{
  "best_metric": 0.7649769585253456,
  "best_model_checkpoint": "videomae-base-finetuned-subset\\checkpoint-2016",
  "epoch": 59.00780780780781,
  "eval_steps": 500,
  "global_step": 6660,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 7.507507507507508e-07,
      "loss": 1.7044,
      "step": 10
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.5015015015015015e-06,
      "loss": 1.6581,
      "step": 20
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.2522522522522524e-06,
      "loss": 1.5572,
      "step": 30
    },
    {
      "epoch": 0.01,
      "learning_rate": 3.003003003003003e-06,
      "loss": 1.6749,
      "step": 40
    },
    {
      "epoch": 0.01,
      "learning_rate": 3.753753753753754e-06,
      "loss": 1.6511,
      "step": 50
    },
    {
      "epoch": 0.01,
      "learning_rate": 4.504504504504505e-06,
      "loss": 1.6406,
      "step": 60
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.255255255255255e-06,
      "loss": 1.6205,
      "step": 70
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.006006006006006e-06,
      "loss": 1.707,
      "step": 80
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.7567567567567575e-06,
      "loss": 1.5589,
      "step": 90
    },
    {
      "epoch": 0.02,
      "learning_rate": 7.507507507507508e-06,
      "loss": 1.6865,
      "step": 100
    },
    {
      "epoch": 0.02,
      "learning_rate": 8.258258258258259e-06,
      "loss": 1.6074,
      "step": 110
    },
    {
      "epoch": 0.02,
      "eval_accuracy": 0.3686635944700461,
      "eval_loss": 1.5690008401870728,
      "eval_runtime": 253.0884,
      "eval_samples_per_second": 0.857,
      "eval_steps_per_second": 0.217,
      "step": 112
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.00900900900901e-06,
      "loss": 1.5636,
      "step": 120
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.75975975975976e-06,
      "loss": 1.6139,
      "step": 130
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.051051051051051e-05,
      "loss": 1.591,
      "step": 140
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.1261261261261261e-05,
      "loss": 1.5824,
      "step": 150
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.2012012012012012e-05,
      "loss": 1.6132,
      "step": 160
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.2762762762762764e-05,
      "loss": 1.6279,
      "step": 170
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.3513513513513515e-05,
      "loss": 1.6001,
      "step": 180
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.4264264264264266e-05,
      "loss": 1.5873,
      "step": 190
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.5015015015015016e-05,
      "loss": 1.59,
      "step": 200
    },
    {
      "epoch": 1.01,
      "learning_rate": 1.5765765765765765e-05,
      "loss": 1.5691,
      "step": 210
    },
    {
      "epoch": 1.02,
      "learning_rate": 1.6516516516516518e-05,
      "loss": 1.6001,
      "step": 220
    },
    {
      "epoch": 1.02,
      "eval_accuracy": 0.30414746543778803,
      "eval_loss": 1.5782501697540283,
      "eval_runtime": 245.0114,
      "eval_samples_per_second": 0.886,
      "eval_steps_per_second": 0.224,
      "step": 224
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.7267267267267267e-05,
      "loss": 1.5771,
      "step": 230
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.801801801801802e-05,
      "loss": 1.5651,
      "step": 240
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.8768768768768768e-05,
      "loss": 1.4911,
      "step": 250
    },
    {
      "epoch": 2.01,
      "learning_rate": 1.951951951951952e-05,
      "loss": 1.6414,
      "step": 260
    },
    {
      "epoch": 2.01,
      "learning_rate": 2.0270270270270273e-05,
      "loss": 1.5454,
      "step": 270
    },
    {
      "epoch": 2.01,
      "learning_rate": 2.102102102102102e-05,
      "loss": 1.4563,
      "step": 280
    },
    {
      "epoch": 2.01,
      "learning_rate": 2.1771771771771774e-05,
      "loss": 1.4325,
      "step": 290
    },
    {
      "epoch": 2.01,
      "learning_rate": 2.2522522522522523e-05,
      "loss": 1.4738,
      "step": 300
    },
    {
      "epoch": 2.01,
      "learning_rate": 2.3273273273273275e-05,
      "loss": 1.3045,
      "step": 310
    },
    {
      "epoch": 2.01,
      "learning_rate": 2.4024024024024024e-05,
      "loss": 1.2273,
      "step": 320
    },
    {
      "epoch": 2.02,
      "learning_rate": 2.4774774774774777e-05,
      "loss": 1.4193,
      "step": 330
    },
    {
      "epoch": 2.02,
      "eval_accuracy": 0.3824884792626728,
      "eval_loss": 1.4873727560043335,
      "eval_runtime": 251.8516,
      "eval_samples_per_second": 0.862,
      "eval_steps_per_second": 0.218,
      "step": 336
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.552552552552553e-05,
      "loss": 1.5274,
      "step": 340
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.6276276276276278e-05,
      "loss": 1.3287,
      "step": 350
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.702702702702703e-05,
      "loss": 1.4428,
      "step": 360
    },
    {
      "epoch": 3.01,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.2924,
      "step": 370
    },
    {
      "epoch": 3.01,
      "learning_rate": 2.852852852852853e-05,
      "loss": 1.204,
      "step": 380
    },
    {
      "epoch": 3.01,
      "learning_rate": 2.927927927927928e-05,
      "loss": 1.1626,
      "step": 390
    },
    {
      "epoch": 3.01,
      "learning_rate": 3.0030030030030033e-05,
      "loss": 1.5428,
      "step": 400
    },
    {
      "epoch": 3.01,
      "learning_rate": 3.078078078078078e-05,
      "loss": 1.1494,
      "step": 410
    },
    {
      "epoch": 3.01,
      "learning_rate": 3.153153153153153e-05,
      "loss": 1.2899,
      "step": 420
    },
    {
      "epoch": 3.01,
      "learning_rate": 3.2282282282282286e-05,
      "loss": 1.4876,
      "step": 430
    },
    {
      "epoch": 3.02,
      "learning_rate": 3.3033033033033035e-05,
      "loss": 1.398,
      "step": 440
    },
    {
      "epoch": 3.02,
      "eval_accuracy": 0.6405529953917051,
      "eval_loss": 1.0196824073791504,
      "eval_runtime": 247.8505,
      "eval_samples_per_second": 0.876,
      "eval_steps_per_second": 0.222,
      "step": 448
    },
    {
      "epoch": 4.0,
      "learning_rate": 3.3783783783783784e-05,
      "loss": 1.3202,
      "step": 450
    },
    {
      "epoch": 4.0,
      "learning_rate": 3.453453453453453e-05,
      "loss": 1.145,
      "step": 460
    },
    {
      "epoch": 4.0,
      "learning_rate": 3.528528528528528e-05,
      "loss": 1.2636,
      "step": 470
    },
    {
      "epoch": 4.0,
      "learning_rate": 3.603603603603604e-05,
      "loss": 1.2266,
      "step": 480
    },
    {
      "epoch": 4.01,
      "learning_rate": 3.678678678678679e-05,
      "loss": 1.6096,
      "step": 490
    },
    {
      "epoch": 4.01,
      "learning_rate": 3.7537537537537536e-05,
      "loss": 1.1474,
      "step": 500
    },
    {
      "epoch": 4.01,
      "learning_rate": 3.8288288288288285e-05,
      "loss": 1.1879,
      "step": 510
    },
    {
      "epoch": 4.01,
      "learning_rate": 3.903903903903904e-05,
      "loss": 1.0165,
      "step": 520
    },
    {
      "epoch": 4.01,
      "learning_rate": 3.9789789789789796e-05,
      "loss": 1.2291,
      "step": 530
    },
    {
      "epoch": 4.01,
      "learning_rate": 4.0540540540540545e-05,
      "loss": 1.2309,
      "step": 540
    },
    {
      "epoch": 4.02,
      "learning_rate": 4.1291291291291294e-05,
      "loss": 1.0638,
      "step": 550
    },
    {
      "epoch": 4.02,
      "learning_rate": 4.204204204204204e-05,
      "loss": 1.2217,
      "step": 560
    },
    {
      "epoch": 4.02,
      "eval_accuracy": 0.391705069124424,
      "eval_loss": 1.3385610580444336,
      "eval_runtime": 242.8726,
      "eval_samples_per_second": 0.893,
      "eval_steps_per_second": 0.226,
      "step": 560
    },
    {
      "epoch": 5.0,
      "learning_rate": 4.27927927927928e-05,
      "loss": 1.4069,
      "step": 570
    },
    {
      "epoch": 5.0,
      "learning_rate": 4.354354354354355e-05,
      "loss": 1.26,
      "step": 580
    },
    {
      "epoch": 5.0,
      "learning_rate": 4.42942942942943e-05,
      "loss": 1.2424,
      "step": 590
    },
    {
      "epoch": 5.01,
      "learning_rate": 4.5045045045045046e-05,
      "loss": 1.2978,
      "step": 600
    },
    {
      "epoch": 5.01,
      "learning_rate": 4.57957957957958e-05,
      "loss": 1.373,
      "step": 610
    },
    {
      "epoch": 5.01,
      "learning_rate": 4.654654654654655e-05,
      "loss": 1.2995,
      "step": 620
    },
    {
      "epoch": 5.01,
      "learning_rate": 4.72972972972973e-05,
      "loss": 1.053,
      "step": 630
    },
    {
      "epoch": 5.01,
      "learning_rate": 4.804804804804805e-05,
      "loss": 1.1977,
      "step": 640
    },
    {
      "epoch": 5.01,
      "learning_rate": 4.87987987987988e-05,
      "loss": 1.2698,
      "step": 650
    },
    {
      "epoch": 5.02,
      "learning_rate": 4.954954954954955e-05,
      "loss": 1.1812,
      "step": 660
    },
    {
      "epoch": 5.02,
      "learning_rate": 4.9966633299966636e-05,
      "loss": 1.2577,
      "step": 670
    },
    {
      "epoch": 5.02,
      "eval_accuracy": 0.5391705069124424,
      "eval_loss": 1.2195504903793335,
      "eval_runtime": 264.2835,
      "eval_samples_per_second": 0.821,
      "eval_steps_per_second": 0.208,
      "step": 672
    },
    {
      "epoch": 6.0,
      "learning_rate": 4.988321654988322e-05,
      "loss": 1.4038,
      "step": 680
    },
    {
      "epoch": 6.0,
      "learning_rate": 4.97997997997998e-05,
      "loss": 1.4382,
      "step": 690
    },
    {
      "epoch": 6.0,
      "learning_rate": 4.9716383049716386e-05,
      "loss": 1.3709,
      "step": 700
    },
    {
      "epoch": 6.01,
      "learning_rate": 4.963296629963297e-05,
      "loss": 1.0241,
      "step": 710
    },
    {
      "epoch": 6.01,
      "learning_rate": 4.954954954954955e-05,
      "loss": 1.5423,
      "step": 720
    },
    {
      "epoch": 6.01,
      "learning_rate": 4.9466132799466136e-05,
      "loss": 1.2711,
      "step": 730
    },
    {
      "epoch": 6.01,
      "learning_rate": 4.938271604938271e-05,
      "loss": 1.4259,
      "step": 740
    },
    {
      "epoch": 6.01,
      "learning_rate": 4.92992992992993e-05,
      "loss": 1.2684,
      "step": 750
    },
    {
      "epoch": 6.01,
      "learning_rate": 4.921588254921589e-05,
      "loss": 0.9547,
      "step": 760
    },
    {
      "epoch": 6.01,
      "learning_rate": 4.9132465799132463e-05,
      "loss": 1.1885,
      "step": 770
    },
    {
      "epoch": 6.02,
      "learning_rate": 4.9049049049049054e-05,
      "loss": 1.0121,
      "step": 780
    },
    {
      "epoch": 6.02,
      "eval_accuracy": 0.4792626728110599,
      "eval_loss": 1.2319154739379883,
      "eval_runtime": 261.8591,
      "eval_samples_per_second": 0.829,
      "eval_steps_per_second": 0.21,
      "step": 784
    },
    {
      "epoch": 7.0,
      "learning_rate": 4.896563229896564e-05,
      "loss": 1.355,
      "step": 790
    },
    {
      "epoch": 7.0,
      "learning_rate": 4.8882215548882214e-05,
      "loss": 1.4279,
      "step": 800
    },
    {
      "epoch": 7.0,
      "learning_rate": 4.87987987987988e-05,
      "loss": 1.0318,
      "step": 810
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.871538204871539e-05,
      "loss": 1.1375,
      "step": 820
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.8631965298631964e-05,
      "loss": 1.1721,
      "step": 830
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.854854854854855e-05,
      "loss": 1.2527,
      "step": 840
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.846513179846514e-05,
      "loss": 1.0756,
      "step": 850
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.8381715048381715e-05,
