| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9991111111111111, | |
| "eval_steps": 100, | |
| "global_step": 562, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0017777777777777779, | |
| "grad_norm": 0.22311078073034826, | |
| "learning_rate": 8.771929824561404e-08, | |
| "logits/chosen": 14.539060592651367, | |
| "logits/rejected": 15.870795249938965, | |
| "logps/chosen": -470.04345703125, | |
| "logps/rejected": -509.49163818359375, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.017777777777777778, | |
| "grad_norm": 0.1976540584216542, | |
| "learning_rate": 8.771929824561404e-07, | |
| "logits/chosen": 12.96641731262207, | |
| "logits/rejected": 13.155448913574219, | |
| "logps/chosen": -400.9219665527344, | |
| "logps/rejected": -399.60699462890625, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4513888955116272, | |
| "rewards/chosen": -0.00211474671959877, | |
| "rewards/margins": -0.001011626678518951, | |
| "rewards/rejected": -0.0011031196918338537, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.035555555555555556, | |
| "grad_norm": 0.21429385122960512, | |
| "learning_rate": 1.7543859649122807e-06, | |
| "logits/chosen": 13.817936897277832, | |
| "logits/rejected": 13.85405158996582, | |
| "logps/chosen": -420.9461975097656, | |
| "logps/rejected": -404.7037658691406, | |
| "loss": 0.6938, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": -6.537516310345381e-05, | |
| "rewards/margins": -0.0014555875677615404, | |
| "rewards/rejected": 0.0013902120990678668, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.05333333333333334, | |
| "grad_norm": 0.1898289423532557, | |
| "learning_rate": 2.631578947368421e-06, | |
| "logits/chosen": 13.255971908569336, | |
| "logits/rejected": 13.399693489074707, | |
| "logps/chosen": -402.4463806152344, | |
| "logps/rejected": -412.36480712890625, | |
| "loss": 0.6937, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.0020085538271814585, | |
| "rewards/margins": -0.0013788806973025203, | |
| "rewards/rejected": -0.00062967324629426, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.07111111111111111, | |
| "grad_norm": 0.2162185808879369, | |
| "learning_rate": 3.5087719298245615e-06, | |
| "logits/chosen": 13.550105094909668, | |
| "logits/rejected": 13.091300964355469, | |
| "logps/chosen": -379.90777587890625, | |
| "logps/rejected": -373.92138671875, | |
| "loss": 0.6935, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": 0.0008380211656913161, | |
| "rewards/margins": -0.00024208976537920535, | |
| "rewards/rejected": 0.0010801110183820128, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.08888888888888889, | |
| "grad_norm": 0.22810063623939997, | |
| "learning_rate": 4.385964912280702e-06, | |
| "logits/chosen": 13.239973068237305, | |
| "logits/rejected": 13.241312980651855, | |
| "logps/chosen": -396.4941101074219, | |
| "logps/rejected": -430.57049560546875, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.0004129047447349876, | |
| "rewards/margins": 4.834262654185295e-05, | |
| "rewards/rejected": -0.00046124737127684057, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.10666666666666667, | |
| "grad_norm": 0.2512545993725567, | |
| "learning_rate": 4.999564631597802e-06, | |
| "logits/chosen": 13.0819673538208, | |
| "logits/rejected": 13.337594985961914, | |
| "logps/chosen": -388.62628173828125, | |
| "logps/rejected": -414.85382080078125, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -0.002572793047875166, | |
| "rewards/margins": 0.0024341598618775606, | |
| "rewards/rejected": -0.005006952676922083, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.12444444444444444, | |
| "grad_norm": 0.22761734037588918, | |
| "learning_rate": 4.991828966534002e-06, | |
| "logits/chosen": 13.887951850891113, | |
| "logits/rejected": 13.595651626586914, | |
| "logps/chosen": -450.8296813964844, | |
| "logps/rejected": -444.7730407714844, | |
| "loss": 0.6936, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.006924263201653957, | |
| "rewards/margins": -0.001168354763649404, | |
| "rewards/rejected": -0.005755907855927944, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.14222222222222222, | |
| "grad_norm": 0.21722031386603705, | |
| "learning_rate": 4.974452899279292e-06, | |
| "logits/chosen": 13.351213455200195, | |
| "logits/rejected": 12.537083625793457, | |
| "logps/chosen": -425.934814453125, | |
| "logps/rejected": -383.2563781738281, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.015285441651940346, | |
| "rewards/margins": -0.0011091658379882574, | |
| "rewards/rejected": -0.014176277443766594, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 0.27601757784168, | |
| "learning_rate": 4.947503654462277e-06, | |
| "logits/chosen": 13.753236770629883, | |
| "logits/rejected": 13.5679292678833, | |
| "logps/chosen": -435.3922424316406, | |
| "logps/rejected": -426.5663146972656, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4437499940395355, | |
| "rewards/chosen": -0.03359255567193031, | |
| "rewards/margins": -0.0016514979070052505, | |
| "rewards/rejected": -0.03194105625152588, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.17777777777777778, | |
| "grad_norm": 0.23435867570612137, | |
| "learning_rate": 4.911085493475802e-06, | |
| "logits/chosen": 12.843754768371582, | |
| "logits/rejected": 12.69409465789795, | |
| "logps/chosen": -412.6481018066406, | |
| "logps/rejected": -412.96405029296875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -0.0513535737991333, | |
| "rewards/margins": 0.002654131967574358, | |
| "rewards/rejected": -0.05400770157575607, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.17777777777777778, | |
| "eval_logits/chosen": 14.180258750915527, | |
| "eval_logits/rejected": 12.376383781433105, | |
| "eval_logps/chosen": -413.67803955078125, | |
| "eval_logps/rejected": -391.0186462402344, | |
| "eval_loss": 0.6835460066795349, | |
| "eval_rewards/accuracies": 0.6904761791229248, | |
| "eval_rewards/chosen": -0.05108103156089783, | |
| "eval_rewards/margins": 0.021760080009698868, | |
| "eval_rewards/rejected": -0.072841115295887, | |
| "eval_runtime": 90.1605, | |
| "eval_samples_per_second": 11.091, | |
| "eval_steps_per_second": 0.699, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.19555555555555557, | |
| "grad_norm": 0.2803874830493494, | |
| "learning_rate": 4.86533931110987e-06, | |
| "logits/chosen": 13.455083847045898, | |
| "logits/rejected": 13.341041564941406, | |
| "logps/chosen": -415.30401611328125, | |
| "logps/rejected": -425.7816467285156, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": -0.06350487470626831, | |
| "rewards/margins": 0.00028126564575359225, | |
| "rewards/rejected": -0.06378613412380219, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.21333333333333335, | |
| "grad_norm": 0.35378680708306215, | |
| "learning_rate": 4.810442090457072e-06, | |
| "logits/chosen": 12.907583236694336, | |
| "logits/rejected": 13.067194938659668, | |
| "logps/chosen": -389.30303955078125, | |
| "logps/rejected": -398.48333740234375, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.06600853055715561, | |
| "rewards/margins": -0.003265662584453821, | |
| "rewards/rejected": -0.06274287402629852, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2311111111111111, | |
| "grad_norm": 0.2869743496766919, | |
| "learning_rate": 4.7466062181993855e-06, | |
| "logits/chosen": 13.347814559936523, | |
| "logits/rejected": 13.254618644714355, | |
| "logps/chosen": -396.9159851074219, | |
| "logps/rejected": -428.38861083984375, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.08628938347101212, | |
| "rewards/margins": 0.005706036929041147, | |
| "rewards/rejected": -0.09199541807174683, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.24888888888888888, | |
| "grad_norm": 0.3379629862483891, | |
| "learning_rate": 4.6740786629253595e-06, | |
| "logits/chosen": 13.342312812805176, | |
| "logits/rejected": 13.019983291625977, | |
| "logps/chosen": -384.1279602050781, | |
| "logps/rejected": -393.2745666503906, | |
| "loss": 0.6912, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.0947018712759018, | |
| "rewards/margins": 0.0077432007528841496, | |
| "rewards/rejected": -0.10244506597518921, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.26666666666666666, | |
| "grad_norm": 0.350219971213616, | |
| "learning_rate": 4.5931400196566256e-06, | |
| "logits/chosen": 13.074136734008789, | |
| "logits/rejected": 13.148015022277832, | |
| "logps/chosen": -425.1592712402344, | |
| "logps/rejected": -443.909423828125, | |
| "loss": 0.6903, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.11653944104909897, | |
| "rewards/margins": 0.003865143982693553, | |
| "rewards/rejected": -0.12040458619594574, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.28444444444444444, | |
| "grad_norm": 0.37627729798456006, | |
| "learning_rate": 4.504103424280267e-06, | |
| "logits/chosen": 13.34211540222168, | |
| "logits/rejected": 13.010579109191895, | |
| "logps/chosen": -439.13916015625, | |
| "logps/rejected": -433.8727111816406, | |
| "loss": 0.6908, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -0.16217225790023804, | |
| "rewards/margins": 0.007592835463583469, | |
| "rewards/rejected": -0.16976511478424072, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.3022222222222222, | |
| "grad_norm": 0.3180727458224856, | |
| "learning_rate": 4.407313342086906e-06, | |
| "logits/chosen": 13.196157455444336, | |
| "logits/rejected": 13.456674575805664, | |
| "logps/chosen": -426.7943420410156, | |
| "logps/rejected": -437.723388671875, | |
| "loss": 0.6937, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.16415511071681976, | |
| "rewards/margins": 0.009541703388094902, | |
| "rewards/rejected": -0.17369681596755981, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 0.31672781627835284, | |
| "learning_rate": 4.303144235101412e-06, | |
| "logits/chosen": 13.912513732910156, | |
| "logits/rejected": 13.92699146270752, | |
| "logps/chosen": -400.5416564941406, | |
| "logps/rejected": -418.99554443359375, | |
| "loss": 0.6907, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.18776951730251312, | |
| "rewards/margins": 0.00795517023652792, | |
| "rewards/rejected": -0.1957246959209442, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3377777777777778, | |
| "grad_norm": 0.3187670525496018, | |
| "learning_rate": 4.1919991133620705e-06, | |
| "logits/chosen": 12.957951545715332, | |
| "logits/rejected": 12.835040092468262, | |
| "logps/chosen": -428.75921630859375, | |
| "logps/rejected": -429.77801513671875, | |
| "loss": 0.6943, | |
| "rewards/accuracies": 0.4437499940395355, | |
| "rewards/chosen": -0.18785884976387024, | |
| "rewards/margins": -0.00687809195369482, | |
| "rewards/rejected": -0.18098074197769165, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.35555555555555557, | |
| "grad_norm": 0.31966290081913296, | |
| "learning_rate": 4.074307975753044e-06, | |
| "logits/chosen": 13.527238845825195, | |
| "logits/rejected": 13.133882522583008, | |
| "logps/chosen": -411.04052734375, | |
| "logps/rejected": -405.96820068359375, | |
| "loss": 0.689, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.1447705328464508, | |
| "rewards/margins": 0.013265645131468773, | |
| "rewards/rejected": -0.15803618729114532, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.35555555555555557, | |
| "eval_logits/chosen": 13.984086036682129, | |
| "eval_logits/rejected": 12.180322647094727, | |
| "eval_logps/chosen": -422.97613525390625, | |
| "eval_logps/rejected": -403.87432861328125, | |
| "eval_loss": 0.6681899428367615, | |
| "eval_rewards/accuracies": 0.7460317611694336, | |
| "eval_rewards/chosen": -0.1440620720386505, | |
| "eval_rewards/margins": 0.05733573064208031, | |
| "eval_rewards/rejected": -0.20139780640602112, | |
| "eval_runtime": 90.195, | |
| "eval_samples_per_second": 11.087, | |
| "eval_steps_per_second": 0.698, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.37333333333333335, | |
| "grad_norm": 0.2899334647869614, | |
| "learning_rate": 3.950526146422213e-06, | |
| "logits/chosen": 13.182365417480469, | |
| "logits/rejected": 13.274526596069336, | |
| "logps/chosen": -412.0303649902344, | |
| "logps/rejected": -435.66754150390625, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": -0.1916799396276474, | |
| "rewards/margins": -0.0004939109203405678, | |
| "rewards/rejected": -0.19118604063987732, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.39111111111111113, | |
| "grad_norm": 0.3113343433252108, | |
| "learning_rate": 3.821132513220511e-06, | |
| "logits/chosen": 13.472271919250488, | |
| "logits/rejected": 12.98419189453125, | |
| "logps/chosen": -407.59857177734375, | |
| "logps/rejected": -413.1893615722656, | |
| "loss": 0.6852, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -0.18176871538162231, | |
| "rewards/margins": 0.02464126981794834, | |
| "rewards/rejected": -0.2064099758863449, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.4088888888888889, | |
| "grad_norm": 0.37212373445120756, | |
| "learning_rate": 3.686627674977858e-06, | |
| "logits/chosen": 12.922555923461914, | |
| "logits/rejected": 13.052223205566406, | |
| "logps/chosen": -420.0254821777344, | |
| "logps/rejected": -438.8111267089844, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.2267444133758545, | |
| "rewards/margins": 0.00632152333855629, | |
| "rewards/rejected": -0.23306593298912048, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.4266666666666667, | |
| "grad_norm": 0.4194286554865179, | |
| "learning_rate": 3.547532004783539e-06, | |
| "logits/chosen": 13.442713737487793, | |
| "logits/rejected": 12.382542610168457, | |
| "logps/chosen": -448.96044921875, | |
| "logps/rejected": -424.40631103515625, | |
| "loss": 0.6883, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.2144048511981964, | |
| "rewards/margins": 0.04138387367129326, | |
| "rewards/rejected": -0.2557887136936188, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.4444444444444444, | |
| "grad_norm": 0.2831313864511856, | |
| "learning_rate": 3.404383636763809e-06, | |
| "logits/chosen": 13.222406387329102, | |
| "logits/rejected": 12.874628067016602, | |
| "logps/chosen": -460.45782470703125, | |
| "logps/rejected": -457.3854064941406, | |
| "loss": 0.6915, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": -0.22632627189159393, | |
| "rewards/margins": 0.018462661653757095, | |
| "rewards/rejected": -0.24478892982006073, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.4622222222222222, | |
| "grad_norm": 0.3260671824484102, | |
| "learning_rate": 3.2577363841455063e-06, | |
| "logits/chosen": 13.391107559204102, | |
| "logits/rejected": 13.3804931640625, | |
| "logps/chosen": -380.6063537597656, | |
| "logps/rejected": -394.7472839355469, | |
| "loss": 0.6954, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -0.20835450291633606, | |
| "rewards/margins": 0.013799709267914295, | |
| "rewards/rejected": -0.22215421497821808, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 0.3777555315785377, | |
| "learning_rate": 3.1081575966602627e-06, | |
| "logits/chosen": 12.585054397583008, | |
| "logits/rejected": 12.490537643432617, | |
| "logps/chosen": -439.46844482421875, | |
| "logps/rejected": -421.29962158203125, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": -0.19497910141944885, | |
| "rewards/margins": -0.00843017641454935, | |
| "rewards/rejected": -0.18654890358448029, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.49777777777777776, | |
| "grad_norm": 0.2554633592042652, | |
| "learning_rate": 2.9562259655786067e-06, | |
| "logits/chosen": 13.010714530944824, | |
| "logits/rejected": 12.371983528137207, | |
| "logps/chosen": -389.00665283203125, | |
| "logps/rejected": -389.6372985839844, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -0.14177796244621277, | |
| "rewards/margins": 0.008302886970341206, | |
| "rewards/rejected": -0.1500808447599411, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.5155555555555555, | |
| "grad_norm": 0.27965221608690755, | |
| "learning_rate": 2.802529284865863e-06, | |
| "logits/chosen": 12.02336597442627, | |
| "logits/rejected": 12.071900367736816, | |
| "logps/chosen": -402.3577575683594, | |
| "logps/rejected": -393.9205017089844, | |
| "loss": 0.6895, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.13531939685344696, | |
| "rewards/margins": 0.004990140907466412, | |
| "rewards/rejected": -0.14030954241752625, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.5333333333333333, | |
| "grad_norm": 0.2913462405762709, | |
| "learning_rate": 2.6476621771214865e-06, | |
| "logits/chosen": 13.3645601272583, | |
| "logits/rejected": 13.140283584594727, | |
| "logps/chosen": -407.8977966308594, | |
| "logps/rejected": -416.1576232910156, | |
| "loss": 0.6923, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -0.1288231909275055, | |
| "rewards/margins": 0.010681845247745514, | |
| "rewards/rejected": -0.1395050287246704, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5333333333333333, | |
| "eval_logits/chosen": 13.974791526794434, | |
| "eval_logits/rejected": 12.176884651184082, | |
| "eval_logps/chosen": -419.9747009277344, | |
| "eval_logps/rejected": -401.2295227050781, | |
| "eval_loss": 0.6673460006713867, | |
| "eval_rewards/accuracies": 0.7896825671195984, | |
| "eval_rewards/chosen": -0.11404754966497421, | |
| "eval_rewards/margins": 0.06090213730931282, | |
| "eval_rewards/rejected": -0.17494967579841614, | |
| "eval_runtime": 90.2773, | |
| "eval_samples_per_second": 11.077, | |
| "eval_steps_per_second": 0.698, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5511111111111111, | |
| "grad_norm": 0.340661457280574, | |
| "learning_rate": 2.4922237930997435e-06, | |
| "logits/chosen": 12.930778503417969, | |
| "logits/rejected": 12.75629711151123, | |
| "logps/chosen": -408.3033142089844, | |
| "logps/rejected": -429.09954833984375, | |
| "loss": 0.6891, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.1441982239484787, | |
| "rewards/margins": 0.008970534428954124, | |
| "rewards/rejected": -0.15316873788833618, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.5688888888888889, | |
| "grad_norm": 0.3045715209107194, | |
| "learning_rate": 2.3368154937118355e-06, | |
| "logits/chosen": 12.338762283325195, | |
| "logits/rejected": 12.408100128173828, | |
| "logps/chosen": -417.6708984375, | |
| "logps/rejected": -436.1878356933594, | |
| "loss": 0.6906, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -0.1610310822725296, | |
| "rewards/margins": 0.003636928740888834, | |
| "rewards/rejected": -0.164668008685112, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.5866666666666667, | |
| "grad_norm": 0.3261096004039151, | |
| "learning_rate": 2.1820385234773604e-06, | |
| "logits/chosen": 12.432132720947266, | |
| "logits/rejected": 12.333971977233887, | |
| "logps/chosen": -371.9541015625, | |
| "logps/rejected": -377.29327392578125, | |
| "loss": 0.6938, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.14137257635593414, | |
| "rewards/margins": 0.012616041116416454, | |
| "rewards/rejected": -0.15398862957954407, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.6044444444444445, | |
| "grad_norm": 0.3167841738260838, | |
| "learning_rate": 2.02849168442607e-06, | |
| "logits/chosen": 13.544441223144531, | |
| "logits/rejected": 13.638467788696289, | |
| "logps/chosen": -407.0322265625, | |
| "logps/rejected": -412.5428771972656, | |
| "loss": 0.6934, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -0.1435524821281433, | |
| "rewards/margins": 0.008516514673829079, | |
| "rewards/rejected": -0.15206900238990784, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.6222222222222222, | |
| "grad_norm": 0.30772925863397405, | |
| "learning_rate": 1.876769019449141e-06, | |
| "logits/chosen": 12.576923370361328, | |
| "logits/rejected": 12.56715202331543, | |
| "logps/chosen": -379.2666015625, | |
| "logps/rejected": -408.4591979980469, | |
| "loss": 0.6904, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.12999220192432404, | |
| "rewards/margins": 0.009667792357504368, | |
| "rewards/rejected": -0.13966000080108643, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 0.3510251282815541, | |
| "learning_rate": 1.7274575140626318e-06, | |
| "logits/chosen": 12.569269180297852, | |
| "logits/rejected": 12.474992752075195, | |
| "logps/chosen": -426.98651123046875, | |
| "logps/rejected": -420.2184143066406, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": -0.14840544760227203, | |
| "rewards/margins": -0.0038787845987826586, | |
| "rewards/rejected": -0.1445266604423523, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.6577777777777778, | |
| "grad_norm": 0.29359728465406504, | |
| "learning_rate": 1.5811348254745574e-06, | |
| "logits/chosen": 13.278701782226562, | |
| "logits/rejected": 13.138870239257812, | |
| "logps/chosen": -419.22552490234375, | |
| "logps/rejected": -429.7049865722656, | |
| "loss": 0.6907, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.1437017023563385, | |
| "rewards/margins": 0.006204875651746988, | |
| "rewards/rejected": -0.14990659058094025, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.6755555555555556, | |
| "grad_norm": 0.3041311221496272, | |
| "learning_rate": 1.4383670477413676e-06, | |
| "logits/chosen": 12.825994491577148, | |
| "logits/rejected": 12.180424690246582, | |
| "logps/chosen": -393.3042297363281, | |
| "logps/rejected": -379.05279541015625, | |
| "loss": 0.6881, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.1392853856086731, | |
| "rewards/margins": 0.007530958391726017, | |
| "rewards/rejected": -0.14681634306907654, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.6933333333333334, | |
| "grad_norm": 0.3072757460893545, | |
| "learning_rate": 1.2997065216600179e-06, | |
| "logits/chosen": 12.89039134979248, | |
| "logits/rejected": 13.162782669067383, | |
| "logps/chosen": -427.39910888671875, | |
| "logps/rejected": -431.7848205566406, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -0.14786246418952942, | |
| "rewards/margins": 0.0003430729848332703, | |
| "rewards/rejected": -0.14820551872253418, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.7111111111111111, | |
| "grad_norm": 0.27986512697276633, | |
| "learning_rate": 1.165689697868726e-06, | |
| "logits/chosen": 12.779977798461914, | |
| "logits/rejected": 12.16929817199707, | |
| "logps/chosen": -420.75653076171875, | |
| "logps/rejected": -415.25244140625, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -0.14449290931224823, | |
| "rewards/margins": 0.02024017833173275, | |
| "rewards/rejected": -0.16473311185836792, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.7111111111111111, | |
| "eval_logits/chosen": 13.931650161743164, | |
| "eval_logits/rejected": 12.126729011535645, | |
| "eval_logps/chosen": -420.5239562988281, | |
| "eval_logps/rejected": -402.1236267089844, | |
| "eval_loss": 0.665495753288269, | |
| "eval_rewards/accuracies": 0.7698412537574768, | |
| "eval_rewards/chosen": -0.11954014003276825, | |
| "eval_rewards/margins": 0.06435071676969528, | |
| "eval_rewards/rejected": -0.18389087915420532, | |
| "eval_runtime": 90.069, | |
| "eval_samples_per_second": 11.103, | |
| "eval_steps_per_second": 0.699, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.7288888888888889, | |
| "grad_norm": 0.3450004634992278, | |
| "learning_rate": 1.0368350614236685e-06, | |
| "logits/chosen": 12.588386535644531, | |
| "logits/rejected": 12.996353149414062, | |
| "logps/chosen": -403.56207275390625, | |
| "logps/rejected": -424.3435974121094, | |
| "loss": 0.6911, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.15337641537189484, | |
| "rewards/margins": 0.0022048167884349823, | |
| "rewards/rejected": -0.15558123588562012, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.7466666666666667, | |
| "grad_norm": 0.2947231406284332, | |
| "learning_rate": 9.136411258810229e-07, | |
| "logits/chosen": 13.508552551269531, | |
| "logits/rejected": 13.856636047363281, | |
| "logps/chosen": -397.31103515625, | |
| "logps/rejected": -415.14794921875, | |
| "loss": 0.6916, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -0.14091812074184418, | |
| "rewards/margins": 0.009603964164853096, | |
| "rewards/rejected": -0.15052208304405212, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.7644444444444445, | |
| "grad_norm": 0.2949706409885573, | |
| "learning_rate": 7.965845046448659e-07, | |
| "logits/chosen": 12.716641426086426, | |
| "logits/rejected": 12.739425659179688, | |
| "logps/chosen": -413.5958557128906, | |
| "logps/rejected": -418.07379150390625, | |
| "loss": 0.6924, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -0.13801425695419312, | |
| "rewards/margins": 0.0007665277225896716, | |
| "rewards/rejected": -0.13878078758716583, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.7822222222222223, | |
| "grad_norm": 0.29714148256405104, | |
| "learning_rate": 6.861180670424983e-07, | |
| "logits/chosen": 13.228073120117188, | |
| "logits/rejected": 12.513433456420898, | |
| "logps/chosen": -436.03326416015625, | |
| "logps/rejected": -432.7162170410156, | |
| "loss": 0.6882, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.14261861145496368, | |
| "rewards/margins": 0.018722299486398697, | |
| "rewards/rejected": -0.16134092211723328, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 0.35583565190086214, | |
| "learning_rate": 5.826691862609987e-07, | |
| "logits/chosen": 13.008562088012695, | |
| "logits/rejected": 12.5421724319458, | |
| "logps/chosen": -393.10614013671875, | |
| "logps/rejected": -396.7547607421875, | |
| "loss": 0.6858, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -0.14102980494499207, | |
| "rewards/margins": 0.006347469985485077, | |
| "rewards/rejected": -0.14737728238105774, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.8177777777777778, | |
| "grad_norm": 1.2849378777119373, | |
| "learning_rate": 4.866380859233891e-07, | |
| "logits/chosen": 12.925387382507324, | |
| "logits/rejected": 13.122329711914062, | |
| "logps/chosen": -406.62255859375, | |
| "logps/rejected": -432.58575439453125, | |
| "loss": 0.6911, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -0.13780517876148224, | |
| "rewards/margins": 0.009602868929505348, | |
| "rewards/rejected": -0.14740803837776184, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.8355555555555556, | |
| "grad_norm": 0.3135496868277975, | |
| "learning_rate": 3.98396291701183e-07, | |
| "logits/chosen": 13.169825553894043, | |
| "logits/rejected": 12.919093132019043, | |
| "logps/chosen": -420.40594482421875, | |
| "logps/rejected": -424.626220703125, | |
| "loss": 0.6887, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -0.14023754000663757, | |
| "rewards/margins": 0.012833138927817345, | |
| "rewards/rejected": -0.15307065844535828, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.8533333333333334, | |
| "grad_norm": 0.3340559430991029, | |
| "learning_rate": 3.1828519395374095e-07, | |
| "logits/chosen": 13.121709823608398, | |
| "logits/rejected": 13.285209655761719, | |
| "logps/chosen": -428.4640197753906, | |
| "logps/rejected": -445.868408203125, | |
| "loss": 0.6934, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -0.1603536158800125, | |
| "rewards/margins": 0.00828765518963337, | |
| "rewards/rejected": -0.16864125430583954, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.8711111111111111, | |
| "grad_norm": 0.32372602686992025, | |
| "learning_rate": 2.466147269552893e-07, | |
| "logits/chosen": 13.549275398254395, | |
| "logits/rejected": 13.303945541381836, | |
| "logps/chosen": -398.70245361328125, | |
| "logps/rejected": -403.9294128417969, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.1653539538383484, | |
| "rewards/margins": 0.015345364809036255, | |
| "rewards/rejected": -0.18069931864738464, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.8888888888888888, | |
| "grad_norm": 0.28909373758975987, | |
| "learning_rate": 1.8366216981942632e-07, | |
| "logits/chosen": 12.993377685546875, | |
| "logits/rejected": 13.172945976257324, | |
| "logps/chosen": -441.27288818359375, | |
| "logps/rejected": -444.38983154296875, | |
| "loss": 0.696, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -0.17415288090705872, | |
| "rewards/margins": -0.0011280607432126999, | |
| "rewards/rejected": -0.17302480340003967, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.8888888888888888, | |
| "eval_logits/chosen": 13.899726867675781, | |
| "eval_logits/rejected": 12.095231056213379, | |
| "eval_logps/chosen": -421.1900939941406, | |
| "eval_logps/rejected": -403.3279724121094, | |
| "eval_loss": 0.6633419990539551, | |
| "eval_rewards/accuracies": 0.7539682388305664, | |
| "eval_rewards/chosen": -0.1262015700340271, | |
| "eval_rewards/margins": 0.06973244994878769, | |
| "eval_rewards/rejected": -0.19593402743339539, | |
| "eval_runtime": 90.1435, | |
| "eval_samples_per_second": 11.093, | |
| "eval_steps_per_second": 0.699, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.9066666666666666, | |
| "grad_norm": 0.44905140288995754, | |
| "learning_rate": 1.296710737600934e-07, | |
| "logits/chosen": 12.890707015991211, | |
| "logits/rejected": 12.480443000793457, | |
| "logps/chosen": -397.1539611816406, | |
| "logps/rejected": -404.51470947265625, | |
| "loss": 0.6926, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -0.14948078989982605, | |
| "rewards/margins": 0.0065328641794621944, | |
| "rewards/rejected": -0.1560136377811432, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.9244444444444444, | |
| "grad_norm": 0.3012451953425647, | |
| "learning_rate": 8.485031983924558e-08, | |
| "logits/chosen": 13.72374153137207, | |
| "logits/rejected": 14.290632247924805, | |
| "logps/chosen": -409.4125061035156, | |
| "logps/rejected": -427.1560974121094, | |
| "loss": 0.6908, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": -0.15699708461761475, | |
| "rewards/margins": -0.00222613662481308, | |
| "rewards/rejected": -0.15477094054222107, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.9422222222222222, | |
| "grad_norm": 0.3269177206871498, | |
| "learning_rate": 4.93733108466013e-08, | |
| "logits/chosen": 12.78927993774414, | |
| "logits/rejected": 13.388028144836426, | |
| "logps/chosen": -428.2759704589844, | |
| "logps/rejected": -456.5814514160156, | |
| "loss": 0.6895, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": -0.15250369906425476, | |
| "rewards/margins": 0.005445868708193302, | |
| "rewards/rejected": -0.15794958174228668, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 0.33934624748600184, | |
| "learning_rate": 2.3377300437934236e-08, | |
| "logits/chosen": 13.640138626098633, | |
| "logits/rejected": 13.247881889343262, | |
| "logps/chosen": -384.22723388671875, | |
| "logps/rejected": -377.16546630859375, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.14129753410816193, | |
| "rewards/margins": 0.019017567858099937, | |
| "rewards/rejected": -0.16031508147716522, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.9777777777777777, | |
| "grad_norm": 0.31794002037631625, | |
| "learning_rate": 6.962862127343206e-09, | |
| "logits/chosen": 13.490945816040039, | |
| "logits/rejected": 13.803362846374512, | |
| "logps/chosen": -422.62939453125, | |
| "logps/rejected": -427.26959228515625, | |
| "loss": 0.6894, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.167506605386734, | |
| "rewards/margins": 0.010436911135911942, | |
| "rewards/rejected": -0.17794351279735565, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.9955555555555555, | |
| "grad_norm": 0.29176173477304573, | |
| "learning_rate": 1.9350018786556956e-10, | |
| "logits/chosen": 13.073209762573242, | |
| "logits/rejected": 13.310659408569336, | |
| "logps/chosen": -431.8447265625, | |
| "logps/rejected": -418.0799865722656, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.1620911806821823, | |
| "rewards/margins": -0.006803811527788639, | |
| "rewards/rejected": -0.15528738498687744, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.9991111111111111, | |
| "step": 562, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6916975294143703, | |
| "train_runtime": 7520.1237, | |
| "train_samples_per_second": 4.787, | |
| "train_steps_per_second": 0.075 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 562, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "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": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |