{
  "best_metric": 0.9787709497206704,
  "best_model_checkpoint": "resnet-50-LongSleeveCleanedData/checkpoint-2733",
  "epoch": 19.88083416087388,
  "global_step": 2860,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07,
      "learning_rate": 1.7241379310344828e-05,
      "loss": 1.1031,
      "step": 10
    },
    {
      "epoch": 0.14,
      "learning_rate": 3.4482758620689657e-05,
      "loss": 1.102,
      "step": 20
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.998233839632639e-05,
      "loss": 1.0994,
      "step": 30
    },
    {
      "epoch": 0.28,
      "learning_rate": 4.980572235959025e-05,
      "loss": 1.0905,
      "step": 40
    },
    {
      "epoch": 0.35,
      "learning_rate": 4.962910632285412e-05,
      "loss": 1.0828,
      "step": 50
    },
    {
      "epoch": 0.42,
      "learning_rate": 4.945249028611798e-05,
      "loss": 1.0761,
      "step": 60
    },
    {
      "epoch": 0.49,
      "learning_rate": 4.927587424938185e-05,
      "loss": 1.0646,
      "step": 70
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.909925821264571e-05,
      "loss": 1.0606,
      "step": 80
    },
    {
      "epoch": 0.63,
      "learning_rate": 4.892264217590958e-05,
      "loss": 1.0502,
      "step": 90
    },
    {
      "epoch": 0.7,
      "learning_rate": 4.874602613917344e-05,
      "loss": 1.0411,
      "step": 100
    },
    {
      "epoch": 0.76,
      "learning_rate": 4.85694101024373e-05,
      "loss": 1.0358,
      "step": 110
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.8392794065701166e-05,
      "loss": 1.023,
      "step": 120
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.821617802896503e-05,
      "loss": 1.0164,
      "step": 130
    },
    {
      "epoch": 0.97,
      "learning_rate": 4.8039561992228896e-05,
      "loss": 0.9906,
      "step": 140
    },
    {
      "epoch": 0.99,
      "eval_accuracy": 0.6134078212290502,
      "eval_loss": 1.0393954515457153,
      "eval_runtime": 725.2732,
      "eval_samples_per_second": 1.234,
      "eval_steps_per_second": 0.154,
      "step": 143
    },
    {
      "epoch": 1.04,
      "learning_rate": 4.786294595549276e-05,
      "loss": 0.9853,
      "step": 150
    },
    {
      "epoch": 1.11,
      "learning_rate": 4.768632991875663e-05,
      "loss": 0.9606,
      "step": 160
    },
    {
      "epoch": 1.18,
      "learning_rate": 4.750971388202049e-05,
      "loss": 0.9562,
      "step": 170
    },
    {
      "epoch": 1.25,
      "learning_rate": 4.7333097845284354e-05,
      "loss": 0.9219,
      "step": 180
    },
    {
      "epoch": 1.32,
      "learning_rate": 4.715648180854822e-05,
      "loss": 0.9033,
      "step": 190
    },
    {
      "epoch": 1.39,
      "learning_rate": 4.697986577181208e-05,
      "loss": 0.8942,
      "step": 200
    },
    {
      "epoch": 1.46,
      "learning_rate": 4.680324973507595e-05,
      "loss": 0.8555,
      "step": 210
    },
    {
      "epoch": 1.53,
      "learning_rate": 4.662663369833981e-05,
      "loss": 0.8475,
      "step": 220
    },
    {
      "epoch": 1.6,
      "learning_rate": 4.645001766160367e-05,
      "loss": 0.8118,
      "step": 230
    },
    {
      "epoch": 1.67,
      "learning_rate": 4.627340162486754e-05,
      "loss": 0.82,
      "step": 240
    },
    {
      "epoch": 1.74,
      "learning_rate": 4.6096785588131406e-05,
      "loss": 0.8028,
      "step": 250
    },
    {
      "epoch": 1.81,
      "learning_rate": 4.592016955139527e-05,
      "loss": 0.7645,
      "step": 260
    },
    {
      "epoch": 1.88,
      "learning_rate": 4.5743553514659135e-05,
      "loss": 0.7472,
      "step": 270
    },
    {
      "epoch": 1.95,
      "learning_rate": 4.5566937477923e-05,
      "loss": 0.7315,
      "step": 280
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7631284916201118,
      "eval_loss": 0.6789541244506836,
      "eval_runtime": 20.8953,
      "eval_samples_per_second": 42.833,
      "eval_steps_per_second": 5.36,
      "step": 287
    },
    {
      "epoch": 2.02,
      "learning_rate": 4.5390321441186864e-05,
      "loss": 0.757,
      "step": 290
    },
    {
      "epoch": 2.09,
      "learning_rate": 4.521370540445072e-05,
      "loss": 0.7235,
      "step": 300
    },
    {
      "epoch": 2.15,
      "learning_rate": 4.503708936771459e-05,
      "loss": 0.6899,
      "step": 310
    },
    {
      "epoch": 2.22,
      "learning_rate": 4.486047333097845e-05,
      "loss": 0.6831,
      "step": 320
    },
    {
      "epoch": 2.29,
      "learning_rate": 4.468385729424232e-05,
      "loss": 0.6265,
      "step": 330
    },
    {
      "epoch": 2.36,
      "learning_rate": 4.450724125750618e-05,
      "loss": 0.6394,
      "step": 340
    },
    {
      "epoch": 2.43,
      "learning_rate": 4.433062522077005e-05,
      "loss": 0.64,
      "step": 350
    },
    {
      "epoch": 2.5,
      "learning_rate": 4.415400918403391e-05,
      "loss": 0.6319,
      "step": 360
    },
    {
      "epoch": 2.57,
