{
  "best_metric": NaN,
  "best_model_checkpoint": "miner_id_24/checkpoint-25",
  "epoch": 0.05900140128328048,
  "eval_steps": 25,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0011800280256656095,
      "grad_norm": NaN,
      "learning_rate": 5e-05,
      "loss": 332.6675,
      "step": 1
    },
    {
      "epoch": 0.0011800280256656095,
      "eval_loss": NaN,
      "eval_runtime": 2.7317,
      "eval_samples_per_second": 18.303,
      "eval_steps_per_second": 4.759,
      "step": 1
    },
    {
      "epoch": 0.002360056051331219,
      "grad_norm": NaN,
      "learning_rate": 0.0001,
      "loss": 0.0,
      "step": 2
    },
    {
      "epoch": 0.003540084076996829,
      "grad_norm": NaN,
      "learning_rate": 9.990365154573717e-05,
      "loss": 0.0,
      "step": 3
    },
    {
      "epoch": 0.004720112102662438,
      "grad_norm": NaN,
      "learning_rate": 9.961501876182148e-05,
      "loss": 0.0,
      "step": 4
    },
    {
      "epoch": 0.005900140128328048,
      "grad_norm": NaN,
      "learning_rate": 9.913533761814537e-05,
      "loss": 0.0,
      "step": 5
    },
    {
      "epoch": 0.007080168153993658,
      "grad_norm": NaN,
      "learning_rate": 9.846666218300807e-05,
      "loss": 0.0,
      "step": 6
    },
    {
      "epoch": 0.008260196179659268,
      "grad_norm": NaN,
      "learning_rate": 9.761185582727977e-05,
      "loss": 0.0,
      "step": 7
    },
    {
      "epoch": 0.009440224205324876,
      "grad_norm": NaN,
      "learning_rate": 9.657457896300791e-05,
      "loss": 0.0,
      "step": 8
    },
    {
      "epoch": 0.010620252230990486,
      "grad_norm": NaN,
      "learning_rate": 9.535927336897098e-05,
      "loss": 0.0,
      "step": 9
    },
    {
      "epoch": 0.011800280256656095,
      "grad_norm": NaN,
      "learning_rate": 9.397114317029975e-05,
      "loss": 0.0,
      "step": 10
    },
    {
      "epoch": 0.012980308282321705,
      "grad_norm": NaN,
      "learning_rate": 9.241613255361455e-05,
      "loss": 0.0,
      "step": 11
    },
    {
      "epoch": 0.014160336307987315,
      "grad_norm": NaN,
      "learning_rate": 9.070090031310558e-05,
      "loss": 0.0,
      "step": 12
    },
    {
      "epoch": 0.015340364333652923,
      "grad_norm": NaN,
      "learning_rate": 8.883279133655399e-05,
      "loss": 0.0,
      "step": 13
    },
    {
      "epoch": 0.016520392359318535,
      "grad_norm": NaN,
      "learning_rate": 8.681980515339464e-05,
      "loss": 0.0,
      "step": 14
    },
    {
      "epoch": 0.017700420384984143,
      "grad_norm": NaN,
      "learning_rate": 8.467056167950311e-05,
      "loss": 0.0,
      "step": 15
    },
    {
      "epoch": 0.01888044841064975,
      "grad_norm": NaN,
      "learning_rate": 8.239426430539243e-05,
      "loss": 0.0,
      "step": 16
    },
    {
      "epoch": 0.020060476436315363,
      "grad_norm": NaN,
      "learning_rate": 8.000066048588211e-05,
      "loss": 0.0,
      "step": 17
    },
    {
      "epoch": 0.02124050446198097,
      "grad_norm": NaN,
      "learning_rate": 7.75e-05,
      "loss": 0.0,
      "step": 18
    },
    {
      "epoch": 0.022420532487646583,
      "grad_norm": NaN,
      "learning_rate": 7.490299105985507e-05,
      "loss": 0.0,
      "step": 19
    },
    {
      "epoch": 0.02360056051331219,
      "grad_norm": NaN,
      "learning_rate": 7.222075445642904e-05,
      "loss": 0.0,
      "step": 20
    },
    {
      "epoch": 0.0247805885389778,
      "grad_norm": NaN,
      "learning_rate": 6.946477593864228e-05,
      "loss": 0.0,
      "step": 21
    },
    {
      "epoch": 0.02596061656464341,
      "grad_norm": NaN,
      "learning_rate": 6.664685702961344e-05,
      "loss": 0.0,
      "step": 22
    },
    {
      "epoch": 0.02714064459030902,
      "grad_norm": NaN,
      "learning_rate": 6.377906449072578e-05,
      "loss": 0.0,
      "step": 23
    },
    {
      "epoch": 0.02832067261597463,
      "grad_norm": NaN,
      "learning_rate": 6.087367864990233e-05,
      "loss": 0.0,
      "step": 24
    },
    {
      "epoch": 0.02950070064164024,
      "grad_norm": NaN,
      "learning_rate": 5.794314081535644e-05,
      "loss": 0.0,
      "step": 25
    },
    {
      "epoch": 0.02950070064164024,
      "eval_loss": NaN,
      "eval_runtime": 1.1857,
      "eval_samples_per_second": 42.17,
      "eval_steps_per_second": 10.964,
      "step": 25
    },
    {
      "epoch": 0.030680728667305847,
      "grad_norm": NaN,
      "learning_rate": 5.500000000000001e-05,
      "loss": 0.0,
      "step": 26
    },
    {
      "epoch": 0.031860756692971455,
      "grad_norm": NaN,
      "learning_rate": 5.205685918464356e-05,
      "loss": 0.0,
      "step": 27
    },
    {
      "epoch": 0.03304078471863707,
      "grad_norm": NaN,
      "learning_rate": 4.912632135009769e-05,
      "loss": 0.0,
      "step": 28
    },
    {
      "epoch": 0.03422081274430268,
      "grad_norm": NaN,
      "learning_rate": 4.6220935509274235e-05,
      "loss": 0.0,
      "step": 29
    },
    {
      "epoch": 0.035400840769968286,
      "grad_norm": NaN,
      "learning_rate": 4.3353142970386564e-05,
      "loss": 0.0,
      "step": 30
    },
    {
      "epoch": 0.036580868795633895,
      "grad_norm": NaN,
      "learning_rate": 4.053522406135775e-05,
      "loss": 0.0,
      "step": 31
    },
    {
      "epoch": 0.0377608968212995,
      "grad_norm": NaN,
      "learning_rate": 3.777924554357096e-05,
      "loss": 0.0,
      "step": 32
    },
    {
      "epoch": 0.03894092484696512,
      "grad_norm": NaN,
      "learning_rate": 3.509700894014496e-05,
      "loss": 0.0,
      "step": 33
    },
    {
      "epoch": 0.040120952872630726,
      "grad_norm": NaN,
      "learning_rate": 3.250000000000001e-05,
      "loss": 0.0,
      "step": 34
    },
    {
      "epoch": 0.041300980898296334,
      "grad_norm": NaN,
      "learning_rate": 2.9999339514117912e-05,
      "loss": 0.0,
      "step": 35
    },
    {
      "epoch": 0.04248100892396194,
      "grad_norm": NaN,
      "learning_rate": 2.760573569460757e-05,
      "loss": 0.0,
      "step": 36
    },
    {
      "epoch": 0.04366103694962755,
      "grad_norm": NaN,
      "learning_rate": 2.53294383204969e-05,
      "loss": 0.0,
      "step": 37
    },
    {
      "epoch": 0.044841064975293166,
      "grad_norm": NaN,
      "learning_rate": 2.3180194846605367e-05,
      "loss": 0.0,
      "step": 38
    },
    {
      "epoch": 0.046021093000958774,
      "grad_norm": NaN,
      "learning_rate": 2.1167208663446025e-05,
      "loss": 0.0,
      "step": 39
    },
    {
      "epoch": 0.04720112102662438,
      "grad_norm": NaN,
      "learning_rate": 1.9299099686894423e-05,
      "loss": 0.0,
      "step": 40
    },
    {
      "epoch": 0.04838114905228999,
      "grad_norm": NaN,
      "learning_rate": 1.758386744638546e-05,
      "loss": 0.0,
      "step": 41
    },
    {
      "epoch": 0.0495611770779556,
      "grad_norm": NaN,
      "learning_rate": 1.602885682970026e-05,
      "loss": 0.0,
      "step": 42
    },
    {
      "epoch": 0.05074120510362121,
      "grad_norm": NaN,
      "learning_rate": 1.464072663102903e-05,
      "loss": 0.0,
      "step": 43
    },
    {
      "epoch": 0.05192123312928682,
      "grad_norm": NaN,
      "learning_rate": 1.3425421036992098e-05,
      "loss": 0.0,
      "step": 44
    },
    {
      "epoch": 0.05310126115495243,
      "grad_norm": NaN,
      "learning_rate": 1.2388144172720251e-05,
      "loss": 0.0,
      "step": 45
    },
    {
      "epoch": 0.05428128918061804,
      "grad_norm": NaN,
      "learning_rate": 1.1533337816991932e-05,
      "loss": 0.0,
      "step": 46
    },
    {
      "epoch": 0.055461317206283646,
      "grad_norm": NaN,
      "learning_rate": 1.0864662381854632e-05,
      "loss": 0.0,
      "step": 47
    },
    {
      "epoch": 0.05664134523194926,
      "grad_norm": NaN,
      "learning_rate": 1.0384981238178534e-05,
      "loss": 0.0,
      "step": 48
    },
    {
      "epoch": 0.05782137325761487,
      "grad_norm": NaN,
      "learning_rate": 1.0096348454262845e-05,
      "loss": 0.0,
      "step": 49
    },
    {
      "epoch": 0.05900140128328048,
      "grad_norm": NaN,
      "learning_rate": 1e-05,
      "loss": 0.0,
      "step": 50
    },
    {
      "epoch": 0.05900140128328048,
      "eval_loss": NaN,
      "eval_runtime": 1.223,
      "eval_samples_per_second": 40.884,
      "eval_steps_per_second": 10.63,
      "step": 50
    }
  ],
  "logging_steps": 1,
  "max_steps": 50,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 25,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 1,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 1
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 8.430527686115328e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}