      "loss": 1.1513,
      "step": 860
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.82982982982983e-05,
      "loss": 1.0789,
      "step": 870
    },
    {
      "epoch": 7.01,
      "learning_rate": 4.821488154821489e-05,
      "loss": 1.2463,
      "step": 880
    },
    {
      "epoch": 7.02,
      "learning_rate": 4.8131464798131465e-05,
      "loss": 1.2485,
      "step": 890
    },
    {
      "epoch": 7.02,
      "eval_accuracy": 0.7511520737327189,
      "eval_loss": 0.8229795098304749,
      "eval_runtime": 252.4438,
      "eval_samples_per_second": 0.86,
      "eval_steps_per_second": 0.218,
      "step": 896
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.804804804804805e-05,
      "loss": 0.9738,
      "step": 900
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.796463129796463e-05,
      "loss": 0.7946,
      "step": 910
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.7881214547881215e-05,
      "loss": 1.2819,
      "step": 920
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.77977977977978e-05,
      "loss": 1.7354,
      "step": 930
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.771438104771438e-05,
      "loss": 1.161,
      "step": 940
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.7630964297630966e-05,
      "loss": 1.3311,
      "step": 950
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.754754754754755e-05,
      "loss": 0.9299,
      "step": 960
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.746413079746413e-05,
      "loss": 0.9622,
      "step": 970
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.7380714047380716e-05,
      "loss": 0.9557,
      "step": 980
    },
    {
      "epoch": 8.01,
      "learning_rate": 4.72972972972973e-05,
      "loss": 0.9241,
      "step": 990
    },
    {
      "epoch": 8.02,
      "learning_rate": 4.721388054721388e-05,
      "loss": 1.025,
      "step": 1000
    },
    {
      "epoch": 8.02,
      "eval_accuracy": 0.6866359447004609,
      "eval_loss": 0.8022767305374146,
      "eval_runtime": 256.9331,
      "eval_samples_per_second": 0.845,
      "eval_steps_per_second": 0.214,
      "step": 1008
    },
    {
      "epoch": 9.0,
      "learning_rate": 4.7130463797130466e-05,
      "loss": 1.6034,
      "step": 1010
    },
    {
      "epoch": 9.0,
      "learning_rate": 4.704704704704705e-05,
      "loss": 1.0115,
      "step": 1020
    },
    {
      "epoch": 9.0,
      "learning_rate": 4.696363029696363e-05,
      "loss": 1.337,
      "step": 1030
    },
    {
      "epoch": 9.0,
      "learning_rate": 4.688021354688022e-05,
      "loss": 1.351,
      "step": 1040
    },
    {
      "epoch": 9.01,
      "learning_rate": 4.67967967967968e-05,
      "loss": 1.0992,
      "step": 1050
    },
    {
      "epoch": 9.01,
      "learning_rate": 4.6713380046713384e-05,
      "loss": 1.0828,
      "step": 1060
    },
    {
      "epoch": 9.01,
      "learning_rate": 4.662996329662997e-05,
      "loss": 1.0086,
      "step": 1070
    },
    {
      "epoch": 9.01,
      "learning_rate": 4.654654654654655e-05,
      "loss": 0.8658,
      "step": 1080
    },
    {
      "epoch": 9.01,
      "learning_rate": 4.6463129796463134e-05,
      "loss": 0.7355,
      "step": 1090
    },
    {
      "epoch": 9.01,
      "learning_rate": 4.637971304637971e-05,
      "loss": 1.0191,
      "step": 1100
    },
    {
      "epoch": 9.02,
      "learning_rate": 4.62962962962963e-05,
      "loss": 1.0171,
      "step": 1110
    },
    {
      "epoch": 9.02,
      "learning_rate": 4.6212879546212884e-05,
      "loss": 1.2952,
      "step": 1120
    },
    {
      "epoch": 9.02,
      "eval_accuracy": 0.6036866359447005,
      "eval_loss": 0.9130178093910217,
      "eval_runtime": 256.4943,
      "eval_samples_per_second": 0.846,
      "eval_steps_per_second": 0.214,
      "step": 1120
    },
    {
      "epoch": 10.0,
      "learning_rate": 4.612946279612946e-05,
      "loss": 0.8975,
      "step": 1130
    },
    {
      "epoch": 10.0,
      "learning_rate": 4.604604604604605e-05,
      "loss": 1.0215,
      "step": 1140
    },
    {
      "epoch": 10.0,
      "learning_rate": 4.5962629295962635e-05,
      "loss": 0.9562,
      "step": 1150
    },
    {
      "epoch": 10.01,
      "learning_rate": 4.587921254587921e-05,
      "loss": 1.2006,
      "step": 1160
    },
    {
      "epoch": 10.01,
      "learning_rate": 4.57957957957958e-05,
      "loss": 1.2811,
      "step": 1170
    },
    {
      "epoch": 10.01,
      "learning_rate": 4.5712379045712385e-05,
      "loss": 1.0057,
      "step": 1180
    },
    {
      "epoch": 10.01,
      "learning_rate": 4.562896229562896e-05,
      "loss": 1.0403,
      "step": 1190
    },
    {
      "epoch": 10.01,
      "learning_rate": 4.5545545545545545e-05,
      "loss": 0.8273,
      "step": 1200
    },
    {
      "epoch": 10.01,
      "learning_rate": 4.5462128795462135e-05,
      "loss": 0.7807,
      "step": 1210
    },
    {
      "epoch": 10.02,
      "learning_rate": 4.537871204537871e-05,
      "loss": 0.9059,
      "step": 1220
    },
    {
      "epoch": 10.02,
      "learning_rate": 4.5295295295295295e-05,
      "loss": 0.9499,
      "step": 1230
    },
    {
      "epoch": 10.02,
      "eval_accuracy": 0.6036866359447005,
      "eval_loss": 1.0620718002319336,
      "eval_runtime": 255.1418,
      "eval_samples_per_second": 0.851,
      "eval_steps_per_second": 0.216,
      "step": 1232
    },
    {
      "epoch": 11.0,
      "learning_rate": 4.5211878545211886e-05,
      "loss": 1.0751,
      "step": 1240
    },
    {
      "epoch": 11.0,
      "learning_rate": 4.512846179512846e-05,
      "loss": 1.034,
      "step": 1250
    },
    {
      "epoch": 11.0,
      "learning_rate": 4.5045045045045046e-05,
      "loss": 0.6693,
      "step": 1260
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.4961628294961636e-05,
      "loss": 1.0748,
      "step": 1270
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.487821154487821e-05,
      "loss": 0.9789,
      "step": 1280
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.4794794794794796e-05,
      "loss": 0.8083,
      "step": 1290
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.471137804471138e-05,
      "loss": 0.9976,
      "step": 1300
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.462796129462796e-05,
      "loss": 0.7674,
      "step": 1310
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.4544544544544546e-05,
      "loss": 0.8705,
      "step": 1320
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.446112779446113e-05,
      "loss": 1.195,
      "step": 1330
    },
    {
      "epoch": 11.02,
      "learning_rate": 4.437771104437771e-05,
      "loss": 0.8805,
      "step": 1340
    },
    {
      "epoch": 11.02,
      "eval_accuracy": 0.7050691244239631,
      "eval_loss": 0.871336042881012,
      "eval_runtime": 256.7263,
      "eval_samples_per_second": 0.845,
      "eval_steps_per_second": 0.214,
      "step": 1344
    },
    {
      "epoch": 12.0,
      "learning_rate": 4.42942942942943e-05,
      "loss": 0.8859,
      "step": 1350
    },
    {
      "epoch": 12.0,
      "learning_rate": 4.421087754421088e-05,
      "loss": 1.2024,
      "step": 1360
    },
    {
      "epoch": 12.0,
      "learning_rate": 4.4127460794127464e-05,
      "loss": 0.9896,
      "step": 1370
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.404404404404405e-05,
      "loss": 0.9208,
      "step": 1380
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.3960627293960624e-05,
      "loss": 1.2375,
      "step": 1390
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.3877210543877214e-05,
      "loss": 0.9493,
      "step": 1400
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.37937937937938e-05,
      "loss": 0.9476,
      "step": 1410
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.3710377043710374e-05,
      "loss": 0.8137,
      "step": 1420
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.3626960293626964e-05,
      "loss": 0.854,
      "step": 1430
    },
    {
      "epoch": 12.01,
      "learning_rate": 4.354354354354355e-05,
      "loss": 1.0682,
      "step": 1440
    },
    {
      "epoch": 12.02,
      "learning_rate": 4.3460126793460125e-05,
      "loss": 1.2066,
      "step": 1450
    },
    {
      "epoch": 12.02,
      "eval_accuracy": 0.5852534562211982,
      "eval_loss": 0.9363781809806824,
      "eval_runtime": 263.3253,
      "eval_samples_per_second": 0.824,
      "eval_steps_per_second": 0.209,
      "step": 1456
    },
    {
      "epoch": 13.0,
      "learning_rate": 4.3376710043376715e-05,
      "loss": 0.7223,
      "step": 1460
    },
    {
      "epoch": 13.0,
      "learning_rate": 4.32932932932933e-05,
      "loss": 0.7549,
      "step": 1470
    },
    {
      "epoch": 13.0,
      "learning_rate": 4.3209876543209875e-05,
      "loss": 0.9539,
      "step": 1480
    },
    {
      "epoch": 13.01,
      "learning_rate": 4.312645979312646e-05,
      "loss": 0.8501,
      "step": 1490
    },
    {
      "epoch": 13.01,
      "learning_rate": 4.304304304304305e-05,
      "loss": 0.9684,
      "step": 1500
    },
    {
      "epoch": 13.01,
      "learning_rate": 4.2959626292959625e-05,
      "loss": 1.3551,
      "step": 1510
    },
    {
      "epoch": 13.01,
      "learning_rate": 4.287620954287621e-05,
      "loss": 0.9607,
      "step": 1520
    },
    {
      "epoch": 13.01,
      "learning_rate": 4.27927927927928e-05,
      "loss": 0.8907,
      "step": 1530
    },
    {
      "epoch": 13.01,
      "learning_rate": 4.2709376042709376e-05,
      "loss": 1.2041,
      "step": 1540
    },
    {
      "epoch": 13.01,
      "learning_rate": 4.262595929262596e-05,
      "loss": 0.7261,
      "step": 1550
    },
    {
      "epoch": 13.02,
      "learning_rate": 4.254254254254255e-05,
      "loss": 0.9358,
      "step": 1560
    },
    {
      "epoch": 13.02,
      "eval_accuracy": 0.5852534562211982,
      "eval_loss": 0.910686731338501,
      "eval_runtime": 265.3055,
      "eval_samples_per_second": 0.818,
      "eval_steps_per_second": 0.207,
      "step": 1568
    },
    {
      "epoch": 14.0,
      "learning_rate": 4.2459125792459126e-05,
      "loss": 0.9092,
      "step": 1570
    },
    {
      "epoch": 14.0,
      "learning_rate": 4.237570904237571e-05,
      "loss": 0.8376,
      "step": 1580
    },
    {
      "epoch": 14.0,
      "learning_rate": 4.229229229229229e-05,
      "loss": 1.0121,
      "step": 1590
    },
    {
      "epoch": 14.0,
      "learning_rate": 4.2208875542208876e-05,
      "loss": 0.9361,
      "step": 1600
    },
    {
      "epoch": 14.01,
      "learning_rate": 4.212545879212546e-05,
      "loss": 1.053,
      "step": 1610
    },
    {
      "epoch": 14.01,
      "learning_rate": 4.204204204204204e-05,
      "loss": 0.9652,
      "step": 1620
    },
    {
      "epoch": 14.01,
      "learning_rate": 4.1958625291958627e-05,
      "loss": 0.7609,
      "step": 1630
    },
    {
      "epoch": 14.01,
      "learning_rate": 4.187520854187521e-05,
      "loss": 1.0871,
      "step": 1640
    },
    {
      "epoch": 14.01,
      "learning_rate": 4.1791791791791793e-05,
      "loss": 0.9353,
      "step": 1650
    },
    {
      "epoch": 14.01,
      "learning_rate": 4.170837504170838e-05,
      "loss": 0.7793,
      "step": 1660
    },
    {
      "epoch": 14.02,
      "learning_rate": 4.162495829162496e-05,
      "loss": 1.0073,
      "step": 1670
    },
    {
      "epoch": 14.02,
      "learning_rate": 4.1541541541541544e-05,
      "loss": 0.9043,
      "step": 1680
    },
    {
      "epoch": 14.02,
      "eval_accuracy": 0.6359447004608295,
      "eval_loss": 0.9146963953971863,
      "eval_runtime": 259.9831,
      "eval_samples_per_second": 0.835,
      "eval_steps_per_second": 0.212,
      "step": 1680
    },
    {
      "epoch": 15.0,
      "learning_rate": 4.145812479145813e-05,
      "loss": 1.0818,
      "step": 1690
    },
    {
      "epoch": 15.0,
      "learning_rate": 4.137470804137471e-05,
      "loss": 0.9315,
      "step": 1700
    },
    {
      "epoch": 15.0,
      "learning_rate": 4.1291291291291294e-05,
      "loss": 0.8777,
      "step": 1710
    },
    {
      "epoch": 15.01,
      "learning_rate": 4.120787454120788e-05,
      "loss": 1.0695,
      "step": 1720
    },
    {
      "epoch": 15.01,
      "learning_rate": 4.112445779112446e-05,
      "loss": 0.6343,
      "step": 1730
    },
    {
      "epoch": 15.01,
      "learning_rate": 4.1041041041041045e-05,
      "loss": 1.0106,
      "step": 1740
    },
    {
      "epoch": 15.01,
      "learning_rate": 4.095762429095763e-05,
      "loss": 0.9499,
      "step": 1750
    },
    {
      "epoch": 15.01,
      "learning_rate": 4.087420754087421e-05,
      "loss": 0.9108,
      "step": 1760
    },
    {
      "epoch": 15.01,
      "learning_rate": 4.0790790790790795e-05,
      "loss": 0.8515,
      "step": 1770
    },
    {
      "epoch": 15.02,
      "learning_rate": 4.070737404070737e-05,
      "loss": 0.749,
      "step": 1780
    },
    {
      "epoch": 15.02,
      "learning_rate": 4.062395729062396e-05,
      "loss": 0.8383,
      "step": 1790
    },
    {
      "epoch": 15.02,
      "eval_accuracy": 0.6359447004608295,
      "eval_loss": 0.945084810256958,
      "eval_runtime": 259.6462,
      "eval_samples_per_second": 0.836,
      "eval_steps_per_second": 0.212,
      "step": 1792
    },
    {
      "epoch": 16.0,
      "learning_rate": 4.0540540540540545e-05,
      "loss": 0.8241,
      "step": 1800
    },
    {
      "epoch": 16.0,
      "learning_rate": 4.045712379045712e-05,
      "loss": 0.7361,
      "step": 1810
    },
    {
      "epoch": 16.0,
      "learning_rate": 4.037370704037371e-05,
      "loss": 0.978,
      "step": 1820
    },
    {
      "epoch": 16.01,
      "learning_rate": 4.0290290290290296e-05,
      "loss": 0.7197,
      "step": 1830
    },
    {
      "epoch": 16.01,
      "learning_rate": 4.020687354020687e-05,
      "loss": 1.0322,
      "step": 1840
    },
    {
      "epoch": 16.01,
      "learning_rate": 4.012345679012346e-05,
      "loss": 0.8089,
      "step": 1850
    },
    {
      "epoch": 16.01,
      "learning_rate": 4.0040040040040046e-05,
      "loss": 0.8855,
      "step": 1860
    },
    {
      "epoch": 16.01,
      "learning_rate": 3.995662328995662e-05,
      "loss": 0.9764,
      "step": 1870
    },
    {
      "epoch": 16.01,
      "learning_rate": 3.9873206539873206e-05,
      "loss": 0.7759,
      "step": 1880
    },
    {
      "epoch": 16.01,
      "learning_rate": 3.9789789789789796e-05,
      "loss": 0.5536,
      "step": 1890
    },
    {
      "epoch": 16.02,
      "learning_rate": 3.970637303970637e-05,
      "loss": 0.7482,
      "step": 1900
    },
    {
      "epoch": 16.02,
      "eval_accuracy": 0.6221198156682027,
      "eval_loss": 0.876456618309021,
      "eval_runtime": 273.3147,
      "eval_samples_per_second": 0.794,
      "eval_steps_per_second": 0.201,
      "step": 1904
    },
    {
      "epoch": 17.0,
      "learning_rate": 3.9622956289622956e-05,
      "loss": 0.6662,
      "step": 1910
    },
    {
      "epoch": 17.0,
      "learning_rate": 3.953953953953955e-05,
      "loss": 0.857,
      "step": 1920
    },
    {
      "epoch": 17.0,
      "learning_rate": 3.945612278945612e-05,
      "loss": 0.8507,
      "step": 1930
    },
    {
      "epoch": 17.01,
      "learning_rate": 3.937270603937271e-05,
      "loss": 1.1261,
      "step": 1940
    },
    {
      "epoch": 17.01,
      "learning_rate": 3.92892892892893e-05,
      "loss": 0.9851,
      "step": 1950
    },
    {
      "epoch": 17.01,
      "learning_rate": 3.9205872539205874e-05,
      "loss": 0.5859,
      "step": 1960
    },
    {
      "epoch": 17.01,
      "learning_rate": 3.912245578912246e-05,
      "loss": 0.9539,
      "step": 1970
    },
    {
      "epoch": 17.01,
      "learning_rate": 3.903903903903904e-05,
      "loss": 0.8663,
      "step": 1980
    },
    {
      "epoch": 17.01,
      "learning_rate": 3.8955622288955624e-05,
      "loss": 0.9946,
      "step": 1990
    },
    {
      "epoch": 17.01,
      "learning_rate": 3.887220553887221e-05,
      "loss": 0.835,
      "step": 2000
    },
    {
      "epoch": 17.02,
      "learning_rate": 3.878878878878879e-05,
      "loss": 0.9547,
      "step": 2010
    },
    {
      "epoch": 17.02,
      "eval_accuracy": 0.7649769585253456,
      "eval_loss": 0.7997832894325256,
      "eval_runtime": 275.0933,
      "eval_samples_per_second": 0.789,
      "eval_steps_per_second": 0.2,
      "step": 2016
    },
    {
      "epoch": 18.0,
      "learning_rate": 3.8705372038705374e-05,
      "loss": 0.8025,
      "step": 2020
    },
    {
      "epoch": 18.0,
      "learning_rate": 3.862195528862196e-05,
      "loss": 0.9637,
      "step": 2030
    },
    {
      "epoch": 18.0,
      "learning_rate": 3.8538538538538534e-05,
      "loss": 1.1231,
      "step": 2040
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.8455121788455125e-05,
      "loss": 0.6693,
      "step": 2050
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.837170503837171e-05,
      "loss": 1.0535,
      "step": 2060
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.8288288288288285e-05,
      "loss": 0.9,
      "step": 2070
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.8204871538204875e-05,
      "loss": 0.7855,
      "step": 2080
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.812145478812146e-05,
      "loss": 0.7349,
      "step": 2090
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.8038038038038035e-05,
      "loss": 0.9599,
      "step": 2100
    },
    {
      "epoch": 18.01,
      "learning_rate": 3.7954621287954625e-05,
      "loss": 0.6589,
      "step": 2110
    },
    {
      "epoch": 18.02,
      "learning_rate": 3.787120453787121e-05,
      "loss": 0.7028,
      "step": 2120
    },
    {
      "epoch": 18.02,
      "eval_accuracy": 0.6405529953917051,
      "eval_loss": 0.9256826639175415,
      "eval_runtime": 263.4393,
      "eval_samples_per_second": 0.824,
      "eval_steps_per_second": 0.209,
      "step": 2128
    },
    {
      "epoch": 19.0,
      "learning_rate": 3.7787787787787786e-05,
      "loss": 0.5654,
      "step": 2130
    },
    {
      "epoch": 19.0,
      "learning_rate": 3.7704371037704376e-05,
      "loss": 0.8278,
      "step": 2140
    },
    {
      "epoch": 19.0,
      "learning_rate": 3.762095428762096e-05,
      "loss": 1.0629,
      "step": 2150
    },
    {
      "epoch": 19.0,
      "learning_rate": 3.7537537537537536e-05,
      "loss": 1.026,
      "step": 2160
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.745412078745412e-05,
      "loss": 0.8544,
      "step": 2170
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.737070403737071e-05,
      "loss": 0.8669,
      "step": 2180
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.7287287287287286e-05,
      "loss": 0.6044,
      "step": 2190
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.720387053720387e-05,
      "loss": 0.8129,
      "step": 2200
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.712045378712046e-05,
      "loss": 0.5283,
      "step": 2210
    },
    {
      "epoch": 19.01,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 1.0115,
      "step": 2220
    },
    {
      "epoch": 19.02,
      "learning_rate": 3.695362028695362e-05,
      "loss": 0.7743,
      "step": 2230
    },
    {
      "epoch": 19.02,
      "learning_rate": 3.687020353687021e-05,
      "loss": 0.8659,
      "step": 2240
    },
    {
      "epoch": 19.02,
      "eval_accuracy": 0.5852534562211982,
      "eval_loss": 1.0655064582824707,
      "eval_runtime": 271.2554,
      "eval_samples_per_second": 0.8,
      "eval_steps_per_second": 0.203,
      "step": 2240
    },
    {
      "epoch": 20.0,
      "learning_rate": 3.678678678678679e-05,
      "loss": 0.8819,
      "step": 2250
    },
    {
      "epoch": 20.0,
      "learning_rate": 3.670337003670337e-05,
      "loss": 0.7077,
      "step": 2260
    },
    {
      "epoch": 20.0,
      "learning_rate": 3.6619953286619954e-05,
      "loss": 0.737,
      "step": 2270
    },
    {
      "epoch": 20.01,
      "learning_rate": 3.653653653653654e-05,
      "loss": 0.8414,
      "step": 2280
    },
    {
      "epoch": 20.01,
      "learning_rate": 3.645311978645312e-05,
      "loss": 0.8566,
      "step": 2290
    },
    {
      "epoch": 20.01,
      "learning_rate": 3.6369703036369704e-05,
      "loss": 0.8155,
      "step": 2300
    },
    {
      "epoch": 20.01,
      "learning_rate": 3.628628628628629e-05,
      "loss": 0.6651,
      "step": 2310
    },
    {
      "epoch": 20.01,
      "learning_rate": 3.620286953620287e-05,
      "loss": 0.7734,
      "step": 2320
    },
    {
      "epoch": 20.01,
      "learning_rate": 3.6119452786119454e-05,
      "loss": 0.9056,
      "step": 2330
    },
    {
      "epoch": 20.02,
      "learning_rate": 3.603603603603604e-05,
      "loss": 0.9031,
      "step": 2340
    },
    {
      "epoch": 20.02,
      "learning_rate": 3.595261928595262e-05,
      "loss": 0.5591,
      "step": 2350
    },
    {
      "epoch": 20.02,
      "eval_accuracy": 0.576036866359447,
      "eval_loss": 1.2793947458267212,
      "eval_runtime": 265.9827,
      "eval_samples_per_second": 0.816,
      "eval_steps_per_second": 0.207,
      "step": 2352
    },
    {
      "epoch": 21.0,
      "learning_rate": 3.5869202535869205e-05,
      "loss": 0.6663,
      "step": 2360
    },
    {
      "epoch": 21.0,
      "learning_rate": 3.578578578578579e-05,
      "loss": 0.6847,
      "step": 2370
    },
    {
      "epoch": 21.0,
      "learning_rate": 3.570236903570237e-05,
      "loss": 0.8432,
      "step": 2380
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.5618952285618955e-05,
      "loss": 0.6488,
      "step": 2390
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.553553553553554e-05,
      "loss": 0.9092,
      "step": 2400
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.545211878545212e-05,
      "loss": 0.8903,
      "step": 2410
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.5368702035368706e-05,
      "loss": 0.4662,
      "step": 2420
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.528528528528528e-05,
      "loss": 0.9775,
      "step": 2430
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.520186853520187e-05,
      "loss": 0.973,
      "step": 2440
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.5118451785118456e-05,
      "loss": 0.5741,
      "step": 2450
    },
    {
      "epoch": 21.02,
      "learning_rate": 3.503503503503503e-05,
      "loss": 0.8963,
      "step": 2460
    },
    {
      "epoch": 21.02,
      "eval_accuracy": 0.695852534562212,
      "eval_loss": 1.0049302577972412,
      "eval_runtime": 270.1661,
      "eval_samples_per_second": 0.803,
      "eval_steps_per_second": 0.204,
      "step": 2464
    },
    {
      "epoch": 22.0,
      "learning_rate": 3.495161828495162e-05,
      "loss": 0.7073,
      "step": 2470
    },
    {
      "epoch": 22.0,
      "learning_rate": 3.4868201534868206e-05,
      "loss": 0.5915,
      "step": 2480
    },
    {
      "epoch": 22.0,
      "learning_rate": 3.478478478478478e-05,
      "loss": 0.3567,
      "step": 2490
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.470136803470137e-05,
      "loss": 0.7007,
      "step": 2500
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.4617951284617957e-05,
      "loss": 0.4253,
      "step": 2510
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.453453453453453e-05,
      "loss": 0.9297,
      "step": 2520
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.4451117784451123e-05,
      "loss": 0.9365,
      "step": 2530
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.436770103436771e-05,
      "loss": 0.4607,
      "step": 2540
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.4284284284284284e-05,
      "loss": 0.8208,
      "step": 2550
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.420086753420087e-05,
      "loss": 0.8313,
      "step": 2560
    },
    {
      "epoch": 22.02,
      "learning_rate": 3.411745078411746e-05,
      "loss": 0.9221,
      "step": 2570
    },
    {
      "epoch": 22.02,
      "eval_accuracy": 0.6082949308755761,
      "eval_loss": 1.111289620399475,
      "eval_runtime": 272.9644,
      "eval_samples_per_second": 0.795,
      "eval_steps_per_second": 0.201,
      "step": 2576
    },
    {
      "epoch": 23.0,
      "learning_rate": 3.4034034034034034e-05,
      "loss": 0.8386,
      "step": 2580
    },
    {
      "epoch": 23.0,
      "learning_rate": 3.395061728395062e-05,
      "loss": 0.549,
      "step": 2590
    },
    {
      "epoch": 23.0,
      "learning_rate": 3.386720053386721e-05,
      "loss": 1.0528,
      "step": 2600
    },
    {
      "epoch": 23.01,
      "learning_rate": 3.3783783783783784e-05,
      "loss": 0.6398,
      "step": 2610
    },
    {
      "epoch": 23.01,
      "learning_rate": 3.370036703370037e-05,
      "loss": 0.7276,
      "step": 2620
    },
    {
      "epoch": 23.01,
      "learning_rate": 3.361695028361695e-05,
      "loss": 0.8483,
      "step": 2630
    },
    {
      "epoch": 23.01,
      "learning_rate": 3.3533533533533535e-05,
      "loss": 0.5737,
      "step": 2640
    },
    {
      "epoch": 23.01,
      "learning_rate": 3.345011678345012e-05,
      "loss": 0.588,
      "step": 2650
    },
    {
      "epoch": 23.01,
      "learning_rate": 3.33667000333667e-05,
      "loss": 0.6416,
      "step": 2660
    },
    {
      "epoch": 23.01,
      "learning_rate": 3.3283283283283285e-05,
      "loss": 0.7938,
      "step": 2670
    },
    {
      "epoch": 23.02,
      "learning_rate": 3.319986653319987e-05,
      "loss": 0.7154,
      "step": 2680
    },
    {
      "epoch": 23.02,
      "eval_accuracy": 0.6405529953917051,
      "eval_loss": 0.9371297359466553,
      "eval_runtime": 283.5798,
      "eval_samples_per_second": 0.765,
      "eval_steps_per_second": 0.194,
      "step": 2688
    },
    {
      "epoch": 24.0,
      "learning_rate": 3.311644978311645e-05,
      "loss": 0.4492,
      "step": 2690
    },
    {
      "epoch": 24.0,
      "learning_rate": 3.3033033033033035e-05,
      "loss": 0.6304,
      "step": 2700
    },
    {
      "epoch": 24.0,
      "learning_rate": 3.294961628294962e-05,
      "loss": 0.7286,
      "step": 2710
    },
    {
      "epoch": 24.0,
      "learning_rate": 3.2866199532866195e-05,
      "loss": 0.9193,
      "step": 2720
    },
    {
      "epoch": 24.01,
      "learning_rate": 3.2782782782782786e-05,
      "loss": 0.7511,
      "step": 2730
    },
    {
      "epoch": 24.01,
      "learning_rate": 3.269936603269937e-05,
      "loss": 0.7076,
      "step": 2740
    },
    {
      "epoch": 24.01,
      "learning_rate": 3.2615949282615946e-05,
      "loss": 0.7735,
      "step": 2750
    },
    {
      "epoch": 24.01,
      "learning_rate": 3.2532532532532536e-05,
      "loss": 0.9404,
      "step": 2760
    },
    {
      "epoch": 24.01,
      "learning_rate": 3.244911578244912e-05,
      "loss": 0.9553,
      "step": 2770
    },
    {
      "epoch": 24.01,
      "learning_rate": 3.2365699032365696e-05,
      "loss": 1.0067,
      "step": 2780
    },
    {
      "epoch": 24.02,
      "learning_rate": 3.2282282282282286e-05,
      "loss": 0.7712,
      "step": 2790
    },
    {
      "epoch": 24.02,
      "learning_rate": 3.219886553219887e-05,
      "loss": 0.8795,
      "step": 2800
    },
    {
      "epoch": 24.02,
      "eval_accuracy": 0.7235023041474654,
      "eval_loss": 0.6837921738624573,
      "eval_runtime": 264.0243,
      "eval_samples_per_second": 0.822,
      "eval_steps_per_second": 0.208,
      "step": 2800
    },
    {
      "epoch": 25.0,
      "learning_rate": 3.2115448782115447e-05,
      "loss": 0.6295,
      "step": 2810
    },
    {
      "epoch": 25.0,
      "learning_rate": 3.203203203203203e-05,
      "loss": 0.5106,
      "step": 2820
    },
    {
      "epoch": 25.0,
      "learning_rate": 3.194861528194862e-05,
      "loss": 0.6515,
      "step": 2830
    },
    {
      "epoch": 25.01,
      "learning_rate": 3.18651985318652e-05,
      "loss": 0.5182,
      "step": 2840
    },
    {
      "epoch": 25.01,
      "learning_rate": 3.178178178178178e-05,
      "loss": 0.9393,
      "step": 2850
    },
    {
      "epoch": 25.01,
      "learning_rate": 3.169836503169837e-05,
      "loss": 0.6204,
      "step": 2860
    },
    {
      "epoch": 25.01,
      "learning_rate": 3.161494828161495e-05,
      "loss": 0.5407,
      "step": 2870
    },
    {
      "epoch": 25.01,
      "learning_rate": 3.153153153153153e-05,
      "loss": 0.8556,
      "step": 2880
    },
    {
      "epoch": 25.01,
      "learning_rate": 3.144811478144812e-05,
      "loss": 0.6423,
      "step": 2890
    },
    {
      "epoch": 25.02,
      "learning_rate": 3.13646980313647e-05,
      "loss": 0.8295,
      "step": 2900
    },
    {
      "epoch": 25.02,
      "learning_rate": 3.128128128128128e-05,
      "loss": 0.631,
      "step": 2910
    },
    {
      "epoch": 25.02,
      "eval_accuracy": 0.6129032258064516,
      "eval_loss": 1.209276556968689,
      "eval_runtime": 285.1576,
      "eval_samples_per_second": 0.761,
      "eval_steps_per_second": 0.193,
      "step": 2912
    },
    {
      "epoch": 26.0,
      "learning_rate": 3.119786453119787e-05,
      "loss": 0.7184,
      "step": 2920
    },
    {
      "epoch": 26.0,
      "learning_rate": 3.111444778111445e-05,
      "loss": 0.6305,
      "step": 2930
    },
    {
      "epoch": 26.0,
      "learning_rate": 3.103103103103103e-05,
      "loss": 0.7949,
      "step": 2940
    },
    {
      "epoch": 26.01,
      "learning_rate": 3.0947614280947615e-05,
      "loss": 1.2476,
      "step": 2950
    },
    {
      "epoch": 26.01,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.7922,
      "step": 2960
    },
    {
      "epoch": 26.01,
      "learning_rate": 3.078078078078078e-05,
      "loss": 0.581,
      "step": 2970
    },
    {
      "epoch": 26.01,
      "learning_rate": 3.0697364030697365e-05,
      "loss": 0.6329,
      "step": 2980
    },
    {
      "epoch": 26.01,
      "learning_rate": 3.061394728061395e-05,
      "loss": 0.8019,
      "step": 2990
    },
    {
      "epoch": 26.01,
      "learning_rate": 3.053053053053053e-05,
      "loss": 0.7641,
      "step": 3000
    },
    {
      "epoch": 26.01,
      "learning_rate": 3.044711378044712e-05,
      "loss": 0.6169,
      "step": 3010
    },
    {
      "epoch": 26.02,
      "learning_rate": 3.0363697030363696e-05,
      "loss": 1.0489,
      "step": 3020
    },
    {
      "epoch": 26.02,
      "eval_accuracy": 0.5483870967741935,
      "eval_loss": 1.472022294998169,
      "eval_runtime": 262.7206,
      "eval_samples_per_second": 0.826,
      "eval_steps_per_second": 0.209,
      "step": 3024
    },
    {
      "epoch": 27.0,
      "learning_rate": 3.0280280280280282e-05,
      "loss": 0.5064,
      "step": 3030
    },
    {
      "epoch": 27.0,
      "learning_rate": 3.0196863530196866e-05,
      "loss": 0.493,
      "step": 3040
    },
    {
      "epoch": 27.0,
      "learning_rate": 3.0113446780113446e-05,
      "loss": 0.6668,
      "step": 3050
    },
    {
      "epoch": 27.01,
      "learning_rate": 3.0030030030030033e-05,
      "loss": 0.5836,
      "step": 3060
    },
    {
      "epoch": 27.01,
      "learning_rate": 2.9946613279946616e-05,
      "loss": 0.997,
      "step": 3070
    },
    {
      "epoch": 27.01,
      "learning_rate": 2.9863196529863196e-05,
      "loss": 0.8393,
      "step": 3080
    },
    {
      "epoch": 27.01,
      "learning_rate": 2.9779779779779783e-05,
      "loss": 0.609,
      "step": 3090
    },
    {
      "epoch": 27.01,
      "learning_rate": 2.9696363029696367e-05,
      "loss": 0.7111,
      "step": 3100
    },
    {
      "epoch": 27.01,
      "learning_rate": 2.9612946279612947e-05,
      "loss": 0.7491,
      "step": 3110
    },
    {
      "epoch": 27.01,
      "learning_rate": 2.952952952952953e-05,
      "loss": 0.69,
      "step": 3120
    },
    {
      "epoch": 27.02,
      "learning_rate": 2.9446112779446117e-05,
      "loss": 0.5881,
      "step": 3130
    },
    {
      "epoch": 27.02,
      "eval_accuracy": 0.631336405529954,
      "eval_loss": 1.1905186176300049,
      "eval_runtime": 281.4297,
      "eval_samples_per_second": 0.771,
      "eval_steps_per_second": 0.195,
      "step": 3136
    },
    {
      "epoch": 28.0,
      "learning_rate": 2.9362696029362697e-05,
      "loss": 1.0567,
      "step": 3140
    },
    {
      "epoch": 28.0,
      "learning_rate": 2.927927927927928e-05,
      "loss": 0.5978,
      "step": 3150
    },
    {
      "epoch": 28.0,
      "learning_rate": 2.9195862529195867e-05,
      "loss": 0.5434,
      "step": 3160
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.9112445779112447e-05,
      "loss": 0.6441,
      "step": 3170
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.902902902902903e-05,
      "loss": 0.8564,
      "step": 3180
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.8945612278945618e-05,
      "loss": 0.7492,
      "step": 3190
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.8862195528862194e-05,
      "loss": 0.7439,
      "step": 3200
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.877877877877878e-05,
      "loss": 0.731,
      "step": 3210
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.8695362028695365e-05,
      "loss": 0.8052,
      "step": 3220
    },
    {
      "epoch": 28.01,
      "learning_rate": 2.8611945278611945e-05,
      "loss": 0.8463,
      "step": 3230
    },
    {
      "epoch": 28.02,
      "learning_rate": 2.852852852852853e-05,
      "loss": 0.7919,
      "step": 3240
    },
    {
      "epoch": 28.02,
      "eval_accuracy": 0.576036866359447,
      "eval_loss": 1.1292119026184082,
      "eval_runtime": 300.8957,
      "eval_samples_per_second": 0.721,
      "eval_steps_per_second": 0.183,
      "step": 3248
    },
    {
      "epoch": 29.0,
      "learning_rate": 2.844511177844511e-05,
      "loss": 0.9881,
      "step": 3250
    },
    {
      "epoch": 29.0,
      "learning_rate": 2.8361695028361695e-05,
      "loss": 0.4171,
      "step": 3260
    },
    {
      "epoch": 29.0,
      "learning_rate": 2.8278278278278282e-05,
      "loss": 0.4612,
      "step": 3270
    },
    {
      "epoch": 29.0,
      "learning_rate": 2.8194861528194862e-05,
      "loss": 0.6761,
      "step": 3280
    },
    {
      "epoch": 29.01,
      "learning_rate": 2.8111444778111445e-05,
      "loss": 0.4762,
      "step": 3290
    },
    {
      "epoch": 29.01,
      "learning_rate": 2.8028028028028032e-05,
      "loss": 0.6633,
      "step": 3300
    },
    {
      "epoch": 29.01,
      "learning_rate": 2.794461127794461e-05,
      "loss": 0.5644,
      "step": 3310
    },
    {
      "epoch": 29.01,
      "learning_rate": 2.7861194527861196e-05,
      "loss": 0.9361,
      "step": 3320
    },
    {
      "epoch": 29.01,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.8429,
      "step": 3330
    },
    {
      "epoch": 29.01,
      "learning_rate": 2.769436102769436e-05,
      "loss": 0.6224,
      "step": 3340
    },
    {
      "epoch": 29.02,
      "learning_rate": 2.7610944277610946e-05,
      "loss": 0.8489,
      "step": 3350
    },
    {
      "epoch": 29.02,
      "learning_rate": 2.752752752752753e-05,
      "loss": 0.9158,
      "step": 3360
    },
    {
      "epoch": 29.02,
      "eval_accuracy": 0.6359447004608295,
      "eval_loss": 1.0214248895645142,
      "eval_runtime": 281.3806,
      "eval_samples_per_second": 0.771,
      "eval_steps_per_second": 0.195,
      "step": 3360
    },
    {
      "epoch": 30.0,
      "learning_rate": 2.744411077744411e-05,
      "loss": 0.6866,
      "step": 3370
    },
    {
      "epoch": 30.0,
      "learning_rate": 2.7360694027360696e-05,
      "loss": 0.8052,
      "step": 3380
    },
    {
      "epoch": 30.0,
      "learning_rate": 2.727727727727728e-05,
      "loss": 0.6658,
      "step": 3390
    },
    {
      "epoch": 30.01,
      "learning_rate": 2.719386052719386e-05,
      "loss": 0.6167,
      "step": 3400
    },
    {
      "epoch": 30.01,
      "learning_rate": 2.7110443777110443e-05,
      "loss": 0.5109,
      "step": 3410
    },
    {
      "epoch": 30.01,
      "learning_rate": 2.702702702702703e-05,
      "loss": 1.1015,
      "step": 3420
    },
    {
      "epoch": 30.01,
      "learning_rate": 2.694361027694361e-05,
      "loss": 0.7845,
      "step": 3430
    },
    {
      "epoch": 30.01,
      "learning_rate": 2.6860193526860194e-05,
      "loss": 0.6872,
      "step": 3440
    },
    {
      "epoch": 30.01,
      "learning_rate": 2.677677677677678e-05,
      "loss": 0.4606,
      "step": 3450
    },
    {
      "epoch": 30.02,
      "learning_rate": 2.669336002669336e-05,
      "loss": 0.7213,
      "step": 3460
    },
    {
      "epoch": 30.02,
      "learning_rate": 2.6609943276609944e-05,
      "loss": 0.8319,
      "step": 3470
    },
    {
      "epoch": 30.02,
      "eval_accuracy": 0.6682027649769585,
      "eval_loss": 1.2861884832382202,
      "eval_runtime": 276.1791,
      "eval_samples_per_second": 0.786,
      "eval_steps_per_second": 0.199,
      "step": 3472
    },
    {
      "epoch": 31.0,
      "learning_rate": 2.652652652652653e-05,
      "loss": 0.7171,
      "step": 3480
    },
    {
      "epoch": 31.0,
      "learning_rate": 2.6443109776443108e-05,
      "loss": 1.2605,
      "step": 3490
    },
    {
      "epoch": 31.0,
      "learning_rate": 2.6359693026359694e-05,
      "loss": 0.7113,
      "step": 3500
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.6276276276276278e-05,
      "loss": 0.4538,
      "step": 3510
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.6192859526192858e-05,
      "loss": 0.5572,
      "step": 3520
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.6109442776109445e-05,
      "loss": 0.7715,
      "step": 3530
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.6026026026026028e-05,
      "loss": 1.0429,
      "step": 3540
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.5942609275942608e-05,
      "loss": 0.7335,
      "step": 3550
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.5859192525859195e-05,
      "loss": 0.7139,
      "step": 3560
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.577577577577578e-05,
      "loss": 0.422,
      "step": 3570
    },
    {
      "epoch": 31.02,
      "learning_rate": 2.569235902569236e-05,
      "loss": 0.6775,
      "step": 3580
    },
    {
      "epoch": 31.02,
      "eval_accuracy": 0.6405529953917051,
      "eval_loss": 1.0971248149871826,
      "eval_runtime": 295.308,
      "eval_samples_per_second": 0.735,
      "eval_steps_per_second": 0.186,
      "step": 3584
    },
    {
      "epoch": 32.0,
      "learning_rate": 2.5608942275608942e-05,
      "loss": 0.3834,
      "step": 3590
    },
    {
      "epoch": 32.0,
      "learning_rate": 2.552552552552553e-05,
      "loss": 0.5938,
      "step": 3600
    },
    {
      "epoch": 32.0,
      "learning_rate": 2.544210877544211e-05,
      "loss": 0.6666,
      "step": 3610
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.5358692025358692e-05,
      "loss": 0.7156,
      "step": 3620
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.527527527527528e-05,
      "loss": 0.816,
      "step": 3630
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.519185852519186e-05,
      "loss": 0.7641,
      "step": 3640
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.5108441775108443e-05,
      "loss": 0.7681,
      "step": 3650
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.502502502502503e-05,
      "loss": 0.9274,
      "step": 3660
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.494160827494161e-05,
      "loss": 0.8763,
      "step": 3670
    },
    {
      "epoch": 32.01,
      "learning_rate": 2.4858191524858193e-05,
      "loss": 0.5928,
      "step": 3680
    },
    {
      "epoch": 32.02,
      "learning_rate": 2.4774774774774777e-05,
      "loss": 0.7191,
      "step": 3690
    },
    {
      "epoch": 32.02,
      "eval_accuracy": 0.6497695852534562,
      "eval_loss": 1.026401400566101,
      "eval_runtime": 280.0325,
      "eval_samples_per_second": 0.775,
      "eval_steps_per_second": 0.196,
      "step": 3696
    },
    {
      "epoch": 33.0,
      "learning_rate": 2.4691358024691357e-05,
      "loss": 0.9832,
      "step": 3700
    },
    {
      "epoch": 33.0,
      "learning_rate": 2.4607941274607943e-05,
      "loss": 0.7523,
      "step": 3710
    },
    {
      "epoch": 33.0,
      "learning_rate": 2.4524524524524527e-05,
      "loss": 0.6381,
      "step": 3720
    },
    {
      "epoch": 33.01,
      "learning_rate": 2.4441107774441107e-05,
      "loss": 0.7641,
      "step": 3730
    },
    {
      "epoch": 33.01,
      "learning_rate": 2.4357691024357694e-05,
      "loss": 0.5699,
      "step": 3740
    },
    {
      "epoch": 33.01,
      "learning_rate": 2.4274274274274274e-05,
      "loss": 0.5342,
      "step": 3750
    },
    {
      "epoch": 33.01,
      "learning_rate": 2.4190857524190857e-05,
      "loss": 0.4933,
      "step": 3760
    },
    {
      "epoch": 33.01,
      "learning_rate": 2.4107440774107444e-05,
      "loss": 0.7027,
      "step": 3770
    },
    {
      "epoch": 33.01,
      "learning_rate": 2.4024024024024024e-05,
      "loss": 0.5812,
      "step": 3780
    },
    {
      "epoch": 33.01,
      "learning_rate": 2.3940607273940608e-05,
      "loss": 0.5228,
      "step": 3790
    },
    {
      "epoch": 33.02,
      "learning_rate": 2.385719052385719e-05,
      "loss": 0.7662,
      "step": 3800
    },
    {
      "epoch": 33.02,
      "eval_accuracy": 0.6405529953917051,
      "eval_loss": 1.0589454174041748,
      "eval_runtime": 290.7017,
      "eval_samples_per_second": 0.746,
      "eval_steps_per_second": 0.189,
      "step": 3808
    },
    {
      "epoch": 34.0,
      "learning_rate": 2.3773773773773775e-05,
      "loss": 0.5944,
      "step": 3810
    },
    {
      "epoch": 34.0,
      "learning_rate": 2.3690357023690358e-05,
      "loss": 0.588,
      "step": 3820
    },
    {
      "epoch": 34.0,
      "learning_rate": 2.360694027360694e-05,
      "loss": 0.6972,
      "step": 3830
    },
    {
      "epoch": 34.0,
      "learning_rate": 2.3523523523523525e-05,
      "loss": 0.544,
      "step": 3840
    },
    {
      "epoch": 34.01,
      "learning_rate": 2.344010677344011e-05,
      "loss": 0.6615,
      "step": 3850
    },
    {
      "epoch": 34.01,
      "learning_rate": 2.3356690023356692e-05,
      "loss": 0.5002,
      "step": 3860
    },
    {
      "epoch": 34.01,
      "learning_rate": 2.3273273273273275e-05,
      "loss": 0.5654,
      "step": 3870
    },
    {
      "epoch": 34.01,
      "learning_rate": 2.3189856523189855e-05,
      "loss": 0.6728,
      "step": 3880
    },
    {
      "epoch": 34.01,
      "learning_rate": 2.3106439773106442e-05,
      "loss": 0.5296,
      "step": 3890
    },
    {
      "epoch": 34.01,
      "learning_rate": 2.3023023023023026e-05,
      "loss": 0.4234,
      "step": 3900
    },
    {
      "epoch": 34.02,
      "learning_rate": 2.2939606272939606e-05,
      "loss": 0.5872,
      "step": 3910
    },
    {
      "epoch": 34.02,
      "learning_rate": 2.2856189522856192e-05,
      "loss": 0.7313,
      "step": 3920
    },
    {
      "epoch": 34.02,
      "eval_accuracy": 0.5622119815668203,
      "eval_loss": 1.5075870752334595,
      "eval_runtime": 286.0606,
      "eval_samples_per_second": 0.759,
      "eval_steps_per_second": 0.192,
      "step": 3920
    },
    {
      "epoch": 35.0,
      "learning_rate": 2.2772772772772773e-05,
      "loss": 0.5481,
      "step": 3930
    },
    {
      "epoch": 35.0,
      "learning_rate": 2.2689356022689356e-05,
      "loss": 0.7286,
      "step": 3940
    },
    {
      "epoch": 35.0,
      "learning_rate": 2.2605939272605943e-05,
      "loss": 0.795,
      "step": 3950
    },
    {
      "epoch": 35.01,
      "learning_rate": 2.2522522522522523e-05,
      "loss": 0.7738,
      "step": 3960
    },
    {
      "epoch": 35.01,
      "learning_rate": 2.2439105772439106e-05,
      "loss": 0.8864,
      "step": 3970
    },
    {
      "epoch": 35.01,
      "learning_rate": 2.235568902235569e-05,
      "loss": 0.3864,
      "step": 3980
    },
    {
      "epoch": 35.01,
      "learning_rate": 2.2272272272272273e-05,
      "loss": 0.6556,
      "step": 3990
    },
    {
      "epoch": 35.01,
      "learning_rate": 2.2188855522188857e-05,
      "loss": 0.9211,
      "step": 4000
    },
    {
      "epoch": 35.01,
      "learning_rate": 2.210543877210544e-05,
      "loss": 0.6319,
      "step": 4010
    },
    {
      "epoch": 35.02,
      "learning_rate": 2.2022022022022024e-05,
      "loss": 0.6768,
      "step": 4020
    },
    {
      "epoch": 35.02,
      "learning_rate": 2.1938605271938607e-05,
      "loss": 0.7539,
      "step": 4030
    },
    {
      "epoch": 35.02,
      "eval_accuracy": 0.5898617511520737,
      "eval_loss": 1.2265080213546753,
      "eval_runtime": 266.6298,
      "eval_samples_per_second": 0.814,
      "eval_steps_per_second": 0.206,
      "step": 4032
    },
    {
      "epoch": 36.0,
      "learning_rate": 2.1855188521855187e-05,
      "loss": 0.6197,
      "step": 4040
    },
    {
      "epoch": 36.0,
      "learning_rate": 2.1771771771771774e-05,
      "loss": 0.5972,
      "step": 4050
    },
    {
      "epoch": 36.0,
      "learning_rate": 2.1688355021688357e-05,
      "loss": 0.9935,
      "step": 4060
    },
    {
      "epoch": 36.01,
      "learning_rate": 2.1604938271604937e-05,
      "loss": 0.6247,
      "step": 4070
    },
    {
      "epoch": 36.01,
      "learning_rate": 2.1521521521521524e-05,
      "loss": 0.6728,
      "step": 4080
    },
    {
      "epoch": 36.01,
      "learning_rate": 2.1438104771438104e-05,
      "loss": 0.8943,
      "step": 4090
    },
    {
      "epoch": 36.01,
      "learning_rate": 2.1354688021354688e-05,
      "loss": 0.6704,
      "step": 4100
    },
    {
      "epoch": 36.01,
      "learning_rate": 2.1271271271271275e-05,
      "loss": 0.8664,
      "step": 4110
    },
    {
      "epoch": 36.01,
      "learning_rate": 2.1187854521187855e-05,
      "loss": 0.8486,
      "step": 4120
    },
    {
      "epoch": 36.01,
      "learning_rate": 2.1104437771104438e-05,
      "loss": 0.5425,
      "step": 4130
    },
    {
      "epoch": 36.02,
      "learning_rate": 2.102102102102102e-05,
      "loss": 0.571,
      "step": 4140
    },
    {
      "epoch": 36.02,
      "eval_accuracy": 0.6267281105990783,
      "eval_loss": 1.1598420143127441,
      "eval_runtime": 288.0962,
      "eval_samples_per_second": 0.753,
      "eval_steps_per_second": 0.191,
      "step": 4144
    },
    {
      "epoch": 37.0,
      "learning_rate": 2.0937604270937605e-05,
      "loss": 0.2915,
      "step": 4150
    },
    {
      "epoch": 37.0,
      "learning_rate": 2.085418752085419e-05,
      "loss": 0.5187,
      "step": 4160
    },
    {
      "epoch": 37.0,
      "learning_rate": 2.0770770770770772e-05,
      "loss": 0.4894,
      "step": 4170
    },
    {
      "epoch": 37.01,
      "learning_rate": 2.0687354020687355e-05,
      "loss": 0.522,
      "step": 4180
    },
    {
      "epoch": 37.01,
      "learning_rate": 2.060393727060394e-05,
      "loss": 0.3428,
      "step": 4190
    },
    {
      "epoch": 37.01,
      "learning_rate": 2.0520520520520522e-05,
      "loss": 0.4749,
      "step": 4200
    },
    {
      "epoch": 37.01,
      "learning_rate": 2.0437103770437106e-05,
      "loss": 0.9899,
      "step": 4210
    },
    {
      "epoch": 37.01,
      "learning_rate": 2.0353687020353686e-05,
      "loss": 0.7875,
      "step": 4220
    },
    {
      "epoch": 37.01,
      "learning_rate": 2.0270270270270273e-05,
      "loss": 0.4509,
      "step": 4230
    },
    {
      "epoch": 37.01,
      "learning_rate": 2.0186853520186856e-05,
      "loss": 0.5523,
      "step": 4240
    },
    {
      "epoch": 37.02,
      "learning_rate": 2.0103436770103436e-05,
      "loss": 0.3404,
      "step": 4250
    },
    {
      "epoch": 37.02,
      "eval_accuracy": 0.6359447004608295,
      "eval_loss": 1.0306791067123413,
      "eval_runtime": 260.5195,
      "eval_samples_per_second": 0.833,
      "eval_steps_per_second": 0.211,
      "step": 4256
    },
    {
      "epoch": 38.0,
      "learning_rate": 2.0020020020020023e-05,
      "loss": 0.6409,
      "step": 4260
    },
    {
      "epoch": 38.0,
      "learning_rate": 1.9936603269936603e-05,
      "loss": 0.9606,
      "step": 4270
    },
    {
      "epoch": 38.0,
      "learning_rate": 1.9853186519853186e-05,
      "loss": 0.5896,
      "step": 4280
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.9769769769769773e-05,
      "loss": 1.0073,
      "step": 4290
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.9686353019686353e-05,
      "loss": 0.7469,
      "step": 4300
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.9602936269602937e-05,
      "loss": 0.9638,
      "step": 4310
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.951951951951952e-05,
      "loss": 0.2893,
      "step": 4320
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.9436102769436104e-05,
      "loss": 0.9141,
      "step": 4330
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.9352686019352687e-05,
      "loss": 0.4846,
      "step": 4340
    },
    {
      "epoch": 38.01,
      "learning_rate": 1.9269269269269267e-05,
      "loss": 0.2622,
      "step": 4350
    },
    {
      "epoch": 38.02,
      "learning_rate": 1.9185852519185854e-05,
      "loss": 0.5553,
      "step": 4360
    },
    {
      "epoch": 38.02,
      "eval_accuracy": 0.7235023041474654,
      "eval_loss": 0.8180494904518127,
      "eval_runtime": 302.1979,
      "eval_samples_per_second": 0.718,
      "eval_steps_per_second": 0.182,
      "step": 4368
    },
    {
      "epoch": 39.0,
      "learning_rate": 1.9102435769102438e-05,
      "loss": 0.7832,
      "step": 4370
    },
    {
      "epoch": 39.0,
      "learning_rate": 1.9019019019019018e-05,
      "loss": 0.7238,
      "step": 4380
    },
    {
      "epoch": 39.0,
      "learning_rate": 1.8935602268935604e-05,
      "loss": 0.2628,
      "step": 4390
    },
    {
      "epoch": 39.0,
      "learning_rate": 1.8852185518852188e-05,
      "loss": 0.7441,
      "step": 4400
    },
    {
      "epoch": 39.01,
      "learning_rate": 1.8768768768768768e-05,
      "loss": 0.4422,
      "step": 4410
    },
    {
      "epoch": 39.01,
      "learning_rate": 1.8685352018685355e-05,
      "loss": 0.5266,
      "step": 4420
    },
    {
      "epoch": 39.01,
      "learning_rate": 1.8601935268601935e-05,
      "loss": 0.5469,
      "step": 4430
    },
    {
      "epoch": 39.01,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.9752,
      "step": 4440
    },
    {
      "epoch": 39.01,
      "learning_rate": 1.8435101768435105e-05,
      "loss": 1.0334,
      "step": 4450
    },
    {
      "epoch": 39.01,
      "learning_rate": 1.8351685018351685e-05,
      "loss": 0.604,
      "step": 4460
    },
    {
      "epoch": 39.02,
      "learning_rate": 1.826826826826827e-05,
      "loss": 0.7227,
      "step": 4470
    },
    {
      "epoch": 39.02,
      "learning_rate": 1.8184851518184852e-05,
      "loss": 0.8499,
      "step": 4480
    },
    {
      "epoch": 39.02,
      "eval_accuracy": 0.6497695852534562,
      "eval_loss": 1.0074414014816284,
      "eval_runtime": 296.0859,
      "eval_samples_per_second": 0.733,
      "eval_steps_per_second": 0.186,
      "step": 4480
    },
    {
      "epoch": 40.0,
      "learning_rate": 1.8101434768101436e-05,
      "loss": 0.3984,
      "step": 4490
    },
    {
      "epoch": 40.0,
      "learning_rate": 1.801801801801802e-05,
      "loss": 0.5711,
      "step": 4500
    },
    {
      "epoch": 40.0,
      "learning_rate": 1.7934601267934602e-05,
      "loss": 0.5289,
      "step": 4510
    },
    {
      "epoch": 40.01,
      "learning_rate": 1.7851184517851186e-05,
      "loss": 0.7913,
      "step": 4520
    },
    {
      "epoch": 40.01,
      "learning_rate": 1.776776776776777e-05,
      "loss": 0.6343,
      "step": 4530
    },
    {
      "epoch": 40.01,
      "learning_rate": 1.7684351017684353e-05,
      "loss": 0.5266,
      "step": 4540
    },
    {
      "epoch": 40.01,
      "learning_rate": 1.7600934267600936e-05,
      "loss": 0.5792,
      "step": 4550
    },
    {
      "epoch": 40.01,
      "learning_rate": 1.7517517517517516e-05,
      "loss": 0.8629,
      "step": 4560
    },
    {
      "epoch": 40.01,
      "learning_rate": 1.7434100767434103e-05,
      "loss": 0.7761,
      "step": 4570
    },
    {
      "epoch": 40.02,
      "learning_rate": 1.7350684017350687e-05,
      "loss": 0.7784,
      "step": 4580
    },
    {
      "epoch": 40.02,
      "learning_rate": 1.7267267267267267e-05,
      "loss": 0.5036,
      "step": 4590
    },
    {
      "epoch": 40.02,
      "eval_accuracy": 0.631336405529954,
      "eval_loss": 1.1159977912902832,
      "eval_runtime": 266.794,
      "eval_samples_per_second": 0.813,
      "eval_steps_per_second": 0.206,
      "step": 4592
    },
    {
      "epoch": 41.0,
      "learning_rate": 1.7183850517183853e-05,
      "loss": 0.8591,
      "step": 4600
    },
    {
      "epoch": 41.0,
      "learning_rate": 1.7100433767100434e-05,
      "loss": 0.4294,
      "step": 4610
    },
    {
      "epoch": 41.0,
      "learning_rate": 1.7017017017017017e-05,
      "loss": 0.5119,
      "step": 4620
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.6933600266933604e-05,
      "loss": 0.3545,
      "step": 4630
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.6850183516850184e-05,
      "loss": 0.7615,
      "step": 4640
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.6766766766766767e-05,
      "loss": 0.1783,
      "step": 4650
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.668335001668335e-05,
      "loss": 0.6125,
      "step": 4660
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.6599933266599934e-05,
      "loss": 1.0295,
      "step": 4670
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.6516516516516518e-05,
      "loss": 0.6956,
      "step": 4680
    },
    {
      "epoch": 41.01,
      "learning_rate": 1.6433099766433098e-05,
      "loss": 0.6383,
      "step": 4690
    },
    {
      "epoch": 41.02,
      "learning_rate": 1.6349683016349685e-05,
      "loss": 0.814,
      "step": 4700
    },
    {
      "epoch": 41.02,
      "eval_accuracy": 0.695852534562212,
      "eval_loss": 0.903153657913208,
      "eval_runtime": 298.2312,
      "eval_samples_per_second": 0.728,
      "eval_steps_per_second": 0.184,
      "step": 4704
    },
    {
      "epoch": 42.0,
      "learning_rate": 1.6266266266266268e-05,
      "loss": 0.5097,
      "step": 4710
    },
    {
      "epoch": 42.0,
      "learning_rate": 1.6182849516182848e-05,
      "loss": 0.4471,
      "step": 4720
    },
    {
      "epoch": 42.0,
      "learning_rate": 1.6099432766099435e-05,
      "loss": 0.4364,
      "step": 4730
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.6016016016016015e-05,
      "loss": 0.5641,
      "step": 4740
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.59325992659326e-05,
      "loss": 0.8213,
      "step": 4750
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.5849182515849185e-05,
      "loss": 0.7006,
      "step": 4760
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.5765765765765765e-05,
      "loss": 0.5187,
      "step": 4770
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.568234901568235e-05,
      "loss": 0.3012,
      "step": 4780
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.5598932265598936e-05,
      "loss": 0.7188,
      "step": 4790
    },
    {
      "epoch": 42.01,
      "learning_rate": 1.5515515515515516e-05,
      "loss": 0.5198,
      "step": 4800
    },
    {
      "epoch": 42.02,
      "learning_rate": 1.54320987654321e-05,
      "loss": 0.7293,
      "step": 4810
    },
    {
      "epoch": 42.02,
      "eval_accuracy": 0.728110599078341,
      "eval_loss": 0.9331218004226685,
      "eval_runtime": 258.6829,
      "eval_samples_per_second": 0.839,
      "eval_steps_per_second": 0.213,
      "step": 4816
    },
    {
      "epoch": 43.0,
      "learning_rate": 1.5348682015348683e-05,
      "loss": 0.4944,
      "step": 4820
    },
    {
      "epoch": 43.0,
      "learning_rate": 1.5265265265265266e-05,
      "loss": 0.4447,
      "step": 4830
    },
    {
      "epoch": 43.0,
      "learning_rate": 1.5181848515181848e-05,
      "loss": 0.3132,
      "step": 4840
    },
    {
      "epoch": 43.01,
      "learning_rate": 1.5098431765098433e-05,
      "loss": 0.5882,
      "step": 4850
    },
    {
      "epoch": 43.01,
      "learning_rate": 1.5015015015015016e-05,
      "loss": 0.7151,
      "step": 4860
    },
    {
      "epoch": 43.01,
      "learning_rate": 1.4931598264931598e-05,
      "loss": 0.7703,
      "step": 4870
    },
    {
      "epoch": 43.01,
      "learning_rate": 1.4848181514848183e-05,
      "loss": 0.8505,
      "step": 4880
    },
    {
      "epoch": 43.01,
      "learning_rate": 1.4764764764764765e-05,
      "loss": 0.3957,
      "step": 4890
    },
    {
      "epoch": 43.01,
      "learning_rate": 1.4681348014681348e-05,
      "loss": 0.5204,
      "step": 4900
    },
    {
      "epoch": 43.01,
      "learning_rate": 1.4597931264597934e-05,
      "loss": 0.4383,
      "step": 4910
    },
    {
      "epoch": 43.02,
      "learning_rate": 1.4514514514514515e-05,
      "loss": 0.4402,
      "step": 4920
    },
    {
      "epoch": 43.02,
      "eval_accuracy": 0.5668202764976958,
      "eval_loss": 1.4190495014190674,
      "eval_runtime": 275.5299,
      "eval_samples_per_second": 0.788,
      "eval_steps_per_second": 0.2,
      "step": 4928
    },
    {
      "epoch": 44.0,
      "learning_rate": 1.4431097764431097e-05,
      "loss": 0.2849,
      "step": 4930
    },
    {
      "epoch": 44.0,
      "learning_rate": 1.4347681014347682e-05,
      "loss": 0.5245,
      "step": 4940
    },
    {
      "epoch": 44.0,
      "learning_rate": 1.4264264264264266e-05,
      "loss": 0.4884,
      "step": 4950
    },
    {
      "epoch": 44.0,
      "learning_rate": 1.4180847514180847e-05,
      "loss": 0.3824,
      "step": 4960
    },
    {
      "epoch": 44.01,
      "learning_rate": 1.4097430764097431e-05,
      "loss": 0.5175,
      "step": 4970
    },
    {
      "epoch": 44.01,
      "learning_rate": 1.4014014014014016e-05,
      "loss": 0.5526,
      "step": 4980
    },
    {
      "epoch": 44.01,
      "learning_rate": 1.3930597263930598e-05,
      "loss": 0.4134,
      "step": 4990
    },
    {
      "epoch": 44.01,
      "learning_rate": 1.384718051384718e-05,
      "loss": 0.4121,
      "step": 5000
    },
    {
      "epoch": 44.01,
      "learning_rate": 1.3763763763763765e-05,
      "loss": 0.7849,
      "step": 5010
    },
    {
      "epoch": 44.01,
      "learning_rate": 1.3680347013680348e-05,
      "loss": 0.7587,
      "step": 5020
    },
    {
      "epoch": 44.02,
      "learning_rate": 1.359693026359693e-05,
      "loss": 0.4583,
      "step": 5030
    },
    {
      "epoch": 44.02,
      "learning_rate": 1.3513513513513515e-05,
      "loss": 0.4625,
      "step": 5040
    },
    {
      "epoch": 44.02,
      "eval_accuracy": 0.7004608294930875,
      "eval_loss": 1.0268219709396362,
      "eval_runtime": 299.0881,
      "eval_samples_per_second": 0.726,
      "eval_steps_per_second": 0.184,
      "step": 5040
    },
    {
      "epoch": 45.0,
      "learning_rate": 1.3430096763430097e-05,
      "loss": 0.4425,
      "step": 5050
    },
    {
      "epoch": 45.0,
      "learning_rate": 1.334668001334668e-05,
      "loss": 0.6364,
      "step": 5060
    },
    {
      "epoch": 45.0,
      "learning_rate": 1.3263263263263265e-05,
      "loss": 0.386,
      "step": 5070
    },
    {
      "epoch": 45.01,
      "learning_rate": 1.3179846513179847e-05,
      "loss": 0.6462,
      "step": 5080
    },
    {
      "epoch": 45.01,
      "learning_rate": 1.3096429763096429e-05,
      "loss": 0.8925,
      "step": 5090
    },
    {
      "epoch": 45.01,
      "learning_rate": 1.3013013013013014e-05,
      "loss": 0.6358,
      "step": 5100
    },
    {
      "epoch": 45.01,
      "learning_rate": 1.2929596262929598e-05,
      "loss": 0.8605,
      "step": 5110
    },
    {
      "epoch": 45.01,
      "learning_rate": 1.284617951284618e-05,
      "loss": 0.9282,
      "step": 5120
    },
    {
      "epoch": 45.01,
      "learning_rate": 1.2762762762762764e-05,
      "loss": 0.7836,
      "step": 5130
    },
    {
      "epoch": 45.02,
      "learning_rate": 1.2679346012679346e-05,
      "loss": 0.4875,
      "step": 5140
    },
    {
      "epoch": 45.02,
      "learning_rate": 1.259592926259593e-05,
      "loss": 0.2266,
      "step": 5150
    },
    {
      "epoch": 45.02,
      "eval_accuracy": 0.6405529953917051,
      "eval_loss": 1.2808195352554321,
      "eval_runtime": 257.9463,
      "eval_samples_per_second": 0.841,
      "eval_steps_per_second": 0.213,
      "step": 5152
    },
    {
      "epoch": 46.0,
      "learning_rate": 1.2512512512512515e-05,
      "loss": 0.6837,
      "step": 5160
    },
    {
      "epoch": 46.0,
      "learning_rate": 1.2429095762429097e-05,
      "loss": 0.3609,
      "step": 5170
    },
    {
      "epoch": 46.0,
      "learning_rate": 1.2345679012345678e-05,
      "loss": 0.7921,
      "step": 5180
    },
    {
      "epoch": 46.01,
      "learning_rate": 1.2262262262262263e-05,
      "loss": 1.2116,
      "step": 5190
    },
    {
      "epoch": 46.01,
      "learning_rate": 1.2178845512178847e-05,
      "loss": 0.6612,
      "step": 5200
    },
    {
      "epoch": 46.01,
      "learning_rate": 1.2095428762095429e-05,
      "loss": 0.2317,
      "step": 5210
    },
    {
      "epoch": 46.01,
      "learning_rate": 1.2012012012012012e-05,
      "loss": 0.4213,
      "step": 5220
    },
    {
      "epoch": 46.01,
      "learning_rate": 1.1928595261928596e-05,
      "loss": 0.8834,
      "step": 5230
    },
    {
      "epoch": 46.01,
      "learning_rate": 1.1845178511845179e-05,
      "loss": 0.8265,
      "step": 5240
    },
    {
      "epoch": 46.01,
      "learning_rate": 1.1761761761761762e-05,
      "loss": 0.1315,
      "step": 5250
    },
    {
      "epoch": 46.02,
      "learning_rate": 1.1678345011678346e-05,
      "loss": 0.7424,
      "step": 5260
    },
    {
      "epoch": 46.02,
      "eval_accuracy": 0.6497695852534562,
      "eval_loss": 1.1821485757827759,
      "eval_runtime": 277.4384,
      "eval_samples_per_second": 0.782,
      "eval_steps_per_second": 0.198,
      "step": 5264
    },
    {
      "epoch": 47.0,
      "learning_rate": 1.1594928261594928e-05,
      "loss": 0.7674,
      "step": 5270
    },
    {
      "epoch": 47.0,
      "learning_rate": 1.1511511511511513e-05,
      "loss": 0.8093,
      "step": 5280
    },
    {
      "epoch": 47.0,
      "learning_rate": 1.1428094761428096e-05,
      "loss": 0.3749,
      "step": 5290
    },
    {
      "epoch": 47.01,
      "learning_rate": 1.1344678011344678e-05,
      "loss": 0.572,
      "step": 5300
    },
    {
      "epoch": 47.01,
      "learning_rate": 1.1261261261261261e-05,
      "loss": 0.4389,
      "step": 5310
    },
    {
      "epoch": 47.01,
      "learning_rate": 1.1177844511177845e-05,
      "loss": 0.421,
      "step": 5320
    },
    {
      "epoch": 47.01,
      "learning_rate": 1.1094427761094428e-05,
      "loss": 0.7683,
      "step": 5330
    },
    {
      "epoch": 47.01,
      "learning_rate": 1.1011011011011012e-05,
      "loss": 0.333,
      "step": 5340
    },
    {
      "epoch": 47.01,
      "learning_rate": 1.0927594260927594e-05,
      "loss": 0.8005,
      "step": 5350
    },
    {
      "epoch": 47.01,
      "learning_rate": 1.0844177510844179e-05,
      "loss": 0.5426,
      "step": 5360
    },
    {
      "epoch": 47.02,
      "learning_rate": 1.0760760760760762e-05,
      "loss": 0.4852,
      "step": 5370
    },
    {
      "epoch": 47.02,
      "eval_accuracy": 0.6589861751152074,
      "eval_loss": 1.2433668375015259,
      "eval_runtime": 303.2594,
      "eval_samples_per_second": 0.716,
      "eval_steps_per_second": 0.181,
      "step": 5376
    },
    {
      "epoch": 48.0,
      "learning_rate": 1.0677344010677344e-05,
      "loss": 0.9861,
      "step": 5380
    },
    {
      "epoch": 48.0,
      "learning_rate": 1.0593927260593927e-05,
      "loss": 0.4832,
      "step": 5390
    },
    {
      "epoch": 48.0,
      "learning_rate": 1.051051051051051e-05,
      "loss": 0.3998,
      "step": 5400
    },
    {
      "epoch": 48.01,
      "learning_rate": 1.0427093760427094e-05,
      "loss": 0.599,
      "step": 5410
    },
    {
      "epoch": 48.01,
      "learning_rate": 1.0343677010343678e-05,
      "loss": 0.6899,
      "step": 5420
    },
    {
      "epoch": 48.01,
      "learning_rate": 1.0260260260260261e-05,
      "loss": 0.618,
      "step": 5430
    },
    {
      "epoch": 48.01,
      "learning_rate": 1.0176843510176843e-05,
      "loss": 0.5096,
      "step": 5440
    },
    {
      "epoch": 48.01,
      "learning_rate": 1.0093426760093428e-05,
      "loss": 0.379,
      "step": 5450
    },
    {
      "epoch": 48.01,
      "learning_rate": 1.0010010010010011e-05,
      "loss": 0.5432,
      "step": 5460
    },
    {
      "epoch": 48.01,
      "learning_rate": 9.926593259926593e-06,
      "loss": 0.5489,
      "step": 5470
    },
    {
      "epoch": 48.02,
      "learning_rate": 9.843176509843177e-06,
      "loss": 0.523,
      "step": 5480
    },
    {
      "epoch": 48.02,
      "eval_accuracy": 0.6267281105990783,
      "eval_loss": 1.2123322486877441,
      "eval_runtime": 258.3979,
      "eval_samples_per_second": 0.84,
      "eval_steps_per_second": 0.213,
      "step": 5488
    },
    {
      "epoch": 49.0,
      "learning_rate": 9.75975975975976e-06,
      "loss": 0.2441,
      "step": 5490
    },
    {
      "epoch": 49.0,
      "learning_rate": 9.676343009676344e-06,
      "loss": 0.6196,
      "step": 5500
    },
    {
      "epoch": 49.0,
      "learning_rate": 9.592926259592927e-06,
      "loss": 0.3021,
      "step": 5510
    },
    {
      "epoch": 49.0,
      "learning_rate": 9.509509509509509e-06,
      "loss": 0.73,
      "step": 5520
    },
    {
      "epoch": 49.01,
      "learning_rate": 9.426092759426094e-06,
      "loss": 0.5437,
      "step": 5530
    },
    {
      "epoch": 49.01,
      "learning_rate": 9.342676009342677e-06,
      "loss": 0.3555,
      "step": 5540
    },
    {
      "epoch": 49.01,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.5324,
      "step": 5550
    },
    {
      "epoch": 49.01,
      "learning_rate": 9.175842509175843e-06,
      "loss": 0.3838,
      "step": 5560
    },
    {
      "epoch": 49.01,
      "learning_rate": 9.092425759092426e-06,
      "loss": 0.5131,
      "step": 5570
    },
    {
      "epoch": 49.01,
      "learning_rate": 9.00900900900901e-06,
      "loss": 0.5087,
      "step": 5580
    },
    {
      "epoch": 49.02,
      "learning_rate": 8.925592258925593e-06,
      "loss": 0.5122,
      "step": 5590
    },
    {
      "epoch": 49.02,
      "learning_rate": 8.842175508842176e-06,
      "loss": 0.8344,
      "step": 5600
    },
    {
      "epoch": 49.02,
      "eval_accuracy": 0.663594470046083,
      "eval_loss": 1.1888865232467651,
      "eval_runtime": 295.3851,
      "eval_samples_per_second": 0.735,
      "eval_steps_per_second": 0.186,
      "step": 5600
    },
    {
      "epoch": 50.0,
      "learning_rate": 8.758758758758758e-06,
      "loss": 0.6569,
      "step": 5610
    },
    {
      "epoch": 50.0,
      "learning_rate": 8.675342008675343e-06,
      "loss": 0.4491,
      "step": 5620
    },
    {
      "epoch": 50.0,
      "learning_rate": 8.591925258591927e-06,
      "loss": 0.6111,
      "step": 5630
    },
    {
      "epoch": 50.01,
      "learning_rate": 8.508508508508508e-06,
      "loss": 0.2911,
      "step": 5640
    },
    {
      "epoch": 50.01,
      "learning_rate": 8.425091758425092e-06,
      "loss": 0.6463,
      "step": 5650
    },
    {
      "epoch": 50.01,
      "learning_rate": 8.341675008341675e-06,
      "loss": 0.3108,
      "step": 5660
    },
    {
      "epoch": 50.01,
      "learning_rate": 8.258258258258259e-06,
      "loss": 0.4594,
      "step": 5670
    },
    {
      "epoch": 50.01,
      "learning_rate": 8.174841508174842e-06,
      "loss": 0.4389,
      "step": 5680
    },
    {
      "epoch": 50.01,
      "learning_rate": 8.091424758091424e-06,
      "loss": 0.4616,
      "step": 5690
    },
    {
      "epoch": 50.02,
      "learning_rate": 8.008008008008007e-06,
      "loss": 0.2669,
      "step": 5700
    },
    {
      "epoch": 50.02,
      "learning_rate": 7.924591257924593e-06,
      "loss": 0.6648,
      "step": 5710
    },
    {
      "epoch": 50.02,
      "eval_accuracy": 0.6405529953917051,
      "eval_loss": 1.2328165769577026,
      "eval_runtime": 294.1867,
      "eval_samples_per_second": 0.738,
      "eval_steps_per_second": 0.187,
      "step": 5712
    },
    {
      "epoch": 51.0,
      "learning_rate": 7.841174507841174e-06,
      "loss": 0.5881,
      "step": 5720
    },
    {
      "epoch": 51.0,
      "learning_rate": 7.757757757757758e-06,
      "loss": 0.658,
      "step": 5730
    },
    {
      "epoch": 51.0,
      "learning_rate": 7.674341007674341e-06,
      "loss": 0.473,
      "step": 5740
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.590924257590924e-06,
      "loss": 0.561,
      "step": 5750
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.507507507507508e-06,
      "loss": 0.3124,
      "step": 5760
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.424090757424092e-06,
      "loss": 0.6171,
      "step": 5770
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.340674007340674e-06,
      "loss": 0.2376,
      "step": 5780
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.257257257257258e-06,
      "loss": 0.5859,
      "step": 5790
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.173840507173841e-06,
      "loss": 0.3655,
      "step": 5800
    },
    {
      "epoch": 51.01,
      "learning_rate": 7.090423757090424e-06,
      "loss": 0.5225,
      "step": 5810
    },
    {
      "epoch": 51.02,
      "learning_rate": 7.007007007007008e-06,
      "loss": 0.6929,
      "step": 5820
    },
    {
      "epoch": 51.02,
      "eval_accuracy": 0.6129032258064516,
      "eval_loss": 1.3269143104553223,
      "eval_runtime": 258.5914,
      "eval_samples_per_second": 0.839,
      "eval_steps_per_second": 0.213,
      "step": 5824
    },
    {
      "epoch": 52.0,
      "learning_rate": 6.92359025692359e-06,
      "loss": 0.2531,
      "step": 5830
    },
    {
      "epoch": 52.0,
      "learning_rate": 6.840173506840174e-06,
      "loss": 0.6214,
      "step": 5840
    },
    {
      "epoch": 52.0,
      "learning_rate": 6.7567567567567575e-06,
      "loss": 0.6634,
      "step": 5850
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.67334000667334e-06,
      "loss": 0.5057,
      "step": 5860
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.589923256589924e-06,
      "loss": 0.4402,
      "step": 5870
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.506506506506507e-06,
      "loss": 0.9344,
      "step": 5880
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.42308975642309e-06,
      "loss": 0.1245,
      "step": 5890
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.339673006339673e-06,
      "loss": 0.568,
      "step": 5900
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.256256256256257e-06,
      "loss": 0.4524,
      "step": 5910
    },
    {
      "epoch": 52.01,
      "learning_rate": 6.172839506172839e-06,
      "loss": 0.4983,
      "step": 5920
    },
    {
      "epoch": 52.02,
      "learning_rate": 6.0894227560894234e-06,
      "loss": 0.4253,
      "step": 5930
    },
    {
      "epoch": 52.02,
      "eval_accuracy": 0.6820276497695853,
      "eval_loss": 1.1884889602661133,
      "eval_runtime": 278.2143,
      "eval_samples_per_second": 0.78,
      "eval_steps_per_second": 0.198,
      "step": 5936
    },
    {
      "epoch": 53.0,
      "learning_rate": 6.006006006006006e-06,
      "loss": 0.5096,
      "step": 5940
    },
    {
      "epoch": 53.0,
      "learning_rate": 5.9225892559225895e-06,
      "loss": 0.6105,
      "step": 5950
    },
    {
      "epoch": 53.0,
      "learning_rate": 5.839172505839173e-06,
      "loss": 1.0454,
      "step": 5960
    },
    {
      "epoch": 53.01,
      "learning_rate": 5.755755755755756e-06,
      "loss": 0.4144,
      "step": 5970
    },
    {
      "epoch": 53.01,
      "learning_rate": 5.672339005672339e-06,
      "loss": 0.8673,
      "step": 5980
    },
    {
      "epoch": 53.01,
      "learning_rate": 5.5889222555889224e-06,
      "loss": 0.4563,
      "step": 5990
    },
    {
      "epoch": 53.01,
      "learning_rate": 5.505505505505506e-06,
      "loss": 0.5437,
      "step": 6000
    },
    {
      "epoch": 53.01,
      "learning_rate": 5.422088755422089e-06,
      "loss": 0.4049,
      "step": 6010
    },
    {
      "epoch": 53.01,
      "learning_rate": 5.338672005338672e-06,
      "loss": 0.5225,
      "step": 6020
    },
    {
      "epoch": 53.01,
      "learning_rate": 5.255255255255255e-06,
      "loss": 0.4383,
      "step": 6030
    },
    {
      "epoch": 53.02,
      "learning_rate": 5.171838505171839e-06,
      "loss": 0.7003,
      "step": 6040
    },
    {
      "epoch": 53.02,
      "eval_accuracy": 0.7004608294930875,
      "eval_loss": 1.1521508693695068,
      "eval_runtime": 293.3646,
      "eval_samples_per_second": 0.74,
      "eval_steps_per_second": 0.187,
      "step": 6048
    },
    {
      "epoch": 54.0,
      "learning_rate": 5.0884217550884214e-06,
      "loss": 0.6495,
      "step": 6050
    },
    {
      "epoch": 54.0,
      "learning_rate": 5.005005005005006e-06,
      "loss": 0.5755,
      "step": 6060
    },
    {
      "epoch": 54.0,
      "learning_rate": 4.921588254921588e-06,
      "loss": 0.5131,
      "step": 6070
    },
    {
      "epoch": 54.0,
      "learning_rate": 4.838171504838172e-06,
      "loss": 0.5332,
      "step": 6080
    },
    {
      "epoch": 54.01,
      "learning_rate": 4.754754754754754e-06,
      "loss": 0.2762,
      "step": 6090
    },
    {
      "epoch": 54.01,
      "learning_rate": 4.671338004671339e-06,
      "loss": 0.383,
      "step": 6100
    },
    {
      "epoch": 54.01,
      "learning_rate": 4.587921254587921e-06,
      "loss": 0.6731,
      "step": 6110
    },
    {
      "epoch": 54.01,
      "learning_rate": 4.504504504504505e-06,
      "loss": 0.4703,
      "step": 6120
    },
    {
      "epoch": 54.01,
      "learning_rate": 4.421087754421088e-06,
      "loss": 0.3767,
      "step": 6130
    },
    {
      "epoch": 54.01,
      "learning_rate": 4.337671004337672e-06,
      "loss": 0.5669,
      "step": 6140
    },
    {
      "epoch": 54.02,
      "learning_rate": 4.254254254254254e-06,
      "loss": 0.4277,
      "step": 6150
    },
    {
      "epoch": 54.02,
      "learning_rate": 4.170837504170838e-06,
      "loss": 0.4105,
      "step": 6160
    },
    {
      "epoch": 54.02,
      "eval_accuracy": 0.7373271889400922,
      "eval_loss": 1.0037211179733276,
      "eval_runtime": 262.5907,
      "eval_samples_per_second": 0.826,
      "eval_steps_per_second": 0.209,
      "step": 6160
    },
    {
      "epoch": 55.0,
      "learning_rate": 4.087420754087421e-06,
      "loss": 0.2836,
      "step": 6170
    },
    {
      "epoch": 55.0,
      "learning_rate": 4.004004004004004e-06,
      "loss": 0.6498,
      "step": 6180
    },
    {
      "epoch": 55.0,
      "learning_rate": 3.920587253920587e-06,
      "loss": 0.5334,
      "step": 6190
    },
    {
      "epoch": 55.01,
      "learning_rate": 3.837170503837171e-06,
      "loss": 0.7971,
      "step": 6200
    },
    {
      "epoch": 55.01,
      "learning_rate": 3.753753753753754e-06,
      "loss": 0.3278,
      "step": 6210
    },
    {
      "epoch": 55.01,
      "learning_rate": 3.670337003670337e-06,
      "loss": 0.3237,
      "step": 6220
    },
    {
      "epoch": 55.01,
      "learning_rate": 3.5869202535869206e-06,
      "loss": 0.7723,
      "step": 6230
    },
    {
      "epoch": 55.01,
      "learning_rate": 3.503503503503504e-06,
      "loss": 0.6996,
      "step": 6240
    },
    {
      "epoch": 55.01,
      "learning_rate": 3.420086753420087e-06,
      "loss": 0.544,
      "step": 6250
    },
    {
      "epoch": 55.02,
      "learning_rate": 3.33667000333667e-06,
      "loss": 0.6693,
      "step": 6260
    },
    {
      "epoch": 55.02,
      "learning_rate": 3.2532532532532535e-06,
      "loss": 0.5206,
      "step": 6270
    },
    {
      "epoch": 55.02,
      "eval_accuracy": 0.7188940092165899,
      "eval_loss": 1.0913441181182861,
      "eval_runtime": 299.8505,
      "eval_samples_per_second": 0.724,
      "eval_steps_per_second": 0.183,
      "step": 6272
    },
    {
      "epoch": 56.0,
      "learning_rate": 3.1698365031698365e-06,
      "loss": 0.2842,
      "step": 6280
    },
    {
      "epoch": 56.0,
      "learning_rate": 3.0864197530864196e-06,
      "loss": 0.4048,
      "step": 6290
    },
    {
      "epoch": 56.0,
      "learning_rate": 3.003003003003003e-06,
      "loss": 0.6278,
      "step": 6300
    },
    {
      "epoch": 56.01,
      "learning_rate": 2.9195862529195865e-06,
      "loss": 0.4073,
      "step": 6310
    },
    {
      "epoch": 56.01,
      "learning_rate": 2.8361695028361695e-06,
      "loss": 0.7256,
      "step": 6320
    },
    {
      "epoch": 56.01,
      "learning_rate": 2.752752752752753e-06,
      "loss": 0.5157,
      "step": 6330
    },
    {
      "epoch": 56.01,
      "learning_rate": 2.669336002669336e-06,
      "loss": 0.205,
      "step": 6340
    },
    {
      "epoch": 56.01,
      "learning_rate": 2.5859192525859194e-06,
      "loss": 0.513,
      "step": 6350
    },
    {
      "epoch": 56.01,
      "learning_rate": 2.502502502502503e-06,
      "loss": 0.3176,
      "step": 6360
    },
    {
      "epoch": 56.01,
      "learning_rate": 2.419085752419086e-06,
      "loss": 0.4901,
      "step": 6370
    },
    {
      "epoch": 56.02,
      "learning_rate": 2.3356690023356693e-06,
      "loss": 0.7129,
      "step": 6380
    },
    {
      "epoch": 56.02,
      "eval_accuracy": 0.6866359447004609,
      "eval_loss": 1.108277678489685,
      "eval_runtime": 292.9571,
      "eval_samples_per_second": 0.741,
      "eval_steps_per_second": 0.188,
      "step": 6384
    },
    {
      "epoch": 57.0,
      "learning_rate": 2.2522522522522524e-06,
      "loss": 0.2624,
      "step": 6390
    },
    {
      "epoch": 57.0,
      "learning_rate": 2.168835502168836e-06,
      "loss": 0.0914,
      "step": 6400
    },
    {
      "epoch": 57.0,
      "learning_rate": 2.085418752085419e-06,
      "loss": 0.684,
      "step": 6410
    },
    {
      "epoch": 57.01,
      "learning_rate": 2.002002002002002e-06,
      "loss": 0.3441,
      "step": 6420
    },
    {
      "epoch": 57.01,
      "learning_rate": 1.9185852519185853e-06,
      "loss": 0.4939,
      "step": 6430
    },
    {
      "epoch": 57.01,
      "learning_rate": 1.8351685018351686e-06,
      "loss": 0.1774,
      "step": 6440
    },
    {
      "epoch": 57.01,
      "learning_rate": 1.751751751751752e-06,
      "loss": 0.5289,
      "step": 6450
    },
    {
      "epoch": 57.01,
      "learning_rate": 1.668335001668335e-06,
      "loss": 0.9437,
      "step": 6460
    },
    {
      "epoch": 57.01,
      "learning_rate": 1.5849182515849183e-06,
      "loss": 0.7129,
      "step": 6470
    },
    {
      "epoch": 57.01,
      "learning_rate": 1.5015015015015015e-06,
      "loss": 0.1014,
      "step": 6480
    },
    {
      "epoch": 57.02,
      "learning_rate": 1.4180847514180847e-06,
      "loss": 0.4772,
      "step": 6490
    },
    {
      "epoch": 57.02,
      "eval_accuracy": 0.7142857142857143,
      "eval_loss": 1.1276459693908691,
      "eval_runtime": 263.52,
      "eval_samples_per_second": 0.823,
      "eval_steps_per_second": 0.209,
      "step": 6496
    },
    {
      "epoch": 58.0,
      "learning_rate": 1.334668001334668e-06,
      "loss": 0.5678,
      "step": 6500
    },
    {
      "epoch": 58.0,
      "learning_rate": 1.2512512512512514e-06,
      "loss": 0.2869,
      "step": 6510
    },
    {
      "epoch": 58.0,
      "learning_rate": 1.1678345011678347e-06,
      "loss": 0.9173,
      "step": 6520
    },
    {
      "epoch": 58.01,
      "learning_rate": 1.084417751084418e-06,
      "loss": 0.5002,
      "step": 6530
    },
    {
      "epoch": 58.01,
      "learning_rate": 1.001001001001001e-06,
      "loss": 0.4222,
      "step": 6540
    },
    {
      "epoch": 58.01,
      "learning_rate": 9.175842509175843e-07,
      "loss": 0.2793,
      "step": 6550
    },
    {
      "epoch": 58.01,
      "learning_rate": 8.341675008341675e-07,
      "loss": 0.2781,
      "step": 6560
    },
    {
      "epoch": 58.01,
      "learning_rate": 7.507507507507508e-07,
      "loss": 0.4779,
      "step": 6570
    },
    {
      "epoch": 58.01,
      "learning_rate": 6.67334000667334e-07,
      "loss": 0.4709,
      "step": 6580
    },
    {
      "epoch": 58.01,
      "learning_rate": 5.839172505839173e-07,
      "loss": 0.5879,
      "step": 6590
    },
    {
      "epoch": 58.02,
      "learning_rate": 5.005005005005005e-07,
      "loss": 0.4822,
      "step": 6600
    },
    {
      "epoch": 58.02,
      "eval_accuracy": 0.7235023041474654,
      "eval_loss": 1.0919580459594727,
      "eval_runtime": 304.8665,
      "eval_samples_per_second": 0.712,
      "eval_steps_per_second": 0.18,
      "step": 6608
    },
    {
      "epoch": 59.0,
      "learning_rate": 4.1708375041708376e-07,
      "loss": 0.601,
      "step": 6610
    },
    {
      "epoch": 59.0,
      "learning_rate": 3.33667000333667e-07,
      "loss": 0.5131,
      "step": 6620
    },
    {
      "epoch": 59.0,
      "learning_rate": 2.5025025025025023e-07,
      "loss": 0.175,
      "step": 6630
    },
    {
      "epoch": 59.0,
      "learning_rate": 1.668335001668335e-07,
      "loss": 0.4028,
      "step": 6640
    },
    {
      "epoch": 59.01,
      "learning_rate": 8.341675008341675e-08,
      "loss": 0.5526,
      "step": 6650
    },
    {
      "epoch": 59.01,
      "learning_rate": 0.0,
      "loss": 0.6307,
      "step": 6660
    },
    {
      "epoch": 59.01,
      "eval_accuracy": 0.7188940092165899,
      "eval_loss": 1.098652958869934,
      "eval_runtime": 289.552,
      "eval_samples_per_second": 0.749,
      "eval_steps_per_second": 0.19,
      "step": 6660
    },
    {
      "epoch": 59.01,
      "step": 6660,
      "total_flos": 3.3122529003141366e+19,
      "train_loss": 0.7877646436920395,
      "train_runtime": 57566.1475,
      "train_samples_per_second": 0.463,
      "train_steps_per_second": 0.116
    },
    {
      "epoch": 59.01,
      "eval_accuracy": 0.6296296296296297,
      "eval_loss": 0.8929917812347412,
      "eval_runtime": 260.8949,
      "eval_samples_per_second": 0.828,
      "eval_steps_per_second": 0.207,
      "step": 6660
    },
    {
      "epoch": 59.01,
      "eval_accuracy": 0.6296296296296297,
      "eval_loss": 0.8929917812347412,
      "eval_runtime": 274.034,
      "eval_samples_per_second": 0.788,
      "eval_steps_per_second": 0.197,
      "step": 6660
    }
  ],
  "logging_steps": 10,
  "max_steps": 6660,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 500,
  "total_flos": 3.3122529003141366e+19,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}