| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9993060374739764, | |
| "eval_steps": 1000, | |
| "global_step": 720, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0013879250520471894, | |
| "grad_norm": 2.184646703877197, | |
| "learning_rate": 6.9444444444444435e-09, | |
| "logits/chosen": -1.157708764076233, | |
| "logits/rejected": -1.0856982469558716, | |
| "logps/chosen": -392.3704528808594, | |
| "logps/rejected": -422.7169189453125, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.013879250520471894, | |
| "grad_norm": 2.045186292766699, | |
| "learning_rate": 6.944444444444444e-08, | |
| "logits/chosen": -0.9201799035072327, | |
| "logits/rejected": -0.9960015416145325, | |
| "logps/chosen": -403.1753234863281, | |
| "logps/rejected": -406.3473815917969, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.3680555522441864, | |
| "rewards/chosen": -0.0007031817804090679, | |
| "rewards/margins": -0.0005899361567571759, | |
| "rewards/rejected": -0.00011324579827487469, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.027758501040943788, | |
| "grad_norm": 2.0501238605790117, | |
| "learning_rate": 1.3888888888888888e-07, | |
| "logits/chosen": -0.9155173301696777, | |
| "logits/rejected": -0.9719367027282715, | |
| "logps/chosen": -393.12945556640625, | |
| "logps/rejected": -407.33709716796875, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": 0.0007024414371699095, | |
| "rewards/margins": 0.00011867408466059715, | |
| "rewards/rejected": 0.0005837674252688885, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.041637751561415685, | |
| "grad_norm": 2.1931925008563304, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -0.9494641423225403, | |
| "logits/rejected": -1.0066778659820557, | |
| "logps/chosen": -429.01910400390625, | |
| "logps/rejected": -425.8172912597656, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.559374988079071, | |
| "rewards/chosen": 0.0010106399422511458, | |
| "rewards/margins": 0.0010079689091071486, | |
| "rewards/rejected": 2.671018592081964e-06, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.055517002081887576, | |
| "grad_norm": 2.196961132429192, | |
| "learning_rate": 2.7777777777777776e-07, | |
| "logits/chosen": -0.9400655627250671, | |
| "logits/rejected": -1.0188666582107544, | |
| "logps/chosen": -420.2276916503906, | |
| "logps/rejected": -405.5096130371094, | |
| "loss": 0.6923, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.002629968337714672, | |
| "rewards/margins": 0.001661272719502449, | |
| "rewards/rejected": 0.0009686955017969012, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.06939625260235947, | |
| "grad_norm": 2.2620008213708194, | |
| "learning_rate": 3.472222222222222e-07, | |
| "logits/chosen": -0.9268127679824829, | |
| "logits/rejected": -0.9413528442382812, | |
| "logps/chosen": -407.0869140625, | |
| "logps/rejected": -433.43035888671875, | |
| "loss": 0.691, | |
| "rewards/accuracies": 0.703125, | |
| "rewards/chosen": 0.006026268471032381, | |
| "rewards/margins": 0.005162273999303579, | |
| "rewards/rejected": 0.0008639938896521926, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.08327550312283137, | |
| "grad_norm": 2.9642516438971254, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -0.9514287710189819, | |
| "logits/rejected": -1.0300030708312988, | |
| "logps/chosen": -408.86181640625, | |
| "logps/rejected": -417.53106689453125, | |
| "loss": 0.6884, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": 0.01152682676911354, | |
| "rewards/margins": 0.00950426422059536, | |
| "rewards/rejected": 0.002022563014179468, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.09715475364330327, | |
| "grad_norm": 2.0059535680251264, | |
| "learning_rate": 4.861111111111111e-07, | |
| "logits/chosen": -0.9322255849838257, | |
| "logits/rejected": -0.9961916208267212, | |
| "logps/chosen": -416.08197021484375, | |
| "logps/rejected": -440.123291015625, | |
| "loss": 0.6845, | |
| "rewards/accuracies": 0.721875011920929, | |
| "rewards/chosen": 0.02161681093275547, | |
| "rewards/margins": 0.016515256837010384, | |
| "rewards/rejected": 0.005101555027067661, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.11103400416377515, | |
| "grad_norm": 2.959027274351803, | |
| "learning_rate": 4.998119881260575e-07, | |
| "logits/chosen": -0.8885990381240845, | |
| "logits/rejected": -0.9473252296447754, | |
| "logps/chosen": -412.66253662109375, | |
| "logps/rejected": -426.8203125, | |
| "loss": 0.6784, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.03722007945179939, | |
| "rewards/margins": 0.03375672921538353, | |
| "rewards/rejected": 0.0034633490722626448, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.12491325468424705, | |
| "grad_norm": 2.896496746709743, | |
| "learning_rate": 4.990486745229364e-07, | |
| "logits/chosen": -0.9239265322685242, | |
| "logits/rejected": -1.015549659729004, | |
| "logps/chosen": -406.7080078125, | |
| "logps/rejected": -408.1942138671875, | |
| "loss": 0.6692, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": 0.04666774719953537, | |
| "rewards/margins": 0.0515812449157238, | |
| "rewards/rejected": -0.004913502838462591, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.13879250520471895, | |
| "grad_norm": 2.17437482262988, | |
| "learning_rate": 4.977001008412112e-07, | |
| "logits/chosen": -1.0140012502670288, | |
| "logits/rejected": -1.0407038927078247, | |
| "logps/chosen": -416.13800048828125, | |
| "logps/rejected": -424.24468994140625, | |
| "loss": 0.6619, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": 0.06093335896730423, | |
| "rewards/margins": 0.05933469533920288, | |
| "rewards/rejected": 0.0015986515209078789, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.15267175572519084, | |
| "grad_norm": 2.0661816686924537, | |
| "learning_rate": 4.957694362057149e-07, | |
| "logits/chosen": -0.9822956919670105, | |
| "logits/rejected": -1.011765956878662, | |
| "logps/chosen": -417.8843688964844, | |
| "logps/rejected": -422.9627380371094, | |
| "loss": 0.6518, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.05249622464179993, | |
| "rewards/margins": 0.07984234392642975, | |
| "rewards/rejected": -0.027346113696694374, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.16655100624566274, | |
| "grad_norm": 2.1244191981703135, | |
| "learning_rate": 4.932612176449559e-07, | |
| "logits/chosen": -0.9970356225967407, | |
| "logits/rejected": -1.0578349828720093, | |
| "logps/chosen": -407.3737487792969, | |
| "logps/rejected": -426.0777893066406, | |
| "loss": 0.6368, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.04393979534506798, | |
| "rewards/margins": 0.10807327926158905, | |
| "rewards/rejected": -0.06413348764181137, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.18043025676613464, | |
| "grad_norm": 2.0595896023881703, | |
| "learning_rate": 4.901813394291801e-07, | |
| "logits/chosen": -0.9531866908073425, | |
| "logits/rejected": -0.9673601984977722, | |
| "logps/chosen": -416.93572998046875, | |
| "logps/rejected": -443.47894287109375, | |
| "loss": 0.6201, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 0.03279733285307884, | |
| "rewards/margins": 0.1790144443511963, | |
| "rewards/rejected": -0.14621710777282715, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.19430950728660654, | |
| "grad_norm": 2.169975326300987, | |
| "learning_rate": 4.865370392189376e-07, | |
| "logits/chosen": -0.9693125486373901, | |
| "logits/rejected": -0.9949439167976379, | |
| "logps/chosen": -396.1827697753906, | |
| "logps/rejected": -439.85198974609375, | |
| "loss": 0.6004, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 0.005037306807935238, | |
| "rewards/margins": 0.23115964233875275, | |
| "rewards/rejected": -0.22612233459949493, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.2081887578070784, | |
| "grad_norm": 2.439547013921171, | |
| "learning_rate": 4.823368810567056e-07, | |
| "logits/chosen": -1.0007877349853516, | |
| "logits/rejected": -1.0211200714111328, | |
| "logps/chosen": -408.7399597167969, | |
| "logps/rejected": -449.01953125, | |
| "loss": 0.5878, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -0.07737134397029877, | |
| "rewards/margins": 0.25105008482933044, | |
| "rewards/rejected": -0.3284214437007904, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.2220680083275503, | |
| "grad_norm": 2.191957070219856, | |
| "learning_rate": 4.775907352415367e-07, | |
| "logits/chosen": -0.9597848057746887, | |
| "logits/rejected": -0.9754905700683594, | |
| "logps/chosen": -404.9871826171875, | |
| "logps/rejected": -453.3285217285156, | |
| "loss": 0.572, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.14126327633857727, | |
| "rewards/margins": 0.30146196484565735, | |
| "rewards/rejected": -0.44272518157958984, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.2359472588480222, | |
| "grad_norm": 2.8446228567237726, | |
| "learning_rate": 4.723097551340265e-07, | |
| "logits/chosen": -0.8638654947280884, | |
| "logits/rejected": -0.8823292851448059, | |
| "logps/chosen": -451.3196716308594, | |
| "logps/rejected": -487.5765075683594, | |
| "loss": 0.5457, | |
| "rewards/accuracies": 0.784375011920929, | |
| "rewards/chosen": -0.2892600893974304, | |
| "rewards/margins": 0.37542229890823364, | |
| "rewards/rejected": -0.6646823287010193, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.2498265093684941, | |
| "grad_norm": 2.488211046081056, | |
| "learning_rate": 4.6650635094610966e-07, | |
| "logits/chosen": -0.9351462125778198, | |
| "logits/rejected": -0.9417260885238647, | |
| "logps/chosen": -442.58917236328125, | |
| "logps/rejected": -491.6702575683594, | |
| "loss": 0.5337, | |
| "rewards/accuracies": 0.784375011920929, | |
| "rewards/chosen": -0.40233302116394043, | |
| "rewards/margins": 0.40838655829429626, | |
| "rewards/rejected": -0.8107194900512695, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.263705759888966, | |
| "grad_norm": 3.2325426784137417, | |
| "learning_rate": 4.6019416057727577e-07, | |
| "logits/chosen": -0.9113900065422058, | |
| "logits/rejected": -0.9249471426010132, | |
| "logps/chosen": -454.93658447265625, | |
| "logps/rejected": -505.32763671875, | |
| "loss": 0.5093, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.5173531174659729, | |
| "rewards/margins": 0.5457069277763367, | |
| "rewards/rejected": -1.0630600452423096, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.2775850104094379, | |
| "grad_norm": 2.6935717929946374, | |
| "learning_rate": 4.5338801756574185e-07, | |
| "logits/chosen": -0.9233020544052124, | |
| "logits/rejected": -0.936613917350769, | |
| "logps/chosen": -477.2933654785156, | |
| "logps/rejected": -543.9907836914062, | |
| "loss": 0.5088, | |
| "rewards/accuracies": 0.765625, | |
| "rewards/chosen": -0.698850154876709, | |
| "rewards/margins": 0.6066737174987793, | |
| "rewards/rejected": -1.3055237531661987, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2914642609299098, | |
| "grad_norm": 2.6234937800693237, | |
| "learning_rate": 4.461039162298939e-07, | |
| "logits/chosen": -0.9328230619430542, | |
| "logits/rejected": -0.8883553743362427, | |
| "logps/chosen": -474.24176025390625, | |
| "logps/rejected": -560.80810546875, | |
| "loss": 0.4809, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.7412362694740295, | |
| "rewards/margins": 0.7680382132530212, | |
| "rewards/rejected": -1.5092744827270508, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.3053435114503817, | |
| "grad_norm": 3.024837883769618, | |
| "learning_rate": 4.3835897408191513e-07, | |
| "logits/chosen": -0.9258328676223755, | |
| "logits/rejected": -0.902255654335022, | |
| "logps/chosen": -545.10205078125, | |
| "logps/rejected": -611.480224609375, | |
| "loss": 0.4716, | |
| "rewards/accuracies": 0.7906249761581421, | |
| "rewards/chosen": -0.9928766489028931, | |
| "rewards/margins": 0.8125397562980652, | |
| "rewards/rejected": -1.8054163455963135, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.3192227619708536, | |
| "grad_norm": 4.999154415916733, | |
| "learning_rate": 4.301713916019286e-07, | |
| "logits/chosen": -0.864261269569397, | |
| "logits/rejected": -0.8518573045730591, | |
| "logps/chosen": -521.3760986328125, | |
| "logps/rejected": -621.6141967773438, | |
| "loss": 0.4618, | |
| "rewards/accuracies": 0.8343750238418579, | |
| "rewards/chosen": -1.0177581310272217, | |
| "rewards/margins": 1.0421375036239624, | |
| "rewards/rejected": -2.0598955154418945, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.3331020124913255, | |
| "grad_norm": 4.221789042426902, | |
| "learning_rate": 4.2156040946718343e-07, | |
| "logits/chosen": -0.8536814451217651, | |
| "logits/rejected": -0.855249285697937, | |
| "logps/chosen": -554.1260986328125, | |
| "logps/rejected": -660.2291259765625, | |
| "loss": 0.4504, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.2824825048446655, | |
| "rewards/margins": 0.9512511491775513, | |
| "rewards/rejected": -2.233733654022217, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.3469812630117974, | |
| "grad_norm": 2.9920792128032283, | |
| "learning_rate": 4.125462633367959e-07, | |
| "logits/chosen": -0.9102590680122375, | |
| "logits/rejected": -0.8657618761062622, | |
| "logps/chosen": -548.0431518554688, | |
| "logps/rejected": -677.654541015625, | |
| "loss": 0.4387, | |
| "rewards/accuracies": 0.809374988079071, | |
| "rewards/chosen": -1.2187608480453491, | |
| "rewards/margins": 1.2359611988067627, | |
| "rewards/rejected": -2.4547219276428223, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.3608605135322693, | |
| "grad_norm": 3.276352885973741, | |
| "learning_rate": 4.031501362983007e-07, | |
| "logits/chosen": -0.9509126543998718, | |
| "logits/rejected": -0.8994159698486328, | |
| "logps/chosen": -542.1051635742188, | |
| "logps/rejected": -668.0875244140625, | |
| "loss": 0.44, | |
| "rewards/accuracies": 0.815625011920929, | |
| "rewards/chosen": -1.3456058502197266, | |
| "rewards/margins": 1.152848243713379, | |
| "rewards/rejected": -2.4984538555145264, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.3747397640527412, | |
| "grad_norm": 3.433959749322587, | |
| "learning_rate": 3.933941090877615e-07, | |
| "logits/chosen": -0.9043784141540527, | |
| "logits/rejected": -0.8728653192520142, | |
| "logps/chosen": -572.6734619140625, | |
| "logps/rejected": -690.0535278320312, | |
| "loss": 0.4269, | |
| "rewards/accuracies": 0.796875, | |
| "rewards/chosen": -1.49461030960083, | |
| "rewards/margins": 1.185040831565857, | |
| "rewards/rejected": -2.6796510219573975, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.3886190145732131, | |
| "grad_norm": 3.135829679670571, | |
| "learning_rate": 3.833011082004228e-07, | |
| "logits/chosen": -0.9181197285652161, | |
| "logits/rejected": -0.8538551330566406, | |
| "logps/chosen": -580.3753662109375, | |
| "logps/rejected": -723.9002685546875, | |
| "loss": 0.4243, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.563520908355713, | |
| "rewards/margins": 1.419384479522705, | |
| "rewards/rejected": -2.982905387878418, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.4024982650936849, | |
| "grad_norm": 3.4471828982674606, | |
| "learning_rate": 3.728948520138426e-07, | |
| "logits/chosen": -0.8911579847335815, | |
| "logits/rejected": -0.8040667772293091, | |
| "logps/chosen": -581.972412109375, | |
| "logps/rejected": -704.8370361328125, | |
| "loss": 0.4129, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.598099946975708, | |
| "rewards/margins": 1.2950611114501953, | |
| "rewards/rejected": -2.893160820007324, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.4163775156141568, | |
| "grad_norm": 3.4538574024709185, | |
| "learning_rate": 3.6219979505011555e-07, | |
| "logits/chosen": -0.9019178152084351, | |
| "logits/rejected": -0.8638097047805786, | |
| "logps/chosen": -552.6265869140625, | |
| "logps/rejected": -678.7213745117188, | |
| "loss": 0.4264, | |
| "rewards/accuracies": 0.815625011920929, | |
| "rewards/chosen": -1.5648345947265625, | |
| "rewards/margins": 1.3646188974380493, | |
| "rewards/rejected": -2.9294533729553223, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.4302567661346287, | |
| "grad_norm": 4.0164936619199905, | |
| "learning_rate": 3.512410705081684e-07, | |
| "logits/chosen": -0.8717131614685059, | |
| "logits/rejected": -0.8151350021362305, | |
| "logps/chosen": -573.2371215820312, | |
| "logps/rejected": -731.2991333007812, | |
| "loss": 0.4135, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -1.5790045261383057, | |
| "rewards/margins": 1.489929437637329, | |
| "rewards/rejected": -3.0689339637756348, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.4441360166551006, | |
| "grad_norm": 3.5055099662080766, | |
| "learning_rate": 3.400444312011776e-07, | |
| "logits/chosen": -0.9052634239196777, | |
| "logits/rejected": -0.8478328585624695, | |
| "logps/chosen": -601.0723266601562, | |
| "logps/rejected": -743.2720947265625, | |
| "loss": 0.3968, | |
| "rewards/accuracies": 0.809374988079071, | |
| "rewards/chosen": -1.5770965814590454, | |
| "rewards/margins": 1.5724782943725586, | |
| "rewards/rejected": -3.1495749950408936, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.4580152671755725, | |
| "grad_norm": 3.965189906947426, | |
| "learning_rate": 3.286361890379034e-07, | |
| "logits/chosen": -0.8289991617202759, | |
| "logits/rejected": -0.7707743644714355, | |
| "logps/chosen": -592.8557739257812, | |
| "logps/rejected": -759.8111572265625, | |
| "loss": 0.397, | |
| "rewards/accuracies": 0.8031250238418579, | |
| "rewards/chosen": -1.7602096796035767, | |
| "rewards/margins": 1.649737000465393, | |
| "rewards/rejected": -3.409946918487549, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.4718945176960444, | |
| "grad_norm": 6.571433183841374, | |
| "learning_rate": 3.1704315319015936e-07, | |
| "logits/chosen": -0.8428624868392944, | |
| "logits/rejected": -0.807705283164978, | |
| "logps/chosen": -569.0560302734375, | |
| "logps/rejected": -749.4913330078125, | |
| "loss": 0.4054, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -1.711888074874878, | |
| "rewards/margins": 1.6754896640777588, | |
| "rewards/rejected": -3.387377977371216, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.4857737682165163, | |
| "grad_norm": 3.3895657051926116, | |
| "learning_rate": 3.052925670917219e-07, | |
| "logits/chosen": -0.8292319178581238, | |
| "logits/rejected": -0.7592617273330688, | |
| "logps/chosen": -615.757080078125, | |
| "logps/rejected": -775.6668090820312, | |
| "loss": 0.3867, | |
| "rewards/accuracies": 0.7718750238418579, | |
| "rewards/chosen": -1.9108030796051025, | |
| "rewards/margins": 1.574349045753479, | |
| "rewards/rejected": -3.48515248298645, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.4996530187369882, | |
| "grad_norm": 3.9806648944209155, | |
| "learning_rate": 2.934120444167326e-07, | |
| "logits/chosen": -0.822270393371582, | |
| "logits/rejected": -0.7529654502868652, | |
| "logps/chosen": -610.9036865234375, | |
| "logps/rejected": -785.0739135742188, | |
| "loss": 0.3914, | |
| "rewards/accuracies": 0.815625011920929, | |
| "rewards/chosen": -2.06144380569458, | |
| "rewards/margins": 1.6444549560546875, | |
| "rewards/rejected": -3.7058987617492676, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.5135322692574601, | |
| "grad_norm": 4.270423330616301, | |
| "learning_rate": 2.814295041880407e-07, | |
| "logits/chosen": -0.8429604768753052, | |
| "logits/rejected": -0.7863871455192566, | |
| "logps/chosen": -599.1904296875, | |
| "logps/rejected": -767.7243041992188, | |
| "loss": 0.391, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.957845687866211, | |
| "rewards/margins": 1.5971133708953857, | |
| "rewards/rejected": -3.5549590587615967, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.527411519777932, | |
| "grad_norm": 3.711197534301825, | |
| "learning_rate": 2.6937310516798275e-07, | |
| "logits/chosen": -0.8030338287353516, | |
| "logits/rejected": -0.7218085527420044, | |
| "logps/chosen": -635.15380859375, | |
| "logps/rejected": -831.8131103515625, | |
| "loss": 0.3798, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -2.005901336669922, | |
| "rewards/margins": 1.8555805683135986, | |
| "rewards/rejected": -3.8614819049835205, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.5412907702984039, | |
| "grad_norm": 4.065093944544315, | |
| "learning_rate": 2.5727117968577785e-07, | |
| "logits/chosen": -0.7184926867485046, | |
| "logits/rejected": -0.6400619745254517, | |
| "logps/chosen": -647.2830810546875, | |
| "logps/rejected": -867.3894653320312, | |
| "loss": 0.3737, | |
| "rewards/accuracies": 0.8531249761581421, | |
| "rewards/chosen": -2.1855947971343994, | |
| "rewards/margins": 2.0819170475006104, | |
| "rewards/rejected": -4.26751184463501, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.5551700208188758, | |
| "grad_norm": 4.664675650179661, | |
| "learning_rate": 2.4515216705704393e-07, | |
| "logits/chosen": -0.7386522889137268, | |
| "logits/rejected": -0.6554276347160339, | |
| "logps/chosen": -680.4814453125, | |
| "logps/rejected": -875.1360473632812, | |
| "loss": 0.3768, | |
| "rewards/accuracies": 0.809374988079071, | |
| "rewards/chosen": -2.458164691925049, | |
| "rewards/margins": 1.9175488948822021, | |
| "rewards/rejected": -4.37571382522583, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.5690492713393477, | |
| "grad_norm": 4.082366486193136, | |
| "learning_rate": 2.330445467518977e-07, | |
| "logits/chosen": -0.7022604942321777, | |
| "logits/rejected": -0.5973988175392151, | |
| "logps/chosen": -627.4280395507812, | |
| "logps/rejected": -786.0429077148438, | |
| "loss": 0.377, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.2749786376953125, | |
| "rewards/margins": 1.6706234216690063, | |
| "rewards/rejected": -3.94560170173645, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.5829285218598196, | |
| "grad_norm": 3.6701003128420386, | |
| "learning_rate": 2.209767714686924e-07, | |
| "logits/chosen": -0.672639787197113, | |
| "logits/rejected": -0.5875508189201355, | |
| "logps/chosen": -665.8011474609375, | |
| "logps/rejected": -878.91015625, | |
| "loss": 0.3537, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -2.3907861709594727, | |
| "rewards/margins": 2.144279956817627, | |
| "rewards/rejected": -4.5350661277771, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.5968077723802915, | |
| "grad_norm": 3.8553830986355764, | |
| "learning_rate": 2.0897720027066897e-07, | |
| "logits/chosen": -0.67729651927948, | |
| "logits/rejected": -0.5586274862289429, | |
| "logps/chosen": -661.219482421875, | |
| "logps/rejected": -875.0428466796875, | |
| "loss": 0.3715, | |
| "rewards/accuracies": 0.815625011920929, | |
| "rewards/chosen": -2.52473783493042, | |
| "rewards/margins": 2.1635537147521973, | |
| "rewards/rejected": -4.688291072845459, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.6106870229007634, | |
| "grad_norm": 4.080685306612081, | |
| "learning_rate": 1.970740319426474e-07, | |
| "logits/chosen": -0.65470290184021, | |
| "logits/rejected": -0.5402953028678894, | |
| "logps/chosen": -674.3743896484375, | |
| "logps/rejected": -868.76611328125, | |
| "loss": 0.3756, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -2.5917928218841553, | |
| "rewards/margins": 1.9496173858642578, | |
| "rewards/rejected": -4.54141092300415, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.6245662734212353, | |
| "grad_norm": 3.7177874253136913, | |
| "learning_rate": 1.8529523872436977e-07, | |
| "logits/chosen": -0.6337302327156067, | |
| "logits/rejected": -0.5308721661567688, | |
| "logps/chosen": -659.3401489257812, | |
| "logps/rejected": -889.49462890625, | |
| "loss": 0.3548, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.4972152709960938, | |
| "rewards/margins": 2.2179877758026123, | |
| "rewards/rejected": -4.715203285217285, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.6384455239417072, | |
| "grad_norm": 3.781558261259466, | |
| "learning_rate": 1.7366850057622172e-07, | |
| "logits/chosen": -0.627717137336731, | |
| "logits/rejected": -0.5284063816070557, | |
| "logps/chosen": -665.6905517578125, | |
| "logps/rejected": -902.7052612304688, | |
| "loss": 0.3565, | |
| "rewards/accuracies": 0.8031250238418579, | |
| "rewards/chosen": -2.4058690071105957, | |
| "rewards/margins": 2.2597124576568604, | |
| "rewards/rejected": -4.665581703186035, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.6523247744621791, | |
| "grad_norm": 3.624902658770024, | |
| "learning_rate": 1.622211401318028e-07, | |
| "logits/chosen": -0.5916509628295898, | |
| "logits/rejected": -0.48228612542152405, | |
| "logps/chosen": -683.482421875, | |
| "logps/rejected": -909.9552612304688, | |
| "loss": 0.3597, | |
| "rewards/accuracies": 0.840624988079071, | |
| "rewards/chosen": -2.451106071472168, | |
| "rewards/margins": 2.257086992263794, | |
| "rewards/rejected": -4.708193302154541, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.666204024982651, | |
| "grad_norm": 4.657777670569194, | |
| "learning_rate": 1.5098005849021078e-07, | |
| "logits/chosen": -0.5622087717056274, | |
| "logits/rejected": -0.46162405610084534, | |
| "logps/chosen": -661.2738037109375, | |
| "logps/rejected": -859.2135009765625, | |
| "loss": 0.3586, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -2.3591208457946777, | |
| "rewards/margins": 2.061002016067505, | |
| "rewards/rejected": -4.4201226234436035, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.6800832755031229, | |
| "grad_norm": 6.251774756777437, | |
| "learning_rate": 1.3997167199892385e-07, | |
| "logits/chosen": -0.5363645553588867, | |
| "logits/rejected": -0.39701658487319946, | |
| "logps/chosen": -659.409912109375, | |
| "logps/rejected": -862.7174072265625, | |
| "loss": 0.3703, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -2.3470611572265625, | |
| "rewards/margins": 2.0817208290100098, | |
| "rewards/rejected": -4.428781986236572, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.6939625260235948, | |
| "grad_norm": 4.437192094325177, | |
| "learning_rate": 1.2922185017584036e-07, | |
| "logits/chosen": -0.47550851106643677, | |
| "logits/rejected": -0.33102065324783325, | |
| "logps/chosen": -642.2039794921875, | |
| "logps/rejected": -921.5857543945312, | |
| "loss": 0.341, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -2.3800008296966553, | |
| "rewards/margins": 2.652977466583252, | |
| "rewards/rejected": -5.032978534698486, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.7078417765440667, | |
| "grad_norm": 4.426537023215993, | |
| "learning_rate": 1.1875585491635998e-07, | |
| "logits/chosen": -0.5427089333534241, | |
| "logits/rejected": -0.3993372321128845, | |
| "logps/chosen": -642.7872924804688, | |
| "logps/rejected": -917.6959838867188, | |
| "loss": 0.339, | |
| "rewards/accuracies": 0.859375, | |
| "rewards/chosen": -2.4717633724212646, | |
| "rewards/margins": 2.438523054122925, | |
| "rewards/rejected": -4.910286903381348, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.7217210270645386, | |
| "grad_norm": 3.8981108408032368, | |
| "learning_rate": 1.0859828112836539e-07, | |
| "logits/chosen": -0.4982399344444275, | |
| "logits/rejected": -0.3629917502403259, | |
| "logps/chosen": -662.3302001953125, | |
| "logps/rejected": -924.0505981445312, | |
| "loss": 0.3575, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -2.5424540042877197, | |
| "rewards/margins": 2.429400682449341, | |
| "rewards/rejected": -4.9718546867370605, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.7356002775850105, | |
| "grad_norm": 4.062590769922296, | |
| "learning_rate": 9.877299893461455e-08, | |
| "logits/chosen": -0.4675068259239197, | |
| "logits/rejected": -0.32594671845436096, | |
| "logps/chosen": -694.9034423828125, | |
| "logps/rejected": -893.17724609375, | |
| "loss": 0.3619, | |
| "rewards/accuracies": 0.828125, | |
| "rewards/chosen": -2.6744236946105957, | |
| "rewards/margins": 2.1196742057800293, | |
| "rewards/rejected": -4.794097900390625, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.7494795281054824, | |
| "grad_norm": 4.00277278125475, | |
| "learning_rate": 8.930309757836516e-08, | |
| "logits/chosen": -0.4969969689846039, | |
| "logits/rejected": -0.33853963017463684, | |
| "logps/chosen": -672.2896118164062, | |
| "logps/rejected": -913.6337890625, | |
| "loss": 0.3511, | |
| "rewards/accuracies": 0.809374988079071, | |
| "rewards/chosen": -2.6400508880615234, | |
| "rewards/margins": 2.4135284423828125, | |
| "rewards/rejected": -5.053578853607178, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.7633587786259542, | |
| "grad_norm": 4.354943956619566, | |
| "learning_rate": 8.021083116405173e-08, | |
| "logits/chosen": -0.45121559500694275, | |
| "logits/rejected": -0.34631314873695374, | |
| "logps/chosen": -652.8707885742188, | |
| "logps/rejected": -885.6276245117188, | |
| "loss": 0.358, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -2.4929685592651367, | |
| "rewards/margins": 2.304232120513916, | |
| "rewards/rejected": -4.797201156616211, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.7772380291464261, | |
| "grad_norm": 4.835099186465556, | |
| "learning_rate": 7.151756636052527e-08, | |
| "logits/chosen": -0.4754953384399414, | |
| "logits/rejected": -0.3613481819629669, | |
| "logps/chosen": -677.6926879882812, | |
| "logps/rejected": -924.0533447265625, | |
| "loss": 0.3606, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.6459193229675293, | |
| "rewards/margins": 2.297471523284912, | |
| "rewards/rejected": -4.943390846252441, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.7911172796668979, | |
| "grad_norm": 4.2131174135927765, | |
| "learning_rate": 6.324373218975104e-08, | |
| "logits/chosen": -0.44653385877609253, | |
| "logits/rejected": -0.3059902787208557, | |
| "logps/chosen": -659.1906127929688, | |
| "logps/rejected": -913.2021484375, | |
| "loss": 0.3647, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -2.6062073707580566, | |
| "rewards/margins": 2.3299379348754883, | |
| "rewards/rejected": -4.936145782470703, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.8049965301873698, | |
| "grad_norm": 4.3409883619650795, | |
| "learning_rate": 5.5408772018959996e-08, | |
| "logits/chosen": -0.44951170682907104, | |
| "logits/rejected": -0.3005351424217224, | |
| "logps/chosen": -656.5030517578125, | |
| "logps/rejected": -886.4778442382812, | |
| "loss": 0.3535, | |
| "rewards/accuracies": 0.815625011920929, | |
| "rewards/chosen": -2.515747547149658, | |
| "rewards/margins": 2.289696455001831, | |
| "rewards/rejected": -4.805444240570068, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.8188757807078417, | |
| "grad_norm": 13.005915598801213, | |
| "learning_rate": 4.8031097869072225e-08, | |
| "logits/chosen": -0.4361554980278015, | |
| "logits/rejected": -0.3123430609703064, | |
| "logps/chosen": -669.0994873046875, | |
| "logps/rejected": -921.1182861328125, | |
| "loss": 0.3566, | |
| "rewards/accuracies": 0.840624988079071, | |
| "rewards/chosen": -2.5885722637176514, | |
| "rewards/margins": 2.3980278968811035, | |
| "rewards/rejected": -4.986600399017334, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.8327550312283136, | |
| "grad_norm": 4.507720819165163, | |
| "learning_rate": 4.112804714676593e-08, | |
| "logits/chosen": -0.4602123200893402, | |
| "logits/rejected": -0.3107297122478485, | |
| "logps/chosen": -671.499267578125, | |
| "logps/rejected": -886.6002807617188, | |
| "loss": 0.3403, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -2.505514144897461, | |
| "rewards/margins": 2.2939445972442627, | |
| "rewards/rejected": -4.7994585037231445, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.8466342817487855, | |
| "grad_norm": 4.108687652880907, | |
| "learning_rate": 3.4715841901871545e-08, | |
| "logits/chosen": -0.4734552800655365, | |
| "logits/rejected": -0.31184083223342896, | |
| "logps/chosen": -664.9038696289062, | |
| "logps/rejected": -929.7863159179688, | |
| "loss": 0.3487, | |
| "rewards/accuracies": 0.846875011920929, | |
| "rewards/chosen": -2.5965065956115723, | |
| "rewards/margins": 2.5841736793518066, | |
| "rewards/rejected": -5.180680751800537, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.8605135322692574, | |
| "grad_norm": 4.599039860177261, | |
| "learning_rate": 2.8809550705835546e-08, | |
| "logits/chosen": -0.43759673833847046, | |
| "logits/rejected": -0.25794321298599243, | |
| "logps/chosen": -715.013671875, | |
| "logps/rejected": -970.8460083007812, | |
| "loss": 0.3524, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -2.732679843902588, | |
| "rewards/margins": 2.611907482147217, | |
| "rewards/rejected": -5.344587802886963, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.8743927827897293, | |
| "grad_norm": 4.4811322085497105, | |
| "learning_rate": 2.3423053240837514e-08, | |
| "logits/chosen": -0.4395861029624939, | |
| "logits/rejected": -0.27333635091781616, | |
| "logps/chosen": -696.0693359375, | |
| "logps/rejected": -928.7999877929688, | |
| "loss": 0.3694, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.7578296661376953, | |
| "rewards/margins": 2.2534213066101074, | |
| "rewards/rejected": -5.011250972747803, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.8882720333102012, | |
| "grad_norm": 4.763510092667976, | |
| "learning_rate": 1.8569007682777415e-08, | |
| "logits/chosen": -0.4241456985473633, | |
| "logits/rejected": -0.29944029450416565, | |
| "logps/chosen": -663.63330078125, | |
| "logps/rejected": -864.45703125, | |
| "loss": 0.358, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -2.602560043334961, | |
| "rewards/margins": 2.116223096847534, | |
| "rewards/rejected": -4.718783378601074, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.9021512838306731, | |
| "grad_norm": 4.368284739209083, | |
| "learning_rate": 1.4258820954781037e-08, | |
| "logits/chosen": -0.4327312409877777, | |
| "logits/rejected": -0.2682781219482422, | |
| "logps/chosen": -700.5999145507812, | |
| "logps/rejected": -931.2169799804688, | |
| "loss": 0.3583, | |
| "rewards/accuracies": 0.828125, | |
| "rewards/chosen": -2.735988140106201, | |
| "rewards/margins": 2.3255417346954346, | |
| "rewards/rejected": -5.061530113220215, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.916030534351145, | |
| "grad_norm": 4.046149463634508, | |
| "learning_rate": 1.0502621921127774e-08, | |
| "logits/chosen": -0.43741026520729065, | |
| "logits/rejected": -0.2442634403705597, | |
| "logps/chosen": -671.8573608398438, | |
| "logps/rejected": -919.7029418945312, | |
| "loss": 0.3519, | |
| "rewards/accuracies": 0.846875011920929, | |
| "rewards/chosen": -2.6393163204193115, | |
| "rewards/margins": 2.4307634830474854, | |
| "rewards/rejected": -5.070079326629639, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.9299097848716169, | |
| "grad_norm": 3.883568325725545, | |
| "learning_rate": 7.309237584595007e-09, | |
| "logits/chosen": -0.4095235764980316, | |
| "logits/rejected": -0.27290958166122437, | |
| "logps/chosen": -672.8810424804688, | |
| "logps/rejected": -929.7780151367188, | |
| "loss": 0.3433, | |
| "rewards/accuracies": 0.859375, | |
| "rewards/chosen": -2.6669869422912598, | |
| "rewards/margins": 2.416769504547119, | |
| "rewards/rejected": -5.083756446838379, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.9437890353920888, | |
| "grad_norm": 5.273020682836859, | |
| "learning_rate": 4.6861723431538265e-09, | |
| "logits/chosen": -0.42638665437698364, | |
| "logits/rejected": -0.25011223554611206, | |
| "logps/chosen": -685.7161865234375, | |
| "logps/rejected": -900.3049926757812, | |
| "loss": 0.3666, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -2.629455089569092, | |
| "rewards/margins": 2.211458206176758, | |
| "rewards/rejected": -4.84091329574585, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.9576682859125607, | |
| "grad_norm": 5.214078177947886, | |
| "learning_rate": 2.639590354763882e-09, | |
| "logits/chosen": -0.41801247000694275, | |
| "logits/rejected": -0.29952767491340637, | |
| "logps/chosen": -651.67724609375, | |
| "logps/rejected": -901.2472534179688, | |
| "loss": 0.3506, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -2.55751371383667, | |
| "rewards/margins": 2.3576016426086426, | |
| "rewards/rejected": -4.9151153564453125, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.9715475364330326, | |
| "grad_norm": 4.655188648736662, | |
| "learning_rate": 1.1743010517085427e-09, | |
| "logits/chosen": -0.4516308307647705, | |
| "logits/rejected": -0.2846836745738983, | |
| "logps/chosen": -696.3230590820312, | |
| "logps/rejected": -928.4945068359375, | |
| "loss": 0.3491, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -2.7318358421325684, | |
| "rewards/margins": 2.3993608951568604, | |
| "rewards/rejected": -5.13119649887085, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.9854267869535045, | |
| "grad_norm": 3.572705217532298, | |
| "learning_rate": 2.9374783851240923e-10, | |
| "logits/chosen": -0.39585572481155396, | |
| "logits/rejected": -0.26901108026504517, | |
| "logps/chosen": -655.4457397460938, | |
| "logps/rejected": -909.2194213867188, | |
| "loss": 0.3344, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -2.4239559173583984, | |
| "rewards/margins": 2.4643149375915527, | |
| "rewards/rejected": -4.888271331787109, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.9993060374739764, | |
| "grad_norm": 4.876709666740222, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -0.4101489186286926, | |
| "logits/rejected": -0.2496207058429718, | |
| "logps/chosen": -666.4955444335938, | |
| "logps/rejected": -903.6095581054688, | |
| "loss": 0.3544, | |
| "rewards/accuracies": 0.8218749761581421, | |
| "rewards/chosen": -2.61322283744812, | |
| "rewards/margins": 2.3339571952819824, | |
| "rewards/rejected": -4.947180271148682, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.9993060374739764, | |
| "step": 720, | |
| "total_flos": 0.0, | |
| "train_loss": 0.4493948830498589, | |
| "train_runtime": 20175.981, | |
| "train_samples_per_second": 9.142, | |
| "train_steps_per_second": 0.036 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 720, | |
| "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": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |