{
  "best_metric": 0.39415255188941956,
  "best_model_checkpoint": "mikhail-panzo/zlm_b64_le5_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": 4.788218975067139,
      "learning_rate": 2.4500000000000004e-07,
      "loss": 1.1819,
      "step": 50
    },
    {
      "epoch": 0.08375209380234507,
      "grad_norm": 15.618935585021973,
      "learning_rate": 4.95e-07,
      "loss": 1.0042,
      "step": 100
    },
    {
      "epoch": 0.12562814070351758,
      "grad_norm": 2.7821555137634277,
      "learning_rate": 7.4e-07,
      "loss": 0.9676,
      "step": 150
    },
    {
      "epoch": 0.16750418760469013,
      "grad_norm": 3.759263038635254,
      "learning_rate": 9.85e-07,
      "loss": 0.9362,
      "step": 200
    },
    {
      "epoch": 0.20938023450586266,
      "grad_norm": 3.906439781188965,
      "learning_rate": 1.235e-06,
      "loss": 0.8667,
      "step": 250
    },
    {
      "epoch": 0.25125628140703515,
      "grad_norm": 4.397643566131592,
      "learning_rate": 1.485e-06,
      "loss": 0.8455,
      "step": 300
    },
    {
      "epoch": 0.2931323283082077,
      "grad_norm": 2.94077467918396,
      "learning_rate": 1.7350000000000001e-06,
      "loss": 0.8376,
      "step": 350
    },
    {
      "epoch": 0.33500837520938026,
      "grad_norm": 4.161074161529541,
      "learning_rate": 1.985e-06,
      "loss": 0.7521,
      "step": 400
    },
    {
      "epoch": 0.3768844221105528,
      "grad_norm": 2.8088552951812744,
      "learning_rate": 2.235e-06,
      "loss": 0.7444,
      "step": 450
    },
    {
      "epoch": 0.4187604690117253,
      "grad_norm": 2.62283992767334,
      "learning_rate": 2.4850000000000003e-06,
      "loss": 0.7129,
      "step": 500
    },
    {
      "epoch": 0.4187604690117253,
      "eval_loss": 0.6066474914550781,
      "eval_runtime": 213.4911,
      "eval_samples_per_second": 39.763,
      "eval_steps_per_second": 4.974,
      "step": 500
    },
    {
      "epoch": 0.46063651591289784,
      "grad_norm": 3.654665946960449,
      "learning_rate": 2.7350000000000005e-06,
      "loss": 0.6772,
      "step": 550
    },
    {
      "epoch": 0.5025125628140703,
      "grad_norm": 5.1851887702941895,
      "learning_rate": 2.9850000000000002e-06,
      "loss": 0.6552,
      "step": 600
    },
    {
      "epoch": 0.5443886097152428,
      "grad_norm": 4.017187595367432,
      "learning_rate": 3.2350000000000004e-06,
      "loss": 0.6506,
      "step": 650
    },
    {
      "epoch": 0.5862646566164154,
      "grad_norm": 2.1161563396453857,
      "learning_rate": 3.485e-06,
      "loss": 0.6301,
      "step": 700
    },
    {
      "epoch": 0.628140703517588,
      "grad_norm": 1.857313871383667,
      "learning_rate": 3.7350000000000002e-06,
      "loss": 0.6373,
      "step": 750
    },
    {
      "epoch": 0.6700167504187605,
      "grad_norm": 3.6841330528259277,
      "learning_rate": 3.985000000000001e-06,
      "loss": 0.6305,
      "step": 800
    },
    {
      "epoch": 0.711892797319933,
      "grad_norm": 2.6450912952423096,
      "learning_rate": 4.235e-06,
      "loss": 0.6169,
      "step": 850
    },
    {
      "epoch": 0.7537688442211056,
      "grad_norm": 1.7241184711456299,
      "learning_rate": 4.485e-06,
      "loss": 0.5933,
      "step": 900
    },
    {
      "epoch": 0.7956448911222781,
      "grad_norm": 2.602735996246338,
      "learning_rate": 4.735e-06,
      "loss": 0.583,
      "step": 950
    },
    {
      "epoch": 0.8375209380234506,
      "grad_norm": 12.419028282165527,
      "learning_rate": 4.9850000000000006e-06,
      "loss": 0.5922,
      "step": 1000
    },
    {
      "epoch": 0.8375209380234506,
      "eval_loss": 0.4979843497276306,
      "eval_runtime": 214.0502,
      "eval_samples_per_second": 39.659,
      "eval_steps_per_second": 4.961,
      "step": 1000
    },
    {
      "epoch": 0.8793969849246231,
      "grad_norm": 3.200178861618042,
      "learning_rate": 5.235e-06,
      "loss": 0.5678,
      "step": 1050
    },
    {
      "epoch": 0.9212730318257957,
      "grad_norm": 2.458869218826294,
      "learning_rate": 5.485e-06,
      "loss": 0.5558,
      "step": 1100
    },
    {
      "epoch": 0.9631490787269682,
      "grad_norm": 2.341752290725708,
      "learning_rate": 5.735e-06,
      "loss": 0.5666,
      "step": 1150
    },
    {
      "epoch": 1.0050251256281406,
      "grad_norm": 2.9191603660583496,
      "learning_rate": 5.985000000000001e-06,
      "loss": 0.5345,
      "step": 1200
    },
    {
      "epoch": 1.0469011725293131,
      "grad_norm": 2.68806529045105,
      "learning_rate": 6.235000000000001e-06,
      "loss": 0.5653,
      "step": 1250
    },
    {
      "epoch": 1.0887772194304857,
      "grad_norm": 3.278820276260376,
      "learning_rate": 6.485000000000001e-06,
      "loss": 0.5565,
      "step": 1300
    },
    {
      "epoch": 1.1306532663316582,
      "grad_norm": 3.066511869430542,
      "learning_rate": 6.735000000000001e-06,
      "loss": 0.5644,
      "step": 1350
    },
    {
      "epoch": 1.1725293132328307,
      "grad_norm": 4.269862651824951,
      "learning_rate": 6.985000000000001e-06,
      "loss": 0.5356,
      "step": 1400
    },
    {
      "epoch": 1.2144053601340032,
      "grad_norm": 3.6681156158447266,
      "learning_rate": 7.235000000000001e-06,
      "loss": 0.5246,
      "step": 1450
    },
    {
      "epoch": 1.2562814070351758,
      "grad_norm": 3.1227505207061768,
      "learning_rate": 7.485000000000001e-06,
      "loss": 0.5168,
      "step": 1500
    },
    {
      "epoch": 1.2562814070351758,
      "eval_loss": 0.45785218477249146,
      "eval_runtime": 211.6749,
      "eval_samples_per_second": 40.104,
      "eval_steps_per_second": 5.017,
      "step": 1500
    },
    {
      "epoch": 1.2981574539363483,
      "grad_norm": 2.6239662170410156,
      "learning_rate": 7.735e-06,
      "loss": 0.5153,
      "step": 1550
    },
    {
      "epoch": 1.3400335008375208,
      "grad_norm": 2.881901979446411,
      "learning_rate": 7.985e-06,
      "loss": 0.518,
      "step": 1600
    },
    {
      "epoch": 1.3819095477386933,
      "grad_norm": 2.8052310943603516,
      "learning_rate": 8.235e-06,
      "loss": 0.5272,
      "step": 1650
    },
    {
      "epoch": 1.4237855946398659,
      "grad_norm": 2.6334052085876465,
      "learning_rate": 8.485000000000001e-06,
      "loss": 0.5188,
      "step": 1700
    },
    {
      "epoch": 1.4656616415410384,
      "grad_norm": 3.038274049758911,
      "learning_rate": 8.735000000000002e-06,
      "loss": 0.5139,
      "step": 1750
    },
    {
      "epoch": 1.507537688442211,
      "grad_norm": 3.2438852787017822,
      "learning_rate": 8.985000000000001e-06,
      "loss": 0.5092,
      "step": 1800
    },
    {
      "epoch": 1.5494137353433834,
      "grad_norm": 4.168883800506592,
      "learning_rate": 9.235e-06,
      "loss": 0.5103,
      "step": 1850
    },
    {
      "epoch": 1.591289782244556,
      "grad_norm": 2.6399967670440674,
      "learning_rate": 9.485000000000002e-06,
      "loss": 0.5025,
      "step": 1900
    },
    {
      "epoch": 1.6331658291457285,
      "grad_norm": 1.6828950643539429,
      "learning_rate": 9.735e-06,
      "loss": 0.4963,
      "step": 1950
    },
    {
      "epoch": 1.675041876046901,
      "grad_norm": 1.8955535888671875,
      "learning_rate": 9.985000000000002e-06,
      "loss": 0.4954,
      "step": 2000
    },
    {
      "epoch": 1.675041876046901,
      "eval_loss": 0.43827974796295166,
      "eval_runtime": 212.1455,
      "eval_samples_per_second": 40.015,
      "eval_steps_per_second": 5.006,
      "step": 2000
    },
    {
      "epoch": 1.7169179229480735,
      "grad_norm": 3.9634511470794678,
      "learning_rate": 9.953000000000001e-06,
      "loss": 0.4891,
      "step": 2050
    },
    {
      "epoch": 1.758793969849246,
      "grad_norm": 3.5446126461029053,
      "learning_rate": 9.903e-06,
      "loss": 0.5141,
      "step": 2100
    },
    {
      "epoch": 1.8006700167504186,
      "grad_norm": 3.1621015071868896,
      "learning_rate": 9.853e-06,
      "loss": 0.4941,
      "step": 2150
    },
    {
      "epoch": 1.8425460636515911,
      "grad_norm": 2.6622440814971924,
      "learning_rate": 9.803e-06,
      "loss": 0.4979,
      "step": 2200
    },
    {
      "epoch": 1.8844221105527639,
      "grad_norm": 1.7774969339370728,
      "learning_rate": 9.753e-06,
      "loss": 0.4768,
      "step": 2250
    },
    {
      "epoch": 1.9262981574539364,
      "grad_norm": 2.1141610145568848,
      "learning_rate": 9.703000000000002e-06,
      "loss": 0.4834,
      "step": 2300
    },
    {
      "epoch": 1.968174204355109,
      "grad_norm": 3.4576170444488525,
      "learning_rate": 9.653e-06,
      "loss": 0.493,
      "step": 2350
    },
    {
      "epoch": 2.0100502512562812,
      "grad_norm": 2.4238383769989014,
      "learning_rate": 9.603000000000001e-06,
      "loss": 0.4836,
      "step": 2400
    },
    {
      "epoch": 2.0519262981574538,
      "grad_norm": 2.2526774406433105,
      "learning_rate": 9.553000000000002e-06,
      "loss": 0.4832,
      "step": 2450
    },
    {
      "epoch": 2.0938023450586263,
      "grad_norm": 2.8719518184661865,
      "learning_rate": 9.503e-06,
      "loss": 0.4836,
      "step": 2500
    },
    {
      "epoch": 2.0938023450586263,
      "eval_loss": 0.4208170473575592,
      "eval_runtime": 214.437,
      "eval_samples_per_second": 39.587,
      "eval_steps_per_second": 4.953,
      "step": 2500
    },
    {
      "epoch": 2.135678391959799,
      "grad_norm": 3.2671754360198975,
      "learning_rate": 9.453e-06,
      "loss": 0.4775,
      "step": 2550
    },
    {
      "epoch": 2.1775544388609713,
      "grad_norm": 2.454883337020874,
      "learning_rate": 9.403000000000001e-06,
      "loss": 0.4777,
      "step": 2600
    },
    {
      "epoch": 2.219430485762144,
      "grad_norm": 2.901097297668457,
      "learning_rate": 9.353000000000002e-06,
      "loss": 0.4757,
      "step": 2650
    },
    {
      "epoch": 2.2613065326633164,
      "grad_norm": 2.980057954788208,
      "learning_rate": 9.303e-06,
      "loss": 0.4644,
      "step": 2700
    },
    {
      "epoch": 2.3031825795644894,
      "grad_norm": 1.7162193059921265,
      "learning_rate": 9.253000000000001e-06,
      "loss": 0.4714,
      "step": 2750
    },
    {
      "epoch": 2.3450586264656614,
      "grad_norm": 2.912874937057495,
      "learning_rate": 9.203000000000002e-06,
      "loss": 0.4596,
      "step": 2800
    },
    {
      "epoch": 2.3869346733668344,
      "grad_norm": 1.7781060934066772,
      "learning_rate": 9.153e-06,
      "loss": 0.4591,
      "step": 2850
    },
    {
      "epoch": 2.4288107202680065,
      "grad_norm": 5.2557902336120605,
      "learning_rate": 9.103e-06,
      "loss": 0.4652,
      "step": 2900
    },
    {
      "epoch": 2.4706867671691795,
      "grad_norm": 1.7230316400527954,
      "learning_rate": 9.053000000000001e-06,
      "loss": 0.46,
      "step": 2950
    },
    {
      "epoch": 2.5125628140703515,
      "grad_norm": 4.071449279785156,
      "learning_rate": 9.003e-06,
      "loss": 0.4623,
      "step": 3000
    },
    {
      "epoch": 2.5125628140703515,
      "eval_loss": 0.4098353981971741,
      "eval_runtime": 214.5603,
      "eval_samples_per_second": 39.565,
      "eval_steps_per_second": 4.95,
      "step": 3000
    },
    {
      "epoch": 2.5544388609715245,
      "grad_norm": 3.5458014011383057,
      "learning_rate": 8.953e-06,
      "loss": 0.4488,
      "step": 3050
    },
    {
      "epoch": 2.5963149078726966,
      "grad_norm": 1.726624846458435,
      "learning_rate": 8.903000000000001e-06,
      "loss": 0.4536,
      "step": 3100
    },
    {
      "epoch": 2.6381909547738696,
      "grad_norm": 3.204106092453003,
      "learning_rate": 8.853e-06,
      "loss": 0.4654,
      "step": 3150
    },
    {
      "epoch": 2.6800670016750416,
      "grad_norm": 2.245302677154541,
      "learning_rate": 8.803e-06,
      "loss": 0.462,
      "step": 3200
    },
    {
      "epoch": 2.7219430485762146,
      "grad_norm": 2.0797626972198486,
      "learning_rate": 8.753e-06,
      "loss": 0.4599,
      "step": 3250
    },
    {
      "epoch": 2.7638190954773867,
      "grad_norm": 1.9437323808670044,
      "learning_rate": 8.703e-06,
      "loss": 0.4473,
      "step": 3300
    },
    {
      "epoch": 2.8056951423785597,
      "grad_norm": 2.5299806594848633,
      "learning_rate": 8.653e-06,
      "loss": 0.4442,
      "step": 3350
    },
    {
      "epoch": 2.8475711892797317,
      "grad_norm": 3.208460807800293,
      "learning_rate": 8.603e-06,
      "loss": 0.4553,
      "step": 3400
    },
    {
      "epoch": 2.8894472361809047,
      "grad_norm": 2.703047037124634,
      "learning_rate": 8.553000000000001e-06,
      "loss": 0.4532,
      "step": 3450
    },
    {
      "epoch": 2.931323283082077,
      "grad_norm": 1.9532470703125,
      "learning_rate": 8.503e-06,
      "loss": 0.4499,
      "step": 3500
    },
    {
      "epoch": 2.931323283082077,
      "eval_loss": 0.4003259837627411,
      "eval_runtime": 213.8518,
      "eval_samples_per_second": 39.696,
      "eval_steps_per_second": 4.966,
      "step": 3500
    },
    {
      "epoch": 2.9731993299832498,
      "grad_norm": 1.9028701782226562,
      "learning_rate": 8.453000000000002e-06,
      "loss": 0.4391,
      "step": 3550
    },
    {
      "epoch": 3.0150753768844223,
      "grad_norm": 10.080265045166016,
      "learning_rate": 8.403e-06,
      "loss": 0.4462,
      "step": 3600
    },
    {
      "epoch": 3.056951423785595,
      "grad_norm": 2.996328115463257,
      "learning_rate": 8.353000000000001e-06,
      "loss": 0.45,
      "step": 3650
    },
    {
      "epoch": 3.0988274706867673,
      "grad_norm": 2.221226930618286,
      "learning_rate": 8.303000000000002e-06,
      "loss": 0.4469,
      "step": 3700
    },
    {
      "epoch": 3.14070351758794,
      "grad_norm": 2.961524248123169,
      "learning_rate": 8.253e-06,
      "loss": 0.4535,
      "step": 3750
    },
    {
      "epoch": 3.1825795644891124,
      "grad_norm": 2.487060785293579,
      "learning_rate": 8.203000000000001e-06,
      "loss": 0.4444,
      "step": 3800
    },
    {
      "epoch": 3.224455611390285,
      "grad_norm": 1.5449366569519043,
      "learning_rate": 8.153000000000001e-06,
      "loss": 0.449,
      "step": 3850
    },
    {
      "epoch": 3.2663316582914574,
      "grad_norm": 1.6193325519561768,
      "learning_rate": 8.103e-06,
      "loss": 0.4423,
      "step": 3900
    },
    {
      "epoch": 3.30820770519263,
      "grad_norm": 2.442594528198242,
      "learning_rate": 8.053e-06,
      "loss": 0.4456,
      "step": 3950
    },
    {
      "epoch": 3.3500837520938025,
      "grad_norm": 1.759829044342041,
      "learning_rate": 8.003000000000001e-06,
      "loss": 0.4421,
      "step": 4000
    },
    {
      "epoch": 3.3500837520938025,
      "eval_loss": 0.39415255188941956,
      "eval_runtime": 212.9896,
      "eval_samples_per_second": 39.856,
      "eval_steps_per_second": 4.986,
      "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": 3.895851451089946e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}