{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.7476635514018691,
  "eval_steps": 50,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.037383177570093455,
      "grad_norm": 53.10315006693553,
      "learning_rate": 5e-07,
      "logits/chosen": -2.7264351844787598,
      "logits/rejected": -2.7314915657043457,
      "logps/chosen": -233.46450805664062,
      "logps/rejected": -215.2651824951172,
      "loss": 0.6911,
      "rewards/accuracies": 0.29374998807907104,
      "rewards/chosen": 0.011523213237524033,
      "rewards/margins": 0.00106804131064564,
      "rewards/rejected": 0.010455173440277576,
      "step": 5
    },
    {
      "epoch": 0.07476635514018691,
      "grad_norm": 47.36575434115236,
      "learning_rate": 1e-06,
      "logits/chosen": -2.7007861137390137,
      "logits/rejected": -2.6771092414855957,
      "logps/chosen": -243.54736328125,
      "logps/rejected": -216.7264404296875,
      "loss": 0.6571,
      "rewards/accuracies": 0.606249988079071,
      "rewards/chosen": 0.35093382000923157,
      "rewards/margins": 0.08091190457344055,
      "rewards/rejected": 0.270021915435791,
      "step": 10
    },
    {
      "epoch": 0.11214953271028037,
      "grad_norm": 47.16978215311432,
      "learning_rate": 9.995924118521016e-07,
      "logits/chosen": -2.430677652359009,
      "logits/rejected": -2.3965601921081543,
      "logps/chosen": -245.1031951904297,
      "logps/rejected": -206.2293701171875,
      "loss": 0.6246,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 1.259174108505249,
      "rewards/margins": 0.5380627512931824,
      "rewards/rejected": 0.7211112380027771,
      "step": 15
    },
    {
      "epoch": 0.14953271028037382,
      "grad_norm": 43.635723418114395,
      "learning_rate": 9.983703119207998e-07,
      "logits/chosen": -2.1696972846984863,
      "logits/rejected": -2.1355605125427246,
      "logps/chosen": -241.0774383544922,
      "logps/rejected": -203.36190795898438,
      "loss": 0.6793,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": 1.169782280921936,
      "rewards/margins": 0.6064848899841309,
      "rewards/rejected": 0.5632972717285156,
      "step": 20
    },
    {
      "epoch": 0.18691588785046728,
      "grad_norm": 48.55163726023584,
      "learning_rate": 9.963356926598848e-07,
      "logits/chosen": -2.0636093616485596,
      "logits/rejected": -2.068882942199707,
      "logps/chosen": -245.859130859375,
      "logps/rejected": -225.98214721679688,
      "loss": 0.6303,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 1.4211041927337646,
      "rewards/margins": 0.8779617547988892,
      "rewards/rejected": 0.5431426167488098,
      "step": 25
    },
    {
      "epoch": 0.22429906542056074,
      "grad_norm": 42.28573315120618,
      "learning_rate": 9.934918712161414e-07,
      "logits/chosen": -2.1142563819885254,
      "logits/rejected": -2.0855462551116943,
      "logps/chosen": -239.36471557617188,
      "logps/rejected": -208.6727294921875,
      "loss": 0.5921,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 0.9205316305160522,
      "rewards/margins": 0.793667733669281,
      "rewards/rejected": 0.12686386704444885,
      "step": 30
    },
    {
      "epoch": 0.2616822429906542,
      "grad_norm": 43.204924277825036,
      "learning_rate": 9.898434840212305e-07,
      "logits/chosen": -2.1356260776519775,
      "logits/rejected": -2.0982658863067627,
      "logps/chosen": -249.84848022460938,
      "logps/rejected": -232.4295196533203,
      "loss": 0.5949,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 0.6800674200057983,
      "rewards/margins": 0.8049441576004028,
      "rewards/rejected": -0.12487666308879852,
      "step": 35
    },
    {
      "epoch": 0.29906542056074764,
      "grad_norm": 38.34647398013497,
      "learning_rate": 9.853964792326704e-07,
      "logits/chosen": -2.1388490200042725,
      "logits/rejected": -2.106875419616699,
      "logps/chosen": -231.1354217529297,
      "logps/rejected": -210.77197265625,
      "loss": 0.6446,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.6789754033088684,
      "rewards/margins": 1.0355161428451538,
      "rewards/rejected": -0.35654082894325256,
      "step": 40
    },
    {
      "epoch": 0.3364485981308411,
      "grad_norm": 44.46300001215897,
      "learning_rate": 9.80158107036243e-07,
      "logits/chosen": -2.182988405227661,
      "logits/rejected": -2.139224052429199,
      "logps/chosen": -253.672119140625,
      "logps/rejected": -198.7143096923828,
      "loss": 0.5967,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 0.711867094039917,
      "rewards/margins": 0.8050382733345032,
      "rewards/rejected": -0.093171127140522,
      "step": 45
    },
    {
      "epoch": 0.37383177570093457,
      "grad_norm": 36.368588934194,
      "learning_rate": 9.741369078256344e-07,
      "logits/chosen": -2.1803622245788574,
      "logits/rejected": -2.1714465618133545,
      "logps/chosen": -229.3830108642578,
      "logps/rejected": -214.208251953125,
      "loss": 0.577,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.7632301449775696,
      "rewards/margins": 0.9728155136108398,
      "rewards/rejected": -0.20958539843559265,
      "step": 50
    },
    {
      "epoch": 0.37383177570093457,
      "eval_logits/chosen": -2.168900966644287,
      "eval_logits/rejected": -2.154597759246826,
      "eval_logps/chosen": -240.57708740234375,
      "eval_logps/rejected": -220.16871643066406,
      "eval_loss": 0.578223705291748,
      "eval_rewards/accuracies": 0.7250000238418579,
      "eval_rewards/chosen": 0.7959616780281067,
      "eval_rewards/margins": 0.9759488701820374,
      "eval_rewards/rejected": -0.17998719215393066,
      "eval_runtime": 252.8699,
      "eval_samples_per_second": 15.024,
      "eval_steps_per_second": 0.237,
      "step": 50
    },
    {
      "epoch": 0.411214953271028,
      "grad_norm": 36.85921784944549,
      "learning_rate": 9.673426982785825e-07,
      "logits/chosen": -2.1165783405303955,
      "logits/rejected": -2.133802890777588,
      "logps/chosen": -227.85147094726562,
      "logps/rejected": -229.4573516845703,
      "loss": 0.5854,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 0.668209969997406,
      "rewards/margins": 0.9655311703681946,
      "rewards/rejected": -0.29732123017311096,
      "step": 55
    },
    {
      "epoch": 0.4485981308411215,
      "grad_norm": 37.099348027626476,
      "learning_rate": 9.597865553522297e-07,
      "logits/chosen": -2.1299071311950684,
      "logits/rejected": -2.1265180110931396,
      "logps/chosen": -246.6769256591797,
      "logps/rejected": -218.6841583251953,
      "loss": 0.5468,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.3689742088317871,
      "rewards/margins": 1.112032175064087,
      "rewards/rejected": -0.7430580258369446,
      "step": 60
    },
    {
      "epoch": 0.48598130841121495,
      "grad_norm": 35.927665101781095,
      "learning_rate": 9.514807982237785e-07,
      "logits/chosen": -2.298119068145752,
      "logits/rejected": -2.2940573692321777,
      "logps/chosen": -265.78155517578125,
      "logps/rejected": -213.521240234375,
      "loss": 0.5918,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.9371916055679321,
      "rewards/margins": 1.3884761333465576,
      "rewards/rejected": -0.45128464698791504,
      "step": 65
    },
    {
      "epoch": 0.5233644859813084,
      "grad_norm": 39.11873955408514,
      "learning_rate": 9.424389682058886e-07,
      "logits/chosen": -2.3393406867980957,
      "logits/rejected": -2.309872627258301,
      "logps/chosen": -218.0382080078125,
      "logps/rejected": -194.01332092285156,
      "loss": 0.5295,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.9512729644775391,
      "rewards/margins": 1.144668698310852,
      "rewards/rejected": -0.19339559972286224,
      "step": 70
    },
    {
      "epoch": 0.5607476635514018,
      "grad_norm": 37.634486315960864,
      "learning_rate": 9.326758066695624e-07,
      "logits/chosen": -2.34443736076355,
      "logits/rejected": -2.325118064880371,
      "logps/chosen": -259.53143310546875,
      "logps/rejected": -198.3692626953125,
      "loss": 0.5487,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.9949586987495422,
      "rewards/margins": 1.7472797632217407,
      "rewards/rejected": -0.7523208856582642,
      "step": 75
    },
    {
      "epoch": 0.5981308411214953,
      "grad_norm": 35.50580413675745,
      "learning_rate": 9.222072310105126e-07,
      "logits/chosen": -2.3364174365997314,
      "logits/rejected": -2.312894105911255,
      "logps/chosen": -236.32666015625,
      "logps/rejected": -261.32843017578125,
      "loss": 0.6152,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.37070125341415405,
      "rewards/margins": 1.0360088348388672,
      "rewards/rejected": -0.6653076410293579,
      "step": 80
    },
    {
      "epoch": 0.6355140186915887,
      "grad_norm": 33.7374435363093,
      "learning_rate": 9.110503086981955e-07,
      "logits/chosen": -2.2803781032562256,
      "logits/rejected": -2.267977237701416,
      "logps/chosen": -255.5703887939453,
      "logps/rejected": -206.46066284179688,
      "loss": 0.553,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": 0.18914642930030823,
      "rewards/margins": 1.2585302591323853,
      "rewards/rejected": -1.0693838596343994,
      "step": 85
    },
    {
      "epoch": 0.6728971962616822,
      "grad_norm": 39.897529811340966,
      "learning_rate": 8.992232294498169e-07,
      "logits/chosen": -2.1736109256744385,
      "logits/rejected": -2.1623623371124268,
      "logps/chosen": -255.2686309814453,
      "logps/rejected": -225.6684112548828,
      "loss": 0.5359,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": 0.11371274292469025,
      "rewards/margins": 1.4066712856292725,
      "rewards/rejected": -1.2929584980010986,
      "step": 90
    },
    {
      "epoch": 0.7102803738317757,
      "grad_norm": 36.88770484599986,
      "learning_rate": 8.867452755746805e-07,
      "logits/chosen": -2.1795907020568848,
      "logits/rejected": -2.1622931957244873,
      "logps/chosen": -267.36358642578125,
      "logps/rejected": -237.27359008789062,
      "loss": 0.4889,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.24791303277015686,
      "rewards/margins": 1.365724802017212,
      "rewards/rejected": -1.6136376857757568,
      "step": 95
    },
    {
      "epoch": 0.7476635514018691,
      "grad_norm": 29.575103655541206,
      "learning_rate": 8.736367905372246e-07,
      "logits/chosen": -2.1824848651885986,
      "logits/rejected": -2.164578914642334,
      "logps/chosen": -262.33575439453125,
      "logps/rejected": -242.2086639404297,
      "loss": 0.5388,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.11618832498788834,
      "rewards/margins": 1.8072330951690674,
      "rewards/rejected": -1.9234212636947632,
      "step": 100
    },
    {
      "epoch": 0.7476635514018691,
      "eval_logits/chosen": -2.199050188064575,
      "eval_logits/rejected": -2.174029588699341,
      "eval_logps/chosen": -252.93431091308594,
      "eval_logps/rejected": -238.5013885498047,
      "eval_loss": 0.5390673875808716,
      "eval_rewards/accuracies": 0.7479166388511658,
      "eval_rewards/chosen": -0.4397614300251007,
      "eval_rewards/margins": 1.5734889507293701,
      "eval_rewards/rejected": -2.0132501125335693,
      "eval_runtime": 252.4297,
      "eval_samples_per_second": 15.05,
      "eval_steps_per_second": 0.238,
      "step": 100
    }
  ],
  "logging_steps": 5,
  "max_steps": 399,
  "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": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1178822762299392.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}