| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9997382884061764, | |
| "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": 0.59375, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -2.1666858196258545, | |
| "logits/rejected": -2.182244300842285, | |
| "logps/chosen": -12.368609428405762, | |
| "logps/rejected": -24.687644958496094, | |
| "loss": 0.6931, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1, | |
| "use_label": 10.0 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 0.6015625, | |
| "learning_rate": 5.208333333333334e-07, | |
| "logits/chosen": -2.2111542224884033, | |
| "logits/rejected": -2.2718067169189453, | |
| "logps/chosen": -57.56840133666992, | |
| "logps/rejected": -65.20916748046875, | |
| "loss": 0.693, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.2569444477558136, | |
| "rewards/chosen": 0.0011389791034162045, | |
| "rewards/margins": 0.0002508986508473754, | |
| "rewards/rejected": 0.0008880805689841509, | |
| "step": 10, | |
| "use_label": 90.0 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 0.6796875, | |
| "learning_rate": 1.0416666666666667e-06, | |
| "logits/chosen": -2.242893695831299, | |
| "logits/rejected": -2.279961109161377, | |
| "logps/chosen": -56.537681579589844, | |
| "logps/rejected": -68.3794174194336, | |
| "loss": 0.6924, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": 0.006626849062740803, | |
| "rewards/margins": 0.001654049614444375, | |
| "rewards/rejected": 0.004972799215465784, | |
| "step": 20, | |
| "use_label": 242.0 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 0.55078125, | |
| "learning_rate": 1.5625e-06, | |
| "logits/chosen": -2.2637219429016113, | |
| "logits/rejected": -2.2480521202087402, | |
| "logps/chosen": -53.993507385253906, | |
| "logps/rejected": -67.89700317382812, | |
| "loss": 0.6919, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.26875001192092896, | |
| "rewards/chosen": 0.016421381384134293, | |
| "rewards/margins": 0.002580237342044711, | |
| "rewards/rejected": 0.013841142877936363, | |
| "step": 30, | |
| "use_label": 402.0 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 0.6328125, | |
| "learning_rate": 2.0833333333333334e-06, | |
| "logits/chosen": -2.2831993103027344, | |
| "logits/rejected": -2.2760486602783203, | |
| "logps/chosen": -55.59602737426758, | |
| "logps/rejected": -66.58573913574219, | |
| "loss": 0.6909, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.20624999701976776, | |
| "rewards/chosen": 0.018266689032316208, | |
| "rewards/margins": 0.0004533957107923925, | |
| "rewards/rejected": 0.017813291400671005, | |
| "step": 40, | |
| "use_label": 562.0 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 0.6015625, | |
| "learning_rate": 2.604166666666667e-06, | |
| "logits/chosen": -2.344376564025879, | |
| "logits/rejected": -2.3342297077178955, | |
| "logps/chosen": -69.12073516845703, | |
| "logps/rejected": -84.67558288574219, | |
| "loss": 0.6889, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": 0.02673395536839962, | |
| "rewards/margins": 0.00583356199786067, | |
| "rewards/rejected": 0.020900394767522812, | |
| "step": 50, | |
| "use_label": 722.0 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 0.72265625, | |
| "learning_rate": 3.125e-06, | |
| "logits/chosen": -2.3030121326446533, | |
| "logits/rejected": -2.3094825744628906, | |
| "logps/chosen": -82.04167175292969, | |
| "logps/rejected": -90.7291488647461, | |
| "loss": 0.6876, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": 0.036534082144498825, | |
| "rewards/margins": 0.013860121369361877, | |
| "rewards/rejected": 0.022673960775136948, | |
| "step": 60, | |
| "use_label": 882.0 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 0.7890625, | |
| "learning_rate": 3.6458333333333333e-06, | |
| "logits/chosen": -2.345569610595703, | |
| "logits/rejected": -2.3263676166534424, | |
| "logps/chosen": -77.1853256225586, | |
| "logps/rejected": -77.63880920410156, | |
| "loss": 0.685, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.3062500059604645, | |
| "rewards/chosen": 0.025494003668427467, | |
| "rewards/margins": 0.016305232420563698, | |
| "rewards/rejected": 0.009188770316541195, | |
| "step": 70, | |
| "use_label": 1042.0 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 0.81640625, | |
| "learning_rate": 4.166666666666667e-06, | |
| "logits/chosen": -2.241882801055908, | |
| "logits/rejected": -2.195146322250366, | |
| "logps/chosen": -81.66094207763672, | |
| "logps/rejected": -89.08940124511719, | |
| "loss": 0.6805, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": 0.003909807652235031, | |
| "rewards/margins": 0.025169039145112038, | |
| "rewards/rejected": -0.021259231492877007, | |
| "step": 80, | |
| "use_label": 1202.0 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 1.7734375, | |
| "learning_rate": 4.6875000000000004e-06, | |
| "logits/chosen": -2.1871695518493652, | |
| "logits/rejected": -2.2313501834869385, | |
| "logps/chosen": -62.76741409301758, | |
| "logps/rejected": -81.16191101074219, | |
| "loss": 0.6747, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.33125001192092896, | |
| "rewards/chosen": -0.016776535660028458, | |
| "rewards/margins": 0.048332639038562775, | |
| "rewards/rejected": -0.06510917842388153, | |
| "step": 90, | |
| "use_label": 1362.0 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 1.4375, | |
| "learning_rate": 4.9997324926814375e-06, | |
| "logits/chosen": -2.1414177417755127, | |
| "logits/rejected": -2.107236623764038, | |
| "logps/chosen": -78.60578155517578, | |
| "logps/rejected": -81.1384506225586, | |
| "loss": 0.6685, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -0.04031088575720787, | |
| "rewards/margins": 0.052690792828798294, | |
| "rewards/rejected": -0.09300167858600616, | |
| "step": 100, | |
| "use_label": 1522.0 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "eval_logits/chosen": -2.109715223312378, | |
| "eval_logits/rejected": -2.0796475410461426, | |
| "eval_logps/chosen": -71.95718383789062, | |
| "eval_logps/rejected": -84.7625961303711, | |
| "eval_loss": 0.6684110760688782, | |
| "eval_pred_label": 0.0, | |
| "eval_rewards/accuracies": 0.335317462682724, | |
| "eval_rewards/chosen": -0.030566338449716568, | |
| "eval_rewards/margins": 0.06307896971702576, | |
| "eval_rewards/rejected": -0.09364530444145203, | |
| "eval_runtime": 247.4954, | |
| "eval_samples_per_second": 8.081, | |
| "eval_steps_per_second": 0.255, | |
| "eval_use_label": 1856.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 1.71875, | |
| "learning_rate": 4.996723692767927e-06, | |
| "logits/chosen": -2.12998104095459, | |
| "logits/rejected": -2.1109042167663574, | |
| "logps/chosen": -68.2921142578125, | |
| "logps/rejected": -84.99057006835938, | |
| "loss": 0.6713, | |
| "pred_label": 0.07500000298023224, | |
| "rewards/accuracies": 0.3062500059604645, | |
| "rewards/chosen": -0.06523006409406662, | |
| "rewards/margins": 0.0570509135723114, | |
| "rewards/rejected": -0.12228099256753922, | |
| "step": 110, | |
| "use_label": 2185.925048828125 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 1.046875, | |
| "learning_rate": 4.9903757462135984e-06, | |
| "logits/chosen": -2.3605504035949707, | |
| "logits/rejected": -2.243201971054077, | |
| "logps/chosen": -80.2857666015625, | |
| "logps/rejected": -96.13867950439453, | |
| "loss": 0.6667, | |
| "pred_label": 1.875, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -0.062185365706682205, | |
| "rewards/margins": 0.0815814733505249, | |
| "rewards/rejected": -0.143766850233078, | |
| "step": 120, | |
| "use_label": 2344.125 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 1.0078125, | |
| "learning_rate": 4.980697142834315e-06, | |
| "logits/chosen": -2.146286725997925, | |
| "logits/rejected": -2.1618175506591797, | |
| "logps/chosen": -67.6681137084961, | |
| "logps/rejected": -78.9002456665039, | |
| "loss": 0.6675, | |
| "pred_label": 2.075000047683716, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -0.09194014966487885, | |
| "rewards/margins": 0.04880703240633011, | |
| "rewards/rejected": -0.14074717462062836, | |
| "step": 130, | |
| "use_label": 2503.925048828125 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 2.28125, | |
| "learning_rate": 4.967700826904229e-06, | |
| "logits/chosen": -2.1254963874816895, | |
| "logits/rejected": -2.160235643386841, | |
| "logps/chosen": -74.917724609375, | |
| "logps/rejected": -99.1263427734375, | |
| "loss": 0.6684, | |
| "pred_label": 5.050000190734863, | |
| "rewards/accuracies": 0.28125, | |
| "rewards/chosen": -0.15644724667072296, | |
| "rewards/margins": 0.08642110973596573, | |
| "rewards/rejected": -0.2428683489561081, | |
| "step": 140, | |
| "use_label": 2660.949951171875 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 1.359375, | |
| "learning_rate": 4.951404179843963e-06, | |
| "logits/chosen": -2.1338083744049072, | |
| "logits/rejected": -2.0785932540893555, | |
| "logps/chosen": -57.580589294433594, | |
| "logps/rejected": -64.5077133178711, | |
| "loss": 0.6658, | |
| "pred_label": 9.100000381469727, | |
| "rewards/accuracies": 0.26875001192092896, | |
| "rewards/chosen": -0.10085760056972504, | |
| "rewards/margins": 0.09126537293195724, | |
| "rewards/rejected": -0.19212298095226288, | |
| "step": 150, | |
| "use_label": 2816.89990234375 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 2.15625, | |
| "learning_rate": 4.931828996974498e-06, | |
| "logits/chosen": -2.1478817462921143, | |
| "logits/rejected": -2.1238207817077637, | |
| "logps/chosen": -102.6265640258789, | |
| "logps/rejected": -118.9216537475586, | |
| "loss": 0.667, | |
| "pred_label": 15.399999618530273, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -0.21023361384868622, | |
| "rewards/margins": 0.16089467704296112, | |
| "rewards/rejected": -0.37112829089164734, | |
| "step": 160, | |
| "use_label": 2970.60009765625 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 2.140625, | |
| "learning_rate": 4.909001458367867e-06, | |
| "logits/chosen": -1.9664795398712158, | |
| "logits/rejected": -1.9388923645019531, | |
| "logps/chosen": -81.67234802246094, | |
| "logps/rejected": -97.5047836303711, | |
| "loss": 0.6635, | |
| "pred_label": 23.475000381469727, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -0.17715924978256226, | |
| "rewards/margins": 0.13409331440925598, | |
| "rewards/rejected": -0.31125253438949585, | |
| "step": 170, | |
| "use_label": 3122.52490234375 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 2.96875, | |
| "learning_rate": 4.882952093833628e-06, | |
| "logits/chosen": -1.896836519241333, | |
| "logits/rejected": -1.9397751092910767, | |
| "logps/chosen": -82.50892639160156, | |
| "logps/rejected": -105.1452407836914, | |
| "loss": 0.6701, | |
| "pred_label": 29.649999618530273, | |
| "rewards/accuracies": 0.33125001192092896, | |
| "rewards/chosen": -0.20006974041461945, | |
| "rewards/margins": 0.11658792197704315, | |
| "rewards/rejected": -0.3166576623916626, | |
| "step": 180, | |
| "use_label": 3276.35009765625 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 1.5859375, | |
| "learning_rate": 4.853715742087947e-06, | |
| "logits/chosen": -1.8957335948944092, | |
| "logits/rejected": -1.8187646865844727, | |
| "logps/chosen": -101.19456481933594, | |
| "logps/rejected": -109.06144714355469, | |
| "loss": 0.6648, | |
| "pred_label": 35.17499923706055, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -0.27311572432518005, | |
| "rewards/margins": 0.1226036325097084, | |
| "rewards/rejected": -0.39571934938430786, | |
| "step": 190, | |
| "use_label": 3430.824951171875 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 2.125, | |
| "learning_rate": 4.821331504159906e-06, | |
| "logits/chosen": -1.8366466760635376, | |
| "logits/rejected": -1.8133814334869385, | |
| "logps/chosen": -118.866943359375, | |
| "logps/rejected": -123.262451171875, | |
| "loss": 0.676, | |
| "pred_label": 43.42499923706055, | |
| "rewards/accuracies": 0.35624998807907104, | |
| "rewards/chosen": -0.38812780380249023, | |
| "rewards/margins": 0.1128091812133789, | |
| "rewards/rejected": -0.5009369850158691, | |
| "step": 200, | |
| "use_label": 3582.574951171875 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": -1.7319464683532715, | |
| "eval_logits/rejected": -1.6888620853424072, | |
| "eval_logps/chosen": -106.19064331054688, | |
| "eval_logps/rejected": -124.95625305175781, | |
| "eval_loss": 0.6716896295547485, | |
| "eval_pred_label": 61.55555725097656, | |
| "eval_rewards/accuracies": 0.3214285671710968, | |
| "eval_rewards/chosen": -0.372901052236557, | |
| "eval_rewards/margins": 0.12268086522817612, | |
| "eval_rewards/rejected": -0.49558189511299133, | |
| "eval_runtime": 248.0123, | |
| "eval_samples_per_second": 8.064, | |
| "eval_steps_per_second": 0.254, | |
| "eval_use_label": 3898.4443359375, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 2.515625, | |
| "learning_rate": 4.7858426910973435e-06, | |
| "logits/chosen": -1.9356311559677124, | |
| "logits/rejected": -1.9080215692520142, | |
| "logps/chosen": -93.94760131835938, | |
| "logps/rejected": -106.8377456665039, | |
| "loss": 0.6743, | |
| "pred_label": 81.25, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -0.259369432926178, | |
| "rewards/margins": 0.13343419134616852, | |
| "rewards/rejected": -0.3928036093711853, | |
| "step": 210, | |
| "use_label": 4208.75 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 2.34375, | |
| "learning_rate": 4.747296766042161e-06, | |
| "logits/chosen": -1.8437402248382568, | |
| "logits/rejected": -1.810903549194336, | |
| "logps/chosen": -100.77757263183594, | |
| "logps/rejected": -112.38002014160156, | |
| "loss": 0.6651, | |
| "pred_label": 92.5999984741211, | |
| "rewards/accuracies": 0.35624998807907104, | |
| "rewards/chosen": -0.24724093079566956, | |
| "rewards/margins": 0.1804189234972, | |
| "rewards/rejected": -0.427659809589386, | |
| "step": 220, | |
| "use_label": 4357.39990234375 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 1.828125, | |
| "learning_rate": 4.705745280752586e-06, | |
| "logits/chosen": -1.5612363815307617, | |
| "logits/rejected": -1.494425654411316, | |
| "logps/chosen": -108.3369369506836, | |
| "logps/rejected": -121.25785064697266, | |
| "loss": 0.6784, | |
| "pred_label": 107.0, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": -0.29597795009613037, | |
| "rewards/margins": 0.18698883056640625, | |
| "rewards/rejected": -0.482966810464859, | |
| "step": 230, | |
| "use_label": 4503.0 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 3.109375, | |
| "learning_rate": 4.661243806657256e-06, | |
| "logits/chosen": -1.279926061630249, | |
| "logits/rejected": -1.207486629486084, | |
| "logps/chosen": -94.6622314453125, | |
| "logps/rejected": -121.74755859375, | |
| "loss": 0.68, | |
| "pred_label": 124.05000305175781, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -0.30948689579963684, | |
| "rewards/margins": 0.16862434148788452, | |
| "rewards/rejected": -0.4781111776828766, | |
| "step": 240, | |
| "use_label": 4645.9501953125 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 1.8203125, | |
| "learning_rate": 4.613851860533367e-06, | |
| "logits/chosen": -1.4483808279037476, | |
| "logits/rejected": -1.535796880722046, | |
| "logps/chosen": -88.96175384521484, | |
| "logps/rejected": -106.26942443847656, | |
| "loss": 0.678, | |
| "pred_label": 140.64999389648438, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -0.2383408546447754, | |
| "rewards/margins": 0.1766553670167923, | |
| "rewards/rejected": -0.4149962067604065, | |
| "step": 250, | |
| "use_label": 4789.35009765625 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 2.359375, | |
| "learning_rate": 4.563632824908252e-06, | |
| "logits/chosen": -1.5566436052322388, | |
| "logits/rejected": -1.472214937210083, | |
| "logps/chosen": -101.36164855957031, | |
| "logps/rejected": -132.7355194091797, | |
| "loss": 0.6768, | |
| "pred_label": 156.85000610351562, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -0.404205858707428, | |
| "rewards/margins": 0.1837155818939209, | |
| "rewards/rejected": -0.5879215002059937, | |
| "step": 260, | |
| "use_label": 4933.14990234375 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 2.265625, | |
| "learning_rate": 4.510653863290871e-06, | |
| "logits/chosen": -1.5190045833587646, | |
| "logits/rejected": -1.5413776636123657, | |
| "logps/chosen": -123.2553482055664, | |
| "logps/rejected": -132.5965118408203, | |
| "loss": 0.6699, | |
| "pred_label": 174.9499969482422, | |
| "rewards/accuracies": 0.3187499940395355, | |
| "rewards/chosen": -0.478261798620224, | |
| "rewards/margins": 0.1528010368347168, | |
| "rewards/rejected": -0.6310628056526184, | |
| "step": 270, | |
| "use_label": 5075.0498046875 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 1.390625, | |
| "learning_rate": 4.454985830346574e-06, | |
| "logits/chosen": -1.4161837100982666, | |
| "logits/rejected": -1.461897850036621, | |
| "logps/chosen": -97.59378051757812, | |
| "logps/rejected": -113.92098236083984, | |
| "loss": 0.6682, | |
| "pred_label": 182.6750030517578, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -0.36361420154571533, | |
| "rewards/margins": 0.11149580776691437, | |
| "rewards/rejected": -0.4751099944114685, | |
| "step": 280, | |
| "use_label": 5227.3251953125 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 2.0, | |
| "learning_rate": 4.396703177135262e-06, | |
| "logits/chosen": -1.1572140455245972, | |
| "logits/rejected": -1.1582170724868774, | |
| "logps/chosen": -113.84346008300781, | |
| "logps/rejected": -136.17958068847656, | |
| "loss": 0.6783, | |
| "pred_label": 200.5749969482422, | |
| "rewards/accuracies": 0.3687500059604645, | |
| "rewards/chosen": -0.3756815791130066, | |
| "rewards/margins": 0.2912302017211914, | |
| "rewards/rejected": -0.666911780834198, | |
| "step": 290, | |
| "use_label": 5369.4248046875 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 2.59375, | |
| "learning_rate": 4.335883851539693e-06, | |
| "logits/chosen": -0.7739458084106445, | |
| "logits/rejected": -0.8222519159317017, | |
| "logps/chosen": -96.963623046875, | |
| "logps/rejected": -129.59027099609375, | |
| "loss": 0.6728, | |
| "pred_label": 223.0749969482422, | |
| "rewards/accuracies": 0.3187499940395355, | |
| "rewards/chosen": -0.43035492300987244, | |
| "rewards/margins": 0.2433358132839203, | |
| "rewards/rejected": -0.6736907958984375, | |
| "step": 300, | |
| "use_label": 5506.9248046875 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "eval_logits/chosen": -0.7414401173591614, | |
| "eval_logits/rejected": -0.6762140393257141, | |
| "eval_logps/chosen": -116.0198745727539, | |
| "eval_logps/rejected": -145.98529052734375, | |
| "eval_loss": 0.6783695220947266, | |
| "eval_pred_label": 270.96826171875, | |
| "eval_rewards/accuracies": 0.3373015820980072, | |
| "eval_rewards/chosen": -0.4711931049823761, | |
| "eval_rewards/margins": 0.23467905819416046, | |
| "eval_rewards/rejected": -0.7058721780776978, | |
| "eval_runtime": 248.0617, | |
| "eval_samples_per_second": 8.063, | |
| "eval_steps_per_second": 0.254, | |
| "eval_use_label": 5793.03173828125, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 2.21875, | |
| "learning_rate": 4.2726091940171055e-06, | |
| "logits/chosen": -1.1870858669281006, | |
| "logits/rejected": -1.1604619026184082, | |
| "logps/chosen": -93.13883972167969, | |
| "logps/rejected": -115.36534118652344, | |
| "loss": 0.6683, | |
| "pred_label": 314.92498779296875, | |
| "rewards/accuracies": 0.3062500059604645, | |
| "rewards/chosen": -0.4414878487586975, | |
| "rewards/margins": 0.11716248840093613, | |
| "rewards/rejected": -0.5586503148078918, | |
| "step": 310, | |
| "use_label": 6079.0751953125 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 2.34375, | |
| "learning_rate": 4.206963828813555e-06, | |
| "logits/chosen": -0.7778801918029785, | |
| "logits/rejected": -0.8481136560440063, | |
| "logps/chosen": -115.88163757324219, | |
| "logps/rejected": -152.60556030273438, | |
| "loss": 0.6715, | |
| "pred_label": 340.4750061035156, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": -0.41940560936927795, | |
| "rewards/margins": 0.2641361653804779, | |
| "rewards/rejected": -0.6835418343544006, | |
| "step": 320, | |
| "use_label": 6213.52490234375 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 2.8125, | |
| "learning_rate": 4.139035550786495e-06, | |
| "logits/chosen": -0.7590861320495605, | |
| "logits/rejected": -0.6889998316764832, | |
| "logps/chosen": -90.45745086669922, | |
| "logps/rejected": -116.98609924316406, | |
| "loss": 0.6718, | |
| "pred_label": 364.04998779296875, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -0.269859254360199, | |
| "rewards/margins": 0.2574610114097595, | |
| "rewards/rejected": -0.5273202657699585, | |
| "step": 330, | |
| "use_label": 6349.9501953125 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 2.703125, | |
| "learning_rate": 4.068915207986931e-06, | |
| "logits/chosen": -0.6209542155265808, | |
| "logits/rejected": -0.47464966773986816, | |
| "logps/chosen": -102.54121398925781, | |
| "logps/rejected": -130.24276733398438, | |
| "loss": 0.678, | |
| "pred_label": 391.07501220703125, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -0.395341694355011, | |
| "rewards/margins": 0.19944116473197937, | |
| "rewards/rejected": -0.594782829284668, | |
| "step": 340, | |
| "use_label": 6482.9248046875 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 1.8203125, | |
| "learning_rate": 3.996696580158211e-06, | |
| "logits/chosen": -0.7593547701835632, | |
| "logits/rejected": -0.6881019473075867, | |
| "logps/chosen": -92.08587646484375, | |
| "logps/rejected": -110.56968688964844, | |
| "loss": 0.6727, | |
| "pred_label": 411.9750061035156, | |
| "rewards/accuracies": 0.3187499940395355, | |
| "rewards/chosen": -0.31440719962120056, | |
| "rewards/margins": 0.14519965648651123, | |
| "rewards/rejected": -0.4596068263053894, | |
| "step": 350, | |
| "use_label": 6622.02490234375 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 1.9609375, | |
| "learning_rate": 3.922476253313921e-06, | |
| "logits/chosen": -0.996785044670105, | |
| "logits/rejected": -0.9698454737663269, | |
| "logps/chosen": -101.95857238769531, | |
| "logps/rejected": -114.76066589355469, | |
| "loss": 0.6783, | |
| "pred_label": 427.95001220703125, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": -0.32893818616867065, | |
| "rewards/margins": 0.16835859417915344, | |
| "rewards/rejected": -0.4972967505455017, | |
| "step": 360, | |
| "use_label": 6766.0498046875 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 1.484375, | |
| "learning_rate": 3.846353490562664e-06, | |
| "logits/chosen": -1.0720884799957275, | |
| "logits/rejected": -0.859793484210968, | |
| "logps/chosen": -103.12544250488281, | |
| "logps/rejected": -120.26700592041016, | |
| "loss": 0.6626, | |
| "pred_label": 443.875, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -0.29031243920326233, | |
| "rewards/margins": 0.2187139242887497, | |
| "rewards/rejected": -0.509026288986206, | |
| "step": 370, | |
| "use_label": 6910.125 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 2.671875, | |
| "learning_rate": 3.768430099352445e-06, | |
| "logits/chosen": -0.4451161324977875, | |
| "logits/rejected": -0.32113510370254517, | |
| "logps/chosen": -106.78487396240234, | |
| "logps/rejected": -132.46365356445312, | |
| "loss": 0.6759, | |
| "pred_label": 463.2749938964844, | |
| "rewards/accuracies": 0.34375, | |
| "rewards/chosen": -0.4723123610019684, | |
| "rewards/margins": 0.1635245531797409, | |
| "rewards/rejected": -0.6358368992805481, | |
| "step": 380, | |
| "use_label": 7050.72509765625 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 2.875, | |
| "learning_rate": 3.6888102953122307e-06, | |
| "logits/chosen": -0.02978489175438881, | |
| "logits/rejected": -0.03225391358137131, | |
| "logps/chosen": -136.5101318359375, | |
| "logps/rejected": -144.56173706054688, | |
| "loss": 0.6827, | |
| "pred_label": 495.82501220703125, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -0.516473650932312, | |
| "rewards/margins": 0.25756725668907166, | |
| "rewards/rejected": -0.774040937423706, | |
| "step": 390, | |
| "use_label": 7178.1748046875 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 2.515625, | |
| "learning_rate": 3.607600562872785e-06, | |
| "logits/chosen": 0.09610392153263092, | |
| "logits/rejected": 0.09092014282941818, | |
| "logps/chosen": -117.39754486083984, | |
| "logps/rejected": -152.1193084716797, | |
| "loss": 0.6715, | |
| "pred_label": 523.0250244140625, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -0.4636654853820801, | |
| "rewards/margins": 0.2620038390159607, | |
| "rewards/rejected": -0.725669264793396, | |
| "step": 400, | |
| "use_label": 7310.97509765625 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": 0.6419875621795654, | |
| "eval_logits/rejected": 0.764842689037323, | |
| "eval_logps/chosen": -113.52101135253906, | |
| "eval_logps/rejected": -148.9146270751953, | |
| "eval_loss": 0.6811794638633728, | |
| "eval_pred_label": 572.6825561523438, | |
| "eval_rewards/accuracies": 0.3551587164402008, | |
| "eval_rewards/chosen": -0.44620463252067566, | |
| "eval_rewards/margins": 0.2889607846736908, | |
| "eval_rewards/rejected": -0.7351653575897217, | |
| "eval_runtime": 247.9054, | |
| "eval_samples_per_second": 8.068, | |
| "eval_steps_per_second": 0.254, | |
| "eval_use_label": 7595.3173828125, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 2.671875, | |
| "learning_rate": 3.5249095128531863e-06, | |
| "logits/chosen": 0.34008723497390747, | |
| "logits/rejected": 0.07830000668764114, | |
| "logps/chosen": -100.46337890625, | |
| "logps/rejected": -124.54425048828125, | |
| "loss": 0.6735, | |
| "pred_label": 637.2999877929688, | |
| "rewards/accuracies": 0.3812499940395355, | |
| "rewards/chosen": -0.3595535457134247, | |
| "rewards/margins": 0.30695658922195435, | |
| "rewards/rejected": -0.6665101647377014, | |
| "step": 410, | |
| "use_label": 7860.7001953125 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 2.53125, | |
| "learning_rate": 3.4408477372034743e-06, | |
| "logits/chosen": 0.0810169205069542, | |
| "logits/rejected": 0.1531490534543991, | |
| "logps/chosen": -104.22071838378906, | |
| "logps/rejected": -134.35255432128906, | |
| "loss": 0.6774, | |
| "pred_label": 659.75, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -0.4057747721672058, | |
| "rewards/margins": 0.256902813911438, | |
| "rewards/rejected": -0.6626775860786438, | |
| "step": 420, | |
| "use_label": 7998.25 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 2.03125, | |
| "learning_rate": 3.355527661097728e-06, | |
| "logits/chosen": 0.1920831948518753, | |
| "logits/rejected": 0.327668160200119, | |
| "logps/chosen": -127.31324768066406, | |
| "logps/rejected": -138.50869750976562, | |
| "loss": 0.6761, | |
| "pred_label": 684.7249755859375, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -0.5084472894668579, | |
| "rewards/margins": 0.1930726319551468, | |
| "rewards/rejected": -0.7015198469161987, | |
| "step": 430, | |
| "use_label": 8133.27490234375 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 3.46875, | |
| "learning_rate": 3.269063392575352e-06, | |
| "logits/chosen": 0.2943039536476135, | |
| "logits/rejected": 0.04520421102643013, | |
| "logps/chosen": -116.11759185791016, | |
| "logps/rejected": -138.48878479003906, | |
| "loss": 0.6761, | |
| "pred_label": 708.7999877929688, | |
| "rewards/accuracies": 0.3187499940395355, | |
| "rewards/chosen": -0.48671650886535645, | |
| "rewards/margins": 0.26555269956588745, | |
| "rewards/rejected": -0.7522691488265991, | |
| "step": 440, | |
| "use_label": 8269.2001953125 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 2.046875, | |
| "learning_rate": 3.181570569931697e-06, | |
| "logits/chosen": 0.16918572783470154, | |
| "logits/rejected": -0.033099401742219925, | |
| "logps/chosen": -121.46165466308594, | |
| "logps/rejected": -148.68612670898438, | |
| "loss": 0.6878, | |
| "pred_label": 732.25, | |
| "rewards/accuracies": 0.34375, | |
| "rewards/chosen": -0.4864117503166199, | |
| "rewards/margins": 0.24023088812828064, | |
| "rewards/rejected": -0.7266427278518677, | |
| "step": 450, | |
| "use_label": 8405.75 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 2.109375, | |
| "learning_rate": 3.09316620706208e-06, | |
| "logits/chosen": 0.21930424869060516, | |
| "logits/rejected": 0.31035083532333374, | |
| "logps/chosen": -91.38786315917969, | |
| "logps/rejected": -110.15885925292969, | |
| "loss": 0.6718, | |
| "pred_label": 753.5999755859375, | |
| "rewards/accuracies": 0.3062500059604645, | |
| "rewards/chosen": -0.34389492869377136, | |
| "rewards/margins": 0.23962631821632385, | |
| "rewards/rejected": -0.5835212469100952, | |
| "step": 460, | |
| "use_label": 8544.400390625 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 2.453125, | |
| "learning_rate": 3.0039685369660785e-06, | |
| "logits/chosen": 0.31231826543807983, | |
| "logits/rejected": 0.4929059147834778, | |
| "logps/chosen": -109.46476745605469, | |
| "logps/rejected": -141.3244171142578, | |
| "loss": 0.6786, | |
| "pred_label": 777.9000244140625, | |
| "rewards/accuracies": 0.35624998807907104, | |
| "rewards/chosen": -0.4255266785621643, | |
| "rewards/margins": 0.30392220616340637, | |
| "rewards/rejected": -0.7294487953186035, | |
| "step": 470, | |
| "use_label": 8680.099609375 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 1.6328125, | |
| "learning_rate": 2.91409685362137e-06, | |
| "logits/chosen": 0.5682842135429382, | |
| "logits/rejected": 0.4352129399776459, | |
| "logps/chosen": -135.01535034179688, | |
| "logps/rejected": -157.18922424316406, | |
| "loss": 0.6756, | |
| "pred_label": 809.4000244140625, | |
| "rewards/accuracies": 0.3187499940395355, | |
| "rewards/chosen": -0.598731517791748, | |
| "rewards/margins": 0.19433310627937317, | |
| "rewards/rejected": -0.7930646538734436, | |
| "step": 480, | |
| "use_label": 8808.599609375 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 1.8359375, | |
| "learning_rate": 2.8236713524386085e-06, | |
| "logits/chosen": 0.13114799559116364, | |
| "logits/rejected": 0.3516528606414795, | |
| "logps/chosen": -117.50811767578125, | |
| "logps/rejected": -130.34207153320312, | |
| "loss": 0.6686, | |
| "pred_label": 831.2999877929688, | |
| "rewards/accuracies": 0.3062500059604645, | |
| "rewards/chosen": -0.5088413953781128, | |
| "rewards/margins": 0.19557976722717285, | |
| "rewards/rejected": -0.7044212222099304, | |
| "step": 490, | |
| "use_label": 8946.7001953125 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 1.9296875, | |
| "learning_rate": 2.7328129695107205e-06, | |
| "logits/chosen": 0.332313597202301, | |
| "logits/rejected": 0.04164884611964226, | |
| "logps/chosen": -148.10968017578125, | |
| "logps/rejected": -173.51394653320312, | |
| "loss": 0.6744, | |
| "pred_label": 849.9749755859375, | |
| "rewards/accuracies": 0.42500001192092896, | |
| "rewards/chosen": -0.7197140455245972, | |
| "rewards/margins": 0.24881935119628906, | |
| "rewards/rejected": -0.9685333967208862, | |
| "step": 500, | |
| "use_label": 9088.025390625 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "eval_logits/chosen": 0.6149206757545471, | |
| "eval_logits/rejected": 0.7128078937530518, | |
| "eval_logps/chosen": -120.11334228515625, | |
| "eval_logps/rejected": -151.15725708007812, | |
| "eval_loss": 0.672174334526062, | |
| "eval_pred_label": 893.8412475585938, | |
| "eval_rewards/accuracies": 0.341269850730896, | |
| "eval_rewards/chosen": -0.5121279954910278, | |
| "eval_rewards/margins": 0.24546381831169128, | |
| "eval_rewards/rejected": -0.7575918436050415, | |
| "eval_runtime": 247.8447, | |
| "eval_samples_per_second": 8.07, | |
| "eval_steps_per_second": 0.254, | |
| "eval_use_label": 9378.1591796875, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 1.7578125, | |
| "learning_rate": 2.641643219871597e-06, | |
| "logits/chosen": 0.4815472662448883, | |
| "logits/rejected": 0.2771294116973877, | |
| "logps/chosen": -110.22953033447266, | |
| "logps/rejected": -142.9767608642578, | |
| "loss": 0.6765, | |
| "pred_label": 940.7249755859375, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": -0.4598473608493805, | |
| "rewards/margins": 0.2644655704498291, | |
| "rewards/rejected": -0.724312961101532, | |
| "step": 510, | |
| "use_label": 9661.275390625 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 2.890625, | |
| "learning_rate": 2.5502840349805074e-06, | |
| "logits/chosen": 0.290465772151947, | |
| "logits/rejected": 0.05848363786935806, | |
| "logps/chosen": -115.6847915649414, | |
| "logps/rejected": -137.01820373535156, | |
| "loss": 0.6804, | |
| "pred_label": 966.2000122070312, | |
| "rewards/accuracies": 0.34375, | |
| "rewards/chosen": -0.4329158365726471, | |
| "rewards/margins": 0.21230947971343994, | |
| "rewards/rejected": -0.6452253460884094, | |
| "step": 520, | |
| "use_label": 9795.7998046875 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 2.015625, | |
| "learning_rate": 2.4588575996495797e-06, | |
| "logits/chosen": 0.5015053153038025, | |
| "logits/rejected": 0.544513463973999, | |
| "logps/chosen": -124.15202331542969, | |
| "logps/rejected": -145.21820068359375, | |
| "loss": 0.6847, | |
| "pred_label": 992.7999877929688, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -0.49871382117271423, | |
| "rewards/margins": 0.27802106738090515, | |
| "rewards/rejected": -0.7767347693443298, | |
| "step": 530, | |
| "use_label": 9929.2001953125 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 3.796875, | |
| "learning_rate": 2.367486188632446e-06, | |
| "logits/chosen": 0.7326034903526306, | |
| "logits/rejected": 0.5614863634109497, | |
| "logps/chosen": -132.87228393554688, | |
| "logps/rejected": -159.64352416992188, | |
| "loss": 0.6731, | |
| "pred_label": 1025.175048828125, | |
| "rewards/accuracies": 0.3812499940395355, | |
| "rewards/chosen": -0.5613355040550232, | |
| "rewards/margins": 0.30218708515167236, | |
| "rewards/rejected": -0.8635226488113403, | |
| "step": 540, | |
| "use_label": 10056.8251953125 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 2.4375, | |
| "learning_rate": 2.276292003092593e-06, | |
| "logits/chosen": 0.6115967631340027, | |
| "logits/rejected": 0.6694309711456299, | |
| "logps/chosen": -123.36216735839844, | |
| "logps/rejected": -129.92201232910156, | |
| "loss": 0.684, | |
| "pred_label": 1044.2249755859375, | |
| "rewards/accuracies": 0.28125, | |
| "rewards/chosen": -0.553015947341919, | |
| "rewards/margins": 0.14989802241325378, | |
| "rewards/rejected": -0.7029139995574951, | |
| "step": 550, | |
| "use_label": 10197.775390625 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 2.171875, | |
| "learning_rate": 2.1853970071701415e-06, | |
| "logits/chosen": 0.649623453617096, | |
| "logits/rejected": 0.5992484092712402, | |
| "logps/chosen": -117.4646987915039, | |
| "logps/rejected": -130.9988555908203, | |
| "loss": 0.6748, | |
| "pred_label": 1060.5999755859375, | |
| "rewards/accuracies": 0.29374998807907104, | |
| "rewards/chosen": -0.47909289598464966, | |
| "rewards/margins": 0.1609223484992981, | |
| "rewards/rejected": -0.6400152444839478, | |
| "step": 560, | |
| "use_label": 10341.400390625 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 2.5, | |
| "learning_rate": 2.0949227648656194e-06, | |
| "logits/chosen": 0.1570337414741516, | |
| "logits/rejected": 0.2956157624721527, | |
| "logps/chosen": -116.41545104980469, | |
| "logps/rejected": -151.84051513671875, | |
| "loss": 0.6765, | |
| "pred_label": 1081.375, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": -0.49815383553504944, | |
| "rewards/margins": 0.2765403389930725, | |
| "rewards/rejected": -0.7746941447257996, | |
| "step": 570, | |
| "use_label": 10480.625 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 2.09375, | |
| "learning_rate": 2.00499027745888e-06, | |
| "logits/chosen": 0.3431427478790283, | |
| "logits/rejected": 0.18610627949237823, | |
| "logps/chosen": -128.4036865234375, | |
| "logps/rejected": -150.36404418945312, | |
| "loss": 0.677, | |
| "pred_label": 1101.3499755859375, | |
| "rewards/accuracies": 0.33125001192092896, | |
| "rewards/chosen": -0.5349212288856506, | |
| "rewards/margins": 0.18440793454647064, | |
| "rewards/rejected": -0.7193291783332825, | |
| "step": 580, | |
| "use_label": 10620.650390625 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 2.21875, | |
| "learning_rate": 1.915719821680624e-06, | |
| "logits/chosen": 0.18862374126911163, | |
| "logits/rejected": 0.19857950508594513, | |
| "logps/chosen": -134.26577758789062, | |
| "logps/rejected": -165.70481872558594, | |
| "loss": 0.6624, | |
| "pred_label": 1121.5250244140625, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": -0.5457721948623657, | |
| "rewards/margins": 0.30153924226760864, | |
| "rewards/rejected": -0.8473113775253296, | |
| "step": 590, | |
| "use_label": 10760.474609375 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 3.125, | |
| "learning_rate": 1.8272307888529276e-06, | |
| "logits/chosen": 0.5807100534439087, | |
| "logits/rejected": 0.25763237476348877, | |
| "logps/chosen": -133.36561584472656, | |
| "logps/rejected": -180.56822204589844, | |
| "loss": 0.6784, | |
| "pred_label": 1148.5999755859375, | |
| "rewards/accuracies": 0.41874998807907104, | |
| "rewards/chosen": -0.5903924703598022, | |
| "rewards/margins": 0.3091353178024292, | |
| "rewards/rejected": -0.8995277285575867, | |
| "step": 600, | |
| "use_label": 10893.400390625 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": 0.8859585523605347, | |
| "eval_logits/rejected": 0.9939000606536865, | |
| "eval_logps/chosen": -119.97545623779297, | |
| "eval_logps/rejected": -156.7531280517578, | |
| "eval_loss": 0.6791760325431824, | |
| "eval_pred_label": 1206.1904296875, | |
| "eval_rewards/accuracies": 0.3511904776096344, | |
| "eval_rewards/chosen": -0.5107490420341492, | |
| "eval_rewards/margins": 0.30280154943466187, | |
| "eval_rewards/rejected": -0.8135506510734558, | |
| "eval_runtime": 247.9094, | |
| "eval_samples_per_second": 8.067, | |
| "eval_steps_per_second": 0.254, | |
| "eval_use_label": 11169.8095703125, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 3.234375, | |
| "learning_rate": 1.739641525213929e-06, | |
| "logits/chosen": 0.6684261560440063, | |
| "logits/rejected": 0.5376627445220947, | |
| "logps/chosen": -112.9523696899414, | |
| "logps/rejected": -154.13601684570312, | |
| "loss": 0.6711, | |
| "pred_label": 1272.925048828125, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -0.5404548645019531, | |
| "rewards/margins": 0.259638249874115, | |
| "rewards/rejected": -0.8000930547714233, | |
| "step": 610, | |
| "use_label": 11433.0751953125 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 2.296875, | |
| "learning_rate": 1.6530691736402317e-06, | |
| "logits/chosen": 0.6429753303527832, | |
| "logits/rejected": 0.5887765288352966, | |
| "logps/chosen": -125.70455169677734, | |
| "logps/rejected": -175.96624755859375, | |
| "loss": 0.6734, | |
| "pred_label": 1297.300048828125, | |
| "rewards/accuracies": 0.35624998807907104, | |
| "rewards/chosen": -0.6399649381637573, | |
| "rewards/margins": 0.3187193274497986, | |
| "rewards/rejected": -0.9586843252182007, | |
| "step": 620, | |
| "use_label": 11568.7001953125 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 2.046875, | |
| "learning_rate": 1.5676295169786864e-06, | |
| "logits/chosen": 0.8863061666488647, | |
| "logits/rejected": 0.5724608302116394, | |
| "logps/chosen": -149.8594970703125, | |
| "logps/rejected": -178.8180694580078, | |
| "loss": 0.6716, | |
| "pred_label": 1321.8499755859375, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -0.7683452367782593, | |
| "rewards/margins": 0.2745322585105896, | |
| "rewards/rejected": -1.042877435684204, | |
| "step": 630, | |
| "use_label": 11704.150390625 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 2.453125, | |
| "learning_rate": 1.4834368231970922e-06, | |
| "logits/chosen": 0.6763439178466797, | |
| "logits/rejected": 0.8844535946846008, | |
| "logps/chosen": -157.50503540039062, | |
| "logps/rejected": -179.59378051757812, | |
| "loss": 0.6738, | |
| "pred_label": 1351.8499755859375, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": -0.8581286668777466, | |
| "rewards/margins": 0.28607478737831116, | |
| "rewards/rejected": -1.1442034244537354, | |
| "step": 640, | |
| "use_label": 11834.150390625 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 2.921875, | |
| "learning_rate": 1.4006036925609245e-06, | |
| "logits/chosen": 0.5444064736366272, | |
| "logits/rejected": 1.0297753810882568, | |
| "logps/chosen": -149.0248565673828, | |
| "logps/rejected": -186.53790283203125, | |
| "loss": 0.6803, | |
| "pred_label": 1381.949951171875, | |
| "rewards/accuracies": 0.35624998807907104, | |
| "rewards/chosen": -0.7583541870117188, | |
| "rewards/margins": 0.32841619849205017, | |
| "rewards/rejected": -1.0867704153060913, | |
| "step": 650, | |
| "use_label": 11964.0498046875 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 2.21875, | |
| "learning_rate": 1.3192409070404582e-06, | |
| "logits/chosen": 1.0406488180160522, | |
| "logits/rejected": 0.8413463830947876, | |
| "logps/chosen": -115.27984619140625, | |
| "logps/rejected": -135.30836486816406, | |
| "loss": 0.6782, | |
| "pred_label": 1410.574951171875, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -0.5715335607528687, | |
| "rewards/margins": 0.21567881107330322, | |
| "rewards/rejected": -0.7872124910354614, | |
| "step": 660, | |
| "use_label": 12095.4248046875 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 2.28125, | |
| "learning_rate": 1.2394572821496953e-06, | |
| "logits/chosen": 0.3095243275165558, | |
| "logits/rejected": 0.21946246922016144, | |
| "logps/chosen": -121.89112854003906, | |
| "logps/rejected": -147.43978881835938, | |
| "loss": 0.6854, | |
| "pred_label": 1429.925048828125, | |
| "rewards/accuracies": 0.3187499940395355, | |
| "rewards/chosen": -0.5618449449539185, | |
| "rewards/margins": 0.2343917340040207, | |
| "rewards/rejected": -0.7962367534637451, | |
| "step": 670, | |
| "use_label": 12236.0751953125 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 1.9921875, | |
| "learning_rate": 1.1613595214152713e-06, | |
| "logits/chosen": 0.8114501237869263, | |
| "logits/rejected": 0.7095287442207336, | |
| "logps/chosen": -152.98135375976562, | |
| "logps/rejected": -175.5968475341797, | |
| "loss": 0.6781, | |
| "pred_label": 1451.175048828125, | |
| "rewards/accuracies": 0.3687500059604645, | |
| "rewards/chosen": -0.6692850589752197, | |
| "rewards/margins": 0.2704187035560608, | |
| "rewards/rejected": -0.9397038221359253, | |
| "step": 680, | |
| "use_label": 12374.8251953125 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 2.59375, | |
| "learning_rate": 1.0850520736699362e-06, | |
| "logits/chosen": 0.7043443322181702, | |
| "logits/rejected": 0.5973688364028931, | |
| "logps/chosen": -175.5553436279297, | |
| "logps/rejected": -207.06423950195312, | |
| "loss": 0.6749, | |
| "pred_label": 1481.625, | |
| "rewards/accuracies": 0.39375001192092896, | |
| "rewards/chosen": -0.7385014891624451, | |
| "rewards/margins": 0.37226757407188416, | |
| "rewards/rejected": -1.1107690334320068, | |
| "step": 690, | |
| "use_label": 12504.375 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 1.75, | |
| "learning_rate": 1.0106369933615043e-06, | |
| "logits/chosen": 0.44649118185043335, | |
| "logits/rejected": 0.6410871744155884, | |
| "logps/chosen": -136.0747833251953, | |
| "logps/rejected": -162.20260620117188, | |
| "loss": 0.6783, | |
| "pred_label": 1506.0999755859375, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -0.6972768902778625, | |
| "rewards/margins": 0.23985597491264343, | |
| "rewards/rejected": -0.9371329545974731, | |
| "step": 700, | |
| "use_label": 12639.900390625 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "eval_logits/chosen": 1.192717432975769, | |
| "eval_logits/rejected": 1.2994883060455322, | |
| "eval_logps/chosen": -135.23951721191406, | |
| "eval_logps/rejected": -171.3760528564453, | |
| "eval_loss": 0.6756439805030823, | |
| "eval_pred_label": 1558.5238037109375, | |
| "eval_rewards/accuracies": 0.3670634925365448, | |
| "eval_rewards/chosen": -0.6633896827697754, | |
| "eval_rewards/margins": 0.29639023542404175, | |
| "eval_rewards/rejected": -0.9597799181938171, | |
| "eval_runtime": 247.9992, | |
| "eval_samples_per_second": 8.065, | |
| "eval_steps_per_second": 0.254, | |
| "eval_use_label": 12921.4765625, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 3.421875, | |
| "learning_rate": 9.382138040640714e-07, | |
| "logits/chosen": 0.6493266820907593, | |
| "logits/rejected": 0.6850475072860718, | |
| "logps/chosen": -134.1044158935547, | |
| "logps/rejected": -165.8421173095703, | |
| "loss": 0.6766, | |
| "pred_label": 1616.875, | |
| "rewards/accuracies": 0.35624998807907104, | |
| "rewards/chosen": -0.7103394269943237, | |
| "rewards/margins": 0.2792840301990509, | |
| "rewards/rejected": -0.9896234273910522, | |
| "step": 710, | |
| "use_label": 13193.125 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 2.203125, | |
| "learning_rate": 8.678793653740633e-07, | |
| "logits/chosen": 0.7393421530723572, | |
| "logits/rejected": 0.6282132267951965, | |
| "logps/chosen": -108.0786361694336, | |
| "logps/rejected": -142.93173217773438, | |
| "loss": 0.6886, | |
| "pred_label": 1639.2750244140625, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -0.5120818614959717, | |
| "rewards/margins": 0.260955274105072, | |
| "rewards/rejected": -0.7730370759963989, | |
| "step": 720, | |
| "use_label": 13330.724609375 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 2.984375, | |
| "learning_rate": 7.997277433690984e-07, | |
| "logits/chosen": 0.7698175311088562, | |
| "logits/rejected": 0.6512314677238464, | |
| "logps/chosen": -120.15057373046875, | |
| "logps/rejected": -147.58602905273438, | |
| "loss": 0.6705, | |
| "pred_label": 1667.324951171875, | |
| "rewards/accuracies": 0.34375, | |
| "rewards/chosen": -0.49796366691589355, | |
| "rewards/margins": 0.33586567640304565, | |
| "rewards/rejected": -0.8338292837142944, | |
| "step": 730, | |
| "use_label": 13462.6748046875 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 1.8984375, | |
| "learning_rate": 7.338500848029603e-07, | |
| "logits/chosen": 0.9725875854492188, | |
| "logits/rejected": 0.894719123840332, | |
| "logps/chosen": -113.24635314941406, | |
| "logps/rejected": -142.79689025878906, | |
| "loss": 0.6779, | |
| "pred_label": 1692.25, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -0.5262097716331482, | |
| "rewards/margins": 0.26758259534835815, | |
| "rewards/rejected": -0.7937922477722168, | |
| "step": 740, | |
| "use_label": 13597.75 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 2.0, | |
| "learning_rate": 6.70334495204884e-07, | |
| "logits/chosen": 0.8061111569404602, | |
| "logits/rejected": 0.7181490659713745, | |
| "logps/chosen": -142.5282745361328, | |
| "logps/rejected": -173.8203887939453, | |
| "loss": 0.6814, | |
| "pred_label": 1716.25, | |
| "rewards/accuracies": 0.35624998807907104, | |
| "rewards/chosen": -0.6499794125556946, | |
| "rewards/margins": 0.24588195979595184, | |
| "rewards/rejected": -0.8958613276481628, | |
| "step": 750, | |
| "use_label": 13733.75 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 2.203125, | |
| "learning_rate": 6.092659210462232e-07, | |
| "logits/chosen": 0.6536890268325806, | |
| "logits/rejected": 0.6605783104896545, | |
| "logps/chosen": -102.72889709472656, | |
| "logps/rejected": -149.5058135986328, | |
| "loss": 0.6733, | |
| "pred_label": 1741.074951171875, | |
| "rewards/accuracies": 0.3187499940395355, | |
| "rewards/chosen": -0.48761066794395447, | |
| "rewards/margins": 0.28703850507736206, | |
| "rewards/rejected": -0.7746490836143494, | |
| "step": 760, | |
| "use_label": 13868.9248046875 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 2.375, | |
| "learning_rate": 5.507260361320738e-07, | |
| "logits/chosen": 0.6806662082672119, | |
| "logits/rejected": 0.5722958445549011, | |
| "logps/chosen": -151.7313690185547, | |
| "logps/rejected": -176.56260681152344, | |
| "loss": 0.6719, | |
| "pred_label": 1766.5250244140625, | |
| "rewards/accuracies": 0.40625, | |
| "rewards/chosen": -0.6818407773971558, | |
| "rewards/margins": 0.3464636206626892, | |
| "rewards/rejected": -1.0283044576644897, | |
| "step": 770, | |
| "use_label": 14003.474609375 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 3.03125, | |
| "learning_rate": 4.947931323697983e-07, | |
| "logits/chosen": 0.6961285471916199, | |
| "logits/rejected": 0.45934200286865234, | |
| "logps/chosen": -136.76901245117188, | |
| "logps/rejected": -162.4578094482422, | |
| "loss": 0.6822, | |
| "pred_label": 1797.9000244140625, | |
| "rewards/accuracies": 0.34375, | |
| "rewards/chosen": -0.6396945714950562, | |
| "rewards/margins": 0.2714308798313141, | |
| "rewards/rejected": -0.9111254811286926, | |
| "step": 780, | |
| "use_label": 14132.099609375 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 1.8046875, | |
| "learning_rate": 4.4154201506053985e-07, | |
| "logits/chosen": 0.824557900428772, | |
| "logits/rejected": 0.8706857562065125, | |
| "logps/chosen": -113.81227111816406, | |
| "logps/rejected": -128.26260375976562, | |
| "loss": 0.6771, | |
| "pred_label": 1823.199951171875, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -0.4897800385951996, | |
| "rewards/margins": 0.2714278995990753, | |
| "rewards/rejected": -0.7612079381942749, | |
| "step": 790, | |
| "use_label": 14266.7998046875 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 2.453125, | |
| "learning_rate": 3.910439028537638e-07, | |
| "logits/chosen": 0.6243492364883423, | |
| "logits/rejected": 0.7354862689971924, | |
| "logps/chosen": -111.455322265625, | |
| "logps/rejected": -141.9140625, | |
| "loss": 0.6776, | |
| "pred_label": 1847.300048828125, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": -0.5140770673751831, | |
| "rewards/margins": 0.28408390283584595, | |
| "rewards/rejected": -0.7981609106063843, | |
| "step": 800, | |
| "use_label": 14402.7001953125 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": 1.3586419820785522, | |
| "eval_logits/rejected": 1.4788893461227417, | |
| "eval_logps/chosen": -123.90097045898438, | |
| "eval_logps/rejected": -161.67913818359375, | |
| "eval_loss": 0.6800512671470642, | |
| "eval_pred_label": 1900.4920654296875, | |
| "eval_rewards/accuracies": 0.3531745970249176, | |
| "eval_rewards/chosen": -0.5500041842460632, | |
| "eval_rewards/margins": 0.312806636095047, | |
| "eval_rewards/rejected": -0.8628108501434326, | |
| "eval_runtime": 247.6869, | |
| "eval_samples_per_second": 8.075, | |
| "eval_steps_per_second": 0.254, | |
| "eval_use_label": 14683.5078125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 3.0, | |
| "learning_rate": 3.4336633249862084e-07, | |
| "logits/chosen": 0.6561521291732788, | |
| "logits/rejected": 0.6886910200119019, | |
| "logps/chosen": -128.8015899658203, | |
| "logps/rejected": -164.35073852539062, | |
| "loss": 0.678, | |
| "pred_label": 1961.699951171875, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -0.5894675254821777, | |
| "rewards/margins": 0.26825448870658875, | |
| "rewards/rejected": -0.8577221035957336, | |
| "step": 810, | |
| "use_label": 14952.2998046875 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 2.203125, | |
| "learning_rate": 2.98573068519539e-07, | |
| "logits/chosen": 0.7671118974685669, | |
| "logits/rejected": 0.7413855791091919, | |
| "logps/chosen": -113.36724853515625, | |
| "logps/rejected": -125.75687408447266, | |
| "loss": 0.6785, | |
| "pred_label": 1987.5, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -0.533771276473999, | |
| "rewards/margins": 0.18785560131072998, | |
| "rewards/rejected": -0.721626877784729, | |
| "step": 820, | |
| "use_label": 15086.5 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 1.7734375, | |
| "learning_rate": 2.5672401793681854e-07, | |
| "logits/chosen": 0.9424182176589966, | |
| "logits/rejected": 1.0951740741729736, | |
| "logps/chosen": -104.4168472290039, | |
| "logps/rejected": -135.86708068847656, | |
| "loss": 0.6757, | |
| "pred_label": 2014.300048828125, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -0.4637536108493805, | |
| "rewards/margins": 0.3284408450126648, | |
| "rewards/rejected": -0.7921944260597229, | |
| "step": 830, | |
| "use_label": 15219.7001953125 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 3.546875, | |
| "learning_rate": 2.178751501463036e-07, | |
| "logits/chosen": 0.880671501159668, | |
| "logits/rejected": 0.7017362117767334, | |
| "logps/chosen": -107.72086334228516, | |
| "logps/rejected": -113.772216796875, | |
| "loss": 0.6804, | |
| "pred_label": 2034.0250244140625, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -0.48421382904052734, | |
| "rewards/margins": 0.11718887090682983, | |
| "rewards/rejected": -0.6014026999473572, | |
| "step": 840, | |
| "use_label": 15359.974609375 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 2.1875, | |
| "learning_rate": 1.820784220652766e-07, | |
| "logits/chosen": 0.7102145552635193, | |
| "logits/rejected": 0.6271827816963196, | |
| "logps/chosen": -143.84104919433594, | |
| "logps/rejected": -181.9434814453125, | |
| "loss": 0.661, | |
| "pred_label": 2054.27490234375, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": -0.5962380170822144, | |
| "rewards/margins": 0.3919173777103424, | |
| "rewards/rejected": -0.9881553649902344, | |
| "step": 850, | |
| "use_label": 15499.724609375 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 2.03125, | |
| "learning_rate": 1.4938170864468636e-07, | |
| "logits/chosen": 0.856406569480896, | |
| "logits/rejected": 0.9847167730331421, | |
| "logps/chosen": -134.33340454101562, | |
| "logps/rejected": -162.52786254882812, | |
| "loss": 0.6682, | |
| "pred_label": 2086.074951171875, | |
| "rewards/accuracies": 0.42500001192092896, | |
| "rewards/chosen": -0.5313155651092529, | |
| "rewards/margins": 0.33722516894340515, | |
| "rewards/rejected": -0.8685407638549805, | |
| "step": 860, | |
| "use_label": 15627.9248046875 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 2.453125, | |
| "learning_rate": 1.1982873884064466e-07, | |
| "logits/chosen": 0.7517425417900085, | |
| "logits/rejected": 0.7345870137214661, | |
| "logps/chosen": -139.02523803710938, | |
| "logps/rejected": -179.33541870117188, | |
| "loss": 0.681, | |
| "pred_label": 2114.550048828125, | |
| "rewards/accuracies": 0.38749998807907104, | |
| "rewards/chosen": -0.5771310925483704, | |
| "rewards/margins": 0.393027126789093, | |
| "rewards/rejected": -0.9701582193374634, | |
| "step": 870, | |
| "use_label": 15759.4501953125 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 1.3359375, | |
| "learning_rate": 9.345903713082305e-08, | |
| "logits/chosen": 0.8886432647705078, | |
| "logits/rejected": 0.9018303751945496, | |
| "logps/chosen": -118.17820739746094, | |
| "logps/rejected": -171.14111328125, | |
| "loss": 0.6788, | |
| "pred_label": 2148.22509765625, | |
| "rewards/accuracies": 0.4000000059604645, | |
| "rewards/chosen": -0.5542714595794678, | |
| "rewards/margins": 0.41032201051712036, | |
| "rewards/rejected": -0.9645935297012329, | |
| "step": 880, | |
| "use_label": 15885.775390625 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 3.03125, | |
| "learning_rate": 7.030787065396866e-08, | |
| "logits/chosen": 0.9458627700805664, | |
| "logits/rejected": 0.860288143157959, | |
| "logps/chosen": -113.8023452758789, | |
| "logps/rejected": -145.21006774902344, | |
| "loss": 0.68, | |
| "pred_label": 2174.175048828125, | |
| "rewards/accuracies": 0.29374998807907104, | |
| "rewards/chosen": -0.4997434616088867, | |
| "rewards/margins": 0.2181231528520584, | |
| "rewards/rejected": -0.7178665399551392, | |
| "step": 890, | |
| "use_label": 16019.8251953125 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 2.265625, | |
| "learning_rate": 5.0406202043228604e-08, | |
| "logits/chosen": 1.2760592699050903, | |
| "logits/rejected": 1.0304285287857056, | |
| "logps/chosen": -123.82283020019531, | |
| "logps/rejected": -180.8885955810547, | |
| "loss": 0.6751, | |
| "pred_label": 2197.10009765625, | |
| "rewards/accuracies": 0.39375001192092896, | |
| "rewards/chosen": -0.5441134572029114, | |
| "rewards/margins": 0.37965571880340576, | |
| "rewards/rejected": -0.9237691164016724, | |
| "step": 900, | |
| "use_label": 16156.900390625 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "eval_logits/chosen": 1.3693251609802246, | |
| "eval_logits/rejected": 1.4904797077178955, | |
| "eval_logps/chosen": -123.65629577636719, | |
| "eval_logps/rejected": -161.58062744140625, | |
| "eval_loss": 0.678970456123352, | |
| "eval_pred_label": 2251.015869140625, | |
| "eval_rewards/accuracies": 0.3571428656578064, | |
| "eval_rewards/chosen": -0.547557532787323, | |
| "eval_rewards/margins": 0.31426796317100525, | |
| "eval_rewards/rejected": -0.8618254065513611, | |
| "eval_runtime": 247.8741, | |
| "eval_samples_per_second": 8.069, | |
| "eval_steps_per_second": 0.254, | |
| "eval_use_label": 16436.984375, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 1.8984375, | |
| "learning_rate": 3.378064801637687e-08, | |
| "logits/chosen": 0.8874324560165405, | |
| "logits/rejected": 0.9277682304382324, | |
| "logps/chosen": -107.72428131103516, | |
| "logps/rejected": -141.05999755859375, | |
| "loss": 0.6748, | |
| "pred_label": 2314.64990234375, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -0.47065839171409607, | |
| "rewards/margins": 0.3038247525691986, | |
| "rewards/rejected": -0.7744830846786499, | |
| "step": 910, | |
| "use_label": 16703.349609375 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 2.359375, | |
| "learning_rate": 2.0453443778310766e-08, | |
| "logits/chosen": 1.0454901456832886, | |
| "logits/rejected": 1.0777199268341064, | |
| "logps/chosen": -97.26091003417969, | |
| "logps/rejected": -147.07302856445312, | |
| "loss": 0.6799, | |
| "pred_label": 2345.5, | |
| "rewards/accuracies": 0.3187499940395355, | |
| "rewards/chosen": -0.4330506920814514, | |
| "rewards/margins": 0.3231905996799469, | |
| "rewards/rejected": -0.7562412023544312, | |
| "step": 920, | |
| "use_label": 16832.5 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 2.96875, | |
| "learning_rate": 1.0442413283435759e-08, | |
| "logits/chosen": 1.0506960153579712, | |
| "logits/rejected": 0.9065178632736206, | |
| "logps/chosen": -111.4305648803711, | |
| "logps/rejected": -146.4462432861328, | |
| "loss": 0.6772, | |
| "pred_label": 2373.125, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -0.4827675223350525, | |
| "rewards/margins": 0.2878049314022064, | |
| "rewards/rejected": -0.7705724835395813, | |
| "step": 930, | |
| "use_label": 16964.875 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 2.515625, | |
| "learning_rate": 3.760945397705828e-09, | |
| "logits/chosen": 1.234220027923584, | |
| "logits/rejected": 0.9605228304862976, | |
| "logps/chosen": -124.61787414550781, | |
| "logps/rejected": -160.95223999023438, | |
| "loss": 0.6706, | |
| "pred_label": 2397.75, | |
| "rewards/accuracies": 0.34375, | |
| "rewards/chosen": -0.5566731691360474, | |
| "rewards/margins": 0.27601632475852966, | |
| "rewards/rejected": -0.8326894640922546, | |
| "step": 940, | |
| "use_label": 17100.25 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 2.375, | |
| "learning_rate": 4.1797599220405605e-10, | |
| "logits/chosen": 0.7550326585769653, | |
| "logits/rejected": 0.6674235463142395, | |
| "logps/chosen": -136.30941772460938, | |
| "logps/rejected": -160.4894561767578, | |
| "loss": 0.6729, | |
| "pred_label": 2425.39990234375, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -0.5969915390014648, | |
| "rewards/margins": 0.21693304181098938, | |
| "rewards/rejected": -0.8139246106147766, | |
| "step": 950, | |
| "use_label": 17232.599609375 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6760230718482851, | |
| "train_runtime": 20063.9235, | |
| "train_samples_per_second": 3.047, | |
| "train_steps_per_second": 0.048 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 955, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |