{
  "best_metric": 0.8968253968253969,
  "best_model_checkpoint": "dinov2-base-finetuned-har/checkpoint-4103",
  "epoch": 49.55223880597015,
  "eval_steps": 500,
  "global_step": 4150,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.11940298507462686,
      "grad_norm": 37.050025939941406,
      "learning_rate": 1.2048192771084338e-06,
      "loss": 3.2333,
      "step": 10
    },
    {
      "epoch": 0.23880597014925373,
      "grad_norm": 39.30177307128906,
      "learning_rate": 2.4096385542168676e-06,
      "loss": 2.9768,
      "step": 20
    },
    {
      "epoch": 0.3582089552238806,
      "grad_norm": 35.40634536743164,
      "learning_rate": 3.614457831325301e-06,
      "loss": 2.4932,
      "step": 30
    },
    {
      "epoch": 0.47761194029850745,
      "grad_norm": 44.90218734741211,
      "learning_rate": 4.819277108433735e-06,
      "loss": 1.9544,
      "step": 40
    },
    {
      "epoch": 0.5970149253731343,
      "grad_norm": 36.5886116027832,
      "learning_rate": 6.024096385542169e-06,
      "loss": 1.514,
      "step": 50
    },
    {
      "epoch": 0.7164179104477612,
      "grad_norm": 35.85557556152344,
      "learning_rate": 7.228915662650602e-06,
      "loss": 1.2005,
      "step": 60
    },
    {
      "epoch": 0.835820895522388,
      "grad_norm": 33.366024017333984,
      "learning_rate": 8.433734939759036e-06,
      "loss": 1.0445,
      "step": 70
    },
    {
      "epoch": 0.9552238805970149,
      "grad_norm": 34.10721969604492,
      "learning_rate": 9.63855421686747e-06,
      "loss": 0.9155,
      "step": 80
    },
    {
      "epoch": 0.991044776119403,
      "eval_accuracy": 0.8052910052910053,
      "eval_loss": 0.6204278469085693,
      "eval_runtime": 51.558,
      "eval_samples_per_second": 36.658,
      "eval_steps_per_second": 1.164,
      "step": 83
    },
    {
      "epoch": 1.0746268656716418,
      "grad_norm": 38.803226470947266,
      "learning_rate": 1.0843373493975904e-05,
      "loss": 0.8863,
      "step": 90
    },
    {
      "epoch": 1.1940298507462686,
      "grad_norm": 35.97611999511719,
      "learning_rate": 1.2048192771084338e-05,
      "loss": 0.8112,
      "step": 100
    },
    {
      "epoch": 1.3134328358208955,
      "grad_norm": 35.07186508178711,
      "learning_rate": 1.3253012048192772e-05,
      "loss": 0.7862,
      "step": 110
    },
    {
      "epoch": 1.4328358208955223,
      "grad_norm": 41.96965789794922,
      "learning_rate": 1.4457831325301205e-05,
      "loss": 0.8454,
      "step": 120
    },
    {
      "epoch": 1.5522388059701493,
      "grad_norm": 42.4810676574707,
      "learning_rate": 1.566265060240964e-05,
      "loss": 0.8309,
      "step": 130
    },
    {
      "epoch": 1.671641791044776,
      "grad_norm": 28.388389587402344,
      "learning_rate": 1.6867469879518073e-05,
      "loss": 0.754,
      "step": 140
    },
    {
      "epoch": 1.7910447761194028,
      "grad_norm": 35.19267654418945,
      "learning_rate": 1.8072289156626505e-05,
      "loss": 0.7247,
      "step": 150
    },
    {
      "epoch": 1.9104477611940298,
      "grad_norm": 34.15055847167969,
      "learning_rate": 1.927710843373494e-05,
      "loss": 0.749,
      "step": 160
    },
    {
      "epoch": 1.9940298507462688,
      "eval_accuracy": 0.8666666666666667,
      "eval_loss": 0.44325119256973267,
      "eval_runtime": 27.443,
      "eval_samples_per_second": 68.87,
      "eval_steps_per_second": 2.186,
      "step": 167
    },
    {
      "epoch": 2.029850746268657,
      "grad_norm": 27.7845516204834,
      "learning_rate": 2.0481927710843373e-05,
      "loss": 0.7159,
      "step": 170
    },
    {
      "epoch": 2.1492537313432836,
      "grad_norm": 32.60486602783203,
      "learning_rate": 2.168674698795181e-05,
      "loss": 0.6907,
      "step": 180
    },
    {
      "epoch": 2.2686567164179103,
      "grad_norm": 34.04874038696289,
      "learning_rate": 2.289156626506024e-05,
      "loss": 0.7388,
      "step": 190
    },
    {
      "epoch": 2.388059701492537,
      "grad_norm": 44.46305465698242,
      "learning_rate": 2.4096385542168677e-05,
      "loss": 0.7852,
      "step": 200
    },
    {
      "epoch": 2.5074626865671643,
      "grad_norm": 25.580005645751953,
      "learning_rate": 2.530120481927711e-05,
      "loss": 0.6814,
      "step": 210
    },
    {
      "epoch": 2.626865671641791,
      "grad_norm": 38.6807861328125,
      "learning_rate": 2.6506024096385545e-05,
      "loss": 0.7193,
      "step": 220
    },
    {
      "epoch": 2.746268656716418,
      "grad_norm": 29.82606315612793,
      "learning_rate": 2.7710843373493977e-05,
      "loss": 0.7395,
      "step": 230
    },
    {
      "epoch": 2.8656716417910446,
      "grad_norm": 27.26519203186035,
      "learning_rate": 2.891566265060241e-05,
      "loss": 0.7069,
      "step": 240
    },
    {
      "epoch": 2.9850746268656714,
      "grad_norm": 50.2726936340332,
      "learning_rate": 3.012048192771085e-05,
      "loss": 0.8197,
      "step": 250
    },
    {
      "epoch": 2.997014925373134,
      "eval_accuracy": 0.8571428571428571,
      "eval_loss": 0.4825701117515564,
      "eval_runtime": 27.1447,
      "eval_samples_per_second": 69.627,
      "eval_steps_per_second": 2.21,
      "step": 251
    },
    {
      "epoch": 3.1044776119402986,
      "grad_norm": 32.886810302734375,
      "learning_rate": 3.132530120481928e-05,
      "loss": 0.719,
      "step": 260
    },
    {
      "epoch": 3.2238805970149254,
      "grad_norm": 25.671785354614258,
      "learning_rate": 3.253012048192771e-05,
      "loss": 0.6762,
      "step": 270
    },
    {
      "epoch": 3.343283582089552,
      "grad_norm": 33.79856491088867,
      "learning_rate": 3.3734939759036146e-05,
      "loss": 0.6348,
      "step": 280
    },
    {
      "epoch": 3.4626865671641793,
      "grad_norm": 25.73578453063965,
      "learning_rate": 3.4939759036144585e-05,
      "loss": 0.6784,
      "step": 290
    },
    {
      "epoch": 3.582089552238806,
      "grad_norm": 40.199222564697266,
      "learning_rate": 3.614457831325301e-05,
      "loss": 0.6659,
      "step": 300
    },
    {
      "epoch": 3.701492537313433,
      "grad_norm": 30.21065902709961,
      "learning_rate": 3.734939759036144e-05,
      "loss": 0.7042,
      "step": 310
    },
    {
      "epoch": 3.8208955223880596,
      "grad_norm": 32.785037994384766,
      "learning_rate": 3.855421686746988e-05,
      "loss": 0.7352,
      "step": 320
    },
    {
      "epoch": 3.9402985074626864,
      "grad_norm": 27.440940856933594,
      "learning_rate": 3.9759036144578314e-05,
      "loss": 0.6854,
      "step": 330
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8724867724867725,
      "eval_loss": 0.42430025339126587,
      "eval_runtime": 32.3022,
      "eval_samples_per_second": 58.51,
      "eval_steps_per_second": 1.857,
      "step": 335
    },
    {
      "epoch": 4.059701492537314,
      "grad_norm": 28.7294921875,
      "learning_rate": 4.0963855421686746e-05,
      "loss": 0.6307,
      "step": 340
    },
    {
      "epoch": 4.17910447761194,
      "grad_norm": 29.991897583007812,
      "learning_rate": 4.2168674698795186e-05,
      "loss": 0.6336,
      "step": 350
    },
    {
      "epoch": 4.298507462686567,
      "grad_norm": 27.699886322021484,
      "learning_rate": 4.337349397590362e-05,
      "loss": 0.7364,
      "step": 360
    },
    {
      "epoch": 4.417910447761194,
      "grad_norm": 64.75654602050781,
      "learning_rate": 4.457831325301205e-05,
      "loss": 0.6794,
      "step": 370
    },
    {
      "epoch": 4.537313432835821,
      "grad_norm": 28.235536575317383,
      "learning_rate": 4.578313253012048e-05,
      "loss": 0.6682,
      "step": 380
    },
    {
      "epoch": 4.656716417910448,
      "grad_norm": 30.604808807373047,
      "learning_rate": 4.698795180722892e-05,
      "loss": 0.6224,
      "step": 390
    },
    {
      "epoch": 4.776119402985074,
      "grad_norm": 26.9096736907959,
      "learning_rate": 4.8192771084337354e-05,
      "loss": 0.6643,
      "step": 400
    },
    {
      "epoch": 4.895522388059701,
      "grad_norm": 45.720359802246094,
      "learning_rate": 4.9397590361445786e-05,
      "loss": 0.7058,
      "step": 410
    },
    {
      "epoch": 4.991044776119403,
      "eval_accuracy": 0.8592592592592593,
      "eval_loss": 0.434872031211853,
      "eval_runtime": 21.0424,
      "eval_samples_per_second": 89.818,
      "eval_steps_per_second": 2.851,
      "step": 418
    },
    {
      "epoch": 5.014925373134329,
      "grad_norm": 25.985727310180664,
      "learning_rate": 4.99330655957162e-05,
      "loss": 0.7109,
      "step": 420
    },
    {
      "epoch": 5.134328358208955,
      "grad_norm": 38.45668029785156,
      "learning_rate": 4.97991967871486e-05,
      "loss": 0.6265,
      "step": 430
    },
    {
      "epoch": 5.253731343283582,
      "grad_norm": 38.37868881225586,
      "learning_rate": 4.9665327978580993e-05,
      "loss": 0.6318,
      "step": 440
    },
    {
      "epoch": 5.373134328358209,
      "grad_norm": 23.872426986694336,
      "learning_rate": 4.953145917001339e-05,
      "loss": 0.5749,
      "step": 450
    },
    {
      "epoch": 5.492537313432836,
      "grad_norm": 30.97602653503418,
      "learning_rate": 4.9397590361445786e-05,
      "loss": 0.629,
      "step": 460
    },
    {
      "epoch": 5.611940298507463,
      "grad_norm": 54.77509689331055,
      "learning_rate": 4.926372155287818e-05,
      "loss": 0.6286,
      "step": 470
    },
    {
      "epoch": 5.731343283582089,
      "grad_norm": 58.17964553833008,
      "learning_rate": 4.912985274431058e-05,
      "loss": 0.6289,
      "step": 480
    },
    {
      "epoch": 5.850746268656716,
      "grad_norm": 25.494422912597656,
      "learning_rate": 4.8995983935742975e-05,
      "loss": 0.6603,
      "step": 490
    },
    {
      "epoch": 5.970149253731344,
      "grad_norm": 31.255512237548828,
      "learning_rate": 4.886211512717537e-05,
      "loss": 0.6717,
      "step": 500
    },
    {
      "epoch": 5.994029850746268,
      "eval_accuracy": 0.8433862433862434,
      "eval_loss": 0.49840009212493896,
      "eval_runtime": 22.0815,
      "eval_samples_per_second": 85.592,
      "eval_steps_per_second": 2.717,
      "step": 502
    },
    {
      "epoch": 6.08955223880597,
      "grad_norm": 25.194740295410156,
      "learning_rate": 4.872824631860777e-05,
      "loss": 0.5848,
      "step": 510
    },
    {
      "epoch": 6.208955223880597,
      "grad_norm": 53.04393005371094,
      "learning_rate": 4.8594377510040165e-05,
