| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 100, | |
| "global_step": 1751, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 20.572983553522633, | |
| "learning_rate": 2.8409090909090907e-09, | |
| "logits/chosen": -2.7078514099121094, | |
| "logits/rejected": -2.673313856124878, | |
| "logps/chosen": -322.9836120605469, | |
| "logps/rejected": -261.1998291015625, | |
| "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": 19.20958843122381, | |
| "learning_rate": 2.8409090909090908e-08, | |
| "logits/chosen": -2.7531583309173584, | |
| "logits/rejected": -2.715681314468384, | |
| "logps/chosen": -219.50958251953125, | |
| "logps/rejected": -236.27330017089844, | |
| "loss": 0.6934, | |
| "rewards/accuracies": 0.4166666567325592, | |
| "rewards/chosen": -0.0001647783356020227, | |
| "rewards/margins": 6.51524678687565e-05, | |
| "rewards/rejected": -0.00022993076709099114, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 18.221209446502318, | |
| "learning_rate": 5.6818181818181815e-08, | |
| "logits/chosen": -2.736729145050049, | |
| "logits/rejected": -2.7084593772888184, | |
| "logps/chosen": -233.8996124267578, | |
| "logps/rejected": -230.8451385498047, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.00037134578451514244, | |
| "rewards/margins": 2.871244305424625e-06, | |
| "rewards/rejected": 0.000368474458809942, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 19.659186304289324, | |
| "learning_rate": 8.522727272727271e-08, | |
| "logits/chosen": -2.718043804168701, | |
| "logits/rejected": -2.705275535583496, | |
| "logps/chosen": -265.68658447265625, | |
| "logps/rejected": -247.42025756835938, | |
| "loss": 0.6926, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": 0.0039044865407049656, | |
| "rewards/margins": 0.0005032892222516239, | |
| "rewards/rejected": 0.003401197027415037, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 21.248085751648183, | |
| "learning_rate": 1.1363636363636363e-07, | |
| "logits/chosen": -2.722376585006714, | |
| "logits/rejected": -2.7113113403320312, | |
| "logps/chosen": -207.11141967773438, | |
| "logps/rejected": -212.8821563720703, | |
| "loss": 0.6916, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 0.013140101917088032, | |
| "rewards/margins": 0.0030325299594551325, | |
| "rewards/rejected": 0.010107570327818394, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 17.062348015914782, | |
| "learning_rate": 1.4204545454545455e-07, | |
| "logits/chosen": -2.7086739540100098, | |
| "logits/rejected": -2.6903626918792725, | |
| "logps/chosen": -230.5471954345703, | |
| "logps/rejected": -218.68508911132812, | |
| "loss": 0.6905, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": 0.030233308672904968, | |
| "rewards/margins": 0.0065773241221904755, | |
| "rewards/rejected": 0.023655982688069344, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 16.966597240337634, | |
| "learning_rate": 1.7045454545454543e-07, | |
| "logits/chosen": -2.7187132835388184, | |
| "logits/rejected": -2.707646608352661, | |
| "logps/chosen": -231.4257354736328, | |
| "logps/rejected": -222.29745483398438, | |
| "loss": 0.6856, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.05241071432828903, | |
| "rewards/margins": 0.018025759607553482, | |
| "rewards/rejected": 0.03438495472073555, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 19.376648396399435, | |
| "learning_rate": 1.9886363636363636e-07, | |
| "logits/chosen": -2.7627768516540527, | |
| "logits/rejected": -2.699197292327881, | |
| "logps/chosen": -230.32540893554688, | |
| "logps/rejected": -220.99795532226562, | |
| "loss": 0.6812, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.08621755987405777, | |
| "rewards/margins": 0.024967864155769348, | |
| "rewards/rejected": 0.06124969199299812, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 16.574962507205655, | |
| "learning_rate": 2.2727272727272726e-07, | |
| "logits/chosen": -2.7578330039978027, | |
| "logits/rejected": -2.758152723312378, | |
| "logps/chosen": -241.5107879638672, | |
| "logps/rejected": -234.1169891357422, | |
| "loss": 0.6775, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": 0.1304091513156891, | |
| "rewards/margins": 0.03304018825292587, | |
| "rewards/rejected": 0.09736897051334381, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 17.123245271646947, | |
| "learning_rate": 2.5568181818181816e-07, | |
| "logits/chosen": -2.747803211212158, | |
| "logits/rejected": -2.714414119720459, | |
| "logps/chosen": -245.60464477539062, | |
| "logps/rejected": -236.7899932861328, | |
| "loss": 0.6688, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.12192847579717636, | |
| "rewards/margins": 0.04756360873579979, | |
| "rewards/rejected": 0.07436486333608627, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 17.870065861056936, | |
| "learning_rate": 2.840909090909091e-07, | |
| "logits/chosen": -2.7195897102355957, | |
| "logits/rejected": -2.681842803955078, | |
| "logps/chosen": -207.11083984375, | |
| "logps/rejected": -211.34866333007812, | |
| "loss": 0.6628, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.1526772379875183, | |
| "rewards/margins": 0.07189060747623444, | |
| "rewards/rejected": 0.08078663051128387, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "eval_logits/chosen": -2.7152342796325684, | |
| "eval_logits/rejected": -2.6720974445343018, | |
| "eval_logps/chosen": -221.00881958007812, | |
| "eval_logps/rejected": -221.34713745117188, | |
| "eval_loss": 0.6611106395721436, | |
| "eval_rewards/accuracies": 0.6317307949066162, | |
| "eval_rewards/chosen": 0.13367876410484314, | |
| "eval_rewards/margins": 0.08476302027702332, | |
| "eval_rewards/rejected": 0.04891572147607803, | |
| "eval_runtime": 864.7473, | |
| "eval_samples_per_second": 14.398, | |
| "eval_steps_per_second": 0.451, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 17.758173761785073, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.745452642440796, | |
| "logits/rejected": -2.6345908641815186, | |
| "logps/chosen": -211.13497924804688, | |
| "logps/rejected": -245.9041290283203, | |
| "loss": 0.6526, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": 0.10004226118326187, | |
| "rewards/margins": 0.09651143848896027, | |
| "rewards/rejected": 0.0035308226943016052, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 25.3503988826482, | |
| "learning_rate": 3.4090909090909085e-07, | |
| "logits/chosen": -2.6189653873443604, | |
| "logits/rejected": -2.57344388961792, | |
| "logps/chosen": -228.3636932373047, | |
| "logps/rejected": -230.4072265625, | |
| "loss": 0.6476, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.04373621195554733, | |
| "rewards/margins": 0.13258467614650726, | |
| "rewards/rejected": -0.176320880651474, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 25.5123273990743, | |
| "learning_rate": 3.693181818181818e-07, | |
| "logits/chosen": -2.564441204071045, | |
| "logits/rejected": -2.547898292541504, | |
| "logps/chosen": -269.93511962890625, | |
| "logps/rejected": -285.6426086425781, | |
| "loss": 0.6484, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.3578580319881439, | |
| "rewards/margins": 0.16583284735679626, | |
| "rewards/rejected": -0.5236908793449402, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 25.02924344762446, | |
| "learning_rate": 3.977272727272727e-07, | |
| "logits/chosen": -2.5508384704589844, | |
| "logits/rejected": -2.5335073471069336, | |
| "logps/chosen": -274.27508544921875, | |
| "logps/rejected": -272.3787536621094, | |
| "loss": 0.6395, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.3012983798980713, | |
| "rewards/margins": 0.09775562584400177, | |
| "rewards/rejected": -0.3990539610385895, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 23.226996578001714, | |
| "learning_rate": 4.2613636363636364e-07, | |
| "logits/chosen": -2.5063862800598145, | |
| "logits/rejected": -2.4907989501953125, | |
| "logps/chosen": -242.54183959960938, | |
| "logps/rejected": -267.40216064453125, | |
| "loss": 0.6367, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.11261695623397827, | |
| "rewards/margins": 0.19061629474163055, | |
| "rewards/rejected": -0.30323326587677, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 22.32459928056343, | |
| "learning_rate": 4.545454545454545e-07, | |
| "logits/chosen": -2.538865327835083, | |
| "logits/rejected": -2.4915454387664795, | |
| "logps/chosen": -259.973876953125, | |
| "logps/rejected": -284.99273681640625, | |
| "loss": 0.6342, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.06355403363704681, | |
| "rewards/margins": 0.3071693778038025, | |
| "rewards/rejected": -0.3707233965396881, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 28.238806692314206, | |
| "learning_rate": 4.829545454545455e-07, | |
| "logits/chosen": -2.5641493797302246, | |
| "logits/rejected": -2.5063588619232178, | |
| "logps/chosen": -256.0674133300781, | |
| "logps/rejected": -281.95965576171875, | |
| "loss": 0.6207, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.11485638469457626, | |
| "rewards/margins": 0.2587049603462219, | |
| "rewards/rejected": -0.3735613226890564, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 34.83431654717648, | |
| "learning_rate": 4.999920426892062e-07, | |
| "logits/chosen": -2.6064653396606445, | |
| "logits/rejected": -2.600895404815674, | |
| "logps/chosen": -245.4825897216797, | |
| "logps/rejected": -257.6805725097656, | |
| "loss": 0.6398, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.1361248791217804, | |
| "rewards/margins": 0.16745266318321228, | |
| "rewards/rejected": -0.3035775423049927, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 33.80086885729117, | |
| "learning_rate": 4.999025287600885e-07, | |
| "logits/chosen": -2.6627895832061768, | |
| "logits/rejected": -2.6619114875793457, | |
| "logps/chosen": -263.55096435546875, | |
| "logps/rejected": -284.45147705078125, | |
| "loss": 0.6325, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.09487941861152649, | |
| "rewards/margins": 0.21393874287605286, | |
| "rewards/rejected": -0.3088182210922241, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 29.828595993243468, | |
| "learning_rate": 4.997135899956001e-07, | |
| "logits/chosen": -2.638547420501709, | |
| "logits/rejected": -2.5978102684020996, | |
| "logps/chosen": -245.3714599609375, | |
| "logps/rejected": -266.0381164550781, | |
| "loss": 0.6203, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.1466042697429657, | |
| "rewards/margins": 0.28931504487991333, | |
| "rewards/rejected": -0.43591928482055664, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "eval_logits/chosen": -2.6775360107421875, | |
| "eval_logits/rejected": -2.621267080307007, | |
| "eval_logps/chosen": -243.9757843017578, | |
| "eval_logps/rejected": -266.8084411621094, | |
| "eval_loss": 0.6120733618736267, | |
| "eval_rewards/accuracies": 0.660897433757782, | |
| "eval_rewards/chosen": -0.09599092602729797, | |
| "eval_rewards/margins": 0.30970630049705505, | |
| "eval_rewards/rejected": -0.405697226524353, | |
| "eval_runtime": 841.2052, | |
| "eval_samples_per_second": 14.801, | |
| "eval_steps_per_second": 0.464, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 31.207584928179323, | |
| "learning_rate": 4.994253015658708e-07, | |
| "logits/chosen": -2.6432251930236816, | |
| "logits/rejected": -2.6140708923339844, | |
| "logps/chosen": -232.1490936279297, | |
| "logps/rejected": -261.6481018066406, | |
| "loss": 0.6111, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.03973131626844406, | |
| "rewards/margins": 0.33994999527931213, | |
| "rewards/rejected": -0.3796812891960144, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 23.510226746553357, | |
| "learning_rate": 4.990377781677378e-07, | |
| "logits/chosen": -2.674729108810425, | |
| "logits/rejected": -2.621265172958374, | |
| "logps/chosen": -272.51715087890625, | |
| "logps/rejected": -285.8130187988281, | |
| "loss": 0.6261, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.12226327508687973, | |
| "rewards/margins": 0.3179982602596283, | |
| "rewards/rejected": -0.4402615427970886, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 34.69513987573851, | |
| "learning_rate": 4.985511739791128e-07, | |
| "logits/chosen": -2.682669162750244, | |
| "logits/rejected": -2.6345534324645996, | |
| "logps/chosen": -243.78695678710938, | |
| "logps/rejected": -252.6505889892578, | |
| "loss": 0.5961, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.19093939661979675, | |
| "rewards/margins": 0.23702657222747803, | |
| "rewards/rejected": -0.4279659688472748, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 33.90649114668316, | |
| "learning_rate": 4.979656825976425e-07, | |
| "logits/chosen": -2.636639356613159, | |
| "logits/rejected": -2.556272029876709, | |
| "logps/chosen": -267.24310302734375, | |
| "logps/rejected": -295.2035217285156, | |
| "loss": 0.6078, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.3182388246059418, | |
| "rewards/margins": 0.24964241683483124, | |
| "rewards/rejected": -0.5678812861442566, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 27.80845453447812, | |
| "learning_rate": 4.972815369636832e-07, | |
| "logits/chosen": -2.6218130588531494, | |
| "logits/rejected": -2.55488920211792, | |
| "logps/chosen": -299.002197265625, | |
| "logps/rejected": -309.2835388183594, | |
| "loss": 0.6176, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.2978210151195526, | |
| "rewards/margins": 0.2504158914089203, | |
| "rewards/rejected": -0.5482369661331177, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 28.138070133686533, | |
| "learning_rate": 4.964990092676262e-07, | |
| "logits/chosen": -2.6748335361480713, | |
| "logits/rejected": -2.593806028366089, | |
| "logps/chosen": -268.3004455566406, | |
| "logps/rejected": -270.9266662597656, | |
| "loss": 0.6223, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.1147528737783432, | |
| "rewards/margins": 0.2601880431175232, | |
| "rewards/rejected": -0.3749409317970276, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 28.844690039420406, | |
| "learning_rate": 4.956184108416046e-07, | |
| "logits/chosen": -2.7262961864471436, | |
| "logits/rejected": -2.741664409637451, | |
| "logps/chosen": -238.9697723388672, | |
| "logps/rejected": -277.444091796875, | |
| "loss": 0.5909, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.17726224660873413, | |
| "rewards/margins": 0.36171460151672363, | |
| "rewards/rejected": -0.538976788520813, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 34.00184014415281, | |
| "learning_rate": 4.946400920356287e-07, | |
| "logits/chosen": -2.754479169845581, | |
| "logits/rejected": -2.707367181777954, | |
| "logps/chosen": -268.70452880859375, | |
| "logps/rejected": -305.3086853027344, | |
| "loss": 0.5949, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.30848780274391174, | |
| "rewards/margins": 0.48010724782943726, | |
| "rewards/rejected": -0.7885950803756714, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 24.974211972857105, | |
| "learning_rate": 4.935644420781978e-07, | |
| "logits/chosen": -2.6929306983947754, | |
| "logits/rejected": -2.6737914085388184, | |
| "logps/chosen": -283.9693603515625, | |
| "logps/rejected": -305.68646240234375, | |
| "loss": 0.5971, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.3218172490596771, | |
| "rewards/margins": 0.37671035528182983, | |
| "rewards/rejected": -0.6985276341438293, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 22.031451800060307, | |
| "learning_rate": 4.923918889214436e-07, | |
| "logits/chosen": -2.8449320793151855, | |
| "logits/rejected": -2.802635908126831, | |
| "logps/chosen": -249.634033203125, | |
| "logps/rejected": -294.66973876953125, | |
| "loss": 0.6134, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.1441958099603653, | |
| "rewards/margins": 0.37521737813949585, | |
| "rewards/rejected": -0.5194131731987, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "eval_logits/chosen": -2.8550686836242676, | |
| "eval_logits/rejected": -2.79876971244812, | |
| "eval_logps/chosen": -240.60450744628906, | |
| "eval_logps/rejected": -263.5723571777344, | |
| "eval_loss": 0.6074427962303162, | |
| "eval_rewards/accuracies": 0.670192301273346, | |
| "eval_rewards/chosen": -0.06227818876504898, | |
| "eval_rewards/margins": 0.31105825304985046, | |
| "eval_rewards/rejected": -0.37333643436431885, | |
| "eval_runtime": 861.7371, | |
| "eval_samples_per_second": 14.449, | |
| "eval_steps_per_second": 0.453, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 22.548306029706836, | |
| "learning_rate": 4.91122899070868e-07, | |
| "logits/chosen": -2.88305401802063, | |
| "logits/rejected": -2.8267903327941895, | |
| "logps/chosen": -253.9397735595703, | |
| "logps/rejected": -268.42498779296875, | |
| "loss": 0.608, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.02261662669479847, | |
| "rewards/margins": 0.35017159581184387, | |
| "rewards/rejected": -0.3727882504463196, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 26.91606740249196, | |
| "learning_rate": 4.897579773997414e-07, | |
| "logits/chosen": -2.7912163734436035, | |
| "logits/rejected": -2.748124599456787, | |
| "logps/chosen": -260.5610046386719, | |
| "logps/rejected": -272.39898681640625, | |
| "loss": 0.6085, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.2463793307542801, | |
| "rewards/margins": 0.2676694393157959, | |
| "rewards/rejected": -0.5140487551689148, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 25.26707067068886, | |
| "learning_rate": 4.882976669482367e-07, | |
| "logits/chosen": -2.6727445125579834, | |
| "logits/rejected": -2.5968549251556396, | |
| "logps/chosen": -253.74581909179688, | |
| "logps/rejected": -296.7026672363281, | |
| "loss": 0.6019, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.2068738043308258, | |
| "rewards/margins": 0.3490857481956482, | |
| "rewards/rejected": -0.5559597015380859, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 35.48578116044424, | |
| "learning_rate": 4.867425487073786e-07, | |
| "logits/chosen": -2.649488925933838, | |
| "logits/rejected": -2.6318719387054443, | |
| "logps/chosen": -241.77029418945312, | |
| "logps/rejected": -246.7797088623047, | |
| "loss": 0.629, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.01744382455945015, | |
| "rewards/margins": 0.21225237846374512, | |
| "rewards/rejected": -0.22969619929790497, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 32.561918770233014, | |
| "learning_rate": 4.850932413878934e-07, | |
| "logits/chosen": -2.5814261436462402, | |
| "logits/rejected": -2.5403079986572266, | |
| "logps/chosen": -251.0198516845703, | |
| "logps/rejected": -265.89190673828125, | |
| "loss": 0.6163, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.004257055930793285, | |
| "rewards/margins": 0.2787313163280487, | |
| "rewards/rejected": -0.28298836946487427, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 32.76257794045328, | |
| "learning_rate": 4.833504011740522e-07, | |
| "logits/chosen": -2.493020534515381, | |
| "logits/rejected": -2.4786593914031982, | |
| "logps/chosen": -239.0045623779297, | |
| "logps/rejected": -267.716552734375, | |
| "loss": 0.5962, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.0083407461643219, | |
| "rewards/margins": 0.33076852560043335, | |
| "rewards/rejected": -0.33910924196243286, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 25.42313532237448, | |
| "learning_rate": 4.815147214626056e-07, | |
| "logits/chosen": -2.403029680252075, | |
| "logits/rejected": -2.375685930252075, | |
| "logps/chosen": -286.8995666503906, | |
| "logps/rejected": -282.9924621582031, | |
| "loss": 0.5961, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.26247408986091614, | |
| "rewards/margins": 0.31958019733428955, | |
| "rewards/rejected": -0.5820542573928833, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 29.80215735650837, | |
| "learning_rate": 4.795869325869116e-07, | |
| "logits/chosen": -2.416861057281494, | |
| "logits/rejected": -2.4126358032226562, | |
| "logps/chosen": -263.8486022949219, | |
| "logps/rejected": -291.6919250488281, | |
| "loss": 0.6312, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.2886821925640106, | |
| "rewards/margins": 0.29902181029319763, | |
| "rewards/rejected": -0.5877040028572083, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 22.149160653773386, | |
| "learning_rate": 4.775678015263708e-07, | |
| "logits/chosen": -2.5736076831817627, | |
| "logits/rejected": -2.5460524559020996, | |
| "logps/chosen": -223.2324981689453, | |
| "logps/rejected": -245.54092407226562, | |
| "loss": 0.6081, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.09774526208639145, | |
| "rewards/margins": 0.38662397861480713, | |
| "rewards/rejected": -0.2888787090778351, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 28.055639097927603, | |
| "learning_rate": 4.7545813160127845e-07, | |
| "logits/chosen": -2.6072590351104736, | |
| "logits/rejected": -2.505096912384033, | |
| "logps/chosen": -236.22549438476562, | |
| "logps/rejected": -265.6544494628906, | |
| "loss": 0.5967, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.02556512877345085, | |
| "rewards/margins": 0.45649051666259766, | |
| "rewards/rejected": -0.4309254288673401, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "eval_logits/chosen": -2.519050121307373, | |
| "eval_logits/rejected": -2.4576427936553955, | |
| "eval_logps/chosen": -247.5235595703125, | |
| "eval_logps/rejected": -278.0496826171875, | |
| "eval_loss": 0.5991776585578918, | |
| "eval_rewards/accuracies": 0.678205132484436, | |
| "eval_rewards/chosen": -0.1314685195684433, | |
| "eval_rewards/margins": 0.38664132356643677, | |
| "eval_rewards/rejected": -0.5181097984313965, | |
| "eval_runtime": 861.5357, | |
| "eval_samples_per_second": 14.452, | |
| "eval_steps_per_second": 0.453, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 26.14630535783879, | |
| "learning_rate": 4.732587621532214e-07, | |
| "logits/chosen": -2.478102922439575, | |
| "logits/rejected": -2.438175678253174, | |
| "logps/chosen": -250.90457153320312, | |
| "logps/rejected": -284.3564453125, | |
| "loss": 0.5827, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.15595874190330505, | |
| "rewards/margins": 0.3831656873226166, | |
| "rewards/rejected": -0.5391243696212769, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 28.031713354054272, | |
| "learning_rate": 4.709705682111412e-07, | |
| "logits/chosen": -2.477914810180664, | |
| "logits/rejected": -2.4634792804718018, | |
| "logps/chosen": -262.4942321777344, | |
| "logps/rejected": -290.16009521484375, | |
| "loss": 0.5869, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.22279219329357147, | |
| "rewards/margins": 0.4437895715236664, | |
| "rewards/rejected": -0.6665817499160767, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 25.747403286133764, | |
| "learning_rate": 4.68594460143201e-07, | |
| "logits/chosen": -2.5019423961639404, | |
| "logits/rejected": -2.429969310760498, | |
| "logps/chosen": -256.59185791015625, | |
| "logps/rejected": -283.1378173828125, | |
| "loss": 0.5936, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.14861060678958893, | |
| "rewards/margins": 0.4206705093383789, | |
| "rewards/rejected": -0.5692810416221619, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 22.31072442106435, | |
| "learning_rate": 4.661313832945903e-07, | |
| "logits/chosen": -2.504993438720703, | |
| "logits/rejected": -2.467900037765503, | |
| "logps/chosen": -254.93057250976562, | |
| "logps/rejected": -312.25250244140625, | |
| "loss": 0.5925, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.15858420729637146, | |
| "rewards/margins": 0.5125812292098999, | |
| "rewards/rejected": -0.671165406703949, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 28.65488774525744, | |
| "learning_rate": 4.635823176114162e-07, | |
| "logits/chosen": -2.5073506832122803, | |
| "logits/rejected": -2.472411870956421, | |
| "logps/chosen": -284.4822998046875, | |
| "logps/rejected": -320.66925048828125, | |
| "loss": 0.5817, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.37685471773147583, | |
| "rewards/margins": 0.3860743045806885, | |
| "rewards/rejected": -0.7629289627075195, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 34.815824611343665, | |
| "learning_rate": 4.6094827725082684e-07, | |
| "logits/chosen": -2.5682735443115234, | |
| "logits/rejected": -2.501211166381836, | |
| "logps/chosen": -265.11126708984375, | |
| "logps/rejected": -309.05157470703125, | |
| "loss": 0.5814, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.2719506323337555, | |
| "rewards/margins": 0.5007379055023193, | |
| "rewards/rejected": -0.7726885080337524, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 26.238405483406037, | |
| "learning_rate": 4.582303101775248e-07, | |
| "logits/chosen": -2.62959623336792, | |
| "logits/rejected": -2.562079906463623, | |
| "logps/chosen": -251.88674926757812, | |
| "logps/rejected": -288.1053771972656, | |
| "loss": 0.5955, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.16994395852088928, | |
| "rewards/margins": 0.4775424599647522, | |
| "rewards/rejected": -0.6474864482879639, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 30.50178144367043, | |
| "learning_rate": 4.5542949774682956e-07, | |
| "logits/chosen": -2.568668842315674, | |
| "logits/rejected": -2.5604805946350098, | |
| "logps/chosen": -267.90655517578125, | |
| "logps/rejected": -290.76019287109375, | |
| "loss": 0.5991, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.22079896926879883, | |
| "rewards/margins": 0.31227344274520874, | |
| "rewards/rejected": -0.5330723524093628, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 27.535568193967393, | |
| "learning_rate": 4.5254695427445517e-07, | |
| "logits/chosen": -2.637202501296997, | |
| "logits/rejected": -2.6354382038116455, | |
| "logps/chosen": -244.74551391601562, | |
| "logps/rejected": -290.29644775390625, | |
| "loss": 0.5647, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.011266368441283703, | |
| "rewards/margins": 0.5976118445396423, | |
| "rewards/rejected": -0.5863455533981323, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 30.285379350904805, | |
| "learning_rate": 4.4958382659317536e-07, | |
| "logits/chosen": -2.740764617919922, | |
| "logits/rejected": -2.7210445404052734, | |
| "logps/chosen": -272.13702392578125, | |
| "logps/rejected": -297.4922790527344, | |
| "loss": 0.6216, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.258303165435791, | |
| "rewards/margins": 0.32254815101623535, | |
| "rewards/rejected": -0.5808513760566711, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "eval_logits/chosen": -2.7310774326324463, | |
| "eval_logits/rejected": -2.687927484512329, | |
| "eval_logps/chosen": -238.078125, | |
| "eval_logps/rejected": -267.6940002441406, | |
| "eval_loss": 0.5941099524497986, | |
| "eval_rewards/accuracies": 0.6721153855323792, | |
| "eval_rewards/chosen": -0.037014104425907135, | |
| "eval_rewards/margins": 0.37753891944885254, | |
| "eval_rewards/rejected": -0.4145529866218567, | |
| "eval_runtime": 858.9897, | |
| "eval_samples_per_second": 14.495, | |
| "eval_steps_per_second": 0.454, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 23.512447045863215, | |
| "learning_rate": 4.465412935965506e-07, | |
| "logits/chosen": -2.7190139293670654, | |
| "logits/rejected": -2.675178050994873, | |
| "logps/chosen": -217.1962432861328, | |
| "logps/rejected": -246.1102294921875, | |
| "loss": 0.6138, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.14496248960494995, | |
| "rewards/margins": 0.36382409930229187, | |
| "rewards/rejected": -0.21886160969734192, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 23.621464405914903, | |
| "learning_rate": 4.434205657698999e-07, | |
| "logits/chosen": -2.568403720855713, | |
| "logits/rejected": -2.551494836807251, | |
| "logps/chosen": -224.2484588623047, | |
| "logps/rejected": -278.65936279296875, | |
| "loss": 0.6125, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.11122129112482071, | |
| "rewards/margins": 0.36189281940460205, | |
| "rewards/rejected": -0.47311410307884216, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 23.359399767763318, | |
| "learning_rate": 4.402228847087046e-07, | |
| "logits/chosen": -2.5537118911743164, | |
| "logits/rejected": -2.502847194671631, | |
| "logps/chosen": -261.56927490234375, | |
| "logps/rejected": -274.5636291503906, | |
| "loss": 0.622, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.15581832826137543, | |
| "rewards/margins": 0.2899227440357208, | |
| "rewards/rejected": -0.44574108719825745, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 33.81353805062806, | |
| "learning_rate": 4.36949522624633e-07, | |
| "logits/chosen": -2.578953742980957, | |
| "logits/rejected": -2.547420024871826, | |
| "logps/chosen": -236.4100341796875, | |
| "logps/rejected": -284.703857421875, | |
| "loss": 0.5917, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.17350006103515625, | |
| "rewards/margins": 0.42752647399902344, | |
| "rewards/rejected": -0.6010265350341797, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 25.836827126785362, | |
| "learning_rate": 4.33601781839386e-07, | |
| "logits/chosen": -2.516505718231201, | |
| "logits/rejected": -2.4714202880859375, | |
| "logps/chosen": -271.7455139160156, | |
| "logps/rejected": -320.60003662109375, | |
| "loss": 0.583, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.1934550553560257, | |
| "rewards/margins": 0.46987539529800415, | |
| "rewards/rejected": -0.6633304357528687, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 29.317285148378573, | |
| "learning_rate": 4.301809942665625e-07, | |
| "logits/chosen": -2.5027713775634766, | |
| "logits/rejected": -2.4169793128967285, | |
| "logps/chosen": -247.04306030273438, | |
| "logps/rejected": -263.53033447265625, | |
| "loss": 0.6041, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.13370783627033234, | |
| "rewards/margins": 0.29323458671569824, | |
| "rewards/rejected": -0.4269424080848694, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 23.273024353193524, | |
| "learning_rate": 4.2668852088175145e-07, | |
| "logits/chosen": -2.429138660430908, | |
| "logits/rejected": -2.4112179279327393, | |
| "logps/chosen": -261.13232421875, | |
| "logps/rejected": -292.0643005371094, | |
| "loss": 0.613, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.12648805975914001, | |
| "rewards/margins": 0.3239407241344452, | |
| "rewards/rejected": -0.4504288136959076, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 27.396331045603723, | |
| "learning_rate": 4.231257511810618e-07, | |
| "logits/chosen": -2.55159068107605, | |
| "logits/rejected": -2.4876770973205566, | |
| "logps/chosen": -251.07192993164062, | |
| "logps/rejected": -293.2655334472656, | |
| "loss": 0.6128, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.010355165228247643, | |
| "rewards/margins": 0.33735883235931396, | |
| "rewards/rejected": -0.34771400690078735, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 22.123964853357307, | |
| "learning_rate": 4.1949410262830523e-07, | |
| "logits/chosen": -2.5525479316711426, | |
| "logits/rejected": -2.4439752101898193, | |
| "logps/chosen": -233.11337280273438, | |
| "logps/rejected": -262.0377502441406, | |
| "loss": 0.5785, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.024647405371069908, | |
| "rewards/margins": 0.3654420077800751, | |
| "rewards/rejected": -0.3900894522666931, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 29.777775051728057, | |
| "learning_rate": 4.1579502009105164e-07, | |
| "logits/chosen": -2.488654613494873, | |
| "logits/rejected": -2.4201626777648926, | |
| "logps/chosen": -246.08486938476562, | |
| "logps/rejected": -273.70709228515625, | |
| "loss": 0.5919, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.1482144594192505, | |
| "rewards/margins": 0.40206265449523926, | |
| "rewards/rejected": -0.5502771139144897, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "eval_logits/chosen": -2.47454833984375, | |
| "eval_logits/rejected": -2.404360294342041, | |
| "eval_logps/chosen": -249.46990966796875, | |
| "eval_logps/rejected": -283.90716552734375, | |
| "eval_loss": 0.5903669595718384, | |
| "eval_rewards/accuracies": 0.6865384578704834, | |
| "eval_rewards/chosen": -0.15093201398849487, | |
| "eval_rewards/margins": 0.42575252056121826, | |
| "eval_rewards/rejected": -0.5766844749450684, | |
| "eval_runtime": 858.5855, | |
| "eval_samples_per_second": 14.502, | |
| "eval_steps_per_second": 0.454, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 27.691202685227417, | |
| "learning_rate": 4.120299752657827e-07, | |
| "logits/chosen": -2.5120370388031006, | |
| "logits/rejected": -2.4610724449157715, | |
| "logps/chosen": -261.5104064941406, | |
| "logps/rejected": -296.66217041015625, | |
| "loss": 0.592, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.18553677201271057, | |
| "rewards/margins": 0.3792334496974945, | |
| "rewards/rejected": -0.5647702217102051, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 24.36661678943314, | |
| "learning_rate": 4.082004660923702e-07, | |
| "logits/chosen": -2.4631857872009277, | |
| "logits/rejected": -2.373511791229248, | |
| "logps/chosen": -248.58895874023438, | |
| "logps/rejected": -271.892333984375, | |
| "loss": 0.5945, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.06783659011125565, | |
| "rewards/margins": 0.3515480160713196, | |
| "rewards/rejected": -0.419384628534317, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 23.60175971446507, | |
| "learning_rate": 4.0430801615811437e-07, | |
| "logits/chosen": -2.4521324634552, | |
| "logits/rejected": -2.4140613079071045, | |
| "logps/chosen": -231.697021484375, | |
| "logps/rejected": -277.6011962890625, | |
| "loss": 0.5938, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.045631490647792816, | |
| "rewards/margins": 0.44818124175071716, | |
| "rewards/rejected": -0.4938127398490906, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 28.337765984985374, | |
| "learning_rate": 4.00354174091578e-07, | |
| "logits/chosen": -2.4540791511535645, | |
| "logits/rejected": -2.3845746517181396, | |
| "logps/chosen": -229.52133178710938, | |
| "logps/rejected": -277.82452392578125, | |
| "loss": 0.6079, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.0009828682523220778, | |
| "rewards/margins": 0.471419095993042, | |
| "rewards/rejected": -0.47240200638771057, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 27.293175260711642, | |
| "learning_rate": 3.963405129464569e-07, | |
| "logits/chosen": -2.3416523933410645, | |
| "logits/rejected": -2.244310140609741, | |
| "logps/chosen": -261.36871337890625, | |
| "logps/rejected": -316.395263671875, | |
| "loss": 0.5519, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.22408190369606018, | |
| "rewards/margins": 0.5848034024238586, | |
| "rewards/rejected": -0.8088852167129517, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 27.68918047728001, | |
| "learning_rate": 3.922686295757347e-07, | |
| "logits/chosen": -2.336780071258545, | |
| "logits/rejected": -2.2526936531066895, | |
| "logps/chosen": -301.7099609375, | |
| "logps/rejected": -356.14056396484375, | |
| "loss": 0.5933, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.42652687430381775, | |
| "rewards/margins": 0.5547691583633423, | |
| "rewards/rejected": -0.9812959432601929, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 24.193971795815184, | |
| "learning_rate": 3.881401439963666e-07, | |
| "logits/chosen": -2.272087574005127, | |
| "logits/rejected": -2.220580577850342, | |
| "logps/chosen": -229.0252227783203, | |
| "logps/rejected": -266.34405517578125, | |
| "loss": 0.5979, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.23234577476978302, | |
| "rewards/margins": 0.3420170247554779, | |
| "rewards/rejected": -0.5743628144264221, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 33.515555503395184, | |
| "learning_rate": 3.839566987447491e-07, | |
| "logits/chosen": -2.345500946044922, | |
| "logits/rejected": -2.2792842388153076, | |
| "logps/chosen": -268.81317138671875, | |
| "logps/rejected": -309.93463134765625, | |
| "loss": 0.6034, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.14895500242710114, | |
| "rewards/margins": 0.49111080169677734, | |
| "rewards/rejected": -0.6400657892227173, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 22.584232277933243, | |
| "learning_rate": 3.7971995822322947e-07, | |
| "logits/chosen": -2.27201247215271, | |
| "logits/rejected": -2.226531505584717, | |
| "logps/chosen": -262.95782470703125, | |
| "logps/rejected": -290.07696533203125, | |
| "loss": 0.6172, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.2456875741481781, | |
| "rewards/margins": 0.29335761070251465, | |
| "rewards/rejected": -0.5390451550483704, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 28.603695886249437, | |
| "learning_rate": 3.7543160803791545e-07, | |
| "logits/chosen": -2.301879644393921, | |
| "logits/rejected": -2.223996162414551, | |
| "logps/chosen": -261.08795166015625, | |
| "logps/rejected": -307.17010498046875, | |
| "loss": 0.5769, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.22030536830425262, | |
| "rewards/margins": 0.49444580078125, | |
| "rewards/rejected": -0.714751124382019, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_logits/chosen": -2.2923786640167236, | |
| "eval_logits/rejected": -2.2189748287200928, | |
| "eval_logps/chosen": -258.4495849609375, | |
| "eval_logps/rejected": -292.7128601074219, | |
| "eval_loss": 0.5901564955711365, | |
| "eval_rewards/accuracies": 0.6772435903549194, | |
| "eval_rewards/chosen": -0.24072882533073425, | |
| "eval_rewards/margins": 0.4240129590034485, | |
| "eval_rewards/rejected": -0.6647418141365051, | |
| "eval_runtime": 860.8912, | |
| "eval_samples_per_second": 14.463, | |
| "eval_steps_per_second": 0.453, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 26.031424540267988, | |
| "learning_rate": 3.7109335432805e-07, | |
| "logits/chosen": -2.2880516052246094, | |
| "logits/rejected": -2.229980707168579, | |
| "logps/chosen": -265.29754638671875, | |
| "logps/rejected": -275.8728942871094, | |
| "loss": 0.6037, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.11812520027160645, | |
| "rewards/margins": 0.3968770205974579, | |
| "rewards/rejected": -0.5150021910667419, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 22.719027068474563, | |
| "learning_rate": 3.667069230872155e-07, | |
| "logits/chosen": -2.254481554031372, | |
| "logits/rejected": -2.231606960296631, | |
| "logps/chosen": -259.39422607421875, | |
| "logps/rejected": -272.6926574707031, | |
| "loss": 0.5933, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.046945445239543915, | |
| "rewards/margins": 0.3208449184894562, | |
| "rewards/rejected": -0.3677903413772583, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 29.374605264798323, | |
| "learning_rate": 3.6227405947664003e-07, | |
| "logits/chosen": -2.2828311920166016, | |
| "logits/rejected": -2.2138302326202393, | |
| "logps/chosen": -256.0699768066406, | |
| "logps/rejected": -295.7370910644531, | |
| "loss": 0.5858, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.0820874497294426, | |
| "rewards/margins": 0.5297425985336304, | |
| "rewards/rejected": -0.6118300557136536, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 24.91144032228272, | |
| "learning_rate": 3.577965271308771e-07, | |
| "logits/chosen": -2.2585289478302, | |
| "logits/rejected": -2.201707124710083, | |
| "logps/chosen": -266.5726318359375, | |
| "logps/rejected": -301.1708679199219, | |
| "loss": 0.5891, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.2869926691055298, | |
| "rewards/margins": 0.4772910475730896, | |
| "rewards/rejected": -0.7642836570739746, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 31.59826426014115, | |
| "learning_rate": 3.5327610745613546e-07, | |
| "logits/chosen": -2.3542895317077637, | |
| "logits/rejected": -2.242069721221924, | |
| "logps/chosen": -250.84194946289062, | |
| "logps/rejected": -285.4515380859375, | |
| "loss": 0.5963, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.19332554936408997, | |
| "rewards/margins": 0.4167536795139313, | |
| "rewards/rejected": -0.6100791692733765, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 25.89736672152189, | |
| "learning_rate": 3.487145989215391e-07, | |
| "logits/chosen": -2.383197069168091, | |
| "logits/rejected": -2.3221888542175293, | |
| "logps/chosen": -251.72390747070312, | |
| "logps/rejected": -282.80804443359375, | |
| "loss": 0.5814, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.09259854257106781, | |
| "rewards/margins": 0.45896273851394653, | |
| "rewards/rejected": -0.5515612363815308, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 24.879610991566807, | |
| "learning_rate": 3.4411381634359796e-07, | |
| "logits/chosen": -2.389742612838745, | |
| "logits/rejected": -2.3619580268859863, | |
| "logps/chosen": -255.8960418701172, | |
| "logps/rejected": -290.5157165527344, | |
| "loss": 0.5646, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.1797618567943573, | |
| "rewards/margins": 0.44017988443374634, | |
| "rewards/rejected": -0.6199417114257812, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 26.915784595282116, | |
| "learning_rate": 3.3947559016417554e-07, | |
| "logits/chosen": -2.4469265937805176, | |
| "logits/rejected": -2.4010143280029297, | |
| "logps/chosen": -261.87322998046875, | |
| "logps/rejected": -302.0085754394531, | |
| "loss": 0.6045, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.26690053939819336, | |
| "rewards/margins": 0.33789390325546265, | |
| "rewards/rejected": -0.6047943830490112, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 27.76077206811895, | |
| "learning_rate": 3.3480176572223885e-07, | |
| "logits/chosen": -2.5314059257507324, | |
| "logits/rejected": -2.4485652446746826, | |
| "logps/chosen": -238.18112182617188, | |
| "logps/rejected": -297.2239074707031, | |
| "loss": 0.554, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.199807807803154, | |
| "rewards/margins": 0.6575660705566406, | |
| "rewards/rejected": -0.857373833656311, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 24.226296856947638, | |
| "learning_rate": 3.300942025196824e-07, | |
| "logits/chosen": -2.587006092071533, | |
| "logits/rejected": -2.541341781616211, | |
| "logps/chosen": -256.3536376953125, | |
| "logps/rejected": -278.5965270996094, | |
| "loss": 0.5725, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.13480545580387115, | |
| "rewards/margins": 0.4779321551322937, | |
| "rewards/rejected": -0.6127376556396484, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "eval_logits/chosen": -2.573244094848633, | |
| "eval_logits/rejected": -2.527627468109131, | |
| "eval_logps/chosen": -238.99398803710938, | |
| "eval_logps/rejected": -274.5382995605469, | |
| "eval_loss": 0.5882474780082703, | |
| "eval_rewards/accuracies": 0.6836538314819336, | |
| "eval_rewards/chosen": -0.04617273062467575, | |
| "eval_rewards/margins": 0.43682336807250977, | |
| "eval_rewards/rejected": -0.4829961061477661, | |
| "eval_runtime": 862.9985, | |
| "eval_samples_per_second": 14.428, | |
| "eval_steps_per_second": 0.452, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 26.141416208141568, | |
| "learning_rate": 3.253547734815173e-07, | |
| "logits/chosen": -2.607003688812256, | |
| "logits/rejected": -2.554546594619751, | |
| "logps/chosen": -232.52743530273438, | |
| "logps/rejected": -314.36346435546875, | |
| "loss": 0.5801, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.0894065648317337, | |
| "rewards/margins": 0.6495140790939331, | |
| "rewards/rejected": -0.5601075887680054, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 26.097129519417752, | |
| "learning_rate": 3.2058536421071914e-07, | |
| "logits/chosen": -2.5027403831481934, | |
| "logits/rejected": -2.471726417541504, | |
| "logps/chosen": -257.47320556640625, | |
| "logps/rejected": -263.6692810058594, | |
| "loss": 0.5805, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.03840960934758186, | |
| "rewards/margins": 0.3363339304924011, | |
| "rewards/rejected": -0.29792430996894836, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 25.23275414096734, | |
| "learning_rate": 3.1578787223803294e-07, | |
| "logits/chosen": -2.426281452178955, | |
| "logits/rejected": -2.399183750152588, | |
| "logps/chosen": -272.22869873046875, | |
| "logps/rejected": -290.59649658203125, | |
| "loss": 0.6043, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.32681387662887573, | |
| "rewards/margins": 0.41791266202926636, | |
| "rewards/rejected": -0.7447265386581421, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 24.11509059960286, | |
| "learning_rate": 3.109642062670318e-07, | |
| "logits/chosen": -2.5103323459625244, | |
| "logits/rejected": -2.4195475578308105, | |
| "logps/chosen": -244.50863647460938, | |
| "logps/rejected": -288.40069580078125, | |
| "loss": 0.5889, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.2207474410533905, | |
| "rewards/margins": 0.4590122699737549, | |
| "rewards/rejected": -0.6797597408294678, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 33.37568508664132, | |
| "learning_rate": 3.0611628541473074e-07, | |
| "logits/chosen": -2.497424602508545, | |
| "logits/rejected": -2.472369432449341, | |
| "logps/chosen": -266.05230712890625, | |
| "logps/rejected": -307.9107360839844, | |
| "loss": 0.6051, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.008460281416773796, | |
| "rewards/margins": 0.4078546166419983, | |
| "rewards/rejected": -0.39939433336257935, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 24.365317735651118, | |
| "learning_rate": 3.012460384480576e-07, | |
| "logits/chosen": -2.4374663829803467, | |
| "logits/rejected": -2.3599624633789062, | |
| "logps/chosen": -259.6045837402344, | |
| "logps/rejected": -302.10345458984375, | |
| "loss": 0.5678, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.18751467764377594, | |
| "rewards/margins": 0.49369174242019653, | |
| "rewards/rejected": -0.6812063455581665, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 22.6057869532, | |
| "learning_rate": 2.963554030164842e-07, | |
| "logits/chosen": -2.2620365619659424, | |
| "logits/rejected": -2.280235767364502, | |
| "logps/chosen": -270.291748046875, | |
| "logps/rejected": -298.3429870605469, | |
| "loss": 0.6028, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.1897084265947342, | |
| "rewards/margins": 0.3970777988433838, | |
| "rewards/rejected": -0.5867862105369568, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 26.035338300340943, | |
| "learning_rate": 2.9144632488112416e-07, | |
| "logits/chosen": -2.274094343185425, | |
| "logits/rejected": -2.2430529594421387, | |
| "logps/chosen": -270.6689758300781, | |
| "logps/rejected": -314.1228942871094, | |
| "loss": 0.5872, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.21771720051765442, | |
| "rewards/margins": 0.5008502006530762, | |
| "rewards/rejected": -0.7185673713684082, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 22.12211938779161, | |
| "learning_rate": 2.865207571406029e-07, | |
| "logits/chosen": -2.3526949882507324, | |
| "logits/rejected": -2.253375768661499, | |
| "logps/chosen": -269.4740905761719, | |
| "logps/rejected": -295.1158752441406, | |
| "loss": 0.5956, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.25454986095428467, | |
| "rewards/margins": 0.40211352705955505, | |
| "rewards/rejected": -0.6566634774208069, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 26.967953686731473, | |
| "learning_rate": 2.815806594540089e-07, | |
| "logits/chosen": -2.3986849784851074, | |
| "logits/rejected": -2.3849453926086426, | |
| "logps/chosen": -241.43505859375, | |
| "logps/rejected": -276.5702819824219, | |
| "loss": 0.5814, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.002503329422324896, | |
| "rewards/margins": 0.3857470154762268, | |
| "rewards/rejected": -0.38825035095214844, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "eval_logits/chosen": -2.409844160079956, | |
| "eval_logits/rejected": -2.335495948791504, | |
| "eval_logps/chosen": -246.1585693359375, | |
| "eval_logps/rejected": -279.9914245605469, | |
| "eval_loss": 0.5864266753196716, | |
| "eval_rewards/accuracies": 0.6810897588729858, | |
| "eval_rewards/chosen": -0.11781882494688034, | |
| "eval_rewards/margins": 0.41970840096473694, | |
| "eval_rewards/rejected": -0.5375272035598755, | |
| "eval_runtime": 861.4542, | |
| "eval_samples_per_second": 14.453, | |
| "eval_steps_per_second": 0.453, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 24.93686478364852, | |
| "learning_rate": 2.7662799726123454e-07, | |
| "logits/chosen": -2.364938974380493, | |
| "logits/rejected": -2.3040971755981445, | |
| "logps/chosen": -242.66366577148438, | |
| "logps/rejected": -277.5296936035156, | |
| "loss": 0.6005, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.25096192955970764, | |
| "rewards/margins": 0.2948233187198639, | |
| "rewards/rejected": -0.5457853078842163, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 29.741661581034126, | |
| "learning_rate": 2.7166474100101674e-07, | |
| "logits/chosen": -2.362717390060425, | |
| "logits/rejected": -2.29923677444458, | |
| "logps/chosen": -260.1564636230469, | |
| "logps/rejected": -299.52142333984375, | |
| "loss": 0.5972, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.2629412114620209, | |
| "rewards/margins": 0.4658204913139343, | |
| "rewards/rejected": -0.7287617325782776, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 23.70048734007165, | |
| "learning_rate": 2.6669286532698964e-07, | |
| "logits/chosen": -2.4182770252227783, | |
| "logits/rejected": -2.377297878265381, | |
| "logps/chosen": -272.5760192871094, | |
| "logps/rejected": -307.6293640136719, | |
| "loss": 0.5734, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.10112947225570679, | |
| "rewards/margins": 0.5127297639846802, | |
| "rewards/rejected": -0.6138591170310974, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 26.612920994033377, | |
| "learning_rate": 2.6171434832205953e-07, | |
| "logits/chosen": -2.4122025966644287, | |
| "logits/rejected": -2.3759803771972656, | |
| "logps/chosen": -256.7776184082031, | |
| "logps/rejected": -289.4403381347656, | |
| "loss": 0.5657, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.20392732322216034, | |
| "rewards/margins": 0.4127303957939148, | |
| "rewards/rejected": -0.6166577339172363, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 31.14720620364731, | |
| "learning_rate": 2.567311707114157e-07, | |
| "logits/chosen": -2.4095797538757324, | |
| "logits/rejected": -2.354799747467041, | |
| "logps/chosen": -229.2027130126953, | |
| "logps/rejected": -283.2561340332031, | |
| "loss": 0.604, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.197126567363739, | |
| "rewards/margins": 0.4597938060760498, | |
| "rewards/rejected": -0.6569203734397888, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 27.00204010664385, | |
| "learning_rate": 2.5174531507449037e-07, | |
| "logits/chosen": -2.4471428394317627, | |
| "logits/rejected": -2.338510513305664, | |
| "logps/chosen": -234.0675506591797, | |
| "logps/rejected": -259.7816467285156, | |
| "loss": 0.5899, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.190177783370018, | |
| "rewards/margins": 0.3870656490325928, | |
| "rewards/rejected": -0.577243447303772, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 19.932243905164242, | |
| "learning_rate": 2.4675876505618045e-07, | |
| "logits/chosen": -2.4684808254241943, | |
| "logits/rejected": -2.4519195556640625, | |
| "logps/chosen": -262.174560546875, | |
| "logps/rejected": -282.4520263671875, | |
| "loss": 0.6019, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.06612774729728699, | |
| "rewards/margins": 0.45484066009521484, | |
| "rewards/rejected": -0.5209684371948242, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 26.495466395218312, | |
| "learning_rate": 2.4177350457764527e-07, | |
| "logits/chosen": -2.53678822517395, | |
| "logits/rejected": -2.4713311195373535, | |
| "logps/chosen": -253.9574432373047, | |
| "logps/rejected": -298.145263671875, | |
| "loss": 0.6026, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.07538889348506927, | |
| "rewards/margins": 0.39861345291137695, | |
| "rewards/rejected": -0.47400230169296265, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 24.86682957992287, | |
| "learning_rate": 2.367915170469956e-07, | |
| "logits/chosen": -2.4735522270202637, | |
| "logits/rejected": -2.337954044342041, | |
| "logps/chosen": -256.18963623046875, | |
| "logps/rejected": -267.07305908203125, | |
| "loss": 0.5731, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.04039744660258293, | |
| "rewards/margins": 0.35975947976112366, | |
| "rewards/rejected": -0.4001569151878357, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 32.9954052513008, | |
| "learning_rate": 2.3181478457018475e-07, | |
| "logits/chosen": -2.4602394104003906, | |
| "logits/rejected": -2.372180700302124, | |
| "logps/chosen": -259.4634704589844, | |
| "logps/rejected": -298.13836669921875, | |
| "loss": 0.5514, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.039340369403362274, | |
| "rewards/margins": 0.5396673083305359, | |
| "rewards/rejected": -0.5790076851844788, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "eval_logits/chosen": -2.4855430126190186, | |
| "eval_logits/rejected": -2.411501169204712, | |
| "eval_logps/chosen": -252.6514892578125, | |
| "eval_logps/rejected": -291.29022216796875, | |
| "eval_loss": 0.5839190483093262, | |
| "eval_rewards/accuracies": 0.6871795058250427, | |
| "eval_rewards/chosen": -0.18274781107902527, | |
| "eval_rewards/margins": 0.4677676856517792, | |
| "eval_rewards/rejected": -0.6505153775215149, | |
| "eval_runtime": 844.1499, | |
| "eval_samples_per_second": 14.75, | |
| "eval_steps_per_second": 0.462, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 33.832396038055975, | |
| "learning_rate": 2.2684528716241968e-07, | |
| "logits/chosen": -2.528446674346924, | |
| "logits/rejected": -2.4585366249084473, | |
| "logps/chosen": -278.91461181640625, | |
| "logps/rejected": -327.67449951171875, | |
| "loss": 0.5971, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.20916755497455597, | |
| "rewards/margins": 0.4697844088077545, | |
| "rewards/rejected": -0.6789519190788269, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 23.559014104508705, | |
| "learning_rate": 2.218850019604028e-07, | |
| "logits/chosen": -2.5937182903289795, | |
| "logits/rejected": -2.506432056427002, | |
| "logps/chosen": -244.5588836669922, | |
| "logps/rejected": -312.0033264160156, | |
| "loss": 0.5622, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.12260422855615616, | |
| "rewards/margins": 0.6369132399559021, | |
| "rewards/rejected": -0.7595174908638, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 33.04728248607101, | |
| "learning_rate": 2.1693590243571935e-07, | |
| "logits/chosen": -2.641472816467285, | |
| "logits/rejected": -2.5865638256073, | |
| "logps/chosen": -279.6830139160156, | |
| "logps/rejected": -306.81243896484375, | |
| "loss": 0.5842, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.17163579165935516, | |
| "rewards/margins": 0.36918142437934875, | |
| "rewards/rejected": -0.5408172607421875, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 28.35991011420196, | |
| "learning_rate": 2.119999576096832e-07, | |
| "logits/chosen": -2.623781681060791, | |
| "logits/rejected": -2.5274269580841064, | |
| "logps/chosen": -278.32916259765625, | |
| "logps/rejected": -306.47198486328125, | |
| "loss": 0.5772, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.28789210319519043, | |
| "rewards/margins": 0.4908284544944763, | |
| "rewards/rejected": -0.7787205576896667, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 24.13700126033885, | |
| "learning_rate": 2.0707913126995244e-07, | |
| "logits/chosen": -2.555406093597412, | |
| "logits/rejected": -2.5062708854675293, | |
| "logps/chosen": -302.45416259765625, | |
| "logps/rejected": -314.2410583496094, | |
| "loss": 0.579, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.26258963346481323, | |
| "rewards/margins": 0.5782099962234497, | |
| "rewards/rejected": -0.8407996296882629, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 31.0299041852118, | |
| "learning_rate": 2.021753811892281e-07, | |
| "logits/chosen": -2.591373920440674, | |
| "logits/rejected": -2.5653815269470215, | |
| "logps/chosen": -267.46551513671875, | |
| "logps/rejected": -289.0950012207031, | |
| "loss": 0.5749, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.2024902105331421, | |
| "rewards/margins": 0.5493400692939758, | |
| "rewards/rejected": -0.7518302202224731, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 27.46997764843006, | |
| "learning_rate": 1.9729065834634528e-07, | |
| "logits/chosen": -2.5627617835998535, | |
| "logits/rejected": -2.4920713901519775, | |
| "logps/chosen": -259.22216796875, | |
| "logps/rejected": -298.4840393066406, | |
| "loss": 0.5842, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.045971281826496124, | |
| "rewards/margins": 0.47002631425857544, | |
| "rewards/rejected": -0.515997588634491, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 23.969955157596903, | |
| "learning_rate": 1.924269061500678e-07, | |
| "logits/chosen": -2.5113027095794678, | |
| "logits/rejected": -2.500351667404175, | |
| "logps/chosen": -237.98489379882812, | |
| "logps/rejected": -280.7731628417969, | |
| "loss": 0.5878, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.1136646643280983, | |
| "rewards/margins": 0.36606425046920776, | |
| "rewards/rejected": -0.4797288477420807, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 23.516344403891786, | |
| "learning_rate": 1.8758605966589434e-07, | |
| "logits/chosen": -2.5660672187805176, | |
| "logits/rejected": -2.5005881786346436, | |
| "logps/chosen": -275.2914733886719, | |
| "logps/rejected": -298.65155029296875, | |
| "loss": 0.5762, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.09098437428474426, | |
| "rewards/margins": 0.4184696078300476, | |
| "rewards/rejected": -0.32748526334762573, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 29.404877835754245, | |
| "learning_rate": 1.8277004484618357e-07, | |
| "logits/chosen": -2.5443501472473145, | |
| "logits/rejected": -2.4919300079345703, | |
| "logps/chosen": -245.58584594726562, | |
| "logps/rejected": -276.38427734375, | |
| "loss": 0.5946, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.019266966730356216, | |
| "rewards/margins": 0.3828735649585724, | |
| "rewards/rejected": -0.4021405279636383, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": -2.508995532989502, | |
| "eval_logits/rejected": -2.447495937347412, | |
| "eval_logps/chosen": -241.06724548339844, | |
| "eval_logps/rejected": -277.4429626464844, | |
| "eval_loss": 0.5846326351165771, | |
| "eval_rewards/accuracies": 0.6846153736114502, | |
| "eval_rewards/chosen": -0.06690550595521927, | |
| "eval_rewards/margins": 0.4451369047164917, | |
| "eval_rewards/rejected": -0.5120424032211304, | |
| "eval_runtime": 860.4121, | |
| "eval_samples_per_second": 14.471, | |
| "eval_steps_per_second": 0.453, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 28.170030221996313, | |
| "learning_rate": 1.7798077776390603e-07, | |
| "logits/chosen": -2.497800827026367, | |
| "logits/rejected": -2.438910722732544, | |
| "logps/chosen": -266.95526123046875, | |
| "logps/rejected": -271.0229187011719, | |
| "loss": 0.5981, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.20956678688526154, | |
| "rewards/margins": 0.3521861732006073, | |
| "rewards/rejected": -0.5617530345916748, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 25.64342328633492, | |
| "learning_rate": 1.7322016385032536e-07, | |
| "logits/chosen": -2.4583446979522705, | |
| "logits/rejected": -2.400876522064209, | |
| "logps/chosen": -265.85760498046875, | |
| "logps/rejected": -305.25830078125, | |
| "loss": 0.5854, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.011677969247102737, | |
| "rewards/margins": 0.538432240486145, | |
| "rewards/rejected": -0.5501102209091187, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 29.29633909055343, | |
| "learning_rate": 1.6849009713691454e-07, | |
| "logits/chosen": -2.458902597427368, | |
| "logits/rejected": -2.4362378120422363, | |
| "logps/chosen": -273.21710205078125, | |
| "logps/rejected": -307.6544494628906, | |
| "loss": 0.5831, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.21327920258045197, | |
| "rewards/margins": 0.44302114844322205, | |
| "rewards/rejected": -0.65630042552948, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 29.210383152583212, | |
| "learning_rate": 1.6379245950180666e-07, | |
| "logits/chosen": -2.4900028705596924, | |
| "logits/rejected": -2.45780611038208, | |
| "logps/chosen": -270.41424560546875, | |
| "logps/rejected": -287.94708251953125, | |
| "loss": 0.5648, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.17546476423740387, | |
| "rewards/margins": 0.5030435919761658, | |
| "rewards/rejected": -0.6785083413124084, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 25.427016062934616, | |
| "learning_rate": 1.591291199210813e-07, | |
| "logits/chosen": -2.519883871078491, | |
| "logits/rejected": -2.4721555709838867, | |
| "logps/chosen": -239.39413452148438, | |
| "logps/rejected": -290.2775573730469, | |
| "loss": 0.5746, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.21366102993488312, | |
| "rewards/margins": 0.43268680572509766, | |
| "rewards/rejected": -0.6463478207588196, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 27.841026179227654, | |
| "learning_rate": 1.5450193372518438e-07, | |
| "logits/chosen": -2.5061633586883545, | |
| "logits/rejected": -2.446307420730591, | |
| "logps/chosen": -254.13290405273438, | |
| "logps/rejected": -282.93585205078125, | |
| "loss": 0.5938, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.22907939553260803, | |
| "rewards/margins": 0.40173858404159546, | |
| "rewards/rejected": -0.6308180093765259, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 23.462003347624545, | |
| "learning_rate": 1.4991274186077628e-07, | |
| "logits/chosen": -2.528425931930542, | |
| "logits/rejected": -2.4988036155700684, | |
| "logps/chosen": -256.7923583984375, | |
| "logps/rejected": -287.1715393066406, | |
| "loss": 0.5873, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.17279884219169617, | |
| "rewards/margins": 0.4650405943393707, | |
| "rewards/rejected": -0.6378394365310669, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 29.33113372811353, | |
| "learning_rate": 1.4536337015830323e-07, | |
| "logits/chosen": -2.567692279815674, | |
| "logits/rejected": -2.534449815750122, | |
| "logps/chosen": -267.5567626953125, | |
| "logps/rejected": -302.30535888671875, | |
| "loss": 0.5849, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.20980314910411835, | |
| "rewards/margins": 0.41117554903030396, | |
| "rewards/rejected": -0.6209786534309387, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 27.190688664117623, | |
| "learning_rate": 1.4085562860558253e-07, | |
| "logits/chosen": -2.571300506591797, | |
| "logits/rejected": -2.5311543941497803, | |
| "logps/chosen": -264.1678161621094, | |
| "logps/rejected": -294.2384338378906, | |
| "loss": 0.581, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.21912555396556854, | |
| "rewards/margins": 0.4142730236053467, | |
| "rewards/rejected": -0.6333985924720764, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 29.84363481456159, | |
| "learning_rate": 1.3639131062769075e-07, | |
| "logits/chosen": -2.495572805404663, | |
| "logits/rejected": -2.4769346714019775, | |
| "logps/chosen": -259.7516784667969, | |
| "logps/rejected": -298.76519775390625, | |
| "loss": 0.5988, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.23917154967784882, | |
| "rewards/margins": 0.4562154710292816, | |
| "rewards/rejected": -0.695387065410614, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "eval_logits/chosen": -2.5292820930480957, | |
| "eval_logits/rejected": -2.470348596572876, | |
| "eval_logps/chosen": -261.1407775878906, | |
| "eval_logps/rejected": -299.3863525390625, | |
| "eval_loss": 0.5829187631607056, | |
| "eval_rewards/accuracies": 0.6891025900840759, | |
| "eval_rewards/chosen": -0.2676407992839813, | |
| "eval_rewards/margins": 0.4638356566429138, | |
| "eval_rewards/rejected": -0.7314765453338623, | |
| "eval_runtime": 847.4482, | |
| "eval_samples_per_second": 14.692, | |
| "eval_steps_per_second": 0.46, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 28.419557011703112, | |
| "learning_rate": 1.3197219237344155e-07, | |
| "logits/chosen": -2.5275566577911377, | |
| "logits/rejected": -2.50520396232605, | |
| "logps/chosen": -279.53619384765625, | |
| "logps/rejected": -297.91168212890625, | |
| "loss": 0.5781, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.25856587290763855, | |
| "rewards/margins": 0.46807146072387695, | |
| "rewards/rejected": -0.7266373038291931, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 23.724168571838458, | |
| "learning_rate": 1.2760003200873697e-07, | |
| "logits/chosen": -2.5259270668029785, | |
| "logits/rejected": -2.463974714279175, | |
| "logps/chosen": -274.4045715332031, | |
| "logps/rejected": -303.4898986816406, | |
| "loss": 0.5929, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.40674060583114624, | |
| "rewards/margins": 0.42905139923095703, | |
| "rewards/rejected": -0.8357919454574585, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 25.666665518773463, | |
| "learning_rate": 1.2327656901707338e-07, | |
| "logits/chosen": -2.522529125213623, | |
| "logits/rejected": -2.4855215549468994, | |
| "logps/chosen": -291.68292236328125, | |
| "logps/rejected": -332.4886474609375, | |
| "loss": 0.5913, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.3141830861568451, | |
| "rewards/margins": 0.4157239496707916, | |
| "rewards/rejected": -0.7299070358276367, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 27.19771337622201, | |
| "learning_rate": 1.1900352350748024e-07, | |
| "logits/chosen": -2.5396928787231445, | |
| "logits/rejected": -2.527820587158203, | |
| "logps/chosen": -299.08056640625, | |
| "logps/rejected": -341.8636474609375, | |
| "loss": 0.5916, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.3272027373313904, | |
| "rewards/margins": 0.37227770686149597, | |
| "rewards/rejected": -0.6994804739952087, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 33.33117238316685, | |
| "learning_rate": 1.147825955301668e-07, | |
| "logits/chosen": -2.533803939819336, | |
| "logits/rejected": -2.4939873218536377, | |
| "logps/chosen": -256.9107666015625, | |
| "logps/rejected": -292.6522521972656, | |
| "loss": 0.581, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.3282284736633301, | |
| "rewards/margins": 0.4669713079929352, | |
| "rewards/rejected": -0.7951998114585876, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 29.14283955300301, | |
| "learning_rate": 1.1061546440014963e-07, | |
| "logits/chosen": -2.5241713523864746, | |
| "logits/rejected": -2.5124003887176514, | |
| "logps/chosen": -256.1329040527344, | |
| "logps/rejected": -269.87127685546875, | |
| "loss": 0.6125, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.24249228835105896, | |
| "rewards/margins": 0.24842050671577454, | |
| "rewards/rejected": -0.4909127652645111, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 25.742222165117763, | |
| "learning_rate": 1.0650378802913007e-07, | |
| "logits/chosen": -2.4792118072509766, | |
| "logits/rejected": -2.428273916244507, | |
| "logps/chosen": -239.7839813232422, | |
| "logps/rejected": -273.16070556640625, | |
| "loss": 0.5855, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.14450299739837646, | |
| "rewards/margins": 0.43993139266967773, | |
| "rewards/rejected": -0.5844343900680542, | |
| "step": 1270 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 23.149905324082404, | |
| "learning_rate": 1.0244920226588597e-07, | |
| "logits/chosen": -2.5420844554901123, | |
| "logits/rejected": -2.4969589710235596, | |
| "logps/chosen": -246.9677734375, | |
| "logps/rejected": -282.594482421875, | |
| "loss": 0.5884, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.17258282005786896, | |
| "rewards/margins": 0.34093397855758667, | |
| "rewards/rejected": -0.5135167837142944, | |
| "step": 1280 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 24.301291223143128, | |
| "learning_rate": 9.845332024544245e-08, | |
| "logits/chosen": -2.5490660667419434, | |
| "logits/rejected": -2.5159246921539307, | |
| "logps/chosen": -253.81149291992188, | |
| "logps/rejected": -271.77716064453125, | |
| "loss": 0.5834, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.09992814064025879, | |
| "rewards/margins": 0.4652518332004547, | |
| "rewards/rejected": -0.5651799440383911, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 23.714606581482208, | |
| "learning_rate": 9.451773174727873e-08, | |
| "logits/chosen": -2.517934799194336, | |
| "logits/rejected": -2.467642068862915, | |
| "logps/chosen": -253.8499298095703, | |
| "logps/rejected": -283.5979919433594, | |
| "loss": 0.5725, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.028648119419813156, | |
| "rewards/margins": 0.5222169756889343, | |
| "rewards/rejected": -0.5508650541305542, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "eval_logits/chosen": -2.5130560398101807, | |
| "eval_logits/rejected": -2.4589791297912598, | |
| "eval_logps/chosen": -245.44595336914062, | |
| "eval_logps/rejected": -282.7961120605469, | |
| "eval_loss": 0.580857515335083, | |
| "eval_rewards/accuracies": 0.6878204941749573, | |
| "eval_rewards/chosen": -0.11069269478321075, | |
| "eval_rewards/margins": 0.4548812806606293, | |
| "eval_rewards/rejected": -0.5655739903450012, | |
| "eval_runtime": 863.5784, | |
| "eval_samples_per_second": 14.418, | |
| "eval_steps_per_second": 0.452, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 25.51529829677871, | |
| "learning_rate": 9.064400256282755e-08, | |
| "logits/chosen": -2.497252941131592, | |
| "logits/rejected": -2.4359519481658936, | |
| "logps/chosen": -246.2664794921875, | |
| "logps/rejected": -289.52105712890625, | |
| "loss": 0.565, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.18726077675819397, | |
| "rewards/margins": 0.4956758916378021, | |
| "rewards/rejected": -0.6829366087913513, | |
| "step": 1310 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 28.395653589692625, | |
| "learning_rate": 8.68336738725177e-08, | |
| "logits/chosen": -2.5088109970092773, | |
| "logits/rejected": -2.444235324859619, | |
| "logps/chosen": -260.43255615234375, | |
| "logps/rejected": -314.78466796875, | |
| "loss": 0.585, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.09720905870199203, | |
| "rewards/margins": 0.5907143950462341, | |
| "rewards/rejected": -0.6879234910011292, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 34.946146652842195, | |
| "learning_rate": 8.308826163260852e-08, | |
| "logits/chosen": -2.4821484088897705, | |
| "logits/rejected": -2.437180995941162, | |
| "logps/chosen": -284.7655029296875, | |
| "logps/rejected": -339.27410888671875, | |
| "loss": 0.6013, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.38431376218795776, | |
| "rewards/margins": 0.5592700242996216, | |
| "rewards/rejected": -0.9435837864875793, | |
| "step": 1330 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 27.1693846661215, | |
| "learning_rate": 7.940925597206052e-08, | |
| "logits/chosen": -2.5242748260498047, | |
| "logits/rejected": -2.423187732696533, | |
| "logps/chosen": -240.61532592773438, | |
| "logps/rejected": -303.8311462402344, | |
| "loss": 0.5867, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.17216572165489197, | |
| "rewards/margins": 0.4809741973876953, | |
| "rewards/rejected": -0.6531398892402649, | |
| "step": 1340 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 26.85515693070903, | |
| "learning_rate": 7.579812059968014e-08, | |
| "logits/chosen": -2.468845844268799, | |
| "logits/rejected": -2.4025533199310303, | |
| "logps/chosen": -239.42636108398438, | |
| "logps/rejected": -292.7440490722656, | |
| "loss": 0.5718, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.1654459834098816, | |
| "rewards/margins": 0.45221590995788574, | |
| "rewards/rejected": -0.6176619529724121, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 29.04379120857399, | |
| "learning_rate": 7.225629222177715e-08, | |
| "logits/chosen": -2.4990878105163574, | |
| "logits/rejected": -2.483397960662842, | |
| "logps/chosen": -273.4325256347656, | |
| "logps/rejected": -326.87835693359375, | |
| "loss": 0.5592, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.2638954222202301, | |
| "rewards/margins": 0.4499826431274414, | |
| "rewards/rejected": -0.7138780355453491, | |
| "step": 1360 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 27.003396904472925, | |
| "learning_rate": 6.878517997056457e-08, | |
| "logits/chosen": -2.5113110542297363, | |
| "logits/rejected": -2.4679648876190186, | |
| "logps/chosen": -231.7012176513672, | |
| "logps/rejected": -295.1663818359375, | |
| "loss": 0.5706, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.17551007866859436, | |
| "rewards/margins": 0.5275388956069946, | |
| "rewards/rejected": -0.7030488848686218, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 25.178752513980715, | |
| "learning_rate": 6.538616484352902e-08, | |
| "logits/chosen": -2.5119760036468506, | |
| "logits/rejected": -2.473975658416748, | |
| "logps/chosen": -261.88665771484375, | |
| "logps/rejected": -301.8829040527344, | |
| "loss": 0.5784, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.27380552887916565, | |
| "rewards/margins": 0.4960559010505676, | |
| "rewards/rejected": -0.7698613405227661, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 31.62767752494405, | |
| "learning_rate": 6.206059915399495e-08, | |
| "logits/chosen": -2.475964307785034, | |
| "logits/rejected": -2.415445566177368, | |
| "logps/chosen": -300.7445373535156, | |
| "logps/rejected": -290.59259033203125, | |
| "loss": 0.5803, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.21556460857391357, | |
| "rewards/margins": 0.4194510579109192, | |
| "rewards/rejected": -0.6350156664848328, | |
| "step": 1390 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 26.11221463277773, | |
| "learning_rate": 5.88098059931004e-08, | |
| "logits/chosen": -2.513415575027466, | |
| "logits/rejected": -2.431800603866577, | |
| "logps/chosen": -248.53909301757812, | |
| "logps/rejected": -316.1845703125, | |
| "loss": 0.5719, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.17952445149421692, | |
| "rewards/margins": 0.6487424373626709, | |
| "rewards/rejected": -0.8282669186592102, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_logits/chosen": -2.5095772743225098, | |
| "eval_logits/rejected": -2.4584505558013916, | |
| "eval_logps/chosen": -255.48683166503906, | |
| "eval_logps/rejected": -296.0592346191406, | |
| "eval_loss": 0.5792786478996277, | |
| "eval_rewards/accuracies": 0.6894230842590332, | |
| "eval_rewards/chosen": -0.2111014872789383, | |
| "eval_rewards/margins": 0.4871039092540741, | |
| "eval_rewards/rejected": -0.6982054114341736, | |
| "eval_runtime": 847.3731, | |
| "eval_samples_per_second": 14.694, | |
| "eval_steps_per_second": 0.46, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 31.913715183238843, | |
| "learning_rate": 5.563507870339962e-08, | |
| "logits/chosen": -2.4840493202209473, | |
| "logits/rejected": -2.45074725151062, | |
| "logps/chosen": -297.8506164550781, | |
| "logps/rejected": -323.21722412109375, | |
| "loss": 0.5745, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.21714219450950623, | |
| "rewards/margins": 0.49252939224243164, | |
| "rewards/rejected": -0.7096715569496155, | |
| "step": 1410 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 35.605251227958966, | |
| "learning_rate": 5.25376803643007e-08, | |
| "logits/chosen": -2.5487353801727295, | |
| "logits/rejected": -2.4819111824035645, | |
| "logps/chosen": -277.7652282714844, | |
| "logps/rejected": -307.17413330078125, | |
| "loss": 0.5827, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.3039516806602478, | |
| "rewards/margins": 0.4392912983894348, | |
| "rewards/rejected": -0.7432429194450378, | |
| "step": 1420 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 28.285683945551348, | |
| "learning_rate": 4.9518843289544007e-08, | |
| "logits/chosen": -2.5319647789001465, | |
| "logits/rejected": -2.4716343879699707, | |
| "logps/chosen": -246.0936737060547, | |
| "logps/rejected": -295.38726806640625, | |
| "loss": 0.5892, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.1578165590763092, | |
| "rewards/margins": 0.518903911113739, | |
| "rewards/rejected": -0.6767204999923706, | |
| "step": 1430 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 29.113838409092516, | |
| "learning_rate": 4.6579768536920695e-08, | |
| "logits/chosen": -2.5386886596679688, | |
| "logits/rejected": -2.4646518230438232, | |
| "logps/chosen": -255.16152954101562, | |
| "logps/rejected": -311.1474609375, | |
| "loss": 0.6026, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.15220266580581665, | |
| "rewards/margins": 0.46218329668045044, | |
| "rewards/rejected": -0.6143860220909119, | |
| "step": 1440 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 33.086643583472295, | |
| "learning_rate": 4.372162543042623e-08, | |
| "logits/chosen": -2.506279230117798, | |
| "logits/rejected": -2.4417097568511963, | |
| "logps/chosen": -250.19287109375, | |
| "logps/rejected": -287.5571594238281, | |
| "loss": 0.5846, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.07686599344015121, | |
| "rewards/margins": 0.5035240650177002, | |
| "rewards/rejected": -0.5803900957107544, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 22.051531350612663, | |
| "learning_rate": 4.094555109503983e-08, | |
| "logits/chosen": -2.5418009757995605, | |
| "logits/rejected": -2.4571533203125, | |
| "logps/chosen": -257.3257141113281, | |
| "logps/rejected": -294.6969299316406, | |
| "loss": 0.5669, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.17704418301582336, | |
| "rewards/margins": 0.46817469596862793, | |
| "rewards/rejected": -0.6452188491821289, | |
| "step": 1460 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 31.588020383365453, | |
| "learning_rate": 3.825265000431424e-08, | |
| "logits/chosen": -2.513741970062256, | |
| "logits/rejected": -2.4840691089630127, | |
| "logps/chosen": -260.6844177246094, | |
| "logps/rejected": -297.4983825683594, | |
| "loss": 0.5591, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.2122991979122162, | |
| "rewards/margins": 0.5046205520629883, | |
| "rewards/rejected": -0.7169197797775269, | |
| "step": 1470 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 27.48498825785638, | |
| "learning_rate": 3.5643993540955844e-08, | |
| "logits/chosen": -2.5381698608398438, | |
| "logits/rejected": -2.4838833808898926, | |
| "logps/chosen": -266.23724365234375, | |
| "logps/rejected": -313.0609130859375, | |
| "loss": 0.5712, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.22639349102973938, | |
| "rewards/margins": 0.47134050726890564, | |
| "rewards/rejected": -0.697733998298645, | |
| "step": 1480 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 29.66621598419702, | |
| "learning_rate": 3.312061957057061e-08, | |
| "logits/chosen": -2.5433754920959473, | |
| "logits/rejected": -2.4874513149261475, | |
| "logps/chosen": -253.4493408203125, | |
| "logps/rejected": -313.81622314453125, | |
| "loss": 0.5552, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.29640263319015503, | |
| "rewards/margins": 0.5009024739265442, | |
| "rewards/rejected": -0.797305166721344, | |
| "step": 1490 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 25.013543267548187, | |
| "learning_rate": 3.0683532028744184e-08, | |
| "logits/chosen": -2.5175490379333496, | |
| "logits/rejected": -2.4559388160705566, | |
| "logps/chosen": -250.25332641601562, | |
| "logps/rejected": -292.8772888183594, | |
| "loss": 0.5702, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.2427365779876709, | |
| "rewards/margins": 0.4707626700401306, | |
| "rewards/rejected": -0.7134992480278015, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "eval_logits/chosen": -2.5242793560028076, | |
| "eval_logits/rejected": -2.4745876789093018, | |
| "eval_logps/chosen": -261.0100402832031, | |
| "eval_logps/rejected": -301.7152099609375, | |
| "eval_loss": 0.5788725018501282, | |
| "eval_rewards/accuracies": 0.6887820363044739, | |
| "eval_rewards/chosen": -0.26633328199386597, | |
| "eval_rewards/margins": 0.4884316027164459, | |
| "eval_rewards/rejected": -0.754764974117279, | |
| "eval_runtime": 847.3963, | |
| "eval_samples_per_second": 14.693, | |
| "eval_steps_per_second": 0.46, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 27.931120203439026, | |
| "learning_rate": 2.8333700521621955e-08, | |
| "logits/chosen": -2.4960639476776123, | |
| "logits/rejected": -2.465576648712158, | |
| "logps/chosen": -262.0352478027344, | |
| "logps/rejected": -281.86077880859375, | |
| "loss": 0.5803, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.2684056758880615, | |
| "rewards/margins": 0.4416535496711731, | |
| "rewards/rejected": -0.7100592255592346, | |
| "step": 1510 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 33.57688090254581, | |
| "learning_rate": 2.6072059940146772e-08, | |
| "logits/chosen": -2.545090675354004, | |
| "logits/rejected": -2.5021095275878906, | |
| "logps/chosen": -262.41546630859375, | |
| "logps/rejected": -307.7872009277344, | |
| "loss": 0.5798, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.2850339114665985, | |
| "rewards/margins": 0.5246847867965698, | |
| "rewards/rejected": -0.8097187280654907, | |
| "step": 1520 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 25.889341574243268, | |
| "learning_rate": 2.3899510088108587e-08, | |
| "logits/chosen": -2.5121829509735107, | |
| "logits/rejected": -2.485520839691162, | |
| "logps/chosen": -253.7327117919922, | |
| "logps/rejected": -294.1796569824219, | |
| "loss": 0.5486, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.231918066740036, | |
| "rewards/margins": 0.5059068202972412, | |
| "rewards/rejected": -0.7378247976303101, | |
| "step": 1530 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 25.634449566248733, | |
| "learning_rate": 2.1816915324153334e-08, | |
| "logits/chosen": -2.541168212890625, | |
| "logits/rejected": -2.4927926063537598, | |
| "logps/chosen": -277.5816650390625, | |
| "logps/rejected": -327.8016662597656, | |
| "loss": 0.5588, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.2811830937862396, | |
| "rewards/margins": 0.6148523092269897, | |
| "rewards/rejected": -0.896035373210907, | |
| "step": 1540 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 24.433006444286615, | |
| "learning_rate": 1.9825104217894018e-08, | |
| "logits/chosen": -2.5491321086883545, | |
| "logits/rejected": -2.494581460952759, | |
| "logps/chosen": -252.4076690673828, | |
| "logps/rejected": -305.54437255859375, | |
| "loss": 0.5959, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.3050937354564667, | |
| "rewards/margins": 0.45700544118881226, | |
| "rewards/rejected": -0.7620991468429565, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 24.678156956964582, | |
| "learning_rate": 1.7924869220260626e-08, | |
| "logits/chosen": -2.5328431129455566, | |
| "logits/rejected": -2.4817864894866943, | |
| "logps/chosen": -251.4116973876953, | |
| "logps/rejected": -301.2293395996094, | |
| "loss": 0.5487, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.2088792771100998, | |
| "rewards/margins": 0.508228600025177, | |
| "rewards/rejected": -0.7171077728271484, | |
| "step": 1560 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 28.5502202419384, | |
| "learning_rate": 1.6116966348220046e-08, | |
| "logits/chosen": -2.5276474952697754, | |
| "logits/rejected": -2.498145818710327, | |
| "logps/chosen": -244.6429443359375, | |
| "logps/rejected": -293.36865234375, | |
| "loss": 0.56, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.10921525955200195, | |
| "rewards/margins": 0.6631591320037842, | |
| "rewards/rejected": -0.7723743319511414, | |
| "step": 1570 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 30.92180920593632, | |
| "learning_rate": 1.4402114883991318e-08, | |
| "logits/chosen": -2.526020050048828, | |
| "logits/rejected": -2.4712436199188232, | |
| "logps/chosen": -261.9678039550781, | |
| "logps/rejected": -287.0662536621094, | |
| "loss": 0.5647, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.3635459542274475, | |
| "rewards/margins": 0.45373767614364624, | |
| "rewards/rejected": -0.8172836303710938, | |
| "step": 1580 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 29.555421546021286, | |
| "learning_rate": 1.2780997088875866e-08, | |
| "logits/chosen": -2.5242857933044434, | |
| "logits/rejected": -2.475109100341797, | |
| "logps/chosen": -264.02740478515625, | |
| "logps/rejected": -286.8149108886719, | |
| "loss": 0.6074, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.1986238658428192, | |
| "rewards/margins": 0.452684223651886, | |
| "rewards/rejected": -0.6513080596923828, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 26.529175018271285, | |
| "learning_rate": 1.125425793181703e-08, | |
| "logits/chosen": -2.5457723140716553, | |
| "logits/rejected": -2.5002970695495605, | |
| "logps/chosen": -268.09466552734375, | |
| "logps/rejected": -307.01690673828125, | |
| "loss": 0.5854, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.31708940863609314, | |
| "rewards/margins": 0.44073382019996643, | |
| "rewards/rejected": -0.7578231692314148, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "eval_logits/chosen": -2.552327871322632, | |
| "eval_logits/rejected": -2.503713846206665, | |
| "eval_logps/chosen": -257.1977233886719, | |
| "eval_logps/rejected": -298.1694641113281, | |
| "eval_loss": 0.5782989859580994, | |
| "eval_rewards/accuracies": 0.6913461685180664, | |
| "eval_rewards/chosen": -0.22821034491062164, | |
| "eval_rewards/margins": 0.49109748005867004, | |
| "eval_rewards/rejected": -0.7193077802658081, | |
| "eval_runtime": 860.4396, | |
| "eval_samples_per_second": 14.471, | |
| "eval_steps_per_second": 0.453, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 31.28191437907546, | |
| "learning_rate": 9.822504832796036e-09, | |
| "logits/chosen": -2.559476375579834, | |
| "logits/rejected": -2.5392966270446777, | |
| "logps/chosen": -271.2260437011719, | |
| "logps/rejected": -299.50653076171875, | |
| "loss": 0.5778, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.2566366195678711, | |
| "rewards/margins": 0.49396759271621704, | |
| "rewards/rejected": -0.7506042718887329, | |
| "step": 1610 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 30.662407347819215, | |
| "learning_rate": 8.48630742116746e-09, | |
| "logits/chosen": -2.536745071411133, | |
| "logits/rejected": -2.4669063091278076, | |
| "logps/chosen": -273.7347717285156, | |
| "logps/rejected": -307.0451354980469, | |
| "loss": 0.6142, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.28270235657691956, | |
| "rewards/margins": 0.5321863293647766, | |
| "rewards/rejected": -0.8148886561393738, | |
| "step": 1620 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 28.8995046840584, | |
| "learning_rate": 7.246197309029617e-09, | |
| "logits/chosen": -2.5540242195129395, | |
| "logits/rejected": -2.51172137260437, | |
| "logps/chosen": -288.4748229980469, | |
| "logps/rejected": -334.9012451171875, | |
| "loss": 0.5696, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.14674052596092224, | |
| "rewards/margins": 0.5696360468864441, | |
| "rewards/rejected": -0.7163766026496887, | |
| "step": 1630 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 30.322071826617417, | |
| "learning_rate": 6.102667879720164e-09, | |
| "logits/chosen": -2.517651081085205, | |
| "logits/rejected": -2.470822811126709, | |
| "logps/chosen": -270.4826354980469, | |
| "logps/rejected": -301.36944580078125, | |
| "loss": 0.5924, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.3259834349155426, | |
| "rewards/margins": 0.4199187755584717, | |
| "rewards/rejected": -0.7459022402763367, | |
| "step": 1640 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 37.10925817248419, | |
| "learning_rate": 5.056174091521509e-09, | |
| "logits/chosen": -2.585538387298584, | |
| "logits/rejected": -2.5188987255096436, | |
| "logps/chosen": -271.0771179199219, | |
| "logps/rejected": -313.93499755859375, | |
| "loss": 0.5735, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.19604423642158508, | |
| "rewards/margins": 0.5013729333877563, | |
| "rewards/rejected": -0.697417140007019, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 28.879184337464647, | |
| "learning_rate": 4.107132296653548e-09, | |
| "logits/chosen": -2.545414447784424, | |
| "logits/rejected": -2.5129075050354004, | |
| "logps/chosen": -249.78903198242188, | |
| "logps/rejected": -274.78466796875, | |
| "loss": 0.5854, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.20237192511558533, | |
| "rewards/margins": 0.3796917498111725, | |
| "rewards/rejected": -0.5820637345314026, | |
| "step": 1660 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 31.80047095954754, | |
| "learning_rate": 3.255920075626084e-09, | |
| "logits/chosen": -2.501981735229492, | |
| "logits/rejected": -2.5079236030578613, | |
| "logps/chosen": -243.0762176513672, | |
| "logps/rejected": -292.73077392578125, | |
| "loss": 0.5755, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.18036358058452606, | |
| "rewards/margins": 0.4649125933647156, | |
| "rewards/rejected": -0.6452761888504028, | |
| "step": 1670 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 26.04495569695924, | |
| "learning_rate": 2.5028760870168253e-09, | |
| "logits/chosen": -2.5446419715881348, | |
| "logits/rejected": -2.5122885704040527, | |
| "logps/chosen": -276.3479919433594, | |
| "logps/rejected": -304.0161437988281, | |
| "loss": 0.5854, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.3501695692539215, | |
| "rewards/margins": 0.2919423580169678, | |
| "rewards/rejected": -0.6421119570732117, | |
| "step": 1680 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 24.450405436519606, | |
| "learning_rate": 1.8482999327343597e-09, | |
| "logits/chosen": -2.5676443576812744, | |
| "logits/rejected": -2.4829487800598145, | |
| "logps/chosen": -276.8134765625, | |
| "logps/rejected": -306.0779113769531, | |
| "loss": 0.6091, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.22208046913146973, | |
| "rewards/margins": 0.37004339694976807, | |
| "rewards/rejected": -0.5921238660812378, | |
| "step": 1690 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 31.006086476520473, | |
| "learning_rate": 1.2924520388204462e-09, | |
| "logits/chosen": -2.5088953971862793, | |
| "logits/rejected": -2.507538318634033, | |
| "logps/chosen": -248.55526733398438, | |
| "logps/rejected": -294.12591552734375, | |
| "loss": 0.578, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.3303738832473755, | |
| "rewards/margins": 0.3995395600795746, | |
| "rewards/rejected": -0.7299134135246277, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "eval_logits/chosen": -2.5474650859832764, | |
| "eval_logits/rejected": -2.4987149238586426, | |
| "eval_logps/chosen": -255.72315979003906, | |
| "eval_logps/rejected": -296.4235534667969, | |
| "eval_loss": 0.5782065987586975, | |
| "eval_rewards/accuracies": 0.691987156867981, | |
| "eval_rewards/chosen": -0.21346470713615417, | |
| "eval_rewards/margins": 0.4883839786052704, | |
| "eval_rewards/rejected": -0.7018487453460693, | |
| "eval_runtime": 847.8379, | |
| "eval_samples_per_second": 14.686, | |
| "eval_steps_per_second": 0.46, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 27.94865663504945, | |
| "learning_rate": 8.355535518381751e-10, | |
| "logits/chosen": -2.5459036827087402, | |
| "logits/rejected": -2.523585081100464, | |
| "logps/chosen": -267.97119140625, | |
| "logps/rejected": -273.44970703125, | |
| "loss": 0.5877, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.24003572762012482, | |
| "rewards/margins": 0.3651558756828308, | |
| "rewards/rejected": -0.6051916480064392, | |
| "step": 1710 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 26.206122309038193, | |
| "learning_rate": 4.77786250887846e-10, | |
| "logits/chosen": -2.5695338249206543, | |
| "logits/rejected": -2.5357468128204346, | |
| "logps/chosen": -251.841552734375, | |
| "logps/rejected": -304.9547424316406, | |
| "loss": 0.5833, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.12936021387577057, | |
| "rewards/margins": 0.5671738386154175, | |
| "rewards/rejected": -0.6965340971946716, | |
| "step": 1720 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 26.649920950848678, | |
| "learning_rate": 2.1929247528540418e-10, | |
| "logits/chosen": -2.554141044616699, | |
| "logits/rejected": -2.493603467941284, | |
| "logps/chosen": -265.3050231933594, | |
| "logps/rejected": -310.0926208496094, | |
| "loss": 0.5706, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.20094648003578186, | |
| "rewards/margins": 0.47598400712013245, | |
| "rewards/rejected": -0.6769304871559143, | |
| "step": 1730 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 27.75248252937658, | |
| "learning_rate": 6.017506793212779e-11, | |
| "logits/chosen": -2.5380330085754395, | |
| "logits/rejected": -2.4906625747680664, | |
| "logps/chosen": -251.3086700439453, | |
| "logps/rejected": -314.17645263671875, | |
| "loss": 0.5705, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.11332281678915024, | |
| "rewards/margins": 0.5656474828720093, | |
| "rewards/rejected": -0.6789703369140625, | |
| "step": 1740 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 31.342302748055804, | |
| "learning_rate": 4.973343980252398e-13, | |
| "logits/chosen": -2.543778896331787, | |
| "logits/rejected": -2.499126434326172, | |
| "logps/chosen": -258.0113830566406, | |
| "logps/rejected": -294.2593994140625, | |
| "loss": 0.5742, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.23090767860412598, | |
| "rewards/margins": 0.5441216230392456, | |
| "rewards/rejected": -0.7750292420387268, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 1751, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5964619764459262, | |
| "train_runtime": 39064.845, | |
| "train_samples_per_second": 2.868, | |
| "train_steps_per_second": 0.045 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1751, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |