| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9994767137624281, | |
| "eval_steps": 100, | |
| "global_step": 955, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 71.7172276228935, | |
| "learning_rate": 5.208333333333333e-09, | |
| "logits/chosen": -1.4981693029403687, | |
| "logits/rejected": -1.5654948949813843, | |
| "logps/chosen": -257.4665222167969, | |
| "logps/rejected": -99.09321594238281, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 72.96425106358538, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -1.875927209854126, | |
| "logits/rejected": -1.8456478118896484, | |
| "logps/chosen": -326.4024658203125, | |
| "logps/rejected": -141.96922302246094, | |
| "loss": 0.6926, | |
| "rewards/accuracies": 0.5138888955116272, | |
| "rewards/chosen": 4.941297447658144e-05, | |
| "rewards/margins": 0.001581083401106298, | |
| "rewards/rejected": -0.0015316703356802464, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 67.94798943175728, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -1.8996143341064453, | |
| "logits/rejected": -1.8235642910003662, | |
| "logps/chosen": -323.2792663574219, | |
| "logps/rejected": -135.71456909179688, | |
| "loss": 0.6788, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 0.013530316762626171, | |
| "rewards/margins": 0.030556190758943558, | |
| "rewards/rejected": -0.01702587492763996, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 50.35196678274095, | |
| "learning_rate": 1.5624999999999999e-07, | |
| "logits/chosen": -1.9016050100326538, | |
| "logits/rejected": -1.8865772485733032, | |
| "logps/chosen": -306.1601257324219, | |
| "logps/rejected": -138.81069946289062, | |
| "loss": 0.6075, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 0.08184734731912613, | |
| "rewards/margins": 0.19270142912864685, | |
| "rewards/rejected": -0.11085411161184311, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 24.96243157624135, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -1.9789104461669922, | |
| "logits/rejected": -2.000354766845703, | |
| "logps/chosen": -269.0250549316406, | |
| "logps/rejected": -152.91973876953125, | |
| "loss": 0.5009, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 0.12138372659683228, | |
| "rewards/margins": 0.45498305559158325, | |
| "rewards/rejected": -0.3335992991924286, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 14.96717932132009, | |
| "learning_rate": 2.604166666666667e-07, | |
| "logits/chosen": -2.158332109451294, | |
| "logits/rejected": -2.16444993019104, | |
| "logps/chosen": -280.4437255859375, | |
| "logps/rejected": -197.60487365722656, | |
| "loss": 0.4282, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 0.15608879923820496, | |
| "rewards/margins": 0.867284893989563, | |
| "rewards/rejected": -0.7111960053443909, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 10.138930051409568, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.3020527362823486, | |
| "logits/rejected": -2.289559841156006, | |
| "logps/chosen": -255.1080780029297, | |
| "logps/rejected": -206.7790069580078, | |
| "loss": 0.3883, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -0.01837761141359806, | |
| "rewards/margins": 0.9224237203598022, | |
| "rewards/rejected": -0.9408015012741089, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 9.627361782454214, | |
| "learning_rate": 3.645833333333333e-07, | |
| "logits/chosen": -2.3991897106170654, | |
| "logits/rejected": -2.4038243293762207, | |
| "logps/chosen": -295.4974670410156, | |
| "logps/rejected": -273.0234375, | |
| "loss": 0.3499, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.15266235172748566, | |
| "rewards/margins": 1.2677773237228394, | |
| "rewards/rejected": -1.4204397201538086, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 14.313296662619264, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.5073745250701904, | |
| "logits/rejected": -2.553334951400757, | |
| "logps/chosen": -332.49945068359375, | |
| "logps/rejected": -319.8727722167969, | |
| "loss": 0.3368, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.5055555105209351, | |
| "rewards/margins": 1.468584418296814, | |
| "rewards/rejected": -1.9741401672363281, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 23.431306250567218, | |
| "learning_rate": 4.6874999999999996e-07, | |
| "logits/chosen": -2.611389636993408, | |
| "logits/rejected": -2.643505334854126, | |
| "logps/chosen": -355.7128601074219, | |
| "logps/rejected": -361.992431640625, | |
| "loss": 0.309, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.61676025390625, | |
| "rewards/margins": 1.7021185159683228, | |
| "rewards/rejected": -2.318878412246704, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 12.356735796295311, | |
| "learning_rate": 4.999732492681437e-07, | |
| "logits/chosen": -2.621342420578003, | |
| "logits/rejected": -2.6476845741271973, | |
| "logps/chosen": -319.81988525390625, | |
| "logps/rejected": -322.2472229003906, | |
| "loss": 0.3047, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.6059430241584778, | |
| "rewards/margins": 1.4836266040802002, | |
| "rewards/rejected": -2.089569568634033, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "eval_logits/chosen": -2.44050931930542, | |
| "eval_logits/rejected": -2.3863425254821777, | |
| "eval_logps/chosen": -361.08013916015625, | |
| "eval_logps/rejected": -337.77484130859375, | |
| "eval_loss": 0.8551393747329712, | |
| "eval_rewards/accuracies": 0.3203125, | |
| "eval_rewards/chosen": -0.49303656816482544, | |
| "eval_rewards/margins": -0.29053980112075806, | |
| "eval_rewards/rejected": -0.20249679684638977, | |
| "eval_runtime": 97.3683, | |
| "eval_samples_per_second": 20.541, | |
| "eval_steps_per_second": 0.329, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 12.540254444715838, | |
| "learning_rate": 4.996723692767926e-07, | |
| "logits/chosen": -2.7024824619293213, | |
| "logits/rejected": -2.732452154159546, | |
| "logps/chosen": -409.3314208984375, | |
| "logps/rejected": -413.891357421875, | |
| "loss": 0.2958, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.829090416431427, | |
| "rewards/margins": 2.046325206756592, | |
| "rewards/rejected": -2.875415325164795, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 15.24853315289221, | |
| "learning_rate": 4.990375746213598e-07, | |
| "logits/chosen": -2.661564350128174, | |
| "logits/rejected": -2.7124838829040527, | |
| "logps/chosen": -378.96533203125, | |
| "logps/rejected": -433.1641540527344, | |
| "loss": 0.2884, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.9071281552314758, | |
| "rewards/margins": 2.1523900032043457, | |
| "rewards/rejected": -3.059518337249756, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 20.41295055769396, | |
| "learning_rate": 4.980697142834314e-07, | |
| "logits/chosen": -2.693155527114868, | |
| "logits/rejected": -2.738525390625, | |
| "logps/chosen": -401.2513122558594, | |
| "logps/rejected": -464.12353515625, | |
| "loss": 0.2528, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -1.0387437343597412, | |
| "rewards/margins": 2.259091854095459, | |
| "rewards/rejected": -3.2978355884552, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 27.736891370306594, | |
| "learning_rate": 4.967700826904229e-07, | |
| "logits/chosen": -2.6829416751861572, | |
| "logits/rejected": -2.743159770965576, | |
| "logps/chosen": -405.8708801269531, | |
| "logps/rejected": -488.6322326660156, | |
| "loss": 0.2529, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.1785838603973389, | |
| "rewards/margins": 2.3818531036376953, | |
| "rewards/rejected": -3.560436964035034, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 31.193716735224292, | |
| "learning_rate": 4.951404179843962e-07, | |
| "logits/chosen": -2.4764091968536377, | |
| "logits/rejected": -2.481703281402588, | |
| "logps/chosen": -504.5975036621094, | |
| "logps/rejected": -567.6092529296875, | |
| "loss": 0.2358, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -1.8161567449569702, | |
| "rewards/margins": 2.5039477348327637, | |
| "rewards/rejected": -4.320104598999023, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 16.84967305265633, | |
| "learning_rate": 4.931828996974498e-07, | |
| "logits/chosen": -2.4284980297088623, | |
| "logits/rejected": -2.4476141929626465, | |
| "logps/chosen": -411.03546142578125, | |
| "logps/rejected": -520.0714721679688, | |
| "loss": 0.2377, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -1.3872876167297363, | |
| "rewards/margins": 2.5429630279541016, | |
| "rewards/rejected": -3.930250883102417, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 25.1927910072169, | |
| "learning_rate": 4.909001458367866e-07, | |
| "logits/chosen": -2.2736942768096924, | |
| "logits/rejected": -2.292642831802368, | |
| "logps/chosen": -452.697998046875, | |
| "logps/rejected": -563.0865478515625, | |
| "loss": 0.2128, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -1.6114835739135742, | |
| "rewards/margins": 2.781120538711548, | |
| "rewards/rejected": -4.392604351043701, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 18.107740978814483, | |
| "learning_rate": 4.882952093833627e-07, | |
| "logits/chosen": -2.1593220233917236, | |
| "logits/rejected": -2.1532785892486572, | |
| "logps/chosen": -509.1954650878906, | |
| "logps/rejected": -644.2286376953125, | |
| "loss": 0.2037, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -1.7663100957870483, | |
| "rewards/margins": 3.2817070484161377, | |
| "rewards/rejected": -5.0480170249938965, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 22.004607228534194, | |
| "learning_rate": 4.853715742087946e-07, | |
| "logits/chosen": -2.1366734504699707, | |
| "logits/rejected": -2.1492080688476562, | |
| "logps/chosen": -469.94976806640625, | |
| "logps/rejected": -587.4523315429688, | |
| "loss": 0.2059, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -1.6838188171386719, | |
| "rewards/margins": 2.9172627925872803, | |
| "rewards/rejected": -4.601081848144531, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 23.839892866837005, | |
| "learning_rate": 4.821331504159906e-07, | |
| "logits/chosen": -2.0906856060028076, | |
| "logits/rejected": -2.1124696731567383, | |
| "logps/chosen": -466.5874938964844, | |
| "logps/rejected": -608.2426147460938, | |
| "loss": 0.1861, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -1.8450790643692017, | |
| "rewards/margins": 3.033769130706787, | |
| "rewards/rejected": -4.878848552703857, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": -1.5417730808258057, | |
| "eval_logits/rejected": -1.5107452869415283, | |
| "eval_logps/chosen": -450.2716369628906, | |
| "eval_logps/rejected": -421.0933532714844, | |
| "eval_loss": 1.0494639873504639, | |
| "eval_rewards/accuracies": 0.38671875, | |
| "eval_rewards/chosen": -1.3849513530731201, | |
| "eval_rewards/margins": -0.3492693305015564, | |
| "eval_rewards/rejected": -1.0356820821762085, | |
| "eval_runtime": 97.2226, | |
| "eval_samples_per_second": 20.571, | |
| "eval_steps_per_second": 0.329, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 21.766108201350445, | |
| "learning_rate": 4.785842691097342e-07, | |
| "logits/chosen": -1.914615273475647, | |
| "logits/rejected": -1.9393634796142578, | |
| "logps/chosen": -541.3153076171875, | |
| "logps/rejected": -766.5374145507812, | |
| "loss": 0.1892, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.552523374557495, | |
| "rewards/margins": 3.7902703285217285, | |
| "rewards/rejected": -6.342793941497803, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 37.7465298870174, | |
| "learning_rate": 4.7472967660421603e-07, | |
| "logits/chosen": -2.004246234893799, | |
| "logits/rejected": -2.0251355171203613, | |
| "logps/chosen": -528.6552734375, | |
| "logps/rejected": -743.7687377929688, | |
| "loss": 0.1798, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.033097743988037, | |
| "rewards/margins": 4.051178932189941, | |
| "rewards/rejected": -6.0842766761779785, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 37.93991066707489, | |
| "learning_rate": 4.705745280752585e-07, | |
| "logits/chosen": -1.9027751684188843, | |
| "logits/rejected": -1.8593933582305908, | |
| "logps/chosen": -495.2493591308594, | |
| "logps/rejected": -688.589599609375, | |
| "loss": 0.1765, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -1.9664674997329712, | |
| "rewards/margins": 3.703148603439331, | |
| "rewards/rejected": -5.66961669921875, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 19.738306178674453, | |
| "learning_rate": 4.6612438066572555e-07, | |
| "logits/chosen": -1.7326488494873047, | |
| "logits/rejected": -1.585479497909546, | |
| "logps/chosen": -627.2045288085938, | |
| "logps/rejected": -877.15234375, | |