      "learning_rate": 4.3977393147297774e-05,
      "loss": 0.5885,
      "step": 370
    },
    {
      "epoch": 2.64,
      "learning_rate": 4.3800777110561645e-05,
      "loss": 0.6172,
      "step": 380
    },
    {
      "epoch": 2.71,
      "learning_rate": 4.36241610738255e-05,
      "loss": 0.6075,
      "step": 390
    },
    {
      "epoch": 2.78,
      "learning_rate": 4.3447545037089374e-05,
      "loss": 0.5672,
      "step": 400
    },
    {
      "epoch": 2.85,
      "learning_rate": 4.327092900035323e-05,
      "loss": 0.5717,
      "step": 410
    },
    {
      "epoch": 2.92,
      "learning_rate": 4.30943129636171e-05,
      "loss": 0.5241,
      "step": 420
    },
    {
      "epoch": 2.99,
      "learning_rate": 4.291769692688096e-05,
      "loss": 0.559,
      "step": 430
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8547486033519553,
      "eval_loss": 0.47351497411727905,
      "eval_runtime": 19.826,
      "eval_samples_per_second": 45.143,
      "eval_steps_per_second": 5.649,
      "step": 431
    },
    {
      "epoch": 3.06,
      "learning_rate": 4.2741080890144825e-05,
      "loss": 0.5018,
      "step": 440
    },
    {
      "epoch": 3.13,
      "learning_rate": 4.256446485340869e-05,
      "loss": 0.5239,
      "step": 450
    },
    {
      "epoch": 3.2,
      "learning_rate": 4.2387848816672555e-05,
      "loss": 0.5113,
      "step": 460
    },
    {
      "epoch": 3.27,
      "learning_rate": 4.221123277993642e-05,
      "loss": 0.5464,
      "step": 470
    },
    {
      "epoch": 3.34,
      "learning_rate": 4.2034616743200284e-05,
      "loss": 0.5247,
      "step": 480
    },
    {
      "epoch": 3.41,
      "learning_rate": 4.185800070646415e-05,
      "loss": 0.5309,
      "step": 490
    },
    {
      "epoch": 3.48,
      "learning_rate": 4.168138466972801e-05,
      "loss": 0.4472,
      "step": 500
    },
    {
      "epoch": 3.55,
      "learning_rate": 4.150476863299188e-05,
      "loss": 0.4852,
      "step": 510
    },
    {
      "epoch": 3.61,
      "learning_rate": 4.132815259625574e-05,
      "loss": 0.4296,
      "step": 520
    },
    {
      "epoch": 3.68,
      "learning_rate": 4.1151536559519606e-05,
      "loss": 0.4878,
      "step": 530
    },
    {
      "epoch": 3.75,
      "learning_rate": 4.097492052278347e-05,
      "loss": 0.4741,
      "step": 540
    },
    {
      "epoch": 3.82,
      "learning_rate": 4.0798304486047335e-05,
      "loss": 0.4571,
      "step": 550
    },
    {
      "epoch": 3.89,
      "learning_rate": 4.06216884493112e-05,
      "loss": 0.4514,
      "step": 560
    },
    {
      "epoch": 3.96,
      "learning_rate": 4.0445072412575065e-05,
      "loss": 0.4905,
      "step": 570
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8983240223463688,
      "eval_loss": 0.3148014545440674,
      "eval_runtime": 21.1724,
      "eval_samples_per_second": 42.272,
      "eval_steps_per_second": 5.29,
      "step": 575
    },
    {
      "epoch": 4.03,
      "learning_rate": 4.026845637583892e-05,
      "loss": 0.4468,
      "step": 580
    },
    {
      "epoch": 4.1,
      "learning_rate": 4.0091840339102794e-05,
      "loss": 0.4162,
      "step": 590
    },
    {
      "epoch": 4.17,
      "learning_rate": 3.991522430236666e-05,
      "loss": 0.4565,
      "step": 600
    },
    {
      "epoch": 4.24,
      "learning_rate": 3.973860826563052e-05,
      "loss": 0.4272,
      "step": 610
    },
    {
      "epoch": 4.31,
      "learning_rate": 3.956199222889439e-05,
      "loss": 0.3927,
      "step": 620
    },
    {
      "epoch": 4.38,
      "learning_rate": 3.938537619215825e-05,
      "loss": 0.4419,
      "step": 630
    },
    {
      "epoch": 4.45,
      "learning_rate": 3.9208760155422116e-05,
      "loss": 0.4796,
      "step": 640
    },
    {
      "epoch": 4.52,
      "learning_rate": 3.9032144118685974e-05,
      "loss": 0.3772,
      "step": 650
    },
    {
      "epoch": 4.59,
      "learning_rate": 3.8855528081949846e-05,
      "loss": 0.4285,
      "step": 660
    },
    {
      "epoch": 4.66,
      "learning_rate": 3.86789120452137e-05,
      "loss": 0.4611,
      "step": 670
    },
    {
      "epoch": 4.73,
      "learning_rate": 3.8502296008477575e-05,
      "loss": 0.4045,
      "step": 680
    },
    {
      "epoch": 4.8,
      "learning_rate": 3.832567997174143e-05,
      "loss": 0.4066,
      "step": 690
    },
    {
      "epoch": 4.87,
      "learning_rate": 3.8149063935005304e-05,
      "loss": 0.4457,
      "step": 700
    },
    {
      "epoch": 4.94,
      "learning_rate": 3.797244789826916e-05,
      "loss": 0.3465,
      "step": 710
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9363128491620112,
      "eval_loss": 0.22248676419258118,
      "eval_runtime": 20.5776,
      "eval_samples_per_second": 43.494,
      "eval_steps_per_second": 5.443,
      "step": 719
    },
    {
      "epoch": 5.0,
      "learning_rate": 3.7795831861533026e-05,
      "loss": 0.3614,
      "step": 720
    },
    {
      "epoch": 5.07,
      "learning_rate": 3.761921582479689e-05,
      "loss": 0.3649,
      "step": 730
    },
    {
      "epoch": 5.14,
      "learning_rate": 3.7442599788060755e-05,
      "loss": 0.3456,
      "step": 740
    },
    {
      "epoch": 5.21,
      "learning_rate": 3.7265983751324627e-05,
      "loss": 0.3546,
      "step": 750
    },
    {
      "epoch": 5.28,
      "learning_rate": 3.7089367714588484e-05,
      "loss": 0.3802,
      "step": 760
    },
    {
      "epoch": 5.35,
      "learning_rate": 3.6912751677852356e-05,
      "loss": 0.3694,
      "step": 770
    },
    {
      "epoch": 5.42,
      "learning_rate": 3.6736135641116213e-05,
      "loss": 0.3551,
      "step": 780
    },
    {
      "epoch": 5.49,
      "learning_rate": 3.655951960438008e-05,
      "loss": 0.3756,
      "step": 790
    },
    {
      "epoch": 5.56,
      "learning_rate": 3.638290356764394e-05,
      "loss": 0.4003,
      "step": 800
    },
    {
      "epoch": 5.63,
      "learning_rate": 3.620628753090781e-05,
      "loss": 0.4294,
      "step": 810
    },
    {
      "epoch": 5.7,
      "learning_rate": 3.602967149417167e-05,
      "loss": 0.3199,
      "step": 820
    },
    {
      "epoch": 5.77,
      "learning_rate": 3.5853055457435536e-05,
      "loss": 0.3732,
      "step": 830
    },
    {
      "epoch": 5.84,
      "learning_rate": 3.56764394206994e-05,
      "loss": 0.3084,
      "step": 840
    },
    {
      "epoch": 5.91,
      "learning_rate": 3.5499823383963265e-05,
      "loss": 0.3275,
      "step": 850
    },
    {
      "epoch": 5.98,
      "learning_rate": 3.532320734722713e-05,
      "loss": 0.3372,
      "step": 860
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9486033519553073,
      "eval_loss": 0.18386144936084747,
      "eval_runtime": 19.9501,
      "eval_samples_per_second": 44.862,
      "eval_steps_per_second": 5.614,
      "step": 863
    },
    {
      "epoch": 6.05,
      "learning_rate": 3.5146591310490994e-05,
      "loss": 0.3084,
      "step": 870
    },
    {
      "epoch": 6.12,
      "learning_rate": 3.496997527375486e-05,
      "loss": 0.3171,
      "step": 880
    },
    {
      "epoch": 6.19,
      "learning_rate": 3.4793359237018724e-05,
      "loss": 0.3379,
      "step": 890
    },
    {
      "epoch": 6.26,
      "learning_rate": 3.461674320028259e-05,
      "loss": 0.3426,
      "step": 900
    },
    {
      "epoch": 6.33,
      "learning_rate": 3.444012716354645e-05,
      "loss": 0.3396,
      "step": 910
    },
    {
      "epoch": 6.4,
      "learning_rate": 3.426351112681032e-05,
      "loss": 0.3184,
      "step": 920
    },
    {
      "epoch": 6.46,
      "learning_rate": 3.4086895090074175e-05,
      "loss": 0.3158,
      "step": 930
    },
    {
      "epoch": 6.53,
      "learning_rate": 3.3910279053338046e-05,
      "loss": 0.3197,
      "step": 940
    },
    {
      "epoch": 6.6,
      "learning_rate": 3.3733663016601904e-05,
      "loss": 0.3312,
      "step": 950
    },
    {
      "epoch": 6.67,
      "learning_rate": 3.3557046979865775e-05,
      "loss": 0.3256,
      "step": 960
    },
    {
      "epoch": 6.74,
      "learning_rate": 3.338043094312964e-05,
      "loss": 0.3237,
      "step": 970
    },
    {
      "epoch": 6.81,
      "learning_rate": 3.3203814906393505e-05,
      "loss": 0.3162,
      "step": 980
    },
    {
      "epoch": 6.88,
      "learning_rate": 3.302719886965737e-05,
      "loss": 0.2939,
      "step": 990
    },
    {
      "epoch": 6.95,
      "learning_rate": 3.285058283292123e-05,
      "loss": 0.3349,
      "step": 1000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.958659217877095,
      "eval_loss": 0.16170552372932434,
      "eval_runtime": 21.263,
      "eval_samples_per_second": 42.092,
      "eval_steps_per_second": 5.267,
      "step": 1007
    },
    {
      "epoch": 7.02,
      "learning_rate": 3.26739667961851e-05,
      "loss": 0.3104,
      "step": 1010
    },
    {
      "epoch": 7.09,
      "learning_rate": 3.2497350759448956e-05,
      "loss": 0.3108,
      "step": 1020
    },
    {
      "epoch": 7.16,
      "learning_rate": 3.232073472271283e-05,
      "loss": 0.2673,
      "step": 1030
    },
    {
      "epoch": 7.23,
      "learning_rate": 3.2144118685976685e-05,
      "loss": 0.3208,
      "step": 1040
    },
    {
      "epoch": 7.3,
      "learning_rate": 3.1967502649240556e-05,
      "loss": 0.2951,
      "step": 1050
    },
    {
      "epoch": 7.37,
      "learning_rate": 3.1790886612504414e-05,
      "loss": 0.2816,
      "step": 1060
    },
    {
      "epoch": 7.44,
      "learning_rate": 3.161427057576828e-05,
      "loss": 0.2879,
      "step": 1070
    },
    {
      "epoch": 7.51,
      "learning_rate": 3.143765453903214e-05,
      "loss": 0.2891,
      "step": 1080
    },
    {
      "epoch": 7.58,
      "learning_rate": 3.126103850229601e-05,
      "loss": 0.2988,
      "step": 1090
    },
    {
      "epoch": 7.65,
      "learning_rate": 3.108442246555988e-05,
      "loss": 0.3067,
      "step": 1100
    },
    {
      "epoch": 7.72,
      "learning_rate": 3.090780642882374e-05,
      "loss": 0.3065,
      "step": 1110
    },
    {
      "epoch": 7.79,
      "learning_rate": 3.073119039208761e-05,
      "loss": 0.3315,
      "step": 1120
    },
    {
      "epoch": 7.86,
      "learning_rate": 3.0554574355351466e-05,
      "loss": 0.2949,
      "step": 1130
    },
    {
      "epoch": 7.92,
      "learning_rate": 3.037795831861533e-05,
      "loss": 0.3493,
      "step": 1140
    },
    {
      "epoch": 7.99,
      "learning_rate": 3.02013422818792e-05,
      "loss": 0.3159,
      "step": 1150
    },
    {
      "epoch": 7.99,
      "eval_accuracy": 0.9620111731843576,
      "eval_loss": 0.13231398165225983,
      "eval_runtime": 19.7019,
      "eval_samples_per_second": 45.427,
      "eval_steps_per_second": 5.685,
      "step": 1150
    },
    {
      "epoch": 8.06,
      "learning_rate": 3.002472624514306e-05,
      "loss": 0.3193,
      "step": 1160
    },
    {
      "epoch": 8.13,
      "learning_rate": 2.9848110208406928e-05,
      "loss": 0.3245,
      "step": 1170
    },
    {
      "epoch": 8.2,
      "learning_rate": 2.967149417167079e-05,
      "loss": 0.2705,
      "step": 1180
    },
    {
      "epoch": 8.27,
      "learning_rate": 2.9494878134934657e-05,
      "loss": 0.2941,
      "step": 1190
    },
    {
      "epoch": 8.34,
      "learning_rate": 2.9318262098198518e-05,
      "loss": 0.2936,
      "step": 1200
    },
    {
      "epoch": 8.41,
      "learning_rate": 2.914164606146238e-05,
      "loss": 0.3415,
      "step": 1210
    },
    {
      "epoch": 8.48,
      "learning_rate": 2.8965030024726247e-05,
      "loss": 0.2976,
      "step": 1220
    },
    {
      "epoch": 8.55,
      "learning_rate": 2.8788413987990108e-05,
      "loss": 0.2979,
      "step": 1230
    },
    {
      "epoch": 8.62,
      "learning_rate": 2.8611797951253976e-05,
      "loss": 0.3391,
      "step": 1240
    },
    {
      "epoch": 8.69,
      "learning_rate": 2.8435181914517837e-05,
      "loss": 0.3329,
      "step": 1250
    },
    {
      "epoch": 8.76,
      "learning_rate": 2.8258565877781705e-05,
      "loss": 0.2764,
      "step": 1260
    },
    {
      "epoch": 8.83,
      "learning_rate": 2.808194984104557e-05,
      "loss": 0.2915,
      "step": 1270
    },
    {
      "epoch": 8.9,
      "learning_rate": 2.790533380430943e-05,
      "loss": 0.2607,
      "step": 1280
    },
    {
      "epoch": 8.97,
      "learning_rate": 2.77287177675733e-05,
      "loss": 0.2805,
      "step": 1290
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.958659217877095,
      "eval_loss": 0.16597720980644226,
      "eval_runtime": 21.3333,
      "eval_samples_per_second": 41.953,
      "eval_steps_per_second": 5.25,
      "step": 1294
    },
    {
      "epoch": 9.04,
      "learning_rate": 2.755210173083716e-05,
      "loss": 0.2688,
      "step": 1300
    },
    {
      "epoch": 9.11,
      "learning_rate": 2.7375485694101028e-05,
      "loss": 0.2965,
      "step": 1310
    },
    {
      "epoch": 9.18,
      "learning_rate": 2.719886965736489e-05,
      "loss": 0.2919,
      "step": 1320
    },
    {
      "epoch": 9.25,
      "learning_rate": 2.7022253620628757e-05,
      "loss": 0.2912,
      "step": 1330
    },
    {
      "epoch": 9.31,
      "learning_rate": 2.6845637583892618e-05,
      "loss": 0.26,
      "step": 1340
    },
    {
      "epoch": 9.38,
      "learning_rate": 2.666902154715648e-05,
      "loss": 0.2639,
      "step": 1350
    },
    {
      "epoch": 9.45,
      "learning_rate": 2.6492405510420347e-05,
      "loss": 0.3168,
      "step": 1360
    },
    {
      "epoch": 9.52,
      "learning_rate": 2.6315789473684212e-05,
      "loss": 0.2797,
      "step": 1370
    },
    {
      "epoch": 9.59,
      "learning_rate": 2.6139173436948077e-05,
      "loss": 0.2514,
      "step": 1380
    },
    {
      "epoch": 9.66,
      "learning_rate": 2.596255740021194e-05,
      "loss": 0.2826,
      "step": 1390
    },
    {
      "epoch": 9.73,
      "learning_rate": 2.578594136347581e-05,
      "loss": 0.3124,
      "step": 1400
    },
    {
      "epoch": 9.8,
      "learning_rate": 2.560932532673967e-05,
      "loss": 0.27,
      "step": 1410
    },
    {
      "epoch": 9.87,
      "learning_rate": 2.543270929000353e-05,
      "loss": 0.271,
      "step": 1420
    },
    {
      "epoch": 9.94,
      "learning_rate": 2.52560932532674e-05,
      "loss": 0.2657,
      "step": 1430
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9530726256983241,
      "eval_loss": 0.14559155702590942,
      "eval_runtime": 19.8491,
      "eval_samples_per_second": 45.09,
      "eval_steps_per_second": 5.643,
      "step": 1438
    },
    {
      "epoch": 10.01,
      "learning_rate": 2.507947721653126e-05,
      "loss": 0.2865,
      "step": 1440
    },
    {
      "epoch": 10.08,
      "learning_rate": 2.4902861179795125e-05,
      "loss": 0.2725,
      "step": 1450
    },
    {
      "epoch": 10.15,
      "learning_rate": 2.472624514305899e-05,
      "loss": 0.2701,
      "step": 1460
    },
    {
      "epoch": 10.22,
      "learning_rate": 2.4549629106322854e-05,
      "loss": 0.2808,
      "step": 1470
    },
    {
      "epoch": 10.29,
      "learning_rate": 2.437301306958672e-05,
      "loss": 0.2918,
      "step": 1480
    },
    {
      "epoch": 10.36,
      "learning_rate": 2.4196397032850583e-05,
      "loss": 0.291,
      "step": 1490
    },
    {
      "epoch": 10.43,
      "learning_rate": 2.4019780996114448e-05,
      "loss": 0.2738,
      "step": 1500
    },
    {
      "epoch": 10.5,
      "learning_rate": 2.3843164959378316e-05,
      "loss": 0.254,
      "step": 1510
    },
    {
      "epoch": 10.57,
      "learning_rate": 2.3666548922642177e-05,
      "loss": 0.245,
      "step": 1520
    },
    {
      "epoch": 10.64,
      "learning_rate": 2.348993288590604e-05,
      "loss": 0.265,
      "step": 1530
    },
    {
      "epoch": 10.71,
      "learning_rate": 2.3313316849169906e-05,
      "loss": 0.267,
      "step": 1540
    },
    {
      "epoch": 10.77,
      "learning_rate": 2.313670081243377e-05,
      "loss": 0.2262,
      "step": 1550
    },
    {
      "epoch": 10.84,
      "learning_rate": 2.2960084775697635e-05,
      "loss": 0.2396,
      "step": 1560
    },
    {
      "epoch": 10.91,
      "learning_rate": 2.27834687389615e-05,
      "loss": 0.2945,
      "step": 1570
    },
    {
      "epoch": 10.98,
      "learning_rate": 2.260685270222536e-05,
      "loss": 0.2929,
      "step": 1580
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9698324022346368,
      "eval_loss": 0.10857631266117096,
      "eval_runtime": 20.1943,
      "eval_samples_per_second": 44.319,
      "eval_steps_per_second": 5.546,
      "step": 1582
    },
    {
      "epoch": 11.05,
      "learning_rate": 2.2430236665489225e-05,
      "loss": 0.2728,
      "step": 1590
    },
    {
      "epoch": 11.12,
      "learning_rate": 2.225362062875309e-05,
      "loss": 0.2714,
      "step": 1600
    },
    {
      "epoch": 11.19,
      "learning_rate": 2.2077004592016954e-05,
      "loss": 0.2536,
      "step": 1610
    },
    {
      "epoch": 11.26,
      "learning_rate": 2.1900388555280822e-05,
      "loss": 0.2504,
      "step": 1620
    },
    {
      "epoch": 11.33,
      "learning_rate": 2.1723772518544687e-05,
      "loss": 0.2501,
      "step": 1630
    },
    {
      "epoch": 11.4,
      "learning_rate": 2.154715648180855e-05,
      "loss": 0.2787,
      "step": 1640
    },
    {
      "epoch": 11.47,
      "learning_rate": 2.1370540445072413e-05,
      "loss": 0.2295,
      "step": 1650
    },
    {
      "epoch": 11.54,
      "learning_rate": 2.1193924408336277e-05,
      "loss": 0.2616,
      "step": 1660
    },
    {
      "epoch": 11.61,
      "learning_rate": 2.1017308371600142e-05,
      "loss": 0.2796,
      "step": 1670
    },
    {
      "epoch": 11.68,
      "learning_rate": 2.0840692334864006e-05,
      "loss": 0.2889,
      "step": 1680
    },
    {
      "epoch": 11.75,
      "learning_rate": 2.066407629812787e-05,
      "loss": 0.3062,
      "step": 1690
    },
    {
      "epoch": 11.82,
      "learning_rate": 2.0487460261391735e-05,
      "loss": 0.2692,
      "step": 1700
    },
    {
      "epoch": 11.89,
      "learning_rate": 2.03108442246556e-05,
      "loss": 0.2466,
      "step": 1710
    },
    {
      "epoch": 11.96,
      "learning_rate": 2.013422818791946e-05,
      "loss": 0.2763,
      "step": 1720
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.976536312849162,
      "eval_loss": 0.088605597615242,
      "eval_runtime": 21.0879,
      "eval_samples_per_second": 42.441,
      "eval_steps_per_second": 5.311,
      "step": 1726
    },
    {
      "epoch": 12.03,
      "learning_rate": 1.995761215118333e-05,
      "loss": 0.2598,
      "step": 1730
    },
    {
      "epoch": 12.1,
      "learning_rate": 1.9780996114447194e-05,
      "loss": 0.2472,
      "step": 1740
    },
    {
      "epoch": 12.16,
      "learning_rate": 1.9604380077711058e-05,
      "loss": 0.3078,
      "step": 1750
    },
    {
      "epoch": 12.23,
      "learning_rate": 1.9427764040974923e-05,
      "loss": 0.2277,
      "step": 1760
    },
    {
      "epoch": 12.3,
      "learning_rate": 1.9251148004238787e-05,
      "loss": 0.2445,
      "step": 1770
    },
    {
      "epoch": 12.37,
      "learning_rate": 1.9074531967502652e-05,
      "loss": 0.2041,
      "step": 1780
    },
    {
      "epoch": 12.44,
      "learning_rate": 1.8897915930766513e-05,
      "loss": 0.2447,
      "step": 1790
    },
    {
      "epoch": 12.51,
      "learning_rate": 1.8721299894030378e-05,
      "loss": 0.2281,
      "step": 1800
    },
    {
      "epoch": 12.58,
      "learning_rate": 1.8544683857294242e-05,
      "loss": 0.234,
      "step": 1810
    },
    {
      "epoch": 12.65,
      "learning_rate": 1.8368067820558107e-05,
      "loss": 0.2434,
      "step": 1820
