{
  "best_metric": 0.03492419794201851,
  "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-2800",
  "epoch": 6.0,
  "eval_steps": 100,
  "global_step": 4020,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 9.975124378109453e-05,
      "loss": 2.5776,
      "step": 10
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.950248756218906e-05,
      "loss": 2.217,
      "step": 20
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.925373134328359e-05,
      "loss": 1.9279,
      "step": 30
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.900497512437812e-05,
      "loss": 1.6945,
      "step": 40
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.875621890547265e-05,
      "loss": 1.4931,
      "step": 50
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.850746268656717e-05,
      "loss": 1.2644,
      "step": 60
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.82587064676617e-05,
      "loss": 1.1095,
      "step": 70
    },
    {
      "epoch": 0.12,
      "learning_rate": 9.800995024875622e-05,
      "loss": 1.0658,
      "step": 80
    },
    {
      "epoch": 0.13,
      "learning_rate": 9.776119402985075e-05,
      "loss": 0.8875,
      "step": 90
    },
    {
      "epoch": 0.15,
      "learning_rate": 9.751243781094528e-05,
      "loss": 0.8306,
      "step": 100
    },
    {
      "epoch": 0.15,
      "eval_accuracy": 0.90848026868178,
      "eval_loss": 0.791517972946167,
      "eval_runtime": 20.4511,
      "eval_samples_per_second": 58.237,
      "eval_steps_per_second": 7.286,
      "step": 100
    },
    {
      "epoch": 0.16,
      "learning_rate": 9.72636815920398e-05,
      "loss": 0.7402,
      "step": 110
    },
    {
      "epoch": 0.18,
      "learning_rate": 9.701492537313434e-05,
      "loss": 0.7802,
      "step": 120
    },
    {
      "epoch": 0.19,
      "learning_rate": 9.676616915422886e-05,
      "loss": 0.6563,
      "step": 130
    },
    {
      "epoch": 0.21,
      "learning_rate": 9.65174129353234e-05,
      "loss": 0.5579,
      "step": 140
    },
    {
      "epoch": 0.22,
      "learning_rate": 9.626865671641792e-05,
      "loss": 0.6564,
      "step": 150
    },
    {
      "epoch": 0.24,
      "learning_rate": 9.601990049751244e-05,
      "loss": 0.5115,
      "step": 160
    },
    {
      "epoch": 0.25,
      "learning_rate": 9.577114427860697e-05,
      "loss": 0.5087,
      "step": 170
    },
    {
      "epoch": 0.27,
      "learning_rate": 9.552238805970149e-05,
      "loss": 0.4821,
      "step": 180
    },
    {
      "epoch": 0.28,
      "learning_rate": 9.527363184079603e-05,
      "loss": 0.3972,
      "step": 190
    },
    {
      "epoch": 0.3,
      "learning_rate": 9.502487562189055e-05,
      "loss": 0.3828,
      "step": 200
    },
    {
      "epoch": 0.3,
      "eval_accuracy": 0.9269521410579346,
      "eval_loss": 0.42952126264572144,
      "eval_runtime": 18.5768,
      "eval_samples_per_second": 64.112,
      "eval_steps_per_second": 8.021,
      "step": 200
    },
    {
      "epoch": 0.31,
      "learning_rate": 9.477611940298507e-05,
      "loss": 0.4261,
      "step": 210
    },
    {
      "epoch": 0.33,
      "learning_rate": 9.452736318407961e-05,
      "loss": 0.3445,
      "step": 220
    },
    {
      "epoch": 0.34,
      "learning_rate": 9.427860696517413e-05,
      "loss": 0.3152,
      "step": 230
    },
    {
      "epoch": 0.36,
      "learning_rate": 9.402985074626867e-05,
      "loss": 0.3448,
      "step": 240
    },
    {
      "epoch": 0.37,
      "learning_rate": 9.37810945273632e-05,
      "loss": 0.4312,
      "step": 250
    },
    {
      "epoch": 0.39,
      "learning_rate": 9.353233830845772e-05,
      "loss": 0.3472,
      "step": 260
    },
    {
      "epoch": 0.4,
      "learning_rate": 9.328358208955224e-05,
      "loss": 0.4248,
      "step": 270
    },
    {
      "epoch": 0.42,
      "learning_rate": 9.303482587064676e-05,
      "loss": 0.3977,
      "step": 280
    },
    {
      "epoch": 0.43,
      "learning_rate": 9.27860696517413e-05,
      "loss": 0.3164,
      "step": 290
    },
    {
      "epoch": 0.45,
      "learning_rate": 9.253731343283582e-05,
      "loss": 0.3623,
      "step": 300
    },
    {
      "epoch": 0.45,
      "eval_accuracy": 0.9370277078085643,
      "eval_loss": 0.30219167470932007,
      "eval_runtime": 23.4722,
      "eval_samples_per_second": 50.741,
      "eval_steps_per_second": 6.348,
      "step": 300
    },
    {
      "epoch": 0.46,
      "learning_rate": 9.228855721393035e-05,
      "loss": 0.2591,
      "step": 310
    },
    {
      "epoch": 0.48,
      "learning_rate": 9.203980099502488e-05,
      "loss": 0.277,
      "step": 320
    },
    {
      "epoch": 0.49,
      "learning_rate": 9.17910447761194e-05,
      "loss": 0.2457,
      "step": 330
    },
    {
      "epoch": 0.51,
      "learning_rate": 9.154228855721394e-05,
      "loss": 0.1788,
      "step": 340
    },
    {
      "epoch": 0.52,
      "learning_rate": 9.129353233830847e-05,
      "loss": 0.2828,
      "step": 350
    },
    {
      "epoch": 0.54,
      "learning_rate": 9.104477611940299e-05,
      "loss": 0.2579,
      "step": 360
    },
    {
      "epoch": 0.55,
      "learning_rate": 9.079601990049753e-05,
      "loss": 0.2195,
      "step": 370
    },
    {
      "epoch": 0.57,
      "learning_rate": 9.054726368159205e-05,
      "loss": 0.2773,
      "step": 380
    },
    {
      "epoch": 0.58,
      "learning_rate": 9.029850746268657e-05,
      "loss": 0.188,
      "step": 390
    },
    {
      "epoch": 0.6,
      "learning_rate": 9.00497512437811e-05,
      "loss": 0.2623,
      "step": 400
    },
    {
      "epoch": 0.6,
      "eval_accuracy": 0.9269521410579346,
      "eval_loss": 0.2882852256298065,
      "eval_runtime": 19.7432,
      "eval_samples_per_second": 60.325,
      "eval_steps_per_second": 7.547,
      "step": 400
    },
    {
      "epoch": 0.61,
      "learning_rate": 8.980099502487562e-05,
      "loss": 0.2402,
      "step": 410
    },
    {
      "epoch": 0.63,
      "learning_rate": 8.955223880597016e-05,
      "loss": 0.1562,
      "step": 420
    },
    {
      "epoch": 0.64,
      "learning_rate": 8.930348258706468e-05,
      "loss": 0.3121,
      "step": 430
    },
    {
      "epoch": 0.66,
      "learning_rate": 8.905472636815922e-05,
      "loss": 0.2505,
      "step": 440
    },
    {
      "epoch": 0.67,
      "learning_rate": 8.880597014925374e-05,
      "loss": 0.2329,
      "step": 450
    },
    {
      "epoch": 0.69,
      "learning_rate": 8.855721393034826e-05,
      "loss": 0.1912,
      "step": 460
    },
    {
      "epoch": 0.7,
      "learning_rate": 8.83084577114428e-05,
      "loss": 0.2406,
      "step": 470
    },
    {
      "epoch": 0.72,
      "learning_rate": 8.805970149253732e-05,
      "loss": 0.1853,
      "step": 480
    },
    {
      "epoch": 0.73,
      "learning_rate": 8.781094527363185e-05,
      "loss": 0.2545,
      "step": 490
    },
    {
      "epoch": 0.75,
      "learning_rate": 8.756218905472637e-05,
      "loss": 0.2251,
      "step": 500
    },
    {
      "epoch": 0.75,
      "eval_accuracy": 0.9462636439966414,
      "eval_loss": 0.23029384016990662,
      "eval_runtime": 20.0888,
      "eval_samples_per_second": 59.287,
      "eval_steps_per_second": 7.417,
      "step": 500
    },
    {
      "epoch": 0.76,
      "learning_rate": 8.731343283582089e-05,
      "loss": 0.1969,
      "step": 510
    },
    {
      "epoch": 0.78,
      "learning_rate": 8.706467661691543e-05,
      "loss": 0.1476,
      "step": 520
    },
    {
      "epoch": 0.79,
      "learning_rate": 8.681592039800995e-05,
      "loss": 0.2066,
      "step": 530
    },
    {
      "epoch": 0.81,
      "learning_rate": 8.656716417910447e-05,
      "loss": 0.3319,
      "step": 540
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.631840796019901e-05,
      "loss": 0.1591,
      "step": 550
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.606965174129353e-05,
      "loss": 0.1666,
      "step": 560
    },
    {
      "epoch": 0.85,
      "learning_rate": 8.582089552238807e-05,
      "loss": 0.2891,
      "step": 570
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.55721393034826e-05,
      "loss": 0.178,
      "step": 580
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.532338308457712e-05,
      "loss": 0.1507,
      "step": 590
    },
    {
      "epoch": 0.9,
      "learning_rate": 8.507462686567164e-05,
      "loss": 0.2695,
      "step": 600
    },
    {
      "epoch": 0.9,
      "eval_accuracy": 0.9378673383711167,
      "eval_loss": 0.2464931756258011,
      "eval_runtime": 19.6563,
      "eval_samples_per_second": 60.591,
      "eval_steps_per_second": 7.58,
      "step": 600
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.482587064676616e-05,
      "loss": 0.2417,
      "step": 610
    },
    {
      "epoch": 0.93,
      "learning_rate": 8.45771144278607e-05,
      "loss": 0.2455,
      "step": 620
    },
    {
      "epoch": 0.94,
      "learning_rate": 8.432835820895522e-05,
      "loss": 0.1249,
      "step": 630
    },
    {
      "epoch": 0.96,
      "learning_rate": 8.407960199004975e-05,
      "loss": 0.1533,
      "step": 640
    },
    {
      "epoch": 0.97,
      "learning_rate": 8.383084577114428e-05,
      "loss": 0.1838,
      "step": 650
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.360696517412937e-05,
      "loss": 0.1247,
      "step": 660
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.335820895522388e-05,
      "loss": 0.1583,
      "step": 670
    },
    {
      "epoch": 1.01,
      "learning_rate": 8.310945273631841e-05,
      "loss": 0.1024,
      "step": 680
    },
    {
      "epoch": 1.03,
      "learning_rate": 8.286069651741294e-05,
      "loss": 0.0599,
      "step": 690
    },
    {
      "epoch": 1.04,
      "learning_rate": 8.261194029850746e-05,
      "loss": 0.1256,
      "step": 700
    },
    {
      "epoch": 1.04,
      "eval_accuracy": 0.9714525608732157,
      "eval_loss": 0.12489016354084015,
      "eval_runtime": 18.8609,
      "eval_samples_per_second": 63.146,
      "eval_steps_per_second": 7.9,
      "step": 700
    },
    {
      "epoch": 1.06,
      "learning_rate": 8.2363184079602e-05,
      "loss": 0.0541,
      "step": 710
    },
    {
      "epoch": 1.07,
      "learning_rate": 8.211442786069652e-05,
      "loss": 0.0742,
      "step": 720
    },
    {
      "epoch": 1.09,
      "learning_rate": 8.186567164179106e-05,
      "loss": 0.0975,
      "step": 730
    },
    {
      "epoch": 1.1,
      "learning_rate": 8.161691542288558e-05,
      "loss": 0.1011,
      "step": 740
    },
    {
      "epoch": 1.12,
      "learning_rate": 8.13681592039801e-05,
      "loss": 0.1081,
      "step": 750
    },
    {
      "epoch": 1.13,
      "learning_rate": 8.111940298507464e-05,
      "loss": 0.129,
      "step": 760
    },
    {
      "epoch": 1.15,
      "learning_rate": 8.087064676616916e-05,
      "loss": 0.1595,
      "step": 770
    },
    {
      "epoch": 1.16,
      "learning_rate": 8.062189054726368e-05,
      "loss": 0.163,
      "step": 780
    },
    {
      "epoch": 1.18,
      "learning_rate": 8.037313432835821e-05,
      "loss": 0.1464,
      "step": 790
    },
    {
      "epoch": 1.19,
      "learning_rate": 8.012437810945273e-05,
      "loss": 0.0718,
      "step": 800
    },
    {
      "epoch": 1.19,
      "eval_accuracy": 0.9622166246851386,
      "eval_loss": 0.13346683979034424,
      "eval_runtime": 19.3064,
      "eval_samples_per_second": 61.689,
      "eval_steps_per_second": 7.718,
      "step": 800
    },
    {
      "epoch": 1.21,
      "learning_rate": 7.987562189054727e-05,
      "loss": 0.0517,
      "step": 810
    },
    {
      "epoch": 1.22,
      "learning_rate": 7.962686567164179e-05,
      "loss": 0.1143,
      "step": 820
    },
    {
      "epoch": 1.24,
      "learning_rate": 7.937810945273633e-05,
      "loss": 0.062,
      "step": 830
    },
    {
      "epoch": 1.25,
      "learning_rate": 7.912935323383085e-05,
      "loss": 0.2013,
      "step": 840
    },
    {
      "epoch": 1.27,
      "learning_rate": 7.888059701492537e-05,
      "loss": 0.0995,
      "step": 850
    },
    {
      "epoch": 1.28,
      "learning_rate": 7.863184079601991e-05,
      "loss": 0.1009,
      "step": 860
    },
    {
      "epoch": 1.3,
      "learning_rate": 7.838308457711443e-05,
      "loss": 0.1292,
      "step": 870
    },
    {
      "epoch": 1.31,
      "learning_rate": 7.813432835820896e-05,
      "loss": 0.0848,
      "step": 880
    },
    {
      "epoch": 1.33,
      "learning_rate": 7.788557213930348e-05,
      "loss": 0.0323,
      "step": 890
    },
    {
      "epoch": 1.34,
      "learning_rate": 7.7636815920398e-05,
      "loss": 0.0436,
      "step": 900
    },
    {
      "epoch": 1.34,
      "eval_accuracy": 0.979009235936188,
      "eval_loss": 0.09892778843641281,
      "eval_runtime": 20.325,
      "eval_samples_per_second": 58.598,
      "eval_steps_per_second": 7.331,
      "step": 900
    },
    {
      "epoch": 1.36,
      "learning_rate": 7.738805970149254e-05,
      "loss": 0.0345,
      "step": 910
    },
    {
      "epoch": 1.37,
      "learning_rate": 7.713930348258706e-05,
      "loss": 0.0505,
      "step": 920
    },
    {
      "epoch": 1.39,
      "learning_rate": 7.68905472636816e-05,
      "loss": 0.0464,
      "step": 930
    },
    {
      "epoch": 1.4,
      "learning_rate": 7.664179104477612e-05,
      "loss": 0.0654,
      "step": 940
    },
    {
      "epoch": 1.42,
      "learning_rate": 7.639303482587065e-05,
      "loss": 0.0776,
      "step": 950
    },
    {
      "epoch": 1.43,
      "learning_rate": 7.614427860696518e-05,
      "loss": 0.0396,
      "step": 960
    },
    {
      "epoch": 1.45,
      "learning_rate": 7.589552238805971e-05,
      "loss": 0.0373,
      "step": 970
    },
    {
      "epoch": 1.46,
      "learning_rate": 7.564676616915424e-05,
      "loss": 0.083,
      "step": 980
    },
    {
      "epoch": 1.48,
      "learning_rate": 7.539800995024875e-05,
      "loss": 0.1018,
      "step": 990
    },
    {
      "epoch": 1.49,
      "learning_rate": 7.514925373134328e-05,
      "loss": 0.1681,
      "step": 1000
    },
    {
      "epoch": 1.49,
      "eval_accuracy": 0.9496221662468514,
      "eval_loss": 0.18528999388217926,
      "eval_runtime": 19.3383,
      "eval_samples_per_second": 61.588,
      "eval_steps_per_second": 7.705,
      "step": 1000
    },
    {
      "epoch": 1.51,
      "learning_rate": 7.490049751243781e-05,
      "loss": 0.0607,
      "step": 1010
    },
    {
      "epoch": 1.52,
      "learning_rate": 7.465174129353234e-05,
      "loss": 0.1343,
      "step": 1020
    },
    {
      "epoch": 1.54,
      "learning_rate": 7.440298507462687e-05,
      "loss": 0.1163,
      "step": 1030
    },
    {
      "epoch": 1.55,
      "learning_rate": 7.41542288557214e-05,
      "loss": 0.0628,
      "step": 1040
    },
    {
      "epoch": 1.57,
      "learning_rate": 7.390547263681592e-05,
      "loss": 0.1001,
      "step": 1050
    },
    {
      "epoch": 1.58,
      "learning_rate": 7.365671641791046e-05,
      "loss": 0.0764,
      "step": 1060
    },
    {
      "epoch": 1.6,
      "learning_rate": 7.340796019900498e-05,
      "loss": 0.0281,
      "step": 1070
    },
    {
      "epoch": 1.61,
      "learning_rate": 7.315920398009952e-05,
      "loss": 0.0861,
      "step": 1080
    },
    {
      "epoch": 1.63,
      "learning_rate": 7.291044776119404e-05,
      "loss": 0.0305,
      "step": 1090
    },
    {
      "epoch": 1.64,
      "learning_rate": 7.266169154228856e-05,
      "loss": 0.0771,
      "step": 1100
    },
    {
      "epoch": 1.64,
      "eval_accuracy": 0.964735516372796,
      "eval_loss": 0.14002813398838043,
      "eval_runtime": 19.5,
      "eval_samples_per_second": 61.077,
      "eval_steps_per_second": 7.641,
      "step": 1100
    },
    {
      "epoch": 1.66,
      "learning_rate": 7.241293532338309e-05,
      "loss": 0.1127,
      "step": 1110
    },
    {
      "epoch": 1.67,
      "learning_rate": 7.216417910447761e-05,
      "loss": 0.0412,
      "step": 1120
    },
    {
      "epoch": 1.69,
      "learning_rate": 7.191542288557215e-05,
      "loss": 0.1501,
      "step": 1130
    },
    {
      "epoch": 1.7,
      "learning_rate": 7.166666666666667e-05,
      "loss": 0.1287,
      "step": 1140
    },
    {
      "epoch": 1.72,
      "learning_rate": 7.141791044776119e-05,
      "loss": 0.0528,
      "step": 1150
    },
    {
      "epoch": 1.73,
      "learning_rate": 7.116915422885573e-05,
      "loss": 0.0139,
      "step": 1160
    },
    {
      "epoch": 1.75,
      "learning_rate": 7.092039800995025e-05,
      "loss": 0.0465,
      "step": 1170
    },
    {
      "epoch": 1.76,
      "learning_rate": 7.067164179104479e-05,
      "loss": 0.0412,
      "step": 1180
    },
    {
      "epoch": 1.78,
      "learning_rate": 7.042288557213931e-05,
      "loss": 0.0812,
      "step": 1190
    },
    {
      "epoch": 1.79,
      "learning_rate": 7.017412935323384e-05,
      "loss": 0.0815,
      "step": 1200
    },
    {
      "epoch": 1.79,
      "eval_accuracy": 0.9764903442485307,
      "eval_loss": 0.0969606265425682,
      "eval_runtime": 18.9862,
      "eval_samples_per_second": 62.73,
      "eval_steps_per_second": 7.848,
      "step": 1200
    },
    {
      "epoch": 1.81,
      "learning_rate": 6.992537313432836e-05,
      "loss": 0.0212,
      "step": 1210
    },
    {
      "epoch": 1.82,
      "learning_rate": 6.967661691542288e-05,
      "loss": 0.0576,
      "step": 1220
    },
    {
      "epoch": 1.84,
      "learning_rate": 6.942786069651742e-05,
      "loss": 0.0269,
      "step": 1230
    },
    {
      "epoch": 1.85,
      "learning_rate": 6.917910447761194e-05,
      "loss": 0.0557,
      "step": 1240
    },
    {
      "epoch": 1.87,
      "learning_rate": 6.893034825870647e-05,
      "loss": 0.1284,
      "step": 1250
    },
    {
      "epoch": 1.88,
      "learning_rate": 6.8681592039801e-05,
      "loss": 0.0499,
      "step": 1260
    },
    {
      "epoch": 1.9,
      "learning_rate": 6.843283582089553e-05,
      "loss": 0.114,
      "step": 1270
    },
    {
      "epoch": 1.91,
      "learning_rate": 6.818407960199006e-05,
      "loss": 0.0338,
      "step": 1280
    },
    {
      "epoch": 1.93,
      "learning_rate": 6.793532338308459e-05,
      "loss": 0.1359,
      "step": 1290
    },
    {
      "epoch": 1.94,
      "learning_rate": 6.768656716417911e-05,
      "loss": 0.0203,
      "step": 1300
    },
    {
      "epoch": 1.94,
      "eval_accuracy": 0.964735516372796,
      "eval_loss": 0.14076600968837738,
      "eval_runtime": 19.9477,
      "eval_samples_per_second": 59.706,
      "eval_steps_per_second": 7.47,
      "step": 1300
    },
    {
      "epoch": 1.96,
      "learning_rate": 6.743781094527363e-05,
      "loss": 0.0373,
      "step": 1310
    },
    {
      "epoch": 1.97,
      "learning_rate": 6.718905472636815e-05,
      "loss": 0.0751,
      "step": 1320
    },
    {
      "epoch": 1.99,
      "learning_rate": 6.694029850746268e-05,
      "loss": 0.0923,
      "step": 1330
    },
    {
      "epoch": 2.0,
      "learning_rate": 6.669154228855721e-05,
      "loss": 0.0406,
      "step": 1340
    },
    {
      "epoch": 2.01,
      "learning_rate": 6.644278606965174e-05,
      "loss": 0.0333,
      "step": 1350
    },
    {
      "epoch": 2.03,
      "learning_rate": 6.619402985074627e-05,
      "loss": 0.0674,
      "step": 1360
    },
    {
      "epoch": 2.04,
      "learning_rate": 6.59452736318408e-05,
      "loss": 0.0971,
      "step": 1370
    },
    {
      "epoch": 2.06,
      "learning_rate": 6.569651741293532e-05,
      "loss": 0.1701,
      "step": 1380
    },
    {
      "epoch": 2.07,
      "learning_rate": 6.544776119402986e-05,
      "loss": 0.1362,
      "step": 1390
    },
    {
      "epoch": 2.09,
      "learning_rate": 6.519900497512438e-05,
      "loss": 0.0774,
      "step": 1400
    },
    {
      "epoch": 2.09,
      "eval_accuracy": 0.966414777497901,
      "eval_loss": 0.12388637661933899,
      "eval_runtime": 19.0185,
      "eval_samples_per_second": 62.623,
      "eval_steps_per_second": 7.834,
      "step": 1400
    },
    {
      "epoch": 2.1,
      "learning_rate": 6.495024875621892e-05,
      "loss": 0.0306,
      "step": 1410
    },
    {
      "epoch": 2.12,
      "learning_rate": 6.470149253731344e-05,
      "loss": 0.0421,
      "step": 1420
    },
    {
      "epoch": 2.13,
      "learning_rate": 6.445273631840796e-05,
      "loss": 0.0408,
      "step": 1430
    },
    {
      "epoch": 2.15,
      "learning_rate": 6.420398009950249e-05,
      "loss": 0.0333,
      "step": 1440
    },
    {
      "epoch": 2.16,
      "learning_rate": 6.395522388059701e-05,
      "loss": 0.0124,
      "step": 1450
    },
    {
      "epoch": 2.18,
      "learning_rate": 6.370646766169155e-05,
      "loss": 0.0546,
      "step": 1460
    },
    {
      "epoch": 2.19,
      "learning_rate": 6.345771144278607e-05,
      "loss": 0.0088,
      "step": 1470
    },
    {
      "epoch": 2.21,
      "learning_rate": 6.32089552238806e-05,
      "loss": 0.0525,
      "step": 1480
    },
    {
      "epoch": 2.22,
      "learning_rate": 6.296019900497513e-05,
      "loss": 0.0336,
      "step": 1490
    },
    {
      "epoch": 2.24,
      "learning_rate": 6.271144278606965e-05,
      "loss": 0.0723,
      "step": 1500
    },
    {
      "epoch": 2.24,
      "eval_accuracy": 0.982367758186398,
      "eval_loss": 0.06230420619249344,
      "eval_runtime": 20.5805,
      "eval_samples_per_second": 57.87,
      "eval_steps_per_second": 7.24,
      "step": 1500
    },
    {
      "epoch": 2.25,
      "learning_rate": 6.246268656716419e-05,
      "loss": 0.04,
      "step": 1510
    },
    {
      "epoch": 2.27,
      "learning_rate": 6.221393034825871e-05,
      "loss": 0.029,
      "step": 1520
    },
    {
      "epoch": 2.28,
      "learning_rate": 6.196517412935324e-05,
      "loss": 0.0161,
      "step": 1530
    },
    {
      "epoch": 2.3,
      "learning_rate": 6.171641791044776e-05,
      "loss": 0.0307,
      "step": 1540
    },
    {
      "epoch": 2.31,
      "learning_rate": 6.146766169154228e-05,
      "loss": 0.0099,
      "step": 1550
    },
    {
      "epoch": 2.33,
      "learning_rate": 6.121890547263682e-05,
      "loss": 0.0814,
      "step": 1560
    },
    {
      "epoch": 2.34,
      "learning_rate": 6.0970149253731343e-05,
      "loss": 0.052,
      "step": 1570
    },
    {
      "epoch": 2.36,
      "learning_rate": 6.0721393034825867e-05,
      "loss": 0.0437,
      "step": 1580
    },
    {
      "epoch": 2.37,
      "learning_rate": 6.04726368159204e-05,
      "loss": 0.0411,
      "step": 1590
    },
    {
      "epoch": 2.39,
      "learning_rate": 6.0223880597014927e-05,
      "loss": 0.0187,
      "step": 1600
    },
    {
      "epoch": 2.39,
      "eval_accuracy": 0.9731318219983207,
      "eval_loss": 0.0926276370882988,
      "eval_runtime": 20.5962,
      "eval_samples_per_second": 57.826,
      "eval_steps_per_second": 7.234,
      "step": 1600
    },
    {
      "epoch": 2.4,
      "learning_rate": 5.9975124378109457e-05,
      "loss": 0.0374,
      "step": 1610
    },
    {
      "epoch": 2.42,
      "learning_rate": 5.972636815920398e-05,
      "loss": 0.0216,
      "step": 1620
    },
    {
      "epoch": 2.43,
      "learning_rate": 5.94776119402985e-05,
      "loss": 0.0667,
      "step": 1630
    },
    {
      "epoch": 2.45,
      "learning_rate": 5.922885572139304e-05,
      "loss": 0.0691,
      "step": 1640
    },
    {
      "epoch": 2.46,
      "learning_rate": 5.898009950248756e-05,
      "loss": 0.0364,
      "step": 1650
    },
    {
      "epoch": 2.48,
      "learning_rate": 5.87313432835821e-05,
      "loss": 0.0212,
      "step": 1660
    },
    {
      "epoch": 2.49,
      "learning_rate": 5.8482587064676616e-05,
      "loss": 0.0209,
      "step": 1670
    },
    {
      "epoch": 2.51,
      "learning_rate": 5.823383084577114e-05,
      "loss": 0.012,
      "step": 1680
    },
    {
      "epoch": 2.52,
      "learning_rate": 5.7985074626865676e-05,
      "loss": 0.0118,
      "step": 1690
    },
    {
      "epoch": 2.54,
      "learning_rate": 5.77363184079602e-05,
      "loss": 0.055,
      "step": 1700
    },
    {
      "epoch": 2.54,
      "eval_accuracy": 0.9756507136859782,
      "eval_loss": 0.09100380539894104,
      "eval_runtime": 19.4994,
      "eval_samples_per_second": 61.079,
      "eval_steps_per_second": 7.641,
      "step": 1700
    },
    {
      "epoch": 2.55,
      "learning_rate": 5.7487562189054736e-05,
      "loss": 0.0161,
      "step": 1710
    },
    {
      "epoch": 2.57,
      "learning_rate": 5.723880597014926e-05,
      "loss": 0.0071,
      "step": 1720
    },
    {
      "epoch": 2.58,
      "learning_rate": 5.699004975124378e-05,
      "loss": 0.0103,
      "step": 1730
    },
    {
      "epoch": 2.6,
      "learning_rate": 5.674129353233831e-05,
      "loss": 0.0706,
      "step": 1740
    },
    {
      "epoch": 2.61,
      "learning_rate": 5.6492537313432836e-05,
      "loss": 0.0523,
      "step": 1750
    },
    {
      "epoch": 2.63,
      "learning_rate": 5.624378109452737e-05,
      "loss": 0.0853,
      "step": 1760
    },
    {
      "epoch": 2.64,
      "learning_rate": 5.5995024875621896e-05,
      "loss": 0.027,
      "step": 1770
    },
    {
      "epoch": 2.66,
      "learning_rate": 5.574626865671642e-05,
      "loss": 0.0283,
      "step": 1780
    },
    {
      "epoch": 2.67,
      "learning_rate": 5.549751243781095e-05,
      "loss": 0.0135,
      "step": 1790
    },
    {
      "epoch": 2.69,
      "learning_rate": 5.524875621890547e-05,
      "loss": 0.0304,
      "step": 1800
    },
    {
      "epoch": 2.69,
      "eval_accuracy": 0.9722921914357683,
      "eval_loss": 0.12775705754756927,
      "eval_runtime": 18.9368,
      "eval_samples_per_second": 62.893,
      "eval_steps_per_second": 7.868,
      "step": 1800
    },
    {
      "epoch": 2.7,
      "learning_rate": 5.500000000000001e-05,
      "loss": 0.0089,
      "step": 1810
    },
    {
      "epoch": 2.72,
      "learning_rate": 5.475124378109453e-05,
      "loss": 0.0697,
      "step": 1820
    },
    {
      "epoch": 2.73,
      "learning_rate": 5.4502487562189055e-05,
      "loss": 0.0094,
      "step": 1830
    },
    {
      "epoch": 2.75,
      "learning_rate": 5.4253731343283585e-05,
      "loss": 0.0103,
      "step": 1840
    },
    {
      "epoch": 2.76,
      "learning_rate": 5.400497512437811e-05,
      "loss": 0.0399,
      "step": 1850
    },
    {
      "epoch": 2.78,
      "learning_rate": 5.3756218905472645e-05,
      "loss": 0.0181,
      "step": 1860
    },
    {
      "epoch": 2.79,
      "learning_rate": 5.350746268656717e-05,
      "loss": 0.0129,
      "step": 1870
    },
    {
      "epoch": 2.81,
      "learning_rate": 5.325870646766169e-05,
      "loss": 0.0096,
      "step": 1880
    },
    {
      "epoch": 2.82,
      "learning_rate": 5.300995024875622e-05,
      "loss": 0.0215,
      "step": 1890
    },
    {
      "epoch": 2.84,
      "learning_rate": 5.2761194029850745e-05,
      "loss": 0.0068,
      "step": 1900
    },
    {
      "epoch": 2.84,
      "eval_accuracy": 0.9680940386230059,
      "eval_loss": 0.12825354933738708,
      "eval_runtime": 20.818,
      "eval_samples_per_second": 57.21,
      "eval_steps_per_second": 7.157,
      "step": 1900
    },
    {
      "epoch": 2.85,
      "learning_rate": 5.251243781094528e-05,
      "loss": 0.0513,
      "step": 1910
    },
    {
      "epoch": 2.87,
      "learning_rate": 5.2263681592039805e-05,
      "loss": 0.0949,
      "step": 1920
    },
    {
      "epoch": 2.88,
      "learning_rate": 5.201492537313433e-05,
      "loss": 0.0435,
      "step": 1930
    },
    {
      "epoch": 2.9,
      "learning_rate": 5.176616915422886e-05,
      "loss": 0.0382,
      "step": 1940
    },
    {
      "epoch": 2.91,
      "learning_rate": 5.151741293532338e-05,
      "loss": 0.0579,
      "step": 1950
    },
    {
      "epoch": 2.93,
      "learning_rate": 5.126865671641792e-05,
      "loss": 0.0151,
      "step": 1960
    },
    {
      "epoch": 2.94,
      "learning_rate": 5.101990049751244e-05,
      "loss": 0.051,
      "step": 1970
    },
    {
      "epoch": 2.96,
      "learning_rate": 5.0771144278606964e-05,
      "loss": 0.0474,
      "step": 1980
    },
    {
      "epoch": 2.97,
      "learning_rate": 5.05223880597015e-05,
      "loss": 0.0227,
      "step": 1990
    },
    {
      "epoch": 2.99,
      "learning_rate": 5.027363184079602e-05,
      "loss": 0.0319,
      "step": 2000
    },
    {
      "epoch": 2.99,
      "eval_accuracy": 0.9865659109991604,
      "eval_loss": 0.0589967742562294,
      "eval_runtime": 19.1134,
      "eval_samples_per_second": 62.312,
      "eval_steps_per_second": 7.796,
      "step": 2000
    },
    {
      "epoch": 3.0,
      "learning_rate": 5.0024875621890554e-05,
      "loss": 0.013,
      "step": 2010
    },
    {
      "epoch": 3.01,
      "learning_rate": 4.977611940298508e-05,
      "loss": 0.0053,
      "step": 2020
    },
    {
      "epoch": 3.03,
      "learning_rate": 4.952736318407961e-05,
      "loss": 0.036,
      "step": 2030
    },
    {
      "epoch": 3.04,
      "learning_rate": 4.927860696517414e-05,
      "loss": 0.005,
      "step": 2040
    },
    {
      "epoch": 3.06,
      "learning_rate": 4.902985074626866e-05,
      "loss": 0.0061,
      "step": 2050
    },
    {
      "epoch": 3.07,
      "learning_rate": 4.8781094527363184e-05,
      "loss": 0.052,
      "step": 2060
    },
    {
      "epoch": 3.09,
      "learning_rate": 4.8532338308457714e-05,
      "loss": 0.0093,
      "step": 2070
    },
    {
      "epoch": 3.1,
      "learning_rate": 4.8283582089552244e-05,
      "loss": 0.0055,
      "step": 2080
    },
    {
      "epoch": 3.12,
      "learning_rate": 4.803482587064677e-05,
      "loss": 0.0194,
      "step": 2090
    },
    {
      "epoch": 3.13,
      "learning_rate": 4.77860696517413e-05,
      "loss": 0.0054,
      "step": 2100
    },
    {
      "epoch": 3.13,
      "eval_accuracy": 0.982367758186398,
      "eval_loss": 0.07611096650362015,
      "eval_runtime": 19.2367,
      "eval_samples_per_second": 61.913,
      "eval_steps_per_second": 7.746,
      "step": 2100
    },
    {
      "epoch": 3.15,
      "learning_rate": 4.753731343283582e-05,
      "loss": 0.0433,
      "step": 2110
    },
    {
      "epoch": 3.16,
      "learning_rate": 4.728855721393035e-05,
      "loss": 0.0047,
      "step": 2120
    },
    {
      "epoch": 3.18,
      "learning_rate": 4.703980099502488e-05,
      "loss": 0.005,
      "step": 2130
    },
    {
      "epoch": 3.19,
      "learning_rate": 4.67910447761194e-05,
      "loss": 0.0087,
      "step": 2140
    },
    {
      "epoch": 3.21,
      "learning_rate": 4.654228855721393e-05,
      "loss": 0.0045,
      "step": 2150
    },
    {
      "epoch": 3.22,
      "learning_rate": 4.6293532338308456e-05,
      "loss": 0.0229,
      "step": 2160
    },
    {
      "epoch": 3.24,
      "learning_rate": 4.6044776119402986e-05,
      "loss": 0.0043,
      "step": 2170
    },
    {
      "epoch": 3.25,
      "learning_rate": 4.5796019900497516e-05,
      "loss": 0.006,
      "step": 2180
    },
    {
      "epoch": 3.27,
      "learning_rate": 4.554726368159204e-05,
      "loss": 0.0361,
      "step": 2190
    },
    {
      "epoch": 3.28,
      "learning_rate": 4.529850746268657e-05,
      "loss": 0.0136,
      "step": 2200
    },
    {
      "epoch": 3.28,
      "eval_accuracy": 0.982367758186398,
      "eval_loss": 0.07589124888181686,
      "eval_runtime": 18.4299,
      "eval_samples_per_second": 64.623,
      "eval_steps_per_second": 8.085,
      "step": 2200
    },
    {
      "epoch": 3.3,
      "learning_rate": 4.50497512437811e-05,
      "loss": 0.0046,
      "step": 2210
    },
    {
      "epoch": 3.31,
      "learning_rate": 4.480099502487562e-05,
      "loss": 0.0136,
      "step": 2220
    },
    {
      "epoch": 3.33,
      "learning_rate": 4.455223880597015e-05,
      "loss": 0.0042,
      "step": 2230
    },
    {
      "epoch": 3.34,
      "learning_rate": 4.4303482587064676e-05,
      "loss": 0.0344,
      "step": 2240
    },
    {
      "epoch": 3.36,
      "learning_rate": 4.4054726368159206e-05,
      "loss": 0.0112,
      "step": 2250
    },
    {
      "epoch": 3.37,
      "learning_rate": 4.3805970149253736e-05,
      "loss": 0.0401,
      "step": 2260
    },
    {
      "epoch": 3.39,
      "learning_rate": 4.355721393034826e-05,
      "loss": 0.0255,
      "step": 2270
    },
    {
      "epoch": 3.4,
      "learning_rate": 4.330845771144279e-05,
      "loss": 0.0053,
      "step": 2280
    },
    {
      "epoch": 3.42,
      "learning_rate": 4.305970149253731e-05,
      "loss": 0.0223,
      "step": 2290
    },
    {
      "epoch": 3.43,
      "learning_rate": 4.281094527363184e-05,
      "loss": 0.0136,
      "step": 2300
    },
    {
      "epoch": 3.43,
      "eval_accuracy": 0.9865659109991604,
      "eval_loss": 0.04480349272489548,
      "eval_runtime": 19.7958,
      "eval_samples_per_second": 60.164,
      "eval_steps_per_second": 7.527,
      "step": 2300
    },
    {
      "epoch": 3.45,
      "learning_rate": 4.256218905472637e-05,
      "loss": 0.0552,
      "step": 2310
    },
    {
      "epoch": 3.46,
      "learning_rate": 4.2313432835820895e-05,
      "loss": 0.0041,
      "step": 2320
    },
    {
      "epoch": 3.48,
      "learning_rate": 4.2064676616915425e-05,
      "loss": 0.004,
      "step": 2330
    },
    {
      "epoch": 3.49,
      "learning_rate": 4.181592039800995e-05,
      "loss": 0.0041,
      "step": 2340
    },
    {
      "epoch": 3.51,
      "learning_rate": 4.156716417910448e-05,
      "loss": 0.012,
      "step": 2350
    },
    {
      "epoch": 3.52,
      "learning_rate": 4.131840796019901e-05,
      "loss": 0.0039,
      "step": 2360
    },
    {
      "epoch": 3.54,
      "learning_rate": 4.106965174129354e-05,
      "loss": 0.0052,
      "step": 2370
    },
    {
      "epoch": 3.55,
      "learning_rate": 4.082089552238806e-05,
      "loss": 0.0049,
      "step": 2380
    },
    {
      "epoch": 3.57,
      "learning_rate": 4.0572139303482585e-05,
      "loss": 0.0297,
      "step": 2390
    },
    {
      "epoch": 3.58,
      "learning_rate": 4.0323383084577115e-05,
      "loss": 0.0086,
      "step": 2400
    },
    {
      "epoch": 3.58,
      "eval_accuracy": 0.9848866498740554,
      "eval_loss": 0.08157633990049362,
      "eval_runtime": 19.2031,
      "eval_samples_per_second": 62.021,
      "eval_steps_per_second": 7.759,
      "step": 2400
    },
    {
      "epoch": 3.6,
      "learning_rate": 4.0074626865671645e-05,
      "loss": 0.0201,
      "step": 2410
    },
    {
      "epoch": 3.61,
      "learning_rate": 3.9825870646766175e-05,
      "loss": 0.0043,
      "step": 2420
    },
    {
      "epoch": 3.63,
      "learning_rate": 3.95771144278607e-05,
      "loss": 0.0037,
      "step": 2430
    },
    {
      "epoch": 3.64,
      "learning_rate": 3.932835820895522e-05,
      "loss": 0.0038,
      "step": 2440
    },
    {
      "epoch": 3.66,
      "learning_rate": 3.907960199004975e-05,
      "loss": 0.0698,
      "step": 2450
    },
    {
      "epoch": 3.67,
      "learning_rate": 3.883084577114428e-05,
      "loss": 0.0037,
      "step": 2460
    },
    {
      "epoch": 3.69,
      "learning_rate": 3.858208955223881e-05,
      "loss": 0.0039,
      "step": 2470
    },
    {
      "epoch": 3.7,
      "learning_rate": 3.8333333333333334e-05,
      "loss": 0.0267,
      "step": 2480
    },
    {
      "epoch": 3.72,
      "learning_rate": 3.808457711442786e-05,
      "loss": 0.0035,
      "step": 2490
    },
    {
      "epoch": 3.73,
      "learning_rate": 3.783582089552239e-05,
      "loss": 0.0212,
      "step": 2500
    },
    {
      "epoch": 3.73,
      "eval_accuracy": 0.9848866498740554,
      "eval_loss": 0.057428497821092606,
      "eval_runtime": 19.0519,
      "eval_samples_per_second": 62.513,
      "eval_steps_per_second": 7.821,
      "step": 2500
    },
    {
      "epoch": 3.75,
      "learning_rate": 3.758706467661692e-05,
      "loss": 0.0045,
      "step": 2510
    },
    {
      "epoch": 3.76,
      "learning_rate": 3.733830845771145e-05,
      "loss": 0.0036,
      "step": 2520
    },
    {
      "epoch": 3.78,
      "learning_rate": 3.708955223880598e-05,
      "loss": 0.0033,
      "step": 2530
    },
    {
      "epoch": 3.79,
      "learning_rate": 3.68407960199005e-05,
      "loss": 0.0033,
      "step": 2540
    },
    {
      "epoch": 3.81,
      "learning_rate": 3.6592039800995024e-05,
      "loss": 0.0033,
      "step": 2550
    },
    {
      "epoch": 3.82,
      "learning_rate": 3.6343283582089554e-05,
      "loss": 0.0033,
      "step": 2560
    },
    {
      "epoch": 3.84,
      "learning_rate": 3.6094527363184084e-05,
      "loss": 0.0034,
      "step": 2570
    },
    {
      "epoch": 3.85,
      "learning_rate": 3.584577114427861e-05,
      "loss": 0.0031,
      "step": 2580
    },
    {
      "epoch": 3.87,
      "learning_rate": 3.559701492537314e-05,
      "loss": 0.0034,
      "step": 2590
    },
    {
      "epoch": 3.88,
      "learning_rate": 3.534825870646766e-05,
      "loss": 0.0392,
      "step": 2600
    },
    {
      "epoch": 3.88,
      "eval_accuracy": 0.982367758186398,
      "eval_loss": 0.07654974609613419,
      "eval_runtime": 20.1007,
      "eval_samples_per_second": 59.252,
      "eval_steps_per_second": 7.413,
      "step": 2600
    },
    {
      "epoch": 3.9,
      "learning_rate": 3.509950248756219e-05,
      "loss": 0.0307,
      "step": 2610
    },
    {
      "epoch": 3.91,
      "learning_rate": 3.485074626865672e-05,
      "loss": 0.0075,
      "step": 2620
    },
    {
      "epoch": 3.93,
      "learning_rate": 3.4601990049751244e-05,
      "loss": 0.0325,
      "step": 2630
    },
    {
      "epoch": 3.94,
      "learning_rate": 3.4353233830845774e-05,
      "loss": 0.0032,
      "step": 2640
    },
    {
      "epoch": 3.96,
      "learning_rate": 3.41044776119403e-05,
      "loss": 0.0031,
      "step": 2650
    },
    {
      "epoch": 3.97,
      "learning_rate": 3.385572139303483e-05,
      "loss": 0.0151,
      "step": 2660
    },
    {
      "epoch": 3.99,
      "learning_rate": 3.360696517412936e-05,
      "loss": 0.0048,
      "step": 2670
    },
    {
      "epoch": 4.0,
      "learning_rate": 3.335820895522388e-05,
      "loss": 0.003,
      "step": 2680
    },
    {
      "epoch": 4.01,
      "learning_rate": 3.310945273631841e-05,
      "loss": 0.003,
      "step": 2690
    },
    {
      "epoch": 4.03,
      "learning_rate": 3.286069651741294e-05,
      "loss": 0.0029,
      "step": 2700
    },
    {
      "epoch": 4.03,
      "eval_accuracy": 0.979009235936188,
      "eval_loss": 0.073957659304142,
      "eval_runtime": 19.9521,
      "eval_samples_per_second": 59.693,
      "eval_steps_per_second": 7.468,
      "step": 2700
    },
    {
      "epoch": 4.04,
      "learning_rate": 3.261194029850746e-05,
      "loss": 0.0036,
      "step": 2710
    },
    {
      "epoch": 4.06,
      "learning_rate": 3.236318407960199e-05,
      "loss": 0.0039,
      "step": 2720
    },
    {
      "epoch": 4.07,
      "learning_rate": 3.2114427860696516e-05,
      "loss": 0.0026,
      "step": 2730
    },
    {
      "epoch": 4.09,
      "learning_rate": 3.1865671641791046e-05,
      "loss": 0.0031,
      "step": 2740
    },
    {
      "epoch": 4.1,
      "learning_rate": 3.1616915422885576e-05,
      "loss": 0.0031,
      "step": 2750
    },
    {
      "epoch": 4.12,
      "learning_rate": 3.13681592039801e-05,
      "loss": 0.0027,
      "step": 2760
    },
    {
      "epoch": 4.13,
      "learning_rate": 3.111940298507463e-05,
      "loss": 0.0029,
      "step": 2770
    },
    {
      "epoch": 4.15,
      "learning_rate": 3.087064676616915e-05,
      "loss": 0.0029,
      "step": 2780
    },
    {
      "epoch": 4.16,
      "learning_rate": 3.062189054726368e-05,
      "loss": 0.0028,
      "step": 2790
    },
    {
      "epoch": 4.18,
      "learning_rate": 3.037313432835821e-05,
      "loss": 0.0521,
      "step": 2800
    },
    {
      "epoch": 4.18,
      "eval_accuracy": 0.9924433249370277,
      "eval_loss": 0.03492419794201851,
      "eval_runtime": 20.5925,
      "eval_samples_per_second": 57.837,
      "eval_steps_per_second": 7.236,
      "step": 2800
    },
    {
      "epoch": 4.19,
      "learning_rate": 3.012437810945274e-05,
      "loss": 0.0029,
      "step": 2810
    },
    {
      "epoch": 4.21,
      "learning_rate": 2.9875621890547266e-05,
      "loss": 0.0029,
      "step": 2820
    },
    {
      "epoch": 4.22,
      "learning_rate": 2.962686567164179e-05,
      "loss": 0.0027,
      "step": 2830
    },
    {
      "epoch": 4.24,
      "learning_rate": 2.937810945273632e-05,
      "loss": 0.0029,
      "step": 2840
    },
    {
      "epoch": 4.25,
      "learning_rate": 2.912935323383085e-05,
      "loss": 0.0026,
      "step": 2850
    },
    {
      "epoch": 4.27,
      "learning_rate": 2.8880597014925376e-05,
      "loss": 0.004,
      "step": 2860
    },
    {
      "epoch": 4.28,
      "learning_rate": 2.8631840796019905e-05,
      "loss": 0.0027,
      "step": 2870
    },
    {
      "epoch": 4.3,
      "learning_rate": 2.838308457711443e-05,
      "loss": 0.0028,
      "step": 2880
    },
    {
      "epoch": 4.31,
      "learning_rate": 2.8134328358208955e-05,
      "loss": 0.0029,
      "step": 2890
    },
    {
      "epoch": 4.33,
      "learning_rate": 2.7885572139303485e-05,
      "loss": 0.0025,
      "step": 2900
    },
    {
      "epoch": 4.33,
      "eval_accuracy": 0.9865659109991604,
      "eval_loss": 0.06394174695014954,
      "eval_runtime": 19.857,
      "eval_samples_per_second": 59.979,
      "eval_steps_per_second": 7.504,
      "step": 2900
    },
    {
      "epoch": 4.34,
      "learning_rate": 2.7636815920398012e-05,
      "loss": 0.0028,
      "step": 2910
    },
    {
      "epoch": 4.36,
      "learning_rate": 2.7388059701492542e-05,
      "loss": 0.0027,
      "step": 2920
    },
    {
      "epoch": 4.37,
      "learning_rate": 2.7139303482587065e-05,
      "loss": 0.0026,
      "step": 2930
    },
    {
      "epoch": 4.39,
      "learning_rate": 2.689054726368159e-05,
      "loss": 0.0026,
      "step": 2940
    },
    {
      "epoch": 4.4,
      "learning_rate": 2.664179104477612e-05,
      "loss": 0.0025,
      "step": 2950
    },
    {
      "epoch": 4.42,
      "learning_rate": 2.6393034825870648e-05,
      "loss": 0.0025,
      "step": 2960
    },
    {
      "epoch": 4.43,
      "learning_rate": 2.6144278606965178e-05,
      "loss": 0.0027,
      "step": 2970
    },
    {
      "epoch": 4.45,
      "learning_rate": 2.58955223880597e-05,
      "loss": 0.0026,
      "step": 2980
    },
    {
      "epoch": 4.46,
      "learning_rate": 2.5646766169154228e-05,
      "loss": 0.0025,
      "step": 2990
    },
    {
      "epoch": 4.48,
      "learning_rate": 2.5398009950248758e-05,
      "loss": 0.0024,
      "step": 3000
    },
    {
      "epoch": 4.48,
      "eval_accuracy": 0.9882451721242653,
      "eval_loss": 0.050867918878793716,
      "eval_runtime": 19.0308,
      "eval_samples_per_second": 62.583,
      "eval_steps_per_second": 7.829,
      "step": 3000
    },
    {
      "epoch": 4.49,
      "learning_rate": 2.5149253731343288e-05,
      "loss": 0.0033,
      "step": 3010
    },
    {
      "epoch": 4.51,
      "learning_rate": 2.490049751243781e-05,
      "loss": 0.0026,
      "step": 3020
    },
    {
      "epoch": 4.52,
      "learning_rate": 2.465174129353234e-05,
      "loss": 0.0026,
      "step": 3030
    },
    {
      "epoch": 4.54,
      "learning_rate": 2.4402985074626868e-05,
      "loss": 0.0025,
      "step": 3040
    },
    {
      "epoch": 4.55,
      "learning_rate": 2.4154228855721394e-05,
      "loss": 0.0024,
      "step": 3050
    },
    {
      "epoch": 4.57,
      "learning_rate": 2.390547263681592e-05,
      "loss": 0.0025,
      "step": 3060
    },
    {
      "epoch": 4.58,
      "learning_rate": 2.3656716417910448e-05,
      "loss": 0.0025,
      "step": 3070
    },
    {
      "epoch": 4.6,
      "learning_rate": 2.3407960199004977e-05,
      "loss": 0.0361,
      "step": 3080
    },
    {
      "epoch": 4.61,
      "learning_rate": 2.3159203980099504e-05,
      "loss": 0.0024,
      "step": 3090
    },
    {
      "epoch": 4.63,
      "learning_rate": 2.291044776119403e-05,
      "loss": 0.0024,
      "step": 3100
    },
    {
      "epoch": 4.63,
      "eval_accuracy": 0.9899244332493703,
      "eval_loss": 0.0467989481985569,
      "eval_runtime": 19.003,
      "eval_samples_per_second": 62.674,
      "eval_steps_per_second": 7.841,
      "step": 3100
    },
    {
      "epoch": 4.64,
      "learning_rate": 2.2661691542288557e-05,
      "loss": 0.0037,
      "step": 3110
    },
    {
      "epoch": 4.66,
      "learning_rate": 2.2412935323383087e-05,
      "loss": 0.0125,
      "step": 3120
    },
    {
      "epoch": 4.67,
      "learning_rate": 2.2164179104477614e-05,
      "loss": 0.0025,
      "step": 3130
    },
    {
      "epoch": 4.69,
      "learning_rate": 2.191542288557214e-05,
      "loss": 0.0026,
      "step": 3140
    },
    {
      "epoch": 4.7,
      "learning_rate": 2.1666666666666667e-05,
      "loss": 0.0026,
      "step": 3150
    },
    {
      "epoch": 4.72,
      "learning_rate": 2.1417910447761194e-05,
      "loss": 0.0023,
      "step": 3160
    },
    {
      "epoch": 4.73,
      "learning_rate": 2.1169154228855724e-05,
      "loss": 0.0022,
      "step": 3170
    },
    {
      "epoch": 4.75,
      "learning_rate": 2.092039800995025e-05,
      "loss": 0.0024,
      "step": 3180
    },
    {
      "epoch": 4.76,
      "learning_rate": 2.0671641791044777e-05,
      "loss": 0.0023,
      "step": 3190
    },
    {
      "epoch": 4.78,
      "learning_rate": 2.0422885572139307e-05,
      "loss": 0.0024,
      "step": 3200
    },
    {
      "epoch": 4.78,
      "eval_accuracy": 0.9882451721242653,
      "eval_loss": 0.052428748458623886,
      "eval_runtime": 20.9205,
      "eval_samples_per_second": 56.93,
      "eval_steps_per_second": 7.122,
      "step": 3200
    },
    {
      "epoch": 4.79,
      "learning_rate": 2.017412935323383e-05,
      "loss": 0.0132,
      "step": 3210
    },
    {
      "epoch": 4.81,
      "learning_rate": 1.992537313432836e-05,
      "loss": 0.0023,
      "step": 3220
    },
    {
      "epoch": 4.82,
      "learning_rate": 1.9676616915422887e-05,
      "loss": 0.0024,
      "step": 3230
    },
    {
      "epoch": 4.84,
      "learning_rate": 1.9427860696517413e-05,
      "loss": 0.0023,
      "step": 3240
    },
    {
      "epoch": 4.85,
      "learning_rate": 1.9179104477611943e-05,
      "loss": 0.0022,
      "step": 3250
    },
    {
      "epoch": 4.87,
      "learning_rate": 1.8930348258706466e-05,
      "loss": 0.0025,
      "step": 3260
    },
    {
      "epoch": 4.88,
      "learning_rate": 1.8681592039800996e-05,
      "loss": 0.0028,
      "step": 3270
    },
    {
      "epoch": 4.9,
      "learning_rate": 1.8432835820895523e-05,
      "loss": 0.0022,
      "step": 3280
    },
    {
      "epoch": 4.91,
      "learning_rate": 1.818407960199005e-05,
      "loss": 0.0021,
      "step": 3290
    },
    {
      "epoch": 4.93,
      "learning_rate": 1.793532338308458e-05,
      "loss": 0.0022,
      "step": 3300
    },
    {
      "epoch": 4.93,
      "eval_accuracy": 0.9899244332493703,
      "eval_loss": 0.0504080168902874,
      "eval_runtime": 19.0283,
      "eval_samples_per_second": 62.591,
      "eval_steps_per_second": 7.83,
      "step": 3300
    },
    {
      "epoch": 4.94,
      "learning_rate": 1.7686567164179106e-05,
      "loss": 0.0021,
      "step": 3310
    },
    {
      "epoch": 4.96,
      "learning_rate": 1.7437810945273633e-05,
      "loss": 0.0023,
      "step": 3320
    },
    {
      "epoch": 4.97,
      "learning_rate": 1.718905472636816e-05,
      "loss": 0.0022,
      "step": 3330
    },
    {
      "epoch": 4.99,
      "learning_rate": 1.6940298507462686e-05,
      "loss": 0.0021,
      "step": 3340
    },
    {
      "epoch": 5.0,
      "learning_rate": 1.6691542288557216e-05,
      "loss": 0.002,
      "step": 3350
    },
    {
      "epoch": 5.01,
      "learning_rate": 1.6442786069651742e-05,
      "loss": 0.0021,
      "step": 3360
    },
    {
      "epoch": 5.03,
      "learning_rate": 1.619402985074627e-05,
      "loss": 0.0021,
      "step": 3370
    },
    {
      "epoch": 5.04,
      "learning_rate": 1.5945273631840796e-05,
      "loss": 0.0021,
      "step": 3380
    },
    {
      "epoch": 5.06,
      "learning_rate": 1.5696517412935326e-05,
      "loss": 0.0022,
      "step": 3390
    },
    {
      "epoch": 5.07,
      "learning_rate": 1.5447761194029852e-05,
      "loss": 0.0022,
      "step": 3400
    },
    {
      "epoch": 5.07,
      "eval_accuracy": 0.9890848026868178,
      "eval_loss": 0.048807818442583084,
      "eval_runtime": 20.0305,
      "eval_samples_per_second": 59.459,
      "eval_steps_per_second": 7.439,
      "step": 3400
    },
    {
      "epoch": 5.09,
      "learning_rate": 1.5199004975124379e-05,
      "loss": 0.002,
      "step": 3410
    },
    {
      "epoch": 5.1,
      "learning_rate": 1.4950248756218907e-05,
      "loss": 0.0021,
      "step": 3420
    },
    {
      "epoch": 5.12,
      "learning_rate": 1.4701492537313432e-05,
      "loss": 0.0023,
      "step": 3430
    },
    {
      "epoch": 5.13,
      "learning_rate": 1.4452736318407962e-05,
      "loss": 0.0022,
      "step": 3440
    },
    {
      "epoch": 5.15,
      "learning_rate": 1.4203980099502487e-05,
      "loss": 0.002,
      "step": 3450
    },
    {
      "epoch": 5.16,
      "learning_rate": 1.3955223880597015e-05,
      "loss": 0.002,
      "step": 3460
    },
    {
      "epoch": 5.18,
      "learning_rate": 1.3706467661691543e-05,
      "loss": 0.0022,
      "step": 3470
    },
    {
      "epoch": 5.19,
      "learning_rate": 1.345771144278607e-05,
      "loss": 0.0021,
      "step": 3480
    },
    {
      "epoch": 5.21,
      "learning_rate": 1.3208955223880598e-05,
      "loss": 0.0021,
      "step": 3490
    },
    {
      "epoch": 5.22,
      "learning_rate": 1.2960199004975123e-05,
      "loss": 0.0021,
      "step": 3500
    },
    {
      "epoch": 5.22,
      "eval_accuracy": 0.9899244332493703,
      "eval_loss": 0.04841174930334091,
      "eval_runtime": 19.9366,
      "eval_samples_per_second": 59.739,
      "eval_steps_per_second": 7.474,
      "step": 3500
    },
    {
      "epoch": 5.24,
      "learning_rate": 1.2711442786069651e-05,
      "loss": 0.002,
      "step": 3510
    },
    {
      "epoch": 5.25,
      "learning_rate": 1.246268656716418e-05,
      "loss": 0.002,
      "step": 3520
    },
    {
      "epoch": 5.27,
      "learning_rate": 1.2213930348258706e-05,
      "loss": 0.002,
      "step": 3530
    },
    {
      "epoch": 5.28,
      "learning_rate": 1.1965174129353235e-05,
      "loss": 0.002,
      "step": 3540
    },
    {
      "epoch": 5.3,
      "learning_rate": 1.1716417910447761e-05,
      "loss": 0.002,
      "step": 3550
    },
    {
      "epoch": 5.31,
      "learning_rate": 1.146766169154229e-05,
      "loss": 0.0021,
      "step": 3560
    },
    {
      "epoch": 5.33,
      "learning_rate": 1.1218905472636816e-05,
      "loss": 0.0021,
      "step": 3570
    },
    {
      "epoch": 5.34,
      "learning_rate": 1.0970149253731343e-05,
      "loss": 0.0021,
      "step": 3580
    },
    {
      "epoch": 5.36,
      "learning_rate": 1.0721393034825871e-05,
      "loss": 0.002,
      "step": 3590
    },
    {
      "epoch": 5.37,
      "learning_rate": 1.04726368159204e-05,
      "loss": 0.0019,
      "step": 3600
    },
    {
      "epoch": 5.37,
      "eval_accuracy": 0.9899244332493703,
      "eval_loss": 0.04795178398489952,
      "eval_runtime": 19.4137,
      "eval_samples_per_second": 61.348,
      "eval_steps_per_second": 7.675,
      "step": 3600
    },
    {
      "epoch": 5.39,
      "learning_rate": 1.0223880597014926e-05,
      "loss": 0.002,
      "step": 3610
    },
    {
      "epoch": 5.4,
      "learning_rate": 9.975124378109452e-06,
      "loss": 0.002,
      "step": 3620
    },
    {
      "epoch": 5.42,
      "learning_rate": 9.72636815920398e-06,
      "loss": 0.002,
      "step": 3630
    },
    {
      "epoch": 5.43,
      "learning_rate": 9.477611940298507e-06,
      "loss": 0.002,
      "step": 3640
    },
    {
      "epoch": 5.45,
      "learning_rate": 9.228855721393036e-06,
      "loss": 0.002,
      "step": 3650
    },
    {
      "epoch": 5.46,
      "learning_rate": 8.980099502487562e-06,
      "loss": 0.0019,
      "step": 3660
    },
    {
      "epoch": 5.48,
      "learning_rate": 8.73134328358209e-06,
      "loss": 0.0019,
      "step": 3670
    },
    {
      "epoch": 5.49,
      "learning_rate": 8.482587064676617e-06,
      "loss": 0.0019,
      "step": 3680
    },
    {
      "epoch": 5.51,
      "learning_rate": 8.233830845771144e-06,
      "loss": 0.002,
      "step": 3690
    },
    {
      "epoch": 5.52,
      "learning_rate": 7.985074626865672e-06,
      "loss": 0.0019,
      "step": 3700
    },
    {
      "epoch": 5.52,
      "eval_accuracy": 0.9899244332493703,
      "eval_loss": 0.04765713959932327,
      "eval_runtime": 19.0716,
      "eval_samples_per_second": 62.449,
      "eval_steps_per_second": 7.813,
      "step": 3700
    },
    {
      "epoch": 5.54,
      "learning_rate": 7.7363184079602e-06,
      "loss": 0.002,
      "step": 3710
    },
    {
      "epoch": 5.55,
      "learning_rate": 7.487562189054727e-06,
      "loss": 0.0019,
      "step": 3720
    },
    {
      "epoch": 5.57,
      "learning_rate": 7.238805970149254e-06,
      "loss": 0.002,
      "step": 3730
    },
    {
      "epoch": 5.58,
      "learning_rate": 6.990049751243781e-06,
      "loss": 0.0019,
      "step": 3740
    },
    {
      "epoch": 5.6,
      "learning_rate": 6.741293532338309e-06,
      "loss": 0.0019,
      "step": 3750
    },
    {
      "epoch": 5.61,
      "learning_rate": 6.492537313432837e-06,
      "loss": 0.002,
      "step": 3760
    },
    {
      "epoch": 5.63,
      "learning_rate": 6.243781094527364e-06,
      "loss": 0.002,
      "step": 3770
    },
    {
      "epoch": 5.64,
      "learning_rate": 5.995024875621891e-06,
      "loss": 0.002,
      "step": 3780
    },
    {
      "epoch": 5.66,
      "learning_rate": 5.746268656716418e-06,
      "loss": 0.002,
      "step": 3790
    },
    {
      "epoch": 5.67,
      "learning_rate": 5.4975124378109455e-06,
      "loss": 0.0019,
      "step": 3800
    },
    {
      "epoch": 5.67,
      "eval_accuracy": 0.9899244332493703,
      "eval_loss": 0.04748144745826721,
      "eval_runtime": 19.2405,
      "eval_samples_per_second": 61.901,
      "eval_steps_per_second": 7.744,
      "step": 3800
    },
    {
      "epoch": 5.69,
      "learning_rate": 5.248756218905473e-06,
      "loss": 0.002,
      "step": 3810
    },
    {
      "epoch": 5.7,
      "learning_rate": 5e-06,
      "loss": 0.0019,
      "step": 3820
    },
    {
      "epoch": 5.72,
      "learning_rate": 4.751243781094528e-06,
      "loss": 0.0019,
      "step": 3830
    },
    {
      "epoch": 5.73,
      "learning_rate": 4.5024875621890544e-06,
      "loss": 0.0019,
      "step": 3840
    },
    {
      "epoch": 5.75,
      "learning_rate": 4.253731343283583e-06,
      "loss": 0.0018,
      "step": 3850
    },
    {
      "epoch": 5.76,
      "learning_rate": 4.004975124378109e-06,
      "loss": 0.002,
      "step": 3860
    },
    {
      "epoch": 5.78,
      "learning_rate": 3.7562189054726368e-06,
      "loss": 0.0018,
      "step": 3870
    },
    {
      "epoch": 5.79,
      "learning_rate": 3.5074626865671646e-06,
      "loss": 0.0019,
      "step": 3880
    },
    {
      "epoch": 5.81,
      "learning_rate": 3.2587064676616916e-06,
      "loss": 0.0019,
      "step": 3890
    },
    {
      "epoch": 5.82,
      "learning_rate": 3.009950248756219e-06,
      "loss": 0.0018,
      "step": 3900
    },
    {
      "epoch": 5.82,
      "eval_accuracy": 0.9899244332493703,
      "eval_loss": 0.0473811998963356,
      "eval_runtime": 22.4733,
      "eval_samples_per_second": 52.996,
      "eval_steps_per_second": 6.63,
      "step": 3900
    },
    {
      "epoch": 5.84,
      "learning_rate": 2.7611940298507465e-06,
      "loss": 0.002,
      "step": 3910
    },
    {
      "epoch": 5.85,
      "learning_rate": 2.512437810945274e-06,
      "loss": 0.0018,
      "step": 3920
    },
    {
      "epoch": 5.87,
      "learning_rate": 2.263681592039801e-06,
      "loss": 0.0019,
      "step": 3930
    },
    {
      "epoch": 5.88,
      "learning_rate": 2.0149253731343284e-06,
      "loss": 0.0018,
      "step": 3940
    },
    {
      "epoch": 5.9,
      "learning_rate": 1.7661691542288559e-06,
      "loss": 0.002,
      "step": 3950
    },
    {
      "epoch": 5.91,
      "learning_rate": 1.517412935323383e-06,
      "loss": 0.0019,
      "step": 3960
    },
    {
      "epoch": 5.93,
      "learning_rate": 1.2686567164179105e-06,
      "loss": 0.0019,
      "step": 3970
    },
    {
      "epoch": 5.94,
      "learning_rate": 1.019900497512438e-06,
      "loss": 0.002,
      "step": 3980
    },
    {
      "epoch": 5.96,
      "learning_rate": 7.711442786069652e-07,
      "loss": 0.0019,
      "step": 3990
    },
    {
      "epoch": 5.97,
      "learning_rate": 5.223880597014926e-07,
      "loss": 0.0018,
      "step": 4000
    },
    {
      "epoch": 5.97,
      "eval_accuracy": 0.9899244332493703,
      "eval_loss": 0.047312360256910324,
      "eval_runtime": 19.2838,
      "eval_samples_per_second": 61.762,
      "eval_steps_per_second": 7.727,
      "step": 4000
    },
    {
      "epoch": 5.99,
      "learning_rate": 2.736318407960199e-07,
      "loss": 0.0019,
      "step": 4010
    },
    {
      "epoch": 6.0,
      "learning_rate": 2.4875621890547262e-08,
      "loss": 0.002,
      "step": 4020
    },
    {
      "epoch": 6.0,
      "step": 4020,
      "total_flos": 4.98440050928135e+18,
      "train_loss": 0.10442606876627426,
      "train_runtime": 2908.5673,
      "train_samples_per_second": 22.112,
      "train_steps_per_second": 1.382
    }
  ],
  "logging_steps": 10,
  "max_steps": 4020,
  "num_train_epochs": 6,
  "save_steps": 100,
  "total_flos": 4.98440050928135e+18,
  "trial_name": null,
  "trial_params": null
}