| "loss": 0.164, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -3.1001508235931396, | |
| "rewards/margins": 4.346390247344971, | |
| "rewards/rejected": -7.446540832519531, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 20.99257000654717, | |
| "learning_rate": 4.6138518605333664e-07, | |
| "logits/chosen": -1.7733532190322876, | |
| "logits/rejected": -1.5429089069366455, | |
| "logps/chosen": -534.9285888671875, | |
| "logps/rejected": -766.8218383789062, | |
| "loss": 0.1779, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.3250820636749268, | |
| "rewards/margins": 4.1490068435668945, | |
| "rewards/rejected": -6.4740891456604, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 24.636105883207083, | |
| "learning_rate": 4.5636328249082514e-07, | |
| "logits/chosen": -1.7743794918060303, | |
| "logits/rejected": -1.6784734725952148, | |
| "logps/chosen": -502.83929443359375, | |
| "logps/rejected": -686.4469604492188, | |
| "loss": 0.1666, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -2.020374298095703, | |
| "rewards/margins": 3.5390701293945312, | |
| "rewards/rejected": -5.559444427490234, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 26.442049627508165, | |
| "learning_rate": 4.510653863290871e-07, | |
| "logits/chosen": -1.8089313507080078, | |
| "logits/rejected": -1.504288673400879, | |
| "logps/chosen": -553.7681884765625, | |
| "logps/rejected": -773.63671875, | |
| "loss": 0.163, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.3301241397857666, | |
| "rewards/margins": 4.031432628631592, | |
| "rewards/rejected": -6.361557483673096, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 21.752850388248913, | |
| "learning_rate": 4.4549858303465737e-07, | |
| "logits/chosen": -1.792066216468811, | |
| "logits/rejected": -1.5596303939819336, | |
| "logps/chosen": -521.0010986328125, | |
| "logps/rejected": -791.414794921875, | |
| "loss": 0.1514, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -2.1834633350372314, | |
| "rewards/margins": 4.415928363800049, | |
| "rewards/rejected": -6.599390983581543, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 21.62051159722618, | |
| "learning_rate": 4.396703177135261e-07, | |
| "logits/chosen": -1.573697805404663, | |
| "logits/rejected": -1.1857482194900513, | |
| "logps/chosen": -611.6795654296875, | |
| "logps/rejected": -855.2744140625, | |
| "loss": 0.1493, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.884495258331299, | |
| "rewards/margins": 4.296299934387207, | |
| "rewards/rejected": -7.180795192718506, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 16.812667194402408, | |
| "learning_rate": 4.335883851539693e-07, | |
| "logits/chosen": -1.738490343093872, | |
| "logits/rejected": -1.7563718557357788, | |
| "logps/chosen": -494.66253662109375, | |
| "logps/rejected": -731.6812744140625, | |
| "loss": 0.1608, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -1.9538425207138062, | |
| "rewards/margins": 3.9912962913513184, | |
| "rewards/rejected": -5.945138931274414, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "eval_logits/chosen": -1.4367032051086426, | |
| "eval_logits/rejected": -1.4021562337875366, | |
| "eval_logps/chosen": -454.9446105957031, | |
| "eval_logps/rejected": -422.9683532714844, | |
| "eval_loss": 1.0910040140151978, | |
| "eval_rewards/accuracies": 0.39453125, | |
| "eval_rewards/chosen": -1.4316813945770264, | |
| "eval_rewards/margins": -0.37724918127059937, | |
| "eval_rewards/rejected": -1.0544321537017822, | |
| "eval_runtime": 97.2015, | |
| "eval_samples_per_second": 20.576, | |
| "eval_steps_per_second": 0.329, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 23.495796172679743, | |
| "learning_rate": 4.272609194017105e-07, | |
| "logits/chosen": -1.7330036163330078, | |
| "logits/rejected": -1.6350265741348267, | |
| "logps/chosen": -519.0903930664062, | |
| "logps/rejected": -720.5287475585938, | |
| "loss": 0.1552, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.7445989847183228, | |
| "rewards/margins": 4.112780570983887, | |
| "rewards/rejected": -5.857378959655762, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 16.298880424996614, | |
| "learning_rate": 4.2069638288135547e-07, | |
| "logits/chosen": -1.7347389459609985, | |
| "logits/rejected": -1.7288787364959717, | |
| "logps/chosen": -510.26483154296875, | |
| "logps/rejected": -713.36572265625, | |
| "loss": 0.157, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -1.7514524459838867, | |
| "rewards/margins": 3.96923828125, | |
| "rewards/rejected": -5.720690727233887, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 19.47125814373079, | |
| "learning_rate": 4.139035550786494e-07, | |
| "logits/chosen": -1.1601016521453857, | |
| "logits/rejected": -0.8797602653503418, | |
| "logps/chosen": -555.39990234375, | |
| "logps/rejected": -796.0065307617188, | |
| "loss": 0.1561, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -2.3634750843048096, | |
| "rewards/margins": 4.329623222351074, | |
| "rewards/rejected": -6.693098545074463, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 28.613677806731356, | |
| "learning_rate": 4.0689152079869306e-07, | |
| "logits/chosen": -1.2572317123413086, | |
| "logits/rejected": -0.9607254862785339, | |
| "logps/chosen": -606.7630615234375, | |
| "logps/rejected": -846.2305908203125, | |
| "loss": 0.1416, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.5160038471221924, | |
| "rewards/margins": 4.53468132019043, | |
| "rewards/rejected": -7.050685882568359, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 17.707474904024725, | |
| "learning_rate": 3.99669658015821e-07, | |
| "logits/chosen": -1.3649070262908936, | |
| "logits/rejected": -0.9824110269546509, | |
| "logps/chosen": -558.0640869140625, | |
| "logps/rejected": -775.701171875, | |
| "loss": 0.1563, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -2.48624849319458, | |
| "rewards/margins": 4.034348964691162, | |
| "rewards/rejected": -6.5205979347229, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 17.329304967511433, | |
| "learning_rate": 3.92247625331392e-07, | |
| "logits/chosen": -1.5912964344024658, | |
| "logits/rejected": -1.5140416622161865, | |
| "logps/chosen": -473.7013244628906, | |
| "logps/rejected": -705.4163818359375, | |
| "loss": 0.1391, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -2.0327391624450684, | |
