{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.1008827238335435,
  "eval_steps": 10,
  "global_step": 40,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "computed_total": 0.45715999603271484,
      "epoch": 0.0025220680958385876,
      "grad_norm": 102.097412109375,
      "kl_loss": 6.5083294309431494e-09,
      "learning_rate": 0.0,
      "loss": 4.3237,
      "sft_loss": 0.45650917291641235,
      "step": 1
    },
    {
      "computed_total": 0.555546760559082,
      "epoch": 0.005044136191677175,
      "grad_norm": 87.49225616455078,
      "kl_loss": 5.4945092919922445e-09,
      "learning_rate": 4.000000000000001e-06,
      "loss": 4.3036,
      "sft_loss": 0.5549973249435425,
      "step": 2
    },
    {
      "computed_total": 0.5587829947471619,
      "epoch": 0.007566204287515763,
      "grad_norm": 64.52155303955078,
      "kl_loss": 3.4872137177899276e-08,
      "learning_rate": 8.000000000000001e-06,
      "loss": 3.5946,
      "sft_loss": 0.5552957653999329,
      "step": 3
    },
    {
      "computed_total": 0.4053139090538025,
      "epoch": 0.01008827238335435,
      "grad_norm": 79.32300567626953,
      "kl_loss": 7.649157964806363e-07,
      "learning_rate": 1.2e-05,
      "loss": 3.3472,
      "sft_loss": 0.3288223147392273,
      "step": 4
    },
    {
      "computed_total": 0.3922015130519867,
      "epoch": 0.012610340479192938,
      "grad_norm": 49.438594818115234,
      "kl_loss": 9.65095068750088e-07,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 3.0682,
      "sft_loss": 0.29569199681282043,
      "step": 5
    },
    {
      "computed_total": 0.3126932382583618,
      "epoch": 0.015132408575031526,
      "grad_norm": 36.26791000366211,
      "kl_loss": 9.634337629904621e-07,
      "learning_rate": 2e-05,
      "loss": 2.7815,
      "sft_loss": 0.21634985506534576,
      "step": 6
    },
    {
      "computed_total": 0.2588472068309784,
      "epoch": 0.017654476670870115,
      "grad_norm": 28.21232795715332,
      "kl_loss": 7.276251494658936e-07,
      "learning_rate": 1.9948979591836737e-05,
      "loss": 2.499,
      "sft_loss": 0.18608468770980835,
      "step": 7
    },
    {
      "computed_total": 0.32910704612731934,
      "epoch": 0.0201765447667087,
      "grad_norm": 24.3712158203125,
      "kl_loss": 8.822735821922834e-07,
      "learning_rate": 1.9897959183673473e-05,
      "loss": 2.4776,
      "sft_loss": 0.24087969958782196,
      "step": 8
    },
    {
      "computed_total": 0.3658309280872345,
      "epoch": 0.02269861286254729,
      "grad_norm": 31.98260498046875,
      "kl_loss": 7.195029070317105e-07,
      "learning_rate": 1.9846938775510205e-05,
      "loss": 2.3718,
      "sft_loss": 0.2938806414604187,
      "step": 9
    },
    {
      "computed_total": 0.22178040444850922,
      "epoch": 0.025220680958385876,
      "grad_norm": 34.0799674987793,
      "kl_loss": 6.166483217384666e-07,
      "learning_rate": 1.979591836734694e-05,
      "loss": 2.4129,
      "sft_loss": 0.1601155698299408,
      "step": 10
    },
    {
      "computed_total": 2.1096696853637695,
      "epoch": 0.025220680958385876,
      "eval_loss": 1.9816868305206299,
      "eval_runtime": 67.6013,
      "eval_samples_per_second": 10.429,
      "eval_steps_per_second": 5.222,
      "kl_loss": 3.9526437944914505e-07,
      "sft_loss": 2.070143222808838,
      "step": 10
    },
    {
      "computed_total": 0.4089103043079376,
      "epoch": 0.027742749054224466,
      "grad_norm": 26.954879760742188,
      "kl_loss": 4.996533107259893e-07,
      "learning_rate": 1.9744897959183677e-05,
      "loss": 2.277,
      "sft_loss": 0.35894498229026794,
      "step": 11
    },
    {
      "computed_total": 0.23354634642601013,
      "epoch": 0.03026481715006305,
      "grad_norm": 18.464841842651367,
      "kl_loss": 5.376580816118803e-07,
      "learning_rate": 1.969387755102041e-05,
      "loss": 2.2583,
      "sft_loss": 0.17978054285049438,
      "step": 12
    },
    {
      "computed_total": 0.30147990584373474,
      "epoch": 0.03278688524590164,
      "grad_norm": 26.595727920532227,
      "kl_loss": 5.573554062721087e-07,
      "learning_rate": 1.9642857142857145e-05,
      "loss": 2.2624,
      "sft_loss": 0.24574437737464905,
      "step": 13
    },
    {
      "computed_total": 0.24872392416000366,
      "epoch": 0.03530895334174023,
      "grad_norm": 16.613079071044922,
      "kl_loss": 4.973371687810868e-07,
      "learning_rate": 1.9591836734693877e-05,
      "loss": 2.1151,
      "sft_loss": 0.19899021089076996,
      "step": 14
    },
    {
      "computed_total": 0.19776779413223267,
      "epoch": 0.03783102143757881,
      "grad_norm": 15.080662727355957,
      "kl_loss": 4.196848806259368e-07,
      "learning_rate": 1.9540816326530613e-05,
      "loss": 2.0897,
      "sft_loss": 0.15579931437969208,
      "step": 15
    },
    {
      "computed_total": 0.284699946641922,
      "epoch": 0.0403530895334174,
      "grad_norm": 11.973522186279297,
      "kl_loss": 4.3317794506947394e-07,
      "learning_rate": 1.948979591836735e-05,
      "loss": 2.216,
      "sft_loss": 0.2413821518421173,
      "step": 16
    },
    {
      "computed_total": 0.21680015325546265,
      "epoch": 0.04287515762925599,
      "grad_norm": 11.934453964233398,
      "kl_loss": 3.477235850368743e-07,
      "learning_rate": 1.9438775510204085e-05,
      "loss": 1.9235,
      "sft_loss": 0.18202780187129974,
      "step": 17
    },
    {
      "computed_total": 0.18643273413181305,
      "epoch": 0.04539722572509458,
      "grad_norm": 10.074520111083984,
      "kl_loss": 2.590618919384724e-07,
      "learning_rate": 1.9387755102040817e-05,
      "loss": 1.8522,
      "sft_loss": 0.16052654385566711,
      "step": 18
    },
    {
      "computed_total": 0.22932234406471252,
      "epoch": 0.04791929382093316,
      "grad_norm": 15.449592590332031,
      "kl_loss": 2.511018522000086e-07,
      "learning_rate": 1.9336734693877553e-05,
      "loss": 1.7982,
      "sft_loss": 0.20421215891838074,
      "step": 19
    },
    {
      "computed_total": 0.29722726345062256,
      "epoch": 0.05044136191677175,
      "grad_norm": 12.395018577575684,
      "kl_loss": 2.6658420892999857e-07,
      "learning_rate": 1.928571428571429e-05,
      "loss": 1.817,
      "sft_loss": 0.27056884765625,
      "step": 20
    },
    {
      "computed_total": 1.6116889715194702,
      "epoch": 0.05044136191677175,
      "eval_loss": 1.5718247890472412,
      "eval_runtime": 69.8078,
      "eval_samples_per_second": 10.099,
      "eval_steps_per_second": 5.057,
      "kl_loss": 3.4212027344437956e-07,
      "sft_loss": 1.577476978302002,
      "step": 20
    },
    {
      "computed_total": 0.1908552497625351,
      "epoch": 0.05296343001261034,
      "grad_norm": 23.406896591186523,
      "kl_loss": 2.7672922442434356e-07,
      "learning_rate": 1.9234693877551024e-05,
      "loss": 1.7865,
      "sft_loss": 0.163182333111763,
      "step": 21
    },
    {
      "computed_total": 0.2098204642534256,
      "epoch": 0.05548549810844893,
      "grad_norm": 12.387871742248535,
      "kl_loss": 1.9662114425500477e-07,
      "learning_rate": 1.9183673469387756e-05,
      "loss": 1.6599,
      "sft_loss": 0.19015835225582123,
      "step": 22
    },
    {
      "computed_total": 0.27859383821487427,
      "epoch": 0.058007566204287514,
      "grad_norm": 17.064393997192383,
      "kl_loss": 2.6223989380014245e-07,
      "learning_rate": 1.9132653061224492e-05,
      "loss": 1.7114,
      "sft_loss": 0.25236985087394714,
      "step": 23
    },
    {
      "computed_total": 0.2661365270614624,
      "epoch": 0.0605296343001261,
      "grad_norm": 14.536602973937988,
      "kl_loss": 3.302961033568863e-07,
      "learning_rate": 1.9081632653061225e-05,
      "loss": 1.7381,
      "sft_loss": 0.23310692608356476,
      "step": 24
    },
    {
      "computed_total": 0.21774020791053772,
      "epoch": 0.06305170239596469,
      "grad_norm": 10.612632751464844,
      "kl_loss": 2.2093004758971801e-07,
      "learning_rate": 1.903061224489796e-05,
      "loss": 1.6493,
      "sft_loss": 0.195647194981575,
      "step": 25
    },
    {
      "computed_total": 0.21532197296619415,
      "epoch": 0.06557377049180328,
      "grad_norm": 13.36475944519043,
      "kl_loss": 1.9050941091336426e-07,
      "learning_rate": 1.8979591836734696e-05,
      "loss": 1.65,
      "sft_loss": 0.19627103209495544,
      "step": 26
    },
    {
      "computed_total": 0.21690420806407928,
      "epoch": 0.06809583858764187,
      "grad_norm": 13.064225196838379,
      "kl_loss": 2.1687725393348956e-07,
      "learning_rate": 1.892857142857143e-05,
      "loss": 1.7006,
      "sft_loss": 0.19521647691726685,
      "step": 27
    },
    {
      "computed_total": 0.19565589725971222,
      "epoch": 0.07061790668348046,
      "grad_norm": 12.320195198059082,
      "kl_loss": 2.17976335648018e-07,
      "learning_rate": 1.8877551020408164e-05,
      "loss": 1.6652,
      "sft_loss": 0.17385827004909515,
      "step": 28
    },
    {
      "computed_total": 0.27470341324806213,
      "epoch": 0.07313997477931904,
      "grad_norm": 11.224297523498535,
      "kl_loss": 2.520193618238409e-07,
      "learning_rate": 1.88265306122449e-05,
      "loss": 1.6883,
      "sft_loss": 0.24950148165225983,
      "step": 29
    },
    {
      "computed_total": 0.1886218637228012,
      "epoch": 0.07566204287515763,
      "grad_norm": 14.4179048538208,
      "kl_loss": 3.655400746538362e-07,
      "learning_rate": 1.8775510204081636e-05,
      "loss": 1.7232,
      "sft_loss": 0.1520678550004959,
      "step": 30
    },
    {
      "computed_total": 1.3417655229568481,
      "epoch": 0.07566204287515763,
      "eval_loss": 1.4369149208068848,
      "eval_runtime": 67.0852,
      "eval_samples_per_second": 10.509,
      "eval_steps_per_second": 5.262,
      "kl_loss": 2.1116991888447956e-07,
      "sft_loss": 1.3206485509872437,
      "step": 30
    },
    {
      "computed_total": 0.19783087074756622,
      "epoch": 0.07818411097099622,
      "grad_norm": 14.451069831848145,
      "kl_loss": 1.4020483263266215e-07,
      "learning_rate": 1.8724489795918368e-05,
      "loss": 1.6225,
      "sft_loss": 0.18381038308143616,
      "step": 31
    },
    {
      "computed_total": 0.2664089798927307,
      "epoch": 0.0807061790668348,
      "grad_norm": 10.6675443649292,
      "kl_loss": 2.558473113367654e-07,
      "learning_rate": 1.8673469387755104e-05,
      "loss": 1.6656,
      "sft_loss": 0.24082423746585846,
      "step": 32
    },
    {
      "computed_total": 0.16216596961021423,
      "epoch": 0.0832282471626734,
      "grad_norm": 10.89920711517334,
      "kl_loss": 3.102907157881418e-07,
      "learning_rate": 1.862244897959184e-05,
      "loss": 1.7422,
      "sft_loss": 0.13113689422607422,
      "step": 33
    },
    {
      "computed_total": 0.18052271008491516,
      "epoch": 0.08575031525851198,
      "grad_norm": 10.312054634094238,
      "kl_loss": 2.5258012215090275e-07,
      "learning_rate": 1.8571428571428575e-05,
      "loss": 1.5284,
      "sft_loss": 0.1552647054195404,
      "step": 34
    },
    {
      "computed_total": 0.13403812050819397,
      "epoch": 0.08827238335435057,
      "grad_norm": 9.937405586242676,
      "kl_loss": 1.9699928088812158e-07,
      "learning_rate": 1.8520408163265307e-05,
      "loss": 1.6276,
      "sft_loss": 0.11433819681406021,
      "step": 35
    },
    {
      "computed_total": 0.21145765483379364,
      "epoch": 0.09079445145018916,
      "grad_norm": 10.810340881347656,
      "kl_loss": 2.5251029001083225e-07,
      "learning_rate": 1.8469387755102043e-05,
      "loss": 1.5674,
      "sft_loss": 0.1862066239118576,
      "step": 36
    },
    {
      "computed_total": 0.28667497634887695,
      "epoch": 0.09331651954602774,
      "grad_norm": 13.322245597839355,
      "kl_loss": 2.524957380956039e-07,
      "learning_rate": 1.8418367346938776e-05,
      "loss": 1.6924,
      "sft_loss": 0.2614254057407379,
      "step": 37
    },
    {
      "computed_total": 0.17698872089385986,
      "epoch": 0.09583858764186633,
      "grad_norm": 12.67447566986084,
      "kl_loss": 1.8715130067903374e-07,
      "learning_rate": 1.836734693877551e-05,
      "loss": 1.6326,
      "sft_loss": 0.1582735925912857,
      "step": 38
    },
    {
      "computed_total": 0.218657985329628,
      "epoch": 0.09836065573770492,
      "grad_norm": 10.07801342010498,
      "kl_loss": 2.1152843032723467e-07,
      "learning_rate": 1.8316326530612247e-05,
      "loss": 1.5745,
      "sft_loss": 0.1975051462650299,
      "step": 39
    },
    {
      "computed_total": 0.1585373878479004,
      "epoch": 0.1008827238335435,
      "grad_norm": 10.709808349609375,
      "kl_loss": 1.5567046318665234e-07,
      "learning_rate": 1.826530612244898e-05,
      "loss": 1.4899,
      "sft_loss": 0.14297033846378326,
      "step": 40
    },
    {
      "computed_total": 1.3272018432617188,
      "epoch": 0.1008827238335435,
      "eval_loss": 1.3992165327072144,
      "eval_runtime": 67.3288,
      "eval_samples_per_second": 10.471,
      "eval_steps_per_second": 5.243,
      "kl_loss": 1.587673352787533e-07,
      "sft_loss": 1.3113250732421875,
      "step": 40
    }
  ],
  "logging_steps": 1,
  "max_steps": 397,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 10,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 7317006936207360.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}