{
  "best_metric": 0.8123041987419128,
  "best_model_checkpoint": "./output/training_results/C013_llama3-8b-base_instruct_20240428_005832/checkpoint-15",
  "epoch": 4.0,
  "eval_steps": 5,
  "global_step": 192,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.020833333333333332,
      "grad_norm": 0.0,
      "learning_rate": 0.0,
      "loss": 0.9805,
      "step": 1
    },
    {
      "epoch": 0.020833333333333332,
      "eval_loss": 0.9736970067024231,
      "eval_runtime": 2.153,
      "eval_samples_per_second": 157.916,
      "eval_steps_per_second": 1.393,
      "step": 1
    },
    {
      "epoch": 0.10416666666666667,
      "grad_norm": 14.850728211706278,
      "learning_rate": 1.5e-06,
      "loss": 0.9446,
      "step": 5
    },
    {
      "epoch": 0.10416666666666667,
      "eval_loss": 0.9454841613769531,
      "eval_runtime": 2.0973,
      "eval_samples_per_second": 162.11,
      "eval_steps_per_second": 1.43,
      "step": 5
    },
    {
      "epoch": 0.20833333333333334,
      "grad_norm": 4.950599387514031,
      "learning_rate": 5.25e-06,
      "loss": 0.8481,
      "step": 10
    },
    {
      "epoch": 0.20833333333333334,
      "eval_loss": 0.8153812289237976,
      "eval_runtime": 2.0923,
      "eval_samples_per_second": 162.499,
      "eval_steps_per_second": 1.434,
      "step": 10
    },
    {
      "epoch": 0.3125,
      "grad_norm": 4.621063619275185,
      "learning_rate": 9e-06,
      "loss": 0.7794,
      "step": 15
    },
    {
      "epoch": 0.3125,
      "eval_loss": 0.8123041987419128,
      "eval_runtime": 2.1028,
      "eval_samples_per_second": 161.686,
      "eval_steps_per_second": 1.427,
      "step": 15
    },
    {
      "epoch": 0.4166666666666667,
      "grad_norm": 4.141809373286457,
      "learning_rate": 1.275e-05,
      "loss": 0.7798,
      "step": 20
    },
    {
      "epoch": 0.4166666666666667,
      "eval_loss": 0.8410752415657043,
      "eval_runtime": 2.0891,
      "eval_samples_per_second": 162.747,
      "eval_steps_per_second": 1.436,
      "step": 20
    },
    {
      "epoch": 0.5208333333333334,
      "grad_norm": 4.211750921552142,
      "learning_rate": 1.3195176200175283e-05,
      "loss": 0.8576,
      "step": 25
    },
    {
      "epoch": 0.5208333333333334,
      "eval_loss": 0.8676239848136902,
      "eval_runtime": 2.0885,
      "eval_samples_per_second": 162.793,
      "eval_steps_per_second": 1.436,
      "step": 25
    },
    {
      "epoch": 0.625,
      "grad_norm": 4.126229536438554,
      "learning_rate": 9.515676612044427e-06,
      "loss": 0.8852,
      "step": 30
    },
    {
      "epoch": 0.625,
      "eval_loss": 0.867268979549408,
      "eval_runtime": 2.0839,
      "eval_samples_per_second": 163.157,
      "eval_steps_per_second": 1.44,
      "step": 30
    },
    {
      "epoch": 0.7291666666666666,
      "grad_norm": 4.316589185885892,
      "learning_rate": 6.797580677308734e-06,
      "loss": 0.8529,
      "step": 35
    },
    {
      "epoch": 0.7291666666666666,
      "eval_loss": 0.8560981154441833,
      "eval_runtime": 2.1307,
      "eval_samples_per_second": 159.573,
      "eval_steps_per_second": 1.408,
      "step": 35
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 4.0216031828158005,
      "learning_rate": 4.808575415542887e-06,
      "loss": 0.8224,
      "step": 40
    },
    {
      "epoch": 0.8333333333333334,
      "eval_loss": 0.8470456004142761,
      "eval_runtime": 2.0873,
      "eval_samples_per_second": 162.886,
      "eval_steps_per_second": 1.437,
      "step": 40
    },
    {
      "epoch": 0.9375,
      "grad_norm": 4.316706720311178,
      "learning_rate": 3.3676619069852654e-06,
      "loss": 0.8536,
      "step": 45
    },
    {
      "epoch": 0.9375,
      "eval_loss": 0.8378292918205261,
      "eval_runtime": 2.0847,
      "eval_samples_per_second": 163.089,
      "eval_steps_per_second": 1.439,
      "step": 45
    },
    {
      "epoch": 1.0416666666666667,
      "grad_norm": 3.7957934185208795,
      "learning_rate": 2.334947896124909e-06,
      "loss": 0.662,
      "step": 50
    },
    {
      "epoch": 1.0416666666666667,
      "eval_loss": 0.8293696045875549,
      "eval_runtime": 2.0835,
      "eval_samples_per_second": 163.187,
      "eval_steps_per_second": 1.44,
      "step": 50
    },
    {
      "epoch": 1.1458333333333333,
      "grad_norm": 3.4155908301931186,
      "learning_rate": 1.603233215095547e-06,
      "loss": 0.437,
      "step": 55
    },
    {
      "epoch": 1.1458333333333333,
      "eval_loss": 0.8531150817871094,
      "eval_runtime": 2.1006,
      "eval_samples_per_second": 161.859,
      "eval_steps_per_second": 1.428,
      "step": 55
    },
    {
      "epoch": 1.25,
      "grad_norm": 3.377214905899517,
      "learning_rate": 1.0911174606561334e-06,
      "loss": 0.4402,
      "step": 60
    },
    {
      "epoch": 1.25,
      "eval_loss": 0.8569180369377136,
      "eval_runtime": 2.0899,
      "eval_samples_per_second": 162.69,
      "eval_steps_per_second": 1.436,
      "step": 60
    },
    {
      "epoch": 1.3541666666666667,
      "grad_norm": 4.018786896199577,
      "learning_rate": 7.373930741131784e-07,
      "loss": 0.4244,
      "step": 65
    },
    {
      "epoch": 1.3541666666666667,
      "eval_loss": 0.8569238185882568,
      "eval_runtime": 2.0969,
      "eval_samples_per_second": 162.148,
      "eval_steps_per_second": 1.431,
      "step": 65
    },
    {
      "epoch": 1.4583333333333333,
      "grad_norm": 4.3050060673581205,
      "learning_rate": 5.374210410959207e-07,
      "loss": 0.4495,
      "step": 70
    },
    {
      "epoch": 1.4583333333333333,
      "eval_loss": 0.8547163605690002,
      "eval_runtime": 2.0852,
      "eval_samples_per_second": 163.056,
      "eval_steps_per_second": 1.439,
      "step": 70
    },
    {
      "epoch": 1.5625,
      "grad_norm": 3.8753963390823842,
      "learning_rate": 3.6222476698215175e-07,
      "loss": 0.4689,
      "step": 75
    },
    {
      "epoch": 1.5625,
      "eval_loss": 0.8493571877479553,
      "eval_runtime": 2.1006,
      "eval_samples_per_second": 161.855,
      "eval_steps_per_second": 1.428,
      "step": 75
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 3.2777220151938935,
      "learning_rate": 2.462755297384099e-07,
      "loss": 0.4309,
      "step": 80
    },
    {
      "epoch": 1.6666666666666665,
      "eval_loss": 0.846055269241333,
      "eval_runtime": 2.0775,
      "eval_samples_per_second": 163.657,
      "eval_steps_per_second": 1.444,
      "step": 80
    },
    {
      "epoch": 1.7708333333333335,
      "grad_norm": 3.25027538013195,
      "learning_rate": 1.7088740175034947e-07,
      "loss": 0.4299,
      "step": 85
    },
    {
      "epoch": 1.7708333333333335,
      "eval_loss": 0.8445951342582703,
      "eval_runtime": 2.0859,
      "eval_samples_per_second": 163.002,
      "eval_steps_per_second": 1.438,
      "step": 85
    },
    {
      "epoch": 1.875,
      "grad_norm": 3.841600887262257,
      "learning_rate": 1.228102956599465e-07,
      "loss": 0.4461,
      "step": 90
    },
    {
      "epoch": 1.875,
      "eval_loss": 0.8440027832984924,
      "eval_runtime": 2.099,
      "eval_samples_per_second": 161.984,
      "eval_steps_per_second": 1.429,
      "step": 90
    },
    {
      "epoch": 1.9791666666666665,
      "grad_norm": 4.633157495322692,
      "learning_rate": 9.279207916081227e-08,
      "loss": 0.4474,
      "step": 95
    },
    {
      "epoch": 1.9791666666666665,
      "eval_loss": 0.8438854217529297,
      "eval_runtime": 2.094,
      "eval_samples_per_second": 162.368,
      "eval_steps_per_second": 1.433,
      "step": 95
    },
    {
      "epoch": 2.0833333333333335,
      "grad_norm": 3.3543713588136885,
      "learning_rate": 7.448002404850094e-08,
      "loss": 0.3614,
      "step": 100
    },
    {
      "epoch": 2.0833333333333335,
      "eval_loss": 0.8445320725440979,
      "eval_runtime": 2.0778,
      "eval_samples_per_second": 163.634,
      "eval_steps_per_second": 1.444,
      "step": 100
    },
    {
      "epoch": 2.1875,
      "grad_norm": 3.5776096289343053,
      "learning_rate": 6.35920070839697e-08,
      "loss": 0.3861,
      "step": 105
    },
    {
      "epoch": 2.1875,
      "eval_loss": 0.8457441926002502,
      "eval_runtime": 2.1055,
      "eval_samples_per_second": 161.484,
      "eval_steps_per_second": 1.425,
      "step": 105
    },
    {
      "epoch": 2.2916666666666665,
      "grad_norm": 3.811456756438563,
      "learning_rate": 5.7299804687499997e-08,
      "loss": 0.3829,
      "step": 110
    },
    {
      "epoch": 2.2916666666666665,
      "eval_loss": 0.847288191318512,
      "eval_runtime": 2.083,
      "eval_samples_per_second": 163.223,
      "eval_steps_per_second": 1.44,
      "step": 110
    },
    {
      "epoch": 2.3958333333333335,
      "grad_norm": 3.1978758437608823,
      "learning_rate": 5.37771434967624e-08,
      "loss": 0.3764,
      "step": 115
    },
    {
      "epoch": 2.3958333333333335,
      "eval_loss": 0.8487641215324402,
      "eval_runtime": 2.1168,
      "eval_samples_per_second": 160.617,
      "eval_steps_per_second": 1.417,
      "step": 115
    },
    {
      "epoch": 2.5,
      "grad_norm": 3.472352228062058,
      "learning_rate": 5.187403540619925e-08,
      "loss": 0.3655,
      "step": 120
    },
    {
      "epoch": 2.5,
      "eval_loss": 0.8499611020088196,
      "eval_runtime": 2.0908,
      "eval_samples_per_second": 162.615,
      "eval_steps_per_second": 1.435,
      "step": 120
    },
    {
      "epoch": 2.6041666666666665,
      "grad_norm": 3.2298459394815793,
      "learning_rate": 5.088648238966908e-08,
      "loss": 0.4243,
      "step": 125
    },
    {
      "epoch": 2.6041666666666665,
      "eval_loss": 0.8510637879371643,
      "eval_runtime": 2.0941,
      "eval_samples_per_second": 162.36,
      "eval_steps_per_second": 1.433,
      "step": 125
    },
    {
      "epoch": 2.7083333333333335,
      "grad_norm": 3.7544587648641756,
      "learning_rate": 5.039701925276604e-08,
      "loss": 0.3884,
      "step": 130
    },
    {
      "epoch": 2.7083333333333335,
      "eval_loss": 0.8520172238349915,
      "eval_runtime": 2.1032,
      "eval_samples_per_second": 161.66,
      "eval_steps_per_second": 1.426,
      "step": 130
    },
    {
      "epoch": 2.8125,
      "grad_norm": 3.5032769257867695,
      "learning_rate": 5.0166900048082497e-08,
      "loss": 0.3634,
      "step": 135
    },
    {
      "epoch": 2.8125,
      "eval_loss": 0.8528143763542175,
      "eval_runtime": 2.0786,
      "eval_samples_per_second": 163.568,
      "eval_steps_per_second": 1.443,
      "step": 135
    },
    {
      "epoch": 2.9166666666666665,
      "grad_norm": 3.023294292675947,
      "learning_rate": 5.0065147322870076e-08,
      "loss": 0.3846,
      "step": 140
    },
    {
      "epoch": 2.9166666666666665,
      "eval_loss": 0.8537066578865051,
      "eval_runtime": 2.0903,
      "eval_samples_per_second": 162.659,
      "eval_steps_per_second": 1.435,
      "step": 140
    },
    {
      "epoch": 3.0208333333333335,
      "grad_norm": 3.1767015238154075,
      "learning_rate": 5.002328628528332e-08,
      "loss": 0.3872,
      "step": 145
    },
    {
      "epoch": 3.0208333333333335,
      "eval_loss": 0.8547406196594238,
      "eval_runtime": 2.0891,
      "eval_samples_per_second": 162.748,
      "eval_steps_per_second": 1.436,
      "step": 145
    },
    {
      "epoch": 3.125,
      "grad_norm": 3.1942747338221045,
      "learning_rate": 5.0007484528133236e-08,
      "loss": 0.3869,
      "step": 150
    },
    {
      "epoch": 3.125,
      "eval_loss": 0.8557960391044617,
      "eval_runtime": 2.0819,
      "eval_samples_per_second": 163.312,
      "eval_steps_per_second": 1.441,
      "step": 150
    },
    {
      "epoch": 3.2291666666666665,
      "grad_norm": 3.815918812229993,
      "learning_rate": 5.0002110817570477e-08,
      "loss": 0.3876,
      "step": 155
    },
    {
      "epoch": 3.2291666666666665,
      "eval_loss": 0.8566272854804993,
      "eval_runtime": 2.0781,
      "eval_samples_per_second": 163.61,
      "eval_steps_per_second": 1.444,
      "step": 155
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 3.4577646975309366,
      "learning_rate": 5.0000504842356326e-08,
      "loss": 0.3844,
      "step": 160
    },
    {
      "epoch": 3.3333333333333335,
      "eval_loss": 0.8572790026664734,
      "eval_runtime": 2.0811,
      "eval_samples_per_second": 163.373,
      "eval_steps_per_second": 1.442,
      "step": 160
    },
    {
      "epoch": 3.4375,
      "grad_norm": 3.274685205370877,
      "learning_rate": 5.000009745562451e-08,
      "loss": 0.3535,
      "step": 165
    },
    {
      "epoch": 3.4375,
      "eval_loss": 0.8578632473945618,
      "eval_runtime": 2.0918,
      "eval_samples_per_second": 162.539,
      "eval_steps_per_second": 1.434,
      "step": 165
    },
    {
      "epoch": 3.5416666666666665,
      "grad_norm": 3.246459205886974,
      "learning_rate": 5.0000014077810156e-08,
      "loss": 0.3488,
      "step": 170
    },
    {
      "epoch": 3.5416666666666665,
      "eval_loss": 0.85884028673172,
      "eval_runtime": 2.1178,
      "eval_samples_per_second": 160.545,
      "eval_steps_per_second": 1.417,
      "step": 170
    },
    {
      "epoch": 3.6458333333333335,
      "grad_norm": 3.3944513203963504,
      "learning_rate": 5.0000001343508807e-08,
      "loss": 0.3464,
      "step": 175
    },
    {
      "epoch": 3.6458333333333335,
      "eval_loss": 0.8598365783691406,
      "eval_runtime": 2.0828,
      "eval_samples_per_second": 163.238,
      "eval_steps_per_second": 1.44,
      "step": 175
    },
    {
      "epoch": 3.75,
      "grad_norm": 3.258773113208273,
      "learning_rate": 5.000000006747581e-08,
      "loss": 0.361,
      "step": 180
    },
    {
      "epoch": 3.75,
      "eval_loss": 0.8606703877449036,
      "eval_runtime": 2.1172,
      "eval_samples_per_second": 160.588,
      "eval_steps_per_second": 1.417,
      "step": 180
    },
    {
      "epoch": 3.8541666666666665,
      "grad_norm": 3.586703083699586,
      "learning_rate": 5.0000000001094325e-08,
      "loss": 0.3674,
      "step": 185
    },
    {
      "epoch": 3.8541666666666665,
      "eval_loss": 0.8611735701560974,
      "eval_runtime": 2.0956,
      "eval_samples_per_second": 162.243,
      "eval_steps_per_second": 1.432,
      "step": 185
    },
    {
      "epoch": 3.9583333333333335,
      "grad_norm": 3.5661429802112616,
      "learning_rate": 5.000000000000139e-08,
      "loss": 0.3988,
      "step": 190
    },
    {
      "epoch": 3.9583333333333335,
      "eval_loss": 0.8612277507781982,
      "eval_runtime": 2.0853,
      "eval_samples_per_second": 163.045,
      "eval_steps_per_second": 1.439,
      "step": 190
    },
    {
      "epoch": 4.0,
      "step": 192,
      "total_flos": 5363820134400.0,
      "train_loss": 0.5090278356025616,
      "train_runtime": 6014.9673,
      "train_samples_per_second": 2.031,
      "train_steps_per_second": 0.032
    }
  ],
  "logging_steps": 5,
  "max_steps": 192,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 5,
  "total_flos": 5363820134400.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}