{
  "best_metric": 0.05167504400014877,
  "best_model_checkpoint": "./swim2-base-model/checkpoint-6000",
  "epoch": 3.0,
  "eval_steps": 100,
  "global_step": 6564,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 7.610350076103501e-07,
      "loss": 0.7091,
      "step": 10
    },
    {
      "epoch": 0.01,
      "learning_rate": 1.5220700152207001e-06,
      "loss": 0.7069,
      "step": 20
    },
    {
      "epoch": 0.01,
      "learning_rate": 2.2831050228310503e-06,
      "loss": 0.7092,
      "step": 30
    },
    {
      "epoch": 0.02,
      "learning_rate": 3.0441400304414002e-06,
      "loss": 0.6853,
      "step": 40
    },
    {
      "epoch": 0.02,
      "learning_rate": 3.80517503805175e-06,
      "loss": 0.6791,
      "step": 50
    },
    {
      "epoch": 0.03,
      "learning_rate": 4.566210045662101e-06,
      "loss": 0.678,
      "step": 60
    },
    {
      "epoch": 0.03,
      "learning_rate": 5.32724505327245e-06,
      "loss": 0.6356,
      "step": 70
    },
    {
      "epoch": 0.04,
      "learning_rate": 6.0882800608828005e-06,
      "loss": 0.6258,
      "step": 80
    },
    {
      "epoch": 0.04,
      "learning_rate": 6.849315068493151e-06,
      "loss": 0.6107,
      "step": 90
    },
    {
      "epoch": 0.05,
      "learning_rate": 7.6103500761035e-06,
      "loss": 0.5804,
      "step": 100
    },
    {
      "epoch": 0.05,
      "eval_accuracy": 0.7388454335132754,
      "eval_loss": 0.5394906997680664,
      "eval_runtime": 86.631,
      "eval_samples_per_second": 419.111,
      "eval_steps_per_second": 6.557,
      "step": 100
    },
    {
      "epoch": 0.05,
      "learning_rate": 8.37138508371385e-06,
      "loss": 0.5194,
      "step": 110
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.132420091324201e-06,
      "loss": 0.4818,
      "step": 120
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.89345509893455e-06,
      "loss": 0.431,
      "step": 130
    },
    {
      "epoch": 0.06,
      "learning_rate": 1.06544901065449e-05,
      "loss": 0.3844,
      "step": 140
    },
    {
      "epoch": 0.07,
      "learning_rate": 1.1415525114155251e-05,
      "loss": 0.3404,
      "step": 150
    },
    {
      "epoch": 0.07,
      "learning_rate": 1.2100456621004567e-05,
      "loss": 0.3175,
      "step": 160
    },
    {
      "epoch": 0.08,
      "learning_rate": 1.2861491628614916e-05,
      "loss": 0.2857,
      "step": 170
    },
    {
      "epoch": 0.08,
      "learning_rate": 1.3622526636225266e-05,
      "loss": 0.2735,
      "step": 180
    },
    {
      "epoch": 0.09,
      "learning_rate": 1.4383561643835617e-05,
      "loss": 0.2653,
      "step": 190
    },
    {
      "epoch": 0.09,
      "learning_rate": 1.5144596651445967e-05,
      "loss": 0.2244,
      "step": 200
    },
    {
      "epoch": 0.09,
      "eval_accuracy": 0.8787319598986449,
      "eval_loss": 0.30570992827415466,
      "eval_runtime": 88.9341,
      "eval_samples_per_second": 408.257,
      "eval_steps_per_second": 6.387,
      "step": 200
    },
    {
      "epoch": 0.1,
      "learning_rate": 1.5905631659056316e-05,
      "loss": 0.2628,
      "step": 210
    },
    {
      "epoch": 0.1,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.2083,
      "step": 220
    },
    {
      "epoch": 0.11,
      "learning_rate": 1.7427701674277015e-05,
      "loss": 0.2146,
      "step": 230
    },
    {
      "epoch": 0.11,
      "learning_rate": 1.8188736681887367e-05,
      "loss": 0.2373,
      "step": 240
    },
    {
      "epoch": 0.11,
      "learning_rate": 1.8949771689497718e-05,
      "loss": 0.1777,
      "step": 250
    },
    {
      "epoch": 0.12,
      "learning_rate": 1.9710806697108066e-05,
      "loss": 0.1866,
      "step": 260
    },
    {
      "epoch": 0.12,
      "learning_rate": 2.0471841704718417e-05,
      "loss": 0.1661,
      "step": 270
    },
    {
      "epoch": 0.13,
      "learning_rate": 2.1232876712328768e-05,
      "loss": 0.1571,
      "step": 280
    },
    {
      "epoch": 0.13,
      "learning_rate": 2.1993911719939116e-05,
      "loss": 0.2228,
      "step": 290
    },
    {
      "epoch": 0.14,
      "learning_rate": 2.2754946727549467e-05,
      "loss": 0.134,
      "step": 300
    },
    {
      "epoch": 0.14,
      "eval_accuracy": 0.9129392971246006,
      "eval_loss": 0.22183451056480408,
      "eval_runtime": 86.9027,
      "eval_samples_per_second": 417.801,
      "eval_steps_per_second": 6.536,
      "step": 300
    },
    {
      "epoch": 0.14,
      "learning_rate": 2.351598173515982e-05,
      "loss": 0.1564,
      "step": 310
    },
    {
      "epoch": 0.15,
      "learning_rate": 2.4277016742770166e-05,
      "loss": 0.1485,
      "step": 320
    },
    {
      "epoch": 0.15,
      "learning_rate": 2.503805175038052e-05,
      "loss": 0.1303,
      "step": 330
    },
    {
      "epoch": 0.16,
      "learning_rate": 2.579908675799087e-05,
      "loss": 0.1608,
      "step": 340
    },
    {
      "epoch": 0.16,
      "learning_rate": 2.656012176560122e-05,
      "loss": 0.144,
      "step": 350
    },
    {
      "epoch": 0.16,
      "learning_rate": 2.732115677321157e-05,
      "loss": 0.1399,
      "step": 360
    },
    {
      "epoch": 0.17,
      "learning_rate": 2.808219178082192e-05,
      "loss": 0.153,
      "step": 370
    },
    {
      "epoch": 0.17,
      "learning_rate": 2.884322678843227e-05,
      "loss": 0.1343,
      "step": 380
    },
    {
      "epoch": 0.18,
      "learning_rate": 2.960426179604262e-05,
      "loss": 0.1478,
      "step": 390
    },
    {
      "epoch": 0.18,
      "learning_rate": 3.036529680365297e-05,
      "loss": 0.1786,
      "step": 400
    },
    {
      "epoch": 0.18,
      "eval_accuracy": 0.9373140905585546,
      "eval_loss": 0.15671265125274658,
      "eval_runtime": 87.4216,
      "eval_samples_per_second": 415.321,
      "eval_steps_per_second": 6.497,
      "step": 400
    },
    {
      "epoch": 0.19,
      "learning_rate": 3.112633181126332e-05,
      "loss": 0.1249,
      "step": 410
    },
    {
      "epoch": 0.19,
      "learning_rate": 3.188736681887367e-05,
      "loss": 0.1067,
      "step": 420
    },
    {
      "epoch": 0.2,
      "learning_rate": 3.264840182648402e-05,
      "loss": 0.1295,
      "step": 430
    },
    {
      "epoch": 0.2,
      "learning_rate": 3.3409436834094374e-05,
      "loss": 0.1676,
      "step": 440
    },
    {
      "epoch": 0.21,
      "learning_rate": 3.417047184170472e-05,
      "loss": 0.1248,
      "step": 450
    },
    {
      "epoch": 0.21,
      "learning_rate": 3.493150684931507e-05,
      "loss": 0.1196,
      "step": 460
    },
    {
      "epoch": 0.21,
      "learning_rate": 3.569254185692542e-05,
      "loss": 0.1423,
      "step": 470
    },
    {
      "epoch": 0.22,
      "learning_rate": 3.645357686453577e-05,
      "loss": 0.1194,
      "step": 480
    },
    {
      "epoch": 0.22,
      "learning_rate": 3.7214611872146123e-05,
      "loss": 0.1422,
      "step": 490
    },
    {
      "epoch": 0.23,
      "learning_rate": 3.797564687975647e-05,
      "loss": 0.0924,
      "step": 500
    },
    {
      "epoch": 0.23,
      "eval_accuracy": 0.9463754544453014,
      "eval_loss": 0.1360485702753067,
      "eval_runtime": 87.9764,
      "eval_samples_per_second": 412.702,
      "eval_steps_per_second": 6.456,
      "step": 500
    },
    {
      "epoch": 0.23,
      "learning_rate": 3.873668188736682e-05,
      "loss": 0.12,
      "step": 510
    },
    {
      "epoch": 0.24,
      "learning_rate": 3.949771689497717e-05,
      "loss": 0.1355,
      "step": 520
    },
    {
      "epoch": 0.24,
      "learning_rate": 4.025875190258752e-05,
      "loss": 0.197,
      "step": 530
    },
    {
      "epoch": 0.25,
      "learning_rate": 4.101978691019787e-05,
      "loss": 0.1059,
      "step": 540
    },
    {
      "epoch": 0.25,
      "learning_rate": 4.1780821917808224e-05,
      "loss": 0.1446,
      "step": 550
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.254185692541857e-05,
      "loss": 0.1277,
      "step": 560
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.330289193302892e-05,
      "loss": 0.0974,
      "step": 570
    },
    {
      "epoch": 0.27,
      "learning_rate": 4.406392694063927e-05,
      "loss": 0.1112,
      "step": 580
    },
    {
      "epoch": 0.27,
      "learning_rate": 4.482496194824962e-05,
      "loss": 0.1206,
      "step": 590
    },
    {
      "epoch": 0.27,
      "learning_rate": 4.5585996955859973e-05,
      "loss": 0.1217,
      "step": 600
    },
    {
      "epoch": 0.27,
      "eval_accuracy": 0.9349179244243693,
      "eval_loss": 0.17323894798755646,
      "eval_runtime": 88.1283,
      "eval_samples_per_second": 411.99,
      "eval_steps_per_second": 6.445,
      "step": 600
    },
    {
      "epoch": 0.28,
      "learning_rate": 4.6347031963470325e-05,
      "loss": 0.1138,
      "step": 610
    },
    {
      "epoch": 0.28,
      "learning_rate": 4.710806697108067e-05,
      "loss": 0.1118,
      "step": 620
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.786910197869102e-05,
      "loss": 0.1389,
      "step": 630
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.863013698630137e-05,
      "loss": 0.084,
      "step": 640
    },
    {
      "epoch": 0.3,
      "learning_rate": 4.939117199391172e-05,
      "loss": 0.1383,
      "step": 650
    },
    {
      "epoch": 0.3,
      "learning_rate": 4.99830709327916e-05,
      "loss": 0.1075,
      "step": 660
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.989842559674962e-05,
      "loss": 0.1136,
      "step": 670
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.981378026070764e-05,
      "loss": 0.0847,
      "step": 680
    },
    {
      "epoch": 0.32,
      "learning_rate": 4.9729134924665656e-05,
      "loss": 0.0758,
      "step": 690
    },
    {
      "epoch": 0.32,
      "learning_rate": 4.964448958862367e-05,
      "loss": 0.144,
      "step": 700
    },
    {
      "epoch": 0.32,
      "eval_accuracy": 0.953756747824171,
      "eval_loss": 0.12332048267126083,
      "eval_runtime": 87.5844,
      "eval_samples_per_second": 414.549,
      "eval_steps_per_second": 6.485,
      "step": 700
    },
    {
      "epoch": 0.32,
      "learning_rate": 4.955984425258168e-05,
      "loss": 0.1051,
      "step": 710
    },
    {
      "epoch": 0.33,
      "learning_rate": 4.94751989165397e-05,
      "loss": 0.082,
      "step": 720
    },
    {
      "epoch": 0.33,
      "learning_rate": 4.939055358049772e-05,
      "loss": 0.1475,
      "step": 730
    },
    {
      "epoch": 0.34,
      "learning_rate": 4.9305908244455735e-05,
      "loss": 0.1063,
      "step": 740
    },
    {
      "epoch": 0.34,
      "learning_rate": 4.922126290841375e-05,
      "loss": 0.0788,
      "step": 750
    },
    {
      "epoch": 0.35,
      "learning_rate": 4.913661757237176e-05,
      "loss": 0.0857,
      "step": 760
    },
    {
      "epoch": 0.35,
      "learning_rate": 4.905197223632978e-05,
      "loss": 0.0774,
      "step": 770
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.89673269002878e-05,
      "loss": 0.0949,
      "step": 780
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.8882681564245814e-05,
      "loss": 0.0833,
      "step": 790
    },
    {
      "epoch": 0.37,
      "learning_rate": 4.879803622820383e-05,
      "loss": 0.0917,
      "step": 800
    },
    {
      "epoch": 0.37,
      "eval_accuracy": 0.9379475597664426,
      "eval_loss": 0.16546840965747833,
      "eval_runtime": 86.576,
      "eval_samples_per_second": 419.377,
      "eval_steps_per_second": 6.561,
      "step": 800
    },
    {
      "epoch": 0.37,
      "learning_rate": 4.871339089216184e-05,
      "loss": 0.0699,
      "step": 810
    },
    {
      "epoch": 0.37,
      "learning_rate": 4.862874555611986e-05,
      "loss": 0.0976,
      "step": 820
    },
    {
      "epoch": 0.38,
      "learning_rate": 4.854410022007788e-05,
      "loss": 0.0966,
      "step": 830
    },
    {
      "epoch": 0.38,
      "learning_rate": 4.845945488403589e-05,
      "loss": 0.1021,
      "step": 840
    },
    {
      "epoch": 0.39,
      "learning_rate": 4.837480954799391e-05,
      "loss": 0.0786,
      "step": 850
    },
    {
      "epoch": 0.39,
      "learning_rate": 4.8290164211951925e-05,
      "loss": 0.1091,
      "step": 860
    },
    {
      "epoch": 0.4,
      "learning_rate": 4.820551887590994e-05,
      "loss": 0.1209,
      "step": 870
    },
    {
      "epoch": 0.4,
      "learning_rate": 4.8120873539867957e-05,
      "loss": 0.0724,
      "step": 880
    },
    {
      "epoch": 0.41,
      "learning_rate": 4.803622820382597e-05,
      "loss": 0.1295,
      "step": 890
    },
    {
      "epoch": 0.41,
      "learning_rate": 4.795158286778399e-05,
      "loss": 0.1005,
      "step": 900
    },
    {
      "epoch": 0.41,
      "eval_accuracy": 0.9632037016635452,
      "eval_loss": 0.10467371344566345,
      "eval_runtime": 88.5612,
      "eval_samples_per_second": 409.976,
      "eval_steps_per_second": 6.414,
      "step": 900
    },
    {
      "epoch": 0.42,
      "learning_rate": 4.7866937531742004e-05,
      "loss": 0.0868,
      "step": 910
    },
    {
      "epoch": 0.42,
      "learning_rate": 4.778229219570002e-05,
      "loss": 0.0964,
      "step": 920
    },
    {
      "epoch": 0.43,
      "learning_rate": 4.7697646859658036e-05,
      "loss": 0.0846,
      "step": 930
    },
    {
      "epoch": 0.43,
      "learning_rate": 4.761300152361605e-05,
      "loss": 0.0617,
      "step": 940
    },
    {
      "epoch": 0.43,
      "learning_rate": 4.752835618757407e-05,
      "loss": 0.0833,
      "step": 950
    },
    {
      "epoch": 0.44,
      "learning_rate": 4.744371085153208e-05,
      "loss": 0.1064,
      "step": 960
    },
    {
      "epoch": 0.44,
      "learning_rate": 4.73590655154901e-05,
      "loss": 0.0904,
      "step": 970
    },
    {
      "epoch": 0.45,
      "learning_rate": 4.7274420179448115e-05,
      "loss": 0.0695,
      "step": 980
    },
    {
      "epoch": 0.45,
      "learning_rate": 4.718977484340613e-05,
      "loss": 0.0643,
      "step": 990
    },
    {
      "epoch": 0.46,
      "learning_rate": 4.7105129507364146e-05,
      "loss": 0.1391,
      "step": 1000
    },
    {
      "epoch": 0.46,
      "eval_accuracy": 0.9554092761925747,
      "eval_loss": 0.12814238667488098,
      "eval_runtime": 87.1578,
      "eval_samples_per_second": 416.578,
      "eval_steps_per_second": 6.517,
      "step": 1000
    },
    {
      "epoch": 0.46,
      "learning_rate": 4.702048417132216e-05,
      "loss": 0.0779,
      "step": 1010
    },
    {
      "epoch": 0.47,
      "learning_rate": 4.693583883528018e-05,
      "loss": 0.0691,
      "step": 1020
    },
    {
      "epoch": 0.47,
      "learning_rate": 4.6851193499238194e-05,
      "loss": 0.08,
      "step": 1030
    },
    {
      "epoch": 0.48,
      "learning_rate": 4.676654816319621e-05,
      "loss": 0.0868,
      "step": 1040
    },
    {
      "epoch": 0.48,
      "learning_rate": 4.6681902827154226e-05,
      "loss": 0.0838,
      "step": 1050
    },
    {
      "epoch": 0.48,
      "learning_rate": 4.659725749111224e-05,
      "loss": 0.082,
      "step": 1060
    },
    {
      "epoch": 0.49,
      "learning_rate": 4.651261215507026e-05,
      "loss": 0.1095,
      "step": 1070
    },
    {
      "epoch": 0.49,
      "learning_rate": 4.642796681902827e-05,
      "loss": 0.0835,
      "step": 1080
    },
    {
      "epoch": 0.5,
      "learning_rate": 4.634332148298629e-05,
      "loss": 0.072,
      "step": 1090
    },
    {
      "epoch": 0.5,
      "learning_rate": 4.6258676146944305e-05,
      "loss": 0.0488,
      "step": 1100
    },
    {
      "epoch": 0.5,
      "eval_accuracy": 0.9687947559766442,
      "eval_loss": 0.09652336686849594,
      "eval_runtime": 87.5976,
      "eval_samples_per_second": 414.486,
      "eval_steps_per_second": 6.484,
      "step": 1100
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.617403081090232e-05,
      "loss": 0.0777,
      "step": 1110
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.6089385474860336e-05,
      "loss": 0.0984,
      "step": 1120
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.600474013881835e-05,
      "loss": 0.0871,
      "step": 1130
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.592009480277637e-05,
      "loss": 0.0671,
      "step": 1140
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.5835449466734384e-05,
      "loss": 0.0977,
      "step": 1150
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.57508041306924e-05,
      "loss": 0.1014,
      "step": 1160
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.5666158794650416e-05,
      "loss": 0.0866,
      "step": 1170
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.558151345860843e-05,
      "loss": 0.0788,
      "step": 1180
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.549686812256645e-05,
      "loss": 0.0556,
      "step": 1190
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.541222278652447e-05,
      "loss": 0.0587,
      "step": 1200
    },
    {
      "epoch": 0.55,
      "eval_accuracy": 0.9459623223532004,
      "eval_loss": 0.19257286190986633,
      "eval_runtime": 90.3559,
      "eval_samples_per_second": 401.833,
      "eval_steps_per_second": 6.286,
      "step": 1200
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.532757745048248e-05,
      "loss": 0.0674,
      "step": 1210
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.5242932114440495e-05,
      "loss": 0.1255,
      "step": 1220
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.515828677839851e-05,
      "loss": 0.087,
      "step": 1230
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.5073641442356526e-05,
      "loss": 0.0726,
      "step": 1240
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.498899610631455e-05,
      "loss": 0.0967,
      "step": 1250
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.490435077027256e-05,
      "loss": 0.1071,
      "step": 1260
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.4819705434230574e-05,
      "loss": 0.0541,
      "step": 1270
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.473506009818859e-05,
      "loss": 0.0595,
      "step": 1280
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.4650414762146606e-05,
      "loss": 0.0771,
      "step": 1290
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.456576942610463e-05,
      "loss": 0.0827,
      "step": 1300
    },
    {
      "epoch": 0.59,
      "eval_accuracy": 0.9629833645477581,
      "eval_loss": 0.098239004611969,
      "eval_runtime": 91.5385,
      "eval_samples_per_second": 396.642,
      "eval_steps_per_second": 6.205,
      "step": 1300
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.448112409006264e-05,
      "loss": 0.0613,
      "step": 1310
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.439647875402065e-05,
      "loss": 0.0623,
      "step": 1320
    },
    {
      "epoch": 0.61,
      "learning_rate": 4.431183341797867e-05,
      "loss": 0.0607,
      "step": 1330
    },
    {
      "epoch": 0.61,
      "learning_rate": 4.4227188081936685e-05,
      "loss": 0.0677,
      "step": 1340
    },
    {
      "epoch": 0.62,
      "learning_rate": 4.414254274589471e-05,
      "loss": 0.0592,
      "step": 1350
    },
    {
      "epoch": 0.62,
      "learning_rate": 4.405789740985272e-05,
      "loss": 0.0425,
      "step": 1360
    },
    {
      "epoch": 0.63,
      "learning_rate": 4.397325207381073e-05,
      "loss": 0.0447,
      "step": 1370
    },
    {
      "epoch": 0.63,
      "learning_rate": 4.388860673776875e-05,
      "loss": 0.0375,
      "step": 1380
    },
    {
      "epoch": 0.64,
      "learning_rate": 4.380396140172677e-05,
      "loss": 0.0917,
      "step": 1390
    },
    {
      "epoch": 0.64,
      "learning_rate": 4.3719316065684786e-05,
      "loss": 0.035,
      "step": 1400
    },
    {
      "epoch": 0.64,
      "eval_accuracy": 0.9675553597003416,
      "eval_loss": 0.10114078968763351,
      "eval_runtime": 89.3744,
      "eval_samples_per_second": 406.246,
      "eval_steps_per_second": 6.355,
      "step": 1400
    },
    {
      "epoch": 0.64,
      "learning_rate": 4.36346707296428e-05,
      "loss": 0.0759,
      "step": 1410
    },
    {
      "epoch": 0.65,
      "learning_rate": 4.355002539360081e-05,
      "loss": 0.119,
      "step": 1420
    },
    {
      "epoch": 0.65,
      "learning_rate": 4.346538005755883e-05,
      "loss": 0.0736,
      "step": 1430
    },
    {
      "epoch": 0.66,
      "learning_rate": 4.338073472151685e-05,
      "loss": 0.0765,
      "step": 1440
    },
    {
      "epoch": 0.66,
      "learning_rate": 4.3296089385474866e-05,
      "loss": 0.0843,
      "step": 1450
    },
    {
      "epoch": 0.67,
      "learning_rate": 4.321144404943288e-05,
      "loss": 0.0725,
      "step": 1460
    },
    {
      "epoch": 0.67,
      "learning_rate": 4.312679871339089e-05,
      "loss": 0.0672,
      "step": 1470
    },
    {
      "epoch": 0.68,
      "learning_rate": 4.3042153377348906e-05,
      "loss": 0.095,
      "step": 1480
    },
    {
      "epoch": 0.68,
      "learning_rate": 4.295750804130693e-05,
      "loss": 0.0722,
      "step": 1490
    },
    {
      "epoch": 0.69,
      "learning_rate": 4.2872862705264945e-05,
      "loss": 0.0529,
      "step": 1500
    },
    {
      "epoch": 0.69,
      "eval_accuracy": 0.964580808637215,
      "eval_loss": 0.09844555705785751,
      "eval_runtime": 90.2127,
      "eval_samples_per_second": 402.471,
      "eval_steps_per_second": 6.296,
      "step": 1500
    },
    {
      "epoch": 0.69,
      "learning_rate": 4.278821736922296e-05,
      "loss": 0.0694,
      "step": 1510
    },
    {
      "epoch": 0.69,
      "learning_rate": 4.270357203318097e-05,
      "loss": 0.1012,
      "step": 1520
    },
    {
      "epoch": 0.7,
      "learning_rate": 4.261892669713899e-05,
      "loss": 0.1021,
      "step": 1530
    },
    {
      "epoch": 0.7,
      "learning_rate": 4.253428136109701e-05,
      "loss": 0.0851,
      "step": 1540
    },
    {
      "epoch": 0.71,
      "learning_rate": 4.2449636025055024e-05,
      "loss": 0.0808,
      "step": 1550
    },
    {
      "epoch": 0.71,
      "learning_rate": 4.236499068901304e-05,
      "loss": 0.0546,
      "step": 1560
    },
    {
      "epoch": 0.72,
      "learning_rate": 4.228034535297105e-05,
      "loss": 0.0437,
      "step": 1570
    },
    {
      "epoch": 0.72,
      "learning_rate": 4.219570001692907e-05,
      "loss": 0.0547,
      "step": 1580
    },
    {
      "epoch": 0.73,
      "learning_rate": 4.211105468088709e-05,
      "loss": 0.0688,
      "step": 1590
    },
    {
      "epoch": 0.73,
      "learning_rate": 4.20264093448451e-05,
      "loss": 0.0653,
      "step": 1600
    },
    {
      "epoch": 0.73,
      "eval_accuracy": 0.9666464690977196,
      "eval_loss": 0.08774091303348541,
      "eval_runtime": 89.6473,
      "eval_samples_per_second": 405.009,
      "eval_steps_per_second": 6.336,
      "step": 1600
    },
    {
      "epoch": 0.74,
      "learning_rate": 4.194176400880312e-05,
      "loss": 0.0586,
      "step": 1610
    },
    {
      "epoch": 0.74,
      "learning_rate": 4.185711867276113e-05,
      "loss": 0.0635,
      "step": 1620
    },
    {
      "epoch": 0.74,
      "learning_rate": 4.177247333671915e-05,
      "loss": 0.0916,
      "step": 1630
    },
    {
      "epoch": 0.75,
      "learning_rate": 4.1687828000677166e-05,
      "loss": 0.0587,
      "step": 1640
    },
    {
      "epoch": 0.75,
      "learning_rate": 4.160318266463518e-05,
      "loss": 0.0568,
      "step": 1650
    },
    {
      "epoch": 0.76,
      "learning_rate": 4.15185373285932e-05,
      "loss": 0.0532,
      "step": 1660
    },
    {
      "epoch": 0.76,
      "learning_rate": 4.143389199255121e-05,
      "loss": 0.069,
      "step": 1670
    },
    {
      "epoch": 0.77,
      "learning_rate": 4.134924665650923e-05,
      "loss": 0.0345,
      "step": 1680
    },
    {
      "epoch": 0.77,
      "learning_rate": 4.1264601320467246e-05,
      "loss": 0.1074,
      "step": 1690
    },
    {
      "epoch": 0.78,
      "learning_rate": 4.117995598442526e-05,
      "loss": 0.0749,
      "step": 1700
    },
    {
      "epoch": 0.78,
      "eval_accuracy": 0.9603668612977856,
      "eval_loss": 0.12081305682659149,
      "eval_runtime": 89.5626,
      "eval_samples_per_second": 405.392,
      "eval_steps_per_second": 6.342,
      "step": 1700
    },
    {
      "epoch": 0.78,
      "learning_rate": 4.109531064838328e-05,
      "loss": 0.0744,
      "step": 1710
    },
    {
      "epoch": 0.79,
      "learning_rate": 4.101066531234129e-05,
      "loss": 0.0391,
      "step": 1720
    },
    {
      "epoch": 0.79,
      "learning_rate": 4.092601997629931e-05,
      "loss": 0.0633,
      "step": 1730
    },
    {
      "epoch": 0.8,
      "learning_rate": 4.0841374640257325e-05,
      "loss": 0.0591,
      "step": 1740
    },
    {
      "epoch": 0.8,
      "learning_rate": 4.075672930421534e-05,
      "loss": 0.0438,
      "step": 1750
    },
    {
      "epoch": 0.8,
      "learning_rate": 4.0672083968173356e-05,
      "loss": 0.0598,
      "step": 1760
    },
    {
      "epoch": 0.81,
      "learning_rate": 4.058743863213137e-05,
      "loss": 0.0703,
      "step": 1770
    },
    {
      "epoch": 0.81,
      "learning_rate": 4.050279329608939e-05,
      "loss": 0.0369,
      "step": 1780
    },
    {
      "epoch": 0.82,
      "learning_rate": 4.0418147960047404e-05,
      "loss": 0.0808,
      "step": 1790
    },
    {
      "epoch": 0.82,
      "learning_rate": 4.033350262400542e-05,
      "loss": 0.0686,
      "step": 1800
    },
    {
      "epoch": 0.82,
      "eval_accuracy": 0.9719070177371378,
      "eval_loss": 0.07419364154338837,
      "eval_runtime": 88.1813,
      "eval_samples_per_second": 411.743,
      "eval_steps_per_second": 6.441,
      "step": 1800
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.0248857287963436e-05,
      "loss": 0.0486,
      "step": 1810
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.016421195192145e-05,
      "loss": 0.0591,
      "step": 1820
    },
    {
      "epoch": 0.84,
      "learning_rate": 4.007956661587947e-05,
      "loss": 0.0859,
      "step": 1830
    },
    {
      "epoch": 0.84,
      "learning_rate": 3.999492127983748e-05,
      "loss": 0.0291,
      "step": 1840
    },
    {
      "epoch": 0.85,
      "learning_rate": 3.99102759437955e-05,
      "loss": 0.0774,
      "step": 1850
    },
    {
      "epoch": 0.85,
      "learning_rate": 3.9825630607753515e-05,
      "loss": 0.0702,
      "step": 1860
    },
    {
      "epoch": 0.85,
      "learning_rate": 3.974098527171153e-05,
      "loss": 0.066,
      "step": 1870
    },
    {
      "epoch": 0.86,
      "learning_rate": 3.9656339935669546e-05,
      "loss": 0.0502,
      "step": 1880
    },
    {
      "epoch": 0.86,
      "learning_rate": 3.957169459962756e-05,
      "loss": 0.0513,
      "step": 1890
    },
    {
      "epoch": 0.87,
      "learning_rate": 3.948704926358558e-05,
      "loss": 0.039,
      "step": 1900
    },
    {
      "epoch": 0.87,
      "eval_accuracy": 0.9716591384818772,
      "eval_loss": 0.08289344608783722,
      "eval_runtime": 89.3988,
      "eval_samples_per_second": 406.135,
      "eval_steps_per_second": 6.354,
      "step": 1900
    },
    {
      "epoch": 0.87,
      "learning_rate": 3.9402403927543594e-05,
      "loss": 0.0462,
      "step": 1910
    },
    {
      "epoch": 0.88,
      "learning_rate": 3.931775859150161e-05,
      "loss": 0.1035,
      "step": 1920
    },
    {
      "epoch": 0.88,
      "learning_rate": 3.9233113255459625e-05,
      "loss": 0.0869,
      "step": 1930
    },
    {
      "epoch": 0.89,
      "learning_rate": 3.914846791941764e-05,
      "loss": 0.0591,
      "step": 1940
    },
    {
      "epoch": 0.89,
      "learning_rate": 3.906382258337566e-05,
      "loss": 0.0523,
      "step": 1950
    },
    {
      "epoch": 0.9,
      "learning_rate": 3.897917724733367e-05,
      "loss": 0.0476,
      "step": 1960
    },
    {
      "epoch": 0.9,
      "learning_rate": 3.889453191129169e-05,
      "loss": 0.0441,
      "step": 1970
    },
    {
      "epoch": 0.9,
      "learning_rate": 3.8809886575249705e-05,
      "loss": 0.0769,
      "step": 1980
    },
    {
      "epoch": 0.91,
      "learning_rate": 3.872524123920772e-05,
      "loss": 0.0332,
      "step": 1990
    },
    {
      "epoch": 0.91,
      "learning_rate": 3.8640595903165736e-05,
      "loss": 0.0607,
      "step": 2000
    },
    {
      "epoch": 0.91,
      "eval_accuracy": 0.9746061474055304,
      "eval_loss": 0.07671281695365906,
      "eval_runtime": 89.4267,
      "eval_samples_per_second": 406.008,
      "eval_steps_per_second": 6.352,
      "step": 2000
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.855595056712375e-05,
      "loss": 0.0564,
      "step": 2010
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.847130523108177e-05,
      "loss": 0.0863,
      "step": 2020
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.8386659895039784e-05,
      "loss": 0.0538,
      "step": 2030
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.83020145589978e-05,
      "loss": 0.0704,
      "step": 2040
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.8217369222955815e-05,
      "loss": 0.0598,
      "step": 2050
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.813272388691384e-05,
      "loss": 0.042,
      "step": 2060
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.804807855087185e-05,
      "loss": 0.0739,
      "step": 2070
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.796343321482986e-05,
      "loss": 0.0594,
      "step": 2080
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.787878787878788e-05,
      "loss": 0.0692,
      "step": 2090
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.7794142542745895e-05,
      "loss": 0.0478,
      "step": 2100
    },
    {
      "epoch": 0.96,
      "eval_accuracy": 0.9724854026660791,
      "eval_loss": 0.07888983935117722,
      "eval_runtime": 89.0735,
      "eval_samples_per_second": 407.618,
      "eval_steps_per_second": 6.377,
      "step": 2100
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.770949720670392e-05,
      "loss": 0.0639,
      "step": 2110
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.7624851870661926e-05,
      "loss": 0.0654,
      "step": 2120
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.754020653461994e-05,
      "loss": 0.082,
      "step": 2130
    },
    {
      "epoch": 0.98,
      "learning_rate": 3.745556119857796e-05,
      "loss": 0.048,
      "step": 2140
    },
    {
      "epoch": 0.98,
      "learning_rate": 3.7370915862535974e-05,
      "loss": 0.0702,
      "step": 2150
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.7286270526493996e-05,
      "loss": 0.0401,
      "step": 2160
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.7201625190452005e-05,
      "loss": 0.0537,
      "step": 2170
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.711697985441002e-05,
      "loss": 0.075,
      "step": 2180
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.703233451836804e-05,
      "loss": 0.0466,
      "step": 2190
    },
    {
      "epoch": 1.01,
      "learning_rate": 3.694768918232605e-05,
      "loss": 0.0408,
      "step": 2200
    },
    {
      "epoch": 1.01,
      "eval_accuracy": 0.9757078329844663,
      "eval_loss": 0.07499607652425766,
      "eval_runtime": 90.4675,
      "eval_samples_per_second": 401.338,
      "eval_steps_per_second": 6.278,
      "step": 2200
    },
    {
      "epoch": 1.01,
      "learning_rate": 3.6863043846284075e-05,
      "loss": 0.0758,
      "step": 2210
    },
    {
      "epoch": 1.01,
      "learning_rate": 3.6778398510242085e-05,
      "loss": 0.0578,
      "step": 2220
    },
    {
      "epoch": 1.02,
      "learning_rate": 3.66937531742001e-05,
      "loss": 0.0509,
      "step": 2230
    },
    {
      "epoch": 1.02,
      "learning_rate": 3.6609107838158116e-05,
      "loss": 0.0797,
      "step": 2240
    },
    {
      "epoch": 1.03,
      "learning_rate": 3.652446250211614e-05,
      "loss": 0.0379,
      "step": 2250
    },
    {
      "epoch": 1.03,
      "learning_rate": 3.6439817166074155e-05,
      "loss": 0.0933,
      "step": 2260
    },
    {
      "epoch": 1.04,
      "learning_rate": 3.6355171830032164e-05,
      "loss": 0.0564,
      "step": 2270
    },
    {
      "epoch": 1.04,
      "learning_rate": 3.627052649399018e-05,
      "loss": 0.0466,
      "step": 2280
    },
    {
      "epoch": 1.05,
      "learning_rate": 3.6185881157948195e-05,
      "loss": 0.049,
      "step": 2290
    },
    {
      "epoch": 1.05,
      "learning_rate": 3.610123582190622e-05,
      "loss": 0.0228,
      "step": 2300
    },
    {
      "epoch": 1.05,
      "eval_accuracy": 0.9772501927949763,
      "eval_loss": 0.07067732512950897,
      "eval_runtime": 88.9232,
      "eval_samples_per_second": 408.308,
      "eval_steps_per_second": 6.388,
      "step": 2300
    },
    {
      "epoch": 1.06,
      "learning_rate": 3.6016590485864234e-05,
      "loss": 0.0202,
      "step": 2310
    },
    {
      "epoch": 1.06,
      "learning_rate": 3.593194514982225e-05,
      "loss": 0.0476,
      "step": 2320
    },
    {
      "epoch": 1.06,
      "learning_rate": 3.584729981378026e-05,
      "loss": 0.0418,
      "step": 2330
    },
    {
      "epoch": 1.07,
      "learning_rate": 3.5762654477738275e-05,
      "loss": 0.0463,
      "step": 2340
    },
    {
      "epoch": 1.07,
      "learning_rate": 3.56780091416963e-05,
      "loss": 0.0272,
      "step": 2350
    },
    {
      "epoch": 1.08,
      "learning_rate": 3.559336380565431e-05,
      "loss": 0.0746,
      "step": 2360
    },
    {
      "epoch": 1.08,
      "learning_rate": 3.550871846961233e-05,
      "loss": 0.0452,
      "step": 2370
    },
    {
      "epoch": 1.09,
      "learning_rate": 3.542407313357034e-05,
      "loss": 0.0539,
      "step": 2380
    },
    {
      "epoch": 1.09,
      "learning_rate": 3.533942779752836e-05,
      "loss": 0.0394,
      "step": 2390
    },
    {
      "epoch": 1.1,
      "learning_rate": 3.5254782461486376e-05,
      "loss": 0.0431,
      "step": 2400
    },
    {
      "epoch": 1.1,
      "eval_accuracy": 0.9786548419081195,
      "eval_loss": 0.0689849704504013,
      "eval_runtime": 89.6221,
      "eval_samples_per_second": 405.123,
      "eval_steps_per_second": 6.338,
      "step": 2400
    },
    {
      "epoch": 1.1,
      "learning_rate": 3.517013712544439e-05,
      "loss": 0.0372,
      "step": 2410
    },
    {
      "epoch": 1.11,
      "learning_rate": 3.508549178940241e-05,
      "loss": 0.0646,
      "step": 2420
    },
    {
      "epoch": 1.11,
      "learning_rate": 3.500084645336042e-05,
      "loss": 0.0558,
      "step": 2430
    },
    {
      "epoch": 1.12,
      "learning_rate": 3.491620111731844e-05,
      "loss": 0.0579,
      "step": 2440
    },
    {
      "epoch": 1.12,
      "learning_rate": 3.4831555781276455e-05,
      "loss": 0.0503,
      "step": 2450
    },
    {
      "epoch": 1.12,
      "learning_rate": 3.474691044523447e-05,
      "loss": 0.0792,
      "step": 2460
    },
    {
      "epoch": 1.13,
      "learning_rate": 3.466226510919249e-05,
      "loss": 0.0431,
      "step": 2470
    },
    {
      "epoch": 1.13,
      "learning_rate": 3.4577619773150496e-05,
      "loss": 0.0294,
      "step": 2480
    },
    {
      "epoch": 1.14,
      "learning_rate": 3.449297443710852e-05,
      "loss": 0.0365,
      "step": 2490
    },
    {
      "epoch": 1.14,
      "learning_rate": 3.4408329101066535e-05,
      "loss": 0.0675,
      "step": 2500
    },
    {
      "epoch": 1.14,
      "eval_accuracy": 0.9773328192133965,
      "eval_loss": 0.07122696191072464,
      "eval_runtime": 89.3187,
      "eval_samples_per_second": 406.499,
      "eval_steps_per_second": 6.359,
      "step": 2500
    },
    {
      "epoch": 1.15,
      "learning_rate": 3.432368376502455e-05,
      "loss": 0.0572,
      "step": 2510
    },
    {
      "epoch": 1.15,
      "learning_rate": 3.4239038428982566e-05,
      "loss": 0.0358,
      "step": 2520
    },
    {
      "epoch": 1.16,
      "learning_rate": 3.4154393092940575e-05,
      "loss": 0.0348,
      "step": 2530
    },
    {
      "epoch": 1.16,
      "learning_rate": 3.40697477568986e-05,
      "loss": 0.0345,
      "step": 2540
    },
    {
      "epoch": 1.17,
      "learning_rate": 3.3985102420856614e-05,
      "loss": 0.0319,
      "step": 2550
    },
    {
      "epoch": 1.17,
      "learning_rate": 3.390045708481463e-05,
      "loss": 0.0345,
      "step": 2560
    },
    {
      "epoch": 1.17,
      "learning_rate": 3.3815811748772645e-05,
      "loss": 0.0737,
      "step": 2570
    },
    {
      "epoch": 1.18,
      "learning_rate": 3.373116641273066e-05,
      "loss": 0.0506,
      "step": 2580
    },
    {
      "epoch": 1.18,
      "learning_rate": 3.364652107668868e-05,
      "loss": 0.054,
      "step": 2590
    },
    {
      "epoch": 1.19,
      "learning_rate": 3.356187574064669e-05,
      "loss": 0.0624,
      "step": 2600
    },
    {
      "epoch": 1.19,
      "eval_accuracy": 0.9639748815688003,
      "eval_loss": 0.11090453714132309,
      "eval_runtime": 89.5399,
      "eval_samples_per_second": 405.495,
      "eval_steps_per_second": 6.344,
      "step": 2600
    },
    {
      "epoch": 1.19,
      "learning_rate": 3.347723040460471e-05,
      "loss": 0.0843,
      "step": 2610
    },
    {
      "epoch": 1.2,
      "learning_rate": 3.3392585068562725e-05,
      "loss": 0.0387,
      "step": 2620
    },
    {
      "epoch": 1.2,
      "learning_rate": 3.330793973252074e-05,
      "loss": 0.032,
      "step": 2630
    },
    {
      "epoch": 1.21,
      "learning_rate": 3.3223294396478756e-05,
      "loss": 0.0342,
      "step": 2640
    },
    {
      "epoch": 1.21,
      "learning_rate": 3.313864906043677e-05,
      "loss": 0.0447,
      "step": 2650
    },
    {
      "epoch": 1.22,
      "learning_rate": 3.305400372439479e-05,
      "loss": 0.03,
      "step": 2660
    },
    {
      "epoch": 1.22,
      "learning_rate": 3.2969358388352804e-05,
      "loss": 0.0246,
      "step": 2670
    },
    {
      "epoch": 1.22,
      "learning_rate": 3.288471305231082e-05,
      "loss": 0.0414,
      "step": 2680
    },
    {
      "epoch": 1.23,
      "learning_rate": 3.2800067716268835e-05,
      "loss": 0.0412,
      "step": 2690
    },
    {
      "epoch": 1.23,
      "learning_rate": 3.271542238022685e-05,
      "loss": 0.0843,
      "step": 2700
    },
    {
      "epoch": 1.23,
      "eval_accuracy": 0.9691803459292718,
      "eval_loss": 0.10772253572940826,
      "eval_runtime": 88.6336,
      "eval_samples_per_second": 409.642,
      "eval_steps_per_second": 6.408,
      "step": 2700
    },
    {
      "epoch": 1.24,
      "learning_rate": 3.263077704418487e-05,
      "loss": 0.0391,
      "step": 2710
    },
    {
      "epoch": 1.24,
      "learning_rate": 3.254613170814288e-05,
      "loss": 0.0644,
      "step": 2720
    },
    {
      "epoch": 1.25,
      "learning_rate": 3.24614863721009e-05,
      "loss": 0.0722,
      "step": 2730
    },
    {
      "epoch": 1.25,
      "learning_rate": 3.2376841036058914e-05,
      "loss": 0.0229,
      "step": 2740
    },
    {
      "epoch": 1.26,
      "learning_rate": 3.229219570001693e-05,
      "loss": 0.0433,
      "step": 2750
    },
    {
      "epoch": 1.26,
      "learning_rate": 3.2207550363974946e-05,
      "loss": 0.0546,
      "step": 2760
    },
    {
      "epoch": 1.27,
      "learning_rate": 3.212290502793296e-05,
      "loss": 0.0389,
      "step": 2770
    },
    {
      "epoch": 1.27,
      "learning_rate": 3.203825969189098e-05,
      "loss": 0.0388,
      "step": 2780
    },
    {
      "epoch": 1.28,
      "learning_rate": 3.1953614355848994e-05,
      "loss": 0.0819,
      "step": 2790
    },
    {
      "epoch": 1.28,
      "learning_rate": 3.186896901980701e-05,
      "loss": 0.0328,
      "step": 2800
    },
    {
      "epoch": 1.28,
      "eval_accuracy": 0.9794535639528479,
      "eval_loss": 0.06627843528985977,
      "eval_runtime": 88.7026,
      "eval_samples_per_second": 409.323,
      "eval_steps_per_second": 6.403,
      "step": 2800
    },
    {
      "epoch": 1.28,
      "learning_rate": 3.1784323683765025e-05,
      "loss": 0.0555,
      "step": 2810
    },
    {
      "epoch": 1.29,
      "learning_rate": 3.170814288132724e-05,
      "loss": 0.0487,
      "step": 2820
    },
    {
      "epoch": 1.29,
      "learning_rate": 3.162349754528526e-05,
      "loss": 0.0367,
      "step": 2830
    },
    {
      "epoch": 1.3,
      "learning_rate": 3.153885220924327e-05,
      "loss": 0.0507,
      "step": 2840
    },
    {
      "epoch": 1.3,
      "learning_rate": 3.145420687320129e-05,
      "loss": 0.0413,
      "step": 2850
    },
    {
      "epoch": 1.31,
      "learning_rate": 3.1369561537159305e-05,
      "loss": 0.0759,
      "step": 2860
    },
    {
      "epoch": 1.31,
      "learning_rate": 3.128491620111732e-05,
      "loss": 0.0186,
      "step": 2870
    },
    {
      "epoch": 1.32,
      "learning_rate": 3.1200270865075337e-05,
      "loss": 0.0291,
      "step": 2880
    },
    {
      "epoch": 1.32,
      "learning_rate": 3.111562552903335e-05,
      "loss": 0.0693,
      "step": 2890
    },
    {
      "epoch": 1.33,
      "learning_rate": 3.103098019299137e-05,
      "loss": 0.0724,
      "step": 2900
    },
    {
      "epoch": 1.33,
      "eval_accuracy": 0.9766167235870883,
      "eval_loss": 0.08111142367124557,
      "eval_runtime": 88.2495,
      "eval_samples_per_second": 411.424,
      "eval_steps_per_second": 6.436,
      "step": 2900
    },
    {
      "epoch": 1.33,
      "learning_rate": 3.0946334856949384e-05,
      "loss": 0.0622,
      "step": 2910
    },
    {
      "epoch": 1.33,
      "learning_rate": 3.08616895209074e-05,
      "loss": 0.049,
      "step": 2920
    },
    {
      "epoch": 1.34,
      "learning_rate": 3.0777044184865416e-05,
      "loss": 0.0641,
      "step": 2930
    },
    {
      "epoch": 1.34,
      "learning_rate": 3.069239884882343e-05,
      "loss": 0.0507,
      "step": 2940
    },
    {
      "epoch": 1.35,
      "learning_rate": 3.060775351278145e-05,
      "loss": 0.0434,
      "step": 2950
    },
    {
      "epoch": 1.35,
      "learning_rate": 3.052310817673946e-05,
      "loss": 0.0544,
      "step": 2960
    },
    {
      "epoch": 1.36,
      "learning_rate": 3.043846284069748e-05,
      "loss": 0.0326,
      "step": 2970
    },
    {
      "epoch": 1.36,
      "learning_rate": 3.0353817504655495e-05,
      "loss": 0.0524,
      "step": 2980
    },
    {
      "epoch": 1.37,
      "learning_rate": 3.0269172168613514e-05,
      "loss": 0.0923,
      "step": 2990
    },
    {
      "epoch": 1.37,
      "learning_rate": 3.0184526832571523e-05,
      "loss": 0.0385,
      "step": 3000
    },
    {
      "epoch": 1.37,
      "eval_accuracy": 0.9731739561529139,
      "eval_loss": 0.08201687037944794,
      "eval_runtime": 89.2335,
      "eval_samples_per_second": 406.887,
      "eval_steps_per_second": 6.365,
      "step": 3000
    },
    {
      "epoch": 1.38,
      "learning_rate": 3.0099881496529542e-05,
      "loss": 0.0525,
      "step": 3010
    },
    {
      "epoch": 1.38,
      "learning_rate": 3.0015236160487558e-05,
      "loss": 0.0377,
      "step": 3020
    },
    {
      "epoch": 1.38,
      "learning_rate": 2.9930590824445577e-05,
      "loss": 0.0384,
      "step": 3030
    },
    {
      "epoch": 1.39,
      "learning_rate": 2.9845945488403593e-05,
      "loss": 0.0658,
      "step": 3040
    },
    {
      "epoch": 1.39,
      "learning_rate": 2.9761300152361606e-05,
      "loss": 0.0304,
      "step": 3050
    },
    {
      "epoch": 1.4,
      "learning_rate": 2.967665481631962e-05,
      "loss": 0.0374,
      "step": 3060
    },
    {
      "epoch": 1.4,
      "learning_rate": 2.9592009480277637e-05,
      "loss": 0.0432,
      "step": 3070
    },
    {
      "epoch": 1.41,
      "learning_rate": 2.9507364144235657e-05,
      "loss": 0.0399,
      "step": 3080
    },
    {
      "epoch": 1.41,
      "learning_rate": 2.9422718808193672e-05,
      "loss": 0.0482,
      "step": 3090
    },
    {
      "epoch": 1.42,
      "learning_rate": 2.9338073472151685e-05,
      "loss": 0.0315,
      "step": 3100
    },
    {
      "epoch": 1.42,
      "eval_accuracy": 0.9787925526054864,
      "eval_loss": 0.07114042341709137,
      "eval_runtime": 89.2327,
      "eval_samples_per_second": 406.891,
      "eval_steps_per_second": 6.365,
      "step": 3100
    },
    {
      "epoch": 1.42,
      "learning_rate": 2.92534281361097e-05,
      "loss": 0.0355,
      "step": 3110
    },
    {
      "epoch": 1.43,
      "learning_rate": 2.9168782800067716e-05,
      "loss": 0.0517,
      "step": 3120
    },
    {
      "epoch": 1.43,
      "learning_rate": 2.9084137464025736e-05,
      "loss": 0.0522,
      "step": 3130
    },
    {
      "epoch": 1.44,
      "learning_rate": 2.899949212798375e-05,
      "loss": 0.0435,
      "step": 3140
    },
    {
      "epoch": 1.44,
      "learning_rate": 2.8914846791941764e-05,
      "loss": 0.0431,
      "step": 3150
    },
    {
      "epoch": 1.44,
      "learning_rate": 2.883020145589978e-05,
      "loss": 0.0581,
      "step": 3160
    },
    {
      "epoch": 1.45,
      "learning_rate": 2.8745556119857796e-05,
      "loss": 0.0466,
      "step": 3170
    },
    {
      "epoch": 1.45,
      "learning_rate": 2.8660910783815815e-05,
      "loss": 0.0311,
      "step": 3180
    },
    {
      "epoch": 1.46,
      "learning_rate": 2.857626544777383e-05,
      "loss": 0.0273,
      "step": 3190
    },
    {
      "epoch": 1.46,
      "learning_rate": 2.8491620111731843e-05,
      "loss": 0.0367,
      "step": 3200
    },
    {
      "epoch": 1.46,
      "eval_accuracy": 0.976534097168668,
      "eval_loss": 0.0805906429886818,
      "eval_runtime": 89.6613,
      "eval_samples_per_second": 404.946,
      "eval_steps_per_second": 6.335,
      "step": 3200
    },
    {
      "epoch": 1.47,
      "learning_rate": 2.840697477568986e-05,
      "loss": 0.0289,
      "step": 3210
    },
    {
      "epoch": 1.47,
      "learning_rate": 2.8322329439647878e-05,
      "loss": 0.0367,
      "step": 3220
    },
    {
      "epoch": 1.48,
      "learning_rate": 2.8237684103605894e-05,
      "loss": 0.0825,
      "step": 3230
    },
    {
      "epoch": 1.48,
      "learning_rate": 2.815303876756391e-05,
      "loss": 0.0696,
      "step": 3240
    },
    {
      "epoch": 1.49,
      "learning_rate": 2.8068393431521922e-05,
      "loss": 0.0419,
      "step": 3250
    },
    {
      "epoch": 1.49,
      "learning_rate": 2.7983748095479938e-05,
      "loss": 0.0221,
      "step": 3260
    },
    {
      "epoch": 1.49,
      "learning_rate": 2.7899102759437957e-05,
      "loss": 0.0629,
      "step": 3270
    },
    {
      "epoch": 1.5,
      "learning_rate": 2.7814457423395973e-05,
      "loss": 0.03,
      "step": 3280
    },
    {
      "epoch": 1.5,
      "learning_rate": 2.772981208735399e-05,
      "loss": 0.0349,
      "step": 3290
    },
    {
      "epoch": 1.51,
      "learning_rate": 2.7645166751312e-05,
      "loss": 0.0382,
      "step": 3300
    },
    {
      "epoch": 1.51,
      "eval_accuracy": 0.9611931254819874,
      "eval_loss": 0.14438240230083466,
      "eval_runtime": 88.8423,
      "eval_samples_per_second": 408.679,
      "eval_steps_per_second": 6.393,
      "step": 3300
    },
    {
      "epoch": 1.51,
      "learning_rate": 2.7560521415270017e-05,
      "loss": 0.0647,
      "step": 3310
    },
    {
      "epoch": 1.52,
      "learning_rate": 2.7475876079228036e-05,
      "loss": 0.0695,
      "step": 3320
    },
    {
      "epoch": 1.52,
      "learning_rate": 2.7391230743186052e-05,
      "loss": 0.0502,
      "step": 3330
    },
    {
      "epoch": 1.53,
      "learning_rate": 2.7306585407144068e-05,
      "loss": 0.0571,
      "step": 3340
    },
    {
      "epoch": 1.53,
      "learning_rate": 2.722194007110208e-05,
      "loss": 0.0458,
      "step": 3350
    },
    {
      "epoch": 1.54,
      "learning_rate": 2.71372947350601e-05,
      "loss": 0.0453,
      "step": 3360
    },
    {
      "epoch": 1.54,
      "learning_rate": 2.7052649399018116e-05,
      "loss": 0.0389,
      "step": 3370
    },
    {
      "epoch": 1.54,
      "learning_rate": 2.696800406297613e-05,
      "loss": 0.0279,
      "step": 3380
    },
    {
      "epoch": 1.55,
      "learning_rate": 2.688335872693415e-05,
      "loss": 0.0174,
      "step": 3390
    },
    {
      "epoch": 1.55,
      "learning_rate": 2.679871339089216e-05,
      "loss": 0.024,
      "step": 3400
    },
    {
      "epoch": 1.55,
      "eval_accuracy": 0.973807425360802,
      "eval_loss": 0.10379726439714432,
      "eval_runtime": 87.4819,
      "eval_samples_per_second": 415.035,
      "eval_steps_per_second": 6.493,
      "step": 3400
    },
    {
      "epoch": 1.56,
      "learning_rate": 2.671406805485018e-05,
      "loss": 0.0532,
      "step": 3410
    },
    {
      "epoch": 1.56,
      "learning_rate": 2.6629422718808195e-05,
      "loss": 0.0568,
      "step": 3420
    },
    {
      "epoch": 1.57,
      "learning_rate": 2.654477738276621e-05,
      "loss": 0.0672,
      "step": 3430
    },
    {
      "epoch": 1.57,
      "learning_rate": 2.646013204672423e-05,
      "loss": 0.0911,
      "step": 3440
    },
    {
      "epoch": 1.58,
      "learning_rate": 2.637548671068224e-05,
      "loss": 0.0256,
      "step": 3450
    },
    {
      "epoch": 1.58,
      "learning_rate": 2.6290841374640258e-05,
      "loss": 0.0417,
      "step": 3460
    },
    {
      "epoch": 1.59,
      "learning_rate": 2.6206196038598274e-05,
      "loss": 0.0574,
      "step": 3470
    },
    {
      "epoch": 1.59,
      "learning_rate": 2.612155070255629e-05,
      "loss": 0.0508,
      "step": 3480
    },
    {
      "epoch": 1.6,
      "learning_rate": 2.603690536651431e-05,
      "loss": 0.0438,
      "step": 3490
    },
    {
      "epoch": 1.6,
      "learning_rate": 2.5952260030472318e-05,
      "loss": 0.0331,
      "step": 3500
    },
    {
      "epoch": 1.6,
      "eval_accuracy": 0.9660405420293048,
      "eval_loss": 0.11805976927280426,
      "eval_runtime": 87.8848,
      "eval_samples_per_second": 413.132,
      "eval_steps_per_second": 6.463,
      "step": 3500
    },
    {
      "epoch": 1.6,
      "learning_rate": 2.5867614694430337e-05,
      "loss": 0.0448,
      "step": 3510
    },
    {
      "epoch": 1.61,
      "learning_rate": 2.5782969358388353e-05,
      "loss": 0.0561,
      "step": 3520
    },
    {
      "epoch": 1.61,
      "learning_rate": 2.569832402234637e-05,
      "loss": 0.0513,
      "step": 3530
    },
    {
      "epoch": 1.62,
      "learning_rate": 2.5613678686304388e-05,
      "loss": 0.0391,
      "step": 3540
    },
    {
      "epoch": 1.62,
      "learning_rate": 2.55290333502624e-05,
      "loss": 0.0304,
      "step": 3550
    },
    {
      "epoch": 1.63,
      "learning_rate": 2.5444388014220416e-05,
      "loss": 0.0296,
      "step": 3560
    },
    {
      "epoch": 1.63,
      "learning_rate": 2.5359742678178432e-05,
      "loss": 0.0263,
      "step": 3570
    },
    {
      "epoch": 1.64,
      "learning_rate": 2.527509734213645e-05,
      "loss": 0.0472,
      "step": 3580
    },
    {
      "epoch": 1.64,
      "learning_rate": 2.5190452006094467e-05,
      "loss": 0.0365,
      "step": 3590
    },
    {
      "epoch": 1.65,
      "learning_rate": 2.510580667005248e-05,
      "loss": 0.0419,
      "step": 3600
    },
    {
      "epoch": 1.65,
      "eval_accuracy": 0.9789853475818001,
      "eval_loss": 0.06870424002408981,
      "eval_runtime": 87.1711,
      "eval_samples_per_second": 416.514,
      "eval_steps_per_second": 6.516,
      "step": 3600
    },
    {
      "epoch": 1.65,
      "learning_rate": 2.5021161334010496e-05,
      "loss": 0.0404,
      "step": 3610
    },
    {
      "epoch": 1.65,
      "learning_rate": 2.493651599796851e-05,
      "loss": 0.0305,
      "step": 3620
    },
    {
      "epoch": 1.66,
      "learning_rate": 2.485187066192653e-05,
      "loss": 0.0602,
      "step": 3630
    },
    {
      "epoch": 1.66,
      "learning_rate": 2.4767225325884543e-05,
      "loss": 0.0439,
      "step": 3640
    },
    {
      "epoch": 1.67,
      "learning_rate": 2.4682579989842562e-05,
      "loss": 0.0512,
      "step": 3650
    },
    {
      "epoch": 1.67,
      "learning_rate": 2.4597934653800578e-05,
      "loss": 0.0302,
      "step": 3660
    },
    {
      "epoch": 1.68,
      "learning_rate": 2.451328931775859e-05,
      "loss": 0.0433,
      "step": 3670
    },
    {
      "epoch": 1.68,
      "learning_rate": 2.442864398171661e-05,
      "loss": 0.0245,
      "step": 3680
    },
    {
      "epoch": 1.69,
      "learning_rate": 2.4343998645674622e-05,
      "loss": 0.047,
      "step": 3690
    },
    {
      "epoch": 1.69,
      "learning_rate": 2.425935330963264e-05,
      "loss": 0.0352,
      "step": 3700
    },
    {
      "epoch": 1.69,
      "eval_accuracy": 0.9789302633028534,
      "eval_loss": 0.06866224855184555,
      "eval_runtime": 87.2662,
      "eval_samples_per_second": 416.06,
      "eval_steps_per_second": 6.509,
      "step": 3700
    },
    {
      "epoch": 1.7,
      "learning_rate": 2.4174707973590657e-05,
      "loss": 0.0282,
      "step": 3710
    },
    {
      "epoch": 1.7,
      "learning_rate": 2.4090062637548673e-05,
      "loss": 0.0414,
      "step": 3720
    },
    {
      "epoch": 1.7,
      "learning_rate": 2.400541730150669e-05,
      "loss": 0.031,
      "step": 3730
    },
    {
      "epoch": 1.71,
      "learning_rate": 2.39207719654647e-05,
      "loss": 0.0664,
      "step": 3740
    },
    {
      "epoch": 1.71,
      "learning_rate": 2.383612662942272e-05,
      "loss": 0.05,
      "step": 3750
    },
    {
      "epoch": 1.72,
      "learning_rate": 2.3751481293380736e-05,
      "loss": 0.0627,
      "step": 3760
    },
    {
      "epoch": 1.72,
      "learning_rate": 2.3666835957338752e-05,
      "loss": 0.0511,
      "step": 3770
    },
    {
      "epoch": 1.73,
      "learning_rate": 2.3582190621296768e-05,
      "loss": 0.0363,
      "step": 3780
    },
    {
      "epoch": 1.73,
      "learning_rate": 2.3497545285254784e-05,
      "loss": 0.0312,
      "step": 3790
    },
    {
      "epoch": 1.74,
      "learning_rate": 2.34128999492128e-05,
      "loss": 0.0588,
      "step": 3800
    },
    {
      "epoch": 1.74,
      "eval_accuracy": 0.9804175388344166,
      "eval_loss": 0.06199588254094124,
      "eval_runtime": 87.658,
      "eval_samples_per_second": 414.201,
      "eval_steps_per_second": 6.48,
      "step": 3800
    },
    {
      "epoch": 1.74,
      "learning_rate": 2.3328254613170816e-05,
      "loss": 0.0302,
      "step": 3810
    },
    {
      "epoch": 1.75,
      "learning_rate": 2.324360927712883e-05,
      "loss": 0.0337,
      "step": 3820
    },
    {
      "epoch": 1.75,
      "learning_rate": 2.3158963941086847e-05,
      "loss": 0.0577,
      "step": 3830
    },
    {
      "epoch": 1.76,
      "learning_rate": 2.3074318605044863e-05,
      "loss": 0.07,
      "step": 3840
    },
    {
      "epoch": 1.76,
      "learning_rate": 2.298967326900288e-05,
      "loss": 0.0359,
      "step": 3850
    },
    {
      "epoch": 1.76,
      "learning_rate": 2.2905027932960895e-05,
      "loss": 0.0328,
      "step": 3860
    },
    {
      "epoch": 1.77,
      "learning_rate": 2.282038259691891e-05,
      "loss": 0.0256,
      "step": 3870
    },
    {
      "epoch": 1.77,
      "learning_rate": 2.2735737260876926e-05,
      "loss": 0.041,
      "step": 3880
    },
    {
      "epoch": 1.78,
      "learning_rate": 2.2651091924834946e-05,
      "loss": 0.0164,
      "step": 3890
    },
    {
      "epoch": 1.78,
      "learning_rate": 2.2566446588792958e-05,
      "loss": 0.0313,
      "step": 3900
    },
    {
      "epoch": 1.78,
      "eval_accuracy": 0.9721824391318717,
      "eval_loss": 0.09753405302762985,
      "eval_runtime": 87.938,
      "eval_samples_per_second": 412.882,
      "eval_steps_per_second": 6.459,
      "step": 3900
    },
    {
      "epoch": 1.79,
      "learning_rate": 2.2481801252750974e-05,
      "loss": 0.041,
      "step": 3910
    },
    {
      "epoch": 1.79,
      "learning_rate": 2.239715591670899e-05,
      "loss": 0.0208,
      "step": 3920
    },
    {
      "epoch": 1.8,
      "learning_rate": 2.2312510580667005e-05,
      "loss": 0.0429,
      "step": 3930
    },
    {
      "epoch": 1.8,
      "learning_rate": 2.2227865244625025e-05,
      "loss": 0.0321,
      "step": 3940
    },
    {
      "epoch": 1.81,
      "learning_rate": 2.2143219908583037e-05,
      "loss": 0.0543,
      "step": 3950
    },
    {
      "epoch": 1.81,
      "learning_rate": 2.2058574572541056e-05,
      "loss": 0.0462,
      "step": 3960
    },
    {
      "epoch": 1.81,
      "learning_rate": 2.197392923649907e-05,
      "loss": 0.049,
      "step": 3970
    },
    {
      "epoch": 1.82,
      "learning_rate": 2.1889283900457085e-05,
      "loss": 0.0322,
      "step": 3980
    },
    {
      "epoch": 1.82,
      "learning_rate": 2.1804638564415104e-05,
      "loss": 0.0546,
      "step": 3990
    },
    {
      "epoch": 1.83,
      "learning_rate": 2.1719993228373116e-05,
      "loss": 0.0421,
      "step": 4000
    },
    {
      "epoch": 1.83,
      "eval_accuracy": 0.9803349124159965,
      "eval_loss": 0.058828968554735184,
      "eval_runtime": 87.5956,
      "eval_samples_per_second": 414.496,
      "eval_steps_per_second": 6.484,
      "step": 4000
    },
    {
      "epoch": 1.83,
      "learning_rate": 2.1635347892331136e-05,
      "loss": 0.0344,
      "step": 4010
    },
    {
      "epoch": 1.84,
      "learning_rate": 2.1550702556289148e-05,
      "loss": 0.0322,
      "step": 4020
    },
    {
      "epoch": 1.84,
      "learning_rate": 2.1466057220247164e-05,
      "loss": 0.0423,
      "step": 4030
    },
    {
      "epoch": 1.85,
      "learning_rate": 2.1381411884205183e-05,
      "loss": 0.0509,
      "step": 4040
    },
    {
      "epoch": 1.85,
      "learning_rate": 2.1296766548163195e-05,
      "loss": 0.0374,
      "step": 4050
    },
    {
      "epoch": 1.86,
      "learning_rate": 2.1212121212121215e-05,
      "loss": 0.034,
      "step": 4060
    },
    {
      "epoch": 1.86,
      "learning_rate": 2.1127475876079227e-05,
      "loss": 0.031,
      "step": 4070
    },
    {
      "epoch": 1.86,
      "learning_rate": 2.1042830540037246e-05,
      "loss": 0.0507,
      "step": 4080
    },
    {
      "epoch": 1.87,
      "learning_rate": 2.0958185203995262e-05,
      "loss": 0.0269,
      "step": 4090
    },
    {
      "epoch": 1.87,
      "learning_rate": 2.0873539867953275e-05,
      "loss": 0.0182,
      "step": 4100
    },
    {
      "epoch": 1.87,
      "eval_accuracy": 0.9819323565054533,
      "eval_loss": 0.06010541319847107,
      "eval_runtime": 88.4632,
      "eval_samples_per_second": 410.431,
      "eval_steps_per_second": 6.421,
      "step": 4100
    },
    {
      "epoch": 1.88,
      "learning_rate": 2.0788894531911294e-05,
      "loss": 0.0571,
      "step": 4110
    },
    {
      "epoch": 1.88,
      "learning_rate": 2.0704249195869306e-05,
      "loss": 0.0394,
      "step": 4120
    },
    {
      "epoch": 1.89,
      "learning_rate": 2.0619603859827325e-05,
      "loss": 0.0532,
      "step": 4130
    },
    {
      "epoch": 1.89,
      "learning_rate": 2.053495852378534e-05,
      "loss": 0.0275,
      "step": 4140
    },
    {
      "epoch": 1.9,
      "learning_rate": 2.0450313187743357e-05,
      "loss": 0.0304,
      "step": 4150
    },
    {
      "epoch": 1.9,
      "learning_rate": 2.0365667851701373e-05,
      "loss": 0.0262,
      "step": 4160
    },
    {
      "epoch": 1.91,
      "learning_rate": 2.0281022515659385e-05,
      "loss": 0.0438,
      "step": 4170
    },
    {
      "epoch": 1.91,
      "learning_rate": 2.0196377179617405e-05,
      "loss": 0.0531,
      "step": 4180
    },
    {
      "epoch": 1.91,
      "learning_rate": 2.011173184357542e-05,
      "loss": 0.0314,
      "step": 4190
    },
    {
      "epoch": 1.92,
      "learning_rate": 2.0027086507533436e-05,
      "loss": 0.0323,
      "step": 4200
    },
    {
      "epoch": 1.92,
      "eval_accuracy": 0.9818772722265066,
      "eval_loss": 0.059311188757419586,
      "eval_runtime": 88.9569,
      "eval_samples_per_second": 408.153,
      "eval_steps_per_second": 6.385,
      "step": 4200
    },
    {
      "epoch": 1.92,
      "learning_rate": 1.9942441171491452e-05,
      "loss": 0.0286,
      "step": 4210
    },
    {
      "epoch": 1.93,
      "learning_rate": 1.9857795835449468e-05,
      "loss": 0.0394,
      "step": 4220
    },
    {
      "epoch": 1.93,
      "learning_rate": 1.9773150499407484e-05,
      "loss": 0.0318,
      "step": 4230
    },
    {
      "epoch": 1.94,
      "learning_rate": 1.96885051633655e-05,
      "loss": 0.0107,
      "step": 4240
    },
    {
      "epoch": 1.94,
      "learning_rate": 1.9603859827323515e-05,
      "loss": 0.0366,
      "step": 4250
    },
    {
      "epoch": 1.95,
      "learning_rate": 1.951921449128153e-05,
      "loss": 0.0443,
      "step": 4260
    },
    {
      "epoch": 1.95,
      "learning_rate": 1.9434569155239547e-05,
      "loss": 0.0527,
      "step": 4270
    },
    {
      "epoch": 1.96,
      "learning_rate": 1.9349923819197563e-05,
      "loss": 0.0397,
      "step": 4280
    },
    {
      "epoch": 1.96,
      "learning_rate": 1.926527848315558e-05,
      "loss": 0.0509,
      "step": 4290
    },
    {
      "epoch": 1.97,
      "learning_rate": 1.9180633147113595e-05,
      "loss": 0.0565,
      "step": 4300
    },
    {
      "epoch": 1.97,
      "eval_accuracy": 0.9819874407844001,
      "eval_loss": 0.053711626678705215,
      "eval_runtime": 89.1291,
      "eval_samples_per_second": 407.364,
      "eval_steps_per_second": 6.373,
      "step": 4300
    },
    {
      "epoch": 1.97,
      "learning_rate": 1.909598781107161e-05,
      "loss": 0.0335,
      "step": 4310
    },
    {
      "epoch": 1.97,
      "learning_rate": 1.9011342475029626e-05,
      "loss": 0.0575,
      "step": 4320
    },
    {
      "epoch": 1.98,
      "learning_rate": 1.8926697138987642e-05,
      "loss": 0.0573,
      "step": 4330
    },
    {
      "epoch": 1.98,
      "learning_rate": 1.8842051802945658e-05,
      "loss": 0.0393,
      "step": 4340
    },
    {
      "epoch": 1.99,
      "learning_rate": 1.8757406466903674e-05,
      "loss": 0.0187,
      "step": 4350
    },
    {
      "epoch": 1.99,
      "learning_rate": 1.867276113086169e-05,
      "loss": 0.0421,
      "step": 4360
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.858811579481971e-05,
      "loss": 0.0338,
      "step": 4370
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.850347045877772e-05,
      "loss": 0.0162,
      "step": 4380
    },
    {
      "epoch": 2.01,
      "learning_rate": 1.841882512273574e-05,
      "loss": 0.0351,
      "step": 4390
    },
    {
      "epoch": 2.01,
      "learning_rate": 1.8334179786693753e-05,
      "loss": 0.0266,
      "step": 4400
    },
    {
      "epoch": 2.01,
      "eval_accuracy": 0.9804175388344166,
      "eval_loss": 0.06930837035179138,
      "eval_runtime": 89.2514,
      "eval_samples_per_second": 406.806,
      "eval_steps_per_second": 6.364,
      "step": 4400
    },
    {
      "epoch": 2.02,
      "learning_rate": 1.824953445065177e-05,
      "loss": 0.0227,
      "step": 4410
    },
    {
      "epoch": 2.02,
      "learning_rate": 1.8164889114609788e-05,
      "loss": 0.0128,
      "step": 4420
    },
    {
      "epoch": 2.02,
      "learning_rate": 1.80802437785678e-05,
      "loss": 0.0406,
      "step": 4430
    },
    {
      "epoch": 2.03,
      "learning_rate": 1.799559844252582e-05,
      "loss": 0.0312,
      "step": 4440
    },
    {
      "epoch": 2.03,
      "learning_rate": 1.7910953106483832e-05,
      "loss": 0.0156,
      "step": 4450
    },
    {
      "epoch": 2.04,
      "learning_rate": 1.782630777044185e-05,
      "loss": 0.0355,
      "step": 4460
    },
    {
      "epoch": 2.04,
      "learning_rate": 1.7741662434399867e-05,
      "loss": 0.0301,
      "step": 4470
    },
    {
      "epoch": 2.05,
      "learning_rate": 1.765701709835788e-05,
      "loss": 0.0225,
      "step": 4480
    },
    {
      "epoch": 2.05,
      "learning_rate": 1.75723717623159e-05,
      "loss": 0.0303,
      "step": 4490
    },
    {
      "epoch": 2.06,
      "learning_rate": 1.748772642627391e-05,
      "loss": 0.0374,
      "step": 4500
    },
    {
      "epoch": 2.06,
      "eval_accuracy": 0.9818772722265066,
      "eval_loss": 0.06097808852791786,
      "eval_runtime": 88.9816,
      "eval_samples_per_second": 408.039,
      "eval_steps_per_second": 6.383,
      "step": 4500
    },
    {
      "epoch": 2.06,
      "learning_rate": 1.740308109023193e-05,
      "loss": 0.0496,
      "step": 4510
    },
    {
      "epoch": 2.07,
      "learning_rate": 1.7318435754189946e-05,
      "loss": 0.0304,
      "step": 4520
    },
    {
      "epoch": 2.07,
      "learning_rate": 1.723379041814796e-05,
      "loss": 0.0141,
      "step": 4530
    },
    {
      "epoch": 2.07,
      "learning_rate": 1.7149145082105978e-05,
      "loss": 0.0291,
      "step": 4540
    },
    {
      "epoch": 2.08,
      "learning_rate": 1.706449974606399e-05,
      "loss": 0.0287,
      "step": 4550
    },
    {
      "epoch": 2.08,
      "learning_rate": 1.697985441002201e-05,
      "loss": 0.0205,
      "step": 4560
    },
    {
      "epoch": 2.09,
      "learning_rate": 1.6895209073980025e-05,
      "loss": 0.0478,
      "step": 4570
    },
    {
      "epoch": 2.09,
      "learning_rate": 1.681056373793804e-05,
      "loss": 0.0578,
      "step": 4580
    },
    {
      "epoch": 2.1,
      "learning_rate": 1.6725918401896057e-05,
      "loss": 0.0329,
      "step": 4590
    },
    {
      "epoch": 2.1,
      "learning_rate": 1.664127306585407e-05,
      "loss": 0.0246,
      "step": 4600
    },
    {
      "epoch": 2.1,
      "eval_accuracy": 0.9821526936212405,
      "eval_loss": 0.05802278220653534,
      "eval_runtime": 88.2319,
      "eval_samples_per_second": 411.506,
      "eval_steps_per_second": 6.438,
      "step": 4600
    },
    {
      "epoch": 2.11,
      "learning_rate": 1.655662772981209e-05,
      "loss": 0.0244,
      "step": 4610
    },
    {
      "epoch": 2.11,
      "learning_rate": 1.6471982393770105e-05,
      "loss": 0.0467,
      "step": 4620
    },
    {
      "epoch": 2.12,
      "learning_rate": 1.638733705772812e-05,
      "loss": 0.036,
      "step": 4630
    },
    {
      "epoch": 2.12,
      "learning_rate": 1.6302691721686136e-05,
      "loss": 0.0194,
      "step": 4640
    },
    {
      "epoch": 2.13,
      "learning_rate": 1.6218046385644152e-05,
      "loss": 0.0432,
      "step": 4650
    },
    {
      "epoch": 2.13,
      "learning_rate": 1.6133401049602168e-05,
      "loss": 0.0248,
      "step": 4660
    },
    {
      "epoch": 2.13,
      "learning_rate": 1.6048755713560184e-05,
      "loss": 0.0421,
      "step": 4670
    },
    {
      "epoch": 2.14,
      "learning_rate": 1.59641103775182e-05,
      "loss": 0.057,
      "step": 4680
    },
    {
      "epoch": 2.14,
      "learning_rate": 1.5879465041476215e-05,
      "loss": 0.0371,
      "step": 4690
    },
    {
      "epoch": 2.15,
      "learning_rate": 1.579481970543423e-05,
      "loss": 0.0316,
      "step": 4700
    },
    {
      "epoch": 2.15,
      "eval_accuracy": 0.9804175388344166,
      "eval_loss": 0.06742748618125916,
      "eval_runtime": 88.0446,
      "eval_samples_per_second": 412.382,
      "eval_steps_per_second": 6.451,
      "step": 4700
    },
    {
      "epoch": 2.15,
      "learning_rate": 1.5710174369392247e-05,
      "loss": 0.0483,
      "step": 4710
    },
    {
      "epoch": 2.16,
      "learning_rate": 1.5625529033350263e-05,
      "loss": 0.0454,
      "step": 4720
    },
    {
      "epoch": 2.16,
      "learning_rate": 1.554088369730828e-05,
      "loss": 0.034,
      "step": 4730
    },
    {
      "epoch": 2.17,
      "learning_rate": 1.5456238361266295e-05,
      "loss": 0.0346,
      "step": 4740
    },
    {
      "epoch": 2.17,
      "learning_rate": 1.537159302522431e-05,
      "loss": 0.0204,
      "step": 4750
    },
    {
      "epoch": 2.18,
      "learning_rate": 1.5286947689182326e-05,
      "loss": 0.0393,
      "step": 4760
    },
    {
      "epoch": 2.18,
      "learning_rate": 1.5202302353140344e-05,
      "loss": 0.039,
      "step": 4770
    },
    {
      "epoch": 2.18,
      "learning_rate": 1.5117657017098358e-05,
      "loss": 0.0214,
      "step": 4780
    },
    {
      "epoch": 2.19,
      "learning_rate": 1.5033011681056375e-05,
      "loss": 0.0382,
      "step": 4790
    },
    {
      "epoch": 2.19,
      "learning_rate": 1.494836634501439e-05,
      "loss": 0.0415,
      "step": 4800
    },
    {
      "epoch": 2.19,
      "eval_accuracy": 0.9826484521317616,
      "eval_loss": 0.0569329634308815,
      "eval_runtime": 88.7967,
      "eval_samples_per_second": 408.889,
      "eval_steps_per_second": 6.397,
      "step": 4800
    },
    {
      "epoch": 2.2,
      "learning_rate": 1.4863721008972405e-05,
      "loss": 0.0201,
      "step": 4810
    },
    {
      "epoch": 2.2,
      "learning_rate": 1.4779075672930423e-05,
      "loss": 0.0385,
      "step": 4820
    },
    {
      "epoch": 2.21,
      "learning_rate": 1.4694430336888437e-05,
      "loss": 0.0231,
      "step": 4830
    },
    {
      "epoch": 2.21,
      "learning_rate": 1.4618249534450653e-05,
      "loss": 0.0351,
      "step": 4840
    },
    {
      "epoch": 2.22,
      "learning_rate": 1.4533604198408667e-05,
      "loss": 0.0383,
      "step": 4850
    },
    {
      "epoch": 2.22,
      "learning_rate": 1.4448958862366685e-05,
      "loss": 0.0456,
      "step": 4860
    },
    {
      "epoch": 2.23,
      "learning_rate": 1.4364313526324699e-05,
      "loss": 0.0632,
      "step": 4870
    },
    {
      "epoch": 2.23,
      "learning_rate": 1.4279668190282717e-05,
      "loss": 0.0387,
      "step": 4880
    },
    {
      "epoch": 2.23,
      "learning_rate": 1.4195022854240732e-05,
      "loss": 0.0239,
      "step": 4890
    },
    {
      "epoch": 2.24,
      "learning_rate": 1.4110377518198748e-05,
      "loss": 0.0361,
      "step": 4900
    },
    {
      "epoch": 2.24,
      "eval_accuracy": 0.9840255591054313,
      "eval_loss": 0.05500613898038864,
      "eval_runtime": 87.4849,
      "eval_samples_per_second": 415.02,
      "eval_steps_per_second": 6.493,
      "step": 4900
    },
    {
      "epoch": 2.24,
      "learning_rate": 1.4025732182156764e-05,
      "loss": 0.0284,
      "step": 4910
    },
    {
      "epoch": 2.25,
      "learning_rate": 1.3941086846114778e-05,
      "loss": 0.034,
      "step": 4920
    },
    {
      "epoch": 2.25,
      "learning_rate": 1.3856441510072796e-05,
      "loss": 0.0424,
      "step": 4930
    },
    {
      "epoch": 2.26,
      "learning_rate": 1.3771796174030813e-05,
      "loss": 0.0236,
      "step": 4940
    },
    {
      "epoch": 2.26,
      "learning_rate": 1.3687150837988827e-05,
      "loss": 0.0225,
      "step": 4950
    },
    {
      "epoch": 2.27,
      "learning_rate": 1.3602505501946843e-05,
      "loss": 0.0328,
      "step": 4960
    },
    {
      "epoch": 2.27,
      "learning_rate": 1.3517860165904857e-05,
      "loss": 0.0323,
      "step": 4970
    },
    {
      "epoch": 2.28,
      "learning_rate": 1.3433214829862875e-05,
      "loss": 0.0402,
      "step": 4980
    },
    {
      "epoch": 2.28,
      "learning_rate": 1.3348569493820892e-05,
      "loss": 0.0313,
      "step": 4990
    },
    {
      "epoch": 2.29,
      "learning_rate": 1.3263924157778907e-05,
      "loss": 0.0298,
      "step": 5000
    },
    {
      "epoch": 2.29,
      "eval_accuracy": 0.9829789578054423,
      "eval_loss": 0.057487934827804565,
      "eval_runtime": 87.3704,
      "eval_samples_per_second": 415.564,
      "eval_steps_per_second": 6.501,
      "step": 5000
    },
    {
      "epoch": 2.29,
      "learning_rate": 1.3179278821736924e-05,
      "loss": 0.0426,
      "step": 5010
    },
    {
      "epoch": 2.29,
      "learning_rate": 1.3094633485694938e-05,
      "loss": 0.0345,
      "step": 5020
    },
    {
      "epoch": 2.3,
      "learning_rate": 1.3009988149652954e-05,
      "loss": 0.045,
      "step": 5030
    },
    {
      "epoch": 2.3,
      "learning_rate": 1.2925342813610972e-05,
      "loss": 0.0224,
      "step": 5040
    },
    {
      "epoch": 2.31,
      "learning_rate": 1.2840697477568986e-05,
      "loss": 0.027,
      "step": 5050
    },
    {
      "epoch": 2.31,
      "learning_rate": 1.2756052141527003e-05,
      "loss": 0.0384,
      "step": 5060
    },
    {
      "epoch": 2.32,
      "learning_rate": 1.2671406805485017e-05,
      "loss": 0.029,
      "step": 5070
    },
    {
      "epoch": 2.32,
      "learning_rate": 1.2586761469443035e-05,
      "loss": 0.018,
      "step": 5080
    },
    {
      "epoch": 2.33,
      "learning_rate": 1.250211613340105e-05,
      "loss": 0.0207,
      "step": 5090
    },
    {
      "epoch": 2.33,
      "learning_rate": 1.2417470797359067e-05,
      "loss": 0.0275,
      "step": 5100
    },
    {
      "epoch": 2.33,
      "eval_accuracy": 0.9836124270133304,
      "eval_loss": 0.059953879565000534,
      "eval_runtime": 89.2241,
      "eval_samples_per_second": 406.931,
      "eval_steps_per_second": 6.366,
      "step": 5100
    },
    {
      "epoch": 2.34,
      "learning_rate": 1.2332825461317082e-05,
      "loss": 0.032,
      "step": 5110
    },
    {
      "epoch": 2.34,
      "learning_rate": 1.2248180125275098e-05,
      "loss": 0.0193,
      "step": 5120
    },
    {
      "epoch": 2.34,
      "learning_rate": 1.2163534789233114e-05,
      "loss": 0.0343,
      "step": 5130
    },
    {
      "epoch": 2.35,
      "learning_rate": 1.207888945319113e-05,
      "loss": 0.0158,
      "step": 5140
    },
    {
      "epoch": 2.35,
      "learning_rate": 1.1994244117149146e-05,
      "loss": 0.0194,
      "step": 5150
    },
    {
      "epoch": 2.36,
      "learning_rate": 1.191806331471136e-05,
      "loss": 0.0391,
      "step": 5160
    },
    {
      "epoch": 2.36,
      "learning_rate": 1.1833417978669376e-05,
      "loss": 0.0362,
      "step": 5170
    },
    {
      "epoch": 2.37,
      "learning_rate": 1.1748772642627392e-05,
      "loss": 0.045,
      "step": 5180
    },
    {
      "epoch": 2.37,
      "learning_rate": 1.1664127306585408e-05,
      "loss": 0.0455,
      "step": 5190
    },
    {
      "epoch": 2.38,
      "learning_rate": 1.1579481970543424e-05,
      "loss": 0.0194,
      "step": 5200
    },
    {
      "epoch": 2.38,
      "eval_accuracy": 0.9824556571554478,
      "eval_loss": 0.06781785935163498,
      "eval_runtime": 88.6192,
      "eval_samples_per_second": 409.708,
      "eval_steps_per_second": 6.409,
      "step": 5200
    },
    {
      "epoch": 2.38,
      "learning_rate": 1.149483663450144e-05,
      "loss": 0.0339,
      "step": 5210
    },
    {
      "epoch": 2.39,
      "learning_rate": 1.1410191298459455e-05,
      "loss": 0.0224,
      "step": 5220
    },
    {
      "epoch": 2.39,
      "learning_rate": 1.1325545962417473e-05,
      "loss": 0.0223,
      "step": 5230
    },
    {
      "epoch": 2.39,
      "learning_rate": 1.1240900626375487e-05,
      "loss": 0.0226,
      "step": 5240
    },
    {
      "epoch": 2.4,
      "learning_rate": 1.1156255290333503e-05,
      "loss": 0.0261,
      "step": 5250
    },
    {
      "epoch": 2.4,
      "learning_rate": 1.1071609954291519e-05,
      "loss": 0.0346,
      "step": 5260
    },
    {
      "epoch": 2.41,
      "learning_rate": 1.0986964618249534e-05,
      "loss": 0.0188,
      "step": 5270
    },
    {
      "epoch": 2.41,
      "learning_rate": 1.0902319282207552e-05,
      "loss": 0.0234,
      "step": 5280
    },
    {
      "epoch": 2.42,
      "learning_rate": 1.0817673946165568e-05,
      "loss": 0.0078,
      "step": 5290
    },
    {
      "epoch": 2.42,
      "learning_rate": 1.0733028610123582e-05,
      "loss": 0.0279,
      "step": 5300
    },
    {
      "epoch": 2.42,
      "eval_accuracy": 0.9838327641291176,
      "eval_loss": 0.06075261905789375,
      "eval_runtime": 88.6073,
      "eval_samples_per_second": 409.763,
      "eval_steps_per_second": 6.41,
      "step": 5300
    },
    {
      "epoch": 2.43,
      "learning_rate": 1.0648383274081598e-05,
      "loss": 0.0119,
      "step": 5310
    },
    {
      "epoch": 2.43,
      "learning_rate": 1.0563737938039614e-05,
      "loss": 0.0557,
      "step": 5320
    },
    {
      "epoch": 2.44,
      "learning_rate": 1.0479092601997631e-05,
      "loss": 0.0398,
      "step": 5330
    },
    {
      "epoch": 2.44,
      "learning_rate": 1.0394447265955647e-05,
      "loss": 0.0397,
      "step": 5340
    },
    {
      "epoch": 2.45,
      "learning_rate": 1.0309801929913663e-05,
      "loss": 0.0304,
      "step": 5350
    },
    {
      "epoch": 2.45,
      "learning_rate": 1.0225156593871679e-05,
      "loss": 0.0299,
      "step": 5360
    },
    {
      "epoch": 2.45,
      "learning_rate": 1.0140511257829693e-05,
      "loss": 0.0261,
      "step": 5370
    },
    {
      "epoch": 2.46,
      "learning_rate": 1.005586592178771e-05,
      "loss": 0.0583,
      "step": 5380
    },
    {
      "epoch": 2.46,
      "learning_rate": 9.971220585745726e-06,
      "loss": 0.039,
      "step": 5390
    },
    {
      "epoch": 2.47,
      "learning_rate": 9.886575249703742e-06,
      "loss": 0.0585,
      "step": 5400
    },
    {
      "epoch": 2.47,
      "eval_accuracy": 0.9840255591054313,
      "eval_loss": 0.054785601794719696,
      "eval_runtime": 88.3719,
      "eval_samples_per_second": 410.855,
      "eval_steps_per_second": 6.427,
      "step": 5400
    },
    {
      "epoch": 2.47,
      "learning_rate": 9.801929913661758e-06,
      "loss": 0.0127,
      "step": 5410
    },
    {
      "epoch": 2.48,
      "learning_rate": 9.717284577619774e-06,
      "loss": 0.0391,
      "step": 5420
    },
    {
      "epoch": 2.48,
      "learning_rate": 9.63263924157779e-06,
      "loss": 0.0242,
      "step": 5430
    },
    {
      "epoch": 2.49,
      "learning_rate": 9.547993905535805e-06,
      "loss": 0.0297,
      "step": 5440
    },
    {
      "epoch": 2.49,
      "learning_rate": 9.463348569493821e-06,
      "loss": 0.0345,
      "step": 5450
    },
    {
      "epoch": 2.5,
      "learning_rate": 9.378703233451837e-06,
      "loss": 0.0333,
      "step": 5460
    },
    {
      "epoch": 2.5,
      "learning_rate": 9.294057897409854e-06,
      "loss": 0.0293,
      "step": 5470
    },
    {
      "epoch": 2.5,
      "learning_rate": 9.20941256136787e-06,
      "loss": 0.0261,
      "step": 5480
    },
    {
      "epoch": 2.51,
      "learning_rate": 9.124767225325884e-06,
      "loss": 0.0265,
      "step": 5490
    },
    {
      "epoch": 2.51,
      "learning_rate": 9.0401218892839e-06,
      "loss": 0.0272,
      "step": 5500
    },
    {
      "epoch": 2.51,
      "eval_accuracy": 0.9840531012449047,
      "eval_loss": 0.05645028501749039,
      "eval_runtime": 87.9116,
      "eval_samples_per_second": 413.006,
      "eval_steps_per_second": 6.461,
      "step": 5500
    },
    {
      "epoch": 2.52,
      "learning_rate": 8.955476553241916e-06,
      "loss": 0.011,
      "step": 5510
    },
    {
      "epoch": 2.52,
      "learning_rate": 8.870831217199934e-06,
      "loss": 0.0234,
      "step": 5520
    },
    {
      "epoch": 2.53,
      "learning_rate": 8.78618588115795e-06,
      "loss": 0.0258,
      "step": 5530
    },
    {
      "epoch": 2.53,
      "learning_rate": 8.701540545115965e-06,
      "loss": 0.0437,
      "step": 5540
    },
    {
      "epoch": 2.54,
      "learning_rate": 8.61689520907398e-06,
      "loss": 0.0481,
      "step": 5550
    },
    {
      "epoch": 2.54,
      "learning_rate": 8.532249873031995e-06,
      "loss": 0.0288,
      "step": 5560
    },
    {
      "epoch": 2.55,
      "learning_rate": 8.447604536990013e-06,
      "loss": 0.0173,
      "step": 5570
    },
    {
      "epoch": 2.55,
      "learning_rate": 8.362959200948029e-06,
      "loss": 0.0458,
      "step": 5580
    },
    {
      "epoch": 2.55,
      "learning_rate": 8.278313864906044e-06,
      "loss": 0.0253,
      "step": 5590
    },
    {
      "epoch": 2.56,
      "learning_rate": 8.19366852886406e-06,
      "loss": 0.027,
      "step": 5600
    },
    {
      "epoch": 2.56,
      "eval_accuracy": 0.9839704748264845,
      "eval_loss": 0.05653201788663864,
      "eval_runtime": 87.7404,
      "eval_samples_per_second": 413.812,
      "eval_steps_per_second": 6.474,
      "step": 5600
    },
    {
      "epoch": 2.56,
      "learning_rate": 8.109023192822076e-06,
      "loss": 0.0341,
      "step": 5610
    },
    {
      "epoch": 2.57,
      "learning_rate": 8.024377856780092e-06,
      "loss": 0.0332,
      "step": 5620
    },
    {
      "epoch": 2.57,
      "learning_rate": 7.939732520738108e-06,
      "loss": 0.034,
      "step": 5630
    },
    {
      "epoch": 2.58,
      "learning_rate": 7.855087184696124e-06,
      "loss": 0.0132,
      "step": 5640
    },
    {
      "epoch": 2.58,
      "learning_rate": 7.77044184865414e-06,
      "loss": 0.0215,
      "step": 5650
    },
    {
      "epoch": 2.59,
      "learning_rate": 7.685796512612155e-06,
      "loss": 0.0104,
      "step": 5660
    },
    {
      "epoch": 2.59,
      "learning_rate": 7.601151176570172e-06,
      "loss": 0.0195,
      "step": 5670
    },
    {
      "epoch": 2.6,
      "learning_rate": 7.516505840528188e-06,
      "loss": 0.0369,
      "step": 5680
    },
    {
      "epoch": 2.6,
      "learning_rate": 7.431860504486203e-06,
      "loss": 0.0562,
      "step": 5690
    },
    {
      "epoch": 2.61,
      "learning_rate": 7.3472151684442185e-06,
      "loss": 0.0154,
      "step": 5700
    },
    {
      "epoch": 2.61,
      "eval_accuracy": 0.9818221879475598,
      "eval_loss": 0.0670928880572319,
      "eval_runtime": 87.5965,
      "eval_samples_per_second": 414.491,
      "eval_steps_per_second": 6.484,
      "step": 5700
    },
    {
      "epoch": 2.61,
      "learning_rate": 7.262569832402236e-06,
      "loss": 0.0328,
      "step": 5710
    },
    {
      "epoch": 2.61,
      "learning_rate": 7.177924496360251e-06,
      "loss": 0.0335,
      "step": 5720
    },
    {
      "epoch": 2.62,
      "learning_rate": 7.093279160318267e-06,
      "loss": 0.0194,
      "step": 5730
    },
    {
      "epoch": 2.62,
      "learning_rate": 7.008633824276283e-06,
      "loss": 0.0122,
      "step": 5740
    },
    {
      "epoch": 2.63,
      "learning_rate": 6.9239884882342985e-06,
      "loss": 0.0271,
      "step": 5750
    },
    {
      "epoch": 2.63,
      "learning_rate": 6.839343152192315e-06,
      "loss": 0.0202,
      "step": 5760
    },
    {
      "epoch": 2.64,
      "learning_rate": 6.754697816150331e-06,
      "loss": 0.0125,
      "step": 5770
    },
    {
      "epoch": 2.64,
      "learning_rate": 6.670052480108346e-06,
      "loss": 0.0304,
      "step": 5780
    },
    {
      "epoch": 2.65,
      "learning_rate": 6.585407144066362e-06,
      "loss": 0.0306,
      "step": 5790
    },
    {
      "epoch": 2.65,
      "learning_rate": 6.500761808024378e-06,
      "loss": 0.0315,
      "step": 5800
    },
    {
      "epoch": 2.65,
      "eval_accuracy": 0.9850997025448937,
      "eval_loss": 0.05535769835114479,
      "eval_runtime": 88.575,
      "eval_samples_per_second": 409.913,
      "eval_steps_per_second": 6.413,
      "step": 5800
    },
    {
      "epoch": 2.66,
      "learning_rate": 6.416116471982394e-06,
      "loss": 0.0322,
      "step": 5810
    },
    {
      "epoch": 2.66,
      "learning_rate": 6.33147113594041e-06,
      "loss": 0.0289,
      "step": 5820
    },
    {
      "epoch": 2.66,
      "learning_rate": 6.246825799898426e-06,
      "loss": 0.0264,
      "step": 5830
    },
    {
      "epoch": 2.67,
      "learning_rate": 6.162180463856442e-06,
      "loss": 0.0107,
      "step": 5840
    },
    {
      "epoch": 2.67,
      "learning_rate": 6.077535127814458e-06,
      "loss": 0.0198,
      "step": 5850
    },
    {
      "epoch": 2.68,
      "learning_rate": 5.9928897917724735e-06,
      "loss": 0.0494,
      "step": 5860
    },
    {
      "epoch": 2.68,
      "learning_rate": 5.90824445573049e-06,
      "loss": 0.0226,
      "step": 5870
    },
    {
      "epoch": 2.69,
      "learning_rate": 5.823599119688505e-06,
      "loss": 0.0287,
      "step": 5880
    },
    {
      "epoch": 2.69,
      "learning_rate": 5.738953783646522e-06,
      "loss": 0.0371,
      "step": 5890
    },
    {
      "epoch": 2.7,
      "learning_rate": 5.654308447604538e-06,
      "loss": 0.0351,
      "step": 5900
    },
    {
      "epoch": 2.7,
      "eval_accuracy": 0.9832268370607029,
      "eval_loss": 0.06380907446146011,
      "eval_runtime": 87.6124,
      "eval_samples_per_second": 414.416,
      "eval_steps_per_second": 6.483,
      "step": 5900
    },
    {
      "epoch": 2.7,
      "learning_rate": 5.569663111562553e-06,
      "loss": 0.0252,
      "step": 5910
    },
    {
      "epoch": 2.71,
      "learning_rate": 5.485017775520569e-06,
      "loss": 0.029,
      "step": 5920
    },
    {
      "epoch": 2.71,
      "learning_rate": 5.400372439478585e-06,
      "loss": 0.0165,
      "step": 5930
    },
    {
      "epoch": 2.71,
      "learning_rate": 5.315727103436601e-06,
      "loss": 0.0528,
      "step": 5940
    },
    {
      "epoch": 2.72,
      "learning_rate": 5.231081767394617e-06,
      "loss": 0.0362,
      "step": 5950
    },
    {
      "epoch": 2.72,
      "learning_rate": 5.146436431352633e-06,
      "loss": 0.0496,
      "step": 5960
    },
    {
      "epoch": 2.73,
      "learning_rate": 5.0617910953106484e-06,
      "loss": 0.0351,
      "step": 5970
    },
    {
      "epoch": 2.73,
      "learning_rate": 4.977145759268664e-06,
      "loss": 0.0462,
      "step": 5980
    },
    {
      "epoch": 2.74,
      "learning_rate": 4.892500423226681e-06,
      "loss": 0.0266,
      "step": 5990
    },
    {
      "epoch": 2.74,
      "learning_rate": 4.807855087184696e-06,
      "loss": 0.0216,
      "step": 6000
    },
    {
      "epoch": 2.74,
      "eval_accuracy": 0.9850997025448937,
      "eval_loss": 0.05167504400014877,
      "eval_runtime": 87.9374,
      "eval_samples_per_second": 412.885,
      "eval_steps_per_second": 6.459,
      "step": 6000
    },
    {
      "epoch": 2.75,
      "learning_rate": 4.723209751142713e-06,
      "loss": 0.0087,
      "step": 6010
    },
    {
      "epoch": 2.75,
      "learning_rate": 4.6385644151007284e-06,
      "loss": 0.0106,
      "step": 6020
    },
    {
      "epoch": 2.76,
      "learning_rate": 4.5539190790587434e-06,
      "loss": 0.0107,
      "step": 6030
    },
    {
      "epoch": 2.76,
      "learning_rate": 4.46927374301676e-06,
      "loss": 0.016,
      "step": 6040
    },
    {
      "epoch": 2.77,
      "learning_rate": 4.384628406974776e-06,
      "loss": 0.0299,
      "step": 6050
    },
    {
      "epoch": 2.77,
      "learning_rate": 4.299983070932792e-06,
      "loss": 0.0184,
      "step": 6060
    },
    {
      "epoch": 2.77,
      "learning_rate": 4.215337734890808e-06,
      "loss": 0.038,
      "step": 6070
    },
    {
      "epoch": 2.78,
      "learning_rate": 4.1306923988488234e-06,
      "loss": 0.0479,
      "step": 6080
    },
    {
      "epoch": 2.78,
      "learning_rate": 4.046047062806839e-06,
      "loss": 0.0366,
      "step": 6090
    },
    {
      "epoch": 2.79,
      "learning_rate": 3.961401726764855e-06,
      "loss": 0.0218,
      "step": 6100
    },
    {
      "epoch": 2.79,
      "eval_accuracy": 0.9844386911975322,
      "eval_loss": 0.057378485798835754,
      "eval_runtime": 87.9451,
      "eval_samples_per_second": 412.849,
      "eval_steps_per_second": 6.459,
      "step": 6100
    },
    {
      "epoch": 2.79,
      "learning_rate": 3.876756390722872e-06,
      "loss": 0.0236,
      "step": 6110
    },
    {
      "epoch": 2.8,
      "learning_rate": 3.792111054680887e-06,
      "loss": 0.0317,
      "step": 6120
    },
    {
      "epoch": 2.8,
      "learning_rate": 3.7074657186389034e-06,
      "loss": 0.017,
      "step": 6130
    },
    {
      "epoch": 2.81,
      "learning_rate": 3.6228203825969193e-06,
      "loss": 0.0365,
      "step": 6140
    },
    {
      "epoch": 2.81,
      "learning_rate": 3.5381750465549347e-06,
      "loss": 0.0356,
      "step": 6150
    },
    {
      "epoch": 2.82,
      "learning_rate": 3.453529710512951e-06,
      "loss": 0.0397,
      "step": 6160
    },
    {
      "epoch": 2.82,
      "learning_rate": 3.3688843744709667e-06,
      "loss": 0.0185,
      "step": 6170
    },
    {
      "epoch": 2.82,
      "learning_rate": 3.284239038428983e-06,
      "loss": 0.0244,
      "step": 6180
    },
    {
      "epoch": 2.83,
      "learning_rate": 3.1995937023869984e-06,
      "loss": 0.0289,
      "step": 6190
    },
    {
      "epoch": 2.83,
      "learning_rate": 3.1149483663450147e-06,
      "loss": 0.0324,
      "step": 6200
    },
    {
      "epoch": 2.83,
      "eval_accuracy": 0.9850997025448937,
      "eval_loss": 0.052561402320861816,
      "eval_runtime": 90.1689,
      "eval_samples_per_second": 402.667,
      "eval_steps_per_second": 6.299,
      "step": 6200
    },
    {
      "epoch": 2.84,
      "learning_rate": 3.0303030303030305e-06,
      "loss": 0.0276,
      "step": 6210
    },
    {
      "epoch": 2.84,
      "learning_rate": 2.9456576942610463e-06,
      "loss": 0.0166,
      "step": 6220
    },
    {
      "epoch": 2.85,
      "learning_rate": 2.861012358219062e-06,
      "loss": 0.0449,
      "step": 6230
    },
    {
      "epoch": 2.85,
      "learning_rate": 2.776367022177078e-06,
      "loss": 0.0158,
      "step": 6240
    },
    {
      "epoch": 2.86,
      "learning_rate": 2.6917216861350942e-06,
      "loss": 0.0099,
      "step": 6250
    },
    {
      "epoch": 2.86,
      "learning_rate": 2.60707635009311e-06,
      "loss": 0.0182,
      "step": 6260
    },
    {
      "epoch": 2.87,
      "learning_rate": 2.522431014051126e-06,
      "loss": 0.0416,
      "step": 6270
    },
    {
      "epoch": 2.87,
      "learning_rate": 2.437785678009142e-06,
      "loss": 0.0398,
      "step": 6280
    },
    {
      "epoch": 2.87,
      "learning_rate": 2.3531403419671576e-06,
      "loss": 0.0394,
      "step": 6290
    },
    {
      "epoch": 2.88,
      "learning_rate": 2.2684950059251734e-06,
      "loss": 0.0365,
      "step": 6300
    },
    {
      "epoch": 2.88,
      "eval_accuracy": 0.9852374132422607,
      "eval_loss": 0.05458369106054306,
      "eval_runtime": 88.1555,
      "eval_samples_per_second": 411.863,
      "eval_steps_per_second": 6.443,
      "step": 6300
    },
    {
      "epoch": 2.88,
      "learning_rate": 2.1838496698831896e-06,
      "loss": 0.0234,
      "step": 6310
    },
    {
      "epoch": 2.89,
      "learning_rate": 2.0992043338412055e-06,
      "loss": 0.0391,
      "step": 6320
    },
    {
      "epoch": 2.89,
      "learning_rate": 2.0145589977992213e-06,
      "loss": 0.0438,
      "step": 6330
    },
    {
      "epoch": 2.9,
      "learning_rate": 1.9299136617572375e-06,
      "loss": 0.0448,
      "step": 6340
    },
    {
      "epoch": 2.9,
      "learning_rate": 1.845268325715253e-06,
      "loss": 0.0332,
      "step": 6350
    },
    {
      "epoch": 2.91,
      "learning_rate": 1.760622989673269e-06,
      "loss": 0.0242,
      "step": 6360
    },
    {
      "epoch": 2.91,
      "learning_rate": 1.675977653631285e-06,
      "loss": 0.0296,
      "step": 6370
    },
    {
      "epoch": 2.92,
      "learning_rate": 1.5913323175893009e-06,
      "loss": 0.0269,
      "step": 6380
    },
    {
      "epoch": 2.92,
      "learning_rate": 1.506686981547317e-06,
      "loss": 0.0153,
      "step": 6390
    },
    {
      "epoch": 2.93,
      "learning_rate": 1.4220416455053327e-06,
      "loss": 0.0333,
      "step": 6400
    },
    {
      "epoch": 2.93,
      "eval_accuracy": 0.985402666079101,
      "eval_loss": 0.0523296594619751,
      "eval_runtime": 87.9186,
      "eval_samples_per_second": 412.973,
      "eval_steps_per_second": 6.461,
      "step": 6400
    },
    {
      "epoch": 2.93,
      "learning_rate": 1.3373963094633486e-06,
      "loss": 0.0298,
      "step": 6410
    },
    {
      "epoch": 2.93,
      "learning_rate": 1.2527509734213646e-06,
      "loss": 0.0393,
      "step": 6420
    },
    {
      "epoch": 2.94,
      "learning_rate": 1.1681056373793804e-06,
      "loss": 0.0279,
      "step": 6430
    },
    {
      "epoch": 2.94,
      "learning_rate": 1.0834603013373965e-06,
      "loss": 0.0205,
      "step": 6440
    },
    {
      "epoch": 2.95,
      "learning_rate": 9.988149652954123e-07,
      "loss": 0.034,
      "step": 6450
    },
    {
      "epoch": 2.95,
      "learning_rate": 9.141696292534281e-07,
      "loss": 0.0138,
      "step": 6460
    },
    {
      "epoch": 2.96,
      "learning_rate": 8.295242932114441e-07,
      "loss": 0.0365,
      "step": 6470
    },
    {
      "epoch": 2.96,
      "learning_rate": 7.4487895716946e-07,
      "loss": 0.0288,
      "step": 6480
    },
    {
      "epoch": 2.97,
      "learning_rate": 6.602336211274758e-07,
      "loss": 0.0335,
      "step": 6490
    },
    {
      "epoch": 2.97,
      "learning_rate": 5.755882850854919e-07,
      "loss": 0.0121,
      "step": 6500
    },
    {
      "epoch": 2.97,
      "eval_accuracy": 0.9845213176159524,
      "eval_loss": 0.057029012590646744,
      "eval_runtime": 88.7179,
      "eval_samples_per_second": 409.252,
      "eval_steps_per_second": 6.402,
      "step": 6500
    },
    {
      "epoch": 2.98,
      "learning_rate": 4.909429490435077e-07,
      "loss": 0.0213,
      "step": 6510
    },
    {
      "epoch": 2.98,
      "learning_rate": 4.062976130015236e-07,
      "loss": 0.0243,
      "step": 6520
    },
    {
      "epoch": 2.98,
      "learning_rate": 3.2165227695953954e-07,
      "loss": 0.0147,
      "step": 6530
    },
    {
      "epoch": 2.99,
      "learning_rate": 2.3700694091755545e-07,
      "loss": 0.0379,
      "step": 6540
    },
    {
      "epoch": 2.99,
      "learning_rate": 1.5236160487557138e-07,
      "loss": 0.0237,
      "step": 6550
    },
    {
      "epoch": 3.0,
      "learning_rate": 6.771626883358728e-08,
      "loss": 0.0191,
      "step": 6560
    },
    {
      "epoch": 3.0,
      "step": 6564,
      "total_flos": 1.366461317950577e+19,
      "train_loss": 0.06796654012328891,
      "train_runtime": 7965.932,
      "train_samples_per_second": 52.725,
      "train_steps_per_second": 0.824
    }
  ],
  "logging_steps": 10,
  "max_steps": 6564,
  "num_train_epochs": 3,
  "save_steps": 100,
  "total_flos": 1.366461317950577e+19,
  "trial_name": null,
  "trial_params": null
}