    },
    {
      "epoch": 12.72,
      "learning_rate": 1.819145178382197e-05,
      "loss": 0.2309,
      "step": 1830
    },
    {
      "epoch": 12.79,
      "learning_rate": 1.8014835747085836e-05,
      "loss": 0.2143,
      "step": 1840
    },
    {
      "epoch": 12.86,
      "learning_rate": 1.78382197103497e-05,
      "loss": 0.2691,
      "step": 1850
    },
    {
      "epoch": 12.93,
      "learning_rate": 1.7661603673613565e-05,
      "loss": 0.2371,
      "step": 1860
    },
    {
      "epoch": 13.0,
      "learning_rate": 1.748498763687743e-05,
      "loss": 0.2475,
      "step": 1870
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9731843575418995,
      "eval_loss": 0.1041172444820404,
      "eval_runtime": 19.6519,
      "eval_samples_per_second": 45.543,
      "eval_steps_per_second": 5.699,
      "step": 1870
    },
    {
      "epoch": 13.07,
      "learning_rate": 1.7308371600141294e-05,
      "loss": 0.2237,
      "step": 1880
    },
    {
      "epoch": 13.14,
      "learning_rate": 1.713175556340516e-05,
      "loss": 0.2541,
      "step": 1890
    },
    {
      "epoch": 13.21,
      "learning_rate": 1.6955139526669023e-05,
      "loss": 0.223,
      "step": 1900
    },
    {
      "epoch": 13.28,
      "learning_rate": 1.6778523489932888e-05,
      "loss": 0.2379,
      "step": 1910
    },
    {
      "epoch": 13.35,
      "learning_rate": 1.6601907453196752e-05,
      "loss": 0.2184,
      "step": 1920
    },
    {
      "epoch": 13.42,
      "learning_rate": 1.6425291416460613e-05,
      "loss": 0.1985,
      "step": 1930
    },
    {
      "epoch": 13.49,
      "learning_rate": 1.6248675379724478e-05,
      "loss": 0.2464,
      "step": 1940
    },
    {
      "epoch": 13.56,
      "learning_rate": 1.6072059342988343e-05,
      "loss": 0.2067,
      "step": 1950
    },
    {
      "epoch": 13.62,
      "learning_rate": 1.5895443306252207e-05,
      "loss": 0.2622,
      "step": 1960
    },
    {
      "epoch": 13.69,
      "learning_rate": 1.571882726951607e-05,
      "loss": 0.2322,
      "step": 1970
    },
    {
      "epoch": 13.76,
      "learning_rate": 1.554221123277994e-05,
      "loss": 0.2618,
      "step": 1980
    },
    {
      "epoch": 13.83,
      "learning_rate": 1.5365595196043804e-05,
      "loss": 0.2699,
      "step": 1990
    },
    {
      "epoch": 13.9,
      "learning_rate": 1.5188979159307665e-05,
      "loss": 0.2608,
      "step": 2000
    },
    {
      "epoch": 13.97,
      "learning_rate": 1.501236312257153e-05,
      "loss": 0.2148,
      "step": 2010
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9776536312849162,
      "eval_loss": 0.09550000727176666,
      "eval_runtime": 20.0949,
      "eval_samples_per_second": 44.539,
      "eval_steps_per_second": 5.574,
      "step": 2014
    },
    {
      "epoch": 14.04,
      "learning_rate": 1.4835747085835394e-05,
      "loss": 0.2454,
      "step": 2020
    },
    {
      "epoch": 14.11,
      "learning_rate": 1.4659131049099259e-05,
      "loss": 0.2304,
      "step": 2030
    },
    {
      "epoch": 14.18,
      "learning_rate": 1.4482515012363124e-05,
      "loss": 0.254,
      "step": 2040
    },
    {
      "epoch": 14.25,
      "learning_rate": 1.4305898975626988e-05,
      "loss": 0.2417,
      "step": 2050
    },
    {
      "epoch": 14.32,
      "learning_rate": 1.4129282938890853e-05,
      "loss": 0.235,
      "step": 2060
    },
    {
      "epoch": 14.39,
      "learning_rate": 1.3952666902154715e-05,
      "loss": 0.2349,
      "step": 2070
    },
    {
      "epoch": 14.46,
      "learning_rate": 1.377605086541858e-05,
      "loss": 0.244,
      "step": 2080
    },
    {
      "epoch": 14.53,
      "learning_rate": 1.3599434828682445e-05,
      "loss": 0.2315,
      "step": 2090
    },
    {
      "epoch": 14.6,
      "learning_rate": 1.3422818791946309e-05,
      "loss": 0.219,
      "step": 2100
    },
    {
      "epoch": 14.67,
      "learning_rate": 1.3246202755210174e-05,
      "loss": 0.2722,
      "step": 2110
    },
    {
      "epoch": 14.74,
      "learning_rate": 1.3069586718474038e-05,
      "loss": 0.2546,
      "step": 2120
    },
    {
      "epoch": 14.81,
      "learning_rate": 1.2892970681737905e-05,
      "loss": 0.2179,
      "step": 2130
    },
    {
      "epoch": 14.88,
      "learning_rate": 1.2716354645001766e-05,
      "loss": 0.2407,
      "step": 2140
    },
    {
      "epoch": 14.95,
      "learning_rate": 1.253973860826563e-05,
      "loss": 0.209,
      "step": 2150
    },
    {
      "epoch": 14.99,
      "eval_accuracy": 0.9709497206703911,
      "eval_loss": 0.10605238378047943,
      "eval_runtime": 21.3223,
      "eval_samples_per_second": 41.975,
      "eval_steps_per_second": 5.253,
      "step": 2157
    },
    {
      "epoch": 15.01,
      "learning_rate": 1.2363122571529495e-05,
      "loss": 0.2344,
      "step": 2160
    },
    {
      "epoch": 15.08,
      "learning_rate": 1.218650653479336e-05,
      "loss": 0.2087,
      "step": 2170
    },
    {
      "epoch": 15.15,
      "learning_rate": 1.2009890498057224e-05,
      "loss": 0.1903,
      "step": 2180
    },
    {
      "epoch": 15.22,
      "learning_rate": 1.1833274461321088e-05,
      "loss": 0.2297,
      "step": 2190
    },
    {
      "epoch": 15.29,
      "learning_rate": 1.1656658424584953e-05,
      "loss": 0.2814,
      "step": 2200
    },
    {
      "epoch": 15.36,
      "learning_rate": 1.1480042387848818e-05,
      "loss": 0.2455,
      "step": 2210
    },
    {
      "epoch": 15.43,
      "learning_rate": 1.130342635111268e-05,
      "loss": 0.2502,
      "step": 2220
    },
    {
      "epoch": 15.5,
      "learning_rate": 1.1126810314376545e-05,
      "loss": 0.2388,
      "step": 2230
    },
    {
      "epoch": 15.57,
      "learning_rate": 1.0950194277640411e-05,
      "loss": 0.23,
      "step": 2240
    },
    {
      "epoch": 15.64,
      "learning_rate": 1.0773578240904276e-05,
      "loss": 0.2555,
      "step": 2250
    },
    {
      "epoch": 15.71,
      "learning_rate": 1.0596962204168139e-05,
      "loss": 0.2256,
      "step": 2260
    },
    {
      "epoch": 15.78,
      "learning_rate": 1.0420346167432003e-05,
      "loss": 0.2611,
      "step": 2270
    },
    {
      "epoch": 15.85,
      "learning_rate": 1.0243730130695868e-05,
      "loss": 0.232,
      "step": 2280
    },
    {
      "epoch": 15.92,
      "learning_rate": 1.006711409395973e-05,
      "loss": 0.2425,
      "step": 2290
    },
    {
      "epoch": 15.99,
      "learning_rate": 9.890498057223597e-06,
      "loss": 0.2408,
      "step": 2300
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9743016759776536,
      "eval_loss": 0.07841791957616806,
      "eval_runtime": 19.8182,
      "eval_samples_per_second": 45.161,
      "eval_steps_per_second": 5.651,
      "step": 2301
    },
    {
      "epoch": 16.06,
      "learning_rate": 9.713882020487461e-06,
      "loss": 0.2557,
      "step": 2310
    },
    {
      "epoch": 16.13,
      "learning_rate": 9.537265983751326e-06,
      "loss": 0.2286,
      "step": 2320
    },
    {
      "epoch": 16.2,
      "learning_rate": 9.360649947015189e-06,
      "loss": 0.2412,
      "step": 2330
    },
    {
      "epoch": 16.27,
      "learning_rate": 9.184033910279053e-06,
      "loss": 0.2506,
      "step": 2340
    },
    {
      "epoch": 16.34,
      "learning_rate": 9.007417873542918e-06,
      "loss": 0.261,
      "step": 2350
    },
    {
      "epoch": 16.41,
      "learning_rate": 8.830801836806782e-06,
      "loss": 0.2101,
      "step": 2360
    },
    {
      "epoch": 16.47,
      "learning_rate": 8.654185800070647e-06,
      "loss": 0.2341,
      "step": 2370
    },
    {
      "epoch": 16.54,
      "learning_rate": 8.477569763334512e-06,
      "loss": 0.1933,
      "step": 2380
    },
    {
      "epoch": 16.61,
      "learning_rate": 8.300953726598376e-06,
      "loss": 0.2145,
      "step": 2390
    },
    {
      "epoch": 16.68,
      "learning_rate": 8.124337689862239e-06,
      "loss": 0.2417,
      "step": 2400
    },
    {
      "epoch": 16.75,
      "learning_rate": 7.947721653126104e-06,
      "loss": 0.2545,
      "step": 2410
    },
    {
      "epoch": 16.82,
      "learning_rate": 7.77110561638997e-06,
      "loss": 0.2472,
      "step": 2420
    },
    {
      "epoch": 16.89,
      "learning_rate": 7.594489579653833e-06,
      "loss": 0.2041,
      "step": 2430
    },
    {
      "epoch": 16.96,
      "learning_rate": 7.417873542917697e-06,
      "loss": 0.222,
      "step": 2440
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.9698324022346368,
      "eval_loss": 0.08389975875616074,
      "eval_runtime": 20.7685,
      "eval_samples_per_second": 43.094,
      "eval_steps_per_second": 5.393,
      "step": 2445
    },
    {
      "epoch": 17.03,
      "learning_rate": 7.241257506181562e-06,
      "loss": 0.2161,
      "step": 2450
    },
    {
      "epoch": 17.1,
      "learning_rate": 7.064641469445426e-06,
      "loss": 0.2392,
      "step": 2460
    },
    {
      "epoch": 17.17,
      "learning_rate": 6.88802543270929e-06,
      "loss": 0.2356,
      "step": 2470
    },
    {
      "epoch": 17.24,
      "learning_rate": 6.7114093959731546e-06,
      "loss": 0.2305,
      "step": 2480
    },
    {
      "epoch": 17.31,
      "learning_rate": 6.534793359237019e-06,
      "loss": 0.2393,
      "step": 2490
    },
    {
      "epoch": 17.38,
      "learning_rate": 6.358177322500883e-06,
      "loss": 0.1977,
      "step": 2500
    },
    {
      "epoch": 17.45,
      "learning_rate": 6.181561285764747e-06,
      "loss": 0.1951,
      "step": 2510
    },
    {
      "epoch": 17.52,
      "learning_rate": 6.004945249028612e-06,
      "loss": 0.2562,
      "step": 2520
    },
    {
      "epoch": 17.59,
      "learning_rate": 5.8283292122924765e-06,
      "loss": 0.2741,
      "step": 2530
    },
    {
      "epoch": 17.66,
      "learning_rate": 5.65171317555634e-06,
      "loss": 0.2385,
      "step": 2540
    },
    {
      "epoch": 17.73,
      "learning_rate": 5.475097138820206e-06,
      "loss": 0.2471,
      "step": 2550
    },
    {
      "epoch": 17.8,
      "learning_rate": 5.298481102084069e-06,
      "loss": 0.2479,
      "step": 2560
    },
    {
      "epoch": 17.86,
      "learning_rate": 5.121865065347934e-06,
      "loss": 0.2439,
      "step": 2570
    },
    {
      "epoch": 17.93,
      "learning_rate": 4.945249028611798e-06,
      "loss": 0.208,
      "step": 2580
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.9731843575418995,
      "eval_loss": 0.08734852075576782,
      "eval_runtime": 20.3969,
      "eval_samples_per_second": 43.879,
      "eval_steps_per_second": 5.491,
      "step": 2589
    },
    {
      "epoch": 18.0,
      "learning_rate": 4.768632991875663e-06,
      "loss": 0.2174,
      "step": 2590
    },
    {
      "epoch": 18.07,
      "learning_rate": 4.592016955139527e-06,
      "loss": 0.1979,
      "step": 2600
    },
    {
      "epoch": 18.14,
      "learning_rate": 4.415400918403391e-06,
      "loss": 0.2435,
      "step": 2610
    },
    {
      "epoch": 18.21,
      "learning_rate": 4.238784881667256e-06,
      "loss": 0.2102,
      "step": 2620
    },
    {
      "epoch": 18.28,
      "learning_rate": 4.0621688449311195e-06,
      "loss": 0.2216,
      "step": 2630
    },
    {
      "epoch": 18.35,
      "learning_rate": 3.885552808194985e-06,
      "loss": 0.233,
      "step": 2640
    },
    {
      "epoch": 18.42,
      "learning_rate": 3.7089367714588486e-06,
      "loss": 0.2159,
      "step": 2650
    },
    {
      "epoch": 18.49,
      "learning_rate": 3.532320734722713e-06,
      "loss": 0.2606,
      "step": 2660
    },
    {
      "epoch": 18.56,
      "learning_rate": 3.3557046979865773e-06,
      "loss": 0.1915,
      "step": 2670
    },
    {
      "epoch": 18.63,
      "learning_rate": 3.1790886612504414e-06,
      "loss": 0.2747,
      "step": 2680
    },
    {
      "epoch": 18.7,
      "learning_rate": 3.002472624514306e-06,
      "loss": 0.2493,
      "step": 2690
    },
    {
      "epoch": 18.77,
      "learning_rate": 2.82585658777817e-06,
      "loss": 0.1871,
      "step": 2700
    },
    {
      "epoch": 18.84,
      "learning_rate": 2.6492405510420347e-06,
      "loss": 0.209,
      "step": 2710
    },
    {
      "epoch": 18.91,
      "learning_rate": 2.472624514305899e-06,
      "loss": 0.2328,
      "step": 2720
    },
    {
      "epoch": 18.98,
      "learning_rate": 2.2960084775697633e-06,
      "loss": 0.2214,
      "step": 2730
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.9787709497206704,
      "eval_loss": 0.08888570219278336,
      "eval_runtime": 19.8638,
      "eval_samples_per_second": 45.057,
      "eval_steps_per_second": 5.638,
      "step": 2733
    },
    {
      "epoch": 19.05,
      "learning_rate": 2.119392440833628e-06,
      "loss": 0.2477,
      "step": 2740
    },
    {
      "epoch": 19.12,
      "learning_rate": 1.9427764040974924e-06,
      "loss": 0.2084,
      "step": 2750
    },
    {
      "epoch": 19.19,
      "learning_rate": 1.7661603673613566e-06,
      "loss": 0.2407,
      "step": 2760
    },
    {
      "epoch": 19.26,
      "learning_rate": 1.5895443306252207e-06,
      "loss": 0.205,
      "step": 2770
    },
    {
      "epoch": 19.32,
      "learning_rate": 1.412928293889085e-06,
      "loss": 0.2047,
      "step": 2780
    },
    {
      "epoch": 19.39,
      "learning_rate": 1.2363122571529496e-06,
      "loss": 0.2102,
      "step": 2790
    },
    {
      "epoch": 19.46,
      "learning_rate": 1.059696220416814e-06,
      "loss": 0.2407,
      "step": 2800
    },
    {
      "epoch": 19.53,
      "learning_rate": 8.830801836806783e-07,
      "loss": 0.2137,
      "step": 2810
    },
    {
      "epoch": 19.6,
      "learning_rate": 7.064641469445425e-07,
      "loss": 0.2519,
      "step": 2820
    },
    {
      "epoch": 19.67,
      "learning_rate": 5.29848110208407e-07,
      "loss": 0.2114,
      "step": 2830
    },
    {
      "epoch": 19.74,
      "learning_rate": 3.5323207347227126e-07,
      "loss": 0.2116,
      "step": 2840
    },
    {
      "epoch": 19.81,
      "learning_rate": 1.7661603673613563e-07,
      "loss": 0.2044,
      "step": 2850
    },
    {
      "epoch": 19.88,
      "learning_rate": 0.0,
      "loss": 0.2375,
      "step": 2860
    },
    {
      "epoch": 19.88,
      "eval_accuracy": 0.9743016759776536,
      "eval_loss": 0.0864485576748848,
      "eval_runtime": 21.3872,
      "eval_samples_per_second": 41.847,
      "eval_steps_per_second": 5.237,
      "step": 2860
    },
    {
      "epoch": 19.88,
      "step": 2860,
      "total_flos": 3.399744572389122e+18,
      "train_loss": 0.37100943968846245,
      "train_runtime": 12472.4089,
      "train_samples_per_second": 12.912,
      "train_steps_per_second": 0.229
    }
  ],
  "max_steps": 2860,
  "num_train_epochs": 20,
  "total_flos": 3.399744572389122e+18,
  "trial_name": null,
  "trial_params": null
}