{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 59.881188118811885,
  "eval_steps": 1000,
  "global_step": 11340,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.53,
      "learning_rate": 4.956349206349207e-05,
      "loss": 0.0798,
      "step": 100
    },
    {
      "epoch": 1.06,
      "learning_rate": 4.912257495590829e-05,
      "loss": 0.0816,
      "step": 200
    },
    {
      "epoch": 1.58,
      "learning_rate": 4.868165784832452e-05,
      "loss": 0.0761,
      "step": 300
    },
    {
      "epoch": 2.11,
      "learning_rate": 4.824514991181658e-05,
      "loss": 0.0723,
      "step": 400
    },
    {
      "epoch": 2.64,
      "learning_rate": 4.7804232804232806e-05,
      "loss": 0.0601,
      "step": 500
    },
    {
      "epoch": 3.17,
      "learning_rate": 4.736331569664903e-05,
      "loss": 0.0593,
      "step": 600
    },
    {
      "epoch": 3.7,
      "learning_rate": 4.692239858906526e-05,
      "loss": 0.0692,
      "step": 700
    },
    {
      "epoch": 4.22,
      "learning_rate": 4.648148148148148e-05,
      "loss": 0.0587,
      "step": 800
    },
    {
      "epoch": 4.75,
      "learning_rate": 4.604056437389771e-05,
      "loss": 0.0598,
      "step": 900
    },
    {
      "epoch": 5.28,
      "learning_rate": 4.559964726631393e-05,
      "loss": 0.0559,
      "step": 1000
    },
    {
      "epoch": 5.28,
      "eval_accuracy": 0.9191418886184692,
      "eval_loss": 0.3096904754638672,
      "eval_runtime": 51.3243,
      "eval_samples_per_second": 47.229,
      "eval_steps_per_second": 5.904,
      "step": 1000
    },
    {
      "epoch": 5.81,
      "learning_rate": 4.515873015873016e-05,
      "loss": 0.0552,
      "step": 1100
    },
    {
      "epoch": 6.34,
      "learning_rate": 4.471781305114639e-05,
      "loss": 0.0519,
      "step": 1200
    },
    {
      "epoch": 6.86,
      "learning_rate": 4.428130511463845e-05,
      "loss": 0.0536,
      "step": 1300
    },
    {
      "epoch": 7.39,
      "learning_rate": 4.3840388007054675e-05,
      "loss": 0.0438,
      "step": 1400
    },
    {
      "epoch": 7.92,
      "learning_rate": 4.33994708994709e-05,
      "loss": 0.0436,
      "step": 1500
    },
    {
      "epoch": 8.45,
      "learning_rate": 4.295855379188713e-05,
      "loss": 0.0447,
      "step": 1600
    },
    {
      "epoch": 8.98,
      "learning_rate": 4.2517636684303355e-05,
      "loss": 0.0535,
      "step": 1700
    },
    {
      "epoch": 9.5,
      "learning_rate": 4.207671957671958e-05,
      "loss": 0.0467,
      "step": 1800
    },
    {
      "epoch": 10.03,
      "learning_rate": 4.16358024691358e-05,
      "loss": 0.0557,
      "step": 1900
    },
    {
      "epoch": 10.56,
      "learning_rate": 4.1194885361552036e-05,
      "loss": 0.047,
      "step": 2000
    },
    {
      "epoch": 10.56,
      "eval_accuracy": 0.9191418886184692,
      "eval_loss": 0.34823155403137207,
      "eval_runtime": 51.4649,
      "eval_samples_per_second": 47.1,
      "eval_steps_per_second": 5.888,
      "step": 2000
    },
    {
      "epoch": 11.09,
      "learning_rate": 4.0753968253968256e-05,
      "loss": 0.0397,
      "step": 2100
    },
    {
      "epoch": 11.62,
      "learning_rate": 4.031305114638448e-05,
      "loss": 0.0475,
      "step": 2200
    },
    {
      "epoch": 12.15,
      "learning_rate": 3.987213403880071e-05,
      "loss": 0.0478,
      "step": 2300
    },
    {
      "epoch": 12.67,
      "learning_rate": 3.9431216931216936e-05,
      "loss": 0.0421,
      "step": 2400
    },
    {
      "epoch": 13.2,
      "learning_rate": 3.8994708994709e-05,
      "loss": 0.0408,
      "step": 2500
    },
    {
      "epoch": 13.73,
      "learning_rate": 3.8553791887125224e-05,
      "loss": 0.0368,
      "step": 2600
    },
    {
      "epoch": 14.26,
      "learning_rate": 3.8112874779541445e-05,
      "loss": 0.0379,
      "step": 2700
    },
    {
      "epoch": 14.79,
      "learning_rate": 3.767195767195768e-05,
      "loss": 0.052,
      "step": 2800
    },
    {
      "epoch": 15.31,
      "learning_rate": 3.72310405643739e-05,
      "loss": 0.0424,
      "step": 2900
    },
    {
      "epoch": 15.84,
      "learning_rate": 3.6790123456790125e-05,
      "loss": 0.0402,
      "step": 3000
    },
    {
      "epoch": 15.84,
      "eval_accuracy": 0.9080032706260681,
      "eval_loss": 0.3889801502227783,
      "eval_runtime": 50.7451,
      "eval_samples_per_second": 47.768,
      "eval_steps_per_second": 5.971,
      "step": 3000
    },
    {
      "epoch": 16.37,
      "learning_rate": 3.634920634920635e-05,
      "loss": 0.0417,
      "step": 3100
    },
    {
      "epoch": 16.9,
      "learning_rate": 3.590828924162258e-05,
      "loss": 0.0419,
      "step": 3200
    },
    {
      "epoch": 17.43,
      "learning_rate": 3.54673721340388e-05,
      "loss": 0.0439,
      "step": 3300
    },
    {
      "epoch": 17.95,
      "learning_rate": 3.502645502645503e-05,
      "loss": 0.0446,
      "step": 3400
    },
    {
      "epoch": 18.48,
      "learning_rate": 3.458553791887125e-05,
      "loss": 0.0339,
      "step": 3500
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.414462081128748e-05,
      "loss": 0.0301,
      "step": 3600
    },
    {
      "epoch": 19.54,
      "learning_rate": 3.3703703703703706e-05,
      "loss": 0.0277,
      "step": 3700
    },
    {
      "epoch": 20.07,
      "learning_rate": 3.326278659611993e-05,
      "loss": 0.0333,
      "step": 3800
    },
    {
      "epoch": 20.59,
      "learning_rate": 3.282186948853615e-05,
      "loss": 0.0298,
      "step": 3900
    },
    {
      "epoch": 21.12,
      "learning_rate": 3.2380952380952386e-05,
      "loss": 0.0328,
      "step": 4000
    },
    {
      "epoch": 21.12,
      "eval_accuracy": 0.9150164723396301,
      "eval_loss": 0.37457939982414246,
      "eval_runtime": 50.3165,
      "eval_samples_per_second": 48.175,
      "eval_steps_per_second": 6.022,
      "step": 4000
    },
    {
      "epoch": 21.65,
      "learning_rate": 3.1940035273368606e-05,
      "loss": 0.0298,
      "step": 4100
    },
    {
      "epoch": 22.18,
      "learning_rate": 3.149911816578483e-05,
      "loss": 0.0314,
      "step": 4200
    },
    {
      "epoch": 22.71,
      "learning_rate": 3.105820105820106e-05,
      "loss": 0.0253,
      "step": 4300
    },
    {
      "epoch": 23.23,
      "learning_rate": 3.061728395061729e-05,
      "loss": 0.0339,
      "step": 4400
    },
    {
      "epoch": 23.76,
      "learning_rate": 3.017636684303351e-05,
      "loss": 0.0266,
      "step": 4500
    },
    {
      "epoch": 24.29,
      "learning_rate": 2.973544973544974e-05,
      "loss": 0.0361,
      "step": 4600
    },
    {
      "epoch": 24.82,
      "learning_rate": 2.929453262786596e-05,
      "loss": 0.0305,
      "step": 4700
    },
    {
      "epoch": 25.35,
      "learning_rate": 2.885361552028219e-05,
      "loss": 0.0294,
      "step": 4800
    },
    {
      "epoch": 25.87,
      "learning_rate": 2.8412698412698414e-05,
      "loss": 0.0339,
      "step": 4900
    },
    {
      "epoch": 26.4,
      "learning_rate": 2.797178130511464e-05,
      "loss": 0.0189,
      "step": 5000
    },
    {
      "epoch": 26.4,
      "eval_accuracy": 0.9113036394119263,
      "eval_loss": 0.42735978960990906,
      "eval_runtime": 49.4145,
      "eval_samples_per_second": 49.054,
      "eval_steps_per_second": 6.132,
      "step": 5000
    },
    {
      "epoch": 26.93,
      "learning_rate": 2.7530864197530864e-05,
      "loss": 0.0285,
      "step": 5100
    },
    {
      "epoch": 27.46,
      "learning_rate": 2.7089947089947094e-05,
      "loss": 0.0307,
      "step": 5200
    },
    {
      "epoch": 27.99,
      "learning_rate": 2.6649029982363318e-05,
      "loss": 0.0269,
      "step": 5300
    },
    {
      "epoch": 28.51,
      "learning_rate": 2.6208112874779544e-05,
      "loss": 0.0292,
      "step": 5400
    },
    {
      "epoch": 29.04,
      "learning_rate": 2.5767195767195768e-05,
      "loss": 0.032,
      "step": 5500
    },
    {
      "epoch": 29.57,
      "learning_rate": 2.5326278659611995e-05,
      "loss": 0.0297,
      "step": 5600
    },
    {
      "epoch": 30.1,
      "learning_rate": 2.4885361552028218e-05,
      "loss": 0.027,
      "step": 5700
    },
    {
      "epoch": 30.63,
      "learning_rate": 2.4444444444444445e-05,
      "loss": 0.0258,
      "step": 5800
    },
    {
      "epoch": 31.16,
      "learning_rate": 2.4003527336860672e-05,
      "loss": 0.0237,
      "step": 5900
    },
    {
      "epoch": 31.68,
      "learning_rate": 2.3562610229276895e-05,
      "loss": 0.0187,
      "step": 6000
    },
    {
      "epoch": 31.68,
      "eval_accuracy": 0.9100660085678101,
      "eval_loss": 0.4131234884262085,
      "eval_runtime": 49.5018,
      "eval_samples_per_second": 48.968,
      "eval_steps_per_second": 6.121,
      "step": 6000
    },
    {
      "epoch": 32.21,
      "learning_rate": 2.3121693121693122e-05,
      "loss": 0.0255,
      "step": 6100
    },
    {
      "epoch": 32.74,
      "learning_rate": 2.268077601410935e-05,
      "loss": 0.0244,
      "step": 6200
    },
    {
      "epoch": 33.27,
      "learning_rate": 2.2239858906525572e-05,
      "loss": 0.0273,
      "step": 6300
    },
    {
      "epoch": 33.8,
      "learning_rate": 2.17989417989418e-05,
      "loss": 0.0214,
      "step": 6400
    },
    {
      "epoch": 34.32,
      "learning_rate": 2.1358024691358026e-05,
      "loss": 0.0265,
      "step": 6500
    },
    {
      "epoch": 34.85,
      "learning_rate": 2.091710758377425e-05,
      "loss": 0.0276,
      "step": 6600
    },
    {
      "epoch": 35.38,
      "learning_rate": 2.0476190476190476e-05,
      "loss": 0.0153,
      "step": 6700
    },
    {
      "epoch": 35.91,
      "learning_rate": 2.0035273368606703e-05,
      "loss": 0.0246,
      "step": 6800
    },
    {
      "epoch": 36.44,
      "learning_rate": 1.959435626102293e-05,
      "loss": 0.0266,
      "step": 6900
    },
    {
      "epoch": 36.96,
      "learning_rate": 1.9153439153439153e-05,
      "loss": 0.0203,
      "step": 7000
    },
    {
      "epoch": 36.96,
      "eval_accuracy": 0.9236798882484436,
      "eval_loss": 0.3643423020839691,
      "eval_runtime": 49.3507,
      "eval_samples_per_second": 49.118,
      "eval_steps_per_second": 6.14,
      "step": 7000
    },
    {
      "epoch": 37.49,
      "learning_rate": 1.871252204585538e-05,
      "loss": 0.0225,
      "step": 7100
    },
    {
      "epoch": 38.02,
      "learning_rate": 1.8271604938271607e-05,
      "loss": 0.0296,
      "step": 7200
    },
    {
      "epoch": 38.55,
      "learning_rate": 1.783068783068783e-05,
      "loss": 0.0181,
      "step": 7300
    },
    {
      "epoch": 39.08,
      "learning_rate": 1.7389770723104057e-05,
      "loss": 0.0184,
      "step": 7400
    },
    {
      "epoch": 39.6,
      "learning_rate": 1.6948853615520284e-05,
      "loss": 0.0191,
      "step": 7500
    },
    {
      "epoch": 40.13,
      "learning_rate": 1.6507936507936507e-05,
      "loss": 0.0224,
      "step": 7600
    },
    {
      "epoch": 40.66,
      "learning_rate": 1.6067019400352734e-05,
      "loss": 0.0161,
      "step": 7700
    },
    {
      "epoch": 41.19,
      "learning_rate": 1.562610229276896e-05,
      "loss": 0.0211,
      "step": 7800
    },
    {
      "epoch": 41.72,
      "learning_rate": 1.5185185185185186e-05,
      "loss": 0.0165,
      "step": 7900
    },
    {
      "epoch": 42.24,
      "learning_rate": 1.4744268077601411e-05,
      "loss": 0.0147,
      "step": 8000
    },
    {
      "epoch": 42.24,
      "eval_accuracy": 0.9294554591178894,
      "eval_loss": 0.3574332892894745,
      "eval_runtime": 49.7962,
      "eval_samples_per_second": 48.678,
      "eval_steps_per_second": 6.085,
      "step": 8000
    },
    {
      "epoch": 42.77,
      "learning_rate": 1.4303350970017638e-05,
      "loss": 0.0235,
      "step": 8100
    },
    {
      "epoch": 43.3,
      "learning_rate": 1.3862433862433863e-05,
      "loss": 0.0207,
      "step": 8200
    },
    {
      "epoch": 43.83,
      "learning_rate": 1.3421516754850088e-05,
      "loss": 0.0139,
      "step": 8300
    },
    {
      "epoch": 44.36,
      "learning_rate": 1.2980599647266315e-05,
      "loss": 0.0168,
      "step": 8400
    },
    {
      "epoch": 44.88,
      "learning_rate": 1.253968253968254e-05,
      "loss": 0.0146,
      "step": 8500
    },
    {
      "epoch": 45.41,
      "learning_rate": 1.2098765432098767e-05,
      "loss": 0.0149,
      "step": 8600
    },
    {
      "epoch": 45.94,
      "learning_rate": 1.1657848324514992e-05,
      "loss": 0.0155,
      "step": 8700
    },
    {
      "epoch": 46.47,
      "learning_rate": 1.1221340388007055e-05,
      "loss": 0.0196,
      "step": 8800
    },
    {
      "epoch": 47.0,
      "learning_rate": 1.0780423280423282e-05,
      "loss": 0.0229,
      "step": 8900
    },
    {
      "epoch": 47.52,
      "learning_rate": 1.0339506172839507e-05,
      "loss": 0.0148,
      "step": 9000
    },
    {
      "epoch": 47.52,
      "eval_accuracy": 0.9220296740531921,
      "eval_loss": 0.36532989144325256,
      "eval_runtime": 50.0277,
      "eval_samples_per_second": 48.453,
      "eval_steps_per_second": 6.057,
      "step": 9000
    },
    {
      "epoch": 48.05,
      "learning_rate": 9.898589065255732e-06,
      "loss": 0.0133,
      "step": 9100
    },
    {
      "epoch": 48.58,
      "learning_rate": 9.457671957671959e-06,
      "loss": 0.0131,
      "step": 9200
    },
    {
      "epoch": 49.11,
      "learning_rate": 9.016754850088184e-06,
      "loss": 0.0121,
      "step": 9300
    },
    {
      "epoch": 49.64,
      "learning_rate": 8.575837742504409e-06,
      "loss": 0.0168,
      "step": 9400
    },
    {
      "epoch": 50.17,
      "learning_rate": 8.134920634920636e-06,
      "loss": 0.0148,
      "step": 9500
    },
    {
      "epoch": 50.69,
      "learning_rate": 7.694003527336861e-06,
      "loss": 0.0129,
      "step": 9600
    },
    {
      "epoch": 51.22,
      "learning_rate": 7.253086419753087e-06,
      "loss": 0.012,
      "step": 9700
    },
    {
      "epoch": 51.75,
      "learning_rate": 6.812169312169313e-06,
      "loss": 0.0112,
      "step": 9800
    },
    {
      "epoch": 52.28,
      "learning_rate": 6.371252204585539e-06,
      "loss": 0.0193,
      "step": 9900
    },
    {
      "epoch": 52.81,
      "learning_rate": 5.930335097001764e-06,
      "loss": 0.0137,
      "step": 10000
    },
    {
      "epoch": 52.81,
      "eval_accuracy": 0.9352310299873352,
      "eval_loss": 0.3257134258747101,
      "eval_runtime": 50.886,
      "eval_samples_per_second": 47.636,
      "eval_steps_per_second": 5.954,
      "step": 10000
    },
    {
      "epoch": 53.33,
      "learning_rate": 5.489417989417989e-06,
      "loss": 0.0171,
      "step": 10100
    },
    {
      "epoch": 53.86,
      "learning_rate": 5.048500881834215e-06,
      "loss": 0.0169,
      "step": 10200
    },
    {
      "epoch": 54.39,
      "learning_rate": 4.611992945326279e-06,
      "loss": 0.0136,
      "step": 10300
    },
    {
      "epoch": 54.92,
      "learning_rate": 4.171075837742505e-06,
      "loss": 0.0152,
      "step": 10400
    },
    {
      "epoch": 55.45,
      "learning_rate": 3.7301587301587305e-06,
      "loss": 0.015,
      "step": 10500
    },
    {
      "epoch": 55.97,
      "learning_rate": 3.289241622574956e-06,
      "loss": 0.0136,
      "step": 10600
    },
    {
      "epoch": 56.5,
      "learning_rate": 2.848324514991182e-06,
      "loss": 0.0156,
      "step": 10700
    },
    {
      "epoch": 57.03,
      "learning_rate": 2.4074074074074075e-06,
      "loss": 0.0137,
      "step": 10800
    },
    {
      "epoch": 57.56,
      "learning_rate": 1.9664902998236335e-06,
      "loss": 0.0147,
      "step": 10900
    },
    {
      "epoch": 58.09,
      "learning_rate": 1.525573192239859e-06,
      "loss": 0.0174,
      "step": 11000
    },
    {
      "epoch": 58.09,
      "eval_accuracy": 0.933993399143219,
      "eval_loss": 0.30968689918518066,
      "eval_runtime": 51.0931,
      "eval_samples_per_second": 47.443,
      "eval_steps_per_second": 5.93,
      "step": 11000
    },
    {
      "epoch": 58.61,
      "learning_rate": 1.0846560846560847e-06,
      "loss": 0.0163,
      "step": 11100
    },
    {
      "epoch": 59.14,
      "learning_rate": 6.437389770723105e-07,
      "loss": 0.0096,
      "step": 11200
    },
    {
      "epoch": 59.67,
      "learning_rate": 2.0282186948853617e-07,
      "loss": 0.0121,
      "step": 11300
    },
    {
      "epoch": 59.88,
      "step": 11340,
      "total_flos": 2.20514636224512e+19,
      "train_loss": 0.030625741817122836,
      "train_runtime": 25244.9182,
      "train_samples_per_second": 28.806,
      "train_steps_per_second": 0.449
    },
    {
      "epoch": 59.88,
      "eval_accuracy": 0.9323432445526123,
      "eval_loss": 0.3117374777793884,
      "eval_runtime": 50.8983,
      "eval_samples_per_second": 47.624,
      "eval_steps_per_second": 5.953,
      "step": 11340
    }
  ],
  "logging_steps": 100,
  "max_steps": 11340,
  "num_train_epochs": 60,
  "save_steps": 2000,
  "total_flos": 2.20514636224512e+19,
  "trial_name": null,
  "trial_params": null
}