{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 31.020408163265305,
  "eval_steps": 20,
  "global_step": 380,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.82,
      "grad_norm": 0.18377472460269928,
      "learning_rate": 2.9999999999999997e-05,
      "loss": 1.861,
      "step": 10
    },
    {
      "epoch": 1.63,
      "grad_norm": 0.35202744603157043,
      "learning_rate": 5.9999999999999995e-05,
      "loss": 1.7263,
      "step": 20
    },
    {
      "epoch": 1.63,
      "eval_loss": 1.4457755088806152,
      "eval_runtime": 89.8938,
      "eval_samples_per_second": 4.305,
      "eval_steps_per_second": 0.545,
      "step": 20
    },
    {
      "epoch": 2.45,
      "grad_norm": 0.928983747959137,
      "learning_rate": 8.999999999999999e-05,
      "loss": 1.1718,
      "step": 30
    },
    {
      "epoch": 3.27,
      "grad_norm": 0.253262996673584,
      "learning_rate": 0.00011999999999999999,
      "loss": 0.4789,
      "step": 40
    },
    {
      "epoch": 3.27,
      "eval_loss": 0.3332095146179199,
      "eval_runtime": 89.9804,
      "eval_samples_per_second": 4.301,
      "eval_steps_per_second": 0.545,
      "step": 40
    },
    {
      "epoch": 4.08,
      "grad_norm": 0.12236642092466354,
      "learning_rate": 0.00015,
      "loss": 0.3568,
      "step": 50
    },
    {
      "epoch": 4.9,
      "grad_norm": 0.09160923212766647,
      "learning_rate": 0.00017999999999999998,
      "loss": 0.3256,
      "step": 60
    },
    {
      "epoch": 4.9,
      "eval_loss": 0.2753114104270935,
      "eval_runtime": 90.3206,
      "eval_samples_per_second": 4.285,
      "eval_steps_per_second": 0.543,
      "step": 60
    },
    {
      "epoch": 5.71,
      "grad_norm": 0.10242326557636261,
      "learning_rate": 0.00020999999999999998,
      "loss": 0.2841,
      "step": 70
    },
    {
      "epoch": 6.53,
      "grad_norm": 0.1305350810289383,
      "learning_rate": 0.00023999999999999998,
      "loss": 0.2615,
      "step": 80
    },
    {
      "epoch": 6.53,
      "eval_loss": 0.2476309835910797,
      "eval_runtime": 90.525,
      "eval_samples_per_second": 4.275,
      "eval_steps_per_second": 0.541,
      "step": 80
    },
    {
      "epoch": 7.35,
      "grad_norm": 0.17941106855869293,
      "learning_rate": 0.00027,
      "loss": 0.2216,
      "step": 90
    },
    {
      "epoch": 8.16,
      "grad_norm": 0.20095375180244446,
      "learning_rate": 0.0003,
      "loss": 0.1832,
      "step": 100
    },
    {
      "epoch": 8.16,
      "eval_loss": 0.2350914180278778,
      "eval_runtime": 90.3919,
      "eval_samples_per_second": 4.281,
      "eval_steps_per_second": 0.542,
      "step": 100
    },
    {
      "epoch": 8.98,
      "grad_norm": 0.2600422501564026,
      "learning_rate": 0.00029,
      "loss": 0.1441,
      "step": 110
    },
    {
      "epoch": 9.8,
      "grad_norm": 0.20544037222862244,
      "learning_rate": 0.00028,
      "loss": 0.1186,
      "step": 120
    },
    {
      "epoch": 9.8,
      "eval_loss": 0.23090216517448425,
      "eval_runtime": 90.3144,
      "eval_samples_per_second": 4.285,
      "eval_steps_per_second": 0.543,
      "step": 120
    },
    {
      "epoch": 10.61,
      "grad_norm": 0.2158157229423523,
      "learning_rate": 0.00027,
      "loss": 0.0947,
      "step": 130
    },
    {
      "epoch": 11.43,
      "grad_norm": 0.18916285037994385,
      "learning_rate": 0.00026,
      "loss": 0.0768,
      "step": 140
    },
    {
      "epoch": 11.43,
      "eval_loss": 0.24214179813861847,
      "eval_runtime": 90.2597,
      "eval_samples_per_second": 4.288,
      "eval_steps_per_second": 0.543,
      "step": 140
    },
    {
      "epoch": 12.24,
      "grad_norm": 0.22263498604297638,
      "learning_rate": 0.00025,
      "loss": 0.0615,
      "step": 150
    },
    {
      "epoch": 13.06,
      "grad_norm": 0.21315976977348328,
      "learning_rate": 0.00023999999999999998,
      "loss": 0.054,
      "step": 160
    },
    {
      "epoch": 13.06,
      "eval_loss": 0.25932466983795166,
      "eval_runtime": 89.8439,
      "eval_samples_per_second": 4.307,
      "eval_steps_per_second": 0.545,
      "step": 160
    },
    {
      "epoch": 13.88,
      "grad_norm": 0.18338361382484436,
      "learning_rate": 0.00023,
      "loss": 0.0455,
      "step": 170
    },
    {
      "epoch": 14.69,
      "grad_norm": 0.17157459259033203,
      "learning_rate": 0.00021999999999999995,
      "loss": 0.0393,
      "step": 180
    },
    {
      "epoch": 14.69,
      "eval_loss": 0.27233538031578064,
      "eval_runtime": 90.1364,
      "eval_samples_per_second": 4.293,
      "eval_steps_per_second": 0.544,
      "step": 180
    },
    {
      "epoch": 15.51,
      "grad_norm": 0.1541435867547989,
      "learning_rate": 0.00020999999999999998,
      "loss": 0.0352,
      "step": 190
    },
    {
      "epoch": 16.33,
      "grad_norm": 0.1553652435541153,
      "learning_rate": 0.00019999999999999998,
      "loss": 0.0325,
      "step": 200
    },
    {
      "epoch": 16.33,
      "eval_loss": 0.28704825043678284,
      "eval_runtime": 89.7951,
      "eval_samples_per_second": 4.31,
      "eval_steps_per_second": 0.546,
      "step": 200
    },
    {
      "epoch": 17.14,
      "grad_norm": 0.13403691351413727,
      "learning_rate": 0.00018999999999999998,
      "loss": 0.0297,
      "step": 210
    },
    {
      "epoch": 17.96,
      "grad_norm": 0.14512716233730316,
      "learning_rate": 0.00017999999999999998,
      "loss": 0.0279,
      "step": 220
    },
    {
      "epoch": 17.96,
      "eval_loss": 0.2964874505996704,
      "eval_runtime": 89.7009,
      "eval_samples_per_second": 4.314,
      "eval_steps_per_second": 0.546,
      "step": 220
    },
    {
      "epoch": 18.78,
      "grad_norm": 0.12400835007429123,
      "learning_rate": 0.00016999999999999999,
      "loss": 0.0263,
      "step": 230
    },
    {
      "epoch": 19.59,
      "grad_norm": 0.1139909029006958,
      "learning_rate": 0.00015999999999999999,
      "loss": 0.0246,
      "step": 240
    },
    {
      "epoch": 19.59,
      "eval_loss": 0.30519917607307434,
      "eval_runtime": 89.8387,
      "eval_samples_per_second": 4.308,
      "eval_steps_per_second": 0.545,
      "step": 240
    },
    {
      "epoch": 20.41,
      "grad_norm": 0.12317101657390594,
      "learning_rate": 0.00015,
      "loss": 0.0235,
      "step": 250
    },
    {
      "epoch": 21.22,
      "grad_norm": 0.12494686245918274,
      "learning_rate": 0.00014,
      "loss": 0.0224,
      "step": 260
    },
    {
      "epoch": 21.22,
      "eval_loss": 0.314134418964386,
      "eval_runtime": 89.7974,
      "eval_samples_per_second": 4.31,
      "eval_steps_per_second": 0.546,
      "step": 260
    },
    {
      "epoch": 22.04,
      "grad_norm": 0.1180659756064415,
      "learning_rate": 0.00013,
      "loss": 0.022,
      "step": 270
    },
    {
      "epoch": 22.86,
      "grad_norm": 0.09653373062610626,
      "learning_rate": 0.00011999999999999999,
      "loss": 0.0212,
      "step": 280
    },
    {
      "epoch": 22.86,
      "eval_loss": 0.3175604045391083,
      "eval_runtime": 89.9764,
      "eval_samples_per_second": 4.301,
      "eval_steps_per_second": 0.545,
      "step": 280
    },
    {
      "epoch": 23.67,
      "grad_norm": 0.10445748269557953,
      "learning_rate": 0.00010999999999999998,
      "loss": 0.0208,
      "step": 290
    },
    {
      "epoch": 24.49,
      "grad_norm": 0.09245337545871735,
      "learning_rate": 9.999999999999999e-05,
      "loss": 0.0199,
      "step": 300
    },
    {
      "epoch": 24.49,
      "eval_loss": 0.32360976934432983,
      "eval_runtime": 89.8613,
      "eval_samples_per_second": 4.307,
      "eval_steps_per_second": 0.545,
      "step": 300
    },
    {
      "epoch": 25.31,
      "grad_norm": 0.09468758851289749,
      "learning_rate": 8.999999999999999e-05,
      "loss": 0.0197,
      "step": 310
    },
    {
      "epoch": 26.12,
      "grad_norm": 0.0891977846622467,
      "learning_rate": 7.999999999999999e-05,
      "loss": 0.0192,
      "step": 320
    },
    {
      "epoch": 26.12,
      "eval_loss": 0.3267403841018677,
      "eval_runtime": 90.3063,
      "eval_samples_per_second": 4.285,
      "eval_steps_per_second": 0.543,
      "step": 320
    },
    {
      "epoch": 26.94,
      "grad_norm": 0.08574336767196655,
      "learning_rate": 7e-05,
      "loss": 0.0188,
      "step": 330
    },
    {
      "epoch": 27.76,
      "grad_norm": 0.08517367392778397,
      "learning_rate": 5.9999999999999995e-05,
      "loss": 0.0184,
      "step": 340
    },
    {
      "epoch": 27.76,
      "eval_loss": 0.33063870668411255,
      "eval_runtime": 89.8041,
      "eval_samples_per_second": 4.309,
      "eval_steps_per_second": 0.546,
      "step": 340
    },
    {
      "epoch": 28.57,
      "grad_norm": 0.08357132971286774,
      "learning_rate": 4.9999999999999996e-05,
      "loss": 0.0181,
      "step": 350
    },
    {
      "epoch": 29.39,
      "grad_norm": 0.08679915964603424,
      "learning_rate": 3.9999999999999996e-05,
      "loss": 0.0181,
      "step": 360
    },
    {
      "epoch": 29.39,
      "eval_loss": 0.3352932929992676,
      "eval_runtime": 90.2438,
      "eval_samples_per_second": 4.288,
      "eval_steps_per_second": 0.543,
      "step": 360
    },
    {
      "epoch": 30.2,
      "grad_norm": 0.07208231836557388,
      "learning_rate": 2.9999999999999997e-05,
      "loss": 0.0178,
      "step": 370
    },
    {
      "epoch": 31.02,
      "grad_norm": 0.08611435443162918,
      "learning_rate": 1.9999999999999998e-05,
      "loss": 0.0175,
      "step": 380
    },
    {
      "epoch": 31.02,
      "eval_loss": 0.338119238615036,
      "eval_runtime": 90.2767,
      "eval_samples_per_second": 4.287,
      "eval_steps_per_second": 0.543,
      "step": 380
    }
  ],
  "logging_steps": 10,
  "max_steps": 400,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 34,
  "save_steps": 20,
  "total_flos": 2.1789060364511478e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}