{
  "best_metric": 0.4110700488090515,
  "best_model_checkpoint": "mikhail_panzo/fil_b64_le4_s4000/checkpoint-2500",
  "epoch": 133.33333333333334,
  "eval_steps": 500,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 3.1291983127593994,
      "learning_rate": 2.5e-06,
      "loss": 0.7845,
      "step": 50
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 1.4106616973876953,
      "learning_rate": 5e-06,
      "loss": 0.7033,
      "step": 100
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 3.4416260719299316,
      "learning_rate": 7.5e-06,
      "loss": 0.6535,
      "step": 150
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 1.7668957710266113,
      "learning_rate": 1e-05,
      "loss": 0.5726,
      "step": 200
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 1.3341814279556274,
      "learning_rate": 1.25e-05,
      "loss": 0.5312,
      "step": 250
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 1.4352439641952515,
      "learning_rate": 1.5e-05,
      "loss": 0.5094,
      "step": 300
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 1.0031296014785767,
      "learning_rate": 1.75e-05,
      "loss": 0.495,
      "step": 350
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 3.518950939178467,
      "learning_rate": 2e-05,
      "loss": 0.489,
      "step": 400
    },
    {
      "epoch": 20.0,
      "grad_norm": 2.1400034427642822,
      "learning_rate": 2.25e-05,
      "loss": 0.4816,
      "step": 450
    },
    {
      "epoch": 22.22222222222222,
      "grad_norm": 1.7150920629501343,
      "learning_rate": 2.5e-05,
      "loss": 0.4725,
      "step": 500
    },
    {
      "epoch": 22.22222222222222,
      "eval_loss": 0.4371508061885834,
      "eval_runtime": 8.4206,
      "eval_samples_per_second": 18.882,
      "eval_steps_per_second": 2.375,
      "step": 500
    },
    {
      "epoch": 24.444444444444443,
      "grad_norm": 1.6130620241165161,
      "learning_rate": 2.7500000000000004e-05,
      "loss": 0.4607,
      "step": 550
    },
    {
      "epoch": 26.666666666666668,
      "grad_norm": 2.737618923187256,
      "learning_rate": 3e-05,
      "loss": 0.4634,
      "step": 600
    },
    {
      "epoch": 28.88888888888889,
      "grad_norm": 3.0077686309814453,
      "learning_rate": 3.2500000000000004e-05,
      "loss": 0.4621,
      "step": 650
    },
    {
      "epoch": 31.11111111111111,
      "grad_norm": 1.4074121713638306,
      "learning_rate": 3.5e-05,
      "loss": 0.4611,
      "step": 700
    },
    {
      "epoch": 33.333333333333336,
      "grad_norm": 2.665407657623291,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.4557,
      "step": 750
    },
    {
      "epoch": 35.55555555555556,
      "grad_norm": 2.857210874557495,
      "learning_rate": 4e-05,
      "loss": 0.4493,
      "step": 800
    },
    {
      "epoch": 37.77777777777778,
      "grad_norm": 1.7210990190505981,
      "learning_rate": 4.25e-05,
      "loss": 0.4489,
      "step": 850
    },
    {
      "epoch": 40.0,
      "grad_norm": 1.4690616130828857,
      "learning_rate": 4.5e-05,
      "loss": 0.4494,
      "step": 900
    },
    {
      "epoch": 42.22222222222222,
      "grad_norm": 1.6961876153945923,
      "learning_rate": 4.75e-05,
      "loss": 0.4451,
      "step": 950
    },
    {
      "epoch": 44.44444444444444,
      "grad_norm": 1.8849211931228638,
      "learning_rate": 5e-05,
      "loss": 0.4415,
      "step": 1000
    },
    {
      "epoch": 44.44444444444444,
      "eval_loss": 0.4203202724456787,
      "eval_runtime": 7.7871,
      "eval_samples_per_second": 20.418,
      "eval_steps_per_second": 2.568,
      "step": 1000
    },
    {
      "epoch": 46.666666666666664,
      "grad_norm": 4.348108768463135,
      "learning_rate": 5.25e-05,
      "loss": 0.4444,
      "step": 1050
    },
    {
      "epoch": 48.888888888888886,
      "grad_norm": 2.2666618824005127,
      "learning_rate": 5.500000000000001e-05,
      "loss": 0.436,
      "step": 1100
    },
    {
      "epoch": 51.111111111111114,
      "grad_norm": 2.367915630340576,
      "learning_rate": 5.7499999999999995e-05,
      "loss": 0.4358,
      "step": 1150
    },
    {
      "epoch": 53.333333333333336,
      "grad_norm": 1.8919609785079956,
      "learning_rate": 6e-05,
      "loss": 0.432,
      "step": 1200
    },
    {
      "epoch": 55.55555555555556,
      "grad_norm": 2.736358642578125,
      "learning_rate": 6.25e-05,
      "loss": 0.4373,
      "step": 1250
    },
    {
      "epoch": 57.77777777777778,
      "grad_norm": 2.9132068157196045,
      "learning_rate": 6.500000000000001e-05,
      "loss": 0.4317,
      "step": 1300
    },
    {
      "epoch": 60.0,
      "grad_norm": 1.5804557800292969,
      "learning_rate": 6.750000000000001e-05,
      "loss": 0.4288,
      "step": 1350
    },
    {
      "epoch": 62.22222222222222,
      "grad_norm": 2.7998554706573486,
      "learning_rate": 7e-05,
      "loss": 0.4291,
      "step": 1400
    },
    {
      "epoch": 64.44444444444444,
      "grad_norm": 1.1768637895584106,
      "learning_rate": 7.25e-05,
      "loss": 0.4183,
      "step": 1450
    },
    {
      "epoch": 66.66666666666667,
      "grad_norm": 1.888795018196106,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.423,
      "step": 1500
    },
    {
      "epoch": 66.66666666666667,
      "eval_loss": 0.4168904721736908,
      "eval_runtime": 8.1243,
      "eval_samples_per_second": 19.571,
      "eval_steps_per_second": 2.462,
      "step": 1500
    },
    {
      "epoch": 68.88888888888889,
      "grad_norm": 1.5093848705291748,
      "learning_rate": 7.75e-05,
      "loss": 0.4214,
      "step": 1550
    },
    {
      "epoch": 71.11111111111111,
      "grad_norm": 3.0897819995880127,
      "learning_rate": 8e-05,
      "loss": 0.4217,
      "step": 1600
    },
    {
      "epoch": 73.33333333333333,
      "grad_norm": 4.435822486877441,
      "learning_rate": 8.25e-05,
      "loss": 0.4115,
      "step": 1650
    },
    {
      "epoch": 75.55555555555556,
      "grad_norm": 1.9570446014404297,
      "learning_rate": 8.5e-05,
      "loss": 0.414,
      "step": 1700
    },
    {
      "epoch": 77.77777777777777,
      "grad_norm": 2.8186404705047607,
      "learning_rate": 8.75e-05,
      "loss": 0.418,
      "step": 1750
    },
    {
      "epoch": 80.0,
      "grad_norm": 5.278728485107422,
      "learning_rate": 9e-05,
      "loss": 0.4128,
      "step": 1800
    },
    {
      "epoch": 82.22222222222223,
      "grad_norm": 2.285869598388672,
      "learning_rate": 9.250000000000001e-05,
      "loss": 0.4144,
      "step": 1850
    },
    {
      "epoch": 84.44444444444444,
      "grad_norm": 2.994182586669922,
      "learning_rate": 9.5e-05,
      "loss": 0.4105,
      "step": 1900
    },
    {
      "epoch": 86.66666666666667,
      "grad_norm": 2.9983792304992676,
      "learning_rate": 9.75e-05,
      "loss": 0.414,
      "step": 1950
    },
    {
      "epoch": 88.88888888888889,
      "grad_norm": 3.550610303878784,
      "learning_rate": 0.0001,
      "loss": 0.4108,
      "step": 2000
    },
    {
      "epoch": 88.88888888888889,
      "eval_loss": 0.4182928204536438,
      "eval_runtime": 8.5174,
      "eval_samples_per_second": 18.668,
      "eval_steps_per_second": 2.348,
      "step": 2000
    },
    {
      "epoch": 91.11111111111111,
      "grad_norm": 2.9625656604766846,
      "learning_rate": 9.75e-05,
      "loss": 0.4158,
      "step": 2050
    },
    {
      "epoch": 93.33333333333333,
      "grad_norm": 1.8065454959869385,
      "learning_rate": 9.5e-05,
      "loss": 0.4089,
      "step": 2100
    },
    {
      "epoch": 95.55555555555556,
      "grad_norm": 1.9464865922927856,
      "learning_rate": 9.250000000000001e-05,
      "loss": 0.4043,
      "step": 2150
    },
    {
      "epoch": 97.77777777777777,
      "grad_norm": 3.0861456394195557,
      "learning_rate": 9e-05,
      "loss": 0.4031,
      "step": 2200
    },
    {
      "epoch": 100.0,
      "grad_norm": 1.8622132539749146,
      "learning_rate": 8.75e-05,
      "loss": 0.4017,
      "step": 2250
    },
    {
      "epoch": 102.22222222222223,
      "grad_norm": 2.099729061126709,
      "learning_rate": 8.5e-05,
      "loss": 0.4042,
      "step": 2300
    },
    {
      "epoch": 104.44444444444444,
      "grad_norm": 3.328990936279297,
      "learning_rate": 8.25e-05,
      "loss": 0.4052,
      "step": 2350
    },
    {
      "epoch": 106.66666666666667,
      "grad_norm": 4.531957149505615,
      "learning_rate": 8e-05,
      "loss": 0.3985,
      "step": 2400
    },
    {
      "epoch": 108.88888888888889,
      "grad_norm": 1.7652114629745483,
      "learning_rate": 7.75e-05,
      "loss": 0.3909,
      "step": 2450
    },
    {
      "epoch": 111.11111111111111,
      "grad_norm": 1.0521656274795532,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.3934,
      "step": 2500
    },
    {
      "epoch": 111.11111111111111,
      "eval_loss": 0.4110700488090515,
      "eval_runtime": 8.8533,
      "eval_samples_per_second": 17.959,
      "eval_steps_per_second": 2.259,
      "step": 2500
    },
    {
      "epoch": 113.33333333333333,
      "grad_norm": 2.5307023525238037,
      "learning_rate": 7.25e-05,
      "loss": 0.3946,
      "step": 2550
    },
    {
      "epoch": 115.55555555555556,
      "grad_norm": 1.3955553770065308,
      "learning_rate": 7e-05,
      "loss": 0.3877,
      "step": 2600
    },
    {
      "epoch": 117.77777777777777,
      "grad_norm": 1.8233174085617065,
      "learning_rate": 6.750000000000001e-05,
      "loss": 0.3885,
      "step": 2650
    },
    {
      "epoch": 120.0,
      "grad_norm": 3.06978440284729,
      "learning_rate": 6.500000000000001e-05,
      "loss": 0.3946,
      "step": 2700
    },
    {
      "epoch": 122.22222222222223,
      "grad_norm": 2.5015697479248047,
      "learning_rate": 6.255e-05,
      "loss": 0.3928,
      "step": 2750
    },
    {
      "epoch": 124.44444444444444,
      "grad_norm": 1.035135269165039,
      "learning_rate": 6.005000000000001e-05,
      "loss": 0.3855,
      "step": 2800
    },
    {
      "epoch": 126.66666666666667,
      "grad_norm": 1.8948378562927246,
      "learning_rate": 5.755e-05,
      "loss": 0.3907,
      "step": 2850
    },
    {
      "epoch": 128.88888888888889,
      "grad_norm": 1.7191160917282104,
      "learning_rate": 5.505e-05,
      "loss": 0.3925,
      "step": 2900
    },
    {
      "epoch": 131.11111111111111,
      "grad_norm": 1.152048110961914,
      "learning_rate": 5.255e-05,
      "loss": 0.3872,
      "step": 2950
    },
    {
      "epoch": 133.33333333333334,
      "grad_norm": 1.2533172369003296,
      "learning_rate": 5.005e-05,
      "loss": 0.3821,
      "step": 3000
    },
    {
      "epoch": 133.33333333333334,
      "eval_loss": 0.4164334237575531,
      "eval_runtime": 9.0357,
      "eval_samples_per_second": 17.597,
      "eval_steps_per_second": 2.213,
      "step": 3000
    }
  ],
  "logging_steps": 50,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 182,
  "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": 4.120227645951542e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}