{
  "best_global_step": 2470,
  "best_metric": 0.00025587898562662303,
  "best_model_checkpoint": "./beans_outputs/checkpoint-2470",
  "epoch": 50.0,
  "eval_steps": 500,
  "global_step": 6500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07692307692307693,
      "grad_norm": 28.160919189453125,
      "learning_rate": 1.9972307692307693e-05,
      "loss": 0.9067,
      "step": 10
    },
    {
      "epoch": 0.15384615384615385,
      "grad_norm": 13.613167762756348,
      "learning_rate": 1.9941538461538464e-05,
      "loss": 0.5139,
      "step": 20
    },
    {
      "epoch": 0.23076923076923078,
      "grad_norm": 10.76201057434082,
      "learning_rate": 1.9910769230769232e-05,
      "loss": 0.2543,
      "step": 30
    },
    {
      "epoch": 0.3076923076923077,
      "grad_norm": 44.93894958496094,
      "learning_rate": 1.9880000000000003e-05,
      "loss": 0.5313,
      "step": 40
    },
    {
      "epoch": 0.38461538461538464,
      "grad_norm": 25.037395477294922,
      "learning_rate": 1.984923076923077e-05,
      "loss": 0.2034,
      "step": 50
    },
    {
      "epoch": 0.46153846153846156,
      "grad_norm": 40.23418045043945,
      "learning_rate": 1.9818461538461538e-05,
      "loss": 0.2249,
      "step": 60
    },
    {
      "epoch": 0.5384615384615384,
      "grad_norm": 26.893648147583008,
      "learning_rate": 1.978769230769231e-05,
      "loss": 0.3337,
      "step": 70
    },
    {
      "epoch": 0.6153846153846154,
      "grad_norm": 0.18680469691753387,
      "learning_rate": 1.9756923076923077e-05,
      "loss": 0.0313,
      "step": 80
    },
    {
      "epoch": 0.6923076923076923,
      "grad_norm": 0.3512158691883087,
      "learning_rate": 1.9726153846153848e-05,
      "loss": 0.088,
      "step": 90
    },
    {
      "epoch": 0.7692307692307693,
      "grad_norm": 26.937734603881836,
      "learning_rate": 1.9695384615384616e-05,
      "loss": 0.2438,
      "step": 100
    },
    {
      "epoch": 0.8461538461538461,
      "grad_norm": 37.74310302734375,
      "learning_rate": 1.9664615384615387e-05,
      "loss": 0.3817,
      "step": 110
    },
    {
      "epoch": 0.9230769230769231,
      "grad_norm": 11.974236488342285,
      "learning_rate": 1.9633846153846155e-05,
      "loss": 0.0529,
      "step": 120
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.7709632515907288,
      "learning_rate": 1.9603076923076926e-05,
      "loss": 0.0435,
      "step": 130
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9624060150375939,
      "eval_loss": 0.11930900067090988,
      "eval_runtime": 0.5873,
      "eval_samples_per_second": 226.46,
      "eval_steps_per_second": 28.946,
      "step": 130
    },
    {
      "epoch": 1.0769230769230769,
      "grad_norm": 0.0285137090831995,
      "learning_rate": 1.9572307692307693e-05,
      "loss": 0.1304,
      "step": 140
    },
    {
      "epoch": 1.1538461538461537,
      "grad_norm": 54.899940490722656,
      "learning_rate": 1.9541538461538464e-05,
      "loss": 0.1374,
      "step": 150
    },
    {
      "epoch": 1.2307692307692308,
      "grad_norm": 0.06680646538734436,
      "learning_rate": 1.9510769230769232e-05,
      "loss": 0.2245,
      "step": 160
    },
    {
      "epoch": 1.3076923076923077,
      "grad_norm": 80.54553985595703,
      "learning_rate": 1.948e-05,
      "loss": 0.1663,
      "step": 170
    },
    {
      "epoch": 1.3846153846153846,
      "grad_norm": 41.55508041381836,
      "learning_rate": 1.944923076923077e-05,
      "loss": 0.1449,
      "step": 180
    },
    {
      "epoch": 1.4615384615384617,
      "grad_norm": 0.04248078912496567,
      "learning_rate": 1.941846153846154e-05,
      "loss": 0.2905,
      "step": 190
    },
    {
      "epoch": 1.5384615384615383,
      "grad_norm": 0.199375182390213,
      "learning_rate": 1.938769230769231e-05,
      "loss": 0.3444,
      "step": 200
    },
    {
      "epoch": 1.6153846153846154,
      "grad_norm": 1.8903758525848389,
      "learning_rate": 1.9356923076923077e-05,
      "loss": 0.0755,
      "step": 210
    },
    {
      "epoch": 1.6923076923076923,
      "grad_norm": 0.1028146967291832,
      "learning_rate": 1.932615384615385e-05,
      "loss": 0.3268,
      "step": 220
    },
    {
      "epoch": 1.7692307692307692,
      "grad_norm": 7.126160621643066,
      "learning_rate": 1.929538461538462e-05,
      "loss": 0.3276,
      "step": 230
    },
    {
      "epoch": 1.8461538461538463,
      "grad_norm": 79.28353881835938,
      "learning_rate": 1.9264615384615387e-05,
      "loss": 0.2051,
      "step": 240
    },
    {
      "epoch": 1.9230769230769231,
      "grad_norm": 1.5293940305709839,
      "learning_rate": 1.9233846153846155e-05,
      "loss": 0.1562,
      "step": 250
    },
    {
      "epoch": 2.0,
      "grad_norm": 5.852114677429199,
      "learning_rate": 1.9203076923076923e-05,
      "loss": 0.1536,
      "step": 260
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.0022762541193515062,
      "eval_runtime": 0.5666,
      "eval_samples_per_second": 234.733,
      "eval_steps_per_second": 30.004,
      "step": 260
    },
    {
      "epoch": 2.076923076923077,
      "grad_norm": 62.74606704711914,
      "learning_rate": 1.9172307692307694e-05,
      "loss": 0.1537,
      "step": 270
    },
    {
      "epoch": 2.1538461538461537,
      "grad_norm": 0.01672833040356636,
      "learning_rate": 1.914153846153846e-05,
      "loss": 0.3,
      "step": 280
    },
    {
      "epoch": 2.230769230769231,
      "grad_norm": 0.7310534119606018,
      "learning_rate": 1.9110769230769233e-05,
      "loss": 0.1279,
      "step": 290
    },
    {
      "epoch": 2.3076923076923075,
      "grad_norm": 6.83922004699707,
      "learning_rate": 1.908e-05,
      "loss": 0.0137,
      "step": 300
    },
    {
      "epoch": 2.3846153846153846,
      "grad_norm": 39.29887771606445,
      "learning_rate": 1.904923076923077e-05,
      "loss": 0.2474,
      "step": 310
    },
    {
      "epoch": 2.4615384615384617,
      "grad_norm": 61.3380012512207,
      "learning_rate": 1.901846153846154e-05,
      "loss": 0.2451,
      "step": 320
    },
    {
      "epoch": 2.5384615384615383,
      "grad_norm": 0.08102953433990479,
      "learning_rate": 1.898769230769231e-05,
      "loss": 0.0788,
      "step": 330
    },
    {
      "epoch": 2.6153846153846154,
      "grad_norm": 0.01904252916574478,
      "learning_rate": 1.8956923076923078e-05,
      "loss": 0.0667,
      "step": 340
    },
    {
      "epoch": 2.6923076923076925,
      "grad_norm": 0.021739983931183815,
      "learning_rate": 1.892615384615385e-05,
      "loss": 0.2218,
      "step": 350
    },
    {
      "epoch": 2.769230769230769,
      "grad_norm": 65.27315521240234,
      "learning_rate": 1.8895384615384617e-05,
      "loss": 0.348,
      "step": 360
    },
    {
      "epoch": 2.8461538461538463,
      "grad_norm": 5.799041271209717,
      "learning_rate": 1.8864615384615384e-05,
      "loss": 0.213,
      "step": 370
    },
    {
      "epoch": 2.9230769230769234,
      "grad_norm": 29.148353576660156,
      "learning_rate": 1.8833846153846155e-05,
      "loss": 0.0291,
      "step": 380
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.01673489809036255,
      "learning_rate": 1.8803076923076923e-05,
      "loss": 0.183,
      "step": 390
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.001464208704419434,
      "eval_runtime": 0.7345,
      "eval_samples_per_second": 181.068,
      "eval_steps_per_second": 23.144,
      "step": 390
    },
    {
      "epoch": 3.076923076923077,
      "grad_norm": 0.16125129163265228,
      "learning_rate": 1.8772307692307694e-05,
      "loss": 0.0953,
      "step": 400
    },
    {
      "epoch": 3.1538461538461537,
      "grad_norm": 1.255500316619873,
      "learning_rate": 1.8741538461538462e-05,
      "loss": 0.0718,
      "step": 410
    },
    {
      "epoch": 3.230769230769231,
      "grad_norm": 0.10942380875349045,
      "learning_rate": 1.8710769230769233e-05,
      "loss": 0.2385,
      "step": 420
    },
    {
      "epoch": 3.3076923076923075,
      "grad_norm": 0.38341224193573,
      "learning_rate": 1.8680000000000004e-05,
      "loss": 0.1585,
      "step": 430
    },
    {
      "epoch": 3.3846153846153846,
      "grad_norm": 5.892263412475586,
      "learning_rate": 1.8649230769230772e-05,
      "loss": 0.3054,
      "step": 440
    },
    {
      "epoch": 3.4615384615384617,
      "grad_norm": 90.24942016601562,
      "learning_rate": 1.861846153846154e-05,
      "loss": 0.1194,
      "step": 450
    },
    {
      "epoch": 3.5384615384615383,
      "grad_norm": 0.43507635593414307,
      "learning_rate": 1.8587692307692307e-05,
      "loss": 0.0827,
      "step": 460
    },
    {
      "epoch": 3.6153846153846154,
      "grad_norm": 14.393282890319824,
      "learning_rate": 1.8556923076923078e-05,
      "loss": 0.0025,
      "step": 470
    },
    {
      "epoch": 3.6923076923076925,
      "grad_norm": 0.07155368477106094,
      "learning_rate": 1.8526153846153846e-05,
      "loss": 0.0185,
      "step": 480
    },
    {
      "epoch": 3.769230769230769,
      "grad_norm": 0.0020313099958002567,
      "learning_rate": 1.8495384615384617e-05,
      "loss": 0.0582,
      "step": 490
    },
    {
      "epoch": 3.8461538461538463,
      "grad_norm": 21.025943756103516,
      "learning_rate": 1.8464615384615385e-05,
      "loss": 0.1224,
      "step": 500
    },
    {
      "epoch": 3.9230769230769234,
      "grad_norm": 33.07258224487305,
      "learning_rate": 1.8433846153846156e-05,
      "loss": 0.0161,
      "step": 510
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.0024726390838623047,
      "learning_rate": 1.8403076923076924e-05,
      "loss": 0.2256,
      "step": 520
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.038613125681877136,
      "eval_runtime": 0.5587,
      "eval_samples_per_second": 238.037,
      "eval_steps_per_second": 30.426,
      "step": 520
    },
    {
      "epoch": 4.076923076923077,
      "grad_norm": 112.0116195678711,
      "learning_rate": 1.8372307692307695e-05,
      "loss": 0.1838,
      "step": 530
    },
    {
      "epoch": 4.153846153846154,
      "grad_norm": 0.09052282571792603,
      "learning_rate": 1.8341538461538462e-05,
      "loss": 0.1015,
      "step": 540
    },
    {
      "epoch": 4.230769230769231,
      "grad_norm": 0.09375716745853424,
      "learning_rate": 1.8310769230769233e-05,
      "loss": 0.1176,
      "step": 550
    },
    {
      "epoch": 4.3076923076923075,
      "grad_norm": 101.64923095703125,
      "learning_rate": 1.828e-05,
      "loss": 0.1395,
      "step": 560
    },
    {
      "epoch": 4.384615384615385,
      "grad_norm": 0.0043612755835056305,
      "learning_rate": 1.824923076923077e-05,
      "loss": 0.1515,
      "step": 570
    },
    {
      "epoch": 4.461538461538462,
      "grad_norm": 9.888139724731445,
      "learning_rate": 1.821846153846154e-05,
      "loss": 0.2349,
      "step": 580
    },
    {
      "epoch": 4.538461538461538,
      "grad_norm": 59.99452209472656,
      "learning_rate": 1.8187692307692308e-05,
      "loss": 0.134,
      "step": 590
    },
    {
      "epoch": 4.615384615384615,
      "grad_norm": 42.51639938354492,
      "learning_rate": 1.815692307692308e-05,
      "loss": 0.3639,
      "step": 600
    },
    {
      "epoch": 4.6923076923076925,
      "grad_norm": 44.57368087768555,
      "learning_rate": 1.8126153846153846e-05,
      "loss": 0.2551,
      "step": 610
    },
    {
      "epoch": 4.769230769230769,
      "grad_norm": 9.872251510620117,
      "learning_rate": 1.8095384615384618e-05,
      "loss": 0.2064,
      "step": 620
    },
    {
      "epoch": 4.846153846153846,
      "grad_norm": 79.4139404296875,
      "learning_rate": 1.806461538461539e-05,
      "loss": 0.2031,
      "step": 630
    },
    {
      "epoch": 4.923076923076923,
      "grad_norm": 0.005071394145488739,
      "learning_rate": 1.8033846153846156e-05,
      "loss": 0.0281,
      "step": 640
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.024555927142500877,
      "learning_rate": 1.8003076923076924e-05,
      "loss": 0.0555,
      "step": 650
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.03397071361541748,
      "eval_runtime": 0.5652,
      "eval_samples_per_second": 235.305,
      "eval_steps_per_second": 30.077,
      "step": 650
    },
    {
      "epoch": 5.076923076923077,
      "grad_norm": 0.2948272228240967,
      "learning_rate": 1.7972307692307692e-05,
      "loss": 0.1079,
      "step": 660
    },
    {
      "epoch": 5.153846153846154,
      "grad_norm": 0.010563166812062263,
      "learning_rate": 1.7941538461538463e-05,
      "loss": 0.062,
      "step": 670
    },
    {
      "epoch": 5.230769230769231,
      "grad_norm": 0.1277681589126587,
      "learning_rate": 1.791076923076923e-05,
      "loss": 0.5324,
      "step": 680
    },
    {
      "epoch": 5.3076923076923075,
      "grad_norm": 18.837158203125,
      "learning_rate": 1.788e-05,
      "loss": 0.0799,
      "step": 690
    },
    {
      "epoch": 5.384615384615385,
      "grad_norm": 0.03636413440108299,
      "learning_rate": 1.784923076923077e-05,
      "loss": 0.1586,
      "step": 700
    },
    {
      "epoch": 5.461538461538462,
      "grad_norm": 0.7668882012367249,
      "learning_rate": 1.781846153846154e-05,
      "loss": 0.2337,
      "step": 710
    },
    {
      "epoch": 5.538461538461538,
      "grad_norm": 15.153043746948242,
      "learning_rate": 1.778769230769231e-05,
      "loss": 0.1419,
      "step": 720
    },
    {
      "epoch": 5.615384615384615,
      "grad_norm": 0.12611930072307587,
      "learning_rate": 1.775692307692308e-05,
      "loss": 0.1464,
      "step": 730
    },
    {
      "epoch": 5.6923076923076925,
      "grad_norm": 31.0040225982666,
      "learning_rate": 1.7726153846153847e-05,
      "loss": 0.1024,
      "step": 740
    },
    {
      "epoch": 5.769230769230769,
      "grad_norm": 0.11015983670949936,
      "learning_rate": 1.7695384615384618e-05,
      "loss": 0.1542,
      "step": 750
    },
    {
      "epoch": 5.846153846153846,
      "grad_norm": 0.1276661902666092,
      "learning_rate": 1.7664615384615386e-05,
      "loss": 0.0896,
      "step": 760
    },
    {
      "epoch": 5.923076923076923,
      "grad_norm": 23.176246643066406,
      "learning_rate": 1.7633846153846153e-05,
      "loss": 0.2576,
      "step": 770
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.016006305813789368,
      "learning_rate": 1.7603076923076924e-05,
      "loss": 0.0713,
      "step": 780
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.07276736199855804,
      "eval_runtime": 0.6348,
      "eval_samples_per_second": 209.524,
      "eval_steps_per_second": 26.781,
      "step": 780
    },
    {
      "epoch": 6.076923076923077,
      "grad_norm": 0.00783840287476778,
      "learning_rate": 1.7572307692307692e-05,
      "loss": 0.1411,
      "step": 790
    },
    {
      "epoch": 6.153846153846154,
      "grad_norm": 43.413272857666016,
      "learning_rate": 1.7541538461538463e-05,
      "loss": 0.1076,
      "step": 800
    },
    {
      "epoch": 6.230769230769231,
      "grad_norm": 1.725425362586975,
      "learning_rate": 1.751076923076923e-05,
      "loss": 0.2663,
      "step": 810
    },
    {
      "epoch": 6.3076923076923075,
      "grad_norm": 0.03434762358665466,
      "learning_rate": 1.7480000000000002e-05,
