{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 8.88888888888889,
  "eval_steps": 256,
  "global_step": 10000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.23,
      "learning_rate": 4.872000000000001e-05,
      "loss": 0.8173,
      "step": 256
    },
    {
      "epoch": 0.23,
      "eval_test_accuracy": 0.09690721649484536,
      "eval_test_loss": 0.4266754984855652,
      "eval_test_runtime": 233.5115,
      "eval_test_samples_per_second": 10.385,
      "eval_test_steps_per_second": 1.302,
      "step": 256
    },
    {
      "epoch": 0.46,
      "learning_rate": 4.744e-05,
      "loss": 0.441,
      "step": 512
    },
    {
      "epoch": 0.46,
      "eval_test_accuracy": 0.16742268041237113,
      "eval_test_loss": 0.36029067635536194,
      "eval_test_runtime": 232.4589,
      "eval_test_samples_per_second": 10.432,
      "eval_test_steps_per_second": 1.308,
      "step": 512
    },
    {
      "epoch": 0.68,
      "learning_rate": 4.6160000000000005e-05,
      "loss": 0.397,
      "step": 768
    },
    {
      "epoch": 0.68,
      "eval_test_accuracy": 0.17649484536082474,
      "eval_test_loss": 0.3348633348941803,
      "eval_test_runtime": 227.1069,
      "eval_test_samples_per_second": 10.678,
      "eval_test_steps_per_second": 1.339,
      "step": 768
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.488e-05,
      "loss": 0.3611,
      "step": 1024
    },
    {
      "epoch": 0.91,
      "eval_test_accuracy": 0.23010309278350516,
      "eval_test_loss": 0.3209882974624634,
      "eval_test_runtime": 226.4132,
      "eval_test_samples_per_second": 10.711,
      "eval_test_steps_per_second": 1.343,
      "step": 1024
    },
    {
      "epoch": 1.14,
      "learning_rate": 4.36e-05,
      "loss": 0.3469,
      "step": 1280
    },
    {
      "epoch": 1.14,
      "eval_test_accuracy": 0.19505154639175257,
      "eval_test_loss": 0.31938230991363525,
      "eval_test_runtime": 227.1405,
      "eval_test_samples_per_second": 10.676,
      "eval_test_steps_per_second": 1.338,
      "step": 1280
    },
    {
      "epoch": 1.37,
      "learning_rate": 4.232e-05,
      "loss": 0.3301,
      "step": 1536
    },
    {
      "epoch": 1.37,
      "eval_test_accuracy": 0.21608247422680413,
      "eval_test_loss": 0.3045227825641632,
      "eval_test_runtime": 229.033,
      "eval_test_samples_per_second": 10.588,
      "eval_test_steps_per_second": 1.327,
      "step": 1536
    },
    {
      "epoch": 1.59,
      "learning_rate": 4.104e-05,
      "loss": 0.3194,
      "step": 1792
    },
    {
      "epoch": 1.59,
      "eval_test_accuracy": 0.2354639175257732,
      "eval_test_loss": 0.2938610017299652,
      "eval_test_runtime": 225.016,
      "eval_test_samples_per_second": 10.777,
      "eval_test_steps_per_second": 1.351,
      "step": 1792
    },
    {
      "epoch": 1.82,
      "learning_rate": 3.9760000000000006e-05,
      "loss": 0.3189,
      "step": 2048
    },
    {
      "epoch": 1.82,
      "eval_test_accuracy": 0.23876288659793815,
      "eval_test_loss": 0.2826090455055237,
      "eval_test_runtime": 235.6637,
      "eval_test_samples_per_second": 10.29,
      "eval_test_steps_per_second": 1.29,
      "step": 2048
    },
    {
      "epoch": 2.05,
      "learning_rate": 3.848e-05,
      "loss": 0.3031,
      "step": 2304
    },
    {
      "epoch": 2.05,
      "eval_test_accuracy": 0.23752577319587628,
      "eval_test_loss": 0.276397705078125,
      "eval_test_runtime": 228.4566,
      "eval_test_samples_per_second": 10.615,
      "eval_test_steps_per_second": 1.331,
      "step": 2304
    },
    {
      "epoch": 2.28,
      "learning_rate": 3.72e-05,
      "loss": 0.3008,
      "step": 2560
    },
    {
      "epoch": 2.28,
      "eval_test_accuracy": 0.231340206185567,
      "eval_test_loss": 0.27680695056915283,
      "eval_test_runtime": 232.6253,
      "eval_test_samples_per_second": 10.424,
      "eval_test_steps_per_second": 1.307,
      "step": 2560
    },
    {
      "epoch": 2.5,
      "learning_rate": 3.592e-05,
      "loss": 0.2941,
      "step": 2816
    },
    {
      "epoch": 2.5,
      "eval_test_accuracy": 0.24082474226804124,
      "eval_test_loss": 0.27100667357444763,
      "eval_test_runtime": 234.9203,
      "eval_test_samples_per_second": 10.323,
      "eval_test_steps_per_second": 1.294,
      "step": 2816
    },
    {
      "epoch": 2.73,
      "learning_rate": 3.464e-05,
      "loss": 0.2855,
      "step": 3072
    },
    {
      "epoch": 2.73,
      "eval_test_accuracy": 0.2552577319587629,
      "eval_test_loss": 0.26808857917785645,
      "eval_test_runtime": 233.699,
      "eval_test_samples_per_second": 10.377,
      "eval_test_steps_per_second": 1.301,
      "step": 3072
    },
    {
      "epoch": 2.96,
      "learning_rate": 3.336e-05,
      "loss": 0.2827,
      "step": 3328
    },
    {
      "epoch": 2.96,
      "eval_test_accuracy": 0.23876288659793815,
      "eval_test_loss": 0.26740676164627075,
      "eval_test_runtime": 232.615,
      "eval_test_samples_per_second": 10.425,
      "eval_test_steps_per_second": 1.307,
      "step": 3328
    },
    {
      "epoch": 3.19,
      "learning_rate": 3.208e-05,
      "loss": 0.2824,
      "step": 3584
    },
    {
      "epoch": 3.19,
      "eval_test_accuracy": 0.23670103092783504,
      "eval_test_loss": 0.26493921875953674,
      "eval_test_runtime": 234.3958,
      "eval_test_samples_per_second": 10.346,
      "eval_test_steps_per_second": 1.297,
      "step": 3584
    },
    {
      "epoch": 3.41,
      "learning_rate": 3.08e-05,
      "loss": 0.2699,
      "step": 3840
    },
    {
      "epoch": 3.41,
      "eval_test_accuracy": 0.24082474226804124,
      "eval_test_loss": 0.2653910517692566,
      "eval_test_runtime": 236.6454,
      "eval_test_samples_per_second": 10.247,
      "eval_test_steps_per_second": 1.285,
      "step": 3840
    },
    {
      "epoch": 3.64,
      "learning_rate": 2.9520000000000002e-05,
      "loss": 0.2744,
      "step": 4096
    },
    {
      "epoch": 3.64,
      "eval_test_accuracy": 0.2503092783505155,
      "eval_test_loss": 0.25940871238708496,
      "eval_test_runtime": 230.5104,
      "eval_test_samples_per_second": 10.52,
      "eval_test_steps_per_second": 1.319,
      "step": 4096
    },
    {
      "epoch": 3.87,
      "learning_rate": 2.824e-05,
      "loss": 0.2679,
      "step": 4352
    },
    {
      "epoch": 3.87,
      "eval_test_accuracy": 0.24824742268041236,
      "eval_test_loss": 0.25839442014694214,
      "eval_test_runtime": 279.8888,
      "eval_test_samples_per_second": 8.664,
      "eval_test_steps_per_second": 1.086,
      "step": 4352
    },
    {
      "epoch": 4.1,
      "learning_rate": 2.6960000000000003e-05,
      "loss": 0.27,
      "step": 4608
    },
    {
      "epoch": 4.1,
      "eval_test_accuracy": 0.2465979381443299,
      "eval_test_loss": 0.25590091943740845,
      "eval_test_runtime": 229.037,
      "eval_test_samples_per_second": 10.588,
      "eval_test_steps_per_second": 1.327,
      "step": 4608
    },
    {
      "epoch": 4.32,
      "learning_rate": 2.5679999999999998e-05,
      "loss": 0.2644,
      "step": 4864
    },
    {
      "epoch": 4.32,
      "eval_test_accuracy": 0.23835051546391753,
      "eval_test_loss": 0.25387293100357056,
      "eval_test_runtime": 231.73,
      "eval_test_samples_per_second": 10.465,
      "eval_test_steps_per_second": 1.312,
      "step": 4864
    },
    {
      "epoch": 4.55,
      "learning_rate": 2.44e-05,
      "loss": 0.2664,
      "step": 5120
    },
    {
      "epoch": 4.55,
      "eval_test_accuracy": 0.24783505154639177,
      "eval_test_loss": 0.2522536516189575,
      "eval_test_runtime": 233.9019,
      "eval_test_samples_per_second": 10.368,
      "eval_test_steps_per_second": 1.3,
      "step": 5120
    },
    {
      "epoch": 4.78,
      "learning_rate": 2.312e-05,
      "loss": 0.2557,
      "step": 5376
    },
    {
      "epoch": 4.78,
      "eval_test_accuracy": 0.23628865979381444,
      "eval_test_loss": 0.2543907165527344,
      "eval_test_runtime": 230.6139,
      "eval_test_samples_per_second": 10.515,
      "eval_test_steps_per_second": 1.318,
      "step": 5376
    },
    {
      "epoch": 5.01,
      "learning_rate": 2.184e-05,
      "loss": 0.2544,
      "step": 5632
    },
    {
      "epoch": 5.01,
      "eval_test_accuracy": 0.2445360824742268,
      "eval_test_loss": 0.24817436933517456,
      "eval_test_runtime": 231.9449,
      "eval_test_samples_per_second": 10.455,
      "eval_test_steps_per_second": 1.311,
      "step": 5632
    },
    {
      "epoch": 5.23,
      "learning_rate": 2.0560000000000003e-05,
      "loss": 0.2549,
      "step": 5888
    },
    {
      "epoch": 5.23,
      "eval_test_accuracy": 0.24618556701030928,
      "eval_test_loss": 0.24915936589241028,
      "eval_test_runtime": 226.5831,
      "eval_test_samples_per_second": 10.702,
      "eval_test_steps_per_second": 1.342,
      "step": 5888
    },
    {
      "epoch": 5.46,
      "learning_rate": 1.9280000000000002e-05,
      "loss": 0.2545,
      "step": 6144
    },
    {
      "epoch": 5.46,
      "eval_test_accuracy": 0.2465979381443299,
      "eval_test_loss": 0.2478923350572586,
      "eval_test_runtime": 233.3756,
      "eval_test_samples_per_second": 10.391,
      "eval_test_steps_per_second": 1.303,
      "step": 6144
    },
    {
      "epoch": 5.69,
      "learning_rate": 1.8e-05,
      "loss": 0.2539,
      "step": 6400
    },
    {
      "epoch": 5.69,
      "eval_test_accuracy": 0.24989690721649485,
      "eval_test_loss": 0.24747271835803986,
      "eval_test_runtime": 231.4653,
      "eval_test_samples_per_second": 10.477,
      "eval_test_steps_per_second": 1.313,
      "step": 6400
    },
    {
      "epoch": 5.92,
      "learning_rate": 1.672e-05,
      "loss": 0.2466,
      "step": 6656
    },
    {
      "epoch": 5.92,
      "eval_test_accuracy": 0.24329896907216494,
      "eval_test_loss": 0.24743035435676575,
      "eval_test_runtime": 233.3396,
      "eval_test_samples_per_second": 10.393,
      "eval_test_steps_per_second": 1.303,
      "step": 6656
    },
    {
      "epoch": 6.14,
      "learning_rate": 1.544e-05,
      "loss": 0.2507,
      "step": 6912
    },
    {
      "epoch": 6.14,
      "eval_test_accuracy": 0.25237113402061856,
      "eval_test_loss": 0.24656306207180023,
      "eval_test_runtime": 228.6153,
      "eval_test_samples_per_second": 10.607,
      "eval_test_steps_per_second": 1.33,
      "step": 6912
    },
    {
      "epoch": 6.37,
      "learning_rate": 1.4160000000000002e-05,
      "loss": 0.2456,
      "step": 7168
    },
    {
      "epoch": 6.37,
      "eval_test_accuracy": 0.24701030927835052,
      "eval_test_loss": 0.24853350222110748,
      "eval_test_runtime": 228.5698,
      "eval_test_samples_per_second": 10.609,
      "eval_test_steps_per_second": 1.33,
      "step": 7168
    },
    {
      "epoch": 6.6,
      "learning_rate": 1.288e-05,
      "loss": 0.241,
      "step": 7424
    },
    {
      "epoch": 6.6,
      "eval_test_accuracy": 0.2490721649484536,
      "eval_test_loss": 0.2469691038131714,
      "eval_test_runtime": 225.5172,
      "eval_test_samples_per_second": 10.753,
      "eval_test_steps_per_second": 1.348,
      "step": 7424
    },
    {
      "epoch": 6.83,
      "learning_rate": 1.16e-05,
      "loss": 0.2443,
      "step": 7680
    },
    {
      "epoch": 6.83,
      "eval_test_accuracy": 0.24824742268041236,
      "eval_test_loss": 0.24246640503406525,
      "eval_test_runtime": 226.9617,
      "eval_test_samples_per_second": 10.685,
      "eval_test_steps_per_second": 1.339,
      "step": 7680
    },
    {
      "epoch": 7.05,
      "learning_rate": 1.0320000000000001e-05,
      "loss": 0.2429,
      "step": 7936
    },
    {
      "epoch": 7.05,
      "eval_test_accuracy": 0.2515463917525773,
      "eval_test_loss": 0.24250201880931854,
      "eval_test_runtime": 228.9785,
      "eval_test_samples_per_second": 10.591,
      "eval_test_steps_per_second": 1.328,
      "step": 7936
    },
    {
      "epoch": 7.28,
      "learning_rate": 9.04e-06,
      "loss": 0.241,
      "step": 8192
    },
    {
      "epoch": 7.28,
      "eval_test_accuracy": 0.25195876288659796,
      "eval_test_loss": 0.24321790039539337,
      "eval_test_runtime": 228.3774,
      "eval_test_samples_per_second": 10.618,
      "eval_test_steps_per_second": 1.331,
      "step": 8192
    },
    {
      "epoch": 7.51,
      "learning_rate": 7.76e-06,
      "loss": 0.2418,
      "step": 8448
    },
    {
      "epoch": 7.51,
      "eval_test_accuracy": 0.2490721649484536,
      "eval_test_loss": 0.24221281707286835,
      "eval_test_runtime": 228.7396,
      "eval_test_samples_per_second": 10.602,
      "eval_test_steps_per_second": 1.329,
      "step": 8448
    },
    {
      "epoch": 7.74,
      "learning_rate": 6.48e-06,
      "loss": 0.2378,
      "step": 8704
    },
    {
      "epoch": 7.74,
      "eval_test_accuracy": 0.2465979381443299,
      "eval_test_loss": 0.24190692603588104,
      "eval_test_runtime": 228.3602,
      "eval_test_samples_per_second": 10.619,
      "eval_test_steps_per_second": 1.331,
      "step": 8704
    },
    {
      "epoch": 7.96,
      "learning_rate": 5.2e-06,
      "loss": 0.2388,
      "step": 8960
    },
    {
      "epoch": 7.96,
      "eval_test_accuracy": 0.25237113402061856,
      "eval_test_loss": 0.24081237614154816,
      "eval_test_runtime": 228.954,
      "eval_test_samples_per_second": 10.592,
      "eval_test_steps_per_second": 1.328,
      "step": 8960
    },
    {
      "epoch": 8.19,
      "learning_rate": 3.92e-06,
      "loss": 0.2304,
      "step": 9216
    },
    {
      "epoch": 8.19,
      "eval_test_accuracy": 0.257319587628866,
      "eval_test_loss": 0.24084854125976562,
      "eval_test_runtime": 227.9624,
      "eval_test_samples_per_second": 10.638,
      "eval_test_steps_per_second": 1.334,
      "step": 9216
    },
    {
      "epoch": 8.42,
      "learning_rate": 2.64e-06,
      "loss": 0.2423,
      "step": 9472
    },
    {
      "epoch": 8.42,
      "eval_test_accuracy": 0.2556701030927835,
      "eval_test_loss": 0.2404756098985672,
      "eval_test_runtime": 228.8532,
      "eval_test_samples_per_second": 10.596,
      "eval_test_steps_per_second": 1.328,
      "step": 9472
    },
    {
      "epoch": 8.65,
      "learning_rate": 1.36e-06,
      "loss": 0.2366,
      "step": 9728
    },
    {
      "epoch": 8.65,
      "eval_test_accuracy": 0.2556701030927835,
      "eval_test_loss": 0.24009202420711517,
      "eval_test_runtime": 229.0313,
      "eval_test_samples_per_second": 10.588,
      "eval_test_steps_per_second": 1.327,
      "step": 9728
    },
    {
      "epoch": 8.87,
      "learning_rate": 8e-08,
      "loss": 0.2321,
      "step": 9984
    },
    {
      "epoch": 8.87,
      "eval_test_accuracy": 0.25278350515463915,
      "eval_test_loss": 0.23984022438526154,
      "eval_test_runtime": 231.7368,
      "eval_test_samples_per_second": 10.464,
      "eval_test_steps_per_second": 1.312,
      "step": 9984
    }
  ],
  "logging_steps": 256,
  "max_steps": 10000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 256,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}