{
  "best_metric": 0.33282962441444397,
  "best_model_checkpoint": "mikhail_panzo/zlm_b128_le4_s8000/checkpoint-4000",
  "epoch": 6.701570680628272,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08376963350785341,
      "grad_norm": 2.9895308017730713,
      "learning_rate": 2.4500000000000003e-06,
      "loss": 1.0423,
      "step": 50
    },
    {
      "epoch": 0.16753926701570682,
      "grad_norm": 3.051593542098999,
      "learning_rate": 4.950000000000001e-06,
      "loss": 0.8473,
      "step": 100
    },
    {
      "epoch": 0.2513089005235602,
      "grad_norm": 2.0044381618499756,
      "learning_rate": 7.45e-06,
      "loss": 0.733,
      "step": 150
    },
    {
      "epoch": 0.33507853403141363,
      "grad_norm": 3.4974701404571533,
      "learning_rate": 9.950000000000001e-06,
      "loss": 0.6511,
      "step": 200
    },
    {
      "epoch": 0.418848167539267,
      "grad_norm": 1.854073405265808,
      "learning_rate": 1.2450000000000001e-05,
      "loss": 0.6143,
      "step": 250
    },
    {
      "epoch": 0.5026178010471204,
      "grad_norm": 1.737787127494812,
      "learning_rate": 1.4950000000000001e-05,
      "loss": 0.5909,
      "step": 300
    },
    {
      "epoch": 0.5863874345549738,
      "grad_norm": 2.0971367359161377,
      "learning_rate": 1.745e-05,
      "loss": 0.5684,
      "step": 350
    },
    {
      "epoch": 0.6701570680628273,
      "grad_norm": 1.8380221128463745,
      "learning_rate": 1.995e-05,
      "loss": 0.5472,
      "step": 400
    },
    {
      "epoch": 0.7539267015706806,
      "grad_norm": 3.9271857738494873,
      "learning_rate": 2.245e-05,
      "loss": 0.5287,
      "step": 450
    },
    {
      "epoch": 0.837696335078534,
      "grad_norm": 7.809891700744629,
      "learning_rate": 2.495e-05,
      "loss": 0.5174,
      "step": 500
    },
    {
      "epoch": 0.837696335078534,
      "eval_loss": 0.4793977439403534,
      "eval_runtime": 265.0789,
      "eval_samples_per_second": 32.024,
      "eval_steps_per_second": 4.006,
      "step": 500
    },
    {
      "epoch": 0.9214659685863874,
      "grad_norm": 2.2309463024139404,
      "learning_rate": 2.7450000000000003e-05,
      "loss": 0.5084,
      "step": 550
    },
    {
      "epoch": 1.0052356020942408,
      "grad_norm": 1.8079086542129517,
      "learning_rate": 2.995e-05,
      "loss": 0.4954,
      "step": 600
    },
    {
      "epoch": 1.0890052356020943,
      "grad_norm": 6.239879608154297,
      "learning_rate": 3.245e-05,
      "loss": 0.4954,
      "step": 650
    },
    {
      "epoch": 1.1727748691099475,
      "grad_norm": 12.593622207641602,
      "learning_rate": 3.495e-05,
      "loss": 0.4946,
      "step": 700
    },
    {
      "epoch": 1.256544502617801,
      "grad_norm": 3.1568186283111572,
      "learning_rate": 3.745e-05,
      "loss": 0.4768,
      "step": 750
    },
    {
      "epoch": 1.3403141361256545,
      "grad_norm": 3.9486255645751953,
      "learning_rate": 3.995e-05,
      "loss": 0.4737,
      "step": 800
    },
    {
      "epoch": 1.4240837696335078,
      "grad_norm": 2.641502618789673,
      "learning_rate": 4.245e-05,
      "loss": 0.4706,
      "step": 850
    },
    {
      "epoch": 1.5078534031413613,
      "grad_norm": 1.9798855781555176,
      "learning_rate": 4.495e-05,
      "loss": 0.4605,
      "step": 900
    },
    {
      "epoch": 1.5916230366492146,
      "grad_norm": 2.691363573074341,
      "learning_rate": 4.745e-05,
      "loss": 0.4554,
      "step": 950
    },
    {
      "epoch": 1.675392670157068,
      "grad_norm": 3.704902410507202,
      "learning_rate": 4.995e-05,
      "loss": 0.4561,
      "step": 1000
    },
    {
      "epoch": 1.675392670157068,
      "eval_loss": 0.402164101600647,
      "eval_runtime": 265.618,
      "eval_samples_per_second": 31.959,
      "eval_steps_per_second": 3.998,
      "step": 1000
