{
  "best_metric": 0.4847160577774048,
  "best_model_checkpoint": "./kd_results/microsoft/beit-base-finetuned-ade-640-640_alpha0.5_temp3.0_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": 8.006985664367676,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 1.2585,
      "step": 25
    },
    {
      "epoch": 0.2777777777777778,
      "grad_norm": 8.391579627990723,
      "learning_rate": 6.944444444444445e-06,
      "loss": 1.2561,
      "step": 50
    },
    {
      "epoch": 0.4166666666666667,
      "grad_norm": 6.191210746765137,
      "learning_rate": 1.0416666666666668e-05,
      "loss": 1.2421,
      "step": 75
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 4.974086284637451,
      "learning_rate": 1.388888888888889e-05,
      "loss": 1.2562,
      "step": 100
    },
    {
      "epoch": 0.6944444444444444,
      "grad_norm": 6.58894681930542,
      "learning_rate": 1.736111111111111e-05,
      "loss": 1.2139,
      "step": 125
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 4.742503643035889,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 1.1966,
      "step": 150
    },
    {
      "epoch": 0.9722222222222222,
      "grad_norm": 9.244735717773438,
      "learning_rate": 2.4305555555555558e-05,
      "loss": 1.2201,
      "step": 175
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.3804347826086957,
      "eval_loss": 1.345258116722107,
      "eval_runtime": 72.7512,
      "eval_samples_per_second": 13.91,
      "eval_steps_per_second": 0.44,
      "step": 180
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 8.967430114746094,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.1103,
      "step": 200
    },
    {
      "epoch": 1.25,
      "grad_norm": 6.795831680297852,
      "learning_rate": 3.125e-05,
      "loss": 1.1327,
      "step": 225
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 4.595554828643799,
      "learning_rate": 3.472222222222222e-05,
      "loss": 1.0515,
      "step": 250
    },
    {
      "epoch": 1.5277777777777777,
      "grad_norm": 7.298379421234131,
      "learning_rate": 3.8194444444444444e-05,
      "loss": 1.0006,
      "step": 275
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 8.414870262145996,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.8868,
      "step": 300
    },
    {
      "epoch": 1.8055555555555556,
      "grad_norm": 8.990382194519043,
      "learning_rate": 4.5138888888888894e-05,
      "loss": 0.7949,
      "step": 325
    },
    {
      "epoch": 1.9444444444444444,
      "grad_norm": 8.477620124816895,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.7455,
      "step": 350
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6897233201581028,
      "eval_loss": 0.8167068362236023,
      "eval_runtime": 70.4671,
      "eval_samples_per_second": 14.361,
      "eval_steps_per_second": 0.454,
      "step": 360
    },
    {
      "epoch": 2.0833333333333335,
      "grad_norm": 16.687789916992188,
      "learning_rate": 4.976851851851852e-05,
      "loss": 0.6573,
      "step": 375
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 9.717782020568848,
      "learning_rate": 4.938271604938271e-05,
      "loss": 0.6293,
      "step": 400
    },
    {
      "epoch": 2.361111111111111,
      "grad_norm": 13.879584312438965,
      "learning_rate": 4.899691358024692e-05,
      "loss": 0.6265,
      "step": 425
    },
    {
      "epoch": 2.5,
      "grad_norm": 8.353231430053711,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.5848,
      "step": 450
    },
    {
      "epoch": 2.638888888888889,
      "grad_norm": 5.697619438171387,
      "learning_rate": 4.8225308641975306e-05,
      "loss": 0.6338,
      "step": 475
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 5.058846950531006,
      "learning_rate": 4.783950617283951e-05,
      "loss": 0.5479,
      "step": 500
    },
    {
      "epoch": 2.9166666666666665,
      "grad_norm": 6.882582664489746,
      "learning_rate": 4.745370370370371e-05,
      "loss": 0.5498,
      "step": 525
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7608695652173914,
      "eval_loss": 0.6834536194801331,
      "eval_runtime": 70.9592,
      "eval_samples_per_second": 14.262,
      "eval_steps_per_second": 0.451,
      "step": 540
    },
    {
      "epoch": 3.0555555555555554,
      "grad_norm": 7.2087321281433105,
      "learning_rate": 4.70679012345679e-05,
      "loss": 0.4998,
      "step": 550
    },
    {
      "epoch": 3.1944444444444446,
      "grad_norm": 5.85562801361084,
      "learning_rate": 4.66820987654321e-05,
      "loss": 0.4621,
      "step": 575
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 5.500523567199707,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.4507,
      "step": 600
    },
    {
      "epoch": 3.4722222222222223,
      "grad_norm": 6.355223178863525,
      "learning_rate": 4.591049382716049e-05,
      "loss": 0.4434,
      "step": 625
    },
    {
      "epoch": 3.611111111111111,
      "grad_norm": 7.778648853302002,
      "learning_rate": 4.5524691358024696e-05,
      "loss": 0.4206,
      "step": 650
    },
    {
      "epoch": 3.75,
      "grad_norm": 5.605661392211914,
      "learning_rate": 4.5138888888888894e-05,
      "loss": 0.4684,
      "step": 675
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 4.532389163970947,
      "learning_rate": 4.4753086419753084e-05,
      "loss": 0.4331,
      "step": 700
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7737154150197628,
      "eval_loss": 0.6307458877563477,
      "eval_runtime": 67.4219,
      "eval_samples_per_second": 15.01,
      "eval_steps_per_second": 0.475,
      "step": 720
    },
    {
      "epoch": 4.027777777777778,
      "grad_norm": 5.660678386688232,
      "learning_rate": 4.436728395061729e-05,
      "loss": 0.4223,
      "step": 725
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 5.2751545906066895,
      "learning_rate": 4.3981481481481486e-05,
      "loss": 0.3667,
      "step": 750
    },
    {
      "epoch": 4.305555555555555,
      "grad_norm": 7.975470066070557,
      "learning_rate": 4.359567901234568e-05,
      "loss": 0.3466,
      "step": 775
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 8.613996505737305,
      "learning_rate": 4.3209876543209875e-05,
      "loss": 0.3781,
      "step": 800
    },
    {
      "epoch": 4.583333333333333,
      "grad_norm": 3.808844804763794,
      "learning_rate": 4.282407407407408e-05,
      "loss": 0.3723,
      "step": 825
    },
    {
      "epoch": 4.722222222222222,
      "grad_norm": 5.349509239196777,
      "learning_rate": 4.243827160493827e-05,
      "loss": 0.3355,
      "step": 850
    },
    {
      "epoch": 4.861111111111111,
      "grad_norm": 7.2955217361450195,
      "learning_rate": 4.205246913580247e-05,
      "loss": 0.3533,
      "step": 875
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.271685600280762,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.3415,
      "step": 900
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.791501976284585,
      "eval_loss": 0.5955695509910583,
      "eval_runtime": 67.3511,
      "eval_samples_per_second": 15.026,
      "eval_steps_per_second": 0.475,
      "step": 900
    },
    {
      "epoch": 5.138888888888889,
      "grad_norm": 5.458718776702881,
      "learning_rate": 4.128086419753087e-05,
      "loss": 0.2986,
      "step": 925
    },
    {
      "epoch": 5.277777777777778,
      "grad_norm": 5.4023823738098145,
      "learning_rate": 4.089506172839506e-05,
      "loss": 0.3153,
      "step": 950
    },
    {
      "epoch": 5.416666666666667,
      "grad_norm": 2.750906229019165,
      "learning_rate": 4.0509259259259265e-05,
      "loss": 0.3,
      "step": 975
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 7.01237154006958,
      "learning_rate": 4.012345679012346e-05,
      "loss": 0.3281,
      "step": 1000
    },
    {
      "epoch": 5.694444444444445,
      "grad_norm": 4.9237799644470215,
      "learning_rate": 3.973765432098765e-05,
      "loss": 0.3275,
      "step": 1025
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 3.391829252243042,
      "learning_rate": 3.935185185185186e-05,
      "loss": 0.3064,
      "step": 1050
    },
    {
      "epoch": 5.972222222222222,
      "grad_norm": 3.3089654445648193,
      "learning_rate": 3.8966049382716055e-05,
      "loss": 0.2533,
      "step": 1075
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8112648221343873,
      "eval_loss": 0.5515549182891846,
      "eval_runtime": 68.4995,
      "eval_samples_per_second": 14.774,
      "eval_steps_per_second": 0.467,
      "step": 1080
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 1.8449082374572754,
      "learning_rate": 3.8580246913580246e-05,
      "loss": 0.272,
      "step": 1100
    },
    {
      "epoch": 6.25,
      "grad_norm": 5.354348659515381,
      "learning_rate": 3.8194444444444444e-05,
      "loss": 0.2491,
      "step": 1125
    },
    {
      "epoch": 6.388888888888889,
      "grad_norm": 6.31616735458374,
      "learning_rate": 3.780864197530865e-05,
      "loss": 0.2449,
      "step": 1150
    },
    {
      "epoch": 6.527777777777778,
      "grad_norm": 4.748219966888428,
      "learning_rate": 3.742283950617284e-05,
      "loss": 0.2413,
      "step": 1175
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 4.347348213195801,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.2921,
      "step": 1200
    },
    {
      "epoch": 6.805555555555555,
      "grad_norm": 3.311044692993164,
      "learning_rate": 3.665123456790124e-05,
      "loss": 0.2591,
      "step": 1225
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 7.541372299194336,
      "learning_rate": 3.626543209876543e-05,
      "loss": 0.2366,
      "step": 1250
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8073122529644269,
      "eval_loss": 0.5602309107780457,
      "eval_runtime": 69.0609,
      "eval_samples_per_second": 14.654,
      "eval_steps_per_second": 0.463,
      "step": 1260
    },
    {
      "epoch": 7.083333333333333,
      "grad_norm": 3.477921724319458,
      "learning_rate": 3.587962962962963e-05,
      "loss": 0.2382,
      "step": 1275
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 1.96200692653656,
      "learning_rate": 3.5493827160493834e-05,
      "loss": 0.2236,
      "step": 1300
    },
    {
      "epoch": 7.361111111111111,
      "grad_norm": 3.339064121246338,
      "learning_rate": 3.5108024691358025e-05,
      "loss": 0.2297,
      "step": 1325
    },
    {
      "epoch": 7.5,
      "grad_norm": 3.174384355545044,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.2297,
      "step": 1350
    },
    {
      "epoch": 7.638888888888889,
      "grad_norm": 7.0932464599609375,
      "learning_rate": 3.4336419753086427e-05,
      "loss": 0.2277,
      "step": 1375
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 1.4575105905532837,
      "learning_rate": 3.395061728395062e-05,
      "loss": 0.2362,
      "step": 1400
    },
    {
      "epoch": 7.916666666666667,
      "grad_norm": 4.844512462615967,
      "learning_rate": 3.3564814814814815e-05,
      "loss": 0.2326,
      "step": 1425
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7865612648221344,
      "eval_loss": 0.6457492709159851,
      "eval_runtime": 68.51,
      "eval_samples_per_second": 14.772,
      "eval_steps_per_second": 0.467,
      "step": 1440
    },
    {
      "epoch": 8.055555555555555,
      "grad_norm": 10.99942398071289,
      "learning_rate": 3.317901234567901e-05,
      "loss": 0.2376,
      "step": 1450
    },
    {
      "epoch": 8.194444444444445,
      "grad_norm": 7.1465678215026855,
      "learning_rate": 3.279320987654321e-05,
      "loss": 0.2091,
      "step": 1475
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 2.6103241443634033,
      "learning_rate": 3.240740740740741e-05,
      "loss": 0.2112,
      "step": 1500
    },
    {
      "epoch": 8.472222222222221,
      "grad_norm": 4.026971817016602,
      "learning_rate": 3.2021604938271605e-05,
      "loss": 0.2121,
      "step": 1525
    },
    {
      "epoch": 8.61111111111111,
      "grad_norm": 8.55809211730957,
      "learning_rate": 3.16358024691358e-05,
      "loss": 0.2091,
      "step": 1550
    },
    {
      "epoch": 8.75,
      "grad_norm": 2.3621342182159424,
      "learning_rate": 3.125e-05,
      "loss": 0.2184,
      "step": 1575
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 2.3842403888702393,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.2331,
      "step": 1600
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8023715415019763,
      "eval_loss": 0.5528243184089661,
      "eval_runtime": 70.6799,
      "eval_samples_per_second": 14.318,
      "eval_steps_per_second": 0.453,
      "step": 1620
    },
    {
      "epoch": 9.027777777777779,
      "grad_norm": 2.858403205871582,
      "learning_rate": 3.04783950617284e-05,
      "loss": 0.2233,
      "step": 1625
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 4.148252487182617,
      "learning_rate": 3.0092592592592593e-05,
      "loss": 0.204,
      "step": 1650
    },
    {
      "epoch": 9.305555555555555,
      "grad_norm": 5.163666725158691,
      "learning_rate": 2.970679012345679e-05,
      "loss": 0.1977,
      "step": 1675
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 1.7793688774108887,
      "learning_rate": 2.9320987654320992e-05,
      "loss": 0.2097,
      "step": 1700
    },
    {
      "epoch": 9.583333333333334,
      "grad_norm": 1.8262099027633667,
      "learning_rate": 2.8935185185185186e-05,
      "loss": 0.208,
      "step": 1725
    },
    {
      "epoch": 9.722222222222221,
      "grad_norm": 1.702004075050354,
      "learning_rate": 2.8549382716049384e-05,
      "loss": 0.2025,
      "step": 1750
    },
    {
      "epoch": 9.86111111111111,
      "grad_norm": 1.6006430387496948,
      "learning_rate": 2.8163580246913578e-05,
      "loss": 0.2023,
      "step": 1775
    },
    {
      "epoch": 10.0,
      "grad_norm": 35.382694244384766,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.2284,
      "step": 1800
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.808300395256917,
      "eval_loss": 0.5509985685348511,
      "eval_runtime": 69.5828,
      "eval_samples_per_second": 14.544,
      "eval_steps_per_second": 0.46,
      "step": 1800
    },
    {
      "epoch": 10.13888888888889,
      "grad_norm": 2.435732126235962,
      "learning_rate": 2.7391975308641977e-05,
      "loss": 0.1829,
      "step": 1825
    },
    {
      "epoch": 10.277777777777779,
      "grad_norm": 2.7984702587127686,
      "learning_rate": 2.700617283950617e-05,
      "loss": 0.1903,
      "step": 1850
    },
    {
      "epoch": 10.416666666666666,
      "grad_norm": 3.7362217903137207,
      "learning_rate": 2.6620370370370372e-05,
      "loss": 0.2002,
      "step": 1875
    },
    {
      "epoch": 10.555555555555555,
      "grad_norm": 1.3439286947250366,
      "learning_rate": 2.623456790123457e-05,
      "loss": 0.1858,
      "step": 1900
    },
    {
      "epoch": 10.694444444444445,
      "grad_norm": 1.9526004791259766,
      "learning_rate": 2.5848765432098764e-05,
      "loss": 0.1972,
      "step": 1925
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 2.4614603519439697,
      "learning_rate": 2.5462962962962965e-05,
      "loss": 0.1842,
      "step": 1950
    },
    {
      "epoch": 10.972222222222221,
      "grad_norm": 0.9845628142356873,
      "learning_rate": 2.5077160493827162e-05,
      "loss": 0.1929,
      "step": 1975
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.8152173913043478,
      "eval_loss": 0.5410348176956177,
      "eval_runtime": 69.3792,
      "eval_samples_per_second": 14.586,
      "eval_steps_per_second": 0.461,
      "step": 1980
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 1.6947256326675415,
      "learning_rate": 2.4691358024691357e-05,
      "loss": 0.1758,
      "step": 2000
    },
    {
      "epoch": 11.25,
      "grad_norm": 2.7373292446136475,
      "learning_rate": 2.4305555555555558e-05,
      "loss": 0.184,
      "step": 2025
    },
    {
      "epoch": 11.38888888888889,
      "grad_norm": 1.56728196144104,
      "learning_rate": 2.3919753086419755e-05,
      "loss": 0.1768,
      "step": 2050
    },
    {
      "epoch": 11.527777777777779,
      "grad_norm": 1.053453803062439,
      "learning_rate": 2.353395061728395e-05,
      "loss": 0.1863,
      "step": 2075
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 3.018230676651001,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.1821,
      "step": 2100
    },
    {
      "epoch": 11.805555555555555,
      "grad_norm": 2.7623562812805176,
      "learning_rate": 2.2762345679012348e-05,
      "loss": 0.1847,
      "step": 2125
    },
    {
      "epoch": 11.944444444444445,
      "grad_norm": 4.333892345428467,
      "learning_rate": 2.2376543209876542e-05,
      "loss": 0.1819,
      "step": 2150
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.825098814229249,
      "eval_loss": 0.5152264833450317,
      "eval_runtime": 69.5095,
      "eval_samples_per_second": 14.559,
      "eval_steps_per_second": 0.46,
      "step": 2160
    },
    {
      "epoch": 12.083333333333334,
      "grad_norm": 2.1153132915496826,
      "learning_rate": 2.1990740740740743e-05,
      "loss": 0.1636,
      "step": 2175
    },
    {
      "epoch": 12.222222222222221,
      "grad_norm": 2.431675434112549,
      "learning_rate": 2.1604938271604937e-05,
      "loss": 0.1649,
      "step": 2200
    },
    {
      "epoch": 12.36111111111111,
      "grad_norm": 1.4594920873641968,
      "learning_rate": 2.1219135802469135e-05,
      "loss": 0.1689,
      "step": 2225
    },
    {
      "epoch": 12.5,
      "grad_norm": 2.117074489593506,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.1861,
      "step": 2250
    },
    {
      "epoch": 12.63888888888889,
      "grad_norm": 4.4751410484313965,
      "learning_rate": 2.044753086419753e-05,
      "loss": 0.1832,
      "step": 2275
    },
    {
      "epoch": 12.777777777777779,
      "grad_norm": 0.9905070662498474,
      "learning_rate": 2.006172839506173e-05,
      "loss": 0.1699,
      "step": 2300
    },
    {
      "epoch": 12.916666666666666,
      "grad_norm": 0.8221396207809448,
      "learning_rate": 1.967592592592593e-05,
      "loss": 0.1677,
      "step": 2325
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.8191699604743083,
      "eval_loss": 0.5095157027244568,
      "eval_runtime": 69.4709,
      "eval_samples_per_second": 14.567,
      "eval_steps_per_second": 0.461,
      "step": 2340
    },
    {
      "epoch": 13.055555555555555,
      "grad_norm": 1.3842869997024536,
      "learning_rate": 1.9290123456790123e-05,
      "loss": 0.1763,
      "step": 2350
    },
    {
      "epoch": 13.194444444444445,
      "grad_norm": 2.07989501953125,
      "learning_rate": 1.8904320987654324e-05,
      "loss": 0.1709,
      "step": 2375
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 1.4017869234085083,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.169,
      "step": 2400
    },
    {
      "epoch": 13.472222222222221,
      "grad_norm": 5.464236736297607,
      "learning_rate": 1.8132716049382716e-05,
      "loss": 0.1588,
      "step": 2425
    },
    {
      "epoch": 13.61111111111111,
      "grad_norm": 0.8512153029441833,
      "learning_rate": 1.7746913580246917e-05,
      "loss": 0.1708,
      "step": 2450
    },
    {
      "epoch": 13.75,
      "grad_norm": 1.689468502998352,
      "learning_rate": 1.736111111111111e-05,
      "loss": 0.1669,
      "step": 2475
    },
    {
      "epoch": 13.88888888888889,
      "grad_norm": 1.8449950218200684,
      "learning_rate": 1.697530864197531e-05,
      "loss": 0.1647,
      "step": 2500
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.8310276679841897,
      "eval_loss": 0.49788549542427063,
      "eval_runtime": 69.8025,
      "eval_samples_per_second": 14.498,
      "eval_steps_per_second": 0.458,
      "step": 2520
    },
    {
      "epoch": 14.027777777777779,
      "grad_norm": 1.2574468851089478,
      "learning_rate": 1.6589506172839506e-05,
      "loss": 0.1898,
      "step": 2525
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 2.1918540000915527,
      "learning_rate": 1.6203703703703704e-05,
      "loss": 0.1738,
      "step": 2550
    },
    {
      "epoch": 14.305555555555555,
      "grad_norm": 1.9546629190444946,
      "learning_rate": 1.58179012345679e-05,
      "loss": 0.1653,
      "step": 2575
    },
    {
      "epoch": 14.444444444444445,
      "grad_norm": 1.4839627742767334,
      "learning_rate": 1.54320987654321e-05,
      "loss": 0.164,
      "step": 2600
    },
    {
      "epoch": 14.583333333333334,
      "grad_norm": 0.7947421073913574,
      "learning_rate": 1.5046296296296297e-05,
      "loss": 0.167,
      "step": 2625
    },
    {
      "epoch": 14.722222222222221,
      "grad_norm": 1.2005423307418823,
      "learning_rate": 1.4660493827160496e-05,
      "loss": 0.1608,
      "step": 2650
    },
    {
      "epoch": 14.86111111111111,
      "grad_norm": 1.8710709810256958,
      "learning_rate": 1.4274691358024692e-05,
      "loss": 0.1552,
      "step": 2675
    },
    {
      "epoch": 15.0,
      "grad_norm": 3.051074504852295,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.1669,
      "step": 2700
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.8369565217391305,
      "eval_loss": 0.5143154263496399,
      "eval_runtime": 69.3605,
      "eval_samples_per_second": 14.59,
      "eval_steps_per_second": 0.461,
      "step": 2700
    },
    {
      "epoch": 15.13888888888889,
      "grad_norm": 2.957630157470703,
      "learning_rate": 1.3503086419753085e-05,
      "loss": 0.1555,
      "step": 2725
    },
    {
      "epoch": 15.277777777777779,
      "grad_norm": 1.1109966039657593,
      "learning_rate": 1.3117283950617285e-05,
      "loss": 0.1583,
      "step": 2750
    },
    {
      "epoch": 15.416666666666666,
      "grad_norm": 1.0647658109664917,
      "learning_rate": 1.2731481481481482e-05,
      "loss": 0.1573,
      "step": 2775
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 1.6117775440216064,
      "learning_rate": 1.2345679012345678e-05,
      "loss": 0.1572,
      "step": 2800
    },
    {
      "epoch": 15.694444444444445,
      "grad_norm": 1.2288424968719482,
      "learning_rate": 1.1959876543209878e-05,
      "loss": 0.1619,
      "step": 2825
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 0.905220627784729,
      "learning_rate": 1.1574074074074075e-05,
      "loss": 0.1592,
      "step": 2850
    },
    {
      "epoch": 15.972222222222221,
      "grad_norm": 1.0033453702926636,
      "learning_rate": 1.1188271604938271e-05,
      "loss": 0.1511,
      "step": 2875
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.8448616600790514,
      "eval_loss": 0.4889645278453827,
      "eval_runtime": 69.6086,
      "eval_samples_per_second": 14.538,
      "eval_steps_per_second": 0.46,
      "step": 2880
    },
    {
      "epoch": 16.11111111111111,
      "grad_norm": 0.687929630279541,
      "learning_rate": 1.0802469135802469e-05,
      "loss": 0.1507,
      "step": 2900
    },
    {
      "epoch": 16.25,
      "grad_norm": 0.8268466591835022,
      "learning_rate": 1.0416666666666668e-05,
      "loss": 0.1562,
      "step": 2925
    },
    {
      "epoch": 16.38888888888889,
      "grad_norm": 0.6281347274780273,
      "learning_rate": 1.0030864197530866e-05,
      "loss": 0.1532,
      "step": 2950
    },
    {
      "epoch": 16.52777777777778,
      "grad_norm": 1.8197768926620483,
      "learning_rate": 9.645061728395062e-06,
      "loss": 0.1617,
      "step": 2975
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 0.7512584924697876,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.1518,
      "step": 3000
    },
    {
      "epoch": 16.805555555555557,
      "grad_norm": 0.9672183394432068,
      "learning_rate": 8.873456790123458e-06,
      "loss": 0.1446,
      "step": 3025
    },
    {
      "epoch": 16.944444444444443,
      "grad_norm": 1.07584810256958,
      "learning_rate": 8.487654320987654e-06,
      "loss": 0.1523,
      "step": 3050
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.8359683794466403,
      "eval_loss": 0.49963295459747314,
      "eval_runtime": 70.4402,
      "eval_samples_per_second": 14.367,
      "eval_steps_per_second": 0.454,
      "step": 3060
    },
    {
      "epoch": 17.083333333333332,
      "grad_norm": 1.854743480682373,
      "learning_rate": 8.101851851851852e-06,
      "loss": 0.1536,
      "step": 3075
    },
    {
      "epoch": 17.22222222222222,
      "grad_norm": 0.8545793294906616,
      "learning_rate": 7.71604938271605e-06,
      "loss": 0.1547,
      "step": 3100
    },
    {
      "epoch": 17.36111111111111,
      "grad_norm": 0.5410196185112,
      "learning_rate": 7.330246913580248e-06,
      "loss": 0.1439,
      "step": 3125
    },
    {
      "epoch": 17.5,
      "grad_norm": 1.009651780128479,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.1544,
      "step": 3150
    },
    {
      "epoch": 17.63888888888889,
      "grad_norm": 2.0547895431518555,
      "learning_rate": 6.558641975308642e-06,
      "loss": 0.1463,
      "step": 3175
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 0.9627209901809692,
      "learning_rate": 6.172839506172839e-06,
      "loss": 0.1545,
      "step": 3200
    },
    {
      "epoch": 17.916666666666668,
      "grad_norm": 2.3797316551208496,
      "learning_rate": 5.787037037037038e-06,
      "loss": 0.1562,
      "step": 3225
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.8349802371541502,
      "eval_loss": 0.4912929832935333,
      "eval_runtime": 71.1493,
      "eval_samples_per_second": 14.224,
      "eval_steps_per_second": 0.45,
      "step": 3240
    },
    {
      "epoch": 18.055555555555557,
      "grad_norm": 0.5895190834999084,
      "learning_rate": 5.401234567901234e-06,
      "loss": 0.154,
      "step": 3250
    },
    {
      "epoch": 18.194444444444443,
      "grad_norm": 0.6346850991249084,
      "learning_rate": 5.015432098765433e-06,
      "loss": 0.1423,
      "step": 3275
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 0.6166771054267883,
      "learning_rate": 4.6296296296296296e-06,
      "loss": 0.1493,
      "step": 3300
    },
    {
      "epoch": 18.47222222222222,
      "grad_norm": 1.0588099956512451,
      "learning_rate": 4.243827160493827e-06,
      "loss": 0.148,
      "step": 3325
    },
    {
      "epoch": 18.61111111111111,
      "grad_norm": 0.5394554138183594,
      "learning_rate": 3.858024691358025e-06,
      "loss": 0.147,
      "step": 3350
    },
    {
      "epoch": 18.75,
      "grad_norm": 0.5485161542892456,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 0.1538,
      "step": 3375
    },
    {
      "epoch": 18.88888888888889,
      "grad_norm": 0.6902341842651367,
      "learning_rate": 3.0864197530864196e-06,
      "loss": 0.1474,
      "step": 3400
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.8379446640316206,
      "eval_loss": 0.48764702677726746,
      "eval_runtime": 73.2016,
      "eval_samples_per_second": 13.825,
      "eval_steps_per_second": 0.437,
      "step": 3420
    },
    {
      "epoch": 19.02777777777778,
      "grad_norm": 0.5021217465400696,
      "learning_rate": 2.700617283950617e-06,
      "loss": 0.1545,
      "step": 3425
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 2.773411750793457,
      "learning_rate": 2.3148148148148148e-06,
      "loss": 0.1544,
      "step": 3450
    },
    {
      "epoch": 19.305555555555557,
      "grad_norm": 0.35924020409584045,
      "learning_rate": 1.9290123456790124e-06,
      "loss": 0.1438,
      "step": 3475
    },
    {
      "epoch": 19.444444444444443,
      "grad_norm": 0.6273934245109558,
      "learning_rate": 1.5432098765432098e-06,
      "loss": 0.1465,
      "step": 3500
    },
    {
      "epoch": 19.583333333333332,
      "grad_norm": 0.9444616436958313,
      "learning_rate": 1.1574074074074074e-06,
      "loss": 0.1473,
      "step": 3525
    },
    {
      "epoch": 19.72222222222222,
      "grad_norm": 0.8144278526306152,
      "learning_rate": 7.716049382716049e-07,
      "loss": 0.148,
      "step": 3550
    },
    {
      "epoch": 19.86111111111111,
      "grad_norm": 1.4011549949645996,
      "learning_rate": 3.8580246913580245e-07,
      "loss": 0.1422,
      "step": 3575
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.8211696147918701,
      "learning_rate": 0.0,
      "loss": 0.1422,
      "step": 3600
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.8379446640316206,
      "eval_loss": 0.4847160577774048,
      "eval_runtime": 69.9873,
      "eval_samples_per_second": 14.46,
      "eval_steps_per_second": 0.457,
      "step": 3600
    },
    {
      "epoch": 20.0,
      "step": 3600,
      "total_flos": 0.0,
      "train_loss": 0.32099378168582915,
      "train_runtime": 18098.2461,
      "train_samples_per_second": 6.332,
      "train_steps_per_second": 0.199
    }
  ],
  "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
}