{
  "best_metric": 0.35058465600013733,
  "best_model_checkpoint": "mikhail_panzo/zlm_b64_le4_s12000/checkpoint-4000",
  "epoch": 3.3500837520938025,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04187604690117253,
      "grad_norm": 3.0184223651885986,
      "learning_rate": 2.35e-06,
      "loss": 1.1206,
      "step": 50
    },
    {
      "epoch": 0.08375209380234507,
      "grad_norm": 3.4442005157470703,
      "learning_rate": 4.85e-06,
      "loss": 0.8185,
      "step": 100
    },
    {
      "epoch": 0.12562814070351758,
      "grad_norm": 2.6812493801116943,
      "learning_rate": 7.35e-06,
      "loss": 0.7576,
      "step": 150
    },
    {
      "epoch": 0.16750418760469013,
      "grad_norm": 4.518482685089111,
      "learning_rate": 9.85e-06,
      "loss": 0.6938,
      "step": 200
    },
    {
      "epoch": 0.20938023450586266,
      "grad_norm": 4.683146953582764,
      "learning_rate": 1.235e-05,
      "loss": 0.6386,
      "step": 250
    },
    {
      "epoch": 0.25125628140703515,
      "grad_norm": 2.7392578125,
      "learning_rate": 1.485e-05,
      "loss": 0.6226,
      "step": 300
    },
    {
      "epoch": 0.2931323283082077,
      "grad_norm": 2.017056465148926,
      "learning_rate": 1.7349999999999998e-05,
      "loss": 0.6128,
      "step": 350
    },
    {
      "epoch": 0.33500837520938026,
      "grad_norm": 3.6669230461120605,
      "learning_rate": 1.985e-05,
      "loss": 0.5675,
      "step": 400
    },
    {
      "epoch": 0.3768844221105528,
      "grad_norm": 4.16896915435791,
      "learning_rate": 2.235e-05,
      "loss": 0.5655,
      "step": 450
    },
    {
      "epoch": 0.4187604690117253,
      "grad_norm": 3.691459894180298,
      "learning_rate": 2.485e-05,
      "loss": 0.5487,
      "step": 500
    },
    {
      "epoch": 0.4187604690117253,
      "eval_loss": 0.47458764910697937,
      "eval_runtime": 214.7196,
      "eval_samples_per_second": 39.535,
      "eval_steps_per_second": 4.946,
      "step": 500
    },
    {
      "epoch": 0.46063651591289784,
      "grad_norm": 3.3792996406555176,
      "learning_rate": 2.7350000000000004e-05,
      "loss": 0.5232,
      "step": 550
    },
    {
      "epoch": 0.5025125628140703,
      "grad_norm": 3.6263363361358643,
      "learning_rate": 2.985e-05,
      "loss": 0.5132,
      "step": 600
    },
    {
      "epoch": 0.5443886097152428,
      "grad_norm": 2.652538776397705,
      "learning_rate": 3.235e-05,
      "loss": 0.5067,
      "step": 650
    },
    {
      "epoch": 0.5862646566164154,
      "grad_norm": 3.5355172157287598,
      "learning_rate": 3.485e-05,
      "loss": 0.4981,
      "step": 700
    },
    {
      "epoch": 0.628140703517588,
      "grad_norm": 2.5035619735717773,
      "learning_rate": 3.735e-05,
      "loss": 0.5094,
      "step": 750
    },
    {
      "epoch": 0.6700167504187605,
      "grad_norm": 2.9585518836975098,
      "learning_rate": 3.9850000000000006e-05,
      "loss": 0.5053,
      "step": 800
    },
    {
      "epoch": 0.711892797319933,
      "grad_norm": 9.997116088867188,
      "learning_rate": 4.235e-05,
      "loss": 0.502,
      "step": 850
    },
    {
      "epoch": 0.7537688442211056,
      "grad_norm": 1.8630377054214478,
      "learning_rate": 4.4850000000000006e-05,
      "loss": 0.483,
      "step": 900
    },
    {
      "epoch": 0.7956448911222781,
      "grad_norm": 1.5534099340438843,
      "learning_rate": 4.735e-05,
      "loss": 0.4743,
      "step": 950
    },
    {
      "epoch": 0.8375209380234506,
      "grad_norm": 2.8155503273010254,
      "learning_rate": 4.9850000000000006e-05,
      "loss": 0.483,
      "step": 1000
    },
    {
      "epoch": 0.8375209380234506,
      "eval_loss": 0.4226582944393158,
      "eval_runtime": 213.2339,
      "eval_samples_per_second": 39.811,
      "eval_steps_per_second": 4.98,
      "step": 1000
    },
    {
      "epoch": 0.8793969849246231,
      "grad_norm": 4.2074055671691895,
      "learning_rate": 5.235e-05,
      "loss": 0.4625,
      "step": 1050
    },
    {
      "epoch": 0.9212730318257957,
      "grad_norm": 2.215416431427002,
      "learning_rate": 5.485e-05,
      "loss": 0.461,
      "step": 1100
    },
    {
      "epoch": 0.9631490787269682,
      "grad_norm": 3.3321831226348877,
      "learning_rate": 5.7350000000000005e-05,
      "loss": 0.4807,
      "step": 1150
    },
    {
      "epoch": 1.0050251256281406,
      "grad_norm": 3.5796420574188232,
      "learning_rate": 5.9850000000000005e-05,
      "loss": 0.4456,
      "step": 1200
    },
    {
      "epoch": 1.0469011725293131,
      "grad_norm": 2.3565518856048584,
      "learning_rate": 6.235000000000001e-05,
      "loss": 0.4664,
      "step": 1250
    },
    {
      "epoch": 1.0887772194304857,
      "grad_norm": 3.147794485092163,
      "learning_rate": 6.485e-05,
      "loss": 0.4681,
      "step": 1300
    },
    {
      "epoch": 1.1306532663316582,
      "grad_norm": 3.028409957885742,
      "learning_rate": 6.735e-05,
      "loss": 0.4663,
      "step": 1350
    },
    {
      "epoch": 1.1725293132328307,
      "grad_norm": 4.820915222167969,
      "learning_rate": 6.985e-05,
      "loss": 0.4493,
      "step": 1400
    },
    {
      "epoch": 1.2144053601340032,
      "grad_norm": 1.93254816532135,
      "learning_rate": 7.235000000000001e-05,
      "loss": 0.4413,
      "step": 1450
    },
    {
      "epoch": 1.2562814070351758,
      "grad_norm": 2.237015724182129,
      "learning_rate": 7.485e-05,
      "loss": 0.432,
      "step": 1500
    },
    {
      "epoch": 1.2562814070351758,
      "eval_loss": 0.39829039573669434,
      "eval_runtime": 217.3782,
      "eval_samples_per_second": 39.052,
      "eval_steps_per_second": 4.885,
      "step": 1500
    },
    {
      "epoch": 1.2981574539363483,
      "grad_norm": 1.709120273590088,
      "learning_rate": 7.735e-05,
      "loss": 0.4361,
      "step": 1550
    },
    {
      "epoch": 1.3400335008375208,
      "grad_norm": 3.733328104019165,
      "learning_rate": 7.985e-05,
      "loss": 0.4357,
      "step": 1600
    },
    {
      "epoch": 1.3819095477386933,
      "grad_norm": 4.131324768066406,
      "learning_rate": 8.235000000000001e-05,
      "loss": 0.4445,
      "step": 1650
    },
    {
      "epoch": 1.4237855946398659,
      "grad_norm": 2.2893927097320557,
      "learning_rate": 8.485e-05,
      "loss": 0.4379,
      "step": 1700
    },
    {
      "epoch": 1.4656616415410384,
      "grad_norm": 2.1464734077453613,
      "learning_rate": 8.735000000000001e-05,
      "loss": 0.4393,
      "step": 1750
    },
    {
      "epoch": 1.507537688442211,
      "grad_norm": 2.01698637008667,
      "learning_rate": 8.985e-05,
      "loss": 0.4341,
      "step": 1800
    },
    {
      "epoch": 1.5494137353433834,
      "grad_norm": 2.259115219116211,
      "learning_rate": 9.235000000000001e-05,
      "loss": 0.4324,
      "step": 1850
    },
    {
      "epoch": 1.591289782244556,
      "grad_norm": 2.0072858333587646,
      "learning_rate": 9.485e-05,
      "loss": 0.4327,
      "step": 1900
    },
    {
      "epoch": 1.6331658291457285,
      "grad_norm": 2.060303211212158,
      "learning_rate": 9.735000000000001e-05,
      "loss": 0.4279,
      "step": 1950
    },
    {
      "epoch": 1.675041876046901,
      "grad_norm": 1.2846741676330566,
      "learning_rate": 9.985000000000001e-05,
      "loss": 0.429,
      "step": 2000
    },
    {
      "epoch": 1.675041876046901,
      "eval_loss": 0.3953319191932678,
      "eval_runtime": 217.8108,
      "eval_samples_per_second": 38.974,
      "eval_steps_per_second": 4.876,
      "step": 2000
    },
    {
      "epoch": 1.7169179229480735,
      "grad_norm": 2.253840923309326,
      "learning_rate": 9.953e-05,
      "loss": 0.4264,
      "step": 2050
    },
    {
      "epoch": 1.758793969849246,
      "grad_norm": 1.3114299774169922,
      "learning_rate": 9.903e-05,
      "loss": 0.4355,
      "step": 2100
    },
    {
      "epoch": 1.8006700167504186,
      "grad_norm": 1.8178610801696777,
      "learning_rate": 9.853e-05,
      "loss": 0.4256,
      "step": 2150
    },
    {
      "epoch": 1.8425460636515911,
      "grad_norm": 2.1193735599517822,
      "learning_rate": 9.803e-05,
      "loss": 0.4305,
      "step": 2200
    },
    {
      "epoch": 1.8844221105527639,
      "grad_norm": 1.6871263980865479,
      "learning_rate": 9.753e-05,
      "loss": 0.4106,
      "step": 2250
    },
    {
      "epoch": 1.9262981574539364,
      "grad_norm": 1.2699921131134033,
      "learning_rate": 9.703000000000001e-05,
      "loss": 0.4156,
      "step": 2300
    },
    {
      "epoch": 1.968174204355109,
      "grad_norm": 1.7966769933700562,
      "learning_rate": 9.653000000000001e-05,
      "loss": 0.4257,
      "step": 2350
    },
    {
      "epoch": 2.0100502512562812,
      "grad_norm": 1.7838683128356934,
      "learning_rate": 9.603000000000001e-05,
      "loss": 0.4156,
      "step": 2400
    },
    {
      "epoch": 2.0519262981574538,
      "grad_norm": 1.4161527156829834,
      "learning_rate": 9.553e-05,
      "loss": 0.4096,
      "step": 2450
    },
    {
      "epoch": 2.0938023450586263,
      "grad_norm": 1.5864896774291992,
      "learning_rate": 9.503e-05,
      "loss": 0.4168,
      "step": 2500
    },
    {
      "epoch": 2.0938023450586263,
      "eval_loss": 0.3701198697090149,
      "eval_runtime": 220.4399,
      "eval_samples_per_second": 38.509,
      "eval_steps_per_second": 4.818,
      "step": 2500
    },
    {
      "epoch": 2.135678391959799,
      "grad_norm": 1.2460856437683105,
      "learning_rate": 9.453e-05,
      "loss": 0.415,
      "step": 2550
    },
    {
      "epoch": 2.1775544388609713,
      "grad_norm": 1.994315266609192,
      "learning_rate": 9.403e-05,
      "loss": 0.4075,
      "step": 2600
    },
    {
      "epoch": 2.219430485762144,
      "grad_norm": 2.148324966430664,
      "learning_rate": 9.353000000000001e-05,
      "loss": 0.4077,
      "step": 2650
    },
    {
      "epoch": 2.2613065326633164,
      "grad_norm": 1.7242035865783691,
      "learning_rate": 9.303000000000001e-05,
      "loss": 0.4044,
      "step": 2700
    },
    {
      "epoch": 2.3031825795644894,
      "grad_norm": 1.912110447883606,
      "learning_rate": 9.253000000000001e-05,
      "loss": 0.4075,
      "step": 2750
    },
    {
      "epoch": 2.3450586264656614,
      "grad_norm": 1.3661189079284668,
      "learning_rate": 9.203000000000001e-05,
      "loss": 0.4041,
      "step": 2800
    },
    {
      "epoch": 2.3869346733668344,
      "grad_norm": 1.7083531618118286,
      "learning_rate": 9.153e-05,
      "loss": 0.396,
      "step": 2850
    },
    {
      "epoch": 2.4288107202680065,
      "grad_norm": 1.7829670906066895,
      "learning_rate": 9.103e-05,
      "loss": 0.4042,
      "step": 2900
    },
    {
      "epoch": 2.4706867671691795,
      "grad_norm": 2.188753366470337,
      "learning_rate": 9.053e-05,
      "loss": 0.3963,
      "step": 2950
    },
    {
      "epoch": 2.5125628140703515,
      "grad_norm": 1.2608036994934082,
      "learning_rate": 9.003e-05,
      "loss": 0.4021,
      "step": 3000
    },
    {
      "epoch": 2.5125628140703515,
      "eval_loss": 0.3613213300704956,
      "eval_runtime": 214.7226,
      "eval_samples_per_second": 39.535,
      "eval_steps_per_second": 4.946,
      "step": 3000
    },
    {
      "epoch": 2.5544388609715245,
      "grad_norm": 1.501607060432434,
      "learning_rate": 8.953e-05,
      "loss": 0.3897,
      "step": 3050
    },
    {
      "epoch": 2.5963149078726966,
      "grad_norm": 1.321588397026062,
      "learning_rate": 8.903e-05,
      "loss": 0.3948,
      "step": 3100
    },
    {
      "epoch": 2.6381909547738696,
      "grad_norm": 1.5581566095352173,
      "learning_rate": 8.853000000000001e-05,
      "loss": 0.4005,
      "step": 3150
    },
    {
      "epoch": 2.6800670016750416,
      "grad_norm": 1.3824856281280518,
      "learning_rate": 8.803e-05,
      "loss": 0.3981,
      "step": 3200
    },
    {
      "epoch": 2.7219430485762146,
      "grad_norm": 1.7730051279067993,
      "learning_rate": 8.753e-05,
      "loss": 0.3982,
      "step": 3250
    },
    {
      "epoch": 2.7638190954773867,
      "grad_norm": 1.516622543334961,
      "learning_rate": 8.703e-05,
      "loss": 0.3906,
      "step": 3300
    },
    {
      "epoch": 2.8056951423785597,
      "grad_norm": 1.2814849615097046,
      "learning_rate": 8.653e-05,
      "loss": 0.3859,
      "step": 3350
    },
    {
      "epoch": 2.8475711892797317,
      "grad_norm": 1.239619255065918,
      "learning_rate": 8.603e-05,
      "loss": 0.3942,
      "step": 3400
    },
    {
      "epoch": 2.8894472361809047,
      "grad_norm": 2.24064040184021,
      "learning_rate": 8.553e-05,
      "loss": 0.3938,
      "step": 3450
    },
    {
      "epoch": 2.931323283082077,
      "grad_norm": 1.1459012031555176,
      "learning_rate": 8.503e-05,
      "loss": 0.3925,
      "step": 3500
    },
    {
      "epoch": 2.931323283082077,
      "eval_loss": 0.35087427496910095,
      "eval_runtime": 217.1018,
      "eval_samples_per_second": 39.101,
      "eval_steps_per_second": 4.892,
      "step": 3500
    },
    {
      "epoch": 2.9731993299832498,
      "grad_norm": 1.8844960927963257,
      "learning_rate": 8.453e-05,
      "loss": 0.3803,
      "step": 3550
    },
    {
      "epoch": 3.0150753768844223,
      "grad_norm": 1.5738004446029663,
      "learning_rate": 8.403000000000001e-05,
      "loss": 0.3895,
      "step": 3600
    },
    {
      "epoch": 3.056951423785595,
      "grad_norm": 1.19874906539917,
      "learning_rate": 8.353000000000001e-05,
      "loss": 0.3902,
      "step": 3650
    },
    {
      "epoch": 3.0988274706867673,
      "grad_norm": 2.4810609817504883,
      "learning_rate": 8.303000000000001e-05,
      "loss": 0.3863,
      "step": 3700
    },
    {
      "epoch": 3.14070351758794,
      "grad_norm": 1.4734585285186768,
      "learning_rate": 8.253000000000001e-05,
      "loss": 0.391,
      "step": 3750
    },
    {
      "epoch": 3.1825795644891124,
      "grad_norm": 1.6449118852615356,
      "learning_rate": 8.203000000000001e-05,
      "loss": 0.3888,
      "step": 3800
    },
    {
      "epoch": 3.224455611390285,
      "grad_norm": 2.5607736110687256,
      "learning_rate": 8.153e-05,
      "loss": 0.3893,
      "step": 3850
    },
    {
      "epoch": 3.2663316582914574,
      "grad_norm": 1.0670469999313354,
      "learning_rate": 8.103e-05,
      "loss": 0.3853,
      "step": 3900
    },
    {
      "epoch": 3.30820770519263,
      "grad_norm": 2.191486120223999,
      "learning_rate": 8.053e-05,
      "loss": 0.3832,
      "step": 3950
    },
    {
      "epoch": 3.3500837520938025,
      "grad_norm": 1.0259532928466797,
      "learning_rate": 8.003e-05,
      "loss": 0.3839,
      "step": 4000
    },
    {
      "epoch": 3.3500837520938025,
      "eval_loss": 0.35058465600013733,
      "eval_runtime": 214.7925,
      "eval_samples_per_second": 39.522,
      "eval_steps_per_second": 4.944,
      "step": 4000
    }
  ],
  "logging_steps": 50,
  "max_steps": 12000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 11,
  "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": 3.88890953465351e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}