| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9997382884061764, | |
| "eval_steps": 100, | |
| "global_step": 955, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0010468463752944255, | |
| "grad_norm": 5.5, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -0.32969313859939575, | |
| "logits/rejected": -0.3478139042854309, | |
| "logps/chosen": -282.7626953125, | |
| "logps/rejected": -210.6118621826172, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.010468463752944255, | |
| "grad_norm": 6.0, | |
| "learning_rate": 5.208333333333334e-07, | |
| "logits/chosen": -0.5201655030250549, | |
| "logits/rejected": -0.5292737483978271, | |
| "logps/chosen": -312.750244140625, | |
| "logps/rejected": -275.4716491699219, | |
| "loss": 0.6919, | |
| "rewards/accuracies": 0.4652777910232544, | |
| "rewards/chosen": 0.0034895986318588257, | |
| "rewards/margins": 0.002287701005116105, | |
| "rewards/rejected": 0.001201898674480617, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02093692750588851, | |
| "grad_norm": 5.5, | |
| "learning_rate": 1.0416666666666667e-06, | |
| "logits/chosen": -0.4813454747200012, | |
| "logits/rejected": -0.5021841526031494, | |
| "logps/chosen": -304.3456726074219, | |
| "logps/rejected": -268.962158203125, | |
| "loss": 0.6934, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.0017953884089365602, | |
| "rewards/margins": 0.0013109362917020917, | |
| "rewards/rejected": -0.003106324467808008, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.031405391258832765, | |
| "grad_norm": 5.78125, | |
| "learning_rate": 1.5625e-06, | |
| "logits/chosen": -0.4880186915397644, | |
| "logits/rejected": -0.4969080984592438, | |
| "logps/chosen": -287.50286865234375, | |
| "logps/rejected": -249.9375457763672, | |
| "loss": 0.688, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.010473029688000679, | |
| "rewards/margins": 0.01431797631084919, | |
| "rewards/rejected": -0.003844945225864649, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04187385501177702, | |
| "grad_norm": 5.375, | |
| "learning_rate": 2.0833333333333334e-06, | |
| "logits/chosen": -0.46122869849205017, | |
| "logits/rejected": -0.4958384037017822, | |
| "logps/chosen": -303.1500549316406, | |
| "logps/rejected": -241.998046875, | |
| "loss": 0.683, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.03339455649256706, | |
| "rewards/margins": 0.0322079211473465, | |
| "rewards/rejected": 0.0011866316199302673, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.05234231876472128, | |
| "grad_norm": 4.59375, | |
| "learning_rate": 2.604166666666667e-06, | |
| "logits/chosen": -0.5217183828353882, | |
| "logits/rejected": -0.5492846369743347, | |
| "logps/chosen": -302.4110412597656, | |
| "logps/rejected": -280.4798278808594, | |
| "loss": 0.6784, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.07218249142169952, | |
| "rewards/margins": 0.04362189769744873, | |
| "rewards/rejected": 0.028560593724250793, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06281078251766553, | |
| "grad_norm": 5.03125, | |
| "learning_rate": 3.125e-06, | |
| "logits/chosen": -0.5444208979606628, | |
| "logits/rejected": -0.5450385212898254, | |
| "logps/chosen": -288.3908996582031, | |
| "logps/rejected": -252.4159393310547, | |
| "loss": 0.662, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.10959690809249878, | |
| "rewards/margins": 0.04304002225399017, | |
| "rewards/rejected": 0.0665568858385086, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07327924627060979, | |
| "grad_norm": 5.21875, | |
| "learning_rate": 3.6458333333333333e-06, | |
| "logits/chosen": -0.5294634699821472, | |
| "logits/rejected": -0.5446035265922546, | |
| "logps/chosen": -284.2558288574219, | |
| "logps/rejected": -261.0161437988281, | |
| "loss": 0.6465, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.1622026413679123, | |
| "rewards/margins": 0.1133759468793869, | |
| "rewards/rejected": 0.04882669448852539, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.08374771002355404, | |
| "grad_norm": 4.78125, | |
| "learning_rate": 4.166666666666667e-06, | |
| "logits/chosen": -0.46931973099708557, | |
| "logits/rejected": -0.4959704279899597, | |
| "logps/chosen": -285.55169677734375, | |
| "logps/rejected": -271.7218017578125, | |
| "loss": 0.6312, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.2240675389766693, | |
| "rewards/margins": 0.17874710261821747, | |
| "rewards/rejected": 0.04532044008374214, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.0942161737764983, | |
| "grad_norm": 5.90625, | |
| "learning_rate": 4.6875000000000004e-06, | |
| "logits/chosen": -0.5021494626998901, | |
| "logits/rejected": -0.5237849950790405, | |
| "logps/chosen": -328.7523193359375, | |
| "logps/rejected": -293.5586853027344, | |
| "loss": 0.6197, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.19337503612041473, | |
| "rewards/margins": 0.22977392375469208, | |
| "rewards/rejected": -0.036398887634277344, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.10468463752944256, | |
| "grad_norm": 5.65625, | |
| "learning_rate": 4.9997324926814375e-06, | |
| "logits/chosen": -0.5387491583824158, | |
| "logits/rejected": -0.5343273282051086, | |
| "logps/chosen": -274.4442443847656, | |
| "logps/rejected": -288.30035400390625, | |
| "loss": 0.6297, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.221399188041687, | |
| "rewards/margins": 0.2429426610469818, | |
| "rewards/rejected": -0.0215434692800045, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.10468463752944256, | |
| "eval_logits/chosen": -0.5525645613670349, | |
| "eval_logits/rejected": -0.5478540062904358, | |
| "eval_logps/chosen": -288.4034118652344, | |
| "eval_logps/rejected": -275.7340087890625, | |
| "eval_loss": 0.6140495538711548, | |
| "eval_rewards/accuracies": 0.6959999799728394, | |
| "eval_rewards/chosen": 0.13576927781105042, | |
| "eval_rewards/margins": 0.26342061161994934, | |
| "eval_rewards/rejected": -0.1276513636112213, | |
| "eval_runtime": 351.3471, | |
| "eval_samples_per_second": 5.692, | |
| "eval_steps_per_second": 0.356, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11515310128238682, | |
| "grad_norm": 4.25, | |
| "learning_rate": 4.996723692767927e-06, | |
| "logits/chosen": -0.5775566101074219, | |
| "logits/rejected": -0.603683590888977, | |
| "logps/chosen": -287.82733154296875, | |
| "logps/rejected": -275.31671142578125, | |
| "loss": 0.606, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.15195702016353607, | |
| "rewards/margins": 0.34537211060523987, | |
| "rewards/rejected": -0.1934150755405426, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12562156503533106, | |
| "grad_norm": 4.96875, | |
| "learning_rate": 4.9903757462135984e-06, | |
| "logits/chosen": -0.5295812487602234, | |
| "logits/rejected": -0.5440601110458374, | |
| "logps/chosen": -260.30078125, | |
| "logps/rejected": -251.58401489257812, | |
| "loss": 0.5924, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.02338511124253273, | |
| "rewards/margins": 0.281913697719574, | |
| "rewards/rejected": -0.25852862000465393, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.1360900287882753, | |
| "grad_norm": 5.46875, | |
| "learning_rate": 4.980697142834315e-06, | |
| "logits/chosen": -0.49980980157852173, | |
| "logits/rejected": -0.5171049237251282, | |
| "logps/chosen": -300.2707214355469, | |
| "logps/rejected": -336.19036865234375, | |
| "loss": 0.5933, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.02033509686589241, | |
| "rewards/margins": 0.3162926733493805, | |
| "rewards/rejected": -0.2959575653076172, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.14655849254121958, | |
| "grad_norm": 4.75, | |
| "learning_rate": 4.967700826904229e-06, | |
| "logits/chosen": -0.5714845061302185, | |
| "logits/rejected": -0.5790292024612427, | |
| "logps/chosen": -280.66845703125, | |
| "logps/rejected": -273.6711120605469, | |
| "loss": 0.5811, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.004461909644305706, | |
| "rewards/margins": 0.407376229763031, | |
| "rewards/rejected": -0.4118381440639496, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15702695629416383, | |
| "grad_norm": 4.84375, | |
| "learning_rate": 4.951404179843963e-06, | |
| "logits/chosen": -0.5678334832191467, | |
| "logits/rejected": -0.5395095944404602, | |
| "logps/chosen": -306.054443359375, | |
| "logps/rejected": -276.5906066894531, | |
| "loss": 0.5777, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.13024172186851501, | |
| "rewards/margins": 0.4546026587486267, | |
| "rewards/rejected": -0.3243609070777893, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.16749542004710807, | |
| "grad_norm": 4.375, | |
| "learning_rate": 4.931828996974498e-06, | |
| "logits/chosen": -0.5105162858963013, | |
| "logits/rejected": -0.5018196702003479, | |
| "logps/chosen": -296.04840087890625, | |
| "logps/rejected": -270.5509033203125, | |
| "loss": 0.5478, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.04980681464076042, | |
| "rewards/margins": 0.5175832509994507, | |
| "rewards/rejected": -0.4677763879299164, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.17796388380005235, | |
| "grad_norm": 4.78125, | |
| "learning_rate": 4.909001458367867e-06, | |
| "logits/chosen": -0.5939355492591858, | |
| "logits/rejected": -0.5804780125617981, | |
| "logps/chosen": -287.94708251953125, | |
| "logps/rejected": -275.73980712890625, | |
| "loss": 0.5789, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.19239839911460876, | |
| "rewards/margins": 0.4369952082633972, | |
| "rewards/rejected": -0.6293936967849731, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.1884323475529966, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 4.882952093833628e-06, | |
| "logits/chosen": -0.5993765592575073, | |
| "logits/rejected": -0.5707160830497742, | |
| "logps/chosen": -301.9080505371094, | |
| "logps/rejected": -265.2618713378906, | |
| "loss": 0.555, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.19447076320648193, | |
| "rewards/margins": 0.48417288064956665, | |
| "rewards/rejected": -0.6786437630653381, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.19890081130594087, | |
| "grad_norm": 5.90625, | |
| "learning_rate": 4.853715742087947e-06, | |
| "logits/chosen": -0.5384653806686401, | |
| "logits/rejected": -0.5204354524612427, | |
| "logps/chosen": -275.14886474609375, | |
| "logps/rejected": -281.9067687988281, | |
| "loss": 0.5706, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.033171653747558594, | |
| "rewards/margins": 0.469027042388916, | |
| "rewards/rejected": -0.5021986961364746, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "grad_norm": 6.5625, | |
| "learning_rate": 4.821331504159906e-06, | |
| "logits/chosen": -0.5440086722373962, | |
| "logits/rejected": -0.5563145279884338, | |
| "logps/chosen": -296.43060302734375, | |
| "logps/rejected": -255.0513916015625, | |
| "loss": 0.5676, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.12640251219272614, | |
| "rewards/margins": 0.5171942710876465, | |
| "rewards/rejected": -0.6435968279838562, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "eval_logits/chosen": -0.5116367340087891, | |
| "eval_logits/rejected": -0.4945087134838104, | |
| "eval_logps/chosen": -290.90509033203125, | |
| "eval_logps/rejected": -281.0560302734375, | |
| "eval_loss": 0.5569251179695129, | |
| "eval_rewards/accuracies": 0.699999988079071, | |
| "eval_rewards/chosen": -0.11436504870653152, | |
| "eval_rewards/margins": 0.5454893708229065, | |
| "eval_rewards/rejected": -0.6598544716835022, | |
| "eval_runtime": 349.6273, | |
| "eval_samples_per_second": 5.72, | |
| "eval_steps_per_second": 0.358, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21983773881182936, | |
| "grad_norm": 6.15625, | |
| "learning_rate": 4.7858426910973435e-06, | |
| "logits/chosen": -0.5827375650405884, | |
| "logits/rejected": -0.5831678509712219, | |
| "logps/chosen": -278.0469970703125, | |
| "logps/rejected": -271.9605407714844, | |
| "loss": 0.5508, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.025282684713602066, | |
| "rewards/margins": 0.5209798216819763, | |
| "rewards/rejected": -0.5462625622749329, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.23030620256477363, | |
| "grad_norm": 5.125, | |
| "learning_rate": 4.747296766042161e-06, | |
| "logits/chosen": -0.5575278997421265, | |
| "logits/rejected": -0.5416348576545715, | |
| "logps/chosen": -317.4739990234375, | |
| "logps/rejected": -269.8050231933594, | |
| "loss": 0.5368, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.03362050652503967, | |
| "rewards/margins": 0.569665789604187, | |
| "rewards/rejected": -0.536045253276825, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.24077466631771788, | |
| "grad_norm": 5.6875, | |
| "learning_rate": 4.705745280752586e-06, | |
| "logits/chosen": -0.589636504650116, | |
| "logits/rejected": -0.5510541200637817, | |
| "logps/chosen": -290.55206298828125, | |
| "logps/rejected": -287.7290954589844, | |
| "loss": 0.5581, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.08335064351558685, | |
| "rewards/margins": 0.5679007768630981, | |
| "rewards/rejected": -0.6512514352798462, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.2512431300706621, | |
| "grad_norm": 4.96875, | |
| "learning_rate": 4.661243806657256e-06, | |
| "logits/chosen": -0.5663197636604309, | |
| "logits/rejected": -0.525894045829773, | |
| "logps/chosen": -299.3938903808594, | |
| "logps/rejected": -263.598876953125, | |
| "loss": 0.5563, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.1382952779531479, | |
| "rewards/margins": 0.5307878255844116, | |
| "rewards/rejected": -0.6690832376480103, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.26171159382360637, | |
| "grad_norm": 5.6875, | |
| "learning_rate": 4.613851860533367e-06, | |
| "logits/chosen": -0.5768588781356812, | |
| "logits/rejected": -0.5346202850341797, | |
| "logps/chosen": -293.30059814453125, | |
| "logps/rejected": -259.82177734375, | |
| "loss": 0.5789, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.08053269237279892, | |
| "rewards/margins": 0.5023098587989807, | |
| "rewards/rejected": -0.5828425288200378, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.2721800575765506, | |
| "grad_norm": 5.6875, | |
| "learning_rate": 4.563632824908252e-06, | |
| "logits/chosen": -0.5458053946495056, | |
| "logits/rejected": -0.5105060935020447, | |
| "logps/chosen": -292.6302185058594, | |
| "logps/rejected": -279.1827392578125, | |
| "loss": 0.5404, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.12764385342597961, | |
| "rewards/margins": 0.8003277778625488, | |
| "rewards/rejected": -0.6726840138435364, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.2826485213294949, | |
| "grad_norm": 5.78125, | |
| "learning_rate": 4.510653863290871e-06, | |
| "logits/chosen": -0.545439600944519, | |
| "logits/rejected": -0.5205134749412537, | |
| "logps/chosen": -294.71783447265625, | |
| "logps/rejected": -302.39599609375, | |
| "loss": 0.5465, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.015889674425125122, | |
| "rewards/margins": 0.5484660267829895, | |
| "rewards/rejected": -0.564355731010437, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.29311698508243916, | |
| "grad_norm": 4.78125, | |
| "learning_rate": 4.454985830346574e-06, | |
| "logits/chosen": -0.6070719957351685, | |
| "logits/rejected": -0.5709980726242065, | |
| "logps/chosen": -300.3115234375, | |
| "logps/rejected": -283.37689208984375, | |
| "loss": 0.5681, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.04909560829401016, | |
| "rewards/margins": 0.496112197637558, | |
| "rewards/rejected": -0.5452078580856323, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.3035854488353834, | |
| "grad_norm": 4.65625, | |
| "learning_rate": 4.396703177135262e-06, | |
| "logits/chosen": -0.5642537474632263, | |
| "logits/rejected": -0.5349927544593811, | |
| "logps/chosen": -285.5408935546875, | |
| "logps/rejected": -256.77691650390625, | |
| "loss": 0.5249, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.10653576999902725, | |
| "rewards/margins": 0.6360087990760803, | |
| "rewards/rejected": -0.5294729471206665, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.31405391258832765, | |
| "grad_norm": 5.75, | |
| "learning_rate": 4.335883851539693e-06, | |
| "logits/chosen": -0.5834770798683167, | |
| "logits/rejected": -0.544857382774353, | |
| "logps/chosen": -296.65936279296875, | |
| "logps/rejected": -293.7464599609375, | |
| "loss": 0.5414, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.30995655059814453, | |
| "rewards/margins": 0.6452124714851379, | |
| "rewards/rejected": -0.9551690220832825, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.31405391258832765, | |
| "eval_logits/chosen": -0.4775179624557495, | |
| "eval_logits/rejected": -0.45395201444625854, | |
| "eval_logps/chosen": -293.56982421875, | |
| "eval_logps/rejected": -284.91802978515625, | |
| "eval_loss": 0.5403068661689758, | |
| "eval_rewards/accuracies": 0.7260000109672546, | |
| "eval_rewards/chosen": -0.38083866238594055, | |
| "eval_rewards/margins": 0.6652126312255859, | |
| "eval_rewards/rejected": -1.0460513830184937, | |
| "eval_runtime": 349.5769, | |
| "eval_samples_per_second": 5.721, | |
| "eval_steps_per_second": 0.358, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3245223763412719, | |
| "grad_norm": 6.28125, | |
| "learning_rate": 4.2726091940171055e-06, | |
| "logits/chosen": -0.5361314415931702, | |
| "logits/rejected": -0.5678786635398865, | |
| "logps/chosen": -296.285400390625, | |
| "logps/rejected": -342.7236022949219, | |
| "loss": 0.5056, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -0.19761790335178375, | |
| "rewards/margins": 0.8299154043197632, | |
| "rewards/rejected": -1.0275332927703857, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.33499084009421615, | |
| "grad_norm": 5.21875, | |
| "learning_rate": 4.206963828813555e-06, | |
| "logits/chosen": -0.5878556370735168, | |
| "logits/rejected": -0.5477277040481567, | |
| "logps/chosen": -296.2107849121094, | |
| "logps/rejected": -280.1604919433594, | |
| "loss": 0.524, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.1949264109134674, | |
| "rewards/margins": 0.7900466918945312, | |
| "rewards/rejected": -0.9849729537963867, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.34545930384716045, | |
| "grad_norm": 5.0625, | |
| "learning_rate": 4.139035550786495e-06, | |
| "logits/chosen": -0.6225862503051758, | |
| "logits/rejected": -0.5590678453445435, | |
| "logps/chosen": -289.2728576660156, | |
| "logps/rejected": -260.7643737792969, | |
| "loss": 0.5217, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.2666059136390686, | |
| "rewards/margins": 0.6325812339782715, | |
| "rewards/rejected": -0.8991872072219849, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.3559277676001047, | |
| "grad_norm": 5.96875, | |
| "learning_rate": 4.068915207986931e-06, | |
| "logits/chosen": -0.5779263377189636, | |
| "logits/rejected": -0.5189130902290344, | |
| "logps/chosen": -298.147705078125, | |
| "logps/rejected": -258.82391357421875, | |
| "loss": 0.5224, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.4627823233604431, | |
| "rewards/margins": 0.7694815397262573, | |
| "rewards/rejected": -1.2322638034820557, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.36639623135304894, | |
| "grad_norm": 6.21875, | |
| "learning_rate": 3.996696580158211e-06, | |
| "logits/chosen": -0.5267971158027649, | |
| "logits/rejected": -0.5050491690635681, | |
| "logps/chosen": -335.9749755859375, | |
| "logps/rejected": -290.8840026855469, | |
| "loss": 0.5206, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.16786900162696838, | |
| "rewards/margins": 0.7010205984115601, | |
| "rewards/rejected": -0.8688896894454956, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.3768646951059932, | |
| "grad_norm": 5.3125, | |
| "learning_rate": 3.922476253313921e-06, | |
| "logits/chosen": -0.5035579800605774, | |
| "logits/rejected": -0.5138710141181946, | |
| "logps/chosen": -273.709716796875, | |
| "logps/rejected": -296.85650634765625, | |
| "loss": 0.5587, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.28165164589881897, | |
| "rewards/margins": 0.6714268922805786, | |
| "rewards/rejected": -0.9530784487724304, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.38733315885893743, | |
| "grad_norm": 5.875, | |
| "learning_rate": 3.846353490562664e-06, | |
| "logits/chosen": -0.5329638123512268, | |
| "logits/rejected": -0.5389483571052551, | |
| "logps/chosen": -288.07745361328125, | |
| "logps/rejected": -262.96124267578125, | |
| "loss": 0.5038, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.05612688139081001, | |
| "rewards/margins": 0.8792274594306946, | |
| "rewards/rejected": -0.9353543519973755, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.39780162261188173, | |
| "grad_norm": 5.25, | |
| "learning_rate": 3.768430099352445e-06, | |
| "logits/chosen": -0.5536895990371704, | |
| "logits/rejected": -0.5606611371040344, | |
| "logps/chosen": -304.82574462890625, | |
| "logps/rejected": -277.5342712402344, | |
| "loss": 0.5089, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.04281007498502731, | |
| "rewards/margins": 0.9195000529289246, | |
| "rewards/rejected": -0.9623101353645325, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.408270086364826, | |
| "grad_norm": 6.5, | |
| "learning_rate": 3.6888102953122307e-06, | |
| "logits/chosen": -0.5975215435028076, | |
| "logits/rejected": -0.5678552389144897, | |
| "logps/chosen": -261.8155212402344, | |
| "logps/rejected": -262.86151123046875, | |
| "loss": 0.5554, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.049895767122507095, | |
| "rewards/margins": 0.72125244140625, | |
| "rewards/rejected": -0.7711483240127563, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "grad_norm": 5.40625, | |
| "learning_rate": 3.607600562872785e-06, | |
| "logits/chosen": -0.5524119138717651, | |
| "logits/rejected": -0.5214421153068542, | |
| "logps/chosen": -284.4479675292969, | |
| "logps/rejected": -274.91790771484375, | |
| "loss": 0.5124, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.08701231330633163, | |
| "rewards/margins": 0.7274686694145203, | |
| "rewards/rejected": -0.8144810795783997, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "eval_logits/chosen": -0.45163384079933167, | |
| "eval_logits/rejected": -0.42426639795303345, | |
| "eval_logps/chosen": -292.09857177734375, | |
| "eval_logps/rejected": -284.3532409667969, | |
| "eval_loss": 0.5341187715530396, | |
| "eval_rewards/accuracies": 0.7039999961853027, | |
| "eval_rewards/chosen": -0.23370328545570374, | |
| "eval_rewards/margins": 0.7558707594871521, | |
| "eval_rewards/rejected": -0.9895740151405334, | |
| "eval_runtime": 349.3488, | |
| "eval_samples_per_second": 5.725, | |
| "eval_steps_per_second": 0.358, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.42920701387071447, | |
| "grad_norm": 5.1875, | |
| "learning_rate": 3.5249095128531863e-06, | |
| "logits/chosen": -0.5735105872154236, | |
| "logits/rejected": -0.5243286490440369, | |
| "logps/chosen": -278.07611083984375, | |
| "logps/rejected": -275.5641784667969, | |
| "loss": 0.5147, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.19002243876457214, | |
| "rewards/margins": 0.7818821668624878, | |
| "rewards/rejected": -0.9719046354293823, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.4396754776236587, | |
| "grad_norm": 5.75, | |
| "learning_rate": 3.4408477372034743e-06, | |
| "logits/chosen": -0.5684117078781128, | |
| "logits/rejected": -0.5453870296478271, | |
| "logps/chosen": -308.50128173828125, | |
| "logps/rejected": -296.2371826171875, | |
| "loss": 0.5608, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.22137188911437988, | |
| "rewards/margins": 0.63337242603302, | |
| "rewards/rejected": -0.8547442555427551, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.45014394137660296, | |
| "grad_norm": 6.125, | |
| "learning_rate": 3.355527661097728e-06, | |
| "logits/chosen": -0.5637370347976685, | |
| "logits/rejected": -0.5657171010971069, | |
| "logps/chosen": -280.50177001953125, | |
| "logps/rejected": -281.47467041015625, | |
| "loss": 0.5229, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.39310961961746216, | |
| "rewards/margins": 0.6110748648643494, | |
| "rewards/rejected": -1.0041844844818115, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.46061240512954726, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 3.269063392575352e-06, | |
| "logits/chosen": -0.5321542024612427, | |
| "logits/rejected": -0.5234124660491943, | |
| "logps/chosen": -328.39447021484375, | |
| "logps/rejected": -307.01556396484375, | |
| "loss": 0.5135, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.2074432671070099, | |
| "rewards/margins": 0.7093298435211182, | |
| "rewards/rejected": -0.9167732000350952, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.4710808688824915, | |
| "grad_norm": 6.625, | |
| "learning_rate": 3.181570569931697e-06, | |
| "logits/chosen": -0.5547568202018738, | |
| "logits/rejected": -0.5429819822311401, | |
| "logps/chosen": -286.77447509765625, | |
| "logps/rejected": -282.605224609375, | |
| "loss": 0.5, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.41850847005844116, | |
| "rewards/margins": 0.6984038352966309, | |
| "rewards/rejected": -1.1169124841690063, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.48154933263543576, | |
| "grad_norm": 5.125, | |
| "learning_rate": 3.09316620706208e-06, | |
| "logits/chosen": -0.5107685327529907, | |
| "logits/rejected": -0.5204188227653503, | |
| "logps/chosen": -307.8090515136719, | |
| "logps/rejected": -288.7105407714844, | |
| "loss": 0.4984, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.21866807341575623, | |
| "rewards/margins": 0.888043999671936, | |
| "rewards/rejected": -1.106711983680725, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.49201779638838, | |
| "grad_norm": 5.3125, | |
| "learning_rate": 3.0039685369660785e-06, | |
| "logits/chosen": -0.5120314359664917, | |
| "logits/rejected": -0.46821707487106323, | |
| "logps/chosen": -281.59722900390625, | |
| "logps/rejected": -266.38134765625, | |
| "loss": 0.5331, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.03815505653619766, | |
| "rewards/margins": 0.8301553726196289, | |
| "rewards/rejected": -0.8683103322982788, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5024862601413242, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 2.91409685362137e-06, | |
| "logits/chosen": -0.5207866430282593, | |
| "logits/rejected": -0.5130727887153625, | |
| "logps/chosen": -278.73504638671875, | |
| "logps/rejected": -276.3402404785156, | |
| "loss": 0.5098, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.15987971425056458, | |
| "rewards/margins": 0.769777774810791, | |
| "rewards/rejected": -0.9296576380729675, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.5129547238942685, | |
| "grad_norm": 4.9375, | |
| "learning_rate": 2.8236713524386085e-06, | |
| "logits/chosen": -0.5813151597976685, | |
| "logits/rejected": -0.5460097193717957, | |
| "logps/chosen": -278.7942810058594, | |
| "logps/rejected": -256.22735595703125, | |
| "loss": 0.498, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.02987014129757881, | |
| "rewards/margins": 0.7739948034286499, | |
| "rewards/rejected": -0.8038649559020996, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.5234231876472127, | |
| "grad_norm": 5.59375, | |
| "learning_rate": 2.7328129695107205e-06, | |
| "logits/chosen": -0.4917908310890198, | |
| "logits/rejected": -0.5057145357131958, | |
| "logps/chosen": -264.3317565917969, | |
| "logps/rejected": -273.6480712890625, | |
| "loss": 0.5529, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.05067148059606552, | |
| "rewards/margins": 0.9685169458389282, | |
| "rewards/rejected": -1.0191885232925415, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5234231876472127, | |
| "eval_logits/chosen": -0.4290330708026886, | |
| "eval_logits/rejected": -0.3994762897491455, | |
| "eval_logps/chosen": -291.93804931640625, | |
| "eval_logps/rejected": -284.49481201171875, | |
| "eval_loss": 0.5259639620780945, | |
| "eval_rewards/accuracies": 0.7239999771118164, | |
| "eval_rewards/chosen": -0.2176663726568222, | |
| "eval_rewards/margins": 0.7860648036003113, | |
| "eval_rewards/rejected": -1.0037312507629395, | |
| "eval_runtime": 349.4608, | |
| "eval_samples_per_second": 5.723, | |
| "eval_steps_per_second": 0.358, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.533891651400157, | |
| "grad_norm": 4.75, | |
| "learning_rate": 2.641643219871597e-06, | |
| "logits/chosen": -0.5275259017944336, | |
| "logits/rejected": -0.49047979712486267, | |
| "logps/chosen": -313.4814147949219, | |
| "logps/rejected": -297.6827697753906, | |
| "loss": 0.5085, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.3123074471950531, | |
| "rewards/margins": 0.7204921841621399, | |
| "rewards/rejected": -1.0327996015548706, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.5443601151531012, | |
| "grad_norm": 6.90625, | |
| "learning_rate": 2.5502840349805074e-06, | |
| "logits/chosen": -0.5200182199478149, | |
| "logits/rejected": -0.5098154544830322, | |
| "logps/chosen": -310.73944091796875, | |
| "logps/rejected": -297.84326171875, | |
| "loss": 0.5471, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.353024423122406, | |
| "rewards/margins": 0.8153821229934692, | |
| "rewards/rejected": -1.1684064865112305, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.5548285789060455, | |
| "grad_norm": 5.03125, | |
| "learning_rate": 2.4588575996495797e-06, | |
| "logits/chosen": -0.47494420409202576, | |
| "logits/rejected": -0.48031002283096313, | |
| "logps/chosen": -272.15692138671875, | |
| "logps/rejected": -262.4747009277344, | |
| "loss": 0.5205, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.41294246912002563, | |
| "rewards/margins": 0.8126422166824341, | |
| "rewards/rejected": -1.225584626197815, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.5652970426589898, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 2.367486188632446e-06, | |
| "logits/chosen": -0.5167185068130493, | |
| "logits/rejected": -0.514365017414093, | |
| "logps/chosen": -285.01080322265625, | |
| "logps/rejected": -324.41302490234375, | |
| "loss": 0.5084, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.18025778234004974, | |
| "rewards/margins": 0.8430238962173462, | |
| "rewards/rejected": -1.0232816934585571, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.575765506411934, | |
| "grad_norm": 4.1875, | |
| "learning_rate": 2.276292003092593e-06, | |
| "logits/chosen": -0.5328477621078491, | |
| "logits/rejected": -0.514901340007782, | |
| "logps/chosen": -257.83953857421875, | |
| "logps/rejected": -265.4601135253906, | |
| "loss": 0.4871, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.24479135870933533, | |
| "rewards/margins": 0.8982963562011719, | |
| "rewards/rejected": -1.1430877447128296, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5862339701648783, | |
| "grad_norm": 4.875, | |
| "learning_rate": 2.1853970071701415e-06, | |
| "logits/chosen": -0.5274088978767395, | |
| "logits/rejected": -0.4985182285308838, | |
| "logps/chosen": -278.2223815917969, | |
| "logps/rejected": -279.3916931152344, | |
| "loss": 0.5206, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.12979185581207275, | |
| "rewards/margins": 0.8115113973617554, | |
| "rewards/rejected": -0.9413032531738281, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.5967024339178225, | |
| "grad_norm": 5.84375, | |
| "learning_rate": 2.0949227648656194e-06, | |
| "logits/chosen": -0.5541775822639465, | |
| "logits/rejected": -0.5307848453521729, | |
| "logps/chosen": -294.7757263183594, | |
| "logps/rejected": -261.8687438964844, | |
| "loss": 0.5405, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.2242761105298996, | |
| "rewards/margins": 0.8702728152275085, | |
| "rewards/rejected": -1.0945489406585693, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.6071708976707668, | |
| "grad_norm": 5.78125, | |
| "learning_rate": 2.00499027745888e-06, | |
| "logits/chosen": -0.5282704830169678, | |
| "logits/rejected": -0.5080757141113281, | |
| "logps/chosen": -299.1280212402344, | |
| "logps/rejected": -297.7095031738281, | |
| "loss": 0.5509, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.1550193578004837, | |
| "rewards/margins": 0.7501784563064575, | |
| "rewards/rejected": -0.9051979184150696, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.6176393614237111, | |
| "grad_norm": 5.65625, | |
| "learning_rate": 1.915719821680624e-06, | |
| "logits/chosen": -0.5544322729110718, | |
| "logits/rejected": -0.5079981088638306, | |
| "logps/chosen": -285.83575439453125, | |
| "logps/rejected": -281.488037109375, | |
| "loss": 0.5198, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": 0.1644931137561798, | |
| "rewards/margins": 0.8637059330940247, | |
| "rewards/rejected": -0.6992128491401672, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "grad_norm": 5.34375, | |
| "learning_rate": 1.8272307888529276e-06, | |
| "logits/chosen": -0.47845038771629333, | |
| "logits/rejected": -0.44052690267562866, | |
| "logps/chosen": -261.1684265136719, | |
| "logps/rejected": -279.1067810058594, | |
| "loss": 0.53, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.013519972562789917, | |
| "rewards/margins": 0.8593659400939941, | |
| "rewards/rejected": -0.8728858828544617, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "eval_logits/chosen": -0.4317292273044586, | |
| "eval_logits/rejected": -0.4027506709098816, | |
| "eval_logps/chosen": -290.4488830566406, | |
| "eval_logps/rejected": -283.0403137207031, | |
| "eval_loss": 0.5244117379188538, | |
| "eval_rewards/accuracies": 0.7200000286102295, | |
| "eval_rewards/chosen": -0.06874838471412659, | |
| "eval_rewards/margins": 0.7895320653915405, | |
| "eval_rewards/rejected": -0.8582804203033447, | |
| "eval_runtime": 349.48, | |
| "eval_samples_per_second": 5.723, | |
| "eval_steps_per_second": 0.358, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6385762889295996, | |
| "grad_norm": 4.625, | |
| "learning_rate": 1.739641525213929e-06, | |
| "logits/chosen": -0.510543942451477, | |
| "logits/rejected": -0.5007991194725037, | |
| "logps/chosen": -264.67572021484375, | |
| "logps/rejected": -270.05438232421875, | |
| "loss": 0.4957, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.04250194877386093, | |
| "rewards/margins": 0.8683522343635559, | |
| "rewards/rejected": -0.9108540415763855, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.6490447526825438, | |
| "grad_norm": 4.5625, | |
| "learning_rate": 1.6530691736402317e-06, | |
| "logits/chosen": -0.5225728750228882, | |
| "logits/rejected": -0.49780869483947754, | |
| "logps/chosen": -291.7324523925781, | |
| "logps/rejected": -280.8247985839844, | |
| "loss": 0.5072, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.15649111568927765, | |
| "rewards/margins": 0.8370261192321777, | |
| "rewards/rejected": -0.9935172200202942, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.6595132164354881, | |
| "grad_norm": 5.34375, | |
| "learning_rate": 1.5676295169786864e-06, | |
| "logits/chosen": -0.5341150164604187, | |
| "logits/rejected": -0.4966842532157898, | |
| "logps/chosen": -283.0377502441406, | |
| "logps/rejected": -271.4356994628906, | |
| "loss": 0.5294, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.335578590631485, | |
| "rewards/margins": 0.8176633715629578, | |
| "rewards/rejected": -1.1532419919967651, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.6699816801884323, | |
| "grad_norm": 5.25, | |
| "learning_rate": 1.4834368231970922e-06, | |
| "logits/chosen": -0.5592583417892456, | |
| "logits/rejected": -0.5109054446220398, | |
| "logps/chosen": -287.51068115234375, | |
| "logps/rejected": -272.8183288574219, | |
| "loss": 0.5, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.3240619897842407, | |
| "rewards/margins": 0.7095167636871338, | |
| "rewards/rejected": -1.0335787534713745, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.6804501439413766, | |
| "grad_norm": 5.28125, | |
| "learning_rate": 1.4006036925609245e-06, | |
| "logits/chosen": -0.5300047993659973, | |
| "logits/rejected": -0.49794530868530273, | |
| "logps/chosen": -300.1281433105469, | |
| "logps/rejected": -249.81936645507812, | |
| "loss": 0.5413, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.32234618067741394, | |
| "rewards/margins": 0.8122032880783081, | |
| "rewards/rejected": -1.1345494985580444, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6909186076943209, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 1.3192409070404582e-06, | |
| "logits/chosen": -0.550862193107605, | |
| "logits/rejected": -0.532629668712616, | |
| "logps/chosen": -303.74505615234375, | |
| "logps/rejected": -305.6463928222656, | |
| "loss": 0.5143, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.20528188347816467, | |
| "rewards/margins": 0.7565289735794067, | |
| "rewards/rejected": -0.9618108868598938, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.7013870714472651, | |
| "grad_norm": 6.59375, | |
| "learning_rate": 1.2394572821496953e-06, | |
| "logits/chosen": -0.5544015169143677, | |
| "logits/rejected": -0.5221393704414368, | |
| "logps/chosen": -277.2763977050781, | |
| "logps/rejected": -258.49676513671875, | |
| "loss": 0.5181, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.20379216969013214, | |
| "rewards/margins": 0.7780588269233704, | |
| "rewards/rejected": -0.9818509817123413, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.7118555352002094, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 1.1613595214152713e-06, | |
| "logits/chosen": -0.5822625160217285, | |
| "logits/rejected": -0.540503203868866, | |
| "logps/chosen": -286.454833984375, | |
| "logps/rejected": -275.0696105957031, | |
| "loss": 0.5043, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.32862305641174316, | |
| "rewards/margins": 0.7822999954223633, | |
| "rewards/rejected": -1.1109230518341064, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.7223239989531536, | |
| "grad_norm": 4.3125, | |
| "learning_rate": 1.0850520736699362e-06, | |
| "logits/chosen": -0.5231102705001831, | |
| "logits/rejected": -0.49847880005836487, | |
| "logps/chosen": -340.5146484375, | |
| "logps/rejected": -315.625, | |
| "loss": 0.5348, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.29148069024086, | |
| "rewards/margins": 0.8814651370048523, | |
| "rewards/rejected": -1.1729459762573242, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.7327924627060979, | |
| "grad_norm": 5.625, | |
| "learning_rate": 1.0106369933615043e-06, | |
| "logits/chosen": -0.5652047991752625, | |
| "logits/rejected": -0.5295856595039368, | |
| "logps/chosen": -315.13458251953125, | |
| "logps/rejected": -262.6495361328125, | |
| "loss": 0.5028, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.3104090690612793, | |
| "rewards/margins": 0.7127448320388794, | |
| "rewards/rejected": -1.0231537818908691, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7327924627060979, | |
| "eval_logits/chosen": -0.417924702167511, | |
| "eval_logits/rejected": -0.38744810223579407, | |
| "eval_logps/chosen": -293.1184387207031, | |
| "eval_logps/rejected": -285.8177185058594, | |
| "eval_loss": 0.5190439224243164, | |
| "eval_rewards/accuracies": 0.7319999933242798, | |
| "eval_rewards/chosen": -0.33567702770233154, | |
| "eval_rewards/margins": 0.8003441095352173, | |
| "eval_rewards/rejected": -1.1360211372375488, | |
| "eval_runtime": 349.5304, | |
| "eval_samples_per_second": 5.722, | |
| "eval_steps_per_second": 0.358, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7432609264590422, | |
| "grad_norm": 6.1875, | |
| "learning_rate": 9.382138040640714e-07, | |
| "logits/chosen": -0.5737341046333313, | |
| "logits/rejected": -0.5333853363990784, | |
| "logps/chosen": -264.9908752441406, | |
| "logps/rejected": -278.7557067871094, | |
| "loss": 0.5569, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.3085269331932068, | |
| "rewards/margins": 0.818738579750061, | |
| "rewards/rejected": -1.1272655725479126, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.7537293902119864, | |
| "grad_norm": 5.34375, | |
| "learning_rate": 8.678793653740633e-07, | |
| "logits/chosen": -0.4978073537349701, | |
| "logits/rejected": -0.49744582176208496, | |
| "logps/chosen": -263.29010009765625, | |
| "logps/rejected": -263.02618408203125, | |
| "loss": 0.5204, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.2587752938270569, | |
| "rewards/margins": 0.8211803436279297, | |
| "rewards/rejected": -1.0799555778503418, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.7641978539649307, | |
| "grad_norm": 4.75, | |
| "learning_rate": 7.997277433690984e-07, | |
| "logits/chosen": -0.5180292725563049, | |
| "logits/rejected": -0.47244367003440857, | |
| "logps/chosen": -301.2723693847656, | |
| "logps/rejected": -287.05584716796875, | |
| "loss": 0.5153, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.28081485629081726, | |
| "rewards/margins": 0.7739724516868591, | |
| "rewards/rejected": -1.054787278175354, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.7746663177178749, | |
| "grad_norm": 4.65625, | |
| "learning_rate": 7.338500848029603e-07, | |
| "logits/chosen": -0.47029241919517517, | |
| "logits/rejected": -0.48912668228149414, | |
| "logps/chosen": -290.9686584472656, | |
| "logps/rejected": -274.376953125, | |
| "loss": 0.4939, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.0831630602478981, | |
| "rewards/margins": 0.7876149415969849, | |
| "rewards/rejected": -0.8707780838012695, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.7851347814708192, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 6.70334495204884e-07, | |
| "logits/chosen": -0.5157723426818848, | |
| "logits/rejected": -0.49868693947792053, | |
| "logps/chosen": -323.7619323730469, | |
| "logps/rejected": -285.82916259765625, | |
| "loss": 0.5023, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.11948978900909424, | |
| "rewards/margins": 0.8135945200920105, | |
| "rewards/rejected": -0.9330843687057495, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7956032452237635, | |
| "grad_norm": 5.46875, | |
| "learning_rate": 6.092659210462232e-07, | |
| "logits/chosen": -0.5377079248428345, | |
| "logits/rejected": -0.5193135142326355, | |
| "logps/chosen": -268.98455810546875, | |
| "logps/rejected": -268.32025146484375, | |
| "loss": 0.5455, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.2502976357936859, | |
| "rewards/margins": 0.698966920375824, | |
| "rewards/rejected": -0.949264645576477, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.8060717089767077, | |
| "grad_norm": 5.28125, | |
| "learning_rate": 5.507260361320738e-07, | |
| "logits/chosen": -0.5280566215515137, | |
| "logits/rejected": -0.5302770137786865, | |
| "logps/chosen": -284.72637939453125, | |
| "logps/rejected": -279.0828552246094, | |
| "loss": 0.5093, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.12712079286575317, | |
| "rewards/margins": 0.6862561106681824, | |
| "rewards/rejected": -0.8133770227432251, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.816540172729652, | |
| "grad_norm": 5.21875, | |
| "learning_rate": 4.947931323697983e-07, | |
| "logits/chosen": -0.5104556083679199, | |
| "logits/rejected": -0.48463624715805054, | |
| "logps/chosen": -286.16925048828125, | |
| "logps/rejected": -279.46295166015625, | |
| "loss": 0.5278, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.2269764244556427, | |
| "rewards/margins": 0.6587584018707275, | |
| "rewards/rejected": -0.8857349157333374, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.8270086364825961, | |
| "grad_norm": 5.375, | |
| "learning_rate": 4.4154201506053985e-07, | |
| "logits/chosen": -0.5380119681358337, | |
| "logits/rejected": -0.5271115303039551, | |
| "logps/chosen": -300.27313232421875, | |
| "logps/rejected": -263.89984130859375, | |
| "loss": 0.511, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -0.15483184158802032, | |
| "rewards/margins": 0.8044347763061523, | |
| "rewards/rejected": -0.9592665433883667, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "grad_norm": 4.84375, | |
| "learning_rate": 3.910439028537638e-07, | |
| "logits/chosen": -0.5491563677787781, | |
| "logits/rejected": -0.5051991939544678, | |
| "logps/chosen": -347.644287109375, | |
| "logps/rejected": -302.0996398925781, | |
| "loss": 0.5347, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.08171938359737396, | |
| "rewards/margins": 0.7037029266357422, | |
| "rewards/rejected": -0.7854223251342773, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "eval_logits/chosen": -0.42248398065567017, | |
| "eval_logits/rejected": -0.39241039752960205, | |
| "eval_logps/chosen": -291.1650085449219, | |
| "eval_logps/rejected": -283.87603759765625, | |
| "eval_loss": 0.5191403031349182, | |
| "eval_rewards/accuracies": 0.7319999933242798, | |
| "eval_rewards/chosen": -0.14035792648792267, | |
| "eval_rewards/margins": 0.8014954924583435, | |
| "eval_rewards/rejected": -0.9418535232543945, | |
| "eval_runtime": 349.5564, | |
| "eval_samples_per_second": 5.722, | |
| "eval_steps_per_second": 0.358, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8479455639884846, | |
| "grad_norm": 4.6875, | |
| "learning_rate": 3.4336633249862084e-07, | |
| "logits/chosen": -0.5698053240776062, | |
| "logits/rejected": -0.5044723153114319, | |
| "logps/chosen": -319.55126953125, | |
| "logps/rejected": -288.7781066894531, | |
| "loss": 0.5021, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.17804962396621704, | |
| "rewards/margins": 0.8044875860214233, | |
| "rewards/rejected": -0.9825371503829956, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.8584140277414289, | |
| "grad_norm": 5.25, | |
| "learning_rate": 2.98573068519539e-07, | |
| "logits/chosen": -0.5468612909317017, | |
| "logits/rejected": -0.5319810509681702, | |
| "logps/chosen": -307.5137939453125, | |
| "logps/rejected": -293.56561279296875, | |
| "loss": 0.5358, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.15299567580223083, | |
| "rewards/margins": 0.7571539282798767, | |
| "rewards/rejected": -0.9101495742797852, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.8688824914943732, | |
| "grad_norm": 5.375, | |
| "learning_rate": 2.5672401793681854e-07, | |
| "logits/chosen": -0.565108597278595, | |
| "logits/rejected": -0.5492611527442932, | |
| "logps/chosen": -274.61260986328125, | |
| "logps/rejected": -269.25439453125, | |
| "loss": 0.5048, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.17146927118301392, | |
| "rewards/margins": 0.8092424273490906, | |
| "rewards/rejected": -0.9807117581367493, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.8793509552473174, | |
| "grad_norm": 5.1875, | |
| "learning_rate": 2.178751501463036e-07, | |
| "logits/chosen": -0.533309280872345, | |
| "logits/rejected": -0.5187536478042603, | |
| "logps/chosen": -313.8269958496094, | |
| "logps/rejected": -307.17987060546875, | |
| "loss": 0.517, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.20717601478099823, | |
| "rewards/margins": 0.7723677158355713, | |
| "rewards/rejected": -0.9795438051223755, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.8898194190002617, | |
| "grad_norm": 5.4375, | |
| "learning_rate": 1.820784220652766e-07, | |
| "logits/chosen": -0.5832973122596741, | |
| "logits/rejected": -0.5479104518890381, | |
| "logps/chosen": -345.9014587402344, | |
| "logps/rejected": -279.53985595703125, | |
| "loss": 0.5083, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.0073674083687365055, | |
| "rewards/margins": 0.875013530254364, | |
| "rewards/rejected": -0.8823810815811157, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9002878827532059, | |
| "grad_norm": 5.375, | |
| "learning_rate": 1.4938170864468636e-07, | |
| "logits/chosen": -0.5060287117958069, | |
| "logits/rejected": -0.4896390438079834, | |
| "logps/chosen": -289.37042236328125, | |
| "logps/rejected": -268.93487548828125, | |
| "loss": 0.4806, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.11767210066318512, | |
| "rewards/margins": 0.9038281440734863, | |
| "rewards/rejected": -1.0215001106262207, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.9107563465061502, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 1.1982873884064466e-07, | |
| "logits/chosen": -0.4754942059516907, | |
| "logits/rejected": -0.4765293598175049, | |
| "logps/chosen": -287.2554931640625, | |
| "logps/rejected": -276.70794677734375, | |
| "loss": 0.5263, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.16617272794246674, | |
| "rewards/margins": 0.6353666186332703, | |
| "rewards/rejected": -0.801539421081543, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.9212248102590945, | |
| "grad_norm": 5.75, | |
| "learning_rate": 9.345903713082305e-08, | |
| "logits/chosen": -0.5580819845199585, | |
| "logits/rejected": -0.5499614477157593, | |
| "logps/chosen": -315.15478515625, | |
| "logps/rejected": -280.8697814941406, | |
| "loss": 0.5463, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.1284041702747345, | |
| "rewards/margins": 0.6124891042709351, | |
| "rewards/rejected": -0.7408932447433472, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.9316932740120387, | |
| "grad_norm": 5.125, | |
| "learning_rate": 7.030787065396866e-08, | |
| "logits/chosen": -0.5307958722114563, | |
| "logits/rejected": -0.49479198455810547, | |
| "logps/chosen": -318.7244567871094, | |
| "logps/rejected": -293.06219482421875, | |
| "loss": 0.5198, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.11201741546392441, | |
| "rewards/margins": 0.724387526512146, | |
| "rewards/rejected": -0.8364049792289734, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.942161737764983, | |
| "grad_norm": 4.96875, | |
| "learning_rate": 5.0406202043228604e-08, | |
| "logits/chosen": -0.541701078414917, | |
| "logits/rejected": -0.5375146865844727, | |
| "logps/chosen": -333.6339111328125, | |
| "logps/rejected": -274.9059753417969, | |
| "loss": 0.4783, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.0002611428499221802, | |
| "rewards/margins": 0.9774513244628906, | |
| "rewards/rejected": -0.9777124524116516, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.942161737764983, | |
| "eval_logits/chosen": -0.4219192862510681, | |
| "eval_logits/rejected": -0.391748309135437, | |
| "eval_logps/chosen": -291.15997314453125, | |
| "eval_logps/rejected": -283.916259765625, | |
| "eval_loss": 0.5189567804336548, | |
| "eval_rewards/accuracies": 0.7260000109672546, | |
| "eval_rewards/chosen": -0.13985459506511688, | |
| "eval_rewards/margins": 0.8060198426246643, | |
| "eval_rewards/rejected": -0.9458745121955872, | |
| "eval_runtime": 349.4658, | |
| "eval_samples_per_second": 5.723, | |
| "eval_steps_per_second": 0.358, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9526302015179272, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 3.378064801637687e-08, | |
| "logits/chosen": -0.5818526148796082, | |
| "logits/rejected": -0.5350463390350342, | |
| "logps/chosen": -314.6712341308594, | |
| "logps/rejected": -313.63116455078125, | |
| "loss": 0.5208, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.047819893807172775, | |
| "rewards/margins": 0.7616057991981506, | |
| "rewards/rejected": -0.8094256520271301, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.9630986652708715, | |
| "grad_norm": 4.6875, | |
| "learning_rate": 2.0453443778310766e-08, | |
| "logits/chosen": -0.5121539831161499, | |
| "logits/rejected": -0.48242586851119995, | |
| "logps/chosen": -329.2786865234375, | |
| "logps/rejected": -304.5957336425781, | |
| "loss": 0.5356, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.06458105146884918, | |
| "rewards/margins": 0.869833767414093, | |
| "rewards/rejected": -0.9344147443771362, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.9735671290238157, | |
| "grad_norm": 4.78125, | |
| "learning_rate": 1.0442413283435759e-08, | |
| "logits/chosen": -0.509577751159668, | |
| "logits/rejected": -0.47219425439834595, | |
| "logps/chosen": -317.72943115234375, | |
| "logps/rejected": -278.4623718261719, | |
| "loss": 0.5143, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.05142117291688919, | |
| "rewards/margins": 1.1160078048706055, | |
| "rewards/rejected": -1.167428970336914, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.98403559277676, | |
| "grad_norm": 4.1875, | |
| "learning_rate": 3.760945397705828e-09, | |
| "logits/chosen": -0.5354495048522949, | |
| "logits/rejected": -0.4856652617454529, | |
| "logps/chosen": -290.7127380371094, | |
| "logps/rejected": -260.788330078125, | |
| "loss": 0.487, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.03208203241229057, | |
| "rewards/margins": 1.0025131702423096, | |
| "rewards/rejected": -0.9704310297966003, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.9945040565297043, | |
| "grad_norm": 4.6875, | |
| "learning_rate": 4.1797599220405605e-10, | |
| "logits/chosen": -0.5512427687644958, | |
| "logits/rejected": -0.5348908305168152, | |
| "logps/chosen": -295.21319580078125, | |
| "logps/rejected": -278.82257080078125, | |
| "loss": 0.5053, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.0824163556098938, | |
| "rewards/margins": 0.8457223176956177, | |
| "rewards/rejected": -0.9281387329101562, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.9997382884061764, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 0.543355998318857, | |
| "train_runtime": 19333.2554, | |
| "train_samples_per_second": 3.162, | |
| "train_steps_per_second": 0.049 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 955, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100000, | |
| "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": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |