{
  "best_metric": 0.39516589045524597,
  "best_model_checkpoint": "mikhail_panzo/ceb_b64_le4_s4000/checkpoint-3500",
  "epoch": 138.6138613861386,
  "eval_steps": 500,
  "global_step": 3500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.9801980198019802,
      "grad_norm": 1.725104808807373,
      "learning_rate": 2.5e-06,
      "loss": 0.7809,
      "step": 50
    },
    {
      "epoch": 3.9603960396039604,
      "grad_norm": 1.6038875579833984,
      "learning_rate": 5e-06,
      "loss": 0.7005,
      "step": 100
    },
    {
      "epoch": 5.9405940594059405,
      "grad_norm": 1.6793955564498901,
      "learning_rate": 7.5e-06,
      "loss": 0.6246,
      "step": 150
    },
    {
      "epoch": 7.920792079207921,
      "grad_norm": 1.5051871538162231,
      "learning_rate": 1e-05,
      "loss": 0.5278,
      "step": 200
    },
    {
      "epoch": 9.900990099009901,
      "grad_norm": 1.404683232307434,
      "learning_rate": 1.25e-05,
      "loss": 0.5095,
      "step": 250
    },
    {
      "epoch": 11.881188118811881,
      "grad_norm": 1.248382568359375,
      "learning_rate": 1.5e-05,
      "loss": 0.4814,
      "step": 300
    },
    {
      "epoch": 13.861386138613861,
      "grad_norm": 0.995944082736969,
      "learning_rate": 1.75e-05,
      "loss": 0.4743,
      "step": 350
    },
    {
      "epoch": 15.841584158415841,
      "grad_norm": 1.472835659980774,
      "learning_rate": 2e-05,
      "loss": 0.4631,
      "step": 400
    },
    {
      "epoch": 17.821782178217823,
      "grad_norm": 2.442906618118286,
      "learning_rate": 2.245e-05,
      "loss": 0.4622,
      "step": 450
    },
    {
      "epoch": 19.801980198019802,
      "grad_norm": 1.071074366569519,
      "learning_rate": 2.495e-05,
      "loss": 0.4561,
      "step": 500
    },
    {
      "epoch": 19.801980198019802,
      "eval_loss": 0.41511112451553345,
      "eval_runtime": 7.4907,
      "eval_samples_per_second": 24.03,
      "eval_steps_per_second": 3.07,
      "step": 500
    },
    {
      "epoch": 21.782178217821784,
      "grad_norm": 1.0462185144424438,
      "learning_rate": 2.7450000000000003e-05,
      "loss": 0.4465,
      "step": 550
    },
    {
      "epoch": 23.762376237623762,
      "grad_norm": 1.103574275970459,
      "learning_rate": 2.995e-05,
      "loss": 0.4453,
      "step": 600
    },
    {
      "epoch": 25.742574257425744,
      "grad_norm": 3.00575590133667,
      "learning_rate": 3.245e-05,
      "loss": 0.447,
      "step": 650
    },
    {
      "epoch": 27.722772277227723,
      "grad_norm": 1.786911129951477,
      "learning_rate": 3.495e-05,
      "loss": 0.4351,
      "step": 700
    },
    {
      "epoch": 29.702970297029704,
      "grad_norm": 1.236941933631897,
      "learning_rate": 3.745e-05,
      "loss": 0.4347,
      "step": 750
    },
    {
      "epoch": 31.683168316831683,
      "grad_norm": 1.3743062019348145,
      "learning_rate": 3.995e-05,
      "loss": 0.4319,
      "step": 800
    },
    {
      "epoch": 33.663366336633665,
      "grad_norm": 2.7615420818328857,
      "learning_rate": 4.245e-05,
      "loss": 0.4358,
      "step": 850
    },
    {
      "epoch": 35.64356435643565,
      "grad_norm": 1.662369966506958,
      "learning_rate": 4.495e-05,
      "loss": 0.4276,
      "step": 900
    },
    {
      "epoch": 37.62376237623762,
      "grad_norm": 1.0967382192611694,
      "learning_rate": 4.745e-05,
      "loss": 0.4267,
      "step": 950
    },
    {
      "epoch": 39.603960396039604,
      "grad_norm": 2.530874252319336,
      "learning_rate": 4.995e-05,
      "loss": 0.4179,
      "step": 1000
    },
    {
      "epoch": 39.603960396039604,
      "eval_loss": 0.39941468834877014,
      "eval_runtime": 7.3656,
      "eval_samples_per_second": 24.438,
      "eval_steps_per_second": 3.123,
      "step": 1000
    },
    {
      "epoch": 41.584158415841586,
      "grad_norm": 2.8653476238250732,
      "learning_rate": 5.245e-05,
      "loss": 0.4268,
      "step": 1050
    },
    {
      "epoch": 43.56435643564357,
      "grad_norm": 1.5550223588943481,
      "learning_rate": 5.495e-05,
      "loss": 0.4265,
      "step": 1100
    },
    {
      "epoch": 45.54455445544554,
      "grad_norm": 1.804150104522705,
      "learning_rate": 5.745e-05,
      "loss": 0.4192,
      "step": 1150
    },
    {
      "epoch": 47.524752475247524,
      "grad_norm": 1.9916889667510986,
      "learning_rate": 5.995000000000001e-05,
      "loss": 0.4149,
      "step": 1200
    },
    {
      "epoch": 49.504950495049506,
      "grad_norm": 2.1027019023895264,
      "learning_rate": 6.245000000000001e-05,
      "loss": 0.4203,
      "step": 1250
    },
    {
      "epoch": 51.48514851485149,
      "grad_norm": 1.1542466878890991,
      "learning_rate": 6.494999999999999e-05,
      "loss": 0.4127,
      "step": 1300
    },
    {
      "epoch": 53.46534653465346,
      "grad_norm": 1.8733513355255127,
      "learning_rate": 6.745e-05,
      "loss": 0.4165,
      "step": 1350
    },
    {
      "epoch": 55.445544554455445,
      "grad_norm": 2.544435739517212,
      "learning_rate": 6.995e-05,
      "loss": 0.4156,
      "step": 1400
    },
    {
      "epoch": 57.42574257425743,
      "grad_norm": 2.9764773845672607,
      "learning_rate": 7.245000000000001e-05,
      "loss": 0.4045,
      "step": 1450
    },
    {
      "epoch": 59.40594059405941,
      "grad_norm": 1.334035038948059,
      "learning_rate": 7.495e-05,
      "loss": 0.4075,
      "step": 1500
    },
    {
      "epoch": 59.40594059405941,
      "eval_loss": 0.40177610516548157,
      "eval_runtime": 6.6337,
      "eval_samples_per_second": 27.134,
      "eval_steps_per_second": 3.467,
      "step": 1500
    },
    {
      "epoch": 61.386138613861384,
      "grad_norm": 2.3007051944732666,
      "learning_rate": 7.745e-05,
      "loss": 0.4067,
      "step": 1550
    },
    {
      "epoch": 63.366336633663366,
      "grad_norm": 0.9966986179351807,
      "learning_rate": 7.995e-05,
      "loss": 0.4042,
      "step": 1600
    },
    {
      "epoch": 65.34653465346534,
      "grad_norm": 1.4066482782363892,
      "learning_rate": 8.245e-05,
      "loss": 0.4079,
      "step": 1650
    },
    {
      "epoch": 67.32673267326733,
      "grad_norm": 3.3195865154266357,
      "learning_rate": 8.495e-05,
      "loss": 0.4061,
      "step": 1700
    },
    {
      "epoch": 69.3069306930693,
      "grad_norm": 2.83154559135437,
      "learning_rate": 8.745000000000001e-05,
      "loss": 0.4028,
      "step": 1750
    },
    {
      "epoch": 71.2871287128713,
      "grad_norm": 1.5752816200256348,
      "learning_rate": 8.995e-05,
      "loss": 0.3977,
      "step": 1800
    },
    {
      "epoch": 73.26732673267327,
      "grad_norm": 1.8909986019134521,
      "learning_rate": 9.245e-05,
      "loss": 0.4013,
      "step": 1850
    },
    {
      "epoch": 75.24752475247524,
      "grad_norm": 4.082262992858887,
      "learning_rate": 9.495e-05,
      "loss": 0.3991,
      "step": 1900
    },
    {
      "epoch": 77.22772277227723,
      "grad_norm": 1.8281221389770508,
      "learning_rate": 9.745000000000001e-05,
      "loss": 0.4011,
      "step": 1950
    },
    {
      "epoch": 79.20792079207921,
      "grad_norm": 2.2827675342559814,
      "learning_rate": 9.995e-05,
      "loss": 0.3981,
      "step": 2000
    },
    {
      "epoch": 79.20792079207921,
      "eval_loss": 0.40288153290748596,
      "eval_runtime": 7.4596,
      "eval_samples_per_second": 24.13,
      "eval_steps_per_second": 3.083,
      "step": 2000
    },
    {
      "epoch": 81.18811881188118,
      "grad_norm": 2.13647723197937,
      "learning_rate": 9.755000000000001e-05,
      "loss": 0.3994,
      "step": 2050
    },
    {
      "epoch": 83.16831683168317,
      "grad_norm": 1.8043389320373535,
      "learning_rate": 9.505e-05,
      "loss": 0.3982,
      "step": 2100
    },
    {
      "epoch": 85.14851485148515,
      "grad_norm": 1.0940660238265991,
      "learning_rate": 9.255e-05,
      "loss": 0.3926,
      "step": 2150
    },
    {
      "epoch": 87.12871287128714,
      "grad_norm": 2.0112838745117188,
      "learning_rate": 9.005000000000001e-05,
      "loss": 0.3822,
      "step": 2200
    },
    {
      "epoch": 89.10891089108911,
      "grad_norm": 1.5353419780731201,
      "learning_rate": 8.755e-05,
      "loss": 0.3857,
      "step": 2250
    },
    {
      "epoch": 91.08910891089108,
      "grad_norm": 2.0039117336273193,
      "learning_rate": 8.505000000000001e-05,
      "loss": 0.3887,
      "step": 2300
    },
    {
      "epoch": 93.06930693069307,
      "grad_norm": 1.2098206281661987,
      "learning_rate": 8.255e-05,
      "loss": 0.3874,
      "step": 2350
    },
    {
      "epoch": 95.04950495049505,
      "grad_norm": 0.9372404217720032,
      "learning_rate": 8.005000000000001e-05,
      "loss": 0.3816,
      "step": 2400
    },
    {
      "epoch": 97.02970297029702,
      "grad_norm": 0.9278631806373596,
      "learning_rate": 7.755e-05,
      "loss": 0.3797,
      "step": 2450
    },
    {
      "epoch": 99.00990099009901,
      "grad_norm": 1.0212403535842896,
      "learning_rate": 7.505e-05,
      "loss": 0.384,
      "step": 2500
    },
    {
      "epoch": 99.00990099009901,
      "eval_loss": 0.3953019976615906,
      "eval_runtime": 6.9648,
      "eval_samples_per_second": 25.844,
      "eval_steps_per_second": 3.302,
      "step": 2500
    },
    {
      "epoch": 100.99009900990099,
      "grad_norm": 2.3658339977264404,
      "learning_rate": 7.255e-05,
      "loss": 0.3793,
      "step": 2550
    },
    {
      "epoch": 102.97029702970298,
      "grad_norm": 1.4890061616897583,
      "learning_rate": 7.005000000000001e-05,
      "loss": 0.3823,
      "step": 2600
    },
    {
      "epoch": 104.95049504950495,
      "grad_norm": 2.134577512741089,
      "learning_rate": 6.755e-05,
      "loss": 0.3735,
      "step": 2650
    },
    {
      "epoch": 106.93069306930693,
      "grad_norm": 1.2545772790908813,
      "learning_rate": 6.505e-05,
      "loss": 0.3788,
      "step": 2700
    },
    {
      "epoch": 108.91089108910892,
      "grad_norm": 1.011292815208435,
      "learning_rate": 6.255e-05,
      "loss": 0.3762,
      "step": 2750
    },
    {
      "epoch": 110.89108910891089,
      "grad_norm": 1.4628199338912964,
      "learning_rate": 6.005000000000001e-05,
      "loss": 0.3742,
      "step": 2800
    },
    {
      "epoch": 112.87128712871286,
      "grad_norm": 1.661889910697937,
      "learning_rate": 5.755e-05,
      "loss": 0.3778,
      "step": 2850
    },
    {
      "epoch": 114.85148514851485,
      "grad_norm": 1.1926722526550293,
      "learning_rate": 5.505e-05,
      "loss": 0.3713,
      "step": 2900
    },
    {
      "epoch": 116.83168316831683,
      "grad_norm": 0.9905158877372742,
      "learning_rate": 5.255e-05,
      "loss": 0.3714,
      "step": 2950
    },
    {
      "epoch": 118.81188118811882,
      "grad_norm": 0.7342912554740906,
      "learning_rate": 5.005e-05,
      "loss": 0.3688,
      "step": 3000
    },
    {
      "epoch": 118.81188118811882,
      "eval_loss": 0.39639872312545776,
      "eval_runtime": 7.0442,
      "eval_samples_per_second": 25.553,
      "eval_steps_per_second": 3.265,
      "step": 3000
    },
    {
      "epoch": 120.79207920792079,
      "grad_norm": 0.7711001634597778,
      "learning_rate": 4.755e-05,
      "loss": 0.3657,
      "step": 3050
    },
    {
      "epoch": 122.77227722772277,
      "grad_norm": 0.9125708937644958,
      "learning_rate": 4.5050000000000004e-05,
      "loss": 0.3665,
      "step": 3100
    },
    {
      "epoch": 124.75247524752476,
      "grad_norm": 1.138588786125183,
      "learning_rate": 4.2550000000000004e-05,
      "loss": 0.3714,
      "step": 3150
    },
    {
      "epoch": 126.73267326732673,
      "grad_norm": 0.8590717315673828,
      "learning_rate": 4.0050000000000004e-05,
      "loss": 0.3696,
      "step": 3200
    },
    {
      "epoch": 128.7128712871287,
      "grad_norm": 0.8743947148323059,
      "learning_rate": 3.7550000000000005e-05,
      "loss": 0.3652,
      "step": 3250
    },
    {
      "epoch": 130.69306930693068,
      "grad_norm": 0.8117642998695374,
      "learning_rate": 3.505e-05,
      "loss": 0.3643,
      "step": 3300
    },
    {
      "epoch": 132.67326732673268,
      "grad_norm": 0.9085512161254883,
      "learning_rate": 3.2550000000000005e-05,
      "loss": 0.3637,
      "step": 3350
    },
    {
      "epoch": 134.65346534653466,
      "grad_norm": 0.8282843232154846,
      "learning_rate": 3.0050000000000002e-05,
      "loss": 0.3594,
      "step": 3400
    },
    {
      "epoch": 136.63366336633663,
      "grad_norm": 0.6631731390953064,
      "learning_rate": 2.7550000000000002e-05,
      "loss": 0.3621,
      "step": 3450
    },
    {
      "epoch": 138.6138613861386,
      "grad_norm": 0.8194773197174072,
      "learning_rate": 2.5050000000000002e-05,
      "loss": 0.3628,
      "step": 3500
    },
    {
      "epoch": 138.6138613861386,
      "eval_loss": 0.39516589045524597,
      "eval_runtime": 7.3015,
      "eval_samples_per_second": 24.652,
      "eval_steps_per_second": 3.15,
      "step": 3500
    }
  ],
  "logging_steps": 50,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 160,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 3.78207684271391e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}