      "loss": 0.6182,
      "step": 520
    },
    {
      "epoch": 6.3283582089552235,
      "grad_norm": 35.287010192871094,
      "learning_rate": 4.8460508701472554e-05,
      "loss": 0.6342,
      "step": 530
    },
    {
      "epoch": 6.447761194029851,
      "grad_norm": 37.487796783447266,
      "learning_rate": 4.832663989290496e-05,
      "loss": 0.6249,
      "step": 540
    },
    {
      "epoch": 6.567164179104478,
      "grad_norm": 20.985584259033203,
      "learning_rate": 4.8192771084337354e-05,
      "loss": 0.6296,
      "step": 550
    },
    {
      "epoch": 6.686567164179104,
      "grad_norm": 25.937698364257812,
      "learning_rate": 4.8058902275769744e-05,
      "loss": 0.5756,
      "step": 560
    },
    {
      "epoch": 6.8059701492537314,
      "grad_norm": 24.166778564453125,
      "learning_rate": 4.792503346720215e-05,
      "loss": 0.6639,
      "step": 570
    },
    {
      "epoch": 6.925373134328359,
      "grad_norm": 17.558452606201172,
      "learning_rate": 4.779116465863454e-05,
      "loss": 0.6544,
      "step": 580
    },
    {
      "epoch": 6.997014925373135,
      "eval_accuracy": 0.8544973544973545,
      "eval_loss": 0.47304412722587585,
      "eval_runtime": 24.5582,
      "eval_samples_per_second": 76.96,
      "eval_steps_per_second": 2.443,
      "step": 586
    },
    {
      "epoch": 7.044776119402985,
      "grad_norm": 25.491958618164062,
      "learning_rate": 4.765729585006693e-05,
      "loss": 0.5738,
      "step": 590
    },
    {
      "epoch": 7.164179104477612,
      "grad_norm": 22.450450897216797,
      "learning_rate": 4.7523427041499336e-05,
      "loss": 0.5737,
      "step": 600
    },
    {
      "epoch": 7.2835820895522385,
      "grad_norm": 25.659818649291992,
      "learning_rate": 4.738955823293173e-05,
      "loss": 0.5931,
      "step": 610
    },
    {
      "epoch": 7.402985074626866,
      "grad_norm": 32.22180938720703,
      "learning_rate": 4.725568942436412e-05,
      "loss": 0.5595,
      "step": 620
    },
    {
      "epoch": 7.522388059701493,
      "grad_norm": 23.480642318725586,
      "learning_rate": 4.712182061579652e-05,
      "loss": 0.5705,
      "step": 630
    },
    {
      "epoch": 7.641791044776119,
      "grad_norm": 23.681175231933594,
      "learning_rate": 4.698795180722892e-05,
      "loss": 0.5688,
      "step": 640
    },
    {
      "epoch": 7.7611940298507465,
      "grad_norm": 24.127212524414062,
      "learning_rate": 4.685408299866131e-05,
      "loss": 0.606,
      "step": 650
    },
    {
      "epoch": 7.880597014925373,
      "grad_norm": 23.920379638671875,
      "learning_rate": 4.672021419009371e-05,
      "loss": 0.5718,
      "step": 660
    },
    {
      "epoch": 8.0,
      "grad_norm": 30.988821029663086,
      "learning_rate": 4.658634538152611e-05,
      "loss": 0.5846,
      "step": 670
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8629629629629629,
      "eval_loss": 0.4630894362926483,
      "eval_runtime": 22.8566,
      "eval_samples_per_second": 82.69,
      "eval_steps_per_second": 2.625,
      "step": 670
    },
    {
      "epoch": 8.119402985074627,
      "grad_norm": 23.982349395751953,
      "learning_rate": 4.64524765729585e-05,
      "loss": 0.5197,
      "step": 680
    },
    {
      "epoch": 8.238805970149254,
      "grad_norm": 26.285791397094727,
      "learning_rate": 4.63186077643909e-05,
      "loss": 0.5108,
      "step": 690
    },
    {
      "epoch": 8.35820895522388,
      "grad_norm": 18.24249839782715,
      "learning_rate": 4.61847389558233e-05,
      "loss": 0.5251,
      "step": 700
    },
    {
      "epoch": 8.477611940298507,
      "grad_norm": 26.010107040405273,
      "learning_rate": 4.605087014725569e-05,
      "loss": 0.6023,
      "step": 710
    },
    {
      "epoch": 8.597014925373134,
      "grad_norm": 28.695484161376953,
      "learning_rate": 4.5917001338688086e-05,
      "loss": 0.507,
      "step": 720
    },
    {
      "epoch": 8.716417910447761,
      "grad_norm": 17.62828254699707,
      "learning_rate": 4.578313253012048e-05,
      "loss": 0.4988,
      "step": 730
    },
    {
      "epoch": 8.835820895522389,
      "grad_norm": 35.14638900756836,
      "learning_rate": 4.564926372155288e-05,
      "loss": 0.5591,
      "step": 740
    },
    {
      "epoch": 8.955223880597014,
      "grad_norm": 30.526409149169922,
      "learning_rate": 4.5515394912985275e-05,
      "loss": 0.5207,
      "step": 750
    },
    {
      "epoch": 8.991044776119402,
      "eval_accuracy": 0.8751322751322751,
      "eval_loss": 0.4072401523590088,
      "eval_runtime": 21.1198,
      "eval_samples_per_second": 89.489,
      "eval_steps_per_second": 2.841,
      "step": 753
    },
    {
      "epoch": 9.074626865671641,
      "grad_norm": 20.028371810913086,
      "learning_rate": 4.538152610441767e-05,
      "loss": 0.4773,
      "step": 760
    },
    {
      "epoch": 9.194029850746269,
      "grad_norm": 21.046306610107422,
      "learning_rate": 4.524765729585007e-05,
      "loss": 0.5069,
      "step": 770
    },
    {
      "epoch": 9.313432835820896,
      "grad_norm": 21.227153778076172,
      "learning_rate": 4.5113788487282465e-05,
      "loss": 0.462,
      "step": 780
    },
    {
      "epoch": 9.432835820895523,
      "grad_norm": 27.245750427246094,
      "learning_rate": 4.497991967871486e-05,
      "loss": 0.4378,
      "step": 790
    },
    {
      "epoch": 9.552238805970148,
      "grad_norm": 12.378539085388184,
      "learning_rate": 4.484605087014726e-05,
      "loss": 0.4878,
      "step": 800
    },
    {
      "epoch": 9.671641791044776,
      "grad_norm": 19.924983978271484,
      "learning_rate": 4.4712182061579654e-05,
      "loss": 0.5377,
      "step": 810
    },
    {
      "epoch": 9.791044776119403,
      "grad_norm": 31.331254959106445,
      "learning_rate": 4.457831325301205e-05,
      "loss": 0.5306,
      "step": 820
    },
    {
      "epoch": 9.91044776119403,
      "grad_norm": 25.974903106689453,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 0.4977,
      "step": 830
    },
    {
      "epoch": 9.99402985074627,
      "eval_accuracy": 0.8608465608465609,
      "eval_loss": 0.46266716718673706,
      "eval_runtime": 34.1921,
      "eval_samples_per_second": 55.276,
      "eval_steps_per_second": 1.755,
      "step": 837
    },
    {
      "epoch": 10.029850746268657,
      "grad_norm": 26.603260040283203,
      "learning_rate": 4.431057563587684e-05,
      "loss": 0.4794,
      "step": 840
    },
    {
      "epoch": 10.149253731343283,
      "grad_norm": 28.95216178894043,
      "learning_rate": 4.417670682730924e-05,
      "loss": 0.443,
      "step": 850
    },
    {
      "epoch": 10.26865671641791,
      "grad_norm": 16.765554428100586,
      "learning_rate": 4.4042838018741636e-05,
      "loss": 0.4752,
      "step": 860
    },
    {
      "epoch": 10.388059701492537,
      "grad_norm": 25.070423126220703,
      "learning_rate": 4.390896921017403e-05,
      "loss": 0.5519,
      "step": 870
    },
    {
      "epoch": 10.507462686567164,
      "grad_norm": 18.10160255432129,
      "learning_rate": 4.377510040160643e-05,
      "loss": 0.4328,
      "step": 880
    },
    {
      "epoch": 10.626865671641792,
      "grad_norm": 32.88917922973633,
      "learning_rate": 4.3641231593038825e-05,
      "loss": 0.4394,
      "step": 890
    },
    {
      "epoch": 10.746268656716419,
      "grad_norm": 23.60790252685547,
      "learning_rate": 4.350736278447122e-05,
      "loss": 0.5036,
      "step": 900
    },
    {
      "epoch": 10.865671641791044,
      "grad_norm": 18.67464828491211,
      "learning_rate": 4.337349397590362e-05,
      "loss": 0.4576,
      "step": 910
    },
    {
      "epoch": 10.985074626865671,
      "grad_norm": 20.74434471130371,
      "learning_rate": 4.3239625167336014e-05,
      "loss": 0.4974,
      "step": 920
    },
    {
      "epoch": 10.997014925373135,
      "eval_accuracy": 0.8661375661375661,
      "eval_loss": 0.4599558711051941,
      "eval_runtime": 23.4296,
      "eval_samples_per_second": 80.667,
      "eval_steps_per_second": 2.561,
      "step": 921
    },
    {
      "epoch": 11.104477611940299,
      "grad_norm": 17.247478485107422,
      "learning_rate": 4.3105756358768404e-05,
      "loss": 0.4217,
      "step": 930
    },
    {
      "epoch": 11.223880597014926,
      "grad_norm": 24.32207489013672,
      "learning_rate": 4.297188755020081e-05,
      "loss": 0.426,
      "step": 940
    },
    {
      "epoch": 11.343283582089553,
      "grad_norm": 19.61281394958496,
      "learning_rate": 4.2838018741633203e-05,
      "loss": 0.4266,
      "step": 950
    },
    {
      "epoch": 11.462686567164178,
      "grad_norm": 34.41626739501953,
      "learning_rate": 4.270414993306559e-05,
      "loss": 0.4094,
      "step": 960
    },
    {
      "epoch": 11.582089552238806,
      "grad_norm": 17.059947967529297,
      "learning_rate": 4.2570281124497996e-05,
      "loss": 0.4498,
      "step": 970
    },
    {
      "epoch": 11.701492537313433,
      "grad_norm": 14.022907257080078,
      "learning_rate": 4.243641231593039e-05,
      "loss": 0.4787,
      "step": 980
    },
    {
      "epoch": 11.82089552238806,
      "grad_norm": 19.341840744018555,
      "learning_rate": 4.230254350736278e-05,
      "loss": 0.5272,
      "step": 990
    },
    {
      "epoch": 11.940298507462687,
      "grad_norm": 23.444053649902344,
      "learning_rate": 4.2168674698795186e-05,
      "loss": 0.4502,
      "step": 1000
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.8724867724867725,
      "eval_loss": 0.45481687784194946,
      "eval_runtime": 21.0538,
      "eval_samples_per_second": 89.77,
      "eval_steps_per_second": 2.85,
      "step": 1005
    },
    {
      "epoch": 12.059701492537313,
      "grad_norm": 14.740572929382324,
      "learning_rate": 4.203480589022758e-05,
      "loss": 0.4053,
      "step": 1010
    },
    {
      "epoch": 12.17910447761194,
      "grad_norm": 21.706453323364258,
      "learning_rate": 4.190093708165997e-05,
      "loss": 0.4259,
      "step": 1020
    },
    {
      "epoch": 12.298507462686567,
      "grad_norm": 17.945667266845703,
      "learning_rate": 4.176706827309237e-05,
      "loss": 0.4064,
      "step": 1030
    },
    {
      "epoch": 12.417910447761194,
      "grad_norm": 18.671659469604492,
      "learning_rate": 4.163319946452477e-05,
      "loss": 0.4299,
      "step": 1040
    },
    {
      "epoch": 12.537313432835822,
      "grad_norm": 19.7445068359375,
      "learning_rate": 4.149933065595716e-05,
