{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 29280,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.17,
      "grad_norm": 1.2632079124450684,
      "learning_rate": 6.25e-05,
      "loss": 6.1591,
      "step": 500
    },
    {
      "epoch": 0.34,
      "grad_norm": 1.0422550439834595,
      "learning_rate": 0.000125,
      "loss": 3.6369,
      "step": 1000
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.9257670044898987,
      "learning_rate": 0.0001875,
      "loss": 3.3421,
      "step": 1500
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8972496390342712,
      "learning_rate": 0.00025,
      "loss": 3.1677,
      "step": 2000
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.891448974609375,
      "learning_rate": 0.0003125,
      "loss": 3.0423,
      "step": 2500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4385253446525597,
      "eval_loss": 3.011591911315918,
      "eval_runtime": 3.3039,
      "eval_samples_per_second": 1359.31,
      "eval_steps_per_second": 10.896,
      "step": 2928
    },
    {
      "epoch": 1.02,
      "grad_norm": 0.8652610778808594,
      "learning_rate": 0.000375,
      "loss": 2.9336,
      "step": 3000
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.8467530012130737,
      "learning_rate": 0.00043750000000000006,
      "loss": 2.8609,
      "step": 3500
    },
    {
      "epoch": 1.37,
      "grad_norm": 0.823049783706665,
      "learning_rate": 0.0005,
      "loss": 2.812,
      "step": 4000
    },
    {
      "epoch": 1.54,
      "grad_norm": 0.8464782238006592,
      "learning_rate": 0.0005625000000000001,
      "loss": 2.7718,
      "step": 4500
    },
    {
      "epoch": 1.71,
      "grad_norm": 0.8122602701187134,
      "learning_rate": 0.000625,
      "loss": 2.7306,
      "step": 5000
    },
    {
      "epoch": 1.88,
      "grad_norm": 0.802824079990387,
      "learning_rate": 0.0006875,
      "loss": 2.712,
      "step": 5500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.45938947010381215,
      "eval_loss": 2.786367893218994,
      "eval_runtime": 3.3754,
      "eval_samples_per_second": 1330.491,
      "eval_steps_per_second": 10.665,
      "step": 5856
    },
    {
      "epoch": 2.05,
      "grad_norm": 0.8417896032333374,
      "learning_rate": 0.00075,
      "loss": 2.6766,
      "step": 6000
    },
    {
      "epoch": 2.22,
      "grad_norm": 0.7684197425842285,
      "learning_rate": 0.0008125,
      "loss": 2.6201,
      "step": 6500
    },
    {
      "epoch": 2.39,
      "grad_norm": 0.678827166557312,
      "learning_rate": 0.0008750000000000001,
      "loss": 2.6145,
      "step": 7000
    },
    {
      "epoch": 2.56,
      "grad_norm": 0.7093390822410583,
      "learning_rate": 0.0009375,
      "loss": 2.6112,
      "step": 7500
    },
    {
      "epoch": 2.73,
      "grad_norm": 0.6050966382026672,
      "learning_rate": 0.001,
      "loss": 2.5901,
      "step": 8000
    },
    {
      "epoch": 2.9,
      "grad_norm": 0.6185945272445679,
      "learning_rate": 0.0010625,
      "loss": 2.5896,
      "step": 8500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.469654970483399,
      "eval_loss": 2.6919634342193604,
      "eval_runtime": 3.3644,
      "eval_samples_per_second": 1334.861,
      "eval_steps_per_second": 10.7,
      "step": 8784
    },
    {
      "epoch": 3.07,
      "grad_norm": 0.5936027765274048,
      "learning_rate": 0.0011250000000000001,
      "loss": 2.5412,
      "step": 9000
    },
    {
      "epoch": 3.24,
      "grad_norm": 0.6204653978347778,
      "learning_rate": 0.0011875,
      "loss": 2.5202,
      "step": 9500
    },
    {
      "epoch": 3.42,
      "grad_norm": 0.537083625793457,
      "learning_rate": 0.00125,
      "loss": 2.511,
      "step": 10000
    },
    {
      "epoch": 3.59,
      "grad_norm": 0.5039043426513672,
      "learning_rate": 0.0013125,
      "loss": 2.5106,
      "step": 10500
    },
    {
      "epoch": 3.76,
      "grad_norm": 0.5108810067176819,
      "learning_rate": 0.001375,
      "loss": 2.5283,
      "step": 11000
    },
    {
      "epoch": 3.93,
      "grad_norm": 0.4879595935344696,
      "learning_rate": 0.0014375000000000002,
      "loss": 2.5169,
      "step": 11500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.4748219097863268,
      "eval_loss": 2.643162250518799,
      "eval_runtime": 3.363,
      "eval_samples_per_second": 1335.399,
      "eval_steps_per_second": 10.705,
      "step": 11712
    },
    {
      "epoch": 4.1,
      "grad_norm": 0.45613226294517517,
      "learning_rate": 0.0015,
      "loss": 2.4625,
      "step": 12000
    },
    {
      "epoch": 4.27,
      "grad_norm": 0.43395110964775085,
      "learning_rate": 0.0015625,
      "loss": 2.4584,
      "step": 12500
    },
    {
      "epoch": 4.44,
      "grad_norm": 0.4137372374534607,
      "learning_rate": 0.001625,
      "loss": 2.4641,
      "step": 13000
    },
    {
      "epoch": 4.61,
      "grad_norm": 0.43818387389183044,
      "learning_rate": 0.0016875,
      "loss": 2.4658,
      "step": 13500
    },
    {
      "epoch": 4.78,
      "grad_norm": 0.4149482846260071,
      "learning_rate": 0.0017500000000000003,
      "loss": 2.462,
      "step": 14000
    },
    {
      "epoch": 4.95,
      "grad_norm": 0.37128859758377075,
      "learning_rate": 0.0018124999999999999,
      "loss": 2.4655,
      "step": 14500
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.48043243091607535,
      "eval_loss": 2.604926347732544,
      "eval_runtime": 3.381,
      "eval_samples_per_second": 1328.318,
      "eval_steps_per_second": 10.648,
      "step": 14640
    },
    {
      "epoch": 5.12,
      "grad_norm": 0.40244346857070923,
      "learning_rate": 0.001875,
      "loss": 2.398,
      "step": 15000
    },
    {
      "epoch": 5.29,
      "grad_norm": 0.4036855399608612,
      "learning_rate": 0.0019375000000000002,
      "loss": 2.4045,
      "step": 15500
    },
    {
      "epoch": 5.46,
      "grad_norm": 0.3680976927280426,
      "learning_rate": 0.002,
      "loss": 2.4195,
      "step": 16000
    },
    {
      "epoch": 5.64,
      "grad_norm": 0.2993474304676056,
      "learning_rate": 0.0020625,
      "loss": 2.4271,
      "step": 16500
    },
    {
      "epoch": 5.81,
      "grad_norm": 0.394279807806015,
      "learning_rate": 0.002125,
      "loss": 2.426,
      "step": 17000
    },
    {
      "epoch": 5.98,
      "grad_norm": 0.372591495513916,
      "learning_rate": 0.0021874999999999998,
      "loss": 2.4298,
      "step": 17500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.4831114547555303,
      "eval_loss": 2.5832390785217285,
      "eval_runtime": 3.3575,
      "eval_samples_per_second": 1337.597,
      "eval_steps_per_second": 10.722,
      "step": 17568
    },
    {
      "epoch": 6.15,
      "grad_norm": 0.36848798394203186,
      "learning_rate": 0.0022500000000000003,
      "loss": 2.3586,
      "step": 18000
    },
    {
      "epoch": 6.32,
      "grad_norm": 0.36105024814605713,
      "learning_rate": 0.0023125000000000003,
      "loss": 2.358,
      "step": 18500
    },
    {
      "epoch": 6.49,
      "grad_norm": 0.3024386763572693,
      "learning_rate": 0.002375,
      "loss": 2.3776,
      "step": 19000
    },
    {
      "epoch": 6.66,
      "grad_norm": 0.3114764094352722,
      "learning_rate": 0.0024375,
      "loss": 2.3966,
      "step": 19500
    },
    {
      "epoch": 6.83,
      "grad_norm": 0.35501497983932495,
      "learning_rate": 0.0025,
      "loss": 2.3975,
      "step": 20000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.48376893770042273,
      "eval_loss": 2.5711376667022705,
      "eval_runtime": 3.3747,
      "eval_samples_per_second": 1330.765,
      "eval_steps_per_second": 10.667,
      "step": 20496
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.2878439426422119,
      "learning_rate": 0.0025625,
      "loss": 2.4054,
      "step": 20500
    },
    {
      "epoch": 7.17,
      "grad_norm": 0.31277361512184143,
      "learning_rate": 0.002625,
      "loss": 2.322,
      "step": 21000
    },
    {
      "epoch": 7.34,
      "grad_norm": 0.35922256112098694,
      "learning_rate": 0.0026875000000000002,
      "loss": 2.3412,
      "step": 21500
    },
    {
      "epoch": 7.51,
      "grad_norm": 0.29865553975105286,
      "learning_rate": 0.00275,
      "loss": 2.3535,
      "step": 22000
    },
    {
      "epoch": 7.68,
      "grad_norm": 0.275280237197876,
      "learning_rate": 0.0028125,
      "loss": 2.3584,
      "step": 22500
    },
    {
      "epoch": 7.86,
      "grad_norm": 0.28225448727607727,
      "learning_rate": 0.0028750000000000004,
      "loss": 2.3628,
      "step": 23000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.48590794888113936,
      "eval_loss": 2.5585122108459473,
      "eval_runtime": 3.3611,
      "eval_samples_per_second": 1336.181,
      "eval_steps_per_second": 10.711,
      "step": 23424
    },
    {
      "epoch": 8.03,
      "grad_norm": 0.3388291001319885,
      "learning_rate": 0.0029375,
      "loss": 2.3561,
      "step": 23500
    },
    {
      "epoch": 8.2,
      "grad_norm": 0.2987268567085266,
      "learning_rate": 0.003,
      "loss": 2.2985,
      "step": 24000
    },
    {
      "epoch": 8.37,
      "grad_norm": 0.3162915110588074,
      "learning_rate": 0.002715909090909091,
      "loss": 2.3115,
      "step": 24500
    },
    {
      "epoch": 8.54,
      "grad_norm": 0.32822996377944946,
      "learning_rate": 0.0024318181818181817,
      "loss": 2.3054,
      "step": 25000
    },
    {
      "epoch": 8.71,
      "grad_norm": 0.3083394467830658,
      "learning_rate": 0.002147727272727273,
      "loss": 2.2944,
      "step": 25500
    },
    {
      "epoch": 8.88,
      "grad_norm": 0.24597662687301636,
      "learning_rate": 0.0018636363636363638,
      "loss": 2.2882,
      "step": 26000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.4934400734978268,
      "eval_loss": 2.5040788650512695,
      "eval_runtime": 3.3615,
      "eval_samples_per_second": 1336.001,
      "eval_steps_per_second": 10.709,
      "step": 26352
    },
    {
      "epoch": 9.05,
      "grad_norm": 0.3208218514919281,
      "learning_rate": 0.0015795454545454546,
      "loss": 2.2569,
      "step": 26500
    },
    {
      "epoch": 9.22,
      "grad_norm": 0.29910457134246826,
      "learning_rate": 0.0012954545454545456,
      "loss": 2.1707,
      "step": 27000
    },
    {
      "epoch": 9.39,
      "grad_norm": 0.2915799617767334,
      "learning_rate": 0.0010113636363636364,
      "loss": 2.1683,
      "step": 27500
    },
    {
      "epoch": 9.56,
      "grad_norm": 0.33277857303619385,
      "learning_rate": 0.0007272727272727273,
      "loss": 2.1691,
      "step": 28000
    },
    {
      "epoch": 9.73,
      "grad_norm": 0.2690746784210205,
      "learning_rate": 0.0004431818181818182,
      "loss": 2.1603,
      "step": 28500
    },
    {
      "epoch": 9.9,
      "grad_norm": 0.30213144421577454,
      "learning_rate": 0.0001590909090909091,
      "loss": 2.144,
      "step": 29000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.4994556041216291,
      "eval_loss": 2.47524094581604,
      "eval_runtime": 3.35,
      "eval_samples_per_second": 1340.589,
      "eval_steps_per_second": 10.746,
      "step": 29280
    },
    {
      "epoch": 10.0,
      "step": 29280,
      "total_flos": 2273237316403200.0,
      "train_loss": 2.5690583817945805,
      "train_runtime": 746.5661,
      "train_samples_per_second": 627.473,
      "train_steps_per_second": 39.22
    }
  ],
  "logging_steps": 500,
  "max_steps": 29280,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 2000,
  "total_flos": 2273237316403200.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}