      "loss": 0.0053,
      "step": 820
    },
    {
      "epoch": 6.384615384615385,
      "grad_norm": 0.030778298154473305,
      "learning_rate": 1.7449230769230773e-05,
      "loss": 0.0819,
      "step": 830
    },
    {
      "epoch": 6.461538461538462,
      "grad_norm": 0.064858078956604,
      "learning_rate": 1.741846153846154e-05,
      "loss": 0.0463,
      "step": 840
    },
    {
      "epoch": 6.538461538461538,
      "grad_norm": 10.116390228271484,
      "learning_rate": 1.738769230769231e-05,
      "loss": 0.4003,
      "step": 850
    },
    {
      "epoch": 6.615384615384615,
      "grad_norm": 53.98569107055664,
      "learning_rate": 1.7356923076923076e-05,
      "loss": 0.3175,
      "step": 860
    },
    {
      "epoch": 6.6923076923076925,
      "grad_norm": 9.284001350402832,
      "learning_rate": 1.7326153846153847e-05,
      "loss": 0.2402,
      "step": 870
    },
    {
      "epoch": 6.769230769230769,
      "grad_norm": 0.22148916125297546,
      "learning_rate": 1.7295384615384615e-05,
      "loss": 0.0054,
      "step": 880
    },
    {
      "epoch": 6.846153846153846,
      "grad_norm": 0.05801761895418167,
      "learning_rate": 1.7264615384615386e-05,
      "loss": 0.0524,
      "step": 890
    },
    {
      "epoch": 6.923076923076923,
      "grad_norm": 159.55609130859375,
      "learning_rate": 1.7233846153846154e-05,
      "loss": 0.178,
      "step": 900
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.6641747951507568,
      "learning_rate": 1.7203076923076925e-05,
      "loss": 0.0082,
      "step": 910
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.04108954966068268,
      "eval_runtime": 0.5705,
      "eval_samples_per_second": 233.148,
      "eval_steps_per_second": 29.801,
      "step": 910
    },
    {
      "epoch": 7.076923076923077,
      "grad_norm": 23.453298568725586,
      "learning_rate": 1.7172307692307696e-05,
      "loss": 0.3437,
      "step": 920
    },
    {
      "epoch": 7.153846153846154,
      "grad_norm": 0.03930990770459175,
      "learning_rate": 1.7141538461538464e-05,
      "loss": 0.2261,
      "step": 930
    },
    {
      "epoch": 7.230769230769231,
      "grad_norm": 20.72200584411621,
      "learning_rate": 1.711076923076923e-05,
      "loss": 0.1309,
      "step": 940
    },
    {
      "epoch": 7.3076923076923075,
      "grad_norm": 39.58359909057617,
      "learning_rate": 1.7080000000000002e-05,
      "loss": 0.1134,
      "step": 950
    },
    {
      "epoch": 7.384615384615385,
      "grad_norm": 1.5824700593948364,
      "learning_rate": 1.704923076923077e-05,
      "loss": 0.0574,
      "step": 960
    },
    {
      "epoch": 7.461538461538462,
      "grad_norm": 22.621976852416992,
      "learning_rate": 1.7018461538461538e-05,
      "loss": 0.1681,
      "step": 970
    },
    {
      "epoch": 7.538461538461538,
      "grad_norm": 16.806806564331055,
      "learning_rate": 1.698769230769231e-05,
      "loss": 0.1348,
      "step": 980
    },
    {
      "epoch": 7.615384615384615,
      "grad_norm": 15.27640151977539,
      "learning_rate": 1.6956923076923077e-05,
      "loss": 0.1698,
      "step": 990
    },
    {
      "epoch": 7.6923076923076925,
      "grad_norm": 39.200557708740234,
      "learning_rate": 1.6926153846153848e-05,
      "loss": 0.0372,
      "step": 1000
    },
    {
      "epoch": 7.769230769230769,
      "grad_norm": 2.748015880584717,
      "learning_rate": 1.6895384615384615e-05,
      "loss": 0.0288,
      "step": 1010
    },
    {
      "epoch": 7.846153846153846,
      "grad_norm": 15.786018371582031,
      "learning_rate": 1.6864615384615387e-05,
      "loss": 0.252,
      "step": 1020
    },
    {
      "epoch": 7.923076923076923,
      "grad_norm": 0.021507665514945984,
      "learning_rate": 1.6833846153846158e-05,
      "loss": 0.0895,
      "step": 1030
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.04156835377216339,
      "learning_rate": 1.6803076923076925e-05,
      "loss": 0.0085,
      "step": 1040
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.10019110888242722,
      "eval_runtime": 0.5783,
      "eval_samples_per_second": 229.992,
      "eval_steps_per_second": 29.397,
      "step": 1040
    },
    {
      "epoch": 8.076923076923077,
      "grad_norm": 0.013579009100794792,
      "learning_rate": 1.6772307692307693e-05,
      "loss": 0.1722,
      "step": 1050
    },
    {
      "epoch": 8.153846153846153,
      "grad_norm": 0.0050528221763670444,
      "learning_rate": 1.674153846153846e-05,
      "loss": 0.0511,
      "step": 1060
    },
    {
      "epoch": 8.23076923076923,
      "grad_norm": 0.015883196145296097,
      "learning_rate": 1.6710769230769232e-05,
      "loss": 0.134,
      "step": 1070
    },
    {
      "epoch": 8.307692307692308,
      "grad_norm": 0.1903264969587326,
      "learning_rate": 1.668e-05,
      "loss": 0.0889,
      "step": 1080
    },
    {
      "epoch": 8.384615384615385,
      "grad_norm": 0.02674732729792595,
      "learning_rate": 1.664923076923077e-05,
      "loss": 0.0728,
      "step": 1090
    },
    {
      "epoch": 8.461538461538462,
      "grad_norm": 0.04318476840853691,
      "learning_rate": 1.661846153846154e-05,
      "loss": 0.1575,
      "step": 1100
    },
    {
      "epoch": 8.538461538461538,
      "grad_norm": 0.01189767848700285,
      "learning_rate": 1.658769230769231e-05,
      "loss": 0.0308,
      "step": 1110
    },
    {
      "epoch": 8.615384615384615,
      "grad_norm": 0.1701788455247879,
      "learning_rate": 1.655692307692308e-05,
      "loss": 0.1881,
      "step": 1120
    },
    {
      "epoch": 8.692307692307692,
      "grad_norm": 4.718733787536621,
      "learning_rate": 1.6526153846153848e-05,
      "loss": 0.0678,
      "step": 1130
    },
    {
      "epoch": 8.76923076923077,
      "grad_norm": 0.16493193805217743,
      "learning_rate": 1.6495384615384616e-05,
      "loss": 0.172,
      "step": 1140
    },
    {
      "epoch": 8.846153846153847,
      "grad_norm": 0.013741947710514069,
      "learning_rate": 1.6464615384615387e-05,
      "loss": 0.1544,
      "step": 1150
    },
    {
      "epoch": 8.923076923076923,
      "grad_norm": 0.008533033542335033,
      "learning_rate": 1.6433846153846155e-05,
      "loss": 0.0382,
      "step": 1160
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.01939828135073185,
      "learning_rate": 1.6403076923076922e-05,
      "loss": 0.0733,
      "step": 1170
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.00041508162394165993,
      "eval_runtime": 0.5615,
      "eval_samples_per_second": 236.857,
      "eval_steps_per_second": 30.275,
      "step": 1170
    },
    {
      "epoch": 9.076923076923077,
      "grad_norm": 0.029882851988077164,
      "learning_rate": 1.6372307692307693e-05,
      "loss": 0.1645,
      "step": 1180
    },
    {
      "epoch": 9.153846153846153,
      "grad_norm": 0.02415158972144127,
      "learning_rate": 1.634153846153846e-05,
      "loss": 0.2324,
      "step": 1190
    },
    {
      "epoch": 9.23076923076923,
      "grad_norm": 0.011508429422974586,
      "learning_rate": 1.6310769230769232e-05,
      "loss": 0.0056,
      "step": 1200
    },
    {
      "epoch": 9.307692307692308,
      "grad_norm": 0.49797359108924866,
      "learning_rate": 1.628e-05,
      "loss": 0.0653,
      "step": 1210
    },
    {
      "epoch": 9.384615384615385,
      "grad_norm": 0.024000057950615883,
      "learning_rate": 1.624923076923077e-05,
      "loss": 0.0544,
      "step": 1220
    },
    {
      "epoch": 9.461538461538462,
      "grad_norm": 0.04410143569111824,
      "learning_rate": 1.6218461538461542e-05,
      "loss": 0.1128,
      "step": 1230
    },
    {
      "epoch": 9.538461538461538,
      "grad_norm": 0.10039117932319641,
      "learning_rate": 1.618769230769231e-05,
      "loss": 0.0244,
      "step": 1240
    },
    {
      "epoch": 9.615384615384615,
      "grad_norm": 0.4355177879333496,
      "learning_rate": 1.6156923076923078e-05,
      "loss": 0.1553,
      "step": 1250
    },
    {
      "epoch": 9.692307692307692,
      "grad_norm": 0.01548450905829668,
      "learning_rate": 1.6126153846153845e-05,
      "loss": 0.0277,
      "step": 1260
    },
    {
      "epoch": 9.76923076923077,
      "grad_norm": 0.010753784328699112,
      "learning_rate": 1.6095384615384616e-05,
      "loss": 0.1313,
      "step": 1270
    },
    {
      "epoch": 9.846153846153847,
      "grad_norm": 4.099564552307129,
      "learning_rate": 1.6064615384615384e-05,
      "loss": 0.028,
      "step": 1280
    },
    {
      "epoch": 9.923076923076923,
      "grad_norm": 0.0062192706391215324,
      "learning_rate": 1.6033846153846155e-05,
      "loss": 0.0388,
      "step": 1290
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.002432856010273099,
      "learning_rate": 1.6003076923076923e-05,
      "loss": 0.0215,
      "step": 1300
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.0003331384214106947,
      "eval_runtime": 0.5589,
      "eval_samples_per_second": 237.973,
      "eval_steps_per_second": 30.418,
      "step": 1300
    },
    {
      "epoch": 10.076923076923077,
      "grad_norm": 40.07775115966797,
      "learning_rate": 1.5972307692307694e-05,
      "loss": 0.1465,
      "step": 1310
    },
    {
      "epoch": 10.153846153846153,
      "grad_norm": 0.03718170151114464,
      "learning_rate": 1.5941538461538465e-05,
      "loss": 0.0022,
      "step": 1320
    },
    {
      "epoch": 10.23076923076923,
      "grad_norm": 0.006770229432731867,
      "learning_rate": 1.5910769230769233e-05,
      "loss": 0.1207,
      "step": 1330
    },
    {
      "epoch": 10.307692307692308,
      "grad_norm": 1.803260326385498,
      "learning_rate": 1.588e-05,
      "loss": 0.1151,
      "step": 1340
    },
    {
      "epoch": 10.384615384615385,
      "grad_norm": 34.21731185913086,
      "learning_rate": 1.584923076923077e-05,
      "loss": 0.1446,
      "step": 1350
    },
    {
      "epoch": 10.461538461538462,
      "grad_norm": 0.011409844271838665,
      "learning_rate": 1.581846153846154e-05,
      "loss": 0.0283,
      "step": 1360
    },
    {
      "epoch": 10.538461538461538,
      "grad_norm": 0.010058027692139149,
      "learning_rate": 1.5787692307692307e-05,
      "loss": 0.133,
      "step": 1370
    },
    {
      "epoch": 10.615384615384615,
      "grad_norm": 0.9373985528945923,
      "learning_rate": 1.5756923076923078e-05,
      "loss": 0.0455,
      "step": 1380
    },
    {
      "epoch": 10.692307692307692,
      "grad_norm": 0.008180659264326096,
      "learning_rate": 1.5726153846153846e-05,
      "loss": 0.0249,
      "step": 1390
    },
    {
      "epoch": 10.76923076923077,
      "grad_norm": 66.53052520751953,
      "learning_rate": 1.5695384615384617e-05,
      "loss": 0.1623,
      "step": 1400
    },
    {
      "epoch": 10.846153846153847,
      "grad_norm": 0.008386868052184582,
      "learning_rate": 1.5664615384615388e-05,
      "loss": 0.0183,
      "step": 1410
    },
    {
      "epoch": 10.923076923076923,
      "grad_norm": 6.9249043464660645,
      "learning_rate": 1.5633846153846156e-05,
      "loss": 0.0448,
      "step": 1420
    },
    {
      "epoch": 11.0,
      "grad_norm": 0.0051022847183048725,
      "learning_rate": 1.5603076923076927e-05,
      "loss": 0.0501,
      "step": 1430
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9774436090225563,
      "eval_loss": 0.0634198933839798,
      "eval_runtime": 0.56,
      "eval_samples_per_second": 237.505,
      "eval_steps_per_second": 30.358,
      "step": 1430
    },
    {
      "epoch": 11.076923076923077,
      "grad_norm": 51.06268310546875,
      "learning_rate": 1.5572307692307694e-05,
      "loss": 0.1581,
      "step": 1440
    },
    {
      "epoch": 11.153846153846153,
      "grad_norm": 22.144161224365234,
      "learning_rate": 1.5541538461538462e-05,
      "loss": 0.0681,
      "step": 1450
    },
    {
      "epoch": 11.23076923076923,
      "grad_norm": 2.6330926418304443,
      "learning_rate": 1.551076923076923e-05,
      "loss": 0.0616,
      "step": 1460
    },
    {
      "epoch": 11.307692307692308,
      "grad_norm": 0.04433123394846916,
      "learning_rate": 1.548e-05,
      "loss": 0.1252,
      "step": 1470
    },
    {
      "epoch": 11.384615384615385,
      "grad_norm": 23.456703186035156,
      "learning_rate": 1.544923076923077e-05,
      "loss": 0.0639,
      "step": 1480
    },
    {
      "epoch": 11.461538461538462,
      "grad_norm": 1.3539597988128662,
      "learning_rate": 1.541846153846154e-05,
      "loss": 0.191,
      "step": 1490
    },
    {
      "epoch": 11.538461538461538,
      "grad_norm": 0.002958856290206313,
      "learning_rate": 1.5387692307692307e-05,
      "loss": 0.0002,
      "step": 1500
    },
    {
      "epoch": 11.615384615384615,
      "grad_norm": 0.001320499461144209,
      "learning_rate": 1.535692307692308e-05,
      "loss": 0.1247,
      "step": 1510
    },
    {
      "epoch": 11.692307692307692,
      "grad_norm": 26.650876998901367,
      "learning_rate": 1.532615384615385e-05,
      "loss": 0.2736,
      "step": 1520
    },
    {
      "epoch": 11.76923076923077,
      "grad_norm": 0.6191799640655518,
      "learning_rate": 1.5295384615384617e-05,
      "loss": 0.0123,
      "step": 1530
    },
    {
      "epoch": 11.846153846153847,
      "grad_norm": 0.16651397943496704,
      "learning_rate": 1.5264615384615385e-05,
      "loss": 0.071,
      "step": 1540
    },
    {
      "epoch": 11.923076923076923,
      "grad_norm": 0.5216618180274963,
      "learning_rate": 1.5233846153846154e-05,
      "loss": 0.1478,
      "step": 1550
    },
    {
      "epoch": 12.0,
      "grad_norm": 0.0047162421979010105,
      "learning_rate": 1.5203076923076925e-05,
      "loss": 0.0338,
      "step": 1560
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.024751892313361168,
      "eval_runtime": 0.5573,
      "eval_samples_per_second": 238.65,
      "eval_steps_per_second": 30.504,
      "step": 1560
    },
    {
      "epoch": 12.076923076923077,
      "grad_norm": 0.01325373724102974,
      "learning_rate": 1.5172307692307693e-05,
      "loss": 0.1432,
      "step": 1570
    },
    {
      "epoch": 12.153846153846153,
      "grad_norm": 24.465373992919922,
      "learning_rate": 1.5141538461538463e-05,
      "loss": 0.0911,
      "step": 1580
    },
    {
      "epoch": 12.23076923076923,
      "grad_norm": 0.01066196896135807,
      "learning_rate": 1.5110769230769232e-05,
      "loss": 0.0654,
      "step": 1590
    },
    {
      "epoch": 12.307692307692308,
      "grad_norm": 30.326496124267578,
      "learning_rate": 1.5080000000000001e-05,
      "loss": 0.0937,
      "step": 1600
    },
    {
      "epoch": 12.384615384615385,
      "grad_norm": 15.133655548095703,
      "learning_rate": 1.504923076923077e-05,
      "loss": 0.1265,
      "step": 1610
    },
    {
      "epoch": 12.461538461538462,
      "grad_norm": 0.009278634563088417,
      "learning_rate": 1.501846153846154e-05,
      "loss": 0.0028,
      "step": 1620
    },
    {
      "epoch": 12.538461538461538,
      "grad_norm": 34.179298400878906,
      "learning_rate": 1.498769230769231e-05,
      "loss": 0.0711,
      "step": 1630
    },
    {
      "epoch": 12.615384615384615,
      "grad_norm": 0.539430558681488,
      "learning_rate": 1.4956923076923077e-05,
      "loss": 0.0683,
      "step": 1640
    },
    {
      "epoch": 12.692307692307692,
      "grad_norm": 0.504645586013794,
      "learning_rate": 1.4926153846153848e-05,
      "loss": 0.1046,
      "step": 1650
    },
    {
      "epoch": 12.76923076923077,
      "grad_norm": 36.87862777709961,
      "learning_rate": 1.4895384615384616e-05,
      "loss": 0.0515,
      "step": 1660
    },
    {
      "epoch": 12.846153846153847,
      "grad_norm": 0.005421087611466646,
      "learning_rate": 1.4864615384615385e-05,
      "loss": 0.0943,
      "step": 1670
    },
    {
      "epoch": 12.923076923076923,
      "grad_norm": 54.710540771484375,
      "learning_rate": 1.4833846153846155e-05,
      "loss": 0.2085,
      "step": 1680
    },
    {
      "epoch": 13.0,
      "grad_norm": 0.004657375160604715,
      "learning_rate": 1.4803076923076924e-05,
      "loss": 0.0045,
      "step": 1690
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.09390003234148026,
      "eval_runtime": 0.5702,
      "eval_samples_per_second": 233.258,
      "eval_steps_per_second": 29.815,
      "step": 1690
    },
    {
      "epoch": 13.076923076923077,
      "grad_norm": 0.014458782970905304,
      "learning_rate": 1.4772307692307692e-05,
      "loss": 0.0008,
      "step": 1700
    },
    {
      "epoch": 13.153846153846153,
      "grad_norm": 4.515313625335693,
      "learning_rate": 1.4741538461538463e-05,
      "loss": 0.0232,
      "step": 1710
    },
    {
      "epoch": 13.23076923076923,
      "grad_norm": 20.115612030029297,
      "learning_rate": 1.4710769230769232e-05,
      "loss": 0.0343,
      "step": 1720
    },
    {
      "epoch": 13.307692307692308,
      "grad_norm": 0.0025744156446307898,
      "learning_rate": 1.4680000000000002e-05,
      "loss": 0.2143,
      "step": 1730
    },
    {
      "epoch": 13.384615384615385,
      "grad_norm": 0.005626370664685965,
      "learning_rate": 1.4649230769230771e-05,
      "loss": 0.1297,
      "step": 1740
    },
    {
      "epoch": 13.461538461538462,
      "grad_norm": 1.902580738067627,
      "learning_rate": 1.4618461538461539e-05,
      "loss": 0.0052,
      "step": 1750
    },
    {
      "epoch": 13.538461538461538,
      "grad_norm": 0.0006718473159708083,
      "learning_rate": 1.458769230769231e-05,
      "loss": 0.1279,
      "step": 1760
    },
    {
      "epoch": 13.615384615384615,
      "grad_norm": 0.40659937262535095,
      "learning_rate": 1.4556923076923078e-05,
      "loss": 0.0005,
      "step": 1770
    },
    {
      "epoch": 13.692307692307692,
      "grad_norm": 0.0035146658774465322,
      "learning_rate": 1.4526153846153847e-05,
      "loss": 0.0778,
      "step": 1780
    },
    {
      "epoch": 13.76923076923077,
      "grad_norm": 0.05003712326288223,
      "learning_rate": 1.4495384615384616e-05,
      "loss": 0.002,
      "step": 1790
    },
    {
      "epoch": 13.846153846153847,
      "grad_norm": 0.0015587609959766269,
      "learning_rate": 1.4464615384615386e-05,
      "loss": 0.1025,
      "step": 1800
    },
    {
      "epoch": 13.923076923076923,
      "grad_norm": 0.011278959922492504,
      "learning_rate": 1.4433846153846155e-05,
      "loss": 0.141,
      "step": 1810
    },
    {
      "epoch": 14.0,
      "grad_norm": 0.002003464847803116,
      "learning_rate": 1.4403076923076925e-05,
      "loss": 0.0013,
      "step": 1820
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.03729831054806709,
      "eval_runtime": 0.5723,
      "eval_samples_per_second": 232.408,
      "eval_steps_per_second": 29.706,
      "step": 1820
    },
    {
      "epoch": 14.076923076923077,
      "grad_norm": 0.023831067606806755,
      "learning_rate": 1.4372307692307694e-05,
      "loss": 0.0016,
      "step": 1830
    },
    {
      "epoch": 14.153846153846153,
      "grad_norm": 23.71483612060547,
      "learning_rate": 1.4341538461538462e-05,
      "loss": 0.1814,
      "step": 1840
    },
    {
      "epoch": 14.23076923076923,
      "grad_norm": 0.0006417655386030674,
      "learning_rate": 1.4310769230769233e-05,
      "loss": 0.1044,
      "step": 1850
    },
    {
      "epoch": 14.307692307692308,
      "grad_norm": 0.003414665814489126,
      "learning_rate": 1.428e-05,
      "loss": 0.1385,
      "step": 1860
    },
    {
      "epoch": 14.384615384615385,
      "grad_norm": 0.9071229100227356,
      "learning_rate": 1.4249230769230772e-05,
      "loss": 0.1259,
      "step": 1870
    },
    {
      "epoch": 14.461538461538462,
      "grad_norm": 15.944780349731445,
      "learning_rate": 1.421846153846154e-05,
      "loss": 0.1491,
      "step": 1880
    },
    {
      "epoch": 14.538461538461538,
      "grad_norm": 0.040743470191955566,
      "learning_rate": 1.4187692307692309e-05,
      "loss": 0.0304,
      "step": 1890
    },
    {
      "epoch": 14.615384615384615,
      "grad_norm": 0.09457922726869583,
      "learning_rate": 1.4156923076923076e-05,
      "loss": 0.1211,
      "step": 1900
    },
    {
      "epoch": 14.692307692307692,
      "grad_norm": 0.19244757294654846,
      "learning_rate": 1.4126153846153847e-05,
      "loss": 0.1062,
      "step": 1910
    },
    {
      "epoch": 14.76923076923077,
      "grad_norm": 1.560502529144287,
      "learning_rate": 1.4095384615384617e-05,
      "loss": 0.1089,
      "step": 1920
    },
    {
      "epoch": 14.846153846153847,
      "grad_norm": 0.030071793124079704,
      "learning_rate": 1.4064615384615386e-05,
      "loss": 0.0024,
      "step": 1930
    },
    {
      "epoch": 14.923076923076923,
      "grad_norm": 0.0013134493492543697,
      "learning_rate": 1.4033846153846156e-05,
      "loss": 0.0012,
      "step": 1940
    },
    {
      "epoch": 15.0,
      "grad_norm": 0.0009961730102077127,
      "learning_rate": 1.4003076923076923e-05,
      "loss": 0.0002,
      "step": 1950
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.0515468567609787,
      "eval_runtime": 0.5584,
      "eval_samples_per_second": 238.17,
      "eval_steps_per_second": 30.443,
      "step": 1950
    },
    {
      "epoch": 15.076923076923077,
      "grad_norm": 0.0006362455897033215,
      "learning_rate": 1.3972307692307694e-05,
      "loss": 0.0451,
      "step": 1960
    },
    {
      "epoch": 15.153846153846153,
      "grad_norm": 0.0071708387695252895,
      "learning_rate": 1.3941538461538462e-05,
      "loss": 0.0534,
      "step": 1970
    },
    {
      "epoch": 15.23076923076923,
      "grad_norm": 0.006367762107402086,
      "learning_rate": 1.3910769230769232e-05,
      "loss": 0.0013,
      "step": 1980
    },
    {
      "epoch": 15.307692307692308,
      "grad_norm": 1.7080107927322388,
      "learning_rate": 1.3880000000000001e-05,
      "loss": 0.0584,
      "step": 1990
    },
    {
      "epoch": 15.384615384615385,
      "grad_norm": 48.49470901489258,
      "learning_rate": 1.384923076923077e-05,
      "loss": 0.0883,
      "step": 2000
    },
    {
      "epoch": 15.461538461538462,
      "grad_norm": 0.0020162356086075306,
      "learning_rate": 1.3818461538461541e-05,
      "loss": 0.0147,
      "step": 2010
    },
    {
      "epoch": 15.538461538461538,
      "grad_norm": 0.11068695038557053,
      "learning_rate": 1.3787692307692309e-05,
      "loss": 0.0926,
      "step": 2020
    },
    {
      "epoch": 15.615384615384615,
      "grad_norm": 0.001251698238775134,
      "learning_rate": 1.3756923076923079e-05,
      "loss": 0.011,
      "step": 2030
    },
    {
      "epoch": 15.692307692307692,
      "grad_norm": 0.0013289551716297865,
      "learning_rate": 1.3726153846153846e-05,
      "loss": 0.0048,
      "step": 2040
    },
    {
      "epoch": 15.76923076923077,
      "grad_norm": 87.35831451416016,
      "learning_rate": 1.3695384615384617e-05,
      "loss": 0.0228,
      "step": 2050
    },
    {
      "epoch": 15.846153846153847,
      "grad_norm": 0.009976107627153397,
      "learning_rate": 1.3664615384615385e-05,
      "loss": 0.2276,
      "step": 2060
    },
    {
      "epoch": 15.923076923076923,
      "grad_norm": 0.00704730162397027,
      "learning_rate": 1.3633846153846156e-05,
      "loss": 0.248,
      "step": 2070
    },
    {
      "epoch": 16.0,
      "grad_norm": 0.0022731584031134844,
      "learning_rate": 1.3603076923076924e-05,
      "loss": 0.0074,
      "step": 2080
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.0016808852087706327,
      "eval_runtime": 0.5593,
      "eval_samples_per_second": 237.796,
      "eval_steps_per_second": 30.395,
      "step": 2080
    },
    {
      "epoch": 16.076923076923077,
      "grad_norm": 4.551645278930664,
      "learning_rate": 1.3572307692307693e-05,
      "loss": 0.0013,
      "step": 2090
    },
    {
      "epoch": 16.153846153846153,
      "grad_norm": 0.004595694597810507,
      "learning_rate": 1.3541538461538464e-05,
      "loss": 0.0017,
      "step": 2100
    },
    {
      "epoch": 16.23076923076923,
      "grad_norm": 0.027721809223294258,
      "learning_rate": 1.3510769230769232e-05,
      "loss": 0.0054,
      "step": 2110
    },
    {
      "epoch": 16.307692307692307,
      "grad_norm": 159.5991668701172,
      "learning_rate": 1.3480000000000001e-05,
      "loss": 0.2422,
      "step": 2120
    },
    {
      "epoch": 16.384615384615383,
      "grad_norm": 0.0032202787697315216,
      "learning_rate": 1.344923076923077e-05,
      "loss": 0.0059,
      "step": 2130
    },
    {
      "epoch": 16.46153846153846,
      "grad_norm": 0.004118644632399082,
      "learning_rate": 1.341846153846154e-05,
      "loss": 0.0003,
      "step": 2140
    },
    {
      "epoch": 16.53846153846154,
      "grad_norm": 0.796428382396698,
      "learning_rate": 1.3387692307692308e-05,
      "loss": 0.0837,
      "step": 2150
    },
    {
      "epoch": 16.615384615384617,
      "grad_norm": 0.001536823227070272,
      "learning_rate": 1.3356923076923079e-05,
      "loss": 0.0004,
      "step": 2160
    },
    {
      "epoch": 16.692307692307693,
      "grad_norm": 0.0012222144287079573,
      "learning_rate": 1.3326153846153847e-05,
      "loss": 0.1912,
      "step": 2170
    },
    {
      "epoch": 16.76923076923077,
      "grad_norm": 0.001332849613390863,
      "learning_rate": 1.3295384615384616e-05,
      "loss": 0.0007,
      "step": 2180
    },
    {
      "epoch": 16.846153846153847,
      "grad_norm": 5.964162826538086,
      "learning_rate": 1.3264615384615385e-05,
      "loss": 0.2174,
      "step": 2190
    },
    {
      "epoch": 16.923076923076923,
      "grad_norm": 0.010293848812580109,
      "learning_rate": 1.3233846153846155e-05,
      "loss": 0.0657,
      "step": 2200
    },
    {
      "epoch": 17.0,
      "grad_norm": 0.005218516103923321,
      "learning_rate": 1.3203076923076926e-05,
      "loss": 0.0005,
      "step": 2210
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.058793652802705765,
      "eval_runtime": 0.5588,
      "eval_samples_per_second": 238.01,
      "eval_steps_per_second": 30.422,
      "step": 2210
    },
    {
      "epoch": 17.076923076923077,
      "grad_norm": 0.007321541663259268,
      "learning_rate": 1.3172307692307694e-05,
      "loss": 0.0652,
      "step": 2220
    },
    {
      "epoch": 17.153846153846153,
      "grad_norm": 0.24121706187725067,
      "learning_rate": 1.3141538461538463e-05,
      "loss": 0.1264,
      "step": 2230
    },
    {
      "epoch": 17.23076923076923,
      "grad_norm": 0.014303168281912804,
      "learning_rate": 1.311076923076923e-05,
      "loss": 0.0746,
      "step": 2240
    },
    {
      "epoch": 17.307692307692307,
      "grad_norm": 0.004994900431483984,
      "learning_rate": 1.3080000000000002e-05,
      "loss": 0.0033,
      "step": 2250
    },
    {
      "epoch": 17.384615384615383,
      "grad_norm": 0.0024373421911150217,
      "learning_rate": 1.304923076923077e-05,
      "loss": 0.0008,
      "step": 2260
    },
    {
      "epoch": 17.46153846153846,
      "grad_norm": 0.0018968008225783706,
      "learning_rate": 1.301846153846154e-05,
      "loss": 0.0121,
      "step": 2270
    },
    {
      "epoch": 17.53846153846154,
      "grad_norm": 0.004607335664331913,
      "learning_rate": 1.2987692307692308e-05,
      "loss": 0.0386,
      "step": 2280
    },
    {
      "epoch": 17.615384615384617,
      "grad_norm": 0.0010372723918408155,
      "learning_rate": 1.2956923076923078e-05,
      "loss": 0.0001,
      "step": 2290
    },
    {
      "epoch": 17.692307692307693,
      "grad_norm": 0.1868879795074463,
      "learning_rate": 1.2926153846153849e-05,
      "loss": 0.1025,
      "step": 2300
    },
    {
      "epoch": 17.76923076923077,
      "grad_norm": 0.001327532809227705,
      "learning_rate": 1.2895384615384616e-05,
      "loss": 0.1086,
      "step": 2310
    },
    {
      "epoch": 17.846153846153847,
      "grad_norm": 0.002292274497449398,
      "learning_rate": 1.2864615384615386e-05,
      "loss": 0.1219,
      "step": 2320
    },
    {
      "epoch": 17.923076923076923,
      "grad_norm": 0.0024085345212370157,
      "learning_rate": 1.2833846153846155e-05,
      "loss": 0.2003,
      "step": 2330
    },
    {
      "epoch": 18.0,
      "grad_norm": 0.01654043421149254,
      "learning_rate": 1.2803076923076925e-05,
      "loss": 0.0046,
      "step": 2340
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.07149150967597961,
      "eval_runtime": 0.6008,
      "eval_samples_per_second": 221.365,
      "eval_steps_per_second": 28.295,
      "step": 2340
    },
    {
      "epoch": 18.076923076923077,
      "grad_norm": 0.0006843829760327935,
      "learning_rate": 1.2772307692307692e-05,
      "loss": 0.0572,
      "step": 2350
    },
    {
      "epoch": 18.153846153846153,
      "grad_norm": 0.00446607219055295,
      "learning_rate": 1.2741538461538463e-05,
      "loss": 0.1223,
      "step": 2360
    },
    {
      "epoch": 18.23076923076923,
      "grad_norm": 60.59695053100586,
      "learning_rate": 1.2710769230769231e-05,
      "loss": 0.155,
      "step": 2370
    },
    {
      "epoch": 18.307692307692307,
      "grad_norm": 0.001972057158127427,
      "learning_rate": 1.268e-05,
      "loss": 0.0193,
      "step": 2380
    },
    {
      "epoch": 18.384615384615383,
      "grad_norm": 36.57919692993164,
      "learning_rate": 1.264923076923077e-05,
      "loss": 0.276,
      "step": 2390
    },
    {
      "epoch": 18.46153846153846,
      "grad_norm": 0.3979909420013428,
      "learning_rate": 1.261846153846154e-05,
      "loss": 0.0329,
      "step": 2400
    },
    {
      "epoch": 18.53846153846154,
      "grad_norm": 0.05160733684897423,
      "learning_rate": 1.258769230769231e-05,
      "loss": 0.0179,
      "step": 2410
    },
    {
      "epoch": 18.615384615384617,
      "grad_norm": 51.901119232177734,
      "learning_rate": 1.2556923076923078e-05,
      "loss": 0.0916,
      "step": 2420
    },
    {
      "epoch": 18.692307692307693,
      "grad_norm": 0.018259704113006592,
      "learning_rate": 1.2526153846153848e-05,
      "loss": 0.0002,
      "step": 2430
    },
    {
      "epoch": 18.76923076923077,
      "grad_norm": 0.010010771453380585,
      "learning_rate": 1.2495384615384615e-05,
      "loss": 0.1238,
      "step": 2440
    },
    {
      "epoch": 18.846153846153847,
      "grad_norm": 0.026239359751343727,
      "learning_rate": 1.2464615384615386e-05,
      "loss": 0.0222,
      "step": 2450
    },
    {
      "epoch": 18.923076923076923,
      "grad_norm": 5.157707214355469,
      "learning_rate": 1.2433846153846154e-05,
      "loss": 0.0424,
      "step": 2460
    },
    {
      "epoch": 19.0,
      "grad_norm": 0.02478371374309063,
      "learning_rate": 1.2403076923076925e-05,
      "loss": 0.0618,
      "step": 2470
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.00025587898562662303,
      "eval_runtime": 0.5583,
      "eval_samples_per_second": 238.234,
      "eval_steps_per_second": 30.451,
      "step": 2470
    },
    {
      "epoch": 19.076923076923077,
      "grad_norm": 0.034582603722810745,
      "learning_rate": 1.2372307692307693e-05,
      "loss": 0.0451,
      "step": 2480
    },
    {
      "epoch": 19.153846153846153,
      "grad_norm": 0.0011788305127993226,
      "learning_rate": 1.2341538461538462e-05,
      "loss": 0.1372,
      "step": 2490
    },
    {
      "epoch": 19.23076923076923,
      "grad_norm": 0.00216018408536911,
      "learning_rate": 1.2310769230769233e-05,
      "loss": 0.0001,
      "step": 2500
    },
    {
      "epoch": 19.307692307692307,
      "grad_norm": 0.0017174163367599249,
      "learning_rate": 1.2280000000000001e-05,
      "loss": 0.0522,
      "step": 2510
    },
    {
      "epoch": 19.384615384615383,
      "grad_norm": 73.49071502685547,
      "learning_rate": 1.224923076923077e-05,
      "loss": 0.0448,
      "step": 2520
    },
    {
      "epoch": 19.46153846153846,
      "grad_norm": 109.05286407470703,
      "learning_rate": 1.221846153846154e-05,
      "loss": 0.1094,
      "step": 2530
    },
    {
      "epoch": 19.53846153846154,
      "grad_norm": 46.95461654663086,
      "learning_rate": 1.218769230769231e-05,
      "loss": 0.0483,
      "step": 2540
    },
    {
      "epoch": 19.615384615384617,
      "grad_norm": 0.001702558365650475,
      "learning_rate": 1.2156923076923077e-05,
      "loss": 0.0478,
      "step": 2550
    },
    {
      "epoch": 19.692307692307693,
      "grad_norm": 0.016203153878450394,
      "learning_rate": 1.2126153846153848e-05,
      "loss": 0.0006,
      "step": 2560
    },
    {
      "epoch": 19.76923076923077,
      "grad_norm": 0.001535124727524817,
      "learning_rate": 1.2095384615384616e-05,
      "loss": 0.0169,
      "step": 2570
    },
    {
      "epoch": 19.846153846153847,
      "grad_norm": 0.0012106726644560695,
      "learning_rate": 1.2064615384615385e-05,
      "loss": 0.0004,
      "step": 2580
    },
    {
      "epoch": 19.923076923076923,
      "grad_norm": 17.061466217041016,
      "learning_rate": 1.2033846153846154e-05,
      "loss": 0.1205,
      "step": 2590
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.009339146316051483,
      "learning_rate": 1.2003076923076924e-05,
      "loss": 0.0007,
      "step": 2600
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.06972503662109375,
      "eval_runtime": 0.5563,
      "eval_samples_per_second": 239.076,
      "eval_steps_per_second": 30.559,
      "step": 2600
    },
    {
      "epoch": 20.076923076923077,
      "grad_norm": 0.009359082207083702,
      "learning_rate": 1.1972307692307695e-05,
      "loss": 0.1452,
      "step": 2610
    },
    {
      "epoch": 20.153846153846153,
      "grad_norm": 0.06399708241224289,
      "learning_rate": 1.1941538461538463e-05,
      "loss": 0.014,
      "step": 2620
    },
    {
      "epoch": 20.23076923076923,
      "grad_norm": 0.002059635240584612,
      "learning_rate": 1.1910769230769232e-05,
      "loss": 0.0001,
      "step": 2630
    },
    {
      "epoch": 20.307692307692307,
      "grad_norm": 31.44277572631836,
      "learning_rate": 1.188e-05,
      "loss": 0.1186,
      "step": 2640
    },
    {
      "epoch": 20.384615384615383,
      "grad_norm": 0.004387051798403263,
      "learning_rate": 1.1849230769230771e-05,
      "loss": 0.0004,
      "step": 2650
    },
    {
      "epoch": 20.46153846153846,
      "grad_norm": 0.0017380157951265574,
      "learning_rate": 1.1818461538461539e-05,
      "loss": 0.1127,
      "step": 2660
    },
    {
      "epoch": 20.53846153846154,
      "grad_norm": 0.01089541893452406,
      "learning_rate": 1.178769230769231e-05,
      "loss": 0.0869,
      "step": 2670
    },
    {
      "epoch": 20.615384615384617,
      "grad_norm": 0.0011546051828190684,
      "learning_rate": 1.1756923076923077e-05,
      "loss": 0.2089,
      "step": 2680
    },
    {
      "epoch": 20.692307692307693,
      "grad_norm": 0.003806912340223789,
      "learning_rate": 1.1726153846153847e-05,
      "loss": 0.1001,
      "step": 2690
    },
    {
      "epoch": 20.76923076923077,
      "grad_norm": 0.000958717311732471,
      "learning_rate": 1.1695384615384618e-05,
      "loss": 0.183,
      "step": 2700
    },
    {
      "epoch": 20.846153846153847,
      "grad_norm": 0.18556897342205048,
      "learning_rate": 1.1664615384615386e-05,
      "loss": 0.0003,
      "step": 2710
    },
    {
      "epoch": 20.923076923076923,
      "grad_norm": 0.0016327186021953821,
      "learning_rate": 1.1633846153846155e-05,
      "loss": 0.1264,
      "step": 2720
    },
    {
      "epoch": 21.0,
      "grad_norm": 0.002433057175949216,
      "learning_rate": 1.1603076923076924e-05,
      "loss": 0.0001,
      "step": 2730
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.9774436090225563,
      "eval_loss": 0.11052080243825912,
      "eval_runtime": 0.5588,
      "eval_samples_per_second": 238.023,
      "eval_steps_per_second": 30.424,
      "step": 2730
    },
    {
      "epoch": 21.076923076923077,
      "grad_norm": 0.004819987341761589,
      "learning_rate": 1.1572307692307694e-05,
      "loss": 0.0602,
      "step": 2740
    },
    {
      "epoch": 21.153846153846153,
      "grad_norm": 0.0032386251259595156,
      "learning_rate": 1.1541538461538461e-05,
      "loss": 0.0086,
      "step": 2750
    },
    {
      "epoch": 21.23076923076923,
      "grad_norm": 0.0015244975220412016,
      "learning_rate": 1.1510769230769232e-05,
      "loss": 0.0002,
      "step": 2760
    },
    {
      "epoch": 21.307692307692307,
      "grad_norm": 0.0007208559545688331,
      "learning_rate": 1.148e-05,
      "loss": 0.0108,
      "step": 2770
    },
    {
      "epoch": 21.384615384615383,
      "grad_norm": 2.059755325317383,
      "learning_rate": 1.144923076923077e-05,
      "loss": 0.0232,
      "step": 2780
    },
    {
      "epoch": 21.46153846153846,
      "grad_norm": 0.0012232252629473805,
      "learning_rate": 1.141846153846154e-05,
      "loss": 0.0,
      "step": 2790
    },
    {
      "epoch": 21.53846153846154,
      "grad_norm": 0.07696986198425293,
      "learning_rate": 1.1387692307692308e-05,
      "loss": 0.2859,
      "step": 2800
    },
    {
      "epoch": 21.615384615384617,
      "grad_norm": 61.04636001586914,
      "learning_rate": 1.135692307692308e-05,
      "loss": 0.1596,
      "step": 2810
    },
    {
      "epoch": 21.692307692307693,
      "grad_norm": 0.001321778865531087,
      "learning_rate": 1.1326153846153847e-05,
      "loss": 0.134,
      "step": 2820
    },
    {
      "epoch": 21.76923076923077,
      "grad_norm": 0.08154579252004623,
      "learning_rate": 1.1295384615384617e-05,
      "loss": 0.0001,
      "step": 2830
    },
    {
      "epoch": 21.846153846153847,
      "grad_norm": 0.007910426706075668,
      "learning_rate": 1.1264615384615384e-05,
      "loss": 0.1725,
      "step": 2840
    },
    {
      "epoch": 21.923076923076923,
      "grad_norm": 0.010755512863397598,
      "learning_rate": 1.1233846153846155e-05,
      "loss": 0.0768,
      "step": 2850
    },
    {
      "epoch": 22.0,
      "grad_norm": 12.361237525939941,
      "learning_rate": 1.1203076923076923e-05,
      "loss": 0.0214,
      "step": 2860
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.09297900646924973,
      "eval_runtime": 0.5599,
      "eval_samples_per_second": 237.551,
      "eval_steps_per_second": 30.364,
      "step": 2860
    },
    {
      "epoch": 22.076923076923077,
      "grad_norm": 10.506192207336426,
      "learning_rate": 1.1172307692307694e-05,
      "loss": 0.008,
      "step": 2870
    },
    {
      "epoch": 22.153846153846153,
      "grad_norm": 0.0015517654828727245,
      "learning_rate": 1.1141538461538462e-05,
      "loss": 0.0735,
      "step": 2880
    },
    {
      "epoch": 22.23076923076923,
      "grad_norm": 42.15374755859375,
      "learning_rate": 1.1110769230769231e-05,
      "loss": 0.0698,
      "step": 2890
    },
    {
      "epoch": 22.307692307692307,
      "grad_norm": 0.003171928459778428,
      "learning_rate": 1.1080000000000002e-05,
      "loss": 0.0475,
      "step": 2900
    },
    {
      "epoch": 22.384615384615383,
      "grad_norm": 0.0019311098149046302,
      "learning_rate": 1.104923076923077e-05,
      "loss": 0.0003,
      "step": 2910
    },
    {
      "epoch": 22.46153846153846,
      "grad_norm": 0.0021256860345602036,
      "learning_rate": 1.101846153846154e-05,
      "loss": 0.0308,
      "step": 2920
    },
    {
      "epoch": 22.53846153846154,
      "grad_norm": 0.01151897944509983,
      "learning_rate": 1.0987692307692309e-05,
      "loss": 0.123,
      "step": 2930
    },
    {
      "epoch": 22.615384615384617,
      "grad_norm": 0.009536568075418472,
      "learning_rate": 1.0956923076923078e-05,
      "loss": 0.0018,
      "step": 2940
    },
    {
      "epoch": 22.692307692307693,
      "grad_norm": 0.04015596956014633,
      "learning_rate": 1.0926153846153846e-05,
      "loss": 0.0961,
      "step": 2950
    },
    {
      "epoch": 22.76923076923077,
      "grad_norm": 0.009416976943612099,
      "learning_rate": 1.0895384615384617e-05,
      "loss": 0.1117,
      "step": 2960
    },
    {
      "epoch": 22.846153846153847,
      "grad_norm": 0.0018806488951668143,
      "learning_rate": 1.0864615384615385e-05,
      "loss": 0.0012,
      "step": 2970
    },
    {
      "epoch": 22.923076923076923,
      "grad_norm": 0.006934476085007191,
      "learning_rate": 1.0833846153846154e-05,
      "loss": 0.0016,
      "step": 2980
    },
    {
      "epoch": 23.0,
      "grad_norm": 0.0018909511854872108,
      "learning_rate": 1.0803076923076925e-05,
      "loss": 0.0004,
      "step": 2990
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.027222247794270515,
      "eval_runtime": 0.5445,
      "eval_samples_per_second": 244.243,
      "eval_steps_per_second": 31.219,
      "step": 2990
    },
    {
      "epoch": 23.076923076923077,
      "grad_norm": 0.0007299839635379612,
      "learning_rate": 1.0772307692307693e-05,
      "loss": 0.0456,
      "step": 3000
    },
    {
      "epoch": 23.153846153846153,
      "grad_norm": 0.0018170330440625548,
      "learning_rate": 1.0741538461538464e-05,
      "loss": 0.0001,
      "step": 3010
    },
    {
      "epoch": 23.23076923076923,
      "grad_norm": 0.7225848436355591,
      "learning_rate": 1.0710769230769232e-05,
      "loss": 0.0004,
      "step": 3020
    },
    {
      "epoch": 23.307692307692307,
      "grad_norm": 0.0020151655189692974,
      "learning_rate": 1.0680000000000001e-05,
      "loss": 0.0014,
      "step": 3030
    },
    {
      "epoch": 23.384615384615383,
      "grad_norm": 0.00038547179428860545,
      "learning_rate": 1.0649230769230769e-05,
      "loss": 0.0051,
      "step": 3040
    },
    {
      "epoch": 23.46153846153846,
      "grad_norm": 0.00043245829874649644,
      "learning_rate": 1.061846153846154e-05,
      "loss": 0.1965,
      "step": 3050
    },
    {
      "epoch": 23.53846153846154,
      "grad_norm": 0.46839669346809387,
      "learning_rate": 1.0587692307692308e-05,
      "loss": 0.0002,
      "step": 3060
    },
    {
      "epoch": 23.615384615384617,
      "grad_norm": 0.008953952230513096,
      "learning_rate": 1.0556923076923079e-05,
      "loss": 0.1007,
      "step": 3070
    },
    {
      "epoch": 23.692307692307693,
      "grad_norm": 0.05692081153392792,
      "learning_rate": 1.0526153846153846e-05,
      "loss": 0.0562,
      "step": 3080
    },
    {
      "epoch": 23.76923076923077,
      "grad_norm": 0.004630447365343571,
      "learning_rate": 1.0495384615384616e-05,
      "loss": 0.0013,
      "step": 3090
    },
    {
      "epoch": 23.846153846153847,
      "grad_norm": 0.002413865178823471,
      "learning_rate": 1.0464615384615387e-05,
      "loss": 0.0804,
      "step": 3100
    },
    {
      "epoch": 23.923076923076923,
      "grad_norm": 0.002768196165561676,
      "learning_rate": 1.0433846153846155e-05,
      "loss": 0.0004,
      "step": 3110
    },
    {
      "epoch": 24.0,
      "grad_norm": 0.00019644868734758347,
      "learning_rate": 1.0403076923076924e-05,
      "loss": 0.1619,
      "step": 3120
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.0024267190601676702,
      "eval_runtime": 0.6209,
      "eval_samples_per_second": 214.222,
      "eval_steps_per_second": 27.382,
      "step": 3120
    },
    {
      "epoch": 24.076923076923077,
      "grad_norm": 34.10586166381836,
      "learning_rate": 1.0372307692307693e-05,
      "loss": 0.1111,
      "step": 3130
    },
    {
      "epoch": 24.153846153846153,
      "grad_norm": 0.07522152364253998,
      "learning_rate": 1.0341538461538463e-05,
      "loss": 0.0008,
      "step": 3140
    },
    {
      "epoch": 24.23076923076923,
      "grad_norm": 92.14463806152344,
      "learning_rate": 1.031076923076923e-05,
      "loss": 0.121,
      "step": 3150
    },
    {
      "epoch": 24.307692307692307,
      "grad_norm": 0.0009277680655941367,
      "learning_rate": 1.0280000000000002e-05,
      "loss": 0.1377,
      "step": 3160
    },
    {
      "epoch": 24.384615384615383,
      "grad_norm": 0.006104259751737118,
      "learning_rate": 1.024923076923077e-05,
      "loss": 0.0001,
      "step": 3170
    },
    {
      "epoch": 24.46153846153846,
      "grad_norm": 53.126731872558594,
      "learning_rate": 1.0218461538461539e-05,
      "loss": 0.0853,
      "step": 3180
    },
    {
      "epoch": 24.53846153846154,
      "grad_norm": 0.0030978797003626823,
      "learning_rate": 1.018769230769231e-05,
      "loss": 0.181,
      "step": 3190
    },
    {
      "epoch": 24.615384615384617,
      "grad_norm": 0.6085143685340881,
      "learning_rate": 1.0156923076923077e-05,
      "loss": 0.0002,
      "step": 3200
    },
    {
      "epoch": 24.692307692307693,
      "grad_norm": 0.015882331877946854,
      "learning_rate": 1.0126153846153849e-05,
      "loss": 0.0001,
      "step": 3210
    },
    {
      "epoch": 24.76923076923077,
      "grad_norm": 0.000975192931946367,
      "learning_rate": 1.0095384615384616e-05,
      "loss": 0.0003,
      "step": 3220
    },
    {
      "epoch": 24.846153846153847,
      "grad_norm": 0.0011907311854884028,
      "learning_rate": 1.0064615384615386e-05,
      "loss": 0.0001,
      "step": 3230
    },
    {
      "epoch": 24.923076923076923,
      "grad_norm": 20.971738815307617,
      "learning_rate": 1.0033846153846153e-05,
      "loss": 0.2057,
      "step": 3240
    },
    {
      "epoch": 25.0,
      "grad_norm": 0.0007839695899747312,
      "learning_rate": 1.0003076923076924e-05,
      "loss": 0.0015,
      "step": 3250
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.000277143029961735,
      "eval_runtime": 0.5589,
      "eval_samples_per_second": 237.951,
      "eval_steps_per_second": 30.415,
      "step": 3250
    },
    {
      "epoch": 25.076923076923077,
      "grad_norm": 0.0034783727023750544,
      "learning_rate": 9.972307692307694e-06,
      "loss": 0.059,
      "step": 3260
    },
    {
      "epoch": 25.153846153846153,