      "loss": 0.4439,
      "step": 1050
    },
    {
      "epoch": 12.656716417910447,
      "grad_norm": 20.6423282623291,
      "learning_rate": 4.136546184738956e-05,
      "loss": 0.4782,
      "step": 1060
    },
    {
      "epoch": 12.776119402985074,
      "grad_norm": 20.769142150878906,
      "learning_rate": 4.123159303882196e-05,
      "loss": 0.4127,
      "step": 1070
    },
    {
      "epoch": 12.895522388059701,
      "grad_norm": 31.262954711914062,
      "learning_rate": 4.109772423025435e-05,
      "loss": 0.4051,
      "step": 1080
    },
    {
      "epoch": 12.991044776119402,
      "eval_accuracy": 0.870899470899471,
      "eval_loss": 0.44037628173828125,
      "eval_runtime": 20.5272,
      "eval_samples_per_second": 92.073,
      "eval_steps_per_second": 2.923,
      "step": 1088
    },
    {
      "epoch": 13.014925373134329,
      "grad_norm": 16.262481689453125,
      "learning_rate": 4.0963855421686746e-05,
      "loss": 0.4364,
      "step": 1090
    },
    {
      "epoch": 13.134328358208956,
      "grad_norm": 11.120735168457031,
      "learning_rate": 4.082998661311915e-05,
      "loss": 0.3634,
      "step": 1100
    },
    {
      "epoch": 13.253731343283581,
      "grad_norm": 17.504487991333008,
      "learning_rate": 4.069611780455154e-05,
      "loss": 0.3874,
      "step": 1110
    },
    {
      "epoch": 13.373134328358208,
      "grad_norm": 27.072677612304688,
      "learning_rate": 4.0562248995983936e-05,
      "loss": 0.433,
      "step": 1120
    },
    {
      "epoch": 13.492537313432836,
      "grad_norm": 21.18752670288086,
      "learning_rate": 4.042838018741633e-05,
      "loss": 0.423,
      "step": 1130
    },
    {
      "epoch": 13.611940298507463,
      "grad_norm": 15.781579971313477,
      "learning_rate": 4.029451137884873e-05,
      "loss": 0.4167,
      "step": 1140
    },
    {
      "epoch": 13.73134328358209,
      "grad_norm": 21.992441177368164,
      "learning_rate": 4.0160642570281125e-05,
      "loss": 0.4286,
      "step": 1150
    },
    {
      "epoch": 13.850746268656717,
      "grad_norm": 18.54501724243164,
      "learning_rate": 4.002677376171352e-05,
      "loss": 0.465,
      "step": 1160
    },
    {
      "epoch": 13.970149253731343,
      "grad_norm": 14.206164360046387,
      "learning_rate": 3.989290495314592e-05,
      "loss": 0.3862,
      "step": 1170
    },
    {
      "epoch": 13.99402985074627,
      "eval_accuracy": 0.8772486772486773,
      "eval_loss": 0.4498014450073242,
      "eval_runtime": 21.3493,
      "eval_samples_per_second": 88.528,
      "eval_steps_per_second": 2.81,
      "step": 1172
    },
    {
      "epoch": 14.08955223880597,
      "grad_norm": 16.436885833740234,
      "learning_rate": 3.9759036144578314e-05,
      "loss": 0.4198,
      "step": 1180
    },
    {
      "epoch": 14.208955223880597,
      "grad_norm": 18.03758430480957,
      "learning_rate": 3.962516733601071e-05,
      "loss": 0.3928,
      "step": 1190
    },
    {
      "epoch": 14.328358208955224,
      "grad_norm": 18.285844802856445,
      "learning_rate": 3.949129852744311e-05,
      "loss": 0.41,
      "step": 1200
    },
    {
      "epoch": 14.447761194029852,
      "grad_norm": 25.42701530456543,
      "learning_rate": 3.93574297188755e-05,
      "loss": 0.369,
      "step": 1210
    },
    {
      "epoch": 14.567164179104477,
      "grad_norm": 37.38109588623047,
      "learning_rate": 3.92235609103079e-05,
      "loss": 0.3675,
      "step": 1220
    },
    {
      "epoch": 14.686567164179104,
      "grad_norm": 18.741134643554688,
      "learning_rate": 3.9089692101740296e-05,
      "loss": 0.399,
      "step": 1230
    },
    {
      "epoch": 14.805970149253731,
      "grad_norm": 13.165870666503906,
      "learning_rate": 3.895582329317269e-05,
      "loss": 0.4279,
      "step": 1240
    },
    {
      "epoch": 14.925373134328359,
      "grad_norm": 13.853320121765137,
      "learning_rate": 3.882195448460509e-05,
      "loss": 0.351,
      "step": 1250
    },
    {
      "epoch": 14.997014925373135,
      "eval_accuracy": 0.8677248677248677,
      "eval_loss": 0.48586639761924744,
      "eval_runtime": 23.1342,
      "eval_samples_per_second": 81.697,
      "eval_steps_per_second": 2.594,
      "step": 1256
    },
    {
      "epoch": 15.044776119402986,
      "grad_norm": 17.546924591064453,
      "learning_rate": 3.8688085676037485e-05,
      "loss": 0.3519,
      "step": 1260
    },
    {
      "epoch": 15.164179104477611,
      "grad_norm": 15.299020767211914,
      "learning_rate": 3.855421686746988e-05,
      "loss": 0.3644,
      "step": 1270
    },
    {
      "epoch": 15.283582089552239,
      "grad_norm": 27.95186424255371,
      "learning_rate": 3.842034805890228e-05,
      "loss": 0.3707,
      "step": 1280
    },
    {
      "epoch": 15.402985074626866,
      "grad_norm": 24.1954288482666,
      "learning_rate": 3.8286479250334675e-05,
      "loss": 0.3937,
      "step": 1290
    },
    {
      "epoch": 15.522388059701493,
      "grad_norm": 15.104287147521973,
      "learning_rate": 3.815261044176707e-05,
      "loss": 0.3693,
      "step": 1300
    },
    {
      "epoch": 15.64179104477612,
      "grad_norm": 20.633682250976562,
      "learning_rate": 3.801874163319947e-05,
      "loss": 0.3565,
      "step": 1310
    },
    {
      "epoch": 15.761194029850746,
      "grad_norm": 25.37423324584961,
      "learning_rate": 3.7884872824631864e-05,
      "loss": 0.3697,
      "step": 1320
    },
    {
      "epoch": 15.880597014925373,
      "grad_norm": 15.787097930908203,
      "learning_rate": 3.7751004016064253e-05,
      "loss": 0.3586,
      "step": 1330
    },
    {
      "epoch": 16.0,
      "grad_norm": 23.120027542114258,
      "learning_rate": 3.761713520749666e-05,
      "loss": 0.3807,
      "step": 1340
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.8555555555555555,
      "eval_loss": 0.5188712477684021,
      "eval_runtime": 25.4894,
      "eval_samples_per_second": 74.148,
      "eval_steps_per_second": 2.354,
      "step": 1340
    },
    {
      "epoch": 16.119402985074625,
      "grad_norm": 17.341602325439453,
      "learning_rate": 3.748326639892905e-05,
      "loss": 0.4134,
      "step": 1350
    },
    {
      "epoch": 16.238805970149254,
      "grad_norm": 25.106281280517578,
      "learning_rate": 3.734939759036144e-05,
      "loss": 0.3528,
      "step": 1360
    },
    {
      "epoch": 16.35820895522388,
      "grad_norm": 12.993470191955566,
      "learning_rate": 3.7215528781793846e-05,
      "loss": 0.3435,
      "step": 1370
    },
    {
      "epoch": 16.47761194029851,
      "grad_norm": 14.296526908874512,
      "learning_rate": 3.708165997322624e-05,
      "loss": 0.3641,
      "step": 1380
    },
    {
      "epoch": 16.597014925373134,
      "grad_norm": 17.266536712646484,
      "learning_rate": 3.694779116465863e-05,
      "loss": 0.4076,
      "step": 1390
    },
    {
      "epoch": 16.71641791044776,
      "grad_norm": 18.806644439697266,
      "learning_rate": 3.6813922356091035e-05,
      "loss": 0.312,
      "step": 1400
    },
    {
      "epoch": 16.83582089552239,
      "grad_norm": 20.539464950561523,
      "learning_rate": 3.668005354752343e-05,
      "loss": 0.3232,
      "step": 1410
    },
    {
      "epoch": 16.955223880597014,
      "grad_norm": 22.800634384155273,
      "learning_rate": 3.654618473895582e-05,
      "loss": 0.3538,
      "step": 1420
    },
    {
      "epoch": 16.991044776119402,
      "eval_accuracy": 0.8645502645502645,
      "eval_loss": 0.4959017336368561,
      "eval_runtime": 29.6901,
      "eval_samples_per_second": 63.658,
      "eval_steps_per_second": 2.021,
      "step": 1423
    },
    {
      "epoch": 17.074626865671643,
      "grad_norm": 28.333608627319336,
      "learning_rate": 3.641231593038822e-05,
      "loss": 0.3491,
      "step": 1430
    },
    {
      "epoch": 17.19402985074627,
      "grad_norm": 15.66159439086914,
      "learning_rate": 3.627844712182062e-05,
      "loss": 0.2808,
      "step": 1440
    },
    {
      "epoch": 17.313432835820894,
      "grad_norm": 17.449249267578125,
      "learning_rate": 3.614457831325301e-05,
      "loss": 0.3763,
      "step": 1450
    },
    {
      "epoch": 17.432835820895523,
      "grad_norm": 16.44668197631836,
      "learning_rate": 3.601070950468541e-05,
      "loss": 0.3173,
      "step": 1460
    },
    {
      "epoch": 17.55223880597015,
      "grad_norm": 12.903731346130371,
      "learning_rate": 3.587684069611781e-05,
      "loss": 0.3423,
      "step": 1470
    },
    {
      "epoch": 17.671641791044777,
      "grad_norm": 12.897427558898926,
      "learning_rate": 3.57429718875502e-05,
      "loss": 0.3042,
      "step": 1480
    },
    {
      "epoch": 17.791044776119403,
      "grad_norm": 21.114595413208008,
      "learning_rate": 3.5609103078982596e-05,
      "loss": 0.3684,
      "step": 1490
    },
    {
      "epoch": 17.91044776119403,
      "grad_norm": 27.15323829650879,
      "learning_rate": 3.5475234270415e-05,
      "loss": 0.3181,
      "step": 1500
    },
    {
      "epoch": 17.99402985074627,
      "eval_accuracy": 0.8698412698412699,
      "eval_loss": 0.48311787843704224,
      "eval_runtime": 20.9604,
      "eval_samples_per_second": 90.17,
      "eval_steps_per_second": 2.863,
      "step": 1507
    },
    {
      "epoch": 18.029850746268657,
      "grad_norm": 20.24980926513672,
      "learning_rate": 3.534136546184739e-05,
      "loss": 0.3785,
      "step": 1510
    },
    {
      "epoch": 18.149253731343283,
      "grad_norm": 20.919986724853516,
      "learning_rate": 3.5207496653279785e-05,
      "loss": 0.3465,
      "step": 1520
    },
    {
      "epoch": 18.26865671641791,
      "grad_norm": 18.047338485717773,
      "learning_rate": 3.507362784471218e-05,
      "loss": 0.3131,
      "step": 1530
    },
    {
      "epoch": 18.388059701492537,
      "grad_norm": 18.01116943359375,
      "learning_rate": 3.4939759036144585e-05,
      "loss": 0.3251,
      "step": 1540
    },
    {
      "epoch": 18.507462686567163,
      "grad_norm": 17.661109924316406,
      "learning_rate": 3.4805890227576974e-05,
      "loss": 0.3534,
      "step": 1550
    },
    {
      "epoch": 18.62686567164179,
      "grad_norm": 13.190411567687988,
      "learning_rate": 3.467202141900937e-05,
      "loss": 0.3443,
      "step": 1560
    },
    {
      "epoch": 18.746268656716417,
      "grad_norm": 11.302985191345215,
      "learning_rate": 3.4538152610441774e-05,
      "loss": 0.294,
      "step": 1570
    },
    {
      "epoch": 18.865671641791046,
