{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9345794392523364,
  "eval_steps": 50,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04672897196261682,
      "grad_norm": 55.02521133123827,
      "learning_rate": 5e-07,
      "logits/chosen": -2.7216238975524902,
      "logits/rejected": -2.7209055423736572,
      "logps/chosen": -268.4510192871094,
      "logps/rejected": -203.9590606689453,
      "loss": 0.6914,
      "rewards/accuracies": 0.33125001192092896,
      "rewards/chosen": 0.017813727259635925,
      "rewards/margins": 0.009322145953774452,
      "rewards/rejected": 0.008491581305861473,
      "step": 5
    },
    {
      "epoch": 0.09345794392523364,
      "grad_norm": 50.12554517439661,
      "learning_rate": 1e-06,
      "logits/chosen": -2.660832166671753,
      "logits/rejected": -2.6669700145721436,
      "logps/chosen": -256.93609619140625,
      "logps/rejected": -211.64938354492188,
      "loss": 0.6467,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": 0.4671781659126282,
      "rewards/margins": 0.18480566143989563,
      "rewards/rejected": 0.28237253427505493,
      "step": 10
    },
    {
      "epoch": 0.14018691588785046,
      "grad_norm": 48.457281147169425,
      "learning_rate": 9.993623730611148e-07,
      "logits/chosen": -2.494657278060913,
      "logits/rejected": -2.5133018493652344,
      "logps/chosen": -232.6891632080078,
      "logps/rejected": -212.8677215576172,
      "loss": 0.6204,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 1.2435152530670166,
      "rewards/margins": 0.6126660108566284,
      "rewards/rejected": 0.6308490037918091,
      "step": 15
    },
    {
      "epoch": 0.18691588785046728,
      "grad_norm": 42.23668651632898,
      "learning_rate": 9.97451118516912e-07,
      "logits/chosen": -2.3121209144592285,
      "logits/rejected": -2.302377462387085,
      "logps/chosen": -234.3399658203125,
      "logps/rejected": -191.3181610107422,
      "loss": 0.6223,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 1.3283790349960327,
      "rewards/margins": 0.9071598052978516,
      "rewards/rejected": 0.4212193489074707,
      "step": 20
    },
    {
      "epoch": 0.2336448598130841,
      "grad_norm": 49.68505743626908,
      "learning_rate": 9.94271111036929e-07,
      "logits/chosen": -2.2619333267211914,
      "logits/rejected": -2.2323482036590576,
      "logps/chosen": -230.17385864257812,
      "logps/rejected": -205.64108276367188,
      "loss": 0.616,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 1.3353136777877808,
      "rewards/margins": 1.208017110824585,
      "rewards/rejected": 0.127296581864357,
      "step": 25
    },
    {
      "epoch": 0.2803738317757009,
      "grad_norm": 39.50573463077347,
      "learning_rate": 9.898304612549066e-07,
      "logits/chosen": -2.308243989944458,
      "logits/rejected": -2.2968266010284424,
      "logps/chosen": -241.26632690429688,
      "logps/rejected": -207.3319854736328,
      "loss": 0.5998,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 0.9092999696731567,
      "rewards/margins": 0.9556086659431458,
      "rewards/rejected": -0.046308644115924835,
      "step": 30
    },
    {
      "epoch": 0.32710280373831774,
      "grad_norm": 42.78736900246308,
      "learning_rate": 9.841404950825536e-07,
      "logits/chosen": -2.3728604316711426,
      "logits/rejected": -2.3580102920532227,
      "logps/chosen": -256.8548583984375,
      "logps/rejected": -205.154052734375,
      "loss": 0.5996,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": 1.0152027606964111,
      "rewards/margins": 1.2137099504470825,
      "rewards/rejected": -0.1985071450471878,
      "step": 35
    },
    {
      "epoch": 0.37383177570093457,
      "grad_norm": 41.73908597429494,
      "learning_rate": 9.77215724822721e-07,
      "logits/chosen": -2.4492850303649902,
      "logits/rejected": -2.4539356231689453,
      "logps/chosen": -243.1707763671875,
      "logps/rejected": -213.95166015625,
      "loss": 0.6098,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 1.0353302955627441,
      "rewards/margins": 1.2659428119659424,
      "rewards/rejected": -0.2306123673915863,
      "step": 40
    },
    {
      "epoch": 0.4205607476635514,
      "grad_norm": 41.530064757148224,
      "learning_rate": 9.69073812155662e-07,
      "logits/chosen": -2.5637125968933105,
      "logits/rejected": -2.5535428524017334,
      "logps/chosen": -244.7168731689453,
      "logps/rejected": -205.80923461914062,
      "loss": 0.5974,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.8133939504623413,
      "rewards/margins": 0.9837163686752319,
      "rewards/rejected": -0.17032238841056824,
      "step": 45
    },
    {
      "epoch": 0.4672897196261682,
      "grad_norm": 38.26706141308248,
      "learning_rate": 9.597355230927788e-07,
      "logits/chosen": -2.5823917388916016,
      "logits/rejected": -2.562842607498169,
      "logps/chosen": -240.04067993164062,
      "logps/rejected": -209.23428344726562,
      "loss": 0.572,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": 0.9298027753829956,
      "rewards/margins": 1.2456680536270142,
      "rewards/rejected": -0.3158652186393738,
      "step": 50
    },
    {
      "epoch": 0.4672897196261682,
      "eval_logits/chosen": -2.530949115753174,
      "eval_logits/rejected": -2.529101610183716,
      "eval_logps/chosen": -245.5291748046875,
      "eval_logps/rejected": -217.46429443359375,
      "eval_loss": 0.5720326900482178,
      "eval_rewards/accuracies": 0.7578125,
      "eval_rewards/chosen": 1.0708366632461548,
      "eval_rewards/margins": 1.28933846950531,
      "eval_rewards/rejected": -0.2185017466545105,
      "eval_runtime": 202.2601,
      "eval_samples_per_second": 15.03,
      "eval_steps_per_second": 0.237,
      "step": 50
    },
    {
      "epoch": 0.514018691588785,
      "grad_norm": 40.54073508413725,
      "learning_rate": 9.4922467501275e-07,
      "logits/chosen": -2.495945930480957,
      "logits/rejected": -2.487422466278076,
      "logps/chosen": -250.51620483398438,
      "logps/rejected": -228.90200805664062,
      "loss": 0.5176,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": 1.0155770778656006,
      "rewards/margins": 1.9236654043197632,
      "rewards/rejected": -0.9080885648727417,
      "step": 55
    },
    {
      "epoch": 0.5607476635514018,
      "grad_norm": 38.23797310786567,
      "learning_rate": 9.375680759151206e-07,
      "logits/chosen": -2.474236249923706,
      "logits/rejected": -2.4737977981567383,
      "logps/chosen": -255.09298706054688,
      "logps/rejected": -200.73593139648438,
      "loss": 0.5654,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": 1.0740002393722534,
      "rewards/margins": 1.5434155464172363,
      "rewards/rejected": -0.4694152772426605,
      "step": 60
    },
    {
      "epoch": 0.6074766355140186,
      "grad_norm": 42.648181943788025,
      "learning_rate": 9.247954560462927e-07,
      "logits/chosen": -2.505916118621826,
      "logits/rejected": -2.506608724594116,
      "logps/chosen": -255.432861328125,
      "logps/rejected": -205.4224090576172,
      "loss": 0.5628,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": 1.1411590576171875,
      "rewards/margins": 1.7762504816055298,
      "rewards/rejected": -0.6350914239883423,
      "step": 65
    },
    {
      "epoch": 0.6542056074766355,
      "grad_norm": 38.81572593341751,
      "learning_rate": 9.109393920723001e-07,
      "logits/chosen": -2.4328043460845947,
      "logits/rejected": -2.4342734813690186,
      "logps/chosen": -233.8389129638672,
      "logps/rejected": -212.91085815429688,
      "loss": 0.5378,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.6836588978767395,
      "rewards/margins": 1.282029390335083,
      "rewards/rejected": -0.5983705520629883,
      "step": 70
    },
    {
      "epoch": 0.7009345794392523,
      "grad_norm": 32.317612654080975,
      "learning_rate": 8.960352239917699e-07,
      "logits/chosen": -2.450084924697876,
      "logits/rejected": -2.401425361633301,
      "logps/chosen": -240.6315460205078,
      "logps/rejected": -227.21084594726562,
      "loss": 0.5154,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": 0.6731350421905518,
      "rewards/margins": 1.5562646389007568,
      "rewards/rejected": -0.8831297755241394,
      "step": 75
    },
    {
      "epoch": 0.7476635514018691,
      "grad_norm": 30.17721204804764,
      "learning_rate": 8.801209650009814e-07,
      "logits/chosen": -2.4172046184539795,
      "logits/rejected": -2.400567054748535,
      "logps/chosen": -245.4665985107422,
      "logps/rejected": -214.18515014648438,
      "loss": 0.5206,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": 0.9583255052566528,
      "rewards/margins": 1.7562158107757568,
      "rewards/rejected": -0.7978904843330383,
      "step": 80
    },
    {
      "epoch": 0.794392523364486,
      "grad_norm": 32.24535114623233,
      "learning_rate": 8.632372045409141e-07,
      "logits/chosen": -2.320589065551758,
      "logits/rejected": -2.3311946392059326,
      "logps/chosen": -245.598388671875,
      "logps/rejected": -234.7646026611328,
      "loss": 0.5472,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": 1.1685658693313599,
      "rewards/margins": 1.754003882408142,
      "rewards/rejected": -0.5854381322860718,
      "step": 85
    },
    {
      "epoch": 0.8411214953271028,
      "grad_norm": 34.061291659967246,
      "learning_rate": 8.454270047735642e-07,
      "logits/chosen": -2.329784870147705,
      "logits/rejected": -2.304997682571411,
      "logps/chosen": -238.0483856201172,
      "logps/rejected": -195.24313354492188,
      "loss": 0.5291,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": 0.6335947513580322,
      "rewards/margins": 1.5654070377349854,
      "rewards/rejected": -0.9318124055862427,
      "step": 90
    },
    {
      "epoch": 0.8878504672897196,
      "grad_norm": 31.735542564732725,
      "learning_rate": 8.267357907515661e-07,
      "logits/chosen": -2.298316478729248,
      "logits/rejected": -2.2975010871887207,
      "logps/chosen": -246.3526153564453,
      "logps/rejected": -206.03524780273438,
      "loss": 0.5418,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": 0.5418449640274048,
      "rewards/margins": 1.539952039718628,
      "rewards/rejected": -0.9981070756912231,
      "step": 95
    },
    {
      "epoch": 0.9345794392523364,
      "grad_norm": 32.65915062987667,
      "learning_rate": 8.072112345612433e-07,
      "logits/chosen": -2.2663910388946533,
      "logits/rejected": -2.218681812286377,
      "logps/chosen": -246.5704803466797,
      "logps/rejected": -218.6560821533203,
      "loss": 0.4997,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": 0.8927062153816223,
      "rewards/margins": 2.236736297607422,
      "rewards/rejected": -1.3440301418304443,
      "step": 100
    },
    {
      "epoch": 0.9345794392523364,
      "eval_logits/chosen": -2.2007782459259033,
      "eval_logits/rejected": -2.177567720413208,
      "eval_logps/chosen": -247.5850067138672,
      "eval_logps/rejected": -224.4142608642578,
      "eval_loss": 0.5101521015167236,
      "eval_rewards/accuracies": 0.7864583134651184,
      "eval_rewards/chosen": 0.8652558326721191,
      "eval_rewards/margins": 1.7787574529647827,
      "eval_rewards/rejected": -0.9135015606880188,
      "eval_runtime": 201.8599,
      "eval_samples_per_second": 15.06,
      "eval_steps_per_second": 0.238,
      "step": 100
    }
  ],
  "logging_steps": 5,
  "max_steps": 321,
  "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
}