| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9997382884061764, | |
| "eval_steps": 100, | |
| "global_step": 1910, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 1.75, | |
| "learning_rate": 2.617801047120419e-08, | |
| "logits/chosen": -2.2731704711914062, | |
| "logits/rejected": -2.1761367321014404, | |
| "logps/chosen": -360.11749267578125, | |
| "logps/rejected": -295.71942138671875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 1.6328125, | |
| "learning_rate": 2.617801047120419e-07, | |
| "logits/chosen": -2.259983777999878, | |
| "logits/rejected": -2.1553776264190674, | |
| "logps/chosen": -303.7694091796875, | |
| "logps/rejected": -241.65496826171875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.4097222089767456, | |
| "rewards/chosen": 0.0003990587720181793, | |
| "rewards/margins": 0.00018596058362163603, | |
| "rewards/rejected": 0.00021309818839654326, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 1.5078125, | |
| "learning_rate": 5.235602094240838e-07, | |
| "logits/chosen": -2.2269904613494873, | |
| "logits/rejected": -2.1716256141662598, | |
| "logps/chosen": -262.0898132324219, | |
| "logps/rejected": -249.70352172851562, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.001905500190332532, | |
| "rewards/margins": 0.00033334543695673347, | |
| "rewards/rejected": 0.0015721546951681376, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 1.609375, | |
| "learning_rate": 7.853403141361258e-07, | |
| "logits/chosen": -2.1317379474639893, | |
| "logits/rejected": -2.082672595977783, | |
| "logps/chosen": -277.956298828125, | |
| "logps/rejected": -239.60293579101562, | |
| "loss": 0.692, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": 0.005530746188014746, | |
| "rewards/margins": 0.0024719357024878263, | |
| "rewards/rejected": 0.003058810718357563, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 1.6484375, | |
| "learning_rate": 1.0471204188481676e-06, | |
| "logits/chosen": -2.241527795791626, | |
| "logits/rejected": -2.1675925254821777, | |
| "logps/chosen": -267.1982727050781, | |
| "logps/rejected": -270.7894287109375, | |
| "loss": 0.6901, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.010646065697073936, | |
| "rewards/margins": 0.007375557906925678, | |
| "rewards/rejected": 0.003270507324486971, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 1.578125, | |
| "learning_rate": 1.3089005235602096e-06, | |
| "logits/chosen": -2.269123077392578, | |
| "logits/rejected": -2.1638808250427246, | |
| "logps/chosen": -285.50164794921875, | |
| "logps/rejected": -254.6826629638672, | |
| "loss": 0.6881, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.016856886446475983, | |
| "rewards/margins": 0.012282918207347393, | |
| "rewards/rejected": 0.004573967307806015, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 1.5625, | |
| "learning_rate": 1.5706806282722515e-06, | |
| "logits/chosen": -2.272505283355713, | |
| "logits/rejected": -2.122178316116333, | |
| "logps/chosen": -310.4421081542969, | |
| "logps/rejected": -260.3993835449219, | |
| "loss": 0.6838, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.027886558324098587, | |
| "rewards/margins": 0.01857396773993969, | |
| "rewards/rejected": 0.009312589652836323, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 1.71875, | |
| "learning_rate": 1.8324607329842933e-06, | |
| "logits/chosen": -2.2799620628356934, | |
| "logits/rejected": -2.126312494277954, | |
| "logps/chosen": -266.8229064941406, | |
| "logps/rejected": -234.4290313720703, | |
| "loss": 0.68, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.03913799300789833, | |
| "rewards/margins": 0.0319330058991909, | |
| "rewards/rejected": 0.0072049833834171295, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 1.859375, | |
| "learning_rate": 2.094240837696335e-06, | |
| "logits/chosen": -2.2087223529815674, | |
| "logits/rejected": -2.1472485065460205, | |
| "logps/chosen": -279.0526123046875, | |
| "logps/rejected": -270.34185791015625, | |
| "loss": 0.6778, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.05015086010098457, | |
| "rewards/margins": 0.03647767752408981, | |
| "rewards/rejected": 0.013673178851604462, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 1.75, | |
| "learning_rate": 2.356020942408377e-06, | |
| "logits/chosen": -2.2452147006988525, | |
| "logits/rejected": -2.178189992904663, | |
| "logps/chosen": -250.22409057617188, | |
| "logps/rejected": -241.08779907226562, | |
| "loss": 0.676, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.0462195985019207, | |
| "rewards/margins": 0.03395666554570198, | |
| "rewards/rejected": 0.012262934818863869, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 1.9296875, | |
| "learning_rate": 2.617801047120419e-06, | |
| "logits/chosen": -2.2681682109832764, | |
| "logits/rejected": -2.121201515197754, | |
| "logps/chosen": -244.41537475585938, | |
| "logps/rejected": -206.2124481201172, | |
| "loss": 0.6641, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.027621516957879066, | |
| "rewards/margins": 0.07613887637853622, | |
| "rewards/rejected": -0.048517368733882904, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "eval_logits/chosen": -2.136847972869873, | |
| "eval_logits/rejected": -2.0436248779296875, | |
| "eval_logps/chosen": -271.1658935546875, | |
| "eval_logps/rejected": -254.5337371826172, | |
| "eval_loss": 0.6636302471160889, | |
| "eval_rewards/accuracies": 0.6899999976158142, | |
| "eval_rewards/chosen": 0.00537072541192174, | |
| "eval_rewards/margins": 0.07350712269544601, | |
| "eval_rewards/rejected": -0.06813640147447586, | |
| "eval_runtime": 206.3535, | |
| "eval_samples_per_second": 9.692, | |
| "eval_steps_per_second": 0.606, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 2.125, | |
| "learning_rate": 2.8795811518324613e-06, | |
| "logits/chosen": -2.2474560737609863, | |
| "logits/rejected": -2.1244096755981445, | |
| "logps/chosen": -264.55267333984375, | |
| "logps/rejected": -220.3800506591797, | |
| "loss": 0.6544, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.01394604705274105, | |
| "rewards/margins": 0.09943284094333649, | |
| "rewards/rejected": -0.08548679202795029, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 2.21875, | |
| "learning_rate": 3.141361256544503e-06, | |
| "logits/chosen": -2.269590139389038, | |
| "logits/rejected": -2.1833536624908447, | |
| "logps/chosen": -320.98858642578125, | |
| "logps/rejected": -295.50213623046875, | |
| "loss": 0.6587, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.03138069808483124, | |
| "rewards/margins": 0.07552894949913025, | |
| "rewards/rejected": -0.10690964758396149, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 2.734375, | |
| "learning_rate": 3.403141361256545e-06, | |
| "logits/chosen": -2.2066664695739746, | |
| "logits/rejected": -2.1157479286193848, | |
| "logps/chosen": -286.4569396972656, | |
| "logps/rejected": -271.8417053222656, | |
| "loss": 0.6471, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.026338139548897743, | |
| "rewards/margins": 0.12607838213443756, | |
| "rewards/rejected": -0.09974025189876556, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 3.296875, | |
| "learning_rate": 3.6649214659685865e-06, | |
| "logits/chosen": -2.1494216918945312, | |
| "logits/rejected": -2.084484815597534, | |
| "logps/chosen": -266.2762756347656, | |
| "logps/rejected": -270.8194580078125, | |
| "loss": 0.6385, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.075466588139534, | |
| "rewards/margins": 0.14162591099739075, | |
| "rewards/rejected": -0.21709248423576355, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 3.0625, | |
| "learning_rate": 3.926701570680629e-06, | |
| "logits/chosen": -2.189535617828369, | |
| "logits/rejected": -2.0706353187561035, | |
| "logps/chosen": -291.97283935546875, | |
| "logps/rejected": -291.4541015625, | |
| "loss": 0.6269, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.13837790489196777, | |
| "rewards/margins": 0.16128641366958618, | |
| "rewards/rejected": -0.29966431856155396, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 3.9375, | |
| "learning_rate": 4.18848167539267e-06, | |
| "logits/chosen": -2.3002448081970215, | |
| "logits/rejected": -2.1360554695129395, | |
| "logps/chosen": -293.7611083984375, | |
| "logps/rejected": -274.07025146484375, | |
| "loss": 0.6296, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.17077429592609406, | |
| "rewards/margins": 0.15933959186077118, | |
| "rewards/rejected": -0.3301139175891876, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 3.71875, | |
| "learning_rate": 4.450261780104713e-06, | |
| "logits/chosen": -2.1448490619659424, | |
| "logits/rejected": -2.123392105102539, | |
| "logps/chosen": -263.1458740234375, | |
| "logps/rejected": -278.94635009765625, | |
| "loss": 0.6264, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.16839735209941864, | |
| "rewards/margins": 0.17006024718284607, | |
| "rewards/rejected": -0.3384575843811035, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 2.78125, | |
| "learning_rate": 4.712041884816754e-06, | |
| "logits/chosen": -2.2578933238983154, | |
| "logits/rejected": -2.1124982833862305, | |
| "logps/chosen": -318.94476318359375, | |
| "logps/rejected": -312.0294494628906, | |
| "loss": 0.628, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.23213641345500946, | |
| "rewards/margins": 0.20630350708961487, | |
| "rewards/rejected": -0.43843990564346313, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 3.53125, | |
| "learning_rate": 4.9738219895287965e-06, | |
| "logits/chosen": -2.2287371158599854, | |
| "logits/rejected": -2.1092655658721924, | |
| "logps/chosen": -303.0882873535156, | |
| "logps/rejected": -299.7857360839844, | |
| "loss": 0.6017, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.2376522272825241, | |
| "rewards/margins": 0.2720206081867218, | |
| "rewards/rejected": -0.5096728205680847, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 3.34375, | |
| "learning_rate": 4.999661831436499e-06, | |
| "logits/chosen": -2.1837334632873535, | |
| "logits/rejected": -2.094944477081299, | |
| "logps/chosen": -285.54425048828125, | |
| "logps/rejected": -296.94720458984375, | |
| "loss": 0.6105, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.1814960539340973, | |
| "rewards/margins": 0.24684414267539978, | |
| "rewards/rejected": -0.4283401370048523, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "eval_logits/chosen": -2.0918540954589844, | |
| "eval_logits/rejected": -2.003007173538208, | |
| "eval_logps/chosen": -304.0613098144531, | |
| "eval_logps/rejected": -307.0966796875, | |
| "eval_loss": 0.607542872428894, | |
| "eval_rewards/accuracies": 0.6890000104904175, | |
| "eval_rewards/chosen": -0.32358381152153015, | |
| "eval_rewards/margins": 0.2701820731163025, | |
| "eval_rewards/rejected": -0.5937658548355103, | |
| "eval_runtime": 205.7166, | |
| "eval_samples_per_second": 9.722, | |
| "eval_steps_per_second": 0.608, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 4.9984929711403395e-06, | |
| "logits/chosen": -2.132068395614624, | |
| "logits/rejected": -2.023831844329834, | |
| "logps/chosen": -262.7835693359375, | |
| "logps/rejected": -267.73046875, | |
| "loss": 0.6256, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.24357381463050842, | |
| "rewards/margins": 0.18603582680225372, | |
| "rewards/rejected": -0.4296096861362457, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 3.671875, | |
| "learning_rate": 4.996489634487865e-06, | |
| "logits/chosen": -2.1585371494293213, | |
| "logits/rejected": -2.0781378746032715, | |
| "logps/chosen": -319.03131103515625, | |
| "logps/rejected": -319.93634033203125, | |
| "loss": 0.5928, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.40161317586898804, | |
| "rewards/margins": 0.3085588216781616, | |
| "rewards/rejected": -0.7101720571517944, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 5.65625, | |
| "learning_rate": 4.9936524905772466e-06, | |
| "logits/chosen": -2.099738359451294, | |
| "logits/rejected": -2.0198581218719482, | |
| "logps/chosen": -287.83807373046875, | |
| "logps/rejected": -313.92620849609375, | |
| "loss": 0.5783, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.435781329870224, | |
| "rewards/margins": 0.37423354387283325, | |
| "rewards/rejected": -0.8100149035453796, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 4.28125, | |
| "learning_rate": 4.9899824869915e-06, | |
| "logits/chosen": -2.1551759243011475, | |
| "logits/rejected": -2.0810234546661377, | |
| "logps/chosen": -326.7232360839844, | |
| "logps/rejected": -348.8532409667969, | |
| "loss": 0.5916, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.7081143260002136, | |
| "rewards/margins": 0.3044819235801697, | |
| "rewards/rejected": -1.0125962495803833, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 3.203125, | |
| "learning_rate": 4.985480849482012e-06, | |
| "logits/chosen": -2.1337616443634033, | |
| "logits/rejected": -2.0014939308166504, | |
| "logps/chosen": -374.9393615722656, | |
| "logps/rejected": -349.9031066894531, | |
| "loss": 0.5716, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.6511236429214478, | |
| "rewards/margins": 0.35167860984802246, | |
| "rewards/rejected": -1.0028022527694702, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 3.96875, | |
| "learning_rate": 4.980149081559142e-06, | |
| "logits/chosen": -2.1727969646453857, | |
| "logits/rejected": -2.108790874481201, | |
| "logps/chosen": -314.8628234863281, | |
| "logps/rejected": -321.34710693359375, | |
| "loss": 0.5922, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.4089882969856262, | |
| "rewards/margins": 0.38155826926231384, | |
| "rewards/rejected": -0.7905465960502625, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 3.421875, | |
| "learning_rate": 4.9739889639900655e-06, | |
| "logits/chosen": -2.1201374530792236, | |
| "logits/rejected": -2.074112892150879, | |
| "logps/chosen": -317.30908203125, | |
| "logps/rejected": -349.3733825683594, | |
| "loss": 0.5602, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.4397730827331543, | |
| "rewards/margins": 0.4018850326538086, | |
| "rewards/rejected": -0.8416581153869629, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 3.3125, | |
| "learning_rate": 4.967002554204009e-06, | |
| "logits/chosen": -2.1125736236572266, | |
| "logits/rejected": -2.088714122772217, | |
| "logps/chosen": -330.7603759765625, | |
| "logps/rejected": -371.1779479980469, | |
| "loss": 0.5972, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.5998188853263855, | |
| "rewards/margins": 0.33086952567100525, | |
| "rewards/rejected": -0.9306885004043579, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 3.765625, | |
| "learning_rate": 4.959192185605089e-06, | |
| "logits/chosen": -2.266671657562256, | |
| "logits/rejected": -2.168347120285034, | |
| "logps/chosen": -372.093017578125, | |
| "logps/rejected": -373.88555908203125, | |
| "loss": 0.5818, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.5426467657089233, | |
| "rewards/margins": 0.4017227590084076, | |
| "rewards/rejected": -0.9443694353103638, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 7.375, | |
| "learning_rate": 4.950560466792969e-06, | |
| "logits/chosen": -2.1483025550842285, | |
| "logits/rejected": -2.048117160797119, | |
| "logps/chosen": -341.9796447753906, | |
| "logps/rejected": -351.8575744628906, | |
| "loss": 0.5883, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.6638094782829285, | |
| "rewards/margins": 0.44865164160728455, | |
| "rewards/rejected": -1.1124610900878906, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "eval_logits/chosen": -2.076050281524658, | |
| "eval_logits/rejected": -1.9913941621780396, | |
| "eval_logps/chosen": -342.9188232421875, | |
| "eval_logps/rejected": -360.5767517089844, | |
| "eval_loss": 0.5817497372627258, | |
| "eval_rewards/accuracies": 0.7020000219345093, | |
| "eval_rewards/chosen": -0.7121586203575134, | |
| "eval_rewards/margins": 0.4164075553417206, | |
| "eval_rewards/rejected": -1.1285661458969116, | |
| "eval_runtime": 205.7512, | |
| "eval_samples_per_second": 9.72, | |
| "eval_steps_per_second": 0.608, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 4.9411102806916185e-06, | |
| "logits/chosen": -2.1432528495788574, | |
| "logits/rejected": -2.0720839500427246, | |
| "logps/chosen": -332.2557067871094, | |
| "logps/rejected": -340.0525817871094, | |
| "loss": 0.5635, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.576932966709137, | |
| "rewards/margins": 0.40913018584251404, | |
| "rewards/rejected": -0.9860631823539734, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 4.4375, | |
| "learning_rate": 4.930844783586424e-06, | |
| "logits/chosen": -2.1997363567352295, | |
| "logits/rejected": -2.0350875854492188, | |
| "logps/chosen": -320.69720458984375, | |
| "logps/rejected": -334.16217041015625, | |
| "loss": 0.5567, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.5008949041366577, | |
| "rewards/margins": 0.5016661882400513, | |
| "rewards/rejected": -1.002561092376709, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 4.71875, | |
| "learning_rate": 4.919767404070033e-06, | |
| "logits/chosen": -2.1636505126953125, | |
| "logits/rejected": -2.0376198291778564, | |
| "logps/chosen": -330.29254150390625, | |
| "logps/rejected": -358.37811279296875, | |
| "loss": 0.5787, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.6509741544723511, | |
| "rewards/margins": 0.4970785677433014, | |
| "rewards/rejected": -1.1480529308319092, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 8.5, | |
| "learning_rate": 4.907881841897216e-06, | |
| "logits/chosen": -2.1893808841705322, | |
| "logits/rejected": -2.115548849105835, | |
| "logps/chosen": -376.6805114746094, | |
| "logps/rejected": -387.1194763183594, | |
| "loss": 0.5545, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.8470619320869446, | |
| "rewards/margins": 0.41202712059020996, | |
| "rewards/rejected": -1.2590891122817993, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 3.53125, | |
| "learning_rate": 4.89519206674919e-06, | |
| "logits/chosen": -2.1224799156188965, | |
| "logits/rejected": -2.1183266639709473, | |
| "logps/chosen": -311.69000244140625, | |
| "logps/rejected": -391.626220703125, | |
| "loss": 0.529, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.6741515398025513, | |
| "rewards/margins": 0.5650461912155151, | |
| "rewards/rejected": -1.2391977310180664, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 4.9375, | |
| "learning_rate": 4.881702316907769e-06, | |
| "logits/chosen": -2.1981089115142822, | |
| "logits/rejected": -2.0802032947540283, | |
| "logps/chosen": -349.52618408203125, | |
| "logps/rejected": -364.61285400390625, | |
| "loss": 0.5662, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.5019327998161316, | |
| "rewards/margins": 0.4679330885410309, | |
| "rewards/rejected": -0.9698659181594849, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 6.4375, | |
| "learning_rate": 4.86741709783982e-06, | |
| "logits/chosen": -2.064434051513672, | |
| "logits/rejected": -1.957092523574829, | |
| "logps/chosen": -323.6076965332031, | |
| "logps/rejected": -344.89971923828125, | |
| "loss": 0.5741, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.6532014608383179, | |
| "rewards/margins": 0.5228798985481262, | |
| "rewards/rejected": -1.1760812997817993, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 6.5, | |
| "learning_rate": 4.852341180692471e-06, | |
| "logits/chosen": -2.0200791358947754, | |
| "logits/rejected": -1.9938948154449463, | |
| "logps/chosen": -333.9423522949219, | |
| "logps/rejected": -399.6837463378906, | |
| "loss": 0.548, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.8045186996459961, | |
| "rewards/margins": 0.47954291105270386, | |
| "rewards/rejected": -1.2840616703033447, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 3.65625, | |
| "learning_rate": 4.836479600699579e-06, | |
| "logits/chosen": -2.0388081073760986, | |
| "logits/rejected": -1.9344475269317627, | |
| "logps/chosen": -365.44732666015625, | |
| "logps/rejected": -364.44549560546875, | |
| "loss": 0.5859, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.677007794380188, | |
| "rewards/margins": 0.46105116605758667, | |
| "rewards/rejected": -1.1380589008331299, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 4.819837655500014e-06, | |
| "logits/chosen": -2.023102283477783, | |
| "logits/rejected": -1.9553874731063843, | |
| "logps/chosen": -306.1051330566406, | |
| "logps/rejected": -330.5825500488281, | |
| "loss": 0.5651, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.5321759581565857, | |
| "rewards/margins": 0.4713428020477295, | |
| "rewards/rejected": -1.00351881980896, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": -1.982036828994751, | |
| "eval_logits/rejected": -1.9001150131225586, | |
| "eval_logps/chosen": -350.7093200683594, | |
| "eval_logps/rejected": -376.6873779296875, | |
| "eval_loss": 0.566522479057312, | |
| "eval_rewards/accuracies": 0.7250000238418579, | |
| "eval_rewards/chosen": -0.7900638580322266, | |
| "eval_rewards/margins": 0.4996088147163391, | |
| "eval_rewards/rejected": -1.289672613143921, | |
| "eval_runtime": 205.8214, | |
| "eval_samples_per_second": 9.717, | |
| "eval_steps_per_second": 0.607, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 4.71875, | |
| "learning_rate": 4.802420903368286e-06, | |
| "logits/chosen": -2.045619487762451, | |
| "logits/rejected": -2.054408550262451, | |
| "logps/chosen": -364.0269775390625, | |
| "logps/rejected": -412.7276306152344, | |
| "loss": 0.5841, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -1.106697678565979, | |
| "rewards/margins": 0.3950948119163513, | |
| "rewards/rejected": -1.501792311668396, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 5.34375, | |
| "learning_rate": 4.784235161358124e-06, | |
| "logits/chosen": -2.0790481567382812, | |
| "logits/rejected": -2.003018617630005, | |
| "logps/chosen": -399.572998046875, | |
| "logps/rejected": -433.9266662597656, | |
| "loss": 0.583, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.2603890895843506, | |
| "rewards/margins": 0.41771477460861206, | |
| "rewards/rejected": -1.6781038045883179, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 5.75, | |
| "learning_rate": 4.765286503359632e-06, | |
| "logits/chosen": -2.0433547496795654, | |
| "logits/rejected": -1.968266248703003, | |
| "logps/chosen": -357.1161193847656, | |
| "logps/rejected": -381.88507080078125, | |
| "loss": 0.5707, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.8924205899238586, | |
| "rewards/margins": 0.43997693061828613, | |
| "rewards/rejected": -1.3323975801467896, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 4.1875, | |
| "learning_rate": 4.745581258070654e-06, | |
| "logits/chosen": -1.973879098892212, | |
| "logits/rejected": -1.9118854999542236, | |
| "logps/chosen": -329.8348693847656, | |
| "logps/rejected": -366.04461669921875, | |
| "loss": 0.5512, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.562414288520813, | |
| "rewards/margins": 0.5114336013793945, | |
| "rewards/rejected": -1.073848009109497, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 5.9375, | |
| "learning_rate": 4.725126006883047e-06, | |
| "logits/chosen": -1.8499130010604858, | |
| "logits/rejected": -1.827254056930542, | |
| "logps/chosen": -307.35150146484375, | |
| "logps/rejected": -369.239990234375, | |
| "loss": 0.5526, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.7073386311531067, | |
| "rewards/margins": 0.5170890092849731, | |
| "rewards/rejected": -1.224427580833435, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 4.625, | |
| "learning_rate": 4.70392758168454e-06, | |
| "logits/chosen": -1.9882104396820068, | |
| "logits/rejected": -1.8469337224960327, | |
| "logps/chosen": -375.27935791015625, | |
| "logps/rejected": -379.7239685058594, | |
| "loss": 0.5734, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.9050993919372559, | |
| "rewards/margins": 0.47090214490890503, | |
| "rewards/rejected": -1.3760015964508057, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 4.03125, | |
| "learning_rate": 4.68199306257695e-06, | |
| "logits/chosen": -1.9672870635986328, | |
| "logits/rejected": -1.9157249927520752, | |
| "logps/chosen": -396.733154296875, | |
| "logps/rejected": -421.2018127441406, | |
| "loss": 0.5354, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.9322503805160522, | |
| "rewards/margins": 0.5409501791000366, | |
| "rewards/rejected": -1.4732005596160889, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 5.1875, | |
| "learning_rate": 4.659329775511478e-06, | |
| "logits/chosen": -1.973670244216919, | |
| "logits/rejected": -1.9386718273162842, | |
| "logps/chosen": -343.16387939453125, | |
| "logps/rejected": -386.0146484375, | |
| "loss": 0.556, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.9054447412490845, | |
| "rewards/margins": 0.559832751750946, | |
| "rewards/rejected": -1.4652774333953857, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 5.0, | |
| "learning_rate": 4.635945289841902e-06, | |
| "logits/chosen": -1.9971214532852173, | |
| "logits/rejected": -1.9401931762695312, | |
| "logps/chosen": -368.4125061035156, | |
| "logps/rejected": -396.1431579589844, | |
| "loss": 0.5296, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.8962229490280151, | |
| "rewards/margins": 0.6276494264602661, | |
| "rewards/rejected": -1.5238726139068604, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 5.25, | |
| "learning_rate": 4.611847415796476e-06, | |
| "logits/chosen": -2.0031449794769287, | |
| "logits/rejected": -1.9153926372528076, | |
| "logps/chosen": -377.18121337890625, | |
| "logps/rejected": -411.65234375, | |
| "loss": 0.5136, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -1.0275315046310425, | |
| "rewards/margins": 0.6614707708358765, | |
| "rewards/rejected": -1.6890023946762085, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_logits/chosen": -1.8879700899124146, | |
| "eval_logits/rejected": -1.8081378936767578, | |
| "eval_logps/chosen": -374.9992370605469, | |
| "eval_logps/rejected": -414.1807556152344, | |
| "eval_loss": 0.5519765019416809, | |
| "eval_rewards/accuracies": 0.718999981880188, | |
| "eval_rewards/chosen": -1.0329629182815552, | |
| "eval_rewards/margins": 0.6316434741020203, | |
| "eval_rewards/rejected": -1.6646064519882202, | |
| "eval_runtime": 205.8805, | |
| "eval_samples_per_second": 9.714, | |
| "eval_steps_per_second": 0.607, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 5.3125, | |
| "learning_rate": 4.587044201869378e-06, | |
| "logits/chosen": -1.9341243505477905, | |
| "logits/rejected": -1.8541990518569946, | |
| "logps/chosen": -348.41650390625, | |
| "logps/rejected": -377.76055908203125, | |
| "loss": 0.5499, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.9087320566177368, | |
| "rewards/margins": 0.5346066355705261, | |
| "rewards/rejected": -1.4433386325836182, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 5.40625, | |
| "learning_rate": 4.561543932132574e-06, | |
| "logits/chosen": -1.9719091653823853, | |
| "logits/rejected": -1.8899688720703125, | |
| "logps/chosen": -356.0486145019531, | |
| "logps/rejected": -401.54852294921875, | |
| "loss": 0.5157, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.79421067237854, | |
| "rewards/margins": 0.6548107266426086, | |
| "rewards/rejected": -1.449021577835083, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 6.5625, | |
| "learning_rate": 4.535355123469009e-06, | |
| "logits/chosen": -1.938852071762085, | |
| "logits/rejected": -1.8662010431289673, | |
| "logps/chosen": -384.50897216796875, | |
| "logps/rejected": -447.8426818847656, | |
| "loss": 0.5328, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.1384373903274536, | |
| "rewards/margins": 0.725753664970398, | |
| "rewards/rejected": -1.8641912937164307, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 5.625, | |
| "learning_rate": 4.508486522728037e-06, | |
| "logits/chosen": -1.7794740200042725, | |
| "logits/rejected": -1.712432861328125, | |
| "logps/chosen": -402.6400146484375, | |
| "logps/rejected": -434.7808532714844, | |
| "loss": 0.5375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.2654469013214111, | |
| "rewards/margins": 0.6806719303131104, | |
| "rewards/rejected": -1.9461190700531006, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 5.28125, | |
| "learning_rate": 4.480947103804044e-06, | |
| "logits/chosen": -1.789461374282837, | |
| "logits/rejected": -1.69761061668396, | |
| "logps/chosen": -392.892578125, | |
| "logps/rejected": -422.740478515625, | |
| "loss": 0.5647, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.173139214515686, | |
| "rewards/margins": 0.6945830583572388, | |
| "rewards/rejected": -1.8677222728729248, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 4.09375, | |
| "learning_rate": 4.452746064639239e-06, | |
| "logits/chosen": -1.7166792154312134, | |
| "logits/rejected": -1.6769170761108398, | |
| "logps/chosen": -347.1570129394531, | |
| "logps/rejected": -412.37762451171875, | |
| "loss": 0.5468, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.7677599787712097, | |
| "rewards/margins": 0.697329580783844, | |
| "rewards/rejected": -1.4650894403457642, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 4.423892824151617e-06, | |
| "logits/chosen": -1.834933876991272, | |
| "logits/rejected": -1.6764202117919922, | |
| "logps/chosen": -381.7269592285156, | |
| "logps/rejected": -391.5668029785156, | |
| "loss": 0.5411, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.7797503471374512, | |
| "rewards/margins": 0.558749794960022, | |
| "rewards/rejected": -1.3385001420974731, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 5.6875, | |
| "learning_rate": 4.3943970190891164e-06, | |
| "logits/chosen": -1.7892568111419678, | |
| "logits/rejected": -1.662570595741272, | |
| "logps/chosen": -390.51953125, | |
| "logps/rejected": -397.81451416015625, | |
| "loss": 0.5465, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8887346982955933, | |
| "rewards/margins": 0.6098966002464294, | |
| "rewards/rejected": -1.498631477355957, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 4.90625, | |
| "learning_rate": 4.364268500811025e-06, | |
| "logits/chosen": -1.6196047067642212, | |
| "logits/rejected": -1.5555574893951416, | |
| "logps/chosen": -402.47772216796875, | |
| "logps/rejected": -446.379638671875, | |
| "loss": 0.5403, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.4109909534454346, | |
| "rewards/margins": 0.6602510213851929, | |
| "rewards/rejected": -2.071241855621338, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 6.0, | |
| "learning_rate": 4.333517331997704e-06, | |
| "logits/chosen": -1.6825096607208252, | |
| "logits/rejected": -1.56454336643219, | |
| "logps/chosen": -378.9342041015625, | |
| "logps/rejected": -428.3062438964844, | |
| "loss": 0.5587, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.3179118633270264, | |
| "rewards/margins": 0.6010549068450928, | |
| "rewards/rejected": -1.9189666509628296, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "eval_logits/chosen": -1.5609403848648071, | |
| "eval_logits/rejected": -1.4665495157241821, | |
| "eval_logps/chosen": -403.8534240722656, | |
| "eval_logps/rejected": -448.6079406738281, | |
| "eval_loss": 0.5327390432357788, | |
| "eval_rewards/accuracies": 0.7319999933242798, | |
| "eval_rewards/chosen": -1.3215045928955078, | |
| "eval_rewards/margins": 0.6873735189437866, | |
| "eval_rewards/rejected": -2.008878231048584, | |
| "eval_runtime": 205.8124, | |
| "eval_samples_per_second": 9.718, | |
| "eval_steps_per_second": 0.607, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 4.302153783289737e-06, | |
| "logits/chosen": -1.657099962234497, | |
| "logits/rejected": -1.5476830005645752, | |
| "logps/chosen": -389.5664367675781, | |
| "logps/rejected": -463.8951110839844, | |
| "loss": 0.5259, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.1841762065887451, | |
| "rewards/margins": 0.7400022745132446, | |
| "rewards/rejected": -1.9241783618927002, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 6.53125, | |
| "learning_rate": 4.270188329857613e-06, | |
| "logits/chosen": -1.687072515487671, | |
| "logits/rejected": -1.5819660425186157, | |
| "logps/chosen": -392.8175354003906, | |
| "logps/rejected": -414.7140197753906, | |
| "loss": 0.531, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.9409040212631226, | |
| "rewards/margins": 0.7219224572181702, | |
| "rewards/rejected": -1.6628265380859375, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 6.53125, | |
| "learning_rate": 4.237631647903115e-06, | |
| "logits/chosen": -1.6374574899673462, | |
| "logits/rejected": -1.513041377067566, | |
| "logps/chosen": -367.43792724609375, | |
| "logps/rejected": -422.2757873535156, | |
| "loss": 0.5751, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.0590665340423584, | |
| "rewards/margins": 0.6901532411575317, | |
| "rewards/rejected": -1.749219536781311, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 5.5, | |
| "learning_rate": 4.204494611093548e-06, | |
| "logits/chosen": -1.5762876272201538, | |
| "logits/rejected": -1.439675211906433, | |
| "logps/chosen": -353.2894592285156, | |
| "logps/rejected": -415.69384765625, | |
| "loss": 0.5084, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.0523182153701782, | |
| "rewards/margins": 0.8923817873001099, | |
| "rewards/rejected": -1.944700002670288, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 5.75, | |
| "learning_rate": 4.170788286930024e-06, | |
| "logits/chosen": -1.5430004596710205, | |
| "logits/rejected": -1.5310301780700684, | |
| "logps/chosen": -418.9070739746094, | |
| "logps/rejected": -503.3553771972656, | |
| "loss": 0.5448, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.7990005016326904, | |
| "rewards/margins": 0.6734641194343567, | |
| "rewards/rejected": -2.4724647998809814, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 4.136523933051005e-06, | |
| "logits/chosen": -1.553636074066162, | |
| "logits/rejected": -1.4528017044067383, | |
| "logps/chosen": -416.9197692871094, | |
| "logps/rejected": -488.38824462890625, | |
| "loss": 0.5151, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.6012948751449585, | |
| "rewards/margins": 0.7516334652900696, | |
| "rewards/rejected": -2.352928638458252, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 5.375, | |
| "learning_rate": 4.101712993472348e-06, | |
| "logits/chosen": -1.5965683460235596, | |
| "logits/rejected": -1.5354357957839966, | |
| "logps/chosen": -373.5552978515625, | |
| "logps/rejected": -417.95806884765625, | |
| "loss": 0.5543, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.0864415168762207, | |
| "rewards/margins": 0.6252428293228149, | |
| "rewards/rejected": -1.711684226989746, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 7.46875, | |
| "learning_rate": 4.066367094765091e-06, | |
| "logits/chosen": -1.5761507749557495, | |
| "logits/rejected": -1.4767498970031738, | |
| "logps/chosen": -377.71893310546875, | |
| "logps/rejected": -418.2323303222656, | |
| "loss": 0.536, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.113816261291504, | |
| "rewards/margins": 0.6647173166275024, | |
| "rewards/rejected": -1.7785335779190063, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 4.9375, | |
| "learning_rate": 4.030498042172277e-06, | |
| "logits/chosen": -1.692792296409607, | |
| "logits/rejected": -1.6168203353881836, | |
| "logps/chosen": -386.3111267089844, | |
| "logps/rejected": -467.1307678222656, | |
| "loss": 0.5083, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.1883788108825684, | |
| "rewards/margins": 0.8070133328437805, | |
| "rewards/rejected": -1.9953922033309937, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 8.625, | |
| "learning_rate": 3.994117815666095e-06, | |
| "logits/chosen": -1.5926947593688965, | |
| "logits/rejected": -1.482301950454712, | |
| "logps/chosen": -407.1866455078125, | |
| "logps/rejected": -477.66632080078125, | |
| "loss": 0.5167, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.3628543615341187, | |
| "rewards/margins": 0.8543855547904968, | |
| "rewards/rejected": -2.21724009513855, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "eval_logits/chosen": -1.4903253316879272, | |
| "eval_logits/rejected": -1.3917725086212158, | |
| "eval_logps/chosen": -399.6684265136719, | |
| "eval_logps/rejected": -467.6413269042969, | |
| "eval_loss": 0.5299228429794312, | |
| "eval_rewards/accuracies": 0.7229999899864197, | |
| "eval_rewards/chosen": -1.2796550989151, | |
| "eval_rewards/margins": 0.9195566177368164, | |
| "eval_rewards/rejected": -2.199211597442627, | |
| "eval_runtime": 205.8617, | |
| "eval_samples_per_second": 9.715, | |
| "eval_steps_per_second": 0.607, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 7.125, | |
| "learning_rate": 3.957238565946672e-06, | |
| "logits/chosen": -1.5940111875534058, | |
| "logits/rejected": -1.4753977060317993, | |
| "logps/chosen": -437.0880432128906, | |
| "logps/rejected": -492.3658752441406, | |
| "loss": 0.4612, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -1.347034215927124, | |
| "rewards/margins": 0.9986494183540344, | |
| "rewards/rejected": -2.345684051513672, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 6.96875, | |
| "learning_rate": 3.919872610383831e-06, | |
| "logits/chosen": -1.4148577451705933, | |
| "logits/rejected": -1.3983594179153442, | |
| "logps/chosen": -392.7431640625, | |
| "logps/rejected": -485.83819580078125, | |
| "loss": 0.4803, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.4197003841400146, | |
| "rewards/margins": 1.0409051179885864, | |
| "rewards/rejected": -2.4606053829193115, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 7.0, | |
| "learning_rate": 3.882032428903195e-06, | |
| "logits/chosen": -1.5557122230529785, | |
| "logits/rejected": -1.5083773136138916, | |
| "logps/chosen": -366.6706848144531, | |
| "logps/rejected": -449.88397216796875, | |
| "loss": 0.5389, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.3032705783843994, | |
| "rewards/margins": 0.8904848098754883, | |
| "rewards/rejected": -2.1937553882598877, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 3.84373065981799e-06, | |
| "logits/chosen": -1.607391595840454, | |
| "logits/rejected": -1.5484809875488281, | |
| "logps/chosen": -386.6382751464844, | |
| "logps/rejected": -422.96148681640625, | |
| "loss": 0.4946, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.2972227334976196, | |
| "rewards/margins": 0.7347890734672546, | |
| "rewards/rejected": -2.0320117473602295, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 9.625, | |
| "learning_rate": 3.8049800956079552e-06, | |
| "logits/chosen": -1.5829339027404785, | |
| "logits/rejected": -1.4720735549926758, | |
| "logps/chosen": -412.76531982421875, | |
| "logps/rejected": -479.85272216796875, | |
| "loss": 0.4811, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.502972960472107, | |
| "rewards/margins": 0.8380988240242004, | |
| "rewards/rejected": -2.341071605682373, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 8.375, | |
| "learning_rate": 3.765793678646753e-06, | |
| "logits/chosen": -1.4679361581802368, | |
| "logits/rejected": -1.4195685386657715, | |
| "logps/chosen": -412.76763916015625, | |
| "logps/rejected": -477.0049743652344, | |
| "loss": 0.5676, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.7044397592544556, | |
| "rewards/margins": 0.8208500742912292, | |
| "rewards/rejected": -2.525290012359619, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 6.125, | |
| "learning_rate": 3.726184496879323e-06, | |
| "logits/chosen": -1.4565179347991943, | |
| "logits/rejected": -1.4047472476959229, | |
| "logps/chosen": -431.69512939453125, | |
| "logps/rejected": -482.6431579589844, | |
| "loss": 0.5755, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.7018753290176392, | |
| "rewards/margins": 0.5642359852790833, | |
| "rewards/rejected": -2.2661116123199463, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 3.686165779450619e-06, | |
| "logits/chosen": -1.5907853841781616, | |
| "logits/rejected": -1.5108683109283447, | |
| "logps/chosen": -422.84735107421875, | |
| "logps/rejected": -478.88580322265625, | |
| "loss": 0.4939, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.5510692596435547, | |
| "rewards/margins": 0.7624626159667969, | |
| "rewards/rejected": -2.3135318756103516, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 3.645750892287178e-06, | |
| "logits/chosen": -1.51558518409729, | |
| "logits/rejected": -1.4765126705169678, | |
| "logps/chosen": -420.275634765625, | |
| "logps/rejected": -512.0084838867188, | |
| "loss": 0.5178, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.6825615167617798, | |
| "rewards/margins": 0.7326836585998535, | |
| "rewards/rejected": -2.4152450561523438, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 5.59375, | |
| "learning_rate": 3.604953333633009e-06, | |
| "logits/chosen": -1.5319992303848267, | |
| "logits/rejected": -1.4801173210144043, | |
| "logps/chosen": -434.16082763671875, | |
| "logps/rejected": -493.60308837890625, | |
| "loss": 0.5465, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.7253410816192627, | |
| "rewards/margins": 0.7062736749649048, | |
| "rewards/rejected": -2.431614398956299, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": -1.4641598463058472, | |
| "eval_logits/rejected": -1.3684852123260498, | |
| "eval_logps/chosen": -438.16168212890625, | |
| "eval_logps/rejected": -494.5843505859375, | |
| "eval_loss": 0.5188941359519958, | |
| "eval_rewards/accuracies": 0.7200000286102295, | |
| "eval_rewards/chosen": -1.664587140083313, | |
| "eval_rewards/margins": 0.8040550947189331, | |
| "eval_rewards/rejected": -2.468642234802246, | |
| "eval_runtime": 205.9278, | |
| "eval_samples_per_second": 9.712, | |
| "eval_steps_per_second": 0.607, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 7.59375, | |
| "learning_rate": 3.56378672954129e-06, | |
| "logits/chosen": -1.5462032556533813, | |
| "logits/rejected": -1.4628247022628784, | |
| "logps/chosen": -409.8141174316406, | |
| "logps/rejected": -486.94940185546875, | |
| "loss": 0.5401, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.5370090007781982, | |
| "rewards/margins": 0.756435751914978, | |
| "rewards/rejected": -2.293444871902466, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 6.34375, | |
| "learning_rate": 3.5222648293233806e-06, | |
| "logits/chosen": -1.492166519165039, | |
| "logits/rejected": -1.4682246446609497, | |
| "logps/chosen": -375.8512268066406, | |
| "logps/rejected": -446.55914306640625, | |
| "loss": 0.5386, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.162585973739624, | |
| "rewards/margins": 0.9081439971923828, | |
| "rewards/rejected": -2.070729970932007, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 7.25, | |
| "learning_rate": 3.4804015009566573e-06, | |
| "logits/chosen": -1.5281693935394287, | |
| "logits/rejected": -1.4133872985839844, | |
| "logps/chosen": -343.50860595703125, | |
| "logps/rejected": -365.1877746582031, | |
| "loss": 0.5756, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -1.016842246055603, | |
| "rewards/margins": 0.5352990031242371, | |
| "rewards/rejected": -1.5521411895751953, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 7.90625, | |
| "learning_rate": 3.4382107264527244e-06, | |
| "logits/chosen": -1.5180078744888306, | |
| "logits/rejected": -1.351276159286499, | |
| "logps/chosen": -419.5958557128906, | |
| "logps/rejected": -453.79693603515625, | |
| "loss": 0.4943, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.1950498819351196, | |
| "rewards/margins": 0.8461320996284485, | |
| "rewards/rejected": -2.041182041168213, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 7.0, | |
| "learning_rate": 3.3957065971875387e-06, | |
| "logits/chosen": -1.488877534866333, | |
| "logits/rejected": -1.3683526515960693, | |
| "logps/chosen": -446.36004638671875, | |
| "logps/rejected": -501.3478088378906, | |
| "loss": 0.5352, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.955243468284607, | |
| "rewards/margins": 0.8272415399551392, | |
| "rewards/rejected": -2.782485246658325, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 5.59375, | |
| "learning_rate": 3.352903309194999e-06, | |
| "logits/chosen": -1.5099315643310547, | |
| "logits/rejected": -1.4115077257156372, | |
| "logps/chosen": -438.7635803222656, | |
| "logps/rejected": -531.4382934570312, | |
| "loss": 0.5104, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.076658248901367, | |
| "rewards/margins": 0.8972193598747253, | |
| "rewards/rejected": -2.9738779067993164, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 6.21875, | |
| "learning_rate": 3.309815158425591e-06, | |
| "logits/chosen": -1.3534867763519287, | |
| "logits/rejected": -1.33297598361969, | |
| "logps/chosen": -432.561279296875, | |
| "logps/rejected": -499.51153564453125, | |
| "loss": 0.5148, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.744372010231018, | |
| "rewards/margins": 0.8061767816543579, | |
| "rewards/rejected": -2.550548791885376, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 8.25, | |
| "learning_rate": 3.266456535971654e-06, | |
| "logits/chosen": -1.5983731746673584, | |
| "logits/rejected": -1.5390180349349976, | |
| "logps/chosen": -411.49652099609375, | |
| "logps/rejected": -469.626708984375, | |
| "loss": 0.5347, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.632322907447815, | |
| "rewards/margins": 0.7145469784736633, | |
| "rewards/rejected": -2.346869945526123, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 3.2228419232608692e-06, | |
| "logits/chosen": -1.6237919330596924, | |
| "logits/rejected": -1.5474942922592163, | |
| "logps/chosen": -407.1629943847656, | |
| "logps/rejected": -489.897705078125, | |
| "loss": 0.502, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.5020484924316406, | |
| "rewards/margins": 0.8913022875785828, | |
| "rewards/rejected": -2.393350839614868, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 6.3125, | |
| "learning_rate": 3.1789858872195888e-06, | |
| "logits/chosen": -1.4964396953582764, | |
| "logits/rejected": -1.4070580005645752, | |
| "logps/chosen": -410.7691345214844, | |
| "logps/rejected": -528.3038330078125, | |
| "loss": 0.5002, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.6621105670928955, | |
| "rewards/margins": 1.0146641731262207, | |
| "rewards/rejected": -2.676774501800537, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "eval_logits/chosen": -1.505366325378418, | |
| "eval_logits/rejected": -1.4178627729415894, | |
| "eval_logps/chosen": -450.1383056640625, | |
| "eval_logps/rejected": -519.8884887695312, | |
| "eval_loss": 0.5142305493354797, | |
| "eval_rewards/accuracies": 0.7289999723434448, | |
| "eval_rewards/chosen": -1.784353494644165, | |
| "eval_rewards/margins": 0.9373300075531006, | |
| "eval_rewards/rejected": -2.7216835021972656, | |
| "eval_runtime": 205.8963, | |
| "eval_samples_per_second": 9.714, | |
| "eval_steps_per_second": 0.607, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 7.90625, | |
| "learning_rate": 3.1349030754075945e-06, | |
| "logits/chosen": -1.473548173904419, | |
| "logits/rejected": -1.4354422092437744, | |
| "logps/chosen": -455.54534912109375, | |
| "logps/rejected": -518.474609375, | |
| "loss": 0.5482, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.9718334674835205, | |
| "rewards/margins": 0.7413331270217896, | |
| "rewards/rejected": -2.7131667137145996, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 5.78125, | |
| "learning_rate": 3.0906082111259313e-06, | |
| "logits/chosen": -1.4893120527267456, | |
| "logits/rejected": -1.4045579433441162, | |
| "logps/chosen": -469.7232360839844, | |
| "logps/rejected": -551.2777099609375, | |
| "loss": 0.5324, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.048574209213257, | |
| "rewards/margins": 1.0462497472763062, | |
| "rewards/rejected": -3.0948235988616943, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 13.0, | |
| "learning_rate": 3.046116088499449e-06, | |
| "logits/chosen": -1.518959403038025, | |
| "logits/rejected": -1.447871208190918, | |
| "logps/chosen": -473.65850830078125, | |
| "logps/rejected": -532.4075927734375, | |
| "loss": 0.514, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -2.0019118785858154, | |
| "rewards/margins": 0.7329636216163635, | |
| "rewards/rejected": -2.734875202178955, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 6.375, | |
| "learning_rate": 3.0014415675356813e-06, | |
| "logits/chosen": -1.5328662395477295, | |
| "logits/rejected": -1.4063748121261597, | |
| "logps/chosen": -478.6943359375, | |
| "logps/rejected": -537.5908813476562, | |
| "loss": 0.4481, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.9542016983032227, | |
| "rewards/margins": 0.9170918464660645, | |
| "rewards/rejected": -2.871293544769287, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 7.71875, | |
| "learning_rate": 2.9565995691617242e-06, | |
| "logits/chosen": -1.4631025791168213, | |
| "logits/rejected": -1.3391731977462769, | |
| "logps/chosen": -521.5445556640625, | |
| "logps/rejected": -596.1942138671875, | |
| "loss": 0.4863, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.2124335765838623, | |
| "rewards/margins": 1.047732949256897, | |
| "rewards/rejected": -3.260166645050049, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 5.46875, | |
| "learning_rate": 2.9116050702407706e-06, | |
| "logits/chosen": -1.5032353401184082, | |
| "logits/rejected": -1.4071391820907593, | |
| "logps/chosen": -507.3263244628906, | |
| "logps/rejected": -564.1354370117188, | |
| "loss": 0.5064, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.2760252952575684, | |
| "rewards/margins": 0.9310242533683777, | |
| "rewards/rejected": -3.20704984664917, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 5.96875, | |
| "learning_rate": 2.8664730985699537e-06, | |
| "logits/chosen": -1.5140959024429321, | |
| "logits/rejected": -1.4394776821136475, | |
| "logps/chosen": -478.9674377441406, | |
| "logps/rejected": -564.43212890625, | |
| "loss": 0.5034, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.1803536415100098, | |
| "rewards/margins": 0.8669061660766602, | |
| "rewards/rejected": -3.04725980758667, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 4.15625, | |
| "learning_rate": 2.8212187278611907e-06, | |
| "logits/chosen": -1.5928372144699097, | |
| "logits/rejected": -1.4567886590957642, | |
| "logps/chosen": -487.82489013671875, | |
| "logps/rejected": -544.0657348632812, | |
| "loss": 0.4944, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.9777450561523438, | |
| "rewards/margins": 0.938134491443634, | |
| "rewards/rejected": -2.915879726409912, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 7.125, | |
| "learning_rate": 2.7758570727066843e-06, | |
| "logits/chosen": -1.4225056171417236, | |
| "logits/rejected": -1.3867155313491821, | |
| "logps/chosen": -464.3641052246094, | |
| "logps/rejected": -549.7391357421875, | |
| "loss": 0.5201, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.0725159645080566, | |
| "rewards/margins": 1.0133941173553467, | |
| "rewards/rejected": -3.0859100818634033, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 7.34375, | |
| "learning_rate": 2.730403283530767e-06, | |
| "logits/chosen": -1.4609169960021973, | |
| "logits/rejected": -1.3873465061187744, | |
| "logps/chosen": -514.2507934570312, | |
| "logps/rejected": -596.9881591796875, | |
| "loss": 0.5017, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -2.514453887939453, | |
| "rewards/margins": 0.9918726682662964, | |
| "rewards/rejected": -3.506326198577881, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "eval_logits/chosen": -1.3947843313217163, | |
| "eval_logits/rejected": -1.2972902059555054, | |
| "eval_logps/chosen": -533.4492797851562, | |
| "eval_logps/rejected": -608.9218139648438, | |
| "eval_loss": 0.5058408379554749, | |
| "eval_rewards/accuracies": 0.7360000014305115, | |
| "eval_rewards/chosen": -2.6174628734588623, | |
| "eval_rewards/margins": 0.9945541024208069, | |
| "eval_rewards/rejected": -3.6120169162750244, | |
| "eval_runtime": 205.8968, | |
| "eval_samples_per_second": 9.714, | |
| "eval_steps_per_second": 0.607, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 6.0, | |
| "learning_rate": 2.6848725415297888e-06, | |
| "logits/chosen": -1.4565789699554443, | |
| "logits/rejected": -1.4135055541992188, | |
| "logps/chosen": -489.64874267578125, | |
| "logps/rejected": -567.3189697265625, | |
| "loss": 0.4895, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.4934046268463135, | |
| "rewards/margins": 0.8419889211654663, | |
| "rewards/rejected": -3.3353939056396484, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 7.375, | |
| "learning_rate": 2.639280053601719e-06, | |
| "logits/chosen": -1.573761224746704, | |
| "logits/rejected": -1.4989092350006104, | |
| "logps/chosen": -512.4972534179688, | |
| "logps/rejected": -576.3961181640625, | |
| "loss": 0.4665, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.162745952606201, | |
| "rewards/margins": 1.0285950899124146, | |
| "rewards/rejected": -3.191340923309326, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 6.15625, | |
| "learning_rate": 2.59364104726716e-06, | |
| "logits/chosen": -1.484546184539795, | |
| "logits/rejected": -1.3231149911880493, | |
| "logps/chosen": -528.154541015625, | |
| "logps/rejected": -576.6033325195312, | |
| "loss": 0.512, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.636870861053467, | |
| "rewards/margins": 0.9289640188217163, | |
| "rewards/rejected": -3.5658352375030518, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 4.4375, | |
| "learning_rate": 2.547970765583491e-06, | |
| "logits/chosen": -1.4552520513534546, | |
| "logits/rejected": -1.4171597957611084, | |
| "logps/chosen": -519.861328125, | |
| "logps/rejected": -591.8118896484375, | |
| "loss": 0.5328, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.421581745147705, | |
| "rewards/margins": 0.9120811223983765, | |
| "rewards/rejected": -3.33366322517395, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 7.3125, | |
| "learning_rate": 2.502284462053799e-06, | |
| "logits/chosen": -1.4553632736206055, | |
| "logits/rejected": -1.3192330598831177, | |
| "logps/chosen": -525.8233642578125, | |
| "logps/rejected": -599.904296875, | |
| "loss": 0.4895, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.279545307159424, | |
| "rewards/margins": 1.0983493328094482, | |
| "rewards/rejected": -3.377894639968872, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 7.3125, | |
| "learning_rate": 2.456597395532338e-06, | |
| "logits/chosen": -1.4849998950958252, | |
| "logits/rejected": -1.4006808996200562, | |
| "logps/chosen": -488.8839416503906, | |
| "logps/rejected": -538.5235595703125, | |
| "loss": 0.499, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.3017091751098633, | |
| "rewards/margins": 0.8732277750968933, | |
| "rewards/rejected": -3.1749370098114014, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 6.375, | |
| "learning_rate": 2.4109248251281953e-06, | |
| "logits/chosen": -1.5076830387115479, | |
| "logits/rejected": -1.427677869796753, | |
| "logps/chosen": -512.9411010742188, | |
| "logps/rejected": -549.7650146484375, | |
| "loss": 0.5241, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.252761125564575, | |
| "rewards/margins": 0.8173268437385559, | |
| "rewards/rejected": -3.0700876712799072, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 6.03125, | |
| "learning_rate": 2.365282005108875e-06, | |
| "logits/chosen": -1.5420407056808472, | |
| "logits/rejected": -1.4118328094482422, | |
| "logps/chosen": -469.3882751464844, | |
| "logps/rejected": -503.13726806640625, | |
| "loss": 0.5381, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.117811679840088, | |
| "rewards/margins": 0.8648239970207214, | |
| "rewards/rejected": -2.982635498046875, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 10.9375, | |
| "learning_rate": 2.319684179805491e-06, | |
| "logits/chosen": -1.5114177465438843, | |
| "logits/rejected": -1.377497673034668, | |
| "logps/chosen": -492.01141357421875, | |
| "logps/rejected": -511.5269470214844, | |
| "loss": 0.5381, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.3241496086120605, | |
| "rewards/margins": 0.6907966732978821, | |
| "rewards/rejected": -3.014946460723877, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 6.0, | |
| "learning_rate": 2.2741465785212905e-06, | |
| "logits/chosen": -1.4633898735046387, | |
| "logits/rejected": -1.3916144371032715, | |
| "logps/chosen": -471.26141357421875, | |
| "logps/rejected": -557.3035278320312, | |
| "loss": 0.4966, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -2.141718626022339, | |
| "rewards/margins": 0.8240246772766113, | |
| "rewards/rejected": -2.96574330329895, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "eval_logits/chosen": -1.4740034341812134, | |
| "eval_logits/rejected": -1.3782777786254883, | |
| "eval_logps/chosen": -477.5079650878906, | |
| "eval_logps/rejected": -545.9103393554688, | |
| "eval_loss": 0.504342794418335, | |
| "eval_rewards/accuracies": 0.7369999885559082, | |
| "eval_rewards/chosen": -2.0580503940582275, | |
| "eval_rewards/margins": 0.923851728439331, | |
| "eval_rewards/rejected": -2.9819023609161377, | |
| "eval_runtime": 205.8953, | |
| "eval_samples_per_second": 9.714, | |
| "eval_steps_per_second": 0.607, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 5.75, | |
| "learning_rate": 2.2286844104451848e-06, | |
| "logits/chosen": -1.4495588541030884, | |
| "logits/rejected": -1.3826367855072021, | |
| "logps/chosen": -462.7281799316406, | |
| "logps/rejected": -548.97265625, | |
| "loss": 0.5435, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -2.0098259449005127, | |
| "rewards/margins": 0.95000821352005, | |
| "rewards/rejected": -2.959834337234497, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 7.125, | |
| "learning_rate": 2.183312859572008e-06, | |
| "logits/chosen": -1.4757729768753052, | |
| "logits/rejected": -1.3990167379379272, | |
| "logps/chosen": -496.04339599609375, | |
| "logps/rejected": -542.3977661132812, | |
| "loss": 0.4874, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.0698461532592773, | |
| "rewards/margins": 0.8941160440444946, | |
| "rewards/rejected": -2.9639620780944824, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 9.125, | |
| "learning_rate": 2.1380470796311843e-06, | |
| "logits/chosen": -1.4388748407363892, | |
| "logits/rejected": -1.3238605260849, | |
| "logps/chosen": -490.1044006347656, | |
| "logps/rejected": -552.4682006835938, | |
| "loss": 0.5065, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.0447587966918945, | |
| "rewards/margins": 0.98194420337677, | |
| "rewards/rejected": -3.026703119277954, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 9.8125, | |
| "learning_rate": 2.092902189025507e-06, | |
| "logits/chosen": -1.579469919204712, | |
| "logits/rejected": -1.4409890174865723, | |
| "logps/chosen": -501.4720764160156, | |
| "logps/rejected": -554.6583251953125, | |
| "loss": 0.5142, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.0851852893829346, | |
| "rewards/margins": 0.9953821301460266, | |
| "rewards/rejected": -3.0805675983428955, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 6.71875, | |
| "learning_rate": 2.0478932657817105e-06, | |
| "logits/chosen": -1.5608408451080322, | |
| "logits/rejected": -1.4242637157440186, | |
| "logps/chosen": -509.3160705566406, | |
| "logps/rejected": -564.5472412109375, | |
| "loss": 0.4965, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.316391944885254, | |
| "rewards/margins": 0.892681896686554, | |
| "rewards/rejected": -3.209073543548584, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 6.375, | |
| "learning_rate": 2.0030353425145376e-06, | |
| "logits/chosen": -1.4031821489334106, | |
| "logits/rejected": -1.392533779144287, | |
| "logps/chosen": -465.71734619140625, | |
| "logps/rejected": -529.384033203125, | |
| "loss": 0.5434, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -2.336087465286255, | |
| "rewards/margins": 0.8120396733283997, | |
| "rewards/rejected": -3.1481270790100098, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 7.75, | |
| "learning_rate": 1.958343401405964e-06, | |
| "logits/chosen": -1.4990739822387695, | |
| "logits/rejected": -1.444461703300476, | |
| "logps/chosen": -472.1663513183594, | |
| "logps/rejected": -566.5687255859375, | |
| "loss": 0.533, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.144223690032959, | |
| "rewards/margins": 0.9988595843315125, | |
| "rewards/rejected": -3.1430835723876953, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 6.65625, | |
| "learning_rate": 1.9138323692012734e-06, | |
| "logits/chosen": -1.5581806898117065, | |
| "logits/rejected": -1.4660371541976929, | |
| "logps/chosen": -500.2381286621094, | |
| "logps/rejected": -572.9973754882812, | |
| "loss": 0.4747, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.226621150970459, | |
| "rewards/margins": 0.942787766456604, | |
| "rewards/rejected": -3.1694092750549316, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 5.375, | |
| "learning_rate": 1.8695171122236443e-06, | |
| "logits/chosen": -1.4782741069793701, | |
| "logits/rejected": -1.3444011211395264, | |
| "logps/chosen": -538.46923828125, | |
| "logps/rejected": -594.9655151367188, | |
| "loss": 0.4581, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -2.3236842155456543, | |
| "rewards/margins": 1.0366638898849487, | |
| "rewards/rejected": -3.3603484630584717, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 5.5, | |
| "learning_rate": 1.8254124314089225e-06, | |
| "logits/chosen": -1.4107558727264404, | |
| "logits/rejected": -1.2883371114730835, | |
| "logps/chosen": -511.8663635253906, | |
| "logps/rejected": -562.5142822265625, | |
| "loss": 0.5087, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.319617748260498, | |
| "rewards/margins": 0.765404224395752, | |
| "rewards/rejected": -3.08502197265625, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": -1.4261987209320068, | |
| "eval_logits/rejected": -1.333054780960083, | |
| "eval_logps/chosen": -508.8494567871094, | |
| "eval_logps/rejected": -582.47119140625, | |
| "eval_loss": 0.5040486454963684, | |
| "eval_rewards/accuracies": 0.7450000047683716, | |
| "eval_rewards/chosen": -2.371464729309082, | |
| "eval_rewards/margins": 0.9760459661483765, | |
| "eval_rewards/rejected": -3.347510814666748, | |
| "eval_runtime": 205.8406, | |
| "eval_samples_per_second": 9.716, | |
| "eval_steps_per_second": 0.607, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 6.21875, | |
| "learning_rate": 1.781533057362221e-06, | |
| "logits/chosen": -1.5876115560531616, | |
| "logits/rejected": -1.5332571268081665, | |
| "logps/chosen": -502.91046142578125, | |
| "logps/rejected": -568.5400390625, | |
| "loss": 0.5013, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.2924888134002686, | |
| "rewards/margins": 0.8915454745292664, | |
| "rewards/rejected": -3.1840338706970215, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 7.8125, | |
| "learning_rate": 1.7378936454380277e-06, | |
| "logits/chosen": -1.4910120964050293, | |
| "logits/rejected": -1.3955504894256592, | |
| "logps/chosen": -508.55474853515625, | |
| "logps/rejected": -586.0122680664062, | |
| "loss": 0.5113, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.406351089477539, | |
| "rewards/margins": 1.08707594871521, | |
| "rewards/rejected": -3.49342679977417, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 6.90625, | |
| "learning_rate": 1.6945087708454273e-06, | |
| "logits/chosen": -1.5146963596343994, | |
| "logits/rejected": -1.401847243309021, | |
| "logps/chosen": -513.2926635742188, | |
| "logps/rejected": -559.9019775390625, | |
| "loss": 0.4528, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.2331342697143555, | |
| "rewards/margins": 0.9627434015274048, | |
| "rewards/rejected": -3.19587779045105, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 7.625, | |
| "learning_rate": 1.651392923780105e-06, | |
| "logits/chosen": -1.5016862154006958, | |
| "logits/rejected": -1.433712124824524, | |
| "logps/chosen": -480.95703125, | |
| "logps/rejected": -585.30712890625, | |
| "loss": 0.4813, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -2.1668860912323, | |
| "rewards/margins": 1.0968172550201416, | |
| "rewards/rejected": -3.2637035846710205, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 5.875, | |
| "learning_rate": 1.608560504584737e-06, | |
| "logits/chosen": -1.435272455215454, | |
| "logits/rejected": -1.3818395137786865, | |
| "logps/chosen": -489.82293701171875, | |
| "logps/rejected": -585.4735717773438, | |
| "loss": 0.4916, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.3021085262298584, | |
| "rewards/margins": 1.004734992980957, | |
| "rewards/rejected": -3.3068439960479736, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 5.59375, | |
| "learning_rate": 1.5660258189393945e-06, | |
| "logits/chosen": -1.527229905128479, | |
| "logits/rejected": -1.4203643798828125, | |
| "logps/chosen": -503.65716552734375, | |
| "logps/rejected": -556.7635498046875, | |
| "loss": 0.4766, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.143995523452759, | |
| "rewards/margins": 1.0609943866729736, | |
| "rewards/rejected": -3.2049896717071533, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 7.5625, | |
| "learning_rate": 1.5238030730835578e-06, | |
| "logits/chosen": -1.4004570245742798, | |
| "logits/rejected": -1.3315227031707764, | |
| "logps/chosen": -484.2720642089844, | |
| "logps/rejected": -577.0310668945312, | |
| "loss": 0.4337, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.3353257179260254, | |
| "rewards/margins": 1.0654656887054443, | |
| "rewards/rejected": -3.4007911682128906, | |
| "step": 1270 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 9.1875, | |
| "learning_rate": 1.4819063690713565e-06, | |
| "logits/chosen": -1.44136381149292, | |
| "logits/rejected": -1.3490862846374512, | |
| "logps/chosen": -492.79620361328125, | |
| "logps/rejected": -599.3836059570312, | |
| "loss": 0.5165, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.4487698078155518, | |
| "rewards/margins": 1.134749174118042, | |
| "rewards/rejected": -3.5835189819335938, | |
| "step": 1280 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 6.6875, | |
| "learning_rate": 1.4403497000615885e-06, | |
| "logits/chosen": -1.481796145439148, | |
| "logits/rejected": -1.3834607601165771, | |
| "logps/chosen": -502.568603515625, | |
| "logps/rejected": -589.8240356445312, | |
| "loss": 0.5391, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.3244917392730713, | |
| "rewards/margins": 0.9923933744430542, | |
| "rewards/rejected": -3.316884994506836, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 5.9375, | |
| "learning_rate": 1.3991469456441273e-06, | |
| "logits/chosen": -1.5021679401397705, | |
| "logits/rejected": -1.4300124645233154, | |
| "logps/chosen": -498.32763671875, | |
| "logps/rejected": -587.1395874023438, | |
| "loss": 0.4799, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.2161061763763428, | |
| "rewards/margins": 1.0599734783172607, | |
| "rewards/rejected": -3.2760796546936035, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "eval_logits/chosen": -1.4277222156524658, | |
| "eval_logits/rejected": -1.3340253829956055, | |
| "eval_logps/chosen": -502.36865234375, | |
| "eval_logps/rejected": -582.15625, | |
| "eval_loss": 0.5011327266693115, | |
| "eval_rewards/accuracies": 0.7450000047683716, | |
| "eval_rewards/chosen": -2.306657075881958, | |
| "eval_rewards/margins": 1.0377042293548584, | |
| "eval_rewards/rejected": -3.3443613052368164, | |
| "eval_runtime": 205.8366, | |
| "eval_samples_per_second": 9.716, | |
| "eval_steps_per_second": 0.607, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 14.375, | |
| "learning_rate": 1.3583118672042441e-06, | |
| "logits/chosen": -1.5242735147476196, | |
| "logits/rejected": -1.423452377319336, | |
| "logps/chosen": -495.67608642578125, | |
| "logps/rejected": -586.1196899414062, | |
| "loss": 0.4927, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.3464579582214355, | |
| "rewards/margins": 1.1166086196899414, | |
| "rewards/rejected": -3.463066577911377, | |
| "step": 1310 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 6.625, | |
| "learning_rate": 1.3178581033264218e-06, | |
| "logits/chosen": -1.4835782051086426, | |
| "logits/rejected": -1.394766092300415, | |
| "logps/chosen": -526.1903076171875, | |
| "logps/rejected": -599.1194458007812, | |
| "loss": 0.464, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.623629093170166, | |
| "rewards/margins": 0.9394786953926086, | |
| "rewards/rejected": -3.56310772895813, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 9.8125, | |
| "learning_rate": 1.2777991652391757e-06, | |
| "logits/chosen": -1.4715049266815186, | |
| "logits/rejected": -1.3863550424575806, | |
| "logps/chosen": -536.083984375, | |
| "logps/rejected": -642.5689697265625, | |
| "loss": 0.4522, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.66798996925354, | |
| "rewards/margins": 1.0482239723205566, | |
| "rewards/rejected": -3.7162139415740967, | |
| "step": 1330 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 7.28125, | |
| "learning_rate": 1.2381484323024178e-06, | |
| "logits/chosen": -1.418473482131958, | |
| "logits/rejected": -1.34147047996521, | |
| "logps/chosen": -537.9956665039062, | |
| "logps/rejected": -605.8355102539062, | |
| "loss": 0.4949, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.777344226837158, | |
| "rewards/margins": 1.0527513027191162, | |
| "rewards/rejected": -3.8300960063934326, | |
| "step": 1340 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 11.0, | |
| "learning_rate": 1.1989191475388518e-06, | |
| "logits/chosen": -1.5058234930038452, | |
| "logits/rejected": -1.4633159637451172, | |
| "logps/chosen": -527.7548828125, | |
| "logps/rejected": -597.4782104492188, | |
| "loss": 0.5086, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.516946792602539, | |
| "rewards/margins": 0.9122180938720703, | |
| "rewards/rejected": -3.429164409637451, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 6.03125, | |
| "learning_rate": 1.160124413210918e-06, | |
| "logits/chosen": -1.416601538658142, | |
| "logits/rejected": -1.351701021194458, | |
| "logps/chosen": -547.7350463867188, | |
| "logps/rejected": -607.982666015625, | |
| "loss": 0.5249, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.456645965576172, | |
| "rewards/margins": 0.9266567230224609, | |
| "rewards/rejected": -3.383302688598633, | |
| "step": 1360 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 6.125, | |
| "learning_rate": 1.1217771864447396e-06, | |
| "logits/chosen": -1.5074869394302368, | |
| "logits/rejected": -1.4494173526763916, | |
| "logps/chosen": -547.4534912109375, | |
| "logps/rejected": -611.69482421875, | |
| "loss": 0.4353, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.41982102394104, | |
| "rewards/margins": 1.221834659576416, | |
| "rewards/rejected": -3.641655445098877, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 6.59375, | |
| "learning_rate": 1.08389027490255e-06, | |
| "logits/chosen": -1.4987690448760986, | |
| "logits/rejected": -1.3986529111862183, | |
| "logps/chosen": -490.8663024902344, | |
| "logps/rejected": -572.9545288085938, | |
| "loss": 0.4849, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -2.3217995166778564, | |
| "rewards/margins": 0.8936141729354858, | |
| "rewards/rejected": -3.215414047241211, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 6.78125, | |
| "learning_rate": 1.046476332505036e-06, | |
| "logits/chosen": -1.4200499057769775, | |
| "logits/rejected": -1.355092167854309, | |
| "logps/chosen": -523.000244140625, | |
| "logps/rejected": -604.7027587890625, | |
| "loss": 0.4608, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.414511203765869, | |
| "rewards/margins": 1.0894904136657715, | |
| "rewards/rejected": -3.504002094268799, | |
| "step": 1390 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 6.875, | |
| "learning_rate": 1.0095478552050348e-06, | |
| "logits/chosen": -1.4363720417022705, | |
| "logits/rejected": -1.4057958126068115, | |
| "logps/chosen": -509.18304443359375, | |
| "logps/rejected": -622.3807373046875, | |
| "loss": 0.4606, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -2.538700580596924, | |
| "rewards/margins": 1.1025012731552124, | |
| "rewards/rejected": -3.6412017345428467, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "eval_logits/chosen": -1.4218993186950684, | |
| "eval_logits/rejected": -1.3291493654251099, | |
| "eval_logps/chosen": -521.8630981445312, | |
| "eval_logps/rejected": -603.546875, | |
| "eval_loss": 0.4991470277309418, | |
| "eval_rewards/accuracies": 0.7429999709129333, | |
| "eval_rewards/chosen": -2.5016019344329834, | |
| "eval_rewards/margins": 1.056665062904358, | |
| "eval_rewards/rejected": -3.5582666397094727, | |
| "eval_runtime": 206.8725, | |
| "eval_samples_per_second": 9.668, | |
| "eval_steps_per_second": 0.604, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 8.5625, | |
| "learning_rate": 9.731171768139808e-07, | |
| "logits/chosen": -1.425769567489624, | |
| "logits/rejected": -1.3510525226593018, | |
| "logps/chosen": -534.87353515625, | |
| "logps/rejected": -616.7877807617188, | |
| "loss": 0.5568, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.6500022411346436, | |
| "rewards/margins": 0.9020141363143921, | |
| "rewards/rejected": -3.552016496658325, | |
| "step": 1410 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 6.96875, | |
| "learning_rate": 9.371964648825221e-07, | |
| "logits/chosen": -1.497687578201294, | |
| "logits/rejected": -1.411454439163208, | |
| "logps/chosen": -542.4428100585938, | |
| "logps/rejected": -576.1937255859375, | |
| "loss": 0.5665, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.580232620239258, | |
| "rewards/margins": 0.7872194051742554, | |
| "rewards/rejected": -3.3674521446228027, | |
| "step": 1420 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 6.75, | |
| "learning_rate": 9.017977166366445e-07, | |
| "logits/chosen": -1.4858081340789795, | |
| "logits/rejected": -1.3852221965789795, | |
| "logps/chosen": -526.9737548828125, | |
| "logps/rejected": -635.5704345703125, | |
| "loss": 0.4595, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.443568468093872, | |
| "rewards/margins": 1.2052720785140991, | |
| "rewards/rejected": -3.6488404273986816, | |
| "step": 1430 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 7.40625, | |
| "learning_rate": 8.669327549707096e-07, | |
| "logits/chosen": -1.6181617975234985, | |
| "logits/rejected": -1.5521459579467773, | |
| "logps/chosen": -541.6924438476562, | |
| "logps/rejected": -613.9597778320312, | |
| "loss": 0.5011, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.4517698287963867, | |
| "rewards/margins": 0.8626209497451782, | |
| "rewards/rejected": -3.3143906593322754, | |
| "step": 1440 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 6.65625, | |
| "learning_rate": 8.326132244986932e-07, | |
| "logits/chosen": -1.511328935623169, | |
| "logits/rejected": -1.432902216911316, | |
| "logps/chosen": -508.4786071777344, | |
| "logps/rejected": -612.4136962890625, | |
| "loss": 0.488, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.4070212841033936, | |
| "rewards/margins": 1.1571033000946045, | |
| "rewards/rejected": -3.564124584197998, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 5.75, | |
| "learning_rate": 7.988505876649863e-07, | |
| "logits/chosen": -1.4642016887664795, | |
| "logits/rejected": -1.353859543800354, | |
| "logps/chosen": -487.744140625, | |
| "logps/rejected": -559.90234375, | |
| "loss": 0.5018, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.3719143867492676, | |
| "rewards/margins": 0.841119647026062, | |
| "rewards/rejected": -3.213034152984619, | |
| "step": 1460 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 8.4375, | |
| "learning_rate": 7.656561209160248e-07, | |
| "logits/chosen": -1.430325984954834, | |
| "logits/rejected": -1.3274606466293335, | |
| "logps/chosen": -524.908447265625, | |
| "logps/rejected": -588.8642578125, | |
| "loss": 0.4972, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.478081226348877, | |
| "rewards/margins": 1.044504165649414, | |
| "rewards/rejected": -3.522585391998291, | |
| "step": 1470 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 6.3125, | |
| "learning_rate": 7.330409109340563e-07, | |
| "logits/chosen": -1.4950729608535767, | |
| "logits/rejected": -1.399980902671814, | |
| "logps/chosen": -539.3687133789062, | |
| "logps/rejected": -593.13623046875, | |
| "loss": 0.5227, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.505924701690674, | |
| "rewards/margins": 1.0100219249725342, | |
| "rewards/rejected": -3.515946626663208, | |
| "step": 1480 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 6.5625, | |
| "learning_rate": 7.010158509342682e-07, | |
| "logits/chosen": -1.422590732574463, | |
| "logits/rejected": -1.3758926391601562, | |
| "logps/chosen": -523.3316650390625, | |
| "logps/rejected": -640.554931640625, | |
| "loss": 0.4806, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.6385467052459717, | |
| "rewards/margins": 1.0877034664154053, | |
| "rewards/rejected": -3.726250410079956, | |
| "step": 1490 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 6.65625, | |
| "learning_rate": 6.695916370265529e-07, | |
| "logits/chosen": -1.454978585243225, | |
| "logits/rejected": -1.4019845724105835, | |
| "logps/chosen": -488.15228271484375, | |
| "logps/rejected": -599.3827514648438, | |
| "loss": 0.4763, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.454293727874756, | |
| "rewards/margins": 1.122016429901123, | |
| "rewards/rejected": -3.576310396194458, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_logits/chosen": -1.4325312376022339, | |
| "eval_logits/rejected": -1.3394147157669067, | |
| "eval_logps/chosen": -521.4944458007812, | |
| "eval_logps/rejected": -599.7630615234375, | |
| "eval_loss": 0.4985138177871704, | |
| "eval_rewards/accuracies": 0.746999979019165, | |
| "eval_rewards/chosen": -2.497915267944336, | |
| "eval_rewards/margins": 1.0225143432617188, | |
| "eval_rewards/rejected": -3.5204296112060547, | |
| "eval_runtime": 206.7979, | |
| "eval_samples_per_second": 9.671, | |
| "eval_steps_per_second": 0.604, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 7.3125, | |
| "learning_rate": 6.387787646430854e-07, | |
| "logits/chosen": -1.4497127532958984, | |
| "logits/rejected": -1.3799619674682617, | |
| "logps/chosen": -531.990966796875, | |
| "logps/rejected": -574.7440185546875, | |
| "loss": 0.5299, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.597123861312866, | |
| "rewards/margins": 0.8730913400650024, | |
| "rewards/rejected": -3.470215320587158, | |
| "step": 1510 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 5.6875, | |
| "learning_rate": 6.085875250329401e-07, | |
| "logits/chosen": -1.5374524593353271, | |
| "logits/rejected": -1.4480254650115967, | |
| "logps/chosen": -493.16351318359375, | |
| "logps/rejected": -555.0206298828125, | |
| "loss": 0.4998, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.448246717453003, | |
| "rewards/margins": 0.842933177947998, | |
| "rewards/rejected": -3.291179656982422, | |
| "step": 1520 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 7.21875, | |
| "learning_rate": 5.79028001824894e-07, | |
| "logits/chosen": -1.4127936363220215, | |
| "logits/rejected": -1.4277909994125366, | |
| "logps/chosen": -473.9107360839844, | |
| "logps/rejected": -639.1129150390625, | |
| "loss": 0.4427, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -2.282348155975342, | |
| "rewards/margins": 1.2384718656539917, | |
| "rewards/rejected": -3.520819902420044, | |
| "step": 1530 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 4.5625, | |
| "learning_rate": 5.501100676595761e-07, | |
| "logits/chosen": -1.5411286354064941, | |
| "logits/rejected": -1.4845304489135742, | |
| "logps/chosen": -528.363525390625, | |
| "logps/rejected": -610.4037475585938, | |
| "loss": 0.4969, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.430098056793213, | |
| "rewards/margins": 1.0298941135406494, | |
| "rewards/rejected": -3.4599921703338623, | |
| "step": 1540 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 6.4375, | |
| "learning_rate": 5.218433808920884e-07, | |
| "logits/chosen": -1.5715194940567017, | |
| "logits/rejected": -1.4276337623596191, | |
| "logps/chosen": -515.0184326171875, | |
| "logps/rejected": -564.2301025390625, | |
| "loss": 0.4785, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.4059693813323975, | |
| "rewards/margins": 0.9038017988204956, | |
| "rewards/rejected": -3.3097712993621826, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 6.28125, | |
| "learning_rate": 4.942373823661928e-07, | |
| "logits/chosen": -1.537407398223877, | |
| "logits/rejected": -1.352263331413269, | |
| "logps/chosen": -559.8363037109375, | |
| "logps/rejected": -603.9837646484375, | |
| "loss": 0.5263, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.3580124378204346, | |
| "rewards/margins": 1.0659737586975098, | |
| "rewards/rejected": -3.4239859580993652, | |
| "step": 1560 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 6.40625, | |
| "learning_rate": 4.6730129226114363e-07, | |
| "logits/chosen": -1.4590742588043213, | |
| "logits/rejected": -1.3711886405944824, | |
| "logps/chosen": -499.3367614746094, | |
| "logps/rejected": -608.9996337890625, | |
| "loss": 0.4878, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.5338335037231445, | |
| "rewards/margins": 0.9903131723403931, | |
| "rewards/rejected": -3.524146318435669, | |
| "step": 1570 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 6.28125, | |
| "learning_rate": 4.4104410701222703e-07, | |
| "logits/chosen": -1.4868284463882446, | |
| "logits/rejected": -1.4123101234436035, | |
| "logps/chosen": -530.2494506835938, | |
| "logps/rejected": -610.2762451171875, | |
| "loss": 0.5491, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.6920907497406006, | |
| "rewards/margins": 0.8161090612411499, | |
| "rewards/rejected": -3.508200168609619, | |
| "step": 1580 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 6.8125, | |
| "learning_rate": 4.154745963060197e-07, | |
| "logits/chosen": -1.4255352020263672, | |
| "logits/rejected": -1.3225051164627075, | |
| "logps/chosen": -488.10015869140625, | |
| "logps/rejected": -570.8845825195312, | |
| "loss": 0.4793, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.4687325954437256, | |
| "rewards/margins": 1.027092695236206, | |
| "rewards/rejected": -3.4958252906799316, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 7.84375, | |
| "learning_rate": 3.9060130015138863e-07, | |
| "logits/chosen": -1.4365253448486328, | |
| "logits/rejected": -1.4320509433746338, | |
| "logps/chosen": -510.7242126464844, | |
| "logps/rejected": -598.7286987304688, | |
| "loss": 0.5008, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.55981707572937, | |
| "rewards/margins": 0.9268480539321899, | |
| "rewards/rejected": -3.4866650104522705, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": -1.4415297508239746, | |
| "eval_logits/rejected": -1.3491688966751099, | |
| "eval_logps/chosen": -517.2503662109375, | |
| "eval_logps/rejected": -594.9102172851562, | |
| "eval_loss": 0.4976504147052765, | |
| "eval_rewards/accuracies": 0.7480000257492065, | |
| "eval_rewards/chosen": -2.4554741382598877, | |
| "eval_rewards/margins": 1.016426682472229, | |
| "eval_rewards/rejected": -3.471900701522827, | |
| "eval_runtime": 206.6827, | |
| "eval_samples_per_second": 9.677, | |
| "eval_steps_per_second": 0.605, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 6.5625, | |
| "learning_rate": 3.664325260271953e-07, | |
| "logits/chosen": -1.4340312480926514, | |
| "logits/rejected": -1.381696105003357, | |
| "logps/chosen": -529.0896606445312, | |
| "logps/rejected": -568.31787109375, | |
| "loss": 0.5177, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.3479690551757812, | |
| "rewards/margins": 0.7791107892990112, | |
| "rewards/rejected": -3.127079963684082, | |
| "step": 1610 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 6.6875, | |
| "learning_rate": 3.429763461076677e-07, | |
| "logits/chosen": -1.5787631273269653, | |
| "logits/rejected": -1.4651563167572021, | |
| "logps/chosen": -544.9623413085938, | |
| "logps/rejected": -591.4544677734375, | |
| "loss": 0.4948, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.447205066680908, | |
| "rewards/margins": 0.9208120107650757, | |
| "rewards/rejected": -3.3680167198181152, | |
| "step": 1620 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 6.34375, | |
| "learning_rate": 3.202405945663556e-07, | |
| "logits/chosen": -1.5388927459716797, | |
| "logits/rejected": -1.435225248336792, | |
| "logps/chosen": -527.3236694335938, | |
| "logps/rejected": -591.0799560546875, | |
| "loss": 0.4897, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.576470136642456, | |
| "rewards/margins": 0.910003662109375, | |
| "rewards/rejected": -3.486473798751831, | |
| "step": 1630 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 6.65625, | |
| "learning_rate": 2.982328649595856e-07, | |
| "logits/chosen": -1.4623661041259766, | |
| "logits/rejected": -1.418001651763916, | |
| "logps/chosen": -497.0453186035156, | |
| "logps/rejected": -582.2621459960938, | |
| "loss": 0.5222, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.3540682792663574, | |
| "rewards/margins": 0.9974225759506226, | |
| "rewards/rejected": -3.3514907360076904, | |
| "step": 1640 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 5.1875, | |
| "learning_rate": 2.7696050769026954e-07, | |
| "logits/chosen": -1.4713447093963623, | |
| "logits/rejected": -1.4519026279449463, | |
| "logps/chosen": -513.2701416015625, | |
| "logps/rejected": -621.521484375, | |
| "loss": 0.5271, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.4599874019622803, | |
| "rewards/margins": 1.0760111808776855, | |
| "rewards/rejected": -3.535998582839966, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 7.34375, | |
| "learning_rate": 2.564306275529341e-07, | |
| "logits/chosen": -1.4529814720153809, | |
| "logits/rejected": -1.3669501543045044, | |
| "logps/chosen": -488.9100646972656, | |
| "logps/rejected": -551.0786743164062, | |
| "loss": 0.5005, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.357375144958496, | |
| "rewards/margins": 0.9981120824813843, | |
| "rewards/rejected": -3.355487108230591, | |
| "step": 1660 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 7.34375, | |
| "learning_rate": 2.3665008136077332e-07, | |
| "logits/chosen": -1.577505350112915, | |
| "logits/rejected": -1.4699662923812866, | |
| "logps/chosen": -525.9560546875, | |
| "logps/rejected": -618.0115966796875, | |
| "loss": 0.4965, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.5162107944488525, | |
| "rewards/margins": 1.0870957374572754, | |
| "rewards/rejected": -3.603306293487549, | |
| "step": 1670 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 7.125, | |
| "learning_rate": 2.1762547565553293e-07, | |
| "logits/chosen": -1.4847347736358643, | |
| "logits/rejected": -1.3919851779937744, | |
| "logps/chosen": -512.091552734375, | |
| "logps/rejected": -616.5901489257812, | |
| "loss": 0.459, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -2.374174118041992, | |
| "rewards/margins": 1.1540172100067139, | |
| "rewards/rejected": -3.528191328048706, | |
| "step": 1680 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 7.03125, | |
| "learning_rate": 1.993631645009747e-07, | |
| "logits/chosen": -1.4473907947540283, | |
| "logits/rejected": -1.346949815750122, | |
| "logps/chosen": -485.31512451171875, | |
| "logps/rejected": -593.9915161132812, | |
| "loss": 0.4396, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -2.3521013259887695, | |
| "rewards/margins": 1.203741192817688, | |
| "rewards/rejected": -3.555842161178589, | |
| "step": 1690 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 6.84375, | |
| "learning_rate": 1.818692473606748e-07, | |
| "logits/chosen": -1.4863381385803223, | |
| "logits/rejected": -1.3555645942687988, | |
| "logps/chosen": -513.4937744140625, | |
| "logps/rejected": -581.9398803710938, | |
| "loss": 0.4654, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -2.393968105316162, | |
| "rewards/margins": 1.0673789978027344, | |
| "rewards/rejected": -3.4613468647003174, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "eval_logits/chosen": -1.440211534500122, | |
| "eval_logits/rejected": -1.347786784172058, | |
| "eval_logps/chosen": -516.6851806640625, | |
| "eval_logps/rejected": -594.441650390625, | |
| "eval_loss": 0.49758249521255493, | |
| "eval_rewards/accuracies": 0.7509999871253967, | |
| "eval_rewards/chosen": -2.449821710586548, | |
| "eval_rewards/margins": 1.0173932313919067, | |
| "eval_rewards/rejected": -3.467214822769165, | |
| "eval_runtime": 206.7023, | |
| "eval_samples_per_second": 9.676, | |
| "eval_steps_per_second": 0.605, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 8.8125, | |
| "learning_rate": 1.6514956706084885e-07, | |
| "logits/chosen": -1.5661511421203613, | |
| "logits/rejected": -1.444165825843811, | |
| "logps/chosen": -521.919189453125, | |
| "logps/rejected": -586.3562622070312, | |
| "loss": 0.4662, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.438523769378662, | |
| "rewards/margins": 1.0026253461837769, | |
| "rewards/rejected": -3.4411492347717285, | |
| "step": 1710 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 6.75, | |
| "learning_rate": 1.4920970783889737e-07, | |
| "logits/chosen": -1.4804189205169678, | |
| "logits/rejected": -1.3565007448196411, | |
| "logps/chosen": -517.4879150390625, | |
| "logps/rejected": -590.8859252929688, | |
| "loss": 0.4738, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.4466044902801514, | |
| "rewards/margins": 0.9578853845596313, | |
| "rewards/rejected": -3.4044899940490723, | |
| "step": 1720 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 7.4375, | |
| "learning_rate": 1.340549934783164e-07, | |
| "logits/chosen": -1.4345219135284424, | |
| "logits/rejected": -1.398148775100708, | |
| "logps/chosen": -510.354736328125, | |
| "logps/rejected": -606.5594482421875, | |
| "loss": 0.473, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.461632490158081, | |
| "rewards/margins": 1.1440181732177734, | |
| "rewards/rejected": -3.6056504249572754, | |
| "step": 1730 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 5.3125, | |
| "learning_rate": 1.196904855305961e-07, | |
| "logits/chosen": -1.4843945503234863, | |
| "logits/rejected": -1.3764415979385376, | |
| "logps/chosen": -511.3872985839844, | |
| "logps/rejected": -572.77783203125, | |
| "loss": 0.5174, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.429408311843872, | |
| "rewards/margins": 1.009290099143982, | |
| "rewards/rejected": -3.4386982917785645, | |
| "step": 1740 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 5.3125, | |
| "learning_rate": 1.0612098162470302e-07, | |
| "logits/chosen": -1.5372560024261475, | |
| "logits/rejected": -1.3980759382247925, | |
| "logps/chosen": -542.35400390625, | |
| "logps/rejected": -590.6364135742188, | |
| "loss": 0.5084, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.490891933441162, | |
| "rewards/margins": 0.8898002505302429, | |
| "rewards/rejected": -3.38069224357605, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 5.4375, | |
| "learning_rate": 9.335101386471285e-08, | |
| "logits/chosen": -1.5029523372650146, | |
| "logits/rejected": -1.4057238101959229, | |
| "logps/chosen": -491.0376892089844, | |
| "logps/rejected": -596.9283447265625, | |
| "loss": 0.4623, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.436763286590576, | |
| "rewards/margins": 1.0951788425445557, | |
| "rewards/rejected": -3.5319418907165527, | |
| "step": 1760 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 6.3125, | |
| "learning_rate": 8.138484731612273e-08, | |
| "logits/chosen": -1.4329261779785156, | |
| "logits/rejected": -1.3377002477645874, | |
| "logps/chosen": -494.4960021972656, | |
| "logps/rejected": -598.6796264648438, | |
| "loss": 0.4788, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.4506049156188965, | |
| "rewards/margins": 1.1435714960098267, | |
| "rewards/rejected": -3.5941765308380127, | |
| "step": 1770 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 4.75, | |
| "learning_rate": 7.022647858135501e-08, | |
| "logits/chosen": -1.4736502170562744, | |
| "logits/rejected": -1.451183557510376, | |
| "logps/chosen": -493.92706298828125, | |
| "logps/rejected": -590.5465087890625, | |
| "loss": 0.5134, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.3842227458953857, | |
| "rewards/margins": 1.1362553834915161, | |
| "rewards/rejected": -3.5204784870147705, | |
| "step": 1780 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 7.5625, | |
| "learning_rate": 5.987963446492384e-08, | |
| "logits/chosen": -1.4801084995269775, | |
| "logits/rejected": -1.4299700260162354, | |
| "logps/chosen": -557.2908325195312, | |
| "logps/rejected": -669.438720703125, | |
| "loss": 0.4906, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.4419429302215576, | |
| "rewards/margins": 1.1474285125732422, | |
| "rewards/rejected": -3.5893714427948, | |
| "step": 1790 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 6.90625, | |
| "learning_rate": 5.034777072871394e-08, | |
| "logits/chosen": -1.4195510149002075, | |
| "logits/rejected": -1.336871862411499, | |
| "logps/chosen": -495.6595153808594, | |
| "logps/rejected": -588.1336669921875, | |
| "loss": 0.4854, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.4461569786071777, | |
| "rewards/margins": 1.0101310014724731, | |
| "rewards/rejected": -3.4562880992889404, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "eval_logits/chosen": -1.4365895986557007, | |
| "eval_logits/rejected": -1.34409499168396, | |
| "eval_logps/chosen": -516.9639892578125, | |
| "eval_logps/rejected": -595.033935546875, | |
| "eval_loss": 0.49747607111930847, | |
| "eval_rewards/accuracies": 0.7480000257492065, | |
| "eval_rewards/chosen": -2.4526102542877197, | |
| "eval_rewards/margins": 1.0205274820327759, | |
| "eval_rewards/rejected": -3.473137855529785, | |
| "eval_runtime": 206.9741, | |
| "eval_samples_per_second": 9.663, | |
| "eval_steps_per_second": 0.604, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 6.21875, | |
| "learning_rate": 4.163407093778243e-08, | |
| "logits/chosen": -1.4668951034545898, | |
| "logits/rejected": -1.3559987545013428, | |
| "logps/chosen": -506.38153076171875, | |
| "logps/rejected": -566.1290283203125, | |
| "loss": 0.5032, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -2.4424426555633545, | |
| "rewards/margins": 1.029581904411316, | |
| "rewards/rejected": -3.472024440765381, | |
| "step": 1810 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 7.46875, | |
| "learning_rate": 3.37414453970758e-08, | |
| "logits/chosen": -1.5458420515060425, | |
| "logits/rejected": -1.3794844150543213, | |
| "logps/chosen": -514.1174926757812, | |
| "logps/rejected": -563.9821166992188, | |
| "loss": 0.5023, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.361403226852417, | |
| "rewards/margins": 1.0842626094818115, | |
| "rewards/rejected": -3.4456658363342285, | |
| "step": 1820 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 7.75, | |
| "learning_rate": 2.6672530179410183e-08, | |
| "logits/chosen": -1.4805978536605835, | |
| "logits/rejected": -1.4541041851043701, | |
| "logps/chosen": -495.7486877441406, | |
| "logps/rejected": -565.6021118164062, | |
| "loss": 0.5193, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.410522699356079, | |
| "rewards/margins": 0.8961440324783325, | |
| "rewards/rejected": -3.306666612625122, | |
| "step": 1830 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 8.125, | |
| "learning_rate": 2.04296862450451e-08, | |
| "logits/chosen": -1.4646061658859253, | |
| "logits/rejected": -1.4149842262268066, | |
| "logps/chosen": -559.3178100585938, | |
| "logps/rejected": -612.7825927734375, | |
| "loss": 0.4901, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.692322254180908, | |
| "rewards/margins": 0.8889071345329285, | |
| "rewards/rejected": -3.5812296867370605, | |
| "step": 1840 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 5.6875, | |
| "learning_rate": 1.501499865314171e-08, | |
| "logits/chosen": -1.4032434225082397, | |
| "logits/rejected": -1.3378710746765137, | |
| "logps/chosen": -570.8396606445312, | |
| "logps/rejected": -627.8958129882812, | |
| "loss": 0.4925, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.5015716552734375, | |
| "rewards/margins": 0.943825364112854, | |
| "rewards/rejected": -3.445397138595581, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 6.21875, | |
| "learning_rate": 1.0430275865371265e-08, | |
| "logits/chosen": -1.3695303201675415, | |
| "logits/rejected": -1.251265525817871, | |
| "logps/chosen": -500.8570251464844, | |
| "logps/rejected": -570.6333618164062, | |
| "loss": 0.4912, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.518383502960205, | |
| "rewards/margins": 1.0404582023620605, | |
| "rewards/rejected": -3.5588417053222656, | |
| "step": 1860 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 8.8125, | |
| "learning_rate": 6.677049141901315e-09, | |
| "logits/chosen": -1.4154746532440186, | |
| "logits/rejected": -1.3926751613616943, | |
| "logps/chosen": -495.06915283203125, | |
| "logps/rejected": -597.0353393554688, | |
| "loss": 0.4822, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -2.457148313522339, | |
| "rewards/margins": 0.9585596919059753, | |
| "rewards/rejected": -3.415708065032959, | |
| "step": 1870 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 5.3125, | |
| "learning_rate": 3.756572029968708e-09, | |
| "logits/chosen": -1.5453914403915405, | |
| "logits/rejected": -1.4504231214523315, | |
| "logps/chosen": -514.4291381835938, | |
| "logps/rejected": -585.8607177734375, | |
| "loss": 0.4866, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.3940634727478027, | |
| "rewards/margins": 1.0368168354034424, | |
| "rewards/rejected": -3.430880308151245, | |
| "step": 1880 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 6.1875, | |
| "learning_rate": 1.6698199452053199e-09, | |
| "logits/chosen": -1.4837844371795654, | |
| "logits/rejected": -1.3669326305389404, | |
| "logps/chosen": -526.9951171875, | |
| "logps/rejected": -609.9420166015625, | |
| "loss": 0.4737, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.426424503326416, | |
| "rewards/margins": 1.0135810375213623, | |
| "rewards/rejected": -3.4400055408477783, | |
| "step": 1890 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 6.875, | |
| "learning_rate": 4.1748984585560094e-10, | |
| "logits/chosen": -1.4921762943267822, | |
| "logits/rejected": -1.3556934595108032, | |
| "logps/chosen": -516.9928588867188, | |
| "logps/rejected": -590.8782958984375, | |
| "loss": 0.4879, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.5034995079040527, | |
| "rewards/margins": 1.0504939556121826, | |
| "rewards/rejected": -3.5539937019348145, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "eval_logits/chosen": -1.4358515739440918, | |
| "eval_logits/rejected": -1.3432255983352661, | |
| "eval_logps/chosen": -517.0148315429688, | |
| "eval_logps/rejected": -595.1220703125, | |
| "eval_loss": 0.49738776683807373, | |
| "eval_rewards/accuracies": 0.75, | |
| "eval_rewards/chosen": -2.4531185626983643, | |
| "eval_rewards/margins": 1.020900845527649, | |
| "eval_rewards/rejected": -3.474019765853882, | |
| "eval_runtime": 206.2288, | |
| "eval_samples_per_second": 9.698, | |
| "eval_steps_per_second": 0.606, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 6.90625, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -1.5044982433319092, | |
| "logits/rejected": -1.4231712818145752, | |
| "logps/chosen": -519.6868896484375, | |
| "logps/rejected": -603.6512451171875, | |
| "loss": 0.4895, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.5049242973327637, | |
| "rewards/margins": 0.8985089063644409, | |
| "rewards/rejected": -3.403432846069336, | |
| "step": 1910 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 1910, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5312882837824796, | |
| "train_runtime": 19005.9968, | |
| "train_samples_per_second": 3.217, | |
| "train_steps_per_second": 0.1 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1910, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |