| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.9946666666666668, | |
| "eval_steps": 500, | |
| "global_step": 374, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02666666666666667, | |
| "grad_norm": 91.80353546142578, | |
| "learning_rate": 1.3157894736842104e-08, | |
| "logits/chosen": 0.05448655039072037, | |
| "logits/rejected": 0.061774831265211105, | |
| "logps/chosen": -70.88352966308594, | |
| "logps/rejected": -68.78348541259766, | |
| "loss": 0.6945, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -0.007239150814712048, | |
| "rewards/margins": 0.00464630126953125, | |
| "rewards/rejected": -0.011885452084243298, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.05333333333333334, | |
| "grad_norm": 98.0750732421875, | |
| "learning_rate": 2.6315789473684208e-08, | |
| "logits/chosen": -0.09558521211147308, | |
| "logits/rejected": -0.09429871290922165, | |
| "logps/chosen": -47.07535934448242, | |
| "logps/rejected": -57.48785400390625, | |
| "loss": 0.6855, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.04929858446121216, | |
| "rewards/margins": 0.017685014754533768, | |
| "rewards/rejected": 0.03161356970667839, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 86.46255493164062, | |
| "learning_rate": 3.947368421052631e-08, | |
| "logits/chosen": -0.0949799194931984, | |
| "logits/rejected": -0.08726556599140167, | |
| "logps/chosen": -58.88152313232422, | |
| "logps/rejected": -69.49443054199219, | |
| "loss": 0.6504, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.033757805824279785, | |
| "rewards/margins": 0.1223212480545044, | |
| "rewards/rejected": -0.08856344223022461, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.10666666666666667, | |
| "grad_norm": 85.12960052490234, | |
| "learning_rate": 5.2631578947368416e-08, | |
| "logits/chosen": 0.04131064563989639, | |
| "logits/rejected": 0.047180645167827606, | |
| "logps/chosen": -64.10095977783203, | |
| "logps/rejected": -71.49341583251953, | |
| "loss": 0.622, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.08759395778179169, | |
| "rewards/margins": 0.22742386162281036, | |
| "rewards/rejected": -0.13982990384101868, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.13333333333333333, | |
| "grad_norm": 72.46548461914062, | |
| "learning_rate": 6.578947368421053e-08, | |
| "logits/chosen": 0.07103381305932999, | |
| "logits/rejected": 0.06991696357727051, | |
| "logps/chosen": -71.91819763183594, | |
| "logps/rejected": -76.61006927490234, | |
| "loss": 0.6107, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.14151427149772644, | |
| "rewards/margins": 0.2075047492980957, | |
| "rewards/rejected": -0.06599047034978867, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 71.43992614746094, | |
| "learning_rate": 7.894736842105262e-08, | |
| "logits/chosen": -0.0016497105825692415, | |
| "logits/rejected": -0.003721952438354492, | |
| "logps/chosen": -64.13134765625, | |
| "logps/rejected": -59.88336181640625, | |
| "loss": 0.5604, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.14888896048069, | |
| "rewards/margins": 0.5770025253295898, | |
| "rewards/rejected": -0.4281136095523834, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.18666666666666668, | |
| "grad_norm": 86.11872100830078, | |
| "learning_rate": 9.210526315789473e-08, | |
| "logits/chosen": 0.09134344756603241, | |
| "logits/rejected": 0.08688181638717651, | |
| "logps/chosen": -74.20964050292969, | |
| "logps/rejected": -66.22145080566406, | |
| "loss": 0.5322, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.3988303542137146, | |
| "rewards/margins": 0.3464100956916809, | |
| "rewards/rejected": 0.052420247346162796, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.21333333333333335, | |
| "grad_norm": 83.05272674560547, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.051578253507614136, | |
| "logits/rejected": -0.04844808578491211, | |
| "logps/chosen": -91.24781036376953, | |
| "logps/rejected": -73.64341735839844, | |
| "loss": 0.5272, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.11290506273508072, | |
| "rewards/margins": 0.25997716188430786, | |
| "rewards/rejected": -0.14707207679748535, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 74.43838500976562, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.06837300211191177, | |
| "logits/rejected": -0.06631087511777878, | |
| "logps/chosen": -60.6038818359375, | |
| "logps/rejected": -62.58821487426758, | |
| "loss": 0.5138, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.5276230573654175, | |
| "rewards/margins": 0.8340091705322266, | |
| "rewards/rejected": -0.30638617277145386, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.26666666666666666, | |
| "grad_norm": 62.49097442626953, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.00017045289860107005, | |
| "logits/rejected": -0.0001672551006777212, | |
| "logps/chosen": -82.781005859375, | |
| "logps/rejected": -82.09158325195312, | |
| "loss": 0.4674, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.34503522515296936, | |
| "rewards/margins": 1.0920366048812866, | |
| "rewards/rejected": -0.7470014691352844, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.29333333333333333, | |
| "grad_norm": 80.10330200195312, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.02159346453845501, | |
| "logits/rejected": -0.023507962003350258, | |
| "logps/chosen": -66.51358032226562, | |
| "logps/rejected": -65.78565979003906, | |
| "loss": 0.5481, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.3966673016548157, | |
| "rewards/margins": 0.8023613095283508, | |
| "rewards/rejected": -0.4056939482688904, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 67.92163848876953, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.02508384920656681, | |
| "logits/rejected": -0.023196673020720482, | |
| "logps/chosen": -69.94526672363281, | |
| "logps/rejected": -54.190528869628906, | |
| "loss": 0.5126, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.5231302976608276, | |
| "rewards/margins": 0.935653805732727, | |
| "rewards/rejected": -0.41252344846725464, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.3466666666666667, | |
| "grad_norm": 62.79459762573242, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.14604972302913666, | |
| "logits/rejected": -0.1421947032213211, | |
| "logps/chosen": -71.80470275878906, | |
| "logps/rejected": -62.09037399291992, | |
| "loss": 0.5224, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.6033689379692078, | |
| "rewards/margins": 0.9922472238540649, | |
| "rewards/rejected": -0.3888782560825348, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.37333333333333335, | |
| "grad_norm": 63.457252502441406, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.06079145520925522, | |
| "logits/rejected": 0.07840003073215485, | |
| "logps/chosen": -71.81758117675781, | |
| "logps/rejected": -85.1531982421875, | |
| "loss": 0.39, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.24788689613342285, | |
| "rewards/margins": 1.4833877086639404, | |
| "rewards/rejected": -1.2355008125305176, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 69.04164123535156, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.030724067240953445, | |
| "logits/rejected": 0.037750810384750366, | |
| "logps/chosen": -63.71375274658203, | |
| "logps/rejected": -74.11785888671875, | |
| "loss": 0.501, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.5482760667800903, | |
| "rewards/margins": 1.5258257389068604, | |
| "rewards/rejected": -0.9775495529174805, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.4266666666666667, | |
| "grad_norm": 68.25962829589844, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.0079464977607131, | |
| "logits/rejected": 0.0006744593265466392, | |
| "logps/chosen": -44.04719924926758, | |
| "logps/rejected": -63.342002868652344, | |
| "loss": 0.441, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.5909361839294434, | |
| "rewards/margins": 0.9477832913398743, | |
| "rewards/rejected": -0.3568470776081085, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.4533333333333333, | |
| "grad_norm": 61.337249755859375, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.13302002847194672, | |
| "logits/rejected": -0.12650710344314575, | |
| "logps/chosen": -76.5062255859375, | |
| "logps/rejected": -77.22346496582031, | |
| "loss": 0.3969, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.7510318160057068, | |
| "rewards/margins": 1.1422964334487915, | |
| "rewards/rejected": -0.3912646174430847, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 60.21578598022461, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.08266867697238922, | |
| "logits/rejected": 0.08583595603704453, | |
| "logps/chosen": -85.27580261230469, | |
| "logps/rejected": -83.88575744628906, | |
| "loss": 0.4887, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.2144974172115326, | |
| "rewards/margins": 1.155656337738037, | |
| "rewards/rejected": -0.9411588907241821, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.5066666666666667, | |
| "grad_norm": 78.39936828613281, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.018161656334996223, | |
| "logits/rejected": -0.012231660075485706, | |
| "logps/chosen": -72.49486541748047, | |
| "logps/rejected": -81.66197967529297, | |
| "loss": 0.4794, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.487576961517334, | |
| "rewards/margins": 1.1082786321640015, | |
| "rewards/rejected": -0.6207017302513123, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.5333333333333333, | |
| "grad_norm": 72.66243743896484, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.11775938421487808, | |
| "logits/rejected": -0.10979805141687393, | |
| "logps/chosen": -37.96864318847656, | |
| "logps/rejected": -41.957908630371094, | |
| "loss": 0.4535, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.5172302722930908, | |
| "rewards/margins": 0.8645200729370117, | |
| "rewards/rejected": -0.3472897410392761, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 84.33699798583984, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.07096003741025925, | |
| "logits/rejected": -0.06229569762945175, | |
| "logps/chosen": -79.24551391601562, | |
| "logps/rejected": -73.00336456298828, | |
| "loss": 0.4727, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.48409804701805115, | |
| "rewards/margins": 1.5594733953475952, | |
| "rewards/rejected": -1.0753753185272217, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.5866666666666667, | |
| "grad_norm": 75.8474349975586, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.002287261188030243, | |
| "logits/rejected": 0.011448122560977936, | |
| "logps/chosen": -36.34912872314453, | |
| "logps/rejected": -46.311866760253906, | |
| "loss": 0.5285, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.3635416328907013, | |
| "rewards/margins": 0.3584732413291931, | |
| "rewards/rejected": 0.005068361759185791, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.6133333333333333, | |
| "grad_norm": 75.15863800048828, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.07114384323358536, | |
| "logits/rejected": -0.06986474245786667, | |
| "logps/chosen": -53.876953125, | |
| "logps/rejected": -52.810569763183594, | |
| "loss": 0.4579, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.7901820540428162, | |
| "rewards/margins": 1.1014235019683838, | |
| "rewards/rejected": -0.3112414479255676, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 64.49634552001953, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.023414045572280884, | |
| "logits/rejected": -0.02103838510811329, | |
| "logps/chosen": -58.588645935058594, | |
| "logps/rejected": -54.12468719482422, | |
| "loss": 0.4412, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.6539210081100464, | |
| "rewards/margins": 1.1925280094146729, | |
| "rewards/rejected": -0.5386068224906921, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.6666666666666666, | |
| "grad_norm": 77.18801879882812, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.07651327550411224, | |
| "logits/rejected": 0.09625453501939774, | |
| "logps/chosen": -86.07091522216797, | |
| "logps/rejected": -89.9097671508789, | |
| "loss": 0.449, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.31949272751808167, | |
| "rewards/margins": 1.6389261484146118, | |
| "rewards/rejected": -1.3194334506988525, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.6933333333333334, | |
| "grad_norm": 56.2886848449707, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.029146382585167885, | |
| "logits/rejected": 0.03658589348196983, | |
| "logps/chosen": -73.04118347167969, | |
| "logps/rejected": -76.2184829711914, | |
| "loss": 0.4175, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.45680832862854004, | |
| "rewards/margins": 1.049865484237671, | |
| "rewards/rejected": -0.5930570960044861, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 77.33464813232422, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.075675830245018, | |
| "logits/rejected": -0.0697128027677536, | |
| "logps/chosen": -76.72476959228516, | |
| "logps/rejected": -83.47200012207031, | |
| "loss": 0.4749, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.5277665853500366, | |
| "rewards/margins": 1.3413928747177124, | |
| "rewards/rejected": -0.8136262893676758, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.7466666666666667, | |
| "grad_norm": 86.7103271484375, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.06388665735721588, | |
| "logits/rejected": -0.055876873433589935, | |
| "logps/chosen": -48.2696647644043, | |
| "logps/rejected": -43.421836853027344, | |
| "loss": 0.4583, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.27933669090270996, | |
| "rewards/margins": 0.9728415608406067, | |
| "rewards/rejected": -0.6935049295425415, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.7733333333333333, | |
| "grad_norm": 77.15120697021484, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.04482016712427139, | |
| "logits/rejected": -0.0454244539141655, | |
| "logps/chosen": -58.773719787597656, | |
| "logps/rejected": -54.565284729003906, | |
| "loss": 0.4598, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.33584028482437134, | |
| "rewards/margins": 0.4276786744594574, | |
| "rewards/rejected": -0.09183841943740845, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 62.76329803466797, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.046587713062763214, | |
| "logits/rejected": -0.04582952335476875, | |
| "logps/chosen": -77.17513275146484, | |
| "logps/rejected": -81.3776626586914, | |
| "loss": 0.414, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.00032804012880660594, | |
| "rewards/margins": 0.8608261346817017, | |
| "rewards/rejected": -0.8611541986465454, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.8266666666666667, | |
| "grad_norm": 48.6246337890625, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.0023877639323472977, | |
| "logits/rejected": 0.000560196116566658, | |
| "logps/chosen": -61.36391067504883, | |
| "logps/rejected": -74.22142028808594, | |
| "loss": 0.3678, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.5382565855979919, | |
| "rewards/margins": 1.6945278644561768, | |
| "rewards/rejected": -1.15627121925354, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.8533333333333334, | |
| "grad_norm": 66.40213775634766, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.07235555350780487, | |
| "logits/rejected": -0.06712132692337036, | |
| "logps/chosen": -47.608489990234375, | |
| "logps/rejected": -58.56159591674805, | |
| "loss": 0.4747, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.3097238540649414, | |
| "rewards/margins": 0.753086268901825, | |
| "rewards/rejected": -0.44336241483688354, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 56.45853805541992, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.007908957079052925, | |
| "logits/rejected": 0.008793281391263008, | |
| "logps/chosen": -56.72149658203125, | |
| "logps/rejected": -59.84600067138672, | |
| "loss": 0.4839, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.4165565073490143, | |
| "rewards/margins": 0.6458317637443542, | |
| "rewards/rejected": -0.22927527129650116, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.9066666666666666, | |
| "grad_norm": 56.173946380615234, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.030985862016677856, | |
| "logits/rejected": 0.03558924049139023, | |
| "logps/chosen": -62.272987365722656, | |
| "logps/rejected": -69.59276580810547, | |
| "loss": 0.4139, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.3212319016456604, | |
| "rewards/margins": 1.2905486822128296, | |
| "rewards/rejected": -0.9693166613578796, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.9333333333333333, | |
| "grad_norm": 78.22785949707031, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.0006909176590852439, | |
| "logits/rejected": 0.002517472254112363, | |
| "logps/chosen": -59.8375129699707, | |
| "logps/rejected": -58.26896286010742, | |
| "loss": 0.3862, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.49182993173599243, | |
| "rewards/margins": 1.2024425268173218, | |
| "rewards/rejected": -0.7106126546859741, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 50.518157958984375, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.03367740660905838, | |
| "logits/rejected": 0.04943201318383217, | |
| "logps/chosen": -73.77796173095703, | |
| "logps/rejected": -83.64482879638672, | |
| "loss": 0.4253, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.5700420141220093, | |
| "rewards/margins": 2.157716751098633, | |
| "rewards/rejected": -1.587674856185913, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.9866666666666667, | |
| "grad_norm": 57.97651290893555, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.07821191847324371, | |
| "logits/rejected": -0.07423903793096542, | |
| "logps/chosen": -63.325286865234375, | |
| "logps/rejected": -49.606910705566406, | |
| "loss": 0.4184, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.7217522859573364, | |
| "rewards/margins": 1.4272280931472778, | |
| "rewards/rejected": -0.7054757475852966, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 1.0133333333333334, | |
| "grad_norm": 42.711490631103516, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.011517315171658993, | |
| "logits/rejected": -0.016577688977122307, | |
| "logps/chosen": -63.11487579345703, | |
| "logps/rejected": -71.26173400878906, | |
| "loss": 0.3845, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.48871931433677673, | |
| "rewards/margins": 1.1433701515197754, | |
| "rewards/rejected": -0.654650866985321, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 36.31159973144531, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.006033201701939106, | |
| "logits/rejected": 0.016330499202013016, | |
| "logps/chosen": -57.083030700683594, | |
| "logps/rejected": -71.14306640625, | |
| "loss": 0.2752, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.7944145202636719, | |
| "rewards/margins": 1.1578031778335571, | |
| "rewards/rejected": -0.3633885979652405, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 1.0666666666666667, | |
| "grad_norm": 32.363685607910156, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.09754471480846405, | |
| "logits/rejected": -0.09675069153308868, | |
| "logps/chosen": -50.0403938293457, | |
| "logps/rejected": -50.053855895996094, | |
| "loss": 0.2892, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.6043117642402649, | |
| "rewards/margins": 1.4791052341461182, | |
| "rewards/rejected": -0.8747934103012085, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.0933333333333333, | |
| "grad_norm": 49.40591812133789, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.03340379148721695, | |
| "logits/rejected": 0.03613832965493202, | |
| "logps/chosen": -75.98091888427734, | |
| "logps/rejected": -89.6574935913086, | |
| "loss": 0.2736, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.5458163022994995, | |
| "rewards/margins": 1.8934462070465088, | |
| "rewards/rejected": -1.3476299047470093, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "grad_norm": 33.99880599975586, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.014881300739943981, | |
| "logits/rejected": -0.008952580392360687, | |
| "logps/chosen": -61.28364944458008, | |
| "logps/rejected": -67.7909164428711, | |
| "loss": 0.2599, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.6829001903533936, | |
| "rewards/margins": 2.180811643600464, | |
| "rewards/rejected": -1.4979116916656494, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.1466666666666667, | |
| "grad_norm": 32.734886169433594, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.1162935271859169, | |
| "logits/rejected": -0.1172286868095398, | |
| "logps/chosen": -72.66918182373047, | |
| "logps/rejected": -81.00222778320312, | |
| "loss": 0.2414, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.747806191444397, | |
| "rewards/margins": 2.400660276412964, | |
| "rewards/rejected": -1.652854323387146, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 1.1733333333333333, | |
| "grad_norm": 45.963077545166016, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.03322611004114151, | |
| "logits/rejected": 0.03860603645443916, | |
| "logps/chosen": -67.83540344238281, | |
| "logps/rejected": -77.8966293334961, | |
| "loss": 0.262, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.487581342458725, | |
| "rewards/margins": 2.018979072570801, | |
| "rewards/rejected": -1.531397819519043, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 31.000837326049805, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.06304231286048889, | |
| "logits/rejected": -0.06262607872486115, | |
| "logps/chosen": -72.00182342529297, | |
| "logps/rejected": -69.18010711669922, | |
| "loss": 0.3098, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.16077418625354767, | |
| "rewards/margins": 1.9396165609359741, | |
| "rewards/rejected": -2.100390911102295, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.2266666666666666, | |
| "grad_norm": 41.64716339111328, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.06013431400060654, | |
| "logits/rejected": -0.05363202840089798, | |
| "logps/chosen": -71.9935531616211, | |
| "logps/rejected": -77.51136779785156, | |
| "loss": 0.276, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.6929202675819397, | |
| "rewards/margins": 2.5861668586730957, | |
| "rewards/rejected": -1.8932468891143799, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.2533333333333334, | |
| "grad_norm": 39.7431526184082, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.0543329119682312, | |
| "logits/rejected": -0.04811491817235947, | |
| "logps/chosen": -44.972129821777344, | |
| "logps/rejected": -48.955535888671875, | |
| "loss": 0.3028, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.212233066558838, | |
| "rewards/margins": 2.2813668251037598, | |
| "rewards/rejected": -1.069133996963501, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "grad_norm": 36.94087600708008, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.03034238889813423, | |
| "logits/rejected": -0.02829553559422493, | |
| "logps/chosen": -45.940433502197266, | |
| "logps/rejected": -47.90568161010742, | |
| "loss": 0.2795, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.9621788263320923, | |
| "rewards/margins": 1.6789710521697998, | |
| "rewards/rejected": -0.7167922258377075, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.3066666666666666, | |
| "grad_norm": 34.41165542602539, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.08339103311300278, | |
| "logits/rejected": 0.10816546529531479, | |
| "logps/chosen": -73.56194305419922, | |
| "logps/rejected": -73.25684356689453, | |
| "loss": 0.2708, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.37739673256874084, | |
| "rewards/margins": 2.098477840423584, | |
| "rewards/rejected": -1.7210811376571655, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 1.3333333333333333, | |
| "grad_norm": 31.31505584716797, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.04144307225942612, | |
| "logits/rejected": 0.04325007274746895, | |
| "logps/chosen": -71.2078628540039, | |
| "logps/rejected": -80.70985412597656, | |
| "loss": 0.274, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.5666126012802124, | |
| "rewards/margins": 1.788511872291565, | |
| "rewards/rejected": -1.2218992710113525, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.3599999999999999, | |
| "grad_norm": 35.27589416503906, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.09748046845197678, | |
| "logits/rejected": 0.08802054077386856, | |
| "logps/chosen": -70.91510772705078, | |
| "logps/rejected": -70.45256805419922, | |
| "loss": 0.2802, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.5191822648048401, | |
| "rewards/margins": 2.1395535469055176, | |
| "rewards/rejected": -1.6203714609146118, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 1.3866666666666667, | |
| "grad_norm": 36.5883674621582, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.028733888640999794, | |
| "logits/rejected": -0.022805647924542427, | |
| "logps/chosen": -68.54586029052734, | |
| "logps/rejected": -66.59730529785156, | |
| "loss": 0.3011, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.6113399267196655, | |
| "rewards/margins": 2.464463472366333, | |
| "rewards/rejected": -1.853123664855957, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.4133333333333333, | |
| "grad_norm": 21.833608627319336, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.007651590742170811, | |
| "logits/rejected": 0.011075135320425034, | |
| "logps/chosen": -70.35153198242188, | |
| "logps/rejected": -72.46214294433594, | |
| "loss": 0.2587, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.6863104701042175, | |
| "rewards/margins": 1.9516360759735107, | |
| "rewards/rejected": -1.2653255462646484, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "grad_norm": 42.52511215209961, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.019661933183670044, | |
| "logits/rejected": -0.020269040018320084, | |
| "logps/chosen": -81.03693389892578, | |
| "logps/rejected": -75.59123229980469, | |
| "loss": 0.263, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.1058200597763062, | |
| "rewards/margins": 2.4359288215637207, | |
| "rewards/rejected": -1.3301087617874146, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 1.4666666666666668, | |
| "grad_norm": 54.02488327026367, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.045877061784267426, | |
| "logits/rejected": -0.05564676970243454, | |
| "logps/chosen": -80.84842681884766, | |
| "logps/rejected": -79.160888671875, | |
| "loss": 0.3026, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.76087886095047, | |
| "rewards/margins": 1.8281453847885132, | |
| "rewards/rejected": -1.067266583442688, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 1.4933333333333334, | |
| "grad_norm": 50.83435821533203, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.09148342907428741, | |
| "logits/rejected": -0.08500328660011292, | |
| "logps/chosen": -41.23069763183594, | |
| "logps/rejected": -40.55739212036133, | |
| "loss": 0.3747, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.123250126838684, | |
| "rewards/margins": 1.7874069213867188, | |
| "rewards/rejected": -0.6641567945480347, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "grad_norm": 49.45879364013672, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.001988143427297473, | |
| "logits/rejected": -0.006540440954267979, | |
| "logps/chosen": -50.0777587890625, | |
| "logps/rejected": -53.56150436401367, | |
| "loss": 0.3404, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.7993362545967102, | |
| "rewards/margins": 1.6362117528915405, | |
| "rewards/rejected": -0.8368755578994751, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 1.5466666666666666, | |
| "grad_norm": 47.30949020385742, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.02390037290751934, | |
| "logits/rejected": 0.01714068278670311, | |
| "logps/chosen": -59.122962951660156, | |
| "logps/rejected": -60.09889602661133, | |
| "loss": 0.2568, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.1686420440673828, | |
| "rewards/margins": 2.2462868690490723, | |
| "rewards/rejected": -1.0776447057724, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 1.5733333333333333, | |
| "grad_norm": 34.02714538574219, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.0011189490323886275, | |
| "logits/rejected": 0.0036900490522384644, | |
| "logps/chosen": -56.07664108276367, | |
| "logps/rejected": -59.93505859375, | |
| "loss": 0.2744, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.0613322257995605, | |
| "rewards/margins": 1.9388986825942993, | |
| "rewards/rejected": -0.8775663375854492, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 52.804840087890625, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.017051326110959053, | |
| "logits/rejected": -0.011620084755122662, | |
| "logps/chosen": -53.503883361816406, | |
| "logps/rejected": -56.5186882019043, | |
| "loss": 0.3073, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.9534000158309937, | |
| "rewards/margins": 1.7106940746307373, | |
| "rewards/rejected": -0.7572939395904541, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.6266666666666667, | |
| "grad_norm": 38.153648376464844, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.00943007878959179, | |
| "logits/rejected": 0.028043877333402634, | |
| "logps/chosen": -66.48413848876953, | |
| "logps/rejected": -105.61421966552734, | |
| "loss": 0.2633, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.8428307771682739, | |
| "rewards/margins": 2.7967689037323, | |
| "rewards/rejected": -1.9539384841918945, | |
| "step": 305 | |
| }, | |
| { | |
| "epoch": 1.6533333333333333, | |
| "grad_norm": 61.87904739379883, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.03582911938428879, | |
| "logits/rejected": -0.0231150072067976, | |
| "logps/chosen": -59.17612838745117, | |
| "logps/rejected": -71.96540832519531, | |
| "loss": 0.2505, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.9612646102905273, | |
| "rewards/margins": 2.2090392112731934, | |
| "rewards/rejected": -1.247774600982666, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.6800000000000002, | |
| "grad_norm": 55.832542419433594, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.044271357357501984, | |
| "logits/rejected": -0.021111857146024704, | |
| "logps/chosen": -69.70079040527344, | |
| "logps/rejected": -89.76176452636719, | |
| "loss": 0.2382, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.37344738841056824, | |
| "rewards/margins": 2.854713201522827, | |
| "rewards/rejected": -2.4812657833099365, | |
| "step": 315 | |
| }, | |
| { | |
| "epoch": 1.7066666666666666, | |
| "grad_norm": 47.42922592163086, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.07435096800327301, | |
| "logits/rejected": -0.074483223259449, | |
| "logps/chosen": -77.44403076171875, | |
| "logps/rejected": -71.1507797241211, | |
| "loss": 0.3022, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.5832604169845581, | |
| "rewards/margins": 2.0570945739746094, | |
| "rewards/rejected": -1.4738342761993408, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.7333333333333334, | |
| "grad_norm": 44.8173713684082, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.001697111176326871, | |
| "logits/rejected": 0.004212519619613886, | |
| "logps/chosen": -73.09126281738281, | |
| "logps/rejected": -77.27495574951172, | |
| "loss": 0.2864, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.5175535678863525, | |
| "rewards/margins": 1.9960486888885498, | |
| "rewards/rejected": -1.4784951210021973, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 52.862125396728516, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.05788411572575569, | |
| "logits/rejected": -0.054354745894670486, | |
| "logps/chosen": -77.21604919433594, | |
| "logps/rejected": -90.43736267089844, | |
| "loss": 0.2593, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.5059728026390076, | |
| "rewards/margins": 2.594764232635498, | |
| "rewards/rejected": -2.088791847229004, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.7866666666666666, | |
| "grad_norm": 37.6073112487793, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.029447251930832863, | |
| "logits/rejected": -0.02990039810538292, | |
| "logps/chosen": -61.29267501831055, | |
| "logps/rejected": -83.2472915649414, | |
| "loss": 0.1939, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.3546075224876404, | |
| "rewards/margins": 2.5164620876312256, | |
| "rewards/rejected": -2.1618547439575195, | |
| "step": 335 | |
| }, | |
| { | |
| "epoch": 1.8133333333333335, | |
| "grad_norm": 49.868560791015625, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": 0.016156326979398727, | |
| "logits/rejected": 0.03281542658805847, | |
| "logps/chosen": -81.39087677001953, | |
| "logps/rejected": -81.13011169433594, | |
| "loss": 0.2456, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.4369918704032898, | |
| "rewards/margins": 2.3583126068115234, | |
| "rewards/rejected": -1.9213206768035889, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.8399999999999999, | |
| "grad_norm": 41.97678756713867, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.021811150014400482, | |
| "logits/rejected": -0.02089584432542324, | |
| "logps/chosen": -69.17388916015625, | |
| "logps/rejected": -58.414764404296875, | |
| "loss": 0.3057, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.2403881549835205, | |
| "rewards/margins": 2.0268137454986572, | |
| "rewards/rejected": -0.7864255905151367, | |
| "step": 345 | |
| }, | |
| { | |
| "epoch": 1.8666666666666667, | |
| "grad_norm": 46.311893463134766, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.02446429245173931, | |
| "logits/rejected": -0.014963751658797264, | |
| "logps/chosen": -36.05630874633789, | |
| "logps/rejected": -43.71437454223633, | |
| "loss": 0.2696, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.1039693355560303, | |
| "rewards/margins": 2.3277952671051025, | |
| "rewards/rejected": -1.2238258123397827, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.8933333333333333, | |
| "grad_norm": 44.26567459106445, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.04746484011411667, | |
| "logits/rejected": -0.04588810354471207, | |
| "logps/chosen": -67.33465576171875, | |
| "logps/rejected": -73.4018325805664, | |
| "loss": 0.2557, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.6922093629837036, | |
| "rewards/margins": 1.8044688701629639, | |
| "rewards/rejected": -1.1122596263885498, | |
| "step": 355 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 29.275634765625, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.00010519176430534571, | |
| "logits/rejected": 0.01266922615468502, | |
| "logps/chosen": -63.38561248779297, | |
| "logps/rejected": -76.10896301269531, | |
| "loss": 0.2347, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.5989769697189331, | |
| "rewards/margins": 2.0385549068450928, | |
| "rewards/rejected": -1.4395779371261597, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.9466666666666668, | |
| "grad_norm": 38.70104217529297, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.07840217649936676, | |
| "logits/rejected": -0.06996472924947739, | |
| "logps/chosen": -77.53556060791016, | |
| "logps/rejected": -79.41651916503906, | |
| "loss": 0.23, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.3816269040107727, | |
| "rewards/margins": 2.385305166244507, | |
| "rewards/rejected": -2.003678321838379, | |
| "step": 365 | |
| }, | |
| { | |
| "epoch": 1.9733333333333334, | |
| "grad_norm": 45.61483383178711, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.030093077570199966, | |
| "logits/rejected": -0.03412201628088951, | |
| "logps/chosen": -65.24625396728516, | |
| "logps/rejected": -68.85826873779297, | |
| "loss": 0.3033, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.0214712619781494, | |
| "rewards/margins": 1.9241225719451904, | |
| "rewards/rejected": -0.902651309967041, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.9946666666666668, | |
| "step": 374, | |
| "total_flos": 0.0, | |
| "train_loss": 0.3853599507222201, | |
| "train_runtime": 1020.3175, | |
| "train_samples_per_second": 11.759, | |
| "train_steps_per_second": 0.367 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 374, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "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": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |