{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.981366459627329,
  "eval_steps": 50,
  "global_step": 120,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12422360248447205,
      "grad_norm": 54.367663803058946,
      "learning_rate": 5e-07,
      "logits/chosen": -2.7148144245147705,
      "logits/rejected": -2.7243547439575195,
      "logps/chosen": -242.867431640625,
      "logps/rejected": -227.12136840820312,
      "loss": 0.691,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": 0.009815122000873089,
      "rewards/margins": 0.005822173319756985,
      "rewards/rejected": 0.003992948215454817,
      "step": 5
    },
    {
      "epoch": 0.2484472049689441,
      "grad_norm": 51.804115964444165,
      "learning_rate": 1e-06,
      "logits/chosen": -2.6798529624938965,
      "logits/rejected": -2.703315258026123,
      "logps/chosen": -256.2458190917969,
      "logps/rejected": -217.85592651367188,
      "loss": 0.6456,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.43589210510253906,
      "rewards/margins": 0.08772359788417816,
      "rewards/rejected": 0.3481685519218445,
      "step": 10
    },
    {
      "epoch": 0.37267080745341613,
      "grad_norm": 50.53080123806113,
      "learning_rate": 9.949107209404663e-07,
      "logits/chosen": -2.5799756050109863,
      "logits/rejected": -2.565157651901245,
      "logps/chosen": -237.31692504882812,
      "logps/rejected": -208.06655883789062,
      "loss": 0.6378,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 1.4567431211471558,
      "rewards/margins": 0.7478972673416138,
      "rewards/rejected": 0.7088459730148315,
      "step": 15
    },
    {
      "epoch": 0.4968944099378882,
      "grad_norm": 54.57054056014394,
      "learning_rate": 9.797464868072486e-07,
      "logits/chosen": -2.4506874084472656,
      "logits/rejected": -2.433974027633667,
      "logps/chosen": -247.51657104492188,
      "logps/rejected": -216.2230987548828,
      "loss": 0.6587,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 1.5972602367401123,
      "rewards/margins": 1.0024363994598389,
      "rewards/rejected": 0.594823956489563,
      "step": 20
    },
    {
      "epoch": 0.6211180124223602,
      "grad_norm": 44.117531702150536,
      "learning_rate": 9.548159976772592e-07,
      "logits/chosen": -2.4157333374023438,
      "logits/rejected": -2.3935298919677734,
      "logps/chosen": -231.5720672607422,
      "logps/rejected": -216.5135498046875,
      "loss": 0.6622,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.7699011564254761,
      "rewards/margins": 0.8483353853225708,
      "rewards/rejected": -0.07843427360057831,
      "step": 25
    },
    {
      "epoch": 0.7453416149068323,
      "grad_norm": 46.22604593677178,
      "learning_rate": 9.206267664155906e-07,
      "logits/chosen": -2.4077823162078857,
      "logits/rejected": -2.4088187217712402,
      "logps/chosen": -260.6187744140625,
      "logps/rejected": -240.7838897705078,
      "loss": 0.6343,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.9798136949539185,
      "rewards/margins": 1.1557605266571045,
      "rewards/rejected": -0.17594675719738007,
      "step": 30
    },
    {
      "epoch": 0.8695652173913043,
      "grad_norm": 45.738597782002316,
      "learning_rate": 8.778747871771291e-07,
      "logits/chosen": -2.4890403747558594,
      "logits/rejected": -2.4890661239624023,
      "logps/chosen": -267.4264831542969,
      "logps/rejected": -236.0730743408203,
      "loss": 0.6245,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 0.5028200745582581,
      "rewards/margins": 1.037217617034912,
      "rewards/rejected": -0.5343974232673645,
      "step": 35
    },
    {
      "epoch": 0.9937888198757764,
      "grad_norm": 40.58782325478915,
      "learning_rate": 8.274303669726426e-07,
      "logits/chosen": -2.464543104171753,
      "logits/rejected": -2.45270037651062,
      "logps/chosen": -246.43997192382812,
      "logps/rejected": -244.5944366455078,
      "loss": 0.6076,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": 0.423252671957016,
      "rewards/margins": 0.9229713678359985,
      "rewards/rejected": -0.49971866607666016,
      "step": 40
    },
    {
      "epoch": 1.1180124223602483,
      "grad_norm": 21.75862492001889,
      "learning_rate": 7.703204087277988e-07,
      "logits/chosen": -2.437509059906006,
      "logits/rejected": -2.4511005878448486,
      "logps/chosen": -238.031005859375,
      "logps/rejected": -239.1492462158203,
      "loss": 0.2467,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 1.3256285190582275,
      "rewards/margins": 2.4453110694885254,
      "rewards/rejected": -1.1196826696395874,
      "step": 45
    },
    {
      "epoch": 1.2422360248447206,
      "grad_norm": 22.132976615768026,
      "learning_rate": 7.077075065009433e-07,
      "logits/chosen": -2.484419822692871,
      "logits/rejected": -2.485710859298706,
      "logps/chosen": -241.15200805664062,
      "logps/rejected": -220.26907348632812,
      "loss": 0.2364,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.7490062713623047,
      "rewards/margins": 2.5016419887542725,
      "rewards/rejected": -0.7526359558105469,
      "step": 50
    },
    {
      "epoch": 1.2422360248447206,
      "eval_logits/chosen": -2.5087192058563232,
      "eval_logits/rejected": -2.515753746032715,
      "eval_logps/chosen": -249.35264587402344,
      "eval_logps/rejected": -202.0917205810547,
      "eval_loss": 0.5729268789291382,
      "eval_rewards/accuracies": 0.7291666865348816,
      "eval_rewards/chosen": 1.2346218824386597,
      "eval_rewards/margins": 1.0999401807785034,
      "eval_rewards/rejected": 0.13468176126480103,
      "eval_runtime": 75.094,
      "eval_samples_per_second": 15.181,
      "eval_steps_per_second": 0.24,
      "step": 50
    },
    {
      "epoch": 1.3664596273291925,
      "grad_norm": 17.330600410265617,
      "learning_rate": 6.408662784207149e-07,
      "logits/chosen": -2.4883952140808105,
      "logits/rejected": -2.482597827911377,
      "logps/chosen": -236.4322052001953,
      "logps/rejected": -229.5203094482422,
      "loss": 0.2196,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 2.2716193199157715,
      "rewards/margins": 3.196570634841919,
      "rewards/rejected": -0.9249511957168579,
      "step": 55
    },
    {
      "epoch": 1.4906832298136645,
      "grad_norm": 17.716660362051734,
      "learning_rate": 5.711574191366427e-07,
      "logits/chosen": -2.468207836151123,
      "logits/rejected": -2.4725213050842285,
      "logps/chosen": -222.43896484375,
      "logps/rejected": -200.0598602294922,
      "loss": 0.2119,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 2.1822891235351562,
      "rewards/margins": 2.8318796157836914,
      "rewards/rejected": -0.6495904922485352,
      "step": 60
    },
    {
      "epoch": 1.6149068322981366,
      "grad_norm": 23.32765774015972,
      "learning_rate": 5e-07,
      "logits/chosen": -2.5094847679138184,
      "logits/rejected": -2.5222580432891846,
      "logps/chosen": -236.4397430419922,
      "logps/rejected": -216.79052734375,
      "loss": 0.2118,
      "rewards/accuracies": 0.981249988079071,
      "rewards/chosen": 2.586951494216919,
      "rewards/margins": 3.453221559524536,
      "rewards/rejected": -0.8662700653076172,
      "step": 65
    },
    {
      "epoch": 1.7391304347826086,
      "grad_norm": 25.27024243839641,
      "learning_rate": 4.2884258086335745e-07,
      "logits/chosen": -2.506361484527588,
      "logits/rejected": -2.4916276931762695,
      "logps/chosen": -228.9758758544922,
      "logps/rejected": -225.41006469726562,
      "loss": 0.2294,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 2.325880527496338,
      "rewards/margins": 3.3348236083984375,
      "rewards/rejected": -1.0089433193206787,
      "step": 70
    },
    {
      "epoch": 1.8633540372670807,
      "grad_norm": 27.860922972380834,
      "learning_rate": 3.591337215792851e-07,
      "logits/chosen": -2.5197181701660156,
      "logits/rejected": -2.5090882778167725,
      "logps/chosen": -239.81277465820312,
      "logps/rejected": -230.70059204101562,
      "loss": 0.265,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 2.431699275970459,
      "rewards/margins": 3.363804340362549,
      "rewards/rejected": -0.9321050643920898,
      "step": 75
    },
    {
      "epoch": 1.9875776397515528,
      "grad_norm": 17.61138833178944,
      "learning_rate": 2.922924934990568e-07,
      "logits/chosen": -2.543259382247925,
      "logits/rejected": -2.493022918701172,
      "logps/chosen": -237.87887573242188,
      "logps/rejected": -279.49261474609375,
      "loss": 0.2128,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 2.389310836791992,
      "rewards/margins": 4.381407260894775,
      "rewards/rejected": -1.9920963048934937,
      "step": 80
    },
    {
      "epoch": 2.111801242236025,
      "grad_norm": 13.031340899683215,
      "learning_rate": 2.2967959127220137e-07,
      "logits/chosen": -2.5387518405914307,
      "logits/rejected": -2.5558295249938965,
      "logps/chosen": -225.1177520751953,
      "logps/rejected": -233.97073364257812,
      "loss": 0.1297,
      "rewards/accuracies": 0.981249988079071,
      "rewards/chosen": 2.17976713180542,
      "rewards/margins": 3.992032527923584,
      "rewards/rejected": -1.812265396118164,
      "step": 85
    },
    {
      "epoch": 2.2360248447204967,
      "grad_norm": 16.033450629688048,
      "learning_rate": 1.725696330273575e-07,
      "logits/chosen": -2.5489468574523926,
      "logits/rejected": -2.5377697944641113,
      "logps/chosen": -255.279296875,
      "logps/rejected": -245.3787078857422,
      "loss": 0.123,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 2.2218708992004395,
      "rewards/margins": 4.716561794281006,
      "rewards/rejected": -2.4946906566619873,
      "step": 90
    },
    {
      "epoch": 2.360248447204969,
      "grad_norm": 12.307639352242482,
      "learning_rate": 1.2212521282287093e-07,
      "logits/chosen": -2.5559678077697754,
      "logits/rejected": -2.5754735469818115,
      "logps/chosen": -240.5291748046875,
      "logps/rejected": -270.05230712890625,
      "loss": 0.1073,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 1.9212055206298828,
      "rewards/margins": 4.681941032409668,
      "rewards/rejected": -2.760735511779785,
      "step": 95
    },
    {
      "epoch": 2.4844720496894412,
      "grad_norm": 12.425204577942079,
      "learning_rate": 7.937323358440934e-08,
      "logits/chosen": -2.549752950668335,
      "logits/rejected": -2.5533714294433594,
      "logps/chosen": -235.24368286132812,
      "logps/rejected": -259.5509033203125,
      "loss": 0.1061,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 1.9440858364105225,
      "rewards/margins": 4.815189361572266,
      "rewards/rejected": -2.8711037635803223,
      "step": 100
    },
    {
      "epoch": 2.4844720496894412,
      "eval_logits/chosen": -2.55989933013916,
      "eval_logits/rejected": -2.5775303840637207,
      "eval_logps/chosen": -254.12814331054688,
      "eval_logps/rejected": -212.31497192382812,
      "eval_loss": 0.6158778071403503,
      "eval_rewards/accuracies": 0.7569444179534912,
      "eval_rewards/chosen": 0.7570738792419434,
      "eval_rewards/margins": 1.6447181701660156,
      "eval_rewards/rejected": -0.8876442313194275,
      "eval_runtime": 74.8651,
      "eval_samples_per_second": 15.227,
      "eval_steps_per_second": 0.24,
      "step": 100
    },
    {
      "epoch": 2.608695652173913,
      "grad_norm": 18.01907950225221,
      "learning_rate": 4.518400232274078e-08,
      "logits/chosen": -2.546436309814453,
      "logits/rejected": -2.5362162590026855,
      "logps/chosen": -227.8841552734375,
      "logps/rejected": -236.8609619140625,
      "loss": 0.1288,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 1.770033836364746,
      "rewards/margins": 4.3705244064331055,
      "rewards/rejected": -2.6004908084869385,
      "step": 105
    },
    {
      "epoch": 2.732919254658385,
      "grad_norm": 17.707001673219747,
      "learning_rate": 2.025351319275137e-08,
      "logits/chosen": -2.5263776779174805,
      "logits/rejected": -2.5271897315979004,
      "logps/chosen": -243.36788940429688,
      "logps/rejected": -254.2205352783203,
      "loss": 0.1263,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.7786777019500732,
      "rewards/margins": 4.2846503257751465,
      "rewards/rejected": -2.505972385406494,
      "step": 110
    },
    {
      "epoch": 2.857142857142857,
      "grad_norm": 16.228515302877508,
      "learning_rate": 5.0892790595336575e-09,
      "logits/chosen": -2.5431525707244873,
      "logits/rejected": -2.5375916957855225,
      "logps/chosen": -234.5476531982422,
      "logps/rejected": -246.033447265625,
      "loss": 0.113,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 1.938812494277954,
      "rewards/margins": 4.491750240325928,
      "rewards/rejected": -2.552938461303711,
      "step": 115
    },
    {
      "epoch": 2.981366459627329,
      "grad_norm": 15.40676083530787,
      "learning_rate": 0.0,
      "logits/chosen": -2.529771089553833,
      "logits/rejected": -2.5455727577209473,
      "logps/chosen": -242.18539428710938,
      "logps/rejected": -240.2424774169922,
      "loss": 0.1225,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 2.026702642440796,
      "rewards/margins": 4.205595970153809,
      "rewards/rejected": -2.178893566131592,
      "step": 120
    },
    {
      "epoch": 2.981366459627329,
      "step": 120,
      "total_flos": 1414680891359232.0,
      "train_loss": 0.3313312023878098,
      "train_runtime": 4261.9082,
      "train_samples_per_second": 7.22,
      "train_steps_per_second": 0.028
    }
  ],
  "logging_steps": 5,
  "max_steps": 120,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1414680891359232.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}