      "grad_norm": 17.95700454711914,
      "learning_rate": 3.4404283801874164e-05,
      "loss": 0.3354,
      "step": 1580
    },
    {
      "epoch": 18.98507462686567,
      "grad_norm": 21.134395599365234,
      "learning_rate": 3.427041499330656e-05,
      "loss": 0.3225,
      "step": 1590
    },
    {
      "epoch": 18.997014925373133,
      "eval_accuracy": 0.8804232804232804,
      "eval_loss": 0.4890158772468567,
      "eval_runtime": 21.2829,
      "eval_samples_per_second": 88.804,
      "eval_steps_per_second": 2.819,
      "step": 1591
    },
    {
      "epoch": 19.104477611940297,
      "grad_norm": 20.105192184448242,
      "learning_rate": 3.413654618473896e-05,
      "loss": 0.2791,
      "step": 1600
    },
    {
      "epoch": 19.223880597014926,
      "grad_norm": 19.9129638671875,
      "learning_rate": 3.400267737617135e-05,
      "loss": 0.3031,
      "step": 1610
    },
    {
      "epoch": 19.34328358208955,
      "grad_norm": 18.273677825927734,
      "learning_rate": 3.386880856760375e-05,
      "loss": 0.3341,
      "step": 1620
    },
    {
      "epoch": 19.46268656716418,
      "grad_norm": 22.22028350830078,
      "learning_rate": 3.3734939759036146e-05,
      "loss": 0.2613,
      "step": 1630
    },
    {
      "epoch": 19.582089552238806,
      "grad_norm": 17.404489517211914,
      "learning_rate": 3.360107095046854e-05,
      "loss": 0.3275,
      "step": 1640
    },
    {
      "epoch": 19.701492537313435,
      "grad_norm": 11.650886535644531,
      "learning_rate": 3.346720214190094e-05,
      "loss": 0.3002,
      "step": 1650
    },
    {
      "epoch": 19.82089552238806,
      "grad_norm": 21.051326751708984,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.3075,
      "step": 1660
    },
    {
      "epoch": 19.940298507462686,
      "grad_norm": 11.516342163085938,
      "learning_rate": 3.319946452476573e-05,
      "loss": 0.3257,
      "step": 1670
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.8735449735449735,
      "eval_loss": 0.48166623711586,
      "eval_runtime": 35.9635,
      "eval_samples_per_second": 52.553,
      "eval_steps_per_second": 1.668,
      "step": 1675
    },
    {
      "epoch": 20.059701492537314,
      "grad_norm": 21.99034881591797,
      "learning_rate": 3.306559571619813e-05,
      "loss": 0.2885,
      "step": 1680
    },
    {
      "epoch": 20.17910447761194,
      "grad_norm": 15.522448539733887,
      "learning_rate": 3.2931726907630524e-05,
      "loss": 0.2835,
      "step": 1690
    },
    {
      "epoch": 20.298507462686565,
      "grad_norm": 16.3590030670166,
      "learning_rate": 3.279785809906292e-05,
      "loss": 0.2958,
      "step": 1700
    },
    {
      "epoch": 20.417910447761194,
      "grad_norm": 15.87469482421875,
      "learning_rate": 3.266398929049532e-05,
      "loss": 0.2831,
      "step": 1710
    },
    {
      "epoch": 20.53731343283582,
      "grad_norm": 16.18557357788086,
      "learning_rate": 3.253012048192771e-05,
      "loss": 0.2622,
      "step": 1720
    },
    {
      "epoch": 20.65671641791045,
      "grad_norm": 15.619478225708008,
      "learning_rate": 3.23962516733601e-05,
      "loss": 0.3094,
      "step": 1730
    },
    {
      "epoch": 20.776119402985074,
      "grad_norm": 16.444242477416992,
      "learning_rate": 3.2262382864792506e-05,
      "loss": 0.3312,
      "step": 1740
    },
    {
      "epoch": 20.895522388059703,
      "grad_norm": 14.049259185791016,
      "learning_rate": 3.21285140562249e-05,
      "loss": 0.2667,
      "step": 1750
    },
    {
      "epoch": 20.991044776119402,
      "eval_accuracy": 0.8682539682539683,
      "eval_loss": 0.5198962688446045,
      "eval_runtime": 21.6457,
      "eval_samples_per_second": 87.315,
      "eval_steps_per_second": 2.772,
      "step": 1758
    },
    {
      "epoch": 21.01492537313433,
      "grad_norm": 25.527427673339844,
      "learning_rate": 3.199464524765729e-05,
      "loss": 0.2932,
      "step": 1760
    },
    {
      "epoch": 21.134328358208954,
      "grad_norm": 20.6369686126709,
      "learning_rate": 3.1860776439089695e-05,
      "loss": 0.3097,
      "step": 1770
    },
    {
      "epoch": 21.253731343283583,
      "grad_norm": 17.76145362854004,
      "learning_rate": 3.172690763052209e-05,
      "loss": 0.3055,
      "step": 1780
    },
    {
      "epoch": 21.37313432835821,
      "grad_norm": 17.357999801635742,
      "learning_rate": 3.159303882195448e-05,
      "loss": 0.3102,
      "step": 1790
    },
    {
      "epoch": 21.492537313432837,
      "grad_norm": 16.11004638671875,
      "learning_rate": 3.1459170013386885e-05,
      "loss": 0.2703,
      "step": 1800
    },
    {
      "epoch": 21.611940298507463,
      "grad_norm": 14.16020679473877,
      "learning_rate": 3.132530120481928e-05,
      "loss": 0.3295,
      "step": 1810
    },
    {
      "epoch": 21.73134328358209,
      "grad_norm": 19.754079818725586,
      "learning_rate": 3.119143239625168e-05,
      "loss": 0.2677,
      "step": 1820
    },
    {
      "epoch": 21.850746268656717,
      "grad_norm": 15.820847511291504,
      "learning_rate": 3.105756358768407e-05,
      "loss": 0.2737,
      "step": 1830
    },
    {
      "epoch": 21.970149253731343,
      "grad_norm": 11.689366340637207,
      "learning_rate": 3.092369477911647e-05,
      "loss": 0.2863,
      "step": 1840
    },
    {
      "epoch": 21.99402985074627,
      "eval_accuracy": 0.8682539682539683,
      "eval_loss": 0.48346272110939026,
      "eval_runtime": 21.1384,
      "eval_samples_per_second": 89.411,
      "eval_steps_per_second": 2.838,
      "step": 1842
    },
    {
      "epoch": 22.08955223880597,
      "grad_norm": 9.886032104492188,
      "learning_rate": 3.078982597054887e-05,
      "loss": 0.2935,
      "step": 1850
    },
    {
      "epoch": 22.208955223880597,
      "grad_norm": 26.104557037353516,
      "learning_rate": 3.0655957161981256e-05,
      "loss": 0.2708,
      "step": 1860
    },
    {
      "epoch": 22.328358208955223,
      "grad_norm": 19.244731903076172,
      "learning_rate": 3.052208835341366e-05,
      "loss": 0.3206,
      "step": 1870
    },
    {
      "epoch": 22.44776119402985,
      "grad_norm": 12.097124099731445,
      "learning_rate": 3.0388219544846053e-05,
      "loss": 0.296,
      "step": 1880
    },
    {
      "epoch": 22.567164179104477,
      "grad_norm": 22.592735290527344,
      "learning_rate": 3.025435073627845e-05,
      "loss": 0.2638,
      "step": 1890
    },
    {
      "epoch": 22.686567164179106,
      "grad_norm": 9.944023132324219,
      "learning_rate": 3.012048192771085e-05,
      "loss": 0.3055,
      "step": 1900
    },
    {
      "epoch": 22.80597014925373,
      "grad_norm": 28.136188507080078,
      "learning_rate": 2.9986613119143242e-05,
      "loss": 0.3577,
      "step": 1910
    },
    {
      "epoch": 22.925373134328357,
      "grad_norm": 14.673059463500977,
      "learning_rate": 2.9852744310575638e-05,
      "loss": 0.2527,
      "step": 1920
    },
    {
      "epoch": 22.997014925373133,
      "eval_accuracy": 0.8772486772486773,
      "eval_loss": 0.4763975739479065,
      "eval_runtime": 21.7063,
      "eval_samples_per_second": 87.071,
      "eval_steps_per_second": 2.764,
      "step": 1926
    },
    {
      "epoch": 23.044776119402986,
      "grad_norm": 16.39044189453125,
      "learning_rate": 2.971887550200803e-05,
      "loss": 0.2727,
      "step": 1930
    },
    {
      "epoch": 23.16417910447761,
      "grad_norm": 15.90278148651123,
      "learning_rate": 2.958500669344043e-05,
      "loss": 0.3085,
      "step": 1940
    },
    {
      "epoch": 23.28358208955224,
      "grad_norm": 17.350341796875,
      "learning_rate": 2.9451137884872827e-05,
      "loss": 0.2484,
      "step": 1950
    },
    {
      "epoch": 23.402985074626866,
      "grad_norm": 25.032800674438477,
      "learning_rate": 2.931726907630522e-05,
      "loss": 0.2594,
      "step": 1960
    },
    {
      "epoch": 23.52238805970149,
      "grad_norm": 16.619842529296875,
      "learning_rate": 2.918340026773762e-05,
      "loss": 0.2444,
      "step": 1970
    },
    {
      "epoch": 23.64179104477612,
      "grad_norm": 16.149154663085938,
      "learning_rate": 2.9049531459170017e-05,
      "loss": 0.2535,
      "step": 1980
    },
    {
      "epoch": 23.761194029850746,
      "grad_norm": 14.45046615600586,
      "learning_rate": 2.891566265060241e-05,
      "loss": 0.2122,
      "step": 1990
    },
    {
      "epoch": 23.880597014925375,
      "grad_norm": 12.895037651062012,
      "learning_rate": 2.878179384203481e-05,
      "loss": 0.2739,
      "step": 2000
    },
    {
      "epoch": 24.0,
      "grad_norm": 10.731025695800781,
      "learning_rate": 2.8647925033467206e-05,
      "loss": 0.2657,
      "step": 2010
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.8767195767195767,
      "eval_loss": 0.4650660455226898,
      "eval_runtime": 21.5221,
      "eval_samples_per_second": 87.817,
      "eval_steps_per_second": 2.788,
      "step": 2010
    },
    {
      "epoch": 24.119402985074625,
      "grad_norm": 15.251605987548828,
      "learning_rate": 2.85140562248996e-05,
      "loss": 0.2214,
      "step": 2020
    },
    {
      "epoch": 24.238805970149254,
      "grad_norm": 22.7297420501709,
      "learning_rate": 2.8380187416331995e-05,
      "loss": 0.2274,
      "step": 2030
    },
    {
      "epoch": 24.35820895522388,
      "grad_norm": 15.054763793945312,
      "learning_rate": 2.8246318607764395e-05,
      "loss": 0.2559,
      "step": 2040
    },
    {
      "epoch": 24.47761194029851,
      "grad_norm": 11.38253402709961,
      "learning_rate": 2.8112449799196788e-05,
      "loss": 0.2383,
      "step": 2050
    },
    {
      "epoch": 24.597014925373134,
      "grad_norm": 10.91185188293457,
      "learning_rate": 2.7978580990629184e-05,
      "loss": 0.2419,
      "step": 2060
    },
    {
      "epoch": 24.71641791044776,
      "grad_norm": 12.653990745544434,
      "learning_rate": 2.7844712182061584e-05,
      "loss": 0.2794,
      "step": 2070
    },
    {
      "epoch": 24.83582089552239,
      "grad_norm": 16.15665626525879,
      "learning_rate": 2.7710843373493977e-05,
      "loss": 0.2554,
      "step": 2080
    },
    {
      "epoch": 24.955223880597014,
      "grad_norm": 18.719255447387695,
      "learning_rate": 2.7576974564926374e-05,
      "loss": 0.1995,
      "step": 2090
    },
    {
      "epoch": 24.991044776119402,
      "eval_accuracy": 0.8693121693121693,
      "eval_loss": 0.5079193711280823,
      "eval_runtime": 22.8272,
