{
  "best_metric": 0.12866149842739105,
  "best_model_checkpoint": "ViT_LFW_Model4/checkpoint-1500",
  "epoch": 30.0,
  "global_step": 7230,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04,
      "learning_rate": 0.00019972337482710928,
      "loss": 4.8889,
      "step": 10
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00019944674965421854,
      "loss": 4.6666,
      "step": 20
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0001991701244813278,
      "loss": 4.5092,
      "step": 30
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00019889349930843708,
      "loss": 4.3611,
      "step": 40
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00019861687413554635,
      "loss": 4.108,
      "step": 50
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00019834024896265561,
      "loss": 3.9738,
      "step": 60
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00019806362378976488,
      "loss": 4.0113,
      "step": 70
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00019778699861687415,
      "loss": 3.7319,
      "step": 80
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00019751037344398342,
      "loss": 3.7732,
      "step": 90
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00019723374827109268,
      "loss": 3.4756,
      "step": 100
    },
    {
      "epoch": 0.41,
      "eval_accuracy": 0.6014760147601476,
      "eval_f1": 0.2697704377928782,
      "eval_loss": 2.8778722286224365,
      "eval_precision": 0.846091378546004,
      "eval_recall": 0.34063761097659406,
      "eval_runtime": 5.4803,
      "eval_samples_per_second": 49.45,
      "eval_steps_per_second": 6.204,
      "step": 100
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00019695712309820195,
      "loss": 3.3636,
      "step": 110
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0001966804979253112,
      "loss": 3.371,
      "step": 120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00019640387275242046,
      "loss": 2.9877,
      "step": 130
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00019612724757952973,
      "loss": 3.3375,
      "step": 140
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000195850622406639,
      "loss": 3.0179,
      "step": 150
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00019557399723374826,
      "loss": 3.0167,
      "step": 160
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00019529737206085753,
      "loss": 2.8845,
      "step": 170
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001950207468879668,
      "loss": 2.7591,
      "step": 180
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00019474412171507606,
      "loss": 2.6481,
      "step": 190
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00019446749654218533,
      "loss": 2.6524,
      "step": 200
    },
    {
      "epoch": 0.83,
      "eval_accuracy": 0.7749077490774908,
      "eval_f1": 0.506350370992651,
      "eval_loss": 1.8111718893051147,
      "eval_precision": 0.8298295022595891,
      "eval_recall": 0.5915300546448088,
      "eval_runtime": 5.0674,
      "eval_samples_per_second": 53.479,
      "eval_steps_per_second": 6.71,
      "step": 200
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.00019419087136929463,
      "loss": 2.5118,
      "step": 210
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0001939142461964039,
      "loss": 2.4079,
      "step": 220
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.00019363762102351316,
      "loss": 2.4003,
      "step": 230
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00019336099585062243,
      "loss": 2.3569,
      "step": 240
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0001930843706777317,
      "loss": 2.2189,
      "step": 250
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00019280774550484096,
      "loss": 2.0548,
      "step": 260
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.00019253112033195023,
      "loss": 1.7242,
      "step": 270
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0001922544951590595,
      "loss": 1.9848,
      "step": 280
    },
    {
      "epoch": 1.2,
      "learning_rate": 0.00019197786998616877,
      "loss": 1.813,
      "step": 290
    },
    {
      "epoch": 1.24,
      "learning_rate": 0.00019170124481327803,
      "loss": 1.6994,
      "step": 300
    },
    {
      "epoch": 1.24,
      "eval_accuracy": 0.8450184501845018,
      "eval_f1": 0.6160474839155702,
      "eval_loss": 1.1828616857528687,
      "eval_precision": 0.8065270188221006,
      "eval_recall": 0.7112021857923497,
      "eval_runtime": 5.133,
      "eval_samples_per_second": 52.795,
      "eval_steps_per_second": 6.624,
      "step": 300
    },
    {
      "epoch": 1.29,
      "learning_rate": 0.0001914246196403873,
      "loss": 1.6323,
      "step": 310
    },
    {
      "epoch": 1.33,
      "learning_rate": 0.00019114799446749657,
      "loss": 1.6477,
      "step": 320
    },
    {
      "epoch": 1.37,
      "learning_rate": 0.00019087136929460584,
      "loss": 1.6168,
      "step": 330
    },
    {
      "epoch": 1.41,
      "learning_rate": 0.00019059474412171508,
      "loss": 1.4539,
      "step": 340
    },
    {
      "epoch": 1.45,
      "learning_rate": 0.00019031811894882434,
      "loss": 1.4612,
      "step": 350
    },
    {
      "epoch": 1.49,
      "learning_rate": 0.0001900414937759336,
      "loss": 1.4453,
      "step": 360
    },
    {
      "epoch": 1.54,
      "learning_rate": 0.00018976486860304288,
      "loss": 1.246,
      "step": 370
    },
    {
      "epoch": 1.58,
      "learning_rate": 0.00018948824343015215,
      "loss": 1.3659,
      "step": 380
    },
    {
      "epoch": 1.62,
      "learning_rate": 0.00018921161825726141,
      "loss": 1.1509,
      "step": 390
    },
    {
      "epoch": 1.66,
      "learning_rate": 0.00018893499308437068,
      "loss": 1.3097,
      "step": 400
    },
    {
      "epoch": 1.66,
      "eval_accuracy": 0.922509225092251,
      "eval_f1": 0.7908159406628629,
      "eval_loss": 0.6848821043968201,
      "eval_precision": 0.88076278659612,
      "eval_recall": 0.8486111111111111,
      "eval_runtime": 4.7102,
      "eval_samples_per_second": 57.535,
      "eval_steps_per_second": 7.218,
      "step": 400
    },
    {
      "epoch": 1.7,
      "learning_rate": 0.00018865836791147995,
      "loss": 1.1696,
      "step": 410
    },
    {
      "epoch": 1.74,
      "learning_rate": 0.00018838174273858922,
      "loss": 1.3098,
      "step": 420
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.00018810511756569848,
      "loss": 1.0508,
      "step": 430
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.00018782849239280775,
      "loss": 1.1426,
      "step": 440
    },
    {
      "epoch": 1.87,
      "learning_rate": 0.00018755186721991702,
      "loss": 1.0524,
      "step": 450
    },
    {
      "epoch": 1.91,
      "learning_rate": 0.0001872752420470263,
      "loss": 0.953,
      "step": 460
    },
    {
      "epoch": 1.95,
      "learning_rate": 0.00018699861687413555,
      "loss": 0.9812,
      "step": 470
    },
    {
      "epoch": 1.99,
      "learning_rate": 0.00018672199170124482,
      "loss": 0.9045,
      "step": 480
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0001864453665283541,
      "loss": 0.7454,
      "step": 490
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00018616874135546336,
      "loss": 0.5976,
      "step": 500
    },
    {
      "epoch": 2.07,
      "eval_accuracy": 0.933579335793358,
      "eval_f1": 0.8293473256291592,
      "eval_loss": 0.4778498709201813,
      "eval_precision": 0.9015032679738562,
      "eval_recall": 0.8802777777777778,
      "eval_runtime": 5.9354,
      "eval_samples_per_second": 45.658,
      "eval_steps_per_second": 5.728,
      "step": 500
    },
    {
      "epoch": 2.12,
      "learning_rate": 0.00018589211618257262,
      "loss": 0.6686,
      "step": 510
    },
    {
      "epoch": 2.16,
      "learning_rate": 0.0001856154910096819,
      "loss": 0.6399,
      "step": 520
    },
    {
      "epoch": 2.2,
      "learning_rate": 0.00018533886583679116,
      "loss": 0.6039,
      "step": 530
    },
    {
      "epoch": 2.24,
      "learning_rate": 0.00018506224066390043,
      "loss": 0.6056,
      "step": 540
    },
    {
      "epoch": 2.28,
      "learning_rate": 0.0001847856154910097,
      "loss": 0.5236,
      "step": 550
    },
    {
      "epoch": 2.32,
      "learning_rate": 0.00018450899031811896,
      "loss": 0.5595,
      "step": 560
    },
    {
      "epoch": 2.37,
      "learning_rate": 0.0001842323651452282,
      "loss": 0.5033,
      "step": 570
    },
    {
      "epoch": 2.41,
      "learning_rate": 0.00018395573997233747,
      "loss": 0.3865,
      "step": 580
    },
    {
      "epoch": 2.45,
      "learning_rate": 0.00018367911479944674,
      "loss": 0.4324,
      "step": 590
    },
    {
      "epoch": 2.49,
      "learning_rate": 0.000183402489626556,
      "loss": 0.412,
      "step": 600
    },
    {
      "epoch": 2.49,
      "eval_accuracy": 0.9298892988929889,
      "eval_f1": 0.8000484758297258,
      "eval_loss": 0.4110070765018463,
      "eval_precision": 0.85546875,
      "eval_recall": 0.8987654241864917,
      "eval_runtime": 5.3579,
      "eval_samples_per_second": 50.579,
      "eval_steps_per_second": 6.346,
      "step": 600
    },
    {
      "epoch": 2.53,
      "learning_rate": 0.00018312586445366527,
      "loss": 0.5181,
      "step": 610
    },
    {
      "epoch": 2.57,
      "learning_rate": 0.00018284923928077457,
      "loss": 0.431,
      "step": 620
    },
    {
      "epoch": 2.61,
      "learning_rate": 0.00018257261410788383,
      "loss": 0.3338,
      "step": 630
    },
    {
      "epoch": 2.66,
      "learning_rate": 0.0001822959889349931,
      "loss": 0.3139,
      "step": 640
    },
    {
      "epoch": 2.7,
      "learning_rate": 0.00018201936376210237,
      "loss": 0.299,
      "step": 650
    },
    {
      "epoch": 2.74,
      "learning_rate": 0.00018174273858921164,
      "loss": 0.3637,
      "step": 660
    },
    {
      "epoch": 2.78,
      "learning_rate": 0.0001814661134163209,
      "loss": 0.3153,
      "step": 670
    },
    {
      "epoch": 2.82,
      "learning_rate": 0.00018118948824343017,
      "loss": 0.3319,
      "step": 680
    },
    {
      "epoch": 2.86,
      "learning_rate": 0.00018091286307053944,
      "loss": 0.2579,
      "step": 690
    },
    {
      "epoch": 2.9,
      "learning_rate": 0.0001806362378976487,
      "loss": 0.3165,
      "step": 700
    },
    {
      "epoch": 2.9,
      "eval_accuracy": 0.9261992619926199,
      "eval_f1": 0.7349842967490026,
      "eval_loss": 0.3295341730117798,
      "eval_precision": 0.8107843137254902,
      "eval_recall": 0.8786764705882352,
      "eval_runtime": 5.1954,
      "eval_samples_per_second": 52.161,
      "eval_steps_per_second": 6.544,
      "step": 700
    },
    {
      "epoch": 2.95,
      "learning_rate": 0.00018035961272475797,
      "loss": 0.3355,
      "step": 710
    },
    {
      "epoch": 2.99,
      "learning_rate": 0.00018008298755186724,
      "loss": 0.2537,
      "step": 720
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0001798063623789765,
      "loss": 0.2528,
      "step": 730
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00017952973720608578,
      "loss": 0.1832,
      "step": 740
    },
    {
      "epoch": 3.11,
      "learning_rate": 0.00017925311203319504,
      "loss": 0.1962,
      "step": 750
    },
    {
      "epoch": 3.15,
      "learning_rate": 0.0001789764868603043,
      "loss": 0.1616,
      "step": 760
    },
    {
      "epoch": 3.2,
      "learning_rate": 0.00017869986168741358,
      "loss": 0.1541,
      "step": 770
    },
    {
      "epoch": 3.24,
      "learning_rate": 0.00017842323651452285,
      "loss": 0.1783,
      "step": 780
    },
    {
      "epoch": 3.28,
      "learning_rate": 0.00017814661134163211,
      "loss": 0.1569,
      "step": 790
    },
    {
      "epoch": 3.32,
      "learning_rate": 0.00017786998616874135,
      "loss": 0.1537,
      "step": 800
    },
    {
      "epoch": 3.32,
      "eval_accuracy": 0.9520295202952029,
      "eval_f1": 0.840494227994228,
      "eval_loss": 0.242670938372612,
      "eval_precision": 0.8791666666666667,
      "eval_recall": 0.9333333333333333,
      "eval_runtime": 4.7213,
      "eval_samples_per_second": 57.4,
      "eval_steps_per_second": 7.201,
      "step": 800
    },
    {
      "epoch": 3.36,
      "learning_rate": 0.00017759336099585062,
      "loss": 0.192,
      "step": 810
    },
    {
      "epoch": 3.4,
      "learning_rate": 0.0001773167358229599,
      "loss": 0.1323,
      "step": 820
    },
    {
      "epoch": 3.44,
      "learning_rate": 0.00017704011065006916,
      "loss": 0.1149,
      "step": 830
    },
    {
      "epoch": 3.49,
      "learning_rate": 0.00017676348547717842,
      "loss": 0.1185,
      "step": 840
    },
    {
      "epoch": 3.53,
      "learning_rate": 0.0001764868603042877,
      "loss": 0.112,
      "step": 850
    },
    {
      "epoch": 3.57,
      "learning_rate": 0.00017621023513139696,
      "loss": 0.109,
      "step": 860
    },
    {
      "epoch": 3.61,
      "learning_rate": 0.00017593360995850623,
      "loss": 0.1205,
      "step": 870
    },
    {
      "epoch": 3.65,
      "learning_rate": 0.0001756569847856155,
      "loss": 0.1542,
      "step": 880
    },
    {
      "epoch": 3.69,
      "learning_rate": 0.00017538035961272476,
      "loss": 0.1273,
      "step": 890
    },
    {
      "epoch": 3.73,
      "learning_rate": 0.00017510373443983403,
      "loss": 0.087,
      "step": 900
    },
    {
      "epoch": 3.73,
      "eval_accuracy": 0.9520295202952029,
      "eval_f1": 0.8561827956989246,
      "eval_loss": 0.23734191060066223,
      "eval_precision": 0.8989247311827957,
      "eval_recall": 0.930750350631136,
      "eval_runtime": 4.7854,
      "eval_samples_per_second": 56.63,
      "eval_steps_per_second": 7.105,
      "step": 900
    },
    {
      "epoch": 3.78,
      "learning_rate": 0.0001748271092669433,
      "loss": 0.0805,
      "step": 910
    },
    {
      "epoch": 3.82,
      "learning_rate": 0.00017455048409405256,
      "loss": 0.113,
      "step": 920
    },
    {
      "epoch": 3.86,
      "learning_rate": 0.00017427385892116183,
      "loss": 0.0657,
      "step": 930
    },
    {
      "epoch": 3.9,
      "learning_rate": 0.0001739972337482711,
      "loss": 0.1087,
      "step": 940
    },
    {
      "epoch": 3.94,
      "learning_rate": 0.00017372060857538037,
      "loss": 0.0809,
      "step": 950
    },
    {
      "epoch": 3.98,
      "learning_rate": 0.00017344398340248963,
      "loss": 0.0878,
      "step": 960
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0001731673582295989,
      "loss": 0.0669,
      "step": 970
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00017289073305670817,
      "loss": 0.0965,
      "step": 980
    },
    {
      "epoch": 4.11,
      "learning_rate": 0.00017261410788381744,
      "loss": 0.0601,
      "step": 990
    },
    {
      "epoch": 4.15,
      "learning_rate": 0.0001723374827109267,
      "loss": 0.0728,
      "step": 1000
    },
    {
      "epoch": 4.15,
      "eval_accuracy": 0.948339483394834,
      "eval_f1": 0.8297055375180376,
      "eval_loss": 0.20681600272655487,
      "eval_precision": 0.8815104166666666,
      "eval_recall": 0.9264039855072463,
      "eval_runtime": 5.2179,
      "eval_samples_per_second": 51.936,
      "eval_steps_per_second": 6.516,
      "step": 1000
    },
    {
      "epoch": 4.19,
      "learning_rate": 0.00017206085753803597,
      "loss": 0.0452,
      "step": 1010
    },
    {
      "epoch": 4.23,
      "learning_rate": 0.0001717842323651452,
      "loss": 0.0441,
      "step": 1020
    },
    {
      "epoch": 4.27,
      "learning_rate": 0.0001715076071922545,
      "loss": 0.0446,
      "step": 1030
    },
    {
      "epoch": 4.32,
      "learning_rate": 0.00017123098201936377,
      "loss": 0.0637,
      "step": 1040
    },
    {
      "epoch": 4.36,
      "learning_rate": 0.00017095435684647304,
      "loss": 0.0321,
      "step": 1050
    },
    {
      "epoch": 4.4,
      "learning_rate": 0.0001706777316735823,
      "loss": 0.0326,
      "step": 1060
    },
    {
      "epoch": 4.44,
      "learning_rate": 0.00017040110650069158,
      "loss": 0.0367,
      "step": 1070
    },
    {
      "epoch": 4.48,
      "learning_rate": 0.00017012448132780084,
      "loss": 0.0507,
      "step": 1080
    },
    {
      "epoch": 4.52,
      "learning_rate": 0.0001698478561549101,
      "loss": 0.0707,
      "step": 1090
    },
    {
      "epoch": 4.56,
      "learning_rate": 0.00016957123098201938,
      "loss": 0.0305,
      "step": 1100
    },
    {
      "epoch": 4.56,
      "eval_accuracy": 0.955719557195572,
      "eval_f1": 0.8278565878565878,
      "eval_loss": 0.17585988342761993,
      "eval_precision": 0.8692307692307693,
      "eval_recall": 0.9390746934225195,
      "eval_runtime": 5.0546,
      "eval_samples_per_second": 53.614,
      "eval_steps_per_second": 6.727,
      "step": 1100
    },
    {
      "epoch": 4.61,
      "learning_rate": 0.00016929460580912865,
      "loss": 0.028,
      "step": 1110
    },
    {
      "epoch": 4.65,
      "learning_rate": 0.00016901798063623791,
      "loss": 0.0565,
      "step": 1120
    },
    {
      "epoch": 4.69,
      "learning_rate": 0.00016874135546334718,
      "loss": 0.0615,
      "step": 1130
    },
    {
      "epoch": 4.73,
      "learning_rate": 0.00016846473029045645,
      "loss": 0.0447,
      "step": 1140
    },
    {
      "epoch": 4.77,
      "learning_rate": 0.00016818810511756572,
      "loss": 0.036,
      "step": 1150
    },
    {
      "epoch": 4.81,
      "learning_rate": 0.00016791147994467498,
      "loss": 0.0335,
      "step": 1160
    },
    {
      "epoch": 4.85,
      "learning_rate": 0.00016763485477178425,
      "loss": 0.0216,
      "step": 1170
    },
    {
      "epoch": 4.9,
      "learning_rate": 0.00016735822959889352,
      "loss": 0.0298,
      "step": 1180
    },
    {
      "epoch": 4.94,
      "learning_rate": 0.0001670816044260028,
      "loss": 0.0398,
      "step": 1190
    },
    {
      "epoch": 4.98,
      "learning_rate": 0.00016680497925311205,
      "loss": 0.0277,
      "step": 1200
    },
    {
      "epoch": 4.98,
      "eval_accuracy": 0.9446494464944649,
      "eval_f1": 0.7855656781029915,
      "eval_loss": 0.18792293965816498,
      "eval_precision": 0.8328358208955223,
      "eval_recall": 0.9197490806835388,
      "eval_runtime": 4.8182,
      "eval_samples_per_second": 56.245,
      "eval_steps_per_second": 7.057,
      "step": 1200
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00016652835408022132,
      "loss": 0.027,
      "step": 1210
    },
    {
      "epoch": 5.06,
      "learning_rate": 0.0001662517289073306,
      "loss": 0.0244,
      "step": 1220
    },
    {
      "epoch": 5.1,
      "learning_rate": 0.00016597510373443986,
      "loss": 0.0194,
      "step": 1230
    },
    {
      "epoch": 5.15,
      "learning_rate": 0.00016569847856154912,
      "loss": 0.0199,
      "step": 1240
    },
    {
      "epoch": 5.19,
      "learning_rate": 0.00016542185338865836,
      "loss": 0.0157,
      "step": 1250
    },
    {
      "epoch": 5.23,
      "learning_rate": 0.00016514522821576763,
      "loss": 0.0143,
      "step": 1260
    },
    {
      "epoch": 5.27,
      "learning_rate": 0.0001648686030428769,
      "loss": 0.0178,
      "step": 1270
    },
    {
      "epoch": 5.31,
      "learning_rate": 0.00016459197786998617,
      "loss": 0.0142,
      "step": 1280
    },
    {
      "epoch": 5.35,
      "learning_rate": 0.00016431535269709543,
      "loss": 0.0166,
      "step": 1290
    },
    {
      "epoch": 5.39,
      "learning_rate": 0.0001640387275242047,
      "loss": 0.0126,
      "step": 1300
    },
    {
      "epoch": 5.39,
      "eval_accuracy": 0.959409594095941,
      "eval_f1": 0.8193095793095793,
      "eval_loss": 0.17592698335647583,
      "eval_precision": 0.87,
      "eval_recall": 0.9333333333333333,
      "eval_runtime": 4.7932,
      "eval_samples_per_second": 56.538,
      "eval_steps_per_second": 7.093,
      "step": 1300
    },
    {
      "epoch": 5.44,
      "learning_rate": 0.00016376210235131397,
      "loss": 0.0435,
      "step": 1310
    },
    {
      "epoch": 5.48,
      "learning_rate": 0.00016348547717842324,
      "loss": 0.0141,
      "step": 1320
    },
    {
      "epoch": 5.52,
      "learning_rate": 0.0001632088520055325,
      "loss": 0.0217,
      "step": 1330
    },
    {
      "epoch": 5.56,
      "learning_rate": 0.00016293222683264177,
      "loss": 0.0191,
      "step": 1340
    },
    {
      "epoch": 5.6,
      "learning_rate": 0.00016265560165975104,
      "loss": 0.0152,
      "step": 1350
    },
    {
      "epoch": 5.64,
      "learning_rate": 0.0001623789764868603,
      "loss": 0.0176,
      "step": 1360
    },
    {
      "epoch": 5.68,
      "learning_rate": 0.00016210235131396957,
      "loss": 0.0281,
      "step": 1370
    },
    {
      "epoch": 5.73,
      "learning_rate": 0.00016182572614107884,
      "loss": 0.013,
      "step": 1380
    },
    {
      "epoch": 5.77,
      "learning_rate": 0.0001615491009681881,
      "loss": 0.0134,
      "step": 1390
    },
    {
      "epoch": 5.81,
      "learning_rate": 0.00016127247579529738,
      "loss": 0.0137,
      "step": 1400
    },
    {
      "epoch": 5.81,
      "eval_accuracy": 0.9630996309963099,
      "eval_f1": 0.8395517676767676,
      "eval_loss": 0.15947963297367096,
      "eval_precision": 0.8770833333333334,
      "eval_recall": 0.9440104166666666,
      "eval_runtime": 4.711,
      "eval_samples_per_second": 57.524,
      "eval_steps_per_second": 7.217,
      "step": 1400
    },
    {
      "epoch": 5.85,
      "learning_rate": 0.00016099585062240664,
      "loss": 0.0107,
      "step": 1410
    },
    {
      "epoch": 5.89,
      "learning_rate": 0.0001607192254495159,
      "loss": 0.0118,
      "step": 1420
    },
    {
      "epoch": 5.93,
      "learning_rate": 0.00016044260027662518,
      "loss": 0.0097,
      "step": 1430
    },
    {
      "epoch": 5.98,
      "learning_rate": 0.00016016597510373445,
      "loss": 0.0118,
      "step": 1440
    },
    {
      "epoch": 6.02,
      "learning_rate": 0.00015988934993084371,
      "loss": 0.0111,
      "step": 1450
    },
    {
      "epoch": 6.06,
      "learning_rate": 0.00015961272475795298,
      "loss": 0.0102,
      "step": 1460
    },
    {
      "epoch": 6.1,
      "learning_rate": 0.00015933609958506225,
      "loss": 0.0098,
      "step": 1470
    },
    {
      "epoch": 6.14,
      "learning_rate": 0.00015905947441217152,
      "loss": 0.0089,
      "step": 1480
    },
    {
      "epoch": 6.18,
      "learning_rate": 0.00015878284923928078,
      "loss": 0.0098,
      "step": 1490
    },
    {
      "epoch": 6.22,
      "learning_rate": 0.00015850622406639005,
      "loss": 0.0083,
      "step": 1500
    },
    {
      "epoch": 6.22,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8838383838383838,
      "eval_loss": 0.12866149842739105,
      "eval_precision": 0.9053763440860216,
      "eval_recall": 0.9583333333333333,
      "eval_runtime": 5.0711,
      "eval_samples_per_second": 53.44,
      "eval_steps_per_second": 6.705,
      "step": 1500
    },
    {
      "epoch": 6.27,
      "learning_rate": 0.00015822959889349932,
      "loss": 0.0082,
      "step": 1510
    },
    {
      "epoch": 6.31,
      "learning_rate": 0.0001579529737206086,
      "loss": 0.0089,
      "step": 1520
    },
    {
      "epoch": 6.35,
      "learning_rate": 0.00015767634854771785,
      "loss": 0.0092,
      "step": 1530
    },
    {
      "epoch": 6.39,
      "learning_rate": 0.00015739972337482712,
      "loss": 0.0083,
      "step": 1540
    },
    {
      "epoch": 6.43,
      "learning_rate": 0.0001571230982019364,
      "loss": 0.0083,
      "step": 1550
    },
    {
      "epoch": 6.47,
      "learning_rate": 0.00015684647302904566,
      "loss": 0.0083,
      "step": 1560
    },
    {
      "epoch": 6.51,
      "learning_rate": 0.00015656984785615492,
      "loss": 0.0072,
      "step": 1570
    },
    {
      "epoch": 6.56,
      "learning_rate": 0.0001562932226832642,
      "loss": 0.008,
      "step": 1580
    },
    {
      "epoch": 6.6,
      "learning_rate": 0.00015601659751037346,
      "loss": 0.0078,
      "step": 1590
    },
    {
      "epoch": 6.64,
      "learning_rate": 0.00015573997233748273,
      "loss": 0.0078,
      "step": 1600
    },
    {
      "epoch": 6.64,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.12949958443641663,
      "eval_precision": 0.891005291005291,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.8989,
      "eval_samples_per_second": 55.319,
      "eval_steps_per_second": 6.94,
      "step": 1600
    },
    {
      "epoch": 6.68,
      "learning_rate": 0.000155463347164592,
      "loss": 0.0078,
      "step": 1610
    },
    {
      "epoch": 6.72,
      "learning_rate": 0.00015518672199170126,
      "loss": 0.0073,
      "step": 1620
    },
    {
      "epoch": 6.76,
      "learning_rate": 0.00015491009681881053,
      "loss": 0.0074,
      "step": 1630
    },
    {
      "epoch": 6.8,
      "learning_rate": 0.0001546334716459198,
      "loss": 0.007,
      "step": 1640
    },
    {
      "epoch": 6.85,
      "learning_rate": 0.00015435684647302906,
      "loss": 0.0067,
      "step": 1650
    },
    {
      "epoch": 6.89,
      "learning_rate": 0.00015408022130013833,
      "loss": 0.0073,
      "step": 1660
    },
    {
      "epoch": 6.93,
      "learning_rate": 0.0001538035961272476,
      "loss": 0.0075,
      "step": 1670
    },
    {
      "epoch": 6.97,
      "learning_rate": 0.00015352697095435687,
      "loss": 0.008,
      "step": 1680
    },
    {
      "epoch": 7.01,
      "learning_rate": 0.00015325034578146613,
      "loss": 0.0065,
      "step": 1690
    },
    {
      "epoch": 7.05,
      "learning_rate": 0.00015297372060857537,
      "loss": 0.0064,
      "step": 1700
    },
    {
      "epoch": 7.05,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13224804401397705,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.4748,
      "eval_samples_per_second": 49.5,
      "eval_steps_per_second": 6.21,
      "step": 1700
    },
    {
      "epoch": 7.1,
      "learning_rate": 0.00015269709543568464,
      "loss": 0.0065,
      "step": 1710
    },
    {
      "epoch": 7.14,
      "learning_rate": 0.0001524204702627939,
      "loss": 0.0059,
      "step": 1720
    },
    {
      "epoch": 7.18,
      "learning_rate": 0.00015214384508990318,
      "loss": 0.0055,
      "step": 1730
    },
    {
      "epoch": 7.22,
      "learning_rate": 0.00015186721991701244,
      "loss": 0.0069,
      "step": 1740
    },
    {
      "epoch": 7.26,
      "learning_rate": 0.0001515905947441217,
      "loss": 0.0059,
      "step": 1750
    },
    {
      "epoch": 7.3,
      "learning_rate": 0.00015131396957123098,
      "loss": 0.0059,
      "step": 1760
    },
    {
      "epoch": 7.34,
      "learning_rate": 0.00015103734439834025,
      "loss": 0.0061,
      "step": 1770
    },
    {
      "epoch": 7.39,
      "learning_rate": 0.00015076071922544951,
      "loss": 0.0063,
      "step": 1780
    },
    {
      "epoch": 7.43,
      "learning_rate": 0.00015048409405255878,
      "loss": 0.0063,
      "step": 1790
    },
    {
      "epoch": 7.47,
      "learning_rate": 0.00015020746887966805,
      "loss": 0.0062,
      "step": 1800
    },
    {
      "epoch": 7.47,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.1299365907907486,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.3287,
      "eval_samples_per_second": 50.857,
      "eval_steps_per_second": 6.381,
      "step": 1800
    },
    {
      "epoch": 7.51,
      "learning_rate": 0.00014993084370677732,
      "loss": 0.006,
      "step": 1810
    },
    {
      "epoch": 7.55,
      "learning_rate": 0.00014965421853388658,
      "loss": 0.0054,
      "step": 1820
    },
    {
      "epoch": 7.59,
      "learning_rate": 0.00014937759336099585,
      "loss": 0.0052,
      "step": 1830
    },
    {
      "epoch": 7.63,
      "learning_rate": 0.00014910096818810512,
      "loss": 0.0057,
      "step": 1840
    },
    {
      "epoch": 7.68,
      "learning_rate": 0.0001488243430152144,
      "loss": 0.0055,
      "step": 1850
    },
    {
      "epoch": 7.72,
      "learning_rate": 0.00014854771784232365,
      "loss": 0.0057,
      "step": 1860
    },
    {
      "epoch": 7.76,
      "learning_rate": 0.00014827109266943292,
      "loss": 0.0057,
      "step": 1870
    },
    {
      "epoch": 7.8,
      "learning_rate": 0.0001479944674965422,
      "loss": 0.0052,
      "step": 1880
    },
    {
      "epoch": 7.84,
      "learning_rate": 0.00014771784232365146,
      "loss": 0.0057,
      "step": 1890
    },
    {
      "epoch": 7.88,
      "learning_rate": 0.00014744121715076072,
      "loss": 0.0053,
      "step": 1900
    },
    {
      "epoch": 7.88,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13068939745426178,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.5096,
      "eval_samples_per_second": 49.187,
      "eval_steps_per_second": 6.171,
      "step": 1900
    },
    {
      "epoch": 7.93,
      "learning_rate": 0.00014716459197787,
      "loss": 0.0053,
      "step": 1910
    },
    {
      "epoch": 7.97,
      "learning_rate": 0.00014688796680497926,
      "loss": 0.005,
      "step": 1920
    },
    {
      "epoch": 8.01,
      "learning_rate": 0.00014661134163208853,
      "loss": 0.005,
      "step": 1930
    },
    {
      "epoch": 8.05,
      "learning_rate": 0.0001463347164591978,
      "loss": 0.0049,
      "step": 1940
    },
    {
      "epoch": 8.09,
      "learning_rate": 0.00014605809128630706,
      "loss": 0.0048,
      "step": 1950
    },
    {
      "epoch": 8.13,
      "learning_rate": 0.00014578146611341633,
      "loss": 0.0054,
      "step": 1960
    },
    {
      "epoch": 8.17,
      "learning_rate": 0.0001455048409405256,
      "loss": 0.0049,
      "step": 1970
    },
    {
      "epoch": 8.22,
      "learning_rate": 0.00014522821576763486,
      "loss": 0.0046,
      "step": 1980
    },
    {
      "epoch": 8.26,
      "learning_rate": 0.00014495159059474413,
      "loss": 0.0045,
      "step": 1990
    },
    {
      "epoch": 8.3,
      "learning_rate": 0.0001446749654218534,
      "loss": 0.0049,
      "step": 2000
    },
    {
      "epoch": 8.3,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.12953458726406097,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.8018,
      "eval_samples_per_second": 46.71,
      "eval_steps_per_second": 5.86,
      "step": 2000
    },
    {
      "epoch": 8.34,
      "learning_rate": 0.00014439834024896267,
      "loss": 0.0046,
      "step": 2010
    },
    {
      "epoch": 8.38,
      "learning_rate": 0.00014412171507607193,
      "loss": 0.0044,
      "step": 2020
    },
    {
      "epoch": 8.42,
      "learning_rate": 0.0001438450899031812,
      "loss": 0.0043,
      "step": 2030
    },
    {
      "epoch": 8.46,
      "learning_rate": 0.00014356846473029047,
      "loss": 0.0043,
      "step": 2040
    },
    {
      "epoch": 8.51,
      "learning_rate": 0.00014329183955739974,
      "loss": 0.0045,
      "step": 2050
    },
    {
      "epoch": 8.55,
      "learning_rate": 0.000143015214384509,
      "loss": 0.0044,
      "step": 2060
    },
    {
      "epoch": 8.59,
      "learning_rate": 0.00014273858921161827,
      "loss": 0.0044,
      "step": 2070
    },
    {
      "epoch": 8.63,
      "learning_rate": 0.00014246196403872754,
      "loss": 0.0045,
      "step": 2080
    },
    {
      "epoch": 8.67,
      "learning_rate": 0.0001421853388658368,
      "loss": 0.0045,
      "step": 2090
    },
    {
      "epoch": 8.71,
      "learning_rate": 0.00014190871369294607,
      "loss": 0.0041,
      "step": 2100
    },
    {
      "epoch": 8.71,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.1301640421152115,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.1178,
      "eval_samples_per_second": 52.952,
      "eval_steps_per_second": 6.643,
      "step": 2100
    },
    {
      "epoch": 8.76,
      "learning_rate": 0.00014163208852005534,
      "loss": 0.0046,
      "step": 2110
    },
    {
      "epoch": 8.8,
      "learning_rate": 0.0001413554633471646,
      "loss": 0.0043,
      "step": 2120
    },
    {
      "epoch": 8.84,
      "learning_rate": 0.00014107883817427388,
      "loss": 0.0038,
      "step": 2130
    },
    {
      "epoch": 8.88,
      "learning_rate": 0.00014080221300138314,
      "loss": 0.0041,
      "step": 2140
    },
    {
      "epoch": 8.92,
      "learning_rate": 0.00014052558782849238,
      "loss": 0.0042,
      "step": 2150
    },
    {
      "epoch": 8.96,
      "learning_rate": 0.00014024896265560165,
      "loss": 0.004,
      "step": 2160
    },
    {
      "epoch": 9.0,
      "learning_rate": 0.00013997233748271092,
      "loss": 0.0039,
      "step": 2170
    },
    {
      "epoch": 9.05,
      "learning_rate": 0.0001396957123098202,
      "loss": 0.0039,
      "step": 2180
    },
    {
      "epoch": 9.09,
      "learning_rate": 0.00013941908713692945,
      "loss": 0.0043,
      "step": 2190
    },
    {
      "epoch": 9.13,
      "learning_rate": 0.00013914246196403872,
      "loss": 0.0036,
      "step": 2200
    },
    {
      "epoch": 9.13,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13097554445266724,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.7497,
      "eval_samples_per_second": 57.056,
      "eval_steps_per_second": 7.158,
      "step": 2200
    },
    {
      "epoch": 9.17,
      "learning_rate": 0.000138865836791148,
      "loss": 0.0038,
      "step": 2210
    },
    {
      "epoch": 9.21,
      "learning_rate": 0.00013858921161825726,
      "loss": 0.0041,
      "step": 2220
    },
    {
      "epoch": 9.25,
      "learning_rate": 0.00013831258644536652,
      "loss": 0.0032,
      "step": 2230
    },
    {
      "epoch": 9.29,
      "learning_rate": 0.0001380359612724758,
      "loss": 0.0043,
      "step": 2240
    },
    {
      "epoch": 9.34,
      "learning_rate": 0.00013775933609958506,
      "loss": 0.0037,
      "step": 2250
    },
    {
      "epoch": 9.38,
      "learning_rate": 0.00013748271092669433,
      "loss": 0.0037,
      "step": 2260
    },
    {
      "epoch": 9.42,
      "learning_rate": 0.0001372060857538036,
      "loss": 0.0032,
      "step": 2270
    },
    {
      "epoch": 9.46,
      "learning_rate": 0.00013692946058091286,
      "loss": 0.0039,
      "step": 2280
    },
    {
      "epoch": 9.5,
      "learning_rate": 0.00013665283540802213,
      "loss": 0.0033,
      "step": 2290
    },
    {
      "epoch": 9.54,
      "learning_rate": 0.0001363762102351314,
      "loss": 0.0037,
      "step": 2300
    },
    {
      "epoch": 9.54,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.131111741065979,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.4618,
      "eval_samples_per_second": 49.617,
      "eval_steps_per_second": 6.225,
      "step": 2300
    },
    {
      "epoch": 9.59,
      "learning_rate": 0.00013609958506224066,
      "loss": 0.0036,
      "step": 2310
    },
    {
      "epoch": 9.63,
      "learning_rate": 0.00013582295988934993,
      "loss": 0.0034,
      "step": 2320
    },
    {
      "epoch": 9.67,
      "learning_rate": 0.0001355463347164592,
      "loss": 0.0033,
      "step": 2330
    },
    {
      "epoch": 9.71,
      "learning_rate": 0.00013526970954356847,
      "loss": 0.004,
      "step": 2340
    },
    {
      "epoch": 9.75,
      "learning_rate": 0.00013499308437067773,
      "loss": 0.0037,
      "step": 2350
    },
    {
      "epoch": 9.79,
      "learning_rate": 0.000134716459197787,
      "loss": 0.0032,
      "step": 2360
    },
    {
      "epoch": 9.83,
      "learning_rate": 0.00013443983402489627,
      "loss": 0.0037,
      "step": 2370
    },
    {
      "epoch": 9.88,
      "learning_rate": 0.00013416320885200554,
      "loss": 0.0031,
      "step": 2380
    },
    {
      "epoch": 9.92,
      "learning_rate": 0.0001338865836791148,
      "loss": 0.0032,
      "step": 2390
    },
    {
      "epoch": 9.96,
      "learning_rate": 0.00013360995850622407,
      "loss": 0.0028,
      "step": 2400
    },
    {
      "epoch": 9.96,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.1300639659166336,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.162,
      "eval_samples_per_second": 52.499,
      "eval_steps_per_second": 6.587,
      "step": 2400
    },
    {
      "epoch": 10.0,
      "learning_rate": 0.00013333333333333334,
      "loss": 0.0031,
      "step": 2410
    },
    {
      "epoch": 10.04,
      "learning_rate": 0.0001330567081604426,
      "loss": 0.0032,
      "step": 2420
    },
    {
      "epoch": 10.08,
      "learning_rate": 0.00013278008298755187,
      "loss": 0.0035,
      "step": 2430
    },
    {
      "epoch": 10.12,
      "learning_rate": 0.00013250345781466114,
      "loss": 0.0031,
      "step": 2440
    },
    {
      "epoch": 10.17,
      "learning_rate": 0.0001322268326417704,
      "loss": 0.0032,
      "step": 2450
    },
    {
      "epoch": 10.21,
      "learning_rate": 0.00013195020746887968,
      "loss": 0.0031,
      "step": 2460
    },
    {
      "epoch": 10.25,
      "learning_rate": 0.00013167358229598894,
      "loss": 0.0031,
      "step": 2470
    },
    {
      "epoch": 10.29,
      "learning_rate": 0.0001313969571230982,
      "loss": 0.0027,
      "step": 2480
    },
    {
      "epoch": 10.33,
      "learning_rate": 0.00013112033195020748,
      "loss": 0.0029,
      "step": 2490
    },
    {
      "epoch": 10.37,
      "learning_rate": 0.00013084370677731675,
      "loss": 0.0031,
      "step": 2500
    },
    {
      "epoch": 10.37,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13081514835357666,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.7064,
      "eval_samples_per_second": 57.581,
      "eval_steps_per_second": 7.224,
      "step": 2500
    },
    {
      "epoch": 10.41,
      "learning_rate": 0.00013056708160442601,
      "loss": 0.003,
      "step": 2510
    },
    {
      "epoch": 10.46,
      "learning_rate": 0.00013029045643153528,
      "loss": 0.0028,
      "step": 2520
    },
    {
      "epoch": 10.5,
      "learning_rate": 0.00013001383125864455,
      "loss": 0.003,
      "step": 2530
    },
    {
      "epoch": 10.54,
      "learning_rate": 0.00012973720608575382,
      "loss": 0.0031,
      "step": 2540
    },
    {
      "epoch": 10.58,
      "learning_rate": 0.00012946058091286308,
      "loss": 0.0028,
      "step": 2550
    },
    {
      "epoch": 10.62,
      "learning_rate": 0.00012918395573997235,
      "loss": 0.0028,
      "step": 2560
    },
    {
      "epoch": 10.66,
      "learning_rate": 0.00012890733056708162,
      "loss": 0.0029,
      "step": 2570
    },
    {
      "epoch": 10.71,
      "learning_rate": 0.0001286307053941909,
      "loss": 0.003,
      "step": 2580
    },
    {
      "epoch": 10.75,
      "learning_rate": 0.00012835408022130015,
      "loss": 0.003,
      "step": 2590
    },
    {
      "epoch": 10.79,
      "learning_rate": 0.0001280774550484094,
      "loss": 0.0026,
      "step": 2600
    },
    {
      "epoch": 10.79,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13041570782661438,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.252,
      "eval_samples_per_second": 51.6,
      "eval_steps_per_second": 6.474,
      "step": 2600
    },
    {
      "epoch": 10.83,
      "learning_rate": 0.00012780082987551866,
      "loss": 0.0029,
      "step": 2610
    },
    {
      "epoch": 10.87,
      "learning_rate": 0.00012752420470262793,
      "loss": 0.0026,
      "step": 2620
    },
    {
      "epoch": 10.91,
      "learning_rate": 0.0001272475795297372,
      "loss": 0.0027,
      "step": 2630
    },
    {
      "epoch": 10.95,
      "learning_rate": 0.00012697095435684646,
      "loss": 0.0027,
      "step": 2640
    },
    {
      "epoch": 11.0,
      "learning_rate": 0.00012669432918395573,
      "loss": 0.0026,
      "step": 2650
    },
    {
      "epoch": 11.04,
      "learning_rate": 0.000126417704011065,
      "loss": 0.0028,
      "step": 2660
    },
    {
      "epoch": 11.08,
      "learning_rate": 0.00012614107883817427,
      "loss": 0.0025,
      "step": 2670
    },
    {
      "epoch": 11.12,
      "learning_rate": 0.00012586445366528353,
      "loss": 0.0027,
      "step": 2680
    },
    {
      "epoch": 11.16,
      "learning_rate": 0.0001255878284923928,
      "loss": 0.0025,
      "step": 2690
    },
    {
      "epoch": 11.2,
      "learning_rate": 0.00012531120331950207,
      "loss": 0.0023,
      "step": 2700
    },
    {
      "epoch": 11.2,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.1299184411764145,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.3147,
      "eval_samples_per_second": 50.99,
      "eval_steps_per_second": 6.397,
      "step": 2700
    },
    {
      "epoch": 11.24,
      "learning_rate": 0.00012503457814661134,
      "loss": 0.0026,
      "step": 2710
    },
    {
      "epoch": 11.29,
      "learning_rate": 0.0001247579529737206,
      "loss": 0.0024,
      "step": 2720
    },
    {
      "epoch": 11.33,
      "learning_rate": 0.00012448132780082987,
      "loss": 0.0028,
      "step": 2730
    },
    {
      "epoch": 11.37,
      "learning_rate": 0.00012420470262793914,
      "loss": 0.0025,
      "step": 2740
    },
    {
      "epoch": 11.41,
      "learning_rate": 0.0001239280774550484,
      "loss": 0.0026,
      "step": 2750
    },
    {
      "epoch": 11.45,
      "learning_rate": 0.00012365145228215767,
      "loss": 0.0026,
      "step": 2760
    },
    {
      "epoch": 11.49,
      "learning_rate": 0.00012337482710926694,
      "loss": 0.0026,
      "step": 2770
    },
    {
      "epoch": 11.54,
      "learning_rate": 0.0001230982019363762,
      "loss": 0.0025,
      "step": 2780
    },
    {
      "epoch": 11.58,
      "learning_rate": 0.00012282157676348548,
      "loss": 0.0026,
      "step": 2790
    },
    {
      "epoch": 11.62,
      "learning_rate": 0.00012254495159059477,
      "loss": 0.0024,
      "step": 2800
    },
    {
      "epoch": 11.62,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13152910768985748,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.8081,
      "eval_samples_per_second": 56.363,
      "eval_steps_per_second": 7.071,
      "step": 2800
    },
    {
      "epoch": 11.66,
      "learning_rate": 0.00012226832641770404,
      "loss": 0.0025,
      "step": 2810
    },
    {
      "epoch": 11.7,
      "learning_rate": 0.0001219917012448133,
      "loss": 0.0022,
      "step": 2820
    },
    {
      "epoch": 11.74,
      "learning_rate": 0.00012171507607192253,
      "loss": 0.0022,
      "step": 2830
    },
    {
      "epoch": 11.78,
      "learning_rate": 0.00012143845089903181,
      "loss": 0.0023,
      "step": 2840
    },
    {
      "epoch": 11.83,
      "learning_rate": 0.00012116182572614108,
      "loss": 0.0021,
      "step": 2850
    },
    {
      "epoch": 11.87,
      "learning_rate": 0.00012088520055325035,
      "loss": 0.0023,
      "step": 2860
    },
    {
      "epoch": 11.91,
      "learning_rate": 0.00012060857538035962,
      "loss": 0.0021,
      "step": 2870
    },
    {
      "epoch": 11.95,
      "learning_rate": 0.00012033195020746888,
      "loss": 0.0023,
      "step": 2880
    },
    {
      "epoch": 11.99,
      "learning_rate": 0.00012005532503457815,
      "loss": 0.0022,
      "step": 2890
    },
    {
      "epoch": 12.03,
      "learning_rate": 0.00011977869986168742,
      "loss": 0.0022,
      "step": 2900
    },
    {
      "epoch": 12.03,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13213181495666504,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.9695,
      "eval_samples_per_second": 54.533,
      "eval_steps_per_second": 6.842,
      "step": 2900
    },
    {
      "epoch": 12.07,
      "learning_rate": 0.00011950207468879669,
      "loss": 0.0022,
      "step": 2910
    },
    {
      "epoch": 12.12,
      "learning_rate": 0.00011922544951590595,
      "loss": 0.0025,
      "step": 2920
    },
    {
      "epoch": 12.16,
      "learning_rate": 0.00011894882434301522,
      "loss": 0.0023,
      "step": 2930
    },
    {
      "epoch": 12.2,
      "learning_rate": 0.00011867219917012449,
      "loss": 0.0024,
      "step": 2940
    },
    {
      "epoch": 12.24,
      "learning_rate": 0.00011839557399723376,
      "loss": 0.0025,
      "step": 2950
    },
    {
      "epoch": 12.28,
      "learning_rate": 0.00011811894882434302,
      "loss": 0.002,
      "step": 2960
    },
    {
      "epoch": 12.32,
      "learning_rate": 0.00011784232365145229,
      "loss": 0.002,
      "step": 2970
    },
    {
      "epoch": 12.37,
      "learning_rate": 0.00011756569847856156,
      "loss": 0.002,
      "step": 2980
    },
    {
      "epoch": 12.41,
      "learning_rate": 0.00011728907330567083,
      "loss": 0.002,
      "step": 2990
    },
    {
      "epoch": 12.45,
      "learning_rate": 0.0001170124481327801,
      "loss": 0.002,
      "step": 3000
    },
    {
      "epoch": 12.45,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13211499154567719,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.707,
      "eval_samples_per_second": 57.573,
      "eval_steps_per_second": 7.223,
      "step": 3000
    },
    {
      "epoch": 12.49,
      "learning_rate": 0.00011673582295988936,
      "loss": 0.002,
      "step": 3010
    },
    {
      "epoch": 12.53,
      "learning_rate": 0.00011645919778699863,
      "loss": 0.0023,
      "step": 3020
    },
    {
      "epoch": 12.57,
      "learning_rate": 0.0001161825726141079,
      "loss": 0.0019,
      "step": 3030
    },
    {
      "epoch": 12.61,
      "learning_rate": 0.00011590594744121716,
      "loss": 0.0021,
      "step": 3040
    },
    {
      "epoch": 12.66,
      "learning_rate": 0.00011562932226832643,
      "loss": 0.0021,
      "step": 3050
    },
    {
      "epoch": 12.7,
      "learning_rate": 0.00011535269709543569,
      "loss": 0.0021,
      "step": 3060
    },
    {
      "epoch": 12.74,
      "learning_rate": 0.00011507607192254495,
      "loss": 0.0021,
      "step": 3070
    },
    {
      "epoch": 12.78,
      "learning_rate": 0.00011479944674965422,
      "loss": 0.0018,
      "step": 3080
    },
    {
      "epoch": 12.82,
      "learning_rate": 0.00011452282157676349,
      "loss": 0.0019,
      "step": 3090
    },
    {
      "epoch": 12.86,
      "learning_rate": 0.00011424619640387276,
      "loss": 0.002,
      "step": 3100
    },
    {
      "epoch": 12.86,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13324756920337677,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.108,
      "eval_samples_per_second": 53.054,
      "eval_steps_per_second": 6.656,
      "step": 3100
    },
    {
      "epoch": 12.9,
      "learning_rate": 0.00011396957123098202,
      "loss": 0.0018,
      "step": 3110
    },
    {
      "epoch": 12.95,
      "learning_rate": 0.00011369294605809129,
      "loss": 0.002,
      "step": 3120
    },
    {
      "epoch": 12.99,
      "learning_rate": 0.00011341632088520056,
      "loss": 0.0019,
      "step": 3130
    },
    {
      "epoch": 13.03,
      "learning_rate": 0.00011313969571230983,
      "loss": 0.0017,
      "step": 3140
    },
    {
      "epoch": 13.07,
      "learning_rate": 0.0001128630705394191,
      "loss": 0.0018,
      "step": 3150
    },
    {
      "epoch": 13.11,
      "learning_rate": 0.00011258644536652836,
      "loss": 0.0019,
      "step": 3160
    },
    {
      "epoch": 13.15,
      "learning_rate": 0.00011230982019363763,
      "loss": 0.002,
      "step": 3170
    },
    {
      "epoch": 13.2,
      "learning_rate": 0.0001120331950207469,
      "loss": 0.0019,
      "step": 3180
    },
    {
      "epoch": 13.24,
      "learning_rate": 0.00011175656984785616,
      "loss": 0.002,
      "step": 3190
    },
    {
      "epoch": 13.28,
      "learning_rate": 0.00011147994467496543,
      "loss": 0.0017,
      "step": 3200
    },
    {
      "epoch": 13.28,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13269364833831787,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.7339,
      "eval_samples_per_second": 57.246,
      "eval_steps_per_second": 7.182,
      "step": 3200
    },
    {
      "epoch": 13.32,
      "learning_rate": 0.0001112033195020747,
      "loss": 0.002,
      "step": 3210
    },
    {
      "epoch": 13.36,
      "learning_rate": 0.00011092669432918397,
      "loss": 0.0017,
      "step": 3220
    },
    {
      "epoch": 13.4,
      "learning_rate": 0.00011065006915629323,
      "loss": 0.0017,
      "step": 3230
    },
    {
      "epoch": 13.44,
      "learning_rate": 0.0001103734439834025,
      "loss": 0.0018,
      "step": 3240
    },
    {
      "epoch": 13.49,
      "learning_rate": 0.00011009681881051177,
      "loss": 0.0017,
      "step": 3250
    },
    {
      "epoch": 13.53,
      "learning_rate": 0.00010982019363762104,
      "loss": 0.0018,
      "step": 3260
    },
    {
      "epoch": 13.57,
      "learning_rate": 0.0001095435684647303,
      "loss": 0.0018,
      "step": 3270
    },
    {
      "epoch": 13.61,
      "learning_rate": 0.00010926694329183956,
      "loss": 0.0019,
      "step": 3280
    },
    {
      "epoch": 13.65,
      "learning_rate": 0.00010899031811894882,
      "loss": 0.0017,
      "step": 3290
    },
    {
      "epoch": 13.69,
      "learning_rate": 0.00010871369294605809,
      "loss": 0.0016,
      "step": 3300
    },
    {
      "epoch": 13.69,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13280674815177917,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.7092,
      "eval_samples_per_second": 57.547,
      "eval_steps_per_second": 7.22,
      "step": 3300
    },
    {
      "epoch": 13.73,
      "learning_rate": 0.00010843706777316736,
      "loss": 0.0018,
      "step": 3310
    },
    {
      "epoch": 13.78,
      "learning_rate": 0.00010816044260027663,
      "loss": 0.0015,
      "step": 3320
    },
    {
      "epoch": 13.82,
      "learning_rate": 0.0001078838174273859,
      "loss": 0.002,
      "step": 3330
    },
    {
      "epoch": 13.86,
      "learning_rate": 0.00010760719225449516,
      "loss": 0.0018,
      "step": 3340
    },
    {
      "epoch": 13.9,
      "learning_rate": 0.00010733056708160443,
      "loss": 0.0016,
      "step": 3350
    },
    {
      "epoch": 13.94,
      "learning_rate": 0.0001070539419087137,
      "loss": 0.0017,
      "step": 3360
    },
    {
      "epoch": 13.98,
      "learning_rate": 0.00010677731673582296,
      "loss": 0.0016,
      "step": 3370
    },
    {
      "epoch": 14.02,
      "learning_rate": 0.00010650069156293223,
      "loss": 0.0019,
      "step": 3380
    },
    {
      "epoch": 14.07,
      "learning_rate": 0.0001062240663900415,
      "loss": 0.0016,
      "step": 3390
    },
    {
      "epoch": 14.11,
      "learning_rate": 0.00010594744121715077,
      "loss": 0.0015,
      "step": 3400
    },
    {
      "epoch": 14.11,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13356590270996094,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.4158,
      "eval_samples_per_second": 50.039,
      "eval_steps_per_second": 6.278,
      "step": 3400
    },
    {
      "epoch": 14.15,
      "learning_rate": 0.00010567081604426003,
      "loss": 0.0018,
      "step": 3410
    },
    {
      "epoch": 14.19,
      "learning_rate": 0.0001053941908713693,
      "loss": 0.0016,
      "step": 3420
    },
    {
      "epoch": 14.23,
      "learning_rate": 0.00010511756569847857,
      "loss": 0.0018,
      "step": 3430
    },
    {
      "epoch": 14.27,
      "learning_rate": 0.00010484094052558784,
      "loss": 0.0017,
      "step": 3440
    },
    {
      "epoch": 14.32,
      "learning_rate": 0.0001045643153526971,
      "loss": 0.0019,
      "step": 3450
    },
    {
      "epoch": 14.36,
      "learning_rate": 0.00010428769017980637,
      "loss": 0.0016,
      "step": 3460
    },
    {
      "epoch": 14.4,
      "learning_rate": 0.00010401106500691564,
      "loss": 0.0016,
      "step": 3470
    },
    {
      "epoch": 14.44,
      "learning_rate": 0.00010373443983402491,
      "loss": 0.0014,
      "step": 3480
    },
    {
      "epoch": 14.48,
      "learning_rate": 0.00010345781466113417,
      "loss": 0.0015,
      "step": 3490
    },
    {
      "epoch": 14.52,
      "learning_rate": 0.00010318118948824344,
      "loss": 0.0015,
      "step": 3500
    },
    {
      "epoch": 14.52,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13429729640483856,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.4445,
      "eval_samples_per_second": 49.775,
      "eval_steps_per_second": 6.245,
      "step": 3500
    },
    {
      "epoch": 14.56,
      "learning_rate": 0.0001029045643153527,
      "loss": 0.0016,
      "step": 3510
    },
    {
      "epoch": 14.61,
      "learning_rate": 0.00010262793914246196,
      "loss": 0.0015,
      "step": 3520
    },
    {
      "epoch": 14.65,
      "learning_rate": 0.00010235131396957123,
      "loss": 0.0014,
      "step": 3530
    },
    {
      "epoch": 14.69,
      "learning_rate": 0.0001020746887966805,
      "loss": 0.0016,
      "step": 3540
    },
    {
      "epoch": 14.73,
      "learning_rate": 0.00010179806362378977,
      "loss": 0.0015,
      "step": 3550
    },
    {
      "epoch": 14.77,
      "learning_rate": 0.00010152143845089903,
      "loss": 0.0014,
      "step": 3560
    },
    {
      "epoch": 14.81,
      "learning_rate": 0.0001012448132780083,
      "loss": 0.0014,
      "step": 3570
    },
    {
      "epoch": 14.85,
      "learning_rate": 0.00010096818810511757,
      "loss": 0.0014,
      "step": 3580
    },
    {
      "epoch": 14.9,
      "learning_rate": 0.00010069156293222684,
      "loss": 0.0016,
      "step": 3590
    },
    {
      "epoch": 14.94,
      "learning_rate": 0.0001004149377593361,
      "loss": 0.0015,
      "step": 3600
    },
    {
      "epoch": 14.94,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.1344984471797943,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.1207,
      "eval_samples_per_second": 52.923,
      "eval_steps_per_second": 6.64,
      "step": 3600
    },
    {
      "epoch": 14.98,
      "learning_rate": 0.00010013831258644537,
      "loss": 0.0013,
      "step": 3610
    },
    {
      "epoch": 15.02,
      "learning_rate": 9.986168741355464e-05,
      "loss": 0.0014,
      "step": 3620
    },
    {
      "epoch": 15.06,
      "learning_rate": 9.95850622406639e-05,
      "loss": 0.0013,
      "step": 3630
    },
    {
      "epoch": 15.1,
      "learning_rate": 9.930843706777317e-05,
      "loss": 0.0013,
      "step": 3640
    },
    {
      "epoch": 15.15,
      "learning_rate": 9.903181189488244e-05,
      "loss": 0.0013,
      "step": 3650
    },
    {
      "epoch": 15.19,
      "learning_rate": 9.875518672199171e-05,
      "loss": 0.0013,
      "step": 3660
    },
    {
      "epoch": 15.23,
      "learning_rate": 9.847856154910098e-05,
      "loss": 0.0015,
      "step": 3670
    },
    {
      "epoch": 15.27,
      "learning_rate": 9.820193637621023e-05,
      "loss": 0.0015,
      "step": 3680
    },
    {
      "epoch": 15.31,
      "learning_rate": 9.79253112033195e-05,
      "loss": 0.0013,
      "step": 3690
    },
    {
      "epoch": 15.35,
      "learning_rate": 9.764868603042876e-05,
      "loss": 0.0014,
      "step": 3700
    },
    {
      "epoch": 15.35,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.1343967467546463,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.0524,
      "eval_samples_per_second": 53.638,
      "eval_steps_per_second": 6.73,
      "step": 3700
    },
    {
      "epoch": 15.39,
      "learning_rate": 9.737206085753803e-05,
      "loss": 0.0013,
      "step": 3710
    },
    {
      "epoch": 15.44,
      "learning_rate": 9.709543568464731e-05,
      "loss": 0.0012,
      "step": 3720
    },
    {
      "epoch": 15.48,
      "learning_rate": 9.681881051175658e-05,
      "loss": 0.0014,
      "step": 3730
    },
    {
      "epoch": 15.52,
      "learning_rate": 9.654218533886585e-05,
      "loss": 0.0013,
      "step": 3740
    },
    {
      "epoch": 15.56,
      "learning_rate": 9.626556016597512e-05,
      "loss": 0.0014,
      "step": 3750
    },
    {
      "epoch": 15.6,
      "learning_rate": 9.598893499308438e-05,
      "loss": 0.0014,
      "step": 3760
    },
    {
      "epoch": 15.64,
      "learning_rate": 9.571230982019365e-05,
      "loss": 0.0014,
      "step": 3770
    },
    {
      "epoch": 15.68,
      "learning_rate": 9.543568464730292e-05,
      "loss": 0.0014,
      "step": 3780
    },
    {
      "epoch": 15.73,
      "learning_rate": 9.515905947441217e-05,
      "loss": 0.0014,
      "step": 3790
    },
    {
      "epoch": 15.77,
      "learning_rate": 9.488243430152144e-05,
      "loss": 0.0013,
      "step": 3800
    },
    {
      "epoch": 15.77,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13538949191570282,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.9582,
      "eval_samples_per_second": 54.657,
      "eval_steps_per_second": 6.857,
      "step": 3800
    },
    {
      "epoch": 15.81,
      "learning_rate": 9.460580912863071e-05,
      "loss": 0.0014,
      "step": 3810
    },
    {
      "epoch": 15.85,
      "learning_rate": 9.432918395573997e-05,
      "loss": 0.0014,
      "step": 3820
    },
    {
      "epoch": 15.89,
      "learning_rate": 9.405255878284924e-05,
      "loss": 0.0013,
      "step": 3830
    },
    {
      "epoch": 15.93,
      "learning_rate": 9.377593360995851e-05,
      "loss": 0.0013,
      "step": 3840
    },
    {
      "epoch": 15.98,
      "learning_rate": 9.349930843706778e-05,
      "loss": 0.0013,
      "step": 3850
    },
    {
      "epoch": 16.02,
      "learning_rate": 9.322268326417704e-05,
      "loss": 0.0013,
      "step": 3860
    },
    {
      "epoch": 16.06,
      "learning_rate": 9.294605809128631e-05,
      "loss": 0.0012,
      "step": 3870
    },
    {
      "epoch": 16.1,
      "learning_rate": 9.266943291839558e-05,
      "loss": 0.0012,
      "step": 3880
    },
    {
      "epoch": 16.14,
      "learning_rate": 9.239280774550485e-05,
      "loss": 0.0012,
      "step": 3890
    },
    {
      "epoch": 16.18,
      "learning_rate": 9.21161825726141e-05,
      "loss": 0.0013,
      "step": 3900
    },
    {
      "epoch": 16.18,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13570712506771088,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.3303,
      "eval_samples_per_second": 50.841,
      "eval_steps_per_second": 6.379,
      "step": 3900
    },
    {
      "epoch": 16.22,
      "learning_rate": 9.183955739972337e-05,
      "loss": 0.0013,
      "step": 3910
    },
    {
      "epoch": 16.27,
      "learning_rate": 9.156293222683264e-05,
      "loss": 0.0012,
      "step": 3920
    },
    {
      "epoch": 16.31,
      "learning_rate": 9.128630705394192e-05,
      "loss": 0.0011,
      "step": 3930
    },
    {
      "epoch": 16.35,
      "learning_rate": 9.100968188105118e-05,
      "loss": 0.0011,
      "step": 3940
    },
    {
      "epoch": 16.39,
      "learning_rate": 9.073305670816045e-05,
      "loss": 0.0012,
      "step": 3950
    },
    {
      "epoch": 16.43,
      "learning_rate": 9.045643153526972e-05,
      "loss": 0.0013,
      "step": 3960
    },
    {
      "epoch": 16.47,
      "learning_rate": 9.017980636237899e-05,
      "loss": 0.0011,
      "step": 3970
    },
    {
      "epoch": 16.51,
      "learning_rate": 8.990318118948825e-05,
      "loss": 0.0013,
      "step": 3980
    },
    {
      "epoch": 16.56,
      "learning_rate": 8.962655601659752e-05,
      "loss": 0.0011,
      "step": 3990
    },
    {
      "epoch": 16.6,
      "learning_rate": 8.934993084370679e-05,
      "loss": 0.0012,
      "step": 4000
    },
    {
      "epoch": 16.6,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.1364743858575821,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.1075,
      "eval_samples_per_second": 53.059,
      "eval_steps_per_second": 6.657,
      "step": 4000
    },
    {
      "epoch": 16.64,
      "learning_rate": 8.907330567081606e-05,
      "loss": 0.0012,
      "step": 4010
    },
    {
      "epoch": 16.68,
      "learning_rate": 8.879668049792531e-05,
      "loss": 0.0013,
      "step": 4020
    },
    {
      "epoch": 16.72,
      "learning_rate": 8.852005532503458e-05,
      "loss": 0.0012,
      "step": 4030
    },
    {
      "epoch": 16.76,
      "learning_rate": 8.824343015214385e-05,
      "loss": 0.0012,
      "step": 4040
    },
    {
      "epoch": 16.8,
      "learning_rate": 8.796680497925311e-05,
      "loss": 0.0011,
      "step": 4050
    },
    {
      "epoch": 16.85,
      "learning_rate": 8.769017980636238e-05,
      "loss": 0.0011,
      "step": 4060
    },
    {
      "epoch": 16.89,
      "learning_rate": 8.741355463347165e-05,
      "loss": 0.0012,
      "step": 4070
    },
    {
      "epoch": 16.93,
      "learning_rate": 8.713692946058092e-05,
      "loss": 0.0012,
      "step": 4080
    },
    {
      "epoch": 16.97,
      "learning_rate": 8.686030428769018e-05,
      "loss": 0.0012,
      "step": 4090
    },
    {
      "epoch": 17.01,
      "learning_rate": 8.658367911479945e-05,
      "loss": 0.0011,
      "step": 4100
    },
    {
      "epoch": 17.01,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13566486537456512,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.1645,
      "eval_samples_per_second": 52.473,
      "eval_steps_per_second": 6.583,
      "step": 4100
    },
    {
      "epoch": 17.05,
      "learning_rate": 8.630705394190872e-05,
      "loss": 0.0013,
      "step": 4110
    },
    {
      "epoch": 17.1,
      "learning_rate": 8.603042876901799e-05,
      "loss": 0.0011,
      "step": 4120
    },
    {
      "epoch": 17.14,
      "learning_rate": 8.575380359612725e-05,
      "loss": 0.001,
      "step": 4130
    },
    {
      "epoch": 17.18,
      "learning_rate": 8.547717842323652e-05,
      "loss": 0.0011,
      "step": 4140
    },
    {
      "epoch": 17.22,
      "learning_rate": 8.520055325034579e-05,
      "loss": 0.0011,
      "step": 4150
    },
    {
      "epoch": 17.26,
      "learning_rate": 8.492392807745506e-05,
      "loss": 0.0012,
      "step": 4160
    },
    {
      "epoch": 17.3,
      "learning_rate": 8.464730290456432e-05,
      "loss": 0.001,
      "step": 4170
    },
    {
      "epoch": 17.34,
      "learning_rate": 8.437067773167359e-05,
      "loss": 0.0012,
      "step": 4180
    },
    {
      "epoch": 17.39,
      "learning_rate": 8.409405255878286e-05,
      "loss": 0.0011,
      "step": 4190
    },
    {
      "epoch": 17.43,
      "learning_rate": 8.381742738589213e-05,
      "loss": 0.001,
      "step": 4200
    },
    {
      "epoch": 17.43,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13612236082553864,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.97,
      "eval_samples_per_second": 54.527,
      "eval_steps_per_second": 6.841,
      "step": 4200
    },
    {
      "epoch": 17.47,
      "learning_rate": 8.35408022130014e-05,
      "loss": 0.001,
      "step": 4210
    },
    {
      "epoch": 17.51,
      "learning_rate": 8.326417704011066e-05,
      "loss": 0.001,
      "step": 4220
    },
    {
      "epoch": 17.55,
      "learning_rate": 8.298755186721993e-05,
      "loss": 0.0011,
      "step": 4230
    },
    {
      "epoch": 17.59,
      "learning_rate": 8.271092669432918e-05,
      "loss": 0.001,
      "step": 4240
    },
    {
      "epoch": 17.63,
      "learning_rate": 8.243430152143845e-05,
      "loss": 0.001,
      "step": 4250
    },
    {
      "epoch": 17.68,
      "learning_rate": 8.215767634854772e-05,
      "loss": 0.0011,
      "step": 4260
    },
    {
      "epoch": 17.72,
      "learning_rate": 8.188105117565698e-05,
      "loss": 0.001,
      "step": 4270
    },
    {
      "epoch": 17.76,
      "learning_rate": 8.160442600276625e-05,
      "loss": 0.0011,
      "step": 4280
    },
    {
      "epoch": 17.8,
      "learning_rate": 8.132780082987552e-05,
      "loss": 0.0011,
      "step": 4290
    },
    {
      "epoch": 17.84,
      "learning_rate": 8.105117565698479e-05,
      "loss": 0.001,
      "step": 4300
    },
    {
      "epoch": 17.84,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13639481365680695,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.7347,
      "eval_samples_per_second": 57.237,
      "eval_steps_per_second": 7.181,
      "step": 4300
    },
    {
      "epoch": 17.88,
      "learning_rate": 8.077455048409405e-05,
      "loss": 0.0011,
      "step": 4310
    },
    {
      "epoch": 17.93,
      "learning_rate": 8.049792531120332e-05,
      "loss": 0.0009,
      "step": 4320
    },
    {
      "epoch": 17.97,
      "learning_rate": 8.022130013831259e-05,
      "loss": 0.001,
      "step": 4330
    },
    {
      "epoch": 18.01,
      "learning_rate": 7.994467496542186e-05,
      "loss": 0.0009,
      "step": 4340
    },
    {
      "epoch": 18.05,
      "learning_rate": 7.966804979253112e-05,
      "loss": 0.001,
      "step": 4350
    },
    {
      "epoch": 18.09,
      "learning_rate": 7.939142461964039e-05,
      "loss": 0.0012,
      "step": 4360
    },
    {
      "epoch": 18.13,
      "learning_rate": 7.911479944674966e-05,
      "loss": 0.0011,
      "step": 4370
    },
    {
      "epoch": 18.17,
      "learning_rate": 7.883817427385893e-05,
      "loss": 0.001,
      "step": 4380
    },
    {
      "epoch": 18.22,
      "learning_rate": 7.85615491009682e-05,
      "loss": 0.001,
      "step": 4390
    },
    {
      "epoch": 18.26,
      "learning_rate": 7.828492392807746e-05,
      "loss": 0.001,
      "step": 4400
    },
    {
      "epoch": 18.26,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.137937530875206,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.7146,
      "eval_samples_per_second": 57.48,
      "eval_steps_per_second": 7.212,
      "step": 4400
    },
    {
      "epoch": 18.3,
      "learning_rate": 7.800829875518673e-05,
      "loss": 0.0009,
      "step": 4410
    },
    {
      "epoch": 18.34,
      "learning_rate": 7.7731673582296e-05,
      "loss": 0.0011,
      "step": 4420
    },
    {
      "epoch": 18.38,
      "learning_rate": 7.745504840940526e-05,
      "loss": 0.001,
      "step": 4430
    },
    {
      "epoch": 18.42,
      "learning_rate": 7.717842323651453e-05,
      "loss": 0.0009,
      "step": 4440
    },
    {
      "epoch": 18.46,
      "learning_rate": 7.69017980636238e-05,
      "loss": 0.0009,
      "step": 4450
    },
    {
      "epoch": 18.51,
      "learning_rate": 7.662517289073307e-05,
      "loss": 0.0009,
      "step": 4460
    },
    {
      "epoch": 18.55,
      "learning_rate": 7.634854771784232e-05,
      "loss": 0.0011,
      "step": 4470
    },
    {
      "epoch": 18.59,
      "learning_rate": 7.607192254495159e-05,
      "loss": 0.0009,
      "step": 4480
    },
    {
      "epoch": 18.63,
      "learning_rate": 7.579529737206086e-05,
      "loss": 0.0009,
      "step": 4490
    },
    {
      "epoch": 18.67,
      "learning_rate": 7.551867219917012e-05,
      "loss": 0.001,
      "step": 4500
    },
    {
      "epoch": 18.67,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.13750597834587097,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.0326,
      "eval_samples_per_second": 53.849,
      "eval_steps_per_second": 6.756,
      "step": 4500
    },
    {
      "epoch": 18.71,
      "learning_rate": 7.524204702627939e-05,
      "loss": 0.0009,
      "step": 4510
    },
    {
      "epoch": 18.76,
      "learning_rate": 7.496542185338866e-05,
      "loss": 0.0009,
      "step": 4520
    },
    {
      "epoch": 18.8,
      "learning_rate": 7.468879668049793e-05,
      "loss": 0.0009,
      "step": 4530
    },
    {
      "epoch": 18.84,
      "learning_rate": 7.44121715076072e-05,
      "loss": 0.0009,
      "step": 4540
    },
    {
      "epoch": 18.88,
      "learning_rate": 7.413554633471646e-05,
      "loss": 0.0009,
      "step": 4550
    },
    {
      "epoch": 18.92,
      "learning_rate": 7.385892116182573e-05,
      "loss": 0.0009,
      "step": 4560
    },
    {
      "epoch": 18.96,
      "learning_rate": 7.3582295988935e-05,
      "loss": 0.0009,
      "step": 4570
    },
    {
      "epoch": 19.0,
      "learning_rate": 7.330567081604426e-05,
      "loss": 0.0009,
      "step": 4580
    },
    {
      "epoch": 19.05,
      "learning_rate": 7.302904564315353e-05,
      "loss": 0.001,
      "step": 4590
    },
    {
      "epoch": 19.09,
      "learning_rate": 7.27524204702628e-05,
      "loss": 0.0009,
      "step": 4600
    },
    {
      "epoch": 19.09,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.1373746395111084,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.1005,
      "eval_samples_per_second": 53.132,
      "eval_steps_per_second": 6.666,
      "step": 4600
    },
    {
      "epoch": 19.13,
      "learning_rate": 7.247579529737207e-05,
      "loss": 0.0009,
      "step": 4610
    },
    {
      "epoch": 19.17,
      "learning_rate": 7.219917012448133e-05,
      "loss": 0.0008,
      "step": 4620
    },
    {
      "epoch": 19.21,
      "learning_rate": 7.19225449515906e-05,
      "loss": 0.0009,
      "step": 4630
    },
    {
      "epoch": 19.25,
      "learning_rate": 7.164591977869987e-05,
      "loss": 0.0009,
      "step": 4640
    },
    {
      "epoch": 19.29,
      "learning_rate": 7.136929460580914e-05,
      "loss": 0.0008,
      "step": 4650
    },
    {
      "epoch": 19.34,
      "learning_rate": 7.10926694329184e-05,
      "loss": 0.0008,
      "step": 4660
    },
    {
      "epoch": 19.38,
      "learning_rate": 7.081604426002767e-05,
      "loss": 0.0009,
      "step": 4670
    },
    {
      "epoch": 19.42,
      "learning_rate": 7.053941908713694e-05,
      "loss": 0.0008,
      "step": 4680
    },
    {
      "epoch": 19.46,
      "learning_rate": 7.026279391424619e-05,
      "loss": 0.0008,
      "step": 4690
    },
    {
      "epoch": 19.5,
      "learning_rate": 6.998616874135546e-05,
      "loss": 0.0009,
      "step": 4700
    },
    {
      "epoch": 19.5,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.1374416947364807,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.7128,
      "eval_samples_per_second": 47.438,
      "eval_steps_per_second": 5.952,
      "step": 4700
    },
    {
      "epoch": 19.54,
      "learning_rate": 6.970954356846473e-05,
      "loss": 0.0009,
      "step": 4710
    },
    {
      "epoch": 19.59,
      "learning_rate": 6.9432918395574e-05,
      "loss": 0.0009,
      "step": 4720
    },
    {
      "epoch": 19.63,
      "learning_rate": 6.915629322268326e-05,
      "loss": 0.0008,
      "step": 4730
    },
    {
      "epoch": 19.67,
      "learning_rate": 6.887966804979253e-05,
      "loss": 0.0008,
      "step": 4740
    },
    {
      "epoch": 19.71,
      "learning_rate": 6.86030428769018e-05,
      "loss": 0.0009,
      "step": 4750
    },
    {
      "epoch": 19.75,
      "learning_rate": 6.832641770401106e-05,
      "loss": 0.0009,
      "step": 4760
    },
    {
      "epoch": 19.79,
      "learning_rate": 6.804979253112033e-05,
      "loss": 0.0008,
      "step": 4770
    },
    {
      "epoch": 19.83,
      "learning_rate": 6.77731673582296e-05,
      "loss": 0.0008,
      "step": 4780
    },
    {
      "epoch": 19.88,
      "learning_rate": 6.749654218533887e-05,
      "loss": 0.0009,
      "step": 4790
    },
    {
      "epoch": 19.92,
      "learning_rate": 6.721991701244813e-05,
      "loss": 0.0009,
      "step": 4800
    },
    {
      "epoch": 19.92,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.1381683647632599,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 5.0077,
      "eval_samples_per_second": 54.116,
      "eval_steps_per_second": 6.789,
      "step": 4800
    },
    {
      "epoch": 19.96,
      "learning_rate": 6.69432918395574e-05,
      "loss": 0.0009,
      "step": 4810
    },
    {
      "epoch": 20.0,
      "learning_rate": 6.666666666666667e-05,
      "loss": 0.0008,
      "step": 4820
    },
    {
      "epoch": 20.04,
      "learning_rate": 6.639004149377594e-05,
      "loss": 0.0008,
      "step": 4830
    },
    {
      "epoch": 20.08,
      "learning_rate": 6.61134163208852e-05,
      "loss": 0.0008,
      "step": 4840
    },
    {
      "epoch": 20.12,
      "learning_rate": 6.583679114799447e-05,
      "loss": 0.0008,
      "step": 4850
    },
    {
      "epoch": 20.17,
      "learning_rate": 6.556016597510374e-05,
      "loss": 0.0007,
      "step": 4860
    },
    {
      "epoch": 20.21,
      "learning_rate": 6.528354080221301e-05,
      "loss": 0.0009,
      "step": 4870
    },
    {
      "epoch": 20.25,
      "learning_rate": 6.500691562932227e-05,
      "loss": 0.0008,
      "step": 4880
    },
    {
      "epoch": 20.29,
      "learning_rate": 6.473029045643154e-05,
      "loss": 0.0008,
      "step": 4890
    },
    {
      "epoch": 20.33,
      "learning_rate": 6.445366528354081e-05,
      "loss": 0.0008,
      "step": 4900
    },
    {
      "epoch": 20.33,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.1384662240743637,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 4.7789,
      "eval_samples_per_second": 56.708,
      "eval_steps_per_second": 7.115,
      "step": 4900
    },
    {
      "epoch": 20.37,
      "learning_rate": 6.417704011065008e-05,
      "loss": 0.0008,
      "step": 4910
    },
    {
      "epoch": 20.41,
      "learning_rate": 6.390041493775933e-05,
      "loss": 0.0009,
      "step": 4920
    },
    {
      "epoch": 20.46,
      "learning_rate": 6.36237897648686e-05,
      "loss": 0.0008,
      "step": 4930
    },
    {
      "epoch": 20.5,
      "learning_rate": 6.334716459197787e-05,
      "loss": 0.0008,
      "step": 4940
    },
    {
      "epoch": 20.54,
      "learning_rate": 6.307053941908713e-05,
      "loss": 0.0007,
      "step": 4950
    },
    {
      "epoch": 20.58,
      "learning_rate": 6.27939142461964e-05,
      "loss": 0.0008,
      "step": 4960
    },
    {
      "epoch": 20.62,
      "learning_rate": 6.251728907330567e-05,
      "loss": 0.0008,
      "step": 4970
    },
    {
      "epoch": 20.66,
      "learning_rate": 6.224066390041494e-05,
      "loss": 0.0008,
      "step": 4980
    },
    {
      "epoch": 20.71,
      "learning_rate": 6.19640387275242e-05,
      "loss": 0.0007,
      "step": 4990
    },
    {
      "epoch": 20.75,
      "learning_rate": 6.168741355463347e-05,
      "loss": 0.0007,
      "step": 5000
    },
    {
      "epoch": 20.75,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.138875350356102,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.2432,
      "eval_samples_per_second": 51.686,
      "eval_steps_per_second": 6.485,
      "step": 5000
    },
    {
      "epoch": 20.79,
      "learning_rate": 6.141078838174274e-05,
      "loss": 0.0008,
      "step": 5010
    },
    {
      "epoch": 20.83,
      "learning_rate": 6.113416320885202e-05,
      "loss": 0.0008,
      "step": 5020
    },
    {
      "epoch": 20.87,
      "learning_rate": 6.085753803596127e-05,
      "loss": 0.0008,
      "step": 5030
    },
    {
      "epoch": 20.91,
      "learning_rate": 6.058091286307054e-05,
      "loss": 0.0007,
      "step": 5040
    },
    {
      "epoch": 20.95,
      "learning_rate": 6.030428769017981e-05,
      "loss": 0.0007,
      "step": 5050
    },
    {
      "epoch": 21.0,
      "learning_rate": 6.0027662517289076e-05,
      "loss": 0.0006,
      "step": 5060
    },
    {
      "epoch": 21.04,
      "learning_rate": 5.9751037344398344e-05,
      "loss": 0.0008,
      "step": 5070
    },
    {
      "epoch": 21.08,
      "learning_rate": 5.947441217150761e-05,
      "loss": 0.0007,
      "step": 5080
    },
    {
      "epoch": 21.12,
      "learning_rate": 5.919778699861688e-05,
      "loss": 0.0008,
      "step": 5090
    },
    {
      "epoch": 21.16,
      "learning_rate": 5.8921161825726146e-05,
      "loss": 0.0007,
      "step": 5100
    },
    {
      "epoch": 21.16,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.1390545666217804,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.1771,
      "eval_samples_per_second": 52.346,
      "eval_steps_per_second": 6.567,
      "step": 5100
    },
    {
      "epoch": 21.2,
      "learning_rate": 5.8644536652835414e-05,
      "loss": 0.0007,
      "step": 5110
    },
    {
      "epoch": 21.24,
      "learning_rate": 5.836791147994468e-05,
      "loss": 0.0008,
      "step": 5120
    },
    {
      "epoch": 21.29,
      "learning_rate": 5.809128630705395e-05,
      "loss": 0.0007,
      "step": 5130
    },
    {
      "epoch": 21.33,
      "learning_rate": 5.7814661134163216e-05,
      "loss": 0.0007,
      "step": 5140
    },
    {
      "epoch": 21.37,
      "learning_rate": 5.753803596127248e-05,
      "loss": 0.0007,
      "step": 5150
    },
    {
      "epoch": 21.41,
      "learning_rate": 5.7261410788381744e-05,
      "loss": 0.0007,
      "step": 5160
    },
    {
      "epoch": 21.45,
      "learning_rate": 5.698478561549101e-05,
      "loss": 0.0007,
      "step": 5170
    },
    {
      "epoch": 21.49,
      "learning_rate": 5.670816044260028e-05,
      "loss": 0.0007,
      "step": 5180
    },
    {
      "epoch": 21.54,
      "learning_rate": 5.643153526970955e-05,
      "loss": 0.0007,
      "step": 5190
    },
    {
      "epoch": 21.58,
      "learning_rate": 5.6154910096818814e-05,
      "loss": 0.0007,
      "step": 5200
    },
    {
      "epoch": 21.58,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.13922710716724396,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 4.9996,
      "eval_samples_per_second": 54.204,
      "eval_steps_per_second": 6.801,
      "step": 5200
    },
    {
      "epoch": 21.62,
      "learning_rate": 5.587828492392808e-05,
      "loss": 0.0007,
      "step": 5210
    },
    {
      "epoch": 21.66,
      "learning_rate": 5.560165975103735e-05,
      "loss": 0.0007,
      "step": 5220
    },
    {
      "epoch": 21.7,
      "learning_rate": 5.532503457814662e-05,
      "loss": 0.0007,
      "step": 5230
    },
    {
      "epoch": 21.74,
      "learning_rate": 5.5048409405255884e-05,
      "loss": 0.0007,
      "step": 5240
    },
    {
      "epoch": 21.78,
      "learning_rate": 5.477178423236515e-05,
      "loss": 0.0007,
      "step": 5250
    },
    {
      "epoch": 21.83,
      "learning_rate": 5.449515905947441e-05,
      "loss": 0.0006,
      "step": 5260
    },
    {
      "epoch": 21.87,
      "learning_rate": 5.421853388658368e-05,
      "loss": 0.0007,
      "step": 5270
    },
    {
      "epoch": 21.91,
      "learning_rate": 5.394190871369295e-05,
      "loss": 0.0006,
      "step": 5280
    },
    {
      "epoch": 21.95,
      "learning_rate": 5.3665283540802215e-05,
      "loss": 0.0007,
      "step": 5290
    },
    {
      "epoch": 21.99,
      "learning_rate": 5.338865836791148e-05,
      "loss": 0.0007,
      "step": 5300
    },
    {
      "epoch": 21.99,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.13969561457633972,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 4.812,
      "eval_samples_per_second": 56.317,
      "eval_steps_per_second": 7.066,
      "step": 5300
    },
    {
      "epoch": 22.03,
      "learning_rate": 5.311203319502075e-05,
      "loss": 0.0007,
      "step": 5310
    },
    {
      "epoch": 22.07,
      "learning_rate": 5.283540802213002e-05,
      "loss": 0.0007,
      "step": 5320
    },
    {
      "epoch": 22.12,
      "learning_rate": 5.2558782849239285e-05,
      "loss": 0.0007,
      "step": 5330
    },
    {
      "epoch": 22.16,
      "learning_rate": 5.228215767634855e-05,
      "loss": 0.0006,
      "step": 5340
    },
    {
      "epoch": 22.2,
      "learning_rate": 5.200553250345782e-05,
      "loss": 0.0007,
      "step": 5350
    },
    {
      "epoch": 22.24,
      "learning_rate": 5.172890733056709e-05,
      "loss": 0.0006,
      "step": 5360
    },
    {
      "epoch": 22.28,
      "learning_rate": 5.145228215767635e-05,
      "loss": 0.0007,
      "step": 5370
    },
    {
      "epoch": 22.32,
      "learning_rate": 5.1175656984785616e-05,
      "loss": 0.0007,
      "step": 5380
    },
    {
      "epoch": 22.37,
      "learning_rate": 5.089903181189488e-05,
      "loss": 0.0007,
      "step": 5390
    },
    {
      "epoch": 22.41,
      "learning_rate": 5.062240663900415e-05,
      "loss": 0.0007,
      "step": 5400
    },
    {
      "epoch": 22.41,
      "eval_accuracy": 0.966789667896679,
      "eval_f1": 0.8592271925605259,
      "eval_loss": 0.14013563096523285,
      "eval_precision": 0.8910052910052911,
      "eval_recall": 0.951058201058201,
      "eval_runtime": 4.7358,
      "eval_samples_per_second": 57.224,
      "eval_steps_per_second": 7.179,
      "step": 5400
    },
    {
      "epoch": 22.45,
      "learning_rate": 5.034578146611342e-05,
      "loss": 0.0006,
      "step": 5410
    },
    {
      "epoch": 22.49,
      "learning_rate": 5.0069156293222686e-05,
      "loss": 0.0007,
      "step": 5420
    },
    {
      "epoch": 22.53,
      "learning_rate": 4.979253112033195e-05,
      "loss": 0.0006,
      "step": 5430
    },
    {
      "epoch": 22.57,
      "learning_rate": 4.951590594744122e-05,
      "loss": 0.0006,
      "step": 5440
    },
    {
      "epoch": 22.61,
      "learning_rate": 4.923928077455049e-05,
      "loss": 0.0006,
      "step": 5450
    },
    {
      "epoch": 22.66,
      "learning_rate": 4.896265560165975e-05,
      "loss": 0.0007,
      "step": 5460
    },
    {
      "epoch": 22.7,
      "learning_rate": 4.8686030428769016e-05,
      "loss": 0.0007,
      "step": 5470
    },
    {
      "epoch": 22.74,
      "learning_rate": 4.840940525587829e-05,
      "loss": 0.0007,
      "step": 5480
    },
    {
      "epoch": 22.78,
      "learning_rate": 4.813278008298756e-05,
      "loss": 0.0006,
      "step": 5490
    },
    {
      "epoch": 22.82,
      "learning_rate": 4.7856154910096825e-05,
      "loss": 0.0007,
      "step": 5500
    },
    {
      "epoch": 22.82,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14039519429206848,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.0941,
      "eval_samples_per_second": 53.199,
      "eval_steps_per_second": 6.674,
      "step": 5500
    },
    {
      "epoch": 22.86,
      "learning_rate": 4.7579529737206086e-05,
      "loss": 0.0006,
      "step": 5510
    },
    {
      "epoch": 22.9,
      "learning_rate": 4.7302904564315354e-05,
      "loss": 0.0007,
      "step": 5520
    },
    {
      "epoch": 22.95,
      "learning_rate": 4.702627939142462e-05,
      "loss": 0.0006,
      "step": 5530
    },
    {
      "epoch": 22.99,
      "learning_rate": 4.674965421853389e-05,
      "loss": 0.0006,
      "step": 5540
    },
    {
      "epoch": 23.03,
      "learning_rate": 4.6473029045643156e-05,
      "loss": 0.0006,
      "step": 5550
    },
    {
      "epoch": 23.07,
      "learning_rate": 4.6196403872752424e-05,
      "loss": 0.0006,
      "step": 5560
    },
    {
      "epoch": 23.11,
      "learning_rate": 4.5919778699861684e-05,
      "loss": 0.0007,
      "step": 5570
    },
    {
      "epoch": 23.15,
      "learning_rate": 4.564315352697096e-05,
      "loss": 0.0007,
      "step": 5580
    },
    {
      "epoch": 23.2,
      "learning_rate": 4.5366528354080226e-05,
      "loss": 0.0007,
      "step": 5590
    },
    {
      "epoch": 23.24,
      "learning_rate": 4.5089903181189494e-05,
      "loss": 0.0006,
      "step": 5600
    },
    {
      "epoch": 23.24,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.1403643935918808,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.1567,
      "eval_samples_per_second": 52.553,
      "eval_steps_per_second": 6.593,
      "step": 5600
    },
    {
      "epoch": 23.28,
      "learning_rate": 4.481327800829876e-05,
      "loss": 0.0006,
      "step": 5610
    },
    {
      "epoch": 23.32,
      "learning_rate": 4.453665283540803e-05,
      "loss": 0.0006,
      "step": 5620
    },
    {
      "epoch": 23.36,
      "learning_rate": 4.426002766251729e-05,
      "loss": 0.0006,
      "step": 5630
    },
    {
      "epoch": 23.4,
      "learning_rate": 4.398340248962656e-05,
      "loss": 0.0006,
      "step": 5640
    },
    {
      "epoch": 23.44,
      "learning_rate": 4.3706777316735824e-05,
      "loss": 0.0006,
      "step": 5650
    },
    {
      "epoch": 23.49,
      "learning_rate": 4.343015214384509e-05,
      "loss": 0.0006,
      "step": 5660
    },
    {
      "epoch": 23.53,
      "learning_rate": 4.315352697095436e-05,
      "loss": 0.0005,
      "step": 5670
    },
    {
      "epoch": 23.57,
      "learning_rate": 4.287690179806363e-05,
      "loss": 0.0007,
      "step": 5680
    },
    {
      "epoch": 23.61,
      "learning_rate": 4.2600276625172894e-05,
      "loss": 0.0006,
      "step": 5690
    },
    {
      "epoch": 23.65,
      "learning_rate": 4.232365145228216e-05,
      "loss": 0.0006,
      "step": 5700
    },
    {
      "epoch": 23.65,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14015419781208038,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.2932,
      "eval_samples_per_second": 51.198,
      "eval_steps_per_second": 6.423,
      "step": 5700
    },
    {
      "epoch": 23.69,
      "learning_rate": 4.204702627939143e-05,
      "loss": 0.0006,
      "step": 5710
    },
    {
      "epoch": 23.73,
      "learning_rate": 4.17704011065007e-05,
      "loss": 0.0006,
      "step": 5720
    },
    {
      "epoch": 23.78,
      "learning_rate": 4.1493775933609964e-05,
      "loss": 0.0006,
      "step": 5730
    },
    {
      "epoch": 23.82,
      "learning_rate": 4.1217150760719225e-05,
      "loss": 0.0006,
      "step": 5740
    },
    {
      "epoch": 23.86,
      "learning_rate": 4.094052558782849e-05,
      "loss": 0.0006,
      "step": 5750
    },
    {
      "epoch": 23.9,
      "learning_rate": 4.066390041493776e-05,
      "loss": 0.0006,
      "step": 5760
    },
    {
      "epoch": 23.94,
      "learning_rate": 4.038727524204703e-05,
      "loss": 0.0006,
      "step": 5770
    },
    {
      "epoch": 23.98,
      "learning_rate": 4.0110650069156295e-05,
      "loss": 0.0007,
      "step": 5780
    },
    {
      "epoch": 24.02,
      "learning_rate": 3.983402489626556e-05,
      "loss": 0.0007,
      "step": 5790
    },
    {
      "epoch": 24.07,
      "learning_rate": 3.955739972337483e-05,
      "loss": 0.0006,
      "step": 5800
    },
    {
      "epoch": 24.07,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.1410708874464035,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.0905,
      "eval_samples_per_second": 53.236,
      "eval_steps_per_second": 6.679,
      "step": 5800
    },
    {
      "epoch": 24.11,
      "learning_rate": 3.92807745504841e-05,
      "loss": 0.0006,
      "step": 5810
    },
    {
      "epoch": 24.15,
      "learning_rate": 3.9004149377593365e-05,
      "loss": 0.0005,
      "step": 5820
    },
    {
      "epoch": 24.19,
      "learning_rate": 3.872752420470263e-05,
      "loss": 0.0006,
      "step": 5830
    },
    {
      "epoch": 24.23,
      "learning_rate": 3.84508990318119e-05,
      "loss": 0.0005,
      "step": 5840
    },
    {
      "epoch": 24.27,
      "learning_rate": 3.817427385892116e-05,
      "loss": 0.0006,
      "step": 5850
    },
    {
      "epoch": 24.32,
      "learning_rate": 3.789764868603043e-05,
      "loss": 0.0006,
      "step": 5860
    },
    {
      "epoch": 24.36,
      "learning_rate": 3.7621023513139696e-05,
      "loss": 0.0006,
      "step": 5870
    },
    {
      "epoch": 24.4,
      "learning_rate": 3.734439834024896e-05,
      "loss": 0.0006,
      "step": 5880
    },
    {
      "epoch": 24.44,
      "learning_rate": 3.706777316735823e-05,
      "loss": 0.0006,
      "step": 5890
    },
    {
      "epoch": 24.48,
      "learning_rate": 3.67911479944675e-05,
      "loss": 0.0006,
      "step": 5900
    },
    {
      "epoch": 24.48,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14112421870231628,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.101,
      "eval_samples_per_second": 53.127,
      "eval_steps_per_second": 6.665,
      "step": 5900
    },
    {
      "epoch": 24.52,
      "learning_rate": 3.6514522821576766e-05,
      "loss": 0.0006,
      "step": 5910
    },
    {
      "epoch": 24.56,
      "learning_rate": 3.623789764868603e-05,
      "loss": 0.0006,
      "step": 5920
    },
    {
      "epoch": 24.61,
      "learning_rate": 3.59612724757953e-05,
      "loss": 0.0006,
      "step": 5930
    },
    {
      "epoch": 24.65,
      "learning_rate": 3.568464730290457e-05,
      "loss": 0.0005,
      "step": 5940
    },
    {
      "epoch": 24.69,
      "learning_rate": 3.5408022130013836e-05,
      "loss": 0.0006,
      "step": 5950
    },
    {
      "epoch": 24.73,
      "learning_rate": 3.5131396957123096e-05,
      "loss": 0.0006,
      "step": 5960
    },
    {
      "epoch": 24.77,
      "learning_rate": 3.4854771784232364e-05,
      "loss": 0.0006,
      "step": 5970
    },
    {
      "epoch": 24.81,
      "learning_rate": 3.457814661134163e-05,
      "loss": 0.0006,
      "step": 5980
    },
    {
      "epoch": 24.85,
      "learning_rate": 3.43015214384509e-05,
      "loss": 0.0005,
      "step": 5990
    },
    {
      "epoch": 24.9,
      "learning_rate": 3.4024896265560166e-05,
      "loss": 0.0006,
      "step": 6000
    },
    {
      "epoch": 24.9,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14132976531982422,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.221,
      "eval_samples_per_second": 51.906,
      "eval_steps_per_second": 6.512,
      "step": 6000
    },
    {
      "epoch": 24.94,
      "learning_rate": 3.3748271092669434e-05,
      "loss": 0.0006,
      "step": 6010
    },
    {
      "epoch": 24.98,
      "learning_rate": 3.34716459197787e-05,
      "loss": 0.0005,
      "step": 6020
    },
    {
      "epoch": 25.02,
      "learning_rate": 3.319502074688797e-05,
      "loss": 0.0005,
      "step": 6030
    },
    {
      "epoch": 25.06,
      "learning_rate": 3.2918395573997236e-05,
      "loss": 0.0005,
      "step": 6040
    },
    {
      "epoch": 25.1,
      "learning_rate": 3.2641770401106504e-05,
      "loss": 0.0005,
      "step": 6050
    },
    {
      "epoch": 25.15,
      "learning_rate": 3.236514522821577e-05,
      "loss": 0.0006,
      "step": 6060
    },
    {
      "epoch": 25.19,
      "learning_rate": 3.208852005532504e-05,
      "loss": 0.0005,
      "step": 6070
    },
    {
      "epoch": 25.23,
      "learning_rate": 3.18118948824343e-05,
      "loss": 0.0005,
      "step": 6080
    },
    {
      "epoch": 25.27,
      "learning_rate": 3.153526970954357e-05,
      "loss": 0.0006,
      "step": 6090
    },
    {
      "epoch": 25.31,
      "learning_rate": 3.1258644536652834e-05,
      "loss": 0.0005,
      "step": 6100
    },
    {
      "epoch": 25.31,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14183248579502106,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.2903,
      "eval_samples_per_second": 51.226,
      "eval_steps_per_second": 6.427,
      "step": 6100
    },
    {
      "epoch": 25.35,
      "learning_rate": 3.09820193637621e-05,
      "loss": 0.0006,
      "step": 6110
    },
    {
      "epoch": 25.39,
      "learning_rate": 3.070539419087137e-05,
      "loss": 0.0005,
      "step": 6120
    },
    {
      "epoch": 25.44,
      "learning_rate": 3.0428769017980633e-05,
      "loss": 0.0005,
      "step": 6130
    },
    {
      "epoch": 25.48,
      "learning_rate": 3.0152143845089904e-05,
      "loss": 0.0005,
      "step": 6140
    },
    {
      "epoch": 25.52,
      "learning_rate": 2.9875518672199172e-05,
      "loss": 0.0005,
      "step": 6150
    },
    {
      "epoch": 25.56,
      "learning_rate": 2.959889349930844e-05,
      "loss": 0.0005,
      "step": 6160
    },
    {
      "epoch": 25.6,
      "learning_rate": 2.9322268326417707e-05,
      "loss": 0.0005,
      "step": 6170
    },
    {
      "epoch": 25.64,
      "learning_rate": 2.9045643153526974e-05,
      "loss": 0.0006,
      "step": 6180
    },
    {
      "epoch": 25.68,
      "learning_rate": 2.876901798063624e-05,
      "loss": 0.0005,
      "step": 6190
    },
    {
      "epoch": 25.73,
      "learning_rate": 2.8492392807745506e-05,
      "loss": 0.0006,
      "step": 6200
    },
    {
      "epoch": 25.73,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14200928807258606,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 4.806,
      "eval_samples_per_second": 56.387,
      "eval_steps_per_second": 7.074,
      "step": 6200
    },
    {
      "epoch": 25.77,
      "learning_rate": 2.8215767634854773e-05,
      "loss": 0.0005,
      "step": 6210
    },
    {
      "epoch": 25.81,
      "learning_rate": 2.793914246196404e-05,
      "loss": 0.0005,
      "step": 6220
    },
    {
      "epoch": 25.85,
      "learning_rate": 2.766251728907331e-05,
      "loss": 0.0005,
      "step": 6230
    },
    {
      "epoch": 25.89,
      "learning_rate": 2.7385892116182576e-05,
      "loss": 0.0005,
      "step": 6240
    },
    {
      "epoch": 25.93,
      "learning_rate": 2.710926694329184e-05,
      "loss": 0.0006,
      "step": 6250
    },
    {
      "epoch": 25.98,
      "learning_rate": 2.6832641770401107e-05,
      "loss": 0.0005,
      "step": 6260
    },
    {
      "epoch": 26.02,
      "learning_rate": 2.6556016597510375e-05,
      "loss": 0.0005,
      "step": 6270
    },
    {
      "epoch": 26.06,
      "learning_rate": 2.6279391424619642e-05,
      "loss": 0.0005,
      "step": 6280
    },
    {
      "epoch": 26.1,
      "learning_rate": 2.600276625172891e-05,
      "loss": 0.0004,
      "step": 6290
    },
    {
      "epoch": 26.14,
      "learning_rate": 2.5726141078838174e-05,
      "loss": 0.0005,
      "step": 6300
    },
    {
      "epoch": 26.14,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14206776022911072,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 4.7749,
      "eval_samples_per_second": 56.756,
      "eval_steps_per_second": 7.121,
      "step": 6300
    },
    {
      "epoch": 26.18,
      "learning_rate": 2.544951590594744e-05,
      "loss": 0.0005,
      "step": 6310
    },
    {
      "epoch": 26.22,
      "learning_rate": 2.517289073305671e-05,
      "loss": 0.0005,
      "step": 6320
    },
    {
      "epoch": 26.27,
      "learning_rate": 2.4896265560165977e-05,
      "loss": 0.0005,
      "step": 6330
    },
    {
      "epoch": 26.31,
      "learning_rate": 2.4619640387275244e-05,
      "loss": 0.0006,
      "step": 6340
    },
    {
      "epoch": 26.35,
      "learning_rate": 2.4343015214384508e-05,
      "loss": 0.0005,
      "step": 6350
    },
    {
      "epoch": 26.39,
      "learning_rate": 2.406639004149378e-05,
      "loss": 0.0005,
      "step": 6360
    },
    {
      "epoch": 26.43,
      "learning_rate": 2.3789764868603043e-05,
      "loss": 0.0006,
      "step": 6370
    },
    {
      "epoch": 26.47,
      "learning_rate": 2.351313969571231e-05,
      "loss": 0.0005,
      "step": 6380
    },
    {
      "epoch": 26.51,
      "learning_rate": 2.3236514522821578e-05,
      "loss": 0.0006,
      "step": 6390
    },
    {
      "epoch": 26.56,
      "learning_rate": 2.2959889349930842e-05,
      "loss": 0.0005,
      "step": 6400
    },
    {
      "epoch": 26.56,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14228321611881256,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 4.7827,
      "eval_samples_per_second": 56.663,
      "eval_steps_per_second": 7.109,
      "step": 6400
    },
    {
      "epoch": 26.6,
      "learning_rate": 2.2683264177040113e-05,
      "loss": 0.0006,
      "step": 6410
    },
    {
      "epoch": 26.64,
      "learning_rate": 2.240663900414938e-05,
      "loss": 0.0005,
      "step": 6420
    },
    {
      "epoch": 26.68,
      "learning_rate": 2.2130013831258645e-05,
      "loss": 0.0005,
      "step": 6430
    },
    {
      "epoch": 26.72,
      "learning_rate": 2.1853388658367912e-05,
      "loss": 0.0005,
      "step": 6440
    },
    {
      "epoch": 26.76,
      "learning_rate": 2.157676348547718e-05,
      "loss": 0.0005,
      "step": 6450
    },
    {
      "epoch": 26.8,
      "learning_rate": 2.1300138312586447e-05,
      "loss": 0.0005,
      "step": 6460
    },
    {
      "epoch": 26.85,
      "learning_rate": 2.1023513139695715e-05,
      "loss": 0.0005,
      "step": 6470
    },
    {
      "epoch": 26.89,
      "learning_rate": 2.0746887966804982e-05,
      "loss": 0.0005,
      "step": 6480
    },
    {
      "epoch": 26.93,
      "learning_rate": 2.0470262793914246e-05,
      "loss": 0.0005,
      "step": 6490
    },
    {
      "epoch": 26.97,
      "learning_rate": 2.0193637621023514e-05,
      "loss": 0.0005,
      "step": 6500
    },
    {
      "epoch": 26.97,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14237850904464722,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 4.9266,
      "eval_samples_per_second": 55.007,
      "eval_steps_per_second": 6.901,
      "step": 6500
    },
    {
      "epoch": 27.01,
      "learning_rate": 1.991701244813278e-05,
      "loss": 0.0005,
      "step": 6510
    },
    {
      "epoch": 27.05,
      "learning_rate": 1.964038727524205e-05,
      "loss": 0.0004,
      "step": 6520
    },
    {
      "epoch": 27.1,
      "learning_rate": 1.9363762102351316e-05,
      "loss": 0.0005,
      "step": 6530
    },
    {
      "epoch": 27.14,
      "learning_rate": 1.908713692946058e-05,
      "loss": 0.0005,
      "step": 6540
    },
    {
      "epoch": 27.18,
      "learning_rate": 1.8810511756569848e-05,
      "loss": 0.0005,
      "step": 6550
    },
    {
      "epoch": 27.22,
      "learning_rate": 1.8533886583679115e-05,
      "loss": 0.0005,
      "step": 6560
    },
    {
      "epoch": 27.26,
      "learning_rate": 1.8257261410788383e-05,
      "loss": 0.0005,
      "step": 6570
    },
    {
      "epoch": 27.3,
      "learning_rate": 1.798063623789765e-05,
      "loss": 0.0005,
      "step": 6580
    },
    {
      "epoch": 27.34,
      "learning_rate": 1.7704011065006918e-05,
      "loss": 0.0005,
      "step": 6590
    },
    {
      "epoch": 27.39,
      "learning_rate": 1.7427385892116182e-05,
      "loss": 0.0004,
      "step": 6600
    },
    {
      "epoch": 27.39,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.1428203284740448,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.1918,
      "eval_samples_per_second": 52.198,
      "eval_steps_per_second": 6.549,
      "step": 6600
    },
    {
      "epoch": 27.43,
      "learning_rate": 1.715076071922545e-05,
      "loss": 0.0006,
      "step": 6610
    },
    {
      "epoch": 27.47,
      "learning_rate": 1.6874135546334717e-05,
      "loss": 0.0005,
      "step": 6620
    },
    {
      "epoch": 27.51,
      "learning_rate": 1.6597510373443984e-05,
      "loss": 0.0005,
      "step": 6630
    },
    {
      "epoch": 27.55,
      "learning_rate": 1.6320885200553252e-05,
      "loss": 0.0005,
      "step": 6640
    },
    {
      "epoch": 27.59,
      "learning_rate": 1.604426002766252e-05,
      "loss": 0.0005,
      "step": 6650
    },
    {
      "epoch": 27.63,
      "learning_rate": 1.5767634854771783e-05,
      "loss": 0.0004,
      "step": 6660
    },
    {
      "epoch": 27.68,
      "learning_rate": 1.549100968188105e-05,
      "loss": 0.0005,
      "step": 6670
    },
    {
      "epoch": 27.72,
      "learning_rate": 1.5214384508990317e-05,
      "loss": 0.0005,
      "step": 6680
    },
    {
      "epoch": 27.76,
      "learning_rate": 1.4937759336099586e-05,
      "loss": 0.0005,
      "step": 6690
    },
    {
      "epoch": 27.8,
      "learning_rate": 1.4661134163208853e-05,
      "loss": 0.0005,
      "step": 6700
    },
    {
      "epoch": 27.8,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.1428556591272354,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 4.7602,
      "eval_samples_per_second": 56.931,
      "eval_steps_per_second": 7.143,
      "step": 6700
    },
    {
      "epoch": 27.84,
      "learning_rate": 1.438450899031812e-05,
      "loss": 0.0005,
      "step": 6710
    },
    {
      "epoch": 27.88,
      "learning_rate": 1.4107883817427387e-05,
      "loss": 0.0005,
      "step": 6720
    },
    {
      "epoch": 27.93,
      "learning_rate": 1.3831258644536654e-05,
      "loss": 0.0005,
      "step": 6730
    },
    {
      "epoch": 27.97,
      "learning_rate": 1.355463347164592e-05,
      "loss": 0.0005,
      "step": 6740
    },
    {
      "epoch": 28.01,
      "learning_rate": 1.3278008298755187e-05,
      "loss": 0.0005,
      "step": 6750
    },
    {
      "epoch": 28.05,
      "learning_rate": 1.3001383125864455e-05,
      "loss": 0.0005,
      "step": 6760
    },
    {
      "epoch": 28.09,
      "learning_rate": 1.272475795297372e-05,
      "loss": 0.0005,
      "step": 6770
    },
    {
      "epoch": 28.13,
      "learning_rate": 1.2448132780082988e-05,
      "loss": 0.0004,
      "step": 6780
    },
    {
      "epoch": 28.17,
      "learning_rate": 1.2171507607192254e-05,
      "loss": 0.0005,
      "step": 6790
    },
    {
      "epoch": 28.22,
      "learning_rate": 1.1894882434301522e-05,
      "loss": 0.0005,
      "step": 6800
    },
    {
      "epoch": 28.22,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14278294146060944,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.0544,
      "eval_samples_per_second": 53.617,
      "eval_steps_per_second": 6.727,
      "step": 6800
    },
    {
      "epoch": 28.26,
      "learning_rate": 1.1618257261410789e-05,
      "loss": 0.0005,
      "step": 6810
    },
    {
      "epoch": 28.3,
      "learning_rate": 1.1341632088520057e-05,
      "loss": 0.0005,
      "step": 6820
    },
    {
      "epoch": 28.34,
      "learning_rate": 1.1065006915629322e-05,
      "loss": 0.0005,
      "step": 6830
    },
    {
      "epoch": 28.38,
      "learning_rate": 1.078838174273859e-05,
      "loss": 0.0005,
      "step": 6840
    },
    {
      "epoch": 28.42,
      "learning_rate": 1.0511756569847857e-05,
      "loss": 0.0005,
      "step": 6850
    },
    {
      "epoch": 28.46,
      "learning_rate": 1.0235131396957123e-05,
      "loss": 0.0004,
      "step": 6860
    },
    {
      "epoch": 28.51,
      "learning_rate": 9.95850622406639e-06,
      "loss": 0.0005,
      "step": 6870
    },
    {
      "epoch": 28.55,
      "learning_rate": 9.681881051175658e-06,
      "loss": 0.0005,
      "step": 6880
    },
    {
      "epoch": 28.59,
      "learning_rate": 9.405255878284924e-06,
      "loss": 0.0004,
      "step": 6890
    },
    {
      "epoch": 28.63,
      "learning_rate": 9.128630705394191e-06,
      "loss": 0.0005,
      "step": 6900
    },
    {
      "epoch": 28.63,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.1430046111345291,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.1494,
      "eval_samples_per_second": 52.628,
      "eval_steps_per_second": 6.603,
      "step": 6900
    },
    {
      "epoch": 28.67,
      "learning_rate": 8.852005532503459e-06,
      "loss": 0.0005,
      "step": 6910
    },
    {
      "epoch": 28.71,
      "learning_rate": 8.575380359612725e-06,
      "loss": 0.0005,
      "step": 6920
    },
    {
      "epoch": 28.76,
      "learning_rate": 8.298755186721992e-06,
      "loss": 0.0005,
      "step": 6930
    },
    {
      "epoch": 28.8,
      "learning_rate": 8.02213001383126e-06,
      "loss": 0.0004,
      "step": 6940
    },
    {
      "epoch": 28.84,
      "learning_rate": 7.745504840940525e-06,
      "loss": 0.0004,
      "step": 6950
    },
    {
      "epoch": 28.88,
      "learning_rate": 7.468879668049793e-06,
      "loss": 0.0005,
      "step": 6960
    },
    {
      "epoch": 28.92,
      "learning_rate": 7.19225449515906e-06,
      "loss": 0.0005,
      "step": 6970
    },
    {
      "epoch": 28.96,
      "learning_rate": 6.915629322268327e-06,
      "loss": 0.0005,
      "step": 6980
    },
    {
      "epoch": 29.0,
      "learning_rate": 6.639004149377594e-06,
      "loss": 0.0004,
      "step": 6990
    },
    {
      "epoch": 29.05,
      "learning_rate": 6.36237897648686e-06,
      "loss": 0.0005,
      "step": 7000
    },
    {
      "epoch": 29.05,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14295929670333862,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.4686,
      "eval_samples_per_second": 49.556,
      "eval_steps_per_second": 6.217,
      "step": 7000
    },
    {
      "epoch": 29.09,
      "learning_rate": 6.085753803596127e-06,
      "loss": 0.0004,
      "step": 7010
    },
    {
      "epoch": 29.13,
      "learning_rate": 5.8091286307053945e-06,
      "loss": 0.0005,
      "step": 7020
    },
    {
      "epoch": 29.17,
      "learning_rate": 5.532503457814661e-06,
      "loss": 0.0005,
      "step": 7030
    },
    {
      "epoch": 29.21,
      "learning_rate": 5.255878284923929e-06,
      "loss": 0.0005,
      "step": 7040
    },
    {
      "epoch": 29.25,
      "learning_rate": 4.979253112033195e-06,
      "loss": 0.0005,
      "step": 7050
    },
    {
      "epoch": 29.29,
      "learning_rate": 4.702627939142462e-06,
      "loss": 0.0005,
      "step": 7060
    },
    {
      "epoch": 29.34,
      "learning_rate": 4.4260027662517294e-06,
      "loss": 0.0005,
      "step": 7070
    },
    {
      "epoch": 29.38,
      "learning_rate": 4.149377593360996e-06,
      "loss": 0.0004,
      "step": 7080
    },
    {
      "epoch": 29.42,
      "learning_rate": 3.872752420470263e-06,
      "loss": 0.0005,
      "step": 7090
    },
    {
      "epoch": 29.46,
      "learning_rate": 3.59612724757953e-06,
      "loss": 0.0005,
      "step": 7100
    },
    {
      "epoch": 29.46,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14300456643104553,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 5.0306,
      "eval_samples_per_second": 53.87,
      "eval_steps_per_second": 6.759,
      "step": 7100
    },
    {
      "epoch": 29.5,
      "learning_rate": 3.319502074688797e-06,
      "loss": 0.0004,
      "step": 7110
    },
    {
      "epoch": 29.54,
      "learning_rate": 3.0428769017980635e-06,
      "loss": 0.0004,
      "step": 7120
    },
    {
      "epoch": 29.59,
      "learning_rate": 2.7662517289073306e-06,
      "loss": 0.0005,
      "step": 7130
    },
    {
      "epoch": 29.63,
      "learning_rate": 2.4896265560165977e-06,
      "loss": 0.0005,
      "step": 7140
    },
    {
      "epoch": 29.67,
      "learning_rate": 2.2130013831258647e-06,
      "loss": 0.0005,
      "step": 7150
    },
    {
      "epoch": 29.71,
      "learning_rate": 1.9363762102351314e-06,
      "loss": 0.0004,
      "step": 7160
    },
    {
      "epoch": 29.75,
      "learning_rate": 1.6597510373443984e-06,
      "loss": 0.0004,
      "step": 7170
    },
    {
      "epoch": 29.79,
      "learning_rate": 1.3831258644536653e-06,
      "loss": 0.0005,
      "step": 7180
    },
    {
      "epoch": 29.83,
      "learning_rate": 1.1065006915629324e-06,
      "loss": 0.0005,
      "step": 7190
    },
    {
      "epoch": 29.88,
      "learning_rate": 8.298755186721992e-07,
      "loss": 0.0005,
      "step": 7200
    },
    {
      "epoch": 29.88,
      "eval_accuracy": 0.9704797047970479,
      "eval_f1": 0.8666345999679334,
      "eval_loss": 0.14301735162734985,
      "eval_precision": 0.8962962962962964,
      "eval_recall": 0.955026455026455,
      "eval_runtime": 4.7163,
      "eval_samples_per_second": 57.46,
      "eval_steps_per_second": 7.209,
      "step": 7200
    },
    {
      "epoch": 29.92,
      "learning_rate": 5.532503457814662e-07,
      "loss": 0.0005,
      "step": 7210
    },
    {
      "epoch": 29.96,
      "learning_rate": 2.766251728907331e-07,
      "loss": 0.0004,
      "step": 7220
    },
    {
      "epoch": 30.0,
      "learning_rate": 0.0,
      "loss": 0.0005,
      "step": 7230
    },
    {
      "epoch": 30.0,
      "step": 7230,
      "total_flos": 8.952324879334072e+18,
      "train_loss": 0.18286644016659173,
      "train_runtime": 8158.6801,
      "train_samples_per_second": 14.142,
      "train_steps_per_second": 0.886
    }
  ],
  "max_steps": 7230,
  "num_train_epochs": 30,
  "total_flos": 8.952324879334072e+18,
  "trial_name": null,
  "trial_params": null
}