{
  "best_metric": 1.6451231241226196,
  "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/microsoft/phi_2_amazon/checkpoint-350",
  "epoch": 0.9210526315789473,
  "eval_steps": 50,
  "global_step": 350,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "grad_norm": 223.38046264648438,
      "learning_rate": 4.8684210526315795e-06,
      "loss": 4.075,
      "step": 10
    },
    {
      "epoch": 0.05,
      "grad_norm": 277.74298095703125,
      "learning_rate": 4.736842105263158e-06,
      "loss": 3.6094,
      "step": 20
    },
    {
      "epoch": 0.08,
      "grad_norm": 109.10823059082031,
      "learning_rate": 4.605263157894737e-06,
      "loss": 3.5742,
      "step": 30
    },
    {
      "epoch": 0.11,
      "grad_norm": 155.17674255371094,
      "learning_rate": 4.473684210526316e-06,
      "loss": 3.325,
      "step": 40
    },
    {
      "epoch": 0.13,
      "grad_norm": 141.57481384277344,
      "learning_rate": 4.342105263157895e-06,
      "loss": 3.2062,
      "step": 50
    },
    {
      "epoch": 0.13,
      "eval_accuracy": 0.0922266139657444,
      "eval_f1_macro": 0.06341542983984753,
      "eval_f1_micro": 0.0922266139657444,
      "eval_loss": 3.186594247817993,
      "eval_runtime": 7.2855,
      "eval_samples_per_second": 208.358,
      "eval_steps_per_second": 6.588,
      "step": 50
    },
    {
      "epoch": 0.16,
      "grad_norm": 144.5673370361328,
      "learning_rate": 4.210526315789474e-06,
      "loss": 3.1797,
      "step": 60
    },
    {
      "epoch": 0.18,
      "grad_norm": 124.39603424072266,
      "learning_rate": 4.078947368421053e-06,
      "loss": 3.0828,
      "step": 70
    },
    {
      "epoch": 0.21,
      "grad_norm": 151.6533660888672,
      "learning_rate": 3.947368421052632e-06,
      "loss": 2.9945,
      "step": 80
    },
    {
      "epoch": 0.24,
      "grad_norm": 105.09723663330078,
      "learning_rate": 3.815789473684211e-06,
      "loss": 2.982,
      "step": 90
    },
    {
      "epoch": 0.26,
      "grad_norm": 141.55398559570312,
      "learning_rate": 3.6842105263157896e-06,
      "loss": 2.9492,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_accuracy": 0.152832674571805,
      "eval_f1_macro": 0.10809380474427124,
      "eval_f1_micro": 0.152832674571805,
      "eval_loss": 2.908843755722046,
      "eval_runtime": 7.3105,
      "eval_samples_per_second": 207.647,
      "eval_steps_per_second": 6.566,
      "step": 100
    },
    {
      "epoch": 0.29,
      "grad_norm": 451.0547790527344,
      "learning_rate": 3.5526315789473687e-06,
      "loss": 2.975,
      "step": 110
    },
    {
      "epoch": 0.32,
      "grad_norm": 124.9840316772461,
      "learning_rate": 3.421052631578948e-06,
      "loss": 2.8461,
      "step": 120
    },
    {
      "epoch": 0.34,
      "grad_norm": 199.818359375,
      "learning_rate": 3.289473684210527e-06,
      "loss": 2.8281,
      "step": 130
    },
    {
      "epoch": 0.37,
      "grad_norm": 135.9771728515625,
      "learning_rate": 3.157894736842105e-06,
      "loss": 2.8516,
      "step": 140
    },
    {
      "epoch": 0.39,
      "grad_norm": 106.21376037597656,
      "learning_rate": 3.0263157894736843e-06,
      "loss": 2.6945,
      "step": 150
    },
    {
      "epoch": 0.39,
      "eval_accuracy": 0.22859025032938077,
      "eval_f1_macro": 0.16931697526842224,
      "eval_f1_micro": 0.22859025032938077,
      "eval_loss": 2.6944169998168945,
      "eval_runtime": 7.32,
      "eval_samples_per_second": 207.377,
      "eval_steps_per_second": 6.557,
      "step": 150
    },
    {
      "epoch": 0.42,
      "grad_norm": 163.9792022705078,
      "learning_rate": 2.8947368421052634e-06,
      "loss": 2.682,
      "step": 160
    },
    {
      "epoch": 0.45,
      "grad_norm": 228.08639526367188,
      "learning_rate": 2.7631578947368424e-06,
      "loss": 2.6945,
      "step": 170
    },
    {
      "epoch": 0.47,
      "grad_norm": 237.11058044433594,
      "learning_rate": 2.631578947368421e-06,
      "loss": 2.5578,
      "step": 180
    },
    {
      "epoch": 0.5,
      "grad_norm": 115.18451690673828,
      "learning_rate": 2.5e-06,
      "loss": 2.5758,
      "step": 190
    },
    {
      "epoch": 0.53,
      "grad_norm": 223.4733123779297,
      "learning_rate": 2.368421052631579e-06,
      "loss": 2.457,
      "step": 200
    },
    {
      "epoch": 0.53,
      "eval_accuracy": 0.3372859025032938,
      "eval_f1_macro": 0.25289488806347304,
      "eval_f1_micro": 0.3372859025032938,
      "eval_loss": 2.4136712551116943,
      "eval_runtime": 7.3246,
      "eval_samples_per_second": 207.247,
      "eval_steps_per_second": 6.553,
      "step": 200
    },
    {
      "epoch": 0.55,
      "grad_norm": 209.8923797607422,
      "learning_rate": 2.236842105263158e-06,
      "loss": 2.2898,
      "step": 210
    },
    {
      "epoch": 0.58,
      "grad_norm": 160.7167205810547,
      "learning_rate": 2.105263157894737e-06,
      "loss": 2.2609,
      "step": 220
    },
    {
      "epoch": 0.61,
      "grad_norm": 241.58717346191406,
      "learning_rate": 1.973684210526316e-06,
      "loss": 2.2477,
      "step": 230
    },
    {
      "epoch": 0.63,
      "grad_norm": 188.09275817871094,
      "learning_rate": 1.8421052631578948e-06,
      "loss": 2.0973,
      "step": 240
    },
    {
      "epoch": 0.66,
      "grad_norm": 116.9455337524414,
      "learning_rate": 1.710526315789474e-06,
      "loss": 2.0566,
      "step": 250
    },
    {
      "epoch": 0.66,
      "eval_accuracy": 0.4499341238471673,
      "eval_f1_macro": 0.35408778840994787,
      "eval_f1_micro": 0.4499341238471673,
      "eval_loss": 2.0551609992980957,
      "eval_runtime": 7.3154,
      "eval_samples_per_second": 207.508,
      "eval_steps_per_second": 6.562,
      "step": 250
    },
    {
      "epoch": 0.68,
      "grad_norm": 124.80367279052734,
      "learning_rate": 1.5789473684210526e-06,
      "loss": 2.0695,
      "step": 260
    },
    {
      "epoch": 0.71,
      "grad_norm": 152.9394073486328,
      "learning_rate": 1.4473684210526317e-06,
      "loss": 1.9953,
      "step": 270
    },
    {
      "epoch": 0.74,
      "grad_norm": 146.58970642089844,
      "learning_rate": 1.3157894736842106e-06,
      "loss": 1.8379,
      "step": 280
    },
    {
      "epoch": 0.76,
      "grad_norm": 177.54376220703125,
      "learning_rate": 1.1842105263157894e-06,
      "loss": 1.9055,
      "step": 290
    },
    {
      "epoch": 0.79,
      "grad_norm": 283.8413391113281,
      "learning_rate": 1.0526315789473685e-06,
      "loss": 1.7723,
      "step": 300
    },
    {
      "epoch": 0.79,
      "eval_accuracy": 0.5263504611330698,
      "eval_f1_macro": 0.422505143115707,
      "eval_f1_micro": 0.5263504611330698,
      "eval_loss": 1.7764842510223389,
      "eval_runtime": 7.3136,
      "eval_samples_per_second": 207.558,
      "eval_steps_per_second": 6.563,
      "step": 300
    },
    {
      "epoch": 0.82,
      "grad_norm": 137.7584686279297,
      "learning_rate": 9.210526315789474e-07,
      "loss": 1.7563,
      "step": 310
    },
    {
      "epoch": 0.84,
      "grad_norm": 84.1349105834961,
      "learning_rate": 7.894736842105263e-07,
      "loss": 1.7305,
      "step": 320
    },
    {
      "epoch": 0.87,
      "grad_norm": 113.17578887939453,
      "learning_rate": 6.578947368421053e-07,
      "loss": 1.7602,
      "step": 330
    },
    {
      "epoch": 0.89,
      "grad_norm": 158.471923828125,
      "learning_rate": 5.263157894736843e-07,
      "loss": 1.6398,
      "step": 340
    },
    {
      "epoch": 0.92,
      "grad_norm": 150.00511169433594,
      "learning_rate": 3.9473684210526315e-07,
      "loss": 1.7695,
      "step": 350
    },
    {
      "epoch": 0.92,
      "eval_accuracy": 0.5658761528326746,
      "eval_f1_macro": 0.4654710477619855,
      "eval_f1_micro": 0.5658761528326746,
      "eval_loss": 1.6451231241226196,
      "eval_runtime": 7.4633,
      "eval_samples_per_second": 203.396,
      "eval_steps_per_second": 6.431,
      "step": 350
    }
  ],
  "logging_steps": 10,
  "max_steps": 380,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "total_flos": 2.165493586722816e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}