      "eval_samples_per_second": 82.796,
      "eval_steps_per_second": 2.628,
      "step": 2093
    },
    {
      "epoch": 25.074626865671643,
      "grad_norm": 18.27308464050293,
      "learning_rate": 2.7443105756358774e-05,
      "loss": 0.2431,
      "step": 2100
    },
    {
      "epoch": 25.19402985074627,
      "grad_norm": 13.238642692565918,
      "learning_rate": 2.7309236947791167e-05,
      "loss": 0.2215,
      "step": 2110
    },
    {
      "epoch": 25.313432835820894,
      "grad_norm": 19.07526397705078,
      "learning_rate": 2.7175368139223563e-05,
      "loss": 0.2356,
      "step": 2120
    },
    {
      "epoch": 25.432835820895523,
      "grad_norm": 23.106143951416016,
      "learning_rate": 2.7041499330655956e-05,
      "loss": 0.2428,
      "step": 2130
    },
    {
      "epoch": 25.55223880597015,
      "grad_norm": 16.487586975097656,
      "learning_rate": 2.6907630522088356e-05,
      "loss": 0.2199,
      "step": 2140
    },
    {
      "epoch": 25.671641791044777,
      "grad_norm": 18.651390075683594,
      "learning_rate": 2.6773761713520752e-05,
      "loss": 0.2386,
      "step": 2150
    },
    {
      "epoch": 25.791044776119403,
      "grad_norm": 15.018671989440918,
      "learning_rate": 2.6639892904953145e-05,
      "loss": 0.2338,
      "step": 2160
    },
    {
      "epoch": 25.91044776119403,
      "grad_norm": 15.358821868896484,
      "learning_rate": 2.6506024096385545e-05,
      "loss": 0.2481,
      "step": 2170
    },
    {
      "epoch": 25.99402985074627,
      "eval_accuracy": 0.8698412698412699,
      "eval_loss": 0.5112302899360657,
      "eval_runtime": 21.5553,
      "eval_samples_per_second": 87.682,
      "eval_steps_per_second": 2.784,
      "step": 2177
    },
    {
      "epoch": 26.029850746268657,
      "grad_norm": 15.894144058227539,
      "learning_rate": 2.637215528781794e-05,
      "loss": 0.2625,
      "step": 2180
    },
    {
      "epoch": 26.149253731343283,
      "grad_norm": 16.53984260559082,
      "learning_rate": 2.6238286479250334e-05,
      "loss": 0.2675,
      "step": 2190
    },
    {
      "epoch": 26.26865671641791,
      "grad_norm": 16.194046020507812,
      "learning_rate": 2.6104417670682734e-05,
      "loss": 0.2586,
      "step": 2200
    },
    {
      "epoch": 26.388059701492537,
      "grad_norm": 11.86839771270752,
      "learning_rate": 2.597054886211513e-05,
      "loss": 0.237,
      "step": 2210
    },
    {
      "epoch": 26.507462686567163,
      "grad_norm": 10.908110618591309,
      "learning_rate": 2.5836680053547524e-05,
      "loss": 0.2136,
      "step": 2220
    },
    {
      "epoch": 26.62686567164179,
      "grad_norm": 13.298651695251465,
      "learning_rate": 2.570281124497992e-05,
      "loss": 0.2459,
      "step": 2230
    },
    {
      "epoch": 26.746268656716417,
      "grad_norm": 13.250383377075195,
      "learning_rate": 2.556894243641232e-05,
      "loss": 0.245,
      "step": 2240
    },
    {
      "epoch": 26.865671641791046,
      "grad_norm": 11.171323776245117,
      "learning_rate": 2.5435073627844713e-05,
      "loss": 0.2943,
      "step": 2250
    },
    {
      "epoch": 26.98507462686567,
      "grad_norm": 19.93524932861328,
      "learning_rate": 2.530120481927711e-05,
      "loss": 0.2072,
      "step": 2260
    },
    {
      "epoch": 26.997014925373133,
      "eval_accuracy": 0.8830687830687831,
      "eval_loss": 0.5082184076309204,
      "eval_runtime": 21.764,
      "eval_samples_per_second": 86.841,
      "eval_steps_per_second": 2.757,
      "step": 2261
    },
    {
      "epoch": 27.104477611940297,
      "grad_norm": 9.708877563476562,
      "learning_rate": 2.516733601070951e-05,
      "loss": 0.2333,
      "step": 2270
    },
    {
      "epoch": 27.223880597014926,
      "grad_norm": 17.46422004699707,
      "learning_rate": 2.5033467202141902e-05,
      "loss": 0.2399,
      "step": 2280
    },
    {
      "epoch": 27.34328358208955,
      "grad_norm": 15.301572799682617,
      "learning_rate": 2.48995983935743e-05,
      "loss": 0.2569,
      "step": 2290
    },
    {
      "epoch": 27.46268656716418,
      "grad_norm": 10.849388122558594,
      "learning_rate": 2.4765729585006695e-05,
      "loss": 0.2426,
      "step": 2300
    },
    {
      "epoch": 27.582089552238806,
      "grad_norm": 9.713058471679688,
      "learning_rate": 2.463186077643909e-05,
      "loss": 0.2205,
      "step": 2310
    },
    {
      "epoch": 27.701492537313435,
      "grad_norm": 11.311200141906738,
      "learning_rate": 2.4497991967871488e-05,
      "loss": 0.2402,
      "step": 2320
    },
    {
      "epoch": 27.82089552238806,
      "grad_norm": 16.44293785095215,
      "learning_rate": 2.4364123159303884e-05,
      "loss": 0.2145,
      "step": 2330
    },
    {
      "epoch": 27.940298507462686,
      "grad_norm": 12.815253257751465,
      "learning_rate": 2.4230254350736277e-05,
      "loss": 0.2164,
      "step": 2340
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.873015873015873,
      "eval_loss": 0.5001527070999146,
      "eval_runtime": 24.2145,
      "eval_samples_per_second": 78.052,
      "eval_steps_per_second": 2.478,
      "step": 2345
    },
    {
      "epoch": 28.059701492537314,
      "grad_norm": 16.615896224975586,
      "learning_rate": 2.4096385542168677e-05,
      "loss": 0.2217,
      "step": 2350
    },
    {
      "epoch": 28.17910447761194,
      "grad_norm": 11.069056510925293,
      "learning_rate": 2.3962516733601073e-05,
      "loss": 0.1978,
      "step": 2360
    },
    {
      "epoch": 28.298507462686565,
      "grad_norm": 17.405532836914062,
      "learning_rate": 2.3828647925033466e-05,
      "loss": 0.2542,
      "step": 2370
    },
    {
      "epoch": 28.417910447761194,
      "grad_norm": 13.685500144958496,
      "learning_rate": 2.3694779116465866e-05,
      "loss": 0.2102,
      "step": 2380
    },
    {
      "epoch": 28.53731343283582,
      "grad_norm": 14.143411636352539,
      "learning_rate": 2.356091030789826e-05,
      "loss": 0.2151,
      "step": 2390
    },
    {
      "epoch": 28.65671641791045,
      "grad_norm": 11.186328887939453,
      "learning_rate": 2.3427041499330656e-05,
      "loss": 0.2109,
      "step": 2400
    },
    {
      "epoch": 28.776119402985074,
      "grad_norm": 11.518386840820312,
      "learning_rate": 2.3293172690763055e-05,
      "loss": 0.2041,
      "step": 2410
    },
    {
      "epoch": 28.895522388059703,
      "grad_norm": 14.913932800292969,
      "learning_rate": 2.315930388219545e-05,
      "loss": 0.2198,
      "step": 2420
    },
    {
      "epoch": 28.991044776119402,
      "eval_accuracy": 0.8777777777777778,
      "eval_loss": 0.47848638892173767,
      "eval_runtime": 23.896,
      "eval_samples_per_second": 79.093,
      "eval_steps_per_second": 2.511,
      "step": 2428
    },
    {
      "epoch": 29.01492537313433,
      "grad_norm": 13.153636932373047,
      "learning_rate": 2.3025435073627845e-05,
      "loss": 0.2036,
      "step": 2430
    },
    {
      "epoch": 29.134328358208954,
      "grad_norm": 11.16688346862793,
      "learning_rate": 2.289156626506024e-05,
      "loss": 0.1935,
      "step": 2440
    },
    {
      "epoch": 29.253731343283583,
      "grad_norm": 14.297533988952637,
      "learning_rate": 2.2757697456492638e-05,
      "loss": 0.2095,
      "step": 2450
    },
    {
      "epoch": 29.37313432835821,
      "grad_norm": 15.996660232543945,
      "learning_rate": 2.2623828647925034e-05,
      "loss": 0.1946,
      "step": 2460
    },
    {
      "epoch": 29.492537313432837,
      "grad_norm": 22.007244110107422,
      "learning_rate": 2.248995983935743e-05,
      "loss": 0.2171,
      "step": 2470
    },
    {
      "epoch": 29.611940298507463,
      "grad_norm": 19.894161224365234,
      "learning_rate": 2.2356091030789827e-05,
      "loss": 0.2542,
      "step": 2480
    },
    {
      "epoch": 29.73134328358209,
      "grad_norm": 13.56891918182373,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 0.2392,
      "step": 2490
    },
    {
      "epoch": 29.850746268656717,
      "grad_norm": 15.41037368774414,
      "learning_rate": 2.208835341365462e-05,
      "loss": 0.1737,
      "step": 2500
    },
    {
      "epoch": 29.970149253731343,
      "grad_norm": 12.46347713470459,
      "learning_rate": 2.1954484605087016e-05,
      "loss": 0.2137,
      "step": 2510
    },
    {
      "epoch": 29.99402985074627,
      "eval_accuracy": 0.8888888888888888,
      "eval_loss": 0.5012401342391968,
      "eval_runtime": 24.7826,
      "eval_samples_per_second": 76.263,
      "eval_steps_per_second": 2.421,
      "step": 2512
    },
    {
      "epoch": 30.08955223880597,
      "grad_norm": 12.577794075012207,
      "learning_rate": 2.1820615796519413e-05,
      "loss": 0.1764,
      "step": 2520
    },
    {
      "epoch": 30.208955223880597,
      "grad_norm": 10.100226402282715,
      "learning_rate": 2.168674698795181e-05,
      "loss": 0.2085,
      "step": 2530
    },
    {
      "epoch": 30.328358208955223,
      "grad_norm": 7.882339000701904,
      "learning_rate": 2.1552878179384202e-05,
      "loss": 0.217,
      "step": 2540
    },
    {
      "epoch": 30.44776119402985,
      "grad_norm": 9.06692886352539,
      "learning_rate": 2.1419009370816602e-05,
      "loss": 0.2082,
      "step": 2550
    },
    {
      "epoch": 30.567164179104477,
      "grad_norm": 24.43864631652832,
      "learning_rate": 2.1285140562248998e-05,
      "loss": 0.2089,
      "step": 2560
    },
    {
      "epoch": 30.686567164179106,
      "grad_norm": 15.762516021728516,
      "learning_rate": 2.115127175368139e-05,
      "loss": 0.176,
      "step": 2570
    },
    {
      "epoch": 30.80597014925373,
      "grad_norm": 13.477516174316406,
      "learning_rate": 2.101740294511379e-05,
      "loss": 0.2034,
      "step": 2580
    },
    {
      "epoch": 30.925373134328357,
      "grad_norm": 8.844980239868164,
      "learning_rate": 2.0883534136546184e-05,
      "loss": 0.1936,
      "step": 2590
    },
    {
      "epoch": 30.997014925373133,
      "eval_accuracy": 0.8756613756613757,
      "eval_loss": 0.4961310923099518,
      "eval_runtime": 23.3753,
      "eval_samples_per_second": 80.855,
      "eval_steps_per_second": 2.567,
      "step": 2596
    },
    {
      "epoch": 31.044776119402986,
      "grad_norm": 9.131314277648926,
      "learning_rate": 2.074966532797858e-05,
      "loss": 0.1943,
      "step": 2600
    },
    {
      "epoch": 31.16417910447761,
      "grad_norm": 15.152297019958496,
      "learning_rate": 2.061579651941098e-05,
      "loss": 0.23,
      "step": 2610
    },
    {
      "epoch": 31.28358208955224,
      "grad_norm": 13.151653289794922,
      "learning_rate": 2.0481927710843373e-05,
      "loss": 0.1769,
      "step": 2620
    },
    {
      "epoch": 31.402985074626866,
      "grad_norm": 12.334269523620605,
      "learning_rate": 2.034805890227577e-05,
      "loss": 0.1562,
      "step": 2630
    },
    {
      "epoch": 31.52238805970149,
      "grad_norm": 26.206056594848633,
      "learning_rate": 2.0214190093708166e-05,
      "loss": 0.1864,
      "step": 2640
    },
    {
      "epoch": 31.64179104477612,
      "grad_norm": 13.51135540008545,
      "learning_rate": 2.0080321285140562e-05,
      "loss": 0.1935,
      "step": 2650
    },
    {
      "epoch": 31.761194029850746,
      "grad_norm": 18.67013931274414,
      "learning_rate": 1.994645247657296e-05,
      "loss": 0.1884,
      "step": 2660
    },
    {
      "epoch": 31.880597014925375,
      "grad_norm": 14.84253978729248,
      "learning_rate": 1.9812583668005355e-05,
      "loss": 0.1825,
      "step": 2670
    },
    {
      "epoch": 32.0,
      "grad_norm": 11.984930038452148,
      "learning_rate": 1.967871485943775e-05,
      "loss": 0.2255,
      "step": 2680
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.8788359788359789,
      "eval_loss": 0.4986724257469177,
      "eval_runtime": 21.5574,
      "eval_samples_per_second": 87.673,
      "eval_steps_per_second": 2.783,
      "step": 2680
    },
    {
      "epoch": 32.11940298507463,
      "grad_norm": 8.543514251708984,
      "learning_rate": 1.9544846050870148e-05,
      "loss": 0.1831,
      "step": 2690
    },
    {
      "epoch": 32.23880597014925,
      "grad_norm": 16.21164894104004,
      "learning_rate": 1.9410977242302544e-05,
      "loss": 0.1642,
      "step": 2700
    },
    {
      "epoch": 32.35820895522388,
      "grad_norm": 32.741031646728516,
      "learning_rate": 1.927710843373494e-05,
      "loss": 0.2028,
      "step": 2710
    },
    {
      "epoch": 32.47761194029851,
      "grad_norm": 11.744430541992188,
      "learning_rate": 1.9143239625167337e-05,
      "loss": 0.182,
      "step": 2720
    },
    {
      "epoch": 32.59701492537313,
      "grad_norm": 16.407419204711914,
      "learning_rate": 1.9009370816599734e-05,
      "loss": 0.2066,
      "step": 2730
    },
    {
      "epoch": 32.71641791044776,
      "grad_norm": 13.327704429626465,
      "learning_rate": 1.8875502008032127e-05,
      "loss": 0.1831,
      "step": 2740
    },
    {
      "epoch": 32.83582089552239,
      "grad_norm": 14.911913871765137,
      "learning_rate": 1.8741633199464527e-05,
      "loss": 0.1994,
      "step": 2750
    },
    {
      "epoch": 32.95522388059702,
      "grad_norm": 11.125310897827148,
      "learning_rate": 1.8607764390896923e-05,
      "loss": 0.1818,
      "step": 2760
    },
    {
      "epoch": 32.9910447761194,
      "eval_accuracy": 0.8851851851851852,
      "eval_loss": 0.48399239778518677,
      "eval_runtime": 21.334,
      "eval_samples_per_second": 88.591,
      "eval_steps_per_second": 2.812,
      "step": 2763
    },
    {
      "epoch": 33.07462686567164,
      "grad_norm": 12.510457038879395,
      "learning_rate": 1.8473895582329316e-05,
      "loss": 0.1833,
      "step": 2770
    },
    {
      "epoch": 33.19402985074627,
      "grad_norm": 15.175230979919434,
      "learning_rate": 1.8340026773761716e-05,
      "loss": 0.1788,
      "step": 2780
    },
    {
      "epoch": 33.3134328358209,
      "grad_norm": 8.436702728271484,
      "learning_rate": 1.820615796519411e-05,
      "loss": 0.1814,
      "step": 2790
    },
    {
      "epoch": 33.43283582089552,
      "grad_norm": 11.283780097961426,
      "learning_rate": 1.8072289156626505e-05,
      "loss": 0.1644,
      "step": 2800
    },
    {
      "epoch": 33.55223880597015,
      "grad_norm": 8.91975212097168,
      "learning_rate": 1.7938420348058905e-05,
      "loss": 0.1648,
      "step": 2810
    },
    {
      "epoch": 33.67164179104478,
      "grad_norm": 9.153727531433105,
      "learning_rate": 1.7804551539491298e-05,
      "loss": 0.1749,
      "step": 2820
    },
    {
      "epoch": 33.791044776119406,
      "grad_norm": 11.47763442993164,
      "learning_rate": 1.7670682730923694e-05,
      "loss": 0.1978,
      "step": 2830
    },
    {
      "epoch": 33.91044776119403,
      "grad_norm": 15.642449378967285,
      "learning_rate": 1.753681392235609e-05,
      "loss": 0.1644,
      "step": 2840
    },
    {
      "epoch": 33.994029850746266,
      "eval_accuracy": 0.8862433862433863,
      "eval_loss": 0.46936145424842834,
      "eval_runtime": 22.5625,
      "eval_samples_per_second": 83.767,
      "eval_steps_per_second": 2.659,
      "step": 2847
    },
    {
      "epoch": 34.02985074626866,
      "grad_norm": 11.902774810791016,
      "learning_rate": 1.7402945113788487e-05,
      "loss": 0.1615,
      "step": 2850
    },
    {
      "epoch": 34.149253731343286,
      "grad_norm": 35.593807220458984,
      "learning_rate": 1.7269076305220887e-05,
      "loss": 0.1855,
      "step": 2860
    },
    {
      "epoch": 34.26865671641791,
      "grad_norm": 15.615717887878418,
      "learning_rate": 1.713520749665328e-05,
      "loss": 0.182,
      "step": 2870
    },
    {
      "epoch": 34.38805970149254,
      "grad_norm": 13.55231761932373,
      "learning_rate": 1.7001338688085676e-05,
      "loss": 0.2104,
      "step": 2880
    },
    {
      "epoch": 34.507462686567166,
      "grad_norm": 5.884557723999023,
      "learning_rate": 1.6867469879518073e-05,
      "loss": 0.1501,
      "step": 2890
    },
    {
      "epoch": 34.62686567164179,
      "grad_norm": 15.07630729675293,
      "learning_rate": 1.673360107095047e-05,
      "loss": 0.1585,
      "step": 2900
    },
    {
      "epoch": 34.74626865671642,
      "grad_norm": 9.989094734191895,
      "learning_rate": 1.6599732262382866e-05,
      "loss": 0.144,
      "step": 2910
    },
    {
      "epoch": 34.865671641791046,
      "grad_norm": 11.948739051818848,
      "learning_rate": 1.6465863453815262e-05,
      "loss": 0.1658,
      "step": 2920
    },
    {
      "epoch": 34.985074626865675,
      "grad_norm": 13.029720306396484,
      "learning_rate": 1.633199464524766e-05,
      "loss": 0.1799,
      "step": 2930
    },
    {
      "epoch": 34.997014925373136,
      "eval_accuracy": 0.8915343915343915,
      "eval_loss": 0.4599432647228241,
      "eval_runtime": 21.3881,
      "eval_samples_per_second": 88.367,
      "eval_steps_per_second": 2.805,
      "step": 2931
    },
    {
      "epoch": 35.1044776119403,
      "grad_norm": 12.619296073913574,
      "learning_rate": 1.619812583668005e-05,
      "loss": 0.1677,
      "step": 2940
    },
    {
      "epoch": 35.223880597014926,
      "grad_norm": 13.858939170837402,
      "learning_rate": 1.606425702811245e-05,
      "loss": 0.175,
      "step": 2950
    },
    {
      "epoch": 35.343283582089555,
      "grad_norm": 11.862871170043945,
      "learning_rate": 1.5930388219544848e-05,
      "loss": 0.1736,
      "step": 2960
    },
    {
      "epoch": 35.46268656716418,
      "grad_norm": 7.9211859703063965,
      "learning_rate": 1.579651941097724e-05,
      "loss": 0.1585,
      "step": 2970
    },
    {
      "epoch": 35.582089552238806,
      "grad_norm": 14.684250831604004,
      "learning_rate": 1.566265060240964e-05,
      "loss": 0.1576,
      "step": 2980
    },
    {
      "epoch": 35.701492537313435,
      "grad_norm": 8.843223571777344,
      "learning_rate": 1.5528781793842034e-05,
      "loss": 0.1594,
      "step": 2990
    },
    {
      "epoch": 35.82089552238806,
      "grad_norm": 9.077608108520508,
      "learning_rate": 1.5394912985274433e-05,
      "loss": 0.1745,
      "step": 3000
    },
    {
      "epoch": 35.940298507462686,
      "grad_norm": 14.813329696655273,
      "learning_rate": 1.526104417670683e-05,
      "loss": 0.1624,
      "step": 3010
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.8851851851851852,
      "eval_loss": 0.5122030377388,
      "eval_runtime": 22.4119,
      "eval_samples_per_second": 84.33,
      "eval_steps_per_second": 2.677,
      "step": 3015
    },
    {
      "epoch": 36.059701492537314,
      "grad_norm": 17.36454200744629,
      "learning_rate": 1.5127175368139224e-05,
      "loss": 0.1414,
      "step": 3020
    },
    {
      "epoch": 36.17910447761194,
      "grad_norm": 9.699509620666504,
      "learning_rate": 1.4993306559571621e-05,
      "loss": 0.15,
      "step": 3030
    },
    {
      "epoch": 36.298507462686565,
      "grad_norm": 8.813672065734863,
      "learning_rate": 1.4859437751004016e-05,
      "loss": 0.1924,
      "step": 3040
    },
    {
      "epoch": 36.417910447761194,
      "grad_norm": 12.653448104858398,
      "learning_rate": 1.4725568942436414e-05,
      "loss": 0.1925,
      "step": 3050
    },
    {
      "epoch": 36.53731343283582,
      "grad_norm": 9.532696723937988,
      "learning_rate": 1.459170013386881e-05,
      "loss": 0.1822,
      "step": 3060
    },
    {
      "epoch": 36.656716417910445,
      "grad_norm": 9.864669799804688,
      "learning_rate": 1.4457831325301205e-05,
      "loss": 0.1688,
      "step": 3070
    },
    {
      "epoch": 36.776119402985074,
      "grad_norm": 13.674315452575684,
      "learning_rate": 1.4323962516733603e-05,
      "loss": 0.1568,
      "step": 3080
    },
    {
      "epoch": 36.8955223880597,
      "grad_norm": 8.822213172912598,
      "learning_rate": 1.4190093708165998e-05,
      "loss": 0.157,
      "step": 3090
    },
    {
      "epoch": 36.9910447761194,
      "eval_accuracy": 0.8899470899470899,
      "eval_loss": 0.4546496570110321,
      "eval_runtime": 23.0569,
      "eval_samples_per_second": 81.971,
      "eval_steps_per_second": 2.602,
      "step": 3098
    },
    {
      "epoch": 37.014925373134325,
      "grad_norm": 13.846744537353516,
      "learning_rate": 1.4056224899598394e-05,
      "loss": 0.2011,
      "step": 3100
    },
    {
      "epoch": 37.134328358208954,
      "grad_norm": 12.753867149353027,
      "learning_rate": 1.3922356091030792e-05,
      "loss": 0.1594,
      "step": 3110
    },
    {
      "epoch": 37.25373134328358,
      "grad_norm": 9.448025703430176,
      "learning_rate": 1.3788487282463187e-05,
      "loss": 0.1583,
      "step": 3120
    },
    {
      "epoch": 37.37313432835821,
      "grad_norm": 26.652421951293945,
      "learning_rate": 1.3654618473895583e-05,
      "loss": 0.1584,
      "step": 3130
    },
    {
      "epoch": 37.492537313432834,
      "grad_norm": 13.557148933410645,
      "learning_rate": 1.3520749665327978e-05,
      "loss": 0.1753,
      "step": 3140
    },
    {
      "epoch": 37.61194029850746,
      "grad_norm": 12.321662902832031,
      "learning_rate": 1.3386880856760376e-05,
      "loss": 0.179,
      "step": 3150
    },
    {
      "epoch": 37.73134328358209,
      "grad_norm": 14.634367942810059,
      "learning_rate": 1.3253012048192772e-05,
      "loss": 0.1761,
      "step": 3160
    },
    {
      "epoch": 37.850746268656714,
      "grad_norm": 10.881927490234375,
      "learning_rate": 1.3119143239625167e-05,
      "loss": 0.1583,
      "step": 3170
    },
    {
      "epoch": 37.97014925373134,
      "grad_norm": 13.106158256530762,
      "learning_rate": 1.2985274431057565e-05,
      "loss": 0.2165,
      "step": 3180
    },
    {
      "epoch": 37.994029850746266,
      "eval_accuracy": 0.8835978835978836,
      "eval_loss": 0.5097155570983887,
      "eval_runtime": 21.5365,
      "eval_samples_per_second": 87.758,
      "eval_steps_per_second": 2.786,
      "step": 3182
    },
    {
      "epoch": 38.08955223880597,
      "grad_norm": 7.807220935821533,
      "learning_rate": 1.285140562248996e-05,
      "loss": 0.1416,
      "step": 3190
    },
    {
      "epoch": 38.208955223880594,
      "grad_norm": 9.47774887084961,
      "learning_rate": 1.2717536813922356e-05,
      "loss": 0.1422,
      "step": 3200
    },
    {
      "epoch": 38.32835820895522,
      "grad_norm": 12.726134300231934,
      "learning_rate": 1.2583668005354755e-05,
      "loss": 0.1493,
      "step": 3210
    },
    {
      "epoch": 38.44776119402985,
      "grad_norm": 10.369739532470703,
      "learning_rate": 1.244979919678715e-05,
      "loss": 0.1441,
      "step": 3220
    },
    {
      "epoch": 38.56716417910448,
      "grad_norm": 9.966578483581543,
      "learning_rate": 1.2315930388219546e-05,
      "loss": 0.1566,
      "step": 3230
    },
    {
      "epoch": 38.6865671641791,
      "grad_norm": 7.988857746124268,
      "learning_rate": 1.2182061579651942e-05,
      "loss": 0.13,
      "step": 3240
    },
    {
      "epoch": 38.80597014925373,
      "grad_norm": 6.763943195343018,
      "learning_rate": 1.2048192771084338e-05,
      "loss": 0.1633,
      "step": 3250
    },
    {
      "epoch": 38.92537313432836,
      "grad_norm": 22.221744537353516,
      "learning_rate": 1.1914323962516733e-05,
      "loss": 0.1565,
      "step": 3260
    },
    {
      "epoch": 38.997014925373136,
      "eval_accuracy": 0.8952380952380953,
      "eval_loss": 0.4565820097923279,
      "eval_runtime": 21.5334,
      "eval_samples_per_second": 87.771,
      "eval_steps_per_second": 2.786,
      "step": 3266
    },
    {
      "epoch": 39.04477611940298,
      "grad_norm": 7.209897518157959,
      "learning_rate": 1.178045515394913e-05,
      "loss": 0.1382,
      "step": 3270
    },
    {
      "epoch": 39.16417910447761,
      "grad_norm": 8.185746192932129,
      "learning_rate": 1.1646586345381528e-05,
      "loss": 0.1426,
      "step": 3280
    },
    {
      "epoch": 39.28358208955224,
      "grad_norm": 14.66981315612793,
      "learning_rate": 1.1512717536813922e-05,
      "loss": 0.1291,
      "step": 3290
    },
    {
      "epoch": 39.40298507462686,
      "grad_norm": 7.977197647094727,
      "learning_rate": 1.1378848728246319e-05,
      "loss": 0.1293,
      "step": 3300
    },
    {
      "epoch": 39.52238805970149,
      "grad_norm": 9.289730072021484,
      "learning_rate": 1.1244979919678715e-05,
      "loss": 0.1353,
      "step": 3310
    },
    {
      "epoch": 39.64179104477612,
      "grad_norm": 8.643695831298828,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 0.1558,
      "step": 3320
    },
    {
      "epoch": 39.76119402985075,
      "grad_norm": 10.772273063659668,
      "learning_rate": 1.0977242302543508e-05,
      "loss": 0.18,
      "step": 3330
    },
    {
      "epoch": 39.88059701492537,
      "grad_norm": 16.730188369750977,
      "learning_rate": 1.0843373493975904e-05,
      "loss": 0.1323,
      "step": 3340
    },
    {
      "epoch": 40.0,
      "grad_norm": 16.59427261352539,
      "learning_rate": 1.0709504685408301e-05,
      "loss": 0.1476,
      "step": 3350
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.8915343915343915,
      "eval_loss": 0.45787477493286133,
      "eval_runtime": 20.9113,
      "eval_samples_per_second": 90.382,
      "eval_steps_per_second": 2.869,
      "step": 3350
    },
    {
      "epoch": 40.11940298507463,
      "grad_norm": 13.793547630310059,
      "learning_rate": 1.0575635876840696e-05,
      "loss": 0.1335,
      "step": 3360
    },
    {
      "epoch": 40.23880597014925,
      "grad_norm": 12.108443260192871,
      "learning_rate": 1.0441767068273092e-05,
      "loss": 0.1571,
      "step": 3370
    },
    {
      "epoch": 40.35820895522388,
      "grad_norm": 6.3094024658203125,
      "learning_rate": 1.030789825970549e-05,
      "loss": 0.0985,
      "step": 3380
    },
    {
      "epoch": 40.47761194029851,
      "grad_norm": 12.814749717712402,
      "learning_rate": 1.0174029451137885e-05,
      "loss": 0.1461,
      "step": 3390
    },
    {
      "epoch": 40.59701492537313,
      "grad_norm": 4.608739852905273,
      "learning_rate": 1.0040160642570281e-05,
      "loss": 0.1116,
      "step": 3400
    },
    {
      "epoch": 40.71641791044776,
      "grad_norm": 11.582805633544922,
      "learning_rate": 9.906291834002678e-06,
      "loss": 0.1358,
      "step": 3410
    },
    {
      "epoch": 40.83582089552239,
      "grad_norm": 11.274500846862793,
      "learning_rate": 9.772423025435074e-06,
      "loss": 0.1246,
      "step": 3420
    },
    {
      "epoch": 40.95522388059702,
      "grad_norm": 10.500862121582031,
      "learning_rate": 9.63855421686747e-06,
      "loss": 0.1296,
      "step": 3430
    },
    {
      "epoch": 40.9910447761194,
      "eval_accuracy": 0.8931216931216931,
      "eval_loss": 0.4594806134700775,
      "eval_runtime": 21.5399,
      "eval_samples_per_second": 87.744,
      "eval_steps_per_second": 2.786,
      "step": 3433
    },
    {
      "epoch": 41.07462686567164,
      "grad_norm": 13.058525085449219,
      "learning_rate": 9.504685408299867e-06,
      "loss": 0.1284,
      "step": 3440
    },
    {
      "epoch": 41.19402985074627,
      "grad_norm": 11.443678855895996,
      "learning_rate": 9.370816599732263e-06,
      "loss": 0.1214,
      "step": 3450
    },
    {
      "epoch": 41.3134328358209,
      "grad_norm": 7.747429370880127,
      "learning_rate": 9.236947791164658e-06,
      "loss": 0.1001,
      "step": 3460
    },
    {
      "epoch": 41.43283582089552,
      "grad_norm": 13.356925010681152,
      "learning_rate": 9.103078982597054e-06,
      "loss": 0.1258,
      "step": 3470
    },
    {
      "epoch": 41.55223880597015,
      "grad_norm": 7.233834743499756,
      "learning_rate": 8.969210174029452e-06,
      "loss": 0.1238,
      "step": 3480
    },
    {
      "epoch": 41.67164179104478,
      "grad_norm": 9.122505187988281,
      "learning_rate": 8.835341365461847e-06,
      "loss": 0.1224,
      "step": 3490
    },
    {
      "epoch": 41.791044776119406,
      "grad_norm": 11.804553031921387,
      "learning_rate": 8.701472556894244e-06,
      "loss": 0.1173,
      "step": 3500
    },
    {
      "epoch": 41.91044776119403,
      "grad_norm": 16.831106185913086,
      "learning_rate": 8.56760374832664e-06,
      "loss": 0.1159,
      "step": 3510
    },
    {
      "epoch": 41.994029850746266,
      "eval_accuracy": 0.8883597883597883,
      "eval_loss": 0.4840790629386902,
      "eval_runtime": 21.8549,
      "eval_samples_per_second": 86.48,
      "eval_steps_per_second": 2.745,
      "step": 3517
    },
    {
      "epoch": 42.02985074626866,
      "grad_norm": 7.735755443572998,
      "learning_rate": 8.433734939759036e-06,
      "loss": 0.1211,
      "step": 3520
    },
    {
      "epoch": 42.149253731343286,
      "grad_norm": 9.165584564208984,
      "learning_rate": 8.299866131191433e-06,
      "loss": 0.1441,
      "step": 3530
    },
    {
      "epoch": 42.26865671641791,
      "grad_norm": 11.504267692565918,
      "learning_rate": 8.16599732262383e-06,
      "loss": 0.0981,
      "step": 3540
    },
    {
      "epoch": 42.38805970149254,
      "grad_norm": 13.387564659118652,
      "learning_rate": 8.032128514056226e-06,
      "loss": 0.1017,
      "step": 3550
    },
    {
      "epoch": 42.507462686567166,
      "grad_norm": 13.6415376663208,
      "learning_rate": 7.89825970548862e-06,
      "loss": 0.145,
      "step": 3560
    },
    {
      "epoch": 42.62686567164179,
      "grad_norm": 15.573339462280273,
      "learning_rate": 7.764390896921017e-06,
      "loss": 0.1396,
      "step": 3570
    },
    {
      "epoch": 42.74626865671642,
      "grad_norm": 8.934632301330566,
      "learning_rate": 7.630522088353415e-06,
      "loss": 0.129,
      "step": 3580
    },
    {
      "epoch": 42.865671641791046,
      "grad_norm": 6.53822135925293,
      "learning_rate": 7.4966532797858104e-06,
      "loss": 0.1384,
      "step": 3590
    },
    {
      "epoch": 42.985074626865675,
      "grad_norm": 14.102238655090332,
      "learning_rate": 7.362784471218207e-06,
      "loss": 0.1071,
      "step": 3600
    },
    {
      "epoch": 42.997014925373136,
      "eval_accuracy": 0.882010582010582,
      "eval_loss": 0.4729897379875183,
      "eval_runtime": 23.368,
      "eval_samples_per_second": 80.88,
      "eval_steps_per_second": 2.568,
      "step": 3601
    },
    {
      "epoch": 43.1044776119403,
      "grad_norm": 10.040472030639648,
      "learning_rate": 7.228915662650602e-06,
      "loss": 0.1247,
      "step": 3610
    },
    {
      "epoch": 43.223880597014926,
      "grad_norm": 9.124853134155273,
      "learning_rate": 7.095046854082999e-06,
      "loss": 0.1059,
      "step": 3620
    },
    {
      "epoch": 43.343283582089555,
      "grad_norm": 9.409137725830078,
      "learning_rate": 6.961178045515396e-06,
      "loss": 0.1062,
      "step": 3630
    },
    {
      "epoch": 43.46268656716418,
      "grad_norm": 12.034777641296387,
      "learning_rate": 6.827309236947792e-06,
      "loss": 0.1295,
      "step": 3640
    },
    {
      "epoch": 43.582089552238806,
      "grad_norm": 11.372798919677734,
      "learning_rate": 6.693440428380188e-06,
      "loss": 0.0976,
      "step": 3650
    },
    {
      "epoch": 43.701492537313435,
      "grad_norm": 19.926820755004883,
      "learning_rate": 6.559571619812584e-06,
      "loss": 0.1318,
      "step": 3660
    },
    {
      "epoch": 43.82089552238806,
      "grad_norm": 10.970476150512695,
      "learning_rate": 6.42570281124498e-06,
      "loss": 0.1251,
      "step": 3670
    },
    {
      "epoch": 43.940298507462686,
      "grad_norm": 12.275060653686523,
      "learning_rate": 6.291834002677377e-06,
      "loss": 0.1017,
      "step": 3680
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.8931216931216931,
      "eval_loss": 0.4470039904117584,
      "eval_runtime": 23.3521,
      "eval_samples_per_second": 80.935,
      "eval_steps_per_second": 2.569,
      "step": 3685
    },
    {
      "epoch": 44.059701492537314,
      "grad_norm": 8.43974494934082,
      "learning_rate": 6.157965194109773e-06,
      "loss": 0.1239,
      "step": 3690
    },
    {
      "epoch": 44.17910447761194,
      "grad_norm": 14.687185287475586,
      "learning_rate": 6.024096385542169e-06,
      "loss": 0.1189,
      "step": 3700
    },
    {
      "epoch": 44.298507462686565,
      "grad_norm": 9.467026710510254,
      "learning_rate": 5.890227576974565e-06,
      "loss": 0.1491,
      "step": 3710
    },
    {
      "epoch": 44.417910447761194,
      "grad_norm": 5.684982776641846,
      "learning_rate": 5.756358768406961e-06,
      "loss": 0.0982,
      "step": 3720
    },
    {
      "epoch": 44.53731343283582,
      "grad_norm": 7.686473369598389,
      "learning_rate": 5.622489959839358e-06,
      "loss": 0.1147,
      "step": 3730
    },
    {
      "epoch": 44.656716417910445,
      "grad_norm": 11.38039779663086,
      "learning_rate": 5.488621151271754e-06,
      "loss": 0.1179,
      "step": 3740
    },
    {
      "epoch": 44.776119402985074,
      "grad_norm": 6.498850345611572,
      "learning_rate": 5.3547523427041504e-06,
      "loss": 0.12,
      "step": 3750
    },
    {
      "epoch": 44.8955223880597,
      "grad_norm": 9.847461700439453,
      "learning_rate": 5.220883534136546e-06,
      "loss": 0.11,
      "step": 3760
    },
    {
      "epoch": 44.9910447761194,
      "eval_accuracy": 0.891005291005291,
      "eval_loss": 0.4557030200958252,
      "eval_runtime": 24.0952,
      "eval_samples_per_second": 78.439,
      "eval_steps_per_second": 2.49,
      "step": 3768
    },
    {
      "epoch": 45.014925373134325,
      "grad_norm": 9.032334327697754,
      "learning_rate": 5.087014725568942e-06,
      "loss": 0.1352,
      "step": 3770
    },
    {
      "epoch": 45.134328358208954,
      "grad_norm": 9.14494514465332,
      "learning_rate": 4.953145917001339e-06,
      "loss": 0.1014,
      "step": 3780
    },
    {
      "epoch": 45.25373134328358,
      "grad_norm": 8.348182678222656,
      "learning_rate": 4.819277108433735e-06,
      "loss": 0.1052,
      "step": 3790
    },
    {
      "epoch": 45.37313432835821,
      "grad_norm": 5.649065971374512,
      "learning_rate": 4.685408299866132e-06,
      "loss": 0.1138,
      "step": 3800
    },
    {
      "epoch": 45.492537313432834,
      "grad_norm": 7.422122955322266,
      "learning_rate": 4.551539491298527e-06,
      "loss": 0.1091,
      "step": 3810
    },
    {
      "epoch": 45.61194029850746,
      "grad_norm": 9.711326599121094,
      "learning_rate": 4.417670682730924e-06,
      "loss": 0.1068,
      "step": 3820
    },
    {
      "epoch": 45.73134328358209,
      "grad_norm": 6.854335784912109,
      "learning_rate": 4.28380187416332e-06,
      "loss": 0.1331,
      "step": 3830
    },
    {
      "epoch": 45.850746268656714,
      "grad_norm": 7.027961254119873,
      "learning_rate": 4.149933065595716e-06,
      "loss": 0.0979,
      "step": 3840
    },
    {
      "epoch": 45.97014925373134,
      "grad_norm": 8.980549812316895,
      "learning_rate": 4.016064257028113e-06,
      "loss": 0.126,
      "step": 3850
    },
    {
      "epoch": 45.994029850746266,
      "eval_accuracy": 0.8925925925925926,
      "eval_loss": 0.4585404098033905,
      "eval_runtime": 23.6784,
      "eval_samples_per_second": 79.82,
      "eval_steps_per_second": 2.534,
      "step": 3852
    },
    {
      "epoch": 46.08955223880597,
      "grad_norm": 6.251053810119629,
      "learning_rate": 3.882195448460508e-06,
      "loss": 0.1199,
      "step": 3860
    },
    {
      "epoch": 46.208955223880594,
      "grad_norm": 9.849762916564941,
      "learning_rate": 3.7483266398929052e-06,
      "loss": 0.1068,
      "step": 3870
    },
    {
      "epoch": 46.32835820895522,
      "grad_norm": 4.6796722412109375,
      "learning_rate": 3.614457831325301e-06,
      "loss": 0.0909,
      "step": 3880
    },
    {
      "epoch": 46.44776119402985,
      "grad_norm": 7.510217666625977,
      "learning_rate": 3.480589022757698e-06,
      "loss": 0.1058,
      "step": 3890
    },
    {
      "epoch": 46.56716417910448,
      "grad_norm": 10.298948287963867,
      "learning_rate": 3.346720214190094e-06,
      "loss": 0.1045,
      "step": 3900
    },
    {
      "epoch": 46.6865671641791,
      "grad_norm": 10.80692195892334,
      "learning_rate": 3.21285140562249e-06,
      "loss": 0.1051,
      "step": 3910
    },
    {
      "epoch": 46.80597014925373,
      "grad_norm": 7.66585111618042,
      "learning_rate": 3.0789825970548864e-06,
      "loss": 0.1152,
      "step": 3920
    },
    {
      "epoch": 46.92537313432836,
      "grad_norm": 12.45518970489502,
      "learning_rate": 2.9451137884872824e-06,
      "loss": 0.1079,
      "step": 3930
    },
    {
      "epoch": 46.997014925373136,
      "eval_accuracy": 0.8904761904761904,
      "eval_loss": 0.4550909698009491,
      "eval_runtime": 22.5814,
      "eval_samples_per_second": 83.697,
      "eval_steps_per_second": 2.657,
      "step": 3936
    },
    {
      "epoch": 47.04477611940298,
      "grad_norm": 8.93836498260498,
      "learning_rate": 2.811244979919679e-06,
      "loss": 0.0922,
      "step": 3940
    },
    {
      "epoch": 47.16417910447761,
      "grad_norm": 9.170904159545898,
      "learning_rate": 2.6773761713520752e-06,
      "loss": 0.1062,
      "step": 3950
    },
    {
      "epoch": 47.28358208955224,
      "grad_norm": 8.85617446899414,
      "learning_rate": 2.543507362784471e-06,
      "loss": 0.1051,
      "step": 3960
    },
    {
      "epoch": 47.40298507462686,
      "grad_norm": 12.655815124511719,
      "learning_rate": 2.4096385542168676e-06,
      "loss": 0.1178,
      "step": 3970
    },
    {
      "epoch": 47.52238805970149,
      "grad_norm": 7.482461929321289,
      "learning_rate": 2.2757697456492636e-06,
      "loss": 0.1011,
      "step": 3980
    },
    {
      "epoch": 47.64179104477612,
      "grad_norm": 9.295207023620605,
      "learning_rate": 2.14190093708166e-06,
      "loss": 0.1175,
      "step": 3990
    },
    {
      "epoch": 47.76119402985075,
      "grad_norm": 8.872883796691895,
      "learning_rate": 2.0080321285140564e-06,
      "loss": 0.1046,
      "step": 4000
    },
    {
      "epoch": 47.88059701492537,
      "grad_norm": 11.744367599487305,
      "learning_rate": 1.8741633199464526e-06,
      "loss": 0.0891,
      "step": 4010
    },
    {
      "epoch": 48.0,
      "grad_norm": 14.95531177520752,
      "learning_rate": 1.740294511378849e-06,
      "loss": 0.1194,
      "step": 4020
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.8947089947089947,
      "eval_loss": 0.44007420539855957,
      "eval_runtime": 21.2726,
      "eval_samples_per_second": 88.847,
      "eval_steps_per_second": 2.821,
      "step": 4020
    },
    {
      "epoch": 48.11940298507463,
      "grad_norm": 12.020395278930664,
      "learning_rate": 1.606425702811245e-06,
      "loss": 0.1122,
      "step": 4030
    },
    {
      "epoch": 48.23880597014925,
      "grad_norm": 10.32168197631836,
      "learning_rate": 1.4725568942436412e-06,
      "loss": 0.1076,
      "step": 4040
    },
    {
      "epoch": 48.35820895522388,
      "grad_norm": 12.109696388244629,
      "learning_rate": 1.3386880856760376e-06,
      "loss": 0.1145,
      "step": 4050
    },
    {
      "epoch": 48.47761194029851,
      "grad_norm": 5.61909294128418,
      "learning_rate": 1.2048192771084338e-06,
      "loss": 0.0886,
      "step": 4060
    },
    {
      "epoch": 48.59701492537313,
      "grad_norm": 5.84828519821167,
      "learning_rate": 1.07095046854083e-06,
      "loss": 0.088,
      "step": 4070
    },
    {
      "epoch": 48.71641791044776,
      "grad_norm": 7.48928165435791,
      "learning_rate": 9.370816599732263e-07,
      "loss": 0.0943,
      "step": 4080
    },
    {
      "epoch": 48.83582089552239,
      "grad_norm": 6.499556541442871,
      "learning_rate": 8.032128514056225e-07,
      "loss": 0.104,
      "step": 4090
    },
    {
      "epoch": 48.95522388059702,
      "grad_norm": 8.104216575622559,
      "learning_rate": 6.693440428380188e-07,
      "loss": 0.11,
      "step": 4100
    },
    {
      "epoch": 48.9910447761194,
      "eval_accuracy": 0.8968253968253969,
      "eval_loss": 0.44243764877319336,
      "eval_runtime": 21.6947,
      "eval_samples_per_second": 87.118,
      "eval_steps_per_second": 2.766,
      "step": 4103
    },
    {
      "epoch": 49.07462686567164,
      "grad_norm": 5.112276077270508,
      "learning_rate": 5.35475234270415e-07,
      "loss": 0.0958,
      "step": 4110
    },
    {
      "epoch": 49.19402985074627,
      "grad_norm": 9.789267539978027,
      "learning_rate": 4.0160642570281125e-07,
      "loss": 0.1036,
      "step": 4120
    },
    {
      "epoch": 49.3134328358209,
      "grad_norm": 7.143287658691406,
      "learning_rate": 2.677376171352075e-07,
      "loss": 0.1019,
      "step": 4130
    },
    {
      "epoch": 49.43283582089552,
      "grad_norm": 6.7176384925842285,
      "learning_rate": 1.3386880856760375e-07,
      "loss": 0.1089,
      "step": 4140
    },
    {
      "epoch": 49.55223880597015,
      "grad_norm": 11.27464771270752,
      "learning_rate": 0.0,
      "loss": 0.1104,
      "step": 4150
    },
    {
      "epoch": 49.55223880597015,
      "eval_accuracy": 0.8957671957671958,
      "eval_loss": 0.44136127829551697,
      "eval_runtime": 22.143,
      "eval_samples_per_second": 85.354,
      "eval_steps_per_second": 2.71,
      "step": 4150
    },
    {
      "epoch": 49.55223880597015,
      "step": 4150,
      "total_flos": 5.421822883117793e+19,
      "train_loss": 0.33880508939903903,
      "train_runtime": 14225.4065,
      "train_samples_per_second": 37.644,
      "train_steps_per_second": 0.292
    }
  ],
  "logging_steps": 10,
  "max_steps": 4150,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 5.421822883117793e+19,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}