| "rewards/margins": 3.767451047897339, | |
| "rewards/rejected": -5.800190448760986, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 21.039638710046997, | |
| "learning_rate": 3.846353490562664e-07, | |
| "logits/chosen": -1.3319923877716064, | |
| "logits/rejected": -1.090010404586792, | |
| "logps/chosen": -520.6598510742188, | |
| "logps/rejected": -788.8143310546875, | |
| "loss": 0.1403, | |
| "rewards/accuracies": 0.8687499761581421, | |
| "rewards/chosen": -2.3732516765594482, | |
| "rewards/margins": 4.287914276123047, | |
| "rewards/rejected": -6.661165714263916, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 19.41739908042072, | |
| "learning_rate": 3.768430099352445e-07, | |
| "logits/chosen": -1.2049062252044678, | |
| "logits/rejected": -0.9995657205581665, | |
| "logps/chosen": -531.4547119140625, | |
| "logps/rejected": -863.2029418945312, | |
| "loss": 0.1359, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.5220770835876465, | |
| "rewards/margins": 4.725777626037598, | |
| "rewards/rejected": -7.247855186462402, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 30.846420024935966, | |
| "learning_rate": 3.6888102953122304e-07, | |
| "logits/chosen": -1.4293053150177002, | |
| "logits/rejected": -1.129652976989746, | |
| "logps/chosen": -509.592529296875, | |
| "logps/rejected": -834.6383056640625, | |
| "loss": 0.1217, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.1002399921417236, | |
| "rewards/margins": 4.87478494644165, | |
| "rewards/rejected": -6.975025177001953, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 26.99256302878051, | |
| "learning_rate": 3.607600562872785e-07, | |
| "logits/chosen": -1.3965580463409424, | |
| "logits/rejected": -0.9727011919021606, | |
| "logps/chosen": -591.9832763671875, | |
| "logps/rejected": -926.43701171875, | |
| "loss": 0.1368, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -2.8865580558776855, | |
| "rewards/margins": 5.026923656463623, | |
| "rewards/rejected": -7.913481712341309, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": -1.053754210472107, | |
| "eval_logits/rejected": -1.013121485710144, | |
| "eval_logps/chosen": -520.169921875, | |
| "eval_logps/rejected": -479.64556884765625, | |
| "eval_loss": 1.3009716272354126, | |
| "eval_rewards/accuracies": 0.41015625, | |
| "eval_rewards/chosen": -2.0839340686798096, | |
| "eval_rewards/margins": -0.46272987127304077, | |
| "eval_rewards/rejected": -1.6212042570114136, | |
| "eval_runtime": 97.208, | |
| "eval_samples_per_second": 20.574, | |
| "eval_steps_per_second": 0.329, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 18.264964860581, | |
| "learning_rate": 3.5249095128531856e-07, | |
| "logits/chosen": -1.6575086116790771, | |
| "logits/rejected": -1.4926228523254395, | |
| "logps/chosen": -505.78448486328125, | |
| "logps/rejected": -805.1259155273438, | |
| "loss": 0.151, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -2.203605890274048, | |
| "rewards/margins": 4.460430145263672, | |
| "rewards/rejected": -6.664034843444824, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 19.09047976589624, | |
| "learning_rate": 3.4408477372034736e-07, | |
| "logits/chosen": -1.6263633966445923, | |
| "logits/rejected": -1.4956812858581543, | |
| "logps/chosen": -479.7705993652344, | |
| "logps/rejected": -766.1605224609375, | |
| "loss": 0.1448, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -2.093071937561035, | |
| "rewards/margins": 4.394181728363037, | |
| "rewards/rejected": -6.4872541427612305, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 26.836781627013515, | |
| "learning_rate": 3.3555276610977276e-07, | |
| "logits/chosen": -1.662540078163147, | |
| "logits/rejected": -1.4674957990646362, | |
| "logps/chosen": -527.6272583007812, | |
| "logps/rejected": -827.05224609375, | |
| "loss": 0.1477, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.207035541534424, | |
| "rewards/margins": 4.785927772521973, | |
| "rewards/rejected": -6.9929633140563965, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 24.71697296764681, | |
| "learning_rate": 3.269063392575352e-07, | |
| "logits/chosen": -1.682995080947876, | |
| "logits/rejected": -1.625732660293579, | |
| "logps/chosen": -521.5386352539062, | |
| "logps/rejected": -861.8421630859375, | |
| "loss": 0.1431, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.1410486698150635, | |
| "rewards/margins": 5.073235511779785, | |
| "rewards/rejected": -7.214284420013428, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 35.22220664866878, | |
| "learning_rate": 3.1815705699316964e-07, | |
| "logits/chosen": -1.578249216079712, | |
| "logits/rejected": -1.4307564496994019, | |
| "logps/chosen": -618.6419067382812, | |
| "logps/rejected": -917.3260498046875, | |
| "loss": 0.1288, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -3.068641185760498, | |
| "rewards/margins": 4.769894599914551, | |
| "rewards/rejected": -7.838536262512207, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 15.368374650378705, | |
| "learning_rate": 3.0931662070620794e-07, | |
| "logits/chosen": -1.8674280643463135, | |
| "logits/rejected": -1.8419768810272217, | |
| "logps/chosen": -574.4186401367188, | |
| "logps/rejected": -899.6975708007812, | |
| "loss": 0.1317, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.381072998046875, | |
| "rewards/margins": 5.151331424713135, | |
| "rewards/rejected": -7.53240442276001, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 32.45699366987169, | |
| "learning_rate": 3.003968536966078e-07, | |
| "logits/chosen": -1.779637098312378, | |
| "logits/rejected": -1.7716691493988037, | |
| "logps/chosen": -548.851806640625, | |
| "logps/rejected": -895.9444580078125, | |
| "loss": 0.1467, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.4585914611816406, | |
| "rewards/margins": 5.207429885864258, | |
| "rewards/rejected": -7.666021823883057, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 20.390767179507876, | |
| "learning_rate": 2.9140968536213693e-07, | |
| "logits/chosen": -1.919952630996704, | |
| "logits/rejected": -1.9492056369781494, | |
| "logps/chosen": -536.4271240234375, | |
| "logps/rejected": -884.7716064453125, | |
| "loss": 0.1285, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.535160541534424, | |
| "rewards/margins": 5.057857990264893, | |
| "rewards/rejected": -7.593018531799316, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 23.067489784018093, | |
| "learning_rate": 2.823671352438608e-07, | |
| "logits/chosen": -1.9916563034057617, | |
| "logits/rejected": -2.021616220474243, | |
| "logps/chosen": -518.0189208984375, | |
| "logps/rejected": -824.4202880859375, | |
| "loss": 0.1367, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -2.1029324531555176, | |
| "rewards/margins": 4.754405498504639, | |
| "rewards/rejected": -6.857337951660156, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 25.320584902838384, | |
| "learning_rate": 2.73281296951072e-07, | |
| "logits/chosen": -1.8573198318481445, | |
| "logits/rejected": -1.8858684301376343, | |
| "logps/chosen": -541.9103393554688, | |
| "logps/rejected": -806.6531372070312, | |
| "loss": 0.1364, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.396160840988159, | |
| "rewards/margins": 4.32682466506958, | |
| "rewards/rejected": -6.722985744476318, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "eval_logits/chosen": -1.6466461420059204, | |
| "eval_logits/rejected": -1.6090116500854492, | |
| "eval_logps/chosen": -470.0933532714844, | |
| "eval_logps/rejected": -430.8614196777344, | |
| "eval_loss": 1.1773370504379272, | |
| "eval_rewards/accuracies": 0.37109375, | |
| "eval_rewards/chosen": -1.5831681489944458, | |
| "eval_rewards/margins": -0.44980528950691223, | |
| "eval_rewards/rejected": -1.133362889289856, | |
| "eval_runtime": 97.142, | |
| "eval_samples_per_second": 20.588, | |
| "eval_steps_per_second": 0.329, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 22.389152396843905, | |
| "learning_rate": 2.641643219871597e-07, | |
| "logits/chosen": -1.9270741939544678, | |
| "logits/rejected": -1.8253233432769775, | |
| "logps/chosen": -544.8509521484375, | |
| "logps/rejected": -834.5119018554688, | |
| "loss": 0.1273, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.5420873165130615, | |
| "rewards/margins": 4.583073616027832, | |
| "rewards/rejected": -7.125161170959473, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 29.91148010858245, | |
| "learning_rate": 2.550284034980507e-07, | |
| "logits/chosen": -1.8461729288101196, | |
| "logits/rejected": -1.9183883666992188, | |
| "logps/chosen": -536.2467041015625, | |
| "logps/rejected": -846.56689453125, | |
| "loss": 0.1301, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.3210761547088623, | |
| "rewards/margins": 4.834765911102295, | |
| "rewards/rejected": -7.1558427810668945, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 24.244708458502263, | |
| "learning_rate": 2.4588575996495794e-07, | |
| "logits/chosen": -1.80156672000885, | |
| "logits/rejected": -1.815596580505371, | |
| "logps/chosen": -494.86798095703125, | |
| "logps/rejected": -798.2820434570312, | |
| "loss": 0.1237, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -2.3462131023406982, | |
| "rewards/margins": 4.387810230255127, | |
| "rewards/rejected": -6.7340240478515625, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 19.910786436303987, | |
| "learning_rate": 2.367486188632446e-07, | |
| "logits/chosen": -1.9172916412353516, | |
| "logits/rejected": -1.9127076864242554, | |
| "logps/chosen": -525.8649291992188, | |
| "logps/rejected": -845.7303466796875, | |
| "loss": 0.1138, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.059016704559326, | |
| "rewards/margins": 4.951017379760742, | |
| "rewards/rejected": -7.010035037994385, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 26.12771211985874, | |
| "learning_rate": 2.276292003092593e-07, | |
| "logits/chosen": -1.8340644836425781, | |
| "logits/rejected": -1.8532531261444092, | |
| "logps/chosen": -547.7066650390625, | |
| "logps/rejected": -828.9231567382812, | |
| "loss": 0.1309, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.3669962882995605, | |
| "rewards/margins": 4.561534404754639, | |
| "rewards/rejected": -6.928530216217041, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 23.903525728545898, | |
| "learning_rate": 2.185397007170141e-07, | |
| "logits/chosen": -1.7198997735977173, | |
| "logits/rejected": -1.5677220821380615, | |
| "logps/chosen": -525.4856567382812, | |
| "logps/rejected": -855.2352294921875, | |
| "loss": 0.1157, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.4956746101379395, | |
| "rewards/margins": 4.861161708831787, | |
| "rewards/rejected": -7.356836795806885, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 24.886700178362616, | |
| "learning_rate": 2.094922764865619e-07, | |
| "logits/chosen": -1.7758210897445679, | |
| "logits/rejected": -1.836538314819336, | |
| "logps/chosen": -506.49053955078125, | |
| "logps/rejected": -846.2745971679688, | |
| "loss": 0.1366, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -2.2135753631591797, | |
| "rewards/margins": 4.973353385925293, | |
| "rewards/rejected": -7.186928749084473, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 21.212825758656418, | |
| "learning_rate": 2.0049902774588797e-07, | |
| "logits/chosen": -1.7731355428695679, | |
| "logits/rejected": -1.777611494064331, | |
| "logps/chosen": -581.9010009765625, | |
| "logps/rejected": -871.5733642578125, | |
| "loss": 0.123, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -2.585883378982544, | |
| "rewards/margins": 4.742143154144287, | |
| "rewards/rejected": -7.32802677154541, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 24.501331083650943, | |
| "learning_rate": 1.9157198216806238e-07, | |
| "logits/chosen": -1.7602859735488892, | |
| "logits/rejected": -1.84162175655365, | |
| "logps/chosen": -452.37420654296875, | |
| "logps/rejected": -707.836181640625, | |
| "loss": 0.1305, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.025653600692749, | |
| "rewards/margins": 3.844447374343872, | |
| "rewards/rejected": -5.870100975036621, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 31.082842710792463, | |
| "learning_rate": 1.8272307888529274e-07, | |
| "logits/chosen": -1.5975382328033447, | |
| "logits/rejected": -1.4109210968017578, | |
| "logps/chosen": -539.9990234375, | |
| "logps/rejected": -849.1795654296875, | |
| "loss": 0.1223, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -2.722317934036255, | |
| "rewards/margins": 4.622250556945801, | |
| "rewards/rejected": -7.344568729400635, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": -1.1879667043685913, | |
| "eval_logits/rejected": -1.1541048288345337, | |
| "eval_logps/chosen": -541.48828125, | |
| "eval_logps/rejected": -500.4929504394531, | |
| "eval_loss": 1.3205512762069702, | |
| "eval_rewards/accuracies": 0.4140625, | |
| "eval_rewards/chosen": -2.2971181869506836, | |
| "eval_rewards/margins": -0.46744003891944885, | |
| "eval_rewards/rejected": -1.829677939414978, | |
| "eval_runtime": 97.2287, | |
| "eval_samples_per_second": 20.57, | |
| "eval_steps_per_second": 0.329, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 25.74237035768959, | |
| "learning_rate": 1.7396415252139288e-07, | |
| "logits/chosen": -1.5792185068130493, | |
| "logits/rejected": -1.6271593570709229, | |
| "logps/chosen": -576.2874755859375, | |
| "logps/rejected": -917.5999755859375, | |
| "loss": 0.1228, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -2.802475929260254, | |
| "rewards/margins": 5.048992156982422, | |
| "rewards/rejected": -7.851468563079834, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 29.519623894853613, | |
| "learning_rate": 1.6530691736402316e-07, | |
| "logits/chosen": -1.6389156579971313, | |
| "logits/rejected": -1.6191425323486328, | |
| "logps/chosen": -530.8707275390625, | |
| "logps/rejected": -890.0936279296875, | |
| "loss": 0.1254, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.3094658851623535, | |
| "rewards/margins": 5.2504377365112305, | |
| "rewards/rejected": -7.559903144836426, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 50.42769593704326, | |
| "learning_rate": 1.5676295169786864e-07, | |
| "logits/chosen": -1.7326772212982178, | |
| "logits/rejected": -1.5138975381851196, | |
| "logps/chosen": -561.8060302734375, | |
| "logps/rejected": -941.2083740234375, | |
| "loss": 0.1236, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -2.4162166118621826, | |
| "rewards/margins": 5.7043328285217285, | |
| "rewards/rejected": -8.120549201965332, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 27.38270007193824, | |
| "learning_rate": 1.483436823197092e-07, | |
| "logits/chosen": -1.5837154388427734, | |
| "logits/rejected": -1.4245226383209229, | |
| "logps/chosen": -562.3756103515625, | |
| "logps/rejected": -949.3406982421875, | |
| "loss": 0.1163, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -3.040762186050415, | |
| "rewards/margins": 5.183028697967529, | |
| "rewards/rejected": -8.223790168762207, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 35.27601401211646, | |
| "learning_rate": 1.4006036925609243e-07, | |
| "logits/chosen": -1.7516834735870361, | |
| "logits/rejected": -1.7616602182388306, | |
| "logps/chosen": -558.3771362304688, | |
| "logps/rejected": -879.00341796875, | |
| "loss": 0.1139, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -2.68957257270813, | |
| "rewards/margins": 4.818731784820557, | |
| "rewards/rejected": -7.508304595947266, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 22.77394327537456, | |
| "learning_rate": 1.319240907040458e-07, | |
| "logits/chosen": -1.7446712255477905, | |
| "logits/rejected": -1.6567370891571045, | |
| "logps/chosen": -559.461669921875, | |
| "logps/rejected": -941.3621826171875, | |
| "loss": 0.1117, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -2.522426128387451, | |
| "rewards/margins": 5.564127445220947, | |
| "rewards/rejected": -8.086553573608398, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 30.36694985683218, | |
| "learning_rate": 1.239457282149695e-07, | |
| "logits/chosen": -1.4815702438354492, | |
| "logits/rejected": -1.3673474788665771, | |
| "logps/chosen": -548.4630737304688, | |
| "logps/rejected": -930.783203125, | |
| "loss": 0.1092, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -2.9377596378326416, | |
| "rewards/margins": 5.2314453125, | |
| "rewards/rejected": -8.169205665588379, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 28.757425831930576, | |
| "learning_rate": 1.1613595214152711e-07, | |
| "logits/chosen": -1.5986506938934326, | |
| "logits/rejected": -1.5326309204101562, | |
| "logps/chosen": -513.4288940429688, | |
| "logps/rejected": -861.1954345703125, | |
| "loss": 0.1105, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -2.518094062805176, | |
| "rewards/margins": 4.961692810058594, | |
| "rewards/rejected": -7.4797868728637695, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 23.459702688906162, | |
| "learning_rate": 1.0850520736699362e-07, | |
| "logits/chosen": -1.6289697885513306, | |
| "logits/rejected": -1.527553677558899, | |
| "logps/chosen": -537.8997192382812, | |
| "logps/rejected": -878.1143798828125, | |
| "loss": 0.113, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -2.5290098190307617, | |
| "rewards/margins": 4.979551315307617, | |
| "rewards/rejected": -7.508561611175537, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 25.531163769236304, | |
| "learning_rate": 1.0106369933615042e-07, | |
| "logits/chosen": -1.5353871583938599, | |
| "logits/rejected": -1.3200981616973877, | |
| "logps/chosen": -662.8204956054688, | |
| "logps/rejected": -1062.7938232421875, | |
| "loss": 0.0971, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -3.5107734203338623, | |
| "rewards/margins": 5.729162693023682, | |
| "rewards/rejected": -9.239934921264648, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "eval_logits/chosen": -0.9711934924125671, | |
| "eval_logits/rejected": -0.9391850829124451, | |
| "eval_logps/chosen": -577.3128051757812, | |
| "eval_logps/rejected": -533.4666748046875, | |
| "eval_loss": 1.463817834854126, | |
| "eval_rewards/accuracies": 0.390625, | |
| "eval_rewards/chosen": -2.6553633213043213, | |
| "eval_rewards/margins": -0.4959481358528137, | |
| "eval_rewards/rejected": -2.1594154834747314, | |
| "eval_runtime": 97.1589, | |
| "eval_samples_per_second": 20.585, | |
| "eval_steps_per_second": 0.329, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 43.118837668077326, | |
| "learning_rate": 9.382138040640714e-08, | |
| "logits/chosen": -1.6447147130966187, | |
| "logits/rejected": -1.4446865320205688, | |
| "logps/chosen": -644.3197021484375, | |
| "logps/rejected": -1070.079345703125, | |
| "loss": 0.11, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -3.044210910797119, | |
| "rewards/margins": 6.259350776672363, | |
| "rewards/rejected": -9.303561210632324, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 32.87349411722058, | |
| "learning_rate": 8.678793653740632e-08, | |
| "logits/chosen": -1.7681348323822021, | |
| "logits/rejected": -1.6581518650054932, | |
| "logps/chosen": -622.3300170898438, | |
| "logps/rejected": -1050.2943115234375, | |
| "loss": 0.1069, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.925344944000244, | |
| "rewards/margins": 6.121405601501465, | |
| "rewards/rejected": -9.04675006866455, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 36.992204810711705, | |
| "learning_rate": 7.997277433690983e-08, | |
| "logits/chosen": -1.623159646987915, | |
| "logits/rejected": -1.4811782836914062, | |
| "logps/chosen": -612.2554931640625, | |
| "logps/rejected": -995.8507690429688, | |
| "loss": 0.118, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -3.1586315631866455, | |
| "rewards/margins": 5.526102066040039, | |
| "rewards/rejected": -8.684733390808105, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 31.945634867150687, | |
| "learning_rate": 7.338500848029602e-08, | |
| "logits/chosen": -1.51155686378479, | |
| "logits/rejected": -1.349172592163086, | |
| "logps/chosen": -575.31103515625, | |
| "logps/rejected": -1048.4927978515625, | |
| "loss": 0.1135, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -3.0343711376190186, | |
| "rewards/margins": 6.1768388748168945, | |
| "rewards/rejected": -9.211209297180176, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 26.030459327741653, | |
| "learning_rate": 6.70334495204884e-08, | |
| "logits/chosen": -1.6075376272201538, | |
| "logits/rejected": -1.2717111110687256, | |
| "logps/chosen": -635.0384521484375, | |
| "logps/rejected": -1016.2355346679688, | |
| "loss": 0.0957, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -3.0112128257751465, | |
| "rewards/margins": 5.901360511779785, | |
| "rewards/rejected": -8.912572860717773, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 34.627238331382905, | |
| "learning_rate": 6.092659210462231e-08, | |
| "logits/chosen": -1.448529839515686, | |
| "logits/rejected": -1.4409586191177368, | |
| "logps/chosen": -625.353271484375, | |
| "logps/rejected": -1024.951416015625, | |
| "loss": 0.1095, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -3.2976653575897217, | |
| "rewards/margins": 5.690510272979736, | |
| "rewards/rejected": -8.988176345825195, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 29.968791403480516, | |
| "learning_rate": 5.507260361320737e-08, | |
| "logits/chosen": -1.5798556804656982, | |
| "logits/rejected": -1.5097671747207642, | |
| "logps/chosen": -600.1854248046875, | |
| "logps/rejected": -997.0427856445312, | |
| "loss": 0.1033, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": -3.1436641216278076, | |
| "rewards/margins": 5.432265281677246, | |
| "rewards/rejected": -8.575929641723633, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 23.15982358200479, | |
| "learning_rate": 4.947931323697982e-08, | |
| "logits/chosen": -1.4408910274505615, | |
| "logits/rejected": -1.3072535991668701, | |
| "logps/chosen": -556.1472778320312, | |
| "logps/rejected": -968.1648559570312, | |
| "loss": 0.0969, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.2009377479553223, | |
| "rewards/margins": 5.246421813964844, | |
| "rewards/rejected": -8.447359085083008, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 23.663576108460163, | |
| "learning_rate": 4.415420150605398e-08, | |
| "logits/chosen": -1.5656832456588745, | |
| "logits/rejected": -1.3935534954071045, | |
| "logps/chosen": -601.1109008789062, | |
| "logps/rejected": -999.2574462890625, | |
| "loss": 0.1023, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -3.033841609954834, | |
| "rewards/margins": 5.674118995666504, | |
| "rewards/rejected": -8.70796012878418, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 23.462966354133016, | |
| "learning_rate": 3.9104390285376374e-08, | |
| "logits/chosen": -1.466312289237976, | |
| "logits/rejected": -1.2542184591293335, | |
| "logps/chosen": -619.60009765625, | |
| "logps/rejected": -1041.6527099609375, | |
| "loss": 0.1035, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.135887861251831, | |
| "rewards/margins": 5.890833854675293, | |
| "rewards/rejected": -9.026721954345703, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": -0.9231772422790527, | |
| "eval_logits/rejected": -0.8901588916778564, | |
| "eval_logps/chosen": -569.3816528320312, | |
| "eval_logps/rejected": -532.9067993164062, | |
| "eval_loss": 1.4475353956222534, | |
| "eval_rewards/accuracies": 0.39453125, | |
| "eval_rewards/chosen": -2.5760514736175537, | |
| "eval_rewards/margins": -0.42223480343818665, | |
| "eval_rewards/rejected": -2.1538166999816895, | |
| "eval_runtime": 97.1382, | |
| "eval_samples_per_second": 20.589, | |
| "eval_steps_per_second": 0.329, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 30.508507386848592, | |
| "learning_rate": 3.433663324986208e-08, | |
| "logits/chosen": -1.4120627641677856, | |
| "logits/rejected": -1.3580305576324463, | |
| "logps/chosen": -642.4114990234375, | |
| "logps/rejected": -1046.66357421875, | |
| "loss": 0.1086, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": -3.5207436084747314, | |
| "rewards/margins": 5.544968605041504, | |
| "rewards/rejected": -9.065712928771973, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 15.450308092025814, | |
| "learning_rate": 2.9857306851953897e-08, | |
| "logits/chosen": -1.5075926780700684, | |
| "logits/rejected": -1.286818504333496, | |
| "logps/chosen": -577.96533203125, | |
| "logps/rejected": -1041.372802734375, | |
| "loss": 0.0844, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.9870734214782715, | |
| "rewards/margins": 6.117280960083008, | |
| "rewards/rejected": -9.104352951049805, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 32.428690976585585, | |
| "learning_rate": 2.567240179368185e-08, | |
| "logits/chosen": -1.5189392566680908, | |
| "logits/rejected": -1.3501880168914795, | |
| "logps/chosen": -603.621337890625, | |
| "logps/rejected": -1007.1226806640625, | |
| "loss": 0.0982, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -3.046630620956421, | |
| "rewards/margins": 5.773746967315674, | |
| "rewards/rejected": -8.820378303527832, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 35.07586716910879, | |
| "learning_rate": 2.1787515014630357e-08, | |
| "logits/chosen": -1.406670331954956, | |
| "logits/rejected": -1.1134653091430664, | |
| "logps/chosen": -552.0911254882812, | |
| "logps/rejected": -1012.5986328125, | |
| "loss": 0.1141, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.7988593578338623, | |
| "rewards/margins": 6.124402046203613, | |
| "rewards/rejected": -8.923261642456055, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 35.12000542720853, | |
| "learning_rate": 1.820784220652766e-08, | |
| "logits/chosen": -1.5011646747589111, | |
| "logits/rejected": -1.4681178331375122, | |
| "logps/chosen": -629.8143310546875, | |
| "logps/rejected": -1053.095703125, | |
| "loss": 0.0947, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -3.13895583152771, | |
| "rewards/margins": 5.925333499908447, | |
| "rewards/rejected": -9.064289093017578, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 37.505156397985395, | |
| "learning_rate": 1.4938170864468636e-08, | |
| "logits/chosen": -1.5276473760604858, | |
| "logits/rejected": -1.3087430000305176, | |
| "logps/chosen": -611.999267578125, | |
| "logps/rejected": -982.8528442382812, | |
| "loss": 0.1089, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.328166961669922, | |
| "rewards/margins": 5.248373031616211, | |
| "rewards/rejected": -8.576539993286133, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 33.59787382725122, | |
| "learning_rate": 1.1982873884064465e-08, | |
| "logits/chosen": -1.5041674375534058, | |
| "logits/rejected": -1.186408281326294, | |
| "logps/chosen": -543.9867553710938, | |
| "logps/rejected": -960.0665893554688, | |
| "loss": 0.1023, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -2.6006431579589844, | |
| "rewards/margins": 5.71505069732666, | |
| "rewards/rejected": -8.315693855285645, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 24.198197782455043, | |
| "learning_rate": 9.345903713082304e-09, | |
| "logits/chosen": -1.5898497104644775, | |
| "logits/rejected": -1.4362118244171143, | |
| "logps/chosen": -581.5143432617188, | |
| "logps/rejected": -1014.4615478515625, | |
| "loss": 0.0837, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.9443726539611816, | |
| "rewards/margins": 5.875336647033691, | |
| "rewards/rejected": -8.819709777832031, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 32.4653792997969, | |
| "learning_rate": 7.030787065396865e-09, | |
| "logits/chosen": -1.5886590480804443, | |
| "logits/rejected": -1.2676339149475098, | |
| "logps/chosen": -612.6450805664062, | |
| "logps/rejected": -982.3916015625, | |
| "loss": 0.1056, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -3.1008591651916504, | |
| "rewards/margins": 5.476729393005371, | |
| "rewards/rejected": -8.57758903503418, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 28.89242092867656, | |
| "learning_rate": 5.04062020432286e-09, | |
| "logits/chosen": -1.5399090051651, | |
| "logits/rejected": -1.42655611038208, | |
| "logps/chosen": -595.1466064453125, | |
| "logps/rejected": -1033.9515380859375, | |
| "loss": 0.088, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -2.967661142349243, | |
| "rewards/margins": 6.027240753173828, | |
| "rewards/rejected": -8.994901657104492, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "eval_logits/chosen": -1.1238834857940674, | |
| "eval_logits/rejected": -1.0823395252227783, | |
| "eval_logps/chosen": -554.9180297851562, | |
| "eval_logps/rejected": -517.7515869140625, | |
| "eval_loss": 1.3947275876998901, | |
| "eval_rewards/accuracies": 0.38671875, | |
| "eval_rewards/chosen": -2.43141508102417, | |
| "eval_rewards/margins": -0.4291508197784424, | |
| "eval_rewards/rejected": -2.0022642612457275, | |
| "eval_runtime": 97.126, | |
| "eval_samples_per_second": 20.592, | |
| "eval_steps_per_second": 0.329, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 26.806730695970433, | |
| "learning_rate": 3.3780648016376866e-09, | |
| "logits/chosen": -1.658847451210022, | |
| "logits/rejected": -1.3762922286987305, | |
| "logps/chosen": -603.6043701171875, | |
| "logps/rejected": -1017.3605346679688, | |
| "loss": 0.0898, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.103484630584717, | |
| "rewards/margins": 5.833575248718262, | |
| "rewards/rejected": -8.93705940246582, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 15.136819931212546, | |
| "learning_rate": 2.0453443778310766e-09, | |
| "logits/chosen": -1.627918004989624, | |
| "logits/rejected": -1.4132583141326904, | |
| "logps/chosen": -563.746337890625, | |
| "logps/rejected": -1013.9581909179688, | |
| "loss": 0.0936, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -2.8101205825805664, | |
| "rewards/margins": 6.073071479797363, | |
| "rewards/rejected": -8.88319206237793, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 36.0925338348187, | |
| "learning_rate": 1.0442413283435758e-09, | |
| "logits/chosen": -1.5613000392913818, | |
| "logits/rejected": -1.3666173219680786, | |
| "logps/chosen": -583.2642822265625, | |
| "logps/rejected": -1070.8175048828125, | |
| "loss": 0.0947, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -3.103260040283203, | |
| "rewards/margins": 6.282073020935059, | |
| "rewards/rejected": -9.385333061218262, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 30.829530604495584, | |
| "learning_rate": 3.760945397705828e-10, | |
| "logits/chosen": -1.6098768711090088, | |
| "logits/rejected": -1.4137991666793823, | |
| "logps/chosen": -652.3986206054688, | |
| "logps/rejected": -1045.8623046875, | |
| "loss": 0.1041, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.067474365234375, | |
| "rewards/margins": 6.041169166564941, | |
| "rewards/rejected": -9.108643531799316, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 28.95439781379653, | |
| "learning_rate": 4.17975992204056e-11, | |
| "logits/chosen": -1.6131422519683838, | |
| "logits/rejected": -1.4869133234024048, | |
| "logps/chosen": -605.0281372070312, | |
| "logps/rejected": -1031.686279296875, | |
| "loss": 0.0954, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.1735126972198486, | |
| "rewards/margins": 5.8094305992126465, | |
| "rewards/rejected": -8.982942581176758, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 0.05187356284775659, | |
| "train_runtime": 7314.4586, | |
| "train_samples_per_second": 16.716, | |
| "train_steps_per_second": 0.131 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 955, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |