy": 0.8461610078811646, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6885245901639344, | |
| "grad_norm": 158.1416950642293, | |
| "learning_rate": 2.6672779890178046e-07, | |
| "logits/chosen": -1.1891381740570068, | |
| "logits/rejected": -1.1934657096862793, | |
| "logps/chosen": -1.1909846067428589, | |
| "logps/rejected": -1.4780324697494507, | |
| "loss": 2.6254, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -11.9098482131958, | |
| "rewards/margins": 2.8704779148101807, | |
| "rewards/rejected": -14.780324935913086, | |
| "semantic_entropy": 0.8265172243118286, | |
| "step": 315 | |
| }, | |
| { | |
| "epoch": 0.6994535519125683, | |
| "grad_norm": 117.90491663822478, | |
| "learning_rate": 2.500000000000001e-07, | |
| "logits/chosen": -1.2632883787155151, | |
| "logits/rejected": -1.2213891744613647, | |
| "logps/chosen": -1.2055654525756836, | |
| "logps/rejected": -1.6077735424041748, | |
| "loss": 2.6783, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -12.05565357208252, | |
| "rewards/margins": 4.022082805633545, | |
| "rewards/rejected": -16.077735900878906, | |
| "semantic_entropy": 0.8192012906074524, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.7103825136612022, | |
| "grad_norm": 164.40962706267163, | |
| "learning_rate": 2.3363732751439923e-07, | |
| "logits/chosen": -1.2160775661468506, | |
| "logits/rejected": -1.205322027206421, | |
| "logps/chosen": -1.1661893129348755, | |
| "logps/rejected": -1.5277663469314575, | |
| "loss": 2.6624, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -11.661892890930176, | |
| "rewards/margins": 3.615769863128662, | |
| "rewards/rejected": -15.27766227722168, | |
| "semantic_entropy": 0.823623538017273, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.7213114754098361, | |
| "grad_norm": 107.16070244814924, | |
| "learning_rate": 2.1766367922083283e-07, | |
| "logits/chosen": -1.1548130512237549, | |
| "logits/rejected": -1.1359449625015259, | |
| "logps/chosen": -1.1202858686447144, | |
| "logps/rejected": -1.626604676246643, | |
| "loss": 2.5554, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -11.20285701751709, | |
| "rewards/margins": 5.063187599182129, | |
| "rewards/rejected": -16.26604461669922, | |
| "semantic_entropy": 0.8284898996353149, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.73224043715847, | |
| "grad_norm": 143.54805923440733, | |
| "learning_rate": 2.021023847231202e-07, | |
| "logits/chosen": -1.1354625225067139, | |
| "logits/rejected": -1.1055997610092163, | |
| "logps/chosen": -1.2420397996902466, | |
| "logps/rejected": -1.6342121362686157, | |
| "loss": 2.5194, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -12.42039680480957, | |
| "rewards/margins": 3.9217231273651123, | |
| "rewards/rejected": -16.342121124267578, | |
| "semantic_entropy": 0.8055655360221863, | |
| "step": 335 | |
| }, | |
| { | |
| "epoch": 0.7431693989071039, | |
| "grad_norm": 143.90601843179704, | |
| "learning_rate": 1.869761713800254e-07, | |
| "logits/chosen": -1.159055471420288, | |
| "logits/rejected": -1.1217933893203735, | |
| "logps/chosen": -1.216088056564331, | |
| "logps/rejected": -1.6224826574325562, | |
| "loss": 2.6871, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -12.160881996154785, | |
| "rewards/margins": 4.0639448165893555, | |
| "rewards/rejected": -16.22482681274414, | |
| "semantic_entropy": 0.7940818667411804, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7540983606557377, | |
| "grad_norm": 146.77126754239146, | |
| "learning_rate": 1.7230713111182164e-07, | |
| "logits/chosen": -1.2234665155410767, | |
| "logits/rejected": -1.2274234294891357, | |
| "logps/chosen": -1.2689043283462524, | |
| "logps/rejected": -1.6999661922454834, | |
| "loss": 2.6683, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -12.689043998718262, | |
| "rewards/margins": 4.310617923736572, | |
| "rewards/rejected": -16.99966049194336, | |
| "semantic_entropy": 0.7804916501045227, | |
| "step": 345 | |
| }, | |
| { | |
| "epoch": 0.7650273224043715, | |
| "grad_norm": 156.59132030322806, | |
| "learning_rate": 1.5811668813491696e-07, | |
| "logits/chosen": -1.2142775058746338, | |
| "logits/rejected": -1.1987954378128052, | |
| "logps/chosen": -1.1977766752243042, | |
| "logps/rejected": -1.5605593919754028, | |
| "loss": 2.6148, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -11.977766036987305, | |
| "rewards/margins": 3.62782621383667, | |
| "rewards/rejected": -15.605592727661133, | |
| "semantic_entropy": 0.8086638450622559, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7759562841530054, | |
| "grad_norm": 144.3372503884412, | |
| "learning_rate": 1.4442556767166369e-07, | |
| "logits/chosen": -1.1762679815292358, | |
| "logits/rejected": -1.1479089260101318, | |
| "logps/chosen": -1.193447470664978, | |
| "logps/rejected": -1.6105226278305054, | |
| "loss": 2.5697, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -11.934475898742676, | |
| "rewards/margins": 4.170751094818115, | |
| "rewards/rejected": -16.105228424072266, | |
| "semantic_entropy": 0.8080552220344543, | |
| "step": 355 | |
| }, | |
| { | |
| "epoch": 0.7868852459016393, | |
| "grad_norm": 148.5272314733488, | |
| "learning_rate": 1.312537656810549e-07, | |
| "logits/chosen": -1.135874629020691, | |
| "logits/rejected": -1.1405253410339355, | |
| "logps/chosen": -1.242232322692871, | |
| "logps/rejected": -1.6316314935684204, | |
| "loss": 2.629, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -12.422323226928711, | |
| "rewards/margins": 3.8939926624298096, | |
| "rewards/rejected": -16.316316604614258, | |
| "semantic_entropy": 0.8038153648376465, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7978142076502732, | |
| "grad_norm": 189.12080572010774, | |
| "learning_rate": 1.1862051965451214e-07, | |
| "logits/chosen": -1.216966152191162, | |
| "logits/rejected": -1.2227869033813477, | |
| "logps/chosen": -1.2868871688842773, | |
| "logps/rejected": -1.718698263168335, | |
| "loss": 2.5421, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -12.868871688842773, | |
| "rewards/margins": 4.318110466003418, | |
| "rewards/rejected": -17.18698501586914, | |
| "semantic_entropy": 0.7714166045188904, | |
| "step": 365 | |
| }, | |
| { | |
| "epoch": 0.8087431693989071, | |
| "grad_norm": 130.34992194933218, | |
| "learning_rate": 1.0654428051942138e-07, | |
| "logits/chosen": -1.2347556352615356, | |
| "logits/rejected": -1.2042248249053955, | |
| "logps/chosen": -1.3064343929290771, | |
| "logps/rejected": -1.777515172958374, | |
| "loss": 2.6716, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -13.06434440612793, | |
| "rewards/margins": 4.710807800292969, | |
| "rewards/rejected": -17.7751522064209, | |
| "semantic_entropy": 0.7663524150848389, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.819672131147541, | |
| "grad_norm": 120.52547478396558, | |
| "learning_rate": 9.504268569144763e-08, | |
| "logits/chosen": -1.2310011386871338, | |
| "logits/rejected": -1.179221749305725, | |
| "logps/chosen": -1.2675514221191406, | |
| "logps/rejected": -1.7146999835968018, | |
| "loss": 2.5273, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -12.675516128540039, | |
| "rewards/margins": 4.4714837074279785, | |
| "rewards/rejected": -17.14699935913086, | |
| "semantic_entropy": 0.7719418406486511, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.8306010928961749, | |
| "grad_norm": 134.85584452412903, | |
| "learning_rate": 8.413253331499049e-08, | |
| "logits/chosen": -1.1171958446502686, | |
| "logits/rejected": -1.1349446773529053, | |
| "logps/chosen": -1.2904092073440552, | |
| "logps/rejected": -1.6963565349578857, | |
| "loss": 2.5116, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -12.904090881347656, | |
| "rewards/margins": 4.059475898742676, | |
| "rewards/rejected": -16.96356773376465, | |
| "semantic_entropy": 0.7810186147689819, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8415300546448088, | |
| "grad_norm": 137.9880749898867, | |
| "learning_rate": 7.382975772939865e-08, | |
| "logits/chosen": -1.219100832939148, | |
| "logits/rejected": -1.2109915018081665, | |
| "logps/chosen": -1.3701751232147217, | |
| "logps/rejected": -1.809171438217163, | |
| "loss": 2.8231, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -13.701749801635742, | |
| "rewards/margins": 4.389963626861572, | |
| "rewards/rejected": -18.09171485900879, | |
| "semantic_entropy": 0.7575483918190002, | |
| "step": 385 | |
| }, | |
| { | |
| "epoch": 0.8524590163934426, | |
| "grad_norm": 192.7458723286622, | |
| "learning_rate": 6.414940619677734e-08, | |
| "logits/chosen": -1.2024834156036377, | |
| "logits/rejected": -1.1897705793380737, | |
| "logps/chosen": -1.2859004735946655, | |
| "logps/rejected": -1.8171262741088867, | |
| "loss": 2.4732, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -12.85900592803955, | |
| "rewards/margins": 5.312258243560791, | |
| "rewards/rejected": -18.171262741088867, | |
| "semantic_entropy": 0.77564936876297, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8633879781420765, | |
| "grad_norm": 182.36527261020933, | |
| "learning_rate": 5.5105616925376296e-08, | |
| "logits/chosen": -1.1923558712005615, | |
| "logits/rejected": -1.1765515804290771, | |
| "logps/chosen": -1.3574202060699463, | |
| "logps/rejected": -1.7072795629501343, | |
| "loss": 2.4702, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -13.574200630187988, | |
| "rewards/margins": 3.4985928535461426, | |
| "rewards/rejected": -17.07279396057129, | |
| "semantic_entropy": 0.7606626749038696, | |
| "step": 395 | |
| }, | |
| { | |
| "epoch": 0.8743169398907104, | |
| "grad_norm": 156.32689835666156, | |
| "learning_rate": 4.6711598420656976e-08, | |
| "logits/chosen": -1.144141435623169, | |
| "logits/rejected": -1.1195942163467407, | |
| "logps/chosen": -1.3408299684524536, | |
| "logps/rejected": -1.8310232162475586, | |
| "loss": 2.3262, | |
| "rewards/accuracies": 0.8687499761581421, | |
| "rewards/chosen": -13.408299446105957, | |
| "rewards/margins": 4.901931285858154, | |
| "rewards/rejected": -18.310230255126953, | |
| "semantic_entropy": 0.7442251443862915, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8743169398907104, | |
| "eval_logits/chosen": -1.3813101053237915, | |
| "eval_logits/rejected": -1.3473328351974487, | |
| "eval_logps/chosen": -1.2879669666290283, | |
| "eval_logps/rejected": -1.739721655845642, | |
| "eval_loss": 2.5607941150665283, | |
| "eval_rewards/accuracies": 0.8072289228439331, | |
| "eval_rewards/chosen": -12.879671096801758, | |
| "eval_rewards/margins": 4.517546653747559, | |
| "eval_rewards/rejected": -17.397218704223633, | |
| "eval_runtime": 36.6792, | |
| "eval_samples_per_second": 35.933, | |
| "eval_semantic_entropy": 0.7719007730484009, | |
| "eval_steps_per_second": 2.263, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8852459016393442, | |
| "grad_norm": 148.89570610138972, | |
| "learning_rate": 3.897961019419516e-08, | |
| "logits/chosen": -1.1572606563568115, | |
| "logits/rejected": -1.1004865169525146, | |
| "logps/chosen": -1.2273566722869873, | |
| "logps/rejected": -1.640681266784668, | |
| "loss": 2.4505, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -12.273566246032715, | |
| "rewards/margins": 4.13324499130249, | |
| "rewards/rejected": -16.406810760498047, | |
| "semantic_entropy": 0.7908967137336731, | |
| "step": 405 | |
| }, | |
| { | |
| "epoch": 0.8961748633879781, | |
| "grad_norm": 135.13513698899922, | |
| "learning_rate": 3.192094485859526e-08, | |
| "logits/chosen": -1.164189100265503, | |
| "logits/rejected": -1.1957439184188843, | |
| "logps/chosen": -1.340441346168518, | |
| "logps/rejected": -1.8204329013824463, | |
| "loss": 2.5453, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -13.404413223266602, | |
| "rewards/margins": 4.799916744232178, | |
| "rewards/rejected": -18.204330444335938, | |
| "semantic_entropy": 0.7424927949905396, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.907103825136612, | |
| "grad_norm": 138.1252447502212, | |
| "learning_rate": 2.5545911634565265e-08, | |
| "logits/chosen": -1.2133328914642334, | |
| "logits/rejected": -1.217905879020691, | |
| "logps/chosen": -1.3356597423553467, | |
| "logps/rejected": -1.8359012603759766, | |
| "loss": 2.719, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -13.356595993041992, | |
| "rewards/margins": 5.002416133880615, | |
| "rewards/rejected": -18.359012603759766, | |
| "semantic_entropy": 0.7578593492507935, | |
| "step": 415 | |
| }, | |
| { | |
| "epoch": 0.9180327868852459, | |
| "grad_norm": 133.36843556957237, | |
| "learning_rate": 1.9863821294241522e-08, | |
| "logits/chosen": -1.2045748233795166, | |
| "logits/rejected": -1.188718557357788, | |
| "logps/chosen": -1.3029712438583374, | |
| "logps/rejected": -1.7860324382781982, | |
| "loss": 2.3706, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -13.029711723327637, | |
| "rewards/margins": 4.830612659454346, | |
| "rewards/rejected": -17.86032485961914, | |
| "semantic_entropy": 0.7500916123390198, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9289617486338798, | |
| "grad_norm": 139.42311056578998, | |
| "learning_rate": 1.4882972562753615e-08, | |
| "logits/chosen": -1.2088757753372192, | |
| "logits/rejected": -1.2028796672821045, | |
| "logps/chosen": -1.4182978868484497, | |
| "logps/rejected": -1.899735450744629, | |
| "loss": 2.6805, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -14.182978630065918, | |
| "rewards/margins": 4.814376354217529, | |
| "rewards/rejected": -18.997356414794922, | |
| "semantic_entropy": 0.7167907953262329, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.9398907103825137, | |
| "grad_norm": 138.05690419269808, | |
| "learning_rate": 1.0610639997888915e-08, | |
| "logits/chosen": -1.1360652446746826, | |
| "logits/rejected": -1.1394312381744385, | |
| "logps/chosen": -1.265080451965332, | |
| "logps/rejected": -1.7739555835723877, | |
| "loss": 2.2297, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -12.65080451965332, | |
| "rewards/margins": 5.0887532234191895, | |
| "rewards/rejected": -17.73955535888672, | |
| "semantic_entropy": 0.7776240110397339, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9508196721311475, | |
| "grad_norm": 156.31606318677163, | |
| "learning_rate": 7.053063365559997e-09, | |
| "logits/chosen": -1.1962147951126099, | |
| "logits/rejected": -1.2261369228363037, | |
| "logps/chosen": -1.330643892288208, | |
| "logps/rejected": -1.8400895595550537, | |
| "loss": 2.397, | |
| "rewards/accuracies": 0.8687499761581421, | |
| "rewards/chosen": -13.306437492370605, | |
| "rewards/margins": 5.094459056854248, | |
| "rewards/rejected": -18.400897979736328, | |
| "semantic_entropy": 0.7367173433303833, | |
| "step": 435 | |
| }, | |
| { | |
| "epoch": 0.9617486338797814, | |
| "grad_norm": 137.08276067025474, | |
| "learning_rate": 4.215438526591064e-09, | |
| "logits/chosen": -1.1730583906173706, | |
| "logits/rejected": -1.140211582183838, | |
| "logps/chosen": -1.3757076263427734, | |
| "logps/rejected": -1.7452484369277954, | |
| "loss": 2.5885, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -13.757074356079102, | |
| "rewards/margins": 3.6954073905944824, | |
| "rewards/rejected": -17.452484130859375, | |
| "semantic_entropy": 0.7384070754051208, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9726775956284153, | |
| "grad_norm": 153.84420514847153, | |
| "learning_rate": 2.1019098481337426e-09, | |
| "logits/chosen": -1.2059695720672607, | |
| "logits/rejected": -1.1932657957077026, | |
| "logps/chosen": -1.2917059659957886, | |
| "logps/rejected": -1.7843118906021118, | |
| "loss": 2.4083, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -12.917058944702148, | |
| "rewards/margins": 4.926059246063232, | |
| "rewards/rejected": -17.843120574951172, | |
| "semantic_entropy": 0.7641780376434326, | |
| "step": 445 | |
| }, | |
| { | |
| "epoch": 0.9836065573770492, | |
| "grad_norm": 169.02746393168292, | |
| "learning_rate": 7.155641507955445e-10, | |
| "logits/chosen": -1.122040033340454, | |
| "logits/rejected": -1.1226763725280762, | |
| "logps/chosen": -1.3791415691375732, | |
| "logps/rejected": -1.798413872718811, | |
| "loss": 2.6799, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -13.791415214538574, | |
| "rewards/margins": 4.192723274230957, | |
| "rewards/rejected": -17.98413848876953, | |
| "semantic_entropy": 0.7384847402572632, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.994535519125683, | |
| "grad_norm": 160.91531446057363, | |
| "learning_rate": 5.842620032053824e-11, | |
| "logits/chosen": -1.1409598588943481, | |
| "logits/rejected": -1.139664649963379, | |
| "logps/chosen": -1.3967525959014893, | |
| "logps/rejected": -1.73647940158844, | |
| "loss": 2.866, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -13.96752643585205, | |
| "rewards/margins": 3.3972675800323486, | |
| "rewards/rejected": -17.364791870117188, | |
| "semantic_entropy": 0.7418167591094971, | |
| "step": 455 | |
| }, | |
| { | |
| "epoch": 0.9989071038251366, | |
| "step": 457, | |
| "total_flos": 0.0, | |
| "train_loss": 3.4314852449513107, | |
| "train_runtime": 5934.8281, | |
| "train_samples_per_second": 9.867, | |
| "train_steps_per_second": 0.077 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 457, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000000, | |
| "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": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |