{
  "best_metric": 0.517457902431488,
  "best_model_checkpoint": "./kd_results/microsoft/beit-base-finetuned-ade-640-640_alpha0.7_temp5.0_t2/checkpoint-3600",
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 3600,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1388888888888889,
      "grad_norm": 7.222753047943115,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 1.1565,
      "step": 25
    },
    {
      "epoch": 0.2777777777777778,
      "grad_norm": 8.318293571472168,
      "learning_rate": 6.944444444444445e-06,
      "loss": 1.1518,
      "step": 50
    },
    {
      "epoch": 0.4166666666666667,
      "grad_norm": 5.613315582275391,
      "learning_rate": 1.0416666666666668e-05,
      "loss": 1.1668,
      "step": 75
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 6.886393070220947,
      "learning_rate": 1.388888888888889e-05,
      "loss": 1.1685,
      "step": 100
    },
    {
      "epoch": 0.6944444444444444,
      "grad_norm": 4.675456523895264,
      "learning_rate": 1.736111111111111e-05,
      "loss": 1.1258,
      "step": 125
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 6.820631504058838,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 1.0927,
      "step": 150
    },
    {
      "epoch": 0.9722222222222222,
      "grad_norm": 9.431523323059082,
      "learning_rate": 2.4305555555555558e-05,
      "loss": 1.1235,
      "step": 175
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.3932806324110672,
      "eval_loss": 1.3238857984542847,
      "eval_runtime": 72.8485,
      "eval_samples_per_second": 13.892,
      "eval_steps_per_second": 0.439,
      "step": 180
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 10.79791259765625,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.9992,
      "step": 200
    },
    {
      "epoch": 1.25,
      "grad_norm": 6.876852512359619,
      "learning_rate": 3.125e-05,
      "loss": 0.9959,
      "step": 225
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 5.41601037979126,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.9272,
      "step": 250
    },
    {
      "epoch": 1.5277777777777777,
      "grad_norm": 5.712553024291992,
      "learning_rate": 3.8194444444444444e-05,
      "loss": 0.8841,
      "step": 275
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 6.366713523864746,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.779,
      "step": 300
    },
    {
      "epoch": 1.8055555555555556,
      "grad_norm": 9.597274780273438,
      "learning_rate": 4.5138888888888894e-05,
      "loss": 0.6792,
      "step": 325
    },
    {
      "epoch": 1.9444444444444444,
      "grad_norm": 5.85896635055542,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.6303,
      "step": 350
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6877470355731226,
      "eval_loss": 0.841946542263031,
      "eval_runtime": 68.9674,
      "eval_samples_per_second": 14.674,
      "eval_steps_per_second": 0.464,
      "step": 360
    },
    {
      "epoch": 2.0833333333333335,
      "grad_norm": 5.268279075622559,
      "learning_rate": 4.976851851851852e-05,
      "loss": 0.5683,
      "step": 375
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 7.627471923828125,
      "learning_rate": 4.938271604938271e-05,
      "loss": 0.5024,
      "step": 400
    },
    {
      "epoch": 2.361111111111111,
      "grad_norm": 8.074167251586914,
      "learning_rate": 4.899691358024692e-05,
      "loss": 0.5199,
      "step": 425
    },
    {
      "epoch": 2.5,
      "grad_norm": 6.42832612991333,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.5179,
      "step": 450
    },
    {
      "epoch": 2.638888888888889,
      "grad_norm": 7.448732852935791,
      "learning_rate": 4.8225308641975306e-05,
      "loss": 0.528,
      "step": 475
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 7.955748558044434,
      "learning_rate": 4.783950617283951e-05,
      "loss": 0.4876,
      "step": 500
    },
    {
      "epoch": 2.9166666666666665,
      "grad_norm": 7.4806742668151855,
      "learning_rate": 4.745370370370371e-05,
      "loss": 0.471,
      "step": 525
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7322134387351779,
      "eval_loss": 0.7386972904205322,
      "eval_runtime": 71.6196,
      "eval_samples_per_second": 14.13,
      "eval_steps_per_second": 0.447,
      "step": 540
    },
    {
      "epoch": 3.0555555555555554,
      "grad_norm": 4.674981117248535,
      "learning_rate": 4.70679012345679e-05,
      "loss": 0.4377,
      "step": 550
    },
    {
      "epoch": 3.1944444444444446,
      "grad_norm": 4.392780303955078,
      "learning_rate": 4.66820987654321e-05,
      "loss": 0.4075,
      "step": 575
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 6.68107795715332,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.3958,
      "step": 600
    },
    {
      "epoch": 3.4722222222222223,
      "grad_norm": 8.04480266571045,
      "learning_rate": 4.591049382716049e-05,
      "loss": 0.411,
      "step": 625
    },
    {
      "epoch": 3.611111111111111,
      "grad_norm": 5.624967098236084,
      "learning_rate": 4.5524691358024696e-05,
      "loss": 0.3841,
      "step": 650
    },
    {
      "epoch": 3.75,
      "grad_norm": 5.001491546630859,
      "learning_rate": 4.5138888888888894e-05,
      "loss": 0.3445,
      "step": 675
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 3.7939069271087646,
      "learning_rate": 4.4753086419753084e-05,
      "loss": 0.3553,
      "step": 700
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7737154150197628,
      "eval_loss": 0.6575068831443787,
      "eval_runtime": 73.802,
      "eval_samples_per_second": 13.712,
      "eval_steps_per_second": 0.434,
      "step": 720
    },
    {
      "epoch": 4.027777777777778,
      "grad_norm": 3.7097816467285156,
      "learning_rate": 4.436728395061729e-05,
      "loss": 0.367,
      "step": 725
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 2.837089776992798,
      "learning_rate": 4.3981481481481486e-05,
      "loss": 0.3453,
      "step": 750
    },
    {
      "epoch": 4.305555555555555,
      "grad_norm": 3.6544575691223145,
      "learning_rate": 4.359567901234568e-05,
      "loss": 0.2949,
      "step": 775
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 4.620490550994873,
      "learning_rate": 4.3209876543209875e-05,
      "loss": 0.3194,
      "step": 800
    },
    {
      "epoch": 4.583333333333333,
      "grad_norm": 4.1617937088012695,
      "learning_rate": 4.282407407407408e-05,
      "loss": 0.3188,
      "step": 825
    },
    {
      "epoch": 4.722222222222222,
      "grad_norm": 5.08017635345459,
      "learning_rate": 4.243827160493827e-05,
      "loss": 0.2805,
      "step": 850
    },
    {
      "epoch": 4.861111111111111,
      "grad_norm": 6.545626640319824,
      "learning_rate": 4.205246913580247e-05,
      "loss": 0.2989,
      "step": 875
    },
    {
      "epoch": 5.0,
      "grad_norm": 11.199858665466309,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.3071,
      "step": 900
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.782608695652174,
      "eval_loss": 0.6260281205177307,
      "eval_runtime": 72.9279,
      "eval_samples_per_second": 13.877,
      "eval_steps_per_second": 0.439,
      "step": 900
    },
    {
      "epoch": 5.138888888888889,
      "grad_norm": 8.157869338989258,
      "learning_rate": 4.128086419753087e-05,
      "loss": 0.2608,
      "step": 925
    },
    {
      "epoch": 5.277777777777778,
      "grad_norm": 6.550694942474365,
      "learning_rate": 4.089506172839506e-05,
      "loss": 0.2635,
      "step": 950
    },
    {
      "epoch": 5.416666666666667,
      "grad_norm": 4.43126916885376,
      "learning_rate": 4.0509259259259265e-05,
      "loss": 0.2524,
      "step": 975
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 5.895445823669434,
      "learning_rate": 4.012345679012346e-05,
      "loss": 0.2753,
      "step": 1000
    },
    {
      "epoch": 5.694444444444445,
      "grad_norm": 4.39314079284668,
      "learning_rate": 3.973765432098765e-05,
      "loss": 0.2798,
      "step": 1025
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 7.890945911407471,
      "learning_rate": 3.935185185185186e-05,
      "loss": 0.2533,
      "step": 1050
    },
    {
      "epoch": 5.972222222222222,
      "grad_norm": 3.788822650909424,
      "learning_rate": 3.8966049382716055e-05,
      "loss": 0.2206,
      "step": 1075
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7964426877470355,
      "eval_loss": 0.60780268907547,
      "eval_runtime": 72.669,
      "eval_samples_per_second": 13.926,
      "eval_steps_per_second": 0.44,
      "step": 1080
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 2.5203280448913574,
      "learning_rate": 3.8580246913580246e-05,
      "loss": 0.2321,
      "step": 1100
    },
    {
      "epoch": 6.25,
      "grad_norm": 4.587474822998047,
      "learning_rate": 3.8194444444444444e-05,
      "loss": 0.2245,
      "step": 1125
    },
    {
      "epoch": 6.388888888888889,
      "grad_norm": 3.7738234996795654,
      "learning_rate": 3.780864197530865e-05,
      "loss": 0.2278,
      "step": 1150
    },
    {
      "epoch": 6.527777777777778,
      "grad_norm": 3.1061084270477295,
      "learning_rate": 3.742283950617284e-05,
      "loss": 0.1915,
      "step": 1175
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 3.019678831100464,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.2292,
      "step": 1200
    },
    {
      "epoch": 6.805555555555555,
      "grad_norm": 2.68041729927063,
      "learning_rate": 3.665123456790124e-05,
      "loss": 0.2283,
      "step": 1225
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 4.879120826721191,
      "learning_rate": 3.626543209876543e-05,
      "loss": 0.2027,
      "step": 1250
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8053359683794467,
      "eval_loss": 0.5722182393074036,
      "eval_runtime": 73.8232,
      "eval_samples_per_second": 13.708,
      "eval_steps_per_second": 0.433,
      "step": 1260
    },
    {
      "epoch": 7.083333333333333,
      "grad_norm": 3.7718591690063477,
      "learning_rate": 3.587962962962963e-05,
      "loss": 0.1954,
      "step": 1275
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 2.43890118598938,
      "learning_rate": 3.5493827160493834e-05,
      "loss": 0.1957,
      "step": 1300
    },
    {
      "epoch": 7.361111111111111,
      "grad_norm": 2.6753463745117188,
      "learning_rate": 3.5108024691358025e-05,
      "loss": 0.1994,
      "step": 1325
    },
    {
      "epoch": 7.5,
      "grad_norm": 1.5570522546768188,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.1913,
      "step": 1350
    },
    {
      "epoch": 7.638888888888889,
      "grad_norm": 4.230154037475586,
      "learning_rate": 3.4336419753086427e-05,
      "loss": 0.1889,
      "step": 1375
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 2.0178449153900146,
      "learning_rate": 3.395061728395062e-05,
      "loss": 0.1914,
      "step": 1400
    },
    {
      "epoch": 7.916666666666667,
      "grad_norm": 5.009683609008789,
      "learning_rate": 3.3564814814814815e-05,
      "loss": 0.1922,
      "step": 1425
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8092885375494071,
      "eval_loss": 0.5998839139938354,
      "eval_runtime": 72.2786,
      "eval_samples_per_second": 14.001,
      "eval_steps_per_second": 0.443,
      "step": 1440
    },
    {
      "epoch": 8.055555555555555,
      "grad_norm": 4.476706027984619,
      "learning_rate": 3.317901234567901e-05,
      "loss": 0.1881,
      "step": 1450
    },
    {
      "epoch": 8.194444444444445,
      "grad_norm": 3.35538649559021,
      "learning_rate": 3.279320987654321e-05,
      "loss": 0.1706,
      "step": 1475
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 5.0826826095581055,
      "learning_rate": 3.240740740740741e-05,
      "loss": 0.1843,
      "step": 1500
    },
    {
      "epoch": 8.472222222222221,
      "grad_norm": 2.089994192123413,
      "learning_rate": 3.2021604938271605e-05,
      "loss": 0.1752,
      "step": 1525
    },
    {
      "epoch": 8.61111111111111,
      "grad_norm": 2.181107521057129,
      "learning_rate": 3.16358024691358e-05,
      "loss": 0.1795,
      "step": 1550
    },
    {
      "epoch": 8.75,
      "grad_norm": 2.4433932304382324,
      "learning_rate": 3.125e-05,
      "loss": 0.1806,
      "step": 1575
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 2.5213308334350586,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.1801,
      "step": 1600
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7984189723320159,
      "eval_loss": 0.6050839424133301,
      "eval_runtime": 71.6254,
      "eval_samples_per_second": 14.129,
      "eval_steps_per_second": 0.447,
      "step": 1620
    },
    {
      "epoch": 9.027777777777779,
      "grad_norm": 2.464143753051758,
      "learning_rate": 3.04783950617284e-05,
      "loss": 0.1823,
      "step": 1625
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 1.9522860050201416,
      "learning_rate": 3.0092592592592593e-05,
      "loss": 0.1697,
      "step": 1650
    },
    {
      "epoch": 9.305555555555555,
      "grad_norm": 3.385920524597168,
      "learning_rate": 2.970679012345679e-05,
      "loss": 0.1632,
      "step": 1675
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 3.4966518878936768,
      "learning_rate": 2.9320987654320992e-05,
      "loss": 0.1725,
      "step": 1700
    },
    {
      "epoch": 9.583333333333334,
      "grad_norm": 2.042785406112671,
      "learning_rate": 2.8935185185185186e-05,
      "loss": 0.1652,
      "step": 1725
    },
    {
      "epoch": 9.722222222222221,
      "grad_norm": 1.7008278369903564,
      "learning_rate": 2.8549382716049384e-05,
      "loss": 0.1649,
      "step": 1750
    },
    {
      "epoch": 9.86111111111111,
      "grad_norm": 2.4193341732025146,
      "learning_rate": 2.8163580246913578e-05,
      "loss": 0.1575,
      "step": 1775
    },
    {
      "epoch": 10.0,
      "grad_norm": 9.32138442993164,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.1758,
      "step": 1800
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.8063241106719368,
      "eval_loss": 0.572110652923584,
      "eval_runtime": 73.2828,
      "eval_samples_per_second": 13.81,
      "eval_steps_per_second": 0.437,
      "step": 1800
    },
    {
      "epoch": 10.13888888888889,
      "grad_norm": 2.2760612964630127,
      "learning_rate": 2.7391975308641977e-05,
      "loss": 0.1435,
      "step": 1825
    },
    {
      "epoch": 10.277777777777779,
      "grad_norm": 2.5033085346221924,
      "learning_rate": 2.700617283950617e-05,
      "loss": 0.1507,
      "step": 1850
    },
    {
      "epoch": 10.416666666666666,
      "grad_norm": 3.035231590270996,
      "learning_rate": 2.6620370370370372e-05,
      "loss": 0.1615,
      "step": 1875
    },
    {
      "epoch": 10.555555555555555,
      "grad_norm": 1.2022318840026855,
      "learning_rate": 2.623456790123457e-05,
      "loss": 0.1507,
      "step": 1900
    },
    {
      "epoch": 10.694444444444445,
      "grad_norm": 1.8878600597381592,
      "learning_rate": 2.5848765432098764e-05,
      "loss": 0.1605,
      "step": 1925
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 1.6575684547424316,
      "learning_rate": 2.5462962962962965e-05,
      "loss": 0.1535,
      "step": 1950
    },
    {
      "epoch": 10.972222222222221,
      "grad_norm": 1.5876245498657227,
      "learning_rate": 2.5077160493827162e-05,
      "loss": 0.154,
      "step": 1975
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.8102766798418972,
      "eval_loss": 0.5764340758323669,
      "eval_runtime": 69.3741,
      "eval_samples_per_second": 14.588,
      "eval_steps_per_second": 0.461,
      "step": 1980
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 2.0368149280548096,
      "learning_rate": 2.4691358024691357e-05,
      "loss": 0.1449,
      "step": 2000
    },
    {
      "epoch": 11.25,
      "grad_norm": 1.995473027229309,
      "learning_rate": 2.4305555555555558e-05,
      "loss": 0.1484,
      "step": 2025
    },
    {
      "epoch": 11.38888888888889,
      "grad_norm": 2.1665401458740234,
      "learning_rate": 2.3919753086419755e-05,
      "loss": 0.145,
      "step": 2050
    },
    {
      "epoch": 11.527777777777779,
      "grad_norm": 1.5415103435516357,
      "learning_rate": 2.353395061728395e-05,
      "loss": 0.1472,
      "step": 2075
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 1.7828869819641113,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.1468,
      "step": 2100
    },
    {
      "epoch": 11.805555555555555,
      "grad_norm": 1.8061931133270264,
      "learning_rate": 2.2762345679012348e-05,
      "loss": 0.1542,
      "step": 2125
    },
    {
      "epoch": 11.944444444444445,
      "grad_norm": 1.6420892477035522,
      "learning_rate": 2.2376543209876542e-05,
      "loss": 0.1495,
      "step": 2150
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.817193675889328,
      "eval_loss": 0.5539035201072693,
      "eval_runtime": 72.0331,
      "eval_samples_per_second": 14.049,
      "eval_steps_per_second": 0.444,
      "step": 2160
    },
    {
      "epoch": 12.083333333333334,
      "grad_norm": 1.1013528108596802,
      "learning_rate": 2.1990740740740743e-05,
      "loss": 0.1324,
      "step": 2175
    },
    {
      "epoch": 12.222222222222221,
      "grad_norm": 1.1566153764724731,
      "learning_rate": 2.1604938271604937e-05,
      "loss": 0.1322,
      "step": 2200
    },
    {
      "epoch": 12.36111111111111,
      "grad_norm": 2.9402618408203125,
      "learning_rate": 2.1219135802469135e-05,
      "loss": 0.1395,
      "step": 2225
    },
    {
      "epoch": 12.5,
      "grad_norm": 1.4940177202224731,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.149,
      "step": 2250
    },
    {
      "epoch": 12.63888888888889,
      "grad_norm": 1.2324851751327515,
      "learning_rate": 2.044753086419753e-05,
      "loss": 0.1446,
      "step": 2275
    },
    {
      "epoch": 12.777777777777779,
      "grad_norm": 1.6738492250442505,
      "learning_rate": 2.006172839506173e-05,
      "loss": 0.131,
      "step": 2300
    },
    {
      "epoch": 12.916666666666666,
      "grad_norm": 1.5494078397750854,
      "learning_rate": 1.967592592592593e-05,
      "loss": 0.1334,
      "step": 2325
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.8389328063241107,
      "eval_loss": 0.527151346206665,
      "eval_runtime": 71.8766,
      "eval_samples_per_second": 14.08,
      "eval_steps_per_second": 0.445,
      "step": 2340
    },
    {
      "epoch": 13.055555555555555,
      "grad_norm": 1.056087613105774,
      "learning_rate": 1.9290123456790123e-05,
      "loss": 0.1444,
      "step": 2350
    },
    {
      "epoch": 13.194444444444445,
      "grad_norm": 1.6478755474090576,
      "learning_rate": 1.8904320987654324e-05,
      "loss": 0.1372,
      "step": 2375
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 1.4524025917053223,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.1338,
      "step": 2400
    },
    {
      "epoch": 13.472222222222221,
      "grad_norm": 2.1392762660980225,
      "learning_rate": 1.8132716049382716e-05,
      "loss": 0.1295,
      "step": 2425
    },
    {
      "epoch": 13.61111111111111,
      "grad_norm": 1.3661255836486816,
      "learning_rate": 1.7746913580246917e-05,
      "loss": 0.1326,
      "step": 2450
    },
    {
      "epoch": 13.75,
      "grad_norm": 1.6498640775680542,
      "learning_rate": 1.736111111111111e-05,
      "loss": 0.1299,
      "step": 2475
    },
    {
      "epoch": 13.88888888888889,
      "grad_norm": 2.753546714782715,
      "learning_rate": 1.697530864197531e-05,
      "loss": 0.132,
      "step": 2500
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.8270750988142292,
      "eval_loss": 0.5426644682884216,
      "eval_runtime": 71.9593,
      "eval_samples_per_second": 14.063,
      "eval_steps_per_second": 0.445,
      "step": 2520
    },
    {
      "epoch": 14.027777777777779,
      "grad_norm": 1.3266198635101318,
      "learning_rate": 1.6589506172839506e-05,
      "loss": 0.1408,
      "step": 2525
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 1.1106363534927368,
      "learning_rate": 1.6203703703703704e-05,
      "loss": 0.1331,
      "step": 2550
    },
    {
      "epoch": 14.305555555555555,
      "grad_norm": 1.399956226348877,
      "learning_rate": 1.58179012345679e-05,
      "loss": 0.1302,
      "step": 2575
    },
    {
      "epoch": 14.444444444444445,
      "grad_norm": 1.4343998432159424,
      "learning_rate": 1.54320987654321e-05,
      "loss": 0.1309,
      "step": 2600
    },
    {
      "epoch": 14.583333333333334,
      "grad_norm": 3.668091058731079,
      "learning_rate": 1.5046296296296297e-05,
      "loss": 0.1295,
      "step": 2625
    },
    {
      "epoch": 14.722222222222221,
      "grad_norm": 1.215031623840332,
      "learning_rate": 1.4660493827160496e-05,
      "loss": 0.1281,
      "step": 2650
    },
    {
      "epoch": 14.86111111111111,
      "grad_norm": 1.0562926530838013,
      "learning_rate": 1.4274691358024692e-05,
      "loss": 0.1215,
      "step": 2675
    },
    {
      "epoch": 15.0,
      "grad_norm": 5.751645565032959,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.1284,
      "step": 2700
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.825098814229249,
      "eval_loss": 0.5345058441162109,
      "eval_runtime": 73.3445,
      "eval_samples_per_second": 13.798,
      "eval_steps_per_second": 0.436,
      "step": 2700
    },
    {
      "epoch": 15.13888888888889,
      "grad_norm": 2.1824536323547363,
      "learning_rate": 1.3503086419753085e-05,
      "loss": 0.1217,
      "step": 2725
    },
    {
      "epoch": 15.277777777777779,
      "grad_norm": 0.9013480544090271,
      "learning_rate": 1.3117283950617285e-05,
      "loss": 0.1232,
      "step": 2750
    },
    {
      "epoch": 15.416666666666666,
      "grad_norm": 1.235672950744629,
      "learning_rate": 1.2731481481481482e-05,
      "loss": 0.1213,
      "step": 2775
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 1.3273124694824219,
      "learning_rate": 1.2345679012345678e-05,
      "loss": 0.1242,
      "step": 2800
    },
    {
      "epoch": 15.694444444444445,
      "grad_norm": 1.6189485788345337,
      "learning_rate": 1.1959876543209878e-05,
      "loss": 0.1275,
      "step": 2825
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 1.5683832168579102,
      "learning_rate": 1.1574074074074075e-05,
      "loss": 0.126,
      "step": 2850
    },
    {
      "epoch": 15.972222222222221,
      "grad_norm": 1.721937656402588,
      "learning_rate": 1.1188271604938271e-05,
      "loss": 0.1205,
      "step": 2875
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.8349802371541502,
      "eval_loss": 0.5285242795944214,
      "eval_runtime": 73.3035,
      "eval_samples_per_second": 13.806,
      "eval_steps_per_second": 0.437,
      "step": 2880
    },
    {
      "epoch": 16.11111111111111,
      "grad_norm": 0.7611171007156372,
      "learning_rate": 1.0802469135802469e-05,
      "loss": 0.1188,
      "step": 2900
    },
    {
      "epoch": 16.25,
      "grad_norm": 1.2971806526184082,
      "learning_rate": 1.0416666666666668e-05,
      "loss": 0.122,
      "step": 2925
    },
    {
      "epoch": 16.38888888888889,
      "grad_norm": 0.7731566429138184,
      "learning_rate": 1.0030864197530866e-05,
      "loss": 0.1203,
      "step": 2950
    },
    {
      "epoch": 16.52777777777778,
      "grad_norm": 1.075944185256958,
      "learning_rate": 9.645061728395062e-06,
      "loss": 0.1238,
      "step": 2975
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 0.7787513136863708,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.1179,
      "step": 3000
    },
    {
      "epoch": 16.805555555555557,
      "grad_norm": 1.0278524160385132,
      "learning_rate": 8.873456790123458e-06,
      "loss": 0.1152,
      "step": 3025
    },
    {
      "epoch": 16.944444444444443,
      "grad_norm": 0.9401047229766846,
      "learning_rate": 8.487654320987654e-06,
      "loss": 0.1178,
      "step": 3050
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.8300395256916996,
      "eval_loss": 0.535774827003479,
      "eval_runtime": 73.0785,
      "eval_samples_per_second": 13.848,
      "eval_steps_per_second": 0.438,
      "step": 3060
    },
    {
      "epoch": 17.083333333333332,
      "grad_norm": 0.7830631732940674,
      "learning_rate": 8.101851851851852e-06,
      "loss": 0.1165,
      "step": 3075
    },
    {
      "epoch": 17.22222222222222,
      "grad_norm": 0.6156997680664062,
      "learning_rate": 7.71604938271605e-06,
      "loss": 0.1172,
      "step": 3100
    },
    {
      "epoch": 17.36111111111111,
      "grad_norm": 0.7898209691047668,
      "learning_rate": 7.330246913580248e-06,
      "loss": 0.1109,
      "step": 3125
    },
    {
      "epoch": 17.5,
      "grad_norm": 0.995391845703125,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.1184,
      "step": 3150
    },
    {
      "epoch": 17.63888888888889,
      "grad_norm": 1.293843388557434,
      "learning_rate": 6.558641975308642e-06,
      "loss": 0.1128,
      "step": 3175
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 1.179330587387085,
      "learning_rate": 6.172839506172839e-06,
      "loss": 0.1197,
      "step": 3200
    },
    {
      "epoch": 17.916666666666668,
      "grad_norm": 1.5338051319122314,
      "learning_rate": 5.787037037037038e-06,
      "loss": 0.1199,
      "step": 3225
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.8349802371541502,
      "eval_loss": 0.5189934968948364,
      "eval_runtime": 72.4361,
      "eval_samples_per_second": 13.971,
      "eval_steps_per_second": 0.442,
      "step": 3240
    },
    {
      "epoch": 18.055555555555557,
      "grad_norm": 0.6209427118301392,
      "learning_rate": 5.401234567901234e-06,
      "loss": 0.1192,
      "step": 3250
    },
    {
      "epoch": 18.194444444444443,
      "grad_norm": 0.6023913025856018,
      "learning_rate": 5.015432098765433e-06,
      "loss": 0.109,
      "step": 3275
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 1.2056450843811035,
      "learning_rate": 4.6296296296296296e-06,
      "loss": 0.1159,
      "step": 3300
    },
    {
      "epoch": 18.47222222222222,
      "grad_norm": 1.0217233896255493,
      "learning_rate": 4.243827160493827e-06,
      "loss": 0.1126,
      "step": 3325
    },
    {
      "epoch": 18.61111111111111,
      "grad_norm": 0.8297092318534851,
      "learning_rate": 3.858024691358025e-06,
      "loss": 0.1146,
      "step": 3350
    },
    {
      "epoch": 18.75,
      "grad_norm": 0.5455515384674072,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 0.1192,
      "step": 3375
    },
    {
      "epoch": 18.88888888888889,
      "grad_norm": 0.5246239304542542,
      "learning_rate": 3.0864197530864196e-06,
      "loss": 0.112,
      "step": 3400
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.8310276679841897,
      "eval_loss": 0.5260413885116577,
      "eval_runtime": 73.1437,
      "eval_samples_per_second": 13.836,
      "eval_steps_per_second": 0.437,
      "step": 3420
    },
    {
      "epoch": 19.02777777777778,
      "grad_norm": 0.4471115171909332,
      "learning_rate": 2.700617283950617e-06,
      "loss": 0.1185,
      "step": 3425
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 2.6564996242523193,
      "learning_rate": 2.3148148148148148e-06,
      "loss": 0.1196,
      "step": 3450
    },
    {
      "epoch": 19.305555555555557,
      "grad_norm": 0.7345994710922241,
      "learning_rate": 1.9290123456790124e-06,
      "loss": 0.1108,
      "step": 3475
    },
    {
      "epoch": 19.444444444444443,
      "grad_norm": 0.7227876782417297,
      "learning_rate": 1.5432098765432098e-06,
      "loss": 0.113,
      "step": 3500
    },
    {
      "epoch": 19.583333333333332,
      "grad_norm": 0.8343745470046997,
      "learning_rate": 1.1574074074074074e-06,
      "loss": 0.1131,
      "step": 3525
    },
    {
      "epoch": 19.72222222222222,
      "grad_norm": 0.6085208058357239,
      "learning_rate": 7.716049382716049e-07,
      "loss": 0.1139,
      "step": 3550
    },
    {
      "epoch": 19.86111111111111,
      "grad_norm": 0.9748389720916748,
      "learning_rate": 3.8580246913580245e-07,
      "loss": 0.1093,
      "step": 3575
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.821051836013794,
      "learning_rate": 0.0,
      "loss": 0.1087,
      "step": 3600
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.83399209486166,
      "eval_loss": 0.517457902431488,
      "eval_runtime": 71.867,
      "eval_samples_per_second": 14.082,
      "eval_steps_per_second": 0.445,
      "step": 3600
    },
    {
      "epoch": 20.0,
      "step": 3600,
      "total_flos": 0.0,
      "train_loss": 0.2731148091952006,
      "train_runtime": 18326.3929,
      "train_samples_per_second": 6.253,
      "train_steps_per_second": 0.196
    }
  ],
  "logging_steps": 25,
  "max_steps": 3600,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "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": 0.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}