    },
    {
      "epoch": 1.7591623036649215,
      "grad_norm": 4.97900390625,
      "learning_rate": 5.245e-05,
      "loss": 0.4553,
      "step": 1050
    },
    {
      "epoch": 1.8429319371727748,
      "grad_norm": 1.9889676570892334,
      "learning_rate": 5.495e-05,
      "loss": 0.449,
      "step": 1100
    },
    {
      "epoch": 1.9267015706806283,
      "grad_norm": 1.5135546922683716,
      "learning_rate": 5.745e-05,
      "loss": 0.4353,
      "step": 1150
    },
    {
      "epoch": 2.0104712041884816,
      "grad_norm": 7.610673904418945,
      "learning_rate": 5.995000000000001e-05,
      "loss": 0.4311,
      "step": 1200
    },
    {
      "epoch": 2.094240837696335,
      "grad_norm": 2.049562454223633,
      "learning_rate": 6.245000000000001e-05,
      "loss": 0.4312,
      "step": 1250
    },
    {
      "epoch": 2.1780104712041886,
      "grad_norm": 1.4102027416229248,
      "learning_rate": 6.494999999999999e-05,
      "loss": 0.4282,
      "step": 1300
    },
    {
      "epoch": 2.261780104712042,
      "grad_norm": 1.701119065284729,
      "learning_rate": 6.745e-05,
      "loss": 0.4272,
      "step": 1350
    },
    {
      "epoch": 2.345549738219895,
      "grad_norm": 2.0149667263031006,
      "learning_rate": 6.995e-05,
      "loss": 0.4277,
      "step": 1400
    },
    {
      "epoch": 2.4293193717277486,
      "grad_norm": 2.1658883094787598,
      "learning_rate": 7.245000000000001e-05,
      "loss": 0.4247,
      "step": 1450
    },
    {
      "epoch": 2.513089005235602,
      "grad_norm": 2.6821463108062744,
      "learning_rate": 7.495e-05,
      "loss": 0.4169,
      "step": 1500
    },
    {
      "epoch": 2.513089005235602,
      "eval_loss": 0.3853827118873596,
      "eval_runtime": 269.2607,
      "eval_samples_per_second": 31.527,
      "eval_steps_per_second": 3.944,
      "step": 1500
    },
    {
      "epoch": 2.5968586387434556,
      "grad_norm": 1.4221985340118408,
      "learning_rate": 7.745e-05,
      "loss": 0.4145,
      "step": 1550
    },
    {
      "epoch": 2.680628272251309,
      "grad_norm": 3.651655435562134,
      "learning_rate": 7.995e-05,
      "loss": 0.4129,
      "step": 1600
    },
    {
      "epoch": 2.7643979057591626,
      "grad_norm": 1.2700576782226562,
      "learning_rate": 8.245e-05,
      "loss": 0.4106,
      "step": 1650
    },
    {
      "epoch": 2.8481675392670156,
      "grad_norm": 1.8690059185028076,
      "learning_rate": 8.495e-05,
      "loss": 0.4134,
      "step": 1700
    },
    {
      "epoch": 2.931937172774869,
      "grad_norm": 2.483203649520874,
      "learning_rate": 8.745000000000001e-05,
      "loss": 0.413,
      "step": 1750
    },
    {
      "epoch": 3.0157068062827226,
      "grad_norm": 1.738501787185669,
      "learning_rate": 8.995e-05,
      "loss": 0.4063,
      "step": 1800
    },
    {
      "epoch": 3.099476439790576,
      "grad_norm": 1.837342619895935,
      "learning_rate": 9.245e-05,
      "loss": 0.4049,
      "step": 1850
    },
    {
      "epoch": 3.183246073298429,
      "grad_norm": 1.284974455833435,
      "learning_rate": 9.495e-05,
      "loss": 0.4065,
      "step": 1900
    },
    {
      "epoch": 3.2670157068062826,
      "grad_norm": 2.637281656265259,
      "learning_rate": 9.745000000000001e-05,
      "loss": 0.3986,
      "step": 1950
    },
    {
      "epoch": 3.350785340314136,
      "grad_norm": 1.4775941371917725,
      "learning_rate": 9.995e-05,
      "loss": 0.4054,
      "step": 2000
    },
    {
      "epoch": 3.350785340314136,
      "eval_loss": 0.36490994691848755,
      "eval_runtime": 265.9165,
      "eval_samples_per_second": 31.924,
      "eval_steps_per_second": 3.994,
      "step": 2000
    },
    {
      "epoch": 3.4345549738219896,
      "grad_norm": 2.121384859085083,
      "learning_rate": 9.918333333333334e-05,
      "loss": 0.4058,
      "step": 2050
    },
    {
      "epoch": 3.518324607329843,
      "grad_norm": 1.645984411239624,
      "learning_rate": 9.835e-05,
      "loss": 0.4021,
      "step": 2100
    },
    {
      "epoch": 3.6020942408376966,
      "grad_norm": 1.246239185333252,
      "learning_rate": 9.751666666666666e-05,
      "loss": 0.3991,
      "step": 2150
    },
    {
      "epoch": 3.6858638743455496,
      "grad_norm": 1.9096795320510864,
      "learning_rate": 9.668333333333334e-05,
      "loss": 0.3961,
      "step": 2200
    },
    {
      "epoch": 3.769633507853403,
      "grad_norm": 1.8867601156234741,
      "learning_rate": 9.585000000000001e-05,
      "loss": 0.3904,
      "step": 2250
    },
    {
      "epoch": 3.8534031413612566,
      "grad_norm": 1.7438101768493652,
      "learning_rate": 9.501666666666668e-05,
      "loss": 0.3895,
      "step": 2300
    },
    {
      "epoch": 3.93717277486911,
      "grad_norm": 1.1799490451812744,
      "learning_rate": 9.418333333333334e-05,
      "loss": 0.4027,
      "step": 2350
    },
    {
      "epoch": 4.020942408376963,
      "grad_norm": 1.1952763795852661,
      "learning_rate": 9.335e-05,
      "loss": 0.3893,
      "step": 2400
    },
    {
      "epoch": 4.104712041884817,
      "grad_norm": 2.008756160736084,
      "learning_rate": 9.251666666666667e-05,
      "loss": 0.3878,
      "step": 2450
    },
    {
      "epoch": 4.18848167539267,
      "grad_norm": 2.2693591117858887,
      "learning_rate": 9.168333333333333e-05,
      "loss": 0.3863,
      "step": 2500
    },
    {
      "epoch": 4.18848167539267,
      "eval_loss": 0.3528364896774292,
      "eval_runtime": 272.7627,
      "eval_samples_per_second": 31.122,
      "eval_steps_per_second": 3.893,
      "step": 2500
    },
    {
      "epoch": 4.272251308900524,
      "grad_norm": 1.0409568548202515,
      "learning_rate": 9.086666666666666e-05,
      "loss": 0.3894,
      "step": 2550
    },
    {
      "epoch": 4.356020942408377,
      "grad_norm": 2.047908306121826,
      "learning_rate": 9.003333333333333e-05,
      "loss": 0.3862,
      "step": 2600
    },
    {
      "epoch": 4.439790575916231,
      "grad_norm": 1.767168641090393,
      "learning_rate": 8.92e-05,
      "loss": 0.3832,
      "step": 2650
    },
    {
      "epoch": 4.523560209424084,
      "grad_norm": 1.1918233633041382,
      "learning_rate": 8.836666666666667e-05,
      "loss": 0.3839,
      "step": 2700
    },
    {
      "epoch": 4.607329842931938,
      "grad_norm": 0.9553185105323792,
      "learning_rate": 8.753333333333334e-05,
      "loss": 0.3807,
      "step": 2750
    },
    {
      "epoch": 4.69109947643979,
      "grad_norm": 1.1633609533309937,
      "learning_rate": 8.67e-05,
      "loss": 0.3836,
      "step": 2800
    },
    {
      "epoch": 4.774869109947644,
      "grad_norm": 1.1633822917938232,
      "learning_rate": 8.586666666666668e-05,
      "loss": 0.3815,
      "step": 2850
    },
    {
      "epoch": 4.858638743455497,
      "grad_norm": 1.7152239084243774,
      "learning_rate": 8.503333333333334e-05,
      "loss": 0.3809,
      "step": 2900
    },
    {
      "epoch": 4.942408376963351,
      "grad_norm": 1.2379584312438965,
      "learning_rate": 8.42e-05,
      "loss": 0.3801,
      "step": 2950
    },
    {
      "epoch": 5.026178010471204,
      "grad_norm": 1.1686451435089111,
      "learning_rate": 8.336666666666667e-05,
      "loss": 0.3807,
      "step": 3000
    },
    {
      "epoch": 5.026178010471204,
      "eval_loss": 0.3480445444583893,
      "eval_runtime": 272.7064,
      "eval_samples_per_second": 31.129,
      "eval_steps_per_second": 3.894,
      "step": 3000
    },
    {
      "epoch": 5.109947643979058,
      "grad_norm": 1.5680384635925293,
      "learning_rate": 8.253333333333334e-05,
      "loss": 0.3739,
      "step": 3050
    },
    {
      "epoch": 5.193717277486911,
      "grad_norm": 1.5168578624725342,
      "learning_rate": 8.17e-05,
      "loss": 0.3756,
      "step": 3100
    },
    {
      "epoch": 5.277486910994765,
      "grad_norm": 1.0925545692443848,
      "learning_rate": 8.086666666666666e-05,
      "loss": 0.3739,
      "step": 3150
    },
    {
      "epoch": 5.361256544502618,
      "grad_norm": 0.8022084832191467,
      "learning_rate": 8.003333333333333e-05,
      "loss": 0.3736,
      "step": 3200
    },
    {
      "epoch": 5.445026178010472,
      "grad_norm": 1.6964820623397827,
      "learning_rate": 7.920000000000001e-05,
      "loss": 0.379,
      "step": 3250
    },
    {
      "epoch": 5.528795811518324,
      "grad_norm": 1.0717741250991821,
      "learning_rate": 7.836666666666667e-05,
      "loss": 0.3762,
      "step": 3300
    },
    {
      "epoch": 5.612565445026178,
      "grad_norm": 1.2660809755325317,
      "learning_rate": 7.753333333333334e-05,
      "loss": 0.3708,
      "step": 3350
    },
    {
      "epoch": 5.696335078534031,
      "grad_norm": 1.6065802574157715,
      "learning_rate": 7.670000000000001e-05,
      "loss": 0.3717,
      "step": 3400
    },
    {
      "epoch": 5.780104712041885,
      "grad_norm": 1.2089744806289673,
      "learning_rate": 7.586666666666668e-05,
      "loss": 0.3715,
      "step": 3450
    },
    {
      "epoch": 5.863874345549738,
      "grad_norm": 3.1524300575256348,
      "learning_rate": 7.503333333333333e-05,
      "loss": 0.3708,
      "step": 3500
    },
    {
      "epoch": 5.863874345549738,
      "eval_loss": 0.3431868255138397,
      "eval_runtime": 272.9384,
      "eval_samples_per_second": 31.102,
      "eval_steps_per_second": 3.891,
      "step": 3500
    },
    {
      "epoch": 5.947643979057592,
      "grad_norm": 1.914294719696045,
      "learning_rate": 7.42e-05,
      "loss": 0.3673,
      "step": 3550
    },
    {
      "epoch": 6.031413612565445,
      "grad_norm": 1.3186005353927612,
      "learning_rate": 7.336666666666667e-05,
      "loss": 0.3665,
      "step": 3600
    },
    {
      "epoch": 6.115183246073299,
      "grad_norm": 2.019273519515991,
      "learning_rate": 7.253333333333334e-05,
      "loss": 0.3697,
      "step": 3650
    },
    {
      "epoch": 6.198952879581152,
      "grad_norm": 1.0517597198486328,
      "learning_rate": 7.17e-05,
      "loss": 0.3674,
      "step": 3700
    },
    {
      "epoch": 6.282722513089006,
      "grad_norm": 1.0202686786651611,
      "learning_rate": 7.086666666666666e-05,
      "loss": 0.3706,
      "step": 3750
    },
    {
      "epoch": 6.366492146596858,
      "grad_norm": 1.4179818630218506,
      "learning_rate": 7.003333333333335e-05,
      "loss": 0.3681,
      "step": 3800
    },
    {
      "epoch": 6.450261780104712,
      "grad_norm": 1.3820505142211914,
      "learning_rate": 6.92e-05,
      "loss": 0.3671,
      "step": 3850
    },
    {
      "epoch": 6.534031413612565,
      "grad_norm": 1.3857202529907227,
      "learning_rate": 6.836666666666667e-05,
      "loss": 0.3641,
      "step": 3900
    },
    {
      "epoch": 6.617801047120419,
      "grad_norm": 1.0996108055114746,
      "learning_rate": 6.753333333333334e-05,
      "loss": 0.3685,
      "step": 3950
    },
    {
      "epoch": 6.701570680628272,
      "grad_norm": 0.9405946731567383,
      "learning_rate": 6.670000000000001e-05,
      "loss": 0.366,
      "step": 4000
    },
    {
      "epoch": 6.701570680628272,
      "eval_loss": 0.33282962441444397,
      "eval_runtime": 271.6112,
      "eval_samples_per_second": 31.254,
      "eval_steps_per_second": 3.91,
      "step": 4000
    }
  ],
  "logging_steps": 50,
  "max_steps": 8000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 14,
  "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": 7.164514367145274e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}