{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 11154,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.13448090371167296,
      "grad_norm": 4.317680835723877,
      "learning_rate": 1.9103460641922183e-05,
      "loss": 0.2456,
      "step": 500
    },
    {
      "epoch": 0.13448090371167296,
      "eval_accuracy": 0.9731038192576654,
      "eval_loss": 0.10059116035699844,
      "eval_runtime": 79.8153,
      "eval_samples_per_second": 93.165,
      "eval_steps_per_second": 11.652,
      "step": 500
    },
    {
      "epoch": 0.2689618074233459,
      "grad_norm": 15.100606918334961,
      "learning_rate": 1.8206921283844363e-05,
      "loss": 0.1162,
      "step": 1000
    },
    {
      "epoch": 0.2689618074233459,
      "eval_accuracy": 0.9665142549757935,
      "eval_loss": 0.17202958464622498,
      "eval_runtime": 77.2682,
      "eval_samples_per_second": 96.236,
      "eval_steps_per_second": 12.036,
      "step": 1000
    },
    {
      "epoch": 0.40344271113501884,
      "grad_norm": 0.02275875024497509,
      "learning_rate": 1.731038192576654e-05,
      "loss": 0.0779,
      "step": 1500
    },
    {
      "epoch": 0.40344271113501884,
      "eval_accuracy": 0.957235072619688,
      "eval_loss": 0.22287946939468384,
      "eval_runtime": 86.088,
      "eval_samples_per_second": 86.377,
      "eval_steps_per_second": 10.803,
      "step": 1500
    },
    {
      "epoch": 0.5379236148466918,
      "grad_norm": 0.004053408745676279,
      "learning_rate": 1.6413842567688722e-05,
      "loss": 0.082,
      "step": 2000
    },
    {
      "epoch": 0.5379236148466918,
      "eval_accuracy": 0.956831629908553,
      "eval_loss": 0.2240680307149887,
      "eval_runtime": 76.9592,
      "eval_samples_per_second": 96.623,
      "eval_steps_per_second": 12.084,
      "step": 2000
    },
    {
      "epoch": 0.6724045185583647,
      "grad_norm": 0.0030935597606003284,
      "learning_rate": 1.5517303209610903e-05,
      "loss": 0.0645,
      "step": 2500
    },
    {
      "epoch": 0.6724045185583647,
      "eval_accuracy": 0.9519903173749328,
      "eval_loss": 0.30673694610595703,
      "eval_runtime": 72.7269,
      "eval_samples_per_second": 102.246,
      "eval_steps_per_second": 12.788,
      "step": 2500
    },
    {
      "epoch": 0.8068854222700377,
      "grad_norm": 0.019974075257778168,
      "learning_rate": 1.4620763851533084e-05,
      "loss": 0.059,
      "step": 3000
    },
    {
      "epoch": 0.8068854222700377,
      "eval_accuracy": 0.9909897794513179,
      "eval_loss": 0.050069019198417664,
      "eval_runtime": 79.0838,
      "eval_samples_per_second": 94.027,
      "eval_steps_per_second": 11.76,
      "step": 3000
    },
    {
      "epoch": 0.9413663259817105,
      "grad_norm": 0.008741290308535099,
      "learning_rate": 1.3724224493455265e-05,
      "loss": 0.0582,
      "step": 3500
    },
    {
      "epoch": 0.9413663259817105,
      "eval_accuracy": 0.9911242603550295,
      "eval_loss": 0.03790881112217903,
      "eval_runtime": 77.0838,
      "eval_samples_per_second": 96.466,
      "eval_steps_per_second": 12.065,
      "step": 3500
    },
    {
      "epoch": 1.0758472296933834,
      "grad_norm": 38.54170608520508,
      "learning_rate": 1.2827685135377444e-05,
      "loss": 0.0338,
      "step": 4000
    },
    {
      "epoch": 1.0758472296933834,
      "eval_accuracy": 0.9751210328133405,
      "eval_loss": 0.15974809229373932,
      "eval_runtime": 80.1945,
      "eval_samples_per_second": 92.725,
      "eval_steps_per_second": 11.597,
      "step": 4000
    },
    {
      "epoch": 1.2103281334050564,
      "grad_norm": 29.925457000732422,
      "learning_rate": 1.1931145777299625e-05,
      "loss": 0.03,
      "step": 4500
    },
    {
      "epoch": 1.2103281334050564,
      "eval_accuracy": 0.9916621839698763,
      "eval_loss": 0.046014368534088135,
      "eval_runtime": 57.8477,
      "eval_samples_per_second": 128.544,
      "eval_steps_per_second": 16.077,
      "step": 4500
    },
    {
      "epoch": 1.3448090371167294,
      "grad_norm": 0.007805847562849522,
      "learning_rate": 1.1034606419221806e-05,
      "loss": 0.0215,
      "step": 5000
    },
    {
      "epoch": 1.3448090371167294,
      "eval_accuracy": 0.9772727272727273,
      "eval_loss": 0.15217718482017517,
      "eval_runtime": 74.0806,
      "eval_samples_per_second": 100.377,
      "eval_steps_per_second": 12.554,
      "step": 5000
    },
    {
      "epoch": 1.4792899408284024,
      "grad_norm": 0.0005576165858656168,
      "learning_rate": 1.0138067061143987e-05,
      "loss": 0.0144,
      "step": 5500
    },
    {
      "epoch": 1.4792899408284024,
      "eval_accuracy": 0.9413663259817105,
      "eval_loss": 0.4672979414463043,
      "eval_runtime": 74.077,
      "eval_samples_per_second": 100.382,
      "eval_steps_per_second": 12.555,
      "step": 5500
    },
    {
      "epoch": 1.6137708445400754,
      "grad_norm": 0.5596923232078552,
      "learning_rate": 9.241527703066166e-06,
      "loss": 0.0357,
      "step": 6000
    },
    {
      "epoch": 1.6137708445400754,
      "eval_accuracy": 0.991393222162453,
      "eval_loss": 0.05371003597974777,
      "eval_runtime": 91.3598,
      "eval_samples_per_second": 81.392,
      "eval_steps_per_second": 10.18,
      "step": 6000
    },
    {
      "epoch": 1.7482517482517483,
      "grad_norm": 0.0050299325957894325,
      "learning_rate": 8.344988344988347e-06,
      "loss": 0.0282,
      "step": 6500
    },
    {
      "epoch": 1.7482517482517483,
      "eval_accuracy": 0.9795589026358257,
      "eval_loss": 0.10718287527561188,
      "eval_runtime": 63.8039,
      "eval_samples_per_second": 116.545,
      "eval_steps_per_second": 14.576,
      "step": 6500
    },
    {
      "epoch": 1.8827326519634213,
      "grad_norm": 0.1358013153076172,
      "learning_rate": 7.448448986910526e-06,
      "loss": 0.0271,
      "step": 7000
    },
    {
      "epoch": 1.8827326519634213,
      "eval_accuracy": 0.9616729424421732,
      "eval_loss": 0.21952089667320251,
      "eval_runtime": 62.6082,
      "eval_samples_per_second": 118.77,
      "eval_steps_per_second": 14.854,
      "step": 7000
    },
    {
      "epoch": 2.0172135556750943,
      "grad_norm": 0.0005359902861528099,
      "learning_rate": 6.551909628832707e-06,
      "loss": 0.0151,
      "step": 7500
    },
    {
      "epoch": 2.0172135556750943,
      "eval_accuracy": 0.972027972027972,
      "eval_loss": 0.18190746009349823,
      "eval_runtime": 75.7313,
      "eval_samples_per_second": 98.189,
      "eval_steps_per_second": 12.28,
      "step": 7500
    },
    {
      "epoch": 2.151694459386767,
      "grad_norm": 0.0002543228620197624,
      "learning_rate": 5.655370270754886e-06,
      "loss": 0.0124,
      "step": 8000
    },
    {
      "epoch": 2.151694459386767,
      "eval_accuracy": 0.9796933835395374,
      "eval_loss": 0.12187998741865158,
      "eval_runtime": 73.9844,
      "eval_samples_per_second": 100.508,
      "eval_steps_per_second": 12.57,
      "step": 8000
    },
    {
      "epoch": 2.28617536309844,
      "grad_norm": 0.017442911863327026,
      "learning_rate": 4.758830912677067e-06,
      "loss": 0.0084,
      "step": 8500
    },
    {
      "epoch": 2.28617536309844,
      "eval_accuracy": 0.9803657880580957,
      "eval_loss": 0.133524090051651,
      "eval_runtime": 72.5895,
      "eval_samples_per_second": 102.439,
      "eval_steps_per_second": 12.812,
      "step": 8500
    },
    {
      "epoch": 2.420656266810113,
      "grad_norm": 0.00029117995291016996,
      "learning_rate": 3.862291554599247e-06,
      "loss": 0.0045,
      "step": 9000
    },
    {
      "epoch": 2.420656266810113,
      "eval_accuracy": 0.9868208714362561,
      "eval_loss": 0.084413081407547,
      "eval_runtime": 71.5341,
      "eval_samples_per_second": 103.95,
      "eval_steps_per_second": 13.001,
      "step": 9000
    },
    {
      "epoch": 2.555137170521786,
      "grad_norm": 0.0001818160671973601,
      "learning_rate": 2.9657521965214274e-06,
      "loss": 0.0027,
      "step": 9500
    },
    {
      "epoch": 2.555137170521786,
      "eval_accuracy": 0.9838622915545993,
      "eval_loss": 0.1180611401796341,
      "eval_runtime": 72.5206,
      "eval_samples_per_second": 102.536,
      "eval_steps_per_second": 12.824,
      "step": 9500
    },
    {
      "epoch": 2.6896180742334588,
      "grad_norm": 0.00021849782206118107,
      "learning_rate": 2.069212838443608e-06,
      "loss": 0.0061,
      "step": 10000
    },
    {
      "epoch": 2.6896180742334588,
      "eval_accuracy": 0.9842657342657343,
      "eval_loss": 0.11533553153276443,
      "eval_runtime": 72.894,
      "eval_samples_per_second": 102.011,
      "eval_steps_per_second": 12.758,
      "step": 10000
    },
    {
      "epoch": 2.8240989779451318,
      "grad_norm": 0.0002802180533763021,
      "learning_rate": 1.1726734803657882e-06,
      "loss": 0.0026,
      "step": 10500
    },
    {
      "epoch": 2.8240989779451318,
      "eval_accuracy": 0.9905863367401829,
      "eval_loss": 0.06259341537952423,
      "eval_runtime": 72.005,
      "eval_samples_per_second": 103.271,
      "eval_steps_per_second": 12.916,
      "step": 10500
    },
    {
      "epoch": 2.9585798816568047,
      "grad_norm": 0.00011676761641865596,
      "learning_rate": 2.7613412228796843e-07,
      "loss": 0.0008,
      "step": 11000
    },
    {
      "epoch": 2.9585798816568047,
      "eval_accuracy": 0.983593329747176,
      "eval_loss": 0.13157618045806885,
      "eval_runtime": 95.8085,
      "eval_samples_per_second": 77.613,
      "eval_steps_per_second": 9.707,
      "step": 11000
    }
  ],
  "logging_steps": 500,
  "max_steps": 11154,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1.8301403760969e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}