| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.998691442030882, | |
| "eval_steps": 100, | |
| "global_step": 477, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.002093692750588851, | |
| "grad_norm": 4.870537596158184, | |
| "learning_rate": 4.166666666666666e-09, | |
| "logits/chosen": -7.072341442108154, | |
| "logits/rejected": -6.540944576263428, | |
| "logps/chosen": -346.17401123046875, | |
| "logps/rejected": -373.49456787109375, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02093692750588851, | |
| "grad_norm": 5.556671250581048, | |
| "learning_rate": 4.166666666666667e-08, | |
| "logits/chosen": -6.266731262207031, | |
| "logits/rejected": -6.149946212768555, | |
| "logps/chosen": -350.89019775390625, | |
| "logps/rejected": -355.3866271972656, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.4305555522441864, | |
| "rewards/chosen": 0.000294171943096444, | |
| "rewards/margins": 0.0003653134626802057, | |
| "rewards/rejected": -7.114150503184646e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04187385501177702, | |
| "grad_norm": 5.947484840987328, | |
| "learning_rate": 8.333333333333334e-08, | |
| "logits/chosen": -6.901098728179932, | |
| "logits/rejected": -6.428101539611816, | |
| "logps/chosen": -368.4015197753906, | |
| "logps/rejected": -332.3041687011719, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.000805316842161119, | |
| "rewards/margins": 0.0013799300650134683, | |
| "rewards/rejected": -0.0005746129900217056, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06281078251766553, | |
| "grad_norm": 4.7415695218728455, | |
| "learning_rate": 1.25e-07, | |
| "logits/chosen": -6.833685874938965, | |
| "logits/rejected": -6.713122367858887, | |
| "logps/chosen": -381.8787536621094, | |
| "logps/rejected": -335.8555603027344, | |
| "loss": 0.6922, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.0064066750928759575, | |
| "rewards/margins": 0.0018916327971965075, | |
| "rewards/rejected": 0.004515042062848806, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08374771002355404, | |
| "grad_norm": 4.440630542343055, | |
| "learning_rate": 1.6666666666666668e-07, | |
| "logits/chosen": -7.030734062194824, | |
| "logits/rejected": -6.716243743896484, | |
| "logps/chosen": -374.63531494140625, | |
| "logps/rejected": -338.505126953125, | |
| "loss": 0.6894, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.025013629347085953, | |
| "rewards/margins": 0.007752637378871441, | |
| "rewards/rejected": 0.017260991036891937, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10468463752944256, | |
| "grad_norm": 4.761237773867083, | |
| "learning_rate": 1.9998927475076103e-07, | |
| "logits/chosen": -6.727326393127441, | |
| "logits/rejected": -6.21535587310791, | |
| "logps/chosen": -338.68768310546875, | |
| "logps/rejected": -324.8238830566406, | |
| "loss": 0.6853, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.04892030358314514, | |
| "rewards/margins": 0.014415493234992027, | |
| "rewards/rejected": 0.034504808485507965, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12562156503533106, | |
| "grad_norm": 5.119072453981655, | |
| "learning_rate": 1.9961413253717213e-07, | |
| "logits/chosen": -7.186664581298828, | |
| "logits/rejected": -6.633403778076172, | |
| "logps/chosen": -326.31439208984375, | |
| "logps/rejected": -301.0088195800781, | |
| "loss": 0.6768, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.07451288402080536, | |
| "rewards/margins": 0.03354056924581528, | |
| "rewards/rejected": 0.04097231104969978, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14655849254121958, | |
| "grad_norm": 5.007390249181196, | |
| "learning_rate": 1.9870502626379125e-07, | |
| "logits/chosen": -7.373200416564941, | |
| "logits/rejected": -7.114656925201416, | |
| "logps/chosen": -356.11444091796875, | |
| "logps/rejected": -357.88812255859375, | |
| "loss": 0.6678, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.04506208747625351, | |
| "rewards/margins": 0.0459139421582222, | |
| "rewards/rejected": -0.0008518520626239479, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16749542004710807, | |
| "grad_norm": 5.277515253273024, | |
| "learning_rate": 1.9726682903510838e-07, | |
| "logits/chosen": -8.06989860534668, | |
| "logits/rejected": -7.776799201965332, | |
| "logps/chosen": -383.57763671875, | |
| "logps/rejected": -360.22412109375, | |
| "loss": 0.6567, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.008996413089334965, | |
| "rewards/margins": 0.0934341698884964, | |
| "rewards/rejected": -0.08443775773048401, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.1884323475529966, | |
| "grad_norm": 5.718904864627125, | |
| "learning_rate": 1.9530725005474194e-07, | |
| "logits/chosen": -8.55430793762207, | |
| "logits/rejected": -8.116801261901855, | |
| "logps/chosen": -343.30426025390625, | |
| "logps/rejected": -344.4271545410156, | |
| "loss": 0.6418, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.09966368973255157, | |
| "rewards/margins": 0.11764608323574066, | |
| "rewards/rejected": -0.21730978786945343, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "grad_norm": 6.2104459729271335, | |
| "learning_rate": 1.9283679330160724e-07, | |
| "logits/chosen": -8.740680694580078, | |
| "logits/rejected": -8.447057723999023, | |
| "logps/chosen": -355.4083557128906, | |
| "logps/rejected": -336.2757263183594, | |
| "loss": 0.6273, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.18184241652488708, | |
| "rewards/margins": 0.17646023631095886, | |
| "rewards/rejected": -0.35830265283584595, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "eval_logits/chosen": -9.086737632751465, | |
| "eval_logits/rejected": -8.745743751525879, | |
| "eval_logps/chosen": -391.9050598144531, | |
| "eval_logps/rejected": -385.3993835449219, | |
| "eval_loss": 0.6256434321403503, | |
| "eval_rewards/accuracies": 0.6800000071525574, | |
| "eval_rewards/chosen": -0.2307606041431427, | |
| "eval_rewards/margins": 0.16034501791000366, | |
| "eval_rewards/rejected": -0.391105592250824, | |
| "eval_runtime": 241.3872, | |
| "eval_samples_per_second": 8.285, | |
| "eval_steps_per_second": 1.036, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23030620256477363, | |
| "grad_norm": 6.2752590431966375, | |
| "learning_rate": 1.898687012251826e-07, | |
| "logits/chosen": -9.274371147155762, | |
| "logits/rejected": -9.027790069580078, | |
| "logps/chosen": -357.4116516113281, | |
| "logps/rejected": -360.20465087890625, | |
| "loss": 0.6273, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.2726437747478485, | |
| "rewards/margins": 0.17124271392822266, | |
| "rewards/rejected": -0.44388651847839355, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.2512431300706621, | |
| "grad_norm": 6.345691652928505, | |
| "learning_rate": 1.8641888376168482e-07, | |
| "logits/chosen": -9.697042465209961, | |
| "logits/rejected": -9.629292488098145, | |
| "logps/chosen": -399.30670166015625, | |
| "logps/rejected": -386.90557861328125, | |
| "loss": 0.6059, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.28612643480300903, | |
| "rewards/margins": 0.24121908843517303, | |
| "rewards/rejected": -0.5273455381393433, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2721800575765506, | |
| "grad_norm": 7.047283511946962, | |
| "learning_rate": 1.8250583305165094e-07, | |
| "logits/chosen": -10.05250358581543, | |
| "logits/rejected": -9.847970008850098, | |
| "logps/chosen": -378.30462646484375, | |
| "logps/rejected": -395.7349548339844, | |
| "loss": 0.5913, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.3639756739139557, | |
| "rewards/margins": 0.2538781762123108, | |
| "rewards/rejected": -0.6178538799285889, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29311698508243916, | |
| "grad_norm": 8.250566588104485, | |
| "learning_rate": 1.78150524316067e-07, | |
| "logits/chosen": -10.428335189819336, | |
| "logits/rejected": -10.18110179901123, | |
| "logps/chosen": -423.37896728515625, | |
| "logps/rejected": -418.40814208984375, | |
| "loss": 0.5918, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.36636847257614136, | |
| "rewards/margins": 0.3312264084815979, | |
| "rewards/rejected": -0.6975948214530945, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31405391258832765, | |
| "grad_norm": 7.160872331336618, | |
| "learning_rate": 1.7337630342238038e-07, | |
| "logits/chosen": -10.668364524841309, | |
| "logits/rejected": -10.247892379760742, | |
| "logps/chosen": -438.568603515625, | |
| "logps/rejected": -439.31243896484375, | |
| "loss": 0.5943, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.44665470719337463, | |
| "rewards/margins": 0.35836517810821533, | |
| "rewards/rejected": -0.8050198554992676, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33499084009421615, | |
| "grad_norm": 8.012025922238234, | |
| "learning_rate": 1.682087617430782e-07, | |
| "logits/chosen": -10.813860893249512, | |
| "logits/rejected": -10.625129699707031, | |
| "logps/chosen": -402.5640563964844, | |
| "logps/rejected": -415.65032958984375, | |
| "loss": 0.5833, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.584823489189148, | |
| "rewards/margins": 0.2518795430660248, | |
| "rewards/rejected": -0.8367029428482056, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.3559277676001047, | |
| "grad_norm": 8.371912815572413, | |
| "learning_rate": 1.6267559897763025e-07, | |
| "logits/chosen": -10.72344970703125, | |
| "logits/rejected": -10.272022247314453, | |
| "logps/chosen": -365.57806396484375, | |
| "logps/rejected": -445.9237365722656, | |
| "loss": 0.5881, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.5119096040725708, | |
| "rewards/margins": 0.36096295714378357, | |
| "rewards/rejected": -0.872872531414032, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.3768646951059932, | |
| "grad_norm": 8.291785272423704, | |
| "learning_rate": 1.5680647467311557e-07, | |
| "logits/chosen": -11.077351570129395, | |
| "logits/rejected": -10.810079574584961, | |
| "logps/chosen": -403.3762512207031, | |
| "logps/rejected": -436.90411376953125, | |
| "loss": 0.5611, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.5242770910263062, | |
| "rewards/margins": 0.4779755473136902, | |
| "rewards/rejected": -1.0022525787353516, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.39780162261188173, | |
| "grad_norm": 8.768402559709461, | |
| "learning_rate": 1.506328492394303e-07, | |
| "logits/chosen": -10.848445892333984, | |
| "logits/rejected": -10.747730255126953, | |
| "logps/chosen": -361.96917724609375, | |
| "logps/rejected": -427.53131103515625, | |
| "loss": 0.5678, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.6229602098464966, | |
| "rewards/margins": 0.3965635299682617, | |
| "rewards/rejected": -1.0195238590240479, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "grad_norm": 10.10996155431003, | |
| "learning_rate": 1.4418781531128634e-07, | |
| "logits/chosen": -11.236587524414062, | |
| "logits/rejected": -11.044486045837402, | |
| "logps/chosen": -414.62225341796875, | |
| "logps/rejected": -419.2083435058594, | |
| "loss": 0.5701, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.6426066160202026, | |
| "rewards/margins": 0.4211527705192566, | |
| "rewards/rejected": -1.0637595653533936, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "eval_logits/chosen": -11.197042465209961, | |
| "eval_logits/rejected": -10.930534362792969, | |
| "eval_logps/chosen": -442.0617370605469, | |
| "eval_logps/rejected": -458.7671203613281, | |
| "eval_loss": 0.5679408311843872, | |
| "eval_rewards/accuracies": 0.6800000071525574, | |
| "eval_rewards/chosen": -0.7323274612426758, | |
| "eval_rewards/margins": 0.39245596528053284, | |
| "eval_rewards/rejected": -1.1247833967208862, | |
| "eval_runtime": 238.8989, | |
| "eval_samples_per_second": 8.372, | |
| "eval_steps_per_second": 1.046, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4396754776236587, | |
| "grad_norm": 8.838084489326292, | |
| "learning_rate": 1.375059203609562e-07, | |
| "logits/chosen": -11.455360412597656, | |
| "logits/rejected": -11.19542121887207, | |
| "logps/chosen": -448.05706787109375, | |
| "logps/rejected": -442.106689453125, | |
| "loss": 0.579, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.6968704462051392, | |
| "rewards/margins": 0.5110000967979431, | |
| "rewards/rejected": -1.207870364189148, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46061240512954726, | |
| "grad_norm": 11.66423950154552, | |
| "learning_rate": 1.306229815126159e-07, | |
| "logits/chosen": -10.951299667358398, | |
| "logits/rejected": -10.760453224182129, | |
| "logps/chosen": -387.4031066894531, | |
| "logps/rejected": -438.4293518066406, | |
| "loss": 0.5669, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.6763008832931519, | |
| "rewards/margins": 0.459463506937027, | |
| "rewards/rejected": -1.1357643604278564, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48154933263543576, | |
| "grad_norm": 8.962170857968381, | |
| "learning_rate": 1.2357589355094274e-07, | |
| "logits/chosen": -11.39826488494873, | |
| "logits/rejected": -11.283321380615234, | |
| "logps/chosen": -385.0740966796875, | |
| "logps/rejected": -439.373291015625, | |
| "loss": 0.5679, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.6969426870346069, | |
| "rewards/margins": 0.4374031126499176, | |
| "rewards/rejected": -1.1343457698822021, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5024862601413242, | |
| "grad_norm": 9.228744844021563, | |
| "learning_rate": 1.1640243115310217e-07, | |
| "logits/chosen": -11.526562690734863, | |
| "logits/rejected": -11.38569450378418, | |
| "logps/chosen": -431.7037048339844, | |
| "logps/rejected": -451.6121520996094, | |
| "loss": 0.5411, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.6818773150444031, | |
| "rewards/margins": 0.4736524522304535, | |
| "rewards/rejected": -1.1555297374725342, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5234231876472127, | |
| "grad_norm": 9.475835105940853, | |
| "learning_rate": 1.0914104640422679e-07, | |
| "logits/chosen": -11.973138809204102, | |
| "logits/rejected": -11.678539276123047, | |
| "logps/chosen": -438.6128845214844, | |
| "logps/rejected": -494.81475830078125, | |
| "loss": 0.5498, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.7578285932540894, | |
| "rewards/margins": 0.5199133157730103, | |
| "rewards/rejected": -1.2777419090270996, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5443601151531012, | |
| "grad_norm": 10.55001524910548, | |
| "learning_rate": 1.0183066268176774e-07, | |
| "logits/chosen": -11.676985740661621, | |
| "logits/rejected": -11.725804328918457, | |
| "logps/chosen": -406.6064453125, | |
| "logps/rejected": -454.85015869140625, | |
| "loss": 0.5528, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.6437905430793762, | |
| "rewards/margins": 0.597993791103363, | |
| "rewards/rejected": -1.2417843341827393, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5652970426589898, | |
| "grad_norm": 10.40303673883758, | |
| "learning_rate": 9.451046601356724e-08, | |
| "logits/chosen": -11.600053787231445, | |
| "logits/rejected": -11.353797912597656, | |
| "logps/chosen": -433.164794921875, | |
| "logps/rejected": -443.1178283691406, | |
| "loss": 0.5492, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.8422502279281616, | |
| "rewards/margins": 0.5629727244377136, | |
| "rewards/rejected": -1.40522301197052, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5862339701648783, | |
| "grad_norm": 10.92362637516926, | |
| "learning_rate": 8.721969502803953e-08, | |
| "logits/chosen": -11.430012702941895, | |
| "logits/rejected": -11.168180465698242, | |
| "logps/chosen": -443.4029846191406, | |
| "logps/rejected": -476.8294982910156, | |
| "loss": 0.5538, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.9203767776489258, | |
| "rewards/margins": 0.4110565781593323, | |
| "rewards/rejected": -1.3314332962036133, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6071708976707668, | |
| "grad_norm": 9.76250775652993, | |
| "learning_rate": 7.999743062239557e-08, | |
| "logits/chosen": -11.536198616027832, | |
| "logits/rejected": -11.403619766235352, | |
| "logps/chosen": -425.9610900878906, | |
| "logps/rejected": -467.23291015625, | |
| "loss": 0.5452, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.7125166654586792, | |
| "rewards/margins": 0.6340970396995544, | |
| "rewards/rejected": -1.3466136455535889, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "grad_norm": 12.956036139139355, | |
| "learning_rate": 7.28823864763583e-08, | |
| "logits/chosen": -11.646090507507324, | |
| "logits/rejected": -11.440874099731445, | |
| "logps/chosen": -433.30120849609375, | |
| "logps/rejected": -482.5865173339844, | |
| "loss": 0.5398, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.790636420249939, | |
| "rewards/margins": 0.5609198808670044, | |
| "rewards/rejected": -1.351556420326233, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "eval_logits/chosen": -11.52165412902832, | |
| "eval_logits/rejected": -11.35400676727295, | |
| "eval_logps/chosen": -458.75299072265625, | |
| "eval_logps/rejected": -483.2173156738281, | |
| "eval_loss": 0.5490842461585999, | |
| "eval_rewards/accuracies": 0.6840000152587891, | |
| "eval_rewards/chosen": -0.8992397785186768, | |
| "eval_rewards/margins": 0.47004497051239014, | |
| "eval_rewards/rejected": -1.369284749031067, | |
| "eval_runtime": 222.3735, | |
| "eval_samples_per_second": 8.994, | |
| "eval_steps_per_second": 1.124, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6490447526825438, | |
| "grad_norm": 11.066017903277775, | |
| "learning_rate": 6.591270153428288e-08, | |
| "logits/chosen": -11.68547248840332, | |
| "logits/rejected": -11.603824615478516, | |
| "logps/chosen": -465.458740234375, | |
| "logps/rejected": -478.3704528808594, | |
| "loss": 0.5356, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8401340246200562, | |
| "rewards/margins": 0.5598424673080444, | |
| "rewards/rejected": -1.3999764919281006, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6699816801884323, | |
| "grad_norm": 11.177050758944128, | |
| "learning_rate": 5.912573556804452e-08, | |
| "logits/chosen": -11.293741226196289, | |
| "logits/rejected": -11.169164657592773, | |
| "logps/chosen": -391.907958984375, | |
| "logps/rejected": -435.26190185546875, | |
| "loss": 0.5324, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.7623863220214844, | |
| "rewards/margins": 0.5452743768692017, | |
| "rewards/rejected": -1.3076608180999756, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6909186076943209, | |
| "grad_norm": 13.652856840334575, | |
| "learning_rate": 5.255786891654399e-08, | |
| "logits/chosen": -11.769269943237305, | |
| "logits/rejected": -11.3505220413208, | |
| "logps/chosen": -447.8871154785156, | |
| "logps/rejected": -496.74029541015625, | |
| "loss": 0.537, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8857334852218628, | |
| "rewards/margins": 0.6645745038986206, | |
| "rewards/rejected": -1.5503078699111938, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7118555352002094, | |
| "grad_norm": 11.172216313095676, | |
| "learning_rate": 4.624430747529102e-08, | |
| "logits/chosen": -11.618499755859375, | |
| "logits/rejected": -11.424800872802734, | |
| "logps/chosen": -420.2229919433594, | |
| "logps/rejected": -472.84033203125, | |
| "loss": 0.5363, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.7942701578140259, | |
| "rewards/margins": 0.6123701333999634, | |
| "rewards/rejected": -1.4066402912139893, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7327924627060979, | |
| "grad_norm": 12.255527795835215, | |
| "learning_rate": 4.0218893981385925e-08, | |
| "logits/chosen": -11.575703620910645, | |
| "logits/rejected": -11.422430038452148, | |
| "logps/chosen": -432.9644470214844, | |
| "logps/rejected": -470.8561096191406, | |
| "loss": 0.5185, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.8515139818191528, | |
| "rewards/margins": 0.5291236042976379, | |
| "rewards/rejected": -1.380637526512146, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7537293902119864, | |
| "grad_norm": 12.049137551707718, | |
| "learning_rate": 3.45139266054715e-08, | |
| "logits/chosen": -11.708954811096191, | |
| "logits/rejected": -11.447134017944336, | |
| "logps/chosen": -482.3265686035156, | |
| "logps/rejected": -521.8561401367188, | |
| "loss": 0.547, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.885657787322998, | |
| "rewards/margins": 0.5782418251037598, | |
| "rewards/rejected": -1.4638997316360474, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7746663177178749, | |
| "grad_norm": 10.956723652484643, | |
| "learning_rate": 2.9159985823062993e-08, | |
| "logits/chosen": -11.063775062561035, | |
| "logits/rejected": -11.105363845825195, | |
| "logps/chosen": -469.81201171875, | |
| "logps/rejected": -485.8446350097656, | |
| "loss": 0.54, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.8363800048828125, | |
| "rewards/margins": 0.5975291132926941, | |
| "rewards/rejected": -1.4339090585708618, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7956032452237635, | |
| "grad_norm": 13.069162308279825, | |
| "learning_rate": 2.4185770493280577e-08, | |
| "logits/chosen": -11.262879371643066, | |
| "logits/rejected": -11.126230239868164, | |
| "logps/chosen": -455.2259216308594, | |
| "logps/rejected": -489.8641052246094, | |
| "loss": 0.5364, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.9804290533065796, | |
| "rewards/margins": 0.48143139481544495, | |
| "rewards/rejected": -1.4618604183197021, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.816540172729652, | |
| "grad_norm": 13.1101074898303, | |
| "learning_rate": 1.9617944023656108e-08, | |
| "logits/chosen": -11.852777481079102, | |
| "logits/rejected": -11.575521469116211, | |
| "logps/chosen": -470.88604736328125, | |
| "logps/rejected": -511.36767578125, | |
| "loss": 0.532, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.876979649066925, | |
| "rewards/margins": 0.5152709484100342, | |
| "rewards/rejected": -1.3922507762908936, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "grad_norm": 13.484360798114164, | |
| "learning_rate": 1.5480991445620538e-08, | |
| "logits/chosen": -11.511263847351074, | |
| "logits/rejected": -11.029703140258789, | |
| "logps/chosen": -427.28448486328125, | |
| "logps/rejected": -486.9627990722656, | |
| "loss": 0.54, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.8370214700698853, | |
| "rewards/margins": 0.5594587922096252, | |
| "rewards/rejected": -1.3964803218841553, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "eval_logits/chosen": -11.385127067565918, | |
| "eval_logits/rejected": -11.129400253295898, | |
| "eval_logps/chosen": -459.30474853515625, | |
| "eval_logps/rejected": -487.0408020019531, | |
| "eval_loss": 0.5449032783508301, | |
| "eval_rewards/accuracies": 0.6759999990463257, | |
| "eval_rewards/chosen": -0.9047574996948242, | |
| "eval_rewards/margins": 0.5027627944946289, | |
| "eval_rewards/rejected": -1.4075202941894531, | |
| "eval_runtime": 222.5645, | |
| "eval_samples_per_second": 8.986, | |
| "eval_steps_per_second": 1.123, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8584140277414289, | |
| "grad_norm": 10.485330642188247, | |
| "learning_rate": 1.1797088166794e-08, | |
| "logits/chosen": -11.086836814880371, | |
| "logits/rejected": -10.8326997756958, | |
| "logps/chosen": -452.47235107421875, | |
| "logps/rejected": -453.66986083984375, | |
| "loss": 0.5401, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.9103538393974304, | |
| "rewards/margins": 0.4986189007759094, | |
| "rewards/rejected": -1.4089727401733398, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8793509552473174, | |
| "grad_norm": 10.664608353191511, | |
| "learning_rate": 8.585981103608341e-09, | |
| "logits/chosen": -11.340084075927734, | |
| "logits/rejected": -10.990386962890625, | |
| "logps/chosen": -402.55157470703125, | |
| "logps/rejected": -500.97467041015625, | |
| "loss": 0.531, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.8159112930297852, | |
| "rewards/margins": 0.6232641339302063, | |
| "rewards/rejected": -1.4391753673553467, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9002878827532059, | |
| "grad_norm": 11.582461841758496, | |
| "learning_rate": 5.864882831430273e-09, | |
| "logits/chosen": -11.550569534301758, | |
| "logits/rejected": -11.248498916625977, | |
| "logps/chosen": -449.3023986816406, | |
| "logps/rejected": -463.66510009765625, | |
| "loss": 0.5232, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8865207433700562, | |
| "rewards/margins": 0.47660237550735474, | |
| "rewards/rejected": -1.3631231784820557, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9212248102590945, | |
| "grad_norm": 11.301516689755074, | |
| "learning_rate": 3.6483793195745682e-09, | |
| "logits/chosen": -11.318781852722168, | |
| "logits/rejected": -10.964235305786133, | |
| "logps/chosen": -413.6148986816406, | |
| "logps/rejected": -463.11114501953125, | |
| "loss": 0.5351, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.8221774101257324, | |
| "rewards/margins": 0.663087010383606, | |
| "rewards/rejected": -1.4852644205093384, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.942161737764983, | |
| "grad_norm": 12.872962412868715, | |
| "learning_rate": 1.9483517457776433e-09, | |
| "logits/chosen": -11.27305793762207, | |
| "logits/rejected": -10.98169994354248, | |
| "logps/chosen": -440.1998596191406, | |
| "logps/rejected": -488.0940856933594, | |
| "loss": 0.5273, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.8431445956230164, | |
| "rewards/margins": 0.6032007932662964, | |
| "rewards/rejected": -1.446345329284668, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9630986652708715, | |
| "grad_norm": 11.52042506596002, | |
| "learning_rate": 7.739128092312918e-10, | |
| "logits/chosen": -11.074236869812012, | |
| "logits/rejected": -10.837442398071289, | |
| "logps/chosen": -452.1619567871094, | |
| "logps/rejected": -474.42620849609375, | |
| "loss": 0.5402, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.8461356163024902, | |
| "rewards/margins": 0.6280852556228638, | |
| "rewards/rejected": -1.4742207527160645, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98403559277676, | |
| "grad_norm": 11.458055654937185, | |
| "learning_rate": 1.313578835593465e-10, | |
| "logits/chosen": -11.382277488708496, | |
| "logits/rejected": -11.017861366271973, | |
| "logps/chosen": -448.569091796875, | |
| "logps/rejected": -476.3909606933594, | |
| "loss": 0.5184, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.9083318710327148, | |
| "rewards/margins": 0.5944851636886597, | |
| "rewards/rejected": -1.502817153930664, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.998691442030882, | |
| "step": 477, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5785154289169632, | |
| "train_runtime": 17436.3201, | |
| "train_samples_per_second": 3.506, | |
| "train_steps_per_second": 0.027 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 477, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |