| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 19.45945945945946, | |
| "eval_steps": 500, | |
| "global_step": 540, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.36036036036036034, | |
| "grad_norm": 4.098762512207031, | |
| "learning_rate": 9.259259259259259e-07, | |
| "logits/chosen": -2.332144260406494, | |
| "logits/rejected": -2.3385167121887207, | |
| "logps/chosen": -80.89369201660156, | |
| "logps/rejected": -70.11573791503906, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": 0.00037987352698110044, | |
| "rewards/margins": 0.004227532539516687, | |
| "rewards/rejected": -0.003847658634185791, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.7207207207207207, | |
| "grad_norm": 3.641324281692505, | |
| "learning_rate": 1.8518518518518519e-06, | |
| "logits/chosen": -2.323789119720459, | |
| "logits/rejected": -2.351041793823242, | |
| "logps/chosen": -73.2725601196289, | |
| "logps/rejected": -81.80250549316406, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.0012983012711629272, | |
| "rewards/margins": 0.004207946360111237, | |
| "rewards/rejected": -0.0029096449725329876, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.0810810810810811, | |
| "grad_norm": 3.9276015758514404, | |
| "learning_rate": 2.7777777777777783e-06, | |
| "logits/chosen": -2.3353028297424316, | |
| "logits/rejected": -2.3445916175842285, | |
| "logps/chosen": -69.34381103515625, | |
| "logps/rejected": -74.37530517578125, | |
| "loss": 0.6941, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": -0.01035231165587902, | |
| "rewards/margins": -0.006389107555150986, | |
| "rewards/rejected": -0.00396320316940546, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.4414414414414414, | |
| "grad_norm": 4.809000492095947, | |
| "learning_rate": 3.7037037037037037e-06, | |
| "logits/chosen": -2.343184232711792, | |
| "logits/rejected": -2.360262393951416, | |
| "logps/chosen": -77.91002655029297, | |
| "logps/rejected": -76.27156066894531, | |
| "loss": 0.6902, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.008437180891633034, | |
| "rewards/margins": 0.008391124196350574, | |
| "rewards/rejected": -0.016828304156661034, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.8018018018018018, | |
| "grad_norm": 4.504117012023926, | |
| "learning_rate": 4.62962962962963e-06, | |
| "logits/chosen": -2.3394973278045654, | |
| "logits/rejected": -2.3635268211364746, | |
| "logps/chosen": -83.62376403808594, | |
| "logps/rejected": -267.64569091796875, | |
| "loss": 0.6851, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.01695835217833519, | |
| "rewards/margins": 0.14291717112064362, | |
| "rewards/rejected": -0.12595881521701813, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 2.1621621621621623, | |
| "grad_norm": 4.033559322357178, | |
| "learning_rate": 4.998119881260576e-06, | |
| "logits/chosen": -2.32966685295105, | |
| "logits/rejected": -2.3370490074157715, | |
| "logps/chosen": -78.54629516601562, | |
| "logps/rejected": -82.67992401123047, | |
| "loss": 0.6767, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.03485359251499176, | |
| "rewards/margins": 0.035757362842559814, | |
| "rewards/rejected": -0.07061095535755157, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 2.5225225225225225, | |
| "grad_norm": 4.979142189025879, | |
| "learning_rate": 4.9866405060165044e-06, | |
| "logits/chosen": -2.364291191101074, | |
| "logits/rejected": -2.376107931137085, | |
| "logps/chosen": -70.61842346191406, | |
| "logps/rejected": -81.80282592773438, | |
| "loss": 0.6636, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.06025966256856918, | |
| "rewards/margins": 0.03742799907922745, | |
| "rewards/rejected": -0.09768766909837723, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.8828828828828827, | |
| "grad_norm": 4.0428690910339355, | |
| "learning_rate": 4.964774158361991e-06, | |
| "logits/chosen": -2.3341965675354004, | |
| "logits/rejected": -2.3440909385681152, | |
| "logps/chosen": -86.3591537475586, | |
| "logps/rejected": -77.45347595214844, | |
| "loss": 0.6519, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.09531830251216888, | |
| "rewards/margins": 0.09854079782962799, | |
| "rewards/rejected": -0.19385910034179688, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 3.2432432432432434, | |
| "grad_norm": 4.31919527053833, | |
| "learning_rate": 4.93261217644956e-06, | |
| "logits/chosen": -2.351658821105957, | |
| "logits/rejected": -2.3437318801879883, | |
| "logps/chosen": -77.31346130371094, | |
| "logps/rejected": -80.43277740478516, | |
| "loss": 0.6243, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.1381106823682785, | |
| "rewards/margins": 0.16527524590492249, | |
| "rewards/rejected": -0.3033859133720398, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 3.6036036036036037, | |
| "grad_norm": 5.029748439788818, | |
| "learning_rate": 4.8902889044347e-06, | |
| "logits/chosen": -2.3354241847991943, | |
| "logits/rejected": -2.358518600463867, | |
| "logps/chosen": -75.03588104248047, | |
| "logps/rejected": -86.44483947753906, | |
| "loss": 0.6025, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.22239580750465393, | |
| "rewards/margins": 0.1877792775630951, | |
| "rewards/rejected": -0.41017502546310425, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 3.963963963963964, | |
| "grad_norm": 4.6208648681640625, | |
| "learning_rate": 4.837981131305475e-06, | |
| "logits/chosen": -2.3195366859436035, | |
| "logits/rejected": -2.3129196166992188, | |
| "logps/chosen": -72.09532928466797, | |
| "logps/rejected": -73.18878936767578, | |
| "loss": 0.5955, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.22240504622459412, | |
| "rewards/margins": 0.22803232073783875, | |
| "rewards/rejected": -0.45043739676475525, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 4.324324324324325, | |
| "grad_norm": 4.163040637969971, | |
| "learning_rate": 4.775907352415367e-06, | |
| "logits/chosen": -2.3427720069885254, | |
| "logits/rejected": -2.3731276988983154, | |
| "logps/chosen": -85.9415283203125, | |
| "logps/rejected": -93.53765869140625, | |
| "loss": 0.5506, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.24449148774147034, | |
| "rewards/margins": 0.3563767373561859, | |
| "rewards/rejected": -0.6008682250976562, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 4.684684684684685, | |
| "grad_norm": 4.228254795074463, | |
| "learning_rate": 4.70432685680402e-06, | |
| "logits/chosen": -2.336733341217041, | |
| "logits/rejected": -2.3446521759033203, | |
| "logps/chosen": -81.07231140136719, | |
| "logps/rejected": -90.82849884033203, | |
| "loss": 0.5248, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.005195322446525097, | |
| "rewards/margins": 0.6936509609222412, | |
| "rewards/rejected": -0.6988462209701538, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 5.045045045045045, | |
| "grad_norm": 4.454960346221924, | |
| "learning_rate": 4.623538644118244e-06, | |
| "logits/chosen": -2.3331754207611084, | |
| "logits/rejected": -2.3434836864471436, | |
| "logps/chosen": -83.67604064941406, | |
| "logps/rejected": -82.92774200439453, | |
| "loss": 0.5288, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.2716079652309418, | |
| "rewards/margins": 0.4593236446380615, | |
| "rewards/rejected": -0.7309316396713257, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 5.405405405405405, | |
| "grad_norm": 5.223482608795166, | |
| "learning_rate": 4.533880175657419e-06, | |
| "logits/chosen": -2.362809658050537, | |
| "logits/rejected": -2.3657679557800293, | |
| "logps/chosen": -73.20018768310547, | |
| "logps/rejected": -85.37998962402344, | |
| "loss": 0.4682, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.3221462368965149, | |
| "rewards/margins": 0.5968301892280579, | |
| "rewards/rejected": -0.9189764261245728, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 5.7657657657657655, | |
| "grad_norm": 4.905521869659424, | |
| "learning_rate": 4.435725964760331e-06, | |
| "logits/chosen": -2.3808655738830566, | |
| "logits/rejected": -2.368286609649658, | |
| "logps/chosen": -68.88943481445312, | |
| "logps/rejected": -82.69029235839844, | |
| "loss": 0.4586, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.3172217011451721, | |
| "rewards/margins": 0.7665462493896484, | |
| "rewards/rejected": -1.0837678909301758, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 6.126126126126126, | |
| "grad_norm": 5.399628162384033, | |
| "learning_rate": 4.329486012421531e-06, | |
| "logits/chosen": -2.365935802459717, | |
| "logits/rejected": -2.363004684448242, | |
| "logps/chosen": -70.47642517089844, | |
| "logps/rejected": -84.02542877197266, | |
| "loss": 0.4462, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.45835933089256287, | |
| "rewards/margins": 0.8438631892204285, | |
| "rewards/rejected": -1.302222490310669, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 6.486486486486487, | |
| "grad_norm": 4.843445777893066, | |
| "learning_rate": 4.215604094671835e-06, | |
| "logits/chosen": -2.357231855392456, | |
| "logits/rejected": -2.360239028930664, | |
| "logps/chosen": -78.67561340332031, | |
| "logps/rejected": -88.39659118652344, | |
| "loss": 0.3976, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.4842923581600189, | |
| "rewards/margins": 0.8027322888374329, | |
| "rewards/rejected": -1.2870244979858398, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 6.846846846846847, | |
| "grad_norm": 4.972764015197754, | |
| "learning_rate": 4.094555908876765e-06, | |
| "logits/chosen": -2.3751468658447266, | |
| "logits/rejected": -2.3993237018585205, | |
| "logps/chosen": -73.63652038574219, | |
| "logps/rejected": -278.0970458984375, | |
| "loss": 0.3959, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.4291106164455414, | |
| "rewards/margins": 0.9967883229255676, | |
| "rewards/rejected": -1.4258991479873657, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 7.207207207207207, | |
| "grad_norm": 5.071193218231201, | |
| "learning_rate": 3.966847086696045e-06, | |
| "logits/chosen": -2.3572330474853516, | |
| "logits/rejected": -2.357269763946533, | |
| "logps/chosen": -84.92713928222656, | |
| "logps/rejected": -98.15062713623047, | |
| "loss": 0.3544, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -0.5852295756340027, | |
| "rewards/margins": 1.2983506917953491, | |
| "rewards/rejected": -1.883580207824707, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 7.5675675675675675, | |
| "grad_norm": 5.1891655921936035, | |
| "learning_rate": 3.833011082004229e-06, | |
| "logits/chosen": -2.368424892425537, | |
| "logits/rejected": -2.378568649291992, | |
| "logps/chosen": -72.57874298095703, | |
| "logps/rejected": -84.37443542480469, | |
| "loss": 0.3421, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.48721614480018616, | |
| "rewards/margins": 1.2057541608810425, | |
| "rewards/rejected": -1.6929700374603271, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 7.927927927927928, | |
| "grad_norm": 5.771843433380127, | |
| "learning_rate": 3.693606942594873e-06, | |
| "logits/chosen": -2.3891513347625732, | |
| "logits/rejected": -2.4053854942321777, | |
| "logps/chosen": -75.97737121582031, | |
| "logps/rejected": -97.49588012695312, | |
| "loss": 0.3211, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.6163657903671265, | |
| "rewards/margins": 1.1816037893295288, | |
| "rewards/rejected": -1.7979698181152344, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 8.288288288288289, | |
| "grad_norm": 5.1563029289245605, | |
| "learning_rate": 3.549216974976073e-06, | |
| "logits/chosen": -2.4075605869293213, | |
| "logits/rejected": -2.406411647796631, | |
| "logps/chosen": -82.80142974853516, | |
| "logps/rejected": -96.36463928222656, | |
| "loss": 0.2848, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.8106307983398438, | |
| "rewards/margins": 1.647127389907837, | |
| "rewards/rejected": -2.4577584266662598, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 8.64864864864865, | |
| "grad_norm": 5.483398914337158, | |
| "learning_rate": 3.400444312011776e-06, | |
| "logits/chosen": -2.3797879219055176, | |
| "logits/rejected": -2.362518787384033, | |
| "logps/chosen": -82.14349365234375, | |
| "logps/rejected": -97.63994598388672, | |
| "loss": 0.278, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -0.9469457864761353, | |
| "rewards/margins": 1.488023281097412, | |
| "rewards/rejected": -2.434968948364258, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 9.00900900900901, | |
| "grad_norm": 5.042275905609131, | |
| "learning_rate": 3.2479103935691047e-06, | |
| "logits/chosen": -2.3207201957702637, | |
| "logits/rejected": -2.341810941696167, | |
| "logps/chosen": -85.28227233886719, | |
| "logps/rejected": -116.27372741699219, | |
| "loss": 0.2494, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.0121912956237793, | |
| "rewards/margins": 1.997532606124878, | |
| "rewards/rejected": -3.0097243785858154, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 9.36936936936937, | |
| "grad_norm": 5.468939781188965, | |
| "learning_rate": 3.092252370695298e-06, | |
| "logits/chosen": -2.3408374786376953, | |
| "logits/rejected": -2.366006851196289, | |
| "logps/chosen": -72.05101013183594, | |
| "logps/rejected": -102.21392822265625, | |
| "loss": 0.2457, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.0319160223007202, | |
| "rewards/margins": 1.8719971179962158, | |
| "rewards/rejected": -2.9039134979248047, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 9.72972972972973, | |
| "grad_norm": 6.745687007904053, | |
| "learning_rate": 2.9341204441673267e-06, | |
| "logits/chosen": -2.327451467514038, | |
| "logits/rejected": -2.3463993072509766, | |
| "logps/chosen": -86.53431701660156, | |
| "logps/rejected": -116.40992736816406, | |
| "loss": 0.2059, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.3356889486312866, | |
| "rewards/margins": 1.9991543292999268, | |
| "rewards/rejected": -3.334843397140503, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 10.09009009009009, | |
| "grad_norm": 5.230775833129883, | |
| "learning_rate": 2.7741751485313295e-06, | |
| "logits/chosen": -2.3630144596099854, | |
| "logits/rejected": -2.3630847930908203, | |
| "logps/chosen": -76.57563018798828, | |
| "logps/rejected": -99.20953369140625, | |
| "loss": 0.2034, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.2317285537719727, | |
| "rewards/margins": 1.8740953207015991, | |
| "rewards/rejected": -3.1058237552642822, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 10.45045045045045, | |
| "grad_norm": 6.581757545471191, | |
| "learning_rate": 2.6130845929767662e-06, | |
| "logits/chosen": -2.3247475624084473, | |
| "logits/rejected": -2.3450474739074707, | |
| "logps/chosen": -83.9271240234375, | |
| "logps/rejected": -109.54156494140625, | |
| "loss": 0.174, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.4542334079742432, | |
| "rewards/margins": 2.3127167224884033, | |
| "rewards/rejected": -3.7669498920440674, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 10.81081081081081, | |
| "grad_norm": 5.604727745056152, | |
| "learning_rate": 2.4515216705704396e-06, | |
| "logits/chosen": -2.279327869415283, | |
| "logits/rejected": -2.319913387298584, | |
| "logps/chosen": -78.63652801513672, | |
| "logps/rejected": -115.9185562133789, | |
| "loss": 0.1831, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -1.3030173778533936, | |
| "rewards/margins": 2.5270209312438965, | |
| "rewards/rejected": -3.830038547515869, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 11.17117117117117, | |
| "grad_norm": 4.8863606452941895, | |
| "learning_rate": 2.290161247507733e-06, | |
| "logits/chosen": -2.273766040802002, | |
| "logits/rejected": -2.3243603706359863, | |
| "logps/chosen": -90.69010925292969, | |
| "logps/rejected": -131.49423217773438, | |
| "loss": 0.1513, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.566362738609314, | |
| "rewards/margins": 3.1073012351989746, | |
| "rewards/rejected": -4.673664093017578, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 11.531531531531531, | |
| "grad_norm": 5.772294521331787, | |
| "learning_rate": 2.129677344121879e-06, | |
| "logits/chosen": -2.302643299102783, | |
| "logits/rejected": -2.3125722408294678, | |
| "logps/chosen": -78.91960144042969, | |
| "logps/rejected": -103.53559875488281, | |
| "loss": 0.1624, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.6784160137176514, | |
| "rewards/margins": 2.4029908180236816, | |
| "rewards/rejected": -4.081407070159912, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 11.891891891891891, | |
| "grad_norm": 5.915937423706055, | |
| "learning_rate": 1.970740319426474e-06, | |
| "logits/chosen": -2.275726795196533, | |
| "logits/rejected": -2.302337169647217, | |
| "logps/chosen": -99.52557373046875, | |
| "logps/rejected": -122.73197174072266, | |
| "loss": 0.1348, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.9647445678710938, | |
| "rewards/margins": 2.8115408420562744, | |
| "rewards/rejected": -4.776285648345947, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 12.252252252252251, | |
| "grad_norm": 5.65620231628418, | |
| "learning_rate": 1.8140140709517467e-06, | |
| "logits/chosen": -2.274402379989624, | |
| "logits/rejected": -2.2855653762817383, | |
| "logps/chosen": -86.69510650634766, | |
| "logps/rejected": -116.1146469116211, | |
| "loss": 0.1366, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -1.855790376663208, | |
| "rewards/margins": 2.536457061767578, | |
| "rewards/rejected": -4.392247200012207, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 12.612612612612612, | |
| "grad_norm": 4.642848491668701, | |
| "learning_rate": 1.6601532615711452e-06, | |
| "logits/chosen": -2.2652974128723145, | |
| "logits/rejected": -2.285008192062378, | |
| "logps/chosen": -89.00364685058594, | |
| "logps/rejected": -126.0003890991211, | |
| "loss": 0.1216, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.8776098489761353, | |
| "rewards/margins": 2.8089137077331543, | |
| "rewards/rejected": -4.6865234375, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 12.972972972972974, | |
| "grad_norm": 4.52380895614624, | |
| "learning_rate": 1.509800584902108e-06, | |
| "logits/chosen": -2.263986349105835, | |
| "logits/rejected": -2.2855420112609863, | |
| "logps/chosen": -91.05010986328125, | |
| "logps/rejected": -133.76361083984375, | |
| "loss": 0.1076, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.4504590034484863, | |
| "rewards/margins": 3.495802402496338, | |
| "rewards/rejected": -5.946260929107666, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 13.333333333333334, | |
| "grad_norm": 4.304037094116211, | |
| "learning_rate": 1.3635840807037487e-06, | |
| "logits/chosen": -2.261019229888916, | |
| "logits/rejected": -2.264559268951416, | |
| "logps/chosen": -93.27009582519531, | |
| "logps/rejected": -118.05653381347656, | |
| "loss": 0.1072, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.1757514476776123, | |
| "rewards/margins": 3.157127857208252, | |
| "rewards/rejected": -5.332879066467285, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 13.693693693693694, | |
| "grad_norm": 5.501009941101074, | |
| "learning_rate": 1.2221145114853172e-06, | |
| "logits/chosen": -2.211054563522339, | |
| "logits/rejected": -2.22572660446167, | |
| "logps/chosen": -90.1929702758789, | |
| "logps/rejected": -138.2399139404297, | |
| "loss": 0.0889, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.5342886447906494, | |
| "rewards/margins": 3.2334110736846924, | |
| "rewards/rejected": -5.767699241638184, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 14.054054054054054, | |
| "grad_norm": 5.037735939025879, | |
| "learning_rate": 1.085982811283654e-06, | |
| "logits/chosen": -2.2411131858825684, | |
| "logits/rejected": -2.261753559112549, | |
| "logps/chosen": -98.27137756347656, | |
| "logps/rejected": -134.04779052734375, | |
| "loss": 0.0971, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.593488931655884, | |
| "rewards/margins": 3.3826117515563965, | |
| "rewards/rejected": -5.976100444793701, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 14.414414414414415, | |
| "grad_norm": 5.258338451385498, | |
| "learning_rate": 9.557576172663577e-07, | |
| "logits/chosen": -2.244196653366089, | |
| "logits/rejected": -2.2605862617492676, | |
| "logps/chosen": -93.3480224609375, | |
| "logps/rejected": -145.70272827148438, | |
| "loss": 0.0913, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.5507616996765137, | |
| "rewards/margins": 3.9265968799591064, | |
| "rewards/rejected": -6.477358341217041, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 14.774774774774775, | |
| "grad_norm": 4.47573184967041, | |
| "learning_rate": 8.319828944714508e-07, | |
| "logits/chosen": -2.26932954788208, | |
| "logits/rejected": -2.274758815765381, | |
| "logps/chosen": -102.62467956542969, | |
| "logps/rejected": -132.93116760253906, | |
| "loss": 0.0805, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.376272201538086, | |
| "rewards/margins": 3.278709888458252, | |
| "rewards/rejected": -5.654982089996338, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 15.135135135135135, | |
| "grad_norm": 4.006404399871826, | |
| "learning_rate": 7.151756636052529e-07, | |
| "logits/chosen": -2.240022659301758, | |
| "logits/rejected": -2.2303688526153564, | |
| "logps/chosen": -112.4423828125, | |
| "logps/rejected": -143.22732543945312, | |
| "loss": 0.0796, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.971478223800659, | |
| "rewards/margins": 3.4153618812561035, | |
| "rewards/rejected": -6.3868408203125, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 15.495495495495495, | |
| "grad_norm": 4.521777629852295, | |
| "learning_rate": 6.058238413897052e-07, | |
| "logits/chosen": -2.1890573501586914, | |
| "logits/rejected": -2.241264820098877, | |
| "logps/chosen": -111.05755615234375, | |
| "logps/rejected": -146.14109802246094, | |
| "loss": 0.0724, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.0282044410705566, | |
| "rewards/margins": 3.6133179664611816, | |
| "rewards/rejected": -6.641521453857422, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 15.855855855855856, | |
| "grad_norm": 4.783228397369385, | |
| "learning_rate": 5.043842024802675e-07, | |
| "logits/chosen": -2.1972153186798096, | |
| "logits/rejected": -2.192469358444214, | |
| "logps/chosen": -102.45082092285156, | |
| "logps/rejected": -138.3446502685547, | |
| "loss": 0.0723, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.7581396102905273, | |
| "rewards/margins": 3.8694820404052734, | |
| "rewards/rejected": -6.627622127532959, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 16.216216216216218, | |
| "grad_norm": 4.521092891693115, | |
| "learning_rate": 4.1128047146765936e-07, | |
| "logits/chosen": -2.1869163513183594, | |
| "logits/rejected": -2.182262897491455, | |
| "logps/chosen": -104.68853759765625, | |
| "logps/rejected": -145.3485870361328, | |
| "loss": 0.075, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.978736400604248, | |
| "rewards/margins": 3.886040449142456, | |
| "rewards/rejected": -6.864776611328125, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 16.576576576576578, | |
| "grad_norm": 4.271122932434082, | |
| "learning_rate": 3.269015529333805e-07, | |
| "logits/chosen": -2.2191543579101562, | |
| "logits/rejected": -2.2385966777801514, | |
| "logps/chosen": -91.767578125, | |
| "logps/rejected": -135.73141479492188, | |
| "loss": 0.0701, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.994217872619629, | |
| "rewards/margins": 3.757829189300537, | |
| "rewards/rejected": -6.752047061920166, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 16.936936936936938, | |
| "grad_norm": 4.7400712966918945, | |
| "learning_rate": 2.515999069522676e-07, | |
| "logits/chosen": -2.174427032470703, | |
| "logits/rejected": -2.211961507797241, | |
| "logps/chosen": -100.193603515625, | |
| "logps/rejected": -139.8927001953125, | |
| "loss": 0.0682, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.3388049602508545, | |
| "rewards/margins": 3.942704677581787, | |
| "rewards/rejected": -7.281510353088379, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 17.2972972972973, | |
| "grad_norm": 4.059471130371094, | |
| "learning_rate": 1.8569007682777417e-07, | |
| "logits/chosen": -2.172400951385498, | |
| "logits/rejected": -2.2151694297790527, | |
| "logps/chosen": -109.02021789550781, | |
| "logps/rejected": -408.2373962402344, | |
| "loss": 0.0601, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.6841540336608887, | |
| "rewards/margins": 10.677932739257812, | |
| "rewards/rejected": -14.362088203430176, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 17.65765765765766, | |
| "grad_norm": 3.554147243499756, | |
| "learning_rate": 1.2944737520980883e-07, | |
| "logits/chosen": -2.2005763053894043, | |
| "logits/rejected": -2.234318494796753, | |
| "logps/chosen": -105.55352783203125, | |
| "logps/rejected": -156.26739501953125, | |
| "loss": 0.0667, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.3424956798553467, | |
| "rewards/margins": 3.9150681495666504, | |
| "rewards/rejected": -7.257563591003418, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 18.01801801801802, | |
| "grad_norm": 4.700870990753174, | |
| "learning_rate": 8.310673408334496e-08, | |
| "logits/chosen": -2.2091004848480225, | |
| "logits/rejected": -2.2198190689086914, | |
| "logps/chosen": -123.39643859863281, | |
| "logps/rejected": -159.49868774414062, | |
| "loss": 0.0679, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.437526226043701, | |
| "rewards/margins": 3.751330852508545, | |
| "rewards/rejected": -7.188857078552246, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 18.37837837837838, | |
| "grad_norm": 4.196156024932861, | |
| "learning_rate": 4.6861723431538273e-08, | |
| "logits/chosen": -2.166625738143921, | |
| "logits/rejected": -2.182325839996338, | |
| "logps/chosen": -105.13529968261719, | |
| "logps/rejected": -160.1414031982422, | |
| "loss": 0.0599, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.3666820526123047, | |
| "rewards/margins": 4.2599992752075195, | |
| "rewards/rejected": -7.626680850982666, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 18.73873873873874, | |
| "grad_norm": 4.409626483917236, | |
| "learning_rate": 2.0863742672497244e-08, | |
| "logits/chosen": -2.2174572944641113, | |
| "logits/rejected": -2.229252815246582, | |
| "logps/chosen": -111.33894348144531, | |
| "logps/rejected": -138.5430145263672, | |
| "loss": 0.0639, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.91447114944458, | |
| "rewards/margins": 3.7303109169006348, | |
| "rewards/rejected": -6.644782066345215, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 19.0990990990991, | |
| "grad_norm": 4.13588285446167, | |
| "learning_rate": 5.221388247169945e-09, | |
| "logits/chosen": -2.207942485809326, | |
| "logits/rejected": -2.219350814819336, | |
| "logps/chosen": -100.70853424072266, | |
| "logps/rejected": -141.94883728027344, | |
| "loss": 0.0683, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.074063777923584, | |
| "rewards/margins": 3.970613956451416, | |
| "rewards/rejected": -7.044677734375, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 19.45945945945946, | |
| "grad_norm": 3.9963550567626953, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -2.2123262882232666, | |
| "logits/rejected": -2.2226064205169678, | |
| "logps/chosen": -111.34256744384766, | |
| "logps/rejected": -149.1357421875, | |
| "loss": 0.0683, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.439054012298584, | |
| "rewards/margins": 4.026246070861816, | |
| "rewards/rejected": -7.4653000831604, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 19.45945945945946, | |
| "step": 540, | |
| "total_flos": 1.9727302677684552e+18, | |
| "train_loss": 0.2973077946239048, | |
| "train_runtime": 4131.3591, | |
| "train_samples_per_second": 8.573, | |
| "train_steps_per_second": 0.131 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 540, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.9727302677684552e+18, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |