| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.976501305483029, | |
| "eval_steps": 500, | |
| "global_step": 285, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 6.532262417706838, | |
| "learning_rate": 1.7241379310344825e-08, | |
| "logits/chosen": -2.851747512817383, | |
| "logits/rejected": -2.833996534347534, | |
| "logps/chosen": -165.70089721679688, | |
| "logps/rejected": -198.857666015625, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 5.937988600161895, | |
| "learning_rate": 8.620689655172414e-08, | |
| "logits/chosen": -2.770829200744629, | |
| "logits/rejected": -2.7736237049102783, | |
| "logps/chosen": -171.36782836914062, | |
| "logps/rejected": -172.6230926513672, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.3359375, | |
| "rewards/chosen": -4.8455476644448936e-05, | |
| "rewards/margins": 4.198795068077743e-05, | |
| "rewards/rejected": -9.044347098097205e-05, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 7.225518266540644, | |
| "learning_rate": 1.7241379310344828e-07, | |
| "logits/chosen": -2.7824466228485107, | |
| "logits/rejected": -2.7912747859954834, | |
| "logps/chosen": -190.32244873046875, | |
| "logps/rejected": -194.74627685546875, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": 0.0004952027229592204, | |
| "rewards/margins": -6.670488801319152e-05, | |
| "rewards/rejected": 0.0005619077128358185, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 7.193514853714808, | |
| "learning_rate": 2.586206896551724e-07, | |
| "logits/chosen": -2.8433165550231934, | |
| "logits/rejected": -2.8537657260894775, | |
| "logps/chosen": -199.39505004882812, | |
| "logps/rejected": -188.52256774902344, | |
| "loss": 0.6924, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": 0.0056815436109900475, | |
| "rewards/margins": 0.0018546178471297026, | |
| "rewards/rejected": 0.0038269255310297012, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 6.944403732218155, | |
| "learning_rate": 3.4482758620689656e-07, | |
| "logits/chosen": -2.8533196449279785, | |
| "logits/rejected": -2.8635940551757812, | |
| "logps/chosen": -167.04734802246094, | |
| "logps/rejected": -178.37033081054688, | |
| "loss": 0.6908, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.016954954713582993, | |
| "rewards/margins": 0.003948197700083256, | |
| "rewards/rejected": 0.01300675980746746, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 6.157406869612788, | |
| "learning_rate": 4.310344827586206e-07, | |
| "logits/chosen": -2.823472738265991, | |
| "logits/rejected": -2.8209452629089355, | |
| "logps/chosen": -127.2848129272461, | |
| "logps/rejected": -146.4281463623047, | |
| "loss": 0.6877, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": 0.0418720580637455, | |
| "rewards/margins": 0.011652471497654915, | |
| "rewards/rejected": 0.030219584703445435, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 6.734511284936862, | |
| "learning_rate": 4.999811754597861e-07, | |
| "logits/chosen": -2.804356098175049, | |
| "logits/rejected": -2.8042080402374268, | |
| "logps/chosen": -154.28651428222656, | |
| "logps/rejected": -162.4736328125, | |
| "loss": 0.6832, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.05192985013127327, | |
| "rewards/margins": 0.023836981505155563, | |
| "rewards/rejected": 0.028092866763472557, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 6.678745069229156, | |
| "learning_rate": 4.993226141696726e-07, | |
| "logits/chosen": -2.8617546558380127, | |
| "logits/rejected": -2.856639862060547, | |
| "logps/chosen": -188.62733459472656, | |
| "logps/rejected": -201.29901123046875, | |
| "loss": 0.6737, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.026439663022756577, | |
| "rewards/margins": 0.03700970113277435, | |
| "rewards/rejected": -0.010570039972662926, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 7.370864651344303, | |
| "learning_rate": 4.97725658856945e-07, | |
| "logits/chosen": -2.771888256072998, | |
| "logits/rejected": -2.783984422683716, | |
| "logps/chosen": -136.36361694335938, | |
| "logps/rejected": -159.01470947265625, | |
| "loss": 0.6625, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.035046085715293884, | |
| "rewards/margins": 0.0771055668592453, | |
| "rewards/rejected": -0.11215164512395859, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 10.544998371700682, | |
| "learning_rate": 4.951963201008075e-07, | |
| "logits/chosen": -2.752272844314575, | |
| "logits/rejected": -2.767472505569458, | |
| "logps/chosen": -191.96400451660156, | |
| "logps/rejected": -194.68214416503906, | |
| "loss": 0.6596, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.03316263481974602, | |
| "rewards/margins": 0.0969243049621582, | |
| "rewards/rejected": -0.13008692860603333, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 12.709959506461484, | |
| "learning_rate": 4.91744117761213e-07, | |
| "logits/chosen": -2.7166810035705566, | |
| "logits/rejected": -2.7183845043182373, | |
| "logps/chosen": -185.669677734375, | |
| "logps/rejected": -214.07803344726562, | |
| "loss": 0.6426, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.11343240737915039, | |
| "rewards/margins": 0.12829084694385529, | |
| "rewards/rejected": -0.24172325432300568, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 11.114453412752225, | |
| "learning_rate": 4.873820451482591e-07, | |
| "logits/chosen": -2.5826053619384766, | |
| "logits/rejected": -2.590798854827881, | |
| "logps/chosen": -153.5226287841797, | |
| "logps/rejected": -174.6750030517578, | |
| "loss": 0.6376, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.05409809201955795, | |
| "rewards/margins": 0.1648661345243454, | |
| "rewards/rejected": -0.21896424889564514, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 9.723140130030433, | |
| "learning_rate": 4.821265201183039e-07, | |
| "logits/chosen": -2.6595444679260254, | |
| "logits/rejected": -2.656533718109131, | |
| "logps/chosen": -176.4412384033203, | |
| "logps/rejected": -222.3843994140625, | |
| "loss": 0.6169, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.044831644743680954, | |
| "rewards/margins": 0.22063831984996796, | |
| "rewards/rejected": -0.2654699683189392, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 13.924344202071776, | |
| "learning_rate": 4.7599732328086083e-07, | |
| "logits/chosen": -2.67556095123291, | |
| "logits/rejected": -2.689857006072998, | |
| "logps/chosen": -177.05345153808594, | |
| "logps/rejected": -210.99075317382812, | |
| "loss": 0.6143, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.14976295828819275, | |
| "rewards/margins": 0.26197725534439087, | |
| "rewards/rejected": -0.4117402136325836, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 12.80473542751268, | |
| "learning_rate": 4.690175235488516e-07, | |
| "logits/chosen": -2.61578631401062, | |
| "logits/rejected": -2.6244595050811768, | |
| "logps/chosen": -215.93722534179688, | |
| "logps/rejected": -233.77212524414062, | |
| "loss": 0.6003, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.3880045711994171, | |
| "rewards/margins": 0.21204185485839844, | |
| "rewards/rejected": -0.6000463962554932, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 15.27229317717408, | |
| "learning_rate": 4.612133913124268e-07, | |
| "logits/chosen": -2.563448905944824, | |
| "logits/rejected": -2.5605521202087402, | |
| "logps/chosen": -205.3765106201172, | |
| "logps/rejected": -261.5989685058594, | |
| "loss": 0.5784, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.33516931533813477, | |
| "rewards/margins": 0.44470730423927307, | |
| "rewards/rejected": -0.7798766493797302, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 21.212678962714094, | |
| "learning_rate": 4.526142995631487e-07, | |
| "logits/chosen": -2.5200624465942383, | |
| "logits/rejected": -2.5208845138549805, | |
| "logps/chosen": -216.1806640625, | |
| "logps/rejected": -266.473876953125, | |
| "loss": 0.5752, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.4819144606590271, | |
| "rewards/margins": 0.42951878905296326, | |
| "rewards/rejected": -0.911433219909668, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 22.351019404264836, | |
| "learning_rate": 4.432526133406842e-07, | |
| "logits/chosen": -2.5419774055480957, | |
| "logits/rejected": -2.5438098907470703, | |
| "logps/chosen": -214.19711303710938, | |
| "logps/rejected": -279.48095703125, | |
| "loss": 0.5802, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.5401347875595093, | |
| "rewards/margins": 0.5657010078430176, | |
| "rewards/rejected": -1.1058357954025269, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 23.016308190522402, | |
| "learning_rate": 4.3316356791810316e-07, | |
| "logits/chosen": -2.451939821243286, | |
| "logits/rejected": -2.46024751663208, | |
| "logps/chosen": -215.3175506591797, | |
| "logps/rejected": -288.47430419921875, | |
| "loss": 0.5606, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.4637807309627533, | |
| "rewards/margins": 0.6651399731636047, | |
| "rewards/rejected": -1.1289207935333252, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 29.834339022988733, | |
| "learning_rate": 4.223851361842667e-07, | |
| "logits/chosen": -2.4490389823913574, | |
| "logits/rejected": -2.4603734016418457, | |
| "logps/chosen": -264.7539978027344, | |
| "logps/rejected": -329.4302673339844, | |
| "loss": 0.5496, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.7914533019065857, | |
| "rewards/margins": 0.6750684976577759, | |
| "rewards/rejected": -1.4665216207504272, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 23.231563862948793, | |
| "learning_rate": 4.109578857224478e-07, | |
| "logits/chosen": -2.4296836853027344, | |
| "logits/rejected": -2.422900438308716, | |
| "logps/chosen": -266.4080810546875, | |
| "logps/rejected": -364.8707580566406, | |
| "loss": 0.4442, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.880645751953125, | |
| "rewards/margins": 0.9347571134567261, | |
| "rewards/rejected": -1.8154027462005615, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "grad_norm": 17.716052276536377, | |
| "learning_rate": 3.9892482612310836e-07, | |
| "logits/chosen": -2.305148124694824, | |
| "logits/rejected": -2.2915306091308594, | |
| "logps/chosen": -194.28146362304688, | |
| "logps/rejected": -336.57952880859375, | |
| "loss": 0.4129, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.6234027147293091, | |
| "rewards/margins": 1.2217520475387573, | |
| "rewards/rejected": -1.8451547622680664, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 28.733080634368395, | |
| "learning_rate": 3.863312471055116e-07, | |
| "logits/chosen": -2.1953470706939697, | |
| "logits/rejected": -2.1922059059143066, | |
| "logps/chosen": -237.89230346679688, | |
| "logps/rejected": -355.83233642578125, | |
| "loss": 0.4093, | |
| "rewards/accuracies": 0.8687499761581421, | |
| "rewards/chosen": -0.6553321480751038, | |
| "rewards/margins": 1.1414698362350464, | |
| "rewards/rejected": -1.7968021631240845, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 30.4006533250113, | |
| "learning_rate": 3.73224548057446e-07, | |
| "logits/chosen": -2.0299630165100098, | |
| "logits/rejected": -2.017117738723755, | |
| "logps/chosen": -279.6531982421875, | |
| "logps/rejected": -409.65032958984375, | |
| "loss": 0.4244, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -1.3399035930633545, | |
| "rewards/margins": 1.1194835901260376, | |
| "rewards/rejected": -2.4593873023986816, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 31.108774159483165, | |
| "learning_rate": 3.596540596346319e-07, | |
| "logits/chosen": -1.9845222234725952, | |
| "logits/rejected": -1.984548568725586, | |
| "logps/chosen": -280.70513916015625, | |
| "logps/rejected": -407.49725341796875, | |
| "loss": 0.4322, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.0363457202911377, | |
| "rewards/margins": 1.292150616645813, | |
| "rewards/rejected": -2.3284964561462402, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "grad_norm": 28.53321641298578, | |
| "learning_rate": 3.4567085809127245e-07, | |
| "logits/chosen": -1.9905182123184204, | |
| "logits/rejected": -1.999333143234253, | |
| "logps/chosen": -272.6542663574219, | |
| "logps/rejected": -398.8678283691406, | |
| "loss": 0.4106, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -0.8893561363220215, | |
| "rewards/margins": 1.1745526790618896, | |
| "rewards/rejected": -2.063908815383911, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "grad_norm": 22.972434902434117, | |
| "learning_rate": 3.3132757304056576e-07, | |
| "logits/chosen": -1.915196180343628, | |
| "logits/rejected": -1.919231653213501, | |
| "logps/chosen": -279.16876220703125, | |
| "logps/rejected": -411.39105224609375, | |
| "loss": 0.4151, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -1.1732670068740845, | |
| "rewards/margins": 1.2200413942337036, | |
| "rewards/rejected": -2.393308639526367, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "grad_norm": 22.610383808005285, | |
| "learning_rate": 3.166781893687246e-07, | |
| "logits/chosen": -1.8586881160736084, | |
| "logits/rejected": -1.859628438949585, | |
| "logps/chosen": -345.5313415527344, | |
| "logps/rejected": -487.19793701171875, | |
| "loss": 0.4093, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.7219533920288086, | |
| "rewards/margins": 1.2162812948226929, | |
| "rewards/rejected": -2.938234329223633, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "grad_norm": 24.122351737373137, | |
| "learning_rate": 3.017778440480546e-07, | |
| "logits/chosen": -1.7600257396697998, | |
| "logits/rejected": -1.754926323890686, | |
| "logps/chosen": -377.4764099121094, | |
| "logps/rejected": -506.1446838378906, | |
| "loss": 0.3958, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -2.0154218673706055, | |
| "rewards/margins": 1.3107969760894775, | |
| "rewards/rejected": -3.326218843460083, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "grad_norm": 25.762153559701787, | |
| "learning_rate": 2.8668261861384043e-07, | |
| "logits/chosen": -1.7258059978485107, | |
| "logits/rejected": -1.7285445928573608, | |
| "logps/chosen": -293.87701416015625, | |
| "logps/rejected": -444.28558349609375, | |
| "loss": 0.3768, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -1.5132906436920166, | |
| "rewards/margins": 1.4502919912338257, | |
| "rewards/rejected": -2.9635825157165527, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "grad_norm": 24.379288179820218, | |
| "learning_rate": 2.7144932808610996e-07, | |
| "logits/chosen": -1.8721590042114258, | |
| "logits/rejected": -1.86797297000885, | |
| "logps/chosen": -325.639892578125, | |
| "logps/rejected": -451.44403076171875, | |
| "loss": 0.3765, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -1.4320765733718872, | |
| "rewards/margins": 1.3416272401809692, | |
| "rewards/rejected": -2.7737040519714355, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "grad_norm": 26.000189975206272, | |
| "learning_rate": 2.561353071307281e-07, | |
| "logits/chosen": -1.8530679941177368, | |
| "logits/rejected": -1.847919225692749, | |
| "logps/chosen": -271.6641845703125, | |
| "logps/rejected": -430.631591796875, | |
| "loss": 0.4035, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.2568687200546265, | |
| "rewards/margins": 1.369762659072876, | |
| "rewards/rejected": -2.626631259918213, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "grad_norm": 21.301492047046295, | |
| "learning_rate": 2.4079819426466026e-07, | |
| "logits/chosen": -1.7658283710479736, | |
| "logits/rejected": -1.7549536228179932, | |
| "logps/chosen": -323.47308349609375, | |
| "logps/rejected": -471.88787841796875, | |
| "loss": 0.3771, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.5166125297546387, | |
| "rewards/margins": 1.4360930919647217, | |
| "rewards/rejected": -2.9527056217193604, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "grad_norm": 25.43497345992236, | |
| "learning_rate": 2.2549571491760981e-07, | |
| "logits/chosen": -1.8855187892913818, | |
| "logits/rejected": -1.8896701335906982, | |
| "logps/chosen": -339.31903076171875, | |
| "logps/rejected": -464.59307861328125, | |
| "loss": 0.4002, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -1.6602863073349, | |
| "rewards/margins": 1.2881882190704346, | |
| "rewards/rejected": -2.948474407196045, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "grad_norm": 26.47261622045951, | |
| "learning_rate": 2.1028546416653467e-07, | |
| "logits/chosen": -1.7920734882354736, | |
| "logits/rejected": -1.7969112396240234, | |
| "logps/chosen": -306.63165283203125, | |
| "logps/rejected": -428.21380615234375, | |
| "loss": 0.3874, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.4940059185028076, | |
| "rewards/margins": 1.1801685094833374, | |
| "rewards/rejected": -2.6741740703582764, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "grad_norm": 23.402243080751038, | |
| "learning_rate": 1.9522468996078255e-07, | |
| "logits/chosen": -1.7678264379501343, | |
| "logits/rejected": -1.7712920904159546, | |
| "logps/chosen": -367.00921630859375, | |
| "logps/rejected": -499.3409729003906, | |
| "loss": 0.3647, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -1.829602837562561, | |
| "rewards/margins": 1.4595292806625366, | |
| "rewards/rejected": -3.2891323566436768, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 25.45097439145198, | |
| "learning_rate": 1.8037007765373673e-07, | |
| "logits/chosen": -1.7975505590438843, | |
| "logits/rejected": -1.8053462505340576, | |
| "logps/chosen": -383.27325439453125, | |
| "logps/rejected": -521.484375, | |
| "loss": 0.3859, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": -1.9635961055755615, | |
| "rewards/margins": 1.4095726013183594, | |
| "rewards/rejected": -3.373168468475342, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "grad_norm": 23.89434286774924, | |
| "learning_rate": 1.6577753665194498e-07, | |
| "logits/chosen": -1.7756551504135132, | |
| "logits/rejected": -1.7834348678588867, | |
| "logps/chosen": -367.2814025878906, | |
| "logps/rejected": -514.7007446289062, | |
| "loss": 0.3882, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.9314982891082764, | |
| "rewards/margins": 1.3575429916381836, | |
| "rewards/rejected": -3.2890408039093018, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "grad_norm": 29.667518628507914, | |
| "learning_rate": 1.5150198998473802e-07, | |
| "logits/chosen": -1.8071187734603882, | |
| "logits/rejected": -1.8050048351287842, | |
| "logps/chosen": -344.69696044921875, | |
| "logps/rejected": -520.0059814453125, | |
| "loss": 0.3719, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -1.7867143154144287, | |
| "rewards/margins": 1.6172891855239868, | |
| "rewards/rejected": -3.404003620147705, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "grad_norm": 14.245618025323719, | |
| "learning_rate": 1.3759716758634832e-07, | |
| "logits/chosen": -1.711626410484314, | |
| "logits/rejected": -1.7173089981079102, | |
| "logps/chosen": -298.3590393066406, | |
| "logps/rejected": -468.08642578125, | |
| "loss": 0.2819, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": -1.4959304332733154, | |
| "rewards/margins": 1.685693383216858, | |
| "rewards/rejected": -3.181623697280884, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "grad_norm": 20.768785145747042, | |
| "learning_rate": 1.2411540406857063e-07, | |
| "logits/chosen": -1.7734801769256592, | |
| "logits/rejected": -1.769174337387085, | |
| "logps/chosen": -335.4779052734375, | |
| "logps/rejected": -563.084716796875, | |
| "loss": 0.2211, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.7001539468765259, | |
| "rewards/margins": 2.14009165763855, | |
| "rewards/rejected": -3.840245485305786, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "grad_norm": 22.179102644451653, | |
| "learning_rate": 1.1110744174509951e-07, | |
| "logits/chosen": -1.675746202468872, | |
| "logits/rejected": -1.6720969676971436, | |
| "logps/chosen": -331.4947509765625, | |
| "logps/rejected": -547.267333984375, | |
| "loss": 0.2259, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.7998850345611572, | |
| "rewards/margins": 2.0022103786468506, | |
| "rewards/rejected": -3.802095413208008, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "grad_norm": 20.445357164590412, | |
| "learning_rate": 9.862223964891864e-08, | |
| "logits/chosen": -1.7023229598999023, | |
| "logits/rejected": -1.6805604696273804, | |
| "logps/chosen": -357.96478271484375, | |
| "logps/rejected": -561.7637939453125, | |
| "loss": 0.2174, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -2.055269718170166, | |
| "rewards/margins": 1.9106111526489258, | |
| "rewards/rejected": -3.9658806324005127, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "grad_norm": 18.35482979743842, | |
| "learning_rate": 8.670678926155586e-08, | |
| "logits/chosen": -1.6157821416854858, | |
| "logits/rejected": -1.5976340770721436, | |
| "logps/chosen": -380.42376708984375, | |
| "logps/rejected": -607.9219970703125, | |
| "loss": 0.2369, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -2.2109501361846924, | |
| "rewards/margins": 2.16408109664917, | |
| "rewards/rejected": -4.375030994415283, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "grad_norm": 19.381238020569043, | |
| "learning_rate": 7.540593764775679e-08, | |
| "logits/chosen": -1.6243540048599243, | |
| "logits/rejected": -1.6052665710449219, | |
| "logps/chosen": -393.17193603515625, | |
| "logps/rejected": -632.5490112304688, | |
| "loss": 0.2185, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -2.124055862426758, | |
| "rewards/margins": 2.250814437866211, | |
| "rewards/rejected": -4.374870300292969, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "grad_norm": 19.10044218783973, | |
| "learning_rate": 6.476221866126027e-08, | |
| "logits/chosen": -1.5514600276947021, | |
| "logits/rejected": -1.5442235469818115, | |
| "logps/chosen": -389.6855773925781, | |
| "logps/rejected": -634.3771362304688, | |
| "loss": 0.2148, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -2.2639551162719727, | |
| "rewards/margins": 2.4029526710510254, | |
| "rewards/rejected": -4.666908264160156, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 26.319378436487725, | |
| "learning_rate": 5.481569285697637e-08, | |
| "logits/chosen": -1.5155491828918457, | |
| "logits/rejected": -1.4932228326797485, | |
| "logps/chosen": -343.5061950683594, | |
| "logps/rejected": -606.8117065429688, | |
| "loss": 0.2209, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.116722583770752, | |
| "rewards/margins": 2.46051025390625, | |
| "rewards/rejected": -4.577232837677002, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "grad_norm": 17.545599464428165, | |
| "learning_rate": 4.5603796712104096e-08, | |
| "logits/chosen": -1.6445671319961548, | |
| "logits/rejected": -1.6368439197540283, | |
| "logps/chosen": -397.17108154296875, | |
| "logps/rejected": -666.0770874023438, | |
| "loss": 0.1854, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.1950583457946777, | |
| "rewards/margins": 2.5617446899414062, | |
| "rewards/rejected": -4.756802558898926, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "grad_norm": 17.948454753401542, | |
| "learning_rate": 3.71612017236837e-08, | |
| "logits/chosen": -1.6224250793457031, | |
| "logits/rejected": -1.6266944408416748, | |
| "logps/chosen": -398.54791259765625, | |
| "logps/rejected": -611.5243530273438, | |
| "loss": 0.216, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -2.202561616897583, | |
| "rewards/margins": 2.2089762687683105, | |
| "rewards/rejected": -4.411538124084473, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "grad_norm": 23.409793316564198, | |
| "learning_rate": 2.9519683912911263e-08, | |
| "logits/chosen": -1.5537071228027344, | |
| "logits/rejected": -1.5374433994293213, | |
| "logps/chosen": -386.80120849609375, | |
| "logps/rejected": -654.9160766601562, | |
| "loss": 0.1926, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.3248279094696045, | |
| "rewards/margins": 2.6371538639068604, | |
| "rewards/rejected": -4.961981773376465, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "grad_norm": 21.46300987370094, | |
| "learning_rate": 2.2708004227369403e-08, | |
| "logits/chosen": -1.538722276687622, | |
| "logits/rejected": -1.550920009613037, | |
| "logps/chosen": -395.60992431640625, | |
| "logps/rejected": -598.1488037109375, | |
| "loss": 0.2258, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -2.264441728591919, | |
| "rewards/margins": 2.119832754135132, | |
| "rewards/rejected": -4.384274482727051, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "grad_norm": 44.93508676895646, | |
| "learning_rate": 1.675180029131529e-08, | |
| "logits/chosen": -1.5520565509796143, | |
| "logits/rejected": -1.518765926361084, | |
| "logps/chosen": -401.32269287109375, | |
| "logps/rejected": -650.5725708007812, | |
| "loss": 0.1992, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -2.491032123565674, | |
| "rewards/margins": 2.3947300910949707, | |
| "rewards/rejected": -4.8857622146606445, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "grad_norm": 21.84206710153422, | |
| "learning_rate": 1.1673489911451533e-08, | |
| "logits/chosen": -1.5932598114013672, | |
| "logits/rejected": -1.5699536800384521, | |
| "logps/chosen": -373.67340087890625, | |
| "logps/rejected": -625.8887939453125, | |
| "loss": 0.2074, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -2.152372360229492, | |
| "rewards/margins": 2.410846710205078, | |
| "rewards/rejected": -4.5632195472717285, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.77, | |
| "grad_norm": 18.75966723843524, | |
| "learning_rate": 7.492186701364006e-09, | |
| "logits/chosen": -1.5413577556610107, | |
| "logits/rejected": -1.5304628610610962, | |
| "logps/chosen": -389.6387634277344, | |
| "logps/rejected": -657.2249145507812, | |
| "loss": 0.2081, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -2.3259997367858887, | |
| "rewards/margins": 2.552927017211914, | |
| "rewards/rejected": -4.878927230834961, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 19.72453682602052, | |
| "learning_rate": 4.223628142195928e-09, | |
| "logits/chosen": -1.5502567291259766, | |
| "logits/rejected": -1.519486665725708, | |
| "logps/chosen": -397.04425048828125, | |
| "logps/rejected": -660.490478515625, | |
| "loss": 0.1973, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -2.2548627853393555, | |
| "rewards/margins": 2.6391000747680664, | |
| "rewards/rejected": -4.893963813781738, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "grad_norm": 18.733548354455543, | |
| "learning_rate": 1.880116350322508e-09, | |
| "logits/chosen": -1.587247610092163, | |
| "logits/rejected": -1.5682843923568726, | |
| "logps/chosen": -352.83782958984375, | |
| "logps/rejected": -596.1420288085938, | |
| "loss": 0.2228, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -2.089592456817627, | |
| "rewards/margins": 2.3892369270324707, | |
| "rewards/rejected": -4.478829383850098, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "grad_norm": 25.376336588280168, | |
| "learning_rate": 4.704717749627052e-10, | |
| "logits/chosen": -1.5502140522003174, | |
| "logits/rejected": -1.5279264450073242, | |
| "logps/chosen": -416.4481506347656, | |
| "logps/rejected": -668.2899169921875, | |
| "loss": 0.2221, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.377631425857544, | |
| "rewards/margins": 2.5437731742858887, | |
| "rewards/rejected": -4.9214043617248535, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "grad_norm": 23.399219955803773, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -1.5591785907745361, | |
| "logits/rejected": -1.5488736629486084, | |
| "logps/chosen": -378.39990234375, | |
| "logps/rejected": -633.2120361328125, | |
| "loss": 0.2165, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.2136266231536865, | |
| "rewards/margins": 2.3755829334259033, | |
| "rewards/rejected": -4.58920955657959, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "step": 285, | |
| "total_flos": 0.0, | |
| "train_loss": 0.4176908902954637, | |
| "train_runtime": 7252.3508, | |
| "train_samples_per_second": 5.058, | |
| "train_steps_per_second": 0.039 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 285, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |