{
  "best_metric": 0.9356966199505359,
  "best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned-cxr/checkpoint-1026",
  "epoch": 14.912280701754385,
  "eval_steps": 500,
  "global_step": 1275,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12,
      "learning_rate": 3.90625e-06,
      "loss": 0.4698,
      "step": 10
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.8125e-06,
      "loss": 0.4509,
      "step": 20
    },
    {
      "epoch": 0.35,
      "learning_rate": 1.171875e-05,
      "loss": 0.4222,
      "step": 30
    },
    {
      "epoch": 0.47,
      "learning_rate": 1.5625e-05,
      "loss": 0.405,
      "step": 40
    },
    {
      "epoch": 0.58,
      "learning_rate": 1.953125e-05,
      "loss": 0.374,
      "step": 50
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.34375e-05,
      "loss": 0.3357,
      "step": 60
    },
    {
      "epoch": 0.82,
      "learning_rate": 2.734375e-05,
      "loss": 0.3354,
      "step": 70
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.125e-05,
      "loss": 0.2994,
      "step": 80
    },
    {
      "epoch": 0.99,
      "eval_accuracy": 0.8854080791426217,
      "eval_loss": 0.33369317650794983,
      "eval_runtime": 30.5119,
      "eval_samples_per_second": 39.755,
      "eval_steps_per_second": 1.245,
      "step": 85
    },
    {
      "epoch": 1.05,
      "learning_rate": 3.5156250000000004e-05,
      "loss": 0.3264,
      "step": 90
    },
    {
      "epoch": 1.17,
      "learning_rate": 3.90625e-05,
      "loss": 0.2891,
      "step": 100
    },
    {
      "epoch": 1.29,
      "learning_rate": 4.2968750000000004e-05,
      "loss": 0.296,
      "step": 110
    },
    {
      "epoch": 1.4,
      "learning_rate": 4.6875e-05,
      "loss": 0.3368,
      "step": 120
    },
    {
      "epoch": 1.52,
      "learning_rate": 4.9912816041848304e-05,
      "loss": 0.3715,
      "step": 130
    },
    {
      "epoch": 1.64,
      "learning_rate": 4.94768962510898e-05,
      "loss": 0.2945,
      "step": 140
    },
    {
      "epoch": 1.75,
      "learning_rate": 4.90409764603313e-05,
      "loss": 0.3042,
      "step": 150
    },
    {
      "epoch": 1.87,
      "learning_rate": 4.86050566695728e-05,
      "loss": 0.3061,
      "step": 160
    },
    {
      "epoch": 1.99,
      "learning_rate": 4.8169136878814306e-05,
      "loss": 0.2806,
      "step": 170
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9101401483924155,
      "eval_loss": 0.2670154869556427,
      "eval_runtime": 34.1525,
      "eval_samples_per_second": 35.517,
      "eval_steps_per_second": 1.113,
      "step": 171
    },
    {
      "epoch": 2.11,
      "learning_rate": 4.7733217088055796e-05,
      "loss": 0.3049,
      "step": 180
    },
    {
      "epoch": 2.22,
      "learning_rate": 4.72972972972973e-05,
      "loss": 0.2761,
      "step": 190
    },
    {
      "epoch": 2.34,
      "learning_rate": 4.68613775065388e-05,
      "loss": 0.2674,
      "step": 200
    },
    {
      "epoch": 2.46,
      "learning_rate": 4.642545771578029e-05,
      "loss": 0.3055,
      "step": 210
    },
    {
      "epoch": 2.57,
      "learning_rate": 4.59895379250218e-05,
      "loss": 0.2545,
      "step": 220
    },
    {
      "epoch": 2.69,
      "learning_rate": 4.55536181342633e-05,
      "loss": 0.2684,
      "step": 230
    },
    {
      "epoch": 2.81,
      "learning_rate": 4.51176983435048e-05,
      "loss": 0.266,
      "step": 240
    },
    {
      "epoch": 2.92,
      "learning_rate": 4.4681778552746294e-05,
      "loss": 0.2519,
      "step": 250
    },
    {
      "epoch": 2.99,
      "eval_accuracy": 0.9134377576257213,
      "eval_loss": 0.2494814246892929,
      "eval_runtime": 30.5554,
      "eval_samples_per_second": 39.698,
      "eval_steps_per_second": 1.244,
      "step": 256
    },
    {
      "epoch": 3.04,
      "learning_rate": 4.42458587619878e-05,
      "loss": 0.2312,
      "step": 260
    },
    {
      "epoch": 3.16,
      "learning_rate": 4.3809938971229295e-05,
      "loss": 0.2531,
      "step": 270
    },
    {
      "epoch": 3.27,
      "learning_rate": 4.337401918047079e-05,
      "loss": 0.2232,
      "step": 280
    },
    {
      "epoch": 3.39,
      "learning_rate": 4.2938099389712295e-05,
      "loss": 0.2405,
      "step": 290
    },
    {
      "epoch": 3.51,
      "learning_rate": 4.250217959895379e-05,
      "loss": 0.2657,
      "step": 300
    },
    {
      "epoch": 3.63,
      "learning_rate": 4.2066259808195296e-05,
      "loss": 0.233,
      "step": 310
    },
    {
      "epoch": 3.74,
      "learning_rate": 4.163034001743679e-05,
      "loss": 0.2263,
      "step": 320
    },
    {
      "epoch": 3.86,
      "learning_rate": 4.119442022667829e-05,
      "loss": 0.2659,
      "step": 330
    },
    {
      "epoch": 3.98,
      "learning_rate": 4.0758500435919793e-05,
      "loss": 0.2456,
      "step": 340
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9142621599340478,
      "eval_loss": 0.24495702981948853,
      "eval_runtime": 30.579,
      "eval_samples_per_second": 39.668,
      "eval_steps_per_second": 1.243,
      "step": 342
    },
    {
      "epoch": 4.09,
      "learning_rate": 4.032258064516129e-05,
      "loss": 0.2375,
      "step": 350
    },
    {
      "epoch": 4.21,
      "learning_rate": 3.9886660854402794e-05,
      "loss": 0.2188,
      "step": 360
    },
    {
      "epoch": 4.33,
      "learning_rate": 3.945074106364429e-05,
      "loss": 0.2558,
      "step": 370
    },
    {
      "epoch": 4.44,
      "learning_rate": 3.9014821272885795e-05,
      "loss": 0.2009,
      "step": 380
    },
    {
      "epoch": 4.56,
      "learning_rate": 3.857890148212729e-05,
      "loss": 0.222,
      "step": 390
    },
    {
      "epoch": 4.68,
      "learning_rate": 3.814298169136879e-05,
      "loss": 0.2349,
      "step": 400
    },
    {
      "epoch": 4.8,
      "learning_rate": 3.770706190061029e-05,
      "loss": 0.239,
      "step": 410
    },
    {
      "epoch": 4.91,
      "learning_rate": 3.727114210985179e-05,
      "loss": 0.2094,
      "step": 420
    },
    {
      "epoch": 4.99,
      "eval_accuracy": 0.9258037922506183,
      "eval_loss": 0.2105427235364914,
      "eval_runtime": 34.4964,
      "eval_samples_per_second": 35.163,
      "eval_steps_per_second": 1.102,
      "step": 427
    },
    {
      "epoch": 5.03,
      "learning_rate": 3.6835222319093286e-05,
      "loss": 0.2163,
      "step": 430
    },
    {
      "epoch": 5.15,
      "learning_rate": 3.639930252833479e-05,
      "loss": 0.2347,
      "step": 440
    },
    {
      "epoch": 5.26,
      "learning_rate": 3.596338273757629e-05,
      "loss": 0.2055,
      "step": 450
    },
    {
      "epoch": 5.38,
      "learning_rate": 3.552746294681778e-05,
      "loss": 0.2461,
      "step": 460
    },
    {
      "epoch": 5.5,
      "learning_rate": 3.509154315605929e-05,
      "loss": 0.2147,
      "step": 470
    },
    {
      "epoch": 5.61,
      "learning_rate": 3.465562336530079e-05,
      "loss": 0.2168,
      "step": 480
    },
    {
      "epoch": 5.73,
      "learning_rate": 3.421970357454228e-05,
      "loss": 0.2032,
      "step": 490
    },
    {
      "epoch": 5.85,
      "learning_rate": 3.3783783783783784e-05,
      "loss": 0.1962,
      "step": 500
    },
    {
      "epoch": 5.96,
      "learning_rate": 3.334786399302529e-05,
      "loss": 0.1808,
      "step": 510
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9307502061005771,
      "eval_loss": 0.198392853140831,
      "eval_runtime": 30.5883,
      "eval_samples_per_second": 39.656,
      "eval_steps_per_second": 1.242,
      "step": 513
    },
    {
      "epoch": 6.08,
      "learning_rate": 3.2911944202266785e-05,
      "loss": 0.2103,
      "step": 520
    },
    {
      "epoch": 6.2,
      "learning_rate": 3.247602441150828e-05,
      "loss": 0.1569,
      "step": 530
    },
    {
      "epoch": 6.32,
      "learning_rate": 3.2040104620749785e-05,
      "loss": 0.2053,
      "step": 540
    },
    {
      "epoch": 6.43,
      "learning_rate": 3.160418482999128e-05,
      "loss": 0.211,
      "step": 550
    },
    {
      "epoch": 6.55,
      "learning_rate": 3.116826503923278e-05,
      "loss": 0.1825,
      "step": 560
    },
    {
      "epoch": 6.67,
      "learning_rate": 3.073234524847428e-05,
      "loss": 0.1919,
      "step": 570
    },
    {
      "epoch": 6.78,
      "learning_rate": 3.0296425457715783e-05,
      "loss": 0.2144,
      "step": 580
    },
    {
      "epoch": 6.9,
      "learning_rate": 2.9860505666957283e-05,
      "loss": 0.1959,
      "step": 590
    },
    {
      "epoch": 6.99,
      "eval_accuracy": 0.9258037922506183,
      "eval_loss": 0.2021644413471222,
      "eval_runtime": 30.5368,
      "eval_samples_per_second": 39.723,
      "eval_steps_per_second": 1.244,
      "step": 598
    },
    {
      "epoch": 7.02,
      "learning_rate": 2.942458587619878e-05,
      "loss": 0.2051,
      "step": 600
    },
    {
      "epoch": 7.13,
      "learning_rate": 2.898866608544028e-05,
      "loss": 0.1865,
      "step": 610
    },
    {
      "epoch": 7.25,
      "learning_rate": 2.855274629468178e-05,
      "loss": 0.1819,
      "step": 620
    },
    {
      "epoch": 7.37,
      "learning_rate": 2.8116826503923278e-05,
      "loss": 0.2056,
      "step": 630
    },
    {
      "epoch": 7.49,
      "learning_rate": 2.7680906713164778e-05,
      "loss": 0.211,
      "step": 640
    },
    {
      "epoch": 7.6,
      "learning_rate": 2.724498692240628e-05,
      "loss": 0.197,
      "step": 650
    },
    {
      "epoch": 7.72,
      "learning_rate": 2.6809067131647782e-05,
      "loss": 0.185,
      "step": 660
    },
    {
      "epoch": 7.84,
      "learning_rate": 2.6373147340889275e-05,
      "loss": 0.2003,
      "step": 670
    },
    {
      "epoch": 7.95,
      "learning_rate": 2.5937227550130776e-05,
      "loss": 0.179,
      "step": 680
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9299258037922506,
      "eval_loss": 0.1979929804801941,
      "eval_runtime": 40.2803,
      "eval_samples_per_second": 30.114,
      "eval_steps_per_second": 0.943,
      "step": 684
    },
    {
      "epoch": 8.07,
      "learning_rate": 2.550130775937228e-05,
      "loss": 0.1679,
      "step": 690
    },
    {
      "epoch": 8.19,
      "learning_rate": 2.5065387968613773e-05,
      "loss": 0.1573,
      "step": 700
    },
    {
      "epoch": 8.3,
      "learning_rate": 2.4629468177855277e-05,
      "loss": 0.1776,
      "step": 710
    },
    {
      "epoch": 8.42,
      "learning_rate": 2.4193548387096777e-05,
      "loss": 0.1905,
      "step": 720
    },
    {
      "epoch": 8.54,
      "learning_rate": 2.3757628596338274e-05,
      "loss": 0.1734,
      "step": 730
    },
    {
      "epoch": 8.65,
      "learning_rate": 2.3321708805579774e-05,
      "loss": 0.1632,
      "step": 740
    },
    {
      "epoch": 8.77,
      "learning_rate": 2.2885789014821274e-05,
      "loss": 0.1493,
      "step": 750
    },
    {
      "epoch": 8.89,
      "learning_rate": 2.244986922406277e-05,
      "loss": 0.1915,
      "step": 760
    },
    {
      "epoch": 8.99,
      "eval_accuracy": 0.9307502061005771,
      "eval_loss": 0.18886761367321014,
      "eval_runtime": 32.7541,
      "eval_samples_per_second": 37.033,
      "eval_steps_per_second": 1.16,
      "step": 769
    },
    {
      "epoch": 9.01,
      "learning_rate": 2.2013949433304275e-05,
      "loss": 0.1974,
      "step": 770
    },
    {
      "epoch": 9.12,
      "learning_rate": 2.1578029642545772e-05,
      "loss": 0.179,
      "step": 780
    },
    {
      "epoch": 9.24,
      "learning_rate": 2.1142109851787272e-05,
      "loss": 0.1747,
      "step": 790
    },
    {
      "epoch": 9.36,
      "learning_rate": 2.0706190061028772e-05,
      "loss": 0.1752,
      "step": 800
    },
    {
      "epoch": 9.47,
      "learning_rate": 2.0270270270270273e-05,
      "loss": 0.1856,
      "step": 810
    },
    {
      "epoch": 9.59,
      "learning_rate": 1.983435047951177e-05,
      "loss": 0.1695,
      "step": 820
    },
    {
      "epoch": 9.71,
      "learning_rate": 1.939843068875327e-05,
      "loss": 0.1647,
      "step": 830
    },
    {
      "epoch": 9.82,
      "learning_rate": 1.896251089799477e-05,
      "loss": 0.1855,
      "step": 840
    },
    {
      "epoch": 9.94,
      "learning_rate": 1.8526591107236267e-05,
      "loss": 0.1735,
      "step": 850
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.93239901071723,
      "eval_loss": 0.19306738674640656,
      "eval_runtime": 30.5422,
      "eval_samples_per_second": 39.715,
      "eval_steps_per_second": 1.244,
      "step": 855
    },
    {
      "epoch": 10.06,
      "learning_rate": 1.809067131647777e-05,
      "loss": 0.1436,
      "step": 860
    },
    {
      "epoch": 10.18,
      "learning_rate": 1.7654751525719268e-05,
      "loss": 0.1654,
      "step": 870
    },
    {
      "epoch": 10.29,
      "learning_rate": 1.7218831734960768e-05,
      "loss": 0.1469,
      "step": 880
    },
    {
      "epoch": 10.41,
      "learning_rate": 1.6782911944202268e-05,
      "loss": 0.1824,
      "step": 890
    },
    {
      "epoch": 10.53,
      "learning_rate": 1.6346992153443765e-05,
      "loss": 0.159,
      "step": 900
    },
    {
      "epoch": 10.64,
      "learning_rate": 1.591107236268527e-05,
      "loss": 0.162,
      "step": 910
    },
    {
      "epoch": 10.76,
      "learning_rate": 1.5475152571926766e-05,
      "loss": 0.1818,
      "step": 920
    },
    {
      "epoch": 10.88,
      "learning_rate": 1.5039232781168266e-05,
      "loss": 0.1898,
      "step": 930
    },
    {
      "epoch": 10.99,
      "learning_rate": 1.4603312990409764e-05,
      "loss": 0.174,
      "step": 940
    },
    {
      "epoch": 10.99,
      "eval_accuracy": 0.93239901071723,
      "eval_loss": 0.18723934888839722,
      "eval_runtime": 30.6201,
      "eval_samples_per_second": 39.615,
      "eval_steps_per_second": 1.241,
      "step": 940
    },
    {
      "epoch": 11.11,
      "learning_rate": 1.4167393199651266e-05,
      "loss": 0.16,
      "step": 950
    },
    {
      "epoch": 11.23,
      "learning_rate": 1.3731473408892765e-05,
      "loss": 0.1462,
      "step": 960
    },
    {
      "epoch": 11.35,
      "learning_rate": 1.3295553618134264e-05,
      "loss": 0.1281,
      "step": 970
    },
    {
      "epoch": 11.46,
      "learning_rate": 1.2859633827375764e-05,
      "loss": 0.1675,
      "step": 980
    },
    {
      "epoch": 11.58,
      "learning_rate": 1.2423714036617264e-05,
      "loss": 0.1646,
      "step": 990
    },
    {
      "epoch": 11.7,
      "learning_rate": 1.1987794245858763e-05,
      "loss": 0.1656,
      "step": 1000
    },
    {
      "epoch": 11.81,
      "learning_rate": 1.1551874455100261e-05,
      "loss": 0.1761,
      "step": 1010
    },
    {
      "epoch": 11.93,
      "learning_rate": 1.1115954664341762e-05,
      "loss": 0.167,
      "step": 1020
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9356966199505359,
      "eval_loss": 0.17583782970905304,
      "eval_runtime": 30.3282,
      "eval_samples_per_second": 39.996,
      "eval_steps_per_second": 1.253,
      "step": 1026
    },
    {
      "epoch": 12.05,
      "learning_rate": 1.0680034873583262e-05,
      "loss": 0.131,
      "step": 1030
    },
    {
      "epoch": 12.16,
      "learning_rate": 1.024411508282476e-05,
      "loss": 0.1502,
      "step": 1040
    },
    {
      "epoch": 12.28,
      "learning_rate": 9.80819529206626e-06,
      "loss": 0.1945,
      "step": 1050
    },
    {
      "epoch": 12.4,
      "learning_rate": 9.372275501307761e-06,
      "loss": 0.1496,
      "step": 1060
    },
    {
      "epoch": 12.51,
      "learning_rate": 8.93635571054926e-06,
      "loss": 0.1508,
      "step": 1070
    },
    {
      "epoch": 12.63,
      "learning_rate": 8.500435919790758e-06,
      "loss": 0.1346,
      "step": 1080
    },
    {
      "epoch": 12.75,
      "learning_rate": 8.064516129032258e-06,
      "loss": 0.1429,
      "step": 1090
    },
    {
      "epoch": 12.87,
      "learning_rate": 7.628596338273758e-06,
      "loss": 0.1516,
      "step": 1100
    },
    {
      "epoch": 12.98,
      "learning_rate": 7.192676547515258e-06,
      "loss": 0.1408,
      "step": 1110
    },
    {
      "epoch": 12.99,
      "eval_accuracy": 0.9348722176422094,
      "eval_loss": 0.18904681503772736,
      "eval_runtime": 31.9259,
      "eval_samples_per_second": 37.994,
      "eval_steps_per_second": 1.19,
      "step": 1111
    },
    {
      "epoch": 13.1,
      "learning_rate": 6.7567567567567575e-06,
      "loss": 0.1344,
      "step": 1120
    },
    {
      "epoch": 13.22,
      "learning_rate": 6.320836965998257e-06,
      "loss": 0.1284,
      "step": 1130
    },
    {
      "epoch": 13.33,
      "learning_rate": 5.884917175239756e-06,
      "loss": 0.1333,
      "step": 1140
    },
    {
      "epoch": 13.45,
      "learning_rate": 5.448997384481256e-06,
      "loss": 0.1509,
      "step": 1150
    },
    {
      "epoch": 13.57,
      "learning_rate": 5.013077593722755e-06,
      "loss": 0.1516,
      "step": 1160
    },
    {
      "epoch": 13.68,
      "learning_rate": 4.577157802964255e-06,
      "loss": 0.1469,
      "step": 1170
    },
    {
      "epoch": 13.8,
      "learning_rate": 4.141238012205754e-06,
      "loss": 0.1452,
      "step": 1180
    },
    {
      "epoch": 13.92,
      "learning_rate": 3.7053182214472536e-06,
      "loss": 0.1442,
      "step": 1190
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.93239901071723,
      "eval_loss": 0.18487609922885895,
      "eval_runtime": 30.2992,
      "eval_samples_per_second": 40.034,
      "eval_steps_per_second": 1.254,
      "step": 1197
    },
    {
      "epoch": 14.04,
      "learning_rate": 3.2693984306887534e-06,
      "loss": 0.1337,
      "step": 1200
    },
    {
      "epoch": 14.15,
      "learning_rate": 2.8334786399302533e-06,
      "loss": 0.1253,
      "step": 1210
    },
    {
      "epoch": 14.27,
      "learning_rate": 2.3975588491717523e-06,
      "loss": 0.1407,
      "step": 1220
    },
    {
      "epoch": 14.39,
      "learning_rate": 1.961639058413252e-06,
      "loss": 0.1337,
      "step": 1230
    },
    {
      "epoch": 14.5,
      "learning_rate": 1.5257192676547516e-06,
      "loss": 0.1258,
      "step": 1240
    },
    {
      "epoch": 14.62,
      "learning_rate": 1.0897994768962512e-06,
      "loss": 0.163,
      "step": 1250
    },
    {
      "epoch": 14.74,
      "learning_rate": 6.538796861377506e-07,
      "loss": 0.1077,
      "step": 1260
    },
    {
      "epoch": 14.85,
      "learning_rate": 2.179598953792502e-07,
      "loss": 0.1661,
      "step": 1270
    },
    {
      "epoch": 14.91,
      "eval_accuracy": 0.9266281945589447,
      "eval_loss": 0.1879303902387619,
      "eval_runtime": 30.7227,
      "eval_samples_per_second": 39.482,
      "eval_steps_per_second": 1.237,
      "step": 1275
    },
    {
      "epoch": 14.91,
      "step": 1275,
      "total_flos": 1.261425912606462e+19,
      "train_loss": 0.2096748413291632,
      "train_runtime": 6312.1271,
      "train_samples_per_second": 25.936,
      "train_steps_per_second": 0.202
    }
  ],
  "logging_steps": 10,
  "max_steps": 1275,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 500,
  "total_flos": 1.261425912606462e+19,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}