      "grad_norm": 0.08948632329702377,
      "learning_rate": 9.941538461538463e-06,
      "loss": 0.0159,
      "step": 3270
    },
    {
      "epoch": 25.23076923076923,
      "grad_norm": 3.0690531730651855,
      "learning_rate": 9.910769230769231e-06,
      "loss": 0.0012,
      "step": 3280
    },
    {
      "epoch": 25.307692307692307,
      "grad_norm": 0.02405349537730217,
      "learning_rate": 9.88e-06,
      "loss": 0.1706,
      "step": 3290
    },
    {
      "epoch": 25.384615384615383,
      "grad_norm": 0.003044095588847995,
      "learning_rate": 9.84923076923077e-06,
      "loss": 0.0011,
      "step": 3300
    },
    {
      "epoch": 25.46153846153846,
      "grad_norm": 0.0008050467586144805,
      "learning_rate": 9.818461538461539e-06,
      "loss": 0.0006,
      "step": 3310
    },
    {
      "epoch": 25.53846153846154,
      "grad_norm": 0.0011971103958785534,
      "learning_rate": 9.787692307692308e-06,
      "loss": 0.0533,
      "step": 3320
    },
    {
      "epoch": 25.615384615384617,
      "grad_norm": 0.0010218324605375528,
      "learning_rate": 9.756923076923078e-06,
      "loss": 0.0695,
      "step": 3330
    },
    {
      "epoch": 25.692307692307693,
      "grad_norm": 0.0007865706575103104,
      "learning_rate": 9.726153846153847e-06,
      "loss": 0.0096,
      "step": 3340
    },
    {
      "epoch": 25.76923076923077,
      "grad_norm": 70.63385772705078,
      "learning_rate": 9.695384615384617e-06,
      "loss": 0.1136,
      "step": 3350
    },
    {
      "epoch": 25.846153846153847,
      "grad_norm": 0.0011000982485711575,
      "learning_rate": 9.664615384615386e-06,
      "loss": 0.1563,
      "step": 3360
    },
    {
      "epoch": 25.923076923076923,
      "grad_norm": 0.06354475021362305,
      "learning_rate": 9.633846153846155e-06,
      "loss": 0.1292,
      "step": 3370
    },
    {
      "epoch": 26.0,
      "grad_norm": 0.0007121621747501194,
      "learning_rate": 9.603076923076923e-06,
      "loss": 0.0148,
      "step": 3380
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.9774436090225563,
      "eval_loss": 0.13121125102043152,
      "eval_runtime": 0.5643,
      "eval_samples_per_second": 235.687,
      "eval_steps_per_second": 30.125,
      "step": 3380
    },
    {
      "epoch": 26.076923076923077,
      "grad_norm": 0.6681668162345886,
      "learning_rate": 9.572307692307693e-06,
      "loss": 0.1477,
      "step": 3390
    },
    {
      "epoch": 26.153846153846153,
      "grad_norm": 0.6779555678367615,
      "learning_rate": 9.541538461538462e-06,
      "loss": 0.0002,
      "step": 3400
    },
    {
      "epoch": 26.23076923076923,
      "grad_norm": 0.004583708010613918,
      "learning_rate": 9.510769230769231e-06,
      "loss": 0.1163,
      "step": 3410
    },
    {
      "epoch": 26.307692307692307,
      "grad_norm": 0.5304880738258362,
      "learning_rate": 9.48e-06,
      "loss": 0.0515,
      "step": 3420
    },
    {
      "epoch": 26.384615384615383,
      "grad_norm": 7.967146396636963,
      "learning_rate": 9.44923076923077e-06,
      "loss": 0.0007,
      "step": 3430
    },
    {
      "epoch": 26.46153846153846,
      "grad_norm": 18.322139739990234,
      "learning_rate": 9.41846153846154e-06,
      "loss": 0.0199,
      "step": 3440
    },
    {
      "epoch": 26.53846153846154,
      "grad_norm": 0.0012510762317106128,
      "learning_rate": 9.387692307692309e-06,
      "loss": 0.1079,
      "step": 3450
    },
    {
      "epoch": 26.615384615384617,
      "grad_norm": 0.0020101340487599373,
      "learning_rate": 9.356923076923078e-06,
      "loss": 0.0001,
      "step": 3460
    },
    {
      "epoch": 26.692307692307693,
      "grad_norm": 0.02414599433541298,
      "learning_rate": 9.326153846153848e-06,
      "loss": 0.1481,
      "step": 3470
    },
    {
      "epoch": 26.76923076923077,
      "grad_norm": 0.016970224678516388,
      "learning_rate": 9.295384615384615e-06,
      "loss": 0.0005,
      "step": 3480
    },
    {
      "epoch": 26.846153846153847,
      "grad_norm": 0.1080080196261406,
      "learning_rate": 9.264615384615385e-06,
      "loss": 0.0005,
      "step": 3490
    },
    {
      "epoch": 26.923076923076923,
      "grad_norm": 1.719024658203125,
      "learning_rate": 9.233846153846154e-06,
      "loss": 0.0003,
      "step": 3500
    },
    {
      "epoch": 27.0,
      "grad_norm": 0.0033194362185895443,
      "learning_rate": 9.203076923076924e-06,
      "loss": 0.0482,
      "step": 3510
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.08730395138263702,
      "eval_runtime": 0.5625,
      "eval_samples_per_second": 236.442,
      "eval_steps_per_second": 30.222,
      "step": 3510
    },
    {
      "epoch": 27.076923076923077,
      "grad_norm": 0.0017070352332666516,
      "learning_rate": 9.172307692307693e-06,
      "loss": 0.0003,
      "step": 3520
    },
    {
      "epoch": 27.153846153846153,
      "grad_norm": 0.0074979993514716625,
      "learning_rate": 9.141538461538462e-06,
      "loss": 0.0868,
      "step": 3530
    },
    {
      "epoch": 27.23076923076923,
      "grad_norm": 0.001379862311296165,
      "learning_rate": 9.110769230769232e-06,
      "loss": 0.1036,
      "step": 3540
    },
    {
      "epoch": 27.307692307692307,
      "grad_norm": 0.02854689210653305,
      "learning_rate": 9.080000000000001e-06,
      "loss": 0.0552,
      "step": 3550
    },
    {
      "epoch": 27.384615384615383,
      "grad_norm": 0.012071643024682999,
      "learning_rate": 9.04923076923077e-06,
      "loss": 0.005,
      "step": 3560
    },
    {
      "epoch": 27.46153846153846,
      "grad_norm": 0.0024255982134491205,
      "learning_rate": 9.01846153846154e-06,
      "loss": 0.0001,
      "step": 3570
    },
    {
      "epoch": 27.53846153846154,
      "grad_norm": 0.03879780322313309,
      "learning_rate": 8.987692307692308e-06,
      "loss": 0.0879,
      "step": 3580
    },
    {
      "epoch": 27.615384615384617,
      "grad_norm": 54.51979064941406,
      "learning_rate": 8.956923076923077e-06,
      "loss": 0.1568,
      "step": 3590
    },
    {
      "epoch": 27.692307692307693,
      "grad_norm": 0.0009031574008986354,
      "learning_rate": 8.926153846153846e-06,
      "loss": 0.1377,
      "step": 3600
    },
    {
      "epoch": 27.76923076923077,
      "grad_norm": 68.2950439453125,
      "learning_rate": 8.895384615384616e-06,
      "loss": 0.2338,
      "step": 3610
    },
    {
      "epoch": 27.846153846153847,
      "grad_norm": 33.0212287902832,
      "learning_rate": 8.864615384615385e-06,
      "loss": 0.0164,
      "step": 3620
    },
    {
      "epoch": 27.923076923076923,
      "grad_norm": 0.076235830783844,
      "learning_rate": 8.833846153846155e-06,
      "loss": 0.0296,
      "step": 3630
    },
    {
      "epoch": 28.0,
      "grad_norm": 0.014614281244575977,
      "learning_rate": 8.803076923076924e-06,
      "loss": 0.0001,
      "step": 3640
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.07213037461042404,
      "eval_runtime": 0.5725,
      "eval_samples_per_second": 232.306,
      "eval_steps_per_second": 29.693,
      "step": 3640
    },
    {
      "epoch": 28.076923076923077,
      "grad_norm": 24.32257843017578,
      "learning_rate": 8.772307692307693e-06,
      "loss": 0.1775,
      "step": 3650
    },
    {
      "epoch": 28.153846153846153,
      "grad_norm": 0.010675832629203796,
      "learning_rate": 8.741538461538463e-06,
      "loss": 0.0307,
      "step": 3660
    },
    {
      "epoch": 28.23076923076923,
      "grad_norm": 0.009055128321051598,
      "learning_rate": 8.710769230769232e-06,
      "loss": 0.0192,
      "step": 3670
    },
    {
      "epoch": 28.307692307692307,
      "grad_norm": 0.005634867586195469,
      "learning_rate": 8.68e-06,
      "loss": 0.0775,
      "step": 3680
    },
    {
      "epoch": 28.384615384615383,
      "grad_norm": 16.309144973754883,
      "learning_rate": 8.64923076923077e-06,
      "loss": 0.0635,
      "step": 3690
    },
    {
      "epoch": 28.46153846153846,
      "grad_norm": 4.859419345855713,
      "learning_rate": 8.618461538461539e-06,
      "loss": 0.0248,
      "step": 3700
    },
    {
      "epoch": 28.53846153846154,
      "grad_norm": 0.04865121841430664,
      "learning_rate": 8.587692307692308e-06,
      "loss": 0.0002,
      "step": 3710
    },
    {
      "epoch": 28.615384615384617,
      "grad_norm": 0.009109235368669033,
      "learning_rate": 8.556923076923077e-06,
      "loss": 0.0211,
      "step": 3720
    },
    {
      "epoch": 28.692307692307693,
      "grad_norm": 0.005094151012599468,
      "learning_rate": 8.526153846153847e-06,
      "loss": 0.0187,
      "step": 3730
    },
    {
      "epoch": 28.76923076923077,
      "grad_norm": 95.44650268554688,
      "learning_rate": 8.495384615384616e-06,
      "loss": 0.1308,
      "step": 3740
    },
    {
      "epoch": 28.846153846153847,
      "grad_norm": 0.008243425749242306,
      "learning_rate": 8.464615384615386e-06,
      "loss": 0.0006,
      "step": 3750
    },
    {
      "epoch": 28.923076923076923,
      "grad_norm": 0.008599749766290188,
      "learning_rate": 8.433846153846155e-06,
      "loss": 0.0206,
      "step": 3760
    },
    {
      "epoch": 29.0,
      "grad_norm": 0.0015361636178568006,
      "learning_rate": 8.403076923076924e-06,
      "loss": 0.0954,
      "step": 3770
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.014307127334177494,
      "eval_runtime": 0.6444,
      "eval_samples_per_second": 206.4,
      "eval_steps_per_second": 26.382,
      "step": 3770
    },
    {
      "epoch": 29.076923076923077,
      "grad_norm": 0.003046586876735091,
      "learning_rate": 8.372307692307692e-06,
      "loss": 0.0145,
      "step": 3780
    },
    {
      "epoch": 29.153846153846153,
      "grad_norm": 0.0056587024591863155,
      "learning_rate": 8.341538461538462e-06,
      "loss": 0.0025,
      "step": 3790
    },
    {
      "epoch": 29.23076923076923,
      "grad_norm": 0.0009685749537311494,
      "learning_rate": 8.310769230769231e-06,
      "loss": 0.0664,
      "step": 3800
    },
    {
      "epoch": 29.307692307692307,
      "grad_norm": 0.0018060127040371299,
      "learning_rate": 8.28e-06,
      "loss": 0.114,
      "step": 3810
    },
    {
      "epoch": 29.384615384615383,
      "grad_norm": 0.007853769697248936,
      "learning_rate": 8.24923076923077e-06,
      "loss": 0.0595,
      "step": 3820
    },
    {
      "epoch": 29.46153846153846,
      "grad_norm": 0.0192034263163805,
      "learning_rate": 8.218461538461539e-06,
      "loss": 0.0002,
      "step": 3830
    },
    {
      "epoch": 29.53846153846154,
      "grad_norm": 0.0010173041373491287,
      "learning_rate": 8.187692307692309e-06,
      "loss": 0.0342,
      "step": 3840
    },
    {
      "epoch": 29.615384615384617,
      "grad_norm": 0.0016086545074358582,
      "learning_rate": 8.156923076923078e-06,
      "loss": 0.0524,
      "step": 3850
    },
    {
      "epoch": 29.692307692307693,
      "grad_norm": 3.9697511196136475,
      "learning_rate": 8.126153846153847e-06,
      "loss": 0.0973,
      "step": 3860
    },
    {
      "epoch": 29.76923076923077,
      "grad_norm": 0.005043383222073317,
      "learning_rate": 8.095384615384617e-06,
      "loss": 0.001,
      "step": 3870
    },
    {
      "epoch": 29.846153846153847,
      "grad_norm": 0.004252856131643057,
      "learning_rate": 8.064615384615384e-06,
      "loss": 0.0,
      "step": 3880
    },
    {
      "epoch": 29.923076923076923,
      "grad_norm": 0.001231856644153595,
      "learning_rate": 8.033846153846154e-06,
      "loss": 0.0011,
      "step": 3890
    },
    {
      "epoch": 30.0,
      "grad_norm": 0.01130798552185297,
      "learning_rate": 8.003076923076923e-06,
      "loss": 0.1373,
      "step": 3900
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.04489622637629509,
      "eval_runtime": 0.5598,
      "eval_samples_per_second": 237.587,
      "eval_steps_per_second": 30.368,
      "step": 3900
    },
    {
      "epoch": 30.076923076923077,
      "grad_norm": 0.9912404417991638,
      "learning_rate": 7.972307692307693e-06,
      "loss": 0.0347,
      "step": 3910
    },
    {
      "epoch": 30.153846153846153,
      "grad_norm": 0.0017375171883031726,
      "learning_rate": 7.941538461538462e-06,
      "loss": 0.0016,
      "step": 3920
    },
    {
      "epoch": 30.23076923076923,
      "grad_norm": 124.9094009399414,
      "learning_rate": 7.910769230769231e-06,
      "loss": 0.0308,
      "step": 3930
    },
    {
      "epoch": 30.307692307692307,
      "grad_norm": 0.012814003974199295,
      "learning_rate": 7.88e-06,
      "loss": 0.0045,
      "step": 3940
    },
    {
      "epoch": 30.384615384615383,
      "grad_norm": 0.0006594728329218924,
      "learning_rate": 7.84923076923077e-06,
      "loss": 0.0,
      "step": 3950
    },
    {
      "epoch": 30.46153846153846,
      "grad_norm": 0.002067841589450836,
      "learning_rate": 7.81846153846154e-06,
      "loss": 0.0003,
      "step": 3960
    },
    {
      "epoch": 30.53846153846154,
      "grad_norm": 0.007349399384111166,
      "learning_rate": 7.787692307692309e-06,
      "loss": 0.0139,
      "step": 3970
    },
    {
      "epoch": 30.615384615384617,
      "grad_norm": 0.021420804783701897,
      "learning_rate": 7.756923076923077e-06,
      "loss": 0.1997,
      "step": 3980
    },
    {
      "epoch": 30.692307692307693,
      "grad_norm": 0.001424902817234397,
      "learning_rate": 7.726153846153846e-06,
      "loss": 0.032,
      "step": 3990
    },
    {
      "epoch": 30.76923076923077,
      "grad_norm": 2.7047202587127686,
      "learning_rate": 7.695384615384615e-06,
      "loss": 0.0031,
      "step": 4000
    },
    {
      "epoch": 30.846153846153847,
      "grad_norm": 0.0034344890154898167,
      "learning_rate": 7.664615384615385e-06,
      "loss": 0.0517,
      "step": 4010
    },
    {
      "epoch": 30.923076923076923,
      "grad_norm": 0.020475352182984352,
      "learning_rate": 7.633846153846154e-06,
      "loss": 0.0082,
      "step": 4020
    },
    {
      "epoch": 31.0,
      "grad_norm": 0.0006407782784663141,
      "learning_rate": 7.6030769230769245e-06,
      "loss": 0.0076,
      "step": 4030
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.043454647064208984,
      "eval_runtime": 0.5607,
      "eval_samples_per_second": 237.216,
      "eval_steps_per_second": 30.321,
      "step": 4030
    },
    {
      "epoch": 31.076923076923077,
      "grad_norm": 0.0006706058629788458,
      "learning_rate": 7.572307692307693e-06,
      "loss": 0.0174,
      "step": 4040
    },
    {
      "epoch": 31.153846153846153,
      "grad_norm": 0.002210884587839246,
      "learning_rate": 7.5415384615384624e-06,
      "loss": 0.2683,
      "step": 4050
    },
    {
      "epoch": 31.23076923076923,
      "grad_norm": 0.006247711833566427,
      "learning_rate": 7.510769230769232e-06,
      "loss": 0.1214,
      "step": 4060
    },
    {
      "epoch": 31.307692307692307,
      "grad_norm": 0.0019206746947020292,
      "learning_rate": 7.48e-06,
      "loss": 0.0002,
      "step": 4070
    },
    {
      "epoch": 31.384615384615383,
      "grad_norm": 0.0009853794472292066,
      "learning_rate": 7.44923076923077e-06,
      "loss": 0.1271,
      "step": 4080
    },
    {
      "epoch": 31.46153846153846,
      "grad_norm": 0.01296025887131691,
      "learning_rate": 7.418461538461539e-06,
      "loss": 0.0008,
      "step": 4090
    },
    {
      "epoch": 31.53846153846154,
      "grad_norm": 0.009520775638520718,
      "learning_rate": 7.387692307692308e-06,
      "loss": 0.008,
      "step": 4100
    },
    {
      "epoch": 31.615384615384617,
      "grad_norm": 11.869040489196777,
      "learning_rate": 7.356923076923077e-06,
      "loss": 0.0863,
      "step": 4110
    },
    {
      "epoch": 31.692307692307693,
      "grad_norm": 0.016930609941482544,
      "learning_rate": 7.326153846153847e-06,
      "loss": 0.0001,
      "step": 4120
    },
    {
      "epoch": 31.76923076923077,
      "grad_norm": 0.005094442516565323,
      "learning_rate": 7.295384615384617e-06,
      "loss": 0.133,
      "step": 4130
    },
    {
      "epoch": 31.846153846153847,
      "grad_norm": 0.0016887628007680178,
      "learning_rate": 7.264615384615385e-06,
      "loss": 0.0194,
      "step": 4140
    },
    {
      "epoch": 31.923076923076923,
      "grad_norm": 0.004818159621208906,
      "learning_rate": 7.233846153846155e-06,
      "loss": 0.0,
      "step": 4150
    },
    {
      "epoch": 32.0,
      "grad_norm": 0.0007881343481130898,
      "learning_rate": 7.203076923076924e-06,
      "loss": 0.0028,
      "step": 4160
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.01007868628948927,
      "eval_runtime": 0.592,
      "eval_samples_per_second": 224.646,
      "eval_steps_per_second": 28.714,
      "step": 4160
    },
    {
      "epoch": 32.07692307692308,
      "grad_norm": 0.0010091810254380107,
      "learning_rate": 7.172307692307693e-06,
      "loss": 0.0012,
      "step": 4170
    },
    {
      "epoch": 32.15384615384615,
      "grad_norm": 97.89594268798828,
      "learning_rate": 7.141538461538462e-06,
      "loss": 0.1107,
      "step": 4180
    },
    {
      "epoch": 32.23076923076923,
      "grad_norm": 47.61814498901367,
      "learning_rate": 7.1107692307692314e-06,
      "loss": 0.0909,
      "step": 4190
    },
    {
      "epoch": 32.30769230769231,
      "grad_norm": 0.0031400700099766254,
      "learning_rate": 7.08e-06,
      "loss": 0.001,
      "step": 4200
    },
    {
      "epoch": 32.38461538461539,
      "grad_norm": 0.08583249896764755,
      "learning_rate": 7.049230769230769e-06,
      "loss": 0.0001,
      "step": 4210
    },
    {
      "epoch": 32.46153846153846,
      "grad_norm": 0.001287147868424654,
      "learning_rate": 7.01846153846154e-06,
      "loss": 0.0694,
      "step": 4220
    },
    {
      "epoch": 32.53846153846154,
      "grad_norm": 0.0018081064336001873,
      "learning_rate": 6.987692307692309e-06,
      "loss": 0.0042,
      "step": 4230
    },
    {
      "epoch": 32.61538461538461,
      "grad_norm": 0.0005175307160243392,
      "learning_rate": 6.9569230769230776e-06,
      "loss": 0.0209,
      "step": 4240
    },
    {
      "epoch": 32.69230769230769,
      "grad_norm": 0.0015608895337209105,
      "learning_rate": 6.926153846153847e-06,
      "loss": 0.0,
      "step": 4250
    },
    {
      "epoch": 32.76923076923077,
      "grad_norm": 0.0004342070023994893,
      "learning_rate": 6.895384615384616e-06,
      "loss": 0.0,
      "step": 4260
    },
    {
      "epoch": 32.84615384615385,
      "grad_norm": 0.12884479761123657,
      "learning_rate": 6.864615384615385e-06,
      "loss": 0.0001,
      "step": 4270
    },
    {
      "epoch": 32.92307692307692,
      "grad_norm": 0.0009400132112205029,
      "learning_rate": 6.833846153846154e-06,
      "loss": 0.0045,
      "step": 4280
    },
    {
      "epoch": 33.0,
      "grad_norm": 0.0006513062980957329,
      "learning_rate": 6.803076923076924e-06,
      "loss": 0.0001,
      "step": 4290
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.04143361374735832,
      "eval_runtime": 0.5612,
      "eval_samples_per_second": 237.01,
      "eval_steps_per_second": 30.294,
      "step": 4290
    },
    {
      "epoch": 33.07692307692308,
      "grad_norm": 0.0024744998663663864,
      "learning_rate": 6.772307692307692e-06,
      "loss": 0.0893,
      "step": 4300
    },
    {
      "epoch": 33.15384615384615,
      "grad_norm": 0.002017271937802434,
      "learning_rate": 6.741538461538462e-06,
      "loss": 0.0002,
      "step": 4310
    },
    {
      "epoch": 33.23076923076923,
      "grad_norm": 0.03144347295165062,
      "learning_rate": 6.710769230769232e-06,
      "loss": 0.0002,
      "step": 4320
    },
    {
      "epoch": 33.30769230769231,
      "grad_norm": 0.014736099168658257,
      "learning_rate": 6.680000000000001e-06,
      "loss": 0.0,
      "step": 4330
    },
    {
      "epoch": 33.38461538461539,
      "grad_norm": 0.0004188344464637339,
      "learning_rate": 6.64923076923077e-06,
      "loss": 0.0,
      "step": 4340
    },
    {
      "epoch": 33.46153846153846,
      "grad_norm": 0.08335867524147034,
      "learning_rate": 6.618461538461539e-06,
      "loss": 0.0001,
      "step": 4350
    },
    {
      "epoch": 33.53846153846154,
      "grad_norm": 0.0008053340134210885,
      "learning_rate": 6.587692307692309e-06,
      "loss": 0.0003,
      "step": 4360
    },
    {
      "epoch": 33.61538461538461,
      "grad_norm": 0.0012076394632458687,
      "learning_rate": 6.556923076923077e-06,
      "loss": 0.0001,
      "step": 4370
    },
    {
      "epoch": 33.69230769230769,
      "grad_norm": 0.0048568369820714,
      "learning_rate": 6.5261538461538465e-06,
      "loss": 0.0,
      "step": 4380
    },
    {
      "epoch": 33.76923076923077,
      "grad_norm": 0.0008632199605926871,
      "learning_rate": 6.495384615384616e-06,
      "loss": 0.0002,
      "step": 4390
    },
    {
      "epoch": 33.84615384615385,
      "grad_norm": 0.049176134169101715,
      "learning_rate": 6.4646153846153845e-06,
      "loss": 0.0001,
      "step": 4400
    },
    {
      "epoch": 33.92307692307692,
      "grad_norm": 0.00043606868712231517,
      "learning_rate": 6.433846153846154e-06,
      "loss": 0.0682,
      "step": 4410
    },
    {
      "epoch": 34.0,
      "grad_norm": 0.0006029644282534719,
      "learning_rate": 6.403076923076924e-06,
      "loss": 0.001,
      "step": 4420
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.0017387220868840814,
      "eval_runtime": 0.5604,
      "eval_samples_per_second": 237.329,
      "eval_steps_per_second": 30.335,
      "step": 4420
    },
    {
      "epoch": 34.07692307692308,
      "grad_norm": 0.0005392258754000068,
      "learning_rate": 6.3723076923076935e-06,
      "loss": 0.0178,
      "step": 4430
    },
    {
      "epoch": 34.15384615384615,
      "grad_norm": 0.0017064090352505445,
      "learning_rate": 6.341538461538462e-06,
      "loss": 0.007,
      "step": 4440
    },
    {
      "epoch": 34.23076923076923,
      "grad_norm": 0.0033292374573647976,
      "learning_rate": 6.3107692307692315e-06,
      "loss": 0.007,
      "step": 4450
    },
    {
      "epoch": 34.30769230769231,
      "grad_norm": 31.601030349731445,
      "learning_rate": 6.280000000000001e-06,
      "loss": 0.0654,
      "step": 4460
    },
    {
      "epoch": 34.38461538461539,
      "grad_norm": 0.007135252468287945,
      "learning_rate": 6.249230769230769e-06,
      "loss": 0.0,
      "step": 4470
    },
    {
      "epoch": 34.46153846153846,
      "grad_norm": 81.1984634399414,
      "learning_rate": 6.218461538461539e-06,
      "loss": 0.072,
      "step": 4480
    },
    {
      "epoch": 34.53846153846154,
      "grad_norm": 0.06208951398730278,
      "learning_rate": 6.187692307692308e-06,
      "loss": 0.0,
      "step": 4490
    },
    {
      "epoch": 34.61538461538461,
      "grad_norm": 0.0007352761458605528,
      "learning_rate": 6.156923076923077e-06,
      "loss": 0.1805,
      "step": 4500
    },
    {
      "epoch": 34.69230769230769,
      "grad_norm": 0.00034046368091367185,
      "learning_rate": 6.126153846153846e-06,
      "loss": 0.0092,
      "step": 4510
    },
    {
      "epoch": 34.76923076923077,
      "grad_norm": 0.0068006995134055614,
      "learning_rate": 6.095384615384616e-06,
      "loss": 0.0135,
      "step": 4520
    },
    {
      "epoch": 34.84615384615385,
      "grad_norm": 0.009285767562687397,
      "learning_rate": 6.064615384615386e-06,
      "loss": 0.0443,
      "step": 4530
    },
    {
      "epoch": 34.92307692307692,
      "grad_norm": 0.0005047802696935833,
      "learning_rate": 6.033846153846154e-06,
      "loss": 0.0003,
      "step": 4540
    },
    {
      "epoch": 35.0,
      "grad_norm": 0.0017190852668136358,
      "learning_rate": 6.003076923076924e-06,
      "loss": 0.0055,
      "step": 4550
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.07332061976194382,
      "eval_runtime": 0.5797,
      "eval_samples_per_second": 229.424,
      "eval_steps_per_second": 29.325,
      "step": 4550
    },
    {
      "epoch": 35.07692307692308,
      "grad_norm": 0.014129632152616978,
      "learning_rate": 5.972307692307693e-06,
      "loss": 0.0044,
      "step": 4560
    },
    {
      "epoch": 35.15384615384615,
      "grad_norm": 0.051782090216875076,
      "learning_rate": 5.941538461538462e-06,
      "loss": 0.1579,
      "step": 4570
    },
    {
      "epoch": 35.23076923076923,
      "grad_norm": 0.0010272455401718616,
      "learning_rate": 5.910769230769231e-06,
      "loss": 0.0,
      "step": 4580
    },
    {
      "epoch": 35.30769230769231,
      "grad_norm": 0.0008460541139356792,
      "learning_rate": 5.8800000000000005e-06,
      "loss": 0.0799,
      "step": 4590
    },
    {
      "epoch": 35.38461538461539,
      "grad_norm": 0.0015528828371316195,
      "learning_rate": 5.849230769230769e-06,
      "loss": 0.0585,
      "step": 4600
    },
    {
      "epoch": 35.46153846153846,
      "grad_norm": 0.0016128149582073092,
      "learning_rate": 5.818461538461538e-06,
      "loss": 0.0001,
      "step": 4610
    },
    {
      "epoch": 35.53846153846154,
      "grad_norm": 0.004083054140210152,
      "learning_rate": 5.787692307692309e-06,
      "loss": 0.0001,
      "step": 4620
    },
    {
      "epoch": 35.61538461538461,
      "grad_norm": 0.0318060964345932,
      "learning_rate": 5.756923076923078e-06,
      "loss": 0.0001,
      "step": 4630
    },
    {
      "epoch": 35.69230769230769,
      "grad_norm": 1.8864119052886963,
      "learning_rate": 5.726153846153847e-06,
      "loss": 0.0008,
      "step": 4640
    },
    {
      "epoch": 35.76923076923077,
      "grad_norm": 0.0009333742200396955,
      "learning_rate": 5.695384615384616e-06,
      "loss": 0.0028,
      "step": 4650
    },
    {
      "epoch": 35.84615384615385,
      "grad_norm": 0.0018823591526597738,
      "learning_rate": 5.664615384615385e-06,
      "loss": 0.0,
      "step": 4660
    },
    {
      "epoch": 35.92307692307692,
      "grad_norm": 0.014191100373864174,
      "learning_rate": 5.633846153846154e-06,
      "loss": 0.0064,
      "step": 4670
    },
    {
      "epoch": 36.0,
      "grad_norm": 0.001580247189849615,
      "learning_rate": 5.603076923076923e-06,
      "loss": 0.1471,
      "step": 4680
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.9774436090225563,
      "eval_loss": 0.12211022526025772,
      "eval_runtime": 0.56,
      "eval_samples_per_second": 237.504,
      "eval_steps_per_second": 30.358,
      "step": 4680
    },
    {
      "epoch": 36.07692307692308,
      "grad_norm": 0.0018762449035421014,
      "learning_rate": 5.572307692307693e-06,
      "loss": 0.0,
      "step": 4690
    },
    {
      "epoch": 36.15384615384615,
      "grad_norm": 0.000685960752889514,
      "learning_rate": 5.541538461538461e-06,
      "loss": 0.0867,
      "step": 4700
    },
    {
      "epoch": 36.23076923076923,
      "grad_norm": 0.001832542009651661,
      "learning_rate": 5.5107692307692315e-06,
      "loss": 0.0831,
      "step": 4710
    },
    {
      "epoch": 36.30769230769231,
      "grad_norm": 1.0698580741882324,
      "learning_rate": 5.480000000000001e-06,
      "loss": 0.0027,
      "step": 4720
    },
    {
      "epoch": 36.38461538461539,
      "grad_norm": 0.002061473438516259,
      "learning_rate": 5.44923076923077e-06,
      "loss": 0.0,
      "step": 4730
    },
    {
      "epoch": 36.46153846153846,
      "grad_norm": 0.0007437376189045608,
      "learning_rate": 5.418461538461539e-06,
      "loss": 0.0002,
      "step": 4740
    },
    {
      "epoch": 36.53846153846154,
      "grad_norm": 0.00041431974386796355,
      "learning_rate": 5.387692307692308e-06,
      "loss": 0.0002,
      "step": 4750
    },
    {
      "epoch": 36.61538461538461,
      "grad_norm": 1.0955127477645874,
      "learning_rate": 5.356923076923078e-06,
      "loss": 0.0002,
      "step": 4760
    },
    {
      "epoch": 36.69230769230769,
      "grad_norm": 0.15827512741088867,
      "learning_rate": 5.326153846153846e-06,
      "loss": 0.0002,
      "step": 4770
    },
    {
      "epoch": 36.76923076923077,
      "grad_norm": 0.0007032826542854309,
      "learning_rate": 5.2953846153846156e-06,
      "loss": 0.0033,
      "step": 4780
    },
    {
      "epoch": 36.84615384615385,
      "grad_norm": 0.00039768809801898897,
      "learning_rate": 5.264615384615385e-06,
      "loss": 0.0028,
      "step": 4790
    },
    {
      "epoch": 36.92307692307692,
      "grad_norm": 0.002939037047326565,
      "learning_rate": 5.2338461538461535e-06,
      "loss": 0.0938,
      "step": 4800
    },
    {
      "epoch": 37.0,
      "grad_norm": 0.06541663408279419,
      "learning_rate": 5.203076923076924e-06,
      "loss": 0.0484,
      "step": 4810
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.14731426537036896,
      "eval_runtime": 0.5828,
      "eval_samples_per_second": 228.19,
      "eval_steps_per_second": 29.167,
      "step": 4810
    },
    {
      "epoch": 37.07692307692308,
      "grad_norm": 0.9744657874107361,
      "learning_rate": 5.172307692307693e-06,
      "loss": 0.0206,
      "step": 4820
    },
    {
      "epoch": 37.15384615384615,
      "grad_norm": 27.0574893951416,
      "learning_rate": 5.1415384615384625e-06,
      "loss": 0.0719,
      "step": 4830
    },
    {
      "epoch": 37.23076923076923,
      "grad_norm": 0.029181215912103653,
      "learning_rate": 5.110769230769231e-06,
      "loss": 0.0002,
      "step": 4840
    },
    {
      "epoch": 37.30769230769231,
      "grad_norm": 37.923431396484375,
      "learning_rate": 5.0800000000000005e-06,
      "loss": 0.0878,
      "step": 4850
    },
    {
      "epoch": 37.38461538461539,
      "grad_norm": 0.0028850040398538113,
      "learning_rate": 5.04923076923077e-06,
      "loss": 0.0123,
      "step": 4860
    },
    {
      "epoch": 37.46153846153846,
      "grad_norm": 0.0006236844346858561,
      "learning_rate": 5.0184615384615384e-06,
      "loss": 0.082,
      "step": 4870
    },
    {
      "epoch": 37.53846153846154,
      "grad_norm": 0.0010073287412524223,
      "learning_rate": 4.987692307692308e-06,
      "loss": 0.0027,
      "step": 4880
    },
    {
      "epoch": 37.61538461538461,
      "grad_norm": 0.0008756616152822971,
      "learning_rate": 4.956923076923077e-06,
      "loss": 0.0001,
      "step": 4890
    },
    {
      "epoch": 37.69230769230769,
      "grad_norm": 0.0012337320949882269,
      "learning_rate": 4.926153846153847e-06,
      "loss": 0.0003,
      "step": 4900
    },
    {
      "epoch": 37.76923076923077,
      "grad_norm": 26.195831298828125,
      "learning_rate": 4.895384615384616e-06,
      "loss": 0.0488,
      "step": 4910
    },
    {
      "epoch": 37.84615384615385,
      "grad_norm": 0.002201149705797434,
      "learning_rate": 4.8646153846153846e-06,
      "loss": 0.0,
      "step": 4920
    },
    {
      "epoch": 37.92307692307692,
      "grad_norm": 0.0031797313131392,
      "learning_rate": 4.833846153846154e-06,
      "loss": 0.0,
      "step": 4930
    },
    {
      "epoch": 38.0,
      "grad_norm": 0.001882484881207347,
      "learning_rate": 4.803076923076923e-06,
      "loss": 0.0014,
      "step": 4940
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.07482893019914627,
      "eval_runtime": 0.5718,
      "eval_samples_per_second": 232.616,
      "eval_steps_per_second": 29.733,
      "step": 4940
    },
    {
      "epoch": 38.07692307692308,
      "grad_norm": 0.008910280652344227,
      "learning_rate": 4.772307692307693e-06,
      "loss": 0.0,
      "step": 4950
    },
    {
      "epoch": 38.15384615384615,
      "grad_norm": 0.0004342030733823776,
      "learning_rate": 4.741538461538462e-06,
      "loss": 0.0995,
      "step": 4960
    },
    {
      "epoch": 38.23076923076923,
      "grad_norm": 0.000522003450896591,
      "learning_rate": 4.710769230769231e-06,
      "loss": 0.0,
      "step": 4970
    },
    {
      "epoch": 38.30769230769231,
      "grad_norm": 0.4271712303161621,
      "learning_rate": 4.680000000000001e-06,
      "loss": 0.049,
      "step": 4980
    },
    {
      "epoch": 38.38461538461539,
      "grad_norm": 0.0009867213666439056,
      "learning_rate": 4.6492307692307695e-06,
      "loss": 0.0769,
      "step": 4990
    },
    {
      "epoch": 38.46153846153846,
      "grad_norm": 0.0015043043531477451,
      "learning_rate": 4.618461538461539e-06,
      "loss": 0.0,
      "step": 5000
    },
    {
      "epoch": 38.53846153846154,
      "grad_norm": 0.000573535158764571,
      "learning_rate": 4.587692307692308e-06,
      "loss": 0.0,
      "step": 5010
    },
    {
      "epoch": 38.61538461538461,
      "grad_norm": 0.007249703165143728,
      "learning_rate": 4.556923076923077e-06,
      "loss": 0.0807,
      "step": 5020
    },
    {
      "epoch": 38.69230769230769,
      "grad_norm": 1.4994910955429077,
      "learning_rate": 4.526153846153847e-06,
      "loss": 0.0056,
      "step": 5030
    },
    {
      "epoch": 38.76923076923077,
      "grad_norm": 0.0006923425826244056,
      "learning_rate": 4.495384615384616e-06,
      "loss": 0.0001,
      "step": 5040
    },
    {
      "epoch": 38.84615384615385,
      "grad_norm": 0.010082672350108624,
      "learning_rate": 4.464615384615385e-06,
      "loss": 0.0002,
      "step": 5050
    },
    {
      "epoch": 38.92307692307692,
      "grad_norm": 0.0003873300738632679,
      "learning_rate": 4.433846153846154e-06,
      "loss": 0.0123,
      "step": 5060
    },
    {
      "epoch": 39.0,
      "grad_norm": 387.8450622558594,
      "learning_rate": 4.403076923076923e-06,
      "loss": 0.1825,
      "step": 5070
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.10720350593328476,
      "eval_runtime": 0.558,
      "eval_samples_per_second": 238.346,
      "eval_steps_per_second": 30.465,
      "step": 5070
    },
    {
      "epoch": 39.07692307692308,
      "grad_norm": 0.0003139421751257032,
      "learning_rate": 4.372307692307693e-06,
      "loss": 0.0002,
      "step": 5080
    },
    {
      "epoch": 39.15384615384615,
      "grad_norm": 0.0030731302686035633,
      "learning_rate": 4.341538461538462e-06,
      "loss": 0.0005,
      "step": 5090
    },
    {
      "epoch": 39.23076923076923,
      "grad_norm": 95.7864990234375,
      "learning_rate": 4.310769230769231e-06,
      "loss": 0.1588,
      "step": 5100
    },
    {
      "epoch": 39.30769230769231,
      "grad_norm": 0.045532066375017166,
      "learning_rate": 4.2800000000000005e-06,
      "loss": 0.0003,
      "step": 5110
    },
    {
      "epoch": 39.38461538461539,
      "grad_norm": 5.153738975524902,
      "learning_rate": 4.249230769230769e-06,
      "loss": 0.0012,
      "step": 5120
    },
    {
      "epoch": 39.46153846153846,
      "grad_norm": 0.0010733181843534112,
      "learning_rate": 4.218461538461539e-06,
      "loss": 0.1359,
      "step": 5130
    },
    {
      "epoch": 39.53846153846154,
      "grad_norm": 0.00048660460743121803,
      "learning_rate": 4.187692307692308e-06,
      "loss": 0.0001,
      "step": 5140
    },
    {
      "epoch": 39.61538461538461,
      "grad_norm": 0.13067100942134857,
      "learning_rate": 4.156923076923077e-06,
      "loss": 0.0384,
      "step": 5150
    },
    {
      "epoch": 39.69230769230769,
      "grad_norm": 0.0006858094711787999,
      "learning_rate": 4.126153846153847e-06,
      "loss": 0.0008,
      "step": 5160
    },
    {
      "epoch": 39.76923076923077,
      "grad_norm": 5.363431453704834,
      "learning_rate": 4.095384615384615e-06,
      "loss": 0.0007,
      "step": 5170
    },
    {
      "epoch": 39.84615384615385,
      "grad_norm": 0.0011726649245247245,
      "learning_rate": 4.0646153846153854e-06,
      "loss": 0.0018,
      "step": 5180
    },
    {
      "epoch": 39.92307692307692,
      "grad_norm": 0.21595709025859833,
      "learning_rate": 4.033846153846154e-06,
      "loss": 0.0002,
      "step": 5190
    },
    {
      "epoch": 40.0,
      "grad_norm": 0.0002660917234607041,
      "learning_rate": 4.003076923076923e-06,
      "loss": 0.0,
      "step": 5200
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.06868916749954224,
      "eval_runtime": 0.5588,
      "eval_samples_per_second": 238.025,
      "eval_steps_per_second": 30.424,
      "step": 5200
    },
    {
      "epoch": 40.07692307692308,
      "grad_norm": 0.00026413268642500043,
      "learning_rate": 3.972307692307693e-06,
      "loss": 0.0001,
      "step": 5210
    },
    {
      "epoch": 40.15384615384615,
      "grad_norm": 0.00025818985886871815,
      "learning_rate": 3.941538461538461e-06,
      "loss": 0.0,
      "step": 5220
    },
    {
      "epoch": 40.23076923076923,
      "grad_norm": 0.0034207049757242203,
      "learning_rate": 3.9107692307692316e-06,
      "loss": 0.2677,
      "step": 5230
    },
    {
      "epoch": 40.30769230769231,
      "grad_norm": 22.716407775878906,
      "learning_rate": 3.88e-06,
      "loss": 0.099,
      "step": 5240
    },
    {
      "epoch": 40.38461538461539,
      "grad_norm": 0.001969344448298216,
      "learning_rate": 3.8492307692307695e-06,
      "loss": 0.0,
      "step": 5250
    },
    {
      "epoch": 40.46153846153846,
      "grad_norm": 0.0004705400497186929,
      "learning_rate": 3.818461538461539e-06,
      "loss": 0.0301,
      "step": 5260
    },
    {
      "epoch": 40.53846153846154,
      "grad_norm": 0.0022966957185417414,
      "learning_rate": 3.787692307692308e-06,
      "loss": 0.1202,
      "step": 5270
    },
    {
      "epoch": 40.61538461538461,
      "grad_norm": 0.0005464573623612523,
      "learning_rate": 3.7569230769230773e-06,
      "loss": 0.0,
      "step": 5280
    },
    {
      "epoch": 40.69230769230769,
      "grad_norm": 0.00032725094933994114,
      "learning_rate": 3.7261538461538467e-06,
      "loss": 0.1325,
      "step": 5290
    },
    {
      "epoch": 40.76923076923077,
      "grad_norm": 1.188474416732788,
      "learning_rate": 3.6953846153846156e-06,
      "loss": 0.0002,
      "step": 5300
    },
    {
      "epoch": 40.84615384615385,
      "grad_norm": 0.0010689280461519957,
      "learning_rate": 3.6646153846153846e-06,
      "loss": 0.0522,
      "step": 5310
    },
    {
      "epoch": 40.92307692307692,
      "grad_norm": 0.00040832936065271497,
      "learning_rate": 3.633846153846154e-06,
      "loss": 0.0264,
      "step": 5320
    },
    {
      "epoch": 41.0,
      "grad_norm": 0.0008318335167132318,
      "learning_rate": 3.6030769230769234e-06,
      "loss": 0.0081,
      "step": 5330
    },
    {
      "epoch": 41.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.1147400438785553,
      "eval_runtime": 0.5673,
      "eval_samples_per_second": 234.428,
      "eval_steps_per_second": 29.964,
      "step": 5330
    },
    {
      "epoch": 41.07692307692308,
      "grad_norm": 0.010249706916511059,
      "learning_rate": 3.572307692307693e-06,
      "loss": 0.0003,
      "step": 5340
    },
    {
      "epoch": 41.15384615384615,
      "grad_norm": 0.0003951251273974776,
      "learning_rate": 3.5415384615384618e-06,
      "loss": 0.0082,
      "step": 5350
    },
    {
      "epoch": 41.23076923076923,
      "grad_norm": 0.008920038118958473,
      "learning_rate": 3.5107692307692307e-06,
      "loss": 0.0397,
      "step": 5360
    },
    {
      "epoch": 41.30769230769231,
      "grad_norm": 0.0018836874514818192,
      "learning_rate": 3.48e-06,
      "loss": 0.0,
      "step": 5370
    },
    {
      "epoch": 41.38461538461539,
      "grad_norm": 0.004664571024477482,
      "learning_rate": 3.4492307692307695e-06,
      "loss": 0.1514,
      "step": 5380
    },
    {
      "epoch": 41.46153846153846,
      "grad_norm": 0.0018832337809726596,
      "learning_rate": 3.418461538461539e-06,
      "loss": 0.0476,
      "step": 5390
    },
    {
      "epoch": 41.53846153846154,
      "grad_norm": 0.004951864946633577,
      "learning_rate": 3.387692307692308e-06,
      "loss": 0.0,
      "step": 5400
    },
    {
      "epoch": 41.61538461538461,
      "grad_norm": 0.0004485521058086306,
      "learning_rate": 3.356923076923077e-06,
      "loss": 0.004,
      "step": 5410
    },
    {
      "epoch": 41.69230769230769,
      "grad_norm": 0.0011319700861349702,
      "learning_rate": 3.3261538461538463e-06,
      "loss": 0.0002,
      "step": 5420
    },
    {
      "epoch": 41.76923076923077,
      "grad_norm": 0.1554967314004898,
      "learning_rate": 3.2953846153846157e-06,
      "loss": 0.0001,
      "step": 5430
    },
    {
      "epoch": 41.84615384615385,
      "grad_norm": 0.0006219987408258021,
      "learning_rate": 3.264615384615385e-06,
      "loss": 0.0004,
      "step": 5440
    },
    {
      "epoch": 41.92307692307692,
      "grad_norm": 0.007299738470464945,
      "learning_rate": 3.233846153846154e-06,
      "loss": 0.0001,
      "step": 5450
    },
    {
      "epoch": 42.0,
      "grad_norm": 0.0005407740827649832,
      "learning_rate": 3.203076923076923e-06,
      "loss": 0.0557,
      "step": 5460
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.06299061328172684,
      "eval_runtime": 0.5719,
      "eval_samples_per_second": 232.55,
      "eval_steps_per_second": 29.724,
      "step": 5460
    },
    {
      "epoch": 42.07692307692308,
      "grad_norm": 0.000429264095146209,
      "learning_rate": 3.1723076923076924e-06,
      "loss": 0.0011,
      "step": 5470
    },
    {
      "epoch": 42.15384615384615,
      "grad_norm": 0.0012445776956155896,
      "learning_rate": 3.141538461538462e-06,
      "loss": 0.0215,
      "step": 5480
    },
    {
      "epoch": 42.23076923076923,
      "grad_norm": 0.017115270718932152,
      "learning_rate": 3.110769230769231e-06,
      "loss": 0.0005,
      "step": 5490
    },
    {
      "epoch": 42.30769230769231,
      "grad_norm": 0.0003503753978293389,
      "learning_rate": 3.08e-06,
      "loss": 0.0469,
      "step": 5500
    },
    {
      "epoch": 42.38461538461539,
      "grad_norm": 0.0008315156446769834,
      "learning_rate": 3.049230769230769e-06,
      "loss": 0.0082,
      "step": 5510
    },
    {
      "epoch": 42.46153846153846,
      "grad_norm": 0.0038778912276029587,
      "learning_rate": 3.0184615384615385e-06,
      "loss": 0.0115,
      "step": 5520
    },
    {
      "epoch": 42.53846153846154,
      "grad_norm": 0.0003100531466770917,
      "learning_rate": 2.987692307692308e-06,
      "loss": 0.0091,
      "step": 5530
    },
    {
      "epoch": 42.61538461538461,
      "grad_norm": 0.0004450716369319707,
      "learning_rate": 2.9569230769230773e-06,
      "loss": 0.0042,
      "step": 5540
    },
    {
      "epoch": 42.69230769230769,
      "grad_norm": 0.0008668462396599352,
      "learning_rate": 2.9261538461538463e-06,
      "loss": 0.0002,
      "step": 5550
    },
    {
      "epoch": 42.76923076923077,
      "grad_norm": 0.0007650844636373222,
      "learning_rate": 2.8953846153846153e-06,
      "loss": 0.0019,
      "step": 5560
    },
    {
      "epoch": 42.84615384615385,
      "grad_norm": 0.0027373021002858877,
      "learning_rate": 2.8646153846153847e-06,
      "loss": 0.039,
      "step": 5570
    },
    {
      "epoch": 42.92307692307692,
      "grad_norm": 0.0016651992918923497,
      "learning_rate": 2.833846153846154e-06,
      "loss": 0.0,
      "step": 5580
    },
    {
      "epoch": 43.0,
      "grad_norm": 0.0006611489807255566,
      "learning_rate": 2.8030769230769234e-06,
      "loss": 0.0,
      "step": 5590
    },
    {
      "epoch": 43.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.016152488067746162,
      "eval_runtime": 0.5602,
      "eval_samples_per_second": 237.419,
      "eval_steps_per_second": 30.347,
      "step": 5590
    },
    {
      "epoch": 43.07692307692308,
      "grad_norm": 0.0031338355038315058,
      "learning_rate": 2.7723076923076924e-06,
      "loss": 0.0724,
      "step": 5600
    },
    {
      "epoch": 43.15384615384615,
      "grad_norm": 0.04154682531952858,
      "learning_rate": 2.7415384615384614e-06,
      "loss": 0.0185,
      "step": 5610
    },
    {
      "epoch": 43.23076923076923,
      "grad_norm": 0.0005495785153470933,
      "learning_rate": 2.710769230769231e-06,
      "loss": 0.0,
      "step": 5620
    },
    {
      "epoch": 43.30769230769231,
      "grad_norm": 0.0006096691940911114,
      "learning_rate": 2.68e-06,
      "loss": 0.1738,
      "step": 5630
    },
    {
      "epoch": 43.38461538461539,
      "grad_norm": 0.0008338314946740866,
      "learning_rate": 2.6492307692307696e-06,
      "loss": 0.0,
      "step": 5640
    },
    {
      "epoch": 43.46153846153846,
      "grad_norm": 0.004199822433292866,
      "learning_rate": 2.6184615384615385e-06,
      "loss": 0.0914,
      "step": 5650
    },
    {
      "epoch": 43.53846153846154,
      "grad_norm": 0.000944766215980053,
      "learning_rate": 2.587692307692308e-06,
      "loss": 0.1093,
      "step": 5660
    },
    {
      "epoch": 43.61538461538461,
      "grad_norm": 48.179908752441406,
      "learning_rate": 2.5569230769230773e-06,
      "loss": 0.0042,
      "step": 5670
    },
    {
      "epoch": 43.69230769230769,
      "grad_norm": 0.006739782635122538,
      "learning_rate": 2.5261538461538463e-06,
      "loss": 0.0,
      "step": 5680
    },
    {
      "epoch": 43.76923076923077,
      "grad_norm": 0.0064003062434494495,
      "learning_rate": 2.4953846153846157e-06,
      "loss": 0.0,
      "step": 5690
    },
    {
      "epoch": 43.84615384615385,
      "grad_norm": 0.0006233023013919592,
      "learning_rate": 2.4646153846153847e-06,
      "loss": 0.0237,
      "step": 5700
    },
    {
      "epoch": 43.92307692307692,
      "grad_norm": 0.0021808522287756205,
      "learning_rate": 2.433846153846154e-06,
      "loss": 0.0515,
      "step": 5710
    },
    {
      "epoch": 44.0,
      "grad_norm": 0.002309564733877778,
      "learning_rate": 2.4030769230769235e-06,
      "loss": 0.0,
      "step": 5720
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.9924812030075187,
      "eval_loss": 0.04631290212273598,
      "eval_runtime": 0.5655,
      "eval_samples_per_second": 235.178,
      "eval_steps_per_second": 30.06,
      "step": 5720
    },
    {
      "epoch": 44.07692307692308,
      "grad_norm": 0.0009078089497052133,
      "learning_rate": 2.3723076923076924e-06,
      "loss": 0.0,
      "step": 5730
    },
    {
      "epoch": 44.15384615384615,
      "grad_norm": 0.0005848463624715805,
      "learning_rate": 2.341538461538462e-06,
      "loss": 0.0001,
      "step": 5740
    },
    {
      "epoch": 44.23076923076923,
      "grad_norm": 0.0006252095336094499,
      "learning_rate": 2.310769230769231e-06,
      "loss": 0.0003,
      "step": 5750
    },
    {
      "epoch": 44.30769230769231,
      "grad_norm": 0.0007332940585911274,
      "learning_rate": 2.28e-06,
      "loss": 0.1749,
      "step": 5760
    },
    {
      "epoch": 44.38461538461539,
      "grad_norm": 0.002333400072529912,
      "learning_rate": 2.2492307692307696e-06,
      "loss": 0.0,
      "step": 5770
    },
    {
      "epoch": 44.46153846153846,
      "grad_norm": 0.01608298532664776,
      "learning_rate": 2.218461538461539e-06,
      "loss": 0.0,
      "step": 5780
    },
    {
      "epoch": 44.53846153846154,
      "grad_norm": 0.0004494412860367447,
      "learning_rate": 2.187692307692308e-06,
      "loss": 0.0037,
      "step": 5790
    },
    {
      "epoch": 44.61538461538461,
      "grad_norm": 0.0018300774972885847,
      "learning_rate": 2.156923076923077e-06,
      "loss": 0.103,
      "step": 5800
    },
    {
      "epoch": 44.69230769230769,
      "grad_norm": 0.0017312734853476286,
      "learning_rate": 2.1261538461538463e-06,
      "loss": 0.0001,
      "step": 5810
    },
    {
      "epoch": 44.76923076923077,
      "grad_norm": 0.0011130099883303046,
      "learning_rate": 2.0953846153846157e-06,
      "loss": 0.0001,
      "step": 5820
    },
    {
      "epoch": 44.84615384615385,
      "grad_norm": 0.0020217066630721092,
      "learning_rate": 2.064615384615385e-06,
      "loss": 0.0324,
      "step": 5830
    },
    {
      "epoch": 44.92307692307692,
      "grad_norm": 0.0037438569124788046,
      "learning_rate": 2.033846153846154e-06,
      "loss": 0.0038,
      "step": 5840
    },
    {
      "epoch": 45.0,
      "grad_norm": 0.01541079394519329,
      "learning_rate": 2.003076923076923e-06,
      "loss": 0.0197,
      "step": 5850
    },
    {
      "epoch": 45.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.07567540556192398,
      "eval_runtime": 0.5602,
      "eval_samples_per_second": 237.419,
      "eval_steps_per_second": 30.347,
      "step": 5850
    },
    {
      "epoch": 45.07692307692308,
      "grad_norm": 0.0009596956660971045,
      "learning_rate": 1.9723076923076924e-06,
      "loss": 0.0751,
      "step": 5860
    },
    {
      "epoch": 45.15384615384615,
      "grad_norm": 0.0015320915263146162,
      "learning_rate": 1.941538461538462e-06,
      "loss": 0.0,
      "step": 5870
    },
    {
      "epoch": 45.23076923076923,
      "grad_norm": 8.862515449523926,
      "learning_rate": 1.9107692307692312e-06,
      "loss": 0.0028,
      "step": 5880
    },
    {
      "epoch": 45.30769230769231,
      "grad_norm": 0.00041210712515749037,
      "learning_rate": 1.8800000000000002e-06,
      "loss": 0.0004,
      "step": 5890
    },
    {
      "epoch": 45.38461538461539,
      "grad_norm": 0.000413590605603531,
      "learning_rate": 1.8492307692307692e-06,
      "loss": 0.0344,
      "step": 5900
    },
    {
      "epoch": 45.46153846153846,
      "grad_norm": 14.184650421142578,
      "learning_rate": 1.8184615384615386e-06,
      "loss": 0.0512,
      "step": 5910
    },
    {
      "epoch": 45.53846153846154,
      "grad_norm": 0.025255808606743813,
      "learning_rate": 1.7876923076923078e-06,
      "loss": 0.176,
      "step": 5920
    },
    {
      "epoch": 45.61538461538461,
      "grad_norm": 0.021409081295132637,
      "learning_rate": 1.7569230769230772e-06,
      "loss": 0.0001,
      "step": 5930
    },
    {
      "epoch": 45.69230769230769,
      "grad_norm": 0.0009863406885415316,
      "learning_rate": 1.7261538461538463e-06,
      "loss": 0.0001,
      "step": 5940
    },
    {
      "epoch": 45.76923076923077,
      "grad_norm": 4.161138534545898,
      "learning_rate": 1.6953846153846153e-06,
      "loss": 0.0021,
      "step": 5950
    },
    {
      "epoch": 45.84615384615385,
      "grad_norm": 0.0008297639433294535,
      "learning_rate": 1.6646153846153847e-06,
      "loss": 0.0001,
      "step": 5960
    },
    {
      "epoch": 45.92307692307692,
      "grad_norm": 0.0011634193360805511,
      "learning_rate": 1.6338461538461539e-06,
      "loss": 0.0,
      "step": 5970
    },
    {
      "epoch": 46.0,
      "grad_norm": 0.002662894781678915,
      "learning_rate": 1.6030769230769233e-06,
      "loss": 0.1442,
      "step": 5980
    },
    {
      "epoch": 46.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.09411457926034927,
      "eval_runtime": 0.5682,
      "eval_samples_per_second": 234.062,
      "eval_steps_per_second": 29.918,
      "step": 5980
    },
    {
      "epoch": 46.07692307692308,
      "grad_norm": 0.0025213544722646475,
      "learning_rate": 1.5723076923076925e-06,
      "loss": 0.0,
      "step": 5990
    },
    {
      "epoch": 46.15384615384615,
      "grad_norm": 0.0015282627427950501,
      "learning_rate": 1.5415384615384614e-06,
      "loss": 0.0,
      "step": 6000
    },
    {
      "epoch": 46.23076923076923,
      "grad_norm": 0.017315922304987907,
      "learning_rate": 1.5107692307692308e-06,
      "loss": 0.0003,
      "step": 6010
    },
    {
      "epoch": 46.30769230769231,
      "grad_norm": 0.0011600840371102095,
      "learning_rate": 1.48e-06,
      "loss": 0.0002,
      "step": 6020
    },
    {
      "epoch": 46.38461538461539,
      "grad_norm": 0.10278120636940002,
      "learning_rate": 1.4492307692307694e-06,
      "loss": 0.0801,
      "step": 6030
    },
    {
      "epoch": 46.46153846153846,
      "grad_norm": 0.006299668923020363,
      "learning_rate": 1.4184615384615386e-06,
      "loss": 0.109,
      "step": 6040
    },
    {
      "epoch": 46.53846153846154,
      "grad_norm": 0.00047507460112683475,
      "learning_rate": 1.3876923076923076e-06,
      "loss": 0.018,
      "step": 6050
    },
    {
      "epoch": 46.61538461538461,
      "grad_norm": 0.0016794758848845959,
      "learning_rate": 1.356923076923077e-06,
      "loss": 0.0001,
      "step": 6060
    },
    {
      "epoch": 46.69230769230769,
      "grad_norm": 0.38092464208602905,
      "learning_rate": 1.3261538461538461e-06,
      "loss": 0.0451,
      "step": 6070
    },
    {
      "epoch": 46.76923076923077,
      "grad_norm": 0.0006377891404554248,
      "learning_rate": 1.2953846153846155e-06,
      "loss": 0.0022,
      "step": 6080
    },
    {
      "epoch": 46.84615384615385,
      "grad_norm": 3.781010389328003,
      "learning_rate": 1.2646153846153847e-06,
      "loss": 0.0008,
      "step": 6090
    },
    {
      "epoch": 46.92307692307692,
      "grad_norm": 16.57982063293457,
      "learning_rate": 1.233846153846154e-06,
      "loss": 0.0793,
      "step": 6100
    },
    {
      "epoch": 47.0,
      "grad_norm": 0.0005190350348129869,
      "learning_rate": 1.2030769230769233e-06,
      "loss": 0.0019,
      "step": 6110
    },
    {
      "epoch": 47.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.07604598999023438,
      "eval_runtime": 0.5685,
      "eval_samples_per_second": 233.963,
      "eval_steps_per_second": 29.905,
      "step": 6110
    },
    {
      "epoch": 47.07692307692308,
      "grad_norm": 0.0002213361149188131,
      "learning_rate": 1.1723076923076925e-06,
      "loss": 0.0,
      "step": 6120
    },
    {
      "epoch": 47.15384615384615,
      "grad_norm": 0.003321431577205658,
      "learning_rate": 1.1415384615384617e-06,
      "loss": 0.0015,
      "step": 6130
    },
    {
      "epoch": 47.23076923076923,
      "grad_norm": 0.005034551955759525,
      "learning_rate": 1.1107692307692309e-06,
      "loss": 0.0178,
      "step": 6140
    },
    {
      "epoch": 47.30769230769231,
      "grad_norm": 0.002936753910034895,
      "learning_rate": 1.08e-06,
      "loss": 0.0664,
      "step": 6150
    },
    {
      "epoch": 47.38461538461539,
      "grad_norm": 0.0007592691108584404,
      "learning_rate": 1.0492307692307694e-06,
      "loss": 0.0001,
      "step": 6160
    },
    {
      "epoch": 47.46153846153846,
      "grad_norm": 0.0010430924594402313,
      "learning_rate": 1.0184615384615386e-06,
      "loss": 0.0709,
      "step": 6170
    },
    {
      "epoch": 47.53846153846154,
      "grad_norm": 0.00222417083568871,
      "learning_rate": 9.876923076923078e-07,
      "loss": 0.0001,
      "step": 6180
    },
    {
      "epoch": 47.61538461538461,
      "grad_norm": 0.0008099984261207283,
      "learning_rate": 9.56923076923077e-07,
      "loss": 0.0,
      "step": 6190
    },
    {
      "epoch": 47.69230769230769,
      "grad_norm": 0.005475889425724745,
      "learning_rate": 9.261538461538462e-07,
      "loss": 0.0001,
      "step": 6200
    },
    {
      "epoch": 47.76923076923077,
      "grad_norm": 0.0015080670127645135,
      "learning_rate": 8.953846153846155e-07,
      "loss": 0.0,
      "step": 6210
    },
    {
      "epoch": 47.84615384615385,
      "grad_norm": 0.0009012154769152403,
      "learning_rate": 8.646153846153847e-07,
      "loss": 0.0013,
      "step": 6220
    },
    {
      "epoch": 47.92307692307692,
      "grad_norm": 0.0014631409430876374,
      "learning_rate": 8.338461538461539e-07,
      "loss": 0.1119,
      "step": 6230
    },
    {
      "epoch": 48.0,
      "grad_norm": 0.0026774315629154444,
      "learning_rate": 8.030769230769231e-07,
      "loss": 0.0001,
      "step": 6240
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.08850181102752686,
      "eval_runtime": 0.559,
      "eval_samples_per_second": 237.905,
      "eval_steps_per_second": 30.409,
      "step": 6240
    },
    {
      "epoch": 48.07692307692308,
      "grad_norm": 0.005435484927147627,
      "learning_rate": 7.723076923076923e-07,
      "loss": 0.0017,
      "step": 6250
    },
    {
      "epoch": 48.15384615384615,
      "grad_norm": 0.0006916754064150155,
      "learning_rate": 7.415384615384616e-07,
      "loss": 0.0001,
      "step": 6260
    },
    {
      "epoch": 48.23076923076923,
      "grad_norm": 0.008791380561888218,
      "learning_rate": 7.107692307692309e-07,
      "loss": 0.0,
      "step": 6270
    },
    {
      "epoch": 48.30769230769231,
      "grad_norm": 0.0013945720857009292,
      "learning_rate": 6.800000000000001e-07,
      "loss": 0.0,
      "step": 6280
    },
    {
      "epoch": 48.38461538461539,
      "grad_norm": 0.0008068948518484831,
      "learning_rate": 6.492307692307692e-07,
      "loss": 0.0189,
      "step": 6290
    },
    {
      "epoch": 48.46153846153846,
      "grad_norm": 0.0035911358427256346,
      "learning_rate": 6.184615384615385e-07,
      "loss": 0.0001,
      "step": 6300
    },
    {
      "epoch": 48.53846153846154,
      "grad_norm": 0.02670697495341301,
      "learning_rate": 5.876923076923077e-07,
      "loss": 0.0615,
      "step": 6310
    },
    {
      "epoch": 48.61538461538461,
      "grad_norm": 0.0006088964291848242,
      "learning_rate": 5.56923076923077e-07,
      "loss": 0.0069,
      "step": 6320
    },
    {
      "epoch": 48.69230769230769,
      "grad_norm": 26.212440490722656,
      "learning_rate": 5.261538461538462e-07,
      "loss": 0.0589,
      "step": 6330
    },
    {
      "epoch": 48.76923076923077,
      "grad_norm": 0.0010863661300390959,
      "learning_rate": 4.953846153846155e-07,
      "loss": 0.0001,
      "step": 6340
    },
    {
      "epoch": 48.84615384615385,
      "grad_norm": 0.0006395662785507739,
      "learning_rate": 4.6461538461538465e-07,
      "loss": 0.0025,
      "step": 6350
    },
    {
      "epoch": 48.92307692307692,
      "grad_norm": 0.000708853651303798,
      "learning_rate": 4.3384615384615384e-07,
      "loss": 0.0001,
      "step": 6360
    },
    {
      "epoch": 49.0,
      "grad_norm": 0.00027008148026652634,
      "learning_rate": 4.0307692307692313e-07,
      "loss": 0.0854,
      "step": 6370
    },
    {
      "epoch": 49.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.0787627175450325,
      "eval_runtime": 0.5704,
      "eval_samples_per_second": 233.186,
      "eval_steps_per_second": 29.806,
      "step": 6370
    },
    {
      "epoch": 49.07692307692308,
      "grad_norm": 0.0025157821364700794,
      "learning_rate": 3.7230769230769236e-07,
      "loss": 0.0677,
      "step": 6380
    },
    {
      "epoch": 49.15384615384615,
      "grad_norm": 0.002912031952291727,
      "learning_rate": 3.4153846153846155e-07,
      "loss": 0.0001,
      "step": 6390
    },
    {
      "epoch": 49.23076923076923,
      "grad_norm": 0.000612029223702848,
      "learning_rate": 3.107692307692308e-07,
      "loss": 0.0003,
      "step": 6400
    },
    {
      "epoch": 49.30769230769231,
      "grad_norm": 0.001225666725076735,
      "learning_rate": 2.8e-07,
      "loss": 0.0003,
      "step": 6410
    },
    {
      "epoch": 49.38461538461539,
      "grad_norm": 0.884882390499115,
      "learning_rate": 2.4923076923076926e-07,
      "loss": 0.0231,
      "step": 6420
    },
    {
      "epoch": 49.46153846153846,
      "grad_norm": 0.0010143619729205966,
      "learning_rate": 2.1846153846153847e-07,
      "loss": 0.0,
      "step": 6430
    },
    {
      "epoch": 49.53846153846154,
      "grad_norm": 0.028597572818398476,
      "learning_rate": 1.8769230769230773e-07,
      "loss": 0.0004,
      "step": 6440
    },
    {
      "epoch": 49.61538461538461,
      "grad_norm": 0.0006769668543711305,
      "learning_rate": 1.5692307692307694e-07,
      "loss": 0.0014,
      "step": 6450
    },
    {
      "epoch": 49.69230769230769,
      "grad_norm": 0.0014201749581843615,
      "learning_rate": 1.2615384615384617e-07,
      "loss": 0.0,
      "step": 6460
    },
    {
      "epoch": 49.76923076923077,
      "grad_norm": 0.010460024699568748,
      "learning_rate": 9.53846153846154e-08,
      "loss": 0.1236,
      "step": 6470
    },
    {
      "epoch": 49.84615384615385,
      "grad_norm": 0.036062873899936676,
      "learning_rate": 6.461538461538462e-08,
      "loss": 0.0367,
      "step": 6480
    },
    {
      "epoch": 49.92307692307692,
      "grad_norm": 0.0007868582615628839,
      "learning_rate": 3.384615384615385e-08,
      "loss": 0.0051,
      "step": 6490
    },
    {
      "epoch": 50.0,
      "grad_norm": 0.04576790705323219,
      "learning_rate": 3.0769230769230774e-09,
      "loss": 0.0005,
      "step": 6500
    },
    {
      "epoch": 50.0,
      "eval_accuracy": 0.9849624060150376,
      "eval_loss": 0.07072408497333527,
      "eval_runtime": 0.5576,
      "eval_samples_per_second": 238.518,
      "eval_steps_per_second": 30.487,
      "step": 6500
    },
    {
      "epoch": 50.0,
      "step": 6500,
      "total_flos": 4.0046594160900096e+18,
      "train_loss": 0.0702138227691556,
      "train_runtime": 508.0972,
      "train_samples_per_second": 101.752,
      "train_steps_per_second": 12.793
    }
  ],
  "logging_steps": 10,
  "max_steps": 6500,
  "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": 4.0046594160900096e+18,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}