{
  "best_metric": 0.3940223157405853,
  "best_model_checkpoint": "mikhail-panzo/ceb_b64_le5_s4000/checkpoint-3500",
  "epoch": 158.41584158415841,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.9801980198019802,
      "grad_norm": 2.883972644805908,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.8004,
      "step": 50
    },
    {
      "epoch": 3.9603960396039604,
      "grad_norm": 2.5420947074890137,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.77,
      "step": 100
    },
    {
      "epoch": 5.9405940594059405,
      "grad_norm": 1.3930779695510864,
      "learning_rate": 7.5e-07,
      "loss": 0.7525,
      "step": 150
    },
    {
      "epoch": 7.920792079207921,
      "grad_norm": 1.8414520025253296,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.722,
      "step": 200
    },
    {
      "epoch": 9.900990099009901,
      "grad_norm": 2.0665407180786133,
      "learning_rate": 1.25e-06,
      "loss": 0.7076,
      "step": 250
    },
    {
      "epoch": 11.881188118811881,
      "grad_norm": 1.2595587968826294,
      "learning_rate": 1.5e-06,
      "loss": 0.6757,
      "step": 300
    },
    {
      "epoch": 13.861386138613861,
      "grad_norm": 1.4587587118148804,
      "learning_rate": 1.75e-06,
      "loss": 0.6562,
      "step": 350
    },
    {
      "epoch": 15.841584158415841,
      "grad_norm": 2.241546154022217,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.6016,
      "step": 400
    },
    {
      "epoch": 17.821782178217823,
      "grad_norm": 1.4243369102478027,
      "learning_rate": 2.25e-06,
      "loss": 0.5593,
      "step": 450
    },
    {
      "epoch": 19.801980198019802,
      "grad_norm": 1.6889928579330444,
      "learning_rate": 2.5e-06,
      "loss": 0.536,
      "step": 500
    },
    {
      "epoch": 19.801980198019802,
      "eval_loss": 0.4640245735645294,
      "eval_runtime": 6.7454,
      "eval_samples_per_second": 26.685,
      "eval_steps_per_second": 3.41,
      "step": 500
    },
    {
      "epoch": 21.782178217821784,
      "grad_norm": 1.1601781845092773,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.5198,
      "step": 550
    },
    {
      "epoch": 23.762376237623762,
      "grad_norm": 1.1969332695007324,
      "learning_rate": 3e-06,
      "loss": 0.5114,
      "step": 600
    },
    {
      "epoch": 25.742574257425744,
      "grad_norm": 1.1022064685821533,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.5092,
      "step": 650
    },
    {
      "epoch": 27.722772277227723,
      "grad_norm": 1.2418344020843506,
      "learning_rate": 3.5e-06,
      "loss": 0.4951,
      "step": 700
    },
    {
      "epoch": 29.702970297029704,
      "grad_norm": 1.1083300113677979,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.4911,
      "step": 750
    },
    {
      "epoch": 31.683168316831683,
      "grad_norm": 1.2230323553085327,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.4866,
      "step": 800
    },
    {
      "epoch": 33.663366336633665,
      "grad_norm": 1.1232423782348633,
      "learning_rate": 4.25e-06,
      "loss": 0.4876,
      "step": 850
    },
    {
      "epoch": 35.64356435643565,
      "grad_norm": 0.9751402139663696,
      "learning_rate": 4.5e-06,
      "loss": 0.4772,
      "step": 900
    },
    {
      "epoch": 37.62376237623762,
      "grad_norm": 1.0373197793960571,
      "learning_rate": 4.75e-06,
      "loss": 0.4791,
      "step": 950
    },
    {
      "epoch": 39.603960396039604,
      "grad_norm": 1.0164889097213745,
      "learning_rate": 5e-06,
      "loss": 0.4674,
      "step": 1000
    },
    {
      "epoch": 39.603960396039604,
      "eval_loss": 0.4247289001941681,
      "eval_runtime": 6.5527,
      "eval_samples_per_second": 27.47,
      "eval_steps_per_second": 3.51,
      "step": 1000
    },
    {
      "epoch": 41.584158415841586,
      "grad_norm": 1.087668538093567,
      "learning_rate": 5.2500000000000006e-06,
      "loss": 0.4754,
      "step": 1050
    },
    {
      "epoch": 43.56435643564357,
      "grad_norm": 0.891859769821167,
      "learning_rate": 5.500000000000001e-06,
      "loss": 0.4677,
      "step": 1100
    },
    {
      "epoch": 45.54455445544554,
      "grad_norm": 1.018160343170166,
      "learning_rate": 5.75e-06,
      "loss": 0.4642,
      "step": 1150
    },
    {
      "epoch": 47.524752475247524,
      "grad_norm": 1.4765816926956177,
      "learning_rate": 6e-06,
      "loss": 0.4602,
      "step": 1200
    },
    {
      "epoch": 49.504950495049506,
      "grad_norm": 1.3534387350082397,
      "learning_rate": 6.25e-06,
      "loss": 0.4634,
      "step": 1250
    },
    {
      "epoch": 51.48514851485149,
      "grad_norm": 1.031537413597107,
      "learning_rate": 6.5000000000000004e-06,
      "loss": 0.4567,
      "step": 1300
    },
    {
      "epoch": 53.46534653465346,
      "grad_norm": 1.0172531604766846,
      "learning_rate": 6.750000000000001e-06,
      "loss": 0.4592,
      "step": 1350
    },
    {
      "epoch": 55.445544554455445,
      "grad_norm": 0.8680336475372314,
      "learning_rate": 7e-06,
      "loss": 0.4583,
      "step": 1400
    },
    {
      "epoch": 57.42574257425743,
      "grad_norm": 1.2397912740707397,
      "learning_rate": 7.25e-06,
      "loss": 0.4459,
      "step": 1450
    },
    {
      "epoch": 59.40594059405941,
      "grad_norm": 1.021185040473938,
      "learning_rate": 7.500000000000001e-06,
      "loss": 0.4479,
      "step": 1500
    },
    {
      "epoch": 59.40594059405941,
      "eval_loss": 0.4102235734462738,
      "eval_runtime": 7.6974,
      "eval_samples_per_second": 23.384,
      "eval_steps_per_second": 2.988,
      "step": 1500
    },
    {
      "epoch": 61.386138613861384,
      "grad_norm": 1.005591630935669,
      "learning_rate": 7.75e-06,
      "loss": 0.4469,
      "step": 1550
    },
    {
      "epoch": 63.366336633663366,
      "grad_norm": 0.8193312287330627,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.4425,
      "step": 1600
    },
    {
      "epoch": 65.34653465346534,
      "grad_norm": 0.9143276810646057,
      "learning_rate": 8.25e-06,
      "loss": 0.4424,
      "step": 1650
    },
    {
      "epoch": 67.32673267326733,
      "grad_norm": 1.0127445459365845,
      "learning_rate": 8.5e-06,
      "loss": 0.4446,
      "step": 1700
    },
    {
      "epoch": 69.3069306930693,
      "grad_norm": 1.045551061630249,
      "learning_rate": 8.750000000000001e-06,
      "loss": 0.4417,
      "step": 1750
    },
    {
      "epoch": 71.2871287128713,
      "grad_norm": 0.9629473090171814,
      "learning_rate": 9e-06,
      "loss": 0.4362,
      "step": 1800
    },
    {
      "epoch": 73.26732673267327,
      "grad_norm": 0.9655708074569702,
      "learning_rate": 9.250000000000001e-06,
      "loss": 0.4393,
      "step": 1850
    },
    {
      "epoch": 75.24752475247524,
      "grad_norm": 1.167992115020752,
      "learning_rate": 9.5e-06,
      "loss": 0.4373,
      "step": 1900
    },
    {
      "epoch": 77.22772277227723,
      "grad_norm": 0.9866817593574524,
      "learning_rate": 9.75e-06,
      "loss": 0.4369,
      "step": 1950
    },
    {
      "epoch": 79.20792079207921,
      "grad_norm": 0.8923723101615906,
      "learning_rate": 1e-05,
      "loss": 0.4356,
      "step": 2000
    },
    {
      "epoch": 79.20792079207921,
      "eval_loss": 0.4031977355480194,
      "eval_runtime": 7.0669,
      "eval_samples_per_second": 25.471,
      "eval_steps_per_second": 3.255,
      "step": 2000
    },
    {
      "epoch": 81.18811881188118,
      "grad_norm": 1.0076066255569458,
      "learning_rate": 9.755e-06,
      "loss": 0.4365,
      "step": 2050
    },
    {
      "epoch": 83.16831683168317,
      "grad_norm": 0.8251845836639404,
      "learning_rate": 9.505000000000001e-06,
      "loss": 0.4342,
      "step": 2100
    },
    {
      "epoch": 85.14851485148515,
      "grad_norm": 1.029945731163025,
      "learning_rate": 9.255e-06,
      "loss": 0.4325,
      "step": 2150
    },
    {
      "epoch": 87.12871287128714,
      "grad_norm": 1.0486702919006348,
      "learning_rate": 9.005000000000001e-06,
      "loss": 0.4223,
      "step": 2200
    },
    {
      "epoch": 89.10891089108911,
      "grad_norm": 0.9815123677253723,
      "learning_rate": 8.755e-06,
      "loss": 0.427,
      "step": 2250
    },
    {
      "epoch": 91.08910891089108,
      "grad_norm": 1.0457566976547241,
      "learning_rate": 8.505e-06,
      "loss": 0.4297,
      "step": 2300
    },
    {
      "epoch": 93.06930693069307,
      "grad_norm": 1.0108842849731445,
      "learning_rate": 8.255000000000001e-06,
      "loss": 0.4262,
      "step": 2350
    },
    {
      "epoch": 95.04950495049505,
      "grad_norm": 1.0584673881530762,
      "learning_rate": 8.005e-06,
      "loss": 0.4238,
      "step": 2400
    },
    {
      "epoch": 97.02970297029702,
      "grad_norm": 1.346966028213501,
      "learning_rate": 7.755000000000001e-06,
      "loss": 0.4217,
      "step": 2450
    },
    {
      "epoch": 99.00990099009901,
      "grad_norm": 0.9860787987709045,
      "learning_rate": 7.505e-06,
      "loss": 0.4277,
      "step": 2500
    },
    {
      "epoch": 99.00990099009901,
      "eval_loss": 0.39623910188674927,
      "eval_runtime": 7.3733,
      "eval_samples_per_second": 24.412,
      "eval_steps_per_second": 3.119,
      "step": 2500
    },
    {
      "epoch": 100.99009900990099,
      "grad_norm": 1.5156948566436768,
      "learning_rate": 7.255000000000001e-06,
      "loss": 0.4234,
      "step": 2550
    },
    {
      "epoch": 102.97029702970298,
      "grad_norm": 0.7337579727172852,
      "learning_rate": 7.005000000000001e-06,
      "loss": 0.4243,
      "step": 2600
    },
    {
      "epoch": 104.95049504950495,
      "grad_norm": 1.220702052116394,
      "learning_rate": 6.7550000000000005e-06,
      "loss": 0.4169,
      "step": 2650
    },
    {
      "epoch": 106.93069306930693,
      "grad_norm": 1.0253492593765259,
      "learning_rate": 6.505e-06,
      "loss": 0.4226,
      "step": 2700
    },
    {
      "epoch": 108.91089108910892,
      "grad_norm": 0.848473846912384,
      "learning_rate": 6.255e-06,
      "loss": 0.4213,
      "step": 2750
    },
    {
      "epoch": 110.89108910891089,
      "grad_norm": 0.7380446791648865,
      "learning_rate": 6.005000000000001e-06,
      "loss": 0.4207,
      "step": 2800
    },
    {
      "epoch": 112.87128712871286,
      "grad_norm": 0.9291226863861084,
      "learning_rate": 5.755000000000001e-06,
      "loss": 0.4238,
      "step": 2850
    },
    {
      "epoch": 114.85148514851485,
      "grad_norm": 0.8098958730697632,
      "learning_rate": 5.505000000000001e-06,
      "loss": 0.4185,
      "step": 2900
    },
    {
      "epoch": 116.83168316831683,
      "grad_norm": 0.8403974771499634,
      "learning_rate": 5.2550000000000005e-06,
      "loss": 0.4187,
      "step": 2950
    },
    {
      "epoch": 118.81188118811882,
      "grad_norm": 1.092827558517456,
      "learning_rate": 5.0049999999999995e-06,
      "loss": 0.418,
      "step": 3000
    },
    {
      "epoch": 118.81188118811882,
      "eval_loss": 0.39552977681159973,
      "eval_runtime": 7.0687,
      "eval_samples_per_second": 25.464,
      "eval_steps_per_second": 3.254,
      "step": 3000
    },
    {
      "epoch": 120.79207920792079,
      "grad_norm": 0.8844769597053528,
      "learning_rate": 4.755e-06,
      "loss": 0.4144,
      "step": 3050
    },
    {
      "epoch": 122.77227722772277,
      "grad_norm": 0.7425467371940613,
      "learning_rate": 4.505e-06,
      "loss": 0.416,
      "step": 3100
    },
    {
      "epoch": 124.75247524752476,
      "grad_norm": 0.9394475817680359,
      "learning_rate": 4.255e-06,
      "loss": 0.4214,
      "step": 3150
    },
    {
      "epoch": 126.73267326732673,
      "grad_norm": 0.8024426698684692,
      "learning_rate": 4.005000000000001e-06,
      "loss": 0.4207,
      "step": 3200
    },
    {
      "epoch": 128.7128712871287,
      "grad_norm": 1.2388241291046143,
      "learning_rate": 3.7550000000000005e-06,
      "loss": 0.4156,
      "step": 3250
    },
    {
      "epoch": 130.69306930693068,
      "grad_norm": 1.0179219245910645,
      "learning_rate": 3.505e-06,
      "loss": 0.4152,
      "step": 3300
    },
    {
      "epoch": 132.67326732673268,
      "grad_norm": 0.9490112066268921,
      "learning_rate": 3.255e-06,
      "loss": 0.4152,
      "step": 3350
    },
    {
      "epoch": 134.65346534653466,
      "grad_norm": 0.8918274641036987,
      "learning_rate": 3.005e-06,
      "loss": 0.4115,
      "step": 3400
    },
    {
      "epoch": 136.63366336633663,
      "grad_norm": 0.859157145023346,
      "learning_rate": 2.7550000000000003e-06,
      "loss": 0.4148,
      "step": 3450
    },
    {
      "epoch": 138.6138613861386,
      "grad_norm": 0.7952170968055725,
      "learning_rate": 2.505e-06,
      "loss": 0.4162,
      "step": 3500
    },
    {
      "epoch": 138.6138613861386,
      "eval_loss": 0.3940223157405853,
      "eval_runtime": 6.5916,
      "eval_samples_per_second": 27.308,
      "eval_steps_per_second": 3.489,
      "step": 3500
    },
    {
      "epoch": 140.59405940594058,
      "grad_norm": 0.865683376789093,
      "learning_rate": 2.2550000000000004e-06,
      "loss": 0.4181,
      "step": 3550
    },
    {
      "epoch": 142.5742574257426,
      "grad_norm": 0.7344664931297302,
      "learning_rate": 2.0050000000000003e-06,
      "loss": 0.413,
      "step": 3600
    },
    {
      "epoch": 144.55445544554456,
      "grad_norm": 1.005863904953003,
      "learning_rate": 1.7550000000000001e-06,
      "loss": 0.4115,
      "step": 3650
    },
    {
      "epoch": 146.53465346534654,
      "grad_norm": 0.7481868267059326,
      "learning_rate": 1.505e-06,
      "loss": 0.4098,
      "step": 3700
    },
    {
      "epoch": 148.5148514851485,
      "grad_norm": 0.8728845119476318,
      "learning_rate": 1.255e-06,
      "loss": 0.4108,
      "step": 3750
    },
    {
      "epoch": 150.4950495049505,
      "grad_norm": 1.0313975811004639,
      "learning_rate": 1.0050000000000001e-06,
      "loss": 0.4114,
      "step": 3800
    },
    {
      "epoch": 152.47524752475246,
      "grad_norm": 0.8399845361709595,
      "learning_rate": 7.550000000000001e-07,
      "loss": 0.4062,
      "step": 3850
    },
    {
      "epoch": 154.45544554455446,
      "grad_norm": 0.8254792094230652,
      "learning_rate": 5.05e-07,
      "loss": 0.4126,
      "step": 3900
    },
    {
      "epoch": 156.43564356435644,
      "grad_norm": 0.8597183227539062,
      "learning_rate": 2.55e-07,
      "loss": 0.4174,
      "step": 3950
    },
    {
      "epoch": 158.41584158415841,
      "grad_norm": 0.8257197737693787,
      "learning_rate": 5e-09,
      "loss": 0.4082,
      "step": 4000
    },
    {
      "epoch": 158.41584158415841,
      "eval_loss": 0.3941545784473419,
      "eval_runtime": 6.5183,
      "eval_samples_per_second": 27.615,
      "eval_steps_per_second": 3.529,
      "step": 4000
    }
  ],
  "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": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4.322114324975938e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}