{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 29.387755102040817,
  "eval_steps": 20,
  "global_step": 360,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.82,
      "grad_norm": 0.1748005896806717,
      "learning_rate": 2.1e-05,
      "loss": 4.8309,
      "step": 10
    },
    {
      "epoch": 1.63,
      "grad_norm": 0.19310609996318817,
      "learning_rate": 3.2999999999999996e-05,
      "loss": 5.0159,
      "step": 20
    },
    {
      "epoch": 1.63,
      "eval_loss": NaN,
      "eval_runtime": 53.474,
      "eval_samples_per_second": 7.237,
      "eval_steps_per_second": 0.916,
      "step": 20
    },
    {
      "epoch": 2.45,
      "grad_norm": NaN,
      "learning_rate": 4.2e-05,
      "loss": 4.8025,
      "step": 30
    },
    {
      "epoch": 3.27,
      "grad_norm": NaN,
      "learning_rate": 5.399999999999999e-05,
      "loss": 4.1014,
      "step": 40
    },
    {
      "epoch": 3.27,
      "eval_loss": NaN,
      "eval_runtime": 53.4777,
      "eval_samples_per_second": 7.237,
      "eval_steps_per_second": 0.916,
      "step": 40
    },
    {
      "epoch": 4.08,
      "grad_norm": 0.0,
      "learning_rate": 7.199999999999999e-05,
      "loss": 4.7645,
      "step": 50
    },
    {
      "epoch": 4.9,
      "grad_norm": 0.0,
      "learning_rate": 9.3e-05,
      "loss": 4.3747,
      "step": 60
    },
    {
      "epoch": 4.9,
      "eval_loss": NaN,
      "eval_runtime": 53.6778,
      "eval_samples_per_second": 7.21,
      "eval_steps_per_second": 0.913,
      "step": 60
    },
    {
      "epoch": 5.71,
      "grad_norm": NaN,
      "learning_rate": 0.000102,
      "loss": 5.0924,
      "step": 70
    },
    {
      "epoch": 6.53,
      "grad_norm": NaN,
      "learning_rate": 0.00011099999999999999,
      "loss": 4.3234,
      "step": 80
    },
    {
      "epoch": 6.53,
      "eval_loss": NaN,
      "eval_runtime": 53.7614,
      "eval_samples_per_second": 7.198,
      "eval_steps_per_second": 0.911,
      "step": 80
    },
    {
      "epoch": 7.35,
      "grad_norm": 0.0,
      "learning_rate": 0.00012599999999999997,
      "loss": 4.1684,
      "step": 90
    },
    {
      "epoch": 8.16,
      "grad_norm": 0.0,
      "learning_rate": 0.00014399999999999998,
      "loss": 4.3018,
      "step": 100
    },
    {
      "epoch": 8.16,
      "eval_loss": NaN,
      "eval_runtime": 53.6674,
      "eval_samples_per_second": 7.211,
      "eval_steps_per_second": 0.913,
      "step": 100
    },
    {
      "epoch": 8.98,
      "grad_norm": NaN,
      "learning_rate": 0.000162,
      "loss": 4.8328,
      "step": 110
    },
    {
      "epoch": 9.8,
      "grad_norm": NaN,
      "learning_rate": 0.00017099999999999998,
      "loss": 3.8634,
      "step": 120
    },
    {
      "epoch": 9.8,
      "eval_loss": NaN,
      "eval_runtime": 53.8298,
      "eval_samples_per_second": 7.189,
      "eval_steps_per_second": 0.91,
      "step": 120
    },
    {
      "epoch": 10.61,
      "grad_norm": 0.0,
      "learning_rate": 0.00018299999999999998,
      "loss": 4.8214,
      "step": 130
    },
    {
      "epoch": 11.43,
      "grad_norm": 0.0,
      "learning_rate": 0.000195,
      "loss": 5.3188,
      "step": 140
    },
    {
      "epoch": 11.43,
      "eval_loss": NaN,
      "eval_runtime": 53.6779,
      "eval_samples_per_second": 7.21,
      "eval_steps_per_second": 0.913,
      "step": 140
    },
    {
      "epoch": 12.24,
      "grad_norm": NaN,
      "learning_rate": 0.00020999999999999998,
      "loss": 3.932,
      "step": 150
    },
    {
      "epoch": 13.06,
      "grad_norm": NaN,
      "learning_rate": 0.00022799999999999999,
      "loss": 4.5087,
      "step": 160
    },
    {
      "epoch": 13.06,
      "eval_loss": NaN,
      "eval_runtime": 53.7126,
      "eval_samples_per_second": 7.205,
      "eval_steps_per_second": 0.912,
      "step": 160
    },
    {
      "epoch": 13.88,
      "grad_norm": 0.0,
      "learning_rate": 0.00023999999999999998,
      "loss": 4.1573,
      "step": 170
    },
    {
      "epoch": 14.69,
      "grad_norm": NaN,
      "learning_rate": 0.000249,
      "loss": 3.9824,
      "step": 180
    },
    {
      "epoch": 14.69,
      "eval_loss": NaN,
      "eval_runtime": 53.7148,
      "eval_samples_per_second": 7.205,
      "eval_steps_per_second": 0.912,
      "step": 180
    },
    {
      "epoch": 15.51,
      "grad_norm": NaN,
      "learning_rate": 0.000261,
      "loss": 3.9201,
      "step": 190
    },
    {
      "epoch": 16.33,
      "grad_norm": 0.0,
      "learning_rate": 0.00028199999999999997,
      "loss": 3.9784,
      "step": 200
    },
    {
      "epoch": 16.33,
      "eval_loss": NaN,
      "eval_runtime": 53.6797,
      "eval_samples_per_second": 7.209,
      "eval_steps_per_second": 0.913,
      "step": 200
    },
    {
      "epoch": 17.14,
      "grad_norm": 0.0,
      "learning_rate": 0.0003,
      "loss": 4.5915,
      "step": 210
    },
    {
      "epoch": 17.96,
      "grad_norm": NaN,
      "learning_rate": 0.00029699999999999996,
      "loss": 4.8878,
      "step": 220
    },
    {
      "epoch": 17.96,
      "eval_loss": NaN,
      "eval_runtime": 53.8508,
      "eval_samples_per_second": 7.187,
      "eval_steps_per_second": 0.91,
      "step": 220
    },
    {
      "epoch": 18.78,
      "grad_norm": NaN,
      "learning_rate": 0.000294,
      "loss": 3.5563,
      "step": 230
    },
    {
      "epoch": 19.59,
      "grad_norm": 0.0,
      "learning_rate": 0.000289,
      "loss": 4.5682,
      "step": 240
    },
    {
      "epoch": 19.59,
      "eval_loss": NaN,
      "eval_runtime": 53.6594,
      "eval_samples_per_second": 7.212,
      "eval_steps_per_second": 0.913,
      "step": 240
    },
    {
      "epoch": 20.41,
      "grad_norm": 0.0,
      "learning_rate": 0.00028199999999999997,
      "loss": 4.3472,
      "step": 250
    },
    {
      "epoch": 21.22,
      "grad_norm": NaN,
      "learning_rate": 0.00027699999999999996,
      "loss": 4.911,
      "step": 260
    },
    {
      "epoch": 21.22,
      "eval_loss": NaN,
      "eval_runtime": 53.6876,
      "eval_samples_per_second": 7.208,
      "eval_steps_per_second": 0.913,
      "step": 260
    },
    {
      "epoch": 22.04,
      "grad_norm": NaN,
      "learning_rate": 0.000274,
      "loss": 4.5261,
      "step": 270
    },
    {
      "epoch": 22.86,
      "grad_norm": NaN,
      "learning_rate": 0.00027299999999999997,
      "loss": 166.5225,
      "step": 280
    },
    {
      "epoch": 22.86,
      "eval_loss": NaN,
      "eval_runtime": 52.9157,
      "eval_samples_per_second": 7.314,
      "eval_steps_per_second": 0.926,
      "step": 280
    },
    {
      "epoch": 23.67,
      "grad_norm": NaN,
      "learning_rate": 0.00027299999999999997,
      "loss": 0.0,
      "step": 290
    },
    {
      "epoch": 24.49,
      "grad_norm": NaN,
      "learning_rate": 0.00027299999999999997,
      "loss": 0.0,
      "step": 300
    },
    {
      "epoch": 24.49,
      "eval_loss": NaN,
      "eval_runtime": 52.9419,
      "eval_samples_per_second": 7.31,
      "eval_steps_per_second": 0.926,
      "step": 300
    },
    {
      "epoch": 25.31,
      "grad_norm": NaN,
      "learning_rate": 0.00027299999999999997,
      "loss": 0.0,
      "step": 310
    },
    {
      "epoch": 26.12,
      "grad_norm": NaN,
      "learning_rate": 0.00027299999999999997,
      "loss": 0.0,
      "step": 320
    },
    {
      "epoch": 26.12,
      "eval_loss": NaN,
      "eval_runtime": 52.9285,
      "eval_samples_per_second": 7.312,
      "eval_steps_per_second": 0.926,
      "step": 320
    },
    {
      "epoch": 26.94,
      "grad_norm": NaN,
      "learning_rate": 0.00027299999999999997,
      "loss": 0.0,
      "step": 330
    },
    {
      "epoch": 27.76,
      "grad_norm": NaN,
      "learning_rate": 0.00027299999999999997,
      "loss": 0.0,
      "step": 340
    },
    {
      "epoch": 27.76,
      "eval_loss": NaN,
      "eval_runtime": 52.861,
      "eval_samples_per_second": 7.321,
      "eval_steps_per_second": 0.927,
      "step": 340
    },
    {
      "epoch": 28.57,
      "grad_norm": NaN,
      "learning_rate": 0.00027299999999999997,
      "loss": 0.0,
      "step": 350
    },
    {
      "epoch": 29.39,
      "grad_norm": NaN,
      "learning_rate": 0.00027299999999999997,
      "loss": 0.0,
      "step": 360
    },
    {
      "epoch": 29.39,
      "eval_loss": NaN,
      "eval_runtime": 52.8275,
      "eval_samples_per_second": 7.326,
      "eval_steps_per_second": 0.928,
      "step": 360
    }
  ],
  "logging_steps": 10,
  "max_steps": 400,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 34,
  "save_steps": 20,
  "total_flos": 9.114505187898163e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}